BIOLOGICAL BULLETIN

OF THE

fiOarinc Kiolocjical Xaborator^

WOODS HOLL, MASS.

Editorial Staff

E. G. CONKLIN The University of Pennsylvania.
JACQUES LOEB The University of California.
T. H. MORGAN Columbia University.
W '. M. WHEELER American Museum of Natural

History, New York.

C. O. WHITMAN The University of Chicago.
E. B. WILSON Columbia University.

FRANK R. LILLIE The University of Chicago.

VOLUME XII.

WOODS HOLL, MASS.
DECEMBER, 1906, TO MAY, 1907.

PRESS OF

NEW ERA PRINTING COMPANY
^ANCASTER PA.

CONTENTS OF VOL XII.

No. i. DECEMBER, 1906 P AGE .

T. H. MONTGOMERY, JR.: The Oviposition, Cocooning and Hatch-
ing of an Aranead, Theridinm tepidariorum C. Koch. i
C. V. MORRILL, JR.: Regeneration of Certain Structures in Fun-

dulus heteroclitus 1 1

E. P. LYON : Note on the Geotropism of Arbacia Larva 21

E. P. LYON: Note on the Heliotropism of Palcemonetes Larva... 23
M. LOUISE NICHOLS : Chromosome Relations in the Spermatocytes

of Oniscus 26

C. H. TURNER: A Preliminary Note on Ant Behavior 31

J. F. McCLENDOX : On the Development of Parasitic Copepods

Part I. 37

No. 2. JANUARY, 1907

j; F. McCLENDON : On the Development of Parasitic Copepods.

Part II 53

C. M. CHILD : Studies on the Relation Between Amitosis and

Mitosis. I. Development of the Ovaries and Oogenesis in

Moniezia 89

T. H. MONTGOMERY, JR.: Probable Dimorphism of the Eggs of

an Aranead 115

K. FOOT and E. C. STROBELL : The "Accessory Chromosome"

of Anasa tristis 119 "

No. 3. FEBRUARY, 1907

OILMAN A. DREW : The Habits and Movements of the Razor-she/!

Clam, Ensis directus, Con 127

J. F. McCLENDON : Experiments on the Eggs of Chcetopterus and

Asterias in which the Chroma fin was Removed 141

H. D. SENIOR : The Con/is Arteriosus in Tarpon atlanticits ( Citvier

e^ Valenciennes} X 4^

VERNON L. KELLOGG: Some Silkworm Moth Reflexes 152

EDWIN LINTON : An Abnormal Cestode Proglottid 155

S. J. HOLMES: Observations on the Young of Ranatra quadri-

dentata Stal... i5 8

OSCAR RIDDLE: A Study of the Fundamental Bars in Feathers. 165
C. M. CHILD: Studies on the Relation Between Amitosis and
Mitosis. II. Development of the Testes and Spermato-
gencsis in Moniezia 175

No. 4. MARCH, 1907

C. M. CHILD : Studies on the Relation Between Amitosis and
Mitosis. II. Development of the Testes and Spermato-

genesis in Moniezia 191

GILMAN A. DREW : The Circulatory and Nervous Systems of the
Giant Scallop {Pecten tenuicostatus, Mighels), with Re-
marks on the Possible Ancestry of the Lamellibrancliiata,
and on a Method for Making Series of Anatomical Draw-
ings 225

WILLIAM B. KIRKHAM: The Maturation of the Mouse Egg 259

SHINKISHI HATAI : On the Zoological Position of the Albino Rat. 266

CHAS. W. HARGITT : Notes on the Behavior of Sea-Anemones 274

L. C. SHUDDEMAGEN : On the Anatomy of the Central Nervous
System of the Nine-banded Armadillo ( Tatu novemcinc-
tum Linn.^) 285

No. 5. APRIL, 1907

EDMUND B. WILSON: Note on the Chromosome- Groups of Meta-

podius and Banasa 303

H. H. NEWMAN : Spawning Behavior and Sexual Dimorphism

in Fit ndit Ins heteroclitus and Allied Fish 314

No. 6. May, 1907

CHARLES T. BRUES : Systematic Affinities of the Dipterous Family

Phoridce 349

ALBERT C. EYCLESHYMER : Some Observations and Experiments
on the Natural and Artificial Incubation of the Egg of the
Common Fowl 360

CLARA HEPBURN : A Peculiar Pelvic Attachment in Necturus

maculatits 375

H. D. SENIOR : Note on the Conus Arteriosits of Megalops cypri-

noidcs Broussonct 378 V

YKKNON L. KELLOGG : Sex Differentiation in Larval Insects 380

G. G. SCOTT : Further Notes on the Regeneration of the Fins

of Fu ndit lus heteroclitus 385

Vol. XII. December, 1906. No. i

BIOLOGICAL BULLETIN

THE OVIPOSITION, COCOONING AND HATCHING
OF AN ARANEAD, THERIDIUM TEPIDARIORUM

C. KOCH. 1

THOS. H. MONTGOMERY, JR.

During the course of collecting a series of accurately timed
stages of the eggs of one of our most familiar spiders, the
Theridiuin tcpidariornni, I had occasion to accumulate quite a
number of observations upon the reproductive habits. Pub-
lished accounts of such habits in spiders are very few, there is
much still unknown, so that it seemed worth while to write up
these notes since they are the most extensive yet made upon any
single species.

In an earlier contribution 2 I presented an account of this
species along with others, and described the moulting, copula-
tion, sperm-induction by the male, cocooning and care of the
young. I then described three cases of cocooning, and timed
the sequence of cocoons for eight females.

The present observations were made at Woods Hole, Mass.,
from the fifth to the twenty-sixth of August of the past summer.
An unusually large colony of these spiders was found upon an
old stone wall in a wood, to which they had probably strayed
from some buildings adjacent to one end of the wall. About
the first of August the spiders were beginning to form their first
cocoons, and upon each of most of the webs of females one or
more males were present. Two or three weeks later most of
the males had disappeared. My previous study showed that

1 Contributions from the Zoological Laboratory of the University of Texas, No. 76.

2 "Studies on the Habits of Spiders, particularly those of the Mating Period,"
Proc. Acad. Nat. Sci. Philadelphia, 1903.

2 THOS. H. MONTGOMERY, JR.

there may be frequent copulations before the first act of oviposi-
tion, and even between successive acts. But since the males
begin to disappear before the full series of cocoons is made, and
especially because my captive females produced successive co-
coons without the presence of males, and yet these eggs were
fertile, it is apparent that copulation before the first act of ovipo-
sition suffices for the fertilization of the later sets of eggs. Nearly
150 females were kept in small cages consisting of wooden and
paper boxes covered on one side by glass ; there was very little
mortality among them, and at the end of the month they were
manumitted in good condition.

The processes of oviposition and cocooning were observed in
part or completely in a large number of cases, and may be sum-
marized as follows. The female commences the cocooning by
biting through certain threads in a particular region of the web ;
she then gathers the cut ends together with her legs, thus mak-
ing a free space in which to work and at the same time forming
a composite thread that serves for the initial suspension of the
cocoon as well as of herself during the process of construc-
tion. She spins upon the lower end of the suspension thread,
and combs out the new line into a fluffy ball ; this is the
beginning of the base of the cocoon, and the spider employs for
the combing mainly the third pair of legs, but to some extent
the fourth also. From this point in the operations until the
cocoon is nearly completed the spider hangs to lines of the sur-
rounding web by her first and second pairs of legs, with her
cephalothorax directed vertically and above the abdomen.
When the fluffy fundament of the cocoon base has a diameter
somewhat larger than the length of the cephalothorax of the
spider, the latter rotates this loose textured ball with her palpi
and third pair of legs, and with the fourth pair in alternate appli-
cation draws a thread out from her spinnerets and applies it to the
lower rim of the fluffy ball. She continues this motion until the
base of the cocoon has attained the form of an inverted cup, the
upper convexity of which is attached to the suspensory line.
Towards the close of the base-making her spinning becomes
gradually slower. Then she proceeds to oviposit. She lifts
herself slightly by contracting the anterior pairs of legs, then

COCOONING AND HATCHING OF AN ARANEAD. 3

presses her genital aperture against the concavity of the base ; a
yellow drop of fluid pours out and immediately adheres to the
concavity of the base, the eggs rapidly flowing out into it until
the whole reaches a size quite equal to that of the spider's abdo-
men. The spider touches the surface of the egg mass only with
the lower surface of her abdomen ; during the oviposition there
are repeated pressures of the abdomen against the egg mass.
The yellow fluid is at first of thin consistency and soon dries
upon the eggs ; but in drying it does not glue them together.
Shortly after the discharge of all the ripe eggs the spider by a
few quick contractions detaches her epigynum from the surface
of the egg mass, and immediately starts to make the cover of
the cocoon, spinning first upon the exposed surface of the egg
mass until the latter is covered, then upon the surface of the base
also. From her spinnerets she draws out a continuous, com-
pound thread with the fourth legs used alternately, applying this
thread to the growing cocoon, while at the same time rotating
the cocoon slowly with the third pair of legs (with the occasional
help of the second), the spider still hanging to the web above
the cocoon by the first pair. Thus the eggs become quickly
hidden from sight, by the making of a closely knit cover com-
posed of one continuous thread. Throughout the whole opera-
tion the spider keeps pressing the tips of her palpi against the
cocoon, which gives her a knowledge of the progress of the
progress of the work ; there can be no question that she is
guided entirely by the sense of touch, for during her labor she
is is so placed as to be unable to see the cocoon. Towards the
close of the cover making she spins closer, with the result that
the exterior surface of the cocoon has the firmest texture. The
spinning finished, she feels over the entire surface with her palpi.
Rarely does she leave the cocoon hung upon the suspensory
thread, usually she carries it higher up into the web and attaches
it to the object that roofs the latter.

The finished cocoon varies from yellowish to dark brown in
color, and large ones may reach a diameter of 10 mm. It is
quite smooth externally, pyriform and usually pointed at the
upper end where the suspensory thread was first attached.

More than two hundred cocoons were produced by my spiders,

4 THOS. H. MONTGOMERY, JR.

of which 1 88 were marked as to the exact or approximate time
of oviposition. On a few days I did not commence observations
until -6 a. m., and in such cases estimated the time of oviposition
of prior cocoons by the state of completion of the cover ; but in
the majority of cases the precise time of the end of the act of
oviposition was determined. From the accumulated data the
following results were obtained :

1. Time of Day of Oviposition. - - This is probably always in the
morning and usually in the early morning ; in one case a cocoon
was made some time between 1 1 a. m. and 4 p. m., while I was
absent from the laboratory, and in this single case oviposition
may have occurred later than the noon hour. The next latest
case was one at 11.29 a> m - The following table shows the
hours of oviposition and the number of cocoons formed in each :

Before 5 5-6 6-7 7-8 8-9 9-10 10-11 11-12

43 5i 42 3 2 33 9 5 2

In the laboratory the spiders experienced about the same
light conditions as they did upon the stone wall. Though my
lamp was frequently burning at night in the room, and its light
wakened the spiders to activity, yet it never induced them to
start cocooning. The influence that occasions the cocooning
can therefore not be one of amount of light, nor of sudden
change from darkness to light. If it were amount of light they
should be expected to cocoon at twilight as well as at dawn, yet
they never do so. I do not think that cocooning is stimulated
by any light condition, but that the stimulus to oviposition may
be the oncoming of the warmth of the day after the coolness of
the night. My records of cocoons of one and the same indi-
vidual show that successive cocoons may be formed at different
hours of the day, /. e., that one individual does not maintain a
particular cocooning hour.

2. Time Duration of the Act of Oviposition. - - This was timed
in 43 cases, the interval being measured from the moment of com-
mencing extrusion of the eggs to the moment of commencing
spinning of the cover. In one case it lasted just 2 minutes; in
nine cases from 3 to 4 minutes ; in twenty cases, from 4 to 5
minutes; in five cases, from 5 to 6 minutes; in six cases from

COCOONING AND HATCHING OF AN ARANEAD.

6 to 7 minutes ; in one case, 8^< minutes and in one case, where
the spider seemed feeble, 19 minutes. In the majority of cases
it lasts from 3 ^ to 4^ minutes. There is remarkable uniformity
in the length of duration of the act, which seems to be quite
independent of the number of eggs discharged. After all the
eggs have been extruded but while a point of the egg mass sur-
face still adheres to the genital aperture, the spider may continue
reiterated ineffectual efforts to discharge further eggs, and how
long it continues to do so renders the duration of the process
longer or shorter.

3. Time Duration of the Spinning of the Base.- -This was
timed for seventeen cases, showing a variation from 14 to 37
minutes ; in eleven of these cases it was between 20 and 30
minutes.

4. Time Duration of the Spinning of the Cover. This was ob-
served in thirty-two cases. In one case the process lasted 27
minutes ; in five cases, 40 to 49 minutes ; in eleven cases, 50 to 59
minutes ; in two cases, 60 to 69 minutes ; in three cases, 70 to
79 minutes ; in four cases, So to 89 minutes ; in three cases, 90
to 99 minutes ; in one case, I 10 minutes ; and in one case where
the spider lacked the left fourth leg, for 142 minutes. There is,
accordingly, considerable variation in the duration of this oper-
ation, due not at all to the size of the egg mass, apparently also
not to the rate of spinning, but rather to the thickness and firm-
ness of the cocoon which varies greatly. The cover making is
usually uninterrupted, but sometimes the spider may pause, then
evidently from weariness. It may be that high nourishment
would allow the greatest amount of silk secretion, and that in
well fed individuals the cocoons be largest and take the most
time in the making.

5. Abnormalities in Cocooning.-- Abnormal cocoons are not
rare with wild individuals, and relatively more numerous with my
captive specimens. Generally an abnormal cocoon is due to a
defect in the spinning of the base ; the latter may be too small
or too loose, and the spider in working upon it is then liable to
pull it out into an irregular form. Again, the issuing egg mass
may by accident adhere to a leg of the spider, or even a portion
of it become detached from the rest ; when this happens the spider

6 THOS. H. MONTGOMERY, JR.

may cease to spin, seemingly being cognizant of something out
of the ordinary course. Or the cover may be incompletely made,
leaving the egg mass partially exposed, this is apparently due to
lack of sufficient space for the spinning operations. Once a
mistake is made the spider appears to be unable to rectify it, a
fact that I have remarked in my earlier observations (/. r.) ;
the whole process appears to be strictly instinctive and the
spider seems unable to modify it more than quantitatively, and
unable to learn by experience. An individual may make an im-
perfect cocoon, then construct a following one quite perfect ; or the
reverse may happen. When the cocoon is particularly irregular,
as when the egg mass becomes more or less broken, the spider
may either eat the eggs, or may detach them from the web
and allow them to fall to the ground.

6. Time Intervals between Successive Cocoons Formed by the
same Individual. --Of 113 timed intervals between successive
cocoons raised in my cages, there was one interval of 2 days, 2
of 3 days, 13 of 4 days, 26 of 5 days, 18 of 6 days, 19 of 7 days,
12 of 8 days, 10 of 9 days, 3 of 10 days, 4 of 1 1 days, 4 of I 2
days, and I of 1 3 days. More than half of the intervals ranged
from 5 to 7 days. In all probability the rapidity in the rate of
the succession of cocoons depends upon the degree of nourish-
ment of the spider, because some that I purposely starved
furnished no cocoons at all. My captives were of course not as
well fed as they would have been in the wild state, and though
I kept them fairly well supplied with young grasshoppers and
locustids this by no means equalled their natural diet in either
variety or amount. It is probable that these spiders in a natural
state, with normal feeding, would furnish on the average cocoons
at intervals of from four to six days. The shortest time in
which a succession of 4 cocoons was made by any one of my
spiders was 14 days, comprising one interval of 4 days and two
of 5 days. The largest number of cocoons I have found in any
wild web was eight.

7. HatcJiing of the Young. Thirty-nine cocoons were kept
to determine the time of hatching of the young. Of these, 3
failed to hatch; 3 hatched after an interval of 1 1 days ; 3, of 13
days ; 1 8, of 14 days ; 2, of 15 days ; 9, of 16 days ; and I, of

COCOONING AND HATCHING OF AN ARANEAD. /

17 days. The majority, accordingly, hatched at intervals of from
14 to 1 6 days. Different generations of cocoons do not have
different rates of hatching. Cocoons made on the same day
need not hatch at the same time; thus of 8 cocoons made on Au-
gust 20, 3 hatched in 1 1 days, 2 in 13 days, 2 in 14 days, i in
1 6 days ; and of 1 1 cocoons made on the morning of Au-
gust 26, 9 hatched in 14 days, and 2 in 16 days. All these co-
coons were removed from the web, placed in separate glass vials,
and kept together under the same conditions of light, tempera-
ture and moisture. The differences in the rate of hatching are
probably due to the difficulties experienced by the young in
emerging, for where the cocoon is thinnest the hatching is earliest.
Most of the spiderlings emerge in the early morning, but they
may come out in the late morning and the afternoon. Each
of the few most vigorous spiderlings makes a small circular
aperture through the wall of the cocoon, and through these
few openings all the rest find their way. The young are at first
decidedly positively heliotropic, and it is the light shining into
the first made exits that probably guides the less precocious
individuals out of the cocoon. The weakest may not emerge
until several hours or even days after the most vigorous. In
this species, unlike the lycosids, I have found no evidence that
the mother aids the young to hatch ; for I removed from the
webs a number of cocoons immediately after their completion,
thus precluding any opening of the cocoon by the mother, yet
all of the young made their way out. The mother exhibits
some degree of bravery in guarding the cocoon, especially when
the latter is newly formed, though never to the extent of allow-
ing herself to be injured, and if she is roughly handled she invari-
ably drops from the web.

8. Protective Value of tlie Cocoon. I removed eggs, at various
intervals after oviposition, from cocoons and placed them in flat
glass dishes in the ordinary diffuse light of the laboratory ; in all
these cases normal spiderlings resulted. Therefore the presence
of a cocoon is not necessary to normal development and, further,
its value is probably not to exclude the light. One batch of
eggs was opened onto a dish of water, where they floated ; after
a number of days they become covered by a mould that killed

8 THOS. H. MONTGOMERY, JR.

them. Probably the main value of the cocoon is to protect the
eggs from enemies, for were the eggs not enclosed they would
fall singly or in a mass to the ground, become removed from
the guardianship of the mother, and be subjected to a great
variety of rapacious foes.

9. Relation of tJie Cocoon to the other Forms of Aranead Arclii-
tecturc. --On this important question i will dwell only briefly at
this point, reserving a fuller discussion for later. McCook ! con-
cludes : "The spinning work of spiders may be classified gener-
ally as, first, the Snare, spun for the capture of prey ; second, the
Enswathment, by which insects are disarmed and prepared for
food ; third, the Gossamer, used for purposes of aqueous and aerial
locomotion ; fourth, the Cocoon, spun for the propagation and
protection of the species; and fifth, the Nest, which is a domicile
more or less elaborate and permanent within and under which
the aranead dwells for protection against the exigencies of
weather and the assaults of enemies." Menge 2 had previously
called attention to the web that the males of certain species spin
for the deposition of their sperm, preparatory to charging their
palpi with it ; this has also been described by me (/. <r.), and it
may be called the "sperm-web." Then Wagner' in a most
important analysis of the architecture of spiders directed espe-
cially towards its phylogenetic significance, has distinguished the
web spun by some forms for protection during the moult ; and
then divides what McCook calls the " Nest" into : " La retraite-
construction, destinee pour la danciirc dc r araigncc, ou elle passe
tout son temps et qu'elle ne quitte que pour la chasse. A 1'epoque
de la ponte cette retraite pent servir de loge au cocon." And
into " Le 7/zV/- construction, destinee au sejour de la femelle et a
contenir le cocon." As far as I comprehend this distinction of
Wagner's, the "retraite" is a nest made by both sexes, the " nid"
one built by the female only. It seems to me questionable whether
this distinction of Wagner's is a valid one, and McCook's term
"nest" had best be retained without especial subdivision, though
surely different kinds of nests may be distinguished.

'"American Spiders and their Spinning Work," Vol. I, Philadelphia, 1889.
2 " Ueber die Lebensweise der Arachniden," A'ensste Sc/ir. nntitrf. Ges. Danzig,
4, 1843.

3 "L'Industrie des Araneina," Mem. Acad. Sd. Si. Pelcnboni^ (VII.), T. 42.

COCOONING AND HATCHING OF AN AKANEAD. Q

Therefore we have to enumerate: the cocoon, snare, nest, para-
chute (a somewhat more definite term to replace McCook's "gos-
samer "), sperm-web, e/isivat/iincnt and moult-ivcb. Of these
constructions the cocoon occurs in all species, and probably the
nest also ; these may then be said to be the most conservative
or most persistent kinds of architecture in spiders. It is probable
that the nest or protective habitation was the first construction
to be elaborated, perhaps first merely some cranny that later
became lined with silk by the spider ; and that the spider laid its
eggs first without lining against the wall of this nest, then later
spun upon them a cover of silk. This is of course merely an
inference because we know nothing of the ancestors of the modern
spiders. But the cocoon, by which is meant the immediate silken
envelope of the eggs, may in this way be explained as a deriva-
tion of the nest of which it formed at first an integral part ; the
original base of the cocoon would have been the silken investure
of the nest, the cover of the cocoon a later addition. If this
suggestion be correct, then the phylogeny of the cocooning
would have been as follows: (i) oviposition against the wall of
the nest ; (2) the addition of a cover to the eggs ; (3) a special
base spun upon the wall of the nest, the eggs upon which re-
ceived a cover (a condition realized by certain attids, clubionids
and drassids) ; (4) the formation of a cocoon with base and cover
apart from the nest.

Wagner has distinguished two kinds of cocoons : those made
of two pieces, and those of a single piece, a distinction based
upon the appearance of the completed structure. I would main-
tain, on the contrary, that it is probable that in all modern ara-
neads the cocoon is spun of two pieces, first a base, then a cover,
foj- I have observed (/. c.} this process in lycosids, agalenids,
dictynids, theridiids, epeirids and drassids. One may not judge
from the form of the finished cocoon as to its mode of construc-
tion, and, so far as I know, in all cases where the process has
been followed a base is first spun, the eggs oviposited upon it,
then a cover made. There seems to be no positive evidence that
the cocoon is ever made in a single piece. One of the best diag-
nostic characters of Araneads is the making by the mother of a
silken covering, composed of two pieces, for the protective invest-
ment of the eggs.

IO THOS. H. MONTGOMERY, JR.

10. Parasitic Hymeiwpicra in the Eggs of TJieridiuni. In
none of the cocoons raised by my captive spiders were there any
parasites, but about one out of every thirty or forty wild cocoons
was found to be infected with a tiny hymenopteron, black in
color, the males winged and the females wingless, of a length
somewhat less than the diameter of an egg of the spider. Prof.
Wm. H. Ashmead has kindly identified these as Proctrotrypids,
constituting a new species (montgomeryi) of the genus BCEUS. The
males hatch some hours or a day in advance of the females, and
die soon after the momentary copulation with the latter. The
wasps generally hatch out from the cocoons at about the same
time the spiders hatch from the uninfected eggs.

Three cocoons, from four to six hours old, were placed in a
dish containing wasps that had hatched from other cocoons one
or two days previously. The female wasps walked over the
surface of the cocoons, continually tapping them with their
antennae ; most of the females oviposited into the cover of the
cocoons, or at least frequently projected their ovipositors into
them. A smaller number made their way into the interior of
the cocoons, taking about ten minutes to bore, mainly by rota-
tion of the large and sharp-rimmed head, a small circular aper-
ture into the cocoon. After several of the female wasps had
entered in this way I cautiously opened a cocoon and under the
compound microscope could observe the wasps piercing the
spider eggs with their long, needle-like ovipositors ; they seem
not to succeed in piercing every egg that they attempt. A
single wasp inoculates in this way a considerable number of
spider eggs. Only one wasp emerges from any one spider egg,
so it is probable that the wasp places only one egg in a spider
egg, but I hope to decide this point by a careful study of the
infected eggs. Rarely if ever are all the eggs in a cocoon so
infected ; the infected ones after three or four days turn a dull
brown color. If an infected cocoon is kept until hatching in a
closed vial, so that neither the emerging spiders nor the hatching
wasps can escape, the spiderlings that have escaped infection
ultimately catch and eat the adult wasps.

REGENERATION OF CERTAIN STRUCTURES IN
FUNDULUS HETEROCLITUS.

C. V. MORRILL, JR.

It is a fact of common knowledge to all observers of fish kept
in aquaria, that portions of the fins, lost through accident, or pur-
posely removed will be replaced by new structures which cannot
be distinguished, in many cases, from normal fins. This ability
to regenerate has been examined by Duncker (1905), Morgan
(1900, 1902), Nussbaum and Sidoriak ( 1900), Mazza (1890) and
Broussonet (1786). The following experiments were made to
test the effect of cutting off the fins close to the body. The
regenerative powers of the operculum, the lower jaw, the scales
and the lens were also tested.

The various sets of experiments described below were tried on
different fish, no one fish having more than one fin removed,
except in the case of pectorals and pelvics which were removed
from the same fish, the right pectoral and left pelvic being always
selected for convenience.

I wish to express my thanks to Prof. T. H. Morgan, who
kindly suggested this problem and under whose direction it was
completed. My thanks are also due to the staff of the New York
Aquarium, and to Dr. F. B. Sumner, director of the U. S. Fishe-
ries Laboratory at Wood's Hole, Mass., who kindly placed the
necessary tanks and fish at my disposal.

DORSAL FIN.

On October 10, 1905, eighteen fish had the dorsal fin cut oft
as close to the body as possible. These were examined from
time to time until December 29, 1905, when due to various
causes, principally fungal growth, the number had become
reduced to thirteen. All of these with one exception were
regenerating a new fin. The excepted case was kept under ob-
servation three months longer but at the end of that time,
twenty-six weeks after operation, it had developed only a small

1 1

12 C. V. MORRILL, JR.

pointed knob in the middle of the old scar. The remaining
twelve were not kept after December 29 (twelve weeks after
operation). In all of these the regeneration was proceeding
rapidly at this time, some having the new fin half completed.

On February 16, 1906, a second set of twenty-five fish had
the dorsal fin cut off close to the body. These regenerated more
rapidly than the first set and in five weeks' time, twenty-one of
these which still remained, were regenerating along the entire
length of the cut surface, in many cases the new growth being
one eighth of an inch in height. Three weeks later, fifteen only
were still living ; the increase in height of the fins was distinctly
visible in this short time, though more so in some individuals
than in others. Three weeks later (at the time of writing) and
twenty-three weeks after operation, fourteen were still alive.
The new fins in some cases were regenerating evenly and pos-
sessed the normal number of fin rays (eleven). But in other
cases, the new growth appeared along part of the cut surface
only. An examination of the anatomy of the fin and the exact
direction of the cut reveals the cause of this irregularity. Usu-
ally in trimming the fin close to the body, the dermal fin-rays
are severed a short distance distal to their articulation with the
bony supporting rays, beneath the surface. Sometimes, how-
ever, if the flesh at the sides of the fin is pressed down by the
scissors and the fin at the same time held firmly, the cut will
pass, in some places, through the enlarged proximal part of the
dermal fin rays and regeneration in consequence will be much
slower, or may not take place at all in parts so affected. The
cut may even at times remove the dermal rays completely.
These facts, I think, will account for cases in which regeneration
is irregular, i. e., when the new structure does not contain the
normal number of fin-rays. Thus complete or nearly complete
removal of the fin-rays prevents regeneration of the fin.

ANAL FIN.

The anal fins of eighteen fish were cut off close to the body
on October 25, 1905. The wounds healed up rather quickly
but showed no signs of regeneration until two months later,
when of the thirteen fish remaining, five showed slight signs of

REGENERATION IN FUNDULUS HETEROCLITUS. I 3

new growth. Six weeks later the regeneration was fairly started
in five fish which still survived. These five were examined one
week later, i. c., sixteen weeks after removal of the fin and it was
found that three showed slight regeneration while the other two
had a new fin about half the size of the original one. In these
last, the fin was normal in shape and position, though small, and
the rays parallel to one another as in the old fin. Ten weeks
later, twenty-one weeks after removal of original fin, the regen-
eration was almost complete, the new fin having the normal
number of fin-rays and being scarcely distinguishable from ordi-
nary anal fins.

Here as in the case of the dorsal fin, the cut usually passes a
short distance distal to the articulation of the dermal with the
bony fin rays. But it may pass a little further in through the
articulations if not carefully made, as the body is soft in the
region of this fin. A cut of this kind results, as might be ex-
pected, in no regeneration. The cut may, however, pass through
the articulations of some rays and the shafts of others ; then the
new growth takes place from the cut ends of the latter only,
producing an irregular structure with less than the normal num-
ber of fin-rays.

The external prolongation of the oviduct in the females, which
passes down the anterior end of the fin is, of course, severed
with the fin. It regenerates, as might be expected, only when
the anterior fin-ray which supports it, regenerates, not otherwise.

CAUDAL FIN.

On October 5, 1905, the caudal fins were removed from
eighteen fish, the cut being made in a curve so as to remove the
entire fin, without at the same time cutting off any of the tail
vertebrae proper. This operation is rather severe ; there is
usually some bleeding and the edges do not close over the
wound very readily. In consequence the number of fish lost is
usually considerable. After twelve weeks, five fish only re-
mained, all showing regeneration of the fin to some extent. Two
weeks later the regeneration was still proceeding slowly but
somewhat irregularly, the new fin-rays being bent downward,
forming a considerable angle with the original horizontal direc-

14 C. V. MORR1LL, JR.

tion of the rays, and the rays themselves were fewer in number
than in normal fins. This result was due to very close cut-
ting ; entire rays were removed in some places, and in others,
where stumps remained an alteration in direction of these stumps
occurred in some way when the wound healed over. In one
case the new fin grew straight out in the normal direction but at
different rates of growth in different parts so that it presented a
finger-like appearance. This abnormality was probably due to
a jagged cut, the fin-rays being disarticulated in some places
while in others, the stumps were left in situ. In still another
fish no regeneration took place at all although the wound healed
over smoothly. This fish was kept under observation four
months. Here the cut probably either disarticulated the rays or
left too small a stump for regrowth to take place.

On January 9, 1906, the tails of twenty-four fish were cut off
in such a way as to remove a portion of the body of the fish,
including several vertebrae. Obviously this operation was ex-
tremely severe. These fish when returned to water were unable
to assume a horizontal position but remained head downward at
the top of the tank. Three days later eleven were still living.
These latter had succeeded finally in reaching the bottom of the
tank and some were swimming about in a nearly normal position.
At this time, January 12, another set of twenty-four fish, cut as
described above, were put into the same tank with the eleven
just mentioned. Three days later only seventeen were left of the
entire number, thirty-five. This number gradually decreased
until the entire lot died within a month's time.

Another set of twenty-four were cut in the same way but it
was found impossible to keep them alive. The wounds remained
open and the flesh rotted away from the sides of the vertebrae.

In the above sets, various sizes of fish were used. In still
another set ten small fish were employed and -kept in shallow
dishes in the laboratory in order that they might assume a hori-
zontal position. Within a month all of these had died with one
exception. This might have been due to unfavorable conditions
in part, though I have kept fish living in these shallow dishes
over three months. In the exception noted, the wound had
healed over smoothly and a very small white knob of tissue ap-

REGENERATION IN FUNDULUS HETEROCLITUS. I 5

peared over the cut end of the spinal cord. This fish was kept
under observation for three months but during this time the
white knob did not increase perceptibly in size. As far as external
examination could determine no regeneration of bone took place.

PECTORAL AND PELVIC FINS.

On October 25, 1905, twenty-four fish had their right pectoral
and left pelvic fins removed. The cuts were made as close to
the body as possible. In three months all died but three. These
showed no regeneration of the pectoral fin but in one case a slight
indication of a regenerating pelvic. Two weeks later two of these
were still living, both showing the regenerating pelvics well under
way and a well-defined indication of regenerating pectorals.
Unfortunately the two fish died a few days later. Another set
of twelve fish were cut in the same way but all died within a
month, showing no signs of regeneration.

A third set of twenty-five fish cut as above, were started
on February 16, 1906. Nine weeks later eighteen were still
living ; five of these showed regenerating pelvics and one, an
indication of regenerating pectoral ; the rest no regeneration
of either fin. Two weeks later (at the time of writing) seventeen
were living. Two of these showed regenerating pectorals dis-
tinctly, in the shape of a small knob of tissue which no doubt
contained two or three regenerating rays, since the fish moved
the structure to and fro like a fin. Eight of this same lot were
regenerating their missing pelvic fins, the new structure in some
cases being one eighth of an inch long.

An examination of the direction of the cut in the removal of
the pectoral fins shows that in most cases all the dermal fin rays
are removed entire, i. e., disarticulated. Occasionally, however,
the cut does not completely remove the uppermost one or two
rays. These rays extend inward a little further than the others
and so their proximal stumps may remain in situ and give rise to
little pointed structures as described above, but never to a com-
plete fin. In the case of the pelvic fins, it is not possible as a
rule to cut them so close in ; thus it more frequently happens
that stumps of the rays are left, sometimes a few, sometimes all.
Regeneration takes place from the cut ends of these only.

1 6 C. V. MORRILL, JR.

From the foregoing facts, the lack of regeneration of the
paired fins in so many cases is seen to be due to complete
removal of the dermal fin rays. These results fall in line with
those described above for the unpaired fins.

OPERCULUM.

On January 15, 1906, a set of six fish had a piece of the
operculum removed, in order to determine whether the actual
bony structures of the fish would regenerate. The piece re-
moved was about one eighth of an inch broad and half an inch
long, and was taken from the lower posterior border. Six weeks
later two of these fish remained, but showed no signs of regen-
eration. A few days later, these also died. One fish with
operculum cut as above was kept in a dish in the laboratory for
four months, but at the end of this time showed no signs of re-
generation.

The same experiment was tried with two sets of fifteen fish
each at Wood's Hole. These fish were kept from June 20,
1906, to August 25, 1906, but at the end of this period showed
no signs of regeneration.

Fish operated on in the above manner are sometimes difficult
to keep alive since the gills are at first exposed to attacks of
parasites and the nibbling of other fish. In a short time, how-
ever, the branchiostegal membrane spreads over and partly closes
the opening.

LOWER JAW.

On March 23, 1906, pieces about one quarter of an inch long,
were removed from the right-hand side of the lower jaws of
thirty-six fish. Six weeks later, twenty-eight were still living.
The jaws had grown again to such an extent that some of them
could scarcely be distinguished from normal ones. Whether
new bone took the place of that removed, was not accurately
determined, though microscopic examination of sections of three
of these new structures seemed to indicate that cartilage at least
had regenerated.

Attempts were also made to remove larger portions of the
jaw, or the entire jaw, but these always resulted in the death of
the fish even if the wound healed. This may have been due to

REGENERATION IN FUNDULUS HETEROCLITUS. I/

the difficulty the fish experienced in grasping and holding its
food.

SCALES.

On June 20, 1906, the scales were removed from definite areas
of a number of fish. This operation was performed in such a
way as to injure as little as possible the underlying dermis, each
scale being separately pulled out with a pair of fine forceps.
Three sets of fifteen fish each were used in this experiment, each
set of fish having the scales removed from a different part of the
body. In three weeks time the new scales were coming in
rapidly, some of them being one third the normal size, though
very thin and soft. Two weeks later, the new scales were still
only one-half to three quarters the length of a fully grown scale,
though many were as broad. It appears then, that regenerating
scales grow faster in breadth than in length. At this period or
a little earlier the characteristic dotted pigment appears. At
first the dots are comparatively few in number but continue to
increase until the pigmentation, normal, for the position of the
scale is attained. Thus each scale, from whatever region it is
taken, is replaced by a new one pigmented exactly like the
original. The fish in this experiment were kept ten weeks in all
and at the end of that period the regenerated scales were prac-
tically normal in size, thickness and color though sometimes
slightly irregular in shape.

LENS.

On June 21, 1906, the lens was removed from one eye of
a number of fish. The operation was performed by slitting the
cornea with a sharp scalpel and then gently pressing the eyeball
with forceps until the lens popped out. In all thirty-six fish were
used which were divided into two sets of eighteen each. In one
of these sets the iris was slit in addition. These fish were kept
until August 29, about ten weeks, when eleven fish remained in
one set and nine in the other. When the lens is squeezed out
some of the vitreous humor goes with it and after the operation
the eyeball appears flattened. However the cornea heals over
very quickly and the eye soon assumes its regular contour, differ-
ing only from normal apparently, in a slight dilation of the pupil.

1 8 C. V. MORRILL, JR.

The lost vitreous humor is apparently, made good in some way.
An examination of the eye at the end of the experiment (after
ten weeks) revealed not the slightest indication of a new lens.
Regeneration of the lens in Triton and Salainaiidra, has been
obtained by Colucci ('91), Wolff ('95) and Fischel ('98). These
results cannot then be extended to the teleost, Fundiilus lictero-
clitns. However, it still remains to be shown whether the re-
generative powers of the eye are not better developed in other
species of teleosts.

SUMMARY AND CONCLUSIONS.

The general conclusions to be drawn from the foregoing ex-
periments are the following :

1. In Fnnditlus lictcroclitus the dorsal, caudal, anal, pectoral
and pelvic fins \vill regenerate even when cut off close to the
body, provided the proximal ends of the dermal fin-rays remain
in situ. No regeneration takes place except when the stumps
are present, and of sufficient size.

2. The operculum does not regenerate.

3. The lower jaw completes itself if a piece is removed.

4. The scales, if carefully removed, are quickly replaced by
others.

5. The eye does not regenerate a new lens.

The results on the regeneration of fins support in general
Broussonet's earlier work and that of subsequent writers. The
ability to regenerate a portion of the jaw has been found by
Spallanzani (1768) and others in the salamander. The lack of
ability to regenerate a new lens in Fnnduhis, however is peculiar,
since the salamander, an animal of a higher order, possesses it to
a marked degree. It would be interesting in this connection to
test the regenerative powers of the eyes of elasmobranches and
ganoids, as being more generalized types than the teleosts.

In regard to Broussonet's statement that those fins which are
the most useful, regenerate the quickest, it seems more probable
that the conditions of the cut, as described for each experiment
above, are the factors which determine the rate of regeneration,
given equal conditions of temperature, food, vitality, etc. Indeed
it is difficult to estimate the relative rate of growth, for in any

REGENERATION IN FUNDULUS HETEROCLITUS. 19

one set of fish kept in the same tank and having the same fins
removed, the differences in the rate of renewal are so great that
some of the new fins may have completed half their growth
when others are just beginning to show definite regeneration.
Again, the caudal fin though obviously the most important in
many fish does not regenerate, perceptibly, faster than the others.
Finally, leaving aside the caudal it is difficult to say which fin is
the most important. The observations which Osburn ('06) has
recently made on the functions of fins seem to indicate that the
pectorals are more useful than the dorsals and anals while my
own results show that the former regenerate much slower.
Broussonet was probably led to his statement concerning the
relative rate of regeneration in different fins by his acceptance of
the idea of regeneration as an adaptation. A priori, the caudal
fin would regenerate, the most rapidly, on the hypothesis, since
it is the most useful. The facts as I have tried to show, do not
bear this out.

At this point the question naturally arises as to the relation
between regeneration and liability to injury. It is true that the
fins, extending out from the body as they do, are subjected to
injury through brushing against obstructions, attacks of preda-
tory animals, etc. ; therefore, one might say, they acquired the
power to regenerate as an adaptation to these conditions. But
why, on the adaptation hypothesis, should the jaw regenerate
so readily, when it is extremely improbable that this structure
would ever be lost in a natural state ? From the lack of regen-
erative power in the operculum and eye no conclusions can be
drawn for or against this hypothesis, since these structures are
not especially liable to injury under normal conditions. The
power to regenerate scales, however, seems truly adaptive.

ZOOLOGICAL LABORATORY, COLUMBIA UNIVERSITY,
October 6, 1906.

LITERATURE CITED.
Broussonet, M.

1786 Observations sur la regeneration de quelques parties du corps des Poissons.

Hist, de 1'Acad. Roy. des Sciences, 1786.
Colucci, V. S.

'91 Sulla regenerazione parziale dell' occhio nei Tritoni. Mem. della R.
Accad. delle Scienze dell' 1st. di Bologna (Ser. V.), I.

2O C. V. MORRILL, JR.

Duncker, G.

'05 t'ber Regeneration des Schwanzendes bei Syngnathiden. Arch. f. Entw.-

mech., XX., I.
Fischel, A.

'98 Uber die Regeneration der Linse. Anat. Anz., XIV., 13.
Mazza, F.

'90 Sulla regenerazione della pinna caudale in alcuni Pesci. Atti Soc. Ligust.

Sc., n., I.
Morgan, T. H.

'oo Regeneration in Teleosts. Arch. f. Entw.-mech., X., I.

'02 Further experiments on the Regeneration of the Tail of Fishes. Ibid.,

XIV., 3 u. 4.
Nussbaum and Sidoriak.

'oo Beitrage zur Kenntnis der Regenerations-vorgange nach kiinstlichen Ver-
telzungen bei alteren Bachforellenembryonen (Sal/no fario L.). Arch, f
Entw.-mech., X., 4.
Osburn, R. C.

'06 The Functions of the Fins of Fishes. Science, N. S., Vol. XXIII., No.

589-
Spallanzani, L.

1768 Prodromo di un'opera da imprimersi sopra la reproduzioni animali. Mo-
dena, 1768. Cited from the French translation : Programme ou precis
d'un ouvrage sur les reproductions animales ; traduit de I' Italian par M.
B. . . . de la Sabionne, Geneve, 1768.
Wolff, G.

'93 Entwickelungsphysiologische Studien. I. Die Regeneration der Urode-
lenlinse, Arch. f. Entw.-mech., L, 3.

NOTE ON THE GEOTROPISM OF ARBACIA LARVAE.

E. P. LYON.

As soon as the blastulae of the sea urchin begin to swim, they
come to the top of the water. They remain in this locality dur-
ing the blastula and gastrula stages but become scattered when
they reach the pluteus condition.

The gathering at the top is not a light effect, for it takes place
in total darkness.

The gathering is not due to lack of oxygen, for the larvae go
up in an inverted test-tube filled with sea water and remain at
the top a long time. With the dense gathering of thousands
of blastulae at the top of such a tube the gradient of increasing
oxygen supply must be from above downward, yet the larvae do
not go down. A narrow tube closed at one end was nearly
filled with water containing young arbacia. An air bubble was
introduced into the tube about the middle of its length. The
tube was then placed with the closed end directed up. The
larvae went up in both ends of the tube, those above the bubble
going directly away from the oxygen supply.

Experiments in tubes kept at constant temperature showed
that the gathering was not due to currents in the water. Nor
was the gathering due to the larvae being lighter than sea water.
Their specific gravity, on the contrary, is considerably higher,
being about 1.060.

That the gathering of larvae at the top is a true gravity effect
is indicated by their moving against a centrifugal force. The
best way to demonstrate this is by placing sea water containing
many larvae in rather long (say, 25 cm.), narrow tubes. On
rotating these carefully, it will be found that beyond a certain
radius (depending on the rate of rotation) the larvae are precipi-
tated, the acceleration in these parts of the tube being greater
than they can swim against. Inside this critical circle the larvae
will be found to have moved toward the axis of rotation and
to be gathered in dense crowds.

21

22 E. P. LYON.

Paramoecia will rise to the top of a solution of the same den-
sity as themselves and in which they have, therefore, no weight.
This fact, to my mind, is proof that gravity acts within the organ-
ism and not through the relative densities of water and organism.
With the larvae of arbacia, however, I was unable to get any
definite gatherings in a solution of gum arabic and sea water of
the same density as the larvae. The viscosity of this medium is
so great that such minute ciliated organisms as those under con-
sideration can hardly move through it.

While convinced, therefore, that the gathering of the arbacia
larvae at the top of the water is a gravity effect, I am unable at
the present time to say how gravity brings about the response.
The possibility that the orientation is purely passive, a buoyancy
effect, must be kept in mind, as well as the various theories pro-
posed by Jensen and Davenport. However, on the basis of my
paramoecium work, I consider it most probable that gravity acts
directly on the cells.

The blastulae from centrifugalized eggs have all the pigment
on one side and are therefore presumably heavier on that side.
Later examination shows that this pigment may be in any part
of the gastrulae or plutei. Such larvae, although, as stated, pre-
sumably unbalanced, nevertheless come to the top of the water.
I consider this observation opposed to the theory of orientation
by buoyancy, for if the heavier portion were always directed
down and this portion sustained no relation to the direction of
locomotion, the organisms ought to be scattered through the
water.

NOTE ON THE HELIOTROPISM OF PALyE-
MONETES LARV/E.

F. P. LYON.

The striking peculiarity of the light reaction of these larvae
is that they orient themselves with the anterior end away from
the light and then swim backward toward the light. If orienta-
tion alone is the criterion of a tropism, these organisms are nega-
tively heliotropic. But they gather on the side of the disk
toward the light and as ordinarily observed would surely be con-
sidered positive.

Blue light produces these effects as strongly as daylight.
Red light is almost ineffective.

The orientation of the larvae is usually extremely precise.
The instant that light is admitted to a dish containing the larvae,
they all turn so that their heads are directed away from the
source of light and the body axis is in the line of the rays. In
many the abdomen is strongly extended so that the body axis is
a straight line. These swim backward practically in a straight
line. Very frequently they swim on their backs in a perfectly
straight course to the light side of the dish. Others swim with
ventral side directed down.

Besides the individuals which hold their bodies extended and
exhibit very accurate orientation there are many which swim
with the abdomen more or less flexed ventrally. These are
inclined to rotate, but their general course is quite exact toward
the light.

A study of the water currents, when India ink was added,
showed that the backward movement was the result of strong
forward beats of the swimmerets. The details have not been
studied. The animals can move forward, but this form of loco-
motion is quite different from the backward type, being accom-
plished by a series of very rapid jumps or plunges. So quick
were these forward movements that I could not accurately ob-
serve how they were accomplished. I believe that they were

23

24 E. P. LYON.

due to strong extension of the tail, the animal first bending the
body ventrally and then suddenly extending it. The jumping
reaction occurred on strong stimulation of any kind : for example,
on throwing light suddenly into the dish. The orientation with
the head away from the light was accomplished by one or more
leaps, but so accurately that one could hardly speak of trial and
error. When the animals in swimming backward toward the
light suddenly hit any obstruction, the springing reaction carried
them momentarily away from the light. But they immediately
began, under the influence of the swimmerets, to move back-
ward toward the light. So it would often happen that on reach-
ing the side of the vessel the larvae would strike, jump away,
swim back, jump away again, and so on many times without
losing their orientation in the direction of the light rays.

It was therefore of some interest to know how these larvae
would behave if the sense of response to light could be reversed.
I have made only a beginning of this study.

The effects of raising the temperature were not clear. For
the most part the leaping reaction predominated. The animals
darted rapidly about and lost their orientation to light.

Better results were obtained by diluting the sea water. If the
larvae were placed in a mixture of 50 c.c. distilled water and 100
c.c. sea water, a considerable proportion of them went to the
negative side of the dish. This they did without changing their
orientation. In other words they now employed the forward
instead of the backward type of locomotion and thus arrived at
the side of the dish away from the light. Often there seemed to
be a conflict between the two methods of locomotion, the animals
oscillating back and forth like shuttles, but always with the heads
away from the light. Equal quantities of distilled and sea water
gave even more pronounced results.

The responses were not always so regular and machine-like as
I have described. Some of the more pronounced "negative"
individuals moved by an irregular, mostly sidewise motion away
from the light. Taken as a whole the " negative " effect secured
by diluting the sea water was less precise and definite than the
" positive " reaction. Also most of the "negative" individuals
were only temporarily so.

NOTE ON THE HELIOTKOPISM OF PAL/EMONETES. 25

On returning negative individuals to the undiluted sea water
they at once swam to the light side of the dish, but without
changing their orientation, i. e., with head directed away from
the light.

The question of nomenclature which these observations sug-
gest may well await further investigations. The experiments
here recorded were made at the Marine Biological Laboratory at
Woods Hole.

PHYSIOLOGICAL LABORATORY,
ST. Louis UNIVERSITY.

CHROMOSOME RELATIONS IN THE SPERMA-
TOCYTES OF ONISCUS.

M. LOUISE NICHOLS.

Recent work on the germ cells of animals has shown that the
chromosomes contained withia one nucleus frequently are not all
of the same size. It has been known for some time that in insects
and spiders one or more of the chromosomes, the so-called acces-
sory, is smaller than the others and also differs in its behavior.
\Yith regard to chromosomes other than the accessory, Sutton, 1
in his studies on the spermatogenesis of Brachystola discovered a
well-marked difference in size. Montgomery's ~ studies of the
chromosomes of various Hemipteran species also revealed a
variation in size of chromosomes occupying the same nucleus,
and McClung, 3 in a recent paper on the spermatocytes of Orthop-
tera has described a remarkable structure, strikingly larger than
the other chromosomes, which he has named " multiple chromo-
some," and which he interprets as being formed by the associa-
tion of a number of chromosomes. The still more recent work
of Morgan 4 and Stevens 5 on aphids indicates much the same
thing. Morgan studied the eggs of Phylloxerans and shows in his
figures Q{ Phylloxera caryce-globnli a distinct difference in the size
of the chromosomes. Miss Stevens investigated a number of
species of plant lice and concludes that each species is character-
ized by a definite number of chromosomes.

Chromosomes of the same nucleus may differ not only in size
but also in shape. In a paper published in 1902 fi I showed that

'W. S. Sutton, "On the Morphology of the Chromosome Group in Brachystola
ma^na," BIOL. BULL., 4, No. i, 1902.

2 T. H. Montgomery, Jr., "A Study of the Germ Cells in Metazoa," Trans. Am.
Phil. Soc., 20, 1901.

3 C. E. McClung, "The Chromosome Complex of Orthopteran Spermatocytes,"
BIOL. BULL., 9, No. 5, 1905.

4 T. H. Morgan, " Male and Female Eggs of Phylloxerans of the Hickories,'*
BIOL. BULL., 10, No. 5, 1906.

5 N. M. Stevens, " Germ Cells of Aphids," Carnegie Inst., May, 1906.

6 M. L. Nichols, "Spermatogenesis of Onistiis asellits, Linn.," Proc. Am. Phil.
Soc., 41, No. 1 68, 1902.

26

CHROMOSOME RELATIONS IN ONISCUS. 2?

in the equatorial plate of the first maturation division of the male
reproductive cells of the wood-louse, Oniscus, there occur chro-
mosomes of three kinds. Some are straight or dumb-bell shaped,
others are crescent shaped. In these two forms the component
parts of the bivalent chromosomes lie end to end. A third form
exists in which the univalent chromosomes are joined not end to
end, but side by side. In each type a split may be observed run-
ning the length of the chromosomes and the first division is there-
fore reducing. In the prophase the third type is represented
usually by rings, occasionally by V-shaped structures (Figs. 7
and 14), while the crescents appear as curved rods and the re-
mainder as straight or dumb-bell shaped rods. The reduced
number of chromosomes, as far as could be determined, is sixteen.
Of this number two have the ring form, two are crescents and
the others are straight or dumb-bell shaped. The chromosomes
vary somewhat in size, but in Oniscns the differences are not so
great as in some of the insects before mentioned. The largest
have the ring form, the smallest are straight (Figs. 1,2,4, ! 7, !&)
Sometimes one of the two rings is complete or nearly so, the
other incomplete (Figs. I, 11). The rings lie oh opposite sides
of the nucleus, sometimes directly opposite, in other cases ap-
parently shifted to one side or the other (Figs. I, 2, 3, 8, 9, 1 1,
15, 1 6). Since they are the largest and most easily recognized of
the nuclear elements they form convenient points of localization.

The crescent form is the next most easily recognizable. Of
this type there are likewise two, one of them sometimes more
strongly curved than the other. They also occupy opposite
sides of the nucleus and at points between the positions occupied
by the rings (Figs. 3, 4, 8, 9, 10, II, 15, 16, 19).

The relative position of the rings and crescents is more readily
determined than that of the rods, because they are not so numer-
ous and are also distinguished by their shape. Repeatedly how-
ever I have observed an arrangement of three or four straight
chromosomes in a row (Figs. 9, 12, 17). One of the smaller of
the straight chromosomes, too, is often seen lying not far from
one of the crescents (Figs. 8, 13, 17, 18).

These facts point to an individuality of the chromosomes and
also indicate a tendency to localization in the nucleus. The evi-

28 M. LOUISE NICHOLS.

dence for this is further strengthened by the manner in which
the chromosomes develop from the resting nucleus. The resting
nucleus is a network formed by the gradual lengthening and an-
astomosis of threads which retain their individuality until just
before the formation of the nuclear membrane. From it arise the
chromosomes of the first maturation division by a breaking apart
of the threads of this network and their gradual shortening and
thickening.

The question presents itself whether these relations of form and
position of the chromosomes are peculiar to Oniscns or character-
istic of the land isopods as a group. Two other genera I have
examined, Porcellio and Armadillo, The spermatogenesis of these
is similar to that of Oniscns, but the cells are smaller, in Armadillo
so small as to make an examination of the chromosomes very
difficult and uncertain.

The cells of Porcellio more nearly approach those of Oniscns
in size and I was able to compare the prophase of the first division
in the two animals. Although the chromosomes of Porcellio are
smaller than those of Oniscns, the peculiarities of shape and
arrangement are similar. In Porcellio also are present two rings
lying opposite each other with a crescent between. More than
one crescent in a nucleus, however, I have not seen. One at least
of the straight chromosomes appears considerably smaller than
the others and, as in Oniscns, frequently lies not far from the
crescent (Figs. 5 and 6). I was not able to determine the num-
ber of chromosomes in Porcellio, as my preparations contained no
spermatocytic mitoses. Besides the smaller size of the chro-
mosomes would make an accurate determination even more diffi-
cult than in Oniscns. As far as can be judged from the appearance
of the prophase, the number would seem not to be very different
in the two genera.

The presence of chromosomes of the three forms described
within the same germ nucleus is not then peculiar to Oniscns, but
probably is characteristic of the group, the genera being visibly
distinguished from each other by differences in size of the chromo-
somes and possibly by slight differences in the shape of homol-
ogous chromosomes.

Comparing the results obtained for various animals and plants,

CHROMOSOME RELATIONS IN ONISCUS. 29

it seems safe to conclude that any correlation existing between
the shape of the chromosomes and the tangible bodily character-
istics is not very close nor exact. Species quite distinct from each
other may possess chromosomes of almost the same shape and size.
Especially is this true of -plant species. On the other hand,
species closely resembling each other may have chromosomes
of perceptibly different shape or size. This is even more obvi-
ously the case with the number of chromosomes. In some
instances a difference in number may be accompanied by differ-
ence in sex, 1 in others by specific differences. 2 Again the same
number may characterize groups as widely different as angio-
sperms, mammals and insects 3 ; while related species of the same
genus may differ in number.

Nevertheless in a general way there appears to be some cor-
respondence between the bodily attributes of a group and the
character and behavior of the chromosomes of that group.
Chromosomes of a particular shape are more characteristic of
certain groups than of others. Thus, the form in which the
limbs of the bivalent chromosomes are twined about each other
is frequently found in flowering plants and also in amphibia, the
ring form is particularly apt to occur in arthropoda and a shape
somewhat like a shepherd's crook, with a slender body and thick-
ened, curved extremity, is often seen in mollusks and worms.
Again the peculiar chromosome known as the accessory so far
has not been found outside of the Tracheata.

No doubt the shape, possibly also the number of chromo-
somes, is determined by complex physical and chemical condi-
tions which may not have any direct relation to visible bodily
characters and two chromosomes of the same shape and size
may contain potentialities of quite different nature, just as an
apple and its waxen imitation, although of the same size and
shape, have very diverse properties.

A parallel condition may be seen in the relation of the shape
and size of the egg to the bodily characters of the organism
developing from it. Forms differing as widely as fish and

1 E. B. Wilson, "The Sexual Differences of the Chromosome Groups in
Hemiptera," Jour. Exp. Zoo!., III., No. I, 1906.
2 Montgomery, Morgan, Stevens, loc, cit,
3 E. B. Wilson, "The Cell in Development and Inheritance," p. 206, igoo.

3O M. LOUISE NICHOLS.

echinoderms may have eggs of about the same size and shape,
but of varying potentialities. Yet here, too, there are peculiar-
ities of constitution and behavior characteristic of certain groups
rather than of others. There may be as good ground for deter-
mining the relationship of organisms by the behavior of chromo-
somes as by the character of cleavage in the egg, but in the
former case the handicap is severe because of the minute size 01
the elements and the fact that changes of important nature may
take place in the constitution of the chromosome without pro-
ducing any effect tangible for modern microscopic methods. If,
however, specific or other differences happen to coincide with
visible variations in the chromosomes, such objects form con-
venient points of departure for the study of such problems as the
determination of sex and the relation of paternal and maternal
chromosomes in the cells of hybrid offspring.

IO

20

EXPLANATION OF FIGURES.

With the exception of 5 and 6, the figures are prophases of the first maturation
division of Oniscus asellus Linn. Nos. 5 and 6 are from Porcellio. The drawings
were made with the camera lucida, Zeiss oc. 6, obj. T V imm. at the level of the table.

A PRELIMINARY NOTE ON ANT BEHAVIOR.

C. H. TURNER.

There is a wide-spread belief among living entomologists that
ants slavishly follow the odor trail and Bethe insists that ants are
mere reflex machines. For something over three years I have
been conducting a series of experiments on the behavior of ants,
the results of which do not harmonize with the above view.
These experiments, which were performed upon a large number
of southern species and a few northern species, show conclusively :

1. That the movements of worker ants are not controlled by
tropisms.

2. That ants are not guided by a homing instinct.

3. That ants are enabled to take long trips by learning, by ex-
perience, the way and retaining what they have thus acquired.
They learn the way in the same manner that vertebrates learn to
open problem boxes and to run mazes.

4. That ants do not as slavishly follow the odor trail as is
supposed.

5. That Sir John Lubbock was right when he said: "In de-
termining their course ants are greatly influenced by the direction
of the light."

6. The color of the pathway has little or no effect upon the
home-going of ants.

7. That Dr. Wheeler is right in stressing the high develop-
ment of the female ; for the winged females often take part in
the regular duties of the nest. I have had them learn the way
home from new situations and assist the workers in carrying the
pupae home.

8. That the males are stupid and seem unable to solve even
the simplest of problems. They seem to be more or less
heliotropic.

9. The major workers of PJieidole, which Ernest Andre claims
function as soldiers and do not take any active part in the ordin-
ary work of the nest, frequently assist the workers in making

32 C. H. TURNER.

excavations ; and, occasionally, assist in conveying pupae from
one place to another. I have never noticed one continue to carry
pupae for any considerable length of time.

10. Ants are affected by olfactory, optic, tactile, kinesthetic
and probably auditory stimuli.

11. Ants seem to have fairly definite impressions of direction,
both in horizontal and vertical planes, and of distance.

12. Ants have associative memory.

13. Such cases of division of labor as Romanes -- quoting
from Moggridge, Lespes, Belt and Herr Gredler - - describes in
his "Animal Intelligence," are to be looked upon as cases of
incidental coincidence rather than as examples of mutual
cooperation.

14. Ants can be trained to do certain simple things in the
same way that vertebrates are trained.

15. In their activities ants display marked individual variations
In this preliminary paper, space forbids a discussion of the

data upon which the above statements are based ; that will be
done in a paper which I expect to publish in the spring or sum-
mer of 1907. In this communication I merely desire to give
some of the evidence that warrants the fourth and fifth state-
ments.

In these experiments I used a card-board stage six inches
square, from which a card-board incline, three fourths of an inch
wide, led down to the island upon which the nest of the ants
was confined. A great many pupae and worker ants were placed
on this stage. After many random movements the ants learned
the way from the stage to the nest and back. The way once
learned, the ants continued to work until all of the pupae had been
carried into the nest. I then replaced the ants and pupae on the
stage. The pupae were again carried home. This was repeated
over and over again ; until, by their actions, I concluded that the
ants were thoroughly acquainted with the way down the incline
to the nest. Then a second incline was so placed as to lead from
the opposite side of the stage to the Lubbock island. If, after a
lapse of a few minutes, no workers descended this second in-
cline, I concluded that the workers thoroughly knew the way
down the other path. I then placed the first incline, which had

A PRELIMINARY NOTE ON ANT BEHAVIOR. 33

become scented by the foot-prints of the ants, in the place where
the new incline had been and placed a new unscented incline in
the place formerly occupied by the one the ants had been tra-
versing. Thus there was an unscented path in the position of
the old trail and the old familiar scented path was in a new
position.

Now if ants are guided home solely by the sense of smell, they
should either have spent approximately as much time learning
the way down the new incline as they did before; or else, in their
random movements, they should have happened upon the scented
incline and gone down it. In reality, they did neither of these
things. They went almost immediately down the unscented in-
cline which occupied the position in space formerly filled by the
incline down which they had been carrying pupa; to the nest.

Similar experiments, yielding the same results, were conducted
with marked individual ants.

It will be seen at once that these experiments disprove not only
Bethe's double polarized trail hypothesis but also Wasmann's
assumption that their odor tracks have for ants an odor-shape
which guides them home.

According to the current belief, ants going to and from my
stages should have followed the same path. But experiments,
conducted with marked individual ants, showed that this was not
always the case. I found :

1. That the ant had to learn the way, not only from the stage
to the island, but also from the nest back again to the stage and
that it usually required more time to solve the second problem
than it did to solve the first.

2. That sometimes an ant would regularly descend to the
island along the top side of the incline and ascend to the stage
along the under side of the same incline.

3. I have noticed a worker regularly descend from the stage
by way of the incline and ascend the stage by way of its central
support.

4. I have had ants ascend to the stage by way of one incline
and descend by the other.

5. In passing from the nest to the foot of the incline the ant
did not always follow the same path.

34 C. H. TURNER.

6. I have known an ant to regularly drop from the stage with
a pupa and carry it to the nest and then return to the island,
mount a section lifter, which I presented, and rest quietly thereon
until it had been conveyed to the stage. Then it would step off,
pick up a pupa and drop from the stage. This was repeated over
and over again by the ant.

More than a quarter of a century ago Lubbock claimed that
ants were influenced by the direction of the rays of light.
These experiments have been either overlooked or ignored by
recent continental writers. Although Sir John Lubbock's experi-
ments were not so planned as to exclude the possibility of the
effect noticed having been produced by heat or change in the_ in-
tensity of light, yet it is hard to see any reason why they should
have been so completely ignored by European writers. My ex-
periments on the effect of light upon the home-going of ants,
which were conducted under perfectly planned conditions of con-
trol, resemble very much certain of Lubbock's experiments.

Two kinds of tests were conducted : experiments with ants
working in concert, and experiments with marked individual ants
working alone.

In experiments of the first type a card-board stage was used,
from which a card-board incline led down to the island. A
1 6 c. p. incandescent lamp was placed near the side to which the
inclined plane was attached. After the ants had thoroughly
learned the way home, a new incline was attached to the stage on
the side opposite the one to which the old incline was still attached.
If, after a lapse of five minutes, no ants went down the new in-
cline, conditions were considered right for the test. The light
was then transferred to the opposite side of the stage. In each
test the halting movements of the ants showed that they were
much disturbed. In most cases, some of the ants would finally
go down the new incline ; and in a few cases, after the lapse of
several minutes, all of the ants went down the new incline. The
above described experiments with ants working in concert were
conducted with several different species.

The individual experiments to test the same thing were con-
ducted with Formica fusca var. snbscricea Say and Myrmica punc-
tiventris Rog. The apparatus was arranged in the same manner

A PRELIMINARY NOTE ON ANT BEHAVIOR. 35

as for ants working in concert ; but, marked ants were used and
only one was allowed to work at a time. In all essential points,
the results harmonized with those obtained from ants acting in
concert. In most cases, however, after a greater or lesser lapse
of time, the ant would usually find its way down the old incline
to the nest ; and, after a greater lapse of time, find its way back
again to the stage.

That the above effect was not due to heat was proven in the
following manner. At the right and left of the stage used was
placed a tall museum jaw .34 cm. x 16 cm. x I/ cm., filled with
distilled water. These heat filters were so adjusted as to have
their broad sides parallel to the edges of the stage. At the be-
ginning of the experiment, a 32 c. p. incandescent lamp was
placed behind one of these filters (usually behind the one that
was near the incline). After the ants had traveled the path long
enough to make the trips regularly and rapidly, the lamp was
transferred to behind the opposite filter. In every case the
workers were much disturbed in the manner stated above. This
was repeated with seven different colonies of Formica fusca. Since
the heat had been excluded, it is evident that the disturbance
was the result of some form of light stimulus.

To determine whether the intensity or the direction of light
was the determining factor, the experiment was slightly modi-
fied. The stage and incline were arranged as before. Some-
times the heat filters were used, but more often they were not.
To furnish illumination, four different candle powers (4 c. p., 8
c. p., 16 c. p., 32 c. p.) of incandescent lamps were used, one at a
time, in a darkened room. At the beginning of the experiment
a lamp of a certain candle-power was placed near the side of the
stage to which the incline was attached. After the ants had
thoroughly learned the way home a different candle-power was
substituted for the first. After the lapse of a few more minutes
this lamp was transferred to near the opposite side of the stage.
Shortly it was returned to its former position. A few minutes
later a different candle-power was substituted for this. This
performance was repeated over and over again until each candle-
power had been used one or more times.

It was found that substituting a lamp of one candle-power for

36 C. H. TURNER.

a lamp of a different candle-power had no disturbing effect upon
the actions of the ants ; but that by any marked change in the
angular position of the light, no matter what the candle-power,
the ants were much disturbed. This forces us to the conclusion
that the direction of the rays of light, when present, plays a
prominent role in the home-going of ants.

ZOOLOGICAL LABORATORY,
UNIVERSITY OF CHICAGO,
November 10, 1906.

ON THE DEVELOPMENT OF PARASITIC

COPEPODS. 1

PART I.

J. F. McCLENDON.

CONTENTS.

I . I n trod action 37

II. Historical 39

III. Anatomy of the reproductive organs 39

IV. Oogenesis and maturation 42

1 . Oogenesis 42

A . The dichel estid 42

B. Lamargus muricaftis 43

C. Pa n darns sinuatus 44

2. Maturation 44

A. Panda rtts sinuatus 44

B. Lmmargus muricatus 45

C. The dichelestid 45

V. Spermatogenesis 47

VI. Polarity of the egg 49

VII. Summary 50

VIII. Bibliography (see part II. of this paper) 76

IX. Explanation of plate 50

x

The work herein presented was begun at Wood's Hole, 1904,
with a view of determining the effects of pressure on the eggs of
parasitic copepods. Owing to technical difficulties, three sum-
mers have passed with few experimental results, and in conse-
quence the present paper is chiefly of a morphological nature.

I obtained parasitic copepods from fish caught in traps of the
Marine Biological Laboratory and the U. S. Fish Commission,
at Woods Hole. It was usually necessary to go to the trap to
get them fresh, or in the case of some Caligidae to catch them
before they left the fish.

Most of the work was done on Pandants sinnatus Verrill,
Lceniargits muricatus Kroyer (collected by Dr. Conklin), and a
new species (text Fig. 3) of the family Dichelesteidse and of a new

1 Thesis accepted by the Faculty of the Department of Philosophy of the Uni-
versity of Pennsylvania toward the degree of Doctor of Philosophy.

37

38 J. F. McCLENDON.

genus near Kroycria (according to Prof. Charles B. Wilson)
which I will call throughout this paper the dichelestid. The
eggs of Pandarns simtatus are the most difficult to handle, being
very flat and thin and pigmented, but I used them because this
is the most common species at Woods Hole. I used the eggs
of the following for comparison (arranged according to the classi-
fication of Claus) :
Caligidae :

Caligns bonito Wilson (and several undetermined species of

C aligns).

Caligns rapax Milne Edwards.
Lepioptlierins edivardsi Wilson.
Perissopns coimnnnis Rathbun.
Nesippns alatns Wilson.
Cccrops lair cillii Leach.
Lcemargus T muricatus Kroyer (and an undetermined species

of Lcemargus).

Philorthagoriscus scrratus Kroyer.
Pandarns simtatus Verrill.
Dichelesthiidae :

Anthosoma crassnm Abilguard.
Kroycria ?

Eudactylina nigra Wilson and E. sp.?
Lernaeidae :

Penella.

Chondracanthidae :
Chondr acanthus.
Sphyrion ?

For the determination of species I am indebted to Prof. Charles
Branch Wilson, of Massachusetts State Normal School, West-
field, Mass., who has been very kind in his interest in my work.
After my work was nearly finished I received the material col-
lected at Woods Hole in 1899 by Prof. Edw. Rynearson, of
Pittsburgh, with a view of working on this same subject, and
which he kindly turned over to me. I wish to express my thanks
for this abundance of material, which allowed me to confirm

'The word Liemargus was applied to a copepod and a shark in the year 1837
and as yet it is disputed which should claim priority.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 39

many points on which my material was scanty and to add some
new ones.

I am indebted to Dr. Conklin for constant guidance and assis-
tance besides the general direction of my work, and to Mr.
Kribs for the use of Zeiss apochromatic lenses.

Dr. Formad and Dr. Fischelis rendered me invaluable service
in translating Russian.

I am under obligations to the Carnegie Institute for a table in
the Marine Biological Laboratory, Wood's Hole, 1904, and to
the U. S. Fish Commission for a table in the U. S. F. C. Labora-
tory, Wood's Hole, Mass., 1905-6.

The material was fixed in Flemming's fluid, picro-acetic, or cor-
rosive-sublimate-acetic. The first gave the best fixation but the
second was the most convenient. Heidenhain's iron haematoxylin
followed by various counter-stains, and Hermann's safranin-
gentian violet were used most frequently.

II. HISTORICAL.

The ground covered in this paper was almost entirely un-
touched by earlier workers as regards the particular genera treated,
and I will consider here only some papers on related forms.

The free living copepods have long been favorite objects
for study. Gruber ('79) described their reproductive organs.
Weismann and Ischikavva ('88 a, b] and Ischikawa ('92) used
them in studying the question of reduction in number of chromo-
somes. Haecker ('91, '92 a, b, '94 a, b, '95 a, b, '97, '02) found
them favorable objects for study of oogenesis, maturation, reduc-
tion, and unsymmetrical mitoses. Later, Ruckert ('94, '95 a, b,
c) studied oogenesis, maturation, and reduction in these.

Of the parasitic and half parasitic copepods, Heider ('79) notes
the spermatogenesis of Lernantkropus, and Giesbrecht ('82) the
oogenesis of the Notodelphidae.

Charles Branch Wilson in his monograph on the Caligidae
('05) describes the anatomy of the reproductive organs and the life
history.

III. ANATOMY OF THE REPRODUCTIVE ORGANS.

In Paiuiants, Caligns and allied forms the two ovaries lie in
the head (Fig. i) and the two oviducts run backwards to the

J. F. McCLENDON.

genital segment, where they are much convoluted and increase
greatly in diameter and each communicates by an opening, the

os uteri, with an egg string that trails
behind the animal. The ovary is formed
of a much convoluted cord of cells in a
single linear series, small at the oogonial
end and gradually increasing in diameter
until it passes into the oviduct. As these
cells (oocytes) grow in the ovary they
become much compressed in the direc-
tion of the long axis of the cord and the
cell boundaries almost completely dis-
appear. Passing into the oviduct, the
oocytes continue to grow and soon their
boundaries reappear, and they become
more flattened as they grow larger.
There are at least two broods in the ovi-
duct at once, caused by the periodic
activity of the ovary. The younger
brood extends some distance into the
genital segment and is sharply marked
off from the older brood, which occupies
the last coils of the oviduct and consists
of oocytes that have nearly or quite
completed their growth. As the oocytes
pass out through the os uteri they are
fertilized from the seminal receptacles
(Fig. 2, sr") and surrounded with secre-
tion from the cement gland (f-g.~) which
forms the wall of the egg string. The
eggs are distorted when passing through
the thorax, but in most cases regain their
symmetrical form. Embryos in the egg
strings have their heads turned latero-
ventrally and their ventral surfaces ante-
riorly, in relation to the mother. Occa-
sionally one finds an embryo reversed,

e.cx

FIG. I. Female reproduc-
tive organs of Caligus bonito.
( Drawn by Emerton from Wil-
son's CaligidiT.} c. g., ce-
ment gland ; e. c. , external
egg cases ; o , ovary ; o. d. ,
oviduct; s. r., semen recep-
tacle.

or partially rotated.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 4!

In the male the same general arrangement of reproductive
organs exists, save that the seminal receptacles are absent and
the distal ends of the vasa deferentia are enlarged and become
the spermatophore receptacles. The cement glands secrete the
material for the walls of the spermatophores.

In the dichelestid (Fie. 3) the genital segment is very much
elongated and the oviducts are not convoluted. The ovaries have
been carried backward until they lie in the anterior end of the genital

e.g.

FIG. 2. Semen receptacles and vagina of a female Lepeophtheirus. (Partly
after Claus. ) c. g. , cement glands; d. , cement gland duct; e. c., egg cases; s. ,
spermatophores; s. r., spermaries ; v., vulva; va. , vagina; 5, fifth legs. (From
Wilson's CaligidtE.)

segment and the oviducts pass forward to the anterior end of the
thorax, then backward to the attachment of the egg strings. In
other respects the description given for Caligus will hold for this
species. The older brood of oocytes occupies the posterior two-
thirds of the oviduct. When one brood passes into the egg
string the distal eggs of the remaining brood have nothing to
press against and tend to round up and become distorted. The
thickness of the oviduct at its posterior end varies somewhat,
depending on the number of eggs produced in one brood, and
the size of the eggs. When fewer, the eggs are thicker, and
when more numerous, thinner. The egg string is more slender
than the posterior portion of the oviduct, and the contained eggs
are, therefore, thicker.

J. F. McCLENDON.

In Lcemargus the egg strings are packed in loops

under broad lamellar coverings.

IV. OOGENESIS AND MATURATION.

i . Oogenesis.

A. The Dichelestid. - -The oogonia (Fig. i) are
very small isodiametrical cells. The nucleus is
spherical and contains chromatin granules in a
peripheral linin reticulum. Minute nucleoli are also
embedded in this reticulum. The cytoplasm stains
deeply. I have found but few oogonial mitoses in
this species although I have sectioned over a hun-
dred females ; the number of chromosomes is the
same as in the primary germ cells of the embryo
(16) and tivice the reduced number (8).

The primary oocyte is readily distinguished from
the oogonium by its nucleus being about half the
size of the latter (when first formed). The earliest
stage I have is the late telophase, in which the
chromosomes are drawn together as though at the
pole of the spindle, yet the whole mass is surrounded
by a nuclear membrane (Fig. 2). The nuclear
membrane is therefore formed of the fused linin
sheaths of the chromosomes (or from cytoplasm ?).
This mass is placed excentrically in the nucleus
and the individual chromosomes are so pressed
together or fused that only their ends sticking out
can be distinguished separately, so that it resembles
a " synapsis." The cytoplasm stains deeply. The
oocytes are arranged in single linear series as stated
above and are pressed and flattened one against the
other ; the nuclei are close to the free surfaces of
the cells, and that edge of an oocyte containing the
nucleus is thicker than the opposite edge. Soon
the chromosomes swell and the chromatin becomes

FIG. 3. The Dichelestid, ventral view of J . The ovaries and ovi-
ducts are solid black. Portions of the egg strings are still attached.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 43

dispersed as granules in a peripheral achromatic reticulum, as in
the oogonia (Fig. 3). Minute nucleoli appear in this reticulum.
The boundaries between the oocytes become so faint as to be no
longer everywhere perceptible. At the same time the cyto-
plasm begins to increase in volume and the diameter of the egg
cord to increase. This process continues until the egg cord
passes into the oviduct, and for some time thereafter. When
the oocytes have traversed about a fifth of the oviduct, the cell
boundaries reappear and the nucleus migrates to very near the
center, there forming an enlargement in the oocyte (Fig. 4). The
reticulum and the three to four nucleoli have grown considerably
with the growth of the nucleus. A few yolk spherules and fewer
oil globules appear and grow to considerable size. For some
time growth of the oocyte consists almost solely in the addition
of yolk spherules and oil globules and this continues until the
yolk almost obliterates the cytoplasm and closely surrounds the
nucleus. I have seen these spherules and oil globules separated
from the protoplasm with a high speed centrifuge (kindness of
Dr. Lyon), the former are heavier and the latter lighter than the
protoplasm.

B. L<einai'gus uniricatus. The oogenesis resembles that of
the dichelestid. In the oogonial divisions there appear to be
sixteen chromosomes, but these are set so close together that I
cannot be sure by actual count.

In the growth period the reserve materials, yolk and oil
globules, are laid down in ways that show them to be of quite
different consistency. The oil globules (dissolved out in sections)
appear as minute points that grow in size but always retain a
spherical shape. The yolk substances, at least some of them,
are laid down as thin discs. A disc receives new substance on
its flat surfaces only, and the layers of substances are alternately
chromatic and achromatic in staining qualities (when much de-
stained after certain stains). Some of these piles of discs become
spherical, others oblong by addition of more discs. The con-
stituent layers do not mix but remain separate until dissoved in
the segmenting egg or developing embryo. The oil globules
are relatively few in number and large in size. The yolk bodies
vary greatly in size, none of them reaching the size of the oil
globules.

44 J- F - McCLENDOX.

C. Pandarns sinuatits. - - The oocyte of Pandarus si mi at us
differs from that of the dichelestid in that a single nucleolus is
formed, about equal in bulk to that of the several nucleoli of the
latter species together. The oocyte of the former is much more
compressed (thinner) than that of the latter.

2. Maturation.

A. Pandarus sinuatits. - - At the end of the growth period
the nucleus becomes irregular in outline which gives it a shrunken
appearance. The nucleolus becomes vacuolated and the nuclear
sap intensely staining. The location of chromatin cannot be
made out very well, but in thin sections chromatic threads can
be seen radiating from the nucleolus. Soon the nuclear sap
fades enough to show that the chromatic threads have split
(Fig. 6). The nuclear wall is dissolved. Sometimes the vacuoles
in the nucleolus increase very much in size and fuse into one.
In preparations of the entire oocyte the chromatic thread can be
seen to be divided into eight double chromosomes (Fig. 5). By
shortening of these double-rod-shaped chromosomes, ring-shaped
chromosomes are formed (Fig. 7). A transverse constriction
transforms the diad into a tetrad, that is a ring constricted at
four equidistant points, the two opposite constrictions, represent-
ing the divisions between the original rods, being deeper than
the other two. The spindle when first formed is longer than the
shortest diameter of the egg. It is similar to that found in free
living copepods, having no polar rays. The dense protoplasm
at each end of the spindle (Figs. 7 and 9) may possibly be de-
rived from archoplasm of preceding divisions, and it shows a
striking resemblance to the pole plates of the dividing nucleus in
Protozoa. The spindle is at first parallel to the flat surfaces of
the egg and rotates to an almost vertical position (Fig. 8). The
first polar body is very small and is extruded between the egg
and its neighbor at about the center of the flat surface that is
posterior (in relation to the mother) (Figs. 9 and 10). The
second polar spindle is smaller than the first. It is at first
parallel to the flat surfaces of the egg and rotates nearly to a
perpendicular position under the first polar body (Figs. 9 and
10). I have not seen the second polar body being cut off and

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 45

have distinguished only one polar body in later stages, which on
account of its size I regard as the first polar body or both polar
bodies fused (Fig. 49, P\ It may be that the second polar body
remains in the egg as in Cyclops and is not easily distinguishable,
or more probably it is extruded immediately under the first polar
body and the two being pressed together and becoming very flat,
appear as one. The chromosomes in the first polar body do not
swell and fuse to form a nucleus.

I have made no observations on the entrance of the sperm.
Vide under Polarity of the Egg.

B. L&margus muricatus (Fig. 18). - - Maturation in this species
is very similar to the process in Pandarns. The disintegration
of the single large nucleolus shows many stages which appear
to be peculiar to this species. A number of vacuoles (globules
of achromatic fluid) appear in it, and these sometimes fuse into
a single mass. The chromatic substance does not remain as a
continuous peripheral layer but rounds up into four or more
masses whose inner surfaces are hemispherical and outer surfaces
form portions of the general surface of the nucleolus. In some
cases there are two fluids (one colorless and the other staining
very faintly) formed within the nucleolus. The colorless fluid
forms a large sphere in the center and the faintly staining fluid
occupies the periphery and contains spheres of the original chro-
matic substance of the nucleolus. The chromosomes do not go
as far in metamorphosis as in Pandarus. In the equatorial plate
the chromosomes appear as short double rods (Fig. 18), and
if we regard the spireme formed of chromosomes joined end to
end and cut into half the usual number of pieces the first matu-
ration division is equational and the second reducing. On the
other hand if chromosomes pair in the synapsis stage and lie
parallel to form double rods, the first division is reducing and
the second equational. The chromosomes cannot be distin-
guished all through the resting period, and whereas I am inclined
toward the view that each rod of the double rod represents a
chromosome, I have not sufficient direct evidence to be sure that
this is the case.

C. The Dichelestid. - -Toward the close of the growth period the
nucleus undergoes an enormous change. The nucleoli become

46 J. F. McCLENDON.

vacuolated and irregular in outline. The reticulum containing
chromatin begins to break up and the nuclear sap stains intensely,
probably due to the solution of chromatin. The nucleus be-
comes irregular in outline, probably due to pressure of yolk.
From one to three spheres of protoplasm (Fig. 1 1) are found in
some eggs near the wall of the oviduct, which appear to be
abnormal structures. With the breaking up of the reticulum,
the chromosomes are formed, but the manner in which this occurs
is obscured by the intense staining of the nuclear sap. (This
process shows better in Pandams.'] In order to see any struc-
tures in the nucleus it must be de-stained until it is very faint.
The nuclear membrane disappears and the karyoplasm increases
in volume and soon some of the chromosomes are distinctly
double. The nucleoli disappear suddenly. The karyoplasm
fades somewhat and becomes filled with alveoles, some of which
are very large (Fig. 12). Each chromosome, of which there are
eight (Fig. 13), is surrounded by a sphere of homogeneous pro-
toplasm that stains slightly deeper than the surrounding cyto-
plasm. Each chromosome is a tetrad and often opens out into
a ring constricted at four equidistant points. Two opposite con-
strictions are deeper than the other two and are to be regarded
(from comparison with L&margns and Pandarus) as the first di-
visions formed, and probably divisions between whole chromo-
somes. When seen on edge the tetrad usually appears dumb-bell
shaped. There is some variation in the shape of the chromo-
somes in early prophases (Fig. 1 2) but in later stages they appear
quite uniform in size and shape (Fig. 14). The spheres of homo-
geneous protoplasm surrounding the chromosomes fuse into one
mass. A colorless area appears around each chromosome (Fig.
13) which makes it appear as though each chromosome was
enclosed in a linin sac.

These linin " sacs " are each drawn out in the form of a spindle,
and these spindles, lying parallel, form the first maturation
spindle. Spindle fibers develop (Figs. 13 and 14) and some of
them become attached to the chromosomes. The spindle is
elongated, and at each pole the protoplasm stains more intensely
resembling the pole plate in protozoon mitosis (Fig. 16). The
paler protoplasm forms a sphere around each pole of the spindle,

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 47

(Fig. 1 6). From these spheres radiate strands of protoplasm
simulating astral rays, but not so dense. The spindle is formed
parallel to the flat surfaces of the egg and then begins to rotate
to a perpendicular position, at the same time shortening (Fig. 17).
In this state it remains until fertilization, and further than this I
have not followed it. The behavior of the linin sheath of the
chromosome is very peculiar, and it seems to be more conspic-
uous than in any other instance I know of. The spindle, in
general appearance,, resembles the first cleavage spindle of
Cyclops (Hacker).

V. SPERMATOGENESIS.
Lcemargus mnricatns. (Preliminary Notice^

The spermatogonia are small cells with comparatively large
spherical nuclei, and are arranged in single linear series. The
nucleus contains chromatin granules in a peripherical linin retic-
ulum. When preparing for mitosis these granules become
arranged into looped rows, but I have not ascertained whether
these rows form a continuous spireme or not. There is a single
large nucleolus. In the last spermatogonic divisions there are 16
chromosomes (twice the number of the spermatocytic divisions).
The preliminary spermatocytes are half the size of the spermato-
gonia when first formed, and are arranged in single linear series.
Whether there is a single chain or cord of cells in the testes as
in the ovaries, I cannot tell, owing to the many convolutions,
but think there are at least several such chains or cords. The
regular growth of these causes the testis to be divided into zones.
During the growth period the cells grow to the size of the sper-
matogonia and cannot be distinguished from them save by their
position in the testis (growth zone). When preparing for division
the cells lose their linear arrangement ; the chromatin forms
eight double rods which lie close together (synapsis stage).
Each double rod becomes ring shaped and the rings contract
until only a small lumen is left. An additional pair of constric-
tions converts the ring into a tetrad. The division forms two
secondary spermatocytes, each with eight diads, and another
division immediately following forms four spermatids, each with
eight chromosomes. The spermatids are grouped in fours. In

48 J. F. McCLENDON.

some of these groups, each of the cells becomes filled with an
achromatic substance which presses the chromatin and proto-
plasm against the cell wall. This substance increases greatly in
quantity and becomes more and more like yolk (called Austreibe-
stoffby C. Heider, '79). In very rare cases in some very young
spermatids I have seen a small achromatic sphere in the cyto-
plasm and have found a still smaller sphere in the cytoplasm of
some primary spermatocytes. On the other hand the substance
in question at first appears to lie within the nucleus and is closely
surrounded by chromatin. A plausible explanation might be
that the small globule seen in some very young* spermatids
moved up to and indented the nucleus, there growing until it
became larger than the original nucleus yet I have been unable
to find steps in such a process, in fact I believe the substance
arises in the nucleus.* The cells containing these spheres nourish
the spermatozoa, and may be called nutritive cells.

Going back to the spermatids, many groups degenerate, be-
coming much shrunken, while those that will form spermatozoa
collect into larger groups, the cells of which begin to elongate
radially. As the spermatids become longer they come to lie
nearly parallel. The nucleus elongates into a spindle shaped
deeply staining fiber, covered by a thin layer of achromatic sub-
stance. The remaining protoplasm fuses with that of adjacent
spermatozoa and forms a mass in the center of the group. These
groups sometimes lie with one end against a nutritive cell, and
when the groups break up the spermatozoa collect around nutri-
tive cells until the cytoplasm of the latter disintegrates and the
chromatin collects into rounded masses that float about, leaving
only a sphere of a yolk-like substance, the achromatic substance,
which has now developed an affinity for plasma stains. As the
elements pass down the vas deferens the nutritive spheres lie
near its walls and on reaching the spermatophore the nutritive
spheres (Austreibekorperchen) form a layer several spheres
deep, and become pressed together into polyhedrons. Most of
the chromatic spherules form a layer within the nutritive layer but
some chromatic spherules caught between the nutritive spheres,

* Compare formation of " Glanzkorper " from the plasmosomes of Pelomyxa :
Goldschmidt, Arch. f. Protistenkunde 5, p. 130.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 49

become elongate curved bodies that remain for a long time.
The spermatozoa fill the center of the vas deferens and lie parallel
to one another. In the spermatophores they extend radially
from the nutritive walls. In spermatophores that have been
attached to the female a long time, the nutritive material has
disappeared, leaving a mass resembling evacuated cell walls.

The same description of the spermatogenesis holds in general
for Pandams siiniatns, though there are many minor differences.
As regards the behavior of these nutritive bodies I have observed
one similar instance, in Peripaius, but in Peripatns the nutritive
bodies are nuclei of degenerating cells. In these copepods the
nutritive body appears to form in or in close relation to the
nucleus, but too little is known both of the nutritive bodies in
Peripatns and copepods to suppose them genetically related.

VI. POLARITY OF THE EGG.

The fertilized egg and embryo in all cases in which the egg
string extends straight behind the animal and contains a single
linear series of eggs, is definitely oriented in relation to the
mother. The animal pole of the egg is posterior and the first
protoplasmic cell and resulting head end of the embryo is latero-
ventral. The chief axis of the egg is manifested in the ovary in
the flattening of the egg (being the shortest axis). In any stage
in which the primary oocyte is considerably thicker than the
diameter of the nucleus, the latter lies nearer the animal pole.
The head end of the embryo coincides with the region in which
the first protoplasmic cell is formed, and this is probably deter-
mined by the point of entrance of the sperm. The seminal
receptical opens into the oviduct by a small orifice which would
lead sperm to the egg at or near the position of the future head
of the embryo. Cases of rotation of the long axis of the embryo,
which sometimes occur, might be due to the spermatozoon getting
around the egg before entering.

The first polar body is extruded in a slightly eccentric position
on that flat side of the egg directed toward the free end of the
egg string. Thus the chief axis of the egg does not exactly
coincide with the shortest axis, but is a little inclined toward the
anterior end, yet not enough to cause the first polar body to lie

5O J. F. McCLENDON.

in the first cleavage furrow. This is probably due to the great
inequality in the first cleavage. Furthermore, the blastopore
does not close at a point diametrically opposite to the first polar
body, but considerably posterior to such a point. From a study
of the literature on gastrulation in Crustacea, I am led to believe
that the blastopore in the majority of cases in this group closes
posterior to the vegetal pole and such a character would be
accentuated by flattening of the egg. As the eggs occur in single
linear series in an egg string surrounded on all sides by sea
water, there is probably nothing in the surroundings that could
determine the axes of the egg, and we should regard them as
probably determined by the structure of the protoplasm.

VII. SUMMARY.

1. In the maturation of the eggs of parasitic copepods the
behavior of the chromosomes in regard to the question of reduc-
tion is very similar to the same process in the free living copepods,
yet I differ from Haecker as to the reducing division, considering
the first maturation division most probably to be the reducing
division.

2. In the spermatogenesis only a small proportion of the
spermatids become spermatozoa. Many spermatids degenerate,
others become metamorphosed into peculiar nutritive cells. The
protoplasm of the nutritive cells degenerates leaving only a sphere
of deutoplasm " Austreibekorperchen," which C. Heider ('79)
thought was a glandular secretion.

IX. ABBREVIATIONS.

n = Nucleolus.
/ = First polar body.

s = Sphere of protoplasm, probably an abnormal structure
x -= Darkly staining cell in yolk.
y = Yolk spherule.

52 J. F. McCLENDON.

EXPLANATION OF PLATE I.
. 1-14, 7Vie Dichelestid. )

FIG. I. An Oogonium.

FIG. 2. Three consecutive primary oocytes in the synapsis stage. Optical section
of the cord of ovarian cells.

FIG. 3. Early growth stage of same, partial disappearance of cell walls.

FIG. 4. Middle portion of oocyte containing the nucleus, from longitudinal section
of oviduct. Commencement of formation of yolk spherules (}>). Two nucleoli are
seen ().

(Figs. J-io, Pandarus siitttatus J'ern//.)

FIG. 5. A little later stage. Nucleus viewed from the animal pole. Two of the
eight double rods, or diads, are joined together end to end (to the right), so that it
is difficult to distinguish them separately.

FIG. 6. The same in a section through the short axis containing six of the diads.
Above at n, the nucleolus of another egg is represented to show a later stage in the
disintegration of the nucleolus.

FIG. 7. The first maturation spindle.

FlG. 8. Later stage of same, in which it has rotated through almost a right angle
and is pushing through the egg membrane to extrude the first polar body.

FlG. 9. Late metaphase o/ the second polar spindle. The first polar body is rep-
resented at/. ( Constructed from two sections of the series. )

FlG. 10. The same stage from a surface preparation of the egg. The wall of the
oviduct was torn off and the eggs allowed to separate in sea water (two hours). In
separation the first polar body is pulled out into a stalked structure.

( Figs. //-//, The Dichelestid. )

FlG. II. Less magnified figure of prophase showing sphere of cytoplasm (.r) near
periphery of egg.

FIG. 12. A later stage. The nuclear sap has faded, it is vacuolated, and around
each chromosome stains darker (_/") than elsewhere. The two chromosomes that
are stippled are out of focus. The chromosomes are ring-shaped tetrads.

FlG. 13. The chromosomes are each enclosed in a linin sac, and these sacs have
begun to elongate.

FIG. 14. The same or a little later stage. The elongated linin sacs lie parallel,
with spindle fibers developed between them. The incipient spindle is formed in
dense (dark) protoplasm while this latter is surrounded by looser (paler) protoplasm.

FIG. 15. Equatorial plate, same stage as 14.

FIG. 16. The spindle has elongated perpendicular to the short axis of the egg.
The dense protoplasm forms the two poles of the spindle. While the loose proto-
plasm forms astrosphere-like structures enclosing the poles of the spindle.

FlG. 17. The spindle is half rotated toward the short axis of the egg and has con-
siderably shortened in doing so.

FlG. 18. Lcemargus muricatus Kroyer. Central portion of the egg containing the
first maturation spindle. Metaphase. The spindle fibers do not show in this prepara-
tion. Three of the eight chromosomes are in focus, one being seen from the end and
appearing smaller than the others.

BIOLOGICAL BULLETIN, VOL. XI

PLATE I

Vol. XII. January, /po/. No. 2.

BIOLOGICAL BULLETIN

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 1

PART II.

J. F. McCLENDON.

CONTENTS.

I. Cleavage of the egg 57

1 . Lcemargiis muricatus 57

A. First cleavage 57

B. Second cleavage 5&

2. The Dichelestid 60

A. First cleavage 60

B. Second cleavage 60

C. Third cleavage 60

D. Fourth cleavage 60

E. Fifth cleavage 61

F. Sixth cleavage 61

II. On the nature of the cleavage process 62

III. The mesoblast 63

1. The nauplius mesoblast 63

A. The germ cells 63

a. Pandarus sinnatus 63

b. The dichelestid 65

B. Mesoblastic rudiments of the nauplius appendages 65

2. Post-nauplius mesoblast 67

IV. The entoblast 68

V. Polyspermy 69

VI. Relation of pressure, polyspermy, etc., to the type of cleavage 7

VII. Summary 74

VIII. Explanation of plates 75

IX. Bibliography 76

The egg is covered by a chitinous chorion and the eggs are
packed (and much compressed) in a tough chitinous tube (e. c.,
Text Fig. i), which is attached to the mother. In order to allow

1 Thesis accepted by the Faculty of the Department of Philosophy of the Univer-
sity of Pennsylvania toward the degree of doctor of philosophy.

53

54 J- F - McCLENDON.

removal from the tube, the eggs must be fixed in fluids which
cause them to draw away from the chorion (partly by increasing
the osmotic pressure inside the chorion) and which do not make
them friable. The only fluids which served contained alcohol
and nitric acid. Nitric acid of 5 per cent.-io per cent, in alcohol of
about 30 per cent, served well enough for cell lineage but allowed
the chromosomes to swell (partly re-dissolve in water ?). The
addition of chromic acid (Perenyi's formula) prevented the swell-
ing of the chromosomes and made a good fixitive if manipulated
properly. The chromic acid in this mixture is changed to blue
chromic oxide, and P. Mayer in the first German Edition of Lee's
" Vade Mecum," ' says it contains 30 per cent, alcohol, 5 per cent,
nitric acid and a little nitric ether and chromic oxide, the last two
having no effect in fixation. However, it has been my experi-
ence that the chromic oxide was necessary to prevent swelling
of the chromosomes. Fischer ('99) says that nucleinic acid is
not precipitated by dilute nitric acid, and that its precipitate
formed by alcohol is soluble in water. It is possible that the
swelling of the chromosomes is due to the solution of nucleinic
acid or its compounds in the aqueous staining bath. In addition
to its other qualities, nitric acid bleaches the eggs of Pandarus,
and others that contain pigment, and although it is difficult to
wash out (in 70 per cent, alcohol) it was found to be an indis-
pensable ingredient.

I tried various standard fixatives for eggs to be sectioned, and
found them little better than Perenyi's fluid ; but for whole adults
to be sectioned for the ovaries, etc., the latter fluid seemed to be
much inferior to some others. It was probably too much diluted
by the body fluids.

For staining whole eggs Delifield's hsmatoxylin diluted and
acidulated (Conklin's formula) was the most convenient, but for
sections various stains were used. In Hermann's safranin-gen-
tian violet, besides the usual differentiations of chromatin, the
centrosomes (centrioles) stained red and the archoplasm blue
(unless the sequence of stains was reversed). Iron ha^matoxylin
g-ave the sharpest stains for chromosomes.

It is probable that there can be no fixation without some

1 " Grundziige der Mikroskopischer Tecknik," by Lee and Mayer, Berlin, 1898.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 55

artefacts, and that the differentiation shown by many stains is due
to differences in the size and density of particles of protoplasm
coagulated by fixation as claimed by some writers. I have paid
special attention to the structure of coagulated proteids, and
repeated experiments of Fischer and others by producing granu-
lar and " curdled " precipitates, and aster-like formations in albu-
min, and by staining the same. These led me to believe that
probably none of the finest structures seen in the fixed proto-
plasm could be relied upon as representing structures in the living
cell, but that such large bodies as chromosomes, spheres, etc.,
could not be considered artefacts, though their finer structure
may be changed.

In the cell lineage I have used the quartet system of nomen-
clature (of Kofoid, '94) as applied by Bigelow to Lepas (the only
crustacean whose cell lineage has been described beyond the 16
cell stage) to facilitate comparison among Crustacea, but do not
think this type of cleavage closely related to that of annelids and
molluscs, in fact the cleavage of the parasitic copepods does not
follow a quartet system and I hope no one will be misled by the
inappropriate nomenclature.

The cells of the 4 cell stage are designated a, b, c, d in a
dextral order, a being the left anterior cell. An exponent
denotes the order of the generation starting with the ovum as
the first. A second exponent is used to distinguish a cell from
other cells of the same generation and derivation. The odd
numbers refer in cases of equatorial division to cells nearer the
vegetal pole ; of transverse, to cells nearer the anterior end ; of
longitudinal, to cells nearer the sagittal plane, or in case the
cleavage coincides with that plane, the right side. Thus equa-
torial refers to the equator of the chief axis of the egg while sagit-
tal and transverse to the axes of the embryo that will develop
therefrom, but which may be distinguished in the egg as early
as the 2 cell stage.

To determine the second exponent of the two daughter cells
of any cell division, multiply the second exponent of the mother
cell by two and the product is the second exponent of that
daughter cell which has an even number for this exponent, and is
one greater than the second exponent of the daughter cell which
has an odd number for a second exponent.

56 J. F. MtCLENDON.

For cell lineage, whole eggs had to be used, and it was
exceedingly difficult to get them out of the egg tube. The best
way is to separate the eggs by cutting the tube between them
with a sharp "spear head " dissecting needle under the micro-
scope, which increases in difficulty in proportion to the flattening
of the eggs. When the eggs are separated the polar bodies are
lost and other means of orientation are necessary. In stages
before the origin of the primary germ cell it was necessary to
lay the eggs on the slide with a determined pole uppermost.
The eggs, except when abnormally placed, have the vegetal pole
turned toward the mother, and by placing the mother and
attached egg strings in cedar oil on a slide under a Zeiss binocu-
lar dissecting microscope, it was possible to lay the eggs with
vegetal pole up as they were separated, place a cover glass over
them to prevent turning, and run balsam under from one side.

Schimkewitz ('96, '99) concluded that pressure was an im-
portant factor in determining the form of the cleavage of para-
sitic copepods.

Pedaschenko ('93, '97, '98) worked for a number of years on
the embryology of Lerncsa branchialis and traced the cell lineage
to the 16 cell stage, but was mistaken in the orientation, thinking
the first protoplasmic cell to be formed at the animal pole and
not distinguishing between the two flat surfaces (dorsal and ven-
tral) of the egg in early stages. He found the germ cells to
arise from four cells at the edge of the blastopore and consid-
ered two of these to be male and two female. The identity of
two of these cells was lost (incorporated in the other two) and
the remaining two gave rise to the sex glands. If such were the
case, it seems to me that we should expect frequent occurrence
of bilateral androgyny (hermaphroditism). The close relation of
Lernaea to the forms I studied has made Pedaschenko's work of
great service as a hand-book.

Grobben ('79) had long before found the germ cells to arise
from four cells of the anterior lip of blastopore of the phyllopod,
Moina.

C. B. Wilson ('05) includes in his excellent monograph of the
Caligidae, a description of the general embryology of these para-
sitic copepods.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 57

The only crustacean whose cell lineage has hitherto been
carried beyond the 16 cell stage is Lepas, as described by Biglow
('02). Pedaschenko pointed out the resemblance between the
segmentations of Lepas and parasitic copepods and I believe this
resemblance is fundamental. The endoblast arises one generation
earlier in Lepas \ha.n in parasitic copepods but this may be due to
larger amount of yolk in the latter, which causes a retardation in
the segregation of organ-forming substances, and of gastrulation.

Canu ('92) published a paper which I have not seen, which
included embryology of copepods. Further consideration of the
literature may be found in the text.

I. CLEAVAGE OF THE EGG.

i . Licinargns muricatus Kroyer.

A. First Cleavage. --At the earliest stage I have (Fig. 19)
the male and female pronuclei lie side by side at the center of
the egg and at the equator of the spindle. At this stage the egg
throws out a number of yolk spherules into the space between the
egg and the chorion, but the exact nature of this process seems
obscure. Each pronucleus contains a nucleolus, and the chro-
matin is being aggregated into chromosomes. The pronuclei are
of the same size and apparently similar in every respect. There
is a deeply staining centriole at each pole of the spindle, sur_
rounded by a layer of hyaloplasm that is drawn out into astral
rays connected with the surface of the egg, and mantle fibers
connected with the pronuclei. The astral rays of one pole are
thicker (stronger) than those of the other. Where the mantle
fibers come in contact with the nuclear membrane, the latter is
pushed in (and partially dissolved?) and finally becomes dis-
solved, and the mantle fibers become attached to the chro-
mosomes. The spindle thus formed is elongated. The astral
rays of one pole shorten more rapidly than those of the other,
drawing this pole nearer one edge of the egg than the other.
The hyaloplasm is drawn from between the yolk globules and
the astral rays increase in thickness. Not only is the hyaloplasm
drawn into the astral rays, but small lumps of hyaloplasm adhere
to their surfaces and move toward the centrosomes. Thus the

58 J. F. McCLEXDON.

spheres increase in size by thickening of the hyaloplasm layer
around the centrosome and become surrounded by protoplasm
drawn in along the rays. One sphere grows larger than
the other and moves to the surface of the egg. Some yolk
granules are caught between the astral rays and form a clear
space between the sphere and the egg membrane and push the
astral rays outward. We thus have a central space almost free
from rays, surrounded by an annular area in which the rays are
especially aggregated. The center is bulged out and the annular
area sunken in by the stress (Fig. 20). Some astral rays connect
with those of the other pole, forming "spindle fibers" outside
the mantle fibers. The sphere at the surface of the egg soon
pulls all the hyaloplasm from between the yolk granules in that
half of the egg and the astral rays that pass through the yolk
break and are drawn into the sphere. The mantle fibers con-
nected with the outer pole shorten more than those of the oppo-
site pole, and the equatorial plate moves to the plane of the
ensuing cell division. The cell division is very unequal, sepa-
rating a cell containing very little yolk (ah' 1 } from one containing
practically all the yolk (cd z , Fig. 21).

Going back a little, by the dissolution of the nuclear mem-
branes a good deal of nuclear sap is liberated. This fluid is
hard to follow, but I have some slides that seem to show that
most of it goes toward the sphere that reaches the surface, and
is therefore included in the cell ab'-. The first cleavage plane is
parallel to the chief axis, but is very eccentric because of the
great inequality of the division. (See the section on orientation
of the egg.)

B. Second Cleavage. After fusion of the chromosomal ves-
icles in the two cell stages (Fig. 21), the nuclei thus formed
remain connected for a short time by interzonal fibers. I have
no stages in the division of the centrosome, but soon after this
division the two centrosomes are at the ends of a spindle shaped
sac or centrodesmus (Fig. 21, small figure to right below).
Mantle fibers become attached to the nuclear membrane and
the chromosomes gather in that side of the nucleus nearest these
points of attachment. From this stage on, the histories of the
protoplasmic cell and the yolk cell are different and will be
treated separately.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 59

The nuclear membrane of the protoplasmic cell dissolves in
the region of attachment of the mantle fibers, which then become
attached to the chromosomes. The remainder of the nuclear
wall, and the nucleolus dissolve and the chromosomes are ar-
ranged in the equatorial plate (Fig. 23), which is at first far
removed from the central spindle, but later the central spindle
assumes its normal position in the center of the peripheral spindle
(Fig. 24). The division divides the cell by a meridional (sagit-
tal) cleavage into equal daughter cells (a* and #*).

In the yolk cell, as the attraction spheres separate they grow
in size (Figs. 21-24). The central spindle presses against the
nucleus, forming a groove (Fig. 22) which makes it appear as
though the nuclues was divided (maternal and paternal elements
distinct), but sections show that this division does not pass com-
pletely through the nuclues. The nucleus is drawn out to about
four times its original length (Fig. 23), one sphere moving faster
than the other and reaching the surface of the egg, on the right
side of the protoplasmic cell.

This elongated nucleus is bent considerably and suggests that
it is being elongated by a force applied internally, and is bent by
external resistance, but I think the bending may be due to the
unequal pressure of the yolk, and the elongation of the nucleus
may be due wholly or in part, to the contraction and separation*
of the mantle fibers. The nuclear membrane dissolves, and
the equatorial plate is formed (Fig. 24). It is to be noted that
whereas the elongated nucleus is bent the fully formed spindle
is straight. In Fig. 23 it is seen that the end of the nucleus
attached to the peripheral sphere is enlarged and nearer to its
sphere than the other end is, probably due to increased tension
of the mantle fibers at this end, accompanied by pressure of the
yolk on the sides of the nucleus.

On dissolution of the membrane all the nuclear sap goes into the
peripheral sphere. A yolk spherule is often caught between the
astral rays and the cell wall (Fig. 24, r 5 ). Protoplasm migrates
along the astral rays to the spheres. The division cuts off a small
protoplasmic cell from a large cell containing practically all the yolk
(d*). I have not worked out the cell lineage any further in this spe-
cies, though it appears to be essentially the same as the dichelestid.

* By elongation of the central spindle?

6O I. F. McCLENDON.

2. The Diclidestid.

A. First Cleavage.- -The earliest stage I have of this is an
anaphase of the first cleavage (Fig. 25). It is similar to the
same stage in the preceding species save that the centrosomes if
they exist at all are larger and less dense, and the sphere reach-
ing the surface collects a considerable mass of cytoplasm around
it. The cleavage plane is " meridional " or more correctly, it is
perpendicular to the equator of the egg, but owing to the great
difference in size of the protoplasmic and yolk cells thus formed,
it does not pass through the animal or vegetal pole (Fig. 26).

B. Second Cleavage. - - The yolk cell (cd~} is sometimes re-
tarded in division in Fig. 26 its nucleus is yet a mass of chro-
mosomal vesicles while that of the protoplasmic cell (cd~} has
reached a late prophase. Already a thickened layer of proto-
plasm marks the place where c 3 will be cut off.

The protoplasmic cell (ah 2 } divides by a meridional (sagittal)
furrow into two cells, a s and 3 , almost equal in size (Fig. 27).
The yolk cell produces an elongated spindle similar to that in
the preceding species, one pole of which reaches the surface of
the egg to the right (left, when viewed from the vegetal pole)
of $ 3 (Fig. 27). The protoplasmic cell that is cut off (C 3 ) often
contains a considerable quantity of yolk (Fig. 28). Already a
thickened layer of protoplasm (Fig. 33) marks the place where
d^-' 1 will be cut off.

In this and the two succeeding cleavages, the poles of the
yolk cell spindle are differentiated by the appearance of larger
granules on the astral rays of the posterior side of the sphere
that is to remain in the yolk (Fig. 27). This probably occurs
also in the first cleavage but I have not the right stage to show it.
These granules are probably homologous to lumps of cytoplasm
on the astral rays of Ltcmargns inuricatus.

C. Third Cleavage. The division of the protoplasmic cells
<z 3 , ti\ and c 1 is equatorial (parallel with the face of the disc)
(Fig. 28). The yolk cell gives off a protoplasmic cell, d^~ to
the left of a*. Large granules appear on the astral rays of the
posterior side of the sphere left in the yolk. A thickened layer
of protoplasm marks the place where d : '-~ will be cut off.

D. Foitrth Cleavage (Figs. 2930).- - In this cleavage the divi-

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 6 1

sion of the yolk cell (c/ u ) cuts off a protoplasmic cell d 5 - 2 (Fig.
31) to the right of the cap of protoplasmic cells; d 4 - 2 divides
equatorially, C 4 and ^ 4>1 transverselsy ; a 4A divides obliquely
but the daughter cells (a r>A and <? 5 - 2 , Fig. 31) come to lie one
behind the other.

On the dorsal side <?'- 2 , < 4 - 2 , and O " divide transversely.

Fig. 30 is an enlarged view of the spindle in the yolk cell, con-
structed from the two consecutive sections. The astral rays are
only partially shown, and, connecting them, the alveolar (or
reticular ?) hyaloplasm between the yolk spheres is represented
by dotted lines. This is a little later stage than Fig. 29 and the
spindle has shortened more. The distinctness of the centro-
somes is exaggerated in the figure, in fact it is doubtful whether
we deal here with centrosomes, but that the sphere is denser in
the center can be shown in some cases with Hermann's safranin-
gentian violet stain.

E. FiftJi Cleavage (Figs. 3 1-35). --Of this cleavage I have
not enough stages to be sure of the lineage of every cell. There
are many disarrangements due to the cells extending over the
yolk and slipping on one another, which makes their lineage ex-
tremely difficult to follow. In the figures I have divided the
derivatives of a, b, c and d by heavy lines, and by comparing
Figs. 31 and 32, one can see the great change that has come
about.

In this cleavage the yolk cell divides, giving off (near the
center of the ventral side) the last protoplasmic cell d 6 - 2 (Fig. 32),
which is the primary germ cell.

F. Sixth Cleavage. In the fifth cleavage the divisions were
not synchronous, in the sixth cleavage the division of two cells,
d^ (the primary germ cell) and d 6 - 1 (the primary entoderm cell)
is delayed until some of the other cells are dividing for the
eighth time. After cleavage of the majority of the cells (Fig.
33) the blastoderm stretches over the yolk until it has half cov-
ered the latter (Fig. 34). During this process some derivatives
of d at each side of the egg and at the end of the blastoderm,
come to lie under the others and give rise to mesoderm. The
yolk cell (entoderm) divides totally by a sagittal furrow (Figs.
34, 35) and the primary germ cell sinks beneath the blastoderm
and divides by a sagittal furrow into two cells of unequal size.

62 J. F. McCLENDON.

Egg

W<

X 5 ' 2

/a' 6 ' 2 (Primary germ cell).

\/6-> (Entoblast).
TABLE OF CELL LINEAGE TO 32 CELL STAGE.

II. ON THE NATURE OF THE CLEAVAGE PROCESS.

Since Van Beneden in 1883 put forward the hypothesis that
separation of the chromosomes and division of the cell was
caused by contraction of the fibers of the karyokinetic figure, the
question of the mechanics of mitosis has aroused a great deal of
interest. The large size and peculiar form of the spindles in the
early cleavage of this egg, make them favorable objects for obser-
vations on this point. I have attempted to harmonize observa-
tions on the structure of coagulated colloids, and the modern
theory of the ultramicroscopic structure of colloids, with the
observations on artificially produced asters in colloids, and
asters and spindles appearing during mitosis in living cells.
A theory of Rhumbler and others as to the mechanics of the
formation of asters seems in general to be the only one applicable
to the observations I have made, yet I do not believe that this
theory is inseparable from the alveolar theory of the structure of
protoplasm. Asters can be produced in colloids which we have
no reason to believe have the alveolar structure in the strict sense.
According to Mann ('06), colloids consist of minute or ultrami-
croscopic particles suspended in a thin fluid. On congealing
(Hardy, Jour, of Physiol., V., 24), these particles by mutual at-
traction form rows which make up a meshwork (or interalveolar
structure ?) giving consistency to the mass. When colloids are
coagulated with substances (electrolytes) that act strongly and
quickly, the particles are large enough to be seen with the micro-
scope and are at first distributed homogeneously through the
fluid, but soon arrange themselves in rows which make up
a meshwork (" gerinnselbilder ' of Fischer). This passing of
a colloid from the "sol' to the "gel" or congealed state
may be hastened by addition of a fragment of coagulated colloid
to the former, in which case the rows of drops or particles
arrange themselves radially around the fragment, and an aster is

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 63

formed. Fischer ('99) varied this experiment in a number of
ways, and one of his experiments . modified slightly, might be
tried by every one interested in the subject without much
trouble : Spread a layer of egg albumin (which consists chiefly
of albumin and a little globulin) on a slide and through a capil-
lary tube introduce a small drop of a fixing solution into the
albumin, observing the changes that take place under the micro-
scope. The albumin immediately around the drop is coagulated
into a membrane through which the fixing solution diffuses and
from which radiations begin to form, giving the whole structure
the appearance of an aster with the drop of fixing solution and
the membrane around it as the centrosome. If the rays form,
as they seem to, by mutual attraction of the drops or particles
in the fluid, such rays or rows of drops would exert a pulling
force, and if their ends were released should shorten by synaeresis
into a spherical mass. This may be the nature of the fibers of the
karyokinetic figures in the cleavage of these copepods but does
not explain the direction of movement of the asters.

III. MESOBLAST.

i. Nauplius Mesoblast (Pandarus sinuatns, PI. IV. and V.).

When the cap of the protoplasmic cells has covered about one
third of the yolk some of the marginal cells (lip of the blasto-
pore) become differentiated as mesoblast. Of these one or more
on the right and left edge will give rise to mesoderm of the first
and second antennae, and one near the middle of the ventral side
and distinguished by its large nucleus (Fig. 37) will give rise to
the sex or germ cells.

A. The Gcnn Cells. This primary germ cell is turned under
the rim of the blastopore (Fig. 38) and divides by a sagittal fur-
row into two (Figs. 39-40), which lie about the center of the
ventral side just under the ectoderm. About the time of the
closure of the blastopore these two divide by transverse furrows
into four (Fig. 41). This group of four cells rotates until one
cell is anterior, two lateral and one posterior. (In Fig. 42 the
rotation is not quite completed.) But there is considerable vari-
ation in the amount of rotation (Figs. 41-48). The four germ

64 J. F. McCLENDON.

cells lie just beneath the ectoderm until the Metanauplius stage,
when by concentration of the ventral nerve chain, the latter is
pushed under them and they rest on top of it (Fig. 56). I have
said the germ cells have large nuclei --the nucleus is further
characterized by the fact that the chromosomes remain distinct
as oval masses just inside the nuclear membrane (Fig. 52). The
chromosomes can be counted and are sixteen, just twice the
number in the female pronucleus. The cytoplasm is much
vacuolated. The germ cells are flattened against each other,
and are flattened against the ectoderm in the early stages (Figs.
49, 50). In later stages they detach from the ectoderm, and
round up (Figs. 52, 56). And still later (during the Metanau-
plius stage from 24 to 72 hours after hatching of the larva) they
separate and pass laterally and upwards into the yolk and two
of them come together dorsal to the intestine, and I have not
traced them further than the fourth day after the larva hatched,
when they were still two in number. Pedaschenko says that two
of the four genital cells pass to the right and two (one lateral and
one median) to the left. Each pair fuse and, probably by degen-
eration of one nucleus, becomes a single cell, which finds its way
upwards and posteriorly and by division forms the ovary of that
side. The fusion of each pair he considers of great significance
and the basis of a theory on the origin of the sex of the adult.
He believes that one cell of each pair is male and one female and
the one whose nucleus persists after fusion of the cytoplasm de-
termines the sex of the animal. His belief that two cells of the
four are male and two female is based on comparison with O.
Hertwig's account of Sagitta in which this condition exists, with
difference however in the later history. In Sagitta the two
female cells give rise to the ovaries (in the anterior part of the
animal) and the two male cells give rise to the testes (in the pos-
terior part of the animal).

Sex is said to be determined in some animals by amount of
food (of the individual, the parents, or the grandparents) in others
by fertilization vs. parthenogenesis, in others by dimorphism of
egg or spermatozoon, in others by temperature, etc. Peda-
schenko proposes an additional factor.

Haecker ('97) found in Cyclops the primary germ cell differen-

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 65

tiated from the somatic cells at the close of the fourth cleavage
or one generation earlier than in the parasitic copepods.

Boveri ('92) in Ascaris, and Haecker ('97) in Cyclops traced
the " Keimbahn " from the first cleavage. In Ascaris the visible
difference between germ cell and somatic cell was in the chromo-
somes, in Cyclops in the cytoplasm. Early differentiation of the
germ cells has been noticed in a large number of animals, but
the causal factors in their differentiation are yet unknown. From
Boveri's account of Ascaris, it seems that the cells of the " Keim-
bahn " preserve all the characters of the fertilized egg, while the
somatic cells lose some characters. Yet the mature ova and
spermatozoa of most animals are possibly as highly differentiated
as any somatic cell.

In the dichelestid the germ cells have the same origin as in
Pandarns simtatns but they differ in appearance. Fig. 57 shows
the primary germ cell beginning to be turned under the blasto-
poral rim. Fig. 58 shows a stage after the division of the germ
cell into two cells (of unequal size). If we followed Pedaschenko's
theory we might consider the large cell as female and the small
cell as male as it is always true that one is larger than the other.
The nuclei of the two germ cells lie in their ends that are nearest
the free border of the blastoderm (blastopore). These two cells
divide into four and the nuclei of two are larger than of the
other two, but the cell boundaries between them are extremely
difficult to make out.

B. The Mesoblastic Rudiments of the Nanpliits Appendages
(Pandarus sinuatus} arise from cells turned under the rim of the
blastopore during epibole. When the cap of protoplasmic cells
has covered about one third the yolk (Fig. 38) a few cells are
turned under the rim at the extreme right and left, that is to
say at the edge of the disc shaped egg. These cells are the
mesoblastic elements of the first and second antennae and divide
on each side into two masses (Fig. 40, an 1 , ati 2 }. The time
of this division varies slightly, the elements being sometimes
widely separated before closure of the blastopore (Fig. 40) and
sometimes close together just after the closure of the blastopore
(Fig. 41, an 1 , atr). Just before closure of the blastopore, a
few cells are turned under its lip on each side (Fig. 40, uui] and

66 J. F. McCLENDON.

are the mesoblastic rudiments of the mandibles. This completes
the rudiments of the nauplius appendages. After the close of
the blastopore the post nauplius segments are laid down by
teloblastic growth at the posterior end, and the nauplius is pushed
(compressed) forward, carrying the rudiments of the second an-
tennae and mandibles forward (Figs. 47-48), and causing the
three pairs of appendages to lie closer together. In stage D
(Figs. 456) the appendages begin to grow out and at the same
time the muscle cells elongate into fibers. I think it more profit-
able to follow these latter backward in development, as it seems
doubtful whether they have a single or a double origin. Observe
the muscle cells in Fig. 45 elongating radially and attached peri-
pherally to the rudiments of the appendages. In Fig. 44 (Stage
C] the muscle cells (one shown at in] are just beginning to
differentiate from the mesoblastic rudiments of the appendages,
and two of them have begun to elongate (compare Fig. 50, ;//).
The question arises whether all or only some of these muscle
cells arose from the mesoblastic rudiments of the appendages.

Just after the closure of the blastopore a few cells similar to
these muscle cells are seen considerably removed from the
mesoblastic rudiments of the appendages (Fig. 41, ;;/). And
just before closure of the blastopore minute cells with scarcely
any cytoplasm are seen budding off from the ectoderm in this
region, (Figs. 40, 49, -i'). There is a slight probability that
some of the cells in arise by growth of the cells x which would
be a case of muscle cells arising from ectoderm as in ccelenterata,
etc. But small cells with hardly any cytoplasm are found in the
yolk at many stages of the embryo (Figs. 44 and 47, .r) and
although I have not closely traced them from cells like x in Fig.
40, I think their resemblance in structure indicates a likeness in
origin. I think the evidence indicates that all the mesoblast
arises from cells turned in from the lip of the blastopore, as is
the case in other copepoda, phyllopoda, decapoda and cirripedia.

The muscle cells when first elongated push the ectoderm
toward the center and mass it in a sort of structure which some-
what resembles the "dorsal organ" which disintegrates, and the
elements of which wander into the yolk. The muscle cells are
thus arranged radially just beneath the extremely thin dorsal

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 6/

ectoderm (Fig. 45) but the forward movement of the appendages
carries their peripheral ends forward (Figs. 4548) until they
assume a longitudinal direction. Contractile fibrilLne begin to
form in the muscle cells in stage E (Fig. 47) and the nucleus
and undifferentiated cytoplasm is pushed to one side. In the
liberated nauplius the muscle fibers run almost the whole length
of the animal and show cross striations (Fig. 5 i). Each append-
age then has at least one muscle fiber attached' to the anterior
and one to the posterior border of its base. Muscle cells that
go into the hollow appendages as they grow out, form muscles
attached to the bifurcated ends of the appendages (Fig. 5 1 , left
side).

The same description in general holds good for the dichelestid
and Lceniargns. In these the mesoderm of the appendages is
clearly derived only from marginal cells. In Lamargus the
ectoderm massed in the middle of the dorsal side by growth of the
dorsal muscle fibers forms a more conspicuous "dorsal organ"
than in the other species and the elements arising from its dis-
integration are more numerous.

In relation to the formation of the appendages might be men-
tioned the segmentation of the nauplius of L^margns. Soon
after the closure of the blastopore the embryo is divided by bands
of thinner ectoderm into three segments corresponding to the
three pairs of nauplius appendages. This segmentation slowly
disappears with the development of the nauplius. Other species
show it but to a less degree than Lceinargus. This segmentation
might be used as evidence that the nauplius of ancestors of crus-
tacea was segmented or it might be considered as coenogenetic
and associated with the development of the appendages and
neuromeres of the nauplius.

2. Post nan pi ins mesoblast (Pandarus si nit at us).

At the closure of the blastopore some of the marginal cells are
turned in (Fig. 49, Mp) and become the mesoblast of the post
naupliar segments. These cells are much larger than the sur-
rounding cells (Fig. 41) and form a mass at the posterior end of
the animal that is destined to develop mesoblastic somites by
teloblastic growth. By rapid division the cells become small and

68 J. F. McCLENDON.

by this time the ectoderm has completely closed over them
(Fig. 42). This mass of cells divides in the sagittal plane into
two masses (Fig. 44), which begin to grow forward as a pair of
broad bands under the neural thickenings of the ectoderm (Figs.
46, 47, Mp]. From the anterior ends of these bands oval masses
are cut off that are the mesoblastic somites (Fig. 48).

IV. ENTOBLAST.

The entoblast is segregated one generation later than in Lepas.
In the 32 cell stage the entoblast consists of one cell that con-
tains practically all the yolk and which does not divide until the
majority of the cells have completed the seventh cleavage and
some are in the eighth. It then forms a very long transverse
spindle with the poles inclined anteriorly. The daughter nuclei
are widely separated, but in Pandams, Lannargiis and other
Caligidffi the yolk does not segment. The next (second) division
occurs about the time the "blastoderm" has covered half the
yolk. The spindles extend longitudinally and the poles are
inclined outward. Each of the two spindles is shorter than that
of the previous division (Figs. 35, 38). The next (third) division
occurs about the time of the closure of the blastopore (Fig. 40).
There is much variation in the direction and curvature of each of
the four spindles, but the daughter nuclei are about equally dis-
tributed through the yolk as they are after each division. The
fourth division occurs in stage B (Fig. 42) and the fifth in stage

<^(Fig. 43)-

In Eudactylina the yolk segments in the first three cleavages
of the entoblast, (forming eight cells) after which the entoblast
forms a syncytium. In the dichelestid the yolk divides into
four cells and is then transformed into a syncytium. In the
remaining species studied a syncytium is formed from the first.
This omitting of the cleavage of the yolk is probably not entirely
due to the amount of yolk present, which is as great in the
dichelestid as in Lczinargns, but largely due to the extent of
compressions of the egg, for it has gone farther in those eggs
which are compressed the most. The entoblast nuclei migrate
to the surface of the yolk and form the enteron or mid gut, as
described by Pedaschenko.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 69

V. POLYSPERMY.

In Lcemargus inuricatus I have found many eggs into which a
number of spermatozoa had entered. In one case the whole egg
string was of such eggs ; in the other cases only a few such eggs
were found in a string. The " development " of these eggs falls
under three classes :

1. The 9 pronucleus and the c? pronuclei fuse to form one
nucleus in the center of the egg which does not develop further.

2. In the center of the egg a multipolar spindle is formed
usually of three principal poles and one minor pole. The result-
ing division in all observed cases cleaves the egg into three sub-
equal cells, in each of which a bipolar spindle with a very large
number of chromosomes is formed. Further development is
very irregular.

3. A bipolar spindle with an immense number of chromosomes
is formed in the center of the egg. Apart from the number of
chromosomes the cleavage approaches the normal type, especi-
ally up to the 4 cell stage after which it diverges more and
more from the normal type. I am led to believe by certain eggs
that show an intermediate stage between a multipolar and a
bipolar spindle, that the bipolar spindles in the first cleavage of
these eggs are formed out of multipolar spindles.

As all of these eggs were already mounted (by Professor Rynear-
son) I was not able to observe whether the axes of these poly-
spermous eggs that approached the normal type in development
were the same as in normal eggs. I have not observed whether
maturation takes place in polyspermous eggs the cleavage
spindles are very different from normal cleavage spindles, and
are very similar to normal maturation spindles. This may be due
to a tendency to throw out the excess of chromatin, and in some
cases I have found a mass of very small cells extruded from the
egg, not always, however, in the position in which polar bodies
normally form. There are often many asters in the egg uncon-
nected with chromosomes, and this may account for rounded
masses of yolk that are sometimes cut off from the egg.

7O J. F. McCLENDON.

VII. RELATION OF PRESSURE, ETC., TO THE TYPE OF CLEAVAGE.

When the eggs are released from the oviduct in sea water,
they begin slowly to round up and separate one from another.
The eggs adhere together so strongly that their tendency to
assume a spherical form is greatly impeded, and it always takes
several hours for them to round up. The majority of the eggs
liberated begin to disintegrate before they proceed very far toward
becoming isodiametrical. This is due to their very low surface
tension, their cohesion being less than their adhesion for the
surface film of sea water or for glass. This is shown by the fact
that the eggs tend to stick to the bottom of the glass dish con-
taining the sea water, and when the dish is tilted so that some
eggs come in contact with the surface of the water, they quickly
spread out over that surface. All these experiments support
the direct observation that the oocyte is surrounded by no other
membrane than its surface film.

If eggs are left standing in sea water more than two to four
hours their surfaces begin to disintegrate. This is probably caused
by partial solution in sea water. The nuclei remain intact after
a great deal of the egg has disintegrated. If the eggs are placed
in hypotonic solutions they swell, if in hypertonic solutions, they
shrink, without any other change that can be observed. I tried
solutions of magnesium chloride, ether, and sodium hydroxide,
of varying strengths in sea water containing eggs alone 01 eggs
and sperm but could neither induce parthenogenesis nor fertiliza-
tion. The spermatozoa are very similar to those of cirripedia,
being thread-like and each containing a homogeneous thread of
chromatin running the entire length. The sperm of many crus-
tacea are non-motile when examined in sea water or serum, but
some of them have been observed to perform movements in the
female genital ducts. Cano ('93) saw decapod spermatozoa
move lively in the Rec. seminis. It is therefore probable that I
did not find the proper stimulus to cause fertilization in sea
water. Immediately after fertilization and passage into the egg
strings the egg secretes a chitinous chorion that resists all
attempts at freeing the eggs so that they will round up, without
mutilating them, so I had to resort to looking for eggs that by
accident were not flattened in the usual manner. In the Di-

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 7 1

chelestid the egg at each end of the string is hemispherical in
shape, due to the fact that it is pressed on only one side (Fig.
36). In the proximal egg the ventral side is rounded and in the
distal egg the dorsal side is rounded. The first protoplasmic
cell (ah) is cut off at one edge of the hemisphere. The second
cleavage results in the formation of three protoplasmic cells (a,
b, c) whose centers form the apices of a triangle on the spherical
surface at its edge (Fig. 36, A and B}. We should assume that
this arrangement is nearer the ancestral type, which was prob-
ably a sphere, and that the first three protoplasmic cells being
in the equatorial plane (Fig. 27) is due to the pressure. Fig.
36, A and B, shows a similar arrangement of cells to the same
stage in the cleavage of Lepas as figured by Biglow, save that in
the dichelestid the yolk is much greater in amount and one side
of the egg is flat. In both cases d (the yolk cell) extends under
a, b, and c but in the dichelestid the yolk cell is so large as to
push c over b (in the distal egg).

This altered arrangement of the protoplasmic cells does not
seem to affect the normal development of the embryo. The
ectoderm grows over the yolk in the usual manner, except that
it is stretched more on the rounded side of the egg (Fig. 36, C).
The four entoderm cells are thicker, and in the distal egg of
more volume, than normally and after the entoderm forms a
syncytium the nuclei have not exactly their normal arrangement,
but when the ensuing nauplius escapes from the egg membrane
everything is apparently restored to its normal relation, save that
a nauplius developing from a distal egg is larger.

This is contrary to the idea of Schimkewitz, who attributes
many abnormalities in parasitic copepod embryos to slight differ-
ences in pressure in the egg string ; but the eggs he studied had
less yolk than those considered in this paper. Differences in
pressure in the dichelestid egg result principally in differences in
form of the yolk mass. This yolk mass does not, save to very
small extent, enter as such into the composition of cells, but is
dissolved and used as food by the cells. The protoplasmic cells
always being on the surface of the yolk, their relation to the
food supply remains unchanged.

Experiments on the effect of unequal compression on cleavage

J. F. McCLENDON.

have been made on Ascaris eggs by Auerbach ('74) ; on am-
phibian eggs by Pfluger ('84), Roux ('85, '93), Born ('93, '94),
O. Hertwig ('93) ; on echinoderm eggs by Driesch ('92, '93, '95),
Morgan ('93), Ziegler ('94) ; on ctenophore eggs by Ziegler
('94); and on Nereis eggs by Wilson ('95). These experi-
ments show that if the egg is pressed
more on certain sides than on others, as
when it is pressed between two plates of
glass and forced to assume a flattened
shape, that the direction of the cleavage
spindles (and consequently cleavage fur-
rows) will be affected. Hertwig formu-
lated the law that the spindle lies in the
longest axis of the protoplasmic mass of
the cell. This rule probably applies in
the majority of cases, but there may be
some exceptions, and there are evidently
other and unknown factors which enter
into the polar differentiation of the cell.
Bigelow found in Lepas ('02) that the
polarity of the egg was not affected by
the oval form of the rigid chorion. The
principal axis of the egg coincided with
the long axis when the chorion was se-
creted and during the prophase of the
first cleavage spindle the egg rotated

through a right angle so that the first cleavage spindle was made
to coincide with the long axis of the egg determined by the form
of the chorion, and the principal or primary axis was perpen-
dicular to it.

I found a similar condition in Endactylina nigra Wilson. If the
egg string of this copepod be placed in sea water under the micro-
scope during the first cleavage stage, the majority of the eggs will
have their spindle axes, or in case the division is complete, common
cell axis, nearly in the same plane. Often, however, some of these
axes are considerably inclined to this plane as is shown in Fig. 4, A.
If the egg string be pressed between slide and cover glass the above
axes will rotate sufficiently to bring them all in a plane midway

FIG. 4. Egg strings of
Endaclylina nigra Wilson.
The protoplasm is stippled
and the yolk white, the dis-
tinctness between the two
being accentuated. A, Liv-
ing egg string during first
cleavage. B. The same com-
pressed between the slide
and cover glass.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 73

between slide and cover glass and therefore in the same plane
(Fig. 4, -B). Thus the first cleavage obeys Hertwig's rule
whether the compression be applied during or after the meta-
phase and possibly during the ensuing resting stage.

Hertwig says that if frog eggs are thus compressed normal
embryos will develop, although a totally different distribution of
nuclei results.

Born found that if a frog's egg were inverted before formation
of first cleavage spindle the relative density of protoplasm and
yolk would cause streaming movements in the egg, the proto-
plasm and nucleus rising through the yolk to the upper pole.
But it is probable that these streaming movements would be
hindered by the astral rays after formation of the spindle.

In the parasitic copepods the direction of many of the spindles
is influenced by the pressure, and Hertwig's law applies in most
cases. But the peculiar form of cleavage seems well adapted to
variations in pressure. The blastoderm lying on the yolk may
be compared to a rubber bag divided by lines into polygonal
areas. The bag may be pressed into various shapes without
altering the mutual relation of adjacent polygons. The only cell
whose form is changed very much by pressure is the yolk cell.

I have said that the distal egg (the one at the free end of the
egg string) is larger than the others. This is due to the fact
that in the oviduct it presented more surface for absorption of
nutriment through the wall of the oviduct. While the other
eggs present only a thin edge toward the source of food, this
last egg of the series presents this edge and one whole flat side
in addition. Usually it does not remain flat, but becomes more
or less hemispherical on the free side while still in the oviduct.

The cytoplasm of many cells is formed in large part from the
substances that escape from the nucleus at the first maturation
and early cleavage divisions. Dr. Conklin traced a similar proc-
ess in gasteropods and in the living eggs of ascidians, and it may
be a general phenomenon. In other words, a quantity of chro-
matin is dissolved and escapes into the cytoplasm in many and
perhaps all cell divisions.

In Paramcecium the macronucleus and a large part of the
substance of the micron ucleus escapes into the cytoplasm at each
copulation and may constitute necessary ingredients of the cyto-

74 J- F. McCLENDON.

plasm, as copulation is necessary to long continued existence of
Paramcecia. It is probably chiefly thus that the heritable quali-
ties residing in the chromosomes are conveyed to the cytoplasm.
I do not mean to say that this is the only way the nucleus affects
the cytoplasm, for with a few exceptions (7. e., red blood cor-
puscles of higher animals) cytoplasm not containing a nucleus
soon dies, but if the heritable qualities are stored up in the chro-
matin, part of this chromatin bearing these qualities could be
transferred to the cytoplasm more easily during the absence of a
nuclear membrane.

After the close of the fifth cleavage (32 cells) the embryo is
composed of three types of cells that differ visibly.

1. The primary germ cell.

2. The primary entoblast cell.

3. Thirty cells of the blastoderm all similar in appearance.
The primary germ cell when first separated from the entoderm

looks like the other cells of the blastoderm, but during the rest
grows larger than its neighbors and is delayed in mitosis. In
this character of delayed mitosis it resembles its sister cell (pri-
mary entoderm cell). There is nothing characteristic of its posi-
tion that could cause it to become different from its neighbors,
so we must ascribe this difference to the difference in the sub-
stances entering into it which in turn may be caused by unequal

cleavage.

VII. SUMMARY.

i. My observations on the cell lineage agree in general
with those of Pedaschenko (who worked it out to the 16 cell
stage) save in regard to the orientation. Pedaschenko used no
means to distinguish between the two flat sides of the egg and
was mistaken in regard to the location of the animal pole, as I
have shown <p. 50). At the fifth cleavage the yolk cell buds
off the last protoplasmic cell, which is the primary germ cell.
After extrusion of the germ cell (32 cell stage) the yolk cell is
purely entoblastic. The segregation of the entoblast takes place
one generation later than in Lcpas (Biglow '02), and the segrega-
tion of the germ cells one generation later than in Cyclops
(Haecker). The delay in the segregation of these two elements
is probably due to the large amount of yolk present and the
compressed condition of the egg, which cause delay in gastrula-

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 75

tion (epibole). It is probable that all the mesoderm arises from
cells turned under the lip of the blastopore.

2. The entrance of supernumerary spermatozoa into the egg
so greatly disturbs the process of development that the latter is
either prevented or so distorted that it never progresses very far,
and then in an abnormal manner.

3. The compressed condition of the egg affects the cleavage :
(i) by altering the arrangement of the protoplasmic cells, (2) by
necessitating increased length of the spindles in the yolk cells,
(3) by preventing cleavage of the yolk, and (4) by increasing the
surface of the egg and retarding gastrulation (epibole). But it is
very improbable that slight alterations in the amount and direc-
tion of compression have as great an influence on development
as supposed by Schimkewitz. I found nauplii which were appar-
ently normal (save perhaps in size) hatching from hemispherical
eggs. As the nauplius hatches it immediately rounds up, and
assumes the same form whether it arise from a hemispherical or

from a very flat egg.

VIII. EXPLANATION OF PLATES.

ABBREVIATIONS.

an ' = First antenna.

an 2 = Second antenna.

b - Blastopore.

e - Entoblast cell.

en - Entoblast nucleus.

ec = Ectoderm cell.

f = Deeply staining protoplasm.

g = Germ cell.

m = Muscle cell.

M 2 = Mesoblast of first antenna.

>n z - Mesoblast of second antenna.

m* = Mesoblast of mandible.

w s-io _ Mesoblast of post nauplius appendages.

md Mandible.

nip - Postnauplius rnesoblast.

' = Procerebrum.

n 2 Neuromere of first antenna.

n 3 - Neuromere of second antenna.

w 4 = Neuromere of mandible.

n 5 ~ 10 = Neuromeres of post nauplius segments.

o = Rudiment of mouth.

0' = Rudiment of lateral eye.

om = Rudiment of median eye.

x - Darkly staining cell in yolk.

76 J. F. McCLENDON.

IX. BIBLIOGRAPHY (OF BOTH PARTS).

Auerbach, L.

'74 Organologische Studien. Breslau, 1874.
Berg, Walt.

'05 Weitere Beitrage zur Theorie der histologischen Fixation Versuche an Nu-

cleinsauren Protamin. Arch. Anat., 65.
Berthold, G.

'86 Studien iiber Protoplasmamechanick. Leipzig, 1886 (Felix).
Bigelow, M. A.

'02 The Early Development of Lepas. A Study of Cell Lineage and Germ

Layers. Bull. Museum Comp. Zoo. Harvard College, XL., No. 2.
Born, G.

'93 Ueber Druckversuche an Froscheiren. Anat. Anz. 8.

'94 Nene Compressionsversuche an Froscheiren. Jahr. Schles. Ges. f. Vaterl.

Cultur. Zoo. Bot. Sect., 1894.
Boveri, Th.

'92 Die Entstehung des Gegensatzes zwischen den Geschlechtzellen und den
somatischen Zellen bei Ascaris megalocephala. Sitzungsber, Ges. fiir Morph.
Physiol. Miinchen, Bd. 8.
Brauer, A.

'92 Ueber das Ei von Branchipus grubei von der Bildung bis zur Ablage. Anhandl.

Akad. Wiss. Berlin, 1892.
Biitschli, 0.

'91 Ueber die sogenannten CentralkSrper. Verh. d. Nat. Med. Ver. zu Heidele

berg, IV.
'92 Untersuchungen iiber mikroscopische Schaume und das Protoplasma. Leip-

sig, 1892.

'98 " Untersuchungen iiber Structuren, inbesondere, iiber S. nichtzelliges Erzen.
guisse des Organismus und iiber ihre Beziehungen zu Structuren, welche
ausserhalb des Organismus entstehen. Leipzig, 1898.
'oo Bemerkungen iiber Plasmastromungen bei der Zelltheilung. Arch. Ent-

wichlungmechanic, X.
Cano, G.
'93 " Svilluppo dei Dromidei. Atti R. Acad. Sc. Fis. et Mat., Vol. 6, Ser. 2,

No. 2, Napoli, 1893.
Canu, E.

'92 Les Copepods du Boulonnais Trav. Lab. z. Wimereux-Ambleteuse Tome 6.

Lille.
Carnoy, J. B.

'85 La Cytodierese chez les Arthropodes. La Cellule, Vol. I.
Casteel, C. D.

'04 The Cell Lineage and Early Development of Fiona marina, a Nudibranchiate

Mollusc. Proc. Acad. Nat. Sc. Phila., '04.
Conklin, E. G.

'97 The Embryology of Crepidula, a Contribution to the Cell Lineage and Early

Development of some Marine Gasteropods. Jour. Morph., 13.
'02 Karyokinesis and Cytokinesis in the Maturation, Fertilization and Cleavage
of Crepidula and other Gasteropoda. Jour. Acad. Nat. Sc. Phila., XII.
(2d ser. ).

ON THE DEVELOPMENT OF PARASITIC COPEPODS. J"J

'05 The Organization and Cell Lineage of the Ascidian Egg. Jour. Acad. Sc-

Phila., XIII., pt. I.
Crampton, H. E.

'99 Studies upon the Early History of the Ascidian Egg. Jour. [Morph., XV.

Supplement.
Driesch, H.

'92 Entwicklungsmech. Studien, IV. Zeit. Wiss. Zoo., 55.

'93 Zur Verlagerung der Blastomeren des Echinideneis. Anat. Anz., 8.

'05 Die Entwicklungsphysiologie von 1902 bis 1905. Ergenbuisse der Anat. u.

Entw. , 14, '04.
Eigenmann, C. H.

'91 On the Precocious Segregation of the Sex Cells in Micrometrus aggregatus

Gibbons. Jour. Morph. Boston, Vol. 5.
Fischel, A.

'99 Ueber vitale Farbung von Echinoderm eiern wahrend ihrer Entwickelung.

Anat. Hefte, I Abt., II.
Fischer, A.

'99 Fixierung, Farbung und Bau des Protoplasmas, Jena.
Foot, K.

'96 Yolk Nucleus and Polar Rings. Jour. Morph., XII.
Fiirth, 0. von.

'03 Vergleichende Chemische Physiologic der niederen Tiere. Jena (Gustav

Fisher), 1903.
Gerstaecker, A.

'66 Copepoda. Bronn's " Klassen und Ordnungen des Thierreichs " V., pt. 2.
Giard.

'87 Sur un Copepod parasite de 1'Amphiura squamata. Comp. Rend., Tome

104, pp. 1189-92.
Giesbrecht, W.

'82 Beitrage zur Kenntniss einiger Notodelphyiden. M it. Zoo. St. Neapel, 3 Bd.
Gilson, S.

'86 Etude cornparee de la spermatogenese chez les Arthropodes. La Cellule,

Tome i and 2, p. 140.
Grobben, C.

'79 Die Entwickelungsgeschichte der Moina rectirostris. Arb. Zoo. Ins. Wien,

I Bd., 2 Hft.
'81 Die Entwichelungsgeschichte von Cetochilus septentrionalis. Arb.Zoo. Ins.

Wien, 3 Bd.
Gruber, Aug.

'79 Beitrage zur Kenntniss der Generationsorgane der freilebenden Copepoden.

Zeit. Wiss. Zool., 32 Bd., p. 407.
Haecker, V.

'91 Die Richtungskorperbildung bei Cyclops und Canthocamptus. Preliminary.

Ber. Nat. Ges. Freiburg, 6. Bd., Bio. Cent., u.
'92 Die Kerntheilungsvorgange bei der Mesoderm und Entodermbildung von

Cyclops. Arch. Mikr. Anat., 39 Bd.
'92, a Die Eibildung bei Cyclops und Canthocamptus. Z. Jahrb. Morph., Abt.,

5 Bd., p. 211.
'93 Das Keimblaschen, seine Elemente und Lagerveranderungen I. Uber die

78 j. F. MCCLENDON.

biologische Bedeutung des Keimblaschenstadiums und iiber die Bildung der
Viercrgruppen. Arch. Mikr. Anat., 41 Bd., p. 452.
'94 iJie Entwicklung der Wintereier der Daphniden. Ber. naturf. Gessell.

Freiburg, 8 Bd.
'94, a Ueber generative und erabryonale Mitosen, sowie iiber pathologische

Kerntheilungsbilder. Arch. Mikr. Anat., 43 Bd.

'95 Ueber die Selbstandigkeit der vaterlichen und miitterlichen Kernbestandtheile
wahrend der Embryonalentwickelung von Cyclops. Arch. Mikr. Anat., 46
Bd., p. 579.

'95, a Die Vorstadien der Eireifung. Arch. Mikr. Anat., 45 Bd., p. 200.
'97 Die Keimbahn von Cyclops. Arch. Mikr. Anat., 49 Bd., p. 35.
'02 Ueber das Schicksal der elterlichen und groselterlichen Kernantheile. Mor-
phologische Beitrage zum Ausbau der Vererbungslehre. Jena. Zeit. Natur..
37 Bd., p. 297.
Hansen, H. J.

'oo Danmarks Stilling og Tilstand. 2. Del. kongelige Danske Videnskabernes

Gelskab. Kjobenhavn, 214 pp., 10 figs.
Heath, H

'99 The Development of Ischnochiton. Zoo. jabr., XII.
Heider, K.

'79 Die Gattung Lernanthropus. Arb. Zoo. Ins. Wien, 2 Bd., 3 Hft.

'92 Crustacea. Lehrbuch der vergleichenden Entwickelungsgeschichte der wir-

bellosen Thiere. Jena, 1892.
Hermann, S.

'83 Sur la Spermatogenese des Crustace Podopthalmes, specialment des Deca-

podes. Compt. Rend., Tome 97, p. 958.
Hertwig, 0.

'93 Ueber den Werth der ersten Furchungszellen fur die Organbildung des Em-
bryo. Arch. f. Mic. Anat., 42.
Ishikawa, A.

'92 Studies in Reproductive elements. I. Spermatogenesis, Oogenesis and Fer-
tilization in Diaptomus sp. jour. Coll. Sc. Japan, Vol. 5-
Jennings, H. S.

'96 The Early Development of Asplanchna herrickii. Bull. Mus. Comp. Zoo.,

XXX., i.

'04 Physical Imitations of the activities of Amoeba. Amer. Nat., XXXVIII.
Kofoid,C. A.

'94 On some Laws of Cleavage in Limax. (Preliminary.) Proc. Am. Acad.

Arts and Sc., V., 29.
Korscheldt & Heider.

'o2-'o3 Lehrbuch der vergleichenden Entwicklungsgeschichte der wirbellosen

Thiere. Allegemeiner Theil., ]ena.
Kroyer, H.

'37-8 Om Snyltekrebsene, isaer med. Hensyn til Dansk. Faune. Naturhistorisk

Tidsskrift I. and II.
'63 Hidrag til Kundskab om Snyltekrebsene. Naturhistorisk Tidsskrift, Tredie

Raekke. Anclet Hind., pp. 75-426.
Labbe, A.

'04 Sur la formation des tetrades et les divisions maturatives dans le testicule du
Hornard. C. R. Acad. Sc., Paris, Tome 138, p. 96.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 79

Lee, A. B.

'05 The Microtomist's Vade-Mecum. 6th Ed., Phila., '05.
Lerat, Paul.

'02 La premiere cinese de maturation dans 1'Ovogenese et la spermatogenese du

Cyclops strenuus. Note preliminaire. Anat. Anz., 21 Bd., p. 407.
Lillie, F. R.

'95 The Embryology of the Unionidre. A Study in Cell Lineage. Jour.

Morph., 10.

'01 The Organization of the Egg of Unio. Jour. Morph., XVII.
Mathews, A. P.

'98 A Contribution to the Chemistry of Cytological Staining. Am. Jour, of

Phys., I., No. 4.
Mann, G.

'06 Chemistry of the Pretoids.
Mark, E. L.

'81 Maturation, Fecundation and Segmentation of Limax Campestris. Bull.

Mus. Comp. Zoo. Harvard College, Vol. VI., No. 12.
Mead, A. D.

'97 The Early Development of Marine Annelids. Jour. Morph., XIII.
Meyer, E.

'oi Studien iiber den Korperbau der Anneliden. V. Das Mesoderm der Ringel-

wiirmer. Mitt. Zoo. Sta. Neapel, 14 Bd.
Morgan, T. H.

'93 Experimental Studies on Echinoderm Eggs. Anat. Anz., 9.

'96 The Production of Artificial Astrospheres. Arch. Entwickelungsmechanik,

III.
Montgomery, T. H.

'98 The Spermatogenesis of Pentatoma. Zoo. lahrb., XII.

'99 Comparative Cytological Studies with special reference to the Morphology of

the Nucleolus. Jour. Morph., XV.

'oo The Spermatogenesis of Peripatus. Zoo. Jahrb., XIV.
Nelson, J. N.

'04 The Early Development of Dinophilus. Pro. Acad. Natl. Sc. Phila., Oct.,

'04.
Nusbaum, and Schreiber, W.

'98 Beitrage zur Kenntnis der sogen. Ruckenorgane der Crustaceenembryonen.

Bio. Centralbl., i8Bd.,p. 736.
Norman, W. W.

'96 Segmentation of the Nucleus without Segmentation of the Protoplasm.

Archiv. f. Entwickelungsmech., III.
Pauli, W.

'02 Allgemeine Physiko-Chemie der Zellen und Gewebe. Ergebnisse der

Physiologic Wiesbaden, I Jahr., I. Abt., Bio-Chemie.
Pedaschenko, D.

'93 Sur la segmentation de 1'oeuf et la formation des feuillets embryonnaires chez
la Lernae branchialis. (Preliminary) Revue Sc. N. Peterbourg Tome 37.
'99 Embryonalentwickelung und Metamorphose von Lernae branchialis, L.
Trav. Soc. Nat. Petersbourg, Vol. 26.

8O J. F. McCLENDON.

Pfliiger, E.

'84 Ueber die Einwirkung der Schwerkraft und anderer Bedingungen auf die

Richtung der Zelltheilung. Arch. Ges. Physiol., 34.
Rath, 0. Vom.

'92 Zur Kenntniss der Spermatogenese von Gryllotalpa vulgaris. Arch. Miks.

Anat. , 40 Bd. , p. 103
Reed, Margaret.

'05 Formation of the interior cells in the segmentation of the Frog's egg. Biol.

Bull., Feb., 1905, Vol. 8, No. 3.
Reinke, F.

'oo Ueber den Mitotischen Druck. Arch. f. Entwickelungsmech., IX.
Rhumbler, L.

'96 Versuch einer mechanischen Erklarung der indirekten Zell und Kerntheilung.

I. Die Cytokinese. Arch. Entwickelungsmechanik Organ., 3.
'05 Zur Theorie der Oberflachenkrafte der Amoben. Zeit. Wiss. Zoo., 83.
Robert, A.

'03 Recherches sur le development des Troques. Arch. Zoo. Exper. Serie X.
Roux.

'85 Beitr. z. Entw-Mech. d. Embryo. III. Bresl. Aerzth. Zeitschr., 1885,

Ges. Abb.., II. Bd., No. 20.
'93 Ueber Mosaikarbeit und neuere Entwicklungshypothesen. Anat. Heftes,

1893, Ges. Abh., II. Bd., No. 27.
Ruckert, J.

'94 Zur Ereifung bei Copepoden. Anat. Hefte, I Abth., 4 Bd., p. 261.

Ueber das Selbstandigbleiben der vaterlichen und miitterlichen Kernsubstanz
wahrend des ersten Entwickelung des befruchteten Cyclops Eis. Arch.
Mikr. Anat., 45 Bd., p. 339.
'95, a Zur Kenntnis des Befruchtungsvorganges. Sitz. Ber. Akad. Miinchen,

25 Bd.

'95, b Zur Befruchtung von Cyclops strenuus. Anat. Anz., 10 Bd., p. 708.
'96 Nochmals zur Reductionsfrage. Arch. Mikr. Anat., 47 Bd., p. 386.
Schimkewitz, W.

'96 Studien iiber parasitische Copepoden. Zeit. Wiss. Zoo., 2, 61 Bd., p. 339.
'99 Einige Worte iiber die Entwickelung der parasitischen Copepoden. Z. An-

zeiger, 22 Bd., p. 14.
Schlapfer.

'05 Eine physikalische Erklarung der acromatischen Spindelfigure, etc. Arch.

Entw. Mech., 19, 1905.
Steuer, A

'03 Mytilicola intestinalis n. g., n. sp. Arb. Z. Ins. Wien., 15 Bd.
Tread well, A. L.

'01 The Cytogeny of Pdarke obscura. Jour. Morph., 17, No. 3.
Urbanowitz, F.

'85 Beitrage zur Entwickelungsgeschichte der Copepoden. Kosmos Lemberg 10
Jahrg. & Berichten VVarsch. Universitat, 1885, and Arch. Slav. Biol., Tome
I, p. 663 (Review, '86).
Wagner, J.

'93 Einige Betrachtungen iiber die Bildung, der Keimblatter, der Dotterzellen und
der Embryonalhiillen bei Arthropoden. Bio Centralbl., 14 Bd., p. 361.

ON THE DEVELOPMENT OF PARASITIC COPEPODS. 8 I

Weismann & Ischikawa.

'88 Ueber die Bildungder Richtungskorper bei thierischen Eiern. Ber. Nat. Ges.

Freibourg, 3 Bd.
'88, a Weitere Untersuchungen zum Zahlengesetz der Richtungskorper. Z.

Jahrb. Morph. Abth., 3 Bd.
Wheeler, Wm. M.

'93 A contribution to Insect Embryology. Jour. Morph., Vol. 8.

'97 The Maturation, Fecundation and early Cleavage in Myzostoma. Arch, de

Biol., XV.
Wierezjski, A.

'05 Embryologie von Physa fontinalis. Zeit. Wiss. Zoo., 83.
Wilson, C. B.

'05 New Species of Parasitic Copepods from the Massachusetts Coast. Pro. Bio-
logical Soc. Washington.

05, a North American Parasitic Copepods belonging to the Family Caligidre. Pt. I.
The Caliginse. Pro. U. S. Nat. Museum, XXVIII., pp. 479-672, pi.
V. XXIX.
Wilson, E. B.

'92 The Cell Lineage of Nereis. Jour. Morph., VI.

'95 On Cleavage and Mosaic-work. Appendix to Crampton. Arch. f. Entw

Mech., 3.
Wright, R. R.

'83 Notes on American Parasitic Copepoda, I. Proc. Canad. Ins. (2), Vol. I.,

P- 243-
Ziegler, H. E.

'94 Ueber Furchung unter Pressung. Verb. Anat. Gessellsch. 8 Vers. Strass-

burg, 1894 (Anat. Anz. Supple.).
'98 Ueber den derzeitigen Stand der Colomfrage. Verh. der deutsch. Zoo. Ges.,

VIII.

82 J. F. McCLENDON.

PLATE II.

(Figs. 19-24, Lamargus muricatus Kroyer.}

FIG. 19. First cleavage spindle, prophase.

FIG. 20. First cleavage spindle, metaphase.

FIG. 21. Early prophase of second cleavage. To the right, below, is a highly
magnified section of the centrosomes in which the centrodesmus and nucleus of the
yolk cell are shown.

FIG. 22. A little later prophase of the same. In the protoplasmic cell the nuclear
membrane has begun to dissolve.

FIG. 23. Later prophase showing the elongation of the nucleus of the yolk cell.
Viewed from the animal pole.

FIG. 24. Late prophase (the protoplasmic cell is in the metaphase) viewed from
the vegetal pole.

figs. 25-30, The Dichelestid. All eggs viewed from vegetal pole, )

FIG. 25. Anaphase of the first cleavage (fixation poor?).

FIG. 26. Two cell stage. The protoplasmic cell is in the anaphase of the second
cleavage.

FIG. 27. Anaphase of the second cleavage viewed from the vegetal pole (the
protoplasmic cell is in the telophase).

FIG. 28. Prophase of third cleavage.

FIG. 29. Late prophase of fourth cleavage a 4 - 2 , 3 4 - 2 and c 4 - 2 are almost completely
hidden by cells lying over them.

FIG. 30. Spindle in the yolk cell, metaphase of fourth cleavage, from two con-
secutive sections and magnified more highly than Fig. 29. Stained with safranin-
gentian-violet. The distinctness of the "centrosomes" is exaggerated. The dark
granules on the astral rays of the sphere to the right are lumps of hyaloplasm. The
delicate network of hyaloplasm between the yolk spherules is represented by dotted
lines but the yolk itself is not shown.

BIOLOGICAL BULLETIN, VOL. XII

19

PLATE II

84 J. F. McCLENDON.

PLATE III.

(In Figs. 31, 32 and 33 the derivatives of a, b, c and d are separated by heavy
lines. )

FIG. 31. Sixteen cell stage. The cells of the animal-pole-side are shown by dotted
lines, a 3 - 1 , b 5 - 1 and b ^- 2 are in the metaphase of the fifth cleavage.

FIG. 32. Thirty two cell stage. The cells of the animal-pole-side are shown by
spaced lines.

FIG. 33. Sixty two cell stage. The primary entoblast, d 6 - 1 , and the primary
germ cell, d 6 - 2 , have not begun the sixth cleavage while fz' 7 - 10 , d 1 - 11 and d 1 ^ 2 are in
the metaphase or anaphase of the seventh.

[Figs. 34-36, The Dichelestid. Both embryos seen from ventral (vegefal) side.}

FIG. 34. A later stage than the one shown in Fig. 36, B, Plate VI. The pri-
mary entoblast cell is dividing. The primary germ cell (,v } has grown to large size
and the blastoderm is beginning to grow over it. At the sides of the figure some
mesoblast cells have been turned under the rim of the blastopore (/ 2 + 3 ).

FIG. 35. A later stage than Fig. 34. The entoblast is in the telophase of the second
division. The yolk is cut through completely by both divisions of the entoblast. The
primary germ cell has divided (g) and the blastoderm has grown over it.

FIG. 36. Hemispherical eggs from the ends of egg strings.

A. Dorsal view.

B. Anterior view.

C. Lateral view of stage in which the blastoderm (stippled) has covered half the
yolk ; the entoblast nuclei are stippled heavily.

(Figs. 4956, Panda i us simialus. )

FIG. 49. Sagittal section of gastrula just before the closure of the blastopore (/;).
^= first polar body. ;/> = acell of the post nauplius mesoblast. a- is taken from
another section of the same series and shows a small cell budded off into the yolk
from an ectoderm cell, g is from one of the same series of sections near the median
line, and represents a germ cell in its relation to the ectoderm.

FIG. 50. Part of a median cross-section of an embryo of stage C (Fig. 44). The
section passes through two germ cells (g) a muscle cell (in) and two entoblast
nuclei. The thickened portion of the ectoderm on the ventral side is the gang! ionic
rudiment of the second antenna.

FIG. 51. The nauplius just hatched. The ventral aspect is shown in the left, the
dorsal in the right half of the figure. The median eye is seen at am, and the stomo-
daeum at a. The rudiments of the post nauplius ganglia (w 5 - 10 ) and appendages
(w r '- in ) are clearly differentiated. The entoblasts (en) are still scattered through
the yolk.

FiG. 52. Section of one of the four germ cells of the nauplius showing the sixteen
chromosomes.

FIG. 53. Enlarged view of section of divided cell from rim of blastopore in Fig.
40. All the chromosomes are not included in the section.

FIG. 54. Prophase from same region.

FiG. 55. Telophase.

FIG. 56. Cross-section through germ cells, ventral ganglia, and ectoderm of nauplius
twenty-four hours after hatching.

BIOLOGICAL BULLETIN, VOL. XII

PLATE III

86 J. F. McCLENDON.

PLATE IV.

(Pandarus sinitatits Verrill. f'Jgs. 37 JQ, viewed from the vegetal pole.}

FIG. 37. Shows the metaphase of the first cleavage of the primary ectoblast cell.
To the right below is a more highly magnified view of the spindle of the same, in the
prophase. g primary germ cell.

FIG. 38. Shows the prophase of the second cleavage of the entoblast. The blasto-
derm is growing over the primary germ cell (,") Some mesoblast cells have been
turned under the rim of the blastopore at the sides of the figure.

FIG. 39. About the same stage as Fig. 38, but the entoblast has not begun its
second cleavage and the primary germ cell (g) is in the metaphase of division.

[In Figs. 40-42 half of the dorsal (animal) side of the egg is shown to the left
and half of the ventral (vegetal) side to the right.]

FIG. 40. Just before closure of the blastopore (6). X= small cell budding into
the yolk. An = mesoblastic rudiment of the first antenna. An 2 = mesoblastic rudi-
ment of second antenna. Md = mesoblastic rudiment of the mandible. The germ
cells are shown (stippled) beneath the ectoderm they have separated from one
another.

FIG. 41. Stage A. Just after closure of the blastopore. The dorsal and ventral
ectoderm is omitted save in the anterior and posterior portions, t, M = mesoblast
(muscle?) cells, just beneath the ectoderm. The germ cells have divided.

FIG. 42. Stage A a little later than 41. The eight stippled rods are the spindles
of the entoblast. The other stippled areas are mesoblast.

BIOLOGICAL BULLETIN, VOL. XII

37

"Tr&'>l'5i%- o -o., ,o

U O () _- o -A" O ",.: ... o 6 '

7

- w,^ . T/TS *> : -Ti ' -O ,

:2;^:?:^f

A'a .. -- : o -;l/ - -,'';?.- - : .?1

88 J. F. McCLENDON.

PLATE V.

{Pandarus siniiatus Verrill.}

(One half of each figure shows the dorsal, the other half the ventral aspect. In
Figs. 43 and 45 the ventral aspect is to the right and in the remaining figures to the
left.)

FIG. 43. Staged. The rudiments of the ganglia, w'- 4 , are shown. The strippled
rods are spindles of entoblast nuclei.

FIG. 44. Stage C. At x are minute (mesoblast?) cells in the yolk. Some muscles
in the right side of the figure are beginning to elongate. Over the ganglion of the
second antenna is a cluster of mesoblast cells of unknown history.

FIG. 45. Stage D. The appendages are beginning to bud out. The ganglia have
become connected by thickened ectoderm (outlined by a dotted line). The muscle-
cells are beginning to elongate radially, and above them the ectoderm cells are
elongated in the opposite direction (some of them are represented by dotted outlines,
ec). The entoblast is not represented, but a few small dark cells are shown in the
yolk.

FIG. 46. Stage D, later than in Fig. 45. The ectoderm thickening to form the
stomadceum is shown at o.

FIG. 47. The nauplius a short time before hatching.

FIG. 48. The nauplius just before hatching. In the left half of the figure the
ganglia and mesoblastic rudiments of the post nauplius appendages are shown
(stippled) in process of formation.

BIOLOGICAL BULLETIN, VOL. XII

STUDIES ON THE RELATION BETWEEN AMITOSIS

AND MITOSIS.

I. DEVELOPMENT OF THE OVARIES AND OOGENESIS IN

MONIEZIA.

C. M. CHILD.

I. INTRODUCTION.

Some years ago, during an examination of certain abnormali-
ties l in Moniesia, the apparent infrequency or total absence of any
evidence of mitosis even in regions where rapid growth was taking
place attracted my attention. Further investigation showed what
various authors had already noted, viz., that anything even re-
motely resembling mitotic divisions was either entirely absent or
extremely rare in most of the growing regions. Nevertheless, a
rapid multiplication of nuclei was occurring in these regions as
could readily be determined merely by the examination of sec-
tions of successive proglottids or more exactly by counting
nuclei in well-defined corresponding regions of different proglot-
tids. It was soon possible to establish, beyond a doubt, the fact
that the characteristic form of nuclear division was amitotic, not
mitotic, no single case of mitosis ever having been seen in most
parts of the body.

Naturally the next step was to determine whether the devel-
opment of the germ cells followed the same course. The present
and following papers are devoted to a consideration of the results
of these investigations.

The species used are Moniezia expansa (Blanchard) and Mon-
tezia pianissimo. (Stiles and H assail) both tapeworms of the
sheep, which may be obtained in great abundance from the
Chicago Stockyards at certain seasons. The work was begun
without any preconceived opinions, but as it progressed it be-
came evident that the facts could not be made to agree with the
views commonly accepted among cytologists. For this reason

'Child, "Abnormalities in the Cestode Moniezia Expansa, I. and II.," BlOL.
BULL., Vol. I., Nos. 5 and 6, 1900; III., ibid., Vol. III., Nos. 3 and 4, 1902.

89

9O C. M. CHILD.

the utmost care as regards observations and conclusions has been
taken and publication has been delayed from year to year to per-
mit the examination of new material. In 1904 a brief account
of amitosis in the early stages was published. 1 Since that time,
however, a large amount of new material has been prepared and
much of the old reexamined, but without essentially altering my
conclusions.

Because of the general significance of the data to be described,
and in order to forestall possible objections, it has seemed advis-
able to give in some detail the methods of preparation and pro-
cedure. Fresh material was fixed in May or June of four dif-
ferent years. In all, except the first lot of material, the animals
were fixed within five minutes after the sheep had been killed.
The first material was fixed about two hours after removal from
the intestine of the newly killed sheep, but all specimens were
alive and apparently in good condition when fixed, and results
showed that this material was as satisfactory as that fixed imme-
diately after removal from the host.

A considerable variety of fixing agents was employed in
order to discover the best possible fixation and to eliminate the
possible effect of particular fixing agents. The fluids used in-
clude Hermann's, Chichkoff's, Gilson's, Perenyi's, Merkel's,
HNO five per cent., HNO two per cent., HNO two per cent, for
one or two minutes followed by Merkel for twenty-four hours or
more, aqueous saturated solution of sublimate, sublimate with
one per cent, acetic acid and Graf's chrom-oxalic mixture.
While certain methods of fixation proved more satisfactory than
others the results obtained were essentially identical in all cases
so far as the points in question were involved. In the work of
the later years Hermann's fluids, chrom-oxalic and aqueous sat-
urated sublimate, were most frequently used as they had been
shown to be satisfactory. The chrom-oxalic does not preserve
the delicate cytoplasmic structures in the young testes and ova-
ries as well as does the sublimate, but the nuclear structures are
equally clear in both cases.

Various methods of staining were also employed, and while
the essential features are visible after almost any fairly good

1 Child, " Amitosis in Moniezia," Anat. Am., 1904.

THE RELATION BETWEEN AMITOSIS AND MITOSIS. QI

nuclear stain, Heidenhain's iron-haematoxylin was most com-
monly used because of its well-known sharp definition. Un-
doubtedly this stain is of little or no value for determining
differences of chemical composition but it is certainly unsurpassed
for bringing out physical differences. Almost the whole body
of one worm from the scolex to the region where the uterus
contained late cleavage stages was sectioned. From other chains
pieces containing three or four proglottids were taken at intervals
of from ten to forty proglottids, throughout the body. Pieces
separated by thirty or forty or even a larger number of pro-
glottids give stages so near together that nothing is lost, but
since it was necessary to make certain that mitosis did not occur
periodically in certain regions the intervals between pieces were
made much shorter in certain cases. The pieces intervening
between those sectioned were numbered and kept and frequently
some of them were sectioned later when a larger number of
cases of certain stages was needed or some point remained to be
settled. Every piece sectioned included at least two complete
proglottids and usually more, in the anterior regions often fifteen
or twenty. Sections were cut 35 // thick.

New material was sectioned and examined in four different
years and each year the old material was reexamined. Some
four hundred camera drawings of nuclei or groups dividing
amitotically have been made and in all cases the greatest care
has been taken to select those cases which were most clear and
convincing. Cases which might possibly be interpreted as
amitosis were recorded only when they formed part of a region
included in a drawing made for other reasons. The cases
recorded by special drawings were usually those which it seemed
impossible to interpret in any other way. Many of the cases
from which drawings were made were examined by other persons
and their interpretation agreed with my own. The observations
were 'all made with a 2 mm. oil immersion lens. With this
power some of the divisions may be seen with almost diagram-
matic clearness provided fixation and staining are satisfactory.

So far as I am aware I have employed every possible means
to establish my observations and to eliminate errors. Amitotic
division is not as readily distinguished as mitosis, for there are

p2 C. M. CHILD.

no visible characteristic stages of preparation and reconstruction
and no clearly visible chromosomes and spindles. The dividing
nucleus usually does not stain differently from any other, and
after division there is in many cases no demonstrative evidence
that the nuclei have arisen by division. Certain critical stages in
the division must be found, viz., those in which separation is just
beginning and the two parts are manifestly connected. More-
over, such stages must be found frequently in order to establish
the presumption that anything more than an abnormal or perhaps
a degenerative process is involved. My observations have ful-
filled these conditions. Thousands of cases of the critical stages
have been observed. Cases more or less similar to every case
figured in the drawings have been observed repeatedly. As re-
gards a possible failure to recognize mitoses when they occur it
should perhaps be said that perfectly distinct and characteristic
mitoses do occur at certain times and places and that there is no
chance for confusion of these with anything else : chromosomes,
spindle, centrosome, equatorial plate, division of the chromo-
somes, etc., all are visible and instantly recognizable. In fact it
has been possible in the germ cells of one species to establish
the number of chromosomes with a considerable degree of ac-
curacy. It is very certain that the form of division which I
have designated amitosis in these species cannot be interpreted as
mitosis indistinctly visible or of peculiar form. All drawings are
made from camera drawings and schematization of the actual
cases of division and "improvement" of the camera-drawings
have been avoided as far as possible.

Many of the figures are schematic in that non-essentials are
omitted and simple methods of representation are employed, but
every case of amitosis figured is as nearly like the observed case
as it was possible to make it after the most careful examination.
No attempt has been made to represent the parenchymal sub-
stance in which the cells lie in the earlier stages of development,
and when, as is often the case, no well-marked area of cytoplasm
appears about the nucleus, no cell-boundaries are indicated and
the nucleus alone is drawn. As a matter of fact the distinction
between cytoplasm and parenchymal substance, at least during
the earlier stages, is not nearly as sharp as the figures indicate, for

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 93

no visible cell membrane is present. For the sake of simplicity,
however, the approximate area of the cytoplasm about the nucleus
is indicated by a line.

To give a detailed description of the development of the repro-
ductive organs in Monicsia is beyond the present purpose. The
morphological features concern us only secondarily and will be
considered only so far as may be necessary for the understanding
of other matters.

II. THE NUCLEUS AND THE TYPES OF DIVISION.

The nuclei of most of the somatic structures and of the young
germ cells do not differ widely in appearance. The nucleus con-
tains a deeply staining " nucleolus " which appears to be at least
in large part chromatic in composition and might perhaps more
properly be called a karyosome. For the present, however, I
prefer to use the term nucleolus. After the usual degree of
extraction this is very commonly the only element stained in the
nucleus which appears entirely homogenous except for this body.
With less extraction other granules are visible scattered here and
there through the nucleus, but a distinct reticular structure does
not appear. This type of nucleus is shown in most of the figures
of earlier stages of ovarian development. In various figures there
are cases where the nucleus shows a few small granules in addi-
tion to the nucleolus (Figs. 8,A,&,d, e, etc., 13, C, etc.). Nuclei
are frequently found with two nucleoli both of which may be of
equal size (Figs. 8, A,/ ; 14, A, a, etc.) or they maybe unequal
(Figs. 14, D ; 15, />, a; 16, b, etc.). The question as to
whether the two nucleoli always arise by the division of one it
has been impossible to settle. Sometimes (Fig. 15, A, a) a
minute nucleolus is found apparently in contact with a much
larger one and occasionally (Figs. 7, c ; 9, B] two nucleoli ap-
parently connected by a strand of stained substance are seen.
On the other hand, in many cases the two nucleoli of very differ-
ent size are widely separated (Fig. 8, A, a and g] as if one were
arising de novo. It seems impossible to decide such questions
as this until our methods of study of the cell are greatly im-
proved. From my own observations I should conclude pro-
visionally that both methods of origin exist. It is certain, how-

94 C. M. CHILD.

ever, that elongation and division of the nucleolus is not a typical
feature of amitosis here. The nucleolus is always spherical or
nearly so.

When amitosis occurs each part of the dividing nucleus usually
very probably always -- contains a nucleolus. Occasionally
it cannot be found but its apparent absence may be due to too
great extraction or to loss from the section. It is probable that
in Moniczia the formation of two nucleoli in a nucleus will be
followed sooner or later by division, though division need not
necessarily occur at once.

The process of amitosis is simple as far as visible features are
concerned, but various apparent modifications occur. In some
cases a constriction in the nuclear membrane appears, extending
about the whole circumference of the nucleus or limited to one
side (Figs. 3, b, c, d, e; 5, a, b ; 7, e; 13, A, ., etc.). Fre-
quently there is a faint extension from the deepest part of the
constriction partly or wholly across the nucleus (Figs. 3, b, c ;
5, />, c ; 7, e, etc.). In other cases the formation of a nuclear
plate or membrane across some part of the nucleus takes place
before anything more than a very slight constriction appears in
the old nuclear membrane (Figs. 7, /, g ; 8, A, e ; 13, C ;
13, D, etc.). In such cases it is possible by careful focusing to
follow the new membrane across the whole diameter of the
nucleus.

The method of separation of the products of division also varies
in accordance with the differences in the earlier stages. The
nuclei sometimes separate from one side (Figs. 3, c ; 5, c;

7, e ; 15, B, b, etc.), such cases being presumably the result of
formation of the constriction from one side. Frequently also the
constriction appears to deepen uniformly about the whole cir-
cumference (Fig. 3, c/, e, etc.), separation being completed at or
near the middle. In those cases where a distinct nuclear plate
or partition forms across the whole nucleus separation seems to
occur simultaneously or nearly so over the whole surface (Figs.

8, A, c, i ; 9, A, b, etc.). These cases are perhaps the most
demonstrative of all, for the flattened surfaces of the two nuclei
and occasionally their contact at one margin (Figs. 8, A, c ;

9, A, b) leave no room for doubt that division has actually

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 95

occurred. Occasionally the margins of the two nuclei are the
last portions to separate (Fig. 16, a). The flattened surfaces are
undoubtedly soon lost after division. Some of the cases of sepa-
ration from one side show very clearly that the separated parts
of the surface begin to become convex before separation is com-
pleted (Fig. 5, c).

The division of the cytoplasm also varies to some extent, but
is more difficult to observe since the cytoplasm is usually with-
out any sharply defined boundary. In some cases a constriction
of the cytoplasm follows the constriction of the nucleus (Figs. 3,
a ; "J,f, i ; 9, A, d ; 29, /;, etc.), and in others the nuclear divi-
sion may be completed before the cytoplasm shows any trace of
constriction. (Figs. 7, d, g ; 8, A, c, c, i; 11, A, b ; 14, A, b,
etc.). Occasionally nuclear and cytoplasmic division are appar-
ently almost simultaneous (Figs. 7, h, i ; 8, A, //, etc.). A few
cases have been noted where the division of the cytoplasm is
"endogenous." Such a case is shown in Fig. \i,A,a; and
another in Fig. 21, a. Such cases as Figs. 2, a ; 5, d, and 29, a,
seem to indicate that a " cell-plate " may sometimes be formed
across the cytoplasm, though after the appearance of a cell-plate
it is of course impossible to be absolutely certain that the two
cells are the product of a division. But the fact that the nuclei
with their surrounding areas of cytoplasm are usually isolated
renders it probable that such cases represent division.

Not infrequently one part of the nucleus stains more deeply
than the other. In such cases the stain is uniformly distributed
in each part but the boundary line between the darker and lighter
portion is sharp. Figs. 7, b, 14, B, a, and 16, c, show cases of
this kind. This difference in staining is of some importance as
indicating that there are differences of some sort in the two
regions and probably also that the two are functioning more or
less independently.

Occasionally a case of what appears to be an " endogenous"
division is found (Fig. 13, E). In such cases the nuclear plate or
new membrane does not appear as a simple partition but two dis-
tinct membranes more or less convex toward each other are
formed, while the old nuclear membrane appears to surround the
whole. Some cases of this sort of almost diagrammatic clearness

96 C. M. CHILD.

have been observed and in one of the turbellaria, a form with much
larger nuclei, I have recently seen something similar and am there-
fore inclined to believe that cases of this sort actually do occur.
In later stages the old membrane seems to disappear leaving two
separate nuclei.

The occurrence of amitosis is probably to be regarded as the
result of the establishment of more or less independent functional
regions in different parts of the nucleus and the consequent forma-
tion of a membrane about each of these. The details of the proc-
ess must of course differ according to conditions so that many
different forms of amitosis may occur, all due primarily to the
same factors. The "endogenous" method of division is not
perhaps so widely different from the others as might appear.
The old nucleus is so large or the new " functional " nuclei so
small that parts of the old nucleus are left out when the new
membranes are formed. Ordinarily the new membranes are
formed in direct contact, here they are merely formed separately.

As regards the process of mitosis but little need be said here.
The maturation mitoses will be described later but the tissue
mitoses and those in the germ mother-cells differ more or less
from these. In preparation for mitosis the nucleus stains more
deeply and traces of a spireme are sometimes seen but the prepa-
ration first becomes readily recognizable when the chromosomes
are formed (Fig. 8, A). The largest number of chromosomes
counted is fourteen but accuracy is out of the question here.
Fig. 8, B, shows a case where twelve were clearly seen. A
spindle in metaphase is seen in Fig. 8, A ; it is usually possible
to distinguish dark bodies at the poles but much depends on the
degree of extraction ; the spindle fibers are very delicate and
astral radiations are not certainly distinguishable. A later stage
is seen in a, Fig. 5. Any stage of the division from the forma-
tion of the chromosomes to and including the late anaphase is as
readily recognizable in the sections as in the figures.

The occurrence of these two forms of division side by side is
an indubitable fact, but further data will be given in following
papers. There can be little doubt that each form of division is a
reaction to special conditions. Certain observations of my own
indicate that amitosis occurs more frequently in very rapid and

THE RELATION BETWEEN AMITOSIS AND MITOSIS.

97

mitosis more frequently in slower growth. But until other data
are described theoretical considerations are out of place.

III. THE EARLY STAGES OF THE FEMALE REPRODUCTIVE

ORGANS.

The ovary itself does not appear in the earliest stages of devel-
opment of the female organs. It is in fact formed only after a
considerable portion of the ducts has differentiated. The earliest
visible stage in the development of the female organs is the
increase in the number of nuclei in a region immediately adjoin-
ing the longitudinal nephridial canals. Viewed from the surface
it appears as in Fig. i, A. A somewhat later stage is shown in
Fig. i, B. The nuclei are more closely packed together here

than elsewhere consequently the region stains more deeply. This
region differentiates later on into the middle region of the repro-
ductive ducts. Sections through this region at this stage show a
'number of nuclei often surrounded by more or less cytoplasm
without definite boundary. These nuclei are indistinguishable
from the other parenchymal nuclei except that many of them are
smaller. Parenchymal fibers can often be traced from the cells
and the region is not marked off in any way from the parenchyma.
But the most interesting point forpresent purposes is the appar-
ent absence of mitosis in these regions. Although the writer has
examined hundreds of sections of these early stages he has never
seen a single case of mitosis. Yet it is very evident that an ex-
ceedingly rapid multiplication of nuclei is taking place, for the
size of the area and the number of nuclei increases rapidly with

9 8

C. M. CHILD.

increasing distance from the scolex. Figs. 2 and 3 show small
portions of this proliferating region. Fig. 2 is from Jlf. planis-
siuia, Fig. 3 from M. expansa. The cytoplasm about the nucleus

FIG. 2.

FIG. 3.

is often so slight in amount as to be almost'invisible. All the
cells lie free in the parenchyma. A number of nuclei in both
figures show various stages of amitosis and others have appar-

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 99

ently only recently separated. From this stage up to the period
when the ducts begin to assume definite form the appearance of
these regions is much the same. The course of development of
this area certainly favors the view that the proliferation is due to
some localized stimulus or condition rather than to anything in-
herent in the cells themselves. In the central portions of the
area the nuclei have divided so rapidly that they are very small ;
nearer the periphery they are larger and about the outside of the
area are nuclei of the same size as other parenchymal muclei.
Divisions likewise decrease in frequency from the center toward
the periphery. Fig. 2 shows this difference to some extent.
The nuclei near the center of the figure which represents approxi-
mately the center of the proliferating area are much smaller than
those about the periphery. The two large nuclei in the upper
right corner are about the size of typical parenchymal nuclei.
Fig. 3 is a smaller area from a somewhat later stage and entirely
within the proliferating region.

The later development of the ducts will be considered more
fully in connection with other somatic structures. It is typically
a process of continued amitotic division although in some indivi-
duals an occasional case of typical mitosis occurs.

IV. THE FORMATION OF THE OVARY.

From the region of its first appearance near the nephridial
canals (Fig. i) the proliferating area gradually extends somewhat
toward the median plane and toward one surface of the proglot-
tid known as the ventral surface. So far as can be determined
there is no appreciable migration of cells through the parenchyma ;
each part seems to be formed in si/u, the stimulus to proliferation
continually involving more of the parenchymal nuclei and extend-
ing in a more or less definite direction.

As the inner and ventral end of the proliferating area approaches
the inner layer of circular muscles it spreads out into a flat-
tened somewhat disc-like area exactly as if it had encountered
resistance to its growth in the original direction and so had be-
gun to spread out in other directions. Fig. 4 shows a stage
soon after the disc-like flattened terminal region has appeared.
This disc-like terminal portion indicated by o is the ovary. In

IOO

C. M. CHILD.

connection with it and indicated diagramatically in the figure is
the oviduct. The outer portion of the female duct, i. e., that
leading from the region of the nephridial canals, is still in an
early stage of development and the genital opening has not yet
appeared. This sequence in the formation of parts is rather
peculiar ; the first portion to appear becomes the middle region
of the ducts while inner and outer terminal portions, including
the ovary appear considerably later.

FIG. 4.

In its earliest stages the ovary does not differ very widely in
appearance from the early stages of the ducts. It consists merely
of a number of nuclei in the parenchyma surrounded by more or
less cytoplasm and undergoing frequent amitotic divisions. But
the divisions seem to be somewhat less frequent here than in the
regions of the ducts and the nuclei never become so reduced in
size. But one difference between the ovary and the early stages
of duct-formation exists : viz., the occasional occurrence of a case
of mitosis in the ovary. Fig. 5 illustrates very clearly the occur-
rence side by side of the two forms of division. It represents a
portion of a longitudinal section through the inner end of the
oviduct and the ovarian region. The smaller cells (od~} on the
left in the upper part of the figure with elongated cytoplasmic
areas represent the terminal portions of the oviduct and the
region between these and the muscles, which are represented by
small circles, the ovarian region. In this section only a few of
the cells involved in the formation of the ovary appear ; they are

THE RELATION BETWEEN AM1TOSIS AND MITOSIS.

IOI

d

FIG. 5.

mostly in other sections. On the right is a typical case of
mitosis in late anaphase ; this is the earliest case of mitosis seen
in the ovary ; c is a very distinct case of amitosis, b is a second
case. The smaller nuclei also show two cases of amitosis.
Whenever mitosis occurs in the developing ovary other nuclei
dividing amitotically are found near it.

V. GROWTH OF THE OVARY.

The further development of the ovary consists in increase in
size of the proliferating area and a little later the outgrowth
from its margins of finger-like follicles which elongate and give
the ovary its characteristic form. A follicular membrane differ-
entiates, apparently from the cytoplasm of the cells about the
periphery of the proliferating region, thus separating the ovary
from the general parenchyma. Fig. 6 shows a stage in the
later development of the ovary in which the follicles have attained
almost complete development.

102

C. M. CHILD.

The posterior portion of the disc -like proliferating area which
terminates the oviduct at the stage of Fig. 4 is indistinguishable
from other parts in the earlier stages but in the later stages shows
smaller cells and smaller follicles than the ovary and forms the
vitellarium. It is not shown in Fig. 6.

FIG. 6.

After the development of the ovary is completed nuclear di-
vision ceases and the mother cells enter upon the first stage of
their development as ova. As long as division continues, how-
ever, it is predominantly amitotic. The following figures taken
from various stages between the formation of the ovary and the
completion of its development in both species will show very
clearly the prevalence of this process.

During the growth of the ovary the parenchymal substance
remaining within the ovary gradually disappears. In many
cases, especially after chrom-oxalic, the cytoplasmic areas about
the nuclei are indistinct. The cytoplasm is usually not sharply
marked off from the parenchymal substance still present so that
the whole often appears as a syncytium. After sublimate fixa-
tion the cytoplasmic regions appear more distinct. Most of the
figures are drawn from sublimate preparations.

Fig. 7 is from a stage slightly later than Fig. 5, before the
separate follicles have appeared and before an ovarian membrane
has formed. The width of the figure represents the whole

THE RELATION BETWEEN AMITOSIS AND MITOSIS.

103

width of the young ovary. In the five nuclei d, e,f, g, /there
can be no doubt regarding the occurrence of amitosis. Each
one of these nuclei was examined at all levels and in such cases
as f and g the membrane can be followed through the whole
nucleus. The. case indicated by b is one of those often found
where the two parts of the nucleus stain differently. At f the
two nuclei have apparently recently separated for they are nearly
hemispherical in form, their flattened surfaces are parallel and
closely approximated and are not visibly covered by cytoplasm.

FIG. 7.

The two nuclei at / are apparently also the products of a recent
division. The cytoplasm is still continuous on the upper side
and the nuclei are flattened on opposing faces. The figure also
shows one case of mitosis.

Fig. 8, A, is from the same proglottid as Fig. 7. At this
stage muscle fibres are still visible passing directly through the
ovary and are indicated in the figure. It contains six cases of
mitosis, the largest number observed in any area of similar size.
But in this section are also two very clear cases of amitosis, (c
and z). The cases b, c, li and j are also undoubtedly amitoses
and several other nuclei in the figure are probably also dividing
amitotically. The number of mitoses in this section is of interest
since it exceeds so greatly the number seen in any/ other similar

IO4

C. M. CHILD.

case. Commonly section after section may be examined without
seeing a case of mitosis or occasionally one or two may be
found. In the ovary shown in Fig. 8, A, however, several con-
secutive sections showed frequent mitoses particularly in one
region of the ovary. Amitosis was common everywhere as

FIG. 8, A, 8, B.

usual. Discussion as to possible determining conditions is post-
poned for the present. Fig. 8, B, is a case of mitosis in which
twelve chromosomes were clearly visible probably not the
whole number. It was usually impossible to determine with
accuracy the number of chromosomes in these divisions.

Fig. g, A, is also taken from the same stage and across the
whole width of the ovary near its base. At least four perfectly
clear cases of amitosis (a, b, c, d] are present besides two other

THE RELATION BETWEEN AMITOSIS AND MITOSIS.

105

FIG. 9, A, 9, B.

doubtful cases. Fig. 9, B, is a cell from the same ovary with
two nucleoli apparently connected.

The Figs. 7-9 are taken from M. expansa. Since the course
of development is identical in both species so far as can be
observed, it is not necessary to duplicate all stages in two species,
but a few typical cells and cell-groups from the ovary of M.
pianissimo, are figured. Fig. 10, A, and 10, B, both show
characteristic cases of amitosis in the earliest stages of ovarian
development. Fig. 1 1, A, shows two cases and Fig. 1 1, B, one.
The case a in Fig. n, A, represents a rather interesting stage

II

12

C

D

13

FIGS. 10, A, 10, B, 11, A, n, B. FIG. 12, A-\2, C. FIG. 13, ^-13, E.

IO6 C. M. CHILD.

and one that possibly does not always occur. The nuclei have
separated and between them there is a lenticular space appar-
ently empty, but probably containing fluid in the living condi-
tion. This space is apparently still intracellular for the cytoplasm
is clearly continuous across its ends. In most cases of amitosis
the two nuclei separate at one end first so that no such space is
formed.

In Fig. 12, A and B, the same nucleus is shown at different
levels, A being the upper portion and B the lower. C is prob-
ably an early stage of amitosis. Both are from slightly later
stages of development than Fig. 11.

Cases from a still later stage are shown in Fig. 13, AE. The
chain from which Figs. 10-13 were taken showed fewer mitoses
in development of all organs than any other examined. Mitosis
was scarcely ever seen in the ovary. That individual differences
do exist in this respect can scarcely be doubted.

After the individual follicles begin to form, the divisions seem
to be more frequent in them, and especially near their tips, than
in the central portions of the ovary. Fig. 14, AD, show
groups with characteristic amitoses from the developing follicles
of J/. cxpansa. Fig. 15, A, and 15, B, are similar, the latter
showing a few nuclei from the extreme tip of the young follicle
not yet enclosed by a follicular membrane. In Fig. 16 the ter-
minal portion of another follicle is shown with three cases of
amitosis.

An early stage in the development of the follicles in M. pianis-
simo, is shown in Fig. 17. The numerous cases of amitosis are
clearly visible. At a in the left follicle is what appeared to be
a group of small nuclei. It may be a multiple amitosis or may
possibly represent a stage in reconstitution after mitotic division.
The large nuclei in two of the follicles are frequently found along
the axes of developing follicles. One of them (<) is apparently
dividing into three parts. The nuclear divisions which have been
described thus far are in reality the oogonial divisions. The ami-
totic divisions certainly constitute a normal feature in the history
of the ova, as there is no evidence that the nuclei which have
divided mitotically have a different fate from the others. The
relative frequency of mitoses varies not only in different chains

FIG. 18.

FIG. 20. D. 20.

-J'

FIG. 19, A, 19, B. FIGS. 19, C, 20,
FIG. 21. FIG. 22.

23B

FIG. 20, , 20, C.
FIG. 2?. ,4. 23. .#.

THE RELATION BETWEEN AMITOSIS AND MITOSIS. I Op

but in different proglottids. The chain in which mitosis was
almost never seen in the ovarian development produced appar-
ently as many eggs as others and these developed in the normal
manner in the uterus.

VI. THE GROWTH PERIOD, " SYNAPSIS " AND YOLK FORMATION.

Although the later stages in the formation of the ova are not
concerned directly with amitosis a brief description of these stages
is added in order to demonstrate that nuclei which have passed
through a long series of amitoses are quite capable of exhibiting
the characteristic phenomena of typical germ cells. These
stages in the development of the ova also appear to be identical
in both species. All the figures except 22, 23 and 30, AD, are
from M. cxpansa.

The oogonia at the end of the period of division are seen in
Fig. 1 8. There is little difference in appearance between them
and the dividing cells during ovarian development. The amount
of cytoplasm is perhaps slightly greater, but this difference is not
marked.

But now the nuclei undergo a sudden and remarkable change.
The only deeply staining portions of the nucleus up to this time
have been the nucleolus and frequently a few other granules
(Fig. 1 8). Now the nucleus develops rapidly a large amount of
chromatin. The earliest observed stages in this development are
shown in Fig. 19, Aig, C. This change in the nuclear sub-
stance is accompanied by a great increase in size of both nucleus
and cytoplasm. The chromatin soon shows itself in the form of
a typical spireme (Figs. 20, A-2O, E, 23, A, 23, B] which is com-
monly massed at one side of the nucleus (Figs. 20 A, 20, C, 2O,
E, 23, A, 23, E] and in most cases is visibly connected with the
nucleolus. So far as can be determined the spireme does not
appear to be formed from the substance of the nucleolus, since
the latter increases in size like the other elements of the cell.
This is clearly a typical case of what is commonly known as
synapsis.

'The appearance of these stages varies considerably with the
degree of extraction of the stain. If extraction is carried farther
than in Fig. 20 only small portions of the spireme or often only

HO C. M. CHILD.

the karyosome retain the stain and the presence of the spireme
would not be suspected. The writer is firmly convinced by long
experience with this stain that it is of little or no value in de-
termining differences of chemical constitution ; size, density, and
permeability seem to be the chief factors determining which parts
shall retain the stain.

The spireme appears first in the central portions of the ovary
adjoining the end of the oviduct and proceeds in all directions dis-
tally along the follicles. The follicle tips continue, however, to
divide amitotically for some time after the other portions of
the ovary have entered the spireme-stage. Fig. 21 shows the
tip of a follicle at this stage. In the same ovary the cells near
the oviduct were like those in Fig. 20. Fig 22 is a nearly longi-
tudinal section through the terminal portion of a follicle showing
the different stages at different levels. In all but one of the five
cases with spiremes most of the chromatin is outside the plane of
section. In this ovary all the cells have entered the spireme-
stage except those in the terminal region of the follicles and
division has already ceased there.

Together with the nuclear changes occur marked changes in
the cytoplasm. The amount of cytoplasm increases in marked
degree as the figures show. It also stains more deeply than
before. During synapsis the nucleus is usually more or less
asymmetrical in position. In many cases at least the greater
amount of cytoplasm is on that side of the nucleus against which
the spireme is massed. (Figs. 20, A, 20, C, 20, D, 20, E, 23, A,
23, B}. It cannot be stated positively that this is always the case
but it is certainly of frequent occurrence. After fixation in chrom-
oxalic a regional differentiation of the cytoplasm is visible in many
cases and is perhaps a characteristic feature. As noted above,
the spireme usually becomes massed at one side of the nucleus :
this is followed by the appearance in that part of the cytoplasm
nearest the spireme of an area staining more deeply than any
other part of the cytoplasm (Fig. 24, A]. If extraction is carried
too far this area is scarcely or not at all visible. It is compar-
able to certain of the differentiations which have been called yolk
nuclei in other eggs and its appearance is followed almost im-
mediately by the formation of yolk granules which in Moniczia

THE RELATION BERVVEEN A MITOSIS AND MITOSIS.

I I I

are contained in the egg-cell itself. The first yolk granules to
appear, however, are not confined to this region but are more or
less scattered (Fig. 24). The natural conclusion from the
sequence of events is that the change in nuclear condition is in
some way correlated with the cytoplasmic changes and since the
former precedes, that it is in some way responsible for the latter.
Nuclear changes connected with yolk-formation have been de-

24

FIGS. 24, 25, 26, 27, 28.

scribed by many authors but it is necessary to review the various
accounts. It seems probable these nuclear changes indicate an
alteration in the metabolic processes and that they are concerned
primarily with the increase of the cytoplasm and the deposition
of yolk.

Figs. 24-27 show successive stages in yolk-formation. The
granules formed first increase in size and others appear. Fusion

112 C. M. CHILD.

of the smaller to form larger masses also occurs frequently. Dur-
ing yolk-formation the nucleolus remains apparently unchanged,
but the spireme soon loses its deeply staining character and its
substance appears to spread throughout the nucleus in a more or
less recticular condition so that the nucleus resembles the nuclei
of most eggs before maturation.

It is almost impossible to reproduce the nuclear structure of
this period with any degree of exactness : moi cover, it is not cer-
tain how far the visible structure is characteristic of the living cell
and how far it is the product of fixation. An attempt has been
made, however, in Figs. 2427 to indicate the changes in nuclear
structure. Although the spireme of earlier stages has disappeared
the nucleus differs in appearance from the ovarian nuclei before
the growth period in that a reticulum is now visible while in the
earlier stages the nuclei were almost entirely homogenous in ap-
pearance except for the nucleolus and other granules. An
account of maturation and fertilization will be given in another
paper. At this time it need only be said that typical maturation
spindles appear, but that it has been impossible to follow the
process of chromosome-reduction. Fig. 28 shows the first polar
spindle with its enormous centrosomes.

VII. THE DEVELOPMENT OF THE VITELLARIUM.

The " yolk-gland " differentiates from the posterior portion of
the ovarian mass of proliferating cells. During the earlier stages
its cells are indistinguishable from those of the ovary, but later
they can be distinguished by their slightly smaller size. The
nuclear division is almost wholly amitotic. Occasionally, how-
ever, though less frequently than in the ovary, a case of mitosis
is seen. Division seems to be somewhat more rapid than in the
ovary, amitoses being commonly more numerous in a given area.
As in the ovary the nuclei of earlier stages are surrounded by very
little cytoplasm and lie in the parenchymal tissue. Fig. 29 rep-
resents a group of nuclei at the stage when the yolk-gland is first
distinguishable from the ovary. Numerous amitoses are visible.
Fig. 30 shows several cases of division from a later stage. The
process of yolk formation in these cells differs from that in the egg
cells. It is not preceded by any marked increase in size of the

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 113

nucleus or cytoplasm. No spireme has been observed but a
number of deeply staining granules appear in the nucleus in ad-
dition to the karyosome. The yolk appears first in the form of
small granules which increase in size and fuse, until the cell con-
tains a single large spherical mass of yolk and the greatly reduced
nucleus is flattened at one side. Successive stages of yolk-de-
velopment are shown in Figs. 31-35. These cells apparently

FIGS. 29-35.

arise from the same primordium as the ova but have become
specialized in the direction of yolk-production. The process of
yolk-formation is not identical with that in the egg but this dif-
ference is probably correlated with the extreme specialization.

VIII. CONCLUSION.

Extended discussion is better postponed until other data have
been presented, but the most important facts of this paper may
be briefly stated as follows.

114 C - M - CHILD.

Nuclear division in the development of the female reproductive
organs of Moniezia cxpansa and M. pianissimo, is predominantly
amitotic though typical mitoses occur, their frequency varying
apparently in different chains and in different proglottids.

After the long period of repeated amitotic division the nuclei
pass through the characteristic mitotic maturation divisions and
the cells form typical ova.

The process of amitosis consists in the formation of a con-
striction in some part of the nucleus or of a "nuclear plate" or
membrane across some portion of the nucleus and the separation
of the nuclei thus marked off and later of the cytoplasm about
each. Each part thus separated usually possesses a visible
nucleolus at the time of formation of the membrane.

The fact that the female germ-cells may arise by a long series
of divisions almost wholly amitotic is of considerable theoretical
importance. There is no room for doubt that the fate of a cell
may be the same, whether it divides mitotically or amitotically
during developmental stages. It is also very difficult to under-
stand how anything like individuality of the chromosomes can be
maintained in this case during the development of the germ-cell.

The two forms of division may occur side by side in the same
tissue and at the same stage, but their relative frequency may
vary in different individuals and in different proglottids. It seems
probable, therefore, and other data will confirm this conclusion,
that the form of division is determined by the conditions to which
the cell is subjected.

HULL ZOOLOGICAL LABORATORY,
UNIVERSITY OF CHICAGO,
August, 1906.

PROBABLE DIMORPHISM OF THE EGGS OF AN

ARANEAD. 1

THOS. H. MONTGOMERY, JR.

On comparing the eggs from cocoons raised in captivity of the
common spider, Theridinui tepidariorum, C. K., 2 I was struck by
the fact that eggs of the same age but from different cocoons
may be of distinctly different volumes. That is to say, all the
eggs of one cocoon may be larger or smaller than all the eggs
of another cocoon made by the same spider.

In this comparison only such egg batches were considered, of
which I had records as to the exact hour of oviposition. Further,
comparisons were made only of eggs preserved by the same fix-
ative, hardened in the same way, and preserved in the same
grade of alcohol ; in all cases the cocoons were opened and the
eggs dropped into the fixative. Then all the spiders came from
one locality, and all lived under the same conditions of captivity
in the month of August, 1906.

The following cases exhibited different volumes of the eggs of
successive cocoons made by particular spider individuals :

1S3.

i. One spider formed three successive cocoons, nos. 752 (pth
August), 818 (Hth August), and 872 (igth August). Every
egg of no. 752 (fixed at the age of twelve hours) was markedly
larger than any egg of no. 818 (fixed at the age of fifteen hours)
and no. 872 (fixed at the age of nineteen hours). The difference
in these egg sizes is shown in the accompanying figures.

Contributions from the Zoological Laboratory of the University of Texas, no. Si.
^Vide ray preceding paper in this journal, "The Oviposition, Cocooning and
Hatching of an Aranead, Theridium tepidariorum, C. Koch."

"5

Il6 THOS. MONTGOMERY, JR.

2. Another spider made two cocoons, nos. 725 (8th August)
and 821 (i5th August). All the eggs of no. 725 (fixed at the
age of two hours and twenty minutes) were markedly larger than
any of the eggs of no. 821 (fixed at the moment of oviposition).

3. Another spider made cocoons nos. 770 (iith August) and
905 (iSth August) ; the eggs of no. 770 (fixed at the age of 50
minutes) were all markedly smaller than the eggs of no. 904
(fixed at the age of 60^ hours).

4. A fourth spider furnished cocoons no. 726 (8th August) and
no. 850 (i7th August). The eggs of no. 726 (fixed at the age
of 3 y 2 hours) were all markedly smaller than those of no. 850
(fixed at the age of 3 hours).

The preceding series of cases show that successive cocoons
may have eggs of the same size ; or, and this is what more par-
ticularly interests us, that all the eggs of one may be larger or
smaller than all the eggs of another. It will be also noticed
(cases i and 4) that eggs of younger age may be larger than
eggs of maturer age. Indeed, there is probably no change in
the volume of a given egg from the time of oviposition up at
least to the time of appearance of the limbs ; accordingly, indi-
vidual growth of an egg is not a factor entering in to disturb our
conclusions as to these voluminal differences, since we are con-
sidering only stages antecedent to the appearance of limbs. 1

To estimate the comparative egg volumina I placed the eggs
side by side under the lens, and judged their difference ocularly,
always comparing the smallest of a larger batch with the largest
of a smaller batch. It would not be possible to estimate relative
volumes by determining the number required to fill a given
space, without first dissecting off the envelopes of each egg.

Now the variability of volume is usually of small amount in
any given cocoon, that is, all the eggs of a cocoon are large
or all small in most instances. But there are frequent excep-
tions to this. Thus one female produced three cocoons : nos.
722 (8th August), 800 ( 1 2th August) and 856 (i7th August),
all the eggs of no. 856 (fixed at the age of 26 hours) and

1 During oviposition the eggs are polygonal, but usually within a few minutes all
become rounded (slightly ovoidal), as has been noted by Balbiani : Memoires sur le
developpement des Araneides, 1873. Bibl. de /' Ecole des Hautes Etudes, T. 7.

PROBABLE DIMORPHISM OF THE EGGS OF AN ARANEAD. I I/

almost all the eggs of no. 800 (fixed at the same age) were
larger than the eggs of no. 722 (fixed at the age of I ^ hours),
but a few eggs of no. 800 were as small as the eggs of no. 722.
The same held for the successive cocoons of four other spiders.
Then in one cocoon collected in the wild state, not raised in cap-
tivity, there were 596 eggs (the largest number I have found in
any single cocoon) ; most of these could, with certainty, be
ranked as large eggs (about 478) while about 118 of the eggs
were clearly small eggs, but a few were intermediate in size be-
tween these two groups. The latter case is important in showing
that while intermediate sizes may occur between the large and
the small eggs, in the same cocoon or in successive cocoons, the
intermediates are very few in number compared with the ex-
tremes, a condition that would not occur in simple individual
variation.

The conclusions permitted by these observations are as follows :
This species of Theridium produces large eggs and small eggs ;
in one cocoon all may be large or all may be small, or in any one
cocoon both kinds may occur ; intermediates in size are relatively
very infrequent. Such a difference of volume might be termed
" dimegaly " for convenience, especially when no structural dif-
ferences are found or known to accompany this difference in
volume. But there is a possibility if not a probability that this
dimegaly may be dimorphism, and that females develop from the
large eggs and males from the small ones. The adult female
spiders are considerably larger than the adult males, notably
with regard to the dimensions of the abdomen. The occurrence
of occasional intermediates between the two kinds of eggs may be
readily explained by the fact that in each kind of eggs there is
always some individual variation in volume, in conjunction with
the assumption that the smallest extremes of the larger kind
may not be larger than the largest extremes of the smaller. In
those cases where some cocoons contain only large eggs, others
only small ones, there would then be instances where some
cocoons produce only females and others only males. Further,
if this dimegaly is really dimorphism, a conclusion, that we are
tentatively maintaining, then in a succession of cocoons made by
the same spider there would be batches of female eggs only

Il8 THOS. MONTGOMERY, JR.

alternating with batches of male eggs only ; I did not have a
sufficient number of cocoons from any one female to determine
what is the regularity of this succession. Or again both kinds
of eggs may occur in the same cocoon, and perhaps future ob-
servations will show that the first cocoons contain only large (or
small) eggs, the next succeeding eggs of both kinds, and the
last cocoons only small (or large) eggs.

Whether this dimegaly is true sexual dimorphism can be
decided only by examining the genital organs of the hatching
spiderlings since there are no external sexual differences apparent
in the young, which would require much labor ; or by raising all
the spiderlings to maturity, a method that would require still
more time and patience. But one of these methods must be
tried in order to finally demonstrate whether this is true sexual
dimorphism of the eggs.

If the small and large eggs of Theridium are really male and
female eggs, and it must be admitted that there is a probability
of this, then here is another instance of two kinds of eggs to be
added to those already known, namely, the cases of the Aphids,
Rotatoria and Dinophihts apatris. Adult sexual differences in size
are very marked in many spiders, the male is probably always
somewhat the smaller, and in many species, particularly among
the Argiopidse, Theridiidas and Thomisidae the disparity in size
of the sexes is most striking. It would be of interest to examine
this point in the case of the common orb-weaver Argiope
cophinaria (Walck.), where the male may be less than one
fiftieth the volume of the grown female, here if any where there
should be marked dimorphism of the eggs ; and in species of the
genus Acrosoma, The common Epcira labyrinthea Hentz would
be especially favorable because it places its cocoons in a string
in the order of their making.

THE "ACCESSORY CHROMOSOME" OF ANASA

TRISTIS.

KATHARINE FOOT AND E. C. STROBELL.

In the Quart. Jonrn. Mic. Set., Vol. 48, 1905, Professor]. E.
S. Moore and L. E. Robinson writing on the spermatogonesis of
Periplanata Americana claim that the nucleolus of the first
spermatocyte is undoubtedly the homologue of the structure de-
scribed by Paulmier ('99), Montgomery ('01), and McClung ('02),
in different forms as one or two of the spermatogonial chromo-
somes. It's morphological resemblance to a chromosome shown
by its frequent elongate form, Moore and Robinson attribute to
mechanical influences and claim that normally it is spherical like
the nucleolus. In three recent papers Professor E. B. Wilson
('05, '06) has given special attention to this structure in a num-
ber of forms and the above interpretation of Moore and Robinson
he ascribes to superficial work.

A study of the spermatogenesis of Anasa tristis 1 has convinced
us that for this form the interpretation of Moore and Robinson
is correct, that the nucleolar-like structure of the rest stage is
the homologue of the nucleolus of the egg, that it is not a chro-
mosome, as claimed by the three cytologists who have investi-
gated this form.

In 1899 Paulmier identified this body of the rest stage with
the two small spermatogonial chromosomes which Wilson has
aptly named the " microchromosomes."

Montgomery in 1901 supported Paulmier in this interpreta-
tion, but in 1906 he changed his position and now supports
Wilson in identifying this structure the so-called "chromo-
some nucleolus " as one of the larger spermatogonial chro-
mosomes an unpaired spermatogonial chromosome, called by
Wilson the odd or " heterotropic " chromosome. Both these in-
vestigators claim that it divides only in the first division, in the
second division, passing undivided to one pole of the spindle.

1 We are indebted to the courtesy of Dr. P. R. Uhler for identifying our material.

I2O KATHARINE FOOT AND E. C. STROBELL.

In Wilson's "Studies on Chromosomes, Nos. II. and III.," he
has published several text figures of Anasa tristis, including the
stages from the "contraction phase" of the first spermatocyte
to the anaphase of the second spindle, and also four spermato-
gonial groups in which he figures an odd number of spermato-
gonial chromosomes, i. e., 10 pairs, and one large unpaired
univalent, the odd or " heterotropic " chromosome.

A study of a large number of smear preparations of the testes of
Anasa tristis has forced us to the conclusion that in the spermato-
genesis of this form there is no so-called "accessory chromosome "
no odd, "heterotropic" chromosome that the so-called
" chromosome nucleolus " of the rest stage is the homologue of the
nucleolus of the egg, that in its form and time of disappearance it
bears a striking resemblence to the plasmosome of the egg of
Allolobophora fcetida. Our observations and interpretations are
so at variance with the conclusions reached by the three cytolo-
gists who have studied this form, that we would hesitate to take
issue with such competent authority were we not able to support
our observations by a very large number of photomicrographs
of the preparations. We have already nearly 200 photographs
which seem to demonstrate beyond question the following points.

That the so-called " chromosome nucleolus " of the resting
spermatocyte is morphologically the equivalent of a nucleolus,
that it is not a chromosome. Wilson has emphasized the evi-
dence of morphological likeness in his Fig. 2, b and c (" Studies
on Chromosomes, II."), in which he shows a structure which he
interprets as a chromosome and which has a marked morpho-
logical resemblance to his sketch of the " chromosome nucleolus "
in his Fig. a.

It is significant that in Wilson's three figures of this early pro-
phase only 6 of the 1 1 bivalents are shown in two of his draw-
ings and only 7 in the third. In fact, not one of the investiga-
tors of this form has given a single figure of this stage in which
all the eleven chromosomes are shown.

In our photographs, on the contrary, all the 1 1 bivalents are
in evidence, and not one of them resembles in the least the
" heterotropic " chromosome figured by Wilson. This holds
true for hundreds of cells in which all the eleven bivalents are
present and clearly defined.

ACCESSORY CHROMOSOME IN ANASA TRISTIS. 121

It is only when the chromosomes or parts of chromosomes are
abnormal that they show a condensed chromatin mass, or masses,
suggesting a resemblance to a nucleolus. In many cases one of
the arms of a cross-shaped chromosome will resemble a round
dense nucleolus and this may appear in from one to five of the
crosses, and again both arms, or the entire cross may have de-
generated into a compact, deeply staining mass of chromatin.
We have a number of photographs of connecting stages between
these extremes, and they leave no doubt that the normal chro-
mosome resembles in no way a nucleolus.

Our preparations also demonstrate that the " chromosome nu-
cleolus " like the plasmosome of the egg of Allolobopliora fcctida
has disappeared, as a rule, when the chromosomes are formed
(early prophase) very rarely persisting until after the chromosomes
have attained their definite shape.

Our smear preparations further demonstrate the absence in the
resting spermatocyte of any other structure which can be inter-
preted as a nucleolus. We approached the study of this form with
the hope of being able to identify a structure in the male cell
which could be interpreted as the homologue of the " accessory
nucleolus " of the egg, 1 but we have found no structure suffi-
ciently pronounced or constant to justify our interpreting it as
an " accessory nucleolus."

Paulmier ('99), Montgomery ('01) and Wilson ('o5-'o6) have
all indicated a second nucleolar-like structure in the resting sper-
matocyte which they interpret as the true plasmosome, but we
have been unable to demonstrate a second nuleolar-like structure
in our smear preparations. However in sections of testes fixed
with Hermann's fluid and stained with iron haematoxylin and with
anilin stains, we often find nuclei of resting spermatocytes in
which a second nucleolar-like structure is differentiated, but the
complete absence of such a feature in our smear preparations,
makes us hesitate to interpret these two structures as the homo-
logues of the plasmosome and accessory nucleolus of the egg.

Again our preparations demonstrate that the so-called uni-
valent " heterotropic " chromosome is distinctly a bivalent. Its
constant bivalent character indicates that it represents in value

'Foot & Strobell, 1905.

122 KATHARINE FOOT AND E. C. STROBELL.

two spermatogonial chromosomes and not one, and when this
chromosome is first formed its bivalent character is much more
pronounced than at the later prophase stages. Our photographs
however support Wilson in his claim that it appears only excep-
tionally as a tetrad - - as a rule this and the micro-chromosomes
appear bivalent, while all the others show a marked tetrad char-
acter. The frequent eccentric position of this bivalent chromo-
some, outside the characteristic ring arrangement of the chro-
mosomes in the late prophase, seems to warrant suggesting
"eccentric" chromosome as a convenient descriptive name for
this special chromosome.

Individuality of the Chromosomes. Our preparations show a
marked individuality ot the chromosomes, and in this support
the observations of Paulmier, Montgomery and Wilson. Sev-
eral of the chromosomes can often be clearly identified during
the prophases, metaphase and anaphase, though a comparison of
a large number of photographs demonstrates that \\\z form is not
constant. For example, at a definite prophase, 7, 8 or 9 of the
1 1 bivalents may be clear and sharply defined crosses, while
again in the same stage we may have all rods or only I, 2 or 3
crosses, this indicating that the cross type is not invariably asso-
ciated with any one chromosome.

During the growth period the chromosomes certainly lose
their individuality as completely as in the case of Allolobophora
fcetida, and we have therefore no positive proof that each biva-
lent of the prophase represents the same chromatin that formed
a pair of the spermatogonial univalents. There is a certain
degree of constancy in the relative sizes of the chromosomes,
although a definite chromosome may differ greatly in size in dif-
ferent cells at the same stage of development. This may be due
in some degree to the technique, but this difference is often so
great that we feel convinced it is probably due at least in
part, to an actual difference in the size of the individual chromo-
somes.

Montgomery ('06) observed an inequality in the size of the
two microchromosomes, but in our preparations we do not find
any support for this observation.

Plane of the First Division. In many cells in which all the

ACCESSORY CHROMOSOME IN ANASA TRISTIS. 123

chromosomes are clearly defined a transverse division for each
chromosome can be plainly demonstrated by one photograph.
In many other cells, however, it can be as clearly determined
that the " eccentric " chromosome divides longitudinally while
all the others divide transversely. It may be stated, as a rule,
that the "eccentric" chromosome divides longitudinally, though
many exceptions can be demonstrated.

Tlic Lagging Chromosome of the First Division. At the late
anaphase or early telophase of many of the first divisions, it can
be demonstrated that one of the chromosomes has divided at a
later period than the others. It may be seen between the two
poles in all stages of separation --sometimes the entire bivalent
will be between the poles, its two univalent halves having just
separated or one univalent may have reached one pole while the
other half still lies midway between the poles. We have several
photographs in which a lagging chromosome is shown while all
the other chromosomes can be counted, thus an error in inter-
pretation is quite impossible. As a rule this lagging chromosome
appears to be the "eccentric" chromosome, though it cannot be
demonstrated that it is invariably the same chromosome which
lags in division. We have not found this phenomenon with
sufficient frequency to justify our interpreting it as a constant
feature of this division. We think the condition exceptional,
though not necessarily pathological.

The Lagging Chromosome of the Second Division. The second
division shows a phenomenon which appears to us to be the
equivalent of the one just described for the first division. It is
more frequently found for the second spindle than for the first
and much more difficult to interpret, as these spindles are so
exceedingly small and the chromosomes so closely crowded
together, that cases are rarely found in which all the chromo-
somes are in evidence and the true value of the lagging chromo-
some can be safely interpreted. We have a few photographs
which we think throw some light upon this point. All the
chromosomes at each pole are demonstrated and the lagging
chromosome lying midway between the poles, in several cases
shows a distinct transverse constriction. In one preparation the
two halves have separated, and in another the two halves have

124 KATHARINE FOOT AND E. C. STROBELL.

reached opposite poles of the spindle, though in these cases the
chromosomes at the poles are too crowded to be counted. We
interpret this lagging chromosome as a univalent, being equal in
value to all the other chromosomes of the second spindle, just as
in the first spindle we interpret the lagging chromosome as a
bivalent both cases indicating simply a retarded division of
one of the chromosomes.

Our preparations do not support Wilson and Montgomery in
their observation that the lagging chromosome goes over undivided
to one pole of the second spindle, and we are therefore unable to
follow them in supporting McClung's theory of the dimorphism
of the spermatozoa.

If these authors are correct in interpreting this lagging chromo-
some as only half of a univalent, what is the significance of the
frequent transverse constriction? What can such a constriction
mean but foreshadowing a division ? and this interpretation is
supported by the cases in which the division of this chromosome
is actually demonstrated. We do not interpret the presence of
a lagging chromosome in the first or second spindle as neces-
sarily an abnormal condition though it may be a step in that
direction, for we have seen unmistakably pathological spindles
where sometimes one and sometimes two chromosomes pass to
one pole undivided. We have photographs of some of these
spindles and their pathological character can be readily recog-
nized. We also have examples of such unequal, abnormal
separation of the chromosomes in the first and second spindles
of Allolobophora fa'tida.

Spennatogonial Chromosomes. Paulmier ('99) who was the
first to study the spermatogenesis of this form interpreted the
number of spermatogonial chromosomes to be 22 and in his
Fig. 9 has reproduced one of his sections in which 22 chromo-
somes are clearly represented.

Montgomery ('01) supports Paulmier in his estimate of the
number of spermatogonial chromosomes and in his Fig. 74 gives
a very clear demonstration of this number.

Wilson in his "Studies on Chromosomes, No. I," corrects
this original count of the spermatogonial chromosomes with such
positive assurance ' that we have great hesitation in questioning

ACCESSORY CHROMOSOME IN ANASA TRISTIS. 125

his results and would not presume to do so were it not possible
to corroborate our observations with photomicrographs in which
22 chromosomes can be counted without any question. We
realize in common with all cytologists the difficulty of getting a
correct count of so large a number of small bodies crowded into
a contracted space. If 2 or more chromosomes are in such close
contact that their line of separation is obscured a correct count
is impossible. It is certainly possible to find cells in which only
2 1 chromosomes can be differentiated and still easier to find cells
in which only 20 or 19 are defined. It is much more difficult to
find each chromosome so distinctly isolated that all can be demon-
strated in one photograph.

Wilson ('c>5-'o6) corroborates his original count of 21 sperma-
togonial chromosomes and illustrates this point in four sketches.
Montgomery ('06) in his last paper withdraws his earlier estimate
of the number and supports Wilson's results, figuring 21 chromo-
somes in his sketch No. 151.

In view of this weight of authority we do not feel inclined to
be in the least dogmatic in our estimate of the count of the
chromosomes, but our preparations certainly justify us in main-
taining that it is possible to demonstrate 22 spermatogonial
chromosomes in Anasa tristis, as we shall show later in photo-
micrographs of the preparations themselves.
December I, 1906.

1 " Since this paper was sent to press I have determined beyond the possibility of
doubt, I think, that the number of spermatogonial chromosomes in Aiiasa tristis is 21
not 22 as given by both Paulmier and Montgomery. This result is based on a study
of a large number of preparations and careful camera drawings of more than 20 per-
fectly clear metaphase figures have been made. All without exception show 2 1 chromo-
somes, and I have sought in vain for even a single cell that shows 22."

126 KATHARINE FOOT AND E. C. STROBELL.

BIBLIOGRAPHY.
Foot & Strobell.

'05 Prophase and Metaphase of the first Maturation Spindle of Allolobophora
fcetida. The Amer. Journ. Anat., Vol. IV., No. 2.

McClung, C. E.

'02 The Accessory Chromosome Sex-determinant? Biol. Bull., Vol. III.,
No. 2.

Montgomery, Thomas H.

'01 A study of the chromosomes of the Germ Cells of Metazoa. Trans. Amer.

Philos. Soc., Vol. XX.
'06 Chromosomes in the spermatogenesis of the Hemiptera heteroptera . Trans.

Amer. Philos. Soc., Vol. XXI.

Paulmier, F. C.

'99 The Spermatogenesis of Anasa tristis. Journ. Morph., Vol. XV.

Wilson, E. B.

'05 Studies on Chromosomes, I. Journ. Exper. Zool., Vol. II., No. 2.

'05 Studies on Chromosomes, II. Journ. Exp. Zool., Vol. II., No. 4.
'06 Studies on Chromosomes, III. Journ. Exp. Zool., Vol. III., No. 4.

Vol. XII. February, 1907. No.

BIOLOGICAL BULLETIN

THE HABITS AND MOVEMENTS OF THE RAZOR-
SHELL CLAM, ENSIS DIRECTUS, CON.

OILMAN A. DREW.

Many of the older naturalists have called attention to the sen-
sitiveness and remarkable activity of this form, and the conse-
quent difficulty that is sometimes experienced in capturing it.
Some of the observations are not strictly accurate but they
have served to call attention to its adaptation for a burrowing
life. Even the uncommon shape of the animal indicates this
adaptation

The species under consideration is to be found more or less
abundantly all along the eastern coast of the United States. It
is best known on sandy flats from which most of the water flows
at low tide, where specimens may be dug with spade or clam-
hoe. In some localities, as in restricted areas around Woods
Holl and North Falmouth, Massachusetts, where most of these
observations were made, the animals are quite abundant, and in
such places one may find the protruding posterior ends of the
shells, or see the siphon openings of the undisturbed individuals.
They readily take alarm and even a slight jarring of the mud of
the bottom in their vicinity serves as a signal for them to in-
stantly disappear. It is this sudden disappearance that has
attracted wide attention, and has given the impression that the
animals are exceptionally hard to capture (5 and 6).

The species is probably not restricted to very shallow water.
Specimens are not often taken in a dredge, but the position that
they occupy buried in the rather hard sand or mud of the bottom,
makes their capture unlikely. Young specimens, from a milli-
meter to a centimeter in length have been taken in large numbers

127

128 OILMAN A. DREW.

near South Hapswell, Maine, in from ten to thitry feet of water,
by means of a fine wire dredge. Supposedly where conditions
are right for individuals to grow to a centimeter in length, larger
ones would thrive also.

The usual position for undisturbed specimens is with the pos-
terior ends of the shells protruding just above the surface of the
mud. Sometimes specimens are found with several centimeters
of their shells protruding, but this is not very common. Again
all that indicates the presence of a specimen, may be the depres-
sion at the spot where the animal has disappeared. The most
conspicuous parts of a specimen in its normal position are the
siphons, the openings of which appear as nearly round apertures
surrounded by tentacles (Fig. 5). Close observation is necessary
to see such specimens, as the color of their siphons is almost
exactly the same as the bottom where they occur.

Specimens that have their posterior ends protruding above the
surface of the mud seem to be more common where the water
has drained entirely away than where the flats are still covered
with water. Shells are not uncommon on the flats and shores,
but they are seldom found embedded in the position that the
animals occupy during life. These observations are of especial
interest in view of the fact that animals in aquaria quite uni-
versally push up out of the mud before they die. The heat of
the sun on the bare mud flats is probably disturbing an'd may
cause them to react in the same way they do before they die.
If animals react in nature as they do in aquaria, it is natural that
the shells of dead animals should be found on the surface.

Specimens are not easily studied in their native places because
of ripples on the water and because the character of the bottom
makes it hard to approach closely without disturbing them.
Walking on the mud near them will cause them to withdraw
their siphons or disappear beneath the surface. It has accord-
ingly been more satisfactory to study specimens in aquaria that
contain several inches of sand or mud from the bottom from
which the specimens were obtained. The animals do not live
well in aquaria, even when supplied with running water, but for
several hours after they are collected, they remain very active
and seem to be quite normal.

If a specimen is touched it will either withdraw its siphons anp

HABITS AND MOVEMENTS OF THE RAZOR-SHELL CLAM. 129

remain quiet until disturbed again, or it will immediately vanish
beneath the surface of the mud. If a specimen is grasped and
pulled upward there is an immediate response that is so power-
ful that the animal frequently escapes and disappears. This fact
Tryon (6) mentions, saying : " It may often be seen at low tide
projecting a little above the level of the sand but, if touched or
disturbed, it descends with astonishing rapidity and force, much
to the amazement of him who may lay hold of it thinking to
make an easy capture." These observations indicate that the
animal probably habitually keeps its foot protruded some distance
out of the shell to be ready for disturbances.

Specimens taken in firm sandy soil, where the depth can be
noted, are frequently found several inches below the surface.
Verrill and Smith (7) report that this species digs somewhat
permanent burrows that extend nearly perpendicularly into
the sand to the depth of three feet, and Woodward (8) states
that the animals never voluntarily leave their burrows. I have
never been able to demonstrate permanent burrows in the locali-
ties where I have worked, but the usual muddy character of the
bottom was not satisfactory for the purpose. I doubt, however,
if such burrows are habitually constructed. The character of the
bottom where they live is frequently not suitable for permanent
burrows unless something like a tough secretion is added to the
mud to keep it in place, and such a secretion does not seem to be
formed. Specimens in aquaria never seem to construct anything
like permanent burrows.

As individuals are known to burrow to some depth it is prob-
able that in digging for them, those in the immediate vicinity are
disturbed and burrow beneath the reach of the shovel. Verrill
and Smith (7) call attention to how easily disturbed they are and
say : " When thus alarmed it is generally useless to try to dig
them out, for they quickly descend beyond the reach of the
spade." Dr. J. Gwynn Jeffreys as quoted by Tryon (6) reports
them as having exceptional powers for detecting disturbances.
He says : " They are evidently sensible to vibratory movements
in the air, as well as on ground, taking alarm at greater or less
distances according to the state of the atmosphere and the direc-
tion of the wind." It is hard to verify these observations and I
am inclined to think they are not accurate, but there can be no

I3O OILMAN A. DREW.

doubt that specimens are easily disturbed by vibrations of the
bottom. That they habitually burrow to some depth is indicated
by the fact that after the first half dozen trials, specimens are not
usually obtained without going to another spot several feet away,
although the first trials may have resulted in one or more speci-
mens each.

When dug from the mud, individuals frequently leap or swim.
In leaping the shell may be thrown several inches by the action
of the foot. In swimming the animal progresses posterior end
first and large specimens may swim several feet before stopping.
The foot is always active while the animal is swimming. These
activities, together with the movements of burrowing, will receive
attention later.

Before describing the movements it is desirable to call atten-
tion to some points of anatomy.

The shell is of nearly even diameter both dorso-ventrally and
laterally (Figs. I and 4), throughout its length, except very near
its ends. Anteriorly it contracts in both directions a little, but
the anterior margins of the shell valves remain wide apart even
when they are in contact along their ventral borders (Fig. 6).
This leaves ample space for the protrusion or withdrawal of the
foot when the shell is closed. The anterior margins of the lobes
of the mantle are thickened and extended, so, when the foot is
withdrawn into the shell, these flaps cover the opening between
the shell valves (Fig. 6). When the foot is extended the flaps
are spread apart and form a collar around the foot, the free
margins of which are in contact with the foot (Figs. I and 4).
The collar is thick and muscular, being well supplied with the
radial pallial muscles, and as it is held tightly against the foot,
forms a very effective scraper, that cleans the foot so mud is not
drawn into the shell with it. The cilia covering the foot no
doubt aid in loosing the mud so it is easily scraped off. The
flexible collar adapts itself to the shape of the foot so its margin
is applied to the surface of the foot until its very extremity is
drawn into the shell (Figs. 6 and 7).

The posterior end of the shell narrows laterally, but here again
the margins of the shell valves are wide apart when the ventral
edges of the valves are in contact (Fig. 5). This makes it pos-
sible for the siphons to be at least partially extended when the

HABITS AND MOVEMENTS OF THE RAZOR-SHELL CLAM. 13!

shell is closed. The reason for this arrangement is found by
studying the movements of burrowing and swimming.

The siphons have nearly circular or slightly elliptical openings,
and are separated near their extremities. The whole posterior
end of the mantle, bearing the siphons, may be protruded some
distance beyond the posterior end of the shell. Sense tentacles
surround the siphons near their bases and occur on the mantle
dorsally and ventrally near the posterior end, where the mantle
is exposed between the shell valves (Figs. 2 and 5). Small
sense tentacles occur on the surfaces of both siphons, and the
branchial siphon bears a number along its extremity that tend to
radiate in over the opening of this siphon. The margin of the
cloacal siphon has no tentacles. Verrill and Smith have (7) de-
scribed a definite arrangement for the tentacles but it is doubtful
if this arrangement always holds.

The ventral margins of the mantle lobes are united throughout
their length except near the middle of the length of the animal,
where a small opening remains that is situated just posterior to
the retracted foot (Fig. 4). This opening is surrounded by a
single row of sense tentacles. Except for this opening, the
opening through which the foot is protruded, and the openings
of the siphons, the mantle forms a closed chamber.

The united mantle margins are very muscular, being provided
with strong circular and radial pallial muscles that are very simi-
lar to the muscles of the mantle margins in Solenoinya (2) where
they serve very much the same purpose, that is, to close the
shell tightly and to obliterate a portion of the mantle chamber.
Like Solenoinya the valves of the shell are covered with a very
heavy, elastic cuticle that is extended beyond the calcareous mar-
gins. Mud does not readily adhere to this cuticle. When the
valves are closed the cuticle is bent in over the hard margins of
the shell (Fig. 8), thus allowing the united margins of the mantle
to be withdrawn. Probably this elastic cuticle aids in opening
the mantle chamber when the muscles relax, as is undoubtedly
the case with Solenoinya (2), but the effect in this form is cer-
tainly much less than in Solenoinya.

Both adductor muscles are present. The anterior adductor is
very large and strong. The posterior adductor is quite small
and does not seem to function actively. The united margins of

132 OILMAN A. DREW.

the mantle posterior to the ventral opening are especially mus-
cular and seem to replace the posterior adductor in function to a
marked extent.

The foot, when retracted into the shell, is nearly cylindrical
and together with other organs completely fills the part anterior
to the ventral opening in the mantle. The dorsal portion of its
extremity is pointed, and a slight ridge marks the boundary of
what may be called the sole (Fig. 2). The foot is very power-
ful and remarkably active, its movements being very unlike the
slow movements of the foot in most lamellibranchs. It may be
thrust from the anterior end of the shell to a distance exceeding
one half the length of the shell, and in this position the end may
be swelled into a knob or bulb that considerably exceeds the
diameter of the shell (Fig. i). The knob is not cylindrical but
is extended dorso-ventrally and laterally and the free extremity
or sole is comparatively flattened. In this swollen condition the
end of the foot forms a very efficient anchor, as will be found by
grasping a shell and trying to withdraw it from the mud. The
resistance of the expanded foot is so great that the foot is fre-
quently torn away from the shell when the shell is jerked
quickly (7). The foot is attached to the shell by two pairs of
foot muscles, both of which are strong and aid in withdrawing the
foot into the shell. With the end of the foot anchored, the ob-
vious result of the contraction of these muscles is to pull the
shell into the mud up to the position of the bulbous portion of
the foot.

For our present purpose it is not necessary to give more atten-
tion to the anatomy of the animal, and we will proceed at once to
the study of the movements.

Burrowing.- -The movements of burrowing may be best
studied either in specimens placed in shallow dishes of sea water,
which are very likely to execute the movements soon after they
are placed in the water, or in specimens held with the anterior
end pointing downward and stimulated to activity by stroking
the sense tentacles around the ventral opening in the mantle and
around the siphons. Specimens in the water are more normal in
their movements than the specimens held and stimulated as de-
scribed. Apparently the action of gravity may cause the held
specimen to protrude its foot and to partially expand the end of

HABITS AND MOVEMENTS OF THE RAZOR-SHELL CLAM. 133

it, in which position the foot may remain quiet for some time.
When the movements are active they are essentially the same in
both cases. The foot is slowly protruded with the pointed tip
working as if trying to bore into the mud, ending each time with
a dorsal thrust. These movements are continued until the foot
is fully extended. During this extension the end of the foot is
kept small, the point is directed well forward, and the general
diameter of the protruded part of the foot is decidedly less than
its normal diameter when at rest. When the foot reaches its
greatest extension, the end is suddenly swelled into a great
bulb, more than twice the diameter of the remainder of the
foot (Figs, i and 4) and the whole foot becomes very rigid.
That this result is attained by injecting blood into the foot may
be readily proved by sticking spring forceps into the end of the
foot so the spring will hold the wound open, and stimulating the
foot to activity by stroking the tentacles as before described.
When the foot starts to become active the wound begins to
bleed rapidly, and when the final effort to swell the end of the foot
is made, the blood rushes out in a great jet, but the swelling is
slight. A simple incision does not answer as well, as the con-
tracting muscles seem to close the wound more or less perfectly.

The instant that the swelling of the end of the foot is com-
plete, a process that takes place so rapidly as to be almost start-
ling, the retractor muscles pull the foot back to the shell with a
jerk, the end remaining swollen until it reaches the shell (Fig. 7).
It is then reduced in size and either withdrawn into the shell or
extended in the beginning of a new burrowing movement.

While the foot is being extended the shell valves are allowed
to gap apart and the siphons and ventral opening in the mantle
are kept more or less widely open. Just before the final sudden
retraction, the siphons and ventral opening are all tightly closed,
and kept so until retraction is complete. The result is that the
water in the mantle chamber is discharged through the opening
through which the. foot is extended, between the collar and the
foot. Whether the water escapes all around the foot or only
ventral to it, where the contact of the collar is poorest, has not
been determined, but the jet of water is quite powerful. When
the shell is embedded in the mud, each retraction of the foot,

134 GILMAN A. DREW.

squirts the water against the mud ahead of the shell, the shell is
decreased in diameter by being closed, and the mud is dislodged
and washed up the sides of the shell where it may be seen raising
after each downward movement of the shell. The action is
similar to the pile driver that opens a way for the pile by a
somewhat similar stream of water.

The burrowing movements may follow each other quite rapidly
but the extension of the foot is never very rapid, as it must be
carefully worked into the mud to keep from forcing the shell
back. The opening of the shell just before the extension of the
foot, tends to embed it more firmly and thus to hold it in position
while the foot is being worked into the mud.

A specimen laid on its side on mud, has no difficulty in gaining a
hold with its foot that enables it to right itself and start the anterior
end into the mud. Burrowing is then normal, and the shell is soon
completely buried. The time necessary for a specimen to com-
pletely bury itself varies with the character of the mud. In soft
mud the thrusts may be rapid and few are needed, in hard sand
the thrusts will necessarily be slower and more movements are
required, but even in such material the animal will disappear very
promptly. When the animal is laid on its side on such sandy
mud as that in which it usually lives, one movement will fre-
quently suffice for it to right itself, and four or five more will
carry it out of sight. The time necessary for this may be less
than half a minute. Embedded as the animal usually lives, a
single retraction takes it out of sight and away from enemies.

Swimming. It is more difficult to study the movements of
swimming, as animals swim only occasionally, and then generally
immediately after being dug, and the movements of parts of the
animal, and the animal as a whole, are so rapid as to make ac-
curate observations difficult. The following points have been
determined however and, from these, conclusions may be drawn :
(i) The animal progresses posterior end foremost; (2) move-
ments are by jerks, each jerk carrying the animal one or more
times its length ; (3) the foot is very active, being thrust out and
withdrawn repeatedly. The outward thrust is comparatively
slow but the withdrawal is extremely rapid ; (4) apparently the
valves of the shell are drawn together every time the foot is re-

HABITS AND MOVEMENTS OF THE RAZOR-SHELL CLAM. 135

tracted ; (5) each movement of the animal as a whole, corre-
sponds to the period of retraction of the foot.

In describing the movements of burrowing it has already been
mentioned that water is thrown from the shell, through the
opening through which the foot is protruded, every time the foot
is retracted into the shell. Each jet is caused by closing the
other openings into the mantle chamber and driving the water
out by pulling the foot in, by closing the shell by the contraction
of the adductor muscles and the united margins of the lobes of
the mantle, and by drawing the mantle margins, with the shell
cuticle to which they are attached, into the mantle chamber
(Fig. 8). The resultant action is to drive out most of the water
that was between the valves of the shell, as nearly all of the
space is now occupied by organs of the body. As all of the
openings except the one around the foot are held closed, a very
strong jet of water must be forced out around the sides of the
foot. This is sufficient to cause the movement of the animal in
the opposite direction. Many muscles, all of which are power-
ful, are used in this action, and as the water is thrown through
a small opening between the muscular collar and the foot, the
resulting force is considerable. The action, so far as movement
is concerned, is similiar to what is so well known in the squid,
and differs from the movement of Solenomya only in direction (2).
Here the movement is posterior, in Solenomya the movement is
anterior. Here the water is admitted through the siphons, and
possibly also around the foot, and then, with the siphons closed,
the water is thrown from the anterior end of the animal. In
Solenomya the water is admitted around the foot, which opening is
then closed and the water is thrown through the posterior
opening. The method of forming the jet is quite the same
in both animals. In both forms the same organs are used, but
in Solenomya more use is made of the mantle margins and less of
the retraction of the foot than is the case in this form.

Throwing strong jets of water from the siphons must aid
lamellibranchs in keeping their mantle chambers clean. Some
forms need to throw more powerful jets than others because the
conditions under which they live demand it. A diversion of this
use is apparently to be seen in the forms that swim either by

136 OILMAN A. DREW.

clapping the shell valves together, as in the case of Pccten (3), or
by the more complicated method used by Solenomya and the
form under consideration. It is likely that the jets thrown by
this form are of very secondary importance, so far as swimming
is concerned, and that their chief function is to aid the animal in
burrowing. This is indicated by the fact that the jets are thrown
from the anterior end of the shell, while in forms that use them
for cleaning the mantle chamber only, they are thrown from the
siphons.

Leaping. Leaping may consist simply of a sudden, powerful
protrusion of the foot, in which case the animal generally turns
so as to lie somewhat nearly on its dorsal margin and catches the
tip of its foot in the mud as it is protruded. The shell is thus
thrown posteriorly. Generally, however, the foot is bent back
under the shell, which is turned partly over towards its dorsal
margin (Fig. 3) and is then suddenly made rigid with the result
that it straightens out with great rapidity. This may result in
projecting the animal backward, or in certain cases the foot may
catch so as to turn the shell more or less completely end for end.
Leaping movements are usually rapidly repeated several times
when they are once begun. In many ways they resemble similar
movements in Yoldia and Solenomya (2), but the foot of this form
is so much longer that the impression of much greater activity is
left with the observer.

The perfection of the movements of burrowing by a form that
lives in the mud, so it may be able to escape its enemies, is of
so much importance as to need no comment. When combined
with sense organs that give immediate information of the presence
of enemies, and with protective coloration that hides it from its
enemies until they shall have given it warning, the rapid burrow-
ing movements form a striking adaptation.

The uses of the swimming and leaping movements are not quite
so evident. Small razor-shell clams have been taken in tow nets
at the surface of the sea. The ability to swim is, then, sufficient
to make- it possible for the young specimens, at least, to change
their positions after settling to the bottom, and after the larval
locomotor organ, the velum, has been lost. If the first location
does not offer the necessary food or bottom conditions, it is pos-

HABITS AND MOVEMENTS OF THE RAZOR-SHELL CLAM. 137

sible to move. Although the machinery for this is rather clumsy,
and was primarily designed for another purpose, that of burrow-
ing, it might be the deciding factor in the struggle of life. The
fact that young specimens may be taken at the surface of the
water, even the fact that the animals are able to swim at all, indi-
cates that they probably occasionally change their positions. The
statement made by Woodward (8) that the animals never volun-
tarily leave their burrows seems doubtful. They certainly do
leave the mud when they are about to die, and there is no reason
to believe that they might not voluntarily move from one place
to another should occasion require.

Leaping may aid the animals in getting free from certain kinds
of bottom or even occasionally in escaping enemies, should they
be removed from the mud in any manner. Not infrequently
specimens that have been swimming might become lodged so
they could not burrow without changing their positions, and
then the leaping movements would be of advantage.

SUMMARY.

The animal is very active, burrows with great rapidity, and
may also swim and leap.

In burrowing, the foot is worked into the mud as it is protruded,
the end is then swelled into a knob, and by its sudden withdrawal
the shell is drawn to the position of the anchored end of the foot.
Simultaneous with the retraction of the foot, a strong jet of water
is thrown from the anterior end of the shell, so the mud is soft-
ened or even washed away as the shell descends, an action
similar to that of some of the modern pile drivers. A collar,
formed by the anterior end of the mantle, surrounds the foot
and acts as a scraper that prevents mud from being drawn into
the shell with the foot.

The animal is able to swim by throwing jets of water from the
anterior end of the shell, thus progressing backward by a series
of jerks.

By the uncommon activity of the foot the animal is able to
throw itself about on the bottom.
UNIVERSITY OF MAINE, ORONO, MAINE.

138 OILMAN A. DREW.

LITERATURE.

1. Cambridge Natural History. Mollusca.

2. Drew.

'oo Locomotion in Solenomya and its Relatives. Anat. Anz., Bd. XVII., 1900.

3. Drew.

'06 The Habits, Anatomy and Embryology of the Giant Scallop (Pecten tenui-
costatus, Mighels). Univ. of Maine, Studies, No. 6, 1906.

4. Gould and Binney.

'70 Invertebrata of Massachusetts. 1870.

5. Ingersoll.

'87 The Oyster, Scallop, Mussel, and Abelonie Industries. Fisheries and Fish-
ery Industries of the U. S., Sec. II., Vol. V., Pt. XX., 1887.

6. Tyron.

'82 Structural and Systematic Conchology. 1882.

7. Verrill and Smith.

'74 Report upon the Invertebrate Animals of Vineyard Sound and Adjacent
Waters. Rept. U. S. Com. Fish and Fisheries on the Conditions of the Sea
Fisheries of the South Coast of New England in 1871 and 1872. 1874.

8. Woodward.

'71 Manual of the Mollusca. 1871.

BIOLOGICAL BULIE1IN VOl . XII

8

I4O OILMAN A. DREW.

EXPLANATION OF PLATE II.

FIG. i. A specimen with the foot extended and the end of the foot swelled, the
instant previous to withdrawal. Drawn with the aid of an instantaneous photograph
made by Mr. J. G. Hubbard of a specimen held with the anterior end down, and
stimulated into activity. Natural size.

FIG. 2. A specimen showing the usual position assumed in a dish of sea-water.
The siphons may be extended more than is shown in this figure. Drawn from ob-
servations. Natural size.

FIG. 3. A specimen showing the position assumed just before leaping. Drawn
from observations. Natural size.

FIG. 4. Ventral view of a specimen that has the foot extended to the position
shown in Fig. I. Drawn from observations. Natural size.

FIG. 5. Direct view of the posterior end of an animal that has the siphons
extended. Drawn from observations. Natural size.

FIG. 6. Direct view of the anterior end of a specimen that had withdrawn all but
the tip of the foot, showing the adjustment of the collar. Drawn from observations.
Natural size.

FIG. 7. Anterior end of a specimen showing the position and shape of the foot at
the end of a burrowing movement. Drawn from observations. Natural size.

FIG. 8. Diagrammatic cross section of an animal, taken just posterior to the ventral
opening in the mantle, to show how the mantle chamber is diminished by the contrac-
tion of the united mantle margins. Twice natural size.

-

EXPERIMENTS ON THE EGGS OF CHyETOPTERUS
AND ASTERIAS IN WHICH THE CHROMATIN

WAS REMOVED.

J. F. McCLENDON.

Last July and August, while in the U. S. Bureau of Fisheries
Laboratory at Wood's Hole, Mass., I devised an apparatus for
removing blastomeres from ascidian eggs, but as the chorion of
these eggs was too tough for a successful operation, I tried to
use the apparatus for other purposes. I found that with it I
could remove parts of the unsegmented eggs of echinoderms
and annelids, and, in the few
weeks at my disposal, performed
the experiments mentioned be-
low.

THE APPARATUS.

To a Greenough binocular
stand (see accompanying figure)
I attached a Spencer mechanical
stage supplied with a fine adjust-
ment screw (<?) made by Wie-
back and Pietzsch, Philadelphia.
The addition of this screw al-
lowed movement in three dimen-
sions, and the fine adjustment (a)
carried a tube (^) drawn to a cap-
illary opening at the lower end,
and with the upper end attached
by a rubber tube to the glass tube
(r), which I held in my mouth
during the operation. The lenses used were Zeiss binocular ob-
jective <? 3 and a pair of Zeiss orthomorphic oculars, 4, giving a
magnification of 65 diameters. Eggs were placed in 23 mm.
depth of sea water in a glass dish on the stage of the microscope
and the capillary end of the tube ($) inserted into the water above

141

142 J. F. MCCLENDON.

them and allowed to remain until the water rose in it as high as
it would by capillarity. The dish was moved until an egg of the
most favorable orientation was brought in the center of the field.
The capillary tube was then brought against the proper spot on
the egg and a portion of the egg sucked away through the tube
(<:) which was held in the mouth. Immediately the operated egg
was lifted from the dish by means of a capillary pipette just large
enough to admit it easily, and placed in a watch glass half full
of sea water. This capillary pipette does not need a bulb, as
water enters it by capillarity. The size of the capillary end ot
the tube (&) was determined for the egg of each species by trial
Some parts of the mechanical stage were an unnecessary hind-
rance and were removed.

EXPERIMENTS ON THE EGGS OF CH^TOPTERUS PERGAMENTACEUS

CUVIER.

Mead ('gS 2 ) and J. Loeb ('01) found that differentiation went
on in the unfertilized egg of Chcstoptervs in solutions of KC1, and
F. R. Lillie showed that this differentiation bears a semblance
to the normal development save that cell division is usually
suppressed.

I removed the chromosomes during the maturation divisions
of the unfertilized egg in KC1 solutions by sucking away the part
of the egg immediately under the first polar body, in order to
determine the effect of the chromatin on differentiation. A small
percentage of eggs were so injured by the operation that they
remained inactive until disintegration. Others, however, per-
formed those " amceboid " movements which Loeb mistook for
cleavage, just as in operated eggs. Further than this no changes
were visible externally, but sections showed that irregular flow-
ing of substances had taken place within the egg. None of the
operated eggs developed cilia, and we might conclude that the
chromatin is necessary for the formation of cilia, a function that
has been attributed to the archoplasm. But as the centrosomes
were probably removed with the chromosomes, and as the ele-
ment of injury could not be excluded, this negative evidence
should not be considered conclusive. Nearly one hundred eggs
were operated on, and, although I do not think it advisable to

EXPERIMENTS ON EGGS OF CH^TOPTERUS. 143

record the individual experiments, I will state the precautions
against error that were used :

The majority of the Cliatoptents females were gathered by me,
or under my direct supervision, and were never placed in the
same vessel with males. Each female was washed in a stream
of fresh water before eggs were removed from it. Except dur-
ing two days no males were allowed in the same laboratory with
the females. One check of eggs in sea-water and another in sea-
water to which KC1 had been added, were kept to each experi-
ment. None in the check in plain sea-water developed, while
many in the KC1 sea-water developed cilia. In each lot of eggs
the operations were continued from the first appearance of the
first polar body to the formation of the second polar body, and
sometimes a few minutes later. Some eggs were fixed and sec-
tioned immediately after the operation, others at later intervals up
to twenty-four hours, and in this way was determined whether
any chromatin had been left in the egg.

EXPERIMENTS ON THE EGGS OF ASTERIAS FORBESII.

Eggs from which the chromosomes were removed and eggs
from which the whole nucleus was removed were fertilized with
sperm of the same species. These operated eggs did not show
any differences from the normal save in number of chromosomes,
in increased tendency toward polyspermy and in increased mor-
tality. This last prevented my ascertaining whether any differ-
ences would appear in the later development.

An attempt was made to remove some of the chromosomes
during the first cleavage, but the spindle was so much more viscid
than the surrounding yolk that no part of it could be sucked out
without removing all of it. In some sections the spindle was
shown pulled to the surface of the egg but not very much
distorted.

Eggs from which chromosomes were removed were mixed with
sperm of a species of Synapta, but did not show signs of devel-
opment save for the separation of the " fertilization membrane "
from the egg.

In the experiments on the Asterias egg the following precau-
tions were used : The starfish were washed in a strong stream of

144 J- F - MCCLENDON.

fresh water before removal of the sexual organs, and after a male
had been opened my instruments and hands were washed in like
manner, to remove all spermatozoa. A check of unfertilized
eggs was always kept, and in case some eggs developed in the
check, the material was thrown out. In a number of experi-
ments the eggs were fixed and sectioned as a further check.

EXPERIMENTS ON EGGS OF CH/ETOPTERUS. 145

LITERATURE.

Garbowski, Th.

'04 Ueber Blastomerentransplantation bei Seeigeln. Bull, de 1'Acad. des Sc.

de Cracovie.
Hirase, S.

'97 Untersuchungen ueber das Erhalten des pollens von Ginko biloba. Bol.
Centrb., XIV., 2, 3.

Lillie, F. R.

.'02 Differentiation without Cleavage in the Egg of Chcetopterus pergamenta-
ceus. Arch. Entwickmech., XIV., 3, 4.

Lillie, F. R.

'06 Observations and Experiments Concerning the Elementary Phenomena of

Embryonic Development in Chsetopterus. Jour. Exp. Zoo., III., No. 2.
Loeb, J.

'01 Experiments on Artificial Parthenogenesis in Annelids (Chsetopterus) and the
Nature of the Process of Fertilization. Am. Jour. Physiol., IV., No. 9,
January, 'oi.

Mead, A. D.

'98' The Origin and Behavior of the Centrosomes in the Annelid Egg. Jour.
Morph., 14, No. 2.

Mead, A. D.

'g8 2 The Rate of Cell Division and Function of the Centrosome. Bio. Lect.

Wood's Hole ('96-7), published 1898.
Morgan, T. H.

'95 Studies of the "Partial" Larvaj of Sphaerechinus. Arch. Entwickmech.,
II., p. 82.

BIOLOGICAL LABORATORY,

RANDOLPH MACON COLLEGE, ASHLAND, VA.,
October, 1906.

THE CONUS ARTERIOSUS IN TARPON ATLANTICUS
(CUVIER & VALENCIENNES).

H. D. SENIOR, M.B., F.R.C.S. -
ASSOCIATE IN ANATOMY, WISTAR INSTITUTE OF ANATOMY, PHILADELPHIA.

The tubular prolongation of the arterial end of the heart
furnished with numerous valves, and known as the conus arteri-
osus, is one of the characteristic features of elasmobranchs and
ganoids. In Aniia calva, the conus arteriosus is relatively shorter
than in other ganoids, and its valves are reduced to three trans-
versely arranged tiers. It may be said that the absence of the
conus arteriosus as a separate structure is a characteristic of the
teleostean heart, and that this fact is emphasized by a few recorded
exceptions, all of which occur in teleostean families more or less
closely related to Anna.

Among these exceptional teleosts only one has been hitherto
described, which possesses more than one tier of conus valves,
this is Butirimis (Albula) which has two, and to it may now be
added Tarpon atlanticns.

The heart from which the following description is taken, was
sent to me by Mr. Charles H. Townsend, director of the New
York Aquarium. It comes from a specimen 5 feet 4 inches in
length. 1 I take this opportunity of thanking Mr. Townsend for
his courtesy in sending this heart, also an entire Tarpon, 4 feet
4 inches long, the heart of which is shown in Fig. 3.

The conus of Tarpon atlanticus resembles that of Aniia calva
in form, but differs from it in being proportionately smaller, in
having two tiers of valves instead of three, and in appearing to
have been driven into the heart towards the apex, so that,
instead of projecting freely from the ventricle, as in Aniia, it is
more or less buried in the latter.

In the natural position, the conus of Tarpon is a horizontally
placed longitudinal tube, elliptical in transverse section. The
longest diameter of the ellipse is dorso- ventral, and measures 16

1 Measurements include caudal fin.

146

CONUS ARTERIOSUS IN TARPON ATLANTICUS.

mm., the shortest measures 11 mm., and is transverse. The
conus wall is slightly under 2 mm. in thickness. The total
length of the conus varies according to the site of measurement,
at the mid-dorsal and mid-ventral lines it measures 8 mm.,
laterally its measurement increases, until at the mid-lateral line
on either side it becomes 10 mm.

FIG. I. The heart has been opened by a mid-ventral sagittal incision, and the
parts widely separated (natural size). A, aorta; A V, atrio- ventricular valve; B,
bulbus arteriosus ; C, conus arteriosus ; D, left distal conus valve ; P, r 'ght proximal
conus valve ; V, wall of ventricle.

In order to compare the relative lengths of conus and ventricle
in Tarpon and Amia, the length of the ventricle has been meas-
ured by plunging a needle into the apex of the ventricle, in such
a direction that its point emerges where the ventricle and conus

148

H. D. SENIOR.

blend. In this particular Tarpon, the ventricle thus measured
is 41 mm. long, and taking 9 mm. as the average length of the
conus, the proportion of the conus length to ventricle length,
becomes I to 4. 5. Six Amia hearts measured in the same way
yield an average proportion of conus to ventricle of I to 1.76.

The exterior of the conus presents
relations which differ in different re-
gions. At the mid-lateral line, and
ventral to this, the ventricle covers
the conus completely. Dorsal to the
mid-lateral line, the ventricle recedes
rapidly, so as only to overlap the conus
for a short distance on either side ; in
the interval, the conus is incom-
pletely covered by the atrium. The
area uncovered by ventricle and at-
rium measured back from the bulbus,
is about 3 mm. in the midline, and
lateral to this about 4 mm., it is cov-
ered by visceral pericardium. (These
relations are indicated in Fig. 2.)

The conus is everywhere overlaid
by a distinct layer of loose connec-
tive tissue, which separates it from
the structures which cover it, and
renders its outline very distinct in
sections of the heart. Owing to the
looseness of its connection with
neighboring structures, the entire
conus is easily exposed from the
outside by incising the pericardium
at the base of the bulbus, and strip-
ping it away from the adjacent parts
of the ventricle and atrium.

The conus valves are disposed in two transverse rows. Each
row consists of a right and left cusp symmetrically placed with
regard to the median dorso-ventral plane of the conus. Seen
rom the lumen of the heart (as in Fig. i) the valve cusps of

FIG. 2. Diagrammatic right
lateral surface-view of bulbus,
conus and ventricle. The atrium
is represented as incised mesi-
ally, and the right half removed
(natural size). The line across
the conus indicates the site of re-
flection of visceral pericardium
on to the atrium. The broken
line indicates the site section in

Fig- 3-

CONUS ARTERIOSUS IN TARPON ATLANTICUS.

149

the proximal l and distal l rows appear to be approximately equal
in size, this however is far from being the case, those of the
distal row having a capacity far exceeding that of the proximal.
The proximal valves are extremely fleshy at their attached
margins, and shade rapidly into a thin semilunar area near 'the
free edge ; the edge itself is marked by a cord-like thickening,
and is quite unattached, except at either end, where, having
blended with the corresponding extremity of the other cusp, it is
attached dorsally and ventrally to the mid-line of the bulbus a
short distance beyond the conus.

The distal valves are not so fleshy as the proximal, and the
marginal semilunar area is very thin and profusely perforated.

The margins are free except
at their extremities, the dorsal
ends of the right and left valves
blend at the mid-longitudinal
region of the bulbus, and be-
come continuous with an elas-
tic cord, the other end of
which is attached to the dor-
sal bulbus wall at its distal
extremity. The ventral ex-
tremities of the distal valves
blend at their point of attach-
ment in the mid-line at the
junction of the proximal and
middle thirds of the ventral
bulbus wall.

That the capacity of the dis-
tal valves is much greater than
that of the proximal, is shown
by the fact that a probe can
be passed 1 1 mm. into the
former, and only 3 mm. into

the latter. To illustrate this point, a frontal section passing ap-
proximately through the mid -lateral line of another heart is
shown in Fig. 3.

1 The terms proximal and distal are used with regard to the ventricle.

FIG. 3. Ventral face of a frontal sec-
tion of the heart from the smaller, Tarpon
its approximate source is indicated by the
broken line in Fig. 3 ( X | ) &> lumen
of bulbus arteriosus ; C, conus arteriosus ;
D, cavity of right distal conus valve ;
P, cavity of left proximal conus valve ; V t
wall of ventricle.

I5O H. D. SENIOR.

The conus in Tarpon appears to differ from that of Bntirinus
(Albuld) described by Boas ('80) in that it is less overlapped by
the bulbus arteriosus, and more deeply buried in the ventricle,
also in that it shows no diminution in length dorsally, as com-
pared to the ventral measurement. The two subsidiary valves
between the larger ones of the proximal row in Albitla do not
occur in Tarpon.

The sinu-atrial valves are two, with strong tensor muscles.
There are four atrio-ventricular valves of which two are of large
size, and two somewhat smaller. The hepatic vein, at its junc-
tion with the sinus venosus, is of almost cartilaginous rigidity, the
size of the orifice is reduced by a thin fold of intima on either
side, these almost meet mesially to convert the circular orifice into
a vertical slit. The folds of intima appear to have no valvular
action.

It is singular that since the appearance of Stannius's paper
('46) Albula should have enjoyed the reputation of being the
only teleost provided with a conus having two rows of valves ;
whether the heart of Megalops cyprinoides will also prove to have
more than one row of valves is an open question. So far as I
am aware a description has not been recorded.

Of the other fishes showing evidence of near relationship to
Aniia the following have been examined with a negative result :
Elops sannts by J. Mueller ('46), Hyodon by Mueller ('46)
and Boas ('80), Osteoglossum by Mueller ('46) and Boas ('80),
Notopterns by Boas ('80), Mormyrops by Mueller ('46). I have
also examined Elops saurns (for a specimen of which I hereby
beg to thank the authorities of the U. S. National Museum)
Hyodon tergisus and Notopterus borniensis.

The original opinion of Gegenbaur ('66) which has been restated
and amplified by Hoyer ('oo) that the conus, although it has ceased
to exist as a separate structure in the ordinary teleost heart, is
represented by the portion of the myocardium adjacent to the
aortic valves, is well illustrated by the conus relations in Tarpon.
One has only to imagine the connective tissue layer between
the exterior of the conus and the ventricle to have disappeared,
allowing the conus muscle to be merged into the general myo-
cardium, and the transition is complete ; the relation of the myo-

CONUS ARTERIOSUS IN TARPON ATLANTICUS. 151

cardium to the distal valve will be similar to that generally found
in teleosts. An interesting transitional stage can be seen in the
heart of Dorosoma cepediamtm, where there is an extremely thin
but distinct streak of connective tissue projecting into the myo-
cardium for sufficient distance .to clearly separate the areas of
original conus and original ventricle.

PHILADELPHIA,
October, 1906.

LITERATURE CITED.
Boas, J. E. V.

'80 Ueber den Connus Arteripsus bei Butirinus und bei anderen Knockenfischen.

Morph. Jahrb., Bd. 6, p. 527.
Gegenbaur, C.

'66 Zur vergl. Anatomic des Herzens. Jenaische Zeitschrift, Bd. 2, p. 365.
Hoyer, H.

'oo Bulletin international de 1'academie des Sciences de Cracovie, No. 7, 1900,

p. 263.
Mueller, J.

'46 Ueber den Bau und die Grenzen der Ganoiden. Berlin, 1846.
Stannius.

'46 Berraerkungen tiber das Verhaltniss der Ganoiden zu den Clupeiden, insbe-
sordere zu Butirinus. Rostock, 1846.

SOME SILKWORM MOTH REFLEXES.

VERNON L. KELLOGG.

Silkworm moths, Bombyx mori, are sexually mature and eager
to mate immediately on issuing from the pupal cocoon. They
take no food (their mouth parts are atrophied), they do not fly,
they are unresponsive to light ; their whole behavior, in fact, is de-
termined by their response to the mating and egg-laying instincts.
We have thus an animal of considerable complexity of organiza-
tion, belonging to a group of organisms well advanced in the
animal scale, in a most simple state for experimentation.

The female moth, nearly immobile, protrudes a paired scent-
organ from the hindmost abdominal segment, and the male, walk-
ing nervously about and fluttering its useless wings, soon finds
the female by virtue of its chemotactic response to the emanating
odor. Males find the females exclusively by this response, but
orient themselves for copulation (after reaching the female) by
contact. When two males accidentally come into contact in their

moving about they try persistently to copulate.

A male with antennae intact, but with eyes blackened, finds
females immediately and with just as much precision as those
with eyes unblackened. A male with antennae off and eyes un-
blackened does not find females unless by accident in its aimless
moving about. But if a male with antennae off does come into
contact, by chance, with a female it always (or nearly so) readily
and immediately mates. The male is not excited before touching
the female, but is immediately and strongly so after coming in
contact with her. Males with antennae on become strongly ex-
cited when a female is brought within several inches of them.

o

The protruded scent-glands of the female are withdrawn into
the body immediately on her being touched by a male. If the
scent-glands are cut off and put wholly apart from the female,
males are as strongly attracted to these isolated scent- glands as
they are to unmutilated females ; on the contrary they are not
at all attracted to the mutilated females. If the cut-out scent-
glands are put by the side of and but a little apart from the

152

SOME SILKWORM MOTH REFLEXES. 153

female from which they are taken, the males always neglect the
near-by live female and go directly to the scent-glands. Males
attracted to the isolated scent-glands remain by them persistently
trying to copulate with them, moving excitedly around and
around them and over and over them with the external genitalia
vainly trying to seize them.

The behavior of males with the antenna of only one side re-
moved is striking. A male with left antenna off when within
three or four inches of a female (with protruded scent- glands)
becomes strongly excited and moves energetically around in re-
peated circles to the right, or rather in a flat spiral thus getting
(usually) gradually nearer and nearer the female and finally com-
ing into contact with her, when he is immediately controlled by
the contact stimulus. A male with right antenna off circles or
spirals to the left. It is a curious sight to see two males with
right and left antenna off, respectively, circling violently about
in opposite directions when the immobile female a few inches
removed protrudes her scent-glands. This behavior is quite in
accordance with Loeb's explanation of the forward movement of
bilaterally symmetrical animals.

The results of all the experiments tried show how rigorously
the male moths are controlled by the scent attraction (chemo-
tropism) and how absolutely dependent mating (the one adult
performance of the males) is on this reaction. If we can find
specialized animals in a condition where all attractions and re-
pulsions (stimuli) but one are eliminated we may readily perceive
the rigorous control exercised by this remaining one. We are,
unfortunately, in the general circumstances of animal life too
much limited to the use of very simply organized animals for
reaction and reflex experimentation. This tends to make it diffi-
cult to carry over to the behavior of complexly organized ani-
mals the plrysico-chemical interpretation which is steadily gaining
ground as the key to the understanding of the springs and char-
acter of the behavior of the simplest organisms. But where the
complex stimuli and reactions that determine the behavior of
complexly organized forms can be isolated and studied the inevi-
tableness of much of this behavior can be recognized.

Reflexes of Moths Without Ccpha/ic and Thoracic Ganglia.

154 VERNON L. KELLOGG.

A number of experiments was made to determine the need, or
absence of need, of the principal ganglia of the central nervous
system in the performance of the two chief reflexes in the silk-
worm moth's life, viz., mating and egg-laying.

Males mate with headless females, and the headless females,
after mating, lay a few eggs which develop normally, that is be-
come fertilized by the release of spermatozoa from the sperma-
theca in the female's body, are oviposited by the repeated extru-
sion and retraction of the ovipositor, and make the usual color
changes (from yellow to cherry-red and then to lead-gray) inci-
dental to normal development. But in no case did a headless fe-
male lay her full complement of eggs, in fact in no case were more
than a score of eggs laid (the normal number is from 200 to 350).
Headless females (and headless males) usually live as long as
unmutilated individuals, i. e., from a week to two weeks.

Females with head and thorax cut off (and even part of the
abdomen) can be mated with by males, and this fractional part
of the female can fertilize and oviposit a few eggs which begin
normal development. In one case 10 eggs, of which 8 are now
normally developing were oviposited by such an impregnated
part of female abdomen, this abdominal relict remaining alive
(!), /. e., flexible and responsive to stimulus and capable of
extruding the ovipositor and laying eggs, for forty hours.

Males with head removed cannot find females, nor can they
mate if placed in contact with them. When the head or head
and prothorax of a male is cut off immediately after the male and
female are in copula the female, although uninjured, lays no eggs.
If heads of both males and females in copula are removed no eggs
are laid although both moths remain alive usually as long as do
unmutilated individuals.

A silkworm moth can maintain- itself right side up with anten-
nae off or with antennae off and eyes blackened, but with head
off one position seems indistinguishable from another to it, i. e., it
lies on one side or the other, on the venter or dorsum equally
willingly. The organs of equilibrium are not on the antennas,
then, but are lost when the rest of the head is removed.

STANFORD UNIVERSITY, CALIF.,
October 15, 1906.

AN ABNORMAL CESTODE PROGLOTTID. 1

EDWIN LINTON.

In the summer of 1905, while engaged in work for the Bureau
of Fisheries at the Woods Hole Laboratory I found an interest-
ing cestode abnormality, a description of which is here given.

Among a lot of free segments of the cestode Calyptrobothrium
occidental? Linton from the torpedo, I noticed one which had
two reproductive apertures upon one of the lateral margins.
Upon flattening this segment under a cover-glass it was found to
be double. The reduplication can be best explained in general
by saying that it is such as would be formed if two normal pro-
glottids were to grow together by their anterior ends.

The specimen was fixed over a flame, and after the customary
hardening was stained with borax carmine and mounted in bal-
sam. Fig. i was sketched from the mounted specimen. The
specimen measured 4 millimeters in length and 2 millimeters in
breadth.

As may be seen by an examination of the figure the specimen
would give rise to two complete proglottids if it were transversely
bisected midway between the two reproductive apertures. At
the opposite ends will be seen the lobed ovaries (o, o) arranged
symmetrically on either side of the median line, with the shell
gland (sg, sg) between the two lateral masses of the ovaries and
toward the extremity of the proglottid.

The reproductive apertures are on the same lateral margin,
each being approximately not far from one third the total length
of the double segment from one end.

Each of the two parts, considered by itself is a normal segment,
so far as the general arrangement of the various organs is con-
cerned, except that there is relatively less space taken up by the
testes (f) which are massed together in the median region and
possessed in common by the two component parts of the double
segment.

The vitelline glands (vg, vg) are continuous along the margin

1 Published by permission of the Bureau of Fisheries.

i- 5 6

EDWIN LINTON.

opposite the reproductive apertures. On the opposite margin a
vitelline mass separates the two vaginse and, like the testes, belongs
to both components of the double segment.

There is no line of demarcation between the two components.
The margins are continuous and smooth everywhere except
between the reproductive apertures where the minute serrations,
which are characteristic features of younger portions of the

strobile, are seen. These ser-
rations are inclined towards the
extremity of the smaller part,
and may be taken as belonging
to that part.

There is no indication as to
which is the older of the two
parts, except a slight difference
in size. Evidently the reversal
took place very soon after the
primary segment was formed.
Although the segment was not
seen attached to its strobile it
is not conceivable that the ab-

normal condition was assumed
after separation from the stro-
bile.

It is to be noted that there
is a reversal in a dorso-ventral
direction also, the vagina and
oviduct lying above the uterus
in one part and below the ute-
rus in the other. Or, to state

FIG. 2. After Blanchard. Abnorm- thig compar j son in another Way,
ality of genital organs of Tcenia saginata.
A, C, normal segment ; B, abnormal
segment.

if the specimen were folded to-
gether on a hinge-line crossing
transversely between the vaginse

the two components would not be symmetrical to the plane of
their apposed faces, but one would correspond in the position
and arrangement of the parts to the other.

Abnormalities in cestodes are of common occurrence and the

{T^

.

o

-od

''<&<
iSi

.A- v ifL>

wf^st^

"^^^1
>F>vr, ^ f

FIG. I Abnormal segment of Calyptrobothriutii occidental \. inton from the torpedo. Ac-
tual length 4 millimeters, <c, reproductive aperture ; sij, shell gland ; c, cirrus ; /, testes ; o,
ovary ; it, uterus ; o<f, oviduct ; v, vagina; 7^. vitelline gland.

AN ABNORMAL CESTODE PROGLOTTID. 157

literature of the subject is considerable. The only case, how-
ever, which at all resembles the subject of this sketch, that I
find in the literature of the subject, to which I have access, is the
one recorded by Blanchard for Tcenia saginata {Bulletin de la
Societe Zoologique de France, 1890, XV., 166-168, and reprinted
in Pr ogres Medical, July, 1 894).

My friend Dr. Stiles informs me that, while he has repeatedly
found abnormalities in cestode segments, he has not made any
record of any cases such as is described in this paper.

On account of the unique character of this abnormality, there-
fore, and for purposes of comparison I reproduce Blanchard's
figure (Fig. 2).

In Blanchard's specimen the ovaries and vitelline glands are
at opposite ends of the abnormal segment ; the testes are con-
tinuous along the lateral margins ; the reproductive apertures
are on opposite lateral margins ; the uterus is common to both
parts. At about one third of the length from the anterior end
there is a transverse incision which reaches to the median line,
and is followed on its side by a perfectly normal segment. If
this incision were to extend somewhat diagonally across, so as
to reach the other margin behind the reproductive aperture,
there would then be a single inverted segment intercalated
between two normal segments.

It should be stated here that the abundance of material found
in the summer of 1905 furnishes additional data, which will make
it necessary to separate the two varieties noted in the original
description (see Bulletin of the United States Fish Commission
for 1899, p. 298) into two distinct species.

The abnormal segment belongs to the larger variety which
will retain the name C. occidental.

WASHINGTON AND JEFFERSON COLLEGE,
December I, 1906.

OBSERVATIONS ON THE YOUNG OF RANATRA
QUADRIDENTATA STAL.

S. J. HOLMES.

In two previously published papers l I have described certain
features of the behavior of the water scorpion, Ranatra, and as
opportunity presented itself this last spring of obtaining young
Ranatras in abundance attention was devoted to a comparison of
the behavior of the young and the mature forms. The various
stages in the metamorphosis of the nymph were followed, but as
these have recently been described by Torre-Bueno 2 the obser-
vations made on this subject are omitted.

Ranatra lays its eggs in the spring commonly in the stems of
aquatic plants or floating pieces of wood. The eggs are inserted
in the material containing them so that they are nearly buried.
They are of cylindrical form, rounded at either end, and provided
at the outer end with a pair of slender filaments of uncertain
function. When one sees a Ranatra which has recently emerged
from the egg he cannot but be surprised that the young insect
should have been enclosed in so small a receptacle. The body of
the young Ranatra is as broad as the egg and over twice as
long. The length of the young, measured from the end of the
proboscis to the tip of the breathing tube is frequently 8 mm.,
while the egg itself is only about 3 mm.

The young in general appearance closely resemble the adult
form. The prothorax, however, is relatively shorter, the wings
are entirely absent, and the breathing tube, or what functions as
such, 3 is relatively short, being about one fourth of the length of
the body, while in the adult it is about four fifths the length of
the body. At first the young are very soft. The slender legs
bend with the greatest ease, and if the insect is taken out of the

'Holmes, "The Reactions of Ranatra to Light," Jour. Comparative Neurology
and Psychology, Vol. 15, p. 305. " Death Feigning in Ranatra," /. f., Vol. 16, p. 200.

2 Torre-Bueno, Canadian Entomologist, Vol. 38, p. 242, 1906.

3 Torre-Bueno, /. c., has shown that the so-called breathing tube of the young
nymph differs essentially in structure from that of the adult insect.

I 5 8

OBSERVATIONS ON YOUNG OF RANATRA QUADRIDENTATA 159

water it is unable to support itself and wriggles about helplessly.
Its integument rapidly hardens, however, and in the course of a
few hours it is able to move around out of water as well as an
older individual.

When first hatched the young are pale in color, but they soon
become much darker, many specimens becoming quite dark in
less than a day.

The manner of walking, swimming, turning over when placed
on the back, and the attitudes assumed when resting at the sur-
face of the water, or in contact with objects below the surface,
are very nearly the same as in the mature insect. Soon after
hatching the young Ranatras take up a position at the surface
of the water with the tip of the breathing tube just projecting
through the surface film and the body inclined obliquely down-
ward. The second and third pairs of legs are held in a sprawled
out position, while the first pair is held in front, and bent upward
at the middle, with the claw held open. The position is one of
readiness for quickly seizing any small object that passes within
reach. Young Ranatras are remarkably active in the capture of
prey. Any small object that strikes the outstretched arms is
grabbed at with surprising quickness. If a small insect or crus-
tacean is seized, it is drawn towards the mouth, usually with the
assistance of the other arm, and the proboscis is moved about
over it in the endeavor to find a soft spot through which it can
penetrate. When this is found the juices of the body are grad-
ually sucked out and the rest of the prey rejected. Young Ran-
atras will seize and suck out almost any animal not of too large
size. I observed one not a day old deftly catch a small ostracod
that happened to swim against one of its outstretched arms. The
ostracod tightly closed the valves of its shell, but the Ranatra
turned it over and over, exploring all sides of it with the tip of
its proboscis and endeavoring in vain to force an entrance between
the valves. After the round smooth ostracod was rolled about
for several minutes it slipped from the grasp of its captor and
swam away.

Ranatras may readily be fed by seizing a small organism in a
pair of pincers and carefully bringing it up to them. Animals as
large as themselves are successfully coped with. Large Hyalcl-

l6o S. J. HOLMES.

las are seized and sucked out without much difficulty. While
mature Ranatras will live together peaceably for a long time the
young readily attack and devour one another. If several young
are kept in the same dish it will be found that, in the course of
a few days, a majority of them will have fallen victims to a few
successful combatants. If a young Ranatra seizes another near
the middle of the body it is usually able to bring its victim up to
its proboscis, the tip of which is moved about in search of a soft
spot in the armor of the unfortunate individual, whose blood is
then deliberately sucked out notwithstanding the creature's
struggles. Often a Ranatra is seized by one of its legs. This
is usually not resented until its captor, after pushing the tip of
its proboscis along the leg until it finds one of the joints, begins
to insert its piercing stylets through the soft integument, when a
vigorous struggle ensues.

Young Ranatras are exceedingly voracious creatures, as they
will kill and suck out several insects as large as themselves in
the course of a day. Their food consists mainly of small swim-
ming forms, chiefly small crustaceans and insects, which come
near the surface of the water. Like the adults, they are very
efficient enemies of the larvae of mosquitoes. They do not pur-
sue their prey, and they seldom catch forms that keep in close
contact with solid objects. They are like so many traps set
ready to seize anything that comes in contact with them. Often,
however, an object is grabbed at if it passes near a Ranatra with-
out coming into actual contact with it. This action is probably
a response to the impact of the water. If a Ranatra is hungry,
touching the surface film with a needle near the insect will often
cause it to grab about wildly in the effort to seize whatever may
have caused the disturbance. An object of too large or threat-
ening appearance causes the young Ranatra to jerk back its first
pair of legs, but there are no efforts to swim away from danger.
When this reaction occurs the insect cannot be induced to take
food for some time.

The reactions of young Ranatras to light are not nearly so
vigorous and decided as those of the adult. A feeble positive
phototaxis is manifested the first day after hatching and increases
gradually as the insect grows older. Individuals a v/eek old are

OBSERVATIONS ON YOUNG OF RANATRA OUADRIUENTATA l6l

very often found swimming on the side of the dish towards the
light ; if the dish is turned about they quickly swim back again
to the light side. When out of water they are comparatively
irresponsive to light a fact in marked contrast to the behavior
of the mature insects. Movements of the head in response to
changes in the position of the light, which are so pronounced in
the adults are manifested in the young of a week old or even
less, but they are not very pronounced. When out of water the
young could not be induced to walk toward the light or respond
to it in any other way than by making rather feeble movements
of the head. While contact stimuli applied to the mature insects
when in the water cause a negative phototaxis, they failed in the
forms experimented with to produce this effect in the young.

The death feigning of young Ranatras is not so decided or pro-
longed as in older specimens, and it also differs in certain other
particulars. Young Ranatras when taken out of the water and
laid on a table frequently become immobile in whatever position
they may happen to lie. Neither in the young nor the mature
form do the appendages assume any definite position such as they
do in the death feint of many other insects. The feint is shown
during the first day of free life. The muscular system gives evi-
dence of a certain degree of rigidity, but owing to the flexibility
of the appendages this is not so clearly manifested as in some-
what older individuals.

In specimens five days old the death feint is more decided. Sev-
eral specimens of this age were taken out of a dish and laid on a
table. Immediately they all became immobile. They could be
picked up by one of the slender legs and held out without caus-
ing a bend in any of the joints, thus showing that the muscles
were in a state of extreme contraction. Many specimens would
endure considerable handling and poking about without making
any response. In some cases such treatment would bring them
quickly out of the feint, and all the forms experimented with were
brought out of it by more prolonged stimulation. In this respect
the young differ from the mature insects which will endure a
great deal of maltreatment without making any response.

In specimens coming out of the death feint handling or rubbing
them with a dry camel's hair brush produced in some cases a

1 62 S. J. HOLMES.

resumption of the feint ; in others it had no effect. This, too, is
different from the characteristic reaction of the adults which, when
they awaken from the feint, can readily be caused to resume
feigning many times in succession by handling or gently stroking
them. And if an adult is picked up while feigning it is rarely
brought out of its feint by this means.

It is a curious fact that while the mature Ranatra will endure
all sorts of maltreatment during the death feint, even suffering its
legs to be cut off one by one or its body cut in two without the
least response, the moment the insect is placed in the water the
death feint entirely disappears. There is no way in which the
feint can be terminated so quickly and completely as by this
means. Nor is it possible by any sort of manipulation to cause
the insect to feign death so long as it is in the water. It is cer-
tainly remarkable that an insect that will feign, it may be for
hours, with its muscular system tense so that all its appendages
are perfectly stiff should so completely and suddenly change its
behavior when it is placed in another medium.

In Ranatras of five days old or less the death feint often per-
sists for a time after they are placed in water. They do not, as
a rule, swim directly away, as the adults do, but frequently remain
motionless and apparently still for several seconds, or in some
cases for over a minute. When in the air the duration of the
death feint of the young is increased if they are kept wet. Speci-
mens that refuse to feign when rubbed with a dry camel's hair brush
can commonly be made to do so by rubbing them with a brush
dipped in water. And when specimens cannot be induced to
feign by this means they can generally be thrown into a feint by
dipping them into water and then leaving them on the table.
Dryness produces in the young a restlessness that is not shown
by the adult, a result not improbably due to their less ability
to withstand lack of water.

The young, like the mature forms, can be cut in two while in
the death feint without causing any response. A specimen which
was cut across near the middle of the body showed no movement
in either piece ; the legs were rigid, but after being poked about
for some time they began to move. Handling the anterior piece
failed to cause it to feign again, but when it was dipped into the

OBSERVATIONS ON YOUNG OF RANATRA OUADRI DENT ATA. 163

water and placed back on the table it feigned for several minutes,
the legs giving the same signs of muscular rigidity as before.
Several times it came out of its feint and as many times it was
caused to resume feigning by dipping it into water and placing it
on the table. Contact with a wet camel's hair brush would
readily cause it to feign, but a dry brush would produce no feint,
or but a very short one. Other experiments gave very similar
results.

It is difficult to understand how the death feint in Ranatra
can be of much value to it. While the European Ranatra
lincaris has been known to fly to lights at night (a rare occur-
rence apparently) I have been unable to obtain any evidence of
such a habit in any of our American species. In fact Ranatra very
seldom leaves the water of its own accord on account of any sort
of inducement, and one is. therefore strongly inclined to believe
that the death feint which is manifested only when the insect is
in the air is rather an incidental result of certain physiological
peculiarities of the organism than an instinct which has been
built up by natural selection for the benefit of the species. We
must adopt such a view, I think, regarding hypnotism in the
higher animals and in man ; for what selective value can it be to
a species, such as our own for instance, to possess the capacity
of being thrown into the hypnotic state ? The instinct of feign-
ing death is of unquestionable service to many forms, and it is
possible that in rare instances it may have proven of selective
value to Ranatra, but it is open to serious question if the instinct
in this form has been evolved because of its importance as a
means of protection.

The strong and at times almost violent positive phototaxis
which Ranatra exhibits presents another problem of the same
kind. Certain of the most striking features of this reaction are
manifested only when the insect is out of water. As a rule
Ranatra inhabits more or less shaded retreats among submerged
grass or weeds near the water's edge. It is kept in such situa-
tions, partly through the direct effect of its positive thigmotaxis,
and partly because contact stimuli (as shown in a previous paper)
cause it to become negatively phototactic. The positive photo-
taxis which appears when the insect is swimming freely in the

164 S. J. HOLMES.

water is certainly of no service to it in leading it into its accus-
tomed habitat if it should,, for any reason, become removed from
it. In the air and near a bright light Ranatra becomes, sooner
or later, strongly positive, often being wrought up to the highest
pitch of excitement in its efforts to reach the light. Of the util-
ity of such curious behavior it is indeed difficult to conceive a
reasonable explanation.
UNIVERSITY OF WISCONSIN,
MADISON, Wis.

A STUDY OF FUNDAMENTAL BARS IN FEATHERS.

OSCAR RIDDLE.

The structure and development of feathers have been studied
by many investigators. The pigments of feathers have also
been the subject of a very great number of researches. In spite,
however, of these numerous studies of feather structures and
pigments, we know almost nothing of structural differences be-
tween pigmented and non-pigmented areas, and nothing at all
of the causes which lead to the orderly and definite distribution
of pigment into the often complex color-patterns commonly
found in birds. In connection with a research directed to these
two points it was thought advisable to make a study of certain
defects which were known to appear occasionally in feathers.
It is a preliminary account of my results in this more restricted
field which, however, proved to be a rather significant one
that is to be found in this paper. The work was undertaken in
1904 under the direction of Professor C. O. Whitman in the
zoological laboratory of the University of Chicago. It is a
pleasure to acknowledge the help, encouragement and criticism
which Professor Whitman has given in connection with this work.

At the time I took up the study of the defects under consid-
eration, they had been reported but once, and this report had to
do with but a single specimen, a single plumage, and a single
defect in each feather. This was an account by R. M. Strong l
of "A Case of Abnormal Plumage" found in a hybrid pigeon.
Dr. Strong described and figured two types of abnormalities, and
concurred with Professor Whitman, who had reared the bird, in
the opinion that the defects were probably caused by malnutrition
during the growth of the Juvenal plumage. Besides this case,
Professor Whitman had observed these or similar defects in the
feathers of several of his birds, and my problem was to learn the
extent of their occurrence and to determine their cause. In
the course of my studies I have found still other abnormalities

BIOLOGICAL BULLETIN, November, 1902.

165

1 66

OSCAR RIDDLE.

or rather, different forms of the same defect and I shall first

give a description of the nature of the defects, and later consider

the question of their extent and cause.

Defects in Adult Morphology. In the adult expanded feather,

I have found five types of defects in structure. In Fig. I, a is

shown the first type. There is in
this case a sharply defined area ex-
tending entirely across the feather-
vane, in which there are no, or
very few, perfect barbules. A
cross-section of the feather at this
point would show only shaft and
barbs. One such area in the entire
length of the feather was one of the
types described by Strong. I find,
however, an abundance of cases
where such areas occur at regular
intervals practically throughout the
length of the feather. This regu-
larity in the spaces separating the
defects, indeed, furnished the clue
to the nature of the latter. It will
be seen that these areas cross the
Feather from a poorly barbs in such a way as to form

nourished chick showing abnormal!- a l mO st a right angle with them.

ties. (?, abnormal area ; b, ' funda- , . ,. .

f ,,,,,, .,. The same thing is true of the other

mental bar (a day s growth) ; c,

constrictions ;</, region in which de- types of defects and argues for

fective lines showed plainly in this their standing primarily all for the

feather. ( V 2 - ) 4.1

same thing.

The second type represents the greatest extreme to be met
with among these abnormalities. The feather in the abnormal
region has been reduced to shaft only ; both barbules and barbs
are gone. The second of the defects described by Strong evi-
dently belonged to this type, though he states that there was no
shaft present in his material and that its place was taken by a
small cylinder of fused barbs. I have not seen just such a
structure as he describes ; but it is rare and doubtless is to be
regarded as a sort of record of the very severest conditions which

FIG. i.

A STUDY OF FUNDAMENTAL BARS IN FEATHERS. l6/

a bird can encounter and endure. In types one and two the
barbs and shaft are often bent or kinked in the abnormal

region.

The third type of defect is something very much less conspicu-
ous than either of the two types already considered. It could
not be represented in the drawing. It is a very minute depres-
sion extending across the upper and dorsal surface of the feather.
It is not always easy, however, to determine that it is a depres-
sion at all. It often seems a line, or simply the point of union
of a distal with a proximal part of the feather-vane. This line
crosses a series of barbs making with them a right angle as did
the defective area of type one. These lines or depressions are
usually so inconspicuous that even close observation may not re-
veal them. Yet they exist and can be demonstrated in all feathers,
and at any level throughout the length of the feather.

The existence of these depressions as normal occurrences in
the feather is apparently nowhere mentioned in the literature.
Certainly their significance has not been made known. As may
be inferred from my classification of them, I have
found them to bear a close relation to the defective
feather areas. These lines which are thoroughly
characteristic of feathers are properly classified among
feather defects, for, it is always at these lines that the
defects like those of types one and four appear and,
moreover, they show all possible gradations into types
one and four. I shall hereafter speak of defects of
this type as defective lines, or depressions ; those of
type one as defective areas ; those of type four as con- FIG. 2. En-

Strictions. tire feather-

It is those feather-vanes which are made up of a germ rora

Car din arts

series of deep depressions or constrictions that show v i r<r j n i am{S
the defects of type four. I shall say nothing here showing con-
of the conditions represented in this type, but per- strict ions.

, r i r i (Actual length

haps an idea can be had from the feather germ shown ^ mm .
in Fig. 2.

Of type five, I have seen but a single example. In this the
defect extends vertically or the long way of the feather. The
barbs of one half of the vane have their distal portions broken

1 68 OSCAR RIDDLE.

away at even distances from the shaft. Duerden l states that this
abnormality is also very rarely met with among ostriches.

The recognition of the defective lines in all feathers throws a
new light on abrasion and wear in feathers. That there are birds
which "normally " have the barbules broken off at certain fairly
definite points in the more distal barbs has been observed by
Meves, 2 Chapman, 3 Dwight, 4 Strong 5 and others. Meves and
Chapman have noted, too, that the barb itself may be broken
near the distal end. I have seen several cases of the breaking of
a series of barbs at the point where they were crossed by the
same defective line, and I believe that further study will prove
that most feather abrasions occur by the space between two de-
fective lines breaking away as a single piece.

The Defects in Feather-germs. I have been able to observe
the defects in several formative stages. I shall say only a word
concerning them here. The prominent defects in the unexpanded
germ are easily recognized by the unaided eye ; sometimes they
appear as definite constrictions (Cardinalis virginianus, see Fig.
2, .), but very often as points of a different color (rectrices of
Turtur risorius). A microscopic examination of a region which
would develop a defective area shows a reduction of cell-growth
and division particularly in the region of the barbules.

EXTENT AND DISTRIBUTION ON THE ABNORMALITIES.

In the Bird Groups. In looking for the cause of the defects
one turned naturally to the birds to find whether they were wide-
spread or restricted phenomena. I stated that at the time the
present work was begun, there was in the literature but a single
account of them, and that account had to do with a single speci-
men a hybrid pigeon. Recently Professor J. E. Duerden has

Duerden, "Bars in Ostrich Feathers," Agr. Jour. Cape of Good Hope, May
1906.

2 Meves, W. , " Uber die Farbenveranderung der Vogel," Jour, fur Ornith.,
Bd. 3, 1855.

"Chapman, F. M., "On the Changes of Plumage in the Snowflake," Amer.
Mus. Nat. Hist., vol. 8.

1 Dwight, J., Jr., " The Sequence of Plumages and Moults in the Passerine Birds
of New York," Ann. N. Y. Acad. Set., vol. 13, No. I.

"> Strong, R. M., "The Development of Color in the Definitive Feather," Bull.
Mus. Comp. Zool., vol. 40, no. 3.

A STUDY OF FUNDAMENTAL BARS IN FEATHERS. 169

reported the abnormality in the ostriches, particularly in those
of South Africa. I learn from him by letter that he has under-
taken a thoroughgoing research to determine the cause of the
" barring " so prevalent in the ostriches. He estimates that the
value of the ostrich plumes from South Africa alone are from
this cause depreciated in value to the extent of ,250,000
annually.

The defects are, however, not confined to hybrid pigeons and
domesticated ostriches. I find them in the most widely separated
bird groups ; in primitive and in recent birds ; in land and in
water birds ; in domesticated and in wild birds ; in birds from the
arctic and from the torrid zone, etc. I have been able, owing to
the courtesies extended by Professor C. B. Cory and Dr. Ned
Dearborn, of the Field Columbian Museum in Chicago, to exam-
ine a very great variety of birds belonging to the Museum. I
find that although it is not easy to see evident defects in every
specimen, it is easy to find them in every species. We may con-
clude therefore, that they are to be found in all birds.

It is a fact, and a significant one I think, that the defects are,
in general, more common in domesticated and caged birds than
in wild birds. In this connection, however, it should be stated
that the defects appear indifferently in pure breeds, hybrids and
mongrels. At any rate I have verified this in a number of
our domesticated birds.

On Individual Birds. --\ have found the defects in all of the
plumages of the birds, with the possible exception of the first
or downy plumage. In some birds the defects seem to occur
more frequently in the juvenal (of D wight) than in the others.
The emphasized defects appear in all the feather-tracts or pter-
ylae ; but in a particular bird, and usually in a particular species,
certain tracts show them in greater numbers than do others.

/// an Individual Feather. In the feather there may be pro-
duced at any point in its length, either of the five types of ab-
normality. In some birds (Callus] the distal part of the feather
oftener shows the defective areas ; the proximal end, the deep
constrictions, while we get defective lines in one form or another
at every point in the feather's length.

I/O OSCAR RIDDLE.

THE MEANING AND CAUSE OF THE DEFECTIVE LINES AND
OF THE SPACES BETWEEN THEM.

We may now consider the significance of this blocking out of
the feather from end to end into bands, "bars," or plane feather-
elements, separated from each other by extremely faint depres-
sions or constrictions --for, my studies demonstrate that this is a
true conception of feather structure.

That the feather from tip to tip does not represent a perfect,
uniform continuity, but is made up of an apposed series of faint
"fundamental bars" is a conception which I owe to Professor
Whitman. I have proved absolutely that the defective lines, or
points of apposition of the "fundamental bars " are the points at
which all of the defects appear, and are therefore, really minia-
ture representatives of the defective areas and constrictions of
types one and four. I think I have also proved that each block,
segment or " fundamental bar " of the feather represents a day of
growth, and this is at the same time the amount of feather-growth
between two low blood-pressures. Further, I have abundant
evidence that the defective lines and areas represent points devel-
oped under a diminished rate of cell-growth and cell-division,
brought about by a reduced nutrition, which is in turn the result
of a daily lowering of the blood-pressure. This low blood-pres-
sure doubtless occurs between one o'clock A. M. and six A. M.

The evidence that a single "fundamental bar" and a single
defective line or area are laid down each day, and that this is the
total of a day's growth is conclusive. In very favorable material
I have been able to show, for example, that a feather 56 days
old shows 56 "fundamental bars" and 56 defective lines, areas
and constrictions. That the defective area is laid down at night
and during a period of low blood-pressure, I have demonstrated
twice experimentally. A chick was kept on two succeeding
nights, from 8 o'clock P. M. till 8 A. M. in an atmosphere con-
taining amyl nitrite (which lowers the blood-pressure). 1 This
bird later showed two emphasized defective areas in the region of
the feather produced during the two days of the experiment, and
these areas occupied the region normal to the defective lines and

1 The effect of several drugs on the blood-pressure of birds has been investigated
by Dr. S. A. Matthews and the writer. Our results are soon to be published.

A STUDY OF FUNDAMENTAL BARS IN FEATHERS. I/ 1

did not appear in the territory occupied by a " fundamental bar."
Since these defective lines are laid down at approximately the
same time each day as is proved by the regularity in the dis-
tances separating them we are forced to the conclusion that
the defective lines are normally laid down at night, and that a
lowering of the blood-pressure is associated with the production
of defective areas, and, therefore of defective lines, for, that the
defective line stands for the initial stage of the defective area is
as certain as that an area has more dimensions than a line. The
evidence in part is, that one sees all possible intergradations, that
each marks off a day's growth, that when the area occurs it
always falls in the place for the line, that a certain part of the
line only may be transformed into the obviously defective area,
etc. That there is a reduction of cell-growth and cell-division in
the defective area is proved absolutely by an examination of the
adult morphology of an exaggerated defect, as it is also by the
histology of the defects in the feather-germ.

That the low blood-pressure occurs at night is evidenced by
the experiment of the chick in the amyl nitrite. That it occurs
between midnight and six in the morning may be inferred from
the fact that the lowest daily temperature in birds falls between
these hours. Reasoning from the facts known in mammals we
may assume that the minimum blood-pressure coincides in point
of time with the minimum temperature. I have not been able to
get the daily blood-pressure curve of birds, owing to the diffi-
culty of doing so in birds of small size. The ostriches might
well be used for that purpose. At this point I may suggest that
the ostriches will doubtless cease to interpolate defective areas in
their plumes as soon as they can find the perfect diet, and the
various life conditions which will give them well-nourished bodies
and strong, effective circulations. After all, these two are one.

THE RELATION OF NUTRITION TO THE DEFECTS.
At the very beginning of this study, it was thought that the
defective areas stood in a certain relation to a faulty nutrition. A
number of experiments were made to determine this. A number
of young ring doves were alternately starved and fed, with the
result that in these birds the defective areas appeared in the

172 OSCAR RIDDLE.

juvenal plumage in great numbers. The same experiment was
tried on young chicks with the same result. It was noticed,
however, that notwithstanding the careful and plentiful feeding of
the control, an occasional defect could be found in their feathers
too. These experiments l showed that malnutrition is beyond
doubt the important factor in the production of the defective
areas, but apparently not the only one. An experiment was then
carried through to learn whether the defects found in the control
could be produced by the usual handling, of the birds, and per-
haps slightly crumpling their feather-germs. The results were
negative ; but it was found that when the feather-germs were
strongly cm inpled and broken in the region of feather growth, the
defects were readily produced.

During the progress of an experiment on some young chicks
(carried on for a quite different purpose) it was found that chicks
which were fed on the fat stain Sudan III produced the defective
areas in much greater numbers than did the control birds. It
was determined that the ordinary variations of light, temperature,
etc., did not cause the defects. The net result so far of all the
experiments for the determination of the cause of the defects,
indicated that those things which interfere with the nutrition of
the feather-germ will produce the defects, while those things not
capable of affecting the nutrition will not produce them. It is
easy to understand how a crumpling of the feather-germ would
temporarily interfere with the circulation within it. In the chicks
fed with Sudan III. it was evident that a sort of " starving "
effect was produced by it. By the time the experiments had
proceeded thus far I knew that a day of normal growth in a
feather is represented by a "fundamental bar" and a defective
line, and also that a defective line stands in close relation to a de-
fective area. This suggested that the defective areas in the con-
trol, and the defective lines in all feathers, are produced by an
internal factor with a definite rhythm, and that the rhythm is able,
like my experiments, to effect the nutrition of the feather ele-
ments. This recommended blood-pressure to me, and the ex-
periments were made with the result stated above.

Blood-Pressure and Temperature Rhythms. - - 1 shall not here

1 Partial results of these feeding experiments were communicated by letter to
Professor Duerden and were published in his paper, cited elsewhere.

A STUDY OF FUNDAMENTAL BARS IN FEATHERS. 1/3

attempt to explain the causes for the nightly fall of the blood-
pressure in birds. Let it suffice to say that we find a parallel
phenomenon in mammals. I wish further to call attention to the
fact that my demonstration that the lowest blood-pressure in
birds falls at night is evidence that the blood-pressure and tem-
perature curves of birds are similar curves, as they are known
to be in mammals. Of course, I have not showed a blood-
pressure curve for birds ; I have, however, located in a general
way the time of its minimum. I am in a position to confirm the
observations of Conn and Van Beneden l as to the temperature
cu-rves of birds. They worked
with pigeons. I have tempera-
ture curves essentially similar to
theirs, from ducks, ring-doves, P"\
and from chicks both old and z&J
young. The lowest temperature
occurs at about four to five o'clock

Loiv Blood-Pressure and the Nu-
trition of the Feather Elements. FIG. 3. Cross-section of a feather-
There remains to be indicated i erm in the re ion of growth. (Semi-

, , . . , . diagrammatic, magnified about loo dia-

some oi the histological relations , , , ,,

meters.) f>, barb-forming cells; ble,

of the capillaries and the feather- barbule-fonning cells; /.pigment cell;
elements which suffer from the ea t> capillaries; //, P ul p; /, outer

i r , i i sheath ; i.s/i. inner sheath.

lowering of the vascular tension.

I shall also outline the way in which the low pressure probably

acts.

Just as among the vertebrates we know that certain tissues,
e. g-., the liver cells, are kept always on the verge of asphyxia-
tion, so I believe are the epidermal cells of the growing feather-
germ taxed to their utmost to secure from the blood enough
nourishment to allow the rapid cell-division to proceed in full
swing. Where else in an adult vertebrate do we find a more
rapid growth and differentiation of tissue than we find in the
moulting of certain birds ? We may then expect to find here a
struggle for food when this becomes reduced in amount, and
those parts nearer the blood-supply should fare better than parts

1 Corin, G. , and Van Beneden, A., "La Regulation de la temperarure chez les
Pigeons," Archives de Biologie, Vol. VII., pp. 265-276, 1887.

174 OSCAR RIDDLE.

more removed. Now this is exactly what happens. The cap-
illaries (Fig. 3 cap] of the feather-germ lie nearest those cells
which enter into the formation of the barbs (Fig. 3, b] and these
are able to continue to grow even with a weakened food-supply ;
they too, though, are suppressed in cases of extreme starvation.
The cells which form the barbules (Fig. 3, ft) are not in contact
with capillary walls, and can utilize only the surplus of food which
filters through the barb-forming cells. With this fact in mind it
is clear that we should expect a diminished food-supply to first
check the growth in the barbules and that still further reduction
is necessary to check the growth of the barbs. Experience proves
that this is true (I use the words " food " and " nutriment " in a
broad sense, and oxygen is to be read into them). It is conceiv-
able that a reduced oxygen-supply is here playing a part, since
in all my experiments and in any normal lowering of the blood-
pressure, the available oxygen is decreased.

From what has just been said of the filtration method by which
the barbule cells receive their nutriment, we can now see how it
is that blood-pressure plays so important a part in the produc-
tion of defective areas. It is well known that when a period of
low blood-pressure sets in, the lymph begins to flow from the
spaces between the cells of the body into the capillaries. Thus,
by withdrawing a quantity of food from the immediate environs
of the cell, a low blood-pressure affects the cell in the same way
as does an actual reduction of the amount of nourishment circu-
lating in the blood.

It appears, therefore, that the feather germ as it grows and
unfolds, spreads before us a record of some earlier significant oc-
currences within. Indeed, it now seems certain that the delicate
filaments, so admirably interlaced to form a feather-vane, are as
capable as a revolving drum of recording the important changes
in vascular pressure. To be sure, the tracings on the plumes are
not to be measured as symmetrical curves with a definite num-
ber of millimeters of daily variation, but are written in "funda-
mental bars " separated by areas more or less imperfect which are
to be read in terms of cell-growth and cell-division.

HULL ZOOLOGICAL LABORATORY,
UNIVERSITY OF CHICAGO,
December, 1906.

STUDIES ON THE RELATION BETWEEN AMITOSIS

AND MITOSIS.

II. DEVELOPMENT OF THE TESTES AND SPERMATOGENESIS IN

MONIEZIA.
C. M. CHILD.

The material for this study was obtained from the same spe-
cies, vis., Moniezia expansa and Moniezia pianissimo,, as that of
the first paper of this series * and most of it, in fact from the
same chains. The methods of fixation, observation and record
are the same as those already described in that paper. Various
details in the spermatogenesis are only briefly considered, since
they are not directly connected with the chief purpose of the
paper.

I. The Formation of the Te&tes.

The testes develop from cells of the parenchyma which do not
differ visibly from other cells of the same region. They appear in
the dorsal region of the central parenchyma. Before their appear-
ance two kinds of cells are visible in the parenchyma : one of
these is smaller and surrounded by more or less cytoplasm ap-
parently without definite boundary and usually elongated in the
dorso-ventral direction, with one or more fibrillar extensions at
each end. Figs. I, A-D (PI. VII.), show parenchymal cells of this
kind in the earliest stages of testis formation. Amitotic division
of the nucleus is occurring in each case. Fig. I, A, shows a case
in which the two parts of the dividing nucleus stain differently
Fig. i, C, a case of the endogenous form of nuclear division
which was described in the preceding paper, while Figs, i, B,
and i, D, show late stages in division, the one by constriction,
the other by formation of a nuclear plate. For a description of
these and other forms and stages of amitosis in Moniezia the
reader is referred to the preceding paper of this series.

The other form of cell existing in the parenchyma before testis

1 Child, " Studies on the Relation between Amitosis and Mitosis I. Development
of the Ovary and Oogenesis in Moniezia," BIOL. BULL., Vol. XII., No. 2, 1907.

175

1/6 C. M. CHILD.

formation is much larger, and its cytoplasm is highly vacuolated
and more distinctly bounded from the parenchymal matrix about
it. These cells appear always to be connected with deeply
staining fibers which extend dorso-ventrally across the central
region of the proglottid and which I take to be dorso-ventral
muscle fibers since they and the cells connected with them are
similar to those of the muscular layers. Fig. 2, A (PL VII.),
shows one of these cells with a portion of its fiber. The fiber
passes directly through the cell-body on one side of the nucleus
and the cytoplasm extends visibly for a considerable distance
along the fiber. The evidence that these cells develop into
testes is very strong. Fig. 2, B (PI. VII.), represents a section
through nucleus and body of one of these cells in w r hich the
nucleus is apparently undergoing amitosis. Whether this par-
ticular case would have developed into a testis it is of course
impossible to determine. But Figs. 2, C, and 2, D(P\. VII.), rep-
resent characteristic cases slightly more advanced. Here sev-
eral nuclei are contained in a space which corresponds closely
with the form of the muscle cell and contains what seem to be
strands of the old vacuolated cytoplasm, while about some of
the nuclei a layer of more deeply staining cytoplasm is visible,
apparently in process of formation. Through the space which
apparently represents the region previously occupied by the body
of the muscle cell passes the fiber. The presence of the fiber
and the well marked outline of the space seem to me to consti-
tute very strong evidence in favor of the conclusion that each of
these groups of nuclei have arisen by the division of a nucleus
of a muscle cell. That these groups develop into testes there
can be no doubt. Their development can be followed from pro-
glottid to proglottid without the slightest difficulty, and there
are no other similar groups of nuclei in the parenchyma. Fig.
2, E (PI. VII.), shows a case in which the muscle fiber is appar-
ently undergoing degeneration. In all of these cases amitotic
division of the nuclei is taking place. The figures give only a
few examples of the cases observed. In a brief account of the
history of these cells already published additional figures are
given. 1 If these observations are correct, and I have, so far as

1 Child, "The Development of Germ Cells from Differentiated Somatic Cells in
Moniezia." Anat. Anz., Bd. XXIX., Nos. 21 and 22, 1906.

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 1/7

I am aware, taken all possible precautions to assure myself that
they are, it seems impossible to escape the conclusion that in
Moniesia the male germ-cells may develop from cells which
have previously been differentiated and functional in the soma.
Generalization from this conclusion is, however, no more justi-
fiable than is that so often made from observations which seem
to point in the opposite direction. It does not seem probable
that uniformity exists here any more than in other features of
development in regard to which premature generalizations mark
like wrecks the dangers along the channel of biological thought.

In slightly later stages it is difficult or impossible to determine
with certainty which testes have arisen from the smaller and
which from the larger cells. Figs. 3, A-%, F(P\. VII.), represent
young testes which probably developed from the smaller paren-
chymal nuclei : amitosis is visible in all cases except Fig. 3, F, in
which mitosis is occurring. In the hundreds of testes examined
at this stage four cases of mitotic division have been observed of
which one is shown in this figure. All of the observed cases
were found in a single chain of Moniesia expansa, another fact
which seems to indicate that the relative frequency of the two
forms of division may vary. It was not possible to determine
the number of chromosomes with certainty, but it was more than
twelve, the number shown in the figure.

Figs. 4, A 4, D (PI. VIII.), represent cases which probably de-
veloped from the large cells-- muscle cells. Such stages are
found in the same proglottids as the stages shown in Figs. 3,
A F, and are clearly larger and contain more nuclei than the
latter. Fig. 4, A, shows a case in which a small nucleus, appar-
ently not a part of the large mass, is also seemingly involved in the
development. Cases of this sort are not infrequent, and the
small nucleus often becomes one of the membrane-nuclei, though
the latter appear in many cases to arise from the same primordium
as the germ-cells themselves. Here, as in the development of the
female organs, it is difficult to resist the impression that the de-
velopment of these organs is the result of some localized stimu-
lus or condition and that any cells within reach of this factor may
become involved.

At this stage the nuclei of the developing testis lie in a con-

1/8 C. M. CHILD.

tinuous mass of cytoplasm, which is more or less distinctly
marked off from the parenchymal substance, but still shows
fibrous extensions into the parenchyma. The development of the
testes differs from that of the ovary in that none of the parenchy-
mal substance is included within the testis. This difference is
merely the consequence of the fact that the testes usually develop
from a single nucleus and its surrounding cytoplasm, while the
ovary develops from a large number of the parenchymal nuclei.

II. The Gro-ii'tli of tlic Testes Preceding iJie Spireine Stage.

During the period preceding the appearance of the spireme in
a part of the cells the development of the testes consists of in-
crease in size as the result of numerous divisions, chiefly amitotic
and of formation of the membrane about the testis and of the
vasa efferentia.

The formation of the membrane occurs at an early stage from
the cytoplasm of cells about the periphery of the proliferating
mass. Fig. 5, . / (PI. YIII.), shows one of these cells with the
membrane forming an extension of the cytoplasm. The vas
efferens is formed in the same manner (Fig. 6, A, PI. VIII.). The
membrane-forming cells are lew in number and the nuclei appar-
ently undergo degeneration in later stages, for they are very rarely
fou ml in the fully developed testis.

The contents differ somewhat in appearance according to the
method of fixation employed. After Hermann or chrom-oxalic
the nuclei appear in most cases to be imbedded in a syncytial mass
of cytoplasm, cell boundaries being indistinguishable or some-
times faintly visible, though cavities or vacuoles are frequently
observed. After sublimate and some of the sublimate mixtures
in all but the earlier stages the cytoplasm appears to have under-
gone shrinkage and to be more or less definite!}' concentrated
about each of the nuclei. In later stages the individual cells ap-
pear more distinct after any fluids. The earliest stages of testis-
development are certainly syncytial and without doubt the indi-
vidualization of the cells takes place gradually. In consequence
of the shrinkage of the delicate and probably highly fluid proto-
plasm caused by the sublimate fluids the distinctness of the cells
is exaggerated. Most of the figures of these stages (Figs. ;,

THE RELATION BETWEEN AM1TOSIS AND MITOSIS. 1/9

A-6, B, PI. VIII., 7, A-S, B, PI. IX.) are taken from the chrom-
oxalic preparations and cell-boundaries do not appear in most
cases though they may be present. Fig. 9, D (PI. IX.) is taken
from a sublimate preparation at a rather late prespireme stage
and shows the cells as distinct.

But for present purposes the nuclei are of chief importance.
Figs. 5, A$, C (PI. VIII.) represent stages just after the forma-
tion of the membrane. In Figs. 5, A, and 5, B, amitoses are
visible and in Figs. 5, B, and 5, C, two of the very infrequent
cases of mitosis in these stages are shown. Figs. 6, A, and 6, B
(PI. VIII.) are from slightly later stages : Fig. 6, A, is a section
through one side of a testis and does not show the full size.
Both figures show amitoses and Fig. 6, B, shows one case of
mitosis. Figs. 7, A, and 7, B (PI. IX.), are from still later
stages. In the latter figure one case of division of a nucleus
into three parts is shown. The nuclei in which such divisions
take place are usually larger than the others, often lie near the
center of the testes and are similar in appearance to the large
nuclei along the axis of the developing ovarian follicles which
were mentioned in the preceding paper. In Figs. 8, A, and 8, B
(PI. IX.), stages just preceding the first appearance of the spireme,
each with several amitoses, are shown. In Fig. 8, A, one of the
large nuclei dividing into three parts is seen near the opening of
the vas efferens.

In Fig. 9, A-Q, D (PI. IX.), cells or cell-groups from various
stages containing amitoses of special interest are shown. Figs.
9, A, and 9, C, represent cases of the form of amitosis designated
endogenous in the preceding paper in which the two nuclei result-
ing from division do not occupy the entire space within the old
membrane. Numerous cases of this sort have been observed in
the testes and have been examined with great care. Fig. 9, B, is
a case of triple division and Fig. 9, D, a case of cytoplasmic divi-
sion proceeding from within outward. This figure is from a sub-
limate preparation.

During these stages mitoses are rare. Very often not a single
case is found in any of the numerous testes of a proglottid. In
other proglottids a number of testes may show one or more each.
In general the relative frequency of mitosis appears to vary in

ISO C. M. CHILD.

different chains and in different proglottids. In one chain of M.
pianissimo, for example mitosis has been observed only very
rarely though the chain has been carefully examined ; in another
it was found to be much more frequent. In all cases, however,
amitosis is the predominant form of division in these stages.

V

III. Formation of the Spireme and the Growth Period.

In the stages before spireme-formation, or as it has often been
called, synapsis, all the nuclei in the testes are similar in appear-
ance and contain a large deeply staining nucleolus with perhaps
a few smaller granules.

Suddenly a part of the nuclei begin to increase in size and a
spireme appears (Figs. 10, A-IT,, B, PI. X.). The formation of
the spireme takes place in the manner described for the ovary in
the preceding paper of this series. The change does not appear
to begin in any particular region of the testis. Sometimes differ-
ent groups of cells in different regions of the same testis give rise
to a spireme while about them and between them lie others still
unchanged and undergoing amitosis. From the first appearance
of the spireme in the testes until the formation of spermatozoa is
completed the multiplication of the spermatogonia which remain
in the prespireme stage goes on, chiefly or wholly by amitosis
and some of the cells thus produced are continually passing into
the spireme stage. Consequently the stage is not characteristic
of any particular period of development of the testis as a whole
after its first appearance ; in the older testes some groups of cells
in the spireme stage and some groups of spermatogonia in pre-
spireme stages are always to be found.

In some testes before the spireme stage appears and frequently
afterward some of the cells are seen to be more or less pear-
shaped in form with the pointed ends radially arranged about a
center and united by strands of cytoplasm (Figs. 10, A, 11, B,
PI. X.). The number of cells in a group of this kind varies from
three or four to eight or ten. All the cells of a group pass into
the spireme stage simultaneously. Whether such groups are due
to the persistence of cytoplasmic connections from previous divi-
sions or to the formation of new connections it has been impos-
sible to determine, but it seems possible from the varying size of

THE RELATION BETWEEN AMITOSIS AND MITOSIS. iSl

the groups that they are merely the result of connection of cells
lying near each other.

The grouping of the cells is not by any means a characteristic
feature in these stages. Frequently cells which are entirely
separate from each other pass into the spireme stage simultane-
ously (Figs. 10, B, n, A, PI. X.). As will appear below, the
grouping is merely the first step in a process characteristic of
spermatogenesis and observation indicates that in some cells it
begins before the spireme appears in others not until later.

As in the ovary the spireme is usually massed at one side of
the nucleus as in many other cases of synapsis and is often visibly
connected with the nucleolus (Figs. 12, A-12 B, PI. X.), which,
however, does not decrease in size but increases as the nucleus
grows larger (Figs. 10, A-1%, B, PI. X.).

The appearance of the spireme is accompanied by an increase
in the amount of cytoplasm. Comparison of the cells in this
stage with those in earlier stages in Figs. 10, A, 10, B, and 1 1, A
(PI. X.), shows this difference clearly. The cytoplasm in the
spireme stages also appears somewhat more dense in structure
and stains a little more deeply. Here as in the ovary the ap-
parent connection of the nuclear changes with the growth of the
cytoplasm is most striking.

As is usual, this stage in the testes is not accompanied by such
extreme growth of the cytoplasm as in the ovary nor by any
formation of yolk, but is soon followed by the spermatogenetic
divisions.

But the stages following the spireme are not the same in all
cells ; the later development follows two very different lines. In
the later stages of the spireme period it is possible to distinguish
two different sorts of nuclei. In the one (Fig. 13, A, PI. X.)
the nucleolus has disappeared and the spireme is very dense and
occupies almost the whole periphery of the nucleus. Careful
examination and comparison has convinced me that these nuclei
are in preparation for the first spermatocytic mitosis. The nuclei
of the other sort are considerably larger (Fig. 13, B, PI. X.), the
spireme is much less dense and more irregular in form and does
not occupy the whole periphery but is still massed more or less
at one side and the nucleolus is still intact. I am confident that

l82 C. M. CHILD.

these nuclei do not represent earlier stages than^those in Fig. 13,
A, for they are always larger than the latter and the irregularity
of the spireme is not found in the earlier stages. These nuclei
are the first stages in a remarkable process of fragmentation
which will be described in a later section. Their relative fre-
quency as compared with the others appears to vary in different
chains, proglottids and regions. Sometimes they seem to be
more, sometimes less numerous than the others.

(To be continued.)

184 C. M. CHILD.

EXPLANATION OF PLATES.

PLATE VII.

FIG. i. A-D, testes forming from parenchymal cells_: amitotic division in all
cases.

FIG. 2. A, muscle cell ; B, muscle cell undergoing amitosis ; C, D, E, muscle
cells developing into testes ; in E the muscle fiber is apparently undergoing de-
generation.

FIG. 3. A-F, young testes; in Fa case of mitosis.

BIOLOGICAL BULLETIN, VOL. XII

PLATE VII

B

D

3

186 C. M. CHILD.

PLATE VIII.

FIG. 4. A-D, young testes showing amitosis.

FIG. 5. A-C, young testes after formation of membrane and vas efferens ; B and
C contain mitoses.

FIG. 6. A, B, developing testes.

BIOLOGICAL BULLETIN, VOL. XII

B

5

B

1 88 C. M. CHILD.

PLATE IX.

FIG. 1 . A, B, developing testes.

FIG. 8. A, B, developing testes in later stages.

FIG. 9. A-D, amitoses from developing testes.

BIOLOGICAL BULLETIN, VOL. XII

PLATE IX

7

B

19 C. M. CHILD.

PLATE X.

FIG. 10. A, B, Fig. ii; A, B, the spireme stage. Figs. 10, A, and II, B,
show the first stages in fusion of the spermatocytes to form a cytophore.

FIG. 12. A, B, spireme stages.

FIG. 13. A, preparation for first spermatocytic mitosis; B, preparation for frag-
mentation.

BIOLOGICAL BULLETIN, VOL. XII

PLATE X

12

Vol. XII. March, 1906. No. 4.

BIOLOGICAL BULLETIN

STUDIES ON THE RELATION BETWEEN AMITOSIS

AND MITOSIS

II. DEVELOPMENT OF THE TESTES AND SPERMATOGENESIS IN

MONIEZIA (Continued}.

C. M. CHILD.

IV. The Spennatocytic Mitoses.

The appearance of the chromosomes of the first maturation
division follows the stage shown in Fig. 13, A (PI. X.). Figs. 14,
A, 14, B, and 15 (PI. XI.) show this stage in M. e.vpansa, Figs.
16, A-i6, C (PI. XL), in M. pianissimo,. In the former species
eight of these chromosomes have been counted in the nuclei in
some twenty- five cases (Fig. 14, A] and in no case have more
than eight been found. In many cases, however, it has been im-
possible even with the utmost care in examining successive sec-
tions, to find eight, only seven or six (Fig. 14, B] being visible.
As regards the number in M. pianissimo, the results are less defi-
nite. In some cases eight (one of the nuclei in Fig. 16, C) in
others nine (the other nucleus in Fig. 16, C) have been counted
and in one (Fig. 16, B} case thirteen distinct masses of chroma-
tin were visible in one nucleus. This, however, was probably an
earlier stage for some of these masses appear to be grouped in
pairs. Fig. 16, A, shows two nuclei of M. pianissimo, in which
five of these chromosomes are visible. It was impossible to de-
termine whether parts of these nuclei were in the next section.
One reason for the greater uncertainty in regard to M. pianissimo,
lies in the fact that these stages were much less frequently seen
in my sections of this species than in those of M. expansa. I
believe that the utmost caution should be used in observations of

191

C. M. CHILD.

this kind. In a study of cytological literature it is difficult to
resist the impression that some of the apparent uniformity in
cytological phenomena which appears so remarkable and mys-
terious is in reality the result of the selection of certain " typical '
cases and the discarding of others. While I should not venture
to assert positively that the number of chromosomes in either of
these species of Moniezia is actually variable, this certainly appears
to be the case, and the evidence given in support of the conclusion
that the number of chromosomes in a given species or variety is
invariable has not always carried conviction to my mind.

However, the description of the spermatocytic divisons is not
the chief object of the present paper and although I have spent
much time in endeavoring to reach' well-founded conclusions in
regard to the number of chromosomes in Moniczia it has not
been possible for M. pianissimo.. In M. expansa, as noted above,
the maximal number counted during maturation was eight.

In some of the nuclei at this stage these chromosomes appear,
as in many other forms in a more or less regular grouping about
the periphery of the nucleus. The figures do not show this con-
dition particularly well but Figs. 14, A, 14, B, 16, A, 16, C (PI.
XL), were drawn from nuclei in which the chromosomes were
thus arranged. Such nuclei are recognizable at once in a section.
This condition, however, was not very commonly found ; much
more frequent was the condition shown in Fig. I 5 where they were
irregularly disposed, some at the periphery, others near the cen-
ter. It is possible that the stage of peripheral arrangement is of
relatively short duration and is therefore less frequently seen, but
there is also a possibility that it is not of universal occurrence.
The question may be left open for the present.

In Fig. 17 (PI. XI.) two of these chromosomes are shown, both
from nuclei with membranes still intact. At the stage figured
the fusion of the two parts is not yet complete. So far as actual
observation goes these chromosomes of the first maturation divi-
sion are dyads, not tetrads, no indication of quadrivalence hav-
ing been observed at any stage, and although they correspond as
regards ultimate fate with the tetrads of various authors, I prefer
to designate them as what they appear to be --dyads. In con-
sequence of their small size, almost spherical form in late stages,

THE RELATIONS BETWEEN AMITOSIS AND MITOSIS. 1 93

and close approximation in the spindle no data regarding the
direction of the two divisions could be obtained.

As stated in the preceding section the spermatocytes often
appear in groups, the pear-shaped members of which are con-
nected at the pointed end by strands of cytoplasm. As the first
spindle forms the fusion of the cells proceeds (Fig. 18, PI. XI.) until
in many cases the metaphases of a number of nuclei lie in a con-
tinuous mass of cytoplasm, the cytophore (Fig. 19, PI. XL). Often,
however, the central region of the mass consists for a consider-
able time of a large space traversed by strands of cytoplasm (Figs.
1 8, 20, 23 (PI. XL), 25 (PI. XII.). The strands of cytoplasm are
represented in a somewhat diagrammatic manner.) In other
cases the first spermatocytes are entirely isolated (Fig. 21, PI. 1 1).

The spindle appears to be formed largely from the nuclear sub-
stance. The fibers are very delicate and no asters are visible
(Figs. 1 8 and 19, PI. XL). At the poles are very minute, but
fairly distinct deeply-staining, centrosomes. In the masses re-
sulting from the fusion of several spermatocytes the spindles are,
so far as observed, always tangential or nearly so (Figs. 19 and
20, PL XL). Just before their division the dyads assume the
form seen in Fig. 22 (PI. XL), where one is viewed from the side,
the other from the surface. Fig. 20 (PI. XL) shows the ana-
phase. Division of the centrosomes has not been observed, as
they cannot be distinguished after the chromosomes have ap-
proached the poles. Fig. 23 (PI. XL) shows the telophase of the
first spermatocytic division in a group. The cytoplasm of the
cells which formed the group now constitutes a cytophore in
which the masses of chromatin lie. It is less dense in appear-
ance and stains less deeply than in earlier stages, where it was
concentrated about particular nuclei. The chromatin masses
remain at the periphery and the central region is often still more or
less vacuolated. Where isolated spermatocytes divide the result
is the same : no division of the cytoplasm follows nuclear
division, and a cytophore differing from the group-cytophore
merely as regards size is formed.

The second division follows the first, in most cases apparently
without the formation of a " resting nucleus " between the two.
In a few cases, however, nuclei larger than spermatid nuclei and

194 c - M - CHILD.

containing irregular strands and masses of chromatin as if in
preparation for a division were formed about the periphery of a
cytophore (Fig. 24, PL XL). These were very probably stages
following the first spermatocytic division.

Fig. 25 (PI. XII.) shows the metaphase of the second sperma-
tocytic division in a group cytophore. Here again the divisions
are more or less nearly tangential. Fig. 26 (PI. XII.) shows the
anaphase in a portion of a group.

V. The Fragmentation of Spermatocyte Nuclei.

In Section III. (pp. 181-182) it was stated that certain of the
nuclei of the first spermatocytes become larger than the others, and
that the spireme instead of becoming more dense and giving
rise to chromosomes becomes less dense and very irregular
and the nucleolus remains instead of disappearing (Fig. 13,
B, PI. X.). These nuclei constitute the earliest recognizable
stage in a remarkable process of fragmentation which is appar-
ently a normal phenomenon in the testes of both species of Mon-
iezia. The process is so entirely different from anything de-
scribed in the spermatogenesis of other forms that at first I
regarded it as a form of degeneration. But repeated examina-
tion of old and new material during four successive years has
convinced me that the process gives rise to nuclei which are in-
distinguishable from the spermatid nuclei which have arisen by
mitosis. Whether these "spermatid nuclei" resulting from the
fragmentation of spermatocyte nuclei actually take part in the
formation of functional spermatozoa it is impossible to determine,
but that spermatozoa structurally similar to those developing
from nuclei which arise in the ordinary manner develop from
these spermatids is, as will appear probable.

This process, like the typical spermatogenesis, occurs in iso-
lated cells or simultaneously in groups. In the latter case the
different members of the group fuse together, in the manner de-
scribed in the preceding section, and this cytoplasm forms a
cytophore. It is rare, however, that groups so regular in form
and arrangement as that shown in Fig. 27, A (PI. XII.), are found
More commonly only two or three cells fuse together or the
process occurs in isolated cells. Fig. 27, A (PI. XII.) shows a

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 195

group of the nuclei in most of which the spireme is apparently
breaking up and disappearing. It stains less deeply than in earlier
stages and appears to be separating into irregular masses. Fig.
27, j5(Pl. XII.), shows an exceptionally clear case, in which some
portions of the spireme are apparently breaking up into granules.
Figs. 27, C, and 27, D (PI. XII.), show still other cases. It will
be observed that the nucleolus retains its staining power : as a
matter of fact it undergoes no visible change.

Except for the spireme or its remnants the nucleus shows with
the usual degree of extraction no visible structural features. As
it increases in size the nuclear membrane becomes increasingly
difficult to distinguish and finally disappears, so that what was
formerly the nucleus appears merely as a cavity in the cytoplasm
containing the nucleolus and irregular shreds and granules of
chromatic substance usually situated at one side (Figs. 27, B, 27,
C, 27, D, 28, PI. XII.). The whole appears as if degenerating.
In these and the following figures the boundaries of the nuclear
cavities and the spaces in the cytoplasm which remain in the
later stages are indicated by broken lines.

But the observations now to be described afford very strong
evidence in favor of the view that these nuclei do not undergo
complete degeneration. In and about these cavities containing
the remains of the spireme and the nucleolus very small new
nuclei with one or a few small granules of chromatin appear to
be formed. In Fig. 28 (PI. XII.), one such nucleus is shown
in one of the cells : in Figs. 29, A, and 29, B (PI. XII.), two appear
in each cell : in Fig. 29, C (PI. XII.), is one a little larger than
the preceding ; in this case the fragmenting nucleus lies in a
cytophore with spermatids already formed and from one of which
a spermatozoon is developing : in Fig. 29, D (PI. XII.), are again
two of the small nuclei. The nature of the process appears to
be somewhat as follows : as the old nuclear membrane breaks
down the cytoplasm encroaches more or less upon the nuclear
cavity and a new membrane forms about some of the particles or
masses of chromatin. This is not very different from what occurs
in various forms when a single chromosome or a few chromo-
somes are separated from the rest during division.

But the new nuclei are not always so small as these shown in

196 C. M. CHILD.

Figs. 28 and 29, A-2g, D. Figs. 30, A-C (PI. XIII.), show
cases in which they are somewhat larger, and in which the
method of their formation is more clearly visible. In these cases
the new nuclei are nearly hemispherical in form and appear as if
growing or budding out from the cavity representing the old
nucleus into the cytoplasm about it. Whether the method of
formation is actually the same in all cases is doubtful ; some-
times the new nuclei seem to lie wholly within the cavity as in
Figs. 29, A and 29, B (PI. XII.). But there can be no doubt of
their formation. The newly formed nuclear membrane is quite
distinct and the whole process presents a very characteristic
appearance.

The number of deeply staining bodies visible within these
small nuclei depends on the fixation and staining. If extraction
is carried beyond a certain point only one dark body retains the
stain : otherwise several may be visible, one or two of which are
usually larger than the others. After chrom-oxalic the stain is
more readily extracted from all except the one body and with
the usual degree of extraction only the one body appears. Figs.
29, ^29, D, are from such preparations. In nearly all cases
the nuclei stain somewhat more deeply throughout than the
cytoplasm as is indicated in the figures by stippling.

Whether the number of new nuclei formed is always the same
is uncertain but it is probable that as many as three or four and
perhaps more new nuclei may arise from one old nucleus. In
Fig. 29, E (PI. XII.), four small nuclei lie near the lower nuclear
cavity but some of these may have arisen from another nucleus.

Figs. 3136 (PI. XIII.), show what I regard as later stages in
this process. In the group shown in Fig. 3 1 two spermatocytes
were apparently involved and the two old nuclear cavities are
still visible, though very irregular in form and divided by strands
of cytoplasm. One of the old nucleoli is visible, the other lying
in another section. Judging from their position three of the
small nuclei were formed about the cavity on the right and one
on the other : in the next section more small nuclei were found
which probably belonged to the nuclear cavity on the left but in
this section other adjoining nuclear cavities with other small
nuclei appeared, so that certainty was impossible. It is of in-

THE RELATION BETWEEN AMITOSIS AND MITOSIS.

terest to note that some of the nuclei in the figure are distinctly
hemispherical as if they had been formed in the manner repre-
sented in Figs. 30, ^-30, C.

In Fig. 32 (PI. XIII.) three cavities appear, the old nucleolus
being visible in the one on the right. In the middle cavity there
are four nuclei. Fig. 33 (PI. XIII.) shows another case in which
portions of two nuclear cavities are visible, each with its nucleo-
lus, and seven nuclei, four about one cavity, three about the other.

In Figs. 34, 35 and 36 (PI. XIII.) groups of nuclei are shown
lying in cavities of the cytophore : nucleoli were not visible in
this section and it was impossible to determine how many sper-
matocytes were involved in the formation of these groups. That
such groups as these are formed by fragmentation there can, I
think, be no doubt. In all cases where the spermatocytes divide
mitotically in groups the spindles lie peripherally and nearly
tangentially in the cytophore (Figs. 19, 20, 23, PI. XI.; 25, 26,
PI. XII.), and the nuclei formed lie at or near the periphery
(Figs. 24, PI. XL; 42, 43, 44, PI. XIV.). There are no grounds
for supposing that they migrate from the periphery and return
to it ; moreover, it is difficult to see how nuclei grouped as those
in Figs. 34 and 35 (PI. XIII.) are grouped qould have been formed
in situ by mitosis. It seems probable that this difference in posi-
tion is quite sufficient to distinguish the nuclei formed by frag-
mentation of spermatocytes from those formed by the spermato-
genetic mitoses.

Assuming that these nuclei have arisen by the fragmentation
of spermatocytes, the number of chromatin granules has in-
creased beyond that contained in some of the nuclei at the time
of their formation and the nuclei are now somewhat larger than
some of those. No further growth takes place, however. The
nuclei themselves are indistinguishable from those spermatid nu-
clei which lie at the periphery of the cytophores and are doubt-
less the products of mitoses (Figs. 42, 43, PI. XIV.).

Figs. 37 (PI. XIII.) and 38 (PI. XIV.) represent larger areas
of two testes showing how the nuclei which, judging from their
position in the cytophore, have arisen by fragmentation are situ-
ated in relation to other stages. No trace of definite position or
arrangement of the various stages in the testis could be found.

o o

198 C. M. CHILD.

But these nuclei apparently do not remain massed together in
the cytophore. There is every reason to believe that they mi-
grate to the periphery. Very frequently cytophores with several
old nucleoli imbedded in different regions of their cytoplasm, in-
dicating that fragmentation has occurred, show all their nuclei at
the periphery as in the case of spermatids formed by mitosis.
Since I have never found first spermatocytic mitoses and frag-
mentation occurring in the same cytophore and since the nucle-
olus disappears before mitosis it seems probable that the presence
of the old nucleoli which stain very characteristically and are
readily recognizable is sufficient to identify a particular cytophore
and its spermatid nuclei, for such, I believe we may call them, at
least so far as appearance goes, as the result of fragmentation.

Fig. 39 (PI. XIV.) is probably a stage in the migration to the
periphery of spermatid nuclei formed by fragmentation. At the
right of the figure one of the old nucleoli, itself near the peri-
phery, and a part of an old nuclear cavity containing a few
granules are visible : another lies near the middle of the cyto-
phore and still others were found in the same cytophore in
other sections. Some of the nuclei have already reached the
periphery of the cytophore and the cytoplasm about them bulges
from its surface. Most of these nuclei are more or less hemi-
spherical with convex surface toward the periphery and a space
adjoins the flattened side : this condition is probably connected
with the migration of the nuclei toward the periphery. Indeed, it
is not impossible that this migration may not be a bodily movement
through the cytoplasm of the cytophore, but rather the continued
formation of new nuclear membrane on the peripheral side and
its continued disappearance on the proximal side. In any case
the change of position is probably a " tactic " reaction of some
sort.

In two testes cytophores have been found containing nuclei
intermediate in size between those of the spermatocytes at the
time of fragmentation and those of spermatids, but which ap-
peared as if preparing for fragmentation. Fig. 40 shows one of
these cases. These two groups may possibly be cases of second-
ary fragmentation. Occasionally as in Fig. 30, A (PI. XIII.), the
nuclei arising by fragmentation are unusually large and perhaps

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 199

undergo fragmentation again. At any rate such cases are rare
and not of fundamental importance. In one case a mitotic spindle
was observed in a cytophore whose nuclei had apparently arisen
by fragmentation since two nucleoli were present and the nuclei
were not at the periphery (Fig. 41, PI. XIV.). This would ap-
pear to be a case where a nucleus resulting from fragmentation
divides mitotically afterward : it might, however, be a case where
a spermatocyte undergoing mitosis had fused with a group under-
going fragmentation : if that were the case the size of the spindle
would seem to identify it as the second spermatocytic division.
But whatever its interpretation, this, too, is clearly an exceptional
case. It is probable that a large proportion of the nuclear sub-
stance of the spermatocyte passes into the cytoplasm at the time
of fragmentation. It may be that a kind of reduction is accom-
plished in this manner.

The relative frequency of fragmentation and mitosis seems to
vary in different chains, proglottids and testes. In some chains
the spermatogenetic mitoses were rarely seen except in the older
proglottids, yet spermatozoa were produced in the younger pro-
glottids as abundantly apparently as elsewhere. In some other
cases mitosis is more frequent. From examination of the testes
one gains the impression that the mitoses are not in any case
sufficiently numerous to account for the large number of sperma-
tids and spermatozoa formed. I regard it as at least probable
that spermatozoa are produced from the "spermatid" nuclei
which arise by fragmentation, as well as even those which arise
by mitosis. As will appear in a later section, some cells undergo
degeneration in almost all or all testes and it is of course im-
possible to prove that these particular spermatid nuclei which arise
by fragmentation do not undergo degeneration. Still, developing
spermatozoa have been found on cytophores containing the old
spermatocyte nucleoli.

VI. The Formation of the Spermatozoon from the Spermatid.

In consequence of the difficulty of observation of details in
these exceedingly minute structures, which is farther increased
by the massing together of the spermatids in cytophores and the
condensation of these in later stages, it has not been possible
to reach positive conclusions on all points. It can scarcely be

2OO C. M. CHILD.

doubted from my own observations as well as those of others
that the development of the spermatozoon in these forms differs
in certain respects from the typical method. Although the
greatest caution has been observed throughout, the observations
are given with a certain reserve, because it was impossible to
attain complete certainty on many points and because they are
not in accord with commonly accepted opinions regarding the
development of the spermatozoon. I believe, however, that a
careful investigation of spermatogenesis in the cestodes will prove
of interest.

Except in the early stages when the spermatid nuclei produced
mitotically appear in pairs about the cytophore (Fig. 42, PI. XIV.)
and those arising by fragmentation of the spermatocytes are
massed in the interior of the cytophore (Figs. 34-36, PI. XIII.)
there is no certain criterion for distinguishing the two kinds. The
presence of old nucleoli in a cytophore render it probable that
all or a part of the spermatid nuclei of that cytophore have
arisen by fragmentation, but beyond this no means of identifying
the nuclei of different origin has been discovered.

The following account concerns the spermatids without respect
to origin. If my conclusions are correct, however, many of
these may be the results of fragmentation. As was noted above,
developing spermatozoa not different in appearance from others
have often been found on cytophores containing the old nucleoli.

Figs. 42 and 43 (PI. XIV.) show the newly formed spermatid
nuclei after mitosis. In Fig. 43, from M. pianissimo,, five chro-
matin granules are distinctly visible in each of the nuclei.
These may represent five chromosomes : it is possible that the
spermatocytes of this species contained only five dyads (Fig.
16, A, PI. XI.) and that the cases where a larger number seemed
to be present (Figs. 16, B, 16, C, PI. XI.) were only earlier
stages before the chromatin had become massed in the dyads.

After the formation of the spermatid nuclei their peripheral
position on the cytophore becomes more and more marked until
finally each is borne on a short peduncle or stalk of cytoplasm
(Fig. 44, PI. XIV.). The cytophores differ greatly in size
according as they were formed from a single spermatocyte or a
larger number : in fact in older testes single isolated spermatids

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 2OI

are sometimes found which have apparently become entirely
separated from the cytophore.

The spermatid nuclei contain at this time only a few very dis-
tinct deeply staining granules (some of the nuclei in Fig. 44, PI.
XIV.) : in cases where extraction is carried to extremes only two
granules, one at the peripheral end of the nucleus, the other
near the middle or at one side of the nucleus (some of the
nuclei in Fig. 44, PI. XIV.). The peripheral granule is closely
applied to the nuclear membrane, so closely indeed that it is
often difficult to determine whether it is inside or outside the
nucleus. In some cases, however, it is clearly inside the nucleus
(Fig. 44, PI. XIV.) and this is probably its position in all cases.

The first visible step in the formation of the spermatozoon is
the appearance at the periphery of the cytoplasm peripheral to
the nucleus of a minute deeply staining granule. In position
this granule corresponds to the peripheral centrosome which
enters the middle piece in the spermatozoa of many other forms.
It has been impossible in consequence of the small size of these
cells to obtain any data regarding its origin in this case. If the
spermatids arising by fragmentation do produce spermatozoa the
question as to its origin in those cases is of some interest. This
peripheral body which apparently lies in contact with the border
of the cytoplasm appears to be connected by a very delicate
cytoplasmic strand or fiber with the granule at the peripheral end
of the nucleus (Figs. 44 and 45, PI. XIV.). Whether there is
another cytoplasmic granule in contact with or near the nucleus
corresponding in position to the other centrosome of other forms
could not be determined. From the peripheral granule in the
nucleus the delicate fiber appears to continue through the nucleus
usually to the second granule (some of the nuclei in Fig. 44,
also Fig. 45, PI. XIV.). This continuation of the fiber within
the nucleus has been a matter of the most careful examination
and I can say regarding it only that I have seen it in the nuclei
of practically every cytophore examined and under the most
various conditions of fixation and staining so that if present
methods of technique permit trustworthy conclusions in regard to
such matters its existence seems beyond doubt. The figures ex-
aggerate its distinctness to some extent. It does not stain as

2O2 C. M. CHILD.

deeply as the granules themselves but this is very likely due to its
smaller diameter.

The next step is the formation of the tail which appears first
as a delicate thread extending from the granule at the border of
the cytoplasm (Fig. 46, PL XIV.).

Figs. 47, AAr7, (PL XIV.) show the developing spermato-
zoa after different methods of fixation and staining : Fig. 47, A,
is from M. expansa after sublimate and Delafield's haematoxylin ;
Fig. 47, B, M. expansa after sublimate and iron-haematoxylin ;
Fig. 47, C, M. planissima, after chrom-oxalic and iron-haema-
toxylin ; Fig. 47, D, M. expansa, after Hermann and iron-hasma-
toxylin. Fig. 47, E, is from M. planissima after sublimate and
iron-haematoxylin, but with extraction stopped at an earlier
stage. One interesting point in this figure as compared with the
others is the much larger size of the peripheral cytoplasmic gran-
ule and the fiber connecting it with the nucleus --an excellent
illustration of the uncertainty attending the use of iron-
haematoxylin.

The tail of the spermatozoon grows to a very great length.
Fresh spermatozoa obtained by teasing living proglottids in indif-
ferent fluids are 0.3-0.4 mm. in length. Most or all the tails aris-
ing from one cytophore usually lie parallel in the testis, and
since their length is much greater than the diameter of the testis
they become coiled in the spaces between the cells or along the
wall of the testis. The tail is very delicate and without visible
differentiation in structure.

As regards the formation of the head of the spermatozoon
Moniezia does not seem to agree with other species described. At
least I know of no other case in which the sperm-head, if it can
be called a head, is formed in the manner described below.

In my study of spermatogenesis I was for a long time puzzled
by the fact that all of the sperm nuclei appeared to degenerate
after the tails were formed. Masses like Fig. 49 (PL XV.)
consisting of degenerating nuclei and condensed cytophore cyto-
plasm can be found in every older testis. At first I concluded
that these were probably the spermatids formed by fragmenta-
tion which began the development of spermatozoa but were un-
able to complete it. But another feature made the matter still

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 2O3

more puzzling. The most careful examination, under varied con-
ditions of fixation and staining, of spermatozoa in the male ducts
and in the seminal receptacle of the female ducts, which becomes
greatly distended with them at a certain stage, failed absolutely
to reveal the existence of a head differing in appearance from the
tail. The examination and staining of fresh spermatozoa from
the seminal receptacle and ducts of living proglottids led to the
same result. The spermatozoa appeared as very long thread-
like structures perhaps slightly larger at one end than at the other
but without the least trace of a physically or chemically differen-
tiated head.

Then the question arose as to whether the eggs were actually
fertilized by these spermatozoa. As will be described in the fol-
lowing paper, the spermatozoa were found entering the eggs as
these passed the opening from the seminal receptacle on their
way to the uterus, and nuclei which could be nothing else than
male pronuclei unless these eggs differ from other known cases
in their maturation and fertilization stages were found. Return-
ing to the developing spermatozoa the most careful study was
made of the various stages and especially of the masses like Fig.
49 (PI. XV.) which were apparently undergoing degeneration. It
is very difficult to distinguish details in these masses for they
stain more deeply as they condense and the nuclei especially
become more or less filled with deeply staining granules and
masses. In the course of time certain apparently favorable cases
were found some of which are shown in Figs. 48, A-^S, C (PI.
XV.). These seem to indicate that the "head" of the sperma-
tozoon, /. e., the part arising from the nucleus is formed from the
two nuclear granules, the peripheral and the other which may be
central or proximal, together with the connecting strand, and
furthermore, that when degeneration of the other parts of the
nucleus begins the spermatozoon is set free. Figs. 48, ^-48, C,
show examples of the early stages of nuclear degeneration in
sperm cytophores. In Fig. 48, C, the spermatozoon head is
apparently in the act of escaping from the degenerating nucleus.
The peripheral portion of the nuclear membrane has disappeared
but the peripheral nuclear granule is still recognizable. Figs.
48, A, and 48, B, are apparently somewhat earlier stages in

2O4 C. M. CHILD.

which the nuclear membrane is still intact. I am forced there-
fore to the conclusion that only a part of the nucleus is concerned
in the formation of the sperm-head, the remainder undergoing
degeneration. Fig. 49 (PI. XV.) represents a cytophore after con-
densation. For the sake of clearness only a few of the nuclei
which cover the surface of the mass are represented as they
appear, the others being indicated by dotted lines. In several
cases what seems to be the sperm-head is visible in the degen-
erating nucleus.

Fig. 50 (PI. XV.) represents a case in which the spermatozoa
are apparently just separating from the cytoplasm of the cyto-
phore which contains the deeply staining remains of the nuclei.
In this case I convinced myself that these were the anterior
ends of the spermatozoa, by following the tails throughout their
whole length in the testis. The diameter of the ends shown in
the figure was distinctly though only slightly greater than that
of the other ends, but the change in diameter is very gradual.

In the free spermatozoa no differentation in staining of the
head-region is visible. The whole spermatozoon stains uniformly
and less intensely than the nuclear granules or masses of earlier
stages. In a few cases I believed I had distinguished slight traces
of the two nuclear granules in the fully developed spermatozoon
but these observations were so doubtful that no figures are given.

In the fresh spermatozoa obtained by teasing in indifferent
fluids no visible head and no movement was ever observed. An
examination of the bibliography of the subject afforded scanty
results. So far as I have been able to determine no full account
of the spermatogenesis of the cestodes exists. Among the older
papers several give brief descriptions of the formation of the
spermatozoa but these are either very incomplete or incorrect in
consequence of the technique employed and need not be reviewed
in detail. In one point, however, the early observations agree
fairly well : the head of the spermatozoon is described and fig-
ured as exceedingly minute or is said to be absent. Sommer
and Landois l in describing the testes of Bot/iriocepkahis latus
mention spermatozoa bearing at one end " ein kleines, stark
lichtbrechendes Kopfchen."

1 Sommer and Landois, " Ueber den Bau der geschlechtsreifen Glieder von Both-
riocephalus latus Bremser," Zeitschr.f. wiss. Zoo!., Bd. XXII., 1872.

THE RELATION BETWEEN AM1TOSIS AND MITOSIS. 2O5

Two years later Salensky l states that in Ainf/iiliua the nuclei
disappear completely in the formation of the spermatozoa and that
" die Kerne bei der Bildung der Spermatozoen keine Rolle
spielen." Regarding the fully formed spermatozoa he says :
" Die Faden sind sehr lang, ungefahr 0.27 mm. und an einen Ende
etwas gekrummt. Diese Kriimmung soil aber nicht als Kopfchen
angesehen werden, indem die spermatozoen in ihrer ganzen Lange
gleich dick sind."

As regards Tcema uiediocancllata and Ta*nia soliuin Sommer 2
speaks of the bundles of spermatozoa which hang from certain
large cells (in reality the cytophores) and "mit ihren ausserst
feinen, glanzenden Kopfchen noch in Zellenprotoplasma stecken."
These " Kopfchen " are probably the nuclear granules which
present this appearance in unstained or slightly stained prepara-
tions, the nuclear membrane not being clearly visible. " Zwischen
diesen Samenfaden producirenden Zellen findet man gleichzeitig
im Hodenkorperchen kleine Anhaufungen freier, heller, scharf
contourirter und blaschenformiger Kerne. Einzelne derselben
haben an ihrem Grenzrande noch Spuren von Protoplasma in
welchen mit seinem glanzenden punctformigen Kopfchen ein
Samenfadchen haftet." These masses are perhaps degenerating
cytophores. In another paragraph he describes the formation of
the large multinucleate cells which give rise to the spermatozoa
and says : " An der Peripherie dieser grossen Zellen geht von
irgend einer Stelle die Bildung der Samenfaden aus. Letztere
entstehen lediglich aus dem Protoplasma der Zelle ; eine Betheili-
gung der Kerne dabei findet nicht statt. In demselben Maasse
wie mit der Bildung der Samenfaden das Protoplasma der Zelle
schwindet, werden die eingelagerten Kerne frei, erscheinen dann
scharfer berandet wie fruher, etwas aufgeblaht oder gequollen,
homogen und wasserhell, dann fallen sie zusammen, collabiren,
wie wenn sie einen flussigen Inhalt entleert hatten und gehen
zu grunde, oder werden, wenn sich inzwischen Samengaiige ge-

1 Salensky, " Ueber den Bau und die Entwickelungsgeschichte der A mph ilina,
G. Wagen (Monostomum foliacium Reed)," Zeitschr. f. wiss. Zool., Bd. XXIV.,
1874.

2 Sommer, "Ueber den Bau und die Entwickelung der Geschlechtsorgane von
TcEiiia mediocanellata (Kiichenmeister) und Toenia solium (Linne)," Zeitschr. f.
wiss. Zool., Bd. XXIV., 1874.

2O6 C. M. CHILD.

bildet haben mit den Samenfaden fortgespielt." Sommer's prep-
arations were obtained by maceration and teasing and without
staining. As described, the fate of the nuclei does not differ
very widely from that described in the present paper except for
the fact that Sommer failed to observe that any portion of the
nucleus took part in the formation of the spermatozoon. Con-
sidering the methods employed this failure is not strange.

Moniez l describes the formation of large multinucleate cells
and the protrusion from their surface of the nuclei which are united
with the body of the cell by pedicels (these are evidently the
spermatids on the cytophore). He continues as follows : " Ces
nouvelles formations qui rayonnent de la cellule-mere sont les vrais
spermatozoides : leur flagellum se forme a la partie peripherique,
tandis qu'ils sont encore fixe par 1'autre extremite ; c'est apres
qu'ils se sont detaches que leur tete s'atrophie comme Ton sait."
These facts he describes as common to a number of species
among them Tcenia expansa, i. e., Moniezia expansa as it is now
known.

In describing the spermatozoa of Tcenia saginata Leuckart 2
speaks of the " freilich kaum ausgezeichneten " head.

From all of these observations it is evident that where a dis-
tinct head is visible it is exceedingly minute and the observations
of Salensky, Sommer, and Moniez seem to indicate that the sper-
matozoa of several species are without visible heads. The
description of the collapse and degeneration of the nuclei bySom-
merand the mention of atrophy of the head by Moniez appear to
be somewhat closely in line with my own observations. But until
other species have been examined with the aid of present cytolog-
ical methods general conclusions are impossible. I am convinced,
however, that if the spermatozoa of Moniezia possessed distinct,
visibly differentiated heads I should have seen them in some cases
at least. Comparative study of other species will undoubtedly
prove of interest.

VII. The Degeneration of Cell-Groups in the Testis.

During almost the whole period of existence of the testis groups
of cells undergo degeneration from time to time. Cells in any

1 Moniez, " Sur les Spermatozoides des Cestodes," Comptes Rendus, 1878.

2 Leuckart, " Die Parasiten des Menschen," 1879-1886.

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 2O?

stage of development from the spermatogonia to the developing
spermatozoa except during the spireme stage are subject to this
fate. The proportion of degenerating cell-groups varies greatly
in different chains, proglottids and testes. In some chains only
one or two cases of degeneration preceding the first appearance
of the spireme stage have been observed. In others degenerat-
ing groups are found in almost every testis in the spireme period.
During the earlier stages of the process the degenerating cells
form rounded masses : later these break up and become distrib-
uted through the testis and are apparently absorbed by the cyto-
plasm of other cells. While it is impossible to assign positively
a definite reason for this degeneration I am inclined to believe
that it results from differences in physiological condition which
may in turn be correlated with differences in nutrition. Careful
examination of regions of rapid growth in many forms often shows
a certain proportion of cells which are undergoing degeneration.
Undoubtedly in such regions the intensity of certain stimuli or
conditions carries some cells beyond the point where physiolog-
ical equilibrium can be regained and they degenerate, serving per-
haps as food for the others. There can be little doubt that the
testis is a region of this sort. The great variation in the fre-
quency of degeneration in different chains may indicate that it is
connected with nutritive conditions. Apparently more cells are
produced than can be sustained and some are eliminated.

The fact that no case of degeneration beginning during the
spireme stage has been observed may be of some interest. It is
not improbable that this stage is relatively independent of external
conditions, i. e., that a cell having entered this stage is capable of
completing it without the intervention of external factors. To
judge from appearances this stage is a readjustment or the estab-
lishment of a new condition of equilibrium in the cell and it may
represent a reaction from previously existing conditions which
have disturbed the previous equilibrium of the cell. There can
be little doubt that in many respects the life of the cell possesses
a cyclical character. One complex of processes or reactions con-
tinues until it brings about a reversal in reaction or initiates a
different complex, etc.

That this degeneration has any connection with amitosis is

2O8 C. M. CHILD.

extremely improbable. In no case has a whole testis been found
undergoing degeneration, yet in all the testes most of the divi-
sions before spermatogenesis proper were amitotic and in the
great majority the first divisions certainly were amitotic. As was
suggested in the preceding section degeneration of cell-groups in
post-spireme stages may be connected with the fragmentation of
spermatocyte nuclei though this seems improbable, and more-
over, it does not explain degeneration in pre-spireme stages. I
believe, though I see no way of demonstrating it, that the method
of origin of the cell-groups in the testis has no connection with
their degeneration.

The degenerating cell-groups vary greatly in appearance
according to the stage at which degeneration begins and the dif-
ferent stages of degeneration itself. In many cases, though not
always, it is possible to determine from the appearance of the
degenerating mass approximately the stage at which degenera-
tion began. In some cases cells in the same stage of develop-
ment undergo two different processes of degeneration.

Some of the characteristic forms and stages of degenerating
cell groups are shown in the following figures : in these figures
no attempt has been made to represent the cytoplasmic back-
ground. This varies somewhat in density and staining in differ-
ent cases. Vacuoles and spaces are indicated by broken lines.
The method of reproduction exaggerates the depth of shade in
the more deeply staining portions. Fig. 5 I (PI. XV.) shows a
small group of cells from a young testis in the first stages of
degeneration. The first evidence of degeneration in these cases
is a condensation of the cytoplasm and a massing together of the
nuclei, and the degenerating group becomes quite distinct from
other cells, usually lying in a space. Fig. 52 (PI. XV.) shows a
later stage of this form of degeneration ; the nucleoli increase in
size and stain very deeply, the nuclear membrane becomes indis-
tinct, and the whole mass stains more intensely. Later, as shown
in Fig. 53 (PI. XV.), the mass breaks up into irregular deeply
staining fragments and strands which are distributed through the
testis and are often found in the cytoplasm of other cells sur-
rounded by small vacuoles ; a few of these fragments in the cyto-
plasm are shown in the figure.

THE RELATION BETWEEN AMITOSIS AND MITOSIS.

Fig. 54 (PI. XV.) represents a form of degeneration in cells
in prespireme stages which sometimes occurs in old testes.
Here the nuclei form irregular densely staining masses and finally
the whole breaks up and is absorbed.

Fig. 55 (PI. XV.) shows degeneration of a group of spermatid
nuclei in a cytophore. The deeply staining granules and masses
in the nucleus increase in number, the nuclear membrane breaks
down, and the granules are distributed through the cytoplasm.
In Fig. 56 (PI. XV.) another form of spermatid degeneration is
seen and a later stage in Fig. 57 (PI. XV.). Fig. 58 (PI. XV.)
represents a still later stage : vacuoles usually containing a single
granule still indicate the position of the nuclei ; the mass stains
only very faintly at this stage and seems to decrease gradually
in size until finally it becomes imbedded in a cytophore and
gradually disappears.

Fig- 59 (PI- XVI.) represents a form of degeneration of the
spermatids which usually occurs only after the spermatozoa have
begun to develop. About the periphery of each nucleus a large
amount of deeply staining substance develops and appears to
flow toward the center of the cytophore. In Fig. 60 (PI. XVI.)
is shown a later stage of this form of degeneration. Here the
cytoplasm stains rather more deeply than that of the normal
cytophore, the deeply staining substance has disappeared entirely
from the peripheral regions except in a few radiating strands and
the positions of the nuclei are indicated only by vacuoles. In
still later stages (Fig. 61, PI. XVI.) the deeply staining substance
gradually breaks up into granules (Fig. 64, k, PI. XVI.), loses its
staining power and finally disappears, and the whole cytophore
becomes highly vacuolated, breaks up into irregular masses and
shreds (Fig. 64 /, PI. XVI.) and is apparently absorbed.

Other modifications of the process of degeneration are occa-
sionally seen but these are the principal ones. The apparent
variation which these processes exhibit is of some interest as
indicating that differences in the processes of degeneration like
differences in development are undoubtedly determined by differ-
ences in the condition of the cells or of the environment. At
present, however, even a surmise as to the nature of these dif-
ferences is of little value.

2IO C. M. CHILD.

VIII. The Full-Gr&wn Testis.

The testis continues to increase in size for a considerable time
after spermatogenesis begins. Only a part of the spermatogonia
enter the spireme stage at any one time, the others continuing to
divide amitotically. After the appearance of the spermatogenetic
divisions in a testis, I have never seen a case of mitosis in the
spermatogonia, but amitoses are frequent. Figs. 62, A, 62, B
(PI. XVI.) represent groups of spermatogonia in full grown testes.
In the same testes all stages of spermatogenesis and fully devel-
oped spermatozoa may be found. At this period the spermato-
gonia are usually found in small groups near the periphery.
Figs. 63, A 63, D(P\. XVI.), show cases of amitosis in spermato-
gonia from full grown testes, including nearly all the modifica-
tions of the process observed.

In Fig. 64 (PL XVI.), one half of a full-grown testis is shown
on a scale half as large as that of the other figures. The different
stages shown are as follows : at a is a group of spermatogonia
still in the prespireme stage and showing one amitosis ; b shows
the earliest stages of the spireme, c, c, two groups with fully-
developed spireme, while at d some cells are preparing for the
first spermatocytic mitosis ; at c is a cell in which the dyads are
formed, part of a group which appears in adjoining sections ;
f,f, are cytophores with spermatid nuclei and developing sper-
matozoa ; g, is a cytophore in which degeneration of the nuclei
has begun, but the spermatozoa are still attached ; at Ji is seen
part of a bundle of free spermatozoa which can be followed in
other sections ; k, k, represent two degenerating cytophores in
which the nuclei have already vanished : the shreds of cytoplasm
and the debris from earlier cytophores are indicated at /. Al-
though this one section does not show all the stages in the his-
tory of the cells, it serves to indicate the promiscuous distribu-
tion of different stages.

IX. Conclusion.

The point of chief importance in the present paper is the fact
that typical mitosis and amitosis may appear together and ap-
parently under identical conditions in the development of the male
as well as of the female germ cells. The relative frequency of

THE RELATION BETWEEN AMITOSIS AND MITOSIS. 211

the two forms of division varies in different chains, proglottids
and regions. Observations and experiments to be described later
will show very clearly, however, that amitosis as well as mitosis
is an important factor in growth, not only in Moniezia but in
many other forms and that in some cases at least either form of
division may be changed into the other by altering the conditions.

These facts are of considerable importance as bearing upon
certain hypotheses regarding the significance of the chromosomes.
At present it seems improbable that the views held by certain
authors regarding the individuality of the chromosomes can be
reconciled with them. Extended discussion is, however, post-
poned until other facts have been presented.

The most important features in the development of the male
germ cells in Moniezia are as follows :

The testes apparently arise from cells which are already differ-
entiated as muscle-cells, as well as from other cells of the
parenchyma. The earlier divisions are almost entirely amitotic,
mitosis being rarely seen.

The growth of the testis up to the time when spermatogenesis
proper begins is almost wholly by amitotic division. In the full-
grown testis the remaining spermatogonia still continue to divide
amitotically. After the spireme stage the spermatocytes follow
two very different lines of development. In some of them typi-
cal dyads are formed and the two usual spermatogenetic mitoses
follow : the spermatid nuclei are usually situated about the peri-
phery of large masses of cytoplasm, cytophores formed by fusion
of the spermatocytic cytoplasm, but may be isolated.

In the other spermatocytes the nucleus increases in size, the
spireme breaks up into granules and masses and loses most of its
staining power, the old nuclear membrane disappears, and new
nuclear membranes form about small fragments of the chromatin :
each spermatocyte may give rise to several small nuclei : in ap-
pearence these nuclei are indistinguishable from the spermatid
nuclei produced mitotically. When first formed they are massed
in groups in the interior of the cytophore about spaces which
indicate the former positions of the spermatocyte nuclei. The
nucleolus does not take part in this fragmentation but remains in
the cytoplasm of the cytophore for some time. The nuclei thus

212 C. M. CHILD.

formed gradually make their way to the periphery of the cyto-
phore and probably give rise to spermatozoa, though this cannot
be demonstrated with absolute certainty.

Apparently only a part of the nucleus is involved in the for-
mation of -the anterior end of the spermatozoon in which no
" head " is visible. The sperm- head is apparently represented by
two granules in the nucleus, one peripheral, one more or less
nearly central and a less deeply staining fiber which connects
them, these being in most cases the only deeply staining portions
of the nucleus. When development of the spermatozoon is
completed the nuclear portion is apparently set free from the
remainder of the nucleus by the degeneration of the latter.

Groups of cells in all stages of development except the spireme
stage are frequently attacked by degenerative processes probably
because of insufficient nutrition or exhaustion.

HULL ZOOLOGICAL LABORATORY,
UNIVERSITY OF CHICAGO,
September, 1906.

214 C - M - CHILD.

EXPLANATION OF PLATES.

PLATE XI.

FIG. 14. A, B, the dyads grouped about the periphery of the nucleus. M. ex-
pansa.

FIG. 15. Dyads irregularly disposed in nucleus. M. expansa.

FIG. 1 6. A-C, dyads. M.planissima.

FIG. 17. Two dyads more highly magnified.

FIGS. 18, 19, 20, 21. Different stages of first spermatocy tic mitosis.

FIG. 22. Two dyads in metaphase, more highly magnified.

FIG. 23. After the first spermatocytic division.

FIG. 24. Probably resting nuclei after first spermatocytic mitosis.

BIOLOGICAL BULLETIN, VOL. XII

2l6 C. M. CHILD.

PLATE XII.

FIGS. 25, 26. The second spermatocytic mitosis.

FIG. 27. A-D, early stages of fragmentation of nuclei of first spermatocytes.

FIG. 28. Fragmentation. One small nucleus forming.

FIG. 29. A-E, the formation of nuclei by fragmentation.

BIOLOGICAL BULLETIN, VOL. XII

PLATE XII

B

2l8 C. M. CHILD.

PLATE XIII.

FIG. 30. A-C, the formation of nuclei by fragmentation.
FIGS. 31-36. Nuclei and old nucleoli in the cytophores after fragmentation.
FIG. 37. Nuclei and old nucleolus after fragmentation with spireme stages ad-
joining.

BIOLOGICAL BULLETIN, VOL. Xh

J7

22O C. M. CHILD.

PLATE XIV.

FIG. 38. Nuclei and old nucleolus after fragmentation with other stages ad-
joining.

FIG. 39. Probably migration of nuclei to periphery after fragmentation.

FIG. 40. Possible case of secondary fragmentation.

FIG. 41. A case of mitosis in a cytophore which was probably formed by frag-
mentation.

FIG. 42. Spermatid nuclei formed by mitosis.

FIG. 43. Spermatid nuclei.

FIGS. 44 and 45. Early stages in the development of the spermatozoa.

FIG. 46. Development of the spermatozoa.

FIG. 47. A-E, development of the spermatozoa after different methods of fixation
and staining ; A, M. expansa, sublimate and Delafield's haematoxylin ; , M. ex-
pansa, sublimate and iron-hsematoxylin ; C, M. pianissimo, chrom-oxalic and iron-
haematoxylin ; D, M. expansa, Hermann and iron-hsematoxylin ; , M. planissima,
sublimate and iron-haematoxylin, extraction stopped at an early stage.

BIOLOGICAL BULLETIN, YOL. XM

PLATE XIV

G

OG

46 \

47 C

222 C. M. CHILD.

PLATE XV.

FIG. 48, A-C. The spermatozoon and the degenerating portions of the spermatid
nucleus.

FIG. 49. A cytophore with spermatozoa and degenerating spermatid nuclei.
FIG. 50. Spermatozoa becoming free from degenerating cytophore.
FIGS. 51-54. Degeneration of cell groups in prespireme stages.
FIGS. 55-58. Degeneration of cell groups in spermatid stage.

BIOLOGICAL BULLETIN, VOL. XII

53

224 c - M - CHILD.

PLATE XVI.

FIGS. 59, 60. One form of degeneration of cytophores after development of
spermatozoa.

FIG. 61. Late stage of degenerating cytophore.

FIGS. 62, A, B ; 63, A-D. Spermatogonia dividing amitotically ; from full grown
testes.

FIG. 64. Half of full grown testis ; scale one half that of other figures ; a, sperma-
togonia ; b, early stages of spireme formation ; f, c, spireme stage ; d, preparation for
first spermatocytic mitosis ; e, dyads before first spermatocytic mitosis ; f, f, cytophores
with spermatids ; g, cytophore with degenerating nuclei and spermatozoa; h, free
spermatozoa ; k, k, degenerating cytophores; /, shreds of protoplasm from earlier cyto-
phores which have undergone degeneration.

The following figures are from Moniezia expansa : I B, I D, 3 A, 3 B, 3 C, 4 B,
4 C, 9 D, 10 A, ii A, ii B, 12 A, 13 A, 13 B, 14 A, 14 B, 15, 17, 18, 20, 21, 23,
26, 27 D, 28, 29 E, 30 A, 30 C, 31, 32, 34, 35, 36, 37, 38, 39, 40, 42, 44, 45, 46
47 A, 47 B, 47 D, 48 A, 48 B, 48 C, 49, 51, 55, 57, 58, 59, 60, 61, 62 ^, 62 ^,
63 A, 63 5, 63 C, 63 A 64.

The others are from M. pianissimo.

BIOLOGICAL BULLETIN, VOL. XII

THE CIRCULATORY AND NERVOUS SYSTEMS OF
THE GIANT SCALLOP (PECTEN TENUICOSTATUS,
MIGHELS), WITH REMARKS ON THE POSSIBLE
ANCESTRY OF THE LAMELLIBRANCHIATA, AND
ON A METHOD FOR MAKING SERIES OF ANA-
TOMICAL DRAWINGS.

OILMAN A. DREW.

In considering any system of organs it is essential that we
should bear in mind the modifications of the possessor of the
organs, that adapt it to its particular life.

Pecten is one of the ablest swimmers among lamellibranchs.
The whole structure of the animal is modified for this purpose.
The valves have become rounded in outline, flattened, and com-
paratively light. The anterior adductor muscle has been lost,
and the posterior adductor muscle, which is very powerful, is
situated near the middle of the body. The cartilage has become
well developed, so the shell may be opened quickly when the
muscle relaxes, and the hinge line is straight, so there may be no
unnecessary strains in opening and in closing the shell. Each
gill is attached by one lamella only, so water in the temporary
cloacal chamber may be thrown out without injuring the gills,
and the gills and margins of the mantle are provided with
muscles to withdraw them from the margins of the shell when
the shell is closed. Furthermore the margins of the mantle are
provided with infolded ridges and with circular muscles so it is
possible to direct the current of water which issues from the shell
in the required direction.

To fit the animal to a life of such activity, and to enable it to
live in the comparatively exposed positions that it inhabits, an
abundance of sense organs, tactile and probably visual, have
been developed. These are placed in the most exposed posi-
tions, where they may give warning to their possessor, and are

1 Free use has been made of both descriptions and figures published as No. 6 of
the University of Maine Studies, under the title of "The Habits, Anatomy, and
Embryology of the Giant Scallop (Pecten tenuicostatus, Mighels)."

225

226 G1LMAN A. DREW.

accordingly borne along the margins of the lobes of the mantle.
It is not entirely certain what relationship Pecten bears to the
usual form of lamellibranch as regards positions of parts. In
lamellibranchs that are supplied with two practically equal
adductor muscles, a line connecting the two adductors runs
nearly lengthwise of the animal. In such a case the hinge line
is more or less dorsal, one end is anterior, and the other pos-
terior. When one of the muscles disappears, as is the case with
Pcctcn, one of the landmarks disappears and it becomes more
difficult to locate the direction of parts. Inasmuch as the hinge
line is usually dorsal, it is very natural to look at the hinge line
of this form as dorsal, and for matters of description it is con-
venient to so consider it. If, however, the position that the
anterior adductor would have occupied, had it been retained, be
considered, the position of the mouth, foot and heart indicate
that it would have to be placed much nearer the hinge line than
the present position of the posterior adductor muscle, the muscle
that is retained. If this is the case, it becomes evident that the
loss of the anterior adductor muscle has been accompanied by a
general reduction of the anterior part of the body, so a large part
of the body of Pecten is to be considered morphologically pos-
terior. This supposition seems to be borne out by the nervous
system, and the vascular system of the mantle, as well as by the
extent and position of organs. In most forms the margin of
each lobe of the mantle is supplied with a posterior and an ante-
rior pallial nerve of approximately equal size. These nerves
supply the muscles and sense organs of the margins, and, in
many forms at least, unite with each other so they form a con-
tinuous connective between the cerebral and the visceral ganglia.
In Pcctcn, not only is this the case, but the nerve in the margin
of the mantle is joined at intervals for nearly its whole length
by nerves from the visceral ganglion (Fig. 6). On the other
hand, it is joined only in the region of the anterior ear by
nerves from the cerebral ganglion. The visceral ganglia are the
important ganglia of the animal, and both the cerebral and pedal
ganglia are greatly reduced.

The blood is supplied to the mantle very largely by the pos-
terior pallial arteries (Fig. 5). The anterior pallia! arteries are

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 22/

comparatively small, and while they are connected with the pos-
terior pallial arteries, the size and character of the vessels indi-
cates that the junction is probably very near the anterior ear.

Considering everything, it seems likely that the longitudinal
axis of the body could be morphologically represented by a line
drawn from near the hinge extremity of the anterior ear to the
middle of the adductor muscle, and that a very small portion of
the scallop is anterior.

CIRCULATORY SYSTEM.

The animal is large enough to allow one to successfully inject
the chief vessels with starch or gelatin injecting mass, and then
by dissection and microscopic preparations to trace the dis-
tribution of the vessels of the different organs and to determine
quite definitely the course taken by the blood in its circulation.

The heart is a typical, symmetrical lamellibranch heart, with
two auricles and one ventricle (Figs. I and 3), the latter per-
forated by the intestine which enters it near one end and leaves
it near the other end (Fig. 2). Dorsally the ventricle is prolonged
somewhat, posterior to the intestine, where the morphologically
anterior aorta is given off, and ventrally to a less extent it is
prolonged anterior to the intestine, where the much smaller
morphologically posterior aorta is given off. The walls of the
ventricle are of about even thickness throughout their extent,
and are quite smooth outside and inside. The auricles join the
ventricle on each side near its middle, are somewhat triangular
in shape, with the most acute angle receiving blood from the
gills and mantle at a point dorsal to the adductor muscle and
directly ventral to, but some distance from, the cartilage. The
opening of each auricle into the ventricle is near the middle of
the side of the auricle that lies next to the ventricle and farthest
away from the opening where the auricle receives its blood.
The muscles around the openings of the auricles into the ven-
tricle, and to a less extent around the openings through which
the auricles receive blood, are well developed and must act as
sphincters that tend to keep the blood from being regurgitated
The walls of the auricles, unlike those of the ventricles, are
roughened by pits that open into the cavities of the auricles.

228 OILMAN A. DREW.

Both auricles and ventricle are composed of interlacing muscle
fibers, and are capable of great extension. In preserved speci-
mens, the heart is usually contracted and is not very conspic-
uous. In such contracted hearts the cavities of auricles and
ventricle are practically obliterated.

The heart lies in a somewhat triangular, spacious pericardial
cavity that is dorsal to the posterior half of the adductor muscle,
and ventral to the posterior portion of the liver. Posteriorly it
is covered only by a somewhat thick, muscular membrane which
separates it from the mantle chamber.

As already mentioned, two blood vessels leave the ventricle
(Figs, i and 3), one from each end. Although they are not so
placed in reference to the ways the terms are generally used in
describing Pccten, the two ends correspond to the anterior and
posterior ends of the ventricle in most forms of lamellibranchs.
The posterior aorta is much the smaller of the two, leaves the
heart ventral to the intestine (actually anterior to it) and divides
immediately after leaving the heart, into two vessels, one of
which, the smaller, follows along the intestine, supplying it and
surrounding portions with blood. The other vessel turns almost
at right angles upon leaving the aorta and enters the adductor
muscle, where it divides into a system of vessels that supply the
muscle with blood.

The anterior aorta is much larger than the posterior aorta,
and supplies all of the remainder of the body. It leaves the
ventricle dorsal to (actually posterior to) the intestine and very
soon gives rise to a vessel which passes into and supplies the
wall that separates the pericardial cavity from the mantle cham-
ber. From the pericardium the anterior aorta follows along the
postero-dorsal border of the liver to the base of the ear. Here
it gives rise to the branch (Fig 3, ppa) which passes posteriorly
to the extreme upper margin of the mantle that lines the ear,
giving off along its course a number of branches, which supply
this portion of the mantle. Here it divides into two vessels, a
right and a left, each of which bends abruptly ventrally (Fig.
5, ppa) and follows along the margin of the respective mantle
lobe about opposite the line of attachment of the infolded ridge
of the mantle, alongside but external to the pallial nerve.

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 22Q

Very fine branches are given off from these vessels all along
their courses, which further divide to form systems of capillary
spaces that are finest and most numerous near the margins.
Some of these capillary spaces are large enough to be injected
with starch mass, and I have a preparation of the mantle lobe
from which only the infolded ridge has been removed, that was
dehydrated, cleared and mounted in balsam, in which the whole
system of vessels can be traced. A gelatin mass not only fills
the spaces mentioned, but passes out between the cells so that
in sections it may be seen to be diffused throughout the tissue.
This seems to hold good for all other parts of the body with the
exception of the gills, in which organs the mass is more com
pletely, but not entirely, confined to the blood spaces. The in-
dication therefore is, that the blood spaces are not confined ves-
sels, and that the blood functions as both blood and lymph.
The posterior pallial vessel may be traced far anteriorly, grad-
ually diminishing in size along its course. Here it finally joins
the anterior pallial vessel. The anterior pallial artery (Fig. 3,
apa) leaves the anterior aorta very near the cartilage and runs
directly to the anterior border of the hinge region of the mantle,
giving off vessels to this portion of the mantle on the way.
Here it branches into right and left vessels, each of which bends
abruptly ventrally (Fig. 5, apa] and pursues a course along the
anterior border of the mantle similar to that taken by the pos-
terior pallial artery at the other extremity of the animal.

Along the anterior border of the mantle, near the dorsal line,
the vessel is rather small and slightly broken in its course. It
may be possible that this represents the border line between the
posterior and the anterior pallial arteries. There are other reasons
for believing that a large share of the animal is morphologically
equivalent to the posterior portions of other forms, and that the
anterior portion is greatly reduced. This has received attention
in another place.

Several vessels leave the anterior aorta to supply the liver and
stomach. Most prominent among these is a vessel which leaves
the aorta between the points of origin of the anterior and posterior
pallial arteries. This bends out toward the left side of the liver,
where, in injected specimens, it is very conspicuous, passes ven-

230 OILMAN A. DREW.

trally and sends branches to the major part of the liver and to
the stomach.

A short distance in front of the cartilage the anterior aorta
bends ventrally, passes through the liver and gives off a few small
branches to it, sends a vessel to the palps in passing, and passes
on to supply the foot and the visceral mass. The vessel that
supplies the foot (Fig. 3, fa) leaves the aorta a short distance
ventral to the mouth, passes along the body wall until the foot is
reached, and extends into the foot along its dorsal border. Just
before entering the foot this, the pedal artery, gives rise to a
small vessel that passes posteriorly along the single retractor
muscle of the foot, supplying it with blood. From the point of
origin of the pedal artery the aorta extends into the visceral mass,
following along the enlarged portion of the intestine that leads
away from the stomach, and supplying this and other portions of
the intestine and the reproductive organs with small and with
large branches. The enlarged portion of the intestine that comes
from the stomach is especially well supplied (compare Figs. 2 and
3), there being numerous small branches that are given off
directly from the aorta, and large branches that follow along on
the different sides of this portion of the intestine and likewise
supply it with branches. A short distance ventral to the foot a
large branch leaves the aorta and passes postero-ventrally to
divide again and form small branches that supply the remaining
loops of the intestine and the postero-ventral portions of the
reproductive organs.

This completes what might be called the systemic arterial sys-
tem. Beginning with the heart the system ends in the capillary
spaces of the various organs. This system is most easily injected
through the vessel in the suspensory membrane of the gills that
is farthest from the adductor muscle (Fig. i, bv], with a hypo-
dermic syringe, injecting toward the heart. If a starch mass that
will not pass through the capillary spaces is used, all of the ves-
sels thus far described will be injected, as will also the veins that
return blood from the gills, as this vessel is the one that returns
blood from the gills to the heart. If a gelatin mass is used all
of the systems may be injected, but as the injecting mass may
pass out of the spaces, between the cells of the various organs,
such injection does not aid in tracing the course of blood flow.

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 23!

The systemic veins (Fig. 4) that collect the blood that is sup-
plied by the systemic arteries, from the various organs of the body,
may be injected from several different vessels. They may be in-
jected by pushing the needle beneath the membrane that covers
the posterior surface of the adductor muscle. A large blood
space occupies this position, into which the needle is inserted and
the mass injected fills the systemic veins. Another point from
which these veins may be injected is from one of the superficial
vessels of the visceral mass. These vessels are very conspicuous,
and may be very easily picked up with the needle. Still another
vessel is the vein that returns blood from the liver, which may
be seen on the left side of the animal anterior to, but near the
large artery that supplies the liver. Injecting any one of these
vessels will to a greater or less extent inject the others, but there
does not seem to be an entirely free communication between
them. They all carry blood to the kidneys, and seem to empty
into a common sinus on either side, that lies alongside the kidneys
in the walls of the visceral mass. The sinuses of the two sides
are connected beneath the adductor muscle, but it frequently hap-
pens that a complete injection of the system is not obtained from
an injection from any one of the veins mentioned. Just where
the obstruction lies in such cases has not been determined. It
has been noticed that obstructions are more likely to be encoun-
tered in injecting from the veins of the visceral mass than in in-
jecting from any of the others.

Inasmuch as blood spaces are cut in removing the muscle from
the shell, it has been found desirable in injecting this system of
vessels to wedge the valves open and to inject from the posterior
surface of the adductor muscle. In injecting after the animal is
removed, a considerable quantity of the injecting mass is sure to
escape at the ends of the muscle.

The position of the veins may be seen in Fig. 4. A large vein
comes from the liver, another from the foot, and the veins in the
muscle unite to form a more or less definite sinus along the dor-
sal border of the muscle, and two smaller ones on the antero-ven-
tral side of the muscle. These sinuses unite near the anterior ends
of the kidneys. A series of vessels from the visceral mass unite
along the borders of the kidneys and finally connect with these

232 OILMAN A. DREW.

sinuses. Most of the blood from all of these organs is distributed
to the kidneys through systems of capillary spaces. The branch-
ing of these vessels is not conspicuous on the surface of the kid-
neys, but is better seen by cutting the kidneys open. That not all
of the blood necessarily traverses the capillary spaces of the kid-
neys is indicated by the fact that injections of the systemic veins
frequently fill the veins that carry the blood away from the kid-
neys as well as those leading to it. This is much more frequently-
the case when injecting from the posterior surface of the adductor
muscle than when injecting from other places, and seems to be
dependent upon a direct connection between the vessel in ques-
tion and the sinuses on the antero-ventral surface of the adductor
muscle near the dorsal ends of the kidneys.

Of the blood that leaves the heart, only that which goes to
the mantle remains to be accounted for. This is collected and
returned directly to the heart (Fig. 5, pv.}

All of the blood that leaves the kidneys is conducted to the
gills. The blood from each kidney is collected into a sinus that
runs along the border of the kidney that is applied to the adductor
muscle. This sinus, which also seems to receive blood from the
sinuses on the anterior and ventral surfaces of the adductor muscle,
bends abruptly ventrally over the anterior end of the kidney and
is continued on the lower border of the suspensory membrane of
the gill (Fig. I, ba) to the posterior end of the gill, supplying
the gill with branches throughout its length.

Blood vessels leave the vessel that carries blood from the
kidney, opposite each of the inter-lamellar junctions of each of
the gills supported by the suspensory membrane. Each of these
branches is continued along the free border of the membrane
that forms the inter-lamellar junction (Fig. 7, ba') until it reaches
the free edge of the lamella, the edge that is not attached to the
suspensory membrane. That, is, if the branch supplies an outer
gill, it leaves the suspensory membrane along the free border of
an inter-lamellar junction and crosses over to the free border of
the outer lamella of this gill. Here the vessel is continued
down the enlarged, modified filament that is concerned in the
formation of the inter-lamellar junction (Fig. 7, ba"} giving out
side branches through each of the inter-filamentar junctions

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 233

(as long as these are composed of tissue that can carry blood
vessels) 1 and so supplies the various filaments of the lamella.
The blood thus distributed finds its way around the margin of
the gill through small blood spaces and is continued up the other
lamella of the gill, the blood of the small filaments being gradually
collected through the vessels of the inter-filamentar junctions into
the vessels of the large filaments (Fig. 7, &/), and by these poured
into a vessel that lies just beneath the vessel that supplies the
gill and runs parallel with it (Fig. i, h>). This vessel receives all
of the blood from both of the gills of the side, and carries it di-
rectly to the corresponding auricle of the heart. Just before the
vessel empties into the heart it receives a rather large vessel from
the corresponding lobe of the mantle, which returns the blood
that was sent to the mantle back of the heart.

To sum up the course of the circulation of the blood briefly,
it will be seen that of the blood that leaves the heart only that
which is sent to the mantle is returned to the heart after travers-
ing a single set of capillary spaces ; that a small portion of the
blood sent to the adductor muscle (that which is collected by
the sinuses on the antero-ventral portion of the muscle) may be
returned after traversing two sets of capillaries those of the
adductor muscle and those of the gills ; and that the greater por-
tion is returned only after traversing three sets of capillaries
those of the general system, those of the kidneys, and those ot
the gills.

The reasons for this arrangement of the circulatory system
are at least in part not hard to find. The blood which
passes to the mantle loses some of its nourishing materials,
but as the mantle lobes are thin and are bathed over such a
large portion of their surfaces by a current of water, in which
there is an abundance of dissolved oxygen, respiration, no doubt,
takes place direct, and the blood has no need to pass through
the gills to get a supply. Again the work of the mantle is not
of such an active nature as to load the blood with nitrogenous
wastes. It seems likely that the amount of nitrogenous waste
in the blood that has traversed the mantle is so small that it

'The inter-filamentar junctions near the free margins of the gills are composed of
cilia only.

234 OILMAN A. DREW.

would diminish the proportion of nitrogenous waste in the blood,
if this blood were added to the blood that passes through the
kidneys.

The blood that goes to the general system must in its progress
lose a considerable portion of its oxygen, and in all portions
except around the alimentary canal (where there is, of course, a
decided gain) also food materials, and gain from the tissues
a considerable amount of nitrogenous and carbonaceous wastes.
It is then essential that such blood should go to the excretory
and respiratory organs to get rid of these waste products and to
gain oxygen. Inasmuch as the heart provides for but a single
circulation it is necessary that the capillaries of these organs be
traversed before the blood is returned to the heart. Why it is
arranged so part of the blood may dodge the kidneys and be
carried directly to the gills is not nearly so evident. Possibly the
periodically great activity of the adductor muscle causes the blood
to move through it so rapidly that the small kidneys cannot
take care of it and properly perform their function, and the
other channel is provided to carry the surplus away to the
comparatively extensive gills where the increased flow can be
taken care of with greater ease. It is, of course, essential that
the amount of oxygen in the blood at such times shall not be
reduced. It is at any rate evident that there is a possibility that
part of the blood that is returned from the muscle, liver, etc.,
may not pass through the kidneys, for when starch injecting mass
is injected through a vessel that carries blood from one of the
kidneys to the gills, not only are the kidney and the gills injected,
but part of the mass usually finds its way into the adductor
muscle, liver, and other organs of the body.

The rate of the heart beat is slow, and as in other lamelli-
branchs is, no doubt, dependent upon the temperature of the
animal as well as on other factors. The auricles and ventricle
become very greatly distended during diastole, and contract so
that their cavities are almost entirely obliterated in systole.

NERVOUS SYSTEM.

The three pairs of ganglia that are usually found in lamelli-
branchs are present in this form, but they differ greatly in size and
they are not all placed in the usual positions.

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 235

The cerebral ganglia (Fig. 6, eg) are placed some distance
ventral to the mouth, just beneath the outer covering of the body.
They, like the other ganglia, are yellowish in color, and may
frequently be faintly seen through the covering of the body.
Each cerebral ganglion is somewhat elliptical in outline with the
long axis directed dorso-ventrally and has a rather distinct
swelling on the ventral (actually anterior) and outer side (the
side away from the median plane of the body) (Fig. 9, eg). The
anterior end of each cerebral ganglion presents a forked appear-
ance, due to the origin of two large nerve cords. The inner and
ventral one of these two cords (Figs. 8 and 9, cc) is the commis-
sure that joins the two cerebral ganglia. As the ganglia lie
some distance ventral to the oesophagus, this commissure forms a
long loop that passes dorsally around the oesophagus just pos-
terior to the mouth. The outer and posterior of the two large
cords that leave the anterior end of each ganglion is the anterior
pallial nerve (Figs. 6, 8 and 9, apii). This runs parallel with
the commissure as far as the oesophagus and is then continued
along the side of the liver, and in the mantle, to the margin of
the mantle in the region of the anterior ear of the shell, where it
joins by several branches the circumpallial nerve (cpii) that fol-
lows along the margin of the mantle near the bases of the ten-
tacles and eyes. The circumpallial nerve will receive attention
later.

Between the points of origin of the cerebral commissure and
the pallial nerve, a small nerve (Figs. 8 and 9, /;/) leaves the
ganglion to be continued dorsally, and to supply the labial palp.

From the inner, ventral surface of each cerebral ganglion, a
little in front of the middle, the cerebro-pedal connective leaves
to join the pedal ganglion of the same side. The cerebro-pedal
connective is smaller near the cerebral than the pedal ganglion
(Fig. 9, cpc] and bears a ganglionic swelling on its outer side
very near the pedal ganglion.

In the acute angle formed by the surface ot the cerebral
ganglion with the cerebro-pedal connective, a small nerve (otii),
the otocystic nerve, leaves the ganglion to be continued around
the dorsal surface of the cerebro-pedal connective to the otocyst
of the same side.

236 GILMAN A. DREW.

Posteriorly the cerebral ganglia taper rather gradually into the
cerebro-visceral connectives, which run along the sides of the
visceral mass very near the adductor muscle, until the visceral
ganglia are reached.

The pedal ganglia lie very near each other (Fig. 9, pg), so the
commissure that connects them is short and broad and presents
ordinary ganglionic structure. They are separated from the
cerebral ganglia only by a short interval, and lie anterior and
slightly ventral to them, some distance dorsal to the base of the
foot. They lie so near the surface that their color may fre-
quently be distinguished through the body wall beneath the
mouth. Two large nerves (fit) leave each pedal ganglion to be
continued into the foot, where they supply the muscles of the
foot and probably the byssal gland. The swellings on the cere-
bro-pedal connectives near the pedal ganglia have already been
described. The otocystic nerves, which usually leave the cerebro-
pedal connectives near the pedal ganglia, in this form originate
directly from the cerebral ganglia near the point where the con-
nectives leave the ganglia.

The visceral ganglia (Figs. 6, 8 and 10, vg) are by far the
largest and most complicated of the ganglia, and from them nerves
are sent to most parts of the body. They are situated on the
antero-ventral surface of the adductor muscle, nearly opposite the
external openings of the kidneys. They are imbedded in a mass
of connective tissue and are fused to each other, so the commis-
sure that connects them is nearly as broad as the ganglia them-
selves and shows ganglionic structure. The chief indication of
the presence of a pair of ganglia is the arrangement of the nerves
that leave them, and of the cerebro-visceral connectives that join
them. The ganglia are divided into very definite regions, each
of which is connected with definite bundles of nerve fibers and, no
doubt, has a particular function to perform. I have not had time
to make a detailed study of the structure and nerve tracts of the
ganglia, but I am satisfied that there is much more complexity
than is ordinarily attributed to the ganglia of lamellibranchs.
The dorsal surfaces of the ganglia are quite smooth, but when
seen from the ventral surface (Fig. 10) the regions that are indi-
cated in the figure are always visible. On each cerebro-visceral

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 237

connective, just before it joins the ganglion proper, there is a
ganglionic swelling (x) that supplies one of two roots of a
nerve (Figs. 6, 8, and 10, b)i) that leaves in an antero-dorsal direc-
tion along the border of the excretory organ, to bend ventrally
and posteriorly in the suspensory membrane of the gills, and sup-
ply the gills of the corresponding side. Between the points
where the cerebro-visceral connectives join the visceral ganglia
on the ventral side, there are four rather distinct swellings, with
three less distinct swellings posterior to them. Extending later-
ally from the outer side of each ganglion is a somewhat flattened
ridge (Fig. 10, j') from which all of the pallial nerves from this
ganglion originate. These nerves (Figs. 6 and 8, //w)pass laterally,
posteriorly and anteriorly along the surface of the adductor muscle,
to meet the mantle lobes and to be continued to the margins,
where they unite with the circumpallial nerves. It will be noticed
that they unite with the circumpallial nerve at intervals through-
out the greater length of this nerve. As the pallial nerves
that leave the visceral ganglia in most forms pass directly to
the posterior portion of the mantle, the distribution in this
form may be looked upon as evidence that all of this portion of
the mantle belongs morphologically to the posterior portion of
the animal.

Other nerves leave the dorsal surface of the visceral ganglia
near their posterior ends, and enter the adductor muscle directly.
The nerves that supply the posterior division of the muscle are
continued along the ventral surface of the anterior portion of
the adductor muscle until this posterior portion is reached.
Small nerves also leave the ventral side of the ganglia and pene-
trate the visceral mass.

All of the ganglia are well supplied with nerve cells, there
being very many large polar cells present, but the number of the
cells is far greater and their arrangement more complicated in the
visceral than in any of the other ganglia.

Nerve cells are also to be found in the circumpallial nerves
and in the branchial nerves. So abundant are the nerve cells
in the circumpallial nerves that they assume the structure of
ganglia. The nerves by which they are connected with the
visceral and cerebral ganglia contain no ganglionic cells. From

238 OILMAN A. DREW.

the structural standpoint we would accordingly be justified in
considering the circumpallial nerves as separate ganglia, and the
nerves connecting them with the visceral and cerebral ganglia
as connectives.

The circumpallial nerves of the two lobes of the mantle are
connected with each other anteriorly and posteriorly near the
hinge line (Fig. 8, cpn]. They are not of constant diameter, but
suddenly increase or diminish in size so that they have a rather
irregular appearance. They lie just inside, that is, toward the
median plane of the body, of the large pallial arteries that sup-
ply the mantle margin, about opposite the line of attachment of
the infolded ridge. From them nerves are sent to the eyes and
tentacles, to the infolded ridge and to the pallial muscles. Very
likely the pallial muscles are partially supplied from the pallial
nerves that come from the visceral ganglia, but of this I am not
sure.

It seems probable that the ganglionic structure of these nerves
has been developed to meet the needs of the very complex mar-
gins of the mantle. The development of ganglia in the immedi-
ate region of the sense organs is an indication of the ease with
which such centers may be established when need arises.

The branchial nerves are supplied with ganglionic cells through-
out their length. These are present not only along the borders
of the gills, but from the points where the nerves originate to
their extremities. The almost constant activity of the gills no
doubt renders such an arrangement desirable. No other nerves
or connectives in the body seem to be abundantly supplied with
ganglion cells.

The whole nervous system is modified to meet the special
needs of the animal. The cerebral and pedal ganglia are small,
corresponding with the slight development of the anterior parts
of the body and of the foot. The visceral ganglia are highly
developed, corresponding to the excessive development of the
parts that are supplied by these ganglia. Accessory centers
have also been developed in the margins of the mantle and in
the gills.

It seems that many students of Mollusca hold that the lamelli-
branch ganglia have been derived from a gastropod-like type, a

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 239

type that possesses at least one pair of ganglia, the pleural, that
are not commonly found in lamellibranchs. This view seems to
be based largely upon the acceptance of a hypothetical type for
a primitive mollusk that seems to me to be a much better ancestor
for the gastropods than for the other classes of the Mollusca.

PHYLOGENY.

The hypothetical primitive mollusk that has persistently been
offered for our consideration, and has found its way into a num-
ber of text-books, among which is Lang's " Text-Book of Com-
parative Anatomy," has the dorsal portion of the body covered
by a conical shell, the foot flattened and adapted for creeping, a
head fold that may be protruded from beneath the shell, a pair
of plumose gills, and a nervous system with at least four pairs of
definite ganglia, cerebral, pleural, pedal and visceral. Distinctly
gastropod throughout.

If the development of animals is to be considered of any
importance in pointing their possible lines of descent, and as
long as embryo chicks have gill arches our belief has good
foundation, it would seem that in those mollusks whose eggs
are not loaded with yolk, whose embryos are not modified
for protection in brood pouches, and do not have long larval his-
tories that call for special modifications to enable them to cope
with enemies and to get food, the embryos might be suggestive.

The presence of unlimited food and protection always tend to
destroy characters. Thus we find that parasitic forms may have
entirely lost organs that must have been well developed before
the animals took to parasitic lives. The presence of a quantity
of yolk furthermore frequently must have mechanical effects on
the developing embryo that cause direct modification. Again
those embryos that pass through long larval histories exposed to
the competition of forms that would eat their food and other
forms that would eat them, must necessarily be exposed to the
same evolutionary factors, whatever they may be, that adult ani-
mals are exposed to and we would accordingly expect adaptive
modifications in them.

There are many lamellibranchs, and not a few gastropods,
that do not seem to be seriously modified by any of these fac-

240 OILMAN A. DREW.

tors and when their embryos are examined every one must be
struck with their close resemblances. These embryos would
seem to point to a free swimming ancestral form that obtained its
food by means of surface cilia.

The first living forms that made their appearance on the earth
must have used non-living substances for food. What the nature
of these substances were, whether they were of a comparatively
simple nature, like those that are used by our green plants
to-day, or whether they were of an entirely different nature, we
have no means of knowing, but it is evident that their food was
not alive.

Then came the discovery by some form that the protoplasm
of other forms could be used for food. This must have been the
first great factor that led to the competition of forms and called
for the improvement of bodily machinery among living things, to
aid in the struggle thus begun, the struggle to get food and to
escape from being used as food. As Professor Brooks has indi-
cated, 1 this would naturally lead to the discovery and colonization
of the bottom of the ocean because of the greater advantages it
offered both for capturing food and in affording means of protec-
tion. This introduces the further element into the competition,
of some positions being far more favorable than others, and as
the struggle for position increased, a struggle that has never
ceased, the competition, especially between close relatives, must
have become very severe.

These factors, with the struggle dependent upon them, must
have caused changes in structure (in the improved machinery that
aids forms in getting food and in keeping from being used as
food) to change very rapidly and it seems very plausible that in
a comparatively short time in those days when forms were of
simple structure and this keen competition was begun, the founda-
tions of the great types of animal structure were laid.

We know that among our earliest fossils are to be found both
lamellibranchs and gastropods, and it is back in the earlier time
that we must look for the changes that have resulted in the for-
mation of these classes.

1 Brooks, " The Origin of the Oldest Fossils and the Discovery of the Bottom of
the Ocean," Smithsonian Report for 1894 (also Salpa).

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 24!

We may possibly conceive that the ancestor of the Mollusca
was among the early ones to recognize the advantages of the
ocean bottom, and that its race soon developed a protective shell,
if this had not started to form before it became a dweller on the
bottom. The shell would offer protection, but would, because of
weight, interfere with rapid movement. As enemies became able
to get beneath its armor the shell became thickened and was
made to cover the animal more completely, but the added weight
interfered still more with rapid movement.

At this time we need not suppose that the animal had more
than the very simplest nervous system, hardly more than that
needed by a trocophore larva, for it would probably be depend-
ent upon simple bands of cilia, or at the most a movable mouth
portion, for getting its food. There is no reason for supposing
that this animal had yet developed gills, or if gills were present
they would hardly be more than simple folds of the mantle.

As competition became more severe, animals of this kind were
in need of better protection, and it is possible to conceive that
there might have been evolved two types, one that inclosed itself
in a bivalve shell, crawled into the mud, and obtained its food by
capturing the forms brought to it in a current of water of its own
creation, the other, more like the hypothetical primitive niulliisk
that IMS been described, which retained a single shell and got its
food by creeping over the bottom and picking it up directly.
The first form would still have a simple head apparatus and
would need new nervous centers only to provide for the mech-
anism necessary to crawl into the mud and the mechanisms
necessary to create the current of water and capture the living
forms from it. The second form would have a more complicated
head apparatus and would need nervous centers to supply it and
to supply the organ by means of which it was enabled to creep.
In these differences in life, and in the consequent differences in
structure, it seems reasonable to look for the differences in their
nervous systems. If this conception is anything like true, from
very early times there was no similarity in the method these two
groups used in getting food. One has finally developed a re-
markably satisfactory method of straining out living particles that
serve it as food, from a current of water of its own formation, and

242 OILMAN A. DREW.

is thus able to leave little of its surface exposed to the attacks of
enemies. The other has developed one of the most complicated
of machines in connection with its mouth to aid it in getting food.
As the head apparatus of the one type has increased in com-
plexity, there has been greater need of ganglia to supply it, but
in the whole line of development of the other type there has
been no complicated head apparatus. About all of the actual
evidence that we have of the presence of pleural ganglia in lamel-
libranchs is that given by Pelseneer, 1 who finds in Nucnla and
some other forms, that each anterior ganglionic mass is so shaped
that it is possible to consider it as two ganglionic masses, and
further that the connective that runs from this mass to the pedal
ganglion is connected with this mass by two roots. The inter-
pretation that he has put on this is that the two apparent divisions
of the ganglion represent respectively the cerebral and pleural
ganglion, and that the roots of the connective represent the
cerebro-pedal and pleuro-pedal connectives that have become
fused before reaching the pedal ganglion. My own view, dis-
cussed in another paper 2 is that the apparent division into two
ganglionic masses is superficial, and due to the swellings accom-
panying the origins of nerves, and that one of the cerebral ends
.of the connective may be the central end of the otocystic nerve
which is fused for the greater part of its length with the connec-
tive, but, unlike most forms, is free near the ganglion. This view-
seemed to me most reasonable as Stempell 3 has found that in
Solcynia togata, a supposed near relative of Nitcula, the otocystic
nerve arises directly from the cerebral ganglion and is separate
from the connective throughout its length. So far as I know,
the instance given by Stempell is the only one that has hereto-
fore been reported where the otocystic nerves originate from the
cerebral ganglia, and are free from the cerebro-pedal connectives
throughout their length. Pecten temiicostatus has the same
arrangement. In this form the position of the ganglia, connec-
tives and otocysts is such that it is a very simple matter for the

JPelseneer, "Contribution a 1'etude des Laruellibranchs," Arch, de BioL, XL,
1891.

2 Drew, "The Life-History of Nucula delphinodonta" Quart. Jour, of Micro.
Sfi., Vol. 44, Part 3, New Series, 1901.

'Stempell, " 7-ur Anatomic von Solemya togata" Zool. Jahrb., Bd. XIII. , 1899.

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 243

otocystic nerves to make direct connection with the cerebral
ganglia, but they do not join the ganglia at their nearest points.
Instead they are continued around the connectives to join the
ganglia in contact with, and posterior to them.

To me it seems probable that the separation into the two
groups that have developed into the classes Lamellibranchiata
and Gastropoda took place at an early date in the history of the
Mollusca, probably before a complicated head apparatus was de-
veloped, and while the nervous system was of a very simple nature.
If this was the case, we have no reason to search for pleural
ganglia in lamellibranchs, for it is very probable that they never
had them. In fact were ganglia ever present in this region in
lamellibranchs, it would be more reasonable to view them as new
formations for special purposes than as direct descendants from,
and accordingly homologous with, the pleural ganglia of gastro-
pods. The gastropod and lamellibranch are so different in struc-
ture and habits that we may reasonably expect important differ-
ences in their nervous systems. Gastropods and cephalopods
possess accessory ganglia that have evidently been developed to
perform special functions. That such centers may be compara-
tively easily developed is indicated by the fact that the circum-
pallial nerves of the scallop are essentially such centers. Is it
not then more likely that pleural ganglia have been developed in
the groups that need them than that lamellibranchs, which, so far
as we know, have never been more complicated than they are
to-day, should have formerly possessed these ganglia and have
since quite uniformly lost them?

ANATOMICAL DRAWINGS.

It sometimes happens that in making a series of drawings in-
tended to illustrate different organs of the same animal, consid-
erable labor can be saved by using a combination of photograph
and ink. The figures of the present paper illustrate this saving
much better than is usually the case. To draw the margin of
the mantle, with its large number of sense organs, requires both
time and patience, and were it necessary to draw it for each of
the figures where it necessarily occurs, one would be tempted to
abandon it altogether.

244 OILMAN A. DREW.

It occurred to me, while engaged in drawing this margin, that
possibly it could be photographed on a paper of a quality that
would allow pen drawing and thus save redrawing it. After some
trials a platinum paper was found that met the requirements but
I was surprised to find how much blacker Higgins ink was than
the blackest print I could make.

Evidently, however, any mark that would take at all in mak-
ing a zinc etching would print the same color as the rest when
being put through the press, so one of the poorest of these photo-
graphs was finished with Higgins ink and sent away to have a
zinc etching made from it. The result was perfectly satisfactory.
It will be seen that the margins on Figs, i, 2, 3, 4, 6, 1 1, and 12
are all alike. The margin of Fig. I is the only one that was
made with pen and ink. Fig. 1 1 is a print of a negative made
from this margin before the rest of the animal was drawn.
Taking a print similar to that shown in Fig. 1 1, with pen and ink
there was drawn into it the organs shown in Fig. 12. Fig. 2,
before the alimentary canal was added, was the figure from which
the photograph resulting in Fig. 12 was taken. The margin of
Fig. 12 is then a photograph of a photograph of an ink drawing.
The original of Fig. 12 was then worked on to form Fig. 2 just
as the original of Fig. 1 1 was worked into Fig. I. Figs. 3, 4 and
6 are all worked onto prints similar to that shown in Fig. 12. In
the original paper in which these figures were published a num-
ber of others were based on photographs in a similar way. The
saving of time in the paper probably amounted to more than one
half, and certainly may be of importance to others. I have no
doubt that photographs may also be made the basis of brush
work, but great care will be necessary in such cases in getting
the proper printing value. While the figures accompanying this
paper show no evidence that the photograph and the ink had
different printing valves, they would have been very unsatisfactory
had they been reproduced by some other processes.

SUMMARY.

Circulatory System. The large size of the animal makes it
possible to inject the vascular system successfully. Blood from
the mantle is returned immediately to the heart. Most of the

CIRCULATORY SYSTEM OF THE GIANT SCALLOP. 245

blood from other portions is carried to the kidneys, from which
it is carried to the gills, and then back to the heart. A portion
may dodge the kidneys and go to the gills. Blood seems to act
both as blood and lymph. (See pp. 227-234 and Figs. 3, 4, 5
and 7.)

Nervous System. The cerebral and pedal ganglia are small
and somewhat removed from their usual positions. The visceral
ganglia are very large and complicated in structure. The circum-
pallial nerves and the branchial nerves have ganglion cells
throughout their length. The otocystic nerves originate directly
from the cerebral ganglia. (See pp. 234239 and Figs. 6, 8,
9 and 10.)

Phylogeny. Ontogeny and the probable conditions that have
resulted in the complication of structure, both seem to indicate
that the division of the Mollusca into lamellibranchs and gas-
tropods, took place at an early time, before the ancestors had
attained much complexity of structure.

There seems to be no reason for believing that lamellibranchs
ever had more complicated head machinery than they have at
the present time. If this is true they probably have never had
need of more anterior ganglia than they now generally have.
(See pp. 239-243.)

Anatomical Draivings. A combination of photographs and
drawings may sometimes save much time and tedious work.
(See pp. 243 and 244.)

UNIVERSITY OF MAINE,
ORONO, MAINE,

November 15, 1906.

246 GILMAN A. DREW.

PLATE XVI I.

FIG. i. Animal as seen from the left side with the left shell valve and mantle
lobe removed and with a portion of the pericardial wall cut away. A few of the
blood vessels are shown. Two thirds natural size.

FIG. 2. Animal as seen from the left side with the left shell valve and mantle
lobe removed, with the alimentary canal shown. Two thirds natural size.

, auricle ; ba, branchial artery ; bv, branchial vein ; c, cartilage ; e, excretory
organ ; /, foot ; fe, free edge of the unattached lamella of the gill ; g, gill ; i, intes-
tine ; Ip, labial palp ; m, mantle ; /<?, posterior adductor muscle ; s, stomach ; v,
ventricle ; nil, visceral mass.

BIOLOGICAL BULLETIN, VOL. XII

PLATE XVII

pa

m

ba

FIG. i.

pa

FIG. 2.

Drew del.

248 OILMAN A. DREW.

PLATE XVIII.

FIG. 3. Animal as seen from the left side with the left shell valve and mantle
lobe removed. Drawn to show the arterial system of blood vessels. Two thirds
natural size.

FIG. 4. Animal as seen from the left side with the left shell valve and mantle
lobe removed. Drawn to show the systemic veins. Two thirds natural size.

a, auricle ; aa, anterior aorta ; apa, anterior pallial artery ; e, excretory organ ;
fa, foot artery ; fv, loot vein ; ha, hepatic artery ; hv, hepatic vein ; pa, posterior
adductor muscle ; paa, posterior adductor artery ; pav, posterior adductor vein ;
ppa, posterior pallial artery; v, ventricle; i>a, visceral arteries; vm, visceral mass.

BIOLOGICAL BULLETIN, VOL. XII

PLATE XVIII

PPa

FIG. 3.

FIG. 4-

Drew del.

25O OILMAN A. DREW.

PLATE XIX.

FIG. 5. Outer surface of the left lobe of the mantle showing the arrangement of
blood vessels. Two thirds natural size.

FIG. 6. Animal as seen from the left side with the left shell valve and mantle lobe
removed. Drawn to show the nervous system. Two thirds natural size.

apa, anterior pallial artery; apn, anterior pallial nerve; bn, branchial nerve;
cc, cerebral commissure ; eg, cerebral ganglion ; cpn, circumpallial nerve ; cvc, cere-
bro-visceral connective; of, otocyst ; pa, posterior adductor muscle (anterior por-
tion) ; pa' , posterior adductor muscle (posterior portion) ; pg, pedal ganglion ; ppa,
posterior pallial artery ; ppn, posterior pallial nerve ; pv, pallial vein ; vg, visceral
ganglion.

BIOLOGICAL BULLETIN, VOL. XII

PLATE XIX

apa

,pa'

FIG. 5.

cvc

oo

cpn

ppn

FIG. 6-

Drew del.

252 OILMAN A. DREW.

PLATE XX.

FIG. 7. A portion of a gill showing the arrangement of parts. The figure indicates
the inter-lamellar junctions cut at different levels. The further lamella is the one
that was attached to the suspensory membrane and the vessel (f>a / ) was directly con-
nected with the vessel that supplied the gill with blood (&a, Fig. l). This vessel
follows along the edge of the inter-lamellar junction to the free edge of the unattached
lamella (the one on the side nearest the observer in the figure), where it bends back
and passes down the modified filament as the vessel ba ff . Branches are given off
from this vessel through the inter-filamentar junctions to supply the filaments. The
vessel bv' is the vessel into which the blood that has traversed the gill is collected.
It in turn communicates with the vein of the gill (bv, Fig. l). Magnified about
seventy diameters.

ba> ', branch of the branchial artery; ba" ', branch of the branchial artery in the
modified filament ; bv' ', branch of the branchial vein ; ft; chitinous rod ; gf, gill
filament ; ifj, inter-filamentar junction ; ilj, inter-lamellar junction ; io, inhalent
ostium.

2 U

254 OILMAN A. DREW.

PLATE XXI.

Fig. 23. Nervous system as seen from in front and a little to one side. Natural
size. (Diagrammatic.)

apn, anterior pallial nerve; bn, branchial nerve; cc, cerebral commissure; eg,
cerebral ganglion ; epc, cerebro-pedal connective ; cfn, circumpallial nerve ; cvc,
cerebro-visceral 'connective; of, otocyst ; pg, pedal ganglion ; pn, palp nerve ; ppn,
posterior pallial nerve ; vg, visceral ganglion.

BIOLOGICAL BULLETIN, VOL. XII

opc

cpn

apn

Drew del.

256 OILMAN A. DREW.

PLATE XXII.

Fig. 9. Cerebral and pedal ganglia with their nervous connections, as seen
from the antero-ventral position. These ganglia and the otocysts lie in a mass of
connective tissue and may be dissected out and mounted for study without injury
Magnified about fifteen diameters.

Fig. 10. Visceral ganglia seen from the ventral side. These may easily be ex-
posed for study by stripping the thin muscular covering from their ventral surfaces.
They are hard to separate from the adductor muscle but they may be mounted with a
thin piece of the muscle and studied in position. Magnified about fifteen diameters.

apn, anterior pallial nerve; bn, branchial nerve; cc, cerebral commissure; eg,
cerebral ganglion ; cpc, cerebro-pedal connective ; cvc, cerebro-visceral connective ;
fn, foot nerve; of, otocyst ; otc, otocystic canal; otn, otocystic nerve ; pg, pedal
ganglion ; /, palp nerve; ppn, posterior pallial nerves; x, swelling on the visceral
ganglion from which the anterior root of the branchial nerve originates ; y, swelling
on the visceral ganglion from which the posterior pallial nerves originate.

BIOLOGICAL BULLETIN, VOL. Xll

PLATE XXII

apn

Drew del.

258 OILMAN A. DREW.

PLATE XXIII.

Fig. n. Etching made from a photograph of the margin of Fig. I, before that
figure had been completed. It will be noticed that the same margin occurs on all of
the figures that show this portion of the mantle.

Fig. 12. Etching made from a photograph of a combination of a photograph and an
ink drawing. The photograph was made from Fig. 2 before the alimentary canal
had been worked in. Fig. 2 was drawn on a print like Fig. II. Figs. 3, 4 and 6
are etchings of drawings made by adding various organs on prints like Fig. 12.

BIOLOGICAL BULLETIN, VOL XII

PLATE XXIII

Fig, 12,

THE MATURATION OF THE MOUSE EGG.

WILLIAM B. KIRKHAM.

Sobotta ('95) after careful study of a very large number of
preparations of the egg and ovary of the white mouse came to
the conclusion that in nine tenths of these eggs the maturation
processes involve the suppression of the first polar spindle, and
the formation of only a single polar body. Gerlach ('06), after
a study of preparations made at least as early as 1890, has re-
vived Tafani's theory that in the majority of mouse eggs the second
polar body is suppressed. Gerlach's conclusion is that when a
spermatozoon enters an egg sometime after it has formed the
second polar spindle, the second polar body fails to develop, and
the spindle degenerates within the egg.

These results are at variance with the majority of opinions
reached, before and since, by investigators of the eggs of other
animals, vertebrate and invertebrate, and a reinvestigation of the
maturation processes in the egg of the white mouse has brought
it into line with most other metazoon eggs.

Material and Method. The mice used have been killed dur-
ing the period of most active breeding, namely, April, May, June
and September, and serial sections made of the ovaries and Fallo-
pian tubes. Ovulation, during the spring months, occurs very
soon after parturition, independent of copulation, as observed by
Rubaschkin ('05) in the guinea-pig.

When observed to be pregnant, the females were mated, and
killed, some a few days or hours before parturition, others during
that process, and still others at intervals from a few minutes to
thirty hours after giving birth to a litter. The tissues were killed
with a variety of the more generally used cytological fluids, and
the following is a brief summary of the results obtained : All
the ovaries contained some eggs with the second polar spindle
and accompanied by the first polar body, and a majority of the
series revealed ovarian eggs at the end of the spireme or with the
first polar spindle. The eggs observed in the Fallopian tube fall
into two main groups : those which had not been fertilized, and

259

260

WILLIAM B. KIRKHAM.

therefore retained the second polar spindle some being accom-.
panied by the first polar body, more without it and those which
had been fertilized. The latter included stages from the entrance
of the spermatozoon through the cleavage stages.

FIG. i. Ovarian egg showing first polar spindle. Zona pellucida represented by
double line. X 1200.

First Polar Spindle (Fig. i). - - The preparations in which there
are stages immediately preceding the formation of the first polar
spindle have not been fully studied, but there is evidence of a
precocious division, the number of chromatin masses being be-
tween twelve and twenty-four.

The first polar spindle when first formed lies with its axis per-
pendicular to the radius of the egg, as found by Rubaschkin
('05) in the egg of the guinea-pig, and later one pole swings some-
what toward the center of the egg. The chromosomes of the
first polar spindle are short and thick (Fig 2), and vary greatly
in size. The spindle fibers come to more or less of a focus,
and centrioles have often been seen at the poles of this spindle,
where they are made up of several distinct, eccentrically placed
granules.

THE MATURATION OF THE MOUSE EGG. 26 1

First Polar Body (Figs. 3 and 5).- -The study of many
preparations reveals the following facts : None of the eggs in the
Fallopian tube have failed to develop at least to the formation of
the second polar spindle, and all the
ovarian eggs which by their size,
slightly denser protoplasm and large
follicles appear to be nearly ripe, have
already extruded the first polar body.
The conclusion arrived at is, that ap-
parently every egg which is capable
of further development forms a first
polar body within the ovary. This
agrees with the observations of Ru-

FIG. 2. Diagram of chromo-

baschkin ('05) upon the guinea-pig somes in first po ] ar sp i n dle. Note

egg, and those of Van der Stricht great variation in size. Four
('Ol) upon the egg of a bat, Vesper- more chromatin masses in adjacent

sections.

it go noctwa.

This point established, it is next necessary to explain the dis-
appearance of the first polar body in the majority of eggs seen in
the Fallopian tube. The zona pellucida may persist in the mouse
egg, undiminished, through the early cleavage stages, but in the
majority of instances during the process of ovulation the first
polar body is either forced through a weakened part of the zona,
or frees itself by amoeboid movements, and comes to lie outside
the zona, as described and figured by Van der Stricht ('04).

The first polar body is usually oval -in form, and is character-
ized, as found by Van der Stricht ('04) in the egg of K noctnla,
by often possessing a little maturation spindle of its own, and in
other instances having its chromosomes scattered. In some of
these cases which possess a spindle, the first polar body would
probably have divided mitotically, as observed by Sobotta ('95)
in the mouse egg, and once by Rubaschkin ('05) in the egg of
the guinea-pig. The polar bodies vary somewhat in size, and in
one series of ovarian eggs there have been found first polar bodies
of about four times the average volume. The number of chromo-
somes in the first polar body is twelve (dyads).

Second Polar Spindle (Fig. 3). Immediately after the forma-
tion of the first polar body, the twelve dyads remaining in the

262

WILLIAM B. KIRKHAM.

egg are drawn into the equator of a new spindle, split longitudi-
nally, and the twenty-four daughter, univalent chromosomes

FIG. 3. Ovarian egg showing first polar body and second polar spindle. Seven-
teen masses of chromatih, some of which are undivided dyads, are scattered through
the first polar body ; twenty-four univalent chromosomes appear in the equator of the
second polar spindle. Certain chromosomes have been added from adjacent sections.
A minute centriole appears at each pole of the second spindle. The zona pellucida
is represented by a double line. ( 1200.

FIG. 4. Diagram of univalent chromosomes in second polar spindle, indicating
difference in size.

lengthen out into filaments of various sizes (Fig. 4). Like the
first polar spindle the second varies in size, and lies with its axis at

THE MATURATION OF THE MOUSE EGG.

263

right angles to the radius of the egg, usually near the first polar
body. Centrioles, similar to those described above for the first
polar spindle, have frequently been observed in second polar
spindles, and in some cases a few radiating aster fibers have been
seen at the poles. In attempting to determine whether a given
polar spindle is first or second, the character of the chromatin
has always been found a positive guide.

FIG. 5. Egg in Fallopian tube showing both polar bodies. Note spindle in first
polar body. The sperm head appears at left, the female pronucleus at right, in the
egg. X 1200.

Mature eggs which are retained within the ovary, together with
such as are discharged and fail to be fertilized, degenerate with
the second polar spindle, as found by Rubaschkin ('05) in the
case of the guinea-pig egg.

Second Polar Body (Figs. 5 and 6). Only one spermatozoon
enters an egg, and it carries in most, if not all of its tail, a fact
observed by Van der Stricht ('04) in the egg of V. noctnla. When
fertilized the egg at once forms its second polar body. This is
more or less nearly spherical, smaller than the first polar body,
and, as stated by Van der Stricht ('04) for V. noctnla, generally

264

WILLIAM B. KIRKHAM.

has its chromosomes gathered into a single compact mass. It
quickly forms a resting nucleus, possessing compact masses of
chromatin, and is usually the only polar body seen during the
early cleavage stages. In one instance (Fig. 6.) a second polar
body was observed which had just been constricted off, and in
consequence showed the separate chromosomes, twelve in num-

FlG. 6. Egg in Fallopian tube showing second polar body. First polar body has
disappeared. At left in the egg is seen the sperm nucleus, and above it the separated
tail of the spermatozoon ; at right appears the egg nucleus, surrounded by delicate
radiating fibers. X 1200.

ber, and another preparation showed the second polar body form-
ing the resting nucleus.

The mouse egg is thus shown to be no exception to the general
rule, that the maturation process in the metazoon egg involves
the formation of two polar bodies.

In closing, I desire to express my gratitude, and great indebt-
edness to Professor Wesley R. Coe for his constant oversight
and encouragement.

THE MATURATION OF THE MOUSE EGG. 265

BIBLIOGRAPHY.
Gerlach.

'06 Ueber die Bildung der Richtungskorper bei Mus musculus. 4 Wiesbaden,
Bergmaiin.

Heape.

'05 Ovulation and Degeneration of Ova in the Rabbit. Proc. R. Soc. London,
Vol. 76 B.

Rubaschkin.

'05 Ueber die Reifungs- und Befruchtungs-processe des Meerschweincheneies.
Anat. Heft., Bd. 29.

Sobotta.

'95 Die Befruchtung und Furchung des Eies der Maus. Arch. f. Mikr. Anat.,
Bd. 45.

Van der Stricht.

'01 La pont ovarique et 1'histogenese du corps jaune. Bull. d. L'Acad. R. d.

Med. d. Belgique, 1901.

'04 Une anomalie tres interessante concernant le developpement d'un oeuf de
mammifere. Ann. d. 1. Soc. d. Med. d. Gand, Vol. LXXXIV.

ON THE ZOOLOGICAL POSITION OF THE ALBINO

RAT. 1

SHINKISHI HATAL, PH.D.
ASSOCIATE IN NEUROLOGY AT THE WISTAR INSTITUTE.

According to Leunis ('83) the black rat ( Mns rattits} was known
in Europe as early as the twelfth century, while the Encyclo-
pedia Britannica (Ol field Thomas, '86) states the appearance of
the black rat to be at least as early as the thirteenth century.
Although the statements by the different writers as to the ap-
pearance of the black rat in Europe do not quite agree, yet it is
clear that the arrival of the black rat was much earlier than that
of the brown rat (Mi is norvegicusf which, according to various
records, appeared in Europe at about the middle of the eighteenth
century, or a little earlier.

Although both species of rats are described as originally
natives of Central Asia, yet they are everywhere enemies. By
the incessant competition between these two forms, the black
rats were almost exterminated, first from Europe, and later from
the greater part of North America, and at the end of the eigh-
teenth century, the brown rats were alone found in abundance
in these regions.

It is often stated that the white rat at present found in cap-
tivity, is the albino of Mits rattits. In support of this view there
are a number of statements to be found in the older literature
(Donndorff, 1792). (No effort has been made to examine the
records previous to Linneus).

It is apparently on the basis of these records in the older liter-
ature that the current statements in popular natural histories and
in encyclopedias are based.

On the other hand, in the zoological literature in the nine-
teenth century, there are numerous statements which refer to the
albino rats as a variety of Mus dccitinainis.

1 From the Wistar Institute of Anatomy and Biology at Philadelphia.

2 Ahts tinn'e^i, us, Krxleben = Mus dc.ntiianiis Pall, of older Zoological Litera-
ture. Norvegieus has priority, and has come into general use within the last two or
three years.

266

ZOOLOGICAL POSITION OF THE ALBINO RAT. 267

Von Fischer ('69) in a catalogue of the mammals of the St.
Petersburg Government, makes the following statement :

" Die Wanderratte, Mus dcciunainis Pall, (russisch Kryssa-
Kryssa heist eigentlich Mus rattus, diese art ist bekannt unter
dem namen Passjuck) kommt ueberall massenhaft vor in alien
Farben ; schwartz, schmutziggrau bis rostgelb, weissgescheckt
und auch ganz weiss. " Die Hausratte, Mus rattns L., habe ich
nie gefangen, weshalb ich annehmen zu durfen glaube, dass diesse
Ratte hier auch nicht' vorkommt."

Von Fischer ('74) used a white Mns norvegicns in his experi-
ments on the production of hybrids. Later Crampe ('85) also
used a white Mus noruegicus in experiments of the same nature.

Haacke ('95) and Bateson ('03) studied the crosses between
the white Mns norvegicus and the common brown rat. None
of the authors, however, describe in detail the white forms which
they employed.

Despite the general belief to the contrary, there are many re-
ports in recent literature indicating that groups of Mus rattttsa.ro.
still to be found in a number of localities, both in Europe and the
United States.

In the United States, Mus rattns is reported from Texas, Florida
and other southern states, and also from Iowa. Rhoads ('03)
reports a number of new localities in the States of Pennsylvania
and New Jersey. It has been learned through Director Dr.
Seitz that in Germany the black rat is present in large numbers
in the buildings connected with the zoological garden in Frank-
furt a/m.

It may be interesting to note that the occurrence of white rats
in a wild state has been reported from two localities in Iowa, by
students working in the neurological laboratory at the University
of Chicago. There are no means of determining, however,
whether these were albinos of the black or brown rat. From this
review it is evident, therefore, that there are, or have been, at least
two forms of albino rats.

Since 1893 a colony of albino rats has been maintained in the
neurological laboratory at the University of Chicago, and in 1906
a similar colony was established at the Wistar Institute of Anat-
omy at Philadelphia.

268

SH1NKISHI HATAI.

These colonies have been recruited for the most part from the
northern states of the Atlantic seaboard, but some specimens
have come from as far south as Missouri. All the rats received
from these various localities have appeared to be of the same
variety, and have always bred true.

Heretofore, the specific similarity of the albinos and the other
forms has been concluded from observation of the external char-
acters only. Wishing more exact information as to the zoologi-
cal relation of the rats composing these colonies, the present
investigation was undertaken to determine whether we were deal-
ing with an albino variety of Mus rattus or Mus decumanus.

Externally, Mus rattus is usually distinguished from Mus uorve-
gicus by the following specific characters :

Mus rattus is smaller in size. The tail of Mns rattus is con-
siderably longer than the body, while in Mus norvegicus it is
either shorter or only slightly longer than the body, but not rela-
tively as long as that of Mus rattus.

The following measurements, though incomplete, serve to in-
dicate this relation :

TABLE SHOWING LENGTH OF BOBY AND OF TAIL.

Mus r ail us-

M

us norvegicu

j.

Observer.

Length
Body.

Length
Tail.

No. of
Obser.

No. of
Obser.

^Length
jTail.

Length
Body.

New Interna-
tional Encycl..
Leunis

21 cm.
1 6 cm.

19 cm.

IQ cm.

27 cm.
24 cm.

Hatai

27 males

21 cm.

4 cm.

The general shape of the head (see Fig. i) of 3 fits rattus is
slender, the nose is sharper, and the ear is both wider and longer
than in Mus norvegicus. It may be worth while to mention that
the so-called Alexandrian rat (Mus alc.vandrinus) is said to have
external characters similar to those of the black rat (Mus rattus)
and these two species are only distinguished by their coloring,
Mus alexandrinus having a brown colored coat.

If we compare the external bodily characters of the albino rat
found in our rat colonies, with those of the brown rat, we are
surprised by their close similarity. All these characters of the
brown rat are also characters of the albino rats composing our

ZOOLOGICAL POSITION OF THE ALBINO RAT.

269

colonies. In other words, the common brown and our albino
rats cannot be distinguished from one another by their external
characters.

It is nevertheless true that the albino rats which we have ex-
amined, are smaller in size than the brown rats in the same locali-
ties. In fact, the absolute size of the albino rat is nearly inter-
mediate between Mus rattus and Mus uorvcgicus. It is possible

FIG. I. Copied from "Encyclopedia Britannica," in order to show the shape of

the heads of the brown and black rats. I . Mus rattus.

2. Jhts norvegicus.

that the confinement in which these albinos have been reared,
accounts for their smaller size, as the result of lack of exercise
and altered conditions of life. It is possible also that we have here
a phenomenon similar to that described by Semper ('Si) and De
Varigny ('94) on snails, where the size of the animals diminished
with the size of the vessels in which they were reared.

It was thought that the character of the skull misfht serve for

<-> o

a more exact distinction of the forms under discussion. We
therefore examined and compared the skulls of Mus rattns, Mus
norvegicus, and of the albinos. 1

1 In order to make this comparison, it was necessary to examine as many skulls as
possible, and I am indebted to Professor J. A. Allen, American Museum of Natural
History, at New York, Professor Elliot, Field Columbian Museum at Chicago, Dr.
Greenman, The Wistar Institute of Anatomy at Philadelphia, and Professor Merriam,
National Museum at Washington, for putting at my disposal various series of skulls,
possessed by their several institutions.

2/O SHINKISHI HATAI.

To illustrate the differences found, both photographs and draw-
ings have been made.

On comparing the skull of Mns rat/its with the brown rat, the
general un likeness can be seen in Fig. 2.' The most noticeable
difference is in the shape of the cranium.

When viewed from the dorsal aspect, the cranium of Mus
rattns is oval in the outline, while that of Mus norvegicus is some-
what rectangular. Moreover, the dorsal aspect of the cranium
in Mns rattns is decidedly convex, while in Mus norvegicus it is
nearly flat. In Mns rattns the os nasale as compared to the
entire length of the skull, is relatively shorter than Mus norvegicus.
In J///Y rattns, the outline of the os interparietale is somewhat
semilunar in shape, while in Mus decumanus it is rectangular.
In J///.v rattns, the os parietale is broader as compared with its
length, than in Mus decumanus. In Mus rattns, the foramen
magnum is subcircular in outline, while in Hfus norvegicus it is
somewhat rectangular. On the ventral aspect of the skull, the
large tympanic bullae in Mus rattus are more conspicuous and
eminent than in Mus norvegicus.

The junction point of the os basi-sphenoidale and os basi-
occipitale is flat in Ulns rattns, and protrudes in Mns norvegicus.
The anterior end of the maxilla which forms the lateral wall of
the infraorbital fissure, is blunter in M'us rattns, than in Mus
norvegicus. The skulls of our albino rats are very similar in the
above characters to those of Mus norvegicus, and the description
of J///.S- norvegicus may be taken to apply to them.

In connection with the shape of the skulls, the determination
of a cranial index has been made. The index used, was that
obtained by dividing maximum width of the cranium by the
length of the fronto-occipital line. (See Fig. 3.) On account
of the small number of specimens measured, the accompanying
table is to be considered as merely preliminary, but as it stands
it shows a similarity in this index between Mus norvegicus and
the albino rats, and a difference between these two forms and
M'ns rattus. The cranial index will be made the object of a more
extended investigation.

1 Care has been taken to use only the skulls of fully matured animals. See J. A.
Allen ('94) and II. C. Merriam ('95).

ft

c b a

FIG. 2. Shows the skulls of Mus norvegicus (a), albino rat (/;) and Mus ratlus (t). The skulls were

photographed from two different aspects, in order to show various views of the skulls for a comparison.

The upper row was taken from the dorsal aspect, and the lower from the ventral. The figures are about
the natural size.

ZOOLOGICAL POSITION OF THE ALBINO RAT.

271

FIG. 3. (X two diameters.)' The measurement of frontal-occipital length was
determined in the following way :

Since the length measured from the tip of the nose to the posterior end of the inter-
parietal bone, is not always equal to the length measured from the tip of the nose to
the end of the occipital bone, both measurements were taken. First, the measure-
ment from the tip of the nose to the end of the occipital bone, and second, that from
the tip of the nose to the end of the inter-parietal bone. The difference thus obtained,
was added to the length of the frontal-interparietal line, and the sum was called
frontal -occipital length.

The width of the cranium was determined by taking a maximum width between
jhe two points (right and left) where the zygomatic bones rest on the lateral walls of
the cranium.

We conclude therefore, that the albino rats composing the
colonies at Chicago and Philadelphia, are similar to JMits nor-
vegicns in their bodily proportion, and in their cranial characters.
They are however, smaller in size than the specimens of J\fns
norvegicus usually found.

2/2

SHINKISHI HATAI.

TAF.LE SHOWING CRANIAL INDEX.

Males.

Cranial Index
Average.

Extremes.

No. of Rats Used.

Mil s rat tits

.60

58- 62

8

Mus norvegiciis. . .

S4

. CI-. CC

12

Albino rat

.S4

. so- s6

I 3

Nevertheless this form is to be regarded as an albino variety
of that species and to be designated Mas uorvegicus var. albns
(oculis rubicundis}.

ZOOLOGICAL POSITION OF THE ALBINO RAT. 273

LITERATURE CITED.
Allen, J. A.

'94 Cranial variations in Neotoma micropus due to growth and individual differ-
entiation. Bull, of the Amer. Museum of Natural History, Vol. VI.
Bateson, W.

'03 The present state of knowledge of color heredity in mice and rats. Proc.

Zool. Soc., London.
Bechstein.

'oo Pennant : Allgemeine Uebersicht der vierfiissige Thiere. II., p. 494.
Donndorff, J. A.

'92 Beitrage zur XIII. Ausgabe des Linneschen Natursystem, 2 volumes,
Leipzig. Donndorff refers to the following authors : Schreber, Saugethiere,
IV., p. 649. Naturforscher, I., p. 63, N. 15. Wolf, Reis. nach Zeilan, p.
128. Beckmann, Phys. Okon. B. bl.. V., p. 102, Vergl. mit II., p. 588.
von Fischer, John.

'69 Die Saugethiere des St. Petersburg Governments. Zool. Garten., Vol. X.
'74 Beobachtungen iiber Kreuzungen verschiedener Farbenspielartan innerhalb

einer Species. Zool. Garten., Vol. 15.
Haacke, V. W.

'95 Ueber Wesen, Ursachen, und Vererbung von Albinismus, etc. Biol.

Centralbl.
Kolozy.

'71 On the habits, lameness, and prolificness of the albino Mus rattus. Verh.

Zool. Bot. Gessel. Wien., pp. 731-734.
Leunis, John.

'83 Synopsis der Thierkunde.
Merriam, H. C.

'95 Monographic Revision of the pocket gophers. U. S. Dept. of Agriculture.
Rhoads, S. N.

'03 The mammals of Pennsylvania and New jersey. Privately printed.
Semper, Karl.

'81 Animal Life. New York.
Sterndale.

'84 Natural History of Indian Mammalia.
De Varigny, H.

'94 Recherches sur le nanisme experimental contribution a 1' etude de 1' influence
de milieu sur les organismes. Journ. de 1'Anat. et Physiol., Paris.

NOTES ON THE BEHAVIOR OF SEA-ANEMONES. 1

CHAS. W. HARGITT.

During the summer of 1901 while keeping a few sea-anemones
in the aquarium for the purpose of studying their general habits,
particularly those of feeding, my attention was drawn to the
interesting phenomenon that certain species appeared more
alert during the night, closing up more or less during the
day. This was more noticeable in the large sand-anemone,
Eloactis proditcta, whose peculiar habit of burrowing in the sand,
enabled it to withdraw entirely when disturbed, or under other
unfavorable conditions.

Having secured several specimens of this anemone they were
placed in an aquarium, the bottom of which had been covered with
sand to the depth of some six inches or more. The specimens,
true to their habit, soon burrowed deeply in the sand, and lining
the burrows with a slimy excretion they soon seemed quite at
home. During the day they would be found with only the whorl
of tentacles quietly protruding at the surface of the burrow,
where their colors so closely conformed to that of the sand that
the casual observer would hardly notice their presence. Going
into the laboratory at night I was interested to see the specimens
greatly extended, half of the body protruding beyond the bur-
rows and tentacles raised in an attitude to seize passing prey.
This was frequently observed afterward, and notes made of it at
the time were recorded in which it was remarked that " these
creatures are probably nocturnal in their habits."

At the same time I had under observation another anemone,
Sagartia leucolena, a very common species about Woods Holl,
and it was seen to migrate at times into darker portions of the
aquarium, even creeping under bits of rock or other objects.

No further observations were made on the subject till the cur-
rent summer. About a dozen specimens of Eloactis were col-
lected and placed in the aquarium as before, and with the same

1 Contributions from the Zoological Laboratory, Syracuse University.

274

BEHAVIOR OF SEA-ANEMONES. 275

result that such specimens as found sand proceeded to cover
themselves as far as possible. In the light of current activity
and interest in the matter of behavior it occurred to me to sub-
ject these creatures to a few experiments with a view of testing
their reaction to light, and perhaps a few other environmental
factors.

An examination of the available literature has brought to light
but few instances in which any observations have been made
concerning the behavior of actinians in relation to light. The
Hertwig brothers, '79 (" Die Actinien Anatomisch und Histo-
logisch," p. 191), cite brief observations made by Quatrefages
on species of Edivardsia in 1842, and by Haime on Cerianthus
in 1854, and include likewise brief references to their own obser-
vations on a deep sea-anemone, Cladactis costce, in the Naples
aquarium.

Quatrefages found that when a ray of light from a lamp
was condensed upon the specimens by means of a lens they
partially retracted. Haime observed that in bright sunlight
species of Cerianthus contracted within their tubes and later ex-
panded when the light became less intense. The Hertwigs record
simply the fact the specimens during full daylight were more or
less contracted and expanded as the light became less intense.
" Im tageslicht zieht sie ihren Korper stark zusammen und erst
wenn es zu dunkeln beginnt, dehnt sie sich auf das Vier- bis
Funffache aus und entfaltet ihre Tentakeln, di zuvor eingezogen

waren.'

Jourdan has recorded a similar observation (" Les Sens chez
Les Animaux Inferieurs," Paris, 1889), made upon a species of
Paractis in which similar behavior was exhibited. " J'ai pu voir
moi-meme, sur des Actinies du genre Paractis, des manifestions
evidentes de cette sensibilite speciale. Des Orties de mer restent
fermees aussi longtemps qu'on les expose a une lumiere trop vive ;
alles ne o'epanouissent que lorsqu'on les met a 1'abri des rayons
lumineux " (p. 221).

Eloactis producta. My first observations were made to con-
firm those already cited, namely, to clearly demonstrate their
nocturnal habit. Placed in the aquaria of the general labora-
tory, and in a few cases in smaller jars in my private laboratory,

2/6 CHAS. W. HARGITT.

their behavior was closely watched after the specimens had be-
come adjusted to their new habitat. In this connection should
be mentioned the fact, to be discussed later, that some specimens
were much less prompt in burrowing, a few remaining more or
less indifferently upon the surface of the sand and showing but
slight attempts to bury themselves.

It only required a few observations to determine beyond any
doubt that only in light of low intensity, such as twilight, or in
the aquarium under the rather dim light of an incandescent lamp at
some distance, did the specimens protrude their oral portions and
tentacles and show any degree of activity. To further demon-
strate that these seemingly nocturnal activities were not merely
a periodic response made at more or less definite intervals, the
following experiment was made. A tall glass jar, some twenty
inches in depth, the lower third of which was filled with sand, in
which had been placed several anemones some two days pre-
vious, was so placed on a laboratory table that it was freely ex-
posed to the diffused light of the room. Over the jar was
placed about mid forenoon, when the creatures were securely
withdrawn in the burrows, a blackened chamber or dark hood,
so arranged as to exclude more or less perfectly the light. Re-
moving the hood at the end of an hour it was found that the
creatures were quite extended as at night. And it was soon evi-
dent, that with the removal of the hood and the admission of
light, they were at once aware of the change and promptly began
to show signs of irritation, which ended within five minutes in
every specimen having retracted into its tube. To make certain
that the response had not been induced by some mechanical
stimulus, such as the tremor of passing steps, or an accidental
disturbance of the table or the water in the jar, the experiment
was repeated within a half hour and under conditions which
made it possible to observe the phases of the response.

Within fifteen minutes after the chamber had been placed over
the jar it became quite evident that the change had been recognized
by the specimens. This was shown first by the extension of the
tentacles, and next by a slow protrusion of the oral region by
degrees, till within about half an hour the body was extended
an inch beyond the surface, as before. Again removing the

BEHAVIOR OF SEA-ANEMONES. 2J J

chamber and thus exposing the specimens to the light, within
two minutes, indeed, almost immediately, they began to retract.
This reaction is not sudden or general at once, as in such creatures
as the earthworm, but begins in a somewhat indefinite movement
of the body, accompanied by similar movements of the tentacles,
followed very soon by a slow but definite retraction of the entire
body within the tube, often including likewise the tentacles as
well.

The experiment was later repeated in a room where it was
possible to utilize direct sunlight. Under these conditions the
reaction was much more energetic and definite, as might be ex-
pected. Variously modified, the experiments were performed
repeatedly, perhaps fifty times, and with substantially the same
results, though, as will be noted in a later connection, exhibiting
variations of response. In some cases the reaction was so defi-
nite and prompt as to leave the impression on the observer that
the creature was possessed of something akin to visual sensation.
At other times the reactions were indefinite, sluggish, variable,
and less convincing, though in the end resulting in the retraction
of the specimen as before.

The following experiment was made to determine the extent
of the sensory area, or in other words whether all portions of
the body were similarly responsive to light. A specimen which
had been quietly expanded on the surface of the sand for some
time, being one of those which had shown less aptitude for bur-
rowing, was so placed as to make it possible to reflect a harrow
ray of light upon sharply defined parts of the body or tentacles.
It was found that the oral region, including about one third of
the body, was distinctly more sensitive than was any other.
Light concentrated on the aboral portion seemed to have no
effect at all, or so slight as to be indistinguishable. The tentacles
were apparently less responsive than the immediately adjacent
oral part of the body. This is slightly different from the con-
dition found in Sagartia modesta, as will be noted later, and was
a matter of some surprise, since the pigmentation of the tips of
these organs might be thought to have some relation to sensory
functions.

In a general way these results confirm the histological studies

2/8 CHAS. W. HAKGITT.

of the Hertwigs (op. cit., p. 22), as to the distribution of the
sense cells in actinians. They also agree substantially with some
of their experimental observations as to the unequal distribution
of the sensory areas, though on this point they gave slight atten-
tion to the effect of light as a stimulus (ibid., p. 190).

Sagartia inodcsta. - This anemone has much in common
with the former species. It is a creature having its habitat in the
sand just below or near low tide line. Like the former it takes
somewhat readily to the artificial environment of the aquarium,
though seems somewhat less hardy under these conditions. I
first studied this species in its native haunt, having found several
specimens on an accessible beach. I first found them just before
twilight, and in the shadow of a large boulder which still further
reduced the light, with the tentacles extended very much as in
the case of Eloactis ; the body was not protruded beyond the tube.
Going again in the brightness of early morning they were not to
be seen, no sign of tentacles even in the partially closed burrow.
I made these observations several times, and concluded that they
were probably also nocturnal.

Specimens were collected and taken to the laboratory and
placed in the same general conditions as were the former species.
Experiments similar to the former were performed, but with
much less promptness or clearness in reactions. Placed under
the dark chamber there was not the ready extension of the body
as in Eloactis. Further, on removal of the hood the response
was much less sharp and convincing, though quite evident.
Placed on a table upon which a beam of sunlight could be
reflected it was found when the ray was reflected upon the
numerous tentacles that there was immediate reaction. It should
be stated that in this species the tentacles are very numerous,
even a hundred or more, and form a dense crown in expansion
covering the oral region like an umbrella, while in the former
species these organs are but twenty in number and rather short.
In Scgartia the tentacles seemed more sensitive than in the
former species, or than the oral region, but this may be due in
some measure to their numbers, and to the general relations they
sustain to the oral portion of the body, especially the region
just below the tentacles. Still the results agree again with the

BEHAVIOR OF SEA-ANEMONES. 2/9

views of the Hertwigs, as expressed in the following words :
" Die Sinneszellen finden sich im Ektoderm der Mundscheibe
und der Tentakeln, wie uns schien, liberal 1 ziemlich gleichmassig
vor ; nur an der Spitze der Tentakeln mochten sie vielleicht in
grosserer Anzahl vorhanden sein " (op. cit., p. 22).

Similar experiments were made on three other species of
anemones, namely, Sagartia Icncolcna, Sagartia Inche, and
Mctridiuin inarginatnin. These species are all more or less free,
and variable as to habitat. The first, S. leucolcna, is fairly com-
mon at various points along the shore-lines of the region of
Vineyard Sound and southward. Its usual habitat seems to be
under rocks near low tide, though taken also on the piles of
docks. It seems to seek the under sides of rocks, or settles
among masses of Molgula, sponges, etc., on piles, thus more or
less secluded, and seldom seen by the casual observer.

On the other hand, v$. lucice seems to be equally at home almost
anywhere in shallower pools, on fucus, piles of docks, etc., some-
times in shaded places, but oftener in the open sunlight on rocks,
fucus, etc. About the same may be said of Metridium. While
more common from deeper water than either 'of the others, it is
yet quite common just below tide line on rocks, piles, etc.

The experiments on these species were made under the en-
vironment of the aquarium, but were sufficiently varied to give
fairly satisfactory tests as to their reactions to this class of stimuli.

From what has been said as to the habitat just given it might
be inferred that Sagartia leucolena would prove the more respon-
sive to the tests, and such was found to be the case without ex-
ception, though as in the former cases, with considerable individual
differences.

Verrill long ago pointed out that this species was more active
when in dimly lighted aquaria, or at night. However, I have not
found that specimens in the general light of the laboratory
showed any very evident light reactions. But when an aquarium
was placed in direct sunlight there was an almost uniform attempt
on the part of specimens to escape from the direct rays. As a
rule this was done by slowly creeping over the edge of the stone
or shell into a less exposed position. Specimens which were in
glass jars, and attached to the sides or bottoms of the jars, when

28O CHAS. W. HARGITT.

brought into direct sunlight soon closed up entirely, withdrawing
even the tentacles, and assuming a more or less hemispherical
shape. Taken from the direct light into the diffused light of the
room they promptly expanded and remained so until again
placed in the sunlight. This experiment was repeated again and
again, and with substantially the same results. It was also found
that the degree of contraction was very closely an expression of
the degree of light intensity.

Many specimens were brought to the laboratory adhering to
small rock fragments, bits of shells, etc. In a few cases when
such specimens were exposed to direct light they would creep
over to the shaded side of the rock, and during the night return
to apparently the exact spot previously occupied. This might be
taken to suggest some such sense of position as is known to be
had by certain gasteropods ; but the tests were not sufficiently
numerous nor constant to warrant any definite statement.

With Sagarlia lucice and Metridimn the case was very different,
as might be expected. Specimens of these anemones placed
under the same conditions as the former, indeed in many cases
when occupying the same aquarium, were found to be almost
without exception, quite indifferent to light. Placed for some
time under a dark hood and suddenly exposed to direct sunlight
there was not the slightest evidence that there was any sense of
the change. The experiment was made in various ways. Some-
times as just suggested. Again, a beam of strong light was
reflected directly on the specimen as it was quietly expanded on
the table, but so far as >S. Incice was concerned, always and with-
out exception, with negative results. Occasionally, though
always doubtfully, Metridinm would show some slight sensory
movements of the tentacles. But specimens have been subjected
to the reflection of a strong beam of light directly upon the oral
surface for ten minutes at a time without the slightest response.

I have had a few specimens of Edivardsia clcgaiis in the
aquarium but for some reasons they did not seem at ease under
these conditions, and exhibited no distinct evidence of any photic
sensibility. I have seen but once any living specimens of Ccri-
antJins at Woods Holl and then only under circumstances which
rendered any observations impracticable. I regret therefore,

BEHAVIOR OF SEA-ANEMONES. 28 I

not to have been able to test the sensory behavior of these
species.

The only other aspects of behavior which have been observed
are those of feeding, and the very variable reactions concerned in
tube-building.

Concerning the former my first experiments were made several
years since. At that time I tested their feeding propensities by
trying in various ways to induce them to take food. At various
times during their aquarium life I tried to feed the creatures with
bits of crab meat, bits of fish, clam, etc., but in no case was I
able to induce the creatures to take the bait. During the pres-
ent summer I observed that specimens of Eloactis which had been
dug up and placed in a pail along with specimens of Balanoglossus
were found devouring the latter alive. This was so unlike the
former behavior that one was tempted to wonder whether they
might have peculiarities of diet, and that their habitat on these
sand flats, where likewise Balanoglossus has its home, might
sustain some relation thereto. I therefore repeated the former
experiment of offering them shreds of crab and fish meat and
with the same negative results. I then tested them again with
the Balanoglossns and found that it was taken quite readily by
the same specimens which had refused the other bait. Leaving
them for several days they were again tested with the same foods
and with the same negative results. Having no specimens of
Balanoglossns at hand some annelids, Hydroides, were offered
them alive, and they were readily taken by three out of four
tested.

No further qualitative tests were made along this line, but it
would seem as if they were rather partial as to feeding habits, and
particularly as to whether it be living or otherwise.

Limited tests were made as to their reaction to such substances
as blood of crabs, clams, etc., but there seemed hardly any defi-
nite reactions indicative of olfactory, or gustatory sensibility.
The swallowing reaction of Eloactis is much as in other species
of actinians, namely that it consists largely of oral efforts. The
tentacles play but little part in the reaction, though serving to
press the food down upon the oral margins or lips. The swal-
lowing act in these creatures involves something of a peristalsis of

282 CHAS. \V. HARGITT.

the esophagus. It was observed in several instances that any
considerable irritation of a specimen during the swallowing proc-
ess was almost invariably followed by a reversion and ejection
of the food. A worm three fourths swallowed would be ejected
by a sort of antiperistalsis, which was more rapid than the swal-
lowing had been.

From what has just been stated it need hardly be observed
that attempts to feed specimens with bits of blotting paper, or
other such materials, were uniformly negative in character.

The feeding experiments with other species were too limited to
justify any special attention in this connection. In most cases no
difficulty was encountered in inducing species of Sagartia to take
food of almost any sort.

Bitrroi^itig Reaction. - -Attention has been directed in an
earlier connection to the fact that considerable variability is evi-
dent among various specimens as to the matter of burrowing, or
tube-building. It may not be without some interest to briefly
cite a few details along this line. It is one of the curious features
in the activities of Eloactis that among a dozen specimens put into
an aquarium the most remarkable difference of behavior in this
respect may be seen. Most will show early signs of activity,
and soon bury their bodies as completely as possible, and assume
an erect position. Others appear to go through, the efforts but
in a most futile way. Left over night the aquarium will show in
the tracks over the surface of the sand the varied movements
made^ in this way. Still other specimens seem to show no effort
whatever to burrow, but lie indifferently upon the surface, hardly
showing signs of life except as they are stimulated by some
means. This may continue somewhat indefinitely. But after a
time a change may come over one of these sluggish specimens
and it sets about constructing a burrow all at once, as it were,
and within a night will have taken up the characteristic attitude
of its kind. If now it be dug out and left again upon the sand it
may promptly readjust itself again in a burrow, or it may remain
for some days in the same indifferent aspect. Specimens which
first bury themselves are usually prompt to build fresh burrows
if dug out of the earlier ones.

The facts herein portrayed suggest several interesting infer-
ences and inquiries by way of conclusion.

BEHAVIOR OF SEA-ANEMONES. 283

1. It seems clear that in the behavior of actinians toward light
one is forced to recognize that certain species have sensory per-
ceptions of photic stimuli quite as well defined as exist in such
organisms as the earthworm, clam, etc. And while in this group
of ccelenterates no such definite sensory organs are known as
those found in many medusae, the Hertwigs have described cer-
tain ectodermal cells which they have designated as sensory in
function. It is not without some warrant that we may conclude
that the various aspects of behavior under consideration are more
or less definitely correlated with sensory structures and perhaps
nerve cells.

2. Loeb, who has studied certain aspects of the behavior of
Ceriantlins ineinbrauaceiis (" Physiology of the Brain," pp. 5659),
attributes them to the influence of two tropic forces, namely,
geotropism and heliotropism. " Positive geotropism and positive
stereotropism cause the Cerianthi to burrow in the sand ver-
tically, and positive geotropism keeps them permanently in the
burrow."

I have elsewhere shown the inadequacy of this explanation as
applied to tube-dwelling annelids. I believe the facts under
review may likewise be better understood and more consistently
explained by other modes. Certainly the factor of light must be
reckoned with as potent in the behavior of the several species
studied. Again the variable behavior of these creatures in their
burrowing habits is not easily accounted for on the usual theory
of tropisms. Furthermore, it seems highly probable that in some
cases the food-taking habit may sustain a relation to the general
tube-dwelling habit.

3. Finally, as one considers the interesting facts as to the distri-
bution of these light-reacting anemones the foregoing inferences
are strongly corroborated. It is not necessary to review these
facts in detail. It will be recalled that the observations of Quat-
refages and Haime, already cited, had to do with species of Cerian-
tlins and Edzvardsia both of burrowing habit. Those of the Hert-
wigs were made on a species of Cladactis, an inhabitant of the deep
sea. The observations of Jourdan were made on a species of
Paractis, whose habit is not given, though species of this genus
taken by the Challenger Expedition were also from the deep sea.

284 CHAS. W. HARGITT.

Of the species which have come under my own observations as
light-perceptive, two are tube-dwelling, and one free-living, but
secreting itself u