Skip to main content

Full text of "The Early Development of a Polyclad, Planocera Inquilina Wh"

See other formats








Materials and Methods. 


The Living Eggs. 

General Account of tlie Embryology. 

The Unsegmented Egg. 

The First Cleavage. 

The Second Clea^'age. 

Segregation of the iu'toblast. 

Third Cleavage. 

Fourth Cleavage. 

Fifth Cleavage. 

Formation of the Fourth Quartet. 

History of the First Quartet. 

The Apical Cells. 

Further Divisions of the First Quartet. 

History of the Second and Tliird Quartets. 

History of the Fourth Quartet. 

The Mesoblast. 

The Entoblast. 



Literature List. 

Explanation of Plates. 

Tables of Cell Lineage. 

It gives me great pleasure to express my deep indebtedness to Prof. 
E. G. Conklin for his maiw suggestions and kind encouragement 
throughout this and other work. I also wish to acknowledge the 
many favors I have received from the University of Penns 3 dvania by 
which this work was made possible. ]\Iy thanks are also tendered to 
the donor of the University of Pennsylvania room at the ^larine 
Hiological Laboratory at Woods IIoll during the summer of 1906. 


Our knowledge of the development of the polyclad worms is based 
chiefly on the admirable works of Arnold Lang. In his extensive 
monograph of this class of the Turbellaria, published in 1SS4, Lang has 
not only summarized the results of all previous investigators, but has 
himself added very materially to our knowledge of their embryology. 




rrcvioiis students of i)()lyclad cml)rvolo»:y have shown the undoubt- 
edly spiral natur(‘ of the eleava.i^e up until a late .stage of segincntation. 
In this respect the development of these platodes corresponds closely 
with the cleavage of annelidan and inolluscan eggs. Wilson (92) says 
(p. 4o9): '‘Up to a late stage in the spiral period (twenty-eight cells) 
every individual blastomere and every cell division is represented by 
a corresponding blastomere and a corresponding cell division in the 
embryo of the polyclad and in that of the gasteropod.'' The same 
practically may be said of the annelid. 

This striking resemblance in the early ontogeny of three great groups 
of animals should not be without some significance. Yet, according 
to the accounts of these earlier investigators, the later history of the 
cells in the polyclad embryo differs very greatly from that of the 
apparently homologous cells in the annelids and mollusks. The 
difference is so marked that Conklin (97) has characterized it as ''very 
great, perhaps irreconcilable.’^ Wilson (92) has summarized this 
difference as follows (p. 411): "In the polyclad the first group of 
micromeres gives rise to the entire ectoblast, the second and third 
groups to the mesoblast, the macromeres to the entoblast. In the 
mollusk and annelid, on the other hand, the second and third groups of 
micromeres give rise to the ectoblast, like the first set, and the meso- 
derm arises subsequently.^' 

The formation of the ectoderm from the first quartet alone and of the 
mesoderm from the whole of the second and third quartets has been a 
serious stumbling block to those embryologists who have attempted to 
establish cellular homologies. Vfilson in a later j^aper (94) has cited 
this case as the climax in the contradictions of comparative embry- 
ology. A nunilxT of recent writers have expressed a doubt as to the 
correctness of Lang’s interjiretations. Thus i\Iead (97) writes (p. 
2S9): "I am not convinced that the cells described by Lang do give 
rise to the mesoderm, and 1 Iielieve it jiossible that the mesoderm of the 
polyclad is formed in the same manner and from exactly the same cell 
as in the annelids with unequal cleavage." 

This was the state of affairs until 189S when Prof. Wilson published 
in his paper on “Cell Lineage and Ancestral Reminiscences" some 
observations on a Pacific coast species of Leptoplana, Wilson's 
investigations showed that Lang was wrong in certain particulars, but 
did not fulfill .Mead's jirediction with regard to the mesoblast. Wilson 
found that each of the first three quartets of micromeres contributed 
to the formation of ectoderm, but also found that the mesoderm arose 
by inbiidding from cells of the second quartet and possibly some from 




the third. Wilson’s work was not a detailed study of the develo])- 
ment, and it leaves the impression that there may he something still 
undiscovered in the enilnyology of these interesting worms. Most of 
Wilson’s work and all of Lang’s was on the living egg. Every student 
of embryology knows how unsatisfactory this method is for the late 
stages of segmentation, unless checked by ])ro])erly fixed and stained 

In view of these facts it has seemed worth while to enter into a mor(‘ 
or less detailed account of the cell lineage of this form and gain, if 
possible, data to sup})ort or refute the theory of cellular homology. 
Besides, such data should throw some light on the phylogeny of this 
class of the Turbellaria. 

J^ang has given so complete a review of the literature on }M)lyclad 
embryology previous to 1884 that it would be mere repetition to go 
over that in detail here. The earliest investigators were Girard (184G- 
1854), Vailliant (18GG-1S68) and Keferstein (18G8). (3f these the wnvk 

of Keferstein was by far the best. Three other investigators bad 
studied polyclad development previous to 1884. The first of these 
was Ilallez (1878-1879). Hallez recognized but one quartet of micro- 
meres. From this the ectoderm arose. The macromeres liudded off 
four small cells at the oral pole which he believed formed the mesoderm. 
He descril)ed four other later buds at the oral pole which formed the 
wall of the gut. The work of Goette (1878-1882) and Selenka (1881) 
followed close on that oi Hallez. Goette also observed but one ({uartet 
of micromeres in Stj/Inchvs piUidium. When the ectoderm had 
reached the equator of the egg two to four small cells were formed at 
the oral pole (lower endoderm). Later the large macromeres l)udded 
large cells towards the aboral pole. These cells formed the upper 
endoderm. From the upper and lower endoderm the wall of tlie 
alimentary canal was formed, while the large macromeres *l)ecame 
food yolk. Goette found no mesoderm. 

Selenka (1881) determined that two ({uartets were given off. Ac- 
cording to him the first formed ectoderm and the second the mesoderm. 
Four small cells were formed at the oral ])ok‘ (lower endoderm of 
Goette), from which he believed the entire wall of the alimentary canal 
was formed. He found no upper endoderm. 

Lang (1884) found that three? quartets of micromeres were foi'nuHl. 
As already stated he believed the first formed the ectoderm, the second 
and third the mesoderm, t'rom the large macromeres four small cells 
were formed at the oral pole (lower endoderm). Then each of the large 
macromeres (fourth c[uartet) budded towards the aboral ])ole an upper 

1907 .] 



<Mi(Iod('nu coll, as (loolt(' hud found. Ih^ (l(’riv(‘(l tlu* alinioiitary canal 
from tho ii])por and lower ondod(M’in (‘(‘ll.s, while llie macroiiKU'cs (middle 
endoderm) broke uj) into food yolk and w('re al).sorI)ed by the other 
cells. All of the last four writers found that the ])osterior macromcre 
behaved differently from the other three in that it divided with its 
nuclear s])indle lyin<»‘ horizontally, thus ;:!;i\'i!ig I’ise to fi\'o macromercs 
(fourth (piartet). 

Material and Mititiods. 

The following paj)er is based ui>oii the study of a species of polyclad, 
Planoa m inquilina, described by Ih-of. Wheeler in 1S94. The material 
was obtained at the Marine Hiolo^-ical Laboratory at Woods lIol(\ 
.Massachusetts, during the months of July and August of DOG. 

Planoccm iuqiiilina is ])eculiar among ])olyclads in that it leads an 
apj^areiitly parasitic life. Tliese worms are found in the branchial 
chamber of the large whelk, Sijcotj/])iis c(ni(dicul(itiis Gill. As Wheeler 
suggests, it seems jirobable that they live on the excretory or waste 
})roducts of this gasteroj)od. Xo evidences that they feed on the 
tissues of the host have been found. The adult worms were obtained 
in considerable abundance, averaging about three worms for ever}^ 
whelk opened. The adult polyclads were transferred to dishes of sea 
water, in which the water was changed by means of a s}'stem of balanced 
syphons. These syphons served to keep the water free from sand and 
dirt, and also j^revented the overflow of the water and the escu])e of the 
worms. The animals usually laid c'ggs soon after l)eing brought into 
the laboratory. As described by Wheeler, the eggs are laid in spiral, 
gelatinous capsules containing aii}'where from 100 to 2,000 eggs each. 
Kiwh egg is surrounded l)y a separate egg membrane and the whole is 
iml)edded in the capsule material. In many cases two eggs are 
deposited in a single egg membrane, l^oth of which develop normally. 
'This is the usual way in which polyclad eggs are deposited. The tough 
capsule is difhcult to penetrate with fixing and staining reagents. 
Tins fact no doul)t is one of the chief reasons why so few embryologists 
have worked on polyclad development. The eggs of Planoccra 
inquilina seem more favorable in this respect, and with care it is possible 
to get very good preparations. The egg capsules were deposited 
against the sides of the dishes, and it was necessary to cut them away 
with a scalpel. 

Wheeler (94) did not succeed in getting the eggs of this species to 
develop under laboratory conditions. On the contraiy 1 experienced 
no difliculty of this kind. Stages from the maturation to the free 




swimming, ^Hiller's larvce were ol^tained without difficulty. I do not 
knoM^ wherein my methods differed from Wheeler’s. He suggested 
that the water in the laboratory was too warm. However I made no 
attempt to keep it cool, and in some cases the sun shone directly on the 
dishes without apparently affecting the eggs. The adult animals 
however would live for only a few da}^s. After the first day they 
became very sluggish and their bodies l^egan to break up in a manner 
similar to that described b}" Wheeler. 

Although I studied the question a good deal, I have never been able 
to ascertain where the eggs of this pol}^clad are laid under natural 
conditions. The animals were laying throughout the entire summer 
from June to September, yet I have never found a single capsule except 
when deposited in my dishes. I have repeatedly searched the interior 
of the branchial chamljer of the whelks in which adult worms were 
found, but to no avail. I have also examined carefully the shells of 
these gasteropods, l)oth inside and out, but no evidence of egg capsules 
was found. I found that the worms always laid soon after being 
removed from the whelk to the dishes of sea water, and it is possible 
that this is the normal stimulus to egg deposition. If such is the case 
the adult animals must deposit their eggs on stones or other smooth 
objects on the bottom. In such a case both adult and young would 
have to run the risk of again finding a Sycotj/pus and entering its 
branchial chamber. The risk seems to be consideralde, and the number 
of eggs deposited by an individual is perhaps hardly sufficient to 
warrant such an hypothesis. 

The early divisions up to the forty-eight- or fifty-cell stage were 
followed and figured in the living egg. The eggs are rather opaque 
and it is difficult to be certain concerning main^ of the divisions. This 
whole portion of the cell lineage was later gone over in the stained 
preparations and the previous obseiu^ations on the living material 
were verified or corrected. 

Eggs were fixed in various solutions, among which were sublimate 
acetic both aqueous and in 95 per cent, alcohol, (bison’s murcuro- 
nitric, picro-sulphuric, picro-acetic, Perenne 3 *’s fluid and Flemming’s 
solution. Of these the sublimate-acetic mixtures and Gilson’s fluid 
proved most valualDle. For staining whole mounts Conklin’s (02) 
picro-haanatoxlyn was used. Slightly stronger Inematoxlyn than 
recommended by Conklin was found better for these particular eggs. 
The eggs were then cleared in xylol and mounted in l^alsam. It was 
impossible on account of their small size to remove the eggs from the 
capsules, but it was found that they cleared l)etter if the capsules were 
torn into small i)ieces. 




In studying the cell linea^^e the chi(‘f (liflicnlt}^ experienced was in 
not beiniz: able to rotate tlie e^^s under the cover ^lass. The eg^^s are 
not orientated in any definite direction within the capsules, and it was 
necessary to ]uck out for study those eggs which were favorably 
oriented. Besides it is ])articularly diHicult to determine the lineage 
of certain cells if one is able to view them from one side only. The 
fact that the eggs could not be rotated accounts for some of the draw- 
ings being from a somewhat oblirpie view. 

The results obtained from studying the whole mounts were checked 
as far as possible b}" the use of serial sections. It was found necessary 
to bleach the Flemming material with peroxide of hydrogen before 
sectioning. A number of stains were used for the sections, but JJehi- 
fiekFs iLTinatoxlyn, either in toto or on the slide, proved most useful. 
A combination of thionin and acid fuchsin also gave good results. 
There is too much yolk in these eggs to use Ilaidenhaiids iroii-alum- 
hiematoxlyn to advantage. 


The S}'stem of nomenclature followed in the cell lineage of this i)aj)er 
is that used by Chabry (87), Wilson (92) and Conklin (97), with slight 
modifications. This system is the same as has been used by Child 
(1900), Treadwell (1901), Casteel (1904), Nelson (1904), and many 
others. For the sake of convenience the chief points are repeated here. 
Each of the four ciuadrants of the egg is denoted by one of the first 
four letters of the alphabet. Tlie left quadrant is A, the anterior i>, 
the right C, and the posterior D. The four macromeres form the basal 
quartet; the first group of four micromeres to be separated from these 
is the first quartet, and so on. A micromere is denoted by a lower case 
letter, while the capital letters are reserved foi* the corresponding 
macromeres. The number of the quartet to which a micromere 
belongs is indicated by a coeflicient, while the cell generation is shown 
by the exjionent. Of the two cells of any division of a micromere 
(except 4d), the one lying nearer the animal pole is regarded as the 
stem cell and recei\TS the smaller exponent. Thus 2ab a cell of the 
second quartet in the A quadrant, will divide into 2a^ ^ and 2a^-^. 
lies nearer the animal pole than 2rd “, will divide into 2r/^-^-^ and 

2a' In the case of the divisions of the mesentoblast, 4d, the lower 
cell is regarded as the stem cell and receives the smaller exponent 
(Conklin, 97). 

A division is to the right, dexiotropic, if the upper cell lies to the 
right of the lower when viewed by an imaginary observer sit\iated at 




the animal j)ol(‘ and facing the cell in ([iiestion. If the npj^er cell is to 
the left of the lower the division is heotropic (Lillie, 95). If the s})indle 
is horizontal, i.c., the cleavage meridional, the cell to the right receives 
the smaller ex|)onent, 

FolIoAving Child (1900), the macromeres receive the coeflicient of the 
quartet to which they last contributed. 'Hius 2.1 gave rise at its last 
division to Furtlier details of the system will become evident 
reference to the tables of cell lineage and to the figur(\s. 

The Limx(; Fee. 

The living egg of Planocera inquiUna consists of a uniformly dense 
mass of granules which vary only slightly in size. Between these 
granules is a light coloretl fluid substance. When the one-celled egg 
is strongly centrifuged for some time the yolk granules are compacted 
to one side and a cap nf the light colored fluid, in Avhich are only a few 
granules, lies at the opposite side. This cap of fluid occupies ])erha))s 
one-fourth or a little less of the entire egg. When the egg is crushed 
under a cover glass and examined with an immersion lens, minute' 
bodies (microsomes?) are found in the fluid ])ortiou. These small 
bodies exhil)it a constant “l)rowniaiC movement. When the egg is 
entire, however, no motion of any kind can be discerned. 

The eggs of this sj^ecies of Poly clad ap))ear ]jerfectly uniform through- 
out. Selenka, Cloette ami Lang have found that in many polyclad 
eggs there is a darker inner ])ortion and a lighter outer layer to the 
eggs. 1 could make out no such differentiation in these eggs, liang 
(lid not find this separation of sul)stance in Discocalis tiqn'uu. 

Considerable time was spent during the summer attemjTing to 
exi)eriment on these eggs. But in all cases where the conditions were 
varied from the normal the egg died in a short time and no results were 

All the early cleavages as well as the maturation divisions occur at 
intervals of about one hour. The whole develojunent ])roceeds rather 
more ra])idly than in most ])olyclads. At the end of the second day 
or at the beginning of the third the embryo is conqdetc'ly covered by 
the small ectodermal cells. Cilia soon begin to form on these and by 
the third da}^ the embryo begins to slowly rotate within the capsule. 
During the next day or two the cilia become l)etter developed and the 
eml)ryo rotates faster and faster. The rotation takes place first in one 
direction and tlum after a short pause in another. Occasionally they 
c(‘ase moving for some time. The eye S]iots appear aliout the fourth 
day. During the fourth and fifth days a number of homogeneous 




\'olk splionilcs (‘nn Ix' soon inside' tli<‘ embryo. I>y the end of th(‘ 
fiftli day the (*iliat(Ml proec'sse's eharaeleM'istic* of tlie Mullei’’.s larvm bo^in 
to appear and the einl)ryo (*xhibits frexpienl (‘ontraction.s of its l)ody. 
< )n tlie sixth day tlie larva' be^in to burst through the egg membraiu's 
and to swim about as typical Miilh'r’s larva'. I did not succeed in 
kee|)ing these larva' more tlian two or thre'e days, during which tini(‘ 
tlu'V scMMiK'd to undergo but little change'. 


l''()llowing the exam])leof many write'rs on embryology, it seems that 
the later detailed account can be made l)riefer and more readily 
undc'rstood if it is prefaced by a brief gc'iK'ral sketch of the deve'loj)- 
ment. The segnu'iitation of the ('gg is total and slightly tme([iial. 
I'roni the first two divisions four cells result, of which one, the posterior, 
is slightly the' largc'st. Three (piartets of micromeres are then given 
off in alternating dexiotropic' and heotropic directions. The large cells 
of the basal (piartet then l)ud off at their lower, vegatative pole four 
v(‘i’v small cells, which are to 1 h' rc'garded as the macromeres. Th(' 
large upjier cells of this division form the fourth (piartet. Tlu' large 
posterior cell of this (piartet, 4r/, lieliaves vc'ry diffen'iitly in its future 
divisions from the other three. We may designate tliis cell as the 
“ mesentoblast,'' following Conklin’s nomenclature. At the stage 
with forty-four cells this mesentolilast buds into the interior of the 
egg a large C('1I, 4(/“. Both of the mesentoblast cells then divide. In 
these divisions the nuclear spindles lie horizontally. From the lower 
pair of cells the greater part of the alimentary canal is derived. The 
upper pair probably contribute a small amount to the aliinentaiy 
canal, while the larger portion goes to form the UR'soderm of the body. 

In th(' later development tlie chief axis of the ('gg, i.c., tlie axis from 
the animal to the vegetative pole, becoim's bent, so that the animal 
pole conies to lie at the anterior end of the embryo. 

From the first (piartet arisevs tin' ectod(‘rm, covering the anterior 
and dorsal portions of the Ijody. From cells of this (juartet four 
strings of cells bud into the interioi* of the embryo and form the gang- 
lion. The e}a's arise in ectodermal c(‘lls of this (piartet. The’ second 
(piartet gives rise' to the larger portion of the ectoderm on the ventral 
and ])o.>terior regions of the body, f rom (*ells of tliis quartet is formed 
most of the ectc)dernial pharvnx. A jxirtion of the second (piartet is 
budded into the embryo and forms niesorlenn. From this source 
arises probably only that mesod(‘rm found around the blastopore and 
which is later concerned in tin' structure's of the pharynx. 




The third quartet consists of small cells from which apparently 
only ectoderm is derived. The individual divisions of these cells have 
not been traced very far, but there is every reason to believe that they 
form ectoderm onl 3 ^ 

The history of the fourth quartet is peculiar. As already stated,, 
the posterior cell 4d is the mesentoblast, from which the alimentarv 
canal and a portion of the mesoderm arises. The other three cells of 
the fourth quartet, 4a, 46 and 4c, do not divide as long as their history 
can be traced. They, however, break up into a large number of 
homogeneous yolk spheres which are absorbed b}^ the endoderm cells. 
The large nuclei of these three cells can be traced until the alimentar}^ 
canal is partly formed. 

The nuclei of the small macromeres show evidences of degeneration. 
These do not divide as long as they can be followed. The}" are carried 
into the embryo by the phar}"ngeal invagination of ectoderm, and it 
seems probable that the}" degenerate without gi^diig rise to an}" morpho- 
logical structure. 

The Unsegmented Egg. 

The unsegmented egg of Planocera inqiiilina is nearly spherical in 
shape and measures al^out one-tenth of a millimeter in diameter. In 
many cases, however, the eggs are pressed out of their normal shape 
by crowding within the capsule. The eggs when laid possess a large 
germinal vesicle which lies slightly to one side of the centre (fig. 1). 
This statement is not remarkable in itself, were it not for the fact that 
the eggs in the uterus of the adult possess a well-developed spindle 
with equatorial plate, centrosomes, etc. Apparently this spindle 
never goes farther than the equatorial plate stage and then degenerates 
so that the egg when deposited possesses a germinal vesicle. This 
phenomenon was first observed in certain polyclad eggs by Selenka 
(81d), and later WTeeler (94) has recorded it for this species. Gardiner 
(99) has studied this phenomenon in Polycherus candaius and concludes 
that the uterine spindle is due to abnormal conditions of the adult. I 
have not attempted to study this phenomenon in detail, but a casual 
survey shows that in animals which were fixed as soon as possible after 
removal from the whelk this uterine spindle was well developed. 
Since other animals from these same lots laid eggs which developed 
normally, one must conclude that if it is not a normal phenomenon it 
at least does not interfere with the later development. Selenka (81c/) 
suggests that this spindle is of use in bringing the yolk granules to 
the centre of the egg, In it, as Wheeler has noted, such could hardly l:>e 



the case here, since tlie distribution of the yolk granules is uniform 
throii^liont tlie e^^. 

Wheeler (91) has stated that the impregnation is probably what 
Whitman (!)l) has called ‘diypodcTinic.” J have several times 
observed animals ap{)arently in copulation. In this act the two 
animals remain in contact for some time and move about to,c;ether. 
Most fre((uently the ventral sides of the animals were in contact. 
Fertilization is m'cessarih^ internal, although the means b}^ which th(^ 
sperm roach the e^, 2 ;s is not known. 

Two maturation divisions occur after the e^^ is deposited. The 
first occurs about one hour after deposition and the second about one 
hour later. At each of these divisions the egg goes through some of the 
most remarkable contortions (fig. 2), The egg l:)ecomes veiy irregular, 
and i)rocesses occur from all sides. These processes consist of the more 
fluid substance and contain few yolk granules. In many cases parts 
of the egg are cut off entirely. In one case observed, the egg was 
actually cut into two, so that I first mistook it for a two-cell stage, but 
later these parts fused and it then underwent normal development. In 
some cases minute pieces seem to be cut off which do not fuse with the 
egg, for some time at least, but continue to float about between it and 
the egg membrane. 

The movement of the egg substance is very slow. It takes about 
twent}^ minutes for an egg to pass through such a contortion and 
regain its normal splierical shape. The phenomenon is the same at each 
of the two maturation divisions. Similar phenomena have been 
observed by Ilallez (79), Goette (S2) and Selenka (81) in other poh’clad 
eggs. Such amteboid movements are also quite common in the eggs 
of other animals, especiall}" annelids. 

The First Cleavaije. 

About one hour after the second maturation division the first cleav- 
age furrow makes its appearance. The spindle for the first cleavage 
lies near the centre of the egg. The first two blastomeres are not of 
equal size, although the difference is slight (fig. 4), In order to make 
certain that there is a recognizable difference, I have made a number 
of camera drawings both of living and stained eggs. In all cases where 
it was not evident that the egg was pre.ssed out of shape in its membrane 
the difference in size is quite easily recognized. 

According to J.ang (S-1) this difference in size of the first two blasto- 
meres is ver}" constant in ])ol}'clads. Lang sa}^s (p. 330): “Ich habe 
diese allerdings wenig auffallende Verschiedenheit in der Grosse 




der zwei ersteii Hlastoincren, die v^elcnka hei Th}'sanozoon niid Eur}-- 
lepta eoiistatirte, nicht mir bei Discocclis tigrina, sondern aucb bei 
alien I^seiidoceriden imd Kurylepta nachweison kbuncn. Icli glaiibe, 
(lass sie aiieli bei alien Leptoplaniden existirt, obsclioii sie bier scbwer 
nacbweisbar ist.’^ 

Since tbe polar bodies do not remain attached to tbe animal pole, 1 
have been unable to ascertain wlietber there is a rotation of tbe spheres, 
indicating that the first cleavage is spiral, as Conklin (97) lias shown 
for Crepidula. With the separation of the two cells their outlines 
tiecome more or less irregular. Esjiecially along their line of contact 
delicate protoplasmic processes extend outward. This is very miicli 
less marked than in the case of the maturation divisions, but there 
seems little doubt but that it is due to the same internal causes, what- 
e^^er those may be. The same phenomenon may also be observed in 
several of the next succeeding cleavages, i.e., at the four- or even 
eight-cell stage. But with each successive cleavage the processes are 
smaller and more delicate and the phenomenon less marked. 

The Second Cleavage — Two to Four Cells. 

A little less than an hour after the first two cells have separated a 
furrow appears in each. These furrows often appear simultaneously, 
but in many cases one cell begins to divide in advance of the other. 
Lang finds that in Discoca’lis this is always the larger. “ Die Theilung 
erfolgt aber nicht ganz gleichzeitig, die grdssere Furchungskugel theilt 
sich vielmehr etwas friiher als die kleinere.’' In Planoceni this suc- 
cession is not so marked as in Discoccelis. In many cases the two 
spindles are in the same phase at the same time. 

The cells resulting from this second cleavage are much more unequal 
in size than those of the first division. Each of the two cells buds off 
a smaller cell in a la^otropic direction (FI. XXX\^, figs. 5 and 6). As 
Lang has found in Discoadis, the two spindles of thisdi\usion cross at 
an angle. Viewed fi'om the side these spindles form an X (fig. 5). It 
thus happens that the two smaller cells lie at a higher level than the 
two larger. Tlie former tend to meet in a point at the animal }:)ole 
(fig. 6). Before the next division they move downwards a short 
distance, so that a portion of the first furr(nv is visible between them 
and forms a short polar furrow. Mewed from the vegetati^^e pole the 
two larger blastomeres always meet in a line the so-called vegetative 
polar furrow (‘Dh-echtmgslinie,^^ Raiiber (S2), or “ Querfurche,^’ Rabl 
(79)). As in the case all dextral mollusks and in annelids, this polar 
furrow turns to the right when viewed in the plane of the first cleavage. 


NATUUAI. SriKXCKS OK imiiladklphia. 


TIip two smaller hlastomeros. .1 and C, are ai)i)roximately equal in 
size (/!<>:. 0). while of the remainin'^ two, one is significantly lar» 2 ;er tlian 
the other. The lar»'er of these, i.c., the largest of the four cells, is the 
one denot('(l by /). and lies on the posterior side of the The next 

largest cell, li, is anterior, while the two smaller cells, A and C\ are 
lateral in j^osition. It thus happens that the lar^(‘r of the two hlasto- 
nieres in the two-cell sta^e « 2 ;ivcs rise to the jiosterior and ri^ht blasto- 
meres, D and C, while the smaller forms the anterior and left blasto- 
meres, /? and A, Thus the first cleava<:o se[)arates the posterior and 
riirht sides from the anterior and left. 

Tniun Clkavaok — Fouk to th(mT Celus. 

The third cleavage is strongly dexiotro})ic. The resulting cells, la, 
15, Ic and Id, come to lie. when completely sej^arated, in the furrows 
betwe(Mi the macromeres (fig. S), They retain this position only until 
after the next cleavage. As lias been found in the case of other spirally 
cleaving eggs, the ri'sulting jiosition of the blastomercs here is not due 
to surface tension alone. The s]iindles from the moment of the break- 
ing of the nuclear membranes indicate clearly the direction the cleavage 
is to take (comp. fig. 12). It is thus a jihenomenon inherent in the cell 
structure, although surface tension no doubt plays a part in gi\’ing the 
blastoiiKTes their final ]iosition. 

The divisions of the four blastomeres do not usually take ]ilace 
synchi'onously. Lang gi\'es th(' following rhythm for Disroralis. The 
largest cell (D) divides first, next the anterior large cell (/f). and aftc'i* 
these ha\’e divided the lateral cells A and C divide at about the same 
tini(‘. In PUnwcera inquiUna as a greneral rule the same rhythm holds 
true, although there are often exc('|)tions to it, d'he spindles are usually 
[iresent in all four of the cells at the same time, but those of the two 
larger cells are nearly always more acU’anced. Tlie fact that in the 
great majority of cases this rhythm holds true is exceedingly interesting. 
Lang (S4) was the first to ])oint out that there was such a constant 
rhythm in the divisions of embryonic cells. Lang found that not only 
did this succession hold good for divisions of the blastomeres where 
there is an actual differ(‘iic(' in size, but also that the descendants of 
these cells divided in the same <n‘d('r (see quotation p. 526). 

Since that time mant' other ol)servers have found a more or less 
similar and constant rhythm in various animals, r.//.. Lillie (95) in 
Unio, .lennings (9(i) in Afiphuiclnia , Child (00) in .-Irc^nVo/a, and Xelson 
(04) in Dinophiliis. In Cnio, Areuicola and Diuoplnluti the order is 
!), C, A, B] in Af^plnnrhna it is D, (\ B, A : while in Discocaiis and 




Planocera it is D, B (^1, C), the latter two cells dividing at about the 
same time. It is evident that this is correlated with the larger size of 
certain blastomeres. It is thus in contradiction to Balfour’s (SO) law, 
that a greater amount of 3mlk retards the rapidity of cleavage. Kofoid 
(94) has suggested that this difference in rapidity of cleavage is due to 
the greater absolute amount of protoplasm in the larger cells. 

In general this rhythm can be recognized for six or seven cleavages 
in Planocera j although, as stated, there are often exceptions to it. 

The cells of the first quartet are only slightly smaller than the cells 
of the basal quartet (fig. 8). The inequalit}^ while sufficient to be 
easil}^ recognized, is not so great in this species as in Discocodis and 
most other polyclads. Girard (54) has described the cleavage of 
Planocera elliptica as total and equal. In all others so far as known it 
is unequal. All the cells of the first quartet are approximately equal 
in size (fig. 8), notwithstanding the inequalit}^ of the cells from which 
they arose. Of these cells in Discocoelis Lang says (p. 331) : Anschein- 
end sind die vier Ur-Ectodermzellen [first quartet] gleich gross; es ist 
aber sehr leicht mfiglich, dass sie in Wirklichkeit ahnliche Grossen- 
unterschiede zeigen, wie die vier grossen Blastomeren .... weil sie 
bei ihren weiteren Theilungen ganz genau demselben Rhythmiis folgen, 
wie die vier grossen Blastomeren.” 

Fourth Cleavage — Eight to Sixteen Cells. 

The next cleavage is initiated by the division of the largest of the 
macromeres, ID. This division is followed veiy closely by the cleavage 
of ID (PI. XXXVI, fig. 9). Before the daughter cells 2d and 2b have 
been completely separated spindles appear in the two lateral cells lA 
and 1C. The formation of this second quartet takes place in a lieo tropic 
direction (fig. 9). By the separation of these cells, the cells of the 
first quartet are pushed towards the left, thus in the opposite direction 
from which they were given off. Vffiile the macromeres are dividing 
the first quartet begins to divide, also in a lieotropic direction (fig. 9). 
These cells divide very nearly synchronously, but in ver}^ nearl}^ all cases 
the spindle of the posterior cell Id is further advanced than that of the 
others (fig. 9). The result of this division of the first quartet is eight 
cells very nearly equal in size (fig. 10). When the divisions are 
completed the upper or apical cells have rotated almost 45 degrees to 
the left, so that they now occupy a position very nearly over their 
respective macromeres (comp, fig, 12). They would no doubt occupy 
such a position were it not for the inequality of the macromeres, which 




becomes more marked with the separation of each quartet. The 
distal cells of this division, viz., 16 ^^ Ir^ aiul l^P, correspond in their 
method and time of ori^dn to the ‘‘ Trochohlasts’' of annelids (Wilson, 
02) or the “Turret cells^' of mollusks (Conklin, 07). Their future 
divisions corresjioiid more closely perhai)s to the trochoblasts. Since, 
however, no structure directly homologous with the prototroch is 
reco^ 2 ;nizable in the polyclad larvie, we cannot say that these cells are 
homologous with those of annelids. 

llie cells of the second quartet are slightly smaller than the original 
cells of the first quartet. 

Fifth Cleavage — Sixteen to Tiiihty-two Cells. 

In the next cleavage each of the sixteen cells divides so that the 
ideal thirty-two-cell stage is realized. Again the cells of the posterior 
<iuadrant 7) divide slightly in advance of the others. The first cell to 
divide is the large macromere 2L), which sends off in a dexiotropic 
direction a small cell 3c/; 35 of the anterior cpiadrant is next separated 
off. The other two cells of the third quartet may not come off until 
after the cells of the first and second quartets have divided (fig. 14). 

The cells 2d and 2b begin to divide, shortly after the furrows for M 
and 35 are formed. The divisions of the second quartet are nearly 
meridional, but, as PI. XXX Vli, figs. 16 and 19 show, the right cell is 
slightly higher than the left, so that the division is dexiotropic. The 
lower distal cells, 2cr-2c/~, are the larger (fig. 16). Before this division 
has proceeded far the apical or stem cells of the first quartet divide in 
a dexiotropic manner (fig. 12). The distal cells, la^ “-lr/‘'“, are some- 
what smaller than the remaining apical cells (fig. 13). The former 
lie in the angles between the latter, and the cells la^-lc/-are pushed out 
until they lie opposite the ends of the apical cells. By this time the 
other two cells of the third quartet, 3a and 3c, have usually been 
constricted off (fig. 14). Xext the remaining cells of the first quartet, 
Xa^-UP (trochoblasts), divide in a dexiotropic manner. In this case 
the lower distal cells, are somewhat smaller than their stem 

cells (fig. 13). 

At this stage the cells of the third quartet, 3a-3r/, are the smallest 
cells present. The formation of this third quartet in polyclad eggs 
was first observed by Fang, the older workers having overlooked it. 
Lang calls this quartet the primitive mesoderm of the second order. 

The thirty-two cells are distributed as follows: 




First quartet 16 cells. 

Second qufirtet, 8 “ 

Third quartet, 4 “ 

Basal quartet, 4 “ 

From the above it is noticeable that the cells of the first (juartet 
show a tendency towards rapid division. In a number of inollusks, 
e.g., Neritina (Blochman, 81), Unio (Lillie, 9,5) , Crepidula (Conklin, 97) 
and Fiona (Casteel, 04), the first quartet has divided only once liefore 
the third quartet is formed. In Umbrella (Ileymens, 93) and Urofial- 
pinx (Conklin, 91) the first quartet does not divide at all until after the 
third is formed. On the other hand, in the polyclads so far studied 
(Lang, 84), in Ncries (Wilson, 92), Limax (Kofoid, 94), and in 
Dinophilus (Nelson, 04), the first quartet divides twice before or at the 
time the third is forming. As Conklin (97) has pointed out, this indicates 
the general rate of development of the upper hemisphere. In Plano- 
cern this is exceedingly rapid as its further history will show. 

Formation of the Fourth Quartet — Thirty-two to Forty Cells. 

After the thirty-two-cell stage the divisions Ijecome more or h'ss 
irregular; certain cells divide several times before others divide at alL 
After the completion of the thirty-two-cell stage the next cells to 
divide are the large basal quartet cells, 3D and 37L The nuclei of these 
cells have moved from the upper to the lower edge of cells, and when 
the spindle forms it reaches from the centre of the cell downward to its 
lower margin (figs. 15 and 16). These spindles are very nearly radial 
in position. Concerning them in Discocoelis Lang says (p. 335): “Nur 
sehr schwach ist die Abschniirung in der Richtung einer rechts gewun- 
denen Spirale (weim wiv das Ei so orientiren, <lass der orale Pol uiiten, 
der aborale oben liegt, und der Beol)achter in der Achse des lues steht) 
aiigedeutet.^^ By comparing figs. 20 (PL XXX\4I), 25 and 26 (I L 
XXXVIIl) and 30 (PL XXXIX), it will be seen that the small macro- 
meres show a twisting towards the left as seen from below, thus showing 
that the fourth quartet arose liy tlexiotro])ic cleavage. As may l)e 
seen from the sjiindles in fig. 15, the turning of the siiindle itself is very 
slight indeed. 

As ligs. 20, 25, 26 and 30 show, these lower cells are very small com- 
jiared to the c(4Is from Avhich thev arose. Wilson (98) has designated 
these small cells the macromeres, and the large cells from which they 
arose th(‘ fourth (piartet. As will become evident later on. 1 have- 

1907 .] 

XATCKAL srn:.\('i:s of piiiladklpiiia. 


additionnl (‘vid(‘ii(*(' which pcmits to this as th(‘ true interprotatioii. 
Althoui:;h the iianui “ inacronuM’c' ” is (‘vidciitly a inisnoiiier in this 
case*, it nevertheless seems W(‘ll to retain the name for these small 
cells. Wilson (98, p. 21) calls attention to the fact that these macro- 
meres ‘‘are ndatively not very much smaller than in some of tlie 
mollusks/’ for exanij)le Plnuorhis (Kabl, <S0). Jud^in^; from the 
tii^iires, howe\er, the relative size of the cells varies considei’ahly in 
different species of polyclads. In Planoccra iiiquilina these cells are 
relativ(‘ly very small as compared to the foiirtli (piartet. 

ilallez (79) first interpreted these small cells at the oral pole (macro- 
meres) as mesoderm. Selenka in his earlier work called them pharyn- 
geal cells, and in his later papers primitive endoderm cells. Goelte 
(82) and Lang (84) designate them as lower endodi'rm. and believe that 
they give rise to part of the alimentary canal. These four cells form 
one of the chief landmarks up until a very late stage of segmentation. 
At a time Avhen the ectoderm is well established in a la}'er, and just 
before it begins to invaginate at the lower pole to form the phar}uix, 
one can still make out these four cells. Lp until this time they have 
not divided. Their later historv is exceedingly diflicult to follow, i 
am inclined to the view that these <’ells do not take part in the forma- 
tion of any structures in the em1)ry(j of PUiuocera, but that they degen- 
erate and that their sul)stance is absorbed by the endoderm cells. The 
reasons for this view are, first, that they cannot be traced to any organ, 
and, secondly, that tlu' nuclei of these cells from the time of their 
formation until the last trace that can be found of them show increasing 
evidence of degeneration. Soon after these cells are formed it can l)e 
seen that the chromatin is n:iassed on one side of the nucleus in a dark 
staining mass (figs. 17 and 20, Id. XXXVl 1, and fig. 25, Tl. XXX\ 11,. 
etc.). It is this marked evidence of degenerating nuclei that enables 
one to follow these cells even until the ectoderm has reached this pole 
of the egg. In the other nuclei of the egg the chromatin is more or less 
evenly distributed in granules often along distinct linin threads. Wc 
thus have the remarkable and unique pheiiomena of the macromei'cs^^ 
deqeuerallnq u'iihoui (jiving vise to any pari of the embryo. Evidently 
the function of the macroiiK'res in this case has been taken over by the 
cells of the fourth quartet. These latter cells, as tlie figures show, are 
by far the largest cells in the embryo and contain the great bulk of the 
food yolk. In fact three of these cells, 4a, \b and -Ic, i)robably function 
(Mitirely as the bearers of yolk, and if they do take part in the formation 
of the alimentary canal it is only a minor ])art. The history of the 
other member of the fourth (piartet, 4d. is of esj^ecial interest and will 




bo dealt with later. From the fact that this cell gives rise to both 
eiidoderm and mesoderm we may call it the ^^MesentoblasV’ (Conklin, 

History of the First Quartet — Apical Cells. 

After the first two macromeres, AB and 4Z), are formed and before 
the other two cells, oA and 3C, have divided, spindles appear in the 
four stem cells of the first quartet, h The division of these 

cells takes place in a very marked Imotropic direction. The results of 
this division are four very small cells, lying at the proximal 

end of the spindle or just over the animal pole, and four much larger 
distal cells, As fig. IS, PL XXXVII, shows, the Iseotropic 

direction of the spindle is so marked that the cell comes to lie in 

front of and in front of and so on, so that there is a 

rotation of 45 degrees. 

These small cells at the animal pole have been observed by Selenka 
arid Lang in polyclads, and designated by Selenka as the apical or crown 
cells (Scheitelzellen). Selenka described these cells as arising from the 
stem cells of the first quartet, i,c,, as I have done. Lang, 

on the other hand, says that they arise from the cells which alternate 
with these stem cells, Lc., uCg, be^, ce^ and de^ of his system, or 
and Id^-^ of our system. He states that these cells 
send in processes between the stem cells, and from the ends of these 
processes the small apical cells arise. Lang describes this process in 
considerable detail, and while it is possible he is right for Discocodis 
tigrhia, I think it is very unlikely. 

Lang's observations were all made on the living eggs and not checked 
with stained material. 1 first observed these divisions in the living 
egg of Planoccra and came to the same conclusion as Lang, viz., that 
the apical cells came from the cells It was only later when 

I studied the preserved material of this stage that it became unmis- 
takably evident that it was the stem cells that were dividing 

(fig. 18). The appearance in the living egg is very deceptive owing 
to the great angle through which the spindles turn and to the small size 
of the resulting apical cells. 

These apical cells consist of a well-formed, normal-sized nucleus and 
a very small amount' of cytoplasmic material. In the stained eggs it 
is only seldom that one can see anything at all of the cell boundaries. 
Consequently in most of the drawings only the nuclei of these cells are 

^ In their method and time of origin and in their relative size these 




colls oorres])on(l closely with tli(' “npical rosotte^’ of annelids. In 
Ditiopldlus (Nelson, 01) th(‘S(' c(‘lls aris(‘ by exactly the same division 
as in Planoccra. In Xnris (Wilson, 02) tiny arise one division earlier, 
t.c., by the division of kd-ldh In many molliisks, cells of -this same 
lineage la* ‘ are not formed until much later in the develop- 

ment. c.f/., in ('rcpidula at the ei^»:hty-eight-cell sta^e, in Fiona at the 
sixty-four-cell staii:e, etc. The formaticni of these small apical cells in 
the ])olyclad embryo is an interesting and significant phenomenon, 
marking in another detail the close resemblance between the early 
cleavage of these platodes and that of annelids and mollusks. The 
fate of the apical rosette in the majority of annelids is the formation of 
an apical sense organ with an apical tuft of cilia. In Capitella (Eisig, 
98) and Dinophilus (Nelson, 04) the apical ])late does not bear a tuft 
of cilia. Conklin (97) finds that these cells form an apical sense plate 
in the Crepidula vcligcr, without, however, bearing a bunch of large 

As to the ultimate fate of these cells in Planoccra, I must again 
disagree with the previous students of polyclad embryology. Both 
Selenka (81) and bang (84) state that these cells sink down and are 
finally covered over by the other ectoderm cells. Thus Lang (p. 337) 
sa}"s: Hire Abschnurunggeschieht nicht ganz gegen den aboralen Pol 

zu, sondern etwas naeh innen, gegen die Eurchungshbhle, so dass sie, 
wie auch Selenka bemerkt, den Boden einer napfartigen Ahrtiefung 

am aboralen Pol bilden Da spiiter in der Niilie der Stelle, an 

der sich die Scheitelzellen gebildet haben, in besonderen Zellen des 
Ectoderms die Augen entstehen, so ware es moglich, dass aus ihnen 
Theile des Xervens}^stems, vielleicht der sensorielle Theile des Gehirns 
(obercs Schlundganglion?) entstanden.'' Again, I believe the decep- 
tive conditions of the living egg have led both Selenka and Lang 
astray. It is true that by reason of the small size of these apical cells 
the surrounding cells somewhat overtower them and form a “bowl-like 
depressioiP^ (IM. XXXVII, fig. 21). However 1 do not hud, as Lang 
states, that by the further division of the first quartet cells the apical 
cells are covered over and so sink beneath the surface of the ectoderm. 
In focusing on an egg from the animal pole in stages shown in PL 
XXXVIII, figs. 22, 23 and 24, the first nuclei to come into view are 
those of the ajiical cells. Furthermore these cells tend to move out 
from the animal j)ole and remain on the surface (figs. 24, 29, 31). In 
the later stages of segmentation they cannot be definitely distinguished 
from other cells of the same size which come to lie near them. The 
further division of the ectoderm cells in this region do not show any 




indications of covering over these cells. 1 believe that these apical 
cells form a part of the ectodermal covering of this region of the 
embiyo. And furthermore I believe that these are the first of a 
series of divisions occurring from nowon in all the ectodermal cells, by 
which a small epithelial cell is cut off towards the exterior and a larger 
cell remains somewhat deeper down. It is l)v this kind of divisions 
that nearly all of the later ectodermal layer is formed. These divisions 
will be discussed farther on in this paper. 

AMiether these apical cells form a definite sense organ or not I am 
unable to say. It is in this region, as Lang points out, that the eyes 
arise and that later the tentacles of the adult appear. However, the 
exact fate of these cells is mere speculation, since I have found it 
impossible to distinguish them in the late segmentation. 

Further History of First Quartet. 

We have so far followed the divisions of the first quartet until it is 
composed of twenty cells, and to the time when there are forty cells in 
the embryo. The last cells to be formed were the small apical cells, 
^ b Very soon after this the cells divide in a 

dexiotropic manner into two almost equal cells. The lower cells, 
j^p. 2 . 2 _i^i. 2 . 2 ^ appear slightly smaller than the upper cells (figs. 21, 22). 
Next, and sometimes coincident with the last division, spindles appear 
in the large cells These divide in a dexiotropic direction 

into two very unequal cells (figs. 22, 23). The upper cells, 

are much the smaller and lie on the surface of the egg between 
the cells (fig. 24). While this division is occurring the 

cells are dividing in a laeotropic direction (figs. 22, 23). 

Shortly after these divisions are completed the cells divide 

dexiotropically (fig. 24). In this case the lower cells, 
form very small cells which lie on the surface of the egg (figs. 24, 27). 

At this time there are thirty-six cells in the first quartet and about 
sixty-six cells in the entire embryo. This fact indicates clearly the 
very rapid development of the upper liemisphere. In Crepidula at 
the sixty-eight-cell stage there are only sixteen cells in the first quartet. 
In Fiona at a similar stage there are twenty cells in this quartet. In 
Nereis at the fifty-eight-cell stage there are thirt 3 ^-two cells in the 
first quartet, and a similar number inDinophilus at the sixty-five-cell 

Soon after the last division the cells divide again in 

a slightly dexiotropic direction (fig. 28). This time the distal cells, 
^^p.i. 2 . 2 . 2 _j^^i.i. 2 . 2.2 smaller and lie opposite the cells from which 

11 ) 07 .] 

XATUUAL S(’li:xci:s OF I’lIILADFfvPfllA. 


tlu'V liavo l)cen derivc'O (IM. XX A' IX, 29 , 31 ). At this time tliere 
arc tlirce circles of four small c*clls each, lyiu^i; on tli(‘ surface of the c^g 
and arranged around the animal i)olo (figs. 29 , 31 ). These twelve cells 
have been derived by three successive' divisions of the large stem cells 
of the first quartet, l(d-^-ld‘ h The inner circle consists of the small 
apical cells already described. The middle circle consists of the cells 
l^p.i.2.i_lji.i.2.q ^vere derived next and are the largest of these 

twelve cells. The outer circle was derived at the last division described. 
At the next division of these stem cells four more small cells are cut off 
to the exterior. 

At the same time that the last division described is occurring, the 
cells ‘ divide in a heotropic direction (fig. 28 ) into nearly 

equal moieties. AVith the completion of the above divisions there are 
forty-four cells in the first quartet, eleven in each quadrant. After a 
short resting period the cells ji.^ which have been so 

actively dividing, prepare to divide again. This time they bud into the 
interior of the egg four comparatively large cells, 1.2.2. 1.2^ 

the primitive ganglion cells. The outer smaller cells of this division. 
-|^p. 1.2.2. 1.2.2. 1.1^ form the fourth circle of small cells about the 
animal pole. This is as far as 1 have been able to follow accurately the 
divisions of these cells, and it is possible that the cells here designated 
primitive ganglion cells may still bud one or two generations of ecto- 
derm cells to the exterior. The four cells which arc to form the gang- 
lion divide repeatedl}\ Their individual divisions have not been 
traced, but four strings of cells can be distinguished for some time, each 
of which is the result of the subdivision of one of these primitive 
ganglion cells. These ganglion cells lie at first just above the meso- 
derm cells, and “ if i^ extremely difiicult to differentiate 
the later divisions of these cells. 

In the divisions of this first cpiartet we hav(' had a number of ex- 
amples of the process already alluded to, in which a very small cell is 
budded to the exterior of the egg and partially covers the larger, deeper 
lying moiety. Other examples of this same phenomenon will be 
described in the histoiy of the second quartet. The four sets of four 
small cells each which, beginning with the apical cells, are budded off 
in rapid succession from the large stem cells of the first quartet 
1 d}-^, are among the more striking examples of this phenomenon. These 
sixteen cells (and possibly more) very ncarh^ cover the aboral surface 
of the egg, and by their further divisions the ectoderm of this region is 
formed. Other cells of this first quartet show the same process more 
or less strikingly. For instance, by the divisions of a very 




small cell is cut off at the lower edge (figs. 24, 27). The same is true of 
the divisions of On the other hand the cells 

do not show snch an unequal division. These latter cells have been 
traced accurately through only three or four divisions, but the resulting 
cells are all more or less equal in size and apparently will enter directly 
into the formation of the ectoderm. This latter process, f.e., the equal 
division of the ectoderm cells, is much more in accord with what has 
been found in annelids and mollusks than is the former. 

History of the Second and Third Quartets. 

Just after the formation of the apical cells, and at about the time the 
large cells 3A and 3C are dividing, the larger cells of the second quartet, 
2a^-2iP, divide in a nearl}^ vertical direction (fig. 17). In many cases 
the spindles show a slight dexiotropic turn. Since the resulting distal 
cells 2a^-^-2r/“'2 come to lie in the furrows between the large cells of the 
fourth quartet, they are somewhat shifted in position to accommodate 
themselves to the inequalities of these large cells. 

It thus hai)pens that after the cells are separated some may be to the 
right of the upper cell and others to their left. The division is always 
nearly radial, but is still to be considered as belonging to the series of 
spiral divisions. 

With this we have traced the divisions of the second quartet until 
there are twelve cells present, three in each quadrant and all on the 
surface of the egg. 

Just after the mesentoblast cell has divided for the first time the cells 
2a^-2d} divide almost vertically but in a slightly la^otropic direction. 
The lower or distal cells, 2(d*^-2d^-^, are very much smaller than the 
upper ones (fig. 21). The former lie on the surface of the latter. This 
is another examjde of the cutting off of a small cell to the exterior. 
At about this time, sometimes before and sometimes afterwards, the 
small cells 2cr-^-2d“ ", lying in the furrows between the four primary 
cells of the fourth quartet, divide. The direction of the nuclear spindle 
is nearh^ vertical (figs. 26, 27). The two cells seem nearly equal in size. 
The lower cells, 2a^’^’^-2fP-“-q reach almost to tlie u])per edges of the 
macromeres (figs. 26, 30). The descendants of these cells are the first 
to reach the lower ])ole of the egg, and form without a doubt a consider- 
able ])ortion of the ectodermal pharynx. 

Shortly after these two divisions the large cells of the second quartet, 
2a?'^~2ip-^, divide in a dexiotrojiic manner, the two cells varying only 
slightly in size (figs. 25, 27). At this time there are twenty-four cells 
in the second (piartet, six in each (luadrant, and about seventy-eight 

1907 .] 



cells in the entire embryo. So far all the (*(‘lls of this quartet lie on the 
surface of the e<xg. At the next division in tliis quartet liowever the 
cells (*nt off a small cell to the exterior and the major portion 

of the larger cell pushes inwards (fi^s. 30, 32), so that they become 
almost covered by the surrounding cells, rnfortunately t have not 
been able to follow (he further divisions of these cells with certainty. 
At a later stage one finds one or two more small cells lying over these 
larger, deeper lying ones. For this reason 1 am led to suspect that these 
cells bud off one or two more ectoderm cells. The major portions of 
these four cells remain on the interior of the egg and form a portion of 
the mesoderm. Since these cells when first budded in are well towards 
the lower (oral) side of the egg, it is very probabh' that they form a 
portion of the mesoderm around the blastopore. In later stages (Id. 
XL, figs. 36-39) a considerable amount of mesoderm is found in this 
region. This later supplies the muscles and other mesodermal 
structures connected with the ])harvnx. The other mesoderm cells 
found just beneath the ectoderm in these stages are derived from the 
divisions of 4d, as will be described shortly. 

This account of the second quartet agrees closely in its main features 
with that of Wilson (98) for Lc])toplana. Wilson, however, believed 
that these second quartet cells, on the interior, multiplied rapidly and 
formed the entire mesoderm of the body. The development of the 
remainder of tlie mesoderm will be dealt with farther on, and it need 
only be pointed out here that this account confirms Wilson^s with 
regard to ectoderm arising from the second quartet. In contradiction 
to Lang (84), who believes that the whole of the second and third 
quartets formed mesoderm, we find here only a relatively small portion 
of the second quartet budding to the interior. By far the greater bulk 
of this quartet is ectodermal. 

With regard to the third quartet, this is in all ])robability entirely 
ectodermal. The cells of this quartet when first formed are relatively 
very small. These cells divide in a nearly radial direction (slight!}" 
dexiotropic) at about the seventy-four-cell stage (fig. 26). Further 
di\'isions of this (piartet have not been traced accurately. At later 
stages, howev(‘r, one or both of tluise cells in each quadrant have 
divided and all their ])rogeny remain on the surface. There is no 
indication that any of th(‘se c(‘lls pass to the interior. Their small 
size and epithelial character indicate that they are purely ectodermal. 

History of the ForuTii Quartet — The MEsemLAST. 

Shortly after the a])ical cells are formed and the cells 2cr-2(P have 

t ^ t. 

■ " " — ^ 


- ir^\ 

r** rz. "ZL-^ ~ •' c z;»C'_- 

£>r r^~r~ 

.:l ltc f.*r:-» _ 4 
> izi iz-.zi; -- 

53 S 



and 4 c are shown in dotted outline. Just posterior to the nucleus of 
4b lie the two small cells 4cP'^-^ and with the darker staining 

nuclei. Just })Osterior to these cells are the slightly larger cells 4(^2. 1.2.1 
and “ mentioned above. Finally, extending dorsally and anter- 
iorly from each of these are two cells which represent the results of the 
division just mentioned and for which the spindles are shown in figure 
34 . The resemblance to the so-called ‘Tnesoblast bands” of annelids 
and mollusks is, I think, evident. In many cases, however, the 
^^band” formation is not so marked as in this egg, which was chosen 
for drawing on account of its regularity. In many cases the meso- 


Fig. 2. — Schematic optical section of an egg viewed from the posterior side. Ta 
show the relation of the ectoderm, mesoderm and endoderm. Mes. 1, Meso- 
derm derived from 4d^. Mes. 2, Mesoderm derived from second quartet 
cells, ii'nd., Endoderm from 4dh e, Entoblasts from g. c., Pnmiti vt- 

ganglion cells. 

blast cells do not form such evident l)ands. Instead the nuclear 
spindles in the cells 4cP-i*2*2 and 4C/ varying angles. This is 

shown by the direction of the spindles in these cells in fig. 34 . The 
result tends to l)e a cluster of cells rather than symmetrical bands. 
The lineage of these cells becomes too involved to trace further with 
any accuracy. They lie well towards the dorsal side of the egg, and 
become more or less confused with the cells which are now budding 
in from the first quartet to form the ganglion. Nevertheless a fairl}^ 
definite group of cells can be found in this region during several of the 
succeeding stages. I have attempted to represent something of the 



relation of tlicse cells in the dia‘z;raniatic 0 ])tical section (text fi^. 2). 
Text 3 re|)resents an actual oj)ticaI section of a somewhat later 
sta^e. Here the large nuclei of 4a, 4h, and 4c are again shown, while 
lying just above them is a group of mesoderm c(‘Ils derived from the 
division of 4</^. The mesoderm now consist of a large number of cells. 
ITom this time on these begin to spread o\ii and gradually form a 
layer of cells lying just beneath the ectoderm. 

While I have not been able to follow the exact lineage of these cells, 
I think there is no doubt that the greater portion of the mesoderm of 
the body arises from the j)osterior mesoblasts, while as stated before 
only the mesodermal structures in the region of the pharynx arise 
from cells of the second quartet. 


nuclei of the four macromeres. Near the middle of the egg are the three 
large nuclei of 4a, and 4r. Above these are shown the nuclei (stipled) 
of a few of the nie.soderm cells derived from \d‘, I. Toward tlie vegeta- 
tive pole are mesoderm cells derived from tlu? second ([uartet, mes. 2. 

Text fig. 3 sliows a number of cells lying further towards the vegeta- 
tive pole than the derivatives of 4cF have yet reached. These cells 
undoubtedly arise from the further ])roliferation of the second quartet 
cells, which, as we have shown (p. 535), were budded into 

the interior of the embryo (cf. fig. 32). These second quartet cells, 
2a*-^-^-2td-^-^, when first formed lie considerably below the equator of 
the egg. With the overgrowtli of the ectoderm they are carried 




further clown to the region of the l)lasto})ore. Thus the large group of 
mesoderm cells found later around the pharynx arise, in large part at 
least, from the second quartet cells. It is, Innvever, quite possible that 
some of the derivatives of Adr later wander int(^ this region also. 

We have now followed the histor}" of tlie mesoderm from its origin 
until it forms a layer of cells about the emluyo just beneath the ecto- 
derm. We may leave the discussion and comparison with other forms 
until after we ha^'e traced the history of the endoderm. 

The Exto blast. 

The alimentaiy canal of the ]3olyclads has l)een derived in various 
wa}^s by different investigators. Vaillant (6S) and Keferstein (68) 
observed that at the time the embiyo began to rotate within its mem- 
branes there was a mass of roundish fat-like splieres on the interior. 
Ilallez (78 and 79) observed drops of an egg- white-like substance on 
the inside of the embryo, surrounded by a one-cell layer of the alimen- 
tary canal. The cellular elements of the canal he derived rather doubt- 
fully from the four small cells at the oral pole (macromeres). Seleiika 
(Sic) came to a similar conclusion, he., that the entire canal arose from 
tlie four small ‘^primitive endoderm cells’^ (macromeres). These were 
carried into the eml)ryo by the [diaryiigeal invagination and rapidlv 
spread over the large yolk cells (fourth cpiartet). According to this 
investigator these yolk cells break up without nuclear division into a 
large number of yolk spherules and serve as food for the developing 
endoderm. They give rise to no morphological structure of the embryo. 
Selenka (81) sa}^s that as soon as ‘Tlie Nahrungsdotterzelleii in ein 
Dutzend oder inehr ungleich grosse kernlose Kugeln zerfallen sind, 
beginnen die vier Ur-Eiitodermzellen (macromeres) ihre Theilung und 
Wanderung. Zunachst strecken sie sich in die Lange, entsenden 
Ausliiufer und breitem sich aiif den benachbarten Dotterkugeln aus.^’ 

Goette (82) in Stylochus pUlidium finds, as already mentioned, upper 
endoderm cells. From these and from the middle endoderm ceils 
(fourth quartet) small cells are separated, which at first contain some 
of the fat-like yolk. This soon disappears in these cells and the}" 
come to form a definite layer of endoderm. The small lower endoderm 
cells (macromeres) also take part in forming this layer. Large drops 
of the homogeneous }"olk substance separate from the large middle 
endoderm cells and are gradually absorbed by the cells of the canal. 

Lang (84) in Discocaiis, it will be remembered, also finds an upper 
endoderm as well as a lower and a middle layer. Like Selenka, he no 
longer finds any nucleus in the large cells (fourth quartet) after the 


XATUUAL scii:xri:s of piiiladflpiiia. 


upper and lower endoderm hav(‘ separated off. With regard to the 
middle oiidodenn cells he says, ]>. ooO: ‘M)ie in ihnen enthaltenen 
grol)en Ootterkonier v^eheinen init(*inander zu verschmelzen so dass 
die in l^'rage steheiiden Zeilen das Aiissehen von beinahe homogenen, 
stark lichtbrt'ehender Fettkugeln bekamen. Ich habe in diesen 
Kngeln bei ihrem zerfall nie AmphiastiT sieh bilden sehen, obschon 
ieh aufmerksam danaeh gesueht habe.’^ 

The breaking np of these yoHv cells is very irregular. The wall of the 
alimentary canal, according to Lang, is formed by the cells of the 
iip]ier and lower endoderm. I'hese increase 1)}^ division and extend 
over the yolk s])heres. Finally they unite to form a definite la}^er in 
which the cell ])oundaries cannot be distinguished. The endoderm cells 
have a more or less amceboid character and send out protoplasmic 
processes over the numerous }a)lk spheres. According to him the 
middle endoderm (fourth quartet) contains only 3 mlk granules and 
does not take part directly in the formation of any organ. 

Throughout all these accounts one or two phenomena arc constant. 
c.g.j the yolk breaks up into a large number of spherules which arc later 
absorbed b\' the endoderm cells. With regard to the development 
of the canal itself there is some variation. In all the accounts at least 
a portion of the canal is derived from tiie lower endoderm (macro- 
meres). In some cases {Stylochus and Discocoelis) upper endoderm 
cells are formed from the large yo\k cells. 

The account which 1 have to offer of the development of tlie ali- 
mentary canal in Planocem inquiU)i(i differs from aii}^ of the above in 
se\'eral particulars. At the time when the mesoblasts, 4d-'^*- and 
are preparing to divide, the two large entoblasts, 4d^-^ and 4f/^*‘, 
are dividing (fig. 133). By this division two large cells, and 
are budded into the interior of the embiA'o. Soon after this these cells 
divide again. At a considerabl}" later stage the two cells and 

^^^ 1 . 2 . 1 , ^viiich remained on the surface after the last division, divule 
again, budding two more large cells into the lower part of the embr}m. 
At this time or shortly afterwards the ectoderm has covered this 
region, and all six cells and their descendants originally derived from 
the two primitive entoblasts, 4c/^-^ and 4d^-^, are on the interior of 
the egg. 

Text fig. 3 shows in o])tical section an egg of a considerablv later 
stage, in which a number of cells are lying just above the vegetative 
pole. These cells, of which there are several more in the egg, all came 
from the primitive entoblasts, 4(/^-^ and 4d^ -. By examining a large 
nunil)('r of eggs it is found that lh(‘S(' c(‘lls are in verv active division 




and .soon a lar«;e group of cells is found in this region (cf. fig. 36). A 
careful stud}" of the eggs themselves leaves no doubt Ijiit that they all 
arise from the primitive entoblasts. PI. XXXIX, fig. 33, shows the 
spindles for the first division of these cells. 1 have repeatedly found 
eggs sliowing the second and later divisions of these cells, but in every 
case it has been found impossible to make an intelligible drawing of the 
egg. The fact that these eggs cannot be rotated under the cover glass 
has added greatly to the labor of finding eggs suitable for study or 
drawing. In this case it has effectually blocked all attempts to por- 
tray a certain stage. The material itself, however, leaves no doubt 
that the divisions take place as described. 1 have attempted to 
embody the essential facts of these divisions in the schematic text 
fig. 2. 

During all this time the very large nuclei of the three anterior cells 
of the fourth quartet, 4a, 46 and 4c, can still be found (fig. 34). Pre- 
vious to this time there has been some shifting of the relative positions 
of the fourth quartet cells. At the time the mesentoblast cell is budded 
into the interior (fig. 19) all the four cells of this quartet lie in nearly 
the same plane. The two lateral cells, 4a and 4c, are perhaps slightly 
higher than the others. When 4ri^ divides bilaterally these two cells 
overlap 4a and 4c. With the further development of the mesento- 
blast on the interior of the egg and the ectoblast on the outside, the 
large cell 46 is pressed upwards. A narrow process from this cell runs 
along the centre of the egg, and in this, reaching almost from one side 
of the process to the other, lies the enormous nucleus of 46 (figs. 32 and 
34). The nuclei of 4a and 4c are also pressed up along the sides, iDut 
not so high as that of 46. These three cells are closely crowded 
together, and while their boundaries remain distinct for some time 
(fig. 32) they tend to become obliterated. 

In the meantime an interesting process has been going on within 
these cells. As has been noted by all previous students, the yolk 
granules tend to fuse together, thus forming homogeneous fat-like 
drops. As Selenka and Lang have noted, this breaking up of the yolk 
is not accompanied by nuclear division. The process is more or less 
irregular, but one or two regular features can usually be recognized. 
The first one of these s])herules to be formed is from the anterior and 
ventral portion of the cell 46. At first the centre of these spheres is 
composed of granules, while around its periphery the granules dissolve 
into a fluid substance. Soon afterwards smaller sjiheres appear in this 
cell and in 4a and 4c. No nuclear division is concerned in this process, 
for the undivided large nuclei of all three of these cells can be followed 




to n inucli later jx'i'iod. In or lu^ar the caMitre of many of th(‘so yolk 
8[)hcrcs a (liffuscdy stainin.ii; suhslanco can 1 k‘ found wlnai tin* is well 
stained and cleared. This is not a niicUnis, for it has no regular 
honndaries and often sevc'ral such bodies arc seen in a single yolk 
spherule. They do not stain deeply, hut appear as a sort of cloudy 
material. These bodies may Ix' merely the small j)ortions of cyto- 
plasm remainin*;' in these Iar^(' ct'lls or, what seems more likely to me, 
they may be of the nature of mich^ar sap. If tliey were of a cytoplasmic 
riature there is no reason why they should not show in earli(T stages. 
Instead they became e\ddent only after the yolk begins to break up 
into spherules. In many respis'ts they r(‘S('ud)le the archiplasmic 
material often found around a nucleus just after the nuclear mem- 

Fig. 4. — Transverse section of an embryo, showing the ciliated lumen of the 
alimentary canal. The endocierm cells are sj'jreading over the yolk spheres. 
end., endoderm; mes., mesoderm; /)h., phiirynx; y., yolk. 

brane has broken. Further, the large nuclei, which are at first spherical, 
in the late stages become irregular in shape or even very much shrunken 
(fig. 35). That these yolk granules should be broken up through the 
action of enzymes coming from the nucleus is not at variance with the 
modern views of nuclear activity. 

In such a case it seems jirobable that this material would become 
aggregated at those places where the most rapid dissolving action is 
going on. This offers an explanation for the relatively enormous size 
of the nuclei of these three cells. These nuclei do not show evidences 
of degeneration; instead the chromatin granules can be seen scattered 
through the nucleus, often along distinct linin threads. 

After the ectoderm has covered the lower hemisphere of the egg it 




begins to iiivagiiiate in the region of tlie four small inacroineres. The 
ectoderm pushes in and forms a small tul)e, which later becomes the 
])harynx. Previous to this invagination the endoderm cells derived 
from the divisions of and have formed a solid mass of cells 
in the lower part of the embryo (fig. 30). Soon after the invagination 
starts, the beginning of a lumen in the endoderm cells is apjiarent by the 
separation of the cells (fig. 37). This lumen rapidly becomes large and 
ciliated throughout (PI. XL, figs. 38, 39, text fig. 4). The canal is at 
first bent towards the posterior side of the larva (fig. 38), but with its 
further development and enlargement it pushes forward under the 
ganglion (fig. 39). Distinct cell boundaries can seldom be made out 
in the endodermal portions of the canal. The inner borders of these 
cells surround the large yolk spherules. In many cases amoeboid cells 
can be seen spread out on the surface of these yolk spheres. In the 
oldest larvre which I was able to obtain (fig. 39) the canal showed no 
indications of the lateral branches which become evident in the adult 
worm. In these larva3 some of the reduced yolk spheres are still 

From the account I have given it will be seen that practically all 
of the alimentary canal is derived from the two j:)rimaiy entoblasts, 
4d^-^ and 4d^-\ Certainly the larger portion of the canal has such an 
origin. There are three other possible sources of a portion of the canal,, 
although no one of these forms any considerable amount of it. One 
is that the pair of small cells, 4d^*^ Land 4d“-“ \ derived from the further 
division of the mesentoblast, may form a small portion of the endoderm. 
The chief reason for suspecting this, is that these cells are formed and 
remain exactly in the path of the future canal. The second possibilit}^ 
is that the three anterior cells of the fourth quartet, 4a, 46, and 4d, may 
contribute a small amount to the canal at a late period. I believe, 
however, that sueh is not the case. The shrunken nuclei of these 
three cells, which can be seen at the time the canal is forming (fig. 36), 
indicate that these cells degenerate without dividing further. That 
the shrunken condition of these nuclei is not a preparation for karyo- 
kinetic division is, I believe, fully established by the fact that these 
nuclei remain in this shrunken condition for a very long time. Their 
abilit}" to take up the stain gradually becomes less and less, and the last 
that can be seen of them (fig. 36) shows a faint, ^^eiy irregularly out- 
lined nucleus, not at all resembling one about to divide by mitosis. 

The third possibility in this connection is that the four small macro- 
meres may, instead of degenerating, contribute a portion to the 
endodcTin. The degenerating character of the nuclei of these cells and 




tlic very small amount of cytoplasm seems to me to preclude such a 
fate in Planocd'a. 1 cannot find or at least cannot recognize these 
cells after the ectoderm has overgrown tlie lower pole of the egg, and so 
1 am unable to follow their later liistoiy. Text fig. ?> gives an accurate 
representation of the condition of the nuclei of these cells at that stage, 
which is about as late as I am able to follow them. These nuclei have 
shown this same condensed condition of the chromatin almost from 
the time of their formation (cf. figs. 17, 19, 20, 25, 20, 27). These 
cells arc probably carried in with the iin aginating pharynx and then 
absorbed b}^ the endoderm. I have sometimes found what I thought 
were remains of these cells, but of this I cannot be certain. 


We may now return to a comparison of tlie observations of previous 
students with regard to the fourth quartet, and especially 4d, in other 
polyclad eggs. As has been already noted (p, 530) the peculiar bilateral 
division of the posterior cell, 4c/, has been known since the work of 
llallez (79). Ilallez believed that this posterior cell divided without 
an amphiaster, and regarded the product as not equivalent to the other 
four cells, but in the nature of cell sap. Goette (82) states that in some 
cases the cells 4a and 4c of our nomenclature also divide, so that there 
are seven cells, on the surface of the embryo, formed from this quartet. 
This happened£only in some of the eggs of Stylochus (Stylochopis) 
'pillidium. In other eggs of the same species only the posterior cell 
divided. In such cases Goette found that this cell often divided twice, 
forming four cells, all on the surface of the egg. 

Lang (84) finds that in Discocedis all the cells of this fourth quartet 
divide at about the same time, but in different directions. The pos- 
terior cell divides horizontally as described, while each of the other 
three^cells buds into the interior of the egg a cell which he calls upper 
endoderm. Lang sa}"s (p. 337): ‘‘Es treten in ihnen [fourth quartet of 
our system] Iliclitungsspindehi auf, und zwar wieder in der oft ange- 
fiihrten Reihenfolge. Die Riehtuiigsspind(4 der grossten Stammzelle 
des Entoderms [4a] verliingert sich excentrisch in der peripherischen 
Vcrlangerung der l^ibene, welche man sich durch diese Stammzelle und 
die Hauptachse des Eies gclegt denken kann, und welche der ^lediane* 
bene entspricht.^^ .... (P. 338): Lnmittclbar befor sich diegrosste 

Entodermstammzelle [4f/] in Hire zwei seitlichen Ilalfteii getheilt hat, 
zeigen sich auch in den drei Uebrigen, Richtungsspindeln, die aber eine 
gaiiz andere Direction haben. Sie liegen niimlich parallel zur Ilaiip- 
tachse, d. h. sie zeigen eine dorsoveiitrale Hichtung. Die drei 




erwahnten Stammzellen zichen sich in der That gegen den aboralen 
Pol zii aiis, iind schnlii*eii schliesslicli je eine kleine Zelle ab, welche 
unterdie ]\lesodermzellen [second and third quartets] zii liegen kommt.^^ 
These three cells Lang calls the upper endoderm, while the five cells 
below he designates as the ^^mittler Entoderm/^ Goette (82) (p. 9) 
also states that six or seven cells are budded into the interior of the 
embr}^o of Stylochus as upper endoderm, but does not give further 
information as to their exact origin. 

In Planocera inquilina, on the other hand, the three anterior cells of 
the fourth quartet, viz., 4a, 45, and 4c, do not divide at this time nor 
for a long time afterwards, if ever. The very large nuclei of these three 
cells can be traced up until just before the formation of the alimentary 
canal (fig. 36), and at this time they have not divided. The nuclei of 
these cells, especiall}" 45, become exceedingly large (fig. 32) and form- 
excellent landmarks in the later stages. 

Wilson (98) finds and figures the bilateral division of 4d, He states 
that the division of this cell into equal halves is an exception to the 
rule, and that in about 90 per cent, of the eggs of Leptoplana the division 
is markedly unequal (cf. his fig. 6, D, E and F), Wilson did not follow 
the future division of these cells. He says (p, 22): ^^As regards the 
fate of these cells, the inequality of 4d^ and 4^^ [4d^-^ and 4d^-^] (often 
very marked) is itself indirect evidence that they do not give rise to 
symmetrical mesoblast bands as in the higher types, and I find no 
evidence that either of them gives rise to mesoblast cells. Both seem to 
have the same fate as the other entoblast cells, with which they exactly 
agree in deutoplasmic structure, and enter into the formation of the 
archenteron, as Lang has shown in the case of DiscocoelisP 

It is peculiar that Lang should have observed ‘cells budding into the 
interior of the egg from three cells of the fourth quartet and not from 
its other member, while I find that it is only this latter cell which 
divides towards the interior, or in fact the only cell of this quartet that 
divides at all. Lang gives figures of the spindles in the three anterior 
cells of the fourth quartet in Discoccelis and also a detailed description 
of these divisions. Although this work was done on living eggs, it 
does not seem probable that so careful an observer would be mistaken 
in the facts. We must conclude, then, that the three upper endoderm 
cells of Discoccelis are in fact endoderm, and that this species differs 
from Planocera in the division of the three anterior cells of the fourth 
quartet. We would not be surprised to find such a coenogenetic 
difference in different species of polyclads. Discoccelis perhaps shows 
a more primitive condition in this respect, in that cells which are to 

1007 .] 



form part of the aliment ary canal arisen early from these three cells of 
the fourth quartet. In Phinoccra apparcaitly most of the protoplasmic 
material has been separated from the cells 4 r, 4b, and 4c in the previous 
divisions, and they now contain little more than a mass of yolk granules. 

Lang, no doubt, overlooked the internal budding of the posterior C(‘ll, 
4d, if such occurs in Discocalis. This internal division of 4d in Plano- 
cera is very evident and striking. In other particulars the cleavage of 
Planoccra inqniUna is so similar to that of other polyclads that it is 
difficult to believe that this species differs so fundamentally in respect 
to 4r/. As stated before, practically all the previous work on polyclad 
embryology has been done on the living eggs alone. Such a division 
as the internal budding of 4d might easily be overlooked in the living 
opaque eggs. On this basis we might well conclude that in all prob- 
ability such a division was overlooked l)y Lang and his predecessors. 

Wilson, however, who undoubtedly was on the lookout for just 
such a division, did not find it in Leptoplana. Wilson sa}"s that 
he docs not attempt to ‘T1 escribe the cleavage of Leptoplana in 
detail, but onh" indicate its leading features.'' I cannot but 
believe that in this statement is contained the reason wh}’ Wilson 
did not find mesoderm arising from 4d. A process so evident and 
significant as these divisions of 4d in Planocera can scarcely be con- 
ceived of as a coenogenetic character of one or even a few species of 
polyclads. In annelids and mollusks the bilateral division of 4d and the 
origin of mesoderm bands from these cells is without exception in the 
numerous species so far studied. Both Wilson and myself have now 
shown that the ectoderm of polyclads is segregated in the first three 
quartets of micromeres, and that the second quartet gives rise to some 
mesoderm. This process, so exactly paralleled in molluslvs and annelids, 
can now scarcely be doubted as constant in its main features for all 
polyclads. Whether such a uniformity in the origin of the mesoderm 
from 4d will be found to hold throughout these Turbellaria can onl}" be 
proven by further investigations on other species. I l)clieve that 
eertainl}^ the weight of evidence is in favor of this imiformity. 

The resemblance in the behavior of 4d in the polyclads to the homo- 
logous cell in annelids and molluslcs becomes only more striking as we 
consider the details of its divisions. It is true that in annelids and 
mollusks 4d divides into two bilateral halves before it buds cells into 
the interior, while in Planocera 4d first buds a single cell into the seg- 
mentation cavity and then each of the two divides bilaterally (figs. 
19, 25, 20). In either case exactly the same result is reached, and the 
•delay of the bilateral cleavage for one-cell generation can certainly be 
very easily accounted for as a coenogenetic modification. 




Conklin (97) was the first to point out that in the gasteropod Crepi- 
dulas the cell 4d gave rise to both endoderin and niesodenn. In Crepi- 
dula four approximately equal cells are at first formed from 4^/. The 
two lower and external are pure entoblasts. I^lach of the two upper 
cells later gives off anotlier entoblastic cell before they give rise to the 
mesoblast bands. Since that time numerous observers have found that 
in both annelids and mollusks the cell Ad is mesentoblastic. Wilson 
(98), in the paper so often referred to above, shows that a reinvestigation 
of Nereis proves that a iiumlxu' of small eiitoblast cells are budded off 
from the two halves of 4d before these form the mesodermal bands. 
He also shoAvs that a series of stages may be found in different annelids 
and mollusks, ranging from a single pair of minute vestigiaD^ entoblast 
cells in Aricici and Spio to Nereis Avhere from six to ten small cells are 
budded off, and to the condition in Crepidula where more than half the 
bulk of 4d forms endoderm. Around these and other facts Wilson has 
elaborated a beautiful theoiy of ancestral reminiscence. To this series, 
agreeing very closely with Crepidula, wq ma}" add the pol 3 Tlad Plano- 
cera inquilina. Here, as in Crepidula, the two loAA’er superficial cells 
derived from 4r/ are pureW entoblastic and, as has been shown, give rise 
to very nearly all of the alimentarA" canal. Taa o more small cells, 
and 4d?'^'^, deriA^ed from the tAvo upper cells of 4d, are probabl}' added to 
the endoderm, AAdiile the remainder forms mesoderm. This close, almost 
astonishing, agreement of Planocera Avith annelids and mollusks 
cannot be Avithout some significance. As Wilson (98) (p. 13) sa>^s: 
“If Ave accept Langts vieAV, Avhich is supported b}" a large amount of 
evidence, that the platodes are not verv far remoA'ed from the ancestral 
prototype of annelids ami mollusks, Ave should expect to find in the 
polyclad a mode of cleavage to AAdiich that of the higher forms can in its 
main features be reduced.’^ 

Before Wilson^s Avork this resemblance betAA'ceii pol^'clads and higher 
forms had seemed to be “onl\^ in the jonn of cleavage and not, so to 
speak, in its substance.’’ I IjelieA^e that now this difficulty has been 
entire!}^ removed, and the i)olyclad cleavage not onl}" conforms to the 
higher types in its “main features” but, as we have seen, in many of 
its details. These facts liere set forth cannot but lend additional 
weight to the view already expressed on comparative anatomical 
grounds that the pol 3 "clads reju'esent an offshoot from the same 
ancestral branch Avhich later gave rise to the annelids and mollusks. 
On the other hand, it is a remarkable and interesting fact that phyla 
which must have separated from the common stock and from each 
other long ages ago still sIioav such remarkable resemblances in their 




early oiito^'eiiies. Vet I tliink this is the only int(‘rj)retation we can 
put upon the facts. With oiir present knowledge of ‘MMitwickelun^s- 
ineclianik^’ we cannot interpet these resenildances as clue to external, 
mechanical conditions. Why the ectobhist should be sep;regated in 
three and only three quartets in polyclads, annelids, mollusks and, as 
Biglow (02) has shown, in some crustaceic, and why in these same 
groups mesoderm should arise from the posterior or left })OSterior cell 
of the fourth quartet and from no otlicM* of this quartcd, are questions 
which do not readily lend themselves to a mechanical explanation 
under our ])resent theories. We must for the present at least regard 
these resemblances as facts of heredity and hence of })hylogentic value. 

The origin of the alimentary canal in Plauocera is unique among 
animals, so far as 1 am aware. 'I' he whole of the alimentary canal arises 
from a portion of the posterior cell of the fourth <iuartct, while the other 
three cells of this (jiiartet and all four of the macromeres are used as food 
or deycncrate and give rise to no morphological structure. Not one of 
the last seven cells mentioned ever dirides after its formation at the thirty- 
two-cell stage. Surprising and unique as this phenomenon may be, it 
does not necessarily invalidate our present conception of the develop- 
ment of germ layers or their organs. Since the establishment of cell- 
lineage work it lias become well known that many embryonic cells are 
formed in early stages which are destined never to divide again, nor to 
take any further part in the organization of the embryo. Compare 
for example the '' turreC’ cells of certain mollusks as Crepidula{Conk\in, 
97). As has been pointed out. it is well known that in many annelids 
and molluslcs the cell 4c/ gives rise to a ]iortion of the alimentary canal. 
In these animals however the other three cells of this C[uartet as well 
as the macromeres take part in the formation of the digestive tract. 
Indeed these latter cells furnish the major portion of the alimentary 
tissue. With these facts in mind, we must regard the condition of the 
macromeres and the three anterior cells of the fourth quartet in Plano- 
cera as reminiscent of a time when all eight of these cells took part in 
the formation of the alimentary tra(*t. Thus the embryology shows 
that this worm is specialized in this respect and must long ago have 
left the track which led on to annelids and mollusks. 

ddiis peculiar development of the alimentary canal in the })olyclads 
<)ff(‘rs the suggestion that it may be a step toward the development of 
such forms as the acoelous rhabdocoeles, in which the alimentaiy canal 
is altogether absent. But the embryology of these as well as other 
turbellaria present such great variations from the type found in the 
])olyclads that it is useless to speculate along this line. 





The cleavage of the eggs of Planoccra inquilina Wh. until a late stage 
(forty-four cells) is strictly spiral in the dextral sequence. 

Three quartets of ectomeres are given off in alternating clexiotropic 
and lieotropic directions. At the next division a fourth quartet is 
formed, the cells of which are of very large size and contain most of 
the yolk. The “macromeres^^ are veiy minute cells which remain at 
the vegetative pole until the closure of the blastopore. The marked 
degenerative character of their nuclei and the small amount of cyto- 
plasm indicate that they degenerate without giving rise to any struc- 
ture (p. 529). 

At the stage with forty cells there are formed at the animal pole 
four small ‘^apical” cells, which in their method and time of origin 
correspond closely to the cells of the same name in annelids and 
mollusks (p. 530). 

At the forty-four-cell stage the posterior cell of the fourth quartet, 
4d, buds a single large cell into the interior of the embryo. Both of 
these cells, and 4cP, next divide bilaterally (p. 536). 

Of these four cells the two upper and inner give rise to a portion of 
the mesoderm and possibly a small part of the endoderm (p. 537). The 
lower pair of cells, lying on the surface of the embryo, give rise to prac- 
tically all of the endodermal part of the alimentary canal (p. 541). 
Thus the history of this cell, 4d, shows a remarkable resemblance to 
its homologue in mollusks and annelids. 

The three anterior cells of the fourth quartet, 4a, 45 and 4c, seem to 
function only as the bearers of food yolk and apparently give rise to 
no morphological structure. The very large nuclei of these cells can 
be followed until the beginning of the pharyngeal invagination. The 
yolk in these cells breaks up into spherules, probably through the 
action of enzymes from the large nuclei. This liquified yolk is later 
absorbed by the endoderm cells (p. 542). 

A large portion of the ectoderm is formed by the successive budding 
or delimination of small cells from larger, deeper l^dng ones (p. 533). 

A portion of the mesoderm, chiefly that part lying around the 
pharynx, is derived from cells of the second quartet, and thus corres- 
ponds with the secondary mesoblast or ^darvah^ mesenchyme of 
annelids and mollusks (p. 535). 

In the spiral cleavage, the segregation of the ectoblast in three 
quartets, the formation of a large part of the mesoderm from 4d, the 
formation of the apical cells, and in many other details the development 




of these platodes corresponds to that of annelids and mollusks. These 
facts must tend to confirm the view that in their early history these 
platodes were closely related to the two last mentioned phyla. 

On the other hand, in the development of the entire alimentary 
canal from a portion of the mesentoblast, 4d, and in the consequent 
degeneration of the “macromeres^^ and of the remaining cells of the 
fourth quartet, this potyclad is unique. This peculiar development of 
the alimentary tract shows that the cleavage of the polyclads, while 
closely resembling that of other groups, is not a generalized type. 

1907 .] 





+ 001 










Bigfow, M. a. 1902. The Karly Development of Lepas. Bull. Mus. Comp. 
Zoo. Harvard College, XL, Xo. 2. 

Blochmann, F. 1883. Beitriige zur Keimtriiss der Entwickeluiigs der Gas- 
teropoden. Zeit, /. ]Viss. ZooL, XXXVIII. 

Casteel, C. D. 1904. The Cell Lineage and Early Development of Fiona 
Marine, a Nudibranchiate Mollusk. Proc. Acad. Nat. Sci. Phila., 1901. 
Conklin, E. G. 1891. Note on the Embryology of Crepidula fornicata and 
Urosalpinx cinera. John Hop. Univ. Circ., X, No. 88. 

1892. The Cleavage of the Ovum of Crepidula fornicata. Zool. A nz., XV. 

1897. The Embryology of Crepidula. Journ. Morph., XIII. 

1902. Kar 3 mkinesis and Cytokinesis in the Maturation, Fertilization and 
('leavage of Crepidula and other Gasteropods. Journ. Acad. Nat. Sci. 
Phila., 2d Senes, XII, 1. 

— - 1905. The Organization and Cell Lineage of the Ascidian Egg. Journ. 

Acad. Nat. Sci. Phila., XIII, pt. 1. 

CuiLD, C. M. 1900. The Early Development of Areiiicola and Sternapis. 
Arch. Entw. Mech., IV. 

Eisiq, Hugo. 1898. Zur Entwickelungsgeschichte der Capitelliden. Mitth. 
Zool. Stat. Neap., XIII. 

Francotte, R. 189jL Recherches sur la Mat. la Fecond. et la segm. chez des 
Pol}^elads. Arch, de Zool. Exp. et Gen., 3® ser., VI. 

Gardiner, E. G. 1895. The EarL^ Development of Polychoerus eaudatus 
Mark. Journ. Morph., XI. 

1898. The Growth of the Ovum, Formation of the Polar Bodies and the 

Fertilization in Polyehoerus eaudatus. Journ. Morph., XV. 

Girard, Charles. 184G-7. Sur rembrjmgenie des Planaires. Bull, de la 
Societe des Sci. N^at. de Neuchdtel, II. 

1848. On the Development of Planocera elliptica. Proc. Bost. Soc. Nat. 

Hist., III. 

1854. Researches upon Nemerteans and Planarians. I. Embryological 

Development of Planocera elliptica. Journ. Acad. Nat. Sci. of Phila., II. 
Goette, Alexander. 1878. Zur Entwickelungsgeschichte der Seeplanarian. 
Zool. Anz., I. 

1881. Zur Entwickelimgsgeschi elite der Wiirmer. Zool. Anz., \. 

— - 1882. Abhandlungen zur Entw. der Thiere — -1 : Uiitersiiehungen zur 

Entw\ der Wiirmer. Leipsig, 8®. 

Graff, L. V. 1892. Pelagisehe Polveladen. Zeit. j. IFfss. Zool., 55. 

Hallez, Paul. 1878. a. Contrilnitions a I’histoire des Tiirbellaries, P* note 
sur la d^veloppement. Bull. Scient. du Depart, du Nord. 2*^"® ser., P® aimee, 
pp. 193-195. 

1878. b. Considerations au sujet de la segmentation des oeufs. Ibidem, 

pp. 227-229. 

— - 1878. c. Contributions a I’histoire des Tiirbellaries. Ibidem, pp. 2 o0-2(j0. 

1878. d. Considerations sur la determination des plans de segm. dans 

I’emb. du Leptoplana tremellaria. Ibidem, pp. 264-266. 

— - 1879. Contributions a Thistoire nat. des Tiirbellaries. Lille; also in 

Trovaux de V Inslit. Zool. de Lille et de la Stat. marit. de Wimereux, Fas. II. 
Heath, Harold. 1899. The Development of Isehnoehiton. Zool. Jahrb., 
Abth. f. Anat. u. Ontog., XII. 

Heymons, Richard. 1893. Zur Entwick. von Umbrella medit. Lam. Zeit. 
f. IFnss. Zool., .56. 

Holmes, S. J. 1897. Secondary IMesoldast in the Mollusca. Science, N. S., VI. 

1900. The Early IVvclopment of Planorbis. Journ. Morph., XVI. 

Jennings, H. S. 1895. The Early Development of Asplanchna Herricki 
DeGuerne. Bull. Mus. Comp. Zool. Harv. Coll., XXX. 

Kbferstein. Welh. 1868. Beitrage zur Anat. u. Entw. einiges Seeplanarian 
von St. Malo. Abhnnd. d. Konigl. Gesell. der TFiss. zu Gottingen, XIV. 
Kofoid. C. a. 1891. On Same Laivs of Cleavage in Limax. Proc. Am, Acad. 
Arts and Sci., N. S., XXL 




KoitsciiEL’r, 10., and IIkidek, K. 1S93. l.chrbueli (Ut vergleiehcMulcn Entw. 

dcr Wirbellossen 'riiiere. Sp. Jena. 

Lang, Ahnold. 18S L Die Kolyeladen, des CJolfe.s von Xeapel. Fauna, u. 
Flora d. (Joljes v. Xcap., XI. 

Lillie, E. H. 189a. Tlie Einbryolog}' of the Fnionid:e. Journ. Morph., X. 
Mead, A. D. 1897. Tlie Earl}' Development of Mariin* Annelids. Journ. 
Morph., Xlll. 

Meyek, E. 1887. Studien iiben den Ivorperbaii der Anneliden. Mith. Zool. 
Slat. Neap., VII. 

1890. Die Abstammung der Anneliden. IFol. Cent., X. 

Nelson, J. N. 1904. Tlie iOarly De\(‘l()j)nient of Dinopliibis. I*roc. Acad. 
Nat. iSci. Phila., October, 1904. 

I’eueyaslawyew, Mlle. S. 1885. Siir le dcAeloppement des Turbellari<^s. 
Zool. Anz., \TI1. 

Uepiachoff, \V. 18S5. Xacbtrag zu Pen'vuslawvew’s Devel. de.s Tiirbel. 

Zool. Anz., VIIL 

Selenka, Emil. 1881. a. Die K(“iinbliltter d(*r Planarien. Sitznmj.djer. der 
Fhysik-Medic. Soc. zu Frlanyer, No. 14. Trans, in Bull. Sri. Dipt, du Xord. 
1881. b. Tur Entwiekelungsgesehiehte der Secplanarien. JUol. Cent., 

Bd. 1. 

1881. c. Zool. Studien — II: Znr I^nlw. der Seeplanaricn. Eiii Beitrag 

zur Keiinblatterlehre iind Deseendenz Tlu‘orie. l.eipsig. 

1881. d. Ueber cine eigen tliunilielie Art der Kerninetainorphose. Biol. 

Cent., Bd. 1. 

Surface, F. M. 1907. Note on the Origin of tin* MesodiTiu of the Poly elad 
Planocera inquilina \Vh. Science, N. S., Vol. XXV, p. 732. 

Tourey, J. C. 1902. The Early Development of the xMesoblast in Thalas.sema. 
Ann/. Anz., XXI. 

1903. The Early Embryology of Thalasseina Mellita (Oonn). Ann. X. 

Y. Acad. Sci., XIV. 

Tkedwell, a. L. 1901. The Cytogenv of Podarke obseura \'er. Journ. 
Morph., XVII. 

\’aillant, Leon. 180G. Siir le developpeinent du Polycelis laivigatus, Quatref. 
Ulnstitute, I Sect., Tome XXXDL 

1808. Kemarques sur la developpeinent d'une Planarize dendrocAe de 

Quatref. Mem. de VAcad. des Sci. et Lettres de Montpellier, T. VII. 

^'A.N Beneden, Edouard. 1870. Keserehes sur la eompo.sition et la Significa- 
tion de Poeuf. Memoires, couromes et mem. des Savants etranger publics par 
VAcad. Royal, dc Belgique, Tome 34, Bruxelles. 

Wheeler, W. M 1894. Planocera inquilina, a Polyclad inhabiting the 
Branchial Chamber of Sycotypus eanaliculatus (ull. Journ. Morph., IX, 
Whitman, C. O. 1878. The Embryology of Clcp.sine. Quart. Journ. Micr. 
Na*., XVIII. 

- 1891. Spermatophorcs as a means of Hypodermic Imj)rcgnation. Journ. 

Morph., IV. 

WiLso.v, E. B. 1892. The C(4l Lineage of Nereis. Journ. Morph., VI. 

1894. The Embryologieal Criterion of Homology. IPood.-? IIoll Biol. 
Lect., 1894. 

1898. Considerations on CAl Lineage and Anee.stral Reminiscences. 
Ann. AL Y. Acod. Sci., XI. 

1899. Cell Lineage and Ancestral Keminiseenees. iroed.-? Holt Biol. 
Lect., 1899. 


All figures of fixed and .stained eggs, except fig. 35, were’dra\ni with Zeiss 
Camera lueida at table leyel under Leitz obj. oc. 2. Figures of living eggs 
(vTz., figs. 1, 2, 3, 4, 6, 8, 10 and 11) were drawn as above, but with Leitz obj. 7; 
oc. 2. Fig. 35 was drawn with B and L. obj. ; oe. 1. All drawings have, 

been reduced oiu'-tliird. 




Reference Letters. 

e., cntoblasts and 7iAb., nucleus of the cell 46. 

end., endoderni. j^h,, phaiynx. 

ei/., e}'e. r., rhabdites. 

g., ganglion. g., yolk, 

mes., mesoderm. 

Plate XXX^^ — Fig. 1. — Living egg before fu'st maturation division, showing 
large germinal vesicle. 

Fig. 2. — Living egg duiing the second maturation division, shoeing amoeboid 

Fig. 3. — Jjiving egg after second maturation division. 

Fig. 4.— Living egg during the first division. The cell C-D is slightly larger 
‘than A~B. 

Fig. 5. — Stained egg in the second di^dsion. From left side, showing the 
crossing of the spindles. 

Fig. 6. — Living egg in the four-cell stage, from the animal pole. 

Fig. 7. — Formation of the first quartet. From right side, showing dexio- 
tropic cleavage. 

Fig. 8. — Living egg in the eight-cell stage. From animal pole. 

Plate XXXVI. — Fig. 9. — L^eotropic ch\ision of the preceding eight cells. The 
spindles in the D quadrant show" an advance over the others. From 
the animal pole. 

Fig. 10. — Living egg with sixteen cells. From the animal pole. 

Fig. 11. — Same as fig. 10, but from vegetative pole. 

Fig. 12. — Dexiotropic divisions of cells la^-ld^ and of 2a-2d. From animal 

Fig. 13. — Tlhrty-tw "0 cells. Dexiotropic divisions of la“'^-ld^ b From 
animal pole. 

Fig. 14. — Thirty-two cells, showing dexiotropic formation of 3a and 3c. 
Vegetative pole. 

Fig. 15. — Thirty-tw"o cells from vegetative pole, showing spindles for forma- 
tion of 46 and 4d. 

Plate XXXVII. — Fig. 16. — Same as Fig. 15, from right side. 

Fig. 17. — Thirty-six cells from near vegetative pole. Spindle in 2c^. 2a^ 

and 2d} have divided. 

Fig. IS. — Thirty-nine cells. From animal pole, showing Iseotropic forma- 
tion of apical cells. 

Fig. 19.— Forty-five cells, from the right posteiior side, Show"s the first 
division of the mesentoblast Ad. 2d'- also di^iding. 

Fig. 20. — Forty-seven cells. From vegetative pole. 

Fig. 21. — About the same stage as fig. 20,drawm from the right side, show"s 
the unequal di\ision of 2c^ and 26b Also spindles in Ic^-^ and 16^'b 
la^*^-ld^-^ have divided dexiotropically (see fig. 22). 

Plate XXXVIII. — Fig. 22. — Slightly older egg than fig. 21, drawn from left 
upper side. Shows division of la^-^'^-ld^-^*^ and of 16^*^. lol*^ and 16^-^ 
have divided previously. Also show"s Iseotropic division of Icd'^-ld}-^. 

Fig. 23. — Still older egg, showing the same division as fig. 22, but more 
advanced. From animal pole. 

Fig. 24. — Egg from animal pole, showing preceding divisions completed and 
dexiotropic division of Id^*^. Also the bilateral division of 4d^. 

Fig. 25. — Similar stage from vegetative pole. Show"s spindle for bilateral 
di\ision of 4db Mso Iseotropic di^ision of 2a^-^~2tP'b 3d preparing to 

Fig. 26. — Shows di^ision of 4d^ completed and spindle in 4db Also shows 
di\ision of 2d} and 26^, of 3d and 36, and of 2a^'b 

Fig. 27. — Similar stage from the left side. Show"s di\fision of 2a^*^, 4d^ and 
26“-^; also of la^-^*b 

Plate XXXIX. — Fig. 28. — Egg from animal pole, showing di^^sions of 

ldl-l.2.2; lc^-2'k 




Fijx. 20. — Similar ^^ow of slightly older egg, showing preceding divisions 
completed and the resulting small superficial cells 

Fig. .‘30. — Fgg from near the vegetative pole, showing division of Aid} com- 
pleted and budding in of the cells 26^-* and 2d‘*h 

Fig. 31. — Egg from animal pole at slightly later stage than fig. 29. Shows 
divisions of Adr'^ and The resulting small cells and 4d-'^-^ 

are })robably cntoblasts. 

Fig. 32. — Optical section from animal pole, showing the cells 2a|‘‘-2d‘-^ 
budding mesoderm to the interior. Also show^s the large nuclei of 46, 
4a and 4c. 

Fig. 33. — Later stage, showing the internal divi.sions of the entoblasts 4d^ 
and 4d-. Not all the cells on the interior of the egg are showm. 

Plate XL. — Fig. 34. — Optical section of egg from the posterior side. The 
entoblasts 4d^-^ and 4d^'^ have divided oncc’and possibly t^vice. The 
tw’o small entoblavsts and are shown with dark nuclei. 

Above these are the two small mesoblasts, wiiile the larger inesoblasts 
4 ^p. 1.2.3 4^^.3.2.2 again dividing. Above these are a number of 

cells from the first quartet wiiich later form the ganglion. The large 
nuclei of 4a, 46 and 4c are showai in dotted outlines. 

Fig. 3"). — Optical section of an egg view'cd from near the vegetative pole. 
The derivatives of 4d^ are stippled. The division of and 4<p*2.2.3 

indieated in fig. 34 are now^ completed and the beginning of the meso- 
tlermal bands is evident. In man}^ eggs the divisions are not so regular 
as showm here. Instead the mesoderm tends to form clusters of cells 
rather than bands. 

Fig. 36. — Optical section of a much later stage. From the left side. The 
mass of endoderm derived from 4d^ lies just above the future pharynx. 
The mesoderm cells have passed to the periphery and the centre of the 
egg is filled with the homogeneous yolk spheres (y), in w’hich the shrunken 
nucleus of 46 can be seen (a.46.). The two small cells marked (e) are 
probably the entoblast 4d-*^*^ and 

Fig. 37. — Optical section of an older embryo. From the posterior side. 
* Shows ectodermal phar 3 mx and beginning of lumen in the endoderm. 

Fig. 3S. — Actual longitudinal section of a still older embryo, showing the 
backw^ard turn of the alimentary canal in this stage. 

Fig. 39. — Longitudinal section of a Muller’s larva. Nearly all the yolk 
spheres have disappeared by this time and the alimentary canal is 
greatly enlarged but still unbranched.