mm
Journal of Experimental Zoology
THE JOURNAL
OF
EXPERIMENTAL ZOOLOGY
EDITED BY
WILLIAM K. BROOKS
Johns Hupkins University
WILLIAM E. CASTLE
Harvard University
EDWIN G, CONKLIN
University of Pennsylvania
CHARLES B. DAVENPORT
Carnegie Institution
HERBERT S. JENNINGS
University of Pennsylvania
FRANK R. LILLIE
University of Chicago
JACQUES LOEB
University of California
THOMAS H. MORGAN
Columbia University
GEORGE H. PARKER
Harvard University
CHARLES O. WHITMAN
University of Chicago
EDMUND B. WILSON, Columbia University
AND
ROSS G. HARRISON
Johns Hopkins University
Managing Editor
VOLUME I
THE JOURNAL OF EXPERIMENTAL ZOOLOGY
BALTIMORE
1904
ys?o"'V
4/7
BINDERVNH!:^BER_
I a.
MARINE BtOLOGlCAL
WOODS HOL^jJiAi?^
691
*^
O^
^
2-^ OCT 101
VVITH THE VOUJf>|IE^^^____ ^ "
. . . I . n \ i:t.tki;,
5itive Action of the i;r»ntmii liax^nn llegenerationin Planarians.
191
No. 2.— August, 1904.
Edmund B. Wilson,
Experimental Studies on Germinal Localization. II. Experiments on the
Cleavage-Mosaic in Patella and Dentaliimi. With 118 figures 197
A. J. Carlson,
Contributions to the Physiology of the Ventral Nerve Cord of Myriapoda
(Centipedes and Millipedes) . With 6 figures 269
Frank W. Bancroft,
Note on the Galvanotropic Reactions of the Medusa Polyorchis penicillata,
A. Agassiz. With 4 figures 289
Charles Zeleny,
Experiments on the I^ocalization of Developmental Factors in the Nemer-
tine Egg. With 19 figures 293
T. H. Morgan and Abigail C. Dimon,
An Examination of the Problems of Physiological "Polarity " and Electrical
Polarity in the Earthworm 331
Abigail C. Dimon,
The Regeneration of a Heteromorphic Tail in Allolobophora foetida 349
KJ?^''V
^/7
•s^
CONTENTS.
No. 1.— May, 1904.
Edmund B. Wilson,
Experimental Studies on Germinal Localization. I The Germ Regions
in the Egg of Dentalium. With 100 figures 1
Charles W. Hargitt,
Regeneration in Rhizostoma Pulmo. With 6 figures 73
C. M. Child,
Studies on Regulation. IV. Some Experimental Modifications of Form-
Regulation in Leptoplana. With 53 figures 95
T. H. Morgan,
Self-Fertilization Induced by Artificial Means .• 135
John Bruce MacC.\llum,
The Influence of Calcium and Barium on the Secretory Activity of the
Kidney 179
Charles Russell Bardeen and F. H. Baetjer,
The Inhibitive Action of the Rontgen Rays on Regeneration in Planarians . . 191
No. 2.— August, 1904.
Edmund B. Wilson,
Experimental Studies on Germinal Localization. II. Experiments on the
Cleavage-Mosaic in Patella and Dentalium. With 118 figures 197
A. J. Carlson,
Contributions to the Physiology of the Ventral Nerve Cord of Myriapoda
(Centipedes and Millipedes). With 6 figures 269
Frank W. Bancroft,
Note on the Galvanotropic Reactions of the Medusa Polyorchis penicillata,
A. Agassiz. With 4 figures 289
Charles Zeleny,
Experiments on the Localization of Developmental Factors in the Nemer-
tine Egg. With 19 figures 293
T. H. Morgan and Abigail C. Dimon,
An Examination of the Problems of Physiological "Polarity "and Electrical
Polarity in the Earthworm 331
Abigail C. Dimon,
The Regeneration of a Heteromorphic Tail in Allolobophora fcetida 349
Vernon L. Kellogg,
Restorative Regeneration, in Nature, of the Starfish Linckia Diplax (Miiller
and Troschel). With 6 figures 353
Vernon L. Kellogg and R. G. Bell,
Notes on Insect Bionomics 357
No. 3.— November, 1904.
Florence Peebles,
The Location of the Chick Embryo upon the Blastoderm. With 2 plates
and 15 figures in the text 369
T. H. Morgan,
Regeneration of Heteromorphic Tails in Posterior Pieces of Planaria
Simplicissima. With 20 figures 385
Harry Beal Torrey,
Biological Studies on Corymorpha. I. C. Palma and Environment. With
5 figures 395
Gary N. Calkins,
Studies on the Life History of Protozoa. IV. Deatli of the A Series.
Conclusions. With 3 plates and 3 figures in the text 423
C. M. Child,
Studies on Regulation. V. The Relation between the Central Nervous Sys-
tem and Regeneration in Leptoplana: Posterior Regeneration. With
47 figures 463
No. 4. — December, 1904.
C. M. Child,
Studies on Regulation. VI. The Relation Between the Central Nervous
System and Regulation in Leptoplana: Anterior and Lateral Regenera-
tion. With 64 figures 513
T. H. Morgan and N. M. Stevens,
Experiments on Polarity in Tubularia. With 5 figures 559
T. H. Morgan,
An Attempt to Analyze the Phenomena of Polarity in Tubularia 589
Vernon L. Kellogg,
Regeneration in Larval Legs of Silkworms. With 10 figures 593
Influence of the Primary Reproductive Organs on the Secondary Sexual
Characters 599
C. B. Davenport and Marian E. Hubbard,
Studies in the Evolution of Pecten. IV. Ray Variabihty in Pecten varius . 607
CORRIGENDA.
p. 200, 1. 4: for "vegetable" read "vegetative."
p. 201, 1. 27: after "both" insert "somatoblasts."
1.30: omit "CD."
p. 209, 1. 9: before "pro totrochal" insert "the."
p. 210, 1. 7: for "circle" read "row."
p. 212, 1. 4: for "size-relation" read "size-relations."
p. 217. 1. 7: for "ciliation" read "division."
p. 236, 1. 14: for "are" read "is."
p. 237, 1. 13: insert period after "forms."
1. 34: for " disintegrated" read " disintegrate."
p. 244, 1. 30: for " essentially" read " in many of its features."
p. 247, 1. 12: omit "the importance of."
1.13: omit "of."
1. 14: for "has" read "have."
p. 248, 1. 16: for " casual" read " causal."
p. 253, 1.33: insert hyphen after " entoblast."
p. 260, 1. 8: for " Ganzbeziehungsweise " read "Ganz- beziehungsweise.
p. 267, 1. 1: for "Studien" read "Stadien."
EXPERIMENTAL STUDIES ON GERMINAL
LOCALIZATION.
BY
EDMUND B. WILSON..
I. THE GERM-REGIONS IN THE EGG OF DENTALIUMi.
With ioo Figxjres.
CONTENTS.
I. Introduction.
II. Preliminary Observations on the un segmented Egg and the normal
Development.
III. Effect of removing the Polar Lobe.
(a) General history of the lobeless Embryo, with a Comparison of
isolated Blastomeres.
(b) The Mesoblast Question.
IV. Localization of the apical Organ, and its Correlation with the post-trochal
Region.
V. Localization in the unsegmented Egg.
(a) Development of Fragments obtained by horizontal or oblique Sec-
tion.
(b) Development of Fragments after vertical Section.
VI. Observations on enucleated Fragments of fertilized Eggs and on the
isolated Polar Lobe.
VII. Comment.
VIII. Summary.
INTRODUCTION.
The following experimental studies are offered as a contribu-
tion to the theory of "Organbildende Keimbezirke" or germinal
prelocalization, especially as applied to the cytoplasmic regions of
the unsegmented egg. Following the enunciation of the principle
of "precocious segregation" by Ray Lankester, in 1877, the im-
iThis work was carried out at the Naples Zoological Station between February
and August, 1903, on a grant from the Carnegie Institution of Washington, in
which was included the use of one of the tables subscribed for by the Institution.
My best thanks are due to the administration of the station for the unfailing
efficiency and courtesy with which my work was aided in every possible way.
2 Edmund B. JVilson.
portance of the cytoplasmic factors of localization and differen-
tiation was early recognized by Whitman in his remarkable paper
on Clepsine ('78) and emphasized by him in later papers. Sim-
ilar views were more or less clearly expressed by Van Beneden,
Flemming, Platner and others prior to the definite formulation of
the mosaic-theory of development by Roux in 1888.^ Roux
himself recognized from the first, as a prominent factor in his
theory, the importance of a definite topographical grouping of
specific cytoplasmic materials in the unsegmented egg; though
unfortunately this was complicated, then and in later discussions,
by the hypothesis of qualitative nuclear division, which has since
been shown to be untenable and has now been relinquished by its
author (Roux, 1903). Since that time the evidence, both cyto-
logical and experimental, has steadily increased that a prelocaliza-
tion of the morphogenic factors in the cytoplasmic regions is a
leading factor in the early development; and it has become evi-
dent that this is true not only in such "mosaic eggs" as those of
mollusks or ctenophores, but even in those of echinoderms or
nemertines, where an isolated blastomere or an egg-fragment may
produce a perfect dwarf embryo. It has become of high im-
portance to determine experimentally in what degree such pre-
localization or cytoplasmic "organization" may exist in the un-
segmented egg, and to what extent it may vary in different forms.
It is even more important for our general conception of develop-
ment to determine by the same method whether the prelocaliza-
tion of the morphogenic factors, in whatever degree it may occur,
exists from the beginning, or whether, as the cytological evidence
seems to show, it is established by a progressive process; for in
the latter case, as is hardly necessary to point out, prelocalization,
even in the unsegmented egg, may be brought under the category
of epigenetic phenomena ("epigenetic qualities" as distinguished
from "preformed qualities"^), and falls into harmony with hy-
potheses that assume the nucleus to be the primary determining
factor.
The present studies, which are a continuation of the preceding
1 Cf. my work on The Cell.
2 Boveri ('03 ), p. 356.
Experimental Studies on Germinal Localization. 3
ones on the nemertlne egg (Wilson, '03) bear upon both these
questions. In that paper I approached especially the second ques-
tion in an experimental study of the egg of Cerehratulus, which
has since been extended by the work of Yatsu ('04). My
results clearly showed that in this egg the cleavage-factors are
not definitely localized until after the completion of the ma-
turation of the egg, but they gave no definite evidence regarding
the localization of the morphogenic factors (as distinguished
from those of cleavage) at this period; it was, however, shown
that in the comparatively young blastula, before the formation of
the mesoblast, morphogenic localization, as shown in the pre-
determination of the gut and apical organ, has become much
more definite than in the unsegmented egg. Yatsu subsequently
obtained evidence, in the same species, that the localization of the
morphogenic factors is a progressive process even in the stages
preceding cleavage, since the percentage of normal larvae ob-
tained from egg-fragments at successive periods steadily dimin-
ishes from the first discharge of the eggs (when maturation be-
gins) up to the period immediately preceding the first cleavage;
and the nature of the defective larvae, correlated with the plane
of section, pointed to a increasingly definite localization, in the
later stages preceding cleavage, of the bases of several important
organs, such as the apical organ, gut, and ciliated lobes of the
pilidium. I am now able to offer an experimental analysis along
the same lines — perhaps I should say the beginning of such an
analysis — of the molluscan egg, in which pure observation of
the cell-lineage has produced such convincing evidence of mosaic
development, sustained by Crampton's initial experimental exam-
ination of the gasteropod egg ('96), and by the interesting cyto-
logical work of Lillie ('01) and Conklin ('02) on the cyto-
plasmic regions of the unsegmented and segmenting egg. The
cytological and experimental results coincide in demonstrating
in this egg (specifically in Dentalium) the existence of a very
definite prelocalization of some of the most important factors
both of cleavage and morphogenesis, which here closely coincide.
They show conclusively also, contrary to what the nemertine
experiments had led me to expect, that in its main features this
4 Edmund B. JVilson.
prelocalization exists In the egg at the time It leaves the ovary,
and probably much earlier, and long before even the Initial
stages of maturation and fertilization. Nevertheless, progressive
changes take place during and subsequent to maturation, which,
when compared with those occurring In other forms, show this
egg, as I believe, to be only the extreme of a series that connects
it with such forms as the nemertlne or echlnoderm, and brings
them under one point of view.
The present paper deals mainly with the development of frag-
ments of the unfertilized egg of DentaUum, the eggs being cut
singly with the scalpel under the microscope and subsequently
fertilized, following the method of Delage ('99). I shall here
consider the development of isolated blastomeres only incidentally
for the sake of comparison, reserving a fuller account for a second
paper. It may be stated here, however, that the experiments on
this part of the subject demonstrate, even more conclusively than
do those of FIschel for the ctenophore-egg, that the cleavage of
the ovum, In both DentaUum and Patella, Is In fact what the
normal cell-lineage so clearly indicates, essentially a mosaic-work,
In accordance with Crampton's earlier experiments on Ilyanassa.
Blastomeres Isolated at any stage from the 2-cell onward con-
tinue to segment as if still forming part of a complete embryo;
and apart from the changes due to shifting of the cells, which, as
in the ctenophore, often lead to the displacements of the larval
structures and to the closing of the partial embryos, undergo
essentially the same differentiation as If united to their fellows.
Thus, the first two blastomeres, upon separation, give rise to two
dissimilar larvae, each of which Is defective and represents es-
sentially the same structures as would have been produced had the
two cells remained united; In hke manner, of the isolated cells
of the 4-cell stage, the larva from the D-quadrant possesses cer-
tain structures that are lacking in the other three; and the dif-
ferences among the larvae from cells of the 8- or i6-cell stages are
still greater. Cells procured by successive Isolations up to the
64-cell stage, or later, differentiate singly, according to their na-
ture, into actively swimming trochoblasts of three kinds; into
ordinary ectoblast- or entoblast-cells, into sensory cells bearing
Experimental Studies on Germinal Localization. 5
the characteristic sensory hairs of the apical organ; and even into
what I believ^e to be muscle-cells and mesenchyme-cells, though,
unlike the foregoing cases, the precise origin of these was not
traced. These eggs thus represent the opposite extreme to such
forms as those of Amphioxiis, the echinoderm, or the nemertine,
and give a result which, apart from the hypothesis of qualitative
nuclear division, agrees essentially with Roux's original conception
of mosaic-development, with the conclusions of many students
of cell-lineage, with the experimental results of Crampton on
the gasteropod-egg, and with those of Fischel regarding the
ctenophore-egg.^
11.
PRELIMINARY OBSERVATIONS ON THE UNSEGMENTED EGG AND
THE NORMAL DEVELOPMENT.
The egg of Dentalium, like that of Cerebratuliis, possesses
certain features by means of which the axis may be determined
in the living egg from the moment of its release from the ovary.
The egg is more or less deeply pigmented, perfectly opaque, and
of a color that varies in different individuals from light olivaceous
to reddish brown or almost brick red. When first set free the
egg is somewhat irregular, but quickly becomes more rounded.
It is then seen to be very considerably flattened, so as often to
be almost biscuit-shaped, one side being always more flattened
than the other, and often more or less Irregular In contour.
Viewed by reflected light the central region of each of the flat-
tened sides is seen to be occupied by a very distinct, though vaguely
bounded, white area, nearly or quite free from pigment (Fig. i) ;
these areas, as shown by the subsequent development, correspond
with the two poles of the egg, and the more flattened side, which
1 The eggs of Patella, which were employed mainly for a study of the iso-
lated blastomeres, were available from the middle of March until the latter
part of May. Those of Dentalium, which were used especially for the develop-
ment of egg-fragments, first became mature at the beginning of June, when
less than two months remained for their study. The shortness of this period
accounts for some of the obvious gaps in my work. The complexity of the
subject, and the practical difficulties presented by the material are such that
more extended work, with additional material, will be required for its com-
pletion.
Edmund B. JVilson.
is the side of attachment in the ovary, is found to represent the
lower or vegetative hemisphere.
Fig. I.
Cleavage, from living Eggs.
I, Outline of egg soon after release, in polar view, showing white polar area;
2, the same egg, 20 minutes later, after throwing off the membrane; 3, similar
egg, from the side ; 4, egg one hour after fertilization, with fertilization-mem-
branei and polar bodies; 5, beginning of the first cleavage, formation of the
polar lobe; 6, trefoil, i^ hours after fertilization; 7, resulting 2-cell stage;
8, beginning of second cleavage from the side, second polar lobe forming; 9,
second cleavage at its height.
1 Accidentally omitted by engraver.
Experimental Studies on Germinal Localization. 7
During the 20-30 minutes following Its release the ripe, un-
fertilized egg becomes nearly spherical (and hence appears con-
siderably smaller In polar view), the membrane by which it is
at first surrounded separates more widely from the egg, finally
ruptures suddenly, and then quickly draws together at one side,
where It Is thrown off as a mass of debris attached to the egg
(Fig. 2).^ Following this, a substance which at first surrounds
the egg as a thin, transparent layer swells up to form a jelly, which
raises the egg slightly from the bottom. The wall of the ger-
minal vesicle breaks down at about this period (20-30 m.), leav-
ing a clearer space In which the first maturation-figure appears.
The white polar areas are still clearly visible, and the egg, still
unfertilized, now gives the appearance of being surrounded by a
very broad, horizontal pigment-ring, which, though often faint
and with vague boundary, is always distinctly visible (Figs. 3,
4). The ring recalls that described by Boverl ('01) in the egg
of Strongylocentrotiis , though relatively broader. The egg of
Dentaliiim thus shows a visible stratification of material analo-
gous to the zones seen In Strongylocentrotiis ; but, unlike the lat-
ter, the zones of Dentaliiim clearly pre-exist before even the pre-
paratory changes of maturation take place.
Sections and total preparations of the flattened egg, fixed shortly
after Its discharge or removal from the ovary, show that a distinct
structural modification exists in each of the white areas, at this
period much more marked in case of the lower or vegetative area.
Surrounding the lower pole (Fig. 10) is a very distinct mass of
dense almost homogeneous protoplasm, of approximately the same
1 All the figures were outlined as accurately as possible with the camera, and
with the exception of Figs. 10-13 and 33, 38-41, are enlarged to the same scale
(150 diameters). They are only schematized in that the pigment is represented
by stippling, whereas the color does not actually appear in the form of dis-
tinct granules, but as a nearly uniform hue. The stippling somewhat exag-
gerates the distinctness of the pigment as seen in most individuals; though in
the most deeply pigmented ones, viewed under strong direct light, the color
appears with great distinctness and its limits may be clearly seen. The opera-
tion of cutting usually leads to disturbances in the arrangement of the pig-
ment, so that frequently no definite color-pattern can be clearly made out in
the dwarf embryos. I have only represented the pigment in cases where its
boundaries could actually be seen.
8 Edmund B. Jf'ilson.
extent as the white area seen in the living egg; this contains no
yolk-spheres, and stains with great intensity with a strong plasma-
stain like Congo red. This mass, sharply marked off from
the surrounding yolk, bulges slightly outward at the surface and
at the margin is continuous with a very thin ectoplasmic zone that
entirely surrounds the egg, but is only clearly visible in sections.
Internally this mass is confluent with a somewhat narrow zone of
similar finely granular protoplasm that extends upwards partly
around the germinal vesicle. It is probably to the presence of
this remarkable protoplasmic mass that the appearance of the
lower white area is due, though the latter may have a different
cause. In a general way, the lower protoplasmic area is un-
doubtedly comparable with the lower zone, composed of green
material, seen in the egg of Myzostoma (Beard, Wheeler, and
Driesch), as is proved by its later history. Comparison of my
Fig. lo with Wheeler's Fig. 2 ('97), will show how closely
similar the relations of the lower protoplasmic area in the two
eggs are.^
The upper white area cannot be distinguished as such in the
fixed eggs, and is apparently produced by a different cause from
the lower one. Exactly at the upper pole is a very small, super-
ficial disc of clear, dense, intensely staining protoplasm, which,
like the lower protoplasmic mass, is continuous at its margin
with the general ectoplasmic layer (Fig. 10), This upper
disc is so small as readily to escape observation; but suf-
ficiently careful examination invariably reveals its presence, which
is furthermore frequently indicated by a slight indentation of
the egg-periphery at this point. It varies considerably in thick-
ness and extent in different specimens, but is always very small
at the beginning.^ Evidently, the upper protoplasmic disc is not
large enough to account for the appearance of the upper white
area in the living egg, which must be due to some other cause.
1 Compare also Driesch, '96, Fig. 12.
2 Sfctions of the ovary show that both the upper disc and the lower proto-
plasms area are present while the egg is still attached to the ovarian wall. The
eggs are greatly distorted in shape, but in a general way are pyriform, and at-
tached by the narrow end. The lower protoplasmic area occupies the narrower
end, by which the eggs are attached ; the upper disc is at the opposite point.
Experimental Studies on Germinal Localization. 9
perhaps to a lighter tint In the deutoplasm In this region. In the
following account, accordingly, it will be necessary always to
distinguish clearly between the upper white area, or polar area,
and the upper protoplasmic disc or area.
%
10
II
12 13
Fig. II.
Vertical Sections of the Normal Egg.
Fig. 13 directly from section (picro-acetic) ; outlines of Figs. 10-12 (sublimate-
acetic) from optical section of total preparations, details from actual sections.
The peripheral zone of deeply staining yolk shown in Fig. 13 occurs in all these
stages after picro-acetic fixation, but not after sublimate-acetic.
10, Unfertilized egg, five minutes after release, showing both protoplasmic
areas; chromosome-like bodies in the nucleolus; 11, fertilized egg, 30 minutes alter
-fertilization , first polar spindle ; 12, fertilized egg, 60 minutes after fertilization ,
initial stage in formation of polar lobe; 13, first cleavage, 68 minutes after fer-
tilization, just before the complete trefoil stage.
10 Edmund B. Wilson.
I shall here give only a very general account of the later his-
tory of the two protoplasmic areas, which will require a thorough
cytological study for its full elucidation. As the egg, still unfer-
tihzed, lies in sea-water, the ectoplasm in the region of the upper
disc slowly increases in amount, and in some cases this region
shows a faintly radiating appearance around its periphery as If
clear hyaloplasm were flowing into it from the surrounding region.
I am uncertain whether in this process the original disc itself
enlarges or is only surrounded by an accumulation of hyaloplasm
— a point of Importance for the comparison with the upper polar
ring of the annelid egg that is drawn further on. I shall continue
to speak of the ectoplasmic thickening at the top of the egg as the
"upper protoplasmic area," but would call attention especially
to the fact that the original disc is composed of very dense homo-
geneous protoplasm that differs markedly in character from th*:
alveolar protoplasm of the ectoplasmic thickening that afterwards
extends over the whole upper surface of the egg.^
When the germinal vesicle breaks down, the maturatlon-splndle,
which is relatively small. Is formed just below this protoplasmic
area, rotating into a radial position and moving towards the
periphery so that Its outer end lies in or just below It (Fig. ii).
In this position it remains, in metaphase, until the egg is fertilized,
when the divisions proceed, the polar bodies being successively
extruded exactly at the upper pole, at the centre of the upper pro-
toplasmic area (which Is now rapidly extending and shows no defi-
nite boundary), and hence at the centre of the upper white area
(Fig. 4). At this period the protoplasmic area comes Into con-
nection by a rather narrow neck of hyaloplasm, in which the spin-
dle lies, with the central mass left after the germinal vesicle breaks
down. After the polar bodies are formed this connection is sev-
ered, and the upper protoplasimc area spreads out still more
1 The general ectoplasmic layer can in the earlier stages hardly be seen in
total preparations, but appears clearly in sections either after staining with
haematoxylin and a strong plasma-stain such as Congo red (when it appears
clear red) or after borax carmine. It is at first much thinner and less defi-
nitely bounded than, for instance, in Rhynchelmis as figured by Vejdovsky,'88 (in
the recent paper of Vejdovsky and Mrazek, '03, it is represented as much thinner
than in the earlier paper), but later becomes very conspicuous.
Experimental Studies on Germinal Localization. 1 1
widely so as to appear as a general thickening of the ectoplasmic
layer over the whole upper hemisphere (Figs. 12, 13). This
thiclcening is most marked near the animal pole, where it is very
conspicuous at the time of cleavage, extending thence approxi-
mately to the equator of the egg, or slightly below it, but without
any very definite margin. It stains deep red in Congo red and
shows a finely alveolar structure quite unlike that of the original
disc.
During the foregoing stages marked changes occur also in the
lower protoplasmic area, and it is evident that active movements
of its material take place. These are perhaps due in part to the
entrance of the spermatozoon at the lower pole, but in part also
to the fact that upon the breaking down of the germinal vesicle
the finely granular material derived from It becomes more or less
definitely confluent with the lower area (as Wheeler describes
in Myzostoma) , so that an irregular pillar of protoplasm, sur-
rounded on all sides by yolk, now extends from the lower pole
nearly to the upper protoplasmic area (Fig.ii) and ultimately
becomes connected with the latter as the first maturation spindle
moves upwards.^ In vertical section it may very clearly be seen
that the material of the upper part of this pillar differs markedly
from the lower, both In texture and in staining capacity (the two
regions show a rather distinct boundary, indicated by the dotted
line in Fig. 11), the lower region being very dense and staining
in the double stain clear red, the upper one much looser (alveo-
lar?) in structure and staining purple or blue. During the polar
body formation the lower area changes its form, often becoming
irregular and sometimes elongate or sickle-shaped. It is a note-
worthy fact that at the time each polar body Is extruded the egg
becomes irregular in contour or almost amoeboid, at the center
of the loiier polar area, afterwards resuming Its even outline.^
After formation of the polar bodies the upper part of the proto-
plasmic pillar retreats from the periphery, while the yolk again
extends across the upper region above the egg-nucleus. In the
upper part of the internal protoplasmic region conjugation of the
1 Cf. Wheeler's Fig. 10 or 16.
2 This was figured by Lacaze Duthiers ('57 ) nearly fifty years ago.
1 2 Edmund B. JVilson.
germ-nuclei takes place. At the period shortly preceding the
first cleavage, when the upper disc has been replaced by the very
broad ectoplasmic thickening described above, the lower proto-
plasmic area, as seen in surface views of total preparations, varies
a good deal in appearance in different individuals, being some-
times rounded and fairly well circumscribed, sometimes irregular,
or even broken up so as to present a mottled appearance.
The first cleavage, which occurs about thirty minutes after the
extrusion of the second polar body, is characterized by a trefoil
stage, like that occuring in many gasteropods, lamellibranchs
and annelids (Figs. 5, 6). Exactly surrounding the lower pole
is formed, by a horizontal constriction, a large lobe, into which
passes the whole of the lower white polar area, and M^hich, like
the area itself, appears pure white in the living object. Since
the surface of the lobe is much larger than that of the original
lower polar area from which it arises, it is evident that material
from the interior of the egg must How into the lobe as it form.s.
Vertical sections of the egg as the polar lobe begins to form show
somewhat varying appearances, due in part to differences in the
plane of section, but also in part to varying conditions in the
protoplasmic area itself. The rather small cleavage-figure, at
this period entirely surrounded by deutoplasm, lies in late ana-
phase or early telophase slightly above the centre of the egg.
At the lower pole the dense protoplasm of the lower area is now
spread out, more or less irregularly, to form a thick peripheral
layer that fades away insensibly into the yolk-bearing region.
Frequently, as in Fig. 12 {cf. Wheeler's Fig. 46) this thickening
appears fairly regular and symmetrical and suggests the ecto-
plasmic thickening that precedes the formation of a pseudopod
in Amceha; sometimes it is less regular than this, and occasion-
ally gives the appearance of an asymmetrical wedge-shaped mass
extending into the yolk. As the lobe forms it receives this clear
protoplasm, accompanied by an inflow of yolk that seems to in-
vade the clear substance more or less; so that in section scattered
yolk-granules are found in the lobe and frequently no definite
boundary of the clear substance can be distinguished (Fig. 13).
In any case it is certain that the whole of the lower protoplasmic
Experimental Studies en Germinal Localization, 13
area passes into the lobe (like the green material of the Myzos-
tovia egg) to constitute its main bulk, precisely as Wheeler shows
in Myzostoma {cf. his Fig 47). The term "yolk-lobe" em-
ployed by a number of earlier observers is therefore as mislead-
ing as it is inappropriate and may be replaced by the term "polar
lobe." For reasons given in the discussion at the end, I believe
it very probable that at least the lower protoplasmic area, and
probably also the upper disc, are in a general way comparable to,
if not identical with, the polar rings observed in the eggs of cer-
tain leeches and oligochaetes.
Immediately after the polar lobe is formed a vertical furrow
cuts into the egg from the upper pole, dividing the upper white
area into equal parts and forming with the polar lobe a trefoil,
of which the two upper lobes are of exactly equal size and contain
all of the pigment, while the unpigmented polar lobe is consider-
ably less than half the bulk of each of the others (measurements
give a ratio of i to 0.32-0.46, Fig. 6) . At the height of its form-
ation the trefoil appears at first sight to consist of three separate
spheres. Close examination invariably shows however that the po-
lar lobe Is united to one of the upper lobes by a very narrow pedicle
which is never severed; and as the cleavage proceeds these two
lobes completely fuse while the remaining upper lobe Is cut off as
a separate blastomere. Thus Is formed a characteristic unequal
2-cell stage (Fig. 7), consisting of a smaller anterior cell, AB,
and a larger posterior one, CD, which differ In volume by ex-
actly the bulk of the polar lobe. Each of these cells has at the
upper pole a white area, representing half the original upper
polar area. The lower polar area, on the other hand, is con-
fined to the larger cell, and obviously represents that part of the
substance of the fused polar lobe that appears at the surface, a
part having again moved into the interior of the egg.^ Upon
the 2-cell stage thus formed is moulded the entire subsequent
development, which in Its general outline Is of essentially the same
type as In such forms as Unio or Nereis.
The experiments recorded In this paper relate mainly to the
significance of the material of the lower polar area, and of the
polar lobe, and form a continuation of those begun by Crampton
1 Cf. Wheeler's Fig.
H
Edmund B. Wilson.
in his interesting experimental paper on Ilyanassa, published in
1896. In order to understand the significance of the experiments
to be described it will be necessary to trace briefly the subsequent
development. The second cleavage is ushered in by the reap-
pearance of the polar lobe at the vegetative pole of the larger
cell, CD, of the same size and form as before, and again consist-
ing entirely of white material (Figs. 8, 9). The cleavage in
this cell, whether separated from its fellow or remaining united
Fig. III.
Cleavage, from living Eggs.
14, Four-cell stage, from lower pole; 15, beginning of third cleavage, from
lower pole, third polar lobe; 16, eight-cell stage, from lower pole; 17, beginning
of fourth cleavage , first somatoblast in formation ; 18, sixteen-cell stage, from
lower pole; 19, view from lower pole, after the formation of the third quartet;
19a, D (pigmented) and 4d, immediately after division ; surface view.
with it, follows the same general course as in the first cleavage of
the entire egg, the polar lobe finally fusing with one of the cells,
namely, D, the left posterior quadrant, where it again forms a
very definite lower polar white area.^ The anterior cell, AB,
1 Cf. Wheeler's Fig. 49, Driesch's ('96 ) Fig. 12, of Mysostoma.
Experimental Studies on Germinal Localization. 15
in the meantime divides equally, without the formation of a polar
lobe. In the 4-eell stage, accordingly, the large posterior cell,
D, exceeds A, B or C, by exactly the volume of the lobe, and
the lower white area appears only in D (Fig. 14). On the other
hand, the substance of the original upper white area is equally
distributed among the four; but it is evident that the amount of
white material visible at the surface has somewhat increased.
The 4-cell stage shows the characteristic relations of the blasto-
meres observed in so many other eggs of this type. The two
lateral cells, A and C, He at a higher plane, and are In contact
along the upper side by an upper "cross-furrow." B and D, on
the other hand, are in contact along a longer transverse lower
cross-furrow; and these characters, together with the large size
of the posterior cell, D, thus give an immediate means of orienta-
tion from this time forwards.
As the egg prepares for the third cleavage the upper white
material shifts slightly towards the left upper angle in each quad-
rant, anticipating the formation of the first quartet of ectomeres
by the usual dexlotropic cleavage. These cells, which are of
equal size and In the A, B and C quadrants are not much smaller
than the basals, are formed entirely from the white material of
the upper polar areas; and it Is here again evident that an ex-
tensive flow of this material must take place from the interior
of the egg. Their formation does not, however, exhaust the
white substance of the upper areas, which still remain In the
upper regions of the four basals. During this division the polar
lobe forms for the third and last time, from the white material
of the lower area, in the D-quadrant; but it Is now noticeably
smaller than before, and does not constrict so deeply (Fig. 15).
After the completion of the cleavage the lobe again fuses with
D, in which, as the egg enters Into the "resting stage," the lower
white area still appears; though this soon undergoes a great
change (Fig. 16).
The fourth cleavage is of especial interest, since a large part
of the substance of the lower white area now passes into the first
somatoblast, id, or X, and Is thus for the first time actually cut
off from the pigmented region. This cleavage is preceded and
1 6 Edmund B. Wilson.
accompanied by an extensive shifting of the cytoplasmic mater-
ials in all of the cells. In the three basals, A, B and C, the white
material towards the animal pole moves over towards the upper
right angle of the cell and Increases in amount, extending so far
down the egg that in some individuals it may be seen, when the
egg is viewed from the vegetative pole, as a narrow white cres-
centic area (Fig. i6). A similar process takes place in D, but
in addition to this a great change takes place in the white material
of the lower polar area, which leaves its position at the lower pole,
moves over towards the same side as the upper white area, and
finally fuses with it, while the pigmented part becomes lighter in
color, often irregular or mottled in appearance, and extends into
the area formerly occupied by the lower white substance. In
the ensuing cleavage, D is usually the first to divide, giving
rise by a leiotropic cleavage to the large first somatoblast, 2d or
X (Figs. 17, 18). This cell consists almost entirely of white
material which is certainly derived in large part from the orig-
inal lower white area, but undoubtedly also in part from the upper
white area, which, as stated above, fuses with the lower area in
the period preceding this cleavage. In some cases X receives
also a small amount of the pigment (Fig. 18), in others it seems
to be composed entirely of white material. The other members
of the second quartet, 2a, 2b, and 2c, are much smaller than X,
and each is formed mainly from the white material of the upper
polar area, but as a rule, perhaps always, each receives also a
variable amount of pigment. During the foregoing changes the
upper quartet divide leiotropically in the usual fashion, to form
the four primary trochoblasts, which are slightly smaller than
the upper cells. Owing to the foregoing changes the pigment,
which in the unsegmented egg extended far up towards the animal
pole, has been moved downwards so as to lie below the
equator of the egg, most of it being contained in A, B and
C, some in D, a little in 2a, 2b and 2C, and sometimes also a little
in 2d. The pigment becomes still more restricted during the
fifth cleavage, since the micromeres of the third quartet are again
mainly composed of white substance.
Experimental Studies on Germinal Localization. 17
The fifth cleavage, dexlotropic in all the cells, produces the third
quartet, each cell of which is considerably smaller than the corre-
sponding basal (Fig. 19). Qf these cells 3d is much the largest,
and is usually composed entirely of white material, while 3a, 3b
and 3c usually, perhaps always, receive a certain amount of
pigment. At the end of the cleavage the macromeres rapidly
diminish in apparent size, evidently owing to their passing more
deeply into the egg, and the color-pattern becomes more or less
confused, though A, B and C still show the greatest amount of
pigment, while D distinctly shows a white area on the side turned
towards X, where 4d is subsequently formed. I have not been
able to observe the formation of the entire fourth quartet satis-
factorily, either in the opaque living object or in preparations.
I can however state positively that as seen in surface-view of the
living egg, 4d is very small (smaller than 3d and very much
smaller than 2d) and appears pure white (Fig 19, a). I have
been unable to determine whether the white material of this cell
is derived from that of the original lower white area; though,
as will appear hereafter, the experimental evidence indicates that
such is the case. At this period the four basals appear much
smaller, having evidently retreated into the interior.
Beyond this point it is not necessary at this time to trace the
cleavage. The foregoing observations clearly show that, in Den-
talium the freshly discharged egg, prior to maturation or fer-
tilization, shows a definite segregation of zisibly different ma-
terials which accurately foreshadows a corresponding distribution
of these materials among the hlastomcres during cleavage. Of
the three zones of material superficially visible in the living egg,
the upper one (upper white area) is allotted to the first three
quartets of ectomeres, apparently in equal amount in each quad-
rant; the middle pigmented zone is mainly allotted to the four
basal entomeres, though a portion also passes into ectomeres of
the second and third quartets; while the lower zone (lower white
area) certainly passes mainly into the first somatoblast, 2d, or
X, probably in part into the second somatoblast, 4d, or M,
and possibly in part into the left posterior micromere, 3d, of the
third quartet. This agrees in general with the history of the
1 8 Edmund B. Wilson.
zones visible in the living egg of Myzostoma, as observed by
Driesch ('96), where the lower polar area is represented by a
green substance, the upper one by a reddish material, and the
pigment zone of Dentalium by a zone of clear protoplasm. It
is important not to confuse the above-described distribution of
white and pigmented material with that of protoplasm and deu-
toplasm. As shown on a preceding page the upper white area is
not, like the lower one, free from yolk; and in point of fact all
the cells contain a large amount of yolk. The pigment-pattern is
only a visible expression in the living object of a distribution of
specific materials that can only in part be distinguished in sections.
We may now briefly consider the main outlines of the larval
development. In warm weather the embryos become ciliated at
about the ninth or tenth hour, and at the end of twenty-four hours
are well developed trochophores that swim very actively at the
surface, progressing in a spiral curve and rotating from right to
left as seen from the side. At this period (Fig. 29) the body is
of a blunt spindle-shape, encircled at the equator by a very broad
prototroch composed of three principal rows of large trocho-
blasts which bear three corresponding rows of powerful cilia
completely encircling the body and leaving no dorsal gap (as is
also the case in Patella). The pre-trochal and post-trochal re-
gions, while somewhat variable, are at this period nearly similar
in form and size, being roughly conical and rounded at the tip.
The pre-trochal region is wholly covered with very short vibra-
tile cilia and bears at its apex a very long and well-defined tuft
of flexible, but not vibratile, flagelliform sensory hairs. In total
preparations, or in longitudinal sections, it may be seen with
great clearness that the apical tuft is borne upon a large and
definitely circumscribed apical thickening or plate, sharply marked
off from the surrounding cells. The post-trochal region is not
ciliated, but bears at its posterior extremity a small bunch of
sensory hairs, which differ from those of the apical tuft in being
quite stiff, and radiating from the common point of attachment.
The ahmentary canal at this period forms a closed sac divided
into two chambers, into one of which at a slightly later period
opens the mouth, formed immediately below the prototroch, but
Experimental Studies on Germinal Localization. 19
the anus does not yet exist. The post-trochal region already shows
the mantle fold and the beginning of the shell-gland. On either
side the gut may be seen an irregular mass of small cells which
I believe to represent the coelomesoblast, though I have not yet
traced them to the pole-cells. These masses are not to be con-
founded with two masses lying further forward that are pro-
liferated off from the ectoblast in two symmetrically placed lateral
areas in the pre-trochal region and perhaps represent a part of
the paedomesoblast (ectomesoblast) or perhaps the foundations
of* the cerebral ganglia. These areas, which are figured by Ko-
welevsky ('83, Figs. 32, 37, 55) are shown in the lobeless em-
bryos (Figs. 33, 40).
The ensuing changes take place very much more rapidly In
the Naples species (D. entalis) than in the northern form stud-
ied by Lacaze Duthiers ('57), which is probably due in a meas-
ure to the higher temperature. By the 30th hour the post-trochal
region has considerably elongated and the pre-trochal region is
somewhat diminished (Fig. 30). In the course of the ensuing
twelve hours the pre-trochal region wholly disappears from view,
being withdrawn into the interior, while the post-trochal region
becomes still more elongated and the larva sinks to the bottom,
where it swims only sluggishly. About this time the body becomes
surrounded by .an extremely delicate hyaline shell into which
the greatly diminished prototroch can be withdrawn; and by
the end of the second day the foot appears on the median ventral
side. By the end of the third day the foot has become a large
protrusible organ, trilobed towards the free end, and the pro-
trotroch is still smaller (Fig. 31, which closely agrees with
Lacaze's Fig. i, Plate VIII). In many cases the metamorphosis
is complete by the end of the fifth day, the prototroch having
disappeared, the otocysts and pedal ganglia being clearly visible,
and the young Dentaliiim assumes the condition figured by Lacaze
on Plate 8, Figs. 2, 3 — a larva of 20-25 days ( !).
Many details have been omitted from the above account that
have already been described In the well-known memoirs of Lacaze
Duthiers ('57) and Kowalevsky ('83). Many others will re-
quire for their full elucidation much more extended study than
20 Edmund B. Wilson.
I have thus far been able to devote to the subject. The greatest
gap in my work thus far is the failure to trace the connected
history of the mesoblast, which can only be done by a complete
study of the cell-lineage. This presents considerable obstacles
owing to the difficulty of obtaining good total preparations at
every stage (the eggs and embryos stain diffusely in most dyes, and
the great abundance of deeply staining yolk in all the cells renders
it difficult to get clear pictures) , and my time was so taken up with
the study of the living material that I had not opportunity to
work out a really satisfactory method. For sectioning the best
results were given by sublimate-acetic, the sections being stained
with thionin, which gives a sharp nuclear stain without coloring
the yolk. The best total preparations were obtained by mount-
ing in balsam without staining. Apart from the technical diffi-
culties, the object is Itself difficult, in the earlier larval stages on
account of the difficulty of distinguishing between mesoblastic
and entoblastic elements in the crowded mesentoblast-mass, in the
later ones by reason of the complication introduced by the folding
of the mantle and the shell-gland.
EXPERIMENTAL PART.
The ease with which the eggs of Dentaluim may be operated
recalls the remark of Lacaze that "L'embryon du Dentale est un
de ces exemples f aits pour I'etude du developpement" ('57, p. 196).
For experimental purposes however it presents certain difficulties
that should carefully be borne in mind in considering the results
of the operations. First, there is a certain amount of variation,
not wide but still noticeable, in the size of the eggs and the re-
sulting larvae, and in the relative size of the polar lobe and of the
blastomeres during the cleavage-stages. Second, a certain pro-
portion of the entire eggs sooner or later develop abnormally,
which results in an Increasing mortality from day to day. Third,
and most important, the percentage of monstrous forms, and the
mortality. Is always very large in the development of egg-frag-
ments and of Isolated blastomeres. This Is undoubtedly due in
part to the abnormal conditions under which the larvae are placed
In the aquarium, In part to the shock of the operation, and in part
Experimental Studies on Germinal Localization. 21
to the changed condition of surface-tension in the dwarf embryos
and larvae, as is shown by the readiness with which they dis-
integrate. (I have several times seen an actively SAvimming dwarf
larva suddenly fly to pieces on coming in contact with an obstacle
or even with the surface of the water.) For these reasons, de-
spite the great ease with which the eggs may be operated, it is
difficult to base trustworthy conclusions regarding the more spe-
cial features of the egg-localization on the defects observ^ed in
the individual partial larvae. I have therefore in the following
work restricted my account in the main to the results that appear
with unmistakable clearness, and appear in so large a proportion
of the larvae as to remove all reasonable doubt. Beyond this,
owing to the importance of following the development of the
living larvae as far as possible, the number preserved for section-
ing was not very large, and the technical difficulties indicated
above, in case of the normal larvae, here appear in aggravated
form. This explanation is necessary to account for certain ob-
vious gaps in the work, which I hope to fill out by further in-
vestigation, especially those relating to the mesoblast, regarding
which I can at present offer only somewhat provisional conclu-
sions.
III.
EFFECT OF REMOVING THE POLAR LOBE.
(a) General History of the lobeless Larvae. — During the
trefoil stage of the first cleavage the polar lobe may easily be
removed, wholly in part, by means of a fine scalpel. Complete
removal of the lobe produces a highly characteristic and constant,
though in one respect very unexpected, result. Exactly as Cramp-
ton earlier found in Ilyanassa, the egg continues to segment after
this operation quite symmetrically, in a manner similar to the
normal cleavage of such forms as Patella or Lymnaea, giving rise
by typically alternating spiral cleavages to successive symmetrical
quartets of micromeres (Figs. 20-26). These cleavages differ
constantly in two respects from the normal, namely, that ( i )
no trace of a polar lobe is formed at either the second or the
third cleavage, and (2) the members of the D-quadrant are no
22
Edmund B. Wilson.
larger than the others. Correlated with this is the fact that these
embryos show no lower white area, all the basal quadrants being
uniformly pigmented over the lower pole (Figs. 23, 24), which
sometimes shows a large opening into the cleavage-cavity (Fig.
Fig. IV.
Cleavage after Removal of the Polar Lobe.
20, Two-cell stage and polar lobe after removal of the latter ; 21, four-
cell stage of same, from upper pole ; 22, eight-cell stage of same, from upper
pole ; 23, sixteen-cell stage of lobeless embryo from lower pole , symmetrical
second quartet; 24, similar view of the same stage, open type; 25, sixteen-cell
stage, from upper pole ; 26, lobeless embryo from lower pole, after formation
of the third quartet; 27, second cleavage, from the side, of egg from which
about three- fourths of the first polar lobe had been removed ; 28, a similar
form, viewed from the lower pole, after removal of about one-half of the first
lobe.
Experimental Studies on Germinal Localization. 23
24) . The embryos gastrulate and develop with great regularity
into larvae that swim in the same characteristic progressive spiral
course as that of the normal ones. These larvae (Fig. 32) differ
from the normal ones in two obvious respects, namely, ( i ) the
post-trochal region is absent, or represented only by a smoothly
rounded surface from which no outgrowth takes place, and (2)
they show no trace of an apical organ. . The first of these results
fully accords with expectation; for studies in cell-lineage have
shown, both in annelids and in mollusks, that in forms possessing
a typical trochophore larva the ectoblast and mesoblast of the post-
trochal region are mainly derived from the two somatoblasts,
and I have shown that the first of these cells is certainly and the
second probably, derived mainly from the polar lobe (or lower
white area). The second result, on the other hand, is astonish-
ing, since the region that has been removed is diametrically oppo-
site to that from which the apical organ develops; but a large
number of operations have not shown one exception in this re-
spect and the most convincing corroborative evidence is afforded
by other experiments presently to be described.
The structure and subsequent history of these larvae is very
widely different from that of the normal forms. As the cleavage
advances the symmetrical cells of the second and third quartets
close in around the lower pole, frequently followed in greater
or less degree by the cells of the prototroch; and after the gas-
trulation this region (the posterior region of the larva) becomes
somewhat expanded, so that the larva assumes a pyriform shape,
actvely swimming with the narrower end in front, and rotating
from right to left like a normal larva. The narrower anterior
region is uniformly covered with fine vibratile cilia which are
slightly longer near the anterior pole (as in a normal larva —
Fig. 32) ; but an examination of more than fifty such larvae
failed to show a single case in which a true apical tuft was present.
Sections and total preparations reveal the remarkable additional
fact that in such larvae, at least in many cases, no apical plate is
formed, though the lateral areas of proliferation, referred to
above, are present, as shown in Fig. 40, a, a. In a few cases
I have found a somewhat vague thickening at the apical pole,
24
Edmund B. JVils
on.
Experimental Studies on Germinal Localization. 25
Fig. V.
Normal Mefamorpliis and lobeless Larvae.
(Excepting Fig. 33 these figures were drawn from living larva, the cilia
being added from formol preparations and the inner outlines from specimens
mounted in balsam.)
29, Normal trochophore of 24 hours (a rather large specimen) ; 30, normal
trochophore of 32 hours; 31, normal larva of 72 hours, showing foot and shell;
2)2, larva of 24 hours, after removal of first polar lobe ; 33, vertical section of
lobeless larva of 24 hours, showing entoblast-plug protruding through the
blastopore; 34, larva of 72 hours, after removal of first polar lobe; 35I larva of
24 hours, produced from a form like Fig. 28, after removal of about half the
polar lobe ; 36, larva of 24 hours, after removal of second polar lobe ; 37, CD
half-larva, after removal of second polar lobe, 24 hours.
1 This figure has been turned upside down by the engraver.
26 Edmund B. Wilson.
but never one that could be mistaken for a typical apical plate.
In others, however, the apical ectoderm does not differ from that
by which the whole pre-trochal region is surrounded. I feel
justified therefore in the statement that the lobeless larvae typically
fail to develop the apical organ at any period, individuals having
been reared up to the fourth day, when the metamorphosis of the
normal larvae was well advanced. {Cf. Figs. 31 and 34.) Dur-
ing the development, probably owing to the deficiency of material
present in the D-quadrant, the trochoblasts often become more or
less displaced towards the posterior pole, and in greater or less
degree lose their regular arrangement. In many specimens never-
theless the typical prototrochal belt of three rows of cilia is formed
(Fig. 34) , though even in these the rounded posterior region often
also bears patches of cilia. In others no definite belt can be made
out, and such individuals often give the appearance, when alive,
and even after being killed with formol, of being ciliated over the
whole posterior region. In preparations, however, the cilia of
such forms may almost always be seen to be arranged in patches,
leaving non-ciliated regions between them, which are doubtless
occupied by cells derived from the second and third quartets.
It is probable, therefore, that the appearance of uniform clllation
is misleading, and is caused by the confusion of separate tufts
lying at different levels. In cases where no displacement of the
trochoblasts occurs, the posterior region is covered by cells de-
rived from the second and third quartets.
As the development proceeds there is no attempt to regenerate
the missing post-trochal region or apical organ, and the later
history of these larvae differs totally from that of the normal
ones. The pre-trochal region shows an increase, instead of a
decrease, in size, and is not withdrawn Into the interior, but gives
rise to a more or less irregular vesicular structure directed for-
wards as the embryo swims. Such larvae were reared until the
beginning of the fourth day (Fig. 34), after which they in-
variably became more and more irregular and finally disinte-
grated. At this period they present a most remarkable contrast
to the normal control larvae of the same age. There is still no
trace of a post-trochal region, no shell, no foot, and no apical
Experimental Studies on Germinal Localization. 27
organ. Sections show that these larvae have formed no shell-
gland, no mantle-fold, and apparently also no mouth.
The foregoing account applies to the great majority of the
lobeless larvae; but occasionally an apparent exception occurs, the
careful examination of which only serves to confirm the rule.
In these exceptional cases a more or less reduced post-trochal
region appears to be present, and one individual was obtained
that in life seemed to possess this region in a fully developed con-
dition. Sections of these embryos show, however, that what
appears to be a post-trochal region is in reality a plug of ento-
blast cells, projecting through the blastopore-region, that arises
through defective gastrulation (Fig. 33). Such embryos some-
times show towards the upper pole a much larger cleavage-cavity
than in the normal form, — obviously a result of the failure of
the entoblast-cells to invaginate completely. This is conspicuously
shown in the larva, referred to above, which appeared to have a
fully developed post-trochal region. This larva, cut into longi-
tudinal serial sections, shows very clearly the failure of the ento-
blast-cells to invaginate properly, a large space being left in the
upper hemisphere above the archenteron. For this very reason
this larva showed very clearly, both as a total preparation and
after sectioning, the entire lack of an apical organ.
The foregoing observations fully establish the conclusion, I
believe, that the material of the polar lobe is indispensable for the
formation of the post-trochal region and the apical organ, and as
shown beyond they give considerable reason for extending this
conclusion also to the ccelomesoblast. . That the failure to produce
a normal larva is not due to the lack of sufficient material, is con-
clusively shown by several additional facts. First, in Patella
the D-quadrant is no larger than the others, yet a post-trochal
region is formed that is relatively as large as in Dentalium.
Second, as will be described in Part V, much smaller larvae, pos-
sessing all of the typical parts, may be produced from fertilized
egg-fragments. Third, the same conclusion is afforded by the
history of isolated blastomeres, which also fully corroborates the
results obtained by removing the polar lobe from an entire egg.
If in the 2-cell stage the two blastomeres, AB and CD, be sep-
28 Edmund B. Jfilscn.
arated, both continue to segment for a time as if still forming
part of an entire embryo, the second and third polar lobes form-
ing in normal fashion in the CD half; but in the end both com-
pletely close, gastrulate, and form activ^ely swimming larvae. The
two larvae agree in possessing a closed, though often somewhat
asymmetrical or confused prototroch, but otherwise show the fol-
lowing characteristic and constant differences. The AB (smaller)
larva, closely resembles, except in size, that derived from an
entire egg from which the polar lobe has been removed, invaria-
bly lacking a post-trochal region and apical organ (Fig. 46).
The CD (larger) larva, on the other hand, possesses both these
structures, both of which may be as large as in a whole embryo
(Figs. 42-45). These larvae vary greatly in form, but in gen-
eral are asymmetrical and, as may be seen by a comparison of
Figs. 45 and 29, possess a post-trochal region that is almost in-
variably relatively too large, and a pre-trochal region relatively
too small as compared with a normal larva. As in the AB half,
the prototrochal cilia frequently show a confused arrangement,
the regular rings of the normal larva being more or less broken
up. In like manner, if the four blastomeres of the 4-cell stage
be isolated, only the larva from the D (largest) quadrant develops
these two structures (Fig. 47), while those from A, B or C are
nearly like those derived from the AB half, though only half as
large (Figs. 48, 51). Like the CD ^-larvae the D ^^-forms are
variable in form; but whenever they complete what may be con-
sidered their normal development they show the post-trochal region
very much too large, and the pre-trochal region much too small
(Fig. 47)-
All these larvae show a very high mortalit}^ but I have kept the
^-larvae as late as the beginning of the fourth day ( Fig. 51), and
the 54-larvae nearly as long. The smaller larvae (the AB half,
or the small quarters) show a greater tenacity of life, swim more
actively, and become less irregular than the larger ones. In the
end, however, all the forms become irregular and finally wholly
disintegrate, without producing normally formed trochophores or
regenerating the missing structures. The CD ^-larvae of 24
hours sometimes approach the form of normal larvae of the same
age, though always showing the false proportions of the pre-tro-
Experimental Studies on Germinal Localization. 29
chal and post-trochal regions described above. Like the AB halves
and the j4-larvae, they often swim actively at the surface, rotating
in the same way as an entire larva; though the progressive move-
ment is almost always slower and less regular than that of the
smaller halves. These forms, however, seem to live no longer than
the less regular ones, and in spite of every precaution they become
more and more irregular and finally disintegrate in the same
aquaria containing the normally developing whole larvae. Those
that lived to the end of the second day invariably became mon-
strous in form and showed no resemblance to a normal larva. The
history of the AB halves or the smaller quarters in general very
closely resembles that of the lobeless larvae, the pre-trochal
region enlarging, becoming irregular, and finally disintegrating,
often while the embryo is still actively swimming by means of
the trochoblasts, which, as Fischel has observed in case of the
swimming cells of ctenophores, are most tenacious of life of all
the cells.
The relative volumes of protoplasmic substance contained by
these various forms of larvae, may be determined either by meas-
uring the volumes of the blastomeres after isolation by means of
calcium-free sea-water, or by measuring the polar lobe and es-
timating the other volumes, the two methods giving fairly consist-
ent results. It should be remembered, howev^er, that both the
whole eggs and the relative size of the polar lobe (and hence of
the blastomeres) vary somewhat, both in the eggs produced by
a single female, and to some extent in those produced by dif-
ferent females. I observed one. lot of eggs, for instance, the
greater number of which produced lobes considerably smaller than
usual. Measurements of the lobe in typical average trefoils give
a value ranging from one-fifth to one-sixth that of an entire
egg. A typical case gave a volume of 0.18 for the lobe, from
which the other volumes are as follows :
Entire embryo i-oo
Embryo without polar lobe 0.82
CD y2 embryo 0.59
AB y2 embryo 0.41
D y^ embryo 0.385
, A, B or C 0.205
30 Edmund B. Wilson.
Since the CD >4 larva is less than ^ and the D }i larva less
than ^ the volume of the lobeless embryo, yet both produce
apical organ and post-trochal region, the conclusion is unavoid-
able that the failure to form these structures after removal of
the polar lobe must be due to a qualitative and not a quantitative
difference; in other words, the material of the lobe must be spe-
cifically different from the remaining material, and as such is
the determining cause of the development of the structures in
question.
The above conclusion is fully sustained by the effect of cutting
off only a part of the polar lobe. In such embryos during the
second and third cleavages the polar lobe is correspondingly
diminished in size (Figs. 27, 28), and the D-quadrant is too
small by the same amount. Such eggs produce larvae with a cor-
responding reduction in the post-trochal region (Fig. 35) and
these larvae sometimes possess, sometimes lack, the apical organ.
It is not improbable therefore that further experiments of this
kind may show a localization, within the polar lobe itself, of the
determining materials of the apical organ and of the post-trochal
region. This experiment adds to the foregoing the important
result that after the polar lobe has formed there is a direct quan-
titative relation between the amount of specific material it con-
tains and the size of the post-trochal region, there being appar-
ently no regulative process in the later stages (though I have not
yet sufficiently examined this latter point). As will appear in
Part V, this conclusion does not apply to the material of the lower
polar area before the formation of the lobe.
(b) The mesoblast question. — We may now consider what
is in some respects the most interesting, as it is certainly
the most difficult, of the questions relating to the lobeless
larvae, namely, that of the mesoblast. The fact that cer-
tainly the first and probably the second somatoblast is de-
rived mainly from the substance of the polar lobe, and
that after the removal of this substance the post-trochal region
fails to develop, suggests that the material of the coelomesoblast
as well as of the ectoblastic structures, is localized in the polar
lobe and hence in the original polar area. In point of fact
Experimental Studies on Germinal Localization. 31
Crampton ('96) in his interesting paper on Ilyanasssa, found
that after removal of the polar lobe the second somatoblast (4d)
differs from the normal not only in being no larger than the
other members of the quartet, but also in texture, being filled
with yolk-spheres instead of being mainly composed of clear
protoplasm as in the normal, and it also lies at first at the sur-
face, exactly like 4a, 4b and 4c. This observation I can confirm
from a reexamination of the original preparations, kindly placed
at my disposition for this purpose by Dr. Crampton. He found
further, that the larvae produced from such eggs lacked the
mesoblast-bands present in the normal larva, 4d apparently enter-
ing, like its fellow-members of the same quartet, into the forma-
tion of the archenteron.
This highly interesting result, which has atracted considerable
attention, was based on the examination of total preparations
only; and the desirability of a more adequate study of the matter
by means of sections has long been obvious. I have accordingly
given especial attention to this point as far as my material would
allow; but must admit that neither in point of abundance nor
of fixation is this material quite adequate for the full investigation
of the question, which indeed would demand a complete study of
the cell-lineage, both in the normal and in the lobeless forms.
Nevertheless such evidence as I hav^e obtained is distinctly in
favor of the correctness of Crampton's result.
The mesoblast may be most clearly seen in the normal larvae
in cross sections through the region of the prototroch, where the
gut shows two chambers and the complication produced further
back by the shell-gland and mantle-folds are not present. In such
a section (Fig. 38) the gut appears in the form of two distinct
chambers, the wall of the ventral one being a little further back
intimately connected with the stomodaeal invagination (Fig. 39')
though its cavity does not yet appear to communicate with the
outside. The walls of both chambers are composed of large
cells, more or less columnar and radially disposed, completely
filled with yolk-spheres (as are all the cells at this time) and with
large nuclei. On either side is a loose group of much smaller
cells with small nuclei, that appear irregular or often spindle-
32
Edmund B. Wilson.
Fig. VI.
Sections of normal and lobeless Larvae.
(Each of these is drawn from a single section, supplemented by a few details
from the two adjacent sections of the series. The deutoplasm is only shown in
the entoblast and mesoblast.)
38, Slightly oblique cross-section of normal larva, 24 hours, just anterior to
the mouth ; 39, cross-section through the mouth ; 40, vertical section of lobeless
larva, 30 hours; 41, cross-section through prototroch-region of lobeless larva,
48 hours.
Experimental Studies on Germinal Localization. 33
shaped. There can, I think, be no doubt that these are meso-
blast cells,^ though I have not determined whether they are the
products of the second somatoblast, 4d, or arise from another
source. A possibility of error on this point is given by the fact,
already referred to, that just anterior to the prototroch on either
side are two lateral ectoblastic areas of proliferation (of unknown
significance) that may contribute to the small cells in question.
In any case these lateral masses of mesoblast fail to appear in the
lobeless embryos of corresponding age or older. . In the earlier
stages, of which Fig. 33 is an example, it is impossible to de-
termine this point with any degree of certainty, owing to the
crowding together of the entoblast cells in a compact mass in
which frequently no cavity can be seen. In later stages, however,
both longitudinal and transverse sections give pretty clear evi-
dence that the small mesoblast-cells are either wholly absent or
very few in number. Fig. 40 is from a complete series of lon-
gitudinal sections of a lobeless embryo of 30 hours. This shows
the gut as a two-chambered sac directly applied to the ectoblast
with no sign of smaller cells between them, though both the
anterior ectoblastic areas of proliferation are shown (a, a). It
might well be supposed that the small cells are present in a dif-
ferent plane, as would be the case in Fig. 38 if cut in the sagittal
plane ; but their absence appears no less clearly in cross-section, as
shown in Fig. 41 (from a complete transverse series). This
embryo of 48 hours swam actively and normally. Though not
so well fixed as the preceding one, it clearly shows the gut as a
simple sac, enclosing a single cavity that opens at the posterior pole
and anteriorly is nearly filled with a thickening bulging inward
from the wall at one side. I am quite sure that no mesoblast-
cells are present in this embryo unless at the extreme anterior end,
where the layers are cut tangentially and cannot be clearly an-
alyzed. The sections of this embryo clearly show further
1 The relations as figured by Kowalewsky ('83, Fig. 48) in the Marseilles
species are essentially similar to those here shown, except that the mesoblast-
cells are shown very much larger and fewer. This is stated to be from a larva
of 24 hours, but probably represents a relatively earlier stage of development
than mine. Compare the mesoblast-cells in Kowalewsky's Fig. 66, from a larva
of 38 hours.
34 Edmund B. JVilson.
the absence of any structure comparable with the foot, mantle-
folds, shell-gland, or mouth (unless the posterior opening can be
so considered) though all these structures are present in the
normal control embryos. The absence of an apical organ is
shown as in other series, by the two from which Figs. 33 and 40
are taken.
I would not speak too positively before examining additional
material, for in some of the other series a few small cells appear
that may be of the same nature as those seen in the normal
embryos, though they are far less numerous; yet the foregoing
evidence is sufficient to create a strong presumption that Cramp-
ton's result was correct. Crampton showed due caution In guard-
ing against the conclusion, from his observations, that the polar
lobe "contains prelocallzed mesoblast material," being probably
Influenced by the fact that in Ilyanassa the lobe appears to be
composed mainly of deutoplasm. He only concluded "that the
presence of the yolk mass In the cell D may be the stimulus which
causes that cell to act differently from the other macromeres, A,
B and C" ('96, p. 14). I believe, however, the facts brought
forward In this paper render it probable that the polar lobe (and
hence the cell D) does in fact contain a specific kind of cytoplasm
which. If not actually "prelocallzed mesoblast-material" is f.he
direct and necessary antecedent of that material.
IV.
LOCALIZATION OF THE APICAL ORGAN AND ITS CORRELATION
WITH THE POST-TROCHAL REGION.
The failure of the AB half-larva to produce an apical organ,
though wholly consistent with the history of the lobeless embryos,
was to me a surprising fact; for the development of this organ
in other forms Indicates that all of the four quadrants contribute
to Its formation ; and in point of fact I had found in Patella that
not only do both the AB and CD halves produce an apical organ,
but also any of the ^-embryos, and even any Isolated micromere
of the first quartet. I therefore turned with much Interest to a
more detailed examination of the localization of this organ in
Dentalium; and this Involved the inquiry whether the correla-
Experimental Studies on Germinal Localization. 35
Fig. VII.
Larvae from isolated Blastomeres.
42, 43, 44, Various forms of larvae from isolated CD halves, 24 hours; 45, 46,
twin larvae from the isolated CD and AB halves of the same egg, 24 hours;
47, larva from isolated D-quadrant, 24 hours; 48, larva from isolated C-quadrant
of the same egg, 24 hours; 49, larva from isolated posterior micromere, id, of
8-cell stage, 24 hours; 50, larva from isolated micromere, ic, of the same egg,
24 hours; 51, one-fourth larva from one -of the small quadrants (A, B or C),
72 hours.
36 Edmund B. Wilson.
tion between apical organ and post-trochal region is direct or in-
direct — i. e,, whether the development of the one depends on that
of the other, or whether the development of the two is only con-
nected through their common relation to the polar lobe. Further
experiments conclusively show that the latter is the case; for in
several ways larvae may be produced that possess the apical organ
but lack the post-trochal region. My first experiment to test
this consisted in the isolation, separately, of the four micromeres
of the first quartet (la, ib, ic, id), which may easily be effected
by means of Herbst's calcium-free sea-water. The result of this
experiment, several times repeated, is that while all four of these
micromeres may develop into actively swimming ectoblastic em-
bryos, the one derived from the D quadrant ( id) , and this alone,
develops an apical organ (Figs. 49, 50) . All of these four small
embryos are of approximately the same size, ovoidal or some-
what pear-shaped in form, with a group of active trochoblasts at
the larger (posterior) end. The anterior region is covered with
fine cilia (as in the AB ^-larya or the A, B or C ^4 -larva) ;
but only the id larva bears in addition the characteristic apical
tuft, which is nearly or quite as large as in a whole embryo, and
is borne upon the usual ectoblastic thickening or apical plate.
None of these larvae gastrulate or develop a post-trochal region;
from which it follows that after the completion of the third cleav-
age not only is the development of the apical organ independent
of that of the post-trcchal region, but at this time the posterior
micromere of the first quartet, id, is already definitely specified for
the formation of that organ, independently of its relation to the
remainder of the embryo. The result of isolating the cells of the
4-cell stage is entirely in harmony with this, as already mentioned.
The A, B or C 34 develops into a closed pyriform larva swimming
normally with the smaller and turned forwards, but entirely
devoid of apical organ or post-trochal region (Fig. 48). The
D ^, on the other hand, though often distorted, shows typically
the apical organ, and an exaggerated and usually irregular post-
trochal region. (Fig. 47.) This result is in striking contrast
to the fact, mentioned above, that in Patella, each of the quadrants,
whether of the 4-cell stage or of the first quartet, may develop an
Experimental Studies on Germinal Localization. 37
apical organ. The only conclusion that can be drawn from this
contrast is that the definitive basis of the apical organ is more
closely localized in Dentalium than in Patella, being concentrated
in a single cell.
The above results prove that the determination of the develop-
ment of the apical organ takes place at some period between
the first and the third cleavages. Further experiments fix the
period of determination still more nearly. If the egg be allowed
to advance as far as the second cleavage and the polar lobe formed
at that time be removed, the egg continues to segment in a manner
indistinguishable from that of an egg from which the lobe has
been removed at the time of the first cleavage. From such eggs
arise larva agreeing exactly with those arising after removal of
the first polar lobe in every respect save one, namely, that the
apical organ is typically present, though this is not invariably
the case. (Fig. 36.) Sections clearly show that the apical
tuft is borne upon a very definite apical plate, in striking con-
trast to the larvae arising after removal of the first polar lobe.
It is thus possible to produce at will larvae which lack the
post - trochal region and either possess or lack the apical or-
gan; and the determination of the apical organ is thus proved
to be effected during the short period between the first and sec-
ond cleavages. Complete corroboration is given by removal of
the second polar lobe from the isolated CD ^ during its first
division. The resulting larva resembles that arising from the
AB half in having no post - trochal region, but possesses an apical
organ as well developed as though the polar lobe had not been
removed. (Fig. 37.)
The experiments just described prove, first, that the correla-
tion between post-trochal region and apical organ is due to their
common determination by the first polar lobe. The second polar
lobe, though apparently precisely similar to the first, has no longer
any influence on the apical organ, though it still determines the
development of the post-trochal region. It seems impossible to
explain these facts, save under the assumption that the first polar
lobe contains specific stuffs that are in some manner essential to
the formation of both structures, and that during the period
38 Edmund B. Wilson.
between the first and second cleavages the "apical stuff" (if such
a term be allowed) exerts once for all its specific effect. The
most natural explaaation of this is given by the hypothesis that
this stuff moves upward to the apical pole, to be isolated in the
large posterior quadrant, D, during the second cleavage, and
subsequently in the corresponding micromere, id, during the
third cleavage. The basis of correlation between post-trochal
region and apical organ may thus be sought in the physical as-
sociation of the corresponding specific stuffs in the first polar lobe,
while the specification of the posterior micromere, id, is due to
the final isolation within it of the "apical stuff."
V.
LOCALIZATION IN THE UNSEGMENTED EGG.
The preceding sections are in a measure only preliminary to
the present one which includes the most important part of the
present paper, namely, the results of experiments on the localiza-
tion of the polar lobe, and of the structures that it involves, in
the unsegmented egg. As has already been stated, the clear sub-
stance forming the polar lobe is already visible in the egg prior
not only to cleavage, but even to fertilization and maturation.
Experiments on the unsegmented egg show with great clearness
that this area possesses in a general way the same promorpho-
logical value as the polar lobe itself, though at this early period
the egg possesses a greater regulative capacity than at later stages.
The unfertilized living eggs of Dentalium may readily be cut in
two with the scalpel under the microscope, and the plane of sec-
tion determined with considerable accuracy not only during the
operation but by a subsequent examination of the fragments in
which the polar areas are often still clearly visible. As Yves
Delage first showed, such fragments when fertilized may segment
and give rise to ciliated embryos and in certain cases even to
dwarf trochophores. In a considerable proportion of such ex-
periments, both fragments develop. For convenience of descrip-
tion I shall divide them into two classes, including (a) those
obtained by horizontal or oblique section, and (b) those obtained
b2 65
Fig. VIII.
Development of Egg-fragments after horizontal or oblique ScctionA
52, 53, Twins, after oblique or horizontal section near upper pole ; 54, trefoil,
lower half, horizontal section above equator; 55, 56, twins, after oblique section,
larger lower, smaller upper fragment; 57, 58, 59, twins, after slightly unequal
oblique section; 57, trefoil from lower fragment, 58, upper fragmertt (failed to
segment), 59, trochophore of 24 hours developed from 57; 60, 61, 62, twins,
horizontal section below equator ; 60, trefoil, lower fragment, 61, 2-cell stage
of same, 62, upper fragment, 2-cell stage; 63, trefoil, lower fragment, horizontal
section , polar lobe too small ; 63a, trefoil lower fragment, with polar lobe
slightly too large ; 64, 65, twins, plane uncertain, 64 undeveloped fragment, 65
trochophore, 24 hours.
1 In these and the following figures the plane of section is indicated by the
small accompanying diagram, the fragment studied being marked with a bL.ck
dot.
40 Edmund B. Wilson.
by exactly vertical section passing through the axis and bisecting
the polar areas.
{a) Fragments obtained by horizontal or oblique section. —
Under this heading may be grouped all fragments obtained by
sections passing in such a plane as to separate the polar areas, so
that one fragment contains only the upper, the other only the
lower, of these areas. These may be designated respectively as
the upper and the lower fragments. Before maturation I have
not found it possible to distinguish the upper from the lower
fragment; but as soon as the polar bodies form, the upper frag-
ment may be at once identified with certainty, since it alone pro-
duces these bodies. I have not thus far observed any difference
between the results of horizontal and oJ& oblique sections.
The upper and lower fragments differ in a characteristic way,
both in the form of cleavage and In the structure of the resulting
larvae; though it should be added that this appears most clearly
In the cleavage-process, since many of the embryos die before
reaching the trochophore stage, and many of the remainder be-
come wholly monstrous in form. Nevertheless the main result Is
given with great consistency by a comparison of the larvae. This
contrast is especially striking when two fragments from the same
egg are compared; and within rather wide limits It Is Independent
of the plane of section and the size of the piece, certainly as far
as the form of cleavage Is concerned, and apparently also as re-
gards the larval type. Whether large or small the upper frag-
ment forms the polar bodies In normal fashion, and In many
cases segments in essentially the same way as an egg from which
the polar lobe has been removed. The first cleavage takes place
without the formation of a polar lobe and Is Invariably equal
(Figs. 53, 56, 62, etc.), and the same applies to the second cleav-
age. Frequently the two pairs of cells shift during or after the
second cleavage, so as to produce a "cross-form," the succeeding
divisions of which are difficult to analyze. In many cases, how-
ever, the four cells remain In .nearly the same plane ; and In such
cases the succeeding divisions conform to the regular rule of
spiral cleavage, quartets of micromeres being found by alternat-
ing dexlotropic and lelotropic divisions. (Fig. 69.)
Experimental Studies on Germinal Localization. 41
A considerable proportion of these embryos fail to develop
into larvae, breaking up sooner or later Into loose groups of
cells that perish. Many, however, develop Into actively swim-
ming larvae, but these, whether large or small, are never normal
trochophores. While showing many variations, and often being
more or less Irregular In form, these larvae tend In general
towards, and sometimes agree precisely with those derived from
whole eggs minus the polar lobe, from the AB half, or the A, B
or C quarters (Figs. 70, 86). They are in general more or
less distinctly pyriform, swimming actively by the long cilia that
are more or less irregularly disposed about the posterior enlarged
region. A typical case is shown in Fig. 70 (from a preparation,
the cilia from the living larva) produced from the upper two-
thirds of an egg after exactly horizontal section. The cleavage
of this fragment was similar to that shown In Fig. 69. This
larva Is In every respect closely similar to the lobeless larva, though
the pre-trochal region is more expanded than usual, forming
a large hollow vesicle enclosing a few loose cells, and with a
slight thickening at the anterior pole, but without anything like
a true apical organ. The posterior region Is filled with a crowded
mass of rounded cells. Transverse sections of this larva show that
this mass Incloses a very small central cavity; but It is Impossible
to determine whether mesoblast cells are present or not. In a
very few cases an apical organ Is present in such larvae; but this
is so rare that I attribute Its occasional presence to the fact that
the plane of section was not quite correctly determined, a portion
of the lower polar area having been in fact included in the piece.
Another possibility is that the specific material of the polar lobe
extends so far up into the interior as to be removed by a section
that externally passes quite outside the polar area. This Inter-
pretation is supported by the fact that In a very few cases, when
the upper fragment Is considerably larger than the lower one, I
have seen the upper fragment form a very small polar lobe.
The development of the lower fragment — i. e., one that In-
cludes the lower polar area — differs In a remarkable way from
that of the upper one, both In the form of cleavage and in the
end-result. Whether obtained by horizontal or oblique sections.
42 Edmund B. Wilson.
and (within rather wide limits) whatever its size, this fragment
may segment in every detail like an entire egg of diminished
size, forming the polar lobe in normal fashion, and may give rise
to a dwarf larva nearly or quite normal in form and possessing
an apical organ. The study of a large number of these fragments
shows that while there is considerable variation in the size of the
polar lobe it is as a rule of approximately and often exactly,
of the correct proportional volume; and this is true eveij after a
horizontal section that. passes quite outside the limits of the polar
area. By varying the plane of section it is thus possble to obtain
a graduated series of forms leading down from a full-sized em-
bryo to one not more than one-fourth this size, the fragments from
the other halves forming a similar series grading in the opposite di-
rection. That the form of cleavage Is within wide limits, inde-
pendent of the size of the piece, is thus strikingly demonstrated.
Such a graded series of trefoils and the corresponding equal
2-cell stages, is shown in Figs. 52-60, the last of these showing
the smallest one observed. Regulation of the size of the polar
lobe sometimes fails however, examples being shown in Fig. 63,
where, even after horizontal section, the lobe is too small (this
egg produced a larvae possessing an apical organ, but with the
post-trochal region greatly reduced). Fig. G^a, where it is slightly
too large, and Fig. 66^ where it is much too large; but these are
exceptional. It is hardly possible that this apparent regulation
is owing to the fact that the specific polar material extends so
far up into the interior of the egg that a section in almost any
plane includes the right amount of material to form a normally
proportional lobe. Such an explanation is rendered very improb-
able by the usual failure of the upper fragment to form a lobe
even after horizontal section far down in the vegetative hemi-
sphere or after oblique section; and still more improbable by the
fact that so many of the fragments form a normally proportioned
lobe, whatever be the plane of section. The conclusion therefore
appears unavoidable that the size of the polar lobe, and hence
of the structures dependent upon it, is subject to a regulative
process, from which it follows that the predetermination of the
region of the polar lobe is qualitative, not quantitative, or if
Experimental Studies on Germinal Localization. 43
Development of Egg-fragments
66, 67, Lower fragment, oblique section; 66, trefoil, polar lobe much too large,
67 resulting 2-cell stage, CD half too large ; 68, 69a,-69c, cleavage of twin frag-
ments, oblique section, 68 trefoil, from smaller lower fragment, 69a-69c, sym-
metrical cleavage of larger upper fragment, 4-cell to i6-cell stages; 70, larva
of 24 hours, from upper fragment of a case exactly similar to 69; 71, 72, twins,
oblique section, nearly equal fragments, 71, 2-cell stage of upper fragment. 72 and
72, 2-cell and 8-cell stages of lower fragment; 73, 74, twins, probably oblique
section near upper pole, 74, upper fragment, 4-cell stage, 73a, normal third cleav-
age of lower fragment from lower pole, 73b, resulting 8-cell stage, 73c, begin-
ning of fourth cleavage, formation of first somatoblast.
44 Edmund B. JVilson.
quantitative, it is still subject to the operation of a regulative factor
that lies behind the topographical distribution of the egg-mate-
rials. This appears to me one of the most significant results that
my experiments have yielded.
The embryos may in succeeding stages cleave in every detail
like whole eggs. Typical 4-cell stages are shown in Figs. 73a,
78b, 83, 8-cell stages in Figs. 72, 73b, 84, and the fourth cleav-
age, with the formation of the first somatoblast, in Fig. 73c. Fig.
73a shows the third cleavage with the formation of the third
polar lobe. Many individuals were observed showing the forma-
tion of the second polar lobe in normal fashion, though none are
figured.
The larvae arising from fragments of this type differ as mark-
edly from those derived from the upper fragments as does the
cleavage. Although many of the embryos perish, and of those that
live many are abnormal, they frequently possess both the apical
organ and a post-trochal region; and occasionally a dwarf larva
is produced that is essentially similar, except in size, to an entire
trochophore. One of the best of these is shown in Fig. 59, which
arose from a lower fragment obtained by oblique section, slightly
smaller than half the volume of the egg, and including the whole
of the polar area. The typical trefoil stage of this larva is shown
in Fig. 57; it has exactly the normal proportions, and segmented
normally in later stages. This larva Is somewhat less pointed
posteriorly than the normal, but the whole larvae vary consid-
erably in this regard. It swam in quite normal fashion. Another
larger larva from a lower fragment in shown in Fig. 6^. The
total preparation of this larva shows with great clearness a typ-
ical apical plate at the upper pole. Out of a very large num-
ber of operations I have obtained altogether not more than five
or six such perfect larvae, at least half the embryos dying during
the cleavage, and a large proportion becoming abnormal during
the later development.
That so large a proportion of the embryos die or develop
abnormally is to be expected when we consider the very different
mechanical conditions of surface-tension and the like in these small
embryos. The fact remains that abnormal larvae may be pro-
Experimental Studies on Germinal Localization. 45
duced from lower fragments less than half the size of the egg;
and that such larva may possess a typical apical organ when the
section passes far away from the apical pole; while in no case does
the upper fragment produce a larva that ever approaches the
normal form. It may therefore safely be concluded that the
dwarf trochophores obtained by Yves Delage ('99) arose from
fragments including at least a part of the lower polar area.
The abnormalities observed in larvae from the lower frag-
ments range from only slight defects to wholly irregular and
monstrous forms, and thus far do not permit any more detailed
conclusions regarding the prelocalization than those stated above.
A common defect, illustrated by the pair of twins shown in Figs,
85, 86, is a more or less imperfect development of the post-trochal
region, even when the whole lower area is included in the frag-
ment, and sometimes this region appears to be wholly lacking.
Much more rarely the apical organ is lacking while the post-
trochal region is in greater or less degree developed. Such a
case is shown in Fig. 87 (from a preparation), the absence of the
apical tuft having been certainly determined in the living larva.
As in the case of the lobeless larvae, the experiments dem-
onstrate that the failure of the upper fragment to produce the
missing structures is not due to an insufficient mass of proto-
plasm; for I have obtained larvae showing the characteristic de-
fects from upper fragments fully two-thirds the bulk of the egg
(Fig. 70), and perfect dwarfs from much smaller fragments
(Fig. 59). The conclusion is therefore unavoidable that, like
the polar lobe to which it gives rise, the lower polar area contains
specific materials that are essential for the formation of the apical
organ, and of a post-trochal region; and that it is these materials
that enter into the formation of the polar lobe, as simple observa-
tion of the normal development indicates.
{b) Fragments obtained by ^vertical section through the axis.
— In view of the foregoing results we should expect to find that
when the egg is cut exactly vertically, so as to bisect the lower
polar area, both fragments should form the polar lobe; and such
is in fact the case. The experiments of this type were not very
numerous, and only a few cases were obtained In which both frag-
46 Edmund B. Wilson.
ments developed. I have only one pair of camera sketches to
show the polar lobes in such a case (Fig. 75,76). In both these
the lobe is relatively too small, as if produced from insufficient
material; but this not always the case (as shown beyond), and
it should be remembered that the polar lobe is sometimes too
small even in a lower fragment containing the whole of the lower
polar area (Fig. 62,). Figs. 77a, 78a show a pair, one of which
has a lobe of normal proportions ; the »ther is a very nearly nor-
mally formed 2-cell stage, though the larger cell is perhaps a trifle
too small. Both these produced nearly normally proportioned 4-
cell stages (Figs. 77b, 78b). Several other cases, in which only
one fragment developed, showed a normal trefoil. These data are
somewhat meagre, yet they justify the conclusion, I believe, that
after vertical section bisecting the lower polar area both frag-
ments may segment like whole eggs of half size.
The above conclusion renders it probable that by such vertical
section two perfect dwarf trochophores may be produced from
a single egg, which is apparently impossible when one fragment
alone contains the lower polar area. In point of fact, I have
never obtained even a single wholly normal larva after such sec-
tion; but in view of the comparatively small number of successful
operations and the very small number of such larvae obtained
by section in other planes this is not surprising. A number of
larvae from more or less nearly vertical sections is shown in the
following figures. Fig. 88 is a nearly normally formed larva
with two apical organs, from an oblique section passing outside
the lower white area. Fig. 89 is a nearly normal larva from a
section that removed a part of the lower area. Fig. 93 is from
an exactly vertical section bisecting both areas. In section this
larva Is closely similar to a normal one, and seems to show that
the trochoblasts are as large as in a whole embryo. Fig. 90 Is
from the smaller fragment after a slightly oblique section bi-
secting the lower area ; a very distinct apical organ is present and
also an abnormally formed post-trochal region. Figs. 91, 92 are
twins from a slightly unequal vertical section (developed from the
respective twin fragments 81, 82), the post-trochal region is
lacking In both, while one lacks an apical organ.
Experimental Studies on Germinal Localization. 47
Fig. X.
Development of Egg-fragments after vertical Section.
75, y6, Equal twins, respectively in trefoil and polar lobe-formation; lobes
too small ; 77, 78, equal twins, nearly correct proportions ; 77a, 77b, typical trefoil
and 4-cell stages of one fragment; 78a, 78b ,typical 2-cell, slightly abnormal
4-cell stages of the twin fragment; 79, 80, twins, from a fertilized egg, 79,
nearly normal trefoil, 80, the twin, with reduced polar lobe; 81, 82, nearly
equal twins. 81, typical 4-cell stage, 82, its twin, nearly typical 2-cell stage;
83, 84, typical 4-cell and 8-cell stages, from upper pole, of the same fragment
48 Edmund B. Wilson.
It may be pointed out that not one of these larvae shows a
fully developed post - trochal region, though 91 and 92 arose
respectively from 2- and 4-cell stages that show nearly the normal
proportions and must have been produced from nearly normal
trefoils. This may seem to contradict the conclusion, drawn
above, that the predetermination of the lower polar area is not
quantitative ; but a similar reduction sometimes exists in this region
when the whole polar area is present (as in Fig. 85), and I do
not think a trustworthy conclusion can be drawn without addi-
tional data.
I may add that after a large number of unsuccessful attempts
I obtained two nearly normal dwarf trochophores from frag-
ments of the unsegmented egg of Patella. One of- these, which is
about half the volume of a normal larva, clearly shows the cells
of the prototroch. In the full - sized normal trochophore of
Patella the prototroch, as may be seen with the greatest clearness
in total preparations, consists of a closed principal ring of cells
that vary in number (as seen in optical section) from 19 to 21.
In the dwarf the cells are more variable in size and less regularly
arranged, but on the average as large as in the normal individual;
equatorial optical section of this larva shows 13 cells in the prin-
cipal ring.
VI.
OBSERVATIONS ON FRAGMENTS OF THE FERTILIZED EGGS AND
ON THE ISOLATED LOBE.
Extremely interesting and curious results are obtained by a
comparison of the behavior of fragments of fertilized eggs, and of
the isolated polar lobe, with that of fragments of the unfertilized
eggs described above.
{a) The behavior of fragments of fertilized eggs obtained
before cleavage. — In order to malce sure that the eggs were fer-
tilized the operation was delayed until one or both polar bodies
had been formed, and the egg was then cut as nearly as possible
horizontally, so as to separate the lower polar area from the
nucleated part. As already described, if this operation be per-
formed on the unfertilized egg, and the two fragments be fer-
tilized, both may, and frequently do, develop. When, however,
Experimental Studies on Germinal Localization. 49
Fig. XI.
Larvae of 24 Hours, from Egg-fragments.
85, 86, Twin larva of 24 hours, oblique section passing outside lower polar
area, 85, the lower, 86, the npper larva; 87, larva from lower two-thirds, hori-
zontal section, without apical organ; 88, larva from lower two-thirds, oblique
section, two apical organs; 89, larva from nearly vertical section; 90, larva
from smaller fragment, slightly oblique section bisecting lower area; 91, 92,
twin larvae, produced from 81 and 82 respectively, vertical section; 93, larva
from exactly vertical section.
50
Edmund B. Wilson.
Experimental Studies on Germinal Localization. 51
Fragments of fertilised Eggs, horizontal Section; isolated Polar Lob'es and/
Fragments of Lobes.
94, 95, Equal twins from the same egg ; 9Sa, 95b, upper half, 2- and 8-cell
stages ; 943-941, successive changes in the lower enucleated fragment ; 94a, soon
after operation; 94b, first polar lobe (drawn immediately before 9Sa) ; 94c, first
resting stage, upper fragment in 2-cell stage (23 m. after b) ; 94d, second polar
lobe (21 m. after c, the upper fragment just divided into 4) ; 94e, second rest
(16 m. after d) ; 94f, third polar lobe (i6m. after last, at nearly .the same time
with 9Sb) ; 94g, third rest (7 m. after last) ; 94h, fourth lobe (46 m after last,
fourth cleavage in progress in upper fragment) ; after 16 minutes the' fragment
appeared to be divided into two and so remained ; 94i, the same four hours
later; 96, lower fragment, like last, but showing correct proportions of first
polar lobe; 97, successive changes in isolated polar lobe from the individual
shown in Fig. 21; 97a, soon after removal; 97b, first active period (16 m., the
egg just divided into 4) ; 97c, ensuing first resting period (14m. after b) ; in the
second period of activity (not sketched), 15 m. later, as the eggs divided into
8, the lobe constricted as in b, but not so deeply, and again became spherical in
a second resting period ; 97d, e, third period of activity, 74 and 78 ni. after the
first period; 97f, final result, 36 m. later; 98, another isolated lobe, 98a, third
resting period; 98b, second lobe; 98c, final result, in which condition it re-
mained without further change; 99, 100, two fragments obtained by cutting
a polar lobe in two (the original lobe was slightly larger than usual), showing
active changes shortly after division of the egg into 4; 99a, looa, are shown
42 m. after 99 and 100; 99b, loob, 8 m. later; 99 remained in this condition,
while 100 again became spherical by fusion of the two halves and underwent
no further change.
52 Edmund B. JVilson.
a fertilized egg is thus sectioned only the nucleated {i. e., the
upper) fragment develops — a result that agrees with my ob-
servations on the nemertine egg and that of Renilla ('03), and
with the earlier ones of Delage ('01) on those of echinoderms.
This fragment has essentially the same mode of development
as a corresponding fragment of an unfertilized egg, segmenting
equally into two and four without the formation of a polar lobe,
forming successive symmetrical quartets of micromeres by alter-
nating spiral cleavages (Fig, 95), and producing a larva that is
either an irregular monster or a pyriform larva closely similar
to those arising from the lobeless egg or the AB half. This is
what would be expected in view of the preceding results; but
the behavior of the non-nucleated lower half is most remarkable
in that it forms three times in succession a polar lobe from the
white area at the same time that the nucleated half is dividing^
becoming spherical after each period of activity without dividing.
When this was first observed, I believed that I must in some way
have confused the fragments with those of unfertilized eggs; but
repetitions of the experiment under conditions that precluded all
error, gave the same result. A typical case is shown in Fig. 94,
from consecutive camera drawings of the same fragment. The
first lobe is shown (Fig. 94b ) about 15 minutes after the opera-
tion, while the nucleated half (Fig. 95) has just divided into
equal halves. Twenty-three minutes later the fragment was again
perfectly spherical (94c), while the upper fragment was in a
resting 2-cell stage. The second lobe (94d) was formed 44
minutes after the first, while the upper fragment was dividing into
4 equal cells, after which the lower fragment again became spher-
ical (94e, 16 minutes later than 94d). The third lobe (94f)
was formed 32 minutes after the second, and was considerably
smaller than either the first or the second, as in a whole egg;
the upper fragment meanwhile divided into eight cells (Fig. 95b) .
A third period of rest followed (Fig. 94g) . Following the fourth
cleavage of the upper fragment the lower one passed through a
change no less remarkable than the preceding (it is at this
period in the normal development that a large part of the
lower polar area passes into the first somatoblast) . This
Experimental Studies on Germinal Localization. 53
change begins with the formation of a fourth lobe, com-
posed of white material, which is at first much smaller than
any of the preceding (94h, 46 minutes after 94g). Unlike the
preceding lobes this one was not resorbed into the fragment, but
was permanent, slowly increasing in size until after two or three
hours it was nearly as large as the remaining portions, the frag-
ment now appearing as if divided into two (94i).
This case is fairly typical of several that were followed through
the entire cycle of changes, and one or more of the stages were
seen in many individuals. The lobes are not always so distinctly
formed as in the one figured, and the final stage, though usually
like that described, varies considerably in appearance.
{b) Behavior of the isolated polar lobe. — Previous to mak-
ing the observations just described, I had several .times observed
changes of form in the isolated polar lobes after their removal
from the trefoil stage. On reexamining the matter I found that
these changes are also periodic, taking place approximately at the
same time as the cleavage in the lobeless nucleated portion. The ac-
tivities of the isolated lobe at these periods vary considerably in
different individuals. Sometimes the activity is no more than a
slight change of form, the spherical lobe becoming slightly pyri-
form or even almost amoeboid. Frequently, however, the isolated
lobe actually forms a smaller lobe by a process that closely sim-
ulates the formation of a polar lobe by a whole egg or an egg-
fragment. In any case, each period of activity is followed by a
spherical resting-stage that coincides approximately in time with
the resting stages of the segmenting lobeless portion. I regret
that I had not time to study this remarkable phenomenon with
sufl[icient care, but give series of sketches illustrating two particular
cases. Fig. 97a shows a lobe soon after its removal; 97b, the
same, 16 minutes later just after the egg had divided into four;
97c, the ensuing resting stage, 14 minutes after 97b; a second
period of activity followed, in which the lobe again constricted,
but not so deeply as at 97b, followed by a second spherical stage;
97d and 97e show the third active period, and 97f the final result,
after which no further change occurred. In 98 is shown the final
active period of a lobe, which resulted in the permanent apparent
54 Edmund B. Wilson.
division of the lobe into two. Even if the lobe be cut in two
after its removal, the fragments likewise pass through alternating
periods of activity and rest closely similar to those of the whole
lobe, as is shown in Figs. 99, 100 (the original lobe was somewhat
larger than in the other cases shown). This proves that the
power of a rhythmic change of form involving the temporary
formation of lobe-like structures, is not a property of the lobe as
a whole, or of the lower polar area, but is inherent in the sub-
stance of which it is composed. It would be interesting to com-
pare in this respect the behavior of the isolated lobe, or fragment
of a lobe, with fragments from other regions of the fertilized egg.
Such fragments would probably also exhibit rhythmic changes, but
I hazard the conjecture that their activity would be found to
differ in some definite way from that of the lobe-fragments.
The phenomena above described, which deserve further careful
study, are of interest both cytologically and embryologically.
First, since both the nuclei and the centrosomes are absent, it
follows with great probability that even in the cleavage of a
whole egg the constriction of the cell that leads to the formation
of the polar lobe takes place wholly independently of either these
structures or the astral rays, which suggests the possibility that
the same may be true of the constrictions that lead to complete
cell-division. Second, since the rhythm in the formation of the
polar lobes in the enucleated fragment coincides with that shown
in the division of the nucleated fragment, it is clear that as far
as the lobe-formation is concerned the cytoplasmic division rhythm
is quite independent of that of either the centrosome or the chro-
mosomes. This fact may be placed behind the one earlier de-
termined by Boveri ('97), Zlegler ('98) and myself ('01), that
the rhythmic activities of the chromosomes and of the cen-
trosomes are likewise independent, or at least separable. But
beyond this it Is remarkable that the periodic activity in the non-
nucleated fragment is not merely of a rhythmic character, but
changes its character at the time of the fourth cleavage when in
the normal development the material of the polar lobe no longer
forms a merely temporary structure, but is permanently cut off
by a cell-division. We here catch a glimpse, as It were, of a
Experimental Studies on Germinal Localization. 55
definite order of events predetermined in a particular cytoplasmic
area and wholly independent of the immediate action of nucleus
or centrosome. An additional point of great embryological in-
terest is the fact, shown by a comparison of Fig, 6 with Fig. 94,
that in these fragments the polar lobe is, at least in some cases,
nearly or quite as large absolutely as in one entire egg; whereas
in the lower fragment of an unfertilized egg it is typically re-
duced to the correct proportional volume of the lobe in a whole
egg. This is however not invariable, for in some cases, an ex-
ample of which is shown in Fig. 96, the lobe is reduced to its
proper proportional size. I have not accurately studied this mat-
ter in a sufficient number of cases to speak very positively; yet I
feel confident that the contrast in this respect between the lower
fragments from unfertilized and fertilized eggs is a general,
though not an invariable rule. The interest of this fact is
pointed out in the sequel.
VII.
COMMENT.
Without undertaking at this time a complete discussion of the
foregoing observations, I may briefly indicate their bearing on
the general questions referred to at the beginning.^ My observa-
tions demonstrate conclusively, I think, both the mosaic character
of cleavage in these eggs, and the definite prelocalization of some
of the most important morphogenic factors in the unsegmented
egg. The Dentalium egg shows, even before it breaks loose from
its attachment in the ovary,and long before even the initial changes
of maturation, a visible definite topographical grouping of the
cytoplasmic materials. This is proved by the experiments to
stand in definite causal relation to the subsequent differentiation
of the embryo in such wise that the removal of a particular cyto-
plasmic area of the unsegmented egg results in definite defects
in the resulting embryo that are not restored by regenerative or
other regulative processes within the time-limits of the experi-
ment. Since both the egg-fragments and the isolated blastomeres
1 A more general discussion of the mosaic-theory of development, with a
fuller review of the literature, will be given in a following paper.
56 Edmund B. fVilson.
become perfectly spherical before development proceeds, the re-
sulting defects cannot be due to a failure of regulation traceable
to the shape of the fragment, as was formerly assumed by- several
writers. Neither are they due to insufficient mass; for perfect
dwarfs may arise from fragments much smaller than those that
show the characteristic defects. Further, these facts, like those
earlier determined by Crampton ('96) in the gasteropod egg, and
by Driesch and Morgan ('95) and more recently by Fischel
('98) in the ctenophore egg, are fatal to the view that embryonic
differentiation is brought about through quahtative nuclear divi-
sion during the cleavage. The conclusion is therefore unavoidable
that the specification of the blastomeres in these eggs is due to
their reception, not of a particular kind of chromatin, but of a par-
ticular kind of cytoplasm; and that the unsegmented egg con-
tains such different kinds of cytoplasm in a definite topographical
arrangement. How many such specific stuffs exist in the unseg-
mented egg of Dentalhim and what is their arrangement it is
impossible at present to say; for the pigment-band and the two
polar areas can only be considered as an outward sign of an or-
ganization that for the most part doubtless escapes the eye. My
experiments have only positively determined the cytoplasmic pre-
localization in the lower polar area of material essential for the
development of that complex of structures that I have included
in the term "post-trochal region," and of one other structure, the
apical organ. The first of these includes material that is essential
to the development of the typical larval form, including the foot,
to certain characteristic ectoblastic structures of the post-trochal
region, such as the shell-gland, mantle-fold, and probably also the
pedal ganglia ; it also appears probable that it includes material
essential for the formation of the coelomesoblast. I do not doubt
that further experiments on this egg will show a still more definite
and detailed prelocalization ; though, as already stated, it is
not easy to determine this, owing to the difficulty of distinguishing
between defects in the partial larvae that result directly from the
plane of section and those that are due to other causes.
Two additional facts clearly appear from the experiments, on
which I would lay stress. First, the amount of material removed
Experimental Studies on Germinal Localization. 57
with the polar lobe or lower polar area is wholly disproportionate
to the effect produced. The polar lobe includes less than one-
fifth the volume of the egg; yet its removal does not merely cause
a structural defect of like extent, but inhibits the whole process
of growth and differentiation in the post-trochal region and the
concomitant withdrawal of the pre-trochal region. The cleavage
of the lobeless embryos shows that both the second and the third
quartets are formed; and it is fair to conclude that certainly in
the AB half of the embryo, and probably also in the CD half,
these cells contain ectoblastic material, which in a normal embryo
would contribute to the formation of the post-trochal region.
These cells, as stated above, close in around the posterior region,
and perhaps are partially turned in with the invaginating ento-
blast-cells. In any case, however, the power of active growth in
the post-trochal region, so conspicuous in the normal larva, is
wholly lost with the removal of the excess of material in the D
quadrant. It does not seem possible that this loss in power of
growth is due to mechanical obstacles, since the same defects exist
in fragments of the unsegmented egg from which the lower
polar area has been removed and which are free to segment as
best they can. The conclusion therefore appears unavoidable that
the material of the lobe is not only specifically necessary for the
formation of the bases of the post-trochal structures, but also for
the whole growth-process that is here brought to a focus. Apart
from its more general bearings, this conclusion is important from
the light that it may throw on the teloblastic growth of annelids
and other segmented forms, and it seems altogether probable
that if the polar lobe could be removed from such an egg as that
of Sabellaria or Myzostoma the resulting larva would fail to
develop a metameric trunk-region.
A second point of interest that clearly appears from the ex-
periments is that the topographical grouping of specific materials
in the unsegmented egg may be in its ensemble widely different
from that of the definitive bases of the organs which they de-
termine; for the experiments demonstrate that the apical organ,
lying at the upper pole, is determined by material originally lying
far down in the vegetative hemisphere in the lower polar area.
58 Edmund B. Wilson.
On this point an analogous result has recently been obtained by
Yatsu, who has shown with great probability that in the unseg-
mented nemertine egg the basis of the apical organ does not lie
at the upper pole, where we should expect to find it, but in, or
slightly above, the equatorial region.
These facts have an important bearing on our interpretation of
development in general. In my previous paper on the nemertine
egg I have developed an hypothesis of differentiation agreeing
broadly with Sach's well-known theory of formative stuffs, and
with the general conclusions regarding mosaic development inde-
pendently published by Fischel ('03) nearly at the same time, the
essential assumptions being that the prospective value of a cell is de-
termined by its cyVoplasmic content, that this content is de-
termined by the form of cleavage in connection with an antecedent
formation and segregation of specifically different materials
(which may Itself determine the form of cleavage), and that
the morphogenic function of cleavage, so to say, is to isolate
the materials thus segregated. This conception, it is hardly neces-
sary to point out, receives very definite support by the observations
now brought forward; but I wish to bring them more closely into
relation with those made on the nemertine and echinoderm eggs,
especially with regard to the general question of progressive (i. e.,
epigenetic) localization in the egg. In the nemertine {Cerebra-
tulus) I found that either an isolated blastomere or a fragment
from any region of the unsegmented egg may produce a perfect
dwarf larva; but the two differ In the form of cleavage, the
blastomere segmenting as if still forming part of a whole em-
bryo and producing an open blastula (as in the echinoderm),
while the egg-fragment segments like a whole egg and produces
a closed blastula — that is, it develops as a whole from the be-
ginning. I explained the contrast In development between the
two as the result of a regrouping of the egg-materials, occurring
during and subsequent to the process of maturation and fertiliza-
tion, which Initiates the morphogenic process and determines also
the form of the earlier cleavages. I pointed out that such re-
grouping of materials is known to occur at the maturation-period
of many eggs — for Instance, in the sea-urchin — and suggested
Experimental Studies on Germinal Localization. 59
that the contrast between the development of an egg-fragment in
the nemertine and in a sea-urchin (where it segments like a whole
egg only after section in certain planes) is owing to the fact that
in the latter, egg-fragments have only been obtained in the period
subsequent to maturation when the regrouping has been effected.
Localization of the cleavage-factors was thus conceived, essen-
tially in agreement with Roux's early conclusions regarding the
frog's egg, as a progressive (i. e., epigenetic) process, and the
same conception was applied to the general morphogenic process
which, as is shown with especial clearness by the facts here brought
forward, may be so closely connected with the cleavage-process.
As far as the progressive character of localization is con-
cerned, the result obtained in Dentalium may seem at first sight
to be in disagreement with the conclusions just reviewed, for the
germ-regions are here defined by a definite segregation of ma-
terials that exists even in the attached ovarian egg long before
either maturation or fertilization, and the isolated blastomere
is not capable of producing a complete embryo. But the contra-
diction disappears upon comparison with certain other forms,
which are intermediate in character between the extremes repre-
sented by Dentalium and the nemertine or echinoderm egg; and
this comparison demonstrates, as I believe, the validity of the
theory of "precocious segregation," formulated as a pure specula-
tion by Ray Lankester in 1877. I have already expressed the
opinion that the horizontal stratification of the egg expressed by
the three zones of material visible in Dentalium or Myzostoma is
comparable, or at least analogous, to that which finds an expres-
sion in the formation of the well-known polar rings of leeches and
oligochaetes. This comparison is based both on the position and
mode of formation of these rings and on their fate. Vejdovsky
('88) very clearly shows that in Rhynchelmis both the polar rings
arise as local thickenings of a general ectoplasmic layer, and both
assume at one period the form of protoplasmic discs lying at
either pole of the egg (as Whitman also observed in Clepsine) .
Except for the fact that the upper and lower protoplasmic areas
have not at any period been seen to appear in the form of actual
6o Edmund B. JVilson.
rings, the resemblance to these relations of those observed in
Dentaliiim is unmistakably obvious. It is entirely possible that
the correspondence is not complete; but that in a general way
the resemblance indicates a similar form of stratification in the
molluscan and annelidan egg, seems hardly open to question; and
the comparison is sustained by the fact that in Clepsine both rings
were traced by Whitman into the AB half, and the upper one
into the D quadrant, while in Rhyfichelviis Vejdovsky traced both
rings into the D quadrant, where the material of the two fuses into
one mass in the 4-cell stage and later passes into the mesomeres,
which are undoubtedly to be identified with the somatoblasts.^
If this comparison be admitted a further comparison of these
and some other forms is highly significant. In Dentaliiim three
structural zones are present from the beginning, the lower one
coinciding in extent with the lower white area, the upper one
lying at the centre of the upper white area, at first very small, but
rapidly increasing in extent during and after the maturation
period. A condition similar to this exists in Sternaspis, where
Vejdovsky ('81) showed that a distinct protoplasmic area, which
he compares to a polar ring ('88, p. 122) lies at each pole of
the ovarian egg, the upper one being much smaller than the lower
one, though larger than in Dentaliiim. . In Clepsine and Rhyn-
chelmis three structural zones are likewise present, but tJiese first
appear during the maturation period with the development of the
polar rings, like the three zones described by Boveri ('01) in the
Strongylocentrotus egg. The egg of Myzostoma occupies, at
least in some respects, an intermediate position. No upper pro-
toplasmic disc has here been observed as yet, but the lower proto-
plasmic area is obviously represented by the green mass, which, as
Driesch ('96) has shown passes into the polar lobe, and subse-
quently certainly in part into the first somatoblast, and probably
in part into the second somatoblast, precisely as in Dentaliiim.
The interest of this case, compared with the foregoing, lies in the
fact observed by Driesch (which I can confirm) that before ma-
1 Cf. Vejdovsky and Mrazek ('03, p. 454) ; see also the highly interesting
statement (p. 534) that the dense protoplasm of the polar rings ("Polplasmen")
can be recognized as such "in den Zellen des Mesoblasts inbesondere in den
grossen Mesomeren."
Experimental Studies on Germinal Localization. 6 1
turation the egg shows at first but two colored zones, of which
the lower green one exactly represents the lower white area of
Dentalium, while the upper one first segregates during maturation
into an upper red zone and an equatorial colorless one. Like
the lower zone the two upper ones correspond very closely in fate
to those in Dentalium; for the upper (red) area passes into the ec-
tomeres, like the upper white area of Dentalium, while the middle
(colorless) zone passes into the entomeres, as is the case with the
greater part of the middle (pigmented) zone in Dentalium. It is
possible that sufficiently careful search may reveal the presence
in Myzostoma of an upper protoplasmic disc, comparable with
a polar ring; and as far as the visible colored zones are concerned,
it is evident that the Myzostoma egg stands midway between those
of Dentalium and Strongylocentrotus , and it is probably inter-
mediate also between Dentalium or Sternaspis and Clepsine or
Rhynchelmis.
It seems a legitimate interpretation of the foregoing series
that these eggs present an essentially similar form of stratification
which is attained at different periods in the ontogeny, and that
as compared with the leech or oligochaete, Myzostoma and Den-
talium or Sternaspis represent two earlier stages in the precocious
segregation of specific cytoplasmic materials that have a like pros-
pective value in the development.^ But if this be admitted, it
follows that in none of these cases can the segregation in question
be considered as a primary character or "preformed quality" of
the^egg. .Upon this secondary localization of material, as my
experiments prove, depend many of the most important features
of the later morphogenic localization; and I think a presumption
Is thus established that cytoplasmic prelocallzation is in general of
like secondary or epigenetic origin, though to what extent this
holds true can only be determined by further experiment.
Although the characteristic segregation is In its main outlines
effected very early In the egg of Dentalium, It may be pointed out
that, like so many other eggs, there is the clearest evidence of
1 Cf. Vejdovsky "Wahrend aber bei Sternaspis die Concentration des Bil-
dungsplasma an beiden Polen bereits im Laufe der Eibildung stattfindet, sam-
melt sich dasselbe bei Rhynchelmis erst nach der Polzellenbildung und dem Ein-
dringen des Spermatozoon in das Ei an" ('88, p. 122.)
62 Edmund B. JVilson.
later movements and progressive segregation of the cytoplasmic
materials. I will only call attention, among these, first, to the
determination of the apical organ by material originally lying In
the lower polar area, which, if my interpretation of the experi-
ments is valid, moves upwards to the apical pole in the period
between the first and second cleavages. That such a movement
occurs is only a matter of inference ; but this interpretation appears
to me far simpler and more intelligible than to assume a brief
"Fernwirkung," or the like emanating from the first but not the
second polar lobe. It Is however not a matter of inference but
of fact that the remaining material of the lower white area moves
upwards and towards one side in the 8-cell stage preceding the
fourth cleavage, when it apparently fuses with the material of the
upper white area in the D-quadrant. It Is interesting to com-
pare this with the facts described by Vejdovsky in Rhynchelmis,
where the remains of the upper and lower polar rings fuse in the
D-quadrant at the 4-cell stage.
I have endeavored to show that cytoplasmic prelocalizatlon
in Dentalium differs only In degree from the conditions existing in
such eggs as those of the nemertine or sea-urchin. The same may
be said, I think, of the development of isolated blastomeres, de-
spite the fact that in Dentalium such blastomeres are incapable
of producing complete dwarf embryos. As In the nemertine or
sea-urchin, although the isolated blastomere segments as a part
and not as a whole, the embryo finally closes, in the course of
which process structures like the prototroch, the post-trochal and
pre-trochal regions, and the gut, close to form whole structures.
That this process, which in the case of the nemertine I compared
to Morgan's "morphallaxis" In regenerating planarians or hy-
drolds, falls short of producing a complete embryo in Dentalium,
may be due to different causes in different cells. In the AB half
or one of the smaller quarters this Is obviously due in the main
to lack of the specific material of the lower polar area. The
failure of the CD half or the D quarter may In part be due to
a like cause; but since these embryos contain the materials (those
contained in the lower polar area) that are missing In the other
cases, their failure may be due to a different cause. The CD half
Experimental Studies on Germinal Localization. 63
larvae are sometimes nearly normally formed except for the false
proportions of the post-trochal and pre-trochal regions. Their In-
variable subsequent degeneration Into Irregular and monstrous
forms Is not Improbably due to the abnormal mechanical condi-
tions created by their mode of development. It seems possible,
however, that if these larvae could sustain themselves sufficiently
long they might in some cases succeed in attaining a normal con-
dition. They die before attaining this end; and hence succeed no
better than the AB halves In the "attempt" to produce a perfect
embryo.
One cause of the difference between the isolated blastomeres of
the nemertlne or sea-urchin and the mollusk thus doubtless lies
in a difference In the segregation-pattern such that In the former
the specific materials are symmetrically divided between the first
two blastomeres, while in Dentalium such is not the case. In the
former, accordingly, the earlier cleavages are purely quantitative,
but in the latter are qualitative as far as the cytoplasm is con-
cerned, and to this extent produce from the first cleavage onward
a mosaic-work in entire accordance with Roux's general concep-
tion, as I long since indicated in the case of Nereis ('94). But
beyond this the results especially of Driesch's later studies on the
isolated blastomeres of sea-urchins indicate that here, although
a definite polarized segregation of material has taken place at the
time of the earlier cleavages (directly proved by Boveri's ('01)
observations on Strongylocentrotus, indirectly by Driecsh's ('00)
comparison of the development of the upper and lower quartets of
the 8-cell stage) this segregation Is not only symmetrical with
respect to the axis but is also less definite or less complete than
In the molluscan egg, — again a difference which finds its nat-
ural explanation In the theory of precocious segregation (or dif-
ferentiation). I should therefore interpret the differences be-
tween the isolated blastomeres of the mollusk and those of the
sea-urchin or nemertlne as due to a difference, on the one hand.
In the pattern, on the other hand in the degree, of segregation.
It is hardly necessary to point out that the foregoing conclusions
will in large measure reconcile the apparent conflict between the
fact of cytoplasmic prelocallzation and the continually Increasing
64 Edmund B. Jf^ilson.
evidence that the primary determining factors of development
are to be sought in the nuclear organization. The well-known
hybridization experiments of Boveri ('92, p. 469) and Driesch
('98) on sea-urchins have shown that the earlier cleavage-factors
conform to the maternal type and hence must be predetermined
in the egg-cytoplasm; and up to the blastula-stage, at least, the
embryos remain of the pure maternal type. But the same ex-
periments demonstrate no less clearly that the nucleus begins to
affect the cytoplasmic phenomena at least as early as the late
(prismatic) gastrula, and according to Boveri's latest work ('03)
as early as the mesenchyme-formation, though the latter point is
disputed by Driesch ('03). It therefore appears possible, not
to say probable, that every cytoplasmic differentiation, whether
manifested earlier or later, has been determined by a process in
which the nucleus is directly concerned, and that the regional
specifications of the egg-substance are all essentially of secondary
origin.
Another question, which has been often discussed, is raised
by these observations, namely, as to the relation in the regenerative
process between the moulding of the mass as a whole (which
falls under the general conception of Roux's "Umordung der
Zellen" or Morgan "morphallaxis) and the specification of the
individual cells. Like the facts determined by Fischel ('98) in
the ctenophore egg (following the earlier work of Driesch and
Morgan) those observed in Dentalium bring out with great clear-
ness the independence, in this case, of the two groups of factors
by which these are determined. It is a very noteworthy fact that
all the partial larvae that lack the lower polar area, whatever
their size or mode of origin, tend to assume the same form, and
all are alike devoid of further regenerative capacity. The larvae
arising from entire eggs after removal of the polar lobe only,
the CD half from which the second polar lobe has been removed,
the AB half, the A, B or C quarter, or an upper fragment, of
any size, of the unsegmented egg — all these typically assume
the characteristic pyriform shape with the trochoblasts surround-
ing the larger posterior end. This form, which results after
closure of the embryos and gastrulation, is essentially a prolate
Experimental Studies on Germinal Localization. 6^
spheroid modified by the presence at one end of the large tro-
choblasts which have not like the other cells the power of con-
tinued multiplication, and it evidently represents a state of equilib-
rium towards which any segmented mass of the egg tends that
is devoid of the lower polar area. Whether the closure of the
embryos (which in the case of isolated blastomeres are at first
strictly partial structures) to produce this form should be con-
sidered as a regulation or regenerative process is largely a ques-
tion of definition.^ In any case the facts very clearly show that
the process is not perceptibly influenced by the nature of the cells
individually considered; nor does it, on the other hand, appear to
exert any appreciable effect on the nature of the individual cells
("Umdifferenzierung" of Roux), as will be more clearly shown
in my second paper.- Certainly the closing of the embryos does
not lead to the least perceptible tendency towards the restoration
of the missing structures that are dependent on the material of
the lower polar area.^ I am in agreement with the opinion of
Fischel ('98) that, whether a regulative process or not, the
closing in to form a closed structure is probably explicable as a
result of relatively simple physical factors, though I doubt whether
the explanation is as simple as Fischel assumes in the case of the
ctenophore."* It is difficult to avoid the conclusion that these
same factors are operative in the establishment of the normal
form in a whole embryo; but to them are add in the material
of the lower polar area a far more complex group of factors, at
present not analyzable, that involve the whole process of growth
and metamorphosis. That a mass of cytoplasm so small should
1 Roux ('93, p. 837) interpreted the closure of the open blastula as part of the
regenerative process, in opposition to Driesch ('92, p. 585), who asserted that
this had nothing to do with the regenerative process proper ; though he after-
ward took the ground that it should be considered as an initial regulative pro-
cess ('96, p. 88). Morgan ('01, p. 13, etc.) classes morphallaxis under the
head of regeneration, though not the closing in of a cut surface, which is con-
sidered as a preliminary process. Cf. Child, on "Mechanical Regulation" ('02).
2 Cf. Crampton, '97, p. 55.
3 Cf. the remark of Driesch, based especially on Crampton's experiments on
Ilyanassa; "1st, wie bei Gastropoden und Anneliden, echte Lokalisation der Bil-
dungsfaktoren im Ei anztmehmen, so schliesst das eine Regulation zum Ganzen
wirklich aus." ('96, p. 89.)
■* Cf. Rhumbler, '02, Zur Strassen, '03.
66 Edmund B. Wilson.
exert so great an effect on the morphogenic process is a most
convincing piece of evidence in favor of the theory of specific
formative stuffs in development. The only intelligible view of
the polar lobe seems to me to be that it is, so to say, a reservoir
of such stuffs destined for allotment to particular cells which
thereby become definitely specified, irrespective of their subsequent
relation to the embryo as a whole. This is a very different result
from the oft-quoted one of O. Hertwig that the lineage of par-
ticular structures from particular blastomeres is nothing more
than an incidental result of the continuity of development. It is
equally opposed to the conclusions of other writers who have too
hastily rejected the principle of mosaic development for which
Roux and others have contended.
Lastly I may point out that in so far as these observations show
the course of differentiation, and the correlation of parts, to be
determined by a preexisting topographical grouping of specific
egg-materials they sustain an essentially mechanistic (as opposed
to a vitalistic) interpretation of development. To conclude how-
ever that these eggs are devoid of regulative capacity would be
to overlook some of the most striking of the phenomena I have
described. The experiments give clear evidence that a power of
regulation exists in the unsegmented egg that is no less striking
in form, if more limited in degree, than in the nemertine or echlno-
derm. As in the case of the nemertine, the typical spiral cleavage,
alternately dexiotropic and lelotropic, is not affected by section in
any plane. Far more striking is the fact that in the cleavage of an
egg-fragment the size of the polar lobe, on which the proportions
of the trochophore largely depend, is proportional to the size of
the piece. Since this Is true even after horizontal section, when the
whole of the lower polar area is included in the piece, it follows that
the predetermination of this area is qualitative, but not quantita-
tive, or only quantitative in so far as it is subject to regulative
control by other factors. This conclusion receives further sup-
port from the one reached above that the material of the lower
polar area Is as such specifically concerned not merely with the
formation of the structures that arise from it but with the form
of growth that results In the metamorphosis. But if this par-
Experimental Studies on Germinal Localization. 67
ticular area shows such a qualitative, as distinguished from a quan-
titative, pre-determination, one is led to suspect that a like con-
clusion may apply to other egg-regions, such as those that form
the gut, the prototroch, and the like ; and to conclude that however
detailed a prelocalization may exist in the form of regional seg-
regations of material, a regulative factor may always be present
that controls their normal combination. In this respect the un-
segmented egg, may, I believe, be directly compared with such
an adult animal as a planarian or hydroid, which, while possess-
ing more or less definitely specified tissues. In a typical grouping,
nevertheless may possess a high regulative capacity shown in the
process of regeneration after injury.
The facts observed give as little clue to the nature of the regu-
lative factors by which the quantitative relations are determined
in the egg-fragment as in the fragment of a planarian or hydroid;
but one or two considerations deserve brief mention. It is note-
worthy that although the polar lobe regularly forms in the non-
nucleated vegetative half of a fertilized egg it is as a rule, though
not always, not reduced, but nearly or quite as large as in a whole
egg, whereas in a fertilized fragment, representing the same re-
gion of an unfertilized egg, the lobe is as a rule reduced to its
proper proportional volume. While I would not lay too much
stress on this without further study, it seems to indicate that the
power of regulation, on which the size of the lobe depends, is
more complete in a nucleated fragment than in an enucleated one.
Second, when once the polar lobe has formed, the power of regu-
lation seems to be lost, at least temporarily; for if a part of it
be cut away the second lobe is of correspondingly reduced size,
as is also the post-trochal region of the resulting larva. This
result is supported by the fact that, like the post-trochal region
to which it gives rise, the polar lobe in the first (virtual second)
division of the isolated CD half, though sometimes slightly re-
duced, is in general nearly or quite as large as in a whole embryo.
These facts prove that the size of the lobe is not determined
merely by the size of the piece, but by more complex conditions
existing apparently for only a brief period, and apparently also
more effective in a nucleated than in a non-nucleated protoplasmic
68 Edmund B. PFilson.
mass. This sufficiently Indicates the complexity of the problem
with which we are dealing, and the importance of further more
precise studies of the facts. At the same time, it seems clear that
the problem of proportionate development in a fragment of an
organism here appears in a much simpler form than In a blastula-
fragment, or a piece of an adult organism such as a planarlan or
a hydra ; and I think we should not abandon the hope of finding
for It a relatively simple solution. While I am not able to offer
such a solution, It seems to me that it would be rash to deny its
possibility, not merely in the present instance, but In all analogous
processes, even when they take place under the more complex
conditions existing In multicellular masses.
VIII.
SUMMARY.
1. The Dentalium egg shows from the beginning three hori-
zontal zones, an equatorial pigment-zone and two white polar
areas. Each of the polar areas includes a specially modified pro-
toplasmic area probably comparable to a polar ring.
2. During cleavage the pigmented zone Is allotted mainly to
the entomeres, the upper white area to the ectomeres, the lower
white area to the first and probably also the second somatoblast.
At the first, second and third cleavages the lower white area tem-
porarily passes into the "yolk-lobe" or polar lobe.
3. Removal of the first polar lobe leads to a symmetrical
cleavage without the subsequent formation of polar lobes, and to
the formation of a larva devoid of post-trochal region and apical
organ. Removal of a portion of the first lobe produces a larva
with reduced post-trochal reglon,^ and with or without apical organ.
Removal of the second polar lobe produces a larva without post-
trochal region but with an apical organ.
4. The lobeless larvae undergo no metamorphosis, form no
foot, shell-gland or shell, no mantle-folds, no pedal ganglia, ap-
parently no mouth, and probably no coelomesoblast-bands.
5. The isolated AB half or A, B, or C quarter, produces a
closed larva closely similar except In size, to the lobeless ones.
The isolated CD half or D quarter produces a larva possessing a
Experimental Studies on Germinal Localization. 69
post-trochal region as large as In a normal larva, and an apical
organ, which dies without undergoing metamorphosis. The CD
half from which the second polar lobe is removed produces a larva
like that from an AB half, but possesses an apical organ.
6. The isolated micromere id produces a mass of ectoblast-
cells bearing an apical organ, while la, ib, ic produce no apical
organ,
7. Fertilized fragments of the unsegmented unfertilized egg,
obtained by horizontal or oblique section, differ in development
according as they do or do not contain the lower white area.
The upper fragment segments symmetrically without the forma-
tion of polar lobes and produces a larva similar to the lobeless
ones. The lower one segments like a whole egg of diminished
size, and may produce a normally formed dwarf trochophore.
8. Fragments obtained by vertical section through the lower
white area may segment like whole eggs and may produce nearly
normally formed dwarf trochophores.
9. Enucleated fragments, containing the lower white area,
of fertilized eggs, pass through alternating periods of activity
and quiescence corresponding with the division-rhythm of the
nucleated half, and form the polar lobes as if still forming part
of a complete embryo. The same is true of the isolated polar
lobe.
10. The foregoing observations demonstrate the prelocallza-
tlon of specific cytoplasmic stuffs In the unsegmented egg and
their isolation in the early blastomeres. The lower white area
contains such stuffs that are essential to the formation of the
apical organ and the complex of structures forming the post-
trochal region, including the shell-gland and shell, the foot, the
mantle-folds and probably the coelomesoblast. These stuffs are
contained in the first polar lobe, but the second lobe no longer
contains those necessary for the basis of the apical organ. Pro-
gressive changes therefore occur in the original distribution of
the specific cytoplasmic materials.
11. Comparison Indicates that the conditions observed In the
molluscan egg differ only In degree from those In the nemertlne
or echlnoderm. These differences reduce themselves to differ-
70 Edmund B. PFilson.
ences in the period of segregation (or differentiation) and in its
pattern, and are explicable under the general theory of precocious
segregation.
12. The early development of egg-fragments indicates that
the specification of the cytoplasmic regions is primarily qualitative,
but not quantitative, or if quantitative is still subject to a regulative
process that lies behind the original topographical grouping of
the egg-materials.
13. The development of the molluscan egg is in its essential
features a mosaic-work and sustains the theory of "Organbiln
dende Keimbezirke."
LITERATURE.
BovERi, Th., '92. — Befruchtung : Merkel u. Bonnet, Ergebn. I. '9i-'92.
'97. — Zur Physiologic der Kern- und Zelltheilung: Sitzungsber. d.
phys.-med. Ges. Wiirzburg.
'01, I. — tjber die Polaritat des Seeigeleies : Verb, phys.-med. Ges. Wiirz-
burg. (N. F.) XXXIV.
'01, 2. — Die Polaritat von Ovocyte, Ei und Larve des Strongylocen-
trotus lividus: Zool. Jahrb. (Anat.-Ontog.) XIV, 4.
'02. — tJber mehrpolige Mitosen. Verb, phys-med. Ges. Wiirzburg. (N.
F.) XXXV.
'03. — liber den Einfluss der Samenzelle auf die Larvencharaktere der
Echiniden : Arch. Entwm., XVI, 2.
Child ,C. M., '02. — Fission and Regulation in Stenostoma : Arch. Entwm.,X V,2.
CoNKLiN, E. G., '99. — Protoplasmic Movements as a Factor in Differentiation :
Wood's Hole Biol. Lectures. 1898.
'02. — Karyokinesis and Cytokinesis in the Maturation, Fertilization and
Cleavage of Crepidula and other Gasteropoda : Journ. Acad. Nat.
Sci., Phila. II. Ser. XII, i.
Crampton, H. E., '96. — Experimental Studies on Gasreropod Development ; Arch.
Entwm. III.
'97-— The Ascidian Half-Embryo : Ann. N. Y. Acad. Sci., X. •
Delage, Yves, '99. — fitudes sur la merogonie : Arch. Exp, Zool. (Ser. III.), VII.
'01. — fitudes experimentale sur la Maturation cytoplasmique chez les
Echinodermes: Arch. Exp. Zool. (Ser. Ill) IX.
Driesch, H., '92. — Entwickelungsmechanisches : Anat. Anz., VII, 18.
'96. — Betrachtungen iiber die Organisation des Eies und ihre Genese :
Arch. Entwm. IV.
'98. — tlber rein -miitterliche Charaktere an Bastardlarven von Echiniden :
Arch. Entwm., VII.
'99. — Resultate und Probleme der Entwickelungsphysiologie der Thiere :
Merkel u. Bonnet, Ergebn. VIII.
'00. — Die isolirten Blastomeren des Echinidenkeimes : Arch. Entwm. X,
Experimental Studies on Germinal Localization. 71
2, 3.
'02. — Neue Antworten und neue Fragen : Ergebnisse, Merkel u. Bon-
net, XI.
'03. — Ueber Seeigelbastarde : Arch. Entwm. XVI, 4.
Driesch and Morgan^ '95. — Zur Analysis der ersten Entwickelungsstudien des
Ctenophoreneies: Arch. Entwm. II.
FiscHEL, A., '97. — Experimentelle Untersuchungen am Ctenophorenei, I : Arch.
Entwm. VI, i.
'98.— Id., II-IV; Ibid. VII., 4.
'03. — Entwickelung und Organ-Differenzirung : Ibid., XV.
KowALEWSKY, A., '83. — fitude sur I'embryogenie du Dentale : Ann. Mus. d'Hist.
Nat. de Marseille. Zool. I, 7.
Lacaze, Duthiers, '57. — Histoire de I'organisation et du developpement du Den-
tale : Ann. Sci. Nat. IV. Ser. VI, VII.
Lankester, E. Ray, 'jy. — Notes on Embryology and Classification : Q. J. M. S.
XVII.
LiLLiE, F. R., '01. — The Organization of the Egg of Unio; Journ. Morph.
XVII, 2.
Morgan, T. H., 'oi. — Regeneration: Columbia Biological Series. VIII.
Rhumbler, L., '02. — Zur Mechanik des Gastrulationsvorganges, etc. : Arch.
Entwm., XIV.
Roux, W., '85. — tJber die Bestimmung der Hauptrichtungen des Froschembryo
im Ei, etc. : Gesammelte Abhandlungen, II. 20.
'88. — tJber die kiinstliche Hervorbringung "halber" Embryonen, etc. :
Ges. Abh., II. 22.
'93. — Ueber Mosaikarbeit und neuere Entwickelungshypothesen : Ges.
Abh., 27.
'95. — Gesammelte Abhandlungen. II. 2>3- Nachwort.
'95. — Uber die verschiedene Entwickelung isolirter erster Blastomeren:
Arch. Entwm. I.
'03. Ueber die Ursachen der Bestimmung der Hauptrichtungen des
Embryo im Froschei : Anat. Anz., XXIII, 4-7.
Vejdovsky, F., '81. — Untersuchungen iiber die Anatomic, Physiologic und Ent-
wickelung von Sternaspis : Denkschr. d. Akad. Wien, XLIII.
'88-'92. — Entwickelungsgeschichtliche Untersuchungen : Prag.
Vejdovsky und Mrazek, '03. — Umbildung des Cytoplasma wahrend der Be-
fruchtung und Zellteilung: Arch. Mik. Anat. LXII, 3.
Wheeler, W. M., '97. — The Maturation, Fecundation and Early Cleavage of
Myzostoma glabrum : Arch. Biol. XV.
W HITMAN, C. O., '78.— The Embryology of Clepsine : Q. J. M. S., XVIII.
Wilson, E. B., '92. — The Cell-lineage of Nereis : Journ. Morph. VI.
'93. — Amphioxus and the Mosaic Theory of Development: Journ. Morph.
VIII.
'94. — The Mosaic Theory of Development : Wood's Hole Biol. Loct.
Lect. II. 1893.
'g6, I. — On Cleavage and Mosaic-work: Arch. Entwm. III.
'96, 2.— The Cell: ist Ed. New York.
'01. — Experimental Studies in Cytology, II : Archiv. Entwm. XIII., 3.
72 Edmund B. Wilson.
'03. — Experiments on Cleavage and Localization in the Nemertine-egg:
Arch. Entwm. XVI. 3.
'03.— ^Notes on Merogony and Regeneration in Renilla : Biol. Bull. IV. 5.
Yatsu, N., '04. — Experiments on the Development of Egg-fragments in Cere-
bralatus: Biol. Bull., VI, 3.
ZiEGJ-ER, H. E., '98. — Experimentelle Studien iiber die Zelltheilung, I : Arch.
Entwm. VI. 2.
'98. Die Furchungszellen von Beroe ovata : Ardi. Entwm. VII.
ZuR Strassen, O., '03. — Ueber die Mechanik der Epithelbildung; Verb. d.
Deutsch. Zool. Ges., 1903.
REGENERATION IN RHIZOSTOMA PULMO.
BY
CHARLES W. HARGITT.
With 6 Figures.
I. INTRODUCTORY.
The several experiments, of which this paper presents a resume,
were conducted during the early summer of 1903, at the Naples
Zoological Station, while occupying the table of the Smithsonian
Institution, for the courtesy of which it is a pleasure to express
my obligations.
The primary object of the experiments was to test the regen-
erative capacity of the Scyphomedusae and to institute certain
comparisons between these results and those obtained by similar
experiments previously made upon the Hydromedusae. So far
as I am aware no similar experiments have been made upon the
Scypliomedusae with the definite purpose of testing this particu-
lar aspect of their physiological constitution. Romanes in his
experiments upon " Primitive Nervous Systems," '85, has record-
ed incidentally the fact that certain mutilations of medusae are
promptly healed, but gave no details. Eimer, '78, has also
carried on similar experiments and with the same general purpose
of testing the character and distribution of nervous centers, but
makes no reference to the matter of regeneration. And quite
recently Uexkiill, '00, has likewise reviewed these experiments
of Romanes and Eimer and carried them somewhat farther than
they had done. But while arriving at somewhat different con-
clusions, drawn from a series of experiments in some features
coincident with those to be described now, he makes no reference
to any regenerative processes, devoting attention almost exclusively
to the movements, specially those of rhythmic character, and seek-
ing physical explanations of them.
74 Charles W . Hargitt.
The earlier references of Haeckel to the capacity of larvae of
certain medusae to regenerate entire organisms are likewise in-
definite. Morgan in referring to the subject in his recent book on
"Regeneration," 'oi, merely remarks that among Scyphozoa 'the
jelly-fishes belonging to this group have a limited amount of re-
generative power."
I very much regret that an unusual scarcity of material compels
me to leave several points somewhat less fully considered than is
desirable, but I trust they are not of sufficient gravity to seriously
mar the general value of the results as a whole.
In one respect this scarcity of material, making necessary suc-
cessive experiments on the same specimen in many cases, proved
fortunate rather than otherwise, since facts of importance were
thus brought to light which might otherwise have been overlooked.
Some of these will be referred to specifically in another connec-
tion.
.II. EXPERIMENTAL.
The experiments were performed upon Rhizostoma pulmo, one
of the most common of the Mediterranean medusae. Both in
size and vigor this medusae affords one of the most satisfactory
forms for experimentation which has come under my observation.
It seems likewise to suffer less under the somewhat artificial con-
ditions of the aquarium than any other which I have had occasion
to use. As compared with Aurelia and Cyanea of New England
waters it is incomparably superior in every way, but particularly
in its ability to thrive for weeks in an environment which would
prove fatal to the others in as many days. With the single excep-
tion of Gonionemus I know of no other medusa which affords
so good a type for this sort of observation and experimentation.
It was not unusual to have specimens under direct observation in
the ordinary aquaria of the laboratory rooms for from four to
six weeks and without apparent deterioration, even in some cases
under the severe tax of extensive mutilation made necessary by
the experiments to which they were subjected. It should be stated
however that as a rule younger and smaller specimens proved
much better than those of larger size; the latter, on account of
Regeneration in Rhizostoma Puhno. 75
their greater mass, are inclined in most cases to sink toward the
bottom of the tanks, where after a time certain disorganizing
influences appeared to set up pathologic conditions which seemed
to deplete their vigor and at the same time render their regen-
erative processes less satisfactory.
The experiments were directed to three ends, namely to deter-
mine: I, The capacity of the medusae to reproduce lost parts, or
to recover from such Injuries as might ordinarily happen to them
in a state of nature, such as the battering effects of waves, the
injuries inflicted by enemies, etc.
2, The comparative powers of the various regions to regenerate,
or In other words, the relation of the regenerative capacity to
liability to Injury.
3, The capacity to regenerate such highly specialized organs as
rhopalla, or other sensory structures.
The experiments included specimens of sizes from about 20
m/m to 125 m/m In diameter, and while all proved to have
unexpected powers of regeneration those of medium size, from
40 to 70 m/m, proved very much more satisfactory than those of
larger size both in convenience and In their promptness In re-
sponding to the several sorts of operations, and they apparently
were more healthy and vigorous during the progress of the ex-
periments than were those of larger size. Those having a size
of 100 m/m or more In diameter proved to be much less prompt
In regeneration and, as will be seen in the records of experiments,
were much more liable to deteriorate or utterly collapse than were
the smaller specimens. This is only what might be more or less
expected, and Is quite In keeping with observations on other
classes of organisms. The same tendency was more or less evi-
dent in specimens on exhibition in the public aquarium in which
of course no mutilations or similar Injuries had occurred. In this
connection may be noted a somewhat anomalous pathological phe-
nomenon observed in large specimens both in the exhibition aqua-
ria and in the small aquaria during the course of experlmentaton,
namely, the appearance of whitish blotches, or patches of disin-
tegrating tissues at various places on the exumbrella of the animal
which sooner or later affected its health and general behavior.
76 Charles W . Hargitt.
The matter will be referred to in further detail in another con-
nection and some reference made as to its probable significance
and cause.
In all cases the primary experiments were made as soon as
possible after the medusae were brought into the laboratory. I
have said the primary experiments. This refers to the fact al-
ready alluded to, that in several cases experiments were variously
repeated upon the same specimen. This was in part for the pur-
pose of testing the conclusiveness of preceding experiments, and
in part owing to the fact that there was an insufficient supply of
material to serve the demands of the course of experiments under
way. Details as to these aspects will be given in connection with
the several experiments described.
The first experiment was made upon a large specimen, and in
order to determine at the outset whether the earlier observations
of Romanes and others, that complete removal of the marginal
sense organs resulted in complete paralysis of the medusa, these
organs were carefully removed by means of triangular incisions
as indicated in Figure i, a. The results were substantially con-
firmatory of the earlier records, the medusa becoming more or
less passive, except for an occasional single contraction at very
irregular intervals. This experiment was made on May ii, and
the following series of observations will suffice to show the general
course of events. It should be added in this connection that along
with the excision of the rhopalia several other marginal excisions
were made, and that three of the oral arms were cut off close
below the region of the gastric enlargement. The aspect of the
specimen on the next day was practically the same. While there
was an occasional contraction of the bell accompanied by certain
'movements of the body, there were no Indications of rhythm.
May 13th. — The medusa, while apparently in perfect health
and vigor of general functions, was still unable to originate any
definitely rhythmic movements, though responding to various
mechanical stimuli, such as a strong current of water from the
tap, or the touch of a glass rod. At various times during the
day there was evident a rather marked tendency toward sponta-
neous movements, and occasionally something very like a rhythm,
Regeneration in Rhizostoma Pulmo.
11
several contractions following each other in regular succession,
though never continuing beyond three or four pulsations.
May 14th. — The medusa, while still more or less passive as
before, was yet apparently recovering more of the power of spon-
taneity, several pulsations occurring at more frequent intervals,
but these were not of sufficient vigor to produce any locomotion.
--ai
Fig. I.
Diagram showing methods of excising rhopalia.
a, usual triangular excision; ai, excision of larger mass; b, rectangular form
of excision; c, circular form of excision; d, form of rhopalium and lappetts.
May 15th. — During this and the following day there was an
apparent relapse of the medusa to the condition of the first day.
There was also less vigor apparent, such stimuli as those referred
to above producing but slight effects. This condition continued
during the i8th, 19th and 20th.
7 8 Charles JV. Hargitt.
May 2 1 St. — The medusa seemed to have recovered the vigor
or tone to which reference has been made above. There was
also a very evident rhythm in the contractions, often as many as
ten or more regularly recurring pulsations occurring at irregular
intervals during the day. As before, however, they were not of
sufficient force to secure the locomotion of the animal. The same
condition was observable during the following day.
May 23d. — There was again a marked decline in both vigor
and general tone of the body, which showed evident signs of de-
generation. This condition continued during the following day,
and on the morning of the 25th the medusa was found to have
died during the preceding night.
Upon careful examination it was found that wherever tissue
had been mutilated or excised there had been a definite healing
of the wounds and in the case of the oral arms there were indica-
tions of new growth. I was not able to distinguish that there had
been any regeneration of the sensory organs, and this will appear
somewhat surprising in the light of the following experiments.
Whether there had really been no regeneration at all, or that I
had overlooked the new organs, or whether they may have dis-
integrated during the night following the death of the medusa I
am unable to say. Certain it is, however, that if regeneration
had gone forward as markedly as in the following cases one could
hardly have failed to distinguish it. I am inclined to believe
that the paralysis following the total removal of these organs
may have served to delay or inhibit active regeneration.
The next series of experiments differed materially from the
former, particularly in that care was taken to retain certain of
the rhopalia in order to insure continued activity of the organisms
during the progress of the experiment. The number of rhopalia
retained varied from one to eight, the latter case serving as a
means of testing the relative influence of these bodies on the
behavior of the animals and the rate of regeneration.
On May 12th several specimens, averaging only about half the
size of the preceding, namely, about 50 m/m in diameter, were
experimented upon. In the first one all the rhopalia were retained,
but marginal notches were made of varying sizes between the
Regeneration in Rhizostoma Pulmo. 79
sensory bodies, and several of the oral arms were excised. In
other specimens a varying number of the rhopalia were excised,
and in one case all the oral arms were cut off close to the gastric
enlargement and on one side including a portion of this organ
itself.
I shall not undertake to transcribe in detail the records of each
day, but give rather summaries of results as briefly as is com-
patible with clearness, trusting that nothing of importance may
be sacrificed in the attempt to bring the records within as brief
compass as possible.
One of the first effects distinguishable in these and following
experiments was the evident quickening of the pulsations of the
medusae by the process of excision of the organs, or similar op-
eration. Not only was the rate of the rhythm greatly increased,
passing from about seventy pulsations per minute as an average
for medusae of this size, to ninety, or even one hundred per
minute. And this rate continued during the entire day, or at
every observation, which was quite frequent, and well on into
the second day, when the rate fell to ninety and later to eighty;
but it was not till the third day that the rate had fallen to the
normal of seventy per minute. An examination at this time
showed an evident healing of the wounds and some signs of re-
generation. Had this been restricted to the sensory bodies it
might have been interpreted as signifying some important rela-
tion of these organs to rhythmic activity, but the fact that similar
effects were produced upon specimens which had not been de-
prived of their rhopalia would sufficiently negative such an in-
ference.
Eimer, '74, had noted such an effect following a division of
medusae, particularly those which had been divided into halves
or fourths, and had undertaken to show that it was chiefly an
expression of the reduced size of the organism due to its division,
citing the normal rhythm of specimens of varying size as strongly
suggesting such an inference.
Romanes, '85, however, was not able to confirm Eimer's con-
tention either in reference to matter of fact or the cause assigned.
Romanes, while citing the variation as to the rate of rhythm in
8o Charles IF. Hargitt.
specimens of similar size, is inclined to emphasize what he terms
the prepotent influence of certain of the lithocysts (rhopalia) in
coordinating the rate of movement, and the presence or absence of
such prepotent organs in the portions of medusae under exam-
ination.
Forbes, '48, had long previous called attention to the fact of
these quickened movements under the influence of various stimuli,
citing particularly a result of an experiment which he had made
of a similar character to those which I have cited above. In an
experiment in which he had, as he expresses it, "paralyzed one
half of the animal" by cutting out the rhopalia from one side, he
finds "that the other half contracted as usual, though with more
rapidity, as if the animal were alarmed or suffering." He remarks
farther that "all medusae when irritated become much more rapid
in their movements and contract or expand their disks or bodies In
a hurried and irregular manner, as if endeavoring to escape from
their persecutors." (Naked Eyed Medusae, p. 3.)
While in certain details the conclusions of Forbes may be ques-
tioned, of his general observations as to matters of fact there
can hardly be doubt. Furthermore, whether the suggestions of
either Eimer or Romanes are more than approximate guesses,
the later observations of Uexkiill have rendered doubtful. So far
as my own experiments have gone they hardly touch the problem
of the cause of such reactions. We may safely conclude that, in
any case, they are of the nature of responses to any continued
physical stimulus, such as the experiments under consideration cer-
tainly were. With the healing of the wounds there would of
course ensue a decline of the irritation, which in turn would be
followed by a return to the normal rate of rhythm.
On May 26th, or two weeks following the operation, the me-
dusae had measurably regenerated all the excised organs. The
notches cut In the umbrella margins had grown out to complete
the normal symmetry and there had been developed in the areas
the characteristic purple pigment, differing from the color of the
uninjured portions only in its Intensity. The new rhopalia were
apparently normal in everything save size and pigmentation.
It is rather noteworthy that in these experiments certain of the
Regeneration in Rhizostoma Piilmo. 8i
organs which among the Hydromedusae are most promptly re-
generated are here among the most slow to develop ; such, for
example, as the oral arms and gastric lobes. The fact that in the
rhizostomous medusae these organs have no very active function
in the capture of food might apparently afford some plausible
reason for this difference in the rate of regeneration. In Goni-
onemus the gastric and oral organs are among the most prompt
in regeneration, and are, of course, also among the most important
in the functional activities of the animal. That this, rather than
liability to injury, should be a predisposing factor in regeneration
would seem to be confirmed in the case of Rhizostoma, for as will
appear in later experiments there seems to be no good reason to
suppose that the liability to injury, to which these organs are
constantly exposed, has anything to do with the capacity for rapid
or perfect regeneration.
Additional experiments were begun on May 28th and 30th. In
this series the specimens varied in size from 20 to 60 m/m in
diameter. As remarked above there was in these cases the same
degree of promptness in the responses, which was markedly in
contrast with that shown by specimens of considerably larger
size, but in the present cases there was also apparent a somewhat
less favorable response In the very small specimens. This fact
considered in connection with the difficulty of operating easily
upon small specimens, emphasizes the value of animals of medium
size for such experiments. This conclusion was emphasized
throughout the entire course of experimentation.
In part of the specimens of this series only three rhopalia were
excised, in others four, in others five. In some the rhopalia were
all removed from one side, while in others only alternate organs
were removed. In some specimens the same order was observed
as to excision of mouth arms and other similar operations. One
of the specimens of the series had only one full-sized mouth arm,
while the others were in what seemed to be various stages of
regeneration. As is well known these organs among medusae of
this type are among the most open to accident from attack of
fishes or other predatory enemy. The specimen under considera-
tion would seem to confirm the results of these experiments that
82 Charles W. Hargitt. ' »
these organs are readily regenerated, and that in a state of nature
as well as under the artificial conditions of the laboratory. An
examination made with the hand lens on June 2d, or only four
or five days following the operation, showed the first indication
of regenerating rhopalia. As the organ first makes its appearance
it is a very minute papilla-like body, and in these cases at the
inner, or upper edge of the notch made by the incision. Ex-
amined under the compound microscope the papilla appears as
a minute, solid bud growing out from the terminal region of
the radial canal, though it does not at first seem to be a direct
outgrowth of that organ. Very soon, however, there is established
a direct connection with the canal, and it is quite easy to dis-
tinguish the circulation of the gastric fluid in the little bud, which
becomes definitely vesicular, as shown in Figure 2. The growth
of the organ, after its vesicular stage is established, is quite rapid
and there can soon be distinguished the thickening of the terminal
portion to form the lithocysts. Coincident with this stage of de-
velopment there is discernible the development of the new hood
and lappets, accessory organs, and as will be shown in connec-
tion with a study of the histology of these organs, the correspond-
ing development of the so-called olfactory and ocellar pits.
In connection with the present series the following experiments
were made with a view to demonstrate that, not only in form but
in function, the new rhopalia were perfect organs. From one
of the specimens just described in which three rhopalia had been
originally excised the other five were excised on June 5th, or seven
days after the original experiment. If the three regenerated
organs had not yet attained to functional utility the effect of re-
moving the others would, of course, result in the typical paralysis,
as in the first experiment already described. As was anticipated,
the careful removal of all the rhopalia except the three regenerated
ones did not in the least interrupt the normal rhythm or activity
of the creature, save to act as a stimulus to quicken it, as already
cited in connection with a previous series. This experiment was
repeated upon several others of this as well as subsequent series,
and always with the same results, except in a single case which
may as well be cited in this connection, though coming under later
experiments.
Regeneration in Rhizostoma Piihno.
83
In this case the original operation had removed six rhopalia,
leaving but two. Soon after the appearance of the new rhopalia,
but before they had begun to approach complete development, or
before there was any indication of the presence of lithocysts or
pigment, the two original organs were carefully removed, and in
this case with what might likewise have been anticipated, namely,
the complete inhibition of the normal rhythm and the consequent
paralysis of the organism. This inhibition continued during the
Fig. 2.
Section of rhopalium in early stage of regeneration, ect, ectoderm; ent, ento-
derm ; h, hood ; mgl, mesogloea ; r. c, radial canal ; s. e., sensory epithelium.
following two days. With the continued development of the
new rhopalia activity was recovered, though, owing to the Inter-
position just at this juncture of an unhealthy condition of the
medusa, it failed to entirely recover the usual vigor or tone which
the others had shown.
These experiments, abundantly corroborated by subsequent ones,
leave no shadow of doubt, it seems to me, as to the capacity of
84 Charles W. Hargitt.
these organisms to regenerate in the last detail one of the most
highly specialized organs known among Coelenterata. This will
be shown more fully in connection with the later account of the
histology of the regenerated organs.
Other series of experiments, continued to June 20, while varied
In some aspects of detail, were of substantially the same character
and with results quite similar to the preceding.
In several of the experiments care was taken to so modify the
form and extent of the excised portions as to secure evidence as
to the Influence of contiguous tissues or parts upon the regenerat-
ing organs. In Figure i is shown, for example, several aspects of
the mode of excising the rhopalia. For the most part the excision
was In the form of a triangular cut from the margin Inward toward
the radial canal, as shown in the figure. The dotted line a^ will
show also in the same connection the occasional extension of the
cut to Include twice the usual mass. In Figure i, h, will be seen
another form of operation. In this case the portion cut out was
Fig. 3.
Twin rhopalia regenerated in place of the single original one.
rectangular Instead of triangular, as In the former. The mass
excised in the operation also varied as before. At c, In the same
diagram, may be seen another form of excision In which the
cut was circular instead of angular, as in the former cases. It Is
Interesting to note that, so far as I was able to determine, the
form of the excision had no perceptible effect upon the form or
rate of regeneration. In the case of the rectangular or circular
excisions the new organ appeared in Its typical place at the median
position of the upper portion of the notch. In the case of the
large or small portions excised In the triangular cuts not the
slightest difference could be distinguished. With the exceptions
of some two or three cases to be considered, there was not the
Regeneration in Rhizostoma Puhno. 85
slightest evidence of any deviation from the exact position occu-
pied by the original organ.
The apparent exceptions referred to are as follows : First, that
in at least two cases twin rhopalia were developed instead of the
single original one which had been excised. This is well shown in
Figure 3. Second, that in one case two rhopalia were regenerated
instead of the one originally excised, but unlike the preceding,
they appeared at different points — one in the usual position at
the upper angle of the notch, the other at the lower, or marginal
portion of the notch, as shown in Figure 4.
The mere fact of the occurrence of double rhopalia during re-
generation instead of single ones is not of itself particularly re-
markable, for the occurrence of such features is not an unusual
one in a state of nature, both ephyrae and adult medusae being
occasionally found with such double organs. Some further in-
quiry should, however, be directed to the peculiar position in
which the organ noted in Figure 4, at a, occurs, namely, at one
side of the notch and near the margin instead of the usual posi-
tion. On the assumption that these organs are of sensory func-
tion and correlated with marginal nerve centers it might be
a'
Fig. 4.
Two regenerated rhopalia; a, near the margin.
thought that in regeneration they would be likely to occur in close
relation with such centers, and that the case under consideration
might be thus explained. The fact is very clear, however, that such
is not the case with the vast majority of the experiments where ap-
parently the relation of nerve centers had nothing whatever to do
with their position in regeneration. And when furthermore we
reflect that these are not nervous organs in any true sense, either
in their origin or development, though possibly correlated with
86 Charles W . Hargitt.
some sensory function, it must be more or less evident that such
an explanation of the single case cited would hardly hold.
Nor would it perhaps be more satisfactory to appeal to what has
been designated as polarity in explaining either series. The occur-
rence of the organs in conjunction with the radial canals and their
apparent differentiation from terminal portions of these structures
would seem to afford a much more probable explanation of their
regeneration at these apparently predetermined positions. And
may we not find in this view a simple explanation of the occur-
rence of the anomalous case referred to in Figure 4, a, for we
find near the margins a more or less complex network of anas-
tomosing canals, the presence of one of which may have been
the inciting cause of the development of a sensory body at this
particular point.
It is interesting to note in this connection that no appearance
of heteromorphism occurred during the entire series of experi-
ments. This feature I have referred to in a previous paper, '97.
in connection with similar work on Hydromedusae. On the as-
sumption that these organs are metamorphosed tentacles we migh
naturally look for heteromorphic phenomena similar to that re-
corded among the Crustacea, in which occasionally instead of an
eye an antenna develops. Nothing of the sort, however, occurred.
There seems in every organ and tissue a remarkably inflexible
physiological constancy. This is the more remarkable when con-
trasted with the highly flexible character of the polyp phase of the
group among which are found the widest range and variety of
heteromorphism.
The fact is not overlooked that Rhizostoma is devoid of ten-
tacles, which might be assumed as sufficient reason why hetero-
morphism of this sort was not manifested. The fact remains,
however, that Its polyp has the typical tentacular equipment, and
that In Its metomorphism they are resorbed and possibly take
the usual course, some of them contributing toward the formation
of rhopalia. It might be an Interesting problem to determine
in detail just the extent of this supposed metomorphism of the
polypal tentacles Into rhopalia. May it not be possible that the
supposed metamorphism Is In reality a resorption and that only,
Regeneration in Rhizostoma Puhno. 87
and that the rhopalia are essentially independent dev-elopments
such as are found during the process of regeneration? I merely
raise the suggestion as it has been forced upon my attention in
course of these experiments. It seems worth farther investigation.
In this connection may be briefly described a phenomenon which
only came under critical observation late in the course of the
experiments, and which for lack of material it was impossible to
follow out to conclusive results. Among the last of the series
two large specimens were operated upon as follows : In the
first all but one of the rhopalia were excised, while in the second
all but two were removed. In both cases there was distinctly
noticeable an aberrant, rotary sort of swimming movement, the
animal revolving in an irregular circle, instead of directly for-
ward or upward as is usual. Examination showed that this in-
clination of the body in swimming was constantly in the direction
of the remaining rhopalia, which would seem to suggest that
perhaps they functioned something after the nature of equilibrium
organs. I do not recall that this feature has been referred to by
the investigators previously cited, and very much regret that it
was not practicable for me to carry out such additional experi-
ments as would have afforded more definite conclusions. It must
suffice to merely mention the matter, hoping that at some time
someone may be able to secure definite conclusions by extended
experiments not only upon this medusa but perhaps on others
as well.
III. ABNORMALITIES.
In connection with observations upon several specimens which
had become degenerate or perhaps pathologic, resulting from un-
favorable conditions of some of the aquaria, or perhaps in some
cases due to the depleting effects of the experiments, as in the
case of the first experiment cited in this paper, occasion was taken
to examine somewhat in detail the observations and experiments
of Uexkiill and to compare cases coming under my own observa-
tions during the course of the experiments.
In one specimen which had shown evident decline of vigor and
upon which there appeared certain exumbrellar blotches or cor-
88 Charles W. Hargitt.
roslon patches, similar to those mentioned in the earher portion
of this paper and comparable in general aspects to cases men-
tioned by Uexkiill, it was found that after all the rhopalia had
been removed the specimen yet exhibited certain convulsive con-
tractions which at times simulated an irregular rhythm. I there-
fore undertook to repeat several of this observer's experiments
as to the effects of certain chemical stimuli, specially that of com-
mon salt, NaCl. Small crystals of this salt were carefully placed
on definite parts of the sub-umbrellar musculature, and I was able
thereby to confirm in the main his results. There was a very evi-
dent white coloration of the adjacent tissues, and this was followed
by a more or less definite, though somewhat irregular, rhythmic
contraction of the umbrella which continued for perhaps five
minutes. The experiment was repeated several times and upon
different specimens and with usually similar results, though dif-
fering as to vigor or continuity.
Uexkiill had concluded that the recovery of a similar rhythm
In specimens upon which he had experimented by excising the
rhopalia was due, not to any direct restoration of nervous or
other normal equilibrium, but to certain pathologic conditions
which had Intruded themselves, and among which he was spe-
cially Impressed by these corrosion abscesses or disease patches,
to which reference has been made. Doubting whether an agent
of this sort, affecting particularly the exumbrella, could have any
very definite Importance as a center of stimulus, It occurred to me
to vary the experiment by applying the salt to the exumbrellar
region Instead of the musculature of the sub-umbrella, and though
variously repeated the results were uniformly negative In char-
acter, no conclusive responses of any sort being obtained. Nor
was there observed any of the whitening effects which were so
evident In the previous experiments. We may conclude, it seems
to me, that the effects produced by the salt in arousing a simulated
rhythm of contraction was due to the direct action of the sub-
stance on the musculature Itself, and not to any general effect
produced upon the coordinating centers of the medusa. These
stimulating effects of sodium chloride upon muscular tissue are too
well known to call for any special mention In this connection.
Regeneration in Rhizostoma Piilmo. 89
It would seem, therefore, that in the light of these facts one
may well question the validity of Uexkiill's conclusions, or rather
inferences. The mere presence of whitish blotches on an organism
would hardly justify, without the most conclusive demonstration,
the inference that the presence of similar effects produced by some
reagent proved them identical or even analogous. That there
may have been certain pathologic conditions operating upon these
medusae of which the whitish blotches were in some respects ex-
pressions may have some measure of probability. But that these
blotches were in themselves the inciting stimuli giving rise to the
simulated rhythm must be regarded as doubtful, if not indeed,
highly improbable. Such a conclusion could hardly have been
suggested had it been observed that the same whitish blotches are
not unusual on specimens which have been for some time in
aquaria. Moreover, their presence on such specimens has not in
the least, so far as my own observations have gone, served to
Introduce any variation of the normal rhythm, a condition which
might not be unusual on the assumption of these disease patches
becoming sources of abnormal stimuli, and thereby Introducing
erratic or conflicting factors into the physiological processes of
the organism. It Is well that attention should be directed to
disturbing conditions of this character in order that undue weight
be not given to a single factor in determining so Important a prob-
lem. On the other hand It may be quite as Important that In
discrediting one conclusion there Is not substituted another of
even less value.
One might be tempted in this connection to go somewhat out
of the way to consider Uexkiill's conclusions as to the purely
mechanical function of the rhopalia In relation to the rhythmic
action of the umbrella of medusae. If they might be supposed to
act after the fashion of the clapper of a bell, using his figure of
comparison, In the case of such medusae as Rhizostoma what ex-
planation shall we have for the Identical rhythm exhibited by
many other medusae entirely devoid of rhopalia or any equivalent
organ? Many other objections will immediately arise when one
reflects upon the very different histological conditions of structure
found in these organs in various medusae, but to take up any one
90 Charles W . Hargitt.
of these and other phases of the problem would lead too far
afield, and we must satisfy ourselves for the time by the reflection
that while such speculations are interesting as well as ingenious
they are far from demonstrations.
IV. HISTOLOGY.
A brief study of the histology of the regenerated organs shows
the various stages of the process and establishes beyond doubt a
true histogeny, though it has not been possible to demonstrate the
details of mitosis in the proliferating cells. This may be due in
part to lack of just those refinements of technique necessary to
bring out these features. Some of the tissues were fixed by means
of Flemming's solution, some by corrosive-acetic acid, and still
others in lo per cent, formol in water. I have not been able
to distinguish that there was any appreciable advantage in the
one over the others, the formalin seeming to afford equally good
fixation and preservation. Heidenhain's iron haematoxylin and
an aqueous solution of haematein both afforded fairly good differ-
entiation, though they failed as to the nervous tissues, a result
which was not unexpected.
In Figure 2 is shown a longitudinal section of a regenerated
rhopalium at a comparatively early stage, when first distinguish-
able as a minute papilla. In an earlier part of the paper I have
referred to its early appearance as having the character of a solid
bud from the upper angle of the notch made in the process of
excising the organ. From an examination of this figure, which
is among the earliest stages I have been able to satisfactorily sec-
tion, it would seem that in its origin it probably follows the usual
process of the regeneration or development of such organs in the
coelenterates, namely, that of budding, involving both ectoderm
and entoderm. As shown in the figure, there is here a typical
outgrowth from the distal end of the radial canal and, as also
mentioned in another connection, it was easy to demonstrate at
about this stage of development in the living medusa an active
circulation in the bud. The cells of the ectoderm at this stage
are of approximately uniform size over the entire organ, and
the same is also the case with the cells of the entoderm. There
Regeneration in Rhizostoma Pulmo. 91
seems also to be present the middle lamella, though less sharply
defined than at a somewhat later period. There appears to be a
rapid proliferation of the cells of the entoderm near the terminal
portion where they form a mass as shown in the figure, though, as
mentioned above, it was not possible to distinguish evidence of
mitosis.
It is interesting to note at even this early stage the incipient
phases in the regeneration of the sensory areas (Fig. 2, s. e.) just
above and below the rhopalium. The regeneration of the hood
is also shown at h.
\ — eel.
mgl.
s.e. ent.
Fig. 5.
Section of regenerating rhopalium. rh. c, rhopalial canal ; other letters
as in Fig. 2.
In Figure 5 is shown a section taken in the same plane as the
former, but at a somewhat later stage of development. The
rhopalium has apparently attained nearly full size, but lacking as
yet any development of otoliths, though in the network shown at
rh. c. there is apparently evidence of a differentiation preparatory
thereto. There is also shown here the thinning out of the ecto-
derm of the distal portion of the organ as seen at ect.
The sensory areas and epithelium above and below are here
seen to have acquired almost their typical form and character as
92
Charles JV. Hargitt.
shown at s. e. The hood is also shown at h, not having appar-
ently kept pace with the growth of the other organs. Here as
before the direct connection of the radial canal with the rho-
palium is quite broad and characteristic. The middle lamella, or
mesogloea, is shown at mgl. above and below, in the latter ele-
ments of a loose network being traceable, with the embedded cells,
which can also be found indefinitely scattered throughout the jelly.
In Figure 6 we have a section through an almost mature rho-
palium, taken in the same plane as the others, only the organ
n.f s.e.
ect
Fig. 6.
Section of regenerated rhopalium, approaching maturity. It, lithocysts ; nf,
nerve fibers ; other letters as in Figs. 2 and 5.
itself with the terminus of the hood being shown. The ectoderm
has become practically uniform over the entire distal portion of
the organ, but as it approaches the base of the area of the litho-
cysts, shown at It. and generally throughout the entire distal part,
it becomes columnar. At 5. e. it forms a definitely arched por-
tion, the sensory epithelmm, beneath which at n. f. is the so-called
nerve fiber area of the nerve center of this region. While it Is
quite possible to distinguish a more or less fibrous character as
Regeneration in Rhizostoma Pulmo. 93
shown In the figure, It has not been possible to trace these fibers
into any cellular plexus, or ganglion, such as has been claimed to
exist here. Since however my observations In the present instance
have been almost entirely restricted to phases of regeneration, it
will not be pertinent to discuss the question farther.
As will be seen there Is still a continuous connection between
the cavity of the distal portion of the organ and the radial canal.
This connection Hesse, '95, has shown in figures of normal or-
gans in maturity, but In the present examinations I have found it
when fully regenerated to become entirely solid throughout the
llthocyst region, the radial canal ending abruptly at its basal end,
which is shown almost closed In the figure under consideration.
In the rhopalial cavity, rh. c, which at this time Is nearly spher-
ical, there is present a radiating network of delicate fibers, poorly
shown in the figure, which seem to diverge from a point on the
lower surface and extend entirely across the cavity apparently
attaching to the opposite wall. I should consider these fibers of
the same nature as those shown In Figure 5 near the terminus of
the line rh. c. Though It has not been possible to critically trace
the details of the process it seems entirely probable that the ento-
dermic epithelium of this region becomes gradually differentiated
Into fibers which form the Intricate network within which the
llthocysts are later deposited. Within this network may be found
during the various stages of development the gradual metamor-
phosis of this entodermic cell mass, the nuclei of the cells often
remaining as permanent elements of the organ. Some of the
more prominent of these are shown In the figure, and phases of
the metamorphosis may be detected near the narrow slit-like
canal, just beyond the terminus of the radial canal.
Within the network may also be traced the deposition of the
pigment characteristic of the organ.
Concerning the histology of the regenerated oral and gastric
organs it has not seemed essential to make special Inquiry, since
in what has already been shown In connection with the more
highly differentiated tissues of the marginal organs It would seem
that no serious doubt can remain as to normal histogenic pro-
94 Charles W . Hargitt.
cesses probably occurring throughout every regenerating organ in
this medusa.
It was pointed out in connection with the description of certain
experiments that both in form and in function we have among
the Scyphomedusae a regenerative capacity extending to the most
highly specialized organs. In the subsequent account of the his-
tology of the regenerated organs it has been shown that the pro-
cess is a perfectly normal and characteristic one, conforming in
apparently every detail to the course of development of the em-
bryonic history of the several organs.
LITERATURE CITED.
EiMERj Th. — tJber Kiinstliche Theilbarkeit von Aurelia aurita und Cyanea
capillata in Physiologische Individuen. Wurzburg, 1874.
" Die Medusen Physiologisch u. Morphologisch auf ihr Nervensystem.
" Tubingen, 1878.
" Organic Evolution. English Translation. Cunningham, 1890.
Forbes, Edw. — British Naked Eyed Medusae. London, 1848.
Hargitt, C. W. — Recent Studies in Regeneration. Biol. Bui. Vol. I, 1897.
" Experimental Studies upon Hydromedusae. Biol. Bui., 1899.
Hargitt, G. T. — Notes on Regeneration in Gonionemus. Biol. Bui., 1902.
Hesse, R. — Uber das Nervensystem u. d. Sinnesorgane v. Rhizostoma Cuvier.
Zeits. f. wiss. Zool., 1895.
Morgan, T. H. — Regeneration. Macmillan, 1901.
Romanes, G. J. — Jelly-fish, Star-fish and Sea Urchins. 1885.
Uexkull, J. VON. — Die Schwimmbewegung von Rhizostoma pulmo. Mitt. Zool.
Sta. Neapel., XIV, 1900.
Syracuse University,
The Zoologicai Laboratory,
January 20, 1904.
STUDIES ON REGULATION. IV.
SOME EXPERIMENTAL MODIFICATIONS OF
FORM - REGULATION IN LEPTOPLANA.
BY
C. M. CHILD.
With 53 Figures.
INTRODUCTION.
The following observations and experiments on Leptoplana
constitute a part of a series of investigations of form-regulation
undertaken at the Zoological Station at Naples in 1902-3 during
occupation of a table granted by the Smithsonian Institution. A
part of the work on Cerianthus has already appeared (Child '03a,
'03b, '04a.)
The work was undertaken primarily for the purpose of examin-
ing the relations between form-regulation and the mechanical
tensions resulting from the creeping and other movements in some
of the polyclad turbellaria. For this purpose is was desirable
that the direction of movement should be relatively definite in at
least some one of the forms studied. Many of the polyclads
show little definiteness of direction in their movements under
ordinary conditions. This is notably the case in the species of
Stylochus in which, as might be expected, corresponding form-
changes are almost completely absent. Leptoplana proved to be
the only form readily obtainable in large numbers at Naples
which fulfilled these conditions. The creeping movements of this
form are relatively definite in direction.
If form in the Turbellaria is in any way dependent upon the
mechanical strains to which the tissues are subjected, it is to
be expected that in forms with tough, resistant tissues, the changes
in shape resulting from changes in mechanical conditions will
96 C. M. Child.
be less clearly marked than in those with relatively plastic
tissues. As is well known, the tissues of many of the poly-
clads are extremely tough and resistant. Stylochus, Thysanozoon,
and many other forms might be mentioned as examples of this
condition. As regards this feature also Leptoplana proved to be a
favorable form since its tissues are relatively soft and plastic,
though' much firmer than those of Planaria. With this form
as a basis it was possible to make some very interesting com-
parative observations upon various other species, which will be
discussed later.
Nearly all the specimens used were collected about the Castel
dell' Ovo and were presumably Leptoplana tremellans. Since It
is Impossible according to Lang ('84, p. 482) to distinguish with
certainty the species of Leptoplana except by examination of the
copulatory organs In serial sections, my material may have In-
cluded other species — L. alcinoi and L. pallida. Nearly all the
specimens used, however, resembled closely the type of L. tremel-
laris represented In Lang's Figure i, Tafel IIL In no case did
individual specimens exhibit characteristic differences in the regu-
lative processes that could be regarded as specific, so the question
as to the species Is In any event of minor importance for the pres-
ent purpose.
Specimens and pieces were kept Isolated or several together
according to the experiment. In Stender dishes of various sizes,
covered to exclude dust. The water was changed twice a week
or oftener, but the animals proved extremely hardy and capable
of living in small dishes even during summer for much longer
periods without change of water.
An attempt was made early In the course of my work to de-
termine whether the activity of the animals was affected by light.
So far as I could determine, specimens kept In darkness were
slightly more active, but the difference was not sufficiently great
to exert any marked Influence on regulation. Later most of the
specimens were kept In darkness except when under examination.
All were kept without food.
Extensive series of measurements were necessary In the study
of the form changes, and Leptoplana, like most of the turbellaria,
Studies on Regulation. IV. 97
Is not a favorable form for exact measurement. It was necessary
to repeat all measurements several times In order to be certain
that they were approximately correct. In all cases the attempt
was made to secure the measurements while the animal was In
the fully extended condition. In order to accomplish this It was
often necessary to stimulate the specimens to movement and then
to measure them while moving. A small millimeter scale which
could be immersed In the water was used for whole animals and
the longer pieces, while the smaller pieces were measured under a
low power of the microscope with the aid of an ocular micrometer.
In all cases the measurements were reduced to millimeters.
The measurements which were commonly made are as follows :
length of whole body, distance from anterior end of head to
middle of group of eyes, distance from anterior end of head to
anterior end of pharynx, length of pharynx, width of head in
region of the eyes — In whole animals this is usually the widest
part of the body, but In short pieces undergoing regulative changes
of form, the widest part Is anterior to the eyes; in such cases meas-
urement both of the widest part and the eye-region was made — and
finally the width of the body at the posterior end of the pharynx.
In regenerating pieces with new tissue the dimensions of the new
tissue and the position of regenerating pharynx and genital ducts,
If present, were also carefully determined by measurement. Since
a cut surface undergoes marked contraction In Leptoplana and
the new tissue arising from It consequently occupies only a part
of the area of the original cut surface it was necessary In many
cases to determine with great care the width of the body just
anterior to the cut, the width of the new tissue at its origin, the
difference of these measurements representing the degree of con-
traction of the cut surface. Measurements of pieces without ce-
phalic ganglia are not strictly comparable with those of pieces
In which the ganglia are present, since the former rarely extend
fully. It will not be necessary in most cases to give all these meas-
urements in detail since the figures, which are drawn from them
in almost every case, will show the changes with sufficient clear-
ness.
98 C. M. Child.
In all cases where the form of the pieces rendered it necessary
figures were drawn in my notes on the basis of the measurements
and the living specimen. By this means a record was kept not
only of the principal dimensions, but also of any special features,
e. g. the angle between the axis of the new and old tissues, the
curved contours, etc. The actual form-relations and contours are,
I think, shown by the figures as exactly as is possible in a case
where alteration of indiv^idual form is so great. Except where
otherwise stated the figures are about seven times the natural
size. The various internal organs are represented so far as
necessary in a conventional manner.
REGULATION, NUTRITION AND USE OF PARTS.
As a preliminary to the descriptive part of the paper a brief
discussion of certain phases of the problem in hand is desirable
in order to clear the ground.
In no case was the attempt made to feed the pieces employed
for experiment. In consequence of the absence of food a marked
decrease in size occurred during the course of the experiments.
There is no doubt, however, that the results of feeding would be
similar to those obtained by Morgan with Planaria (Morgan,
'oo), for specimens were occasionally found among the worms
collected which had regenerated after a loss of a part of the
body. In these specimens the amount of new tissue formed was
much greater than in the pieces kept without food, and there is
no reason to believe that in pieces where regeneration is possible
growth to the full size may not occur, provided enough material
is at hand. As Morgan ('98 ) has pointed out, the material used
in the formation of new tissue in starving pieces is obtained from
the substance of the piece itself or from its reserve supplies, and
the bulk of the old tissue is reduced to a greater or less extent
by the formation of the new tissue. When the pieces are fed
the amount of new tissue formed is more or less increased and
the old tissues not only do not decrease in size, but may grow
larger.
The fact that formation of new tissue may occur, not only
once but repeatedly, in pieces which have been for weeks without
Studies on Regulation. IF. 99
food indicates that the stimulus which brings about the regen-
eration is sufficiently powerful to deprive the old portions of ma-
terial. There is little doubt that this difference indicates a dif-
ference in metabolic activity between the old and the growing
regions. If this conclusion be correct it follows that in the pres-
ence of nutrition the new parts will grow more rapidly than the
old. Moreover, we find that in the absence of food growth of
the new tissue may cease long before the amount of tissue removed
has been replaced. We must conclude therefore either that the
stimulus to regeneration decreases as regeneration proceeds, or
that the old portions give up material less rapidly as the process
continues (Child, '03d). That there is an actual difference in
quality between the new and old tissue is clearly shown by ob-
servations which I have made repeatedly, viz., that in regenerat-
ing pieces kept without food until death occurs from starvation,
infection or other causes, the old parts usually disintegrate before
the new. In many cases I have seen the old tissue disintegrate al-
most completely in pieces of Leptoplana, while the new tissue re-
mained alive and apparently healthy for a considerable time after-
ward. Moreover, the nev/ parts in regenerating specimens show a
greater degree of muscular and other functional activity than do
the old parts.
For reasons which I hope to state in full at some future time, I
believe that this difference is at least in part dependent upon
functional conditions, viz., that it concerns the use or ac-
tivity of the parts. In the earlier stages of regeneration
other factors are very probably concerned in greater or less
degree. The presence of a cut surface places the cells adjoining
it under conditions widely different from those existing before
the cut was made. The equilibrium in physical conditions is de-
stroyed by the removal of the part and the absence of pressure
from other parts on one side may itself be sufficient to bring about
a migration or growth of tissue outward from the cut surface. It
is extremely difficult in such cases to determine how much of the
" new tissue " is the result of migration and how much of actual
proliferation. But if the attempt is made by the animal to use
the ' new tissue " thus formed in the manner characteristic of
loo C. M. Child.
the part which it represents new conditions of pressure and ten-
sion arise as well as powerful nervous stimuli which probably
affect growth and differentiation directly or indirectly.
When for instance the posterior part of the body of one of
these worms is removed the animal continues to move about and
" attempts " to carry out the same movements as when the pos-
terior end was present, but in the absence of the parts the move-
ments fail more or less completely of success. In fact observation
of these cases leads me to believe that in the absence of the part
the attempts to attain the usual result are often more powerful
than when it is present. For example, a specimen of Leptoplana
with the tail removed makes violent attempts to hold to the
substratum by the cut posterior end of the body as well as by
other parts; a specimen with the lateral lobes of the head removed
makes violent but unsuccessful attempts to swim; and finally, to
take another case somewhat removed from the present considera-
tions, a fish with the tail or part of it removed uses the remaining
stump much more vigorously than would be the case if the whole
were present.
It may appear at first glance that these statements involve
unwarranted assumptions regarding the psychological activities of
forms as low in the scale as the Turbellaria, but I believe such a
conclusion is not justified. Conscious recognition of the successful
or unsuccessful character of the movement is by no means neces-
sary, but on the other hand it is diflicult to understand how these
creeping worms could advance in a regular, definite manner if
the movement over the substratum or the movement of the parts
of the body upon each other did not afford certain characteristic
stimuli. The movements of these forms as well as those of higher
animals are coordinated, and for coordination some stimulus re-
resulting from the movement seems to be necessary. Removal of
a part, e. g., the posterior end by which the animal has been ac-
customed to attach itself, must bring about a change in the rela-
tion of the various stimuli. The animals behave in such cases
as if they were moving over surfaces to which their bodies do
not adhere readily. They appear to make violent efforts to use
the parts which are'missing. These changes in behavior are dis-
Studies on Regulation. IV. lOi
tinct from any irritation due to a wound, for they continue after
the wound has closed and new tissue has appeared. There is
no doubt, I think, that a modification of the motor stimuU occurs
in the absence of a part important to locomotion.
The outgrowth of new tissue from the cut surface is probably,
in its earlier stages, the result of the alteration in local conditions
consequent upon the removal of a part. But the position of the
new tissue, i. e., its connection with a particular part of the old
body determines the conditions to which it is subjected in connec-
tion with the functional activities of the old differentiated parts.
As differentiation in the new tissue proceeds, motor activity ap-
pears and soon the movements of the new part are coordinated
more or less completely with those of adjoining old parts. Thus
the conditions to which the new part is subjected become similar
to those which were present in the part removed. These condi-
tions or some of them are undoubtedly formative factors in many
cases. In Stenostoma (Child, '02, '03) the development of the
tail depends in large degree upon their presence.
Now when the new part first shows characteristic coordinated
motor activity it is much smaller than the part removed, yet func-
tionally it supplies the place of the other, though at first very
imperfectly. But the smaller the size and the more imperfect
the formation of the new part, the greater the activity, i. e., the
"attempt" of the animal to use it. Thus the new part is visibly
more active than the old and if we admit that the conditions con-
nected with this activity are "formative factors" it is easy to see
why in a starving piece the new part continues for a longer or
shorter time to increase in size at the expense of the old tissue.
As the new part increases in size and its coordinations become
more perfect the degree of motor activity decreases, approaching
that of the old parts. The extent of regeneration in starving
pieces is probably determined by the relative functional activity
of the new and old parts. As long as the more intense metabol-
ism of the new part enables it to deprive the old part of material,
so long will it continue to increase in size. It is also probable
that the old tissue gives up material less and less readily as the
encroachments continue. The final result depends on the condi-
tions of the individual case.
io2 C. M. Child.
It is possible that the effect of the functional conditions may
be in many cases largely mechanical, i. e., that in consequence
of the use or attempt at use of a growing part, e. g., a
regenerating tail, it is subjected to certain mechanical condi-
tions of tension and pressure and that these mechanical condi-
tions themselves constitute in reality the chief "formative factor,"
acting either mechanically or as physiological stimuli to growth.
In many cases, however, there is no doubt that other internal
stimuli bring about growth, but even in such cases mechanical
conditions must usually play a certain part in the final arrange-
ment of the material produced. In short there must usually and
perhaps always be a mechanical factor of more or less import-
ance in regulative morphogenesis. I think it probable that in
the lower animals this mechanical factor is relatively simple but
of great importance, while with increasing complexity it becomes
more complex and more difficult of analysis, though perhaps not
less important.
The alteration in general outline and proportion of pieces,
especially of the old portions, called by Morgan morphallaxis,
which occurs during regulation in such forms as Planaria (Mor-
gan, 'oo, 'oi), Stenostoma (Child, '02, '03) and Leptoplana, I
believe to be primarily due to mechanical factors connected with
locomotion and acting very probably both in a simple mechanical
manner and as stimuli to growth, though there is some reason
to believe (Child, '02) that the direct mechanical effect is pre-
dominant in many cases. We cannot conclude, however, that all
phenomena which have been designated as morphallaxis are due
to similar conditions. The changes in form of pieces of the
medusa Gonionemus for instance (Morgan, '99) cannot be due
to the factors which cause the change of form in Stenostoma and
Planaria, but are very probably due to physical conditions in the
tissues whose equilibrium is destroyed by a removal of a part,
and so may be comparable to the inrolling which occurs in pieces
of Cerianthus (Child, '04a). In dealing with problems of so
great complexity generalizations are safe only so far as the actual
facts go. Nothing is gained by referring these diverse phenom-
ena to an inherent capacity in pieces for returning to the original
form. Such an explanation leaves us exactly where we started.
Studies on Regulation. IV.
103
If these views are correct it follows that these form-changes, at
least in the old parts, and often in the new as well, must occur
to a greater extent when regeneration is not quantitatively com-
plete, or must be more evident, since the growth of the
parts may mask it to a greater or less extent. This is actually
.>-p
Fig. 1.
the case, as Morgan's experiments have shown (Morgan, '00).
Since my primary object in investigating the regulative processes
in Leptoplana was the examination of the alterations in propor-
tion, the most favorable* conditions for this purpose, viz., absence
of food, were desirable.
I04 C. M. Child.
On the other hand the study of regeneration in the stricter
sense is not at all impossible under those conditions. The failure
of the new portion to attain full size is a minor matter. Indeed
the presence of food is a complicating factor in the study of reg-
ulation of lower forms, since it renders less possible the distinc-
tion between ordinary processes of growth and the regulative
processes.
THE MOVEMENTS OF LEPTOPLANA.
In Leptoplana, as in Stenostoma (Child, '02, '03a, '03b),
there is a close relation between form-regulation and movement.
A description of the characteristic methods of movement may
properly, therefore, precede the account of experiments.
Locomotion in Leptoplana tremellaris is accomplished in two
ways, by swimming and by creeping. Lang ('84, pp. 634-636)
has described the movements of the polyclads and among them
those of Leptoplana. I desire, however, to consider these move-
ments with special regard to their mechanical effect upon the
tissues and for this purpose Lang's description does not suffice.
Figure I shows the outline and proportion of a specimen in fully
extended condition as when creeping.
Swimming is accomplished by an undulating movement, dorso-
ventrally directed, proceeding posteriorly from the anterior end
of the lateral regions of the head and anterior portions of the
body, the median portions remaining meanwhile almost motion-
less. This method of swimming is called by Lang the flying
movement. In various other polyclads it appears in much more
extreme form than in Leptoplana and in some involves not only
the anterior regions but the whole lateral region of the body as
in Thysanozoon.
It is interesting to note that in all cases where this undulating
movement extends over only a part of the lateral region of the
body, the region involved is the broadest portion of the body.
In Leptoplana it is not sharply marked off from other regions
posterior to it, as is the case in some forms, e. g., Stylochoplana
agiUs (Lang, '84, Fig. 2, Tafel II, also pp. 457 and 636). In
correspondence with the absence of sharp demarcation of the
undulating region in Leptoplana we find that the undulating
Studies on Regulation. IV. 105
movements do not cease abruptly as they pass posteriorly but
gradually decrease in amplitude until no longer visible. During
extreme activity they may extend much further posteriorly than
under ordinary conditions and frequently slight undulations of
the margins appear along the sides and pass posteriorly even
when the animal is creeping. During swimming the anterior
region of the body is considerably broader than in Figure i.
It can scarcely be doubted that these movements play a part
in shaping the regions in which they occur. A comparison be-
tween frequency, amplitude, and force of the undulating move-
ments and the degree of lateral development in the regions in
which they occur is most striking. According to the usual point
of view this correlation between structure and function is merely
one of the many remarkable cases of adaptation, but in my opin-
ion it is, at least in part, the direct result of function in the indi-
vidual. Some experimental evidence bearing on this point will
be offered elsewhere.
As regards the manner in which the movement may affect the
tissues it is not difficult to see that the movement of these parts
to and fro through the water must subject them to tension in the
in the lateral direction. This must affect in greater or less degree
the distribution and arrangement of the plastic tissues composing
the parts. A very simple physical experiment serves to illustrate
this point. A cylindrical or square stick of sealing-wax moved
to and fro in one plane in water sufficiently warm to soften it
will undergo flattening in a plane at right angles to the direction
of movement. The change in form is more strikingly shown if
a rigid axis is present; a mass of wax molded in cylindrical form
about a stiff wire will become in a few minutes a thin, flat plate
decreasing in thickness towards the edges and with a rounded
outline. The mechanical conditions resulting from the move-
ment of the wax through the water are not widely different from
those which the undulating margins of Leptoplana produce. If
the wire axis of the wax be considered as the longitudinal axis the
effect of movement through the water is lateral extension. In
Leptoplana the undulating movement is confined chiefly to the
lateral regions in the anterior third of the body and it follows that
the conditions described are limited chiefly to these parts.
io6 C. M. Child.
There can be little doubt, in my opinion, that these mechanical
conditions constitute a factor in the formation of the broad lateral
regions in Leptoplana and more especially in other forms in which
the undulating movements of these parts occur. In other words
the form is in some degree the result, not the cause, of the char-
acteristic method of activity. The experimental data to be de-
scribed support this view.
In addition to its power of swimming, Leptoplana is able to
creep over surfaces rapidly and in a definite direction. Both
muscular and ciliary activity are concerned in the movements,
but one or the other may predominate according to conditions.
When the animal is moving quietly, as for instance after a
slight stimulation, the cilia afford the chief motive power, although
the slight muscular movements of the margins of the body are
almost constant, portions being lifted from the substratum,
brought forward, and again attached. This muscular play of
the margins is especially marked in the anterior regions but extends
in some degree along the whole side of the body.
After a strong stimulus the movements take on a different char-
acter, becoming chiefly muscular. The portions of the body In
which the undulating movements occur during swimming furnish
under these conditions the chief motive power. Parts of the
margin are lifted slightly, extended in the antero-lateral direction,
and attached to the substratum : contraction of the muscles fol-
lows and the body is drawn forward. These movements occur
in rapid alternation on the two sides of the body and the similar-
ity between this mode of progression and the use of legs can-
not escape the observer. The animals appear almost as if walk-
ing forward.
At all times during creeping movements the body adheres
closely to the substratum as may be demonstrated by sudden at-
tempts to dislodge it. The chief regions of attachment are the
lateral margins and the posterior end. Frequently during creep-
ing small portions of the body margin which adhere more closely
than other parts are stretched posteriorly to a considerable degree
before they are torn away from the substratum. As In many
other Turbellaria the posterior end is an important organ of
Studies on Regulation. IV. loy
attachment although in Leptoplana it is not so exclusively em-
ployed for this function as in many other forms.
Leptoplana differs from many other species of polyclads in the
definite direction of its movements. In some forms, e. g. Stylo-
chus, the direction of movement is very indefinite, movements in
other directions being almost as frequent as anteriorly directed
movements. In Leptoplana, however, the deviation from the
longitudinal direction is slight.
As a general rule the more posterior portions of the margin
and the posterior end itself are used more frequently as organs
of attachment than the more anterior regions.
In consequent of the adhesion to the substratum by the margins
and posterior end, the body of Leptoplana is subjected to mechan-
ical tension in the longitudinal direction, often visibly in a con-
siderable degree, during creeping. As in the case of Stenostoma
(Child, '02, '03), this longitudinal tension constitutes a factor
in determining the general form and outline of the body. The
fact that the margins as well as the posterior end are employed
as organs of attachment accounts for certain characteristic fea-
tures in connection with the form.
From the facts above cited regarding movement we must con-
clude that the posterior portions of the body are subjected more
frequently than the anterior parts to longitudinal tension In con-
sequence of their poslton and more frequent use for attachment,
and moreover, that the tension is greater than that In the anterior
regions since all ciliary impulses and muscular contractions aiding
in forward movement anterior to the point of attachment com-
bine to produce it. This statement is correct in a simple case
but frequently various points along the margin may become at-
tached simultaneously and the tension Is distributed among them.
The continual muscular play of the margins, the rapid transitions
which a given region undergoes from attachment to reattachment
are of course accompanied by great variation in mechanical con-
ditions. The Important point is that the tissues are subjected to
longitudinal tension and the posterior regions more than the an-
terior.
io8 C. M. Child.
This case differs from that of Stenostoma in which the posterior
end alone is the chief organ of attachment. Reference to my paper
on Stenostoma (Child, '02) will show clearly how the character-
istic differences of external form betewen Stenostoma and Lepto-
plana may be correlated with the differences in the mechanical
conditions to which the tissues are subjected.
My experiments also indicate that the use of the anterior por-
tions of the lateral margins in drawing the body forward consti-
tutes a factor in their development. In consequence of these char-
acteristic, frequently repeated movements these parts are sub-
jected to characteristic physical conditions, which, like the longi-
tudinal tension, must exert some influence upon the arrangement
of the cells and tissues.
Anyone who observes the creeping movements of different poly-
clads cannot fail to note the close correlation between the gen-
eral outline of the body and the character of the movement. In
general the forms which advance in a definite direction are more
slender than those like Stylochus whose movements are very in-
definite in direction. In the last mentioned form lateral move-
ment occurs almost as often as longitudinal, a part of the body-
margin being advanced and the other portions drawn up to it by
contraction. The breadth of the body is almost as great as the
length in Stylochus. I am forced to the belief that the forms of
the various species are determined in greater or less degree by the
conditions of tension, resulting from swimming and creeping
movements, to which the tissues are subjected. The experiments
to be described afford strong support to this view.
THE LIMITS OF REGENERATION.
This section includes merely a brief preliminary statement
concerning the power of regeneration in Leptoplana. The phe-
nomena will be treated more at length in other connections.
Complete anterior regeneration never occurs in Leptoplana
when the cephalic ganglia are removed. Removal of all portions
of the head anterior to the ganglia and even including the anterior
part of the ganglia Is followed by rapid and complete regenera-
tion.
Studies on Regulation. IV. 109
When the cut is made at any level posterior to the ganglia
neither the ganglia themselves nor the head are regenerated (Cf.
Lillie, '01).
Posterior regeneration is qualitatively complete at all levels
posterior to the ganglia whether the ganglia are present or absent
in the piece, but pieces cut anterior to the ganglia never regenerate
the ganglia nor the posterior parts.
Lateral regeneration is qualitatively complete when the ganglia
are present, but when they are absent neither they nor the lateral
part of the head removed are regenerated though lateral regen-
eration of other parts may be more or less complete in the ab-
sence of the ganglia. Removal of the right or left half of the
ganglia is followed by complete regeneration from the remaining
half.
In general the amount of tissue regenerated in pieces kept with-
out food is much less than that removed, though all the organs
may be present. The amount of posterior regeneration varies
inversely as the distance of the cut surface from the anterior end.
The size of the piece does not affect the quality of regeneration
and affects the amount only slightly, except on approach to the
minimal size, when a marked decrease in the amount of regenera-
tion occurs. The minimal size of pieces capable of qualitatively
complete posterior regeneration was not determined with exact-
ness, but transverse pieces less than one tenth the length of the
body are still capable of qualitatively complete posterior regenera-
tion and pieces even smaller than this, but containing the cephalic
ganglia, regenerate completely in all directions.
REGENERATION AND MOVEMENT.
Considering first one of the simplest cases, viz., posterior re-
generation from a transverse cut surface we find that in Lepto-
plana, as in Planaria and other Turbellaria, the new tissue which
makes its appearance on the cut surface assumes a rounded outline
and grows or extends posteriorly in the direction of the longi-
tudinal axis, becoming more slender and tapering as regeneration
proceeds. Figures 2 — 4, drawn from careful measurements, will
serve as an illustration of the course of regeneration in such cases.
I lO
C. M. Child.
The cut surface in this piece was a short distance posterior to thc^
cephalic ganglia. Figure 2 represents the piece five days after
section, Figure 3 ten days after section, and Figure 4 twenty-seven
days after section. The new tissue is bilaterally symmetrical at
all times and growth appears to occur most rapidly along the me-
dian plane. The gradual decrease in size of the whole is due of
course to the absence of food. The course of posterior regenera-
tion in Planaria, as described by Morgan and others, is similar,
Fig-. 2.
Fig. 3.
Fig-. 4.
F^g. 5. Fig. 6. .Fig. /,
though the amount of new tissue formed is relatively less than in
Leptoplana; similar results have also been obtained by others
with various forms.
If the cut surface from which regeneration occurs be oblique
instead of transverse the course of regeneration differs in some
respects from that just described. Figures 5-7 illustrate the his-
tory of such a piece, begun on the same day as the pre-
Studies on Regulation. IV. 1 1 1
ceding and examined at the same Intervals. Figure 5 shows the
piece five days after section, Figure 6 ten days after section, and
Figure 7 twenty-seven days after section. In Figure 5 the new
tissue Is symmetrical with respect to the contracted cut surface
but not with the median plane of the animal. As growth proceeds
however, a gradual change In direction of the axis of the new
tissue occurs (Figure 6), until finally this corresponds with the
median plane and approximate bilateral symmetry of the whole
results, though the new tissue Is still unsymmetrical In form since
the surface from which It arose Is oblique.
This change in the direction of regeneration Is also familiar
to students of regeneration, having been described by Morgan
and others for Planaria and other forms. It seems to bear the
stamp of a true regulative process for it brings the parts Into the
position which they must occupy In order to produce a bilaterally
symmetrical whole.
During observations on Planaria in which the change Is well-
marked, the possibility suggested itself that it was primarily due,
not to some Internal factor operating In such manner as to pro-
duce the typical form of the species or an approximation to it, but
rather to the locomotion of the animal in the direction of the lon-
gitudinal axis. It appeared probable that since the new parts
were used for attachment and thus subjected to tension in the di-
rection of the longitudinal axis they were gradually drawn out In
this direction and so a symmetrical whole was produced. This
view was supported by the fact that the change seemed to begin
when the new part became functional. When the new tissue first
appears In these forms It Is apparently little used for attachment
or at least without complete success. Within a few days, how-
ever, the specimens can be seen to adhere closely to the substratum
by means of It, and it is at this time that the apparent change In
the direction of growth first becomes conspicuous.
The question as to the effect of altering the direction of loco-
motion In pieces at once presented itself to me and fortunately I
found In Leptoplana a favorable form for experiments of this
kind. Short pieces from the body of Leptoplana containing the
cephalic gaoglia or a considerable portion of them move In circles
I 12
C. M. Child.
when one side of the body is cut away, since the axis usually be-
comes bent and there is nothing to counterbalance the effect of the
cilia and muscular movements of the opposite side. The results
of experiments with pieces of this kind demonstrated in a most
satisfactory manner the correctness of my belief. Numerous ex-
periments were performed, the results in all cases being unequivo-
cal. In the following sections some of these experiments are de-
scribed.
CIRCULAR LOCOMOTION AND REGENERATION IN PIECES
OF HEADS.
A very satisfactory method of obtaining pieces which move in
curves is that of separating the anterior end by a cut a short dis-
tance posterior to the cephalic ganglia and splitting this piece lon-
gitudinally in half at or near the median line. This method of
preparation is illustrated by Figures 8 and 9. Figure 8 shows the
Fig. 9.
direction in which the cuts are made and Figure 9 the piece after
contraction of the cut surfaces has taken place. It is evident from
the latter figure that the contraction is an important factor in
bringing about circular locomotion. The longitudinal axis of the
piece becomes bent toward the cut side and movement in a straight
line is impossible. The curve of locomotion approximates more
or less closely the curve of the axis but does not necessarily coin-
cide with it since the irregular form of the piece often alters the
direction. The tissue giving rise to the new posterior region soon
begins to show the effect of the direction of movement and the tail
forms at an angle with the old parts. The description of the fol-
lowing series will serve to illustrate the course of regeneration.
Studies on Regulation. IV.
113
I. August 31, 1903. A specimen of average size was prepared
as shown In Figure 10. The greater part of the body was re-
moved by a transverse cut about 2 mm. posterior to the cephalic
gangha and the anterior piece thus obtained was split longitudin-
ally. In this case the longitudinal cut appeared to be coincident
with the median plane, but the differences In behavior of the two
pieces after section Indicated that the left cephalic ganglion was
injured to a greater extent than the right. After section the cut
Fig. 10. Fig, 11.
Fig. 13.
Fig. 16.
Fie- 15.
Fig. 17.
Fig. 18.
surfaces soon contracted thus bending each piece Into a curved
form resembling Figure 9 and In both pieces locomotion diverged
constantly toward the cut side, the pieces thus moving in circles,
as Indicated by the arrows accompanying the figures.
September 3: 3 days after section:
The pieces have assumed the forms shown In Figures 1 1 and
12. New tissue has appeared almost uniformly over the whole
extent of the cut surface. The left plecp (Figure 12) is consid-
erably more contracted than the right piece (Fig. 11) and moves
114 CM. Child.
somewhat more slowly. Numerous experiments to be discussed
later have shown that with increasing injury to the cephalic gan-
glia the rapidity of movement and the degree of extension de-
crease, hence it is probable that the ganglion in the left piece has
suffered greater injury than that in the right.
September 6 : 6 days after section :
At this stage the regenerating posterior end has become distinct
and the first traces of the new pharynx are visible (Figs. 13 and
14). The posterior outgrowth is directed toward the cut side
and Its axis coincides with the curve of locomotion. The new
tissue is now functional to some extent, the tail being employed
by both pieces for attachment.
September 20: 20 days after section:
The two pieces are shown in Figures 15 and 16. The axis of
the new body is distinctly curved in both and the pharynx shows
in each case some degree of curvature. In each the part of the
ganglia removed is regenerating and In connection with It are a
few eye spots. The small new ganglion was distinctly visible in
the living specimens from the dorsal side. Both pieces continue
to move in circles, though with a somewhat larger radius than be-
fore, the change being due to the development and use of the new
tissue along the side of the head, which now evidently aids In loco-
motion and thus counterbalances in some degree the effect pro-
duced by the old parts. Portions of the margin of both the new
tissue and the old can be extended antero-laterally and attached
to the substratum, and tension Is exerted upon the other parts by
muscular contraction of these regions. But the power of the new
portions Is still much less than that of the old parts. Similarly,
in consequence of the curvature of the old parts, itself due to the
contraction following section, the effect of the cilia on these parts
is such as to cause movement in a curved line which Is not yet
counteracted by the cilia on the new parts.
The larger right piece was accidentally Injured at this time and
its fupther history could not be observed.
October 12 : 42 days after section:
Figure 17 represents the left piece at this stage. Considerable
reduction in size has occurred, but the amount of regenerated tis-
sue is relatively much greater than in Figure 16. The direction
Studies on Regulation. IV. 115
of movement diverges less from a straight line than before, and
correspondingly the angle between the longitudinal axis of the head
and new body is decreasing. The new cephalic ganglion is nearly
as large as the old, but the eyes are still less numerous in the new
tissue than in the old. The regeneration of the lateral regions
of the head has proceeded so far that anterior to the eyes the new
portion is nearly as broad as the old.
The change in form of the regenerating lateral margin of the
head is the most conspicuous feature of this stage (compare Figs.
16 and 17) . It has now acquired almost its typical form. More-
over, the curvature of the longitudinal boundary between the new
and old portions, /. e. the longitudinal cut surface, is decreasing.
Observation of the movements of this piece at this stage showed
that the functional activity of the regenerated margin of the head
was very great. It was much used in locomotion, portions being
extended anteriorly or antero-laterally, attached, and then con-
tracted, thus drawing the body forward. Swimming movements
were also often made, though short pieces of this kind do not suc-
ceed in swimming to any extent, being apparently unable to main-
tain their equilibrium in the absence of posterior parts of normal
size. There can be little doubt that the functional activity of
this region has brought about the change in form. Characteris-
tic movements have produced a characteristic arrangement of the
tissues. Moreover, the frequent extension of the margin anteri-
orly followed by attachment and contraction has undoubtedly
aided in forcing the anterior part of the old tissue toward the left
and thus straightening the outline of the cut surface.
October 24: 54 days after section:
As indicated in Figure 1 8 the changes described above con-
tinue. The piece moves still more nearly in a straight line than
twelve days ago and the form is correspondingly altered. The
present stage exhibits one interesting effect of the change in direc-
tion of the tension upon the tissues. During locomotion the
margin at a", including both new and old tissue, is thrown Into
small wrinkles or folds, while the right side of the body Is very
evidently stretched. The folds are clearly the result of the
altered direction of tension In the adjoining parts. The body
grew out in a direction differing considerably from that in which
ii6 C. M. Child.
it extends at present and with the change in position the tension
on the tissues at this point has decreased until now it has become
pressure and these parts are "too long" for the position they must
occupy under the altered conditions, A comparison of Figure i8,
the form of the piece during locomotion, and Figure 19, the form
during rest, when the parts are not subjected to longitudinal ten-
sion, renders it still more evident that the tension due to move-
ment is the cause, not the effect of the change in form. When the
piece is at rest the angle between the original axis and the axis of
the new body is always greater than during locomotion and the
folds at X disappear. In other words the change in direction of
the new body does not precede but follows, and does not even keep
pace with the change in direction of locomotion. These facts
leave no room for doubt that the tension due to locomotion is the
efficient factor.
In a previous section the fact was noted that locomotion in Lep-
toplana is chiefly ciliary when the animal moves quietly, but that
when strongly stimulated the movements are to a large extent
muscular. The same is true of these pieces. When stimulated only
slightly they progress at a uniform rate, largely by means of the
cilia, but under stronger stimulation the margins of the head are
used in the manner described and the body is drawn forward by
strong muscular contractions usually alternating on the two sides.
A marked difference in direction between the two kinds of loco-
motion was observed in this piece and indeed in many other similar
pieces. The direction of locomotion by muscular contraction was
sometimes after strong stimulation in a curve to the left while that
of ciliary locomotion was always toward the right. During loco-
motion the muscular activity of the right side of the head — the
new tissue — appears to be greater than that of the old tissue on
the left. In ordinary locomotion the muscular play of the mar-
gins of this part is much more conspicuous than on the left. Ap-
parently the new parts are in a more active condition functionally
than the old, and doubtless under strong stimulation are capable
of more work. When the piece turns to the left after strong
stimulation the difference in muscular activity between the two
sides Is much more marked, that of the right side being clearly
much greater.
Studies on Regulation. IF.
117
November 8: 69 days after section:
At this time the locomotion of the piece diverges only occasion-
ally and then slightly from a straight line, except sometimes after
strong stimulation when the piece turns to the left. The form is
still more nearly symmetrical (Fig. 20). The folds which were
visible at x in Figure 18 have disappeared in consequence of rear-
rangement or resorption of the superfluous parts (atrophy from
disuse?), though when the piece turns to the left after strong
stimulation folds appear temporarily In this region.
The longitudinal boundary between new and old tissue is now
almost a straight line, i. e. the contraction of the cut surface which
occurred after removal of the right side is now scarcely percept-
Fig. 22.
Fig. 21
Fig 23
X-.
Fig. 24
Fig. 25.
Fig. 26
-■XX
Fig. 27.
ible. This change Is probably, like the contraction Itself, due
primarily to an alteration in mechanical condition, i. e. the changes
in mutual pressure and tension.
II. August 31, 1902. The body was removed by a transverse
cut about I mm. posterior to the cephalic ganglia and the anterior
piece tlius obtained was split longitudinally (Fig. 21 ) . The lon-
gitudinal cut passed a little to the left of the median plane and so
through the left cephalic ganglion. In consequence of the con-
traction of the worm during the operation the course of the cut
ii8 C. M. Child.
was curved as shown in the figure. Only the piece on the right
of the cut will be considered here. After section the cut surfaces
contracted and the anterior end bent over so far that the outline
of the anterior region became almost symmetrical (see the outline
of the old tissue in Fig. 22).
September 3 : 3 days after section :
Figure 22 shows the piece as it appears at this stage. The
contraction has brought the two cut surfaces, originally at right
angles into almost the same plane. New tissue has begun to ap-
pear but there is no marked difference in amount in different re-
gions. That portion of the left cephalic ganglion which remained
in the piece protrudes slightly from the cut surface and is indicated
in the figure by deep shading. The piece moves in rather small
circles as indicated by the arrow.
September 6: 6 days after section:
As indicated in Figure 23 the new tissue, probably the ectoderm,
has united with the protruding portion of the left cephalic gan-
glion and is thus prevented from extending at this point. An-
terior and posterior to this region growth has occurred and in a
curious manner. It appears as if two posterior ends were forming,
one from the lateral region of the head, the other from the poster-
ior cut surface, both of them corresponding in direction to the ten-
sion resulting from locomotion. Both adhere to the surface to
some extent, but the posterior one somewhat more firmly. Several
factors combine to produce this peculiar condition : the piece has
contracted In such a manner that the cut surface from which the
left side of the head would normally regenerate faces somewhat
posteriorly; the union between the protruding nervous tissue and
the new tissue divides the growing region into two parts; and
finally the tissue representing the left side of the head is just be-
coming functional so that its margin reacts to the contact of the
substratum but does not yet extend anteriorly and contract strongly
and aid in locomotion, being instead stretched postero-laterally
since It adheres to the substratum until the forward movement
loosens It. This condition shows very clearly how effective me-
chanical tension may be as a ''formative factor."
Studies on Regulation. IV. 119
September 10: 10 days after section:
Unfortunately the condition described above did not continue,
for the ganglionic mass became separated from the right ganglion
and remained united with the new tissue on its dorsal surface near
the left margin (Fig. 24). In consequence of this change the
regions of new tissue before separated are now continuous and the
outline of the margin is rapidly undergoing alteration. More-
over, the new lateral tissue is now further developed functionally
and its margin reaches forward, attaches itself and contracts in
the characteristic manner, this assisting in the locomotion which
consequently becomes slightly less curved in direction. Corres-
ponding with this increase in characteristic functional activity is
the convex outline of the lateral margin of this region. As long
as it was being subjected to postero-lateral tension this portion of
the margin was in part slightly concave like the sides of a growing
tail or body ( Fig. 23 ) .
The mass of ganglionic substance affords a landmark which
enables us to determine that the terminal portions of the body are
formed first, a conclusion agreeing with that of various authors
in regard to soft parts at least in other forms.
September 20: 20 days after section:
Figure 25 represents the condition at this stage. The direction
of movement is still far from a straight line, though the curvature
is decreasing. The curved body and pharynx require no special
comment. The new lateral region of the head now functions
very activ-ely and a comparison of Figures 23, 24 and 25 shows
that the curve of contraction of the original cut surface is becom-
ing less marked, i. e. the old tissue is being pressed back toward
the right at the anterior end by the active new tissue. Regenera-
tion of the left cephalic ganglion is taking place.
October 12: 42 days after section:
The new lateral region of the head is now so active that it
counterbalances the old part to a considerable extent and the di-
rection of movement is less curved. Figure 26 shows the speci-
men at this stage. The change in form and the growth anteriorly
of the lateral region is marked (compare Figures 25 and 26).
Small folds at x during locomotion indicating the pressure exerted
120 C. M. Child.
by the functional activity of the new part. The angle between
the body and the original longitudinal axis is decreasing, i. e., the
body is swinging into typical position.
October 24: 54 days after section:
At this stage the direction of locomotion approaches still more
closely a straight line, and the form is correspondingly changed
(Fig. 27). The small folds at x due to the pressure of the new
lateral region against the old parts at the anterior end are still
visible, and similar folds appear at xx in consequence of the change
In position of the body. The left margin of the head shows great
activity In the region where the lateral outgrowth Is greatest, and
frequently performs swimming or "flying" movements of some
amplitude. The old portions, on the other hand, are less active.
The regenerated cephalic ganglion is nearly as large as the other
and eyes are present In connection with it.
Loss of the piece a few days later prevented completion of the
record.
Fig. 29.
Fig. 28.
Fig. 30. ^^i-21-
III. August 31, 1902. A specimen was prepared in the manner
described for Series I and II, the longitudinal cut being made
as nearly as possible in the median plane. Probably, however,
it was actually a little to the right of the median plane, since, as
In Series II, a part of what seemed to be the right cephalic gan-
glion protruded from the cut surface of the left piece, the part
used.
Apparently the left cephalic ganglion was more or less injured
by the operation, for during the first two weeks the piece showed
Studies on Regulation. IV. 121
little motor activity. As regeneration proceeded, however, it be-
gan to revolve in circles scarcely greater in diameter than its own
length, appearing almost as if revolving on a pivot.
The extreme curvature of the direction of locomotion renders
the piece of interest and certain points require consideration.
During the first period after section when locomotion was slight
regeneration was almost uniform over the whole cut surface,
i. e., there was no marked extension of the portion representing
the posterior region. Figure 28 represents the piece ten days
after section. A comparison of this figure with Figure 24 the
corresponding stage of Series II In which active locomotion has
occurred during the ten days suggests the possibility that the
tension or other conditions connected with locomotion may not
only determine the direction of outgrowth of new tissue but may
also affect the amount of regeneration, a point which will be dis-
cussed more fully elsewhere.
Within the next few days the piece began to move in the manner
described above and twenty days after section appeared as repre-
sented in Figure 29. The axis of the regenerated portion forms
an angle of more than ninety degrees with the original longitud-
inal axis. The protruding mass of nerve tissue has united with
the new tissue and delayed growth at that point. The rapid
change of form which has occurred in the new tissue during ten
days indicates the marked effect which use of the parts exerts
upon regeneration.
Twenty-two days later, forty-two days after section, the form
was much the same. Figure 30 represents the piece during or-
dinary locomotion. Frequently the piece assumed the form shown
in Figure 3 1 in which the tip of the tail was overlapped by the
lateral margin of the head. Figure 32 represents the form as-
sumed when the piece attached Itself by the tail and contracted,
drawing Itself backward; In this condition folds appear at x,
indicating that pressure Instead of tension occurs In that region.
About ten days later the piece died without further changes.
This case differs from the preceding in that no marked change in
the direction of the body-axis occurred during the whole history,
although the piece lived as long as many others in which the
122 C. M. Child.
change occurred. The continued circular locomotion is of course
directly responsible for the absence of change in form and this
in turn may be due to the delayed regeneration of the right ce-
phalic ganglion and consequent imperfect coordination of the
new lateral margin. As a matter of fact the right lateral margin
of the head appeared much less functionally activ^e than in the
other cases described. Regeneration of the ganglion was delayed
by the presence of the old ganglionic tissue which did not lose its
connection with the left ganglion until about four weeks after sec-
tion. And finally this long-continued attachment of the injured
ganglionic tissue to the left ganglion is doubtless to be ascribed
to the fact that scarcely any locomotion occurred during the first
two weeks after section, so that the new tissues with which the
ganglionic tissue was united were not subjected to tension which
would aid in removing this tissue from the region where its pres-
ence interfered with regeneration. I have no doubt that had the
piece lived sufficiently long before exhaustion occurred, the right
lateral margin of the head would have acquired its characteristic
activity and so would have counterbalanced the motor effect of
the old tissue, thus bringing about the change in direction of the
body-axis which occurred in Series I and II.
In all of the cases described thus far the transverse cut surface,
originally posterior, becomes oblique in consequence of the con-
traction and in most cases the outgrowth of new tissue forming
the body occurs in a direction nearly perpendicular to this surface,
though there is considerable variation in different cases. Thus
in Figures 15 and 16 of Series I the angle between the axis of
the new body and the cut surface is somewhat more than 90°,
while in Figures 23, 24 and 25 of Series II it is approximately
90°, and in Figures 28, 29 and 30 of Series III It is again more
than 90°. These cases therefore are open to the objection that
the direction of growth may have been determined in some degree
by the direction of the cut surface rather than by the tension due
to movement, for it is a well known fact that in many cases regen-
eration takes place chiefly at right angles to the cut surface. Al-
though I did not consider this objection valid I prepared other
series in which the posterior cut surfaces of the pieces were
Studies on Regulation. IV.
123
strongly oblique (Fig. 33) in order to obtain experimental evi-
dence on the question. A few cases from one of these series are
described. These cases show that the direction of regeneration
Fig. 33.
Fig. 39.
Fig. 42
Fig. 34.
Fig. 35.
Fig. 40.
Fig. 41.
Fig. 45.
Fig. 43, Fig. 44.
is not determined, except of course in the early stages, by the sur-
face or surfaces from which it occurs.
124 C". M. Child.
IV. September 3, 1902. Four large specimens were cut in the
manner represented in Figue 33. By this method the relation
between the plane of the cut surface and the direction of locomo-
tion is different in the right and left pieces.
Two pieces of each set are described, the others showing no
additional features of importance.
1. A piece from the left side (see Fig. 33) : Figure 34, seven
days after section; Figure 35, twenty days after section; Figure
36, thirty days after section. In this case the cut surfaces re-
mained nearly in their original relations and the piece was not
greatly bent. Consequently the curvature of the axis and of the
direction of locomotion was not as great in many cases. The
angle at which the body appears corresponds with the direction
of locomotion, but the axis of this region is far from perpendicular
to the posterior cut surface. In later stages this specimen became
symmetrical.
2. A piece from the left side: Figure 37, seven days after sec-
tion; Figure 38, twenty days after section; Figure 39, thirty
days after section. This piece became so bent during contraction
that the posterior cut surface faced somewhat toward the right
instead of to the left as originally (compare Figs. 33 and 37) and
the direction of locomotion was correspondingly curved. The
longitudinal cut was a little to the left of the median plane, thus
injuring the left cephalic ganglion to some extent. The piece
was consequently less active In locomotion and the right margin
did not acquire full functional activity as soon as In many other
cases. There was therefore no marked change in the direction
of locomotion and the curvature of the axis persisted to a great
extent up to the time of death. In this case also It Is evident that
the outgrowth forming the posterior region Is not perpendicular
to the posterior cut surface.
3. A piece from the right side of the head (see Figure 33) :
Figure 40, seven days after section; Figure 41, twenty days after
section; Figure 42, thirty days after section. In this case the
contraction of the piece, though not great, brought the two cut
surfaces almost Into line. The curvature of the regenerating
body is clearly shown in Figure 41. The outgrowth is more
Studies on Regulation. IV. 125
nearly perpendicular to the posterior cut surface in this case than
in the two preceding cases, but this is to be expected from the
position of the latter. In this piece the regeneration of the left
cephalic ganglion was not delayed, the left margin of the head
became functionally active within a month after section (Fig.
42) and reduction of the curvature began and was completed
before death.
4. A piece from the right side of the head: Figure 43, seven
days after section ; Figure 44, twenty days after section ; Figure
45, thirty days after section. The longitudinal cut in this case
injured the right ganglion to some extent. The piece became
greatly bent and simply revolved within a space little greater than
its own size. The posterior cut surface was brought into line
with the longitudinal surface. The posterior region grew out
toward the anterior tip of the head, but not at right angles to the
plane of the cut surface. Within the month the left cephalic
ganglion was partially regenerated and the left margin of the
head attained some degree of functional activity thus reducing
the curvature of locomotion and the axis of the body began to
straighten (Fig. 45). The piece did not, however, attain any-
thing like symmetrical form before death.
A comparison of these four cases renders it sufficiently evident
that the angle between the plane of the posterior cut surface and
the regenerating body may vary greatly, while, on the other hand,
the relation between the direction of the outgrowth and the direc-
tion of locomotion is evident.
In the cases described the relation between the functional ac-
tivity of the regenerating margin of the head and the direction of
locomotion has been pointed out. Attention has also been di-
rected to an apparent relation between the development of this
functional activity and the regeneration of the cephalic ganglion
of that side. The relation of the nervous system to regeneration
will be discussed elsewhere, but the fact may be noted here that
the presence or absence of the one ganglion appears to affect the
regeneration of the lateral margin of the head, but not that of
the posterior region. The latter may be formed in the usual man-
ner when only one ganglion is present, though it is usually not as
126 CM. Child.
long under these conditions as when both are present. I think
there is little doubt that the difference is connected with the func-
tional activity of the parts. Each region develops its character-
istic form only as it is used in the cjiaracteristic manner. In all
cases marked growth anteriorly and laterally of the new margin
of the head has been observed as soon as the animal begins to use
it in the manner characteristic of these regions, while the case
represented in Figure 23 shows that so long as this region is not
used in the ordinary manner it may develop posteriorly. The
posterior region of the body can perform its functions to some
extent in the absence of the cephalic ganglia as will be shown else-
where, and moreover, in the cases under consideration the regen-
erating posterior region is undoubtedly innervated from the part
of the nervous system present. It therefore performs its usual
functions, though perhaps less perfectly, before the other ganglion
regenerates, and its development proceeds in the typical manner.
To put It briefly, the margin of the head develops a characteristic
form because used in a characteristic manner, and the body de-
velops a different form because it is used differently. The preced-
ing experiments are sufficient to show that mechanical tension Is
an important factor In morphogenesis in these animals. No one
would admit more readily than myself, however, that many other
factors may be concerned here and that in other cases the factors
may be wholly different.
Attention has been called In several cases to the apparently
greater functional activity of the regenerated parts as compared
with the old In later stages. This difference indicates, I believe,
a real physiological difference. The old portion decreases in size
In consequence of loss of material while the new parts increase
In absolute size in the earlier stages and in relative size in later
stages. It is not at all Improbable that the functional condition
of the reduced old part differs widely from that of the new part.
The former, reduced to a fraction of Its former size, Is certainly
less plastic mechanically and probably less sensitive to stimuli.
The new part Is to be regarded as possessing the qualities of a
young and growing organism, the old on the other hand as ap-
proaching exhaustion. The difference observed in functional ac-
Studies on Regulation. IV. 127
tivity between new and old portions agrees well with the fact
already mentioned in the section on "Regulation, Nutrition and
Use of Parts," that in starving pieces the old part usually dies and
disintegrates before the new, which may live and remain appar-
ently healthy for several days after the loss of the old part.
CIRCULAR LOCOMOTION AND REGENERATION AFTER
OBLIQUE SECTION.
Another method employed for bringing about circular loco-
motion was that of making oblique cuts at various levels posterior
to the cephalic ganglia. The results obtained by this method are
in some respects less striking than those described in the preceding
section, but since the cephalic ganglia are not injured in any way
by this method, a possible objection to the preceding series is ren-
dered invalid.
The change in direction of the longitudinal axis of the regen-
erating tissue arising from a posterior oblique cut surface has
been mentioned (see also Figures 5-7). In this case the direc-
tion of locomotion was not altered by the cut, and, as might be
expected, the axis of the new tissue soon became coincident with
that of the old. In the course of similar experiments I found,
however, that if the cut were very oblique the contraction of the
cut surface following the operation might bring about circular
locomotion. Figure 46 represents the manner in which such a
cut is made, and Figure 47 the piece after contraction. It Is
evident that contraction produces a marked curvature In the lon-
gitudinal axis, and therefore the piece In advancing turns con-
stantly toward the cut side. It was found necessary to make the
cut in the anterior region of the body in order that markedly
circular locomotion might occur, because if the cut were made
near the middle of the body or in the posterior half, the part
of the body in which the axis was not affected by the cut was so
long that the bilaterally symmetrical Impulse to movement from
this region nearly or quite counterbalanced the effect of the bent
portion, and the regenerating tissue grew out in the direction of
the old axis. In all cases where the circular locomotion was well
marked the cut was made either just anterior to the pharynx as
128
C. M. Child.
Fig. 4.6.
Fig. 49.
Fig. 4/.
Fig-. 50.
Fig. 48.
Fig. 51.
Fig. -52.
Fig. 53.
Studies on Regulation. IV. 129
In Figure 46 or through Its anterior portion. Within certain
hmlts the circular locomotion Is more marked as the obliquity of
the cut Increases, the reason being clear. If, however, the cut
be nearly longitudinal the slender strip on the longer side Is likely
to roll up and may act as a drag, thus complicating locomotion
and delaying or preventing typical regeneration.
Figures 48-53 Illustrate the history of a piece cut in the manner
indicated in Figure 46 and moving In a curve toward the right.
Even as early as three days after section (Fig. 48) the new tissue
is symmetrical with respect to the cut surface, evidently In con-
sequence of the effect of locomotion. Figure 49 shows the con-
dition six days after section. Here the curvature of the new
tissue is becoming evident. In Figure 50 — tw^elve days after
section — the curvature of the regenerating part Is still more con-
spicuous. Twenty-eight days after section the tissue has acquired
the form shown In Figure 51. Now that the regenerated tissue
has attained a considerable length and its posterior region con-
stitutes the chief organ of attachment some degree of straight-
ening occurs. The manner in which this takes place is shown in
Figures 52 and 53, both of which represent the same stage — fifty-
eight days after section. Figure 52 shows the piece In ordinary
locomotion. Here the margins as well as the tip of the tail, or
frequently only the margins or certain regions of them, adhere to
some extent, and the resulting tensions cannot cause straightening
of the longitudinal axis since they follow approximately the same
curve. Frequently, however, only the posterior end of the body
adheres, or the margins of the head and the posterior end, and at
such times the piece assumes temporarily the form of Figure 53,
in which the axis is nearly straight. The longitudinal tension to
which the body Is subjected straightens it, but at the same time
bends the contracted posterior part of the old tissue to the left so
that the outline becomes convex at the left of this region. At the
same time small folds appear at x, indicating that In this region
the tissues are subjected to pressure instead of tension. This posi-
tion Is never maintained for any length of time, and as soon as
the tension ceases the piece resumes the form of Figure 52. There
can be no doubt, however, that if this position is taken sufficiently
130 C. M. Child.
often straightening will occur. This I believe is the chief factor in
the change from the curved to the straight bilaterally symmetrical
form. This piece died after sixty-six days without having become
completely symmetrical.
The history of other pieces prepared in a similar manner Is
essentially the same, though, as has been mentioned, the degree of
curvature of the new tissue varies within certain limits with the
obliquity of the cut and with the level at which the cut is made.
The reason for variation in curvature with the angle of the cut
lies in the fact that the more oblique the cut the more the axis of
the piece is bent and consequently the greater is the curvature
of locomotion.
As regards the level, the curvature of the new tissue decreases
as the distance of the cut from the anterior end increases, because
the region of the body in which the axis is not bent by the con-
traction of the cut surfaces increases and counteracts the asym-
metrical motor effect of the bent region more and more com-
pletely. This variation of curvature with the level of the cut
renders it evident that the cut surface itself has little influence
upon the direction of growth except in the earliest stages, for a
piece cut at a given angle near the anterior end, e. g. as in Figures
46-53, will give rise to new tissue with a marked curvature, while
in another specimen cut at the same angle but farther posteriorly
the new tissue will show much less curvature while in still another,
cut at the same angle in the posterior pharyngeal region, the
new tissue will grow out in the direction of the longitudinal axis
or will very soon acquire this direction, simply because the direc-
tion of locomotion is curved only very slightly or not at all.
In each case the direction of outgrowth coincides with the line
of locomotion,
GENERAL CONSIDERATIONS.
The bearing and the significance of the experiments described
is sufficiently clear to render extended discussion unnecessary. They
may, I think, be regarded as demonstrating the fact that the ex-
tended form of the body in Leptoplana is determined in large
degree by mechanical conditions. It is difficult to describe ac-
Studies on Regulation. IV. 131
curately the continually changing movements of these animals, but
I think no one who actually observes the pieces can fail to be
convinced of their importance in determining form. It is true of
Leptoplana as of Stenostoma (Child, '02) that it has no "normal
form," t. e., no definite hard and fast form inherited and devel-
oping In the Individual Independent of physical conditions. What
these animals do possess Is a capacity for certain kinds of activity.
These are given potentially in the chemical and physical structure
of the protoplasm, which to my mind represents rather capacity
for functional activity in the broadest sense than form. As my
experiments prove, certain elements of form In the morphological
sense develop incidentally as the result of functional activity In
in a given environment. These elements have been commonly
regarded as typical and determined by heredity because they are
common within certain limits of variation to all individuals of
a species, but when we consider that under natural conditions
both functional activity and environment are essentially similar
in different individuals of the species the reason for likeness In
these form-elements becomes clear. It is only when we can alter
the functional activity as I have done experimentally in the case
of Leptoplana, or the environment, as I succeeded in doing for
Stenostoma (Child, '03a) that the dependence of these elements
of form upon these two factors becomes clearly evident.
But these experiments concern only morphological characters
of a certain kind. Experiments of others have already shown that
"formative factors" are many and various, and generalizations
from the consideration of a single group of characters are unsafe.
It win never be possible to explain form on the basis of a single
principle. All the complex activities of which organisms are
capable are "formative factors" : when we can view all of these
In their complex interrelations and know the part which each
plays, then and only then shall we "understand" organic form.
The relation between form and heredity has never been satis-
factorily determined. With the advance in our knowledge the
fact becomes more and more evident that the organism is not
merely a complex of structural elements ready made by heredity
for certain functional activities, but rather a complex of ac-
132 CM. Child.
tlvlties in consequence of which morphological structure develops
Physical and chemical structure of protoplasm must not be con
fused with morphological structure : the distinction between the
two is important though often overlooked. As regards the indi-
vidual the former represents capacities for activity, i. e., for
transformation and transference of energy, or in short, functional
activity in the broadest sense. Form in the morphological sense,
is the combined result of this activity and the erxvironment, ex-
ternal or internal. According to this view, it is functional capacity
that is inherited rather than form : heredity is, strictly speaking,
a physiological and not a morphological problem.
SUMMARY.
1. Locomotion in Leptoplana is accomplished by two meth-
ods, swimming and creeping. In swimming the lateral regions
of the anterior part of the body perform undulating movements
in a dorso-ventral direction. Creeping movements are both ciliary
and muscular, the muscular movements consisting chiefly of an
alternate extension anteriorly or antero-laterally of the margins
of the head, adhesion to the substratum and muscular contrac-
tion, thus drawing the body forward. In creeping the margins
and posterior end of the body are used as organs of attachment.
2. In consequence of the typical movements the tissues of
the body are subjected to typical mechanical tensions and pres-
sures which constitute formative factors.
3. The effect of these mechanical conditions upon the tissues
is at least in part directly mechanical, but they may also act as
physiological stimuli to growth (formative stimuli).
4. The effect of the mechanical conditions incident to locomo-
tion may be demonstrated experimentally by various methods.
The method described in this paper consists in making the pieces
of such a form that the direction of locomotion becomes curved
instead of straight. In these experiments the regenerating part
grows in the direction of the principal tension, even though this
form an angle of 90° with the typical direction of growth.
5. The experiments lead to the conclusion that in Leptoplana
the regions of the body develop in a characteristic form because
they function or attempt to function in a characteristic manner.
Studies on Regulation. IV. 133
BIBLIOGRAPHY.
Child, C. M., '02. — Studies on Regulation. I. Fission and Regulation in Ste-
nostoma. 4, Archiv. f. Entwickelungsmech., Bd. XV., H. 2 & 3, 1902.
'03a. — Studies on Regulation. II. Experimental Control of Form-Regu-
lation in Zooids and Pieces of Stenostoma. Archiv f. Entwick-
elungsmech., Bd. XV., H. 4, 1903.
Child, C. M., '03b. — Studies on Regulation. III. Regulative Destruction of Zo-
oids and Parts of Zooids in Stenostoma. Archiv. f. Entwick-
elungsmech., Bd. XVII., H. I, 1903.
'03c. — Form-Regulation in Cerianthus. I. The Typical Course of Re-
generation. Biol. Bull. Vol. v., No. 5, 1903.
'03d.— Form-Regulation in Cerianthus. II. The Effect of Position, Size,
and other Factors upon Regeneration. Biol. Bull., Vol. V., No.
6, Vol. VI., No. I, 1903.
'04a. — Form-Regulation in Cerianthus. III. The Initiation of Regenera-
tion. Biol. Bull., Vol. VI., No. 2, 1904.
Lang, A., '84. — Fauna und Flora des Golfes von Neapel. XI. Die Polycladen.
Leipzig, 1884.
LiLLiE, F. R., '01. — Notes on Regeneration and Regulation in Planarians. Amer.
Journ. of Physiol., Vol. VI., No. 2, 1901.
Morgan, T. H., '98. — Experimental Studies of the Regeneration of Planaria
maculata. Archiv. f. Entwickelungsmech., Bd. VII., H. 2 & 3, 1898.
'99. — Regeneration in the Hydromedusa Gonionemus vertens. The
Amer. Nat., Vol. XXXIIL, 1899.
'00. — Regeneration in Planarians. Archiv. f. Entwickelungsmech. Bd
X., H. I, 1900.
'01. — Regeneration. New York, 1901.
SELF-FERTILIZATION INDUCED BY ARTIFICIAL
MEANS.
BY
T. H. Morgan.
It has long been known that the pollen of some plants will
not fertilize the ovules of the same plant. The cause of this
impotence has not yet been detected.
It is also known that pollen from another plant is often pre-
potent in those cases where normal self-fertilization may occur.
It has further been shown, especially by Darwin, that the offspring
from self-fertilized ovules are in general not so vigorous as those
from cross-fertilized ones.
There are here two problems, which, even if they should prove
to be fundamentally related, can be most profitably examined
separately; — first, the problem of the Inability of the male ele-
ment to fertilize the female germ-cells of the same individual;
and, second, the effect of self-fertilization (in those cases In which
It occurs) on the offspring. Both problems appear to be within
the range of experimental examination.
There are only a very few cases known amongst animals where
conditions similar to those in plants have been found to prevail,
although very few hermaphroditic animals appear to have been ex-
amined in this respect. Close inbreeding, which is commonly sup-
posed to bring about deterioration in some cases. Is perhaps not
very dissimilar to self-fertilization. Whether in the case of in-
breeding there is ultimately a loss of power to fertilize the egg,
or whether the egg fails to develop after It has been fertilized,
has not, so far as I know, been determined.
Castle discovered in the ascidian, Ciona intestinaUs, a case ap-
parently similar to those in plants. The eggs are generally In-
capable of self-fertilization, yet can be readily cross-fertilized; i.e.,
136 T. H. Morgan.
the spermatozoa of an Individual will not fertilize the eggs of
that individual, but have the power to fertilize the eggs of any
other individual.
My object in undertaking a study of this problem was, in the
first place, to determine if possible the nature of the conditions
that prevent or interfere with self-fertilization; and in the second
place, I was not without hope of being able to find some way
in which self-fertilization could be artificially Induced. As will
appear in the sequel, these two questions are not two sides of
the same problem ; for, while it has been possible to discover the
means of bringing about self-fertilization, it still remains to be
definitely determined what conditions In the egg normally prevent
the entrance of the spermatozoa of the same Individual.
Since Castle's observations had shown that the ascidlans offer
favorable material for a study of this sort, I first turned my
attention to this group, using the three most available species
found at Woods Hole, or in the vicinity; namely, Ciona intes-
tinalis, Molgula manhattensis, and Cynthia partita {Styela sp.).
The work was done while holding the Bryn Mawr Table at the
Marine Biological Laboratory, from June to September, 1903.
Owing to the scarcity of Ciona I have not been able to work out
completely a number of Important problems connected with one
of the two main questions that I examined. In the near future
I shall hope to complete this side of the investigation.
EXPERIMENTS WITH CIONA INTESTINALIS.
The ovary of Ciona Is a sac-shaped body of fair size lying
loosely attached in the coil of the Intestine. It can easily be
removed without cutting Into the testis. Its lumen contains some
of the ripe eggs, but the majority of these are in the oviduct.
The oviduct can readily be opened and the eggs set free without
cutting into the vas deferens, which follows a course parallel to
the oviduct. If the animal is kept Isolated for 24 hours the ovi-
duct becomes greatly distended with eggs, and after another 24
hours even more eggs may have accumulated. The eggs are
laid normally In the early morning, at dawn, and Castle has re-
corded that Ciona deposits its eggs and sperm with the regularity
Self -Fertilization Induced by Artificial Means. 137
of the rising sun. The rough handling incidental to removal and
isolation appears to cause Ciona to retain its eggs for several
days. The individuals to be used were isolated, as a rule, from
24 to 48 hours, and in most cases were rinsed in fresh water before
opening. It was not found necessary to boil the water; for check
experiments "showed that eggs left to themselves were never fer-
tilized by stray spermatozoa in the sea-water. Since Ciona de-
posits its eggs only in the very early morning, the chances are
very slight that functionally active spermatozoa would be present
in the sea-water in the late morning and in the afternoon when
the experiments were carried out.
The eggs of Ciona are surrounded by a rather thick membrane.
Standing out like broad spikes over the surface of the membrane,
and forming a beautiful aureole around the egg, are the trans-
parent follicle cells, each with a shining drop in its outer end.
A'number of preliminary experiments confirmed Castle's con-
clusion that self-fertilization is rarely possible in Ciona intes-
tinalis. The evidence, however, on which Castle based this con-
clusion is not altogether satisfactory, since he records many cases
in which self-fertilization occurred. Instances are cited in which
isolated individuals gave 90, 25, 16, 5, 4, o per cent, of self-
fertilized eggs. Castle supposes that, in the first of these cases
at least, the spermatozoa of one day fertilized the eggs of the
next, but it has not been shown that the spermatozoa have this
power if left so long in sea-water. The same individuals that
had been used for these isolation experiments were killed (after
being washed in 90 per cent, alcohol), and the eggs and sperm
of each taken out and mixed together. The results gave 50, 4,
I, Ys, o per cent, of self- fertilized eggs. The same exj^rriment
repeated with fresh individuals gave 50, 12^, 10, 5, 2, o per cent,
of self-fertilized eggs. From these figures it is clear that in some
cases a considerable amount of self-fertilization occurred, unless
there was some source of error in the experiment. In fact. Castle
believes that in those cases where a large number of eggs were
fertilized there was some contamination. My own results with
Ciona have never given so large a percentage of self-fertilized
eggs, and I am inclined to attribute this result in part to the
138 T. H. Morgan.
precaution that I took to isolate the individuals the day before
they were to be used. I have rarely seen more than from i to 10
per cent, of self-fertilized eggs segment, and in the greater num-
ber of cases not a single egg segmented. On the other hand I
found, as did Castle, that as a rule 100 per cent, of cross-fertilized
eggs develop, to which statement I should add, provided the
spermatozoa are in "good" condition.
What is the meaning of these remarkable facts? Why do
not the sperm fertilize the eggs produced by the same individual,
and yet fertilize those of any other individual? A number of
possibilities readily suggest themselves, and since the following
pages record an attempt to test these suggestions they may be
briefly mentioned here :
1. That the spermatozoa are not made sufficiently active by
secretions from the eggs of the same individual, but by those from
the eggs of any other individual.
2. That the spermatozoa are not "attracted" to the eggs of the
same individual.
3. That the egg contains, or secretes some substance that les-
sens the activity of the spermatozoa of the same individual.
4. That some mechanical difficulty prevents the spermatozoon
from entering the egg of the same individual.
5. That even if the spermatozoon enters, it can not fertilize
the egg of the same individual, in the sense of causing the egg
to begin to develop.
In order to discover if the lack of power to self-fertilize the
eggs is due to the absence of some substance around the eggs that
excites the spermatozoa, the following experiment was carried
out. The eggs of an individual (A) were taken from the ovi-
duct. Similarly the eggs of another individual (B) were also
taken out. Then the ovary of (A) and that of (B) were
crushed separately, and a little sea-water was added. The eggs
of (A) were then allowed to soak in the crushed ovary extract
of (B) and those of (B) in the extract of (A). After a short
time the sperm of (A) with a little water was added to the (A)-
eggs, and the sperm of (B) to the (B)-eggs. If the sojourn
of the eggs in the extract of the ovary of another individual has
Self-Fertilization Induced by Artificial Means. 139
the postulated effect, or if the presence of the extract of the ovary
of another individual has the postulated effect on the sperm, fer-
tilization ought to have occurred. The results showed, however,
that fertilization did not take place.
This experiment was performed four times, giving eight sets
in all. In six of these sets not a single egg segmented. In two
others a very few eggs segmented (6 per cent, in one, 5 per cent,
in the other), but this sometimes occurs in self-fertilized eggs
not treated in any special way. Moreover there may have been
contamination in the latter case.
Another experiment similar in some respects to the last was also
carried out. The heart of one individual was opened and the
blood collected. The eggs of another individual were put into
this blood and allowed to stand. Later, sperm of the same indi-
vidual was added in sea-water, but no fertilization occurred in
one set and only one per cent, in the other. Check eggs were also
kept in this experiment to make certain that no sperm had acci-
dentally gotten into the blood. That none were present was
shown by the fact that no fertilization took place. It is evident
from this experiment that self-fertilization can not be brought
about by soaking the eggs in the extract from the ovary or in the
blood of another individual, although the somewhat high per-
centage of self-fertilized eggs that segmented in two cases after
treatment with the ovarian extract may have resulted from the
influence of the extract on the spermatozoa.
If the spermatozoa are excited to greater activity by the pres-
ence of the eggs of another individual it seemed not improbable
that this might be directly observed. Therefore, I placed some
of the sperm with the eggs of another individual and more
of the same sperm with the eggs of the same individual, and
compared the two preparations under the microscope. The sper-
matozoa of Ciona are not very active as a rule, nor do they
accumulate in crowds around the eggs, as they do in many other
animals, or at least not to any marked extent. It seemed to me
in both cases that sometimes the spermatozoa were more active
immediately in the vicinity of the eggs, And in the spaces between
the follicle cells, but as they also show the same activity around
140 T. H. Morgan.
pieces of the tissue of the body of the same or of another In-
dividual I have not laid much stress on this observation, or ac-
credited the results to the presence of an exciting substance. At
times I have thought that the spermatozoa were more active
around the eggs of another Individual than around the eggs of the
same Individual, but as there Is no very accurate means of determin-
ing their relative motility, unless very marked, I should not wish,
as yet, to give a final answer to this question. It is certain that there
Is no such great difference In the behaviour of the spermatozoa
In the presence of the eggs of the same and of another individual
as to suggest that the difference In the result Is connected with this
factor. And even if this were the case, the Influence probably
extends for only a short distance from the surface of the egg, as
the following experiment shows.
The eggs were taken from the oviduct, great care being taker
not to injure the sperm-duct. The eggs from another Individual
were collected In the same way. An equal number of eggs from,
each were put together and fertilized with the sperm from one
of the individuals. In another dish another lot of the same eggc
were mixed half and half, and these fertilized with the spern
from the other Individual. In each of these two sets half at
least of the eggs should be fertilized by the other sperm, but half
should not be fertilized unless the eggs of one Individual exer.
some Influence that causes the sperm to fertilize the eggs of the
same individual also. It was found that only about half of the
eggs were fertilized. This result shows that the fertilization I
probably not due to some substance set free by the eggs that act?
on the sperm or at least that if such a substance is set free Its
action Is confined to the Immediate vicinity of the egg. The
experiment does not show, however, whether the egg, or its
membranes, may not contain some substance that prevents the
spermatozoa from entering the eggs of the same Individual. Ever
If such a substance Is set free from the eggs it may not have had
time In my experiment to accumulate sufl'iclently In the surround-
ing water to have prevented the spermatozoa from fertilizing the
other eggs, which may be quickly entered. This view can be
tested by letting eggs stand In a small amount of water for c
Self -Fertilization Induced by Artificial Means. 141
long time, then taking out some sperm from the same individual,
first making it active by placing it in sea-water, and then putting
it into the water in which the eggs have stood. On the hypothesis
these sperm should soon be brought to rest, and if then the eggs
of another individual are added, they should not be fertilized, o
at least not in the same proportion as when the sperm is taker
directly from the oviducts, put into sea-water, and then added t
the eggs.
THE INFLUENCE OF ETHER ON CROSS-FERTILIZATION.
EXPERIMENTS WITH CIONA.
My first experiments with ether were made in order to deter-
mine whether when eggs are etherized it might not be possible to
self-fertilize them. The results turned out somewhat differently
from what I had anticipated, for although I found that it was
possible to self- fertilize the eggs in ether-solutions, the result
seemed to be due to the action of the ether on the sperm rather
than on the eggs.
The experiment was first made with Cynthia, which in most
cases has very sluggish spermatozoa. I observed that the first
effect of the ether was to make the sluggish sperm very active, and
even greatly quickened the activity of already active sperm. Fur-
thermore I found that spermatozoa that scarcely moved at all
in sea-water became active in the ether-solutions. Finally I found
that in ether-solutions of certain strengths the eggs of Cynthia and
of Ciona could be self-fertilized. The eggs behave in this respect
so capriciously that I was obliged to carry out a large number of
experiments in order to determine the conditions that lead to the
self-fertilization of eggs in ether-solutions. The outcome was
only partially satisfactory, but the experiments opened up a field
for research, in which it may be possible to obtain further results
of interest.
The experiments with ether were carried out as follows : At
first I used a nearly saturated solution of ether and diluted it a
half, or a fourth, etc. In the later experiments I used solutions
of known strength. It was found by trial that the solutions were
effective between 0.25 and 5 per cent. Some of the results may
now be given in detail.
142 T. H. Morgan.
Experiment I. The eggs were removed from an individual that
had been isolated 20 hours. The sperm was also taken out, and,
together with the eggs, was put into ether-solutions, 5, 2, i, 0.7,
0.5 per cent, in sea-water. After 5 minutes, and again after 10
minutes, the eggs were removed to pure sea-water. The eggs
were injured by the ether in the strongest solution, but neverthe-
less one segmented. In all of the other solutions about 80 per
cent, of the eggs divided; the most in the weaker solutions.
Experiment II. In this experiment the eggs and the sperm
were put together into ether solutions of 5, 2, i, 0.7, 0.5 per cent.
Some of the eggs were transferred to water after 5 and 10 min-
utes, but others were left in the solutions. In the strongest solu-
tion the eggs w^ere killed. In the others the following results were
obtained:
Eggs segmented Eggs segmented.
Ether 5 minutes in ether. 10 minutes in ether.
2. percent. 20 per cent. 25 per cent.
I. " 5 " 75 "
0.7 " 2 " 2
0.5 " 10 " 5
.J't is clear that the stronger solutions gave the best results, and
that ten minutes immersion was better than five minutes. None
of the eggs that were left in the ether-solutions segmented. This
does not mean that they were not fertilized, but that the ether so
injured the eggs after a long immersion, that they failed to de-
velop. Several check experiments were also made in this case.
In one the eggs were not self-fertilized but were put into a 5
per cent, ether-solution, and transferred after ten minutes to
sea-water. They did not segment, nor did a few that were left
behind in the ether solution. In another check series the eggs
were not fertilized, and were left in sea-water. None segmented,
which shows clearly that the ether in the preceding experiment
was in some way responsible for the self-fertilization of the eggs.
It should also be recorded that tadpoles developed from all the
fertihzed eggs that had been in the ether-solutions.
Experiment III. The eggs and sperm of the same individual
Self-Fertilization Induced by Artificial Means. 143
were put into ether-solutions of 3, 2, i, 0.7, 0.5 per cent., and
were removed to sea-water after 10, 20 and 30 minutes.
Ether
10
min.
20
min.
30
min.
3-
P'
er
cent.
20
P
er cent.
20
per cent.
per
• cent.
2.
u
33
u
17
((
90
((
I.
u
50
u
25
a
10
((
0.7
u
50
a
20
u
u
0.5
u
20
a
20
u
12
((
The table shows that eggs segmented in all of the solutions,
best however in the stronger solutions, although in one case the
eggs became so injured by the ether that they did not develop
further than the segmentation stages. In a check series, in which
the self-fertilized eggs were put into sea-water, about ten per cent,
of the eggs segmented. There may have been some source of con-
tamination, or else, and this seems more likely since the indi-
vidual had been isolated 20 hours, self-fertilization took place
on a larger scale than usual.
Experiment IF. Eggs and sperm were mixed in ether-solutions
of 4, 2, I, 0.5 per cent.
Ether 2 min. 4 min.
4. o o ■ «*-,-
2. 40 I
I- 35 30
0.5 5 I
These results show that the 4 per cent, solution was too strong,
while the 0.5 per cent, solution appears to have been too weak.
The injurious action of the 4 per cent, solution appears to have
been mainly on the sperm rather than on the eggs, for these
eggs after they had been in sea-water 4 hours were capable of
being cross-fertilized, and 25 per cent, of them developed.
Experiment V. Eggs and sperm were put into a 2 and into a
0.5 per cent, ether-solution and removed after 5 minutes.
ther
5 mm
2.
90
0.5
90
144 T. H. Morgan.
It Is Interesting to note In this case, In which so large a per-
centage of the eggs were self-fertUIzed In ether, that of several
hundred eggs of the same Individual, to which sperm was added,
but which were kept In sea-water, not one segmented. It was
also found, and will be referred to again later, that when the
sperm alone was put into a 2 per cent, soluton, of ether for five
minutes, and was then added to eggs of the same individual In
sea-water, 70 per cent, of the eggs segmented.
Experiment VI. In this experiment the eggs and the sperm
were put into ether-solutions of 2 and of 0.5 per cent, and
removed after ten minutes. In one lot the eggs were self-fertil-
ized, in the other they were cross-fertilized.
Self-fert. Cross-fert.
• Ether 10 min. 10 min.
2. o 100
0.5 o 100
In this case although no self-fertihzation took place, all the
crossed eggs which had also been In the ether-solution developed,
showing that the solutions have no baneful effect on cross-fertiliza-
tion. The lack of self-fertilization shows that the sperm were
not sufficiently acted upon by the ether-solutions employed to effect
self-fertilization.
Experiment VII. This experiment shows how slight a differ-
ence in the conditions may cause great differences In the result.
The Individuals had been Isolated 48 hours. One lot of self-
fertilized eggs was kept in water and allowed to stand there 20
minutes. The eggs with the surrounding sperm were then put
Into ether-solutions of 2, i and 0.5 per cent, for 20 minutes, and
then returned to sea-water. None of these segmented. Another
lot of eggs from this individual were mixed with sperm of the
same individual and put into a 2 per cent, ether-solution for
ten minutes and then carried back to sea-water. Here 95 per
cent, of the eggs segmented. On the other hand some of these
same eggs taken from the ether after 5 minutes did not divide.
The following experiments were also carried out with other self-
fertilized eggs of the same individual.
Self -Fertilization Induced by Artificial Means. 145
Ether 15 min. 30 min.
2. 45 o (only 3 eggs)
I. 30 40
Experiment VIII. The eggs and the sperm of one individual
were mixed in ether-solutions of 4, 3, 2, i, 0.7, 0.5 per cent., and
removed after 10, 20, 30, 60 minutes to sea-water. It was
noticed that the spermatozoa were very sluggish in sea-water, and
although somewhat more active in the ether solutions, yet their
activity was not marked. Of the eggs, which appeared to be in
excellent condition, only three segmented, two in the i per cent.
(20 minutes) and one in the 0.5 per cent. (20 minutes.)
The eggs that had not segmented after the ether treatment
were fertilized, after they had stood 6 hours, with sperm from
another individual, and three quarters of an hour later nearly all
had divided normally into two cells. It was observed that the
spermatozoa of this second individual, used for cross-fertilization,
were also very inactive in their own fluid. They seemed to be
more active in the extract from the ovary of the first individual.
This ovary had also stood 6 hours in sea-water.
At the same time another experiment was made in which the
eggs of another individual, that had been isolated for 24 hours,
were put into ether solutions of 3, 2, i per cent, for 10, 20, and
60 minutes, and then returned to sea-water. None of these eggs
divided, except one in the 3 per cent, solution (60 minutes). It
was observed in this case that the sperm was inactive even in the
ether-solutions, but nevertheless this same sperm, not in ether,
cross-fertilized the eggs of the other individual in the preceding
series, and also, as stated, appeared to be somewhat active in
the extract of the ovary of the other individual.
Experiment IX. The eggs and sperm together were put into
ether-solutions of 5, 4, 0.7 per cent. The sperm were active
even in sea-water. The individual had been isolated for two days.
Ether.
5-
4-
0.7
:o mm.
20 mm.
30 mm.
12
10
5
(7 out of 8 eggs)
12
10
(9 out of ID eggs)
146 T. H. Morgan.
There was another series in this set in which the ether was
stronger (about one-half saturated). None of the eggs from this
solution segmented, but they became filled with clear spots. In
another check series, self-fertilized but kept in sea-water, none
of the eggs developed.
Experiment X. Sperm alone was put into ether solutions of
6, 4, I, 0.5 per cent. It was removed (along with some of the
surrounding fluid) and added to the eggs after 2 and 10 minutes.
Ether 2 min. lomin.
6. o o
4. o o
I. 20 10
0.5 4 50
It appears from this experiment that it suffices to put only the
sperm into the ether-solutions to bring about self-fertilization, but
it should not be overlooked that a certain amount of the ether is
carried over with the sperm when the latter is added to the eggs.
The amount, it is true, will be small, since the eggs stand in
water which further dilutes the ether, but so long as this source
of error is present, and it is very difficult to remove it entirely,
the result does not show conclusively that the ether acts on the
sperm alone, although I think this is the more probable inter-
pretation.
A check series of experiments was also made in which both
eggs and sperm were put into solutions of the same strength as
those given above, for 15 minutes and then removed to water.
Ether. 15 min.
6. o
4. o
I. 100 (but only ten eggs present.)
0.5 90
It is evident from both of the foregoing tables that only the
weak solutions were effective, and from the first table it appears
that this must have been the result of injury to the sperm. It
can easily be seen that the eggs also are killed in a few minutes
by a 6 per cent, solution of ether.
Self -Fertilization Induced by Artificial Means. 147
Experiment XI. Eggs from the oviduct were put into ether
solutions of I and 0.5 per cent, for one-half and three-quarters
of an hour; then washed in a small amount of fresh water and
fertilized with the sperm from the vas deferens of the same
individual. The experiment was carried out primarily in order
to see if the eggs were affected by the solutions, so that they
could be subsequently self-fertilized, but it is obvious that this
test is not a good one, since the eggs will carry with them, de-
spite the partial washing in water, some of the ether which may
then act on the sperm. The results were as follows :
Ether ^ hour ^ hour
i.o o o
0-5 4 5
Without a check series, which unfortunately was not made, it
is difficult to decide whether the small number of eggs that were
self-fertilized was due to the action of the ether on the eggs
or on the sperm. The experiment must be repeated on a more
elaborate scale.
Experiment XII. In this experiment with two individuals,
weaker solutions of ether were used. In one lot the sperm alone
was put into ether, and then added to the eggs. In the other
lot both eggs and sperm were put together into the ether. I
omitted recording the time in the ether, but it was probably about
five minutes.
Ether
Sperm and Eggs in Ether
Sperm only in Ether
1.0
0.5
2
0.25
Ether
Sperm and Eggs in Ether
Sperm only in Ether
1.0
20
0.5
50
0.25
85
4
This experiment shows that the sperm of the first individual
was incapable of self-fertilization, even with the ether present.
In the other individual, the sperm was good, and there was a
148 T. H. Morgan.
great deal of it; hence, no doubt, the excellent results in the first
column. What is especially significant is that the best results
were obtained when the eggs and the sperm were put at the same
time into the solution together. This may mean the ether has
some effect on the eggs as well as on the sperm, or that the most
effective period of activity for the sperm is immediately after
it comes into contact with the ether. My experiments do not
sufl^ce to settle this point, but that the spermatozoa are still
capable of cross-fertilizing, after they have been in the ether
for some time, is shown by the following result. After four hours
the eggs of the first individual were mixed with the eggs and the
sperm of the second individual. Later it was found that all
the unsegmented eggs had been fertilized. The ether had no
doubt largely evaporated.
The preceding twelve experiments with ether-solutions gave
definite results, although in a few cases the number of eggs self-
fertilized was small. It should be stated that there were ten
other individuals in which self-fertilization in ether did not take
place. This does not detract, I think, from the value of the
successful experiments, because, as has been shown, the sperm
is sometimes incapable of fertilizing even the eggs of another
individual. The following experiments were carried out in order
to examine this question further. It will be observed that parallel
experiments with ether were also performed.
In each series five individuals were used. The eggs of each
were fertilized with the sperm of every other individual. The
following scheme shows the order in which the eggs were crossed.
An individual having been opened, the eggs were removed from
its oviduct and distributed in five dishes, A-A. Another individual
was then opened (using, of course, different scissors, pipettes, etc.)
and Its eggs distributed to the next line of dishes, B-B. The same
method was followed for the other three individuals. The sperm,
a, of the first individual was then taken out and put into a small
amount of water. It was then distributed to one set of eggs from
each of the other Individuals, B, C, D, E; then the sperm of B
was taken out and applied to another set of eggs. The process
was repeated until all the eggs were supplied with sperm. The
Self -Fertilization Induced by Artificial Means. 149
sperm in each case is indicated in the table by the small letter
used as an exponent. The first set of A-eggs was as a rule fer-
tilized with the e-sperm and the last set of E-eggs with the a-sperm.
Experiment XIII. —
&100 EnOO E'^lOO EnOO E^' 85
D^OO DnOO DaOO D^OO D^lOO
O 98 0100 C 75 O 75 OlOO
B^ 99 BaOO B-^lOO B<^ 100 B'^ 100
A^(adtsptm) ^-'G'niil) A<^100 A"^ 85 A^ 99
In this experiment practically all of the eggs were fertilized
by the sperm of another individual. When fewer than the total
number segmented (fertilized), immature eggs may have been
present. As a check series A, B, C, D and E were self-fertilized.
None segmented, except in E, where two eggs out of the twenty
present, i. e., 10 per cent, divided.
The two ether series (self-fertilized) of these same eggs gave
the following results :
Ether
A
B
C
D
E
0.5
10
I
80
30
I.O
50
2
Experiment XIV. An experiment similar to the last was carried
out, with five other individuals, and gave the following results:
E^
E^
E^ 30
En 00
E^ 4
D^
D"
D^ c:.)
■P)e /no npe\
^-^ V eggs )
D^OO
C^
C
pd / no V
cnoo
0100
B^
B'^
B^
B'^ 50
B^OO
A«
A''
A.^
AnOO
A^OO
Despite a few slight discrepancies in this table, the main result
is clear. In only two of the five individuals was the sperm capable
of cross-fertilization, namely, the e-sperm and the d-sperm.
There was also a self-fertilized series of these eggs, and in this
not any of the eggs segmented. The ether series gave the fol-
lowing results :
T. H. Morgan
A
BCD
o
o o o
o
? 2
150
Ether A B C D E
0.5 00 o o 10 (only ten eggs.)
i.o o o ? 2 50
It becomes evident from this result that the frequent failure
of the sperm and eggs (mixed together) to self-fertilize in ether
is due to the poor quality of the sperm. The poor sperm does
not cross-fertilize, and presumably for the same cause it can not
always be made to self-fertilize even in the ether. That sperm
that is too poor to cross-fertilize may sometimes self-fertilize with
the help of ether I hold to be possible. I regret that I did not at-
tempt to determine whether poor sperm, that will not cross-fer-
tilize, can be made to do so by means of ether, but other experi-
ments lead me to think that it would often do so.
Experiment XV. In the following experiment all of the sperm
appears to have been good except that of A, whose eggs, how-
ever, were in excellent condition.
E^
2
E" 85
E^ 90
£<• (^;^
E^'
D=^
1
D^OO
D^ 20
D^ 90
D^lOO
C^
70
C" 100
C'^lOO
C'-lOO
C^ 90
B^
5
B^ 100
B^OO
BnOO
B^ 100
A^
A" 70
AaOO
AnOO
A^lOO
It is clear that the a-sperm was poor, although it did well
In C"" in which 70 per cent, of the eggs divided.
There was also a self-fertilized series in which none of the
eggs segmented, except 5 per cent. In B. (In C, 90 per cent, of
the eggs divided, but this may have been due to accidental con-
tamination.) In the ether series the following results were ob-
tained.
Ether
A
B
C
D
E
0.5
25
50
2
1.0
10
In this case although the spermatozoa of B, C, D, E were capa-
ble of crossing, they self-fertilized In ether very poorly, except In
C, where good results followed.
Experiment XVI. In this experiment again only the first in-
B
C
D
E
4
2
30
12 Utls)
10
30
Self -Fertilization Induced by Artificial Means. 151
dividual produced poor sperm, yet It did fairly well In one case,
and In ether gave some results.
E^i 5 ^. .^JE^IOOO^^E^ 80:;':^^ E" 100 E^ 70
D- OriH-'DnOO^r D'^ lO(Ss) D^lOO D^ 30
O; [^ 0100 0100 0100 C 100
B^ 25Gtj,) B^c-) B^G-) B^ (,- ) B^ ( -j
A^ 75 AMOO A<^100 A" 100 A^OO
In the self-fertilized series no eggs segmented. The ether
series gave the following results :
Ether A
0.5 2
1.0 50 (;Ss)
The experiments recorded in Exp. XIII to XVI show that
the sperm is at fault when cross-fertilization does not take place.
In fact, eggs in the oviduct seem always to be capable of cross-
fertilization. It is also evident that it Is more difficult to get
results with ether when the sperm does not cross-fertilize well,
than when it does act well In this way. From this It seems to me
very probable that when the ether fails to bring about self- fer-
tilization the fault lies with the sperm. We may perhaps even
go further and conclude that the action of the ether in bringing
about the self-fertilization Is on the sperm alone, but I am not
In position to prove positively that the action of the ether on the
eggs may not also enter Into the result.
In concluding my account of these experiments on Clona, I
should like to point out that I had constantly in mind the possi-
bility that the ether might produce parthenogenetic segmentation,
and that the sperm had in reality nothing to do with the result.
It was abundantly shown, however, that this was not the case,
and in the few experiments in which I put this view to the test,
by keeping eggs without sperm In ether - solutions of various
strengths, I got no results when the eggs were returned to water.
It should be noted in this connection that Lyon^ has recently
1 American Journal of Physiology, IX, July, 1903.
152 T. H. Morgan.
recorded that he was unable to cause artificial parthenogenesis In
Ciona intestlnalls at Naples by any of the ordinary means that
excite this development in other eggs.
I shall discuss later the view as to whether eggs may be entered
by the sperm of the same individual, but fail to develop unless
incited to do so by some external agent.
It has been pointed out in the preceding pages that the eggs of
Ciona may be fertilized after they have been in sea-water several
hours. I made a test of this again in the following experiment:
Experiment XVII. Some eggs were cross-fertilized at once,
others after 30, 80, 125 minutes, with fresh sperm from the same
Individual. All the eggs developed. A striking fact was ob-
served in this case. The eggs fertilized late began to segment
after a shorter interval than did those fertilized at once, so that at
the 32-cell stage those fertilized last were only one division behind
the first set, and no doubt soon caught up. It appears that a
ripening process goes on in the egg as it stands in the sea-water,
so that it begins to segment more quickly after it is fertilized
than does an egg fertilized as soon as removed from the oviducts.
It even appeared that after the first cleavage the rhythm of divi-
sion was quicker in the eggs whose fertilization had been de-
layed, but this point needs a special examination which I have not
yet made. The discovery is all the more significant because the
first polar spindle Is already formed in Ciona while the egg is in
the oviduct, and the spindle remains resting in the equatorial plate
stage until the egg is fertilized; hence the difference In time of
segmentation can not be accounted for by the time required for
the breaking down of the egg-nucleus and for the formation of
the polar spindle after the egg has been removed from the ani-
mal. Some change must take place In the sea-water, which, while
it does not cause the polar spindle to pursue Its development, yet
causes the developments that take place after the spermatozoon
enters to go on more rapidly.
EXPERIMENTS WITH CYNTHIA
The ovaries of Cynthia extend far forward, and have a very
short oviduct. Each ovary — there appear to be two in each In-
Self -Fertilization Induced by Artificial Means. 153
dividual — is double, the halves being united at the distal end.
Owing to the close proximity of the ovary to the surrounding
tubes of the testis, it is possible only by very careful manipulation
to get the eggs out of the cavity of the ovary without cutting into
the testicular tubes. When it was necessary to separate the eggs
from the sperm of the same individual, I have carried out this
operation, but in general the ovaries and the testes were cut up
together.
For the purpose of studying the effects of self-fertilization
Cynthia is in many respects inferior to Ciona because self-fer-
tilization takes place to a very large extent. On the other hand,
if check experiments are use'd for each individual, this factor can
be estimated, and the very fact that Cynthia does self-fertilize
its own eggs to such an extent gives an opportunity to examine
other aspects of the problem. A much more serious difficulty
is met with in that artificial cross-fertilization is often unsuccessful
in this species. Even when the eggs and sperm from a large
number of individuals are mixed together, fertilization may not
take place; but in curious contrast to this result are the following
observations on the egg-laying processes of this animal kept in
aquaria. On several occasions a number of individuals were put
together in the same dish. About 5 o'clock in the afternoon one
after another began to send out jets of eggs and of sperm pro-
ducing the effect of a lively cannonading. Under these circum-
stances it was found that every single egg was fertilized. Perhaps
only ripe individuals sent out their eggs and sperm, or perhaps the
eggs were mature in all individuals, and the sperm from one or
two individuals may have sufficed to fertilize all of the eggs. In
general it is, I think, the sperm of Cynthia that is not good.
Certainly the spermatozoa are often very sluggish when taken
from the testis and put into water. May it not be possible that
when the eggs are laid, Cynthia secretes some other fluid that
makes the sperm active? This point needs further investigation.
The best means that I found to determine the extent to which
self-fertilization of the eggs of Cynthia may take place was to
isolate some of the individuals early in the day, and observe in
those that emitted eggs and sperm in the late afternoon the per-
154 T. H. Morgan.
centage of eggs that segmented. The following four records
were obtained in this way: For August ii — 33, 10, 100, 95, 95,
75, 10 per cent. For August 16 — 30, 30, 10, i, 75, 90, 85 per
cent. For August 19 — 33, o, 10, 4, 4, o. For August 20 — 12, 4.
A much larger number of individuals gave off neither eggs
nor sperm, and some produced sperm and no eggs, and vice versa.
The results in the above list show all conditions from perfect
self-fertility to absolute self-sterility, although some of the latter
cases may have been due to no sperm being given off.
A few preliminary trials were made with two (A and B), and
with three (A, B, and C) individuals. The scheme of crossing
is given in the following diagrams:
For Two Individuals. For Three Individuals.
A^
B"
A^
B^
C<^
A^
B^
B^
A'
B=
C^^
C
A^
A few examples
of the results with two Individuals are as
follows :
A^
B" A^
B''
A" few
B^ 15 A" 10
B^IO
A^
B^ Outln.) A^
B^ 00
A" few
B^ 20 A" few
A^ B^OC^^)
A^ 50 B^
B^-few
Comparing the self-fertilized eggs with the crossed-eggs. It
Is clear that while self-fertilization did not take place in nine
cases, and In only one egg in the other case, yet cross-fertilization
more frequently occurred, but never so completely as when many
individuals normally deposited their eggs and sperm together.
In addition to these cases there were three others in which none
of the eggs, neither self- nor cross-fertilized, segmented. One
of the results with three individuals is given In the next table :
A^ B*^ 2 C^
B^ few A'' very few B"^ 25
75 Orare A^ 4
B'^
C^
A^ 50 CIT)
B^ 1
C" very few
A'^SO
Self -Fertilization Induced by Artificial Means. 155
In this experiment the a- and c-sperm did not self-fertilize, but
the former did well with C- and the latter with B-eggs. The
b-sperm self-fertilized to a slight extent, but did no better with
the A- and with the C-eggs.
In the next series the results are more striking:
A^
B^95
O50
Here none of the sperm self-fertilized the eggs. The a-sperm
did quite well with the B- and C-eggs (95 and 50 per cent) . The
b-sperm did well with the A-eggs, but not with the C-eggs. The
c-sperm did well with the A-eggs, but not with the B-eggs. It
may appear from the preceding table that there is something more
involved than simply the question of good sperm, for the same
sperm appears to act differently with different eggs.
Another experiment with three individuals gave no eggs self-
fertilized, but good cross-fertilizations with the c-sperm; less good
with the b-sperm. These experiments should be carried out on
a larger scale, and at different times of the year, but they suffice
to show that self-fertilization is very infrequent when the process
is an artificial one. It takes place to a considerable extent in some
cases when eggs are normally laid. Moreover the artificially
crossed eggs do not segment nearly so well in Cynthia as in Ciona.
The next experiment shows the action of ether on self- and
cross-fertilized eggs. Some of the eggs and sperm of one in-
dividual, A, were removed and put into sea-water. Other eggs,
A% were self-fertilized in an ether-solution, and a third lot, A^,
were crossed with sperm from B (A-sperm was also present).
The same process was carried out with B which was crossed with
sperm from A.
A BO
A^' few B'' few
A'' very few B^ very few
The results show that the self-fertilized eggs in ether did as
well as those that were crossed, but none of the eggs in water alone,
156 T. H. Morgan.
with their own sperm, segmented. Another similar experiment
with two other individuals gave the following results :
A BO
A^' B"
A" 50 B' r^'^^T)
In another set the ether appears to have been too strong, yet
50 per cent, of A'' divided.
In another experiment, 10 per cent, of the self-fertilized eggs
in ether segmented, and 50 per cent, of the crossed.
In another, 5 per cent, of the self-fertilized eggs in ether seg-
mented, and 75 of the crossed.
The next set is more instructive :
A BO
A^' 100 B*^
A^ 2 B^ 4
It is clear that the ether had a marked effect in A"*, making all
of the eggs self-fertilize. This is all the more interesting be-
cause none of the eggs without ether self-fertilized. Both eggs
and sperm of the B- set appear to have been in poor condition, so
that the sperm did not cross-fertilize, or the eggs become cross-
fertilized, to any extent.
In searching for other substances that might act on the sperma-
tozoa as does tht ether, I tried, amongst other things, a solution
of ammonia in sea-water, and this I found made the spermatozoa
even more active than the ether. Dilute solutions of alcohol
from I to 10 per cent, also excite the spermatozoa to greater ac-
tivity. Certain salt-solutions, ammonium chloride (i, ^, ^ per
cent.), magnesium chloride (2 per cent.), and sodium chloride (i
per cent.) appeared also to act on the sperm, but much less ef-
fectively than does ether, alcohol, or ammonia. In the alcohol
series of i, 3, 5, 6, 10 per cent., it was found that i per cent, made
the sperm very little more active; 3 per cent, more so; 5 per cent,
most active; 6 per cent, less; 8 per cent, no effect; 10 per cent., no
effect. The last two solutions undoubtedly injured the sperm.
In another series, 7 per cent, gave the best results.
Self-Fertilization Induced by Artifidal Means. 157
A few experiments were carried out in order to see if the sperm
made active by the alcohol, would self-fertilize the eggs when
it would not do so without the stimulus. Here, as in the preceding
series, the same lettering will be used in the tables. A^ self-
fertilized in sea-water, A^ self-fertilized in alcohol-solution, A''
crossed in alcohol-solution.
A^ B'^O
A^ [Alcohol] few B" [Alcohol] several
A" 20 B^ 50
In this experiment while no eggs were self-fertilized in sea-
water, a few or several (the percentages were not recorded) were
self-fertilized in alcohol, but even more developed in the crossed
lots.
In another experiment only one individual was used. The
eggs, self-fertilized in sea-water, did not segment, but 10 per cent,
did so in a 3 per cent, soluton of alcohol, and 50 per cent, in a 5
per cent, solution of alcohol.
Solutions of ammonia gave similar results. Sperm and eggs
were mixed together in very dilute solutions of ammonia. Many
eggs divided and of these most appeared, from their method of
division into several cells at once, to be polyspermic. Some of
the sperm from the last lot was added to eggs in sea-water.
Fewer eggs were fertilized, but several that were fertilized were
polyspermic. Eggs (not separated from their own sperm) were
crossed in ether. All of these were polyspermic. Another set
gave almost identical results.
It is clear from these experiments that those solutions that
make the spermatozoa more active often induce fertilization of
the eggs, when such a fertilization does not take place without
the use of the solutions. The activity of the sperm and the fer-
tilization of the egg appear to be directly connected. This point
will be more fully discussed later.
EXPERIMENTS WITH MOLGULA.
On each side of the body of Molgula there is an ovary sur-
rounded by a testis. It is very easy to open the central cavity
of the ovary, and remove the eggs without cutting the testis.
158 T.H.Morgan.
A few preliminary experiments showed that the sperm of Mol-
gula fertilizes the eggs of the same individual. The -following
illustrations will show the great powers of self-fertilization of
this species:
A^ 85
A"^ 90
B" 90
B^ 100
A^lOO
A" 100
B" 2
B^" 90 [irregular
A=^ 100
AMOO
B^ 100
B^ 100
A^
A" few
BMOO
B^
A^ 90
A" 100
B"
B^"
0^
100
These cases make it clear that the sperm is capable of fer-
tilizing the eggs of the same individual. Whether the sperm
of another individual is prepotent I did not attempt to determine.
There were only a few cases in which neither self- nor cross-
fertilization was effective, and whenever good crossing was accom-
plished self-fertilization was also realized, showing that when the
sperm is good, it will readily fertilize the eggs of the same
individual. Since similar results were obtained when three in-
dividuals were used it will not be necessary to give the latter
cases. The experiments were not extensive enough to show whether
good sperm affects the eggs of certain individuals better than it
does others, but Molgula is not well suited to test this point.
It occurred to me as possible that in Cynthia and in Molgula
the power to self-fertilize the eggs might be due to the eggs
coming from the ovary on one side of the body, and the sperm
from the other side. Conversely, if this were true, the lack of
self-fertilization in Ciona might be connected with the presence
of only one ovo-testis. I examined this possibility for Molgula.
The eggs from the small ovo-testis were fertilized with sperm
from the same side, and other eggs with the sperm from the other
side. In both cases all the eggs were fertilized. Conversely, the
eggs from the large ovary were fertilized with sperm from the
1 In B the sperm was probably bad. The Ab must therefore have been self-
fertilized. The same conditions hold also for the second couple.
Self -Fertilization Induced by Artificial Means. 159
same side, and others with sperm from the opposite side. Here
also all the eggs segmented. It is perfectly evident, therefore,
that the question of self-fertilization in Molgula is not connected
with the double condition of the ovo-testis.
EXPERIMENTS WITH OTHER FORMS.
In order to find out how generally ether, alcohol, and ammonia
excite to greater activity the movements of cilia, of flagella, and
of the spermatozoa of other animals, I made a few experiments
on certain protozoa and on the spermatozoa of the frog and of
the rat.
Ether, 5 per cent, stops the movements of paramoecium, and
kills stentor; 3 per cent, slows up the movements of the former,
and causes stentor to throw off its outer layer; the movements of
free swimming v^orticellae seemed to be increased; 2 and i per
cent, hasten the movements of paramoecium and of stentor.
Alcohol of 6 and of 8 per cent, slow down the movements of
paramoecium and stylonychia, and cause stentor to disintegrate;
10 per cent, kills; 4 per cent, appears to be near the limit, and
seems to increase their activity; 2 per cent, clearly increases their
activity.
Ammonia 1/200 per cent, kills paramoecium, stentor, and sty-
lonychia; and even 1/2000 also kills; 1/5000 per cent, seems to
make these protozoa somewhat more active, but I have not suf-
ficiently tested this solution.
Some of the same solutions were used with euglena, which
moves by means of an anteriorly directed flagellum. Ether 5
per cent, makes them somewhat more active; 3 per cent, less so,
and 2 per cent, gives no very noticeable effect. Alcohol 10 per
cent, kills; 8, 6, and 4 per cent, make them swim more actively;
2 and I per cent, give no definite result. Ammonia 1/200 kills;
1/2000 per cent, does not appear to make euglena more active,
but other strengths should be tried.
A male spotted frog [Rana halecina) was killed in No-
vember; its testes opened, and the immobile sperm squeezed
out into normal salt-solution. It was found that it took
some minutes to get a noticeable effect. Ether 5 and 2
i6o T. H. Morgan.
per cent, caused the spermatozoa to show some movement
in the course of 15 minutes. Alcohol gave better results. A
10 and a 6 per cent, solution awakened the spermatozoa to ac-
tivity; a 4 per cent, gave the best results of all. In no case, how-
ever, was the activity very great. No movements were detected
in ammonia-solutions, but only two strengths were used.
These scattering and incomplete observations show that these
substances are in all probability general stimulants for protoplas-
mic activity of certain kinds. *
I have also made a few experiments with the spermatozoa
of mice. The spermatozoa were taken directly from the testis
of a mouse that had just been killed. The solutions were added
to a drop of the sperm squeezed out from the testis into a drop
of physiological salt-solution, consequently the dilution is greater
than actually given by the percentage. In certain strengths of
ether (5 per cent.) and of alcohol (8 per cent.) it appeared that
the movement was increased; with ammonia I did not get satis-
factory results. The observations are made more uncertain here
because, when the testes are opened, spermatozoa in all stages
of development are found, and are consequently acted upon dif-
ferently by the solutions. It would be more satisfactory to use
a larger animal and take the spermatozoa from the vasa deferentia,
where they arc all fully formed. It is certain, however, that al-
cohol and ether do not produce as great effects on these sperma-
tozoa as they do on the spermatozoa of the ascidians and of
some other marine animals that I have examined.
In one of the preparations of the mouse testis the water began
to run out at one side and it became apparent at once that the
spermatozoa all turned and headed up-stream. It has been re-
corded by Kraft that spermatozoa swim in the opposite direction
to that in which the cilia of the oviducts act. My observation
suggests that movement in this direction is not due to the sper-
matozoa swimming against the direction of the greatest action
of the cilia, but against the stream that is produced by the cilia.
The movement may be a simple physical phenomenon — the lighter
tails of the spermatozoa being swept backwards by the current
so that the heads are turned up-stream, and the contraction of
the tail then causes the spermatozoon to travel in this direction.
Self -Fertilization Induced hy Artificial Means. i6i
In later experiments the sperm was taken from the vasa defer-
entia, and put first into distilled water where the spermatozoa
remained quiescent. If a drop of salt-solution (water loo, NaCl
0.75) was added to a drop of water containing the spermatozoa,
they became active in the course of a minute or less, and their
activity continued to increase for several minutes longer, when
they remained active for some time. If a drop of a 5 per cent,
ether solution is added to a drop of water containing quiescent
spermatozoa, no result is seen at first, but after ten minutes I
have observed a slight vibration of the spermatozoa. If now
after the ether has been added, a drop of the salt-solution is also
added,the spermatozoa become active, but it is difficult to determine
whether they become more active than when the salt-solution alone
is present. Certainly there is no marked difference. If a drop
of 8 per cent, alcohol is added to a drop of water containing the
spermatozoa no activity is observable, but if then a drop of salt-
solution is also added the spermatozoa begin to swim, showing
that the alcohol had not injured them, although it had failed to
arouse them to activity. Several strengths of KOH (3 per cent,
and weaker) were tried, but without effect; yet if salt-solution was
added later some slight activity was seen.
In another series of experiments the spermatozoa quiescent in
water were first made active by adding the salt solution. If ether
was then added no decided effect on the sperm could be seen
when their activity was compared with that of check preparations
of salt-solution only. It appeared sometimes as though the ether
did make the activity more pronounced, and the movement of the
spermatozoa appeared somewhat different in the two cases. In
the ether the motion was more jerky, and in the salt solution more
sinuous and normal.
The following. solutions were also tried: The sperm was first
put into a drop of water, and then a drop of the solution was
added, NaHCO^ 0.625 per cent, caused the sperm to vibrate
rather actively; NasCOg, 5.0 per cent, caused a little activity after
five minutes; KCl, 0.5 per cent, caused greater activity than
did the sodium carbonate, while CaCl caused somewhat less vibra-
tion.
1 62 T. H. Morgan.
These, and some other experiments that need not be described
here, show that salt-solutions of various kinds have a marked
effect in arousing to activity the inactive spermatozoa of the vasa
deferentia. They also make active, spermatozoa that are quies-
cent in distilled water. On the other hand ether, alcohol, and
ammonia, which proved so efficient for the spermatozoa of the sea-
urchin and starfish, appear to have little effect on the spermatozoa
of the mouse.
The more fundamental physiological question as to the nature
of the action of these different substances I shall not attempt to
discuss without a further basis of observation and experiment to
go upon. Enough has been seen, however, to suggest that the
substances act as a "stimulus," which is perhaps not dissimilar in
kind from that which causes some eggs to begin to develop, or
a nerve impulse to start, or a muscle to contract. Here also we
may urge, as I have urged elsewhere^ in opposition to Loeb's
conclusion in regard to the action of certain agents in causing
artificial parthenogenesis, that the nature of the stimulus is of
such a kind that the result depends much more on the structure
or the composition of the living thing than upon the kind of
stimulus employed. So unstable is the living organization that
the sHghtest change brought about in it by chemical or by physical
means suffices to set into action a perfectly definite and pre-
arranged series of events.
HISTORICAL REVIEW.
The action of ether, ammonia and alcohol on the speramtozoa
of Ciona, arousing them to greater activity and thus, under certain
conditions, bringing about the fertilization of the egg, raises the
question as to whether in the higher animals a similar action may
not result from the application of these and of other substances,
and also whether the secretions of some of the glands connected
with the reproductive system may not have a similar effect on the
spermatozoa.
When I tried to find some substances that might bring about
self-fertilization in Ciona I was not aware that there had already
1 Science. N. S. XI. 1900. Pp. 178-180.
Self -Fertilization Induced by Artificial Means, 163
been made several experiments on the action of solutions on the
spermatozoa of other animals. I find that there are quite a num-
ber of observations of this sort, although none of the observers
have had in view the same question with which I was especially
concerned.
Kolliker in 1856 carried out an extensive series of experiments
on the effect of different solutions on the spermatozoa of the
bull, dog, rabbit, horse, and also made a few observations on the
spermatozoa from a human cadaver. He found that water alone
quickly brings spermatozoa to rest, but does not kill them. They
can be aroused to activity by adding, for instance, a 10 per cent,
solution of disodium phosphate.^ Many other substances were
found favorable to the activity of the spermatozoa, such as blood-
serum, sugar in certain strengths, sodium chloride, caustic pot-
ash, etc.
The caustic alkalies (potassium, sodium, and ammonium hy-
droxide) were found to be especially powerful excitants. Kol-
liker also tried a number of other solutions, such as three different
kinds of sugar, glycerine, gum, etc., which in certain strengths
cause increased activity; also urea, gall, morphine, strychnine
(nitricum), which have an indifferent effect. He also tried al-
cohol, creosote, chloroform, ether, alkaloids, and tannin, which
have an injurious effect. Kolliker also examined the action of the
secretions of the glands of the male reproductive organs — the
uterus masculinus, prostate and Cowper's glands. He found
that these secretions excite the spermatozoa to greater activity.^
The much more recent experiments of Steinach bear even more
directly on the present problem. He found that after removal of
the glandulae vesiculares ("receptaculum seminalis" of some writ-
ers) of the male rat, that, although the sexual instinct remained,
the number of young that were born was much decreased. When
1 Kolliker gives the formula 2Na0H0P0g, which is no doubt disodium
phosphate, now written Na HPO .
2 Moleschott and Richetti are quoted by Kolliker as recognizing the favorable
action of sodium salts on the spermatozoa. Quatrefages found that the sper-
matozoa of the weasel showed a "surexcitation" in 64 parts sea water to one
part sea salt. Newport found that potassium carbonate, and also 1/480 of
potassium salt made the activity of the spermatozoa of the frog greater.
164 T. H. Morgan.
this gland, as well as the prostate, was removed no young at all
were born, although frequent union with the females took place.
The results may be due to the semen being insufficiently diluted
when it is not mixed with the secretions of the glands, or else to
the absence of proper excitation of the spermatozoa when the
gland-secretion is removed. That the spermatozoa may be nor-
mally acted upon by the secretion of the glands was shown by
Steinach in the following way: Sperm from the vas deferens was
mixed with a physiological salt-solution. A drop was placed
under a cover slip and the edges sealed to prevent evaporation.
The preparation was kept at a temperature of 35° to 37° C. A
similar preparation was made with the secretion of the prostate.
In the former the spermatozoa began to lose their activity in one
and a half hours, and after three hours had come completely
to rest. In the other preparation, that containing the extract
from the prostate gland, the spermatozoa were active after 1 1
hours, and ceased to move altogether only after 22 hours. This
experiment shows that the secretion of the gland prolongs greatly
the period of activity of the spermatozoa. Whether it excites
them to greater activity is not stated, but Kolliker's results leave
no doubt on this score. The decrease in the fertilizing power
when the glands were removed may well be connected, as sug-
gested above, with the lessened activity of the spermatozoa.
Duller has recently studied the question as^ to whether the
spermatozoa of the sea urchin are attracted to the egg, — in other
words, whether, as some authors have assumed off-hand to be the
case, there is a chemotactic action of the egg on the spermatozoa.
He points out that although Strasburger claimed that the egg
of Fucus excretes a substance that attracts the spermatozoon from
a distance of two diameters of the egg, Bordet and Buller himself
have failed to confirm this statement. Massart thinks that In
the case of the frog the meeting of the spermatozoon and the egg
Is purely accidental. Buller finds for the sea-urchins, Arbacia,
Echinus, and others, that when the spermatozoa are set free near
the egg they show no tendency to swim towards it. The dense
collection of spermatozoa that forms around the egg is due to
those that happened to run into the jelly sticking there. These
Self-Fertilization Induced by Artificial Means. 165
spermatozoa then proceed to bore into the jelly; most of them
in a radial direction, although a few can be seen to go in obliquely,
or tangentially. The same phenomenon occurs in unripe eggs,
and in eggs that have been killed in weak osmic acid and the
acid washed out. It is improbable, therefore, that chemotaxis
has anything to do with the result.
In order to see if any substance is given off by the eggs that
attracts the spermatozoon, the eggs were taken from the ovary,
carefully washed, and allowed to stand for 2 to 12 hours in a
small amount of sea-water. Capillary tubes were then filled with
this water and placed in a drop containing spermatozoa. The
spermatozoa did not show any tendency to collect around the
openings of the tubes. Several other substances were tried in
the tubes in the same way, — salts, sugar, ferments, acids, alcohol,
etc. — but no chemotaxis was discovered.
The spermatozoa of the sea-urchins swim in spirals. Coming
into contact with a surface, the spiral is changed to a circular
movement due to contact. Buller considers whether the radial
path taken by most of the spermatozoa after they have entered
the jelly is due to stereotropism. He reaches the conclusion that
while theoretically this assumption will explain the phenomenon,
yet conclusive evidence in favour of this view is lacking. He
suggests that it may be possible to find a purely physical solution
of the problem.
Von Dungern has examined the question of cross-fertilization
from the point of view of the different substances contained in the
egg, and has reached some conclusions of the greatest interest.
He finds that the egg of the starfish, Jsterias glacialis, contains a
substance that acts as a poison on the sperm of the sea-urchin
(Echinus or Sphaerechinus) . The minimal lethal dose for the
sperm mixed with 2 ccm of sea-water varies considerably with the
individual; for Echinus between 1/800 to 1/6400 part is fatal
in half an hour. Von Dungern tried to obtain an antitoxin from
the blood of the rabbit that would neutralize the effect of the
poison of the eggs, hoping that it might be possible in this way
to bring about the cross-fertilization of the egg of the starfish
by the spermatozoon of the sea-urchin. He found, however, that
1 66 T. H. Morgan.
the serum of the normal rabbit already contains a substance that
has a powerful antitoxic action on the poison of the starfish, so
that it was not necessary to obtain an antitoxin by injecting the
poison into the rabbit. The antitoxin of the rabbit's serum was
added to water containing the eggs of Asterias, and then sperm
from a sea-urchin was supplied. Von Dungern often obtained
two- and four-cell stages in this way, but the results were uncertain,
and he could not decide whether fertilization had or had not
taken place. It seems not improbable, I think, that the outcome
may have been due to artificial parthenogenesis which occurs
very readily in the eggs of certain starfish; in fact, it is very
difficult to prevent its occurrence, unless the eggs are very care-
fully handled.
The same poison that is present in the eggs of the starfish is
also secreted by the skin. It is also rendered harmless by the
rabbit's serum. In the sea-urchin there is a poisonous substance
in the gemmiform pedicellariae, which is very injurious to the
sperm of the starfish. If lOO of the pedicellariae of Sphaerechi-
nus are rubbed up in one ccm of sea-water, the solution will de-
stroy in a quarter of an hour the sperm contained in ten to tw^enty
litres of sea-water. The minimal lethal dose for 2 ccm is 1/5 120
to 1/ 1 6240 ccm. The spermatozoa of Sphaerechinus itself are
killed by this fluid, but a much stronger dose is necessary. On
the other hand an extract of the egg of Echinus, Sphaerechinus,
Strongylocentrotus, or Arbacia does not kill the spermatozoa of
the starfish even in the strongest solutions. What then prevents
the spermatozoa of the starfish from entering the eggs of these
sea-urchins? There Is another factor. Von Dungern thinks, that
interferes with this combination. The egg membrane of these
urchins has an agglutinizing effect on the spermatozoa of the
starfish. This agglutinizing effect appears to be the same phe-
nomenon as that seen "whenever cells of any kind are introduced
into the body of another animal." So far as this process is in-
volved in the union of germ-cells. Von Dungern thinks that under
certain conditions it might assist the fusion, while under others
it might interfere with it. Thus two naked and equivalent cells
might be helped to unite, while an egg surrounded by an agglu-
Self -Fertilization Induced by Artificial Means. 167
tinizing jelly would fail to be fertilized. The substance in the
sea-urchin's egg that agglutinizes the starfish sperm can be ren-
dered ineffective by the rabbit's serum. Not all starfish sperma-
tozoa are agglutinized by the jelly or by the egg-substance of all
the different sea-urchins. In Sphaerechinus it fails to occur.
Therefore in this case the failure to cross-fertilize must be due
to some other factor, and, in fact. Von Dungern claims to have
found still another substance in the sea-urchin's egg that excites
to greater activity the immature and quiescent spermatozoa of
the starfish. These immature sperm, made active by this sub-
stance, are then capable of fertilizing the eggs of the starfish.
He found that weak doses of chloral hydrate and of cocaine also
make these quiescent spermatozoa active, and that rabbit's serum
has a marked effect. Von Dungern believes further that these ex-
citing substances may actually prevent, in certain cases, the cross-
fertilization, because they may change the kind of reaction shown
by the sperm. He observed that the spermatozoa of those
species that do not normally show rotational movements when
they come in contact with surfaces, usually do so when the excit-
ing substances just mentioned are present. It does not appear
to me, however, that this is an altogether satisfactory explanation
of the failure of cross-fertilization in these cases.
Von Dungern also examined the question as to whether the
egg secretes a substance that favours fertilization by its own
sperm. He believes that he has also discovered such a substance.
The eggs of Echinus (or of Sphaerechinus) are rubbed up and
mixed with pieces of jelly that have been carefully washed. When
sperm is added to the water in which such pieces lie they stand
vertically to the surfaces of the pieces. If on the other hand the
pieces of jelly are not mixed with the substance from the egg, the
spermatozoa simply rotate on the surface of the jelly, and do not
stand vertically. Starfish spermatozoa with Arbacia jelly be-
have as with simple jelly alone, i. e.^ they do not stand vertically.
The vertical position of the spermatozoa is due. Von Dungern
thinks, to the presence of some substance in the extract that lowers
the excitability of the spermatozoon to contact, and hence it takes
a vertical position. He also points out that this same substance
1 68 T. H. Morgan.
causes the spermatozoa to lose their power of movement in a
short time. Thus, while Von Dungern finds no evidence of a
substance in the egg that attracts the sperm, he believes that there
may be present in some eggs a substance that favours the fer-
tilization of the egg, by causing the spermatozoon to assume that
position in the jelly that is most likely to bring them to the surface
of the egg.
Loew has attempted to show by an experiment, which is not, I
think, well suited to prove his point, that the spermatozoa of the
rat are attracted to, i. e., that they are positively chemotactic to,
the slime layer of the uterus and also to the alkaline mucosa of
the digestive tract, but not to the acid slime of the vagina. His
method of experimenting was as follows : A piece of the mucosa
of the uterus was put on one side of a slide and a piece of the
vagina on the other. A drop containing the sperm was placed in
the middle of a cover-slip, and this put over the pieces on the
slide. It was found that the sperm collected more on the side
near the piece of the uterus, and from this Loew infers that they
have been attracted to this side. In the light of the other experi-
ments described above it will be clear, I think, that the greater
accumulation of the sperm on one side by no means establishes the
conclusion that they have been attracted to this side. Loew tried
to show that filter paper saturated with alkaline substances acts che-
motactically on the spermatozoa, in the sense that they move to-
wards such substances, but, as in the preceding case, it does not nec-
essarily follow because spermatozoa collect around or in certain
substances, therefore they must have moved towards these sub-
stances. The recent work of Jennings on the protozoans shows that
their accumulations in certain areas is not due to the action of sub-
stances that cause the individuals to swim towards those sub-
stances, but on the contrary to their action being such that those
individuals that enter areas containing these substances are unable
to leave them. The result is the same as when the spermatozoa
touch the jelly of the egg and stick to it, although the means by
which the accumulations are formed in the two cases are entirely
different. It would be interesting to see if spermatozoa may
not behave towards certain solutions as do the protozoans.
Self -Fertilization Induced by Artificial Means. 169
THEORETICAL.
It has been often assumed by embryologists that there exists
some sort of attraction between the eggs and the spermatozoa of
the same species. This idea would readily suggest itself to anyone
who saw spermatozoa collecting in crowds around the eggs, but
it by no means follows that this phenomenon is really due to an
attracting substance emanating from the egg. The result may
be due to the membrane of the egg, to which those sperma-
tozoa stick that come accidentally into contact with it. In
fact I have observed similar collections of spermatozoa in the
ascidian around pieces of the body tissue, where the result had
every appearance of being due to some sticky substance, exuding
from the piece, rather than to an attraction exerted by the piece
on the spermatozoa.
Pfeffer's oft-qiioted experiment M'ith tlie antherozooids of ferns,
liverworts, etc., appears to support the idea that the antherozo-
oids are attracted to the malic acid that is present in the neck of
the archegonia, but in the light of the recent experiments of Jen-
nings and others, as to the way in which unicellular forms accu-
mulate in a drop of acid, we can readily see that the results may
have a very different interpretation from that usually given to
them. Confining our discussion to the results obtained with the
ascidians, I offer the following tentative analysis of the problem :
The failure of the spermatozoon of Ciona to enter the egg of
the same individual may be conceived as due to some physical ob-
stacle. It Is conceivable that pores may exist In the egg-membrane,
or even In the surface of the egg itself. This is the argument
that Ptliiger^ used in the case of cross-fertilization of the frog's
egg. If in the ascidian there existed a correlation of such a sort,
that the size of a spermatozoon of a given individual Is always
greater than the pores of the eggs of the same individual, then
self-fertilization could not take place. That this Is not the real
explanation Is shown by the fact that good spermatozoa are ap-
parently capable of fertilizing the eggs of all other individuals.
This would certainly not be the case if the exclusion of the sperma-
1 Archiv. f. die gesammte Physiologic, XXIX., 1882.
lyo T. H. Morgan.
tozoon from the egg of the same individual was due to the size
of the pores, because there would be eggs of some other individuals
having pores as small or smaller. Another possibility that sug-
gests itself is that the surface tension of the egg is of such a sort
that it excludes the spermatozoa of the same individual, but this
idea does not appear to give a satisfactory solution, for, aside
from the fact that it is difficult to imagine how such a relation
could exist, there would also occur cases in which the surface ten-
sion of the eggs of other individuals would exclude certain sperm,
and this does not appear to be the case. It is true that the ad-
dition of the ether to the water may cause a difference in the
surface tension of the egg, and it might be made to appear that
this was the way in which the self-fertilization is effected in the
ether-solutions, but I can not believe that this is the explanation
of the results, because other experiments show that a considerable
amount of ether is necessary to cause self-fertilization.
It seemed to me that violent shaking might so affect the sur-
face of the egg that self-fertilization might take place. A
number of eggs from the oviduct were violently shaken for a
few minutes in a small vial, and then sperm from the same in-
dividual was added. No segmentation took place, and the pre-
sumption is therefore that the eggs were not fertilized.
Turning to the chemical side we find a number of possibilities
that demand consideration. The inactivity of the immature sper-
matozoa, and the lack of power of such sperm to fertilize the egg,
their becoming active in certain solutions, and their power then
to fertilize eggs that they did not fertilize before, as best shown
in Cynthia, suggests that normally the eggs may secrete certain
substances that make more active the spermatozoa, which then be-
come capable of fertilizing the eggs. This view appears all the
more attractive in the present case on account of the observed leth-
argy of the spermatozoa of these ascidians, and the apparent con-
nection in such cases between this condition and the impotence of
such sperm in fertilization. Yet after careful consideration I am
not prepared to advocate this view as the only solution, although I
realize that it might be made to give the appearance of a ready
explanation of my results. Not that this induced activity may not
Self-Ferttlization Induced by Artificial Means. 171
be one of the factors to be taken into consideration, but it is not,
I think, the whole explanation. My reasons for regarding this
view as insufficient are the following: It was found that sperm
that appeared to be very little active was sometimes capable of
cross-fertilizing the eggs of another individual. Possibly this
may be due to somewhat greater activity induced by something
secreted by the eggs of the other individual, yet on the whole I
can not claim that direct obsei-vation gave any convincing evi-
dence in favour of this assumption. More significant are the
results of the experiment of mixing eggs from two Individuals,
and subsequently fertilizing them with the sperm from one of.
the Individuals. Half only of the eggs segmented, presumably
those cross-fertilized. If some substance that makes the sperm
active were really thrown out by the eggs, then we should expect
that all the eggs would have been fertilized, unless indeed the se-
cretion loses its power a short distance from the surface of the
egg that secretes it; but this does not seem to be a probable Inter-
pretation.
A different point of view Is that the egg secretes some sub-
stance that attracts the spermatozoa. On this view we must
suppose that the substance secreted by the egg of Ciona has no
attraction for the spermatozoa of the same individual.
The little evidence that I have to' offer, based on experiments
with ascidians. Is not favorable to this idea, that the cross-fertiliza-
tion is due to some attractive substance secreted by the egg. In
the species that I have examined there is no such marked
accumulation of spermatozoa around the eggs as is seen in many
other animals, and nothing in the behaviour of the cross- and self-
fertilized egg to suggest that the difference in the results is due
to an attraction in the one case, and to the absence of an attraction
In the other. In other forms where there is a better opportunity
for examining this question the most recent observations go to
show, as has been pointed out in detail above, that there Is no suf-
ficient evidence for the view that the egg attracts the spermato-
zoon.
Conversely, it may be supposed that the egg secretes some sub-
stance that repels the spermatozoa of the same individual. I
172 T. H. Morgan.
observed nothing that would support such a conclusion, and this
interpretation of the process would be foreign to what we find
in general in connection with fertilization even in cases where the
sperm of one species does not fertilize the eggs of another.
We come now to a more subtile argument, and one that we
are scarcely in position to discuss profitably in our present state
of ignorance concerning the union of egg and spermatozoon.
It may be assumed that there is some sort of "chemical affinity"
between the egg and the spermatozoon that causes the two to
unite when they come together. On this assumption we should
have to suppose in Ciona that this affinity does not exist, or at
least is less strong, between the egg and the spermatozoa of the
same individual than between those of different individuals. Such
a statement carries us no further, however, than the facts, and in
the case of Cynthia we should have to assume that the affinity is
so nicely balanced that sometimes the spermatozoon can unite, and
sometimes it can not. In the case of Molgula the affinity must
be assumed to suffice to bring about self-fertilization. Until we
can give some more tangible form to this idea it does not appear
to have any greater value, than the mere statement of the facts,
and indeed may have less value, since it may give a wrong im-
pression as to the real factors at work.
Finally there might be advanced what may be called the electro-
chemical hypothesis. The union of the egg and the spermatozoon
may be supposed to be an electrical phenomenon, connected with
a difference in the chemical composition of the two elements. The
sperm head is almost pure nuclear chromatin, while the surface
of the egg is protoplasmic. Possibly the spermatozoon and the
egg have different electrical charges and unite with each other
if brought near enough for the charges to become effective. But
on this supposition it is not clear why the eggs and the sperm of
the same individual would not unite. Here also we get no light
on the absence of self-fertilization in Ciona.
I have kept constantly in mind while at work on this problem
the possibility that the spermatozoon may really enter the egg, but
fail to develop there, or fail to start the development of the egg,
because, coming from the same individual, it was not sufficiently
Self -Fertilization Induced by Artificial Means. 173
different In composition to supply the necessary stimulus. The
ether might be supposed to make the sperm sufficiently different
from the egg to start the cleavage, or the ether might itself supply
the stimulus which is capable of starting the development of the
egg after the spermatozoon has entered.
The test of this view should be found in direct . observation
of the eggs themselves. I prepared therefore a series of eggs
of Ciona, some unfertilized for check series, others self-fertilized,
but not put into ether, and others like the last, but put into ether.
The difficulties of determining whether the spermatozoa can
enter the eggs of the same individual, but fail to start the devel-
opment, are greater than may appear at. first sight. The sperm
head Is so minute that If after it entered no changes were af-
fected In the protoplasm about it, Its presence might be readily
overlooked, and since the spermatozoon of Ciona^ enters the egg
in a granular zone that colors more deeply in certain stains than
does the rest of the egg, the difficulty is thereby increased. Of
course I have been on my guard against cases where the sur-
rounding sperm have floated over the section, as sometimes hap-
pens, or have been carried over it by a defect In the knife, and I
have also been careful to exclude all cases where specks of foreign
matter may have been on the slide, or in the fixative. There are
also two further precautions to be taken. When the egg with-
draws from the membrane and the test-cells are extruded, as it
were, from the outer zone of the egg, the protoplasm is some-
times drawn out in mamiliform processes that stain deeply and
resemble the entrance cone formed by the spermatozoon pene-
trating certain eggs. Even when the protoplasm does not pro-
trude, deeply staining spots are generally present and are espe-
cially obvious after Iron haematoxylin. Careful staining with
Delafield's haematoxylin shows clearly that these spots have noth-
ing to do with the entrance of spermatozoa. Furthermore these
spots are found in unfertilized eggs. After iron haematoxylin
minute deeply staining bodies, flattened against the outer surface
of the egg, can generally be found, and these strongly suggest
spermatozoa. That they are not such is shown by their presence
in unfertilized eggs, and by their absence after the Delafield
174 T. H. Morgan.
stain. I mention these points because they might easily lead one
after only a casual examination to conclude that spermatozoa
enter the eggs. My best results have been obtained by drawing
out the iron haematoxylin until the protoplasm has lost all of its
color, or better still by using the Delafield stain, and also thor-
oughly extracting the color from the protoplasm.
Although I have examined a large number of preparations I
have not seen a single definite case without ether in which a
spermatozoon has entered the egg of the same individual. Diffi-
cult as it admittedly is to be absolutely certain on this point, yet if
the spermatozoa had entered and had begun to enlarge I feel
certain that I should have detected their presence. That un-
developed sperm-heads may be present I must admit as a pos-
sibility, but I have not detected them, and believe that I should
have been able to do so were they present. It is also a point of
some importance that I have not found any spermatozoa within
the egg membrane, although quantities of them may lie outside.
There is a further point in this connection, the importance of
which I did not appreciate until I had closed the experimental
part of my work. In the eggs of many animals a change takes
place In the egg, after the penetration of one spermatozoon, of
such a sort that the entrance of more spermatozoa is prevented,
I have found In Ciona that, after the sperm has stood with the
eggs of the same individual and has failed to fertilize them, these
eggs could still be readily fertilized by spermatozoa from another
individual. If a spermatozoon of the same individual really
enters the egg it does not In consequence bring about such a change
in the egg that other spermatozoa can not enter, and therefore
many spermatozoa of the same Individual from which the eggs
were taken should be expected to gain entrance, but I am quite
certain that this, at least, does not occur. From this consideration
also It may be Inferred that the spermatozoa do not normally pene-
trate the eggs of the same individual.
In the light of these observations it seems probable that whenever
a spermatozoon enters the egg, the egg begins to develop regard-
less of whether the spermatozoon comes from the same or from
another Individual. The ether must therefore induce a change
Self -Fertilization Induced by Artificial Means. 175
of some sort that directly effects the entrance of the spermatozoon
into the egg, and at present I see no other interpretation that is
left than that this entrance is due to the greater activity of the
spermatozoon that causes it to overcome some resistance, either
on the surface of the egg itself, or in the membrane surrounding
it. The nature of this resistance I did not detect, and this must
be the next step in the analysis. One method by which this view
may be tested is obvious, and has already been referred to. The
spermatozoa made active by sea-water must be placed in an ex-
tract of the eggs (or body-tissues) of the same individual, and
then, after a time, the eggs of another individual added. On
the hypothesis these eggs should be less likely to become fertihzed
than eggs placed directly in contact with the fresh sperm.
It has been found that certain substances secreted by the glands
of the reproductive organs of the male mammal arouse the sper-
matozoa to greater activity. It has also been found that many
other substances have a similar effect on spermatozoa. It would
be equally interesting to discover if the secretions of other parts
of the genital ducts of the male or of the receptacula of the fe-
male, when such are present, may not bring the spermatozoa
to rest, or keep them quiescent until some other exciting agent
arouses them. It seems almost certain that this must be the case
in those animals in which the spermatozoa of the male are stored
up in receptacula of the female, as for instance in the honey bee,
or in such a hermaphroditic animal as the earthworm. The
length of life of the spermatozoa in some of these forms would
seem to make some assumption of this sort necessary. Experi-
ments can easily be made that would decide this question. Kol-
liker has shown, in fact, that water quiets the spermatozoa of
mammals without killing them.
In the ascidians it is probable that the spermatozoa in the vas
deferens are quiescent. It is significant that in these hermaphro-
ditic forms the oviduct in which the eggs are stored takes a course
parallel to the male duct. Possibly the proximity of the two
ducts may be connected with the lack of power of self-fertilization
of the eggs, because the egg may be saturated with the same
substances that keep the sperm quiescent. It may be, however,
176 T. H. Morgan.
that this relation is more fundamental, and the particular substance
is one peculiar to the whole body. That the reaction must be
something quite specific is shown by the fact that the spermatozoa
are able to enter eggs of any other individual.
It appears probable that of all the different substances that
excite the spermatozoa to activity the secretions of the glands
connected with the male reproductive organs may be the most
efficient. From a statement of KoUiker's it seems not improbable
that the substance secreted in the glands of one species may be
also efficient for the spermatozoa of other species. Whether
by the use of the substances from the glands of another mammal
it might not be possible to excite human spermatozoa to greater
activity and thus assist materially in bringing about fertilization
in cases where the impotence is on the side of the male remains
to be examined. There is here a question that may have an im-
portant practical aspect.
The lack of power to self-fertilize in plants may also be due
to the inability of the pollen tube to penetrate sufficiently far into
the stigma and style. It appears that penetration does actually
begin in some cases that have been observed, but possibly the
growth may be arrested further down in the style. The pre-
potency of other pollen would then find its explanation in the
more rapid growth of this foreign pollen. Here again is an op-
portunity for future work.^
In attempting to formulate a theory to account for the deter-
mination of sex, Castle assumes that there are two kinds of sperma-
tozoa, male and female, and that there are also two kinds of eggs,
male and female. He also assumes that a female egg can be
fertilized only by a male spermatozoon and that a male egg only
by a female spermatozoon. I have already pointed out elsewhere^
that my results do not support this assumption. Castle appealed
to the case of Ciona as one in favour of his contention, for the
eggs here can not be fertilized by the sperm of the same individual.
It is not explicitly stated to the contrary, and the reader might be
led to infer from the context that in Ciona all the eggs and all
1 The experimmts of Myoshi should be especially considered.
2 Popular Science Monthly. Dec, 1903.
Self -Fertilization Induced by Artificial Means. 177
the spermatozoa of one individual must be supposed to be male,
and in another individual the reverse; but certainly this is not the
case, and could not have been Castle's meaning, for if it were so
then half of the individuals would be infertile with the sperm of
the other half, and this is not so. I have pointed out that
my results with ether, etc., do not support Castle's assumption,
although it might, of course, be claimed that the ether causes
the spermatozoa to lose, as it were, their homosexual repugnance.
However this may be, I have found that no such lack of power
to self-fertilize is found in some other ascidians, as in Molgula
for example. If my supposition is correct, that self-fertilization
in Ciona is due to the presence in the eggs of some substance that
brings the spermatozoa to rest, the whole question assumes a very
different aspect and does not appear to have any connection with
the question of the determination of sex.
LIST OF REFERENCES.
BoRDET. — Contribution a I'Etiide de I'lrritabilite des Spermatozoides chez les
Fuccacees. Bull, de I'Acad. Belgique 3W(?. ser. XXVII. 1894.
BuLLER, A. H. — The Fertilization Process in Echinoidea. Report 70. Meeting
Brit. Assoc, ipoi.
Is Chemotaxis a Factor in the Fertilization of the Eggs of Animals.
Quart. Journ. Micro. Sc, XLVI. 1903.
BuLLER, A. H. R. — Contributions to our Knowledge of the Physiology of the
Spermatozoa of Ferns. Ann. of Botany, XIV. 1900.
Castle, W. E.— The Heredity of Sex. Bull. Mus. Comp. Zool.. XL. 1903.
Dewitz, J. — tjber Gesetzmassigkeit in der Ortsveranderimg der Spermatozoen
und in der Vereinigung derselben mit dem Ei. Archiv. f. die gesammte
Physiologic. XXXVIII. 1886.
V. DuNGERN, E. — Die Ursachen der specifitat bei der Befruchtung. Centralbl. f.
Phys. XIIL 1901.
Neue Versuche zur Physiologic der Befruchtung. Zeitschr. f. allgem.
Physiologic. I. 1902.
IvANOFF. — Journal de Physiologic et de Pathologic general. II. 1900.
KoLLiKER, A. — Physiologische Studien ucbcr die Samcnflussigkeit. Zeitschr. f.
wiss. Zool. VII. 1856.
Kraft, H. — Zur Physiologic des Flimmerepithels bei Wirbelthieren. Archiv. f. d.
gesammte Physiologic. XLVII. 1890.
LiDFORSS. — iJber den Chemotropismus der PoUenschlauche. Ber. d. Deutsch.
Bot. Gcsell. XVII. 1895.
178 T. H. Morgan.
LoEW, O. — Die Chemotaxis der Spermatozoen im weiblichen Genitaltract. Sitz-
ungsber. d. Wiener Akad. ; Math.-naturw. CI. CXI. 1903.
Massart, J. — Sur rirritabilite des Spermatozo'ides de la Grenouille. Bull, de
I'Acad.roy. de Belgique. 2'me Sir. XV. 1888.
Sur la Penetration des Spermatozoides dans I'Oeuf de la Grenouille.
Bull, de I'Acad. roy. de Belgique. Zme. Ser. XVIII. 1889.
MiYOSHi, C. — tjber Reizbewegungen der Pollenschlauche. Flora. LXXVIII. 1894.
MoLiscH, H. — tJber die Ursachen der Wachtumsrichtunger bei Pollenschlau-
chen. Sitzungsber. d. k. Acad. d. Wiss. in Wien. 1889, 1893.
Morgan, T. H. — Recent Theories in Regard to the Determination of Sex. Pop-
ular Science Monthly. Dec, 1903.
Pfeffer, W. — Locornotorische Richtungsbewegungen durch chemische Reize.
Untersuchungen aus d. Bot. Inst, zu Tiibingen. I. 1884.
Steinach, E. — Untersuchungen zur vergleichenden Physiologic der mannlichen
Geschlechtsorgane insbesondere der accessorischen Geschlechtsdriisen.
Archiv f. d. gesammte Physiologic. LVI. 1894.
Strasburger, a. — Das botan Prakticum, 2 Auf 1887. Page 402.
i
THE INFLUENCE OF CALCIUM AND BARIUM ON
THE SECRETORY ACTIVITY OF THE KIDNEY.^
BY
JOHN BRUCE MacCALLUM, M. D.
From the Rudolph Spreckels Physiological Laboratory of the
University of California.
In previous publications^ it was shown that subcutaneous or
intravenous injections of small quantities of solutions of certain
salts, including the saline purgatives, produce not only increased
peristalsis, but also an increased secretion of fluid into the in-
testine. This was found to be true also when the solutions were
applied locally to the peritoneal surfaces of the intestine. It was
suggested that the main actions of saline purgatives consist in
the production of increased peristaltic movements, and of in-
creased secretion of fluid into the intestine; and that the semi-
fluid foeces which are produced by saline purgatives are the
result not of decreased power of absorption by the intestine,
but of an increased secretion of fluid into the intestine. It was
further shown that the administration of calcium or magnesium
chloride tends to suppress the peristaltic movements and the secre-
tory activity of the intestine. Attention was specially called to the
marked action of barium chloride in the production of violent peri-
staltic movements and ringlike constrictions in the intestine, and
also in the production of an increased flow of fluid into the intes-
tine. It was also pointed out that the production of these activi-
ties in the intestine by the purgative salts, and their suppression
by calcium and magnesium is analogous to the production and
suppression of rhythmical contractions in voluntary muscles de-
1 A preliminary report of these experiments was published in the University
of California Publications, Physiology, January 15, 1904, Vol. I., No. 10, p. 81.
2 MacCallum, J. B. — American Journal of Physiology, Vol. X., No. III.,
p. loi, 1903, and Vol. X, No. V, p. 259, 1904.
i8o Joliii Bruce MacCallinn, M. D.
scribed by Loeb^. The antagonism which has been shown by
Loeb to exist between the actions of many sodium salts on the
one hand, and calcium and magnesium salts on the other was
further illustrated by these experiments.
It seemed possible then in the light of these facts that the activ-
ities of the kidney might be controlled in the same way as those
of the intestine. Since it is well known that many sodium salts
have a distinct diuretic action, it seemed conceivable that calcium
or magnesium might act as an antidiuretic. In order to decide
this point I have made a series of experiments in which I have
found that the relation of many of the salts to the activity of the
kidneys is similar to that which they bear to the glandular activity
of the intestine.
METHODS.
The following experiments were carried out mainly on rabbits;
a few dogs also were used. In all cases morphine was given
as an anaesthetic. The rabbits received 3-5 cc. 1% solution of
morphine hydrochlorate subcutaneously ; the dogs in addition
to this dose of morphine were given ether when necessary.
The urine was collected by catheterising the ureters or by tying
a cannula in the bladder. The latter method was employed in
all cases except those in which it was necessary to observe the dif-
ference between the amounts secreted by the two kidneys. In
placing a cannula in the bladder a small incision was made in the
abdominal wall. The bladder, which usually contains a consid-
erable quantity of urine, was then lifted out of the body cavity,
and the abdominal wall sewed up around the neck of the bladder
so that the intestines could not be forced out. A purse-string su-
ture was then made in the fundus of the bladder and an incision
made in the bladder wall within the suture. In this way the urine
could be removed, and the cannula securely tied in. Care was
taken to allow no urine to collect in the bladder, so that the meas-
urements given in the tables represent all the urine that was se-
creted during each period. The simple catheterisation of the
bladder through the urethra may be easily performed in rabbits,
1 Loeb, J. — Festschrift fiir Fick, 1899; Archiv. fiir die gesammte Physiologic.
1902, XCI, p. 248.
Influence of Calcium and Barium on the Kidney. i8i
but this method Is unsatisfactory when it is necessary to obtain
the exact amount of urine secreted in a given time since it is im-
possible to tell whether the bladder is at any time entirely empty.
Solutions of the salts whose actions were tested were introduced
into the body intravenously. In rabbits a hypodermic needle was
placed in a vein of the ear; in dogs the fluid was forced into one
of the superficial veins of the lower limb. When small quantities
were injected a hypodermic syringe was used; when larger amounts
were introduced a pressure apparatus was employed. This was
the apparatus commonly used in injection work, consisting of a
pressure bottle connected on one side with a water tap, and on
the other with a graduated bottle containing the solution to be
injected. In this way a constant pressure could be obtained, and
the quantity of fluid injected in a unit of time accurately con-
trolled. By causing the fluid to pass through a coil of rubber
tubing immersed in hot water before reaching the needle, the
solution could be kept constantly at the body temperature. Some
of the details of the apparatus were suggested to me by an ap-
paratus used by Dr. M. H. Fischer in this laboratory. For such
Infusions into the blood only 'Vs solutions were employed; In
subcutaneous injections stronger solutions were used. Except In
those cases where It was necessary to obtain the eftect of the salt
on the normal flow of urine, the secretion was considerably raised
and kept constant by the uniform infusion of ""/g NaCl solution
throughout the experiment. The effect of the other salts was
then obtained by allowing small quantities of '"/s solutions to flow
into the vein along with the NaCl solution. In other Instances
these salts were Injected into a vein of one ear while the NaCl solu-
tion was at the same time flowing Into the opposite ear. In these
experiments the ear of the rabbit was securely tacked to the board,
and the needle kept from slipping out of the vein by means of
bull-dog forceps.
EXPERIMENTS.
The results of the experiments on the actions of calcium and
barium may be best seen In the following tables :
I. Dog — small terrier — cannula placed In right ureter.
Urine secreted In ist lo minutes 3.6 cc.
2d 10 minutes 3-6 "
1 82 John Bruce MacCallum, M. D.
8 cc ""/s CaClg injected into vein of leg.
Urine secreted in ist lo minutes 2.4 cc
2d 10 minutes 2.2 "
3d 10 minutes 1.8 "
4th 10 mniutes 1.6 "
5th 10 minutes i-4 "
10 cc ""/s sodium citrate injected subcutaneously.
Urine secreted in ist 10 minutes 1.6 "
2d 10 minutes 2.3 "
3d 10 mniutes 3.1 "
4th 10 minutes 3.6 "
In this case the secretion of urine gradually decreased after the
injection of calcium chloride until the amount collected in a unit
of time was less than half of the initial amount. The addition of
sodium citrate to the blood counteracted this effect so that the
rate of secretion again approached the normal. These effects
are more striking when the quantity of urine secreted is increased
by the introduction of normal salt solution into the blood as shown
in the following experiment:
2. Rabbit — cannula placed in bladder. No urine flowed in
the first or second periods of 10 minutes before the NaCl solution
was injected.
Salts other than "Vg NaCl in- Urine in
Time. NaCl injected. jected in cc. cc.
10.10 10
10.15 10
10.20 5 0.5
10.40 10 0.8
11.00 10 0.5
11.20 5 i.o
11.40 10 2.8
12.00 10 6.0
12.00 5 cc Vs CaCla intravenously
12.05 5 cc ^"/s CaClg subcutaneously
12.20 5 0.2
12.40 10 1.8
i.oo 10 0.8
1 .00 5 cc ^/g sodium citrate intravenously
1.20 10 2.2
1.40 5 3.6
Influence of Calcium and Barium on the Kidney. 183
In this experiment, although the flow of urine has been consider-
ably increased by the injection of "Vs NaCl, the introduction of
CaClo markedly suppresses the secretion. The flow of urine remains
small for an hour, although a somewhat greater quantity of fluid
is forced into the blood than in the previous hour. This suppres-
sion of urine is at once counteracted by the injection of sodium
citrate.
The following table (3) which represents only the latter half
of an experiment shows roughly the duration of the action of
smaller doses of calcium.
3. Rabbit — cannula in bladder — injections intravenous.
Salts other than Vg NaCl in- Urine in
Time. NaCl injected. jected in cc. cc.
9.25
1.40 150 64.5
1.45 10 6.G
1.50 10 5.6
1.55 10 6.2
2.00 10 7.4
2.05 10 9.5
2.05 5 cc Vs CaCls
2.10 5 2.2
2.15 10 0.8
2.20 10 1.2
2.25 10 1.6
2.30 10 2.8
2.35 8 3.0
2.40 5 4.5
2.45 o 4.8
2.50 o 5.1
2.55 O 6.2
As shown here and in other experiments, the action of calcium
is only temporary. I have found also that magnesium chloride
in many cases has cin antidiuretic action similar to that of calcium
chloride. The suppression of urine, however, is not so marked
as with calcium.
As shown in the following experiment (4) barium chloride in
very small doses has a strong diuretic action. Although it is
much more powerful in this respect than sodium citrate, the in-
creased flow of urine which it causes may be suppressed by the
injection of calcium chloride.
184 John Bruce MacCallum, M. D.
4. Rabbit; injections intravenous.
Salts other than ""/g NaCl in- Urine in
Time. NaCl injected. jected in cc. cc.
10.20
10.30 20
10.40 10
10.50 20 1.2
11.00 -: 32 2.8
II. 10 28 5.8
1 1.20 20 6.1
11.30 10 8.2
11.40 10 8.3
11.40 yi ccVs BaCls
11.50 10 14.4
12.00 10 18.0
12.10 10 12.4
12.10 >^ cc '"/s BaCls
12.20 10 18.4
12.30 10 16.4
12.30 5 cc ""/s CaCL
12.40 10 8.6
12.50 10 4.0
1. 00 10 2.0
1. 10 ID 2.4
1.20 5 3.4
1.20 J4 cc Vs BaCla
1.30 8 6.4
1.40 10 8.2
1.50 10 8.6
1. 50 ^ cc ™/8 BaCla
31.55 10 i.8(
1 2.00 10 0.6 f
2.10 10 0.0
2.20 0.0
2.30 0.0
In the uniform injection of considerable quantities of normal
salt solution into the blood, the flow of urine, after about an
hour, becomes fairly constant. If an average amount of i cc.
in I minute be introduced, the secretion of urine during the first
two or three hours is usually slightly less than the amount of
fluid injected. After this time, when no other salts have been
added, the quantity injected and the quantity secreted may become
Influence of Calcium and Barium on the Kidney. 185
approximately equal. As shown in experiment 4 however the addi-
tion of a minute quantity of BaCL (less than y^ cc ""/§ solution) to
the blood causes the flow of urine to increase markedly, so that the
quantity secreted in a unit of time is far in excess of the quantity
of fluid introduced into the blood. If, however, while this active
secretion is going on, 5 cc. '"/g CaClg solution be injected into
the blood, the flow of urine rapidly decreases, although the total
quantity of fluid added to the blood remains constant. The fur-
ther addition of BaCU again increases the secretory activity so
that the quantity secreted in 10 minute periods which has fallen
from 16.4 to 2, under the influence of CaCL is again raised to 8.6
by the injection of the barium salt. An apparently contradictory
thing, however, happens when a larger amount of barium chloride
is suddenly added to the-blood. As shown in the foregoing table,
while yi cc. BaCL largely increases the urinary secretion, the
injection of ^ cc. in addition to that already present, causes an
entire cessation of the flow of urine. In some cases this suppres-
sion of the flow of urine is quite abrupt; in other instances it is
more gradual, a few drops of urine flowing from the cannula at
intervals. As shown in the following experiment, the injection
of CaCL sometimes counteracts this action of larger doses of
BaClg and causes the urine to flow again.
5. Rabbit — cannula in bladder; injections intravenous.
Salts other than "Vg NaCl in- Urine in
Time. NaCl inj"ected. jected in cc. cc.
9-55
10.00 10
10.15 15
10.30 15
10.45 15
11.00 15
11.00 I cc Vs BaCU + 4 cc Vs NaCl
II. 15 10
11.30 15
11.45 15
11.45 5 cc. Vs CaCla
12.00 10
12.15 • .15
12.30 15
I
.0
3
•4
5'
.2
5'
. I
4
.8
2 ,
■4
0,
.2
•3
2,
.0
3
.8
4
.0
1 86 John Bruce MacCallum, M. D.
In this case i cc "/s BaClg gradually suppresses the flow of
urine, and no trace of the strong diuretic action of barium is seen.
And, further, calcium chloride has here an action which seems at
first glance entirely opposed to that which it ordinarily has. As
shown in the previous experiments, calcium characteristically sup-
presses the secretion of urine. In this case the flow of urine
increases after its administration. These apparent contradictions
may be explained in the following way. In discussing the actions
of calcium and barium on the intestine, it was pointed out that
barium chloride, like the other saline purgatives, affects the in-
testine in two ways, namely, by increasing the peristaltic move-
ments and by increasing the secretion of fluid into the lumen. At-
tention was further called to the violent character of the muscular
contractions in the intestine caused by barium, which may so con-
strict the lumen of the intestine that fluid cannot pass from one
part to another. It was also shown that calcium to some extent
counteracts the action of barium both on the muscle, and on the
glands of the intestine. It seems therefore probable that the
increase in the flow of urine caused by small doses of barium
chloride {yi cc. ""/« solution) is due to an increase in the secretory
activity of the kidney entirely analogous to that which is pro-
duced in the intestine by the same salt. The cessation of the flow
of urine however which follows the administration of larger doses
of barium chloride (icc. "/g solution) is in all probability due
to the action of the barium on the muscle coats of the urinary pas-
sages, especially those of the calyces and pelvis of the kidney, and
those of the ureter. Since all of these various parts of the urinary
passages are surrounded by thick, circular and longitudinal muscle
coats, not unlike those of the intestine, it seems conceivable that
a strong contraction of these coats, such as barium is capable of
causing in the intestine might effectually shut off the lumen so
that no urine could pass. Furthermore the action of calcium in
renewing the flow of urine under these circumstances is quite
analogous to its action in suppressing the peristaltic movements
or in relieving the constrictions in the intestine caused by barium.
The actions of calcium and barium which are shown in Table 5,
are on the muscle coats of the urinary passages. It is quite conceiv-
Influence of Calcium and Barium on the Kidney. 187
able however that this action of calcium may coexist with its char-
asteristic action in diminishing the secretory activity of the kidney.
In both the intestine and the urinary apparatus (kidney, and
urinary passages) barium stimulates the glandular and the mus-
cular tissues to activity. Calcium on the other hand uniformly
suppresses these activities.
It must be pointed out however that the suppression of the
flow of urine which follows a relatively large dose of barium
chloride cannot always be relieved by calcium. As was found to
be true In the intestine, the action of barium Is seldom completely
counteracted by calcium. In many cases the barium stops the
flow of urine entirely so that it is not possible to start it again.
This is shown In the following experiment (6) where relatively
large quantities of calcium chloride are incapable of reestablishing
the flow of urine. This naturally suggests the idea that the large
doses of barium may stop the secretion of urine by injuring the
cells of the kidney, or perhaps Indirectly by a constricting influ-
ence on the blood vessels. These possibilities must be taken into
consideration; but the fact that calcium sometimes causes the
urine to flow again after it has been Inhibited by barium speaks
strongly In favor of the theory advanced above, that the inhibiting
action of barium on the flow of urine Is an action on the mus-
cular tissue of the urinary passages.
6. Rabbit — cannula In bladder; Injections Intravenous.
Salts other than '"/s NaCl In- Urine in
Time. NaCl Injected. jected In cc. cc.
9.50
10.20 23
10.30 20
10.40 25
10.50 28
11.00 20
II. 10 16
11.20 10
11.30 10 ,
11.40 12
11.50 15
1 1.5 I ^ ccVs BaCl2
11.55 i-o
0.
.8
I ,
.8
2 ,
.8
4
.6
5'
•4
5'
.8
5'
.2
7
. 2
1 88 John Bruce MacCaUum, M. D.
12,00 15 ,. 14-4
12.00 . Yz cc
12.05 —
12.10 —
12.20 —
12.30
12.32 5 cc "Vj
12.40
/s oaCla
1.6
0.0
14
12
0.0
12
0.0
; CaCl^
15
0.0
CaClg
13
0.0
10
0.0
15
20
0.0
0.0
12.50 5 cc
1. 00 —
1. 10 —
1.20 -
It will be noticed in this experiment (6) that immediately after
the injection of ^ cc. ""/g BaClg solution there is a marked diminu-
tion in the flow of urine followed within a few minutes by a very
considerable increase. This partial cessation of the flow imme-
diately following the injection is due probably to a temporary
action of the barium on the muscle coats of the urinary passages.
The subsequent increase is the result of the diuretic action of
barium on the kidney as described above.
In considering the actions of calcium and barium we must
therefore take into account not only their influence on the glandular
tissue, but also their effect on the muscular tissue of the body.
In all cases these salts are antagonistic in their action ; and their
influence on the secretory activity of the kidney and on the flow
of urine is entirely analogous to their influence on the glandular
and muscular activities of the intestine. With regard to its action
on the kidney calcium chloride may be properly termed an anti-
diuretic.
Attention must be again called to the extremely poisonous nature
of barium chloride. A subcutaneous injection of 3 cc 'Vs BaCU
solution is usually sufficient to kill a rabbit. Intravenously it
should always be injected with four or five times its volume of
'Vs NaCl solution.
CONCLUSIONS.
I. In dogs and rabbits the quantity of urine secreted in a unit
of time may for a time be markedly diminished and in some cases
almost entirely inhibited by the Introduction of calcium chloride
into the circulation.
Influence of Calcium and Barium on the Kidney. 189
2. Calcium chloride diminishes not only the normal flow of
urine, but also that which is caused by the administration of saline
diuretics. For example, the rate of secretion which has been
largely increased by the intravenous injection of normal salt so-
lution may be temporarily lessened to a marked extent by the in-
troduction of CaClo into the blood.
3. In all cases "/g solutions were used, and ""/g NaCl solution
was introduced into the blood at a constant rate throughout the
experiments. After a short time the rate of secretion became
constant. It was then found in rabbits, that the addition of a
small quantity of BaCL (^ cc ""/g solution) to the blood causes
a marked increase in the flow of urine, so that the amount of
fluid secreted may considerably exceed that which is introduced
mto the blood during the same period of time.
4. This action of barium is counteracted by the injection of
CaCls.
5. If a larger quantity of BaCL ( i cc ""/g solution) be added
to the blood, the flow of urine ceases and often complete anuria
ensues. In some cases the injection of CaCL abolishes this in-
hibitory action so that the urine flows again. Usually however
the action of barium persists.
6. The fact that barium when given in smaller and in larger
doses may thus apparently have opposite effects on the flow of
urine may be explained by analogy with Its action on the intes-
tine. Barium chloride causes not only an increase in the secretion of
fluid into the intestine, but also active peristaltic movements, and
violent local constrictions of the Intestine. Similarly very small
doses of BaClg increase the secretory activity of the kidney. It
seems^ probable however that the cessation of the flow of urine
which follows the injection of larger quantities of the salt is due
not to an inhibition of secretion, but to the action of the barium
on the muscular coats of the urinary passages, especially those
of the calyces and pelvis of the kidney and those of the ureter.
This action would bring about a constriction of the tubes and a
closure of the lumen. The fact that calcium counteracts both
effects of the barium supports this explanation.
190 John Bruce MacCallum, M. D.
7. The influence of calcium and barium on the flow of urine
is in every way analogous to their action on the intestine, which
I have previously described. The suppression of the urinary
secretion by calcium is also analogous to the suppression of twitch-
ings in voluntary muscles by calcium, which has been described
by Loeb.
In conclusion it is a pleasure to thank Professor Loeb for the
interest which he has taken in these experiments. I am indebted
also to Dr. Theo. C. Burnett, who has assisted me in many of
the experiments.
THE INHIBITIVE ACTION OF THE ROENTGEN
RAYS ON REGENERATION IN
PLANARIANS.
BY
CHARLES RUSSELL BARDEEN, M. D.,
Associate Professor of Anatomy, the Johns Hopkins University, Baltimore.
AND
F. H. BAETJER, M. D.,
Director of the Roentgen Apparatus, the Johns Hopkins Hospital, Baltimore.
The great capacity of regeneration possessed by fresh-water
planarians is well known because of the number of investigations
lately devoted to the subject.^ Complicated portions of the body,
like the head and pharynx, when removed, are restored within a
few days or weeks. In several species a new individual may
be regenerated from a very small, isolated piece. We have found
that this power of regeneration may be completely destroyed by
exposing planarians to the action of the Roentgen rays.
Our experim.ents have been conducted upon two species of
planarians which have especial regenerative capacity, P. maculata
and P. lugubris.. Specimens were placed in shallow, open dishes
about ten to fifteen centimeters below the vacuum tube and were
exposed from ten to twenty minutes each day for varying periods
of time. We made use of a twenty-centimeter coil with an inter-
rupter of the electrolytic type, a ten-Inch coil with a mechanical
Interrupter, and several styles of vacuum tubes. The vacuum of
each tube was so arranged that rays of "medium-soft" quality
were obtained.
1 In addition to the literature quoted in Morgan's "Regeneration," New York,
1901, articles have appeared as follows: F. R. Lillie, American Journal of
Physiology, VL, p. 129, 1901 ; E. Schultz, Zeitschrift f. wissenschaftliche Zo-
ologie, LXXII, p. I, 1902; N. M. Stevens, Archiv. f. Entwichelungsmechanilk
XIIL, p. 396, 1901 ; T. H. Morgan, Archiv. f. Entwichelungsmechanik, XIIL, p.
179, 1901 ; Biological Bulletin IIL, p. 132, 1902 ; H. F. Thacher, American Nat-
uralist, XXXVL, p. 633, 1902; C. R. Bardeen, Biological Bulletin III., 262, 1902;
Archiv. f. Entwichelungsmechanik, XVP., p. i, 1903.
192 Charles Russell Bardeen, M. D.
Our first experiments were upon worms from each of which
the anterior region of the body was removed Immediately before
the first exposure. The cut edges became closed In by muscular
contraction and the extension of epithelium In the usual manner.
For some days there was a slight production of new tissue near
the cut surfaces, but this soon ceased. No new heads were pro-
duced and no new pharynges. In one specimen of P. maculata,
however, an Imperfect eye was regenerated on the left side at the
junction of the old tissue with the new, and a very small eye-spot
appeared on the right side, but the anterior end of the piece
at no time assumed the normal shape of a head. The specimens
were subjected to thirteen exposures. All died between the twen-
tieth and twenty-second days after the first exposure. Control
specimens regenerated In the usual manner. A "tail-piece" of
P. maculata had become a worm of perfect form and proportions
on the fifteenth day after the operation, and a "tail-piece" of
P. lugiibris had regenerated a well proportioned new head and a
new pharynx at that time.
Several experiments were made to test the effect of the
rays on uninjured worms. Specimens were exposed from
twelve to eighteen times and were then kept under as hy-
gienic conditions as possible. Some of these Individuals lived
a month after the first exposure. During this period they
reacted normally to light, to mechanical and to chemical (food)
stimuli. Microscopical preparations made from a few specimens
at varying periods after the last exposure showed no marked
alterations In the muscular, nervous and Intestinal apparatus. The
cutaneous epithelium seemed to be normal except for a few areas
where the cells were shorter and broader than usual. The cUIa
of the ciliated cells were Intact. In specimens with well-developed
genitalia the cells of the testes showed no karyoklnesls. On the
contrary most of them seemed to be undergoing a degenerative
change. In corresponding control specimens mitosis was most
active In these cells. No clear Instances of nuclear degeneration
were observed. Death In these specimens resulted from a de-
generative process which began In the region of the head and
extended slowly back. This degeneration was probably parasytic
Inhibitive Action of Roentgen Rays. 193
in nature and seemed due possibly to insufficient protection offered
by the thinned epithelium. Young individuals, of which we had
several hatched out from an egg-capsule of P. lugubris, a few
weeks before the experiments began, were affected like the ma-
ture specimens, but more quickly.
From worms exposed from ten to fifteen times to the rays
pieces were cut immediately in some instances, and in some in-
stances a week or ten days after the last exposure. In each instance
the cut surfaces were closed by muscular contraction and mechan-
ical extension of the surface epithelium, but in no instance was
there subsequently seen any sign of the production of new tissue
at the cut surface or in a region of the body where under normal
conditions a new pharynx would be formed. Microscopical sec-
tions o'f a specimen killed within the second twenty-four hours
after the isolating cut was made showed no signs whatever of
cell-division, either direct or indirect. In control specimens mitosis
was most active in the tissue-forming parenchymal cells at this
period. In the exposed specimens the epithelium where it had
extended out to cover a cut surface remained a flat, thin mem-
brane as long as the specimen lived. In the control specimens
it was quickly restored to its normal columnar form. The ex-
posed Individuals lived for from twenty-five to thirty days from
the time of the first exposure. One piece of P. maculata, from
which the head had been remov^ed, lived for forty-one days. Re-
actions to light and to mechanical and chemical stimuli seemed
normal in all the specimens.
From these experiments it is evident that the Roentgen rays
have a powerful inhibitive effect upon cell-reproduction in pla-
narians. It may be entirely stopped by sufficient exposure. No
effect was noticed in the physiological activities or in histological
structure of the highly differentiated tissues such as those of the
nervous system and the musculature. The effects of the rays do
not appear for some days after the first exposure. Thus there
is a slight production of regenerative material at a cut surface
in a specimen sectioned before exposure to the rays. The sub-
sequent differentiation of an imperfect eye in one specimen in-
dicates that the rays have effect not so much upon tissue differ-
194 Charles Russell Bardecn, M. D.
entiation as upon cell-reproduction. The spreading out of the
surface epithelium so as to cover a cut surface, whereby columnar
epithelium becomes transformed into pavement epithelium also
indicates this. Death in exposed specimens may possibly be due
to a necessity on the part of the organism for a certain amount
of cell-reproduction.
The effects of the Roentgen rays on planarians thus tend to
support the view of those investigators who regard its effects
upon the tissues of other animals as due primarily to its action
on cells capable of reproductive activity. Scholtz in his excellent
clinical and experimental studies on the effects of X rays on the
mammalian skin^ concludes that both the nuclei and the cell-
protoplasm of the epithelial cells are injured by the rays, but
that the effect on the connective tissues, elastic tissue, muscu-
lature and cartilage is very slight if any. The skin on both sides
of a rabbit's ear may be affected when it is exposed to rays on one
side only.
The effect is not, however, a direct one upon the actual process
of cell-division. This is shown in planarians by the production
of tissue at a cut surface during the first few days of exposure
to the rays. It is indicated also by the work of Gilman and Baet-
jer on chick embryos^ which showed that exposed hen's eggs
develop even faster than control eggs for a few days although
subsequently development is markedly altered and checked.
One of us, likewise, found that exposure to Roentgen rays re-
peated frequently throughout the day for several days failed to
prevent the normal course of development in the eggs of certain
sea-urchins and teleosts during the peroid of exposure. The
latent period between exposure to the rays and the development
of a burn is well known to clinicians.
1 "Ueber den Einfluss der Roentgenstrahlen auf die Haut in gesunden und
kranken Zustande," Archiv f. Dermatologie und Syphilis, LIX, pp. 87, 241, 421,
X902.
^ Some effects of the Roentgen rays on the development of embryos, Amer-
ican Journal of Physiology, X, p-. 222, 1904.
InJiibitive Action of Roentgen Rays. 195
While the effect of the Roentgen rays is seen chiefly in the
inhibition or alteration of reproductive activity in the cells of
animal tissues it is improbable that it is limited to the results
of such action. Schaudinn has shown^ that individuals of sev-
eral species of protozoa may be killed by exposure to the Roentgen
rays for a few hours. Other forms are not, however, thus sus-
ceptible.^
1 Archiv f. die gesammte Physiologic, LXXVII, p. 29, 1899.
2 Schwarz in a recent interesting paper (Uber die Wirknng der Radium-
strahlen, Archiv. f. die gesammte Physiologic C, 532, 1903) concludes that the
action of radium rays is due to a decomposition similar to that of a dry dis-
tillation brought about in albunienoid bodies of the cell. He explains their
effect on rapidly growing tissues as due to their special power to decompose
lecithin.
n ^
EXPERIMENTAL STUDIES IN GERMINAL LOCALI-
ZATION.
EDMUND B. WILSON.
II. EXPERIMENTS ON THE CLEAVAGE-MOSAIC IN
PATELLA AND DENTALIUM.
CONTENTS.
I. Introduction. Methods.
II. Preliminary Notes on tlie normal Development.
III. The Development of isolated Blastomeres.
A. Analysis of the first Quartet.
(i) General development of isolated micromeres of the first quar-
tet (la, lb, etc.).
(2) The primary trochoblasts and their products (i^, i^.i^ 12.1.1^
etc.).
(3) Development of the i^ cells (1/16).
(4) Development of isolated apical cells and secondary trocho-
blasts.
(5) Development of the isolated entire first quartet.
(6) Summary on the first quartet.
B. Experiments on Cells of the lower Hemisphere.
(i) The isolated i/8-macromere (lA, iB, etc.).
(2) The isolated i/i6-macromere (2A, 2B, etc.).
(3) The isolated second quartet-cells (2a, 2b, etc.).
(4) Isolated cells obtained by maceration en masse.
(5) Summary on the lower hemisphere.
C. Development of isolated 1/2 and 1/4 blastomeres.
(i) The partial cleavage in Doitalhim.
(2) The partial cleavage in Patella.
(3) The half and quarter larvae.
IV. Summary.
V. Discussion of Results.
198 Edmund B. fFilson.
The first of these studies {Joiirn. Exp. Zoology, I, i, 1904)
was especially concerned with the question of cytoplasmic ^re-
localization in the unsegmented molluscan egg, and gave only
an incidental account of experiments on the cleavage. In that
paper both cytological and experimental evidence was presented
to show that the Dentaliiim egg contains from the beginning defi-
nitely specified regions, consisting of visibly different materials,
which stand in such a relation to the morphogenic process that
the removal of particular areas of the unsegmented egg produces
corresponding definite defects in the resulting larva. It was
shown, further, that during the cleavage process these materials
are definitely distributed to the blastomeres of the early embryo,
and that when these blastomeres are isolated they give rise al-
ways to defective larvae, showing the same general character as
those derived from the corresponding regions of the unsegmented
egg. I therefore concluded that the development of these eggs
sustains His's theory of germinal prelocalization ("Organbil-
dende Keimbezirke") as applied to the unsegmented egg, and
Roux's mosaic theory as applied to the cleavage process, and is in
harmony with the theory of formative stuffs.
In that paper, the evidence for the mosaic character of the
cleavage was given only in part, including only a brief account
of the general development, in Dentaliiim, of isolated blastomeres
from the 2-cell and 4-cell embryos, and of isolated micromeres
of the first quartet. The present paper offers more detailed evi-
dence in the same direction, derived mainly from experiments on
Patella ca^nilea. The comparative ease and certainty with
which blastomeres of any desired stage may be obtained by means
of Herbst's calcium-free sea-water led me to hope that a fairly
complete experimental analysis of the potencies of the cleavage
cells might be carried out; and I do not doubt that in time such
an analysis can be effected. Various practical diflRculties, how-
ever, have rendered the analysis here offered incomplete in sev-
eral directions. Nevertheless the positive results attained form
the most detailed and, as I think, convincing evidence of mosaic
development thus far produced, and in my judgment clearly dem-
onstrate this general principle in the molluscan egg. Despite the
Experimental Studies en Germinal Localization. 199
obvious gaps that they show In some directions, I therefore pub-
lish the results as they stand, with the hope that they may be
extended hereafter/
METHODS.
The eggs ,of Patella cocrulea were obtained in a mature state from March
until June, those of Dentalium entalis during June and July. Artificial fertiliza-
tion is easily effected in Dentalium, but is much more difficult in Patella. In
the latter case I found, after many trials, that the eggs fertilized more readily
if first placed for half an hour in sea-water rendered slightly alkaline by the
addition of 4-6 drops of a 5% solution of potassium or sodium hydrate to
half a litre of sea-water (the slight precipitate first formed quickly dissolves
upon agitation). The spermatozoa were also placed in the alkalized water
for the same length of tiime. From 15 to 20 minutes after fertilization (in the
same water) the eggs were as a rule transferred to a large quantity of pure sea-
water brought from the open sea.
In both forms the opaque tgg is at first surrounded by a very distinct
membrane, whrich, in the case of the ripe eggs, disappears as the eggs lie in
water before fertilization, in Patella by gradually dissolving away and disinte-
grating at several points, in Dentalium by suddenly bursting and being thrown
ofif. Double fertilization occurs rarely in Dentalium, but very frequently in
Patella, so that ^in the latter case it is essential to pick out the normal eggs one
by one with a pipette at the 2-cell stage. In both forms the blastomeres can
be separated with the greatest ease by means of Herbst's calcium-free sea-
water — indeed, the action is so energetic that better results are obtained if
it is restrained somewhat by mixing the artificial water with a certain amount
of normal sea-water. The eggs were placed in the artificial water shortly
after both polar bodies had formed, and after division the blastomeres were
carefully separated under the lens with a fine scalpel and immediately isolated
in normal sea-water. Even so, however, the blastomeres often continue to
separate in the normal water, and the best results for the earlier stages were
obtained by not employing the artificial water, but by separating the cells
with the scalpel in normal water. This is difficult in Patella, but very easy in
Dentalium. in the earlier stages. For somewhat later stages the artificial water
must be used; but this can often be successfully accomplished by transferring
the 2-cell stages to normal water and separating the blastomeres at the proper
stage. The tendency to separate after transference to normal water steadily
decreases as the development proceeds; hence good results for still later stages
are obtained by allowing the eggs to segment in the artificial water up to the
16-32-64-cell stages, before isolation and transfer to normal water. For greater
certainty of identification the best plan is to separate and isolate the blasto-
iLike the preceding work, this was done at the Naples Zoological Station
between February and the end of July, 1903, on a grant from the Carnegie In
stitution of Washington. I would again express my great indebtedness to
the administration of the Station for the unremitting care and efficiency with
which my work was aided.
200 Edmund B. JVilson.
meres after each division, transfering them to normal water at the stage de-
sired and all my critical results have been thus attained. The mortality is very
large, since the blastomeres seem to suffer severely in the change from the ar-
tificial to the normal water, and is greatest in cells from the vegetable hemi-
sphere; hence my failure thus far to isolate successfully the second somatoblast
(or primary mesoblast-cell, 4d), in some respects the most interesting of all
the cells. For the latest stages I did not endeavor to isolate the cells at all,
but allowed the eggs to develop for 24 hours in the artificial water, from time
to time separating the cells by jets from a fine pipette.
Most of the studies on isolated blastomeres were made on Patella, since
with Dentalimn most of my time was given to experiments on egg-fragments.
For preparation of the Patella eggs I found no better method than the simple
one employed by Patten ('85) ot acetic acid and glycerine. The eggs were
placed in a watch-glass nearly filled with sea-water, two to four drops of glacial
acetic acid added, followed by successive additions of dilute glycerine gradu-
ally replaced with strong glycerine. This renders the embryos perfectly trans-
parent, with sharply marked cell-boundaries, and often gives preparations of
admirable clearness. A slight stain with acetic carmine often adds consider-
ably to the effectiveness of the preparation for a time, though they subse-
quently deteriorate, and for most purposes the stain is superfluous.
II.
PRELIMINARY NOTES ON THE NORMAL DEVELOPMENT.
Unfortunately the cell-lineage of neither Patella nor Dental-
iiim has been worked out. Patten's early paper on the embry-
ology of Patella ('85), excellent as it is in many respects, leaves
this part of the development nearly untouched, and the same is
true of the still earlier paper of Lacaze-Duthiers ('57) and that
of Kowalewsky ('83) on Dentalium. Everyone familiar with
work of this type will appreciate the fact that to work out the
cell-lineage fully would require prolonged study, and both the
forms here dealt with present peculiar difficulties in the later
stages. The time at my disposal has only allowed me to deter-
mine the main outlines of the cell-lineage, including details es-
sential to the interpretation of the more important experimental
results. Fortunately, however, Robert ('02) has recently pub-
lished a detailed study of the cell-lineage of TrocJiuSj which agrees
so closely with that of Patella that it may be taken as a standard
Experimental Studies on Germinal Localization. 201
of comparison. To facilitate the comparison I shall employ
Robert's nomenclature, which combines certain advantageous
modifications, suggested by Conklln, Mead and Child, of the sys-
tem I used In 1892 in describing the cell-lineage of Nereis. The
primary quadrants are designated as A, B, C and D (D being the
posterior one), the corresponding micromeres as a, b, c and d;
the coefficient (i, 2, 3 or 4) designates the number of the quar-
tet, or In case of the basals (macromeres) the number of divi-
sions they have undergone; each exponent denotes a subsequent
division, i designating the cell nearer the animal pole, 2 the sis-
ter-cell nearer the "lower pole. Thus, starting with the 4-cell
stage, D divides Into iD and id; iD into 2D below and 2d
above; id into id^ above and id' below (the primary trocho-
blast) ; id^ Into id^-^ above (primary rosette-cell at the upper
pole) and id^, •- below (primary cross-cell) ; 2D Into 3D and 3d;
2d Into 2d^ and 2d", etc. Since in Patella the quadrants cannot
be distinguished by simple inspection before the 32-cell stage I
shall in general, where the quadrant is unknown, omit the letter.
Thus the primary trochoblast is i", the primary rosette-cell i^-\
a primary quartet-cell i, 2, 3 or 4, and so on.
Both Patella and Dentalium are typical examples of the spiral
type of cleavage, the former being of the symmetrical type (like
Crepidiila, Trochus, Hydroides or Polygordius) In which the
four quadrants are of nearly or quite equal size, the latter of the
asymmetrical type (like Nassa, Ilyanassa, Unio, Nereis or Jm-
phitrite) In which the first division Is unequal and the posterior
quadrant Is larger than the others until after both have been
formed. Dentalium, further. Is characterized by the formation
during the first three cleavages of a large polar lobe which after-
wards fuses with the posterior cell, CD, the egg passing at the
first cleavage through the characteristic "trefoil stage" that so
commonly occurs among mollusks (Nassa, Ostrea, etc.) and oc-
casionally In annelids (Myzostoma, Sahellaria, Chaetopterus) .
In a preceding paper ('04) I have sketched the early cleavage of
Dentalium and will here describe primarily that of Patella.
The egg of Patella first divides Into equal quadrants, without
the formation of a polar lobe; and the 4-cell stage Is remarkable
202
Edmund B. inison.
Fig. I.
Experimental Studies on Germinal Localization. 203
Normal Development of Patella.
(From Acetic-Glycerine Preparations; x20o).
I, 4-celI stage, from upper pole; 2, 8-cell stage, from upper pole, prep::ring
for fourth cleavage; 3, i6-cell stage, from the side (primary trochoblasts
shaded); 4, 32-cell stage, from the side; 5, 48-cell-stage (transitional to 56-
cell stage); 6, 48-cell stage (transitional to 52-cell), from upper pole; 7, 58-
cell-stage, from upper pole, after division of the rosette-cells and establish-
ment of the primary cross: 8, ctenophore-stage, about 10 hours, from upper
pole, primary trocboblasts ciliated; 9, 52-cell stage, from lower pole; 10, em-
bryo of about II hours, from the right side, showing three secondary trocho-
blasts in the lateral gap; 11, the same embryo from the left side; 12, portion
of the same, anterior view; at the opposite end are two secondary trochoblasts
(primary trochoblasts stippled, secondary unshaded).
204 Edmund B. Wilson.
from the fact that in the early stages it often shows no cross-
furrow (thus differing from Trochiis) ^ or if one is present it is
very short (Fig. i), though in the 32-cell stage a characteristic
cross-furrow is present at the lower pole (Fig. 13). It is there-
fore impossible to identify the quadrants in the earlier stages
without having observed the divisions from the beginning. As
usual, three quartets of ectomeres are successively formed by al-
ternating dexiotropic and leiotropic spiral or oblique divisions.
The micromeres of the first quartet, often only slightly displaced
towards the left, are considerably smaller than the basal cells,
but are relatively larger than in Trochiis (Fig. 2). The fourth
cleavage is closely similar to that of Trochiis^ each of the upper
cells dividing slightly unequally and each of the basals somewhat
more unequally to form the second quartet. In the i6-cell stage
(Fig. 3) the egg consists as usual of four large basal cells (2A,
2B, 2C, 2D), four smaller upper cells (la^ — id^) and eight
alternating cells surrounding the equatorial region. Four of
these, of equal size, form the second quartet (2a — 2d). The al-
ternating four, which are somewhat smaller (la- — id"), are the
primary trochoblasts,^ by two successive equal divisions of which
arise the sixteen cells of the primary prototroch. The i6-cell stage
is thus closely similar to that of Trochus, except that the basal
cells are relatively smaller while all the others are relatively
larger.
The fifth cleavages are dexiotropic and symmetrical through-
out the embryo, and again agree in the main with those of Tro-
chus. Each of the basals divides unequally to form a cell of the
third quartet, relatively somewhat larger than in Trochiis^ while
each cell of the second quartet divides nearly equally (in Trochus
this division is distinctly unequal). The primary trochoblasts di-
vide equally, the upper cells unequally, so as to form at the upper
pole a rosette of smaller cells (Fig. 6) almost identical with those
in Trochus, but slightly larger.
The 32-cell stage thus attained (Fig. 4) is at first perfectly
radially (spirally) symmetrical. From the four large symmet-
^These cells, and their products, are stippled in all of the figures
Experimental Studies on Germinal Localizaton. 205
rically placed basal cells arises the ento-mesoblast, while as usual
the 28 remaining cells constitute the ectoblast.
The sixth cleavages (32-64 cells) are in the main oblique and
leiotropic; but unlike Trochus the posterior micromeres of the
third quartet depart more or less widely from the type. Proceed-
ing from the upper pole downwards the divisions are as follows.
The rosette-cells (i/"^) divide nearly equally in regular spiral
order exactly as in Trochus^ so as to form a symmetrical group of
eight small cells at the upper pole (Fig. 7) which form, certainly
in part and probably as a whole, the basis of the apical organ.
Nearly at the same time the i^- cells divide nearly equally, so as
to form the primary "cross," which, as in Trochus, has at this
period spirally curved arms (Figs. 6, 7). The trochoblast-pairs
(i"-^ and I'-) divide equally, somewhat earlier than the fore-
going, so as to produce four symmetrically placed groups of four
equal cells (Figs. 5-7). This division takes place much earlier
than in Trochus, and no further division occurs In the products,
which become ciliated from the eighth to the tenth hour and form
the primary prototroch. The second quartet cells divide at about
the same time in a very characteristic fashion that is almost identi-
cal with that occuring In the nemertine egg and nearly similar to
that of Trochus. The upper left cell (2.^) divides slightly un-
equally, the smaller cell lying above and between the two ad-
joining trochoblast groups (Fig. 5). The lower right cell (2-)
divides still more unequally, the smaller lower cell (2--) lying
below against the corresponding macromere, and between the two
adjoining cells of the third quartet. In Trochus this cell Is smaller
still. The egg thus attains a 56-cell stage, at which a slight pause
occurs, and In the meantime a marked change occurs in one of the
macromeres which, I think. Is undoubtedly the posterior one, 3D.
This cell rapidly passes into the Interior, Its outer end becoming
greatly reduced, and being connected with a narrow neck with
a swollen interior portion, the nucleus however still lying at the
surface (Figs. 9, 13). The next cells to divide are those of the
third quartet. The two anterior ones divide leiotropically, like
the preceding micromeres. In the two posterior ones, however,
the spindles assume a bilateral position, with the central poles
2o6 Edmund B. JFilson.
close against the outer end of 3D (Figs. 9, 13) ; and while I
have not actually seen the division, it is nearly certain from the
position of the spindles that the division is unequal. The study
of a good many preparations of this stage leads me to believe that
this is a constant relation.
The last cells to divide in the sixth cleavage are the macro-
meres, and of these 3D is the first. At the time of its division it
is only connected with the surface by a very narrow neck, as-
suming the extraordinary appearance shown in Fig. 14. The
result of this division is to form a large rounded cell, that lies
quite in the upper hemisphere (shown in Figs. 15, 16) and a
more superficial cell. From the conditions observed at a slightly
later stage I believe the former to be 4D, the latter the primary
somatoblast 4d or M ; but I am not entirely certain of this identi-
fication. Slightly later the remaining macromeres divide some-
what unequally, the cells in the meantime undergoing consider-
able shiftings and extending further up into the egg, so that it is
exceedingly difficult to identify them individually. The ectoblast-
cap has now extended far down towards the lower pole, so that
the macromeres are connected with the surface by narrow necks.
The cell I believe to be 4d now divides symmetrically into two
to form two large symmetrical cells lying between the entomeres
and the ectoblast (Figs. 15, 16), which correspond with the
mesoblast pole-cells as figured by Patten (<?. g., in his Figs. 27,
36). I have not positively traced these cells into the coelome-
soblast, but believe there can hardly be a doubt as to their na-
ture.^ At this period the large inner cell (identified as 4D) is
still undivided (Fig. 15) the primary trochoblasts have devel-
oped cilia, and the apical tuft is present (10-12 hours).
Beyond this point I shall not for the present attempt to trace
the general cleavage, but will pass on to some points in the later
development. Patten has given figures of the trochophore of
^Sections of the trochophores of 24 hours clearly show two large meso-
blastic pole-cells (one of which appears in Fig. 17) near the posterior end,
from which two mesoblast-bands extend forward as figured by Patten, c. g.,
in his Fig. 50.
Experimental Studies on Germinal Loealization. 207
Patella with which in the main my observations agree, though
the arrangement of the cells of the prototroch is somewhat
schematized. The embryo becomes ciliated at about eight
to ten hours (depending on the temperature) the first cells to
acquire cilia being the sixteen primary trochoblasts. For a brief
period the prototroch consists of only those sixteen cells, still ar-
ranged in four separate groups (Fig. 8). The cilia are from the
first arranged, not in vague patches or tufts, but in very definite
oblique transverse rows, which bear a marked resemblance to
the swimming plates of a ctenophore — indeed, it hardly seems
forced to compare the embryo directly at this period to a larval
ctenophore. At the same time, or a little later, the group of
small cells at the apical pole, derived mainly, if not wholly, from
the apical rosette, develops a tuft of flexible but non-vibratile sen-
sory flagella, and constitutes the apical organ.
The ctenophore-stage is of short duration. In two or three
hours several cells lying in the gaps between the four groups of
primary trochoblasts also become ciliated and ultimately enter the
prototroch as secondary trochoblasts. These cells are not more
than half the size of the primary trochoblasts (a point of im-
portance in connection with the experimental results) , and at
first bear much smaller cilia. There are at least three and prob-
ably four of these trochoblasts in each quadrant (with the pos-
sible exception of the posterior group, in which I have only
certainly seen two of these cells), (Figs. 10-12), giving a total
of 28 to 32 cells in the prototroch, to which possibly still others
may be added. While I have not traced step by step the exact
origin of these cells, their position in the embryo leaves little
doubt that in each quadrant two of them are derived from the
first quartet (/. e.^ from derivatives of the i'- cells), and this
is demonstrated to be the case by the experimental evidence. The
position of the third cell (Fig. 10) shows almost beyond a doubt
that it is derived from the second quartet, /'. e., from 2^ and
probably from 2.^\ The experimental evidence again proves
that at least one, and in some cases two, trochoblasts are derived
from the second quartet. As may be seen in Fig. 10, a second
cell lies next to the one described, the position of which indicates
20^
Edryiund B. Wilson.
Fig. II.
Experimental Studies on Germinal Localization. 209
Normal Development and Larva from Egg-Fragment, Patella; x 200.
13, 48-cell stage, lower pole: 14, larva of 9 hours, sagittal optical section,
division of 3D; 15, larva of 12 hours, optical sagittal section, showing left
primary mesoblast (?) in division; 16, the same larva, in frontal optical sec-
tion; 17, trochophore of 30 hours, from the left side (from a total preparation,
shell-gland (s. g.) and primary mesoblast (m) from a corresponding actual
section), prototrochal cells in surface-view, body-wall in section; 18. larva of
20 hours, from upper pole, showing the cells as accurately as possible (some
of those just anterior to prototroch could not be clearly seen and have been
omitted); 19, optical section at the level of the prototroch of normal larva
of 24 hours; 20, larva of 24 hours from fertilized egg-fragment that segmented
like a whole egg, prototrochal cells in surface-view; 21, optical section of the
same larva at the level of the prototroch; in all these figures the prototrochal
cells are shown as acurately as possible.
2IO Edmund B. ffllson.
that it may also enter the prototroch.^ The 28 (32?) trocho-
blasts are at first arranged in two roughly alternating rows en-
circling the embryo slightly above the equator; and the ciliary
plates of contiguous cells are still not united to form a continu-
ous ciliary girdle (Figs. 10-12). Later, extensive shif tings of
the cells occur in such wise that a principal circle of trochoblasts
is formed in a single circle completely surrounding the embryo,
bearing a perfectly continuous series of powerful cilia (Figs.
17-19). The cells in this row vary in number from 19 to 21 —
a fact of which no doubt is left by the study especially of acetic-
glycerine preparations, in which the cells may be seen with sche-
matic clearness. Posterior to this row lies a second row of smaller
elongated trochoblasts, which in the dorsal region become as
large as those of the principal row (Fig. 17). At this point,
therefore, where in so many trochophores a gap exists in the
prototroch, the ciliated belt is not only closed, but broader than
at any other point. At this point the prototroch is often three
cells wide; elsewhere I have not been able to distinguish three
rows of cilia as figured by Patten, though three such rows are
certainly present in Dentalhim.
The trochophore of 24-30 hours (Fig. 17) is in the main sim-
ilar to that of Dentalium, as described in my former paper, but
the post-trochal region is relatively larger, the pre-trochal region
less pointed, the apical tuft shorter and broader, and the apical
plate less clearly marked off from the surrounding ectoblast as
may be very clearly seen in sagittal section. In this respect my
^This derivation of the prototroch in Patella agrees closely with that of
Isclmochitou (Heath, 99), where two cells in each quadrant are likewise con-
tributed from i^-, and two from the second quartet except in the D-quad-
rant, wbere a non-ciliated dorsal gap exists from the first. I have determined
beyond doubt, I think, that at least two secondary trochoblasts are formed
in the mid-dorsal line, as shown in Figs. 10-12, where there are three such
trochoblasts in three of the quadrants and two in the fourth. There is further
no doubt whatever that the completed prototroch is closed in the mid-dorsal
line (Figs. 17-19). Robert describes the prototroch of Troclius as agreeing ex-
actly with that of Am[>hitrite and Arenicola, no cells being derived from the
first quartet except the primary trochoblasts. It appears to me, however, that
his observations do not fully establish this. {Cf. the useful comparative table
given by Robert at p. 420).
Experimental Studies on Germinal Localization.
211
larvae seem to differ somewhat from those figured by Patten,
which show an extremely distinct apical plate. The later stages
are in the main similar to those described by Patten, and need
not here be considered.
III.
THE DEVELOPMENT OF ISOLATED BLASTOMERES.
In general, as Crampton ('96) found for the 2- and 4-cell
stages of lUyanassa, the isolated blastomere, at whatever stage
it be separated' from its fellows, continues to segment essentially
in the same way as if forming a part of a whole embryo; but a
point on which I would lay stress is that there is a tendency for
all unequal divisions to be less unequal than in the normal devel-
opment, though this is by no means always the case, and the
isolated blastomere often divides exactly as in a whole embryo.
The partial character of the cleavage is also frequently masked
by shifting of the cells, and the partial embryos often close, some-
times at a very early period. Such shifting or closure appears,
however, to have no effect on the differentiation of the cells, as
is shown with especial clearness by the history of the trochoblasts.
Differentiation takes, in the main, the same course as if the cell
had remained united to its fellows, and gives rise to structures
that agree in a general way, and sometimes exactly with the parts
to which the cells would have given rise in a complete embryo.
For the sake of clearness I shall not follow the most logical or-
der, but will present first the cases that most completely sustain
the above statement — namely, the blastomeres of the first quartet.
It may be premised that all of the isolated blastomeres assume a
nearly or quite spherical form before division occurs, showing
no trace of flattening on one side; and they are indistinguishable
from one another except in size, and in the slightly greater trans-
parency of the micromeres. It is also necessary to bear in mind
that both in Dentalium and in Patella the eggs from different fe-
males vary very considerably in size, so that exactly correspond-
ing blastomeres from different eggs likewise present consider-
212 Edmund B. Wilson.
able size variations. This accounts for certain discrepancies in
the figures, which represent blastomeres from many different eggs,
and possibly even from different species, though most of them
are from P. cceridea. The typical size-relation in this species,
from the eggs of a single female, are shown in Fig. loo, the
successive concentric outlines representing the entire egg, the J-2-
blastomere, M-blastomere, >^-macromere and >^-micromere. Dis-
tinct deviations from these mean volumes will be observed in the
figures.
A. ANALYSIS OF THE FIRST QUARTET.
I. General development of isolated micromeres of the first
quartet (^s -embryos) .
As described in my preceding paper, the development of the
posterior micromere of this quartet (id) in Dentalium differs
from that of the others in being the only one to form an apical
organ. In Patella this is not the case, each micromere giving rise
to a closed ectoblastic structure, bearing at the posterior end a
group of active trochoblasts and at the anterior end an apical
organ (Figs. 28-29).^
The first cleavage is slightly unequal (sometimes nearly or
quite equal), (Figs. 22-24). I ^t first supposed the smaller cell
to be the primary trochoblast (i") since in the whole embryo
this cell gives the appearance of being slightly the smaller (in
Trochus this division is described as "nearly equal"). When,
however, the entire j4-blastomere segments in the calcium-free
water it may clearly be seen, at least in some cases, that the larger
cell is the lower one (i'), and I believe therefore the primary
trochoblast is slightly larger than its fellow in the normal devel-
opment. This is typically followed by an equal division of the
trochoblast, and an unequal division of the other cell, giving a
group that closely represents one quadrant of the first quartet in
^This has not been proved in Putclla by isolation of all four of the micro-
meres from one t^^g (as was done in Dentalium') ; but among the numerous
larvae obtained of this type all that were closely examined possessed the api-
cal organ.
Experimental Studies on Germinal Localization. 213
3/ ^32 ^33
Fig. III.
Development of Isolated ^s Micromeres.
(Figs. 22-27 >^ 250; -Figs. 28-30 X 290)
22, isolated micromere; 23. 24, first division; 25, 26, two different indi-
viduals, 32/8 stage, each with two trochoblasts, one rosette-cell, and one pri-
mary cross-cell; 27, an entire quadrant segmenting after removal from cal-
cium-free water, products of first and second quartets somewhat separated
from the basal (the first quartet group seen from the outside, the others from
the inside). 28, larva of 24 hours, from J/g-micromere, from the side, showing
trochoblasts below, apical cells above; 29, similar larva (with less regular pre-
trochal region), from below, showing both primary and secondary trocho-
blasts; 30, loose group, from J^-micromere, after 24 hours in water nearly
free from calcium, primary and secondary trochoblasts, apical cells, pre-
trochal ectoblast-cells; 31, group, with two apical cells, from ^-micromere.
after 24 hours in calcium-free water; 32-33, isolated apical cells from similar
culture.
2 14 Edmund B. Ifllson.
a normal 32-cell stage — z. e., consists of two trochoblasts, one
rosette cell (i^'^), and its larger sister cell (i^"'), from which
one arises one arm of the cross (Figs. 25-27). It should be
noted that the rosette-cell almost always appears somewhat too
large, which is owing In part to the fact that it is less crowded
than in a whole embryo, but undoubtedly in part also as to a les-
sened inequality in the division of i\ Such embryos give rise to
actively swimming partial larvae, similar in a general way to the
corresponding ones in DentaJ'uim. These embryos do not gas-
trulate, but close to form pyriform ectoblastic larvae, which bear
at the larger end a group of large ciliated trochoblasts, and at
the narrower end an apical organ consisting of a group of cells
bearing stiffish motionless sensory hairs (Figs. 28-29). It may
be clearly seen that the larger end of the embryo is formed of
four primary trochoblasts, each bearing a row of powerful cilia,
while just above these at one side are two somewhat smaller sec-
ondary trochoblasts. The apical organ at this period appears,
in most cases at least, to include only two cells from which the
sensory hairs radiate like a fan. This differs from the normal
apical organ, in which the sensory hairs form a thick tuft directed
straight forwards. The radiating arrangement of these hairs
in the partial embryos appears to be due to the fact that the apical
cells do not extend so deeply below the surface, and retain a
rounded form, so that the sensory hairs spread apart like a fan,
while in the normal embryo they are crowded together and as-
sume a pyramidal shape, the free surface being considerably re-
duced. In the partial larvae, too, the sensory hairs appear rela-
tively shorter and more rigid than in the normal organ.
The composition of these larvae is shown with great clearness
by allowing the isolated ^s-micromere to develop in calcium-free
water, the action of which is more or less restrained by the ad-
dition of a certain amount of normal sea-water. All degrees of
dissociation may thus be obtained, and among the resulting cell-
groups may be found forms like Fig. 30, in which the cells lie in
a loose group, yet approximately retain their normal position. It
is evident that each of these larvae represents one quadrant of the
products of the first quartet, including four primary trochoblasts,
Experimental Studies on Germinal Localization. 215
two secondary ones, one-fourth of the apical organ and a group
of small ectoblast cells derived from i^-^, the whole structure
closing to form a morula or blastula-like structure, but otherwise
differentiating typically without gastrulating. In the aquarium
these larvas gradually disintegrate in the course of the second or
third day, the trochoblasts being always the longest-lived of the
cells, and often continuing to swim actively when the remainder
of the larvae has gone to pieces. I have not followed the details
of the development of the corresponding Isolated cells of Denta-
lium; but it is clear that their general development is closely simi-
lar. The one important difference, pointed out above, is that in
Dentaliiim only the micromere from the D-quadrant develops an
apical organ. As my experiments on Dentaliiim showed, the de-
velopment of the apical organ in this form is determined by ma-
terial that originally lies In the polar lobe, and no other conclu-
sion seems possible than that this material is in Dentaliiim finally
isolated In the posterior micromere (id), while in Patella the
corresponding stuff is equally distributed among the four mlcro-
meres. This is doubtless due to a different relation. In the two
cases, of the original segregation pattern to the first two cleavage
planes, and is perhaps connected with the absence of a polar lobe
In Patella.
2. The primary trochoblasts (1-16, 1-32, i-64-embryos) .
Exceedingly clear and interesting results are given by the Isola-
tion of the primary trochoblasts (i") or their products. If a
single trochoblast be isolated at the i6-cell stage It divides equally
twice in succession, but no further division takes place (Figs.
34-39) . From the eighth to the tenth hour each of the four cells
becomes ciliated and the group begins to swim. After twenty-
four hours the group Is swimming with great activity, and each
cell is found to bear a series of powerful cilia arranged in a
transverse row, exactly as In a normal embryo (Figs. 40-42).
The cells vary In arrangement, sometimes lying in a single plane,
sometimes having shifted so as to interlock in a rounded mass.
Exactly as in a normal embryo the action of the cilia is more or
2l6
Edmund B. Wilson.
43^
Fig. IV.
Isolated Primary and Secondary Trochoblasts.
(Figs. 34-39 X 250; Figs. 40-48 x 290).
34, primary trochoblast (1/16, i-), obtained by successive isolation; 35, 36,
result of first division; 37-39, various forms after second division; 40, product
of Figs. 34, 36, 37, after 24 hours; 41, 42, similar individuals of the same age
and history; 43, pair of primary trochoblasts, 24 hours, the products of i^-i or
i2.2 (1/32); 44, 45, single primary trochoblasts. 24 hours, products of i^-i-i,
i2-^-2, etc. (1/64); 46, pair of secondary trochoblasts, 24 hours, the products of
1^-2; 47, 48, single secondary trochoblasts, 24 hours.
less intermittent, sometimes ceasing wholly and again suddenly
being resumed. Sudden mechanical shock frequently causes a
sudden suspension of activity, followed immediately by a more
vigorous activity than before. Careful study of these embryos,
especially when they are dying, shows that the cilia in each cell
beat in the same direction; but owing to the fact that the rows
Experimental Studies on Germinal Localization. 217
of cilia rarely coincide in direction the group does not, as a rule,
rotate in a constant direction, but irregularly.
If now the two products (i^-^ and i"-) of the first division of
the primary trochoblasts be separated, each divides once, and only
once, thus giving a pair of cells that become ciliated and swim
together like the above-described group of four (Fig. 43). If,
finally, these two cells be separated at the time of ciliation — i. e.,
at a period corresponding with the 64-cell stage, no further divi-
sion occurs, but In due time each trochoblast develops its row
of cilia (Figs. 44-45) and swims singly with the fullest vigor
and activity. Such single trochoblasts often rotate steadily in a
nearly constant direction, proving that the action of the cilia
is normally coordinated. They may live for two days or more
when the action gradually ceases and disintegration occurs.
The history of these cells gives indubitable evidence that they
possess within themselves all the factors that determine the form
and rhythm of cleavage, and the characteristic and complex dif-
ferentiation that they undergo, wholly independently of their re-
lation to the remainder of the embryo. Roux's "self-differentia-
tion" here appears in the clearest and most unmistakable form.
Similar results were obtained in Dentalium, but I did not in this
case attempt to isolate the trochoblasts individually, but merely
allowed entire eggs, or isolated ^ or ^-blastomeres to continue
their development In the calcium-free water. As in Patella, the
result at the end of 24 hours is a chaos of more or less com-
pletely separated cells of different forms and sizes, among which
are trochoblasts, actively swimming, singly or In groups. These
trochoblasts fall roughly into three groups, large (Figs. 49-51),
medium (Figs. 52-53) and small (Figs. 54-55). The large
trochoblasts, which are considerably larger than in Patella, are
probably primary ones, the medium and small forms secondary
ones; and the difference in size among the latter suggest that as
in Patella they may arise from different quartets. It is worthy
of note that the cilia in all the trochoblasts are considerably longer
than in Patella, and are also relatively less numerous and
crowded. The small trochoblasts sometimes have as few as six
cilia (Fig. 55). Sections of the entire larvae show that the cilia
2l8
Edmund B. Wilson.
Fig. V.
Isolated Troclioblasts of Dcntalium and Mcsoblast-likc Cells of Patella 24 Hours:
X 290.
(All these obtained by leaving entire embryos in the calcium-free water for
24 hours).
49-51, large (primary) trochoblasts; 52, 53, medium (secondary?) trocho-
blasts; 54, 55, small (secondary?) trochoblasts; 56, group of mesenchyme-Hke
cells; 57, group of muscle-like cells.
Experimental Studies on Germinal Localization. 219
are grouped in small tufts (an arrangement I failed to note in
the isolated cells) at the base of each of which is a very distinct
deeply staining basal body; and I believe this would be an ex-
cellent object for the cytological study of the possible relation of
these bodies to the centrosome.
3. Development of the sister-cell (i^) of the primary trocho-
blast ( i/i6-embryos) .
The development of this cell, which, except for its slightly
smaller size, is indistinguishable in appearance from the primary
trochoblast, differs totally from the foregoing in the form and
rhythm of cleavage, and in the course of its differentiation. After
isolation this cell typically divides unequally to form a single
rosette cell (i^-^) (though here, too, the inequality is often less
marked than in the normal embryo, and sometimes disappears),
and its larger sister (i^"), (Figs. 58-60) ; and this is typically fol-
lowed by a nearly equal division of both these cells to form a
group of four, two of which obviously represent daughter-rosette
cells (Figs. 61, 62). The divisions do not cease here, however,
but continue, and at the end of 24 hours a larva is produced that
consists of many cells and is somewhat similar to those arising
from an entire micromere of the first quartet; this larva is, how-
ever, only about half as large, and lacks the four large trocho-
blasts at the posterior end (Figs. 63-64). At this end of the
embryo are two trochoblasts, much smaller than those of the pri-
mary group, which obviously represent the secondary trocho-
blasts of the first quartet. At the narrower anterior end is an
apical organ precisely like that of the micromere j/^-embryo,
while the middle region consists of small ectoblast cells usually
larger on one side than on the other. These embryos swim ac-
tively, but less vigorously than the J^-forms, and, like the lat-
ter, perish in the course of the second or third day. It is obvious
that the development of the i^ cell is, except for its closure, es-
sentially the same when isolated as when it forms part of a whole
embryo; and its remarkable contrast with its sister-cell, the pri-
mary trochoblast, shows most convincingly that despite their
220
Edmund B. JVilson.
Fig. VI.
Isolated i^ Cells and Isolated First Quartet, Patella.
(Fig-s. 58-62, 65, 66 X 250; Figs. 63, 64 X 290).
58, isolated iM 59, 60, two examples of first division (rosette-cell slightly
too large in both); 61, 62, two examples of second division; C>3, product, 12
hours, apical cells and secondary trochoblasts (somewhat smaller cells on re-
verse side); 64, similar larva of 22 hours; 65, first division of isolated first quar-
tet; 66, second division; 67, product, 24 hours. (Some of the cells have been
lost).
Experimental Studies on Germinal Localization. 221
closely similar external appearance, each is from its first forma-
tion definitely specified, irrespective of its connection with its fel-
lows.
4. Development of isolated apical cells, and secondary trocho-
b lasts.
In spite of many attempts, I did not succeed in rearing singly
one of the rosette cells; but isolated products of these cells, as
well as isolated secondary trochoblasts, were obtained in another
way. This was by allowing isolated micromeres of the first quar-
tet to continue their development in the calcium-free water, some
of the individuals being left undisturbed, others shaken to pieces
from time to time by a stream from a fine pipette. In those left
undisturbed for 24 hours all degrees of disintegration were ob-
served, loose masses being found from which the trochoblasts
often had broken away and were swimming about singly. In
many such loose masses the characteristic apical cells could often
be observed, loosely attached to their fellows at the end opposite
the trochoblasts (as in Fig. 30). Those that had been shaken
to pieces showed a collection of more or less completely separated
rounded cells, among which spherical trochoblasts of two sizes
(the larger evidently being the primary, the smaller (Figs. 46-
48), the secondary ones) were actively swimming singly or in
groups. Among these cells occur forms that are evidently sin-
gle apical cells, since they agree exactly in size and structure with
those attached to the loose masses referred to above. These cells
(Figs. 31-33) are ovoidal in form, and bear on one side the
characteristic non-vibratile radiating sensory processes or hairs.
There is no possibility of mistaking these cells for either kind of
trochoblast, since they are considerably smaller, and the appear-
ance and arrangement of the sensory hairs (apart from their im-
mobility) is entirely different from that of the cilia. There can,
therefore, be no doubt that both the secondary trochoblasts and
typical sensory cells of the apical plate may undergo their charac-
teristic differentiation when entirely isolated fro?fi their fellows.
Mingled with the foregoing cells are rounded, non-ciliated cells
22 2 Edmund B. JVilson.
of various sizes that are obviously isolated cells of the general
pras-trochal ectoblast.
The relatively large size of the apical cells, whether completely
isolated or forming part of the ]'i or ^o larvae, is a matter I
have not yet fully cleared up. The apical pole of the normal
larva of 24 hours (Fig. 18), seen in surface view, gives some-
what varying appearances, but clearly shows a central group of
four to six larger cells. This does not exactly agree with what
appears in the partial larvae, whiqh, as a rule, show two large
apical cells, apparently of nearly equal size; it is, however, dif-
ficult to determine the exact size of the cells in the normal larva,
owing to their crowding together. It seems probable that this ap-
parent discrepancy may be due to the fact that, as stated above,
the primary rosette cell is so frequently too large in the }i and
tV embryos, and that this results in a slightly abnormal later
development of its products. It is clear, however, that this does
not affect in any essential way the differentiation of the apical
cells.
5. Development of the isolated entire first quartet.
In the sea-urchin Driesch (1902) has shown that both the
upper and the lower quartets of the 8-cell stage may produce
complete dwarf larvae, though the two quartets show certain char-
acteristic differences in development, proving that they are not
identical. In Patella it is difficult to perform this operation, and
still more difficult to rear the larvae, since the cells always separ-
ate more or less after replacement in normal water. I neverthe-
less succeeded in obtaining a few cases. The cleavage of the iso-
lated first quartet is essentially the same as in a whole embryo.
The first division, leiotropic in all the cells, produces four upper
cells (i^) and the four trochoblasts (i") alternating with them,
the eight cells forming a nearly flat plate (Fig. 65). The sec-
ond division, dexiotropic in all the cells, produces a plate of 16
cells (Fig. 66) in the centre of which is the rosette, around which
lie the four i^-^ cells and, at the margin, four groups each of two
trochoblasts (i^S i'"). As shown in the figures, the form and
Experimental Studies on Germinal Localization. 223
grouping of these cells is essentially the same as in the upper half
of a normal 32-cell stage; though, owing to the flattening out
of the group, the normal position of the cells is somewhat modi-
fied and (as in the case figured) the cells are often rather loosely
connected.
In later stages such groups invariably broke up more or less,
and no larvae were obtained which had not lost some of the cells.
Nevertheless these larvae close up more or less completelv, form-
ing irregularly pyriform structures with an apical organ at the
smaller end and a group of trochoblasts at the larger one. The
largest of these larvae obtained probably represents at least three-
fourths of the first quartet. This individual (24 hours) is shown
in Fig. 67, drawn from a preparation; in life it swam very ac-
tiv^ely about. This larva is clearly a purely ectoblastic structure,
and shows no trace of archenteron. It is of an irregular flattened
pyramidal form, with an irregular group of apical cells at the
narrow end, forming an unmistakable apical organ. The larger
end is occupied by a group of trochoblasts, which form a some-
what irregular series around the margin, but also extend some-
what over the base. The remainder of the embryo is formed of
small ectoblast cells that have on the lower side extended more
or less into the basal region. The exact number of trochoblasts
cannot be determined, but there are at least 14, and probably a
larger number. It is clear that this larva represents a partly
closed and distorted prae-trochal region, with that part of the
prototroch derived from the first quartet, minus a certain number
of cells that have separated. The other larvce were similar in
type, but evidently represent a smaller portion of the same
region. This case, taken In connection with the other facts de-
termined, renders it practically certain that the first quartet as a
whole is here incapable of producing a complete dwarf, but gives
rise to essentially the same ectoblastic structures as in a whole
embryo. This result is entirely in agreement with that after-
wards obtained on Cerehratiiliis by Zeleny ('04), who, at my sug-
gestion, undertook a comparison in this form of the upper and
lower quartets of the 8-cell stage — a question particularly inter-
esting in this case since the upper quartet is larger than the lower.
2 24 Edmund B. Wilson.
The constant result of this experiment was, that while both quar-
tets produce closed hlastulas, only the lower one gastrulates, while
only the upper one produces an apical organ.
6. Summary on the first quartet.
The foregoing observations are sufficient, I believe, to establish
definitely the mosaic character of cleavage and differentiation in
the first quartet. This is strictly proved for the primary trocbo-
blasts (i^), for their first (i"-^ and i'-) and second (I'-^'S I^•^■^
etc.) products, and for the i^ cells; it is less strictly but hardly less
convincingly established for the apical cells and the secondary
trochoblasts. Excepting the secondary trochoblasts, the other
products of i^ do not show sufficiently definite characters to allow
of a similar definite proof, but it can hardly be doubted that the
same conclusion applies to them, and also to the first quartet taken
as a whole.
B. EXPERIMENTS ON CELLS OF THE LOWER HEMISPHERE.
Isolated blastomeres of the lower hemisphere are in general
much less tenacious of life than those of the upper quartet (a fact
parallel to that observed by Driesch in sea-urchins), and my ob-
servations are here much less detailed. In a more general way,
however, the results are entirely in agreement with the conclu-
sions reached in case of the upper quartet.
I. Development of the isolated Ys-macromere.
These cells divide, at least as far as the 64/8-cell stage, as if
forming part of a complete embryo. At the first division a cell
of the second quartet is formed (Fig. 69-70), which at the en-
suing cleavage divides nearly equally (into 2^ and 2-), while a
cell of the third quartet is produced from the basal in the proper
position (Fig. 71). At the next cleavage 2^ divides nearly
equally, 2" very unequally, forming below the characteristic small
cell 2--, that lies against 3. The latter cell then divides equally
or unequally (the quadrant probably being in the former case one
Experimental Studies on Germinal Localization. 225
of the anterior, in the latter case one of the posterior ones). A
group is thus produced (Figs. 72-73) which, excepting the usually
lessened inequality of 2^-^ and 2^', is practically identical with one
quadrant of the lower hemisphere in the 64-cell stage (Cf. Figs.
4-5). This is followed by a division, sometimes distinctly un-
equal, sometimes nearly equal, of the basal to form a cell of the
fourth quartet (Fig. 73).
Fig. VII.
Isolated 1/8 Basal, Patella; x 250.
68, isolated basal (a rather small example; cf. Fig. 100); 69, 70, examples
of first division; 71, 32/4-cell stage, typical; 72, 56/8-cell stage, typical 2-
group, equal divison of 3; 73, 64/8-cell stage, after formation of 4 (in this ex-
ample the 2-group has rotated into an abnormal position); 74, product, 24
hours, showing two secondary trochoblasts (products of 2i-i) and two feebly
ciliated cells (pre-anal cells?).
226 Edmund B. fnison.
It is exceedingly difficult to rear these embryos, many of them
dying, while most of those that live break up into smaller masses.
A few larvae were nevertheless obtained, the best of which is
shown in Fig. 74. Like the others, this larva has evidently gas-
trulated (though the entoblast-mass is relatively small, perhaps
owing to the loss of some of the cells). Its most interesting fea-
tures are the presen-ce, at one end, of two cells bearing powerful
and active cilia by the activity of which the larva rotates irregu-
larly, while near the opposite end are two cells bearing much
smaller and feebler ones. It is evident that the two anterior cells
are secondary trochoblasts, undoubtedly those derived from the
second quartet; and the fact that there are two of these may be
taken as evidence that two trochoblasts are contributed to the
prototroch, at least in some of the quadrants, by the 2^-^ cells, as
is indicated by a study of the normal embryos. The two weakly
ciliated cells were to me at first a puzzle, since I failed to observe
anything corresponding to them in the normal embryos. But it
may be recalled that Patten describes and figures a ventral reg-
ion, covered with fine short cilia, just anterior to the pre-anal
sense-organ ('85, Figs. 47, 48, 57, etc.), and while I became
aware of this when it was too late to re-examine the normal larvae,
it seems very probable that it is these cells that appear in the
posterior ciliated tract of the >^-macromere larva. This is sus-
tained by the development of the yV basal cells about to be de-
scribed.
2. Development of the i/i6-macromere.
The tV basal cell, obtained by successive isolations, divides
unequally to form the third quartet cell (Fig. 75), which after-
wards divides into two, while still later the fourth quartet-cell is
produced from the basal (Fig. 76). Only two or three such
cases were obtained, one of which developed into the larva shown
in Fig. 77. While this larva could not be very clearly analyzed,
it evidently consisted of an internal mass of cells (entomesoblast
or entoblast) surrounded by a superficial layer of ectoblast-cells.
At one end the ectoblast-cells were larger and at least one of
Experimental Studies on Germinal Localization. 227
these bore a tuft of short, weak ciHa, by means of which the
larva very slowly rotated. In view of the structure of the Ys-
macromere larva, it is probable that this ciliated cell (or cells)
represents at least a part of the posterior group of cilia which I
have assumed to contribute to the ventral ciliated tract in the nor-
mal larva.
Fig. VIII.
(Figs. 75-81 X 250; Fig. 82 X 300).
Isolated 1/16 Basal, and Isolated Cell of Second Quartet, Patella.
75, first division of 1/16 basal; "jG, second division (64/16), 4 formed; 77,
product, 24 hours; 78-81, cleavage of isolated 2; 82, product, 24 hours, with two
trochoblasts and pre-anal (?) cells.
3. Isolated blastomeres of the second quartet ( i/i6-embryo) .
These blastomeres, obtained by successive isolations, divide like
the foregoing as if still forming part of a complete embryo. The
first division is nearly or quite equal. In the second division one
of the cells divides nearly equally, the other very unequally. Thus
arises a 64/16-stage (Figs. 78-81) that is closely similar to the
228 Edmund B. Wilson.
corresponding group in a whole embryo (Cf. Fig. 5), though
these divisions are often (as in all the foregoing cases) less un-
equal than in a whole embryo, and their arrangement is fre-
quently modified by shifting of the cells. At the end of 24 hours
these groups produce closed ovoidal or irregular ectoblastic vesi-
cles that swim rather slowly by means of a tuft of cilia at one
end. In some cases these cilia seem to be borne by a single cell;
in others I am sure there are two of these cells (Fig. 82). These
cells are evidently secondary trochoblasts; and their presence is
entirely in agreement with the facts observed in the J^-macro-
mere-larva described under ( i ) , and with the fact that the nor-
mal larva clearly shows the derivation of at least one, and prob-
ably two, secondary trochoblasts from the second quartet.
There are two additional noteworthy points in these larvse.
One is the fact that, in addition to the one or two secondary
trochoblasts, some of them, at least, show one or two other small
patches of short and feeble cilia like those seen in the }i or tV
macromere larva. If my interpretation of these cells is correct,
this may be taken as evidence that the ventral ciliated tract arises
from derivatives of both the second and third quartets.
A noteworthy point in these embryos is the presence, in some
of them, of loose groups of rounded cells lying within the cavity
(Fig. 82). These cells, considerably smaller than the entoblast-
cells, are not improbably mesenchyme cells of the "larval mesen-
chyme" (paedomesoblast or ectomesobast) ; and it may here be
recalled that in Crepidiila, according to Conklin ('97), the ec-
tomesoblast is derived from the second quartet. Soon after the
stage described the embryos died and disintegrated without fur-
ther noticeable change,
4. Observations on isolated cells obtained from larvae that
have developed' continuously in calcium-free water.
Beyond the facts recorded above, I have not traced the devel-
opment of isolated blastomeres from the lower hemisphere. Sev-
eral times I succeeded by successive isolation in separating single
cells of the third and fourth quartets, and of the corresponding
Experimental Studies on Germinal Localization. 229
basals; but in every case the embryos became abnormal or died
without division. I therefore resorted to the method of allow-
ing the eggs to continue their development for 24 hours in the
calcium-free water, separating the cells from time to time by
directing a rather strong jet of water upon them by means
of a fine pipette. In this way the cells may be almost com-
pletely separated so as to produce what is in effect a pro-
gressive maceration of the larva without killing the cells. The
result is most striking. At the end of 24 hours the whole em-
bryo is disintegrated into its constituent cells, some of them lying
in small groups, but in favorable cases many are completely iso-
lated. The greater number of these cells are motionless and
perfectly spherical, of many different sizes, and still appear to be
living and in a healthy condition. Among these are swimming
with great vigor numbers of trochoblasts, singly, in pairs, or
sometimes in groups of four or three (Figs. 49-55). Measure-
ments of these trochoblasts show that in Patella they are of two
sizes, in Dentaliinn, as pointed out above, of three, the larger one
agreeing perfectly with the primary trochoblasts obtained by in-
dividual isolation, the smaller with the secondary trochoblasts.
Here and there can sometimes be seen a single apical cell, with
its chai'acteristic radiating sensory hairs.
Among the motionless rounded cells it is impossible to dis-
tinguish the different categories by their structure, since all have
the same form and all are filled with yolk spheres. In view of
the foregoing results, however, it can hardly be doubted that the
largest ones are isolated entoblast-cells. But the most interest-
ing cells are those which are not rounded but of a different form.
Two kinds of such cells can be distinguished, both in Patella and
in Dentaliinn, namely, spindle-shaped cells (Fig. 57), and branch-
ing mesenchyme-like cells (Fig. 56). The cells of both forms,
are relatively small, less heavily laden with yolk, and more trans-
parent than the others. /;/ all these respects these cells are closely
similar to the mesoblast-cells, as seen in total preparations or sec-
tions of the normal trochophore of the same age.
These facts must be interpreted with considerable reserve; for
it is well known that isolated cleavage-cells often become irregu-
230 Edmund B. JFilson.
lar or even amoeboid, and I have sometimes observed even trocho-
blasts of very irregular form. But this is not the case with most
of the isolated cells in Patella and Dentalhim, and I am inclined
to accept the probability that the cells in question may really be
mesenchyme- and perhaps actually muscle-cells, that have dif-
ferentiated in more or less complete isolation from their fel-
lows. If this be considered an improbable conclusion, it should
be recalled that a trochoblast is probably, to say the least, as
highly differentiated as a mesenchyme cell; yet it has been strictly
proved that such a cell may undergo its normal differentiation
and continue for a time to perform its normally coordinated ac-
tivities when completely isolated from the time of its formation.
Further research specifically directed to this point will, I believe,
give a positive result on this very interesting question.
5. Summary en isolated cells from the lower hemisphere.
The evidence derived from these cells is less detailed and com-
plete than that derived from the first quartet; but as far as it
goes gives the same general conclusion. The isolated >^-macro-
mere, yV-macromere or second quartet-cell segments as if form-
ing part of a whole embryo, and shows more clearly than do the
first quartet cells that not only the form, but also the rhythm of
cleavage is maintained (precisely as I showed in the nemertlne) ;
for In the cleavage of both the ^s- and the yV-macromere the
fourth quartet cell is the last to form. Only the embryos con-
taining derivatives of the second quartet produce secondary
trochoblasts, namely, those arising from the ^-macromere or the
second quartet-cell. While all the embryos close, only those gas-
trulate that contain the basal region (/'. e., the entoblast region).
All of the three types examined develop one or two feebly cil-
iated cells that probably represent cells of the pre-anal ventral
ciliated tract. Finally, there is some evidence, though only of an
inferential character, that Isolated mesoblast-cells may develop
into mesenchyme-cells, possibly into muscle-cells. We may, there-
fore conclude that, speaking broadly, the development of cells
of the lower hemisphere, like that of the upper, conforms to the
mosaic principle.
Experimental Studies on Germinal Localization. 231
C. DEVELOPMENT OF ISOLATED CELLS OF THE TWO- AND FOUR-
CELL STAGE.
I have purposely left to the last an account of the development
of the half or quarter embryos, since this is in Patella in some
respects the least satisfactory part of the work. This is owing
especially to the great susceptibility of the Yi and M-larvse, which
frequently break up into smaller fragments, go to pieces, or be-
come quite abnormal during the cleavage process, so that very
few satisfactory larvae were obtained. In Dentalium the results
are much better, since the blastomeres can be easily separated
without the use of the calcium-free water; but even here my
fixed material has proved insufficient for a satisfactory analysis
of the internal phenomena. For these reasons the following ob-
servations remain somewhat fragmentary and must await a sup-
plementary study in these or other forms.
I. Tlie partial cleavage in Dentalium.
In my preceding paper I have described in a general way the
development of isolated halves and quarters in Dentalium, and
will here only add some details regarding their mode of cleavage,
which are hardly more than a confirmation of Crampton's results
on Ilyanassa. As in Patella, these earlier blastomeres, like the
later ones, become perfectly spherical after isolation before cleav-
age begins; their characteristic partial cleavage must therefore
be due to internal factors and not to their shape.
The AB (anterior) half, which shows only an upper white
polar area (Fig. 83), segments equally into two, with no trace
of a polar lobe (Figs. 84-85) , and then forms by dexiotropic divi-
sions two slightly smaller micromeres of the first quartet, which
are composed entirely of white material (Fig. 86). The fol-
lowing division (Fig. 87) is like that occurring in half an egg,
the upper cells dividing slightly unequally to form below the
two primary trochoblasts ( la^ and ib") and above the two upper
cells (la^ and ib^). The lower cells in the meantime produce
the two cells of the second quartet (2a, 2b) in characteristic
232
Edmund B. JVilson.
fashion, these being likewise composed mainly or wholly of white
material (Fig. 87). Beyond this point (16/2 stage) I have not
followed the divisions.
Fig. IX.
Cleavage of Isolated Blastomercs in Dentalium.
83, 88. isolated AB and CD halves, before division; 83-87, cleavage of AB-
half; 84, 85, 4/2-cell stage, from, the side; 86, 8/2-cell stage, from the side; 87,
16/2-cell stage, from the side; 88-93, cleavage of CD-half; 88-90, first cleavage
(second polar lobe) and resulting 4/2-cell stage, from the side; 91, second
cleavage (third polar lobe) from the side; 92, resulting 8/2-cell stage, oblique
view, from the side and above; 93, 16/2-cell stage, with first somatoblast, 2d
(X) obliquely from the side; 94, 2-cell stage of C-fourth; 95-99. cleavage of
D-fourth; 95, 96, trefoils; 97, 98, 8/4-cell stages; 99, 16/4-cell stage (seen from
the inner side, so as to appear reversed).
Experimental Studies on Germinal Localization. 233
The CD half, which clearly shows both upper and lower white
polar areas (Fig. 88), forms a polar lobe from the lower white
area and passes through a trefoil stage, nearly similar to that of
a whole egg (Fig. 89). Measurements show, however, that the
polar lobe is always proportionately larger than in a normal tre-
foil, being often as large as in a whole egg, though sometimes
more or less reduced. The lobe subsequently fuses with the pos-
terior cell, D, producing a 4/2-cell stage closely similar to a nor-
mal 2-cell stage, except that the inequality is greater (Fig. 90).
At the second cleavage the polar lobe forms again (Fig. 91)
from the larger cell, D, which divides unequally and dexiotropl-
cally to form id, while the smaller cell, C, divides slightly un-
equally to form ic. As in the whole egg the polar lobe then
fuses with D, producing an 8/2-cell stage that is essentially like
the posterior half of a normal 8-cell stage (Fig. 92). The fol-
lowing cleavage is especially interesting, corresponding again
with the divisions in the posterior half of a whole egg (Fig. 93).
All the divisions are leiotropic. The two upper cells divide
slightly unequally to form the two primary trochoblasts ( ic^, id^)
and the slightly larger upper cells (ic^ and id^). From iC
arises the right cell (2c) of the second quartet, while from iD
arises the first somatoblast (2d), which Is as large as in a whole
embryo, and in like manner is mainly formed from the lower
white area in \D. The 16/2-cell stage has, therefore, exactly
the same origin and composition as the posterior half of a whole
egg, consisting of six white ectomeres (ic\ id\ ic", id", 2c and
2d), of which 2d is the largest, and of two macromeres (2C,
2D), which contain all of the pigment and show each an upper
white area (Fig. 93).
The history of isolated 54-blastomeres is entirely analogous.
The A, B or C quadrant typically divides slightly unequally, with-
out a polar lobe (Fig. 94), the smaller cell being composed of
white material and the pigment remaining in the larger; but cases
are not infrequent In which the division is nearly or quite equal.
The D-quadrant, on the other hand, forms a polar lobe, which,
as in a whole embryo, is typically much smaller than either the
first or the second (Figs. 95-96). The 2-cell stage is very un-
234 Edmund B. JFilson.
equal, the small cell (id) being pure white, the larger showing
both upper and lower polar areas (Figs. 97-98). At the sec-
ond division (virtual fourth) the second somatoblast (2d) forms
from the lower polar area, while the micromere produces the
single trochoblast (id"), and the corresponding larger upper cell
id^ (Fig. 99). Beyond this the cleavage was not followed in
detail. It is noteworthy that in the divisions both of the halves
and the fourths the normal inequality of the cells is frequently
reduced, and this is frequently expressed by a reduction in the size
of the polar lobe, both in the CD-half and the D-fourth — in-
deed, I have seen only a few D-fourths In which the polar lobe
was of full normal size, and the first division of these cells is fre-
quently irregular and abnormal. This is doubtless due in part
to shock, perhaps also to the effect of the calcium-free water when
this is used. Nevertheless, I think it probable that the effect may
also be due in part to disturbances in the arrangement of the
cytoplasmic materials, which may possibly be interpreted as a
regulative phenomena.
2. The partial cleavage in Patella.
The cleavage of isolated halves or fourths in Patella is entirely
in agreement with the foregoing in being strictly partial in charac-
ter, but I wish especially to emphasize the fact that, precisely as
I showed in Cerebratulus, two general types exist, in one of which
the cells so shift as to produce a closed embryo from the begin-
ning, while in the other the blastula is at first widely open on one
side. The point is important because the effect of the displace-
ment in the closed type is to shift the primary trochoblast-groups
more or less widely, sometimes to opposite sides of the embryo,
while in the open type they remain in nearly their normal position.
Nature thus performs an experiment in the displacement of the
blastomeres closely similar to those carried out by Fischel on the
ctenophore-egg, and a corresponding result is produced that
clearly shows the differentiation of the cells to be independent of
their position in the embryo.
Experimental Studies on Germinal Localization. 235
In the following description the typical case is described; at-
tention is again called to the fact that the unequal divisions are
frequently less unequal than in the normal and sometimes be-
come quite equal.
Fig. X.
Isolated 1/2-BIastovicres, Patella; x 250
100, successive camera outlines, showing relative sizes of whole egg, and
the y2, % and ^^-'blastomeres; loi, 4/2-ceIl stage; 102, 8/2-cell stage; 103,
16/2-cell stage, nearly typical open type, from above; 104, slightly less open
form from the side; 105. 106, 16/2-cell stages, closed type; 107, 32/2-cell stage,
closed type; 108, open blastula, from the open side.
236 Edmund B. Wilson.
The Isolated ^^-blastomere first divides equally (Fig. loi),
then unequally and dexiotropically, so as to form two slightly
smaller micromeres, displaced towards the left (Fig. 102). Up
to this point the embryo remains strictly a half of the correspond-
ing 8-cell stage. At the succeeding division the differences be-
tween the open and closed types become apparent. In the former
case, as shown in Figs. 103-104, the divisions may occur nearly
typically, though frequently the cells become more or less dis-
placed. In the second case the cells shift during the division, so
as to fit accurately together; and, as is clearly shown in Figs. 105,
106, the two trochoblasts (shaded) may thus come to lie on op-
posite sides of the embryo, as is also the case with the two cells
of the second quartet. I have not followed out in full the later
cleavage of these larvae, which are very puzzling in both cases,
owing to either the initial or subsequent shiftings. But so much
is certain, that from the open type may arise an open blastula
(Fig- 108), while the closed type remains closed; and the effects
are clearly shown in the resulting larvae. Fig. 107 shows a closed
32/2-cell stage, with the polar body in position. This embryo is
diflicult to analyze in detail, but very clearly shows two rosette-
cells above, with the corresponding i^- cells, and on each side
are two cells that doubtless represent the daughter-trochoblasts.
The eight cells of the lower hemisphere are more difficult to
identify, and the cell-connections shown in the figure are only in-
ferred.
The isolated M blastomere first divides unequally, forming a
micromere above, a macromere below (Fig. 109) ; and this is
followed by a leiotropic division identical with that occurring
in a single quadrant of a whole embryo (Fig. no). The 16/4-
cell stage then divides dexiotropically, producing a 32/4-cell stage
that may pretty accurately correspond with a single quadrant of
the normal 32-cell stage. Fig. in shows this stage of the same
individual shown in iio; this differs from a single quadrant of a
whole 32-cell stage only in the fact that 2^ has extended upwards
somewhat, so as to separate i^- from i"". Fig. 27 shows a 32/4-
cell stage that has separated somewhat (the cells are shown ex-
actly as they lay). Every cell is of correct proportion and po-
Experimental Studies on Germinal Localization. 237
sitlon, except that one of the groups of four has turned over
(doubtless during the removal with the pipette), so that the
upper group presents to view the outer, the lower one the inner,
side; while the rosette-cell is somewhat too large. Both the cases
figured represent the open type, which appears to be the rule in
the quarter cleavage.
3. The half and quarter larva in Patella.
The detailed study of the larvae derived from the 3^ or K-
blastomeres presents many practical difficulties. While the early
cleavage of these embryos is easily determined, the later stages
are exceedingly difficult to follow, owing to the shiftings of the
cells, the more or less complete closure of the embryos, and the
great number of defective or monstrous forms, I must admit that
as far. as Patella is concerned, and in some respects in Dentalium
also, the following account is far from satisfactory, especially in
regard to the most interesting question of all, that of the meso-
blast; but since I may have no opportunity to complete it at pres-
ent, I desire to record some observations which may at least open
the way for a more adequate study in the future.
The most essential point has been recorded in my preceding
paper, namely, that in Dentalium neither the >4- nor the 34-blas-
tomere is able to produce a perfect dwarf larva; and, further,
that the AB and the CD halves show definite and constant dif-
ferences, the former lacking both the post-trochal region and the
apical organ, while both these structures are present in the CD
larva. In like manner, among the quarter larvae only the D-
fourth produces these two structures, which are entirely lacking
in the A, B or C-fourths.
In Patella the corresponding comparison is far more difficult,
owing partly to the equal size of the halves or quadrants, but
more especially to the even greater difficulty of rearing the larvae,
which very frequently go to pieces during the late cleavage stages,
and invariably become irregular and monstrous during the sec-
ond day, and finally disintegrated before the larval characters be-
come clearly marked. My observations clearly show one point,
238 Edmund B. JVihon.
however, in the comparison of the two halves from the same egg,
in which Patella differs from Dentalium, namely, that both halves
develop an apical organ; and while this has not been directly
proved for the four quarters, the fact, described above, that any
isolated micromere of the first quartet may develop an apical or-
gan leaves practically no doubt that the same is true for the quar-
ters. The basis of the apical organ in Patella must, therefore,
be symmetrically divided by the first two cleavages, while it re-
mains undivided in Dentalium, remaining as a whole in the D-
quadrant. This is possibly correlated with the fact that the apical
rosette, formed at the fifth cleavage of Patella nnd Trochiis, fails
to appear in Dentalium, where the i^-^ cells are as large as their
sister-cells i^".
In Dentalium, as described in my first paper, the larvae in-
variably close sooner or later, and the prototroch, in most if not
all cases, closes also to form a complete belt encircling the body.
In Patella, however, this is not always the case; and frequently
the >4-larv2e of 24 hours show the prototroch as an area of char-
acteristic trochoblasts extending around one side only, terminating
abruptly to leave a space occupied by much smaller non-ciliated
cells. ^ (Figs. 117, 118). In other half larvae the prototroch ap-
pears as a complete belt, in still others as a more or less irregular
or interrupted structure.
An examination of the earlier ciliate^l stages, combined with
the results obtained with isolated trochoblasts, gives the obvious
explanation of these differences. In those of the earlier larva?
(8-10 hours) that are still open on one side (and hence must
have been derived from the open type of cleavage) two adjoin-
ing groups of trochoblasts are found on one side, leaving a space
on the opposite side free from trochoblasts (Fig. 114). In the
closed embryos, on the other hand, two corresponding trocho-
blast groups are formed on opposite sides of the embryo, with
only rather narrow gaps between them (Figs. 11 3- 11 5- 116).
Both these types may be represented in twins from the same egg,
a case which I am fortunately able to show by Figs. 113 and 114
iln agreement with Crampton's observation that the i^-larvse of Ilyanassa
form "a partial circle of cilia" (96, p. 9).
Experimental Studies on Germinal Localization. 239
(from acetic-glycerine preparations). Of these twins one (Fig.
113) is closed and shows the gastrulation well advanced (the su-
perficial ectoblast-cells of the lower hemisphere are only in part
shown). This larva shows very clearly the two groups of pri-
mary trochoblasts on opposite sides of the egg, at t and t, with
at least two secondary trochoblasts lying between them on each
side, the general arrangement being similar to that shown in Fig.
116.^ The twin larva (Fig. 114) is still widely open on one
side; and while the small ectoblast cells have closed in to fill the
gap above, the two primary trochoblast groups lie at one side,
leaving a wide gap occupied by smaller cells. Fig. 115 is a J-^-
larva, which, though somewhat asymmetrical, is clearly of the
closed type (the superficial post-trochal ectoblast-cells are shown
only in optical sections at the sides) ; and here, too, the primary
trochoblast groups lie on opposite sides of the larva.
It is hardly possible to doubt that these two types of larvae
arise from the open and closed types of cleavage, the trochoblasts
having undergone their normal differentiation whether displaced
or not. This has not been strictly proved by isolation experi-
ments ; but in view of the demonstrated fact that the trochoblasts
differentiate typically If wholly separated from their fellows,
there can be no doubt, I think, of the interpretation offered. It
is quite clear that in this case the prospective value of the cell is
not a function of its position, but is dependent on its internal or-
ganization irrespective of its position. This result is exactly anal-
ogous to those obtained by Fischel ('98) by displacing the micro-
meres of the ctenophore egg — an operation that, as he shows in
the most convincing manner, leads to a correspoiiding displace-
ment of the rows of swimming plates in the larva.
With the M-larvas in Patella I had little success, since they al-
most invariably broke apart into smaller fragments. A very few
nearly complete larvae of 24 hours were, however, obtained, one
of which is shown in Fig. 112. This larva shows a central mass
of rather large rounded cells completely surrounded by ectoblast,
and has evidently gastrulated. At one side is a very distinct group
iThe latter larva apparently shows five primary trochoblasts on one side
— a fact for which I cannot account.
240
Edmund B. Ifllson.
Fig. XI.
Experimental Studies on Germinal Localization. 241
1/4 and 1/2 Larva. Patella; x 250.
109-111, cleavage of isolated 1/4, open type; 112, resulting larva, 24 hours;
113, 114, tw^in larvae. 9 hours, 113 of the closed type, 114 of the open; 115,
closed J^-larva, 9 hours; 116, closed J/^-larva, 11 hours, apical view; 117, 118,
products of open type, 24 hours.
242 Edmund B. Wilson.
of six trochoblasts. Four of these, bearing powerful rows of
cilia, are evidently primary trochoblasts; two, lying in front of the
last, are much smaller, and are probably secondary trochoblasts
from the first quartet. Those of the second quartet have either
failed to develop or have broken away from their connections (as
very often occurs with all the trochoblasts owing to their ac-
tivity). No apical organ was seen in this larva; but I observed
an apical organ in several less normally developed individuals,
and since the apical organ constantly appears in the J/g-micromere
larva, there can be no doubt that it may appear also in the J4-
larvae.
Perhaps the most interesting question presented by these larvas
is whether the AB and the CD half-larvae differ in respect to the
mesoblast; for if the mosaic principle holds for this structure,
one should expect to find coelomesoblast only in the CD half.
For the present I can give no certain answer to this question, fur-
ther than to state that in DentaJium the two larvae certainly dif-
fer to some extent in respect to the mesoblast, and there is possibly
some reason to conclude that they do also in Patella.
In the latter form some of the larvae show a large rounded cell
in the upper region of the central mass (the dotted outline in
Figs. 113 and 115) which does not appear in others; and this
difference distinctly appears between the two twin larvae shown
in Figs. 113 and 114. This cell is possibly the primary meso-
blast, 4d; but it may also represent the large rounded cell which
I have considered to be 4D In the normal larva (Figs. 15-16).
This evidence, unsatisfactory as It is. Is mentioned as an Indica-
tion that the Internal structure of the two half-larvae shows dis-
tinct differences In Patella. In Dentalium the evidence is some-
what better, but still far from adequate, owing to paucity of ma-
terial and the confused appearance of the inner cell-mass, as seen
either In total preparation or In sections. Sections of the CD
larvae nevertheless show groups of smaller and irregular cells
lying between the large entoblast-cells and the ectoblast, and
there is little doubt that these represent in part the coelomeso-
blast. Sections of the AB larvae are in general closely similar to
those of the lobeless larvae described in my preceding paper, but
Experimental Studies on Germinal Localization. 243
in some cases distinctly show a few small cells lying between the
gut and the ectoblast. The only conclusion that I am justified in
drawing is that the mesoblast cells are more numerous in the CD
larva than in the AB, and additional material will be necessary
to determine the point. When, however, we consider the evi-
dence, not entirely conclusive but still fairly definite, given in my
preceding paper, that the material of the polar lobe (which
passes only into the CD half) is necessary for the production of
the coelomesoblast, I think It may be concluded with some prob-
ability that the mesoblast-cells (If they be such) found in the AB
half represent a portion of the larval mesoblast or ectomesoblast,
and that the coelomesoblast Is represented only in the CD half.
I hope in the near future to obtain additional material that may
afford a more definite conclusion.
IV. SUMMARY.
{This Applies Primarily to Patella.)
1. Isolated blastomeres of any stage segment essentially in
the same manner as if still forming part of a complete embryo,
with a tendency, however, for all unequal divisions to be less un-
equal than In the normal. The partial form of cleavage is fre-
quently masked by shifting of the cells.
2. All of the partial embryos, if of sufficient size, tend to close
to form morula- or blastula-like structures ; but these only gastru-
late if they contain entoblast material from the basal cells. Apart
from such closure all of the cells, and their products, as far as
examined, differentiate typically, regardless of their relative po-
sition or of complete Isolation from their fellows.
3. Isolated >^-mIcromeres produce pyriform larvs, bearing
at one end an apical organ, at the other a group of four primary
and two secondary trochoblasts. In Dentalium the apical organ
is produced only by the posterior micromere, id.
4. Isolated primary trochoblasts (i') divide twice and pro-
duce four typical ciliated prototrochal cells. Isolated first prod-
ucts of the primary trochoblasts (i"S i"'") divide once and pro-
244 Ed?fiiind B. inisoH.
duce a pair of typical prototrochal cells. Isolated second prod-
ucts (i"-^"S i"'^", etc.) undergo no further division, but different-
iate singly into typical prototrochal cells.
5. Isolated i^ cells produce embryos bearing at the anterior
end an apical organ, at the other two secondary trochoblasts.
6. Isolated products of the i^ cells differentiate into typical
sensory cells of the apical organ, into secondary trochoblasts, and
into less differentiated ectoblast cells.
7. Isolated >^-macromeres produce closed embryos that gas-
trulate and bear at one end one or two secondary trochoblasts,
and at some other point a small group of feebly ciliated cells,
probably representing the pre-anal ciliated cells of the normal
larva.
8. Isolated yV-macromeres produce closed embryos that
gastrulate, bear no trochoblasts, but have feebly ciliated cells, as
in 7.
9. Isolated cells of the second quartet produce closed ecto-
blastic embryos bearing one or two secondary trochoblasts, and
one or two feebly ciliated cells, that probably also represent part
of the pre-anal tract. These embryos do not gastrulate, but may
form mesenchyme-like cells.
10. Isolated >4-blastomeres produce embryos that gastrulate,
produce four primary trochoblasts, at least two secondary ones,
and an apical organ.
11. Isolated ^-blastomeres produce, in Patella, larvae bear-
ing an apical organ, and a prototroch, either open or closed, ac-
cording to the mode of cleavage. In DentaUiim only the CD
half produces an apical organ and a post-trochal region, and
probably only this half produces caelomesoblast.
12. The development of both Patella and Dentaliuvi is essen-
tially a mosaic-work of self-differentiating cells.
v. DISCUSSION OF RESULTS.
The experimental results brought forward in this paper and the
preceding one seem to me to establish definitely the principle of
mosaic development in the case of the mollusks Dentalium and Pa-
Experimental Studies on Germinal Localization. 245
tella, and to place the study of cell-lineage on a new and firmer
basis. Clearly as the exquisite adjustment between the cleavage-
process and the operations of morphogenesis has been revealed
by the descriptive-comparative study of cell-lineage, it appears in
still stronger relief in the light of the experimental proof that
the cleavage-pattern, as a whole and in detail, is the visible ex-
presson of an actual distribution of specific morphogenic factors
among the cells.
Although Crampton's initial, and hitherto almost unique, ex-
periments on this type of development had led' to the expectation
that some evidence of cell-specification and self-differentiation
would be found, I confess that I was not prepared to find that evi-
dence so circumstantial and consistent. The evidence in Patella
that the cleavage-cells are definitely specified from the time of
their first formation, and. that they undergo self-differentiation
without essential modification through their relation to the other
cells, is demonstrative in the case of the cells of the first quartet, at
least as far as the i6-cell stage, as shown by the development of
isolated entire micromeres at the 8-cell stage, and of their prod-
ucts i^ and i^ at the i6-cell stage. It is no less demonstrative in
the case of the products of the primary trochoblasts isolated at
the 32- and 64-cell stages; and inasmuch as cells of the apical or-
gan derived from the i^-^ cells, and secondary trochoblasts derived
from the i^- cells, also differentiate typically when the isolated
micromere is allowed to segment continuously in the calcium-
free water, and the cells are separated more or less completely
after every division, the conclusion is unavoidable that these cells,
too, may undergo their characteristic development in complete
isolation from their fellows. Less detailed, but hardly less con-
vincing, is the evidence derived from the isolated ]4. basal, the
tV basal, or the isolated second quartet-cell ; and it can hardly
be doubted that the individual products of these respective cells
are, like those of the first quartet, definitely specified in greater
or less degree.
The general conclusion thus reached in the case of Patella is
sustained by the development of larger masses of cells derived
from the earlier stages both of Patella and of Dentaliiim. The
246 Edmund B. JFilson.
entire first quartet of Patella, when isolated, produces a mass of
ectoblast-cells, which, though It closes, does not gastrulate, but
undergoes essentially the same differentiation as if It formed the
upper hemisphere of a complete larva. The isolated quadrant
of a 4-cell stage gastrulates, produces a group of trochoblasts and
an apical organ, the latter structure appearing apparently in any
of the quadrants in Patella, while in Dentaliuin It Is restricted to
the D-quadrant. In Dentaliiim, further, only the D-quadrant
produces a post-trochal region, which Is due to the fact that this
quadrant alone contains the material of the lower polar area from
which arises the somatoblasts. Finally, the two halves of the 2-
cell stage gastrulate, but (at least in Dentalhim) differ widely in
their later development. Both in Dentalium and in Patella the
half-embryo forms a prototroch, which In the former seems al-
ways to close to form a complete ring, but In Patella frequently
remains open at one side, forming a half ring. In Patella both
halves form an apical organ; in Dentalium only the CD-half.
In Dentalium, finally, only the CD-half forms a post-trochal
region, for the same reason as in case of the D-quadrant. It is
probable, further, that only the CD-half and the D-quadrant pro-
duce coelomesoblast. This conclusion has not thus far been sat-
isfactorily established by direct examination of the half-embryos,
but is Indirectly rendered very probable through the observations
on the lobeless larvae recorded in my preceding paper.
The foregoing facts constitute a strong body of prima facie
evidence that the entire cleavage-pattern in the molluscan egg
represents (with certain reservations considered beyond) a mosaic-
work of self-differentiating cells, exactly in the sense of Roux's
general conception.^ The proof is indeed entirely complete In
^Here, and in all that follows, I exclude from that conception the hypothe-
sis of qualitative nuclear division. It should be borne in mind that Roux
himself expressly stated as early as 1893 that this hypothesis did not form a
necessary part of his conception. "Die beiden AmuiJimefi" (nuclear idioplasm,
distribution by qualitative division) "sind jedoch nicht unerlasslich noth-
wendige Glieder mciner in iliren wesentlicJten Theilen experimentell erwles-
enen Auifassung;" ('93.2, p. 874, Italics in the original). This fact has been
ignored by many of Roux's critics, in spite of the fact that some of his most
important contributions to experimental embryology have been specifically
Experiniental Studies on Germinal Localization. 247
the case of only a few kinds of cells. It is evident that the limi-
tations of potency vary in different cells — the i^ cells, for exam-
ple, contain more complex potencies than the 1= — and It is quite
possible that dependent or correlative differentiation may play a
larger part in the development than my experiments have thus
far shown. But the proof by experiment of definite specification
and self-differentiation in only a few categories of the early cleav-
age-cells establishes a principle that is to be reckoned with as a
most important factor in the whole problem of embryonic differ-
entiation. If, in considering some aspects of this problem, I
again take up a discussion that has been so prolonged, it is because
I believe that the importance of the principle of mosaic-develop-
ment, and of the nearly related one of specific formative or deter-
mining stuffs, has received insufficient recognition by many embry-
ologists, and by some has been prematurely discredited. That a
reaction is well under way will be evident to everv reader of Fis-
chel's ('03) excellent recent discussion of development and dif-
ferentiation, the essential conclusions of which agree closely with
those earlier stated in a brief form in my paper on cleavage and
mosaic-work ('96, appended to Crampton's paper), and more
fully considered in the discussion of my nemertine results ('"03) ;
the agreement with the conclusion reached in this and the preced-
ing paper is still closer. In full harmony with the same general
conception are the important cytological results of Lillie ('99,
'01) and Conklin ('98, '99, '02) cited in the two preceding
papers.
The long continued discussion of the mosaic-theory of develop-
ment that followed its first definite formulation by Roux in 1888,
in the course of which Roux so ably defended his position, has
been greatly prejudiced by the fact that the experimental analysis
of cleavage was at first confined to the so-called "indeterminate"
types of cleavage, such as those of the echinoderm, the medusa
and Amphioxus ( it may for the time be left an open question
whether that of the frog should be placed in the same class) . The
directed towards the role played in development by the segregation and locali-
zation of cytoplasmic materials. Roux himself ('03) has now abandoned the
second of these "Annahmen" (qualitative nuclear division).
248 Edmund B. PFilson.
earlier results formulated for these types seemed wholly subver-
sive of the mosaic-principle and of the nearly related one of ger-
minal prelocalization. In the sea-urchin egg, the first in which
an isolated blastomere was shown to be capable of producing a
complete dwarf larva, the experiments seemed at first to show
that the blastomeres are composed of indifferent material, so that,
to cite an early statement of Driesch's, "Durch die Theilung bei
der Furchung vollig gleichwerthige, zu allem fahige (indiffer-
ente) Stiicke geschaffen werden" ('92.2, p. 36), forming a ma-
terial "welches man in beliebiger Weise, wie einen Haufen Kug-
eln durch einander werfen kann, ohne dass seine normale Ent-
wicklungsfahigkeit darunter im mindesten leidet" {op. cit. p.
25). Despite the fact that Driesch early recognized that the
cytoplasm of this egg is not isotropic, he considered that his ex-
periments definitely overthrew His's principle of "Organbildende
Keimbezirke," or at least deprived it of all casual significance in
the echinoderm egg ('92.1, p. 178, '93, p. 243). Specification of
the early cleavage cells was denied ('92.2, p. 22), as was also the
principle of mosaic development as applied to this egg (1. c, p.
36). Again, in the paper of 1899, on "Die Localization mor-
phogenetischer Vorgange," where Driesch's theory of vitalism is
first definitely formulated, the ground is taken that "Darin nam-
lich, das jeder beliebige Eitheil, sowie das Eiganze in beliebiger
Verlagerung eine ganze Larve llefern, also jede 'Organization,'
die postulierte Vorbedingung zum Eintritt lokalisirten speci-
fischen Geschehens iiberhaupt, nach Storung, regulatorisch wieder
herzustellen vermag, kommt zum Ausdruck, das eben die 'Struck-
tur' des Eies nicht aus mannigfach-verschiedenen Elementen in
irgendwie typisch-specifischer Lagerung aufgebaut sein konne, die
etwa zu den spateren Differenzierungen in irgend einer Bezie-
hung stiinde" {op. cit., p. 43). It is hardly necessary to point
out how greatly all this has been changed by Boveri's discovery
of the fact that the sea-urchin egg does in point of fact contain
"mannigfach-verschiedene Elemente" disposed in a "typisch-
specifischer Lagerung," which are proved by the experiments of
both Boveri and Driesch to stand in definite relation to the sub-
sequent process of cleavage and differentiation. The relation of
Experimental Studies on Germinal Localization. 249
these facts to those determined in the nemertine and mollusk is
considered beyond. I will at this point only express my agree-
ment with the conclusion of Fischel, that "Sowohl bei den Echin-
odermen, wie bei den speciell so-genannten 'Mosaik-eier' erfolgt
die normale Entwickelung im Wesentlichen als Mosaik-arbeit"
(Fischel, '03, p. 728). I am convinced that had the experimental
analysis of cleavage been first undertaken in the case of such a
determinate type as that of the gasteropod or annelid, and had
Roux not handicapped his theory with a purely speculative hy-
pothesis of differentiation, which proved to be untenable, the
whole discussion would have taken a very different course ; and I
believe it would from the first have been recognized that the
mosaic-principle holds true in greater or less degree for every type
of development, not excepting the most "indeterminate" forms
of cleavage.
My experiments on the unsegmented egg of Dentalinm have
added fresh proof to that obtained by Fischel and his predeces-
sors in the ctenophore, that the cleavage-mosaic is a mosaic of
specifically different cytoplasmic materials, in which are somehow
involved corresponding morphogenic factors. In this egg, con-
firming and extending the earlier work of Crampton on the gas-
teropod-egg, I was able to show even more definitely than has been
done in the ctenophore-egg the existence in the unsegmented egg
of prelocalized cytoplasmic regions, distinguishable by the eye,
that stand in some necessary relation to the formation of the
structures to which they give rise in the normal development; for
if one of these areas (the lower polar area) be removed, the
structures to which it is destined to give rise fail to develop, while
if this area remains while other areas are removed the structures
in question make their appearance. His's principle of "Organ-
bildende Keimbezirke," which he developed in a purely descrip-
tive sense, is thus shown to have a true causal significance. Since,
further, this area contains no nucleus, the conclusion is unavoid-
able that here, as in the ctenophore — and as we are now able to
say, even in the echinoderm — there is a localized distribution to
some extent, of the factors both of cleavage and of differentia-
tion in the cytoplasm before development begins. A no less sig-
250 Edmund B. Wilson.
nificant fact, proved by these experiments in connection with ob-
servations by many observers of normal cell-lineage, is that the
germ regions prelocalized in the unsegmented egg are, at least
in the case of certain cells, accurately marked off by the subsequent
lines of cleavage. This is shown with great clearness by the his-
tory of the lower polar area in DentaUum (or the analogous
lower green area in Myzostoma or the lower polar ring in Rhyn-
chelmis), which, although it lies primarily at the center of the
lower hemisphere, is not bisected by the first or the second vertical
cleavage (despite the fact that both the cleavage furrows first lie
exactly in the egg-axis), but is moved to one side so as to pass
bodily into one of the cells at each division. Here is an adjust-
ment, of admirable accuracy, by which a specific prelocalized area
is handed on from cell to cell, to be finally assigned to its proper
positon in the cell-mosaic; and if such be the case with one such
specific germ area, we have strong ground to infer that it is also
so with others. In such cases as these it is evident that the fac-
tors of cleavage run so accurately parallel to those of differentia-
tion that they must be referred to a common determining cause,
and may be treated as practically identical.
But even in cases where the adjustment is less evident, or less
precise (as appears to be the case, for example, in the third cleav-
age of the echinoderm egg, considered beyond) we shall not, I
believe, escape the conclusion that cleavage involves a definite
distribution of specific morphogenic factors among the cleavage
cells. The facts, proved by my experiments, that these factors
may be completely separated and isolated by cell-division, and
may retain their specific character after isolation of the cells, are
only intelligible under the assumption that they are somehow in-
volved in specific materials or stuffs which differ in a definite way
and have a specific topographical grouping in the undivided egg.
This conclusion is not to be avoided by assuming that the visible
cytoplasmic differences are only an accompaniment or consequence
of an invisible ulterior structure or organization. Admitting this,
and even admitting, for the sake of argument, that the localized
cytoplasmic factors are not definitely characterized chemical ma-
terials, but only local physical or structural conditions, established
Experimental Studies on Germinal Localization. 251
by virtue of the relation of the particular cytoplasmic regions to
the egg as a whole : the fact remains that the cytoplasmic sub-
stance possesses different specific qualities in different regions, and
that these differences persist after the regions have lost their re-
lation to the whole. Only by a play upon words, therefore, can
the conclusion be escaped that the cytoplasmic regions consist of
specifically different substances having a definite morphogenic
v^alue. The question whether these substances are to be consid-
ered as preformed building materials, or rather as specific deter-
mining materials^ (such, for example, as enzymes) is a second-
ary one, on which I do not propose to enter here. Holding both
these possibilities in view, I can see no valid objection to the frank
adoption, in a provisional sense, of the term "formative stuffs"
in the general spirit of the Bonnet-Sachs hypothesis, awaiting fu-
ture research, to determine what is their mode of action. We
must, therefore, conclude that the cleavage-pattern represents lit-
erally a mosaic-work of such formative stuffs that have been dis-
tributed by the cleavage process, and that the specification of the
cells is within certain limits determined by their inclusion of these
stuffs. If for the conceptions of qualitative and quantitative
nuclear divisions we substitute those of qualitative and quantitative
cytoplasmic divisions, a very large part of the development that
Roux has given to his theory in his long controversy with Driesch,
O. Hertwig and other writers, is, I believe, entirely valid. I
shall not undertake to go over the whole of this ground again,
but will apply these terms to a specific interpretation of certain
facts.
In my preceding paper I have suggested that the difference in
behavior between isolated blastomeres of different forms is pri-
marily due to differences in the initial form and degree of segre-
gation. The possibility of the production of a perfect larva from
either half of any quarter in the egg of Amphioxus, Echinus or
Cerebratulus is given by the symmetrical or purely quantitative
distribution of materials by the first and second cleavages.' In
Dentalium both halves are not able to produce perfect larvae,
'^Cf. Morgan, Regeneration, p. 89.
Kf. Fischel, 03, p. 7^2>-
252 Edmund B. JFUsoti.
owing to an asymmetrical distribution of material, the cleavage
being visibly qualitative from the beginning; and it is impor-
tant to note that this asymmetry of distribution is effected by the
process of cleavage itself, since the primary segregation-pattern,
as far as can be determined, is symmetrical with respect to the
axis. In the nemertine or sea urchin the first qualitative division
occurs at the third cleavage (which is also qualitative in the mol-
lusk)^ which for the first time separates ectoblast-stuff from en-
toblast-stuff. A comparison of the different forms indicates, how-
ever, that in respect to this cleavage they differ somewhat in de-
gree. In Patella the cells of the first quartet are from the first
completely specified, whether as a group or individually, and pro-
duce purely ectoblastic embryos that never show any tendency to
gastrulate. The same is true in Cerebratiilus, according to the
recent work of Zeleny ('04), which shows that if in the 8-cell
stage the upper and lower quartets be separated along the line
of the third cleavage, both quartets develop into closed swimming
embryos, but the upper one (although the larger) does not gas-
trulate, though it produces an apical organ, while the lower one
gastrulates but produces no apical organ." In the; sea urchin,
however, a small proportion of the upper cells (20-25%) are able
to gastrulate (Driesch, '00. '02.2) ; and this can only mean that
the third cleavage is less strictly qualitative or not invariably so.
lAs was also assumed by Samassa in the case of the frog's egg, "Diese ver-
schiedenen Entwickelungsbedingungen konnen aber nur in verschieden Sub-
stanzen Hegen, die bei der qualitativ ungleiche Theilung der dritten Furchung
den beiden Zellarten zufallen" ('96, p. 386).
2In the light of Zeleny's observations on the 8-cell stage, and in spite
of his apparent confirmation of my own preceding results on the blastula
stage, it seems to me very probable that the gastrulas I obtained from upper
fragments of blastulas in Ccrebratuhis were obtained by slightly oblique sec-
tion, so that a small group of entoblast cells were included in the upper frag-
ment. I have since observed that the entoblast-plate extends nearly to the
equator of the egg, so that even a slight obliquity in the plane of section
might give a misleading result. A similar interpretation not improbably may
apply to the upper fragments of echinoderm blastulae, cut in two just before
gastrulation, which were observed by Driesch ('95) to gastrulate; but these
were cut eii masse without individual orientation, and the experiments evi-
dently do not exclude the possibilty that the upper fragments may have con-
tained a part of the entoblast-region. A repetition of this work on both forms
by means of individual operation is much to be desired.
Experimental Studies on Germinal Localization. 253
Boverl ('oi.i) has In fact shown experimentally that the ability
to gastrulate depends on the presence of a certain amount of the
pigment-band that approximately coincides with the entoblast-
zone; and the variation in this regard is explicable under the as-
sumption either of a varying position of the third cleavage-plane
with respect to the entoblast-zone or of a variation in the de-
gree of concentration of the entoblast-stuff. While Boveri adopts
provisionally the former of these alternatives, he also suggests
that the formation of entoblast and mesenchyme is not absolutely
predetermined in the plasma, but occurs at the "most vegetative"
point, which is the lower pole. Driesch ('02.2) adds the sug-
gestion that the frequent failure of the animal larvae to gastrulate
may be due, not to absolute lack of "vegetativlty" ("Um einen
nicht sehr schonen aber deutllchen Ausdruck zu gebrauchen") ,
but to its Insufficient degree; and he has recently shown ('03) by
an experiment of admirable ingenuity that artificial displacement
of the third cleavage-furrow towards the vegetative pole causes
a large Increase in the proportion of gastrulas produced by the
isolated upper cells. This interpretation becomes perfectly intel-
ligible If stated frankly in terms of the formative stuff hypothesis;
and it harmonizes with my conclusion regarding the Dentalium
egg that the influence of the specific stuffs is within certain limits
qualitative rather than quantitative, which was based on the fact
that If the upper part of the egg be cut away, leaving the whole
of the lower pole area the polar lobe typically is reduced to the
correct proportional volume, and the resulting larva has a post-
trochal region of the proper size. This conclusion Is in agree-
ment with that of both Boverl and Driesch, that the plasma struc-
ture plays "nur eine determlnirende, kelne fixierende Rolle"
(Driesch, '02.2, p. 522).
In the ctenophore It appears from Fischel's observations ('98)
that the first qualitative division is the fourth, which first sep-
arates ectoblastic micromere material from the entoblast con-
taining basal cells. In the whole series up to this point we have
been considering a segregation that in its initial form is vertical
and symmetrical about the axis, though in the moUusk and an-
nelid it becomes asymmetrical in the course of the first cleavage
254 Edmund B. Jf'Uson.
(prov^ed by direct observation in Clepsine, Rhynchelmis, Myzos-
toma and DentaUiim) . In the medusae it appears from the work,
of Zoja ('95) and Maas ('01), that the primary segregation is
not visibly polarized, but concentric; and qualitative division (de-
lamlnation) does not take place before the fifth cleavage.^ It
seems evident that in these differences of form and degree of the
initial segregation pattern we find the leading principle for an
explanation of the differences in mode of development shown by
isolated blastomeres of the various forms ; though as pointed out
beyond, a complete explanation Is not given by these facts alone.
To consider one or two more detailed Instances, the first division
of the first quartet cells in Patella Is qualitative, not merely In a
descriptive or prospective sense, but actually, as Is proved by ex-
periment. By the same standard, the second division of these
cells Is qualitative In the upper cell (i^), but only quantitative
in the lower one (i"). Such facts as these give the strongest
ground for the conclusion that all the divisions that would be
considered as qualitative or quantitative from thfc point of view
of descriptive cell-lineage, are really such as regards the inherent
factors of differentiation. The descriptive and comparative study
of cell-lineage represents something more, therefore, than a mere
enumeration of successive cell divisions and their geometrical re-
lations, and has the value of a direct examination of the normal
morphogenic process.
These conclusions may appear to conflict with the view that
has been frequently urged by embryologists In late years that the
organism develops essentially as a whole, and that cell-formation
plays but a subordinate part in the morphogenic process. The
conflict is, I believe, only a seeming one. Roux has repeatedly
pointed out that the mosaic-principle Is by no means Irreconcll-
'^Cf. Maas: "Wenn die (cytoplasmic) Substanzen in alien Radien, resp.
Axen, gleichmassig verteilt sind, wie bei den kugeligen Eiern von Medusen.
dann und nur dann hat man ein wirklich isotropcs Ei; in anderen Fallen, wo ein
polare Anordnung festgestellt werden kann, wie bei den echinodermen. bes-
teht die Isotropic niir urn eine hcstimmtc Axe; in weiteren Fallen kommt durch
Gestalt des Eies, wie bei den Cephalopoden, oder durch Lagerung der Substan-
zen wie bei den Ampbibien, eine bilateral-symmetrisch Anordnung zu Stande
und in anderen Fallen ist diese Anordnung noch etwas komplizierter (siehe
Z. B. Mycosotoina)." ('03, p. 72.)
Experimental Studies on Germinal Localization. 255
able with such a view, and he has steadily maintained the position
that the development of every animal presents a combination of
self-differentiation and correlative or dependent differentiation,
the relation between which varies more or less widely in different
cases/ Only the most thorough experimental study can deter-
mine what this relation is in any individual case. The hypothesis
of qualitative nuclear division is no doubt responsible for the dis-
favor with which the conception of self-differentiation was re-
ceived by many writers, who either relegated it to a position of
quite minor importance or rejected it in toto, adopting only hy-
potheses of correlative differentiation, or advocating a less clearJv
defined conception of the "organism as a whole," to which the
differentiation of the cells was assumed to be subject. O. Hert-
wig's theory of cellular interaction is a clearly formulated conr
ception of this type, cleavage being assumed to be merely a multi-
plicative process, producing qualitatively equivalent blastomeres
that differentiate by cellular interaction {e. g., '92, p. 481, '93, p.
793). "Die Zellen determiniren sich zu ihrer spateren F^igenart
nicht selbst, sondern werden nach Gesetzen, die sich aus dem Zu-
sammenwirken aller Zellen auf den jeweiligen Entwicklungsstu-
fen des Gesammtorganismus ergehen, determinirt" ('98, p. 144).
Cell-lineage, therefore, has only an incidental significance, arising
from the continuity of development, which involves the deriva-
tion of each part from an earlier group of cells, itself in turn the
product of a still earlier one ('92, p. 479). Whitman ('93), in
his singularly thoughtful and suggestive essay on the "Inadequacy
of the Cell Theory of Development," while repudiating the
theory of cellular interaction as such, urged with great force the
subordination of the Individual cells in development to the or-
ganization of the embryo as a whole — a conception which, though
differing widely in its form of expression, has, I think, much in
common with Driesch's theory. The same general view is very
specifically interpreted in Child's valuable descriptive paper on
the cell-lineage of Arenicola ('00), in such statements as the fol-
^C/^. Roux, '88, p. 455, and elsewhere. Heider, in his suggestive survey
of the determination-problem ("00), and Fischel, in his more recent discussion
(03), takes the same ground. See also Korschelt and Heider, '02.
256 Edmund B. JVilson.
lowing: "The differences between the quiescent trochoblasts and
the other cells does not necessarily signify that the former con-
tain a special substance which makes them distinctively trocho-
blasts from the time of their formation. Of course, at some time
they do become distinctly trochoblasts, but simply because of their
relation to the whole" (p. 664). I have cited this particular
case, since It Is precisely In the case of the trochoblasts that experi-
ment most Indubitably demonstrates self-differentiation Inde-
pendently of the position of the cell In the embryo. To
cite a more general statement, "The material separated
as the result of precocious segregation may, I believe, be
perfectly Indifferent material except as regards position" (p.
682). "Certain amounts, rather than certain kinds of material,
are stored up in certain cells just where they will be in position
to produce by coordinated action the 'desired result' " (p. 679.
Italics mine). I must own to some difficulty In grasping the con-
ception of a "precocious segregation of perfectly Indifferent ma-
terial"; but, this aside, it is clear that differentiation Is considered
to be effected, not through the specific and Inherent nature of the
substance of the individual cell, but through correlative action,
the hypothesis even being advanced that an important function
of the spiral type of cleavage Is to provide for this purpose the
most direct and intimate possible communication between the
blastomeres (p. 658, etc.).
Llllle, who has contributed such valuable observations on the
progressive segregation and organization of the egg-substance,
and has recognized In the fullest degree the complexity of that
organization and the Importance of precocious segregation, never-
theless casts considerable doubt on the conception of prelocalized
germ areas ('01, p. 269), and feels constrained to take the po-
sition "That the entire organism in every stage of its develop-
ment exercises a formative influence on all its parts, appears to
me an absolutely necessary hypothesis" ('01, p. 273). I do not
doubt, as will appear beyond, that this position, with proper quali-
fication, is well grounded; but do not the phenomena of self-dif-
ferentiation, as shown in the Independence of grafts or in the
typical differentiation of Isolated blastomeres in Patella, show that
Experimental Studies on Germinal Localization. 257
as thus stated the conclusion is somewhat misleading? I cannot
think otherwise. The fundamental conception that the develop-
ment of every part is conditioned by that of the organism as a
whole is one that every embryologist must accept; but it seems
to me that Driesch, whom no one will consider a partisan of the
mosaic-theory, expresses the truth when he says (Analytische
Theorie, p. 94), "In diesem Sinne ist nun Selbst-differenzieruno-
einmal angelegter Teile ein wesentliches Merkmal der Ontoge-
nese; ja sie ist in Hinsicht auf die spatere Einheitlichkiet und das
physiologische Zusammenwirken unaphangig entwickelter Gebilde
von einem ganz eigenartigen Interesse" (Italics original). The
fact must be recognized that the developmental energies and poten-
cies undergo a secondary distribution among the cells or tissues
at an earlier or later period, and in varying degrees, which in-
volves corresponding limitations in the secondary centers thus
created. We have long been familiar with such limitations in the
case of the "germ-layers," though the experimental evidence has
shown that they are here less rigid than was formerly supposed.
They have been experimentally shown with great clearness by
Driesch ('95) in the structures of the blastula, gastrula, and
young larva of the echinoderm at successive stages. The ex-
perimental results demonstrating the mosaic-character of cleav-
age have merely shown that similar restrictions of potency may
occur still earlier, so as to become manifest even in the early
cleavage-cells. Now, it is clear that the primary localization of
formative stuffs in the unsegmented egg is essentially an act of
the "organism as a whole;" and even though a complete preform-
ation and prelocalization of specific stuffs for every cell and tissue
were assumed — and I believe with Boveri and Fischel that such
an assumption is not necessary or even probable — we should not
escape the necessity for assuming such action of the whole. That
the egg undergoes a definite development during its ovarian his-
tory and in the stages preceding cleavage, we have evidence both
cytological and experimental. The character of the primary seg-
regation-pattern thus determined is indeed determined by the egg
as a whole, and the localization thus initiated forms the primary
basis of the specification of the blastomeres and organs that de-
258 Edmund B. JVihon.
velop from the various egg regions. This is quite in harmony
with Whitman's contention that "organization precedes cell-form-
ation and regulates it" {op. dt.^ p. 115). But, while in agree-
ment with the general spirit of his conclusions, as I understand
them, it seems to me that Whitman's statement does not sufficiently
recognize, first, the fact that the differentiating factors may un-
dergo so accurate and complete a distribution among the cells,
and be so largely emancipated from the general control as is
proved by my experiments — in other words, sufficient weight is
not given to the effects of precocious segregation; second, (and
here I should more distinctly take issue with him) that the cyto-
plasmic segregation or "organization" is a progressive or epigen-
etic process.
As regards this second point, in my preceding paper I have en-
deavored to show that the Dentaliiim egg presents a form of pre-
cocious segregation (and localization) which in other forms, such
as the eggs of certain annelids, is acquired at a later period. The
facts observed by Boverl on the Strongylocentrotns egg, and the
experimental results of Yatsu, Zeleny and myself on Cerehratu-
lus clearly indicate that in these forms, too, the cytoplasmic seg-
regation is gradually effected, and at the time of the third cleav-
age has progressed further In the nemertine than in theechlnoderm.
There is, therefore, a legitimate basis for the conclusion that the
degree as well as the form of segregation existing at the begin-
ning of cleavage may vary more or less widely; and hence for the
further assumption that the mosaic character of the early cleav-
age stages may be expressed in different degrees. For this rea-
son, in so far as the term "organization" as used by Whitman is
applied to the cytoplasmic conditions, I am unable to accept his
conclusion that the eggs of different forms do not differ In degree
of "organization;" or that "Cell-orientation may enable us to
Infer organization, but to regard It as a measure of organization
Is a serious error" {op. cit., p. 109). Such a conclusion appears
to me a petitio principii in regard to the relation between the
nuclear and the cytoplasmic organizations, and that between "pre-
formed" and "epigenetic" qualities in the cytoplasm;^ and this
^Cf. Boveri, 1902; Wilson, 1904.
Experimental Studies on Germinal Localization. 259
question Is one to be answered, not by a priori considerations, but
by observation and experiment. The facts determined by both
these methods coincide in showing that the internal factors of
cleavage are in a great number of cases so accurately adjusted to
the morphogenic factors that they may be treated practically as
identical with them. A highly differentiated initial cleavage-pat-
tern is, therefore, ipso facto evidence of a high degree of initial
cytoplasmic localization ; and the fact that the form of cleavage
may be artificially altered without affecting the end result is in
no manner opposed to this conclusion (as is pointed out in my
nemertine paper at p. 455).
I cannot better express the general conclusion which the facts
seem to me to justify than by citing the following statement from
Fischel's able general discussion ('03, p. 734) . "Der Unterschied
zwischen den verschieden Eiarten ist demnach nur ein gradueller,
in einzelnen Fallen vielleicht ein graduell sehr erheblicher, aber
doch kein essentieller. Ueberall ist das Griindprincip der (nor-
malen) Entwickeliing Mosaikarheit, und die besondere Unter-
scheidung einer Gruppe von Eiern als 'Mosaik'-Eier ist nur in
dem Sinne zulassig, als bei ihnen die Mosaikarheit besonders in
Erscheinung tritt; Mosaikeier sind jedoch in gewissem Sinne auch
alle iibrigen Eier.
"Im Besondern ist aber noch betont, dass wohl stets niir die
Primitivorgane des Embryo (materiell) in der Eizelle praform-
irt enthalten sind, und dass — ganz besonders wohl bei den sogen-
annten Regulationseiern — die materiallen Substrate fiir die Dif-
ferenzirung der specialleren Organe wahrscheinlich iiberall erst
wdhrend der spdteren Entwickelungsstadien gebildet werden.^
Im Verlaufe der Entwickelung werden stets neue und mannig-
fache Komplikationen (in erster Linie wohl durch Stoftwechsel-
processe, dann durch Lagebeziehungen u. a. m.) gesetzt, durch
welche erst bestimmte Zellgruppen des Keims, und zwar vorwie-
gend durch materielle Umwandlung oder Beimischung oder Dif-
ferenzirung nach bestimmten Richtungen hin specificirt werden ;
gerade dadurch aber verlieren sie die ihrem Miittergewebe friiher
zugekomine Fdhigkeit mehr als eben nur jene specifischen Differ-
^Cf. Wilson, '03, p. 453.
26o Edmund B. IVilson.
enziriingen gegenbenen Falles zu hesorgen, imd aiif diese JVeise
gehen Beschriinkung der Potenz tind Specialisiriing fiir Organ-
bildung einander parallel}^ * *
"Die hier entwickelte Auffassung betont also gegeniiber jener,
welche den verschieden Eiarten nur eine verschiedengradige
'Regulations fahigkeit zum Ganzen' zuerkennt, die verschieden-
stufige Abliiingigkeit zwischen Organogenese iind Eimaterialien.
Die Entstehung von Ganzbeziehungsweise Halbbildungen aus
Theilstiicken des Eies muss danach nicht auf jene verschieden-
gradige 'Regulationsfahigkeit' zuriickfiihrt werden, sondern erk-
lart sich vor Allem aus der Abhangigkeit der Differenzirung
von der 'Qualitat ihres Ausgangsgebildes' (Boveri), d. h. von
dem mehr oder minder vollstandigen Schichtenaufbau des betref-
fenden Eistiickes. Der Satz von Driesch 'Jedes Element kann
Jedes' ist demnach nur mit dem Zusatze richtig: Vorausgesetzt,
dass dieses Element alle zur Bildung des 'Jeden' nothwendigen
(im Ei vorgebildet enhaltenen) Plasmaqualitaten besitzt" (Ital-
ics in the original) .
I should only modify the above statement by recognizing the
probability that in such extreme mosaic-eggs as those of mol-
lusks or annelids the prelocalization may be much more detailed
than Fischel admits, so that for example, the ectoblast may be
represented very early in the cleavage, if not at the beginning,
not by a single equipotent ectoblast-stuff, but by a number of such
stuffs already specified for the production of various categories
of ectoblast-cells (trochoblasts, apical cells, etc.). But admitting
even to this degree the principles of prelocalization, self-differen-
tiation, and mosaic development, it is still impossible to escape the
parallel principle of correlative or dependent differentiation — i. e.,
the influence of the totality of the organism upon the development
of the individual cells. For, however definitely specified a cell
or cell-group may be, its behavior when isolated differs In some
measure from that shown when in its normal relation to its fel-
lows. The nature of the response to the change of conditions,
as the facts show, is, however, conditioned and limited by the
factors inherent in the cell or group. The further conclusions are
'^Cf. Wilson, '93, p. 610.
Experimental Studies on Germinal Localization. 261
justified, I believe, that these factors differ in different types of
eggs from the beginning, and that they become steadily more
specialized and limited as the development progresses. With the
advance of development, accordingly, the response becomes cor-
respondingly more limited.
This is shown by such a series as the following, including the
sea-urchin, nemertine and mollusk. In all these the J/2 or Ya-
blastomere produce an embryo that closes and gastrulates. In
the nemertine or sea-urchin any of these embryos may undergo
complete development, since the first two cleavages are symmetri-
cal and quantitative, distributing to each cell all the elements of
the original system. In the mollusk, however, the AB-half or
the A-, B- or C-quadrant, though undergoing certain characteris-
tic differentiations, is unable to produce a complete embryo, owing
to the absence of necessary specific material contained only in the
D-quadrant or the AB-half.^ Beyond the 4-cell stage all of the
forms exhibit limitations of potency, not primarily due to decrease
of size (as is proved by Zeleny's observations on the upper quar-
tet of Cerebratulns) , but to qualitative Internal factors. Cells of
the upper quartet, or the entire quartet, produce closed embryos
which in the nemertine or mollusk are unable to gastrulate (again
owing to the lack of specific material), but in the sea-urchin may
do so provided the third cleavage does not exclude a certain
amount of entoblast-stuff from the upper cells. The isolated pri-
mary trochoblast of the i6-cell stage completes its predestined
twQ divisions and differentiates typically except for slight changes
in the relative position of the resulting cells; but the remarkable
change of position, which in the complete embryo leads to the ac-
curate fitting together of the rows of cilia, at first disconnected,
to form continuous ciliated rings, fails to take place. Its sister-
cell (i^) likewise divides and differentiates typically; but owing
ij pointed out in the preceding paper that the failure of the CD-half to
produce a perfect larva may not improbably be owing to the fact that owing
to its great susceptibility, the larva is unable to sustain itself long enough to
assume the normal conditon. It is theoretically possible that the same may
be true of the AB-half; but the actual facts are that the latter shows from the
first certain definite defects that do not exist in the former, the CD-larvce
showing merely a lack of the proper proportions.
262 Edmund B. Wilson.
to the greater number of cells produced, it gives rise to an embryo
that closes to form a morula- or blastula-like structure. The single
trochoblast of the 64-cell stage of Patella, finally, accomplishes no
more than a simple rounding out to a spherical form, without un-
dergoing further modification of its predestined development.
Each of the reactions in this series of forms must be considered
as a response to the change of conditions that results from a de-
struction of the relation of the part to the whole, and it seems to
me the different cases must be considered as differing not in kind
but in degree. In any one of these cases the inability to produce a
perfect larva is due, as I believe, not to absolute lack of regula-
tive capacity, but to lack of necessary material, which, as far as
the experiments show, the cell is not able to manufacture anew;
and the degree of regulative response may be considered, other
things equal, as inversely proportional to the degree of segrega-
tion that has taken place. Only, therefore, in a qualified sense,
and in a more or less limited degree, can the prospective value of
a cell be considered a function of its position.^ The sense in which
this saying applies to the upper group of four in an 8-cell stage of
Cerebratulus is far more limited than that in which it applies to
a lateral group of four from the same stage {cf. Zeleny) . As ap-
plied to an isolated primary trochoblast of Patella^ it becomes so
limited as to be largely deprived of its original meaning. The
same discrimination is necessary in considering the matter of dis-
tribution of potencies in the cleavage-pattern. When, for ex-
ample, Driesch asserts that "Furchungsmosaik brauch kein Mo-
saik der Potenzen zu bedeuten" ('99, p. 729, and elsewhere), he
is stating a fact that is incontestible as far as the 2- and 4-cell
stages of the sea-urchin or nemertine egg are concerned, and which
appears to apply to the medusa egg as far as the i6-cell stage;
but when in a later paper he advances to the statement, "Fur-
chungsmosaik ist kein Mosaik der Potenzen" ('02, i, p. 812),
an assertion is made that is contrary to the results of experiment,
not only on the molluscan egg from the beginning, but even on
the nemertine or echinoderm, as soon as the 8-ceIl stage is reached.
From the facts thus far determined the conclusion seems jus-
iC/^. Wilson, '93, p. 610.
Experimental Studies on Germinal Localization. 263
tified that the power of an Isolated blastomere to produce a com-
plete embryo depends upon three conditions : first, upon its vol-
ume; second, upon the presence of all the essential elements (and
apparently of the cytoplasmic elements) of the system, and third,
upon the effectiveness of the regulative process. The production
of a complete embryo involves the regrouping of these elements in
a disposition essentially like that of an entire embryo, and I sec
no escape from Driesch's contention that this is a typical act of
regulation that cannot be explained without recourse to a factor
that lies behind the primary topographical grouping of cytoplas-
mic stuffs/ My observations on Amphioxus, the accuracy of
which I see no reason to doubt, seem to show that this regrouping
may be effected immediately upon the isolation of the cell, as
would also seem to be the case in the inverted single blastomeres
of the frog's egg observed by Morgan (though observations on
the cleavage in this case are still lacking). In the greater number
of cases thus far observed the cleavage-factors, and hence as I
think we now may say, probably also the morphogenic factors,
do not undergo immediate readjustment; and it is still quite an
open question to what extent the cells formed in the ensuing par-
tial cleavage undergo changes of prospective potence. But even
though all the essential elements of the system be present, in a
mass of sufficient volume, a failure of regulation may occur, per-
haps owing to merely physical obstacles. As pointed out in my
preceding paper, this is not improbably the explanation of the
failure of the CD-half in Dentalium to produce a perfect larva.
The ctenophore-egg is of exceptional interest in this direction; for
there is good reason to conclude (since both the cleavage and the
larva are bi-radially symmetrical) that the vertical cleavages —
i. e.^ the first and second, and perhaps also the third — are not
qualitative, yet, notwithstanding the closure of the embryo pro-
duced by the >4 or M-blastomere, the larva remains defective.
Driesch's explanation of a failure of the regulative process owing
to a "rigidity" of organization or of protoplasmic texture seems
in this case perfectly valid; but such explanation must be consid-
^Cf. Lillie, '01, p. 269; Driesch, '02.1, '02.2; Wilson, '03,. p. 456.
264 Edmund B. fVihon.
ered inadequate for the cases of qualitative division reviewed
above.
As regards the relation between self-differentiation and depen-
dent or correlative differentiation, our only guide must be the
indirect evidence derived from the response of the cell to the
change of conditions when its typical relation to the whole Is de-
stroyed by isolation or displacement from Its normal position.
For It Is perfectly obvious that If the "atypical" or secondary
changes characteristic of an isolated blastomere do not take place
In a complete embryo it is because of the relation of the cell to
the whole of which It forms a part; and It Is this "relation" that
renders the developing organism a unit, even in the most highly
differentiated type. As to what this "relation to the whole"
really Is we know practically nothing; but even though we employ
a phrase of vague and uncertain content it Is of use as Indicating
a unity or harmony of organization that is not destroyed by the
secondary distribution of the factors of differentiation among
localized centers.
It is obvious that the differentiation even of such cells as the
primary trochoblasts, which possess so high a degree of self-dif-
ferentiation, must be definitely coordinated In some way with the
development of the embryo as a whole, as Is shown for Instance
by the remarkable manner in which the rows of cilia, at first dis-
connected, are ultimately fitted together to form continuous rings
In the prototroch; but It seems equally obvious that In such cases
corrflative differentiation subsequent to division plays but a minor
part In the Internal transformation of the cell. I may here point
out, however, that the lessened Inequality of division so frequently
observed In the Isolated blastomeres is posibly an indication of
regulative response on the part of the internal factors of cell-dif-
ferentiation. It is clear that the position of the spindle — and
hence the character of the ensuing division — is definitely cor-
related with the segregation-pattern ; and in the moUuscan egg
many, probably all, of the earlier unequal divisions are qualitative
In character. It is, therefore, a fair hypothesis that in these cases^
^The unequal division of teloblasts shows that the statement should not be
made general without further evidence.
Experimental Studies on Germinal Localization. 265
the inequality is caused by, or at least correlated with, a preceding
segregation of different materials in the cell before division.
Hence it is an interesting fact that all the typically unequal divi-
sions of the normal development show a tendency to become less
unequal upon isolation of the cell. This has been observed in
the first division of the isolated M> of the 3^-micromeres, of the
i' cells, of the 5^-micromeres, the yV-macromeres, and in the
CD ^-blastomere of Dentaliiim (where It is expressed by a re-
duction in the size of the polar lobe), all of which are qualitative
divisions. This may be explicable as a result of relatively simple
physical conditions, but it is at the same time not improbably an
expression of a tendency for the segregation to recede, as it were,
towards a less definitely localized condition. The possibility is
thus suggested that the segregative process in the cells when in
their normal position in the whole embryo may, even in relatively
late stages, be in some measure influenced by their relation to their
fellows or to the whole. I believe that important light may be
thrown on this question by an accurate comparison of the later
development of isolated blastomeres that vary in this respect.
The most important question remaining is whether after com-
plete segregation and isolation of specific cytoplasmic stuffs has
been once effected by qualitative division the missing materials
may be restored by regulative metabolic processes. Such remark-
able facts as those determined in regard to the regeneration of
the lens in Triton, or Spemann's hardly less striking results on the
formation of double-headed monsters in the same animal, leave
no doubt that specific cell-characters may, within the limits of the
germ-layers, be very widely altered through a response to a local
defect, or to a change as simple as a mere mechanical alteration
of form in the growing mass ; and the facts of regeneration even
seem to show that one of the differentiated primary germ-layers
may produce structures which in the typical development arise
from a different layer. If the hypothesis of formative cytoplas-
mic stuffs be valid there seems to be no escape from the conclu-
sion that in such cases the necessary formative stuffs may be form-
ed anew. But if the potentiality of the cytoplasmic system be
primarily given in the nuclear organization, and if this be the
266 Edmund B. Wilson.
primary determining source of the initial cytoplasmic localization
in the unsegmented egg, this presents no insuperable difficulty.
It is obvious, however, that this question is one not for specula-
tion but for further experiment.
Zoological Laboratory, Columbia University.
March 29th, 1904.
LITERATURE.
BovERi, Th., '01, I. — Ueber die Polaritat des Seeigeleies: Verb. phys. med.
Ges. Wiirzburg, N. F., XXXIV.
Id., '01, 2. — Die Polaritat von Ovocyte, Ei und Larve des Strongylocentrotus
lividus: Zool. Jabrb.. Anat. Abth., XIV.
Child, C. M., '00. — The Early Development of Arenicola and Sternaspis:
Arch. Entv^rm., IX. 4.
CoNKLiN, E. G., '97. — The Embryology of Crepidula: Journ. Morph. XIII, 1.
Id., '98. — Cleavage and Differentiation: Wood's Holl Biol. Lectures, 1898.
Id., '99. — Protoplasmic Movement as a Factor in Differentiation: Ibid., 1899.
Id., '02. — Karyokinesis and Cytokinesis, etc.: Journ. Acad. Nat. Sci., Phil.,
II Sen. XII, I.
Crampton, H. E., '96. — Experimental Studies on Gasteropod Development:
Arch. Entwm., III.
Driesch, H., '92, I. — Der Werth des beiden ersten Furchungszellen in der
Echinodermenentwicklung: Zeitschr. w^iss Zool LIII.
Id., '92, 2. — Ueber einige allgemeine Fragen der theoretischen Morphologic:
Zeitschr. wiss Zool. LV.
Id., '93. — Ueber einige allgemeine entwicklungsmechanische Ergebnisse:
Mitth. zo51. Station Neapel, XI, 2.
Id., '94. — Analytische Theorie der organischen Entwickelung: Leipzig.
Id., '95. — Zur Analysis der Potenzen embryonaler Organzellen: Arch. Entwm.,
II.
Id., '96. — Betrachtungen iiber die Organization des Eies und ihre Genese:
Arch. Entwm., IV.
Id., '99, I. — Die Localisation morphogenetischer Vorgange: Arch. Entwm.,
VIII.
Id., '99, 2. — Resultate und Probleme der Entwickelungs physiologie der Thiere:
Merkel u. Bonnet, Ergebn., VIII.
Id., 'go. — Die isolirten Blastomeren des Echinidenkeimes: Arch. Entwm., X,
2, 3-
Id., '02, I.— Neue Antworten und neue Fragen: Ergebnisse, Merkel u. Bon-
net, XI.
Id., '02, 2. — Neue Ergangungen zur Entwickelungsphysiologie des Echiniden-
keimes: Arch. Entwm., XIV, 3, 4.
Experimental Studies on Germinal Localization. 267
Id., '03. — Drei Aphorismen zur Entwickelungsphysiologie ji'ingster Studien:
Arch. Entwm., XVII, i.
FiscHEL, A., '97. — Experimentelle Untersuchungen am Ctenophorenei. i:
Arch. Entwm. VI, i.
Id., '98.— Ibid., II-IV: Ibid, VII, 4.
Id., '03. — Organbilding und Differenzirung: Ibid. XV.
Heath, Harold, 99. — The Development of Ischnochiton: Z06I. Jahrb., Anat.
Ont, XII.
Heider, K., '00. — Das Determinationsproblem: Verb. d. Zool. Ges. 1900.
Hertwig, O., '92. — Urmund und Spina bifida: Arch. mikr. Anat., XXXIX.
Id, '93.— Ueber den Wert den ersten Furchungszellen flir die Organbildung
der Embryo: Ibid, XLII.
Id., '98. — Die Zelle und die Gewebebe, II: Jena, 1898.
Korschelt and Heider, '02. — Lehrbuch der vergleichenden Entwicklungs-
geschichte, Allgemeiner Theil: Fischer, Jena.
KowALEWSKY, A., '83. — Etude sur I'embryogenie du Dentale: Ann. Mus.
d'Hist. Nat. de Marseille. Zool. I, 7.
Lacaze-Duthiers, '57. — Histoire de I'organisalion et du developpement du
Dentale: Ann. Sci. Nat. IV, Ser. VI, VII.
Lillie, F. R., '99. — Adaptation in Cleavage; Wood's HoU Biol. Lectures: 1899.
Id., '01. — The Organization of the Egg of Unio: Journ. Morph. XVII, 2
Maas, O., '01. — Experimentelle Untersuchungen iiber die Eifurchung: Sitzb.
Ges. Morph. u. Phys. Miinchen, 1901.
Id., '03. — Einfiihrung in die experimentelle Entvvickelungsgeschichte. Wies-
baden.
Patten, W., '85. — The Embryology of Patella: Arb. Zool. Inst. Wien, VI, 2.
Robert, '02. — Recherches sur la developpement des Troques: Arch. Zool.
Exp., 3e Serie, Vol. X.
Roux, W., '85. — Ueber die Bestimmung der Hauptrichtungen des Frosch-
embryo im Ei, etc.: Gesammelte Abhandlungen, II, No. 20.
Id., '88. — Ueber die kiinstliche Hervorbringung "halber" Embryonen, etc.*
Ibid, II, No. 22.
Id., '93, I. — Ueber Mosaikarbeit und neuere Entwickelungshypothesen: Ibid,
No. 27.
Id., '93, 2. — Ueber die Specification der Furchungszellen: Ibid, No. 28.
Id., '95. — Gesammelte Abhandlungen, II, No. ^,2,: Nachwort.
Id., '95. — Ueber die verschiedene Entwickelung isolirter erster Blastomeren:
Arch. Entwm. I.
Id., '03. — Ueber die Ursachen der Bestimmung der Hauptrichtungen des
Embryo im Froschei: Anat. Anz., XXIII, 4-7.
Samassa, p., '96. — Studien iiber den Einfluss des Dotters auf die Gastrulation,
etc.; II. Amphibia: Arch. Entwm., II.
Whitman, C. O., '93. — The Inadequacy of the Cell-Theory of Development:
Journ. Morph., VIII.
Wilson, E. B., '92. — The Cell-Lineage of Nereis: Journ. Morph., VI.
Id., '93. — Amphioxus and the Mosaic Theory of Development: Journ. Morph.,
VIII.
2 68 Edmund B. JrUson.
Id., '94. — The Mosaic Theory of Development; "Wood's Holl Biol. Lect.,
II. 1893.
Id., '96. — On Cleavage and Mosaic Work: Arch. Eiitwm., III.
Id., '03. — Experiments on Cleavage and Localization in the Nemertine Egg;
Arch. Entwm., XVI, 3.
Id., '04. — The Germ-Regions in the Egg of Dentalium: Journ. Exp. Zool. I, i.
Y.\TSU, N., '04.— Experiments on the Development of Egg Fragments in
Cerebratulus: Biol. Bull, VI, 3-
Zeleny, C, '04. — Experiments on the localization of developmental Factors in
the Nemertine-Egg: Journ. Exp. Zool., I, 2.
ZqjA, R., '95. — Sullo sviluppo dei blastomeri isolati delle uova di alcune
meduse; Arclr. Entwm., I, II.
CONTIRBUTIONS TO THE PHYSIOLOGY OF THE
VENTRAL NERVE-CORD OF MYRIAPODA
(CENTIPEDES AND MILLIPEDES).
(Six Figures.)
A. J. CARLSON.
(From the physiological laboratory of Leland Stanford, Jr., University.)
I. The Rate of Propagation of the Nervous Impulse in the
Ventral Nerve-Cord.
The measurements of the rate of propagation of the nervous
Impulse were made in the fall of 1902. Their publication has
been delayed with the v- iew of obtaining large centipedes from the
tropics for the work, so as to exclude a possible source of error in
the measurements when smaller specimens are made use of. The
attempts to obtain larger centipedes than those available here in
California have not proved successful, and the further work must
therefore be postponed till more favorable material becomes avail-
able.
The structure and relations of the central nervous system of
the centipedes and millipedes are essentially the same as in the
annelid worms. Each segment Is provided with a pair of ganglia,
which are connected by transverse commissures and by longitudi-
nal commissures with the neighboring anterior and posterior pairs
of ganglia. This nerve-cord is situated ventral to the gut. In
the anterior or head segment It Is connected by a commissure on
either side of the oesophagus with the supra-cesophageal ganglion
or "brain."
270
A. J. Carlson.
The method of measuring the rate of propagation of the nerv-
ous impulse through this nerve-cord was essentially the same as
that employed by Dr. Jenkins and myself in the similar work on
the ventral nerve-cord of worms. ^ The preparation and arrange-
ment of the animal for the experiment are shown in the diagram
in Fig. I. The centipede was placed with its dorsal side next
to the platform or removable floor of the moist-chamber and se-
E / El
I I /^
3
Fig. I.
Diagram illustrating the method of measuring the rate of propagation of
the nervous impulse in the ventral nerve-cord of centipedes. A, friction wheel;
B, pins fixing one end of the reacting portion to the platform; E, electrodes;
L, recording-lever; P, platform or floor of moist-chamber.
■cured to the board by pins, care being taken not to injure the nerve-
cord. In Scolopendra morsitans and Scolopocryptops sexpinosus
three or four segments are sufl'icient as reacting or contracting
portion, in the long and very slender centipede Himantarium taen-
iopse"^ eight to ten segments must be used, while in the millipede
{Jules sp.), in which the actual lengthening or shortening of
any part of the body is very slight, eight to ten segments must be
employed in order to give sufl'iclent amplitude to the excursion of
the recording-lever. The segment next to the reacting portion
was fixed to the board by means of two pins in the manner shown
ijenkins and Carlson, Journal of Comparative Neurology, XIII, p, 259, 1903.
-These centipedes w^ere identified for me by Mr. R. V. Chamberlin, of Cor-
nell University. The centipede Stylolccinus, sp., made use of in studying the
reflexes, was identified by Mr. R. E. Snodgrass. of Stanford University.
Physiology of Ventral Nerve Cord of Myriapoda. 271
In Fig. I, so that the contractions of the body anterior to this
point could not be communicated to the lever. The freeing of
the nerve-cord for the application of the distal and proximal elec-
trodes is a very difficult undertaking, and in no instance was it
done as completely as indicated in Fig. i, especially in the slender
Himantariiim, in which the nerve-cord is correspondingly slender,
and in the millipede, in which the dissection is rendered difficult
by the very thick chitenoid epidermis. The dissection for the
proximal electrodes was in every case made at least two or three
segments from the reacting portion of the body, to avoid escape
of the current directly to the reacting musculature. In Himan-
§mm
Fig. 2. — Scolopendra.
Tracings of the contraction of the posterior segments on stimulation of
the cord at the distal and the proximal points. Length of cord, 5cm. Trans-
mission time of the impulse, 0.02 sec. Rate, 2.50m. per sec. Time, 100 d v.
per sec.
tarium six to ten segments were allowed to intervene between the
point of stimulation and the reacting portion.
No anaesthetics were used, but prior to fixing the animal to
the platform the head segment, including the cerebral ganglion,
was usually removed.
The posterior or tail segments of the decapitated centipede
which has been fixed to the board and prepared in this manner
usually become quiescent after a few minutes, and remain quies-
cent during the intervals between the stimulation of the cord,
provided the tension from the recording-lever is not too great.
272
A. J. Carlson.
When the tension due to the weight of the lever is considerable
the segments are kept in constant motion until exhausted. And
the same is true if the segmental appendages or legs are able to
reach or touch any object. The contact of the legs with any
solid object evidently starts reflex movements of locomotion, and
for that reason the preparation does not become quiescent until
nearly exhausted when fixed to the platform ventral side down
so that the ambulatory appendages are in contact with the board.
When the anterior end of the centipede serves as the reacting
portion the reflex restlessness is much greater than when the pos-
terior segments are employed. This is true whether the head seg-
ment is removed or not. The measurements of the rapidity of
conduction of the postero-anterior impulses in the cord by the pres-
ent method is therefore attended with greater diflliculties than the
Fig. 3. — Scolopaidra.
Tracings of the contraction of the posterior segments on distal and proxi-
mal stimulation of the cord. Length of cord, 6 cm. Transmission time of the
impulse, 0.022 sec. Rate, 2.70 m. per sec. Time, 100 d. v. per sec.
measurement of the antero-posterior rate. In the millipede the
union of the segments admits of only slight elongation and con-
traction of the body, but the body may be coiled by contraction
of the ventral muscles in the segments. The amplitude of this
contraction is much greater in the posterior than in the anterior
Physiology of Fentral Nerve Cord of Myriapoda. 273
portion of the body, and for that reason the postero-anterior rate
of the nervous Impulse cannot very well be determined with the
present method.
In the centipedes Scolopendra and Scolopocryptops a single in-
duced shock of moderate intensity applied to the nerve-cord either
at the anterior or at the posterior end of the body produces con-
traction of every segment in the body. In the work on these ani-
mals the break induced shock was therefore used as the stimulus.
This reaction to the single induced shock is not obtained in the
long and slender centipede Himantarium or in the millipede. In
Fig. 4.— Scolopendra.
Tracings of the contraction of the anterior segments on proximal and dis-
tal stimulation of the cord. Length of cord, 4.5 cm. Transmission time of the
impulse, 0.03 sec. Rate, 1.50 m. per sec. Time, 100 d. v. per sec.
Himantarium a single induced shock ev^en of very great intensity
applied to the anterior or posterior end of the nerve-cord does
not always produce a contraction that extends over the whole ani-
mal. The contraction extends further from the point of stimu-
lation the stronger the induced shock, but rarely from one end of
the animal to the other. When the cord is stimulated with three
or four weak induced shocks that follow one another in rapid
succession the contraction involves every segment in the body.
In the experiments on this centipede short series of the interrupted
274 ^- J- Carlson.
current was therefore used as stimuli. A single induced shock
applied at one end of the nerve-cord of the millipede Jules pro-
duces progressive movements of the ambulatory appendages or
legs from the point of stimulation to the opposite end of the ani-
mal, but the contraction of the muscles moving the body seg-
ments is confined to the immediate vicinity of the point of stimu-
lation; but a short series of the tetanizing current produces con-
traction of these muscles in all the segments of the body. A
similar condition was found by Dr. Jenkins and myself to obtain
in the marine annelid Aphrodite, in which a single induced shock
applied to the ventral nerve-cord produced contraction of the
muscles that move the setae, but a tetanizing current was required
to produce contraction of the muscles moving the segments. It
is therefore probable that the nervous mechanism of the setae in
Aphrodite and of the legs in the millipede Is less complex and
more readily excited than is the nervous mechanism in connection
•with the muscles that move the segments. If one of the setae in
the worm and one of the legs of the millipede could be used for
raising the lever and the rapidity of transmission of the impulse
In this nervous mechanism thus measured, it would undoubtedly
be found to be several times greater than that in the nervous
mechanism to the segmental muscles.
The character of the records produced by the contraction of
the reacting portion on stimulation of the nerve-cord may be
gathered from the typical tracings reproduced in Figs. 2 to 6.
Only the first part of the tracings showing the latent period and
the amplitude of contraction is given, as these are the only points
with which we are concerned. In the records from the millipede
(Fig. 6) the rising curves represent the gradual bending ventral-
wards of the reacting portion, the movements of each segment
fusing into one, apparently continuous, contraction. Each stimu-
lation of the cord by a tetanizing current of short duration
usually produces but one such movement. The records from the
centipedes are more irregular from the fact that each stimulation
of the cord usually starts a series of movements or rather con-
tractions and relaxations which may last for a minute or two in
the fresh preparations.
Physiology of Ventral Nerve Cord of Myriapoda. 275
Because of the very complex nature of the muscular part of
the preparation the character of the curves, that is, the rapidity
and the amplitude of the contraction is not a very accurate guide
in determining the admissability of individual records. For ex-
ample, two successive tracings produced by stimulation of the
cord at the distal or at the proximal point may show great diverg-
ence in the amplitude of the contraction and yet exhibit the same
latent period or they may be nearly identical in the amplitude
and rapidity of the contraction and yet show a difference in the
latent period of from 15 to 25%. The tracings that showed a
greater difference in the amplitude of the contractions than is
exhibited by the records in Fig. 3 were usually excluded.
Fig. 5. — Himantarium.
Tracings of the contraction of the posterior segments on distal and proxi-
mal stimulation of the cord. Length of cord, 14 cm. (120 segments). Trans-
mission time of the impulse, 0.52 sec. Rate, 27 cm. per sec. Time, 50 d. v.
per sec.
Of the centipedes worked on the best preparation for these ex-
periments is obtained from Scolopendra. The largest specimens
yield a length of nerve-cord between the distal and the proximal
points of stimulation of from 5 to 6 cm. This centipede is rela-
tively stout and the reacting segments amply able to lift the re-
cording-lever. Himantarhim is more than twice as long as Scolo-
pendra, but is so slender that it is even difficult to fix the specimen
to the platform without injuring the nerve-cord with the pins.
For the experiments on this centipede the recording-lever had to
be very light.
276
A. J. Carlson.
It was stated that the point of application of the proximal
electrodes to the cord was always three or more segments distant
from the reacting portion. This was done with two ends in
view, namely, to prevent escape of the current directly on to
the muscle and to prevent errors in the measurements from stimu-
lation of a more direct nervous mechanism on proximal than on
distal stimulation. In the annelids the cell bodies of the motor
neurones to the musculature of any one segment are situated in
the ganglia of the same segment as well as in the ganglia of the
adjoining anterior and posterior segments. The conditions are
in all probability the same in the nerve-cord of the centipedes and
Fig. 6. — Juks.
Tracings of the contraction of the posterior segments on distal and proxi-
mal stimulation of the cord. Length of cord, 5 cm. Transmission time of the
impulse, 0.24 sec. Rate, 21 cm. per sec. Time, 50 d. v. per sec.
the millipedes. Now, if the cord is stimulated in the segment next
to the reacting portion it is probable that some of the neurones
to the reacting musculature are stimulated directly, while when
the cord is stimulated at a point 5 to 14 cm. further away these
neurones are probably stimulated indirectly; in other words, there
is probably "synapses" at the junction of the longitudinal con-
ducting paths in the cord and the motor cells to each segment. At
such junctions the propagation of the nervous impulse is in all
probability retarded. If therefore the latent time in the records
Physiology of Ventral Nerve Cord of Myriapoda. 277
on distal stimulation includes this delay while the records obtained
on proximal stimulation do not, it is obvious that the rate of prop-
agation of the impulse as calculated from the latent periods of
these records would be less than the actual. For that reason it
would be desirable to check up the measurements on these com-
paratively short centipedes by experiments on larger representa-
titves from the tropics, as in larger specimens this possible source
of error can be practically excluded.
To give an idea of the variability of the latent time in the
records obtained by this method, three series of experiments are
given in detail in Tables I, IV, and V. All of the experiments are
summarized in Tables II, III, V and VII. The character of the
tracings has already been referred to. It is amply illustrated in
figs. 2 to 6.
TABLE I.
Scolopendra morsitans. Antero-posterior. Detail of experi-
ment No. 2, Table II, October 17, 1902. Temperature, 16° C.
TOTAL LATENT TIME IN SECONDS.
Distal. Proximal.
0.045 0.028
0.047 ^-^^5
0.048 0.027
0.047 0.026
0.045 0.025
0.047 °-°^5
Average. . . 0.046 0.026
Transmission time 0-02 sec.
Length of nerve-cord 5 ^"^•
Rate 2.50 m. per sec.
TABLE IL
Summary of the measurements of the antero-posterior rate in
the nerve-cord of Scolopendra (No. 1-8) and Scolopocryptops
(No. 9-13). The length of nerve-cord involves from 13 to 17
segments.
278 A. J. Carlson.
No. of pairs Transmission Length of cord
of records. time in sec. in cm. Rate in cm,
I 8 0.020 5 2.50
2 6 0.020 5 2.50
3 II 0.030 6.5 2.16
4 13 0-023 5 2.15
5 4 0.025 ^-5 2.60
6 ..... . 2 0.020 5 2.50
7 3 0.019 6 3.15
8 2 0.026 5 1.94
9 4 0.015 5 3.33
10 4 0.015 4 2.64
II 3 0.016 4 2.40
12 8 0.017 4.5 2.60
13 3 0.024 6 2.46
Mean rate 2.50 m. per sec.
TABLE III.
Summary of measurements of the postero-anterior rate in the
nerve-cord of Scolopendra.
No. of pairs Transmission Length of cord
of records. time in sec. in cm. Rate in cm.
I 5 0.040 6 1.50
2 4 0.032 4.5 1.40
3 4 0.040 7 1.75
4 3 0.037 4 i-o8
5 4 0.040 6 1.50
Mean rate i .40 m. per sec.
TABLE IV.
Himantarium taeniopse. Antero-posterior. Detail of experi-
ment No. 2, Table V, November 5, 1902. Temperature 18" C.
Physiology of Ventral Nerve Cord of Myriapoda. 279
TOTAL LATENT TLME IN SECONDS.
Distal. Proximal.
0.46 o.io
0.48 0.1 1
0.51 0.13
0.40 0.10
0.42 0.09
0.46 O. II
0.45 0.09
0.45 0.13
0.43 0.13
0.45 O.I I
0.46 0.13
Average. . . 0.45 o.i i
Transmission time 0-34 sec.
Length of cord ( 100 segm.) 10 cm.
Rate 26.4 cm. per sec.
TABLE V.
Summary of the measurements of the antero-posterior rate in
the nerve-cord of Himantarium.
No. of pairs Transmission Length of cord
of records.
time in sec.
in cm.
Rate in cm.
I .
9
0-43
12
^ 100 segm.)
27.6
2.
II
0.34
10
[ 100 segm.)
26.4
3-
7
0.46
14
^ no segm.)
29.4
4-
8
0.37
12
1 15 segm.)
32-5
5-
8
0.52
14
^120 segm.)
28.0
6.
18
0.49
13
125 segm.)
27.0
Mean rate
28.5
cm.
per sec.
28o A. J. Carlson.
TABl.E VI.
Jules sp. Antero-posterior. Detail of experiment No. i, Table
VII, November i6, 1902. Temperature 16" C.
Distal. Proximal.
0.40 0.16
0.37 0.18
0.37 0.17
0.36 0.14
0.36 0.15
0.37 0.14
0.39 0.13
0.37 0.12
0.40 0.16
0.40 0.12
0.42 0.16
0.44 0.15
Average. . . 0.38 0.15
TOTAL LATENT TIME IN SECONDS.
Transmission time 0-23 sec.
Length of nerve-cord 6 cm.
Rate 25.8 cm. per sec.
TABLE VII.
Summary of the measurements of the antero-posterior rate in
the nerve cord of the millipede Jules. The length of cord involves
32 to 37 segments.
No. of pairs Tra
of records. t
I 12
2 15
3 4
4 2
5 5
6 7
7 6
8 6
9 7
ID 2
Mean rate 20 cm. per sec.
lission
Length of cord
m sec.
m cm.
R
ate m cm
0.23
0.36
6
6
25.8
16.8
0.28
0.38
5-5
6
19.8
16.2
0-35
6
17-4
0.17
5
30.0
0.22
U.30
5.5
6.
24.7
20.0
0.29
6
20.4
0.30
5-5
18. 1
Physiology of Vetitral Nerve Cord of Myriapoda. 281
The rapidity of propagation of the antero-posterior nervous
impulse in the cord is the same in the two centipedes Scolopendra
and Scolopocryptops. These two centipedes are also closely alike
in the number of segments and in the swiftness of their reactions
and movements. The rate is lower than one might have expected,
judging by the quick movements of these animals. While it is
higher than the rate in the ventral nerve-cord of some of the
worms, it is only about one-half that in the nerve-cord of the higher
marine annelids Glycera, Eunice and Bispira (one of the Sahel-
lidae) .
The great difference between the rate in Scolopendra and Scolo-
pocryptops on the one hand and that in Himantarium on the other
is probably due to a greater number of "synapses," that is, a great-
er complexity of the conducting path in the cord of the latter. Him-
antarium exhibits a much greater segmental independence than do
the other two centipedes. In Himantarium the progression of the
contraction from the point of stimulation is slow enough to be ob-
served by the eye, while in Scolopendra every segment of the body
seems to contract at the same time on stimulation of the nerve-cord
at any one point. In view of the relatively low rate even in
Scolopendra it seems to me probable that the conducting path in
the cord is not made up of a system of uninterrupted nerve-fibers,
although it is evidently less complex than the corresponding con-
ducting path in Himantarium.
The rate in the nerve-cord of the millipede is the lowest of all,
or only 20 cm. per sec. This is only one-third that of the lowest
rate recorded in the nerve-cord of the annelids, namely in the
leech (56 cm. pr sec), and in the marine worm Aphrodite (55
cm. per sec). The reactions and movements of Jules are also
much slower than those of Scolopendra or Scolopocryptops.
From the fact that the rate of conduction of the impulse in the
nerve appears to stand in direct relation to the rapidity of the
processes of contraction in the muscle supplied by the nerve, ^ it
seems probable that the difference in the rate in Scolopendra and
Jules is not solely apparent and due to the greater complexity of
the conducting path in the latter animal.
^Carlson, American Journal of Physiology, 1904. IX, p. 401.
282 A. J. Carlson.
A comparation of Tables II and III leaves no doubt that in
Scolopendra the rapidity of conduction of the impulse through
the cord is greater in the antero-posterior than in the postero-an-
terior direction. A similar condition exists in the case of the
spinal cord of the California Hagfish {Bdellostoma) and there
are indications of the same condition in the spinal cord of the
snake/ In the annelid Glycera, on the other hand, the rate in the
ventral nerve-cord is the same in both directions. ■ It is difficult
to understand how this difference in the rate of conduction of the
postero-anterior and the antero-posterior nervous impulses has
come about in the course of development. For the preservation
of the individual it would seem that a rapid transmission of the
nervous impulse is just as essential over the sensory part of the
reflex arch as over the motor part.
II. The Reflex Functions of the Ventral Nerve-Cord and the
Segmental Ganglia.
The great difference in the rate of propagation of the nervous
impulse in the cord of Scolopendra and Himantariiim lead to the
study of the reactions and locomotions of these animals under
natural conditions as well as of the reflexes exhibited after sever-
ance of the head segment, together with the supra-oesophageal
ganglion or "brain," in order to determine whether these ani-
mals exhibit other differences in conformity with the difference
in the rate of the nervous impulse.
Himantarium has two modes of locomotion, namely, by means
of its legs and by means of series of contraction waves passing
from one end of the. body to the other exactly as in the worms.
These movements are so identical with those of the worms that
the muscular mechanisms are probably the same or at least simi-
lar. The centipede works its legs at the same time that it resorts
to the other method of getting over the ground. The worm
method of locomotion comes into play only when the anmal is in
a hurry to get away from an enemy. It is made use of with
^Carlson, Archiv fiir die gesammte Physiologic, 1904, Ci, p. 231.
-Jenkins and Carlson, loc. cit.
Physiology of Ventral Nerve Cord of Myriapoda. 283
equal adaptation in moving either forwards or backwards, just
as in the worms. In Scolopendra and Scolopocryptops the legs
are the exclusive means of locomotion, whether the progression is
hurried or slow. The chitenoid epidermis attains also a greater
development in these centipedes.
Himantarium moves backwards or forwards with equal facility
and rapidity. When at rest and touched anteriorly it runs back-
wards; on being touched posteriorly it proceeds forwards. Scolo-
pendra or Scolopocryptops does not move backwards for any
length of time, and nev^er when making haste to escape from dan-
ger, as their backward locomotion is much slower than their pro-
gression. When Himantarium is beheaded its body keeps run-
ning backwards continuously for ten to fifteen min. before it starts
to move in either direction, while the decapitated Scolopendra
keeps running forwards, no matter what obstacles are placed in
its way, and it is very difficult to induce it to walk backwards,
even after the excitation from the injury has partly subsided. It
is therefore plain that Himantarium and related genera exhibit a
less degree of antero-posterior differentiation than do the shorter
and stouter centipedes. This is further shown by the fact that
when the quiescent Himantarium, which is usually coiled up in a
bunch, is gently disturbed by light or by touching it, the two ends
of the animal will often be found to crawl or move in opposite
directions at the same time, that is, the head end walks forwards,
the hind end backwards, till the body is straightened out, when
either end may take the lead. This was never observed in Scolo-
pendra or Scolopocryptops.
When Scolopendra or Scolopocryptops are decapitated by re-
moving the anterior segment, inflicting as little injury as possible
to the body, the body usually continues to move forwards inces-
santly and rapidly for five to ten min., lifting the anterior three
or four segments next to the wound high up from the ground.
After the elapse of a few minutes the body becomes relatively
quiescent, usually moving only when stimulated or touched. If
placed on its dorsal side, the decapitated animal straightway
turns over on its legs. When the posterior part of the body is
touched, it either springs forwards or brings the anterior end of
284 A.J. Carlson.
the body around as if to bite, reactions identically the same as
those of the intact animal. When these centipedes are cut in two
in the middle the posterior half exhibits the reactions just de-
scribed, with the exception that it does not turn over on its ventral
side so readily when placed on its baclc, but it attempts to do so
in every case. The number of segments may be further reduced
without destroying the coordinating mechanism of locomotion. If
the sections are made with a razor or a pair of sharp scissors,
the whole body may be divided Into portions of three or four seg-
ments in length, each portion still retaining coordination to the
extent that it walks across the t^ble and keeps up locomotion for
three to four minutes, but it exhibits no sense of equilibrium —
that Is, attempting to turn over on Its ventral side when placed on
Its back. The direction of the locomotion In these small portions
of the body Is almost invariably forwards. The beheaded Scolo-
pendra or Scolopocryptops live and react In this manner for three
to four days. After the initial restlessness, evidently due to the
stimulation from the lesion, it scarcely stirs If left undisturbed, al-
though its excitability Is retained apparently unimpaired for 24
to 48 hours. It does not react to light. When placed in a glass
jar provided with sand or moist earth in one corner It usually
comes to rest on these places rather than on the glass.
The beheaded Hiinantariiim lives and reacts for seven to eight
days, showing much more "spontaneous" activity than the decapi-
tated Scolopendra. An 8 to 10 mm. long portion of the body
usually exhibits the same reflexes and degree of coordination as
the Scolopendra deprived of only Its head segment. A portion
of that length walks forwards or backwards with apparently per-
fect coordination of its legs, and it turns over on Its ventral side
when placed on Its back, keeping up these reactions for 24 to 48
hours after being isolated from the rest of the body. A portion
of three segments walks in either direction, the usual tendency
being to forward progression. A portion of five to six segments
exhibits the equilibrium reflex in attempting to regain its natural
position when placed on Its back. Longer portions turn over
promptly.
Physiology of Ventral Nerve Cord of Myriapoda. 285
The loss of excitability and death of the decapitated Himan-
tarium proceeds antero-posteriorly. When the animal is simply
cut in two in the middle the anterior half with the head intact
dies sooner than the posterior half. The same is true when this
centipede has been bitten in the middle by a Scolopendra or a Scold-
pocryptops, in which case the posterior half of the body usually
lives for from 12 to 24 hours while the head end ceases to react to
stimuli withm 2 to 6 hours. The poison of these centipedes is
also fatal when introduced into their own bodies. When a Scolo-
pendra is seized at its middle by a pair of forceps it usually turns
about and bites the forceps, but occasionally It will bite the pos-
terior part of its own body, and always with fatal results, the
symptoms of the poisoning appearing in gradual loss of coordi-
nation and power of locomotion, death following within 10 to 15
hours.
The decapitated Stylolaemiis lives and reacts even longer than
Himantarhim, or for 12 to 14 days. The only difference in the
behavior of the decapitated and the intact Stylolaemiis seems to
be the absence of the reaction to light in the former. The wounds
of the decapitated Himantarhim and Stylolaemiis that lived for
8 to 14 days healed in some cases completely. There was no in-
dication of a regeneration of the lost part. The death was prob-
ably due to starvation rather than to infection from the wound.
When a number of specimens of Himantariiim and Scolopen-
dra or Scolopocryptops are confined together where they can be
readily observed, it will be seen that Himantariiim jerks back and
makes haste to get away whenever any portion of Its body comes
in contact with the other two centipedes. And it has good rea-
sons to do so, as it Is an easy prey for these strong and ferocious
centipedes. A similar but much less pronounced jerking back of
the body is exhibited by all the centipedes studied when they come
In contact with the bodies of other Individuals of even their own
species, especially when the animals are much excited and mov-
ing about rapidly, but In no case Is it as pronounced as in Himan-
tariiim on coming In contact with the aforementioned species.
The decapitated Himantariiim exhibits this very same reaction.
Especially if the posterior end of the headless body comes In con-
286 A.J. Carlson.
tact with the centipedes, the body jerks back, and both modes of
locomotion are usually employed in getting away. That the re-
action is more pronounced when the posterior end of the body
makes the contact is probably due to reduced excitability of the
anterior segments next to the wound. The decapitated animal
continues to react in this manner for several days.
The decapitated Himantarium retreats from water just as the
intact individual, but on coming in contact with other objects in
its path it simply walks over or around them. When, however,
a solid object, like a pencil or a pair of forceps, is moved towards
the crawling centipede and the contact thus made, the decapitated
animal usually retreats. When the body comes in contact with
an object which is moving towards it, the impact is necessarily
stronger than when the object is stationary and the centipede alone
moving, hence the difference in the motor reaction is probably
due to the quantitative difference in the sensory impulses. But
the decapitated Himantarium jerks back and retreats from Scolo-
pendra and Scolopocryptops even when these latter lie perfectly
dormant, so that the reaction cannot be explained on that ground.
One further possibility must be investigated before this reaction
can be ascribed to a qualitative discrimination in the motor reac-
tions to touch impressions on the part of the decapitated centi-
pede. The touch impressions may namely be supplemented by
those of temperature. I have made no measurements of the body
temperature of these animals, and until such determinations are
made this interesting point must be left undecided.
Cross-section of the ventral nerve-cord in any part of the body
destroys the coordination between the two ends of the body on
either side of the lesion just as effectively as when the whole body
is cut transversely and the two parts rejoined by a thread or a wire.
The lesion does not destroy the coordinated locomotion of either
half, but the direction of the locomotion of the anterior half may
or may not be the same as that of the posterior half. When the
direction is not the same, a "tug of war" ensues, in which the
portion having the greatest number of segments or having the
most favorable ground for contact for Its legs comes out victo-
rious. Scolopendra usually turns about and bites its refractory
hind body repeatedly.
Physiology of Ventral J\ crze Cord of Myriapoda. 287
When the milllped': Jules is cut transversely in the middle the
coordination is destroyed in the posterior half. The anterior
portion continues to move about for a short time but loss of co-
ordination and death ensue within 10 to 20 min., and the same is
true when the animal is decapitated. This animal is therefore
not suited for the study of the reflexes and the relative indepen-
dence of the coordinating mechanisms of the segmental ganglia.
To recapitulate : Locoitiotion, movements to regain normal
posture, as well as all contact reactions in the centipedes are ob-
viously reflex movements not dependent on the cesophageal nerv-
ous complex or "brain," as the decapitated centipede exhibits the
same reactions and movements as the intact animal, save that it
does not avoid light and cannot feed or make passages for itself
in the ground. The decapitated centipede is not abnormally rest-
less, so that any inhibitory functions can be ascribed to the cesoph-
ageal nervous complex, nor is it quiescent to the extent that so-
called "spontaneous" movements may be said to be wanting. The
bending of the anterior part of the body preparatory to bite the
object touching the posterior part is a reflex not dependent on the
"brain." The maintenance of the body ventral side down is also
a reflex through the segmental ganglia, the turning of the body
to the ventral side when placed on its back probably d^iyending
not so much on the touch impressions on the dorsal side as the
absence of the normal touch impressions from the contact of the
legs with the ground. The relatively great segmental indepen-
dence of this equilibrium reflex and especially of the reflex and co-
ordinating mechanisjtis of locomotion is shown by the fact that
these are exhibited by any portion of the body measuring not less
than three intact segments in length.
The short and stout centipedes {Scolopendra, Scolopocryptops)
exhibit a greater antero-posterior differentiation and a less degree
of segmental independence than do the long and slender centi-
pedes {Himantarium, Stylolaemus) . These latter centipedes re-
tain the annelid mode of locomotion , and the transmission of the
nervous impulse through their ventral nerve-cord is slower.
NOTE ON THE GALVANOTROPIC REACTIONS OF
THE MEDUSA POLYORCHIS PENICILLATA
A. AGASSIZ.
BY
FRANK W. BANCROFT.
(From the Rudolph Spreckels Laboratory of the University of California.)
Comparatively few papers on the galvanotropic reactions of
coelenterates have been published and so far as I know there are
only two bearing directly on the questions here considered. The
first is by PearP, who finds that when any, except the very strong-
est, galvanic currents are passed transversely through hydra the
animal contracts most strongly on the anode side so that the
free end — which may be either oral or aboral — swings around
and points towards the anode. The tentacles, however, behave
differently. With weak currents only those tentacles which are
parallel to the current lines contract, but of these the one towards
the cathode has a tendency to contract most strongly. When the
whole animal has become oriented the tentacles curve slightly so
as to become concave on the side towards the cathode. The sec-
ond observation is by Greeley and will be considered in detail
later on.
The tentacles and manubrium of Polyorchis penicillata, which
occurs abundantly in San Francisco Bay during certain seasons
of the year, furnish excellent material for the demonstration of
galvanotropic reactions, responding to the current in some re-
spects like the tentacles of hydra, but with greater distinctness.
The method of experimentation consisted in cutting the medusae
iPearl, 1901, Studies on the Effects of Electricity on Organisms. II. — The
Reactions of Hydra to the Constant Current. Amer. Jour. Physiol, Vol. V,
PP- 301-320. ,
290
F. W. Bancroft.
in various ways and placing the pieces in a trough of sea water
through which the galvanic current was conducted with non-polar-
izable electrodes. The current strength varied from 25 to 200 'J.
The responses were usually distinct with 25 '^, but became more de-
cided as the current was increased.
If a meridional strip passing from the edge on one side through
center of the bell to the other edge be prepared and the current
passed through it transversely, tentacles and manubrium turn and
point towards the cathode (Fig. I) . A reversal of the current in-
r^
+
Fig. I.
itiates a turning of these organs in the opposite direction, which
is usually completed in a few seconds. This can be repeated many
times and the tentacles continue to respond after hours of ac-
tivity. The manubrium, however, tires sooner and fails to re-
spond. If the strip is placed with its subumbrellar surface up-
wards and extended in a straight line parallel to the current lines
the making of the current causes the tentacles at the anode end to
-h
Fig. 2.
turn through an angle of 180 degrees and point towards the
cathode. The tentacles at the cathode end become more crowded
together, reminding one of the tip of a moistened paint brush,
and also point more directly towards the cathode (Fig. 2) . The
experiment may be varied in still other ways by cutting smaller
or larger pieces from the edge of the swimming bell, but the re-
sponse is always the same. The tentacles wherever possible, and
to a less extent, the manubrium, bend so as to point towards the
Galvanotropic Reactions of Polyorchis. 291
cathode. The response depends In no way upon the connection
of these organs with the swimming bell, muscles or nerve-ring,
for it is obtained equally well with isolated tentacles and pieces
of tentacles. Isolated tentacles when placed transversely to the
current lines curve so as to assume a more or less complete U-
FlG. 3.
shape, with their concave side towards the cathode (Fig. 3).
When placed parallel to the current the tentacles do not curve
(Fig. 3» «)•
If the tentacles are relaxed the making of the current causes
them to contract rapidly. Subsequently they turn their concave
side towards the cathode, and remain contracted for a consider-
able period. But if the current is continued long enough through
the isolated tentacles a partial relaxation comes on which is sud-
denly followed by another rapid contraction; so that we have, as
this process repeats itself, a slow and irregular rhythmic contrac-
tion caused as in the case of the quiescent frogs ventricle by the
constant flow of the galvanic current. If the current is continued
Fig. 4.
still longer in some cases a local anodal relaxation occurs and the
isolated tentacles then have the appearance of Fig. 4. As the
figure shows, this relaxation is at the bend of the U in the curved
tentacles and at the anodal end in those which were parallel to
the current lines and did not curve.
292 F. W . Bancroft.
It is evident that these phenomena lend themselves very nicely
to Loeb's^ explanation of galvanotropism, which he considers de-
pends on similar changes in the tension of associated groups of
muscles. The constant flow of the current brings about an in-
crease of tension on the cathodal side of tentacles and manubrium,
as a result of which this part is more strongly contracted than the
anodal portion. When the tentacles are exhausted the anodal
part may even be completely relaxed,
Greeley- has stated that when "Gonionemus was exposed to the
constant current, rhythmical contractions began always on the
cathodal side when the medusa was immersed in normal sea
water, but that the contractions began on the anodal side in acidu-
lated sea water." A series of experiments was made on Polyor-
chis to test its behavior in acid and alkaline sea water, but as long
as the tentacles were sufficiently uninjured so that they responded
at all to the current, they behaved as above described, «o matter
what the reaction of the water. The influence of acid and alka-
line media on the contraction of the muscles was also tested, but
Greeley's results could not be confirmed. Usually a change in
the reaction of the sea water made no difference, and even when
it did the change in the electrical response was sometimes in one
direction and sometimes in another, so that no significance could
be attached to it.
As a rule the muscles of a meridional strip of Polyorchis do not
behave towards the galvanic current as described by Greeley for
Gonionemus in normal sea water; for the place of maximum re-
sponse is the anode. It is here that the contractions usually start
and here that the most rapid rate of the rhythmic contractions is
usually seen. But there is such an abundant opportunity for stim-
ulation at secondary cathodes that I am not yet prepared to say
that we have here an exception to Pfliiger's law.
Berkeley, April 9, 1904.
_ ^Loeb. J., 1897, Zur Theorie der physiologischen Licht und Schwerkraft-
wirkungen. Pfliiger's Archiv. Bd. 66, p. 440:
-Trelease, W., 1903. Report of a meeting of the Academy of Sciences of
St. Louis. Science, N. S., Vol. XVIII, p. 753.
EXPERIMENTS ON THE LOCALIZATION OF DEVEL-
OPMENTAL FACTORS IN THE NEMER-
TINE EGG.
CHARLES ZELENY.
CONTENTS.
1. Introduction 293
2. Method 294
3. Normal Development 205
4. Experimental Results 298
I. Unfertilized Egg 299
II. Fertilization to Complete Separation of First Polar Body . . . 302
III. First Polar Body to Complete Separation of Second Polar Body 302
IV. Second Polar Body to Beginning of Lateral Elongation of Egg 306
V. Elongated Egg to Completion of First Cleavage 308
VI. Two-Cell Stage 309
VII. Four-Cell Stage 313
VIII. Eight-Cell Stage 315
IX. Sixteen-Cell Stage 316
X. Blastula 320
5. Summary 2>22
6. General Discussion 323
The general problem of the localization of developmental fac-
tors within the egg has received an important addition in a recent
paper by Professor E. B. Wilson, giving strong experimental evi-
dence of a progressive character in the localization of materials
in the egg of Cerebratuhis lacteus.^
The present paper is a description of a similar series of experi-
ments on the Mediterranean species, Cerebratuhis marginatus,
carried on at Naples during April and May, 1903, at the sug-
lExperiments on clevage and localization in the nemertine-egg. Archiv. f.
Entw. der Organismen. Bd. 16 Heft 3, 1903, pp. 411-460.
2 94 Charles Zeleny.
gestlon of Professor Wilson/ The special aim of the experi-
ments was two-fold. In the first place it was desired to throw
some light upon the character of changes in localization which
take place between the time of fertilization of the egg and the
completion of the first cleavage. Since a fragment of the unfer-
tilized egg segments as a whole while an isolated blastomere of
the two-cell stage segments as a half, fragments of the egg taken
at intermediate stages must yield interesting results. In the sec-
ond place a comparative study was made of the characteristics ex-
hibited by larvae developed from different portions of the egg
isolated at the eight-cell stage. The three portions thus compared
are the upper and the lower quartets obtained by a horizontal cut
and the lateral four-cell groups obtained by a vertical cut.
Clear results were obtained on these two points. For the first
it is shown that there is a progressive localization of the cleavage
factors between the time of fertilization and the completion of the
first cleavage. For the second a definite differentiation along the
polar axis of the egg is made out at the eight-cell stage. This
differentiation occurs in such a way that while a lateral four-cell
group remains totipotent, the upper and lower quartets are no
longer so, one lacking the ability to form an enteron and the other
the ability to form an apical organ.
2. Method.
The method of operation was a very simple one. The eggs
were placed on a glass slide and the water was withdrawn until
they were slightly flattened. The cut was made with a fine-bladed
scalpel under a dissecting microscope. The resulting parts were
then placed in individual dishes, where they were allowed to de-
velop. In fragments of the undivided egg the segments of the
sphere thus obtained retained their shape for several minutes, but
gradually assumed the spherical form. Fragments of unfertil-
ized eggs were fertilized after the spherical shape had been as-
il wish to express my great obligation to Professor Wilson for invaluable
advice during the progress of the experiments, and to the members of the
staflf at the Naples Zoological Station for continued kindnesses.
Localization of the Nemertine Egg. 295
sumed. This method of cutting was found to be very successful
for segmented eggs as well. Even at the eight-cell stage the upper
and lower or the lateral groups of fours may be separated in a
considerable percentage of cases without injury to the individual
blastomeres notwithstanding the interlocking of the cells.
3. Normal Development.
The orientation of the egg in Cerehratiihis marginatus, as in
C. lacteits, is made easy by the presence of a basal protuberance
before and for some time after fertilization and later by the pres-
ence of the polar bodies at the opposite pole. The protuberance
is still evident when the first polar body is formed, but usually dis-
appears at about the time of the formation of the second polar
body. A considerable difference was noted in the ability to with-
stand cutting at different periods. Before fertilization the egg
could be cut very readily, an extra-ovate being formed in relatively
few cases. After the first polar body had been formed, however,
the texture of the protoplasm seemed entirely different, the eggs
going to pieces in the great majority of cases immediately after
the cut was made. Again after the second polar body had been
formed the cutting property seemed much better, the quality of
the cut resembling that of the unfertilized egg. However, it is
very hard to draw any definite conclusion as to the comparative
texture of the eggs at these different stages because the amount of
flattening of the eggs on the slide, the sharpness of the scalpel
and practice in handling the latter may have had a great deal to
do with the cleanness of the cut. The general impression, leav-
ing out as far as possible these disturbing factors, is that the
protoplasm is much more liquid at the stage with one polar body
than that at either the unfertilized stage or the two polar body
stage. The maturation divisions seem therefore to be accom-
panied by a profound change in the nature of the cytoplasm.
When the egg is first removed from the body of the animal
there is a large germinal vesicle. This is usually situated on the
polar axis of the egg near the side farthest from the basal pro-
tuberance (Fig. 3, p. 301). The outline of the germinal vesicle
296
Charles Zeleny.
Localization of the Nemertine Egg. 297
Fig. I.
Early Cleavage of Entire Egg of Cerebratultis marginatv^.
A. two-cell stage; side view. B, four-cell stage; side view just after the
completion of the second cleavage. C, four-cell stage; from upper pole, slightly
later than B. D, eight-cell stage; from lower pole. E, eight-cell stage; side
view; the relations of the larger upper quartet to the smaller basal quartet are
shown in D and E. F, sixteen-cell stage; side view; the outline of a quadrant
is indicated by the heavier line. G, twenty-eight-cell stage; side view, slightly
from above; the first break in the rhythm of division is shown in the lagging
behind in each quadrant of the cell of the intermediate group which had been
derived from the basal cell of that quadrant. H, twenty-eight-cell stage; from
lower pole.
(The present figures, as well as all the following ones, unless otherwise
described, were drawn from preparations with the aid of the camera lucida.)
298 Charles Zeleny.
soon fades away and within half an hour the only sign of it is a
clear area which has collected near the pole opposite the pro-
tuberance. In this clear area is the spindle of the first polar di-
vision. The egg remains in this condition unless fertilized. In
the latter case the first and the second polar bodies are formed in
succession (Figs. 6, 7 and 9). The cell then elongates, exter-
nally a cleavage furrow appears in the vicinity of the polar bodies
and later another one at the opposite pole. These constrict the
egg Into two equal parts. The second furrow, at right angles to
the first, also passes through the polar bodies, and the two to-
gether divide the egg into four equal parts with no cross furrow
or very little indication of one. After the third cleavage, dexio-
tropic as usual, the cells of the upper quartet are distinctly larger
than those of the lower quartet. The cleavage goes on as a per-
fect Illustration of the spiral type. The cells of the eight-cell
stage give out smaller cells by leiotropic cleavages, taking place
simultaneously. The further divisions of the cells of the six-
teen-cell stage thus formed are not simultaneous. As in C. lac tens,-
the cell of the Intermediate group in each quadrant which had
been given off by the basal cell of that quadrant lags behind the
others, so that there is a distinct twenty-eight-cell stage before the
tardy cells divide to form the thirty-two-cell stage. The method
of cleavage as here described is very constant for normal whole
eggs, variations being extremely rare.
The character of the normal pilidium is too well known to need
any description here. The essential features of the early cleav-
ages are given in Figure i, p. 296, and three stages of the larval
development are shown In Figure 2, p. 300.
4. The Experimental Results.
0. Introduction. The experiments are described here in turn
according to the period at which the operation was performed.
Ten such periods are recognized:
1. Unfertilized egg.
2. Fertilization to complete separation of the first polar body.
3. First polar body to complete separation of the second polar
body.
Localization of the Nemertine Egg. 299
4. Second polar body to beginning of the lateral elongation of
the egg.
5. Elongated egg to the completion of the-first cleavage.
6. Two-cell stage.
7. Four-cell stage.
8. Eight-cell stage.
9. Sixteen-cell stage.
10. Blastula.
The limits of these periods are fairly well given in their titles
and further definition is added later under each head. It may
be stated that in every case where the operation was performed
after fertilization the sperm had been added to the eggs approxi-
mately half an hour after removal of the latter from the animal
so that the first polar spindle was already in the metaphase in all
cases before the entrance of the spermatozoon. This treatment
gave a greater uniformity to the relations of maturation and fer-
tilization than would otherwise have been possible and the exter-
nal evidences of internal change as given by the polar bodies serve
as good landmarks. The various groups of experiments include
both the cases observed for the cleavage factors and those for the
morphogenic factors. Those on the unsegmented egg were de-
signed mainly to bring out the cleavage factors; those on the two
and four-cell stages were intended both for cleav^age and morpho-
genic factors; while those on the eight and sixteen-cell stages and
blastulae were designed wholly for the morphogenic problems.
I. Fragments of Unfertilized Eggs. The cuts in this case
were made at periods ranging from a half hour to one and a half
hours after removal from the mother animal, the eggs being al-
lowed to lie in a dish of sea-water in the interval. Twenty-one
specimens were operated on, the cuts being made in the three
planes shown in Figure 3. Six of the cuts were horizontal, four-
teen vertical and one oblique.
The localization of cleavage factors (Figures 4A, B, C). In
neither of the three groups was there an indication of a localiza-
tion of the cleavage factors. The fragments segmented in the
regular manner described for normal whole eggs. There is no
cross furrow, or only a very short one, and the normal rhythm,
300
Charles Zeleny.
Fig. 2 (x 216).
Normal Larva Dez'chpcd from Whole Eggs.
A, larva of 233^ hours, showing the beginning of the archenteric invagina-
tion. B, larva of 34 hours, showing enteron, mesenchyme cells and apical plate.
C, larva of 49 hours, showing apical plate, mesenchyme cells, enteron, and one
of the two lappets.
size and position relations are preserved. The only abnormal case
is the irregular flat plate of seven cells shown in Figure 4C. When
therefore the cases are taken as a whole, the conclusion is very
evident that the experiments give no indication of a localization
of cleavage factors at this stage/
iJn connection with the early stages of cleavage the following subsidiary
points are to be noted:
1. The nucleated fragments formed polar bodies as in the whole egg, while
none were formed in the non-nucleated fragments.
2. No difference can be made out between nucleated and non-nucleated
fragments as regards character of the cleavage, each group showing similar
features as far as the limited data go.
3. Probable polyspermy, as indicated by multiple division, was shown in
two eggs.
4. The direction of the cut has no influence upon the character of the
cleavage. Of course, this necessarily follows from the conclusion as stated
above, that no localization of cleavage factors is shown in the group.
Localization of the Nemertine Egg.
301
Diagram of the egg just after removal from the animal, showing germinal
vesicle and basal protuberance. HH, VV, and OO indicate, respectively, the
directions of the horizontal, vertical and oblique cuts.
Fig. 4 (x 216).
Cleavage of Fragments of the Unfertilized Egg.
A, eight-cell stage, from nucleated fragment (r=4/5 of egg), obtained by a
horizontal cut; viewed from lower pole. B, sixteen-cell stage, from non-
nucleated fragment (^3/5 of egg), obtained by a horizontal cut; side view.
C, seven-cell embryo, from non-nucleated fragment (=2/5 of egg), obtained
by a vertical cut; viewed from convex side.
302 Charles Zeleny.
The localization of morphogenic factors. Only two fragments
were allowed to develop Into larvae, and neither of these showed a
sufficient differentiation for our purpose. One larva Is a solid
Fjg. 5 (x 216).
A, larva developed from a one-half vertical fragment of an unfertilized
egg; age 49 hours. B, eight-cell stage of a fragment of a fertilized egg ( — 2/3
of an egg), obtained by a vertical cut before the formation of the first polar
body; except for the unusual cross furrows the cleavage resembles a normal
whole one.
ciliated mass of cells and the other, shown In Figure 5A, has an
Internal cavity with a few free mesenchyme cells.
II. Fertilization to complete separation of the first polar body.
The limits include the period between the entrance of the sperma-
tozoon and the complete separation of the first polar body (Fig.
6). Seven eggs were operated on, but only two of these devel-
oped beyond the two-cell stage. The one clear case bearing on
the localization of cleavage factors (Fig. 5B) shows a typical
whole cleavage at the eight-cell stage. The morphogenic factors
receive no light from the one early blastula obtained.
III. First polar body to complete separation of second polar
body. The limits of this period are represented by the separa-
tion of the first polar body on the one hand and of the second
polar body on the other (Fig. 7), Nineteen eggs were operated
on, seven by horizontal, ten by vertical and two by oblique cuts.
The cases that bear on cleavage factors show In nearly every In-
stance some departure from the normal whole cleavage as regards
size, position or division rhythm of cells. The direction of the
cut seems, however, not to influence the character of the defect,
Localization of the Nemertine Egg.
V
303
V
Fig. 6 (x 180).
Diagram of Egg Soon After Fertilization and Before the Fai-mation of the First Polar
Body.
The dotted line at one pole incloses a clear area containing the first polar
spindle. The basal protuberance is still very prominent; HH, VV and OO
represent, respectively, the directions of the horizontal, vertical and oblique
cuts.
Fig. 7 (x 180).
Diagram of Egg with One Polar Body.
The basal protuberance is still evident, but not as prominent as in the for-
mer stages; HH, VV and OO represent, respectively, the directions of the
horizontal, vertical and oblique cuts.
at least as far as the present experiments go. The most common
irregularity is a simple departure from the normal division rhythm
(Fig, 8D), which in other cases is accompanied by a displacement
of some of the cells (Figs. 8 A, B, C). An interesting plate form
was obtained from a vertical fragment (Figs. BE, F, G). Its
special interest comes from the fact that notwithstanding the ir-
regularity it formed a SAvimming larva on the second day, 24
hours after fertilization.
304
Charles Zeleny.
Localization of the Nemertine Egg. 305
Fig. 8.
Fragments Obtained from Eggs with One Polar Body.
A, 8-i6-cell stage of a fragment (=14/5 of egg), obtained by a vertical cut;
side view; variations from the normal whole in rhythm of division, in size re-
lations and in position of blastomeres are to be noted. B, side view of same
egg from other side after a horizontal rotation through 180 degrees. C, same
egg viewed from lower pole. D, 8-i6-cell stage of the lower fragment ( — 6/7
of egg), obtained by a horizontal cut; side view; variations are to be noted in
rhythm of division and size relations of cells. Judging by the character of the
cut, the fragment probably contained the sperm nucleus without the egg
nucleus. E, 4-6-cell stage of fragment (=^3/5 of egg), obtained by a vertical
cut. F, 8-cell curved plate form of same. G, lo-cell curved plate form of
same. H, larva from fragment (=zupper 2/3 of egg), obtained by a horizontal
cut.
3o6
Charles Zeleny.
None of the three larvae which were allowed to develop showed
a sufficient differentiation of parts to be of service in determining
the localization of the morphogenic factors at this stage. The
one figured (8H) is a large blastula, the cavity of which is filled
with free spherical cells. The other two larvae, one of which was
just mentioned above, both developed after extremely irregular
cleavages, and are interesting because they show the presence of
an extremely high power of regulation. However, because of the
lack of data as regards their structure no definite conclusion can
be drawn even here.
IV. Second polar body to beginning of lateral elongation of
the egg. The period is limited on the one hand by the complete
separation of the second polar body from the egg and on the
other by the division of the cleavage nucleus and the accompany-
ing elongation of the cell preparatory to the first cell division
(Fig. 9). Eleven eggs were operated on, four by horizontal.
V
Fig. 9 (x 180).
Diagram of Egg zvith Tzvo Polar Bodies.
The basal protuberance has disappeared; HH, VV and OO represent,
respectively, the directions of the horizontal, vertical and oblique cuts.
four by vertical and three by oblique cuts. The fragments do not
segment as wholes, but shotv very evident departures from that
mode. However, there is a wide difference in the extent of this
departure in different cases. As determined by the experiments,
the range of the variation is from a possible whole cleavage,
through cases with a slight disturbance in size and position of
Localization of the N emertine Egg.
307
cells or rhythm of division (Figs. loA, B, C) up to a case with
an open cup-shaped blastula of a purely partial type (Fig. loE).
The results give no definite relation between the position of the
removed portion of the egg and the character of the resulting
defect in cleavage. A possible instance of such a relation is shown
in a sixteen-cell stage developed from a vertical-oblique fragment,
which show^s a corresponding flattening of one side of the embryo
(Fig. loC). ■
Fig. 10 (x 216).
Fragments Obtained from Eggs 7vifli Two Polar Bodies Before the Lateral Elonga-
tion of the Cell.
A, eight-cell ( — ) stage of fragment (3=2/3 of egg), obtained by a vertical
(slightly oblique) cut; oblique view. Note that the cells of the upper quartet
are smaller than those of the lower (the reverse of the normal condition) and
that one of the quadrants is behind the others in its division. B, 15-16-ceIl stage
of fragment (=1/2 of egg), obtained by an oblique cut; side view. Note that
the cells differ from normal whole ones in rhythm of division and size rela-
tions. C, i6-cell stage of a fragment (=22/;^ of egg), obtained by vertical-
oblique cut; side view. Note oblique flattening of egg. D, larva (50 hours
old) from fragment (:r=2/3 of egg), obtained by a vertical cut; side view.
Note solid enteron and absence of apical plate. E, open, partial blastula from
a fragment (3=2/3 of egg), obtained by a vertical cut; oblique view from
open side.
3o8
Charles Zeleny.
The localization of morphogenic factors is not particularly
elucidated by the two larvae obtained. The one represented in
Figure loD has a solid interior cell mass, evidently an archen-
teric ingrowth. There is no other structure sufficiently differen-
tiated for our purpose.
V. Elongated Egg to completion of first cleavage. The period
is limited on the one hand by the beginning of the lateral elonga-
tion of the egg and on the other by the completion of the first
cleavage. Five eggs were operated on. In every case the frag-
ments obtained show a partial cleavage from the start, even
though in several instances the cleavage furrow was slight at the
time of the operation, and there was still a broad connecting band
between the two parts of the egg. This band was in every case
equal to one-half or more of the diameter of a blastomere of the
two-cell stage. The two fragments from one of the eggs are
shown in Figures 1 1 A and B at the four-cell stage. It is evident
Fig. II (x 216).
Fragments from Eggs Bctzveni the Beginning of Lateral Elongation and the Coin-
pktion of the First Cleavage.
A, four-cell stage of fragment {=111/2 of egg), obtained by a vertical cut.
Polar bodies are attached. The second furrow is equatorial. B, four-cell
stage from the other half of the same egg.
that the cleavage is a partial one resembling closely that of iso-
lated blastomeres of the two-cell stage to be described later (p.
309) . One of the fragments has the two polar bodies still attach-
ed, and it is evident that the second cleavage furrow is equatorial
Localization of the N emertine Egg. 309
and not vertical as in the whole egg. The fragments from three
of the eggs developed into cup-shaped half blastulae again, re-
sembling the similar embryos arising from isolated blastomeres
of the two-cell stage. There is, therefore, at this period a defi-
nite localization of cleavage factors.
As regards the localization of the morphogenic factors no gen-
eral statement can be made. The two larvae obtained did not
show a sufficient differentiation to be of value.
In the experiments on unsegmented eggs a study of the locali-
zation of the cleavage factors has been the main object in view,
the few and unsatisfactory isolated observations on larvae develop-
ing from the fragments being incidental and subsidiary to the
main point. In the following experiments, however, the study
of the localization of the morphogenic factors is definitely taken
up, the most extended series and the one yielding the most inter-
esting results being on the eight-cell stage. The experiments on
the localization of the cleavage factors are continued for the two-
cell and four-cell stages.
VI. Two-cell stage.
I. Experiments on the localization of the cleax-age factors in
isolated blastomeres. The blastomeres were isolated in twenty-
eight eggs of the two-cell stage. In the majority both blastomeres
segmented, a minority showing no cleavage of one of the parts.
In nearly every case the cleavage could be recognized as a partial
one corresponding with that of a lateral half of the whole egg.
At the four-cell (8/2) stage there is a wide cross furrow and the
cells are not in the same plane. In fact they appear very much
as if they had been removed from the whole eight-celled embryo
by a vertical cut (Figure 12 A, B). The different forms of cleav-
age described for isolated blastomeres of the two-cell stage of
C. lact-ens by Professor Wilson were found here also. Their
characteristics are especially prominent during the eight-cell
(16/2) and the sixteen-cell (32/2) stages (Figs. i2AtoI, 13A).
The most numerous are the cup-shaped embryos resembling a
geometrical half of a w^hole blastula of the corresponding age
(Figs. 12I, 12H). On the one hand the cups are replaced by
3IO
Charles Zeleny.
Fig. 12 (x2i6).
Cleavage Stages of Isolated Blastomeres of the Two-Cell Stage.
A, four-cell stage from an isolated blastomere. Note wide cross furrow;
also, that the blastomeres are not in one plane. B, four-cell stage. C, eight-
cell stage. D, eight-cell stage of plate type. E, eight-cell stage of slightly-
curved plate type; viewed from convex side. F, eight-cell stage of curved
plate type; side view. G, eight-cell stage of very shallow cup type; viewed
from convex side. H, eight-cell stage of shallow cup type. I, 16-32-cell stage
of cup shaped type (i^geometrical half of normal blastula) ; side view.
Localization of the Nemertine Egg. 3 1 1
flat plate-like forms, there being all gradations between the em-
bryos curved into a deep cup through those showing only a slight
curvature up to perfectly flat plates (Figs. 12D to G and 13 A).
On the other hand there is a similar graded series from the cup
forms up to perfectly closed spherical half-embryos usually con-
taining a very small cavity or none at all (Fig. 12C). It seems
probable that these differences of form are the result of slight
changes in the surface tension relations between the cells, as Pro-
fessor Wilson has suggested, and this view is strengthened by
my observation of the development of a plate form and a cup
form from the two blastomeres of a single egg.
2. Experiments on the localization of the morphogenic fac-
tors. The localization of the morphogenic factors in the two-
cell stage was not made out as fully as could have been wished.
The larvae were killed in most cases at too early a stage to deter-
mine the necessary difi^erentiation of organs. The two blasto-
meres were separated in each of sixteen eggs and in thirteen larvae
were obtained. Most of these were about 33 hours old when
killed, only three being older than this. The thirteen individuals
are divided into groups of similar cases in the following descrip-
tion.
In three cases observations were made on the activity of the
embryos, but the embryos themselves were lost during transfer-
ence to the preserving liquid. An interesting fact in connection
with these, and this holds also for other one-half as well as one-
fourth embryos, is the abnormally great rapidity of rotation in
most of the cases.
Another group is formed by isolated blastomeres from five eggs.
The larvae were distinguished by rapid rotation in life and by a
dense ingrowth of cells from one pole, which entirely filled the
blastocoele and came close up against the ectoblastic wall around
the whole surface of the egg (Fig. 13B, C, D). No apical plate
was made out in any of them, but in one case there was a single
lappet (Fig. 13B).
312
Charles Zelen
i
Fig. 13 (x 216).
Cleavage and Larval Stages from Isolated Blastonieres of the Tivo-Ccll Stage.
A, nine-cell stage (rrrcup sbaped type); view from concave side. The cut
was made at one side of cleavage plane so that the fragment included one
blastomere plus part of the other. B, larva (age=z:33^ hours). Note single
lappet, solid enteron and absence of apical organ. C, larva (age=:33 hours).
Note solid enteron and absence of apical organ. D, larva (age^33^ hours).
E, larva (age=:r33 hours). F, larva (agez=33^ hours). G, larva (age=:33
hjours). The blastomeres were not completely separated and may have fused.
H, larva (age=47 hours). Note apical organ and small enteron.
In three cases the embryos resemble the five just mentioned, ex-
cept that the archenteric mass is not as large and a slight blasto-
coele, crescentic in vertical section, is present (Figs. 13E, F) .
This blastocoele contains rounded and irregular mesenchyme cells.
There is no apical organ. The larvae do not differ widely from
the normal larva of about 24 hours though their age is 33 hours.
Localization of the Nemertine Egg. 313
In one of the two remaining cases the blastomeres were not
completely separated. The result was two connected partial em-
bryos in the early stages, which evidently later fused to form a
single individual. The resulting larva (age 33 hours) shows a
large blastocoele, two apical organs and a solid enteric mass grow-
ing in at the base. The blastocoele contains free rounded cells,
and there are no lappets (Fig. 13G).
Finally there is the one case which was allowed to develop for
a sufficient length of time (47 hours) to give the organs a chance
to differentiate. The resulting larva (Fig. 13H) rotated rapidly
in life. It has a large blastocoele, a small enteron, an apical plate
and a thickening in the wall at the side of the mouth opening,
probably representing the basis of the ectodermal invagination at
this point. There are no lappets. With the exception of the small
size of the enteron and the absence of the lappets, the larva does
not differ widely from a normal larva.
Summary of the results on the localization of morphogenic fac-
tors. The larvae developed from isolated blastomeres of the two-
cell stage do not show any constant defects except possibly as re-
gards the lappets, organs which in C. marginatus are developed
at a comparatively late period. Of the instances here cited only
two can be considered as old enough to have formed the lappets.
At any rate we must consider the larva developed from an iso-
lated blastomere of the two-cell stage to be retarded in develop-
ment as compared with a normal one of the same age, though this
view does not serve to explain completely the characteristics of
several of the larvae.
VII. Four-cell stage. The experiments at this period come
under two heads. In one series the segmenting egg was divided
into two groups of two cells each, and in the other the four blasto-
meres were isolated.
The isolated blastomeres segment in every respect as quadrants
of the whole egg. It will be remembered that the whole egg of
Cerehratulus goes through a definite twenty-eight-cell period be-
cause one of the cells of each quadrant lags behind the others in
its division as the egg passes from sixteen to thirty-two cells (see
Fig. iG). Correspondingly the isolated blastomere of the four-
314
Charles Zeleny.
cell stage passes through a definite seven-cell (28/4) stage. Such
a stage is represented in Figures 14B and C.
Fig. 14 (x 216).
Cleavage Stages and Larva from Two-Cell Groups and Isolated Blastomcres of the
Four-Cell Stage.
A, seven-cell stage from fragment (two cells+) of tgg. B, seven-cell
stage from isolated blastomere of an egg. C, seven-cell stage from other
isolated blastomere of egg shown in B. D, larva (ager=33 hours) from two-
cell group. Note apical plate, solid enteric invagination and large balstoccele
with numerous free cells. E, larva (age=r33 hours) from isolated blastomere.
Note the very large solid enteron nearly filling the blastocoele.
Ten eggs were used for the experiments on the localization of
the morphogenic factors. The only larva developed from a two-
cell fragment was asymmetrical and swam in a small circle.
Thirty-three hours after fertilization it had an apical plate and
cilia, the beginning of the ingrowth of the archenteric mass and a
large blastocoele containing rounded and irregular mesenchyme
cells (Fig. 14D). There is thus no definite specification of the
morphogenic factors in a two-cell group of the four-cell stage.
In six cases the four blastomeres were isolated, and in five of
them very rapidly rotating larvae resulted. The observations on
these were made in most cases thirty-three hours after the fertili-
Localization of the Nemertine Egg. 315
zation. The larvae resemble very much the solid larvae of the
isolated blastomeres of the two-cell stage, but are of smaller size
(Fig. 14E). All the one-fourth larvs are like the one figured.
There is a solid archenteric growth, but no sign of an apical organ.
The larvae from isolated blastomeres of the four-cell stage,
therefore, give no indication of a definite localization of the mor-
phogenic factors, though they do not develop in an entirely nor-
mal manner. The latter statement holds also for the larvae de-
veloped from the isolated blastomeres of the two-cell stage, as has
already been stated (p. 311).
VIII. Eight-cell stage. The results yielded by the experiments
on this stage are perhaps the most important of all those given.
The group of experiments included sixty-eight four-cell groups.
These groups were separated by a careful cut with the fine scalpel
blade used in all the experiments. In most cases the knife blade
passed between the cells, and the latter were entirely uninjured by
the operation. In a few, however, the protoplasm was cut, and
these will be mentioned in the descriptions. The operations in-
clude a series of horizontal cuts separating the upper from the
lower quartet, and a series of vertical cuts separating the two
lateral four-cell groups, each of the latter containing two cells of
the upper and two of the lower quartet. There are thus three
kinds of four-cell groups, the larvae from which are to be com-
pared: (i) Upper quartets, (2) lower quartets, and (3) lateral
four-cell groups. The experiments yield a very definite and posi-
tive result. The larva developing from the upper quartet have
an apical organ, but no archenteron, those from the lower quar-
tet have an archenteron, but no apical organ, while those from
lateral four-cell groups have both apical organ and archenteron.
The natural conclusion '4o be drawn from these results is that
certain organ-forming materials are definitely separated by the
third cleavage plane, and the larvae developing from the lower
or the upper quartet have not the power of making up the lack-
ing material. The lateral four-cell groups, howev^er, possess both
kinds of materials and are, therefore, able to develop both archen-
teron and apical organ, though the larvae are usually asymmet-
rical. ,
3i6 Charles Zeleny.
Some of the larvae are shown in Figures 15, 16 and 17, and it
will not be necessary to describe the individual cases in detail, as
the results are very definite and clear. The figures give charac-
teristic types of larvae developing from the upper quartet (Fig.
15), from the lower quartet (Fig. 16), and from lateral four-cell
groups (Fig. 17A, B, C). Figure 17D shows a case in which
six of the eight cells were represented, two of the lower quartet
having been destroyed.
IX. Sixteen-cell stage. Five eggs of the sixteen-cell stage suc-
cessfully withstood an operation, and larvae from three of these
were studied.
In one egg equal upper and lower portions were obtained by a
horizontal cut, but there was not a separate identification of them,
and they were placed in one dish. At forty-eight hours after fer-
tilization both resultant embryos were ciliated. They showed a
difference in that one had ragged edges and swam in a circle,
while the other had even edges and remained stationary. The
embryos were lost.
In two cases the upper four cells were successfully separated
from the lower twelve. The two cases are taken up in turn. In
the first one the division was very clear without injury to any of
the cells. At forty-six and a half hours after fertilization the
upper four cells had 'formed a small distinctly outlined spherical
ciliated embryo, with no rotation or forward motion of the body.
There is a distinct blastocoele containing rounded cells, a large
apical organ and no enteron or lappets (Fig. 18C). At the same
time the lower twelve cells have formed a ciliated rotating embryo,
with a large solid archenteron entirely filling up the cavity of the
blastocoele. Neither apical organ nor lappets are present. The
two embryos thus show a very pronounced difference, the one
formed from the upper four cells containing an apical organ and
no archenteron, and the other, from the lower twelve cells, con-
taining an archenteron and no apical organ.
In the remaining case the upper four cells were separated from
the lower twelve as before. One cell in the former was slightly
injured, but all the cells of the latter were left in good condition.
From the upper four cells at forty-six and a half hours after fer-
Localization of the Nemertine Egg.
317
Fig. 15 (x 216).
Larva from U[^pcr Quartets of the Eight-Cell Stage.
A, larva (age^44 hours) from complete upper quartet. B, larva (age
=:4S hours) from upper quartet, with one cell injured. C, same larva rotated
horizontally. D, larva (age^45 hours) from complete upper quartet. The
egg already showed the cell constrictions for the next (i6-cell) division. E,
same larva rotated horizontally. F, larva (age:=45 hours) from complete
upper quartet. The inner cell mass does not connect with the side of the
larva (i. e., it is free). G, larva (ager=23 hours) from complete upper quartet.
H, larva (age:=33 hours) from isolated blastomere of the upper quartet. I.
double larva (age=46 hours) from upper (?) quartet. Note presence of apical
organ and absence oi enteron in Numbers A to F.
3i8
Charles Zeleny.
Fig. i6 (x 216).
Larvcc from Lower Quartets of the Eight-Cell Stage.
A, larva (age:i=46 hours) from lower (?) quartet. Note large enteron and
absence of apical organ. B, larva (ager=46 hours) from lower quartet. C,
same larva rotated horizontally: Note solid enteron and absence of apical
organ. Ahte large archenteric ingrowth and absefice of apical organ in all cases.
Fig. 17 (x 216).
Lan'cc from Portions of the Eight-Cell Stage.
A, larva (age=:48 hours) from lateral four-cell group. Note presence of
both enteron and apical organ. B, same larva rotated horizontally. C, larva
(age=r:33 hours) from lateral four-cell group. Note presence of both enteron
and apical organ. D, larva (age^48 hours) from upper quartet plus two cells
of lower quartet. Note three apical organs, large blastocoele, small enteron
and two ectodermal invaginations at sides of enteron.
Localization of the N emertine Egg.
319
Fig. 18 (x 216).
Larvce from Portions of the Egg oi tin Sixtecn-Ccll Stage.
A, larva (agc:r=46j^ hours) from the upper four cells (one injured).
Note presence of apical organ and absence of enteron. B, larva (age^46y2
hours) from the lower twelve cells of the same egg. Note that egg is a solid
mass with division into ectodermal and endodermal cells. C, larva (age 46^4
hours) from upper four cells. Note presence of apical organ and absence of
enteron. D, larva from lower twelve cells of the same egg; same age. Note
solid enteron and absence of apical organ.
tilization an elongated very actively swimming larva with a long
apical cilium had developed. The embryo showed after stain-
ing and mounting a well developed apical plate. The anterior
end of the body is occupied by a blastocoele containing a few scat-
tered free cells. The posterior end is a dense mass of cells, with
no signs of an ingrowth of these to form an archenteron (Fig.
18A). From the lower tweh^e cells at the same time there was
developed a ciliated elongated embryo, with only a slight rotation,
and no forward movement of the body. The embryo is a solid
mass of cells, the only differentiation visible being a difference be-
tween the cells at the two ends. Those near one end have the
typical histological endoderm characters of the normal larva,
while those near the other end have ectoderm characters (Fig.
18B).
The characters of the two larvae in this case again show the
presence of the apical-basal differentiation described for the last
320 Charles Zeleny.
specimen. The experiment seems to indicate that the basis of the
apical organ is found in the four upper cells of the sixteen-cell
stage. In connection with this result Yatsu's observations on the
unsegmented egg of C. lacteus are interesting. He localizes the
basis of the apical organ in a broad band just above the equator
of the egg.
X. Blastula stage. Successful operations were made on three
blastulae.
The first one was divided by a horizontal cut into an upper part
( = 2/3 of blastula) and a lower part ( = 1/3 of blastula). The
orientation was made certain by the presence of the polar bodies.
The upper part broke up into two portions, each of which at
twenty-four hours had developed into an embryo with an apical
cilium. At the same time the embryo from the lower one-third
of the blastula was ciliated but had no apical organ. At forty-
seven and a half hours the embryo from the lower one-third and
one of the upper ones were dead. The other upper embryo has
two apical plates, one a well developed and the other a small one,
an invaginated ectodermal sac, a large and well developed en-
teron, a blastocoele with free cells in its cavity, and no lappets.
In fact, except for the absence of the lappets and the presence of
two apical organs, it has all the characters of a typical whole
larva (Fig. 19C; only one of the apical plates is shown). How-
ever, at twenty-four hours, as stated above, there is a distinct dif-
ference between the upper embryos and the lower one because of
the presence of the apical organ in the former and its absence in
the latter.
A second blastula was cut into equal upper and lower parts by
a horizontal cut, but the two were not kept separate. One of the
halves died. The other developed all the organs of the normal
pilidium, except the lappets. There is a large blastocoele, two
apical organs, one in the normal position and one asymmetri-
cally placed and not shown in the figure, and a large long enteron
straighter than in the normal larva (Fig. 19B).
A third blastul-a was cut into two unequal parts equal respec-
tively to two-thirds and one-third of the blastula, by a cut of un-
known direction. One portion, the larger one judging by the
Localization of the Nemertinc Egg.
321
Fig. 19 (x2i6).
LarvcE Developed from Blastula Fragments.
A, larva (age 351^^ hours) from a fragment (1^:2/3 of blastula) ; direction
of cut is not known. B, larva (age 3514 hours) from upper or lower half of
blastula. Note presence of both enteron and apical organ. C, larva (age3r48
hours) from the upper 2/3 of blastula. Note presence of both enteron and
apical organ.
size, was the only one alive thirty-five and a half hours after fer-
tilization. Its cilia were waving, but there was no motion of the
animal as a whole. The body is spherical, with a large solid en-
teron nearly filling the cavity of the blastocoele. The chief defect
is in the absence of the apical organ, as it is too early (35/^
hours) for the lappets to appear (Fig. 19A). The direction of
the cut is not known and the defect, therefore, cannot be corre-
lated with any definite portion of the blastula.
The experiments on blastulae give only one organ which can be
considered as definitely specialized. The apical plate is developed
in each of two embryos from the upper two-thirds of a blastula,
while it is absent in those developed from the lower one-third.
No explanation can be given of the apparently greater regulative
322 Charles Zeleny.
€
power of blastula fragments as compared with those of the eight-
cell and sixteen-cell stages. A similar fact was noted by Professor
Wilson for C. lacteus, and he supposes that a possibility of error
in orientation of the blastulae may account for the result. For this
reason I took special care in determining the orientation, and in
two of the three cases I think there is little doubt of the correct-
ness of the determination.
5. Summary of results.
1. Both nucleated and non-nucleated fragments of the unfer-
tilized eggs of Cerebratulus marginatus segment as wholes.
2. Isolated blastomeres of the two-cell stage segment as if the
other blastomere were still in its place, i. e., they segment as ver-
tical halves.
3. Fragments obtained during the stages between the fertili-
zation of the egg and the completion of the first cleavage show a
progressive specification of the cleavage factors as evidenced by
abnormalities in rhythm of division, size relations of cells and
position of cells. After the separation of the cleavage nuclei and
when the cytoplasm of the two cells is still widely connected, the
two halves when cut apart may already show all the characters of
half cleavages.
4. Isolated blastomeres of the four-cell stage segment as
fourths of the whole cleavage pattern.
5. Larvae developed from the upper quartet of the eight-cell
stage always possess an apical organ and lack an enteron, those
developed from the lower quartet always possess an enteron and
lack an apical organ, while those developed from lateral four-
cell groups containing two cells of the upper and two cells of the
lower quartet always possess both apical organ and enteron.
6. Larvae developed from the upper four cells of the sixteen-
cell stage lack an enteron, but possess an apical organ and blasto-
coele. Those developed from the lower twelve cells have a large
enteron, but no apical organ or blastocoele.
7. Two embryos developed by a secondary division from the
upper two-thirds of a blastula both developed apical organs. The
Localization of the Nemertine Egg, 323
embryo developed from the lower one-third of the same blastula
developed no apical organ.
6. General discussion.
The points brought out by the present experiments are of con-
siderable general interest. In the first place in agreement with the
results of Professor Wilson on C. lacteus, it is found that while an
egg fragment of an unfertilized egg segments as a whole an iso-
lated blastomere of the two-celled stage segments as a half. In the
intermediate stages there is a gradually increasing departure from
a whole cleavage in the fragments as we pass from the first men-
tioned stage to the latter. This is contrary to the statement made
by Yatsu in his recently published paper on C. lacteus. In prin-
ciple, however, it agrees with the progressive increase of defects
found by him in larvae developing from fragments taken at simi-
lar stages.
Though the observations naturally suggest the view that there
is a progressive localization of materials in the egg from one
period to the other, such a conclusion does not necessarily follow
from the experiments themselves without further data. Because,
considering the power of regulation of the embryo shown at all
stages studied, it must be admitted that there remains the possi-
bility of regulation of the unfertilized fragment to form a com-
plete whole cleavage and later a complete larva. For the earlier
the operation be performed the greater the time which must elapse
before the fragment divides, and consequently the greater the
chance for regulation to a whole cleavage pattern. The experi-
ments of Schultze on inversion of whole frog's eggs at the two-
celled stage and the corresponding ones of Morgan on the isolated
blastomeres of the same stage, show that the rearrangement of
materials due to difference in specific gravity gives opportunities
for regulation to a whole development. Observations on the nor-
mal eggs of a great many animals during the maturation period
show a very extensive series of streaming movements in the proto-
plasm at this time. May not these furnish a similar opportunity
for regulation to a whole cleavage and whole development? For
324 Charles Zeleny.
as the materials In the Isolated blastomere of the frog's egg are
undoubtedly specialized so as to form a half cleavage pattern and
a half embryo under the ordinary conditions, the re-adjustment of
materials due to Inversion gives the necessary conditions for regu-
lation. May not the unfertilized egg of Cerebratulus likewise
show a localization of formative factors so that a fragment is a
true portion of a mosaic, but needs only the conditions accompany-
ing the streaming during the maturation stages to accomplish a
readjustment to a whole? Evidently there Is no means of de-
termining this point, because If we assume the possibility of a re-
adjustment during the maturation stages, we remove our only hope
of a direct method of deciding the question as to the presence or
absence of developmental specification before fertilization. The
only remaining method lies In Indirect Inferences from the ob-
served localization of visible materials during these stages. There
Is abundant proof of such a progressive localization during this
time, and the conclusion that there is an arrangement of forma-
tive materials Into a definite pattern at this period Is a natural one.
For there can be no valid objection to the association of the two
parallel processes of localization of visible materials and of re-
sultant cleavage and morphogenic factors.
But why Is there a progressive localization of the morpho-
genic factors In the unsegmented egg, as Indicated by Yatsu's
work, while at the same time the Isolated blastomere of the two-
celled stage develops Into a perfect larva ? The progressive local-
ization of cleavage factors as shown in my experiments Is naturally
to be expected, since there is a gradual passage from a whole
cleavage on the one hand to a half cleavage on the other. We
may assume a gradual localization of materials controlling these
factors, or a greater opportunity for regulation In the earlier as
compared with the later stages,' or both, to account for the data.
With the progressive localization of morphogenic factors, as de-
scribed by Yatsu, there Is no such sequence. Starting with com-
plete larvae developing from the fragments of unfertilized eggs,
there is a gradual increase in the defects in the larvae up to the
completion of the first cleavage. Then very suddenly, as soon as
the cleavage is completed, there Is a return to whole larvae. I
Localization of the Nemertine Egg. 325
think the apparent contradictions may be explained in the fol-
lowing way, though the purely speculative character of all the
discussions is always to be kept in mind.
Numerous recent observations, especially those of Lillie and
Conklin, indicate that cleavage is an accurate means of separating
materials already localized. My experiments on the eight-celled
stage of Cerebratiihis show that the first localization of materials
is in an apical-basal (polar) direction. It is probable, therefore,
that at the four-cell and even at the two-cell stage this same ten-
dency is the predominant one, so that at the four-cell stage there
are four equivalent parts and at the two-cell stage two equivalent
parts. However, in each of these two parts (taking the two-cell
stage as an example) there is an apical-basal differentiation. A
separation of a blastomere at this stage causes at first a half cleav-
age, but the materials retain a relation to each other very similar
to the normal whole relation as far as the apical-basal axis is con-
cerned along which differentiation is assumed; the embryo, there-
fore, can readily adjust itself to form a whole larva, having all
the necessary materials present in the proper relations.
In the fragment of the unsegmented egg this is not true. Here,
according to all indications, there is a great activity in the mate-
rials of the egg. If the egg is cut at an early stage (as in the un-
fertilized egg) there is yet a considerable period of activity and
movement of materials through which the egg must pass before
the first cleavage takes place ; and, therefore, on the one hand a
whole cleavage results, and on the other a normal whole larva
is formed. Such is not the case, however, if we take the egg for
instance after maturation not long before the cleavage. The egg
is nearly ready for the first cleavage, the materials are arranging
themselves for an equal distribution and the proper physical ten-
sions for such a division are present. The egg is now cut and a
portion of it removed. The cleavage ensues very quickly, for the
physical machinery was already starting to act when the cut was
made. The different materials are not separated in a precise way,
even if the cut is vertical, for the cytoplasm is semi-liquid, and
the materials, especially along the injured side, come into abnor-
mal relations with each other, which cannot be regulated as in
326 Charles Zeletiy.
the earlier stages, because of the lack of the opportunity which,
as stated above, is afforded the earlier ones {c. f. again the ex-
periments on the frog's egg). The defects in the larvae have a
definite relation to the position of the removed part of the egg
because the disturbance of the protoplasm is greatest in the region
near the cut. The already differentiated materials may thus be
separated in an abnormal relation to each other and become un-
naturally grouped by the cell walls of the ensuing divisions. In
this manner, on the one hand the abnormality of the resulting
cleavages, and on the other the defects in the larvze developed
from the fragments, may be explained. The first normal cleav-
age, however, divides the cell into two similar parts, each of
which retains a relation between its differentiated materials very
much like that of the whole egg. The conclusion is therefore
reached that the relations of the materials in the isolated blasto-
mere of the two-cell stage are more normal (/. e., more like those
of the whole egg) than are those of fragments of the two-polar-
body stage, and therefore the capacity of regulation to form a
whole larva is greater in the former than in the latter. The
mechanism of division which is disturbed by the cut in the un-
segmented egg, and is capable of regulation if the cut is early
but is disturbed if the cut is late, is also not disturbed in the iso-
lated blastomeres of the two-cell stage, the cell division goes on
as if the other blastomere were present, and a partial (one-half)
cleavage results.
It therefore seems probable that while in normal development
cleavage is an aid in differentiation, in development after re-
moval of a portion of the unsegmented egg (or segmented egg)
it is a distinct detriment in so far as the attainment of the normal
relations of the parts is concerned. For while on the one hand it
isolates materials and allows a more accurate differentiation, on
the other it restricts the power of regulation of the organism.^
^Yatsu has hinted at an explanation somewhat similar to the above. He
says that the differences between the lai'vse derived from fragments of the un-
segmented egg and those from isolated blastomeres of the two-cell stage may
be due to dififerences in the accuracy of the separation of the materials in the
two cases.
Localization of the N emertine Egg. 327
While the larvae developed from Isolated blastomeres of the
two and four-celled stage show certain organic differences from
the normal whole larvse, there is no indication in them of a specific
local defect. The first trace of such a defect is reached in the
eight-cell stage. Here, while larvse developed from lateral four-
cell groups containing two cells of the upper quartet and two cells
of the lower show the characters of a normal larva (except for
asymmetry in arrangement) larvae from the upper quartet al-
ways possess an apical organ and lack an enteron, and those from
the lower possess an enteron and lack an apical organ. There is
thus a very distinct differentiation along the apical-basal (polar)
axis. It is an interesting fact that this first distinctive morpho-
genic localization is coincident with the first inequality in cleavage,
the Inequality being In the same direction. Projecting backward
this differentiation — that Is, assuming that an apical-basal differ-
entiation has been going on for some time before the eight-cell
stage — naturally there would be no indication of It In the Isolated
blastomeres of the two or four-cell stages because the cleavages
are vertical. Likewise there may be a similar differentiation in
the unsegmented egg; for, while my results on cleavage defects
cannot be analyzed as showing any specific relation to the indi-
vidual kinds of egg defects, the observations of Yatsu on morpho-
genlc defects do show such a relation.
The experiments on the eggs of Cerebratuhis marginatits, to-
gether with the former ones on C. lacteus, seem therefore to indi-
cate that at the eight-cell stage the formative materials of the
egg are definitely localized in an apical-basal direction, and the
experiments of Yatsu on morphogenic defects In larvae resulting
from unsegmented eggs of the later maturation stages show a
similar apical-basal differentiation. That this process of apical-
basal differentiation is a progressive one In the unsegmented egg
Is Indicated by the whole character of the cleavage in fragments
of unfertilized eggs, and by the progressive departure from this
character up to the first cleavage, and the corresponding increase
In defects of larvae developed from the fragments, though In the
latter case the continuity of the result seems to be masked by the
development of whole larvae from isolated one-half and one-fourth
328 Charles Zeleny.
blastomeres. An explanation of this has, however, been offered.
The first two cleavages being perfect apical-basal ones, the iso-
lated blastomeres cannot be expected to show other than perfect
larvae, assuming only a slight regulation overcoming lateral
asymmetry, notwithstanding the partial cleavage, which after all
is only quantitatively partial (=^ or 34 of a pattern). At the
same time the fragments of unsegmented eggs can never be said
to contain the materials divided accurately with respect to an
apical-basal axis because, in the first place, the cut is never per-
fectly vertical, and in the second place, the rounding in of the
edges after such a cut causes a disarrangement of the materials
which must result in unequal distribution at the first cleavage. The
ability to regulate such an unequal distribution must of course
largely depend upon its extent and character. The greater op-
portunity given to fragments of unfertilized eggs to regulate such
differences in distribution (if any) before cleavage takes place
may to some extent explain the whole character of the cleavage
in such fragments without the assumption of a perfectly Isotropic
egg, an assumption which is contradicted by the evident polarity
of the egg at this period as indicated by the eccentricity of position
of the nucleus and the presence of the basal protuberance. The
existence of such an apical-basal differentiation in the unsegmented
egg was indeed already indicated by Professor Wilson's result on
certain eggs in which the basal portion was removed by a hori-
zontal cut and which showed the basal quartet of the resulting
eight-cell stage much smaller in comparison with the upper than
In the normal whole egg.
The data on localization of formative factors in the egg be-
fore cleavage and during the early segmentation stages may,
therefore, be provisionally stated In the following form for Cere-
bratuliis, If an Intimate relation is assumed between localization
of visible materials and localization of formative factors. The
unfertilized egg before the beginning of maturation already shows
evidences of a polarization which necessitates the assumption of
a heterogeneity In material. Upon this basis, and In the same
apical-basal direction, later differentiation proceeds.
During the preliminary maturation stages and after fertlliza-
Localization of the Nemertine Egg. 329
tion there are profound changes in the distribution of materials
in the egg, and these changes seem to be accompanied by an in-
creased apical-basal differentiation. The first two cleavages being
vertical and equal merely effect a quantitative and not a qualita-
tive separation of materials, but the third plane of division, a
horizontal one, bringing about an unequal division, separates the
egg into two qualitatively different parts. That such is the case
is absolutely demonstrated by my experiments on the eight-cell
stage in which I obtained complete larvae from lateral four-cell
groups, larvse with an apical organ but without an enteron from
the upper quartet and larvae with an enteron but without an apical
organ from the lower quartet.
Hull Zoological Laboratory, University of Chicago.
February 29, 1904.
AN EXAMINATION OF THE PROBLEMS OF PHYSIO-
LOGICAL "POLARITY" AND OF ELECTRICAL
POLARITY IN THE EARTHWORM.
T. H. MORGAN AND ABIGAIL C. DIMON.
The so-called "polarity" exhibited in the regeneration of ani-
mals has suggested the idea to a number of writers that the phe-
nomenon might be related to, or the outcome of differences in
potential in different regions; or, in other words, of electrical polar-
ity. The term "polarity" itself, which has been generally adopted
to express a sort of stereometric relation in the regeneration of
living things, suggests in certain striking ways the polar relations
observable in many electrical phenomena, and invites a direct com-
parison between the two.
The only experiments that have been undertaken to test directly
this question are the recent ones by Mathews'^ on certain hydroids
and on the tail of Fundulus. The interesting results reached by
Mathews, while leaving the problem, so far as the main issue is
concerned, still an open one, showed the importance of .further
examination of the subject. Mathews avoids, it is true, making
a direct comparison between physiological "polarity" and the
polarity present in electrical phenomena, and speaks rather of the
rate of growth of certain regions in comparison with others; but
if there is in reality any fundamental relation between the phe-
nomena in question, we should expect to find some expression of
it in the polar, or, more generally, the stereometric relations of
the parts.
If, for instance, the development of a head at the anterior end
of a piece and of a tail at the posterior end is connected with dif-
ference of potential in the two regions, we might hope to get evi-
iRlectrical Polarity in the Hydroids. A. P. Mathews. Am. Journ. Physi-
ology, Vol. VIII, No. IV, Jan. i, 1903, pp. 294-299.
332 T. H. Morgan and Abigail C. Dimon.
dence of this by means of the galvanometer. If this relation
should be found to exist, there is a further opportunity of test-
ing the validity of the conclusion in the case of axial heteromor-
phosis. For this reason we selected the earthworm for our study,
since in the earthworm it had been shown by one of us that there
regenerates from the anterior cut surface of posterior pieces not a
head, but a tail. We should expect to find under these circum-
stances a reversal of the potential in this region,* when compared
with an anterior cut surface in the more anterior regions of the
worm. The following pages give the results of our examination.
METHODS.
Both Lumbricus terrestris and Allolobophora foetida were used.
Since the results appeared to be similar for both species, Lum-
bricus, being larger and showing on the whole greater differences
of potential, was preferred when available. In all, sixty-four
worms were used, the number of tests made upon any one varying
from one to seventeen. Differences of potential were detected by
a Rowland-d'Arsonval galvanometer, connected with a pair of
non-polarizable electrodes. Since the regulation of these electrodes
was found to be troublesome, and since they were liable to intro-
duce a source of error into the readings, their manufacture and
regulation had best be described. Two glass tubes about three
inches long, somewhat smaller at one end, were plugged at the
small end with kaolin or filter paper, moistened with normal
(0.85%) salt solution. The plug extended about a half Inch be-
yond the end of the glass tube. When made of kaolin, the end,
after being used, could be easily broken off and re-formed from
fresh material. When made of filter-paper, the tip, if it became
contaminated from touching the worm, could be cut off. Above
the plug each tube contained a saturated solution of zinc sulphate,
into which projected through a cork a small amalgamated zinc
electrode, connected by a wire with one of the poles of the gal-
vanometer. Since the instrument was so sensitive that the slight-
est loss of equilibrium was at once registered by a deflection of the
mirror, great care had to be exercised in keeping the junctions of
wires and zinc electrodes dry, and in balancing the other elements
Physiological ''Polarity" and Electrical Polarity. 333
of the electrodes. As the electrodes were found to deteriorate
rapidly when used, it was necessary to examine them at frequent
intervals, and when the deterioration became so great as to affect
seriously the value of the readings on the worm, to readjust the
parts. This could be accomplished in several ways, of which some-
times one and sometimes another was most effective. A fresh
zinc electrode could be substituted, the zinc electrodes could be
freshly amalgamated, or they could be polished by rubbing them
with sandpaper. The current could sometimes be affected by
moving one zinc so that a greater or less surface was immersed
in the zinc sulphate. Putting fresh zinc sulphate solution into
the tube nearly always produced a distinct effect. All these means
of regulation proved more or less temporary, and often, when
they all failed to balance the electrodes, fresh ones had to
be made. The electrodes in which filter-paper was used proved
much more constant and easy to regulate than the clay ones, and
if they were often renewed, it seemed safe to employ them.
A positive deflection of the galvanometer meant that the poten-
tial at the right hand electrode was higher than that at the left
hand, i. e., the current through the galvanometer flowed from
right to left. Since in nearly all cases the left hand electrode was
placed on the worm anterior to the right hand one, a positi\"e de-
flection meant that the current flowed through the galvanometer
from the posterior to the anterior electrode.
The electrodes were applied, sometimes both to the dorsal sur-
face of the worm, sometimes to the ends made by transverse sec-
tions through the body, and sometimes one to the surface, and the
other to a cut end. The results will be considered according to
the position of the electrodes.
ONE ELECTRODE ON THE SURFACE AND THE OTHER ON THE
CROSS-SECTION.
It will be seen from the records given in the following selected
tables that however general certain results appear to be, never-
theless some individuals show irregularities. This uncertainty in
the results can be in part at least accounted for by the following
considerations. Secretions in different regions of the body, the
334 T. H. Morgan and Abigail C. Dimcn.
flow of blood from cut surfaces, the excretion of slime from the
skin, local or general muscular contractions might all tend to af-
fect the results. Such effects could sometimes be distinctly seen
in an increase or decrease in the deflection of the galvanometer.
It seemed possible that the presence or absence of food in the
digestive tract might in itself or by stimulating the flow of digest-
ive fluids influence the distribution of electric potential, but this
was not observable from tests made with worms that had been
starved for several days.
Experiment i. Lumbricus terrestris starved two days.^ Cut
ends anterior at several levels. The left hand electrode was ap-
plied to the cross-section and the right hand electrode to the dor-
sal surface one-third to one-half inch behind the section. The
zero point of the galvanometer was 27.1, and the effect of the
electrodes varied, deflecting it between the limits 26.1 and 29.3,
both of which deflections are less than those caused by the worm
itself, and may therefore be disregarded. The galvanometer
readings at different levels on the worm were as follows :
Cut at fourth segment. ... a. 45.0+ (off scale) Surface positive
Cut at fourteenth segment. b. 35.5
Successive sections between 1 c. 45.0+ (off scale)
the fourteenth segment J d. 42.5
and the middle of the 1 e. 40.0
worm I r. 43.5
Successive sections posterior I g. 11.5 Cut end positive
to the middle J h. 34.0 Surface positive
[ 1- 32.2
The readings are all definite and represent the state of affairs
in a majority of the worms examined. From these data it will
be seen that when a worm is cut in two it is found that in the
anterior regions of the worm the anterior cut end of the posterior
piece is negative with respect to the near-lying surface. In the
posterior regions of the worm where there was more variation the
differences in potential were usually less, and sometimes reversed in
^Since the results seemed not to be affected by the absence of food from
the digestive tract, this specimen was chosen because a more complete series
of sections were made from it than from any unstarved worm.
Physiological "Polarity" and Electrical Polarity. 335
direction. Irregularities were often observed at about the fif-
teenth segment, where the male reproductive organs open to the
exterior and the crop-gizzard region begins. Irregularities also
occurred at the girdle. In the worm used for this experiment
negative deflection of the galvanometer occurred only once, im-
mediately behind the middle of the worm, but the positive deflec-
tion was diminished at the fourteenth segment and at the other
cut ends in the posterior regions of the worm.
Experiment 2. Lumbricus terrestris. Cut end posterior. The
zero point of the galvanometer was 27.3, and the electrodes
caused a deflection to 27.5 at the beginning and to 30.0 at the end
of the experiment, which does not affect the sense of the deflection
for any reading. The galvanometer record for the different levels
is as follows :
[a. 28.0 Cut end positive
Successive sections from the posteriori L. 19.5 Surface positive
end to the middle of the worm. . . c. 23.0
I d. 23.5 ;; ;;
Section about the middle e. 24.0
Section anterior to middle f. 25.0
Section at 23rd segment g. 39.0 Cut end positive
Section at 15th segment h. 20.0 Surface positive
Section anterior to 15th segment. ... i. 17.7 "
The surface has a higher potential than the cut end at all but
two levels, one where only a few segments are cut oflf from the pos-
terior end, where the difference of potential is small; and the other
at the twenty-third segment. The reversal of current at the extreme
posterior end occurred in all the worms in which a full series of
sections was made. The twenty-third segment is between the fif-
teenth segment and the girdle. In this region two other worms
showed possible cases of reversal of current, though the more usual
condition was that the current was reversed at the fifteenth seg-
ment or at the girdle, or at both points, but between them flowed
in the same direction as in the rest of the worm. The worm used
in this experiment showed the same distribution of potential that
is found in a majority of individuals, except at the fifteenth seg-
ment, where one-half of the worms tested showed a reversal of
336 T. H. Morgan and Abigail C. Dimon.
current, the other half agreeing with this experiment in showing
no reversal.
From experiments i and 2 it may be assumed that when an
earthworm is cut in two, the transverse section commonly repre-
sents a point of lower potential than the uninjured surface near to
it. At the cut end chemical changes no doubt take place as a re-
sult of the fluids there set free, and of the general breaking down
of tissues. These conditions might be expected to alter the elec-
trical potential at the cut end, and presumably, where these changes
are greatest, the alteration of potential will be greatest. Since,
however, the transverse section is usually lower in potential than
the uninjured surface near it, whether the section be at the anterior
or posterior end of the piece, the difference in potential cannot bear
any relation to the kind of regeneration that is to take place.
Having illustrated the more regular and more usual conditions
of distribution of potential between surface and cross section, two
illustrations of the exceptions to this condition, sometimes met
with, will now be given. In the first a distribution of potential
different from the average occurred throughout the whole worm.
In the second, a short piece of a worm showed unusual conditions.
Experiment 3. Lumbricus terrestris, young. Cut end posterior.
The zero point of the galvanometer was 27.4, and the electrodes
deflected it to 26.3 at the beginning of this experiment, and to 28.9
at the beginning of the next experiment made that day. Some
doubt may, therefore, be thrown on two of the readings given
below — those at the girdle and at the fifteenth segment. The direc-
tion of deflection in these cases is probably correct, but the amount
is small. The readings were as follows :-
Posterior to middle a 24.0 Surface positive
Just back of girdle b 29.0 " "
A .^ • ^ • J, ( c 33. c Cut end positive
Anterior to girdle 3 -^-^ -i ^ ^^
] d 34.6
At 15th segment. e 26.5 Surface positive
I f 32.8 Cut end positive
Anterior to icth segment ' ? ^o-O
^ ' h 34.9
i 35-0
Physiological "Polarity" and Electrical Polarity. 337
Here we see the cut end with a higher potential than the surface,
except at the girdle and in front of the fifteenth segment, where
the current is reversed. The worm was small and immature, but
as other young worms gave the same sort of readings as the ma-
jority of mature worms, the peculiar results cannot be due to im-
maturity. In fact, it is difficult even to guess why this worm should
give such different responses from the others. Another case
was also recorded in which the cross section was anterior and of
higher potential than the neighboring surface for a series of seven
sections.
Experiment 4. Lumbricus terrestris. Piece one inch long from
near the posterior end of the worm, with a cut surface at each end
of the piece. Zero point of galvanometer, 28.6; electrodes, 30.0.
Electrodes at anterior end and
middle of piece 26.3 Anterior cut end positive
Electrodes at posterior end and
middle of piece 27.1 Surface positive
Electrodes at both ends 27.0 Anterior cut end p^ositive
In this experiment the anterior electrode is positive with respect
to the posterior one, whether it be on an end or on a surface, and
we get a constant direction of current from before backward. The
worm from which the piece was cut gave the usual results for
the other readings made on it. The piece was cut out by two
consecutive cuts with no appreciable time between them, so that
the freshness of the cut could not, as it does in other cases given
later, influence the result. One other case resembled this one,
while two cases showed conditions in which the middle was posi-
tive with respect to both ends, and three cases showed conditions
in which the middle was negative with respect to both ends.
It has been stated that marked changes in the distribution of
potenial between the surface and cross section often occur at the
fifteenth segment and at the girdle region. When a worm is cut
in two at the fifteenth segment the cut end has usually a higher
potential than the surface either anterior or posterior to it, or when
short pieces of a worm are cut with one end at the fifteenth seg-
ment, that end is positive to the other, whether it be an anterior
338 T. H. Morgan and Abigail C. Dimon.
or a posterior end of the piece. When, however, a worm is cut
in two immediately anterior or immediately posterior to the girdle,
the girdle has a lower potential than either cut end. Both these
regions, therefore, show a state of affairs different from that in
other parts of the worm. When worms were cut in two at a
series of points it was found that at the fifteenth segment there was
a change in direction of the current in twelve cases out of nine-
teen, and at the girdle in eight cases out of fourteen, four cases
in which the change was not very pronounced being included in the
first series, and two in the second. The change in direction of cur-
rent, though by no means uniform, is rather more likely to occur
than not, and may perhaps be connected with substances secreted
by the organs at these levels.
If the distribution of potential in the earthworm resembles the
distribution of potential in a resting muscle, we might expect that
the difference of potential between the electrodes would vary ac-
cording to the position of the electrodes. A series of experiments
was tried, in which one electrode was kept stationary at a trans-
verse section, and the other moved along to different positions —
usually one near, one half way between the ends, and one on the
skin at the end opposite the transverse section. Great variation
in the deflection of the galvanometer was always observed for
these different positions, but it was by no means regular. The
most common case was that the deflection- was greatest when the
electrodes were near one another, decreasing as they moved away,
and sometimes even changing to an opposite direction when they
were at opposite ends of the worm. In this series of experiments
we have not only the complicating conditions already mentioned,
but also the factor of resistance which would be approximately
proportional to the distance between the electrodes, and if appre-
ciable would modify the results in the way stated. The problem
is one of greater complexity than that of the distribution of poten-
tial in the comparatively homogeneous tissue of the muscle, where
the resistance is small. If the earthworm were homogeneous as
regards electrical conductivity, and a difference of potential were
set up by means of a transverse section, the point of lowest poten-
tial would be in the middle of the cut end, and of highest potential
Physiological "Polarity" and Electrical Polarity. 339
at the opposite end of the worm. If, however, resistance varied
in different parts of an unhomogeneous tissue, the difference of
potential observed between two points would be a resultant be-
tween a tendency to a regular rise of potential and an irregular
distribution of resistance, and the recorded distribution of poten-
tial would, therefore, be irregular. In point of fact, however, as
mentioned above, other causes of irregularity may be added to
those due to resistance, and many arrangements of potential were
observed in the seventeen cases tested.
ELECTRODES APPLIED TO TWO POINTS ON THE SURFACE.
Experiment 5. Lumbricus terrestris. The zero point of the
galvanometer was 28.6, and the electrodes deflected it to 27.9 or
27.8, in an opposite direction from deflections caused by the worm.
This experiment was made in order to see what the electrical con-
ditions are on the surface of the worm. Three readings were
taken, as follows :
a. Earthworm cut in two at the mid-
dle.
Posterior piece. One electrode on
the surface at the section and the
other on the surface, posterior to
the section, but near It 30.3 Posterior positive
b. Same worm. Short piece cut from
posterior half of animal.
One electrode on the surface at the
anterior section, and one on the
surface posterior but near 33.9 " "
One electrode on the surface at the
anterior section, and one on the
surface at the posterior section. 35.0 " "
The direction of current was the same as would be expected if
one electrode were applied directly to the cut end, and the other to
the uninjured surface near the end. On the assumption that at
any level the conditions of the surface fairly represent those in the
interior of the worm, by testing the distribution of potential at
the surface of an uninjured specimen It may be possible to get some
340 T. H. Morgan and Abigail C. Dimon.
idea of the distribution M'ithin the worm. This was done in the
following experiment.
Experiment 6. Lumbricus terrestris. Both electrodes applied
to the surface. The zero point of the galvanometer was 26.9,
and the electrodes varied from 26.7 to 22.0 after the third read-
ing, when they were regulated to 24.3, and at the end registered
only 26.0. The readings of the galvanometer were as follows:
One electrode at anterior end, the other
near 32.0 Head end negative
One electrode at anterior end, the other
at middle 30.0, then up, off scale.
Head end negative
One electrode at anterior end, the other
at posterior end 16.0 " " positive
One electrode at posterior end, the
other at middle 28.0 Tail end positive
One electrode at posterior end, the
other near i8.oi ,, ,,
negative
22.5^
It is difficult to explain these data so that they are consistent.
There Is no one point that has a high potential relative to all
others, though the deflections are sufficiently strong to indicate
that they are not due to variability In the electrodes themselves.
The variation may perhaps be due partly to local muscular con-
traction and partly to the excretion of slime at points on the sur-
face, for in Allohophora, where the body cavity fluid extruded
through the dorsal pores is yellow and noticeable. Its excretion
was observed to have a great effect upon the deflection of the
galvanometer.
Different worms, too, show the greatest differences as to their
reactions when electrodes are touched to different parts of their
surface. In general, the two ends tend to have a lower potential
than other parts of the surface, and the middle tends to have a
higher potential with respect to points on either side of It. At
the girdle and at the fifteenth segment, however, the results are
more definite, as Is shown in the two following experiments.
Physiological "Polarity" and Electrical Polarity. 341
Experiment J. Lumbricus terrestrls. Zero point of galvanom-
eter, 28.6; deflection caused by electrodes to 30.5.
One electrode at girdle, the other anterior
to it, and near 20.5 Girdle negative
One electrode at girdle, the other posterior
to it, and near 39.0 " "
Experiment 8. Lumbricus terrestris. Zero point of galvanom-
eter, 28.6; deflection caused by electrodes to 30.0, at end of
experiment.
One electrode at 15th seg-
ment, the other ante
rior to it and near. . 29. S 15th segment negative, probably
One electrode at 15th seg-
ment, the other poste-
rior to it and near. . . .31.5 " " "
The girdle is definitely of a lower potential than the surface
near it, anterior or posterior, and this was found to be the case
for four worms tested. At the fifteenth segment the difference
was not so great, and though this region was negative with respect
to a surface posterior to it, with respect to one anterior it was only
very slightly, or, perhaps, not at all so. In another worm the fif-
teenth segment was evidently positive with respect to a surface an-
terior to it.
ELECTRODES APPLIED AT TWO TRANSVERSE SECTIONS.
Experiment 9. Lumbricus terrestris. Pieces cut out from
worm. The zero point of the galvanometer was 27.3. The
electrodes deflected it to varying amounts. In two cases, namely
the fourth and the seventh readings in the table, where the elec-
trodes deflected the galvanometer to 26.1 and to 29.5, respectively,
these deflections come near the readings given by the worm. In
the other cases it is not necessary to take deflection caused by the
electrodes into account, since they would not affect the direction of
the reading. The data are as follows :
342 T, H. Morgan and Abigail C. Dimon.
(a) Anterior half of worm, both ends cut
(long piece) 16.6 Anterior positive
(b) Short piece from anterior part of an-
terior half of worm, the posterior
end cut somewhat later than the
anterior end 22.3
(c) Same. Anterior end freshly cut. . . . 32.5 Posterior positive
(d) Short piece from middle part of an-
terior half of worm, posterior end
more freshly cut 27.1
27.0
27.9
(e) Same piece, anterior end freshly cut. 36.2 " "
( f ) Short piece from anterior part of pos-
terior half of worm 33-0 "
(g) Short piece from middle of posterior [ 28.5 Anterior end
half of worm \ 24.0 positive probably
(h) Anterior half of (g), posterior end
freshly cut 22.0 Anterior positive
(i) Posterior half of (g), anterior end
freshly cut 34-0 Posterior positive
In this worm, when the two ends were cut at approximately
the same time, which happened in (a), (f) and (g), the piece
from the anterior half had its anterior end positive, and the two
pieces from the posterior half had their posterior ends positive.
In the majority of worms tested, when the two ends of a piece
were cut at the same time, the anterior end was positive, regard-
less of the position of the piece on the worm.^ If, however, the
two ends were cut at different times, which in this worm occurred
in six pieces, the end which' had been cut most recently generally
had a lower potential than the other. Since the testing with the
electrodes on a transverse section and an uninjured surface near
that section the end was usually found to be at a lower potential
than the surface, the fall of potential was supposed to be due to
^When, however, the pieces were long (one-half the worm or more), in a
majority of cases the posterior end was positive with respect to the anterior.
Physiological "Polarity" and Electrical Polarity. 343
changes accompanied by the escape of body fluids at the cut end.
Since, when the electrodes are on two transverse sections, the one
that is more recently cut is of a lower potential than the other, it
would appear that the causes that determine the fall of potential
are such as decrease in the course of a short time. The fluids
that escape from the cut end dry rather rapidly, whereas the tissue
cells, breaking down, are not built up for several days. The first
fall of potential, then, is probably largely due to escape of blood
or other fluids from the section, while the slighter permanent effect
may be due to the breaking down of tissue cells. When there
was but a short time between the two cuts the lower potential did
not always occur at the fresher one, which may have been partly
because fresh fluids were still coming from the earlier section. In
the anterior half of the worm, also, there was great irregularity,
which may be partly due to difi^erent digestive fluids at different
regions of the digestive tract producing varying electrical activi-
ties.
The differences of potential between two transverse sections are
probably a resultant of the factors that cause difference of poten-
tial between a section and a surface. If a piece be cut out from
an earthworm by two transverse sections, there will be a difference
of potential between each end and the surface between the ends.
If the differences of potential between each end and the surface
are equal and opposite they will balance one another, and there
will be no difference of potential when electrodes are applied to
the two ends. If, however, the differences of potential between
the two ends and the surface are unequal, their resultant will de-
termine a difference of potential between the two ends. This is
illustrated by experiment 10.
Experiment 10. Lumbricus terrestris. Short piece. The zero
point of the galvanometer was 28.6, and the electrodes registered
slightly above this at the beginning of the experiment. The read-
ings were as follows :
(a) One electrode applied to anterior end, the other to the
middle of the piece 25.0
(b) One electrode applied to posterior end, the other to the
middle of the piece 30.8
344 T. H. Morgan and Abigail C. Dimon.
(c) One electrode applied to one end, the other to the other
end of the piece 28.0
If we disregard the deflections caused by the electrodes, the
following are the departures from the normal: (a.) == — 3.6,
(b.) = +2.2, (c) = — 0.6. Theoretically, if (c.) were the re-
sultant of (a.) and (b.), it would equal — 1.4, but considering
how variable the conditions were always found to be, — 0.6 pre-
sents a fairly close agreement. In all other cases but one, where
similar tests were made, the results agreed in like manner with the
theory.
REGENERATING WORMS.
In addition to the readings made from worms that had been
freshly cut in two, a series of readings were made on worms in
which the process of regeneration had proceeded for a number
of days. The regenerating worms were divided into four groups :
( I ) those in which a few anterior segments had been cut off and
regeneration was taking place at the anterior end of the long
piece; (2) those in which the worm had been cut in two in the
middle and regeneration was taking place at the posterior end of
the anterior piece; (3) those in which the worm had been cut in
two in the middle and regeneration was taking place at the an-
terior end of the posterior piece; (4) those in which the worm had
been cut in two at the fifteenth segment and regeneration was tak-
ing place at the anterior end of the posterior piece. They were
allowed to regenerate from twenty-five to thirty-two days, and in
the course of that time five or six tests were made on most of
them at intervals of a few days. One electrode was applied to the
regenerating tip, and the other to the old surface a short distance
from the tip. In all one hundred and fourteen readings were re-
corded, the results of which may be summarized as follows :
Group (i) 31 cases, end positive ; 9 cases, end negative
Group (2) 20 " " " 18 "
Group (3) 15 II
Group (4) 8 " " " 2 "
Total 74 " " " 40 "
Physiological "Polarity" and Electrical Polarity. 345
If we regard only the readings made when regeneration had
proceeded more than twenty-one days, in thirty-six cases the end
was positive with respect to the surface, and in eleven cases it was
negative, a much more definite result than when all cases are con-
sidered.
The conditions in a regenerating tip do not, therefore, agree
with those at a freshly cut end, for the current flows in an oppo-
site direction. To be sure, the processes occurring during regen-
eration are not the same as those occurring during the breaking
down of tissues, and the latter may be the predominant ones im-
mediately after a cut is made. The subject, however, needs fur-
ther investigation before the causes for this reversal can be more
than surmised.
ELECTRICAL POLARITY AND RATE OF REGENERATION.
If we look for a relation between electrical polarity in the worm
and rate of regeneration, as Mathews has suggested, we find it
as difficult to demonstrate as the difference between electrical and
physiological polarity. If the average deflection of the galvanom-
eter was greater at certain levels where regeneration is known
to be rapid than at other levels where it is slow, the connection
would be established. For instance, the regeneration of a tail at
the posterior end of a worm when only a few posterior segments
are cut off, is exceedingly rapid, whereas the regeneration of a
heteromorphic tail at the middle of a worm is very slow. The
average deflection of the galvanometer in the former case for
three readings is, however, 3.4, with the cut end positive instead of
negative in every case. In the latter case the average deflection
for seven readings is 3.6 (with the end negative) , with no extreme
readings to brtng it up. The region in the middle of the worm,
where a tail is to regenerate from the posterior end of the anterior
piece, gives an a^^erage deflection of 4.7 for seven readings from
different worms. When five or six segments are cut from the an-
terior end of the worm the average deflection for seven cases was
4.6 at the anterior end of the long piece. At the posterior end of
the short piece, regeneration would be very slow, and at this end
346 T. H. Morgan and Abigail C. Dimon.
only two readings were made, one giving a very slight result, the
other deflecting the galvanometer to 5.2.
If we attack this subject by another method, namely, by mak-
ing a direct comparison between two freshly-cut ends on one worm,
the results are equally indefinite. When five or six segments were
cut from each end of the worms, of a series of fifteen readings on
different worms made with one electrode at the anterior-cut end
and the other at the posterior-cut end, the posterior end was posi-
tive in eleven and the anterior end positive in four. When the
worm was cut through the middle and at the posterior end, the
posterior end was positive in three cases, and the middle in one
case. When the cuts were made through the middle and anterior
end, in five cases the middle was positive, and in two the anterior
end positive. From these considerations it would therefore ap-
pear that no invariable connection between rate of regeneration
and electrical polarity exists in the earthworm, at least as measured
on a freshly cut surface.
From the foregoing experiments we conclude :
( 1 ) That a freshly cut end of an earthworm is generally nega-
tive with respect to a near-lying uninjured surface.
(2) That the freshness of the cut surface has an Important in-
fluence In determining the amount of difference of potential.
(3) That the result is often complicated by the presence of se-
cretions or exudations on the surface, or by the presence of certain
organs at the cut end, or by the contractions of the worm, etc.
(4) That in the region of the girdle and also in the region of
the fifteenth segment (near which the crop and gizzard lie), the
results are often different from those elsewhere.
(5) That there Is no apparent relation between the differences
In potential at freshly cut surfaces and the kind of regeneration
(head or tail) that occurs.
(6) That cut surfaces from which heteromorphic growth
would take place show the same sort of differences in potential
as those from which orthomorphic regeneration occurs.
(7) That the differences in potential present when a cut sur-
face Is exposed can probably be accounted for by the chemical
changes taking place at the surface; and these need have, and do
Physiological "Polarity" and Electrical Polarity. 347
not appear to have, any relation to the kind of regeneration that
takes place.
(8) That when, on the other hand, the cut surface is allowed
to heal, and when later a new structure has begun to appear, the
differences in potential between the new and the old parts {as
measured on the surface only) are not such as can be made to ac-
count for the difference in the kind of part (head or tail) that is
regenerating. Here also many complications enter into the re-
sult and make it difficult to draw satisfactory conclusions.
(9) No definite relation was found between the rate of growth
and the fall of potential between an uninjured surface and a cut
end.
THE REGENERATION OF A HETEROMORPHIC TAIL
IN ALLOLOBOPHORA FOETIDA.
ABIGAIL CAMP DIMON.
In a paper by Professor Morgan^ an account was given of an-
terior regeneration from three different levels in earthworms.
The results seemed to show that the internal factor determining
the formation of a heteromorphic tail might be the presence of
the stomach-intestine at the regenerating surface, and at Professor
Morgan's suggestion and under his direction, the following ex-
periments were undertaken. An attempt was made to test this
view by means of more exactly localized sections made near the
level of the beginning of the stomach-intestine.
In Allolohophora foetida, the species used, the oesophagus ex-
tends to the fifteenth segment, the crop lies in the fifteenth and
sixteenth, the gizzard in the seventeenth and eighteenth, and at
the nineteenth begins the stomach-intestine, which extends pos-
terJorally through the rest of the worm. The external openings
of the vasa deferentia on the fifteenth segment served as con-
venient landmarks for determining the level of the section. The
worm was cut in two, the short anterior piece dropped into alcohol,
and its number of segments counted so that the exact level of the
cut could be recorded. The posterior piece was then left to re-
generate from forty-eight to one hundred and twenty days, when
it was killed, and sections made for study. In some cases a new
head, and other cases a new tail regenerated from the anterior end
of the posterior piece. In the regenerating head the new stomo-
daeum usually did not open into the old digestive tract, which
closed anteriorly, and no definite pharynx formed, A dorsal brain,
connected with the ventral nerve cord was usually present. Since
these conditions represented the most usual form of head regenera-
lExperimental Studies of the Internal Factors of Regeneration in the Earth-
worm. Arch, fiir Entwickelungsmech. der Organ. Bd. XIV. pp. 562-591.
350
Abigail C. Dimon.
tion at the levels of these experiments, the cases in which they ex-
ist are classified in the table as a separate group under Head A.
Cases where the brain lies anterior and even ventral to the level
of the digestive tract; where the nerve cord ends without form-
ing a brain; or where there is no mouth invagination, though the
brain is well developed, are classified as Head B. The heads of
group B look less like a normal head than those of group A, and
yet are very evidently to be classified as heads rather than as tails.
The distinctive features indicating a heteromorphic tail are the
formation of a number of segments, the opening of the digestive
tract to the exterior through a new anus, and the ending of the
nerve cord ventrally, without a brain. Tails possessing these fea-
tures are put in group A, while those in which any of these features
are absent, are put in group B/
In all, one hundred and seventeen worms were examined, with
the results given in the table. The number of the segment given
at the head of each column locates the level at which the worm was
cut in two, and both the actual number of worms and the percent-
ages are given under each class.
12th-
13 h
Segment
Between
14th
Segment
and
15th
Segment
Between
15th
Segment
and
16th
Segment
Between
16th
Segment
and
17th
Segment
Between
17th
Segment
and
18th
Segment
Between
18th
Segment
and
19th
Segment
Back of
1 *h
Segment
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
Head, A..
Head, B..
2
100
7
3
70
30
11
11
46
46
7
4
59
33
6
6
46
46
6
20
1 ,
1
5
18
61
3
3
15
4
5
7
7
23
17
Tail, B...
22
Tail, A. . .
30.5
Uncertain
2
8
1
8
1
8
30.5
Total . .
2
100
10
100
24
100
12
100
13
100
33
100
100
iJn only twenty-four cases out of one hundred and seventeen did the old di-
gestive tract open to the exterior through the new mouth or anus. This occurred
ten times in head regeneration, ten times in heteromorphic tail regeneration,
and four times in cases classified as uncei^tain. Seven of the twenty-four cases
occurred when the worm was cut in two in front of the sixteenth segment, and
the other seventeen when it was cut behind the eighteenth segment.
Heteromorphic Tail in Allolohophora. 351
Since there were but few worms cut further back than the eigh-
teenth segment, and since the stomach-intestine begins at this level,
all the observations made on worms cut posteriorly to the eigh-
teenth segment were brought into one class. It is worth noting,
however, that of the four heads regenerating at these levels, three
formed from worms cut at the nineteenth segment, while the
fourth formed from a worm in which the exact level of the cut
was not noted. The percentages In the different classes, though
based on a small number of cases, yet bring out clearly one or two
points. When a worm was cut in two In front of the stomach-
intestine, In no case was a heteromorphic tail formed. The per-
centage of cases In which a head was formed grows less as the sec-
tion Is made further back on the worm, the fall of percentage
being very great Immediately behind the gizzard. This tends to
support the hypothesis that the formation of a heteromorphic tail
is favored by the presence of the stomach-intestine near the cut
end, though when the section is not more than one segment back
of the gizzard a head is sometimes formed.
Though the preceding experiments seem to show that the devel-
opment of a heteromorphic tail is connected with internal struc-
tures in the worm, they leave untouched the question of the kind
of regeneration that takes place from posterior ends of anterior
pieces cut anterior to the stomach-intestine. This point should be
determined, and I hope In the future to undertake a set of experi-
ments in which the posterior regeneration from anterior pieces
will be observed.
RESTORATIVE REGENERATION IN NATURE OF
THE STARFISH LINCKIA DIPLAX (MULLER
AND TROSCHEL).
BY VERNON L. KELLOGG, STANFORD UNIVERSITY, CALIF.
On the surface of the coral reefs guarding the harbor of Apia
(Samoa) the five-rayed starfish, Linckia pacific^, with its long,
slender, smooth, sky-blue arms, is the most conspicuous and abun-
dant echinoderm in a place where echinoderms abound. Associ-
ated with it, and similarly blue and conspicuous, although smaller,
Fig. I. Linckia diplax, regener-
ating from a single arm; note these
new arms and new disc with madre-
porites.
Fig. 2. Linckia dip/ax, re-
generating from a single arm;
note four new arms and new
disc with madreporites.
is the species L. diplax. Both for number of species and wealth
of individuals, the Apia reefs are distinguished by their star-
fish, sea-urchin and holothurian fauna. In collecting on these
reefs during several weeks in the summer of 1902, as a member of
354
Vernon L. Kellogg.
the U. S. Bureau of Fisheries' Samoan Explorations party, my
attention was particularly attracted by the many examples of star-
fishes with regenerating arms, and I gave some special care to pick-
ing up such specimens. From this material the figures here pre-
sented have been drawn and in themselves tell how effectively this
capacity for restorative regeneration obtains in this species.
Morgan calls attention in his "Regeneration" (1901, p. 102
and elsewhere) to the assertions of some authors that starfishes
Fig. 3. Linckia dip lax,
regenerating from a sin-
gle arm.
Fig. 4. ((?) Linckia dip-
lax, a single arm broken
at both ends regenerat-
ing. (^) Aspect of proxi-
mal end of arm.
Fig. 5, Linckia. pacifica,
regenerating from a sin-
gle arm, broken off ob-
liquel}' from the original
disc; note four new arms
and disc, the outer arms
larger than the two inner
ones.
can regenerate a new disc and other arms from an arm torn olf
without any part of the disc attached, and to the denials by other
authors that such radical restoration can take place. In the case of
Linckia diplax there seems to be no doubt of the capacity of an
arm torn off at some distance from the disc to regenerate a com-
plete new animal from its proximal surface. The possibility that
these arm pieces were thrown off by autotomy instead of being
torn off by enemies may be noted, but such a condition makes the
Restorative Regeneration of Linckia.
355
regenerative phenomena none the less interesting. I have seen no
example of the regeneration of several new arms (or a new disc
and arms) from the distal end of a mutilated arm, as observed by
the Sarasins in Linckia multifera (Ergeb. Naturforsch. auf Ceylon,
1884-85, I, Wiesbaden, 1888). In all cases of regeneration from
the distal end of an arm noted among the Apia reef starfishes,
simply a continuation, in straight line, of the tapering tip oc-
curred. Among the figures will be noted the illustrations of three
specimens in which the regenerating arm has had its distal end
Fig. 6. Linckia diplax, {a) a specimen regener-
ating parts of two arms: {b) the aspect of a normal
madreporite (compare with the regenerated madre-
porite shown in figures i and 2).
torn off (or thrown off) as well as having been Itself broken off
from its basal extremity, and thus freed from the rest of the body
to which it originally belonged. In all of these cases of mere seg-
ments of a single arm regeneration is proceeding at both mutiliated
ends.
In Figures i and 2 a new mouth and both^ madreporites are in
the regenerated part. In Figure 3 a new mouth has been already
regenerated, but no madreporite as yet. In Figure 4 is shown an
^Linckia diplax is characterized by the possession of two madreporites.
356 Vei-non L. Kellogg.
arm torn off at some distance from the disc, just beginning to re-
generate. The cut end has "calloused" over, apparently by the
inbending of the edges of the body wall, but in the center is left
a small opening (serving as mouth ?). No protuberance indi-
cating new disc or arms has yet appeared. The arm segment,
which is regenerating at both ends, shown in Figure 5, is of an-
other species of Linckia, probably pacijica, and had an obliquely
cut surface at the proximal end, and the two outer arms of the four
regenerating ones, that is, those nearest the parent arm, are about
twice as well developed (as far as size goes) as the other two.
No madreporite is yet developed on the new discal portion. The
specimen illustrated in Figure i is regenerating but three new
arms instead of the normally missing four. In all the specimens
illustrated by Figures i, 2, 3, 4 and 5 the arms were undoubtedly
broken off without any part of the disc attached.
NOTES ON INSECT BIONOMICS.
BY V. L. KELLOGG AND R. G. BELL, STANFORD UNIVERSITY,
CALIFORNIA.
In connection with the experimental breeding and rearing under
controlled conditions of food supply of many lots of silkworms
{Boinbyx mori) during the last three years, the writers have made
certain observations and experiments incidental to the main object
of the investigation, the results of some of which may be here
briefly abstracted.
Food Conditions in Relation to Sex Diferentiation.
It has been assumed by some authors that poor nutrition of de-
veloping organisms is an extrinsic influence tending to determine
the sex of the organism to be male and good nutrition an influence
tending to produce females. The most important part of the as-
sumption is the idea that sex is subject to control by the environ-
ment of the organism — that sex is not inherently predetermined in
the germ.
From the notes of the writers recording the results of an ex-
perimental rearing of numerous lots of silkworms on reduced ra-
tions in 1 901, 1902 and 1903, the following data are extracted
touching the problem of the relation of nutrition to sex differen-
tiation. From an inspection of these data it will be noted that a
test is included of the possible influence of poor nutrition of the
parents (and grandparents) in determining the sex character (if
predetermined) of the germ cells, as well as of the possible imme-
diate influence of nutrition in determining the sex of developing
individuals. It will be noted also that we have had in mind the
justly made criticism of most observations on the food and sex
problem, namely, that no attention is paid in records of an appar-
ent overproduction of males following poor nutrition, to the deaths
which ensue before the count is made, and that, as the females (it
being assumed) actually require more food to complete their de-
358 Fernon L. Kellogg and R. G. Bell.
velopment, the preponderance of males is due to the untimely
death of the females.
A series of lots of ten individuals each were reared in 1903,
with the specific intention of testing the assumed influence of nu-
trition on sex determination. These lots included: (a) a lot un-
derfed during the whole of larval existence; (b) a lot underfed
during the second to fifth intermoulting periods, inclusive; (c) a
lot underfed during the third to fifth intermoulting periods; (d)
a lot underfed during the fourth and fifth intermoulting periods;
(e) a lot underfed during the first intermoulting period only; (f)
a lot underfed during the second intermoulting period only;
(g) a lot underfed during the third intermoulting period only; (h)
a lot underfed during the fourth intermoulting period only; (i)
a lot underfed during the fifth intermoulting period only. From
the rearing of such lots it was hoped to determine what influence
reduced rations might have on the determination of sex, and also,
if any, at what time in the larval life the influence was most potent.
A consideration of the records of the rearing of these lots at the
end of the season compels us to say : that the lots were much too
small to aftord trustworthy generalizations; that dissections of
the larvae at various ages reveal an unmistakable differentiation in
sex (indicated by gross differences in the reproductive glands) at a
time as early at least as the beginning of the third intermoulting
period, so that experimental lots c, d, g, h and i were distinctly
superfluous; but finally, that it may be affirmed from the meager
data afforded by experimental lots a, b, e and /, that the reduction
of the food supply (this reduction brought as near as possible to
a living minimum) did not produce any unmistakable results in
the way of an overproduction of males.
Data of more interest are those derived from an inspection of
the records of the experimental rearing of various larger lots of
silkworms in 1901, 1902 and 1903. In 1901 the records for five
lots of twenty larvae each may be referred to :
Lot I — Fed optimum food; no deaths before emergence of
moths ; produced 8 males, 1 2 females.
Lot 2 — Fed optimum food; 2 deaths before maturity; produced
7 males, 1 1 females.
Notes on Insect Bionomics. 359
Lot 3 — Fed one-half (approx.) of optimum of food; 4 deaths
before maturity; produced 10 males , 6 females.
Lot 4 — Fed living minimum of food; 3 deaths before maturity;
produced 10 males, 7 females.
Lot 5 — Fed living minimum of food; 6 deaths; produced 9
males, 5 females.
Four lots of twenty larvae each reared in 1902 may be referred
to.^ These larvae were the offspring of parents of the variously
fed 1 90 1 lots, and the character of the food supply of the parents
is indicated as well as that of the larvae themselves.
Lot I — Fed optimum; born of optimum food parents; no
deaths before maturity; produced 12 males, 9 females (21 indi-
viduals in this lot by mistake) .
Lot 2 — Fed minimum food; born of optimum food parents; 7
deaths before maturity; produced 8 males, 5 females.
Lot 3 — Fed optimum food; born of minimum food parents; 1 1
deaths before maturity; produced 6 males, 3 females.
Lot 4 — Fed minimum food; born of minimum food parents; 3
deaths before maturity; produced 11 males; 6 females.
The records of eight lots of twenty-five larvae each reared in
1903 may be referred to. The food supply condition of the par-
ents and grandparents, as well as of the 1903 progeny, are given.
("O" indicates optimum food, "M" indicates minimum food).
Deaths
Lots Fed Parents Grand- before Males Females
parents maturity produced produced
I O O O 2 13 10
2 M O O • 2 14 9
3 O M O 3 8 14
4 M M O 6 8 II
5 O O M o 15 10
6 M O M o II 14
7 O M M 20 2 3
8 M M M 2 1 - 2 2
^Because of backward season all 1902 larva; were fed for their first 20 days
(=:rabont one-third of whole larval life) on food of a poor quality, namely,
lettuce and mulberrv buds.
360 Venion L. Kellogg and R. G. Bell.
The writers present these figures, actual data, for what they
may be worth. Like the data of the smaller lots previously re-
ferred to, they at least show that individuals living through their
whole post-embryonic life on the smallest food supply capable of
sustaining life, a supply varying from M to >^ of the supply nor-
mally used by individuals of the species, do not necessarily become
males. Whether the figures indicate an appreciable influence of
this nutrition on the determination of sex can be determined by the
readers as well as by the writers. In the rearing season (March
to June) of this year ( 1904), the writers purpose devoting much
larger lots of individuals to the continuation of the experiment.
Forced Pupation.
Experiments were made to determine how early in larval life
the food supply could be cut off without stopping the metamor-
phosis (development) of the silkworm, whether such forced ab-
breviation of the food-taking period results in any unusual struc-
tural or physiological modification in the stages which follow the
withdrawal of food, and whether the metamorphosis (in particu-
lar, pupation) is hastened when food is withdrawn in late larval
life, an adaptation often assumed to be possessed by Lepidoptera.
Such an adaptation would obviously be of real advantage, as it
might often save individuals from death due to a sudden disap-
pearance of the food supply, or to a sudden accidental incapacity
to gain access to the food supply.
The silkworm spends normally about sixty days in the larval
(feeding) stage, divided into five actively feeding intermoulting
periods of about ten days each, by four brief two-day moulting
periods, during which no food is taken. On the eleventh or
twelfth day (from 270 to 300 hours) after the fourth moult, the
larva "spins up" and pupates.
Twenty healthy silkworms were selected at random from a
large lot (several hundred) which had been reared in one tray,
all the individuals, of course, under the same condition of food
supply, temperature, humidity, light, etc. Of the twenty, one
was fed as long as it would take food; the other nineteen were de-
Notes on Insect Bionomics. 361
prived of food variously from the time of the fourth moult, from
one day after the fourth moult, from two days after, from three
days after, and so on until individuals were obtained representing
a withdrawal of food supply for a period of but a day before the
normal time of giving up eating to begin spinning, through periods
of two days before, three days before, four, five and so on to
twelve days before, the twelve-day period being the whole of the
feeding period normally lasting from the fourth moulting up to
spinning time. The following table displays the conditions and
results of the experiment :
362
Vernon L. Kellogg and R. G. Bell.
O -^^
oot- 'bof'boofo'b
COO
-* 10
bC be bC be ^ bC be
1-H (M r^ M jy 10 rf
(M ^ ^ ^ ^ iO(N
«4-l 0)
OOt^ 2cOO0O000l>(M(M g
C CC CI Pl CI PI fl cj fl-c!
^s-^33;3Sci33v:;
•-»'-5Rl-5^-?^-Sl-5l-sl-5l-,0
o 9 »<
"3 1=1
«!_
bebCbCbCbjObCfaCbrbCbe
IOIOOIOC0005»00^
(M00C0TftOOO(M'-H.-i
05005000CDOOCOOOO
»>» ^ ^ ^ C3 ^ ^ j^
O 4,
u to
T3 O
Cj3
3a
O O
£3
ja o
'H S
« a
s-
p M
> a
a>"
2 a
did,
a a a
°'-5
03 O)
Q
2 o
- - . « ^ ^ . » . (D
l:^cO*"t>»OcDI>CCcDCO+^
t^ K*^ r*^ ^ ^ r^ r^ „^
O fl
s'a
aa a
c3
Qa
G (-'^
Q 0^ O; G OJ
(M(N(N(N(M(N(M(M -----
O O
Cl CI CI
bo be be
.a .3 .2
eeoJcSojojojcSesSaoSy, 9^9^0^000
%-i o
(U
03
■*0000000<M(MGOTjHCCOCO(M(MOOGOTi<
(N (N (N IM (N (M rH i-H rH tH ,-1 ,
*
03 OJ
0.2
a a a a a a a a a a a a a
a a a a
cJojiScSoSoS&i&iejaS&HOjrt
1— i^T— 1^1— i^iOiC lOi-ii-H
§§§§§gg§§§gg§§§§g§§
rH(Mfo-<tiio<©i>ooo;iOT-i(Nro'*iocot^ooo5
Notes on Insect Bionomics. 363
Note I — Bottom of bottle in which larva was confined was
lined with threads May 26; threads extended up one side May
27; threads swung across bottle from side to side May 28, and
larva actively spinning very damp threads; on May 29 larva was
spinning a closely woven circular carpet on bottom of bottle, and
on May 30 larva pupated on this carpet (no cocoon).
Note 2 — Bottom of bottle lined with threads May 26; larva
still slowly spinning random threads May 28.
Note 3 — Bottom of bottle lined with silk coating May 18.
Note 4 — Larva lined bottom of bottle with stray threads be-
fore dying.
Note 5 — Slight progress in the spinning by June 6, P. M.; bot-
tom of bottle lined with threads.
From these results it may be said that silkworms may be cut
off from a food supply nearly seven days before the normal limit
of their feeding time and yet complete their development (spin,
pupate and emerge as imago) . These seven days represent a little
more than half of the last intermoulting actively feeding period,
or about one-ninth of the whole larval (feeding) life. The depri-
vation of food for from one to four days seems neither to hasten
the metamorphosis nor to modify it appreciably, nor to result in
the production of a moth of lessened size or lessened fertility. The
larvae deprived of food not more than four days before normal
close of feeding time do not immediately spin and pupate, but
wait restlessly for the normal time of pupation (approxim^aMly
twelve days after the fourth moulting), and then normally spin
and pupate. If deprived of food for more than four days and
less than seven, the larvas shorten their last intermoulting stage
to about seven days, forming, however, a normal cocoon and trans-
forming into a normal moth. If the larvae are deprived of food
eight days or more before their normal splnning-up time, they in-
variably die without forming a cocoon, and in only one case was
pupation accomplished. A beginning at spinning (see notes) is
made by larvae fed for more than two days after the fourth moult-
ing, but no spinning at all is done by larvae deprived of food from
the day of fourth moulting or from the first or second day there-
after.
364 Vernon L. Kellogg and R. G. Bell.
The twentieth larv^a of the lot was to be deprived of food 216
hours after the fourth moult, but it began spinning up in 200
hours (eight days) after, and pupated on the following day. Here
is a normal variation of four days out of the usual twelve of the
last feeding stage, just about as much shortening as the extreme
that could be induced by actual deprivation of food.
Loss of Weight Daring Pupal Life.
A belief among commercial breeders of silkworms that there
is a loss in weight of the cocoons (silk) accompanying pupal life
is indicated by their recognized wish to make an early sale of the
cocoon product. This loss is generally attributed to "evaporation
from the cocoon." The question arose as to whether the loss in
weight of the pupa-containing cocoon might be not a loss in weight
of silk but an accompaniment of developmental changes in the
pupa, a process in which stores of nourishment (in the larval body)
are being converted into moth with chemical changes which might
occasion some loss in weight. Therefore in four individuals the
cocoon and pupa were weighed separately once each day from the
time of pupation to time of emergence of the moth, while at the
same time the daily weights of the naked chrysalids of three other
lepidopterous species were determined to see if a loss of weight
accompanied pupal aging in them as well as in the silkworm
moth. The following table shows plainly the results of these ob-
servations :
Notes on Insect Bionomics,
365
Oi-H
O"
CO CO
r-H
(M CO
q r-;
oc^
^ c*-
q.-
CO
CO
t^ CO
q r-;
QOO
<M CO
qrt_
00 Cs
q.-
00
CO
O)
CO
>> •
05 1C
0^
oco
100
CO CO
10 CO
q i-H
CO
CO
CO
Oi-;
00 --H
05 CO
o^_
T— 1
(>>
^ CO
q r^_
CO
q rH
00
CO
CO
10
May 24
(6 p. m.)
9".
1— 1 1—1
oco
-* CO
q ^
co^
r^ CO
q rt_
CO
q
S CO
CDIO
Or-;
co^
iM CO
CO
CO
CO
00
00
>> •
03 &<
S CO
10 iC
00 (M
q^
oco
coco
T-l 1— (
*
CO
CO
00
May 20
(4 p. m.)
CO >o
Or-;
I— 1
CO
00
00
CO
00
CO
CO
10
q
2?
03 &,
05 IM
0^
"* CO
CO
.—1
CO
03' ci
bi) bJD
00
oco
CO(M
lO CO
1— 1
CO
10
CO
CO
I— 1
C
c
C
d
c
c
c
C
CO
6
CO
■■
Q
d
rH 03
s g
a!
CI
■< u
. c3
-(^ ,—
_Q CO
S^
p
IS
^
P-,
^'-^
^
o3 g
II
0;
C
cm e
^ s
-2 §
PI S
t-i — '
3
1
m ^
366
Vernon L. Kellogg and R. G. Bell.
•-5
O 3
^ CO
CO
4:J-d :
^.2 :
OO
00 CO
00 .-H
^-6
cot^
oOi-;
Ol-H
T-H CO
CO
o
CO(M
00 l-H
lOi-l
coco
05.-H
(M
CI
to 05
00 rH
l-H (M
CD CO
OJl-H
00
O 13
i>o
<£) CO
OOr-J
r-co
C-l
June 1
(3 p. m.)
05,-;
oo
00 CO
QOr-H
OS CO
05 1-;
CD
May 31
(5 p. m.)
CD (M
OO
cft CO
00 .-H
Oco
Oi-H
o
p
May 30
(4 p. m.)
00 (M
02 T-H
0--I
<N CO
q.-H
^ o
oco
i^CO
Ol-H
o
00
05 ^•
t-H
05^
1—1
<N O
--H CO
l-H (M
COCO
Or-;
l-H
o
CM
o
CO
6
>
)
)
6
C
1
5
)
3
CO
d
c
c
6
c
o 5:
o
6
c3
O Sq
go
o3
O <o
1
03 d
Notes on Insect Bionomics. 367
From this table it is apparent that the silken cocoon loses a
very small amount, about 4 per cent., of its weight in the first day
after its completion, and then loses no further weight; that the
pupa loses weight slightly but persistently and steadily from day
to day throughout its entire duration, the total loss amounting to
about 14 per cent. ; and that the pupae of three other lepidopterous
insects, namely, the tent caterpillar {Clisiocampa sp.), checker-
spot butterfly {Melitaea sp.) , and mourning-cloak butterfly (Eiiva-
nessa antiopa) also steadily lose weight from day to day, this loss
being very considerable in two of these species, viz., about 35 per
cent, in the case of one and 65 per cent, in the case of the other.
THE LOCATION OF THE CHICK EMBRYO UPON THE
BLASTODERM.
BT
FLORENCE PEEBLES, Ph.D.
With 2 Plates and 15 Figures in the Text.
An experimental study of the avian egg has led me to examine
the following points:
1. The location of the embryo in the material of the unincubated
blastoderm.
2. The direction of growth before, and after the appearance of
the primitive streak.
3. The origin of the material from which the later embryo
arises.
According to Kopsch^ this third point has been definitely
settled. He concludes that nearly all of the embryo develops
from the primitive streak. I quote his own words: "Somit
entsteht der Embryo, mit Aussnahme des praechordalen Teils,
des Kopfes, durch Umwandlung des Primitivstreifens."
P have already mentioned some experiments, and will describe
others, in the following pages, which seem to prove that only the
trunk and caudal regions of the embryo arise from the material
of the primitive streak.
The methods used by Assheton, myself, and Kopsch are practi-
cally the same, and therefore require little explanation. I have
^Kopsch, Fr. Ueber die Bedeutung des Primitivstreifens beim Hiihnerembryo.
Leipzig, 1902.
^Peebles, F. A Preliminary Note on the Position of the Primitive Streak, and its
Relation to the Embryo of the Chick. Biolog. Bulletin, Vol. IV, No. 4, 1903.
370 Florence Peebles.
again used the method described in my earlier work/ A small
window was made in the shell just above the blastoderm, and the
operation performed, after which the opening was closed by a
piece of shell, sealed with strips of the shell membrane. All in-
struments used in the experiments were carefully sterilized, and
the shells of freshly opened eggs used for closing the windows.
The loss of eggs through infection was small. In most of the
experiments one egg in each set was opened and then sealed again,
without operating upon it, in order to have a check with which to
compare the eggs upon which experiments were made. In this
way it was possible to determine roughly, after further incubation,
whether abnormalities were due to opening the egg or to the
operation performed upon the blastoderm. In general it was
found that the development of eggs in which windows were made
was delayed about two to four hours.
I. THE LOCATION OF THE EMBRYO IN THE MATERIAL OF THE
UNINCUBATED BLASTODERM
In 1896, Assheton^ described some experiments that he made
on the unincubated blastoderm of the chick. Sable hairs were
inserted at various points and their position determined after
periods of incubation varying from eighteen to forty hours.
Assheton proved that Duval's^ theory of the formation of the
primitive streak is incorrect, that instead of forming by the con-
crescence of the posterior margin of the blastoderm, the primitive
streak appears in the region of the unincubated blastoderm which
lies between the center and the posterior margin of the area
pellucida. I have repeated Assheton's experiments, making the
injuries with a hot needle instead of a hair, without removing the
egg from the shell. The results agree with those of Assheton, as
the following experiments show:
'Peebles, Florence. Some Experiments on the Primitive Streak of the Chick.
Archiv. fur Entwickelungsmech. der Organismen. VII Band. 1898.
^Assheton, R. An Experimental Examination into the Growth of the Blastoderm
of the Chick. Proceedings of the Royal Soc, Vol. 63, 1896.
^Duval. De la Formation du Blastoderme dans I'Oeuf d'Oiseau. Annales des
Sciences Naturelles, Zoologie, Vol. 18.
The Location of the Chick Embryo. 371
Experiment I. A small window was made in the shell of an egg
a few hours after it was laid. The blastoderm measured 2.8 mm.
in diameter and the area opaca and area pellucida were faintly
defined. A hot needle (No. 12) was inserted in the center of the
blastoderm (Text-fig. i, x) and quickly withdrawn. The shell
was sealed and the egg put in the incubator, the temperature of
which varied from 37°-39° Centigrade. At the end of twenty
hours the egg was opened, and the blastoderm killed, removed
and stained. The primitive streak was clearly defined (PI. I,
Fig. i) extending from the posterior margin of the area pellucida
to the point of injury (x). The cells around the wound seemed
greatly increased in number and showed evidence of forward
growth which must have been stopped by the injury.
Other eggs, injured in the same way (Text-fig. i) were left in
-f---x
2
the incubator for a longer period, from thirty to forty-eight hours.
PI. I, Fig. 2 is a surface view of one of these embryos after forty-
eight hours' incubation. The embryo is well developed, fourteen
pairs of somites are present, and the heart is forming. The injured
area lies dorsal to the heart on a level with the anterior somites.
The brain region has failed to develop.
From Assheton's results, and from these just described, we
must conclude that the primitive streak and the greater part of the
later embryo form from that region of the unincubated blasto-
derm which lies behind the center, between it and the posterior
margin of the area pellucida. The question arises whether or
not the posterior margin of the area pellucida is a fixed region in
all eggs, and what the relation of the long axis of the embryo is to
the long axis of the shell. In Text-fig. 2, an egg is represented
372 Florence Peebles.
as opened above the blastoderm. The air chamber, which Hes in
the blunt end of the shell, is at the left, and the pointed end at the
right. The chalazae extend on each side of the yolk in the long
axis of the shell. In this position the blastoderm may be divided
into right and left halves, the arrow a-b indicating the median
plane of the bi-laterally symmetrical embryo. We then speak of
the region c as the anterior border, and d as the posterior border
of the blastoderm.
Assheton, in another series of experiments, has made two
injuries in the unincubated blastoderm (Text-fig. 3) one in the
center (x) and the other in the posterior border [y) of the area
pellucida. He found that the primitive streak appeared later
between these two injuries, and he concluded from this that the
point (jy) marks the posterior end of the embryo.
In order to distinguish the region of the primitive streak from
the rest of the area pellucida, I shall call it the radius x-y. This
radius with the corresponding one anterior to the center make the
diameter which represents the median longitudinal axis of the
embryo. In order to determine the constancy of the occurrence
of the embryo in this position I have kept the record of 100 eggs.
The eggs were taken from the nest on the same day that they were
laid. They were placed in the same position in a basket from
which they were transferred to the incubator. After incubation
for eighteen to forty-eight hours the embryo in every fertile egg,
with the exception of two, was found in the median line (Text-fig.
2 a-b). The two exceptions are shown in Text-figs. 4 and 5.
The first embryo (Text-fig. 4) was incubated eighteen hours.
At the end of this time the primitive streak had formed, but in-
stead of lying on the radius x-y it extended from the center to the
right side of the blastoderm and was bent towards the posterior
margin. The second egg (Text-fig. 5) was incubated for a period
of twenty-eight hours. The normal embryo lay at right angles to
the line a-b. In both of these eggs the chalazae were found in
abnormal positions, and the yolk membrane was wrinkled in many
places, showing that the yolk had been abnormally twisted in its
passage through the oviduct.
After I had discovered that the position of the normal embryo,
when undisturbed by twisting or shaking, is constant, I deter-
The Location of the Chick Embryo.
373
mined to find out, if possible, whether the embryo would form on
any other part of the blastoderm if development on the radius x-y
was prevented.
Experiment II. The blunt end of the egg was held in the left
hand so that the blastoderm lay on top of the yolk; a small window
was made immediately above it, and a series of injuries were made
with a hot needle in the radius x-y. The number of injuries was
dependent upon the size of the area pellucida. Usually there is
space enough to insert the tip of a No. 12 cambric needle in three
places between x and y (Text-fig. 6) before all the cells are
destroyed.
At the end of eighteen hours the eggs were killed, but no trace
of primitive streak in another region was found. About 60 per
cent, of the blastoderms showed a large hole where the area pellu-
cida had stretched apart in the growth of the blastoderm. No
5 6
evidence of the formation of the embryo around the margin of the
hole could be found.
The experiment was repeated and the eggs were incubated
from thirty to forty hours. An examination of these embryos
showed no development around the margin of the wounded area,
but in front of it and posterior to it some development had taken
place. In PI. I, Fig. 3, a surface view of an embryo of forty hours'
incubation is given. The brain is abnormal, but shows no lack
of material, the notochord is present, but greatly reduced in
length. There is some trace of the heart lying on each side of the
notochord, but back of it none of the embryo has formed. There
is no evidence of growth of the area pellucida in a posterior direc-
tion, but anteriorly it is the normal size and shape.
Another embryo incubated thirty hours is shown in PI. I, Fig.
4. In this embryo no brain developed but growth from the heart
374 Florence Peebles.
region caudad is evident. The injured area (w) is surrounded
by thickened ridges and back of the hole made by the wound the
notochord is present. Behind the notochord Hes the posterior end
of the primitive streak. No mesoblastic somites are present.
The object of these experiments vv^as to prevent development
along the radius x-y by killing the cells in the region w^here the
primitive streak develops. In this w^ay it was hoped that the
primitive streak might be formed in some other part of the area
pellucida. The results show very clearly that no other part of the
blastoderm is capable of forming the primitive streak. They also
show that the region of the unincubated blastoderm along the
radius x-y is the region from which the mesoblastic somites
develop, /. ^., the trunk region of the embryo. From these ex-
periments it seems evident that the position of the embryo upon
the blastoderm is determined before the egg has been incubated,
and probably before segmentation is completed, for some of the
eggs which I used were operated upon within two hours from the
time that they were laid.
II. THE DIRECTION OF GROWTH OF THE EMBRYO BEFORE AND
AFTER THE FORMATION OF THE PRIMITIVE STREAK.
MarshalF describes the growth of the blastoderm from the be-
ginning of incubation as follows: "After incubation has com-
menced, the blastoderm spreads rapidly, retaining its circular
shape. By the end of the first day of incubation it is about the
size of a sixpence, and by the end of the second day it has
extended nearly halfway round the egg."
According to DuvaP the edge of the blastoderm advances over
the egg at every point except at the posterior margin, and the
edges on each side of this point meet each other in the middle line
to form the primitive streak ("plaque axiale"). Assheton's^
experiments have proved, however, that the growth is symmetrical
as Marshall states.
'Marshall. Vertebrate Embryology.
^Duval. Loc. cit.
^Assheton. Loc. cit.
The Location of the Chick Embryo. 375
While the margin of the area opaca is symmetrical, that of the
area pellucida is not. During the first few hours of incubation
the two areas increase uniformly, but towards the fifteenth hour
the area pellucida begins to extend posteriorly, the anterior region
remaining spherical in outline.
Experiment /. The uniformity of growth in the anterior half
of the blastoderm can be seen in the following experiment. The
unincubated blastoderm was injured at three points (Text-fig. 7,
X, p and 0) ; the needle was inserted at the center (x) at the middle
point of the anterior margin (/>) and at the right margin of the
area pellucida (0). The injuries in the margin were at equal
distances from the center. After eighteen hours' incubation the
distance between x and p was the same as the distance between x
and (PI. I, Fig. 5) showing that the lateral and anterior growth
were the same. The primitive streak was formed, but its posterior
end (jy) was much further from x than x was from p, while the
distances before incubation were equal.
The results of earlier experiments^ led me to believe that the
region immediately in front of the primitive streak represents an
area of rapid growth, because an injury made in this region did
not affect one structure alone, but disturbed the organ covering a
large area. This is also true when the center of the unincubated
blastoderm is killed (PI. I, Fig. 2).
Experiment II. In order to determine the extent of growth in
an interior direction from the center of the blastoderm I injured
a point in the center (Text-fig. 8, x) and one on the same level at
the side (o). The eggs were incubated thirty-six to forty hours.
PI. I, Fig. 6 is a surface view of an embryo at the end of thirty
hours. The injury in the center of the blastoderm produced
great disturbance in the development of the embryo anterior to
the heart. No forward growth took place in the median line.
The wound (0) at the side, which did not move forward, is at a
level with the anterior somites. The normal growth of the margin
of the area pellucida on the side of the injury did not take place.
The margin is irregular and a peculiar rod of cells extends from
the marginal wound to the median line of the embryo.
^Peebles. Loc. cit.
376
Florence Peebles.
Experiment III. In another set of experiments the two injuries
were made about .5 mm. further forward (Text-fig. 9, x and 0).
The position of the injuries after forty hours is seen in PI. I, Fig.
7. The wound at the side (0) has advanced with the growth of the
blastoderm but the wound (x) in front of the somites has prevented
the formation of the head, and the embryo is reduced in length
anteriorly, the trunk and caudal regions are about the normal
length.
From these experiments it seems evident that the region in front
of the middle point of the area pellucida is the seat of active
growth in an anterior direction.
Experiment IV. In order to determine the extent of growth
posteriorly, two injuries were made, one in the center of the unin-
cubated blastoderm and the other in its posterior margin (Text-
fig. 3, X and y). The embryos were incubated thirty-six hours to
two days. They developed somites and medullary folds in the
area between the wounds. PI. II, Fig. 8, represents a surface view
at the end of thirty-six hours. Notochord and somites have
developed between the two wounds. The actual distance from
X to y before the experiment was i mm. After incubation it was
3 mm. showing an increase in length of only 2 mm. The normal
embryos at this age measure 4 mm. from heart to caudal end.
The results of these three sets of experiments show that the
embryo may be greatly reduced in length by preventing growth
anteriorly with the wound x and posteriorly with the wound y^
and that the area pellucida grows less rapidly at the sides than in
the median line.
Up to this time the experiments which I have described have
been made upon the unincubated blastoderm. The change in
the size and the shape of the area pellucida is comparatively
The Location of the Chick Embryo. 377
slight before the appearance of the primitive streak when the area
becomes pear-shaped.
Kopsch^ has found that when two wounds are made in an em-
bryo of twenty-four hours' incubation, at a distance of 2 mm., one
at the anterior, and the other at the posterior end of the primitive
streak, the embryo does not reach its normal size in later develop-
ment. The entire body is much shortened, and lies between the
two wounds. I have repeated this experiment, and have obtained
the same result, the primitive streak, in the eggs upon which I
have worked, is much longer (3 to 3.5 mm.) in a twenty-four hour
chick, and the anterior end is no longer visible, the head process
and the notochord are present.
Another series of experiments was made by Kopsch when the
primitive streak measured about 4 mm. A series of five injuries,
at 1.5 mm. spaces, were made along the side of the primitive
streak, and parallel with it. The embryos were incubated fifty
and one-half hours, and at the end of this time the regions of the
five wounds were located. Growth in length was greater in the
region back of the anterior end of the primitive streak than it was
in front of it.
I have already described experiments which I have made upon
the primitive streak and have tried to show that the anterior end
of the primitive streak of sixteen to eighteen hours represents the
region of the later embryo which lies back of the heart between the
anterior somites.
Experiment V. These experiments were repeated with some
modifications. Instead of injuring the anterior end alone, a
second wound was made at the posterior end (Text-fig. 10, x and
y). The embryo at the time of the operation was from sixteen to
eighteen hours old. After forty hours a normal embryo developed
but instead of extending posteriorly to the usual length it was
shortened 2 mm. Another egg injured in the same way (Text-
fig. 10) developed into an interesting embryo (PI. II, Fig. 10).
The posterior wound {y) healed so that no trace of it could be
discovered, but the anterior wound (x), through the further
^Kopsch. Loc. cit.
378
Florence Peebles.
growth of the embryo, was left on one side. The only disturbance
evident was in the medullary folds and somites on the side of the
injury.
Experiment VI. A wound at the posterior end of the primitive
streak is alone sufficient to shorten the embryo caudad. In PI. II,
Fig. II, a surface view of an embryo of forty hours' incubation is
shown. The wound (y) was made at the posterior end of the
primitive streak of eighteen hours. Fourteen pairs of somites are
present, and the embryo measures 3 mm. from the heart to the
anterior border of the brain. This region is normal, but growth
in a posterior direction has been stopped, by the wound, and the
length is reduced 1.5 mm.
Summary. The results from these experiments show that in
the formation of the third-day chick neither head nor tail region
can be taken as fixed points, indeed no one point on the blastoderm
can be said to be fixed. In the series of diagrams (Text-fig. 11,
A-G) I have indicated, in a schematic way, the method of growth
from the beginning of incubation until the third day. The growth
of the area opaca is symmetrical therefore it is not included in the
diagram. The line I-J represents the plane dividing the unin-
cubated blastoderm into anterior and posterior halves, and passes
through the region in the older embryos which corresponds to the
middle point of the area pellucida before incubation. From this
point growth proceeds in all directions in the plane of the blasto-
derm. The growth from the first to the twelfth hour is sym-
The Location of the Chick Embryo. 379
metrical. From the twelfth to the eighteenth hour the area pellu-
cida increases in length posteriorly. From the eighteenth to the
twenty-fourth hour growth continues posteriorly and also proceeds
in an anterior direction. From the end of the first day to the end
of the second day it advances from the heart in both directions,
more rapidly caudad than cephalad. After this time the tail and
head are folded off from the surface of the blastoderm.
III. THE ORIGIN OF THE MATERIAL FROM WHICH THE LATER
EMBRYO ARISES.
I have already spoken of Kopsch's conclusions as to the material
from which the embryo arises, so that I shall merely mention my
own results. If, as Kopsch says, the primitive streak represents
the entire embryo with the exception of the pre-chordal head
region, then the destruction of definite areas of the primitive
streak should result in a failure to develop the parts which arise
from the injured area.
Experiment I. The first experiment consisted in destroying all
of the primitive streak except its anterior end (Text-fig. 12).
This operation is very likely to kill the entire embryo as injury
to so large an area usually results in a spreading apart of the mar-
gins of the wound. The further development of an embryo
injured in this way may be seen in PI. II, Fig. 12. The embryo
is abnormal, but shows structures which indicate that when de-
prived of all of its material except the anterior end the primitive
streak gives rise to the first few pairs of somites; and that the brain
and notochord develop. The somites are much thinner than in
the normal embryo.
Experiment II. In another series of experiments the posterior
third of the primitive streak was destroyed (Text-fig. 13).
The destruction of this region resulted in an embryo (PI. II, Fig.
13), in which the entire caudal region was abnormal. The heart
and brain, which are not represented in the figure, were normal,
and fifteen to eighteen pairs of somites were formed in the anterior
trunk region. This result agrees with Kopsch's view that the
posterior third of the primitive streak represents the caudal region
of the embryo from the twentieth somite to the posterior end.
38o
Florence Peebles.
Experiment III. In a third series of experiments the middle
part of the primitive streak was killed (Text-fig. 14), leaving some
of the material in front, and some back of the v^ound. According
to Kopsch, in the later embryo the region from the first to the
twentieth somites should be lacking.
Nearly all of the embryos which I operated upon, in this way,
were so greatly disturbed by the wound that all development was
checked. In PI. II, Fig. 14, a surface view of the body region of
one of the embryos which developed further is shown. The brain
and heart were normal, therefore they are not included in the
figure. Posteriorly the wound {w) stretched apart, but anteriorly
medullary folds and ten or twelve pairs of somites are present.
This result indicates that at least ten or twelve pairs of the first
twenty somites come from the material in the anterior third of
the primitive streak.
IS 13 14 15
Experiment IV. Finally, the anterior third of the primitive
streak was killed (Text-fig. 15). After further incubation the
embryo developed a normal brain and heart in front of the wound.
The trunk region (without the brain and heart) of one of these
embryos is shown in PI. II, Fig. 15. Back of the wound (w)
eleven to fourteen pairs of somites are present. By comparison
with normal embryos of the same age I conclude that these somites
represent approximately, the tenth to the twentieth pairs, therefore
all of the somites between the first and tenth pairs have been
destroyed by injuring the anterior one-third of the primitive
streak. The notochord is also lacking in these embryos.
It is evident from these results that the primitive streak of eigh-
teen hours represents the material from which the trunk and tail
regions of the later embryo develop; that the posterior third of
the primitive streak represents the region back of the eighteenth
The Location of the Chick Embryo. 381
pair of somites, the middle third represents roughly, from the
twelfth pair to the eighteenth, while the anterior third supplies
material for those structures which lie between the heart and the
twelfth pair of somites, but does not include the chordal region of
the brain.
SUMMARY AND CONCLUSIONS.
I. The central point of the unincubated blastoderm represents
the anterior end of the primitive streak, and later, the region just
back of the heart; therefore, the greater part of the embryo
develops in the posterior half of the blastoderm.
2. The region midway between the center of the unincubated
blastoderm and its anterior border represents the head region of
the later embryo.
3. The position of the embryo on the area pellucida is fixed.
The long axis of the future embryo divides the unincubated
blastoderm into right and left halves and a line drawn through the
blastoderm in the long axis of the shell divides it into anterior and
posterior halves.
4. Destruction of the material of the unincubated blastoderm
between the center and its posterior margin does not result in the
formation of the primitive streak on any other radius.
5. The growth of the blastoderm is uniform up to the eighth
to tenth hour, and this uniformity is preserved in the later growth
of the area opaca, but from the tenth hour the area pellucida
begins to grow more rapidly in a posterior direction, then later it
advances anteriorly until it assumes an oval form. Up to the third
day the region immediately back of the heart (the anterior end
of the early primitive streak) is the center of growth in all four
directions, anteriorly, to the left, and to the right, and to a much
greater extent posteriorly.
6. Injury to the center and posterior margin of the unincubated
blastoderm results in a shortened embryo.
7. Injury at the posterior margin alone will shorten the embryo
by preventing growth in a posterior direction.
382 Florence Peebles.
8. Neither head nor tail region of the embryo can be taken as
fixed points, the growth at each end proceeds until the head and
tail become folded off from the blastoderm.
9. After destruction of all of the material of the primitive streak
except its anterior end a small embryo with eight to ten pairs of
somites develops.
10. The posterior third of the primitive streak furnishes the
material for the caudal region of the later embryo. The middle
third represents the trunk region, and the anterior third that part
of the embryo which lies between the heart and the tenth to
twelfth pairs of somites. The material of the primitive streak
does not enter into the formation of the brain.
The Woman's College,
Baltimore, June 1, 1904.
The Location of the Chick Embryo. 383
EXPLANATION OF PLATES.
Plate I.
Fig. I. Blastoderm 20 hrs. old. ;f, point of insertion of hot needle before incubation.
Fig. 2. Ventral view of embryo 48 hrs. old. Injury made in center of unincubated blastoderm
lies back of heart. Brain undeveloped.
Fig. 3. Surface view of forty-hour embryo in which the material along the radius x—y had been killed.
Heart, brain and notochord are present.
Fig. 4. Embryo 30 hrs. after operation described for Fig. 3. Heart and posterior body region
present.
Fig. 5. Primitive streak 18 hrs. old. The three black areas indicate the positions of the injuries
made upon the unincubated blastoderm.
Fig. 6. Surface view of embryo 36 hrs. old. The black areas indicate the wounds x and made
in the blastoderm before incubation.
Fig. 7. Embryo 40 hrs. old. The openings x and indicate the wounds made in the blastoderm.
Plate II.
Fig. 8. Embryo 36 hrs. old. The position of the wounds is indicated by the black areas x and y.
Fig. 9. Embryo 40 hours after injuries were made in the anterior and posterior end of the 18
hr. primitive streak.
Fig. 10. Embryo 12 hours older than that in Fig. 9 after the same operation.
Fig. II. Forty-hour embryo in which an injury had been made in the posterior end of the primitive
streak of 18 hrs. The black region (y) indicates the wound.
Fig. 12. Embryo incubated 36 hrs. after four-fifths of the material of the primitive streak was
destroyed, leaving only the anterior end.
Fig. 13. Fifty-hour embryo in which the posterior third of the primitive streak (w) was
destroyed. The brain which was normal is not shown.
Fig. 14. Surface view of embryo 30 hours after the middle part of the primitive streak was
destroyed. The normal brain is not shown.
Fig. 15. Embryo of same age as preceding one. The anterior third of the primitive streak
was destroyed. Heart and brain which are not given are normal, w in these figures indicates region
of injury.
THE LOCATION OF THE CHICK EMBRYO UPON THE
BLASTODERM.— FLORENCE PEEBLES.
PLATE I
The Journal of Experimental Zoologij. Vol. I
THE LOCATION OF THE CHICK EMBRYO UPON THE
BLASTODERM— FLORENCE PEEBLES.
PLATE II
The Journal of Experimental Zoology. Vol. I
REGENERATION OF HETEROMORPHIC TAILS IN
POSTERIOR PIECES OF PLANARIA
SIMPLICISSIMA/
BY
T. H. MORGAN,
With 20 Figures.
The regeneration of a heteromorphic head from the posterior
end of short cross-pieces of Planaria maculata. Figs. 1-5, led me
to ex2im'ine Planarta stTnplifissima^ in order to see if the same result
could be obtained here when short cross-pieces of the worm were
made. The regularity with which a heteromorphic head can be
obtained in the latter species when the old head is cut off just
behind the eyes, Fig. 10, led me to expect that short cross-pieces
from the body would behave in the same way as do similar pieces
o( Planaria maculata. The results have proven, however, in part
otherwise, for while heteromorphic heads do appear on short
cross-pieces from the anterior regions of the worm, Fig. 11, none
such develop from the posterior end of short cross-pieces from the
more posterior regions of Planaria snnplictssitna. On the contrary
these pieces regenerate a structure from the anterior cut surface
that appears to be a heteromorphic tail, and another tail from the
posterior cut surface, Figs. 12, 13. The result is a two-tailed
and not a two-headed piece. In order to determine if the
new anterior structure is really a tail, and not simply an undevel-
oped head, a number of experiments were carried out during the
winter and spring of 1903-04.
Before describing the results certain general considerations
must be spoken of that are intimately connected with the question
'The principal facts recorded in this paper were reported at the Christmas meeting
of the American Zoological Society, 1903.
^This is the same species which, in my earlier papers, has figured as Planaria
Itigubris. Stevens has recently determined that this worm is P. sim'plicissima.
386
T. H. Morgan.
of heteromorphosis in regions posterior to the old pharynx. An
important side light is thrown on the problem of axial polarity
and heteromorphosis by these relations.
Cross-pieces of Planaria si7nplicissima, that are not too short,
from the region between the head and the pharynx-chamber regen-
erate a head on the anterior and a tail on the posterior cut surface,
Fig. 14. The new pharynx is always situated at the posterior
edge of the old material. Similar cross-pieces from the region of
the pharynx-chamber also produce a head at the anterior end and a
tail at the posterior end. The new pharynx develops in the middle
of the piece in connection with the old chamber. Cross-pieces
from the region behind the old pharynx also regenerate a head at
Regeneration of Heteromorphic Tails. 387
the anterior and a tail at the posterior end. The new pharynx
always lies at the anterior end of such pieces, i. e., at the edge of the
old tissue, and therefore, as it were, in the posterior part of the
new head that has developed. Fig. 15. It is this relation of the
new pharynx to the old part that first demands especial considera-
tion, for, at first sight it is not clear why the pharynx in these pos-
terior cross-pieces should shift to the anterior end, and not lie, as
in the more anterior pieces, at the posterior edge. If it did so it is
obvious that it would appear in a region posterior to that in which
the normal pharynx lies in the old worm, and it seems that this
cannot take place. The posterior cut surface can form only that
part of the tail that lies behind it in the old worm. The anterior
cut surface can also produce all that lies in front of it in the old
worm, including the pharynx, although the proportionate distances
apart of the new structures may be at first very different from
those in the adult or embryonic worm. We touch here on one of
the fundamental questions of polarity to which I shall hope to
return at another time.
If a heteromorphic tail were to develop on the anterior cut
surface of a short posterior cross-piece, what should we anticipate
in regard to its relation to a pharynx .? Should we expect to find
a pharynx in the new tail turned in the opposite direction, /. e.,
pointing towards the tip of the tail .? But if the new structure at
the anterior end is a heteromorphic tail why should it develop
a pharynx at all, since this never develops at the posterior
end of cross-pieces from this region .? Should we not rather
expect a heteromorphic tail to behave in this respect in the same
way as the orthomorphic structure .^ This appears to me to be
the correct point of view and the results of experiment seem to
bear out this anticipation.
Let us apply the same point of view to the regeneration of a
pharynx in heteromorphic heads from cross-cut pieces of the ante-
rior regions of the worm. It has been pointed out above that the
new pharynx appears at the posterior end in case a tail develops
at this end. Suppose, however, a heteromorphic head instead of
an orthomorphic tail develops at the posterior end. It is clear,
from our point of view, that no pharynx should develop, and I
388
T. H. Morgan.
have found that none such is present as a rule. In the few cases
in which a pharynx appeared in the middle of the piece, the piece
may have come from the region near to or through the pharynx-
chamber of the old worm.
Turning now to the results of regeneration of very short cross-
pieces of Planaria stmplicissima, it was found that double headed
pieces are sometimes obtained from the more anterior regions,
Fig. II, as in P. ?naculata, Figs. 2-5. When short cross-pieces
of P. simplicissima are cut off posterior to the pharynx-chamber a
number of them produce a head at the anterior end and a tail at
the posterior end, especially if they are rather long. Fig. 15; but
short pieces and sometimes some of the longer pieces also produce
quite often a pointed structure at the anterior end. In the ma-
jority of cases these anterior structures never develop into any-
thing different, and resemble a tail in all respects. In a few cases,
however, the pointed structure may become a head after some
Regeneration of Heteromorphic Tails. 389
time. The possibility that all the anterior pointed structures
may be only undeveloped heads must therefore be given seri-
ous consideration. That they are not such in many cases is
shown, I think, by the following facts. In the first place the
movement of a piece in which the anterior head is undeveloped
is very different from that of the two-tailed pieces, and reveals the
nature of the new part; for, while the former crawls forward as do
the pieces when first cut from the worm and as do those that
develop an anterior head, the two-tailed pieces remain fixed in one
place in the dish, and, if disturbed, fail to move in any definite
direction. This is what we should anticipate if two tails were
present working in opposite directions.
In the second place an orthomorphic pharynx appears as a
rule when the head is delayed in its development while none such
appears in the two-tailed pieces. In the third place the peculiar
motion of the anterior end when it is irritated is similar to that of
a tail and not like that of a head. Finally, the development in
one case to be mentioned below of a two-tailed piece with pharynx
in each tail shows, beyond a doubt, the possibility of the develop-
ment of a heteromorphic tail in these worms.
After the short cross-pieces have been cut off for some time it
is difficult to distinguish the anterior from the posterior end and to
know which is the anterior heteromorphic and which is the normal
posterior orthomorphic tail. In order to distinguish these apart,
I cut off pieces obliquely at one end; in one set the anterior end
being the oblique one. Fig. 16, in the others the posterior, Fig. 18.
This necessitated increasing somewhat the length of the pieces and
brought about in consequence an increase in the number of the
pieces that regenerated a head at the anterior end. The record of
one set of experiments of this sort is given here.
On May 7, twelve short cross-pieces were cut off just behind the
pharynx. The anterior end of each was oblique. On May 20
there were alive three two-tailed pieces; the rest having died.
Twelve short tail-ends regenerated new tissue at the anterior
end which in two cases at least appeared to be tails.
In another series twelve short cross-pieces were cut off behind
the pharynx. The anterior end was square and the posterior end
390 T. H. Morgan.
oblique. Of the eight survivors all had pointed anterior and
posterior ends.
The pieces just behind the last set (with oblique anterior ends)
gave five two-tailed pieces, and one piece with head a.nd tail.
The tail-ends of this set had the posterior tip cut off. The three
that remained alive developed a pointed anterior end.
In another series like the last, the first pieces produced six two-
tailed forms; the second six two-tailed forms; and the tail-ends five
two-tailed forms.
In several other cases I allowed one end of the piece to close
for two or three days before cutting it off. In this way the mor-
tality of very short pieces, which is otherwise very great owing to
their immediate disintegration, is lessened. Some of these pieces
also gave two-tailed forms.
In only one case did I obtain a piece in which a pharynx was
present in each tail, and in each turned outward toward the tip
of the tail, as shown in Fig. 12. The exact location of this piece is,
I am sorry to say, uncertain. It came, in all probability, from
the region just behind, or including some of, the old pharynx
region. I am inclined to think that the latter is the more probable
location, since the cut may sometimes include somewhat more or
less than is intended. The direction taken by the pharynges in
these pieces shows beyond a doubt that one of the two tails is a
heteromorphic structure, and this lends support to the interpre-
tation that I have given to the other cases, in which there is no
pharynx in the heteromorphic tail, and where none should be
expected to be present on theoretical grounds.
Two other kinds of apparently heteromorphic structures have
been met with in carrying out these experiments. In one case a
piece of P. maculata, whose posterior end was cut off very
obliquely, regenerated one head on the anterior cut surface and
another also on the right side of the posterior cut surface, as shown
in Fig. 8. A tail also developed on the posterior cut surface at the
left side, which is also the more posterior end of this surface. In
this case we must look upon the long edge of new tissue on the
posterior surface as producing a head at one end and a tail at the
other, very much as occurs when a longitudinal piece of the worm
«
Regeneration of Heteromorphic Tails. 391
is removed. The case recalls those in which the worm is split
from the posterior end far forward and a head develops at the
anterior end of the cut surface on one or on both sides. I have
discussed the meaning of this case elsewhere.^ In a strict use of
the word heteromorphosis, as I have tried to use it for purposes
of greater clearness, neither this case nor that of the split-worm
can be looked upon as an example of axial heteromorphosis, since
the result depends largely, apparently, on the new part alone
without relation to the old, and the head and tail are orthomor-
phic from this point of view.
In some other cases in which the anterior end is very oblique,
two structures appear on the anterior edge, as shown in Figs. 19-
20. One of these is a head and the other appears from its struc-
ture and movements to be a tail. If so, the case is comparable to
the last one, and shows the converse condition. Here the tail on
the side of the anterior cut surface cannot be looked upon as an
example of axial heteromorphosis. It is rather an orthomorphic
structure, since it stands in this relation to the remainder of the
new material on the anterior cut surface. Both cases, however,
present something of a paradoxical relation.
The results described in the first part of this account recall cer-
tain conditions that I have recently described in connection with
the regeneration of Dendro caelum lacteum. It had been shown
by Lillie that posterior pieces cut off just in front of, or through,
or behind the pharynx-chamber do not regenerate an anterior
end. A histological examination of the anterior end of such
pieces showed me that a certain amount of new tissue is formed
at the anterior cut surface, and it was not apparent why the re-
generation should not go further and produce a new anterior end.
The results with Planaria simplicissima suggest, although they by
no means prove, that the anterior part that regenerates in Dendro-
ccelum may be a heteromorphic tail. For the present, however,
I wish to leave this question open, until further work reveals the
nature of the anterior part in this worm. There are some gen-
eral considerations in connection with the problem of polarity
^Regeneration. 1900.
392 1 . H. Morgan.
and of heteromorphosis that may be very briefly touched upon at
this time. Although I have not hesitated in earlier papers to
speak of polarity as a factor in regeneration, I have always
tried to be careful to state that we are really entirely ignorant in
regard to its nature. When we see the polarity suddenly reversed
in cases of axial heteromorphosis it appears that this ought to
throw some light upon the nature of the factor itself, yet despite
the numerous surmises that have been made of a material, — chem-
ical, or electrical nature — we still remain totally in the dark as
to what factors determine the stereometrical relations of the new
part. The following facts appear, nevertheless, to have an im-
portant bearing on this topic, and while they do not offset an imme-
diate solution of the problem, yet they may point in the direction
in which an analysis may ultimately be undertaken.
In the more highly specialized forms the question of what re-
generates appears, in part, to be connected with the nature of the
material, or with the kinds of the material that give rise to the
new cells, and the relation of direction is less apparent. The tail
of a tadpole regenerates only a tail, even at its anterior end. The
same appears to be true for the leg of the salamander from certain
results that I have obtained, which are as yet unpublished. In
the earthworm as shown by Morgan^ and by Dimon^ the
regeneration of an orthomorphic head is connected with the
presence of the anterior structures of the worm, while from the
part containing the intestine — including by far the greater length
of the worm — only a tail is, as a rule, regenerated, even from the
anterior cut end. In these cases it appears that the nature of the
material must decide the character of the new part, and the polar
relations do not come conspicuously to the front, although that
something of the sort still enters into the problem is shown by the
slower rate, and, in some cases, by the less perfect form of the
heteromorphic growth.
On the other hand, in less specialized forms the polar relations
appear to play a more conspicuous role. In Lumbriculus a head
'Anatomischer Anzeiger. Bd. 15, 1899.
^Journal of Experimental Zoology. Vol. I, No. 2, 1904.
Regeneration of Heteromorphic Tails. 393
may regenerate from an anterior cut surface, and a tail from a
posterior cut surface throughout a very considerable region of the
body. In planarians and in hydra similar facts are known. That
the specification of the tissues or parts plays a role even in these
cases is probable, as shown by the cases of heteromorphosis that
I have described. To many writers it has seemed that the factor
of polarity maybe something in the nature of a crystallizing force —
to use the nearest analogy at hand — a sort of perfecting or com-
pleting principle. Newer results have modified our ideas as to
this form of explanation, if such an analogy can be called at all
an explanation. The fact, for example, that in the earthworm
and in planarians the new head may be very short in comparison
to the part that is missing indicates that a completing force cannot
be acting from the cut surface forwards, but whatever the nature
of the factor it must in large part work from without (surface)
inward (/. e., toward the cut end). This point has been already
urged by myself, and by Driesch.
It is very significant, I think, to find that in planarians the
shortness of the piece is a factor that enters into the problem as
to the character of the new part. I have suggested tentatively
that this means that in Planaria maculata the tendency is stronger
for the new structure to become a head than a tail, and that
when the influence of polarity is removed a head appears on each
end of short cross-pieces. In other worms, as in Planaria sim-
plicissiTua, the tendency in certain posterior regions to produce a
tail is stronger than that to produce a head, and two tails appear
when the polarity is reduced or removed. Why should the length
of the piece be so important a factor .? Can it be that there is
a greater difference, chemical or physical, between the two ends
of a longer piece, so that a stronger polarity is present ^ In short
pieces, from this point of view, the ends being near together are so
much alike that the polarity is correspondingly reduced, and,
under these conditions, the specification of the material of the old
part is not sufficiently strong to determine the nature of the new
part. These and many other equally obscure questions remain
for future investigation to explain.
BIOLOGICAL STUDIES ON CORYMORPHA.
I. C. PALMA AND ENVIRONMENT.
BY
HARRY BEAL TORREY.
With 5 Figures.
CONTENTS.
Introduction.
I. Description of C. palma.
II. Habitat; food.
III. Activities of the polyp.
a. Muscular movements.
b. Geotropism; axial ceUs.
c. Locomotion; amoeboid cells.
d. Circulation; cilia.
IV. The young hydroid.
Summary.
INTRODUCTION.
This is one of a series of papers which deal with some of the
phenomena of growth, differentiation and development in
Corymorphay from different points of view. The study of de-
velopmental mechanics has long since ceased to consist merely in
an analysis of the development of egg and embryo. That
regenerative development must be included goes without saying;
and it is with the feeling that the normal activities not ordinarily
considered in the category of developmental processes should be
included also, that I have incorporated much of what may at first
sight appear to be purely physiological material.
Corymorpha is an exceptionally attractive basis for such an
investigation. In the first place, it combines remarkable powers
of regeneration with a simple development from the egg, the non-
396 Harry Beal Torrey.
sexual origin of sexual individuals, and powers of movement that
are unusual for a hydroid. In the second place, very little is
known of its biology, and its value for experimental work has not
yet been generally realized. No account of its regeneration has
been published, with the exception of a brief reference to it in a
former paper by the present author ('02, p. 41). Up to this time
our knowledge of the egg-development has been based upon three
stages described and figured by Allman ('71):^ the sessile planula,
the polyp with six proximal tentacles, and the polyp with sixteen
to twenty. These Allman took for stages in the development of
what he called "frustules," minute bodies cut off from the pro-
cesses that develbp near the base of the stem and really give rise
to the filaments of the hold-fast. Had he seen the eggs on the
medusae in his aquarium, this pardonable error would not have
been made. Agassiz ('62) and Allman ('63) have given brief
accounts of the natural history of the hydroid and the develop-
ment of the medusa.
These works, with several of a taxonomic character, and a
recent paper by May ('03), comprise the scant publications on
Corymorpha relating to the subject of this paper.
I. DESCRIPTION OF C. PALMA.
The nutritive polyps of C. palma are solitary. The stem may
reach the length of ten centimeters, tapering gradually from a
diameter of perhaps six millimeters, near the base, to a narrow
neck which supports the hydranth. It is covered for about its
proximal third with a thin, non-supporting layer of perisarc.
Within is a solid axis of immense vacuolated cells. These have
almost obliterated the cavity of the stem, which persists as a num-
ber of small, longitudinal canals lying immediately under the thin
mesogloea, and usually made conspicuous by their green tinted
walls.
'A Monograph of the Gymnoblastic or Tubularian Hydroids. London, 1871.
The section in this magnificent monograph devoted to Corymorpha is a reprint of
Allman's paper in the Annals and Magazine of Natural History for January, 1863,
p. 1, with but slight verbal changes and the addition of figures.
Biological Studies on Corymorpha. 397
The hydranth has a single whorl of eighteen to thirty proximal
tentacles with a spread of more than twenty-five millimeters.
The proboscis, terminating in the mouth, is crowned with forty to
sixty distal tentacles. Just within the proximal tentacles are
several peduncles which bear numerous medusoid gonophores.
The stem is anchored by a tangle of filaments which arise on
the longitudinal canals, beneath the perisarc, usually in pairs. ^
II. habitat; food.
Corymorpha palma is a semi-tropical species, dwelling farther
to the south than any of the other North American species of the
genus. It has been found as yet only in two localities: in San
Diego and San Pedro harbors, both on the southern coast of
California. It lives under similar conditions in both places. At
San Diego it was found in a slough near the mouth of the harbor,
on a muddy bottom which was exposed at mean low water. At
San Pedro it has flourished at various points in the harbor, always,
however, on muddy flats. It occurs usually in definitely circum-
scribed patches, which change their position apparently with
much caprice from year to year. A favorite location is along some
small stream that drains the mud flats as the tide ebbs.
Copepods are numerous on the mud, which often carries
patches of green composed of diatoms and other chlorophyl-
bearing protista. All of these organisms seem to serve as food for
Corymorpha, though the copepods form the staple article of diet.
III. activities of the polyp.
Corymorpha captures crawling diatoms and copepods by bend-
ing its column in a half circle and sweeping the sand with its
tentacles. Floating organisms are caught when the column is in
its usual erect position, with proximal tentacles fully extended.
The oral tentacles are almost always active, bending restlessly
now outward, now inward, now moving simultaneously, now
independently. The proboscis is extremely mobile, capable of
lengthening into a narrow stalk or contracting into a sphere, or
'For a diagnosis of tlie species, see my paper just referred to.
398 Harry Beal Torrey.
turning inside out, or carrying the mouth with its prehensile
tentacles to the bases of the proximal tentacles on all sides. The
proximal tentacles are comparatively quiet. For minutes at a
time they may be held in full extension, motionless, curved grace-
fully back from the proboscis, on a vertical stalk. Now and then
one may twitch toward the mouth. Occasionally all may wave
inward together, grasping the proboscis tightly. The points of
the tentacles may be, but usually are not, carried directly toward
the mouth.
a. Muscular Movements.
So far as I am aware, the reactions of hydroids (with the ex-
ception of Hydra) to different sorts of stimuli have never been
studied. Medusae, on the contrary, have been the subjects of
extended investigations by Romanes ('76, '']']), Eimer ('78), whose
paper I have not seen, and Nagel ('93, '94)- In certain respects,
Corymorpha and some of the craspedote medusae (Carmarina,
Sarsia) respond similarly to similar stimuli. For instance, the
proboscis of each may move toward a point of stimulation not on
it; and increasing the stimulation of a tentacle may increase the
number of tentacles taking part in the response, and leads finally
to the contraction of the body of the animal. Hydra and
Corymorpha^ however, resemble each other more closely in their
responses than either resembles a medusa. In general, similar
structures respond similarly, but the tentacles of neither Hydra
nor Corymorpha react to odorous substances, while according to
Nagel ('93), the tentacles of Carmarina hastata do. Such excep-
tions, coupled with obvious differences in structure and habits
between polyp and medusa, make it necessary to treat each case
individually.
The large size of Corymorpha makes it an unusually favorable
object for experimentation in this direction. Experiments with
mechanical, chemical and thermal stimuli brought out the follow-
ing facts :
Mechanical Stimuli. Each proximal tentacle responds to a
touch or pinch from forceps by contracting in the same direction
with the same strength, whether the stimulus be applied at the
Biological Studies on Corymorpha. 399
base or the tip, on the oral, aboral or lateral surfaces. The re-
sponse is always a bend inward, never outward. In this respect
it differs from the tentacular responses in some anemones (Cri-
brina, Sagartia), where a tentacle commonly reacts to a slight
touch by bending sharply at and toward the point stimulated.
This reflex is clearly advantageous to the anemone, which it
enables, to a limited extent, actually to pursue its prey. It is
supplemented by another. As soon as the tentacle, which is
adhesive, seizes the stimulating object, it contracts, carrying its
capture to the mouth, over which it bends. Then, by means of
the cilia with which the tentacle is covered, the object, at least if
available for food, is swept off the end of the tentacle and dropped
upon the lips.
While the general direction of the movement of the tentacles
does not vary, the intensity of the contraction varies with the
intensity of the stimulus. A touch or slight pinch produces a
waving of the tentacle toward the proboscis, though without
reaching it; and the tip of the tentacle is not directed toward the
mouth. A stronger stimulus may cause the tip to touch the distal
tentacles, may even cause the tentacle to coil against the proboscis.
When the stimulus reaches a greater intensity, it may induce
simultaneous movements in several or all the proximal tentacles.
Before this point is reached, however, it is able to set up move-
ments in the distal tentacles and proboscis. If a proximal ten-
tacle contracts, an effect is often evident among the distal ten-
tacles, even though the proximal tentacle has not touched them.
This effect is manifested either by a simultaneous downward
movement or an indeterminate waving of all the tentacles, or by a
downward motion of a few nearest the tentacle stimulated.
These movements of the distal tentacles may occur without any
apparent movement in the proboscis. If, however, the stimula-
tion of the proximal tentacle is increased (occasionally a very
slight stimulus is sufficient to produce the movement) the pro-
boscis also may bend, carrying the mouth and distal tentacles
toward the tentacle stimulated. This is a coordinated reflex of
the same purposive aspect as the movements of the proboscis of
the medusae Sarsia, Tiaropsts (Romanes, '77) and Carmarina
400 Harry Beal Torrey.
(Nagel, '94) toward stimulated points on the sub-umbrella.
While the excitation may be transmitted by means of the nerves of
the tentacles and proboscis, certain facts indicate that the direct
pull of the tentacle on the base of the proboscis serves at least to
reinforce the impulse and aid in guiding the tentacles and pro-
boscis in the proper direction. For example, the proboscis never
bends until the stimulated tentacle contracts, although this con-
traction may be delayed half a second or a second after the
stimulus is applied — an unusual reaction time; it does bend, how-
ever, immediately upon the contraction of the tentacle. Again,
when a simple grip of the forceps does not cause a movement of
the proboscis, the movement may be induced by adding to the
tactual stimulus a definite tension stimulus by pulling the tentacle
or preventing it altogether from shortening.
Not only may stimulation of a proximal tentacle be followed by
movements of distal tentacles and proboscis, but by movements
of the stem as well, which contracts strongly when the stimulation
is vigorous. Only that part of the stem ordinarily contracts
which is not invested with perisarc, though the latter is too thin to
be an effectual hindrance to contraction in the basal region, which
at times does shorten considerably.
When one distal tentacle is pinched, the response is similar to
what occurs when the disal tentacles respond reflexly to a stimu-
lation of a proximal tentacle; that is, several or all the distal ten-
tacles may wave outward and downward together, or indiscrimi-
nately outward. In response to this stimulus, the movement is
always away from, never toward the mouth; in this respect it is
contrary to the direction of the movement of the proximal ten-
tacles. After the first reaction, however, the tentacles may move
actively and singly toward and away from the mouth. This is
the characteristic reaction when the stimulation is prolonged.
The presence of a large food organism in the proboscis cavity will
cause such movements, which will persist until it has been
entirely swallowed. They are only moderately efficient, for the
outward movement of each tentacle is quite as strong as the in-
ward movement, and the tentacle retains its hold on the captured
organism for an instant only. They are indeed far less efficient
Biological Studies on Corymorpha. 401
than the tenacious tentacles of anemones with their more definite
movements.
When the proboscis is pinched at its base, it bends toward the
point of stimulation, the distal tentacles waving. A pinch at any
point, of a sufficient intensity, will induce characteristic move-
ments of the distal and proximal tentacles and a shortening of the
stem. If the stimulus is not too strong, only the proximal and
distal tentacles in the vicinity of the point of stimulation will react
simultaneously.
A slight stimulation of the stem may produce characteristic
movements of both sets of tentacles. The effect is not related to
the position of the point stimulated; none of the parts reacting
appear to distinguish the direction from which the impulse comes.
The stem may also shorten, even to half its original length. The
shortening usually takes place in the distal naked portion. The
proximal third, however, which is covered with perisarc, may also
contract; the perisarc is very delicate and in no way interferes with
this or any other movement of the stem.^
Chemical Stimuli. Several substances were used: flesh in the
shape of pieces of the shore gastropods Littorina and Acmcea, and
boiled ham; clove oil, alcohol and acetic acid. In no case did the
meat juices have the slightest appreciable effect on the hydroid;
the same may be said of the clove oil. Only when touched by a
stream of strong alcohol or acetic acid from a pipette did tentacles
or column respond; the acid killed the former almost instantly.
This response is evidently of the tactile order — as when an
irritating fluid is poured on the hand. Substances which have
for our perception odors and flavors, appear to produce no
reactions.^
^The proportion of stem covered by perisarc is based on measurements of
expanded individuals, under normal conditions. When a hydroid has been standing
in the same water for a week or two, it usually becomes much attenuated, and the
part of the stem invested with perisarc then often appears longer than the distal
naked portion. Often the ratio of the covered to the naked portion of the stem may
become that of the larva (Fig. 5).
^Loeb ('95) has already criticised the use of the words "olfactory" and "gusta-
tory" to describe the reactions to chemical stimuli of animals of whose consciousness
we are as ignorant as we are of the consciousness of the Coelenterata.
402 Harry Beal Torrey.
Thermal Stimuli. A rapid rise in temperature of several de-
grees, caused by flooding the hydroid gently with warm water from
a pipette, produced a general contraction of the same character as
the response to a strongtactuai stimulus.^ Gradual changes in tem-
perature aflPect both irritability and the rate of growth, increase of
temperature resulting in increased irritability and more rapid
growth, and vice versa; the limits, however, were not determined.
The reactions of Corymorpha to the various stimuli considered
above may be summarized as follows: All parts of the hydroid
are very sensitive to mechanical stimuli, irritating chemicals and
abrupt increases of temperature. Proximity to odorous sub-
stances, especially flesh, which might serve as food, awakens no
appreciable response until the substances are actually touched.
Food organisms, therefore, are probably detected only when they
strike the hydroid. The mechanism for capturing them is inter-
esting on account of the definite but dissimilar responses of the
two sorts of tentacles and the coordination exhibited in the activi-
ties of all the parts. The proximal tentacles with their great
spread (which sometimes almost equals the length of the stem)
serve as the chief means of advertising the presence of food and
carrying it to the mouth. These functions are sufl&ciently well
discharged b)^ a movement in one direction only — toward the
mouth; but the absence of the preliminary movement in the direc-
tion of the stimulus, which has been noted among the anemones,
entails a certain loss of efficiency. This loss of efl&ciency is com-
pensated for to some extent by the movements of the distal ten-
tacles and the proboscis. The stimulus which causes the move-
ments is in the great majority of cases liable to be applied to the
proximal tentacles, on account of their relatively much greater
spread. And apparently because of this, whether acquired by
habit or selection, the first movements of the distal tentacles in
response to direct or indirect stimulation are downward and out-
ward, toward the proximal tentacles; that is, toward the usual
^This contraction is the result of muscular activity, does not concern the axial
endoderm (to be especially considered later under Geotropism) and is not to be com-
pared, therefore, with such growth processes as were shown by True ('95) to follow
in radicles of seedlings, transference from water at 0°C. to water at 18°-21° C.
Biological Studies on Corymorpha. 403
point of stimulation and away from the mouth. These move-
ments, together with the tendency of the proboscis to bend toward
the point of stimulation, carrying the distal tentacles with it,
undoubtedly supplement the movements of the proximal tentacles
in bringing food to the mouth, and raise the average of efficiency
of the prehensile mechanism.
The contractions of the stem muscles, determining a limited
range of movement for the hydranth, may be of advantage to
Corymorpha. The rapid shortening of the stem following strong
stimulation, however, can have no value as a part of the
mechanism of prehension, nor does it have any apparent useful-
ness as a means of defense against predatory enemies.
h. Geotropts?n; Functions of the Axial Endoderm.
Up to this point we have been considering the effects produced
by the contractions of muscles, "^ in tentacles, proboscis and column,
under certain sorts of stimulation.
We may now consider another type ot motor reaction induced
by another sort of stimulus which appears ultimately to affect
another tissue element. This is the tendency of the stem, in
assuming its most characteristic attitude, to turn directly away
from the center of the earth, by what seems to be a change in
the turgidity of the axial endoderm cells incited by the stimulus
of gravity.
It will be unnecessary to enter into an extended discussion of
the phenomena of geotropism," which are familiar to all. A few
words will suffice for the purposes of this paper.
^There are both longitudinal (ectodermal) and circular (endodermal) muscle
fibers in both proximal and distal tentacles, proboscis and column. As might be
expected from their activities, the circular fibers in the proximal tentacles form a
much weaker sheet than the longitudinal, except where each tentacle joins the body
of the hydranth. There the circular fibers are aggregated into a strong bandlike
sphincter, and there the tentacles are wont to break away from the hydranth under
unfavorable conditions. Such a habit of casting the tentacles seems to be character-
istic of certain anemones, notably Bolocera, and is accomplished by a similar
sphincter.
^Davenport has distinguished between the responses of free and fixed organisms
to gravity, following Schwartz in applying to the former the term "geotaxis," and
applying to the latter the term " geotropism." With the facts which follow in mind,
it will be difficult, I believe, to see any advantage in this distinction, and it has
accordingly been disregarded.
404 Harry Beal Torrey.
Gravity affects both free and fixed organisms, and, so far as we
are concerned with it, determines orientation, direction of loco-
motion, and direction of growth. In free organisms, orientation
may or may not be accompanied by locomotion. Davenport has
cited the infusorian Spirostomum as an organism which may
belong in the latter category. In this case, orientation is finally
due to the action of cilia with which the animal is clothed; if loco-
motion is associated with orientation here, it is very slight and
inconspicuous. Cerianthus, whose negative geotropism was first
considered by Loeb ('91), orients itself by means of muscular
action. Though a free organism, it pursues a sedentary habit.
The same tendencies are manifested by sand-dwelling anemones.
Among fixed forms may be considered (i) those which are
attached aborally, but are also capable of some degree of loco-
motion, such as most of the anemones and the hydroid Corymor-
pha; (2) those which are permanently attached, such as most of
the hydroids; and with these must be classed plants, especially
seedlings. I have recently referred ('04) to the geotropism of the
anemone Sagartia davisi, the orientation being accomplished, as
in Cerianthus, by muscles. In the discussion of the geotropism
of Corymorpha to follow, it will be shown that the orientation of
the column is probably accomplished, not by muscles, but by
means of growth processes comparable with the growth processes
responsible for the orientation of seedlings and, presumably,
of geotropic hydroids.
That the characteristic position of the stem in Corymorpha is
not due to a difference in the specific gravities of the distal and
proximal regions is apparent when it is seen that not only is the
hydroid both proximally and distally heavier than water, but
distally it is heavier than it is proximally. If a hydroid is placed
in a jar of water after having been slipped out of its proximal
investment of perisarc, weighted down as that is by sand clinging
to the filaments of the hold-fast, it sinks at once, hydranth first, and
lies upon the bottom until the proximal end becomes attached.
When this occurs, the stem begins to rise, and in an hour is erect.
The result is in the end the same whether the hydranth is
present or absent, whether the stem is cut so that the proximal
Biological Studies on Corymorpha,
405
portion is two-thirds, one-third, or even one-eighth the original
length of the stem. Evidently the stem is generally responsive to
the geotropic stimulus. The only difference lies in the time con-
sumed in reaching the vertical position. The longer the stem the
shorter the period.
The result is also in the end the same whether the hydroid is
hung vertically upside down, by the proximal extremity, or right
side up, by the "neck," just below the hydranth, or by the middle.
And it matters not whether the hydranth and foot are both or
either one present or absent.
Fig. 1.
Three vertical stems cut at different levels, which were parallel with S one hour
before.
. Numerous experiments justify this summary. Typical cases
will be described. To exclude the possible influence of light and
oxygen on the direction of orientation, the hydroids were com-
pelled to orient themselves in sealed jars quite full of water, which
were placed in dark closets. Check experiments in the light and
in open aquaria gave identical results. Further, about a dozen
individuals were subjected for three hours to light coming from but
one direction, without any observable result on their orientation.
4o6 Harry Beal Torrey.
Several uninjured hydroids and three (Fig. i) which had been
cut respectively two-thirds, one-third, and one-eighth the length
of the stem from the proximal end, all weighted with sand as
usual, were placed in a dark closet in a jar full of water. As soon
as they became erect, the jar was tilted at an angle of forty-five
degrees, the stems of the hydroids remaining parallel with its
sides. In an hour all the stems had become erect. The distal
pieces cut from the three mutilated stems were still lying on the
bottom, unable to rise for lack of hold-fasts. The jar was brought
back to the vertical, and in another hour the stems had swung
through forty-five degrees to the vertical, in the opposite direction.
The time required for such changes varies considerably, but in the
same experiment the shorter the piece the longer the time — ten to
twenty minutes longer in the case described. The movement was
constantly toward, never away from, the vertical position finally
assumed, and did not suggest in the slightest the method of trial
and error (Jennings, '04).
The time required for an inverted hydroid to right itself is much
longer. Two hydroids, hung vertically by a string tied to their
proximal ends, were horizontal within seven hours, inclined up-
ward at an angle of forty-five degrees in thirty hours, and vertical
in their normal position in forty-eight hours. In another experi-
ment the hydranths and hold-fasts were removed from two
hydroids which were hung on a thread piercing them near their
proximal ends. They righted themselves in twenty-four hours.
When hung from strings around their necks, the stems remained
as they were, vertical, whether they possessed hydranths or not.
A stem lacking both hydranth and hold-fast was pierced through
the middle by a glass needle which was suspended horizontally.
In an hour the stem was vertical, distal end up, and remained thus
for several days.
Fig. I shows fairly well the important fact that it is unmistak-
able in the animals themselves, that the stem in turning toward
the vertical does not bend locally but generally. Corymorpha
resembles the stems of plant seedlings in this respect, as well as
in the preliminary bending of the stem beyond the vertical.
Whether the bending travels progressively from oral to aboral end
Biological Studies on Corymorpha. 407
was not determined. There would seem to be more than an
accidental association in these widely separated organisms of
similar phenomena with similar structures.
Another experiment demonstrated that a long exposure in an
inverted position to the influence of gravity has no effect on the
response of the individual when returned to normal conditions.
Two hydroids were suspended vertically upside down in glass
tubes, to prevent them from righting themselves. At the end of a
week they were freed, and oriented themselves normally.
There are two elements in the stem to which the foregoing
results might be referable: the muscles, and the axial endoderm.
To solve the problem which thus presented itself the following
typical experiments were performed.
A hydroid was decapitated and three wounds made at moderate
intervals half through the stem on one side. The stem bent
toward the opposite side, showing the greater potency of the un-
harmed muscles. When, however, it was laid upon the bottom
of the aquarium, wounded side uppermost, it assumed an erect
position in about an hour. It moved toward the muscularly
weaker side as rapidly as it would have done it had the stem been
intact. In whatever relation to the bottom the wounds were
placed, the stem regained a vertical position in about the same
time — in all cases very gradually. Another individual was cut in
a similar manner, though in this case there were eight or nine cuts
alternately on one side and the other. These cuts interrupted
the continuity of all the muscles except for very short distances on
the stem, which lay quite limp on the floor of the aquarium imme-
diately after the operation. Within two hours, however, it had
stiffened into an erect posture. The wounds in these cases did
not close for many hours after the stem had become erect.
Only the continuity of the longitudinal muscles was broken by
the wounds, whose edges were drawn apart by the contracting
muscles. The axial cells not only maintained a continuous
column, but bulged out into the gaping wounds in the wall, under
considerable internal pressure. Since a stem mutilated on one
side may right itself when it is much contracted, and in this con-
dition the muscles as a whole on the wounded side are weaker than
4o8 Harry Beal Torrey.
those on the other side, it seems highly probable that the orientation
is not the result of muscular activity, but must be due to changes in
relative volume of the vacuolated endoderm cells on opposite sides
of the stem. And since these cells are exceedingly large, v^ith
excessively thin walls and almost no protoplasm, the changes in
volume appear to be due to changes in the turgidity of the cells.
This conclusion is borne out by the facts that the stem may not
only shorten w^ithout increasing its diameter, but may lengthen
yvhWe. actually increasing its diameter, results possible only through
a variation in the turgidity of the axial cells. A complete demon-
stration that muscles do not take part in the geotropic response
is lacking, because in spite of the numerous transverse cuts made
in the stem, the latter was still able to shorten (thickening at the
same time), showing that the muscles were not rendered entirely
impotent. But the slowness of the response and its occurrence
while the wounds gaped and the muscles on the upper side of
the stem were manifestly weaker than those on the lower side,
strongly support the view that they were not concerned in the
result. I think we may say that the muscles produce the
movements of the tentacles, proboscis, and all save the geotropic
movements of the stem, including shortening and possibly length-
ening (by means of the circular muscles) while the axial cells
cause the geotropic orientation as well as lengthening of the
stem.^
If organic growth is increase in volume,^ then the changes in
turgidity which affect the orientation and length of the stem must
be reckoned among growth processes, and as such they will be
found to differ in no fundamental respect from those growth pro-
cesses in plants and in all probability the fixed hydroids also which
accomplish the orientation of these organisms with reference to
gravity. This statement requires some comment.
A comparison of the phenomena of geotropism in the stems of
plant seedlings and Corymorpha brings out points both of resem-
blance and difference. The cells reacting to the geotropic stimu-
^The skeletal function of the axial cells, correlated with the alienee of a sup-
porting perisarc, will be considered in a subsequent paper.
^Davenport, '97.
Biological Studies on Corymorpha. 409
lus are in both cases strikingly similar in structure, being large,
with a relatively small amount of protoplasm and large vacuole,
and they bring about the bending of the stem by changing their
volume. In the seedling, the cells on all sides of the stem increase
in volume, but in those on the lower side the increase is greater
than in the cells on the upper side. In Corymorpha, a similar
differential between the upper and lower cells is established,
though this may not involve a change in volume of cells on both
sides of the stem. Just how it is established cannot be deter-
mined at present, for reasons which bring about an interesting but
not fundamental difference between the responses in plants and
Corymorpha. The increase in volume of the plant cells — their
growth — is permanent, because it includes growth of skeletal cell
walls which prevent the return of the cells to their previous size,
and it can be readily measured. The cells of Corymorpha have
no such walls, and can change their size without difficulty; their
growth is temporary and cannot be measured in the same way,
because the stem is liable to frequent non-geotropic changes in
length. The presence or absence of skeletal cell walls determines
whether the growth is to be permanent or transitory. This con-
sideration leads to the discussion of the geotropisms of such fixed
hydroids as the sertularians, some of which are known to be
geotropic, and all of which are provided with stout perisarcal
skeletons.
The experiments of Driesch ('92) on species of Sertularella,
brought out the facts that the geotropic bending of the stem is not
general, but localized in the growing region at the end of the stem,
and that the growth which accompanies the bending is permanent.
My observations ('02) of Sertularia furcata and .S". argentea in
nature are in harmony with these results. "The San Francisco
colonies [of S. furcata] were growing on erect stalks of Phyllospa-
dix. The stems are short and project from all sides of the eel-
grass. Each stem leaves the eel-grass at an angle of about thirty
degrees, then bends quickly away so that for the most part it
makes an angle of seventy degrees with the stalk. The hydro-
thecae of the first, and often of the second pair as well, are not in
contact. Those of succeeding distal pairs are not only in contact
410 Harry Beal Torrey.
for half their length but tend much more strongly toward the upper
side of the stem than do the proximal hydrothecae. This would
seem to be an instance of the effect of gravity upon the direction of
hydranth buds. The farther the stems diverge from the vertical,
the more closely do the hydrothecae of each pair crowd each other
on the upper side of the stem" (p. 66). "The habit of the San
Francisco colonies of iS". argentea seems to be controlled in an
interesting fashion by gravity. The branches are borne on all
sides of the stems, which were fastened by their bases to the per-
pendicular side of a shore boulder. Each stem had curved up-
ward, so that while the basal portion was nearly horizontal, the
terminal fourth or fifth was approximately vertical. In this
terminal vertical portion the branches and the hydrothecae on
them were arranged symmetrically with respect to the axis of the
colony; and in this region the axis of the colony and the lines of
force of gravity were parallel. At the base, where they were not
parallel, branches and hydrothecae were oriented with respect to
the force of gravity alone. Both hydranth and branch buds, as
well as the stem, thus appear to be more or less negatively geo-
tropic, the hydranths always being borne on the upper sides of
the branches; the latter grow away from the center of the earth
but never become parallel with the main stem" (p. 68).
It is not difl&cult to explain why the geotropic bending is at the
end of the stem. It is only at the growing tips of stems and
stolons that the perisarc is dissolved. Elsewhere the cells of the
coenosarc may contribute perisarc, but are ordinarily unable to
dissolve it. At the tip, then, either by means of muscular activity
{cf. the free Cerianthus) or, more probably, by growth processes
similar to those taking place in Corymorpha — it is not yet de-
termined which — the stem assumes an orientation which is
temporary at first, becoming permanent only when the harden-
ing of the perisarc about it prevents further bending.
Enough has been said to show that no fundamental distinctions
can be made between the geotropism of permanently fixed, tem-
porarily fixed and permanently free organisms, such as plants,
hydroids, anemones, protista. Any theory, then, which seeks to
offer a thoroughly satisfactory interpretation of geotropism in one
Biological Studies on Corymorpha.
411
of these organisms, must be similarly satisfactory for all. The
following characteristics of the geotropism of Corymorpha appear
to find no adequate explanation in existing theories.
Suppose a Corymorpha stem (Fig. 2) to be cut at x into two
segments, A and B. Now, at whatever point A be supported
above the proximal end, the latter will seek the center of the earth.
On the other hand, at whatever point B be supported, its distal
Fig. 2.
Diagram of stem, to illustrate geotropism.
end (cut at the same level as the proximal end of A) is strongly
negatively geotropic. If the cut were made at x^, the shaded por-
tion which, as a part of B would have been negatively geotropic,
would now appear to be positively geotropic. I say "appear to
be," for while the negative response is due to a change in the
turgidity of the axial cells, the positive response may not be truly
geotropic, but may mark an unresponsive period in the axial
412
Harry Beal Torrey.
cells, induced by a suspension of the stem from its distal end, in
which case the stem would come to a vertical position of its own
weight. This statement may be nearer the facts and yet not lead
us appreciably nearer an explanation. Why should the same cell
react in one way when the stem is attached proximally, and another
way or not at all when the stem is attached distally ?
In seeking an answer for this question, it should be observed
that gravity may conceivably stimulate the stem in several ways:
(i) through the difference in the mechanical stresses on the two
Fig. 3.
Diagram to illustrate geotropism. Cc, compression, greater and less; Tt, tension,
greater and less.
sides of the stem, (2) through the difference in the resistance en-
countered by the organism according as it goes upward (frictions-
weight) or downward (friction— weight) — Davenport's theory as
applied to free organisms, (3) by redistributing the contents of the
axial cells so that in any but a vertical position of the stem the cells
would be in a state of unstable equilibrium with respect to the
geotropic function.
With reference to the first hypothesis, it may seem that the
difference in the response of the same cell may depend upon cer-
Biological Studies on Corymorpha. 413
tain differences in the stresses when the stem is suspended from
one end or the other (Fig. 3). If we assume the axis to be rigid
to some extent, then, when the stem is anchored proximally
(a and b) its weight may tend to compress its elements in the direc-
tion of its axis when it is vertical (a, C); when it is not vertical (b),
there may be added to this compression a tension of the elements
on the upper side of the stem (Ct), and an increased compression
of those on the lower side (Cc). When the stem is hung from its
distal end (c and d), a tension may take the place of the com-
pression (d, T), and if the stem be not vertical (r), the tension may
be increased on its upper side {Tt), a degree of compression added
to the tension on its lower side (Tc). There may be, then, a
degree of tension on the upper side and a degree of compression on
the lower side of each stem; in which case the differences would be
differences of degree only. That differences of degree do not
modify the reactions of the axial cells is evident when it is remem-
bered that a stem hung vertically from its proximal end begins to
right itself when, according to the hypothesis, the tension factor is
strong on both sides of the stem, and continues in the same direc-
tion after it has passed the horizontal, /. e., after the tension has
ceased on the lower side and become much reduced on the upper
side.
The inadequacy of the first hypothesis may be shown further,
in the discussion of the second. This view was formulated to
explain the orientation of free organisms only. It assumes that
negatively geotropic organisms tend to move in the direction of
greater resistance; being heavier than water, they would meet with
greater resistance in going upward than in going downward.
"Another stimulus," says Davenport, "which is probably asso-
ciated with this, depends upon the fact that an unsymmetrical
body, heavier than water, tends to fall with its larger end down."
That this view cannot explain the phenomena of orientation in
Corymorpha will be clear from the following considerations:
First, it presupposes locomotion, while locomotion is not con-
cerned in the orientation of Corymorpha. Second, if the stem
moves in the direction of greater resistance, a stem hung from its
distal end ought to move in the same direction as a stem hung from
414
Harry Ben I Torrey.
its proximal end and parallel with the first (Fig. 4). Instead of
moving in the same direction, however, they move away from
each other as indicated by the arrows, A in the direction of less,
B o^ greater, resistance.
It is clear that the factor of external resistance does not govern
such behavior, nor does the mechanical factor of tension, as has
been shown above. It is equally difficult to explain the geo-
tropic reactions of Corymorpha on obviously mechanical grounds
by means of the third hypothesis; for the response of a given cell
may be different, according as the stem is hung by its proximal
or distal end, though the contents of the cell be distributed by
Fig. 4.
Diagram to illustrate geotropism.
gravity in the same way in the two cases. This hostility to the
familiar mechanical explanations which appear to account for
the facts in other geotropic organisms urges upon me the desir-
ability of repeating and extending my experiments as soon as oppor-
tunity is afforded. It is certain, however, that the stem as a whole
orients itself negatively to gravity, without regard to' the point at
which it is supported. And the reactions of the axial cells are
unquestionably associated with the polarity of the stem. A
change in the polarity of a region of the latter is always accom-
panied by a change in the reactions of the axial cells in this region.
Biological Studies on Corymorpha. 415
For instance, the cells in .v-^i (Fig. 2), if forming part of the piece
A, would as a whole be positively geotropic so long as the aboral
end of ^ tended to develop a hold-fast (thus preserving the original
polarity of the piece). If, however, this end should develop a
hydranth, the behavior of these cells would be reversed; they
would exhibit, as a whole, negative geotropism.
There are, then, two manifestations of polarity of apparently
different sorts: first, polarity expressed through a special mechan-
ism involving a single tissue, in terms of osmotic pressure and
consequent movements of geotropic orientation ; second, polarity ex-
pressed through a mechanism involving many tissues, in terms of
regenerative development and differentiation. However different
these may seem, they are undoubtedly referable to the same funda-
mental causes — the causes of polarity in general, which involve
internal factors at present objects of speculation only. Yet it
may be possible to determine these internal factors more easily
by means of the facts of what may be called functional polarity
than by the relatively complex morphological phenomena of devel-
opment and differentiation. The simpler mechanism and the
simpler effects of polarity as manifested in the axial cells, are
bound to bring a true explanation of polarity nearer our compre-
hension, although it may still be unattainable. How an organic
membrane or its contents may change in order to produce a change
in osmotic pressure, while an enormously difficult problem, is yet
more hopeful of solution than the problem of how several tissues
simultaneously differentiate in different directions to produce a
complicated regeneration.
c. Locomotio72; Amoeboid Cells.
We come now to a third type of movement, with a new cause.
The ectoderm cells of the proximal end of the stem are capable
of amoeboid movements, by the aid of which the hydroid may
slowly change its location. In this regard Corymorpha closely
resembles Hydra. On a horizontal surface, whatever locomotion
there is takes place in any direction, with the one qualification
that the stem moves always out of its perisarcal investment,
which it leaves behind.
41 6 Harry Beal Torrey.
The rate of locomotion is slow. Half an inch in twenty-four
hours is a maximum rate. On a vertical surface the movement is
always directly upward. Gravity evidently determines the direction.
The value of locomotion of this sort, and especially its negatively
geotropic character, would seem to lie in providing a means
whereby the hydroid may keep above the surface of the shifting
sands.
The filaments of the hold-fast are also furnished with amoeboid
cells by which they are enabled to move out amongst the sand
grains to which they cling and anchor the stem {cf. amoeboid
movements of the tips of stolons of Campanularian hydroids).
One set of observations gave a rate of nine microns per minute at
the tip and five microns per minute halfway to the base of a fila-
ment several millimeters long. The free end is swollen and club-
shaped, with well developed ectoderm which not only provides
amoeboid cells but gland cells, which secrete the perisarc in which
the final strength of the filament as an anchor lies. The ecto-
derm of the remainder of each filament is attenuated almost to the
limit of visibility; the whole filament appears to be upon the
stretch, pulled out by the creeping club-shaped end. The endo-
derm of the filament is composed of a single column of cells such
as is characteristic of the endoderm of the tentacles of Campanu-
larian hydroids. These cells, under the tension, may become
much longer than broad, and retain these proportions whether the
filament is attached or free. A "setting" process seems to have
followed the stretching here, effecting the permanence of the
attenuation without cell division.
The direction of locomotion of the filaments is always outwardy
but appears to be otherwise indeterminate. Arising below the
perisarc on the peripheral canals of the stem as solid outgrowths
with a deflection toward the proximal end, they creep along the
stem for a short distance, closely in contact after the manner of
stolons, and then push outward, secreting perisarc as they go. If
the stem is hung freely in the water, the filaments extend in all
directions. If it is in contact with the substratum, however, they
creep along the latter as soon as they come in contact with it.
If the substratum is sand, a filament pushes its way between the
Biological Studies on Coryniorpha. 417
grains, as a plant root pushes its way through the earth. In no
case, however, does it appear to respond to the stimulus of gravity,
or any other stimulus, except that of contact. Resistance seems
to incite movements which overcome it. This is probably the
reason why the filaments leave the easy path between stem and
perisarc to push out against the resisting wall of the latter.
d. Circulation; Cilia.
A fourth type of motion is found in the currents set up in the
cavities of the digestive tract by means of cilia. The cilia are
borne on the lining cells of the proboscis of the hydroid and the
epithelium bounding the peripheral canals on their outer side.
They are present throughout the peduncles bearing the medusae,
and the manubria of the latter.
There are variations in the currents, particularly in the peri-
pheral canals. At times there may be no current at all. At
others the current may be setting very rapidly in the same direc-
tion in all the canals visible. Abrupt reversals occur under these
conditions, which can hardly be explained by ciliary action, but
are rather the result of expansions or contractions of the proboscis
and stem, which produce changes of pressure in the canals.
IV. THE YOUNG HYDROID.
The eggs are laid by medusae which are never set free from
the hydroid. They are small but heavy with yolk and fall directly
to the bottom in quiet water, adhering by their delicate coats to
the first object they touch.
As soon as the egg is attached, its free life is practically over.
The embryo is never ciliated^ and has no free-swimming phase in
its existence. It is capable of very slow creeping movem'ents,
however, by means of which it often comes in contact with other
embryos and forms with them temporary associations of as many
as six, ten, twelve individuals. Often it will travel many times its
own length, leaving behind a narrow collapsed tube of perisarc
iC/. Hypolytus peregrinus, Murbach (Q. J. M. S., XLII, 1899, p. 341), which it
resembles in this and other respects.
4i8
Harry Beal Torrey.
which it has secreted and which is continuous with the egg case.
As we have seen, the hydroid never loses its power of locomotion,
even after the development of the filaments of the hold-fast.
As the embryo leaves its egg case, it elongates, and an anterior
(oral) end can be distinguished from the narrower posterior
(aboral, proximal) end. The anterior end soon elevates itself,
and the embryo now touches the substratum by one side of the
aboral region only.
For about thirty hours, or up to the time when the hydranth is
beginning to form, the embryo is completely covered by an
Fig. 5.
Young Corymorpha.
extremely delicate layer of perisarc. From this time the perisarc
is frequently limited entirely to the stem. Before the formation
of the hydranth, the perisarc covering the anterior end of the
embryo, and secreted by glandular cells of the ectoderm, is not
permanent, being dissolved as the stem progresses, probably by
the secretion of other cells in this region. As the hydranth begins
to develop, its ectoderm ceases to manufacture perisarc, which
henceforth is deposited by cells beginning at the aboral limit of
Biological Studies on Corymorpha. 419
the hydranth. The perisarc is hardly sturdy enough at any time
to afford any support to the stem. Its adhesive character, how-
ever, serves to attach a portion of the latter to the substratum, over
which the coenosarc creeps.
Amoeboid ectoderm cells are responsible for the locomotion of
the young Corymorpha (Fig. 5) as of the adult, though they are
not confined to the proximal end of the stem. Often the latter
clings for half its length and may perform looping movements,
much less pronounced, however, than those of Hydra} The
direction of locomotion is also determined by the same factors
which regulate it in the adult. On a horizontal surface the direc-
tion is indeterminate, though the stem always moves out of its
investment. On an oblique surface it tends to move upward by
the nearest route. Young hydroids are often found adhering to
the stem of an adult, the relation of the axis of the attached por-
tion of each to the adult axis varying with the inclination of the
latter. If it is vertical, they are parallel with it, vertical also; and
the rest of the young stem will be nearly vertical, but not quite so
since the distal portion of the stem seems to shun any contact
(negative thigmotropism). If the orientation of the adult is
altered, the young hydroid will gradually take up a new position
in which the most distal point of attachment will be the greatest
possible distance above its proximal end.
Not only, therefore, is the larva negatively geotropic with regard
to orientation, but this has a directive effect upon locomotion.
It is probably due to the effect of the stimulus of gravity on
the endoderm cells which line the single cavity and from which
the axial cells of the adult are derived. These cells do not con-
tain such enormous vacuoles as those in the axial endoderm, and
are ciliated. In these respects they resemble the parietal endo-
derm cells of the peripheral canals of the adult, which are their
descendants also.
With reference to the amoeboid cells which produce locomotion
in Corymorpha, it may be recalled that the ectoderm of hydroids
is not uncommonly amoeboid. To cite but a single instance, not
'C/. MarshaU. Zeitschr. f. w. Zool., XXXVII, p. 664.
420 Harry Beat Torrey.
only is the cauline coenosarc in Campanularian hydroids, e. g.,
Obelia, fastened to the perisarcal tube here and there by multi-
cellular amoeboid processes of the ectoderm, but the anterior ends
of growing stolons exhibit amoeboid changes of form which ac-
count for their creeping movements and produce the tension often
manifested in the coenosarc which is fixed farther back on the
stem. The coenosarc is literally dragged out of the perisarc.
A similar tension has already been noted in the filaments of the
hold-fast, due to a similar cause. And it is probable that the
proximal end of the larva may be dragged along at times after the
more distal attached portion.
The active muscular movements discussed at length for the
adult need not be considered here, as the young hydroid appears
to respond similarly in all respects, with the one exception that the
reaction times are somewhat greater.
The absence of a free-swimming larval form seems to account
for the tendency of Corymorpha palma to dwell in communities,
as previously mentioned. The power of locomotion is too slight
to have any effect on the distribution of individuals, which is
accomplished by tidal currents and the shifting of surface sands.
Occasionally an individual may be washed away from its anchor-
age, and begin a new community in a new locality.
SUMMARY.
Corymorpha is unusually active for a hydroid. It is every-
where sensitive to mechanical stimuli, irritating chemicals and
abrupt changes in temperature, nowhere to "odorous" substances.
The prehensile mechanism is composed of proximal tentacles,
which move toward the mouth in response to all eff'ective stimuli;
distal tentacles, which move away from the mouth in their initial
response to stimuli; and proboscis, which may move toward the
point stimulated. These movements, as well as shortening and
possibly lengthening the stem, are performed by muscles.
The stem of the adult responds to the stimulus of gravity, by
means of a change in the turgidity of the vacuolated axial cells.
The response of these cells varies according as the stem is attached
Biological Studies on Corymorpha. 421
proximally or distally, and according as it is heteromorphic or not.
The polarity of the stem is expressed, not only by the regenerative
development but by the changes in the axial cells.
Locomotion is accomplished by amoeboid cells located at the
proximal end in the adult, more generally distributed in the larva,
and covering the club-shaped ends of the filaments of the
hold-fast.
Cilia are present on the epithelial cells lining the hydranth
cavity and peripheral canals. Supplemented by contractions and
expansions of the hydranth cavity, they provide for the circula-
tory currents through the body.
Eggs are laid both in summer and winter, usually during the
morning hours. They have adhesive coats. The planulae are
never ciliated, and their locomotion is limited to very slow creep-
ing movements. The larvae are geotropic.
BIBLIOGRAPHY.
Agassiz, a., 1862. — Contributions to the Natural History of the United States.
Vol. IV, p. 276.
Allman, G. J., 1863. — Notes on the Hydroida. Structure of Corymorpha nutans.
Ann. and Mag. Nat. Hist., 1863, p. 1.
1871. — A Monograph of the Gymnoblastic or Tubularian Hydroids.
London.
CzAPEK, Fr., 1898. — Weitere Beitrage zur Kenntniss der geotropischen Reizbewe-
gungen. Jahrb. wiss, Bot., Vol. XXII, p. 175.
Davenport, C. B., 1897-9. — Experimental Morphology. Parts I and II. New-
York.
Davenport and Perkins, 1897. — A Contribution to the Study of Geotaxis in the
Higher Animals. Jour. Physiol., Vol. XXI, p. 22.
Driesch, H., 1892. — Kritische Erorterungen neuerer Beitrage zur theoretischen
Morphologie. II. Zur Heteromorphose der Hydroidpolypen. Biol.
Centr., Vol. XII, No. 18, p. 545.
EiMER, Th., 1878. — Die Medusen, etc. Tubingen.
Haberlandt, G., 1901. — Ueber Reizleitung im Pflanzenreich. Biol. Centr., Vol.
XI, p. 369.
Holmes, S. J., 1904.— The Problem of Form Regulation. Arch. Entmckl., Bd.
XVII, p. 265.
Jennings, H. S., 1902. — Studies on Reactions to Stimuli in Unicellular Organisms.
IX. On the Behavior of Fixed Infusoria (Stentor and Vorticella)
with Special Reference to the Modifiability of Protozoan Reactions
Am. Jour. Physiol., Vol. VIII, p. 23.
1904. — Contributions to the Study of the Behavior of Lower Organisms.
Carnegie Institution of Washington. Publ. No. 16.
422 Harry Beal Torrey.
LoEB, J., 1891. — Ueber Geotropismus bei Thieren. Arch. f. ges. Physiol., Bd. XLIX.
1891a. — Untersuchungen zur physiologischen Moiphologie der Thiere. I.
Ueber Heteromorphose. Wiirzburg.
1892. — Untersuchungen zur physiologischen Morphologic der Thiere. II.
Organbildung und Wachsthum. Wiirzburg.
1894. — On Some Facts and Principles of Physiological Morphology. Woods
Hole Biol. Lect.
1895.- — Zur Physiologic und Psychologic der Actinien. Arch. f. ges.
Physiol., Bd. LIX, p. 415.
1902. — Comparative Physiology of the Brain and Comparative Psychology.
New York.
'Mast, S. O., 1903. — Reactions to Temperature Changes in Spirillum, Hydra, and
Fresh-Water Planarians. Am. Jour. Physiol., Vol. X, p. 165.
May, a. J., 1903. — A Contribution to the Morphology and Development of Cory-
morpha pendula Ag. Am. Nat., Vol. XXXVII, No. 441.
Morgan, T. H., 1901. — Regeneration. New York.
1901a. — Regeneration in Antennularia. Biol. Bull., Vol. II, No. 6.
MuRBACH, L., 1899. — Hydroids from Woods Hole, Mass., etc. Q. J. M. S., n. s.,
Vol. XLII, p. 341.
Nagel, W., 1893. — Versuche zur Sinnesphysiologie von Beroe ovata und Carmarina
hastata. Arch. f. ges. Physiol., Bd. LIV.
1894. — Vergleichend physiologische und anatomische Untersuchungen
ueber den Geruchs-und Geschmackssinn und ihre Organe. Bibl.
Zool., Heft 18.
1894a. — Experimentelle sinnesphysiologische Untersuchungen an CcBlen-
teraten. Arch. f. ges. Physiol., Bd. LVII, p. 495.
Nemec, B., 1901. — Ueber die Art der Wahrnehmung der Schwerkraftreizes bei den
Pfianzen. Ber. deut. Bot. Ges., Bd. XVIII, p. 241.
Parker, G. H., 1896. — The Reactions of Metridium to Food and Other Substances.
Bull. Mus. Comp. Zool., Vol. XXIX, p. 107.
Platt, J., 1899. — On the Specific Gravity of Spirostomum, Paramoecium and the
Tadpole in Relation to the Problem of Geo taxis. Am. Nat., Vol.
XXXIII, p. 31.
Romanes, G. J., 1876. — Preliminary Observations on the Locomotor System of
Medusa;. Phil. Trans. Roy. Soc. Lond., Vol. CLXVI.
1877. — Further Observations on the Locomotor System of Medusse. Ihid.,
Vol. CLXVII.
Sars, M., 1853. — On the Nurse Genus Corymorpha and its Species, together with the
Medusse Produced from Them. Ann. and Mag. Nat. Hist., Vol.
VIII, p. 353.
Torrey, H. B., 1902. — The Hydroida of the Pacific Coast of North America. Un.
of Cal. Publ. Zool., Vol. I, p. 1.
1902a.— American Naturalist, Vol. XXXVI, p. 987.
1904. — The Habits and Reactions of Sagartia davisi. Biol. BuU., Vol. VI,
No. 5, p. 203.
True, R. H., 1895. — On the Influence of Sudden Changes of Turgor and of Tempera-
ture on Growth. Ann. of Bot., Vol. IX, p. 365.
Verworn, M., 1899. — General Physiology. Trans, by F. S. Lee. New York.
STUDIES ON THE LIFE HISTORY OF PROTOZOA.
IV. DEATH OF THE A SERIES.
CONCLUSIONS.
BY
GARY N. CALKINS.
With 3 Plates and 3 Figures in the Text.
Experiments on the life-history of Paramcecium caudatum have
now been carried on continuously for 29 months. Two series,
designated as the "A series" and the "B series," were started on
the first of February, 1901, with individuals from different
sources. The B series died out in May, 1902, in the 570th genera-
tion; the A series on December 19, 1902, in the 742d genera-
tion. A third series — "C" was started in June, 1902, with an
individual from Cambridge, Mass., and died out in June, 1903, in
the 379th generation. The progress of the first two series has
been recorded from time to time,^ and in the present paper I
wish to give the history of the last cycle of the A series and to
consider the results in relation to some general biological prob-
lems and theories.
I. THE JUNE AND DECEMBER (19O2) PERIODS OF DEPRESSION.
As described in the earlier Studies (I and III) the general
vitality of the two series, A and B, as expressed by the daily
division rate, underwent periodic cycles of vigor and depression.
^(1) Studies on the Life History of Protozoa. I. The Life Cycle of Paramcecium
caudatum. Archiv. f. Entwk. XV, 1, 1902.
(2) Studies, etc. II. The Effect of Stimuli on the Life Cycle of Paramcecium
caudatum. (With C. C. Lieb). Arch. f. Protistenkunde. I, 1, 1902.
(3) Studies, etc. III. The 620th Generation of Param. caud. Biol. Bull. Ill, 5,
1902.
424 Gary N. Calkins.
The early curves appeared to indicate a periodicity of three-
month intervals, and this v^as taken to be the time of the usual
life cycle in culture of Paramcecium candatum; this conclusion w^as
based partly upon my ow^n results and partly on those of Jou-
kowsky and of Simpson, both of whom found that cultures of this
infusorian died out after three months of treatment. It w^as
found, however, and it may be seen from the now completed
curve of the A series (see Diagram I) that trimonthly periods of
depression were not fatal and that recovery occurred without
purposeful stimulation. Thus in the first apparent depression
(May, 1901,) the recovery was thought to be due to the stimula-
tion by jolting on a railroad trip of six hours; another in March,
1902, was considered due to a slight rise in temperature. These
periods of depression differ markedly from those of August and
December, 1901, and of June, 1902, when the individuals con-
tinued to die at a high rate, notwithstanding repeated jolting
experiments, increase in temperature, and the like, and the race
was saved only by change to a special diet after numerous
attempts and failures with foods of different kinds. The well-
marked cycles, therefore, with periods of depression which de-
manded stimulation of a decided character, were approximately
of six months' duration, while intermediate cycles of less impor-
tance were about three months long. The first of the six-month
cycles ran from February i, i90i,to August i, 1901, (see Diagram
I); the second from August 15 to January i, 1902; the third from
January i to July i, and the last from July to December 19, 1902.
During the first three cycles the number of generations was nearly
the same (200, 198 and 193, respectively), the last, on the other
hand, was much less, the individuals dividing only 126 times.
The stimulation which resulted in the renewal of vitality after
the periods of depression in August and December, 1901, was due
to the change from hay infusion diet to beef extract for a
limited period (see Studies I and III). The same change failed to
work in the July, 1902, period of depression, and after the race
had become reduced to only six individuals, a successful sub-
stitute for the beef extract was found in the extracts of pancreas
and brain (see Studies III). Recovery, however, was not so
Studies on the Life History of Protozoa. 425
successful as in the previous periods and the organisms were much
less vigorous than at similar periods in previous recoveries. The
division rate, furthermore, slowly fell from the relatively high
point in August, and gradually decreased during the fall months
until the A series died on the 19th of December. The B series
had succumbed in the 570th generation, in June, before the right
stimulus was found. Except for the slowness of divisions the
organisms appeared perfectly healthy during the summer and fall
of 1902, although microscopical study of preparations made during
this period showed characteristic changes in the protoplasmic
structure (see Figs. 18 to 21). The organisms were plump and
moved freely about the slide, responding with customary vigor to
stimuli of diflperent kinds. Every precaution was taken during
this period to invigorate the race and every experiment that my
ingenuity could devise wks executed; some appeared to give a
temporary improvement but none was permanent, and the last
individual of the A series finally died after 23 months of continued
daily observation, without, however, any morphological evidence
of general senility. (See Diagram I.)
II. UNSUCCESSFUL ATTEMPTS TO REJUVENATE THE A SERIES
IN THE FINAL PERIOD OF DEPRESSION.
Artificial rejuvenation of the A series was successfully accom-
plished three times. The experiments and results have been
described in other places, and it will be sufficient here to merely
point out that after considerable experiment, beef extract was
successful in the first two cases and pancreas and brain extract in
the third, the result being due, probably, to the change in salt
contents of the medium. The approaching end of the series was
indicated some time in advance by the reduced division rate
during the fall of 1902, and efforts were continuously made to
rejuvenate them during this period. For these experiments all of
the stock of the regular series was maintained, and the number of
lines under observation frequently ran up to twelve or more.
The results of all experiments were tabulated and the effects of
the stimuli used were noted for comparison with the regular
series. The general result may be seen upon the diagram which
426
Gary N. Calkins.
i'illfllflltB
n '■'<
% r
--', 'V-" "
'"■; " 1 "
i
..'. .:::^::i..
kiMf-i---
.._!.'.. _:l _..:
....- r
' [-:■
M ___i....
is ^— ].:._..
i ^....__,..
-r-
""1 :::}EB
[
1
a:
!:::.:i:.::::...
CI -'
O ^
a> o
C
-5= r^
_ o
Oi ^
1^ t-i
e3
S .2
o be
a a
:3
Hi
T-i
n>
13
-M
e
C4-I
<J
<u
<U
id
ID
^ '2
S -S
CD
05 03 •-'
B
-(J
4-1
c,
o
Ol
^
Pi
>
a1
0)
ri
-iJ
>
-o
<T)
a
0)
51
(D
02
o
,a
rt
rt
tH
-f-*
O .5
tin H
la
Studies on the Life History of Protozoa. 427
hows that, despite all efforts to stimulate, the race rapidly
weakened and ultimately died.
I. Experiments with Extracts.
a. Beef Extract. During the last autumnal period, beef extract
was used at different times in the same way that it had been used
successfully during previous periods of depression, and for the same
length of time, twenty-four hours. The organisms were immersed
in the fluid full strength in the majority of cases, but experiments
with the half strength were also made. The failure of the beef
treatment in June, 1902, has already been described (Studies III)
and I shall consider in this place only the experiments subsequent
to that time. In general it may be stated that beginning with*
the treatment in May, the effect of the beef extract was nil. On
the 19th of June A3 and A4 were immersed as usual and for the
same length of time. Both of them died before the 27th. Again
on the 22d Ai and A2 were treated, and both of these died on
the 26th. Similar results were obtained in all later experiments,
as shown by the following resume:
Both died before the 15th.
Ai and A2 died on 25th. A5 on
the 26th.
Both died before the 19th.
Both died before the 13th.
Died on the 27th.
Died on the 29th.
Died December 2.
The short time in beef may not have been long enough to make
the change beneficial; with this in mind I kept the last few
individuals in for three days (Nov. 26, 27, 28). Not one of them
divided more than once and all died within a week. Beef extract,
therefore, had lost its potency as a rejuvenating medium.
The effect of beef extract upon the body structures was to
increase the number of gastric vacuoles; while, in some cases, the
micronuclei were caused to divide (Figs. 5 and 6). Even in May,
1902, there was an indication of the endoplasmic concentration
July. 6.
A3 and
A4 treated
July 24.
Ai, A2,
A5 "
Aug. 2.
Ai and A2 "
Aug. 2.
A5 and A6 "
Nov. 26.
A5
((
Nov. 26.
A;
i<
Nov. 26.
An
<(
Nov. 28.
A6,A8,
Aq and Ai:
428 Gary N. Calkins.
which accompanied depression at this period. The dense con-
dition of the protoplasm is better shown in Fig. 6 which represents
an individual twenty-four hours after transference from beef
extract into hay infusion. It may be noted here that, at this
period, the beef extract failed to reduce the dense endoplasmic
condition to one of tenuity which seems to be the normal
condition.
b. Extract of Pancreas. Extract of pancreas was made in the
same way as the beef extract. A fresh sheep's pancreas was Gut
in small pieces and brought to boiling point in water. After
filtering and cooling, the Paramoecia were placed in it and left for
24 hours, as in the beef. At first it proved a good substitute for
the beef and the organisms appeared to thrive on it; but later, in
November and December, it was as useless as the beef. The
following records show this fact:
June 27. A3 treated. Divided twice in 24 hours. Forms
the regular series from this time.
June 29. Ai and A2 " Divided a few times. Died out
on July 14.
July 16. A4 " Died the next day without divi-
sion.
July 17. A6 " Divided twice the next day; lived.
July 18. Ai,A2 and A4 " All died the next day.
July 20. Ai,A2andA4" Lived. Given mutton broth on
23d. Died on 24th.
Aug. 20. A2 " Died on 28th.
Aug. 20. A8 " Died on the 2ist.
Dec. 8. A5, A6, A7 and A8 treated. No divisions. All died out
on 19th before or after treatment with various other substances.
At the period in June when recovery was effected by using the
extract of pancreas, the organisms of both series were in the con-
dition represented by Fig. 7. The endoplasm was densely granular
and homogeneous, and had a curiously "stuffed" appearance.
This condition was relieved by using extract of pancreas, whereas
beef extract, made with the same water and in the same way, was
ineffectual. Figs. 8, 9 and 10 show the general course of the
Studies on the Life History of Protozoa. 429
action of the pancreas extract. Fig. 8 represents an individual
twenty-four hours after treatment, i. e., after change from pancreas
extract into hav infusion.^ The characteristic dense structure is
distinctly shown, but in the center there is unmistakable evidence
of the normal condition. Figs. 9 and 10 represent two indi-
viduals forty-eight hours after treatment with the pancreas extract.
In the former, the characteristic dense structure is still visible at
the two ends, but the center is clearing. In the latter, new gastric
vacuoles have appeared in the endoplasm, the animal being well
on toward recovery when killed.
c. Extract of Sheep's Brain. This was made in the same way
as the other meat extracts, and the animals were similarly treated
with it. It was not efficient as a permanent stimulant, and was
discarded in subsequent treatment.
d. Extract of Mutton. '' Mutton Broth.'' This extract was also
tried in the summer (July 20 and 23), but in no case was it
successful, the organisms invariably dying within 24 hours.
e. Lecithin. A trace of pure lecithin was put into the regular
hay infusion during the week of August 20. The organisms were
apparently not injured by the change, but did not live more than
48 hours after the treatment.
/. Pineapple Extract. With the view of ascertaining if some
of the vegetable ferments might not prove beneficial, I tried
extract (juice) of fresh pineapple, and of fresh apple. A4 was
put into dilute pineapple juice July 27. The reaction was well
marked, as shown by decided increase of movement and by three
divisions in the ensuing 48 hours. The experiment was repeated
the next day with a like result. It was repeated again August 3,
but was unsuccessful, the organisms dying two days after treat-
ment. The stimulation was temporary in all cases, and it should
be noted that the organisms were in a period of increasing vitality
when the first pineapple treatment was given (see Diagram I).
^ The hay infusion was made every day, the same amount of hay and water being
taken each time and raised to the boiUng point. This method was never varied
during the entire period of the cultures and the salt content of the water, as shown
by weekly analyses, did not vary beyond a very slight fraction of one part to one
hundred thousand.
430 Gary N. Calkins.
g. Apple Juice. A piece of fresh Porter apple was allowed to
lie for a few minutes in the hay infusion. In this case the result
was well marked, and a decided stimulus was noted. Again, on
Sept. 20, A5, A6, A7 and A8 were all put into one drop of apple
juice to 12 drops of hay infusion and left for thirty to forty-five
minutes. They were then transferred to clear hay infusion and
left. All divided the next day. The experiment was repeated on
the 2 1st with a like result. In some cases the organisms died
immediately, showing that the strength used was too great. When
properly diluted, however, apple seemed to give a satisfactory
temporary stimulus, although in no case did the stimulation last
for more than forty-eight hours. The same experiment tried in
October gave no results; the organisms died.
In addition to the above, various proprietary mixtures were
tried from time to time. Among these were phospho-albumin,
and nuclein-albumin; none gave satisfactory results.
2. Experiments with Acids and Salts.
In view of the successful results which have followed experiments
with ions in connection with egg development, it was thought that
perhaps dilute acids or salts would have a beneficial result in the
case of these weakened infusoria. Normal solutions were made in
each case and various strengths were tested from those that would
kill to those that only slightly stimulated. The organisms were left
in the fluids for only a short time (20 to 30 minutes) and were then
transferred to fresh hay infusion. Attention may be called here
to the fact that potassium phosphate when used in this way was
successful in restoring the vitality of weakened Paramcecium in
the preceding December cycle, the "rejuvenation" which resulted
was directly comparable with that eff^ected by the beef extract.
There was reason, therefore, to believe that the repeated use of
various salts would give satisfactory results in the last period of
weakness of the race. This expectation, however, was not real-
ized for none of the chemicals used in the fall and winter of 1902
was successful in this way; all were as futile as the beef and pan-
creas extract, as shown by the following experiments:
Studies on the Life History of Protozoa. 431
A. Potassium Salts.
a. K^HPO^. On the 8th of June, 1902, one individual from
the line of A2 was treated for 30 minutes with a solution of dibasic
phosphate of the strength of one drop of ^^^n to six drops of usual
hay infusion. The result was a marked increase in the rate of
division for a considerable period as compared with the control
series, as follows:
Average daily division-rate for 5-day periods, June 8-July 5.
Stimulated Aj. Control Series.
1st Period 8 8
2d *' 8 6
3d " 8 6
4th " i.o 2
5th " 1.6 o
On the 27th of June the above experimental series was
substituted in the regular culture series and the descendants
of these individuals formed the regular lines until the final
extinction, subject, of course, to the other experiments as stated
elsewhere.
A stronger solution (1-5) and for a longer period (i hour) was
used with A3 on June 26. The individual died in three hours.
On August 2 1 the same strength was used, but for only 25 minutes;
the individual died in four days without dividing. The general
effect of this salt was, therefore, favorable, with evidence that a
certain optimum strength is alone beneficial. The beneficial effects
upon the endoplasmic structure are shown in Figs, il and 13.
b. KH^PO^. Experiments with the monobasic salt were also
made and various proportions were used, but none was successful.
c. KCN. Various proportions of "0 of this salt were used, the
most successful being one drop of the solution to twelve drops of
hay infusion. This was not strong enough to kill the bacteria
which afford the only food for Paramcecium. Four individuals
were immersed October 29 in the mixture and left for 24 hours. At
the end of this time each had divided once, while none of the con-
trol series had divided. Of the eight individuals resulting from
this treatment, four were placed again in the KCN solution,
432 Gary N. Calkins.
(made fresh), while the other four were placed in hay infusion
without the salt. Of the former set, each individual divided once
in 48 hours, and of the latter set, one died, two divided once, and
one divided twice in the same period. The regular control series
did not divide at all during this time. Both sets were placed in
hay infusion on the fourth day and neither set continued to live,
all died before the sixth of November. Another set of four in-
dividuals were treated with the salt every day for the same period.
After the first 24 hours none had divided; after the second 24
hours each one had divided once. This was November 2. On
the 4th all had divided twice, on the 6th only one had divided
again, on the 7th another one had divided, on the 8th none had
divided again, on the 9th one died, while the rest did not divide,
on the loth two others died, and the remaining one was placed in
the usual hay infusion without the salt, having been treated
daily for ten days with it. It did not divide again until the i8th,
and finally died out on the 21st. Others, however, that had been
treated, and had been placed earlier in hay, continued to live and
supplied the regular lines of the experiment. The use of the
KCN therefore can be said to have been successful to a limited
extent, and, possibly, to have prevented an earlier extinction of
the race. The effect on the curve of the life cycle is shown by the
temporary rise during the last period in October and the first
period in November.
d. KOH. This was tried only once with four individuals on
the 28th of October. ^^^-^ i part to 4 was used for 30 minutes.
On the first of December two had died and one had divided once.
None divided again and all of the individuals experimented with
died before the fourth of November.
B. Sodium Salts.
a. Dibasic Sodium Phosphate. Three individuals, Al, A3
and A4 were placed for 30 minutes in ^S^^ NajHPO^. All
died without dividing by the 27th. The experiment was not
repeated.
b. Sodium Tartrate. July 10 two individuals were placed for
30 minutes in -^^ sodium tartrate, one drop to five of hay infusion.
Studies on the Life History of Protozoa. 433
They died in twenty-four hours. The experiment was repeated
on July 14, three drops to five of hay infusion being used. On
the 15th they did not divide, on the 17th they divided once, and
died on the i8th. Experiment not repeated.
c. Sodium Chloride. This salt was used on several occasions with
negative results as a rule. (See, however, table below.) In
September, 1902, when the race was comparatively vigorous, an
individual was treated for 30 minutes with f^ NaCL, one drop of
the salt to twelve of the hay infusion. At the end of 48 hours it
had divided once, but died within five days without further
division. The effect upon the protoplasmic structure was not
particularly noticeable (see Fig. 18).
The following table gives a comparative view of the efficiency
of different salts on the division rate for thirty days subsequent to
treatment. Several individuals of the A series were treated on
the 20th of March with potassium phosphate and the progeny of
one of these in the 78th generation were again treated in part on
May 6 with potassium phosphate, and in part, with potas-
sium chloride, magnesium chloride, sodium chloride and calcium
chloride with the strengths, and for the times indicated. The
following notes were made at the time of the treatment. "When
the individual was put into the potassium chloride it began at
once to swim backward with great rapidity, and continued this
for about five minutes. It then straightened out and appeared
perfectly normal in the solution. When returned to the hay in-
fusion at the end of the treatment, it went through the same con-
vulsions but soon became normal, perhaps slightly swollen and
transparent." Again: "When the individual (another individual
of course) was put into the magnesium chloride solution it was
hardly affected in any way, a very slight increase in movement
being noticed." Again: "Treatment with NaCl did not affect the
individual, it appears fat and happy in the hay-infusion." Again:
"Very much affected by the CaCl2 solution. One of the three
specimens died; the other two were distorted and badly shrunken,
this lasted for at least fifteen minutes after they had been trans-
ferred to the hay infusion."
434 Gary N. Calkins.
Average number of divisions per day after stimulation.
K.HPO^. KCI. MgCU. CaCl,. NaCl.
May 6-10 1.60 1. 00 1. 00 0.60 1.20
May 11-15 1. 00 1.20 0.60 1.20 1.20
May 16-21 1.50 1.50 1. 00 1.50 1.66
May 22-26 1.80 2.00 1.80 1.80 2.00
May 27-June I . . . . 1.25 1.03 1.25 0.75 1.50
June2-6 1.33 1.33 1.33 0.16 1.33
June 7-12 0.20 0.40 1.60 dead 0.40
dead dead living dead
All were normal solutions, diluted 25 times, one drop to twelve
drops of hay infusion and the treatment lasted for 30 minutes in
case of KCI, and for 25 minutes in each of the other solutions.
Definite conclusions cannot be drawn from one set of compari-
sons for it may have been pure accident that the magnesium
chloride specimens continued to live. The effect of MgClj upon
the protoplasmic structures is shown in Fig. 12.
Comparatively few experiments were made with acids. Hydro-
chloric and nitric acids were tried during the period of depression
in October, 1902, but the results were negative, the individuals
dying within twenty-four to forty-eight hours. An interesting
effect was produced by treatment with dilute phosphoric acid.
The dense endoplasm was broken up and with it the macro-
nucleus which, after the treatment, appeared as many small
fragments (see Fig. 16).
Of the other unsuccessful attempts to rejuvenate the race during
the last period of depression I will mention only those with gal-
vanic stimuli, with nitro-glycerine, and with dried and powdered
Paramoecium of an entirely foreign race.
3. Galvanic Stimuli.
A small cell was made and connected with two Mesco batteries.
Four individuals were treated on November 28, three different
times to the full current and for a period of one minute each time.
The usual reaction followed the treatment, migration to the nega-
tive pole, and when the current was reversed, migration from the
Studies on the Life History of Protozoa. 435
positive to the negative pole. At the end of the treatments the
four individuals appeared normal. On the following day one had
died, another on the ensuing day, and the last two on the fourth
day. Another time the same experiment was tried but with only
one minute of exposure. The result was the same, death without
division. The death of these organisms at this time cannot neces-
sarily be ascribed to the treatment, for a glance at the diagram
shows that the entire race was dying and that divisions were infre-
quent in all cases.
4. Nitro-Glycerine.
At the suggestion of Professor Wilson, and as a last resort, I
tried two experiments when the race appeared to be dying out in
December. Nitro-glycerine in very weak solution (unfortunately
I have no record of the strength used) was put into the hay
infusion. It made no appreciable difference in the final result
and the organisms did not divide.
Professor Wilson's other suggestion seemed more hopeful, on
the a priori ground that renewal of vitality is effected by the union
of two individuals. A culture of Paramcecium fresh from
pond water was made, and hundreds of individuals were allowed
to dry in a small drop of water in a watch crystal. When dried
the remains were scraped together and pulverized, the powder
thus formed being added to the hay infusion in which the weak-
ened Paramcecium were kept. Although this extremely ingenious
suggestion was worthy of a fruitful result, the outcome of the
experiments was the same as with all the rest, and not a single
individual lived after the 19th of December, one week after six-
day treatment with the dried Paramcecium.
There remain many experiments that might have been tried,
and that might possibly have accomplished the same results that
were obtained in the earlier periods of depression when the race
was successfully reinvigorated by artificial means, and even the
experiments that were tried might have been successful if different
strengths, or times of action, had been used. Many suggestions
were made by my colleagues and other friends, especially in
regard to the trial of some chemical compound. I am pleased to
436 Gary N. Calkins.
acknowledge the friendly and scientific interest which prompted
these suggestions, and desire to state that if they were not always
carried out, it was because of the limits of my time and of the con-
stantly decreasing number of individuals left to experiment with.
It was my desire to try as many classes of experiments as possible,
and some of these might have been successful if tried at an earlier
time or if carried out on a sufficiently large scale, but here again
the scarcity of living material would not allow the continued
experimentation along lines that were fruitless on the first trial.
It must be remembered that such experiments, to be of any value
in a work like this, had to be made on the material that had been
under constant observation for nearly two years, and preliminary
experiments with fresh forms from the ponds were valueless so
far as indicating the effect on the vitality of the race under
observation.
III. PROTOPLASMIC STRUCTURES OF PARAMCECIUM.
I. The Norrnal Paramceciutn.
The usual size of a normal Paramcecium is from 200 to 300
microns, and the form is fairly constant, warranting the designa-
tion "slipper animal." In all of the preserved specimens that I
have made from time to time, the fixing fluid was saturated cor-
rosive sublimate to which was added 10 per cent of glacial acetic
acid. Having a common method of fixation the different indi-
viduals can be compared point by point.
a. The Endoplasm. The endoplasm of a normal form is made
up of various granules of diff'erent sizes, of vacuoles and crystals
(Fig. i). When the animal is moving about in a nutrient
medium it constantly takes in food with the water absorbed.
The food of Paramcecium consists of bacteria, and these accumu-
late in a gastric vacuole until the latter has attained a certain size
when, according to Wallengren,^ it is caught up in the endo-
plasmic flow and carried to the posterior end of the body. It then
moves anteriorly toward the left side, ultimately passing over to
the right and then down on the right side. In this migration of
^ H. Wallengren. Inanitionserscheinungen der Zelle. Zeit. f. Allg. Physiologie
I.. 1, 1901.
Studies on the Life History of Protozoa. 437
the vacuole the food is brought into the immediate vicinity of the
macronucleus where the effect of the nuclear environment is
shown by the immediate acid reactions with congo-red of the
vacuole contents (Wallengren). The food material in such a
vacuole is massed into a more or less homogeneous body corre-
sponding to Greenwood's observation on Carchesium, and in this
condition the digestive fluids work upon it to resolve it into
digestible and indigestible parts. After this the soluble portions
are absorbed and the residue defecated. The soluble portions
pass into the endoplasm to be stored up as reserve food (Wal-
lengren) from which they are taken as the need comes to be
made into living molecules.
The processes of digestion thus given rise to definite elements
in the endoplasm, elements which react to stains in characteristic
ways. In addition to these, however, we might expect to find
waste matters due to incomplete oxidation as well as final products
of metabolism in the form of crystals, etc. The various possibili-
ties of this nature have given rise to different interpretations upon
which my own observations throw but little additional light.
With neutral-red acting upon the organism when alive, Prowazek^
distinguished three kinds of granules in the endoplasm: (i) The
food balls; (2) Small round granules which are distributed
throughout the periphery more or less uniformly (Prowazek
actually found them at the two extremities and about the mouth);
(3) Minute granules distributed throughout the endoplasm and
all over the body.
The minute peripheral granules (No. 2) are interpreted by
Prowazek in the same way that Wallengren had previously inter-
preted similar bodies in Pleurocoptes hydractinicB, viz., as excre-
tory vacuoles with a solidified granule of excreta within them.
Piitter,2 on the other hand, interpreted them as basal bodies of
centrosome nature lying at the bases of cilia. Wallengren subse-
quently showed, however, that the granules in question are not
at the bases of cilia but lie beside the cilia, and that rows of these
granules alternate with rows of cilia. He interpreted them as the
^ Vitalfiirbung mit Neutralrot an Protozoen- Z. wiss. Zool. Bd. 63, 1898.
^ Studien iiber Thigmotaxis bei Protisten. Arch. f. Anat. u. Phys. 1900.
438 Gary N. Calkins.
papilliform external swellings of the trichocysts and as merely
condensed peripheral portions of the cortical plasm. My own
observations support those of Wallengren.
The third type of granule is interpreted by Prowazek as a
ferment or enzyme bearer. Putter, on the other hand, believes
them to be "respiratory granules" owing their staining capac-
ity to the contained carbon dioxid. Wallengren's observations
on starving forms led him to the belief that neither interpretation
is correct, for, he argued, these granules being the first to disappear
in hungry forms must be of the nature of stored food (see Figs.
23 and 24).
The crystals which are found in well-fed forms were identified
by Schewiakoff as metaphosphate of calcium. They are of
various forms and sizes and are confined to the endoplasm; being
crystalline in nature they cannot be mistaken. They are now
generally regarded as late metabolic products resulting from pro-
teid digestion.
b. The Ectoplasm. As in the majority of holo- and hetero-
trichida the ectoplasmic modifications are well diflPerentiated from
the endoplasm. A cuticle and underlying cortical plasm may be
made out, the latter consisting of a much more dense substance
than the endoplasm, analogous, probably, to the ectoplasm of an
amoeba. In it are embedded the characteristic trichocysts
which ordinarily project ever so slightly from the surface, giving
rise to the minute papillae which may be distinguished in profile
between the furrows of the cilia (shown in Fig. 20). In Para-
moecium taken fresh from the pond water, the fixing agent which
I have used, preserves the trichocysts within the cortical plasm,
but after a few months under cultivation these organs cannot be
made out, and seem to have been discharged and lost under the
stimulation of the fixing fluid. Wallengren believes that they are
taken into the endoplasm and digested as food in starving forms,
but in preparations made from my cultures they are absent in the
well-fed forms as in the degenerate ones. In all cases the spaces
that were occupied by the trichocysts are present in the cortical
plasm as vacuoles, and it is in this state that the relation to the
peripheral papillae can be easily made out (<:/. Figs. 13, 18, 20
Studies on the Life History of Protozoa. 439
and 21). The difficulty appears to be that the cortical plasm is
incapable of holding the trichocyst threads after expulsion,
for the threads may be easily seen as a cloud around the
animal immediately after fixation, while the after-treatment always
dislodges them in the cultivated forms, but not in the wild forms.
c. The Macronucleus. The structure of the normal macro-
nucleus of ParamcBctum aurelia was described by Hertwig in 1889
and the nucleus of P. caudatum agrees so closely with it, that
further details are hardly necessary. It is an elongate, ellip-
soidal body, usually with a smooth contour and without breaks
of any kind save the minute impression made by the micro-
nucleus. It frequently lies in a vacuole which is caused by
the action of the fixing fluids, for in life the macronucleus is in
immediate contact with the endoplasm. The contraction is
probably in the endoplasm away from the nucleus rather than a
contraction of the latter. Often there is a depression in the
macronucleus due to the pressure of the contractile vacuole, and
food vacuoles may also press against it, as Wallengren suggests,
and distort it out of the normal proportions.
In its finer structure the macronucleus is granular with the
irregular granules densely packed together, giving the appearance
of a homogeneous mass. ■
d. The Micronucleus. The micronucleus is usually embedded
in the material of the larger nucleus, but may be separated from
the latter, even in the resting stages, by a considerable distance,
while in the dividing stages it is usually separated. Its finer
structure consists of a more or less homogeneous mass of chro-
matin frequently arranged in lines, while at one end is an accu-
mulation of "achromatic" material in regard to which there is
some diff"erence of opinion. In size the micronucleus is about
1 1 microns, but in the difi^erent phases the size diff"ers so that this
characteristic has but little weight.
e. The Contractile Vacuoles. In the normal individual these
are situated in the anterior and the posterior parts of the body, and
about one-third of the length of the body from the ends. They
are fed by radiating canals which are conspicuous in the living
animals. The pulsation is regular as a rule, but this becomes
440 Gary N. Calkins.
spasmodic after prolonged captivity under a cover glass, and the
irregularity is an index of the ultimate disintegration.
2. Structure of Paronicecium m Depression Periods.
a. Starvation and its Effects. The periodic depressions which
were noted in the experiments, and which appeared at more or
less regular intervals (viz: about every six months) were note-
worthy because not always accompanied by the same type of
degeneration as that characteristic of starved forms.
A most comprehensive study of the structures of starved Para-
mcecium was made by Wallengren, while various observers have
called attention to the characteristic vacuolization which the cell
protoplasm undergoes during starvation or at degeneration
periods in any culture. In general, Wallengren found that the
animals first use up the food material that is stored in granular
form, in the endoplasm, and that when this reserve is used, the
animals in lieu of other food, burn up first their endoplasm and
then the cortical plasm. There results from this destruction,
great vacuoles in the cell body which increase in size until the
entire organism is distorted through the pressure of one confluent,
or two or three great vesicles. Wallengren obtained his material
by transferring the Paramoecia to tap water again and again, and
thus ridding the medium of the customary bacterial food in a very
short time. My own experiments to this end consisted in leaving
the ciliates in a culture glass such as I have used throughout my
experiments, until all the bacteria had been eaten and the culture
medium had cleared. Thus a hundred or more individuals would
be left for a period of a month or six weeks in the culture chambers
where evaporation was prevented, and here they were watched
daily until they ultimately died of starvation. While Wallen-
gren's experiments were undertaken for the purpose of deter-
mining the efi^ect of starvation upon all of the protoplasmic
structures of these forms, mine were done for the purpose of
studying the effects of such treatment upon the nucleus and endo-
plasm, and general vitality. Wallengren found the following
effects in the protoplasm of Paramcecium after starvation for a
period of from 8 to lo days, which he designates as the "first
Studies on the Life History of Protozoa. 441
period" in the inanition phenomena: "All gastric vacuoles and
food balls disappear during the first period. After this the small
endoplasmic granules are used. As a result of this the quantity
of endoplasm becomes much reduced. Toward the end of this
time the living substance of the endoplasm itself is used, in part
at least, to supply fuel for the continuing metabolism. Owing
to the disappearance of the inclusions of the endoplasm and to the
use of endoplasmic substance itself, the body form becomes more
or less distorted or changed. But even in those individuals in
which this has taken place and in which the form is considerably
changed, the ectoplasm with its trichocysts, the contractile vacuole
and the cilia are not altered in any noticeable manner. This
shows, therefore, that in the first period of inanition the first
materials to be used are the reserve stuffs which are normally
utilized for the ordinary fuel (life processes). Only when all of
the reserve material is used and when the endoplasm itself is first
attacked, and only when all food whatsoever is gone, will other
parts of the protoplasmic structures be attacked. When this time
comes the second period is inaugurated." (Loc. cit., p. 87.)
In the second period of inanition there are more fundamental
changes, and the remainder of the protoplasmic structures are
involved. Ultimately the nucleus is affected and when this goes
the organisms are doomed. Wallengren's conclusions as to this
period are as follows: "The endoplasm, which at the beginning of
this last period was already considerably reduced, now appears
strongly vacuolized. The shining vacuoles which are probably
filled with the products of degeneration of the endoplasmic con-
tents, may attain to a considerable size. Along with this vacuol-
ization the ectoplasm becomes more and more absorbed and as a
result of this, the trichocysts are drawn into the endoplasm
streams and are probably digested. Along wnth them the small
papilliform swellings on the outer surface disappear, and with
these the small granules which in the living animal stain with
neutral red. The contractile vacuoles and their feeding canals
become reduced in the same proportion as the thinning of the
ectoplasm. A number of cilia are probably absorbed as a result
of the decreasing size of the whole animal, and the remainder of
442 Gary N. Calkins.
them are shorter than the normal. Owing to the inner changes
the whole organism may at this time be so modified that it is
unrecognizable."
"Thus during the continued inanition of the body, first one part
and then another becomes absorbed, first the endoplasm, next the
ectoplasm, the trichocysts and the cilia in part, all to maintain as
long as is possible the vital functions. In the meantime, however,
the nucleus has not escaped without changes as follows:" (loc.
cit., p. 98) . . . "In the inside of the macronucleus a
rounded mulberry-like mass is developed. Its alveolar structure
has changed at the same time, and in the center there are usually
one or two small central bodies (Binnenkorper). The high
pressure which is developed in the decreasing body form and due
to the enlarging vacuoles, causes the nucleus to become greatly
deformed and compressed. The various parts of the nucleus are
broken up into fragments which may probably be used more or
less as food ( t). Of the former large macronucleus there is now
left unchanged only the nuclear body which has been formed and
this lies between the broken down nuclear parts." (Id., p. 112.)
"In the micronucleus no destructive changes are mani-
fested during the hunger degeneration. It is the one part of the
body which is apparently not affected by the conditions of the
experiments, a not unnatural result considering the importance
which this organ of these cells has in rebuilding the macronucleus
after conjugation. Of all organoids the micronucleus would thus
seem to be the most important of the cell." (Id., p. 114.)
These careful observations and clear results of Wallengren,
most of which I have been able to verify, offer a good basis for the
comparison of structures obtained in the different stages of the
life history of Paramcectum {cf. Figs. 22 and 23). We may
distinguish two types of degeneration changes in the series from
the start to the finish. One set accompanies starvation, and was
characteristic of the first two periods of depression, the other
accompanies physiological depression of a different type at the
last two periods. In the former the changes in structure had to
do mainly with vacuolization of the endoplasm and rupture of the
macronucleus, while in the latter the endoplasmic portions were
Studies on the Life History of Protozoa. 443
degenerated in a different way. The ectoplasmic parts and the
micronuclear structures were not affected until the last depression
period.
The first clearly marked period of depression came in July, about
six months after the cultures were started. It was characterized
by a well-defined reduction in size (down to 109 microns; see
Fig, 3), and by vacuolization of the endoplasm while the ecto-
plasm did not appear to be much involved. Many of the individ-
uals were characterized by great vacuoles similar to those in
starved forms, which distorted the body almost out of recognition,
in others the nuclei were fragmented into two or three parts, and
in all there was a marked absence of the larger food granules and
gastric vacuoles which characterize the normal animals, and this,
notwithstanding the fact that bacterial food was present in abund-
ance (see Studies I). As stated in these Studies (III) the
organisms under these conditions still take food and in some cases
the endoplasm appears opaque with the undigested food balls, but
the decrease in size continues and the endoplasmic vacuolization
is not prevented by the presence of the food. It is the digestive
function, apparently, which becomes ineffective at such periods,
and if this is a correct assumption, this function can be stimulated,
as I have shown by the experiments.
Identical results were obtained in the period of depression in
December, 1901, a depression which was again overcome by the
use of beef extract, while the individuals of the series which had
been continued on the hay diet, all died. These became smaller
and smaller, and again gave morphological indications of starva-
tion, notwithstanding the fact that the individuals which had been
stimulated with the beef extract were living and reproducing
normally in the same food medium. They became much reduced
in size, the endoplasm became distorted with vacuoles, and they
died with absolutely no indication of disease through parasites.
These observations show, therefore, that starvation effects may
be produced even though the animals are living in a medium rich
in food. It is trite to say that to prevent starvation we must have
not only food but the ability to digest and assimilate it, yet com-
mon as this observation is, it is important in the present connec-
444 Gary N. Calkins.
tion and involves a factor which cannot be overlooked in any dis-
cussion on old age.
In the June period, as stated previously, the same conditions
w^ere not observed, for the organisms, in part at least, had been
treated with the beef extract every week during the first three
months, since the previous period of depression. The division
rate began to run down in the case of the B series in April, in the
A series in May, and in all of thq material that had been continued
on the beef, the characteristic structure was a densely granular
endoplasm (Fig. 7). In the specimens that had not been treated
with the beef since the preceding December, this character of the
endoplasm was not noted. These unstimulated individuals died
out in about the 508th generation (B series) after becoming much
emaciated and reduced in size, and with reduced nuclei. The
nature of the protoplasmic changes is indicated, in one case at
least, in Fig. 14. Here the macronucleus has entirely disap-
peared, not even a granular trace remaining, while the endoplasm
is crowded with vacuoles of considerable size. The micronucleus
is slightly hypertrophied and has a very peculiar outer membrane
within which the chromatin and achromatic material lie in what
appears to be the real nuclear membrane. The dense granules
characteristic of the beef-fed individuals are absent. The un-
stimulated A series did not die out until about two weeks later.
At the time when the B individual described above died (May 12)
the unstimulated A series was characterized by somewhat reduced
size, a declining division rate, and absence of the dense protoplas-
mic granules. In the stimulated A series, on the other hand,
(Ai and A2) of about the 560th generation, the structures were
normal, gastric vacuoles were numerous and divisions were fre-
quent. Towards the end of June, however, when the A series
nearly died out in the 620th generation, the conditions were very
different. Fig. 7 is from a specimen in the 615th generation. Its
size is below the normal; its endoplasm is choked up with granules
and there is no trace of vacuoles save the contractile vacuole near
one end. The macronucleus is definitely granular, and its con-
tour 'is irregular as though devoid of nuclear membrane. The
micronucleus is elongate and spindle-formed. The ectoplasm is
Studies on the Life History of Protozoa. 445
not deformed and save for the absence of trichocysts it appears to
be normal. This was the condition of the protoplasm when the
usual large number of culture individuals was reduced to 6 A's
and no B's, and a condition from which the A series were rescued
only with the greatest difficulty by the use of pancreas extract.
Figs. 8, 9 and 10 represent individuals that had been in extract of
pancreas for 48 hours, and then transferred to hay infusion. They
are identical, therefore, with the individuals that lived and carried
the race to the 742d generation. In these forms the endoplasm
in most cases is normally vesicular in the center and gastric
vacuoles are common, while the ends alone still retain the dense
granular aspect.
From this time until the race died out the division rate was
sluggish. The conditions of the protoplasm in the later individ-
uals was decidedly characteristic (Figs. 17, 19, 20, 21 and 22).
Throughout the fall, individuals would appear with densely granu-
lar protoplasm, which is invariably the sign of death, unless the
animals are stimulated in some way. In such forms the macro-
nucleus may or may not be normal, whereas the micronucleus as
a rule becomes hypertrophied and the ectoplasm full of great
vacuoles. Fig. 22 is a good representation of the conditions
at this time. The endoplasm is apparently normal; there are
food vacuoles and endoplasmic granules, and vesicular structure,
but the micronucleus is spherical and vesicular, has lost its usual
place in a niche in the macronucleus and shows evidence
of granular modification of the previously homogeneous chro-
matin.
The sister-cell of the one pictured in Fig. 22, and one of the two
oldest of the A series (742 generations), showed the following
points while alive: "A12 was alive this morning and was picked
out for examination. It had two contractile vacuoles situated
dorsally and close together. The astral canals were absent; in
their place was a row of dorsal feeding canals, such as those
characteristic of the more generalized holotrichida {e.g., Chlamy-
dodontida:). The rest of the body contained eight or ten large
vacuoles not contractile. The macronucleus was slightly hyper-
trophied, and visible, indicating the approach of disintegration.
44^ Gary N. Calkins.
The papillae of the cuticle were plainly visible and what T have
taken to be apertures of the trichocysts were more or less
numerous. (This is shown in the preserved sister-cell, Fig. 22.)
A few trichocysts remained in the cortical plasm, but there were
many vacuoles in this layer indicating that when the trichocysts
were discharged they were not re-formed. The peristome was
normal and the mouth had a vigorous oral membrane. The
size was large, fully as great as any of the preparations that
had been made at any time during the 742 generations. Move-
ments vigorous to slow, with a tendency on the part of the animal
to remain stationary." ^
It was while the organisms were in this structural condition that
the many attempts to rejuvenate the race were made as described
in the previous pages, and it was in this condition of the proto-
plasm that the race finally died out from exhaustion. Before
dying, however, the individuals, as indicated in the above para-
graph from my notes, were of full size and were filled with gastric
vacuoles and partly digested food, while the body form was
normal, (compare Figs. 2 and 21).
It must be admitted that these forms were capable of individual
growth at this period and, since the macronucleus was normal in
the last individuals luhile the micronucleus was considerably
changed, it must be further admitted that the vegetative metabolic
processes were presumably reinvigorated; on the other hand, the
functions of reproduction; that is, of division, were degenerated
possibly, if not probably, because of the apparent degeneration of
the micron-ucleus and of the cortical plasm, whose functions were
not reinvigorated by the artificial means which were tried.
IV. GENERAL DISCUSSION.
Although only a beginning has been made to determine the
objects for which this series of experiments was started, it is
advisable to bring together the results thus far attained and to see
how they conform with the a prion conceptions which were cur-
rent at the outset of the experiments.
1 From my note book.
Studies on the Life History of Protozoa. 447
It is not out of place to consider first the initial objects of the
undertaking, although at the risk of again repeating what has been
often stated.
1. The first aim of the experiments was to get light upon the
general phenomenon of conjugation and through this, upon fertil-
ization in general.
2. To determine whether conjugation is imperatively necessary
for rejuvenescence.
3. To determine whether artificial rejuvenescence is possible.
4. To determine the conditions, antecedent and subsequent to
conjugation.
5. To determine, if possible, the significance of rejuvenescence.
6. To determine, finally, whether protoplasm in these simple
forms is capable of indefinitely continued life without conjugation,
or whether it is subject to the conditions of "old age."
On none of these points can a definitely positive answer be
given, and further experiments must be undertaken to clear them
up. The fact that, after a continuous cultivation of 742 genera-
tions, covering a period of 23 months, the race died out apparently
from exhaustion, shows that under the conditions, continued life
was impossible, and if this conclusion, which seems to be the only
one justified by the results, be granted to obtain in nature, then
we must agree with Maupas and others that the indefinite con-
tinuance of life without conjugation, is improbable.
I. The Conditions of the Experiments.
The question has been raised whether the conditions under
which the experiments were undertaken were in any way abnor-
mal to Paranicecium, and whether, from the results obtained, we
are justified in carrying the conclusions to the free-living forms,
and to similar types in general.
It might be objected that the space allowed was inadequate;
or, that the light conditions were abnormal; or, that the water
would get foul; or, that they were given only one kind of food;
or, that they were subjected to pressure. If we examine these
objections critically we shall find that they have little basis.
Let us consider first the matter of space, for this involves some
of the other objections, viz: pressure, volume conditions, and the
448 Gary N. Calkins.
like. The actual amount of water that was used for each isolated
individual was one-half a cubic centimeter. This was contained
in a small chamber consisting of a hollow-ground slide, two glass
supports about 3 mm. thick, and a thin glass cover. The Para-
incecium had ample room, therefore, for free movement, and an
actual depth of water of more than an eighth of an inch. Pres-
sure, therefore, was out of the question. In such a slide chamber
individuals were kept (/. e.^ extra individuals from the "stock")
for periods considerably longer than six weeks without change of
water, showing that the mere quantity was sufficient in order to
keep the animals alive. Foulness of the water, accumulation of
carbon dioxid, lack of oxygen, etc., were all guarded against
by the almost daily transfer of the culture individuals into fresh
hay infusion. The salt content of the water remained practically
constant, for fresh hay infusion was used each time with the same
amount of hay from the same source while the weekly analysis
of Croton water shows only minor fluctuations in the small
quantity of salts in solution. The gradual decrease in vitality
cannot be attributed to these causes, a similar phenomenon being
a matter of common observation and noticeable in any culture of
protozoa, no matter how large the vessel, nor what the species.
The light conditions were similar to those in any laboratory, the
culture vessels being kept before a window with north exposure.
In regard to the possible objection that the Paramcecium ob-
tained only one kind of food, and therefore that the conditions
were abnormal in this respect, it may be stated that such a condi-
tion of treatment is a sine qua non of the experiments, and the only
possible means of controlling the results, and as I have demon-
strated, it is by a change of diet, including salt constituents, that
the periods of depression are overcome. This objection, there-
fore, begs the question of an object of the investigation.
It seems quite unnecessary to repeat again that the only normal
life possible to Paramcecium caudatum is in the ponds where it is
subject to the changes in chemical composition of the water, to
the exigencies of drought, of heat, of freezing, and of rest by
encystment or lack of food. In the laboratory the protoplasmic
activities get no rest, but day after day they are maintained at the
Studies on the Life History of Protozoa. 449
optimum rate and such conditions can by no stretch of the imagin-
ation be called identical with those of the ponds. Yet the "nor-
mal conditions" may, after all, be but a matter of definition. If
we leave a hay infusion to stand exposed to the air, Paramcecium
will ultimately appear in it, and will ultimately die out from it.
The appearance and disappearance cannot be called artificial, it is
as much normal for Paramcecium to appear in such an infusion as
it is normal for the bacteria upon which the animals feed to be
there. City life for man may be called an artificial life as opposed
to the "normal" original or pastoral life, but it is no less normal
now than the primitive life was, even if it is found that the average
length of life is shorter for urban than for country-dwelling people.
The course of human life, or the history of the race, physiologi-
cally speaking, is no less normal for being rapid. In the same
way we may argue for the race of Paramcecium and its life in the
culture chambers of the laboratory. The life pursues a normal
course, although possibly faster than in nature, and the ultimate
results obtained in cultures may be confidently expected to obtain
sooner or later in the natural habitats. Seven hundred and forty-
two generations represent a long time for organisms to live and
develop in a medium that is not normal, and the mere fact that
they do so live is sufliicient evidence to prove the point. It seems
to me, therefore, perfectly legitimate to take the phenomena of
vitality in Paramcecium in culture as practically identical in
outcome with the phenomena in natural surroundings, and as
indicative of what goes on in living protoplasm under "normal
conditions."
Looked at from this point of view, the experiments teach (i)
that a given form together with the race derived from it will
exhibit periodic depressions in vital activity; (2) That such
depressions can be overcome by artificial means ( and probably
but not surely, in nature by opportune changes in the immediate
environment). Further than this, however, the experiments
teach, (3) that these depressions are not all of the same type, nor
due to the same causes. They give reason for the belief that
periods of depression may ensue wherein different functions give
out, and that when this occurs, as for example when the cortical
45 o Gary N . Calkins.
plasm and micronucleus show evidences of degeneration, all of the
experiments that we may try to artificially reinvigorate them,
will probably be futile. This may indicate one of two things,
viz: that under natural conditions changes m immediate environ-
ment would be insufficient to rejuvenate when the organisms are
in this ultimate state of exhaustion, unless, indeed, the experi-
ments failed to eliminate all of the chances such an organism bas-
in nature; or, conjugation is a necessary condition of continued
protoplasmic activ.ty.
I am inclined to the belief that some material might ultimately
have been found which would have helped the Paramoecia over
this period of extremity and would have stimulated micronucleus
and cortical plasm to continued work. The failure to find it,
however, indicates a like difficulty in nature and makes the
a priori reason most probable that the phenomena with which we
are familiar, namely, the processes of conjugation, have been
essential in mamtaming the races of Paramecium up to the pres-
ent time, and in keeping them from extinction.
2. Does Protoplasm Grozu Old?
The above considerations lead to the discussion of age in a
simple cell organism. In higher forms old age is manifested by
the gradual weakening of the vital functions, waste matters are
inadequately disposed of, or are retained in one form or another
in the cells and tissues; this involves the physical impairment of
organs and enhances the difficulties of their functional activities
until, by the accumulation of such mutually aggravating processes,
the organism ultimately dies of "old age." In Paramcectum
there is little morphological evidence of the onset of old age,
although, if we accept the impairment of the vital functions as an
index, we must conclude that diminution of the division rate,
decrease in size, etc., are evidences of this phenomenon in pro-
tozoa. So far as the accumulation of waste matters is concerned,
there is morphological evidence to indicate that this takes place
more frequently at periods of depression. There was no sign of the
crystals which frequently accumulate in the protoplasm of various
protozoa, and in the last specimens of the race (742d generation)
both endoplasm and macronucleus were normal in structure.
Studies on the Life History of Protozoa. 45 1
The surest evidence of what may be considered old age
in this form, was therefore, functional, and was expressed by
diminished division rate and by the increased frequency of ab-
normal binary fission. Abnormal division, as a matter of fact,
like nuclear hypertrophy, may occur at all periods and marks
some particular weakness of the single individual; occurring more
frequently, however, at certain periods of depression, such ab-
normalities give evidence of general protoplasmic weakness.
The various types of incomplete division are very instructive and
a more prolonged study than I was able to give to them might
afford positive evidence of the nature of the pathological changes
involved. In some of the specimens which T obtained during
periods of depression, the macronuclei and micronuclei appear
normal; in others there is a macronucleus in each of the daughter
individuals, but the micronucleus is undivided; in others the
macronucleus is divided but remains in one individual, the micro-
nucleus is undivided and remains with the original macronucleus,
while the daughter individual has no trace of nuclei. In all cases,
finally, of pathological division the cortical plasm appears ab-
normal and vacuolar, while the endoplasm is very frequently
disintegrated and abnormal (Figs. 25 and 26).
While these observations are too few to permit far-reaching
conclusions, they are sufficient to indicate that some protoplasmic
changes have taken place, and further, that the cortical plasm has
become modified in some way. Indeed, the inability completely
to divide may be accounted for by the loss of vitality in this par-
ticular part of the protoplasm, for in the majority of cases the
initial stages of division are safely passed, the final separation
alone being retarded and usually omitted altogether, so that
monsters of three or four individuals may be formed through the
continued incomplete division of the original degenerate specimen
(Fig. 25). As is well known, the cortical plasm is the seat of the
myoneme formation, of the cilia, and of other motile organs, and,
in general, may be said to possess kinetic or motor functions.
That this portion of the protoplasm is subject to change is shown
by the fact that at certain times the outer protoplasm becomes
452 Gary N. Calkins.
sticky or plastic and to such an extent that two individuals upon
meeting, fuse together in plastogamy. This, which I have
termed elsewhere the "miscible state," may be so marked that
groups consisting often of from five to eight aggregated individuals
are occasionally seen. It is analogous, apparently, to the plas-
togamy so often seen in the fresh water testacea such as Diffliigia
or Arcella, which Schaudinn has recently shown to have no
connection with conjugation in these instances. In Paramcecium
during this miscible state, conjugations are for the only times
possible, and many complete conjugations are found together
with the fused multiple individuals. There is no doubt, then, that
the cortical plasm changes in physical condition, and there is equal
reason to believe that at periods of depression when these abnormal
divisions are more frequent, the cortical plasm shows degenerate
conditions^ or, possibly, a condition of old age.
There is therefore some significance in the fact that the cortical
plasm gives out; some significance connected with the diminishing
division rate and with advancing old age as evidenced by dimin-
ishing activity.
While there may be some uncertainty as to whether the decreas-
ing vitality of a race of Paramcecium is evidence of normally
decreasing functions indicative of protoplasmic old age, or of
some other cause of degeneration, there is absolutely no reason to
believe that it is due to a parasite of any kind, nor to any harmful
substances in the medium in which they live. In the earlier
periods of depression there seems to have been a gradual loss of
powers connected with metabolism, and of something which was
vitally important to the race, for unless the individuals were stimu-
lated, they inevitably died. This was strikingly demonstrated in
the period of depression in December, 1901, when a number of
individuals of the regular series were continued on the usual hay
infusion, while others were treated with beef extract for 24 hours,
and still others with salts of different kinds for not more than 30
minutes. The non-stimulated forms showed increasing sluggish-
ness and depression, and all died in the course of two weeks, while
the sister-cells which had been stimulated, lived with varying
fortunes until a year from then (see Diagram I). The pertinent
Studies on the Life History of Protozoa. 453
questions may be asked was it old age from which the organism
died ? and, if so, what form did it take ? They were fed daily with
the same food upon which the stimulated sister cells thrived, but
they could not assimilate it and would not grow nor divide. In
similar cultures which had been carried to a like point by previous
observers, the entire race died, and although no evidence of
structural degeneration was evident, it has been taken for granted
that their organisms died from exhausted vitality, or in other
words, of old age. In my cultures there was some evidence of
degeneration, especially in the endoplasmic structures and in the
macronucleus.
The fact that stimulation was successful in carrying the race
through this earlier period of depression indicates either that the
conditions are not the same as those accompanying old age in
metazoa, or else that such conditions may be satisfactorily over-
come. I believe the conditions are more or less the same in both
cases, and that in senile Paramcecta certain functions have become
retarded, possibly by the accumulation of useless protoplasmic
elements too minute to be detected, or by some less mechanical
cause connected with the molecular structure of protoplasm and
which, therefore, affords no morphological evidence of change.
Such an hypothesis would explain the difference in length of time
required to get positive results in the stimulation experiments.
For example, in August, 1901, after the race had been on hay
infusion continuously for 7 months, it was necessary to keep the
single individuals on beef extract for three weeks before they
would live again in the hay infusion. But in December it was
necessary to keep them on the stimulant only a day or two to get
the desired result. The short treatment at this period sufficed,
because they were not allowed to become weakened to the same
extent as in the preceding period of depression. This result
points to some physical condition of the protoplasm, possibly to
the accumulation of some protoplasmic product or products which
lead to diminished vigor and to death. Reinvigoration after a
more or less prolonged treatment with the beef extract and stimu-
lation by this and other means indicates that such materials are
disposed of, or, more generally speaking, and to use a phrase which
454 Gary N. Calkins.
Wilson in The Cell attributes to O. Hertwig, that the condition of
stabiHty is changed into one of protoplasmic lability.
While such an hypothesis accounts for the first two periods of
depression, it fails to account for that of June and of December,
1902. In the interval between June and the preceding December,
the race in part, had been treated weekly with beef extract until
the first of April, after which the organisms had been fed with the
usual hay infusion. In June they began to degenerate, and from
this time on, treatment with the beef extract was futile, and the
race was finally saved only by using extract of pancreas and of
brain. This, however, gave only temporary relief and complete
activity was never again recovered and the division rate remained
below the average, until the race finally became extinct in Decem-
ber, 1902, and this despite the fact that, morphologically, the
endoplasm and macronucleus were restored.
Was the last period of depression running from June until
December, 1902, an expression of old age ? From the structures
of the organism and their behavior, there is no doubt that the
ailment at this period was different from that of the earlier
periods of depression, and there is no doubt again that the reme-
dies which had succeeded at the earlier periods failed completely
at this. The final depression of vital activities may be accounted
for by one of two assumptions : (i) There was an accumulation of
waste material of a different kind from that of the earlier periods,
or a different physical condition, and a weakening of different
functions, or (2) certain elements in the protoplasm endowed with
a given potential of activity used up that potential and failed to
recover it by artificial stimulation. Or a third hypothesis may be
conceived which embodies both of these. The morphological
structure at the final period shows that some different elements of
the body were involved in the last period of depression, and that
the elements which had given out in the previous periods were
satisfactorily reinvigorated even in the last individuals of the
race. Thus the micronucleus and the cortical plasm showed
unmistakable signs of degeneration in the last few individuals of
the race, while the endoplasm and macronucleus were perfectly
normal in appearance, and metabolism, which these elements of
Studies on the Life History of Protozoa. 455
the cell appear to control, seemed to be equally normal, since the
organisms were of full size, while the endoplasm was full of
partly digested food. It appears, then, that the experiments
were successful in reinvigorating the elements of the cell that had
given out in previous periods of depression, but that other ele-
ments were now involved which all my experiments failed to
reach. Here a more deeply-lying malady had to be met, and the
experiments being unsuccessful in meeting it, the entire race
died out. This series of facts appears to warrant the assump-
tion that there is a fundamental difference in the protoplasmic
elements which go to make up the body of a protozoan, one of which
IS to be compared with the somatic cells of metazoa, the other tvith
germ cells; the one connected with the vegetative functions of
metabolism, the other with reproduction; the one may give out and
so lead to "physiological death" (Hertwig) or it may be restimu-
lated; the other may give out and so lead to "germinal death" of
the race.
It is not outside the range of possibility that the last depression
period might have been overcome by some suitable experiments,
and the fact that we did not succeed in finding a suitable stimulant
does not justify us in assuming that this period represents the last
vital spark of this protoplasm, any more than we are justified in
assuming that the earlier periods of depression represented this
condition. If, however, some element or elements of the proto-
plasm become exhausted and all experiments to replace them fail,
then we might justly speak of exhaustion or "old age" of these
elements of the protoplasm and affirm that old age in one form,
characterized the organisms during the first two periods of de-
pression, while it took another form in the final period.
3. Conjugation and Rejuvenescence.
"Old age," then, appears to be a natural condition of living
protoplasm and we may ask, is there any experimental evidence
to show that this condition may be overcome by natural means ?
It has been generally assumed by biologists that conjugation
brings about rejuvenescence in the conjugating individuals, and
so imparts to the ex-conjugants and to their immediate descend-
ants a high potential of vigor. During the process of conjugation
456 Gary N. Calkins.
there is a complete change of materials or of protoplasmic make-
up, and a thorough "reorganization," to use the excellent term
proposed by Engelmann. Not only is there a new conjoint micro-
nucleus with its chemical compounds derived from the union of
two nuclei from individuals of diverse environment, but the endo-
plasm and cortical plasm must receive new materials through the
disintegration and the absorption of the old macronucleus, and of
at least three-quarters of the old micronucleus. If, as I have long
maintained, there is a specific "kinetic" substance in the proto-
zoon nucleus, a substance which in the centro-nuclei forms the
division-center and which is found in the micronucleus of Para-
moecium, then the cytoplasm of Paramcecium must receive a
certain amount of "kinoplasm" at each period of conjugation
and from the experiments, enough to carry the race through a
complete cycle. In my cultures such reorganization by conjuga-
tion was prevented in the straight line of the experiments, and the
only opportunity for reorganization came with the change in diet.
This, indeed, seemed to be operative for some time, but ultimately
failed, as we have seen. In the stock material, however, material
left over after the individuals had been selected for the cultures,
conjugation experiments were frequently tried during the course
of the experiments, and the results have been given (Studies I).
Some of the results are very suggestive in the present connection,
for it was found that only a few of the ex-conjugants continued to
live, approximately 6 per cent of them. This result may be due, as
I have previously stated, to the fact that both of the gametes had
been under identical conditions of food, etc., and no new sub-
stances were formed by the union of similar nuclei and protoplasm.
Or the result may be due to the fact which Stevens^ calls attention
to, that conjugation is an exhausting process, and that, being
weakened through long cultivation in cultures, these ex-conjugants
did not have sufficient vitality to recover. This suggestion does
not set aside all of the difficulties, however, for we have still to ex-
plain the large number of cases where the ex-conjugants have lived
^N. M. Stevens Further Studies on the Ciliate Infusoria, Lichnophora and
Boveria. Arch. f. Protistenk., Ill, 1903.
Studies on the Life History of Protozoa.
457
DEC. JAN. 1 rea.
1 1902 1 1 1
MAR. . APR. 1 MAY. 1 JUMB . JUIY
Aue.
1 1
1 ' 1 1
1 1
1
1 ,
'
1 '
1
1 1
1
1 ,
1
1
1
1
1 '
1
1
1
1
tr
1
r— 1
1
1 1
)
1
1 1
1
1
1
n
' 1
' 1 1 1 1
III
' [ 1 1 1
■ 1 ' 1 j 1
1
1 ' 1 1 1
1
1 1 1 1 1
1 1 ■ 1 !
Diagram II.
Complete history of the endogamous ex-conjugant by ten-day periods,
and periods the same as in Diagram I
Ordinates
2.4
1 1 1 1902 1 1 1
} JULY 1 AUG. 1 SIPT. 1 OCT. | NOV. I DEC. | JAM.
FKB.
90S • 1 "
MAR. j APR. 1 MAY 1
n r
III
1
2.2
2.1
2.0
1.9
1.8
i.T
' '
r
t 1 ! • '
1
n
! ' :
' ' '
'
'III'
'Hi: I
1 L_r-
! 1 ■ 1
1
l.C
_
! ' ! 1
1
1 r
1
1 1
' 1
1
1.3
1.2
' 1 1 1 '
■
, ,
III'*
;
1.0
1 ' '
_!
1
' 1 ' 1 I ' 1
J
1
|-H
1
1 ' 1 1 1- ' '
1
i ' 1 1 ' ' •
0.5
1 ' ' 1 I ! !
1— ,
'Ill',
1
0.3
1 1 1 j ■
'
1
1
1 . 1 1 1 1 1
1
r
1
I ! • '
1
tr-.
1
! 1
Diagram III.
Complete history of the third series (C) by ten-day periods Ordinates and
periods the same as in Diagram I, As in Diagram I, this curve- represents the
average division rates of four lines of individuals
458 Gary N. Calkins.
and reproduced for from 8 to 20 generations, and with apparently
well-organized bodies/
One of these successful cases was an ex-conjugant from an
endogamous union of two individuals which were separated by
not more than eight or ten generations from the ancestral A in the
354th generation of my cultures. The other ex-conjugant died
out in the iith generation, while the successful one ran through
376 generations before showing signs of debility. It went through
eight months in culture without beitig stimulated, and died out
finally at the end of 376 generations, which was exactly three
generations less than the life of the third series of Paramcecium
(C series) which I started on June 18, 1902, and carried along in
culture until May 30, 1903, when it died out in the 379th generation
(see Diagram III).^
Unfortunately, this ex-conjugant has not an absolutely clear
record, for the first day after the pair had separated, I placed them
both in beef extract for 24 hours (December 9, 1901). This
experiment had failed a number of times, and I had no reason to
believe that it would succeed this time, and, as stated above, one
of the two ex-conjugants thus treated died after eleven generations.
Although at first I attributed the successful result to this treat-
ment, I do not now believe that the beef extract had anything
to do with the vigor of the race that followed, and believe
that rejuvenescence was accomplished by the conjugation and
nothing else. This conclusion is based upon the following facts:
(i) Other ex-conjugants similarly treated with the beef extract
failed to live; (2) the non-con jugating individuals of the regular
series which were treated with the beef extract at the same time
'See Studies I, table of conjugations opposite p. 174.
^We have, then, the interesting coincidence of an individual running through 354
generations in culture, conjugating with one of its OAvn close relations, and then, as
an endogamous ex-conjugant, running through 376 generations more, a total of 730
generations. Against this we must set the 742 generations of the main culture
series, and the 379 generations of the third series (C) . The close connection between
the 379 and 376 is very significant, and were it not for the fact that the first two
series, A and B, were at a fatal period of depression at the end of 200 and 190 genera-
tions, we might conclude that 370 more or less is the normal length of life of Para-
mcecium in culture.
Studies on the Life History of Protozoa. 459
had a lower division rate and died out before May 5; that is,
after running five months {cf. Diagrams I and Iiy It follows,
therefore, that something luas operative in the ex-conjugant that
was absent in the stimulated form, and this something could
be nothing else than the reorganization which follows conjuga-
tion. The accompanying curves show that the periods of
depression and death which menaced the regular series in
December, 1901, and again in June, 1902, were not paralleled
in the cultures of the descendants of the ex-conjugant, and
the conclusion is obvious that conjugation provided some stimu-
lus which enabled this line of Paramcecium to live through
periods in which the allied races were saved only by vigorous
treatment and stimulation. There is no doubt at all that, had
I tried to revive the race of the ex-conjugant by beef extract
at: the end of August, 1902, I could have done so, for there
was nothing serious in the nature of the depression at this time,
when I allowed them to die without making an effort to save
the race. It is now a matter of deep regret to me that I did
not try to save them, and see if they would live beyond the time
when the allied lines died out in December, 1902. Had they done
so, it would have been still more convincing proof that conjuga-
tion does actually rejuvenate and overcome the conditions of
so-called "old age." I believe that the evidence which I have
outlined above is quite sufficient, however, to establish this point,
the one questionable factor being the beef extract, and even this,
as I have shown, could have only a limited bearing and does not
at all outweigh the positive evidence in favor of the conclusion.
Columbia University,
April, 1904.
460 Gary N. Calkins.
EXPLANATION OF PLATES.
All of the photographs were taken by Dr. Edward Learning from permanent preparations of Para-
mcecium caudatum, stained with picro-carmine. All are equally magnified and the relative sizes
represent absolute differences.
Plate I.
Figs. I and 2. Two normal specimens B series (107th generation and after three months of cul-
ture in hay infusion. These do not differ from typical Paramoecium from the ponds, and have many
endoplasmic vacuoles, alveolar protoplasm, and homogeneous nuclei.
Fig. 3. A typical individual of the B series during the first period of depression. The ectoplasm
is fully as clearly defined, and as thick as in the largest forms, indicating that this portion at least, has
not suffered from degeneration, a result differing from that in starved forms. (Compare Figs. 23
and 24).
Fig. 4. An individual from the B series in the 306th generation, stimulated with beef extract in
August, fed continuously with hay infusion for three months until killed. The endoplasm is filled with
gaGtric vacuoles and with partly digested food, the dissociated or "labile" condition of the endoplasm
shown here is characteristic of Paramoecium under normal conditions.
Fig. 5. An individual from the A series during the third cycle (550th generation), and twenty-four
hours after treatment with beef extract. The endoplasm is filled with gastric vacuoles, the macronu-
cleus is normal, but the micronucleus has divided three times and a clump of six nuclei may be seen at
the lower end. There is a tendency toward a denser structure of the endoplasm, especially at the
two extremities, this being indicative of approaching physiological depression.
Fig. 6. An individual from the A series in the 560th generation. Treated 48 hours before fixation
with beef extract. Gastric vacuoles are abundant in the upper portion, but in the lower part the
characteristic density which marks the climax of physiological depression is shown, /. e., an apparently
general "loading" of the protoplasm with inert material.
Fig. 7. An individual from the A series in the 615th generation killed at a time of general depres-
sion. It shows the typical condensed appearance when the power of division is lost and leads to death
after several days without division.
Fig. 8. An individual from the A series in the 623d generation (June, 1902,) and 24 hours after
successful stimulation with extract of pancreas. The condition shown in Fig, 7 has been successfully
overcome, and activity renewed by this treatment. This and the two following figures show stages in
the breaking up of this dense, endoplasmic mass. The macronucleus is divided while the ends alone of
the animal still retain the densely granular character.
Figs. 9 and 10. Two individuals from the A series 48 hours after successful stimulation with
pancreas extract. The endoplasm is now in a "labile" condition, although the extremities are still
dense. The individual shown in Fig. 10 is further advanced in recovery than that shown in Fig. 9,
but both are sister cells of individuals that carried the race to the 742d generation.
Plate II.
Fig. II. Two individuals of the A series in the 604th generation, two weeks prior to the fatal
depression of June, 1902. These were treated with a weak solution of dibasic potassic phosphate (see
paper) for 30 minutes and then transferred to hay infusion and killed 24 hours afterwards.
Fig. 12. Two individuals of the A series treated at the same time as the preceding, with a weak
solution of magnesium chloride for 25 minutes. The protoplasmic structures are normal and the endo-
plasm has the typical alveolar appearance.
Studies on the Life History of Protozoa. 461
Fig. 13. Another individual treated with potassium phosphate. The micronucleus is separated
from the macronucleus. At the lower end the focus is sufBciently sharp to show the characteristic
papilli-form structure of the trichocyst spaces in the ectoplasm.
Fig. 14. One of the last individuals of the B series (504th generation) in which the macronucleus
is entirely gone. The micronucleus is distinct, and has its chromatin massed near one pole. The place
which held the macronucleus is marked by a large vacuole. There are no observations to indicate the
fate of this macronucleus, the break at the left side indicates that it may have dropped out at some period,
although this did not happen during the course of the treatment, because the same condition was
observed during its life, and immediately after killing.
Fig. 15. An individual from the B series in the 502d generation after treatment with beef extract.
The characteristic dense endoplasm is still present but there are many gastric vacuoles, while the micro-
nucleus has divided three or more times and the daughter nuclei have accumulated at one end.
Fig. 16. An individual of the A series in the 6o2d generation treated for 25 minutes with phos-
phoric acid. It was transferred to hay infusion and killed 24 hours afterwards. The macronucleus is
broken into fragments; the micronucleus has divided and one part (left center) seems to be forming a
new macronucleus. (This individual offers the only evidence obtained of nuclear fragmentation and
reconstruction through artificial means.)
Fig. 17. An individual from the A series in the 718th generation, killed October 20 after living six
days without division. The endoplasm shows a general absence of the larger granules, indicating
starvation; the micronucleus (dimly visible at the lower end of the macronucleus) is hyaline and without
chromatin, evidently degenerated.
Fig. 18. Two individuals of the A series in the 6o2d generation, treated with a dilute solution of
sodium chloride (see description), for 25 minutes. Transferred to hay infusion 24 hours afterward,
and killed.
Plate III.
Figs. 19, 20, 21 and 22. Four individuals at the end period of the A series. Note that in all of
these the macronucleus is not abnormally large as compared with normal individuals. This was true
throughout the entire race at this period and contradicts Hertwig's recent theory of the size-relations at
periods of depression. Fig. 19 represents an individual in the 720th generation, unusually small and
unlike the remainder of the culture at this time. Fig. 20 represents an individual in the 725th
generation, with conspicuously dense endoplasm and macronucleus. The latter bulges out towards the
observer and the effect of the ectoplasm about it is that of a special nuclear capsule. Fig. 21, an
individual in the 741st generation showing the looser texture of the endoplasm, gastric vacuoles and other
characters, indicating that these organs had been restored by stimulation. The micronucleus is hyper-
trophied, the macronucleus is normal. Fig. 22 represents an individual in the 742d generation, the
oldest of the race. It shows the reorganized endoplasm, gastric vacuoles, and the like, but ectoplasm
and micronucleus are degenerated. The former by vacuolization (note punctate appearance on right
of macronucleus) the latter by hypertrophy and loss of chromatin.
Figs. 23 and 24. These represent individuals which were starved for two and four weeks respect-
ively. Those in Fig. 23 were fed on beef extract August 19, transferred to hay August 20th and left
unchanged until September 19, when they were killed. The individual shown in Fig. 24 was not given
beef extract, but was left in hay infusion for two weeks unchanged, when it was killed. In Fig. 23
the spots at the lower ends represent the micronuclei, in Fig. 24 the upper elongated granule is the
micronucleus.
Fig. 25. A triple monster from an individual 72 hours after conjugation, with many nuclear frag-
ments and evidence of two incomplete divisions.
Fig. 26. A double monster from the A series in September, 1901. The micronucleus is undivided,
the macronucleus is deeply cleft and the individual on the right has no trace of nuclei.
LIFE HISTORY OF PROTOZOA. G. N. CALKINS
PLATE I
The Journal of Experimental Zoology
LIFE HISTORY OF PROTOZOA. G. N. CALKINS
PLATE II
The Journal of Experimental Zoology
LIFE HISTORY OF PROTOZOA. G. N. CALKINS
PLATE III
The Journdl of Experimental Zoology
STUDIES ON REGULATION.
V. THE RELATION BETWEEN THE CENTRAL NERVOUS
SYSTEM AND REGENERATION IN LEPTOPLANA:
POSTERIOR REGENERATION.
BY
C. M. CHILD.
With 47 Text Figures.
A. INTRODUCTION.
It is beyond the purpose of the present paper to review the whole
question of the relation between the nervous system and morpho-
genesis or the sustaining effect of "trophic" stimuli upon form.
It is hoped, however, that the observations and experiments to be
described, together with the interpretation offered, may serve
to throw some light upon this most interesting, but difficult
problem.
The question of the relation of the central nervous system to
regeneration in the lower animals has been touched upon by vari-
ous authors. As regards Planaria, a species which has been the
object of study by many investigators, opinions differ to some
extent at the present time. It is well known that in this form
removal of the cephalic ganglia does not interfere with complete
regeneration, the ganglia themselves being regenerated from por-
tions of the nervous system which may be present. As regards
other parts of the central nervous system, however, Bardeen ('03)
holds to the opinion that some portion of the nerve cords or of one
of them must be present in order that regeneration may occur,
while Morgan ('98, '00, '01, p. 44,) believes that regeneration may
occur in pieces from the lateral region of the body which contain
no part of the longitudinal cords. In an interesting paper dealing
464 C. M. Child.
with Phagocata and Dendroccelum, Lillie ('01) records the fact
that while Phagocata equals Planaria in its regenerative power,
the conditions in Dendroccelum are widely different. In this
form the capacity for regeneration of a head is limited to the
anterior third or fourth of the body, pieces from levels posterior
to this failing to regenerate. Lillie noted that the pieces of Den-
droccelum which were incapable of regenerating a head showed
a marked difference in reactive power from those in which such
regeneration was possible. On the other hand it is known (Loeb,
'94,'99, Parker and Burnett, '00) that pieces of Planaria deprived
of the cephalic ganglia react to stimuli in much the same manner
as normal animals. In view of these differences in reactive power
Lillie suggests that the stimulation of the normal movements may
determine the fate of the undifferentiated mass of new tissue, the
head failing to regenerate in the absence of the characteristic
stimuli. This suggestion is, I think, an important one.
As Lillie points out, Dendroccelum resembles in this respect
the earthworm Allolohophora joetida, in which, according to
Morgan ('97), regeneration of a head does not usually occur
posterior to the fifteenth segment. In later work upon this form
Morgan ('02) has discovered that the regeneration of the head
appears to be closely connected with the presence of an anterior
cut surface of the nerve cord, so that if two such surfaces are pre-
sented by removing the head and then cutting out a small piece
of the nerve cord a short distance posterior to the cut end, a head
will regenerate from each of the cut surfaces.
In my previous paper on Leptoplana (Child, '04) I suggested
that the nervous stimuli in the region of a cut surface may exercise
either directly or indirectly some influence upon the growth of new
tissue from this region, and, moreover, that after removal of a
part they may even be increased in intensity because of the more
or less ineffectual attempts of the animal to perform the character-
istic movements.
The facts and conclusions cited, together with many others,
such as the cases described by Herbst ('96a, '96b, '99) of the sub-
stitution of an antenna-like organ for an eye in the absence of the
optic ganglion from the eye-stalks of certain decapod Crustacea
Studies on Regulation. 465
and the extensive literature of the interesting, although at present
somewhat confused question of the relation between the nervous
system and the formation and development of the voluntary
muscles (Herbst, '01, Neumann, '01, '03, Goldstein,^ '04) all
afford evidence that there is a relation of some sort between the
nervous system and the formation of certain structures, at least
in some stages of development.
Regarding the nature of this relation various opinions exist.
The question as to the " trophic " influence of the nervous system is
exceedingly obscure; the formative stimuli of Herbst and others
are apparently regarded as entirely distinct from nervous func-
tional stimuli. But that some relation exists between functional
stimuli and the development and continued existence of certain
structures cannot be doubted.
The occurrence of regeneration in plants. Protozoa and other
forms and in stages in which there is no visible diff^erentiation of
the nervous system is of course no argument against the influence
of the nervous system where it is present. For the development
of the nervous system does not add anything to the protoplasm
which is fundamentally diff^erent from what already exists there.
The nervous system is simply a more or less highly diff^erentiated
structure which accomplishes the transference and transforma-
tion of stimuli, but in its absence some method of transference,
however diff^use, must exist m the protoplasm.
The following study of regeneration and other regulative phe-
nomena in relation to the nervous system endeavors to present
certain phases of the problem which seem to me important for the
form considered.
The figures are diagrammatic but are drawn from careful meas-
urements in nearly all cases. In a number of cases the extent of
the intestinal branches is indicated in the figure in a simple
manner, no attempt being made to show the actual course of
branches in particular individuals. The ganglia are drawn,
where present, but the nerve cords are not indicated. The size
of the pharynx is shown as exactly as possible: in most cases the
^ Further references to the literature of this subject may be found in Goldstein's
paper.
466 C. M. Child.
genital ducts are indicated only by the "genital area" posterior
to the pharynx as this was all that could be distinguished with
certainty except when the ducts were filled with sexual products.
B. GENERAL CONSIDERATIONS REGARDING THE RELATION
BETWEEN THE CENTRAL NERVOUS SYSTEM AND
MORPHOGENESIS.
The existence of a relation between the nervous system and both
morphogenesis and the maintenance of form has been established
or regarded as probable in various cases, some of which have
already been mentioned. Some authors postulate the existence
of special nervous "formative stimuli" and "trophic" nervous
stimuli have been much discussed. But the relation between the
nervous system and morphogenesis is of a problematic character,
though the existence of a relation of some sort can scarcely be
denied in many cases.
This relation may conceivably be either direct or indirect. In
the first case particular nervous stimuli of some sort are to be
regarded as constituting in themselves formative factors. In the
second case in consequence of certain nervous stimuli a particular
part may be subjected to certain conditions which may be the
formative factors, though themselves wholly different in character
from nervous stimuli. The conditions connected with and
resulting from a particular functional activity of a motor organ
constitute a good example of the indirect relation. In general the
functional activity of a motor organ is determined and controlled
more or less completely by the nervous stimuli which affect it and
adjoining regions. In consequence of these stimuli it functions
more or less perfectly in a particular manner. The functional
activity subjects the tissues of the part to a great variety of con-
ditions, physical and chemical, external and internal, which,
however, considered as a whole constitute a characteristic com-
plex. Change in the kind or degree of functional activity is of
course accompanied by changes in the complex of functional con-
ditions to which the part is subjected. If these conditions play any
part in the morphogenesis or form-maintenance a relation between
the nervous system and form will appear to exist in such a case, but
Studies on Regulation. 467
upon analysis will be found to be indirect rather than direct. It
is to be remembered, however, that even in cases of this kind a
direct relation may also exist, /. e., the functional nervous stimuli
themselves may conceivably exercise a direct influence of some
sort upon the form.
The question as to whether the complex of functional conditions
exclusive of nervous stimuli may effect form is undoubtedly to be
answered in the affirmative. The existence of a relation between
these conditions and form has been established with more or less
certainty for various cases by Roux and others. Little attempt
has been made, however, to analyze these conditions or analysis
has usually not proved very successful. The best examples of
so-called functional structures are to be found in the tissues con-
nected with movement. In these structures the arrangement of
parts is very closely dependerit upon the conditions resulting from
use of the organs in a characteristic manner.
In most of the Turbellaria as well as in many other forms the
whole body is more or less involved in the characteristic move-
ments and thus becomes in a sense a complex motor organ. It is
not improbable therefore that the various conditions to which the
tissues are subjected in consequence of the characteristic move-
ments are in certain cases important formative factors. I have
already shown that such conditions are concerned in form
regulation in Stenostoma and Leptoplana (Child, '02, '03a, '04).
As will appear, the description in the following section of the rela-
tion between the nervous system and motor activity in Lepto-
plana is a necessary preliminary to the experiments to be presented.
To what extent the functional conditions may constitute forma-
tive factors in cases where motor activity is not concerned is a
problem regarding which the data are at present few. I am in-
clined to believe, however, that we shall find form to be essentially
functional in very many cases where it is not at present so re-
garded. Indeed in one sense all organic form is functional.
Among the conditions resulting from functional activity me-
chanical conditions are important. Their importance has been
recognized in connection with the structure of bone, muscle and
connective tissue, but I think they are important factors in many
468 C. M. Child.
other cases also. The formative effect of these conditions may
conceivably be twofold; they may act as stimuli to growth or other
changes, i. e., they may exert a "trophic" effect as Triepel and
others have pointed out, or they may act in a direct mechanical
manner, bringing about a particular arrangement of material.
Both of these methods of action are important but the second
has been much neglected in the analysis of formative conditions.
The direct mechanical effect of pressure and tension upon the
form of parts is, I believe, of great importance and may afford in
some cases a simple explanation of phenomena which appear
inexplicable from other points of view. A good case in point is
the change of form called by Morgan "morphallaxis" in regu-
lating pieces of Planaria and other Turbellaria. In the case of
Stenostonia I have shown this change to be primarily mechanical
in nature (Child, '02, '03a) and there is no doubt that in other
forms the same factors are effective. In the case of Leptoplana
the effect of mechanical conditions has already been shown in the
preceding paper (Child, '04), and will be further considered in the
present paper.
But in many cases an indirect relation exists between the nerv-
ous system and the mechanical conditions, as in the cases of Sten-
ostoma and Leptoplana above mentioned, since the mechanical
conditions effective here depend upon the use of the parts in a
characteristic manner during locomotion. It is thus easy to see
how factors, simple in themselves and entirely independent of the
nervous system, may apparently stand in relation to it. The same
is of course true with regard to other functional conditions as well
as the mechanical factors.
Even in cases where a direct relation between the nervous system
and form may be shown to exist I see no necessity for assuming
the existence of special "formative stimuli" or "trophic stimuli"
as distinct from the functional stimuli. Moreover, extreme
caution is necessary before concluding that a direct relation exists.
The problem of organic form is undoubtedly the most complex
and difficult of all biological problems. I do not think that the
suggestions made here tend toward its simplification. The factors
of organic form include all the activities of organic substance as
Studies on Regulation. 469
well as the environmental factors in varying degree. Indeed, in
most cases, if not in all, we may regard organic form as the visible
effect upon the protoplasm of functional activity in the widest sense,
occurring in a given environment. But the basis of this functional
activity is to be found in the composition of the protoplasm to-
gether with environmental factors. I believe this distmction be-
tween protoplasmic composition and organic form is important.
In general the composition of the protoplasm determines — not
form but functional activity of some sort, and m consequence of
the internal or external conditions connected with the activity and
produced by it form appears. We may say that morphological
form is the visible expression of protoplasmic activity in a given
environment.
If my experiments succeed in establishing for certain cases cer-
tain definite factors in the complex of conditions upon which form
depends, something has been gained, especially when we con-
sider the vagueness or the anthropomorphic character of many
hypotheses concerning form, and when we remember for instance
that Driesch has made certain aspects of the problem of form the
basis of his theory of the autonomy of the vital processes, while
certain other authors hold that the problem is at present insoluble.
If it has proven insoluble thus far I believe it is because of the
methods employed rather than the nature of the problem.
C. EXPERIMENTAL PART.
I. The Central Nervous System in Relation to Behavior.
The characteristic movements of the normal animal (Child,
'04) are coordinated in such manner that definite characteristic
results are obtained: locomotion in a definite direction is possible
and the motor reactions to various stimuli possess a definite
character. Removal of the cephalic ganglia brings about a
marked chang-e in the character of the movements. Pieces with-
out the cephalic ganglia appear at first glance to be in great degree
incapable of movement. Careful observation of the pieces shows,
however, that they are capable of at least many of the character-
istic movements of the species but that those movements are much
less powerful and lack coordination.
470 C. M. Child.
But another important feature of the movements in the absence
of the cephahc gangha must be noted, viz: that different pieces
differ from each other in the degree of coordination, power, and
frequency of their movements. Pieces from vs^hich the anterior
end has been removed by a cut only a short distance posterior to
the cephaHc gangha are capable of a somev^^hat greater degree of
activity than those from which the anterior half or two-thirds of
the body has been removed. In general it appears that the greater
the remaining portion of the central nervous system the more
complete the activity.
We may consider first the case of a specimen from which the
anterior end has been removed by a transverse cut two or three
millimeters posterior to the cephalic ganglia. Such a piece is
capable of locomotion but the advance is very slow and uniform.
In my account of the normal movements (Child, '04) I called
attention to the fact that locomotion in Leptoplana is accom-
plished both by means of cilia and by muscular contraction, parts
of the margin being extended and attached to the substratum and
then undergoing contraction, thus dragging the body forward.
The muscular factor is especially conspicuous after strong stimu-
lation. In the specimen deprived of the cephalic ganglia, how-
ever, progression is accomplished largely by means of cilia, hence
the slow, uniform, gliding character of the movement. The
specimen is apparently capable of performing all the muscular
movements necessary for muscular locomotion but they appear to
lack perfect coordination. Occasionally the piece seems to suc-
ceed in using its muscles in some degree effectually, but it is
probable that these instances are simply due to chance coincidence
of particular muscular contractions. As the piece is more and
more strongly stimulated the muscular contractions become more
and more violent, although not coordinated, until finally the whole
piece is involved in convulsive movements during which it may
roll up and unroll or twist and squirm about, often turning over
with ventral surface uppermost.
Use of the posterior margins and posterior end of the body as
organs of attachment occurs to some extent in these pieces. As
the piece glides over the substratum parts of these regions can be
Studies on Regulation. 471
seen to attach and free themselves in the characteristic manner,
though here the muscular pi,ay of the margins is much less marked.
The piece as a whole adheres much less closely to the substratum,
however, than the normal animal. It is not at all difficult to
detach these pieces by means of a current of water from a pipette,
while the normal animal adheres so closely that detachment by
this method is often almost impossible.
According to these observations pieces without the cephalic
ganglia show both a quantitative and qualitative difference from
normal animals as regards motor activities. All motor activities
appear to be less intense than under ordinary conditions and the
imperfect coordination in muscular movements alters the charac-
ter of the movements very greatly.
In these pieces the margins of the head, apparently the chief
tactile organs, are of course absent and other parts of the body
are less sensitive than these. Reaction to tactile stimulation of
the lateral and other regions of the body is, however, less intense
and definite than in pieces containing the cephalic ganglia. The
eyes are also absent in these pieces and there is no marked re-
action to light, though in a few cases, I thought I could observe
some slight reaction (compare Parker and Burnett, '00).
Individual differences in the behavior of pieces without ganglia
are often observed even where the cuts removing the head were at
the same level. Some pieces seem capable of more complete
coordination than others, as is clearly seen, for example, by the
rapidity with which they right themselves. These individual
differences are of most frequent occurrence when the cut is not
far from the ganglia and may be due to slight differences in level
of the cut, one piece retaining some parts of the nervous system
absent in others. Occasionally, however, they occur when the
cut was some distance posterior to the ganglia, and in such cases
must probably be ascribed to some structural or physiological
difference of which at present we know little. The fact of the
existence of such differences is however of interest as probably
indicating the existence of marked variations of some kind in the
nervous system.
There seems to be some degree of correlation between size and
472 CM. Child.
the ability to perform coordinated movements in these pieces
deprived of gangha. Of two pieces with anterior ends at the same
level the longer seems to show a slightly greater degree of coordina-
tion; it is sometimes able to advance more rapidly than the shorter
piece and in general appears to be less completely helpless. The
piece which has lost the greater posterior part of its body does not
make up for this loss by greater use of the regenerating part to any
such extent as does the piece with ganglia, but is simply more help-
less than the piece which has lost only a small part. Exceptions are
frequent but I think that a real difference does exist. It is neces-
sary to distinguish two factors here, viz: the power of motor
activity in general, i. e., the power of performing movements of
any kind, and the power of coordinate functional activity. The
smaller pieces usually appear to be more active than the larger
but their activity seems to be less perfectly coordinated and so
less effective as regards locomotion, etc. I am inclined to believe
that the greater activity of the smaller pieces is connected with the
loss of a large part of the body as is the case in similar pieces with
ganglia, while on the other hand the lack of ability to coordinate
is probably due to the small portion of the central nervous system
present.
The question as to whether the pieces deprived of the cephalic
ganglia retain the power of "spontaneous" movement is some-
what difficult to answer, since no sharp distinction can be made
between spontaneous movements so-called and complex series of
movements following particular stimuli; indeed in my opinion no
distinction save one of degree exists. The pieces without cephalic
ganglia are certainly much less active than normal animals, react
more slowly and less strongly to stimuli and, as has been men-
tioned, are unable to a large extent to coordinate their muscular
movements. But even when apparently undisturbed such pieces
are often found moving slowly about and performing indefinite
muscular movements similar in character to those of normal
animals but not correlated. I am inclined to believe that the loss
of the cephalic ganglia means essentially the loss of the connections
with the principal sense organs, i. e.^ the organs for the reception
of stimuli, and the loss of a part of the more or less complex con-
Studies on Regulation. 473
ducting paths. This being the case we should expect to find less
power of reaction to stimuli, less complexity and a lower degree of
correlation in the movements. These are exactly the conditions
that we do find.
It is, I think, desirable to avoid the use of the word "sponta-
neous" in this connection since the difference between sponta-
neous and non-spontaneous movements seems to be merely one
of degree of complexity and correlation or coordination. Re-
moval of the principal paths by which stimuli enter and a part of
the structures which connect these paths with other parts oi the
nervous system must reduce the complexity of structure and con-
sequently of the visible activities dependent upon this structure.
Among these observations the most important point for the
present consideration is the presence of the power of progressive
locomotion in some degree in pieces deprived of the ganglia.
Loeb ('94, '99) found that in the case of Thysanozoon loss of
the power of progressive locomotion resulted from the removal of
the cephalic ganglion. This is certainly not the case in Lepto-
playia, and indeed experiments of my own upon Thysanozoon led
me to the conclusion that even here the pieces without the ganglia
still possessed some slight power of locomotion, though much less
than that of the normal animal. In both Thysanozoon and Lepto-
plana these pieces are capable of righting themselves after being
turned over, but the change in position is much less rapid than in
normal animals and frequently is accomplished only after re-
peated attempts, or in some cases does not succeed at all, and the
piece gradually becomes quiet.
One other point of considerable interest must be considered.
In Leptoplana I observed a marked difference in the power of
locomotion and of coordination in general in pieces cut at different
levels, the activity decreasing as the portion of the body removed
with the cephalic ganglia increased. If, for instance, an individual
was cut transversely two or three millimeters posterior to the
ganglia the posterior piece was much more active and was capable
of more rapid locomotion and more perfect coordination than a
posterior piece obtained by a cut posterior to the middle of the
body. In general the greater the distance between the cut and
474 C. M. Child.
the ganglia the less the activity and the more irregular and im-
perfect are the movements. The difference between a piece ob-
tained by a cut just posterior to the ganglia and one from the
region posterior to the pharynx is striking. The latter scarcely
reacts at all to stimuli, is almost v^holly incapable of progressive
locomotion and rarely succeeds in righting itself — though this
last may be due in part to the fact that such pieces are necessarily
short— while the activity of the former is much greater in all
respects though far below that of the normal animal.
So far as I am aware the case of Dendroccelum mentioned by
Lillie ('oi) is the only one in which observations of this kind have
been made. In Detidroccelum Lillie found that posterior pieces
obtained by section in the anterior third or fourth of the animal
reacted to light like normal animals though more slowly, while
pieces from levels posterior to this did not react. He does not
mention any degree of reactive power corresponding to difference
of level of the cut in pieces capable of reacting, but since other
Turbellaria which I have observed resemble Leptoplana more or
less closely, I am inclined to think that such a difference may
possibly be present, though I have not had the opportunity of
examining Dendrocoelum in sufficient numbers to decide this
point.
The differences in pieces from different levels are not due simply
to differences in size, for a short piece from the region just pos-
terior to the cephalic ganglia is much more active than a piece of
the same size from the posterior region.- It is apparently not
simply the amount of nerve tissue present that determines the
degree of motor activity, but rather the quality of this tissue which
differs in different regions. With our present knowledge of the
nervous system only vague surmises as to the nature of this differ-
ence are possible. It may be, at least in part, a difference in
structural complexity, or the difference in the quantity of energy
transformed by the stimuli or it may be something different from
either of these: as a matter of fact the nerve cords in the Turbel-
laria diminish in size toward the posterior end of the body. But
the fact that a difference exists is important. My observations
also indicate that the case is somewhat similar as regards the ceph-
Studies on Regulation. 475
alic ganglia: in general the motor activity falls further and further
below the normal as the portion of the ganglia removed or injured
increases. It is often difficult with the present technique of
operation upon these forms to determine the extent of injury to
the small ganglia, but notwithstanding this difficulty my observa-
tions indicate very clearly that a relation exists between coordi-
nated motor activity and the amount of ganglionic tissue present.
When the cut passes through the middle of the ganglia both pieces
separated behave essentially like normal animals, but when less
than half of the ganglionic tissue remains intact the piece behaves
much like specimens without ganglia and if the portions of the
ganglia remaining are very small there is almost no motor activity
except the ciliary movement, unless the piece is strongly stimu-
lated. Since the ganglia are small and the difficulty of making a
section in them at exactly the level desired is great, and since it is
often difficult to determine after section just what parts of the
ganglia remain, the results of these experiments are not exact.
But the fact that the two pieces of an individual separated by a
cut through the middle of the ganglia both behave like normal
animals shows that the removal of half of the ganglionic tissue
does not affect the behavior appreciably. Moreover, it makes
no difference in such cases whether the cut is-longitudinal or other-
wise. Anterior and posterior halves and right and left halves of
the ganglia seem to be essentially alike in this respect.
Pieces from the region anterior to the ganglia show almost no
motor activity except that of the cilia, which continue to beat, and
some degree of contraction after strong stimulation. Such pieces
die in the course of two or three days.
These relations between the various regions of the nerve cords
and the cephalic ganglia and coordinated motor activity will be
illustrated in the consideration of individual cases. The fact of
the relation is of interest and indicates, in my opinion, that co-
ordination is connected in these forms rather with a certain extent
and structural complexity than with certain definite organs or
centers. Certainly the cephalic ganglia are more important for
motor activity and coordination than the other portions of the
nervous system, but it is possible that their connection with the
476 C. M. Child.
chief sense organs, z. e., the paths by which more or less definitely
localized stimuli enter the nervous system, is the primary factor
in their predominance.
With regard to the existence of "centers" in the nervous system
I agree essentially M^ith Loeb ('99) and I think the relations above
described support this view. Coordinated movements are the
result of series of interrelations and exist after mutilation in the
degree in which the interrelations remain intact or are reestab-
lished.
The case of Leptoplaria as cited affords strong support to the
view that the difference between "spontaneous" and "non-
spontaneous" motor-activity is simply one of degree. Moreover,
it is impossible to say that one part of the central nervous system
in Leptoplana is necessarily connected with coordinated move-
ment while another is not. It is rather the amount of nervous
tissue — in all probability the completeness of the system of con-
nections of parts — than the presence of any one portion which
determines the results.
2. The Relation Between the Central Nervous System and
Posterior Regeneration.
From Schultz's ('02) account of regeneration in Leptoplana
atomata, it is evident that there is but little difference as regards
regeneration between this species and L. tremellaris, but in the only
case in which Schultz and I are really concerned with the same
problem our interpretations of the facts differ widely.
As regards the limits of regeneration in Leptoplana a brief
preliminary statement will suffice here. Posterior regeneration
from a cut surface is qualitatively complete at all levels posterior
to the cephalic ganglia whether these are present or not and an-
terior regeneration is complete only when the ganglia are present
at least in large part, /. e., only anterior to them. In other words,
regeneration of a head is impossible in the absence of the cephalic
ganglia but posterior regeneration occurs whether they are present
or not. In the absence of food the size of the new part is never as
great as that of the part removed, but this is not of great impor-
tance.
Studies on Regulation.
All
The course of regeneration in the posterior direction from a
level between the cephalic ganglia and the pharynx is illustrated
in Figs. 1-5. On these figures the organs are indicated in a
somewhat diagrammatic manner. The intestine is not drawn in
the old parts, but the general distribution of its branches is indi-
cated in the regenerated parts. Fig, i indicates the level of
the cut and the shape ol the anterior end before section. After
section the cut surface contracts and becomes concave posteriorly,
and within two or three days new unpigmented tissue appears. In
Fig. 2 the condition of the piece ten days after section is indi-
cated. An outgrowth of new tissue tapering posteriorly is present,
into which intestinal branches extend from the old part — and it
may be mentioned in passing that the intestine in regenerated
areas apparently always arises in connection with the old parts
present. In the median line is a small ill-defined area which
represents the developing pharynx. Fig. 3, sixteen days after
section, shows a more advanced condition. The regenerated
area is longer and the pharynx is distinct. From this time on a
marked decrease in size occurs but the old part is much more
affected than the new, as is indicated by Fig. 4 twenty-seven
days after section. Here the new and old parts are of equal
length, the new being longer, though perhaps not containing more
material than in Fig. 3, and the old shorter. The pharynx
has increased in size and beyond it a small clear area, which may
be called the genital area, indicates that regeneration of the genital
ducts is taking place. Fig. 5 shows a stage fifty-one days after
478 CM. Child.
section. The old parts have continued to decrease in size more
rapidly than the new, but otherwise there is little difference be-
tween this and the preceding stage. This piece remained alive
during another month, the change in relative size of old and new
parts continuing, together with reduction in size of the whole.
The history of this piece is typical for posterior regeneration
from this level of the body. Differences in the amount of regen-
eration occur in different individuals, but in all cases regeneration
may be said to be qualitatively complete in that the characteristic
organs of the part removed are regenerated. As the size of the
part removed decreases so in general does the amount of regenera-
tion. The significance of this fact will be discussed more fully
later. When parts of the pharynx or genital ducts are present
the regeneration apparently always begins from the old part, but
when such organs are wholly removed they are formed anew. In
general the level at which regeneration occurs, the presence or
absence of food, and individual differences affect the regeneration
quantitatively but not qualitatively.
a. Experiments on the Relation between the Cephalic Ganglia
and Posterior Regeneration at Various Levels behind the
Head.
Mention was made above of thefact that Leptoplana is not cap-
able in any case of regenerating a head in the absence of the cephalic
ganglia. This apparent dependence of anterior regeneration upon
the cephalic ganglia has been established for a number of forms
but the question as to the relation between the cephalic ganglia
and posterior regeneration in the Turbellaria has received little
attention. In order to examine this problem I prepared series of
pieces as follows: a certain number of specimens of as nearly as
possible the same size were cut at a given level and from half of
these the cerebral ganglia were removed by a transverse cut just
posterior to them; the regeneration of the two sets was then com-
pared at stated intervals with respect to rapidity, amount, and
quality of regeneration and the form of the new part. In several
cases also series prepared for other purposes proved of value in
this connection and could be compared with other pieces cut at
Studies on Regulation.
479
the same level which were not originally intended as controls for
them. These experiments were performed during the winter
when the temperature of the water was much lower than in sum-
mer and the total amount of regeneration in the various cases is
less than in summer experiments. In all cases the pieces were
kept until regeneration ceased, in order that comparison of the
total regeneration might be made.
Series 73. Six pieces, each representing the region of the
body between the cephalic ganglia and the pharynx, were
obtained by the two transverse cuts indicated in Fig. 6.
Series 82. Five pieces were obtained by transverse cuts just
anterior to the pharynx (the lower line in Fig. 6) but the head
and cephalic ganglia were left intact. This series was not origi-
nally intended as a control for Series 73 consequently the inter-
vals between examinations are somewhat different though not
enough to prevent comparison.
Fig. 7 shows the condition of the pieces of Series 73 eighteen
days after section and Fig. 8 the posterior ends of the pieces of
Series 82 fourteen days after section. In Series J^ the contraction
of the cut surface is greater, the new tissue contains fewer intes-
tinal branches, and the amount of the new tissue is somewhat less
than in Series 82.
The different pieces of each series were so closely similar that
these two will serve as examples.
48 o C. M. Child.
In Figs. 9 and lo the condition of the pieces of Series 73
thirty-eight days after section is indicated. In one of the pieces
the new tissue showed the tapering form of Fig. 10, the other
pieces resembhng Fig. 9. The former piece was capable of
more rapid locomotion than the others. In all the pharynx and
genital area are visible and an axial intestine with short branches
extends down the middle of the pieces.
The condition of the pieces of Series 82 thirty-four days after
section is indicated in Fig. 11. Different pieces differed
slightly as regards the length of the new tissue, but other differ-
ences were not observed. Pharynx and genital area were present
and the new tissue was well-filled with intestinal branches.
The pieces of Series 82 differ markedly, however, from those of
Series 73 in that the amount of regeneration is much greater in the
Series 82, where the cephalic ganglia are present. Moreover,
comparison of the Figs. 9 and 10 with Fig. 11 shows that the
pharynx is longer and the intestinal branches much more abun-
dant in Series 82.
After this time there was no further advance in regeneration.
The pieces of both series had already begun to decrease in size
and continued to do so, but the decrease was somewhat more rapid
in Series 82 than in Series 73.
In these two series the differences seem to be wholly quantita-
tive. The pieces in which the cephalic ganglia are intact regen-
erate more rapidly, at least in the later stages; the amount of new
tissue formed is greater; the pharynx is larger; and the intestinal
branches are more numerous.
Series 78. Five specimens were cut transversely through the
middle of the pharynx (Fig. 12) the anterior part with head and
cephalic ganglia intact being used.
Series 79. Five specimens were cut at the same level as in
Series 78 but in these the head was removed by a second cut just
posterior to the ganglia (Fig. 12).
Figs. 13 and 14 indicate the condition of the posterior ends in
the two series fourteen days after section. In Fig. 14 (Series 79,
without cephalic ganglia) the contraction of the cut surface is
greater and the new tissue contains fewer intestinal branches
Studies on Regulation.
481
than in Fig. 13 (Series 78, ganglia present). There is no marked
difference in the amount of regeneration.
Thirty-four days after section the pieces of Series 78 have at-
tained the condition represented in Figs. 15 and 16, and Figs.
17 and 18 represent the condition of Series 79. Here, as in the
preceding case there is a marked difference between the two
series. In the series containing the cephaHc gangha, the amount
of regeneration is greater, the posterior portion of the new pharynx
has regenerated in the new tissue to a much larger extent, and
intestinal branches fill the new tissue much more completely.
Later stages afford no additional features of interest.
J.4
Series 80 and 81. In these two series the level from which pos-
terior regeneration occurred was at the posterior end of the
pharynx (Fig. 19). Series 80 consisted of five pieces with head
and cephalic ganglia intact and Series 81 of five pieces from which
the head and ganglia had been removed (Fig. 19).
Figs. 20 and 21 represent the condition of Series 80 fourteen
days after section and Figs. 22 and 23 of Series 81 after the
same interval. There is little difference between the two series
at this stage except that in Series 80 the intestinal branches have
penetrated further into the new tissue.
482
C. M. Child.
Figs. 24 and 25 show the two extremes of Series 80 thirty-
four days after section and Figs. 26 and 27 those of Series 81
after the same interval. Comparing the two pieces showing least
regeneration in the two series (Figs. 24 and 26) the amount of
regeneration in the piece from Series 80 (Fig. 24) is the greater
(Fig. 26), and a similar difference, though less marked, occurs in
the pieces showing the maximum regeneration (Figs. 25 and 27).
m
22
w
23
Moreover, three pieces of Series 80 were essentially like Fig. 25,
while only one in Series 81 was like Fig. 27. As regards the
intestinal branches all pieces of Series 80 are in advance of Series
81. No further changes occurred in later stages.
Thus in these series as in the others above described the pieces
containing the cephalic ganglia are quantitatively in advance of
those without ganglia.
Studies on Regulation. 483
b. Discussion of the Experiments.
These three pairs of experiments at three different levels of the
body all afford similar results. In all the series containing the
ganglia regeneration is quantitatively more complete than in
those where the ganglia are absent. Moreover the difference is
greatest between Series 73 and 82, is less but still considerable
between Series 78 and 79, and is only slight between Series 80 and
81, /. e., the difference decreases with increasing distance of the
cut surface from the anterior end. And finally, there is in general
a decrease in the absolute amount of regeneration in all series with
approach of the cut surface to the posterior end.
In order to forestall the possible objection that differences in
thickness in the dorso-ventral dimensions of the new tissue have
not been taken into account it should be said that frequent ob-
servations upon this point showed no marked difference in thick-
ness, though usually the new tissue in pieces without ganglia was
not as thick as in the others, probably because the cut surface
undergoes greater contraction dorso-ventrally as well as in other
directions in such pieces without ganglia.
In all three cases the pieces without ganglia are smaller than
the others, since the whole head was removed with the ganglia,
but only in the first case (Series 73 and 82) is the difference in size
very great; here the pieces without ganglia are only about half the
size of the others. As a matter of fact, however, the differences in
size do not appreciably affect the result, as numerous experi-
ments have shown me. Other things being equal a smaller piece
becomes exhausted and dies sooner than a larger piece, but
except in the case of minute pieces both live several months.
Moreover, in the above experiments, the smallest pieces without
ganglia (Series 73, Figs. 9 and 10) show more regeneration than
either of the other series without ganglia (Series 79, Figs. 17 and
18; Series 81, Figs. 26 and 27) and the same is true of the pieces
with ganglia. The long pieces with ganglia of Series 80 (Figs. 24
and 25) show less regeneration than do the small pieces without
ganglia of Series 73 (Figs. 9 and 10). It is evident that the
differences in size of the pieces cannot account for the results of
these experiments.
484 C. M. Child.
Is it then possible that certain "formative stimuli" which affect
posterior regeneration are connected in some manner with the
presence of the cephalic ganglia ? If such exist they certainly do
not concern particular organs for regeneration of all the organs
characteristic of the part removed takes place in the absence of
the ganglia, though these organs are of smaller size or less com-
plex in arrangement than when the ganglia are present, as for
instance the pharynx and intestinal branches. Moreover, the
differences between the two groups differ to a considerable extent
with the region from which regeneration takes place. It is clear
that we cannot suppose that any particular "formative stimuli"
are connected with the presence of the ganglia. We may, how-
ever, take the position that all stimuli to growth are more power-
ful when the ganglia are present and so bring about a greater
amount of regeneration.
But can we proceed a step further and reach any conclusions
as to the character of these stimuli .? I believe that this is
possible, and moreover, that these experiments afford valuable
data for the interpretation of certain regulative processes in these
forms.
But first it is necessary to recall what was said in an earlier
section (pp. 470 and 471) regarding the behavior of the pieces
deprived of the cephalic ganglia. The imperfect coordination of
movement and the less intense and complex activity are the chief
points in which these pieces differ from those in which the ganglia
are present. Locomotion is slow and chiefly ciliary, all movements
are weaker and there is in general much less movement of all kinds
in the absence of the ganglia. These facts are of great impor-
tance for the consideration of this problem as the following para-
graphs will show.
The chief points for consideration are as follows: first, what is
the reason for the difference in regenerative power between pieces
with and those without ganglia ? Second, why is this difference
most marked in the anterior regions of the body .? Third, why
does it appear chiefly during the later stages of regeneration and
only to a slight extent in the earlier .? Fourth, why does the form
of the regenerated part as a whole differ in the two cases ^ Fifth,
Studies on Regulation. 485
why are certain structures formed in the new tissue of smaller
size or of less complexity in the absence of the ganglia ? It is
hoped that the following discussion of these points may throw
some light upon the problem and afford some hints for future
investigation.
The only satisfactory answer to the above questions is to be
found in the differences in functional activity of the parts in the
two pieces. In the pieces with cephalic ganglia, the locomotion
being much more rapid and all movements more intense, the pos-
terior parts including the new tissue are used to a much greater
extent than in the pieces without ganglia. This greater functional
activity comprises many elements, attachment of the margins
and posterior end to the substratum and consequent tension
upon the parts; the constantly changing but characteristic me-
chanical conditions to which the parts are subjected in consequence
of the coordinated muscular activity and the pressure of the
intestinal contents and perhaps of other internal fluids resulting
from the movements; and the motor stimuli which may possibly
influence growth — these are some of the conditions which accom-
pany the motor activity of a given part, many of which in my
opinion may be formative factors. All of these conditions are
present in much greater degree in the pieces with intact ganglia,
and since the posterior part of the body has been removed, the
region of the cut surface and the new tissue as it appears must
be especially affected by them for these parts so far as function
is concerned supply the place of the parts removed. For example
in a case where the greater part of the body has been removed as
in Series 73 and 82, the region of the cut surface and the new tissue
arising from it are used by the animal, or at least the attempt is
made to use them, as the part removed would be used if it were
present. The greater the degree of functional activity affecting
these parts the greater the stimuli to growth and the more power-
ful the mechanical factors which assist in arranging the new
material or perhaps themselves stimulate growth. Thus the first
of the above questions finds its answer in the fact that the func-
tional activity of the regenerating region is greater in the pieces
with ganglia than in those without. I think no one who com-
486 C. M. Child.
pares the behavior of two such pieces can fail to be convinced of
this fact and of its importance in connection with regeneration.
The second question — why the difference in the amount of
regeneration in the two sets of pieces is most marked in the an-
terior regions of the body follows directly from the answer to the
first. It is evident that when a large part of the body has been
removed the regenerating tissue which arises in its place must
be subjected to a much greater degree of functional activity than
when it supplies the place of only a small part. For example if
the whole body posterior to the anterior end of the pharynx be
removed as in Series 73 and 82 the new tissue which grows out
from the cut surface is the functional representative of the part
removed. In locomotion it serves not only as a posterior end for
'attachment but also takes the place of the lateral margins of the
long piece removed. There can be little doubt that in such a
case all conditions correlated with functional activity must be
present in much greater degree than in new tissue which grows
out from a cut surface near the posterior end and which supplies
the place of only a small portion of the body. If these conditions
constitute factors in regeneration posterior regeneration must
decrease in amount with the approach of the cut surface to the
posterior end. The three sets of pieces with ganglia, Series 82
(Fig. 11), Series 78 (Figs. 15 and 16), and Series 80 (Figs. 24 and
25) show this difference in the amount of regeneration very clearly.
In the pieces without ganglia, however (Series 73, Figs. 9 and 10;
Series 79, Figs. 17 and 18; Series 81, Figs. 26 and 27), the differ-
ence is much less marked. In the first series in which the cut
surface was near the anterior end (Figs. 9 and 10) the amount
of regeneration is slightly greater than in the other two series
(Figs. 17 and 18 and 26 and 27) but between these two there is
little difference.
It is necessary here to recall what was said in an earlier section
(pp. 471 and 472) on the relation between size and coordination in
pieces deprived of ganglia. Of two pieces with anterior ends at the
same level the larger piece seems to be slightly less helpless than the
smaller piece though the latter is often more active. In view of
these facts we cannot expect to find any such difference between
Studies on Regulation. 487
the pieces of Series 73 (Figs. 9 and 10) and those of Series 81
(Figs. 26 and 27) as between Series 82 (Fig. 11) and Series 80
(Figs. 24 and 25) in regard to the functional activity of the new
tissue. The greater activity of the smaller pieces without gan-
glia is, to a certain extent, counterbalanced by their more imper-
fect coordination. Numerous comparative observations were
made upon the pieces of the three series in order to determine if
possible whether actual differences in motor activity did occur,
and I concluded that in the pieces of Series 73 (Figs. 9 and 10)
the new parts were somewhat more active than in the other two
series, though the coordination of movements appeared to be
more imperfect. Thus it is evident that these pieces considered
by themselves afford few data of importance for or against the
present view, since the differences in activity are at best slight.
But the fact that the difference in the amount of regeneration at
different levels is much less marked in pieces without ganglia than
in those with ganglia is what might be expected according to the
views above expressed, since the differences in motor activity in
the different cases are certainly much less in the former than in the
latter series.
The consideration of the third question is next in order, viz:
why the difference in the amount of regeneration between pieces
with and those without ganglia appears chiefly during the later
stages of regeneration and only to a slight extent or not at all
during the earlier stages. Comparison of Figs. 8, 13 and 20
and 21 (pieces with ganglia) with Figs. 7, 14 and 22 and 23
(pieces without ganglia) shows that during the first two weeks
the amount of regeneration does not differ greatly in the two sets.
In most cases the amount of new tissue seems to be slightly greater
in the pieces with ganglia, but the difference is not very marked.
It is not possible at present to reach definite conclusions, but
certain probable reasons for this condition suggest themselves.
In the first place it seems probable that the first outgrowth of new
tissue from a cut surface in cases of this kind is determined by
factors different from those which determine the later regeneration.
Indeed it is by no means certain as yet how far this apparent
formation of new tissue is due to actual proliferation and how far
488 CM. Child.
to change in position of cells from the old part. The removal of a
part leaves a more or less widely open wound and the soft tissues
of the body may gradually migrate or flow out in consequence of
altered conditions of surface tension or capillarity, or may be forced
out by internal pressure in consequence of muscular contractions
in the old part. The rounded form of this new tissue suggests the
possibility that surface tension may play a part in its formation.
At the same time it is probable and indeed certain in many cases
that rapid proliferation of the cells near the cut surface does
occur. This multiplication has been ascribed in a general and
somewhat vague manner to the altered conditions at the cut sur-
face, doubtless a correct conclusion as far as it goes. The possi-
bility that altered conditions of surface tension and pressure
resulting from the removal of the part which originally adjoined
the cut surface may themselves bring about multiplication has,
however, received little attention, although various experimenters
have shown that cell division is influenced by changes in these con-
ditions. Various other physical and chemical factors resulting
from the injury may also be concerned in this process, but the point
to which I desire to call especial attention is that the appearance
of new tissue from the cut surface is not primarily a regeneration
of anything in particular, but may be largely a flowing out of the
soft viscid contents of the body in consequence of reduced pres-
sure in this direction or altered conditions of surface tension ac-
companied by a more or less rapid multiplication of cells which is
itself the result of the actual conditions. In short the factors con-
cerned are local and are mostly present whether the parts are used
in a particular manner or not. The relative amounts of trans-
position and of proliferation doubtless differ widely in different
species according to their consistency and reactive capacity.
If we admit that the first appearance of ''new tissue " from the
cut surface is determined by these relatively simple local factors
there is no obvious reason why in a given species the amount of this
new tissue formed from a cut surface at a given level in a given time
should not be approximately the same in different individuals,
whether the ganglia are present or not. The only way in which
the presence or absence of the ganglia might affect the result lies,
Studies on Regulation. 489
so far as I can see, in the frequency of muscular contraction and
the consequent internal pressure upon the regions adjoining the
cut; it is possible that more frequent contraction and movement
might force the tissues out through the wound more rapidly,
though this is not a factor of great importance in any case.
In the cases under consideration at present I think we may
regard the new tissue present two weeks after section as repre-
senting to a large extent this first stage. At this time the new
parts are functional in movement to only a very slight degree and
may be regarded as practically undifferentiated outgrowths. It
is possible that the slightly greater amount of this tissue in the
pieces with ganglia may be due to the more frequent movements
which have so to speak forced more material out through the
wound, or it may be that we have here the beginnings of the differ-
ence which is more marked in later stages. In earlier stages than
those figured the differences were, as might be expected, even less
marked.
For the purpose of analysis we may consider the stage of differ-
entiation as following the first indifferent stage which we have
considered. As a matter of fact the two overlap in varying
degree and manner according to circumstances. The fate of the
new material must be regarded as depending essentially upon its
relations to the old parts, or in the words of Driesch, its fate is a
function of its position, since these relations determine what
stimuli it receives and to what conditions external and internal,
it is subjected. Under ordinary circumstances the new part
supplies functionally the place of the part removed though very
imperfectly at first; or, in other words, the animal or piece attempts
to use it as it would use the part removed if this were present.
I am inclined to believe that this "attempt at use" is an important
factor if not the most important in determining what the new part
shall become. But from the time when the use of the part begins,
or let us say, when functional stimuli reach it, it is subjected to
conditions differing widely from those of the first stage. By these
conditions the "indifferent" material is moulded into its definitive
form and structure. The influence of the conditions connected
with locomotion on the form of regenerating posterior regions has
490 C. M. Child.
been considered by me in several papers (Child, '02, '03a, '04).
It is these same conditions which we have to consider in the pres-
ent case, and the present question becomes from this stage on
identical with the first, which has already been discussed.
According to this view then regeneration in the earlier stages is
about the same in pieces with and those without ganglia because it
depends to a large extent upon local factors connected with the
absence of the part removed and the presence of the cut surface
while in later stages the special functional conditions connected
with the use of the part in a characteristic manner, determined
essentially by its position in relation to the whole, constitute the
important factors in determining both the amount of regeneration
and the structural differentiation.
The fourth question has special reference to the general form or
outline of the regenerated part which is as a rule more slender and
tapering in the pieces with ganglia than in those without, the
difference being greatest when the cut surface is near the anterior
end.
The answer to this question is simple. I believe that the differ-
ence in form is due primarily to the tension in the direction of the
longitudinal axis exerted upon these parts in consequence of the
use of the posterior end during locomotion as an organ of attach-
ment. In Leptoplana both the posterior end and the lateral
margins are employed for attachment to the substratum and are
in consequence subjected to tension as the animal moves forward
holding by one part or another of this region. Brief examination
of creeping specimens is sufficient to show that these conditions
exist. In pieces containing the ganglia the use of these parts in
this manner is much more frequent and the tension is much greater
since locomotion is much more rapid than in pieces without gan-
glia. Undoubtedly these conditions play a part in the arrange-
ment of the physically plastic new material, as was very clearly
shown in the preceding paper of this series (Child, '04) and it is
not improbable that they also serve as stimuli to growth. The
form of the new part in Figs. 11, 15, 16, and especially in Figs.
2-5 (summer experiments) is very evidently a form resulting
from mechanical tension exerted chiefly at the posterior end.
Studies on Regulation. 491
It is difficult to understand otherwise why the margins of this
region are concave instead of convex or straight.
But the form differs according to the level from which regenera-
tion occurs (compare Figs. 11, 15 and 16, 24 and 25). The
tapering outline with concave margins is most marked when
regeneration takes place from a level near the anterior end. This
difference is of course connected with the fact that the amount or
regeneration is greater in anterior regions but that difference does
not explain why the new tail should be more slender in the one
case than in the other. Evidently this difference in form of the
new part at different levels is mechanical. In the first place,
when the regenerating part represents a large part or all of the
region of the body used for attachment it performs the functions
of that part and is subjected to the tensions resulting from this
function. When, however, as in Figs. 24 and 25, it represents
only the posterior region of the part used for attachment the lateral
margins of the old portion anterior to it perform in large degree
the function of attachment and hence the regenerating part is sub-
jected only to a relatively slig-ht degree of tension. Therefore, it
is less elongated, less concave laterally and more blunt posteriorly.
In the pieces without ganglia these differences are naturally
much less marked since the tension upon the parts resulting from
locomotion is relatively slight in all cases. But even here there is
a difference, at least between Series 73 and the others (compare
Figs. 9 and 10, Series J^, with Figs. 17 and 18, Series 79, and
Figs. 26 and 27, Series 81). The peculiar form of the regenerating
part in Fig. 9 seems to be due to the fact that this piece used one
part of the margin as often as another for attachment, whereas
usually the median posterior region is used more than other parts.
The form of the regenerating part shown in this figure is common
in such pieces and may, I think, be regarded as due primarily to
lack of coordination; any part of the margin which happens to
come into close contact with the substratum or which is stimulated
in any other way becomes attached and consequently the part
does not taper posteriorly. The piece shown in Fig. 10 was
able to progress more rapidly and in general showed a higher
degree of coordination than the other. The difference in the
492 CM. Child.
form of the new part in these two pieces corresponds very closely
with the differences in its use.
The last of the points to be considered in connection with these
pieces is that which concerns the size and complexity of the char-
acteristic organs of the regenerated part, especially the pharynx
and the intestine. In the Series Ji, and 82, in which regeneration
occurred from a level near the anterior end, the pharynx is much
larger and the intestinal branches are much more fully regenerated
in the pieces with ganglia (Fig. 11) than in these without (Figs.
9 and 10).
In the second pair of series, in which regeneration occurred
from a level near the middle of the body, the regeneration of the
posterior end of the pharynx is much more complete in the pieces
with ganglia (Figs. 15 and 16) than in those without (Figs. 17 and
18), and here again we find a difference in the extent of the
intestinal branches similar to that in the preceding sets.
In the third pair of series, in which regeneration occurred from
a level at the posterior end of the pharynx, regeneration of the
pharynx does not take place but there is the same difference in the
number and extent of the intestinal branches that has been ob-
served in the other cases (compare Figs. 24 and 25 with Figs.
26 and 27).
As regards the pharynx the difference in size in the two sets is
doubtless an expression of the proportionality characteristic of
regenerating parts; the greater the amount of regeneration the
larger the pharynx. But why does such a proportionality exist ^
This is a difficult question to answer and Driesch has even as-
serted that it cannot be answered in physico-chemical terms. It
is true that we are at present unable to analyze the factors con-
cerned, but I see no reason for assuming an autonomistic or
vitalistic principle. The conditions of the case seem to me to be
somewhat as follows : The regenerating part represents a certain
region of the body and its relations to the old part determine that
it shall function in a characteristic manner, /. e., in the manner of
the part which it represents. Consequently the area affected by
particular functional stimuli will be more or less nearly propor-
tional to the size of the regenerated portion.
Studies on Regulation. • 493
Or we may put the case in a somewhat different form : admitting
that the total amount of regeneration is proportional to the degree
in which the conditions for regeneration are present, each part of
the regenerating portion attains a certain size and form which
represent in a way the proportionality between the stimuli affect-
ing it and those affecting other parts. If, as I have suggested,
these stimuli are at least in part the functional stimuli, we may say
that the size of the pharynx in a regenerating part is dependent
upon the size of the part and upon the relation between functional
conditions affecting the pharynx and those affecting other regions.
In short, the difference in size of the pharynx in the pieces with
and those without ganglia is simply a particular expression of the
factors which determine the difference in the amount of regenera-
tion in the two cases, viz: the functional conditions. If, as
Bardeen ('01, '03) supposes, the pressure of the intestinal contents
is a factor in the formation of the pharynx, it is evident that this
factor must be present in much greater degree in the pieces with
ganglia since the movements and muscular contractions of the
various parts which force the intestinal contents out of the
branches toward the central parts are much more frequent and
intense in these pieces than in the others. At any rate it is evident
that the regenerating part is functionally active in much greater
degree in the presence of the ganglia, and since I believe that the
various conditions connected with functional activity are impor-
tant factors in the production and maintenance of organic form,
the difference in size of the regenerated pharynx in the two sets is
in full agreement with the other facts already discussed.
The extent to which the intestinal branches regenerate also
differs in the pieces under consideration. Examination of the
figures will show that in every case, even in the earlier stages, the
regeneration of the intestinal branches is more advanced in the
pieces with ganglia. Compare Figs. 8 and 1 1 (ganglia present)
with Figs. 7, 9 and 10 (ganglia absent). Figs. 13, 15 and 16
(ganglia present) with Figs. 14, 17 and 18 (ganglia absent), and
Figs. 20, 21, 24, 25 (ganglia present) with Figs. 22, 23, 26, 27
(ganglia absent).
The intestinal branches appear to rise in all cases from the cut
494 C. M. Child.
ends of parts of the old intestine and extend from these regions
into the new tissue. In another polyclad I have obtained very
strong experimental evidence w^hich I hope to present at another
time in favor of the view that the pressure of the intestinal contents
upon the walls of the intestine is a factor of great importance in
the formation of intestinal branches or the growth of the intes-
tine into new parts. The differences in the degree of intestinal
regeneration in the pieces now under consideration lend strong
support to this interpretation. The difference is not merely in
proportion to the difference in size, but in the pieces without
ganglia the regeneration of the intestine is relatively less complete
than in the others. In consequence of the more powerful and
frequent muscular contractions of the pieces with ganglia the
intestinal contents are pressed against the closed cut ends of the
intestinal branches, or, in later stages, into the extensions of these
in the new parts with greater force and frequency than in the
pieces without ganglia. It is not difficult to observe this differ-
ence. The intestinal branches in the new tissue are much more
frequently distended by the contents of the intestine in the pieces
with ganglia. Thus in regard to these structures as well as those
already discussed the differences in the tissues are, I think, ex-
pressions of the differences in functional activity, and especially
in this case, of muscular activity. The formative conditions in
this case, however, are themselves mechanical.
As regards the size of the regenerated genital area posterior to
the pharynx there are no marked differences in the two sets. Its
size remains almost the same in all cases whether the total amount
of regeneration is large or small. This uniformity in size is to be
interpreted as indicating that conditions which give rise to these
organs are wholly or largely independent of the muscular activity.
We know little regarding the conditions which may be concerned
in the regeneration of these parts, but they are doubtless connected
with the presence of other portions of the ducts in the old part and
also with certain obscure physiological conditions on which the
presence and periodical activity of the sexual organs is also de-
pendent. The period at which these experiments were performed
was not that of greatest sexual activity; and the organs remained
Studies on Regulation.
49 S
in a more or less rudimentary condition. In a considerable
number of other cases, however, animals were used for experi-
ment during the height of the breeding season and the results in
these cases are interesting.
Fig. 28 represents diagrammatically the course of the genital
ducts; the female ducts are indicated in broken lines. The
496 C. M. Child.
regeneration of many specimens in which these ducts were of large
size and filled with the sexual products was observed and it was
found that the regeneration of these structures was as complete in
the absence of the ganglia as when they were present, though
usually somewhat less rapid. The following case will serve as an
example: A specimen in the height of sexual activity was cut as
in Fig. 28 through the anterior end of the pharynx and again
somewhat posterior to the middle of the pharynx, the piece be-
tween the two cuts being used for experiment. It will be observed
that this piece contained only the anterior portion of the vasa
deferentia and a part of the uterus. In Fig. 29 the condition
of the piece seventeen days after section is indicated. At this
time the ducts were extending into the new tissue but the amount
of their regeneration was not clearly visible. Sixty-five days after
section the piece presented the appearance of Fig. 30. Com-
plete regeneration of the ducts and the terminal organs had taken
place and the ducts were widely distended by sexual products.
Similar results were obtained in other cases where the animals
used were in a condition of sexual activity. The piece just
described was without ganglia, but in other pieces with ganglia
the result was the same, though it occurred in a somewhat shorter
time.
When nearly all of the body was removed from breeding speci-
mens, 7. e., when regeneration occurred from a level only slightly
posterior to the ganglia, nothing more than a small genital area,
as in Figs. 10 or 11, was ever regenerated even though the
ganglia were present. There can be little doubt, I think, that the
complete regeneration of the ducts and terminal organs of large
size in the case just described was due first to the fact that parts of
the ducts remained in the old tissue, and second, and probably
chiefly, to the fact that the ducts were in an active functional con-
dition, I. e.^ filled with sexual products. Here again, as in the
case of the intestine, the possible effect of the pressure of the con-
tents of the ducts upon the walls and of other functional condi-
tions upon regeneration must be taken into account. In my
opinion these are important factors. In the cases where all parts
of the ducts were removed and thus the question as to the effect
Studies on Regulation. 497
of the contents was eliminated the whole apparatus was regener-
ated in small size and more or less rudimentary form as above
stated. But in these experiments the physiological condition of
the regenerating specimens as regards sexual activity was the
same as in the cases where complete regeneration of the genital
apparatus occurred. The only difference is that in the one case
all parts of the ducts and all or nearly all of the gonads were
removed and there was no possibility of the extreme functional
activity of the ducts until new gonads were formed and matured,
a process which did not occur in the specimens without food.
In many of my earliest experiments (August) no genital area
was observed in the regenerated specimens, though it may have
been present in some cases. But later in the autumn, as the breed-
ing period approached its height, the genital area appeared in all
cases. This difference indicates that there was some physiologi-
cal difference between the specimens at different seasons, a differ-
ence undoubtedly correlated in some way with the reproductive
organs.
In the case shown in Figs. 28-30 it is interesting to note that
regeneration of the ducts apparently occurred partly within the
old tissue and partly within the new. During the course of
regeneration the cut end of the pharynx was gradually retracted
from the cut surface, leaving a space behind it in which the copu-
latory organ appeared. This contraction of the pharynx is
probably essentially atrophy from disuse; it occurs only in pieces
without ganglia. The region posterior to it is filled in either by
cells which have migrated or flowed in from the sides or by pro-
liferated cells or probably by both, and in this the copulatory
organ appears. The posterior vasa deferentia extend a consider-
able distance anteriorly into the old part before they unite with the
anterior ducts and the whole system of terminal organs is situ-
ated much further anteriorly than originally (compare Figs. 28
and 30). The large size of the regenerated organs is undoubtedly
due to the large quantity of the sexual products contained in them
and entering them during the course of the experiments. The
retraction of the pharynx afforded space for growth, hence the
position and extent of the organs.
498 C. M. Child.
But the primary object in introducing this case was to show how
completely these organs can regenerate in the absence of the gan-
glia. So far as can be seen they are as perfect as in pieces con-
taining the ganglia.
The conclusions reached from these experiments may be
summed up as follows: the influence of the cephalic ganglia upon
posterior regeneration is not "formative" for the same organs
are regenerated whether the ganglia are present or absent. The
amount of regeneration and the size and extent of various organs
are, however, greater in the presence of the ganglia. This
quantitative diff^erence is probably due to the fact that all func-
tional stimuli and conditions connected with muscular activity
and especially those connected with the coordinated muscular
activity of locomotion occur with much greater frequency and
intensity when the ganglia are present. Probably other func-
tional conditions whose existence is less easily determined are also
present in greater degree when the ganglia are present.
c. Posterior Regeneration in the Absence of the Cephalic Gan-
glia and Parts of the Longitudinal Nerve Cords.
The object of these experiments was to compare the posterior
regeneration at a given level in pieces with anterior ends at different
levels posterior to the ganglia, in order to determine whether the
removal of a considerable part of the nerve cords in addition to
the ganglia had any effect on regeneration. In the section on the
nervous system and behavior mention was made of the fact that
pieces from the posterior region of the body show less motor
activity than pieces of the same size from the anterior regions,
posterior to the cephalic ganglia. These facts would seem to
indicate that the anterior regions of the longitudinal nerve cords i
differ in some respect — perhaps in complexity — from the posterior
regions.
In order to test still further the hypothesis of the relation
between posterior regeneration and motor activity pieces with
posterior ends at the same level of the body and anterior ends at
different levels posterior to the ganglia were used. Certain diffi-
culties were encountered in obtaining definite results from these
Studies on Regulation. 499
experiments. If the level of the posterior ends of these pieces was
posterior to the pharynx the amount of regeneration was not very
great in any case, and differences were comparatively slight,
though apparently in agreement with the hypothesis, i. e., the
shorter pieces, those from which larger portions of the nerve cord
had been removed, seemed to show somewhat less regeneration
than the others, though it was often difficult to be certain that
chance individual differences were not concerned. The reasons
for the small amount of posterior regeneration from levels near
the posterior end of the body have been discussed in the preceding
section. It was also found that small pieces from the posterior
regions of the body live at most only a few weeks after section,
and for some time before death are much contracted and show
scarcely any motor activity. The contracted condition of these
pieces renders exact measurements impossible.
In order to obtain results at all satisfactory it was necessary to
use pieces whose posterior ends were somewhat anterior to the
posterior end of the pharynx. With such pieces only the anterior
half of the body could be examined in this way, but we are justi-
fied in concluding that the same relations between functional
activity and regeneration exist in the different regions of the body.
Two series of these experiments afforded definite results. In
one of these (Series 69) five pieces were cut as in Fig. 28, the
posterior ends being somewhat posterior to the middle of the
pharynx and the anterior ends at the anterior end of the pharynx.
In the other series (Series 71) five pieces were cut with posterior
ends at the same level but with anterior ends just posterior to the
cephalic ganglia (see the dotted line in Fig. 28). The two
series differ in respect to the region between the cephalic ganglia
and pharynx, which is present in one series and absent in the
other. Some difference in the motor activity of the two series
was noted, the pieces of Series 71 being considerably more active
and more successful in locomotion : the differences were, however,
not very great. Figs. 31 and 32 represent the two extremes
found among the pieces of Series 69 seventeen days after section
and Fig. 33 shows the condition of the pieces of Series 71 at the
same time, there being little difference among these. In general
500 C. M. Child.
the pieces of Series 71 show sHghtly more regeneration than the
others, though the difference is not great.
Thirty-two days after section the pieces of Series 69 had
attained the condition of Fig. 34 and those of Series 71 the con-
dition of Fig. 35, the amount being somewhat greater in the
latter series. After this time no marked alteration in propor-
tions occurred.
The results obtained from these pieces are not very striking but
according to the hypothesis we cannot expect anything more than
slight differences in such pieces. Such differences would scarcely
attract attention of themselves but when we consider them in
connection with the experiments described in the preceding sec-
tion their significance is apparent. That they are real is well
shown by the fact that in certain series which were used for other
purposes but which happened to be of such a kind that they could
be used in this connection my notes show a difference in the
w
33 ^-^
amount of regeneration similar to that just described, though
when the experiments were made the significance of such a differ-
ence had not occurred to me and the difference did not attract my
attention until examination of my notes and drawings was made
several months later.
These results, like those described in previous sections, show a
close correlation between the regeneration and the motor activity
of the pieces, and I believe that in both cases the conditions con-
nected with the functional activity of the parts are the factors
which determine the differences in regeneration.
A comparison of Fig. 35 with Figs. 17 and 18 (Series 79) shows
that the amount of regeneration in that piece is greater than
in these, although the level from which regeneration occurred is
somewhat farther posterior (compare Figs. 12 and 28). If the
conclusions reached above are correct less rather than more
regeneration should be expected in Series 71 than in Series 79.
I believe, however, that a difference in temperature is chiefly
Studies on Regulation. 501
concerned in this difference: in the first place Series 71 was begun
October 29 and Series 79 January 9. At the first date the tem-
perature of the water was much higher than on the second and all
specimens were much more active. The regenerative power in
pieces with ganglia differs markedly with season, /. <?., with tem-
perature, as is shown by comparison of Figs. 4 and 5 and 40
and 46, both of which are summer experiments, with Fig. 11, a
winter experiment, from nearly the same level. Pieces without
ganglia may be expected to show similar though less marked
differences.
It is possible that another factor is also concerned: in Series 71
the anterior cut was made as near the ganglia as possible (Fig. 28,
dotted line), while in Series 79 it was somewhat further posterior
(Fig. 12), the chief object being in this case to re 