BIOLOGICAL BULLETIN

OF THE

flDarine Biological laboratory

WOODS HOLE, MASS.

Editorial Staff

E. G. CONKLIN — Princeton University,

JACQUES LOEB — The Rockefeller Institute for Medical Research

T. H. MORGAN — Columbia University,

W. M. WHEELER — Harvard University,

E. B. WILSON — Columbia University.

flDanasino BMtor

FRANK R. LILLIE — The University of Chicago.

VOLUME XX.

WOODS HOLE, MASS.
DECEMBER, 1910, TO MAY, 1911

PRESS or

THC New ERA PRINTING COMPANV
LANCASTER. PA.

CONTENTS OF VOLUME XX.

No. i. DECEMBER, 1910.

CHILD, C. M. Further Experiments on Adventitious Reproduction

and Polarity in Harenactis I

PEEBLES, FLORENCE. On the Interchange of the Limbs of the Chick

by Transplantations 14 \f

BROWNE, ETHEL N. The Relation Between Chromosome Number

and Species in Notonecta 19

GOODALE, H. D. Some Results of Castration in Ducks *; . . 35

TOWER, WM. L. The Determination of Dominance and the Modifi-
cation of Behavior in Alternative (Mendelian) Inheritance, by
Conditions Surrounding or Incident upon the Germ Cells at
Fertilization 67

No. 2. JANUARY, 1911.

MONTGOMERY, THOS. H., JR. Certain Habits, Particularly Light Re-
actions, of a Littoral Aranead 71 N

ISELY, F. B. Preliminary Note on the Ecology of the Early Juvenile

Life of the Unionida1 77

GARY, LEWIS R. A Study of Pedal Laceration in Actinians 81 S

STEVENS, N.M. Further Studies on Heterochromo somes in Mosquitoes. 109 -
STEVENS, N. M. Preliminary Note on Heterochromosomes in the

Guinea Pig 121 /

No. 3. FEBRUARY, 1911.

RICHARDS, A. The Method of Cell Division in the Development of

the Female Sex Organs of Moniezia 123

COKER, ROBERT E., AND SURBER. THADDEUS. A Note on the Meta-
morphosis of the Mussel Lampsilis lavissimus 179 S

MclNDOO, NORMAN E. Notes on Some Arachnids from Ohio Valley

Caves 183 v

HARGITT, CHARLES VV. A Further Note on Keratosum complexum i87\/

No. 4. MARCH, 1911.

SMITH, BERTRAM G. The Nests and Larva of Necturus 191

DANFORTH, C. H. A 74 mm. Polyodon 201

KING, HELEN DEAN. Studies on SexDetermination in Amphibians, IV 205 ^

iii

IV CONTENTS OF VOLUME XX.

No. 5. APRIL, 1911.

HEGNER, ROBERT YV. Experiments with Chrysomelid Beetles. III.

The Effects of Killing Parts of the Eggs of Leptinotarsa decem-

lineata 237 *•

SCOTT, JOHN \V. Further Experiments on the Methods of Egg-laying

in Amphitrite 252 •

KEPNER, WILLIAM A. Nematocysts of Microstoma 266

TANQUARY, MAURICE C. Experiments on the Adoption of Lasius,

Formica and Polyergus Queens by Colonies of Alien Species. . . . 282-

No. 6. MAY, 1911.

CHILD, C. M. Experimental Control of Morphogenesis in the Regu-
lation of Planaria 309

COCHRAN, M. ETHEL. The Biology of the Red-backed Salamander

(Plethodon cinereus erythronotus Green) 332

OSBORN, HENRY L. On the Distribution and Mode of Occurrence in
the United States and Canada of Clinostomitm marginatum, a
Trematode Parasite in Fish, Frogs and Birds 350

COCKERELL, T. D. A. The Scales of Freshwater Fishes 367

Vol. XX. December, 1910. No. i

BIOLOGICAL BULLETIN

FURTHER EXPERIMENTS ON ADVENTITIOUS

REPRODUCTION AND POLARITY IN

HARENACTIS.

C. M. CHILD.
WITH ELEVEN FIGURES.

INTRODUCTORY.

In a paper which appeared in IQO91 I described the adventitious
formation of new axes in the Calif ornian actinian, Harenactis
attenuate,. These new axes have thus far been observed only
under certain conditions, viz., in isolated pieces prepared in such
manner that the oral cut surface of the body wall unites with
the aboral about the whole circumference. In my earlier paper
such pieces were designated as "rings."

During the past summer I again enjoyed the privileges of the
laboratory at La Jolla and take this opportunity of expressing
my thanks to the director, Professor Ritter, and to the trustees
of the San Diego Marine Biological Association. During my
stay at La Jolla I obtained results with "rings" from Harenactis
which supplement my earlier work and extend and confirm my
conclusions. These results form the subject of the present paper.

As described in my earlier paper, the rings are produced by
isolating rather short, cylindrical pieces from the body of Harenac-
tis by means of transverse cuts (Fig. i), and then removing more
or less completely from these pieces the mesenteries and mesen-
terial muscles. The region of the body proximal to the esophagus
is the most suitable for such operations, since here the mesenteries
are well developed, bear large retractor muscles and extend into
the enteron with free borders instead of being attached to the

JCh Id, 'ogb, "Factors of Form Regulation in Harenactis attenuata, III., Regu-
lation in 'Rings,' " Journ. Exf>. Zoo/., VII 2, 1909.

I

C. M. CHILD.

esophagus, as in the more distal regions, or being only slightly
developed, as in the more proximal, attenuated region.

The removal of the free borders brings about longitudinal
contraction of the injured mesenteries and so approximates the
oral and aboral cut surfaces of the body wall. This contraction

4

FIGS. 1-4.

is merely a special case of the wound contraction so characteristic
of the actinians, which I have described for Cerianthus and
Harenactis.1 In my earlier experiments I removed merely such

JChild, '03, "Form Regulation in Cerianthus, I., The Typical Course of Re-
generation," BIOL. BULL., V., 5, 1903.

'04, "Form Regulation in Cerianthus III., The Initiation of Regenera-
tion," BIOL. BULL., VI., 2, 1904.

'08, "Form Regulation in Cerianthus czstuari," BIOL. BULL., XV., i,
1908.

ADVENTITIOUS REPRODUCTION IN HARENACTIS. 3

portions of the mesenteries as protruded from the cut ends of
the pieces, or accomplished the same result by other means.1
In the experiments of this summer greater care was taken to
make certain that all the muscles were removed, and the free
borders of the mesenteries — in most cases the greater part of
the mesentery as well — were cut away. After these operations
the pieces consisted of the body wall with only the bases of the
mesenteries, i. e., those portions which lie nearest the line of
attachment to the body wall. The nature of this operation is
indicated approximately in the diagrammatic Fig. 2. This repre-
sents a typical longitudinal section of the body wall of the
cylindrical piece: all the internal mesenterial structures between
the two broken lines are removed, leaving only narrow strips of
mesenterial tissue next to the body wall.

Such pieces form rings in almost every case, and nearly all of
the rings give rise to tentacle groups. The result of the removal
of the mesenteries is the approximation and contact of the oral
and aboral cut ends of the body wall, as indicated in the optical
section, Fig. 3. Union occurs between these cut surfaces and
the body wall thus forms a closed ring with an opening through
the center and with no connection between the enteron and the
exterior. The enteric cavity becomes more or less distended
with water after closure, undoubtedly in consequence of the
passage of water through the body wall, and sooner or later a
peculiar rotation of the parts about a circular axis situated in
the enteric cavity of the ring occurs, as described in my earlier
paper ('096, pp. 356-357). The result of this rotation is a change
in the position of the line of union between the oral and aboral
ends from its original position around the central opening to the
outer surface of the ring (cf. Figs. 3 and 4 of the present paper;
in Fig. 4 the region of union between oral and aboral ends is
indicated by the very thin body wall which is formed of new
tissue). Very commonly the line of union, which is marked by
the formation of more or less new tissue, finally comes to occupy
a position somewhat upon the upper surface of the ring as this

'oga, "Factors of Form Regulation in HarenaC'is attenuata, I., Wound
Reaction and Restitution in General and the Regional Factors in Oral Restitution,"
Journ. Exp. Zoo/., VI., 4, 1909.

, 'ogb, Figs. 2-5, Fig. 31, also pp. 373~37-4-

4 C. M. CHILD.

lies on the bottom of the dish (Fig. 4). If the ring is turned over,
a new rotation often occurs, which brings the line of union into
approximately the same relative position. As stated in my earlier
paper ('096, p. 357), this appears to be an "attempt" at orienta-
tion with respect to the substratum, but since the original distal
and proximal ends of the piece are united with each other and
since, as is evident from the later processes in these pieces, the
original polarity is decreased or almost eliminated by this union,
any complete orientation is of course impossible.

Most of these rings give rise in the course of several weeks to
groups of tentacles or single tentacles which lie along the line of
union, but which may arise wholly on one side or the other, i. e.,
wholly from the original distal or the original proximal parts,
or may involve both in the formation of a single tentacle group.

In some cases these groups of tentacles are without any close
resemblance to the normal disc of Harenactis and they often
show a more or less marked bilateral or biradial symmetry with
respect to the line of union (Child, '09^, Fig. 10). More com-
monly, however, each group shows a more or less distinct radial
symmetry, similar except as regards the number of radii, to the
symmetry of the animal in nature (Child, 09^, Figs. 12, 15,
I7~3°)- In these radially symmetrical tentacle groups mesen-
teries could in many cases be seen distinctly upon the small
discs, their arrangement corresponding with the intervals be-
tween the tentacles. In some cases these were clearly continua-
tions of the old mesenteries (Child, '09^, Figs. 18-20), but in
other cases I was unable to find any such connection and was
uncertain whether the formation of new mesenteries had occurred
or not.

In my earlier experiments the largest number of tentacles in a
radially symmetrical group was seven ('096, Fig. 29^) and in no
case were mouth and esophagus present, though in one ring
several groups showed a tentacle-like outgrowth in the center of
the disc, where the mouth would normally appear. I suggested
that this outgrowth might possibly represent an everted esopha-
gus, but could not be certain as to its nature.

These tentacle groups were interpreted as representing new
polarities, new body-axes, developing adventitiously in the region

ADVENTITIOUS REPRODUCTION IN HARENACTIS. 5

of union of the original distal and proximal ends of the piece,
but failing to attain the usual form because of the peculiar con-
ditions under which they arose.

In returning to these experiments this summer, I hoped to
obtain tentacle groups which should approach more closely the
usual form of Harenactis and thus leave no doubt that these
outgrowths really represent new polarities, in short, that these
phenomena in rings are really a form of asexual reproduction related
to certain cases of the formation of adventitious structures in
plants. These hopes were fully realized in one case, to be de-
scribed below, and other tentacle groups among the large number
obtained showed various points of interest.

DESCRIPTION OF EXPERIMENTS.

The case of greatest interest is one in \vhich only one group of
tentacles appeared upon the whole ring. This group was situated
wholly upon one side — probably the oral — of the line of union,
was radially symmetrical from the beginning and at first pos-
sessed seven or eight tentacles. During its growth, howrever,
new tentacles appeared until the number reached sixteen. The
tentacles were regularly arranged about a disc similar in appear-
ance to that of the normal animal and in the center of the disc
a mouth, opening into a short esophagus, appeared. On the
disc sixteen mesenteries could be counted, most or perhaps all
of these being new, as was evident both from their appearance
and relation to other parts and from the impossibility that sixteen
of the twenty-four old mesenteries should be involved in this
new development from a small area localized on one side of the
ring. Fig. 5 shows this new disc with the sixteen tentacles as it
appeared twenty days after the operation. The line of union
between the original oral and aboral ends of the piece is indicated
by the dotted line. This new individual differs from the full
grown animal in nature in the smaller number and unequal length
of the tentacles and in the absence of a proximal end, but there
can be no doubt that it represents a new body-axis, a new polarity.

After the formation of the disc it gradually became elevated
above the surface of the ring by the growth of a cylindrical
column beneath it. Fig. 6 is a diagrammatic optical section

O C. M. CHILD.

through the side of the ring on which the disc appeared; it shows
the formation of the column beneath the disc. After twenty-five
days the old tissues of the ring began to show signs of disintegra-
tion and the new individual was either poisoned or infected by
the dead and putrefying tissue, and died. In no case thus far
has it been possible to keep these outgrowths of the rings per-
manently alive, for the parts not involved in the new outgrowths
always degenerate sooner or later and finally kill the new parts.
In the case just described it was my intention to separate the
new individual from the ring in order that it might if possible
complete its development, but death occurred before this was
done. If the opportunity of still further experiment arises, I

5

FIGS. 5-6.

shall try the experiment of separating some of these adventitious
buds, for they are such, from other parts of the ring. I have no
doubt that under proper conditions they can be kept alive and
fed so that development will proceed.

This case as it stands is, however, sufficient to demonstrate
that new polarities may arise from these rings, from which the
old polarity is in large measure eliminated. As was pointed out
in my earlier paper, this is a form of reproduction which resembles
more or less closely certain cases of the formation of adventitious
structures in plants. Moreover, there can be no doubt that
each new polarity arises in relation to local conditions in the
growing tissues along the line of union between the original oral
and aboral ends of the piece, though exactly what these condi-
tions are, we do not at present know.

In the course of my experiments a large number of other
adventitious tentacle groups was obtained, many of them with

ADVENTITIOUS REPRODUCTION IN HARENACTIS. 7

new mesenteries, but none with so large a number of tentacles
as in the above case and none with mouth and esophagus. Be-
sides the distinctly radially symmetrical groups, others, showing
the most various degrees of radial and bilateral symmetry and
asymmetry, appeared. Brief descriptions and figures of a few
characteristic cases are added here.

Fig. 7 shows a case of considerable interest: here a disc with
four radially arranged tentacles and with four well-developed
mesenteries, but no mouth, appeared on one side of the line of
union and was gradually elevated by the growth of a long column

FIGS. 7-11.

beneath it. No further tentacles were added during the life of
this specimen and it finally died, apparently from the same causes
as others.

C. M. CHILD.

In Fig. 8, a ring with three tentacle-bearing outgrowths and
one without tentacles is shown. The tentacles on two of the
three outgrowths are radially symmetrical, though unequal in
length and the discs show well-developed, radially arranged
mesenteries, but are without mouths. The third group is less
symmetrical and consists of only four tentacles. All the out-
growths appeared on one side, apparently the oral, of the line
of union.

Fig. 9 shows a case, viewed obliquely from one side, in which
two groups of four tentacles each appeared, one on either side
of the line of union. This is the only case observed thus far in
which distinct groups appeared on different sides of the line of
union in a single ring and it is figured on that account.

A case with seven distinct tentacle groups is shown in Fig. 10.
This is the largest number of distinct groups observed in any
single case. Apparently the formation of tentacle groups is in
general closely connected with the growth of new tissue in the
region of union, for they appear to be localized at points where
the areas of new tissue are greatest. At some points the two
cut edges may unite with scarcely any new tissue between them,
and in such regions tentacles are less likely to appear than in
regions with more new tissue. In the case shown in Fig. 10 an
unusually large amount of new tissue appeared as a more or less
continuous band about most of the circumference and in this
the tentacle groups developed. The area of the new tissue is
approximately indicated in the figure by the two dotted lines.

And finally, Fig. II shows a case in which the symmetry of
the groups is mostly bilateral rather than radial. Here the cor-
responding parts of each group arise on opposite sides of the
line of union. I believe that groups of this kind appear when the
old organization is less injured and the old polarity less com-
pletely eliminated than in other cases. Although tentacles ap-
pear in such cases on the aboral as well as on the oral side of the
line of union, nevertheless the development of such rows of ten-
tacles is manifestly a less extreme change from the "normal"
arrangement with respect to the cut surface of the body wall
than the formation of distinct, radially symmetrical groups on
one side or the other of the line of union. It will be observed

ADVENTITIOUS REPRODUCTION IN HARENACTIS. 9

from the figure that in two cases the tentacles which apparently
arise directly upon the line of union are forked at their tips: this
means that they began their development as two separate ten-
tacles, one on each side of the line of union, and later united to
form a single tentacle. Similar conditions are shown in Fig. 8, b,
and in Fig. 10 of my earlier paper ('096).

CONCLUSION.

As regards the interpretation of these peculiar structures, there
is little to add except by way of confirmation to the conclusions
of my preceding paper on the subject. The outgrowths on the
rings unquestionably represent a kind of reproduction which
results from a certain degree of physiological isolation of parts,
in other words from the decrease in physiological correlation,
which is itself the result of the removal of important elements
of the original organization and of the union of the oral with
the aboral end. Moreover, it seems evident that the localization
of the tentacle groups along the line of union is due to "chance"
factors, probably in part to differences in the amount and rapidity
of the growth of new tissues and on the other hand to differences
in the degree of injury to the old organization.

That the new polarities have no direct relation to the original
polarity is evident from the fact that they may arise from either
the oral or the aboral side of the line of union, or from both sides.
From these facts we must conclude that they are the result of
local conditions in the regions concerned, and of conditions
which are not primarily concerned with the original organization.

As regards symmetry, the new growth may be either bilaterally
or radially symmetrical or asymmetrical. Clearly the symmetry,
as well as the polarity, has nothing to do with the original
organization.

And finally, these regulatory phenomena cannot be satisfac-
torily interpreted by the hypothesis which regards polarity and
symmetry as essentially a summation of the polarities and sym-
metries of ultimate oriented particles or molecules. How, for
example, shall we account on the basis of this hypothesis for
the appearance of bilaterally symmetrical groups of tentacles in
an organism which is naturally radially symmetrical. Further-

IO C. M. CHILD.

more, how, unless we accept Driesch's entelechy, can we account
for the complex changes in orientation of the particles which
must occur in the development of a new polarity and radial
symmetry from a small area of the oral or aboral end of the piece.
If we turn to the process of crystallization for assistance, we
meet difficulties, for we should expect, if the process of organic
development is analogous to crystallization, that Harenactis
material would in general conform to a particular type of arrange-
ment, at least within certain limits of environmental change.

On the other hand, if wre regard polarity and symmetry as
molar phenomena resulting from the localization by any condi-
tions or agents of certain metabolic processes differing in degree
or kind, together with the physiological correlation resulting from
such localization, we can readily understand how new polarities
may arise without reference to the old, in response to local factors,
and how bilateral symmetry may appear in one case and radial
in another in the same organism.

The greater the extent to which we interfere with or destroy
the old polarity and symmetry by removing or altering the locali-
zation of the original metabolic processes, the greater the possi-
bility of the origin of new polarities and symmetries in response
to local conditions.

In the recent experiments of Lillie, Morgan and others on the
effects of centrifuging eggs, the polarity and symmetry of the
eggs are, at least in certain cases, apparently not altered by the
displacement of the visible granules. These results have led
various experimental embryologists to adopt the hypothesis of
orientation of molecules or particles as the basis of polarity and
symmetry. As a matter of fact, however, the visible granules
which are displaced in these experiments are merely the inactive
or relatively inactive products of metabolism and so long as they
are visible, have reference primarily to past activities. Their
localization, in so far as they are localized with respect to an
axis, suggests the more or less sharp localization along this axis
of metabolic processes differing in degree or kind. There is no
reason to believe that displacement of the granules by centri-
fuging alters essentially the localization of the processes, con-
sequently the polarity and symmetry of the egg may remain

ADVENTITIOUS REPRODUCTION IN HARENACTIS. I I

unchanged, whatever the positions of the visible granules. As
a matter of fact most, if not all eggs actually show a more or less
sharp localization along the polar axis of processes differing in
degree or kind : in general the region of the animal pole reacts
more rapidly in various ways than the vegetative pole and be-
tween these two poles a gradation apparently exists. In every
fragment of such eggs above a given size this gradation must
exist in some degree, i. e., such fragments will retain the original
polarity of the whole. If, however, we could control the size
of isolated egg fragments and if it were possible to reduce the
size of nucleated fragments indefinitely we should undoubtedly
find a limit below which polarity would no longer be apparent.
In every case wrhere control of the size of isolated pieces is possible,
such a limit has been found : unfortunately the egg cell is exceed-
ingly unfavorable material for experiment along this line. It
is at present extremely difficult and often impossible to isolate
pieces of the egg of a certain desired size or from a certain desired
region: moreover, if the isolated fragment is to live and show
any developmental processes it must possess a nucleus. Mani-
festly the possibilities of investigating the polarity and symmetry
of the egg by means of isolated fragments are very narrowly
limited, as compared with the possibilities which simple organisms
with elongated axes present, and we have no right to draw con-
clusions as to the nature of polarity from experimentation on
eggs alone. Until we can actually prove that in eggs polarity
does not decrease with decrease in size below a certain limit, or
until we can demonstrate the orientation of particles or molecules
by means of the polariscope, we have no adequate grounds for
regarding polarity as anything but a molar phenomenon. More-
over, the fact already mentioned, that in all cases wyhere control
of the size of the isolated piece is possible, and where extended
experimentation has been made, the phenomena of polarity do
decrease with decreasing size, should suggest the necessity of
caution in drawing conclusions from eggs alone, where both
technical and natural obstacles to exact experimentation along
these lines exist. Moreover, I fail to see how we can avoid
accepting the evidence which the polariscope gives as possessing
much greater weight than the results of experiment with our

12 C. M. CHILD.

present crude technique, and this evidence is, except for certain
highly specialized structures, e. g., the nerve fiber, negative, so
far as I am aware. If an orientation of particles is the basis of
organic polarity and symmetry, surely the polariscope must have
given us more positive evidence of such orientation than we have
yet obtained. In the absence of such evidence the orientation-
hypothesis possesses a metaphysical rather than a scientific char-
acter. We may juggle with the ultimate particles as we please
and we may assume that they do all that is required to attain a
certain result: unquestionably we shall succeed in interpreting
all the phenomena of polarity in this way, but the value of our
interpretations is chiefly academic, and scientific proof or dis-
proof is impossible.

SUMMARY.

1. The pieces of the actinian, Harenactis attenuata, which form
"rings" by the union of oral and aboral ends about the whole
circumference, after more or less complete removal of mesenteries
and mesenterial muscles, may produce new discs with radially
arranged new mesenteries and tentacles and mouth-opening and
esophagus. Thus far the usual number of tentacles, twenty-four,
has not been attained in any case, the largest number being
sixteen. Such discs with from three to eight tentacles are of
frequent occurrence, but the formation of mouth and esophagus
has been observed only once. After the formation of the discs
they may be gradually elevated from the surface of the rings by
the development of a cylindrical column beneath them.

In addition to the well-developed discs, radially as well as
bilaterally symmetrical and asymmetrical tentacle groups may
arise along the line of union on either side, or the tissue of both
sides may take part in the formation of a single group.

2. These further experiments with rings extend and confirm
the earlier work. The new outgrowths on the rings represent a
more or less close approach to new individuals and involve the
establishment of new polarities and symmetries. They are to
be regarded as a form of reproduction related to the formation
of adventitious structures in plants. The localization of the
outgrowths, as well as their polarity and symmetry, have no
relation to the original polarity and symmetry, but are due to

ocal conditions.

ADVENTITIOUS REPRODUCTION IN HARENACTIS. 13

The results of these experiments can be more readily inter-
preted in accordance with the hypothesis that polarity and sym-
metry are essentially molar localizations or gradation of processes
along an axis or axes, than with that which regards polarity and
symmetry as the effect of a summation of the individual polarities
and symmetries of protoplasmic particles or molecules which
are definitely oriented with respect to an axis or axes.

HULL ZOOLOGICAL LABORATORY,
UNIVERSITY OF CHICAGO,
November, 1910.

ON THE INTERCHANGE OF THE LIMBS OF THE
CHICK BY TRANSPLANTATION.

FLORENCE PEEBLES, PH.D.

The experiments of Lillie ('04) and of Shorey ('09) have demon-
strated that the injury of certain parts, and the removal of organs
in the early stages of the embryonic development of the chick
are rarely followed by regeneration of the lost part. Lillie found
some evidence of the regeneration of the notochord, but a new
wing did not develop after removal of the bud. From these
experiments he drew the conclusion that the embryo of the chick
possesses little more power of regeneration than the adult.
Shorey also removed the wing buds of embryos of three to six
days' incubation, and although the region healed and develop-
ment of other parts proceeded normally, the wing buds did not
show any sign of regenerative activity.

In the winter of 1908 I undertook a series of experiments, first
to find out if the limb buds after removal could be grafted on
again, and if so whether a leg bud grafted on the proximal part
of the wing would develop into a wing or a leg, and vice versa,
if a wing bud when grafted on the proximal part of the leg would
develop into a leg or a wing.

Experiments of this kind on the chick are necessarily attended
with many difficulties, so that even under the most favorable
conditions, the percentage of successful operations is exceedingly
small. My results, therefore, are largely negative, but it may
be of interest to note what has been accomplished in the hope
that more perfect methods will enable some one to obtain more
satisfactory results.

Two general methods were followed. In the first series of
experiments the eggs were left in the shell. A window was made
above the embryo and after the operation it was sealed by the
method which I ('98) used in experiments on the primitive streak.
This method has since been used by other investigators with
success. In the later experiments the eggs were removed from

14

ON THE LIMBS OF THE CHICK.

the shell at the end of the forty-eighth hour of incubation, and
during the further development they were kept in porcelain cups
in a moist chamber. The latter method was found more prac-
ticable, as the first appearance and the subsequent growth of
the buds could be observed, and also the changes in the embryo
after the operation could be watched from time to time.

The limb buds are not large enough for removal until the
beginning of the fourth day, and allowing for the delay in develop-
ment caused by the operation, at least four additional days of
incubation are needed before the wings and legs show marked
characteristics. At the end of the fourth day the embryo lies
on its left bide; for this reason the operations were made on the
right side. The buds were removed by means of a curved knife

FIG. i. Two views of a chick on the seventh day. A, the right side showing
the stump of the wing W and the leg L with the line of the graft G. B, the normal
side.

made by heating and bending a no. 12 cambric needle into a
hook and sharpening it at the curve. The hook was inserted
under the bud and drawn up quickly, thus removing the tip
with the least possible disturbance to the surrounding parts.
The buds were then carried on the needle to the position desired,
and held in place for a few minutes until they adhered. The
greatest difficulty was met with in the effort to keep the grafts
together. Some were fastened by fine glass threads, but this
was ot found satisfactory as the insertion of the thread tore the
tissues. In many of the embryos the grafts came apart, and the
buds floated in the albumen or sank in the yolk. Another dif-

i6

FLORENCE PEEBLES.

ficulty arose through e? cessive bleeding after the operation. This
usually esulted in the death of the embryo.

A brief description of the behavior of one or two of the embryos
will suffice to show the general results.

A. The leg and wing buds of an embryo, incub ted two days
in the shell and then two days in a porcelain cup, were removed.
The leg bud was grafted on the proximal part of the wing, and
the wing bud on the proximal pa, i of the leg. The embryo was
then placed in a moist chamber in the incubator where it was
left for three days. On removal it was found that the graft in
the wing region had separated, a d that the bud on the proximal
end of the leg was only partly attached (Fig. i, A). The left
side appeared as shown in Fig. i, B. There was no regeneration
in the wing region, and no apparent development in the leg
region. Sections of this embryo showed the normal limbs to be
composed of a homogeneous mass of mesoderm surrounded by

FIG. 2. Two views of a chick on the eighth day.
grafts; B the normal side.

A, the right side showing

ectoderm. The limbs on the right side were similar in structure.
None of them showed distinguishing characteristics.

B. The same experiment was performed on this embryo as on
the one just described. In this case the chick lived four days
after the grafts were made. Both unions were complete, as

ON THE LIMBS OF THE CHICK.

shown in Fig. 2, A, G. In Fig. 2, B, the left side is indicated.
Sections of this chick show concentrated areas where the skeleton
is beginning to develop. These areas are indicated in Fig. 3,
A, B, C, and D by dots. A section of the normal side of the
embryo is given in Fig. 3, D.

I have not succeeded in keeping any embryos alive beyond
the ninth day of incubation, but a normal embryo of this age
possesses wrings and legs which are readily distinguishable. My

B

D

FIG. 3, A and B. Sections of the grafted wing W showing rudiments of the
skeleton R; C, section of the leg L with the rudiments of the skeleton R; D,
section of the left side. The leg L appears in cross section, the wing W in longi-
tudinal section.

results show that the operation greatly retards the development.
If the embryos could be kept alive a few days longer it would,
no doubt, be possible to determine positively whether or not
the grafted tips become a part of the limb to which they are
attached irrespective of their former position. The results of
my experiments indicate that they do. The failure to keep the
embryos alive is probably due to disturbance of the development
of the allantois.

In regard to regeneration the results obtained from the removal

1 8 FLORENCE PEEBLES.

of the leg and wing buds fully confirm those of Lillie and Shorey.
In no case was the slightest sign of regeneration observed.

SUMMARY.

1. It is possible for chick embryos to develop in porcelain cups
in a moist chamber at the proper temperature, up to the ninth
day, although the development is delayed.

2. The leg bud when removed may be grafted on the proximal
part of the wing and the wing bud may be grafted on the proximal
portion of the leg without permanently injuring the embryo.

3. The results indicate that when the tip of a young bud is
grafted on the proximal portion of another limb it becomes a
part of the appendage to which it is attached instead of retaining
the character of the part it is destined to become.

4. No regeneration of the limbs takes place after the removal
of the buds.

BRYN MAWR COLLEGE,
BRYN MAWR, PA.,

November 12, 1910.

LITERATURE.
Lillie, F. R.

'04 Experimental Studies on the Development of the Fowl (Callus domeslicus).
Biological Bulletin, Vol. VII.. No. i.

Peebles, F.

'98 Some Experiments on the Primitive Streak of the Chick. Archiv fur
Entwickelungsmechanik der Organismen, Vol. VII.

Shorey, M. L.

'07 The Effect of the Destruction of Peripheral Areas on the Differentiation
of the Neuroblasts. Journal of Experimental Zoology, Vol. VII., No. i.

THE RELATION BETWEEN CHROMOSOME-NUMBER
AND SPECIES IN NOTONECTA.1

ETHEL NICHOLSON BROWNE.

After working for several years on the spermatogenesis of
Notonecta, some very interesting facts have recently come to
light which I wish to present in a preliminary paper. The ma-
terial, collected at Woods Hole, consists of three species, Notonecta
undulata Say, N. insulata Kirby and N. irrorata Uhler. The
character of the chromosome groups is quite distinctive for each
species, TV. insulata representing a transition stage between TV.
undulata with a larger number of chromosomes and N. irrorata
with a smaller number.

In all three species there is present a pair of idiochromosomes,
which divide separately in the first spermatocyte-division, as
described by Wilson2 for Euchistus, Ccenus, etc. Accordingly,
there is one more than the reduced or haploid number in the first
division. Owing to the fact that the conjugation of the idio-
chromosome-pair is often still further delayed in Notonecta, the
two unequal mates frequently lie side by side on different spindle
fibers in the second division (Figs. 7, 9, 16, 18, 33, 35). They
subsequently pass to their respective poles without ever having
conjugated. As a result of the second division, two classes of
cells are produced, one containing the small idiochromosome and
the other the large one (Figs. 10, 19, 36). Accordingly, the
spermatozoa are of two kinds which, from analogy with other
insects, are the male-producing and the female-producing sperm-
atozoa respectively.

In N. undulata there are 14 chromosomes in the first spermato-
cyte-division, consisting of a ring of 12 chromosomes with two
small ones in the center (Figs. 1-3). These small ones frequently

JI wish to express my thanks to Professor E. B. Wilson for his helpful sugges-
tions and criticism during the course of the work, to the director of the Marine
Biological Laboratory and the Wistar Institute of Anatomy for the facilities
offered me at Woods Hole, and to Mr. E. P. Van Duzee for identifying the species.

2Wilson, E. B., 1905, "Studies on Chromosomes," I., Journ. Exp. ZooL, II., 3.

'9

2O ETHEL NICHOLSON BROWNE.

give the appearance of being on the same spindle fiber (Fig. 4).
There are 13 chromosomes in the second division, a ring of 12
chromosomes surrounding the idiochromosome-pair in the center
(Figs. 5-10). The spermatogonial number is 26, including two
pairs of small chromosomes, the same ones, evidently, which lie
in the center in the first division (Fig. n).

In N. irrorata there are but 13 chromosomes in the first division,
a ring of 12 surrounding one small chromosome in the center
(Figs. 12-14). In the second division there are 12 chromosomes
including the idiochromosome-pair in the center (Figs. 15-19).
The spermatogonial number is 24, the smallest pair of chromo-
somes corresponding to the small chromosome within the ring
in the first division (Fig. 20).

In N. insulata there are two types of first division groups,
•occurring in the same testis, in approximately equal numbers.
One type has 14 chromosomes, a ring of 12 surrounding two small
ones (Figs. 21-23); the other type has but 13 chromosomes, a
ring of 12 surrounding only one small one (Figs. 24-26). The
apparent discrepancy is accounted for by the fact that the second
small chromosome which lies in the center in the 14-group is
frequently found in the ij-grotip attached to the largest chromosome.
Serial sections of four spindles as seen in side view are given
in Figs. 27-30. In the first three series (Figs. 27, 28, 29), there
is only one small chromosome in the center of a ring of 12 chromo-
somes; the second small one is attached to the largest chromo-
some (Ma). In the fourth series (Fig. 30), the two small chro-
mosomes lie free in the center of a ring of 12. A polar view
showing the compound character of the large chromosome in the
13-group is given in Fig. 31 (Ma). No case of such a compound
chromosome has been found associated with the 14-group, and
some fifty clear cases have been observed in the 13-group. In
many cases, however, the compound nature of the large chromo-
some could not be determined in the 13-group, the two com-
ponents having evidently fused beyond recognition. The fusion
must take place in all cases before the second division, for there
are always, so far as I have observed, 12 chromosomes in this
division, including the idiochromosome-pair in the center of the
group (Figs. 32-36). All trace of the original composition of

CHROMOSOME-NUMBER IN NOTONECTA. 21

the largest chromosome has been lost. It has a decided quad-
ripartite appearance, as though each element into which it divides
were composed of two equal parts. Its appearance suggests that
of the large chromosome described by Wilson1 in the second divi-
sion of Nezara, except that in Nezara the two parts are somewhat
unequal. Unfortunately, no satisfactory spermatogonial groups
have been found, but the expectation would be either 26 or 24,
in the latter case, two of them being compound in character.

DISCUSSION.

The chief interest in Notonecta lies in the fact that there is
a definite relation between the chromosome-number and the
species, and that the change in number can be attributed to the
behavior of a particular chromosome. In N. undulata this chro-
mosome is present as a separate element, together with another
small one, in the center of the spindle in the first division, and in
the peripheral ring in the second division. In TV. irrorata, this
chromosome is lacking throughout both divisions, there being
only one small one in the center of the spindle in the first division.
TV. insulata gives a transition stage between the two; the small
chromosome is present in the first division, either free in the
center as in TV. undulata, or fused with the largest chromosome;
and it is apparently absent in the second division as in TV. irrorata.
Every chromosome cannot be homologized individually in the
three species, for the size relations are different, especially in the
case of the idiochromosomes. The largest chromosome and one
small one can be followed throughout in the three species. Like-
wise, we may safely compare, I think, the other small chromo-
some in TV. undulata with the one of like size and position that
sometimes occurs in TV. insulata in the first division; and since
we can follow the steps of its fusion with the large chromosome
in TV. insulata, we may, I think, reasonably attribute its absence
in TV. irrorata to its permanent fusion with the large chromosome.
A schematic representation follows, only the chromosomes that
are comparable being separated from the ordinary chromosomes
or autosomes, which are designated A. The large and small

'Wilson, E. B., 1910, "Note on the Chromosomes of Nezara," Science, May 20,
1910.

ETHEL NICHOLSON BROWNE.

idiochromosomes are represented by / and i respectively, the
largest chromosome, or macrochromosome by M, the small chrom-
osomes, or small autosomes by a.

Products of the First Division. Products of the Second Division.
N.undulata M+gA+I+i+a+a (14) ( M+gA +1 +a +a (13)

and

(M+gA+i+a+a (13)
either
N. insulata M+gA+I+i+a+a (14) "j (Ma+gA+I+a (12)

or and

Ma+gA+I+i+a (13)) (.Ma+gA+i+a (12)

N.irrorala Ma+gA+I+i+a (13) (Ma+gA+I+a (12)

and
(Ma+gA+i+a (12)

TV. insulata evidently represents a transition stage between TV.
undulata with a larger number of chromosomes and TV. irrorata
with a smaller number. It is not possible to say which is the
primitive condition. We might assume that TV. undulata is the
original species, from which, by a process of progressive fusion,
have arisen the conditions seen in TV. insulata and TV. irrorata.
We might, however, assume that the reverse process has taken
place; or, that TV. insulata arose together with one of the other
species from the third species.

The somatic characters give no clue as to which of these inter-
pretations should be adopted. We could easily derive the wing
coloring of TV. insulata from that of TV. undulata by substituting
brown pigment for white, and the wing coloring of TV. irrorata
from that of TV. insulata by substituting black pigment for brown,
but leaving some of the brown as mottling. But TV. irrorata is
intermediate in size between the small species TV. undulata and
the large species TV. insulata. Of course the most striking somatic
differences are not necessarily those correlated with the fusion
or separation of the two chromosomes; these differences may be
correlated with other chromosomes which, as I have stated, differ
in size in the three species.

The fact that a transition stage has been found in one species
of Notonecta between two other species of the same genus in
respect to a fusion and separation of two chromosomes lends
support to the views advanced regarding the origin of two or more
chromosomes from a single chromosome where the transition

CHROMOSOME-NUMBER IN NOTONECTA. 23

stages do not exist. Payne1 has concluded that the multiple
X-element in the Reduviidae has arisen from a single X-element
of an ordinary idiochromosome-pair by a separation into two or
more parts. Wilson2 has accounted for the double accessory in
Syromastes by assuming that it has arisen from a single chromo-
some that has split into two parts. These hypotheses are sus-
tained by the evidence of Notonecta, two species of which show
permanent fusion and separation, and the third species the inter-
mediate condition.

Besides the interesting correlation of the change in chromo-
some-number and change in species in Notonecta, the condition
of temporary fusion and separation of two chromosomes in one
species, TV. insulata, is of considerable interest, especially in com-
parison with the facts found in other insects. Sinety3 has found
in the Phasmidae and McClung4 in other orthoptera that the
accessory is sometimes attached to another chromosome ; in Mer-
meria, McClung found this compound chromosome further as-
sociated with another chromosome. Wilson5 describes in Meta-
podius a coupling of the supernumeraries with the idiochromo-
somes during part of the maturation-process. Wallace6 has found
in the spider that the accessory is made up of two components,
sometimes united, sometimes separate. Similarly, in Syromastes.
Wilson7 has found that the accessory consists of two unequal
elements which are closely united in the maturation-divisions.
In the Phylloxerans, Morgan8 describes two accessories, some-
times separate, sometimes united. It is to be noted, however,

!Payne, F., 1909, "Some New Types of Chromosome Distribution and Their
Relation to Sex," BIOL. BULL., XVI., 3, 4.

2Wilson, E. B., 1909, "The Female Chromosome Groups of Syromastes and
Pyrrochoris," BIOL. BULL., XVI., 4.

3Sinety, R. de, 1901, "Recherches sur la Biologic et 1' Anatomic des Phasmes,"
La Cellule, XIX.

4McClung, C. E., 1905, "The Chromosome Complex of Orthopteran Sperma-
tocytes," BIOL. BULL., IX., 5.

6Wilson, E. B., 1909, "Studies on Chromosomes," V., Journ. Rxp. Zoo/., VI., 2.

6Wallace, L. B., 1909, "The Spermatogenesis of Agalena ncevia," BIOL. BULL.,
XVII. , 2.

7Wilson, E. B., 1909, "Studies on Chromosomes," IV., Journ. Rxp. Zool.,
VI., i.

8Morgan, T. H., 1909, "A Biological and Cytological Study of Sex Determination
in Phylloxerans and Aphids," Journ. Rxp. Zool., VII., 2.

24 ETHEL NICHOLSON BROWNE.

that in all these cases, the temporary separation or union is
connected with the sex-chromosomes, whereas in N. insulata
it concerns two ordinary chromosomes or autosomes.

MARINE BIOLOGICAL LABORATORY,
WOODS HOLE, MASS.,
September, 1910.

26 ETHEL NICHOLSON BROWNE.

EXPLANATION OF PLATE I.

Notonecta undulata. Figs, i, 2, metaphase of first division showing 14 chromo-
somes, two small ones in the center. Figs. 3, 4, initial anaphase of first division,
side view, two small pairs in the center. Fig. 5, metaphase of second division
showing 13 chromosomes, large idiochromosome in the center. Figs. 6, 7, the
same, but both idiochrosomes showing in the center. Fig. 8, initial anaphase
of second division, side view, idiochromosome-pair in the center. Fig. 9, the
same, but idiochromosomes on different fibers. Fig. 10, A and B, sister anaphase
groups of second division, from the same spindle. Fig. n, spermatogonial group
showing 26 chromosomes.

All the figures were drawn with a camera, 1/12 oil immersion and a 12 com-
pensation ocular, then reduced in the engraving to two thirds.

BIOLOGICAL BULLETIN, VOL. XX.

PLATE I.

•

*

r l I

'//,

//

II 1

.'•«

• • •

*% •

A*«. •»•
ro

B

fTHEL N. BR..WNE.

28 ETHEL NICHOLSON BROWNE.

EXPLANATION OF PLATE II.

Notonecta irrorata. Figs. 12, 13, metaphase of first division showing 13 chromo-
somes, one small one in the center. Fig. 14, initial anaphase of first division,
side view, one small pair in the center. Fig. 15, metaphase of second division
showing 12 chromosomes, large idiochromosome in the center. Fig. 16, the same,
but both idiochromosomes showing in the center. Fig. 17, initial anaphase of
second division, side view, idiochromosome-pair in the center. Fig. 18, the same,
but idiochromosomes on different fibers. Fig. 19, A and B, sister anaphase groups
of second division, from the same spindle. Fig. 20, spermatogonial group showing
24 chromosomes.

BIOLOGICAL BULLETIN, VOl. XX.

PLATE II.

*• • i

12

'//y

ill

1'ife Wk

*

/5 A ** 19

B

ETHEL N. BROWNE.

3O ETHEL NICHOLSON BROWNE.

EXPLANATION OF PLATE III.

Notonecta insulata. Fig. 21, metaphase of first division showing 14 chromo-
somes, two small ones in the center. Figs. 22, 23, anaphase of first division, side
view, two small pairs in the center. Figs. 24, 25, metaphase of first division showing
13 chromosomes, one small one in the center. Fig. 26, initial anaphase of first
division, side view, one small pair in the center.

BIOLOGICAL BULLETIN, VOL. XX.

PLATE III.

\ •

fl

21

•• 25

1

22

23

ETHEl. N. BROWNE.

32 ETHEL NICHOLSON BROWNE.

EXPLANATION OF PLATE IV.

Nonecta insulata. Serial side views of four spindles of first division, as seen
in adjacent sections. Figs. 27-29, .4, B, C, three entire spindles each showing
13 chromosomes, one small pair free in the center in B, and a compound chromo-
some in C (Ma) consisting of the large and the other small chromosome united.
Fig. 30, A, B, C, entire spindle showing 14 chromosomes, two small pairs free in
the center in B; the two elements of the compound chromosome are separate as
M and a in C and B.

Some of the chromosomes have been displaced on the spindles so as to lie side
by side instead of one above another, for the sake of clearness.

•BIOLOGICAL BULLETIN, VOL. XX.

m\
ffl'j

/!/<
4/,

m

I

If

'/

Ma

m

3o c

ETHEL N. BROWNE.

34 ETHEL NICHOLSON BROWNE.

EXPLANATION OF PLATE V.

Notonecta insulata. Fig. 31, metaphase of first division showing 13 chromo-
somes, one small one in the center, the other small one attached to the large one
forming the compound chromosome Ma. Figs. 32, 33, metaphase of second
division, showing 12 chromosomes, both idiochromosomes showing in the center,
Fig. 34, initial anaphase of second division, side view, idiochromosome pair in the
•center. Fig. 35, the same, but idiochromosomes on different fibers. Fig. 36,
A and B, sister anaphase groups of second division, from the same spindle.

•BIOLOGICAL BULLET, N, VOL. XX.

PLATE V.

Ma

f

* *

9
»
32

33

/f/l^

§!

i/lr

•
•
»
•

% •• *

B

iCTHEL N. BROWNE.

SOME RESULTS OF CASTRATION IN DUCKS.1

H. D. GOODALE.

Introduction. — The present paper is largely in the nature of a
first report and will be followed by others as the experiments
warrant. In addition to a brief review of some other studies on
castration, I have considered, also very briefly, some problems
relating to sexual dimorphism, sex-limited inheritance and the
determination of sex.

It is a matter of rather common knowledge that females of
birds and mammals, which have become sterile from any cause,
may assume the plumage and other secondary sexual characters
of the male. On the other hand, the male when castrated, often
does not develop his usual trappings. This has often been inter-
preted as a development of the female characters. It has,
however, been pointed out that castration of the young male
causes rather a retention of his own youthful characters.

Castration, besides its effect on the secondary sexual charac-
ters, may lead to certain specific changes in the organism, not
correlated with secondary sexual characters. Thus the capon
grows larger than the cock and is more sluggish. Tandler and
Grosz found in man, among other things, that there was a ten-
dency for fat to develop in certain regions, and either a failure,
or poor development of body, axial, and pubic hairs.

The present paper deals almost wholly with the effect of cas-
tration on the secondary sexual characters.

The breed known as Rouens was selected for this work, largely
because they are strongly sexually dimorphic in plumage.2

The Rouens are probably derived from the mallard, Anas
boschas. Coues says, "nearly everywhere domesticated, being
the well known original of the barnyard duck." In some respects
the Rouens differ from the mallard. The latter are much smaller.

'For the suggestion of studying the effects of castration in ducks, I am indebted
to Professor T. H. Morgan.

2The individuals used in these experiments were reared from eggs secured from
the White Birch Poultry Farm, Bridgewater, Mass.

35

36 H. D. GOODALE.

The tail of the male has some white, while in summer plumage
he resembles the female Rouen. (For description of latter see
below.) The female mallard, though of the same general color, is
of much lighter tone throughout than the Rouen female, which
might easily pass as a melanistic variety of the former. For prac-
tical purposes we are dealing with a wild species.

In addition to a general description, I have given a somewhat
detailed description of the feathers of the various sections of the
birds. There is, however, so much variation in the patterns of
the feathers of the female and the male in his summer plumage,
that these descriptions and figures illustrate only a few types.
This variation exists, not only between different birds, but even
in a single section of one individual. Though two feathers are
rarely exactly alike, the majority of feathers in a single section
have a general resemblance. On the other hand the feathers
of each section of the male's breeding plumage are very con-
stant in type.

GENERAL DESCRIPTIONS.

Male in Breeding Plumage (Fig. i). — Bill: greenish yellow.
Head and upper part of neck: rich metallic green. Then comes
a narrow white ring often not quite complete dorsally. Ventral
side of remainder of neck and a large part of breast: rich claret.
Remainder of ventral surface: iron gray, becoming lighter toward
the anus. Posterior to the last it becomes darker. Back: very
dark gray in neck region, becoming still darker in the middle and
greenish black on the rump. The two median tail feathers are
strongly curved antero-dorsally. They have the same color as
the rump. The next two are often curved, but may be duller in
color. These four form the so-called sex feathers. The rest of
the tail and the upper surface of wing is brownish black. The
latter has a speculum of iridescent purple. Under surface of
wing: white. The drake's voice is a soft qua, either rapidly re-
peated, with a slight pause after every other note, or else a some-
what louder, single note, much prolonged. In some varieties the
voice and sex feathers are the only conspicuous secondary sexual
characters.

Female (Fig. 2). — Throughout a mixture of buff and dark
brown. Bill: brownish black, usually mottled with yellow. Occu-

RESULTS OF CASTRATION IN DUCKS.

37

FIG. i. Male in breeding plumage.

FIG. 2. Normal female.

38 H. D. GOODALE.

pying the dorsal side of the head and upper neck is a broad dark
stripe, tinged with green and inconspicuously striped with buff.
The sides of the head are buff with small dark flecks. Passing
across the side of the head from the nostril through the eye to
the dorsal surface and from the angle of the jaw to the eye, are
two dark bands. Sides of upper neck: mixed buff and dull black.
Throat: buff. Dorsal surface of body often much darker than
ventral, and usually less conspicuously striped. Speculum and
under surface of wing as in the male. Sex feathers absent.
Voice a loud quack. At times it is modified into a softer, rapidly
repeated quac.

Summer Plumage. — At the close of the breeding season,

FIG. 3. Male in summer plumage.

which comes early in July, both sexes molt. The new coat of the
male, when complete, is different from the old (Fig. 3). The
molt, however, takes place piece-meal, certain sections being
nearly completed before others begin. Consequently there is
only a very brief period at the end of the summer when the new
coat exists as a whole. Then the transition begins again to the
usual coat which is completed early in October. In the modi-
fied coat the male is said to be in "summer plumage," or in a
"state of eclipse."

Head: similar to that of female. Breast: mottled dull black

RESULTS OF CASTRATION IN DUCKS. 39

and reddish buff. Keel: dull slate with huffish streaks. Re-
mainder not conspicuously modified. Sex feathers absent.

The female, so far as I have observed, undergoes no radical
change in coat color, though in one case, at least, the patterns of
the feathers were considerably different in August from those of
October. Compare Fig. 4, E with F; Fig. 6, D with F.

As a result of the molt in early summer, both sexes of the adult
and the young are quite similar for a time. Its meaning is un-
certain. Newton ascribes the assumption of the summer plumage
by the drake to the loss of the power of flight temporarily, since
the remiges all fall out together. But in other cases where a
similar double molt occurs, the male retains the modified plumage
for several months; nor does he become incapable of flight. As
will be shown later, the presence of the active testis is necessary
for the drake to assume this plumage. Conditions in the mallard,
as far as I can find out, appear to be somewhat different from
the Rouen. The summer plumage seems to persist longer. The
one male in the American Museum in a transition stage, had a
brown edging to the feathers of regions, which, in the male
Rouen, have no such edging.

Young. — -When first hatched the young are brownish black
with a few yellowish stripes or spots. While they are acquiring
their first coat of feathers both sexes resemble the female in gen-
eral appearance. This summer, however, I found certain char-
acters of the plumage by which the sexes could be distinguished
at a very early age (Fig. 8, D, E, from birds about 7 weeks old).
The young "peep" for several weeks after hatching. The voice
of the young female is modified into that of the adult a few weeks
before that of the young drake. The latter is about four months
old when he takes on the adult plumage. At this time the
testes are still very small.

Description of Feathers.1 — Male in breeding plumage. — Head
and upper neck: distal half, metallic green; proximal half, dull
black with narrow white base; on the ventral side of head, this
base may be more extensive. Neck ring: entirely white. The
feathers of regions immediately adjoining the neck ring may be
many kinds of mixtures of white, red and black. Ventral surface.

JThe words distal and proximal refer to the feathers.

4O H. D. GOODALE.

-Lower throat: distal half, deep claret; proximal half slate, some-
times stippled or spotted with buff. Breast: distal half claret;
remainder slate, becoming white at base, usually stippled with
buff (Fig. 4; A)', sometimes numerous transverse vermiculations
are present (Fig. 4, B} ; less frequently, still other slight modifica-
tions may occur. Keel (Fig. 5, A, B] and laterals:1 light gray be-
coming darker proximally; distal half traversed by numerous
transverse slate colored vermiculations;2 near the breast region
the tip of the feather may be shaded with faint claret (Fig. 5, B}.
Posterior to vent: (Fig. 8, /) similar to preceding, but the slate
vermiculations are much broader and often become confluent,
forming patches. Sides posterior to folded wing : (Fig. 8, /) similar
to keel but much lighter ; vermiculations often broken ; proximally :
slate at base, but more distally often changes abruptly to white.
Under tail converts: black becoming lighter toward the base; ex-
posed surface with metallic sheen. Dorsal surface.— Lower neck
near body : slate, becoming somewhat lighter toward the base ; light
gray vermiculations distally (Fig. 6, A). Passing back along the
back the gray vermiculations tend to disappear except at the edge
of the feathers. The slate also gradually gives place to black which
as it nears the rump becomes metallic blue or green on its exposed
surface. Rump (Fig. 7, A), upper tail coverts and sex feathers:
black with metallic green or blue on the exposed surface; vermi-
culations rarely occur. Main tail, primaries and primary coverts r
dull brownish black. Secondaries: a narrow white band across
the tip; remainder, outer surface, dorsal side of vane: dull
brownish black; ventral side of vane: next to white tip a narrow
velvety black band; the rest, except for a narrow dull band
near the base is brilliant changeable purple, blue or green; inner
surface, dull brownish black. The innermost secondaries: slate,,
often with velvety black. Secondary coverts: velvety black
band across the tip, then a narrow white band, remainder brown-
ish black. Remainder of upper surface of wing: brownish black.
The feathers at the anterior edge have gray vermiculations which
disappear posteriorly. Scapulars: various modifications of the

JOn the sides of the body just beneath the edge of the folded wing is an elongated;
patch of very large feathers, which I have thus designated.

2I have arbitrarily called the narrowest bands vermiculations.

RESULTS OF CASTRATION IN DUCKS. 41

uniform slate and gray vermiculated types. There is also a
peculiar tuft of small feathers just anterior to the junction of
the wing and body. Each feather (Fig. 8, A} has a narrow band
of slate and light gray vermiculations across the tip, then a
curved band of light claret, the remainder being a mixture of
slate and gray.

Male in Summer Plumage. — Top of head: almost identical
with normal type, but much less brilliant. Light line above ever
distal half mostly buff with small irregular blotches of black;
base, dull black; sometimes the entire feather is nearly black.
Beneath the eye: dull black; distal half with buff edges. Throat:
reddish buff with irregular dull black marks; basal half slate.
Upper breast: (Fig. 4, C) bands of reddish buff or claret and dull
black; though irregular they are more or less transverse; basal
half: slate. Regular patterns may occur (Fig. 4, D). Lower
breast: buff in place of reddish buff; otherwise as preceding;
sometimes feathers like those of one of the female types are
found (Fig. 5, D}. Anterior keel: (Fig. 5, C) mostly slate; mar-
gin of varying width vermiculated with gray; tip usually of rufus
shade. Posterior keel : dull slate, vermiculated margins reduced
or absent. Posterior to vent: dull black with vermiculated mar-
gins. Under tail coverts: dull black with narrow buff margin on
exposed part, sometimes vermiculated. Sides of body posterior
to wing: (Fig. 8, H) slate with narrow gray vermiculations.
Dorsal surface of body: modifications often slight throughout;
the feathers at union of neck with body (Fig. 6, C) lose most of
their vermiculations and may have buff edges. Scapulars: dull
slate. Laterals somewhat modified, often with a distinct tend-
ency for the formation of transverse reddish buff and black
bands across the tip, similar to those of the breast; vermicu-
lations reduced or absent.

Female. — Top of head: distal half, dull greenish black with
buff margins; basal half slate. Line above eye: reddish buff
with median dark stripe; basal half slate. Throat: reddish buff
with gray base. Upper neck at junction with next section: buff
with irregular longitudinal dark stripes; base slate. Breast:
buff, sometimes reddish, with dark patches, arranged in numerous
patterns (Fig. 4, E-H). Keel: (Fig. 5, E, F) similar to preceding

42 H. D. GOODALE.

but usually duller and if anything less regular in pattern. Near
the anus the feathers are apt to show a more regular pattern
similar to Fig. 4, E. Dorsal surface: anteriorly (Fig. 6, D, E),
rather brighter colors than elsewhere and often a distinctly more
regular pattern. Fig. 6, D, closely approximates the fancier's
ideal. Posteriorly the pattern becomes variable again and colors
duller (Fig. 7, C, E). In some females the entire dorsal surface
is almost black, as the buff markings become faint (Fig. 7, E).
Main tail: brownish black, with narrow buff margins. Junction
of wing with body: (Fig. 8, C) light rufous with irregular black
spots. Scapulars: various modifications of "ideal" type. Upper
surface of wing not including remiges: dull black with buff
margins. Remainder of wing like males except that inner secon-
daries have some tendency toward developing white bands.

Young. — Unfortunately I have only a few notes on the feathers
of the young. In one young female, the breast (Fig. 8, E) and
keel feathers were dull brownish black with a buff margin, which
is very narrow at the apex. The rump feathers were black with
a narrow subapical buff band bounded by a margin of black.

In two young males the breast feathers were dull black, the
distal half margined with buff and having two narrow transverse
subapical buff bands (Fig. 8, D}. The keel feathers were like
those of the young female. The rump feathers were black like
those of the adult male.

The general appearance of the young female, even before she
gets a full coat of feathers, is like that of the adult. The young
male, too, resembles the female, until one learns the characters
which distinguish him from his sisters.

The Relation between the Summer Plumage of the Male and the

Female's Plumage.

These descriptions show that the male in summer plumage
merely mimics the female. He does not even become entirely
like her. In certain sections, as pointed out, there are no modi-
fications toward the female type. In others, i. e., the head,
breast and keel region, the feathers of the male become quite
like those of the female. Individual feathers may be indistin-
guishable from female feathers. But the male in addition has

RESULTS OF CASTRATION IN DUCKS. 43

types of feathers which the female lacks. Thus, the feathers
of the top of the head have no buff edges. Those of the posterior
keel region may have gray vermiculations on the margins. Those
of the throat and upper neck regions, however, appear to be
nearly identical with female feathers of the same sections. The
breast feathers present some complications. Certain types of
these are common to both sexes, but the male has certain forms
not possessed by the female. The differences, though elusive,
may perhaps be stated as follows: The apical part of the male's
breast feather is claret; that of the female is buff. There is a
tendency in the male for the red or buff bands to increase from
two to three and to become transverse. In the female the ten-
dency is for the bands to become reduced in number and to
become longitudinal. Only a statistical study could determine
just the exact relations between the two sexes.

This discussion, I think, makes it clear that the modifications
in the plumage of the male are, on the whole, of a type peculiar
to his sex. It can hardly be maintained that this is an example
of the assumption by the male of the female's plumage, especially
as I shall show later, the presence of the testis is necessary for
its appearance. The power to change his plumage is, indeed, a
sort of physiological secondary sexual character.

EXPERIMENTS.

Thus far I have castrated 7 males and 5 females.1 Three
males and two females lived for considerable periods. No. 47Ocf
and 49 were castrated in early spring, 1909, when a little less
than a year old. The right testis of No. 190? was removed at
this same time, but the left was not removed till August 8. The
left testis of No. icf (hatched spring 1908) was removed August
8, but owing to excessive bleeding, the right was well ligatured
and allowed to remain. These two males were in summer plu-
mage when operated on. No. 249 , 12 weeks old, was castrated
August 13. No. icf, No. 249 and No. 49 are still alive.

Results. — -Males: No. 470 did not take on the summer plumage
in 1909. He was unfortunately killed by dogs in spring of 1910.
No. I molted in August, 1910, but did not take on the summer

JNot including several others which died in the early part of the work.

44 H- D. GOODALE.

plumage except for a few feathers of head, throat and breast.
No. 19 with one testis and this injured assumed the summer
plumage in 1909. He did not long survive the second operation.

Evidently the presence of the active (?) testis is necessary
for the male to assume the summer plumage. The presence of
the ligatured testis of No. I indicates that ligaturing is as effective
as removal. I do not know the present condition of the testis
of this male. There was no assumption of female characters
by any of these males.

Females: No. 4 and No. 24 were the only females that lived
more than a few days after the operation. October 20, 1909,

FIG. 9. Female 4, castrated.

they showed no marked modifications, though a small sample of
No. 24*5 breast feathers, saved at that time, are now found to
be quite like those of the male in summer plumage. Those of
No. 4 were of the usual female type. Through the winter and
spring, these two, with several other females were confined in a
rather dark pen. During this time, only cursory observations
were made, for it was believed that the experiments had failed.
July 4, 1910, this flock, for the first time in several months, was

RESULTS OF CASTRATION IN DUCKS.

45

given their liberty out of doors. This procedure at once revealed
that two individuals carried sex feathers. As no males were
with the flock, they were immediately examined carefully. Their
band numbers, of course, provided the necessary identification.
No. 4 (Fig. 9) in addition to the sex feathers, had breast feathers
similar to those of the male's breast in summer plumage. In a
few feathers, gray vermiculations were present. Such other

FIG. 10. Female 24, castrated. Photographed August.

modification as had taken place (e. g., very little buff in dorsal
regions) was not beyond the range of variation in normal females.
No. 24, Fig. 10, was rather more strongly modified. She had a
very narrow white neck ring. The feathers of the breast were
distinctly of the male type (Fig. 4, /, J). Those of the dorsal
side of the lower neck region were of a type occasionally occurring
in the same region in males. In other parts of the body there
was a distinct tendency for the buff bands to become transverse.
Compare Fig. 6, G, H. Feathers, more or less vermiculated,
were quite common. A few found in the anal region, were in-
distinguishable from similar male feathers. Compare Fig. 8, G
with /.

46

H. D. GOODALE.

The next molt of these birds, begun in September, is now
(November) well advanced. On the whole No. 4 has made little
advance toward the male type. There is much more buff in the
dorsal regions than previously, but as stated above, the diminu-
tion of buff as noted in the summer was not beyond the range
of variation in normal females. On the other hand, a distinct
neck ring is now present where earlier there was only a trace.

No. 24 (Fig. n), on the contrary, has made a distinct ad-
vance toward the male type. Compare Fig. 10 with Fig. n. A
comparison of Fig. 1 1 with Fig. 3 shows how closely she resem-

FIG. ii. Female 24, castrated. Photographed November.

bles the male in summer plumage, even though there are still
many unshed feathers of last summer's coat.

Description.- — Head mostly that of female type but in addition
numerous brilliant green feathers just like those of the male
have developed, though at present they are confined to the dorsal
surface. Neck ring better developed than previously. The
feathers of the dorsal surface are very similar to those of the male
(Fig. 6, /, Fig. 7, G, H), though on the back towards the neck,
feathers like those shown in Fig. 6, G, J, are common. The new
feathers of the keel are shown in Fig. 5, H, I. The upper one

RESULTS OF CASTRATION IN DUCKS. 47

is the more common. Its transverse buff bands recall those of
the young male. The new feathers around the vent and sides
of body posterior to wings (Fig. 8, G, F) are largely of the male
breeding plumage type, though those of the sides are somewhat
darker, more like the feathers of the same place in summer
plumage. The new laterals are of the male summer type. Con-
ditions in the breast region are rather complex. The upper part
as a whole, is a deep claret, gradually becoming lighter ventrally
and shading into the keel region. There is no sharp boundary
as in the male. A feather from the upper (anterior) region is
shown in Fig. 4, L. Further down on the breast, types like that
shown in Fig. 4, K, are very common. The types shown in Fig.
4, B, Fig. 5, I, and Fig. 6, G, also occur, the last being especially
frequent at the sides.

This female evidently is still undergoing modifications and in
due course of time may assume the complete plumage of the
male. At present she has the following distinctively male char-
acters. Brilliantly green feathers on head; white neck ring;
much claret in breast and some feathers indistinguishable from
the male's; numerous vermiculated feathers, often identical with
those of the male; rump feathers like that of male; and sex
feathers. I have not seen feathers like those shown in Fig. 5,
H, I; Fig. 6, G, in either normal sex, though feathers suggesting
the last have been seen in the male: The few feathers of purely
female type remaining are in the head and upper surface of
wing. The color of the bill has not changed. The voice, too,
is still the female's, though it has a tendency to break.

All that can be said in regard to sexual behavior is that the
castrated females seem quieter than the others. Fuller obser-
vations will be made when the breeding instincts of both sexes
become more active.

While the difference in plumage between these females may
be due to the difference in age when castrated, no discussion of
this point will be attempted in the present paper.

Discussion. — The results of these experiments show that re-
moval of the testes does not bring about the assumption of the
female characters, but at most results in the loss of a male char-
acter. The loss of the power of taking on the summer plumage

48 H. D. GOODALE.

is similar to the loss of power by the castrated stag of renewing
the antlers each year. Whether castration of the drake, very
early in life, will prevent the assumption of the breeding plumage
will be determined later.

Removal of the ovary, however, has an entirely different re-
sult. The female after a time may gradually lose her normal
characters and assume those of the male. The results entirely
confirm previous observations along this line. Darwin records
a duck which in old age assumed the perfect winter and summer
plumage of the drake. Korscheldt also records a similar case.

Similar changes occur in fowls. The capon is well known.
He is larger, heavier, "softer" and more sluggish than the normal
male. He is said never to crow. Comb and wattles are poorly
developed, but the other secondary sexual characters may be fully
developed. Shattock and Seligmann, however, appear to have
found that this development takes place only when some testicular
material remains behind. But the age of the bird when castrated
may have an important bearing on this point. Stags when cas-
trated very young develop only a very small spike, though the
castrated adult retains large antlers. Likewise, a bull, castrated
when mature (called a "stag"), differs little from the normal adult.
The bull, castrated as a calf, on maturing differs in several
respects from the normal adult. Castration of the cockerel,
then, has two results. First, it may result in the failure of some
of the secondary sexual characters to develop; second, it brings
about certain other modifications not associated with secondary
sexual characters. There is an interesting peculiarity of the
capon. He is said to be capable of being trained to brood and
care for chicks. It is doubtful, though, if this is the acquirement
of a female sexual character, for he apparently does not become
broody, but merely is easily trained to care for the chicks.
Poulards are less well known, but may be expected to develop
male characters. Waterton records such a case, not resulting,
however, from surgical castration (cited by Darwin).

In mammals the general results seem to be quite similar. The
castrated female is commonly supposed to assume male char-
acters. The male on the other hand merely fails to develop
his secondary sexual characters, which remain in a more or less

RESULTS OF CASTRATION IN DUCKS. 49

youthful condition. In addition castration may have specific
results, not associated with secondary sexual characters. Tandler
and Grosz have recently made an extensive study of eunuchs
which may be briefly and incompletely summarized as follows:
The individual tends to retain youthful characters. Thus the
larynx is like that of a boy; the voice high pitched and likely
to break similarly to that of a boy at puberty; beard, body,
axial and pubic hairs sparse or wanting. The epiphyses remain
open for a long time, consequently the eunuch is usually very
long limbed. Fat tends to develop in certain parts of the body.
The skin is usually soft, poor in pigment, and of a peculiar yel-
lowish color. Facial wrinkles of a peculiar type are also found.
The effects of castration of male deer and cattle have been noted.
Less is known of the doe, but horns sometimes develop, appar-
ently as the result of abnormal ovaries (Rorig). Similarly in
the Hardwick breed of sheep, which are horned in the male
only, the horns fail to develop after castration (Shattock and
Seligmann). In reindeer, however, where both sexes are horned,
castration does not affect the development or renewal of the
horns. Other vertebrates have been little studied. Nussbaum,
however, finds that the nuptial organs of the male frog do not
develop on castration.

Experimental castration in insects has been studied by Oude-
mans, Kellogg, Regen, Meisenheimer and Kopec. The last two
have also transplanted the gonads from one sex to the other.
They agree that no modifications of the secondary sexual char-
acters occur. Wheeler, who has recently made an extensive
review of the subject, including its physiological side, reaches
the same conclusion.

In some other arthropods, however, very different results have
been obtained. Giard, Smith and Potts have studied parasitic
castration in various Crustacea. Their results may be summar-
ized thus : The secondary sexual characters of the female remain
unaltered, but the male becomes more or less modified in the
direction of the female. In extreme cases, after recovery, ova
have been found in the testes, the individual having become
hermaphroditic.

This brief review shows that castration varies in its influence

5O H. D. GOODALE.

in different groups. In a given species, castration remodels the
secondary sexual characters of only one sex. The case of the
stylopized Andrenidae recorded by Perez, in which reciprocal
modifications took place, seems in the light of present knowledge
to be somewhat doubtful (Wheeler) . At present, then, the effects
of castration, so far as it affects the secondary sexual characters,
agree with breeding experiments in indicating that one sex only
is heterozygous for sex.

The view that the male in birds owes his secondary sexual
characters to the addition of something to the female type, is
not supported by these experiments on ducks, for it is quite
clear that the female owes her color to the ovaries or something
associated with them, which previous to castration, suppressed
the male characters and which insured the development of her
own type of plumage. The female's plumage is obviously a
protective adaptation. From an evolutionary standpoint, it is
quite as conceivable, that selection should operate to pick out the
inconspicuously colored females, as that the greater vitality of the
male, or a selection of the more brilliantly colored males by the
females, should bring about an addition to the female's type of
plumage.

For the present, I think it makes little difference whether we
assume that the ovaries themselves or a modifier always trans-
mitted with the ovaries, prevents the development of the male
characters. The evidence of the possible presence of a modifier,
however, renders possible a somewhat different suggestion re-
garding the mode of sex-limited inheritance than has previously
been expressed.1 It involves also a number of possibilities re-
garding the inheritance of sexual dimorphism of plumage.

The Brown Leghorns are highly sexually dimorphic in plumage.
In cross breeding it is found that the female transmits this
color to her male offspring only, though the male transmits it
to all his offspring. Barring, likewise dimorphic though in less
degree, is transmitted is the same way. But that this is a
universal rule for poultry with sexually dimorphic plumage ap-
pears to be somewhat doubtful. The penciling of the Dark

^ex-limited inheritance appears to occur in ducks. A paper on this subject
in preparation.

RESULTS OF CASTRATION IN DUCKS. $1

Brahma female seems to be transmitted to the female offspring
only, without regard to whether the Brahma is mother or father
(Davenport). But there are some possible complications result-
ing from the particular matings made, which make this point
uncertain. Thus far, however, the study of sex limited inherit-
ance including the invisible factor "D" described by Bateson
for the Brown Leghorn and the castration experiments on ducks
all point in the same direction.

Some of the possibilities in the mode of inheritance of sexually
dimorphic plumage are as follows:

1. In the female we may assume that the plumage occurs as a
typical Mendelian heterozygote of male color and female color,1
the former being considered recessive. The male, then, must be
a homozygous recessive. But such assumptions do not agree
with the results of breeding experiments as mentioned above
for they show that in the female, a color factor occurs in only
half her gametes.

2. Among the various other ways of representing the mode of
inheritance of sexually dimorphic plumage, the following seems
the nearest approach to our present knowledge. Assume that
the male is homozygous for sex, the female heterozygous; that
the male color (or color factor) me also is homozygous in the male,
but in the female heterozygous, the other element of the pair
being an invisible modifier, M, which always couples with female-
ness; or, assume that the male color always couples with male-
ness. Then the male is me me, gametes all me; the female is

d1 c? cf

me M, gametes me, M. The breeding experiments mentioned
d" 9 cf 9

above afford the reasons for assuming that the male is homozy-
gous for sex, and the female heterozygous, and for the assumed
couplings. The castration experiments give the reason for as-
suming a male color and a modifier. The expression, male color,
however, is used in a purely descriptive sense. I do not wish to
be understood as giving it any special significance.

3. By the additional assumption of selective fertilization, the
male can be shown to be heterozygous instead of the female,

1E. g., male Brown Leghorn color and female Brown Leghorn color as such.

52 H. D. GOODALE.

thus: Male me me, gametes me, me-, female me M, gametes

cf 9 (*) d" 9 (*) 9 (*) 9 (*)

me, M. The only matings to be regarded as possible are me by
9(«) 9(«) 9 f*)

JW and me by me.
9 (*) 9 (*) <?

The # in parenthesis is given as an alternative mode of repre-
senting sex by the ^-element. This mode of representation would
bring the experimental results into line with Guyer's discovery
of an accessory in fowls.

The modifier assumed can hardly be specific for every type
of dimorphism, else in cross breeding the later generations of
females should show various modifications of the original female
type. Of this there is no evidence. The modifier may, of course,
be femaleness itself, in which case a somewhat different mode of
representation should be employed.

It is obvious that Guyer's discovery of an accessory in fowls is
contradictory to the results of experimental study which indicate
the correctness of the Bateson-Spillman theory of sex in fowls,
unless selective fertilization is assumed. Even if this be done, the
male color obviously cannot be transmitted by the idiochromo-
somes. There is, of course, the possibility that the .r-element is-
not a sex determiner at all, but merely a further indication of sex
dimorphism. On the other hand the discovery of cocks which in
cross breeding transmit some character to their female offspring
only, would disprove the Bateson-Spillman theory. Hagedoorn
seems to have found such a case, but he does not seem to have
noticed its implications. A reexamination is desirable.

Before proceeding further, a brief account will be given of
some modifications of sex.

Braem, while studying regeneration, removed the last 22 seg-
ments of 35 in a female Ophryothrocha. Later he found that the
ova in the anterior sections were degenerating and spermatozoa
developing. Hermaphrodites often occur in Ophryotrocha, but
in these the male germ cells always develop first, exactly the
reverse of the present case.

During the past summer, my attention was attracted to certain
changes in sex in maize, which is monoecious. The maize plant
consists of a single stalk, at the apex of which is the tassel bearing

RESULTS OF CASTRATION IN DUCKS. 53

the anthers. The female reproductive organs, forming the "ears,"
arise from axils of the leaves about midway of the stalk. In
some varieties secondary stalks called suckers may arise near
the base of the plant. They are usually devoid of ears, but
almost always produce a tassel.

Sometimes, however, kernels of corn develop on the tassels.
These appear to be always associated with a smut fungus. These
kernels were often very numerous on the tassels of suckers, oc-
casionally a fair resemblance to an ear being formed.

Stamens are less easily found on the ear but they may occur
at its apex. This may mean that the germ cells, which normally
would produce ova, develop into male cells, but since the ear is
morphologically the equivalent of a branch it may mean only
an abortive attempt at the development of the branch. Since
these observations were made, Professor Morgan has kindly
called my attention to the splendid work of Bleringhem on maize.
My results as far as they go entirely confirm Bleringhem, though
the alteration in sex patency appears to be brought about by
a different cause.

The case of Lychnis dioica, studied by Giard, is interesting.
'The young is hermaphrodite, but in certain individuals the
ovaries abort, in others the stamens remain rudimentary."1 Other
instances in plants are cited by Bleringhem.

These instances are sufficient to show that what seems to be
only one sex may really be both sexes with one of them suppressed.
Compare also Smith's studies on Inachus.

At the present time, the evidence available in regard to the
nature of sex and its determination is very contradictory. The
observations of Smith on parasitic castration in Crustacea, the
experiments of Morgan with Drosophila, and the studies of Wilson
and others indicate that the female is pure for femaleness, but
that the male is a sex hybrid of maleness (dominant) and female-
ness. Diametrically opposite evidence is afforded by the experi-
ments of Doncaster on Abraxis, followed by experiments on other
forms by several other investigators, which favor Bateson's the-
ory of a sex heterozygous female (femaleness dominant) and a sex
homozygous male. There are still other experiments which favor
neither view.

JThe quotation is taken from Wheeler's paper.

54 H. D. GOODALE.

In spite of these contradictions, sex is such a well nigh universal
attribute of living beings, that there would seem to be some
equally common differences between the sexes. It hardly seems
reasonable to suppose that femaleness, for example, is one thing
in one group and something else in another. But whatever the
common difference between maleness and femaleness as such, it
by no means follows that this difference of itself determines the
sex of an individual. Instead, other factors may be responsible.

The observations on Ophryotrocha, Inachus, Lychnis and maize
afford evidence of the presence of factors which control the
patency of sex, so that what is really hermaphrodite may appear
to be unisexual. The case of Siline inflata, cited by Correns
from Schultze, also points toward the presence of factors which
modify hermaphrodites. There are five different types, males,
andromonoecious (males and hermaphrodites on the same plant),
hermaphrodites, females and gynomoncecious. Moreover, ac-
cording to Correns and Shull, crossing hermaphrodites and uni-
sexual forms results in dominance of the unisexual forms.

If we assume that hermaphroditism, which is very much more
common in plants and the lower animals than in the higher ani-
mals, is the primitive condition, we have a basis from which
may be derived the present diverse modes of sex determination.
During the course of evolution, various changes have occurred
in the gametic constitution of some individuals or modifying
factors may have developed, which brings about our present
diverse modes for the determination of sex in the individual.
That the determination of sex, even from a Mendelian stand-
point, is not brought about by the simple interaction of maleness
and femaleness, is shown, I think, by the deviations from equality
in the sex ratios. In man, for instance, there is a slight but
constant excess of males at birth, whereas on the assumption
that one sex is heterozygous, the other homozygous, either equal-
ity should be expected, or that sometimes the number of female
births should exceed the male births.

Though our present evidence undoubtedly favors the view
that the female fowl is a sex heterozygote, it does not necessarily
follow that she is therefore a potential hermaphrodite. The
hermaphrodite, it is true, may be considered as a sex heterozygote,

RESULTS OF CASTRATION IN DUCKS. 55

but with both sexes patent. On the other hand maleness and
femaleness may each exist in the same individual in a homozy-
gous condition, which may be expressed in Mendelian form, thus:

22 forming gametes all *?• This formula may represent only

the primitive hermaphrodite, other formulae serving for existing
forms.

SUMMARY.

1. The castrated drake retains his secondary sexual characters,
except the ability to assume the summer plumage.

2. The castrated duck (female) assumes, more or less com-
pletely, the secondary sexual characters of the drake. The
change, however, is very gradual.

3. It is suggested that the female owes her color to the presence
of some modifying element, which prevents the development of
the male color. It is also suggested that the modifier may
sometimes be responsible for sex limited inheritance.

4. Cases in which the patency of sex has changed are pointed
out. This suggests that a common basis for the present contra-
dictory evidences regarding the determination of sex may be
found in the hermaphroditic condition.

LITERATURE.
Bateson.

'09 Mendel's Principles of Heredity. Cambridge, 1909.
Bleringhem.

'07 Action des traumatismes sur la variation et 1'heredite. Lille, 1907.
Braem.

'08 LTber die Aenderung des Geschlects bei Ophryotrocha. Anat. Anz., Bd. 33 ,

1908.
Correns.

'07 Bestimmung und Vererbung des Geschlechtes. Leip/ig, 1907.
Coues.

'96 Key to North American Birds, 1896.
Darwin.

Animals and Plants under Domestication, Vol. II.
Davenport.

'06 Inheritance in Poultry. Carnegie Inst. Washington, 1906.
Doncaster and Raynor.

'06 Breeding Experiments with Lepidoptera. Proc. Roy. Soc., 1906.
Goodale.

'10 Breeding Experiments in Poultry. Proc. Soc. Exp. Biol. Med., Vol. 7, 1910.
Guyer.

'09 Spermatogenesis of the Domestic Chicken. Anat. Anz., Bd. 34, 1909.

56 H. D. GOODALE.

Hagedoorn.

'09 Mendelian Inheritance of Sex. Arch. Ent. Org., Bd. 28, 1909.
Kellogg.

'04 Influence of the Primary Reproductive Organs on the Secondary Sexual

Characters. J. Exp. Zool., Vol. i, 1904.
Kopec.

'10 Ueber Morphologische und Histologische Folgen der Kastration und

Transplantation bei Schmetterlingen. Bull. Acad. Sc. Cracovie, 1910.
Meisenheimer.

'09 Experimentelle Studien zur Soma und Geschlechts Differenzierung. Jena,

1909.
Morgan.

'10 Sex Limited Inheritance in Drosophila. Science, Vol. 32, 1910.
Nussbaum.

'05 Innere Sekretion und Nerveneinfluss. Erg. Anat. Ent., Bd. XV., 1905.
Oudemans.

'99 Falter aus castrierten Raupen. Zool. Jahrb. Abt. Syst., Bd. XII., 1899.
Potts.

'09 Some Phenomena Associated with Parasitism. Parasitology, Vol. II., 1909.
Regen.

'10 Kastration und ihre Folgeerscheinungen bei Gryllus campestris. Zool.

Anz., Bd. 34, 1909; Bd. 35, 1910.
Rorig.

'99 Welche Beziehungen bestehen zwischen den Reproduktion Organen der
Cerviden und der Geweihbildung derselben. Arch. Ent. Org., Bd. 8, 1899.
Shattock and Seligman.

'03 Observations on the Acquirement of Secondary Sexual Characters. Proc.

Roy. Soc. London, Vol. 73, 1903.
Shull.

'10 Inheritance of £ex in Lychins. Bot. Gaz., Vol. 49, 1910.
Smith.

'06 Rhizocephala. F. F. Golf Neapel., Mon. 29, 1906.
Tandler and Grosz.

'10 Ueber den Einfluss der Kastration auf den Organismus. Arch. Ent. Org.,

Bd. 30, 1910.
Wheeler.

'10 The effects of Parasitic and Other Kinds of Castration in Insects. J. Exp.

Zool., Vol. 8, 1910.
Wilson.

'09 Studies on Chromosomes, IV. J. Exp. Zool., Vol. 4, 1909.

58 H. D. GOODAI.E.

PLATE I.

FIG. 4.1 Breast feathers from: A, male No. 29 in breeding plumage; B, oc-
casional variant from A. C, same male; D, a different male, both in summer
plumage. E-H, normal females; E, from female No. 26 in November; F, from
same female in August; G and H, from two other females. I-L, from female
No. 24, castrated; / and J, August; K and L, November; J and L were from
the upper part of the breast; / and A" from the lower part.

xThe buff or reddish bands in the originals of Fig. 4, D, F, K; Fig. 6, C, F, I,
J; Fig. 7, D, E, F failed of reproduction in the plates.

. BIOLOGICAL BULLETIN, VOL. XX.

PLATE

o

£

LU

H. D. GOODAI.E.

60 H. D. GOODALE.

PLATE II.

FIG. 5. Feathers from near the front end of keel. A, B, 29 cf, breeding plumage,
A being the more common type. C, D, 29 cf, summer plumage; C is usual type.
E, F, normal females; E, 26 9 ; .F, 41 9 . G-7, 24 9 , castrated; G, August; this
feather from middle keel region; H, I, November, H being the common type.

BIOLOGICAL BULLETIN, VOL. XX.

PLATE II.

I

LU

.

•->•>

»A*"

>r

Q

CQ

H. 0. GOODALE

62 H. U. GOODALE.

PLATE III.

FIG. 6. Feathers from dorsal surface near juncture of neck with body. A, B,
29 cf, breeding plumage; A, common type. C, 29 cf, summer plumage. D-F,
normal females; D, 26 9, November; F, 26 9, August; £.41 9. G-J, 24 9,
castrated; G, H, August; I, J, November; / is more common than J.

BIOLOGICAL BULLETIN, VOL. XX

PLATE III

LU

OQ

H. D. GOODALE.

64 H. D. GOODALE.

PLATE IV.

FIG. 7. Feathers from rump. A, 29 cf, breeding plumage; B, summer
plumage. C-E, normal females; C, 26 9 , November; D, 26 9 , August; £,41 9.
F-H, 24 9, castrated; F, August; G, H, November.

BIOLOGICAL BULLETIN, VOL. XX.

uu

H D. GOODALE.

66 H. D. GOODALE.

PLATE V.

FIG. 8. A-C, at juncture of wing and body: A, 29 cf ; B, 24 9, castrated;
C, 26 9 , normal. £> and E, breast feathers: D, young male; E, young female.
G, I, posterior to vent: G, 24 9 castrated; /, 29 cf , breeding plumage. F, H, J,
sides of body dorsal to vent: F, 24 9 castrated; H, 29 cf , summer plumage; J,
29 cf, breeding plumage. For normal female feathers compare Fig. 5, E, F.

BIOLOGICAL BULLETIN, VO1 . XX.

PLATE V.

-~-

.

H. 0. GOODALE.

THE DETERMINATION OF DOMINANCE AND THE
MODIFICATION OF BEHAVIOR IN ALTERNATIVE
(MENDELIAN) INHERITANCE, BY CONDITIONS
SURROUNDING OR INCIDENT UPON THE GERM
CELLS AT FERTILIZATION.

WILLIAM LAWRENCE TOWER.

A CORRECTION AND ADDENDUM.

In the preparation of the paper which appeared in this journal
under the above title, certain minor experiments were taken
from a larger series and combined to illustrate a general point
in the behavior of alternative characters in inheritance. Through
some unhappy oversight the data and plate of Ex. No. H. 410
was introduced into the printer's copy in place of the experiment
intended for the position and indicated by the remainder of
the paper. In this correction and addendum I have supplied
the necessary data of the proper experiment with the corre-
sponding illustration, which replaces Plate II. in the original
paper. I am indebted to Professor Cockerell for calling my
attention to this misplacement.

On page 294 of Vol. XVIII. of this journal, substitute the
following for the account given under Ex. No. H. 410.
L. signaticollis 9 X c" L. diversa.
Ex. No. H. 411.

When a cross was made between a female L. signaticollis and
a male L. diversa of exactly the same stocks as described in Ex.
No. H. 409, under the following conditions a quite different
result was obtained.

Food: normal and uniform.

T. R. H.

Day Av. 75° F. =*= 5° F. Day Av. 50 per cent. ±10 per cent.

Night Av. 50° F. ± 5° F. Night Av. 80 per cent. ±15 per cent.

In the FI generation of this experiment there was only one
class of adults and these were all of the female type and exhibited
a narrow range of variability. These when inbred, gave in F2

6;

WILLIAM LAWRENCE TOWER.

F*-

142

pair

? parent
L. signaticollis.

cf parent
L. diver sa.

This illustration replaces Plate II. of the original paper in Vol. XVII., p. 340.
Arranged to show the results obtained in crossing L. signaticollis 9 X cf L. diversa
under the conditions of Ex. No. H. 411. The differences in results shown between
H. 410 and H. 411 are due to factors not considered in the original paper.

DETERMINATION OF DOMINANCE. 69

the same type without the least trace of separation into categories,
and with the same low range of variability. When again inbred,
they gave in 7% the same type without the least trace of the
attributes of the male parent stock. Altogether I have repeated
this experiment six times with identical results. All have been
carried to F3 and some of them to F5 and Fe. In the plate which
accompanies this correction is shown the behavior of F\ and TV
On page 295, line 25, for Ex. No. H. 410 read Ex. No. H. 411

304

3 '

' 410

' 4ii

304 "

12 "

" 410 "

" " 411

304 "

14 " "

' 410

" " " 411

304

19 '

' 410

411

330 "

12 "

" 410 "

" " " 411

334 "

4 " "

" 410

" " " 411

Throughout the paper the references are to Ex. No. H. 411
instead of H. 410.

The description and plate of Ex. No. H. 410 are correct as
are the conditions of experiment as far as given. This experi-
ment, however, is from a second series of cultures parallel to
the one given, but in which there are other factors involved,
which in H. 410 are productive of a typical Mendelian behavior.
I do not at this time care to make any statement of what these
factors are, nor of their relation to the behaviors given in the
H. 409, H. 411, H. 409/11 series which are the simplest and most
easily presented series obtained in the crossing of L. signaticollis
and L. diver sa.

DEPARTMENT OF ZOOLOGY,

THE UNIVERSITY OF CHICAGO,
December 10, 1910.

Vol. XX. January, ign. No. 2

BIOLOGICAL BULLETIN

CERTAIN HABITS, PARTICULARLY LIGHT REACTIONS,
OF A LITTORAL ARANEAD.

THOS. H. MONTGOMERY, JR.

This tiny beach-comber has been identified by Mr. Nathan
Banks as Grammonota inornata (Emert.), a species of theridiid,
for which I am much indebted to him. It lives at Woods Hole,
Mass., and in its vicinity, upon the sandy beaches along Vineyard
Sound and Buzzards Bay. Here it is found, however, in only
small scattered colonies for it occurs only on those sand beaches
where the eel grass, thrown up by the sea, lies in more or less
permanent lines upon the beach. Thus it does not exist upon
gravelly or rocky beaches, nor yet upon sandy beaches devoid
of the eel grass. The area of its distribution is limited to this
cast-up eel grass, from a line slightly above high tide level land-
ward as far as the patches of eel grass extend, a distance rarely
exceeding ten or twelve feet and determined more or less by the
degree of slope of the beach. In this peculiar home it obtains
perpetual moisture and darkness, and does not migrate further
up the beach nor into the salt marshes. In the same places
occur a variety of animal forms in great individual abundance.
Of this fauna the gammarids constitute the only truly marine
element, the others being mainly terrestrial forms: certain other
spiders, of which a drassid and the young of Trochosa cinerea
are most abundant, staphylinid and chrysomelid beetles, an
acarine, a pseudoscorpion and a chilopod. These are all normal
constituents of this fauna, not waifs drifted in by the sea.1

My interest was drawn to them by the observation that when

JThe elements of the "upper beach fauna" have been interestingly characterized
by Davenport, "The Animal Ecology of the Cold Spring Sand Spit," Dec. Publ.
Univ. Chicago, 1903. But he does not mention the particular species we are now
considering.

71

72 THOS. H. MONTGOMERY, JR.

they are disturbed by the removal of the eel grass, they all run
directly landward provided only the sun is shining from an angle.
Fifty or a hundred or more of these spiders may be set into
commotion by the lifting up of a bunch of the eel grass, and
under the conditions stated the large majority, sometimes all,
of them invariably run in a straight course right up the beach and
do not stop until they reach shelter beneath another mound of
the eel grass. This habit is the more striking because none of
the other species that live with them exhibit any such regular
landward migration.

Direction and velocity of the wind have nothing to do with
determining this course of running, for the landward migration
occurs whether the wind blows upon the spiders from in front,
or behind or from the side, or when there is no wind. Further,
the spiders are not guided by any moisture sense, for (i) they do
not regularly run landward when the sun is obscured, and (2) if
the observer pours out sea water at a higher level and lets if
flow down towards the spiders, they nevertheless continue their
landward course. Again, this is not a geotropism, a tendency
of the spiders to run up hill, nor an orientation to the rolling of
sand grains. For the observer can dig a deep trench to land-
ward of the spiders, and they will go down its declivity and
up its opposite side without changing their course; or one
can implant a board vertically in the sand in front of them, when
they will climb up one surface of it and down the other without
changing their general direction of movement.

Therefore the landward course of these spiders is not caused
by any influence of wind or moisture or slope of the beach, but
is probably a light reaction as the following observations would
show.

When the sun is shining and when its rays come from an
angle, at any time but noon, the spiders regularly run up the
beach; this I have tested many times and on different beaches
facing different points of the compass. That is, the spiders run
up the beach whether the sun shines in front of, behind or to one
side of them. But at noon, when the sun is nearly vertical,
they run in all directions ; and that they do also when the observer
on a sunny day shades them with an umbrella.

LIGHT REACTIONS OF A LITTORAL ARANEAD. 73

Then a number of experiments were made to test the reactions
of the spiders to the rays of an ordinary student oil lamp placed
one foot or a foot and a half away from them, all these experi-
ments being made at night. Numbers were caught and placed
in small vials, which can be done without injury to them for
they are hard-bodied. Then a piece of smooth white paper was
laid horizontally upon the table, and kept free from sand or
other particles; on this paper four quadrants were marked,
the one nearest the lamp marked A, the one most distant from
it D and the other two quadrants C and B. A vial was then
uncorked and the spiders allowed to drop in succession gently
upon the center of the paper, each falling slowly on its own
drag line. With a pencil the course of each spider was then
marked upon the paper, which made a permanent record of all
movements of all individuals. On different nights a total of
107 spiders were thus tested, with the following results: 3 ran
into quadrant A, directly towards the light; 45 ran into quadrant
D, directly away from the light; 28 turned into quadrant B
and 31 into quadrant C, these accordingly at right angles to the
direction of the light. This proves negative phototropism to
the rays of an oil lamp, for only 2.8 per cent, of the individuals
ran towards the light. In these experiments it was purely a
matter of chance how the spider was oriented when it first
touched the paper. Of those that happened to touch it facing
the light almost all moved through an arc of a circle so as to get
into another direction and then continued their courses in almost
straight lines; there was a marked turning away from the source
of the light.

Another experiment was made with diffuse daylight. Eleven
spiders were dropped in similar manner upon a piece of white
paper placed on a table in the northwest corner of a room, sun-
light entering the east window but not reaching the table. Ten
of the spiders turned directly away from the sunlight, and one
at right angles to it. This was then negative phototropism.

In a third experiment spiders were dropped upon a piece of
white paper on which the sun shone directly, the sun being then
near the meridian (10.45 A.M.). In this case ten moved towards
the sunlight, one at right angles to it, and eight away from it.

74 THOS. H. MONTGOMERY, JR.

This result seems at first sight to be contradictory to the pre-
ceding experiments. But I believe it is explainable by the nearly
vertical position of the sun's rays; for on the beach, at noon when
the sun is nearly vertical, the spiders do not run in any regular
direction.

It thus seems that these spiders are decidedly negatively
phototropic to lamplight and to diffuse daylight, when these
impinge upon them from an angle; but that when sunlight falls
upon them nearly vertically they do not orient themselves to it.
How is it then on the beach that the spiders run so regularly
landward whenever the sun is shining from an angle, irrespective
of the position of the sun, which may be in front or behind or to
the right or the left of them? One would not suppose it to be a
case of negative phototropism when they run straight towards
the sun. Yet I believe it is this nevertheless, in that the spiders
may orient themselves not to the sun directly but to the light
from the water. For the area of light in the sky and clouds
above the water is certainly greater, and the light more intense,
than that to landward, owing to the great amount of reflection
from the water. It is well known that a man becomes more
sunburnt upon the water, a test that the sunlight is more effective
there. It would seem that only in this way can we correlate
the experiments made with the oil lamp and with diffuse sunlight,
with the observations recorded upon the sea beach. That animals
react to area as well as to intensity of light is now well known.
Thus G. H. Parker1 showed that Vanessa, when in the open,
always comes to rest with its head away from the sun, and that
it thereby "reacts positively to large patches of bright sunlight
rather than to small ones, even though the latter, as in the
case of the sun, may be much more intense." Cole2 has con-
firmed Parker's results, and concluded from his own pains-
taking observations on several forms, that animals with direc-
tion eyes, as those without eyes, respond to light intensity only;
and that animals with image-forming eyes when positively

Phototropism of the Mourning Cloak Butterfly, Vanessa antiopa Linn.,"
Mark. Anniv. Vol., 1903.

2 "An Experimental Study of the Image-forming Powers of Various Types of
Eyes," Proc. Amer. Acad. Arts and Sci., 42, 1907.

LIGHT REACTIONS OF A LITTORAL ARANEAD. 75

phototropic react to the size of the luminous field or to definite
objects in the visual field.

Our Grammonota probably possesses an image-forming eye,
but it is negatively phototropic. Whether it reacts to intensity
of the light or to its area remains to be determined. I have
made no attempt to decide whether it reacts rather to the light
rays than the heat rays, and I was not interested to determine
this, for in the state of nature heat and light, so far as they affect
these animals, are probably always associated.

Under natural conditions the masses of eel grass that shelter
these spiders would be disturbed only by unusually high tides,
perhaps occasionally also by violent winds. If the sun were
shining during such a disturbance the spiders, from their negative
phototropism, would attempt to run landward, and thus some
of them might escape the action of the waves; but it is question-
able whether this light reaction would be of any great benefit
to them at times when the waves play havoc on the beach.

From the scattered distribution of these spiders one would
suppose they might be readily transported by water. I ac-
cordingly made a few experiments to ascertain how sea water
affects them. A number were dropped upon clean sea water
placed in a glass dish; they were able to stand upon the surface
film, but not to run unless there were fine dust particles upon it.
But so soon as the water was violently agitated they sank below
the surface and were unable to rise again. Therefore they
could not long remain upon the top of a wave, but would become
quickly submerged unless they clung to some floating vegetation.
This is because the body is only slightly pilose. They become
quiet after a few minutes of submergence, as though partially
suffocated, yet they can withstand submergence for some hours.
Thus I kept seven beneath water for five and a half hours, then
placed them upon blotting paper to dry, when four revived;
and I kept another lot beneath water for sixteen hours, and one
of these revived after drying.

A number were placed in isolated vials, each with a few
drops of sea water to furnish the necessary moisture, in order
to observe the cocooning, but I was not so fortunate as to see
this process. The cocoon is lenticular, snow-white and relatively

76 THOS. H. MONTGOMERY, JR.

very large — its diameter considerably greater than the length
of the spider. Nine cocoons were made in these vials, by as
many females, and all at some time between 10:30 P.M. and
7 A.M., the time of cocooning would then seem to be nearly
morning. On previous occasions I have indicated how regular
this time is in spiders, and how it varies with different species.
Males are abundant through the summer months, but the
mating was seen only in part although many of both sexes were
kept in observation cages. A male and female were seen to
move directly towards each other, and on meeting each elevated
its head region so that the line of the body made an angle of
about 45° with the floor, when the male extended his cephalo-
thorax somewhat beneath and to one side of that of the female.
This attitude is similar to that of Dictyna.

UNIVERSITY OF PENNSYLVANIA.

PRELIMINARY NOTE ON THE ECOLOGY OF THE
EARLY JUVENILE LIFE OF THE UNIONID.E.1

F. B. ISELY.

During the past four years the writer has given considerable
time to the field study of the Unionidae. A number of puzzling
questions have arisen as to the relative importance of the various
ecological factors that contribute to the environmental complex
of these animals. Among the most important of these factors,
we may mention bottom conditions, water content, stream vol-
ume, stream fluctuations, relation to fish and natural enemies.

In connection with this work, much difficulty was experienced
in finding young mussels for study and experimentation. I have
collected many specimens from the size of a nickel to a quarter,,
but mussels under the size of a dime have been rare. A number
of experienced field workers have spoken to me of a similar
difficulty in finding juvenile specimens.

In order to avoid any misunderstanding, I may state that by
"early juvenile life" I mean the period following the time when
the mussel completes the parasitic stage and leaves the fish to
lead an independent life, until it is about the size of a dime or
about fifteen millimeters in length. This would cover, in most
species, approximately the first year of independent existence.
Other periods may be designated as later juvenile and adult life.

The glochidium stage has received considerable study. Our
knowledge of the parasitic period has been recently cleared up
and extended by the careful investigations of LeFevre and Curtis.2
The later juvenile and adult life has received the attention of
hundreds of students. The scattering and meager references
in the literature, to the early juvenile stages, give us little infor-
mation on this important period in the life history of the mussel.

During the past summer, while working on the Red River Mussel

1 Published by permission of the U. S. Commissioner of Fish and Fisheries.

2 "Reproduction and Parasitism in the Unionidse," George LeFevre and Winter-
ton C. Curtis, Journal of Experimental Zoology, Vol. IX., No. i, September, 1910,
pp. 79-115.

77

78 F. B. ISELY.

Survey for the Bureau of Fisheries, I found a fairly good number
of species in the early juvenile period of development. The
distribution and ecology of these young mussels seem to me to
be of special interest. Furthermore, the facts in regard to their
habitat seem to clear up some ecological perplexities concerning
the adult ecology of the Unionidse.

Thirty-two specimens are here especially considered, these
were attached to rocks and pebbles by a functional byssus.
These mussels, representing nine species, were found in three
different rivers and in four localities. The situations, however,
were similar. Twenty-nine of these specimens weigh less than
five decigrams; of this number twelve weigh under one decigram,
and ten are under nine millimeters in length.

The thirty- two specimens represent the following species:1
(i) Lampsilis luteola, two; (2) Lampsilis fallaciosa, one; (3)
Lampsilis parva, four; (4) Lampsilis gracilis, three; (5) Plagiola
elegans, one; (6) Plagiola donaciformis, sixteen; (7) Anodonta
imbecillis, two; (8) Ptychobranchus phaseolus, two; (9) unnamed
species, one.

The first of these specimens was secured on August 20, 1910,
in the Kiamichi River, near Fort Towson, Oklahoma. While
working in a gravel bed, Owen Home, one of our party, called
attention to an unusually small mussel attached to pebbles by
a byssal thread. After a search for about two hours we secured
eight specimens with byssus, representing four species. These
specimens were found in a coarse gravel bed, the pebbles being
from one fourth to one inch in diameter. The water was fairly
swift, and from one to two feet in depth. The byssus of these
specimens was variable in length, sometimes several inches long,
and often connecting three or four small pebbles. Sometimes
the byssus spread into several branches at the place of attach-
ment. The young mussels were best secured by taking up hand-
fuls of gravel and looking for the thread. The byssus is strong
enough to support the mussel in a rapid current, and will sustain
the weight of a number of small pebbles, without breaking.

On August 30, 19 io; five more specimens were secured in the

1 Most of the specimens have been examined by F. C. Baker and L. S. Frierson,
making sure of the identifications.

ECOLOGY OF EARLY JUVENILE LIFE OF UNIONID^. 79

Little River, near Garvin, Okla. One of these specimens, found
by E. C. Johnston, of our party, was attached by the byssus to
the shell of a large Quadrula pustulosa. An observation of this
kind is described by Kirtland.1

During the afternoon of August 30, we again investigated a
portion of the Kiamichi River, near Roby, Okla. Here we secured
fourteen specimens; ten of these I secured in about half an hour,
one time bringing up three specimens with a handful of gravel.
Here again the environment was typical, fairly swift water,
coarse gravel, and rocks.

A search for young mussels was also made in Blue and Boggy
rivers, but failed to yield specimens with functional byssus.
However, a number of mussels under twelve millimeters were
secured, representing Quadrula pustulosa, Quadrula lachrymosa,
Quadrula coccinea. In the Washita River, near Davis, Okla.,
on September 2, we secured the last of our young mussels for the
summer. Five specimens were found in swift water about two
feet deep. One of these was under three millimeters in length
and weighed .005 of a gram.

In the following table I show species, locality, weight in grams,
length, height, and breadth in millimeters, of ten of the thirty-
two specimens:

Species.

Weight.

Length.

Height.

Breadth.

Locality.

i. L. luteola

.41

IA ?

92

t o

Kiamichi R Aug 30 1010

2. L. fallaciosa . . . .
^. L. fiarva

03 =;

13-5

7

6
^ 8

3-4

2 1

Kiamichi R., Aug. 20, 1910.

4. L. gracilis

.47

22 ?

90

A I

5. A. imbecillis . . . .
6. P. phaseolus ....
7. P. elegans

•055

•IS
.I7C

7-i
12.4
0 6

3-5
5-6

6 Q

I.I
2.7

A <S

Kiamichi R., Aug. 30, 1910.
Little R., Aug. 30, 1910.
Kiamichi R Aug 30 1910

8. P. donaciformis .
9. P. donaciformis.
10. Unnamed sp. . . .

.005
.01
.014

2.9

5-6

5

i-5
3
3-3

i.i

1.9
2.3

Washita R., Sept. 2, 1910.
Kiamichi R., Aug. 30, 1910.
Little R., Aug. 30, 1910.

The facts noted above are closely related, not only to the
ecology of the juvenile mussel, but also to the ecology of the
adult.

i . They indicate the conditions essential for the most successful
growth and early development of the Unionidae. This kind of

^'Fragments of Natural History," J. P. Kirtland, American Journal of Science,
Vol. 39, 1840, pp. 166-168.

80 F. B. ISELY.

an environment gives a constant supply of oxygen, and sufficient
food; is frequented by suitable fish; is free from shifting sand
and silt accumulation. Those mussels that drop from the fish
in these favorable situations develop in large numbers, while
the less fortunate, that drop in shifting sand and silt, die early.

2. In the study of the ecological factors that are inimical to
mussel life, more attention should be given to the consideration
of the juvenile habitat. Absence of gravel bars and stony
situations may sometimes explain the scarcity of the Unionidse
in certain streams and lakes where frequently water content has
been thought the chief unfavorable factor.

3. It is a well known fact that in many streams certain stretches
of mud-bottom are found loaded with mussels, while other areas,
in the same stream, equally favorable from the standpoint of
the habitat of the adult mussels, have only scattering specimens.

This distribution of the adult may be explained by the as-
sumption1 that the average mussel seldom travels far up or down
the stream from the place where it begins successful develop-
ment. Stretches favorable for juvenile development thus come
to be the centers of dispersal in the streams where they occur.
As a result, areas of mud-bottom near these favorable habitats,
become loaded with mussels by migration.

4. In the study of the life history of the Unionidae, we may
consider the embryonic, the glochidia, the parasitic, the early
juvenile, and the adult as distinct periods for separate and special
study.

UNIVERSITY PREPARATORY SCHOOL,
TONKAWA, OKLAHOMA.
November 25, 1910.

1 This assumption is fairly well established, by experimental study that I have
been carrying on for some time, and will be discussed in a later paper.

A STUDY OF PEDAL LACERATION IN ACTINIANS.

LEWIS R. GARY.

Since the appearance of Andres's ('82) classical paper "In-
torno alia Scissiparita della Attinae," the only accounts of the
process of pedal laceration have been very casual, or have been
given only for the sake of comparison with the phenomena of
induced regeneration. Carlgren ('04), ('09), who alone has
studied this form of reproduction in the genus Aiptasia, makes
mention only of the sequence of the appearance of the mesenteries
and tentacles, but gives no general account of the process, nor
does he mention any of the histological features.

Since the last-mentioned phase of the subject is of considerable
interest as a comparison with the development of normal em-
bryos, it has seemed advisable to publish the following account
to supplement the previous contributions to the knowledge of
this type of reproduction which is so common among the
actinozoa.

The observations herein recorded are based upon the study of
four species of actinians: Aiptasia pallida Agg., A. tageies D.
& M., A. annulata Andres and Cylista leucolena Agg.1 Of these
species A. pallida and Cylista were obtained in abundance at
Beaufort, N. C.,2 during the summer of 1903 and 1904, where
they were studied both in their natural environment and in
aquaria in the laboratory.

Specimens of the first-mentioned species were transferred to
Baltimore and kept during two years in aquaria in the zoological
laboratory of the Johns Hopkins University. By means of the
diatom method described by Dr. Caswell Grave for rearing
echinoderm larvae, the actinians were kept in good condition

1 Professor McMurrich informs me that he has conclusive evidence that the
form described by Verrill from Beaufort, N. C., as C. leucolena is not the C. leucolena,
of A. Agassiz, but in the absence of any published statement to that effect Verrill's
terminology will be used.

- 1 am indebted to the officials of the U. S. Bureau of Fisheries for the use of a
table, and equipment at their Beaufort, N. C., laboratory during the two seasons
above mentioned.

81

82 LEWIS R. GARY.

during this time and they continued to reproduce by laceration
throughout the year, although no embryos were ever set free
during either of the breeding seasons.

The specimens of A. tagetes studied were obtained in 1909,
during a stay at the Bermuda Biological Station.

A. annulata was obtained during the past summer while the
writer was at the laboratory of the Carnegie Institution of
Washington at Dry Tortugas, Florida.

The two last-mentioned forms were studied in aquaria in the
laboratory only.

Material of the four species was preserved, after being narcot-
ized in magnesium sulfate, in a number of fixing reagents; of
which Petrunkewitsch's fluid and Bouin's fluid gave the best
results. All of the Aiptasmt cut readily after being imbedded
in paraffin in the usual manner, while the mesoglea of the Cylistas
proved so refractory that double imbedding in celloidin and
paraffin was necessary in order to secure an unbroken series of
sections.

In common with the tissues of most actinians, many of the
hsematoxylin stains were highly selective, staining the ectoderm
much more heavily that the endoderm so that by using an alco-
holic solution of eosin as a counter stain after Delafield's hsema-
toxylin there resulted an absolutely certain color differentiation
between the two tissues by means of which any question of the
origin of a certain structure could be determined by its staining
reaction.

GENERAL ACCOUNT OF THE PROCESS OF LACERATION.

As has been described by Andres (1882), the beginning of the
process of laceration is characterized by a certain part of the
base of an actinian becoming very firmly attached to the sub-
stratum while the animal as a whole moves away from the point
which is thus relatively immovably attached. As a result the
tissues about this point become strongly stretched and attenuated.
With the continuation of the contraction, there finally comes a
rupture of the tissues at some little distance from the free border
of the disc so that the resulting laceration piece has usually an
appearance such as is shown in Fig. i, Plate I., which represents
a fragment from the base of a specimen of A . pallida immediately

PEDAL LACERATION IN ACTINIANS. 83

after its separation from the parent individual. The acontia,
which in laceration pieces of this form are at first so prominent,
are later either withdrawn inside the cavity of the fragment, or
else they are broken off before the open side of the piece has closed
to any considerable extent.

Immediately after the separation of the fragment there takes
place a rolling-in of the free edges, due principally to the elasticity
of the mesoglea, so that for some little distance within the cavity
is lined with ectoderm (Fig. 7. PI. II.). For some hours after
the separation of the fragment there is no visible external change
other than an increasing loss of color about the orifice of the tear
where the laceration piece was separated from the parent indi-
vidual.

The next noticeable change appears as a turning upward of that
portion of the fragment in which is situated the original opening,
which by this time has become markedly reduced in extent; al-
though, contrary to the results of Hazen (1902), who worked with
fragments obtained by cutting off pieces including portions of the
column wall and pedal disc from adult specimens of Sagartia
IUCCB, in none of my specimens did the opening become entirely
obliterated. The new growth which brings about this change
in the shape of a laceration piece is much the more rapid on the
lower-pedal disc-side of the aperture. So that for some time the
elevated area is situated near the former internal edge of the
fragment. The wall that has come from the up-growth from
the pedal disc is nearly perpendicular, while the opposite side of
the fragment slopes away much as in its former manner (Fig. 2,
PI. I.).

In a stage of development such as is represented in Fig. 2
there has already appeared the beginning of the tentacle buds,
and internally, of course, the earliest mesenteries.

In a later stage, as in Fig. 3, PI. I., the readjustment of mate-
rials had proceeded farther in what is practically the same course.
The fragment as a whole has become much higher. The oral
end has nearly the same diameter, while farther down toward
the pedal disc the diameter is relatively much decreased. At
the base, however, the fragment still covers practically the same
area as when it was torn off from the parent animal. When

84 LEWIS R. GARY.

viewed from the side, as in this figure, the lines corresponding
to the insertion of the old mesenteres are seen to extend only
a short distance upward from the base and to be less distinct
toward that side which is most nearly perpendicular, i. e., which
has arisen from the tissue which was originally part of the pedal
disc The tentacle buds have increased very slightly in size,,
but at the oral end of the fragment, which now has the appearance
of a partly contracted Aiptasla there can be made out through
the almost transparent column wall, the outlines of the stomo-
deum and the lines of insertion of the developing mesenteries.

Figs. 4 and 5 were, in order to show the arrangements of the
tentacles, drawn in a nearly oral view, so that the upper part of
the column is not visible. The fact that the lines of insertion
of the old mesenteries appear to extend nearly the whole length
of the column is thus to be explained by the foreshortening in
the figure and not by any difference as regards the mesenteries
between these specimens and that shown in Fig. 3. In the speci-
men shown in Fig. 4, there are eight tentacles, of which one is
considerably longer than the other seven, all of which are of
approximately the same length. In practically all instances one
tentacle outstrips the others in its development so that for a
considerable time it may be distinguished from the others. On
the other hand there seems to be no such definiteness in the de-
termination of just what tentacle shall first appear as there is in
many actinian embryos where the tentacle above the endoccele
between the two dorsal directive mesenteries almost invariably
makes its appearance first.

In later stages there is often a very marked difference in the
sizes of the tentacles of the first series, as in the specimen repre
sented in Fig. 5. Here a single tentacle has far outstripped the
others, while of the remaining seven three are decidedly larger
than the remaining four. In Fig. 5 the tentacles of the second
set are seen to be arising in pairs at each side of several of those
of the first set. While that tentacle of the first set which is
uppermost in the figure is by far the longest, there is on either
side of it a single short tenacle of the first set, and no indication
of the paired tentacles of the second series. The tentacles on
the opposite side of the disk have already acquired their definite
shape and appearance.

PEDAL LACERATION IN ACTINIANS. 85

The young actinian shown in Fig. 6 has already acquired the
characteristic appearance of all small specimens of this species,
so that, from an external examination, it would be impossible to
determine whether it had come directly from an egg or from a
laceration piece. As compared with any of the figures of earlier
stages the proportionate diameter of the column is very much
reduced. This has been brought about by a continuation of the
same processes that are shown at an earlier stage in Fig. 3. The
basal portion of the column is now of no greater diameter than
the upper part, while the whole animal has become considerably
elongated. The first eight tentacles are now of practically the
same length as are also the eight pairs of secondary tentacles.
Internally the stomodeum can be seen to have increased greatly
in length. The mesenteric filaments, not previously distinguish-
able, have now become very prominent structures on the first
eight mesenteries. The column wall has become thin and has
lost its pigment clear to the base through which are seen the points
of insertion of the first eight mesenteries, while those mesenteries
that came over from the parent actinian in the laceration piece,
have entirely disappeared, even at the extreme base of the young
specimen.

In following through the development of any single laceration
piece from the time of its separation from the parent until it
becomes a typical young actinian it is apparent that there has
been very little, if any, increase in its bulk, and that the change
in shape has come about almost entirely through the readjust-
ment of the materials already present and not to any considerable
extent through the metabolism of new material.

Frequency of Reproduction by Laceration among the

Forms Studied.

The following table shows the frequency of this method of
reproduction among the four species studied and the number of
young that has arisen from a single parent individual within
a given time.

In another instance many thousand individuals of A. pallida
were examined casually, but without any records being kept,
and here again it was found that the percentage of those which

86

LEWIS R. GARY.

had undergone laceration was very high. In fact this form of
reproduction was much more frequent in the ^last-mentioned
instance — on the government jetties at Cameron, La., than in
the case of the same species from which the records were made
at Beaufort.

Species.

Number
Examined.

Number
that had
Given Rise
to Lacera-
tion Pieces.

Number of Lacera-
tion Pieces.

Locality.

C, leucolena

30O

292 \

Beaufort, 1903.

C. leucolena

I.OOO

QA2 J

11.32 (0-30)

Beaufort, 1904.

A. pallida

ISO

78

6 (0-14)

Beaufort, 1904.

A tagetes

22O

187

10.4 (0-26)

Bermuda, 1909.

A. annnlala

52

20

3 (o-7)

Tortugas, 1910.

INTERNAL CHANGES DURING LACERATION.
Under this head will be discussed the following topics:
(a) The formation of the stomodeum.
(&) The development of the mesenteries.

(c) The development of the mesenteric filaments and acontia.
(J) General histological changes in the tissues during develop-
ment.

(a) The Formation of the Stomodeum.

As mentioned previously in the general account of the process
of laceration, the first noticeable change in the appearance of
a laceration piece after its separation from the parent consists
in the rolling in of the edges of the torn side of the fragment,
so that from the first the opening into the cavity of the fragment
is lined for some distance with ectoderm (Fig. 7, PI. II.). At
first the ectoderm extends forward for only a short distance on
the upper side of the opening, while on the lower side the inrolling
has progressed considerably farther.

In a section through a stage such as is shown in Fig. 2, PI. I.,
the stomodeum has the appearance such as is shown in Fig. 8,
PI. II. Here the free edges about the orifice of the fragment
have become thinner, the mesoglea is reduced to a thin sheet
of tissue and the cavity of the stomodeum has become vertical
instead of nearly horizontal as at first. At the lower, free, border
of the stomodeum the ectoderm has become turned outward
and extends around the border of the endoderm onto the internal
wall of the stomodeum.

PEDAL LACERATION IN ACTINIANS. 87

In Fig. 9, PI. II., which represents a section through a stage
slightly older than that shown in Fig. 3, PI. I., the stomodeum
has reached nearly its normal length and, on the left side in the
figure, its ectodermal lining is shown extending down along the
border of a mesentery to form the mesenteric filament. On
the opposite side where the section passes through the chamber
between mesenteries, the ectoderm forms a flat plate extending
across the free border of the stomodeum.

The Development of the Mesenteries.

Andres (loc. cit.}, in his figures illustrating the rearrangement
of the mesenteries in the young fragments, shows only the pedal
disc and apparently understood that the mesenteries which come
over in the laceration piece from the parent were rearranged to
become the permanant mesenteries of the young actinian. He
makes no mention of any new growth of mesenteries or stomo-
deum at the oral end of the fragment as it becomes elongated
and assumes the characteristic shape of an actinian.

Carlgren, 1904, 1909, takes up to a considerable length the
question of the arrangement of the mesenteries in laceration
embryos. He distinguishes four types of arrangement of the
first twelve mesenteries in such embryos: depending both on the
sequence in the appearance of the mesenteries, and on whether
or not any of these mesenteries, are situated in old tissue. "Was
erstens die Entwicklungstypus der Mesenterien in den Lacera-
tionstiickchen anbelangt, so haben die 1904 und die hier oben
mit geteilten Untersuchungen gezeugt dass es verschiedene
Variationen derselben giebt. Wie gross die Variation auch ist, so
kann mann doch deutlischerweise in dem Lacerationstiickchen ,
d. h. in solchen Stiickchen, die einen Teil der Fussscheibe und
einen der Korperwand umfassen (Text Fig. I , b] , drei Verschiedene
Haupttypen fur die Mesenterien Entstehung unterscheiden, und
zwar erstens einen bilateralen mit drei primaren, vollstandigen,
gleich, orientierten Mesenterienpaaren (II., 4.), Zweitens ein
bilateralen mit nur zwei ebenfalls gleich (II., i), angeordneten
Paaren und schliesslich drittens einen birddialen (II., 5 and 8),
welche letzteren zwei Neubildungszonen enthalt, die in ihren
am besten entwickelten Gestalt (No. 8), Spiegelbilde zu einander

LEWIS R. GARY.

sind, indem jede aus zwei bilateral angeordrjeten, vollstandigen
Mesenterienpaaren besteht." Besides these three types Carl-
gren recognizes a mixed type which cannot be put in any one
of the above-mentioned classes.

All the specimens of Aiptasia studied by Carlgren fell, with
three exceptions which were very irregular in their growth, into
his group three, the biradial type, in which there are two zones
of new growth. All of the specimens of this genus that I have
sectioned show this arrangement of the newly formed mesenteries,
which, thus, seems to be characteristic of the genus Aiptasla.

The specimens of Cylista also showed the same arrangement of
the first mesenteries.

H. V. Wilson, 1888, states that in Macina areolata the ceso-
phageal invagination is at first centrally placed in the larva, but
that soon "it begins to travel toward one side of the larva."
This lateral motion of the stomodeum continues until that organ
comes to lie close against the outer wall of the larva. Finally
as the lateral movement progresses — from above downward —
the endoderm is pushed before it until the two ectodermal layers,
of the stomodeum and of the body wall, are in contact, or sepa-
rated only by the thin layer of mesoglea secreted between them.
Soon following this stage,- . . . "the oesophagus has moved away
from the surface ectoderm, but while doing so has remained
connected with it by a band of supporting lamella." The mes-
entery arises by the endoderm growing up into the space between
the stomodeum and the outer wall of ectoderm as the stomodeum
is drawn toward the side opposite the first mesentery, for the
formation of the second one.

Appellof, 1900, mentions the same sort of displacement of the
stomodeum previous to the formation of the first mesentery.
He states, on the other hand, that: "Die exzentrische Lage des
Schlundrohres ist in alien Entstehungsphasen desselben konstant
und deutlich nachzuweisen und ist — so nehme ich wenigstens
an — auf ungleichseitiges Wachstum der Larve zuriickzufuhren."

In laceration embryos the stomodeum is from its first appear-
ance excentric in position. As the change in the form of the
piece takes place so that the original orifice, where the piece was
torn off from the parent, becomes carried up toward the highest

PEDAL LACERATION IN ACTINIANS. 89

part of the fragment and becomes smaller and more rounded, the
stomodeum on that side which has come from the growth of the
pedal disc comes into contact with the body wall. The first
mesentery makes its appearance at the point where the forming
stomodeum is in contact with the column wall; although in none
of my sections is there evidence of any marked displacement of
the endoderm such as would be necessary to bring the two meso-
gleal surfaces into contact. As is always true in the formation
of a mesentery, the growth of the mesogleal lamella here begins
at the extreme oral surface and proceeds downward along the
stomodeum and column wall. In laceration embryos the endo-
derm is carried down on both sides of the mesogleal lamella so
that from its earliest appearance, the mesentery is structurally
complete. As the growth leading to the acquisition of the
characteristic actinian shape of the laceration piece goes on the
stomodeum becomes situated more centrally in the oral disc.
While this adjustment is taking place the first mesentery becomes
lengthened in the horizontal direction to keep pace with the
change in the position of the stomodeum.

The method of the formation of all those mesenteries formed
after the stomodeum has become central in position is the same.
The first indication of their appearance is seen as a slight ridge
of mesoglea extending aborally from the oral disc at the point
where the latter joins the column wall. As the growth of the
mesentery continues the mesogleal lamella extends farther from
its point of origin, both horizontally along the oral disc and
proximally along the column wall. When by its horizontal
growth the mesogleal lamella comes in contact with the mesoglea
of the stomodeum a rapid downgrowth of the mesentery takes
place. Throughout the development of a mesentery the growth
is the most rapid where there is contact between the growing
mesogleal lamella and that of the body wall, oral disc, column
wall or pedal disc. In all such mesenteries the growth of the
endoderm keeps pace with that of the mesogleal lamella so that
the latter never extends to the free border of the mesentery.

In Figs. 10 to 14, Pis. II. and III., are shown a number of
sections through the same laceration embryo at different levels.
In Fig. 10, which passes close to the oral disc, only four complete

90 LEWIS R. GARY.

mesenteries are present while a member of the third pair has made
its appearance on one side of the longitudinal axis of the sto-
modeum.

In Fig. n, a section taken through the embryo near the base
of the stomodeum, seven mesenteries are present. The same
four as noted in the previous section are complete while both
those of the third pair are present and one of the fourth pair is
seen as a slight prominence of the endoderm containing a very
thin mesogleal fold. In Fig. 12 only six mesenteries are present,
but it seems that one of the third pair is lacking instead of the
single one of the fourth pair which was seen in the previous section.

The first pair have already acquired their mesenteric filaments,
which at this stage extend only a short distance below the
stomodeum.

Farther down, at the level of the section shown in Fig. 13,
only four new mesenteries are present, while seven of those which
came over in the laceration piece from the parent are seen.
This last-mentioned section was taken about one third of the
distance from the oral disc to the base of the embryo, and in this
series the newly formed mesenteries are found in only four of
the following sections.

In the sections shown in Fig. 14, all of the mesenteries present
are those which were in the laceration piece at the time when it
was separated from the parent. Fig. 14, although taken at
some distance from the base shows, in comparison with the pre-
vious sections, that the basal portion of the embryo is of con-
siderably greater diameter than the more distal portions.

A comparison of the three last-mentioned figures illustrates
one phase of the redistribution of materials in the growth of a
laceration embryo. In Fig. 13, that side where the new mesen-
teries are being formed is very thin-walled, the mesoglea having
become reduced to a thin sheet throughout that half of the body.
In the remaining half, on the other hand, the tissue relations of an
adult Aiptasia are still maintained. In the sections through the
more proximal part of the body the adult tissue relations are
apparently undisturbed, at least morphologically. It will be
seen, however, that in this region there has been coincident with
the elevation of the oral end of the embryo a drawing together of

PEDAL LACERATION IN ACTINIANS. 9!

the basal portion so that it is now nearly circular in outline.
In Fig. 15 it may be seen that the old mesenteries which formerly
occupied only one side of the fragment — at the time of its sepa-
ration from the parent — have now come to be distributed so
that they now arise from about four fifths of the c'rcumference.

In sections through an older specimen, as shown in the series
of figures from Fig. 15 to Fig. 18, the second set of mesenteries
has made its appearance about the inner surface of the oral disc,
and eighteen of them have reached the stomodeum. Farther
down along the stomodeum (Fig. 17), the twelve primary mesen-
teries are complete while only one other pair has made sufficient
downward growth to appear in such a section.

In Fig. 18, where eleven of the twelve primary mesenteries are
present, none of those which came over in the laceration piece
from the parent animal are in evidence. Eight of the new mesen-
teries have acquired their filaments, but only those on the first
pair of mesenteries are trilobed.

As a basis for the comparison of this series with the previous
one attention may be called to the appearance of the mesoglea
in Fig. 18. At a corresponding height on the column of the
individual shown in the former series (Fig. 14), none of the newly
formed mesenteries were present, while in Fig. 13, where only
four of them had appeared, that side of the column on which they
were situated showed all of the characteristics of newly formed
tissue.

This shows beyond question that in the older specimen the
new mesenteries have extended down to a part of the body where
there were formerly some of the older mesenteries which by this
time have been entirely resorbed. It will be noticed that in this
part of the body the tissue readjustment has proceeded until
the relation between the layers characteristic for the adult has
been reached.

Carlgren, 1904, in discussing a series of sections through an
embryo in which eight of the new mesenteries had reached the
stomodeum figures two sections and makes statements from which
I can infer only that he understands, either that some of the
old mesenteries are brought into the permanent systen of mesen-
teries of the laceration embryo, or that some of the new mesen-

92 LEWIS R. GARY.

teries arise at a level below the lower end of the stomodeum.
Either of these conclusions is at variance with the results which
I have obtained in the study of the four species of actinians on
which the present paper is based.

Carlgren's figures and the statements directly relating to them
are as follows:

Ein querschnitt in der Schlundrohrsregion (Fig. 4, Taf. IX.)
zeigt ein biradial Anordnung der vollstandigen Mesenterien mit
den beiden Enden des spaltformigen Schlundrohrs sind zwei
Richtungsmesenterienpaare vereinigt. An jede Seite derselben
liegen zwei vollstandige Mesenterien (a) ich nenne sie spater
kurzlich seitliche Mesenterien — die ihre Langsmuskelen gegen
einander kehren. In allem sind also 8 Mesenterien vollstandig.

Rm

cl

a

Rm
FIG. i. (Carlgren's Fig. 4, Taf. IX.)

(Wenn ich unter von einer biradialen Anordnung sprache, sind
die vollstandigen Mesenterien wie hier grupiert.) Zwischen den
Seitlichen Mesenterien an jeder Seite sind verscheidene, unvoll-
standige Mesenterien entstanden, die in dem proximal korper-
theilen (Fig. 7, Taf. IX.), ziemlich staak sind; in den distalen an
der einen Seite verschwinden, an der anderen nur unbedeutend
entwickelt sind oder ganz fehlend (Fig. 4, Taf. IX.). Die mit
(b) bezeichneten Mesenterien bilden mit ihren Partneren an je
ein Mesenterienpaar mit zugewandten Langsmuskeln. Die
Mesenterien des Stuckchens sind also nach der Sechszahlange-
ordenet, obgleich von dem ersten Cyklus vier Mesenterien
unvollstandig sind. Die zwischen den Mesenterien (6), liegenden

PEDAL LACERATION IN ACTINIANS.

93

Mesenterien gehoren einem zweiten Cyklus an. An jeder Seite
der Richtungsmesenterien findet man auch unvollstandige
Mesenterien eines zweiten Cyklus. Auch einige Mesenterien
der dritten Ordnung sind angedeutet (Fig. 7, Taf. IX.). An

Rm

Rm

Rm

Rm

FIG. 2. (Carlgren's Fig. 7, Taf. IX.)

der Bases sind die mit (a) bezeichneten Mesenterien am Starksten,
die Mesenterien (a) an der rechten Seite der Sagittalachse laufen
fast mit den entsprechenden Mesenterien an der anderen zusam-
men (Fig. 7, Taf. IX.).

His Fig. 4, Taf. IX., my text Fig. 2, shows a condition of the
mesenteries which is common to all laceration embryos in a cor-
responding stage of development. His Fig. 7, Taf. IX , text
Fig. 3, shows twenty-five mesenteries, eight of which he considers
to be of the first cycle (mesenteries marked a) while of the re-
maining seventeen six are according to his statement of the
second cycle (marked b~) and the remainder of a third cycle. It
will be noticed that several of the mesenteries in this figure-
one pair of those marked Rm and the longest ones at one end of
the figure — are very broad in section with a thick mass of endo-
derm extending far beyond the mesoglea. None of the entire
number of mesenteries have any trace of mesenteric filaments,
nor by their structure, in so far as it can be made out from the
original figures, is there any indication that they contain actively

94 LEWIS R. GARY.

growing tissue as is always found in the new mesenteries in the
forms that I have studied. Comparing these two figures with
those of the series shown in my Figs. 10 to 14, Pis. II. and III.,
it seems very evident that all of the mesenteries shown in Carl-
gren's Fig. 7, Taf. IX., are old ones which have come over in
the fragment from the parent individual and which will never
come to be a part of the permanent system of mesenteries of the
actinian arising from the laceration embryo. In the younger
embryo shown in my series of figures just mentioned the first
formed pair of mesenteries has already acquired their mesenteric
filaments, which are as yet, of course, in an undifferentiated
condition; but nevertheless clearly distinguishable as mesenteric
filaments.

In the embryo from wrhich my series of Figs. 16 to 18 was
taken, which has at the level of the stomodeum only four more
mesenteries than that from which Carlgren's figures under discus-
sion were taken, eight of the new mesenteries, at the level at which
Fig. 1 8 was taken, are provided with filaments of which those on
the first pair of mesenteries are already trilobed, while several of
the others show the first steps in the process of differentiation
by which the trilobed condition is reached.

(c) The Development of the Mesenteric Filaments.

The question of the origin of the mesenteric filaments among
the anthozoa has been discussed to considerable length by most
workers on the development of these forms. With the exception
of E. B. Wilson, 1884, all of the previous investigators have
confined themselves to a consideration of the embryonic develop-
ment. The above-named investigator worked with the polyps
arising by budding in several alcyonarians, but not with embryos
of the same forms.

E. B. Wilson (loc. tit.} reached the conclusion that in the
Alcyonaria the dorsal directive filaments arise as downgrowths
of the ectoderm of the stomodeum, while all the other filaments
arise through the differentiation of the endoderm at the free
border of the mesentery. He would homologize the dorsal
filaments of the Alcyonaria with the lateral ciliated bands (flim-
merstriefen) of the trilobed actinian filament. The filaments

PEDAL LACERATION IN ACTINIANS. 95

of the remaining mesenteries he would homologize with the glan-
dular streak (Nesseldriisenstreif) of the trilobed filament. Thus,
according to his interpretation, the actinian filament is a double
structure.

H. V. Wilson (loc. tit.'} working on the coral Manitina, where
the filament consists of a single lobe, reached the conclusion
that this type of filament was comparable to the trilobed actinian
filament and is entirely of ectodermic origin.

McMurrich, 1891, describes a double origin for the trilobed
filaments of Rhodactis and Aulactinia; after the manner suggested
by E. B. Wilson.

Duerden, 1899, working on Lebrunia, where the mesenteries
become developed before the stomodeum becomes hollowed out,
concluded that in this case part, at least, of the filaments are of
endodermic origin since he finds stages where filaments are present
on some of the mesenteries which as yet have no direct con-
nection with the stomodeum.

Appellof (loc. tit.) concludes that the mesenteric filaments
are developed entirely from the ectodermal downgrowth from
the lining of the stomodeum. He mentions having observed
a number of instances where the ectodermal layer had become
very much attenuated just below the level of the stomodeum, and
indeed in some instances the continuity of this layer had been
broken. He argues from this fact that some of the instances
observed by previous workers who, on the basis of a lack of con-
tinuity between the mesenteric filament and the lining of the
stomodeum, maintained the endodermic origin of the filaments
might be simply cases where the continuity of the tissue had been
broken and in no way fundamentally different from the ordinary
course of events as he has described them. He also makes the
suggestion that the method of development described by E. B.
Wilson for the alcyonarian filaments may be confined to those
individuals which arise as buds and not a true explanation of what
takes place during embryonic development, which he tacitly
assumes would be similar to plans followed in other Anthozoan
forms.

In all descriptions of the formation of the mesenteries, and
particularly of the mesenteric filaments, except that of Duerden

96 LEWIS R. GARY.

(loc. cit.} for Lebrunia, it is recorded that the filaments appear
first on the upper part of the mesenteries at the level of the free
border of the stomodeum, and usually there can be observed
a direct continuity between the ectodermal lining of the stomo-
deum and the filaments from the first appearance of the latter.

The mesenteries acquire their filaments in the same order
as that in which they arose, so that the dorso-laterals would
first show these structures. In Fig. 12, PI. II., this pair of
mesenteries alone shows the beginning of the filaments. Farther
up in this series of sections the tissue forming the filaments is
directly continued into the lining of the stomodeum. In Fig. 9,
PI. II., a longitudinal section through a young laceration
embryo, the filament extends for a short distance down along the
free border of the mesentery, and here there is no histological
differentiation to mark the lower limits of the stomodeum. Its
actual limit can, however, be determined by a comparison writh
the other side of the figure where the section passed between
two mesenteries so that the border of the stomodeum hangs free
"in the gastroccel. H. V. Wilson (loc. cit.) has pointed out that
in Manicina the filaments of all the first twelve mesenteries, save
the first pair, arise from a fold of the ectoderm- at the free border
of the stomodeum which bends up and outward into the gas-
trocoel pointing toward the oral disc. In this manner the fila-
ment begins to grow down the mesentery before the edge of the
latter has in its upper part come into contact with the stomodeum.

In the development of laceration pieces of Aiptasia there is the
same folding backward of the wall of the stomodeum at an early
stage in the development (Fig. 8), so that the course of the
development of the filaments in this form follows that already
described for Manicina.

In an older specimen, Fig. 18, PI. II., the filaments of the
first pair of mesenteries have assumed the trilobed form charac-
teristic of the filaments of adult Aiptasias, while the remainder of
the mesenteries have at this level either simple filaments or
none at all. It is thus apparent that the change in the structure
of the filaments follows the same order as does the appearance of
the filaments themselves.

In this transition from the simple to the tri-lobed structure

PEDAL LACERATION IN ACTINIANS. 97

there is in these forms no indication of the origin of the ciliated
bands as an acquisition of new or different tissues, as McMurrich
has suggested in arguing for a dual origin for the parts of the
complex filament, but, on the other hand, the lateral lobes of the
complex filament arise by a very slight differentiation of the cells
already present in the filament from the time of its origin from
the stomodeal ectoderm. In Fig. 17, PI. III., the filaments of
the dorsal directive, the ventro-lateral and the ventral directive
mesenteries all show in their centers the beginning of the growth
which will give to the mesogleal layer its characteristic branched
appearance. Each of these branches will become the center
and support, of one lobe of the adult trefoil filament. In the
dorsal directive mesenteries especially, it may be noticed that
the tissues of the filament have extended out around the edges
of the mesentery so that the filament is nearly circular in outline.
Originally the extension of the stomodeal ectoderm — first appear-
ance of the filament — was in the form of a flat sheet (Fig. 12,
PL II.). The separation of this sheet of tissue into the three
lobes of the adult filament apparently takes place through an
unequal growth from three points in the nearly circular filament,
as shown in Fig. 18, PI. III.

As may be seen by comparing Fig. 9 and Fig. 12, PI. II.,
with the fully formed filament (e. g., the filaments of the dorso-
lateral mesenteries in Fig. 17, PI. III.) the histological changes
which have taken place in the transformation of the "embryonic"
tissue of the stomodeal ectoderm to those of the tri-lobed filament
have been in the nature of a separation of the cellular elements
of the original tissue so that one type of cells (ciliated) is cen-
tralized in the lateral lobes; while the other types (gland and nettle
cells) are concentrated in the median lobe. The apparent
change is greatest in the case of the median lobe, since the un-
differentiated stomodeal ectoderm is at the time of the first
appearance of the filaments poor in gland cells and nematocysts.
This last-mentioned fact is also true of the activity growing
ectoderm of the new part of the column wall and oral disc, and in
all parts of the body there is a marked increase in the actual as
well as in the proportionate number of both these types of cel-
lular elements after the normal body proportions have been
established.

98 LEWIS R. GARY.

(d) Histological Changes During Development by Laceration.

In considering the histological changes undergone by the
tissues during development by pedal laceration those changes
which take place in the basal portions of the parent individual
before the fragment is torn off may be first mentioned. Andres
(loc. cit.) mentions that in Aiptasia lacerata the rim of the basal
portion of the animal becomes very opaque, and that in section
there seems to be an unusual number of gland cells in this region.
The endoderm becomes very much thickened in this region so
that when the fragment becomes torn off there is only a small
cavity left in its interior (Fig. 7, PI. II.). As may be seen in this
section, the thickening of the endoderm is very much greater
on the upper side of the fragment, i. e., that which has come from
the original column well.

In the period immediately following upon the separation of the
fragment from the parent, there is a rapid change in the character
of the tissues about the orifice. All of the newly forming tissue
becomes very light in color due to the comparative paucity of
the zooxanthellae in the endoderm of the active area. The
mesoglea becomes comparatively thin in this region, but seems
always to arise as a direct prolongation of that already present
and not as a new secretion between the ectoderm and endoderm
of the new parts.

In the cellular layers the changes are more profound. In
the ectoderm (Fig. 19, PI. IV.), a section through a young tentacle
bud, the boundaries of all of the cells except the nematocysts
and gland cells becomes lost so that the whole layer constitutes
a syncytium. The cytoplasm is uniformly granular throughout.
The nuclei are scattered through the cytoplasm, but still indicate
by their positions the fact that only definite cell walls are wanting
in order that the usual cell structure should be apparent. At
the internal boundary of the entoderm there is distinguishable
a layer of very fine muscle fibers. When compared with the
size of the nuclei and nematocyst cells, these fibers are seen to be
several times less than the diameter of normal muscle fibers of
this species of actinian. Since in the fully developed Aiptasia
muscle fibers exist both as processes from epithelio-muscle cells
and as nearly the entire component of "muscle cells" in which
the cytoplasmic cell body is reduced to a small mass containing

PEDAL LACERATION IN ACTINIANS. 99

the nucleus, it is impossible in this case to determine whether we
have to do with fibers of the first or of the second of the two
above-mentioned kinds. In other words the fibers in question
may be in the process of becoming differentiated from epithelial
cells, or they may be arising through the division of previously
existing muscle cells of the second type. At this stage there
is no indication of interstitial cells or of developing gland cells
in the ectoderm.

In the endodermal layer of the same section there is exhibited
a similar condition of the tissue: Cell walls are entirely wanting;
the nuclei, however, by their position indicate a potential cellular
arrangement of the cytoplasm. A particularly striking feature
as regards this layer is brought out by a comparison of the
figure under discussion and Fig. 7, PI. II. In the last-mentioned
figure the endoderm, throughout most of its extent, is consid-
erably thicker than the ectoderm, while in the new tissue the
former is generally so thin that the single zooxanthellse are greater
in diameter than the layer of cytoplasm making up the endoderm
and consequently they cause a prominence extending into the
cavity of the tentacle wherever they are situated. In some few
instances the zooxanthellae are placed one above the other, and
n such cases there is either a generally thickened area of the cyto-
plasm or else a cylindrical prominence, as is seen on one side of
the figure. All cellular differentiation is entirely lacking in
the endoderm. At no place about the circumference of this
layer is there any indication of the layer of endodermal (circular)
muscles, which are in normal Aiptasia tissues, when seen in trans-
verse sections of the animal, much more prominent than the
ectodermal muscles.

At the base of the fragment where the rearrangement of the
materials is proceeding at a relatively slow rate there seems never
to be a breaking down of the tissues comparable to that just
described. The mesogleal layer becomes thinned out progres-
sively so that at first it is of its usual thickness on one side of the
fragment — that side originally opposite the point where the frag-
ment was torn off from the parent animal — while on the opposite,
more actively growing, side it has become relatively thin (Fig. 13,
PI. III.). The cellular layers retain practically their original
characteristic appearance, although, as is shown in Fig. 15,

IOO LEWIS R. GARY.

PL III., the ectoderm on the side where the new growth is the
most rapid, i. e., where there are no old mesenteries in the figure,
is much lower than on the opposite side.

When in the later development the cellular elements of the
ectoderm and endoderm become delimited by the appearance
of cell walls it is noticeable that all of the cells show at first an
embryonic character in that they are relatively wide in proportion
to their height (compare Figs. 10 and n with the old part of the
tissue in Fig. 14). In the ectoderm of the column wall in Figs. 10
and ii, and more especially in the lining of the stomodeum of
this individual, there is a very marked paucity of the nematocysts
and gland cells, so that it is apparent that there has not been
an increase in these cellular elements to keep pace with the
increase in the epithelial cells. In the older embryo from which
the series of figures from 15 to 18 was taken the nematocysts and
gland cells have reached nearly their normal number and dis-
tribution throughout the newly formed tissues.

In the sections shown in Figs. 10 and n the endoderm cells
are in general broader than the ectoderm cells and not generally
of greater height. In the older tissue (Fig. 18) there are no
apparent cell boundaries in the endoderm and the zooxanthellse

are much more numerous.

SUMMARY.

1. The change in the form of a laceration piece, leading up to
the acquisition of the typical actinian shape, takes place through
the upgrowth of the tissue about the orifice where it was torn
off from the parent.

2. The permanent mesenteries arise as new growths in the
undifferentiated tissues of the oral end of the laceration piece.

3. The first twelve mesenteries do not appear in the sequence
followed by those in egg embryos.

4. As development goes on the old mesenteries — those brought
over from the parent when the fragment was torn off — become
restricted to a proportionately shorter and shorter part of the
base of the young actinian until they are finally entirely resorbed.

5. The mesenteric filaments are formed, just as in egg embryos,
from a downgrowth of the ectodermal lining of the stomodeum.
Their trilobed condition arises through the differentiation of this
tissue.

PEDAL LACERATION IN ACTINIANS. IOI

6. The tissues of the most actively growing part of a laceration
piece become very thin ; the ectoderm and the entoderm lose all
apparent cell boundaries; the mesoglea arises as a direct continu-
ation of that present in the older tissues.

7. The newly formed tissues contain a very small number of
gland cells and nematocysts. These two types of cells are de-
veloped to the usual number after the tissue relations have be-
become stable.

BIOLOGICAL LABORATORY,
PRINCETON UNIVERSITY.

BIBLIOGRAPHY.
Andres, A.

'82 Intorno alia Scissiparita della Attinae. Mittheilungen a. d. Zool. Sta. z.

Neapel., Bd. III.
Appellof, A.

'oo Studien iiber Actinien-Entwicklung. Bergens Museum Aarbog f. 1900.
Carlgren, O.

'04 Studien iiber Regenerations und Regelationserscheinungen. i. Uber die
correlationen zwischen der Regulation und der Symmetrie bei den Actini-
arien. Svenska. Akad. Handl., Bd. 37, No. 8.
'09 Studien iiber Regenerations und Regulationserscheinungen. Erganzende

Untersuchungen an Actiniarien. Ibid., Bd. 43, No. 9.
Child, C. M.

'o3-'os Form Regulation in Cerianthus. Biol. Bull., vols. 5, 6, 7, 8.
Duerden, J. E.

'99 The Edwardsis stage of the Actinian Lebrunia, and the Formation of the

Gastro-co3lomic Cavity. Jour. Linn. Soc., Vol. XXVII.
Hazen, Anna P.

'02 The Regeneration of an CEsophagus in the Anemone, Sagartia luciae. Arch.

f. Entw. Mech., Bd. XIV.

'03 Regeneration in the Anemone, Sagartia luciae. Ibid., Bd. XVI.
Hertwig, O. und R.
Die Actinien.
Parker, G. H.

'97 The Mesenteries and Siphonoglyphs in Metridium marginatum M. Edw.

Bull. Mus. Comp. Zool. Harvard, vol. 30.
'99 Longitudinal Fission in Metridium marginatum Milne Edwards. Bull.

Mus. Comp. Zool. Harvard, vol. 35.
Torrey, H. B.

'98 Observations on Monogenesis in Metridium. Proc. Calif. Acad. Sci., 1898.
Wilson, E. B.

'82 The Development of Renilla. Philosophical Trs. Roy. Sci., 1883.

'84 The Mesenterial Filaments of the Alcyonaria. Mitt. Zool. Sta. Neapel,

Bd. V., 1884.
Wilson, H. V.

'88 On the Development of Manicina areolata. Jour. Morph., Vol. II., No. 2,
1888.

IO2 LEWIS R. GARY.

EXPLANATION OF PLATES.

All the figures were drawn from specimens of A. pallida; those of the entire
laceration pieces from specimens grown in the laboratory; the sections from speci-
mens obtained at Beaufort, N. C.

The ectoderm is stippled in all the figures, the endoderm is distinguished by the
presence of the zooxanthellae, the mesoglea has been represented by either a single
line in the actively growing parts of the embryo or by a clear space between the
cellular layers in the older tissues.

PLATE I.

The figures on this plate represent a series of stages in the development of a
laceration piece, from the time of its separation from the parent until it has assumed
the form of a typical young actinian.

BIOLOGICAL BULLETIN, VOL. XX

LEWIS R. CARV

IO4 LEWIS E. GARY.

PLATE II.

FIG. 7. A section through a laceration piece soon after its separation from the
parent.

FIG. 8. Part of a vertical section through a young bud to show the formation
of the stomodeum.

FIG. 9. Part of a longitudinal section through a young laceration embryo,
such as is shown in Fig. 4, Plate I., to show the downgrowth of the ectodermal
lining of the stomodeum to form the mesenteric filaments.

FIGS. 10, n, 12 (with FIGS. 13 and 14, PLATE III.). A series of sections at
different lev-els taken through the same embryo to show the relation of the new to
the old mesenteries. (The orientation is the same in all the figures.)

BIOLOGICAL BULLETIN, VOL. XX

PLATE II

10

LEWIS R. CARY

IO6 LEWIS R. GARY.

PLATE III.

FIGS. 13 and 14 are part of the series described in the last paragraph. They
complete the series which began with Fig. 10.

FIGS. 15, 16, 17 and 18. A series of sections, at different levels, through an older
embryo in which all of the twelve primary mesenteries have reached the stomodeum.

BIOLOGICAL BULLETIN, VOL. XX

13

- ' <"•••' "'- -

i?

,

LfcWIS R. CARY

108 LEWIS R. GARY.

PLATE IV.

FIG. 19. A section through a very young tentacle bud to show the character-
istics of the tissues in an actively growing region of the developing laceration embryo.

BIOLOGICAL BULLETIN, VOL. XX

PLATE IV

LEWIS B. GARY

FURTHER STUDIES ON HETEROCHROMOSOMES IN

MOSQUITOES.

N. M. STEVENS.

In my study on the chromosomes in the germ cells of Culex,
published in March, 1910 ('10), there were three figures (n, 20
and 21) which differed from all others for Culex, in showing an
unequal pair of chromosomes in the spermatogonia (Fig. u),
and a condensed pair in the growth stage of the spermatocytes
(Figs. 20 and 21). At the time when the proof of that paper was
corrected, it was suspected that the three figures referred to
belonged to another species, probably Anopheles punctipennis, a
few larvae of Anopheles having, it was supposed, been collected
in the net with Culex, and having escaped detection in both larva
and pupa stage.

In October of this year (1910) an abundant supply of Anopheles
larvae was secured and a satisfactory study of both male and
female germ cells made, showing without question that in the
male Anopheles punctipennis there is present an unequal pair of
heterochromosomes, attached to one of the equal pairs, but other-
wise exactly comparable to the unequal pair described by the
author for nine species of Muscidae ('08). A careful reexami-
nation of the germ cells of Culex pipiens, and a study of Culex
tarsalis and Theobaldia incidens in California during the summer,
revealed no such heterochromosomes in those species. A de-
scription of the chromosomes in the germ cells of Anopheles
punctipennis and Theobaldia incidens will be given in the fol-
lowing pages.

Methods. — The germ glands of both Anopheles and Theobaldia
were studied both in-acetocarmine preparations and in sections
of material fixed in Flemming and stained with iron-haematoxylin
or thionin. The latter stain gave very satisfactory results with
this material, and saved much time in preparing the sections for
study.

100

IIO N. M. STEVENS.

Anopheles punctipennis.

Very clear oogonial plates were found in several ovaries, and
two of these are shown in Figs. I and 2. There are two equal

<?

5 6

FIGS, i and 2. Anopheles punctipennis, chromosomes of the oogonia. Mag,

'OO.

FIGS. 3 and 4. Anopheles punctipinnis, chromosomes of the spermatogonia.

me mag.

FIG. 5. Culex pipiens, chromosomes of an oogonium. Same mag.

FIG. 6. Culex pipiens, chromosomes of a spermatogonium. Same mag.

pairs of chromosomes very similar to those found in other Diptera,
and a third pair which is clearly made up of a longer and a shorter
pair. Corresponding figures of spermatogonial plates are shown
in Figs. 3 and 4. Here the composite pair consists of a longer
equal pair, as in the oogonia, united with a smaller unequal pair,
the larger member of which is of the same relative size as the
members of the corresponding small pair in the oogonial plates.
All of the spermatogonial plates in this Anopheles material were
of this type. In Culex, on the other hand, no such inequality or
composite condition of one pair of chromosomes could be de-
tected. Figs. 5(9) and 6 (d1) are new figures of the oogonial
and spermatogonial chromosomes of Culex pipiens, and those of
Culex tarsalis are of the same general character.

HETEROCHROMOSOMES IN MOSQUITOES. Ill

In resting spermatogonia of Anopheles (Figs. 7 and 8) a con-
densed mass of chromatin varying in form is found associated
with a large plasmosome near the nuclear membrane. In Culex
there is nothing of the kind — only a large central plasmosome
surrounded by an even spireme. In Anopheles there is a very
conspicuous contraction or synizesis stage (Fig. 9) in which the

••' •

• _ • •' — •*<""»* f

•-'^

8

FIGS. 7 and 8. Anopheles punctipennis, resting spermatogonia. X =the hetero-
chromosome.

FIG. 9. Synizesis stage.
FIGS. 10-12. Growth stages.
FIG. 13. Prophase.

heterochromosome appears as a deeply staining mass of chromatin
resting against a ball of paler chromatin threads. The corre-
sponding stage in Culex ('10, Figs. 12 and 13) showed a ball of
chromatin thread wound about a plasmosome. In the later
growth stages of the spermatocytes of A nopheles, the heterochro-
mosome pair assumes a great diversity of forms. Occasionally,
especially in early stages, the unequal ends of the pair only are
fully condensed as shown in Fig. 10, which is a tangential section
of the nucleus. Fig. 1 1 shows a late spireme stage with hetero-
chromosome (x), plasmosome (p~) and parts of a pale spireme.
Fig. 12 is a late growth stage taken from a cyst where the cells
were mostly in metaphase of the first spermatocyte mitosis;
here the heterochromosome pair shows nothing of its composite

112

N. M. STEVENS.

nature, and the members of the pair are in about the position
which they assume in the spindle. Throughout the growth
stage the composite pair of chromosomes is found, more or less
condensed, but always distinguishable from the remainder of the
chromatin — against the nuclear membrane, twisted or coiled or

20

FIGS. 14—17. Anopheles punctipennis, first spermatocyte metaphase. X=the
unequal pair.

FIG. 1 8. Anaphase.

FIGS. 19 and 20. Culex pipiens, first spermatocyte metaphase and anaphase.

FIG. 21. Anopheles punctipennis, second spermatocyte metaphase.

FIG. 22. Spermatid nuclei. All figures same magnification — 2,000.

knotted into such shapes that it is usually impossible to dis-
tinguish the equal and unequal pairs of which it is composed.
Prophase stages are evidently much briefer than in Culex,
for it was difficult to find any clear figures — Fig. 13 shows the
composite heterochromosome pair (x) and the other two chro-
mosomes not yet separated into their parallel components.

HETEROCHROMOSOMES IN MOSQUITOES. 113

The character of the chromosome pairs as they appear in the first
maturation mitosis is well shown in Figs. 14-17, where the chro-
mosomes are represented in one plane for the sake of clearness.
In each figure x is the composite pair. The spindle fibers appear
to be attached at the meeting point of the smaller and larger
components, the smaller, unequal pair forming the outer ends
of the figure and appearing as a conspicuous hook at one end and
a rounded knob at the other. No distinct separation of the
members of the composite pair, such as is always seen in the
spermatogonia (Figs. 3 and 4) was ever observed in the spermat-
ocyte metaphase. The other two chromosomes frequently
assume the form of rings in the early metaphase (Figs. 16 and 17)
and in the anaphase (Fig. 18) they separate as V's with equal
arms, while the heterochromosomes form a pair of hooks with
equal long arms and unequal short arms (Fig. 18).

Fig. 19 shows the chromosomes of a first spermatocyte meta-
phase of Culex pipiens and Fig. 20 an anaphase. The chromo-
somes correspond very closely in form and relative size to those
of Anopheles with the exception that all three pairs are equal.
Both cells and chromosomes are much larger in Culex than in
Anopheles.

In the second spermatocytes the chromosomes are already
divided when they come into the spindle and are usually so
tangled as to defy any attempt to distinguish the two classes or
make clear drawings. In Fig. 21 we have an unusually clear
metaphase. An aggregation of chromatin, apparently the hetero-
chromosome, is found in the spermatids (Fig. 22), lying against
the nuclear membrane, and in some cases the dimorphism can be
distinguished.

Theobaldia incident.

Theobaldia incidens is one of the most common mosquitoes in
the Santa Clara Valley, California. The larvae were found in
abundance in a bucket of water which had been allowed to stand
partly covered for two or three weeks. A few bread crumbs
thrown into the glass jars in which the larvae were kept under
observation, enabled them to grow rapidly and transform into
fine large pupae whose germ glands yielded all desired stages,
from dividing oogonia and spermatogonia to ripe spermatozoa.

N. M. STEVENS.

\

27

/"

32

I *"•<*.
» *

30

33

\

34

FIG. 23. Theobaldia incidens, oogonial prophase.

FIGS. 24-26. Chromosomes of the spermatogonia.

FIG. 27. Spermatogonial prophase.

FIG. 28. Synizesis stage.

FIGS. 29 and 30. First spermatocyte growth stages.

FIGS. 31 and 32. First spermatocyte prophase and metaphase.

FIGS. 33 and 34. Second spermatocyte prophase.

HETEROCHROMOSOMES IN MOSQUITOES. 115

During the summer only acetocarmine preparations were used
and the material resembled Culex so closely that only a few
figures were made. Other figures were taken from sections of
ovaries and testes fixed in Flemming and stained with thionin.
An oogonial prophase is shown in Fig. 23, and the chromosomes
from three spermatogonia in Figs. 24, 25 and 26. Fig. 27 is the
nucleus of a spermatogonium in prophase. In no one of these
figures is there any indication of inequality in the members of
any pair of chromosomes. Neither in resting spermatogonia
nor in the growth stages of the spermatocytes could anything
resembling the heterochromosomes of Anopheles be found. In
all of these stages a central plasmosome is surrounded by a
spireme as in Culex. Figs. 28 to 30 show a contraction or synizesis
stage, and two later growth stages. As in Culex the plasmosome
stains a deeper blue in late growth stages than in the spermato-
gonia and synizesis stages. Fig. 31 is a characteristic prophase
of the first spermatocyte mitosis. In the metaphase two of the
chromosme pairs nearly always appear in the form of rings and
separate as V's; the third pair is usually separated and its uni-
valent components already divided longitudinally, when the
rings are just dividing (Fig. 32). The second spermatocyte
propha'ses (Figs. 33 and 34), metaphases and anaphases resemble
closely those of Culex.

The testes of Theobaldia differed from those of Culex in having
much larger spermatogonia and in lacking a prolonged prophase
of the first spermatocyte mitosis. The first maturation division
also shows characteristic differences in the behavior of the
chromosome pairs in metakinesis (compare Figs. 19 and 32).
Otherwise both ovaries and testes showed similar conditions in
the two species, Culex pipiens and Theobaldia incidens; i. e.,
three equal pairs of chromosomes, no one of which behaves in
any respect like the heterochromosomes of other insects, including
Anopheles punctipennis.

Anopheles sp.?

On March 30 and again on May 31, a few pupae were obtained
from the same pond where I obtained my material in 1909.
All of the males in these collections had six chromosomes in the
spermatocytes and a distinct heterochromosome in the growth

Il6 N. M. STEVENS.

stages. The pupae were thought to be Anopheles but no larvae or
adults were secured for identification. It was supposed that the
few specimens obtained must have wintered over, probably in

36

38

FIG. 35. Anopheles sp.? First spermatocyte growth stage. X = the hetero-
chromosome.

FIG. 36. First spermatocyte prophase.

FIGS. 37 and 38. Chromosomes of the first spermatocyte metaphase. Mag.
2,000 in all figures.

the larval stage. The net was used in the pond many times during
March, April and May, but no mosquito larvae or pupae secured
except on the two dates given above.

Fig. 35 shows the heterochromosome in the nucleus of a growth
stage, Fig. 36 a prophase of the first maturation mitosis, and Figs.
37 and 38 the six chromosomes from two first spermatocyte
spindles. The largest pair is slightly unequal but not composite
as in Anopheles punctipennis.

DISCUSSION.

As we have already seen, the various species of Culicidae, thus
far studied, present peculiarities not hitherto encountered in the
study of heterochromosomes in the various groups of insects.

In Anisoloba maritima (Randolph, '08) and several species
of Lepidoptera (Stevens, '06; Dederer, '07; Cook, '10) there
is an equal pair of heterochromosomes distinguishable as such by
their condensed condition in the growth stages of the first
spermatocytes. In Drosophila ampelophila the unequal pair of
heterochromosomes of the maturation mitoses is not condensed

HETEROCHROMOSOMES IN MOSQUITOES. 1 17

during the growth stages (Stevens, '08). In Metapodius ter-
minalis (Wilson, '09) we find either an unpaired heterochro-
mosome or an unequal pair of idiochromosomes, and a varying
number of supernumeraries within the same species. In no
case before have undoubted heterochromosomes, following the
usual rule as to sex distribution, been found in one or more genera
or species and entirely absent in nearly related species of the
same family. In Anopheles punctipennis the heterochromosomes
can be traced from the youngest spermatogonia observed, through
the two generations of spermatocytes to the spermatids and even
to a stage in which the spermatids are transforming into sper-
matozoa. The unequal pair in the male is represented by an equal
pair in the female, and as in other similar cases, it is evident
that an egg fertilized by a spermatozoon containing the smaller
heterochromosome must produce a male, while an egg fertilized
by a sperm containing the larger heterochromosome will develop
into a female.

Now in Culex and Theobaldia all of the chromosomes behave
alike so far as condensation is concerned, and there is neither an
unpaired heterochromosome nor an unequal pair to be detected
in any stage. In other words there is no visible differentiation
of one pair of chromosomes which can be associated with the
determination of sex. There seems to be absolutely no reason
for supposing that the mechanism by which sex is determined
is not identical in these three species of mosquitoes; and the inevi-
table conclusion would seem to be either that the heterochro-
mosomes of the male are not a necessary factor in sex determi-
nation, or that the sex-determining chromosomes are not neces-
sarily differentiated as heterochromosomes.

I have kept no record of the sex of the pupae dissected, but
should say that the numbers of males and females were about
equal for all of the species studied. In a large collection of
Theobaldia larvae which, judging by their size, might have all
come from one laying, most of the males pupated a day or two
sooner than the females. All of the pupae were taken out and
dissected each day. The first day nearly all were males and
the last day nearly all females. An average for the four or five
days, I feel sure, must have given nearly equal numbers.

Il8 N. M. STEVENS.

Wilson's recent hypothesis ('09) that in some way two X
chromosomes determine the female sex, and one X chromosome
or an X and a F chromosome determine the male sex would
apply perfectly to Anopheles.

Egg^T+sp. X=9,
EggX+sp. F=d\

But what shall we say of Culex and Theobaldia where there is
no differentiation of X and F chromosomes? In forms which
have an equal pair of heterochromosomes, showing the same
general characteristics as the unpaired heterochromosomes and
the idiochromosomes, one naturally infers that the members of
the equal pair in the male have X and F values respectively,
but shall we go further and argue that likewise in Culex and Theo-
baldia the mature eggs all have an X chromosome and the sper-
matozoa are dimorphic, one half having an X chromosome and
the other half a F chromosome, although nothing of the kind
can be traced? If we make this assumption, then Culex and
Theobaldia may stand at one end of a series, as examples of
absence of heterochromosomes associated with sex, while forms
with an unpaired heterochromosome come at the other end of
the series, and all the varieties of paired heterochromosomes
between, the mechanism of sex determination being the same
in all, but evident to the eye only when differential heterochro-
mosomes are present.

The case of the aphids and phylloxerans has been the strongest
argument for the hypothesis that two X chromosomes give a
female and one X or XY a male, since the rejection of one X
chromosome from a parthenogenetic egg is followed by the devel-
opment of a male, but in Culex and Theobaldia it is evident that
sex determination is not dependent on the presence of X and F
chromosomes, although in Anopheles these chromosomes are at
least closely correlated with sex-determination.

At present, the all-important questions seem to me to be:
What is the meaning of the differentiation of heterochromosomes
in one form and not in others closely related? WThat has been
the history of such differentiation where we have an unpaired
heterochromosome or an unequal pair of heterochromosomes?

HETEROCHROMOSOMES IN MOSQUITOES. 1 19

In Metapodius Wilson ('09) traces the origin of an unpaired
heterochromosome and also of the supernumeraries to an irregular
mitosis in which both idiochromosomes go to one daughter cell.
In the Diabroticas the supernumeraries probably owe their origin
to an irregular division of the unpaired heterochromosome.

But in no case are we able to say when or how or why certain
spermatogonial chromosomes became specially differentiated
as heterochromosomes.

We have associated the X and Y chromosomes of the male
with sex-determination, but possibly they have some other
meaning and are only so correlated with the sex-determining
mechanism that they have the same distribution as the sex
characters. The case of the mosquitoes certainly indicates
that we must study the heterochromosomes, apart from the idea
of sex-determination, more intensively, in order to determine
if possible why we have such chromosomes differentiated in
some forms and not in others.

In such cases as Anopheles the sex characters may be supposed
to be located in the unequal heterochromosomes or in the larger
equal pair with which they are connected. This may also be
true of the composite heterochromosomes of Drosophila am-
pelophila and Hesperotettix (McClung, '05). In the ordinary
cases of an unpaired heterochromosome or an unequal pair, we
may suppose the sex characters to be located in the heterochro-
mosomes, but to be independent of the special differentiation,
which may be entirely absent as in Culex and Theobaldia.

If we apply Morgan's new sex formula—

FmFm = 9 , Fmfm = cf ,

to the cases under discussion, m may be located in any equal
pair but F must be either in the heterochromosomes or in chro-
mosomes correlated with them. Sex-limited inheritance in
the male (Morgan, '10) would then seem to be possible only
where unpaired or unequally paired heterochromosomes are
present, and not in such cases as, Anisoloba, the Lepidoptera,
Culex and Theobaldia.

It occurs to me as possible that heterochromosome differenti-
ation may be directly related to sex-limited inheritance of certain

I2O N. M. STEVENS.

characters. If so we should expect to find an unpaired or an
unequally paired heterochromosome in the female of Abraxas
where there is sex-limited inheritance in the female, and also to
find sex-limited inheritance of some character- or characters in
the female of those echinoderms in which Baltzer ('09) and
Heffner ('10) have shown that an unequal pair of chromosomes
is present in the female. Apparently we have reached a point
where the experimental breeder and the cytologist must work
together in order to solve the problems of sex-determination,
sex-limited inheritance and heterochromosome differentiation.

BRYN MAWR COLLEGE,
December 20, 1910.

LITERATURE.
Baltzer, F.

'09 Die Chromosomen non Strongytocentrotus lividus und Echinus micro-

tuberculatus. Arch. f. Zellf., II.
Cook, M. H.

'10 Spermatogenesis in Lepidoptera. Proc. Phila. Acad. Sci., April, 1910.
Dederer, P. H.

'07 Spermatogenesis in Philosamia cynthia. Biol. Bull., XIII.
Heffner, B.

'10 A Study of Chromosomes of Toxopneustes variegatus which show Individual

Peculiarities of Form. Biol. Bull., XIX.
McClung, C. E.

'05 The Chromosome Complex of Orthopteran Spermatocytes. Biol. Bull.,

IX.
Morgan, T. H.

'10 Sex-limited Inheritance in Drosophila. Science, N. S., XXXII.
Randolph, H.

'08 On the Spermatogenesis of the Earwig Anisoloba maritima. Biol. Bull..

XV.
Stevens, N. M.

'06 A Comparative Study of the Heterochromosomes in Certain Species of
Coleoptera, Hemiptera and Lepidoptera, with especial Reference to Sex
Determination. Carnegie Inst. Pub. No. 36, II.

'08 A Study of the Germ Cells of Certain Diptera, with Reference to the
Heterochromosomes and the Phenomena of Synapsis. Journ. Exp. Zool., V.
'10 The Chromosomes in the Germ Cells of Culex. Jour. Exp. Zool., VIII.
Wilson, E. B.

'05 The Behavior of the Idiochromosomes in Hemiptera. Journ. Exp. Zool., II.
'09 Recent Researches on the Determination and Heredity of Sex. Science,

N. S.. XXIX.

'09 The Chromosomes of Metapodius, a Contribution to the Hypothesis of
the Genetic Continuity of Chromosomes. Jour. Exp. Zool., VI.

PRELIMINARY NOTE ON HETEROCHROMOSOMES IN

THE GUINEA PIG.

N. M. STEVENS.

It is the purpose of this preliminary note on the spermato-
genesis of the common guinea-pig merely to state some of the
most obvious facts concerning the unequally paired heterochro-
mosomes as they appear in the first spermatocytes. In the
later growth stages a characteristic heterochromosome (Fig. I , x)

-.-x

X I

a D

'W
4 5

FIG. i. First spermatocyte growth stage. X = the heterochromosome. Mag.
2,000.

FIG. 2. Prophase.

FIG. 3. First spermatocyte metaphase, tangential section of spindle showing
the heterochromosome X.

FIG. 4. Anaphase.

FIG. 5. The heterochromosome; a, b, c, from sections of Flemming material;
c, d, from acetocarmine preparations. Mag. 2,000 for all of the figures.

121

122 N. M. STEVENS.

is easily distinguished both in acetocarmine preparations and in
sections of Flemming material stained with thionin. The
plasmosome (p) is even paler than the spireme. In Fig. 2, a
prophase, the heterochromosome (*) is also shown. Fig. 3, a
tangential section of a first spermatocyte spindle, shows the usual
form of the heterochromosome in metaphase, and in Fig. 4 we
have it divided into its unequal components. Fig. 5, a, b, c,
shows slight variations in the metaphase form of the hetero-
chromosome as seen in sections, and d, e the metakinesis of this
chromosome, the large size of these two figures being due to the
fact that the figures were drawn from acetocarmine preparations.
Further details will be given later.

BRYN MAWR COLLEGE,
January 5, 1911.

Vol. XX. February, 1911. No. 3

BIOLOGICAL BULLETIN

THE METHOD OF CELL DIVISION IN THE
DEVELOPMENT OF THE FEMALE SEX
ORGANS OF MONIEZIA.1

A. RICHARDS.

TABLE OF CONTENTS.

Introduction 123

Review of the Literature 126

Material and Methods 130

Fixatives 132

Observations 134

The Primary Anlage 135

The Oogonial Period 141

The Growth Period and the Behavior of the Nucleolus 142

Synapsis 144

Yolk Production 144

The Ovarian Egg 145

The Vitellarium 145

The Female Genital Ducts 147

Fertilization and Maturation 150

Cleavage 152

Summary of the Evidence • 155

Discussion 158

Comparison with Tcenia .' 160

Conclusions 161

Literature References 162

Explanation of Plates 164

INTRODUCTION.

Cestode histology has been the subject of several investigations
in recent years, some of the results of which are of far-reaching
importance. Throughout the development of Moniezia Child
has reported the regular occurrence of amitosis; he has described
it as present not only in the somatic tissues but also in the germ
cells. It was this latter phase of his results which occasioned
the present work.

Zoological Laboratory, Princeton University, January 26, 1911.

123

124 A- RICHARDS.

There are inherent in the very nature of the problem involved
in the recognition of cases of amitotic cell division as such certain
difficulties which render the study of the question upon fixed
material more or less unsatisfactory. Mitoses are, of course,
well adapted for such a study, although they may be obscured
in various ways. The presence of definite chromatin bodies and
the absence of a nuclear membrane at most stages makes their
recognition in properly treated material comparatively easy.
But positive evidence of amitosis in fixed material is difficult to
obtain. One cannot always be sure that a nuclear constriction
is not a product of mechanical compression, e. g., by yolk granules,
or an artifact due to reagents rather than a stage in division.
A strand of linin stretched across a nucleus with chromatin
granules upon it often gives the appearance of a membrane
dividing the nucleus amitotically (endogenous division?). Nu-
clear lobations are frequently seen in cells which are known to
multiply by mitosis. Finally, a reduplication of nuclei for any
of several reasons — disintegration, accessory sperm nuclei in
egg cells, the male and female pronuclei themselves, lagging of
cytoplasmic division behind nuclear mitosis, incompletely fused
chromosomal vesicles, etc., easily presents a simulacrum of
amitosis which is quite misleading. These and other factors1
all render the amitosis problem difficult of solution.

Clearly, however, a definition of what one means by amitosis
and of what he will accept as evidence of that process is impera-
tive. Dahlgren and Kepner (p. 38) define amitosis as division
"by a series of autoconstrictions of first the nucleolus, then the
nucleus, and lastly the cytoplasmic body." The nucleus itself
divides in three ways: the constriction of its body, the formation
of a "nuclear plate" in the plane of division and subsequent
separation, and the endogenous method of Child. Wilson (p.
114) describes the process as follows: "First the nucleus remains
in the resting stage (reticulum), and there is no formation of a
spireme or of chromosomes. Second, division occurs without
the formation of an amphiaster. . . . The nuclear substance
undergoes a division of its total mass but not of the individual
elements or chromatin granules." It will be seen that these

1 Mitosis may be periodic and of very short duration.

METHOD OF CELL DIVISION IN MONIEZIA. 125

definitions are not sufficiently diagnostic to be of use in the work
here undertaken.

For amitosis there is but one absolutely certain criterion, the
observation of living material and subsequent study of material
fixed under observation, and this is, of course, impossible in most
cases. In the absence of such a criterion only complete series
of all stages in the constriction and subsequent division of the
nucleus and attendant cytoplasm would seem sufficient warrant
for the assertion that amitosis is the prevailing form of division.1
Even complete series are no assurance that part of the cells do
not divide mitotically or that mitosis does not occur periodically.
But given a complete series — assumed or proven — and taking no
account of periodicity, etc., the interpetation of doubtful cases
on the basis of the complete series follows logically. As will
appear later such a complete series is extremely difficult to estab-
lish in the cestodes; its assumption, therefore, in the face of the
above mentioned difficulties is ultra radical. The burden of
proof is upon the shoulders of him who makes the assumption.

It is true, however, that a spirit of open-mindedness demands
an impartial consideration of uncertain cases, incomplete stages
and doubtful processes in the hope that an unprejudiced decision
may be reached. The refusal to consider such cases is as un-
scientific as is the drawing of sweeping generalizations from them.
In such a spirit of investigation I propose to set forth the facts
as I see them in the case of the anlage of the female sex organs
and its derivatives in Moniezia. I then wish to sum up the
evidence for and against the occurrence of amitosis in this form
and to consider various a priori suggestions which have been
made by investigators to cover this case. Finally, I wish to
make a comparison of these results with those of Tcenia upon
which I previously worked. In the development of the anlage of
the female sex organs the history of the sex cells themselves, of
the vitellarium and of the ducts and canals will be considered.

Interesting in this connection are the recent comments upon
Child's position made by Boveri.2

lTo forestall criticism of this point, let me say that this is in no sense a precon-
ceived hypothesis. I had worked on the amitosis question nearly two years before
the significance of this contention was fully borne in on me. It is given thus early
in the discussion merely to make the account more intelligible.

2"Zellenstudien," 6, p. 235.

126 A. RICHARDS.

"If, his (Child's) observations are to prove his contentions
then it must be shown: (i) that the binucleate condition
which he finds really depends on a division; (2) that about each
of these nuclei a part of the protoplasm is marked off, and (3)
that the cells so arising again divide by mitosis and possess the
normal number of chromosomes."

REVIEW OF THE LITERATURE.

The literature of late histological researches upon the tre-
matodes and cestodes deals with the general development of the
flat-worm tissues, especially with the cuticle and the paren-
chyma. The more recent papers also touch upon the question
of the occurrence of amitosis in the life history of the various
cells.

Cestode histology according to this late work presents some
very unique features. All investigators who have given the
subject their attention seem to concur in the view that tissue
growth here differs in character from the usual method of cell
multiplication. All agree that karyokinetic figures are not to
be found as numerously as in other tissues; some assert their
entire absence; while one goes so far as to claim that nuclei arise
de novo. On the other hand, some papers make no mention of
amitosis, possibly due to the fact that the authors have not
interested themselves with the problem. A comparison of the
cases in which amitosis occurs is difficult to make because many
reports, of thirty or more years ago, previous to the introduction
of refined methods of technique, have not been substantiated by
later work; but, nevertheless, it is safe to say that in no other
group of animals have such unique growth phenomena been found
by reputable workers. Granting for the moment that these
phenomena have been correctly reported, one is compelled to
ask what the relation is between them and the parasitic mode of
existence which is universal among cestodes. While as yet there
is no basis for speculation even on this relation, the modifications
of gross structure due to parasitism are so profound in the ces-
todes that one may well be deterred from generalizing from
observations of peculiar histological conditions in them.

Bugge ('02) was the first to note the absence of mitosis in
cestode tissues in his work upon the flame cells.

METHOD OF CELL DIVISION IN MONIEZIA. 127

Child in 1904 called attention to the lack of mitoses in the
reproductive organs of the cestode Moniezia and asserted that
amitosis takes its place. He further elaborated this conception
and furnished more data in 1906 and 1907. As the object of
this research is either to corroborate or to disprove his conclusions,
and as it deals with them at length later, no further attention
will be given to his earlier papers in this place. In a later paper
('io) he brings forward no new evidence but merely reiterates
his former position. This work consists largely of an attack on
my observations on Tcenia ('09). He criticises my paper as
presenting evidence of a negative character only; as being upon
an entirely different genus and perhaps on different stages from
those in which he observed amitosis; as prematurely generalizing
from incomplete data; as using preconceived hypotheses; and
finally, as being the work of a novice who "is likely to be guided
to a greater or less extent by the views of his instructor." I may
venture a reply to the main points of this list. I regret to have
my paper taken solely as a criticism of Child's work; it recounted
the facts as I believe them to exist in Tania; its field, it is true,
is limited to the oogenesis and cell multiplication of certain so-
matic tissues in this worm, but no attempt is made to apply its
conclusions to any of Child's work not upon a similar field (a
fact which he does not seem to have appreciated). I found no
evidence of amitosis in that field and as Child had covered a
similar one it was not only natural but necessary that I should
compare my results with his. As to the negative character of
my evidence, I would ask whether the failure to find evidence of
a certain process, using proper methods and exercising due dili-
gence, is not positive evidence of the lack of that process. Of
what more would positive evidence consist ? As to my not having
studied the proper stages little need be said ; no method of pro-
cedure other than beginning at the scolex and working backward
ever occurred to me. As to the fact that I studied a different
genus, he says, "Such records (referring to the results published
in his previous papers) cannot be controverted ... by observa-
tion upon another genus and species. . . ." His views upon this
point, it seems, have undergone a change. In his "General
Conclusions" to "Amitosis in Moniezia" ('04) he said, "It is

128 A. RICHARDS.

scarce y possible that Moniezia differs from the majority of
other forms in the amitotic multiplication of nuclei." There are,
however, a considerable number of differences in the results ob-
tained in the two allied forms, differences which I hope to make
clear in the course of this discussion.

Spengel in 1905 before the Deutsche Zoologische Gesellschaft1
stated that neither mitotic nor amitotic cell division had been
observed in cestode development.

C. v. Janicki ('07) in an important paper upon the embryonic
development of T&nia serrata states that he finds no pronounced
division figures up to the growth period of the oogonia, but
during the growth period he describes spiremes; he does not
suggest amitosis as the explanation of his failure to find mitoses
but merely says that the divisions must have passed very quickly.
In the cleavage stages achromatic parts of the division figures
are often hard to demonstrate.

In so far as it touches the same points my own work ('09) on
this form agrees with that of v. Janicki. I recorded my failure
to find amitosis and noted that mitoses are more infrequent than
one would expect.

In 1908 appeared the paper by Young upon the histogenesis
of Cysticercus pisiformis, the cysticerus stage of Tcenia serrata.
This paper is revolutionary indeed, and agrees in practically no
cytological points with the work of v. Janicki and myself. Not
only does he accept Child's conclusions and their bearing upon
hereditary problems but he refuses to subscribe to Virchow's
dictum "omnis cellula e cellula." He believes that nuclei arise
de novo in accordance with a tentative hypothesis, of which his
statement is as follows: "I believe that the nucleus in these
forms is not a morphological but a physiological entity; that the
nuclear granules are fundamentally the same as the remaining
protoplasm of the cell, but are differentiated therefrom under
physiological conditions which wre do not at present understand ;
that these granules are perhaps reserve material stored up in the
nucleus for future use, the entire cell body being occasionally
converted into a nucleus; and that the nucleus varies in structure
from time to time in response to the varying physiological de-

Child, 'o^a, p. 277.

METHOD OF CELL DIVISION IN MONIEZIA.

mands made upon it. . . . Further, if my interpretation of my ob-
servations be correct, then distinction between germ and somatic
plasm is obviously impossible; a special vehicle for the trans-
ference of heredity qualities is entirely wanting; such qualities
must be transmitted by the undifferentiated protoplasm; cell
lineage is manifestly lacking; a mosaic theory is plainly un-
tenable; and the fate of any given embryonic element — whether
it shall form parenchyme, muscle, nerve, etc. — must be deter-
mined by physiological causes alone." In an appendix referring
to Child's later papers Young says that he has verified Child's
conclusions upon the reproductive organs of TcEnia serrata. This,
of course, is exactly opposed to the work of v. Janicki and myself
upon T. serrata. I am not aware that he has published any
further account of these more recent observations.1

Balss ('08), in his extensive researches on the development
of the sex ducts of Anoplocephala and Solenophorns did not, I
believe, direct his attention to the method of nuclear division,
nor do his figures throw any light upon the question. His paper
is extremely valuable in making a study of the early stages of
sex organ formation.

The last of the important papers on cestode histology is that
of Spatlich ('09) on Tetrabothrius. This author states that he
finds cases of amitosis and gives several figures of the cells in
question. He does not doubt that the cells are in the process o
amitotic division and certainly that is the easiest explanation
of the figures; but to one skeptically minded the proof will not
appear as conclusive. The cell in his Fig. 54, for example, might
easily be interpreted as a case of compression; while there is no
evidence that either of Figs. 54, 55 and 56 (drawn out nuclei, be-
coming dumb-bell shaped) would actually have completed the divi-
sion. Nothing is said of a complete series of stages. Further-
more, these are cases in the development of the " Dotterzelle "
and may perhaps have been at the end of their life cycle.

In glancing over the list of tape-worms in which amitosis or
other irregularities of cell multiplication have been described, one

'In Science of February 17, Young gives a very brief resume of this work.
Although he finds no mitosis, too few amitoses occur to account for the neces-
sary cell multiplication; he thinks that the "increase is partly due to development
of nuclei either de novo or from chromidial extrusions of preexistent nuclei."

I3O A. RICHARDS.

is struck with the fact that according to the classification usually
given only members of one family, the Taenidae, are included in it.
This, of course, may be the merest coincidence owing, perhaps,
to the fact that members of the Taenidse are very readily secured.
The warning uttered by Young with regard to applying con-
clusions drawn from cestode observations to other groups, is,
however, more strongly emphasized. Not only are the results
which he had in mind deduced from a single class of animals
which are degenerate in a high degree, but they are drawn from
a single family of that degenerate class.

MATERIAL AND METHODS.

At the suggestion of Professor E. G. Conklin, the genus Tcenia,
upon which I had begun investigation with a view to obtaining
further evidence as to the occurrence of amitosis in cestode
tissues, was dropped for the present as a subject for research in
favor of Moniezia. The results which have been obtained upon
this latter genus by Child are of such unusual nature compared
with those of other forms that a reexamination would seem
desirable. In the light of my investigations upon Moniezia and
of Child's recent criticism I have again gone over my old material
of Tcenia as well as some newly prepared.

In making these studies I am indebted to Professor E. G.
Conklin and to Professor Ulric Dahlgren for much kindly assist-
ance. My thanks are due to the management of the Wistar
Institute of Anatomy for assistance in obtaining material. Not
only were the facilities of the laboratory extended me for making
collections but my first lot of material was obtained and fixed
by Dr. Helen Dean King. To Dr. King I also owe thanks for a
number of technical suggestions. Professor C. M. Child, of the
University of Chicago, has also kindly placed facilities at my
disposal for obtaining specimens.

Specimens of Moniezia expansa and of M. planissima have
been secured and fixed in a variety of fixing fluids: Flemming,
Hermann, Graf's chrom-oxalic, saturated aqueous solution of
corrosive sublimate, formol-sublimate,1 Kleinberg's picro-sul-

1 Physiological salt solution saturated with corrosive sublimate to which a few
drops of neutral formol are added.

METHOD OF CELL DIVISION IN MONIEZIA. 131

phuric, Zenker's fluid, Zenker plus an excess of acetic acid, and
a modification of Zenker's fluid by Dr. King.1

None of my sections have been cut over 4 micra thick; many
of them are thinner. Pieces of the worms about 3-6 mm. in
length were embedded serially beginning at the scolex or the
neck region and extending back some 35-40 cm. usually, in-
cluding the beginning of the proglottids with fully formed
embryos. Pieces were then selected at intervals and cut;
later, if examination warranted, the intervening pieces were cut.
In this way practically one entire worm has been cut and studied,
while from a considerable number of other worms incomplete
series have been made as described. I have used both longi-
tudinal and cross sections.

A number of staining methods have been tried on Moniezia,
some of which gave good results, while others succeeded only
indifferently. I have been unable to get good preparations
with Kernschwarz, although this stain served me well on Tcenia.
The other stains used include Delafield's hsematoxylin, haem-
alum, thionin, Bismarck brown, iron haematoxylin alone and
counterstained with Lichtgriin or Bordeaux R, a modification
of Conklin's picro-haematoxylin method for staining eggs in
toto? Ehrlich-Biondi, and the Flemming triple stain, safranin
gentian-violet and orange G.

Of these stains, iron haematoxylin is the best general stain and
most reliable, but for certain phases of the work it is not suffi-
ciently differential. I have found picro-hsematoxylin and the
safranin gentian-violet mixture especially useful in connection
with it. (In most cases I find that the orange G of the triple
stain detracts from rather than adds to its value.)

1 Dr. King uses two solutions, as follows: (i) Four per cent, potassium bichromate;
(2) corrosive sublimate, 4 gr., glacial acetic acid, 20 c.c., water, 76 c.c. The two
solutions are mixed equally and allowed to act for one to two hours as the mixture
penetrates rapidly. For convenience, I shall hereafter call this mixture "King's
fluid."

2 This is an application of the in toto method to sections containing eggs. It
consists of mordanting the sections for a short time — too long is injurious — in
aqueous picric acid solution. After staining the sections may be decolorized in
this same solution and then blued in tap water. It gives the desired differentiation
between yolk and chromatin, although not with the characteristic colors of Conk-
lin's stain.

132 A. RICHARDS.

Fixatives. — To show exactly the various effects of the fixatives
used and to show that the results obtained are not due to the
exclusive use of one reagent, I wish to give here a detailed state-
ment of those effects.

In his work on Moniezia Child used some eleven fixatives of
which the most successful were Hermann's fluid, saturated
aqueous solution of corrosive sublimate, and Graf's chrom-
oxalic mixture. I, therefore, used these three fluids and in
addition certain others which my own experience suggested.
These were, as previously stated, Flemming's and Zenker's
fluids, Zenker plus an excess of acetic acid, and formol-sublimate.
Dr. King fixed some of my first lot of material in her modification
of Zenker's fluid. I have no hesitation in saying that this fluid
gives the best results for Moniezia of any which I have tried.
Subsequently, I myself have used this solution with equally
satisfactory results.

The results of my experiments with the various fluids follow.
Chrom-oxalic is not, I find, a good general fixative for Moniezia.
The general cytoplasmic structures, parenchyma, etc., are not
well preserved and the achromatic substances of the nuclei are
dissolved out. The chalk bodies1 also are not retained. Testis
cells are more satisfactorily fixed, as are also the cleavage cells
and young embryos. But in the case of spermatozoa and the
young female germ cells the mixture is not successful. Nuclei
are everywhere swollen, but in these special cases they are also
distorted. I should not care to trust any results based on the
observation of early oogonia in my chrom-oxalic material.
Many of Child's results are founded on chrom-oxalic material;
doubtless he had greater success with this mixture than I did.

Hermann's is a good fixative for the developing germ cells,
both male and female; for parenchyma the results are slightly
less satisfactory. Nuclear contents are well preserved in this
solution — much better than in the preceding one. It also gives
a good preparation of the cuticle. With Flemming's, Hermann's
shares the drawback of fixing the calcareous bodies so that they

JAs part of their "generic diagnosis" Stiles and Hassal give this statement,
"Calcareous bodies absent from the parenchyma." They, however, studied only
sublimate material.

METHOD OF CELL DIVISION IN MONIEZIA 133

stain black with certain dyes. This is a serious drawback to the
use of osmic acid fixatives, for the chalk bodies are so numerous
as to obscure much of the development of the reproductive
organs.

The remarks concerning Hermann's apply with equal force
to Flemming's although the latter gives a somewhat more
nearly perfect fixation considering the tissue as a whole. For the
cuticle Flemming's gives better results than any other reagent
I have tried.

My experiments with corrosive sublimate in aqueous solution
have proven a complete failure. Circumstances connected with
the experiments were such, however, that I am not disposed to
regard my results as typical; had I not already had a great
number of successful preparations I should have given it further
trial.

With formol-sublimate I have obta:ned good preparat ons. It
gives quite good nuclear and cytoplasmic fixation as well as a
very fair preservation of the parenchyma. Nuclei and cytoplasm
stain dark after it; chromatin is well fixed but achromatic struc-
tures are not easily differentiated owing to the homogeneous
condition of the nuclei ; cytoplasm, too, is of a dense homogeneous
nature not unlike that of certain nerve cells. Calcareous bodies
are not present in sublimate material. Formol-sublimate is,
however, a good general fixative for Moniezia.

Picro-sulphuric, while giving somewhat better results than
aqueous sublimate (the young embryos are fairly well preserved)
is also a failure. It gives a better fixation of the parenchyma
than of other tissues.

There yet remain the Zenker solutions. These give the best
fixations, and of them the modification by Dr. King is the most
satisfactory of all the fluids I have used on Moniezia. I find,
too, that I am able to get better staining reactions on the material
fixed in this solution. My study has been made chiefly on this
material but I have in every case compared the results with those
of the other fixatives. The fixations by chrom-oxalic, Hermann's,
formol-sublimate, and King's solution all depend on different
actions; therefore, it would seem that the phenomena common
to these four reagents cannot be viewed as artifacts due to
fixation.

134 A- RICHARDS.

OBSERVATIONS.

Morphological discussions are without the scope of this re-
search, but a knowledge of the derivation is helpful to an under-
standing of the development of the structures dealt with. Ex-
cluding the cuticle — the writer has not given much attention to
the "ectoderm question," but his view, in so far as his observa-
tions warrant one, is that the cuticle also is of parenchymal
origin — all of the adult structures of the tape-worm are derived
from the parenchyma. Parenchyma is physiologically similar to
embryon:c mesenchyme in that from it are derived muscle fibers,
flame cells, etc. The testes arise from the parenchyma first in
the region of the longitudinal excretory tubes and later appear
medianward, occupying the side of the worm commonly desig-
nated as dorsal. The animal is incompletely protandrous for
the ovaries arise later and many of the testes have completed
their development and degenerated before the female germ cells
have progressed very far. The last of them do not complete
their cycle until most of the ova have already been fertilized
but all ultimately degenerate and their place is taken by the
branches of the uterus.

The first anlage of the female reproductive apparatus arises
also in the neighborhood of the excretory tubes and grows both
laterally and medianward. Its lateral growth gives rise to the
seminal receptacle and the vagina. Its median growth is the
beginning of the oviduct, which later branches to form the uterus,
of the vitellarium and shell glands, and of the ovary. Thus the
entire female apparatus arises from an originally single, indiffer-
ent anlage. This shortly differentiates into the separate anlagen
each of which from now on pursues its own individua' course.
That of the ovary is the innermost part of the original anlage.
It develops its investing membrane and begins a growth, due to
the growth of the germ cells composing it, which ends only when
the ova have all attained their complete size and developed the
needed amount of yolk. These in a rapid passage through the
oviduct are fertilized and enclosed in a thin shell. Maturation
is completed in the uterus where cleavage occurs and early
embryonic life is passed. When all of the ova have been fertilized
the reproductive mechanism except the uterus degenerates. The

METHOD OF CELL DIVISION IN MONIEZIA. 135

entire process of development from primitive germ cells to fully
formed embryos is one of comparative simplicity and is quite
easily followed with the exception of the cleavage stages. These
are difficult to understand in section, and whole mounts of them
are not readily prepared.1

The vitellarium arises from the middle part of the primary
indifferent an age, and in its early stages cannot be distinguished
from the ovary other than by its position. But its growth is
of a different character and soon permits of a distinction between
the cells of the two regions. It acquires the character of a gland
and thereafter its activities are those connected with secretion.

The development of the various female organs will now be
considered in detail.

The Primary Anlage. — The indifferent or primary anlage be-
gins merely as a thickening of cells median and slightly ventral
to the longitudinal nephridial canals. This thickening is not
marked off in any way from the surrounding parenchyma; in
fact there is a gradation from one into the other. There is a
slight difference in the character of the cells of the an'age and,
of the adjacent parenchyma but it is only slight. The cytoplasm
is perhaps a little more dense in the anlage, and the nuclei show
evidence of greater activity. The cells of the parenchyma are
clear and usually have but one nucleus. The cells of the anlage
stain darker, and their nuclei frequently possess two or more
nucleolus-like bodies, of which probably only one is a true nu-
cleolus, and show a more or less distinct reticulum, except when
fixed with chrom-acetic. The ratio of nucleus to cytoplasm is
at first somewhat greater but soon becomes less in the case of .the
anlage than of the parenchyma. Nuclei in the former are often
somewhat smaller than in the latter.

To avoid confusion with regard to my use of the term "cell,"
I will here state my view of the cellular nature of the parenchyma.
I do not consider the parenchyma to be a syncytium as that term
is generally interpreted. Nuclei are always surrounded by a
definite amount of cytoplasm, but that cytoplasm is lacking in
a definite limiting membrane or cell wall.2 The remainder of

JCf. for this general development photomicrographs I. to IV.

2 My former statement that "the parenchyma cells of Tcenia have definite cell
boundaries" was not intended to mean that they have cell walls; I have not observed
that the two genera differ in this respect.

136 A. RICHARDS.

the parenchyma tissue is made up of loose intercellular material
(of cellular or'gin, of course). Thus a cell consists of a nucleus
and its attendant cytoplasm supported by a large mass of inter-
cellular material. Usually the cells are somewhat distant from
each other. Should two or more of them come to lie closely side
by side, it is conceivable that their cytoplasmic masses might
easily fuse and a true syncytium be formed. It will be seen
that this is a possible explanation of binucleate parenchyma cells
where such occur (Fig. 2). The cellular character of the primary
anlage is similar to that of the parenchyma, although, of course,
the anlage has very little intercellular material and many cells.

Child describes the testes as arising from visibly differentiated
muscle cells; he did not, however, observe that muscle cells take
part in the formation of the ovary, "probably because these cells
do not occur in the region where the ovaries develop" (Child,
'06). Muscle cells certainly do occur in the region of the
ovaries1 but I have never seen any evidence that they take
any part whatever in the development of the female germ cells.

Since the primary anlage arises from a parenchyma which has
no visible differentiation there would seem to be no visible basis
for a continuity of germ plasm here. The expression "primordial
germ cell" has little meaning therefore when applied to a cestode
for there is no apparent difference, except of position,2 in the
cells which will go to form the ova and those which will form the

v

vitellaria or the walls of the ducts.

The primary anlage grows both outward toward the edge of
the proglottid and inward toward the ventral surface. Its out-
ward growth is in the form of a cord of cells being extended to
the exterior. Before the solid cord reaches the exterior its middle
portion becomes hollowed out to form a lumen. This lumen
subsequently develops a cuticle which to all appearances is just
like that on the surface of the body. This outwrard growth
becomes the vagina and the receptaculum seminis. Its develop-

'See photomicrographs for circular and longitudinal layers; scattered fibers
also occur.

2 This difference of position, however, is a very constant one. The cells from which
the ova develop are early marked off by their position and never take part in any
other development. The facts are not opposed to the continuity theory; they simply
afford no evidence for it.

METHOD OF CELL DIVISION IN MONIEZIA. 137

ment takes place somewhat more slowly than that of the inward
proliferation of cells, due, doubtless, to the fact that its function
is exercised only relatively late.

The inward "proliferating area gradually extends somewhat
toward the median plane and somewhat toward one surface of
the proglottid known as the ventral surface. ... As the inner
and ventral end of the proliferating area approaches the inner
layer of circular muscles it spreads out into a flattened somewhat
disc-like area exactly as if it had encountered resistance to its
growth in the original direction and so had begun to spread out
in other directions."1 A little later the posterior portion of the
anlage of the ovary becomes the anlage of the vitellarium which
develops about the middle part of the oviduct. This origin
shows that morphologically the vitellarium is part of the ovary
and gives credence to the view expressed for similar cells of other
animals (test cells of ascidians, etc.), that the cells are rudi-
mentary eggs.

There are two regions concerning whose method of growth it is
especially important to have definite knowledge. These are the
primary anlage and the early embryos. Concerning the former
there are four possibilities as to the growth of the cells: origin
de novo, by migration from the surrounding parenchyma, and
by cell division direct and indirect. The first of these as Child
remarks ('10, p. 112) is "an alternative which most of us would
hesitate to accept without proof of the most conclusive char-
acter."

Migration of nuclei to this region, if one could show that it
occurred, would explain the relative lack of mitoses in the an-
lagen. It would then be necessary, of course, to seek for mitoses
or amitoses in the neck region to determine the method of origin
of the cells. But actual evidence of migration is as impossible
to obtain as is evidence of amitosis. At best only indirect evi-
dence of it can be secured.

If abundant divisions occurred there would be little probability
of migration, although it might occur. Abundant divisions are
not to be seen, however. On the other hand, the fact that many
of the nuclei in the primary anlage are smaller than the paren-

1 Child, '076, p. 99.

138

A. RICHARDS.

chymal nuclei suggests that they have divided. Still a third
line of evidence is obtained by a study of the number of nuclei in
given regions before and after the formation of the anlage. To
make such a study it is not sufficient to count the nuclei in a
single section of each proglottid and compare the counts directly,
for the proglottid grows in three dimensions. Either the entire
number of nuclei in all the sections to be compared must be
counted or the ratio determined by a mathematical calculation
based upon counts from different fields of the two sections, meas-
urements of the inner parenchymal ellipses (that part of the
parenchyma enclosed by the circular muscle layers) and the
number of sections of the proglottid.

Such a calculation has been made and its results are given in
the following table. While the problem does not lend itself fully
to the rigor of a mathematical treatment, it is believed that the
results are approximately accurate, at least sufficiently so for
the sake of a comparison. In each case two proglottids from the
same chain, one near the head, the other farther back, are com-
pared; by nuclear ratio is meant the ratio of the number of
nuclei within the circular muscle layer of the one to that of the
other; the volumetric ratio represents the ratio of the volumes
of the parenchyma within the muscle layers for the same pro-
glottids.

Case.

Stage of Development.

Nuclear Ratio.

Volumetric
Ratio.

Al
A2

Testes formed, ovary anlage separate.
Growth period.

6 : 10

7 : 10

Bi

B2

Ovary anlage just separated.
All organs formed, merely growing.

i : 2

i : 5

Ci

C2

Primary anlage stage.
Late growth period.

i : 7

i : 17

Di

D2

D3

Early primary anlage.
Ovary nearing growth period, testes almost
all formed.
All organs formed.

2 : 9 : 10

2:9 = "^

Ei

E2

Early primary anlage.
Testes formed, ovary in growth period.

i : 6

I : 7

Fi

F2

Primary anlage stage.
Testes formed, ovary in early growth period.

i : 7

i : 7

METHOD OF CELL DIVISION IN MONIEZIA. 139

Of the six cases here given, believed to be fairly representative,
in only one, Ai and A2, have the nuclei increased in greater
proportion than the volume, and even here the increase is only
slightly greater; in two, Di and D2, and Fi and F2, the two
ratios have kept pace with each other; while in four, Bi and B2,
Ci and C2, Di and D3, and El and £2, the nuclear ratio has
fallen more or less behind the volumetric. Now evidences of
division are no more — and no less — common in the parenchyma
cells than in the anlage so the result of these four cases favors
somewhat the notion of migration of the parenchyma nuclei to
the anlage for there are relatively fewer nuclei in the older pro-
glottids and it is hardly probable that degeneration, the opposing
factor, would play a part in the decrease in the number of nuclei
in as young proglottids as these are. But on the whole the evi-
'dence for migration is not strong; certainly this factor cannot
have been the important one in the formation of the anlage of
the female organs.

It would seem, then, that here as is general throughout the
animal kingdom germ cells must arise by cell division. It is
the inconspicuousness of these processes that occasions the pres-
ent discussion.

Mitosis unquestionably occurs. Child records and figures oc-
casional cases. Although he holds that they are too infrequent
to account for the cell proliferation he finds them in nearly every
stage of development. He has, however, never found them in
the early stage of the duct formation, in the primary anlage,
unless his recent account of a single individual consisting of
scolex, neck region, and a few of the youngest proglottids in
which nearly every nucleus is dividing mitotically may include
such cases. He has given no definite statement of the stages
included in this worm, but certainly the primary anlage appears
in the proglottids which are quite young. In these earliest
stages I have found a considerable number of mitoses although
they are by no means so abundant as to be seen upon a cursory
examination. Let me call attention to the fact that in almost
any tissue the ratio of resting to dividing cells will be found to be
very great. This ratio is a surprising one.

In the oogenesis of Planorbis, a form in which divisions are

140 A. RICHARDS.

of the usual mitotic type is shown in a research, as yet unpub-
lished, by Conklin, the division figures are very difficult to find.
One may examine section after section without seeing a single
case of mitosis. Although I have not made a statistical study
of the two cases I am of the opinion, from the examination of the
slides which Professor Conklin has kindly shown me, that division
figures are more rare in the oogonia of Planorbis than in those of
Moniezia. The ratio of resting cells to dividing cells is large.
Yet there is no evidence for the occurrence of amitosis here.
Another better known example is found in the oogonia of Ascaris.
Any one who has searched for oogonial divisions in this animal
knows how few cells show them in proportion to the number
present.

As development progresses mitoses occur as stated up to the
growth period, but, it is true, not as frequently as one would
expect from conditions in other tissues. I do not find that the
figures are always as easily recognizable as are mitoses in many
tissues.

That mitoses occur throughout the development of the female
germs cells in considerable numbers, there is not the slightest
doubt. The evidence for this is positive. Furthermore, they
always occur in the peripheral, i. e., the growing portion of the
anlage, a very significant fact. It requifes merely a little careful
search upon properly fixed and stained material to discover them.
Tissues fixed in King's fluid and stained with the Flemming
triple stain are especially favorable for finding the mitotic figures.
They do not occur as frequently as one would a priori expect,
but, in the light of the fact which will appear immediately, i. e.,
that there is no good evidence for the occurrence of amitosis,
we are merely forced to readjust our ideas of the number of
divisions necessary to produce a tissue. The mitotic frequency
has not been sufficiently considered. I regard it as also highly
probable that periodicity of division is an important factor in
tissue production.

There are to be seen here many cases of nuclei which lie close
beside one another with only a narrow layer of cytoplasm be-
tween them. There have been some few cases in which no layer
could be distinguished, but the character of the material is such

METHOD OF CELL DIVISION IN MONIEZIA. 14!

that this easily lies within the limits of observational error.
These cases may be considered as evidences of recent division;
they are not, however, distinctive of either direct or indirect
division. If there were abundant evidence of the former they
would be regarded as adding to it. On the other hand, if spindles
were very numerous this arrangement in pairs would be looked
upon as the final stage in the division of the cells by mitosis.
Cases of juxtaposition of nuclei are seen in Figs. 2, 12 and 17.

One may occasionally see in some sections a few nuclei darker
than their neighbors (Fig. 18) scattered among the pre-oogonia.
They have no visible significance; they differ from other nuclei
only in being darker. I have never seen one half of a nucleus
darker than the other. I may here add that there is never a
dark body, a "Nebendotter" occupying a portion of an ovum
as in Tfsnia.

Rarely, indeed, does one find constricted nuclei in my prepara-
tions. Shrunken and irregular nuclei — artifacts — occur; strands
of linin across a nucleus often resemble the formation of a plate
(as in Figs. 41 and 42) ; the edges of adjacent nuclei overlap and
sometimes the overlapping parts are so thin that the true con-
dition is not at first apparent; but the various stages of plate
formation, constriction, etc., and especially of "endogenous"
division as set forth by Child I have never seen. One does,
however, occasionally find a dumb-bell-shaped nucleus; but so
infrequent are they that I am unable to determine any signifi-
cance for them.

I find myself, therefore, unable to confirm Child's conclusions
for the various reasons given. His position, however, that the
"relative frequency of mitosis and amitosis in certain species,
and even in single individuals may vary greatly according to
conditions" ('10, p. 116) is an unassailable one with our present
data.

The Oogonial Period. — From the disc-like anlage on the ventral
side of the proglottid upward projections or follicles develop and
give to the ovary its characteristic form.1 About the time of full
development an enveloping membrane appears probably as a
differentiation of the surrounding parenchyma. For the female

xCf. Photomicrograph III.

142 A. RICHARDS.

germ cells from the differentiation of the ovarian anlage to the
formation of the membrane of the ovary the term " pre-oogonia "
has been (aptly, I think) suggested by Professor Dahlgren (see
Figs. 9-14). It is only at the end of the pre-oogonial period that
the cells which are to form ova are certainly distinguishable
from all others. The convenience of the use of this term will be
apparent. The formation of the follicular membrane seems to
mark the time of cessation of active proliferation on the part of
the oogonia. I have seen no sign whatever of cell division in
stages following this. There are no longer many parenchymal
elements to be seen in the ovary and each oogonium becomes
sharply marked off from its neighbors. The subsequent growth
of the ovary is merely that attendant upon the growth period
of the oogonia. The cell divisions are, of course, limited to the
maturation and cleavage divisions.

The Growth Period and Behavior of the Nudeolus. — The growth
period in the cestode as in most animals is relatively long; many
more proglottids have the ova in this condition than in any
other previous to the cleavage stages. The actual duration of
the growth period is probably three or four months.1 It is
well known that sheep older than yearlings are rarely infested
with tape-worms. The parasites occur only in the young, but
nevertheless they attain a length of 4-5 meters in the case of
M. expansa and of 2 meters in the case of M. planissima. De-
veloping embryos first occur at a distance of 100-110 cm. from
the head; the primary anlage appears 7-8 cm. from the head;
and the growth period of the oogonia begins about 20 cm. from
the head. This short duration of the stage in which multiplica-
tion of the oogonia is taking place must be kept in mind when
investigating the method of cell division here. To Child's de-
scription of the growth period little can be added. Briefly, there
is accomplished during the period the great growth of the eggs,
synapsis, the formation of the yolk globules, and certain con-
temporary changes in the nucleolus to be described immediately.
During this period of no divisions the oogonia after synapsis
are characterized by a minimum amount of chromatin, faintly
staining, distributed peripherally, although strands may reach
the nucleolus.

1 These time limits are tentative for I have been able to approximate them, only

METHOD OF CELL DIVISION IN MONIEZIA. 143

Concerning the structure of the nucleolus more can be seen
during the growth period than at any other stage of development.
In favorable preparations, the nucleolus in oogonia, oocytes and
segmentation cells shows a peripheral portion stain:ng violet
with the triple stain, and a central portion, the so-called vacuole,
staining with safranin. During the growth period additional
details appear. These are (Figs. 28-30), within the vacuole, a
small more or less refractive body, the endonucleolus (cf. Mont-
gomery, '99), staining very dark with gentian violet, a reticulum
supporting this body, and in the outer peripheral portion of the
nucleolus certain darker bodies. Certain appearances here very
much resemble the conditions described in the karyosome cycle
of Diplodiscus by Cary ('09). There is, however, no parallel in
the deve opment.

Only one true nucleolus is present in a cell. Besides this there
are frequently one or more chromatin bodies, karyosomes, which
with iron haematoxylin stain like nucleoli. But the gentian violet
or Ehrlich-Biondi preparations never show more than a single
true nucleolus.

As to the nature of the true nucleolus there is not much evi-
dence. Young deplores the use of the term "nucleolus" and
calls the structure a "nuclear granule." It is from the nuclear
granules, he holds, that the nuclei develop; the granules in turn
arising de novo (probably) from a common "cytoblastema."
"Containing, as they do in many cases, most of the staining
matter of the nucleus, they represent rather the chromatin than
the true nucleoli." I am obliged to dissent from this view; they
are without doubt true nucleoli, as differential staining shows.
Nevertheless, in addition to their nucleolar character they may
well serve as chromatin reservoirs during the resting period of
the nucleus. Indeed, the peripheral portion stains with gentian
violet just as does chromatin, and the darker bodies easily sug-
gest chromatin masses. The fact that the nucleoli increase in size
like the other elements of the cell during synapsis is explained by
the swelling of the vacuole, for this occurs markedly; in many
cases several vacuoles are present. Furthermore, the nucleoli
are always in intimate connection with the spireme, and of course
they disappear when mitosis comes on. C. v. Janicki ('03) finds

144 A> RICHARDS.

that in the trematode, Gyrodactylus , the chromatin goes into the
nucleolus during the resting stage.

Synapsis. — Whether or not chromatin actually comes from the
nucleolus, the nucleus in the early part of the growth period
acquires an abundance of chromatin and a spireme stage at
once ensues. The spireme, in contact with the nucleolus, de-
velops rapidly and becomes massed on one side of the nucleus
just as in the "bouquet" or "synapsis" stage of the first matura-»
tion prophase. Although some time intervenes between this
and the prophase of the first oocytic division, I regard this as
probably a true synapsis. What seems to be conjugation of the
chromosomal loops occurs (Fig. 26) and all the characteristics
of a synapsis are present. Furthermore, precocious synapses are
not unknown (e. g., see Montgomery on Euchistus, '01). Be-
ginning in the medullary portion of the ovary, synapsis spreads
rapidly over the entire organ, involving nearly all the oogonia
at once, although they are in different stages in different parts.
At the same time the cytoplasm is increasing in amount but there
is no evidence of a causal relation between the two phenomena
except that they are synchronous. The nucleus is as a rule
excentrically placed and the "bouquet" is frequently found on
the side of the nucleus next to the greatest cytoplasmic mass,
but there are many exceptions to this arrangement. Following
the "bouquet" stage the spireme spreads throughout the nucleus,
loses its property of staining very densely and takes on the
appearance of the typical resting nucleus about to undergo ma-
turation. It remains in this stage for a relatively long period
during which yolk production is accomplished and the ovum
attains its full size.

Yolk Production. — -Yolk production rarely begins before the
"bouquet" stage has passed off. Usually upon the same side of
the nucleus as that where the "bouquet" occurred, but always in
a part where the cytoplasm is abundant, an area staining more
deeply than the rest of the cytoplasm is visible. "It is com-
parable to certain of the differentiations which have been called
yolk nuclei in other eggs and its appearance is followed almost
immediately by the formation of yolk granules which are con-
tained in the egg cell itself" (Child). Often two or more of these
yolk-producing areas are to be seen in the same ovum.

METHOD OF CELL DIVISION IN MONIEZIA. 145

In certain favorable material I can see a central dark granule
about which appears a clear brownish court (iron haematoxylin
material) and outside of this the yolk granules develop (Figs.
31 and 32). Whether this is general I am unable to decide;
probably, it is. Child thinks, and it seems a logical conclusion,
that the bouquet stage is responsible in some way for the yolk
production. Similar conditions have been described by Janicki
for Tcsnia serrata, and by several authors, particularly Gold-
schmidt, for trematodes. Goldschmidt1 thinks that during the
spireme stage a separation of somatic and germinal substance
occurs and that the "trophochromatin" escapes from the nucleus
to reappear later in the cytoplasm as yolk nuclei. Somewhat
analogous accounts have been given for the eggs of animals of
other groups.2

The yolk of Moniezia differs from that of Tcenia in that in
the former it occurs as spherules filling the entire cytoplasm
while in the latter there is only a single mass (in the earlier stages
more) as large or larger than the nucleus lying in the cytoplasm.

The Ovarian Egg. — The end result of these processes is the
ovarian egg. It is irregular in shape due to crowding by sur-
rounding cells and neighboring eggs, and contains a large germinal
vesicle located asymmetrically. Filling the cytoplasm more or
less completely are yolk globules of various sizes. The nucleolus
is now comparatively small and is located near the periphery of
the nucleus. Chromatin is distributed throughout the nucleus
in the form of fine granules attached to a linin reticulum. I
cannot say that centrosomes and centrospheres occur but I have
seen some structures resembling them (Fig. 34). In this condi-
tion the egg is ready for passage through the oviduct where
fertilization occurs. The lumen of the oviduct where it enters
the ovary is much smaller than the diameter of the egg so that
pressure is exerted on the egg during its passage causing it to
change its shape. After reaching the uterus it regains its spher-
ical form.

The Vitellarium. — Early in the development of the anlage of
the female sex organs there is set aside a certain portion of the

1 See Janicki, p. 695.

2 Cf. Conklin on Crepidula, '02.

146 A. RICHARDS.

cells to form the vitellarium and shell gland. It lies posterior
to the part which becomes the ovary arising from a group of cells
branched off from the middle portion of the anlage — that portion
which becomes the oviduct. The medullary portion develops
first becoming the- shell gland; from the periphery of this newly
differentiated anlage are proliferated cells which become the
vitellarium proper.

The morphological relations of this organ and the ovary suggest
that the vitellaria are fundamentally ova specialized along an-
other line than reproduction. This view, I think, is clearly
supported by the evidence.1 I have, therefore, endeavored to
find a parallel in their development; it is not, however, a very
close one. They agree in that they early pass through their
division stages and do not proliferate during the production of
yolk. (I have seen only a single case of division, and that not
clear, in a nucleus of that period.) They differ in that the oogo-
nial nuclei continue to increase in size long after the vitellarium
nuclei have reached their full growth and the former are always
more chromatic than the latter. The vitellaria never go through
the synapsis stage and the method of yolk formation is unlike
that of the oogonia. Small spherules which fuse with one another
are formed in the cytoplasm. The mass thus arising grows larger
pushing the nucleus to one side. The completed cell looks not
unlike a fat cell of the "signet ring" type.

Like the ovary the vitellarium at the time when cell multiplica-
tion stops develops a membrane separating it from the other
organs of the complex. Previous to this stage the cytoplasmic
boundaries have not been as clearly defined as those of the other
cells of the primary anlage, but now they become quite distinct
and the cells more dense. The nuclei develop more chromatin,
not in the form of a spireme as in the case of the ova, but as
enlarged granules giving to the nucleus the appearance of several
nucleoli. Sometimes strands may be seen extending from the
nucleoli out to the periphery where the yolk masses are forming.

The further history of the cells of the vitellarium is of great
interest from the standpoint of the histology of secretion but is
outside the scope of the present problem.

!Cf. Child, 'o-jb, p. 113.

METHOD OF CELL CIVISION IN MONIEZIA. 147

Although the vitellarium and the ovary arise from the same
anlage and at first are distinguishable only by their position they
soon acquire histological differences. The nuclei of the former
contain very lightly staining chromatin in the form of a reticulum
but there is a large nucleolus. It frequently occurs that there
are two or more nucleolar-like structures, karyosomes, in a
nucleus. The appearance of the nucleus itself is that of an
almost empty court surrounding the nucleolus or nucleoli. The
nuclei of the oogonia, on the other hand, do contain a reticulum,
rather lightly staining, it is true, but consistently present. While
the cytoplasm is more indefinite in amount in the vitellarium it
can be seen that the ratio of nucleus to cytoplasm, if the inclusions
are not taken into consideration, remains much more constant;
if the yolk mass be considered the second term of the ratio is
much increased. At no time after they become clearly dis-
tinguishable are the cells as large as the oogonia, and at matura-
tion the oocytes are more than twice their size.

As to cell multiplication I am convinced that after a com-
paratively early period division ceases. During this early period
clear cases of mitosis are to be seen, but, it is true, as Child says,
less frequently than in the ovary. On the other hand, the ar-
rangement of nuclei in pairs is, perhaps, more in evidence here
and indented nuclei are somewhat more numerous. The in-
dentations do not, however, give a series of autoconstrictions
from slight indentation up to complete division. If mitosis is
not clearly proved as the sufficient cause of cell multiplication
amitosis is certainly less so; for there is positive evidence that
some mitoses do occur, but for amitosis there is only negative
evidence.

The Female Genital Ducts. — From the same anlage which gives
rise to the ovary and the vitellarium, the female1 genital ducts
develop. Therefore, although they are strictly speaking somatic
structures, their cells will be considered here, with regard to the
method ot cell division, as derivatives of the female anlage.
Their early history, therefore, is the same as that of the ovary
and vitellarium. A thickening of nuclei appears median to the
longitudinal excretory tubes and gradually extends itself both

!The development of the male ducts will not be touched upon here.

148 A. RICHARDS.

outward and inward. The outward extension becomes the
vagina, seminal receptacle and fertilization canal; the inner not
only gives rise to the ovary and vitellarium but also to the re-
maining part of the oviduct, the vitello duct and the uterus.
These various structures are in no sense an invagination from
the outside for their development begins in the middle part of
their length and the connection with the outside, through the
cirrus pouch, is quite secondary to the formation of the lumen
of the tube. The lining is, of course, an epithelium; but it is
not an invagination from the exterior. That of the cirrus
pouch, however, is an invagination and might be looked upon
as an ectoderm. But in the other parts of the sex ducts the
epithelium is clearly not ectodermic in origin although histo-
logically it resembles the cuticle and is connected with it through
the cirrus pouch.

In their histological structure, the walls of the uterus, oviduct
and seminal receptacle are made up of a simple layer of flattened
cells probably with a basement membrane. The oviduct fur-
nishes exception to this statement at its opening into the ovary
where a layer of circular muscle fibers is demonstrable, and in
that part of its course which is known as the fertilization canal.
In this latter canal the structure more nearly resembles that of
the vagina. The vagina beginning at the lumen consists of
ciliary projections, a layer of nuclei embedded in a homogeneous
cuticle, longitudinal and circular muscle fibers and a cellular
layer.

These canals develop fairly quickly and have all (uterus ex-
cepted) finished their growth as far as cell division is concerned
long before the sex products call for their use. Upon the passage
of the eggs from the ovary they degenerate and their place is
taken by the embryo filled uterus.

Whether the ducts grow by more and more parenchymal nuclei
becoming involved in the proliferating area, as Child thinks,
rather than by the actual outgrowth of the original anlage by the
multiplication of its cells is not clear to me. It seems probable
that the extension occurs by both methods of growth. The ovi-
duct and the vitello duct are unquestionably formed by the
differentiation of part of the cells in the primary anlage for it is

METHOD OF CELL DIVISION IN MONIEZIA. 149

in the middle of the anlage that these ducts develop. In the
vagina it is not certain that the parenchyma cells whose position
it occupies were numerous enough to provide all the cells in its
walls without the aid of some outward growth from the region
already formed. There is no fundamental difference, however,
in the two methods for in either case the cells are of parenchymal
origin.

The conditions in the early primary anlage have already been
set forth. Mitoses are few, but they do occur. There is no
certain evidence for amitosis although many nuclei are arranged
in pairs. I have previously given my reasons for the view that
amitosis does not occur at this stage of development.

But in the stages beginning with those in which the anlagen
of the various organs can be distinguished the method of cell
division can be conjectured only. Mitoses are very few, in-
deed, and occur only in the borders of the proliferating region.
On the other hand, the nuclei are small and irregular in size and
are frequently closely crowded together in groups of two, three or
even more. My preparations do not show actual cases of ami-
tosis, however, although conditions would seem to be quite
favorable for its occurrence. Other than the rare mitotic figures
which occur there is absolutely no evidence as to the method
of nuclear multiplication (see Figs. 51 and 52).

The opinions with regard to the regions in which active cell
divisions are occurring expressed by Child ('07^) in the following
statements do not seem to be well founded according to my
observations. He says: "In the early stages there is a marked
difference in the size of the nuclei in the central and the peripheral
portions of the area in which proliferation is occurring. . . .
Evidently proliferation is much more active in the central than
in the peripheral regions of the proliferating area. In somewhat
later stages the rapidity of division apparently decreases and the
nuclei of the central regions gradually increase in size until they
are almost or quite as large as those about the periphery. . . .
From this stage on the differentiation of the walls of the ducts
gradually takes place: muscle fibers develop, a lumen appears,
and nuclear divisions become less and less frequent."

In this quotation (the omitted parts have to do only with

I5O A. RICHARDS.

details not bearing upon the point which I wish to consider)
the criterion of proliferation seems to be smallness of nuclei,
although in the figures illustrating these statements constricted
nuclei, amitoses, are present. As previously stated, these regions
in my sections do not show amitoses. But mitoses are to be
seen (rarely, it is true) in the borders of the proliferating region.
The presence of mitotic figures in this region, it seems to the
present writer, is a most significant fact — where but in the borders
of a region which is extending its area should division figures be
found? — and the small size of the nuclei at the center is of little
importance. These central cells are merely beginning differen-
tiation while the proliferation takes place at the periphery of the
sex duct anlage.

The method of cell multiplication in the female sex ducts of
Moniezia cannot at the present time be positively stated. I find
no certain evidence for amitosis and that for mitosis is, perhaps,
insufficient to account for the growth which has taken place.
Nevertheless, I have seen some cases of mitosis here.

Fertilization and Maturation. — The process of fertilization in
the eggs of Moniezia can be followed only with the greatest dif-
ficulty. Although one may examine carefully a great number of
proglottids he will rarely find a stage showing the entrance of
the sperm and its course prior to the maturation divisions. I
am, therefore, unable to amplify the account given by Child
('oje) of the early stages of fertilization.

From the ovary the passage through the oviduct to the uterus
is so rapid that eggs are seldom found in that part of the duct.
Linton ('08) has observed the process of fertilization in a live
trematode, and found it to be of very short duration (less than
40 seconds) . In Moniezia one often finds segments with embryos
in the uterus undergoing maturation and eggs in the oviduct
ready for fertilization, but the cases are rare in which there are
eggs in the fertilization duct. Doubtless here also the passage
is very rapid and probably, as Child thinks, periodical.

The eggs pass from the ovary into the uterus through a more
or less coiled oviduct. They meet the spermatozoa at the mouth
of that branch of the oviduct which comes from the seminal
receptacle and are there fertilized. Near this point the vitello

METHOD OF CELL DIVISION IN MONIEZIA. 151

duct after passing through the vitellarium and shell gland joins
the oviduct — the point of union being known as the ootyp — and
this latter duct Continues to the uterus.

Cases of polyspermy frequently occur; that is, some eggs show
more than one male pronucleus. I have not observed the en-
trance of several spermatozoa into the same egg, but this is not
surprising since the entrance of a single spermatozoon is rarely
seen. In many proglottids some of the segmenting eggs ulti-
mately degenerate ; possibly they include the cases of polyspermy.

In certain proglottids ova are found in the seminal receptacle,
sometimes in quite large numbers. As the seminal receptacle
is not ruptured in many of these cases it would seem that in
passing through the ootyp the eggs had crowded past the branch
of the oviduct leading to the uterus and forced their way through
the fertilization canal into the seminal receptacle. There they
degenerate. It is, of course, very probable that polyspermy
occurs here but so densely packed are the spermatozoa about
the egg that the process cannot be followed.

For figures of the fertilization stages as well as for additional
figures of the maturation divisions the reader is referred to
Child's paper ('070).

Child's account of the maturation divisions agrees in essentials
with my observations, and I shall, therefore, merely give a brief
resume of this stage of development.

The entrance of the spermatozoon furnishes the stimulus for
the formation of the vitelline membrane and for the beginning
of the maturation divisions. The first polar body is usually
formed by the time the egg has entered the first loops of the
uterus and the second polar body follows soon after. The first
division requires more time than the second if the relative fre-
quency in which the two stages occur may be regarded as a
criterion. Many more first divisions occur in my material than
second.

The characteristics of these divisions are large globular centro-
somes — ring-shaped in section — faint astral radiations,1 and a
very long spindle on which small chromosomes are to be seen

1 Child stated that he was unable to see asters but thought they probably occur.
I have succeeded in distinguishing faint radiations in numerous maturation stages.

152 A. RICHARDS.

distinctly. An equatorial plate is not usually formed as the
chromosomes pass to the poles irregularly. Figs. 53 and 54
illustrate these facts.

The reconstruction of the female pronucleus is shown in Fig.
55. These stages suggest amitotic division of the ovum — some
more than that figured; their true nature, however, is readily
ascertained.

I am unable to see that the maturation of Moniezia has any
bearing whatever on the "hypothesis of individuality" of chro-
mosomes, which Child says these mitoses do not appear to
strengthen, for it does not seem to me that evidence either pro
or con is presented here. Indeed, the whole question involved
is not so much whether the chromosomes maintain their indi-
viduality as it is whether the mechanism of division will give
an exact sorting of the male and female elements in the germ
cells or a mere separation of unequal parts of the nuclei. If
mitosis could be established as the universal method of cell
division that fact alone would by no means warrant the assump-
tion of chromosome continuity, as Child would seem to imply
that it would do. Among those who believe that mitosis will
be found general in germ cells are many who do not accept the
individuality hypothesis as it is now put forward. Furthermore,
proof of that hypothesis will demand observations upon a more
favorable object than Moniezia. I would emphasize the fact,
therefore, that Moniezia presents no evidence upon the indi-
viduality hypothesis.

Cleavage. — The steps of the cleavage process in Moniezia were
discussed in 1881 by Moniez from an embryological point of view,
and by Child in 1907 from the point of view of cell division.
The results of Moniez are very suggestive and deserve to be
repeated in the light of more recent discoveries on cleavage and
with later methods of technique. Since the cleavage of Moniezia
is deserving of this broader treatment I shall not here attempt
other than the most general statement of the process but shall
concern myself with the method by which segmentation is ac-
complished.

The type of cleavage, as will be seen from a glance at the fig-
ures, from 56 on, is that of a large macromere giving off small

METHOD OF CELL DIVISION IN MONIEZIA. 153

micaomeres at the animal pole.1 The later divisions bring about
the formation of a morula which gives off two layers by delamina-
tion and then develops into the hexicanth embryo. This cleavage
especially in its earlier course differs markedly from that in
Tcenia serrata (cf. Janicki's figures with mine).

As cleavage proceeds many embryos become elongated, a con-
dition which Child suggests is due to pressure of the uterine walls.
This, of course, does not represent the future axis of elongation.
That Child's suggestion is probably the correct one is shown by
Moniez's figures from entire embryos which are all spherical
and do not include any elongated stages.

During the early part of my study while I was working upon
material in which the results of fixation were good but by no
means perfect I thought I had found the syncytial condition of
which Child speaks : "In most cases a number of nuclear divisions
occur before cell boundaries become visible in the egg. ... As
cleavage proceeds the egg is gradually divided into blastomeres
containing yolk and blastomeres without yolk. In earlier stages
the yolk bearing blastomeres often contain two or more nuclei,
but in the later stages after cytoplasmic cleavage is more ad-
vanced they usually contain one relatively large nucleus. In
other words as these yolk bearing blastomeres are gradually
reduced in size by successive cleavages the cytoplasmic cleavages
keep pace more nearly with the nuclear divisions. In the yolk-
less portion of the egg, however, nuclear division continues far
in advance of cytoplasmic division as far as the cleavage has
been followed ; the consequence is that each blastomere contains
several or many nuclei of relatively small size."

Later, in studying material in which fixation had been more
nearly perfect, I saw at once that the supposed syncytial con-
dition was in reality an artifact. From the very first the micro-
meres can be distinguished as such, an observation corresponding
with the figures of Moniez made about thirty years ago. The
cytoplasmic membrane becomes completely constricted even be-
fore the telophase is finished as in Fig. 57. In properly fixed
material I have never seen an egg syncytium. The work of

Assuming that, as in all known cases except in the ctenophores according to
Korschelt and Heider, the polar bodies indicate the animal pole.

154 A. RICHARDS.

Moniez contains much which a reinvestigation with modern
methods of technique will substantiate.

As regards the method by which segmenting eggs divide, the
following sentences are illustrative of Child's position:

"But although the first cleavage is usually or always mitotic,
there can be no doubt that amitotic division appears very early
in the course of cleavage.

"Cases of mitosis are rarely seen after the first cleavage but
amitosis is of frequent occurrence.

"Rarely a case of mitosis is observed: in all the hundreds of
eggs in cleavage stages which have been examined, not more
than a dozen cases of mitosis have been seen in stages later than
the first cleavage. When mitosis occurs it apparently always
involves one of the larger nuclei. Mitotic divisions of the small
nuclei in stages like those shown in Plate VI. have never been
observed. The smaller nuclei are, without doubt, dividing more
rapidly than the larger, and we are probably justified in con-
cluding that amitosis occurs in those regions of the egg where
division is most rapid, while mitosis is found, when it occurs at
all, among the nuclei which are dividing more slowly."

With the first of these statements only can I agree. Figs. 56
to 63 show the successive steps in the cleavage process up to the
eight cell stage and every division is by mitosis. Fig. 64 show's
several cells from an embryo of thirty cells, among which are
two blastomeres in mitosis. Fig. 65 is drawn from a single
blastomere in an embryo of from fifty-five to sixty cells; it shows
an early phase of mitosis. These stages are not rare or difficult
to find in my material. I have examined segmenting eggs of
numerous other animals and do not hesitate to say that the
mitotic figures in Moniezia are as frequent as in other lots of
eggs selected at random.1 Indeed, in five sections cut four micra
in thickness, in which the uterus did not contain a relatively
large number of eggs over twenty cases of cleavage mitoses
which were deemed worthy of sketching were found. In another
case, in 125 embryos (all that were present in the given region)

JOf course in those cases where artificial fertilization is practical the eggs are
selected in the desired stage; such cases may not fairly be compared statistically
with the one under discussion.

METHOD OF CELL DIVISION IN MONIEZIA. 155

there were 28 cleavage mitoses; in another of 211 embryos were
53 mitoses; in still another, 124 embryos showed 40 mitoses.
The evidence that these cleavages occur by mitosis is of the most
positive and convincing character. Cleavage mitoses are more
abundant in my material than the second maturation division
which is, of course, mitotic.

Furthermore, I have seen only one condition which could
possibly be interpreted as amitosis; it is shown in Fig. 59. This
figure represents the stage following the telophase of the division
cutting off the third micromere. The same condition obtains
following other telophases, however. The nucleus after the
telophase at first takes on an irregular appearance but later
becomes rounded out into the typical shape. This stage is
exactly comparable to the reconstruction stage of the female
pronuculeus after maturation. It is to be interpreted as the
final phase of a mitosis rather than the precursor of an amitotic
division.

In the later cleavages, although I have not as yet followed out
all of the cell lineage, I have never found any reason to doubt
that mitosis is the method of cell division. Mitosis certainly
occurs frequently and I have seen no evidence of amitosis. The
two figures given are of typical cases; further illustrations would
be mere reduplication of the evidence.

In the smaller blastomeres I have never seen any sign of direct
division; on the other hand, Figs. 60 and 62 show cases of mitosis
in these cells.

To the present writer the facts in the cleavage of Moniezia
seem to admit of no other interpretation than that mitosis is
the method by which segmentation is accomplished.

SUMMARY OF THE EVIDENCE.

This summary will include the evidence gained from the study
of the female sex products from the time of their appearance as
the primary anlage of the ovary to the completion of segmenta-
tion. It will not include the vitellarium of sex ducts since the
evidence given by them is not thought sufficient to be conclusive
in favor of either view.

It has been stated that the cells composing the primary anlage

156 A. RICHARDS.

might arise de novo, by migration, by amitosis and by mitosis.
The first of these possibilities was dismissed as highly improbable;
the burden of proof is upon him who upholds this view. The
second method of origin was admitted to be possible, but it was
thought to play only an unimportant role, if any at all, in the
development; and even if the assumption of migration as a
factor — positive proof would scarcely be procurable — were found
to be necessary it would merely throw the question of division
back into the earlier stages. It was held, therefore, that the
cells under discussion must arise in situ by direct or indirect
division. It is the purpose of this summary to present the evi-
dence for the two views as concisely as may be.

It has been shown, furthermore, that at two periods of de-
velopment only may active growth by cell division be looked for:
namely, during the period of pre-oogonial and oogonial divisions
and during the cleavage of the embryo; the rest of the growth
involved is that of increase in size and in differentiation. The
discussion, therefore, limits itself to the question: is amitosis or
mitosis the method by which the pre-oogonial and oogonial and
the cleavage divisions occur?

The evidence, as I have found it, favoring the view that ami-
tosis occurs during the periods of development mentioned is as
follows :

First, there are in the development of the oogonia fewer cases
of mitosis to be found than one would a priori expect, and it
might be claimed that the number present is not sufficient to
account for the cell multiplication which must have occurred.

Second, the arrangement of the nuclei in pairs suggests in
the light of the fact just stated that the cells may have arisen
by direct division. Two nuclei often occur close together but
separated more or less from their neighbors.

Third, constricted and indented nuclei do occur in the early
stages of development of the oogonia; but they occur in my
material only in cases of imperfect fixation, so they lend very
little support to the amitosis view.

Fourth, in the cleavage stages but one condition (unless ac-
cessory sperm nuclei might be confused with direct cell division)
has been seen in properly fixed material which might be inter-

METHOD OF CELL DIVISION IN MONIEZIA. 157

preted as favoring the amitosis view: the condition represented
in Fig. 59.

The evidence opposed to amitosis and favoring the view that
mitosis is the method by which divisions occur is as follows:

First, mitotic figures do occur throughout all stages of the
development of the tissues considered, and they occur in the
peripheral (or growing) regions of the organs. Analogy may be
drawn from cases like Planorbis and Ascaris in which the oogonial
divisions are very difficult to find, and yet are known to occur
regularly.

Second, stages of nuclear constriction are not found in suf-
ficient number to account for the arrangement of the nuclei in
pairs if amitosis be the method by which they arise. It is logical
to expect that steps in the process would be found if the arrange-
ment in pairs is the end result of amitotic division. As has
been repeatedly stated, it is not possible to establish a complete
series of stages in the auto-constriction of nucleus and cell body.
If periodicity is a factor in cell division (see below) the arrange-
ment may be looked upon as favoring the mitosis view. Finally,
it is most probable, although impossible to prove, for the pre-
oogonia and oogonia are not favorable for such study, that there
may be some movements of the halves toward each other anal-
ogous to the movements of telekinesis in other tissues, for in all
cases carefully studied with regard to this point, including de-
veloping eggs, epithelial tissues, etc., such movements have been
found. I have myself found evidences of them (not shown in the
present series of figures) in the cleavage of Moniezia.

Third, the occasional cases of constriction and indentation
which occur have not been shown to be normal steps in division.

Fourth, the condition represented in Fig. 59 is, as previously
stated, the final stage of mitotic division rather than the be-
ginning of amitosis.

Fifth, the evidence favors periodicity as a factor in mitotic
divisions. (See below, under discussion.)

Sixth, that the maturation and first cleavage alone should
occur by mitosis as described by Child is difficult to reconcile
with amitotic divisions elsewhere throughout the series.

Seventh, the evidence that the cleavage divisions occur by

158 A. RICHARDS.

mitosis is of the most positive character. It cannot be too
strongly emphasized.

Eighth, I have after diligent search upon carefully prepared
material been unable to establish a series of stages in the auto-
constriction and subsequent division of the nucleus and cell body
by amitosis.

Considering the evidence as set forth, it seems to the writer
that one is forced to the conclusion that mitosis is the method
by which pre-oogonial, oogonial and cleavage divisions are accom-
plished. These are the divisions to which the chief interest

attaches.

DISCUSSION.

Since the account of Child appeared describing amitosis in
Moniezia and other forms there have been a number of a priori
suggestions offered by various workers as possible explanations
of the conditions which he describes. It will be of interest to
examine some of these suggestions in the light of the facts which
I have observed in Moniezia.

Gary on the basis of his observations on Diplodiscus suggests
that the telophase of an intranuclear mitosis would give all the
appearances of true amitoses. I have carefully searched for
evidences of intranuclear mitoses in Moniezia but have found
none. I have seen cases which at first resembled a faint intra-
nuclear spindle, but upon more careful study it became clear
that they were merely chance arrangements of linin threads and
chromatin granules and not mitosis. Furthermore, as already
shown, distinct cases of mitosis of the regular type are not difficult
to find. It is, of course, true that the condition which Cary
describes would resemble amitosis, but I have not found it
present in Moniezia.

Boveri calls attention to the fact that figures similar to those
found in Amblystoma (Child, o~a) are also found in Triton. In
this latter form Rubaschin found that after mitosis the nuclear
material does not fuse to form a single vesicle in the blastomeres
but two are always formed. They never separate but go through
mitosis together. In Moniezia there is no evidence of this kind
of phenomenon.

As there is no " Nebendotter " in the eggs of Moniezia my

METHOD OF CELL DIVISION IN MONIEZIA. 159

former suggestion that that structure might give a misleading
appearance of amitosis is not borne out.

As polyspermy occurs here accessory sperm nuclei might be
wrongly interpreted as amitosis. The condition, however, is not
difficult to distinguish.

Cell migration has already been discussed in detail as a method
of origin for the pre-oogonia. As stated before there is not much
evidence that it occurs. It certainly is not the exclusive factor
and probably not even an important one.

There yet remains my suggestion that mitosis may be of short
duration and occur in waves or at more or less definite periods
depending upon some unknown physiological factor — an inter-
petation which Child refuses to accept. He says ('10, p. 113):
" Not only the cells of the neck region but the cells of the scolex,
which does not take part in the growth of other regions are
undergoing mitosis in this specimen.1 Moreover, I am as yet
unable to convince myself that the many cases of apparent
amitosis which I have observed in the neck region of other
individuals, are errors of observation, or something else than
nuclear division." Nevertheless, is not the position that "the
relative frequency of mitosis and amitosis in certain species and
even in single individuals may vary greatly according to con-
ditions" in its practical application an admission of my sugges-
tion?

This suggestion is based first upon certain a priori consider-
ations. Beckwith described cleavage mitoses in Pennaria and
Clava occurring from 4 to 6 A.M. only; in certain insects and
many plants mitosis occurs at night only; in onion root-tips
there are two periods of mitotic activity daily, at I P.M. and at
II P.M. and if the specimens are collected at any other times
than these there will be found a paucity of division figures. I
have seen grass root-tips in which not a mitotic figure could be
seen although young cells of the proper age for multiplication
were abundant. Amitosis has never been suggested for such
cases as this. Secondly, the suggestion has support in certain

*A young specimen, found subsequent to his earlier accounts, consisting of
scolex, neck region and a few of the youngest proglottids in which almost every
nucleus is undergoing mitosis.

I6O A. RICHARDS.

observational evidence not the least of which is Child's young
worm in which almost every nucleus is in mitotic division. The
fact that I am unable to convince myself that amitosis occurs
and the fact that mitosis does occur in some specimens in
greater abundance than in others (this applies to Tcenia as well
as to Monlezia) and even some organ systems show more figures
than others (I have one specimen in which the majority of the
testes cells are in mitosis) point strongly to periodicity as a
factor in growth. The arrangement of nuclei in pairs may be
regarded as a point in favor of this view. That the scolex of
this worm should be in division is of no significance for even if
it takes no part with the neck region in the formation of new
proglottids, and this is not demonstrated, its own growth must
be accounted for. The scolex of a young worm is smaller than
that of an older one.

Furthermore, if it can be shown that the eggs pass through
any one stage periodically it becomes all the more probable
that they also pass through others periodically. The evidence
is strong that the eggs undergo maturation in this manner, and
that fact is most easily explained by periodicity of oogonial
divisions.

Child holds that mitoses decrease in frequency with increasing
age. Let me suggest that perhaps that factor, whatever it may
be, that operates to free the intestine of adult sheep from tape-
worms may also operate to check mitotic division during the
latter part of the period in which the worms can yet live there.

The evidence, I believe for the various reasons given, strongly
favors periodic mitoses as an important factor in the growth
of the cestode, Moniezia.

COMPARISON OF Tcenia AND Moniezia.

Throughout the course of this discussion it has been made
clear that Moniezia differs from Taenia serrata in numerous
details, both anatomical and cytological. Yet it is also clear
that the differences, while often striking, are differences not of
physiological activites and principles but of structural features
and structural differentiation.

Most striking of all the differences is that of gross structure

METHOD OF CELL DIVISION IN MONIEZIA. l6l

as seen in the female reproductive apparatus. Such wide vari-
ations from a common type are not usually found within a single
family of animals. In both the uterus is single1 and median but
the other organs are arranged in pairs, there being two ovaries
and two sets of auxiliary glands and ducts as well as two genital
pores in each proglottid in Moniezia while in Tcenia the ovaries
are double but they discharge into a common duct and the other
parts of the reproductive organs are single.

But the most interesting comparison of the two forms for
present purposes is that regarding cytological characteristics.
In both the anlagen of the organs arises as a thickening of the
parenchyma cells and the period of cell proliferation is strictly
comparable on the two. In Tcenia, however, it is of less duration
as seen by the fact that it extends over fewer proglottids. The
method of yolk formation is dissimilar in appearance only, for
if the yolk granules of Moniezia were to accumulate on one side
of the nucleus and fuse the condition would be exactly that of
Tcenia. I have not discovered that there is any difference in
the principles of division operative in the two forms.

Between the types of cleavage and early development of the
embryos of the two forms there are some noteworthy differences
but as they are embryological in character they may not properly
be included here. It will be seen from this comparison that
there are sufficient differences between Tcenia and Moniezia,
although they are members of one family, to justify the claim
that observations on the former cannot be used a priori as
evidence against the latter; but that actual study does not
disclose any real differences in the principles of growth and
development in the two cases.

CONCLUSIONS.

In conclusion, the evidence presented in this investigation goes
to show: (i) that it cannot now be positively affirmed whether
mitosis or amitosis is the method which obtains in the develop-
ment of the vitellarium and female genital ducts of Moniezia ;
(2) that in the early stages of sex cell development mitosis un-
questionably occurs (probably periodically), while amitosis is not

1 Stiles and Hassal state that the uterus of Moniezia is ontogenetically double.

1 62 A. RICHARDS.

evident in my preparations; and finally, (3) that there cannot
be the slightest doubt that the cleavage of the ovum takes place
by mitosis.

NOTE.

While this paper was in press an article by Gough ["A Monograph of the Tape-
worm of the Subfamily Avitellininae, being a revision of the Genus Stilesia, and
an account of the Histology of A vitelli na centri punctala (Riv.)"] appeared in the
Quarterly Journal of Microscopical Science for February, 1911, wherein he describes
the development of the eggs of Avitellina. He, too, finds Zenker's fluid best for
cestodes. As in Moniezia, yolk cells do not become attached to the eggs to form
compound structures. The large mass which he calls chromatin (Figs. 57-63)
apparently correspond to my nucleoli (my Fig. 30) ; there can be no doubt that in
Moniezia these are nucleoli and that no other nucleolar structures occur, as shown
by differential staining. The extrusion into the cytoplasm of chromatin which
breaks up to form the yolk nucleus (see p. 373) has not been observed in Moniezia.
I would suggest that his Fig. 63 perhaps corresponds to my Fig. 26 which is a
synapsis stage not merely the first step in a mitosis. Gough speaks of the
maturation mitoses as occurring very rapidly. As Stilesia, and, therefore,
Avitellina is quite closely related to Moniezia the comparison of the two is of
interest, particularly with regard to the fact that compound eggs do not occur
in either

LITERATURE REFERENCES.
Balss, H. H.

'08 Ueber die Entwicklung der Geschlectsgange bei Cestoden nebst Bemerkung

zur Ectodermfrage. Zeitschr. f. wiss. Zool., Bd. 91.
Boveri, Th.

'07 Zellenstudien, 6, p. 235. G. Fischer, Jena.
Bugge, G.

'02 Zur Kenntnis des Excretionsgefass-Systems der Cestoden und Trematoden.

Zool. Jahrb., Vol. 16, Anat., s. 177.
Gary, L. R.

'09 The Life History of Diplodiscus temporatus Stafford. Zool. Jahrb. (Anat.

und Ont.), Bd. 28.
Child, C. M.

'04 Amitosis in Moniezia. Anat. Anz., Bd. 25.

'06 The Development of Germ Cells from Differentiated Somatic Cells in

Moniezia. Anat. Anz., Bd. 29.
'070 Amitosis as a Factor in Normal and Regulatory Growth. Anat. Anz.,

Bd. 30.
'07 Studies on the Relation between Amitosis and Mitosis. In Biol. Bull.,

Vols. 12 and 13.
'076 I. Development of the Ovaries and Oogenesis in Moniezia.

II. Development of the Testes and Spermatogenesis in Moniezia.

III. Maturation, Fertilization, and Cleavage in Moniezia.

IV. Nuclear Divisions and Somatic Structures in the Proglottids of
Moniezia.

'10 The Occurrence of Amitosis in Moniezia. Biol. Bull., XVIII.

METHOD OF CELL DIVISION IN MONIEZIA. 163

Conklin, E. G.

'02 Karyokinesis and Cytokinesis. Jour. Acad. Nat. Sci. of Phila., XII., Pt. I.
Dahlgren and Kepner.

'08 Principles of Animal Histology. New York, The Macmillan Co.
Janicki, C. v.

'03 Beziehungen zwischen Chromatin und Nucleolen wahrend der Furchung

des Eies von Gyrodactylus elegans von Nordm. Zool. Anz., Bd. 26.
'07 Ueber die Embryonalenentwicklung von Taenia serrata Goeze. Zeitschr.

f. wiss. Zool., Bd. LXXXVII.
Linton, Edwin.

'08 The Process of Egg Making in a Trematode. Biol. Bull., XV.
Moniez, R.

'81 Memoires sur les Cestodes. Travaux de 1'Institut Zoologique de Lille,

T. III.
Montgomery, T. H.

'99 Comparative Cytological Studies with especial Regard to the Morphology of

the Nucleolus. Jour. Morph., XV., 2.
'01 A Study of the Chromosomes of the Germ Cells of Metazoa. Trans. Amer.

Philos. Soc., Vol. XX.
Richards, A.

'09 On the Method of Cell Division in Taenia. Biol. Bull., Vol. 17.
Rubaschin, W.

'05 Ueber doppelte und polymorphe Kerne in Triton Blastomeren. Arch, f.,

Mikr. Anat., Bd. LXVI.
Spatlich, Walter

'09 Untersuchungen iiber Tetrabothrien. Zool. Jahrb., Bd. 28.
Stiles, C. W., and Hassal, Albert.

'93 A Revision of the Adult Cestodes of Cattle, Sheep, and Allied Animals.

Bull. 4, Bureau of Animal Industry, U. S. Dept. of Agriculture.
Ward, H. B.

'94 The Parasitic Worms of Man and the Domestic Animals. Nebr. State

Board of Agriculture Report for 1894.
Wilson, E. B.

'04 The Cell. New York, The Macmillan Co.
Young, R. T.

'08 The Histogenesis of Cysticercus pisiformis. Zool. Jahrb. (Anat. und Ont.),
Bd. 26.

164 A. RICHARDS.

EXPLANATION OF PLATE I.

Photomicrographs were made with a Leitz no. 3 objective and no. 2 ocular except
IV. in which a no. i ocular was used. All figures were drawn with a Zeiss compen-
sating ocular no. 12 and a 1.5 mm. objective at table level with the aid of a camera
lucida, giving a magnification of about 3,500 diameters.

PHOTO. I. End of a section through a young proglottid, showing the primary
anlage of the female reproductive organs.

PHOTO. II. Primary anlage differentiated into the anlagen of the ovary, vitel-
larium and oviduct.

PHOTO. III. Later stage showing finger-like processes of the ovary. Dorsal
to the middle of the ovary is the tip of the vitellarium.

PHOTO. IV. Stage of maximum development, exc, excretory canal; od,
oviduct; v.t, vitellarium; vt.d, vitello duct; ov, ovary; ut, uterus; /, testes; f.d, fertil-
ization duct; s.r, seminal receptacle.

BIOLOGICAL BULLETIN, VOL. XX

exc.

6d

A. RICHARDS

1 66 A. RICHARDS.

EXPLANATION OF PLATE II.

FIGS, i and 2. Parenchyma cells for comparison with the cells of the primary
anlage; from young proglottids. Fig. 2 illustrates the arrangement of the cells in
pairs common in all undifferentiated tissue of the cestode. Cytoplasmic boundaries
are difficult to make out in such cells as these in all but the best fixed specimens.

FIGS. 3-5. From primary anlage in same section as Fig. 2. Fig. 3 is from the
inner growing end of the anlage, to be the ovary; Fig. 4 is from the middle region,
to develop into the vitellarium, etc.; Fig. 5 is from the outer growing tip, to form the
vagina. Hermann, iron haematoxylin.

FIGS. 6 and 7. From the primary anlage of the same section from which Fig. I
was taken; Fig. 6 from the middle, Fig. 7 from the outer end. Zenker, Ehrlich-
Biondi.

FIG. 8. A case of mitosis in the primary anlage. King, iron haematoxylin.

FIG. 9. Mitosis in a pre-oogonium. Zenker, safranin and gentian violet.

FIG. 10. Polar view of a mitosis from the same anlage as Fig. 9.

FIG. ii. Cell from near the growing tip of the ovary anlage; a pre-oogonium.
Zenker, Ehrlich-Biondi.

FIG. 12. Group of pre-oogonia just before the formation of the follicular mem-
brane. Cytoplasmic boundaries are not well preserved in all cases. The over-
lapping of such nuclei as are shown here gives a misleading suggestion of amitosis.
King, iron haematoxylin.

FIG. 13. A stage in the mitosis of a late pre-oogonium before the condensation
of the chromosomes. Hermann, iron haematoxylin.

FIG. 14. The most distinct case of mitosis found throughout the oogenesis;
spindle fiber bundles are very strong. A late pre-oogonium. King, Ehrlich-Biondi.

FIG. 15. Pre-oogonium, slightly later than Fig. 12. King, iron haematoxylin.

FIG. 16. A young oogonium in mitosis; polar view. Zenker, iron haematoxylin.

BIOLOGICAL BULLETIN, VOL. XX

A. RICHARDS

168 A. RICHARDS.

EXPLANATION OF PLATE III.

FIG. 17. Young oogonia, showing three sets of nuclei arranged in pairs. Only
one nucleolus is present in each, the large black granular masses being chromatin.
Cytoplasmic boundaries not well preserved. King, iron haematoxylin.

FIG. 1 8. Oogonia; the upper nucleus stained very heavily. King, iron haema-
toxylin.

FIG. 19. An oogonium beginning growth. King, iron haematoxylin and Bis-
marck Brown.

FIG. 20. Mitosis in an oogonium just before the growth period begins. King,
iron haematoxylin.

FIG. 21. An oogonium beginning growth; the chromatin is beginning to con-
dense in preparation for the synapsis stage. King, iron hsematoxylin.

FIG. 22. Oogonium from the same ovary, slightly later in development.

FIGS. 23-28. Progressive stages of synapsis, 23-26 from the same slide. King,
iron haematoxylin. 27, King, iron haematoxylin. 28, King, Ehrlich-Biondi.
In Fig. 26 are to be seen the beginnings of the yolk nuclei. They do not regularly
appear at this stage.

BIOLOGICAL BULLETIN, VOL.. XX

PLATE III

A RICHARDS

I7O A. RICHARDS.

EXPLANATION OF PLATE IV.

In the latter part of the growth period, following synapsis, yolk is produced
(Figs. 31-35) and the nucleoli exhibit certain characteristic features. Figs. 28,
29 and 30, selected from the oocytes, show a nucleolus proper, paranucleolus, en-
donucleolus and certain peculiar dark (chromatic?) bodies. These bodies give
the characteristic staining reaction of chromatin, very dark with gentian violet.
The nucleolus with the triple stain is violet, paranucleolus pink, endonucleolus
dark. In the paranucleolus a reticulum supporting the endonucleolus can often
be seen (29^).

FIG. 29. King, Ehrlich-Biondi.

FIG. 30. King, triple.

FIG. 31. Oocyte showing yolk nucleus. The nucleus is not well fixed in this
material. Chrom-oxalic, iron haematoxylin.

FIG. 32. Sections above the nucleus, passing through the yolk nuclei. Chrom-
oxalic, iron haematoxylin.

BIOLOGICAL BULLETIN, VOL. XX

PLATE /V

A. RICHARDS

I?2 A. RICHARDS.

EXPLANATION OF PLATE V.

FIGS. 33-35. First oocytes, the ovarian eggs. Yolk globules shown in black
in Fig. 35 and indicated by dotted circles in the other figures; chromatin in the
characteristic arrangement. Two peculiar black bodies in Figs. 34 and 35 may
perhaps be centrosomes or spheres. Fig. 33, King, iron haematoxylin; 34, Zenker,
Ehrlich-Biondi; 35, King, iron haematoxylin.

FIGS. 36-50. Development of the vitellarium. Character of the early devel-
opment is shown in Figs. 4 and 6 from the primary anlage.

FIG. 36. Corresponds to Fig. n in time of development. Zenker, Ehrlich-
Biondi.

FIG. 37. Same stage as Fig. 12. King, iron haematoxylin.

FIG. 38. Growing vitellarium cell from some section as Fig. 15. King, iron
haematoxylin.

FIG. 39. From the medullary portion of the anlage, the shell gland from the
same section as Fig. 37. Cytoplasm drawn out into fibers.

FIG. 40. Shell gland region; same slide as 41 and 42.

FIGS. 41 and 42. Vitellarium cells; both cells simulate mitosis due to the
heavy linin strands across their centers. Fig. 42 was at first taken for a case of
"endogenous division." King, iron hasmatoxylin.

FIG. 43. From growing vitellarium, about the same stage as Fig. 20, a charac-
teristic mitosis. King, iron haematoxylin.

FIG. 44. Just before the formation of the yolk body. The dense cytoplasmic
mass is probably a yolk nucleus. King, iron haematoxylin and Bismarck Brown.

BIOLOGICAL BULtETiN, VOL. XX

PLATE V

A. RICHARDS

174 A- RICHARDS.

EXPLANATION OF PLATE VI.

FIGS. 45-49. Stages in the development of the yolk body in the vitellarium
cells. King, Conklin's stain.

FIG. 50. The appearance of a much decolorized yolk mass.^ b is the same yolk
mass shown in a at a lower level; probably the central bodies here are remnants
of the yolk nucleus. King, iron hcematoxylin.

FIG. 51. Cells early in the development of the female genital "ducts. Cyto-
plasmic boundaries are not well fixed in this preparation. The nuclei show no evi-
dences of division either direct or indirect. Arrow points in the direction of the
axis of the duct. King, iron haematoxylin.

FIG. 52. A cross section from a somewhat later stage in oviduct formation. A
case of mitosis is shown here. Formol-sublimate, iron hsematoxylin.

FIG. 53. First maturation division of an oocyte. King, iron hsematoxlyin.

FIG. 54. Second maturation division. Zenker, iron haematoxylin.

BIOLOGICAL BULLETIN. VOl. XX

PLATE VI

A RICHARDS

1 76 A. RICHARDS.

EXPLANATION OF PLATE VII.

FIG. 55. Reconstruction of the female pronucleus; male pronucleus near. The
other polar body in the next section. Zenker, Conklin's stain.

In the remaining illustrations the blastomeres not shown in the figure were
to be found in the next section.

FIG. 56. First cleavage division. Zenker, safranin and gentian violet.

FIG. 57. Division cutting off the second micromere. Three-cell stage. Zenker,
safranin and gentian violet.

FIG. 58. Formation of the third micromere. Dotted line indicates the position
of the second micromere. Zenker, safranin and gentian violet.

FIG. 59. Four-cell stage, showing reconstruction of macromere nucleus after
the last mitosis. Zenker, safranin and gentian violet.

FIG. 60. Formation of the fourth micromere. Zenker, safranin and gentian
violet.

BIOLOGICAL BUILETIN, VOl . XX

PLATE VII

A. RICHARDS

178 A. RICHARDS.

EXPLANATION OF PLATE VIII.

FIG. 61. Six cell stage. Next section shows seven chromosomes belonging to
this spindle.

FIG. 62. Seven cell stage. Formol-sublimate, safranin and gentian violet.

FIG. 63. Division in the eight cell stage, only one end of the spindle shown in
this section. Zenker, safranin and gentian violet.

FIG. 64. From an embryo of about 30 cells, showing two mitoses. Zenker,
safranin and gentian violet.

FIG. 65. A blastomere beginning mitosis from an embryo of 55 to 60 cells. Zen-
ker, safranin and gentian violet.

BIOLOGICAL BULLETIN, VOL. XX

PLATE VIII

A. RICHARDS

A NOTE ON THE METAMORPHOSIS OF THE
MUSSEL LAMPSILIS L^VISSIMUS.1

ROBERT E. COKER AND THADDEUS SURBER.

While making a collection of the glochidia of various species
of mussels used in this laboratory certain observations have been
made which prompt the publication of a preliminary note on the
peculiar type of glochidium seen in Lamps His Icevissimus and
other species, and the metamorphosis during the period of
parasitism.

i. TYPE OF GLOCHIDIUM AND POSSIBLE SIGNIFICANCE.

The glochidium of Lampsilis alatus has been observed by
Lefevre and Curtis, 2 who describe it as an " axe-head " glochidium.
"This possesses hooks which are not homologous with those of
the anodonta type and is to be regarded as more nearly related
to the hookless forms, an interpretation which is borne out by
the fact that the 'axe-head' can be readily imagined as a
modification of the glochidial outline seen in some species of
Lampsilis, which, like subrostratus, show an approach to a
rectangular outline."

This glochidium (Figs. 3 and 30) is certainly of a very special-
ized form. In our laboratory we have observed an almost
exactly similar glochidium in L. capax (Figs. 4 and 4a), while in
L. Icevissimus (Figs, i and ia) the same shape is 'observed but
the hooks are wanting. It is remarkable that a larva of so
specialized a form should be found in two species such as alatus
and capax, the adult forms of which are so extremely opposed
in general shape; alatus is one of the most compressed forms in
the genus, while capax is the most inflated (Figs. 4^ and 4c).

L. Icevissimus (Figs, ib and ic) and L. gracilis are forms sug-
gestive of alatus in the compressed character of the shell. They
are somewhat more compressed than alatus and are much thinner

1 Published by permission of the U. S. Commissioner of Fish and Fisheries.
Read by title before the American Society of Zoologists, December, 1910.

2 " Reproduction and Parasitism in the Unionidae," Jour. Exp. Zoo/., Vol. IX..
No. i, p. 94-

179

ISO ROBERT E. COKER AND THADDEUS SURBER.

shelled. These two species seem almost to intergrade (inform of
shell) so that one sometimes hesitates in the identification of a
specimen of intermediate form. On the other hand, the glo-
chidia of these two species show a striking contrast. The glo-
chidium of gracilis (Figs. 2 and 20} is oval-pear-shaped and very
similar to that of ligamentinus or ventricosus, or to what we
generally regard as a typical form; while that of l&vissimus (Figs.
I and la) is of the "axe-head" type. Comparison of Figs. la
and 2<3 may suggest, however, that the two species are not so
dissimilar as at first appears, and Icevissimus may be thought of
as intermediate betwreen alatus and gracilis.

The remarkable fact, yet, is that capax should have this type
of glochidium. Capax has always been grouped with ventricosus
and ovatus, but ventricosus, at least, has a glochidium of the usual
form. Whether the greater significance be attached to resem-
blances in larval form or in adult form, it is a suggestive revelation
of the adaptability of fresh-water mussels and the inadequacy of
superficial diagnosis. Undoubtedly in the classification of fresh-
water mussels an entirely undue stress has been laid upon
characters of a superficial nature. The ultimate system will be
based on a much more thorough analysis of the actual anatomy
of mussels in young and adult stages than has ever been at-
tempted in systematic work on this group. We have not yet a
sufficient knowledge of the comparative anatomy of glochidia
to judge of their relative value in the study of relationships,
but certain considerations derived from the present case are
sufficiently striking to be recorded for their suggestive value.

Ventricosus is certainly closely related to ligamentinus and
luteohis, as judged by the external form. The glochidia, too,
are much like those of ligamentinus. The adult form is special-
ized most noticeably in being very inflated. Gracilis is special-
ized in a different way, being exceedingly compressed laterally
and having compressed teeth; Icevissimus (Figs. i& and ic) is a
more extreme form of the gracilis type, the teeth are more blade-
like, while the shell is equally compressed from side to side and
has well-developed wings. The glochidium, while suggestive of
gracilis (which is of the "typical" form), is of the "axe-head"
form. The adult alatus is of somewhat the same compressed

METAMORPHOSIS OF LAMPSILIS L/EVISSIMUS. l8l

form, broadly winged, and with "axe-head" glochidia still further
modified by the development of hooks. Capax has the same
type of glochidium, but what of the adult form (Figs. 4& and 4^)?

In external shape capax and ventricosus are so closely alike
that it has sometimes been doubted if the two species were
properly separable. However, the species capax is now uni-
versally accepted, for the reason that capax is clearly distinguish-
able from ventricosus in: (i) the polished almost rayless character
of its epidermis, (2) the compressed form of the teeth, and (3)
in the relative thinness of its shell, wrhich is inclined to pinkness
in color of nacre. Now these, it is significant to remark, are all
characters of Icevissimus. In fact, (i) above is the most certain
means of distinguishing, l&vissimus from gracilis.

If, then, we were to draw an inference from the glochidia as
to a relationship between Icevissimus, alatus and capax, there
would be strong corroborative evidence in the adult characters,
in spite of the fact that Icevissimus and capax are the two ex-
tremes in degree of inflation. The similar degree of inflation of
capax and ventricosus would offer only a striking instance of
convergence in one character.

2. CHANGE OF FORM DURING PARASITISM.

The change undergone by Lampsilis Icevissimus during para-
sitism is most striking. As a general rule, it has been held that
there is practically no growth in size during the period of para-
sitism— simply a metamorphosis from glochidium to young mus-
sel in rudimentary form. Our observations show a notable
exception to this general rule.

While examining the gills of a specimen of the sheep's-head,
Aplodinotus grunniens, several specimens of encysted glochidia
were found in an advanced stage of development. The infection
occurred in nature, so that the age of the mussels cannot be
stated.

Figs. 5, 6 and 7 show not only the striking change in form, but
the enormous increase in size as well. Fig. 5 shows a specimen
in side view; the "axe-head" glochidium shell is evident, but
the mussel is now nearly circular in outline and several times
larger than the glochidium. The specimen shows considerable

1 82 ROBERT E. COKER AND THADDEUS SURBER.

inflation. Fig. 7 shows the dorsal aspect of a more advanced
specimen; the degree of inflation is apparent; the glochidial
shell is insignificant in size as compared with the mussel; it is
like a narrow saddle which extends only about half way over the
side of the mussel shell. The position of the glochidial shell
valves in Figs. 5 and 6 will be better understood from a compari-
son with this figure. The specimen of Fig. 6 is viewed from a
ventro-lateral aspect. The following measurements (in decimal
parts of millimeter) were made from the specimen represented
by Fig. 6 :

Length of glochidium 0.095 mm.

Width (height) 0.15 mm.

Length of mussel 0.320 mm.

Width (height) of mussel 0.215 mm.

The mussel, as compared with glochidium, is nearly three and
one half times as long and nearly one and one half times as wide.
It is evident that there must have occurred a material increase
in the size of the cellular cyst of the tissue of the host which
encloses the young mussel.

The shape of the mussel in the stage shown by Fig. 7 led to
the suspicion that the glochidium was of capax instead of Icevis-
simus, but the form of the glochidium is that of l&vissimus, and
there is no sign of teeth on the shell.

It is not known that such a growth in size and alteration of
form occurs in any other species during the period of parasitism.
What further change in form or growth in size would occur
before the liberation of the mussel cannot be determined without

additional material, not now available.
BIOLOGICAL STATION,
FAIRPORT, IOWA.

December 23, 1910.

BIOLOGICAL BULLETIN, VOL. XX

\a

It.

4h

I c.

ROBERT E. COKER AND THADDEUS SURBER

NOTES ON SOME ARACHNIDS FROM OHIO
VALLEY CAVES.1

NORMAN E. McINDOO.

In September, 1909, immediately after the expiration of my
year at the Indiana University's Cave Farm,2 I spent two weeks
collecting arachnids in the following caves: Marengo, Spring,
Wyandotte, Little Wyandotte, Sibert's Well, Saltpeter and
Mammoth Cave. In all these caves 268 individuals were caught.
The expenses in part were defrayed by a grant from the American
Association for the Advancement of Science.

Linyphia Weyeri Emerton3 was taken only from Marengo Cave.
This species in this cave is as abundant as Troglohyphantes (Wil-
libaldia) cavernicola Keys, is in the Shawnee Cave. They are
most abundant at a place 200 feet from the old entrance, or 500
from the new one. None were seen farther that 1,200 feet from
the old entrance. Insects from this locality to the end of the
cave are very rare, while near the entrance they are very abun-
dant. The habitat, habits, webs and cocoons of this spider are
similar to those of Troglohyphantes. Some were collected on
their snares in the angles formed by the floor and wall, some
along clay banks, some in little pits in the sand floor, and others
under rotten boards and debris with thysanurans, beetles, diptera
and myriopods. They were not sensitive to my carbide light,
but were easily irritated by blowing on the web. The snares
were quite abundant and were constructed like those of Tro-
glohyphantes. Several cocoons wrere collected; one contained
newly hatched spiderlings which were white in color. One
spider was caught eating a thysanuran. Since the entrance to
this cave is an artificial one securely made in a sink hole, we may
expect an even temperature throughout the cave. And since
the spiders are most abundant near the entrance we may at-
tribute this phenomenon to the great number of insects.

1 Contribution No. 117 from the Zoological Laboratory of Indiana University.

2 "Biology of the Shawnee Cave Spiders," BIOL. BULL., Vol. XIX., No. 6, No-
vember, 1910.

3 These various species were identified by Dr. Alexander Petrunkevitch.

184 NORMAN E. MCINDOO.

Phanetta subterranea Emerton. — A few were collected 300 feet
from the entrance of Spring Cave (a wet cave one fourth mile
west of Marengo Cave). These were found under flat rocks and
rotten boards on the floor in damp places. They were in com-
pany with thysanurans. Several were caught in all parts of
Little Wyandotte; however, none nearer the entrance than 75
feet. A few of these were observed under flat stones, but most
of them were seen on the base, or in the cracks of stalagmites.
Those on the base were usually running about, while those in the
cracks were hanging to the underside of their tiny webs.1 The
snare is a flat sheet. It generally droops a little in the center.
The meshes are very minute and the number of attachment
threads depends on the surroundings. This species is a swift
runner and the young are entirely white. Small diptera, thy-
sanurans and myriopods were rather plentiful. This arachnid
was abundant in Saltpeter Cave under flat rocks in moist places.
They could not be affected by carbide light. Diptera and thy-
sanurans were common.

Anthrobia mammouthia Tellkampf was found only in Mam-
moth Cave. Eleven specimens were caught. Since I had to
keep in company with a crowd of cave visitors who were led by
a guide, I was unable to stop more than one or two minutes at a
place. However, I found specimens at various localities and
undoubtedly they may be taken at any place in the cave where
there is sufficient moisture. They were always found under flat
rocks, on old mouldy paper and among debris. One small white,
disk-like cocoon was observed under a flat stone. These spiders
are rather active, not fast runners, and are not difficult to catch
for they cling tightly to the rocks and to one's fingers. They
vary in color from white to a light brown.2 In most places
insect life seemed to be comparatively scarce. Perhaps these
arachnids eat old mouldy paper, remains of lunch and the limited
number of small insects.

Meta menardi Latreille. — They are rather common at the
mouth of Spring Cave. One was observed in Saltpeter Cave and
five were caught at the end of the main channel in Mammoth

1 Blatchley says these are wandering spiders and spin no webs. " Indiana Caves
and their Fauna," Rep. Ind. Geol. Surv., XXI., p. 204.

2 Hitherto they have always been described as being only white.

ARACHNIDS FROM OHIO VALLEY CAVES. 185

Cave. According to the guide this locality is two and one half
miles from the entrance. Concluding from the number of webs
at this place, Meta must be common. Since Meta is an outside
form and is ordinarily never found very far from the entrance of
a cave, there is evidently an opening somewhere near this place.
Here as elsewhere in this cave, cave crickets, small diptera and
blind beetles were often seen.

Scotolemon flavescens Cope. — This cave harvestman was very
abundant in Wyandotte. They were found some distance from
the entrance on the damp floor where visitors walk, and often
at the base of damp stalagmites. They were always found in
the same places as thysanurans and where candle drippings were
common. They probably eat the candle drippings and each
other, for three couples were seen righting and several were found
dead. They are the least active arachnids I have ever seen and
are not affected by carbide light. The second pair of legs, ex-
ceedingly long, are used as tactile organs. Several individuals
of this species were caught in Little Wyandotte.

Phalangodes armata Tellkampf. — Two specimens of this cave
harvestman were taken in Mammoth Cave, one at River Styx
and the other at end of main cave. It is much larger and more
active than the preceding species. It is usually found on the
walls and not on the floor.

Chtonius Packardii Hagen. — One specimen of this small semi-
blind pseudoscorpion was taken in Wyandotte, two in Little
Wyandotte and one in Mammoth. I have also taken it in
Shawnee Cave. It is usually found under damp rocks. It
moves along slowly with its chelae held in the air in front, and is
very difficult to find.

In order to keep the specimens from any of these caves alive
very long, it was necessary to place them in a saturated atmo-
sphere. They were most conveniently kept in small vials. One
individual with a drop of water was placed in each vial. In such
confinement several died in a few days, but the majority sur-
vived for a month or more. Light experiments like those with
Troglohyph antes were prosecuted with Theridium porteri Banks
and Erigone infernalis Keys., collected in Mayfield's Cave. They
were decidedly negatively phototropic. Various outside forms

1 86 NORMAN E. MCINDOO.

were likewise experimented with. Some of these were negatively,
others positively phototropic. The kind of phototropism de-
pends on the natural environments of the specimen. It was
observed that the specimens collected on this trip always pre-
ferred the dark end of the vials. All true cave spiders are more
or less negatively phototropic. The distribution and number
of all true cave spiders are controlled by even temperature and
the abundance of food. They probably have no enemies other
than themselves and are rarely seen righting. They soon die
outside the caves unless kept in a saturated atmosphere.

A FURTHER NOTE ON KERATOSUM COMPLEXUM.

CHARLES W. HARGITT.

In my original description of this anomalous hydroid,1 it was
stated that not only was it necessary to create for it a new genus
as well as species, but that there "might be the necessity of
establishing for it a new family." It was further said: "Con-
cerning the family relations I am not disposed in this connection
to enter into any critical review. While the Perisiphonidae would
be the only one under which it might be placed, still this family
as at present denned . . . would by no means provide for the
species. For example, while there is an axial tubular mass, as
showrn in Fig. 9, there is no single one of these which bears the
hydrothecae as called for by the definition referred to. However,
the species may be left under this family till such time as adequate
revision may be undertaken, when needed modifications may
be provided."

I have since found access to a paper by W. Baldwin Spencer,
entitled "A New Family of Hydroidea, Together with a Descrip-
tion of the Structure of a New Species of Plumularia," pub-
lished in The Transactions of the Royal Society of Victoria, 1890,
a publication of very limited circulation. Had this account
been available to me at the time of my perplexity concerning
the family relations of Keratosum it would have greatly facili-
tated my insight into many features which were extremely
puzzling. I take the first opportunity to acid to my earlier
account what seems to be a solution of the point left in abey-
ance pending further knowledge.

Spencer established the family Hydroceratinidae for a rather
remarkable hydroid obtained from Port Phillip having much in
common with Keratosum in both structure and habit, though
with very sharp differences, the details of which need not be
given here since only the matter of its family features are in
question. One interesting point of coincidence may be cited,
namely, that Baldwin at first referred his specimen to the family

1 BIOL. BULL., Vol. XVII., p. 379.

I87

1 88 CHARLES W. HARGITT.

Ceratelladse, Gray, thought to belong to the sponges, but later
recognized as hydroids. Further comparison convinced the
author that his species could not be included under the Ceratel-
ladce, hence his institution of the Hydroceratinidae. The follow-
ing is his definition of the family:

"Family HYDROCERATINID/E."

"Hydrophyton consisting of a mass of entwined hydrorhiza,
with a skeleton in the form of anastomosing chitinous tubes: the
surface is studded with tubular hydrothecse, into which the
hydranths can be completely retracted. Hydranths sessile and
connected with more than one hydrorhizal tube, claviform with
a single verticil of filiform tentacles. Defensive zooids present
with a solid entodermal axis and nematocysts borne at the distal
end."

To one who has any considerable knowledge of hydroid mor-
phology it will hardly be necessary to point out the more obvious
features in this definition which directly or approximately em-
body corresponding features as described in the account of
Keratosum. There are, however, certain points in which I am
not certain that the definition of Hydroceratinidae would wholly
apply to Keratosum. For example, it is stated that "hydranths
[are] connected with more than one hydrorhizal tube." I have
directed attention to the complex anastomoses of the siphcnal
(this term seems more accurate than Spencer's hydrorhizal)
tubes but I have not been able to confirm the condition diagram-
matically portrayed by his Figs. 3 and 13. It must be noted,
however, that my material was not in that vital condition
rendering easy and certain the demonstration of a point like the
one in question. Still the tubular and hydrothecal relations
were such as to render it perfectly sure that the conditions
figured by Spencer are lacking in Keratosum.

But with this difference granted it does not seem sufficiently
great to vitiate the many points of agreement which are more
fundamental and characteristic as family features. There are
also apparent differences as to the nematophores of the two
species, yet these are rather specific than even generic and may
be disregarded in this connection. Furthermore, there is not in

A FURTHER NOTE ON KERATOSUM COMPLEXUM. 1 89

Keratosum any tendency toward a bilaterality in the growth
habit such as described for Clathrozoon wilsoni. But here again
we have a feature which need hardly be considered incompatible
in the family aspects of the species. One finds such differences
of aspect in genera and species of Gorgonidae, to which Clathro-
zoon bears some superficial resemblance. I am constrained to
believe that in its family relations Keratosum is more closely
akin to Hydroceratinida? than to any existing family of Hydrozoa,
and am quite prepared to propose that it be so designated, at
least until stronger reasons are found for a different disposition.

In closing, attention may be called to a still further coincidence
between Keratosum and Clathrozoon, namely, the absence in both
of any trace of reproductive organs. I had already emphasized
this in the original description of Keratosum, stating that while
an absence of germ cells might call for no surprise, yet "if
gonangia are an organic part of the skeleton one might expect
some trace of them, . . . but none could be recognized."

This further inquisition into the morphology of these hydroids
tends to confirm and emphasize what had been said in concluding
my original description, that "we have in this hydroid one of
the most interesting, and in some ways anomalous, of this re-
markable group of organisms." The foregoing comparison of
the species with that from Australia, Clathrozoon wilsoni, tends
to further accentuate this impression.

NAPLES ZOOLOGICAL STATION,
January 3, 1911.

Vol. XX. March, 1911. No. 4.

BIOLOGICAL BULLETIN

THE NESTS AND LARV^: OF NECTURUS.

BERTRAM G. SMITH.

THE NESTS.

I. Nests in a Lake Habitat. — Through the courtesy of Professor
Bennet M. Allen I recently became acquainted with the spawning
grounds of Necturus maculosus Rafinesque in Lake Monona, Wis.
During the latter part of June and the early part of July, 1910,
several trips were made to the locality for the purpose of ob-
taining embryological material.

The "nests" were found in water about 3-5 feet deep and
about 50-100 feet from the shore, in a locality where the bottom
was strewn with loose flat stones of various sizes. The largest
of these stones, about 1^-2 feet in diameter, frequently served as
cover for the eggs of Necturus. The eggs are attached by the
slender stalks of the gelatinous envelopes singly to the under
sides of these stones, distributed over an area about 8-IO
inches in diameter (see Fig. i). The presence of minute algse,
etc., in the water made it so opaque that it was impossible to
see the bottom; the eggs were obtained by wading in the water,
feeling about with the feet for a large flat stone, then bringing
it to the surface.

Eycleshymer ('06) describes nests of eggs attached to the
under sides of logs, boards, pieces of tin, canvas, etc., but does
not mention finding nests under stones. Doubtless any con-
venient object may be selected as cover.

The number of eggs present in a nest was determined in five
cases as follows: 18, 61, 80, 84, 87. The average is 66. The
nest photographed contained 84 eggs.

The first embryos were obtained on June 22; these were in

an advanced stage of development, with head well formed and a

191

192

BERTRAM G. SMITH.

small tail rudiment. The latest embryos were collected on July
5; at this time nearly all the embryos had hatched, only a few
unhatched embryos being found in each of several nests. The
empty capsules remained attached by their stalks.

Eycleshymer ('06) states: "The time of egg-laying varies in
different lakes, depending upon the time when the temperature

FIG. i. "Nest" of eggs of Necturns. The stone to which the eggs are attached
has been removed from the water and set on edge on the wharf; it is about 16 inches
in diameter. The embryos are in an advanced stage of development.

of the water reaches a certain degree. In the larger, deeper
lakes with bold shores this is much later than in those possessing
wide shoals. . . . According to Professor Whitman's and my
own experience the best time for collecting is during the middle
and latter parts of the month of May. The writer has collected
eggs as early as May 3, and as late as June 5, but these extremes
mark the beginning and closing of the early and late seasons."
Unless the dates given are for newly-laid eggs, which would

THE NESTS AND LARV/E OF NECTURUS. 193

hardly be inferred from the text, my experience shows that the
collecting season may be much later than Eycleshymer recorded.

A noticeable feature of the development as compared with
other amphibians that I have studied, is the uniformity in the
stage of development of embryos found in different nests in the
same locality. On each of the following dates from four to seven
nests were secured : June 22, 25, 29, July 5. On each date all the
eggs were found so nearly in the same stage of development that
only slight differences could be detected in eggs from different
nests. This uniformity points to a very short spawning season-
perhaps two or three days — in this locality; it would seem that
all the eggs in a restricted area are laid at nearly the same time.
But Eycleshymer ('06) says: "The eggs are first deposited in
those localities where the water is shallow and exposed for the
greater part of the day to the rays of the sun. The period of
egg-laying usually covers two or three weeks. There is no foun-
dation whatever for the statement made by Hans Virchow that
the animals deposit their eggs so to speak at the same hour."

I had no success in keeping embryos of Necturus alive in the
laboratory, although later in the year larvae of Cryptobranchus
thrived under the same conditions.

2. Nests in a Stream Habitat. — I am not aware of any published
observations on the nesting of Necturus in streams, hence the
following notes on the subject may be of interest.

During the late summer and early autumn of the past five
years, while searching for adults, eggs and larvae of Cryptobran-
chus, I have had occasion to overturn numberless stones in the
bottom of a large creek tributary to the Allegheny River, in
northwestern Pennsylvania. This has resulted in the frequent
finding of specimens of Necturus.

All the specimens found were small, none exceeding 20 cm.
(8 in.) in length and most of them were much smaller. The
smallest, taken September 13, 1906, was 35 mm. long (see Fig. 6).
This specimen was one of a group of six or seven found under the
same stone; the others, aided by the swift current, escaped.

These circumstances led me to suspect that the stream was
a spawning ground for Necturus. This suspicion received con-
firmation when, on August 24, 1910, I found attached to the

194 BERTRAM G. SMITH.

under side of a large rock about fifteen or twenty empty egg
capsules of Nectums, all in good condition. At the distal end
of each capsule was a large round hole through which the embryo
had escaped.

THE LARV.E.

i. The Embryo at the Time of Hatching. — Fig. 2 is from a
photograph of an embryo of Necturus obtained from Lake Mon-
ona on July 5, 1910. This embryo was apparently ready to
hatch, since nearly all the other embryos in the nest had already
hatched out.

FIG. 2. Necturus embryo ready to hatch, killed in Tellyesnicky's fluid and pre-
served in formalin, July 5, 1910. (X 3%-) From Lake Monona.

The embryos of Necturus are hatched in a less advanced stage
of development than is the case with Cryptobranchus (see Smith,
'07, Fig. 9); but in Cryptobranchus at least there is considerable
variation in the time of hatching and the figure referred to repre-
sents one of the more advanced of the newly-hatched embryos,
hence the difference may be less than the figures indicate.
Necturus is hatched in water of a much higher temperature than
is the case with Cryptobranchus, and this would naturally tend
to soften the gelatinous envelope and aid in the early escape of
the embryo.

In Necturus the large yolk content at the time of hatching is
even more noteworthy than in Cryptobranchus. Though set

THE NESTS AND LARVJE OF NECTURUS. 195

free as a larva, the animal is really in an embryonic state so far
as its food supply is concerned, until long after its liberation.
The only advantages secured by hatching in this condition would
seem to be better aeration and opportunities for exercise; these
advantages must be in part offset by the increased danger of
capture by some larger animal.

Aside from differences in the stage of development, the em-
bryos of Necturus and Cryptobranchus at about the time of hatch-
ing are so much alike that it would be difficult to tell them apart
were it not for the noticeably greater bulk of the latter. At the

FIG. 3. Necturus larva reared from advanced embryo obtained from Lake Mon-
ona, and preserved in formalin July 31, 1910. (X

time of hatching the Necturus embryo measures about 18 mm. in
length, the Cryptobranchus embryo about 24 mm.; this difference
may be due in part to the more advanced condition of the latter,
but throughout the entire embryonic history the size of embryos
of corresponding stages is so much greater in the case of Crypto-
branchtis that though eggs from different nests of the same
species may vary slightly in size, the extremes of variation of the
two species do not overlap.

2. Larval Development. — I have examined a series of early larval
stages kindly loaned me for the purpose by Dr. Bennet M. Allen.
The series comprises twelve specimens raised from advanced
embryos obtained from Lake Monona and preserved in formalin
on the following dates: July 31, August 19, 21, 25, 1910.

196 BERTRAM G. SMITH.

Specimens preserved July 31 (see Fig. 3) average about 25
mm. long. As compared with Cryptobranchus larvae of the same
age after hatching, in Nectunts the general form of the body is
more slender, except that the yolk sac is of relatively greater
size. The absolute size of the Cryptobranchus larva is much
greater. The pigmentation in Necturus is less intense, and the
general appearance is that of a less advanced stage of develop-
ment.

These specimens show the beginning of the dorso-lateral
stripes, which attain great prominence in later stages and will
be described in the stage in which they are most marked.

In the larvae preserved during the latter part of August, the
yolk sac still persists, though its size is so reduced that the
abdomen is only slightly distended' by it. The lateral stripes
are quite distinct, though not so conspicuous as in slightly later
stages. The larvae average about 30 mm. in length. In Crypto-
branchiis larvae of the same age after hatching, the yolk sac is
so reduced that the abdomen is no more distended than in the
adult; the color is much darker than in Necturus, the form of
the body stouter, and the absolute size nearly twice as great.

For a larval specimen 34 mm. long, I am indebted to Mr.
L. W. Harrington, formerly an assistant in the Zoological Labora-
tory of the University of Michigan. This specimen was obtained
by Mr. Harrington from fishermen in the Detroit River on
November 24, 1906, and was examined by me on the same day,
before preservation. The transparency of the ventral abdominal
wall enables one readily to note that the yolk has been entirely
absorbed. Since the coloration conforms accurately to that of
the Lake Monona specimens, a description of the color pattern
will serve for all the western larvae examined by me.

As contrasted with the adult, the most striking feature of
the 34-mm. larva is the presence of a conspicuous dorso-lateral
longitudinal stripe, light yellow in color, on each side of the
body. The lateral margin of each stripe is metamerically crenate.
The body stripes continue unbroken and without fusion to the
tip of the tail; anteriorly, they are separated by a slight break
from similar stripes along the margin of the dorsal surface of the
head ; these latter are sometimes connected at their anterior ends
by a faint transverse bar over the tip of the snout.

THE NESTS AND LARV.E OF NECTURUS.

The ground color of the dorsal and lateral surfaces is a dark
brown, with small scattering spots of lighter color especially
noticeable along the sides of the body and tail. As compared
with the earlier stages the pigmentation is much more intense,
but this merely serves to accentuate the light-yellow stripes.
It is to be particularly noted that in all the western larvae that I
have examined, the dark color of the dorsal surface of the body
is continued along the dorsal edge of the tail between the lateral
stripes (see Fig. 4). The ventral surface is pale yellow, almost
transparent.

FIG. 4. FIG. 5.

FIG. 4. Caudal portion of a 34-mm. Neclurus larva taken from the Detroit
River, showing color pattern of the tail. The specimen is somewhat shrivelled
from partial drying before preservation; in particular the ventral margin of the tail
is upturned. Photographed after preservation in formalin. (X 3M-)

FIG. 5. Caudal portion of a 35-mm. larva of Necturus taken from a stream in
northwestern Pennsylvania, showing color pattern of the tail. Photographed from
a formalin specimen. (X 3/^.)

Mention has already been made of larvae taken from a stream
habitat in northwestern Pennsylvania. No attempt was made
to capture all the specimens found, but a series was preserved
representing a gradation in size from the smallest, 35 mm. long,
to the largest, 20 cm. in length. Of these the fourteen smallest,
all under 15 cm. in length, show larval characteristics in the
color pattern.

The smallest specimen, 35 mm. in length (see Fig. 6), was
taken on September 13, 1906. This larva is younger and mor-
phologically less advanced than the slightly smaller larva ob-
tained from the Detroit River; but on account of an important
difference in the color pattern its description has been deferred.

198

BERTRAM G. SMITH.

By dissection it was found that the yolk sac, though reduced
almost to the form of a digestive tube, still contained a small
amount of yolk. Assuming that this larva had been hatched
about 8-10 weeks, we find a similar condition of the yolk sac
in Cryptobranchus larvae of the same age after hatching.

FIG. 6. Living larva of Neclurus, 35 mm. long. (X J.)

This specimen differs from the western larvae in that the dorso-
lateral stripes unite in the median line at the base of the tail,
to be continued as a single stripe along the dorsal edge of the
tail (see Fig. 5). Since this peculiarity is present in all the
fourteen larvae that I have examined from the eastern habitat,
it would appear to be a constant difference between the eastern

FIG. 7. Larva of Necturus, 55 mm. long, photographed after preservation in
ormalin. Nearly actual size.

and western forms. Otherwise the color and color pattern of
both eastern and western larvae are the same.

The larger specimens may be described as follows: A 42-mm.
larva taken on August 29, 1907, differs morphologically from
the 35-mm- larva in its slightly greater yolk content; evidently
it is younger, though of greater size. A specimen 55 mm. long

THE NESTS AND LARY/E OF NECTURUS. 199

(Fig. 7), taken on August 20, 1906, agrees in its color and color
pattern with the 35-mm. larva; the yolk has been entirely ab-
sorbed. Another 55-mm. larva taken September 13, 1906, is in
the same condition. A 6o-mm. specimen taken August 19, 1910,
shows a slight loss in the distinctness of the stripes.

The striped pattern characteristic of the larvae reaches its
culmination in specimens of about 55 mm. body length. From
this time on it gradually disappears, the light yellow stripes
being obscured by dark pigment; during the same period the
large black spots, which give the mottled appearance to the color
pattern of the adult, become established. With the attainment
of a body length of 15 cm. the larval color pattern is entirely
replaced by that of the adult. Several specimens with a body
length of 20 cm. were dissected and found to be sexually mature.

Eycleshymer ('06), after describing the variation in color of
the adult, continues: "It is probable that these variations in
color are responsible for a number of specific names. As an
instance I might state that some years ago Dr. Gamier ('88)
described a small Necturus, taken from the Maitland and Luck-
.now Rivers in Ontario, to which he gave the name Menobranchus
lateralis, var. latastei. 'The coloration above was black, the
abdomen sooty and the gular fold white.' During the summer
of 1904 the writer was fortunate enough to secure twro young
animals which measured about 4 and 6 inches respectively. The
smaller corresponds closely to the description given by Dr.
Gamier and there is every reason for believing that the animal
in question is the young of Necturus maculosus. The older of
the two presents the general coloration of the adult. That
Necturus should undergo such striking changes in color may ap-
pear remarkable to one who has not studied the early stages but
when one has followed the changes in color pattern during growth
he finds that they are no less striking and remarkable than in
the birds/'

While I do not doubt the evidence of great variability in
the color and color pattern during growth, furnished by those1

1 In a letter from Dr. Whitman to the writer, dated April 22, 1907, he says:
"I have reared Necturus from the egg, and I can assure you that now and then a
single dark brown individual is found among the striped ones. I raised one from the
egg, and have had two captured by net. They are rare, but are unquestionably
Necturus in the cases mentioned."

2OO BERTRAM G. SMITH.

who have reared Necturus from the eggs, the fact remains that
in all the specimens studied by me not a single non-striped larva
has been found, nor is any specimen really black. Moreover,
I would point out that Dr. Garnier's description above quoted
would make an excellent account of the coloration of a year-old
larva of Cryptobranchus allegheniensis, having external gills and
with a body length of 8 cm.

ZOOLOGICAL LABORATORY,

UNIVERSITY OF WISCONSIN.

LITERATURE.
Eycleshymer, Albert C.

'06 The Habits of Necturus maculosus. Amer. Nat., Vol. XL., No. 470.
Gamier, J. H.

'88 On a New Species of Menobranchus. Proc. Can. Inst., Ser. 3, Vol. 5, pp.

218-219.
Smith, Bertram G.

'07 The Life History and Habits of Cryptobranchus allegheniensis. Biol.
Bull., Vol. XIII.. No. i.

A 74 MM. POLYODON.

C. H. DANFORTH.
(From the Anatomical Laboratory of Washington University.)

The developmental stages of Polyodon have long been sought
by biologists but, so far as published accounts show, neither the
fertilized eggs nor the young embryos have ever been seen.
For this reason it seems advisable to announce the capture of a
specimen smaller than any previously recorded.

The individual in question was taken from the Mississippi
River near St. Louis on July 12, 1910. It was immediately
killed and fixed in Zenker's fluid and hardened in alcohol. In the
latter medium it measured 74 mm. from the tip of its snout to
the tip of its tail. The figures show three views of this embryo
after preservation.

The features in which it seems to differ most from the adult
are the relatively large barbels, the rostrum, and the fins and tail.
The general outline of the body, too, appears rather more fusi-
form, but a number of measurements failed to bring out any
marked peculiarities in this respect.

The barbels, represented in Figs. I and 2, are approximately
a millimeter long, or .013 of the total length of the fish. This is
relatively many times their adult size. They are, however, small
and apparently rudimentary even in the young, and a macro-
scopic examination of them revealed no new points of interest.
In fish of 170-175 mm. they are still relatively large as compared
with the adult.

The rostrum of the embryo is somewhat unlike that of the
adult in outline. In the former (cf. Figs. 2 and 3) it tapers in
width from the nostril to the tip with only a slight constriction
near the middle, whereas in the adult there is a characteristic
and generally well marked constriction of the proximal half and
usually a distinct dilation of the distal half, giving the whole its
typical paddle-shaped appearance. In an embryo of 89 mm.
the rostrum is similar to that of the 74-mm. specimen, but in

201

202

C. H. DANFORTH.

individuals of 170 mm. and over the organ is much more truly
spatulate in outline. With the exception of the specimen under
consideration, the width of the rostrum was found to be relatively
greater in the small fish than in the larger ones, but width of
snout and contour are not very closely correlated so that some
rather wide snouts are not very spatulate in outline.

The length of the rostrum at different ages is shown in the

FIG. i. Lateral view of a specimen of Polyodon s/>a//n</a measuring 74 mm.
total length.

accompanying table based on measurements of twenty specimens,
ten varying in length from 74 mm. to 200 mm., and ten from
1,000 mm. to 1,300 mm. Intermediate stages were not available.
In this table the second column gives the total length from tip
of snout to tip of tail, the third gives the length of the rostrum
measured from its tip to a line passing between the anterior
corners of the eyes, and the fifth column show's the width of the
rostrum at the beginning of its distal third. The decimals in^he

FIG. 2. Ventral view of the specimen shown in Fig. i.

sixth column are obtained in each case by dividing the length
of the rostrum by the total length of the fish. Professor Stock-
ard1 measured in this way a number of adult Polyodon from 24
in. (about 600 mm.) to 69 in. (about 1,725 mm.) and found an
interesting correlation between the size of the fish and the length

JStockard, Charles R., "Observations on the Natural History of Polyodon
spathula," Amer. Nat., Vol. XLI., pp. 753-766, 1907.

A 74 MM. POLYODON. 2O3

of the rostrum. His tables show a steady increase in the relative
length of the rostrum in passing from larger to smaller individuals.
My own measurements reveal the presence of somewhat striking
individual variations, but it will be seen, nevertheless, on ex-
amining the table that in the first ten the length of the snout is,
on the whole, gradually increasing while with the latter ten the
reverse is the case. The smallest fish measured by Professor
Stockard was approximately 600 mm. in length and had a ros-
trum represented by the fraction .333. The largest of the small
fish measured for the table presented herewith was 200 mm. long
with a rostrum represented by .390. The available data, there-
fore, seem to indicate that the rostrum attains its maximum

FIG. 3. Dorsal view of the same fish.

development at a time when the fish is from 200 mm. to 600 mm.
in length. The biology of the species is too little known to
enable one to say whether or not there is any functional signifi-
cance :n such a development at this time.

In adult Polyodon the fins show a marked angularity. In the
embryo they are rounded and much less distinctive. This dif-
ference will become evident on comparing the figures accompany-
ing Professor Stockard 's article with those presented here. The
position of the various fins does not seem to differ greatly between
the small and the large fish.

The tail of the embryo is rather strikingly unlike that of the
adult. It is much more obviously heterocercal and the dorsal
part is somewhat lobed as shown indistinctly in Fig. I. The
small terminal lobe, which is not found in the adult, seems to
be either without rays or with rays only feebly developed. It is
a fleshy mass with apparently no tendency to become elongated
into a slender filament. It is not clear from the material at

204

C. H. DANFORTH.

TABLE OF MEASUREMENTS.

Fixing Fluid.

Total
Length,
mm.

Length of
Rostrum,
mm.

Length of
Operculum,
mm.

Width of
Rostrum,
mm.

Rostrum by
Total Length,
Decimal.

Zenker's

74

22

19

6

.297

Formalin

89

26

21

9

.292

Zenker's

9i

27

22

8

.297

Formalin

104

34

27

ii

.327

107

34

28

12

.318

140

48

35

16

•343

144

49

36

16

•340

170

58

46

18

•341

175

58

46

18

•331

200

78

50

20

•390

1,000

260

300

72

.260

1,030

290

280

80

.282

1,050

310

300

75

•295

1,070

260

290

72

•243

1,090

300

300

78

•275

1,180

320

340

80

.271

1,200

340

300

80

.283

i

1,210

330

370

90

• 273

t

1,210

310

330

80

.256

t

1,300

330

370

75

•254

hand whether this portion disappears with growth or becomes
incorporated with the rest.

Measurements were made upon several other parts of the body
but the data obtained do not seem to indicate any special
peculiarities beyond those already mentioned. Of these meas-
urements the length of the operculum is alone included in the
table, the figures in the fourth column representing the distance
in millimeters from the posterior corner of the eye to the tip of
the opercular flap.

The age of the specimen cannot, of course, be stated, but it
seems probable that it is from the brood of the current season.
Professor Stockard's observations indicate an early spawning.
During the summer some specimens only slightly larger than
the one described wrere taken, while others had reached 175 mm.
or more in length, and in the winter specimens from six to ten
inches long are captured. Consequently, if these fragmentary
observations are to be trusted, it seems probable that these fish
hatch early in the spring and grow rapidly during the first year,
attaining a length of possibly ten inches in this time.

The writer is indebted to Drs. R. J. Terry and V. E. Emmel
for securing and preserving the specimens studied. The drawings
are the work of Mr. W. T. Oliver.

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS.

IV. THE EFFECTS OF EXTERNAL FACTORS, ACTING BEFORE
OR DURING THE TIME OF FERTILIZATION, ON THE SEX
RATIO OF Bufo lentiginosus.

HELEN DEAN KING,
THE WISTAR INSTITUTE.

At the present time practically all of the investigators who are
working on the problem of sex-determination believe that the
sex of a zygote is fixed at a very early period, and that it is not
dependent upon environmental factors acting during embryonic
development. There is considerable diversity of opinion,
however, regarding the time when sexual differentiation takes
place and the way in which it is brought about.

A number of investigators, among whom may be mentioned
Rauber ('oo), Beard ('02), Schultze ('03) and Russo ('09) main-
tain that sex is determined in the ovary, chiefly by nutritive
conditions. According to this view mature eggs are either male
or female, and the spermatozoan can take no part whatever in
determining the sex of the egg it fertilizes.

Other investigators, prominent among whom are Stevens ('05),
Wilson ('06) and Morgan ('07, '09, '10), are inclined to the opinion
that sex is not determined until the time of fertilization, and that
it is the spermatozoan that definitely and unalterably fixes the
sex of the individual.

These two views are apparently irreconcilable, and such an
array of seemingly indisputable facts has been brought forward
in support of each of them that it may be, as Jordan ('09) has
recently suggested, that "sex has been attained by several paths,
and is now determined in different modes and at different times
in the different groups of animals and plants." If this be true,
it will be futile to try to find a single factor, or set of factors,
that is the fundamental cause of sex in all organisms ; and in-
vestigations in this field must be confined to an attempt to dis-

205

2O6 HELEN DEAN KING.

cover the conditions that determine sex in the particular form
selected for study.

The important experiments of Tower ('10) with the eggs of
the potato-beetle, Leptinotarsa, and also those of Tennent ('10)
with cross-fertilized echinoderm eggs, have shown that very
definite alterations in the structure of the developing organism
can be produced by changing the external conditions that sur-
round the egg at the time of fertilization. The recent investi-
gations of Shull ('ioa, '106) and those of Whitney ('10) on the
rotifer, Hydatina senta, indicate that the chemical composition
of the water in which the cultures live determines the appear-
ance of the sexual cycle. But what the substances are that
bring about the appearance of the sexual forms, and how they
act, has not yet been ascertained.

All of these experiments show that environmental factors
can greatly change the course of development even if they act
for only a comparatively short period on the unsegmented egg:
whether they can influence anything as fundamental as sex,
however, remains to be determined. It was for the purpose
of testing this point that a series of experiments on the eggs of
the toad, Biifo lentiginosus, was undertaken two years ago.
The first of these experiments showed pretty conclusively that
temperature, acting at the time of fertilization is not the domi-
nant factor in the determination of sex in this form (King, '10).
The present paper records the results of the second series of
experiments along this line. It is hoped that the method of
investigation employed will eventually settle the point as to
whether sex in Bufo is determined at or before the time of fer-
tilization, even if it gives no clue to the factors that are involved
in this process.

The Wistar Institute of Anatomy and Biology has provided
an equipment for the carrying on of these investigations that is
nearly ideal for the purpose. This equipment consists of two
iron racks, each of which supports three tiers of twenty galvanized
iron trays 9X12X23/2 inches: dishes of this size are convenient
for handling and each gives ample space for fifty large tadpoles.
A series of %-inch iron pipes, connected with the city water
system, are attached to the supporting frame five inches above

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 2O7

each tier of trays; and by means of small stop-cocks each dish
is supplied with running water. The depth of the water in the
trays is kept at about two inches by means of constant level
siphons, the water running from each siphon into a trough that
empties into a large exit pipe. By the use of this apparatus
large numbers of eggs, subjected to very different influences at
the time of fertilization, can be reared under similar external
conditions. Marked variations in the results of the several
series of experiments must be attributed to the different conditions
under which the eggs were fertilized; they cannot be due to
differences in the environmental conditions under which the
embryos developed.

The apparatus described above was used for all of the
experiments made during the spring of 1910. An abundant
supply of varous kinds of water plants was always kept in the
dishes containing the tadpoles. All of the tadpoles received
similar food, which consisted of small pieces of cooked meat or
of frog muscle, yolk of egg, cereal and so-called "fish food."
The water was allowed to run slowly in each dish for about six
hours every day, and was then turned entirely off at night to
prevent a possible accident through a change in the water pres-
sure. To prevent any loss of tadpoles through a fouling of
the water by food and impurities that could not be carried away
through the siphon, the inner end of which was covered with
fine netting, each dish was thoroughly cleaned once a week
during the first month after the experiment was started, and much
oftener when the tadpoles had become larger and the weather
warm.

As many as possible of the individuals used in each series of
experiments were carried through to metamorphosis and their
sex ascertained: the methods used in ascertaining sex being the
same as those previously employed (King, '07).

In the various tables given in this paper the figures in the
second column refer in every case to the number of individuals
alive one week after the experiment began; they do not indicate
the number of eggs that were subjected to the unusual conditions
at the time of fertilization. In investigations of this kind it is
very important, as I am fully aware, that the mortality in the

2O8 HELEN DEAN KING.

various series should be known and taken into account in con-
sidering the results. With the large numbers of eggs that were
necessarily used in these studies it was quite impossible to take
the time to make an accurate count of the eggs when the experi-
ments were made. For the eggs of the toad, when taken from
the uterus, are so closely packed together in long strings of thick
jelly-like substance that they are greatly distorted in shape, and
a rather careful examination with a lens would be necessary to
enable one to tell the exact number of eggs in any given lot.

It is possible that distinct variations in the sex ratios might
be obtained when batches of eggs from different females are
fertilized with sperm from different males. A control for each
series of experiments was therefore considered necessary, and
was obtained by fertilizing eggs from the same female with sperm
from the same male in ordinary tap water. The percentage
of females in this control lot was taken as the standard by which
to judge whether the sex ratio had been altered by the changed
environmental conditions to which the particular lot of eggs
had been subjected at the time of fertilization.

i. THE INFLUENCE OF ALCOHOL ON THE SEX RATIO OF Biifo.

To subject eggs to the action of solutions that would produce
marked alterations in the cell structure would seem to be an
easy way of ascertaining whether sex can be influenced by ex-
ternal conditions acting at the time of fertilization. Unfortu-
nately the unfertilized eggs of amphibians, as well as those of
the different invertebrates that are extensively used for experi-
mental purposes, are very sensitive to the action of various sub-
stances, even when used in very dilute solutions for only short
periods of time. It is therefore not possible to put the eggs of
Bufo in any solution strong enough to greatly modify the cell
structure without causing immediate death or inducing changes
that prevent normal development.

Experiments made by Morgan and Tsuda ('93) show that the
eggs of the frog, and also those of the toad, can develop normally
for some time if they are placed in a 2.5 per cent, solution of
alcohol, although they are not able to live in a 5 per cent, solu-
tion. As eggs and sperm, according to Overton ('02), are equally

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 209

permeable to the lower alcohols and are not greatly injured by
them, the idea was suggested that it might be possible to fertilize
the eggs of Bufo in solutions of alcohol that would be strong
enough to enter the eggs and, by combining with the fats, bring
about an alteration in the cell structure that would later manifest
itself in a change in the normal sex ratio. Batches of eggs from
the same female were therefore placed with sperm from the same
male in alcohol solutions of the following strengths: 10, 5, 2, I,
.5, .25, .13 per cent. In each case the eggs remained in the
solution for one half hour and were then transferred into fresh
water where they continued their development.

None of the eggs that were subjected to the action of the 10
per cent, and the 5 per cent, alcohol segmented. This result can
probably be attributed to the injurious effects of the solutions
on the spermatozoa which were thus rendered incapable of ferti-
lizing the eggs. Overton ('02) has shown that a 5 per cent, solu-
tion of alcohol anaesthetizes muscle, and presumably a solution
of this strength would have the same effect on the spermatozoa
of Bufo; no investigation of this point was made. At least
one half of the eggs that were fertilized in the 2 per cent, and in
the I per cent, solutions of alcohol segmented normally and con-
tinued their development: in the remaining series practically
all the eggs were fertilized.

Table I. shows the results obtained in this series of investi-
gations.

TABLE I.

EGGS FERTILIZED IN SOLUTIONS OF ALCOHOL.

Per Cent. Alcohol.

Total No.
Individuals.

No. Sex
Ascertained.

Males.

Females.

Per Cent.
Females.

2

398

134

68

66

49-25

I

326

130

69

61

46.92

• 5

260

177

81

96

54-23

•25

244

188

95

93

49-52

• 13

311

209

too

109

52.15

Control,

2IO

134

65

69

5I-4I

The sex ratios obtained in this series are very uniform, since
in no case is there a deviation of even five from the percentage
of females found in the control lot. Somewhat lower percentages
of females were found among the individuals that had developed

210 HELEN DEAN KING.

from eggs that were fertilized in the stronger solutions; but the
deviations of these percentages from that of the control are much
too small to be considered as significant, particularly as the lowest
proportion of females occurs in the lot in which the eggs has
been fertilized in I per cent, alcohol and not in the lot where
fertilization of the eggs took place in 2 per cent, alcohol. The
relatively high percentage of females (54.23) that is found where
the eggs had been fertilized in .5 per cent, alcohol is quite within
the limits of normal variations in the sex ratio of Bufo. For
previous studies in which the sex of nearly 10,000 young toads
was ascertained (King, '07, '09, '10) show that in any large lot
of individuals there is a slight excess of females, as a rule, the
proportion of females varying from 51 to 56 per cent. It is
evident from the sex ratios obtained in this series of experiments
that alcohol, in the strengths used and for the time it acted, has
no influence on the determination of sex in Bufo.

The growth of the tadpoles that developed from the eggs that
were fertilized in 2 per cent, alcohol was greatly retarded. This
was not apparent during the first month after the experiment
was made, but became very noticeable later. At the time of
metamorphosis the majority of these tadpoles had a body length
of only 7-8 mm., although the body length of tadpoles of Bufo
at this period is normally 1 1-14 mm. During early development
the mortality among these individuals was comparatively slight,
being not more than 10 per cent. At the approach of meta-
morphosis the tadpoles, especially the smaller ones, began dying
in great numbers; and during the last week of June over fifty
of them died from no apparent cause. In no other series ot
experiments was the mortality among mature tadpoles anywhere
near as great as in this case. Alcohol, therefore, although acting
for only a short time at the fertilization period in but a 2 per cent,
solution, must have so affected many of the tadpoles that they
were rendered incapable of undergoing metamorphosis. All of
the tadpoles in this lot were normal as far as could be determined
without a microscopic examination of the various organs. None
of them showed defects in the central nervous system comparable
to those that Stockard ('09, !io) has found can be induced in
embryos of fundulus by subjecting the fertilized egg to the action

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 211

of 3 per cent, alcohol. As many as possible of the tadpoles that
died at the time of metamorphosis were preserved; and sections
were made of their gonads in the hope that it would be possible
to ascertain the sex of the individuals by this means. In very
few cases, however, were the gonads sufficiently developed to
make a definite determination of sex possible.

The stunting of the individuals and the mortality at the time
of metamorphosis was much less evident where the eggs had
been fertilized in I per cent, alcohol; while in the other lots in
the series the tadpoles were as large and as vigorous as were
those in the control. In lots of tadpoles that have developed
from normally fertilized eggs under similar external conditions
there is always considerable variation in the size of the individuals
of the same age. Although, as a rule, most of the tadpoles are
of medium size, it not infrequently happens that a few giant in-
dividuals may be found that are nearly twice as large as the
smallest individuals in the lot. It is, therefore, only when the
majority of the individuals show marked deviations from the
usual size that one can safely attribute the change to the unusual
conditions to which the eggs have been subjected.

2. THE INFLUENCE OF THE SPERMATOZOAN ON THE DETER-
MINATION OF SEX IN Bufo.

In his latest contribution to the subject of sex-determination
Hertwig ('07) describes a series of experiments in which the eggs
of two different frogs were divided into six lots and fertilized with
sperm from six different males. The mortality in several of
these lots of eggs was very great, and there was no uniformity
in the sex ratios that were obtained. Some cultures gave over
90 per cent, of females; others produced an excess of males. In
spite of this lack of uniformity, Hertwig considers that the results
indicate that the male has some influence in determining sex
in Rana, although he is of the opinion that, as a rule, "die Eier
zur Zeit der Befruchtung sexuell in so hohem Grad determiniert
sind, dass der relativ geringe Einfluss des Samens gar nicht zur
Geltung kommen wurde."

Morgan ('08) has suggested that, if there be a dimorphism of
the spermatozoa of amphibians that is associated with sex-de-

212

HELEN DEAN KING.

termination, it is possible that more sperm of one kind are pro-
duced in some individuals than in others, and that distinct dif-
ferences in the sex ratios might be obtained if lots of eggs from
the same female are fertilized with sperm from different males.
There is the possibility, also, that the spermatozoa of one kind,
if two kinds exist, are produced only in one testicle, or perhaps in
considerably greater numbers in one testicle than in the other.
It seemed advisable, therefore, in investigating the possible
influence of the spermatozoan on the determination of sex in
Bufo to carry out the experiment in the following way: a batch
of eggs from the right uterus of a female was divided into four
lots of from 300-400 eggs each. Each lot of eggs was then fer-
tilized with sperm from the right or from the left testicle of one
of three different males.

The sex ratios obtained in this series of experiments are indi-
cated in Table II. In this table the males from wrhich the sper-
matozoa were obtained are designated as I, 2, 3; while the letters
R.T. and L.T. refer to the right and to the left testicle respec-
tively.

TABLE II.
EGGS FERTILIZED WITH SPERM FROM DIFFERENT MALES.

Males Used.

Total No.
Individuals.

No. Sex
Ascertained.

Males.

Females.

Per Cent.
Females.

i (L.T.)

255

126

66

60

47.61

i (R.T.)

403

224

108

116

51.78

2 (L.T.)

380

301

143

158

52.49

3 (R-T.)

221

124

55

69

55-64-

Somewhat different percentages of females were obtained in
the various lots, as might be expected in any case. As the series
stands in the above table, the greatest difference between the
percentages of females in any two lots is but 8.03 per cent. If
the percentage of females in all of the individuals that developed
from eggs that wTere fertilized with sperm from male i (50.28
per cent.) is substituted for the relatively low percentage (47.61)
found in the lot where sperm from only the left testicle had been
used for fertilization, the difference between the percentages of
females at the two extremes of the series falls to 5.36 per cent.,
which is only 1.19 per cent, greater than that (4.17) found be-

STUDIES ON SEX-DETERMINATIOX IN AMPHIBIANS. 213

tween the two lots in which the eggs had been fertilized with
sperm from the right and from the left testicle of the same indi-
vidual (male i). Variations in the percentages of females as
great as those shown in the above table have been obtained in
other cases where batches of eggs from the same female were
fertilized with sperm from the same male (Table I.)- The
results of this experiment, therefore, give no very definite support
to Morgan's contention that the male is probably the sex de-
termining factor in amphibians as it appears to be in some of
the lower forms. On the other hand, the experiment proves
nothing against this theory. For a comparison of the percent-
ages of females in the lot of individuals produced from eggs that
were fertilized with sperm from the left testicle of male I with
that of the corresponding lot in which sperm from the left testicle
of male 2 was used showrs a difference of but 4.88 per cent.; in
the two instances in which sperm from the right testicle was used
to fertilize the eggs, the difference in the percentages of females
among the young toads is 3.86 per cent.; where one lot of eggs
was fertilized with sperm from the right testicle and another lot
with sperm from the left testicle of the same individual, there
is later a difference of but 4.17 per cent, in the percentages of
females in the two lots. It seems evident, therefore, that both
kinds of spermatozoa, if two kinds exist, must be produced in
approximately equal numbers in each testicle of every male.
On the theory that the male is the sex-determining factor in
amphibians one would therefore expect to find practically equal
proportions of the sexes in any large lot of individuals, whether
the eggs had been fertilized with sperm from the same or from
different males.

The view that males are produced by the fertilization of eggs
with spermatozoa from the right testicle, while females result
if eggs are fertilized writh spermatozoa from the left testicle
has recently been advocated by Von Seligson ('950, '95&). This
theory, which is generally credited to Hippocrates (460-377 B.C.),
is not adequate to explain the determination of sex in Bufo
since, as shown in the above table, males and females are pro-
duced in approximately equal numbers when lots of eggs from
the same female are fertilized with sperm from the right or from
the left testicle of the same individual.

214 HELEN DEAN KING.

Of the many^ investigators who maintain that sex is already
determined in the ovary, only von Seligson and Dawson ('09)
have advocated the view that the right ovary produces male
eggs exclusively and that eggs which develop into females must
be derived from the left ovary. Von Seligson believes that only
the spermatozoa from the right testicle are able to fertilize
male eggs, and, conversely, that eggs from the left ovary must
be fertilized by spermatozoa from the left testicle. This, of
course, is selective fertilization for which at present there is but
little evidence. Dawson is of the opinion that the spermatozoan
takes no part whatever in determining sex. All of the eggs that
were used in the experiments described above were taken from
the right uterus of the same female. They should, therefore, ac-
cording to Dawson, have produced a great preponderance of males.
On the theory advocated by von Seligson two of the lots should
have produced a great excess of males; and, unless we assume a
struggle for existence between the sex-determining factors in
the egg and those in the sperm with a final dominance of one sex
or the other, it is difficult to see how the tadpole in the remaining
two lots could have any sex at all.

I have not suggested the possibility that in any of these lots
the individuals would be exclusively males according to the views
of von Seligson and of Dawson, since it is very probable that
the right uterus of the toad, and also that of the frog, contains
eggs from the left as well as from the right ovary. The eggs of
these amphibians break through the wralls of the ovary shortly
after the time that the germinal vesicle has disappeared prepara-
tory to the formation of the first polar spindle, and they come to
lie freely in the body cavity from whence they pass through the
oviducts into the uteri. It is very probable that all of the mature
eggs are expelled from both ovaries at the same time; and it is
not at all unlikely that some of the eggs originating in the left
ovary may pass into the right oviduct and vice versa. Pre-
sumably, however, the great majority of eggs from the right ovary
pass into the right uterus as, owing to the enormous numbers
of eggs that are produced, it does not seem possible that many of
the eggs can be moved from one part of the body cavity to the
other. The results of this series of experiments, which confirm

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 215

those previously obtained when batches of eggs from each uterus
of a female were fertilized with sperm from the same male
(King, '09), show that if two kinds of eggs are produced in the
ovaries of Bufo they must be produced in approximately equal
numbers in each ovary. Experiments similar to these can there-
fore give no possible clue to the time or to the manner in which
sex in Bufo is determined.

Hertwig ('07) found that the different lots of frog tadpoles
that developed from eggs of the same female that were fertilized
with sperm from different males showed marked differences in
their growth energy. A similar phenomenon was observed in
the various lots of tadpoles of Bufo that belonged to this series
of experiments. Eggs which were fertilized with sperm from the
left testicle of male 2 produced tadpoles which, as a rule, were
larger, more uniform in size, and apparently more vigorous than
were the individuals in the other lots. The mortality among these
individuals was comparatively low. The great majority of the
tadpoles that were produced from eggs that were fertilized with
sperm from the right testicle of male 3 were somewhat under-
sized ; and they were backward in developing, as only 39 of them
had completed metamorphosis by June 22. The mortality in
this lot was considerable. Both lots of tadpoles derived from
eggs that were fertilized with sperm from the right and from
the left testicle of male i developed at about the same rate and
were of average size. The mortality in these two lots was about
10 per cent.

Owing to a series of accidents a number of individuals in each
of these lots were lost; and therefore the comparatively high
death rate, as shown in Table II., can be ascribed in great part
to this cause. To a similar cause can also be attributed at least
10 per cent, of the mortality in the various other series of ex-
periments described in the present paper.

2l6 HELEN DEAN KING.

3. THE EFFECTS OF CHANGING THE WATER CONTENT OF THE

EGG AT THE TIME OF FERTILIZATION ON THE SEX

RATIO OF Bufo lentiginosus.

I. Extracting Water from an Egg, or Preventing the Absorption
of Water by an Egg, during the Fertilization Period.

If the sex of an embryo is definitely fixed at the time that
the egg is fertilized, it is conceivable that the sex-determining
process may be of such a character, and so evenly regulated, that
placing the egg under conditions that would cause it to lose
water or that would prevent its absorption of water during the
fertilization period might turn the balance in favor of one sex
or the other. Normally, as shown by the investigations of
Backman and Runnstrom ('09), the osmotic pressure of the
ovarian egg of Rana, and presumably also that of Bufo, drops
from .48° to .045° when the egg has been fertilized but is still
unsegmented. This lowering of the osmotic pressure must
be due in great part to the absorption of water by the egg as
Bialaszewicz ('08) has recently maintained, although the change
is ascribed by Backman and Runnstrom chiefly to an alteration
in the condition of the colloids that is induced by fertilization.
Any cause that would tend to prevent the normal increace in
the water content of the egg at the time of fertilization might,
therefore, produce changes that would be far reaching in their
results.

In attempting to ascertain whether extracting water from the
egg, or preventing the egg from absorbing water during the fertili-
zation period, would produce any alteration in the sex ratio,
two different methods of investigation were employed. For
convenience in description these two sets of experiments will be
designated as series A and series B.

Series A. — -While making experiments with the eggs of Bufo
in the spring of 1909, a pair of toads was placed in an empty
aquarium and left undisturbed for several hours. During this
time the female laid a small batch of eggs which were found to
be segmenting when the toads were removed from the aquarium
for use. The eggs which had thus been normally fertilized out
of water were placed at once in a dish of tap water and allowed to

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 217

continue their development. The female that had deposited
the eggs was killed and her remaining eggs were used for the
temperature experiments summarized in Table II. of a previous
paper (King, '10). Many of the eggs "fertilized dry" died
during the gastrulation period, and only 44 individuals lived until
it was possible to ascertain their sex. Of these 44 individuals
17 were males and 27, or 61.36 per cent, were females. The lot
of eggs from the same female that serves as a control for this
experiment, since they were fertilized with sperm from the same
male, produced 51.39 per cent, of females.

The outcome of this single experiment has little significance,
as the number of individuals in which sex was ascertained is too
small to give results of value, especially as it is not known how
long the eggs had been laid and fertilized before they were found
and placed in water. From the condition of the eggs when they
were discovered it was evident that they had been laid at least
three hours. The possibility of obtaining normal tadpoles from
eggs that had been fertilized out of water having been demon-
strated by this experiment, a further series of experiments was
made this past spring in the following way: Strings of ripe eggs
were removed from the uterus of a female and spread out in an
empty glass dish. Then, by means of a pipette, a few cubic
centimeters of a concentrated solution of spermatozoa were dis-
tributed drop by drop over the eggs in as uniform a manner as
possible. The dish was then covered with a glass plate and set
aside. Four hours later part of the eggs (lot X) were placed
in water. An examination of the eggs at his time showed that
many of them had segmented in an apparently normal manner
and were already in the 4-8 cell stage of development. The
remaining eggs (lot Y) were transferred into fresh water at the
end of seven hours, when a considerable number of them were in
the 32 cell stage. A number of eggs in both lots failed to seg-
ment, possibly because the lack of water prevented the free move-
ment of the spermatozoa.

In these experiments the loss of water from the eggs as a result
of evaporation was probably very slight, since the thick jelly-
like substance surrounding the eggs would serve to protect them
during the time that they were kept out of water. The condi-

218

HELEN DEAN KING.

tions under which fertilization was effected, however, prevented
the eggs from absorbing any considerable amount of water during
this period. The early development of these eggs was not inter-
fered with by the abnormal conditions under which they had
been fertilized, as in both series segmentation was normal and but
slightly later than that of the control lot. The mortality among
the older embryos was considerable; and somewhat greater in
lot Y where the eggs had remained the longer time out of water.
The results obtained in this series of experiments are brought
together in Table III.

TABLE III.
EGGS FERTILIZED DRY.

Total No.
Individuals.

No. of Sex
Ascertained.

Males.

Females.

Per Cent.
Females.

Lot X (4 hrs.)
Lot Y (7 hrs.)

243
280

US
72

45

21

70
51

60.86
70.83

Control

201

140

65

75

53-57

In each case, as the above table shows, a percentage of females
was obtained that is considerably higher than that in the con-
trol; the higher percentage of females being found among the
individuals of lot Y which had developed from the eggs that had
remained the longer time out of water. The possible bearing of
these results will be considered in connection with the results of
the experiments in series B.

Series B. — The second method used to ascertain whether les-
sening the water content of the egg at the time of fertilization
would produce any alteration in the sex ratio was suggested to
me by Dr. E. G. Conklin, and consisted in subjecting the unfer-
tilized eggs to the action of salt or of sugar solutions. These
solutions, being hypertonic, would tend to extract water from
the eggs or to check the absorption of wrater by the eggs. To
attempt the fertilization of eggs in these solutions would be futile,
since Loeb ('92) has shown that the addition of even 2 per cent.
NaCl to sea-water anaesthetizes the spermatozoa of Arbacia,
and a much weaker solution would probably have a similar effect
on the spermatozoa of amphibians.

Several preliminary series of experiments made during the
spring of 1909 were complete failures, since either the eggs were

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 2IQ

not fertilized at all after being subjected to the action of solutions
of different strengths for varying lengths of time, or practically
all of the eggs that segmented died during the early stages of
development. The unfertilized eggs of Bufo, as well as those of
Rana according to the investigations of Hertwig ('95), are very
sensitive to the action of hypertonic solutions. It was necessary,
therefore, to make other experiments this past spring in order
to ascertain the maximum strength of solutions that could be
employed and the length of time they could act without injuring
the eggs beyond the possibility of their being fertilized and con-
tinuing their normal development. The experiments from which
results were finally obtained were made as follows : a batch of ripe
eggs was transferred directly from the uterus of a female into a
2.5 per cent, solution of c.p. cane sugar and allowed to remain there
for ten minutes. During this time the eggs were kept in constant
motion, either by stirring them with a glass rod or by gently
shaking the dish. The eggs were then placed in a small amount
of tap water containing a quantity of spermatozoa. A second lot
of eggs from the same female were subjected to the action of a
2.5 per cent, solution of c.p. sodium chloride for the same length
of time and were then fertilized with spermatozoa from the male
used in the previous case.

Even with the weak solution that was used and for the short
time that it acted many of the eggs that had been put in the sugar
solution were killed, or at least rendered incapable of being
fertilized; not more than one half of the eggs subjected to the
action of the NaCl segmented, and many of these did not go
beyond the gastrulation stage. Morgan ('06) has shown that
the fertilized eggs of the frog can develop normally in a 6 per
cent, solution of cane sugar, but that the maximum strength of
NaCl that will permit of normal development is not much above
2 per cent. The more injurious action of the NaCl on the
eggs of the frog, as well as on those of the toad, is doubtless
due, in great part, to the fact that its osmotic pressure is several
times greater than that of sugar.

As the unfertilized eggs of Bufo are so quickly injured by
solutions of salt and of sugar it is difficult to decide what effects
these solutions could have had in the short time the eggs remained

22O

HELEN DEAN KING.

in them. Presumably a slight amount of water was extracted
from the eggs by the salt solution; but, as the osmotic pressure
of sugar is so low, it is very probable that the chief action of the
sugar solution would be to prevent the eggs from absorbing water
during the fertilization period.

The mortality among the tadpoles reared from the eggs that
had been subjected to the action of the salt solution was much
greater than when a sugar solution had been used. In both cases
the greatest mortality occurred during the first fortnight after
the experiment began. The embryos that lived beyond this
time seemed for the most part to have recovered from any inju-
rious effects of the treatment to which the eggs had been sub-
jected before fertilization. In their later development the
tadpoles in both lots appeared to be perfectly normal and of
average size. None of them showed any specific abnormalities
such as those Hertwig ('95), Gurwitsch ('95, '96) and Morgan
('03) produced by subjecting the fertilized eggs of the frog or
of the toad to the action of solutions of various salts.

The eggs used in these experiments were taken from the left
uterus of the female whose eggs, from the right uterus, were
employed in the experiments described in section 2; and they
were fertilized with spermatozoa from the male designated as
no. 2 in those experiments. The sex ratio in the lot of indi-
viduals, belonging to the former series, that developed from
eggs that were fertilized with spermatozoa from the same male
serve as a control for the present experiments (Table II.).

Table IV. shows the results obtained in this investigation.

TABLE IV.
EGGS TREATED WITH HVPERTONIC SOLUTION'S BEFORE FERTILIZATION.

Total No.
Individuals.

No. Sex -UaW
Ascertained.

Females.

Per Cent.
Females.

Salt (2.5 per

cent.)

264

91

25

66

72.52

Sugar (2.5 per

cent J

463

130

39

Qi

7O.OO

Control

380

301

143

158

52.49

Both lots, as the above table shows, gave a percentage of
females that is considerably higher than that in the control.

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 221

The percentage of females among the individuals that developed
from the eggs that were subjected to the action of the salt solu-
tion before fertilization is greater than that in the lot where
a sugar solution had been used. The difference between the
two is only 2.52 per cent., however, so it is evident that the
sugar solution had practically as great an effect on the eggs
as had the salt solution, in spite of the fact that the latter
solution has much the higher osmotic pressure.

Although the methods employed to bring about a change in
the water contents of the egg at the time of fertilization were
very different in the two sets of experiments just described,
there is a very striking uniformity in the results. Both series,
comprising five lots of eggs from three different females, show a
percentage of females much above the upper limit of the range
which is apparently normal for the species and which in three
instances is nearly 20 per cent, greater than that in the control.
These results suggest, although they by no means prove, that
extracting water from the egg at the time of fertilization, or
preventing the absorption of water by the egg during this period ,
in some way affects the sex-determining process and tends to
the production of relatively more females.

The manner in which the experiments were made, particu-
larly those in series B, seem to preclude the possibility that the
spermatozoa and not the eggs could have been affected by the
conditions under which fertilization occurred. The results
obtained suggest also that the sex-determining mechanism is
in the egg and that it can be influenced by external agencies
acting during the fertilization period.

II. The Effects of the Absorption of Water by the Egg at the

Time of Fertilization.

In connection with the experiments just described it seemed
worth while to ascertain whether any alteration in the sex ratio
of Bufo could be induced by subjecting eggs at the time of
fertilization to conditions that would tend to cause them to take
up more than the normal amount of water. An increase in the
amount of water absorbed could presumably be brought about
by fertilizing the eggs in solutions of acid or of alkali, since,

222 HELEN DEAN KING.

according to Loeb ('06), both these substances cause eggs to
absorb water, the quantity of water taken up increasing with the
quantity of acid or of alkali used.

Of the many preliminary experiments made in the spring of
1909, only one gave results that could be considered of any value.
In this case one lot of eggs was artificially fertilized in a .01 per
cent, solution of hydrochloric acid: another lot of eggs from the
same female was fertilized with spermatozoa from the same male
in a .01 per cent, solution of ammonium hydrate. In each case
the eggs remained in the solution for one half hour, and were then
transferred into fresh water. The mortality in both lots of eggs
was very high, and was probably at least 50 per cent. No record
was made of the number of individuals with which the experiment
was started. The results obtained are indicated in Table V.

TABLE V.

EGGS FERTILIZED IN A SOLUTION OF ACID OR OF ALKALI.

Solution.

No. Sex

Ascertained.

Males.

Females.

Per Cent.

Females.

Acid

12 A

72

.

J T O3

Alkali ....

126

crc

71

ZA 76

Control .

2OO

OA

106

ST.OO

The percentage of females obtained from the lot of eggs that
was fertilized it the acid solution is somewhat low, yet one
might attribute it to a chance variation in the sex ratio: the
percentage of females obtained in the lot of individuals derived
from eggs that were fertilized in the alkaline solution is but
slightly greater than that in the control, and therefore calls for
no comment. These results afford little evidence that the sex
ratio of Bufo can be altered by subjecting the eggs to environ-
mental conditions at the time of fertilization that might cause
them to absorb a greater amount of water than usual.

Hydrochloric acid is inorganic, and ammonium hydrate is
known to have a very injurious action on developing eggs. It
was thought, therefore, that possibly different results might,
be obtained if eggs were fertilized in solutions of other sub-
stances. The experiments were repeated on a much larger
scale this past year; acetic acid and sodium hydrate being

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 223

used in making the solutions. Two strengths of each substance
were used; a .01 per cent, solution and a .0025 per cent, solution.

In each case eggs and spermatozoa were placed together in
the solution and left for one half hour; the dish containing them
being kept in constant motion during this time. As the experi-
ment was carried out, therefore, the solutions might have acted
on the eggs alone, on the sperm alone, or on both eggs and sperm.
After remaining in the solution the given length of time, the
eggs were washed in several changes of fresh water and placed
in the dishes in which they were to continue their development.
The percentage of eggs that failed to develop was about the same
in the lots subjected to the action of the acid solutions as where
alkaline solutions had been employed; and was greater, as one
would expect, where the stronger solutions had been used. It is
not possible to say whether the failure of so many of the eggs
to segment is to be attributed to the injurious action of the solu-
tions on the eggs, or whether a large proportion of the spermato-
zoa were rendered incapable of fertilizing the eggs.

For these investigations the eggs of two different females
(a and &) were used. Female a was the one from which the eggs
were taken that were used in the experiments summarized in
Table III., and the control for the former experiments also serves
as control for these. Eggs from female b were used for the study
of the influence of the spermatozoan on the determination of sex
in Bufo, and also for the experiments in which eggs wire subjected
to the action of salt and of sugar solutions before fertilization
(Table IV.). In the present case eggs from female b were fer-
tilized with spermatozoa from both testicles of the male desig-
nated as no. i in the former investigation (Table II.). The
percentage of females that serves as a control in this case was
obtained by taking the percentage of females in the total number
of individuals in the former experiments that developed from
the eggs that were fertilized with spermatozoa from male I.
Only the stronger solutions were used on the eggs taken from
female b.

The sex ratios obtained in the various lots of individuals
derived from eggs that were fertilized in acid solutions are indi-
cated in Table VI.

224

HELEN DEAN KING.

TABLE VI.
EGGS FERTILIZED IN SOLUTIONS OF ACETIC ACID.

Eggs from

Strength
Solution.

Total No.
Individuals.

No. Sex
Asc.

Males.

Fe-
males.

Per Cent.

Females.

Per Cent.
Females
(Control).

Strong (.01

per cent.)

210

79

46

33

41-77

Female a

53-57

Weak (.0025

per cent.)

145

66

42

24

36.30

Female b

Strong (.01

per cent.)

258

71

42

2Q

40.84

50.28

613

216

130

86

39.81

In all cases the number of individuals in which it was possible
to ascertain sex was less than one half of that which the lots
contained one week after the experiments were started. In spite
of this great mortality, due in some cases to accidents, the results
are about the same as that obtained where eggs were fertilized in
a solution of hydrochloric acid (Table V.). In this series each
lot of individuals gave a percentage of females that is consid-
erably lower than that found in the control, and also much below
that which appears to be normal for the species. Curiously
enough the smallest percentage of females appears in the lot of
individuals reared from the eggs of female a which had been fer-
tilized in the weaker solution. This result may mean that a
solution of the strength employed in this instance is more favor-
able to the production of males, but more probably it is merely
a chance variation in the sex ratio.

The results of this series of experiments are especially inter-
esting and suggestive when compared with those obtained in
other cases where eggs from the same females were fertilized
under very dissimilar conditions. Eggs from female a were
fertilized out of water with sperm from the same male used in
fertilizing the eggs in this series. In the one case a great excess
of females was produced (Table III.); in the other case the per-
centage of females falls much below normal. Between the two
extremes of the series there is a difference of 24.53 per cent.
Eggs from female b when subjected to the action of a salt or of
a sugar solution before fertilization gave in both instances 70.00
per cent, of females; when other eggs from this female are fertil-

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS.

225

ized in acid solutions the percentage of females falls to 40.84
per cent. Between the extremes of the series there is a difference
of 31.68 per cent. These variations in the sex ratios are uniform
in the different series: they are apparently beyond the limits of
normal variations in the sex ratio of Bufo, and they are too large
to be justly attributed to mere chance variations in the sex
ratios of different lots of individuals.

The percentages of females obtained in the series of experi-
ments in which eggs were fertilized in solutions of sodium hydrate
are indicated in Table VII.

TABLE VII.
EGGS FERTILIZED IN SOLUTIONS OF SODIUM HYDRATE.

Eggs from

Strength Solu-
tion.

Total No.
Individuals

No. of
Sex
Asc.

Males.

Fe-
males.

Per Cent.

Females.

Per Cent.

Females
(Control).

Strong (.01

per cent.)

175

49

2O

2Q

59.18

Female a

53-57

Weak (.0025

per cent.)

252

98

42

56

56.12

Female b

Strong (.01

per cent.)

215

75

33

42

56.00

50.28

642

222

95

127

57-11

In this series the apparent effect produced on the sex ratio by
fertilizing eggs in alkaline solutions is the reverse of that which
seemingly results when eggs are fertilized in acid solutions, since
in each case a percentage of females is found which is somewhat
above that in the control lot. The percentages of females ob-
tained show very much less deviation from those of the control
than was the case in the acid series; and in no lot is the proportion
of females more than 6 per cent, above that of the control. It
seems probable, therefore, that these percentages, although some-
what high, have no especial significance, and that they are chance
variations in the sex ratio.

The great difference between the results obtained when eggs of
Bufo are fertilized in acid solutions and those found when eggs
are fertilized in solutions of sodium hydrate would seem to
indicate that these solutions have no effect whatever on the
sex-determining process, since both solutions would presumably
affect the eggs in the same way and tend to cause them to absorb

226 HELEN DEAN KING.

water. The rather unusual sex ratios obtained in the acid series
would be explained, on this assumption, as due either to chance
or to the action of the solution on the spermatozoa. There is,
however, another possible explanation that would serve to bring
these results in line with those obtained by subjecting eggs to
conditions which tend to lessen the water contents during the
fertilization period. It is known that egg membranes are more
readily permeable to weak acid than to weak alkaline solutions,
and as the jelly-like mass that surrounds the unfertilized eggs of
Bufo is relatively thick, as is also the zona pellucida that incloses
each individual egg, it is possible that in these experiments only
the acid solutions were able to penetrate the membranes and
act on the eggs. That this interpretation of the results is at
least plausible is indicated from the sex ratios obtained in the
series of experiments about to be described where fertilizing eggs
in acid solutions again leads to a seeming increase in the relative
proportion of males, while practically normal sex ratios are
found where the eggs were fertilized in solutions of NaOH.

In a former study of the effects of temperature on the deter-
mination of sex on Bufo (King, '10), the results obtained seemed
to indicate the possibility that temperature, acting at the time
of fertilization, might indirectly influence the sex-determining
process; a high temperature favoring the development of females,
a low temperature tending to the production of relatively more
males. These experiments showed, however, that the temper-
ature at which the eggs are fertilized cannot be the dominating
factor in determining sex in Bufo, and that at most its influence
would be to facilitate or to hinder certain processes taking place
in the egg that are concerned with sex-determination.

In the experiments just described the various lots of eggs
were fertilized in solutions that were kept at room temperature
(20° C.) during the time that they were allowed to act on the eggs.
In another series of experiments a batch of eggs from the female
whose eggs were used for the alcohol series of experiments (Table
I.) was divided into lots of several hundred eggs each. Each
lot of eggs was then fertilized with spermatozoa from the same male
in a .0050 per cent, solution of acetic acid or of sodium hydrate
that was kept at a temperature of 28°, 20° or 11° C. during the

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS.

227

half hour that the eggs remained in it. It was hoped that
by this means more certain evidence of the action of temperature
in determining sex in Bufo might be obtained, since a high tem-
perature would probably facilitate the penetration of a solution
into the egg and a low temperature would tend to retard it.
Any change in the sex ratio greater than that shown in the other
series of experiments in which eggs were fertilized in acid or in
alkaline solutions might therefore be attributed to the effects
of the temperature at which the solution acted.

A small proportion only of the eggs that were used in these
various experiments segmented; and, as in other experiments
in which eggs were subjected to abnormal conditions at the time
of fertilization, many of the tadpoles died during the early
growth period. The results obtained in the acid series are
brought together in Table VIII.; the percentage of females given
as a control being the same as that for the alcohol series (Table I.).

TABLE VIII.
EGGS FERTILIZED IN ACID SOLUTIONS AT DIFFERENT TEMPERATURES.

Temp, of
Solution.

Total No.
Individuals.

No. Sex
Asc.

Males.

Females.

Per Cent.
Females.

Per Cent.

Females
(Control).

11° C.

156

35

23

12

34-28

20° C.

104

69

46

23

33-33

5I-4I

28° C.

135

73

50

23 3I-50

395

177

HQ

58

32.76

In spite of the comparatively small number of individuals
in which sex was ascertained, very uniform sex ratios were found
in all the various lots, and in every case the percentage of females
falls far below that which is apparently normal for the species.
The smallest proportion of females was found in the lot in wrhich
the eggs had been fertilized at the highest temperature; and,
conversely, the greatest number of females appears where fertili-
zation was effected at a temperature of n°C. This is the result
one would expect if the higher temperature had facilitated the
penetration ityto the egg of a solution that could influence the
sex-determining mechanism in such a wray as to increase its
tendency to produce a male. The difference between the per-
centages of females in the two lots in which the eggs were fertil-

228

HELEN DEAN KING.

ized at the extreme temperatures is only 2.78 per cent., however,
so it is evident that the temperature at which the solution acted
had very little, if any, influence on the results.

No evidence regarding the influence of temperature on the
determination of sex in Bnfo is given by the results of these
experiments, and their chief interest lies in the fact that they
accord so well with the results of the other investigations in which
eggs were fertilized in acid solutions. If we consider the series
as a whole and disregard the possible influence of temperature
on the results, it is found that in the total of 177 individuals in
which sex was ascertained only 58, or 32.76 per cent, are females.
This is 18.65 per cent, lower than the percentage of females in
the control lot and 7.05 per cent, less than the percentage of
females obtained in the former series of experiments where eggs
were fertilized in acid solutions at room temperature (Table VI.).

The results of the experiments in which eggs were fertilized
in solutions of NaOH that were kept at different temperatures
during the period that the solutions acted are shown in Table IX.

TABLE IX.

EGGS FERTILIZED IN ALKALINE SOLUTIONS AT DIFFERENT TEMPERATURES.

1

Per Cent

Temp, of
Solution.

Total No.
Individuals.

No. Sex
Asc.

Males.

Females.

Per Cent.
Females.

Females
(Control).

11° C.

144

54

2?

27

50.00

20° C.

ISO

66

28

38

57-57

5I-4I

28° C.

255

126

65

61

48.41

549

246

I2O

12*6

51.21

In this series also the sex ratios obtained indicate that the
temperature at which the solutions acted has no appreciable
influence on the results, although here, as in the acid series, a
slightly lower percentage of females was obtained where the eggs
had been fertilized at the highest temperature.

Both lots of eggs that were fertilized at the extremes of tem-
perature used gave a percentage of females slightly below that
of the control and some 8 per cent, lower than that in the lot
where the eggs had been fertilized at room temperature. In
the latter case the relatively high percentage of females cannot be
considered as significant, as it is balanced by the low percentages

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 22Q

of females obtained in the other two lots. In the total of 246
individuals in this series in which sex was ascertained 126 were
females. The proportion of females in this series is therefore
51.21 per cent., which is remarkably close to that in the control
lot (51.41 per cent.).

In none of these experiments does the observed percentage of
females differ in any great degree from that of the control. The
results therefore agree very well with those of the other series in
which eggs were fertilized in solutions of NaOH (Table VII.).
Since the fertilization of eggs in solutions of NaOH produce no
apparent effect on the sex ratio of Bufo, it is evident that either
the solutions used were not strong enough to penetrate the egg
membranes and act on the egg in the short time in which the
eggs remained in them, or that NaOH is not a substance that can
act on the egg in such a way as to increase its tendency to pro-
duce a male. A further series of experiments will be made in
which the eggs are subjected to stronger solutions of NaOH
before fertilization. It is probable that the eggs are less readily
injured than the spermatozoa by various solutions, and it may be
possible to subject them to the action of a comparatively strong
solution without serious injury.

DISCUSSION.

The possibility that the relative amount of water in the am-
phibian egg at the time of fertilization may have some influence
in determining sex seems to me to afford a plausible explanation
of the unusual sex ratios obtained by Hertwig ('06, '07), and by
Kuschakewitsch ('10), in lots of frogs developed from eggs that
were overripe when fertilized. In several of his experiments
Hertwig obtained from 70-90 per cent, of males; while in one
experiment made by Kuschakewitsch, in which the fertilization
of a batch of eggs of Rana esculenta was delayed for 89 hours,
all of the resulting individuals were males. As the mortality
in this last lot was only from 4-6 per cent, it is not possible to
explain the results as due to selective mortality. Hertwig at-
tributes the great excess of males in these experiments to the
"Kernplasmarelation" existing in the eggs at the time that they
were fertilized. He believes that young eggs, and also those that

230 HELEN DEAN KING.

are overripe when fertilized, tend to produce males because they
are relatively richer in nuclear substance; females, on the other
hand, are produced when the egg is relatively poor in nuclear
substance at the time that it is fertilized.

It is very probable that the eggs of the frog normally absorb
a slight amount of water during their passage down the oviducts
and while they were retained in the uteri. If the eggs remain
in the uteri two or three days beyond the time that they would
normally have been deposited and fertilized it is not improbable
that they continue to absorb water during this time. Over-
ripe eggs, therefore, would contain relatively more water than
those that were laid at the normal time. The sex ratios found
when such eggs are fertilized accord very well with those that
I have obtained by fertilizing the eggs of Bufo in acid solutions.
In both cases it seems probable that the water contents of the
eggs at the time they were fertilized was higher than usual;
and one is strongly inclined to attribute the great excess of males
to this cause.

Experiments made two years ago with the eggs of Bufo to
ascertain whether delaying fertilization would produce an alter-
ation in the sex ratio (King, '09) gave negative results. The
eggs of Bufo are not as favorable as are those of Rana for experi-
ments of this kind, as it is not possible to delay the deposition of
the eggs more than a few hours. In the experiments referred to
a female was killed as she was about to deposit her eggs. The
body was not opened until seven hours later when the eggs were
artificially fertilized. Post-mortem changes had evidently begun
as, although the majority of the eggs segmented in a more or
less regular manner, many of them failed to develop beyond the
gastrulation stage. The sex ratios were practically normal in
the lot of individuals in which sex was ascertained. In the light
of these recent experiments the negative results that were obtained
can perhaps be atributed to the fact that either the eggs were
unable to absorb water during the time that they remained in
the uteri of the dead female, or that they were not kept a suffi-
ciently long time to enable them to take up enough water to pro-
duce any effect on the sex-determining process.

Unfortunately there is one source of error in all of the experi-

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 23!

ments described in this paper that makes it impossible to draw
any positive conclusions from the results obtained, however defi-
nite they may appear in some cases. The mortality in the various
lots of individuals was very great and only a small percentage of
the eggs with which any experiment was started were carried
through to metamorphosis and their sex ascertained. The possi-
bility exists, therefore, that the results may be due solely to chance
or to selective mortality, and that in no case was the normal sex
ratio really changed by subjecting eggs to the influence of various
external factors at or before the time of fertilization. There is, how-
ever, no evident relation between mortality and sex in tadpoles
reared under artificial conditions, as shown by the investigations
of Pfliiger ('82) and also by my former work. It does not seem
probable, therefore, that in certain lots subjected to similar treat-
ment at the time of fertilization more males than females would
always die; while in other lots, where the eggs had received dif-
ferent treatment, the mortality would invariably be greater among
the females. Practically normal sex ratios were obtained in all
of the control experiments, in the alcohol series, and also in that
in which lots of eggs from the same female were fertilized with
spermatozoa from different males. This uniformity in the re-
sults of so many series of experiments makes it even more im-
probable that mortality was selective in other cases ; neither does
it seems possible that the unusual sex ratios obtained in some
instances could be chance variations.

The results of the various experiments in which eggs were
fertilized in solutions of acetic acid are so uniform that they
seem to me very suggestive, even if no definite conclusions from
them are possible. Altogether a total of seven lots of eggs from
four different females were fertilized in acid solutions. In every
instance the percentage of females that was obtained was from
10-20 per cent, lower than that which is probably normal for
the species. There is apparently something in this series that
stands for a tendency to the production of males rather than of
females. These results taken in connection with the equally
clear-cut results that were obtained in the five experiments in
which water was extracted from the egg, or the egg prevented
from absorbing water, would seem to indicate that the relative

232 HELEN DEAN KING.

amount of water in the egg at the time of fertilization has some
influence in determining sex; an increase in water content tending
to produce a male; a lowering of the water content favoring the
development of a female.

It is, of course, possible to explain these results in other ways
besides attributing them to chance or to selective mortality.
On the assumption that the male is responsible for sex-deter-
mination in Bufo the results of the acid series of experiments
can be attributed to the fact that the solution killed more of
the female-producing than of the male-producing spermatozoa;
why this should occur is not at all clear. The explanation of
the relatively high percentages of females among the individuals
derived from eggs that were subjected to the action of a salt
or of a sugar solution before fertilization must be ascribed, ac-
cording to this view, to the fact that the solutions rendered the
eggs more resistant to the entrance of one kind of spermatozoa
than to the other. This would be, in effect, selective fertilization
for which at present there is no evidence, unless we accept the
case of the pigeon. According to the investigations of Aristotle,
of Fluorens ('64) and of Cuenot ('99) the first of the two eggs
laid by the pigeon almost always produces a male. Selective
fertilization therefore very probably occurs in the pigeon if the
male is responsible for sex in this form.

That the results of any of these experiments are conclusive,
I do not for a moment maintain. They seem to me to be unusual
enough, however, to warrant the continuation of investigations
along these lines. As they stand the results strongly suggest
that sex in Bufo is determined at or near the time of fertilization,
and that external factors acting during this period may influence
the sex-determining mechanism in such a way as to cause it to
produce one sex or the other. The results also seem to indicate
that in Bufo sex is determined in the egg, and that it may depend
in some way on the relative amount of water in the egg at the
time of fertilization. This suggestion, if capable of more general
application, does not exclude the possibility that in some cases
the spermatozoan may influence sex. It is known that the
unfertilized eggs of the bee produce only males, and that fertilized
eggs almost invariably develop into females. In this case the

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 233

entrance of the spermatozoan changes the sex, although, strictly
speaking, it cannot be said to determine sex, since the egg has
a sex mechanism and is capable of developing without fertiliza-
tion into a male. If the eggs of the vertebrates could develop
parthenogenetically it is very unlikely that the embryos would
be without sex. It seems probable, therefore, that in all eggs
there exists a sex mechanism that alone is capable of giving sex
to the resulting embryo ; but in many forms besides the bee the
spermatozoan may be a sex-changing factor if not a sex-deter-
mining one.

LITERATURE CITED.
Aristotle.

'62 History of Animals. Translated by R. Cresswell. Published by Henry

Bohn, London.
Backman, E. L., und Runnstrom, J.

'09 Physikalisch-chemische Faktoren bei der Embryonalentwicklung. Der
osmotische Druck bei der Entwicklung von Rana temporaria. Biochem.
Zeitschr., Bd. 22.
Beard, J*

'02 The Determination of Sex in Animal Development. Zool. Jahrbiicher, Bd.

16.
Bialaszewicz, K.

'08 Beitrage zur Kenntniss der Wachstumvorgange bei Amphibienembryonen.

Bull. Inter, de 1'Akad. Sci. de Cracovie. Math.-Natur. Classe.
Cuenot, L.

'99 Sur la determination du sexe chez les animaux. Bull. Sci. de la France et

de la Belgique, t. 32.
Dawson, E. R.

'09 The Causation of Sex. H. K. Lewis, London.
Fluorens, M.

'64 Remarque a 1'occasion d'une communication de M. Coste concernant des
observations faites sur les ceufs de poule d'une meme ponte. C. R. Acad.
Sci. Paris, t. 58.
Gurwitsch, A.

'95 Ueber die Einwirkung des Lithionchlorids auf die Entwicklung der Frosch-

und Kroteneier (R. fusca und Bufo vulgarise. Anat. Anz., Bd. n.
'96 Ueber die formative Wirkung des veranderten chemischen Mediums auf

die embryonale Entwicklung. Arch. Entwicklungsmech., Bd. 3.
Hertwig, O

'95 Beitrage zur experimentellen Morphologic und Entwicklungsgeschichte.
I., Die Entwicklung des Froscheies unter dem Einflus schwacherer und
starkerer Kochsaltzlosungen. Arch. mikr. Anat., Bd. 44.
Hertwig, R.

'06 Weitere Untersuchungen ueber das Sexualitatsproblem. Verhandl.

deutsch. zool. Gesellsch.

'07 Untersuchungen ueber das Sexualitatsproblem, III. Verhandl. deutsch.
zool. Gesellsch.

234 HELEN DEAN KING.

Jordan, H. E.

'09 Facts Concerning the Determination and Inheritance of Sex. Popular

Sci. Monthly.
King, Helen Dean.

'07 Food as a Factor in the Determination of Sex in Amphibians. Biol Bull.,

Vol. 13.

'09 Studies on Sex-determination in Amphibians, II. Biol. Bull., Vol. 16.
'10 Temperature as a Factor in the Determination of Sex in Amphibians.

Biol. Bull., Vol. 1 8.
Kuschakewitsch, S.

'10 Die Entwicklungsgeschichte der Keimdrtisen von Rana esculenta. Festschr.

fiir R. Hertwig, Jena.
Loeb, J.

'92 Experiments on Cleavage. Jour. Morph., Vol. 7.

'06 The Dynamics of Living Matter. The Macmillan Co., New York, N. Y.
Morgan, T. H.

'03 The Relation between Normal and Abnormal Development of the Embryo
of the Frog, as Determined by the Effects of Lithium Chloride in Solution.
Arch. Entwicklungsmech., Bd. 16.

'06 Experiments with Frogs' Eggs. Biol. Bull., Vol. n.
'07 Experimental Zoology. The Macmillan Co., New York, N. Y.
'08 The Determination of Sex in Frogs. Amer. Nat., Vol. 42.
'09 A Biological and Cytological Study of Sex Determination in Phylloxerans

and Aphids. Jour. Exp. Zool., Vol. 7.
'10 Chromosomes and Heredity. Amer. Nat., Vol. 44.
Morgan, T. H., and Tsuda, Ume.

'93 The Orientation of the Frog's Egg. Quart. Jour. Micr. Sci., Vol. 35.
Overton, E.

'02 Beitrage zur allgemeinen Muskel- und Nervenphysiologie. Arch, gesammte

Phys., Bd. 92.
Pfliiger, E.

'82 LTeber die das Geschlecht bestimmenden Ursachen und die Geschlechts-

verhaltnisse der Frosche. Arch, gesammte Phys., Bd. 29.
Rauber, A.

'oo Der Ueberschluss an Knabengeburten und seine biologische Bedeutung.

Leipzig, 1900.
Russo, A.

'09 Studien iiber die Bestimmung des weiblichen Geschlechtes. Jena.
Schultze, O.

'03 Zur Frage von den Geschlechtsbildenden Ursachen. Arch. mikr. Anat.,

Bd. 43-
von Seligson, E.

'953 Willkiirliche Zeugung von Knaben oder Madchen. Miinchen.

'95b Zur Bestimmung und Entstehung des Geschlechts. Centralb. Gyna-

kologie, Bd. 19.
Shull, A. Franklin.

'ioa Studies in the Life Cycle of Hydatina senta. I., Artificial Control of
the Transition from the Parthenogenetic to the Sexual Method of Re-
production. Jour. Exp. Zool., Vol. 8.

STUDIES ON SEX-DETERMINATION IN AMPHIBIANS. 235

Shull, A. Franklin,
'rob The Artificial Production of the Parthenogenetic and Sexual Phases of

the Life Cycle of Hydatina senta. Amer. Nat., Vol. 44.
Stevens, N. M.

'05 Studies in Spermatogenesis with Especial Reference to the "Accessory

Chromosome." Publ. Carnegie Inst., Washington.
Stockard, C. R.

'09 The Influence of Alcohol and other Anaesthetics on Developing Embryos.

Proc. Soc. Exper. Biol. and Med., Vol. 7.

'10 The Influence of Alcohol and other Anaesthetics on Embryonic Develop-
ment. Amer. Jour. Anat., Vol. 10.
Tennent, D. H.

'10 The Dominance of Maternal or of Paternal Characters in Echinoderm

Hybrids. Arch. Entwicklungsmech., Bd. 29.
Tower, W. L.

'10 The Determinance of Dominance and the Modification of Behavior in
Alternative (Mendelian) Inheritance, by Conditions Surrounding or In-
cident upon the Germ Cells at Fertilization. Biol. Bull., Vol. 18.
Whitney, D. D.

'10 The Influence of External Conditions upon the Life Cycle of Hydatina

senta. Science, Vol. 32.
Wilson, E. B.

'06 Studies on Chromosomes. II., The Sexual Differences of the Chromosome
Groups in Hemiptera, with Some Considerations of the Determination
and Inheritance of Sex. Jour. Exp. Zool., Vol. 3.

Vol. XX. April, 1911. No. 5.

BIOLOGICAL BULLETIN

EXPERIMENTS WITH CHRYSOMELID BEETLES.

III. THE EFFECTS OF KILLING PARTS OF THE EGGS OF
Leptinotarsa decemlineata.1

ROBERT W. HEGNER.

CONTENTS.

1. Introduction 237

2. Killing the germ cell determinants 240

3. Killing the primordial germ cells 243

4. Killing parts of freshly laid eggs 244

5. Killing parts of eggs in the blastoderm stage 245

6. Killing parts of young embryos 247

7. Killing parts of old embryos 250

8. Summary 250

i. INTRODUCTION.

The method used in the experiments described in this paper,
i. e., killing parts of the egg, has been employed by a number of
investigators in many ways and for many purposes. Heat and
electricity are the agents which have been most frequently
applied. So far as I know insects' eggs have heretofore never
been operated upon in this way. The results obtained by the
use of these agents are quite similar to those so frequently brought
about by removing parts of the egg or embryo, or by isolating
blastomeres. The latter process is of course impossible in the
case of the insect's egg, since cleavage is superficial, but material
has been removed successfully .from different parts of beetles' eggs
in various stages of development.2

Three years ago a preliminary report was made of experiments
in removing portions of the eggs of Calligrapha multipunctata,

1 Contributions from the Zoological Laboratory of the University of Michigan,
No. 131. Parts I. and II. of "Experiments with Chrysomelid Beetles" appeared
in the BIOLOGICAL BULLETIN, Vol. XIX., June, 1910, pp. 18-30.

237

238

ROBERT W. HEGNER.

khbh

C. bigsbyana, C. hmata and Leptinotarsa decemlineata.1 Before
the results of those experiments and the data contained in the

present contribution can be
understood clearly, a brief
account of the structure of
freshly laid chrysomelid
eggs and of the principal
stages in their normal em-
bryonic development is
necessary.

The freshly laid egg (Fig.
i) consists of a large central
\— -gn. mass of yolk globules (y)
and a comparatively thin
superficial layer of cyto-
plasm, the " Keimhaut-
blastem " (khbl). Two en-
velopes cover the egg, the
vitelline membrane (vm)
and the chorion. Polar
bodies are usually present
at this time, and the egg
nucleus is in the act of
union with the sperm nu-
cleus (gn), or else cleavage
has already begun. Em-
bedded in the "Keimhaut-
blastem" at the posterior
end of the egg is a disc-
shaped mass of darkly
staining granules, which I
have called the pole-disc,
or germ cell determinants
(gcd}.

FIG. i. A longitu inal section through
an egg of Calligrapha c. tgsbyana four hours
after deposition. gcd, germ cell determi-
nants; gn, germ nuclei copulating; khbl,
" Keimhautblastem "; p, posterior; vm, vitel-
line membrane; y, yolk. The eggs of Lep-
tinotarsa decemlineata are not visibly different
from those of C. bigsbyana.

As cleavage progresses,
a separation of the cleavage products into sections occurs; the

iHegner, R. W., 'o8b, "The Effects of Removing the Germ Cell Determinants
from the Eggs of Some Chrysomelid Beetles," BIOL. BULL., Vol. 16

EXPERIMENTS WITH CHRYSOMELID BEETLES.

239

nuclei of one group form a more or less regular layer equidis-
tant from the periphery, and later migrate outward, fuse with
the " Keimhautblastem " and become the blastoderm (Fig. 2, bT).
The nuclei of the other group remain behind in the yolk (Fig. 2,
v), which it is their duty to dissolve. Eight of the cleavage
products which reach the posterior end do not help to form the
blastoderm, but gather the germ cell determinants about them
and continue their migration until they are entirely separated
from the egg; these are the primordial germ cells (Fig. 2, pgc).

Pgc

FIG. 2. A longitudinal section through an egg of Leptinotarsa decemlineata
one day after deposition when in the blastoderm stage, bl, blastoderm; pgc,
primordial germ cells; v, vitellophag; y, yolk.

FIG. 3. Superficial view of the right side of an egg of Leptinotarsa decemlineata
thirty-six hours after deposition, gb, germ band; pgc, primordial germ cells; pi,
procephalic lobes; s, stomodeum; vg, ventral groove.

FIG. 4. Surface view of right side of an egg of Leptinotarsa decemlineata forty-
eight hours after deposition, a, antenna; in, mandible; ?«', first maxilla; m2, second
maxilla; t'-l3, thoracic appendages; //, tail fold.

They remain quiescent for a considerable period and then
migrate into the embryo, which has developed in the meantime.
At the end of two days the germ band appears on the ventral
surface of the egg (Fig. 3, gb); this lengthens within the next
twenty-four hours, growing forward to the extreme anterior end,
and posteriorly until the tail-fold reaches over half way up on
the dorsal surface (Fig. 4, tf). During this elongation the embryo
segments and the appendages of the head and thorax grow out.
The entire embryo then contracts antero-posteriorly and begins
to grow around the yolk (Fig. 5). This contraction continues

240

ROBERT W. HEGNER.

until the end of the tail-fold coincides with the posterior end of
the egg (Fig. 6). The larva (Fig. 7) usually appears in five or
six days.1

2. KILLING THE GERM CELL DETERMINANTS.
The method used in the experiments mentioned above to
remove the germ cell determinants was to prick the posterior
end of the freshly laid egg with a needle and allow them to flow
out. These experiments were not considered entirely successful,
since I was unable to determine in any case whether all of the
germ cell determinants had been removed. Some of the embryos
which developed from eggs operated upon in this way produced

FIG. 5. Ventral view of an egg of Leptinotarsa decemlineata sixty hours after
deposition, ab, abdomen; h, head; t, thoracic appendages; tf, tail fold.

FIG. 6. Ventral view of an egg of Leptinotarsa decemlineata seventy-two hours
after deposition, ab, abdomen; h, head; t, thoracic appendages.

FIG. 7. Side view of a newly hatched larva of Leptinotarsa decemlineata.

either a lesser number of germ cells than normal, or else no germ
cells at all. The conclusion was reached that if all of the germ
cell determinants are removed from the egg no germ cells will be
produced, and that the granules which constitute the pole disc
aj;e really germ cell determinants.2

Two series of experiments (L.D. 09 and L.D. 018) were carried

iHegner, R. W., '080, "Observations on the Breeding Habits of Three Chry-
somelid Beetles, Calligrapha bigsbyana, C. multipunctata and C. lunata, Psyche
Vol. 15.

2 For a discussion of this subject see Hegner, R. W., 'n, "The Germ Cell De-
terminants in the Eggs of Chrysomelid Beetles," Science, Vol. XXXIII., pp. 71-72.

EXPERIMENTS WITH CHRYSOMELID BEETLES.

241

out in order to learn whether or not germ cells would appear in
embryos if the germ cell determinants were prevented from taking
part in their development. Certain data regarding these experi-
ments are listed in Tables I. and II. Eggs were oriented as
soon as laid and the central region of the posterior end of each
was touched with a hot needle, thus killing the protoplasm just
beneath in which the germ cell determinants were embedded.
Since it was impossible to see the germ cell determinants in the
living egg, a rather large area had to be killed in order to be certain
that all of them had been reached. As noted above, the experi-
ments in removing the germ cell determinants by pricking the
egg and allowing them to flow out were not entirely successful
because it was never possible to tell whether all of them had
been obtained.

TABLE I.

EXPERIMENTS IN KILLING THE GERM CELL DETERMINANTS.
Leptinotarsa decemlineata — Series L.D. 09.

Number of
Experiment.

Stage when
Operated Upon.

Nature of
Operation.

Interval between
Operation
and Fixation.

Remarks.

L.D. 09 A

Control.

Hatched in 6 days.

L.D. 09 Bi

**

O

4 cleavage nuclei present.

L.D. 09 62
L.D. 09 63
L.D. 09 64
L.D. 09 65
L.D. 09 B6

Immedi-
ately after
deposition
(see Fig. i).

Posterior
end killed
with
hot needle.

i day.
2 days.
3 days.
4 days.
5 days.

See Fig. 8.
See Fig. n.

L.D. 09 By

-

_

7 days.

TABLE II.

EXPERIMENTS IN KILLING THE GERM CELL DETERMINANTS.
Leptinotarsa decemlineata — Series L.D. 018.

Number of
Experiment.

Stage when
Operated Upon.

Nature of
Operation.

Interval between
Operation
and Fixation.

Remarks.

L.D. 018 A
L.D. 018 Bi
L.D. 018 62
L.D. 018 63
L.D. 018 64
L.D. 018 65
L.D. 018 B6

Control

Immedi-
ately after
deposition
(see Fig. i).

Posterior
[end killed
1 with
1 hot needle.

O

i day.
2^2 days.
3 days.
4 days.
6 days.

Hatched normally.
2 cleavage nuclei found.

See Fig. 9.

Many of the eggs operated upon in experiments L.D. 09 and
L.D. 018 did not develop at all or developed abnormally. Under

242

ROBERT W. HEGNER.

ordinary circumstances eggs sometimes become shapeless masses
of tissue or remain in the condition of the freshly laid egg, but
this fact does not explain the great number of cases observed
among the operated eggs. It is evident that many of the oper-
ated eggs failed to develop because of the conditions of the ex-
periments. For this reason only a few of the best defined em-
bryos have been selected for descriptive purposes.

b|

m

ab

ec

9

FIG. 8. Longitudinal secion through an egg of Leptinotarsa decemlineata
(L.D. 09 62). The posterior end was killed with a hot needle just after the egg
was laid (see Fig. i); the egg was then allowed to develop for twenty-four hours.
bl, blastoderm; gcd, germ cell determinants; k, portion of egg killed.

FIG. 9. Side view of an egg of Leptinotarsa decemlineata (L.D. 018 63). The
posterior end was killed with a hot needle just after deposition (see Fig. i); the
egg was then allowed to develop for sixty hours, ab, abdomen; h, head; I, thoracic
appendages; tf, tail fold.

FIG. 10. Longitudinal section through the tail fold of a normal embryo of
Leptinotarsa decemlineata sixty hours old, showing the germ cells (gc). ec, ecto-
derm; m, malpighian tubules.

Fig. 8 is from a longitudinal section of an egg (L.D. 09 62)
which was operated upon just after deposition and was then
allowed to develop for twenty-four hours. This should be com-
pared with the normal egg at a similar stage of development
(Fig. 2). In the egg shown in Fig. 8 the germ cell determinants
(gcd) and a portion of the neighboring yolk and cytoplasm (k}
have been prevented from taking part in development. That
part of the egg which remained alive produced a blastoderm of
a single layer of cells (bl). The principal difference to be noted

EXPERIMENTS WITH CHRYSOMELID BEETLES.

243

between this preparation and that of a normal embryo is the
absence of the primordial germ cells at the posterior end (see
Fig. 2, pgc). Three perfect series of sections were cut from eggs
at this stage, but none disclosed any germ cells.

Fig. 9 is a sketch of embryo L.D. 018 63 (see Table II.). It
was killed two and one half days after the operation and is a
stage a trifle younger than that shown in Fig. 5. The tail fold
(//), however, is not fully developed as in Fig. 5. Longitudinal
sections were cut through two embryos like that shown in Fig. 9,
but no germ cells could be found. Fig. 10 was drawn from a
section through the tail fold of a normal embryo two and one
half days old ; it indicates where the germ cells (gc) are situated
at this time. If germ cells had been present in the embryo shown
in Fig. 9, they would certainly have been found.

These experiments demonstrate that germ cells are not produced
when the extreme posterior end of the egg is prevented from
taking part in development, and it seems probable from the
method of origin of the germ cells that the destruction of the
germ cell determinants is the real cause of their absence.1

3. KILLING THE PRIMORDIAL GERM CELLS.
Table III. gives the data for the experiments performed in
series L.D. oil. Sixty-nine eggs were laid at n A.M. June 24,

TABLE III.

EXPERIMENTS IN KILLING THE PRIMORDIAL GERM CELLS.
Leptinotarsa decemlineala — Series L.D. on.

Number of
Experiment.

Stage when
Operated Upon.

Nature of
Operation.

Interval between
Operation
and Fixation.

Remarks.

L.D. oil A

Control.

Hatched normally.

L.D. on Bi

^

"

0

Like Fig. 2.

L.D. on 62
L.D. on 63

One day

Posterior
end

i day.
2 days.

L.D. on 64
L.D. on 65
L.D. on B6

deposition
(see Fig. 2).

> killed
with hot
needle.

3 days.
4 days.
5 days.

One hatched.

L.D. on B?

-

-

6 days.

Several hatching.

:The writer's papers on "The Origin and Early History of the Germ Cells in
some Chrysomelid Beetles" (Journ. Morph., Vol. 20) and on "Germ Cell Deter-
minants and Their Significance" now in press (American Naturalist) discuss these
points fully.

244 ROBERT W. HEGNER.

and were allowed to develop until n A.M. June 25. The pos-
terior end of sixty-five of them was then killed with a hot needle.
The four controls hatched on June 29. The eggs when operated
upon were in a stage like that shown in Fig. 2. One of the
operated eggs hatched on June 30 (L.D. on B6), several were
hatching on the following day, and a number of others were ready
to hatch at that time. The one that hatched and three of those
ready to hatch were examined and then sectioned and stained.
Superficially they resembled the larva shown in Fig. 7, the only
difference being the absence of the last two posterior segments
which are indicated by the letter x in Fig. 7; one possessed all
but the last segment. The sections showed that none of these
larvae contained germ cells.

It is evident from these experiments that the primordial germ
cells were killed by the operation and no new ones were produced
by the developing embryos. This is, I believe, the earliest
stage at which surgical castration has been performed among
the Insecta. The influence of this operation upon secondary
sexual characters could not be determined, since none of these
characters make their appearance in the larvae.

4. KILLING PARTS OF FRESHLY LAID EGGS.

The experiments described under heading 2, "Killing the Germ
Cell Determinants," might be included in this part of the paper,
but they were considered of sufficient importance to warrant a
special account. In performing the experiments in series L.D.
09 (Table I.) it was found impossible to regulate with any degree
of exactness the amount of the egg killed, and, since it was ab-
solutely necessary that all of the germ cell determinants be killed,
a larger portion of the egg was killed than desired. Some of these
eggs, however, when allowed to continue their development,
provided data with regard to the effects of killing a large part of
the posterior end of freshly laid eggs.

Fig. II was drawn from an egg from series L.D. 09 64. The
portion killed by the operation is labelled k; the material that
remained alive produced the head (/?) and part of the thorax (/)
of an embryo. These parts resemble the corresponding parts
of a normal embryo at a like age (see Fig. 6). Apparently that

EXPERIMENTS WITH CHRYSOMELID BEETLES.

245

part of the " Keimhautblastem " which remained alive after the
operation, became supplied with nuclei, was broken up into cells,
and proceeded to develop that particular part of the embryo
to which it would have given rise if the rest of the egg had not
been killed.

Two other series of experiments, L.D. 07 and L.D. 08, were
performed to test these results and in every case the living part

of the egg developed as though the entire
egg were intact. None of the tissue that
is normally produced by the killed portion
was regenerated by the living material.
The embryos developed up to the time of
hatching, but were unable to break out of
the chorion.

KILLING PARTS OF EGGS IN THE
BLASTODERM STAGE.

FIG. ii. Ventral
view of an egg of Lep-
tinotarsa decemlineata
(L.D. 09 64). The pos-
terior end (k) was killed
just after deposition (see
Fig. i); a normal head
(/z) and part of the tho-
rax (/) developed from
the material which re-
mained alive, y, yolk.

The eggs used for the experiments
designated as series L.D. 04 (Table IV.)
were laid at 4 P.M. on June 16, and
operated upon at 4 P.M. June 17. They
were, at the time of the operation, in a
stage similar to that of the egg shown in
Fig. 2. As indicated in Table IV., two
kinds of operations were performed; the
anterior part of one half of the eggs was
killed with a hot needle, and the posterior part of the other half
was killed in a like manner. Three preparations have been
selected to show the results of these experiments, (i) L.D.
04 A2, Fig. 12, (2) L.D. 04 AS, Fig. 13, and (3) L.D. 04 63,
Fig. 14.

The embryos shown in Figs. 12 and 13 developed from eggs
which had their posterior parts killed, and were fixed four days
and seven days later respectively. One of these embryos (Fig.
12) consists of a head and thorax which appear to be normal in
every respect, and are as fully developed as these parts in a
normal embryo at a similar age (five days). The abdomen of this

246

ROBERT W. HEGNER.

TABLE IV.

EXPERIMENTS IN KILLING PARTS OF EGGS IN THE BLASTODERM STAGE.
Leptinotarsa decemlineata — Series L.D. 04.

Number of
Experiment.

Stage when
Operated Upon.

Nature of
Operation.

Interval between
Operation
and Fixation.

Remarks.

L.D. 04 Ai

Control.

Hatched normally.

L.D. 04 A.2

-

I Posterior

4 days.

See Fig. 12.

L.D. 04 A3

Blastoderm

1 part killed.

7 days.

See Fig. 13.

L.D. 04 Bi
L.D. 04 B2
L.D. 04 B3

. stage
(see Fig. 2).

I Anterior
part

killpd

o
3 days.
4 days.

See Fig. 14.

L.D. 04 64

7 days.

embryo is entirely missing. The conclusion is reached that the
part of the blastoderm which would have produced the abdomen
of the embryo was killed in the operation, and that none of this
region was regenerated by the tissue which remained alive.

12 '*

FIG. 12. Ventral view of an egg of Leptinotarsa decemlineata five days old
(L.D. 04 A2). The posterior end of the egg (&) was killed when in the blastoderm
stage (see Fig. 2) ; the blastoderm which remained alive produced a normal head
(h) and thorax (/).

FIG. 13. Ventral view of an egg of Leptinotarsa decemlineata eight days old
(L.D. 04 A3), operated upon as in Fig. 12. Only a head (h) developed from the
tissue which remained alive.

FIG. 14. Side view of an egg of Leptinotarsa decemlineata five days old (L.D.
04 63). The anterior end of the egg (fe) was killed when in the blastoderm stage
(see Fig. 2) ; the blastoderm which remained alive produced a normal abdomen (ab)
and part of the thorax (0.

Fig. 13 was drawn from an egg that was fixed three days later
than that of Fig. 12, i. e., at the age of eight days. Only the
head of this embryo developed. It is apparent that not only

EXPERIMENTS WITH CHRYSOMELID BEETLES. 247

the tissue destined to produce the abdomen, but also that set
aside to form the thorax was killed by the operation. The region
of the blastoderm which normally develops into the head covers
a considerable area at the time the operation was performed.
This area lessens in extent when the germ band arises (Fig. 3, pi),
and, after the cephalic appendages appear (Fig, 4, a, m, m1, w2),
the anterior end of the embryo shortens until the mouth parts
are closely crowded together (Figs. 5, 6 and 7). In egg L.D. 04
A3 (Fig. 13) the shortening of the embryo has resulted in the
uncovering of a large yolk area (y), and, though a comparatively
small part of the egg was killed (&), this portion bore the blas-
toderm which in normal eggs gives rise to the larger part of the
embryo, i. e., the thorax and abdomen.

When the blastoderm surrounding the anterior end of the egg
is killed, only the posterior embryonic region develops from the
part which remains alive. This is shown in egg L.D. 04 63,
Fig. 14. Here the normal number of abdominal segments appear
as well as two thoracic segments (/),one of which has developed
a normal pair of legs.

6. KILLING PARTS OF YOUNG EMBRYOS.

The series of figures numbered 15 to 18 show what takes place
when parts of young embryos are killed and are thus prevented
from continuing development. Table V. gives the data of the
operations. The eggs, fifty-two in number, were laid at 10 A.M.
June 26; the operations were performed at 10 A.M. June 28, at
which time the eggs bore embryos similar to that shown in Fig. 4.
Four of the eggs were kept as controls; these hatched on July 2.
Approximately one half of the anterior end of twenty-four of
the eggs was killed with a hot needle; the posterior half of the
other twenty-four eggs was killed in like manner. In every
instance the part of the embryo which remained alive devel-
oped as though the egg had not been disturbed.

Figs. 15 and 16 show two stages in the development of the
posterior part of the embryo, and Figs. 17 and 18 show corre-
sponding stages in the development of the anterior part of the
embryo. It is interesting to note that in the egg shown in Fig. 15
not only the anterior end of the embryo, but also the extreme

248

ROBERT W. HEGNER.

TABLE V.

EXPERIMENTS IN KILLING PARTS OF YOUNG EMBRYOS.
Leptinotarsa decemlineata — Series L.D. 016.

Number of
Experiment.

Stage when
(Operated Upon.

Nature of
Operation.

Interval between
Operation
and Fixation.

Remarks.

L.D. 016 A

Control.

Hatched normally.

L.D. 016 Bi

™

I

O

See Fig. 4.

L.D. 016 62

Anterior

i day.

See Fig. 15.

L.D. 016 63

part

2 days.

L.D. 016 64

Young

killed.

3 days.

L.D. 016 BS

>• embryo

.1

4 days.

See Fig. 16.

L.D. 016 Ci
L.D. 016 C2

(see Fig. 4).

} Posterior

I day.
2 days.

See Fig. 17.
See Fig. 18.

L.D. 016 C3

part

L-i'Ilorl

3 days.

L.D. 016 €4

•

4 days.

end of the tail fold (//) was killed by the operation, and that the
intermediate region consisting of two thoracic segments and eight
abdominal segments continued to develop normally except for
the mechanical difficulties interposed by the killed material.

15

16

FIG. 15. Side view of an egg of Leptinotarsa decemlineata three days old.
(L.D. 016 62). The anterior end (k) was killed when the embryo had reached the
stage shown in Fig. 4; the posterior end continued to develop, ab, abdomen; h,
head; t, thoracic appendages; //, tail fold.

FIG. 16. As in Fig. 15, six days old (L.D. 016 BS). e, enteron; hi, hind in-
testine.

An in toto preparation of an older embryo from this material
(Fig. 1 6) shows that the living part of the embryo succeeded in
growing around the yolk and developing a hind intestine (hi)
which grew forward toward the yolk-filled enteron (e).

EXPERIMENTS WITH CHRYSOMELID BEETLES.

249

The preparation shown in Fig. 17 is from an egg fixed one day
after the posterior end had been killed, and is the same age as
that of Fig. 15, i. e., three days old. It is of special interest,
since the end of the tail fold (//), which was not killed by the
operation, has continued to develop, although it is an extremely
small piece of tissue and was separated from the rest of the living
embryo by a considerable amount of yolk. The anterior part
of the embryo consisting of the cephalic region and the first
thoracic segment, developed normally. During the twenty-four
hours between the operation and fixation, the living part of the

ab

17

FIG. 17. Side view of an egg of Lepiinotarsa decemlineala three days old (L.D.
016 Ci). The posterior end (k) was killed when the embryo had reached the stage
shown in Fig. 4; the anterior end continued to develop and has come to lie on the
right side of the egg. ab, abdomen; h, head; t, thoracic appendages; tf, tail fold;
y, yolk.

FIG. 18. As in Fig. 17 four days old (L.D. 016 €2).

embryo contracted and left a large yolk space (3;) between it and
the killed material (k). A similar condition was noted above
in series L.D. 04 A3, Fig. 13 (y).

Fig. 1 8 represents an embryo (L.D. 016 C2) which was allowed
to live one day longer than that just described. Here the head
and first thoracic segment have continued to develop reaching
a stage similar to that shown in Fig. 6. This part of the embryo
has changed its orientation since the operation and now lies
on the right side of the egg instead of on the ventral surface.
Several other cases like this were observed in series L.D. 016
and in a number of the embryos from other series of experiments.

250

ROBERT W. HEGNER.

The last egg selected from this series was fixed three days after
the operation at an age of five days. It indicates that develop-
ment of the living part of the embryo proceeds up to the time of
hatching.

7. KILLING PARTS OF AN OLD EMBRYO.

The eggs used for these experiments were laid at 4 P.M.
June 17, and operated upon at 4 P.M. June 20, at the age of
three days. Part of them were kept as controls; the rest were
divided into two lots and operated upon as indicated in Table VI.
The control eggs hatched on June 22.

TABLE VI.

EXPERIMENTS IN KILLING PARTS OF OLD EMBRYOS.
Lepiinolarsa decemlineata — Series L.D. 06.

Number of
Experiment.

Stage when
Operated Upon.

Nature of
Operation.

Interval between
Operation
and Fixation.

Remarks.

L.D. 06 A

Control

Hatched normally.

L.D. 06 Bi

] Posterior

O

See Fig. 6.

L.D. 06 B2

! end

i day.

L.D. 06 63

Old

1 killed.

2 days.

L.D. 06 64

^embryo

J

4 days.

L.D. 06 Ci

(see Fig. 6).

(Anterior

i day.

L.D. 06 C2

end

2 days.

L.D. 06 Cs

.

killed.

4 days.

Fig. 6 shows a normal embryo fixed at the time of the operation.
The eggs under experimentation were fixed at intervals and
stained and mounted. In every case the part of the embryo
that remained alive continued to develop. This was true for
both those with the anterior part and those with the posterior
part killed. There were no signs of regeneration even after the
normal embryonic period had passed. The killed part of the
embryo began to disintegrate immediately after the operation.

8. SUMMARY.

i. If the region of a freshly laid egg of Leptinotarsa decem-
lineata, which contains the germ cell determinants (Fig. I, gcd),
is killed with a hot needle and these granules are thus prevented
from taking part in embryonic development, the embryo pro-
duced by the rest of the egg lacks the germ cells. This supple-

EXPERIMENTS WITH CHRYSOMELID BEETLES. 25!

ments former experiments in removing the germ cell determi-
nants, and indicates that these granules really determine the germ
cells.

2. When the primordial germ cells of Leptinotarsa decemlineata
are killed in the blastoderm stage (Fig. 2, pgc) the resulting
embryos lack germ cells. This is the earliest known stage in
which surgical castration has been performed among the Insecta.

3. When the anterior or posterior parts of freshly laid eggs
(Fig. i) are killed, the material remaining alive develops that
part of the embryo which it would have produced if the eggs
had remained intact (Fig. 1 1) . No regeneration of the part which
would have been produced by the killed region takes place.

4. If the anterior or posterior parts of eggs in the blastoderm
stage (Fig. 2) are killed, the resulting tissue represents the parts
of the embryos which would have been produced by the living
material if the entire egg had been allowed to develop (Figs. 8
and 9).

5. When parts of young embryos (Fig. 4) are killed, the re-
maining tissue develops normally (Figs. 15-18). Even small
pieces of tissue (Fig. 17, //), which are widely separated from the
rest of the embryo, continue to develop normally.

6. Parts of old embryos develop up to the time of hatching.
There is no regeneration of the killed part by the living tissue,

7. The eggs of Leptinotarsa decemlineata, at the time of depo-
sition (Fig. i), are definitely oriented with respect to the future
position of the embryo.1 The areas of the peripheral layer of
cytoplasm (Fig. I, khbl) are already set aside for the production
of particular parts of the embryo, and if these areas are killed,
the parts of the embryo to which they were destined to give rise
will not appear. Likewise areas of the blastoderm (Fig. 2, bl)
are destined to produce certain particular parts of the embryo.

THE UNIVERSITY OF MICHIGAN,
February 6, 1911.

1 Hegner, R. W., '09, "The Effects of Centrifugal Force upon the Eggs of Some
Chrysomelid Beetles." Journ. Exp. Zoo/. Vol. 6.

FURTHER EXPERIMENTS ON THE METHODS OF EGG-
LAYING IN AMPHITRITE.

JOHN W. SCOTT.

INTRODUCTION.

On excursions, taken in connection with the course offered
to students studying invertebrate zoology at Woods Hole, I
have been impressed with the large number of forms not used
for investigation. I believe this must be because little is known
of their life-history and habits, because many of them are fairly
abundant. With this idea in view, I recorded in a recent paper
('09) the results of some observations made upon the egg-laying
habits of one of the marine annelids. In the form studied,
Amphitrite, the germ cells arise in the typical way, i. e., from the
coelomic epithelium. Very early these dehisce into the body
cavity and mix with the coslomic corpuscles. In some annelids
as Nereis, the eggs escape by a rupture of the body wall. In
Amphitrite, however, they pass to the exterior through certain
nephridia that are highly modified to form gonaducts. In the
paper mentioned above it was shown that Amphitrite deposit
eggs at recurring periods that bear a close relation to the time
of spring tides. It was further shown that at the time of ovi-
position, the body cavity contains floating free, not only the
mature eggs and corpuscles but also the younger eggs in various
stages of development. A single period of egg-laying occupies
from 30-60 minutes, and most all eggs when extruded are in
the metaphase of the first polar spindle. A few large, though
unripe, eggs always escape especially toward the latter part
of the period. In Fig. I is shown the comparative size and shape
of various bodies found in the coelome at this time. It will be
noticed that the eggs when first set free in the coelome are smaller
than the red blood cells, yet neither cells nor eggs of this size
ever escape during oviposition. Indeed, comparatively few of
the large unripe eggs pass out through the nephridia. The

252

EGG-LAYING IN AMPHITRITE.

253

interesting question then arises, How are ripe eggs separated from
the other ccelomic bodies in the act of oviposition?

REFERENCES TO PREVIOUS PAPER.

At first an attempt was made to answer this question by
studying the general anatomy of Amphitrite. In this group, as
shown by Meyer, the septa are incomplete or absent, with one
exception. The one complete septum, called the diaphragm, is
near the anterior end and divides the ccelome into two unequal
cavities. Anterior to the diaphragm the nephridia are for ex-

t o fl

FIG. i. To show the comparative size and shape of various bodies found in
the coelome of Amphitrite at the time of egg-laying; a, red blood corpuscles; b,
youngest free egg observed; c, young egg, with little or no yolk; d, unripe egg, nearly
mature; e, mature egg. Both edge and side views shown.

cretion only. The posterior nephridia are modified as gonaducts,
and it is with these last that we are concerned. Their inner
openings are bordered by large folded, or fimbriated, membranes,
covered by strong cilia. Each opening connects with a large
vascular sac, also ciliated. From these sacs tubes pass to the
outer openings of the nephridia which are found in the species
studied on segments 6-10 inclusive. From the dissections I
came to the conclusion that there was no apparatus for sifting
or "straining out" the ripe eggs.

254 JOHN W. SCOTT.

However a theoretical explanation was given in which the
nephridia were regarded as "settling basins," the cilia in this
case preventing the settling of bodies which were not to be ex-
pelled. This explanation was based partly upon the structure
of the nephridia, and partly upon the fact that "when the con-
tents of the ccelome are stirred in sea-water, the largest ova
including the ripe eggs, always settle more quickly to the bottom
of the dish ; the immature eggs then settle and last of all the
ccelomic corpuscles." The important fact here is that gravity
more quickly influences the large ova than the other bodies
floating in the ccelomic fluid. It was suggested that the difference
in the rate of settling was probably due to a greater density of
the ripe eggs, but that the effect was possibly due to a difference
in the shape of the bodies concerned. In either case, the con-
clusion was drawn that gravity is a differential means of separa-
tion. However, this left the matter inconclusive and the question
arose, Is this tendency of large ova to settle quickly due to a greater
specific gravity, or to a greater mass in proportion to the amount of
surface offering resistance? To answer this was the primary object
in the next step of the investigation.

SPECIFIC GRAVITY EXPERIMENTS.

Having set the problem, I began the work of determining the
specific gravity of the eggs and corpuscles found in the fluid of
the body cavity. Incidentally the specific gravities of sperm,
and of eggs in some later stages of development were obtained.

Lyon's plan of getting the density of eggs by centrifuging in
gum arabic solution was tried and, after finding it suited my
purpose, his method was adopted with slight modifications.
First a strong gum arabic solution in sea-water was prepared and
its density carefully determined. From this stock solution a
series of stock solutions was made up of differing densities; this
can be readily done by mixing with the stock solution in proper
proportions sea-water of known density. When ready to cen-
trifuge the capillary tubes of a haematocrit were filled about
three fourths full of gum arabic solution. On top of these
solutions whose density was known was placed the material
whose density was required. In a few cases, in order to guard

EGG-LAYING IN AMPHITRITE. 255

against possible error caused by surface tension, the material
was placed between two solutions of differing densities and then
centrifuged. But surface tension did not appear to offer any
appreciable resistance to the passage of the eggs and corpuscles
either upward or downward. It was found that eighty turns of the
hgematocrit in one and one half minutes was sufficient to produce
thorough separation of materials in the capillary tubes, without
noticeably affecting the material within the egg membrane. This
amount of centrifuging was therefore adopted throughout the
entire series. I shall here present in order my findings in regard
to (i) the specific gravity of eggs, (2) the specific gravity of cor-
puscles, and (3) the specific gravity of sperm, together with
remarks as to the significance of these results.

After a number of preliminary experiments, in order to get
familiar with the method to be used and to guard against possible
sources of error, a series of tests was made upon eggs recently
deposited. The results of some of these tests are shown in
Table I. The female used in this series was found depositing

TABLE I.
To SHOW THE SPECIFIC GRAVITY OF RECENTLY DEPOSITED Amphitrite EGGS

Test
No

Density of
Solution

Material Used.

Result of Centrifuging.

Used.

25

I. OQO

Eggs just deposited.

Eggs on top.

26

1.085

Eggs just deposited.

Eggs on top.

27

I.oSo

Eggs just deposited.

A few begin to sink.

28

I.07S

Eggs just deposited.

All went down.

29

1-0775

Eggs just deposited.

More than one half sink.

30

1-075

Eggs just deposited.

All on bottom.

eggs in a normal manner. Just as quickly as it could be done,
while they were still being deposited, these eggs were centrifuged
with the results shown in tests 25 to 30 inclusive. The table
shows that the specific gravity of recently deposited, that is,
mature Amphitrite eggs, lies between 1.075 and 1.085. They
probably have a mean density very near 1.078; for almost all
eggs remain on top of, and therefore are lighter than a solution
with a specific gravity of 1.080, while more than half sink in a
solution with density at 1.0775. Tests were made on eggs from
other worms and verified the results given here.

256

JOHN W. SCOTT.

Another series of tests was made in order to demonstrate the
changes in density at different stages of development. See
Table II. It was shown that very young eggs, those that had

TABLE II.

To SHOW THE CHANGE IN DENSITY AT DIFFERENT STAGES IN THE DEVELOPMENT

OF EGGS.

Test
No.

Density of
Solution
Used.

Material Used.

Results of Centrifuging.

9

I.O7O

Ccelomic contents. Eggs

Youngest eggs on top.

half-grown and smaller.

12

I 075

Coelomic contents. Eggs

All eggs went down.

nearly mature.

13

1-075

Eggs in various stages.

Unripe eggs on top. Ripe

eggs down.

27

1.080

Mature eggs, just deposited.

Almost all on top, a few sink.

38

I.OSO

Same lot, after being in sea-

One half of eggs sink.

water \y^ hours.

41

I 080

Same lot, fertilized. 8-16-

One fifth of eggs sink.

celled stages.

47

1-075

Same lot, one hour later.

A few eggs sink. Most eggs

partly down.

57

1-075

Same lot, trochophores, age

One third trochophores sink

27 hours.

Best developed on top.

attained one fourth the size of mature eggs and in which there wras
not very much yolk laid down, had a specific gravity less than
1.070 (Fig. I, c). In a word, the density of the egg as a whole
is very noticeably increased as the yolk accumulates. It wras
also learned that allowing the deposited eggs to stand in sea-
water for some time slightly increases their density (test 38).
During segmentation and the formation of the segmentation
cavity, the density grew less as one might expect (tests 41, 47).
It probably continues to decrease slightly, at least as far as the
late trochophore stage (test 57). My experiments did not carry
the work further than this point.

Before the worm used in the first tests mentioned above, was
through depositing eggs, the surface of its body was wiped dry
and the ccelome was opened allowing the contents to escape into
a clean glass dish. An examination with the microscope showed
that the coelomic fluid contained corpuscles and eggs in various
stages of development, some being mature. Tests 32 to 37,
Table III., show typical results in regard to the specific gravity
of the corpuscles. Upon examining these results one finds that

EGG-LAYING IN AMPHITRITE.

257

TABLE III.
To SHOW THE SPECIFIC GRAVITY OF CCELOMIC CORPUSCLES.

The coelomic contents were removed after the worm was about through deposit-
ing eggs.

Test

No.

Density of

Solution
Used.

Material Used.

Result of Centrifuging.

32

1.080

Coelomic corpuscles.

Eggs in

Corpuscles sink. A few ripe

various stages.

eggs sink.

33

I.08S

Coelomic corpuscles.

Eggs in

All corpuscles sink. All eggs

various stages.

on top.

34

I.OQO

Coelomic corpuscles.

Eggs in

Seven eighths corpuscles

various stages.

down. Eggs on top.

35

I.OQ5

Ccelomic corpuscles.

Eggs in

Most corpuscles sink. Eggs

various stages.

on top.

36

I.I05

Coelomic corpuscles.

Eggs in

Two thirds corpuscles sink.

various stages.

Eggs on top.

37

I.I23

Ccelomic corpuscles.

Eggs in

One tenth corpuscles sink.

various stages.

46

I.I23

After standing three

hours in

Nearly one third corpuscles

sea-water.

sink.

the specific gravity of the ccelomic corpuscles varies between wide
limits. However, more than four fifths of them have a density
greater than 1.090 and less than 1.123. Contrary to what my
previous observations had led me to expect the corpuscles have
a density greater than the mature eggs. All corpuscles are
heavier and all eggs lighter than a density of 1.085. After stand-
ing in sea-water for some time the corpuscles appear to increase
in density as shown by test 46, and by other tests not given.
Furthermore, the smaller and what appear to be the younger
corpuscles have a density less than the older ones. Under
these circumstances an examination with the microscope after
centrifuging was of course necessary to distinguish between the
corpuscles. The significance of these results will be explained
later.

In the course of my experiments the ccelomic corpuscles of
both males and females were examined. While not suspected at
the time, upon looking over my notes the rather curious fact came
to light that the mean density of the female corpuscles is slightly
greater than that of male corpuscles. I do not wish to emphasize
this fact, for perhaps the number of individuals examined was
not sufficient on which to base conclusions. Nor yet do I see
whether the significance of these results pertains to nutrition

258

JOHN W. SCOTT.

alone or to differences in reproduction, should they prove gener-
ally true for other annelids.

Another series of experiments showed the extreme lightness of
the sperm; their specific gravity in sea-water being between
1.038 and 1.046. In contrast to eggs which increase in density
as they mature, the sperm masses decrease in density as they
grow larger and finally break up into free-swimming sperm. So
that the density of the ripest sperm is even less than 1.038, as
shown in Table IV. This density is interesting when considered

TABLE IV.
To SHOW THE SPECIFIC GRAVITY OF SPERM.

Test
No.

Density of

Solution

Used.

Material Used.

Result of Cemrifuging.

8

40
42
43
44
45
48
50

52
55
56

1-075

1-075
1.070
1.062
1.051
1.038
1.046
1.041

1.046

1.051

1.038

Contents of ccelome.

Sperm shed in sea-water.
Sperm shed in sea-water.
Sperm shed in sea-water.
Sperm shed in sea-water.
Sperm shed in sea-water.
Sperm shed in sea-water.
Ccelomic contents of worm

that deposited sperm on

previous day.
Coelomic contents of worm

that deposited sperm on

previous day.
Coelomic contents of worm

that deposited sperm on

previous day.
Coelomic contents of worm

that deposited sperm on

previous day.

All sperm on top. All cor-
puscles sink.

All sperm on top.

All sperm on top.

All sperm on top.

All sperm on top.

All sperm at bottom.

All sperm on top.

All unripe sperm sink; a few
nearly ripe on top.

Younger sperm masses sink.
Older sperm masses on top.

Younger sperm masses sink.

A few lightest, ripest sperm
on top.

in connection with fertilization and the habitat in which the
animal lives. Water currents are undoubtedly important factors
in the dissemination of sperm. But the sand flats on which
these animals are found are securely protected from violent cur-
rents, where reefs or eel grass or shoals of adjacent islands break
the force of the changing tide. The limited range of the animal
does not require a wide scattering of the sperm. In fact the
density of the sperm is such that, in the presence of very gentle
currents, their distribution must be very limited at the time of
oviposition. But the lightness of the sperm provides an easy

EGG-LAYING IN AMPHITRITE. 259

means for scattering them by currents over the sand flats, and
their own locomotion aids them in their distribution. They
have sufficient motor power to aid in horizontal distribution,
and have sufficient density to prevent them from rising much
above the bottom where the eggs are found. That sperm have
such a method of distribution was substantiated in a different
way by studying the scattering of sperm in dishes of still water.
A large quantity of sperm was introduced at some point in the
bottom of the dish; then the progress in distribution was ob-
served by noting the advancing cloudiness of the water, reckoned
in various directions from the point of departure. It was found
that the advance in a general horizontal direction was many
times more rapid than the advance in an upward direction. I
take this to be due to the resistance of gravity. Since the sperm
are set free near the bottom under normal conditions, their
lightness tends to bring about more favorable chances for fertili-
zation of the eggs.

We may now sum up the results of these specific gravity ex-
periments in the form of certain explanations and conclusions:

1. The eggs of Amphitrite increase in specific gravity during
growth in the ccelome. This is probably due to an increase in
the amount of yolk. The reason therefore that the larger mature
eggs settle in a dish of sea-water more rapidly than the smaller,
immature ones, is undoubtedly due in part to the greater specific
gravity of the older eggs. And in the process of egg-laying it is
probable that a difference in specific density may act as a means
to separate ripe from unripe eggs.

2. The reason why the larger eggs sink before the ccelomic
corpuscles must be explained in another way, for the specific
gravity of the corpuscles is decidedly greater than that of the
eggs. The difference in behavior between these bodies is to be
explained, I believe, by the flat, oblong shape of the corpuscles;
their shape is such that they offer in settling a much larger resist-
ance in proportion to their mass. Blood cells settle more slowly
than eggs in sea-water because of a difference in shape and not
because of a lesser density. Slight currents in the dish prevent
the corpuscles from settling for a long time, while they hardly
interfere with the downward movement of the eggs. At this

260 JOHN W. SCOTT.

time in my experiments it was thought probable that ciliary
action produces a similar effect in separating eggs and corpuscles
during oviposition. This will be discussed later.

3. While the egg increases in density up to the time of ovi-
position it decreases slightly after fertilization, the decrease
continuing at least as far as the trochophore stage.

4. The comparative lightness of the sperm is undoubtedly
useful as an adaptation for scattering them at the time of fertili-
zation. Their density is just sufficient to tend to keep them near
where the eggs are to be found, but not enough to prevent a
certain amount of locomotion.

DIRECT OBSERVATIONS.

During the past summer I was able to make a series of direct
observations upon the separating process. In these observations
I saw the ripe eggs pass over the fimbriated membrane, along
the grooves, and finally into the vascular, nephridial sac ready
to be expelled. At the same time the blood cells and unripe
eggs were rejected and thrown back into the ccelomic fluid.
The sight was indeed astonishing to see the general precision
of a comparatively simple process.

It was by means of living dissections that the observations
were made possible. First it is necessary to use only ripe females,
and to determine this fact with certainty one needed to wait
until the worm began to deposit eggs; then it was lifted out of
sea-water and the surface of the body quickly dried. Next the
ccelome was opened and its contents carefully drained into a
clean watch glass; this was kept covered to prevent unnecessary
evaporation while the nephridia were being removed. The
worm was then placed in a dry dissecting pan and pinned at
either end, so that it was held in a slightly stretched condition.
The body cavity was opened by making a cut through the
body wrall in the mid-dorsal line; this cut began a little back of
the middle of the body and extended forward to the anterior end.
The enteric canal and accompanying blood vessels were next
divided midway and their anterior portions removed. The
flaps of the body wall on either side were then pinned out. In
order to have these flaps spread out well, and so expose the large

EGG-LAYING IN AMPHITRITE. 26 1

nephridial sacs, it was necessary to cut the series of oblique
muscles which extend from the body wall above the notopodial
cirrus across the cavity to the body wall in the mid-ventral
region. This species has five pairs of nephridia that function
as oviducts. The ccelomic fluid of a ripe female contains some
ripe eggs, many unripe ones, and thousands of corpuscles, mostly
red blood cells. Each nephridium to be observed was removed
entire and placed at once in the watch glass with the ccelomic
fluid. Then the watch glass was placed on the stage of a micro-
scope and the process studied.

As mentioned before the fimbriated membrane is a folded, or
grooved structure covered for the most part with strongly de-
veloped cilia. The stroke of the cilia varies in direction and force
on different J^arts of the membrane and their action furnishes
the motor power that separates the eggs. By using a needle
I could push near the fimbriated membrane any of the bodies
I wished to study. The cilia are in continuous action, and their
effective stroke, so far as I could observe, has a constant general
direction. But the direction of the stroke varies on different
portions of the membrane as follows: (i) On the edge of the
membrane (Fig. 2, a) and near the entrance to the grooves
between the folds, the stroke is such as to send currents toward
or into the grooves. (2) In the grooves (Fig. 2, 6), where the
folds approach somewhat nearer to each other, the stroke at
the bottom of the groove is still onward and inward ; well up on
the sides of the grooves the currents move upward and outward,
and apparently with more force than at the bottom of the groove.
(3) In the main groove (Fig. 2, c] the cilia at the bottom are
arranged in approximately parallel lines that resemble small
grooves, all leading inward and their action comparatively slow.
Above on the sides of the pillars (Fig. 2, d), especially in the
narrow grooves between pillars, the currents move rapidly and
are directed upward and outward. With this explanation of
the ciliary currents, both in regard to position and action, we
may now consider how the different bodies are separated.

The Separation of Blood Cells from Eggs. — This occurs under
the conditions of my experiment principally on the outer edges
of the membrane. The currents are strong enough to lift and

262

JOHN W. SCOTT.

draw in toward the membrane blood cells from a considerable
distance; these cells therefore hit the cilia with some force, but
the major stroke of the cilia is strong enough to bounce them off
again, usually driving them away five to ten times their own
diameter, and frequently beyond the limits of further attraction.
However they may bounce two, three, or four times, that is,

FIG. 2. To show a portion of the fimbriated membrane found at the inner end
of a post-diaphragmatic nephridium of Amphitrite. Not so much magnified as
Fig. i. The membrane is undulating at a, and folding to form a groove at b; at c
is the deep main groove leading into the nephridial sac; at d, are pillar-like projec-
tions overhanging the main groove. For further explanation see text.

until they reach the region of the groove shown at b, Fig. 2 ; then
the cilia on the sides carry them up, out, and away. Only in
very rare instances do blood cells ever reach the main groove;
sometimes they follow the eddy near a mature egg, but they are

EGG-LAYING IN AMPHITRITE. 263

then thrown across the groove and against the pillars where
the currents hurriedly lift them up and drive them away. I did
not see a blood corpuscle pass down the main groove into the
nephridial pouches, so long as the cilia retained their normal
activity. After an hour or so their action becomes slower, but
this is probably due chiefly to the fact that much plasma has
evaporated .

The Separation of Immature from Mature Eggs. — The small,
immature eggs, Fig. I, b and c, as well as some of a larger size
are eliminated in much the same way as blood cells. The task
grows more difficult, however, as the eggs increase in size for
they become too bulky to be bounced off the edge of the mem-
brane like the blood cells. So, as the larger immature eggs (Fig. I ,
d) are carried into the narrowing grooves, they are rolled up on
edge, then lifted out and moved away. For when the sides of
the groove narrow down so that both sides of the flat egg are
affected by the currents produced by cilia on the opposite sides
of the groove, the combined action of these currents is sufficient
to lift even the weight of a large, immature egg. The case
is different with the mature eggs, which differ in shape, are more
plastic, and are somewhat heavier than most immature eggs found
in the ccelome. These eggs are carried into the grooves down
which they slowly pass, their shape and weight evidently being
such that the lateral cilia are unable to lift them out of the groove.
Undoubtedly their plasticity is also an important factor in
resisting the action of the cilia. The side of the egg gives before
the effective stroke of the cilia, so the result is to change the
shape of the ripe egg rather than to push it away, and conse-
quently such eggs move more slowly along the grooves. In
some cases it was observed that the folds of membrane along the
sides of the groove would respond to the presence of a ripe egg
by folding over, virtually forming a closed tube. Inside this
tube-like groove the egg could occasionally be seen changing
shape, as it was forced slowly along toward the nephridial sac.
When the membrane folded it was impossible for unripe eggs,
or corpuscles on the membrane, to pass into the nephridial pouch
with the ripe egg. The behavior of the membrane in the pres-
ence of a ripe egg had all the appearance of a tactile response.

264 JOHN W. SCOTT.

and if any chemical reaction was here involved it could not be
detected with the eye. The separation of these bodies appeared
to be a purely physical process.

Since "the cilia are continuously active, independent of any
egg-laying period, it would seem that the egg-laying reaction
is a direct response to changes in the egg that are chiefly pro-
duced by the breaking down of the germinal vesicle. As noted
in an earlier paper a worm shows unusual muscular activity at
the time of oviposition. This muscular reaction may be due to
chemical changes in the body fluid, the irritation being produced
by materials escaping from the nucleus. However another ex-
planation is possible. As the ripened eggs begin to accumulate
in the nephridial sacs, excretion, the normal function of these
organs, would be hindered or stopped. This would lead to an
accumulation of waste in the part of the coelome posterior to
the diaphragm, and in all probability would act as an irritant
producing more vigorous muscular action until the eggs were
expelled. In this way the wave-like movements of the worm at
oviposition are a physiological response to an interference with
excretion. Though these suggestions are largely theoretical,
I am inclined to believe the latter view is the correct explanation.

SUMMARY AND CONCLUSIONS.

Amphitrite under normal conditions keeps up, with occasional
brief resting periods, a series of wave-like contractions of the body
for the purpose of sending water through its tube. A short time
before oviposition the wave-like movements become stronger and
slightly faster than usual. This excitement indicates the pres-
ence of ripe eggs and marks the beginning of the separating
process. After some time has elapsed, and the nephridial sacs
are well rilled, each contraction wave as it approaches the an-
terior end of the body forces out through the pores small slimy
streams of eggs. Oviposition usually continues from thirty to
sixty minutes. The process of separation has been previously
described. While the cilia furnish the motor power in separating
the ripe eggs, strong wave-like contractions of the body wall
furnish the pressure needed to expel the eggs from the nephridial
pouches. It is suggested that this unusual activity of the worm

EGG-LAYING IN AMPHITRITE. 265

is probably due to interference with excretion, caused by clogging
the nephridial sacs with eggs.

Undoubtedly much the same process occurs in other worms
where eggs accumulate in nephridial sacs before oviposition.
I am told by Downing that eggs accumulate before oviposition
in the nephridial pouches of one species of Arenicola. Gerould
has reported that the eggs of Phascolosoma collect in this way
for several hours before they are laid, and there is little doubt
that the same general method of egg-laying occurs in all worms
with this habit. My results may be summarized in the following
conclusions.

1. The separation of ccelomic corpuscles and unripe eggs, at
the time of oviposition in Amphitrite, is accomplished by physical
not chemical means.

2. The work of cilia on the fimbriated membrane furnishes
the power used in the separating process. Wave-like contrac-
tions of the body aid in expelling the eggs, and it is suggested
that these movements are due to interference with excretion,
caused by clogging the nephridial sacs with ripe eggs.

3. The separating process is aided in several ways. The
shape and arrangement of grooves on the fimbriated membrane,
the size and shape of the bodies separated, and especially the
greater plasticity and greater density of the mature eggs, all
appear to be important factors in separating the bodies found in
the ccelome at the time of oviposition. The action of the grooves
in closing over a ripe egg is also a direct help in keeping other
bodies from the main groove.

4. We are fairly safe in concluding that the method of egg-
laying described for Amphitrite holds good for other worms where
the eggs accumulate in nephridial sacs before or during oviposition.

NEMATOCYSTS OF MICROSTOMA.1

WILLIAM A. KEPNER.

For the past three autumns many brown Hydras and Micro-
stomas were found in a fish pond northeast of the University of
Virginia. In both these forms of animals were present oval,
refractive bodies, which were the nematocysts. These nema-
tocysts when everted have in both cases slender filaments with
three barbules radiating from their base. The filament at its
base is attached to the neck of a pear-shaped "poison-sack"
or capsule. Since Oersted (1844) nematocysts were known to
occur in flat worms as well as in Ccelenterata. Associated with
the nematocyst of Hydra there is a cell which has elaborated the
nematocyst and cares for it until it is discharged. This cell is
known as a cnidoblast or nematocyte. Each cnidoblast of Hydra
has a small spine-like structure — the cnidocil. This is supposed
to receive the stimulus that results in the discharge of the nema-
tocyst. When one examines the living Hydra and the living
Microstoma found in this vicinity no difference can be detected
between the nematocysts of the two forms, except that there is
no cnidocil associated with the nematocysts of the flatworm.
Moreover the nematocysts in both Hydra and Microstoma when
undischarged and when being discharged appear to be normal
parts of the respective animals. This indigenous appearance and
ready discharge of the nematocysts of Microstoma led men to
look upon them as structures elaborated by the cells of Micro-
stoma. Such was my view after I had repeatedly, during three
years, studied the nematocysts of living Microstoma. According
to Martin (1908) until 1903 all zoologists held that the nemato-
cysts of Microstoma were its own products.

Recently the inference has been made that the nematocysts
of Microstoma have been derived from ingested ccelenterata,

1 The author wishes to thank Professor A. H. Tuttle, of this laboratory, Professor
E. G. Conklin and Professor Ulric Dahlgren, of Princeton University, for valuable
suggestions that were helpful in the preparation of this paper. This paper was read
before the Philosophical Society of the University of Virginia.

266

NEMATOCYSTS OF MICROSTOMA. 267

as Grosvenor (1903) had claimed for the nematocysts of aeolids.
Martin (1908) must be given the credit for having first ap-
proached this question of the origin of the nematocysts of
Microstoma in the proper way, i. e., experimentally.

Martin first made the observation that Microstoma do ingest
Hydras. "If a fasting Microstoma is placed in a watch-glass
which contains some small Hydra, it is almost certain in a short
time to come in contact with one of them. If the Microstoma
comes suddenly against the tentacles of the Hydra it contracts
itself immediately, and in this condition it may frequently be
killed by the discharged nematocysts. As a rule, however, the
Microstoma fixes itself for a short time by its posterior end in the
neighborhood of the Hydra, and everts its pharynx to its full
extent.

'The Microstoma then swims over the surface of the Hydra,
usually attacking the lower part of its body with its pharynx
fully everted (vide Fig. 10). The Hydra then usually becomes
strongly contracted, and sweeps its tentacles over to the side on
which it has been attacked, though under these conditions the
tentacles do not grasp the Microstoma, but remain extended
almost parallel with its body, and it would appear as though the
pharyngeal secretion had a paralyzing action on the Hydra.
In many cases, after a time, the Microstoma leaves its prey, and
in such a case the Hydra does not seem much the worse for the
attack, but if the Hydra is of small size, it may be engulfed and
swallowed whole" (Martin, 1908, pp. 267-8).

After this observation Martin's chief evidence is based upon
the histology of Microstoma. He finds that when both Hydra
and Cordylophora are fed to Microstoma there are found within
the tissues of the flatworm nematocysts peculiar to both kinds of
coelenterata that had been fed. Thus in sections of the Rhabdoc<zle
he discovers nematocysts in the endoderm, mesoderm, and ecto-
derm. Martin's methods of procedure had convinced me that
he had established that the nematocysts of Microstoma were
derived from Hydra, until I had considered the details of his work.
I hesitate, for example, to follow Martin when he says that
von Graff (1882) was wrong in saying that each nematocyst lies
in a single cell and that "Hallez had correctly described the

268 WILLIAM A. KEENER.

nematocysts as lying in vacuoles" (Martin, 1908, p. 263). It had
been the experience of students in my classes to find discharged
nematocysts enclosed within a single cell (Fig. i). Such isolated
nematocysts were obtained by adding a drop of i per cent,
acetic acid to a slide containing specimens of living flatworms.
Some time after a cell of this kind had lain in the dilute acetic
acid a large vesicle appears at the pole from which the discharged
nematocyst must emerge (Fig. 2). Such distortion has fre-
quently been observed under similar conditions.

So much for my observations upon the conduct and anatomy
of living specimens and upon temporary preparations.

Before taking up the evidence afforded by permanent histologi-
cal preparations I must interpolate my hesitancy, at least, to
follow Martin when he says, "I have found very young Micro-
stoma in which nematocysts were already present, and at first
this presented a difficulty to this theory of the derivation of
nematocysts, but I do not believe this is insuperable when we
consider that nematocysts in the case of an animal which has
fed largely on Hydra can be found in almost any tissue of he
body. I have found them in the testes (Fig. 8), and although I
have not yet found them in an ovum, I believe that the yolk-
cells might readily carry them into the cocoon thus causing the
infection of the young forms" (Martin, 1908, p. 271). It does
not appeal to my judgment that nematocysts with cnidoblasts
should be able to "infect" the young forms through the yolk-cells.
Nematocysts within vacuoles, without attending cells to propo-
gate them, infecting young forms through yolk-cells seem to be
altogether improbable.

When I consider this last statement of Martin and note his
oversight of the cells that in Microstoma do enclose some of the
nematocysts I am led to believe that the final word has not
been spoken concerning the nematocysts of Microstoma.

Such was my attitude when during September, October,
November and early December, 1910, collections of Microstoma
were made with a view to studying their histology.

Many specimens were fixed in chrom-aceto-formal; others
were fixed in saturated solution of corrosive sublimate to which
5 per cent, of glacial acetic acid had been added. Both of

NEMATOCYSTS OF MICROSTOMA. 269

these fixing fluids gave good results. The sections were mounted
serially. Some individuals were cut 3 microns, others at 4, 5, 6
and 7 microns. All the material was stained with Haidenhain's
iron heematoxylin. Some sections were counter stained with
Bordeaux red. From these many individuals a large list of
sectioned nematocysts has been made.

As the specimens were being collected it was not an unusual
experience to find greatly gorged specimens, which upon the
addition of the fixing fluid would burst and thus liberate or expose
a dead Ch<ztogaster such as is represented in Fig. 3. In passing,
this observation is of interest, for Martin says "one of the com-
monest enemies of Microstoma appears to be Chcetogaster, which
devours it greedily" (Martin, 1908, p. 268).

Within the enteron of the sectioned Microstoma, one frequently
finds the large unbroken setae of these ingested oligochaetes (Fig.
4). If according to Martin, as I interpret him, the passage
of the nematocysts from the enteron through endoderm and
mesoderm to ectoderm is passive, I wonder why at times these
harsh setae of Chcetogaster do not passively find their way into
the tissues of the body of the flatworm. Other solid bodies
such as diatom shells, and round or spheroidal objects such as
the shells of Arcella lie within the enteron. These too, if the
migration of the nematocysts be passive, should find their way
to the surface as do the nematocysts; but not a single instance
of such objects lying outside the lumen of the enteron was
observed.

If, therefore, nematocysts do migrate from the cavity of the
enteron to the general surface of the body, their migration seems
to be accomplished through some active agency.

The first tissue of course to be concerned in this selective
function is the endoderm. The cells of the endoderm must dis-
tinguish between the nematocysts and other solid or rounded
objects within the enteron, just as an Amceba distinguishes be-
tween food and non-food. So far as the present series of ob-
servations is concerned twenty endodermal cells were found
in which a nematocyst was enclosed (Fig. 5). In each case the
nematocyst lies within a vacuole near the base of the endodermal
cell.

270 WILLIAM A. KEPNER.

The question at once arises, might not this nematocyst have
wandered from the mesoderm into the endoderm? There is
ample evidence of nematocysts migrating from one region to
another in the bodies of coelenterata. Schneider says: "Alle
Nesselzellen...der Siphonophoren enstehen an localisierten Bil-
dungsherden, von denen sie in einen bestimmen Entwickelungs-
tadium als Wanderzellen auf die Verbrauchstatten uberwandern"
(Boulenger, 1910, p. 764). In this connection "Hadzi's results
are of the greatest interest, as he was able to examine living
tissue as well as preserved material. His main conclusions are
as follows:

"i. The thread-cells of hydroids are not formed 'in situ'
but in the ectoderm of the ccenosarcal branches, where, on ac-
count of the thick perisarcal investment, they can obviously
not become functional.

"2. When completely developed, except for accessory struc-
tures such as the cnidocils and the stalks, they migrate to the
important nematocyst batteries on the tentacles. This migra-
tion can take place in two different manners. In simple forms,
e. g., Campannlaria, the thread cells move actively by means of
their pseudopodia, making their way between the ectodermal
cells of the colony. In Tubularia, however, they adopt a quite
different method of locomotion : from the ectoderm of the ccenos-
arc they force a way through structureless lamella and endoderm
into the cavity of the hollow stem, whence they are carried by
the current caused by the flagella of the endoderm cells to the
hydranths. Here the thread-cells re-enter the tissues and
migrate actively by their own movements to the ectoderm of the
tentacles" (Boulenger, 1910, p. 764). Conklin (1908) likewise
observed the formation of nematocysts \vithin the mesoderm of
Actinia and their subsequent migration to the ectoderm. Bou-
lenger states "that in Mcerisia the nematocysts of the oral
battery of the medusa are developed in the endoderm at the base
of the manubrium ; this does not necessarily imply that the nemato-
blasts are themselves of endodermal origin" (Boulenger, 1910,
p. 767). Thus we have much testimony as to the migration of
nematocysts within the tissues of coelenterata.

It is to be noted, however, that in all these cited examples the

NEMATOCYSTS OF MICROSTOMA. 2"Jl

nematocysts are transported while enclosed by cnidoblasts. In
the twenty nematocysts found within the endoderm my material
shows the stinging-threads lying free within vacuoles at the
bases of the endodermal cells. Martin also describes the nemato-
cysts that he noted within the endoderm of Microstoma as being
free from nematocytes or cnidoblasts. In these cases, therefore,
either the cnidoblasts after having carried the nematocysts
into the endoderm have died or the stinging-cells were not taken
into the endodermal tissue by cnidoblasts. Judging from what
has been seen I am of the opinion that these nematocysts have
found their way into the endodermal cells through the selective
action of the latter. Such selective faculties have also been
observed for the cnidophages of the ceras of aeolids: Foreign
bodies within the lumen of the cnidophore of aeolids have been
found "by Heckt, by Hancock and Embleton, and by Grosvenor.
I have found uninterpretable fragments of tissue, bodies not
tissue-like, and diatoms, in the liver diverticula and in the cnido-
phores. The ciliated canal has no power to distinguish other
indigestible bodies from nematocysts, but this inability to select
is not shared by the cnidophages, for so far as I know, these
ingest only nematocysts" (Boulenger, 1910, pp. 127-8).

The cnidophages of seolids deliver their nematocysts to the
cnidocyst, whereas the endodermal cells of Microstoma deliver
their nematocysts to the mesoderm.

In the material of this laboratory numerous nematocysts are
found within the mesoderm. The nematocysts in the mesoderm
that lie nearest the endoderm are in all cases enclosed in a vacuole.
These vacuoles are similar to those described and figured by
Martin and Vallez. About each vacuole the mesoderm crowds.
Two kinds of cells of the mesoderm are to be found in or near the
walls of these vacuoles : (a) a branched type of cells, which form
the frame-work of the mesoderm. These cells have nuclei
whose chromatin granules are conspicuous but more or less equal
in size. The cytoplasm of this first type of cell is not very dense
and is greatly branched, its branches anastomosing with those of
its fellows (Fig. 6, A) ; (b) cells which are much less frequent in
their occurrence, with denser, more compact cytoplasm than that
of the first type of cell, the nuclei of the second kind of cells being

272 WILLIAM A. KEPNER.

characterized by having a relatively large black nucleolus which
lies in a vacuole (Fig. 6, B] . These are taken to be the wandering
cells or amoebocytes of Microstoma.

When the nematocysts are found lying nearer the ectoderm
cytoplasmic strands appear extending more or less into the
vacuoles about the nematocysts. At P, in Fig. 6, is represented
what appears to be pseudopodia of a wandering cell invading a
vacuole.

Whether or not in this and similar vacuoles there is evidence
of amoebocytes entering the vacuoles, the vacuoles in all cases
are filled by a cell when the nematocysts come to lie near the
ectoderm. About the nematocysts that have but recently reached
the ectoderm the cytoplasm and nucleus of the invading cell
are much more evident than in the cases where the nematocysts
have been at the surface sufficiently long or near to suffer dis-
charge (Figs. 7 and 8). In other words the cells seem to spend
their vitality in caring for these exotic nematocysts and about
the partially discharged nematocysts are to be found cytoplasm
and nuclei which indicate a depleted condition (Figs. 9 and 10).

The nuclei of those cells do not lie at the surface of the vacuole
as Martin shows in his figure number four, plate 14, as he de-
scribes in his legend for this figure. The nucleus in each case
can be seen to lie well within the lumen of the original vacuole.
This may be determined by the study of transverse or longitu-
dinal sections. Figs. 7 and 8 represent two contiguous sections
of a nematocyst near the ectoderm. Fig. 7 was taken through
or near the equator of the nematocyst. Fig. 8 represents the
section taken immediately following that shown in Fig. 7. In
the eighth figure the lower surface of the section is shown. When
the focal plane is raised above this level three microns the nuclei
of the overarching mesodermal cells appear and the nucleus
of the invading cell disappears. Thus it is determined beyond
doubt that there is intimately associated with the exotic nema-
tocysts, when they come to lie near the ectoderm, a cell which
has invaded the vacuole, that had appeared about the nemato-
cyst within the endoderm. This cell can be seen in fresh material
as described earlier in this paper and in all of my permanent
preparations.

NEMATOCYSTS OF MICROSTOMA. 273

The question naturally presents itself as to what kind of cell
it is that has thus entered the vacuole. In all favorable nuclei
that are found within the vacuole there is present a deeply
staining nucleolus about which there is to be seen a vacuole
(Figs. 8 and 10). This makes the nuclei more closely resemble
those of the amcebocytes than those of the other mesodermal cells
of the Microstoma. So I conclude that in each case an amcebo-
cyte enters the vacuole to take charge of the nematocyst as do
the cnidophages of the seolids.

It is interesting to observe how these nematocysts are handled
by the invading cell of each vacuole.1 The nematocysts when
they lie alone within the vacuoles are pointing indifferently
in all directions, but after they have been taken charge of by the
cnidophages they have in all instances their discharging poles
directed towards the exterior of the body. Martin also observed
this final exact orientation of the nematocysts. He says of it
"the large barbed nematocysts in their final position, always lie
so that the thread, when it is discharged, will pass out of the an-
imal, although they may lie pointing in any direction while they
are still in the gut cells or the body-cavity. This rule does not
seem to hold good in the small cylindrical nematocysts, which,
as far as I can see usually lie almost parallel to the surface. It
is very difficult to say how such an orientation can be effected,
but something of the same kind has been detected in ceolids, and
I believe that the same difficulty is present in the nematocysts
of the tentacles in coelenterates " (Martin, 1908, p. 267). It
would indeed be a most difficult matter to understand their
orientation if Hallez and Martin are correct in describing these
exotic nematocysts as always lying within vacuoles. The cni-
dophages appear to be the vital agents which properly orient
the nematocysts.

In my preparation one barbed nematocyst was observed that
had been apparently inaccurately discharged. This was taken
from near the middle of a specimen that when fixed remained
extended. In such distended specimens the mesoderm near the
middle of the animal is reduced to a very thin layer; this together

1 Hereafter we shall speak of this invading cell of the vacuole as a
"cnidophage."

274 WILLIAM A. KEENER.

with a possible shifting of the mesoderm at the exact time of
discharge may reasonably account for this single exception that
is represented in Fig. 1 1 .

There is a second type of nematocyst found in Hydra. These
are the so-called "grappling" nematocysts. By means of them
the Hydra lays hold of or entwines objects of prey; when the prey
is thus seized the tentacles of Hydra sweep it into the mouth of
the polyp. This seems to be the chief function of these nema-
tocysts; but the poison sac at the base of this type of nemato-
cysts suggests that they are something more than mere prehensile
structures. If they are not more than this, I wonder at Micro-
stoma taking them into its tissues and passing them out to its
surface where they are, so far as my observations are concerned,
always oriented as are the barbed nematocysts. Frequent
nematocysts of this type have been found by me within sections
of Microstoma. In some cases these grappling nematocysts
lie within vacuoles in \vhich no attending cell or cnidophage can
be recognized. Fig. 12 shows such a nematocyst lying near a
vacuole from which it may have been discharged. These nema-
tocysts are small and their vacuoles likewise small so that it is
difficult to determine whether cells, associated with them, lie
within their vacuoles or not; but frequent nematocysts of this
type have been found that have attending cnidophages within
their vacuoles (Fig. 13). Moreover all the exotic "grappling"
nematocysts were so oriented that their discharging poles were
directed towards the surface.

I have come to look upon the nematocysts of Microstoma as
being derived from Hydra. This conclusion is based upon the
following facts: (i) that there is a great variation in the number
of nematocysts of Microstome — from very few to many; (2) the
close resemblance between the nematocysts of Hydra and Mi-
crostoma; (3) the absence of cnidocils in Microstoma; and (4)
the absence of cnidoblasts about the nematocysts when found
within the endoderm and deeper mesoderm of Microstoma.

Toward what end is all this careful handling of the exotic
nematocysts of Microstoma directed?

Glaser (1909) has determined by experiment that nematocysts
will not be discharged in the digestive fluids of seolids.

IsEMATOCYSTS OF MICROSTOMA. 275

There would be no protection gained by the flatworm in
passing them into the tissue of its body if in the digestive fluids
of Microstoma the nematocysts are likewise not discharged;
and in case they did discharge within the digestive fluids of Mi-
crostoma they would have done their injury before they had been
taken up by the tissues of the Rhabdoccele.

Glaser (1909) has further demonstrated that when nemato-
cysts pass from digestive fluids into fresh water in all cases they
discharge. Here then are seen grounds for making the inference
that his careful handling of the nematocysts is done in order to
prevent their haphazard discharge in the region of the mouth
as they are egested. Over against this inference is the fact that
in aeolids certain nematocysts pass from the alimentary canal
with the waste food while others are more carefully handled and
discharged by the cnidophages of the cerata (Glaser, 1909):
also wrhen a flatworm is bold enough to attack a living Hydra
with its anterior end it can hardly be considered to be so serious
a matter to have nematocysts discharged casually within the
vicinity of the mouth.

Again the end may be to avoid discharge of nematocysts by
means of distortion within the enteron; for Glaser (1909) has
shown that nematocysts when distorted do discharge. This
suggestion I think can be dismissed with the inference that within
the mesoderm the nematocysts are just as liable to distortion
as within the enteron; for through the muscular actions of the
body the ova are seen at times to suffer considerable distortion.

The third suggestion is that these bodies are taken up to be
used. Grosvenor says: ' 'No one can have witnessed the re-
action of an seolid to various stimuli . . . without being convinced
that the cerata are used as a means of defence' ' (Glaser, 1910,
p. 138). Likewise no one can have witnessed the discharge of
nematocysts of Microstoma when stimulated by pressure or by
acetic acid without looking upon them as organs of defense.

It taxes my faith seriously to believe that endodermal cells
select nematocysts for so remote a purpose as .that of defending
the cell-colony at the ectodermal surface, and yet until further
evidence can be obtained, if I must take a stand, I conclude that
these exotic nematocysts are handled by Microstoma for purposes
of defense.

276 WILLIAM A. KEENER.

If this conclusion is correct and if the nematocysts of aeolids
are derived from ccelenterata, the question can be raised, among
many other more subtle ones, have the seolids acquired their
method of dealing with nematocysts of coelenterata through
flatworm ancestry, or are these cases of parallel development of
reflexes or instincts?

I hope to be able to make a series of experiments upon the
nematocysts of these Rhabdocceles.

SUMMARY.

The nematocysts of Microstoma are derived from Hydra upon
which Microstoma feeds. These ingested nematocysts are de-
livered to the mesoderm by the endoderm of the flatworm.
Within the mesoderm an amoeboid cell takes charge of each
nematocyst and as it is transported to the surface orients it
so that the " sting- thread " has its discharging pole directed
towards the exterior of the animal. Thus Microstoma has come
to use the nematocysts of its prey as do the aeolids use the
nematocysts of their prey.

BIOLOGICAL LABORATORY,

UNIVERSITY OF VIRGINIA,
March, 1911.

BIBLIOGRAPHY.
Boulenger, Charles L.

'10 On the Origin and Migration of the Stinging-Cells in Craspedote Medusae.

Quart. Journ. Mic. Sc., N. S., No. 220, Vol. 55, Part 4.
Conklin, E. G.

'08 Two Peculiar Actinian Larvae from Tortugas, Florida. Pub. Carnegie

Inst., Washington, D. C., No. 103.
Glaser, Otto C.

'09 The Phj-siology of Nematocysts. Journ. Exp. Zool., Vol. 6, No. 3.
'10 The Nematocysts of Eolids. Journ. Exp. Zool., Vol. 9, No. i.
Graff von, L.

'82 Monographic der Turbellarien.
Grosvenor, G. H.

'03 On the Nematocysts of ^Eolids. Proc. Royal Soc., Vol.*»2.
Hadzi, J.

"07 Ueber die Nesselzelhvanderung bei den Hydroidpolypen. Arb. Zool.

Inst. Wien., Tom. XVII.
Hallez, P.

'79 Contributions a 1'Histoire naturelle des Turbellaries. Lille.

NEMATOCYSTS OF MICROSTOMA. 277

Martin, C. H.

'08 The Nematocysts of Turbellaria. Quart. Journ. Mic. Sc., N. S., 206,

Vol. 52, Part 2.
Oersted, A. S.

'44 Plattwurmen. Copenhagen.
Schneider, K. C.

'oo Mittheilungen iiber Siphonophoren: V, Nesselzellen. Arb. jZool. Inst.
Wien., Tom. XII.

278 WILLIAM A. KEPNEF.

EXPLANATION OF PLATE I.

FIG. i. Nematocyst with its associated cell. Obtained by gentle pressure of
cover-glass upon a specimen treated with a drop of i per cent, acetic acid. Free
hand drawing, with aid of micrometer eye-piece. Xi,8oo.

FIG. 2. Same cell after lying ten minutes in the dilute acetic acid. Vesicle
has formed at one end of cell. Free-hand drawing, with aid of micrometer eye-
piece. Xi,8oo.

FIG. 3. Anterior end of Ch&togaster. Free-hand drawing.

FIG. 4. Setae taken from lumen of enteron of sectioned Microstoma. Camera
lucida drawing. Xsoo.

FIG. 5. Endodermal cell. Nematocyst lying within vacuole, V, at base of
cell; M, basement membrane of endoderm. Camera lucida drawing. XI.SOQ.

FIG. 6. Nematocyst within vacuole, V, lying near middle region of mesoderm;
A, supporting cell of mesoderm; B, wandering cells of mesoderm; E, ectoderm; M,
basement membrane of endoderm ; P, pseudopodium (?) of wandering cell. Camera
lucida drawing. X 1,500.

BIOLOGICAL BULLETIN, VOL. XX

PLATE 1

V

6

. p

E

v • ' ^Ste'^^i

v-5$i

M

__V

W. A. KEPNER

28O WILLIAM A. KEPNER.

EXPLANATION OF PLATE II.

FIG. 7. Undischarged nematocyst with its vacuole filled by cytoplasm. It
lies near the ectoderm, E. Camera lucida drawing. X 1,500.

FIG. 8. Section next to that represented in FIG. 7. Shows the nucleus, N,
of cell associated with the nematocyst. Camera lucida drawing. Xi.soo.

FIG. 9. Partially discharged nematocyst. E, ectoderm; M, basement mem-
brane of endoderm. Camera lucida drawing. X 1,500.

FIG. 10. Partially discharged nematocyst. TV, nucleus of cell associated with
nematocyst. Camera lucida drawing. X 1,500.

FIG. ii. Nematocyst discharged approximately parallel to the surface. E,
ectoderm; M, basement membrane of endoderm. Camera lucida drawing. X 1,500.

FIG. 12. Grappling nematocyst, partially discharged. E, ectoderm, M,
basement membrane of endoderm. Camera lucida drawing. X 1,500.

FIG. 13. Grappling nematocyst, partially discharged. An attending cell or
cnidophage is shown within the vacuole within which the nematocyst lies. Camera
lucida drawing X 1,500.

BIOLOGICAL BULLETIN, VOL. XX

PLATE II

E

._- .E

-•p

as^

" ./**£

11

E

8

12

9

W. A. KEPNER

EXPERIMENTS ON THE ADOPTION OF LASIUS, FOR-
MICA AND POLYERGUS QUEENS BY COLONIES
OF ALIEN SPECIES.1

MAURICE COLE TANQUARY.

The experiments recorded in this paper were performed during
the past summer in the laboratory of the Bussey Institution,
Harvard University, at the suggestion and under the direction
of Professor W. M. Wheeler. At the beginning of this particular
series of experiments I had in mind only two species with which
I intended to work, namely, Aphccnogaster tennesseensis and
Aphcenogaster fulva with its subspecies aquia and variety picea,
but a number of field excursions taken during July and August
brought to hand several other species which served as such ex-
cellent material for similar experiments that it was decided to
include them also.

OBJECT OF EXPERIMENTS.

The object of these experiments was to determine whether
the queens of certain species of ants are parasitic upon certain
other species in founding their colonies. The various methods
employed by queen ants in founding their colonies are stated in
Wheeler's "The Founding of Colonies by Queen Ants," and since
the following experiments have a direct bearing upon that ques-
tion, and in fact form a continuation of exactly the same sort
of work, I think I can do no better in the beginning than quote
a few paragraphs from that paper.

"Female, or queen, ants in founding their colonies resort to one of three
methods, which may be known as the usual or typical, the redundant, and the
defective. . . .

i. "The female ant is able by herself alone to start her colony; that is, under
favorable circumstances she can produce and bring to maturity the first brood
of workers and thus insure the further growth and development of the colony.
She is capable of passing many months without nourishment even while she is
feeding her off-spring. Her voluminous fat-body, built up during her larval life
in the maternal nest, together with her degenerating wing-muscles, furnish the sub-

1 Contributions from the Entomological Laboratory of the Bussey Institution,
Harvard University, No. 29.

28l

282 MAURICE COLE TANQUARY.

stances that are converted into food for the young. Although so arduous that few
of the many queens of all that celebrate their nuptial flight during a season ever
succeed in founding a colony, this method is, nevertheless, the one adopted by the
great majority of ants."

The second method refers to the fungus-raising ants (Attii) and need not be
given here.

"3. The female ant, owing to her small and delicate stature, or delayed fertility,
is quite unable to found a colony without the aid of workers of another species.
This method, which is resorted to by parasitic species — using that term in a very
broad sense — appears under three different aspects.

"A. As temporary social parasitism. The female seeks and obtains adoption
in a small queenless colony of another species and permits its alien workers to bring
up her young. When these have matured, they emancipate themselves and become
an independent colony, either by emigration or, more probably, only through the
natural death of the host species."

" B. As permanent social parasitism. The female seeks and obtains adoption
jn a colony of some other species and there permanently resides together with her
offspring. Examples: Aner gates, Strongylognathus, Proiomognathus, Wheeleria, etc.

"C. As dulosus, or slavery. The solitary female enters a small colony of
another species, kills the workers, and seizes and rears the progeny (larvae and pupas)
as a first step towards bringing up her young. The workers produced by
the female subsequently make forays on other colonies of the host species and
appropriate their offspring. While they use a portion of these as food they permit
another portion to develop as 'auxiliaries' or 'slaves' so that the colony preserves
its 'mixed' character."

In the same paper and in others Wheeler has also brought out
the fact that parasitic species are sporadic in their occurrence,,
that they usually produce a very large number of females which
are aberrant in some way, such as being of unusually small size,
sometimes smaller than the workers, in being mimetic in colora-
tion or behavior, in being covered wTith fulvous, myrmecophilous
hairs, etc. Therefore the possession of the above characteristics
to any extent by a species of ant is good grounds for believing
that the queens of that species may be parasitic in founding their
colonies.

A paragraph on page 447 of Wheeler's book, "Ants, their
Structure, Development and Behavior," gives the reasons for
performing adoption experiments with queens of Aphcenogaster
tennesseensis.

"The female of the myrmecine ant Aphcenogaster tennesseensis in being deep
red and of very small size, with a glabrous body and huge, flattened, epinotal spines
protecting the vulnerable abdominal pedicel, is so unlike the females of any other
members of the genus AphcenogasteY that she was originally described by
Mayr as the type of a distinct species (.4. levis). These peculiarities suggest
temporary parasitism and this is borne out by the observations of Schmitt

ADOPTION OF QUEENS BY ALIEN SPECIES. 283

and myself (igoic). Schmitt found near Beatty, Pa., a small mixed colony of A.
tennesseensis and A. picea, a variety of fulva, and one of the commonest ants in
the northern states. He was impressed with the fact that the nest of this colony
was under a stone, because tennesseensis normally nests only in rotting wood.
During the summer of 1902 I found near Rockford, 111., two mixed colonies like that
observed by Schmitt, except that the variety picea was represented by the
variety rudis. Both colonies were of small size and situated under stones. In one
of them a tennesseensis queen was unearthed. There can be little doubt, therefore,
that the glabrous queens seek out small nests of some variety of fulva and start
their colonies in them just as consocians does in the nest of incerta. This habit
is also indicated by the sporadic distribution of tennesseensis and its occurrence
only in localities where some form of fulva is abundant."

The reasons for performing adoption experiments with queens
of the other species will be given as I take up those experiments.

METHOD.

The method used in these experiments consisted in placing
artificially dealated queens of the supposedly parasitic species
in a nest with a few workers and brood of the supposed host
species and observing their behavior under these conditions.
This is the same method used by Wheeler in the experiments
recorded in the paper mentioned above, and possible objections
to it are answered by stating that an artificially dealated queen
behaves in the same manner as a fertilized queen that has dea-
lated herself, and that if one is to conduct such experiments on a
large scale it is necessary to use artificial nests and to prevent the
escape of the introduced queen. For most of the experiments
I used large Petri dishes 4 1/2 and 6 in. in diameter into which I
dumped the ants with their brood and some earth from the nest.
For some of the experiments, however, I used nests of the Fielde
pattern, and in that case I always introduced the queen into the
light chamber and allowed her to find her own way into the dark
chamber where most of the ants stayed, thus imitating natural
conditions a little more closely. When I used Fielde nests,
one chamber was kept covered with a piece of orange-colored
glass. A few of the experiments with each species I watched
very closely, especially when the queen was first introduced,
sometimes watching a nest almost constantly for several hours,
but as I had a large number of experiments going at the same
time, sometimes more than fifty, I had to be content with examin-
ing most of them hastily several times a day and depending

284 MAURICE COLE TANQUARY.

chjefly upon the final result which, after all, is the important
thing.

APH/ENOGASTER TENNESSEENSIS.

As this species does not occur in eastern Massachusetts I
collected part of a large colony from a wood near Urbana, 111.,
June 8 and brought it with me a few days later to the Bussey Insti-
tution, where I transferred the ants to a plaster of Paris nest,
made on the Fielde pattern. The nest contained only workers
and a large number of larvae. By the middle of July winged
females began to appear and during the summer more than a
hundred were produced. These, and a few sent me from Illinois
later in the summer by Messrs. Hugh Glasgow and R. D. Glasgow,
were the ones used in the experiments. Altogether I tried thirty-
five queens of A. tennesseensis with eleven different colonies, two
of which were A .fulva subspecies aquia var. picea, one A. fulva, and
the others A. fulva aquia. The latter is probably more common
where A. tennesseensis occurs. Of these thirty-five queens I
got but one clear case of adoption and unfortunately this was with
one of the colonies for which I have but scanty notes. The notes
of this experiment are as follows:

EXPERIMENT B. 240.

Aug. 18. — 3.00 P.M. I place an artificially dealated queen of Aphcenogaster

tennesseensis in a Petri dish with five workers and about f f ty pupae of A .

fulva aquia.

Aug. 15 — 9.00 A.M. Queen standing by herself.
Aug. 20 — 9.30 A.M. The same.
Aug. 21 — 11.30 A.M. She is standing on the pile of pupae with the workers, none

attacking her.
Aug. 22. The same.
Aug. 23 — 8.30 A.M. She stays in the midst of the workers and pupae and seems

to have been adopted.

Aug. 23 — 3.30. Still with the workers — they never attack her.
Aug. 24 — 8.00 A.M. In the center of the bunch of workers; about three dozen of

the pupae have become callows.
Aug. 28. Still being treated as their own queen.
Sept. 10. I have examined her every day up to the present time, and have never

seen her treated by the workers other than as their own queen. There are

now about sixty workers.

In all the experiments with Aphcenogaster tennesseensis the
queen was attacked by the workers, usually at the very outset,
sometimes not until she had been in the nest for a few minutes.

ADOPTION OF QUEENS BY ALIEN SPECIES. 285

Her behavior was about the same in most cases ; she was timid and
either ran away or crouched down in a supplicatory posture.
Some of the queens, however, showed no fear whatever at first
and approached the fulva workers as though they were members
of the same colony. Constant biting and pulling about by the
legs and antennae, however, soon caused them to try to escape
from their tormentors. In nearly every instance the behavior
of the introduced queen wasv conciliatory and in no case did she
at first seem to resent her harsh treatment but seemed to try
either to conciliate the fulva workers or to escape from them. One
queen, however, as will be shown in the following experiments,
departed widely from what is apparently the normal behavior
under such conditions by seizing a worker and carrying it about
when the others attacked her. At first this behavior seemed to
render her almost immune from further attacks and I began to
think that perhaps only such queens as showed this particular
phase of behavior succeeded in getting themselves adopted.
Later on, however, the attacks were renewed and the queen was
finally killed. I will give one experiment in detail as a type,
and summarize the others.

EXPERIMENT B. 18.

Aug. 9 — 9.00 A.M. I place an artificially dealated queen of Aphcenogaster ten-
nesseensis in the light chamber of a Fielde nest containing about two dozen
workers and a small pile of larvae and cocoons of A. fulva aquia. She goes
at once into the dark chamber. The first few workers that she meets
recognize her at once as a stranger, but do not attack her. Several threaten
her and then pass on. After about five minutes one worker grabs her by
the petiole and carries her about in the nest. She does not seem afraid and
does not try to get away; her manner is conciliatory.

9.10 A.M. She goes out into the light chamber but soon returns.

9.20 A.M. One worker is tugging at her antennae.

9.25 A.M. The same worker is still holding her.

9.27 A.M. She gets free from the worker and runs around in the dark chamber.
Many of the ants she meets do not attempt to attack her.

9.30 A.M. Another worker catches her by an antenna and holds her for a while.

9.44 A.M. Two workers lick her for about two minutes and then go away.

9.47 A.M. A worker seizes her -by one of her tarsi and pulls her about for awhile.

9.48 A.M. Three workers are licking her. She is resting on the sponge about
one half an inch from the brood. A number of workers keep passing by her;
some of them stop and lick her a while, some feel her over with their antennae,
open their mandibles as though to attack her and then go away without doing
so. Sometimes she pats the heads of the workers with her antennae as
though begging food.

286 MAURICE COLE TANQUARY.

10.09 A.M. One worker catches her by a middle leg and pulls her about for
a minute, then steps and licks her thorax and abdomen, then seizes her again
by the leg and pulls her, but apparently not very hard. Again it stops and
licks the queen for a minute and then goes a little to one side and begins
cleaning its own antennae but soon comes back to the queen and repeats
the whole perfcrmance. The worker keeps this up for twelve minutes and
then goes away. The queen has remained all this time in the same position,
one half inch from the brood.

2.00 P.M. She is out in the light chamber but soon goes back.
4.35 P.M. One ant holding her by the petiole, another by one antenna.
6.00 P.M. A worker holding her by the antennae.
8.15 P.M. The same.
Aug. 10 — 7.15 A.M. She is out in the light chamber alone.

9.15 A.M. She is in the dark chamber but not with the workers.

11.30 A.M. She has been in the light chamber by herself most of the morning.

11.50 A.M. Two workers are pulling her about.

2.00 P.M. She is running around in the light chamber, trying to get out. The

workers seem more hostile today than yesterday.
Aug. ii — 8.00 A.M. The queen is dead; I remove her. »

9.05 A.M. I place another queen in the light chamber. She starts into the

dark chamber and is seized in the passage way by two of the workers, one

pulling an antenna, another holding her petiole. They drag her about in

the dark chamber.
9.50 A.M. The queen is on the sponge near the brood and has seized the petiole

of one of the workers in her mandibles. Workers are standing about but

do not attack her; once in a while a worker licks her.
10.30 A.M. A worker seizes her by the hind tarsus and holds her for five minutes.

She is still holding the worker.
10.15 A.M. Still in the same place. She is straddling the head and thorax

of the worker which she is holding by the petiole. Once in a while a worker

seizes her by a leg or an antenna, but some of the workers are licking her.
10.30 A.M. Still holding the worker.

11. 10 A.M. The same.
12. 10 P.M. The same.

1.30 P.M. The same. She is standing in the midst of the brood.

3.50 P.M. Still holding the worker; she is carrying it around now on the sponge

and in the midst of the other workers and brood. She pays no attention

to the other workers, some of which nab her at times. There is no excitement

in the nest.
4.30 P.M. Still carrying the worker around; she does not get far from the

sponge and most of the time is on it with the brood and workers. The

worker seems to be nearly dead and for several hours has made no effort

to escape or defend itself.
5.00 P.M. The same.

6.00 P.M. The same. A worker holding the queen by the antennae.
8.00 P.M. The queen is still on the sponge holding the worker.
9.00 P.M. She is about two inches from the sponge, a worker holding her by

the antennae; she is not holding the worker now.
Aug. 12 — 8.00 A.M. She is again on the sponge with the brood, a worker tugging

at one of her tarsi.

ADOPTION OF QUEENS BY ALIEN SPECIES. 287

9.00 A.M. She is about two inches from the sponge, not holding the worker,
but one of them is holding her by the antennae.

9.15 A.M. She is again on the sponge with the brood and is holding the worker
by a middle leg. The worker seems to be helpless in her grasp and does
not attempt to escape.

9.22 A.M. I change the orange glass in order to make the other chamber the
dark one.

9.35 A.M. The brood has all been removed to the moist sponge in the other
chamber. A few ants still running about in this (now light) chamber. The
queen and her victim still on the sponge. The other ants sometimes come
up and feel them over with their antennae but do not attack the queen.

10.00 A.M. During the past forty-five minutes the queen has been standing
almost motionless. She now begins to move about on the sponge, probably
because all the other ants are gone. At times she stands on her middle and
hind feet and with her front feet seems to try to twist or change the position
of the worker she is holding. She bends under her abdomen as though
spraying the worker with formic acid.

10. 10 A.M. She leaves the sponge and carries the worker about for a minute
and then comes back.

11.30 A.M. She has remained on the sponge all this time holding the worker.

11.40 A.M. She is alone on the sponge. I do not know whether the worker
escaped or whether it died and she dropped it. There are several dead
workers in this chamber.

11.45 A.M. She is running about in the light chamber and seems to be entirely
uninjured.

11.50 A.M. She goes into the dark chamber, seizes a worker by the thorax near
the petiole (almost exactly the same hold she had on the first worker)
climbs on to the sponge and stops just in the middle of the brood and workers
some of which pull her legs or antennas, but she retains her hold on the one
worker and remains in the same position. I remove all the dead workers
from the nest.

1.40 P.M. Still holding the worker in the same way.

3.15 P.M. The queen is still holding the worker and is staying with the brood.

6.15 P.M. The same.

Aug. 13 — 9.15 A.M. The same. There are no dead workers in the nest so she has
not killed any of them.

n.ooA.M. The same.

12. oo M. The same.

1.25 P.M. The queen is not holding the worker but three of the workers are
holding her, one by the petiole, one by an antenna and one by a mouth part.

3.00 P.M. Three workers are holding the queen.

6.00 P.M. The queen is again holding a worker and three workers are holding her.

Aug. 14 — 9.30 A.M. The queen is being stretched out by two workers that are

pulling in opposite directions on her legs. She is not holding the worker now.

10.30 A.M. The queen is running about in the light chamber.

6.00 P.M. She is alone in the light chamber, and has two legs and one antenna

partly gone.

Aug. 15 — 9.00 A.M. In the light chamber alone; she has four legs and both an-
tennae partly gone.

288 MAURICE COLE TANQUARY.

1.50 P.M. The same.

2 50 P.M. The queen is dead and dismembered.

Aug. 16 — 8.15 A.M. I place in another queen. She runs about in the nest
avoiding at first the workers who threaten her. After about four minutes
when she is resting on the sponge (not with the brood), a worker comes up
and begins licking her at once and continues doing so for more than five
minutes.

8.45 AM. The queen is in the corner of the nest feigning death. I thought
she was dead at first; several workers standing around her.

9.00 A.M. She gets up and goes over to the brood; after a few minutes a worker
drags her away.

9.30 A.M. Three workers holding her.

12.00 M. They are still attacking her.

3.00 P.M. She is being stretched out by four workers.

6.00 P.M. The same.
Aug. 17 — 7.30 A.M. Three workers are holding her. One leg partly gone.

5.30 P.M. The queen is dead and dismembered.

On August 19 I placed in another queen. She received the same treatment and
was killed by August 24. On August 25 I placed in another. This one fared a
little better and lived until September 8.

This colony, it will be seen, killed in succession five queens of
Aphtznogaster tennesseensis between August 6 and September 8.
That there is a tendency toward adoption is seen in the fact that
a number of the workers often licked the queen just as they
would their own, while others were attacking her. It seems as
though there is something about the queen that is attractive to
the workers while at the same time the odor of a different species
antagonizes them. In nature the specific odor is probably not
so distinct by the time the tennesseensis queen enters a fulva
nest since she will have been out of the parent nest and running
about on the ground for some little time. In this and in other
adoption experiments only a few ants attacked a queen at once>
usually not more than two or three, and very seldom more than
five or six, even when the colony was a very large one. It would
seem as though the other ants realized that the intruder was
being taken care of.

The variety picea seemed to be more hostile to the tennes-
seensis queen than aqitia, but even here a tendency toward
adoption is shown, as may be seen from the treatment accorded
one of the queens in an experiment from which I take a few notes.
This colony contained seventeen workers and a small pile of
larvae of Aphcenogaster picea and were in a small two-chambered
Santschi nest made of glass and plaster of Paris.

ADOPTION OF QUEENS BY ALIEN SPECIES. 289

July 25 — 2.45 P.M. I place in the first queen. She receives very harsh treatment
and is killed by 4.00 P.M. and her body cut in two at the petiole.

4.10 P.M. I place in a winged queen. She receives no better treatment than

her predecessor and is killed by 6.00 P.M.

July 27 — 9.15 A.M. I place in the light chamber another artificially dealated queen
of A phcenogaster tennesseensis. She runs about in the light chamber for
several minutes, goes into the passage-way several times but not into the
dark chamber. The fulva workers evidently detect her by her odor and one
or two enter the light chamber and dart at her with abdomen turned under
as though to grab her, but simply feel of her with their antennae and allow
her to run away.

9.45 A.M. Still in the light chamber.

4.45 P.M. I was compelled to be away during the day. The queen is now resting
on the brood and is not being molested by the workers. One worker is
licking her, although at times it seems to nab her.

5.30 P.M. The queen is licking the fulva larvae. She moves about among the

workers, none of which now molest her in the least. The pieces of the other

two queens have been distributed among the larva? and are being eaten by

them.

July 27 — 8.00 P.M. The queen is resting on the larva? with the workers and appears

to be perfectly at home.

July 28 — 7.45 A.M. The queen is resting contentedly with the workers and larva?.
I was compelled to leave the experiment for a few days but during my absence
one of the students in the laboratory recorded that on July 28 — 3.00 P.M.
hostilities were again begun against the queen and that she was killed by
9.00 A.M. July 31.

Between August 2 and August 26 I placed in the nest in succession seven more
queens of A phcenogaster tennesseensis, all of which were killed after a longer or
shorter time.

The other colony of Aphcenogaster picea which I used contained
about forty workers, one queen, and a number of larvae and pupae.
I placed the first queen in the nest on July 19 at 2.10 P.M. She
was killed some time during the following night.

July 20 — 8.55 A.M. I place in another queen. She lives
until 10.45 the next day. At 10.50 A.M. I place in a third
queen and she succeeds in living until the morning of July 23.
I also tried two queens of Aph&nogaster tennesseensis with about
twenty workers and some brood of the true A. fulva. Both
queens were killed. The other experiments with colonies of A.
aquia.may be briefly summarized as follows:

B. 240.

Contained ten workers, two winged queens, one male and brood. Between
August 1 8 and September i, I placed in three queens all of which were killed.

B. 24b.

Contained eight workers and a few pupse. Between August 18 and September
7 two queens were killed.

MAURICE COLE TANQUARY.

B.

Contained eight workers and a number of pupae. I placed in a queen August
29, which lived until September 4.

B. 246.

Contained seven workers and a number of pupae. I placed in a queen August 30
and it lived until September 4.

One other experiment with a queen of Aphcenogaster tennessen-
sis consisted in placing a dealated queen in a Petri dish with
about thirty pupae of A. aqiiia. The pupae of this species, as
with other species of myrmicine ants, are naked and do not
require the assistance of workers in hatching. They were col-
lected and placed in a Petri dish with the queen July 23. The
queen paid no attention to them whatever. In a few days the
first callows began to hatch and busied themselves taking care of
the remaining pupae. The queen at first paid no more attention to
the callows than she had to the pupae and in fact stayed in another
part of the nest most of the time. After about a half a dozen
had hatched, however, she began to stay with them more of the
time. The callows were not hostile to her but readily adopted her
as their queen. This, however, is what was to be expected because
of the discovery made by Miss Fielde and recorded in her paper
"Artificial Mixed Nests of Ants." She says: "If one or more
individuals of each species that is to be represented in the future
mixed nest be sequestered within twelve hours after hatching and
each ant so sequestered touch all the others with its antennae
during the ensuing days, these ants will live amicably together
thereafter although they be of different colonies, varieties, species,
genera or subfamilies." At the present time, September 10,
all the pupae have hatched and have become adult workers,
about thirty in all, and are clustered about the tennesseensis
queen exactly as though she were their own.

FORMICA OBSCURIVENTRIS.

This subspecies of rufa occurs throughout the northern states
east of the Mississippi River. It is fairly common in eastern
Massachusetts and has been taken in Illinois. The queens are
large, about the size of those of F. sanguined var. rubicunda,
and have shiny, jet-black abdomens with very little, almost no
pubescence. The thorax and head are a darker red than in F.
sanguined rubicunda. The workers vary greatly in size.

ADOPTION OF QUEENS BY ALIEN SPECIES. 2QI

There are no published accounts of the occurrence of this
species in mixed nests in this country, but Professor Wheeler
found such a nest this last spring near Boston, and has kindly
given me permission to use the note which he made of it at the
time. "April 10, 1910. On the northern slope of Great Blue
Hill, under a stone, I found a mixed colony consisting of a dealated
queen and about 80 workers of Formica obscuriventris with about
100 workers of F. subsericea. There were a few larvae evidently
of the former species in the nest."

On a trip to that same locality with Professor Wheeler, August
17 we came across a very large colony of this species which ex-
tended under several good sized stones from which we secured
10 dealated or partially dealated queens. The fact that they
had been partially dealated indicates that they had probably
been fertilized and readopted by the colony. The particular
slope upon which this nest was found was very rich in nests of
F. subsericea. I brought the queens of F. obscuriventris to the
laboratory and also a large number of workers and brood from
a colony of F. subsericea.

I divided the subsericea colony into five groups of workers
and brood and tried one or more of the obscuriventris queens with
each group. The behavior on the part of the queen and workers
was the same in each case. Some of the workers attacked the
queen, some began at once to lick her and others alternated
between licking and biting her. The queen always seemed to
show surprise at receiving such harsh treatment so different from
what she was accustomed to receiving from the workers in her
own nest. Then she would try to get away and finding that
impossible would turn and<> bite a worker, sometimes producing
a wound that caused death. If the subserica colony was a large
one the queen was soon killed, but in the smaller colonies the
workers, after a time, ceased their attacks and began to tolerate
her presence in the nest, while more and more of them began
to treat her as their own queen. I will give full notes on one
experiment which shows the difference in results attained in
using large and small colonies.

B. 226.

August 18 — 11.30 A.M. I place a queen (I think the mother queen from the
colony) of Formica obscuriventris in light chamber of a Fielde nest containing

MAURICE COLE TANQUARY.

about 75 workers, one male and many pupae of F. subsericea. These
subsericea are all very large individuals.

The workers attack her fiercely when she is in the passageway. Some pull her
toward the dark chamber while some pull the other way. They finally get
her into the dark chamber where six or seven attack her at once.

12.00 M. She is in the dark chamber, workers still attacking her.

1.30 P.M. She is on the sponge in the dark chamber and is motionless. One
worker is holding her by a middle leg, three others are licking her.

1.35 P.M. The one worker has let go her hold and four others are licking her.
She seems to be dead.

2.30 P.M. The queen is dead. I change the light so as to make the other
chamber the dark one.

4.00 P.M. They have moved the dead queen and the pupae into the other
chamber. I change the light again.

4.30 P.M. They again move the dead queen. I change the light again and when
about two thirds of the workers are in the other chamber I block the passage-
way, leaving about two dozen workers, the dead queen and most of the
pupae on one side. I then place in another obscurii>entris queen with these.
They attack her also.

4.45 P.M. Two workers are holding her by the legs and two are licking her.

6.00 P.M. Four workers are holding her by the legs.

August 19 — 8. 20 A.M. Two workers holding her, one by a leg and one .by an
antenna.

11.45 A.M. Two workers holding her.

2.15 P.M. The two workers still holding her.

3.00 P.M. The queen is dead. I remove the plug and let all but twelve of the
workers escape into the other chamber and then replace the plug.

3.50 P.M. I place in another queen with the twelve workers; they attack her.
She does not attempt to bite them except when a worker gives her an un-
usually painful jerk or bite.

4.00 P.M. One worker is licking her and one holding her.
August 20 — 8.00 A.M. One worker holding her by the leg.

2.00 P.M. Three workers standing near her; none holding her.

6.00 P.M. Two workers holding her.
August 21 — ii. oo A.M. The queen is standing alone.
August 22 — n.oo A.M. She is standing with three workers; they do not attack her.

5.30 P.M. Two workers holding her.

August 23 — 9.00 A.M. Not attacking her this morning; two workers staying with
her.

2.00 P.M. Still standing with two of the workers.

Sept. 10. I have examined this nest several times every day
since August 23 and have never seen the queen attacked. There
are now ten workers and a number of larvae and pupae in the nest,
and the queen stays with them all the time exactly as though
she were their own.

The other experiments with queens of F. obscuriventris may
be summarized as follows:

ADOPTION OF QUEENS BY ALIEN SPECIES. 293

B. 22d.

August 18. I placed an obscuriventris queen in a Petri dish with eight workers and
some pupae of F. subsericea. The workers at first attacked her and continued
doing so at intervals until noon the next day. After that they did not
attack her, but tolerated her presence in the nest and she remained standing
by herself until noon August 23. After that she stayed with the workers
all the time and was treated as their own queen.

B. 220.

August 19. I placed an obscurivenlris queen in a Petri dish with eight workers and
twelve pupae of F. subsericea. She was attacked at first just as the others
were and the workers remained hostile until about the middle of the following
day when they became indifferent to her presence and by August 25 they
adopted her and treated her as their own queen. She had received an injury
however in the first attacks from which she did not recover and by the
morning of August 27 was dead. There were then six workers in the nest and
one of them seemed weak. I removed the dead queen and placed in another.
I did not see her attacked after the first day and by a day or two later she
was fully adopted by the five workers, the other one having died.

B. 22d.

August 24. I placed an obscuriventris queen in a nest containing about fifty
workers, one male and many naked pupae of F. subsericea. She was vigor-
ously attacked, but within five minutes after she was placed in, while four
workers were holding her, two others were diligently licking her thorax and
abdomen. There were too many hostile workers in the nest however and
by the following morning she had been killed and beheaded.

B. 226.

August 29 — 4.00 P.M. I placed an obscuriventris queen in a Petri dish with
ten workers and eight pupae of F. subsericea. She was attacked by the
workers that afternoon, but on the following morning was standing by her-
self on the opposite side of the dish from where the subsericea were with
their pupae, and remained thus for two days. On the third day I disturbed
the nest a little, and she moved over and mingled with the workers. They
did not attack her and from then on, she stayed with them most of the time
and is now, September 10, fully adopted.

It will be seen in the above experiments that of the 8 queens
of F. obscuriventris tried with workers of F. subsericea, 5 were
adopted, that these 5 were with colonies of from 8 to 12 workers
and that those that were not adopted were with colonies of from
25 to 75 workers. The ease with which these queens are adopted
in small colonies of subsericea, the decided tendency on the part
of many of the workers, even in the large colonies, to begin licking
and caressing the queens almost from the first, the inquilinous
tendency shown in the behavior of the queens themselves and
especially the finding of the mixed colony mentioned in the note

294 MAURICE COLE TANQUARY.

above, give conclusive evidence of the fact that the queens of
F. obscuriventris are, at least in part, temporary parasites on
F. subsericea, and the fact that a number of partly dealated and
therefore probably fertilized queens had been retained in the
large colony may be taken as an indication that, once formed, the
colony may grow by budding as is the case with the large mound
ant, F. exsectoides, which has also been shown (Wheeler, 1906)
to be temporarily parasitic upon F. subsericea.

LAS! US (ACANTHOMYOPS) LATIPES.

This species not only possesses aberrant females but is peculiar
in the fact that it has two distinct forms of females as shown
in a paper by Wheeler and McClendon, "Dimorphic Queens in
an American Ant." These two queens have been designated
by them as alpha and beta, the beta being the more aberrant
but the more common, and up to the time of the appearance of
the above mentioned paper, the only one known. The alpha
queen, as shown by measurements given in that paper, is almost
exactly intermediate between the beta female and the normal
female of L. claviger. The two more probable of the four possible
hypotheses suggested therein in explanation of the occurrence
of these two forms are

•

1. That the dimorphism may be regarded as the result of
hybridism between L. claviger and L. latipes.

2. That it may be a case of true dimorphism in the female sex.
In a recent paper, "An Aberrant Lasius from Japan"1 Wheeler,

after stating that he had made many observations in the field
which showed that the queens of the following species of Lasius,
L. americanus, L. neoniger, L. nearcticus, L. brevicornis, L.
(Acanthomyops} claviger and L. (Acanthomyops} inter jectus are
able to establish their colonies independently, adds: "But I
have never seen any of the females of our umbratus forms (mixtus
var. aphidicola Walsh, subnmbratus Viereck, minutus Emery
and speculiventris Emery) in the act of founding their colonies
independently, and it is quite probable that they are temporary
parasites on the extremely common L. americanus. Equally
negative have been my observations upon L. (A.) latpies which
BIOLOGICAL BULLETIN, 1910.

ADOPTION OF QUEENS BY ALIEN SPECIES. 295

has the alpha and beta females. . . . That this species is a tem-
porary parasite on L. americanus is indicated by the fact that
near Colebrook, Conn., I found four small mixed colonies."
The four mixed colonies mentioned here are the only ones of which
there is any published account, but Professor Wheeler found
another such colony this last spring and has given me the fol-
lowing note which he made at the time.

"Ellisville, Mass., April 21, 1910 Found a large mixed colony
of L. americanus and Acanthomyops latipes, both about equally
numerous and the latipes were no larger than the americanus.
The nest was under a stone and contained a number of larvae
so young that I could not tell to which species they belonged.
Both species took part in carrying the larvae to a place of safety
\vhen the stone was raised, but the latipes were much more active
in this pursuit than the americanus. The members of the two
colonies were on the most friendly terms and occupied the same
burrows."

The above facts furnished very good reasons for trying latipes
queens with other species of Lasius and especially with L. ameri-
canus. On a field trip with Professor Wheeler in the Litchfield
Hills near Colebrook, Conn., we came across a very large colony
of L. latipes under a stone from which I obtained more than 75
winged females, all but 3 of which were beta females. These,,
and a number collected about two weeks later by Professor
Wheeler, part of them from a colony in the same locality and
part of them found crawling over the ground near Canton, Conn.,
after they had descended from their nuptial flight, were the ones
used in the experiment.

Altogether I tried 79 queens of L. latipes with 28 different
colonies of 8 different species of Lasius divided as follows: 14
colonies of L. americanus, 4 colonies of L. nearcticus, 4 colonies
of L. claviger, I colony of L. claviger var. subglaber, I colony of
L. brevicornis, 2 colonies of L. interjectus, I colony of L. umbratus
var. minutus and I other colony of L. latipes. Out of all these I
got but two clear cases of adoption in which the queen lived,
one of these being an alpha, the other a beta female. Howrever,.
not nearly all the deaths among the queens experimented writh
were due to the hostilities of the other ants, for the queens of

296 MAURICE COLE TANQUARY.

this species do not keep well in confinement and during the time
the experiment was running more than 100 females not used in
experiments died, part of them in the colony with their own
workers and part of them in a nest by themselves. I am quite
sure that at least four or five and probably more of the queens
with L. americanus were already adopted or would have been
adopted had they not died. In the majority of cases when I
removed the dead queen I was unable to find any trace of injury
whatever although in a number of cases the body was dis-
membered and sometimes eaten.

The most noticeable feature about the behavior of the latipes
female when placed in with a small colony of workers was her
desire to be with the brood, although in only one case out of all
79 did I see a queen pick up a larva or cocoon. This queen was
in a Petri dish with a few workers of L. nearcticus and once when
I uncovered the dish she picked up a cocoon and carried it around
in her mandibles for about a minute. The queen was always
attacked, and in the larger colonies of americanus and in the one
of L. umbratns minutiis, very fiercely. I did not see her in any
case attempt to defend herself. In the larger colonies she only
tried to get away and in the smaller colonies she usually got on
the pile of cocoons, while the little workers attacked her and
crawled over her large body in their efforts to kill her. Time
and again when they succeeded in dragging her away she would
return and mount the pile of cocoons. The callows hatching
under these conditions would not, of course, be hostile to her
and as at this time of the year the callow's were emerging in
large numbers there would soon be many workers about her that
would accept her as their owrn queen. Even in the larger colonies
a number of workers could often be seen licking the queen while
others were attacking her. I will give a few notes on the colonies
in wrhich the adopted queens lived, and summarize the others.

B. 19/8 August 9. I place a beta queen of L. latipes in Petri dish with twelve

workers of L. americanus and 150 cocoons.
Aug. 10. The queen is dead.
Aug. 11. I place in another.
Aug. 12. She is on the pile of cocoons with a large number of newly hatched

workers.
Aug. 13. Still resting on the cocoons with the callows.

ADOPTION OF QUEEN BY ALIEN SPECIES. 2Q7

Aug. 14. The queen and callows are clustered together on the brood.

Aug. 15. The queen is almost hidden from sight by the workers that are clustered
about her and over her. They are all resting on the pile of cocoons.

Aug. 16. The same.

Aug. 17. The same. Many of the workers licking her.

Aug. 1 8. She is entirely hidden by the workers clustered over her. I have exam-
ined this nest every day up to the present time, September 12, and the
workers have always been clustered about her. This is much more pro-
nounced than is the case even with the rightful queen in a colony of ameri-
canus. There are now more than 100 workers in the nest.

B. 19/13 Aug. 10. I place a beta queen of L. latipes in Petri dish with 30 workers
and about the same number of cocoons of L. interjectus.

Aug. 11. Dead. I place in another alpha queen.

Aug. 12. She is staying with the workers on the pile of cocoons and seems to be
perfectly at home.

Aug. 13 — 9.00 A.M. She is near the pile of cocoons with a bunch of workers.
I watch them for 30 minutes this morning and do not see the workers show
the slightest sign of hostility although they are touching her and climbing
over her all the time.

6.00 P.M. The workers are licking the queen with every evidence of satisfaction.

I have examined this colony every day up to the present time,
September 12, and the behavior on the part of both the queen
and the workers has always been the same.

In the following experiments not more than one L. latipes
queen was in a colony at the same time, and although in the
summary I make the statement that the queens were killed it
should be remembered that probably the majority of them
would have died even if they had been in artificial nests with
their own workers since more than 100 died that were not
used in the experiment.

B. 9.
Colony of 150 workers and brood of L. americanus; i queen killed.

B. 16.

60 workers and brood of L. americanus; 3 queens killed. The last one stayed in
the nest 17 days; I think she had been adopted.

B. 176.
300 workers and much brood of L. danger; i queen killed.

B. i7c.
25 workers, 3 winged females, 5 males, and cocoons of L. claviger var. subglaber.

3 queens killed, one of these an alpha queen.
B. 1 9/1. 20 workers of L. claviger and brood; 3 queens killed.
B. 19/2. 20 workers and a few cocoons of L. americanus; 3 queens killed.
B. 19/3. 20 workers and a few cocoons of L. americanus; 5 queens killed.
B. 19/4. 40 workers and 100 cocoons of L. americanus; 7 queens killed.
B. 19/5. 20 workers of L. americanus and 200 cocoons; 6 queens killed.

298 MAURICE COLE TANQUARY.

B. iQ/6. 20 workers and 200 cocoons of L. americanus; 6 queens killed.

B. 10/7. 20 workers and 150 cocoons of L. americanus; 6 queens killed.

B. iQ/8. 12 workers and 100 cocoons of L. americanus; 3 queens killed, i an alpha.

B. IQJIO. 30 workers, 7 queens, no brood of L. nearticus; 6 queens killed.

B. IQ\II. About a dozen workers, 30 cocoons, of L. nearticus; 2 queens killed.

B. IQJI2. About a dozen workers, i male and 30 cocoons of L. nearticus; i

queen killed.

B. 10/14. 30 workers, no brood, of L. claviger; two queens killed.

B. IQ!IS. 50 workers, many cocoons of L. nearticus; i killed.

B. iQJi6. 36 workers, many cocoons of L. brevicornis. 2 queens killed.

B. 10/17. 8 workers and many cocoons of L. americanus; 4 queens killed.

B. 10/18. 4 workers and many cocoons of L. americanus; 3 queens killed.

B. iQliQ. 6 workers and a few larvae of L. americanus; 2 queens killed.

B. IQ/2O. 30 workers and many cocoons of L. americanus; 2 queens killed.

B. IQ/2I. 30 workers and no brood of L. minulus; i queen killed.

B. iQ/22. 12 workers and no brood of L. inter jectus; i queen killed.

B. I9J23. 25 workers and brood of L. latipes; i queen killed.

The fact that more queens died in some of the colonies than
others has no significance, for in some of them I did not replace
the first queen.

We thus have two positive cases of adoption of queens of
L. latipes by workers of other species of Lasius and the five records
given above show that latipes occurs in mixed colonies in nature.
At first thought it might seem that 2 adoptions out of 79 at-
tempts is entirely too small a percentage upon which to base
any conclusions whatever regarding the point of temporary
parasitism. Yet considering the fact that nests of L. latipes
are not found abundantly in nature and that the queens are
produced in the colonies in large numbers would we expect a
larger number of adoptions? It is very doubtful whether even
in L. americanus 2 .5 per cent, of the queens succeed in founding
colonies.

Another possible explanation for the mixed colonies, however
may be suggested by an observation which I made on September
5. There had been a nuptial flight of L. americanus and L.
latipes and as I was walking through a park reservation not far
from the Bussey Institution between five and six o'clock in the
evening, I picked up off the ground 30 dealated females of
L. americanus and 5 dealated beta females of L. latipes. As I
had but one box with me I placed the queens of both species
together and upon returning to the laboratory a few minutes
later, dumped them all together in the same nest. After a few

ADOPTION OF QUEEN BY ALIEN SPECIES. 299

minutes the americanus queens were huddled together near a
moist sponge I had provided for them but the latipes queens,
always more restless in confinement, were still running about
in the nest. They were continually running over or through the
bunch of americanus queens and sometimes remained with them
for several minutes, yet I never saw the slightest signs of hos-
tility either on the part of the americanus or the latipes queens.
The next morning 3 of the latipes queens were dead and two days
later the other 2 died, yet I feel quite sure that death was not
caused by any hostility on the part of the americanus queens.
In the same length of time 7 americanus queens died, but there
have been no more deaths up to the present time. I feel sure
that a number of the deaths of the americanus queens were due
to injuries received when I picked them up off the ground. The
fact that the nuptial flight of latipes and americanus may occur
at the same time and that the queens of the two species are not
hostile to one another suggests the possibility of a colony being
founded in common by queens of the two species. This possi-
bility should be tested by experiment. However, I think tem-
porary parasitism a more plausible explanation of the mixed
colonies mentioned above because of the fact that the latipes
queen is of a more nervous temperment, and even though there
were no hostilities between the two queens she would not be
satisfied to settle down in a little cell with the phlegmatic ameri-
canus queen and wait nine or ten months for the appearance
of workers. This nervous disposition, however, is exactly suited
to running about over the ground until the queen happens to
run into a small Lasius colony, and when she gets on to the brood
she is perfectly satisfied to settle down as is shown by the ex-
periments.

The adoption of the latipes queen by the colony of L. inter-
jectus may be looked upon as adding weight to one of the ex-
planations quoted above for the occurrence of the two forms of
females, namely, that the alpha form may be a hybrid between
the beta female and a male claviger. Since adoption occurred
with interjectus it might also be expected to occur with the nearly
related claviger if enough cases were tried. Again the females of
claviger and interjectus are very similar so that the alpha form

3OO MAURICE COLE TANQUARY.

might just as well be a hybrid between latipes and inter jectiis
as between latipes and daviger.

LASIUS UMBRATUS VAR. MINUTUS.

One of the quotations given above shows the lack of evidence
that the queens of this species are able to found their colonies
independently. The lack of such evidence taken together with
the fact that the ant is sporadic in its occurrence, that it produces
an immense number of the sexual forms and that the females
differ from all the other Lasius females in being very small,
no larger than the largest workers, point to temporary paras-
sitism as a method of colony formation.

On August 12 I came across a large mound nest of this species
at the edge of a forest reservation near the Arnold Arboretum.
The mound was in the shape of a very broad dome, about eighteen
inches high and about three feet across at the base. Judging
from the size of the nest, the number of individuals and the way
the grass was shot up through the mound the colony must have
been several years old. With a trowel I took out about a quart
of dirt from the side of the mound and brought it to the labor-
atory. I found that I had about 150 females, an equal number
of males, several hundred workers and many cocoons. As the
amount of earth I removed hardly made an impression on the
mound the colony must have contained several thousand males
and females and a still larger number of workers. Although I
collected rather extensively in that region and in a number of
other localities around Forest Hills, this is the only colony of
minutus that I found during the entire summer. In my ex-
periments writh this species I used 88 queens in 20 different
colonies as follows.

B. 250.
20 workers, 2 winged queens and brood of L. americanus.

B. 256.
8 workers and many pupae of L. americanus,

B. 2Sc.
8 workers and many pupae of L. americanus.

B. 25d.

30 workers and many pupae of L. americanus.
B. 28/1. 7 workers, no brood of L. daviger.
B. 28/2. 24 workers and a few cocoons of L. americanus.

ADOPTION OF QUEENS BY ALIEN SPECIES. 30 1

B. 28/3. 12 workers, and brood of L. americanus.

B. 28/4. 6 workers and a few larvae and pupae of L. americanus.

B. 28/3. 100 workers, and cocoons of L. americanus.

B. 28/6. 100 workers, and cocoons of L. americanus.

B. 28/7. 75 workers, I winged queen and many ocoons of L. americanus.

B. 28 18. 12 workers, no brood of L. inter jectus.

B. 28!$. 25 workers, and brood of L. americanus.

B. 28(10. 25 workers and 5 winged queens of L. nearticus.

B, 28/11. 24 workers, and cocoons of L. americanus.

B. 28/12. 100 workers and many cocoons and young larvae of L. americanus.

B. 28/13. 12 workers, and cocoons of L. americanus.

B. 28/14. 24 workers, and cocoons of L. brevicornis.

B. 28/13. 50 workers, and cocoons of L. nearticus.

B. 28/16. 30 workers and a number of cocoons of L. americanus.

The queen of this species is very active and very timid. I
did not see her attempt to defend herself in any case but always
tried to escape and by her active movements she was usually able
to get away from the workers for quite a while. When caught,
however, she usually succumbed in a much shorter time than
was the case with the other species with which I worked. A few
of them died within an hour or so after being placed in, others
living for several days. Out of all these experiments I got but
one case of adoption. I will give a few notes from that experi-
ment.

B. 2S&.
Aug. 18 — 3.30 P.M. I place a queen of L. umbratus var. minutus in a Petri dish

with 8 workers and a large number of pupa? of L. americanus. The workers

attack her; she does not defend herself but tries to escape from them.
Aug. 19 — 9.05 A.M. She is running around by herself. A worker attacks her

for a little while but she gets away.

Aug. 20 — 9.30 A.M. She is resting on the cocoons with the workers.
Aug. 21 — 9.00 A.M. The same.

n.oo A.M. Still on the cocoons with the workers and seems to be entirely
immune from attack. About a dozen callows have hatched.
Aug. 22. The same.
Aug. 23. The same.
Aug. 24. There are about 40 or 50 workers in the nest now. The queen stays

with them all the time.
Sept. 12. Most of the cocoons have hatched. The queen has been fully adopted.

Although one adoption out of 88 attempts is a small per-
centage, yet I think the ease with which this queen was adopted
is very suggestive, and taken altogether with the facts mentioned
above, namely the sporadic occurrence of the species, the very
large number of females produced, the small size of the females,

3O2 MAURICE COLE TANQUARY.

the fact that these females have not been seen in the act of found-
ing a colony and one additional fact which may be mentioned,
the mimetic coloration of the females (the color of these females
is exactly the same as that of the darker form of americanus) t
I think justifies us in concluding that the queen of this species
is in all probability, temporarily parasitic upon the common
L. americanus.

POLYERGUS LUCIDUS.

This shining slave-maker has been studied by Mrs. Treat,
McCook, Burill, and Wheeler,1 and the European form P. rufes-
cens, by Huber, Forel, Wasmann and Viehmeyer. It differs
from the ants mentioned above in that it is not a temporary
but a permanent parasite or slave-maker, the workers making
raids upon colonies of Formica schaufussi and' its varieties incerta
and nitidiventris.

Studies by Forel, Wasmann and Viehmeyer on the European
P. rufescens tend to show that that form resembles the temporary
parasites in that the queens may be adopted by fusca workers.
In regard to the founding of colonies by P. lucidus Wheeler says:?
"Several experiments in which I introduced artificially dealated
queens of lucidus into nests containing incerta workers with their
brood gave rather conflicting results. In some cases the lucidus
queens behaved like the sanguinea queens under similar conditions
to the extent of killing the alien workers, but they paid absolutely
no attention to the brood. In other cases they were passive
and conciliatory but equally indifferent to the incerta cocoons.
It will be necessary therefore to study this question further
before making definite statements in regard to the methods em-
ployed by our American amazons in establishing colonies."

The queens used in the following experiments were obtained
on the slope of Blue Hill, August 17, the same date upon which
we found the large colony of F. obscuriventris mentioned above.
The nest was along the side of a little-used roadway. On ex-
cavating and bringing it to the laboratory I found the colony
to contain about a dozen workers, an equal number of winged

'See Wheeler, "Ants, Their Structure, Development and Behavior," pp. 482-

487-

2 "Ants, Their Structure, Development and Behavior," p. 486.

ADOPTION OF QUEENS BY ALIEN SPECIES. 303

queens, one dealated queen of P. lucidus, and about 100 very
small workers of F. incerta and a number of cocoons. The
small number of lucidus workers and the comparatively large
number of females indicates that many of the workers were
probably out on a slavemaking raid at the time.

I tried seven queens with colonies of F. incerta and four with
F. schaufussi. Two out of the seven were clear cases of adoption
in which the queens are still living. One of the others I think
was without doubt adopted but died later on from injuries re-
ceived in the first struggles. The two schaufussi colonies con-
sisted of very large individuals and each contained a queen.
They attacked the lucidus queens more fiercely than did the
incertas, and all four were finally killed, yet in both colonies I
saw at times a few of the workers licking the queens and I think
it would be possible if enough cases were tried, to secure adoption
with F. schaufussi, especially if the colonies did not contain
queens of their own.

The behavior of the lucidus queens was like that of the ob-
scuriventris queens in that they did not at first attack the workers
but ran about in the nest avoiding them and trying to escape.
Howrever, when caught and held by the workers they were not
nearly so submissive as were the obscuriventris queens and when
forced to defend themselves they did so with such vigor and
persistence, pierceing the head or thorax of the unfortunate
workers with their long sickle-shaped mandibles so that a very
large number of the incertas or schaufussi workers \vere killed.
In one of the schaufussi colonies the lucidus queen killed every
one of the fifteen workers, leaving only the schaufussi queen.
Later on, however, the lucidus queen herself died. My results
agree with those of Wheeler mentioned above in that I did not
see a queen at any time pay the slightest attention to the brood,
her behavior being different in this respect from F. sanguinea,
which, as Wheeler has shown, kills off the workers, takes posses-
sion of the brood and frees the first callows from their pupal enve-
lopes. I will give a few notes from the two experiments in
which the queens were adopted and are still living.

304 MAURICE COLE TANQUARY.

B. 230.

Aug. 18 — 2.00 P.M. I place a dealated queen of P. lucidus in Petri dish with two
dozen workers and 8 pupae of F. incerta. They attack her at once. She
tries to escape from them and usually manages to do so as she is very vig-
orous. She does not attack them, and bites them only when she cannot
get away.

2. 20 P.M. 3 workers holding her, one by a leg and one by an antenna.
4.00 P.M. Still being held by 2 workers.
5.00 P.M. The same.
6.00 P.M. The same.

Aug. 19 — 8. 20 A.M. 3 workers holding her on the sponge. This hostility was
continued against the queen until the morning of Aug. 23 when I found
her dead.

Aug. 23 — 1.45 P.M. I place in another Polyergus queen. All but one of the
cocoons have hatched. Some of the workers escaped, but two have been
killed. There are now 15 workers in the nest and i of them is injured.
The injured one attacks the queen and is soon killed by her. None of the
other workers attempt to attack the queen. She runs about in the nest
and attempts to escape.

4.00 P.M. 3 workers holding her, 2 by the antennae and I by a leg.
4.30 P.M. 3 workers are licking her.
6.00 P.M. I worker is holding her by a leg.
Aug. 24 — 8.00 A.M. 2 workers are standing by her side; they do not attack her

and she does not avoid them.
3.30. i of the workers is licking her.
4.00 P.M. 3 workers are licking her. A few minutes later one of them holds

her by the tarsus for a while.
Aug. 25 — 8.00 A.M. The Polyergus queen is dead.

2.45 P.M. I place in another queen.

Aug. 26 — 9.00 A.M. She is standing in the midst of the workers and seems to be
entirely uninjured. The workers do not attack her. There are several
dead workers in the nest.
11.45 A.M. She moves around with the workers and does not try to avoid them,

nor do they avoid her nor show any hostility to her.
Aug. 27 — 8.00 A.M. The workers are standing all around her; there are 10 left

alive.

Sept. 12. I have examined the nest several times every day up to the present time
and have never seen her attacked. She has been fully adopted.

B. 236.

Aug. 19 — 1 1. 20 A.M. I place a Polyergus queen in one chamber of a nest containing
the mother queen, 16 workers and a few larvae and pupae taken from a large
colony of F. incerta. The Polyergus queen runs about in the nest trying to
escape. The workers are afraid of her and only a few try to attack her.
They gather up their pupae as they would in case of a raid.

1.45 P.M. All but 6 of the workers escape past the cotton plug into the other
chamber.

6.00 P.M. The Polyergus queen is standing by herself in a corner of the nest
The incerta queen and workers are standing in another corner with the brood.

ADOPTION OF QUEENS BY ALIEN SPECIES. 305

Aug. 20 — 8.00 A.M. The same.
2.00 P.M. The same.

Aug. 21 — 10.30 A.M. The same. The Polyergus queen kept aloof from the in-
cerlas all the time until August 24. When I examined the nest at i.OO
o'clock on that day the Polyergus 'queen was running around in the nest
but avoiding the incertas and they seemed afraid to attack her. About an
hour later, however, I found the incerta queen dead and the Polyergus
queen standing near her. I removed the dead queen and examined her.
I found four cases where her thorax had been punctured by the sharp man-
dibles of the Polyergus queen. From that time on the incerta workers not
only did not attack the Polyergus queen but they ceased to avoid her, and
and on the following day I found one of the workers licking her. Since then
they have treated her as they treated their own queen before. There are
8 workers in the nest now.

Besides the experiments with the above mentioned ants I
tried a few adoption experiments with queens of F. nepticula,
F. sanguined var. rubicunda, and one with a queen of F. difficilis
var. consocians.

The queens of F. nepticula were tried with small colonies of
F. inserta, F. fusca var. subcenescens and F. subpolita var. neo-
gagates. The two queens that I tried with incerta and subce-
nescens gave negative results. The workers attacked the queens
fiercely in each case. The active little queen defended herself,
however, by seizing them with her mandibles. The movements
made the first few minutes were so rapid that the eye could
scarcely follow them. In an hour or so, however, the queen
was killed in each nest. The behavior was the same when I
placed a queen in with a small colony of F. subpolita, but the
subpolita workers being smaller than those of the other two species
were not able to kill the queen so quickly, and after a fierce
struggle of a few minutes she escaped from them. Although she
was attacked again from time to time the attacks were not so
fierce and several times I saw workers licking her; not only
that, but her behavior then became more like what Wheeler has
described for the queens of F. consocians when introduced into
a colony of F. incerta, more insinuating and conciliatory, and
twice I found the nepticula queen feeding a subpolita worker.
Although the two queens which I tried with this subpolita colony
were both finally killed I think the behavior both on the part of
the queen and the workers tends to confirm the conclusions
reached by Wheeler in 1906, that F. nepticula is a temporary
parasite and that its probable host is F. subpolita.

306 MAURICE COLE TANQUARY.

The four queens of F. sanguined var. mbicunda which I tried
with small colonies of F. subsericea were not in very good condi-
tion as I had kept them too long in confinement without food and
with one exception were killed. This one, however, behaved as
did those in the experiments described by Wheeler in 1906. She
killed four of the eight workers in the nest and after some time
took charge of the pupae. The other four workers remained hos-
tile for about two weeks, after which they adopted her and helped
take care of the brood.

The one colony of incerta writh which I tried a queen of F. con-
socians contained the mother queen, about three dozen workers
and several cocoons. I placed the consocians queen in with them
on August 5 at 4.30 P.M. The introduction of the queen caused
very little disturbance. The workers she met nabbed at her
and pulled her legs and antennae a little but not at all violently.
Within an hour after being placed in she was fully adopted and
was going about feeding the workers by regurgitation. The
two queens, incerta and consocians, lived peacefully side by side.
I still have the colony in the laboratory, September 16, and the
yellow consocians queen seems to be perfectly happy with her
strange nest mates. While on a field trip with Professor Wheeler
near Colebrook, Conn., July 30 and 31, he found two mixed
colonies of incerta and consocians and has given me the following
notes which he made of them:

"Colebrook, Conn., July 30, 1910.

"No. i. A mixed colony (in the second year) consisting of a
consocians female and about 100 workers with brood, and a
somewhat smaller number of workers of incerta. This was under
a stone on Mt. Pisgah.

"No. 2. Colebrook, Conn., July 31, 1910. At Beech Hill
near the Massachusetts boundary north of Colebrook I found a
large and flourishing colony of F. incerta containing fully 200
workers and much brood, containing a single consocians queen
that must have been very recently adopted. This is the largest
inserta colony in which I have found a consocians queen."

The finding of mixed colonies of these two species, incerta
and consocians, near Colebrook, Conn., a number of years ago
and subsequent extensive field observations and experiments

ADOPTION OF QUEENS BY ALIEN SPECIES. 307

is what led to the discovery by Wheeler of the phenomenon
of temporary parasitism among ants. At the same time he also
showed that temporary parasitism exists in other American ants
and predicted that it would be found to exist in certain European
ants. These predictions have since been verified by a number
of European myrmecologists. The results of the experiments
recorded in this paper verify his predictions concerning the queens
of A. tennesseensis, L. latipes, L. umbratus minutus and show that
F. obscuriventris is a temporary parasite upon F. subsericea just
as he has shown that the European F. rufa is parasitic upon the
typical F. fusca.1

A question which will naturally arise in the minds of most
people is, "To what extent would it be possible to secure the
adoption of non-parasitic queens by workers of different species
or even of different colonies of the same species?" Much light
would be thrown on the whole subject if an extensive series of
experiments with such queens should be conducted, and in the
future I hope to be able to perform such experiments. Also,
the logician will say that it should not only be proved that queens
that are supposed to be temporary parasites may be adopted by
workers of another species but it should also be proved that such
queens are incapable of founding colonies unaided. That such
is the case with some of the queens experimented upon we have
only such negative evidence as I have given above, and positive
evidence, one way or the other, can be obtained only by an ex-
tensive series of careful experiments. However, positive evi-
dence that any of the queens given above as temporary parasites
are able to establish colonies independently would not neces-
sarily prove that they may not also be in part temporarily para-
sitic upon some other species. It is more than likely that a very
large number, perhaps most of the ants are in some stage of
development toward parasitism. For instance, the above ex-
periments show that the queens of F. obscuriventris are not so
easily adopted as those of F. consocians but much more easily
than those of F. nepiicula or of A. tennesseensis. Wheeler has
shown that the first step toward both temporary and permanent
parasitism from the primitive independent type of colony for-

1 "Observations on some European Ants," 1909.

3O8 MAURICE COLE TANQUARY

mation is the facultative adoption of the queen by workers of
the vsame species, the second the obligatory adoption of the queen
by workers of the same species, and third the obligatory adoption
of the queen by workers of another species. Perhaps the last
species in which one would look for traces of even the first steps
toward parasitism would be our dominant species L. niger var.
americamis, since it is well known that the queens of this species
are able to extablish colonies independently. Workers from one
colony of this species are always very hostile to those belonging
to another colony and still more so towards queens from another
colony, yet, out of eight attempts I succeeded in getting one
young fertilized and dealated queen adopted by a colony of
L. americanus consisting of about three dozen workers, six cocoons
and a number of very small larvae. It is therefore very likely
that many of the queens of americanus establish their colonies
with the aid of wrorkers either of the same or of a different
colony. Thus we see that even in our most dominant species we
may find the first two steps toward parasitism. It is easity seen,
therefore, that there is still a vast amount of work to be done
before the last word can be said upon the interrelations of the
different species of ants.

In conclusion I wish to thank Professor W. M. Wheeler, under
whose direction the work was done, for his many helpful sug-
gestions while the experiments were being conducted and for
reviewing my manuscript.1

1 The colonies containing the adopted queens were left at the Bussey Institu-
tion and Mr. J. W. Chapman, who was kind enough to care for them, wrote me
on November 25, that in each case the adopted queen was treated just as though
she were -the rightful queen, so that there is no question but that the adoption
was complete.

M. C. T.

Vol. XX. May, i9n. No. 6

BIOLOGICAL BULLETIN

EXPERIMENTAL CONTROL OF MORPHOGENESIS IN
THE REGULATION OF PLANARIA.

C. M. CHILD.

The present paper is a brief account of certain results of ex-
perimental work upon Planaria which has extended with inter-
ruptions over the last ten years. The data presented here con-
cern Planaria dorotocephala Woodworth, but P. macnlata is
similar in most respects. A complete account of the work will
appear elsewhere.

i. THE DOMINANCE OF THE ANTERIOR REGION IN REGULATION.

My experiments support very positively the conclusion that
in P. dorotocephala the anterior region, and more specificially
the head region, controls the process of regulation in regions
posterior to it and within a certain distance, i. e., it is physio-
logically dominant over these regions. A limit of effectiveness
for this dominance exists and this varies in general with the rate
or intensity of the processes in the anterior region. Moreover,
not only is the head region dominant over the regions posterior
to it and within the limit of effectiveness, but in general a given
level of the body is dominant over regions posterior to it within
the limit and is dominated by regions anterior to it.

Before considering other facts it is necessary to recall that in
nature all asexual individuals of P. dorocephala above a certain
size consist of at least two zooids and the larger animals of more
than two. This is shown by the following facts: (i) If the ani-
mals are cut into a series of small pieces of equal length from the
anterior end posteriori}', the capacity for head formation de-
creases and disappears in successive pieces from the anterior end
to about the middle of the postpharyngeal region, i. e., the level
where fission occurs, and then increases suddenly. This sudden

310 C. M. CHILD.

increase represents the anterior region of the second zooid. Where
the second zooid is of considerable length a similar decrease in
the capacity for head formation occurs from its anterior end to
about the posterior one half to one third of its length, where
again a sudden increase in the ability of the pieces to form heads
occurs. Posterior to this level there is a more or less diffuse
region of very great capacity for head formation probably in
reality a series of head regions, extending to the end of the body.
Other features of regulation, the rate, the relation between re-
generation and redifferentiation, the size of the head and the
position of the pharynx show corresponding regional changes.
Most of these facts were briefly stated in an earlier paper (Child,
'o6b) and will be considered more fully elsewhere.

Secondly, it is possible by various means to induce fission in a
very short time in animals which are far below the size at which
fission would occur in nature, or which for other reasons would
be incapable of fission under natural conditions. In all cases
where fission can be induced it occurs at a level where sudden
increase .in the capacity for head formation appears. Some of
the methods of inducing fission have been described in another
paper (Child, '106).

And finally, the length of the second zooid, together with other
zooids which may arise from its posterior region varies with the
length of the whole, but not proportionally. In very small
animals the second zooid cannot be found by any methods thus
far employed, but as the animal grows longer, it appears at the
posterior end as a short region of high capacity for head formation
and as growth continues it increases in length more rapidly
than other parts; consequently in very large animals the post-
pharyngeal region is often double or more than double the length
of the prepharyngeal, while in very small animals the prepharyn-
geal region is the longer. In Planaria the second zooid does not
develop a head as visibly differentiated region before its separa-
tion, as do the posterior zooids of the Microstomidae and annelids.
This is because the processes which make this region morpho-
logically posterior are more firmly fixed in Planaria than in
those forms. So long as it remains attached to more anterior
regions, the development of the second zooid can proceed only

CONTROL OF MORPHOGENESIS IN PLANARIA. 311

to a certain stage because existing conditions inhibit it. But,
as I have pointed out, this development does proceed far enough
to permit some degree of independent motor reaction in the
second zooid, and it is this which brings about fission.

From the fact of the existence of the second zooid it follows that
the region which represents physiologically the posterior end of
the first zooid is somewhere near the middle of the postpharyn-
geal region, or in very large animals, in its anterior third. If
we wish to compare anterior and posterior regions of the same
zooid, this fact must be kept in mind. Turning now to the ques-
tion of the dominance of the head region, the facts which indi-
cate this dominance are briefly as follows : A piece above a certain
size from any level of the body is capable of reproducing all
parts which lie posterior to its level whether it produces a head
or not, but it is incapable of giving rise to any of the parts which
lie anterior to its level, unless some approach to head formation
occurs first.

For example, a piece like cd, Fig. I, is capable of producing
a new pharynx and the characteristic postpharyngeal intestinal
branches, even though it may fail to form a head, as is the case
under certain conditions. On the other hand the piece ef, from
the postpharyngeal region, i. e., the posterior region of the first
zooid, never produces even a pharynx unless some approach to
head formation occurs first. If it remains headless as it usually
does, it also remains without a pharynx or mouth and no pre-
pharyngeal intestinal region with an axial intestine ever appears.
Anterior regulation in such cases is limited practically to closure
of the wound.

The dominance of the head region is also ctearly shown by the
relation between the rapidity of development and the size of the
head on the one hand and the position of the pharynx on the
other. In pieces from near the anterior end (e. g., ac or 12, Fig. i)
the head is disproportionately large and develops very rapidly
and the pharynx is situated near the posterior end of the piece
(Fig. 3), while in pieces such as eg, Fig. I, from the posterior part
of the first zooid, the head is relatively small and forms slowly
and the pharynx is anterior to the middle of the piece, Fig. 3.
Moreover, it is possible to demonstrate experimentally in pieces

3I2

C. M. CHILD.

I

4

4

FIGS. 1-5.

CONTROL OF MORPHOGENESIS IN PLANARIA. 313

from the postpharyngeal region that any external conditions
which alter quantitatively the processes which give rise to a
head also alter the position of the pharynx. A single example
is given here to illustrate this relation. The piece eh, Fig. i,
including the whole postpharyngeal region and the second zooid,
always gives rise, under anything like natural conditions, to
animals like Fig. 4 with large heads, normal eyes and the pharynx
somewhat anterior to the middle. If, howrever, we decrease the
rate of the metabolic processes by means of low temperature,
anesthetics, CO2, etc., the head forms much more slowly, remains
much smaller, often possessses only a single median eye and the
pharynx appears close behind the head and remains of small
size. Fig. 5 shows the position of the pharynx in such a piece
after regulation in 1.5 per cent, alcohol, 0.3 per cent, ether or in
water containing CO2. By varying the external conditions
quantitatively all gradations between the conditions of Figs. 4
and 5 can be obtained. Changes in the other direction can be
brought about by high temperature and also by other conditions
which increase the rate of- reactions in the body. In such cases
the head is larger and develops more rapidly and the pharynx lies
further posteriorly.

In general I have found it possible by these and other methods
to alter the localization of the pharynx in pieces within very wide
limits. Since the pharynx is situated at the posterior end of the
median axial intestinal branch of the prepharyngeal region, the
localization of the pharynx in the piece is directly related to the
length of this axial intestine. In these experiments a definite
spatial relation between the rate of the dynamic processes and
the localization and size of organs and regions appears.

Moreover, the dominance of the head region and the limit of
its effectiveness also appear in the conditions determining the
origin of a second zooid. In newly hatched P. maculata — I have
not as yet had opportunity to examine newly hatched P. doro-
tocephala, but have no doubt that they are similar — the second
zooid is not present and even the whole postpharyngeal region
is incapable of forming a head when isolated. In this respect
it is like the posterior region of the first zooid in larger animals.
As the worm grows longer the capacity to produce a head appears

314 C. M. CHILD.

posterior to a certain level, and once established, the second
zooid grows nore rapidly than the first. When the animal is
5-8 mm. in length a very short second zooid is present and in
animals sightly longer than this fission may be induced experi-
mentally (Child, 'io£). In the posterior region of the second
zooid a third zooid arises early because the head region of the
second zooid is not sufficiently active to control more than a
short region; the extreme posterior region of the body often
seems to consist of a series of very short head regions, reminding
one of the series of minute segments which often appear in
the growing posterior region of various annelids.

Under experimental conditions it is possible to delay or inhibit
the appearance of the second zooid in regulating pieces of to
accelerate it. For example, if we isolate the region anterior

7

FIGS. 6-7.

to the level c or 2 in Fig. I, including the old head, it forms an
animal more or less like Fig. 6, in which the head and prepharyn-
geal region are disproportionately large. If now after the growth
and differentiation of the posterior new tissue is completed, we
isolate the posterior end at the level indicated by the dotted line
in Fig. 6, we find it does not produce a new head, i. e., it is physio-
logically as wrell as morphologically a posterior end. If we feed
pieces like Fig. 6 so that growth occurs, we shall find that after they
attain a certain length the second zooid appears and after this
the posterior region of the animal shows a high capacity for head

CONTROL OF MORPHOGENESIS IN PLANARIA. 315

formation. But such animals must grow to a much greater length
than young norrrtal animals before the second zooid appears,
because in them the dominance of the head region is effective
over a greater distance than in young aninals, and a sufficient
degree of physiological isolation (Child, 'iob, 'na) of the pos-
terior region does not occur until a greater length is attained

On the other hand, in pieces of Planaria which remain head-
less second, third and further zooids arise at once and if the pieces
can be induced by stimulation to move about sufficiently, fission
often occurs. The headless piece, unless very small, always
produces a very large amount of new tissue at its posterior end
(Fig. 7). This new tissue consists of new zooids; if we cut it
off at any of the levels b, c, d (Fig. 7) we find that it always pro-
duces a normal whole very rapidly. If we remove the anterior
end of the headless piece at the level a in Fig. 7 after the second
zooid has formed at its posterior end we find that it is now capable
of forming a normal head, but this is possible only when the
second zooid is present, for if we remove both the anterior end and
the second zooid, the remaining piece is incapable of forming a
head. This last experiment as well as others to be described
elsewhere shows that the second zooid influences regions anterior
as well as those posterior to its anterior end.

In general any condition which retards or inhibits the de-
velopment of the head or the dynamic processes in the developed
head favors the development of a second zooid at the posterior
end and any conditions which accelerate the dynamic processes
in the anterior region retard the development of the second zooid.
As noted above, the length which the animal attains before it
gives rise to a second zooid may also be varied experimentally.
In these experiments we have a demonstration of the spatial
limit of effectiveness of physiological correlation and of the re-
lation between this limit and the rate or intensity of dynamic
processes in the region of origin.

One other point may be added as indicating the dominance
of the head region. In animals which decrease in size in con-
sequence of starvation the head region decreases less rapidly
than other parts and so becomes relatively larger the further the
reduction proceeds. Manifestly the head region is able to live

316 C. M. CHILD.

largely at the expense of other parts, to use their substance in
maintaining its own structure.

Only a few few brief suggestions as to the occurrence and general
significance of dominant and subordinate regions in organisms
are possible here. The dominance of the growing tip over other
regions in the plant is a well established fact: in my work on
Tubularia (Child, 'oja, b, c, d) I called attention to various facts
which indicate that the hydranth region is physiologically domin-
ant over other parts and the same is true for Corymorpha and many
other, perhaps all hydroids and for all actinians with which I
have worked. Moreover, when we recall that in most if not all
animals development of the egg begins at the animal pole
and that the distal or anterior region arises at or near this pole,
the possibility that the "animal" distal or anterior region is very
generally dominant becomes at once apparent. I believe that
we have here a very general law of development which has not
been clearly recognized. It is probable, however, that in many
forms the specification of different regions of the egg becomes
fixed to such an extent that their further differentiation is to a
greater or less extent constitutional rather than correlative, but
there are various facts which indicate that within each such
"self-differentiating" part we shall find a condition of dominance
and subordination of parts more or less similar to that which
exists in the whole egg and often even in the adult of some other
less highly differentiated forms.

II. THE REGULATORY DEVELOPMENT OF THE ANTERIOR END.

The "normal" head of P. dorotocephala possesses the form shown
in Fig. i . As in various other species of Planaria the eyes consist
of black pigment spots and unpigmented sensory areas. The
sensory cephalic lobes or auricles are also only slightly pigmented
and are marked off as well-defined light areas from the rest of
the dorsal surface of the head, which is dark brown. The light
areas of the eyes and the auricles are indicated in Fig. I by dotted
lines.

The result of regulation in the larger isolated pieces is usually
a "normal whole" (Fig. 4), i. e., and individual with head essen-
tially like Fig. i, though varying in size, and the other organs

CONTROL OF MORPHOGENESIS IN PLANARIA.

317

characteristic of the species in nature. In pieces below a certain
size which varies with the region of the body and with other
internal and external conditions, normal wholes develop from iso-
lated pieces but rarely or not at all. A brief description of some
of the characteristic results of regulation under these conditions

9

10

one

15

CJc

II

CM;

16

12

20

21

22

23

U

FIGS. 8-27.

is necessary. For convenience these are arranged under a number
of heads, but it must be understood that the "types" thus dis-
tinguished are by no means sharply denned. They grade into
each other in various ways, so that the whole series of types is
actually a continuous series.

I. Tailless. — (Fig. 8.) Very short pieces from the more
anterior regions often give rise, especially under certain experi-

318

C. M. CHILD.

mental conditions to heads without pharyngeal or postpharyngeal
regions.

2. Normal. — (Figs. 2, 3, 4, 6.) In larger pieces these are all
much alike (Fig. 4), but in smaller pieces they differ according
to the region of the body and other internal and external factors.

I/

A

v

Short pieces from the anterior region produce normal wholes
like Fig. 2, those above a certain size from the pharyngeal region
or immediately posterior to it produce wholes like Fig. 3.

3. Teratophthalmic. — (Figs. 9-23.) Abnormalities of the eyes
are frequent in regulation of pieces of Planaria and their occur-

CONTROL OF MORPHOGENESIS IN PLANARIA. 319

rence has been noted by various authors, though no one has
previously attempted to determine the conditions of their ap-
pearance. They are of various kinds, but in most cases fall into
one of the following groups: differences in size or asymmetries
of position (Figs. 9-11) ; approach to the median line (Figs. 1 1-12)
partial union of pigment areas (Figs. 13-19); single median pig-
ment areas with double sensory areas (Figs. 20-21); single
median eyes (Fig. 22) ; three eyes, one median, others right and
left (Fig. 23). In all of these cases the head may be entirely
normal otherwise, though the more extreme types of abnormal
eyes (e. g., Figs. 20-22) usually occur in heads of relatively small
size and slow rate of development.

4. Teratomorphic. — (Figs. 24-27.) Very frequently, however,
the head itself is abnormal in form, the median anterior region
being incompletely developed or absent and the auricles appear-
ing on the anterior end of the head and in various degrees of
approach to the median line: they may be separated by a con-
siderable interval (Figs. 24, 25) , or partially (Fig. 26) or completely
united in the median line (Figs. 27). As soon as the new heads
become pigmented the unpigmented auricular areas are clearly
visible and in such cases as Figs. 26 and 27 it is these areas which
enable us to identify with certainty the auricles. Teratomorphic
heads almost invariably possess only a single median eye, or
they develop a single median eye first and later two more sym-
metrically placed eyes (Fig. 23) ; they are always of small size
and develop slowly.

5. An ophthalmic. — (Figs. 28-34.) These pieces show a dis-
tinct outgrowrth of new tissue at the anterior end, varying widely
in form, but always without eyes. This outgrowth is well
innervated and its motor activities resemble those of a head.
Sometimes the partially or wholly fused auricles appear anteriorly
in the median line on the outgrowth (Figs. 28, 29), as in the
extreme type of teratophthalmic pieces (Figs. 26, 27), but
usually the outgrowth shows no visible differentiations charac-
teristic of a head. In shape the outgrowth ranges from a broad
and blunt form like Figs. 30 and 31 to a very slender form of
varying length, often almost as long as the remainder of the
piece (Figs. 32-34). These pieces always give rise to a new

320

C. M. CHILD.

pharynx whether they arise from the old prepharyngeal, pharyn-
geal or postpharyngeal regions.

The anophthalmic pieces are always more active than the
headless pieces described below and their movements are better
coordinated; moreover, the motor activity of the anterior out-
growth itself shows very clearly that it represents some approach
to head formation.

V

35

u

XT'

36

Fir, 5. 35 37.

6. Headless. — (Figs. 35-37.) The headless pieces are dis-
tinguished from the anophthalmic by the fact that the new
tissue merely fills the contracted wound and does not grow out.
According to the degree of contraction of the wound these pieces
may appear like Fig. 35 or Fig. 36. Headless pieces which arise
from the old prepharyngeal or pharyngeal region always develop
a new pharynx (Figs. 35, 36), but these from the postpharyngeal
region do not give rise to a pharynx (Fig. 37).

It will be observed that in the order figured types 2-6 present
a gradation from the normal head to the headless condition.
The eyes first show irregularities, then appear nearer together
and finally in partial or complete fusion in the median line. In
the one eyed forms the region between the auricles may be re-
duced or absent and the auricles appear in various degrees of
approach to each other, or like the eyes partially or wholly
united in the median line, and this latter condition is sometimes
seen in the anophthalmic forms: in these all possible gradations
appear between an outgrowth resembling a head in form to a
long slender pointed outgrowth or one which is short and small,
and finally this condition passes into the completely headless
condition.

CONTROL OF MORPHOGENESIS IN PLANARIA- 321

The similarity of the first part of this series to the abnormal
and cyclopian fish embryos obtained by Stockard is at once
apparent.

III. EXPERIMENTAL CONTROL OF THE CHARATER
OF THE ANTERIOR END.

Attempts to control the appearance of the various types of
anterior end were first made in my experiments some five years
ago and it was found that various methods of control were possi-
ble. The anesthetics were first used for this purpose almost
three years ago; a brief report of certain phases of this work has
already appeared (Child, 'ioa). Thus far I have been able to
control the process of head formation through various internal
and external factors.

In connection with these experiments it has become necessary
to standardize the material as far as possible, i. e., to obtain some
method of comparing the physiological condition of worms of
different size and age, of well fed and starved animals, of ani-
mals before and after regulation, pieces from different regions,
etc. By the use of alcohol of low concentration (in most cases
1.5 per cent.) I have been able to accomplish this to a consider-
able extent. A part of the results of this work in their bearing
upon the problem of senescence have been described elsewhere
(Child, lib). These experiments confirm the conclusion reached
from experiments on regulation alone, viz., that in general the
capacity for head formation stands in close relation to rate of
the metabolic processes in the piece.

For given conditions a certain rate of reaction is necessary for
the formation of a normal head. When the rate falls below this
critical level teratophthalmic heads appear first, then as the rate
falls still further the teratomorphic and anophthalmic and head-
less'forms appear in succession. Moreover, among the tera-
tophthalmic forms the cases of partial fusion represent decrease
in rate below the critical level and the single median eyes further
decrease. The irregularities of form and position occur most
frequently at levels near the old head in small worms and in
small pieces from large worms.

Since length of the piece, region of the body and physiological

322 C. M. CHILD.

condition are all factors in the capacity for head formation,
pieces which are to be directly compared must be of the same
size, except of course where the size factors are under consideration,
they must be taken from the same region of the body except
where the regional factor is to be analyzed and from individuals
in similar physiological condition except where different physio-
logical conditions are being compared. As noted above a re-
lative measure of physiological condition is possible; it is also
possible to control the size of the piece and the region of the body
from which it comes within certain limits, especially in the
pharyngeal region and near the head where localized morpholog-
ical landmarks exist.

The method of experiment which I have found most satis-
factory is to compare considerable numbers of similar pieces
rather than individuals. For a temperature experiment for
example, animals of the same size and as nearly as possible in the
same physiological condition are taken as the basis: from these
pieces as nearly as possible of the same size and from the same
region of the body are cut and a number, usually fifty, of these
pieces is placed in one temperature and a like number in another.
For an alcohol experiment fifty pieces obtained in a similar
manner are placed in alcohol and fifty in water at the same
temperature and under otherwise similar conditions.

By taking into consideration the factors of size of piece and
region of the body it is possible to cut series of pieces which will
give 100 per cent, or very nearly of normal heads, or on the other
hand 100 per cent, or very nearly of headless forms, or we can
obtain series which in consequence of unavoidable differences
in size and region and differences in the physiological condition
of the animals from which they were taken, produce a certain
percentage of several of the types described above. In all
these cases we can compare the results under the control con-
ditions with those occurring under other definitely known con-
ditions.

For example, pieces above a certain size from the anterior
region of the body, which give 100 per cent, or nearly of normal
heads in well aerated water at a temperature of 2O°C. show a
large percentage of teratophthalmic forms in water at io°C. or in

CONTROL OF MORPHOGENESIS IN PLANARIA. 323

water containing CO2 or in alcohol 1.5 per cent, at the same tem-
perature as the control, and under more extreme conditions
the teratomorphic, anophthalmic and headless forms appear.

On the other hand, if we take pieces which show nearly 100
per cent, of headless forms in well aerated water at 20° C. we
find that similar pieces kept at 30° C. will produce not only
anophthalmic, teratomorphic and teratophthalmic heads, but
a considerable percentage of normal heads. These illustrations
wyill serve to indicate the general character of the experiments
along this line.

Below are given the results of a few series of experiments
on size, region, temperature, alcohol, nutrition and mechanical
stimulation.

I. Size and Region.

We may investigate these factors by cutting the whole body
of worms of the same size and similar condition into one-fourth,
one-sixth, one-eighth, one-twelfth and one-sixteenth pieces,
etc., and comparing the results. The following table is a part
of a series of this kind and shows in somewhat abbreviated form
the results for one-eighth and one-sixteenth pieces of large worms.
The head is not included and the pieces of the body are numbered
consecutively I, 2, 3, etc., from the anterior end backward; the
consecutive numbers in the second column refer to the pieces.
The remaining columns show the percentages of each of the
regulatory types for each set of pieces. In this series each set
consisted of only ten pieces, but the results are almost as uniform
as in much larger series.

In this table the factors of size and region appear clearly. We
see that the capacity for forming normal wholes decreases pos-
teriorly and then in the region of the second zooid increases again
suddenly. The anterior ends of the one-eighth pieces respect-
ively are at approximately the same levels as those of nos. I, 3,
5, 7, 9, n, 13, 15 of the one-sixteenth pieces and it is at once
evident that the longer one-eighth pieces possess a greater capac-
ity for head formation than the shorter one-sixteenth pieces.

We can obtain a more striking expression of this fact by com-
paring very long and very short pieces from similar worms cut
with anterior ends at the same level. One of my series of this

324

C. M. CHILD.

Tailless
Heads.

Normal
Wholes.

Teratophthal-
mic and
Teratomorphic.

Anophthalmic
and
Headless.

Dead.

i

IOO

2

10

90

3

20

80

One-eighth

4

10

90

Pieces .

c

IO

no

D

yu

6

10

2O

70

7

80

2O

8

90

IO

i

70

30

*

2

30

60

10

3

IO

30

60

4

IO

80

IO

5

IOO

6

IOO

7

IOO

One-sixteenth

8

IOO

Pieces . .

80

20

10

10

70

20

n

70

30

12

20

60

20

13

20

60

2O

14

40

50

10

15

80

IO

10

16

QO

IO

Teratoph- Terato-
thalmic. morphic.

Anoph-
thalmic.

Headless.

Dea

O O

0

0

0

0

10

82

6

sort is as follows: from large worms of equal size and similar
conditions fifty pieces were cut including the whole postpharyn-
geal region (eh, Fig. i) ; a second set of fifty short pieces from the
anterior end of the postpharyngeal region (ef, Fig. i), i. e., with
anterior ends at the same level as the larger pieces, was used for
comparison. The following table gives the results in percentages.

Normal.

Long IOO

Short o

In the same way we may compare other levels of the body and
show to what extent the power of head formation, depends upon
the cells near the anterior cut end and to what extent upon their
connection with more posterior regions. My experiments along
this line show that in general the ability to form heads in pieces
from the posterior region of the first zooid is almost wholly de-
pendent upon their connection with more posterior regions, while
in pieces from the anterior end of the body it is very largely
independent of such connection.

CONTROL OF MORPHOGENESIS IN PLANARIA. 325

2. Nutrition.

My experiments along this line are rather extensive and the
results are of considerable interest. I give here only a single
series of experiments. The pieces used included approximately
the region between the lines 2 and 3 in Fig. I ; fifty pieces were
cut from worms which had been fed every 2-4 days for two
months (I), another set of fifty pieces from worms of approxi-
mately the same size which had had no food for forty days before
the experiment began (II). Results are given in percentages.

Normal. Teratoph- Terato- Aboph- Headless. Dead,

thalmic. morphic. thalmic.

1 4 46 6 30 14 o

II o o 2 10 64 24

The capacity for head formation is manifestly much greater
in the well-fed (I) than in the starved pieces (II).

3. Temperature.

Under this head only a single series is given, but the results
are characteristic. In this series fifty pieces from similar worms
including the region between the lines 3 and 4 in Fig. I were
allowed to regulate at each temperature. Results are given in
percentages.

Normal. Teratoph- Terato- Anoph- Headless. Head,

thalmic. morphic. thalmic.

28°-30° 74 22 2 O O 2

l8°-20.° 22 40 2 14 22 O

The death of one individual (2 per cent.) at high temperature
was purely accidental. The percentages show the general
character of the results. With greater differences of temper-
ature greater differences in the results can be obtained.

4. Alcohol.

The experiments with alcohol and other anaesthetics present
certain complicating factors which will be considered fully else-
where, but the effect of alcohol upon the capacity for head for-
mation and upon the character of the head formed is readily
demonstrated.

In the series presented here fifty pieces from similar worms,
including the region between lines 2 and 3 in Fig. I were placed

326 C. M. CHILD.

in alcohol 1.5 per cent, for 48 hours after section and then re-
moved to water; fifty similar pieces were placed in water for con-
trol. Results are in percentages.

Normal. Teratoph- Terato- Anoph- Headless. Dead,

thalmic. morphic. thalmic.

Ale. 48 hrs 14 16 8 18 38 6

Water 20 44 2 26 8 o

A much larger percentage of headless forms appears in alcohol
than in water.

5. Complications in the Effects of Depressing Agents and Con-
ditions.

In the present paper I wish merely to record the fact that under
certain conditions and within certain limits depressing factors
increase the head-forming capacity in pieces, instead of decreasing
it. As I shall show elsewhere, this result can be accounted for
without difficulty. It depends largely upon the fact that the
rate of metabolism in the cells at the anterior end of the piece
which react to the wound increases in the course of this reaction,
while that of more posterior regions of the piece remains essen-
tially the same. The depressing factor decreases the rate in
both regions and with a certain degree of depression the
more posterior region will become almost quiescent while the
anterior region which loses its previous differentiation in conse-
quence of the wound reaction will still be active. Under these
conditions the dominance of the anterior over the posterior region
will be increased and a larger percentage of heads may be formed
in spite of the depressing effect, simply because the anterior
region is more capable of living, growing and differentiating at
the expense of the posterior region. Under these conditions
then the depressing effect will appear to be a morphogenic

stimulus.

6. Mechanical Stimulation.

As is well known, the shorter pieces of the planarian body,
especially those from the more posterior regions of the first
zooid, very commonly move about but little until regulation
has attained a certain stage; after this stage "spontaneous"
movements occur. Larger pieces show these movements at

CONTROL OF MORPHOGENESIS IN PLANARIA. 327

earlier stages and in still larger pieces, e. g., after removal of only
the head region, they are never absent. In general the pieces
which show more motor activity when left undisturbed in the
dishes and which react more readily to slight stimuli are those
which have the greater capacity for producing new heads, i. e.,
for becoming wholes. From a number of pieces it is possible
within a few hours after section to select with a considerable
degree of certainty those which will form heads, or which will
show the greater degree of head forming capacity, merely by
observing the differences in motor activity and reaction to slight
stimuli.

With these and various other facts in mind I carried out ex-
periments to test the effect of mechanical stimulation to movement
upon the formation of heads. The pieces were cut of such size
and such distance from the old head that they showed little or no
"spontaneous" movement during the first few days after section.
Various methods of stimulation were employed: one of these
was that of tilting the broad flat dish in which they were kept
so that each piece was out of water for a few seconds. This
usually serves to induce locomotion, but it does not continue
long after the pieces are again in water. Sometimes the pieces
were loosened from the glass by currents from a pipette or were
turned upon their dorsal surfaces with needles, after which they
usually righted themselves. But the most commonly used and
most effective method was that of stroking or gently pushing
the pieces with a small soft camel hair brush. Usually a few
seconds of such stimulation was followed by relatively long sus-
tained locomotion. The stimulation began a few moments
after section of the pieces and wras repeated at least every hour
and often every half hour from about 8.30 A.M. till 6 P.M. and
again at least twice between 8 and n P.M. each day until the
heads in such pieces as produced heads were wrell developed
and the eyes visible. Each time that the pieces of the one set
were stimulated the water of the dish containing the control
was gently disturbed and stirred in order to avoid as far as
possible differences in aeration. Care being taken not to touch
the control pieces directly, the movement of the water almost
never induced movement of the pieces. The dishes wrere kept

328 C. M. CHILD.

under as nearly as possible identical conditions of light and
temperature.

Two series of experiments are given here. In the first the
pieces included the region between the lines 2 and 3, Fig. I,
fifty pieces for stimulation (I) and fifty for control (II) being
taken from similar worms.

Teratoph- Terato- Anoph-

Normal. thalmic. morphic. thalmic. Headless. Dead.

I o 28 8 32 32 o

II o 8 8 26 58 o

The second series consists of pieces including the region be-
tween the lines 3 and 4 in Fig. I and from the same worms as
the first. These pieces, being posterior to those of the first series
and of approximately the same size produce heads less frequently.
As in the first series, fifty pieces were used in the stimulated set
(I) and fifty in the control (II).

Teratoph- Terato- Anoph-

Normal. thalmic. morphic. thalmic. Headless. Dead.

1 4 10 2 18 64 2

II O 6 2 12 80 o

The two pieces which produced normal heads in the stimulated
set of this series were undoubtedly pieces which included the
anterior end of the second zooid : the presence of this region even
at the posterior end of a piece increases greatly its head-forming
capacity. In both series the effect of stimulation appears un-
mistakably in the increased capacity for head formation. With-
out doubt it is not the actual performance of the movement itself
but the stimulation which is the important factor. Stimulation
means increase in the rate or intensity of the dynamic processes
in the piece and this appears in increased and altered formative
capacity. I believe that these and other similar experiments
to be described elsewhere establish and confirm against all ob-
jections my position that the process of regulation is essentially
a' dynamic, or as I have often called it, a "functional" process
or complex of processes.

7. The Morphogenic Effect of Quantitative Changes in Conditions

in General.

In the temperature, nutrition and stimulation series the exper-
imental conditions differ quantitatively from those of the control.

CONTROL OF MORPHOGENESIS IN PLANARIA. 329

For reasons which I shall state later it seems probable that the
factor of length of the piece is essentially quantitative in its
effect upon head formation, and the regional factor is certainly
very largely quantitative. The effect of alcohol and other anaes-
thetics is also quantitative, at least in large part.

The results obtained in these experiments point to the possi-
bility of controlling and analyzing morphogenesis in a great variety
of ways and of obtaining some insight into the nature of morpho-
genic processes themselves. As regards the data presented,
there are many facts which indicate that the rate of oxidations
or of certain oxidation processes is the chief factor in the results
obtained.

From the morphological point of view the most important
point is that certain organs may alter their localization, their
form and their relation to each other and may finally disappear
as the rate of the processes concerned decreases and vice versa.
It is evident that under given conditions a certain rate of certain
dynamic processes is necessary for the production of certain
morphological effects.

IV. THE EFFECT OF CHANGE OF POSITION OF PARTS

IN THE BODY.

If we cut small pieces from the posterior region of the first
zooid such for example as ef, Fig. I, they will give 100 per cent, or
nearly of headless forms, If, however, we cut out a large piece,
eh (Fig. i) and allow a head to form at its anterior end and then
after a week or ten days cut out small pieces just behind the head,
we shall find that they now produce 100 per cent, or nearly of
forms like Fig. 2, i. e., normal wholes. These pieces represent
approximately the region which when taken from the original
animal gave rise to 100 per cent, or nearly of headless forms.
Their capacity for forming heads has been widely altered by their
changed position in the body and more specifically I believe
by altered correlation with other parts. The cells of such a
region are not directly concerned in the formation of a head, but
the more anterior their position in the new individual, the greater
their head forming capacity becomes.

In a similar manner we can decrease the head-forming capacity

33O C. M. CHILD.

of parts which were originally near the anterior end and possessed
the capacities characteristic of that region. We can make almost
any portion of the prepharyngeal region into a pharyngeal region
with corresponding change in regulatory capacity. The anterior
region of the second zooid, with its very high regulatory capacity
may be changed into the postpharyngeal region of a first zooid
with a great decrease of regulatory capacity. Changes of this
kind may be made almost at will merely by so arranging the ex-
periment that the part in question is brought into a certain
position with regard to other parts. It not only acquires a
different structure but its capacities are altered. These experi-
ments show a wide range of possibilities of influencing the mor-
phogenic capacities of parts indirectly through correlation. I
believe the point is of some general interest. In general terms
these experiments show that the regulatory capacity of a part
of the body may be altered through correlation. As regards a
specific case, the head, they show that a part which when iso-
lated originally possessed little or no capacity to form a head
may have this capacity greatly increased by closer association
with a region where a head is forming, even though the part itself
is not directly concerned in the formation of the head. Such facts
as these, possess, I believe, a certain significance in connection
with the problem of inheritance. In these cases we actually
alter the hereditary capacities of the part in question by altering

its correlation with other parts.

•
* HULL ZOOLOGICAL LABORATORY,

UNIVERSITY OF CHICAGO,
March, 1911.

BIBLIOGRAPHY.
Child, C. M.

'o6a Contributions toward a Theory of Regulation, I. The Significance of
the Different Methods of Regulation in Turbellaria. Arch. f. Entwicke-
lungsmech., Bd. XX., H. 3, 1906.

'o6b The Relation between Regulation and Fission in Planaria. Biol. Bull.,
Vol. XL, No. 3, 1906.

'07 An Analysis of Form Regulation in Tubularia, I., Stolon Formation and
Polarity. Arch. f. Entwickelungsmech., Bd. XXIII. , H. 3. IV., Regional
and Polar Differences in the Time of Hydranth Formation as a Special
Case of Regulation in a Complex System. Ibid., Bd. XXIV., H. i. V.,
Regulation in Short Pieces. Ibid., H. 2. VI., The Significance of Certain
Modifications of Regulation: Polarity and Form Regulation in General.
Ibid., 1907.

CONTROL OF MORPHOGENESIS IN PLANARIA. 33!

'09 The Regulatory Change of Shape in Planaria dorotocephala. Biol. Bull.,

Vol. XVI., No. 6, 1909.
'ioa Analysis of Form Regulation with the Aid of Anesthetics. Biol. Bull.

Vol. XVII., No. 4, 1910.
'iob Physiological Isolation of Parts and Fission in Planaria. Arch. f. Ent-

wickelungsmech., Bd. XXX. (Festband fur Prof. Roux), II. Teil, 1910.
'na Die physiologische Isolation von Teilen des Organismus. Vortr. u. Aufs.

u. Entwickelungsmechanik, H. XI., 1911.
'nb A Study of Senescence and Rejuvenescence, Based on Experiments with

Planaria dorotocephala. Arch. f. Entwickelungsmech., Bd. XXXI., H. 4,

1911.

THE BIOLOGY OF THE RED-BACKED SALAMANDER
(PLETHODON CINEREUS ERYTHRONOTUS

GREEN).

M. ETHEL COCHRAN.

DISTRIBUTION.

Baird says, "Species of the genus Plethodon are found all
across the North American continent." Boulenger makes the
range of the subfamily Plethodontinfe North America, with
possibly one species in the valley of the Rio de le Plata. Of
the subfamily Plethodontinse, Gadow states that it "is almost
entirely American (with one species, Spelerpes fnscus, in
Europe)," while Holmes says that "the Plethodontinse form a
large group which is mainly confined to America." In the
last edition of his "Manual," Jordan gives the range of the
family as "chiefly North America."

The "red-backed" salamander (Plethodon cinereus erythronotus
Green) belongs to the Plethodontinae and according to Jordan
is confined to the eastern United States. Gadow says, "Ple-
thodon erythronotus extends into Canada," while Boulenger
would have it range in the eastern United States and Canada.
Cope claims the species Plethodon cinereus, including all varieties,
has an extreme range, being "found throughout the United
States, east of the Mississippi River. It appears to be more
abundant in the Middle States; its northern range is to the middle
of Maine, Ontario and Michigan." Kingsley says it is the "most
abundant salamander in the eastern United States."

Throughout Worcester County, Mass., this little salamander
has proved very abundant. It is not an uncommon thing to
find twenty or more during a short afternoon's walk. Every
little wood has its dainty, shy, inhabitants who so easily may be

overlooked.

HABITAT.

The tiny creatures are not visible to the casual observer, for
on bright days they are always concealed beneath stones or fallen
logs. Holmes says they are in damp situations under rocks or

332

THE BIOLOGY OF THE RED-BACKED SALAMANDER.

333

decaying logs, while Montgomery found them near West Chester,
Pa., "at all seasons, never in streams or boggy places, but in
woods and hillsides beneath wood and stones, a strictly ter-
restrial species," Kingsley, on the other hand, claims " Ple-
thodon cinereus is found everywhere in woods, under bark, logs
and stone, in comparatively dry places." Miss Whipple speaks
of its "being found far from any water supply."

Overturning a stone discloses a red-back at home. (Photograph by Dr. Miller.)

During a year's study of the species, they have been found
in both damp and dry places, within five feet of a pond's edge
and on rather dry, high slopes. In the daytime, they are always
concealed. In one case, near a spring, a large sawdust pile that
contained several planks furnished an abode for several sala-
manders.

One damp, dark day, when the rain was falling gently, three
medium-sized individuals were found on the surface of the
ground near stones from beneath which they had evidently come.
In the same locality, five were found beneath stones; two were
so placed that it seemed as if I had frightened them in before

334

M. ETHEL COCHRAN.

aware of their presence. This idea was strengthened by finding
one of the two in an ant's nest, a thing not found before. Of
several specimens kept in the laboratory, in vivaria, a few would
venture out of their retreats on dark days.

Rarely, if ever, are the salamanders found in treeless places,
but they seem to have equal preference for pine, birch or mixed
woods. All slopes, if shaded and of not too sandy a nature,
seem to provide suitable dwellings.

The natural crevices beneath stones are utilized by the red-backs. (Photo-
graphed by Miss Gulick.)

It seems that no holes are made by the salamanders themselves,
but that they utilize whatever is at hand. Allen noted holes,
but was not sure whether the salamanders made them or not.
In a pile of stones or in a sawdust heap, there are many natural
openings connecting in labrinthian fashion which are typical.
In the vivaria kept in the laboratory, the pieces of moss were
placed above a layer of sand and charcoal ; a small dish of water
was sunk to the level of the moss and the spaces between the
pieces of moss, between the moss and sand, and between the
"well" and sand were utilized by the red-backs as homes and
highways.

In one instance, where two specimens were found in rather
soft earth, a well-defined hole was followed — it led by a short

THE BIOLOGY OF THE RED-BACKED SALAMANDER.

335

gently-inclined path to a mass of rocks about a foot below ground
where there were natural openings in all directions. The hole
then appeared to be nothing but a natural opening worn smooth
by use. Five other places were carefully studied where natural
crevices seemed to be utilized.

The number of salamanders found under one object is variable;
the most ever found was ten and countless times has only
one been seen. There seems to be no evidence of their being or
living in pairs during fall and early spring.

DESCRIPTION.

Kingsley calls these salamanders "the smallest in the United
States." The largest specimen found measured 9.2 cm., while
from 7.8 cm. seems to be the average length. The tail is about

Red-backs. (Natural size.) (Photographed by the author.)

as long as the entire animal. Of two individuals, one, whose
total length was 9.2 cm., had a tail 4.9 cm. long, while one 7.8
cm. long, had a tail 4.2 cm. in length. Jordan gives their total
length as 3^2 inches (9 cm.). The following measurements
give some idea of the relative length and width of the body:
total length, 8.4 cm., tail 3 cm., costals 2 cm., width of head
5 mm., width of body 4 mm., dorso-lateral diameter 2 mm.

They are slender, with an almost cylindric body and a sharply
pointed cylindrical tail. The legs are thin and weak and the

336

M. ETHEL COCHRAN.

feet proportionately small. The anterior foot has three'toes and
an inner rudimentary one, while the posterior has four with one
inner rudimentary toe. The body is not always lifted from the
ground when the creature is walking — the tail never. Spelerpes
bilineatus seems to be more slender than erythronotus, and^when
a dusky salamander (Desmognathus fusca) is about, the dainti-
ness of a red-back makes it seem as clumsily built as an Ambly-
stoma.

The most distinguishing feature of this species is the broad
dorsal band of color that varies from bright, brownish-red to a
sort of pale, dead-leaf brown. Age seems to make no difference

Red-backs. (Natural size.) (Photographed by Dr. Miller.)

in the color, nor has any sex differentiation been observed in
fall and early spring. Two peculiar specimens found in company
with several normal ones, in widely separated spots, were of a
bright vermilion red, similar to the color of young Diemyctylus,
over all the dorsal surface; the tail alone was a dark brown. It
was probably a chance variation, as it seemed normal in every
other respect.

The sides of the body and much of the tail are of a dark gray-
brown finely flecked with silver spots that grow more numerous
toward the ventral side where the color is almost white, varied
with a few darker spots. The ventral part of the tail and legs
is much darker than the belly, while the dorsal surface of the

THE BIOLOGY OF THE RED-BACKED SALAMANDER. 337

legs is similar to the sides. The eyes are large for the size of the
head, dark, iridescent, and very prominent. The nostrils are
visible; the head is quite flat, the nose rounded, not pointed.

The skin is always moist and some amount of water is needed
in the environment, for a specimen kept in a dry pasteboard
box for a few hours was found in a much shriveled condition,
dead.

The throat is in constant vibration which at times is very
rapid ; various counts were taken from seventy to one hundred
and eighty a minute. The costals or costal folds, from sixteen
to nineteen, are very easily seen. There seems to be no ex-
ternal character during fall and early spring to distinguish the
sexes; in a few specimens dissected, the males were larger than

the females.

ACTIVITIES.

The red-backed salamander is almost entirely nocturnal.
Jordan and Kingsley mention this fact, while of Autodax lugubris
Hallow, which belongs to a genus close kin to Plethodon, Ritter
says it is entirely terrestrial and nocturnal.

This salamander is quick of movement. Often when first
discovered, it is found in a graceful, curled position, and seldom
moves unless touched. If annoyed, it glides rapidly, sometimes
into crevices or holes and sometimes into the moss or leaves,
where it lies quietly hidden. In this rapid gliding movement,
the tail is lashed from side to side, as if aiding the fragile feet
and legs.

It is a great climber — individuals in the laboratory climb up
the sides of a glass vivarium with great ease, using the moist,
adhesive abdomen, as well as the feet and legs. They have
been seen to do this on rather dark days as well as in the evening.

Kingsley mentions finding them out of doors on a "spear
of grass or coiled at the apex of a rachis of a fern at a distance
of from twelve to eighteen inches above ground."

There is a peculiar jumping habit common to the species—
if one is held in the hand, a foot or more above the moss floor
of a vivarium, it seems to gather itself together into a coil, and
using its tail as an aid, it springs lightly to the moss. The
whole process is indescribably quick. Kingsley says it "leaps
by sudden unbending of the coiled body like some caterpillars."

338 M. ETHEL COCHRAN.

The red-back's activities in the water are interesting. Jordan
says it "rarely or never enters water." Miss Whipple has
described it well: "Although certain lungless species may be
more or less aquatic, their activities, even in water, are terrestrial.
Various species will at first, when in an aquarium, swim to the
surface, then around and around the edge of the aquarium as
if seeking a means of escape, but at the instant active swimming
ceases, the body sinks clumsily and heavily to the bottom, where
they remain until disturbed or until another effort is made to
escape. Lungless forms show on the whole little power to adapt
themselves to aquatic life. Most are terrestrial in habit, some,
Plethodon cinereus and Pletliodon glutinosus, being found far
from any water supply.

"The nares close as soon as the animal is submerged in water
and remain so as long as the animal is in water. In a few cases,
there were attempts at a feeble bucco-pharyngeal respiration,
but even then the external nares were closed and water was
drawn in and expelled through the slightly opened mouth."

Salamanders in the laboratory seem able to endure extreme
cold, for in a vivarium the water was found with a thick coating
of ice, but the salamanders were apparently unaffected.

The latest date of finding a salamander out of doors in the
fall of 1909 was November 13. The first specimens this spring
(1910) were found on March 20, on a warm, southern slope free
from snow. At this time there was still much ice and snow in
the woods all about. In midwinter, Montgomery says the sala-
manders are found deeper in the ground.

FOOD AND FEEDING HABITS.

Regarding the frogs, salamanders, etc., Dr. Hodge says in
"Nature Study and Life," "with one or two exceptions . . .
they are all valuable insect destroyers, each for its peculiar
haunts; they should be generally protected and utilized as bene-
ficent forces in nature."

The red-back, like other amphibians, prefers live, moving
food. On several occasions, bits of meat placed in the vivarium
remained untouched for days and became covered with mould
while if a piece were moved before a salamander, it was taken

THE BIOLOGY OF THE RED-BACKED SALAMANDER. 339

eagerly. Small wriggling earthworms are enjoyed. In one
case, a small toad tadpole was offered to a salamander whose
attention was so attracted at once by the struggles of the tadpole
that it promptly ate it. A companion at the same time ate two
tadpoles. Small insects are captured and eaten with avidity.

A moving object is noticed very quickly. The salamander
will follow it fixedly for a moment, then, with gaze still upon it,
creep slowly after until close at hand, when a rapid spring is
made for it. Often it appears to have difficulty in swallowing
food — the eyes close as several slow gulps are made and some
time elapses before another portion is taken. No use of the hand
in feeding has been noted, as is commonly seen in toads and frogs.

As a means of determining the kind of food taken in nature
fifteen stomach-contents were examined. The salamanders
came from several widely separated districts.

In general, specimens taken in the morning showed fuller
stomachs than those taken in the afternoon.

The following specimens were found in the fifteen stomachs
examined: Ants of several kinds, beetles of several kinds, bugs,
caterpillars (cut-worm), centipedes, earthworms, fleas, flies of
several kinds, maggots (rat-tail), midges, mites, mosquito wig-
gler, plant lice, plant remains, pseudo-scorpion, slugs (Limax),
spiders, spring-tail (Thysanura) , sow-bug, wasp-like insects.

Below is given a list of the contents of five stomachs that may
be taken as typical.

A salamander captured May 17, 1910, at four P.M. in a saw-
dust heap near a spring, Auburn, Mass., contained: four snout
beetles, one yellow ant, one small spider and ten mites.

A salamander found under a stone in an ants' nest, April
28, 1910, in the afternoon at Auburn, Mass., contained: one
earthworm (2.75 cm. long), two mites, two small beetles, and
three maggots that were probably young ants.

A salamdander found at Holden, Mass., at eight A.M., on
April 27, 1910, contained: one earthworm (3.5 cm. long), one
centipede, one black fly, one spider, one cutworm, one ichneumon
maggot (probably from the cutwTorm), one pseudo scorpion,,
one sow-bug, and one unknown insect (which had eaten two small
snails).

34° M. ETHEL COCHRAN.

A salamander found on the afternoon of November, u, 1909,
at Gates Lane, Worcester, Mass., contained: three slugs (Limax,
I, 4, 5 mm. long respectively), one plant louse, one small worm,
and plant remains (probably accidental).

A salamander captured in Norcross Woods, Worcester, Mass.,
at eight A. M., May 9, 1910, contained: one earthworm (2.2 cm.
long), one brown spider (3 mm. long), two mites, one beetle,
one spring-tail, one mosquito wiggler, one Syrphus fly, and plant
remains.

ENEMIES.

Snakes undoubtedly are great enemies of the salamanders.
In a study made in Pennsylvania, Surface found that salamanders
formed a large part of the diet of some snakes. The ribbon
snake or striped garter snake (Thamnophis saurita), which be-
longs to the eastern United States and likes to live in shady,
narrow, watered valleys, preys upon beneficial batrachians. "In
four specimens containing food, four salamanders were found,
two of which were Plethodon cinereus; 37.5 per cent, of the food
of this species of snake was found to be salamanders." The
common garter snake was found to destroy toads and salamanders
among which was Plethodon cinereus. The water snake (Natrix
sipedon] and the grass snake (Liopeltis vernalis) also were found
to eat the red-back among other salamanders.

Ditmars states that the following snakes feed on salamanders,
among which it is reasonable to include the red-back, as their
haunts are much the same: ribbon snake (Eutemia saurita Linn.),
mud snake (Seminatrix pyg&a), brown or DeKay's snake (Stor-
eria dekayi Hoi.), green or grass snake (Liopeltis vernalis),
eastern and western ringed-necked snakes (Diadophis punctatus
Linn, and D. amabilis B. & G.).

A hungry, half-grown bull-frog ate a good-sized salamander
with evident relish and a purple grackle treated one like a worm,
beating it and breaking it with his bill before eating it.

PROTECTIVE DEVICES.'

Plethodon cinereus erythronotus, when annoyed, yields a color-
less, glutinous secretion from its tail. A Californian salamander,
Plethodon oregonensis, was found by Miss Hubbard to yield

THE BIOLOGY OF THE RED-BACKED SALAMANDER. 341

An upturned stone, showing cluster of eggs, manner of attachment and brooding
salamander. (Natural size.) (Photographed by Dr. Miller.)

342

M. ETHEL COCHRAN.

on the tail an abundant poisonous fluid which did not prevent
snakes from eating it. Easterly has written of the large granular
glands of this salamander which are poisonous, but to what he
does not state.

Baird speaks of the genus Plethodon whose skin exudes a
highly glutinous secretion, but he makes no statement of its
nature. Gadow says: "Numerous experiments have shown that
the poison of toads, salamanders and newts is capable when
injected, of killing mammals, birds, reptiles and even fishes,
provided of course, the dose be proportionate to the size of the
animal. Small birds and lizards succumb as a rule in a few
minutes; guinea-pig, rabbits and dogs in less than an hour.
This poison of amphibia is not septic, but acts upon the heart

Raised stone showing pair of red-backs with eggs. (About one half natural size.)
(Photographed by Dr. Miller.)

and central nervous system. Some authorities hold that the
poison is an acid, others regard it as an alkaloid."

The fluid from a red-back's tail is very sticky when placed
in the mouth; after a short time, a slight biting sensation is
felt for a few moments.

Several of the salamanders taken from their native haunts
had every mark of possessing a regenerating tail. In one case,
the tail had been cut or broken off nearly to the hind legs and
the bud of a new one had grown out about a quarter of an inch.

THE BIOLOGY OF THE RED-BACKED SALAMANDER. 343

Miss Towle found that Plethodon cinereus had the regenerative
power also in the limbs.

Of Plethodon oregonensis, Miss Hubbard says it practices
"autotomy only as a last desperate resource, and but in one
region (directly behind the anus)." But when in two or three
instances, the tail of a red-back has been held by forceps, the
tail has come off very easily and in no particular region. Those
found wild with incomplete tails appeared to have been broken
in no particular joint. This salamander has not been observed
thus far to snap or bite for protection.

INTELLIGENCE.

The red-backs are easily tamed, and will learn to eat food
when offered on a splint. They object to being handled even
after they have been in captivity a long time, probably because
of the unpleasant warmth and dryness of the hand.

Cluster of eggs attached to a stone. (Two thirds natural size.) (Photographed
by Dr. Miller.)

They appear to be very quick of hearing. Abbott, among
other statements about the salamanders, mentions "their
quickness of hearing." When undisturbed in their haunts, a
whistle, clap of the hands, or a speaking voice sends them away.
Abbott is of the impression that the salamanders give evidence
of greater intelligence than the toads or frogs. His efforts to
prove them possessed of cunning were not successful.

REPRODUCTIVE HABITS.

Baird writes of the genus Plethodon, "The eggs are deposited
in packages or aggregations, in moist situations, under stones

344

M. ETHEL COCHRAN.

Eggs of dusky salamander (two
clusters at the right), and red
back (one cluster at the left.)
(Natural size.) (Photographed
by the author.)

and logs, not however, in the water; and the larvae lose their
branchiae at a very early age."

Montgomery describes one finding of eggs: There were five
eggs under a stone in July with an adult female curled about on
guard. The eggs were large, enclosed in gelatinous envelopes,

and curled about a large, nearly
spherical yolk mass.

Miss Sampson says, "PJethodon
lays its eggs on land."

Ritter and Miller describe the
eggs of Aurodax lugubris Hallow, a
Californian salamander closely re-
lated to Plethodon. They are at-
tached singly to the under surface
of stones. In one instance a fe-
male with fifteen eggs was found
under a platform in front of a
barn, in dry earth next the foun-
dation wall. The eggs are laid in
July, are about 6 mm. in diamaeter and hatch in fifty days.

During the summer of 1910, eggs of the red-backed salamander
were found first on July 5. From then until the latter part of
August, eggs were found. They are in the ordinary habitats
beneath stones and logs. The stones are always those deeply
set into the ground in close contact with it, probably because
there the atmosphere is moist.

The eggs, from five to nine in number, are in grape-like clusters
attached to the under surface of stones or logs, unlike those of
Spelerpes and Autodax lugubris Hallow, which are attached singly.
One pedicel supports a cluster. This pedicel is tough, white, and
seems made of separate threads closely stuck together; it re-
sembles the tough, outer membrane covering each egg. There
is no unifrom arrangement of eggs in the cluster.

The egg of the red-back is spherical, 5 mm. in diameter, has
a prominent yolk and considerable transparent jelly. On the
outside is a thin, tough envelope to which particles of earth
and leaves readily adhere.

The embrvo is coiled about the volk. None were seen in

THE BIOLOGY OF THE RED-BACKED SALAMANDER.

345

young enough stages to note the first appearance of color, but
the red-brown of the dorsal stripe is visible at an early age.

Embryo red back, age unknown. (Enlarged about six times.) (Photographed
by Dr. Miller.)

With each cluster of eggs, there was always one or two adults.
Dissection showed that when there was one only, it was a female,

Embryo red-back taken from egg some days before time for hatching, showing
the three pairs of gills. (Enlarged three times.) (Photographed by the author.)

when two, they were male and female. There is, accordingly, no
evidence that the male takes any active part in the care of

346

M. ETHEL COCHRAN.

the eggs in this species. The egg cluster is always so attached
that one of the adults may coil its body about it. Sometimes
a crevice beside a rock is utilized, the eggs, attached by the
pedicel to a root or stone above, swinging loose. This brooding
by the adults doubtless serves the double purpose of providing
moisture and securing protection from insects.

Several attempts were made to transfer a parent with a cluster
of eggs to the laboratory, that the development and hatching

From left to right: four embryos from the same cluster of eggs, one removed from
the egg, one hatched about twelve hours, one about six hours and one about twenty-
four hours. (Magnified three and one half times.) (Photographed by the author.)

might be observed. Damp moss was tried but it invariably
moulded. A small jar kept moist with a wad of water-soaked
cotton proved no better. Then, because the air of the laboratory
seemed so disastrous, it was suggested opening the bottle, in
which the eggs wrere brought from the field, upside down into an
overturned sterile jar. Within the jar, the eggs were placed
on a strip of filter paper that was so placed on a piece of glass

THE BIOLOGY OF THE RED-BACKED SALAMANDER. 347

above a dish of water that it was constantly wet. No mould
appeared, the moisture was constant, and the eggs developed.

An embryo just out of the egg is 2 cm. long; the gills are present
but not quite so large as they are a day or two earlier. The head
seems larger compared with the rest of the body, than in the
adult stage. There is a large quantity of yolk visible which
persists for several days.

In this cluster of eight eggs, five were allowed to hatch norm-
ally, three were opened for pictures. The hatching of the
five covered a period of twenty-four hours. The actual process
of hatching was not observed.

There is a small amount of jelly and membrane left; this was
noted both in the laboratory and in the field where nests were
under observation. Beneath one stone, the pedicel and gela-
tinous substance was observed after the eggs had hatched. It
had turned dark and was tough and leathery.

The embryos just out are very active and show several char-
acteristics of the species. Salamanders, less than a day old,
avoid water, wriggle if annoyed and cling to the surface of the
glass jar.

The gills shrivel rapidly; at the end of twenty-four hours,
there are but mere stubs.

SUMMARY.

In summing up the results of the observations reported above,
the writer makes the following statements:

1. The red-backed salamander is found throughout all the
eastern United States and Canada.

2. It lives beneath stones and logs, ranging from rather dry
to wet places.

3. It is a small salamander, usually about three and one half
inches in length.

4. It is almost entirely nocturnal. It is quick of motion
and a climber; it cannot live in water.

5. The food consists mostly of live insects, larva?, worms
and spiders.

6. Snakes and frogs are probably its greatest enemies.

7. The protective devices are autotomy and the secreting
of a viscous fluid from the tail.

348 M. ETHEL COCHRAN.

8. The eggs are laid beneath stones in July and early August.
One of the parents, usually the female, broods them. The young
hatch with three pairs of gills which shrivel in about a day.

My acknowledgments are gratefully given to Dr. Clifton
F. Hodge for suggesting the problem, for his active interest
and help during the year, to Miss Louise Gulick for aid in col-
lecting specimens and photographic work, to Mr. William
T. M. Forbes for help in identifying insects, and to Dr. Newton
Miller for his excellent and painstaking work in photography.
I aslo wish to thank the many people, who by their interest,
advice and criticism have aided in the work.

BIBLIOGRAPHY.
Abbott, C. C.

'84 Intelligence of Batrachians. Science, Vol. III., no. 50, Jan. 18, 1884, pp.

66-67.
Allen, Glover M.

'09 Notes on the Reptiles and Amphibians of Intervale, N. H. Proc. Boston

Soc. Nat. Hist., Vol. XXIX., 1909, pp. 62-75.
Baird, Spencer

'69 Inconographic Encyclopedia, Vol. II., New York. 1869,
Boulenger, G. A.

'82 Catalogue of Batrachis Gradientia, London, 1882, p. 53.
Cope, E. D.

'89 Batrachia of North America, Bull, of U. S. Nat. Mus., no. 34, 1889.
DeKay, J. E.

'42 Nat. Hist, of N. Y. Zool. Part III., Reptiles and Amphibia, Albany, N. Y.,

1842, pp. 75.
Ditmars, Raymond L.

'07 The Reptile Book, New York, 1907.
Esterly, C. O.

The Structure and Regeneration of the Poison Glands of Plethodon Ore-
gonensis. Univ. Cal. Pub. Zool., Vol. L, no. 7, pp. 227-228.
Gadow, H.

'01 Cambridge Natural History. Vol. VIII., Amphibia and Reptiles. London,

1901.
Hodge, C. F.

'02 Nature Study and Life. Boston, 1902, pp. 296.
Holmes, S. J.

'06 Biology of the Frog. New York, 1906.
Hubbard, Marian E.

'03 Correlated Protective Devices in Some Calif ornian Salamanders. Univ,

Cal. Pub. Zool., Vol. VI., no. 4, pp. 157-170, Nov. 5, 1903.
Jordan, D. S.

'04 Manual of Vertebrates. Chicago, 1904.

THE BIOLOGY OF THE RED-BACKED SALAMANDER. 349

Kingsley, J. S.

Standard Nat. Hist., Vol. III., p. 313-
Montgomery, T. H.

Peculiarities of the Terrestrial Larva of the Uredelian Batrachian, Ple-

thodon cinereus (Green). Proc. Acad. Nat. Sci. Phil., Vol. VIII., pp. 503-

508.
Ritter, W. E. and Miller, L.

'99 A Contribution to the Life History of Autodax Hallow, a Californian Sala.

mander. Amer. Nat., Vol. XXXIII., no. 393, Sept., 1899, pp. 691-704*
Surface, H. A.

'06 The Serpents of Pennsylvania. Mo. Bull, of Div. of Zool. Penn. Dept-

of Agriculture, Sept. i, 1906, Harrisburg, Pa., pp. 208.
Towle, Elizabeth W.

On Muscle Regeneration in the Limbs of Plethodon. Biol. Bull., no. 2.

pp. 289-305.
Whipple, Inez L.

'96 The Upsiloid Apparatus of Uredeles. Biol. Bull., no. 10, 1896, pp. 255—297..

ON THE DISTRIBUTION AND MODE OF OCCURRENCE

IN THE UNITED STATES AND CANADA OF

CLINOSTOMUM MARGINATUM, A

TREMATODE PARASITIC IN

FISH, FROGS AND BIRDS.

HENRY LESLIE OSBORN.

CONTENTS.

1. Geographical Distribution of Clinostomum marginatum 350

2. On the Cysts in the Black Bass 354

3. Behavior of Worms Liberated from Bass-cysts 357

4. On the Cysts in the Frog 360

5. On the Occurrence of the Worm in the Heron 362

i. GEOGRAPHICAL DISTRIBUTION OF Clinostomum marginatum.
The fluke referred to in this paper was first noticed in this
country in 1856 by Joseph Leidy, in the intestine of the pike of
the Delaware and in cysts attached to the gills of the sun-fish
(Enpomotis vulgaris} near the city of Philadelphia. Leidy applied
the name Clinostomum gracile to it. This generic name after
many years of non-use has come into current usage since the late
revision of the Trematodes which has received such impetus
from the work of Looss ('oo) who recognizes this name and Braun
(!oo) who in a revision of the genus in 1900 recognizes eight species
of the genus Clinostomum, among them C. marginatum. In
1879 R. Ramsey Wright found cysts attached to the gills, branch-
iostegal membranes and pectoral fins of the yellow perch (Perca
flavescens] at Toronto, Ontario, which he recognized as being iden-
tical with the Clinostomum gracile of Leidy and called Distomum
gracile. His observation extended the geographical range of
the species to the St. Lawrence River system. In the same paper
Professor Wright reported the finding of D. gracile in the bittern
(Botaurus minor] a fish-eating bird which was the first information
as to the definitive host of the worm. Looss in 1885 published
an account of the structure of a fluke which had been found en-
cysted in the muscle of a silurid fish from Porto Rico. He gave

350

DISTRIBUTION OF CLINOSTOMUM MARGINATUM. 351

to this fluke the name Distomum reticulatum in allusion to the
very peculiar excretory collecting apparatus. Subsequent
writers, including Looss himself (Looss '99), have referred this
worm to the species Clinostomum marginatum, and while there
are certain points in which it differs from C. marginatum we may
for the present so recognize it. This observation however takes
the fluke a long way out of the region in which we otherwise
knowr the animal and seriously embarrasses the attempt to deal
with the geographical range of the species. In 1888 Leidy re-
ported a fluke from the striped bass (Roccus lineatus) which he
designated Distomum galactosomum. His account of the or-
ganization of this form touches on the form of the oral end of
the body and the network of collecting vessels in the body-wall,
twro features of such peculiarity that we cannot doubt but that
the subject of his study was none other than Clinostomum mar-
ginatum, which he had described in 1856 and evidently forgotten.
His account located the worm in a marine host for the first time,
unless Looss' silurid was marine.

In 1895 MacCallum ('99), of the University of Toronto,
found a trematode encysted in the muscles of the frog which he
regarded as identical with the ones which Wright had found in
the fish and heron and referred to the D. gracile of Leidy. In
his paper of 1899 MacCallum calls the worm C. heterostomum.

His description, while differing in some respects from my
material, leaves no doubt but that our material is identical.
I shall adopt the name C. marginatum in place of the name used
by MacCallum following the lead of Braun ('oo) in his revision
of the group. He also found the same species in the throat of
Ardea herodias at Danville, Ontario.

In 1898 Linton reported and figured this species under the
name of D. gracile from the sun-fishes (Eupomotis pallidus and
Ch&nobryttus gulosus) of Kansas City, Mo. This observation
extended the range of the worm to a new river system, the Mis-
sisippi and gave us the most western point of its distribution.
In 1900 specimens of this fluke were found by myself at Nebish,
Mich., encysted in the muscular tissues of the small-mouthed
black-bass (Micropterus dolomieu], also, though less frequently,
in the muscles of the yellow perch (Percaflavescens). They were

352 HENRY LESLIE OSBORN.

also found later in the same year in the throat of the little blue
heron (Ardea herodias). Nebish is a small camping place situated
on the St. Mary's River, about twenty-five miles below the
outlet of Lake Superior. A new locality was thus added to the
distribution of the worm though one in the same river system
as the Canadian ones.

In 1903 I received, through the kindness of Professor Linton,
some pieces of the small-mouthed black bass infected with this
parasite. The fish had been taken by the Rev. J. H. Young at
Troy, Ohio, on the Miami River, a tributary of the Ohio, and
part of the Mississippi system. Some of the worms were encysted
in muscle, but others were located in the skin on the internal
aspect of the branchiostegal membranes. Their location is
shown in Fig. I of this article.

In the same year my attention was called by Dr. W. S. Nicke-
son, of the University of Minnesota, to certain flukes which he
had found in frogs, and they were at once recognized as specimens
of this species. At about the same time I began to notice them
in the frogs of this region brought to the laboratory for use as
biological material. The details of the occurrence of the fluke
in this region is still under investigation in connection with a
study of the parasites of the frogs of this vicinity, so that a fuller
account of this part of the subject is reserved for a later occasion.
It is important, however, to place on record the occurrence of the
fluke in this region, as it extends the range of the parasite con-
siderably beyond formerly recorded limits.

The latest appearance of this form in literature is the record
of its recognition in the yellow perch of the Montreal markets
by Stafford in 1904, where it is given Leidy's designation Clino-
stomum gracile.

The foregoing facts show that this species is very widely
distributed in this country, having been recognized at Kansas
City on the West and at St. Paul, Michigan, and as far east as
Montreal. It has been seen as far north as St. Paul and as far
south as Philadelphia and in Ohio in the center of this territory.
The occurrence at Porto Rico is anomalous and cannot be con-
sidered until more information is forthcoming. In view of the
wide distribution and large size of this species, it seems strange

DISTRIBUTION OF CLINOSTOMUM MARGINATUM. 353

that we are not in possession of a particle of information regarding
the early stages of its life-history. The worm is fully developed
when it is encountered in the fish or frog and all that' is needed
for its complete maturity is that the vitellaria shall become
active, which they do as soon as the parasite reaches the
heron. The details of structure of the worms of the bass have
been determined both in total preparations and in serial sections.
In the smallest individuals the entire organization is developed.
It follows from this fact that there are three hosts involved in the
life-history of this species of which the fish or frog is the second
and the heron the final one, the first being unknown. The role
of the fish or frog does not appear to be one of importance from
the view point of the ontogeny, but seems merely to be for the
purpose of getting the wTorm to the heron. It thus would appear
that the primary host is not one which serves as food to the heron
but does serve as food for the bass. This would fit the case of
insects as well as any group of invertebrates.

In view of the fact that this parasite infects the edible portion
of such important game and food fishes a knowledge of its life
history is particularly desirable. While the parasite is fortu-
nately one which is not injurious to the human system, at the
same time its presence in the bass and perch unfit them for the
table. It is clear that the worm is already widely distributed
and is a menace to fish-culture, which at any time may become
of great importance. It seems likely that if one were able to
take up the problem of this life-history under favorable con-
ditions for instance at some small body of water where the
infected fish were abundant and to follow it up at various seasons
the missing data could be obtained.

The data given in the foregoing paragraphs are summarized
in the following table.

354

HENRY LESLIE OSBORN.

Date.

Observer.

Name Used.

Host.

iSeat of Infection.' Locality.

1809 Rudo4phi.

Distomum

marginatum.

1856 Leidy.

Clinostomum

Esox.

intestine. Philadelphia.

gracile.

1879

Wright.

Distomum

Perca

branchioste- Toronto, Ont.

gracile.

Jlavescens.

gal mem-

branes.

1879

Wright.

Distomum

Botaurus

mouth.

Toronto, Ont.

heterostomurn.

minor.

1885

Looss

Distomumi

Silurid fish.

muscle

Porto Rico.

reticulatum.

encysted.

1888

Leidy.

Distomum

Roccus

galactosomum.

lineatus

Philadelphia.

i895

MacCallum.

Distomum

Frog.

muscle.

Toronto,

gracile.

Ont.

1897

MacCallum.

Distomum

Ardea hero-

mouth.

Toronto,

heterostomum.

dias.

Ont.

1898 iLinton

Distomum

Eupomotis

pectoral fin.

Kansas City,

gracile.

pallidus.

Mo.

1898

Linton.

Distomum

Chcenobryttus

roof of

Kansas City,

gracile.

gulosus

mouth.

Mo.

1901 ^Osborn.

Clinostomum

Micropterus

muscle.

Nebish,

marginatum.

dolomieu.

Mich.

1901 Osborn.

Clinostomum

Perca

muscle.

Nebish,

marginatum.

Jlavescens.

Mich.

1903 Osborn.

Clinostomum

Rana

ccelom wall.

St. Paul, Mn.

marginatum.

virescens.

1903

Young,

Clinostomum

Micropterus

branchioste-

Troy, Ohio.

marginatum.

dolomieu.

gal mem-

branes.

1903

Young.

Clinostomum

Micropterus \ muscle.

marginatum.

dolomieu.

1904

Stafford.

Clinostomum

Perca

gills.

Montreal,

gracile.

Jlavescens.

Ont.

2. MODE OF OCCURRENCE IN THE BLACK BASS
AND PERCH, AT NEBISH, MICH.

The observations now to be described were made at Nebish
in 1901 and repeated the following year. Nebish is a small
settlement on the American side of the St. Mary's River about
midway between Lake Superior and Lake Huron. At that time
bass and perch were fairly numerous, beginning in the early
part of September and the last of August. As they were not
caught much earlier it was the local opinion that the fish
migrate to the grounds at Nebish from elsewhere, and this is
borne out by the fact that of late years, as a consequence of the
work done in improving the channel of the river for navigation,
the bass have ceased to come in the fall and the "fishing" is
destroyed. I examined as many of the various kinds of fish and
invertebrates as possible during my short stay at Nebish in hope

DISTRIBUTION OF CLINOSTOMUM MARGINATUM. 355

of getting some information which would lead up to the discovery
of the first host. The only forms on which I recognized the
parasites were the bass and the perch. Other fishes which were
investigated with wholly negative results were the grass-pike,
which is very often caught there, and the sun fish, Eupomotis—
also very common. Various gasteropods and insects were
examined but without any result. I have thus far not been able
to obtain any clue to the source of infection of the fish.

The percentage of infected individuals found among the bass
was very great. I do not happen to have any statistical data
bearing on this point, but the parasite was found in nearly every
bass submitted to examination, while in the perch it was much
more rare. The cysts were very easily seen, being large, opaque
and very creamy white, in marked contrast with the darker
semi-translucent muscular tissue in which they lie imbedded.
When the fish were skinned in preparing them for cooking the
cysts walls were often torn open and the conspicuous worm seen
moving on the surface of the meat. The cysts were found in all
parts of the lateral muscles, deep and superficial, dorsal and
ventral and headwards and tailwards. There was no evidence
from their location bearing on the question of the mode of entry
which had been adopted by the worms, but the observations of
various others who have reported the worm from the gills, roof
of mouth, branchiostegal membranes and pectoral fin would
indicate that the worm enters the fish in the head region. If so
it is difficult to imagine how they reach the places where we find
the cysts if they are as large as we find them when they enter.
On the other hand, if they enter as small and immature stages
we should surely be able to find some evidence of it. Some of the
worms should be less fully developed than others, or some of the
waste products of the chemical processes involved in growth
should be recognizable. So that we are unable to find a solution
to this problem with the data at present available.

The number of cysts in single individuals varied greatly. The
minimum number found was seven and the maximum number
was more than one hundred. The appearance of one of the cysts
in situ in the muscle is shown in Fig. 2. They are oval or spher-
ical. They are smaller than those reported by Looss ('85) from

356 HENRY LESLIE OSBORN.

the fish which he examined, being 1.3-3 mm. long as compared
with 3.5-4 mm. in length according to Looss.

The cysts occupy a space in the endomysium among the fibers
of the muscle of the host. As we do not possess many detailed
accounts of the structure of trematode cysts or their relation to
the host tissues, an account of the facts found in this case may
be of some interest. The following methods were employed.
The cyst was carefully separated by teasing away the surrounding
muscle fibers of a piece which had been fixed and hardened in
suitable reagents. The cyst was then cut open and the enclosed
worm removed, after which pieces of the wrall were submitted
to the action of various staining reagents and mounted for
microscopical study. In other cases the muscle and enclosed
cyst was sectioned serially. In such a series we have transverse
sections of the wall and the surrounding muscle tissue, and at
the ends of the series there are tangential sections which can be
compared with the flat preparation just mentioned. Figs. 3
and 4 are low- and highpower views of such sections, and they
show the structural factors involved. The muscle fibers are
bent around the cyst as seen on the sides of the section; those
in the center which seem to end abruptly at the surface of the
cyst do not in reality do so, but are cut off as they run out of the
plane of the section. The worm wholly fills the cavity of the
cyst, the very small space which is seen in the figure being easily
accounted for by the shrinkage of the worm. The structure
of the cyst is seen in Fig. 4 to be merely a membrane produced
by the condensation of fibrous tissue. There are the usual wavy
fibers, thinning out as they reach the endomysium, and contin-
uous with the fibers which fill in the spaces between the muscle
fibers, and between the fibers the customary flattened connective
tissue nuclei. The flat preparations and tangential sections
show very clearly that the cyst is supplied with a capillary
network derived from the vascular supply of the muscle, and
these capillaries and their contained blood corpuscles can be
recognized in the section of the wall as shown in Fig. 4. Looss
states ('85) that his observations led him to conclude that the
cyst is produced by the host and then lined by a second envelope
produced by the worm itself. "Darunter befinded sich eine
zweite Hulle, die, anscheinend ein erhartetes Secret, von dem

DISTRIBUTION OF CLINOSTOMUM MARGINATUM. 357

Wurme um sich erzeugt wurde, und in dieser liegt eingerollt
und zusammengeschlagen das Thier selbst." As to this inner
lining found by Looss no traces of it are found in my material,
and the cyst is wholly a product of activities on the part of the
host tissues. The cyst may be looked upon as a defense made
by the host against the parasite. It would be interesting to know
the nature of the stimuli which have caused this reaction on the
part of the supporting tissues of the muscle whether they are
merely mechanical and acting in the form of pressure or whether
they are chemical or both. The excessive formation of connec-
tive tissue and the growth of a definite and extensive capillary
network have resulted from the presence of the worm. In the
somewhat analogous case of the cavities produced in oak leaves
by the sting of certain hymenoptera for the purpose of housing
their larvae, the normal growth of the leaf is turned aside suffici-
ently to produce cavities for the larva, walled with a material
produced by the leaf from its own substance. In that case the
stimulus would seem to be chemical. Whether it is so in this
case cannot be determined until we are in possession of more
facts connected with the introduction of the parasite to the tissue
of the fish.

In the bass, as already stated, the worm entirely fills the cavity
of the cyst. In order to determine its position within, several
fixed and hardened cysts were carefully dissected out of their
place in the muscle and the wall carefully removed under a dis-
secting microscope. The worms thus found bent and cramped
within were macerated for an hour or less in tepid water till they
became flexible and could be uncoiled as shown on Fig. 5. They
were thus found to be bent twice, both times with the ventral
surface outward, first on the level of the meeting of the first and
second body thirds, and again on the level of the second and third
body thirds. The first of the bends is in the median plane of
the body and brings the dorsal surfaces in contact; the ventral
sucker is thus exposed and serves to identify the surface posi-
ively. In making the second bend the body does not bend
dorsally at once but first there is a twist which brings the dorsal
side over the edges of the first and second thirds of the body
and then the dorsal surface of the last body third lap^ over these

358 HENRY LESLIE OSBORN.

edges. The worms and cysts in the frog are very different from
this. They will receive attention later.

3. BEHAVIOR OF WORMS LIBERATED FROM THE BASS CYST.

The living worms are easily liberated artificially by cutting
into the cyst wall with a sharp instrument. As soon as a small
opening has been made the worm begins to extrude through it
almost as if it had been confined under some degree of pressure.
It continues to make active movements of extension and contrac-
tion which soon result in freeing it from the cyst. It continues
to move actively after gaining its freedom. In the course of
nature the worm on emerging from the fish cyst would find itself
in the stomach or intestine of a fish-eating bird; those which
\vere removed artificially were received and kept in shallow ves-
sels of fresh wTater where their activities could be watched.
The surroundings in which they thus found themselves were
quite unnatural and perhaps it was for this reason that they were
so very active. The movements did not result in locomotion,
though they produced constant changes in the form of the body,,
but the worm did not progress. Some trematodes adhere very
forcibly by their suckers to any object with which they come in
contact but there was no attempt made on the part of these to
use the suckers for that purpose. Some of the various move-
ments were so irregular and indefinite as to preclude the possi-
bility of grouping them under any general head, but there were
others which were definite and constant enough to be classed
under either of two groups which I have called "poses" from the
fact that in each case a series of movements took place culmin-
ating in a certain bodily form or attitude which was only momen-
tarily maintained and after which the body relaxed and fell back
to its ordinary resting posture. These two poses were seen often
enough to justify the conviction that they are a constant feature
of the behavior of the worm and so merit a detailed description.
The twro poses will be designated the "suctorial pose" and the
"swimming pose." Sometimes the same pose is repeated several
times in succession, at other times the two alternate irregularly.

The suctorial pose is represented in Fig. 6 as seen in the living
animal and in Fig. 7, a view of a sagittal section of the anterior
end of the body of an animal which happened to be caught in

DISTRIBUTION OF CLINOSTOMUM MARGINATUM. 359

this pose at the moment of fixation. The region of the body
chiefly active in the production of this pose is the part anterior
to the ventral sucker. Ordinarily this part of the body has the
shape of an obliquely truncated cylinder. In assuming this
pose the worm draws the end of the body back into the interior,
the side walls being thicker and acting as the rim of a sucker
projecting considerably beyond the level of the center. This
inversion of the oral end of the body is brought about by. the
contraction of fibres of the parenchyma muscle system which
run longitudinally in the body. Some of these fibres are shown
in Fig. 7. Their contraction pulls the center of the oral end back
and the margin is left projecting. If this action were to take
place at a moment when the oral end was in contact with a soft
surface, for example, such a surface as the mucous coat of the
stomach or throat of the heron, then the soft substance of the
host would be sucked into the cavity of this sucker and a powerful
adhesion of the parasite to the host would result. A worm liber-
ated naturally from the cyst would meet conditions which would
furnish responses to this movement, so that instead of being
merely a momentary pose it would be useful and so continued.
On the other hand, in the case of the artificially liberated worm
such stimuli being absent the worm returns to its customary
form. The sucker thus formed is additional to the two usual
trematode suckers. We may suppose that in as large a parasite
as Clinostomum and one which lives in such an exposed place
as the gullet where large masses of food are pressed against
it additional adhesion would be needed to protect it against being
dislodged by the pressure of the food as it is being swallowed.
This striking movement had thus evidently a purposive char-
acter, as can be seen if we take into account the environment
for which it has been developed.

The "swimming pose" is shown in Fig. 9. In this case the
body is made broader and flatter than usual, the ventral surface
becoming somewhat concave. The margins of the posterior
part of the body are reduced to fine sharp lines like fins which
do not extend into the front part of the animal but crossing it
ventrally converge toward the ventral sucker. The line shown
in the figure is not visible in any specimens of the worm after
preservation and is only a momentary structure. The flattened

360 HENRY LESLIE OSBORN.

form of this pose too is only momentary, and no preserved speci-
mens show it. In all cross sections of the worm the body is
elliptical and has a thickness of nearly half its breadth.

The transverse parenchyma muscles are the ones which are
used in the swimming pose, the longitudinal muslces being at
rest. Alternate contractions of the transverse and longitudinal
muscles would throw the animal first into one, then into the other
of these positions. The observation of certain leeches furnished
the suggestion for calling this the swimming pose. In the leeches
the margins of the body are thinned in this way and reduced to a

•

thin lateral fin and the worms swim rapidly through the water.
In the specimens of Clinistomum which were under observation
no vibrations of the body were made and the pose was not turned
to any account. It is possible that in nature this body form
would be adapted to progression through the chyle and chyme
of the bird and would be the means by which the worm reaches
the throat toward which it is travelling from the stomach or
intestine. It wrould be very interesting to experiment with
these worms by removing them from cysts and placing them on
surfaces as much like the avian mucous membranes as possible
so as to ascertain whether these poses are as adaptive as they
seem to be.

4. ON THE CYSTS IN THE FROG.

The first mention of the frog as a host of Clinostomum is by Mac-
Callum ('99), who reports it from the pectoral muscles of frogs
found in Ontario. We have been finding it now for several years
in the frogs caught in the vicinity of St. Paul which are used in
our biological courses.1 The mode of infection of the frog is
different from that of the bass. The cysts, instead of being
located in the muscular tissue, are seldom found there, but are
in the peritoneal lining of the ccelom or in the lymph spaces
between the skin and the muscles of various parts. Text-figure I
is drawn from the most considerably infected specimen which
I have met. The cysts are located in the ccelomic cavity;

1 It is most probable that frogs and fishes in many parts of this country are in-
fected by this parasite. Any information as to localities elsewhere where it has
been noticed would be very gratefully received by the writer. A study of the re-
lation of the worm to the frog is now in progress, and it is hoped that clues to the
early history of the worm may be obtained through its connection with the frog.

DISTRIBUTION OF CLINOSTOMUM MARGINATUM.

361

especially in the region dorsal to the heart. They are found here
most frequently. They are also found in the ventral wall of
the coelomic cavity, in the floor of the mouth between the muscles
and the skin and even in the
leg between the muscle and
the skin. In no case have
the cysts been found in the
frogs of this region in the
interior of muscular tissue
surrounded by the fibers as
they are so characteristi-
cally found in the fish. I „. ,../,
have as yet not been able
to find any reason for this
difference in the location of
the cysts with regard to the
host, but the position of the
worm within the cyst is
considerably affected, as
will be explained fully in a
moment.

The ratio of infection of
the frog does not appear to
be as high as it is in the
fishes studied at Nebish.
The percentage of frogs
not infected is very high,
indeed a large number can
be gone through often with-
out any cysts being found.
The number of cysts in a
single individual is usually
not large; the frog figured
is quite exceptional in that
respect. Data for an exact
statement on these points are not at present worked out.

The relation of the worm to the cyst and to the underlying
muscle of the body wall is shown in Fig. 9. The muscle fibers

View of ventral surface after the removal
of the ccelomic viscera, showing the numer-
ous encysted flukes in place. /, lung; mu,
muscles of floor of mouth; p, flukes in place;
v, ventral muscles of coelomic cavity; v sk,
skin of ventral ccelomic wall. X2/3 natural
size.

362 HENRY LESLIE OSBORN.

run along below the cyst, their course entirely uninfluenced
by the presence of the cyst. A comparison of Fig. 9 with
Fig. 2 shows the difference of habit in this respect at a glance.
The preparation from which Fig. 9 was drawn shows five cysts in
an area half an inch wide by a quarter of an inch across. It is
removed from the right half of the body wall of the frog shown on
page 361. The whole piece was stained and mounted in balsam
and one of the cysts drawn in Fig. 9. The cyst is much larger
than the worm and there are spaces within unoccupied by the
worm. The worm in most cases is bent once, only the ventral
surface being turned outward. In some cases the bend is not
in the center, and then the longer end may be bent slightly as
in the one shown in Fig. 9. The frog cysts show the presence
of a rich network of capillaries spread out over their surface.
These are derived from vessels which come from large and con-
spicuous vessels running in the space between the peritoneum
and the muscle layer. The ordinary surface of the peritoneum
does not possess these capillaries, which are evidently a growth
developed as a result of the presence of the cysts.

The specimens of Clinostomum seen in the frogs are all virtually
fully matured, having the full size of the heron specimens. -Their
inner organization too is identical in appearance with that of
specimens from the fish, and excepting as to the vitellaria with
that of the heron. There is thus no room for an hypothesis as
to the frog being the first host and the medium by which the
fish is reached, which hypothesis we might be tempted to frame,
knowing that the frog is one of the foods of the bass, and other
predaceous fishes. We are thus left to suppose a first host,
possibly an invertebrate, followed by a second intermediate host,
the fish which serves as a medium of transmission from the un-
known first host to the heron. The case of a second host which
serves to transmit a parasite from a first host where it develops
to a final host where it matures and where sexual reproduction
takes place has been recognized for a number of trematodes.
A list containing 28 species is given by Braun (93, pp. 864-866),
in which this species is not included. In some of these the frog
or one of the fishes serves as the medium by which the final
host is reached.

DISTRIBUTION OF CLINOSTOMUM MARGINATUM. 363

5. ON THE OCCURRENCE OF THE WORM IN THE HERON.
A single specimen of the heron sent me by express from Nebish
in October happened to be infected with this parasite and fur-
nishes all that I have myself observed on this point. The bird
had been shot and had been dead a day or two before it came into
my possession. The worms were still alive and suitable for
fixation and preservation for histological study. The worms
were found in considerable numbers adhering to the wall of the
throat by means of the anterior end used as a sucker in the
manner referred to above. On being removed they were found
to be mobile but no data as to their movements were recorded.
The stomach and intestine of the heron were carefully examined
to ascertain whether those organs were infected or not. No
worms were found there. The stomach however contained the
remains of the bodies of five fishes in various stages of digestion
all of which were too decomposed for recognition with entire
certainty but their elongate shape, their size and the fauna of
the region made it virtually certain that they were sepcimens of
the yellow perch. This observation furnishes the evidence of a
connection between the fish and the final host. The liberation
of the worms in the stomach and their migration of the worms
in the stomach and their migration and adhesion to the wall of
the throat is of course the first event of their life in the avian
host followed by the production of eggs and their passage to
the uterus where we find them in great numbers in the specimens
taken from the heron. It is not exactly clear why the eggs are
not passed directly from the body of the heron after the manner
general in the flukes. Here however there is a large uterus into
which the eggs pass and where they accumulate, for reasons as
yet unknown. The eggs when they are expelled from the fluke
must pass down the alimentary canal of the heron and fall with
the feces in the water or less probably on the ground, where they
make their way to the first host. The eggs in the uterus show
no signs of development so that the ontogeny remains for the
present a wholly unknown quantity.

SUMMARY.

i. The first host and early life-history of C. marginatum are
entirely unknown.

364 HENRY LESLIE OSBORN.

2. It passes a period of quiescence encysted in the muscle
tissue of various predacious food fishes and in the lymph-spaces
of frogs, in various localities ranging from Missouri through Ohio
to Pennsylvania, and north as far as Minnesota, Michigan and
Ontario, Canada.

3. It finally reaches an avian host which is some fish-eating
bird, such as the bittern or the heron.

4. The cyst in the fish is a connective tissue structure pro-
duced from the endomysium of the host with a special vascular
equipment.

5. The living worms on being artificially liberated from the
cysts perform characteristic movements adapted to finding
locations in the final host.

BIOLOGICAL LABORATORY,
HAMLINE UNIVERSITY,
SAINT PAUL, MINN.,
March 17, 1911.

LITERATURE.
Braun, M.

'93 Bronn's Klassen u. Ordnungen. Bd. IV., pp. 306-924. Trematoden.
Braun, M.

'oo Die Arten der Gattung Clinostomum Leidy. Zool. Jhrb. Syst., XIV.
Leidy, Jos.

'85 A Synopsis of the Entozoa, etc., no. 31. Proc. Ac. Nat. Sci. Phila., pp.
42-58.

'87 Notes on some Parasitic Worms, Proc. Ac. Nat. Sci. Phila., p. 24.
Linton, Edwin.

'98 Notes on the Trematode Parasites of Fishes. Proc. U. S. Nat. Mus., XX.,

PP- 597-548.
Looss, A.

'85 Beit, zur Trematoden, Distomum palliatum n. s. u. D. reticulatum n. s.

Zeitf. w. Zool., XLI.
'99 Weit. Beitr. z. Kenntn. der Trematoden Fauna Aegyptens. Zool. Jahrb.

Syst., XII., pp. 521-748.
MacCallum, W. G.

'99 On the species Clinostomum heterostomum. Am. Jour. Morphol., XV. ,

pp. 697-710.
Osborn, H. L.

'01 Structure of Clinostomum. Science, XIII., p. 378.
Stafford, J.

'94 Trematoda from Canadian Fishes. Zool. Anzeig., XXVII., pp. 681-694.
Wright, R. R.

'79 Contrib. to Americ. Helminthology, No. i. Proc. Canad. Inst., I.

366 HENRY LESLIE OSBORN.

EXPLANATION OF PLATE I.

FIG. i. View of the inner aspect of the right half of the lower jaw of a specimen
of the small-mouthed black bass, from Ohio; showing the white cysts of C. mar-
ginatum at the bases of the branchiostegal rays, and black spots of an unidentified
nature. Natural size.

FIG. 2. A cyst of C. marginatum in place among the muscle fibers of the bass
from Nebish, Mich., from a glycerine preparation after teasing off part of the mus-
cular tissue.

FIG. 3. View of a section passing through a cyst in place in the muscle tissue
of the bass, from Nebish, Mich. Camera lucida, X24 diameters.

FIG. 4. View from a section of a cyst in a plane transverse to the muscle tissue,
showing in detail the histological structure of the wall of the cyst; from bass from
Nebish, Mich. Camera lucida. X340 diameters.

FIG. 5. Drawn from a specimen of C. marginatum, which had been removed
from the cyst after fixation in alcohol and relaxed by slight maceration; from bass
from Nebish, Mich.

FIG. 6. Sketch of living animal in "suctorial pose" as seen from ventral side;
from bass, Nebish, Mich.

FIG. 7. From a sagittal section of anterior end of worm which died in "suc-
torial pose," from heron, Nebish, Mich. Camera lucida, X 38 diameters.

FIG. 8. View of ventral surface of a living animal in the "swimming pose,"
from bass, Nebish, Mich.

FIG. 9. View of a portion of the muscles of the ventral ccelomic wall of frog,
showing a cyst and enclosed worm. From specimen after clearing and mounting
in balsam. St. Paul, Minn. Camera lucida, X ^5 diameters.

BIOLOGICAL BULLETIN,

PLATE i

-•-' --^- 0.5 mm

HENRY LESLIE OSBORN

THE SCALES OF FRESHWATER FISHES.

T. D. A. COCKERELL,
UNIVERSITY OF COLORADO.

Professor N. S. Shaler,1 writing of his experiences as a student
under Agassiz, said: "I acquired a considerable knowledge of
the literature of ichthyology, becoming especially interested in
the sytem of classification, then most imperfect. I tried to follow
Agassiz's scheme of division into the order of ctenoids and ganoids,
with the result that I found one of my species of side-swimmers
had cycloid scales on one side, and ctenoid on the other. This
not only shocked my sense of the value of classification in a
way that permitted of no full recovery of my original respect
for the process, but for a time shook my confidence in my master's
knowledge."

I quote this, because the breakdown of Agassiz's original system
of classification by scales affected not only Professor Shaler,
but many others, to the detriment of lepidology. Dr. Jordan,
in his great work, "A Guide to the Study of Fishes" (1905),
devotes only two pages to the discussion of scales, with a couple
of figures. In the descriptions of fishes published by various
authors, much is made of the number and size of the scales, but
little of their structure, and it rarely happens that an ichthy-
ologist even takes the trouble to remove a scale from the fish
he is studying, in order to determine its characters. These re-
marks apply especially to the teleosts — the ordinary fishes of
modern times, and it is of course true that students of the "ga-
noids" and their allies, mostly fossil, have never neglected the
scales. The ganoids have greatly thickened scales, usually
rhombic in form, which are well preserved in the rocks; as Agassiz
so well insisted, they afford a most important aid to the classi-
fication of these animals, particularly in the numerous cases in
which the skeleton is poorly preserved. Recent researches, while
destructive to certain theories which Agassiz considered im-

1 Atlantic Monthly, February, 1909, p. 222.

367

368 T. D. A. COCKERELL.

portant, have only emphasized his doctrine that the scale is of
great taxonomic value. I have paid little attention to the ganoids
and their relatives, and will not attempt to discuss them, here,
but may call attention to the important memoir by Goodrich1
in which it is shown that the so-called ganoid scales can be divided
into three very distinct groups, called the cosmoid, palaeoniscoid
ganoid and lepidosteoid ganoid. Mr. Goodrich, having distinctly
differentiated these types, applies his work to the classification
of some difficult forms, with striking results.

Teleostean scales, the ordinary imbricated readily removable
scales of fishes, were classified by Agassiz as cycloid and ctenoid,
and his names are still in current use. The cycloid scale is one
in which the apical margin is what a botanist would call entire,
that is without teeth or serrations. The ctenoid or comb-like
scale has the margin dentate, serrate or spiny. It is now wrell-
known that these are not fundamental divisions, some families
having both ctenoid and cycloid types, and as Shaler has stated,
it is possible for both kinds to exist on a single flat-fish. Never-
theless, the distinction is usually an important one, while the
scales have innumerable other characters of value, not used in
Agassiz's sytem of classification.

My own work with fish-scales had what might be called an
accidental beginning. In the course of class-work, my students
and I took occasion to examine the scales of the fresh-water
fishes of Colorado. We soon perceived that they had excellent
distinctive characters, and thereupon sought literature on the
subject. Failing to find anything satisfactory, we wrote to Dr.
D. S. Jordan, who promptly gave us the desired information,
or rather, indicated the lack of it. Girard, in his report on the
fishes of the Mexican Boundary Survey, had given numerous
figures of the scales of fresh-water fishes, but had not discussed
or classified them. About fifty years ago it appears that it was
seriously intended to work up the lepidology of the American
teleosts, but for some reason nothing came of it, beyond the
publication of \he figures mentioned. Later, Dr. Boulenger gave
me access to the work of Fatio (1882), in which the scales of all
the fishes of Switzerland are figured; also to the great work of

lProc. Zool. Soc. London, 1907, pp. 751-774.

THE SCALES OF FRESHWATER FISHES. 369

Sauvage on the fishes of Madagascar, containing numerous figures
of scales of Acanthopterygian fishes. In other writings, here
and there, are various figures of scales, usually with little dis-
cussion. The scales of Gadus (codfish and allies) have been
beautifully figured and fully discussed by Dr. H. W. M. Tims

(I905)-1

Dr. Jordan, with great kindness, sent me a fine series of fresh-
water fishes from the collections of Stanford University, while
Dr. Evermann was equally good in supplying numerous species
from the Bureau of Fisheries. I thus obtained nearly all of the
principal forms of North American Cyprinidae or carp-like
fishes. The following summer I visited the British Museum, and
was indebted to Dr. Boulenger for the opportunity of investi-
gating the scales of nearly all the principal old-world cyprinids,
and a like series of African characinids. I also obtained through
him many African fishes of other families. Quite recently I
have received from Dr. Eigenmann, of the University of Indiana,
an extremely fine series of South American characinids. With
all these, and some others, it has been possible to test rather
thoroughly the value of scale-characters, and the result has been
to show that while they are not rarely deceptive, through con-
vergence, they are on the whole of great taxonomic importance.
As in most other taxonomic work, there are disturbing elements
due to individual variability and differences of age, but while these
are sure to lead to various minor errors, they will not much affect
the broader results. The key to the origin of the sculpture of a
teleostean scale is apparently to be found in that ancient type
the bowfin of North America, Amia calva. Fig. I shows part of
the base of a scale of this fish, which it wrill be observed consists
of longitudinal strands or fibres, separable elements which fray
out basally.2 In the apical field these are directed toward a
broad rough nuclear area. A close approximation to this is found
in a very old type of teleosteans, the lady-fish, Albula. In this,
however, appear also the beginnings of the radial lines, extending
from the nucleus of the scale to the margin. As we go higher

1 Dr. B. L. Chaudhuri informs me that Dr. John McClelland published an account
of the scales of Indian Cyprinidae in the appendix to his work on these fishes, in
1839. This I have not seen.

2 See Smithsonian Misc. Coll., Vol. 56, No. 3, p. 2, 1910.

37O T. D. A. COCKERELL.

in the scale of fish-evolution, these radiating lines or radii
often become very prominent, while the longitudinal strands
usually become united above and below, forming circular
fibers which we have designated circuli. The nomenclature of
these structures was based on a normal highly-developed scale,
in which the circuli deserved their name, and since then the
term has been applied to the same elements wherever found,
so that I have had to refer, rather illogically, to longitudinal
circuli. Perhaps it would be better to call them fibrillae.

On looking at a normal scale, in which the circuli are strictly
circular, it would be natural to regard them as lines of growth,
like those on a snail's shell. There are real lines of growth, how-
ever, and these do not necessarily coincide with or have anything
to do with the circuli.

In the salmon and trout (Salmo) the scales are strictly cycloid,
and have only circuli. Fig. 2 shows a scale of the bluefin,
Argyrosomus nigripinnis, a member or the salmon family. This
showrs a marked deviation from the simple Salmo type, in that
there are distinct laterobasal angles, the circuli of the apical
field are less dense than those of the basal, and there are slight
indications of apical radii. It is no very great step from this to
the scale of Moxostoma aureolum, one of the suckers (Fig. 3),
but it will be noted that in the sucker, as indeed in all members
of the Catostomidae so far examined, there are very distinct
basal radii. In the typical suckers, Catostomus, the scale is oval,
not unlike that of Salmo in shape, and there are radii all around.
This is well shown in Fig. 4, Pantosteus santa-ance, from Califor-
nia. The possession of basal radii separates the Catostomidae
from the great majority of American cyprinids, although a few
genera of the latter family have them, notably Chrosomus (Fig. 5,
C. dakotensis], the scale of which closely resembles that of
Pantosteus in sculpture, though differing in shape. Among the
old world cyprinids (carp family) basal radii are very common,
and it is curious that the common European minnow (Phoxinus
phoxinus) has scales of quite the same type as the American
Chrosomus. There is reason to believe that the suckers are an
ancient family, close to the old stem from which the carp family
arose. They are nearly confined to North America, though spar-

THE SCALES OF FRESHWATER FISHES. 371

ingly represented in eastern Asia. From this circumstance one
might look for the earliest types of Cyprinidae also in America,
but the strong indications are that they are old-world forms,
the modern American cyprinids, with a few possible exceptions,
having arrived in comparatively recent times from Asia. Never-
theless, the carp family in this country must be only of compara-
tively, not actually, recent origin, since it has had time to de-
velop innumerable species, and a considerable series of endemic
genera. The distinctness of the American Cyprinid fauna
has indeed been emphasized by the scale work, it being shown
that the so-called Leuciscus, Rutilus, etc., of America are not con-
generic with the European fishes bearing these names. It is
probable, in fact, that all of the American Cyprinidse are gen-
erically distinct from those of Asia and Europe.

Returning again to the bluefin scale, we can take a new point of
departure in Fig. 6, the Asiatic cyprinid Squaliobarbus curriculus.
Here the general form remains the same, except for a moderate
elongation, the apical circuli become more widely spaced, the
apical radii are evident, but there are no basal radii. A European
scale of nearly the same type is that of Leuciscus illyricus (Fig. 7).
In other cases, the scale may be more parallel-sided, with a broad
truncate base, as is shown in the Asiatic Labeo sladoni (Fig. 8)
and Rhinogobio typus (Fig. 9).

In order to illustrate the value of scale characters within a
limited group, I give a series of figures of species usually referred
to Leuciscus, the dace and its allies. Leuciscus rutilus (Fig. 10)
and L. friesii (Fig. n) are European; L. hakuensis (Fig. 12) and
L. jouyi (Fig. 13) are Japanese. The remaining four are North
American, and because of their very different scales, and other
reasons, I have removed them to another genus, Richardsonius
Girard. R. orcutti from California (Fig. 14) constitutes a distinct
subgenus, the scale having basal radii. R. pulchellus (Fig. 15)
is of the Rocky Mountains. R. carletoni (Fig. 16) is a distinct
type from Maine, while R. thermopliilus (Fig. 17) inhabits warm
springs in Oregon. It will be observed that in the European
forms the large scales have exceedingly fine circuli. In the
American, the scales are very much smaller, and the circuli
are very coarse and widely-spaced. This is in general a charac-

372 T. D. A. COCKERfcLL.

teristic distinction between New World and Old World Cyprin-
ids, although some of the Old-World genera have widely-spaced
circuli. The two Japanese species, as one might expect, are
rather intermediate between the European and American. They
probably should constitute a new subgenus or genus. Leuciscus
illyricus (Fig. 7) has widely spaced apical circuli, and therein
departs in the direction of the Japanese and American fishes.

The Characinidse are a large family of freshwater fishes oc-
curring in the Ethiopian and Neotropical regions, but not else-
where, with the exception of a few which enter the Palsearctic
in the Nile Valley and the Nearctic in southern North America.
This curious distribution has naturally aroused interest among
naturalists, and while it is probable that America was their original
home, it is a question whether they reached Africa across what
is now the Atlantic, or once inhabited the north and passed south-
ward into the Ethiopian area. I should think the latter sup-
position more probable, but for the fact that as yet we have no
trace of fossil characinids outside of their present area. The
characinids belong to the same general group as the cyprinids,
but are unquestionably more primitive, in spite of the fact that
many of their members are exceedingly specialized in certain
particulars. Among the African characinids is a group, the
Alestini, having a very characteristic type of scale.1 This,
which I call the alestiform scale, is more or less hemispherical
in outline, cycloid as to the margin, with a few very strong radii.
It seems to be an old type, because we find it nearly repeated in
certain of the relatively primitive families, as for instance the
Phractolsemidse (Phractolcemus ansorgii, Fig. ija), a rare type from
the Niger and Congo rivers. The Phractolcemus scale, however,
is incipiently ctenoid. A different scale, yet of the general alesti-
form type, is found in Pantodon buchholzi (Fig. 18), another rare
and isolated genus from the Niger and the Congo. On this fish
the circuli have a peculiar and very characteristic bead-like
appearance. Passing to the more specialized cyprinids, we find
the alestiform scale not uncommon, and a good example from an
Asiatic fish (Barbus mahecola, Fig. 19) is given. It will also be
noted that Leuciscus nitilus of Europe (Fig. 10) has much the

1 See Smiths. Misc. Coll., Vol. 56, No. i, plate i, figs. 4, 5, 6.

THE SCALES OF FRESHWATER FISHES. 373

same sort of scales though lacking the characteristic lateral radii
of the Alestini and the Barbus.

In some of the cyprinid alestiform scales the nuclear region
is broken up into polygonal areas, but in the curious African
Mormyridae1 there is a characteristic system of anastomosing
radii, forming a network. This is unique, so far as material seen
by me shows, except for the case of Heterotis niloticus (Fig. 20),
a member of the ancient and now much reduced family Osteo-
glossidae. In this fish, which inhabits tropical Africa north of the
Equator, the network is extremely well developed.2

An extremely distinct and interesting type of scale is found in
the tench of Europe, Tinea vulgaris (Fig. 21). This seems to be
a relatively old type, formerly more abundant. Through the
kindness of Dr. A. S. Woodward, of the British Museum, I
was able to examine the miocene fossil species Tinea furcata of
Agassiz, and the nominal but apparently synonymous T. magna
of Winkler, and found the scales to be wholly characteristic of
the genus. It is a singular thing that in North America there
exists a fish called Algansea tincella,— named tincella by Valen-
ciennes because of its resemblance to a tench, — which has scales
of the same general type though not so extreme. Mr. Regan
has described two other species of the genus Algansea, A. affinis
and A. stigmatura, and he informs me that they have scales like
those of tincella.

On the other hand, there are scales which so far as shape goes
resemble those of Tinea, and yet offer marked differences in the
details of the sculpture. Such is the scale of the Asiatic Schizo-
thorax biddulphii of Gunther (Fig. 22). Among the American
cyprinids, especially the smaller forms commonly known as
minnows, there are many scales which at first sight look alike.
In some cases, there is actually no definable difference, but often
sufficient familiarity with the scales enables one to recognize
them without difficulty. I illustrate a series of such forms,
Lavinia exilicauda (Fig. 23), Phenacobius mirabilis (Fig. 24),
Notropis galacturus (Fig. 25) and Pimephales anuli (Fig. 26) ;
also a similar European scale, Gobio fluviatilis (Fig. 27). In all

Smiths. Misc. Coll., Vol. 56, No. 3, p. 2, figs, i, 2.

2 Since this was written I have determined that all the Osteoglossidae, and also
all living Dipnoans (lung-fishes) have the radial network.

374 T- D- A- COCKERELL.

these scales there is much variation, which is at times confusing,
although it usually does not affect the fundamental character
of the pattern. Thus I described the scale of Chondrostoma soetta,
from Italy, as having few apical radii, distinguishing it thus from
certain Spanish species. Noting later that this did not agree
well with Fatio's figure, I asked Mr. Regan to look into the matter
and he found that on the same fish immediately adjacent scales
had many and few radii respectively. Consequently my diag-
nostic character proves of no value, but it will be noted that the
variation affects the development rather than the character of
the markings. Greater differences are noted when the scales
are taken for different parts of the body, and for purposes of
comparison I always take them from near the middle of the side,
close to the lateral line. Fig. 28 and 29 show an extreme case,
that of the curious little fish Ericymba buccata, from Indiana.
Fig. 28 is a normal scale, and Fig. 29 is from the subdorsal region
about 5 mm. in front of the dorsal fin. These two scales are
from the same fish.

Occasionally very marked abnormalities occur. Thus on a
specimen of Ostebrama fecB I found a greatly enlarged scale on
the middle of the side, below the lateral line. It was about 8.5
mm. long, six times as long as the exposed parts of the normal
scales, but not differing essentially in sculpture The age of
the fish is also often of importance. Fig. 30 shows a scale of
Gila robusta from Arizona. It was taken from a young fish,
and for some time I did not know that the adult scales of Gila
have a characteristic basal lobe, as indeed was figured long ago
by Girard in the report of the Mexican Boundary Survey.

There is an interesting case outstanding, in which the value
of scale characters for the separation of very closely allied fishes
has yet to be determined. From Stanford University I received
a number of specimens of Myloleucus symmetricus , collected in
different localities. These were sent principally because Dr. J. O.
Snyder had already noticed the variability of this fish in other
than scale characters, and it was suspected that it might include
some separable forms. A study of the scales revealed four ap-
parently distinct types, (i) Scale broad, small. Navarro
River, Gualala River (California) and north to Drew Creek in

THE SCALES OF FRESHWATER FISHES. 375

Oregon. (2) Scale broad, large. Napa River (California).
(3) Scale oblong, more weakly sculptured. Ukiah Creek;
Russian River (California). (4) Scale oblong, strongly sculp-
tured. Conchilla Creek (California).

It appears from this that there is a broad-scaled form to the
north coming south as far as Gualala River. From this it is only
a few miles to Russian River, wrhere the fish is oblong-scaled.
Passing beyond, however, we find another broad-scaled variety
in Napa River. It is suggested that the broad-scaled fishes,
when in competition with the oblong-scaled, may get the better
of them, and that the Napa River fish represents an invasion from
the north away from the coast. All this, however, is at present
based on far too little evidence, and is given here partly as an
example of the kind of problem the scale-work brings up, and
partly in the hope that someone will make an exhaustive study
of the matter. Up to the present, I have not been able to get
any further information, beyond what is given by Dr. Snyder
in Bull. Bureau of Fisheries, XXVII, p. 175. In general, Dr.
Snyder's results, based on other than scale characters, agree with
mine, but he finds no difference between the fishes of Russian
and Napa Rivers. No less than five specific names have been
given to M. symmetricus as now understood; probably two or
three of these will be available for the necessary segregates.
While most fish scales fall readily into a systematic arrangement,
we occasionally meet with one which differs greatly from any
other known to us. Such is the scale of Kneria cameronensis
Boulenger (Fig. 31), a small fish belonging to a peculiar family
found only in tropical Africa. In this the radii run from one end
of the scale to the other and the circuli are for the most part
broken up into peculiar marks between them.

The study of fish-scales opens up a vast new field, the scales
of most of the common fishes being still undescribed. Fish scales
also may claim our attention for their beauty as microscopical
objects. Their preparation involves no difficulty; they are
best mounted dry between thin microscopical slides; or when
small, they may be mounted on a slide under a cover glass.
At first, I placed them in balsam, but this greatly obscures the
markings, and is altogether undesirable.

376 T. D. A. COCKERELL.

For the means of obtaining the photographic figures (the work
of Mr. T. C. Black) I am indebted to a grant from the American
Association for the Advancement of Science. The drawings
are by Miss Evelyn V. Moore. In this paper I have described
and figured only scales of fresh-water species. At some future
time, it may be possible to give a similar account of the marine
families, and of several fresh-water groups now omitted.

378 T. D. A. COCKERELL.

EXPLANATION OF PLATE I.

FIG. i. Amia calva. Plymouth, Indiana. Dr. Evermann.

FIG. 2. Argyrosomus nigripinnis. Dr. Graenicher.

FIG. 3. Moxostoma aureolum. Dr. Graenicher.

FIG. 4. Pantosleus santa-ana. Santa Ana River, California. Stanford

University.

FIG. 5. Chrosomus dakotensis. Valentine, Nebraska. Bureau of Fisheries.

FIG. 6. Squaliobarbus curriculus. Mountain stream near Kiu-Kiang (Styan).
British Museum.

FIG. 7. Leuciscusillyricus. R. Tadro, Dalmatia (Dr. Werner). Brit. Museum.

FIG. 8. Labeo sladoni. Mandalay (F. Day). Brit. Museum.

BIOLOGICAL BULLETIN, VOL, XX

1

1 1

T. 0. A. COCKERELL

380 T. D. A. COCKERELL.

EXPLANATION OF PLATE II.

Fie. g. Rhinogobio typus. Kiu-Kiang (Styan). Brit. Museum.

FIG. 10. Leuciscus rutilus. Salisbury, England (Odgen Smith). Brit.
Museum.

FIG. 11. Leuciscus friesii=L. meidingeri. Lake of Derkos, Constantinople.
Brit. Museum.

FIG. 12. Leuciscus hakuensis. Yamada, Japan (R. Gordon Smith). Brit.
Museum.

FIG. 13. Leuciscus jouyi. Sasuma, Japan (Anderson). Brit. Museum.

FIG. 14. Richardsonius (Temeculina) orcutti. Santa Ana River, California.
Stanford University.

BIOLOGICAL BULLETIN, VOL. XX

PLATE II

fc$-*r "tl

v» / H

-,

T. D. A. COCKERELL

382 T. D. A. COCKERELL.

EXPLANATION OF PLATE III.

FIG. !=;• Richardsonins (Tiogma) pulchellus. Alamosa, Colorado. Stanford
University.

FIG. 16. Richardsonius (Cheonda) carletoni. Cross Lake Thoroughfare, Maine.
Bureau of Fisheries.

FIG. 17. Richardsonius thermophilus. Warm Spring, Harney Co., Oregon.
Stanford University.

FIG. 170. Phractolaemus ansorgii. Brit. Museum.

FIG. 1 8. Pantodon buchholzi. Brit. Museum.

FIG. 19. Barbus mahecola. S. Canara (F. Day). Brit. Museum.

FIG. 20. Heterolis nilolicus. Brit. Museum.

BIOLOGICAL BUI LETIN, VOI . XX

20

T D. A COCKERELL

384 T. D. A. COCKERELL.

EXPLANATION OF PLATE IV.

FIG. 21. Tinea vulgaris. Constantinople. Brit. Museum.
FIG. 22. Schizothorax biddulphii. Lake Balyk-ky near Nijo. (St. Peters-
burg Mus.) Brit. Museum.

FIG. 23. Lavinia exilicauda. Coyote Creek, California. Stanford University.

FIG. 24. Phenacobius mirabilis. Bureau of Fisheries.

FIG. 25. Notropis galacturus. Saltville, Virginia. Bureau of Fisheries.

BIOLOGICAL EULLETIN, VOL. XX

PLATE IV

T. D. A. COCKERELL

386 T. D. A COCKERELL.

EXPLANATION OF PLATE V.

FIG. 26. Pimephales anuli. Lunkasoos Lake, Maine. Bureau of Fisheries.
(Note straight edge of skin across middle of scale.)

FIG. 27. Gobio flm'iatilis (G. vulgaris). R. Neckar near Canstatt (Stuttgart
coll.) Brit. Museum.

FIG. 28. Ericymba buccata. Wild Cat Creek, Indiana. Bureau of Fisheries
Figure reversed.

FIG. 29. Ericymba buccata. Wild Cat Creek, Indiana. From subdorsal region,
in front of dorsal fin.

FIG. 30. Gila robusta. Tempe, Arizona. Stanford University.

FIG. 31. Kneria cameronensis. Brit. Museum.

BIOLOGICAL BULLETIN VOL XX

26

PLATE V

•

, 34