BULLETIN

OF THE

MUSEUM OF COMPARATIVE ZOÖLOGY

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. XXX.

CAMBRIDGE, MASS., U. S. A.
1896 -- 1897.

Reprinted with the permission of the original publisher
KRAUS REPRINT CORPORATION
New York
1967

Printed in U.S.A.

CONTENTS.

No. 1.— The Early Development of Asplanchna Herrickii de Guerne; a con-
tribution to Developmental Mechanics. By H. S. Jnnnines. (10 Plates.)
October, 1896 .

No. 2. — Some Variations in the Genus Eucope. By A. Agassiz and W. M.
WoopwonTu. (9 Plates.) November, 1896

No. 3.— Reports on the Results of Dredging in the “Blake” XXXVII.
Supplementary Notes on the Crustacea. By W. Faxon. (2 Plates.) No-
vember, 1896

No. 4. — On the Colors and Color-patterns of Moths and Butterflies. By A.
G MAYR. (10 Plates). February, 1897

No. 5. — The Mesenteries and Siphonoglyphs in Metridium marginatum
Milne Edwards. By G. H. Parker. (1 Plate.) March, 1897 .

No. 6. — Photomechanieal changes in the Retinal Pigment Cells of Palae-
monetes and their relation to the central nervous system. By G. H.

PARKER. (1 Plate.) April, 1807

PAGE

119

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.
Von. XXX. No. 1.

THE EARLY DEVELOPMENT OF ASPLANCHNA
HERRICKII DE GUERNE.

A. CONTRIBUTION TO DEVELOPMENTAL MECHANICS.

By HERBERT S. JENNINGS.

Wırn Ten PLATES.

CAMBRIDGE, MASS., U. S. A:
PRINTED FOR THE MUSEUM.

Ocronxkn, 1896.

No. 1.— The Early Development of Asplanchna Herrickii de
Guerne. A Contribution to Developmental Mechanics] By
HERBERT S. JENNINGS.

CONTENTS.
PAGR PAGE
Introduction . 2| 2. Maturation . » " 19
Part First. Developmental Mechantos . 4| 3. Orientation of the Developing I 1
I. Statement of Problemm 4 biyot Wo 1608 og 14
1. Clea vage boss uw ae eden, eerie en EH 16
A. Direction of Cleavage «is ric Nomenclature . . . + + + 16
(1) Berthold’s principle of least s sur- First Cleavage . « « 18
faces . . 4 Second ‘ Diu DESTIN M
(2) Hertwig's lew of the apial in Third: ut 1 ood e o 2M
the longest axis of the proto- Fourth“ „
ne mass RR, Bin S 33
(3) Braem's theory of separation in Sixth S . 95
the direction of least resistance 5 Seventh and Later Cleavages 45
(4) Roux's theory of a compromise The Ectoderm. x « + + + 8
between the tendency imma- The Entoderm . » . 54
nent in the nucleus and the III. Discussion of the Bearing of the
tendency due to the form of Observations on the Problems . 58
the protoplasmic mass . + 6| 1. Cleavage . » - H onov 88
(5) Heidenhain's problem of a defi- A. Direction of Cleay age 58
nite angle of rotation 6 (1) Berthold's theory of least siii 58
(6) Sachs’s law of the right- Jangted (2) Hertwig's law of the nn in
arrangement of cell walls. . 7 the longaxis . i~i 60
(7) Rauber's theory of inter-attrac- (3) Braem's theory of least alben 69
tion of asters . ERA. ° (4) Roux’s theory . . . 70
(8) Braem’s principle of equi re- (5) Heidenhain's problem ot a defi-
sistance at the two ends of the nite angle of rotation . . . 72
spindle " 7 (6) Sachs's law of the right-angled
B. Equality or Tosanality of Cleav age 7 arrangement of cell walls .. 73
C. Rate of Cleavage . 1 (T) Rauber's theory of the inter-
D. Differentiation during Cleav age. 9 attraction of asters . . . . 74
2. Later Developmental Processes . 9 (8) Braem's principle of equal re-
II. Descriptive Portion . . 10 sistance at the two ends of the
1. Form and Structure of the E gg . 10 Some « 1 € xoa co Ae

1 Contributions from the Zoölogical Laboratory of the Museum of Comparative
Zoölogy at Harvard College, under the direction of E. L. Mark, No. LXX.
Contribution No. LXIX., under the direction of E. L. Mark, was published
under the following title: “Descriptions of Three Species of Sand Fleas (Amphi-
pods) collected at Newport, Rhode Island. By Sylvester D. Judd." Proceed.
U. S. Nat. Museum, Vol. XVIII. No. 1084, pp. 593-603, 11 Figures, August [1], 1896.
VOL, XXX. — NO. 1. 1

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

PAGE PAGE

B. Equality or Inequality of Cleav- 2, Development 89

age Products . . . „ „ 74 = Maturation « e + « 89

C. Rate of Cleavage. . 4 14 3. Cleavage e « v v. 04
D. Differentiations ^ accompanying " Summary on Maturation and

Cleavage 6 o e o o T Cleavage in the Rotifera . . 100

2. Gastrulation . © . . . . . . 80 Part Third. Material and Methods. . 101

3. General Considerations . . . 88 General Summary , 106

Part Second. Discussion of Matters A, Observationgs 106

bearing upon the Morphology of B. Conclusions «e s 1 108

the Rotifera . . + . 8T | Literature Cited ^ v. e. 111

1. Previous Knowledge of Asplanchaa Explanation of Plates . .117

Herſekkll 8

INTRODUCTION.

Tue following pages contain a study of the early development of an
organism, with especial reference to recent theories in regard to the laws
of cleavage and the relation of cleavage to morphogenesis.

Many theories and so called laws have been set forth concerning tho
factors determining the manner and rate of cleavage. These have
taken the form chiefly of theories in regard to the causes of the direc-
tion of the spindle, of the equality or inequality in size of the products
of division, and of the relative rapidity with which the different cleav-
age cells divide. Yet few attempts have been made to interpret con-
sistently the cleavage of any given organism with relation to any or all
of these theories. The sketch of Braem (794) with regard to the
Echinoderm egg, and the recent studies of Ziegler ('95) and zur Stras-
sen (^96)! on the Nematode egg, are almost the only works that can be
cited in which an attempt has been made to show the relation of any
theory or theories to the series of normal cleavages in any animal. In
other discussions the theories have been based upon experimental evi-
dence or upon scattered observations. Yet it is, of course, the normal
processes for which explanations are desired ; scattered observations may
be adduced for almost any view. It seems of the greatest importance,
therefore, to show clearly the exact relation which the theories hitherto
proposed have to the actual series of cell divisions in the development
of particular organisms.

1 In view of the close similarity of some of my conclusions with some of those
in the more recent (96) of two papers by zur Strassen, it may be proper to state
that a copy of the present paper, exactly as here published, with the exception of
some verbal alterations and the addition of a few references, was deposited with
the Faculty of Arts and Sciences of Harvard University on April 80, 1896, while
zur Strassen's (90) paper was not received here till May 13.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 3

Furthermore, there is much discussion of the question as to whether
cleavage is a mere quantitative separation of a single mass into smaller
masses similar in nature to each other and to the original egg, or
whether it is accompanied by a differentiation of the separated blasto-
meres, — as a result either of qualitative division or other changes.

A third question of theoretical interest, somewhat related to the last,
is whether the method of cleavage has a direct mechanical relation to
future morphogenetie processes, or whether it is merely the passing of
partitions through a mass of protoplasm, the order in which this occurs
and the arrangement of the partitions being immaterial. For examplo,
Ts gastrulation a process independent of cleavage and merely requiring
the latter as a prerequisite, — as the planting of seeds must be preceded
by ploughing, — or is gastrulation in some way connected with or depend-
ent upon the manner of cleavage? Stated in the most general terms,
this is the question: Is cell division a direct morphogenetic factor, or
are the real formative processes dependent upon the introduction of
other factors after the cleavage is finished ?

With these questions in mind, I have studied the development of an
organism of the class Rotifera throughout those stages of development
in which it is possible to make the cells the units of observation, — that
is, through cleavage and gastrulation and somewhat later.

Broadly stated, the object of the work may be expressed as the analysis
of the early development of an organism into the simplest factors possible.

The development of Asplanchna Herrickii has not been studied previ-
ously, and in the course of this paper it will be necessary to discuss some
matters which are of importance primarily to persons who are engaged
particularly with the morphology of the Rotifera, and which are not of
especial interest from a morphogenetic standpoint. In order to dis-
tinguish these two lines of discussion, I shall divide the work into two
main portions. Part First will contain all matters bearing upon devel-
opmental mechanics. Here will be found the minute description of the
cleavage, gastrulation, and other processes, as well as a discussion of
their bearing upon the problems of morphogenesis. Part Second will
contain a brief review of previous knowledge of the organism studied,
a comparison of the development, 80 far as traced, with the develop-
ment of other Rotifera, and a discussion of some of the conditions de-
scribed by other authors. These principal parts will be followed by a
third, on material, methods, and other subordinate matters, and the
whole will be closed by a summary of the more important conclusions
arrived at.

BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

PART FIRST.—DEVELOPMENTAL MECHANICS.

I. Statement of Problems.

We shall deal in the following pages with (1) cleavage, (2) gastrula-
tion, and (3) the relation of these to each other.

1. CLEAVAGE.

It will be necessary, in studying the cleavage and the factors deter-
mining it, to enter into minute details as to the movements of asters,
the form and dimensions of cells, and other similar matters; the effort
of following this, in itself somewhat laborious, description will be much
lightened by holding in mind the problems upon which it bears. I
shall therefore give first a statement of the main theories which have
been advanced as to the determining factors in cell division.

Cell division presents three aspects, in each of which its nature is in
some way determined. (A) As to the direction of cleavage : the posi-
tion in which the new septum is to appear. Since this bears a definite
relation, in general, to the position of the spindle leading to the cleav-
age, we may speak of this aspect as the determination of the direction
of the spindle. (B) As to the relative size of the two products;
whether the division is equal or unequal. (C) As to the relative time
of division, or the interval between successive cleavages.

Besides these, we have (D) the question of the qualitative nature of
cleavage. Are all the cells that are produced of similar structural and
material character, or is cleavage accompanied by qualitative differentia-
tion of the blastomeres, — either as a result of qualitative karyokinesis
or otherwise

A. Theories as to the Factors determining the Direction of the Spindle
and the Position of the new Cell Wall.

The theories as to the factors determining the direction of cleavage
are numerous, and have been much discussed of late. General reviews
of these theories will be found in Driesch (792, p. 26), Braem (794,
p. 340), Ziegler (94, p. 136), and McMurrich ('95). I shall give here
as brief and precise a statement of each theory as possible, first in my
own words, then, so far as practicable, in a quotation from the author.

(1) Berthold’s principle of least surfaces. — Berthold ('86) holds that
the form and relative position of cells, and as a consequence their direc-

|
|

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRIOKII. 5

tion of cleavage, is determined, partially at least, by the same factors
which determine the form and relative position of soap bubbles in a mass.
As a result of surface tension, the cells take such forms as to occupy
the given space with the least possible surface areas. New septa will
appear in such positions that their surfaces will be the least possible
areas that could divide the cell into parts of the required size. “Die
Lamellensysteme ordnen sich so an, die einzelnen Lamellen krümmen
sich in der Weise, dass die Summe der Oberflächen aller unter den ge-
geben Verhältnissen ein Minimum wird.” (Berthold, '86, pp. 219, 220.)

(2) Hertwig’s law of the spindle in the longest axis of the protoplasmic
mass. — According to Hertwig’s well known view, as a result of the inter-
action of nucleus and protoplasm, the spindle during division comes to
lie in such a position that its longitudinal axis coincides with the axis
which passes through the greatest protoplasmic mass. “ Es lässt sich
hier das zweite allpemeine Gesetz aufstellen, dass die beiden Pole der
Theilungsfigur in die Richtung der grössten Protoplasmamassen zu
liegen kommen, etwa in derselben Weise, wie die Lage der Pole eines
Magneten durch Eisentheile in seiner Umgebung beeinflusst wird.”
(Hertwig, ’93, p. 175.)

(3) Braem’s theory of separation in the direction of the greatest space
‚for development. — This is a modification of the principle of least pres-
sure, first enunciated by Pflüger (84). Since Pflüger's principle, con-
sidered from a purely mechanical standpoint, seems irreconcilable with
the nature of the material on which it was supposed to act, and since
Braem’s view is based on an essentially different conception of the na-
ture of the phenomena, I have not thought it necessary to take into
direet consideration Pflüger's view.

Braem holds that when an egg is subjected to unequal pressure, the
spindle places itself in such a position that the resulting products shall
have the freest opportunity for development; that is, in the direction of
least resistance. The rule is not the expression of a purely mechanical
force, but is to a certain extent teleological in character. “Die Spindel
eines ungleichem Druck unterliegenden Eies stellt sich in derjenigen
Richtung ein, in welcher der räumlichen Entfaltung der Zelle und ihrer
Teilprodukte der freieste Spielraum geboten ist. Ich glaube, dass diese
Fassung trotz oder vielmehr gerade wegen ihres teleologischen Gehaltes
dem Wesen der Sache besser entspricht als die rein mechanische Deu-
tung.” (Braem, ’94, pp. 341, 342.)

The result is held to be due to a sort of sensory power resident in
the egg, “eine Art Tastsinn, durch den es der Zelle möglich wird, sich

6 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

über ihre unmittelbare Umgebung zu orientieren und demgemiüss ein-
zurichten." (p. 342.)

(4) Loux’s theory of a compromise between the tendency immanent in
the nucleus and the tendency due to the Jorm of the protoplasmic mass.
— Roux holds that the spindle places itself in one of the positions of
stable equilibrium in relation to the protoplasmic mass, — therefore, at
least generally, in the longest axis of the protoplasmic mass, though
sometimes at right angles to that axis, the factor that decides which of
these positions shall be taken being an immanent tendency in the nu-
cleus to divide in a certain direction.

“Richtiger ist es zu sagen: Die Kernspindel der Furchungszellen
stellt sich in die, resp. in eine Richtung festesten Gleichgewichtes der
tractiven Einzelwirkungen der Protoplasmamasse. Diese Richtung
entspricht überwiegend häufig annähernd oder ganz der grössten durch
den Mittelpunkt der Protoplasmamasse gehenden Dimension.

“Diese Richtung des Gleichgewichtes wird aber nicht vollkommen
vom Protoplasma allein bestimmt, sondern sie kann, wie ich bereits 1884
und 1885 auf Grund von Experimenten erschlossen habe, von der Lage der
immanenten Teilungsrichtung des Kernes zu den Hauptrichtungen des
Protoplasmakörpers abhängig sein ; denn ich erhielt bei symmetrisch ge-
stalteten ‘linsenformig’ deformirten, mit den grössten Fläche senkrecht
stehenden Froscheiern zwei Prädietionsrichtungen der Spindeleinstel-
lung: die Richtung der grössten und der kleinsten durch den Massenmit-
telpunkt gehenden Dimension, erstere allerdings wieder die überwiegend
häufige.” (Roux, ’94, p. 152.)

It is to be noted that this theory does not attempt to give any rule
by which the position of the spindle is necessarily determined ; the ten-
dency of the nucleus is simply “immanent,” and its factors unknown.

In addition to these four well characterizod theories, a number of
less definite or partial views have been set forth, — some proposing fac-
tors which may influence, though not alone determine, the position of
the spindle. A number of the more important of these will be
mentioned.

(5) Heidenhain’s problem of a definite angle of rotation ( Prob-
lem der gesetzmässigen Drehungswinkel ). — Heidenhain (94, p. 719)
thinks it probable, or at least possible, that careful investigation will
show that in a given tissue the position which the spindle takes at the
time of division is a result of its rotation through a definite angle, de-
terminable for the given tissue, after the first formation of the spindle
by the separation of the asters. This Separation of the asters is held

|
|
|

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 7

to be at first in a line at right angles to the axis of the preceding spin-
dle; then, by a rotation through an angle characteristic for the tissue,
the definitive position is reached. The position which the spindle is
finally to take is therefore determined at the time the asters separate.
“Soweit ich indessen die Lage übersehen kann, ist die schliessliche
Stellung der Spindel von dem Moment an fest gegeben, in welchem
die Theilung des Muttermikrocentrums stattfand." (Heidenhain, ’95,
pp. 555, 556.)

(6) Sachs’s view, that the walls separating the cells meet one another at
right angles. — This (Sachs, ’78, p. 1070) can hardly be considered as
more than a statement of a condition commonly found, Berthold (’86,
p. 252) and Hertwig (793, p. 177) have endeavored to show that the
condition is explainable as a result of the theories proposed by them.

(7) Rauber (83, p. 276) holds that there is evidence that the asters of
the different blastomeres exercise an attraction for each other in such a way
that, in a given area-composed of a number of cells, the spindles must
take such positions as to bring about a condition of equilibrium among the
asters. — “ Beurtheilt man die Verschiedenheiten der Furchennetzes von
der Stellungen der karyokinetischen Achsen aus, so gewinnt es den
Anschein, als ob die neu entstehenden Centren eines Blastomers auf
diejenigen der angrenzenden Blastomeren einzuwirken vermógen und die
Richtung ihrer Achsen beeinflussen." (Rauber, ’83, p. 280.)

(8) Braem’s principle of equal resistance at the two ends of the spindle.
— Subordinate to his principle of least resistance, Braem holds that the
spindle tends to take such a position that the pressure at the two ends is
the same. “Es ist das Prineip des gleichen Widerstandes, wodurch die
horizontale Lage der Spindel bedingt wird. Wir müssen annehmen, dass
der Kern von vornherein das Bestreben hat, sich gleichmässig nach bei-
den Seiten hin auszudehnen und somit auf eine äquale Zellteilung
hinzuwirken.“ (Braem, ’94, p. 345.)

In the following deseription these theories will be kept in mind, and
the bearing of the observations upon them pointed out. It will appear
that, for certain of the theories, the conditions in the egg of Asplanchna
present crucial tests.

B. Equality or Inequality of Cleavage.

The second aspect under which cleavage is determined is with regard
to the relative size of the two products. What is it that determines
whether the division shall be equal or unequal ?

Concerning the factors which determine the equality or inequality of

8 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

cleavage, two theories have been proposed. According to the view
which is perhaps that most generally known, the cause of unequal
cleavage lies in the relative distribution of yolk material and formative
protoplasm. The interaction between nucleus and cell contents, which
determines the position of the dividing nucleus, exists only between the
nucleus and the formative protoplasm, not between the nucleus and
the yolk material. As a consequence of this interaction, the nucleus
tends to take a position in the centre of the mass of Jormative proto-
plasm. When one region of the cell is composed largely of yolk
material, in a mere meshwork of protoplasm, while another region is
made up entirely of protoplasm, the dividing nucleus must separate
equal masses of formative protoplasm, and thus may divide the entire
mass into very unequal parts, —one containing a certain mass of proto-
plasm only, the other an equal mass of protoplasm and a large additional
mass of yolk material. The theory has recently been formulated by
Hertwig as follows : “Die Folge dieser Wechselwirkung aber ist, dass der
Kern stets die Mitte seiner Wirkungssphüre einzunehmen sucht.. . .
Wechselwirkungen finden zwischen dem Kern und dem Protoplasma,
nicht aber zwischen ihm und dem Dottermaterial statt, wolches boi allen
Theilungsprocessen sich wie eine passive Masse verhält. Ungleichmässig-
keiten in den Protoplasmavertheiluug müssen sich daher auch auf Grund
des obigen Satzes in der Lage des Kerns geltend machen, und zwar muss
derselbe nach den Orten der grósseren Protoplasmaansammlung hin-
rücken." (Hertwig, ’93, pp. 172 and 174.)

Braem’s principle of equal resistance at both ends of the spindle is in
character related to this view of Hertwig. Besides the effect of it in
determining the direction of the spindle, this supposed principle is like-
wise of effect in determining the equality or inequality of cleavage, as
appears from the quotation from Braem given on page 7.

C. Determination of the Time of Division, or the Interval between
Successive Cleavages.

The same factor which is held to determine the relative size of the
cells was also held by Balfour, with whom Hertwig agrees (Hertwig, '93,
p. 180), to determine the relative rapidity of cleavage. The greatest
interval between successive cleavages is found in cells which contain the
greatest amount of yolk relative to the amount of contained protoplasm.
“The rapidity with which any part of an ovum segments varies ceteris
paribus with the relative amount of protoplasm it contains; and the size
of the segments formed varies inversely to the relative amount of the
protoplasm.” (Balfour, '80, p. 99.)

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIL 9

D. Differentiation during Oleavage.

Besides these questions in regard to the form and rate of cleavage,
we have also the question of the qualitative nature of cleavage. ls
cleavage merely a quantitative process, or is it accompanied by a
differentiation of the separated cells? And if the latter is the case, by
what means is this differentiation accomplished !

The view once maintained, that cleavage is entirely unaccompanied by
differentiation of the separated cells, may be said to be nearly or entirely
given up; the questions which remain relate to the means by which this
differentiation is brought about. In regard to this several well defined
views exist.

1. Roux holds that the differentiation aecompanying cleavage is a
result of qualitative karyokinesis; i.e. at a given cell division the two
products receive nuclear material of different nature.

2. Driesch maintains that the differentiation which may accompany
cleavage is due to the specific cytoplasmic structure of the egg, different
parts of the egg being of different constitution, so that when this
differentiated mass is separated into parts, these parts receive different
sorts of cytoplasm. That is, the qualitative division is in the cytoplasm,
not in the nuclear material. “Ich habe schon oben gesagt, dass ich ein
Verschiedenwerden der Furchungszellen während der Furchung gern
zugebe, aber hierin nichts anderes als die Folge eines spezifischen
Plasmabaus des Eies sehe." — (Driesch, '94, p. 100.)

3. According to Wilson and Hertwig the differentiation accompanying
cleavage is duo, largely at least, to the interaction of the blastomeres,
after division has taken place. This does not exclude the possibility of
the existence at the same time of a qualitative division of the cytoplasm,
as stated above (2).

9. LATER DEVELOPMENTAL PROCESSES,

With regard to the later developmental processes, it will not be
necessary to give here a review of the various factors and theories
which have been set forth by different authors. Driesch (94) gives
an extended analysis of the morphogenetic process and its factors,
and Davenport (95) presents a detailed list of the different processes
concerned in development. It is sufficient here to propose a single
question : What is the relation of the cleavage process to the secondary
morphogenetic processes? Driesch's well known experiments indicate
that, in the caso of the sea-urchin, the manner of cleavage is entirely

10 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.
unimportant for the later morphogenetic processes. Gastrulation, for
example, occurs in the same manner, after the most varied and
fundamental alterations of the cleavage. Is this a fact which is
capable of generalization, — of application to different animals and
different methods of gastrulation? Doubtless the only positive answer
to this question must come from experimental studies; but a care-
ful descriptive analysis of the process in Asplanchna gives results
which, if the egg were a mechanism of the ordinary physical sort, would
be definite and conclusive.

II. Descriptive Portion.

1. Form AND STRUCTURE OF THE ÉGG.

The development of the embryo in Asplanchna Herrickii takes place
within the body of the mother, the egg lying enclosed in the enlarged
oviduct, close to the ovary. The chief axis of the developing embryo
bears no relation to the position of surrounding organs of the mother,
the egg lying in the oviduct as it might within a protecting sac of
any foreign material, its position determined by chance circumstances.
In cases where two embryos are present, their axes may make any angle
with each other.

For study it is necessary to dissect out the eggs. <A full account of
the methods of work is given in Part Third; here it is important to note
two facts: (1) All the work was done on preserved material; (2) Each
egg comes from a different individual, and is therefore in at least a
slightly different stage from every other. A considerable number of
eggs showing any given process, as, for example, the first cleavage, gives
therefore a series of stages, so that a complete idea of the changes taking
place may be gained.

The unsegmented egg is approximately an ellipsoid of nearly equal
axes, one end often slightly more pointed than the other. The form and
proportions vary a little, as do also the absolute dimensions. In many
eggs it is difficult to distinguish a more pointed end. The proportion of
the longer to the shorter axis is about as 9 to 8, and the average dimen-
sions of the egg are about 90 u through the longer axis by 80 u through
the shorter. Variations from a minimum of 84 by 70 to a maximum
of 97 u by 83 u were observed.

Whether an egg membrane is present or not is exceedingly difficult to
decide ; and I have not succeeded in thoroughly satisfying myself upon

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 11

that point. The walls of the oviduct invest the egg closely, and gener-
ally cannot be removed, so that a thin membrane, if present, could not
be detected. In most cases where it was possible to remove the walls
of the oviduct entirely, no membrane could be seen either in sections or
whole preparations, the egg appearing to be naked. In a fow cases,
however, in which cleavage had recently taken place and the cleavage
furrows were marked, it could be observed that the smooth outline of
the egg was preserved even above the cleavage furrow, either by means
of a membrane continuing across the furrow, or, what seemed from the
appearance of the preparations more probable, owing to the presence in
the furrows of a fluid mass, perhaps exuded from the egg. Such a case
is shown in Plate 1, Fig. 4. On other grounds, however, it seems pos-
sible that an extremely delicate membrane is present. Lameere (790)
states that in Asplanchna Sieboldii, which is likewise viviparous, it was
possible to observe definitely a very delicate membrane surrounding the
egg, especially at the time of the formation of the polar cell.

This question of the presence or absence of an egg membrane is of
importance from a mechanical standpoint, owing to its bearing upon the
question as to what preserves the ellipsoidal form of the egg. The form
is retained throughout all the early developmental processes; cleaving
cells do not project above tho general surface of the egg, nor do the
products of cleavage become spherical, touching at a few points only, as
is common in the Mollusea and other groups. This retention of the
ellipsoidal shape by the egg compels the cleaving cells to take various
peculiar forms, which allow of a direct test of some of the theories of
cell division above stated. It is also a most important factor in the
process of gastrulation, so that it becomes of great interest to discover
what it is that preserves this form.

It is ovident that surface tension would tend to produce a spherical
rather than an ellipsoidal form. Roux (95) has recently proved that
blastomeres have a direct attraction for each other; but an equal
attraction throughout the mass would produce a spherical form, and an
unequal attraction, such as would produce a regularly ellipsoidal form,
is very difficult to conceive of, especially as this attraction would have
to vary regularly with the shifting of the contents of tho egg. A mem-
brane of equal elasticity in all parts would likewise result in the produc-
tion of a spherical form. The only direct mechanical factor that seems
capable of explaining the continued ellipsoidal form is the presence of a
non-elastic membrane of the exact size and shape of the egg. But
during the later development the embryo enlarges and changes its form ;

12 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the membrane cannot therefore be absolutely inelastic, but might be of
such strength as to act as an inelastic membrane with regard to such
slight forces as are exerted within the egg during its early development.
The existence of a membrane of this peculiar character is, however,
very improbable, and it loses all its explaining power when the egg of
another rotifer, Melicerta ringens, is taken into consideration. In this
species the egg is not a regular ellipsoid or oval, but is of an irregular
shape, one side being curved in profile, the other straight. (See the
figures of Zelinka, 91.) This form is retained during development
exactly as in Asplanchna, yet is not explainable on the assumption of
a membrane. This question is discussed more fully later.

The cytoplasm of the egg is closely filled with fine yolk granules,
These are distributed uniformly throughout the egg (except that they
are not present in the asters), so that there is no visible differentiation
into regions containing greater and less amounts of yolk material.

The development of Asplanchna priodonta Gosse was also examined
for comparison with that of Asplanchna Herrickii. The egg of this
species is similar throughout to that of Asplanchna Herrickii, save that
it is smaller. The average dimensions are about 70% by 60 jp. The
egg of Asplanchna priodonta is shown in Figure 29, Plate 4, drawn to
the same scale as the figures from Asplanchna Herrickii.

2. MATURATION.

The formation of the polar cell in Asplanchna Sieboldii has been
described by Lameere (90) from observations upon the living egg.
The general features of the process are similar in Asplanchna Herrickii,
though the finer nuclear phenomena differ from those described by
Lameere. An account of the finer nuclear phenomena is, however,
foreign to the purpose of this paper: it is necessary to describe merely
the general features of the process, especially concerning the place where
the process occurs, in relation to the later orientation of the embryo.

As is now well known, but a single polar cell is commonly formed in
the parthenogenetically developing eggs of the Rotifera. The subject
has received full discussion, especially by Weismann und Ischikawa (87)
and Lameere (90). It may be noted that Zelinka (91) observed that in
a number of eggs of Callidina two polar cells were formed ; whether these
arose by division of a single one, or whether the two were formed. separ-
ately from the egg, is not stated. In no case have I obtained any evidence

09

indicating the formation of more thau a single polar cell in Asplanchna.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRIOKII. 18

In immature eggs the large germinative vesicle is commonly found in
an eccentric position, with no very apparent relation to the axes of the
egg. In Asplanchna Sieboldii, according to Lameere (90), just before
the movement of the germinative vesicle toward the spot where the
polar cell is formed, it lies in the long axis of the egg nearer one
pole, — in the position where later the first cleavage spindle is located.
This is probably the case also in Asplanchna Herrickii; but before
the germinative vesicle has begun to show the changes indicating
the formation of the maturation spindle it is difficult in preserved
material to get evidence as to the proper sequence of the stages
observed.

Just before the maturation spindle is formed, the nucleus moves
toward the periphery of the egg, and begins to lose its spherical shape.
It takes a position close to the surface, not at the equator, but nearer
one of the poles of the egg, as shown in Figure 1. In cases where the
differentiation into a more pointed pole and a blunter one is visible,
the nucleus always lies nearer the more pointed pole. Here a spindle is
formed, and the maturation division takes place. The polar cell thus
formed does not lie upon the outer surface of the egg as a free body,
but from the first is pressed into the substance of the yolk (Fig. 2),
as if by a firm membrane, in the manner described by Lameere for
Asplanchna Sieboldii. The nucleus begins to withdraw from the
periphery, at the same time resuming the spherical form, leaving
the polw cell a flattened, disk-like body, not projecting above the
general surfaee of the egg. This condition of the egg is shown in
Figure 2.

From the first, therefore, the polar cell is imbedded in the substance
of the egg, so that it cannot suffer displacement during the processes
which follow. As will be shown in the course of this paper, the place
where the polar cell is formed marks the point on the surface of the egg
opposite to that at which gastrulation takes place. This is contrary to
the statement made by Zelinka (91) for the egg of Callidina, and con-
trary to his general statement for the Rotifera as to the relation of the
place of polar cell formation to the later axes of the embryo. As this
matter is not of especial interest, from the standpoint of developmental
mechanies, a full discussion of the difference between my account and
that of Zelinka is reserved for Part Second. There it will be shown by
evidence from Zelinka's own work, as well as that presented here, that
his general statement of the relation of the place where the polar cell is
formed to the axes of the egg in the Rotifera cannot be considered true

14 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.
for other forms than Callidina russeola, whatever may be the conditions
in that species.

3. ORIENTATION OF THE DEVELOPING EMBRYO.

The first cleavage plane is transverse to the long axis of the egg, and
divides it into two unequal parts (Figure 6). The plane passes through
the place where the polar cell was formed ; the smaller cell includes,
therefore, that end of the egg nearest to which the polar cell is located.
As previously stated, this is also the more pointed end of the egg, when
any difference in the two ends is distinguishable.

The second cleavage is approximately at right angles to the first, and
nearly in the long axis of the egg. It also passes through the region
where the polar cell was formed.

As previously stated, the oval or ellipsoidal form of the egg is re-
taimed throughout the early development. This form is independent ot
the precise arrangement of the material of which the egg is composed. It
is as if the egg substance were enclosed in a rigid mould of oval or ellip-
soidal form. Within this mould the (fluid?) contents may shift their
position widely, without influencing the form of tho mould. Thus, at the
first cleavage, the material of the smaller blastomere occupies all of one
end of the egg. In the ten-cell stage (Plate 3, Figs. 20-25) the same
form is still preserved as if in a rigid cast, but the material which
previously formed the smaller of the first two blastomeres has shifted
from the end to one side of the egg. It therefore is necessary to have
some term by which to designate the two ends of this constant form, as
distinguished from the shifting blastomeres themselves. I shall hence-
forth speak of that end of the egg at which lies the smaller cell in the
two-cell stage as the micromere end of the egg, while the opposite region,
where the larger blastomere lies, will be called the macromere end.
These terms refer to the form of the egg, without regard to the shifting
contents.

The orientation which I shall adopt for the egg itself is similar to
that used by Wilson, Heymons, Conklin, Lillie, Kofoid, and other re-
cent workers on cell lineage. The region where the polar cell is formed,
and which afterward lies opposite the blastopore, will be called the ant-
mal pole ; it marks the dorsal surface. The opposite point is the vege-
tative pole, marking the ventral surface, — the position of the future blas-
topore. Dorsad signifies always toward the animal pole, or place where
the polar cell was formed; ventrad, in the opposite direction, toward

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRIOKII. 15

the region where the blastopore is found at a late stage. The orien-
tation is thus based upon the structure of the gastrula. The chief axis
of the gastrula is the line connecting the animal and vegetative poles.

The first cleavage plane, though coinciding with the chief axis of the
gastrula, is, as shown by the later development, transverse to the long
axis of the embryo. The smaller cell of the two-cell stage is anterior,
since its products occupy the anterior margin of the blastopore ; the
larger cell is posterior, its products forming the posterior lip of the
blastopore. The second cleavage plane, though modified in the pos-
terior part of the egg, is approximately longitudinal. In the four-oell
stage (Plate 2, Fig. 8) the two cells 4? and B°’, resulting from the
division of the smaller cell, AB”, are respectively left anterior and right
anterior, while C? and D’, produced by the division of the larger cell,
OD’, are respeotively right and left posterior.

A section taken transversely to the chief axis of the gastrula will be
spoken of as a transverse section. A section at right angles to this,
passing from anterior to posterior and including the animal and vegeta-
tive poles, is a sagittal section. A section at right angles to both of
these, cutting both the animal and the vegetative pole and passing
through the right and left sides, is a frontal section.

As will be seen from the above, in the two-cell and four-cell stages,
the micromere end coincides with the anterior, the maeromere end with
the posterior end.!

The orientation given above is based upon the relation of the egg to
the axes of the gastrula; the same is true of the orientation used in
most of the recent works upon cell lineage. It differs fundamentally
from the orientation used by Zelinka (791) for the developing egg of
the rotifer Callidina russeola. In that species the egg is of the same
form as in Asplanchna, After extensive shifting during development,
the anterior end (in both Asplanchna and Callidina) comes to lie in the
region of that end of the egg which I have called the macromere end.
Zelinka calls this end of tho egg, therefore, the anterior end, the oppo-
site (my micromere end) the posterior end. Anterior and posterior in
Zelinka’s orientation of course remain constant with regard to the form
of the egg, but not with relation to tho parts of the embryo. Thus, if

1 It is of the greatest importance to observe that I do not use the terms “ micro-
mere end” and * maeromere end” in the same sense in which“ micromere pole“
and “macromere pole" are sometimes used, as-synonymous with “animal pole“
and “vegetative pole.” The two terms are used only as a convenient way of
indicating a peculiarity of the rotifer egg.

16 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Zelinka’s orientation were adopted, the point where the polar cell is
formed might be first ventral, then posterior, then dorsal, and later an-
terior, and this would actually be the case in Asplanchna. In a special
study of cleavage, with particular attention to direction, it is necessary
that the orientation should give some constant basis for reference. It is
therefore impossible for me to use Zelinka’s orientation in my work.
The animal pole of the egg does retain, however, a constant relation to
the position of the blastomeres and to the axes of cleavage, so that I
have adopted this relation as a basis for orientation.

4, CLEAVAGE.
Nomenclature.

For accurate comparative study of the direction and sequence of cleav-
age in the different regions of the egg, such a system of nomenclature is
needed as will indicate directly the relationships, and especially the com-
parative age (measured in cell generations) of the blastomeres. The only
system of nomenclature hitherto proposed which fulfils these demands is,
I believe, that of Kofoid (94). I shall therefore use his system in the
following account.

The four blastomeres of the four-cell stage, and the cells derived
from them, are designated respectively by the letters A, B, C, and D,
beginning with the left anterior blastomere and passing around the egg
to the right, i.e. in the same direction as the hands of a watch, — as-
suming the egg to be viewed from the animal pole. The letters thus
represent the same blastomeres as iu Wilson's work (92) on Nereis,
Heymons’s (93) on Umbrella, Lillie's (95) on Unio, and Kofoid’s (95)
on Limax.

After the first equatorial cleavage, at which the four original blasto-
meres are divided into smaller cells, the capital letters A, B, C, and D
will be reserved to indicate respectively all the cells derived from the
corresponding cell of the first four blastomeres ; and such a collection
will be called a quadrant of the egg.“ The separate cells will be desig-
nated by the lower-case letters a, b, c, and d, according to tho quadrant
to which they belong. Each letter will be followed by two exponents.
The first exponent indicates the generation to which the cell belongs,
the unsegmented egg being considered the first generation. "Thus, in

1 Each quadrant from the four-cell stage onward receives a specific color in the
plates, so that the quadrants are instantly distinguishable by their colors.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. ale

the eight-cell stage (fourth generation) we shall have cells af, bt, c$,
and d*, Since, however, in this and in later generations, there are
more than one cell of a given quadrant in a given generation, this first
exponent must be followed by a second, serving to distinguish each cell
from every other of the same quadrant. In “ spiral” cleavage, this sec-
ond exponent indicates the “quartet,” or layer of cells, in the embryo
to which the blastomere belongs, the ventral cell being number 1, the
next dorsal number 2, and so to the most dorsal quartet. In equatorial
cleavages the same relation may be preserved in other types of cleavage
than the spiral. Thus, in the eight-cell stage (fourth generation), the
ventral blastomeres are d, 0^, c., and d, while the corresponding
dorsal cells are a, 0*2, , and d. But in meridional cleavages,
where there is no trace of the so called spiral, this criterion fails, and
the second exponent can be used only for distinguishing the cells, not
for indicating their relative positions. What is required is a rational
System of applying the exponent such that no two cells of the same
quadrant in the same generation shall have the same exponent. Follow-
ing the suggestion of Kofoid (94), I have in meridional cleavages desig-
nated the right derivative with the even exponent in even generations,
and with the odd exponent in odd generations, — the left derivative of
course receiving the reverse designation. This method of application
was designed to preserve any possible homologies of the products of
meridional with those of spiral cleavage, since in normal spiral cleavage
the right derivative lies above the left in even generations, and so re-
ceives the even exponent, while in odd generations the reverse is true.
The results, however, have not shown any striking homologies with
Spiral eleavage, but the method of application has been retained, since
no other seems to have any advantage over it.

In meridional cleavages, the terms right and left will be used as de-
fined by Kofoid (94, p. 180): “A miniature observer is imagined as
placed in the principal (vertical) axis of the egg, with his head at the
animal pole, facing the part or parts under consideration, and the terms
right and left, upper and lower, are used as determined by this observer."

A full account of this system of nomenclature is given by Kofoid
(94). In order to make clear the relation of the succeeding blasto-
meres and their designations by this system, I give here a scheme of the
nomenclature through the sixth generation, modified from that given by
Kofoid. Only the products of quadrant .4 are carried out beyond
the third generation, since the method is the same for the other quad-
rants. Here a9! represents the most ventral cell derived from the

VOL. XXX. — NO. I. 2

18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

blastomere A, while a5? represents the most dorsal one, the others
occupying intermediate positions.

"E
"x {
jia P
nmi
158 {
gob
A3
[o^ | ai |f oy
41 uch
1 p? H 646.2
ABCD i ait
08 ql
OD:
D3
1 cell. 2 cells. | 4 cells. | 8 cells. | 16 cells. | 32 cells.

First Cleavage.

After the formation of the polar cell the nucleus (Plate 1, Fig. 2)
returns to the position formerly occupied by the germinal vesicle in the
longitudinal axis of the egg, lying nearer that end of the ovum in prox-
imity to which the polar cell was formed (the micromere end). It takes
such a position that a plane at right angles to the long axis of the egg
and cutting the polar cell would also cut the nucleus. The distance
from the centre of the nucleus to the nearer end of the egg is about
two fifths of the length of the egg. Here two asters appear on oppo-
site sides of the nucleus, the line joining them being oblique to the
long axis of the egg. Though I have examined a large number of eggs
at this stage, in no case have I been able to observe a stage in the
process of forming the two asters of the first cleavage spindle (Fig. 3)
from the single aster remaining after the formation of the polar cell
(Fig. 2).

Between the two asters a spindle is formed. This lies at first some-
what oblique to the longitudinal axis of the egg, as shown in Figure 3,
but before cleavage takes place the spindle swings into coincidence with
the long axis (Fig. 4). The aster lying at the micromere end of the
spindle is distinctly smaller than the opposite one (Fig. 3). The nucleus
becomes divided into two small masses, which move toward opposite

l
l
l
l

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 19

ends of the egg, but remain connected for a time by a distinct strand
(Fig. 4). Meantime, before the first cleavage plane has appeared in the
cytoplasm, the aster of the smaller blastomere has begun to divide, as
shown in Figure 4. The two resulting asters separate at right angles
to the axis of the first cleavage spindle. In the future larger cell the
aster does not begin at once to divide. Both nuclei begin immediately
to increase in size. The first cleavage plane passes through the point
on the surface of the egg marked by the polar cell, transversely to the
long axis of the ovum, and through about the middle of the strand con-
necting the two nuclei. The strand is slightly thickened at the point
where the first cleavage plane is to meet it (Fig. 4), indicating perhaps

" The cleavage plane is thus

the formation of the “ Zwischenkórper.
perpendieular to the axis of the spindle, and passes through its middle.
I mention this fact on account of the difference between the first cleay-
age of Asplanchna and that of Callidina. In the latter rotifer, ac-
cording to Zelinka (91), the first eleavage plane is oblique to the
spindle, and the spindle itself, even at the time of division, is oblique
to the long axis of the egg. In another rotifer, Eosphora, the first
cleavage plane is likewise oblique to the long axis of the egg (Tessin,
86), while in Melicerta ringens (Zelinka, 91) and Asplanchna Sieboldii
(Lameere, 90) the first cleavage plane is transverse to the long axis,
as in Asplanchna Herrickii.

During and after the passage of the first cleavage plane through the
cytoplasm, the egg retains its ellipsoidal form, and the resulting cells do
not separate and become rounded, as occurs in the eggs of so many
animals, but remain closely pressed together. In a large series of cases
showing the first cleavage in various stages, the only indication of any
change in the form of the egg or its blastomeres is a slight depression
of the surface where the cleavage plane cuts the periphery of the egg,
forming a shallow furrow. Here the edges of the two blastomeres are
slightly rounded off, as shown in Figure 6, instead of fitting squarely
against each other. The retention of its general form by the egg is
characteristic of all cleavage stages. This surface of contact of the two
blastomeres is curved, the smaller cell, AB’, projecting slightly into
the larger.

So far as the direction of the division is concerned, the first cleavage
of Asplanchna evidently fits easily either the surface tension theory of
Berthold, or Hertwig’s theory of the spindle in the long axis of the pro-
toplasmie mass. Comparison with the first division of Callidina russe-
ola as described by Zelinka (’91) develops an interesting fact. In

20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.
Callidina the first cleavage spindle is oblique to the long axis of the egg,
therefore not in agreement with Hertwig's law ; but immediately after
division is finished, a movement of the egg contents takes place in
such a way that the two cells occupy the same relative position as in
Asplanchna, — such a position, therefore, as is demanded by Berthold’s
theory of least surfaces. It thus appears that in Callidina the direc-
tion of division itself is determined neither by the principle of Berthold
nor that of Hertwig, but that the later arrangement of the cells might
be held to be due to the action of Berthold’s principle. It is somewhat
curious that the exact arrangement produced in Callidina by shifting
should in Asplanchna result at once from the position of the spindle at
the time of cleavage.

No cause can be assigned, from the visible structure of the egg, for
the inequality of the cleavage. The yolk granules are distributed uni-
formly throughout the egg, seeming no more abundant in the large
than in the small cell.

Second Cleavage.

As a result of the first cleavage, the egg is now composed of two un-
equal blastomeres, an anterior, AB’, and a posterior, CD" (Figs. 5 and 6).

In the smaller blastomere, as previously stated, the aster has already
divided and the two parts are separating at the time when the first
cleavage plane passes through the cytoplasm (Fig. 4). The line along
which they move apart is perpendicular to the axis of the first cleavage
spindle, aud also at right angles to a line connecting the polar cell with
the centre of the egg. The forming spindle is thus parallel to the
lateral axis of the embryo and consequently perpendicular to its dorso-
ventral axis. The two asters take up their positions on opposite sides
of the nucleus, and the axis of the resulting spindle has a direction
parallel to the line joining the asters at their first separation (Wigs. 5
and 6). Meanwhile the nucleus has steadily inereased in size, up to the
time when it participates in the formation of the spindle.

In the larger cell, CD’, the order of procedure is different. The
nucleus begins to enlarge, as in the smaller blastomere, but the aster
does not at once divide. The nucleus and aster together begin to
migrate to the right. At the same time the aster comes to lie farther
to the right than the nucleus, either because the two rotate on a com-
mon axis, or because the aster, moving faster, creeps around the nucleus
toward the right side of it. "Thus, whatever the mothod, a condition is
reached in which the large nucleus lies in tho right anterior angle of the

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 21

cell OD", with the undivided aster at its right and slightly behind it
(Fig. 5).

By this time the two asters in the cell AB” have completely separated
and lie upon opposite sides of the enlarged nucleus. Thus the prepara-
tory stages for karyokinesis are much more advanced in the smaller
cell, and it would be anticipated that this cell would cleave first.

The single aster in CD now begins to divide. The process seems to
be accomplished very quickly, since in a series of nineteen specimens of
the two-cell stage (each, of course, taken from a different individual)
only one case was found exhibiting a transitional stage between that
shown in Figure 5 and that shown in Figure 6. In this specimen the
single aster had elongated slightly in the direction of the future
spindle. When formed, the spindle takes an oblique position in the
cell, extending from right anterior to left posterior. The aster at the
left posterior end of the spindle is much the larger, in correlation appar-
ently with the larger mass of cytoplasm surrounding it. The nucleus
of CD' has now overtaken in its metamorphosis that of AB’; the
spindles are found in exactly corresponding stages, the chromatin being
in both arranged in an equatorial plate (Fig. 6).

Not only are the two spindles not parallel, as shown in Figure 6, but
they do not lie in the same plane. If the two spindles are viewed
exactly from the anterior or from the posterior end of the egg, the left
aster in AB” and. the right aster in OD" are seen to lie more dorsally
than their mates. Viewed in this direction, the spindles cross each
other at an angle of about twenty-five degrees.

As a result of the dissimilarity in the direction of the two spindles,
the two next cleavage planes, perpendicular to them, will not meet the
first cleavage plane in a common line. The position and direction of
the spindle in OD” are such that the cleavage plane cutting OD?
would probably meet the first cleavage plane to the right of the line
where the plane dividing 42^ would meet it. Since the right aster of
CD' is farther dorsal than the left, the plane of cleavage of OD?
would be inclined to the sagittal plane, — on the dorsal side toward the
left, on the ventral side toward the right.

The cleavage of the two cells now follows at almost or precisely the
same time, the karyokinetie processes being found from this time on in
the same stage. Ina series of thirty-one eggs from different individ-
nals, each containing more than one and less than five cells, none con-
tained exactly three cells.

An examination of the four-cell stage after the completion of division

22 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

(Plate 2, Fig. 8) shows that the cleavage planes have taken the posi-
tions foreshadowed by the arrangement of the spindles. The plane
separating C? from D? lies to the right of the plane separating A? from
B?; the corresponding furrows on the surface are nearer together on
the dorsal than on the ventral side. The blastomeres resulting from
the division of AB” are equal, whereas CD divides very unequally.
The right derivative (C?) is much smaller than the left (D*), and is of
approximately the same size as A’ and 3°. The blastomeres B? and
D? are in contact along the whole distance from the dorsal to the ven-
tral surface of the egg, while 4? and O? do not touch each other at all.
The polar cell lies either at the junction of B°, Cs, and Ds, as shown in
Figure 8, or sometimes at the junction of 4% B and D. The egg is
now markedly unsymmetrical.

It is evident from the above description that this cleavage may be
considered as belonging to the so called spiral type. Since the left end
of the spindle is in each cell the higher, tho cleavage is a left spiral,
like the corresponding cleavage in Discocelis, Nereis, Limax, and indeed
all forms with spiral cleavage except in the reversed cleavage of certain
mollusks. This fact is striking, since the succeeding cleavages in
Asplanchna do not belong to the spiral type.

The relation between the axes of the embryo in later stages and the
first two cleavage planes is as follows. The first furrow separates an
anterior from a larger posterior portion, but the plane of separation of
the parts bears no simple relation to the axes of the later embryo.
(Compare Figure 8 with Figure 75, Plate 9, in which the parts derived
from the first four cells are colored in the same manner as their parent
cells in Figure 8, and note the great shifting.) The later sagittal plane
of the embryo is coincident with a plane passing through the animal pole
and the longest axis of the egy; that is, through the plane separating A’
from B? (Fig. 8), and dividing the larger blastomere D’ into two un-
equal parts. The second cleavage plane therefore divides the right side
from the left in the anterior part of the egg; but in the posterior part it
lies entirely in the right side. It is not until the seventh cleavage that
the division into symmetrical right and left halves takes place on the
posterior side (Plate 7, Fig. 58) ; indeed, certain cells containing material

for both sides of the egg remain undivided till even a later st

go,
Third Cleavage,

Immediately after the second cleavage, the aster in ench of the four

cells produced begins to extend dorso-ventrally, at right angles to the

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 23

axis of the preceding spindle. An optical section of the egg along its
chief axis, showing the asters in the cells B® and Ds, is given in Plate
2, Fig. 10. Spindles are formed in all four cells nearly or quite in
the position indicated by the direction of separation of the asters.

The tendency of the karyokinetie processes in the posterior half of
the egg to gain upon those in the anterior half, shown during the last
division, is continued and accelerated. Spindles appear in C? and Ds,
while the nuclei in A? and B? are still spherical and have distinct mem-
branes. Figure 9 gives a view of this stage from the right side; the
large spherical nucleus of B? is represented by a broken outline. The
spindle in C? has a dorso-ventral direction, and its middle coincides with
the middle of the length of the cell; the two asters are of equal size.
In D’ the spindle is nearer the dorsal side of the egg, and is inclined,
passing from dorsal and anterior to ventral and posterior. The ventral
aster is the larger.

Cleavage takes place first in the larger cell D, separating a large
ventral blastomere, d, from a smaller dorsal one, d. At the same
time the two cells (considered as a whole) elongate dorso-ventrally.
In so doing, the ventral blastomere, d*!, remains nearly stationary,
while d *? moves in the direction of the animal pole of the egg. (Com-
pare Figure 12, a sagittal section of a five-cell stage, with Figure
10, the corresponding section of a four-cell stage, observing the posi-
tion of the cells in relation to the general form of the egg.) As a
result of this, the dorsal end of the cell B, and, to a less degree, the
ends of A? and , are displaced in the same direction; that is, the
whole animal pole moves toward the micromere end of the egg. At
the same time the cells AP, B°, and C? are slightly compressed dorso-
ventrally. This is the beginning of that peculiar rotation of the blasto-
meres in the eggs of Rotifera, described by Zelinka (91) and others,
which eventually results in the process of gastrulation.

C? divides next, the cleavage being equal; the products are c*
End gms

Before the cleavage is finished in D? and Cs, spindles have been
formed in B° and AP, division taking place in them in the order named.

1

The cleavage is equal, as in O°.

The order of cleavage, then, for the four cells, is as follows: D, €, B, A.
This rhythm reappears in later cleavages.
The third cleavage is therefore equatorial, dividing the egg into two
layers of four cells each. The ventral cells are - d, the dorsal cells
d d. The egg is still slightly unsymmetrical.

24 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Views of the eight-cell stage are shown in Figures 15 to 18.

From a cyto-mechanical standpoint, the third cleavage may be char-
acterized as follows. The first division of the asters is along a line at
right angles to the axis of the previous spindles, and indicates the posi-
tion of the spindles for the next cleavage. These lie in the long axes
of the cells, and the cell walls are formed in the position demanded by
the principle of least surfaces.

Fourth Cleavage.

Immediately after the division of D” (Plate 2, Fig. 12), the asters in
d^! and d^? begin to extend laterally, at right angles to the axis of the
preceding spindles, and each becomes divided into the two asters for the
following spindle. In d*', Figure 11, the two asters for the succeed-
ing cleavage are still connected by a striate band. Figure 11 shows a
ventral view of the same egg as Figure 12, the five-cell stage. The
corresponding dorsal view of a slightly later stage is shown in Figure
14. The asters in d^? are moving apart in the same manner as in d^",
save that the line of separation is slightly oblique, the left aster being
higher.

In the same way the asters in the cells a*1—c41 and at? - c*? become
constricted, and divide at right angles to the axis of the preceding
spindles, The dividing asters in c^? are shown in Figures 14 (Plate 2)
and 17 (Plate 3), and those of e*! in the latter figure. Views of the
other four cells would show similar conditions.

From the manner in which the asters separate in all of the eight
cells, one would be led to expect that the next cleavage would be merid-
ional, at right angles to the third cleavage. This expectation is
strengthened by the fact that the lateral dimensions of the cells in
which the asters lie are considerably greater (in the quadrants A, D, and
O, at least) than the opposite measurements (Fig. 17).

Jut in a slightly later stage it is observed that the line joining the
asters in d*! has become oblique, like that joining those of d^? (as
mentioned above). This oblique position of the asters in di is shown
in the ventral view (Fig. 15). The left aster (right side of the figure)
has become ventral, the right one dorsal. The sagittal section (Plate 1,
Fig. 7) of a slightly later stage shows the completion of the rotation
thus begun ; the line connecting the two asters and passing through the
nucleus is now approximately dorso-ventral in direction.

At the same time a similar rotation has taken place in the cell Ur
but the position taken by the two asters is not the same as in d. One

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 25

of the two asters has become central, the other peripheral, as if the cell
were about to divide into a deep and a superficial portion. This condi-
tion also is shown in the sagittal section, Figure 7. The difference in
the position of the asters in d*" and dee is apparently due to simple
mechanical conditions, — the form of the cell ds compelling the asters
to take the position which they have.

At the same time a slight differentiation in the cytoplasm of the cell
di becomes visible. As previously stated, the fine yolk granules are
at first distributed uniformly throughout the egg. In this eight-cell
stage, a slight concentration of the yolk granules in the ventral part of
the cell d*! may be noticed by careful observation. The condition at
this time is shown in Figure 7; in the ventral part of d*! the yolk
granules are a little larger and more numerous. As will be shown, this
concentration of yolk becomes later much more marked, and its history
is peculiar.

A spindle is now formed in d** in the position indicated by the asters
of that cell in Figure 7, — that is, with a dorso-ventral axis, — thus pre-
figuring another equatorial cleavage. The spindle is shown in Plate 2,
Figure 16.

Immediately thereafter the spindle is formed in ds, and it appears
that the position of the asters shown in Figure 7 (Plate 1) is not define-
tive. The asters shift, so that the spindle in d** is dorso-ventral, like
that in d*!, as is shown in Figure 16. Which of the two asters seen in
Figure 7 becomes dorsal and which ventral, I have been unable to deter-
mine. During the formation of the spindle in d*? the cell extends a
little in the direction of the spindle, as is shown by a comparison of
Figure 7 with Figure 16.

Meanwhile, changes have been occurring in the quadrants 4, B, and
©. As the processes are the same in all three, the quadrant C will be
selected as a type.

At first, as described above, the asters separate tangentially, at right
angles to the axis of the previous spindle (Plate 3, Fig. 17). This
position is retained for some time, but in a later stage the line connect-
ing the asters in c has become oblique, as shown in Figure 18, which
exhibits a side view of the egg of which Figure 7 is a section. The
asters in c** still retain their original position.

Now follows the cleavage of the cell d,. This is accompanied by an
increase of the dorso-ventral extent of the two products, as compared
with that of the original cell. The division is unequal; the ventral
cell d5* is much the larger, and retains the whole of the territory con-

26 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

taining the larger yolk granules shown in Figure 7. The larger deriva-
tive retains its position at the macromere end of the egg (Fig. 19).
The smaller cell d' is therefore pushed dorsad, and this, together with
the extension of d*? at the time of the formation of its spindle, dis-
places the animal pole, marked by the polar cell, still farther toward
the micromere end of the egg. As a further result, the cells of the
quadrants A, B, and C are still more compressed dorso-ventrally, so
that, especially in a*! — c. , the lateral extent is much greater than the
dorso-ventral (Fig. 19).

Nevertheless, as Figure 19 shows, the rotation of the future spindle
axis still continues. The line joining the asters becomes dorso-ventral
first in the dorsal cells 4 — , while in a** — c*4 the asters are still
oblique, as shown in c, Figure 19. In quadrant B of this same figure,
the axis has become dorso-ventral in both cells.

Now occurs the cleavage of d*?, with still further elongation, shifting
of the animal pole toward the micromere end of the egg, and resulting
greater compression of the cells of the quadrants A, B, and C (Figs.
20-24). Without regard to this, the asters in the cells of these quad-
rants continue their movements until the future spindle axes are in
every case dorso-ventral. Spindles are now formed in all of the six
cells, the spindle being in every case in the shortest axis of the cell
(Figs. 20-24).

The conditions at this stage are so significant from a cyto-mechanical
standpoint, that I have thought it best to analyze and illustrate with
especial fulness a typical egg at this stage. Figures 20-95 are views
of a single egg. Figure 20 shows the right side (quadrant C), Figure
22 the left side (quadrant A), Figure 21 an anterior view (quadrant B),
and Figure 25 a posterior view (quadrant D). In Figures 23 and 24 are
given respectively sagittal and frontal optical sections, for comparison
with the surface views.

In order to exclude possibility of error, the egg from which the above
figares were taken was moved about so that the six cells belonging to
the quadrants A, B, and C occupied successively the middle of the
upper surface of the egg; careful camera figures of each cell were made
in this position. Then optical sections were taken in the same way,
both along the axis in which the spindles lie, and at right angles to
these. Accurate measurements of the dimensions of the cells could
thus be made; the results are as follows.

JENNINGS: DEVELOPMENT OF ASPLANCHNA IIERRICKII. 27

Dorso-ventral Lateral Ratio of Dorso-ventral to
Cell, Measurement, Measurement, Lateral Measurement.
atl 25u (2041) 52u 1to 2 (about) (2 to 5!)
at2 35 50u in
btl 24u 42u 4“ 7
i2 35 42 DaN
one 30% 49 3 “ 56 (about)
oe 33 65 2 “ 3 (about)

After the spindles become completely formed, the cells begin to
elongate in the direction of the spindles. A slightly later stage than
that just described is shown in Plate’ 4, Figure 26. Comparing this
with Plate 3, Figure 22, it is evident that the cell % has stretched in
the direction of the spindle to such an extent that the difference between
the two axes of the cells is much diminished. Nevertheless, in both this
cell and a*? the axes in which the spindles lie are distinctly the shorter.
This is still true at the time of the division of the cells. Figure 27
(Plate 4) shows the right side of the egg last considered ; in the quad-
rant C the processes are much more advanced than in A. The nuclei
have separated and the cytoplasm is dividing, yet exact measurements
both of surface views and optical sections show that the greater diam-
eter is still at right angles to the line joining the two nuclei. A frontal
seotion, showing the greatest dorso-ventral extent of the cells of the
quadrants A and C of this egg, is given in Fi

887

gure 28.

The two figures last mentioned show another fact of importance. The
divisions do not separate the blastomeres into cells of equal size in the
quadrants 4, , and C. The completed cleavage is shown in Plate 4,
Fig. 30 (anterior view). This, with the figures just cited, shows that the
cells at? — c*? divide very unequally, the dorsal derivatives, a?*— c**, being
very much larger than the ventral ones, a$? — 65.8. The inequality is less
in the division of the ventral cells a*! —c**, Although the ventral de-

1 1

rivatives, a? — c*!, occupy a larger area on the surface of the egg, there

is little difference in actual volume, and such as occurs is in favor of the
more dorsal cells a5:? — £52
The order of division is the same as in the last cleavage ; first, the

quadrant D, then in order C, B, A. In quadrant D tho

divide at the same time,

The important facts in this fourth cleavage, from a cyto-mechanical

standpoint, may be summarized as follows.

! In every ease the flrst measurement was taken through the two asters; in the
(ase of at the real dorso-ventral extent of the cell, into which the spindle later
novcs, is but 205, — so that the ratio is as two to five.

28 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

The asters in all of the eight cells after the third cleavage separate
tangentially, at right angles to the direction of the preceding spindles.
Having taken up positions on opposite sides of the nucleus, in every
case the complex of nucleus and asters rotates in such a way as to bring
the axis of the forming spindle into the same direction as that occupied
by the spindle for the preceding cleavage. In six of the cells, this ro-
tation is from a previous position in the longer axis of the cell to a later
position in the shorter axis. In these six cells spindles are completed in
the shortest axes of the cells, and division ensues in such a way that
the newly formed septa are surfaces of greatest area, and the cells sepa-
rate in the direction of greatest pressure.

The cleavage in d - c*? and in de is markedly unequal.

In two cells of equal ages but unequal size (4% and de) the larger
divides first.

The exact changes in form during the divisions of the cells is a point
worthy of careful attention. As the transformation of the nucleus
giving rise to the spindle takes place, the cell elongates slightly in the
direction of the spindle. (Compare c, Figure 24, with the earlier
stage of the similar cell d in the same figure.) As the spindle nar-
rows and lengthens and the chromosomes bogin to separate, tho cell
continues to elongate (Fig. 26, , compared with Fig. 20, %, and Fig.
22, a^). As the two new nuclei are formed and move apart, and the
cytoplasm becomes constricted, there is a still further extension of the
cells in the direction of the spindle. (Compare c and c*? with a*! and
a*?, in Plate 4, Fig. 28.)

As Heidenhain (94e, p. 154) has recently urged, this elongation of
the cell in the direction of the spindle is a point of great importance
for a proper understanding of the conditions affecting the direction of
cell division. In many later divisions in Asplanchna the spindle is first
formed, as will be shown, in the short axis of the cell, and then this
axis by stretching becomes the longer. It is possible that to this phe-
nomenon is due the apparent general agreement of normal cleavage with
the law of Hertwig, and that careful observation will in many cases,
as in Asplanchna, show the so called law to be of little significance.

A full discussion of the bearing of the facts above described is
reserved until later cleavages have been examined.

The foregoing description is based on a study of forty-two specimens
from different individuals, showing the various phases of the fourth
cleavage ; that is, each containing more than seven and less than six-
teen cells.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 29

This cleavage, as described above, differs in some respects from that
of Callidina as described by Zelinka (91). A discussion of the differ-
enees will be found in Part Second.

The cleavage of Asplanchna priodonta takes place to this stage in
exactly the same manner as that of Asplanchna Herrickii. Figure 29
(Plate 4) shows the egg of Asplanchna priodonta in the 10-cell stage.

Fifth Cleavage.

As a basis for an account of the following cleavage, it will be well to
summarize the divisions which have already taken place, and to take a
careful survey of the structure of the egg at the end of the fourth
cleavage.

The first and second cleavages pass through both animal and vegeta-
tive poles and are therefore meridional. The third cleavage is at right
angles to the dorso-ventral axis and is therefore equatorial. The fourth
cleavage is parallel to the third, thus likewise equatorial.

As a result of these cleavages, the egg now consists of sixteen cells,
arranged in four dorso-ventral rows or quadrants, each quadrant con-
sisting of four cells derived from one of the four blastomeres of the
four-cell stage (Plates 4 and 5, Figs. 30-36). Passing from the ven-
tral side dorsad, we may also distinguish four layers of cells, each layer
containing one cell of each of the four quadrants: The layers may for
convenience be numbered ; I will call the ventral the first layer, the
others following in order to the fourth, which is at the animal pole. As
a result of the shifting during cleavage, the animal pole has now come to
be situated almost exactly at the micromere end of the egg ; the oppo-
site end is occupied by the large cell d** (Figs. 30 and 33, anterior
and posterior views respectively). The dorso-ventral axis therefore
now coincides with the long axis of the egg.

Of the four quadrants, three, A, B, and C, are alike in the size and
arrangement of the cells of which they are composed. (See Fig. 30,
anterior view.) The four cells composing any given one of these quad-
rants differ in size. The cell of the fourth layer (next to the animal
pole) is much the largest, while that of the third layer is much the
smallest. The cells of the first and second layers are nearly equal in
size; that of the first layer covers more of the surface of the egg (Fig,
30), but that of the second layer is deepest (Fig. 32). The cells of the
first, second, and third layers are much compressed dorso-ventrally, so
that the lateral dimensions of the cells are at least twice as great as the
dorso-ventral dimensions. In the third layer especially, the cells are

30 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

deformed by the pressure to such an extent that the surface area ex-
posed resembles the section of a biconvex lens. In the dorsal layer
the compression is less ; the cells are triangular in surface view, and the
dorso-ventral extent is greatest.

All the cells of quadrant D are much larger tham the corresponding
cells of the other quadrants (Figs. 31 and 33). The ventral blastomore,
de, is much the largest cell of the egg, and occupies the entire ventral
end at this period. Its position as shown by the section (Fig. 32) is
worthy of careful attention. Its dorsal or inner surface, like tho outer,
is convex ; anteriorly the cell is partly covered by the ventral cells of
the other quadrants, while the ventral end of cell dë? extends a slight
distance ventrad of the middle dorsal portion of d51,

The cell d°" is distinguished from all the others by a further peculi-
arity. I have shown above (page 25 and Fig. 7) that in the eight-cell
stage there is a slight concentration of yolk material in the ventral
region of the cell dt, where the yolk granules are a little larger and
more numerous. At the division of d*!, this cloud of granules, as a
natural result of its position, remains in the cell d** (Plates 2 and 3,
Figs. 16 and 19). At the same time it becomes more distinctly differ-
entiated. The granules composing the cloud increase in size and range
themselves about tho periphery of the egg, next to its free surface
(Figs. 23, 24, Plate 3, and Fig. 28, Plate 4). A narrow strip of the
posterior margin of the free surface of the cell is without the granules
(Figs. 20, 22, and 23). At a time when the fourth cleavage is entirely
completed, tho granules have withdrawn still farther from the posterior
margin of the cell, and show a tendency to concentrate at the free sur-
face of the cell over its anterior half (Plate 4, Fig. 32). In the other
cells, and in the remaining portions of 4, the original finely granular
cytoplasm is retained, so that I have not thought it necessary to repre-
sent in the figures the yolk conditions in any region except where the
cloud of granules is present.

The cell of the second layer, d*?, is next in size, then the dorsal cell,
dend, while the cell of the third layer is the smallest in quadrant DÐ.
The cells d^? and d*? are very greatly compressed dorso-ventrally and
elongated laterally, so as to form irregular flat plates, extending from the
posterior surface of the ege two thirds of the distance to the anterior
surface. (Compare Figure 31, left posterior surface, with Figure 32, sec-
tion.) The dorsal cell d** is likewise compressed dorso-ventrally, so as
to appear in a sagittal section (Fig. 32) as a low triangle.

As a whole, the form and arrangement of cells are far from what is

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 81

demanded by the principle of least surfaces. Flat plates, such as we sce
in d”? and de, and in all the cells of the third layer, retain their form
in virtue of some force working strongly against surface tension.

After the fourth cleavage, the asters in all the cells at first separate
at right angles to the axes of the preceding spindles, as happened after
the third cleavage. The later changes are essentially the same in all
the quadrants, so far as the asters are concerned, so that quadrant D
may be described as a type.

Figure 31 (Plate 4) shows the conditions in the four cells of quadrant
JD, after the asters have divided. The two asters of each cell lie upon
opposite sides of their nuclei in such a position that, if no change
occurs, the ensuing division will be meridional.

In the cells of the first three layers the asters retain their original
positions. But in ds a rotation takes place, such as occurred in all the
cells in preparation for the preceding (fourth) cleavage, so that the axis
of the spindle in d9? is at right angles to the axes of the spindles in the
other three cells of the quadrant. This condition is shown in Figure
33, and the completion of the division is shown in Plate 5, Fig. 37.

The same processes take place in the other quadrants, so that all tho
cells of the first three layers have spindles extending laterally, while in
the fourth or dorsal layer the spindles are directed dorso-ventrally
(Fig. 40).

We must now eonsider the cleavage in the several cells more in
detail.

As in previous cleavages, division takes place first in the cells of
quadrant D. The nucleus of the large ventral cell, d??, is earliest to
enter upon the karyokinetie process, followed immediately by d°*, and
a little later by dë? and d. Figure 33 (Plate 4) gives a view of the
posterior surface of the ege at this stage, showing the spindles in all the
cells. As this figure shows, the spindles in the three cells d, d5*,
and d? do not lie in the middle of the cells, but nearer the right ends.
(See definition of right and left, page 17.) In the dorsal cell, &., the
spindle is at right angles to those in the other cells. The plane of
cleavage indicated in the three ventral cells is meridional ; in the dorsal
cell it is equatorial, like that of the two preceding cleavages. The division
of ench of the four cells must be considered separately.

To understand the cleavage of the large ventral cell, 457, it is neces-
sary to observe accurately its position and relation to the other cells.
A longitudinal section of about the same stage as that shown in Figure

33 is given in Figure 34. Comparing this with the earlier correspond-

*

2 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

—

ing section, Figure 32, it is seen that with the formation of the spindles
in dë? and d’? these cells have yielded to the well known tendency to
take a more rounded form at the time of karyokinesis; the inner parts
of the cells have been withdrawn toward the surface and used in in-
creasing the dorso-ventral dimensions of the cells. The animal pole
has been thereby pushed still farther in the direction in which it has
been steadily migrating, so that it is actually past the micromere end of
the egg. The cells of the quadrants A, B, and C, being in a “resting ”
condition, give way to the compression, and become much deeper and
flatter than before. The cell . retains its position at the macromere
end of the egg, but lying in a concavity, partly surrounded by the
other cells. The spindle lies in the deeper (more dorsal) parts of the
cell, with its right end (Fig. 33) deepest, and close to the wall of
the cell A view from the ventral end of the egg (Plate 5, Fig. 35)
shows that this “right” end is really anterior, and that the spindle lies
in an antero-posterior plane, coincident with the plane separating the
quadrants A and B. The anterior (inner) end of the spindle lies close
against the boundary between a’! and 054,

The division which now ensues is of an extraordinary character. The
anterior end of the spindle is pressed against the periphery of the cell
at the place above mentioned, and a minute vesicle is given off, which
lies embedded between the cells / and 053, This, after the division
is finished, is shown in Figure 38 (Plate 5), the vesicle being
labelled de..

During division, the granular cloud which was described as occupying
the anterior half of the periphery of the cell moves still farther toward
the anterior margin, and shows a tendency to concentrate into a more
definite group; the individual granules become larger also (Fig. 38).

In d*? and d*? the spindles are parallel to the spindle in d*', the
right ends being nearer the boundary of the cells, and deeper within
the egg. The latter fact is shown in the transverse section, Figure 36,
passing through the cells of the third layer. The divisions are un-
equal, as foreshadowed by the position of the spindles, but tho inequality
is much less than in the case of d*'. The completed division is shown
in Figure 37. The cleavage takes place first in d“.

At about the same time as the division of d*? occurs that of the dor-
sal coll, d*-^, Here the spindle is in the short axis of the cell; the
cleavage is equatorial and unequal, the dorsal cell being much the
smaller (Fig. 37).

During the occurrence of the cleavage of these cells other changes,

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKI. 29

which are of the greatest importance, have been taking place, partly as
a consequence of these cleavages. As the constriction of the cyto-
plasm in dë? and dë? occurs, these cells show in a most pronounced way
the tendency to become of a rounded form. The inner portions of the
cell are withdrawn still more from the centre of the egg, until the an-
tero-posterior measurements are no greater than their dorso-ventral
dimensions, This is shown in the section, Figure 38. In this egg, a
surface view of which is given in Figure 37, the cleavage of dë? is
finished, and the products, des and de, have already passed into the
“resting stage," so that they take whatever form is impressed upon
them by the surroundings. But 45.8 is just dividing into de and ders,
and the form shown in section by d95 in Figure 38, as compared with
the form of des in Figure 32, shows the change which I have been
describing.

At the same time the cells of the other quadrants, A, B. and C, are
entering upon the stages preparatory to karyokinetio division. As a
first step they also retract their deeper parts and bring their protoplasm
into a more compact mass, as shown by a comparison of quadrant B
in Figure 34 (Plate 4) with the similar quadrant, A, in Figure 38
(Plate 5).

As a consequence of this withdrawal of material from the inner parts
of the egg, the large ventral cell de, which has now passed into the
resting stage, moves inward to occupy the space which would otherwise
be vacant, — being forced to do so, of course, by tho greater dorso-ventral
extension of all the other cells. The result is shown in Figure 38.
This partial enclosure of d' by the other cells is of course a stage in
the process of gastrulation.

Before the cleavage of all the cells of quadrant D is finished, the
karyokinetie processes have begun in the other three quadrants. (See
Plate 5, Figs. 39-42.) The first cells to show the characteristie nu-
clear phenomena are those of the fourth or dorsal layer, % — . As
previously stated, the asters at first take up such a position in these
cells as would lead, if unchanged, to a meridional cleavage. But in these
three cells, as in d55, there is a revolution of the asters and nuclei,
resulting in a dorso-ventral position of the spindles. At any given period
the three cells are not in exactly the same phase of division, though very
nearly so; the order, beginning with the most advanced, is c4, 054, ad,
The sequence is thus the same as in previous cleavages. The division is
equatorial and the dorsal product is the smaller, as in the cleavage of
the corresponding cell (dss) of the left posterior quadrant. In tho three

VOL. XXX. — NO. 1. 3

34 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

cells of quadrants A, B, and C the spindles are not in the short axes, as
in d°*, since these cells are not nearly so broad as that one, while the
dorso-ventral dimension is about the same or greater (Plato 4, Figs. 30
and 32). Figure 41 (Plate 5) shows the process of cleavage in these
cells, while Figure 45 (Plate 6) shows the cleavage concluded.

The division of the second layer follows upon that of the fourth, and
again in the order c9?, 052, 4. The division is here meridional, as it is
in the corresponding cells of the quadrant D, and equal, as it is not in the
corresponding cell ot the quadrant D. (See Figs. 39, 40, 43, and 44.)

Next follow the divisions of the cells of tho first layer, in the same
order as in the previous cleavages, and, slightly later, the divisions in
the third layer, also in the same sequence. The cleavages here also are
meridional and equal.

The nuclear conditions leading to these cleavages are shown in Figures
39, 40, 42 (Plate 5), all from the same egg. A somewhat later stage
is shown in Figure 42, which exhibits the conditions in the quadrants
A, B, and C at a time when the divisions just described are completed
in most of the cells. All the divisions are nearly finished except in the
cells of the third layer (a59— 099), which still contain spindles. This
view shows also another peculiar fact. During the cleavages the cells
about the dorsal pole of the egg have shifted, and the cells a9" and c9
have pushed ventrad to such an extent that on the right side tho cells
c5? and b8 have become completely separated, a part of the cell cê"
lying between them. This condition is only transitory, however; the
cell c is very soon pushed dorsad again, and the cells of the third layer
again form a continuous row. (Compare Figure 47, Plate 6.) Figure 46
shows the cells of the quadrant D at the close of this division, while
Figure 45 is a view of the animal pole at the same stage.

The features of the fifth clenvage may be summarized as follows. In
the twelve cells of the three ventral layers, the asters separate after the
fourth cleavage at right angles to the position of the preceding spindles
and retain the position first taken ; the cleavage is therefore meridional.
In all these cells the spindles are in the long axes of the cells. In the
dorsal layer the asters at first assume the same position as in the othor
cells, but later a rotation takes place, and the spindles when formed
have a dorso-ventral direction ; the resulting division is equatorial. Tho
5.4

spindle is in the longer axis of the cells a5, 055, and c!

720 „ in the shorter
Axis in d.
The division is unequal in the four dorsal cells of all the quadrants,

and in all the cells of the quadrant D. In the other cells it is equal.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 35

The sequence of cleavage is the same as in previous divisions, but
with some modifications. In any given layer of cells, the order is D,
C, B, A; a repetition of the sequence established at the third cleavage.
In any given quadrant, the order of cleavage varies with the relative
size of the cells. In the quadrant D the order is (beginning with the
ventral cell) 1, 2, 4, 3, and this is also the order of relative size of the
cells, beginning with the largest. In the other quadrants the order of
cleavage is 4, 2, 1, 3, and this again is the order of comparative size
beginning with the largest, except that 2 and 1 are so nearly of a size
that it is difficult to say from observation that either is the larger.

The fifth cleavage is accompanied, as a result of the changes in form
of the cells during karyokinesis, by a partial enclosure of the ventral
cell of quadrant Y (d) by the other cells.

The above account of the fifth cleavage is based upon an examination
of twenty-five eges, taken from different individuals and showing differ-
ent phases of the division; i.e. each egg contained more than fifteen
and less than thirty-two cells.

Sixth Cleavage.

The first division belonging to the sixth cleavage, that of de, takes
place coincidently with the last division of the fifth cleavage, that of a".
There is thus no resting period between the two cleavages. Neverthe-
less, there is a sufficiently well characterized stage of thirty-one or
5.1

thirty-two cells, just as the cleavage of de occurs, and it will be well

to describe the egg in this condition as a basis for an account of the
sixth cleavage.

Figure 43 (Plate 5) shows the anterior surface just before this stage
is attained ; Figure 47 (Plate G) shows nearly the same surface after the
fifth cleavage is finished. The posterior surface is shown in Figures 46
(Plate 6), 53, and 54 (Plate 7); the animal pole, in Figure 45 (Plate
6). Figure 48 shows a sagittal section, while a transverse section of a
stage just later (looking toward the animal pole) is given in Figure 52.

The principal axis of the egg still coincides with its long axis, the
animal pole lying at or near the micromere end, the vegetative pole at
tlre macromere end (ig. 48).

The egg now consists of (1) a single large cell, d**, embedded within
the other cells and appearing on the surface at the ventral end only
(Plate 5, Fig. 38), and (2) of thirty-one smaller cells, partly surround-
ing the larger cell %%. One of these smaller cells, 4, is a minute

vesicle embedded between the cells 45, 654 and q9* (Figs. 38 and 42).

96 BULLETIN: MUSEUM OF COMPARATIVE ZOÜLOGY.
The other thirty cells show a regular arrangement. The four quadrants
may be distinguished as at the beginning of the fifth cleavage, each
quadrant now containing eight cells, showing a characteristic arrange-
ment. From ventral to dorsal we may now distinguish five layers. The
first three layers contain each eight cells, two from each quadrant. In
the quadrants A, B, and C, the two cells of a given layer are equal
(Plate 6, Fig. 47). In quadrant D there is great irregularity. The
two cells of the ventral layer are extraordinarily unequal, constituting
the large partly interior cell de, and the minute vesicle de., also
enclosed within the other cells (Plate 5, Fig. 38). In the second
and third layers the two cells are likewise unequal, though less
markedly so.

The fourth layer consists of a single large cell from each quadrant,
that of quadrant D being the largest (Plate 6, Fig. 45).

The fifth layer consists of four small cells at the dorsal pole of the
egg (Fig. 45). The arrangement at the animal pole formed at the four-
cell stage (Fig. 8) is still maintained. The quadrants Band D are in
contact for a considerable distance, whereas A and C do not touch. In
one of the points where three cells of different quadrants meet (in this
case 095, £95, and des) Ties the polar cell.

The first cells in which indications of cleavage are observed are again
the large cells of the D quadrant, d?" and d°°. Spindles are formed in
these at about the same time (Plate 6, Fig. 48). The processes taking
place in the two cells differ, and must be considered separately.

[AUNT
Pit

In d, after the giving off of the small vesicle d“, the nucleus very
quickly enlarges to its original size. The aster begins to elongate at
right angles to the position of the previous spindle (Plate 5, Fig. 38).
But at almost the same time a rotation takes place, and by the time
the two asters are fully separated the line connecting them is seen to
be nearly antero-posterior (Plate 5, Fig. 42). The movement continues
until the axis of the complex becomes exactly antero-posterior, and a
spindle is formed in precisely the same position as the spindle for the pre-
ceding eleavage, This spindle is shown in the sagittal section, Figure
48. Its anterior end lies just ventrad of the small vesicle formed at
the previous cleavage. Division now takes place, and a second small
vesicle is given off to the point in the median plane lying just ventrad
of the vesicle formed at the fifth cleavage. Figure 49 shows the pro-
cess of formation of this vesicle, and Figure 50 shows the condition of
affairs after the division is finished. In later stages the two vesicles are
visible, lying beneath the cells of quadrants A and , in the place

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 37

where they were given off. (Compare Figure 50, Plate 6, with Figures
56, Plate 7, and 65, Plate 8.)

Other changes occur at the same time in the large ventral cell. After
the fifth cleavage the granular cloud in the cytoplasm gathered into the
region of the anterior margin of the free surface of the cell (Plate 5,
Fig. 38). As the spindle for the sixth cleavage is formed, the cloud
becomes concentrated over a small area, at a slight distance from the
anterior margin of the cell (Plate 6, Fig. 48). Then, as division takes
place, the cloud moves up to the anterior margin, at the same time
spreading out, and begins to pass beneath the cells of the quadrants 4
and B (Figs. 49 and 50). As the large nucleus moves away from the
wall of the cell where the vesicle was formed, the granular cloud moves
inward (dorsad) and spreads out between the nucleus of the large cell
and the two vesicles (Figs. 51 and 52). The granules at this time have
become very coarse and distinct.

Meantime, cleavage is taking place in the cell d9?, In this cell the
changes occurring in the asters are peculiar.

Immediately after the preceding cleavage, the cell, having passed
into the resting stage, has been pressed into an irregular wedge-shaped
form by the processes occurring in the surrounding cells (Plate 5, Fig.
37, surface view, and Fig. 38, section, from the same egg). ‘The cell
has become very narrow at the level at which the nucleus lies, so that,
apparently, there is not room for the asters to separate at right angles
to the foregoing spindle. The nucleus is pressed closely against the
ventral wall of the cell (Fig. 38), and the aster begins to extend
obliquely along the dorsal side of it, between the nucleus and the dor-
sal wall of the cell. When the aster has become completely divided
and the products are on opposite sides of the nucleus, their common
axis is already in the same direction as the axis of the spindle at the
previous division. The same result is obtained as in the rotation at the

fourth cleavage, though in a different manner. But the final positi
not yet reached.

As now situated, the asters lie in the long axis of the much elongated
cell (Plate 6, Fig. 46). As the active condition preparatory to di-
vision comes on, the cell withdraws its deeper parts (shown in Plate 5,
Fig. 38), and its dorso-ventral dimension increases. Accompanying this
change is a rotation of the nuclear complex, from a position with the
axis in the greatest dimension of the cell, to a position with axis in the
shortest dimension. This change is shown in progress in Plate 7,
Fig. 53. A later stage is shown in Figure 54; here the spindle is

38 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

completely formed. The dorso-ventral axis of the cell has greatly in-
creased, but is still distinetly less than the width of the cell at right
angles to the spindle. The cytoplasm has become grouped symmetri-
cally about the spindle, with the latter in its short axis. A longitudinal
section of the same egg is shown in Plate 6, Fig. 48; the completed
cleavage (d™ and d.) is shown in section in Figure 50, and from the
surface in Plate 7, Fig. 57. The cleavage is unequal, the ventral cell
being much the smaller.

Cleavage in the other cells of this generation takes place in a sequence
that is complicated by various factors, so that the account will be
clearer if the divisions of the cells are described in connection with their
relative positions in the egg, reserving a discussion of the order of
cleavage till the end. The divisions will be taken up according to the

layers of cells, beginning with the ventral layer.

First or Ventral Layer, consisting of the eight cells, de- d, and
OP Deb

The cleavage of 49^ has been described. The small vesicle d does
not divide farther. The other cells of this layer divide equatorially
into cells of equal size. Two of the spindles leading to this cleavage
are shown in Figure 56 (Plate 7), and the completed cleavage in Figure
61. The resulting fourteen cells are a1 - di, a7. d, a8- l.,
and d .
Second Layer, containing the cells aë? — 499 and a't- q94,

The cleavage of d, has been described ; it is equatorial and unequal.
"The remainder of the cells also divide by equatorial furrows, but the
products are equal in size. One of the spindles is shown in c, in Fig-
ure 47 (Plate G), and in Figure 55 (Plate 7) the nearly completed
cleavage; the nuclei in all but the products of c)? are still connected
by interzonal filaments. The same condition of the cell d*4 is shown
in Figure 57 (the products being d^" and dis). In all of these cells,
except d, the spindles lie at first in the shorter axes of the cells, as
indicated in Figure 47 (Plate 6); but as the karyokinetic processes
progress, the cells elongate in the direction of the spindles until the axes

in which the spindles lie are the longer.

The products of this division are a™'—d7™5, ar. d, a1 = d'1, and
ST,

Third Layer, containing the eight cells a95— de., and e - d°.,

In all these cells the division is meridional, not equatorinl, as in the
cells of the first and second layers. One of the spindles (059) is shown

in Figure 55 (Plate 7). The cleavage in d^? and des is shown in Fig-

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 39

ure 57. In this figure, d*5 has already divided into d"? and die, the
nuclei of which are still connected by interzonal filaments. In dere the
spindle is still present. ‘The completed division in the anterior part of
the egg is shown in Figure 61. The cleavage is equal in all the cells of
this layer except des and ; in these it is unequal, The unequal
cleavage of cs is shown in Figure 58, and the same figure shows the
unequal products (Gu and de) resulting from the division of d

3y the division of the third layer a band of sixteen cells is produced,
extending completely around the embryo. The cells composing the
band are a? =d", 7.40 _ Geo, a. 1 dH, and a - a",

Fourth Layer, containing the four cells a°" — d°.

In these cells the cleavage is meridional, as in the third layer, and in
every case equal. The spindles in 097 and c, are shown in Figure 47
(Plate 6).

The eight cells resulting from this cleavage are a^? e and
G4 qa

Fifth Layer, containing the four cells a*5 de, situated at the animal
pole of the egg.

The four small cells at the animal pole of the egg divide equatorially.
The spindle of des is shown in Figure 59 (Plate 7) ; and of de' in Fig-
ure 60. The cleavage products are very unequal; the dorsal cells so
formed are very minute, so that the distinetion between cell body and
nucleus cannot be observed, and the cells cannot be distinguished from
the polar cell lying in the same region. Figure 60 shows tho cleavage
at the animal pole completed except in the cell ds. A group of three
small vesicles, representing the dorsal cleavage products of the cells
50.8 — Ges, lie at the animal pole, surrounded by the four larger cells, —
one of which is the undivided cell des, while the others are the ventral
cleavage products of 095 s.

The eight cells thus produced are a= d:5 and de — dt

This, the sixth, cleavage may be tabulated as follows: —

Direction of Nature of
Layer. Cella, savage, Cleavage, Product,
First, or (4, 6, c,d) ®t and 92 Equatorial Equal (a-d) ™ and 7, (a-c)
Ventral (except d) (exe. di) 73 and 7*4 [and de 2J.
Second (a, b,c, d) 9*9 and &4 Equatorial Equal (a-d) "5 and 79;
(exc. d9:8) (a-d) 77 and 78,
Third (a,b,c,d) *5 and 99 Meridional Equal (a- d) "9 and 710;
(exe. 066, dë) (a — d) TH and 72.
Fourth ele re at Meridional Equal (a-d) 133 and 714,

Fifth (a, O, c, d) 98 Equatorial Unequal (a -d) 15 and 7:19,

40 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Since the minute cell de does not divide, we thus have produced
sixty-three cells instead of the typical number, sixty-four. Such a stage
does not, however, have an actual existence, sinoe some of the divisions
belonging to the seventh cleavage have täken place before all these
cleavages are finished.

Figure 61 (Plate 7) shows the anterior surface of the egg at the end
of the sixth cleavage, Figure 58 the posterior surface, Figure 60 the
dorsal pole, and Figure 63 (Plate 8) the ventral pole.

Sequence of the Sixth Cleavage. — The order in which the cells divide
is, as I have already stated, now complieated by several factors.

(1) The divisions of the first four quadrants of the egg (nt the third
cleavage) were not synchronous, but followed in the order D, O, D A
Other conditions remaining the same, — that is, with equal intervals
between the ensuing cleavages, — the same order would obtain in the
later stages.

(2) As discussed on page 35, the sequence becomes modified during
the fifth cleavage, so that the cells in any given quadrant divide in
nearly or quite the order of size of the blastomeres, beginning with the
largest. In the three quadrants A, B, and ©, the order is (the ven-
tral cell in each case being considered number one) 4, 2, 1, 3, while in
the quadrant D the order is 1, 2, 4, 3. This order would naturally
reappear in the sixth cleavage, other conditions remaining the same.

Both the above factors do influence the fifth cleavage, but with still
further complications. The first factor appears in the fact, that in any
given layer the general order of cleavage of the component cells is
D, ©, B, A.

The second factor is shown by considering the cleavage of a single
quadrant, as D. The order of cleavage for the large left hand cells of
this quadrant is as follows, naming the layers from ventral to dorsal:
1, 2, 1$, — nearly or quite the same as at the last cleavage.

jut a third factor appears in comparing the large left hand cells d.
and , of the D quadrant with their small right band sister cells d&4
and des (Plate 5, Fig. 37, Plate 6, Fig. 46, Plate 7, Figs. 54 and 57).
The large cells of each pair divide first, though the age of the two, being
sister colls, is exactly the same.

Similar relations may be shown for the other quadrants, Two facts
are worthy of particular notice. (I) The large cells a*" — c9" divide
first of all the cells in the quadrants A, B, and €, — and long before
the small cells aê — cs, which are of exactly the same age. (2) There
are some variations which cannot be brought into relation with any of

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 41

tho factors mentioned. "Thus, Figure 56 (Plate 7) shows that tho cell
56. wis dividing before the cells cèt and bee, though the cells are appar-
ently of the same size, and from the sequence of preceding cleavages
the cell ce. would be expected to divide first. However, such variations
may be correlated with differences in the size of the cells, since it is
impossible to calculate precisely the volume of cells which have such
irregular forms, and are subjected to varying conditions with the chan-
ging positions of the surrounding cells.

Certain general facts appear from the preceding discussion of the
sixth cleavage. (Compare the table of this cleavage, on page 39.)

(1) Every cell of any quadrant cleaves with its spindle in the same
direction as the corresponding cell of any other quadrant (except the
large interior cell de.).

The cleavage of a single typical quadrant up to this time is shown in
the annexed diagram (Diagram I.).

Dracram I.

Diagram of quadrant 4, B, or C in
the seventh generation. Only the second
exponent designating the cells appears in
the diagram, the first being in all cases 7.
Thus, the cell labelled 5 represents (4, b,
or c)^5, The arrows connect cells of

common origin, and show the direction
of the spindles at the preceding division.
R signifies right; L, left; D, dorsal;
V, ventral; according to the plan of
orientation explained at page 14.

(2) All the cells in any layer (series of cells occupying the same
relative position between the dorsal and ventral poles of the egg) cleave
with spindles in the same direction (except 0

(3) All the eleavages are equal except in the dorsal (fifth) layer, and
In and Urn

(4) There is a tendeney for the largest cells to cleave fastest. The
minute cell d*? does not cleave at all.

(5) The cell de cleaves in such a manner as to form a marked ex

ception to the method followed by the other cells. Its cleavage is very

42 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

unequal, while all the other cells in the ventral layer cleave equally.
Also, the spindle does not lie in the same direction as in the other cells
of this layer.

The exceptional nature of the division in de evidently demands
explanation. The regularity of the cleavage in the other cells of the
egg is such that some special condition must be correlated with the
markedly differing division of this cell. The cleavage of d** differs
from that of the other cells of the ventral layer in the following points.
(1) Its cleavage is very unequal, while the cleavage of the other cells
is equal. (2) The spindle for the sixth cleavage in de. lies in almost
exactly the same position as did the spindle for the preceding cleavage,
whereas in the other cells of the ventral layer the spindle for the sixth
cleavage is at right angles to the position of the spindle for the fifth
cleavage.

As to the first point, — that of unequal division, — no special cor-
relating factor for this particular case seems necessary, since in many
cases in this and preceding cleavages the cells varied as to the equality
of the division, though alike in other respects. Thus, in the fifth
cleavage, the three ventral cells of the D quadrant cleft unequally,
though in all the other quadrants the division was equal; and in this

*9 and 499 divide unequally, though in all the

sixth cleavage the cells c
other cells of the same layer the division is equal.

But the second point is totally exceptional in the cleavage up to this
time. "The axial relations of the cells appear to be so distinct and con-
stant, and there is such uniformity in the positions of spindles at a
given cleavage among cells of similar origin and relative position, that
one must look for some other marked difference in the cell de that
might oceasion this change of axis.

In what respects does the cell d?! differ from the other cells of the
egg? (1) In its greater size; (2) in its position.

(1) The greater size evidently has nothing to do with the different
direction of cleavage, since the same disparity in size was present in
earlier cleavages, yet this cell divided in the same direction as did the
other ventral cells.

(2) The change of position is such as to bring about a fundamental
change in the relations of the cell to the egg as a whole and to the
other cells. Previously the cell de formed the posterior cell of the
ventral layer. At the time of the sixth cleavage the cell has moved
toward the interior of the egg and its posterior surface is covered as far

ventrad as its anterior surface. The cell is now central, and surrounded

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 43

on all sides by other cells. It thus occupies a position in the egg which
is fundamentally different from that occupied by any other cell. Cor-
related with this fundamentally different position, the cell acquires a
fundamentally different method of division.

It is impossible to say whether any particular feature of the different
position of the cell is the essential one in bringing about this altered
method of cleavage. As the cell moves inward, it very probably
accomplishes a partial rotation (see below); if the axes of the cell are
definite, and determined within the cell alone, then this rotation would
cause a change in the position of the axes of the cell dé in comparison
with the axes of the other cells, and a different direction of the spindle
would result, But any such explanation is hypothetical.

During the later stages of the sixth cleavage, the process of gastrula-
tion has made much progress. At the end of the fifth cleavage the
large ventral cell d^t had already moved some distance toward the in-
terior of the egg (Plate 5, Fig. 38). As the cells des and des now
withdraw successively their deeper parts and increase their surface
extension during division, they push ventrad, displaeing the posterior
part of the ventral cell (now d") (Compare Plate 6, Figs. 48, 50,
and 51.) The large cell therefore pushes dorsad into the interior of
the egg, occupying the space made vacant by the ventral extension
of the other cells. Soon after, the anterior cells, belonging to quad-
rants A, D, and C also enter upon the karyokinetie process, and in so
doing likewise push ventrad (Plate 8, Fig. 64) at the same time vacat-
ing of course a portion of the space before oecupied by them near the
animal pole, The cell d therefore continues to move dorsad, so that
at the end of this cleavage it is almost completely enclosed (Fig. 65).

During this inward movement of the cell du, the cloud of granules
previously described changes its position still further. We had traced
it, after tho sixth cleavage, until it occupied a position between the nu-
cleus of d" and the two vesicles formed at the fifth and sixth cleavages
(Plate 6, Fig. 51). As gastrulation continues, the cloud of granules
migrates still further dorsad, and later even crosses the dorso-ventral
axis, so as to lie posterior to it, surrounding the dorsal aster at the next
division of d" (Plate 8, Fig 64).

This movement of the cloud of granules possibly gives the key to the
change of axis of division in the cell diu. When the posterior cells
extend ventrad during division, as previously described (Plate 6, Figs.
48 and 51), they push against the posterior side of the cell ., thus

displacing the cell inward. But such an impulse from one side only

44 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

would naturally, if the cell be a body not entirely fluid, give it a rotary
motion. As a result of the partial rotation, the cloud of granules is
now found underneath the anterior cells (Plate 6, Fig. 50). When the
anterior cells extend ventrad, the form of di is so changed by its
change of position that the cells do not push against it, but glide over
it, at the same time vacating a part of the dorsal region of the egg
(Plate 8, Fig. 64). Therefore the cell d'1, as it moves inward, con-
tinues its rotation in the same direction as before, and the cloud of
granules is brought to the dorsal end of the cell, and even farther,

Such a rotation would explain clearly the peculiar position of the
spindle in de at the sixth cleavage. To agree with the other cells of
the ventral layer, the spindle would have to take a dorso-ventral posi-
tion. (Compare Fig. 55, Plate 7.) But a rotation from posterior to
anterior just before this cleavage would bring the spindle into an antero-
posterior position, such as actually occurs.

However, the explanation is not very satisfactory, for several reasons.
(1) We have previously seen that the granules forming the granular
cloud do move within the cytoplasm of the cell, so that this change of
position of the granular mass may be due to simple migration through
the protoplasm of the cell. (2) A study of the movements of the
asters after the fifth and sixth cleavages does not give evidence of any
such rotation. (See the account of the movements of the asters before
the sixth cleavage, page 36, and before the seventh cleavage, page 54.)
(3) Such a rotation explains the position of the spindle at the sixth
cleavage only, — while all the later cleavages of the inner cells are like-
wise out of relation to the divisions of the outer cells. (4) The ex-
planation assumes that the position of the axes of cleavage is definite,
and determined within the cell itself, so that, if the cell rotates, the axis
rotates with it, — which is not proved,

The only conclusion in which we are entirely justified is therefore
merely this, that as the relation of the cell dé to the other cells and to
the egg as a whole becomes fundamentally changed, the method of di-
vision likewise becomes fundamentally changed.

At the close of the processes thus far described, the egg has evidently
passed into the “gastrula stage” proper (Plate 8, Fig. 64). The blasto-
pore is still large and lies at the maeromere end of the egg. It is sur-
rounded at first by eight cells, two belonging to each of the four original
quadrants of the egg (Plate 7, Fig. 56). Later, as the sixth cleav-
age is entirely finished, one of the cells (el.) becomes displaced and is
shut out from the margin of the blastopore (Plate 8, Fig. 63), which is

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 4b

itself diminished in size. The interior of the gastrula is occupied by
the large cell d" and the two minute vesicles d*? and d^? (Fig. 64).
The interior cells are surrounded by a single layer of outer cells, except
at the animal pole of the egg, where the small dorsal cells do not reach
to the cells within the gastrula, but lie on the surface, making here a
two-layered region. "This region remains two-layered as long as it is
possible to trace the history of the animal pole of the egg.

The three inner cells di, d., and d"?, with the products of the
former, may henceforth be called the entoderm, the outer layer the
ectoderm.

Seventh and Later. Cleavages.

I have followed the cleavage through another generation, and, for
parts of the egg, much farther. It becomes impracticable, however, to
describe the eleavage according to the layers or series in which it takes
place, as has been done up to this stage, owing to the complicated suc-
cession of the divisions in the different cells, and to the great changes in
position taking place while the cleavages are in progress. I shall there-
fore now describe the processes in the general order in which they
occur, and in so doing I shall consider separately (1) the ectoderm, and
(2) the entoderm.

Tue ECTODERM.

In discussing the changes taking place in the ectoderm, it will be well
to distinguish for convenience of description two regions : (1) the (left)
posterior part of the ectoderm, derived from the quadrant D; (2) the
anterior and right lateral portions of the ectoderm, derived from the quad-
rants A, B, and C. While the phenomena occurring in all of these regions
are reducible to the same ‚general scheme, so far as the method of
cleavage is concerned, the irregularity in the size of the blastomeres
forming quadrant D, their earlier cleavage as compared with the other
quadrants, and the fact that some of the cells have passed inward to
form the entoderm, give this region a peculiar and somewhat irregular
charaeter, which makes it convenient to discuss it separately.

(1) The Quadrant D. — The entoderm cells belong genetically to this
quadrant, but they will be considered later.

In order to understand the conditions in quadrant D, and to see their
relations with the arrangement in the other quadrants, it will be well to
emphasize certain features of the last two cleavages.

At the fifth cleavage, as previously deseribed, the three ventral cells
of quadrant D divided by meridional planes into unequal portions, the
left derivative being in every case larger (Plato 4, Fig. 33, and Plate 5,

46 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Fig. 37). The single dorsal cell divided equatorially into a small dorsal
and a large ventral part. There is appended a diagram of the quad-
rant D after this cleavage (Diagram IL). If the septa in the three
ventral cells were moved to the middle of the cells, the diagram would
represent the condition in any one of the other three quadrants.

At the next division (sixth) the two ventral cells de. and die have
passed inward, becoming the entoderm, so that we may omit them from
the present discussion. Of the other cells, the ventral pair (d? and de.)

R
2 4
"
7 I
$
AY. V
Diacram II. Dracram III.

Diagram of quadrant D in the sixth Quadrant D in the seventh genera-
generation. Only the second exponent tion. Only the second exponent is
designating the cells appears in the expressed, the first being in all
diagram, the first being in all cases 6. cases T.

The arrows connect cells of common origin, and show the direction of the spindles
at the preceding division. R signifies right; L, left; D, dorsal; V, ventral.

divide equatorially, d*-? unequally, d** equally (see Plate 7, Figs. 57
and 58). The two next layers divide meridionally (Fig. 57), the cell
die unequally, the others equally. The dorsal cell divides equatorially
and unequally. Diagram III. shows the ectodermal part of this
quadrant at the end of the sixth cleavage. The actual condition in the
egg at this period is shown in Figure 58, and at a slightly earlier stage
in Figure 57.

Comparison of Diagram III, with the type diagram for the other

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 4T

three quadrants given on page 41, shows that the directions of the cell
walls are tho same in both, the inequality in the size of the cell in
the quadrant D being the only difference.

In the seventh cleavage, the spindle appears first in the cell d7^, as
shown in Figure 58, and the cell is divided meridionally into two equal
cells, d? and dee. The finished division is shown in Plate 8, Figs.
66, 67, and 68. The plane separating these two cells is the median
dorso-ventral plane of the embryo, as will be shown later.

Shortly after this division is completed, spindles appear in CRIS AE
dun, %s, and diu, as shown in Figure 66. In da’ the spindle is
dorso-ventral, hence lying in the shorter axis of the cell; the cell
extends and divides into two equal dorsal and ventral parts, d*? and de
(Figs. 67 and 68). The greater surface extension of d*? in Figure 67

is due to its being spread out in a thin layer over the surface of the

entoderm cell.

In dis and die (Fig. 66) the spindles are also dorso-ventral in
position, and the cells divide equatorially into equal parts, QU qe
ad, and ds. (Figs. 67 and 68).

In d^? and d: the spindles lie at right angles to those just described,
and the cells divide meridionally and equally, forming des, de-, de.,
and d*-?,

Figure 67 is a view of this region after these cleavages are finished.
As shown in this figure, a certain amount of shifting has taken place
during cleavage, by which the cell d™ has been excluded from its share
in the boundary of the blastopore. As the cells divide, they withdraw
their interior parts and extend in the direction of the spindle, as has
been minutely described for other cleavages; in this way the dorso-
ventral extent of quadrant D has been greatly increased. Asa result
the blastopore has been nearly closed (Fig. 63), and at the opposite
end the animal pole has been pushed beyond the micromere end of the
egg to its anterior side (Fig. 65).

Thus far the six larger left hand cells have divided, leaving six smaller
cells (at the right and at the dorsal pole) undivided (Fig. 67).

Next, as shown in this figure, the cells d" and dis form „spindles and
divide. Each cleaves in the same manner as its larger companion cell
has done, d"? equatorially, dis meridionally. The products, dee, dee,
d8, and de, are shown in Figure 68 (compare Diagram IV.).

Next de cleaves equatorially, like its companion cells d"? and dine.
In d^? the division is unequal, the ventral product, d, being much
the smaller (Fig. 68).

48 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

The cleavage of d™ occurs much later, and is likewise equatorial.
I have not thought it necessary to introduce a special figure to show the
cleavage of this minute cell.

In d™ the cleavage also takes place late; it is meridional and equal.
The resulting cells, de and de., are shown in Figure 72 (Plate 9).
The minute dorsal cell de does not divide.

D
„718 D
par 4
"
lan 30 | 29
n *
26 25 | 28 27
26 23 28 +127
22./| 24 18 a
t i — 22 24 18 20 \
1 1
* + b
21 |23 17 Tan 23 17 [E] |
R L L
wi esie it al 4 as 18 WER
jisi dá 2 | to
| m $ 1 1
ie Bt v9 E
y V
Dracram IV. DIAGRAM V.

Quadrant D in the eighth genera- Diagram of quadrant A, B, or C, in
tion, — except the dorsal cell, d7:16, the eighth generation. Only the second
which does not divide farther. In exponent of the cells appears in the dia-
the other cells only the second ex- gram, the first being 8. The small dor-
ponent is expressed, the first being sal cell, (a, b, or c)7*10, does not divide,
in all cases 8. The arrows connect remaining thus in the seventh genera-
cells of common origin, and show the tion. The arrows connect cells of com-
direction of the spindles at the pre- mon origin, and show the direction of
ceding cleavage. the spindle at the preceding division.

R signifies right; L, left; D, dorsal; R signifies right; L, left; D, dorsal;
V, ventral. V, ventral.

There are thus. 23 cells in the ectodermal part of the quadrant D at
the end of the seventh cleavage. A diagram of this stage is annexed
(Diagram IV.). Nearly this stage is represented in Figure 68.

(2) The Quadrants A, B, and C. — Owing to the regularity in the
size of the cells in these quadrants, and the fact that they are purely
ectodermal, the conditions observed are fairly simple as compared with

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 49

those in quadrant D. Figure 61 (Plate 7) shows the anterior surface of
the egg at the end of the sixth cleavage, and a diagram of a single
quadrant at this stage was given on page 41. Comparing either the
figure or the diagram with the scheme (Diagram III.) given on page 46
for the quadrant D at the same stage, the arrangement is seen to be the
same except in two respects, (1) The cells belonging to the same lat-
eral series are equal in A, B, and C, unequal in D. (2) Four ventral
ectodermal cells additional are present in each of the quadrants A, B,
and C; these are represented in D by the cells which have passed
inward to form the entoderm.

As shown by the spindles in Figure 61, the first cells to divide are
a3 718 and ato, Phe cleavage is meridional and equal. The
resulting cells are a% — 8%, 9826. 68.26, a7 — (897. and d. c. They
are shown in Figures 72 and 75 (Plate 9). (Compare Diagram V.)

Next follows the division of the six cells, a^9— 6^9 and a^5— c^, which
together form part of a transverse girdle, surrounding the egg. The

S4. ale, ar

cleavage is meridional and equal. The resulting 12 cells, «
08.12, 815 — C8. 46, and 9919 095 forming as before a transverse girdle, are
shown in Plate 8, Figs. 69 and 70, and Plate 9, Fig. 71.

Next ensues the cleavage of the transverse row containing the six
cells, ac"? and o*7—c7. These cells, as shown in Figure 61 (Plate 7),
are much flattened dorso-ventrally, and are of exactly the same form as
the cells last discussed, which lie immediately dorsad of them. Moreover,
each cell in this row corresponds in origin to a cell of the row last
described, the two rows having been derived from the equatorial division
of a previously existing transverse row, as shown in Figure 55. Ik
mechanieal conditions are decisive in determining the,direction of the
cleavage, these two rows should cleave in the same manner, i. e. both
meridionally. Nevertheless, as shown in Figure 69 (Plate 8), while the
dorsal of the two rows divides meridionally, the cells of this ventral row
all cleave equatorially. The axis of the cell in which the spindle lies is
about half as long as the axis which is at right angles to it. The cells
elongate in the direetion of the spindles, and a very unequal division
ensues, The ventral products are minute, while the dorsal ones are
nearly equal in size to the mother cells. Figure 75 (Plate 9) shows
the anterior surface of the egg after this cleavage. The twelve cells

18.10 8.10 8.13 8.13 8.14 8.14
2 , .

produced are a=, 5.10 — 80. a or Mand us te

The division of the band of twelve small cells composed of a"? — e,
a — .o, 911 gr and a.t? = 012, and shown in Figure 61 (Plate 7),
follows somewhat later. "The cleavage is equatorial and the spindles lie

VOL. XXX. — NO. 1. 4

50 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

in the short axes of the cells, as shown in Figure 75 (Plate 9). Just
before division the cells elongate until the axis in which the spindle lies
is longest, as shown in the cell 0^? of the figure last mentioned.

The cells e e, near the animal pole, cleave meridionally. The
spindle in a’, and the cells 6° and 05.2, 09? and 6*9 are shown in
Figure 72 (Plate 9). The minute cells d — 6739, like die, do not
cleave further. The arrangement at the animal pole is now very
irregular, and owing to this fact and the minuteness of the cells pro-
duced at the last cleavages it is very diffioult to be certain of the exact
origin of any given cell, though the origin of the growp cannot of
course be doubtful.

We have now accounted for all the cells of these quadrants except

2 7.2
,

the four ventral cells of each quadrant, a! e, at? — 672, a" — , and
Aus u. These correspond in origin to the entoderm cells of quadrant
D, and they do not cleave further until they are partly or entirely
enclosed within the embryo, as will be shown later.

A diagram of that part of any one of the quadrants A, B, and C
corresponding to the ectodermal part of quadrant D, and showing the
conditions at the end of the seventh cleavage, is given on page 48 (Dia-
gram V.). A comparison of this diagram with that for the correspond-
ing stage of the quadrant D (Diagram IV. on the same page) shows
that the direction of the cleavage planes, and hence of the spindles, is the
same throughout in all cells of corresponding position, though there are
many differences as to the equality or inequality of the cleavage products.

The general facts which may be deduced from the foregoing study of
the seventh cleavage in tho ectodermal parts of the egg are similar to
those drawn from a study of the sixth cleavage, page 41.

(1) Every cell of any quadrant divides with its spindle in the same
direction as the corresponding cell of any other quadrant.

(2) All the cells in any layer cleave with spindles in the same direc-
tion (in spite of great differences in the form of the cells.)

(3) No general law can be deduced as to the equality or inequality of
the divisions.

(4) There is a tendency for the largest cells to cleave fastest. Certain
very small cells (at the dorsal pole) do not cleave at all.

Other Changes during the Seventh Cleavage. — In almost every cleav-
age which has taken place, whenever the division was equatorial, — the
spindles taking a dorso-ventral position, — it will have been noticed that
the axis in which the spindle was formed was the short axis of the cell.
On the other hand, the cells in which meridional cleavages have taken

JENNINGS: DEVELOPMENT OF ASPLANCHNA IIERRIOKII. 51

place have commonly been already somewhat elongated in the direction
of the spindle. Therefore, at the occurrence of division, the short cells,
cleaving equatorially, have changed form greatly, becoming more elon-
gated dorso-ventrally, while the cells which eleave meridionally have
already been of sufficient, length to permit of the extension of the spin-
dle without much change of form. As a result, there has been a great
extension of the ectoderm dorso-ventrally. This produces first a com-
plete elosuro of the blastopore. (Compare Figure 65, Plate 8, with
Figure 76, Plate 9.) A second result, due to the larger size of the
cells of the quadrant D, and perhaps partly also to the fact that they
oleave first, is tho further displacement of the animal pole of the egg
from the micromere end toward the anterior side (Plate 8, Fig. 65,
Plate 9, Fig. 76).

The closure of the blastopore is not sufficient to provide for the dorso-
ventral extension brought about during the eleavage, so that, as a third
consequence, the cells at the ventral pole of the egg, where the blasto-
poro was previously situated, are pushed over or under one another,
the ectoderm tending to become two-layered in this region. A compari-
son of Figure 63 (Plate 8) with Figure 73 (Plate 9) shows the result-
ing conditions, In the latter figure the entoderm (dt in Fig. 63) is

entirely enclosed, and the ventral cells a't- ce", a? c., at — 013, and

(A- have become crushed together, and several of them, as a^? and
078, are almost hidden by surrounding cells. The beginning of the two-
layer condition is shown in frontal section in Figure 80 (Plate 10),
from the same egg as that shown in Figure 73 (Plate 9).

Meanwhile other cleavages are taking place in quadrant D, leading to
still further modifications of the structure of the egg. It is therefore
necessary to return to a consideration of this quadrant.

Quadrant D. — By reference to the diagram of the cells of this quad-
rant at the end of the seventh cleavage (Diagram IV., page 48), it will
be secn that there are now present, exclusive of the entoderm, twenty-
three cells, arranged somewhat irregularly. Approximately the same
stave, as it actually appears, is shown in Fignre 68 (Plate 8).

In this egg spindles. have appeared in the large cells d* and d*™, and
later they each become divided into two equal cells. It is worthy of
notice that the two spindles are not parallel, but make a slight angle
with each other. The two cells lie on opposite sides of the median
dorso-ventral plane, so that the angle between the spindles indicates the
beginning of a tendency to bilateral cleavage. The inclined position of
the other spindles (in c? and c^) is indicative of tho same fact. The

52 BULLETIN: MUSEUM OF COMPARATIVE ZOÜLOGY.

tendency toward bilaterality is apparently due to the crowding of the
cells from all directions toward the blastoporic region, which lies at
the ventral border of dè? (Fig. 68). The cells resulting from the
division of de and de! are d”, d. 22, d. 28, and dt; they are shown
in Figure 74 (Plate 9).

Next the four cells immediately dorsad of these, de. h, d, de, and
d*-*, develop spindles in their short axes (Plate 8, Fig. 68), which as a
result of extension become the long axes (Fig. 74, Plate 9), — the
cleavage being in each case equatorial and equal.

Now the four large cells de., d, d e and dee, shown in the dorsal
part of Figure 68 (Plate 8), cleave equatorially also.

As a result of these many equatorial cleavages tne quadrant D
becomes greatly increased in dorso-ventral extent. The animal pole is
forced farther upon the anterior side of the egg, toward the blastopore,
so that the cells of the quadrant D come to occupy in the region of a
sagittal section much more than half the circumference of the egg
(Plate 9, Fig. 76).

The two ventral cells dee and dee meanwhile divide meridionally,
completing the separation of quadrant D into two portions lying on each
side of the median line (Fig. 74). However, the egg is not yet com-
pletely separated by cleavage planes into right and left halves, for the
entodermic cell d? occupies the median plane at even a later stage than
this (Plate 10, Fig. 83).

This is the latest stage to which I have traced the cleavage in the
ectodermal part of quadrant D. Diagram VI. shows the condition at
this time.

The dorso-ventral extension of the ectoderm, and consequent erowd-
ing together of the cells in the region of the blastopore, are still furthor
increased by the eighth cleavage of the large cells 8-25 _ 68.25, (48,28... 58,28,
d. 27 eè and 48.28 ce, belonging to the other quadrants, which like-
wise divide equatorially and equally (Plate 9, Fig. 75).

The blastoporic region has now become distinctly two-layered, as
shown in Figure 77. The cells of quadrants A, B, and C are turned
in and pushed dorsad, in the same manner as bappened in early stages
to the large ventral cell of quadrant D. The anterior lip of the blasto-
pore thus becomes two layers thick, while the posterior lip is formed

wg a €

of a single layer of cells from the quadrant D, resting against tho ento-
derm cells. Between these ventral cells of quadrant JD and the infold-
ing cells of the other quadrants, a slight notch appears, marking the
position of the blastopore. (At an earlier stage the blastopore was

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIL 53

entirely closed, as shown in Figure 73.) The blastopore notch lies, not
directly at the macromere end of the egg, but at some distance on the
posterior side of it.

As the cells of the anterior lip of the blastopore are turned inward,
some of them begin to divide. Spindles in these cells are shown in
Plate 9, Figs. 75 and 77. The form and position of the cells have
changed so much at this time that it is impossible to determine with
certainty whether the cleavage should be considered equatorial or merid-
ional. In the cases figured the spindles are nearly or quite transverse,
so that in some of the cells the division is meridional.

Dracram VI.

Diagram of quadrant D at a time when
mest of the cells have passed into the
ninth generation. All cells in the ninth
generation are bounded by continuous
lines, and are designated by the second
exponent belonging to the cell, the first
exponent being in each case 9. In the
other cells both exponents are given.
The arrows connect cells of common
origin, and show the direction of the
spindle at the last division.

R signifies right; L, left; D, dorsal;
V, ventral,

The cleavage of the ectoderm has now been traced to the eighth gen-
eration in all parts of the egg, and in the greater part of the quadrant
D to the ninth generation.

With this ends the account of the cleavage of the ectodermal cells.
The small size and the great displacements of the blastomeres, espe-
cially in the regions of the blastopore and the animal pole, render it
impossible to determine with certainty their identity in later stages,
and the real direction of eleavage is masked by erowding and deforma-
tion of the cells. It would perhaps be of little interest in connection
with the laws of cleavage to carry the study further, as it is scarcely to
be presumed that the later divisions would exhibit any phenomena dif-

54 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

fering fundamentally in kind from those already shown in the earlier
Stages.
We must now turn to a consideration of

Tue ENTODERNM.

The cells which I have called the entoderm — following in this Zelinka
(91) and Tessin (86)

the quadrant D, which passes within the egg in the manner already de-

are those derived from the single large coll of

scribed. I have already given an account of the first cleavages of this
cell after it has become partially covered by the other blastomeres. As
will be recalled, at the fifth and sixth cleavages the spindles occupied
twice in succession the same position, one end lying in the anterior
median line, between the ventral cells of quadrants 4 and B (Plate 5,
Fig. 35, and Plate 6, Fig. 48). Here was given off at each of these
cleavages a minute vesicle, the entire process being comparable in ex-
ternal features to the successive formation of two polar cells at a given
spot on the surface of the egg. The two vesicles thus formed maintain
their position for some time (Plate 7, Fig. 55), but as tho surrounding
cells become invaginated, I have found it impossible to follow their
later fate.

We will follow the cleavage of the large cell 4% (Plate 6, Fig. 50),
which forms the greater bulk of the entoderm.

After the sixth cleavage (Fig. 49), the asters in d™ at once separate
nearly in the dorso-ventral axis of tho egg, as shown in Figure 50. The
line joining them is at first a little oblique, the ventral aster being :
little to the right. This obliquity soon corrects itself, and the asters
come to lie in the sagittal plane. As the spindle is formed, its dorsal
end moves to the posterior side, so that the spindle is no longer in the
dorso-ventral axis of the egg. This stage is shown in Figure 64 (Plate
8); as may be observed in this figure, the spindle is neither in the
longer axis of the cell nor at right angles to it, but oblique. The cloud
of granules, which soon after the last division oceupied a region on the
anterior side of the cell, underneath the two vesicles d? and , now
surrounds the aster at the dorso-posterior end of the spindle.

The cleavage is unequal, separating (Fig. 65) a smaller dorso-posterior
cell, d*?, from a larger anterior one, d®t, The cloud of granules
remains in the smaller, dorsal cell, forming a band about its periphery,
so as to leave a free space surrounding the nucleus. The position of
the animal pole of the egg with reference to this cell should be carefully
nofed, as the relation remains constant, at least for a. time, during the

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 55

considerable shiftings which take place. As seen in Figure 65, the
animal pole lies at the anterior margin of the cell ds.

Before the eighth cleavage takes place, the blastopore has become
closed and its anterior margin has begun to pass into the two-layer
condition, as previously described. The larger cell, ., divides before
its mate, by a spindle at right angles to the previous spindle. The
cleavage is equal, forming the two large right and left cells % and
d*? shown in Plate 10, Fig. 80. The plane separating these cells
coincides with that separating the quadrants A and Z on the anterior
side of the egg, and also with that separating the two cells deu and
de on the posterior side (Plate 8, Fig. 68). As this plane also
passes through the animal pole and the blastopore, it divides the egg
into symmetrical halves, and is the median dorso-ventral or sagittal
plane of the embryo.

Later the small cell, d*:?, develops a spindle in the same direction as
the spindle of the seventh cleavage, in the shorter axis of the cell, and
divides into two very unequal cells. The anterior or ventral cell, d,
is a minute vesicle, whereas the dorsal or posterior blastomere, d?*, is
scarcely smaller than the mother cell. The process of budding off this
small cell is shown in Plate 10, Figure 80. "The vesicle lies between
the two large cells d™! and d., and, like the minute cells d*? and d"?
g lost among the many

5

can be traced but a short distance, soon becomin
cells by which it is surrounded.

In the ninth cleavage, spindles are formed in the cells d? and d?? in
the position foreshadowed by the asters in Figure 80, — that is, antero-
posterior, and at right angles to the preceding spindles, — and the cells
divide equally, forming the four cells d., d'™?, 0.8. and de.. Figures
76 and 81 show the entoderm at the close of these divisions — the
nuclei of the cells in question being still connected in -pairs by inter-
zonal filaments. The blastopore is now present as a distinct notch ; its
anterior or dorsal lip has become two-layered, owing to the folding
inward of the cells of the anterior quadrants. The animal pole (pol.
anm., Vig. 76) has moved a considerable distance on to the previously
anterior surface, lying still at the anterior margin of the entoderm cell
d**, As the side view (Fig. 76) shows, a frontal plane carried through
the long axis of the egg at this stage would cut the nuclei of all the
entoderm cells, as realized in the frontal section, Figure 81 (Plate 10).

As soon as the cells, after the cleavage process is entirely finished,
have lost their strong tendeney to maintain a form as nearly spherical as

possible, a sudden and considerable change of relative position takes

56 BULLETIN: MUSEUM OF COMPARATIVE ZOÜLOuY.
place. The four entodermal cells, d. -d. become much flattened
dorso-ventrally, and the invagination of cells at the anterior lip of the
blastopore increases in extent. These cells press most strongly upon the
anterior surface of the ventral entodermal cells, forcing them toward
the posterior side. An early stage in the process is shown in Plate 9,
Figure 77. Two of the cells of the invaginating ectoderm have flattened
themselves against the entoderm cells in such a way as to form a direct
continuation of the longitudinal series of interior cells, This longitudinal
series is however tending to become curved by the displacement of the
cell 2? in a dorsal and anterior direction, This cell, together with the
animal pole, has moved a very slight distance toward the maeromere
end of the egg. The animal pole now lies (Fig. 78) directly above the
plane separating the cells 4%. and d'?? from die and d., instead of
above the posterior margin of the cells d. and d., as previously. A
frontal plane carried through the long axis of the egg would not now cut
the cell d at all. Figure 82 shows a view from the animal pole, the
outer layer of ectoderm cells being supposed to be removed from the
dorsal half of the egg, while the entoderm cells remain in position.

This process of rotation of the entodermic contents, as one might call
the phenomenon, continues still farther. Figure 79 shows a stage in
which the process is much more advanced. The ectodermal plug at the
anterior lip of the blastopore has become very much thickened, and pro-
jects farther posteriad; the four large entoderm cells d. - e., of which
of course but one pair is shown in the side view, are now so displaced
that the plane separating the pair, which previously lay in the short axis
of the egg (Figs. 76 and 78), now lies in the long axis (Fig. 79). The
line connecting the centres of a given lateral pair is now at right angles
to the line previously connecting them.

At this time the five large cells constituting the entoderm — the
minute cell ds not being traceable farther — begin to undergo karyo-
kinetic changes preparatory to division. The spindles in various stages
are seen in three of the cells in the side view, Figure 79; the same
stage, slightly earlier, is shown from the animal pole in Figure 83
(Plate 10), in which the covering ectoderm is supposed to have been
removed from the dorsal side of the egg. As a comparison of Figures 79
and 83 shows, the spiudles do not lie in parallel planes, so that no single
view can give a complete representation of their positions. Nevertheless,
Figure 83 shows that the arrangement is distinctly bilateral. The cell
des lies in the middle line, with its spindle in the sagittal plane of the

egg'; the spindles in the other cells all radiate outward from the region

JENNINGS: DEVELOPMENT OF ASPLANCHNA IERRIOKII. 57

occupied by d?^, the sagittal plane passing through this cell forming a
plane of symmetry for all. The median plane thus indicated coincides
with that already defined by the line separating the quadrants A and B
anteriorly, and the boundary between the cells d and dds. H posteriorly ;
it passes through the animal pole and the blastopore.

This movement of the entodermal blastomeres is of course simply a
continuation of the rotation inaugurated at the passage from the four- to
the eight-cell stage. The ectodermie cells continually withdraw their
deeper parts and increase their surface area at division; in this way cells
are continually forced in at the blastopore. These press upon the anterior
and ventral aspects of the entoderm cells, forcing them backward, as
already described. The pressure is greatest in the median region, so
that the anterior or ventral ends of the cells on the two sides of the
median plane are forced apart, the axes of the cells become oblique to
the long axis of the embryo, and the oblique position of spindles shown
in Figure 83 results.

Beyond this point it is impossible to trace the development cell by
cell. In Asplanchna there is a period, intervening between the stage to
which it is possible to trace the cleavage step by step (about 120 cells)
and the stage of recognizable differentiation of organs, during which the
cells divide and become very minute, The cells probably reach the
number of from 250 to 500, and the process of extension of cells and
consequent, “rotation ” and invagination of parts of the embryo continues.
A sagittal section of the embryo at about the time of the beginning of
differentiation of organs is shown in Figure 84. It thus becomes
impossible to trace the fato of individual blastomeres, or even, except in
& most general way, the fate of the different regions of the embryo during
the later portion of the cleaving. From the small size of the adult roti-
for, I had hoped that it would prove to be a favorable object for an exact
study of the cytogenetic history of organs ; in Asplanchna this turns out
not to be the ease. But, on the other hand, for a study of the factors in
the early developmental processes it has shown itself well fitted.

My study of the processes in the early development of Asplanchna there-
foro closes with the stages shown in Figures 79 and 83 (Plates 9 and 10).

For critical comparison of my observations with those of Zelinka and
other workers on the development of the Rotifera, the reader is referred
to Part Second.

I shall now proceed to a discussion of the bearing of the foregoing
observations upon the problems already proposed, as well as upon other

related subjects.

58 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

III. Discussion of the Bearing of the Observations on the
Problems.

In the following discussion I shall adopt in general the order pursued
in my “Statement of Problems," taking up successively the various
theories in regard to cleavage and gastrulation, and pointing out what
bearing the observations above detailed have upon these theories, This
will be followed by a résumé of the general conclusions which may be
drawn from the work.

We will therefore first take up a discussion of the cleavage, and of the
theories bearing upon it.

1. CLEAVAGE,
A. The Direction of Oleavage.

(1) Berthold’s theory of surfaces of least’ area. (See page 4.) — The
two- and four-cell stages in. Asplanchna agree well with the conditions
demanded by the law of least surfaces. The peculiar arrangement of
blastomeres in the four-cell stage, agreeing as it does with the four-cell
stage of animals of the most varions systematic positions, and with the
four-cell stage of many plants, seems probably due to some very general
law. In all these cases only three cells meet along one common line.
As this is the arrangement demanded by the principle of least surfaces,
the conclusion seems perhaps justifiable that this principle of least sur-
faces is that common law.

The eight-cell stage also fulfils the requirements of the principle of
least surfaces. But from this stage onward, many of the conditions
found are irreconcilable with the view that this principle is a determining
factor. Six of the cells in the eight-cell stage divide in a manner that
squarely defies the principle of least surfaces. Nor does the arrange-
ment of cells in the resting periods agree better with the principle. As
pointed out on page 30, the flat, almost disk-shaped form taken by the
cells of quadrant D during the ten-cell stage (Figs. 23 and 25, Plate 3)

stage (Figs. 31 and 32, Plate 4) is widely

at varl-

and the sixteen- cell

ance with the demands of the principle of least surfaces. The form
of the cells in quadrants A, B, and O during their resting period in the
sixteen-cell stage (Figs. 30 and 34) is equally impossible of explauation
on the least surfaces ‘theory.

Many other cases could be adduced in
which this principle is contradicted, but a fuller discussion or these cases

will be given under the next theory (Hertwig’s law). In general, any
case which is not in agreement with Hertwig’s law is likewise inexplicable

JENNINGS: DEVELOPMENT OF ASPLANCHNA IIERRICKII. 59

by Berthold’s principle, so that all the cases cited later as opposed to
Hertwig’s law can be utilized equally well against the prineiple of least
surfaces as a determining force.

It should be noted that Derthold did not in any sense maintain that
this principle is the decisive factor in cell division, or the arrangement of
cells in tissues. He recognized that the conditions in a living cellular
body are widely different from those in a simple vesiculated fluid, and
that the conditions actually found in plant tissues are often inexplicable
by the principle of least surfaces, — in many cases, indeed, directly
opposed to it. “ Aber nothwendig ist das in der Zelle nicht, wie bei den
Flüssigkeitslamellen. Denn wir sahen schon früher, dass die Symmetrie-
verhältnisse der Zellen von der äusseren Form oft vollständig unab-
hängig werden, und auch unter Mitwirkung der äusseren Formverhältnisse
können bei dem ineinandergreifen der verschiedenen Factoren sich Thei-
lungsrichtungen ergeben, die mit den Forderungen des Princips der
kleinsten Flächen nicht in HAE ARI: stehen." (Berthold, '86,
p.230.) Berthold ('86, p. 230, Taf. 4) describes and figures many cases
in which the arrangement and division of cells is not in accordance with
the principle of least surfaces,

The fact that a single cell may at one time take such a form as that
which 49? shows in the surface view in Figure 37 (Plate 5), and in
section in ER 38, and at another time have the form exhibited by the
same cell in Plate 7, Figure 54 (surface), and Plate 6, Figure 48 (section),
while the shape of the egg remains unchanged, demonstrates that we are
not here dealing with a problem of the staties of a vesiculated fluid ;
a single simple principle can no more account for the forms taken than
it can for the protean changes of shape of an Amoeba,

By this I of course do not mean to imply that it is not possible, and
perhaps probable, that the laws of surface tension do, within certain
limits, modify the form and arrangement of cells, as maintained and
discussed at length by Berthold. Wherever the arrangement demanded
by the principle of least surfaces is not in conflict with other purposes of
the organism, or, to put it upon a less teleological basis, where it is not
in conflict with stronger influences than the force of surface tension, the
cells probably accommodate themselves to the demands of that principle.
The point of importance, however, is that this is not a decisive factor,
but may at once be overcome when other influences in the organism
antagonize it.

Zimmerman (93) holds that the general arrangement of cells in

accordance with the principle of least surfaces in plant tissues is not

60 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

due to surface tension at all, but to turgor. Turgor, however, can
hardly be a factor in the cleavage of the egg, where no increase in size is
taking place.

(2) Hertwig’s law of the spindle in the longest axis of the protoplasmic
mass. (Compare page 5.) — This is, so far as I know, the only principle
for which the claim is made, that it is % decisive factor in determining
the direction of the spindle. The statement quoted above (page 5) is
from Hertwig's general text-book on the subject, in which it is presum-
able that care would be taken not to mislead the reader unacquainted
with the literature into taking a special phenomenon for a law of general
import. On the same page it is stated that “ Mit diesen Regeln stimmen
die Erscheinungen, wie sie bei der Zelltheilung und insbesondere bei der
Eifurchung beobachtet werden, fast ausnahmslos überein." (Hertwig,
93, p. 175.) "The “law” has been accepted by others in the same gen-
eral bearing. Thus Ziegler (94, p. 154) questions the validity of cases
apparently not in agreement with the rule, holding that they are due
either to inexact study (cylindrical epithelium), or to the difficulties of
determining in the presence of a mass of yolk (amphibian egg) which 2s
the longest dimension of the protoplasmic mass. This principle has
become the most widely known and generally accepted of any of the
principles which have been proposed in regard to the determining factors
in cell division. I shall therefore discuss it at somewhat greater length
than Berthold’s principle, analyzing in detail the evidence on the subject
from my own work, and reviewing that advanced by others.

A comparison of the very first cleavage of Asplanchna with that of
Callidina (Zelinka, '91) shows that in the Rotifera the form of the egg is
not the factor determining the position of the first cleavage spindle.
For in the two forms the first cleavage spindle bears the same relation to
the animal pole, or place of polar-cell formation, but a different relation
to the long axis of the egg. In Asplanchna the spindle at the time of
division lies in the long axis of the egg (though a little earlier it is
oblique), whereas in Callidina (Zelinka,’91, Taf. I. Fig. 5) the spindle at
division is oblique to both the longer and shorter axes, — the place of
polar-cell formation not being the same as in Asplanchna, but much
nearer one end of the egg. The orientation of the spindle in these two
rotifers, then, is constant with reference to the animal pole, but variable
with reference to the form of the egg.

The passage from the eight- to the sixteen-cell stage in Asplanchna is
particularly instructive. The asters of the six cells of quadrants A, B,
and C first separate in such a way that the line joining them lies in the

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKTI. 61

long axis of the cell (Plate 3, Fig. 17), then a rotation takes place
(Fig. 18) by which the line joining the asters, i. e. the axis of the future
spindle, is brought into the shortest axis of the cell (Figs. 19-24). The
six spindles are then formed in the shortest axes of the cells (Figs. 20—
28), and the planes of cleavage accordingly coincide with the long axes
of the cells.

The simple fact that there are divisions in which the spindles lie in
the shortest axis of the cells is of course a direct contradiction of Hert-
wig's law. The case becomes even more striking, however, when the
movements of the asters are taken into consideration. They at first lie
in the position demanded by the law, but move from this position to
that which directly contradicts the law. (See pages 25, 26.)

Hertwig (93, p. 175) has cited a similar phenomenon, described by
Auerbach, as proof of his law. Auerbach (74) observed in the eggs of
Ascaris nigrovenosa, at the time of fertilization, that the two pronuclei
often come together in such a way that the plane separating them lies
in the short axis of the ege. Since the axis of the first cleavage spindle
commonly eoineides with the plano separating tho pronuclei, the result
in the eggs of this species of Ascaris would be that the spindle would
occupy tho short axis of the egg. But the two pronuclei after meet-
ing undergo a rotation through an angle of 90 degrees, thus bringing
the spindle into the dong axis of the egg. Ziegler has recently observed
with even greater clearness the same phenomenon in the eggs and
cleavage cells of other nematodes (Ziegler, 95, Taf. XVIII. Figs. 40-12),
and in the eggs of echinoderms (94). Te observed in nematodes in

some cases that the line joining the two asters on opposite sides of the

nucleus lies in the short axis of the ego, and that then follows a rota-
tion of the whole complex, till the line joining the asters — the axis of
the forming spindle — occupies the longest axis of the egg. Ziegler,
like Hertwig, has interpreted this change of position as a confirmation
of Hertwig’s law, and the interpretation is certainly the most natural
and apparently well grounded that could be given.

Nevertheless we have in Asplanchna an entirely similar phenomenon,
but occurring under such circumstances as to give a direct contradic-
tion, instead of a confirmation, of Hertwig’s law.

It is instructive also to notice that in the eight-cell stage of As-
planchna, notwithstanding the great variety in the form of the cells,
the direction of the cleavage spindles is the same in all the cells. Thus
d (Plate 2, Fig. 15), though irregular in shape, is of such a form

that it is possible to be confident that the spindle does lie in the greatest

62 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

dimension of the cell. The blastomere d*? (Plate 2, Figs. 14 and 16)
is so excecdingly irregular in form, that it is impossible to determine
with certainty which is the longest axis. The cells a*? — c*? are irregu-
larly triangular, and the direction in which the spindle lies is the short-
est line connecting any apex with the middle of the opposite side (Plate
3, Figs. 18, 19). Finally, the cells a*t- c*! are approximately rectan-
gular in form, with one of the axes much longer than the other; the
spindles lie in the shorter axes (Figs. 20-22). In every case the spindle,
whatever the form of the cell, lies in a meridian connecting the animal
pole with the vegetative pole of the egg. The only rational conclusion
from this fact is, that the position of the spindles is determined by
some factor unconnected with the form of the cells.

The theory that the direction of the spindles is due to their taking a
position of equilibrium determined by the mutual attraction of spindle
aud protoplasm, so strongly insisted upon by Ziegler (794, p. 140), is
likewise inconsistent with the movements of the asters in the cight-cell
stage of Asplanchna. Ziegler holds that, since the greater mass of pro-
toplasm must exercise the greater attraction, the spindle in the short
axis of the cell is in a position of unstable equilibrium ; if by any
cause it is moved in the slightest degree to one side or the other, it
must inevitably swing into the long axis of the cell, where alone it can
be in a position of stable equilibrium. This he holds to be the explana-
tion of the movements of asters and nuclei observed by Auerbach in
nematodes, and by himself in nematodes and echinoderms, as mentioned
above. An oblique position of asters and nuclei, such as is shown in
the coll att, Figure 22 (Plate 3), is intelligible on this assumption if
the movement taking place is from the shorter toward the longer axis.
But in this cell, as in the other five of the quadrants 4, B, and C at
this stage, the movement is from the longer axis toward the shorter.
The hypothesis that the movement is due to simple attraction between
the protoplasm and the fundament of the spindle, varying with the mass
of the protoplasm, is totally inconsistent with such a notion.

The passage from eight to sixteen cells is not the only cleavage in
Asplanchna which is irreconcilable with Hertwie's principle. At the
transition from the sixteen-cell to the thirty-two-cell stage, there is a
similar regularity in the position of the spindles coineident with variety
in the form of the cells, The four dorsal cells, a*t- d divide equa-
torially, three of them with spindles in the longer axis; one, %, with
the spindle in the shorter axis (Plate 4, Fig. 33).
iu, in the sixth cleavage the cell d^? shows the same phenom-

JENNINGS: DEVELOPMENT OF ASPLANCHNA. HERRICKII. 63

enon of rotation of the spindle into the shorter axis of the cell (Plate 5,
Fig. 37, Plate 6, Fig. 46, and Plate 7, Figs. 53 and 54). There is the
same regularity in the direction of cleavage as noticed in the preceding
cleavages, although there is variation in the form of the cells. Thus,
d*^, the companion cell of des, whose division with spindle in the short
axis has just been cited, cleaves with the spindle parallel with that in
des, but owing to the form of the cell the spindle in d** occupies the
long axis (Plate 7, Fig. 57, cleavage finished). Tue cells of the ventral
layer (Plate 6, Fig. 47) divide with spindles in tho same direction as
those in the cells of the second layer, although in the one layer the
spindles must thereby occupy the short axes of the cells, in the other
the longer axes.

In the seventh cleavage a still more striking case occurs. The mid-
dle of the embryo is surrounded by two rows of eight cells each, of
precisely the same form and size, the dorsal row composed of the blasto-
meres d. — d'9 and de - d'8, the ventral row of die -d, and at =.
These two rows are shown in Figure 61 (Plate 7), for the quadrants 4,
D, and C, and in Figure 57 for the quadrant D. In both rows, every
cell but one in each row (di and d", Fig. 57) is strongly flattened
dorso-ventrally, so that the lateral extent of the cell is much greater
than the dorso-ventral extent, If the form of the cell determines in
any way the direction of the spindle, it is certainly to be anticipated
that the direction of cleavage will be the same for the cells of both
rows, On the contrary, all the cells of the dorsal row divide meridi-
onally, while all the cells of the ventral row divide equatorially. In every
cell of the dorsal row, except possibly s, the spindle lies in the long
axis of the cell; in every cell of the ventral row, except , the spindle
lies in the short axis of the cell. The exception of a single cell in each
row gives a finishing touch to the proof that the form of the cells is not
the factor determining the direetion of cleavage. The fact that the cell
ds divides with its spindle in the same direction as the spindles of all
the cells in the same row, but not, as in the other cells, in the long axis
of the cell, while d*? likewise divides with its spindle in the same
direction as those of the other cells of dts row, but not, like them, in the
short axis, demonstrates that the dorsal row is not so constituted that
all the spindles must take their positions in the long axis, nor the ven-
tral row so that all must take their positions in the short axis. It
demonstrates, in other words, that the relative dimensions of tho differ-
eni axes of the cells does not determino the direction of the spindles in

either ono way or the other, (Tho cleavage of these two series of cells

64 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

is shown for the quadrants A, B, and C in Plate 8, Figs. 69 and 70;
in d" in Plate 7, Fig. 58; in d'5 in Plate 8, Fig. 66; in d" and d',
in Fig. 67.)

Still other divisions in which the spindle lies in the short axis bave
been followed out in the deseriptive portion of this paper.

We must conclude, therefore, that a very large number of cell di-
visions in the cleavage of Asplanchna directly contradict Hertwig’s law
that the spindle during division comes to lie in such a position that its
axis coincides with the greatest axis of the protoplasmic mass. The
characteristic feature of the cleavage is regularity in the direction of the
spindles, coupled with great variation in the form of the cells, thus
excluding any close relation between these phenomena.

What is the evidence upon which this law has been based ?

It is chiefly experimental, though there is likewise a certain amount
of evidence based upon the observation of normal cleavage.

Let us consider first the evidence derived from experiment. The ex-
perimental studies of Pflüger (84), Roux (85), Driesch (92), Hertwig
(93*), Born (93 and 794), Ryder (793), and Ziegler (904), on the effects
of pressure upon the direction of the spindle, are so well known that it is
not necessary to review them in detail. It is sufficient to state the general
result. With rare exceptions it has been found that when the egg or
the cleavage cell is so modified in form that one of the axes which may
be passed through its protoplasmic mass is distinctly greater than the
others, the spindle at cleavage comes to lie in this axis. I do not pro-
pose to enter upon an analysis of these experiments, nor to attempt
to explain in any different manner the results gained. A study of the
works of the authors above cited, and a repetition of the pressure ex-
periments upon the eges of the toad (Bufo lentiginosus Shaw) during
the spring of 1895, have convinced me that the explanation commorily
given is the one most in agreement with the conditions, and, from the
evidence, most probably correct, for these cases. But whatever we may
hold as to the validity of the explanation for these cases, we know that
the principle upon which it is based cannot be generalized, since in
many other cases it is directly contradicted by the facts. Before sng-
gesting how the experimental results may perhaps be reconciled with
the apparently contradictory phenomena observed in other cases, it will
be necessary to consider the evidence gained by other means, as well as
such experimental evidence as is against the principle.

First, then, we have the fact that the experimental evidence itself is

not concordant upon this point. Roux (85) found that under certain

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIT. 65

circumstances in deformed frogs’ eggs the first cleavage plane some-
times, though rarely (Roux, ’94*, p. 274), passed through the greater
axis of the cell, the spindle therefore lying in the shorter axis. Fycles-
hymer (95) experimented as to the effects of pressure on the eggs of
Amblystoma tigrinum, and found that when the eggs were compressed
laterally to one half their normal equatorial diameter, there was little or
no relation between the direction of cleavage planes and the greater or
less dimensions of the egg. “The first vertical in the thirty-four eggs
examined showed no constant relation to the compressed surfaces, in
seven passing through the longest equatorial diameter ; in nine through
the shortest ; in eighteen between the two.” (Eycleshymer, ’95, p. 353.)
In experiments of a different nature, Morgan (795) observed that the
shaken eggs of Spherechinus often divide at the first cleavage into
three equal blastomeres, and in such cases at the next cleavage the
three spindles lie in the short axes of the cells.

When we turn to the evidence from observation of normal cell
division, there is the same disagreement that is met with in the experi-
mental evidence. Until within a few years investigators in zoölogical
lines have not paid attention to the exact relations of the spindle to the
cell axes, so that little was to be found in the literature to emphasize
the necessity of caution in accepting at once the generalizations from
the first experimental results. In cases where cells containing spindles
were figured, there was generally an apparent agreement with tho
principle that the spindles are in the long axis, but so long as it was not
determined by observation whether this elongation of the cell was a
consequence of the position of the spindle or a cause of it, the evidence
was worthless, as pointed out by Heidenhain (94), and as clearly illus-
trated in the preceding pages.

In botanieal literature the ease was somewhat different, and seven
years before Hertwig (793) had stated that the phenomena observed in
cell division agreed “fast ausnahmslos” with his law, Berthold (86,
p. 230), in his thorough and comprehensive work on the subject, had said :
“Sehr oft theilen sich prismatisehe oder cylindrische Zellen der Länge
nach, wenn das Prinzip [of least surfaces] eine Querwand, der Quere
nach, wenn es eine Lüngswand verlangte. So theilen sich oft die Mark-
zellen, die Zellen der Grundparenchyms sich entwickelnder Blätter
nur quer, obwohl ihre Höhe im Vergleich zur Breite nur gering ist”
[Italics mine]. He had also given many examples of the conditions
thus characterized.

Recently the attention of zoölogists has been directed to a careful

VOL, XXX. — NO. 1. 5

66 BULLETIN: MUSEUM OF COMPARATIVE ZOÜLOGY.

observation of these relations, and a number of facts are now at hand
which bear upon the subject.

Ziegler (95) has studied the cleavage of certain nematodes (Diplo-
gaster longicauda, Rhabdites teres, and Rhabditis nigrovenosa) with
especial reference to the relations of cell form to the direction of cleav-
age, and finds that the conditions in these cases confirm throughout the
law of Hertwig. In the normal cleavage the spindle always places
itself in the long axis of the cell, even though a rotation of the nucleus
and asters from their first position is often necessary to accomplish this
result. And in cases where the egg is deformed by some outer agent,
as pressure by the walls of the oviduct or the like, the normal position
of the spindle is changed to agree with the changed form of the cells,
the spindle lying in the long axis in every case. The agreement with
Hertwig's law is complete.

On the other hand, zur Strassen (95 and '96) has studied the cleav-
age of another nematode, Ascaris megalocephala, with the same ques-
tions in mind, and has come to opposite results. In tho two-cell stage
one cell divides with the spindle in the long axis, the other with the
spindle in the short, axis, and in later stages a similar independence of the
position of the spindle and relative dimensions of the cell is shown. Zur
Strassen (95, p. 86) concludes: “Ich halte vielmehr den Kern für be-
fühigt, vermöge ihm inhaerenter Eigenschaften eine gewollte Theilungs-
richtung herbeizuführen, selbst wenn mechanische Hindernisse von nicht
unbedeutender Höhe dem entgegenstehen.”

Other observations bearing upon this question are much scattered.
Clases aro not uncommon in which authors have fignred spindles in the
shorter axes of the cells, but unless the observer's attention has been
especially directed to the question, such figures are of little value, since
& slight change in the position from which the cell is viewed produces :
foreshortening which gives very different apparent relative dimensions to
the axes. Heidenhain (95, p. 553) gives a number of such cases, from
most of which the evidence is weakened by this consideration. How-
ever, the case mentioned by Heidenhain of the germinal disk of the
cephalopod egg as figured by Watase (91) cannot be explained away
upon this ground.

Some other instances may be mentioned.

In the formation of the “germ bands” in the Polychat Amphitrite,
according to Mead (94, p. 467), “the axes of the spindles in these
divisions lie in the shortest diameter of the cells, and apparently in the

direction of greatest pressure.”

JENNINGS: DEVELOPMENT OF ASPLANCHNA IHERRICKII. 67

Wheeler (95, p. 309) states in regard to the first cleavage spindle of
Myzostoma: “In Myzostoma the spindle does not conform to O. Hert-
wig’s law, but always lies at right angles to the long axis of the often
very narrow protoplasmie pillar of the egg."

Castle (96, p. 250) states that in the division of the entoderm cells
of Ciona the spindles in certain cases lie in the short axes of the cells,
even when a shifting of the asters from a previous position in the long
axis must have occurred to bring about this condition. Castle states
that no mechanical explanation of this phenomenon offers itself, though
he holds that “other things being equal, it is true that the spindle arises
in the longest axis of the cell” (p. 231, note).

Tn the decapod crustacean Virbius, according to Gorham (795), the
egg is ellipsoidal in form, and the first cleavage spindle may occupy the
long axis, or be more or less inclined to it, or may even be nearly at
right angles to it.

The cell divisions in the germ bands of Crustacea as described by
Bergh (795), in which the spindles take the same direction for many
cell generations, should be added here (sce also MeMurrich, 95); though
tho evidence from these must be weakened in the eyes of the upholders
of Hertwig’s law by the fact that before division the cells * wachsen nur
in der Weise, das ihre Längsdurchmesser dem Querdurchmesser ziemlich
gleich wird und dann tritt die Theilung ein.”

The positive evidence upon the question from observations of normal
coll division is thus rather scanty, though doubtless some additions
might be made by a further study of the literature.

From both experimental evidence and observation of normal division
the only conclusion possible is, that in some organisms the spindle does
take a position in the greatest axis of the protoplasmie mass, apparently
without regard to other factors, while in other cases the position of the
spindle is determined by other factors, without regard to the form of
the cell.

The result is at first thought not very satisfactory, but this is not the
only organic phenomenon with regard to which such a conclusion must
be drawn. A few examples will make this clear.

Stahl (85) found that the direction of the first cleavage in the spores
of Equisetum is determined by the direction of the infalling rays of light.
A general statement of the effect of the direction of the light rays on cell
division would take a form similar to the unsatisfactory conclusion above
given for the shape of the cells.

If we leave cleavage and take up other growth phenomena, such as

68 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

the direction of growth of plants and animals, such a conclusion would
be reached with regard to almost every agent that ever exercises a
determining power. The effect of light is to cause certain parts to grow
towards it, others to grow away from it; while in other organisms the
direction of growth is not affected by it at all. The same is true of
gravity, of heat, and of various chemical and physical agents. (See the
extended list of such cases given by Herbst, ’94.) In all these cases
we are dealing with reactions to stimuli. It is only when we attempt to
make one of these agents the only determining factor, and expect to sce
it act always in the same way, as a simple mechanical cause, that our
result becomes unsatisfactory.

It is evident that in this question of the relation of the form of cells
to the direction of cleavage, we are dealing with a problem of a nature
similar to those which I have briefly stated above. Some organisms are
so constituted as to react to rays of light by growing toward them,
others are not. In the same way, some cells or nuclei are so constituted
as to react to the influences determining form by bringing the spindle
into the longest axis of the protoplasmic mass ; others are not. In each
case the result is due to a reaction to stimulus, or to an action of similar
nature, and not to à simple mechanical action of the agent.

In almost or quite all cases of reaction to stimuli, the result may be
shown to be the accomplishment of a certain end that is of importance
for the existence of the organism. The immediate explanation always
takes a teleological form. It is not diffieult to perceive a teleological
aspect of this tendency of many cells to divide with the spindles in the
longer axes. In cases where the purpose of cell division is merely to
double the number of cells without alteration of form, division with the
spindle in the longer axis is obviously the simplest method. A con-
sideration of what would be the result if the opposite method prevailed
shows this clearly. By continued division of a cubical cell and its
products with the spindle in the shortest axis, a series of flat plates
would be produced ; every cleavage would bring about a greater modifica-
tion in the form of the resulting cells, and cells of such form would
undoubtedly be very inconvenient for the purposes of the organism.
Continued division with the spindles in the long axis would result, on tho

other hand, in the production of cells of the same form as the parent
col. It is not remarkable, therefore, if in many cases the cells are so
constructed as to respond to a chauge of form by a corresponding change
of position of the spindle, so that the resulting cells shall be as nearly as
possible of the form of the parent.

|
|

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRIOCKII. 69

But mere inerease of the number of cells without change of form is
not the only object which can be brought about by cell division, and
when other purposes are to be accomplished the cells are so constituted
as to react in a different manner. Thus, I have shown that in Asplanchna
divisions of cells with spindles in the short axis is the method by which
is brought about the continued extension of cells in one direction, with
consequent gastrulation and a later invagination of ectodermal cells.
In Amphitrite, where, according to Mead, the division in the germ
bands is with spindles in the short axes, this method of division is per-
haps necessary to bring about the elongation of the germ bands, and the
same is doubtless true in the germ bands of Crustacea. It is of course
not necessary, nor is it probable, that in these cases the position of the
spindle in the short axis is a reaction to the form of the cell; more
probably the position of the spindle is determined without reference to
this, as I have endeavored to show for the cells of Asplanchna.

The conclusion to which I have come is therefore similar to that
maintained by Braem (94), except that he seems to imply that the pur-
pose of the cleaving cells is always the same, viz. to gain free space for
the development of the products of division, whereas it appears to me
that the facts indicate that the ends to be accomplished may be various,
and the means by which they are brought about equally varied, This
brings us to a nearer consideration of Braem's view.

(3) Braem’s theory of least resistance. (Compare page 5.) — As just
stated, so far as Braem’s view is teleological, I must agree with him; but
in so far as he seems to restriet his teleology to the accomplishment of a
single purpose, — the attainment of the freest space for development, —
it seems to me that the facts are against him. Certainly the principle
of “least resistance“ does not aid in understanding the cleavage of
Asplanchna, where a large number of the divisions take place in what
must. be considered the direction of greatest pressure, Examination of
the figures will show that, as a rule, the blastomeres in the resting
stages are much flattened dorso-ventrally and extended laterally, as if
subjected to great pressure; nevertheless, as shown in detail in the de-
seriptive portion and in the discussion of Hertwig's law, when division
takes place it is very frequently with the spindles in the dorso-ventral
axes. The cleavage in this direction seems to have a purpose, but that
purpose is not the gaining of tho freest space for the development of the
produets of division, but the accomplishment of the process of gastrulation.

Since Braem’s principle is confessedly teleological, it was probably not
intended to be rigidly applicable to all cases ; indeed, the author states

70 BULLETIN: MUSEUM OF COMPARATIVE ZOGLOGY.

as much. But in view of the frequent. departures from the principle of
least resistance, it appears necessary that the object of any given method
of cleavage should be judged in each case for itse If; so that Braem’s con-
tention that his principle is fitted “den Verlauf der normalen F urchung
in wesentlichen Punkten zu erklüren " must be considered unsuccessful.
It may, by calling attention to one method in which cells react to a
stimulus, “explain” the cleavage of some cells in the same sense that
the growth of the stem and root of a plant may be said to be “explained”
by saying that the protoplasm of one is so constituted as to react to light
by growing toward it, the other, on the contr ary, so as to react by grow-
ing away from it. But in other organisms the deter mining stimuli are
of an entirely different nature, and the “explanation” must be sought
anew for every organism.

I am of course fully aware that the view here put forth, that the
position of the spindle must be interpreted teleologically, Wer in many
cases as a reaction to stimulus, is not an *explanation. But I see no
a priori ground for expecting a simple mechanical explanation for the
direction of cell division, any more than we should of the direction of
growth of a plant. As a matter of fact, the phenomena show that such
an explanation is not at present possible for either.

(4) Rouzs theory of a compromise between tho tendency immanent in
the nucleus and the tendency due to the form of the protoplasmic mass.
(Compare page 6.) — As remarked above, this theory is not definite, in
the same sense as the three foregoing, inasmuch as one of its factors —
the immanent tendency of the nucleus — is of an entirely unknown
character. From the foregoing description and discussion it is evident
that I must agree fully with this conception. The further question
comes, In how far do the phenomena in Asplanchna lead to a recogni-
tion of the second factor, — the tendency due to the form of the cells
It is evident that the form of the cell does not determine the main fea-
tures of the direction of division, — the question as to whether the
spindle shall be dorso-ventral or lateral in direction. But are there sub-
ordinate features in which the form of the cell does affect the position
of the spindle, as held by Roux? In other words, does the spindle
always lie in either the long or the short axis of the cell, and never
oblique to both?

An examination of the figures will show that in the large majority of
cases the latter question is to be answered affirmatively. In the earlier
stages, before the great changes in the positions of cells have occurred, the
dorso-ventral axis of the egg commonly coincides with either the greater

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 71

or the lesser axis of the cells, so that from these cases no evidence on
our question can be gained. But in later stages the crowding together
of the cells results in greater or less alteration of the cell axes in rela-
tion to the axes of the egg. In Figure 68 (Plate 8), for example, the
long axis of the cell a?" is not parallel with that of d*?; the two are
wedged apart dorsally by the cell d*", so that neither one points
exactly to the animal pole of the egg. The two spindles are likewise at
an angle with each other, and lie in the long axes of the cells, instead of
exactly dorso-ventral. The same is true of the cells c^? and c™ in the
same ligure. Again, in Figure 74 (Plate 9) the spindles of d and
der, though in general direction dorso-ventral, form an angle with each
other, the one in de being modified in position so as to lie in the
longer axis of the cell, while the one in d' lies in the shorter axis of
the cell. In the entoderm the same thing is strikingly true of the
spindles shown in Plate 9, Fig. 79, and Plate 10, Fig. 83, for there the
spindles form various angles with one another, all lying in the longer
axes of the cells.

jut these cases do not necessarily lead to the view that the form of
the cell modifies even slightly the position of the spindle. It is
possible that in each cell the axis of the spindle is determined other-
wise, so that alteration of the position (not form) of the cell necessarily
produces an alteration in the direction of the spindle. "Thus in Figure
68 (Plate 8), if the cells d* and de. are ellipsoids of fixed form in
which the two ends of the spindles have predetermined positions, in case
the cells are forced apart, as in this instance by d*", tho ends of the
spindles will be forced apart to the same degree. Though we know that
the cells are not ellipsoids “of fixed form," yet we also know that the
form of the cell is greatly influenced by the direction of the contained
spindle, It is possible that the cytoplasm of the cell tends to group
itself symmetrically about the contained spindle, so that the direction
of the spindle is the primary factor, the fact that it lies symmetrically
in one of the axes of the cytoplasmic mass being a secondary result.
This becomes very probable when we examine from this standpoint the
change of form of the very irregular cell des, shown in Plate 5, Fig.
37 (surface view), and Fig. 38 (section), before the formation of the
spindle, and in Plate 7, Fig. 54 (surface), and Plate 6, Fig. 48 (section),
after the formation of the spindle. In this cell, before the forma-
tion of the spindle, the shape is so irregular that it is not possible
to distinguish a definite “ short axis," and no plane would divide the cell

into symmetrical halves. But as the spindle is formed, the cytoplasm

72 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

groups itself symmetrically about it, until just before division the
spindle lies in the shorter axis of the cell, and a plane including the
long axis of the spindle and the centre of the egg would divide the cell
into symmetrical halves.

If this be the true explanation, then the elongation of the cells in
Plate 10, Fig. 83, in various directions, is due to the previous altera-
tion of the predetermined spindle axes. And certainly the elongation
is not greater than is naturally the result of the position of the spindle,
as has been shown for other cells.

This view is strengthened by the fact that in certain cases, where
apparently it is impossible for the cytoplasm to group itself symmetri-
cally about the spindle, the latter takes a position which is oblique to the
axes of the cell. Such a case is shown in the entodermic cell d™! of
Figure 64 (Plate 8); the spindle lies in neither the longest nor the
shortest axis of the cell, but oblique to both.

It must be said, therefore, that the cleavage of Asplanchna gives no
positive evidence of any effect of the form of the cell upon the position
of the spindle ; on the contrary, the evidence on the whole is decidedly
against it. This, of course, cannot be generalized and applied to other
organisms, as on the other hand the observed conditions in other organ-
isms are not capable of giving generalizations which must hold for
Asplanchna. Generalizations in the field of reaction to stimuli, which
includes a very large proportion of organic activities, are exceedingly
unsafe and are justifiable only after exhaustive examination of the phe-
nomena; but in these matters scarcely more than a beginning of such
an examination has been made.

(5) Heidenhain’s problem of a definite angle of rotation. (Compare
page 6.) — This certainly cannot be held to have any especial signifi-
cance for the cleavage of Asplanchna ; it is proposed by Heidenhain as
applicable to systems of tissues in the later stages of organisms. I may
point out some facts which bear upon the “ problem.”

In the cleavage of Asplanchna, the angle which a given spindle makes
with the preceding spindle may be either 0 or 90 degrees ; but it is gen-
erally either one or the other, not some angle intermediate between
these.

The time and method of rotation, when rotation occurs, cannot be
said to be “ gesetzmässig.” Commonly the asters come to lie on oppo-
site sides of the nucleus before the rotation begins, but in the cells d&1
(Fig. 42, Plate 5) and d*? (Fig. 37) the change of position begins at

the same time with the division of the asters, so that by a separation of

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 78

the asters accompanied by motions not in a straight line, the same
result is attained that would otherwise be produced by a rotation after
the asters were fully separated.

In the larger cell at the two-cell stage, the change of position of the
nucleus and spindle is not due to a rotation after the two asters are
formed, but to a change of position of both nucleus and aster before the
aster has begun to divide.

In d*? (Plate 1, Fig. 7) the asters and nucleus at first rotate into a
position whieh is not the final one, so that subsequently rotation in
a different manner is necessary to bring the spindle into its definitive
position (Plate 2, Fig. 16).

In d*? (Plate 5, Fig. 37, Plate 6, Fig. 46, Plate 7, Figs. 53 and 54)
the definitive position taken by the spindle is at right angles to that of
the preceding spindle, so that the simplest method of formation would
be the natural separation at right angles to the axis of the preceding
spindle, as commonly takes place in such cases. But, owing appar-
ently to the peculiar form of the cell (Plate 5, Figs. 37 and 38, surface
and section), the asters separate obliquely, taking up a position such
that the lino uniting them is parallel to the axis of the preceding
spindle; and the definitive position is reached only by a later rotation
through an angle of 90 degrees (Plate 7, Figs. 53 and 54).

There is thus no regularity about the method by which the asters come
to occupy their definitive positions at the ends of the spindles. Appar-
ently the early position of the asters is influenced or determined by the
mechanical conditions within the cell, whereas the later position of the
spindle is largely independent of such conditions.

On the other hand, in the divisions of the ectoderm of Asplanchna,
there is a regularity in the final angle between the axes of two succes-
sive spindles in any given “layer” of cells, the second spindle being in
all the cells of a given layer either parallel with or else at right angles
to the preceding spindle. It is thus quite possible that there may be
whole systems of tissues where there is such a regularity. But the fact
that the definitive position may be reached by such various means
renders the phenomenon of little significance for a mechanical theory
such as that presented by Heidenhain.

(6) Sachs’s view that the walls separating the cells meet one another at
right angles. — An examination of the figures, especially the sections,
shows that the condition above stated is not generally complied with in
the cleaving cells of Asplanchna, so that it is not necessary to enter upon
a discussion of this view. Neither is the regular alternation of spindles

74 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

at right angles to one another — a theory often attributed to Sach — the
rule in Asplanchna.

(7) Rauber’s theory, that the asters of the different blastomeres exercise an
attraction for one another, thus influencing the direction of the spindles. —
There appears to be nothing in the cleavage of Asplanchna that would
lead to such a conclusion as the above. Any attempt to explain the
relative positions of asters in the different cells — in such a case, for ex-
ample, as is shown in Figure 39 (Plate 5) — by the theory of attractions
and repulsions among the asters, will be found to lead to. purely arbitrary
assumptions as to which asters exercise attractions on others and which
do not; any geueral statement of a positive nature will be found to be
inconsistent with the facts. Moreover, the irregularities in the migrations
of the asters, to which attention was especially called in the discussion of
Hoidenhain's * Problem of a regular angle of rotation,” can with difficulty
be harmonized with such a theory.

(8) Braems principle of equal resistance at the two ends of the spindle.
— This principle is discussed in connection with the question which

immediately follows.

B. What determines the Equality or Inequality of Cleavage Products ?
C. What determines the Rate of Cleavage ?

Owing to the similarity of the factors supposed to determine the
relative size of the cleavage products and the time of cleavage, it will be
well to consider together the foregoing questions.

In regard to the first question, we have (1) the theory of Hertwig,
stated on page 8, that the spindle tends to take a position in the middle
of the mass of protoplasm in which it is contained, and (2) the “ principle
of like resistance at the two ends of the spindle,” set forth by Braem,
and stated on page 7. Since these two theories lead to similar results,
they may be considered together.

It certainly does not aid in understanding the cleavage of Asplanchna
to assume that “der Kern von vornherein das Jestreben hat, sich
gleichmässig nach beiden Seiten hin auszudehnen und somit auf eine
üquale Zelltheilung hinzuwirken ” (Braem, 94, p. 345), nor that “der
Kern stets die Mitte seiner Wirkungssphäre einzunehmen sucht?“
(Hertwig, 93, p. Ira INO differentintion into more and less yolk
bearing regions is visible. Yet the first cleavage is unequal, and the
second equal in one cell and unequal in the other; the third likewise
shows both methods, with a preponderance of equal divisions ; and in
the fourth all the divisions are again unequal. In no cleavage is there

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 75

throughout an equality in the size of the two products in all the cells.
The two assumptions above stated are strikingly contradicted, not only
in the formation of the polar cell (Fig. 1), but also in the fifth (Fig.
35, Plate 5) and sixth (Fig. 49, Plate 6) cleavages of the entodermal
cell (di. and de-), and in the division of the entodermal cell d*? (Fig.
80, Plate 10). In these cases the spindle moves against one side of th
cell, and there a small vesicle is produced, the two products of cleavage
being exeessively unequal. "The contradiction is emphasized by tho fact
that in the fifth and sixth cleavages of the entodermal cell (Figs. 35,
48, and 49), although there is a concentration of yolk in ono region of
the cell, it bears no significant relation to the position of the spindle, and
by the further remarkable fact that in the next or seventh cleavage
(Figs. 64 and 65, Plate 8), this yolk cloud passes into the smaller
blastomere.

When one considers the method of formation of polar cell in all eggs,
and the almost unlimited range in the comparative size of cells resulting
from cleavage in a great number of organisms, the grounds for the
generalization, that “der Kern stets die Mitte seiner Wirkungssphüre
einzunehmen sucht,“ are ditficult to comprehend, at least so far as they
relate to the dividing nucleus. As a statement of fact for specific cases,
with the word “stets” omitted, I have no fault to find with it; but as a
generalization it is, so far as it relates to the dividing nucleus, meaningless.
The position of the spindle within the cell must be considered to be re-
lated to the purpose of the ensuing division. It is of course probable
that in many eases the primary purpose is to divide the formative proto-
plasm equally botween the two products, and this may determine the
position of the spindle in, for example, the first, second, and third
cleavages of the frog's egg. But what determines the position of the
spindle in the two divisions immediately preceding these, — in the forma-
tion of polar cells? It seems to have been generally overlooked, even by
those who have pointed out that polar-cell formation is cell division, that
the formation of polar cells must be reckoned with in any general theory
of cell division.

An examination of the cleavage of many invertebrates shows that, as
in Asplanehna, equal eleavage is no more the rule than unequal cleavage,
even where there is no corresponding differentiation into yolk bearing
and purely protoplasmie regions. The statement that the dividing
nucleus tends to take a position in the middle of its sphere of action is
true, if we consider the “middle of its sphere of action” to be, as it
actually is, the point where it divides. But with this interpretation the

76 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

statement is of course utterly without significance. The real question
is, What determines the point where it divides? Why in Asplanchna
does the middle of the sphere of action in the fourth and seventh
cleavages of the large ventral entoderm cell lie in such a position (Fig.
16, Plate 2, and Fig. 65, Plate 8) as to divide the cell into parts which
are only slightly unequal, while in the intervening fifth and sixth
cleavages (Fig. 35, Plate 5, and Fig. 49, Plate G) the middle of the
sphere of action is at the periphery of tho cell! Why at tho first
cleavage of the frog's egg is the middle of the sphere of action in the
centre of the mass of formative protoplasm, whereas at the divisions
immediately preceding — the formation of the polar cells — it is at the
periphery of the egg !

In regard to the question what determines the comparative rate of
‘cleavage the case is similar. In view of the recent discussion of these
two questions — the inequality of cleavage and the rate of cleavage — in
the works of Kofoid (94, p. 196), McMurrich (795), Lillie (95, p. 45),
zur Strassen (95), Ziegler (95), and others, it seems scarcely worth
while to insist upon the fact that the rate of cleavage and the equality
or inequality of the produets are related to the future morphogenetio
processes, and show in many cases no relation whatever to accumulations
of yolk. Yet the so called laws, according to which these matters are
determined by the distribution of yolk, are repeated in O. Hertwig’s
text-book on the cell (Hertwig, 93, pp. 174 and 180), and reaffirmed in
the latest edition of his treatise on embryology (Hertwig, ’96, pp. 67
and 84).

In Asplanchna the main facts in regard to the sequence of cleavage
are as follows.

(1) The order of cleavage is, within very narrow limits, constant. If
a number of eggs are taken showing successive stages of the division of
any given cell, the series will show corresponding successive stages in
the nuclear history of the other cells. i

(2) There is a typical order for similar cells of different quadrants of
the same egg. This order is D, C, B, A.

(3) There is a marked general tendenoy for larger cells to divide
first. In every case where two cells of the same origin are of different
size, the larger divides before the smaller.

This is a very general phenomenon, as has been repeatedly pointed
out of late, even in cases where the larger cell contains the greater
amount of yolk. The “law” recently formulated by Hertwig (96,
p. 67), that “die Schnelligkeit der Furchung proportional ist der Con-

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIT. T

centration des im Theilungsstück befindlichen Protoplasmas,” must be
considered — like the corresponding one in regard to the equality or
inequality of cleavage produets— as an example of the immature
generalizations to which embryology has fallen heir through the acci-
dental circumstance that the amphibian egg was for a long time the chief
object for the study of cell division. It seems to be exceedingly dit-
cult to grasp the fact, every day becoming more evident, that because a
statement is true for the eggs of some of the lower vertebrates it does
not follow that it must be true for the cells of all organisms.

(4) The rhythm of cleavage has an important relation to the other
processes of morphogenesis. If eleavage took place coincidently in all
the cells of the egg, the latter still retaining its form, there could
apparently be no “rotation” of the cells upon one another, and conse-
quently no gastrulation. The tension in all directions would be the
same; none of the cells would be moulded to fit the extension of
neighboring cells, and all would retain approximately the positions held
at the beginning

It would perhaps be possible to carry this into detail, and show that
the earlier cleavage of the large ventral cell, leaving it in a resting
condition, and therefore plastic (as indicated by a comparative study of
the forms taken by resting and by dividing cells) when the other cells
divide and extend, is directly favorable to gastrulation.

The question as to the faetors determining the time and the equality
or inequality of eleavage, ib will be seen, does not at present admit of
any direct simple answer. Cleavage is a part of the process of morpho-
genesis, and its rhythm and other features are related to the nature of
the form to be produced.

D. As to Differentiation accompanying Cleavage.

The facts bearing upon this question to be derived from the observa-
tion of the early development of Asplanohna are few in number. The
principal phenomenon to which attention must be directed is that of the
segregation and migration of the cloud of granules, described on pages
25, 30, 37, and 54. As will be remembered, a concentration of granules
begins in the ventral region of one of the cells at the eight-cell stage
(Fig. 7) and becomes more and more marked as successive cleavages
take place, till a well defined cloud of very large and distinct spherules
occupies the anterior and ventral margin of the cell d* in the stages
immediately succeeding the sixteen-cell condition (Fig. 32, Plate 4, and
Fig. 48, Plate G). Then oceur the remarkable migrations shown in

78 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Figures 49, 50, 51, and 64 (Plates G and 8), until the cloud of granules
is enclosed at the seventh cleavage in the smaller of the two entoderm
cells (Fig. 65). This whole process shows clearly that other changes
of astriking character are taking place at the same time as the division
of the egg into smaller portions; evidently cleavage is not a mere
separation of the egg into smaller masses, each similar to the other and
to the original egg. The final destiny of this granular mass is not
known, but such a peculiar and well characterized phenomenon as it
exhibits cannot be considered meaningless.

The differentiation in this instance is of the kind admitted by Driesch
(see page 9), in that it is cytoplasmic in nature. It is not, however,
a direct consequence of the original distribution of materials within the
egg; the migrations of the granules show that processes are taking
place in the cytoplasm that are only indirectly connected with cell
division.

We have in this case a distinctly visible differentiation accompanying
cleavage. Certain other phenomena give evidence that there are like-
wise invisible differentiations accompanying the process.

At the division of the second “layer” of ectodermal cells in the
sixth cleavage, shown in Figure 55 (Plate 7), the two rows of cells a’ —
c5, gt oe, Tec”, and aus o are produced. The cells of these
two rows, as shown in Figure 61, are of the same size and the same
form, having similar relations to the surrounding cells and to the axis of
the embryo. Yet, as has been repeatedly stated, all the cells of one row
divide meridionally and equally with spindles in the long axes, the cells
of the other row equatorially and unequally with spindles in the short
axes. What causes this difference

The difference must, of course, be due either to a different stimu-
lus from the outside, or to a different structure of the cells. The
problem may be expressed clearly in this way : If one of the cells of tho
more dorsal row, as a"? (Fig. 61), could be removed and placed in
the position now occupied by d, in the more ventral row, would it
change its method of division? That is, would it cleave equatorially
and unequally, with its spindle in the short axis, like the other cells of
the ventral row, instead of meridionally and equally, with its spindle in
the long axis, as it actually does?

There is, of course, no way of answering this question directly. It
scarcely appears probable, however, tbat there is such a difference in the
influences affecting the two cells as to cause so fundamental a difference

in the cleavage. And if there is not, the only alternative is, that there

JENNINGS: DEVELOPMENT OF ASPLANCHNA IIERRICKII. 79

was a qualitative division, nuclear or cytoplasmic, at the preceding
cleavage.

This conclusion is, of course, speculative, but the history of the
cloud of granules direetly proves that in Asplanchna the cleavage is
accompanied by differentiation.

The recent experimental evidence, showing that in certain organisms
the cleaving cells in early stages all possess the inherent capacity to
produce an entire animal, has led to a rather widespread impression
that cleavage has been shown to be a process of little or no significance,
being merely a quantitative division of a mass into smaller masses of a
similar nature. This view apparently receives confirmation from gen-
eralized statements of the results of such experiments; for example,
the following from Driesch (94, p. 69): “Es liegt also nach allem
gesagten in der That kein Grund vor, in der Furchung etwas anderes als
reine Zellteilung zu sehen; ja die Gleichheit der Furchungskerne ist
direkt durch Versuche bewiesen.” A summary of the evidence which
has been adduced in regard to this matter shows that such a statement
as the above conveys only a small part of the truth and must lead to
error unless carefully interpreted. The evidence that cleavage is accom-
panied by differentiation may be summarized as follows.

(1) It is directly proved by observation that in certain cases the nu-
cleus differs in structure in different blastomeres in early cleavage stages,
and that this differentiation is correlated with a different fate of the differ-
ing cells. This Boveri has shown for Ascaris megalocephala (Boveri, 94),
and Meyer (795) for certain other nematodes.

(2) It is directly proved by observation that in certain cases tho cyto-
plasm of the different cleavage cells in early stages is of a different
structure, reacting differently to chemical reagents, and this difference
is correlated with a different fate of the different cells. Thus, in the
sixteen-cell stage of Nereis, “the somatoblast can always be recognized
at a glance” from its different color (Wilson, '92, p. 390). A similar fact
has been shown above in regard to the cytoplasmic differentiation in
Asplanchna, but here I have not been able to determine the fate of the
cell which receives the differentiated granules,

(3) It has been shown that in many cases the different blastomeres of
early cleavage stages give rise to definite structures in the adult. This
fact in itself of course admits of two interpretations, but taken in con-
nection with the facts stated under (1) and (2) it becomes of great
significance,

(4) It has been shown experimentally, that in some organisms sepa-

80 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

rated blastomeres give origin to parts of the embryo only. The most
complete and satisfactory cases are those of ctenophores as described by
Chun (92) and confirmed by Driesch und Morgan (95), and of the
gasteropod Ilyanassa, by Crampton (96).

It is difficult to conceive how a more complete demonstration could be
possible, that cleavage is accompanied in many cases by differentiations
which are not expressed by the phrase “reine Zellteilung," and that
these differentiations are of the utmost significance for the future
development of the organism. Any amount of evidence that in other
cases there is no differentiation cannot in the least shake confidence in
this demonstration

2. GASTRULATION.

In addition to the problems béaring directly upon cleavage, the plan
of the present work included a study of some of the later morphogenetic
processes, affecting masses of cells and leading to the differentiation of
organs, in order to determine the relation of cleavage to these. Of these
processes, gastrulution and the ensuing invagination of ectodermio cells
to form the pharynx (Zelinka, 91) were studied. These are in reality
parts of a single process, so that they may be treated of together under
the title of Gastrulation.

In regard to the relation of cleavage to gastrulation, the result is
evident from the account given in the descriptive portion of this paper.
No separation of the two processes is possible; gastrulation is an
accompaniment and a consequence of cleavage. At the passage from the
four-cell stage to the eight-cell stage begins a displacement of the
blastomeres ; this displacement, or “rotation,” continues in later cleav-
ages in the same direction, and is still in operation at the latest stage
examined, when it is no longer possible to follow the development cell
by cell As one of the phases of this displacement during cleavage,
the large ventral cell of quadrant D gradually moves inward, followed
later by a similar transference of the ventral cells of the other quadrants
to the inside. The entire process has been followed step by step in the
descriptive portion of the paper, so that it is not necessary to go into
details here. In its general features the process is as follows. As the
ectodermal cells begin to pass into the karyokinetic condition, they
withdraw their more internal parts and increase in surface extent. ‘The
egg as a whole retains its form and size, so that the withdrawal of the
internal parts of one cell necessitates an inward movement on the part
of others; the result is à gradual inward migration of the ventral cella.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 81

The inward movement of the large cell of quadrant D, rather than that
of the other cells, seems due to several causes: — (1) The inequality of
the cells. The large ventral cell having a much greater radius of
curvature has less surface tension, and therefore may more easily change
its form. () It thus yields to pressure, and fits itself to the changing
form of the smaller cells. These are thus able to creep over it, as it
were, and surround it. The greater quantity of cytoplasm in the
entoderm cell as compared with the size of the spindle seems also to
result in less change of form at the time of karyokinesis. The large cell,
in virtue of its mere size, conducts itself on the whole passively with
relation to the more active smaller cells. (2) The sequence of cleavage
is possibly connected with this. At a given cleavage the large cell
divides first, so that, when the karyokinetic stretching of the other cells
takes place, the entodermic cell is in a resting condition, and therefore
passive. (3) The direction of the spindles, which is prevailingly dorso-
ventral, results in a continued dorso-ventral extension of the cells, so
that invagination would naturally take place at one of the two ends.
The developing egg may be likened to a fountain in which there is an
upward movement within, an outflow above, at the animal pole, and
a downflow about the periphery.

The enclosure of the large ventral cell of quadrant D is what has been
considered gastrulation proper by Zelinka and Tessin, only the products
of this cell being spoken of as entoderm. But after this enclosure is
complete the process continues, unchanged in character, the ventral cells
of the other quadrants following that of quadrant D to the inside of the
embryo, as shown in Figures 76-79 (Plate 9).

A necessary condition for all this displacement is of course the
retention by the egg as a whole of its form and outline. If the blasto-
meres should separate and project above the general level, in the manner
that is common for the cells of mollusks (see the figures of Unio by Lillie,
'95) at the time of karyokinesis, no compensating inward movement of
the other cells would be necessary, and apparently therefore gastrulation
would not take place. The retention of the regular form appears thus
to be of the highest, importance for the development, and the questioh
arises as to how this form is preserved. As previously stated, no mem-
brane is visible; and any uniform force, such as surface tension or a
centripetal attraction, would produce a spherical instead of an ellipsoidal
form. The development of the egg proceeds as if it were enclosed in a
rigid mould of oval or ellipsoidal form, so that the contents of the mould
are rotated, while the form is retained. The retention of this shape

VOL. XXX. — NO. 1. 6

82 BULLETIN: MUSEUM OF COMPARATIVE ZOÜLOGY.
seems to me inexplicable on any simple mechanical ground ; the form at
this stage apparently must be judged from the same standpoint as that
of the adult, which no one would attempt to refer to a simple mechanical
principle. Of course the assumption of the presence of a non-elastic
non-extensible membrane of ellipsoidal form, which later becomes elastic
and extensible, would explain the retention of the shape, and is open to
any oue who chooses to make it. But several facts speak against this
view, aside from the general improbability of the existence of a membrane
of such peculiar and changing qualities : —(1) The negative evidence
that no such membrane can be demonstrated in preserved material.
(2) In the sea-urchin and in Amphioxus, as shown by Driesch (93) and
Wilson (793), development takes place as well when the membrane is
removed as when present. This of course does not show that the same
is true for Asplanchna, but it does show that the importance of the egg
membrane has been overestimated for some cases. (3) In the rotifer
Callidina, investigated by Zelinka (91), the egg is of the same form as
in Asplanchna, yet the cells sometimes put forth short amœboid pro-
cesses, which of course would be impossible with a close membrane.
(4) In another rotifer, Melicerta ringens, the egg is not a regular oval or
ellipsoid of rotation, but one side is flattened while the other is curved,
and this irregular form is retained during the shifting of the blastomeres,
as is the case in Asplanchna (see tho figures of Zelinka, '91, and of Joliet,
83). Such a form would not be preserved even by such a non-elastic
membrane.

The facts given under (3) and (4) seem to me to render entirely in-
admissible the explanation that the form of the egg in Asplanchna is
due to the presence of a membrane,.since this would leave the exactly
similar phenomenon in the related forms Callidina and Melicerta without
explanation.

The factors concerned in gastrulation may be summarized as follows : —

1. The form of the egg.

2. The change in the form of the cells at cleavage.
3. The direction of cleavago.

4. The inequality of the cleavage.

5. The sequence of cleavage. (7)

The process of which gastrulation is a part begins with the third
cleavage, and continues through all the period in which it is possible to
trace the development cell by cell, and apparently much later.

The process of gastrulation as above described for Asplanchna is
similar to the method briefly set forth by Ziegler (95, p. 402, note) for

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKL. 83

Rhabditis nigrovenosa: “Ich stelle mir die Mechanik des Gastrulations-
vorganges so vor, dass die Ektodermzellen nach ihrer Theilung sich
abflachen und in Folge dessen ausbreiten; dabei schioben sie die
Mesodermzellen über die Entodermzellen herüber. Es kann dies um so
eher geschenen, da die Mesodermzellen zum Zweck der Theilung sich
kugelig zusammengezogen und dabei an die Oberflüche des einschichtigen
Epithels emporgehoben haben."

On the other hand, zur Strassen ('95 and '96), in his careful studies
on the early development of another nematode, Ascaris mogalocephala,
camo to an entirely different conception of the factors at work in the
displacement and extension of the ectoderm cells. Zur Strassen holds
that four cells of common origin constitute an * elementary mechanism,"
the two pairs of which attract each other in such a way as to bring about
the movements which actually occur. The interaction is thought of as
something having the nature of the “cytotropism” of Roux (95). It
is not necessary to discuss the matter further here, since there is no
occasion to call in any such action in the gastrulation of Asplanchna.
Doubtless the inter-attraction of cells exists in the rotifer, as elsewhere,
but it seems to have no determining significanoe for the movements
which take place.

3. GENERAL CONSIDERATIONS.

T shall next take up certain general considerations upon the mechan-
ies of cleavage aud development, which do not naturally fall under the
discussion of any of the theories above considered, together with a
general review of the results gained.

The egg of Asplanehna at the four-cell stage might be compared to
the egg of an echinoderm, with the exeeption that one of the four
blastomeres is much enlarged, and of a different form from the others.
What variation in the cleavage will be induced by the differences in
form ?

As we have seen, the form of the cell in Asplanchna does not affect
the direction of the spindle at cleavage. Indeed, the most character-
istic feature of cleavage in Asplanchna is difference in ferm and size of
the blastomeres, coupled with identity in the direction of spindles, in cells
having the same general relations to the axes of the embryo.

In a closer consideration of the factors determining the position of
spindles, it is evident from the phenomena described that the question
is not a simple one, but must be resolved into (1) what determines the

direction of separation of the asters; (2) what determines the position

84 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

the movements of the asters into the definitive position occupied by
the spindle.

(1)and (2). In regard to the first two questions, two general facts
are worthy of notice.

First, there is a tendency, other things being equaı, for the newly
formed asters to separate at right angles to the axis of the preceding
spindle, and in such a way that the asters are not to be distinguished as
deep and superficial, but either as right and left, or dorsal and ventral.
No explanation for this fact is apparent, and it is not in every case true.

of the asters before the spindle is formed ; and (3) what determines

Thus in the two-cell stage tbe nucleus and aster in the large cell migrate
to the right before the aster divides, and the separation of the two
newly formed asters is not in a plane at right angles to the axis of the
preceding spindle.

Secondly, the position of the asters before the formation of the spindle
may apparently be modified by the simple mechanical conditions sur-
rounding them. Thus, in the cell d*? (Fig. 14, Plate 2) the asters are
modified in position almost immediately after they begin to separate, so
that very soon we actually have in this cell the condition which may be
considered least typical, — an inner and an outer aster (Fig. 7, Plate 1).
In the thin cell d^? (Figs. 37 and 38, Plate 5, and 46, Plate 6) the form
of the cell apparently operates to cause the two asters to separate in
such a way that almost from the first the line joining them has the
same direction as the axis of the preceding spindle. Such faets give
the impression that before the formation of the spindle the position of
the asters is undetermined, and indifferent for the general structure of
the cell, except that the two asters always lie on opposite sides of the
nucleus.

(3) As the karyokinetic changes take place, the asters migrate into
definite positions, apparently by a rotation of the whole complex com-
posed of the nucleus and its two accompanying asters. This rotation
is into a definite position, without regard to either the form of the con-
taining cell, or the previous position of the asters; that is, the end to
be gained is constant, while the means of gaining it vary. Thus, at the
divisions to form the sixteen-cell stage, the line passing through the
asters and nucleus rotates in d*! from a lateral to a dorso-ventral posi-
tion, and into the greatest axis of its cell (Figs. 11, 15, and 16, Plate 2).
In d*? it rotates from a position in which one aster is central, the other
peripheral (Plate 1, Fig. 7), likewise into a dorso-ventral position (Fig.
16). In a and dec all the axes rotate from a position of

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 85

lateral extension to a dorso-ventral position, and thereby from the greater
into the lesser axes of their respective cells (Figs. 17-22, Plate 3). In
des the asters do not separate at right angles to the previous spindle, as
commonly occurs, but the line joining them is parallel to the preceding
spindle, i. c. lateral (Figs. 37 and 38, Plate 5, and 46, Plate 6) ; later,
by a rotation into the short axis of the cell, the dorso-ventral position is
attained (Plate 6, Fig. 48, Plate 7, Figs. 53 and 54). It is not possible
to refer these and the other changes described in the general account
of the development to any simple factors. We can refer the changes in
position of the asters, and consequent manner of cleavage, only to the
structure of the protoplasm and the (molecular?) processes occurring
within it.

The fact that the spindles take definite positions with relation to the
axes of the developing embryo, but without regard to the form of the
cells, scems to indicate that there is some influence governing the egg
as a whole, which is related to its form, and that the position of the
spindles is regulated by this. The determining factors in the position
of the spindles would therefore lie, not within any given cell itself, but
outside of it. But there are certain facts which seem to render this
very doubtful. As discussed on pages 70, 71, in later stages the cells
become displaced by the changes taking place during gastrulation, and
there is a corresponding change in the position of the spindles ; they
are no longer either parallel with or at right angles to the dorso-ventral
axis of the egg. This is shown especially in Figures 68 (Plate 8) and
83 (Plate 10). If a changed position of the cell with regard to the
axis of the embryo results in a correspondingly changed position of the
cleavage spindle, it seems to follow that the position of the spindle is
determined within the cell.

I do not, however, consider this conclusion as well established. It
still scems possible that the spindles are all placed with reference to some
influence resulting from the axial relations of the egg as a whole, —
though not necessarily in all eases either in the dorso-ventral axis or at
right angles to it.

Comparison of the conditions in Asplanehna with those reported by
other observers for other organisms shows that there are cases in which
the form of the cell does determine the position of the spindle. In the
same way we know that there are enses in which the direction of the in-
falling rays of light determines the position of the spindle (Stahl, '85).
But the result is not universal for either agent, so that we must hold

that the effect in both eases is of the nature of a reaction to stimulus.

86 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The form of the cell and the direction of pressure cannot therefore
be used in explaining in any general way the direction of cleavage, as
proposed by Hertwig and Braem. The method of reaction and the
purpose of the reaction must be determined for each class of casos in
itself.

A consideration of the process of gastrulation leads, though from the
opposite direction, to conclusions of a similar nature. The form and
direction of cleavage are related to the later morphogenetic processes.
Gastrulation is a result of the method of cleavage, — and the method of
cleavage must be looked upon as adapted to the purpose of accomplish-
ing gastrulation.

The relation of the form of cleavage to the later morphogenctic pro-
cesses is Shown in a different way in such forms as Nereis (Wilson, ’92)
and Unio (Lillie, *95), where it has been possible to show the exact
relation of later organs to individual blastomeres. Cleavage in many
cases is itself a direct morphogenetic process, the exact method of
which can be referred to no more simple mechanical factors than can
the characteristic form of the adult.

I do not of course wish to generalize this statement; it is evidently
true for many forms, but may not be true for all. The evidence upon
which a contrary view is sometimes maintained for certain forms, as
the echinoderms, seems however inconclusive. The formation of the
micromeres has been shown to be preceded by a differentiation in the
cytoplasm (Morgan, '94), which would naturally lead to the conclusion
that the micromere formation is a process having a definite signification
for morphogenesis. But Driesch (93) showed that the micromeres
might be removed and a normal larva still produced. From this, how-
ever, it does not follow that the formation of micromeres is without
definite morphogenetic significance, any more than it follows that the
fundaments of limbs are of no morphogenctic significance because a
normal larva results after the embryonic limbs have been removed from
a young amphibian. The formation of micromeres apparently segre-

1

gates a certain amount of substance which needs to be localized in a

definite region. If this segregated material is removed, there is no
evidence proving that similar material is not again segregated, at per-
haps a later stage. As Roux insists, it is important to distinguish
the normal method of development from induced development due to

injury.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 87

PART SECOND.—DISCUSSION OF MATTERS BEARING
UPON THE MORPHOLOGY OF THE ROTIFERA.

Our knowledge of the development of the Rotifera is due chiefly to
the work of Zelinka (91). This author has given a full and careful
description of the development of Callidina russeola Zel. from the egg to
the adult form, with a briefer, but still extended, comparative account of
the development of Melicerta ringens. — Earlier works on the embryol-
ogy of the Rotifera are due to Salensky (72), Joliet (83), Zacharias
(84), and Tessin ('86); but all of these works are incomplete and in
many respeets inaceurate, so that they have been almost completely
superseded by the work of Zelinka. In diseussing the development of
Asplanchna I shall therefore restrict myself chiefly to a comparison with
the results of Zelinka, drawing upon the accounts of other authors only
where there is special oceasion.

Since my work has been done primarily from the standpoint of cyto-
mechanics, and not with regard to the morphology of the Rotifera, it has
an entirely different aim from that of Zelinka. It thus results nat urally
that, in giving an account of the bearing of my studies on questions
relating primarily to rotifer morphology, emphasis must be laid chiefly
upon the points in which my results differ from those set forth in
Zelinka’s paper. The plan of my work required a more minute study of
the eleavage than was demanded for Zelinka's purposes, and as a natural
rosult I shall be compelled to eriticise his acconnt of the cleavage in
regard to certain details. Furthermore, it will be necessary to show
that Zelinka has been incousistent in his aecount with regard to the
place where the polar cell is formed, and hence is mistaken in his state-
ment of the axial relations of the egg and embryo in the Rotifera. But
all these are matters of detail, not affecting in any important way
Zelinkws general conclusions, and I wish to say at the beginning that I
fully appreciate the thoroughness and excellence of Zelinka’s researches
upon this difficult group, and make the criticisms and corrections con-
tained in the following pages in no spirit of disparagement.

Asplanchna Herrickii de Guerne, the form upon which my studies have
been made, is not closely related to any of the species of Rotifera whose
devolopment has been previously described. Callidina and Melicerta,
investigated by Zelinka (91), belong respectively to the groups Ddelloidea
and Rhizota of Hudson and Gosse (86). Rotifer and Philodina, studied

by Zacharias (81), belong also to the Bdelloidea. Brachionus, investi-

88 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

gated by Salensky (’72), is one of the Loricata. Only Eosphora, upon
which the work of Tessin (86) was done, belongs, together with
Asplanchna, to the swimming Illoricata. But even these two are widely
separated, Eosphora possessing a foot and anus, like the majority of the
Rotifera, whereas Asplanchna possesses neither.

The only work which has been done upon the early development of
any species of Asplanchna is that by Leydig (54) and Lameere (90).
The latter observed in the living egg of Asplancha Sieboldii the formation
of the polar cell and the first and second cleavages, but did not carry his
work further. Leydig (54) observed the cleavage of the egg, and
figured the four-cell stage in the same species, but does not give any
detailed deseription. I know of no other work dealing with the early
development of any species of Asplanchna.

I have examined the early stages of both Asplanchna Herrickii de
Guerne and Asplanchna priodonta Gosse, of which the material collected
in the Great Lakes contained about equal numbers. Asplanchna Herrickii
was chosen for special investigation on account of the greater size of the
embryos and adults. The development of Asplanchna priodonta was
examined for comparison, and notes upon the embryology of this species
have been given in connection with the fuller discussion of Asplanchna
Herrickii.

1. Previous Knowledge of Asplanchna Herrickii.

It will be well to give here a brief review of previous references in the
literature to the little known species Asplanchna Herrickii. It was
first figured by Herrick (84, Plate V.), and described in the explanation
of plates as “ flask-shaped Rotifer, hermaphrodite, with eges and sperm,"
but no further deseription was given and no name was proposed.
Herrick in a second paper (°85, p. 61) again mentions this form, but
adds nothing to the description.

In 1888, Jules de Guerne ('88) reproduced Herrick’s figure of the
jaws, and on the basis of this held that the species was new, and pro-
posed for it the name Asplanchna Herrickii.

Hudson and Gosse ('89) included Asplanchna Herrickii in their list
of * doubtful species."

Daday (792) admits Asplanchna Herrickii as a valid species.

Up to this time no observer since Herrick had reported finding this
species. But in this same year, 1892, Wierzejski (92) gave an account
of its presence in Galicia, with a description, and figures of the jaws
and the peculiar glandular organ which Herrick had mistaken for

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 89

a testis. Again, in his Rotatoria (Wrotki) Galicyi, Wierzejski (92°)
gives a figure of the entire animal, with special figures of the character-
istic glandular organ, the trocbal field, the jaws, and the excretory
system, together with a briof description in the Polish language.

Asplanchna Herrickii was afterward reported by the author (Jennings,
'04) as occurring in Lake St. Clair, and by Levander (795) as occurring
in the neighborhood of IIelsingfors.

2. Development.

The unsegmented egg of Asplanchna Herrickii is similar in form to
that of Callidina russeola, investigated by Zelinka, but slightly smaller,
the maximum dimensions in Callidina being 130 by 90 a, while the maxi-
mum dimensions observed in Asplanchna Herrickii were 97 u by 83.

No thick-shell * winter eggs" were ever observed by me in tbe
speeimens taken; possibly later in the fall these would have been
found.

In regard to the details of the developmental process, reference must
be made to Part First; the purpose here is merely to point out such
observations as are of importance from the standpoint of rotifer
morphology, noting especially any differences between my account and
those of other writers.

A. MATURATION.

As stated on page 13, the place of polar-cell formation in Asplanchna
has a different relation to the axes of the egg from ‘that ascribed to it by
Zelinka in Callidina. In the ensuing discussion I shall, for convenience
of comparison, use the orientation adopted by Zelinka ; that is, what I
have called the “macromere end” is anterior, the “micromere end”
posterior. That side of the egg upon which later the blastopore is
found, occupied in early stages chiefly by the quadrant D, is ventral, the
opposite side dorsal. These terms have no constant relation to the
terms of orientation employed in Part First.

As previously deseribed and figured (Plate 1, Figs. 1 and 2), the polar
cell is formed near one of the ends of the ellipsoidal egg, and the place of
formation is cut by the first and second cleavage furrows (Figs. 6 and 8).
The same is true for Asplanchna priodonta,

In these two species of Asplanchna, therefore, the place of polar-cell
formation is the same, with reference to the form of the egg, as that
described by Lameere (90) for Asplanchna Sieboldii, and by Zelinka
(91) for Callidina Leitgebii (p. 53) and Melicerta ringeus (p. 117). Iu

90 BULLETIN: MUSEUM OF COMPARATIVE ZOÜLOGY.
Callidina russeola and Discopus synaptie, according to Zelinka (91, p. 53),
the polar cell is formed almost exactly at one of the ends of the ellipsoidal
egg, though a very little to one side. This difference is a point of little
or no significance; an examination of Zelinka’s figures (Taf. I. Figg. 1-5)
shows that the polar cell in Callidina russeola occupies the same position
with regard to the axis of the first cleavage spindle as it does in the three
species of Asplanchna. Possibly the egg of Callidina russeola is forced
by the shell to take such a form that the axes of the egg, as indicated by
the first cleavage spindle, do not coincide with the apparent axes indi-
cated by the shape. The place of formation of the polar cell, as might
be expected, is correlated with the axis indicated by the cleavage spindle.
After the first cleavage a rotation occurs in Callidina russeola, bringing
the apparent axis into agreement with the real axis.

Dut with regard to the place of polar-cell formation in its relation to
the orientation of the egg as shown by later development, a remarkable
disagreement exists between the condition in Asplanchna Herrickii and
Asplanchna priodonta on the one hand, and the description given by
Zelinka of Callidina russeola on the other. The following is Zelinka’s
statement of the orientation of the egg of Callidina with relation to the
place of polar-cell formation : —

* Ws verdient hervorgehoben zu werden, dass von dem Augenblicke
an, als das Richtungskörperchen gebildet wird, sämmtliche Richtungen
im Räderthier-Eie orientirt sind. An dem Pol, in dessen Nähe das
Körperchen austritt, finden wir später das Vorderende, am gegenüber-
liegenden das Hinterende, während die Fläche, in der es erscheint,
zur Rückenfläche wird." (Zelinka, ’91, p. 54.)

Accepting the later orientation of Zelinka, the above statement be-
comes accurate for Asplanchna Herrickii and Asplanchna priodonta if
** Yorderende " and *Hinterende " are interchanged, and ** Bauchfläche ”
is substituted for “Rückenfläche”; in other words, the orientation
of Asplanchna with reference to the polar cell is precisely the opposite
of that of Callidina. The statement for Asplanchna would read: “ At
the pole in the neighborhood of which the polar cell appears, wo find
later the posterior end, at the opposite pole the anterior end, while the
surface on which it appears becomes the ventral surface.” This state-
ment, while correct if we relate the orientation of the animal simply to
the form of the egg, as is done by Zelinka, contains one false implication.
While that surface of the egg on which the polar cell is formed does later
become the ventral surface of the animal,— the same form being re-
tained to a late stage, — yet during the processes of development that

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIL 91

part of the material of the egg immediately surrounding the region of
polar-cell formation is moved to the posterior end of the egg, and, later
across this and even upon the dorsal side. The same is doubtless true
for the corresponding region (opposite the place of polar-cell formation,
according to Zelinka) in Callidina. It marks tho animal or upper pole
in Asplanchna, lying at the opposite end of the egg from the blastopore
during gastrulation, and is the common point of meeting for the blasto-
meres derived from the four quadrants of the egg. In Callidina russe-
ola, according to Zelinka, the region where the polar cell is formed lies,
not at the opposite end of the gastrula from tho blastopore, but at the
dorsal margin of the blastopore, and the cells of this region are lator
invaginated to form the fundament of the pharynx; the real animal
pole of the egg lying at a distance from the point of polar-cell forma-
tion. The whole of Zelinka’s general discussion of the early develop.
ment of the rotifer egg is based upon this peculiar position of the polar
cell. (See Zelinka, ’91, pp. 132-135.) His general statement of the
place of polar-cell formation is as follows: “ Das Richtungskörperchen
kommt an der dorsalen Seite des künftigen Embryo hervor, bei Melicerta
dem späteren hinteren Pole näher, bei Callidina fast am späteren vor-
deren Pole des in beiden Fällen länglichen Kies.”

The difference between our accounts is seen by comparing Figure 6
(Plate 1) and Figure 8 (Plate 2) with Zelinka’s Figures 8, 9, and 10
(Tafel J.). In the two-cell stage in Asplanchna (Fig. 6), when the
smaller cell, AB’, is turned away from the observer, who looks down
upon the polar cell, the spindle in the larger cell is seen to occupy such
& position that the smaller produet of the division of OD’ will lie to
the right, and in Figure 8 this condition is shown to be realized when the
division has taken place, the cell C? lying to the right of the polar cell.

3), derived from

In Zelinka’s figures, however, the small cell II
the division of the larger blastomere of the t»

left of the polar cell, when the same orientation is

of the two-

the smaller blastomere (A
observer. It therefore follows that, if the position of
dorsal in Callidina, it must be ventral in Asplan
orientation. Later stages show the same contrast. Thus Figures 13
and 14 (Plate 2), representations of the eight-cell stage of Asplanchna,
show the polar cell in the position already described, (the dorsal pole of
the egg being directed toward the observer,) whereas Zelinka’s Figures

15, 16, and 18 of Callidina show it at the opposite pole. Figures 19,

20 (Plate 3), 38, 41 (Plate 5), and 59 (Plate 7) show the polar cell at

92 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

later stages in Asplanchna, and demonstrate clearly that it lies at the
animal pole of the egg, opposite the blastopore.

Moreover, a careful examination of Zelinka’s own work shows that
the general statement above cited cannot be considered correct for all

totifera. His general conclusions are based throughout primarily upon
his more complete study of Callidina, and in his “ Theoretischer Theil "
he seems to have overlooked the fact that in Melicerta the place of
polar-cell formation observed by himself was, not only not near the
anterior end, as noted in the above general statement, but also not upon
the dorsal side.

The egg of Melicerta is described (Zelinka, '91, p. 115) as an elon-
gated ovoid with a sharper and a thicker or blunter end ; one side is
cylindrical in form, so as to appear straight in profile, while the other is
swollen, presenting in profile a convex outline. Zelinka says in regard
to the orientation of the egg given by Joliet: “Da ganz richtigerweise
das dickere End als das Kopfende, das dünnere als das Hinterende
bezeichnet wird, sowie ferner dieser Autor dessgleichen richtig die vorge-
bauchte Flüche als die ventrale, die cylindrische als die dorsale ansicht,
so würe die Orientirung durch diese Form des Kies eigentlich erleich-
tert," etc. .Now, in Zelinka's figures of the maturation stages of
Melicerta (Taf. V. Figg. 74-76) the polar cell is clearly shown to be
formed at this convex, and therefore ventral side. Furthermore, his
description of the process of maturation is as follows: * Das Keimbläs-
chen zeigt zuerst noch seine wohlbegrenzte sphärische Gestalt (Fig. 73)
und liegt etwa in der Mitte des Kies, sodann wandert es, indem es leb-
haft seine Gestalt verändert, gegen den hinteren Pol, wird zu einem
halbmondförmigen Fleck mit gekerbten Rändern (Fig. 74), dessen Kon-
vexität der Bauchflüche zugekehrt, welcher es sich rasch nähert. Knapp
unter der Oberflche zerlegen die Kerben den Kern in drei eng an
einander liegende ungleiche rundliche Stücke (Fig. 75), deren der Ober-
fläche zunächst liegendes aus dem Ei gepresst wird.“! (pp. 116, 117.)
The figures show clearly the convexity of the curved nucleus directed
toward the convex ventral surface of the egg, the gradual approximation
of the nucleus to that surface, and the formation there of the polar cell,
as described in the above passage. In the “ Erklärung der Abbildungen,”
these figures (73-77) are said to be “Rechte Seitenansichten," which
would make the convex surface and the place of polar-cell formation
dorsal, as required by Zelinka’s general statement. But this contradicts
the description just quoted, and there is other proof that the orientation

1 'The Italics are mine.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRIOKII. 93

of this first row of figures (73-80) of Melicerta is incorrectly given in
tho * Erklärung der Abbildungen." Thus, Figure 78 is said to be a
* Dorsalansicht,” which would bring the blastomere II (= Q9?) upon
the left side, But this blastomere is, in all other rotifers whose develop-
ment is known, formed on the right side, and Zelinka states concerning
this very figure, “ Zuerst folgt der Kerntheilung die Zelltheilung in dem
grossen Blastomer, die Theilungsebene steht, senkrecht zur Kernspindel
und schneidet ein Stück an der rechten Seite schief heraus (Fig. 78,
J). 1 (p. 117.) Again, Figure 79 is said to be a ventral view, whereas
tho same considerations as in tho last case show it to be a dorsal view.
Figure 80 is said to be a ventral view of the stage shown in Figure 78.
Yet it is evidently a later stage than Figure 78 ; it really represents a
ventral view of a stage similar to Figure 79, though the latter itself is
said in the “Erklärung” to be a ventral view. In the text, Zelinka
states correctly that Figure 79 is a dorsal, Figure 80 a ventral view
(p. 118) of the same stage (p. 117), contrary to the statements in
the “ Erklärung der Abbildungen.”

In Melicerta, therefore, the polar cell is formed in the same position
as in Asplanchna Herriekii and Asplanchna priodonta, and marks the
animal pole of the egg.

It seems very improbable that between these three forms and Calli-
dina there should actually be such a difference in regard to the place of
polar-cell formation as is brought out by the above comparisons. Ze-
linka’s account of Callidina is full and clear upon this point ; in both
his description and his figures the polar cell is traced to a late stage,
when a mistake in the orientation is impossible. There seems, however,
to be one opportunity for error. The polar cell in Callidina is not em-
bedded between the blastomeres, as in Asplanchna, but lies free upon the
surface of the egg. This is shown by Zelinka’s figures as well as by his
descriptions. On page 62 he says, “ Das Richtungskörperchen hat sei-
non Platz, den es früher eingenommen, verlassen und liegt nun ganz auf
den kleinen Zellen.” The egg or Callidina is enclosed by a rigid shell or
membrane, which is separated from the egg itself by a space, except at
the sides. ‘The shell retains constantly the ellipsoidal form, while the
egg itself may change its form and rotate within the shell. The original
place of polar-coll formation, with respect to this shell, is near one end, —
the end next to which Les later the anterior extremity of the embryo.
But immediately after the first cleavage the egg rotates within the sbell
through an are of about 90 degrees, and the region of polar-cell forma-

1 The Italics are mine.

94 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

tion is carried to the equator of the egg. The polar cell itself is shifted
at tho same time, and it seems possible that during this rapid rotation
it may be transferred to a region of the egg different from its original
position. Such shiftings of the polar cell, though to a less extent,
are mentioned by Zelinka (pp. 55, 58). During the “sehr schnell ver-
laufendes Phänomen” of the rotation of the egg within its shell, the
relatively different shifting of the polar cell might have been over-
looked. This is, however, only the suggestion of a possibility, for which
there is no direct evidence in Zelinka's work.

It is to be noted that in the above discussion I have employed
throughout the orientation of the egg used by Zelinka, and not that

adopted in my own account of the development.

D. CLEAVAGE.

The first cleavage in the two species of Asplanchna differs from that
of Callidina russoola in being exactly transverse to the long axis of the
egg. In the latter form, according to Zelinka, the first cleavage plane
is oblique to the spindle, and the spindle itself is oblique to the long
axis of tho egg. By a change of their relative positions immediately
after division, the two cells are later brought into the same position
relative to each other as in Asplanchna, and even at first the cleavage
plane in the two forms oceupies the same position relative to the place of
polar-cell formation. The difference thus is of little importance, except
that, from a cyto-mechanical standpoint, it shows that the form of the
ege does not determine the position of the first cleavage spindle. In
Eosphora also (Tessin, '86) the first cleavage plane is oblique to the
g, whereas in Melicerta ringens (Zelinka, '01) and
Asplanchna Sieboldii (Lameere, ’90) the first cleavage plane is transverse
to the long axis, as in Asplanchna Herrickii and Asplanchna priodonta.

The second and third cleavages in the two species of Asplanchna
(Plate 1, Fig. 6, Plate 2, Figs. 8-16) are essentially similar to the
corresponding cleavages of Callidina and of other rotifers in which
the development has been described. For convenience in compar-

ing the later stages, I give hero a table showing the correspond-

long axis of the eg

ence between the cells of Asplanchna in the eight-cell stage and those

of Callidina.

Asplanchna. Callidina (Zelinka, ’91). Asplanchna. Callidina (Zelinka, ’91).
aus Sed IPIS SUA dL O04 Kae ax b EAM.
GER Doserpkum. n niet Ug ct dins d er Il,
!!!.. ß „ CIEGO I

Qu ma „ .

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 95

Lamcere (90) and Leydig (54), have given figures of the four-cell
stage of Asplanchna Sieboldii, Beyond this stage there are no other
published figures of the cleavage of any species of Asplanchna.

In regard to the fourth eleavage (Figs. 19-30, Plates 3 and 4), a
romarkable difference is to be observed between tho cleavage of As-
planchna and that of Callidina as described by Zelinka. In Asplanchna
the cleavage takes place up to (and beyond) this point with the greatest
regularity both as to direction of spindles and as to sequence. The first
and second cleavages are meridional, the third equatorial, and the fourth
again equatorial. The sequence of cleavage is in every case DO A
(see nomenclature of cleavage, page 16). In Callidina, according to
Zelinka, the rhythm and regularity of the process is destroyed after the
third cleavage by the remarkable circumstance that the cell desi oye
linka) divides twice in succession before the fourth cleavage of any of
the other cells. After these two divisions of d, the six cells of the
other three quadrants are said to divide in the same succession that
occurs in Asplanchna, while it is not until all these eleavages aro finished
that the cell d+? (III, Zelinka) is separated into two blastomeres. Be-
fore this division of d*? takes place, the egg consists in Callidina, as in
Asplanchna, of four rows of four cells each, but in Callidina the method
of origin of tho four cells of quadrant D is stated to be different from
that in the other quadrants. In this quadrant tho three dorsal cells
(posterior, Zelinka) are said to arise by successive cleavages of the
large ventral (anterior, Zelinka) cell, while in each of the other three
quadrants the four cells arise by the halving of the two cells previously
present.

In Molicerta, according to Zelinka, the cleavage up to this point is as
in Asplanchna; the cell I (4*1) divides first, then II (d*?), then the
cells of the other three quadrants. Later the cleavage of Melicerta
differs from that of Asplanchna, but up to the end of the sixteen-cell
stage the processes are the same in the two.

In Fosphora, as described by Tessin (86), the cleavage is like
that of Asplanchna, except in the unessential particular that his cell
a’ (= d*?) divides before the cleavage of a (== dt). The sixteen-cell
stage is reached by the cleavage of the same cells as in the two species
of Asplanchna and in Melicerta.

In view of the regularity of the cleavage in these four forms, one
might be led to suppose that the irregularity deseribed in Callidina by
Zelinka was due to defective observation. Zelinka has noted this point

with particular attention, and states that he is certain of the difference

BULLEIIN: MUSEUM OF COMPARATIVE ZOOLOGY.

96

between Callidina and Eosphora, as described by Tessin. Nevertheless
it is possible that the slight time variation in the cleavage of Eosphora
may have misled Zelinka as to the point which needed especial care.
In Asplanchna Herrickii and Asplanchna priodonta, in Melicerta and in
Callidina, the large cell du divides first, followed immediately (in all
except Callidina, at least,) by the division of d*?. In Eosphora, by a
slight relative delay of the cleavage in % (a, Tessin), the cell % (a,
Tessin) divides first. Zelinka states that he has observed with especial
care that tho first cell (IV, Zelinka) given off in this quadrant takes
origin from d** (I, Zelinka). This is doubtless true; the important
point, however, is the origin of the next cell formed. Though this also
is stated to arise from the ventral cell of the series (I, Zelinka), it
seems possible that Zelinka was thrown off his guard by the supposed
greater care necessary for determining the exact method of the preced-
ing cleavage, and that the statement with regard to this one is really a
mistake. The cell IV in Figure 23 (Taf. II.) of Zelinka's work might
be the same cell as V in his Figure 24, while the cells called III and IV
in Figure 24 might have arisen by the division of the previously exist-
ing cell III (= 4*7) This would bring the conditions in Callidina
into agreement with those in Eosphora, Melicerta, and tho two species
of Asplanehna. This is, of course, a mere suggestion, which indeed is
rendered rather improbable by the nuclear conditions in the cells under
discussion shown in Zelinka's Figure 24. There can, of course, be no
question about the manner of division in Asplanchna, Figure 16 (Plate
2) shows the spindles in & and d^?, and the accomplished division of
d^! into d^! and d*? is shown in Figure 19 (Plate 3), while d*? still
contains a spindle. I have observed similar conditions in many other
specimens.

In view of the essential similarity of the process in Eosphora, as
described by Tessin, to that in Melicerta, as described by Zelinka, and
in the two species of Asplanchna, as observed by me, and in view of the
fact that the cleavage of the quadrant in question (D) in these four
forms may be said to agree completely with the general plan of cleavage
as exhibited in the other three quadrants, — while in Callidina the con-
ditions in this quadrant are anomalous, — the following remark of
Zelinka (91, p. 61) seems hardly justifiable: “ Da, wie später gezeigt
wird, auch Melieerta in der Entwicklung unserer vorliegenden Form
folgt, so muss der Vorgang bei Eosphora als eine bemerkenswerthe
Verschiedenheit aufgefasst werden.” As above shown, the difference

between Kosphora and the other forms consists merely in a slight

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 97

variation as to the relative time of cleavage of two blastomeres (& and
d*?) — 4a phenomenon which is exceedingly common both in different
individuals of the same species and in closely related species, and to
which little or no significance can be attached. It is Callidina which
shows a *bemerkenswerthe Verschiedenheit,” since here the rhythm
and regularity of the cleavage are completely destroyed, if the division
is correctly described by Zelinka.!

Beyond the fourth cleavage it becomes very difficult to compare the
processes in Asplanchna with those described by other observers for
other rotifers. As above deseribed, even in the fourth eleavage one of
the cells of the D quadrant (45.8) was formed in a different manner in
Callidina, according to Zeliuka, from the method in Asplanchna. This
in itself makes an exact comparison of the fifth cleavage in the two
species impossible. But, considering for convenience the cells corre-
sponding in position at the sixteen-cell stage as equivalent in the two
forms, we find the following to be the process in Callidina as described
by Zelinka.

The first blastomere of the sixteen-cell stage to divide in Callidina is
said to be the dorsal cell of the D quadrant, d** (ITI, Zelinka), whereas in
Asplanchna the very unequal division of the ventral cell d°* takes place
first (Fig. 33, Plate 4, and Figs. 35 and 38, Plate 5). The cleavage of
d5* is followed in Callidina by the division of d*? and d5? (IV and V.,
Zelinka). The division is meridional, as in Asplanchna (Fig. 37), but the
products are equal, whereas in Asplanchna they are unequal.

Now, according to Zelinka, the products (at the fifth cleavage) of the
division of d9* (d. and des, Zelinka’s IIT, and IIIa) are themselves
divided by meridional furrows. Thus the sixth cleavage takes place in
these cells before the fifth has been accomplished in any of the other
quadrants.

Following this, the large ventral cell of the D quadrant, d** (J,
Zelinka), divides equatorially, giving off on its dorsal side a small cell,
VI, which lies between the products of the division of d“.

Thus the cleavage of the quadrant D is much less regular than in
Asplanchna, where the ventral cells all clenve meridionally and unequally,
the dorsal cell equatorially and unequally, the direction of cleavage

1 Zelinka’s statement, quoted above, that Melicerta follows the same method of
division as Callidina, depends merely upon his interpretation of a real variation;
the actual divisions to form the sixteen-cell stage are the same in Melicerta as in
Asplanehna and Eosphora, and different from those of Callidina, as may be seen
by consulting Zelinka's ('91, p. 121) description.

VOL. XXX. — NO. 1.

98 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY,

being the same as in the other three quadrants (see Figs. 33 and 37).
In Asplanchna, moreover, the fifth cleavage in the other quadrants is
well advanced before any of the divisions for the sixth cleavage in
quadrant D have taken place.

Zelinka does not follow the cleavage in the other three quadrants cell
by cell. He states that the dorsal cells of these quadrants (all?) are
divided by planes parallel to the long axis of the egg, as is the case for
all but the fourth or most dorsal layer in Asplanchna, and that the
ventral cells a, 05", and c (47, bj, and II, Zelinka) are the last to
divide. In Asplanchna, as shown in Figures 39-44 (Plates 5 and 6),
all the cells of these quadrants divide meridionally except those of the
dorsal layer, which divide equatorially.

It is obviously impossible to compare in detail the cleavage in the two
forms at this time, or to reduce the condition described for Callidina to
the regular scheme of cleavage exhibited in Asplanchna. Certain facts
are perhaps worthy of notice, as showing the possibility that the cleavage
in Callidina is not so different from that of Asplanchna as would be
inferred from what is indicated above. The relation of the cells in
quadrant D are somewhat similar in Figures 30 and 34 of Zelinku's
work to a later condition in Asplanchna, — a condition reached, how-
ever, in a very different way, and shown in Figures 58 (Plate 7) and 66
(Plate 8). Further, the unequal cleavages of this quadrant are very
confusing, and easily overlooked. Zelinka's work was apparently done
almost entirely on living material, which does not lend itself as well as
does preserved material to precise orientation of the object, and to its

y as to permit views from all directions. It was
only by bringing together a complete series, in which the k:

rotation in such a

-yokinetic

process in was represented in various stages by several

3 3 able to determine absolutely the course of events

Specimens,

unequal division of the cell dë! is especially

did not observe it till the break in the rhythm

this point set me at work upon a minute study of a series

of cleavag:

of eggs separated in avage conditions by very short intervals only.
It i:
cleavage to occur, Zolinka figures (Taf. II. Fig. 31) a small vesicle lying

worth noting that in Callidina, shortly after the time for this
almost exactly in the place occupied by the minute vesicle given off at
this cleavage in Asplanchna, viz. between the ventral cells of quadrants
A and B (a, and b, Zelinl

J Zelinka considers this to be the polar
cell, although at a previous stage he had observed that the polar cell

had become displaced and now lay "arther dorsad, on the outer surface

JENNINGS: DEVELOPMENT OF ASPLANCHNA ITERRICKII. 99

of the smaller cells of tho egg. The condition shown in his Figure 31
he (91, p. 62) considers to be an exception, Later, he states that the
polar cell becomes surrounded by small spherules, indicating that it is
degenerating and falling to pieces. As showu in my Figures 51 and
52 (Plate G), the vesicle d*? produced by the division of d'en in
Asplanchna also becomes surrounded by large granules or spherules ;
but these are not of the nature assumed by Zelinka; they are the result
of a concentration of granules in the ventral cell dd, traceable from
the eight-cell stage onward. In view of these facts, it seems possible
that a similar division actually takes place in Callidina, and that the
small cell lying between a and 6, in Zelinka’s Figure 31 is the small
product of this cleavage, — the equivalent of d®? in Asplanchna.

Tessin, in his study of Eosphora, also failed to follow the cleavage
in detail to the 32-cell stage. He (86, p. 282) speaks of “ fortgesetzte
Aequatorialtheilungen " of the cells in the three smaller quadrants ; his
figures show the three quadrants composed each of a single row of six
cells (Figs. 22 and 23). This corresponds to the condition in Asplanchna
at a time when the dorsal cells (454—655) have divided equatorially, but
when the remainder of these three quadrants are as yet undivided.
Next, all the cells, except the large ventral cell d, are said to divide
meridionally. It is probable, therefore, that the formation of the minute
cell by the division of d** was overlooked, and that the cleavage is
essentially as in Asplanehna. The cells of quadrant D are said to
divide unequally at this cleavage, as is also the case in Asplanchna.

The sixth and later cleavages of the ectoderm have not been studied
in detail by other observers, so that a comparison of my results with
observations on other forms is impossible. Diagrams of the sixth,
seventh, and eighth elenvages for Asplanchna are given on pages 4 1, 46,
48, and 53, that for the eighth cleavage being incomplete.

The divisions of the entoderm cells have been followed somewhat further
by Tessin and Zelinka, so that for these a comparison may be made.

Nothing comparable to the unequal fifth and sixth cleavages of the
entoderm cell (forming the small vesicles d and d.), shown in Fig-
ures 38 (Plate 5) and 49 (Plate 6), have been reported by other
observers.

Later than these the cleavage of the entoderm in Eosphora (Tessin,
'86) takes place as in Asplanchna (Figs. 64, 65, and 76-83, Plates 8-10)
up to a stage comparable with that shown in Figures 77 and 78 (Plate 9),
except that no cleavage of the smaller dorsal cell d? corresponding
to the unequal division which I have shown in Figure 80 (Plate 10) was

100 BULLETIN: MUSEUM OF COMPARATIVE ZOÜLOGY.

observed by Tessin. IIe did not follow the divisions of the entoderm
further.

In Callidina a division takes place in the same manner as the seventh
in Asplanchna (Figs. 64 and 65, Plate 8), separating % and de. The
cleavage of de into det and dee also follows, as described above for
Asplanchna.

An unequal cleavage of d*?, as shown in Fig. 80 (Plate 10), was not
observed in Callidina. The cleavages which next ensue are described by
Zelinka as variable. The two cells corresponding to my d?! and d??
divide in the same direction as the corresponding cleavages of Asplanchna
(Fig. 76, Plate 9, and Fig. 81, Plate 10), but the dorsal cells de. and
die. are smaller than the others. The cell d (e, Zelinka) divides by
two successive divisions, at right angles to each other, into four cells;
one of these divisions corresponds to that indicated for this cell in Fig.
83 (Plate 10), while the other is at right angles to this. The order in
which these cleavages occur in Callidina is, however, variable.

According to Zelinka, each of the four cells corresponding to my die,
d??, d*?, and d now divides into three parts, but the details of these
cleavages are not given.

In Melicerta the cleavage of the entoderm is traced by Zelinka to a
four-cell condition, but the process is entirely different from that in
Asplanchna, Eosphora, and Callidina, so that it would not be of interest
to review the facts here.

The process of gastrulation takes place in Callidina and Eosphora, and
probably in all other Rotifera, in a manner essentially similar to that in
Asplanchna; the large ventral eell of the left posterior quadrant is
envoloped by the other cells during the process of cleavage, and becomes

the entoderm.

C. Summary ON MATURATION AND CLEAVAGE IN THE ROTIFERA.

In general, the following facts are shown for the early development
of Asplanchna, as compared with previous accounts of the development
of Rotifera.

1. The polar cell is formed at the animal pole of the egg, at a point
opposite that where the blastopore is later found, and not at the dorsal
or anterior margin of the blastoporie region, as stated by Zelinka for
Callidina and Melicerta.

9. A much greater regularity, and in a certain sense symmetry, are

shown in the direction and rate of cleavage than has been shown for

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIL 101

other Rotifera. Cell lineage is traced to a much later stage than has
been done for other rotifers.

3. In other respects the development of Asplanchna, so far as
observed, agrees in general with that of Callidina as described by
Zelinka (91). The development of organs was not traced in Asplanchna,
the purpose of the work lying chiefly in the domain of cyto-mechanies.

PART THIRD.— MATERIAL AND METHODS.

The material for the studies here presented was collected by means of
towings from Lake Michigan and certain small lakes connected with it,
in August and September of 1894. Such towings were killed and pre-
served by a variety of methods. For killing, the following reagents were
tried: (1) Flemming's stronger chrom-osmo-acetio mixture; (2) Klein-
enberg’s piero-sulphurio mixture, weaker solution; (3) Henneguy’s
fluid, consisting of Kleinenberg's weaker fluid plus 10% glacial acetic
acid; (4) piero-nitrie acid ; (5) alcoholic corrosive sublimate; and (6) a
mixture of corrosive sublimate and formalin. The best results were
gained by the use of Flemming's mixture. ‘The eggs were considerably
darkened, but this defect was easily corrected by bleaching with chlorine
generated from chlorate of potash and HCI. Henneguy’s fluid and
picro-nitric acid also gave good results. By alcoholic corrosive subli-
mate the eges were commonly shrunk, and with Kleinenberg’s fluid the
shrinking was excessive. The towings after killing were preserved partly
in 80% alcohol, and partly in a mixture of equal parts of glycerine,
alcohol, and water. Both these preservatives gave satisfactory results.

As is well known, the development of the embryo takes place in
Asplanchna within the body of the adult. The developing egg lies in
the posterior part of the body of the mother, enelosed in the greatly
distended oviduet or uterus, and with the ovary of the adult closely
applied to it. It was necessary to pick out the Asplanchnas one by one
from the quantities of Crustacea and other plankton with which they
were mingled. This was done by using capillary tubes. It was necessary,
moreover, to assort them with respect to the state of development of the
contained embryo, if an embryo were present. This is a process involy-
ing great labor, as, in order to determine even approximately the stage
of development of the embryo, it is necessary to examine the animal
with the compound microscope. The majority of the specimens contain
an embryo; not rarely two are present, in different stages of development,

and in a single case I observed three.

102 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

In order to study the eggs, it is of course necessary to dissect them
from the mother. This, again, is a tedious and delicate process, and it
is rarely possible to free the egg entirely from the closely applied ovi-
duct and ovary. This fact causes excessive trouble later, since the
fragments of oviduct and ovary attached to the egg prevent one's
plaeing it in any desired position or rolling it about at will, under à
cover-glass.

For surface study of the early stages the eggs were then mounted in
glycerine under a cover-glass supported by bits of capillary tubing thick
enough to allow free motion to be given to the object. It was found
impossible to use to advantage any stain for this study, because all the
stains tried colored the cytoplasm more than the nucleus, and made the
egg so opaque that cell boundaries and nuclei could not be distinguished.
For the stages from one to about sixteen cells, eggs fixed in Flem-
ming's solution are the best, as the slight darkening produced by this
reagent is of advantage in stages where the egg is cleft into but few
cells. For later stages this darkening is a disadvantage ; eggs killed by
other methods, or bleached after fixing with Flemming's fluid, must be
used.

The eggs were moved about by rolling the cover-glass on its rollers,
and drawings were made of different views thus obtained. It is here
that the ovary and bits of oviduct attached to the egg cause infinite
delay and vexation, in preventing the eggs from rolling easily or rest-
ing in any except certain positions. The time required for the work is
certainly doubled, perhaps more than doubled, by this.

With early stages, camera drawings can be made at once from the egg
after a favorable position is gained ; but after the egg has reached a
stage of about thirty cells, it is necessary first to roll the egg and make
many tentative free-hand drawings of the different surfaces, until
together they show the whole surface of the egg and the relation of
every cell to all surrounding cells. The egg is then oriented and a
camera figure made which shows the exact form of the cells in the
middle region of the upper surface, and the position of all the nuclei of
that surface. The cell boundaries about the periphery corresponding
to these nuclei are then supplied from the free-hand sketehes. "This
method, while not giving mechanical accuracy for the form of the cells
about the periphery of a late stage of cleavage, does permit of complete
accuracy 80 far as the relations of cells to one another are conoerned ;
any other method with eggs in which the cell boundaries are so
faintly marked on the surface is impracticable.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 108

Sections were made of numbers of eggs, but optical sections are much
more instructive, permitting exact orientation and revealing the struc-
ture fully as well as actual sections, so that most of the figures in which
sections are represented were made from optical sections.

In the deseriptive portiou of this paper, details have been given of
the movements of asters, nuclei, and other cell contents, as well as
of the cells as units. As the entire account was gained from a study
of preserved material, the question is a justifiable one, — Is there sufli-
cient evidence that the movements actually oceur as above described, or
are the stages figured and described merely chosen at will from a mass
of material and arranged arbitrarily in series?

The number of eggs used in determining the course of events in a
given cleavage has been stated in several cases in the text. Thus, 31
eggs were studied containing more than one and less than five cells;
42 containing more than seven and less than sixteen, etc. In all, more
than 250 eggs from Asplanchna Herrickii and 50 from Asplanchna prio-
donta, between the single cell stage (Fig. 1) and the stage containing
five entoderm cells (Fig. 83), were mounted in glycerine and studied.
Each egg, of course, came necessarily from a different individual, —
since, where two embryos were present in the same adult, one at least
had passed to a stage in the formation of organs. Of many of these
eggs examinations were made which may be called exhaustive; i. e.
every cell with its nuclear conditions was carefully figured. Thus,
from the egg of which Figure 68 (Plate 8) gives one view, at loast
twenty drawings were made, though but one is shown in the plates.
The figures given, therefore, represent by no means even a considerable
part of the evidence upon which the deseription is based. After de-
termination of the exact order of events, drawings of typical cases were
selected for illustrating the paper.

The determination of the sequence of the stages observed is greatly
lightened by the almost entire constancy in the relative order of events
in the different cells. Very slight variations occur in regard to certain
processes, as in the ease of the migration of the eloud of granules, as
mentioned in the Explanation of Plates, under Figure 51. But, in gen-
eral, a number of eggs representing a series of events in a given cell
show corresponding series of events in the other cells. It is not neces-
sary, therefore, to rely upon the conditions within the cell under ex-
amination for determination of the sequence of stages in this cell.
Even this would probably be possible, however, from the fact that the
nucleus in any cell increases in size steadily from the time the cell is

104 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

formed to the beginning of the transformation of the nucleus and asters
into the spindle figure — so that the relative size of the nucleus in the
same cell from different eggs gives a measure of its relative age.

An illustration will make clear the method used. Perhaps'the most
difficult problem was presented by the movements of the asters and
nuclei in the two large ventral cells of the quadrant D, at the fich and
sixth cleavages. These cells are d°- and dë? before the fifth cleavage ;
de and d*? after the fifth cleavage. Figure 31 (Plate 4) shows the
position of the asters in all the ceils of this quadrant in the resting
sixteen-cell stage. Figure 32 shows the nuclei in quadrant B in the
same egg, as seen in a longitudinal section. In the egg shown in Fig-
ure 33 the spindles are already completed in quadrant D, indicating
that this egg is older than that shown in Figures 31 and 32. Figure
35 (Plate 5), a ventral view of this stage, shows that the nuclei in a5.
and 0^! are much larger than the nuclei in the corresponding cells of
Figures 30 and 31, — which likewise indicates that this is an older stage
than is represented in the latter figures. Figure 35 shows the exact
antero-posterior position of the spindle in d*, while in the other quad-
rants the nuclei are shown to be still in a resting condition.

Figures 37 and 38 together show all the cells of this quadrant during
or just after division, the four dorsal cells being still connected by inter-
zonal filaments. It is therefore a stage later than that shown in Fig-
ures 33 and 35. This is in agreement with the larger size of the nuclei
in quadrant A, Figure 37, as compared with those of quadrant 4 in
Figure 33, and also with the presence of a spindle in 4. In this egg
the nucleus in d^! (Fig. 38) has resumed its nearly spherical form, and
at, the side of the cell where the anterior end of the spindle was located
(Fig. 35) is a small vesicle, d*? (Fig. 38). This represents one product
of the division.

In d^? at this stage (Figs. 37 and 38). the single aster is slightly
extended in a direction which is oblique to both the longer and shorter
axes of the cell.

The spindles in the cells of quadrants A, B, and C in Figures 39-43
(all from the same egg) prove that this is a later stage than that last
considered. In this egg the line joining the two asters in d°? (Fig. 42)
is oblique to the antero-posterior plane of the embryo, and nearly dorso-
ventral.

In the egg shown in Figures 43-46, division is completed in some of
the cells of quadrants A, B, and C, proving this to be a later stage than
those shown in Figures 37-42. In this egg the two asters in d^? (Fig.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRIOKII. 105

46) are at opposite sides of the nucleus, the line joining them being
parallel to the lateral axis of the embryo.

In the egg seen in Figure 53 (Plate 7), the fifth division is finished
in all the cells of the egg, (as shown by the study of the other quad-
rants, which are not figured,) so that this is a later stage than that
shown in Figures 43-46. The asters in des have taken an oblique
position.

In the egg of which Figures 48 (Plate 6) and 54 (Plate 7) are repre-
sentations, the nuclei in the recently formed cells of quadrants A, 2,
and O have enlarged, and the spindle is completely formed in d' (being
dorso-ventral in position). Both of these facts show that the egg is older
than the one shown in Figure 53. In this egg (Fig. 48) we find that a
spindle is present in the entoderm cell d**, occupying nearly the position
foreshadowed by the position of the asters in Figure. 42, and almost
exactly the same position as the spindle at the foregoing division (Fig. 35).

The egg seen in Figure 49 is still older, as shown by the presence
of spindles in des and d^? (seen endwise), and the advanced condition of
cleavage in d, Here de is just dividing, forming d"? and the second
small cell, d“.

Figure 50 is older than Figure 49, since d*5, dd, and d*5 have divided.
In this egg we find that d*? has been separated into two cells, ds aud
due, and there is a second small vesicle, die, in the position where it
was seen in tho process of formation in Figure 49.

The above is sufficient to illustrate the method of work ; the rest of
the account might be analyzed in the same way. It is important to
remember, however, that the deseription is not based merely upon the
cases figured. Thus, for the processes just analyzed, more than thirty
eggs, showing various phases of the changes occurring, were studied,
while only eleven different eggs aro represented in the figures of these
stages.

The foregoing work was done in the winters of 1894-95 and 1895-96,
in the Zoölogieal Laboratory of the Museum of Comparative Zoology at
Harvard University. It gives me pleasure to acknowledge my grent
indebtedness to the Director of the Laboratory, Professor E. L. Mark,
for advice and assistance which have been of the greatest value to me

throughout my work.

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

GENERAL SUMMARY.

A. Observations.

l. Many divisions take place during the cleavage of Asplanchna in
which the spindle lies in the shortest axis of the cell, in the direction of
greatest. pressure, and the ensuing division results in the production of
contact surfaces of greatest area.

2. In the cleavage of the ectoderm of Asplanchna any cell of any one
quadrant cleaves in the same direction as the corresponding cell of the
other quadrants, though the forms of the corresponding cells may vary
excessively. Conversely, cells of the same form and with similar relations
to surrounding cells, but belonging to different layers or series, may
divide with spindles in exactly opposite directions.

3. The entodermal cell follows the same rhythm and direction of
cleavage as the other cells, so long as it remains on the exterior and thus
corresponds in position with other cells of the ege. When it becomes
enveloped by the other cells, so as to come into different relations with
the axis of the embryo, its plan of cleavage changes, showing no definite
relation to that of the ectoderm.

4. All the cleavages in the ectoderm are to a late period either
equatorial or meridional, so that the position of any given spindle is
either parallel or perpendicular to that of the preceding spindle.

5. There is no regular alternation in the direction of spindles. Equa-
torial cleavages may follow successively for three or more generations,
and the same is true of meridional cleavages.

6. The position occupied by the two asters after they have passed to
opposite sides of the nucleus does not indicate the direction of the
ensuing spindle. "lhis may oceupy the position indicated by the asters,
or the definitive position may be gained by a rotation of the asters and
nucleus at the passage into the karyokinetie condition.

7. There is no “regular angle of rotation ” (Heidenhain) in a mechani-
cal sense, since (a) in cells of different layers, in one case the augle may
be zero, in the other ease 90 degrees; and (b) even in cells where tho
direction of the previous spindle and the direction of the following spindle
are the same, the asters may move in an entirely different manner. In
one cell the rotation may be directly through an angle of 90 degrees,
and in a single plane, while in another there may be complex movements
and rotation successively in different planos.

|
|
|

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 107

8. The position and movements of the asters in the resting stage seem
partly determined by the form of the cell.

9. The rotation of the nucleus and asters into the definitive position
at the time of karyokinesis often takes place from the longer into the
shorter axis of the cell, and apparently from the direction of least pressure
into the direction of greatest pressure.

10. The form of the cells in many cases does not conform to the law
of minimal surfaces, being (a) changeable, and (b) even in the resting
stage widely at variance with the conditions required by the law.

11. Many of the cleavages are unequal, sometimes extremely so, but
the inequality shows no significant relation to accumulations of yolk
material. (See 16.)

12. The sequence of cleavage is (within very narrow limits of varia-
tion) constant, and shows no relation to accumulations of yolk, There
is a general tendency for larger cells to divide faster, but not all the
facts regarding the succession of cleavages show relation to the com-
parative size of the cells.

13. In the resting stage the cells seem to be passive, taking what-
ever form is impressed upon them by the surrounding cells, As the
cell passes into the karyokinetio condition it becomes more rounded,
the cytoplasm tends to group itself symmetrically about the spindle, and
the cell elongates in the direction of the spindle.

14. The spindle generally (not always) lies in either the longest or
the shortest axis of the cell, as maintained by Roux. But apparently
this is due in Asplanchna to the fact that the cytoplasm tends to group
itself symmetrically about the spindle.

15. A change of the relation of a cell to the axes of the egg, as by a
displacement due to the other cells, results in a change of the position of
the spindle with reference to the axes of the egg.

16. During cleavage a cloud of granules is segregated in a portion of
the cell which is to form the entoderm ; this mass passes from thé an-
terior and ventral side of the entoderm cell to its posterior and dorsal
side, and is there separated off at the seventh cleavage into the smaller
entodermal cell.

17. The egg retains its ellipsoidal form throughout all the processes
of development, up to a late stage, though as oleavage progresses tho
blastomeres shift extensively their positions with relation to this form.
This retention of the ellipsoidal form by the egg cannot be referred to
any simple mechanical factor. (See pages 81, 82.)

18. Gastrulation accompanies cleavage, and advances step by step

108 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.
with the withdrawal of the deep parts of the peripheral cells and their
dorso-ventral extension, consequent upon frequent equatorial divisions.

19. As to facts bearing upon the special morphology of the Rotifera, —

(a) The polar cell is formed at the animal pole of the egg, at the
point opposite that where the blastopore is later found, and not at
the dorsal (or anterior) margin of the future blastoporie region, as
stated by Zelinka (91) for Callidina.

(b) The cleavage of Asplanchna was traced to a later stage than has
been done for other rotifers. A much greater regularity, and in a cer-
tain sense symmetry, are shown in the direction aud rate of cleavage
than has been shown for other species.

B. Conclusions.

20. It results from 1, 2, 3, 5, 7, and 9 that the direction of cleavage
is not determined by any simple mechanical factors or relations of form.
Specifically, the course of cleavage in Asplanchna is inconsistent with
any general validity of (1) Hertwig's law of the spindles in the longest
axis of the protoplasmic mass, (2) Berthold’s law of least areas, and
(3) Braem’s and Pflüger’s principle of least resistance.

21. It results from 11 that no simple factor can account for the
equality or inequality of the cleavage. Specifically, the conditions
in Asplanchna are inconsistent (a) with Hertwig's view that the
dividing nucleus takes a position *in the middle of its sphere of
action," so far as that expression has any definite significance, and
(b) with Braem’s principle of “like resistance " at the two ends of the
spindle.

22. It results fron 12, as well as from a comparison with the cleavage
of many other invertebrates, that no simple factor, such as greater or
less quantity of yolk, will account for the sequence of cleavage.

23. It is a natural conclusion from 15 and the latter part of 14, that
the direction of the spindle is not due to an influence in the egg as!
whole, connected with its axial relations, but is determined within each
cell itself. However, I do not consider this conclusion at all well
established.

24. It results from 5, 6, 7, 8, and 9 that the problem as to what de-
termines the position of the spindle is resolvable into several: (a) What
determines the direction of separation of the newly formed asters?
(b) What determines the position of the asters during the resting stage

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIL 109

of the cytoplasm? (e) What determines the rotation of the asters and
nucleus as the cell passes into the karyokinetic condition ?
EH

25. It may be concluded from 20, 21, 22, and 24 that the final posi-
tion of the spindle and manner of cleavage are causally determined

by processes — of an unknown character taking place within the
protoplasm.

26. The definite relation of the position of the spindle to external

conditions observed in some cases — such as to the form of the cell, the
direction of pressure (), and the direction of the incoming rays of light —
is to be interpreted as a reaetion to stimulus, dependent in every case
upon the specific structure of the protoplasm, and variable with that
Structure.
27. The manner of division is related to the purpose to be attained
by the given division, and to the general morphogenetic changes in the
organism. In Asplanchna the method of cleavage is adapted to bring-
ing about gastrulation,

28. It follows from 16 that cleavage is not merely a quantitative
division into similar units; it is accompanied by other developmental
processes, some of which are distinctly traceable.

29. Gastrulation in Asplanchna is not a process distinct from cleay-
age, but is an accompaniment and a result of cleavage. The process
of which it forms a part begins at the third cleavage and is not finished
until much later than what is commonly spoken of as gastrulation
proper.

30. Gastrulation in Asplanchna may be analyzed into several factors.

(a) The form of the egg, or the influences determining it.
(b) The direction of cleavage.

(c) The inequality of cleavage.

(d) The sequence of cleavage (öh.

(e) The changes in form taking place as the cells divide.

1 Tt may be well to state expressly that I do not consider the above as in any
sense a general explanation of the process of gastrulation. My aim has been to
give as nearly as possible a correct account, from the standpoint of developmental
mechanics alone, of. the facts in regard to the early development of a single form.
The origin of the process of gastrulation in phylogeny is not touched by this
account. It is a common.phenomenon in the organie world, that the same end is
accomplished by different means in different cases; doubtless in many forms gas-
trulation is brought about in a way that bears no resemblance to the process in
Asplanchna. In general, the whole question of the origin of processes to which
an end or purpose can be assigned lies entirely without the fleld of the present
paper.

110 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

AN of these must, according to 17, 25, and 26, be considered as
determined by the unknown (molecular?) structure and activities of
the protoplasm.

31. It follows from 30 that the early development of Asplanchna, to
a stage somewhat beyond gastrulation, may be analyzed into two fac-
tors: (1) the influences determining and preserving the form of the
egg as a whole, and (2) processes occurring in consequence of the spe-
cific (molecular ?) structure and activities of the protoplasm,

Both of these factors, which perhaps should be considered as different
manifestations of one, are from a causal-mechanical standpoint, entirely
unknown. “Damit werden die causalen Bedingungen der Entwicklung
vorzugsweise in das Moleculargeschehen verlegt und entziehen sich
vorderhand grossentheils unserer weiteren Erforschung.” (Roux, '85*,

p. 427.)

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIL. 111

LITERATURE CITED.

Auerbach, L.
74. Organologische Studien. Breslau. 262 pp., 4 Taf.

Balfour, F. M.
80. A ‘Treatise on Comparative Embryology. Vol. I. London. xi + 492
+ xxii pp.
Bergh, R. S.
95. Vorlesungen über allgemeine Embryologie. Wiesbaden. 289 pp.

Berthold, G.
'86, Studien über Protoplasmamechanik. Leipzig. 339 pp., 7 Taf.
Born, G.

'93. Ueber Druckversuche an Froscheiern. Anat. Anzeiger, Bd. VIII. pp.

609-627, 10 Abbild.
Born, G.

94. Ueber neue Compressionsversuche an Froscheiern. 79'* Jahresb. Schlesi-

schen Gesell. für vaterl. Cultur, Zool.-bot. Section, pp. 47-56.
Boveri, T.

'94. Ueber Beziehungen zwischen Zellfunktion und Kernstructur. Sitz-

ungsb. Phys.-med. Gesell. Würzburg, Jahrg. 1894, pp. 94, 95.
Braem, F.

'94. Ueber den Einfluss des Druckes auf die Zellteilung und über die Bedeu-
tung dieses Binflusses für die normale Eifurchung. Biol. Centralbl., Bd.
XIV. pp. 340-353, 5 Figg.

Castle, W. E.

96. The Early Embryology of Ciona intestinalis, Flemming (L.). Bull. Mus.
Comp. Zoól. Harvard College, Vol. XXVII. No. 7, pp. 201-280, Plates
I-XII.

Chun, C.

'92. Die Dissogonie, eine neue Form der geschlechtlichen Zeugung. Festschr.
siebenzigster Geburtstage Rudolf Leuckarts, pp. 77-108, Taf. IX -XIII.
und 3 Figg.

132 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Conklin, E. G.
'92. The Cleavage of the Ovum in Crepidula fornieata. Zool. Anzeiger,
Jahrg. XV. pp. 185-188.
Crampton, H. E.
96. Experimental Studies on Gasteropod Development. Arch. f. Entw.-
mech., Bd. ILI. pp. 1-19, Taf. I.-IV.
Daday, E. v.
92. Revision der Asplanchna-arten und die ungarländischen Repräsentanten.
Math. u. naturw. Berichte aus Ungarn, Bd. IX. pp. 69-89, Taf. II., III.
Davenport, C. B.
'95. Studies in Morphogenesis. IV. A Preliminary Catalogue of the Pro-

cesses concerned in Ontogeny. Bull. Mus. Comp. Zoöl. Harvard College,
Vol. XXVII. No. 6, pp. 171-199, 31 Figs.
Driesch, H.

'92. Jintwicklungsmechanische Studien. IV. Experimentelle Veränderungen
des Typus der Furehung und ihre Folgen. Zeitschr, f. wiss. Zool., Bd.
LV. pp. 10-29, Taf. I-III.

Driesch, H.
93. Entwicklungsmechanische Studien. VIII. Ueber Variation der Mikro-
merenbildung. Mitth. Zool. Stat. Neapel, Bd. XI. pp. 226-232, Taf. XI.
Driesch, H.
'94. Analytische Theorie der organischen Entwicklung. Leipzig. 184 pp.

Driesch, H., und Morgan, T. H.

'95. Zur Analysis der ersten Entwickelungsstadien des Ctenophoreneies.
I. Von der Entwickelung einzelner Ctenophorenblastomeren. Arch. f.
Entw.-mech., Bd. II. pp. 204-215, Taf. XVI.

Eycleshymer, A. C.

'95. The Barly Development of Amblystoma, with Observations on some other

Vertebrates. Journ. Morph., Vol. X. pp. 343-418, Plates XVIII.-XXII.
Gorham, F. P.

'95. The Cleavage of the Egg of Virbius zostericola, Smith. Journ. Morph.,

Vol. XI. pp. 741-746, Plate XXXVII.
Guerne, J. de

88. Note monographique sur les Rotifères de la Famille des Asplanchnide.
Excursions zoologiques dans les fles de Fayal et de San Miguel (Açores).
Paris, 1888. pp. 50-65, 5 Figs. à

Ibid., Ann. and Mag. Nat. Hist., Ser. 6, Vol. II. pp. 28-40, 5 Pigs.
(Cited from the translation.)
Heidenhain, M.
94. Neue Untersuchungen über die Centralkörper und ihre Beziehungen zum
Kern- und Zellprotoplasma. Arch. f. mikr. Anat., Bd. XLIII. pp. 473-758,
Taf. XXV.-XXXI.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKII. 118

Heidenhain, M.

940 “Discussion” of Zieglers Paper on “Furchung unter Pressung.“

Verhandl. Anat. Gesellsch., Ste Versamml., p. 154.
Heidenhain, M.

95. Cytomechanische Studien. Arch. f. Entw.-mech., Bd. I. pp. 473-577»

Taf. XX. und 17 Figg.
Herrick, C. L.

'84. Final Report on the Crustacea of Minnesota. Twelfth Annual Report

of the Geol. and Nat. Hist. Survey of Minnesota. 191 pp., Plates A to V.
Herrick, C. L.

'85, Rotifers of America, Part I. Bull. Seient. Lab., Denison University,

Vol. I. pp. 43-62, Plates II.-IV. and X.
Herbst, C.

'94. Ueber die Bedeutung der Reizphysiologie für die kausale Auffassung von
Vorgängen in der tierischen Ontogenese. Biol. Centralb., Bd. XIV. pp.
657-666, 689-697, 727—744, 753-771, 800-810.

Hertwig, O.
93. Die Zelle und die Gewebe. Jena. 296 pp., 168 Abbild.
Hertwig, O.

'93* Ueber den Werth der ersten Furchungszellen für die Organbildung des
Embryo. Arch. f. mikr. Anat., Bd. XLIL pp. 663-807, Taf. XXXIX.-
XLIV.

Hertwig, O.

'96. Lehrbuch der Entwicklungsgeschichte des Menschen und der Wirbel-

thiere. Fünfte Auflage. Jena. 612 pp., 2 Taf. und 384 Abbild.
Heymons, R.
'93. Zur Eutwieklungsgeschichte von Umbrella mediterranea, Lam. Zeitschr.
f. wiss. Zool., Bd. LVI. pp. 245-298, Taf. XIV.-XVI.
Hudson, C. T., and Gosse, P. H.
86. The Rotifera or Wheel Animaleules. London. 2 vols., 30 col. Plates.
Hudson, C. T., and Gosse, P. H.

89. The Rotifera or Wheel Animaleules. Supplement. London. vii + 64 pp.,

4 Plates.
Jennings, H. S.

94. Rotatoria of the Great Lakes and of some of the Inland Lakes of

Michigan. Bull. Michigan Fish Commission, No. 3. 34 pp., 1 Plate.
Joliet, L.
83. Monographie des Mélicertes. Arch. de Zool. expér. et génér., sér. 2,
Tom. I. pp. 131-224, Pl. XI.-XIII.
Kofoid, C. A.
'94. On Some Laws of Cleavage in Limax. A Preliminary Notice. Proc.
Amer. Acad. Arts and Sci, Vol. XXIX. pp. 180-203, Pl. I. and II.

VOL. XXX. — NO. 1. 8

114 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Kofoid, C. A.
95. On the Early Development of Limax. Bull. Mus. Comp. Zool. Harvard
College, Vol. XX VIL. No. 2, pp. 34-118, Plates I.-VIIL.
Lameere, A.
90. A propos de la Maturation de l'euf parthénogénétique. Bruxelles.
88 pp., 3 Planches.
Levander, K. M.
95. Materialien zur Kenntniss der Wasserfauna in der Umgebung von Helsing-
fors, mit besonderer Berücksichtigung der Meeresfauna. II. Rotatoria.
Acta Societ. pro Fauna et Flora Fennica, Vol. XII. Nr. 3, 72 pp., 3 Taf.
Leydig, F.
'54. Ueber den Bau und die systematische Stellung der Räderthiere. Zeitschr.
f. wiss. Zool., Bd. VI. pp. 1-120, Taf. L.-IV.

Lillie, F. R.
95. The Embryology of the Unionide, Journ. Morph., Vol. X. pp. 1-100,
Pl. L-VI.

McMurrich, J. P.
95. Cell Division and Development. Biological Lectures delivered at the
Marine Biological Laboratory of. Wood's Holl in the Summer Session of
1894. Boston. pp. 125-147.
Meyer, O.
95. Celluläre Untersuchungen an Nematoden-Eiern. Jena. Zeitschr., Bd.
XXIX. pp. 991-410, Taf. X., XI.
Mead, A. D.
94. Preliminary Account of the Cell-Lineage of Amphitrite and other
Annelids. Journ. Morph., Vol. IX. pp. 465-472, 8 Pigs.
Morgan, T. H.
94. Experimental Studies on Echinoderm Eggs. Anat. Anzeiger, Bd. IX.
pp. 141-152, 4 Figs.
Morgan, T. H.
95. A Study of a Variation in Cleavage. Arch. f. Untw.-mech., Bd. II. pp.
72-80, Taf. X.
Pflüger, E.
84. Ueber die Einwirkung der Schwerkraft und anderer Bedingungen auf
die Richtung der Zelltheilung. Arch. f. d. ges. Physiol., Bd. XXXIV.
pp. 607—616.
Rauber, A.
’83. Neue Grundlegungen zur Kenntnis der Zelle. Morph. Jahrb., Bd. VIII.
pp. 233-338, Taf. XI.-XIV.
Roux, W.
'85. Ueber die Bestimmung der Hauptriehtungen des Froschembryo im Ei
und über die erste Theilung des Froscheies. Breslauer ärztliche Zeitschr.,
1885, Nr. 6-9. Also, Gesammelte Abhandlungen, Bd. II. pp. 277-343.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRICKIL 115

Roux, W.

'85* Einleitung zu den Beiträgen zur Entwicklungsmechanik des Embryo.
Zeitschr. f. Biologie, Bd. XXI. pp. 411-428. Also, Gesammelte Abhandl.,
Bd. II. pp. 1-23.

Roux, W.

94. “Discussion” of Ziegler’s Paper on “Furehung unter Pressung,“

Verhandl. Anat. Gesellsch., Ste Versamml., p. 151.
Roux, W.

’94% Die Methoden zur Hervorbringung halben Froschembryonen und zum
Nachweis der Beziehungen der ersten Furehungsebenen zur Medianebene
des Embryo. Anat. Anzeiger, Bd. IX. pp. 248-282. Aiso, Gesammelte

Abhandl., Bd. II. pp. 940- 986.
Roux, W.

95. Ueber den “ Cytotropismus” der Furchungszellen des Grasfrosches
(Rana fusca). Arch. f. Entw.-mech. Bd. I. pp. 43-68, 161-202, Taf.
L-IIL, 3 Figg.

Ryder, J. A.

'93. The Growth of Euglena viridis when confined principally to two Dimen-
sions of Space. Contrib. Zoól. Lab. Univ. Penn., Vol. I. No. 1, pp. 37-
50, Plate II.

Sachs, J.

78. Ueber die Anordnung der Zellen 2 jüngsten Pflanzentheilen. Arb. d.
bot. Inst. Würzburg, Bd. IT. Heft 1. 4/so, Gesammelte Abhandl. über
Pflanzenphysiol, Bd. II. pp. rege Taf. IX., X. (Citations from
the latter.)

Salensky, W.

72. Beiträge zur Entwicklungsges aye te des Brachionus ureeolaris. Zeit-

schr. f. wiss, Zool., Bd. XXII. pp. 455-466, Taf. XXXVII

Einfluss des Beleuchtungsrichti auf die Theilung der Equisetum-

7

sporen. Ber. deutsch. bot. Gesellsch., Bd. III.

Strassen, O. zur.
95. ee Beobachtungen an Ascaris. Verhandl. deutsch.
zool. Gesellsch., Juni, 1895, pp. 83-95, 6 Figg.
Strassen, O. zur.
'96. limbryonalentwick Soe der Ascaris megalocephala, Arch. f. Entw.-
27-105, Taf. V.-IX.

mech., Bd. IIT. pp.
'Tessin, G.
'86. Ueber Eibildung und Entwicklung der Rotatorien. Zeitschr. f. wiss.
Zool, Bd. XLIV. pp. 273-302, Taf, XIX., XX.
Watase, S.
'91. Studies on Cephalopods. I. Cleavage of the Ovum. Journ. Morph.,
Vol. IV. pp. 247-302, Plates IX.--XIL.

116 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Weismann, A., und Ischikawa, C.
'87. Ueber die Bildung der Richtungskörper bei thierischen Eiern. Ber. d.

Naturf. Gesellsch. Freiburg i. Br., Bd. III., Heft 1, pp. 1-44, Taf. I.-IV.
Wheeler, W. M.
'95. The Behavior ot the Centrosomes in the fertilized Egg of Myzostoma
glabrum, Leuckart. Journ. Morph., Vol. X. pp. 305-311, 10 Figs.
Wierzejski, A.
92. Zur Kenntniss der Asplanchna-Arten. Zool. Anzeiger, Jahrg. XV.
pp. 345-349, 2 Figg.
Wierzejski, A.
'92*. Rotatoria (Wrotki) Galicyi, Akad. d. wiss. in Krakau, pp. 1-106,
Tab. IV.-VI.
Wilson, E. B.
'92. The Cell-Lineage of Nereis. A Contribution to the Cytogeny of tlie
Aunelid Body. Journ. Morph., Vol. VI. pp. 361-480, Plates XIIL-XX.

Wilson, E. B.
93. Amphioxus and the Mosaic Theory of Development. Journ. Morph.,

Vol. VIII. pp. 579-630, Plates XXIX.-XXXVIII.

Zacharias, O.
'84. Ueber Fortpflanzung und Entwicklung von Rotifer vulgaris. Zeitschr.

f. wiss. Zool, Bd. XLI. pp. 226-251, Taf. XVI.

Zelinka, C.
91. Studien über Räderthiere. III. Zur Entwicklungsgeschichte der Räder-

thiere nebst Bemerkungen über ihre Anatomie und Physiologie. Zeitschr.
f. wiss. Zool, Bd. LIII. pp. 1-159, Taf. I. VI. Also, Arb. a. d. zool.
Inst. Graz, Bd. IV. pp. 323-481, Taf. I. VI. (Citations from the former.)

Ziegler, H. E.
94. Ueber Furchung unter Pressung. Verhandl. Anat. Gesellsch., Ste

Versamml., pp. 132-146, 13 Abbild. — “ Discussion,” pp. 151-156.

Ziegler, H. E.
'95, "Untersuchungen über die erste
Zugleich ein Beitrag zur Zellenlehre.

pp. 351-410, Taf. XVIL-XIX.

n Eutwicklungsvorgänge der Nematoden.
Zeitschr. f. wiss. Zool, Bd. LX.

Zimmermann, A.
'93, Ueber die mechanischen Erklärungsversuche der Gestalt und Anordnung

der Zellmembranen. Beiträge zur Morphologie und Physiologie der Pflan-

zenzelle, Bd. I., Heft IL. 8, pp. 159-181. Tübingen.

JENNINGS: DEVELOPMENT OF ASPLANCHNA HERRIOKII 117

EXPLANATION OF PLATES.

All the figures represent preparations of the eggs of Asplanchna Herrickii de
Guerne, except Figure 29, Plate 4, which represents the egg of Asplanchna prio-
donta Gosse. All were drawn, with the aid of the Abbe camera lucida, to a mag-
nifieation of 525 diameters.

The four blastomeres of the four-cell stage are distinguished by different colors,
and the same color is retained throughout for all the cells (constituting a “ quad-
rant”) derived from each of the four blastomeres thus distinguished, The quad-
rant A is blue; B, orange; C, yellow; and D, red.

The promihent granulations in the ventral portion of quadrant D are figured ;
but all other granulations are omitted, except in the case of Figure 7, where the
general granulation is also represented, though somewhat diagrammatically.

For an explanation of the system of nomenclature used in lettering the cells see
page 16. In somecases for want of room the letters have been omitted, the expo-
nents only being expressed. In such cases the color indicates to which quadrant
the cell belongs.

In all the figures where it is possible, the animal pole of the egg is above, the
ventral pole below. In views of the dorsal or ventral poles of the egg, the anterior
end is above. Unless otherwise stated, figures represent surface views of more or
less transparent eggs; these are shaded, whereas seetions — optical or aetual — are
not shaded.

ABBREVIATIONS.

„%%% Blastopore. cl. pol. Polar cell. pol. anm, Animal pole.

Fig.

Fig.

Fig.

Fig.

JzNNINGS, — Asplanchna,

T

PLATE 1.

Egg showing the maturation spindle.

Egg slightly older than that shown in Figure 1; the polar cell is formed
and is embedded in the egg; the cleavage nucleus, preceded by the
deep aster of the maturation spindle, is moving from the place of
polar-cell formation toward the interior of the egg.

First cleavage spindle, early stage.

Longitudinal section (actual). The two nuclei formed at the first cleav-
age are separating; the notch and the granules at the periphery
indicate the beginning of the formation of the first cleavage plane.
The aster in the anterior (upper) end of the egg has divided, while
the opposite aster is still entire.

Two-cell stage seen from the dorsal side; the anterior end above.
There are two asters in the cell AB", while the aster in CLD” is
still undivided.

Two-cell stage, from the dorsal side; spindles in both blastomeres.

View of an eight-cell stage; optical section through the quadrants B and
D, showing the distribution of the yolk granules. Observe the con-
centration of granules in the ventral part of d*!, and the position of

the asters in dt! and d*?,

ig. 14.

JENNINGS. — Asplanchna,

PLATE 2.

Four-cell stage, viewed as a transparent object from the animal pole.
The anterior end is above,

Four-cell stage, later than Figure 8, seen from the right side. Spindles
in C? and D’; spherical nucleus in 33,

Four-cell stage, optical section through the cells B? and 13, showing the
division of the asters in preparation for the third cleavage.

Five-cell stage, from the ventral side. The anterior end is above. The
aster in dt1-has divided laterally, the two parts being still connected.
Spindles in 43, 53, and C8 seen endwise.

Optical section, approximately sagittal, through the quadrants B and D,
from the egg shown in Figure 11.

Oblique view of the eight-cell stage, shortly after the third cleavage.
The animal pole is marked by the polar cell; the anterior median
line by the boundary between quadrants A and B.

Dorsal view of the egg represented in Figure 13, showing the oblique
position of the asters in d*?, and the lateral extension of the aster
in c#2,

Ventral view of the egg seen in Figures 13 and 14, showing the oblique
position of the asters in d*!, and the lateral extension of the aster
in c*1,

Optical, nearly sagittal, section of the eight-cell stage, through the quad-
rants B and D, showing the dorso-ventral direction of the spindles in

d^! and d. Notice also the change of form of the quadrant B, as

compared with the same quadrant in Figures 10 and 12.

p

hee

Jenninas, — Asplanchna,

PLATE 3.

Fig. 17. Right side of the eight-cell stage, same egg as that seen in Figures 13, 14,
and 15, showing the lateral extension and beginning of division of the
asters in c* and c#2,

Fig.18. Right side of a slightly older stage than Figure 17, showing the com-
pletion of the division of the asters. In c*1 the line joining the asters
is lateral; in ct? it has already become oblique.

Fig. 19. Right side of a stage later than that shown in Figure 18, containing nine
cells. The line joining the asters in c has become dorso-ventral,
while that joining those of c*! has become oblique. The cleavage of
du into d^! and d^? has occurred (compare Fig. 16), while d*? is
still undivided,

Figs. 20-25. Different views of an egg in the ten-cell stage, to show the position of
the spindles in relation to the exact form and dimensions of the cells.

Fig. 20. Right side of a stage later than Figure 19, but containing still only ten
cells. The line joining the asters in ct! has become dorso-ventral and
the spindle is formed between them. Likewise in c.; di has divided
into dè? and d54,

Fig. 21. Anterior surface of the egg represented in Figure 20; the spindles occupy
the shorter axes of 0* and 042,

Fig. 22. Left side of same egg showing the nuclear conditions in quadrant 4.
'The spindles are not yet formed.

Fig. 29. Right face of optical, nearly sagittal, section, through quadrants B and
D,from the egg shown in the three preceding figures, to exhibit the
exact dorso-ventral extent of the cells 0*1 and 0^?, as compared with

the lateral extent of the same cells in

shown

Fig. 24. Dorso-ventral, approximately frontal,

in Figures 20-28, showing the £ tral extent of the
" g

of

‘ison with the late

20 and 22.

ce

ral dimen

quadrants A and C, for «

sions of the ss

25. Posterior surface re shown in Figures 20-24.

JENNINGS. — Asplanchna.

PLATE 4.

Figs. 26-29. Ten-cell stage.

Fig. 20. Left side of an egg in a ten-cell stage, slightly older than that shown in
Figures 20-25. Note the dorso-ventral elongation of the cells a*'i
and a*?, as compared with the same cells in Figure 22.

Fig. 27. Right side of the egg shown in Figure 20. The cytoplasm is beginning
to become constricted in the cells of quadrant C.

Fig. 28. Dorso-ventral, approximately frontal, optical section of the egg shown in
the two preceding figures, viewed from the anterior side, to show the
greatest dorso-ventral extent of the quadrants A and C at this stage.

Fig. 29. Right side of the egg of Asplanchna priodonta Gosse, in the ten-cell
stage, showing the spindles in the shorter axes of the cells.

Figs. 30-86. Sixteen-cell stage.

Fig. 30. Resting condition, seen from the anterior side.

Fig. 31. Left posterior view of the egg seen in Figure 30, showing the position of
the asters in all the cells of quadrant D.

Fig. 32. Dorso-ventral, nearly sagittal, optical section through the quadrants B
and D, from the egg shown in the two preceding figures.

Fig.98. Posterior view of an egg slightly older than that shown in Figures 80-32,

Fig. 34. Nearly sagittal optical section, through the quadrants B and D, in the

same stage as that shown in Figure 33.

PLATE 5.

Fig. 35. Ventral view of a stage similar to that seen in Figures 33 and 94, Plate 4,
showing the antero-posterior position of the spindle in dl,

Fig. 36. Transverse optical section of the sixteen-cell stage, through the cells of
the third layer, exhibiting the position of the spindle in d93, The
section is viewed from the dorsal side. (Compare Plate 2, Fig. 8.)

Figs. 87-42. Twenty-cell stage,

Fig. 37. Completion of the fifth cleavage in quadrant D. Note the oblique posi-
tion of the elongated aster in d,

Fig. 38. Sagittal optical section of the egg shown in Figure 37. "The section
passes, on the anterior side, between the cells of quadrants A and B,
showing the cells of quadrant A.

Figs. 39-42. Different views of one and the same egg.

Fig. 99. Left side, showing the asters and spindles for the fifth cleavage in quad-
rants A and .

Fig. 40. Right side of the egg shown in Figure 39. Observe the more advanced
karyokinetic stages in quadrant C, as compared with quadrant A,
Figure 39.

Fig. 41. View looking down upon the animal pole.

Fig.42. View of the same egg from the ventral pole, showing the oblique posi-
tion of the asters in qd9-1,

Fig. 43. An older stage than that given in Figures 30-42; the fifth cleavage in

progress; the egg contains twenty-seven cells. The cells có? and

65? are forced apart by the cell c97,

7627

|
|
|

Fig. 44.
Fig. 45.
Fig. 46.
Fig. 47.

Fig. 48.

Fig. 51.

Fig. 52.

JENNINGS, — Asplanchna,

PLATE 6.

Right side of the egg shown in Figure 48, Plate 5, 27-cell stage.

Dorsal pole of the egg shown in Figures 43 and 44.

Posterior view of the egg shown in the three preceding figures. Note
the lateral position of the asters in di-.

Right anterior surface of an egg, showing the quadrants B and C in the
sixth generation.

Sagittal optical section of the 82-cell stage, viewed from the right side,
and showing the spindles in de and di..

Sagittal optical section of an egg slightly older than that seen in Figure
49, showing the process by which the cell d^? is formed at the sixth
cleavage of the entoderm.

Sagittal optical section of an egg a little older than that seen in Figure
49, showing the recently divided and separating asters in d'1, and
the beginning of migration of the cloud of granules which lies at the
anterior ventral margin of the cell 4.

Sagittal optical section of about the same stage as that shown in Figure
50. The nucleus of d has moved away from the periphery of the
cell, and the cloud of granules is distributed between it and the small
cells d^? and di..

Apparently there is some slight variation in regard to the changes
in the entoderm cell as compared with the other cells. From the
condition of the remaining cells of the D quadrant, one would infer
that Figure 51 is younger than Figure 50, though the migration of
the eloud of granules is more advanced.

Transverse optical section of the egg shown in Figure 51, through the
region marked in Figure 51 by the cell d. The section is viewed

from the ventral side.

had

a

a

JENNINGS. — Asplanchna.

63.

PLATE 7.

Quadrant Dat a sta ge a little older than that shown in Plate 6, Figure 46.
The rotation of the asters in the cell dè is in progress. Thirty-two
cells.

Later stage than Figure 53, showing the final dorso-ventral direction of
the spindle in ds. A section of this egg is represented in Plate 6,
Figure 48. Thirty-two cells.

Anterior surface of 49-cell stage, showing the sixth cleavage in progress.

Ventral end of the egg seen in Figure 55, showing the large cell d"
nearly enclosed by the other cells.

The sixth cleavage in quadrant D. 38-cell stage.

Completion of the sixth and beginning of the seventh cleavage in quad-
rant D. Same egg as that shown in Figures 55 and 50.

Dorsal view of the egg seen in Figure 57.

Dorsal view of a stage later than the preceding, representing the same
egg as that shown in Figure 58. Sixth cleavage nearly completed.
Anterior view, showing the completion of the sixth cleavage in the quad-

rar

ts A and B in all the cells except a®ë Sixty-nine cells.

JuxNINGS. — Asplanchna,

PLATE 8.

Fig. 62. Dorsal view of an egg in approximately the same stage as that of Plate
7, Figure 61. The four small unlabelled cells at the point of meeting
of the four quadrants are a716—4716, About 69-cell stage.

Fig. 63. Ventral view of the egg shown in Figure 61, exhibiting the nearly com-
plete enclosure of the entoderm (to which the cell ds. belongs) by
the ectoderm.

Fig. 64. Sagittal optical section of the egg represented in Plate 7, Figures 55, 56,
and 59. The cloud of granules in the entoderm cell d^! now sur-
rounds the dorsal aster of the spindle of the seventh cleavage.

Fig. 65. Sagittal optical section of the egg represented in Figures 61 and 63,
showing the completed seventh cleavage in the entoderm ; the cloud
of granules forms a ring around the nucleus in the smaller entoderm
cell d9?, The ring being cut twice, appears in the form of two groups
of granules, one anterior, the other posterior to the nucleus.

Fig. 66. Posterior view, showing the seventh cleavage in quadrant D.

Fig. 67. Left posterior view; the seventh cleavage nearly finished in quadrant Y;
the sixth not yet finished in quadrant A. From the egg represented
in Figures 61, 63, and 65. Sixty-nine cells.

Fig. 68. Posterior view of a later stage than that represented in Figure 07. The
line separating d? and d®12 is the posterior median line; the spin-
dles are seen to be arranged symmetrically with respect to this line,
so that the cleavage is becoming bilateral. The entoderm is entirely

covered by the ectoderm; the region where a^! and d® are in con-
tact shows the place where the entoderm cells formerly occupied the
surface. 82-cell stage.

Figs. 69-74 give different views of the same egg, a 94-cell stage.

Fig. 69. Anterior surface, showing the finished meridional cleavage forming the
cells a911 — c8-11, q812 _ 8-12, q8-15 68:15. and aS+16 — 68.16, and the spindles
for the equatorial cleavage of a75— 675 and at? .

Fig. 70. Right side of the egg shown in Figure 69.

PLATE 9.

Fig. 71. Left side of the egg shown in Figures 69-74, Ninety-four cells.

Fig. 72. Dorsal view of the egg represented in Figures 60-74. The small
cells in the centre, at the point of meeting of the four quadrants, are
ate — 716,

Fig. 73. Ventral end of the same egg, showing the crowding together of the
cells of the quadrants A, B, and C in this region.

Fig. 74. Posterior view of the same egg, showing spindles for the ninth cleavage
in some of the cells of quadrant D, and the ninth cleavage completed
in other cells of that quadrant.

Fig. 75. Anterior surface of a later stage, containing about 120 cells. At the ven-
tral end (lower part of the figure) the cells are much crowded and
many of them are very small The vesicles immediately below the
cells 58-15, «8:1, and 4914 are the small ventral produets of the cleav-
age of a^ 5- c^ and a^ =, the spindles for which are shown in
Figures 69 and 70, Plate 8.

Figs. 76-79. Successive stages in which the ectoderm is conceived to have been
removed from the right side, to show the entoderm cells.

Fig. 76. Egg at the stage shown in Figure 75. A frontal section of this egg is
given in Plate 10, Figure 81.

Fig. 77. Slightly older stage than Figure 76, viewed in tlie same way.

Fig. 78. Slightly older stage than Figure ‚77, showing the change in the position
of the cells of the entoderm and of those at the animal pole. A view
of this egg from the animal pole is shown in Plate 10, Figure 82.

Fig. 79. Later stage than Figure 78. The entoderm cells have changed position

still further, and are approaching cleavage.

wA /
Lato
V

X d K N

hà

ZI

A

S

TS

828

| *
TET

MA EAS &. he)

geen

pg n

SUE

Fig. 83.

Fig. 84.

JENNINGS. — Asplanchna,

PLATE 10.

Optical section (approximately frontal) of an egg in the stage seen in
Figures 69-74, showing the very unequal division of d$,

Optical section (approximately frontal) of a stage later than Figure 80,
showing the five large entoderm cells. A side view of the same egg
is seen in Plate 9, Figure 76.

Stage slightly later than that of Figure 81. The ectoderm is supposed
to have been removed from the dorso-anterior part of the egg, dis-
closing the position of the entoderm cells. A side view of the same
egg is shown in Plate 9, Figure 78.

Later stage than the preceding, viewed in the same manner. A spindle
has appeared in each of the five large entoderm cells.

Optical sagittal section of anembryo at about the time of the beginning

of the formation of organs.

PO 5
47 9 A i
| PE. y
) |
| j
\ A j ; ~
| | \ af
li N p
YN
| i | |
|
m
RZ

\ \
| i
\

Bulletin of the Museum of Comparative Zoölogy
AT HARVARD COLLEGE.
Vor. XXX. No. 2.

STUDIES FROM THE NEWPORT MARINE LABORATORY.
COMMUNICATED BY ALEXANDER AGASSIZ.

No. XL.

SOME VARIATIONS IN THE GENUS EUCOPE.

Bv ALEXANDER Acassız AND W. MoM. WoopwonRTH.

Wira Nine PLATES.

CAMBRIDGE, MASS., U. S. A.:
PRINTED FOR THE MUSEUM.

Novemper, 1896.

No. 2. — Studies from the Newport Marine Laboratory.
Communicated by ALEXANDER AGASSIZ.

XL.

Some Variations in the Genus Hucope. By ALEXANDER AGASSIZ AND
W. MoM. WoopwonrH.

We examined for various points nearly four thousand specimens of
Eucope (3,917).

Among these we found nine specimens with only three radial canals,
twenty with five, and three with six radial canals.

There were fourteen specimens in which one of the radiating canals
forked, the forking distal or proximal to the genitals being nearly equally
divided.

No less than thirty-nine specimens showed distinct traces of serrations
or spurs from one or more of the radial canals.

In eight specimens the radial serrations or spurs were not well defined,
and the position and number of the radial canals were indistinct.

In eight specimens marginal tentacles were observed, which had be-
come united at the circular canal, sometimes with the tentacle next to
the tentacle with the otolith.

In six specimens there were marked spurs projecting from the base
of some of the marginal tentacles.

In eight specimens there were two otoliths in each sense capsule. In
four there were three.

In the othor specimens the principal variations extended only to the
degree of development of the cycles of the marginal tentacles and of
the genital organs. The latter showed in some cases peculiar leaf-like
expansions extending laterally from the radial canals.

The radial canals were four in number in an overwhelming majority
of the specimens examined.

The study of some of the variations in the genus Eucope was under-
taken with a view of calling attention to the changes undergone in a
species of jellyfish, of which great numbers are always easily obtained

VOL, XXX, — NO, 2. 1

122 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

during the summer months. The size of the full grown Eucope is so
small that with an ordinary hand lens striking variations can at once be
detected, and it is possible with a low power to pass in review with
comparative ease a large number of specimens.

We hope also to call the attention of zoólogists to the advantages
of photography, not only in an investigation of this kind, but also to
its application for ordinary purposes of delineation. (See Plate VI.
Figs. 3-6.)

Dr. Woodworth photographed the specimens reproduced here on
Plates L-VI., and he has written a short account of the methods he
followed.

In reviewing the variations we have observed in one species of Eucope
(E. diaphana), we may call attention to the similarity of these variations
which occur in this simple Medusa to structures found sometimes in
closely allied genera or families, and even in some cases to characters
of groups considered as only distantly related to the genus we have
examined.

The great number of marginal tentacles in Eucope they have in com-
mon with the Aiquoride.

Eucope shares with the Oceanidæ the limited number of marginal
tentacles connected with sense organs. Pendant leaf-like expansions of
the genital organs recall those of Melicertide.

The presence in Eucope of spurs at the base of the marginal tenta-
cles recalls similar structures in Zygodactyla, Halopsis, and the like.

The forking or branching of the radial canals below the genitals is
found also in Willia and in the Berenicide, a family allied to the /Equo-
ride ; the forking is symmetrical in the latter, and asymmetrical in the
former genus.

The increase in number of the radial canals from the pouch at the
base of the manubrium is a structural feature which is characteristic
of the Æquoridæ.

The serration of the radial canals is a generic character of Saphenia
and allied genera.

The branching or sending off spurs from the radial canals of Eucope
is a structural feature found in Gonionemus, Ptychogena, Polyorchis,
and allied forms. The lateral offshoots in Polyorchis being, hawever,
arranged in regular succession on each side of the radial canal, much
like the rounds of a ladder.

The anastomosing of the radial canals is a feature now characteristic

of the Discophores.

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 123

In the increase and special arrangement of the otoliths in the sense
organ of Eucope we find the first trace of the specialization of the sense
organs of such genera as Oceania, Tiaropsis, and the like.

There were no variations noted in the shape of the digestive cavity,
or in the number of actinal lobes of the manubrium, even in specimens
with five or six radial canals in place of the normal number (four) o
radial canals. The actinal folds were always found to be four in number.
In one case only have we found the radiating canal originating from the
eircular canal. (Plate VIII. Fig. 19.)

The origin of the peculiar club-shaped intertentacular appendages
characteristic of Halopsis and Laodicea, as well as the spur at the base
of the tentacles in many Æquoridæ, may be referred to the spur-like
appendages of the marginal tentacles of. Eucope figured in Plate VIII.
Figs. 4-13.

And it may not be far out of the way to look upon the coalescence
of adjoining marginal tentacles with sense organs as the first indication
of such structural features as the radial marginal tentacles of Eucheilota,
or even of Boungainvillia, Margelis, or Nemopsis.

It is interesting to note that in Echinoderms there are five radial canals,
and four or six or more are considered monstrosities, while in Acalephs
four or its multiples are the normal number of radial canals, and five or
less are variations.

'The specimens of Eucope showing numerical or structural variations
were as a rule fully developed males or females, the eggs and sperma-
tozoa being apparently in a healthy condition.

It would be an interesting study in heredity were it possible to breed
the variations in Eucope here enumerated, and ascertain how far the
structural characters acquired in the variations we have observed can be
transmitted, and lead perhaps finally to the formation of types which
we have been accustomed to look upon as having no structural relation
with the genus.

But it is also possible that in a comparatively simple genus like
Eucope these variations are not necessarily to be considered as heredi-
tary; they may indicate possibilities in. mechanical combinations which

1 Variations in the manubrium have been observed in Tubularian Hydroids, such
as Lizzia, Dysmorphosa, Hybocodon, Dipurena, and Sarsia ; but as they are usually
connected with phenomena of reproduction and of budding they have only a dis-
tant connection with the line of the present investigation. See an interesting
paper by Hartlaub on the reproduction of the manubrium of Sarsia, in Verhandl.
d. Deutschen Zool. Gesell., 1890.

124 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

have become characteristic of jellyfishes, and result from their mode of
life and the simplicity of their structure. Their repetition in closely
allied genera may denote structural affinity, while in distantly allied
groups it may be the result of mechanical combinations, and in no way
indicate any affinity.

The four primary segments of Eucope are of uniform size in the ma-
jority of the specimens examined. Whenever there is suppression of a
radial canal, as in the case of specimens with three radial canals, the
segments are sometimes uniform (Plate III. Figs. 3, 6), or one of the
segments, as in Plate III. Fig. 1, is nearly 180°, indicating the total sup-
pression of the fourth radial canal at its normal point of development.
See also Plate III. Fig. 4, in which one of the segments is smaller than
the two from which the fourth section has been cut.

The inequalities which exist in the segments of some of the Specimens
can best be expressed by a table : —

2 3. 3 6. 2.5
I3 13 li-ii l; 1;—1; 3 lur. 1; 1 mE.
5 5 3 4 4 2 83 18
1; 15; 1; e gs Ui r 111-1; 1 113 Pp 11
—
2 3 3 3 er
1515151111 11 PPL 1s 1; 1; 1;
— — —
1. 25, 25 25 1 26 35 86 36. 1 864 4 35
F lee Leeds Ly ener eA
5 2 2
1315151 poh ah 1; T.
The —-—. indicates a fork of the radial canal.
see E5491 T EE
1, m 1197; P AUTT TN a 1 1 1 15

In this table 1 expresses the smallest segment; the approximate

dimensions of the others are represented by multiples of it thus: x indi-

cates that a segment is twice as wide at the periphery as the smallest ;
= that it is three and a half times as wide. Of course, the value of 1

is a different value in each case.

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 125

1; 1; 1; is the formula for the segments of a Eucope in which they
are of equal size, the radial canals forming an angle of 120? at the
centre.

1; 1; 1; 1; and 1; 1; 1; 1; 1; would each denote the formula
for segments of equal dimensions in a Eucope with four and one with
five segments.

The formation of additional radial tubes may be due to the growth
of independent tubes from the pouch at the base of the digestive cav-
ity, or from the forking of tubes, the new canals eventually reaching
the marginal canal. In one case we observed a radial canal which
had its origin at the periphery and did not extend to the base of the
manubrium (Plate VIII. Fig. 19). Such a formation of a new radial
canal from the circular canal suggests a similar structure in the short
canals, in which are found olusters of lasso cells extending at right
angles from the periphery between the primary radial canals of Willia,
and perhaps other Medus, in which we have clusters of lasso cells
extending a short distance on the outer surface of the umbrella from the
marginal canal.

In the great majority of the specimens of Eucope observed, the
radial canals are tubes with walls nearly parallel all the way from the
base of the digestive cavity to the marginal canal (Plates I.-VI.). But
in a great many instances this parallellism does not exist, and we find on
the edge of some of the radial canals slight serrations, as in Plate VII.
Figs. 1-4 and 6. These serrations vary greatly in size, and in some
cases become short spurs (Plate II. Fig. 4, Plate III. Fig. 3, Plate VII.
Figs. 2, 3, 5, 7), or even spurs of considerable length (Plate VII.
Figs. 6, 9, 10); the longer spurs becoming often the forks of the
primary radial canals (Plate III. Figs. 1, 2-5, Plate VI. Figs. 1, 2,
Plate VII. Fig. 5), either above or below the genital pouches. Or the
spurs may form connecting canals between the radial tubes (Plate VII.
Fig. 4, Plate VIII. Fig. 20), or a rudimentary circular canal round the
base of the manubrium (Plate VII. Fig. 8).

Starting with the normal state, in which the genitals are equally
developed, we find five or six variations, which cover by far the greater
number of the specimens examined.

The greatest number of the specimens (622 out of 1146) examined
for variations in the genital organs were normal, the four genitals being
equally developed ; the females were more numerous than the males;
of the latter there were 175, and of the former 447. This stage is
represented in the table by 1; 1; 1; 1; in which 1 means that the

126 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.
genitals are fully developed, and equally so on each radial canal, of
which there are four.

The next stage represented by the formula, 1; 15; 1; 0 ; the genitals
were atrophied on one of the radial canals and equally developed on
the others. Seventy-eight specimens of this stage, 49 females and 29
males.

The next most frequent stage is that in which two adjoining genitals
are fully developed ; the others are of the same size, but less well de-
veloped ; that stage is represented by the formula 1; 1; 2; 2; Out of
the 1,146 specimens examined there were only 74 specimens of this
stage, of which 45 were females and 29 males.

Next comes the stage in which only one of the genitals is fully de-
yeloped ; the others are less so, corresponding to the formula 152; 2; 2;
39 females and 27 males.

In the order of frequency of occurrence comes :

Eucope with the formula 1; 1; 1; 2; — 28 females and 17 males.

Then comes the stage in which the genitals were unequally developed :

20 females and 21 males.

Next,

Eucope with the formula 1; 0; 0; 0; — 24 females and 13 males.
“ “ [T 1;2; 0; 0; — 15 “ « 3 «
[73 [T 40 PARITI 0; — 10 [73 “ 8 66

In specimens with three radial canals we observed only one specimen
in which the genitals were uniformly developed. On Plate III. are seen
(Fig. 1) a specimen in which one of the canals forks at the extremities of
the heart-shaped genitals, forming three primary segments of nearly
equal size extending to the centre of the disk, with a small sector cut
from the outer edge of two adjoining segments.

Figure 4 of the same plate shows a specimen with three radial canals
and four genital pouches, but the canal which forks subdivides above
the genitals so near the centre of the disk as to subdivide the disk into
four nearly equal segments.

A variation similar to that of Plate III. Fig. 1, for a three-rayed
Eucope, has been observed in a four-rayed Eucope (Plate III. Fig. 2), in
which the fifth sector is a comparatively small triangle cut out from the
periphery of two of the adjoining sections, extending to the centre of
the disk.

In Figure 5 of Plate III. the forking of the four-rayed Eucope, taking
place nearer the centre of the disk, subdivides the disk into segments of
more uniform size, and it closely resembles a five-rayed Eucope.

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 127

In five-rayed Eucope only one specimen was observed with equally
developed genitals (Plate V. Fig. 1). Those figured on Plate VI, as
well as the others on Plates IV. and V., all show a very great range
of variation in their development. In two cases (Plate IV. Figs. 3, 6)
two of the genitals are wanting.

In Plate VL, Figures 1 and 2 are those of two specimens with five
radial eanals, both of which fork so near the base of the manubrium
as to divide the disk into six nearly equal segments.

The accompanying table will show the other variations which have
been observed, and their frequency.

1146 SPECIMENS EXAMINED FOR VARIATIONS IN THE DEVELOPMENT OF THE
GENITAL ORGANS ON EACH OF THE FOUR RADIAL CANALS.

With Three Radial Canals, or Three with Fork.

1; 2; 8; 3 females, and 1 male.
BICI a" S "5 1g T
—

,

1; 1; 1; 1 female.
FFF

With Four Radial Canals.

1; 1; 1; 1;* 447 females and 175 males.
III Wr 89 4 "29. Vt
1; 1; 2;2; 46 a "Sa uy
1121/2; 3;. 09 hy T ay

is lelia oas A "en Iscr
11219; 4; 20 2 SR AE. o"
1;0: 0, 0» 2.924 : ts AH

131; 0:0; 21 -T )
1/2 ,99903 19 t à
159404060 18 Sot
Lp ene: 03 10 8

Le 3,900: 6 t 8 *
1314 9 5 T Es"
ai es 5 ‘ £79
as 23.8) à 4 E
1;0;0;2; 8 18
1;0;2;3; 3 2 «
by Deh Ss 4 males

2

* 1, fully developed genitals; 2, genital smaller than 1; 8, genital smaller

f

than 2; 4, genital smaller than 3; 5, genital smaller than 4; 0, genital organ

absent. The semicolon (;) indicates the position of the radial canal. The —-

indicates the forking of the radial canal.

128 BULLETIN :

pi pd pd pd pd pd konn
d o — — 0 — n 8 O
Co do — CQ rto S SPR On 0 O tto

ow

With

wm Co b
— co

Be oe — P.

nw
A Wwhd

eo

mm

—
we

4

bo CoN O c5

MUSEUM OF COMPARATIVE ZOOLOGY.

PRN ROR OWRF BRK WWOW SO

e

ve our» tO

ww
nes
PR TR PE RoR PO

—

|

4 females.
8 “

3 specimens.
2 females.
1 “

1 e

1 [11]

1 male.

1 female
1 «€

1 of

1 “

1 male.

1 female.
1 [1j

1 [11

"ive Radial Canals, or Four and Five with Fork.

1 female.

2 females and 1 male.
1 4€

1 male.

1 female.

1 “

T “

1 male.

1 LI]

1 female.

1
1 male.

4 females and 1 male.

“

3 females and 1 male.

The variations in the shape of the genitals from a circular to an ellip-
tical outline to lobate, foliate, or elongate shape (Plate VIII. Fig. 21) is
usually connected with the growing together of the genitals of adjoin- |
ing radial canals (Plate VIII. Fig. 18) or with their separation into two
distinct bodies when the radial canals fork (Plate III. Figs. I, 5).
Tt will be seen that in Eucope, as has been observed in Aurelia, the
tendency in the numerical variation of the genitals is greatest in the

direction of an increase rather than a diminution in their number.
similar tendency is also noticed in the numerical variation of the radial

canals of the otoliths.

A.

AGASSIZ AND WOODWORTH: VARIATIONS IN RUCO E. 129

The youngest specimens of Eucope observed (eleven in number) had
twelve tentacles in all, one at the base of each of the four radial canals
and two in each quadrant with a seuse organ, and without any tentacles
intermediate between the radial and the sense organ tentacle. t is
possible that these specimens may belong to a different species, as the
youngest specimens of Eucope diaphana I have raised from the hydroid
already had 24 tentacles." But of course it is possible that the above
stage may have dropped from the reproduetive calycle at an earlier stage
of development.

Were the marginal tentacles to develop regularly we should have as
the third stage a Eucope with 48 tentacles, a fourth stage with 96,
and a fifth stage with 192.

But as far as our experience shows, we find normal Eucope only in the
first and second stages, and even then there are numerous variations.

The formule for the normal stages are, calling T any marginal
tentacle,

Te, the primary tentacle at tho base of the radial canals ;

To, the primary tentacle with a sense organ at its base;

U, the first set of marginal tentacles appearing between T, and 79
the stage with 24 marginal tentacles ;

tə, the second set of marginal tentacles on each side of t, in the stage
with 48 marginal tentacles ;

ts, the third set of marginal tentacles on each side of £j, in the stage
with 96 marginal tentacles ;

t4, the fourth set of marginal tentacles on each side of £4, in the stage
with 192 marginal tentacles ;

ts, the fifth set of marginal tentacles on each side of „ in the next
stage with as many (theoretical number 31 tentacles) as 1645, 84% 4ts,
2t,, 14, tentacles between the (T, and T,) radial canal and sense organ
tentacle, or 95 tentacles in each quadrant between the radial canals, or a
total theoretical number of 384 marginal tentacles. l

The formula of the youngest stage, or first stage, would be

X (4) T, + (4) 2 Ty = 12 T,
or two tentacles with a sense organ in each quadrant.
That of the second stage,
X (4) T, + (4) 2 To + (4) 36 = 24 T,
or two tentacles with a sense organ in each quadrant with a tentacle of
the ei order between each and the radial canal and sense organ.

1 Proc. Boston Soc. Nat. Hist., June 4, 1862, p. 92.

130 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

That of the third stage,
3 (4) T, + (4) 2 Ts + (4) 3, + (4) 6t — 48 T,
the same as the second stage, with an additional tentacle of the ty
order intercalated on each side of the ¢, tentacles, making 11 tentacles
in each quadrant, or three between each of the three primary divisions
of the quarter.
The fourth stage,
X (4) T, + (4) 2 To + (4) 34, + (4) 6½ + (4) 124, = 96 T,
or a tentacle of the third order, 4% intercalated on each side of to, thus
making 7 tentacles in each of the primary divisions of the quarter.

The fifth stage, with 192 tentacles, would have for its formula,

2 (4) T, + (4) 2 Ty + (4) 34, + (4) 6ta + (4) 12½ + (4) 24t, — 192 T,
or a tentacle of the fourth order, /, intercalated on each side of th,
thus making 15 tentacles in each of the primary divisions of the quarter.

The theoretical formula for the sixth stage, with 384 tentacles,
would be,

2 (4) T, + (4) 2 T, + (4) 31, + (4) 6t, + (4) 12½ + (4) 24t, + (4) 48t; =
384 T,

a tentacle of the fifth order, ts, having been intercalated on each side

of the tentacles of the f; order, thus making 41 tentacles in each of

the primary divisions of the quarter.

It is not necessary to carry these stages any further, as the largest
number of tentacles observed in any primary division of each quadrant
is not greater than thirteen, which would limit it to a modified stage of
192 marginal tentacles.

The accompanying tables have been arranged so as to indicate the
association, whatever their number may be, with the highest number of
marginal tentacles occurring in any primary quadrant. This number
may be a normal number between any T, and To, or between any To and
To, and may be due to the absence of one of the To tentacles, or of both,
in the quadrant.

A glance at the tables shows how few of the specimens examined
possessed the normal number of tentacles. We should have as the
normal stages of each quadrant for the four stages up to 192 tentacles.

Te, To, To, Je, first stage, with 12 marginal tentacles.

Tortu Toy Ay Doty T, second stage, with 24 marginal tentacles ; or,
asin the tables il ebd Sled Ty 1, 15 of which eleven speci-

meus were seen from the number tabulated for numerical variations.

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 191

The formula for the third stage, with 48 tentacles, is,
To to, ti, boy To, boy ty lo, To, to, 115 ly Te,

or, as in the table, 3, 3, 3; 3, 3, 3; 3, 3, 3; 3, 3, 3; of which fourteen

specimens were found from the number tabulated for variations.

In the fourth stage the formula is,

Ts fa, ta, ta, 115 [m ty, ty, Tos ts; toy ts, tis tz, 125 135 To, ts, to, ts, 115 ts, bo, m Los

or 96 marginal tentaoles.

Of this normal stage only two specimens were observed.

Of the fifth stage, with 192 marginal tentacles, of which the follow-
ing is the formula, not a single normal stage was observed :

To ta, ts, ty, toy ty, lg, Um hy [7n ta, ty, ty, ta, ts, la, lr» la ts, m 125 ty, T tay hy
lp ts ta bay la by, tay To, tay tss ty tos tay tas tas boy tay ts, tas try La, tay La, bo,
ty, lay ta To.

But a few specimens were seen with the normal number of 15 tentacles

in some of the primary divisions of one quadrant.

Of the sixth stage, with 41 tentacies in each of the primary divisions
of the quadrant, not a single specimen was collected.

It is interesting to note that in the specimens which may be said to
belong to the second stage, some of the primary divisions of the quad-
rants remain in the first stage with only one tentacle, and others with
two, the third tentacle not having developed.

Between the second and third stage, in a number of specimens, the
greatest number of tentacles in a primary division of the quadrant is
four, and the smallest one. In a number of cases two or even three of
the quadrants remain in the second stage, and only in one quadrant do
we find four tentacles in a primary quadrantio division. Similarly, we
find a number of specimens with five tentacles as the largest number in
any primary quadrantie division, and some of the quadrants in the second
stage, but none in the first. Even when we come to six tentacles as the
largest number of tentacles in any primary quadrantio division, we still find
some of the primary quadrantie divisions in the second stage, and occa-
sionally a whole quadrant or a quadrantie division above the first stage,
as is the case in the second stage and stages intermediate with the third.

Only four specimens typical of the normal third stage were observed,
and the great majority of the other specimens in which seven tentacles
were found in a quadrantic division belonged to stages approximating
the third more nearly than the second. Only a small proportion of the
specimens were observed in which the quadrantic divisions belonged to

132 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.
the second stage. As will be noticed from the tables, by far the largest
number of specimens examined belonged in or near the category of the
third stage. The number of specimens with more than seven tentacles
in a primary quadrantic division became less as the number of tentacles
increased, and the whole number of tentacles in any quadrant was also
more variable.

No fully developed Eucope was found in the fourth stage with 15
tentacles in the primary divisions of all the quadrants (see table for 15
tentacles, page 136), and when we come to a larger number of tentacles
for each primary quadrantic division the irregularities of the marginal
tentacles leave us in doubt in which stage to place them.

In the more advanced stages sometimes only one quadrant is affected,
the others remaining less advanced. This seems to indicate that the
mere multiplieation (of parts) of the marginal tentacles beyond the
third stage is accompanied with greater and greater irregularities in
the different quadrants as the number of marginal tentacles increases.

A special table has been made to show the variations observed from
the normal number of eight otoliths and of the four radial canals.

In the tables the notation 2, 8, 8; 3, 3,3; 3, 3, 3; 3, 3, 3; the 2
indicates that starting from a radial canal there are two marginal tenta-
cles between it and the marginal tentacle with a sense organ indicated
by a comma ; the 3 indicates three marginal tentacles between it and the
next sense tentacle, and again three marginal tentacles between it and
the radial canal; and so on, the figure always indicating the number of
marginal tentacles, the comma the sense-bearing tentacles, and the
semicolon the tentacles at the base of the radial canal.

Variations in number of intervening marginal tentacles, or of sense
organs, or of radial canals, are thus easily indicated.

Thus,

„ 010,00 1 0, DOES
indicates the presence of six sense-bearing tentacles, there being only
one in two of the quadrants.

))) 9,35109,9095, 10, 254 83485 5012

on the contrary, is a Eucope with five radial canals and nine otoliths,
while

iti all, 253
indicates a Eucope with only three radial canals and seven otoliths.

In the tables, except in the list marked irregular, those specimens
which differ from the normal in the number of their primary quad-

oD
eo
*

EUCOPE.

VARIATIONS IN

AGASSIZ AND WOODWORTH :

rantic divisions or quadrants, or in that of the otoliths, or in that of

the radial canals, are marked by an asterisk.

First Stage.

With only one tentacle in each primary subdivision of the quadrant.

1593098 551,53113 913 023

Second. Stage.

With not more than three tentacles in any subdivision :

e$ oo od o
ao oS

re
—
o O c9 c
eS cd od e$
Sn

o c o oS

With not more than four tentacles in any subdivision :

-
O O9 co co «M CO Cd N oo
S - HS 08 4 M c cd o X
Hoi

Om co co cd Ho cO co Co A
SPs Sod cO cd CO o oS
- o o3 c3 08 c8 ed M e$ od
MH CO X4 CO cO co c3 CO có «Box
o Hed c3 eG ed c3 cO cd cd A
e$ A c Hod Sod c r
69 co c5 c5 C9 cg co Co co co E
CS e$ od ed C$ od c3 cO cO od A
oS oS oS c3 c od o cO oS cO
*

C9 c0 Q4 co XH « CO M co MOX
Sod c3 Ha dS od od có Cd E

o Hd od PP od 08
ocean
Soma «^ cO -M CO co CO co
CS eS od oS oS od e$ cO eG od o
63 o c oS oS oS OS cO cO cO OS

With not more than five tentacles in any subdivision :

- o c0 0.00 -M C3 am «tn CO 19 9
Coit Laid HOI
- aud ad odo ad c8 9 ad aS c8 06 o
MO 15 1D CN CO 6D o 15 H 19 10 CO ai
Hid wed CA OSS aS ds oS
Mad od od i od ad Sid ad id ST
MAR am qq c5 69 00 am H oam co am
Hid ef ad M € ad a v ad ow d a
Cid iw od Odd ac cs ad ad Er
Mam «OP 05 E 0$ am Mam ooo
cs co 09 cd o am cO c) cj
Sted ed Sod cO ME ME Pid od od c

!! T Oem. sie M
Hon c0 € CO H iO cO CO iO cO cq a
- ad e$ 0$ 6 oa ds Hs SSS c8
oS O8 08 ow ad ad có ad o6
*H CO co CO cO CO «HM iO «H «H CO CO CO SH 1D

iG eS. ced CS Sid Tod aS ad E oS oS cS aS

- A dj 08 a5 09 a x oam x ooo
VS oy oy HA ded - cO ad - c8 M cod a

- oS e$ 08 o M aS aS ad ad ad 4
iS CO im «M E OO cO CO CO «M CO XO B CO jS
Ws ed aW Hod c9 c3 Hid € o 08 a
Pw Pod Pod cO cO Ped cO *

All specimens characterized by some peculiar variation are marked *

=
©
Q
—
Be)
o
N
=
>
—
—
-
ES
<
=
=
[e]
m
E
o
E
=)
E
2
=)
=

BULLETIN

134

ision :

my subdiv

With not more than six tentacles in a

6, 6, 4; 4, 4, 8;

6, 4, 3

5;
8,4

„6, 3; 4, 6,

3

MED

,4; 5,4,4;

3

?

5

; 4, 6, 0;

, 6, 6

2.

?

3, 4,
4

x

5

,

4; 5,6,5

5,

3.4,

4

,

4

,

sıon

y subdivi

es in an

th not more than seven tentacl

i

135

5, 0;

€

9,

EUCOPE.

3,4; 3, 4, 5;
„

€

IN
4,

>», 6,

„6
x

t

)

"m

6, 6, 6; 7, 7, 6; 7, 7, 6;

1

VARIATIONS
i 4
‚6;

Dy 5

8, 4, 8; 5, 7,4;

p
4, 6, t; 4,

,

AGASSIZ AND WOODWORTH
6, 7, 6; 6, 6, 4; 5, 7, 5

„6; 6, 6, 5; 5, 6, 5;

)

R
„ U

4

5
T .
7

6,

7,

7,
$159

Senso
or oO
Or we
S To
= p. cO
ern
S = N
M2 p p. 5
S rTopT
N N oA
*

D
S N e
iS am wa

1.6, 7, 75 7, 7, 6

y subdivision :

in any primar

With not more than eight tentacles

TE.

my subdivision : —

m

2, l;

*

D

tacles in

itacles : -

n

ter

"

in nine

an ten te

the
ne
9

1

"

NO.

XXX

=

VOL.

With not more

136 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

With not more than eleven tentacles : —

*11,5; Mp: 5, 5, 5; 6, 5, 4; 3, 11, 10; 6, 4, 8; 9,8, 10; 6, 8, 4;
2, 11, 5; 5, 9, 4; 5, 10, 3; 5, 7, 3; * 4, 6, 5; 2, 8, 6, 7; 1, 11, 2;
8, 7, 9; 9,7, 61 11, 10, 10; 11, 7, 10 3, 5, 53 4, 6, 63 4,11; 11,2;
*3; 11,0; 5, 4, 23 8, 8, 4; 7, 7, 83 7, 8, 11; 10, 7, 7; 11,6, 7; 11, 7, 6;
With not more than twelve tentacles : —
*5,5,5; 5, 6, 6; 7, 6, 6; 12, 7; * 3, 6, 2; 3,3; 10, 9, 9; 10, 7, 8;
* 077509011757 PN pity lan On 8, 7, 12; (2 forks).
*5, 7,4; 4, 7, 6; 5, 7, 4; 12,5;

|
With not more than thirteen tentacles : — |
* 6, 6, 4; 6,18; 6, 7, 4; 6, 7, 7; * 4, 4, 5; 6, 4, 4; 13, 5; 5, 6, 4; |
*8, 3, 13; 5; 11, 8, 11; 12, 7, 10; *9, 7, 8; 10, 7, 10; 18, 9,7;
With not more than fourteen tentacles : —
*7,7,7; 8,7,8; 6, 7, 6; 14,7; *5,7,1; 6,7,6; 7,8,7; 9,14;
MEG, OF (1551915 K104* 1453
Fourth Stage.
With not more than fifteen tentacles : —

„%%% ᷑ 10.0 7 0671019, 27 * 6, 7, 10; 6, 7, 15; 12, 14, 5;
6,6, 6, 10; 0,21 1, 8, 0; WT 5,0: 0,005 9, 105 1,6, 11
*5/0:01 10/03. 0, 1/13 1 0, (1

With not more than seventeen tentacles : —

*17; 15; 15; 16; no otoliths.

With not more than eighteen tentacles : —
*6, 6,6; 7,7; 9; 18,7; *9, 7, 9; 18,7; 11, 17; 11, 18;

With not more than twenty tentacles : —
TOT OI hEr i IUS AU SÍ

With not more than twenty-seven tentacles : —
*9,6,8; 27; 8, 11;

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 137

Table showing Variations in the number of Radial Canals, of Quad-
rants, of Primary Divisions of Quadrants, or of the Otoliths.

Number of Number of

Irregular, Otoliths. . 5 1
6,7: 15 5, 1,0: 0,0 9j . 5 4
11,8, 11,03 5/0 ,0; 0,0; 4] 6 4
5,7,6; 14, TESTEN NOT € "i 4
5,5,5; 5, 6,6; 7 159; OF Fa T 4
8, 5, 8; 6, 3, 4; 4, 5, 5; 8, 4; 7 4
0, 605 ee $ 4
19,8: 0,9, 0) Dy PDs Ones 7 4
Uu T iuc. 51) rf 4
0, 11015719, 1503)0, 1; ey 142/01 7 4
50,919 0,18::0) 7, 450, 7,14 7 4
6, 6, 6; ee 5 4 4
V 6 4
20, 03 2,15, 6, 7; 9,10; 9 33 7 4
8,6, 0,0; 10, 9,97 10, 7,8 8,7 l; 12; 9 5
2, 0,05 6,9: 5,6,5; T, 6, 10; . 7 4
3, 8, 8; 3, 3,4; 3,5,4, 8,8; 7 4
4, Op Od ty Sk de LE NEE € 5 4
% By Oh: 5 950,21 Oe 25 0 8: 9 4
or + rS NS T 5
890,2; 95,1, Oy 55 5 272, 5, 15 9 4
0,0078; 6, 05 0j En 01-90 9, Oy 0 4
6,0, 7; 6,15; 6,2; 1, 8, 6; 6 í
CANA. TE NET 1 7, 4; 9, 4: 7 !
949,11 0,9, 41 129,011 3 8 t
113 or 107 104. . i 0 1
0/9007 Uy) 0s 0705 OF 10705 7 t
446.25 7222826, 7 3511, 91 7 3
4,4,5; 6, 4, 4; 18,5; 6, 6, 43. 7 1
4,48; 5,0; 9,4; 44,5; 6 4
8, 4, 4 10, 1; 8, 5, 5; 4, 6,6 7 4
8, 5, 5; 4,5,6; 4, 11; Il 4: 6 {
e 5 {
O70, Oy 197/07 065 G T8HTS A T 4
5, 05:73 0/0, 21.8, 10; 8; 7; . a 1
1585 015/59, 15:1, 0345 06:7 405 0,03 9 5
6 7 10 6, 7, 155.19, 14, 0; 6 a
9, 7, 8; 10, 7, 10; 18, 9, 7; Lb an 3
ÜriniT Op iy Oe 7,8, 7: 5,7, 7; 9, 14; 9 4
8, 3, 3, 8, 8, 4; 5, 3, 4; 8,6, 7 4
129,95 1, 8,11 OF Gs 8/10] 6 1
Dd 03 Dido Te 6 4

138 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

ap np Number of
Irregular, W Led
74 0908 409 iy Bae Saw lean 20, Be S963 6 5, 1 forking.
9/0/55 265 Oils ST s 3 3
0p 5»007, To Dis DAD: 2, 01 0, 0,0 8 5
7 O25 (5, S: 940:.147 5 5 4
2, 8, 8; 3, 1, 3, 4; 1,8,2; 2, 8, 4; 9 4
6, 6,7; 6,15; 6, 2; 1, 8, 63 6 4

On Plate IX. will be found a number of diagrammatic figures giving
an idea of the irregular growth of the marginal tentacles. The lengths
of the tentacles are drawn as fully expanded, the position of the radial
and circular canals is indicated, and the otolith tentacles are marked by
a cross. The structure and length of the marginal tentacles of Eucope
are such that the comparative length of adjoining tentacles is readily
observed, owing to the slight degree of contraction and expansion they
possess.

In Figures 1 to 6, 10, and 14 (Plate IX.), we have the normal num-
ber of tentacles (seven) in each of the primary quadrantie subdivisions.
It is noticeable that /, can only in the case of Figures 4 and 10 be dis-
tinguished from the two ½, while the four t, are of nearly uniform size
in all the figures except Figure 6.

In Figures 7, 12, 13, 16, 17 (one primary division), 18, 19, and 22—
26, there are only five marginal tentacles in each primary quadrantic
division. In the greater number of these figures it is possible to dis-
tinguish ¢,, or f, and , while the irregularly developed tentacles are
part of the 1, cycle. In Figures 20-22 (one sector) and 27-29 (two
sectors), the cycles ¢, and t, can be distinguished, and the irregular-
ity of development occurs in the t, cycle, which may appear at different
points of the circular canal, as is seen by comparing Figures 20-22, 27,
and 28.

Figures 8-10, 17, 22, 29, and 30 show the irregularity in time of the
development of the marginal tentacles in the different sectors of the same
quadrant, as well as the irregularity in the growth of the three cycles
^4-t, in adjoining sectors. In Figures 8, 29, and 30, the marginal
tentacles between the otoliths are in the same stage of growth, but
the tentacles of the right or left sectors are in very different stages of
growth. In one case (Fig. 30) only one £, is developed in each sector;
in Figure 29 the cycles of 4 and t are normal, while they are most
irregular in the sectors of Figure 8.

When six marginal tentacles occur in one seetor the irregularities in

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 139

the tz and ts cycles are very marked (see Figs. 8, 9, 11, 15, and ya
t, can only be distinguished from the ¢, pair by its position.

The coalescence at the base of adjoining marginal tentacles to form
a double tentacle with two spurs and two lashes is not uncommon.
During the summer there were fourteen specimens met with having
double tentacles ; in all except two cases they were connected with the
tentacle riding a sense organ.

The sensory tentacles usually have only one otolith; we however
observed thirteen cases in which each sense organ contained two (Plate
VIII. Figs. 15, 17), and five in which there were three otoliths (Plate
VIII. Figs. 14, 16), and one in which there were no otoliths in any of
the quadrants.

An examination of the table on page 137, in which the more interesting
of the variations observed have been collected, will show how large a
number of specimens show great variation in the number of the sense-
bearing tentacles. Among specimens of Eucope with the normal number
of quadrants we find the otoliths bearing tentacles vary from eight, the
normal number, two in each quadrant, to three on one side and ten on
the other. As will be noted, there are only five cases in which the sensory
tentacles are greater in number than in the norm, while the number
of cases in which they are suppressed is quite large. Their increase
does not always accompany an inerease in the number of radial canals.
Two out of four specimens with five radial canals possessed nine sensory
tentacles, another only seven, and one eight.

There seems to be no correlation between the number of marginal
tentacles in any sector and the number of sensory tentacles.

The primary sector with the largest number of tentacles has often
only one sensory tentacle, while that with a smaller number has two.

In a specimen with quadrants of unequal size, and with an unequal
number of marginal tentacles, in each of which the formula is

T 109 4 000% 0,9 05
there are six otoliths, one on each of the tentacles at the base of the
radial eanals between the first and second and the second and third
quadrants, and the others as marked by the comma in the third and
fourth quadrants.

In another specimen with very unequal quadrants there are nine oto-
liths, there being three in the largest quadrant, as shown by its formula.
the first quadrant being the smallest, the last the largest :

1, 3, 23 2, 8, 43 2, 3, 33 3, 1, 3, 4;

140 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

In a specimen in which two small quadrants are adjacent, there
are only six otoliths. Taking the two smaller quadrants first, the

formula is
0723. 15:8, Opb bt 5001: 101

The last quadrant is the largest.

The formula of another specimen, with unequal quadrants, is,

„

This shows only four otoliths, one quadrant without any, and two with
only one.

In a similar specimen with a formula of

le s 11; 12,1, 105 SEDES

there are only six otoliths, with three normal quadrants, one not having

any sense organs.
In a specimen with two adjoining radial canals forking below the
genitals, making four quadrants with two small seotors cut out of two
of them, there are ten otoliths, the formula being
— — oe
VVT
In a second specimen, forking similarly at the extremity of one radial
canal, the formula was
— —
VVV
In a specimen with an eccentric digestive cavity and quadrants un-
equally developed, the formula is
La e Ta Beel aD
or only five otoliths, neither the first nor the third quadrant having
otoliths, while the fourth has two side by side.
In a specimen with equally developed quadrants, but with a long
spur above the genitals at right angles to the radial canal, barely reach-

ing the marginal canal, the formula is
7 f. - 7 Q
VVV

there being three tentacles between the fork (spur) and the nearest

canal.
In a very irregularly developed specimen, with the formula

CCC

there were only three otoliths.

AGASSIZ AND WOODWORTH : VARIATIONS IN EUCOPE. 141

In a specimen with only three radial canals, one of which forks above
the genitals, the formula is

TE
2, 15, 6, 7; 3, 10,1; 9, 2; 2, 5, 5;
the otoliths being normal in number, but irregular in their distribution.

In another Eucope with three radial canals, we find seven otoliths
represented in the formula

1,115: B54 1.655 2, 86 7

the last sector being somewhat larger than the others.
In a specimen with three equally developed sectors with the formula of

12, 14, 6; 6, 7, 10; 6, 7, 15;

there are six otoliths, the normal number in each sector, but the other
marginal tentacles are most irregularly developed.

In still another Eucope, similar to the preceding one, with three
sectors of equal size and well developed genitals with six otoliths, two
in each sector, the formula was

10, 7, 10; 10, 9, 7; 9, 7, 8;

Finally, a Eucope with three quadrants of the same size had only
three otoliths ; its formula is

Yi, 213.6

The formation of spurs ee VIII. Figs. 4-13) takes place usually
at the base of the marginal tentaele, at its connection with the circular
canal, but cases have been observed in which the spur shoots off from
the lash of the tentacle (Plate VIII. Fig. 5). The formation of spurs
is often accompanied by the atrophy of the inframarginal knob of the
marginal tentacles (compare Plate VIII. Figs. 9-13, with Figs. 4-8).

It will be noticed that in the numerical variation of the segments the
tendency is not to doubling the number of normal (4) segments, but
either to add one or two, or to reduce the segments to three.

In Sarsia, Agassiz observed a specimen with six radial canals and
Romanes one with five; he also observed a Sarsia with six radial canals,
six ocelli, and six wi like that seen by Agassiz, the only speci-
men in thousands examined.

According to Bateson, the numerical variations in Aurelia tend in two
directions, i. e. to forms with six and twelve segments instead of the
normal echte

The same tendeney in Aurelia to vary in the direetion of six and

142 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.
twelve segments was pointed out by Brown, and was also noted in the
older papers of Ehrenberg and Romanes.

According to Romanes, monstrous forms of Aurelia aurita are of fre-
quent occurrence. Abnormality consisted in multiplicity and abortion
of parts. All cases of asymmetrical multiplication applied to litho-
cysts, and always occurred in the same inanner. When there were nine
instead of eight lithocysts, the extra one was always fully developed
and in close proximity to one of the normal lithooysts.

In symmetrical abnormalities all parts of the organism were equally
affected. Thus all examples of multiplication extended proportionally
to ovaries, nutritive canals, lithocysts, and tentacles, the effect being to
increase the number while adhering to the type of the natural segments.
In all cases the degree of abnormality was the same; e. g. 6 ovaries,
24 unbranched radial tubes, 12 lithocysts, and a six-lobed manubrium.
All parts and segments thus increased one third their normal number.
Romanes calls attention to the fact that this is the same proportional
increase as in Sarsia, with six canals, and explains it as accidental.
Supernumerary lithocysts always occur at the ends of the faintly
colored radial tubes, never at the ends of the darker ones.

Segments and lobes of the manubrium may be multiplied without the
ovaries increasing in number, Again, segments may multiply and manu-
brium and ovaries remain normal. Processes of multiplication may not
extend to all quadrants of the umbrella. Multiplication of parts may
be confined to one side of the umbrella, thus doubling or tripling organs
on one side only.

Abnormalities usually are symmetrical. When they are not, the
manubrium and ovaries are not affected, the segments only being multi-
plied. Abortion of parts takes place in the same symmetrical way
as multiplication : there may be one ovary and six segments, and three
ovaries instead of eight and four. Segments and ovaries may also be

reduced to one half the normal number. In these two cases the
manubrium is not affected. ^ Abortion of parts was observed in the
ovaries only. Partial suppression of ovaries was of frequent occurrence.
The most prevalent case was where one ovary was smaller than the other
three. Reduction also occurs in two alternating ovaries (i. e. opposite ).
Sometimes three adjacent ovaries were reduced in size.

Total suppression of one ovary was more rare. Only in twelve cases
in thousands was total suppression of two ovaries observed: some-
times it was two adjacent ones, and more frequently the two opposite

ones that were absent. In one case three ovaries were absent, the

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 143

specimen being otherwise fully developed. In no case was it observed
that deficiency or absence of ovaries entailed a corresponding deficiency
or absence of other organs. Reduction or suppression does not occur
in any other organ than ovaries in A. aurita.

H. C. Sorby found, among A. aurita collected in Suffolk and Essex, a
“few per thousand” abnormal specimens exhibiting sixfold, fivefold,
threefold, and partial twofold symmetry. References to variations in
Aurelia, Clavatella, Sarsia, and Stomobrachium may be found also
in Bateson's * Materials for the Study of Variation," pp. 421-429.
Edward T. Browne examined 383 specimens of A. aurita. He found
that eight specimens (2.08%) exhibited numerical variations in the geni-
tal sacs, buccal arms, and tentaculocysts. The number of the genital sacs
and of the buccal arms varied from three to six. He concludes that
there appeared to be a correlation between genital sacs and buecal arms,
but that the tentaculocysts vary independently of these. Eighty-seven
cases (22.8%) showed variation in the number of tentaculocysts.
Twenty of these had less, and the remainder more, thau the normal
number. The range of variation in tentaculocysts was 6 to 15.

The preceding observations on the variations of Aurelia show some
striking differences from those we have made on Eucope. While in
Aurelia there is a general correlation between the number of segments
of genital sacs, of buccal lobes, and of tentaculocysts, there is no such
correlation in the variations of Eucope. The sense organs in Eucope
vary, both in number and in position, irrespectively of the number of
radial canals and of segments. Neither multiplication nor abortion of
parts in Eucope is symmetrical. The suppression of genital sacs is quite
common in Eucope, while it is rare in Aurelia. In Eucope suppression
is not limited to genital sacs; as in Aurelia, it extends to the otolith-
bearing tentacles. As far as we have observed, the number of terminal
folds of the manubrium does not vary in Kucope, and is not correlated
to the number of segments.

The apparatus used in making the photographs was the large photo-
mierographie apparatus of Zeiss, with some modifieations, direet sun-
light being employed by means of a heliostat of the automatie kind,
and all exposures were instantaneous, The camera was always used in
the horizontal position, so that with an objective of low power the full
length of the bellows could be employed to obtain sufficient, magnilica-
tion with the least loss of light. The objectives employed in photo-
graphing Eucope were those of 35 and 70 mm. focus, the lower power

being employed with the larger specimens. The exposures were made

144 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

with a “ Low” pneumatic shutter, this model being chosen on account
of the small amount of space it occupies, allowing it to be introduced
between the camera and the microscope. This was accomplished by
clamping the shutter to the collar on the front-board of the camera,
another similar collar being screwed to the front of the shutter for the
light tight connection with the microscope. With the microscope in the
horizontal position the light was taken directly from the mirror of
the heliostat and diffused by means of a disk of blue ground-glass
placed in the substage immediately behind the iris diaphragm, and then
passes through a simple condensing lens to the object on the stage of
the mieroscope. The immediate source of the light, the ground-glass,
is thus brought near to the object to be photographed, giving a brilliant
illumination and permitting the use of a small diaphragm.

The most difficult task was to confine the animals to be photographed,
more particularly with the microscope in the horizontal position. The
device which proved most serviceable for flat or discoidal objects was a
parallel compressor of the model of Hermann Fol Rings were cut
from pure rubber tissue of different thicknesses, the ring to be em-
ployed for amy particular object being a little thicker than the object
itself. The rubber ring was then placed on the lower plate of the com-
pressor and pressed into contact with it by means of the finger. The
object is then brought into the centre of the ring, and water added with
a pipette until the inside of the ring is completely filled, and the upper
part of the compresser carefully screwed down until it comes in contact
with the rubber, the superfluous water being at the same time squeezed
out. If this be done with care, the inside of the ring will be com-
pletely filled with water and contain no air bubbles. There should be
just enough pressure to allow the upper glass of the compressor to come.
in contact with the object. This can be determined by holding the com-
pressor vertically, and screwing down the upper plate until the objeot
ceases to sink. The compressor can now be clamped to the stage of the
microscope in any position. By employing rubber rings of sufficient thick-
ness, aquaria can be contrived in this way one eighth of an inch in depth.
In photographing the rounder and plumper forms, amy pressure upon
tho animal would produce a sensible change in shape. Such forms,
therefore, were placed in small deep watch glasses and confined by
glass rings, the microscope being placed in a vertical position, the camera
however remaining horizontal. The connection between the microscope
and camera was effected by means of the prism end of an Oberhäuser's
camera lucida, to which the light tight collar had been adjusted by

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 145

means of an adapter. The light in this case was centred upon the
substage mirror of the microscope, and thus upward. Jellyfish that
have been for some time in a small quantity of sea water become par-
tially stupefied by the consumption of the air in the water, and are then
more quiet and their tentacles are better extended.

For larger objects, such as Medusi and Ctenophora, a Zeiss series
IL* 1:8 photographic lens with an iris shutter was made use of. A
reversing prism fastened to the front of the lens allows the use of a
horizontal eamera in photographing animals in open dishes. Work of
this description is done out of doors, illumination being obtained by a
series of mirrors, the arrangement of which varies with the nature of
the object and the view desired. The work is still in an experimental
stage, and it is hoped to give in a subsequent paper a more detailed
account of methods and results.

The following authors have noted variations in Acalephs :—

Agassiz, L.
Contributions to the Natural History of the Acalephs of North America.
Mem. Amer. Acad., Vol. IV. p. 248, Pl. IV. Fig. 4, Pl. V. Fig. 5. 1849.

Barsia.
Bateson, W.
Materials for the Study of Variation, pp. 424-429. London, 1894. Aurelia.
Brown, E. T.
Aurelia aurita: Numerical Variation. Nature, Vol. L. No. 1300, p. 524.
1894.
Brown, E. T.

On the Variation of the Tentaculoeysts of Aurelia aurita. Quar. Jour. Mier.
Sci., Vol. XXXVII. Pt. 3, pp. 245-251, Pl. XXV. 1895.
Brown, E. T.
On Variation of Haliclystus octoradiatus. Quar. Jour. Mior. Sci., Vol.
XXXVIII Pt. 8, pp. 1-8, Pl. I. 1890.
Claparede, Ed.
Beobachtungen über Anatomie und Entwicklungsgeschichte Wirbelloser Thiere,
p. 5. Leipzig, 1803. — Eleutheria.
Ehrenberg, C. G.
Ueber die Akalephen des rothen Meeres und den Organismus der Medusen
der Ostsee. Abhandl Akad. Wiss. Berlin, 1835, pp. 199-202, Taf. I., II.
Berlin, 1837. Aurelia, Stomobrachium.

146 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY,

Filippi, F.
Sopra due Idrozoi del Mediterraneo. Mem. R. Acad. Sei. Torino, 9d ser.
Vol. XXIII. pp. 375-383. 1866. Bleutheria.
Forbes, Edw.
Monograph of the British Naked-eyed Medus, p. 65. London, 1848. Sto-
mobrachium.

Herdman, W. A.
Pentamerous Aurelia. Nature, Vol. L. No. 1996, p. 426. 1894.

Hincks, T.
British Hydroid Zoöphytes, p. 71. 1868. Clavatella.

Hornell, J.
Abnormalities in Halielystus (Lucernaria) octoradiatus. Preliminary Note.
Natural Science, Vol. III. p. 33. 1893.
Hornell, J.
The Lucernarians as Degenerate Scyphomedus®. A Note upon the Phylogeny
of the Order. Natural Science, Vol. III. p. 204. 1893. Haliclystus.

Krohn, A.
Beobachtungen über die Fortpflanzung der Eleutheria, Quatref. Arch. für
Naturg., Jahrg. 27, Bd. I. pp. 157-172. 1861.
Romanes, G. J.
An Account of some New Species, Varieties, and Monstrous Forms of Medusæ.
Jour. Linn. Soc., Vol. XII. p. 528. II. Vol. XIII. p. 190, Pls. XV. and
XVI. 1874-76. Aurelia, Sarsia, Stomobrachium, Tiaropsis.
Sorby, H. C.
Symmetry of Aurelia aurita. Nature, Vol. L. No. 1298, p. 476. 1894.

Unthank, H. W.
Pentamerous Aurelia. Nature, Vol. L. No. 1295, p. 413. 1894.

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 147

EXPLANATION OF PLATES.

Plates I. VI., from photographs taken by W. McM. Woodworth.

PLATE I.

To illustrate the variation in the development of the genital organs. The diame-
ter of the disk of the specimens of Eucope figured varied from 3.5 to 4.5 mm.

Fig.
Fig.

Fig. 9
Fig.

Fig.
Fig.

Fig.

Fig.

Fig.
Fig.

Fig.

Fig.

Adult male, with equally developed genitals, their formula being 1, 1, 1, 1,

Adult male, with only three fully developed genitals, their formula being
1, 1, 1, 4,

Male, with the genital formula 1, 2, 3, 4, all in different stages of
development.

Female with two atrophied genitals, formula being 1, 1, 0, 0,

Male of the formula 1, 1, 3, 4,

Male with only one large genital pouch, with the formula 1, 4, 4, 4,

PLATE II.

Male, with two atrophied genital organs, and with unequally developed
radial canals, three of which have spurs running at a slight angle
from them.

Male, with unequally developed genitals, and with one radial canal forking
both above and below the genital pouch.

Young Eucope with undeveloped genital orgaus.

Male Eucope with spurs on two of the radial canals, with unequally de-
veloped genitals.

Female with only two fully developed genital pouches, and one atrophied
radial canal.

Male with genital formula 1, 1, 1, 0,

148

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.
Fig.
Fig.“

Fig.

Fig. 1.
2.

3.

BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

PLATE III.

Figures of Eucope having three or more radial canals, and forking either above

or below the genital organs.

Eucope with three radial canals, one of which forks at the extremities of
the heart-shaped genital pouch.

Male Eucope with four radial canals, also forking at the genital pouch.
Nearly equally developed genitals.

Female Eucope with three radial canals, one of which sends off a prom-
inent spur above therudimentary genital pouch. Genitals very un-
equally developed.

Male with three radial canals, one of which forks well above the genital
organs of the branches. Genital organs nearly equally developed.

Male with four radial canals. The genital organs corresponding to the
forks of one of the canals are barely united. Genital pouch atrophied
in one quadrant.

Male Eucope with three unequally developed genital organs.

PLATE IV.
With figures of Eucope having five or more radial canals.

Female Eucope with five radial canals and the same number of un-
equally developed genital pouches, with the formula 1, 2, 8, 4, 5,
Young male Eucope with five radial canals, unequally developed genitals,
and a sixth rudimentary radial canal.

Female Eucope with five radial canals, two fully developed genital
pouches, one less so, and two radial canals without genital organs.

Female Eucope with five radial canals, and only one fully developed gen-
ital pouch.

Female Eucope with five radial canals, one of which is atrophied below
the genital pouch. Genital organs unequally developed.

Female Eucope with five radial canals and unequally developed genitals,
one of the canals having no genitals, and forking.

PLATE V.

Male Eucope with five radial canals and five nearly equally developed
genital pouches

Male Eucope with five radial canals and traces of a sixth, and very un-
equally developed genitals.

Female Eucope with five radial canals subdividing the disk into very
unequal segments, only two of the genital organs fully developed.
Female Eucope with unequally developed genitals, but with five very
symmetrical segments.

AGASSIZ AND WOODWORTH: VARIATIONS IN EUCOPE. 149

Fig. 5. Female with five radial canals, one of which forks at the periphery, and
five very unequally developed genitals.

Fig. 6. Male with five radial canals, unequally developed genitals, and segments
of unequal size.

PLATE VI.

Fig. 1. Male Eucope with five radial canals, one of which forks above the genitals
close to the base of the digestive cavity, dividing the disk into six
different-sized sectors.

Fig.2. A male Eucope similar to Figure 1, with more fully developed genitals.

Fig. 8. Young Mnemiopsis Leidyi A. Ag., magnified 3.

Fig. 4. Doliolum sp. in profile, magnified %.

Fig. 5. Annellid larva (Aricidea !?), magnified 190,

Fig. 6. Ectopleura ochracea A. Ag., seen in profile, magnified .

PLATE VII.

To illustrate the formation of spurs, branches, and anastomosing canals from the
sides of the radial canals.

Fig. 1. Slightly projecting spurs on side of radial canal.

Figs. 2, 3, 6, 7, and 10 show quite prominent hook-like lateral prolongations from
the sides of the radial canals.

Figs. 4, 5, and 9 show forks of the radial canals in Figure 4 above the genitals, in
Figure 5 below.

Fig. 4 shows a transverse canal connecting adjoining radial canals.

Fig. 8 shows a circular canal connecting the radial canals below the base of the
manubrium.

PLATE VIII.

Fig. 1. Marginal sense-bearing tentacle with three otoliths.

Figs. 2 and 3 show the coalescence of adjoining marginal tentacles, Figure 2 with
one otolith, Figure 3 with two.

Figs. 4-18 show the mode of formation of an abnormal basal spur frequently
seen jutting out from the marginal tentacles, either with or without
otoliths.

Figs. 14-17 show the numerical variation of the otoliths in the sense organ of the
marginal tentacle.

Fig. 18 shows the coalescence of male genital organs of adjoining radial canals.

Fig. 19 shows the formation of a radial canal shooting up from the circular canal.

Fig. 20. An abnormal Eucope with confluent radial canals and ovaries developed
on the main and the lateral eanal.

Fig. 21 shows lateral leaf-like expansions of male genital organs which have not as
yet run together, as in Figure 18.

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

PLATE IX.

To illustrate the irregular development of the marginal tentacles belonging to
the ti, £5, and ts cycles.

Figs. 1-6, 10, and 14 are normal sectors belonging to the third stage of development.

Figs. 7-9, 12, 13, 16-19, and 22-26 show sectors belonging to the same stage of de-
velopment, but there has been great irregularity in the succession and
growth of the third cycle of tentacles.

Figs. 20-22, 27, and 28 show sectors with only one tentacle of the tg cycle in each
primary division.

Figs. 29 and 30 belong to the second stage of development .

EUCOPE

WOODWORTH

PHOTO,

PLATE h

ARTOTYPE, E. BIERSTADT, N. v.

EUCOPE

NI ieni
VOS ce tibi ge

WOODWORTH PHOTO,
ARTOTYPE, f. BIERSTAOT, N. Y

RUCOPE duds

NM

LEERE

— oi

Aura kan, y
m

WOODWORTH PHOTO.

EUCOPE PLATE IV.

WOODWORTH PHOTO.

ARTOTYPE, k. BIERSTADT, Ny v.

EUCOPE

WOODWORTH

PLATE

PLATE VI,

| EUCOPE

P
fien ras

ARTOTYPE, E, BIERSTADT, N, Y

WOODWORTH PHOTO

wu | EC /
| | ji | |
| /
ü | P | | : | | | 5 N
M | | | | | | | / ie | | | | | n1 // ; x] | \ | | i
a. a a | | | | A \ i4 uu
| | Ley \ ». | | X M | X 7 i /

| | |
mE N | \ Y $
\ f I be N
\ / T | | | 3 i * | | ;
NI | | I i, 4 Po ? a] | \ >
V !
| ri [| | uet |

Bulletin of the Museum of Comparative Zoölogy
AT HARVARD COLLEGE.
Vor. XXX. No. 3.

REPORTS ON HE RESULTS OF DREDGING, UNDER THE SUPERVISION
OF ALEXANDER AGASSIZ, IN THE GULF OF MEXICO AND THE
CARIBBEAN SEA, AND ON THE EAST COAST OF THE UNITED
STATES, 1877 TO 1880, BY THE U.S. COAST SURVEY STEAMER
“BLAKE,” LIEUT.-COMMANDER C. D. SIGSBEE, U. S. N., AND COM-
MANDER J. R. BARTLETT, U. S. N., COMMANDING.

[Published by Permission of CARLILE P PATTERSON and W. W. DUFFIELD, Superinten-
dents U. S. Coast and Geodetic Survey.]

XXXVII.

SUPPLEMENTARY NOTES ON THE CRUSTACEA.

Br WALTER Faxon.

Wıru Two PLATES.

CAMBRIDGE, MASS, U.S. A.:
PRINTED FOR THE MUSEUM.
NovEMmBeEr, 1896.

No. 3: — Reports on the Results of Dredging, under the Super-
vision of ALEXANDER AGASSIZ, in the Gulf of Mexico and the
Caribbean Sea, and on the Hast Coast of the United States,
1877 to 1880, by the U. S. Coast Survey Steamer “ Blake,”
Liewl.- Commander. C. D. SIGSBER, U. S. N., and Commander
J. R. BARTLETT, U.S. V., Commanding.

[Published by Permission of CAR P. PAvTERSON and W. W. DUFFIELD,
Superintendents U. S. Coast and Geodetic Survey.]

XXXVII.

Supplementary Notes on the Crustacea, By WAUAjn Faxon.

Tue following notes were made while identifying some of the“ Blake"
Crustacea that were retained as “duplicates” when the bulk of the
collection was sent to A. Milne Edwards in Paris, and some (Macrura)
that were returned by Milne Edwards undetermined. The notes chiefly
consist of hitherto unpublished locality records, which add something
to our knowledge of the distribution of many species. They also in-
clude descriptions of six new species (five Macrura and one Schizopod).
Detailed lists of the dredging stations occupied by the “ Blake " will be
found in the Bulletin of the Museum of Comparative Zodlogy, Vol. VI.
No. 1, and Vol. VIII. No. 4.

DECAPODA.
Anamathia hystrix (Srrwrs.).

Station 300. 82 fathoms. 14d.

Anomalothir furcillatus (Srrwrs.).

Station 159. 196 fathoms. 1 9.
Off Port Royal, Jamaica. 100 fathoms. 19.
VOL, XXX. — NO. 3.

154 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Pericera cornuta ceelata (A. M. Epw.).

Station XX. 50 fathoms. 2 specimens.

Picroceroides tubularis Mirrs.

Station XXI. 33 fathoms. 1d.
The rostral horns and preocular spines are longer than in the male speci-
men figured by Miers.

Lambrous pourtalesii Srrurs.

Station XXX. 51 fathoms. 2 d.

„ ee
Neptunus (Hellenus) spinicarpus (Srrurs.).

Station 149. 60 to 150 fathoms. 1 9.

Achelous spinimanus (Larr.).

Station 144, 21 fathoms, 29.

Calappa flammea (Hynnsr).

Station 144. 21 fathoms. 14,19.

Acanthocarpus alexandri Srimps.

Station 148. 208 fathoms. 1 4,19.
* 149. 60 to 150 fathoms. 1 d.

Myropsis quinquespinosa Srimprs.

Off Port Royal, Jamaica. 100 fathoms. 1 d.

Iliacantha subglobosa Srımrs.

Station X. 103 fathoms 19.

Cyclodorippe antennaria A. M. Epw.

Station 238. 127 fathoms. 1 9.
„% 246. 154 “ 29.
« 274. 209 id. Te.

FAXON: NOTES ON THE CRUSTACEA.

Iconaxius caribbaeus, sp. nov.
Plate I. Figs, 1-4.

Similar to Iconaxius acutifrons Date, but different in the form of the ros-
trum, which is much broader than in J. acutifrons, less triangular in its out-
line, and broadly rounded at the anterior end; the upper border of the
propodite of the larger cheliped, moreover, is entire, not dentieulate as in
I. acutifrons. The eyes are larger, and more heavily pigmented.

The margins of the rostrum are minutely denticulate, as in J. acutifrons,
the median keel entire.

Length, 17 mm.

Station 166. 150 fathoms. I specimen.

L1 232. 88 « l “
% 241. 163 RR
C 1 (type).

Lives as a commensal in Sponges of the genus Farrea.

The genus Zconaatus, of which four species have been previously described,
has a wide distribution in the warm and temperate seas, It has been recorded
from such remote localities as the Arabian Gulf, Banda Sea, Japan, Kerma-
dec Islands, and the Gulf of Panama. It is now for the first time recorded
from the Atlantic.

Polycheles crucifer (W.-S.).

Station 29. 955 fathoms. 3 specimens,
» 135. 450 * 1 "
« 179. Bot u 1 (exuvie),
" 180. gog -of l specimen,
" ie Isle = 1 “a
&g 188. 372 « 1 e
4 190. 542 ac 1

Polycheles agassizii (A. M. Epw.).

Station 129. 314 fathoms, 3 specimens.
4 153. 303 K l *
di 238. 127 * l 2d
2 260. 291 = l AF
Wo XS Vi 3997 ü 1 t

Polycheles sculptus Smitu.

Station 911.

e

357 fathoms.
573

464

3 specimens.
l
« 1

227. “ ji

80. 5

156 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Station 245. 1,058 fathoms. 1 specimen.!
d 257. 553 Su 2 =
„ 265. 576 “ CHIC
a 268. 955 ? 1
V VALE 610 d 1 f^
EEY UA 600 “ 2

Nephropsis agassizii A. M. Epw.

Nephropsis agassizii A. M. Edw., Ann. Sci. Nat., Zool., 6° sér, Vol. IX. No. 2,
1880.
Station 195. 5024 fathoms. 19.
200, 472 a 12.
Boe Oe E Loch,

Nephropsis aculeata Smitu.

Nephropsis aculeata Smith, Proc. U. S. Nat. Mus., Vol. III. p. 431, 1881.

Nephropsis agassizii Smith, Bull. Mus. Comp. Zool, Vol. XV. p. 44, Fig. 240,
1888 (nec A. M. Edw.).

Nephropsis rosea (W .-Suhm MS.) Bate, Rep.“ Challenger” Macrura, p. 178, Fig. 89,
PL XXIII. Fige. 1 % PL XXIV, Fig- ip 1888:

Station 185. 333 fathoms. 3 specimens.

es a Ert
% 889, 499 |." gu
4% 226. 424 pun
% 230. 464 “ bor 5
s abl. dandi sole 2

There are two species of Nephropsis in the West Indian region, N. agassizii
A. M. Edw., with two pairs of lateral spines on the rostrum, and N. aculeata
Smith, with only one pair of rostral spines. V. agassizii was very inade-
quately. described by A. Milne Edwards, and the type specimen, from the
Strait of Florida, 1,500 metres, has never been returned to Cambridge. Soon
after, the other species, N. aculeata, was described by Smith from specimens
obtained off the south coast of New England, in 100 to 126 fathoms. Subse-
quently Smith and other authors supposed that N. aculeata was identical with
N. agassizii. The chief differences between the two species are the following.
In N. agassizii the rostrum is armed with two or two and a half pairs? of
lateral teeth; in V. aculeata there is only one pair of lateral rostral spines ;

! Tdentified as P. agassizii by A. Milne Edwards, and so recorded by him in
Bull. Mus. Comp. Zoöl., Vol. VIIT. p. 66, 1880.

2 The third lateral spine may occur on either the right or the left side of the
rostrum,

FAXON: NOTES ON THE CRUSTACEA. 157

the shell is less coarsely granulated, but more spiny in the former species than
in the latter; the two lines on the proximal half of the rostrum in both species,
widely diverging as they pass backward over the gastric area, are marked by
small tubercles in N. aculeata, by distinct acute spines in N. agassizii ; the
top of the small median tubercle on the gastric area is truncated in N. acu-
kata, while in N. ag«ssimó it is bluntly triangular, passing into a slight
median longitudinal carina both in front and behind ; the abdominal pleure
are produced into longer spines in N. agassizii than in N. aculeata, and the
spines moreover trend more distinetly backward, forming a stronger angle with
the vertical axis of the pleura; the outer surfaces of these pleure are quite
smooth in N. agassizii, while in N. aculeata they are conspicuously granulated
both on their margins and on the distinctly raised central field ; the lateral
borders of the abdominal terga, which form a festoon on each side of the ab-
domen, are more strongly convex in the former species; another distinction is
apparent in the sixth abdominal somite, viz. in N. aculeata the antero-lateral
margin of the pleura is shorter than the postero-lateral border, whereas in the
other species the antero-lateral border is longer than the postero-lateral; the
tergum of this somite in N. aculeata sends off a granulated ridge from near its
posterior lateral angles, — a ridge which runs forward into the upper, depressed
portion of the pleura ; this ridge is not found im N. agassizii.

Nephropsis rosea Bate is without much doubt a young individual of N. acu-
lata. N. atlantica Norman! is very similar to V. agassizii, but has a sharp
spine on the anterior margin of the second abdominal pleura.

Stenopus hispidus (Ov.).

Station 11. 37 fathoms. 1 specimen.

" DECUS m l 7
1 36. 84 = 1 ge
= 132. 115 sy 2 t

Pontophilus gracilis Smits.

Station 43. 339 fathoms. I specimen.

MC AAT BR cree th: acoge got
F E tks
eal, ee join

Prionocrangon pectinata, sp. nov.
Plate II. Figs. 4-7.
Rostrum spiniform, inelined at an angle of 45? to the axis of the body.
Median dorsal line of the carapace armed with a row of eight spiniform teeth,
Proc Roy. Soc. Edinburgh for 1881-82, p. 684; Wood-Mason and Alcock,
Ann. Mag. Nat. Hist., 6th Series, Vol. VII. p. 197, Fig. 4, 1891.

158 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

which exiends backward nearly to the posterior border of the carapace. Antero-
lateral margins of the carapace angulated below the orbit. Telson much
shorter than the appendages of the sixth abdominal somite, broad, with a pair
of dorsal longitudinal ribs, abruptly contracted a short way beyond the mid-
dle; tip truncate, setiferous.

The eyes are absent ; their peduncles are transformed into a pair of closely
apposed trihedral processes, with acute and somewhat divergent tips. The
first segment of the antennule is very long, reaching nearly to the end of the
antennal scale; the second and third segments are, on the other hand, very
short, the third bearing two flagella, the outer of which is very much shorter
than the inner. 'The antennal scale is long and narrow, its outer margin
lightly concave.

Length, 28 mm.

Station 201. Off Martinique. 565 fathoms. 1 9.

The rostrum is proportionally smaller than in P. ommatosteres Wood-Mason,!
while the dorsal teeth of the carapace are larger, more numerous, and extend
farther back on the cephalothorax ; the telson is shorter; the antennal scale is
longer than the proximal segment of the antennule. According to Wood-
Mason, there is no trace of eyes or eye-stalks in P. ommatosteres. In P. pecti-
nata there are distinct rudiments of the eye-stalks, as above described.
P. ommatosteres comes from the Andaman Sea, 405 fathoms, and the Bay of
Bengal, 200-350 fathoms.

Glyphocrangon aculeata A.M. Epw.

Station 29. 955 fathoms. ^ I specimen.
* 103. 769 . 2
„ 115 3/5 n d
“ 190. 542 E 4 5
„ 195. 5025 1
% 260, D/R 4 2
ee Valid, s O10, | 1 d

5

Glyphocrangon spinicauda A. M. Epw.

Station 148. 208 fathoms. 2 specimens.
* 274. 209 « 19 u
„ 275. 218 pite
p 281. 288 M vé 00

1 Ann. Mag. Nat. Hist., 6th ser., Vol. VIII. p. 362, 1891; Journ. Asiatic Soe.
Bengal, Vol LXIII. p. 152, 1894; Ill. Zool. R. I. M. S. Investigator," Crust., Pl.
IX. Fig. 4, 1895.

FAXON: NOTES ON THE ORUSTACEA. 159

Glyphocrangon nobilis A. M. Epw.!

Station 41. 860 fathoms. 6 specimens.

He 080. 1590] s 2 e
162. 734 b: 2 *
fee g, 88 H 4 i
eral Ow «981 ff l young.
6% 179. 824 b 1 specimen.
10% 85, 3888 k 7 K
[11 357 € r

211. 357 " 1 young.
* 222. 422 s 2 v
BET n 1 specimen.

Glyphocrangon neglecta, sp. nov.
Plate I. Figs. 5, 6.

Rostrum longer than the rest of the carapace, trending a little downward for
the anterior half of its length, then curving gently upward to the tip, which is
slender and acute; the anterior half of the rostrum is distinctly carinated in the
median line, but the carina fades away before attaining the base of the rostrum ;

1 The Glyphocrangon doubtfully referred to G. nobilis in my Report on the Stalk-
eyed Crustacea of the“ Albatross ” Expedition of 1891 (Mem. Mus. Comp. Zoöl., Vol.
XVIII. p. 142, 1895) is distinct from G. nobilis, as appears from an examination of a
larger number of specimens of the latter species. In the * Albatross " species, which
may be called Glyphocrangon vicaria, the upper surface of the rostrum is corrugated
on each side of the median carina, in front of the anterior pair of lateral spines; in
G. nobilis this corrugation does not exist. In G. vicaria the anterior moiety of the
fourth or lateral crest of the carapace is broken into two parts by a deep notch;
the part in front of the noteh is produced anteriorly to form a strong spine, while
sthe part behind the notch merely forms a projecting angle or shoulder; in G. nobili
the anterior moiety of the fourth crest is continuous from the posterior end to the
anterior spine. The tubercles of the first and second crests are more prominent and
spiniform in C. vicaria than in G. nobilis. The dorsal carina of the telson are den-
tate anteriorly in C. vicaria, simple in C. nobilis. C. vicaria is even more closely
related to G. longirostris Smith, which it represents on the Pacific side of’ the
American continent, These are the chief differences between the two species: the
rostrum, corrugated above in both species, is narrower in front of the anterior lat-
eral spines in G. vicaria than in G. /ongirostris. The anterior moiety of the fourth
lateral carina is broken into two distinct parts by a notch in the former, while it is
merely sinuate in its outline in the latter. The tubercles on the first and second
erests of the carapace are more prominent and spiny in the former than in the
latter. The median dorsal crest of the abdomen, moreover, is more prominent,
These differences, though very small, appear to be constant, and afford another
instance of a slight divergence between two representative forms on the Atlantic
and Pacific sides of the American continent. The type specimens of G. vicaria were
„Long. 91° 0” W., Albatross " Station 3411.

dredged in 1189 fathoms, Lat, 0° 54

160 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

there are two pairs of lateral rostral spines, one of which lies in advance of the
eyes, the other just behind the posterior wall of the orbit ; on the lower face
of the rostrum there appears just the slightest trace of a median longitudinal
carina. The upper surface of the first or dorsal pair of carinæ is eroded ; be-
hind the cervical groove this pair of carint converge towards one another. Just
in front of this pair of carinæ, lying in the median line at the base of the ros-
trum, is a small tubercle or papilla. In the interval between the first and second
caring on each side are about four faint tubercles on the cardiac region, and on
each side of the gastric region are four larger low tubereles, the hindmost of
which is the largest of all. The anterior moiety of the third carina (adopting
Wood-Mason's terminology) is well developed as a backward prolongation
of the external orbital spine, which is long, acute, and inclined outward and
upward. The fourth carina is also developed both anteriorly and posteriorly
to the cervical groove, its anterior moiety being continuous with the antero-
inferior, or branchiostegian, spine of the carapace, Barring the external orbital
and branchiostegian spines, the anterior moieties of both the third and fourth
carin are entire, without a trace of spine or tooth. The trend of the branchi-
ostegian spine is nearly straight forward, its downward and outward deflection
being very slight. With the exceptions noted above, the spaces between the
carin of the carapace are pretty smooth.

The abdomen is lightly sculptured for the genus to which this species be-
longs. Only the first and sixth segments are conspicuously carinated above.
The pleurm of the second abdominal segment are one-toothed. The telson
exceeds the last pair of abdominal appendages, and is rather abruptly bent
upward at the tip.

Length, 75 mm.; cephalothorax including rostrum, 35 mm.; rostrum, 19 mm.;
telson, 13 mm.

Station 261, off Grenada. 340 fathoms. 1 9 with eggs. Type.
* 1353, off Montserrat. 303 " E
„2060, off Grenada. 291 y 1 young.

This species is peeuliar in having the anterior moiety of the third and fourth
carinte of the carapace well developed and continuous with the external orbital
and branchiostegian spines respectively. In G. gilesii Wood-Mason, which
also has the anterior portion of both the third and fourth crests developed,
these crests are produced anteriorly into small spines independent of the
external orbital and branchiostegian spines.

Stylodactylus serratus A. M. Epw.
Station 205. 334 fathoms. 3 specimens.
* 151. 356 Li l lis
Pantomus parvulus A. M. Epw.

Station 134 248 fathoms. 2 specimens,

FAXON: NOTES ON THE CRUSTACEA. 161

Pandalus longipes A. M. Epw.

Station 274. 209 fathoms. 124- specimens.

© 8005. S05 —" là.» e
«^ $95, 180 2 “
ic^ URS gara on 12^ “

Pandalus ensis A. M. Epw.
Station 208. 213 fathoms. I specimen.
% 258. 159 jid 2 "
Pandalus leptocerus Smrru.

Station 345. 71 fathoms. I specimen.

Heterocarpus levis A. M. Epw.

Station XXVI. 297 fathoms. I specimen.

Heterocarpus alexandri A. M. Epw.

Station 196. 1030 fathoms. I specimen.

Heteroearpus ensifer A. M. Epw.
Station 146. 245 fathoms. 1 specimen.
" 158. 303 s l d.
* 958, 1598 M 3 *
Nematocarcinus cursor A. M. Epw.

Station 151. 356 fathoms. 12+ specimens.

5 160. 393 d 2 "
" 161. 583 5 l a
“ T200, GY E 2 M
€ eA TOTA Y 2 M
" raTa., 200 : 1 T

Hoplophorus gracilirostris A. M. Epw.

Station 100. 250-400 fathoms. 1 specimen.
" 191. 108-250 á 1 we
" 226. 424 Hd l ”
is 930. 464 T l 5
Y 958. 159 M | "
« 971. 458 « pos

162 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Acanthephyra affinis, sp. nov.
Plate II. Fig. 1-3.

Similar to Acanthephyra (Systellaspis) lanceocaudata Bate, but different in
the following regards : the apical tooth of the antennal scale projects forward
far beyond the membranous part of the organ; the telson is shorter than even
the inner branches of the posterior pair of abdominal appendages, and its dorsal
surface is flattened, but not grooved.

The seven teeth that surmount the gastric crest are closely approximated,
and increase in size successively from the first to the fifth. The sixth is about
equal to the fifth, the seventh a little smaller. The ege of this species measures
3 X 2mm.

Length, 100 mm.

Station 258. 159 fathoms. 19.

This species belongs to the subgenus Systellaspis, in which the orbit is con-
tinuous to the first antennal tooth (the orbital tooth being absent), the dorsal
carina of the sixth abdominal somite is wanting, and a prominent angle or tooth
projects from each side of the anterior border of the first abdominal somite,
overlapping the posterior margin of the carapace. The eggs, moreover, are of
large size, indicating a protracted period of intra-oval development.

Acanthephyra debilis A. M. Epw.
Station 107. 428 fathoms. 1 specimen.
Acanthephyra armata A. M. Epw.
Station 135. 450 fathoms. 1 specimen.

" 151. 356 $ 2 *

Sicyonia edwardsii Miers

Station 142. 27 fathoms. 1 specimen.

Sicyonia brevirostris Srrurs.

Station 38. 20 fathoms. 1 specimen.

Peneus brasiliensis Larr.

Station 27. 35 fathoms. 2 specimens.
2
Lu

2, 90b Mu

young.

FAXON: NOTES ON THE ORUSTACEA.

Parapeneus megalops Swrrn.

Station 147. 250 fathoms. 4 specimens.

9" 185 7208 d 4 m
wes LOB m 6 T
E RS si 4 T
„% 281 8 uc 10 d
* 288. 287 Ji L i

Parapeneus politus Swirn.

Station 36. 84 fathoms. 27 specimens.

Haliporus debilis (Swrrn).

Station 47. 321 fathoms. I specimen.

Plesiopeneus armatus (Bark).

Station 31. 1,920 fathoms. 2 specimens.

Hob. 411 st 1 t4

Hemipeneus triton Fax.

Station 227. 573 fathoms. 1 specimen.

Benthesicymus bartletti Surrn.

Station 29. 955 fathoms. 1 specimen.
S 33. 1400-1568 x 1 "
“ 109. 769-878 8 2 "
De 824 * 1 "
* 1900. 549 a 1 "
UT META 573 “a 2 ^"
" 3245 1058 _ 1 -
* 965. 576 = 1 A
R . 399 x 2 d

Sergestes robustus Smita.

Station 205. 334 fathoms. 1 specimen.

Ya: Bor S 1 T
E 280. 901 i l 25
Oh. 410 " l E
Do WOOD Dr oT 2 8
„„ 1 j

163

164 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Sergestes mollis Surg.

Station 30. 968 fathoms. 2 specimens.

SCHIZOPODA.

Lophogaster longirostris, sp. nov.
Plate. II. Figs. 8-10.

Similar to L. typicus Sars, but different in the great length of the median
spine of the rostrum, which far surpasses the antennular peduncle, and almost
attains to the tips of the antennal scales. here are six teeth along the outer
edge of the antennal scale. Length, 27 mm.

Station 50. 119 fathoms. 20 specimens.

Gnathophausia zoéa W.-Sunm.

Station 185. 333 fathoms. 2 specimens.
“ 201. 565 = 1
* 221. 423 A 1 L
l
|
1

ec

„„ S y
eee 100 d D
* 230. 404 " i
„284. 347 g 2 di
„% 288. 399 d 3 A

Eucopia sculpticauda Fax.

Station 30. 968 fathoms. 1 specimen.

Station 29. 955 fathoms
This is the speeimen figured in my Report on the Stalk-eyed Crust:
the “ Albatross" Expedition of 1891, Pl. LIII. Fig. 2 (Mem. Mus. Com]

Vol. XVIII.).

FAXON: NOTES ON THE CRUSTACEA, 165

STOMATOPODA.

Squilla empusa Say.

Station 36. 84 fathoms, I specimen (young).

Pseudosquilla ciliata (Fanm.).

Martinique. I specimen.

ISOPODA.
Bathynomus giganteus A.M. Epw.

Station 179. 824 fathoms. I specimen, 157 X 80 mm.
Hn VII. 610 i: 1 S 107 x 49 *

According to Wood-Mason and Aleock (Ann. Mag. Nat. Hist., 6th Series,
Vol. VII. p. 270, 1891), this remarkable Isopod was captured in the Bay of
Bengal at a depth of 740 fathoms. Dr. Arnold Ortmann! has described a
second species of Bathynomus (B. daderleini), taken on the coast of Japan,
near Enoshima, Sagarui Bay. The depth is not recorded.

1 Proc. Acad. Nat. Sci. Phila., 1894, p. 191.

166

Fig. 1.
Vig. .
Fig. 3.
Fig. 4.

Fig. 5.

Fig. 6.

Ens
Fig. 2.
Fig. 3.
Fig. 4.
Fig. 5.
Fig. 6.
Pg T
Fig. 8.

Fig. 9.
Fig. 10.

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

EXPLANATION OF THE PLATES.

PLAITE:I

Iconaxius caribbeus Fax. M. C. Z., No. 4195. Blake Sta. 283. x 54.

The same. Head, from above. x 51.

The same. Right chela, from the outside. X 54.

Lconaxius caribbæus Wax. —'Telson and posterior pair of appendages.
M. C. Z., No. 4147. Blake Sta. 241. Much enlarged. 1

Glyphoerangon neglecta Fax. Female, dorsal view. M. C. Z., No. 4434.
Blake Sta, 201. X 14.

The same. Lateral view. X 14.

PLATE II.

Acanthephyra affinis Fax. Female. M. C. Z., No. 4410. Blake Sta. 258.
x M.

The same. Telson. X 14.

The same. Antennal scale. x 14.

Prionocrangon pectinata Fax. Female. M. C. Z., No. 4486. Blake Sta.
201. X 4.

The same. Carapace, from above. x 4.

The same. Chela. X 4.

The same. Telson and posterior pair of abdominal appendages. X 4.

Lophogaster longirostris Fax. M. C. Z., No. 4380. Blake Sta. 50. x 4.

The same. Carapace, from above. X 4.

The same. Telson and posterior pair of abdominal appendages. X 4.

PLATE

FAXON, BLAKE CRUSTACEA

r , , cs
7 ,, 4

Mi -

f

i

FAXON, BLAKE CRUSTACEA

Holotype Printing Co.

| 3 Acanthephyra affinis Fax

| 7 Prtonocrangon prelinala Fax

8-10 hophopaster longirostri | Fax

Bulletin of the Museum of Comparative Zoölogy
AT HARVARD COLLEGE.
Vou. XXX. No. 4.

ON THE COLOR AND COLOR-PATTERNS OF MOTHS
AND BUTTERFLIES.

Bv ALFRED GOLDSBOROUGH MAYER.

Wırn Ten PLATES.

CAMBRIDGE, MASS., U. S. A.:
PRINTED FOR THE MUSEUM.

Forruary, 1897.

No. 4.— On the Color and Color- Patterns of Moths and Butter-
flies.

By ALFRED Gorpsnonovan MAYER,

This research is an investigation of the general phenomena of Color
in Lepidoptera, and also a special account of the Color-Patterns of
the Danaoid and Acraeoid Heliconidac, and of the Papilios of
Tropical South America, and has been carried out under the direetion
of my friend and instructor, Dr. Charles D. Davenport ; and the work
was done in connection with one of the courses given by him in
Harvard University in 1894-95. I am indebted to Dr. Davenport
not only for suggesting the subject, but also for his kindness in devot-
ing much time to a criticism of the results.

The paper is divided into three parts. Part A contains an ac-
count of the general phenomena of color in Lepidoptera; Part B
is devoted to a special discussion of the color-variations in the ITeli-
conidae, with special reference to the phenomena of mimiery ; and
Part C consists of a summary of those results which are believed to
be new to science. A Table of Contents is given at the end of the

paper.
PART A.
GENERAL PHENOMENA OF COLOR IN LEPIDOPTERA.
I. CLASSIFICATION OF COLORS.

We follow Poulton (90) in dividing Lepidopterous colors into (1)
pigmental and (2) structural,

(1) Pigmental Colors are due to the presence of an actual pig-
ment within the scales, and although such colors are very common
in the Lepidoptera, it is frequently very difficult to say off-hand
whether a given color is due to a pigment or to some structural effect.
Coste (90-91) and Urech (793) have, however, given criteria for de-
termining whether a color is due to a pigment or to some other eause.
They sueceeded, for example, in dissolving out the eolor in many

1 Contributions from the Zoülogieal Laboratory of the Museum of Comparative Zoöl
ogy at Harvard College, E. L. Mark, Direetor, No. LXXIV,

2 This paper was written in 1895 essentially asit now stands.

10 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

cases, leaving the wing white or colorless. Coste used as solvents
a number of strong acids and alkalis; while Urech confined him-
self to the use of water, hydrochloric acid, and nitric acid. Their
results may be conveniently summarized as follows :—

Black according to Urech is a pigmental color, for it may be dis-
solved out of the wings by means of hydrochloric or nitric acid.

Brown is usually insoluble in water, but is soluble in hydrochloric
or nitric acid.

The red and orange pigments of the Pieridae, Lycaenidae,
Nymphalidae, Zygaenidae, and some Papilios are soluble in water.
They are insoluble in water in the Sphingidae, Arctidae, Bombycidae,
Saturnidae, and Geometridae.

Yellow pigment is acted upon by reagents in almost the same way
as the red and orange, especially if both red and yellow appear upon
the same wing. It is soluble in the Pieridae, Lycaenidae, Nym-
phalidae, Satyridae, and some Papilios, but insoluble in the Sphin-
gidae, Arctidae, Geometridae, and a few Noctuidae,

White is usually a structural color, but can be dissolved out
from the wings of the Pieridae by water, being in this case, of
course, due to a pigment.

Green pigment can be dissolved out by water in the cases of the
Pieridae, Lycaenidae, and Geometridae. In the vast majority of
cases, however, it is a structural color.

Violet and blue are almost always due to structural causes. Ina
few cases, however, as in Smerinthus ocellatus, a blue pigment can
be dissolved out.

We see, then, that black, brown, red, orange, and yellow are
usually due to pigment, while white, green, violet, and blue are gen-
erally due to structural effects.

It is well known that the scales of Lepidoptera are essentially
hollow, flattened sacs often inclosing pigment, and Burmeister (78)
arrives at the conclusion, from a study of the scales in various spe-

cies of Castnia, that the |

ment is for the most part attached to the
upper layer of the scale-sac, rendering it opaque, while the lower
layer receives less pigment and is, in consequence, a little more
translucent.

(2) Structural Colors owe their origin to the external structure of
the scales or wing-membranes and not to the presence of a pigment.
They are often caused by diffraction, due to the seales being covered

with fine, parallel striae. Some of the most splendid colors in the

MAYER: COLOR AND COLOR-PATTERNS. 171

animal kingdom are due to this cause; such are the iridescent and
opalescent hues of many of the Morphos and Indo-Asiatie Papilios.
Very often the scales which display such brilliant colors contain no
pigment whatsoever; for if one will merely soak them in alcohol,
ether, or water, all color disappears, and the scales become as trans-
parent as glass. ‘This test was devised by Dimmock (83), who
used it upon the brilliantly colored scales of many beetles. It
was first discovered by Burgess ('80), and has since been con-
firmed by Kellogg (94), that the striae which produce these structural
colors are all upon the outer surface of the scale, J. e., the surface
which is away from the wing-membrane and exposed to the light.
Kellogg (94) has determined the distance apart of the striae upon
the scales of many species of Lepidoptera. It appears, for example,
that the striae upon the scales of Danais plexippus are 2u apart,
those upon the transparent scales of Morpho sp. 1.54, upon the
pigment-bearing scales of Morpho 0.724, and upon Callidryas
eubule 0.9% apart. It is very evident, then, that the brilliant color-
ation of the scales may be due to this fine striation, for the striae
upon Rowland’s or Rutherfurd’s finest gratings are approximately
1.5 apart, which is about the average distance between the ridges
of the scales,

Structural colors are, however, not always due to diffraction; in
the case of white, for example, the color is almost invariably due to a
reflection of all, or nearly all, the light that impinges upon the scales.
As long ago as 1855 Leydig pointed out that the silvery white color
seen in the scales of some spiders, such as Salticus and Tegenaria,
was due to air contained within them ; and more recently Dimmock
(83) has shown that silvery white and milk-white colorations are
due to optical effeets produced by refleeted light. In the silvery white
scales, however, such as those of the under surface of the hind wings
of Argynnis, there must be a polished reflecting surface toward
the observer, for both silvery and milk white colors appear simply
milk-white by reflected light.

(3) Combination Colors owe their richness and brilliancy to a
combination of structural and pigmental effects. The geranium-red
spots upon the hind wings of the Mexican Papilio zeunis Lucas owe
their red color to pigment, but over this red there plays, in certain
lights, a beautiful pearly irideseence, which, in combination with the
red, greatly enhances its charm, Urech (92) has demonstrated that
in the Vanessas there are seales whieh have chemical eoloring matter

172 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

and interference colors also. In addition, he points out the interest-
ing case of certain Lycaenidae where the scales exhibit to the eye
only interference effects, and yet a pigment can be dissolved out of
them by the use of water.

(4) Quantitative Determination of Pigmental Colors. I have
analyzed the colors of many butterflies by means of the spectroscope,
and also by Maxwell’s discs. As is well known, Maxwell’s discs are
colored circular dises of cardboard, perforated at the center and slit
along a radius so that, two or more of them may be slid over each
other, thus exposing different proportions of each, Then by rapidly
rotating them the colors become blended, and thus it becomes
possible to match any color, and to discover its fundamental con-
stituents. By this means I have determined that the vast majority
of the colors found in Lepidoptera are impure; that is to say, they
contain a large percentage of black.

For example the white of the upper surface of the wings of the
common Pieris rapae consists of: 17% black, 18% emerald-green,
10% lemon-yellow, and 60% white,

Also the so-called “ blacks” found in butterflies are rarely jet-black,
but, almost always, only deep shades of brown. For instance the
deep brown color of the under surface of the wings of Heliconius
melpomene consists of 93% black, 8% lemon-yellow, 3.5% of
Maxwell’s fundamental red (vermilion), and 0.5% of von Bezold’s
fundamental blue-violet.

The purest color I have met with is the canary-yellow ground
color of the wings of Papilio turnus, which seems to consist of
white light with the addition of a little yellow.

Other colors all possess considerable black. Thus the glaucous
green of Colaenis dido consists of black 29%, vermilion 24%,
emerald-green 37%, von Bezold’s blue-violet 10%.

The sepia-brown ground color of Cercyonis alope consists of black
71%, vermilion 21.5%, emerald-green 7.5%.

The tawny rufous color of the wings of Mechanitis polymnia, ete.,
is made up of black 46%, vermilion 40%, lemon-yellow 14%.

The rufous red patch on the upper surface of the fore wings of
Ifeliconius melpomene is made up of black 27%, vermilion 66.5%,
lemon-yellow 6.59%.

The yellow of the fore wings of Mechanitis polymnia consists of

lemon-yellow 67%, emerald-green 14%, and white 19%.

MAYER: COLOR AND COLOR-PATTERNS. 178

(5) Spectrum Analysis of Colors of Lepidoptera. I have made
some spectrum analyses of the light reflected from the wings
of various butterflies, by means of a piece of apparatus most kindly
suggested for the purpose by Prof. Ogden N. Rood of Columbia
Jollege. The arrangement is shown in Figs. 1, 2, Plate 1; Fig. 1
being a perspective view, and Fig. 2a horizontal seetion of the
apparatus, which consists of a rectangular box, blackened upon the
inside, and having a wellfitting cover. A rectangular slit (O) was
cut through one of the long sides of the box, near one end, and the
other end of the same side was perforated in order to allow the
admission of the direct-vision spectroscope (S). Imagine that we
wish to examine the yellow spots from a butterfly’s wing. All of
the yellow spots from the wing are cut out, and pasted upon two
pieces of cardboard so as to make two large unbroken patehes of
color. The pieces of cardboard are then blaekened upon all those
places where the eolored wing was not pasted. One of the card-
boards is then suitably mounted upon the back of the box at B; the
other is placed upon a vertical support (F), the plane of which is
parallel to the back of the box.

The working of the apparatus is as follows: the sunlight enters
by the slit (O) and is reflected and diffused three or four times
between the pieces of colored wing mounted upon the back (B) of
the box, and the vertical support. (F). The manner of this reflec-
tion and diffusion is shown by the dotted lines of Fig.2. After
undergoing several reflections, the light enters the direct-vision
spectroscope (S). The slit of the spectroscope is wide open, and
thus the light which enters it may readily be examined. It was
found that it was necessary that the light be reflected more than
once from the wing before it enters the spectroscope, for the first
reflection shows so much white light that it is usually quite impossi-
ble to analyze the true color of the wing, the predominant colors
being obscured by a continuous spectrum. In general it was found
that the colors of the wings are not simple, but compound; that is
to say, they are made up of a mixture of several different colors.

For example, the spectrum of the rufous ground color of the
upper surface of the wings of Danais plexippus consists of all of the
red and yellow of the spectrum and about 75% of the green,

The red spots upon the upper side of the fore wings of Heliconius
melpomene also consist of the red and yellow and a very faint,

hardly visible, trace of green.

174 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The glaucous green patches on the wings of Colaenis dido are
composed mainly of green and yellow, but there is also a faint develop-
ment of about half of the blue and a still fainter trace of red.

The iridescent blue-green ground color of the upper surface of the
wings of Morpho menelaus, viewed in such a way that the light
makes an angle of about 20? with the normal to the surface of the
wing, gives a spectrum of green and blue about equally developed.

The yellow ground color found on the upper side of the wings of
Papilio turnus shows a continuous spectrum, in which the yellow
seems to be rather more brilliant than in the normal spectrum of
white light.

The sepia-brown ground color of the upper surface of the wings
of Cercyonis alope gives a spectrum which lacks only the blue-green
and blue.

(6) Summary of Results. The researches of Coste (9091) and
Urech (93) have demonstrated that the colors of butterflies and
moths may be produced by two causes: by the presence of an actual
pigment, or by some structural effect. Some colors are due entirely
to pigment, others to structural causes, and still others to a combina-
tion of the two.

Black, brown, red, orange, and yellow are invariably due to
pigment.

Green is usually due to a structural effect, but in a few cases there
is a green pigment present,

White, blue, and violet are almost invariably due to structural
causes.

In addition to these facts I have found that most of the colors
which are displayed by Lepidoptera contain a surprisingly large
percentage of black. Also they are usually not simple colors, but
composed of a mixture of several different colors. It is remarkable
that Natural Seleetion, which is generally assumed to have been one
of the principal factors in bringing about the wonderful develop-
ment of colors in Lepidoptera, has not been potent enough to make
these colors purer than is the case in existing butterflies.

II. Tur ESSENTIAL NATURE OF PIGMENTAL Cotor IN
LEPIDOPTERA.
(1) Pigments of Larvae. Poulton ('85) showed that the phy-
tophagous larvae of Lepidoptera * owe their colour and markings to

|

MAYER: COLOR AND COLOR-PATTERNS. 175

two causes: (1) Pigments derived from their food-plants, chloro-
phyll and. xanthophyll, and probably others; (2) pigments proper
to the larvae, or larval tissues made use of because of some (merely
incidental) aid which they lend to the colouring, e. g. fat.“ Poulton
concludes that all green coloration is due to chlorophyll, and
that nearly all yellows are due to xanthophyll. All other colors,
including black and white and some yellows, are due to pigments
proper to the larvae themselves.

Later, in 1893, Poulton proved that the larvae of Tryphaena
pronuba could transform both etiolin and chlorophyll into a larval
coloring matter, which may be either green or brown. It thus
appears that some brown pigments are derived from food, and are
merely modified plant pigments. Green larvae have green blood,
and this color is due to chlorophyll in solution. It is remarkable
that this chlorophyll solution is stable under the prolonged action of
light, and in this respect is different from any other known solution
of chlorophyll. It is worthy of note, further, that the spectrum
of this green blood shows a great resemblance to that of chlorophyll.
“In fact the two spectra are far nearer to one another than the
ordinary spectrum of chlorophyll in alcoholic solution, is to the
unaltered chlorophyll of leaves.”

2) Pigments of Imagines. In 1891, Urech showed that the
similarity between the color of the urine of butterflies and the
principal color of their scales is so close that it cannot be considered
as accidental, but rather must be regarded as physiological. Urech
compares in a table the color of the urine and that of the scales
of 29 species of Lepidoptera. In all but two species the resem-
blance is very close.!

Urech further shows that the color of the urine (and the corres-
ponding color of the scales) is not dependent upon the kind of food,
for one and the same food plant may be differently digested in
different groups of Lepidoptera. Thus he compares the behavior ofa
Vanessa with that of one of the Microlepidoptera (leaf-rollers). Both
of these feed upon the nettle (Urtica). In the larva of the Vanessa the
contents of the stomach are intensely green, but become red in the
pupa. In the ease of the leaf-roller the contents of the stomach are
never markedly green and become insipid in color during the pupal
stage.

! Likewise, Hopkins (94) has shown that in the Pieridae the urine is tinged by a yellow
substance having exactly the color of the wings.

176 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Poulton has shown that the reddish fluid voided by the Vanessas
immediately after emergence from the chrysalis contains uric acid,
and Hopkins (94) says that when the yellow Pieridae emerge, they
often void from the rectum a large quantity of uric acid. It should
be borne in mind however, as Urech himself suggests, that the pig-

0 > , 5
ment found within the wings may not be identical in chemical com-
position with the similarly colored fluid from the alimentary tract,

Hopkins (89, 91, 94, ’96) has discovered that the white pigment
found in the seales of Pieridae is uric acid, and that the red and
yellow pigments of the Pieridae are due to derivatives of uric
acid. Ile also says, “these uric acid derivatives used in ornamen-
tation, are apparently confined to the Pieridae alone among butter-
flies.” Hence when a Pierid mimics an insect of another family, the
pigments in the two cases are chemically quite distinct. This is well
seen in the genera Leptalis (Pieridae) and Mechanitis (Danaidae).

In addition to this, Griffiths (92) finds that the green pigment
found in Papilio, Parthenos, Hesperia, Limenitis, Larentia, Ino, and
Halias is a derivative of uric acid, to which he gives the name of
*Lepidopterie acid” and assigns the empyrical formula Ci H,, Az,
N; Of

In à paper published in 1896 in the Bulletin of the Museum of
Comparative Zoólogy at Harvard College, Vol. 29, I have shown,
p. 226-230, that the pigments of the seales of Lepidoptera are
derived by various chemieal processes from the blood, or haemo-
lymph, of the pupa, and that the haemolymph is a proteid substance
containing egg-albumen, globulin, fibrin, xanthophyll, orthophos-
phoric acid, iron, potassium, and sodium.

III. DEVELOPMENT OF THE VARIOUS COLORS IN THE PUPAL

WINGS.

A few researches have been carried out upon this interesting
topic, but as the literature is scattered and has never been brought
together, it will perhaps not be amiss to present a brief résumé of
the principal facts which have been already ascertained,

(1) Historical Account of previous Researches. In 1889
Schäffer ('89) discussed the question of the order and time of
appearance of the colors in the pupal wings of several of the
Vanessas. Unfortunately he apparently did not make his obser-

MAYER: COLOR AND COLOR-PATTERNS. Ln

vations at sufficiently close intervals of time, and was, therefore,
led into some misstatements, which have been corrected by van
Bemmelen (89) and Urech (91).

Van Bemmelen carried out an elaborate research upon the
development of the various spots and colors upon the wings of
Pyrameis cardui, Vanessa urticae, V. io, Pieris brassicae, and a few
other forms. He diseusses in detail the time and manner of appear-
ance of all of the different spots upon the wing. Into these details
we shall not follow him, but shall merely present his general con-
clusions regarding the development of the various colors. In Pieris
brassicae it appears that during the first days of. the pupal stage the
wings are eolorless and transparent ; after a few days, however, the
fore wings become opaque, and white; later the hind wing, also,
goes through the same changes. The wings then remain unaltered
until about two days before the butterfly issues. Then, very sud-
denly, the blaek spots and the yellow ground tone of the under
sides appear. White is thus the primary color ; blaek and yellow
secondary. The first color to make its appearance in the ease of
Pyrameis cardui is à brown-yellow ground eolor, whieh may be
observed in pupae four days old. The hind wings are at this time
somewhat darker than the fore wings. The color then changes from
darker brown to einnamon-brown. The black spots appear later upon
this delicate reddish brown ground color. The three fused spots
which form the whitish band in the middle of the front edge of the
fore wing appear during the last days of development, just before
the completion of the final color-pattern.

zoth van Bemmelen and Urech have shown that in Vanessa
urticae the order of appearance of the various colors is the same as
in Pyrameis cardui, "Phe first color to appear in Vanessa urticae is
a faint reddish tinge; this deepens and forms the ground color, and
later the black spots appear upon it.

Urech (91) has made a careful study of the development of the
colors upon the pupal wings of Vanessa io. The wings are at first
wholly white. Then in a restricted area of this white is noticed the
appearance of a yellow, which forms the yellow of the mature wings.
Almost contemporancous with the development of the yellow eomes
the red, which appears in another part of the primitively white field,
and gradually deepens in eolor until it forms the brownish red
ground color of the adult wings. Still later another portion of the

primitive white changes into the black of the mature wing, The

178 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

under side of the mature wings of Vanessa io is mainly uniform
black, and in this case also this color develops from the white at
very rapid rate, near the end of the pupal stage. This development
of the black directly upon the white areas is quite remarkable in
Vanessa io, and very different from that of both Vanessa urticae
and Pyrameis cardui, where the black spots develop upon a field
already tinged with red. Urech points out the fact, that some of
the white spots seen in the mature wings of the Vanessas represent
the “ primitive white " of the pupal wings

Finally, the latest paper upon the wlbjet of the development
of color in the pupa is that of Haase (93), who has examined
the pupae of a number of Papilios (e. g., philenor, mae haon,
asterias, turnus, and podalirius), and finds that during early pupal
life the wings are as transparent as glass; after a time, however,
they change to an impure white, which soon becomes yellowish, and
then the various colors which are destined to adorn the mature
wings begin to appear.

If we are to learn much of fundamental import concerning the
phylogeny of color in Lepidoptera, the researches should be carried
out 9 the lower moths, and not upon such highly specialized
forms of Rhopalocera as the Vanessae,

In my paper on Wing scales, ete. (Mayer, 796, p. 232), I have
come to the conclusion that dull ocher-yellow and drabs are,
phylogenetically speaking, the oldest pigmental colors in the Lepi-
doptera. The more brilliant colors, such as bright yellows, reds,
and pigmental greens, are derived by complex chemical processes,
and are, phylogenetically speaking, of recent appearance,

I have made a study of the development of the colors and pattern
in the wings of Callosamia promethea Linn. and of Danais plexip-
pus Fab.

(2) Development of Color in the Pupal W ings of Callosamia
promethea. "he cocoons of Callosamia promethea are very abundant
during the winter months, when they may be found hanging to the
stems of the food plants of the larvae. The pupal wings remain
perfectly transparent all through the winter, until about ten days
before the time when the moth is destined to issue ; ; they then become
opaque white. An examination of the wings at this period shows
that the scales are perfectly formed (Fig. 25, Plate 3), except for the

MAYER: COLOR AND COLOR-PATTERNS. 179

lack of pigment, which is developed later. If one treats the scales at
this stage with oil of cedar-wood or clove oil, they become practically
invisible under the microscope, thus demonstrating that there is
no pigment within them. Fig. 26, Plate 3, gives the appearance
presented by a scale taken from the light drab-colored margin of
the mature wing. This is about the lightest area upon the wing,
except the white spots; but it will be seen that this scale is much
darker in appearance than the unpigmented one shown in Fig. 25,
The white or unpigmented condition of the wing lasts for about
four days. The wings then become uniformly tinged with an
impure yellow or light drab, and very soon after this the colors
begin to make their appearance. They first appear upon the lower
surface of the wings. Fig. 28, Plate 3, represents the under
surface of the fore wing of a female in a very early stage of color
development; in fact the upper surface shows, as yet, no trace of
the colors. It will be seen that a few dark red streaks have
appeared near the central portion of the wing, and it is worthy of
note that these occupy the interspaces between the nervures. The
ocellus near the apex of the wing appears faintly outlined upon its
background of impure yellow.

Fig. 27, Plate 3, represents the under side of a hind wing of
a male in about the same stage as Fig. 28. Here, again, the red color
oceupies the interspaces, and indeed it is only later that the nervures
become clouded over by it.

Figs. 29 and 30, Plate 3, represent, respectively, the under and
upper sides of the fore wing of a male about five hours after the
first appearance of the colors. Upon the upper side (Fig. 30) we
see two gray streaks near the base of the wing and a light cinnamon-
brown eolor extending from the lower edge toward the middle of
the wing. The ocellus near the apex is now quite apparent, but
still faint in color, On the under surface (Fig. 29) the red markings
have developed to a much greater extent than in Fig. 28. The
outermost of the two white spots which oceupy the center of this red
area becomes the white central spot of the mature wing; the inner-
most one is soon obliterated owing to its becoming clouded over
with red,

Figs. Q7 and 36 represent respectively the upper surface of
the fore wing and the lower surface of the hind wing of a female,

slightly more advanced than in Fig. 30. Fig, 31 represents a male
guy 8 8

and Fig. 88 a female about twelve hours after the first appearance

180

BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

of the color. Itis remarkable that in this stage the male and female
wings are quite similar in general appearance, except that the ground
color of the male is now a dusky gray, while that of the female is a
cinnamon-brown.

From this time onward, however, the wings of the two sexes begin
to differ more and more in appearance, for the ground color of the

male becomes deep black, while that of the female remains einnamon-
brown. This change is well exhibited by Figs. 32 and 59, Plate 3,
which give the appearance of the upper surfaces of the male and
female wings respectively at about twenty hours after the first appear-
ance of the colors. Fig. 33 represents the hind wing of the same male
whose fore wing is shown in Fig. 82, Figs. 34, 35, 40, and 41 give
the appearanee of the pupal wings just before emergence, when
the colors are completely formed.

To summarize; Figs. 27, 29, 33, and 35 give suecessive stages in
the development of color in the male; and Figs. 28, 36-41 give |
similar stages for the female. It becomes evident, from a comparison
of these successive developmental stages, that the colors appear first
upon the central portions of the wings, and that the outer and costal
edges of the wings and the nervures are the last parts to acquire
the mature coloration,

It is worthy of remark that the color-pattern of the mature male
Callosamia promethea is quite a departure from the type of coloration |
which is commonly found among the Saturnidae. The female,
however, conforms very well to the general pattern of the other
species of the family. It is quite evident that the deep black colora- |
tion of the male is, phylogenetically speaking, a new acquisition, and |
that the coloration of the female represents the less differentiated
and therefore, more primitive type.

It is interesting in connection with these facts to observe that the
color-patterns of both male and female develop in almost identical |
ways up to the twelfth hour after the first appearance of the color ;
that then, however, the grayish ground color of the male wings
begins to deepen into the characteristie jet black of the adult, while
the light cinnamon ground color of the female merely becomes
slightly darker as the wings mature.

(3) Development of Color in the Pupal Wings of Danais
plexippus. Figs., 42-45, Plate 3, are intended to illustrate four
stages in the development of color in the pupal fore wings of Danais

plexippus. The pupal stage of this species is of brief duration, last-

MAYER: COLOR AND COLOR-PATTERNS. 181

ing from one to two weeks only, according to the temperature to
which the chrysalis is exposed. For the first few days the wings are
perfectly transparent, but about five days before the butterfly issues
they become pure white. An examination of the scales at this
period shows that they are completely formed and merely lack
pigment. In about 48 hours after this (see Fig. 42) the ground
color of the wings changes to a dirty yellow. It is interesting to
note that the white spots which adorn the mature wings remain
pure white. Fig. 43 illustrates the next stage, where the black has
begun to appear in the region beyond the cell. The nervures them-
selves, however, remain white. Fig. 44 shows a still later condi-
tion, where the dirty yellow ground color has deepened into rufous,
and the black has deepened and increased in area and has also
begun to appear along the edges of the nervures. In Fig. 45 the
black has finally suffused the nervures, the base of the wing and
the submedian nervure being the only parts that still remain dull
yellow. It is apparent that in Danais plexippus, as in Callosamia
promethea, the central areas of the wings are the first to exhibit the
mature colors, and that the nervures and costal edges of the wings
are the last to be suffused.

IV. Tue Laws WHICH GOVERN THE Coron-PaAvTERNS oF BUTTER-

FLIES AND Morus.

(1) Historical Account of previous Researches. The earliest
paper upon this subject is by Higgins (68). He came to the
conclusion, that “the simplest type of color presents itself in the
plain uniform tint exhibited when the scales are all exactly alike.”
He also thought it probable that “the scales growing on the mem-
brane upon or near the veins would be distinguished from the
scales growing on other parts of the membrane by a freer develop-
ment of pigmentary matter, and that in this manner would arise
a kind of primary or fundamental color-pattern, namely, a pale
ground with darker linear markings following the course of the veins,
e. g. Pieris crataegi.” He also attempted to explain the formation of
eye-spots by assuming that crescent-shaped markings migrate out-
wards from the sides of the nervures and meet so as to inclose a

Space,

182 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

It is, however, untrue that there is a freer development of pig-
ment within the scales lying upon the nervures; in fact, the reverse
is the case, as we have seen, in both Danais plexippus and Callo-
samia promethea. Higgins’s explanation of the formation of eye-
spots is also fallacious.

Darwin (71, Vol. 2, p. 133) published four excellent figures from
a drawing by Trimen, illustrating two simple ways in which eye-
spots are actually formed, both diametrically opposed to Higgins’s
hypothesis. Darwin says that in the South African butterfly, Cyllo
leda, “in some specimens, large spaces on the upper surface of the
wings are coloured black, and include irregular white marks, and
from this state a complete gradation can be traced into a tolerably
perfect ocellus, and this results from the contraetion of the irregular
blotehes of colour. In another series of specimens a gradation can
be followed from excessively minute white dots, surrounded by a
scarcely visible black line, into perfectly symmetrical and large
ocelli? with several rings.

Scudder (’88-’89) and, afterwards, Bateson ('94) have shown
that the ordinary eye-spots, such as those found in Morpho and the
Satyridae, are invariably placed in the interspaces between the longi-
tudinal veins of the wings, and also that they are often found repeated
upon homologous places of both pairs of wings. Bateson says that
ocelli are often seen upon both surfaces of the wing, the centers of
the upper and lower ocelli coinciding. In the majority of cases,
however, the upper and lower ocelli, although coincident, have quite
different colors. The simpler sort of ocelli, such as those seen in the
Satyridae or in Morpho, have their centers on the line of the fold-
marks or creases of the wing. It sometimes happens that these
creases seem to begin from the center of an ocellus. As these
creases commonly run midway between two nervures, it usually re-
sults that the center of the eye-spot is exactly half way between two
nervures. The-large eye-spots of Parnassius apollo are an exception
to this rule. In some Morphos, Satyridae, ete., in cell I^ of the hind
wing there are often two creases and two eye-spots, one for each
crease; but if there be only one eye-spot present, its center does not
correspond with the middle of the cell, “but is exactly upon the
anterior of the two creases.” I have observed the same law for the
white marginal spots in cell I* in Ceratinia vallonia, C. fimbria, and
Mechanitis polymnia.

In 1889 Seudder, in his work upon the Butterflies of New

MAYER: COLOR AND COLOR-PATTERNS. 183

England, ealled attention to the following faets: the transverse
series of dark spots so often seen in the body of the wings of
Lepidoptera are invariably placed in the interspaces between the
longitudinal veins, never upon the veins themselves, excepting
only in rare instances, where the spots occur at the extreme margin.
Ile also pointed out that in many types of moths all differentiation
in coloring has been greatly retarded, so far as the hind wings are
coneerned, by their almost universal concealment by day beneath
the overlapping front wings. In these cases “ the simplest departure
from uniformity consists of a deepening of the tint next the outer
margin of the wing." It is but a step from this condition to a
band of dark color or a row of spots parallel with the margin. This

»

explains why the transverse style of markings, for the hind wings
at least, is so common, Scudder showed that “the number of
instances, in butterflies, in which similar markings appear in the
same areas of the two wings, and in the same relative position
in these areas, is far too common to be a mere coincidence. It is
most readily traced in the disposition of the ocelli, which are very
apt to be similar in size and perfection, and to be situated between
the same branches of homologous veins."

(2) Laws of Color-Patterns. As a result of my own study of
the wings of moths and butterflies, I am prepared to propose the
following additional laws of color-patterns. (a) Any spot found
upon the wings of a moth or butterfly tends to be bilaterally symmet-
rical both as regards form and color, the axis of symmetry being
a line passing through the center of the interspace in which the
spot is found, and parallel to the direction of the longitudinal
nervures. For example, in Figs. 6 and 7, Plate 2, each spot is
bilaterally symmetrical about the axis HH. The same law holds
for the spots represented in Figs, 8-14 and 16.

(b) Spots tend to appear not in one interspace only, but as a row
occupying homologous places in successive interspaces. Indeed we
almost always find similar spots arranged in linear series, each sim-
ilar in shape and color to the others and occupying the center of its
interspace, The rows of spots represented in Figs. 8-14 and 16
will suflice to illustrate this law.

It is interesting to notice that bands of color are often made by
the fusion of a row of adjacent spots; and, conversely, chains of
spots are often formed by the breaking up of bands, leaving
a row of spots oceupying the interspaces. Many instances of this

184 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

are to be seen in certain specimens of various species of the
Heliconidae. For example, in Heliconius eucrate (Fig. 58, Plate
4) I have observed that certain specimens show a row of distinct
spots in place of the, usually entire, band which crosses the middle
of the hind wing. In fact, the vast majority of bands can be
analyzed into a series of similar elements, each element occupying an
interspace. Thus, in Plate 2, Fig. 17, which represents a wing of
Saturnia spini, the band seen crossing the wing parallel with its
margin is made up of a series of fused crescents, each crescent
occupying an interspace,

If, on the other hand, this band were to break away from the
nervures, tiie result would be a series of crescent-shaped spots each
occupying the center of an interspace. It is very interesting to
observe the manner in which bands degenerate and disappear.
Numerous opportunities for doing this may be had among the Heli-
conidae. In some species, as in Melinaea parallelis, hardly any two
specimens are alike in the condition of the black band across the
middle of the hind wings. The most common method of disappear-
ance is a shrinking away of the band at one end. This is wellillus-
trated in Figs. 84-87, Plate 7, which represent a sort of ** Mercator’s
Projection " of the wings of Mechanitis isthmia (for explanation of the
plan of projection see page 207.) Fig. 84 represents a male, showing
a well-marked band of hardly separated spots extending across the
middle of the hind wing. Fig. 87 shows a female in which the
spots are thinner and more crescentic and the separations much
more marked. Fig. 85 is also drawn from a female, in which it will
be seen that the band has shrunk away leaving only a portion of it
at the right, and in Fig. 86, which represents another female speci-
men, only one faint spot is left.

It is very common to find bands shrinking away at one end.
Sometimes, however, they shrink away at both ends, and very often
they break up into a row of spots, which may then contract into the
centers of their interspaces and finally disappear. It is worthy of
note that it is very rare to find a band breaking at the middle of its
length and each half receding from the other. Such a case is, how-
ever, shown by Melinaea parallelis (see Fig. 82, Plate 7), where one
sometimes finds specimens in which the black band across the middle
of the hind wings is complete and unbroken; whereas in other
specimens, as in Fig. 82, it is partially broken in the middle, and in
still others the break has become a wide gap by the drawing away of
the halves of the band from each other.

MAYER: COLOR AND COLOR-PATTERNS. 185

We see, then, that it is very common to find bands shrinking
away from either end, but very rare to find them broken in the
middle region. This, however, is only a special case of the law
enunciated by Bateson (794), that the ends of a linera series are more
variable than the middle. Almost any row of spots also exhibits
the same law, in that the spots occupying the middle portions of
the row are similar one to another, while those at the ends of the
series depart more or less from the type. (See Figs. 10-13,
Plate 2.)

The position of spots which are situated near the edge of the
wing is largely controlled by the wing-folds or creases. In Meli-
naea egina (Fig. 96, Plate 8) there is a row of white spots near the
outer edges of the wings, and each of these spots is cut in two by :
narrow black line which extends along the wing-fold. Also in Cera-
tinia vallonia (Fig. 81, Plate 7) and in many other forms of the
Danaoid Heliconidae one often finds two creases in a cell, and in
this case there are two marginal spots, one on each crease. In
many other cases, however, the marginal spots are double in each
cell, although there is but a single wing-fold ; the spots in these
sases are situated at some distance on either side of the fold. (See
Figs. 95, 96, Plate 8.) Another very common condition is exem-
plified in Fig. 83, Plate 7, where there is a single marginal spot
situated upon the wing-fold in each cell.

(3) Detailed Discussion of the Laws of Color-Patterns. Pigs.
6-14 and 16, Plate 2, are taken from special cases which serve
to illustrate the two chief laws of color-pattern, 7. e., that spots tend
to be bilaterally symmetrical about an axis (HH, Figs. 6, 7) passing
through the center of the cell parallel with the nervures; and
also, that spots of similar shape and color tend to be repeated in
a row of adjacent cells.

In Fig. 7 the spots are separated in the middle, but still incline
outward symmetrically from the center; indeed, instances of double
spots are very common. In such cases, however, each half spot is a
refleetion of its mate on the other side of the axis passing through
the center of the cell.

Fig. 8 represents various eye-spots found in the Morphos, and
will serve to illustrate the laws of eye-spots which have been enunci-
ated by Seudder (89) and Bateson (94), These spots occupy the
center of the cells in which they are found. In cell II, for example,
is a large eye-spot with a crescent in. its center, and it will be

186 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.
observed that this crescent follows the general law and is bilaterally
symmetrical about the usual axis.!

Fig. 9 showsthe law of repetition of some very complex spots, each
being bilaterally symmetrical. It is found in Parthenos gambrisius.

Figs. 10 and 11 represent Ornithoptera urvilliana and O. priamus
respectively. In Fig. 10 we see an instance of a spot within a spot,
and in Fig. 11 an even more complex case, for here there are three
systems of spots one within another.

Fig. 12 represents the marginal markings found in Hestia jasonia
and Fig. 13 Hestia leuconoe var. clara. These two examples are
intended to illustrate the fact, that, although the markings are
situated upon the nervures, they are bilaterally symmetrical not
about the nervures as axes, but about the usual axis passing midway
between the nervures. In Fig. 12 it will be seen that the two
curved markings situated upon nervures I" and 2, and projecting
into cell le, are bilaterally symmetrical only in reference to the axis
through the middle of the cell.

In allied species the spot situated upon nervure 1“ is often absent.
The system of markings is therefore undergoing degeneration at this
end (cf. Fig. 13, cell It). The curved mark upon nervure 5 (ig.
12) projecting into cell V is plainly symmetrical with respeet to its
fellow in the opposite side of cell V, and not with its near companion
which projects into cell IV. The same is also true in the case of the
spots in cell VI.

In Fig. 13 the spots appear at the first glance to be bilaterally
symmetrical about both nervures and centers of cells, but in cell IV
the marking situated on nervure 4 does not quite reach to the cen-
ter, and it is interesting to observe that its fellow on nervure 5 also
falls short of reaching the center and is therefore symmetrical with
respect to the other curved spot in cell IV. This case also furnishes
an instance of a break in the middle of a linear series.

Fig. 14 is taken from the under surface of the hind wing of
Papilio emalthion. It serves to illustrate the fusion of two orig-
inally separate rows of spots. In this case the crescent-shaped spots
above have fused with the rectangular ones below, so as to inclose
a portion of the ground color of the wing. Sometimes two rows of

1 A very beautiful exception (Fig. 19, Plate 2) to thisrule for the crescents found in eye-
spots is seen in the under surface of the fore wing of Missanga patinia Moore. It will be
noticed that the large black crescent found in this beautiful eye-spot is 90° away from its
This is the only exception of the sort known to me,

usual. position,

MAYER: COLOR AND COLOR-PATTERNS. 187

spots of different colors fuse, giving a chain of spots which are of
one eolor above and another below.

In Fig. 16 the spots composing the row BB are blue (dark)
above, and red (light) below. It will be observed that the color is
bilaterally symmetrical, as usual, about the axis through the middle
of the cell. Such bicolored spots are often due to a simple fusion,
as before stated; but sometimes they may, perhaps, be intrinsically
bicolored.

Fig. 15 is a beautiful instance of an exception to the general rule
that spots are bilateral about the axis through the center of the cell.
It is taken from Ornithoptera trojana Staudinger. The light spots
represented near the outer edge of the wing are of a brilliant irides-
cent green. It is evident that they are distinetly bilateral with
respect to the nervures y especially is this true of the pair adjacent to
nervure 1. Ornithoptera brookiana Wallace illustrates another
exeeption, though in a less marked degree. Other allied species of
Ornithoptera, however, would seem to show that these apparent
exceptions may have been derived from forms which exhibited two
spots in each cell and followed the usual rule. These are the only
instances of such exceptions known to me. I do not doubt, how-
ever, that further study would reveal others.

In Fig. 17 an example is given of the peculiar kind of eye-spots
found in the Saturnidae. The species from which the figure was
taken is Saturnia spini. It will be seen that this so-called eye-spot
is quite different in formation from the ocelli of butterflies. It is
simply a series of curved cross-bands between nervures, arranged
symmetrically on both sides of the cross vein CC. The “eye-
spots” upon the wings of Attacus lina and in the genus Telea are
to those found

also of this sort. True eve-spots, however, similaı

among the Morphos and Satyridae, occur in moths, as in the apex
of the fore wing of Samia ceeropia, Callosamia promethea, ete.
„False“ eye-spots are also found on the wings of butterflies; in
Vanessa 10, for example, the so-ealled eye-spot of the fore wing has
been shown by Dixey (90) to be made up of a series of fused
spots. Ut will be remembered that Merrifield (94, Plate 9, Fig. 4)
caused this * ocellus " to break up into its constituents by subjecting
the pupa to a temperature of 1° C. The ocellus upon the hind wing
of Vanessa io ts no doubt a true eye-spot; the only evidenee which
1 See Watkins, '01, Plate +.

? See Hewitson, "50-76, Vol. 1.

188 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

might lead one to infer that the ocellus of the fore wing was of the
same character is, that an aberrant form is sometimes found in
nature having the “eye-spots” on both fore and hind wings
obliterated, thus indieating a possible connection between the two
(see South, ’89).

Fig. 18 is intended to illustrate the process of degeneration occur-
ring in bands. Band BB is represented as hreaking down by the rare
method of parting in the middle. Example, Melinaea parallelis.
Band EE is degenerating at one end; this is a very common
method.

Figs. 20-23 represent hypothetical conditions not found in nature ;
all being contrary to the conditions of the laws which have just been
stated.

In Fig. 20 row RR presents three spots for each cell. I believe
this has not been found in nature, but I should not be surprised if
it were discovered, for it is not contrary to any of the laws.

Row CC, on the other hand, is contrary to the law of bilaterality,
the erescents not being bilateral about axes passing through the
middle of the interspaces parallel with the longitudinal nervures.

Fig. 21 is intended to show a series of spots arranged side by
side in twos in each cell, and of different colors. This, I believe, is
impossible, for it is contrary to the law of bilaterality of color
arrangement about the usual axis (HH, Figs. 6, 7).

In Fig. 22 there are several conditions which are impossible; e. y.,
an eye-spot situated upon à nervure is never seen in nature, also
two spots originally side by side, as in cell III, never rotate around
each other so as to come to lie one above the other. Spots often
move, however, as shown by the arrows in cell IV, thus giving
rise to fusions; or they may move away from each other, causing a
wider gap between the rows. In cell I" are shown two looped
spots. One form (A) is quite usual, being found indeed in Cymo-
thoe caenis Drury.! The other form of spot (D) is an impossibility,
not being bilaterally symmetrical.

Fig. 23 illustrates other impossibilities in color-pattern, none of
them, of course, being found in nature. For example, one never
finds a row of slanting spots such as SS. Also one never sees a
row of similar spots in alternate interspaces, such as is shown in
DD, for this would be contrary to the law that similar spots are
repeated in a row of adjacent interspaces. These last four diagrams

1 See Cramer (1779-82), Vol, 2, Plate 146.

MAYER: COLOR AND COLOR-PATTERNS. 189

(Figs. 20-23) have been introduced merely to give an idea of the
curiously strict limitations which nature has imposed upon the differ-
entiation of the color-pattern. Many beautiful effects might have
been produced, such for example as that of alternate interspaces
showing different colors, but this is not seen in nature.

It is interesting to recall the fact, that the colors themselves are
impure and by no means so brilliant as they, perhaps, might have
been, had Natural Selection been more severe in regard to color.

There is doubtless some physiological reason why spots almost
invariably appear and disappear in the middle of the interspaces, and
when we know more of the anatomical and histological conditions
of the wing during the development of the colors, we may be able to
discover it. It will be remembered that in the developing pupal
wings of Callosamia promethea and Danais plexippus I. found that
the colors first made their appearance in the interspaces, and finally

spread out so as to tinge over the nervures.

(4) Origin of Color- Variations, There is every reason to
believe that all kinds of spots and bands, which are essentially
only fused spots may appear or disappear in any individual
specimen without going through a long course of Natural Selection
and slow phylogenetic differentiation. Darwin and Trimen C71)
and Bateson (94) have demonstrated that this is true for eye-spots.
In the IIeliconidae I have found that bands and rows of spots are
very variable in different speeimens of the same species (see Plate 7,
Figs. 84-87).

There is a large and widely seattered literature recording the
appearance and disappearance of colors and markings upon the
wings of Lepidoptera. Limits of time and spaee prohibit my doing
justice to it here, but it may be well to call attention to a very few
of the more recent papers upon the subject, Many of the color-
aberrations recorded in this list of papers may be due to the direct
influence of environmental conditions upon the individual, but others
are no doubt true sports or, to speak crudely, * congenital” variations,
and might under favorable conditions of life become the ancestors
of new varieties or species. It seems highly probable that new
species often arise from just such sports in the manner so frequently

and ably expounded by Bateson,

190 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

PARTIAL BIBLIOGRAPHY OF REMARKABLE COLOR-ABERRATIONS IN

LEPIDOPTERA.

Bairstow, S, D. 77. Ent. Mo. Mag., Vol. 14, p. 67. (Zygaena filipendulae. )

Bran, T. E. 95. Can. Ent., Vol. 27, p. 87-93, Plate 2. (Nemeophila petrosa
and varieties.)

Benson, E. F. '88. Entomologist, Vol. 16, p. 210. (Arge galathea. )

Brzusz, L. 89. Feuille jeun. Natural., 19 Ann., p. 142-142.

Breienet, F. '00. Bull. Soc. Ent. France, (6), Tome 10, p. 29-30. (Thecla
rubi; Melithaea athalia.)

CARRINGTON, J.’78. Entomologist, Vol. 11, p. 97, Fig. (Cidaria suffumata.)

CARRINGTON, J. 83. Entomologist, Vol. 16, p. 1, Fig. (Callimorpha dominula.)

JARRINGTON, J.’88. Entomologist, Vol. 21, p. 73, Fig. Editorial Note. (Arctia
ala.) And numerous other papers in the Entomologist.

CLARK, J. A. 89. Entomologist, Vol. 22, p. 145-147, Plate 6. (Triphaena
comes.)

SOCKERELL, T. D. A. '86. Entomologist, Vol. 19, p. 220-231. (Epinephele
tithonus.)

JOCKERELL, T. D. A. 88. Entomologist, Vol. 21, p. 189. (Pieridae.)

CockERELL, T. D. A. ’89. Entomologist, Vol. 22, p. 1-6, 13, 20-21, 26-29, 54-
56, 98-100, 125-130, 147-149, 185-186, 243-245.

Dewrrz, H.’85. Berlin. Ent. Zeitschr., Bd. 29, p. 142, Taf. 2. (Precis amestris.)

Epirors or Entomorocist, 78. Entomologist, Vol. II, p. 169-170, Plate 2.
( Vanessa atalanta and several Lepidoptera.)

Epwarps, W. II. 68. Butterflies of North America. (Numerous plates.)

Ferrie, F. J. 89. Feuille jeun. Natural., 19 Ann., p. 84. (Variations of
Lepidoptera in Alsace.)

Frrcu, E. A.’78. Entomologist, Vol. 11, p. 50-61, Plate. (Colias edusa.)

Goss, II. 78. Entomologist, Vol. 11, p. 75-74, Fig. (Chelonia villica.)

ÖBERTHÜR, C. 89. Bull. Soc. Ent. France, (6), Tome 9, p. 74-76.

OnrnTHÜR, C.'03. Feuille jeun. Natural, 24 Ann., p. 2-4.

PO HADE, G. A. '01. Ann. Soc; Ent. France, (0), Tome 11, p. 507-598,
Pl. 16. (Thais rumina.)

Iucnanpsow, N. M.’89. Ent. Mo. Mag., Vol. 25, p. 280-201. (Zygaena filipen-
dulae.)

Scupper, S. H. ’89. Butterflies of New England, p. 1213. (Bibliography of
variations of Pieris rapae.)

Sourn, R.’89. Entomologist, Vol. 22, p. 218-221, Plate 8. (Various Vanessidae.)

SpEYER, A. 74. Stettiner Ent. Zeitung, Bd. 35, p. 08-102. ;

Tiree, II. '84. Berlin. Ent. Zeitschr., Bd. 28, p. 161-162, Fig. (Apatura iris.)

Torr, J. W. '80. Entomologist, Vol. 22, p. 15, 160-161. (Melanie Agrotis
corticea and pale variety of Lycaena bellargus.)

(5) Climate and Melanism, Lord Walsingham (85), in his
presidential address before the Yorkshire Naturalists’ Union, brought
forward the idea, that, although Arctic insects might be perfectly

MAYER: COLOR AND COLOR-PATTERNS. 191

able to withstand the most severe cold while in hibernation during
the winter, it is of great importance for them to absorb as much heat
as possible during the short summer, He placed several species of
lepidopterous larvae upon a snow surface exposed to bright sunshine.
The snow melted at different rates under the various larvae, and
in two hours the darkest insect had sunk by far the deepest into the
snow, proving that it was the best absorber of heat. This ingeni-
ous experiment of Lord Walsingham should be made the beginning
of an extensive and careful research.

Chapman (88) has shown that it may be of advantage to moths
inhabiting wet regions to display dark colors, or become melanie.
His observations were made upon Diamea flagella, and he says that
upon one showery afternoon he observed that one side of the tree
trunks was wet and dark in color; the other side being dry was
paler. “ As a consequence, the dark specimens of flagella were very
conspicuous upon the dry portions, hardly visible on the wet, whilst
with the ordinary form the conditions were reversed, those on the
wet bark were conspicuous, those on the dry much less so.” Per-
haps the dull coloration of Arctie moths may be partially due to the
effect of the somber background of rocks in the regions which they
inhabit.

(6) Relation between Climate and Colors of Papilios. It is well
known that the Lepidoptera in the Tropics display the richest
rariety and greatest number of colors. I have counted the colors
exhibited by the 22 species of Papilio enumerated by Edwards as
inhabiting North America north of Mexico, and also those which
are displayed by the 200 species of Papilio named in Schatz's list as

> were determined by com-

found in South America. The “ colors’
parison with the colored plates in Ridgway (86).

In this manner it was determined that the North American
Papilios exhibit 17 colors, viz., black, brown, primrose-yellow, canary-
yellow, sulphur-yellow, orange, white, greenish white, apple-green,
cream-color, azure-blue, sage-green, rufous, pearl-gray, indigo-blue,
iridescent blue, iridescent green.

On the other hand the South American Papilios exhibited 36
colors, viz., black, translucent black, brown, white, canary-yellow,
citron-yellow, olive-yellow, primrose-yellow, chrome-yellow, straw-
yellow, gamboge-yellow, cream-color, greenish white, apple-green,
malachite-green, emerald-green, sage-green, slaty green iridescence,

pea-ereen, azure-blue, iridescent Berlin-blue, indigo, pearl-blue,

192 BULLETIN: MUSEUM OF COMPARATIVE ZOÓLOGY.
glaucous blue, salmon-buff, écru-drab, flesh-color, coral-red, rose-red,
vermilion, rufous, geranium-red, geranium-pink, olive-buff, iridescent
geranium-pink (as in P. zeuxis), and transparent areas.

As 200 species in South Americ: display but 36 colors, while 22
in North Ameriea show 17, it follows that, while the number of
species in South America is 9 times as great as in North America,
the number of colors displayed is only a little more than twice as great.
The richer display of colors in the Tropics, therefore, may be due
simply to the far greater number of species, which gives a better oppor-
tunity for color-sports to arise, and not to any direct influence of the
climate. The number of broods, also, which occur in a year is much
greater in the Tropies than in the Temperate Zones, so that the Trop-
ical species must possess a correspondingly greater opportunity to
vary.

V. Tue CAUSES WHICH HAVE LED TO THE DEVELOPMENT AND
PRESERVATION OF THE SCALES OF THE LEPIDOPTERA.

(1) Experiments und Theory. It is well known that the scales
of Lepidoptera are morphologically identical with hairs. Indeed, a
graded series from simple hairs such as are found covering the
body-surface of most Arthropods, up to perfeetly developed flat
scales bearing well differentiated striae may usually be found upon
one and the same insect.

It is also remarkable that the color-bearing scales of beetles have
been developed in the same manner as those of moths and butterflies,
and that in this case also hairs have become differentiated into scales
which are precisely similar in appearance to those of the Lepidoptera
(see Dimmock, ’83).

This is only another of the numerous instances met with in nature
where similar conditions of selection have developed complex organs
which are similar in appearance, though found in widely separated
groups. A list of papers relating to the development of scales has
been given by Dimmock (83, p. 1-11).

Most of the hairs which cover the body-surface in Arthropods are
true sensory structures, the axis of each of which is a protoplasmic
process from a single eell of the hypodermis, which lies below the
cuticula. They have probably been developed because the cuticula,

MAYER: COLOR AND COLOR-PATTERNS. 193

being hard, chitinous, and inflexible, would serve but poorly as a
taetile or sensory surface.

Of course no one would venture to ascribe any sensory function to
the seales which cover the wing-membranes of the Lepidoptera.
We may, however, make several more or less reasonable hypotheses
concerning the probable uses of the seales, and by testing these sup-
positions arrive perhaps at some plausible explanation of their reten-
tion and the complex development which they have undergone.

(1) They may have caused the wings of the ancestors of the
Lepidoptera to become more perfect as organs of flight, by causing
the frietional resistance between the air and the wing-surface to
become more nearly an optimum.

(2) The appearance and development of the scales may have
served, as Kellogg (9+) has suggested, “to protect and to strengthen
the wing-membranes."

(3) The present development of the scales may be due to the
fact that they displayed colors which were in various ways advan-
tageous to the insects.

Coneerning the first of these three hypotheses, the wing has,
broadly speaking, two chief functions to perform in flight. It must
beat more or less downward against the air, and must, in addition,
glide or cut through the air, supporting the insect in its flight. For
the mere beating against the air a relatively large co-efficient of
frietion between the air and the wing might be advantageous; but
for gliding and cutting through the air a small eo-eflieient of friction
would certainly be an advantage. There must therefore be an
optimum co-eflicient of friction, which lies somewhere between these
two.

In order to determine the co-efficient of friction between the wing
and the air, use was made of a method which, in one form or
another, has long been known to engineers; that is, of observing the
ratio of damping of the vibrations of a pendulum.

It is well known that when a pendulum is swinging free, and
uninfluenced by any frictional resistances, the law of its motion is
expressed by the formula,

. 2m

(dis Asin Tt
where d is the displacement of the pendulum from its middle
position after the interval of time t, A is the maximum displace-
ment and T the time of a complete- vibration, back and forth. If,

194 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

however, frictional resistances interfere, the formula becomes,

r u t
xw i 11
, —log d T, s
(3) Hence, K = A log e sin 2r e , OF if t ils

—log d

K IT loge
where K is a constant dependent upon the friction, e is the base of
the Napierian system of logarithms and T, is the time of a complete
vibration, which may be different from the T, representing the time
of vibration when not under the influence of friction.

The plan was, then, to attach the wing of some large butterfly or
moth to the end of a short, light pendulum in such a way that it
would either fan against the air, or cut through it, and then to
observe the ratio of damping of the pendulum's vibrations. A
drawing of the pendulum with a wing attached is given in Plate 1,
Fig.3. The wing is here shown in the position for “cutting or glid-
ing” through the air. It would be in the position for fanning against
the air, if it were rotated 90°. The pendulum was made of brass
and steel, the ends being of brass and the slender middle portion of
steel. Its vibrations were read off upon an are graduated in milli-
meters. The readings were certainly accurate down to 0.5 mm,
The pendulum was hung upon a steel knife edge (N, N, Fig. 3),
which rested upon firm level glass bearings. The pendulum was
24.21 cm. long, and weighed 19.61 grams. Its time of vibration
(T,) was 0.877 seconds. This rate of vibration was practically
unaltered when a wing was fastened to the end of the pendulum,
the reason being that the wings were very light, the heaviest, that of

Samia cecropia, weighing only 0.038 grams. The wing to be experi
mented upon was fitted into a deep, narrow slot at the free end ol

the pendulum, and then cemented in by means of a little melted

| part of the pendulum

beeswax. It thus became a perfectly
itself.
The pendulum with wi
are, read off upon the millimeter scale, and its reading at the end of
r

; attached was deflected through a known

the first swing carefully observed. Then if A be the initial deflec-

tion, which we may call unity, and if d be the reading after the first
, » Y3 E

. . B L * ^ 0 ë
swing, the ratio of damping is given by the expression A In experi-

menting with a fore wing of Samia cecropia “fanning the air,” it

MAYER: COLOR AND COLOR-PATTERNS. 195

was found, as the mean of many trials, that this ratio of damping
was 0.919, that is to say, the amplitude of the 2d swing was 0.919
as great as the amplitude of the 1st, that of the 3d only 0.919 as
great as that of the 2d, and so on. The scales were then carefully
removed from the wing-membranes, by means of a camel's hair brush,
and by again testing the vibrations it was found that the new ratio of
damping was 0,917. This is so near the value of the ratio of damp-
ing with the seales on (0.919), that it may be considered identical,
the difference being due to errors of experimentation.

Hence we must conclude that the presence of the scales upon the
wing-membrane has not altered, appreciably, the co-efficient of fric-
tion which would exist between scaleless wing-membranes and the
air. The results indicate rather, that when the scales appeared upon
the wings of the scaleless, clear-winged ancestors of the Lepidoptera,
the co-eflicient of friction remained unaltered. This tempts one to
the further conclusions, that the co-efticient of frietion between the
air and the wings was already an optimum in these elear-winged an-
cestors before the appearance of the scales, and therefore that Natural
Selection would operate to keep it unaltered,

A wing of Samia ceeropia eut so as to give it the same shape and
dimensions as one of Morpho menelaus, gave an identical damping
ratio. I conclude that the co-efficient of friction may be the same
for both moths and butterflies, at least for those which move their
wings at about the same rate in flight,

It was found in the ease of the Samia ceeropia wing, that when
it was vibrated in the position for * cutting through” the air, the ratio
of damping was 0.991. It will be remembered that, when the wing
“fanned” the air, this ratio was 0.917. We may find the ratio be-
tween the resistance encountered in “ fanning” and that encountered
in “gliding” through the air by substituting these values in equa-

log d
AT, log e

tion (4), K:

d : :
Thus for fanning, = 0,917 and T, = 0.877. Making A unity,
log 0.917
0.877 log e
: Ab :
In eutting through the air, \ 0.991 and T, as before =
log 0.991

0.877 log e

Ix IB

0.877.

Hence in this ease K = 0.01.

196 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The wing, then, encounters at least 10 times the resistance in fan-
ning that it does in gliding through the air. It should be said that
this last experiment is somewhat crude, for the wing necessarily
could not be made to cut the air with that delicate precision which
is probably realized by the insect in flight. I should not be
surprised, if in nature the insects encountered at least 20 times
the resistance in beating the air, that they do in merely gliding
through it.

Concerning Mr. Kellogg's supposition, that the scales were devel-
oped to “protect and to strengthen the wing-membranes," I will
admit that they may serve in some slight degree to protect the wing-
membranes from scratches, ete.; but I am unable to accept his con-
clusion, that they strengthen the wing-membranes, any more than
that the shingles upon a roof serve to add strength to it. The
wing-membranes themselves are tough, elastic, and not easily torn or
scratched, and the scaleless wings of the Neuroptera and Ifyme-
noptera are very rarely found torn or scratched in nature.

In 1858 Mr. Alexander Agassiz called attention (59) to the fact,
that “the nervures of the wings of butterflies are so arranged as to
give the greatest lightness and strength ; they are hollow, with their
greatest diameter at the base of the wing, the point of greatest
strain, their diameter gradually diminishing to the edge of the
membrane. If a section be made across such a wing parallel to the
axis of the body, we find very much the arrangement which has
been experimentally proved by Fairbain and Stephenson as giving
the greatest strength of beams, as exemplified in the tubular bridge.
We find the strongest nervure placed. either on or near the anterior

edge of the upper wing; there is no such nervure on the lower
wing, all being of nearly the same size, as such a one would have
prevented the elasticity of the wing from assisting the flight to
any considerable extent." Mr. Agassiz has informed me that he
sarried out an extensive series of experiments upon the rigidity of
the wings of various species of Lepidoptera. Ile placed little
platinum strips upon the wings and observed the extent of the
bending produced. His results demonstrated that the Sphinx moths
possess by far the strongest wings, and that the Danaoid and
Acraeoid Heliconidae have very weak wings. ‘The reason for this
probably lies in the fact, that the Sphinx moths move their wings
with great rapidity, while, according to Bates (62 and all sub-

sequent observers, the Heliconidae have a slow flight.

MAYER: COLOR AND COLOR-PATTERNS. 197

As the scales have been developed not because they aided the
insects in flight or strengthened the wings, their retention must
have been due to some other cause, probably to their display-
ing colors which were advantageous to their possessors in various
ways. As Dimmock (83) says, “it is only in insects where certain
kinds of brilliant coloration have been developed that one finds
scales," Indeed, I believe that the vast majority of the scales
found in Lepidoptera are merely eolor-bearing organs. They prob-
ably first made their appearance upon small areas of the wings,
perhaps adjacent to the body, and were merely colored hairs, sim-
ilar to those of the surface of the body, which had grown out upon
the wings. In this position they displayed some color which was
of advantage to the insect; perhaps serving to render it less con-
spieuous than formerly. Under these circumstances they would
naturally be preserved through the operation of selection until
finally they became modified into true scales; just as the hairs in
the Coleoptera have undergone a similar modification, If this
be true, it is easy to see how they might spread out over
the surfaces of the wings until the whole wing became covered
with scales.

(2) Summary of Conclusions. The seales do not aid the insects
in flight, for the wings have preeisely the same efliciency as organs
of flight when the scalesare removed. The phylogenetic appearance
and development of the scales upon the sealeless ancestors of the
Lepidoptera did not in the least alter the efficiency of their wings as
organs of flight. This efficiency of their wing surfaces was probably,
therefore, already an optimum before the scales appeared. The
scales do not appreciably strengthen the wing-membranes, that
function being performed by the nervures. The majority of the
seales are merely color-bearing organs, which have been developed
under the influence of Natural Selection,

PART B.
COLOR-VARIATIONS IN THE ITELICONIDAE.

I. GENERAL Causes WHICH DETERMINE COLORATION IN THE
I rtcoNIDAE.

In 1561, after eleven years of study within the forests of South

America, Bates read his, now classic, paper upon the life and habits

198 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

of the Heliconidae of the Amazon region. In it he first brought
forward his ingenious theory of Mimiery— a theory which, under
the able interpretations of Wallace and Fritz Müller, and in more
recent times, under the impetus of the zeal of their numerous disci-
ples, has yielded so much that is of interest to scientific men.

The Heliconidae are, above all, creatures of the forest, and Bates
found that the number of species increases as one travels inland
from the Lower Amazons towards the eastern slopes of the Andes,
so that the hot Andean valleys near Bogota, or in Ecuador, contain
perhaps the greatest number. In their range they are restricted to
the Tropics of the New World. Only two species, Dircenna klugii
and Heliconius charitonius, extend so far north as the extreme South-
ern States of the United States, and none of them are found much
further south than 30° 8. Lat.

ates and Felder first saw that the Heliconidae were naturally
divided into two distinct groups. One, the Danaoid Heliconidae,
consists of about twenty genera, all more or less elosely related, and
evidently an offshoot from the great universal family, the Danaidae,
members of which are found in both Hemispheres. The other group,
the true Heliconidae, is composed of two closely related genera, Heli-
conius and Eueides. They are allied in structure to the Acrae-
idae and hence their name, Acracoid Helieonidae. Schatz and Röber
(85 92, p. 105) say of the Acraeoid Heliconidae:— They are an
offshoot of the great family Nymphalidae, which have undergone a
remarkable development in the length of the fore wing, and in this
respeet have been developed in a direction parallel with the Danaoid
Ileliconidae. In their structure, however, they are quite distinct
from the Danaoid group.

Schatz has proposed a new classification for the Heliconidae. He
finds that the genera Lycorea and Ituna, which Dates included among
the Danaoid ITeliconidae, are very closely allied to the Danaidae, he
therefore says that Lycorea should be placed among the Danaidae,
while Ituna is clearly midway between the Danaidae and the Dana-
oid Heliconidae, Schatz proposes the name ** Neotropidae ” for the
Danaoid Helieonidae, However, I think the name * Danaoid Ileli-
conidae,” being older and more descriptive of their relationship,

should by all means be retained. In this paper I shall follow Bates’s
classification, and inelude among the Danaoid Heliconidae the twenty
genera: Lycorea, Ituna, Athesis, Thyridia, Athyrtis, Olyras, Eutre

\protopos, Dircenna, Callithomia, Epithomia, Ceratinia, Sais,

SIS, 4

|
|
l
i

MAYER: COLOR AND COLOR-PATTERNS. 199

Scada, Mechanitis, Napeogenes, Ithomia, Aeria, Melinaea, and
Tithorea. The Aeraeoid Heliconidae will then consist of the two
remaining genera, Heliconius and Rueides,

Staudinger 8488) records 453 species belonging to the Danaoid
group, and 150 belonging to the Aeraeoid group.

Nearly all that we know concerning the early stages of the
Heliconidae is due to Wilhelm Müller (86). Ile gives figures and
more or less complete deseriptions of the early stages of Dircenna

xantho, Ceratinia eupompe, Ithomia neglecta, Thyridia themisto,

Mechanitis lysimnia, and also of Heliconius apseudes, II. eucrate, H.
doris, Eueides isabella, E. aliphera, and E. pavana. Bates (762; p.
596) says that he raised the larvae of Heliconius erato and Eueides
lybia. Schatz and Röber ('85—92) figure the larva and pupa of
Ceratinia euryanassa, Edwards has given a detailed aecount of the
early stages of IT. charitonius.

Müller found that the larvae of the Danaoid group feed on various
species of Solanum, while the genera Ilelieonius and Eueides feed
upon the Passifloreae, The larvae are conspicuously colored, and
often gregarious; they seem to take but little pains to hide them-
selves during the ehrysalis stage, for Müller says that he has seen
the silver-spotted, white chrysalids of IIeliconius doris hanging in
great numbers in the near neighborhood of the larval food plant.
The mature inseets also furnish a good example of what Wallace
C67) designated as “warning coloration,” for their tawny orange
and black wings are very conspicuous as they sail slowly around in
circles, settling at frequent intervals in their lazy irregular flight.

Bates was the first to call attention to the eireumstance that they
often possess a rather strong and disagreeable odor, and in 1878
Frit; Müller confirmed this observation for a number of the Heli-
conidae, He found, for example, that the genera Ituna and Ilione
have a pair of finger-like processes near the end of the abdomen,
which ean be protruded and then emit a rather disagreeable odor;
and he also found that the Acracoid Heliconidae, especially the
females, possess a disgusting odor. Seitz C89), however, examined
about fifty species of IlTeliconidae and found that many of them
appear to have no odor. For example, he says that ITeliconius
eucrate and Eueide dianasa have no odor, but that some specimens
of Heliconius beskei, and Eueides aliphera have a horrid odor,

Whether they are odorous or not, it would seem that the Heli

eonidae have but few enemies to fear, for not one of the many

200 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

skilled observers who have studied them in their native haunts has
ever seen a bird attack them, and the only ground for believing that
they are attacked rests upon the rather dubious evidence of a few
specimens found by Fritz Müller having symmetrical pieces
apparently bitten out of the hind wings. Belt (’74) observed that :
pair of birds which were bringing large numbers of dragon-flies and
butterflies to their young never brought amy of the IIelieonidae,
although these were abundant in the neighborhood. In fact, Belt
was able to discover only one enemy of these butterflies, and that
was a yellow and black wasp, which caught them and stored them
up in its nest to feed its young. The ITeliconidae then, in spite of
their weak structure, conspicuous colors, and slow flight, enjoy a
peculiar immunity.

As is well known, Bates (62) first called attention to the fact that
the Heliconidae were “ mimicked ” or imitated both in color-pattern
and shape of wings by a number of other genera of butterflies and
even moths. Bates had no difficulty in showing that this mimiery
might easily be explained upon the ground that the IHeliconidae, on
account of their bad taste and smell, were immune from the attacks
of birds and other insectivorous animals, and that therefore it gave a
peculiar advantage to a butterfly belonging to any other group not
thus protected, to assume the shape and coloration of the Heliconidae ;
for then the birds could not perceive any difference between it and
the true Heliconidae. Bates found that fifteen species of Pieridae
belonging to the genera Leptalis and Euterpe, four Papilios, seven
Eryeinidae, and among diurnal moths three Jastnias and fourteen
Bombycidae imitate each some distinct species of the Heliconidae
occupying the same district. Healso found that all of these insects
were much rarer than the Heliconidae which they imitated. In some
cases, indeed, he estimated the proportion to be less than one to a
thousand. Wallace (89, p. 265), who has added so much to our
knowledge of this subject, aptly defines this kind of mimiery as an
“exceptional form of protective resemblance.”

But by far the most remarkable discovery made by Bates was the
fact, that species belonging to different genera of the ITeliconidae
themselves mimie one another. Neither Bates nor Wallace was
able to give any satisfactory explanation of the cause of this latter
form of mimiery, for all of the genera of the Heliconidae are
immune. They therefore supposed it to be due to “unknown local

causes,” or similarity of environment and conditions of life.

MAYER: COLOR AND COLOR-PATTERNS. 201

Thus the matter rested until 1879, when Fritz Müller brought out
his well-known paper upon “Ituna and Thyridia, a remarkable
example of mimicry,” in which he showed that both of these genera
are protected, yet they mimic each other. He also showed that this
mimicry might be due to Natural Selection brought about in the
following manner. It is possible that young birds, upon leaving the
nest, are not furnished with an unalterable instinct which tells them
exactly what they should and should not eat; so they may try
experiments, and would then in all probability taste a few of the
Heliconidae before finding out that they were unfit to eat. Müller
then demonstrated that, if this supposition be true, it becomes a
decided advantage to the various species of Heliconidae to resemble
one another. His reasoning was as follows: Let it be supposed
that the young and inexperienced birds of a region must destroy
1,200 specimens of any distasteful species of butterfly before it
becomes recognized as such, and let us assume further that there are
in existence 2,000 specimens of species A, and 10,000 of species D ;
then, if these species are different in appearance, each will lose 1,200
individuals, but if they resemble each other so closely that they can-
not be distinguished apart, the loss will be divided pro rata between
them, and A will lose 200, and B 1,000; therefore A saves 1,000 or
50% and B saves only 200 or 2% of the total number of individuals
in the species; hence, while the relative numbers of the two species
are as 1 to 5, the relative advantage derived from the resemblance
is as 25 to 1.

Blackiston and Alexander ('84) have given a complete mathe-
matical statement of Müllers law, and have come to the conclusion
that, if the number of individuals destroyed is small compared with
the number constituting the species, the relative advantage is
inversely as the square of the original numbers; but if the number
destroyed is large compared with the original number, the ratio of
advantage is much greater than the inverse squares of the original
numbers. Their deduction may be briefly stated as follows: —

202 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.
Designation of, Species A B
(4) Original namber sia 1 a > b
(2) Number lost without imitation .. e — fone
(3) Remainders without imitation . . (a—e)
Ri AORA a b
(4) Number lost with imitation ; . . TIRE 12

(5) remainders with imitation: a (1 rb) b (1 1 x)

(6) Excess of remainders cd i
Jur T3 )e ;
to imitation, or “ abso- » ust ‚86
i i a ) m
lute advantage? (3) —(5) j a+b

b e a

e
Jab qoem | arb X boo

(T) Ratio of excess to remainders )
without imitation (6): (3),
proportional advantage

AIT E e

8) Ratio of proportional advan- — —

(8) I oJ N 3 a (a—e) 15 l D
tage of B to proportional " — — — €

: 1 ES (be) 7 e
advantage of A. 1 i

It is evident, then, if e be small compared with a and b, that the
proportional advantage of B is to the proportional advantage of A
as a? is to b?. If, however, the loss (e) is great compared with a or
b, the relative gain for the weaker species becomes even greater than
the ratio of the: squares of b and a.

If it be true, then, that young birds, when they leave the nest, do
not possess a directing instinct telling them what they should and
should not eat, but actually do experiment to some extent upon
various insects which they meet with, Müller's law is amply sufficient
to account for the numerous cases of mimicry and remarkably close
resemblances which are found among the species of the Heliconidae
themselves.

Unfortunately no direct experiments have ever been made upon
‘the feeding-habits of young South American birds, nor have the
contents of their stomachs been examined. There have been a few
experiments, however, which seem to support the idea that some
animals do learn to associate an agreeable or disagreeable taste with
the coloration and appearance of their prey. It is well known that
Weismann (’82, p. 336-339) found that the black and yellow
larvae of Euchelia jacobaeae were refused by the green lizard of

Europe. He then introduced some young caterpillars of Lasiocampa

MAYER: COLOR AND COLOR-PATTERNS. 208

rubi, which are very similar in appearance to those of Euchelia.
The lizards first cautiously examined the larvae, and finally ate them.
After this Weismann reintroduced the E. jacobaeae larvae and the
lizards were seen to taste them, apparently mistaking them for the
edible L. rubi caterpillars.

Poulton ('87) carried out a most eareful and well-conducted
research upon the protective value of color and markings in insects
in reference to their vertebrate enemies. He experimented upon three
species of lizards and a tree-frog. Poulton combines his results with
those of other observers and presents them in the form of a table,
which certainly supports the suggestion of Wallace (67), that
brilliant and eonspicuous larvae would be refused as food by some
at least of their enemies. Poulton also shows that a limit to the
success of this method of defence (conspicuous larvae having
unpleasant taste or smell) would result from the hunger which the
success itself tends to produce, In the Tropics, indeed, where
insectivorous birds and lizards are far more numerous than with us,
and where competition for food is great among them, * we may feel
sure that some at least would be sufficiently enterprising to make the
best of unpleasant food, which has at least the advantage of bemg
easily seen and caught.” This last suggestion of Poulton certainly
seems reasonable ; moreover, it has occurred to me that young birds,
being but little skilled in the art of obtaining their food, might quite
often be forced by hunger to try various kinds of insects, and per-
haps even the Heliconidae themselves.

Beddard (92, p. 153-167) reports the results of an extensive
series of experiments carried out by Mr. Finn and himself upon
marmosets, birds, lizards, and toads. He arrives at conclusions which
are quite different from those of Poulton and others, but it appears
to me that his experiments were by no means so critically performed

as those of Poulton. He frequently threw larvae into a cage con-

taining many birds and observed them struggle for the prey. It
may well be, however, that a bird would be quite willing to swallow
a very unsavory mouthful in order to prevent any of its companions
from, apparently, enjoying it. However, Beddard found that
toads will eat any inseet without hesitation in spite of brilliant
coloration, strong odors, or stings, THe also found that birds and
marmosets would often deyour “conspicuously colored ? larvae with-
out any hesitation, and that some “protectively colored” or incon-
spicuous larvae were refused. There can be no doubt that many

204 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

insectivorous animals pay but little attention to the colors of their
prey; for example, it is well known to anglers that trout and salmon
will snap at the most gaudily colored “flies,” which may or may not
have any counterpart in nature.

The whole question of warning coloration will have to be made
the subject of an extensive research upon both old and young
insectivorous animals before we can safely arrive at any certain con-
clusions respecting it.

II. METHODS PURSUED IN STUDYING THE Coron-PATTERNS OF
THE HELICONIDAEK.

No comparative study of the color-patterns displayed by the
Heliconidae has ever been made. In fact, very few such studies
have been carried out upon any Lepidoptera. The only works I
know of are those of Eimer (89) and Haase (92) upon the colora-
tion of the Papilios, and of Dixey (90) upon the wing-markings of
certain genera of the Nymphalidae and Pieridae. "Phe family of the
Heliconidae with its numerous species and comparatively simple
coloration affords an excellent opportunity for such a research.

In making this study of the Heliconidae I was permitted through
the kindness of Mr. Samuel Henshaw to make free use of the collec-
tion in the Museum of Comparative Zoölogy at Harvard. I also
found the colored figures in the works of the following authors of
great service: Hewitson (56-776), C. und R. Felder (64—67), Hüb-
ner (0625), Humboldt et Bonpland (33), Cramer (1779-’82), Stau-
dinger (8488), Godman and Salvin (79—86), and Ménétriés (63) ;
likewise the following shorter papers published in various serials:
Dates (63, '65),Butler (65, 69, 6074, 77), Druce (76), Godman
and Salvin (80), Hewitson (54), Snellenen van Leeuwen (787), Srnka
(84, 785), Staudinger (82), and Weymer (75, 84). I was thus
enabled to examine the color-patterns of 400 (89%) of the species
of the Danaoid group, and of 129 (86%) of the Acraeoid group,
either from the inseets themselves or from figures given by the
authors named above. The remaining species were either inaccessi-
ble to me, or were so vaguely described as to be unavailable. A
list of the species known to me is given in Table 28.

(1) The Two Types of Coloration in the Danaoid Heliconidae.
It is very remarkable that the color-patterns of all of the Heliconidae

MAYER: COLOR AND COLOR-PATTERNS. 305

may be grouped into two very closely related types. To the one of
these I have given the name “ Melinaea type,” for it is characteristic
of most of the species of the genus Melinaea. It is well represented
by Figs. 46, 48, 49, 51, and 55-57 (Plate 4). The insects which
belong to this type possess wings colored with rufous, black, and
yellow,

The other type I designate as the “ Zthomia type,” for it is very
characteristic of most of the species of the genus Ithomia. Figs. 47
and 52 (Plate 4) afford examples of it. This type differs from the
Melinaea in that the rufous and yellow areas upon the wings have
become transparent.

There are, also, many species, found in numerous genera, which
fall between these two types of coloration, for the yellow and rufous
spots upon their wings have become translucent, so that one may
speak of them as “translucent yellow” and “translucent rufous.”
These spots are, so to speak, in process of becoming transparent, but
a few yellow or rufous scales still remain dusted over the otherwise
clear spaces. Most of the Dircennas are good examples of this type
(Fig. 54, Plate 4).

Of the 400 species of the Danaoid Heliconidae, about 125 belong
to the * Melinaea type." It is well represented by most of the
species of the genera Lycorea, Athyrtis, Ceratinia, Mechanitis, and
Melinaea. About 30 Ithomias and half a dozen Napeogenes also
belong to it. About 160 species belong to the * Ithomia type,” and
of this number fully 120 belong to the genus Ithomia. The others
are found in the genera Ceratinia, Napeogenes, Ituna, and Thyridia,
and many of them resemble the Ithomias so closely that they are
said to mimic them, About 100 species, some of which are found in
almost all of the genera, are intermediate in their color-patterns
between the Melinaea and the Ithomia types. The 15 remaining
species are represented by Melinaea gazoria (Fig. 53, Plate 4),

Jeratinia eupompe, and a few Ithomias, such as Ithomia hemixantho.
In these forms almost all color has disappeared, so that the whole
wing has become of a uniform dull translucent yellow, bordered on
the outer edges by a grayish black.

(2) Detailed Description of the Melinaea Type of Coloration.
Figs. 46, 48, 49, 51, and 55-57 (Plate 4) afford examples of this type
of coloration, In these insects we find the proximal half of the
central cell of the fore wing occupied by a rufous-colored area, which
I call the “inner rufous.” It is marked I in all of the figures.

206 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Beyond the “inner rufous” we find a black spot, marked II in the
figures. It usually occupies the middle region of the cell of the fore
wing, and I have designated it as the “inner black.” Beyond the
“inner black," and occupying most of the outer portion of the cell
of the fore wing, is a light-colored area, marked III in the figures.
This area is rufous in color in Fig. 49, but it is usually yellow, as in
Figs. 46, 48, 51,54—57, and I have called it the “inner yellow.”
Beyond the “inner yellow," and occupying the extreme outer
portion of the cell, lies the * middle black ” (IV). In many species
it is fused, as in Figs. 46-48,

56, 57, with the large black area, the
“outer black” (VII), which occupies the greater portion of the outer
half of the fore wing. Just outside of the cell beyond the “ middle
black” one finds a well-developed yellow area (V), the “ middle
yellow,” and there is sometimes still another yellow patch beyond
this, which is marked VI and called the “ outer yellow." Finally,
one often finds a row of white or yellow spots, the “ marginal spots"
(IX), lying very near the outer margin of the fore wing (see Figs,
47-49, 51, 54, 56). These spots are very well developed in the
genera Ceratinia, Napeogenes, Ithomia, and Melinaca. One more
very characteristic marking of the fore wing remains to be noticed ‘
that is the longitudinal black stripe (VIII). It is also worthy of note
that the front costal edge of the fore wing is almost always tinged
with black.

The pattern of the hind wing is quite simple. The ground color
is usually rufous and a “middle black” band (XI) runs across the
middle of the wing. The outer edge is bordered by the “outer
black” (XIII). Above the “middle black” band lies the “inner
rufous ” (X) of the hind wing, and below the “middle black ” band
one finds the “outer rufous” (XII) of the hind wing. One often
finds a row of white or yellow dots within the outer black border of
the hind wing, and these I designate as the « marginal spots" of the
hind wing.

The /thomia type of coloration, it will be remembered, may be
derived from the Melinaea, by simply imagining the rufous and
yellow areas to have become transparent. Also the outer black
usually suffers a reduction so as to become only a rather narrow
border along the outer margin of the fore wing. Thyridia psidii
(Fig. 47) is a good example of this type. It will be seen that the
black areas remain about the same as in the Melinaea type, but that.

MAYER: ‘COLOR AND COLOR-PATTERNS. 207

the rufous and yellow have become transparent. The middle and
outer yellow areas have also fused into a large transparent patch,

Ithomia sao (Fig. 52, Plate 4) is another good example of the
Ithomia type. In this particular species the “inner black ” of the
fore wing is absent, and the *middle black band " of the hind wing
has disappeared. When we come to consider the other Ithomias,
we shall find that in this genus it has probably fused with the
marginal blaek of the hind wing.

I have made a record of the color-variations that affect the
various charaeteristie areas just considered, and have recorded them
for every one of the species of the Danaoid and Acraeoid Heliconidae
known to me. As these records are too extensive for convenient
inspeetion, I have condensed the results, and they will be found
in Tables 1-27 inclusive. Thus, Table 1 gives the variations in

,

color of the “inner rufous” area of the fore wing for each genus

of the Danaoid Heliconidae ; Table 2 records the variations of the
“inner black“; Table 3 the “inner yellow” area, etc. In Table 1
we find, for example, opposite the genus Ituna, the number 2 in the
column labeled “transparent.” This indieates that in two species of

,

Ituna the “ inner rufous” area is transparent,

In order to facilitate the study of the color-patterns Dr. Daven-
port suggested that I make use of the ingenious projection method
invented by Keeler (93). This method consists in “squaring the
wing“ in the manner shown in Figs. 4 and 5 (Plate 1). In Fig. 4
the large reetangle (A, B, C, D) just at the right of the figure of the
hind wing represents a kind of. Mereator's projection of the wing
itself, The nervures 1°, 15», 2, 3, ete, are represented by the
vertical lines 1°, 15, 2, 3, etc., on the rectangle A, B, C, D. In cells
I, I^, and le, (bounded by nervures 1*, 1^, and 2,) one finds a
sinuous line winding across the middle of the cell. This line
appears in the same relative position upon the rectangle A, B, C, D.
The same is true of the eye-spot found in the cell bounded by
nervures 2 and 3, and of all the other markings of the wings.
The central cell of the wing itself is shown projected in the dotted
rectangle E, F, G, IH.

In the case of the fore wing (Fig. 5), the central cell of the wing
is dotted, and is shown projected upon the similarly dotted area
within the reetangle I, J, K, L. In other respects the method of
projection is the same as in the ease of the hind wing.

In this manner the colors displayed by various species of. Danaoid

208 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

and Acraeoid Heliconidae have been represented in color in Plates
5-8. Each large rectangle upon the left hand side of the Plate
represents a hind wing, the small middle rectangles show the colors
of the cell of the hind wing, and the right hand rectangles give the
fore wings, all being projected in the manner illustrated in Figs. 4
and 5, Plate 1. The chief advantage in Keeler's projection method
lies in the fact, that similar areas in the projection of the wings lie
vertically under or over one another, and thus by merely glancing
up or down the plates one may observe the color-variations which
occur in homologous cells of all the species represented.

IIT. Generar Discussion or THE Coron-PATTERNS AND OF

MIMICRY IN THE GENERA H ELICONIUS AND EUEIDES

Among the species of the genera Heliconius and Eueides we find
remarkably little variation in venation, but great diversity in color-
pattern of the wings, and in this respect they are very different from
the Danaoid Heliconidae, where, it will be remembered, we find fully
twenty different types of venation and only two types of color-
pattern.

(1) The Four Color Types in the Genus Heliconius. Schatz
and Rober ('85—92) divide the species of the genus Heliconius into
four groups based on color differences, as follows:—(1) the
“Antiochus group” (Plate 4, Fig. 50); (2) the * Erato group?
(Fig. 60); (3) i “Melpomene group" (Fig. 59); and (4) the
“Sylvanus group,” a good example of which is Heliconius eucrate
(Fig. 58, Plate 4).

It will become apparent through an inspection of Figs. 50, 60, 59,
and 58, which represent respectively, Heliconius antiochus, H. erato,
II. melpomene, and IT. eucrate, that the first three are quite closely
related in color-pattern, while the fourth (II. euerate) approaches
very closely to the plan of coloration of the Melinaea type of the
Danaoid Heliconidae. In fact this resemblance is so close that it
may be safely said that the members of the “ Sylvanus group," to
which IT. euerate belongs, mimic the Danaoid Heliconidae.

The **Antiochus group" is represented by Helieonius anti-
ochus (Plate 4, Fig. 50, and Plate 5, Fig 62). H. sara, H. galanthus
? , S , 5 ,

and IT. charitonius (Plate 5, Figs. 61, 63, 64) are also members of
this group; other examples are H. apseudes, H. cydno, II. chiones,
IT. hahnesi, IT. sappho, IT. leuce, H. eleusinus, and H. clysonymus.

MAYER: COLOR AND COLOR-PATTERNS. 309

These species are characterized by their blue iridescence, and the
narrow yellow or white bands upon the primaries ; the hind wings
are pointed at the outer apex, and the venation approaches the type
found in Eueides aliphera. II. ricini (Plate 5, Fig. 66) is a good
example of a form intermediate in coloration between group 1 and
the “ Erato group” (2).

The type of group 2 is Heliconius erato (Plate 4, Fig. 60, and
Plate 5, Figs. 67 and 68). This group is closely allied to group 1 in
its characteristics. A good connecting link between groups 1 and 3,
the * Melpomene group," is H. phyllis (Fig. 65).

The third, or * Melpomene group," is represented by H. mel-
pomene, H. callicopis, H. cybele, II. thelxiope, and II. vesta (Plate 6,
Figs. 70-74, and Plate 4, Fig. 59). II. vulcanus, II. venus, II.
chestertonii, H. burneyi, and II. pachinus are also examples of this
group.

(2) Mimicry between the Genus Heliconius and the Danaoid
Group. To Schatz's group 4, the * Sylvanus group,” belong all those
species of Heliconius which have departed widely from the colora-
tion pattern of the other three groups, and have come to resemble
various species of the genera Melinaea, Mechanitis, and Tithorea of
the Danaoid Heliconidae. II. eucoma, II. euerate, II. dryalus, and H.
sylvana (Plate 8, Figs. 88, 89, 91, and 95) are, good examples of
group 4. By glancing at the diagrams on Plate 8 it will be seen
that, H. dryalus resembles Melinaea paraiya very closely ; in faet, the
likeness is so elose that it is almost certain that no eye could distin-
guish between the two insects when they are upon the wing. Another
startling resemblance is that between II. eucrate and Melinaea thera
(Plate 8, Figs. 91 and 92); moreover, there is but little difference
between the color-patterns of II. eucrate, Eueides dianasa, and
Mechanitis polymnia (Figs. 91, 95, and 94). II. sylvana and
Melinaea egina (Figs. 95 and 96) are also said to mimie each other.
The resemblance certainly appears very close at a casual glance, yet
when the colors are plotted, as in Figs. 95 and 96, the differences be-
come quite apparent. II. claudia (Plate 5, Fig. 69) is a good con-
neeting link between the Sylvanus group and the Melpomene group.
In both the Melpomene and Sylvanus groups the venation has departed
from the Eueides aliphera type, and the contour of the hind wings is
much more rounded. and elliptieal than is the case in the Antiochus
and Erato groups. (Compare Figs. 50 and 60 with Figs, 58 and 59,
Plate 4.) There are rather less than twenty species which certainly

210 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

belong to the Sylvanus group; among them may be mentioned,
in addition to those already spoken of, Heliconius numata, which
resembles Melinaea mneme and Tithorea harmonia ; H. zuleica, which
resembles a Mechanitis and is a good copy of Melinaea hezia; and
IT. metalilis, which is said to mimie Melinaea lilis; there are also
striking resemblances between

H. aurora and Melinaea lucifer ; H. messene and Melinaea mesenina ;
H. eucrate and Mechanitis lysimnia ; H. hecalesia and Tithorea hecalesina ;
H. hecuba and Tithorea bonplandii ; H. ethra and Mechanitis nesaea ;

H. formosus and Tithorea penthias ; II. pardalinus and Melinaea pardalis ;
II. telchina and Melinaea imitata ; H. ismenius and Melinaea messatis.

Most remarkable of all perhaps is the close resemblance between
ITeliconius aristiona, Mechanitis methone, and Ithomia fallax of
Staudinger. In fact, Staudinger states in his “ Exotische Schmet-
terlinge” that he hesitated for some time to describe Ithomia fallax
on account of its close resemblance to Hewitson’s Mechanitis methone.
Good lists of the Heliconidae which are said to mimic one another
are given by Wallace (89, p. 250, 251), and by Haase (98%, p.
146, 147).

(3) The Three Color-Types in the Genus Eueides. In the
genus Eueides we meet with three color-types represented by
E. aliphera, E. thales, and E. cleobaea. These insects are dis-
tinctly smaller than the species of the genus Heliconius, and the
yellow spots upon their primaries are more ocherous in color than
in Heliconius. E. aliphera (Plate 6, Fig. 77) represents the most
highly specialized color-type. Eueides mereaui (Fig. 76), however,
is a good connecting link between the color-patterns of E, aliphera
and E. thales (Fig. 75), and E. thales is almost identical in color-
pattern with Heliconius vesta (Fig. 74).

The other type of Eueides is represented by E. cleobaea, E.
dianasa, E. isabella, ete. (Plate 6, Fig. 78, and Plate 8, Fig. 93).
These resemble the Sylvanus group of Heliconius or various Melinacas
and Mechanitis.

(4) Detailed Discussion of Plates 5-8. PLA 5 is intended
to illustrate the types of coloration found in the Antiochus and
Erato groups of the genus Helieonius, In II. sara (Fig. 61) the
wings are suffused with a dark blue iridescence, and some narrow
yellow bands of color are found upon the primaries, In H. antiochus
(Fig. 62) we find similar bands of color upon the primaries, but
they are changed to white. II. antiochus may have descended

MAYER: COLOR AND COLOR-PATTERNS. 211

from an albinie sport of II. sara. In II. galanthus (Fig. 63)
the white areas have greatly increased in size, and the iridescent
blue has become much lighter. In H. charitonius (Fig. 64) we
find the wings erossed by yellow spots and bands, but in some speci-
mens this yellow color exhibits a decidedly reddish tinge. The figure
of H. charitonius in Staudinger’s “ Exotische Schmetterlinge ” illus-
trates this peculiarity ; indeed, spots which are commonly yellow are
often found red, and vice versa. In II. phyllis (Fig. 65) we find
along the upper part of the diagram of the hind wing a yellow mark-
ing, and a similarly shaped red mark is found in its near ally, H.
thelxiope (Fig. 73, Plate 6). The same is also true of II. ricini
(Fig. 66, Plate 5).

H. erato (Figs. 67 and 68, Plate 5, and Fig. 60, Plate 4) is very
remarkable, for there are no less than four distinct color-ty pes
exhibited by different individuals of this species; one of them (Fig.
67) shows the basal half of the hind wing marked by six red tongues
of color edged with iridescent blue, and there is a dark rufous
suffusion upon some parts of the fore wing. In other specimens
(Fig. 68) the red tongues of color which characterized the hind wing
of Fig. 67 are almost absent, and only the blue iridescence is left;
also there is no rufous to be seen upon the fore wing. In another
type the blue iridescence of the hind wing has become green, and in
still other specimens the yellow stripes upon the fore wing have
become white.

As one looks over the diagrams upon Plates 5-8, it becomes evi-
dent that yellow frequently changes to white, for we often find one
or two species of a genus which exhibit white spots identical in shape
and position with spots which are yellow in most of the others. Good
examples of this are H. antiochus (Plate 5, Fig. 62), Melinaea
parallelis and Ceratinia leucania (Plate 7, Figs. 82 and 83) ; likewise
the white spot near the outer apex of the fore wing in II. euerate
(Plate 8, Fig. 91), which is yellow in many individuals. Yellow
areas are also frequently changed to rufous or red ; thus the yellow
basal half of the hind wing of II. eucrate (Plate 8, Fig. 91) is often
found of a rufous tinge in individual specimens of the species, and
among the specimens of this species in the Museum of Comparative
Zoology one ean trace a gradation of this area from bright yellow
to rufous. II. claudia (Plate 5, Fig. 69) is introduced in order to
exhibit some of the differences between the “Sylvanus” group, to
which it belongs, and the * Antiochus? and ** Erato? groups.

212 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Puare 6 is intended to exhibit the characteristic color-patterns
found in the Melpomene group and in the genus Eueides. Fig. 70
represents IT. melpomene, and Fig. 71 its near ally, H. callycopis, in
which the red area of the fore wing has become broken up, and some
red spots have made their appearance near the base of the hind wing.
In the next variety of H. melpomene, H. cybele (Fig. 72), it is
remarkable that the pattern of the fore wing has come to resemble
the Sylvanus type, and is identical in general plan of coloration with
the fore wings of the Melinaeas or Mechanitis (see Figs. 84 or 85,
Plate 7, or Figs. 92 or 94, Plate 8). In its close ally, H. thelxiope
(Fig. 73), a still nearer approach to the ‚Melinaea type has come
about by the development of a black band across the middle of the
hind wing, and one has only to imagine a general fusion of the seven
club-shaped red stripes of the hind wing in Fig. 73, Plate 6, in order
to produce exactly the Melinaea type as exhibited, for example, by
Eueides cleobaea (Fig. 78). In this connection it is worthy of note
that Bates (62) showed that H. thelxiope was derived from H.
melpomene, there being between the two many intermediate forms.

H. vesta (Fig. 74) is evidently a close relative of H. thelxiope,
and what is still more worthy of note is, that it is almost identical in
the general effect of its color-pattern with Eueides thales (Fig. 75)!
The yellow spots upon the fore wing of E. thales are, however, duller
in hue than are those of H. vesta, and the insects are somewhat
different in size, H. vesta spreading 78 mm., while E. thales spreads
only 66 mm. It will be noticed that the chief difference between
the color-patterns of these two species lies in the fact, that, while the
black stripes of the hind wings in H. vesta lie along the nervures, in
Eueides tuales they occupy the middle of the cells themselves. The
general resemblance of the two color-patterns may of course be
merely aecidental. An easy explanation, however, is afforded by
the theory of mimiery, for the two species look very much alike
until one subjects their color-patterns to close analysis, when
remarkable differences appear. E. thales (Fig. 75) may have been
derived from some such form as E. mereaui (Fig. 76), for one has
merely to imagine a greater development of the black and a general
deepening of the rufous upon the hind wing of E. mereaui to make
it resemble E. thales quite elosely. Finally, in E. aliphera (Fig. 77)
the black serrated border of the hind wing is still more reduced, and
the black stripe which crosses the cell of the fore wing in E. mereaui
is mot present.

MAYER: COLOR AND COLOR-PATTERNS. 218

PrATE 7 is intended to illustrate the peculiarities of color-pattern
found among the Danaoid Heliconidae. Thyridia psidii (Fig. 79)
is an example of the transparent type of color-pattern found among
the Danaoid Heliconidae, and especially prevalent among the
Ithomias. It will be seen by comparing Fig. 79 with the other
figures upon Plates 7 and 8, that the chief difference lies in the fact,
that in this type: both the rufous and yellow areas have become
transparent. The black area of the fore wing has also suffered a
reduction, especially along the outer margin of the wing. Inci-
dentally it should be mentioned, that in this particular species the
middle blaek band of the hind wing has become tilted up at a sharp
angle, instead of erossing the wing horizontally. A life-size figure
of the wings of Thyridia psidii is given on Plate 4, Fig. 47.

In Napeogenes cyrianassa (Fig. 80) and Ceratinia vallonia (Fig.
81) portions of the usually yellow and rufous areas have become
transparent.

The spots upon the fore wing of the Melinaeas are usually yellow,
but in Melinaea parallelis (Fig. 82) they are white. It would seem
that this form may have descended from some albinie sport.
Ceratinia leucania (Fig. 83) resembles Melinaea parallelis so closely
in general plan of coloration, that it is very diflicult to distinguish
between them, even when the two insects are seen side by side.
Jeratinia leucania, however, is somewhat smaller than Melinaea
parallelis. Both occupy the same region in Central America, and
the specimens from which the diagrams were drawn came from
Panama.

Figs. 84-87 are drawn from various specimens of Mechanitis
isthmia, all from Panama, They are intended to give some idea of
the range of individual variation which is met with in this extremely
variable form The contraction of the middle black band of the
hind wing in this form has already been noticed in the general
discussion of the laws of color-pattern (see page 184). In Fig 87 it
will be seen that the inner yellow stripe which usually crosses the cell
of the fore wing has become very narrow and changed to a rufous
eolor. However, upon the under surface of the wing it still remains
as a yellow stripe. Indeed, in most color-changes the upper side of
the wing seems to take the initiative, the under surface being more
sonservative. This is not true, however, in the Ithomias, where
the blaek areas of the under side of the wings often are found to be
rufous in color, while they still remain of the normal black upon the

214 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

upper surface. The colors of the under surface are, however,
usually identical with those of the upper, though they are always
duller in hue. This may be due to the fact, that the colors of the
upper surface are more frequently seen than those of the lower, for
these insects often float lazily along with their wings horizontally
extended. The operation of Natural Selection would then be more
severe with the upper surfaces than with the lower.

PLATE 8 gives an analysis of the color-patterns of some of the
Heliconinae and those Melinaeas, etc., which they resemble. H.
eucoma (Fig. 88) is a good example of the Sylvanus type, and with
its rufous, yellow, and black wings it is certainly a wonderfully close
copy of the color-pattern found so commonly among the species of
the genus Melinaea of the Danaoid Heliconidae.

Heliconius dryalus and Melinaea paraiya (Figs. 89, 90) resemble
each other so closely in size, shape, and &oloration, that it must be
impossible to distinguish between them when the butterflies are in
flight; yet an analysis of their color-patterns shows that there are
considerable differences between them. The shape of the yellow
bands upon the fore wings is quite different; the inner black spot
within the cell is double in Melinaea paraiya, and there is also a row
of white spots along the margin of the fore wing.

A much closer resemblance is found between II. eucrate and
Melinaea thera (Figs. 91 and 92), where the Heliconius is almost a
true copy of the Melinaea.

The color-patterns of Kueides dianasa (Fig. 93) and Mechanitis
polymnia (Fig. 94) are also very nearly the same. Both are
common species in Brazil.

Heliconius sylvana is said by Bates and by Wallace to mimic
Melinaea egina. It will be seen by reference to Figs. 95 and 96
that their color-patterns are quite different in detail, yet the insects
look very much alike when placed side by side, and may easily be
mistaken for each other when upon the wing. Melinaea egina is
much more common than Heliconius sylvana.

IV. GeneraL Discussion or tie ConLoR-PATTERNS AND OF
MrwicRY AMONG THE DANAOID HELICONIDAE.

(1) The Origin of the Two Types of Coloration, The character
of the variation in the Danaoid Heliconidae is very different from
that of the genera Heliconius and Eueides, for while there is great

MAYER: COLOR AND COLOR-PATTERNS. 210

diversity of color-pattern and very little variation in venation among
the species of the Acraeoid group, exactly the opposite condition is
met with in the Danaoid group, where we find at least twenty
different types of venation and only two types of color-pattern.
One of these types of coloration is well exemplified by most of
the Melinaeas (Fig. 48, Plate 4), and I have therefore called it the
« Melinaea” type. The other type is exemplified by most of the
Ithomias (Figs. 47 and 52) and has been designated in this paper as
the “ Zthomia" type. In the Melinaeas, it will be remembered, we
find the rufous and black wings erossed by bands of yellow ; while
in the Ithomias, on the other hand, the rufous and yellow areas have
become transparent, often leaving the wing as clear as glass, and the
black, which is so characteristic of the outer half of the wing in the
Melinaea type, has shrunk away until it has come to lie along the
outer margin of the wing only.

By a study of all the genera of Danaoid Heliconidae we gain light
upon the question of the origin of the * Melinaea" and “ Ithomia "
types of coloration, As we have seen (page 198), the Danaoid
Heliconidae are an offshoot from the great family Danaidae. Indeed,
two of the genera, Lycorea and Ituna, are so closely related to the
Danaidae that Schatz and Röber (85—92) propose to include them
within that family. There can be but little doubt that Lycorea and
Ituna are remnants of the ancestral forms which long ago shot off from
the Danaidae to form the Danaoid Heliconidae ; and it is interesting to
note, that in these two patriarchal genera we find the two distinct
types of eolor-pattern which are exhibited by the Danaoid Helico-
nidae, for all of the five known species of Lycorea are good examples
of the Melinaea type (see Lycorea ceres, Fig. 46, Plate 4), while the
four known species of Ituna all exhibit the transparent, or Ithomia,
type of coloration. In fact, in their color-patterns the species of
Ituna remind one of gigantic Ithomias. The species of Lycorea,
however, are colored very much after the pattern of the Danaidae,
and indeed they have departed but little from the type of the
members of the great family whence they sprang. On this account
I believe that the Melinaea type of coloration, which is so charac-
teristic of the species of Lycorea, is phylogenetically older than
the Ithomia type.

In order to account for the origin of the Ithomia type, we may
assume that, shortly after the primeval forms of the Danaoid: Heli-
conidae began to segregate out from thé Danaidae, the species were

216 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

few and probably rare. Under these circumstances any given insect
would gain but little advantage by resembling merely the general
type of the coloration of its fellows. For the relative advantage
gained by such imitation, according to Fritz Müller's law, increases
inversely as the square of the fraction whose numerator is the actual
number of the imitating form and whose denominator is the actual
number of the imitated. Therefore when the insects were still rare
there would be few to imitate and consequently but little advantage
would be gained by the imitation. Imagine, for example, that a
single insect happens to imitate the color-pattern of a group of 100,
and that the advantage gained thereby is represented by the number
1; it is evident from Fritz Miiller’s Jaw that, if it happened to
imitate the coloration of a group of 1,000, its relative advantage
would be 100 instead of 1. We see, then, that mimicry within the
group of the Danaoid Heliconidae became an important factor only
after the group was well established and the insects became common.
During the early history of the race, then, there would be but little
tendency towards conservatism of color-patterns, and when the
*Ithomia" and “ Melinaea” types of coloration made their appear-
ance, they both survived and now serve as the patterns for mimiery ;
and this accounts very well for the remarkable faet, that there
are no other types of coloration than these two to be found within
the whole group with its 450 species!

(2) Mimicry among the Danaoid Heliconidae. The genus
Ithomia with its 230 species is the dominant genus of the Danaoid
group, and in nearly all of the other genera individual species are
found which have departed widely from their generic type of
coloration and have assumed the clear wings of the Ithomias. A
good idea of how far these interesting individuals may depart from
the coloration of their type may be gained by comparing Fig. 53,
Plate 4, which represents Melinaea gazoria, with Fig. 48, which
represents a typical Melinaea (M. paraiya). It is evident that
Melinaea gazoria is startlingly like an Ithomia both in size and
coloration, although it retains the venation and generic charac-
teristics of a Melinaea,

In Mechanitis, which is the most independent genus of the
Melinaea type of coloration, all of the species are fair examples of
the Melinaea type, except Mechanitis ortygia Druce, from Peru.
Druce (’76) in his description of this curious little species states in
astonishment that it possesses the venation of a Mechanitis, but the
size and coloration of an Ithomia !

MAYER: COLOR AND COLOR-PATTERNS. 217

It is quite remarkable that although the genera Melinaea and
Mechanitis serve as models of mimicry for the Acraeoid Heliconidae,
they should themselves mimic Ithomia.

The genus Ithomia is, however, the most independent of all the
genera of the Danaoid group, and I know of remarkably few good
instances in which an Ithomia has apparently departed from the
coloration of its type to assume the guise of the Melinaeas. One good
example of such a change, however, is afforded by Ithomia fallax of
Southern Peru, which resembles either Mechanitis methone or Heli-
conius aristiona of Colombia (see page 210). There is apparently
a difficulty in ascribing this resemblance to mimicry, for the imitator
and imitated do not occupy the same geographical regions.

In direct contrast with the independence of the Ithomias stands
the case of the genus Napeogenes ; for Godman and Salvin (’79-’86)
say of Napeogenes, that nearly every species mimics some Ithomia
which occupies the same district; and thus almost the very existence
of the genus would seem to depend upon its mimicry of Ithomia.

It is not the purpose of this paper to discuss, in detail, the numerous
interesting cases of mimicry which are believed to exist between
members of the Danaoid Heliconidae. An excellent discussion of
such cases, and of the relationships of the various genera, has been
given by Haase ('93*, p. 116-127).

V. QUANTITATIVE DETERMINATION OF THE VARIATIONS OF THE
CHARACTERISTIC WING-MARKINGS IN THE ACRAEOID AND

DANAOID HELICONIDAR.

(1) Variations of “Inner Rufous” Areas of the Fore and
Hind Wings. Table 1 gives the color-variations which are exhibited
by the “inner rufous” area of the fore wings in the Danaoid Heli-
conidae. This area is marked I in all of the figures upon Plate 4.
We learn from an inspeetion of Table 1 that this area is rufous in
color in 124 species of the Danaoid Ieliconidae, transparent in 152,
black in 24, and that in the remainder it is more or less translucent,
and of either a yellowish or rufous tinge.

Table 10 shows the variations which come over the “inner
rufous” area of the hind wings of the Danaoid Helieonidae. This
area is marked X in the figures upon Plate 4. It is apparent at
a glance that the variations which affect the inner rufous areas of

218 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

both fore and hind wings are very similar. In order to exhibit this
fact graphically, the color-variations have been laid off upon the
diagram, Fig. 97, Plate 9. The base line is marked at equal inter-
vals with the words “rufous,” “translucent rufous,” “translucent,
slightly rufous,” “transparent,” etc., and the ordinates show the
number of species which exhibit the various colors, rufous, trans-
lucent rufous, etc. For example, at the point “translucent rufous ”
we find that the ordinate is 23; this indicates that in 23 species
the area is translucent rufous in color. The points, thus found
upon the ordinates are successively joined by straight lines form-
ing a zig-zag figure. The full line represents the fore wing, and
the dotted line the hind wing, and it becomes clearly evident from
the closeness of these two zig-zag lines that the color of the inner
‘rufous area of the fore wing (area I, Plate 4) is almost always
sure to be identical with that of the inner rufous area of the hind
wing (area X, Plate 4). We see, therefore, that whatever color-
variation affects the inner rufous area of the fore wing, this area
in the hind wing is almost always affected in the same manner.

Fig. 99, Plate 9, is derived from Tables 15 and 24, which show
the color-variations in the fore and hind wings of the genera Helico-
nius and Eueides, It is seen that here also the colors of these two
areas in both the fore and hind wings are almost always identical.
We here meet with one of those interesitng physiological laws
which are independent of Natural Selection, and the m 'aning of
which remains a mystery, for surely we can see no reason on the
ground of adaptation why similar areas upon both fore and hind
wing should bear similar colors.

(2) The “ Inner Black” Spot. , Table 2 shows the presence or
absence of the “inner black” spot in the Danaoid Heliconidae.
This spot is marked II in the figures upon Plate 4 When pres-
ent, it is always black in color and is usually found occupying the
middle region of the cell of the fore wing. The table shows that
it is about an even chance whether it be present or not, for it is
absent in 210 species and present in 190. In the genus Ithomia,
however, it is present in only one third of the species. What is
most worthy of note concerning it, is the faet that it almost, always
appears, when present, as a single spot. Indeed, it appears as a
double spot in only 7 species, and 5 of these belong to the genus
Melinaea. A good example of its appearance as a double spot is
found in Melinaea paraiya (Fig. 48, Plate 4). It will be remem-

MAYER: COLOR AND COLOR-PATTERNS. 219

bered that there are 450 species in the Danaoid group; 25 of
these belong to the genus Melinaea; yet among these 25 we find
5 exhibiting this marking as a double spot. Assuming that the
doubling of this spot has arisen in each species as a sport, and that
such a sport is as likely to appear in one species as in any other of
the Danaoid group, then the chances against five such sports
appearing among the 25 Melinaeas is a OY or about
2,830,000 to 1. Indeed, it is probable that all five of the species
of Melinaea which exhibit the doubling of this spot are descend-
ants of a single ancestor in which it appeared for the first time
double, for the mathematical chance that one such ancestor should
appear among the Melinaeas, rather than in any other genus, is

8 r 5 . : € 3

evidently 1 in =, or one chance in eighteen, The chance against
v 5 e E

two such unrelated ancestors is, however, 25 24 or about 336 to

450X449X448 | 6 rà
25 24% » OF 6,560 to 1, ete.

By reference to Table 16 we find that in the genera Heliconius
and Eueides the inner black area is black or iridescent blue in all
of the species of Heliconius, but absent in 5 of the 18 species of
Eueides known to me. These 5 include Eueides aliphera and its
allies. Now there are 150 known species of the Acraeoid Helico-
nidae, and 24 of these belong to the genus Eueides; so it is evident

1, and the chance against three is

that the mathematical chance against the supposition that five sports
arose independently in the genus Eueides, in which the inner black

vegan dat „ 150X149 148K 147K 140 4« Sed
was absent, is given by rg span » or 13,900 to 1. It is there-

fore probable that the five Kueides lacking the inner black are

the descendants of a single ancestor,

(8) Variations of the “Inner Yellow" and “ Middle Yellow”
Areas. Tables 3 and 5, and diagram Fig. 98, Plate 9, show the
color-variations of the “inner yellow” and “middle yellow” areas
in the fore wings of the Danaoid Heliconidae. These areas are
marked III and V, respectively, in the figures upon Plate 4. The

N

“inner yellow ” area, it will be remembered, occupies the outer por-

tion of the cell of the fore wing; while the “middle yellow” is found
in the region just beyond the outer limits of the cell. The two areas
are often fused together as in Figs. 47, 48, 50, 51, and 55, Plate 4.
The inner yellow area is usually smaller than the middle yellow,
and a comparison of Tables 3 and 5 will show that it is much more
frequently obliterated by the encroachment of the rufous or black

220 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

areas which surround it; for example, while the middle yellow is
rufous in color in only 14 species, the inner yellow is rufous in 56;
also the inner yellow area, being usually smaller and less conspicuous
than the middle yellow, is less important in cases of mimicry, and
the diagram Fig. 98, Plate 9, shows that it is much more variable
in color than the middle yellow. The full zig-zag line in this figure
represents the color-variations of the inner yellow, while the dotted
zig-zag line gives the color-variations of the middle yellow. As
there are nine color-variations displayed by each of these two areas,
and as there are 400 species of the Danaoid Heliconidae recorded
by me, it becomes evident that, if there were no color preferences
displayed by these areas, there would probably be about 499, or 44.4,
species which would display it as rufous, 44.4 translucent, 44.4 yellow,
etc. 'The heavy, straight, dotted line (Fig. 98, Plate 9) represents
this ideal condition, which would be approximately realized were
one color as likely to occur as another in the respective areas. Now
it is evident from an inspection of the figure, that the full zig-zag
line, which represents the color-variations of the “inner yellow,”
approaches the straight line condition more nearly than does the
dotted zig-zag line, which represents the middle yellow.! The
inner yellow is therefore more liable to color-variations than the
middle yellow; and this is what we should expect on account of
its comparatively small size and its consequent inconspicuousness
as a characteristic marking in cases of mimicry.

A. comparison of Figs. 97 and 98, Plate 9, is interesting, for it
shows that the color-variations of the inner rufous are quite similar
to those of the inner yellow and middle yellow. This serves to
illustrate the close physiological relationship which exists between
rufous and yellow. The two pigments are probably closely related
chemically, for every ordinarily rufous area is sometimes found to be
yellow, and vice versa. Yellow areas also often change to white.
Rufous, yellow, and white are evidently closely related color-vari-
ations.?

1 This is not true for one color, white.

2 It may be well to mention here that the black areas upon the wings are subject to
very little color-variation, In some cases, however, especially upon the under surfaces
of the wings in Ithomia, the black has changed toa rufous or russet color. For example,
Table 4 shows that the middle black area (IV in the figures upon Plate 4) is rufous in
only 12 species out of the 400 which are recorded, and all of these 12 are Ithomias. Also
Tables 7 and 13 show that the outer black of the fore wing, and the outer black of the
hind wing are russet in 22 and 11 species, respectively, Evidently, black is a far more
conservative color than rufous, yellow, or white, Probably black is also quite different
from the other pigments chemically.

MAYER: COLOR AND COLOR-PATTERNS. 221

Tables 17 and 19 show the color-variations affecting the “inner
yellow ” and * middle yellow” areas of the fore wing in Heliconius
and Eueides, There is but little difference between the two tables,
except that in 15 species of Heliconius the inner yellow is suffused
with blaek or blue, while the middle yellow is never suffused by the
outer black which surrounds it. Fig. 100, Plate 10, exhibits
graphieally the color-variation of these two areas. The “inner
yellow” is represented as a full line, and the * middle yellow” as
a dotted zig-zag. It is evident that here also the inner yellow is
more variable in color than the middle yellow, for not only does the
inner yellow area display two more colors, but its chart is a flatter
zig-zag.

(4) Variations of the * Middle Black” Mark of the Fore Wing.
Table 4 shows the color-variation of the middle black mark (area
IV in figures upon Plate 4). This marking lies along the extreme
outer border of the central cell of the fore wing. It is small in area,
but is rendered very conspicuous from the fact that it is situated be-
tween the inner yellow and middle yellow markings. In spite of
its small size, however, it is a remarkably permanent marking, for
Table 4 shows that it is absent in only 20 out of 400 Danaoid Heli-
conidae. In these 20 it has been obliterated by the fusion of the
inner and middle yellow areas. It is worthy of note that in 12
Ithomias it has become rufous in color. This change to rufous is
the only color-change which the black areas of the wings ever
display.

Table 18 shows the variations of the middle black area for
Heliconius and Eueides.

(b) Variations of the * Outer Yellow” Area of the Fore Wing.
Table 6 shows the variations which affect the outer yellow area
of the fore wings in the Danaoid Heliconidae. This area is m irked
VU in the figures upon Plate 4; it lies beyond the region of the
middle yellow, but is usually more or less fused with it. Table 6 is
only approximately correct, owing to the difficulty in many cases of
deciding whether the middle and outer yellow be really fused or not.
[t will be seen that in the genus Ithomia the middle and outer
yellows are wholly fused in about 200 species. This is one of the
marked eharacteristies of this very independent genus.

Table 90 shows the color-variations of the outer yellow area in
Heliconius and Eueides, This marking is present in 81 and absent
in 48 of the Acracoid group. It is much more widely separated
from the middle yellow than is the ease in the Danaoid group.

222 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

(6) The relative Permanency of the Black Areas upon the Fore
and Hind Wings. A study of the relative permanency of the
various characteristic black markings upon the wings is of interest,
for, if the generally accepted idea concerning the prevalence of
mimicry within the group of the Danaoid Heliconidae be true, we
should expect the most conspicuous markings to be the most perma-
nent, for they are evidently of the most importance for mimicry.
This is, however, not the case for the black markings, A good
example of this fact is afforded by a comparison of the relative
permanency of the black streak which extends along the extreme
costal edge of the fore wing with the inner black spot (II in figures
on Plate 4). The inner black spot is certainly a more conspicuous
marking than this narrow black streak along the costal edge; yet it
is much more variable, for Table 2 shows that it is present in 210
and absent in 190 of the 400 Danaoid Heliconidae. In other words,
it is about as likely to be present as absent. The black streak upon
the costal edge, on the other hand, is much more permanent, for it is
absent in only 52 species out of the 400.

Another good example of the inaccurracy of the supposition that
large and conspicuously colored areas are always less variable than
small ones, is derived from a comparison of the relative variability
of the large outer black of the fore wing with the small outer
black of the hind wing. Although the outer black area of the fore
wing is usually much larger and more conspicuous than the outer
black margin of the hind wing, it is more variable in color, for it is
rufous in 22 species, while the outer black of the hind wing is
rufous in only 11, out of the 400,

In general, however, large colored areas are more permanent than
small ones, as was found in the case of the inner and middle yellow
areas (see page 220). Indeed, a good instance of this greater vari-
ability of small color areas is afforded by the longitudinal black
stripe marked VIII in the figures of Plate 4, for this is more
variable than the larger outer black area of the fore wing.

(7) The “ Middle Black Stripe” of the Hind Wing. In the
genus Ithomia the middle black stripe (XI, Plate 4) has migrated
downward, so that in many species it has become fused with the
outer black margin, as in Ithomia sao (Fig. 52, Plate 4). In other
cases there is still to be seen a narrow line of rufous color between
the middle black band and the outer black margin of the hind
wing. Such is the case in Ithomia nise (Fig. 54, Plate 4). In

MAYER: COLOR AND COLOR-PATTERNS 223

many. other cases the outer black and middle black are completely
fused, so far as the upper surface of the wings is concerned ; but,
hind wings, it will be
sts between the middle

if one examines the, under surface of the

found that a narrow rufous streak still per:
black band and the outer black margin of the hind wing.

(8) Variations of the Marginal Spots of the Fore Wing. The
marginal spots are found very near the outer margin of the fore
wing; they are usually either yellow or white, but in some few
cases they are rufous. It appears from Table 9 that they are
present in 146 and absent in 254 species of the 400 Danaoid Heli-
conidae known to me. Fig. 101, Plate 10, shows graphically
the manner in which these spots occur in those species which
possess them. It is evident from this curve that the number of
these spots is not determined merely by chance, for they show a
marked tendency to appear either as 2 or 3, or as 6 or 7 spots.
It is due to this fact, that there are two maximum points upon
the diagram Fig. 101, Plate 10. In those species which exhibit the
«9. or 3-spot” condition, the spots are found near the front apex
of the fore wing. ‚In the “ 6- or 7-spot” condition they lie all along
the outer margin of the fore wing, one spot in each cell. In the
genera Ithomia, Napeogenes, and especially in Ceratinia these mar-
ginal spots have become large and conspicuous ornaments. (See
Fig. 49, Plate 4.)

Table 22 shows the manner of appearanee of these spots in the
genera Heliconius and Eueides. They are found in only 26 species
of the 129 known to me; and this number is far too small to war-
rant general conclusions concerning the order of their appearance,

(9) The Marginal Spots of the Hind Wing. Table 14 illus-
trates the manner in whieh the marginal, spots of the hind wings
make their appearance. They are absent in 279 and present in
121 of the 400 species of the Danaoid group. Thus they occur
rather less frequently than the marginal spots of. the fore wing. In
the 121 species in which these spots are found they show a decided
tendeney to appear either as 4 or as 5 spots. Fig. 102, Plate 10,
is a graphie representation of the distribution of these spots, derived
from Table 14. It appears that the outline of the figure approaches
a probability curve, and is approximately symmetrical about the

mean ordinate (A, B), situated at 4.54.

224 BULLETIN: MUSEUM OF COMPARATIVE ZOÓLOGY.

VI. COMPARISON or THE CorLoR-VARIATIONS OF THE PAPILIOS OF

SouTH AMERICA WITH THOSE OF THE HELICONIDAE.

In order to emphasize the peculiarities of the coloration of the
Heliconidae, I will conclude by instituting a comparison between
their variations and those of the South. American Papilios. There
are about 200 species of Papilio in South America, and these display
in all 36 distinet colors. The colors have been determined by
reference to the plates in Ridgway’s * Nomenclature of color for
naturalists.” A list of the colors which are displayed by these Papi-
lios has already been given upon page 191.

By exercising a very fine discrimination in distinguishing color we
may count 15 distinct colors which are displayed by the 450 mem-
bers of the Danaoid Heliconidae, as follows: black, brown, translu-
cent black, sulphur-yellow, canary-yellow, citron-yellow, primrose-
yellow, yellow-rufous, reddish rufous, rufous, white, translucent yel-
low, translucent rufous, transparent areas upon the wings, transpar-
ent areas which display iridescence. We see, then, that while the
200 species of Papilio display 36 different colors, the 450 Danaoid
Heliconidae exhibit only 15. In other words, the numbers of the
species and of the colors are almost in inverse ratio in the two
groups; for while the Papilios are only 4 as numerous as the
Danaoid Heliconidae, they display almost 40 times as many colors;
and this is all the more remarkable when we remember that the gen-
eral class of coloration in the Papilios and Danaoid Heliconidae is
apparently the same. "That is to say, in both groups we find all of
the species displaying decidedly conspicuous colors, the coloration
of the upper surfaces of the wings being in both rather more bril-
liant than that of the lower surfaces, but without essential differences
in eolor-pattern. Noris there an attempt in either case at protective
resemblances, such as the imitation of the coloration of bark, leaves,
etc. The color-patterns of the Papilios are, moreover, extremely
complex, and upon comparing the different species, there are seen
to be frequent fusions and obliterations of the characteristic mark-
ings, so that Haase (93), who has made an extensive study of their
color-patterns, is forced to divide them into many small groups
of a few species each. The variation in the form of the wings is
also very great among the Papilios, for while P. protesilaus pos-
sesses upon its hind wings, long tail-like appendages, the hind
wings of P. hahneli are rounded off and without marked appendages.

MAYER: COLOR AND COLOR-PATTERNS. 225

There is, apparently, but one important respect in whieh the
Danaoid Heliconidae are more variable than the Papilios, and that
is size, For example, Lycorea ceres, which is probably the largest
of the Danaoid group, has 2.2 times the spread of wing of Ithomia
nise, which is one of the smallest (see Plate 4, Figs. 46 and 54).
The largest Papilio, P. androgea, on the other hand, spreads only
2.16 times as much as the smallest, P. triopas.

There is another minor respect in which the color-patterns of the
Papilios are different from those of the Heliconidae. In the Heli-
conidae the fore wing slightly overlaps the hind wing, and that por-
tion of the hind wing which is hidden from view is always dull in
color (see Plates 5-8). In the Papilios, however, the fore wing
does not overlap the hind wing to such an extent as in the Helico-
nidae, and it is worthy of note that the costal edges of the hind
wings in the Papilios are as brilliantly colored as are any other por-
tions of the wings.

It is difficult to account for the remarkable conservatism in respect
to color-variations among the Heliconidae, unless we resort to the
explanation afforded by the theory of mimiery; for, while there is
such remarkable simplicity and uniformity of color-pattern through-
out the whole group of the Heliconidae, individual variations
are very common, In the collection at the Museum of Compara-
tive Zoölogy, for example, one finds a regularly graded series of
specimens of Heliconius euerate; at one end of this series the
“inner rufous" area of the hind wing is bright yellow, and at
the other end it is rufous; intermediate specimens are found in
which this area is yellow, but dusted over with rufous scales. Also
the “middle black band” of the hind wings in Melinaea parallelis
is very variable, some specimens showing it broken in the middle
(Plate 7, Fig. 82), and others having it as an entire band. I have
also seen one specimen of H. burneyi in which the commonly
yellow spots upon the under surface of the wings were changed to
white, Another good instance of individual variability is afforded
by H. phyllis (Plate 5, Fig. 69), for in this species the series of
small red spots sometimes found just below the yellow band upon
the hind wing is very variable, and more often absent than present.
Still other instances of individual variability are seen in the yellow
stripes upon the wings of H. charitonius (Plate 5, Fig. 64), which
are often found tinged with rufous. Also the remarkable diversity
in Mechanitis polymnia, and M. isthmia (Plate 7, Figs. 84-87) are

226 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

other examples which show that there is no lack of individual vari-
ability among the Heliconidae. Yet the Danaoid species as a whole
vary but little from the two great types of coloration represented
by Ithomia and Melinaea, and in this respect they are very differ-
ent from the Papilios, where we find a great many color-types and
great diversity in shape of wings. Surely there must be some cause
for the remarkable fact that the Danaoid Heliconidae with their
453 species should display but two types of color-pattern. I ean
think of but one explanation, which is afforded by Fritz Müller’s
theory of mimicry.

In conclusion it gives me pleasure to thank those friends whose
generous aid and kindness have done so much to render the prose-
cution of this research a pleasure to me. I wish to express my
gratitude to Dr. Charles B. Davenport,; who is the real instigator
and promoter of this research ; to Mr. Samuel Henshaw, to whom
I am indebted for numerous kindnesses, and who placed at my dis-
posal the extensive entomological collections and library of the
Museum of. Comparative Zoology at Harvard; to Prof. Edward L.
Mark for his kindness in revising the manuscript of this paper, and
for the numerous valuable suggestions which he has made; to Dr.
Samuel H. Seudder, to whom I am grateful for much kind advice
and for the use of rare works in his library ; to Prof. Ogden N.
Rood for his valuable suggestions in regard to the speetroscopie
apparatus; to Dr. Alpheus Hyatt for his valued and kind advice,
and to my father, Prof, Alfred M. Mayer, for the use of Maxwell’s
dises and the direct-vision spectroscope.

PART C.

GENERAL SUMMARY OF RESULTS BELIEVED TO
BE NEW TO SCIENCE.

(1) The great majority of the colors of Lepidoptera contain a
surprisingly large percentage of black (p. 172).

(2) The colors displayed by the scales are not simple, but com-
pounded of several different colors (p. 173).

(3) The pigments of the scales of Lepidoptera are derived by
various chemical processes from the blood, or haemolymph, of the

MAYER: COLOR AND COLOR-PATTERNS. 221

pupa. The pupal blood of the Saturnidae is a proteid substance
containing egg albumen, globulin, fibrin, xanthophyll, orthophos-
phorie acid, iron, potassium, and sodium (p. 176).

(4) In Callosamia promethea and Danais plexippus the pupal
wings are at first perfectly transparent, then white, then impure
yellow, excepting upon those portions which are destined to remain
white in the mature wing. The mature colors then begin to appear
near the central areas of the wings and between the nervures. Last
of all, the nervures themselves become tinged with the mature
colors. The central portions of the wings acquire their mature
colors before the outer and eostal edges, or the root of the wing
adjacent to the body (p. 178, Plate 3).

(5) The white stage in the development of color in the pupal
wings represents the condition in which the scales are perfectly
formed but lack the pigment which is destined to be introduced
later (p. 178). (See, also, Mayer, '96, p. 230.)

(6) Dull ocher-yellows and drabs are, phylogenetieally speaking,
the oldest pigmental colors in the Lepidoptera. The more brilliant
colors, such as bright yellows, reds, and pigmental greens, are
derived by complex chemical processes and are, phylogenetically
speaking, of recent appearance (p. 178). (See, also, Mayer, ’96, p.
232.)

(7) While the number of species of Papilio in South America
is 9 times as great as in North America, the number of colors which
they display is only twice as great. Hence the greater number of
colors displayed by the tropieal forms may be due simply to the
far greater number of the species, and not to any direet influence
of the climate (p. 191).

(8) The following laws control the color-patterns of butterflies
and moths: (a) Any spot found upon the wing of a butterfly or
moth tends to be bilaterally symmetrical, both a
color; and the axis of symmetry is a line passing through the center

s regards form and

of the interspaee in which the spot is found, parallel to the longi-
tudinal nervures (p. 183). (b) Spots tend to appear not in one
interspace only, but in homologous places in a row of adjacent
interspaees (p. 183). (e) Bands of color are often made by the
fusion of a row of adjacent spots, and, conversely, chains of spots
are often formed by the breaking up of bands (p. 183). (d) When
in process of disappearance, bands of color usually shrink. away at
one end (p. 184). (e) The ends of a series of spots are more

228 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

variable than the middle. This is only a special case of Bateson’s
(94) law (p. 185). (f) The position of spots situated near the
outer edges of the wing is largely controlled by the wing-folds or
creases (p. 185).

(9) The scales in Lepidoptera do not strengthen the wings or
aid the insects in flight. The vast majority of the scales are merely
color-bearing organs, which have been developed under the influence
of Natural Selection. The phylogenetic appearance and develop-
ment of scales upon the originally scaleless ancestors of the Lepi-
doptera did not alter the efficiency of their wings as organs of flight.
It is probable, therefore, that this efficiency was an optimum before
the scales appeared (p. 197).

(10) A systematic study of the Danaoid Heliconidae demon-
strates that their color-patterns can be placed in two types. Type 1,
the more complex, is closely related to the coloration of the Danaidae
from which the Danaoid Heliconidae sprang, and is therefore,
phylogenetically speaking, the older type of coloration. This type
is characteristic of the genera Lycorea, Melinaea, and Mechanitis,
and I have called it the “ Melinaea" type. It is characterized by
the fact that the wings are rufous and black in color, and erossed by
a definite system of yellow bands. Type 2, the “ Ithomia " type, is
characteristic of the genera Ithomia, Ituna, and Thyridia. The
“ Ithomia” type has been derived from the * Melinaea" by the
originally rufous and yellow areas upon the wings having become
transparent (p. 204).

(11) The phylogenetic origin of the * Melinaea " and * Ithomia ”
types of coloration can be accounted for upon the supposition, that
when the species of the Danaoid Heliconidae began to segregate
out from the Danaidae they were for a time rare (p. 215).

(12) A record of the characteristic markings upon the wings of
the Danaoid and Acraeoid Heliconidae shows that, physiologically
speaking, the colors red, rufous, yellow, and white are closely
related, and that black is quite distinct, from these, being the least
variable color of all (p. 220).

(13) In both the Danaoid and Aeraeoid Heliconidae, whatever
color-variation affects that part of the fore wing which is adjacent
to the body of the insect, almost always the same color-variation
affects the homologous area of the And wing in a similar manner
(p. 218, and Fig. 99).

(14) The smaller yellow spots upon the wings of the Heliconi-

bo
bo

MAYER: COLOR AND COLOR-PATTERNS.

dae are more liable to color-variations than are the larger ones.
This is what we should expect, if the theory of mimicry be true;
for large spots are more conspicuous, and therefore their preservation
is more important (p. 220). This rule, however, does not hold for
the black markings of the wing (p. 222).

(15) The mathematical chance against five similar and inde-
pendent color-sports arising in the genus Melinaea is as 2,830,000
to 1. Hence, the five Melinaeas which display the “ inner black ”
as a double spot are probably descended from a single ancestor (p.
219).

(16) The marginal spots of the fore wing in the Danaoid Heli-
conidae show a marked tendeney to appear either as 2 or 3, or else
as 6 or 7 spots (p. 223, Fig. 101). The marginal spots of the hind
wing show a marked tendency to appear either as 4 or 5 spots
(p. 228, and Fig. 102).

(17) The 200 species of Papilio in South America display 36
distinet colors, while the 450 species of Danaoid Heliconidae exhibit
only 15. Hence the numbers of the species and of the colors are
almost in inverse ratio in the two groups. This may be explained
by the fact, that the Danaoid Heliconidae mimie one another, while
the Papilios do not (p. 224).

(18) The colors are dull upon those portions of the hind wing
which are hidden from view by the overlapping fore wing (p. 225).

(19) There is no lack of individual variability among the species
of the Danaoid Helieonidae; yet the species as a whole vary but
little from the two great types of color-pattern represented by
Melinaea and Ithomia. In order to account for this remarkable
fact I am forced to resort to Fritz Müller's theory of mimicry (p.

225).

230 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

TABLE 1.

Showing the Variations in Color of the “ Inner Rufous” (Area
I in Figures on Plate 4) of the fore wing in the Danaoid Heli-
conidae.

Aie; We rog die nin rog eee |
GENERA. | 2 | 2 — e a EMEN
|. | ae | Bee) A (dB. Bm a 8 8
s kane] [tients RER Pr
r | 65 | | |
Ituna | | | ^ ns
Athesis | | | 2 | 1 | |
Thyridia | | | 5 |
Athyrtis ea | | |
Olyras Sree te SUP NT Ai im
TUTO B cial Spey 8 2 | |
e eee i2 1 1 |
Dircenna EMO B 1 3 un 2 | Im
CANEDO «aad Se Fiss Is; | | 2 |
Epithomia . 1 ikea
Ceratinia | 28 ee clear a 2
Sais 7 | |
ee | | | 10 |
Mechanitis | 18 | | | ] 1 4 |
Napeopetes- ine NT 3 29123 5 3
Tr 13 15 | 120 | 26 5 3 |
Aeria | i | 4 | |
Melinaea | 22 | | 2 |
Tithorea | 4 | | 6 |
j | i
Total Wr a e car aa y ar A kei
Excluding Ithomia . | 95 10 8 32 | dnd Tete see Vet oui

Note: The costal edge of the fore wing is usually black; it is rufous or ‘brown, how-
ever, in 47 Ithomias and dull yellow in one; it is rufous in two species of Sais, in one
species of Ceratinia, and in one species of Athesis. Hence it is black in 348 species and
light colored in 52,

MAYER: COLOR AND COLOR-PATTERNS. Zou

TABLE 2.

Showing the Variation (presence or absence) of the “Inner
Black” (Area II) of the fore wing in the Danaoid Heliconidae.

if
GENERA. DA | 2 Remarks.
o | —
* =
Lycorea 5 |
Ituna 8
Athesis 8
Thyridia 5 |
Athy rtis 2
Olyras 4
Eutresis 2
Aprotopos 2 |
Dircenna 8 t
Callithomia à 3
e eee 2
Gehe iue 20 12
JV
l C oed Rel DE
WESC d oo Mn
Napéogened . . . 13 17 | Appears as 2 spots in 1 species.
Ithomia KT E 72 140 | Appears as 2 spots in“ species.
eee eed | 4
Melinaea . . . 21 3 | Appears as 2 spots in 5 species.
MIO Wea jJ Mb ots 10
TOWN E VL mU uU

Excluding Ithomia — . |198 | 50 |

202 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Showing the

III) of the fore

GENERA.

Lycorea
Ituna
Athesis
Thyridia
Athyrtis
Olyras
Eutresis .
Aprotopos
Dircenna
Callithomia
Epithomia
Ceratinia
Sais

Scada
Mechanitis
Napeogenes
Ithomia
Aeria
Melinaea
'l'ithorea

Total

rufous.

Rufous.
Translucent

Excluding Ithomia

| Translucent,
| slightly

TABLE

ru-

fous.

bo

we

Transparent.

~

low.

>

Translucent,
lightly yel-

8

ar

-J

bo

e Danaoid Heliconidae.

'Translucent

in Color of the “Inne

yellow.

No

Yellow.

: Yellow”

ex

Black.

(Area

=
—

E
do
^

2 5
x

80
B

White,

MAYER: COLOR AND COLOR-PATTERNS.

TABLE 4.

Showing the presence or absence of the “Middle Black ” Mark

(Area IV) of the fore wing in the Danaoid IHeliconidae.

GENERA. Present. i Absent,
Lycorea 5
Ituna 3
Athesis 3
Thyridia 6
Athyrtis 2
Olyras 2
Eutresis . 2
Aprotopos 2
Dircenna 11 1
Callithomia 3
Epithomia 2 l
Ceratinia 34 | 7
Sais... 5 |
Scada . g T |
Mechanitis 24
Napeogenes 25 5
Ithomia 194 6
Aeria 4
Melinaea 23 1
Tithorea 10
|
De!!! m 366 | 20
Excluding Ithomia . | 172 14

Present, but changed to
some color other than black.

12 rufous or brown.

12

234

GENERA.

Lycorea
Ituna
Athesis
Thyridia
Athyrtis
Olyras
Eutresis
Aprotopos
Dircenna
Callithomia
Epithomia
Ceratinia
Sais

Scada
Mechanitis
Napeogenes
Ithomia
Aeria
Melinaea
Tithorea

Total

Excluding Ithomia

Rufous.

* OR

TABLE 5.

Showing the Variation in Color of the
(Area V) of the fore wing

in the Danaoid

$ did
93 95, 3 |
BE |3»% S |
AQ 88 A |
2238489 2 |
ge 8 | s |
h 222
BÉ Thn |
BE
| d
9 |
2| 4
89
| 1
p E
E
2| |
|
id 6 |
| |
|
| o| JP
| | € |
| 6 12 | 128
| |
a
| 3

BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

* Middle Yellow ?

Heliconidae.

82 3.
vn Qe |
© P OF |
ZE. E 2o te
T E
Heo We $5
3 gU A
Fin) S^ —
Eon | H D
Bu | & 8
| 5
|
|
1
1. 78
l
2 2 3
| 9
o
1 1
4 5 24
5
6
1 ue
4 4 6
24 8 14
4
2 15
7
35 31 | 108
2

Black.

Band

White,
more or less
translucent.

MAYER: COLOR AND COLOR-PATTERNS. 235

TABLE 6.

Showing approximately the Number of Species in which the
„Outer Yellow” (Area VI) of the fore wing in the Danaoid Heli-
conidae appears as a separated Marking. It is usually fused with
the * Middle Yellow” Area.

Wholly fused Partially fused

GENERA. with middle with middle | Separate. | Absent.
yellow, yellow. | |

m Zei |
Dehn! o] 1l 4 |
e, e 1 2 1
AONE V o cs 3
DN. . 535553 b | |
BOUES . 4 s | 2 |
Olyras pta NER 1 3
Ew . . : » 2
Apropos 2
Dircenna 5 UT 5?
Callithomia . . . . 1 2
BIO $7 l 1
Ceratinia e . . . 22 16 about 3 about
Sais V | 2 8
Scada cia a t 4 3
Mechanitis . . a l | 20 3
Napeogenes . . . . 6 j 24 about
CC 200 about
BOVE Me paired oo Roue: l |
Mang vow | | | 17 6
Tithorea | | 10
Total . . . . . . | about250 about 30 !"about 50 | perhaps 20

236 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

TABLE 7.

Showing the Degree of Development of the * Outer Black” (Area
VII) of the fore wing in the Danaoid ITeliconidae.

Well devel- reduced
(GENERA oped over a to the. outer Present, but changed to
: large area of margin of the another color,
the fore wing. fore wing.

Eyaoren ar He 5
Ituna 3
Athesis 2 1
Thyridia 5
Athyrtis 2
Olyras 3 |
Kutresis | 2
Aprotopos | 1 1
Dircenna | 5 2
Callithomia | 5
Epithomia 2
Ceratinia | 28 13 2 partly rufous.
Sais | 2 3 partly rufous.
Scada | 3 |
Mechanitis 22 2
Napeogenes . . «v | 20 4
Ithomia | 161 54 16 rufous or brown.
MOM 6 e ox ov 4
Melnaees 24 partly rufous.
Tithorea 10

TVT 305 05 22 partly rufous.

MAYER: COLOR AND COLOR-PATTERNS.

TABLE 8.

Showing the presence or absence of the “ Longitudinal Black
Stripe" (Area VIII) whieh runs parallel with the lower Edge of

the fore wing in the Danaoid Heliconidae.

|
| Present and

well developed
as a stripe.

’ | Much
GENERA, | reduced.
| iim

t

Lycorea 3 5

AUD ee ders | 3

Athesis V 3

Thyridia | 5

Athyrtis pue di cd 2

QUEEN A dun | 3

Eutresis 2

Aprotopos l

Dircenna ji 3
DOMIR VV deua

DIMOU ty =e Um 1 | l
Gars 37 | 2
Sais . 3 1
GAG b yas s 6
Mechanftiss LT

Napeogenes . . . . 20 6
Ii, s 200 | 6
A 4 |
Melinaeee a 14 | 5
BRDOTBRE IT Her 5 |

MODA Aa deu / eb 338 | 24

Absent.

Whole area
suffused with
black.

288 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

TABLE 9.

Showing the Manner of Occurrence of the Marginal Spots (Area
IX) of the fore wing in the Danaoid Heliconidae.

it,, 8 laa i | |
|

„ 9
e — | spot. spots. spots. spots. spots. spots. spots. spots. | spots.

| i | | | | |

| | | | |
Lycorea 5 | | |
Ituna 3 | | | | 4 |
Athesis 3 | | | |
Thyridia 4 | „ | | |
Athyrtis E | | | E |
Olyras 3 | 5
Eutresis 2 | | | |
Aprotopos 1 1 | | | | | |
Dircenna . 11 1 | | | | | |
Callithomia 1 | 2 | | | |
Epithomia 2 j | | |
Ceratinia 19 1 507 i inpti qo | Tm
Sais 5| | | | | | |
Scada 4 | | | | 1 iN |
Mechanitis dd T. Io 1 3 | |
Napeogenes 14 pend | Bel E 5 3 8
khona . 137 3 id 14 A-1919 n 8
Aeria 4 | | |
Melinaea 13 Ls | a pap gi |
Tithorea 5 now Ed | 2 | hatti

| |
SEE a = — —

Total |

MAYER:

COLOR

TABLE

AND COLOR-PATTERNS.

10.

239

Showing the Color-V ariations affecting the“ Inner Rufous ” (Area
X) of the hind wing in the Danaoid Heliconidae.

GENERA

Lycorea
Ituna
Athesis .
Thyridia
Athyrtis
Olyras
Kutresis
Aprotopos .
Dircenna
Callithomia
Epithomia .
Ceratinia
Sais

Scada
Mechanitis
Napeogenes
Ithomia
Aeria
Melinaea
Tithorea

Total

Rufous.

a

Translucent
rufous

111
. M Qe [og
= „„
Bee; S | Spe) 38
225 e SSS vm |
Be LL BU. BT EE |
e of 8 e È do
M H | — | M |
8 5 EE |8 |
^ I
| |
2
2 13 |
4 | 1 |
| |
| |
| Lo
|
| |
1 | |
8 g y |
| |
|
|
3 6 6 1 |
| |
| 55
| | í |
| |
„
GSF OP a |
14 1 89% Tt 5 |
| |
| ji 4
| |
|
|
i — —
| 155 | 99 | 99. |
|

Yellow.

Black.

240 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

TABLE. 11.

Showing the Variations of the * Middle Black Stripe" (Area XI)
of the hind wing in the Danaoid Heliconidae,

| Fused with

|
GENERA. | Present. | Absent. | the marginal | Changed color.
Wack,
| Es |
| | |

Lycorea | 5 | | |
Ituna | 2 1 | |
Athesis | 2 l | |
Thyridia i | 2 2 | 1 partially
E en 2 | |

| | \ changed
Olyras | 1 2 | | to translu-

| | | ! cent yellow ?

Eutresis 2 | |
e ee, VERE 2 | | |
Dircenna | 7 | 1 |
Callithomia tae | 8 |
F 2 | |
Ceratinia 21 20 |
Sais | 1 | 4
Scada 7 |
Mechanitis . . . 22 2 | |
NAWEOBENeEB 4 03 p | 15 | 15 |
Ithomia 45 1 | 168 |
Aeria 4 | |
Melinaea | 15 9 | |
Tithorea | 8 2 | |
Total | : |

MAYER: COLOR AND COLOR-PATTERNS.

Showing the Color-Variations of the “ Outer Rufous”

TABLE

19.

of the hind wing in the Danaoid Heliconidae.

GENERA,

Lycorea
Ituna
Athesis
Thyridia
Athyrtis
Olyras
Eutresis
Aprotopos
Dircenna
Callithomia
Epithomia
Jeratinia
Sais
Scada
Mechanitis
Napeogenes
Ithomia
Aeria
Melinaea
Tithorea

Total

Excluding Ithomia

Rufous.

Translucent

rufous.

—1

Translucent,

Transparent.

ae be

Translucent,

ghtly

>

but sli

yellow.

Translucent
yellow.

(Area

Yellow.
Black.

24]

XII)

White,

242

Showing the presence or absence, and Color-Changes of the

TABLE 13.

BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

“Outer Black” (Area XIII) of the hind wing in the Danaoid

Heliconidae.

GENERA.

Lycorea
Ituna
Athesis
Thyridia
Athyrtis
Olyras
Eutresis
-Aprotopos
Dircenna
Callithomia
Epithomia
Ceratinia
Sais

Scada
Mechanitis
Napeogenes
Ithomia
Aeria
Melinaea
Tithorea

Total

1
|

Present,

Absent.

Changed color.

11 changed to rufous or brown.

MAYER: COLOR AND COLOR-PATTERNS.

TABLE 14.

243

Showing the Number of the Marginal Spots of the hind wing in

the Danaoid Heliconidae.

With- 1 2 8 4
GENERA, out

spots. spot. | spots.| spots, spots.
Lyoore8 . .-.. 1
Kundin: zur n 2 1 |
Athösis >, I i
Thyridia C 4 1 |
Vim wo re al
Olyras . uem 3 1
Eutresis REN 1
Aprotopos . . | 2
Dircenna . . . 10 1
Callithomia . . 2 1
Epithomia . . 1
Ceratinia . . . 18 3 l 2 5
Salts oe ws C 4 1
BOAN aai a 9 : 1
Mechanitis . . 18 1 Jl 2
Napeogenes . . 21 1 1 i
Ithomia . . i 164 8 84.12: |
ASTA 0%; 4 |
Melinaea . . . 19 1
Dehne 4 2 |
Totalia qul. 5. 219 6 14 13 0:99

5
spots.

6

spots.

(i

spots.

8

spots.

9
spots.

i
2 Reddish
Rufous. | rufous,

8

8

BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

TABLE 15.

| Yellow. | Ocher, White.
|

11 1 2

Note: The costal edge of the fore wing is always black.
TABLE 16.

fore wing in Heliconius and Eueides.

Showing the Color-Variations of the “Inner Rufous” Area of the
fore wing in Heliconius and Eueides.

Irides-
Black. cent
blue.
29 26
|
3
|
32 Igor

Showing the Variations affecting the “Inner Black” Area of the

Rufous,

Pfous.

Black. Iridescent blue.
85 26
13
98 26

N
|

TABLE 17.

Yellow. | Ocher. White,

ea
E

20

an

5

Note: In 54 species of Heliconius the inner rufous is entirely suffused with black.

Showing the Color-Variations of the “Inner Yellow” Area of the
Jore wing in Heliconius and Eueides.

Irides-

Black. cent
blue.
12 3

MAYER: COLOR AND COLOR-PATTERNS. 245

TABLE 18.

The “Middle Black” Area in the fore wing in Heliconius is
present as a Black or Blue Marking in 99 Species and absent in 12.
It is present as a Black Mark in all 18 Species of Eueides,

TABLE 19.

Showing the Color-V ariation of the * Middle Yellow " Area of the
fore wing in Heliconius and Eueides.

| | |
: Reddisl
| Rufous, MURS 3 Yellow, | Ocher. | White. Black.
i5 | |
Heliconius . 11 e 90 | 28
| | |
Eueides al E | ete de | 1
„„ | | pa :
| |
TOU Stet 5 | 16 | 12 | 65 | 12 | 24 |

TABLE 20.

Showing the Color-V ariations of the “Outer Yellow ” Area of the
Jore wing in Heliconius and Eueides.

| TEM | | Irides-

| „dis! rides

| Rufous, pe qe Yellow, | Ocher. | White. Black. | cent

| . |

| | | | | blue.

F E
Helieonius . . | 2 | 1 | 47 24 33 | 4
Eueides | | 6 1 10
Tdé rpi ae 2 | i$] 4 | 6 25 44 4
|

TABLE 21.

The “ Outer Black ” Area of the fore wing in all the 111 species
of Heliconius known to me is Black or Iridescent Blue.
It is Black in all 18 Eueides.

246

TABLE

Showing the Manner of Occurrence of the Marginal Spo

fore wing in Heliconius and Eueides.

With-| 1

55 Vou
out x dee
Spots. spot. idi spots.
Heliconius 89 b IERS
Eueides 14 8.
I — — — — —À
Total . | 103 cis peti
|
TABLE

BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

22

4 | 5 6
ANE spots.| spots.
3
pl s 4
1
o k a 5
23.

ts of the

m 8 A
|
Spots.' spots. spots,
|
2 | qe
|
|
| |
2 | ng
|

Showing the Variations affecting the * Longitudinal Black Stripe”

of the fore wing in Heliconius and E

Whole area suf-
fused with black.

Heliconius 75

Eueides 2 |

"otal eh
TABLE

Showing the Color- Variations of the * Inner Rufous”

hind wing in Heliconius and Eueides.

rec: "

| 81385 35 8

eae T

| A M" pe S |
ae ehe a PU m dads :
Heliconius : | 42 i 3
Eueides |
Doll! |

ueides.

Well developed as
a black stripe

24.

LT
og | Sg
$5
$5 8 | oe
ut
T OD E | ne
E W | 4
1 16 | 26

Absent (area suf-
fused with rufous).

Area of the

= = ug

a. | ae 38 5
ag | a4 | age
A9 "o 4 O
Qi — ase
$9 | as |aoe
mo mr zen
10 1 >

MAYER: COLOR AND COLOR-PATTERNS. 241
TABLE 25.

Showing the Variations of the «Middle Black Stripe“ of the hind

wing in Heliconius and Eueides.

Absent

| Welldevel- |. MU M Absent

oped as a more Absent (suf- (place taken (place taken

| or less distinct) fused With | by red or ru- by ocher

| stripe. black), ous). color).
Heliconius 47 50 5
LU T cun: ]- Il ec mater cpu cs 6 10 | |
COUN S EUN | 53 | 59 15 | 1

| | | |

TABLE 26.

Showing the Color- V ariations of the “Outer Rufous ” Area of the
hind wing in Heliconius and Eueides.

| Reddisl Irides-
Rufous. | fon Yellow. | White, | Ocher. Black. cent
| | blue.
wis uu prd — ä
Heliconius | so 3 19 8 | 40 6
i 5 |
Bueides 17 | | 1
TOM v SERES 4T 4 | 19 8 50 6
TABLE 27.

The * Outer Black ” Area of the hind wing is Black in 106 species
of Heliconius, White in 4, and Yellow in 1. It is Black in all the 18

species of. Eueides known to me.

248

TABLE

28.

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Showing the Number of Species in each Genus of the Heliconidae
examined, and also the Number known according to the Enumera-

tion of Staudinger (84-88).

|
I
JENERA. |

Lycorea
Ituna .
Athesis
Thyridia .
Athyrtis .
Olyras
Eutresis
Aprotopos
Dircenna
Callithomia .
Epithomia
Ceratinia
Sais

Scada .
Mechanitis
Napeogenes
Ithomia
Aeria .
Melinaea
Tithorea .

Total of Danaoid Heliconidae.

Heliconius
Eueides

Total

amined by me.

Number of species ex-

4 species

and,l var.

| Number of species
| known to Staudinger
(84— 88).

4 species and 1 var.
4

8 4

3 | 4

5 | 4

2 | 2

4 | 5

2 | 2

2 | 4
12 204-

3 8

2 2
41 504-

5 4

7 9

10 species, 14 var. | 10 species, 13 var.

30 | 30
212 | 230+

4 | 4

24 | 25

10 | 18

|
400 453+
5 | —
111 | 130
18 24

MAYER: COLOR AND COLOR-PATTERNS. 249

BIBLIOGRAPHY.

Agassiz, A.

'59. [Mechanism of the Flight of Lepidoptera.] Proc. Bost. Soc. Nat.

Hist., Vol. 0, p. 426-427.
Bates, H. W.

62. Contributions to an Insect Fauna of the Amazon Valley. Lepidoptera :

Heliconidae. Trans. Linn, Soc. London, Vol. 28, p. 495-566, pl. 55-56.
Bates, H. W.

63. On a Gollection of Butterflies brought by Messrs. Salvin and. Godman
from Panama, with Remarks on geographieal Distribution. Proc. Zool.
Soc. London, p. 239-249, pl. 29.

Bates, H. W.

'04—65.. New Species of Butterflies from Guatemala and Panama, collected
by Osbert Salvin and F. Du Cane Godman, Esqs. Ent. Mo. Mag., Vol. 1,
p. 1-6, 31-35, 55-59, 81-85, 113-116, 126-131, 161-164, 178-180, 202-205.
[New Heliconidae, p. 31-35, 55-59.]

Bateson, W.

94. Materials for the Study of Variation treated With especial Regard to
Discontinuity in the Origin of Species. London and New York, 16 4- 598
pp. [p. 288-302.]

Beddard, F. E.
'02. Animal Coloration. London and New York, 8-++ 288 pp., 4 pls.
Belt, T.

74. The Naturalist in Nicaragua. London, 403 pp. Second edition, London,

1888, 33 + 403 pp.
Bemmelen, J. F. van.

89. Ueber die Entwicklung der Farben und. Adern auf den Schmetterlings-
flügeln. Tijdschrift der Nederlandische Dierkundige Vereeniging, Ser. 2,
Deel 2, p. 235-247.

Mackiston, T., and Alexander, T.

84. Protection by Mimicry — A Problem in Mathematical Zoology. Nature,

Vol. 29, p. 405-400.
Burgess, E.

80. Contributions to the Anatomy of the Milk-weed Butterfly Danais archip-

pus (Fabr.). Anniversary Mem. Bost. Soc. Nat. Hist., 1880, 16 pp., 2 pls.
Burmeister, H.

"I8. Lépidoptéres. Description Physique de la République Argentine, Tome

5, 624 pp., 24 pls. [p.21-28.]

250 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Butler, A. G.

65. Description of six new Species of Diurnal Lepidoptera in the British

Museum Collection. Proc. Zool. Soc. London, 1865, p. 430-434, pl. 25.
Butler, A. G.

69. Remarks upon certain Caterpillars, ete., which are unpalatable to their
enemies. Trans. Ent. Soc. London, 1869, p. 27-29. [Feeding habits of
birds; warning coloration.]

Butler, A. G.
69a. Descriptions of several new Species of Nymphalidian Rhopalocera.
Ann. Mag. Nat. Hist., Ser. 4, Vol. 3, p. 17-21, pl. 9. [New Heliconidae.]
Butler, A. G.
6974. Lepidoptera Exotica. London, 190 pp., 64 pls. [New Heliconidae. ]
Butler, A. G.

77. List of Lepidoptera recently collected by Mr. Walter Davis in Peru,
with Descriptions of a new Genus and several new Species. Ann. Mag.
Nat. Hist., Ser. 4, Vol. 20, p. 117-129.

Chapman, T. A.
88. On Melanism in Lepidoptera. Enu. Mo. Mag., Vol. 25, p. 40.
Coste, F. H. P.

90-91. Contributions to the Chemistry of Insect Colours. Eutomologist,
Vol. 23, p. 128-132, 155-159, 181-187, 217-223, 247-252, 283-287, 309-314,
338-343, 870-874; Vol. 24, p. 9-15, 37-40, 53-60, 86-91, 114-119, 132--139,
163-170, 186-192, 206-211. See also Nature, Vol. 45, p. 513-517, 541-542, 605.

Cramer, P.

1779-82. Uitlandische Kapellen. Amsterdam, 4 Vols., 400 pls.
Darwin, C.

"I1. The Descent of Man, and Selection in Relation to Sex. London, 2 Vols.
Dimmock, G.

'88. The Scales of Coleoptera. Psyche, Vol. 4, p. 3-11, 23-27, 43-47, 63-71.
Distant, W. L.

'82-86. Rhopalocera Malayana. London, 16 ＋ 481 pp., 44 pls. [p. 33; 129;
289 469.]

Dixey, F. A.

'00. On the phylogenetic Significance of the Wing-markings in certain
Genera of the Nymphalidae. ‘Trans. Ent. Soc. London, 1890, p. 89-129,
pl. 1-3.

Druce, H.

76, List of the Butterflies of Peru, with Descriptions of new Species. Proc.

Zool. Soc. London, 1876, p. 205—250, pl. 17-18.
Eimer, G. H. T.

^80. Die Artbildung und Verwandtschaft bei den Schmetterlingen. Jena;

12 + 243 pp., 4 Taf.
Felder, C. und R.

3467. Rhopalocera. Reise, Österreichischen Fregatte Novara. Wien,

6 + 548 pp., 140 Taf. [New Heliconidae. |
Godman, F. D., and Salvin, O.

79-86. Lepidoptera-Rhopalocera. Vol. 1. Biologia Centrali-Americana.

Insecta. [London], 487 pp., 47 pls. [New Heliconidae.]

MAYER: COLOR AND COLOR-PATTERNS. 251

Godman, F. D., and Salvin, O.

80. A List of Diurnal Lepidoptera collected in the Sierra Nevada of Santa
Marta, Colombia, and the Vicinity. Trans. Ent. Soc. London, 1880, p. 119-
132, pl. 3-4. [New Heliconidae. }

Griffiths, A. B.

'02. Recherches sur les Couleurs de quelques Insectes. Comptes Rendus

Acad. Sci. Paris, Tome 115, p. 958-959.
Haase, E.

93. Untersuchungen über die Mimicry auf Grundlage eines natürlichen
Systems der Papilioniden. Bibliotheca Zoologica, Heft 8, Theil 1, 120 pp.,
6 Taf. [p. 14-15.]

Haase, E.

93a. Untersuchungen über die Mimicry auf Grundlage eines natürlichen
Systems der Papilioniden. Bibliotheca Zoologica, Heft 8, Theil 2, 161 pp.,
8 Taf. English translation by C. M. Child. Stuttgart, 1896, 154 pp., 8 pls.

Hewitson, W. C.

54. Descriptions of some new Species of Butterflies from South America.
Trans. Ent. Soc. London, New Ser., Vol. 2, p. 245-248, pl. 22-23. [New
species of Heliconidae.]

Hewitson, W. C.

'56—76. Illustrations of new Species of Exotic Butterflies. London, 5 Vols.,

Plates.
Higgins, H. H.

68. On the Colour-Patternsin Butterflies. Quart. Journ. Sci., Vol. 5, p. 92:

829, 1 pl.
Hopkins, F. G.
89. Uric Acid Derivatives functioning as Pigments in Butterflies. Proc.
Chem. Soc. London, 1889, p. 117. Also Nature, Vol. 40, p. 335.
Hopkins, F. G.
'01. Pigment in Yellow Butterflies. Nature, Vol. 45, p. 197.
Hopkins, F. G.

'04. The Pigments of the Pieridae. A Contribution to the Study of Exere-
tory Substances which function in Ornament. Proc. Roy. Soc. London,
Vol. 57, p. 5-6.

Hopkins, F. G.

96. The Pigments of the Pieridae: A Contribution to the Study of Excre-
tory Substances which funetion in Ornament. Philos. Trans. Roy. Soc.
London, Vol. 186, p. 0661-682.

Hübner, J.

06-25. Sammlung exotischer Schmetterlinge. Augsburg, 5 Bde., Tafeln.
Humboldt, A. von, et Bonpland, A.

33. Recueil d'Observations de Zoologie, Latreille. Paris.
Keeler, C. À.

’93. Evolution of the Colors of North American Land Birds. Occas. Papers
Cal. Acad. Sci., Vol. 3, 12 ＋ 301 pp., 19 pls.

Kellogg, V. L.

04. The Taxonomie Value of the Scales of the Lepidoptera, Kansas Univ.

Quart., Vol, 3, p. 45-80, pl. 9-10.

252 BULLETIN: MUSEUM OF COMPARATIVE ZOÓLOGY.

Kirby, W. F.

7177. A synonymic Catalogue of Diurnal Lepidoptera. London, 5-- 690

pp. Supplement. London, p. 691—885.
Leydig, F.

’55. Zum feineren Bau der Arthropoden. Müller’s Archiv, Jahrg. 1855

p. 376-380, Taf. 15-18.
Mayer, A. G.

'06. The Development of the Wing Scales and their Pigment in Butterflies

and Moths. Bull. Mus. Comp. Zoöl. Harv. Coll., Vol. 29, p. 209-236, 7 pls.
Ménétriés, E.

63. Descriptions des nouvelles Espèces de Lépidoptéres. St. Petersbourg.

[New Heliconidae.)
Merrifield, F.
'04. Temperature Experiments in 1893 on several Species of Vanessa and
other Lepidoptera. Trans. Ent. Soc. London, 1894, p. 425-438, pl. 9.
Moore, F.
90-96. Lepidoptera Indica. London, 2 Vols., 190 pls.
Müller, F.

"I8. Notes on Brazilian Entomology. Trans. Ent. Soc. London, 1878, p. 211-

223. [Odors emitted by Butterflies and Moths.]
Müller, F.

79. Ituna und Thyridia. Ein merkwürdiges Beispiel von Mimicry bei
Schmetterlingen. Kosmos, Bd. 5, p. 100-108, fig. 1-4. English translation
by Raphael Meldola. Trans. Ent. Soc. London, 1879, p. 20-29 Proc., fig. 1-4.

Müller, W.

86. Südamerikanische Nymphalidenraupen. Versuch eines natürlichen
Systems der Nymphaliden. Zool. Jahrbücher, Bd. 1, p. 417-678, Taf.
12-15.

Poulton, E. B.

^85. The essential Nature of the Colouring of Phytophagous Larvae (and
their Pupae); with an Account of some Experiments upon the Relation
between the Colour of such Larvae and their Food-plant. Proc. Roy. Soc.
London, Vol. 38, p. 269-315.

Poulton, E. B.

87. The experimui.! Proof of the Protective Value of Colour and Markings
in Insects in Reference to their Vertebrate Enemies. Proc. Zool. Soc.
London, 1887, p. 191-274.

Poulton, E. B.
'00. The Colours of Animals. International Sei. Series, 67. New York,
3 + 360 pp.
Poulton, E. B.

'03. The experimental Proof that the Colours of certain Lepidopterous
Larvae are largely due to modified Plant Pigments derived from Food,
Proc. Roy. Soc. London, 1893, p. 417—430, pl. 3-4.

Ridgway, R.

'86. A Nomenclature of Colors for Naturalists. Boston, 129 pp., 17 pls.
Rippon, R. H. F.

'89-93. Icones Ornithopterorum : A Monograph of the Rhopalocerous Genus
Ornithoptera, or Bird-wing Butterflies. London, parts 1-5, Plates.

MAYER: COLOR AND COLOR-PATTERNS. 253

Schiffer, C.

'89. Beiträge zur Histologie der Insekten. Zool. Jahrbücher, Morph. Abth.,

Bd. 3, p. 611-652, Taf. 20-30. | Reference, p. 647-652.]
Schatz, B., und Röber, J.

'85-02. Die Familien und Gattungen der Tagfalter. (Exotische Schmetter-

linge von Staudinger und Schatz.) Fürth, 2. Theil, 2 ＋ 284 pp., 50 Taf.
Scudder, S. H.

'88-—89. The Butterflies of the Eastern United States and Canada with special

Reference to New England. Cambridge, 3 Vols., 24 + 1958 pp., 89 pls.
Seitz, A.
89. Lepidopterologische Studien im Ausland. Zool. Jahrbücher, Syst. Abth.,
Bd. 4, 771-779, 905-924.
Semper, G.
'86—92. Die Schmetterlinge der Philippinischen Inseln. Wiesbaden, 2 Bde.
Snellen, P. C. T., en Leeuwen, J. van, Jr.

'87. Bijdrage tot de Kennis der Lepidoptera van het Eiland Curaçao.

Tijdsch. voor Ent., Deel 30, p. 9-66, pl. 1-5.
South, R.

80. Notes on some Aberrations in the Genus Vanessa. Entomologist,

Vol. 92, p. 217-221, pl. 8. [Fig. 7.]
Srnka, A.

84. Eine neue Athyrtis. Lepidoptera: Fam. Heliconidae. Berlin. Ent.

Zeitschr., Bd. 28, 'p. 163-165.
Srnka, A.

'W5. Neue Siidamerikanjsche Danaidae und Heliconidae. Berlin. Ent.

Zeitschr., Bd. 29, p. 121-130, Taf. 1.
Staudinger, O.

82. On three new and interesting Species of Rhopalocera. Proc. Zool. Soc.

London, 1882, p. 396-398, pl. 24. [New Heliconidae.]
Staudinger, O.

84 88. Exotische Tagfalter. (Exotische Schmetterlinge von Staudinger

und Schatz.) Fürth, 1. Theil, 333 pp. 100 Taf. [New Heliconidae.]
Urech, F.

'01. Beobachtungen über die verschiedenen Sehuppenfarben und die zeit-
liche Succession ihres Auftretens (Farbenfelderung) auf den Puppen-
flügelehen von Vanessa urticae und io. Zool. Anzeiger, Jahrg. 14, p. 466—
473.

Urech, F.

92, Über Eigenschaften der Schuppenpigmente einiger Lepidopteren-

Species. Zool. Anzeiger, Jahrg. 15, p. 299-306.
rech, F.

'03. Beiträge zur Kenntniss der Farbe von Insektenschuppen. Zeitschr. f.
wiss, Zool., Bd. 57, p. 300-384.

Van Bemmelen. See Bemmelen, J. F. van.
Wallace, A. R.

“ur, eo Or Warning Coloration.] Trans. Ent. Soc. London, Ser. 3,

Vol. 5, p. 80-81, Proc.
Wallace, A. R.
80, Darwinism. London and New York, 16 ＋ 494 pp. (p. 233-267.

254 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Walsingham.
'85. On some probable Causes of a Tendency to melanie Variation in
Lepidoptera of high Latitudes. Trans. Yorkshire Union, pt. 5, p. 113-140.
Cf. Nature, Vol. 31, p. 505.
Watkins, W.
91. Ornithoptera trojana, Staudinger. Entomologist, Vol. 24, p. 178-179,
pl. 4.
Weismann, A.
'82. Studies in the Theory of Descent. ‘Translated and edited by Raphael
Meldola. London, 2 Vols., 28 + 729 pp., 8 pls. [p. 1-158.]
Weymer, G.
75. Exotische Lepidopteren. Stettiner Ent. Zeit., Jahrg. 36, p. 368-385,
Taf. 1-2.
Weymer, G.
84. Exotische Lepidopteren. II. Stettiner Ent. Zeit., Jahrg. 45, p. 7-28,
Taf. 1-2.

MAYER: COLOR AND COLOR-PATTRERNS. 255

TABLE OF CONTENTS.

PART A.
GENERAL PHENOMENA or Coron IN LEPIDOPTERA.

I. CLASSIFICATION or COLORS.

PAGI
(1) Pigmental Colors : i ; ; ; à h ; : t 169
(2) Structural Colors j ‘ à i ; : : 1 : « 170
(3) Combination Colors i $ i ; 1 : : F : 171
(4) Quantitative Determination of Pigmental Colors : ; i 172
(5) Spectrum Analysis of Colors of Lepidoptera y ‘ n 173
(6) Summary of Results |. i 1 : ; ‘ : ‘ , 174

II. Tun ESSENTIAL NATURE or PIGMENTAL Conor IN LEPIDOPTERA.

(1) Pigments of Larvae. i ; x i ; 3 í i ; 174
(2) Pigments of Imagines ‘ . . : : s ; f ` 175

III. DEVELOPMENT OF THE VARIOUS COLORS rv THE Puran Wryas.

(1) Historical Account of previous Researches : : x : 4 176
(2) Development of Color in the Pupal Wings of Callosamia promethea 178
(3) Development of Color in the Pupal Wings of Danais plexippus
(archippus) . : : ; 5 x 4 3 5 ; i 180
IV. Tun Laws WHICH Govern THE COLOR-PATTERNS Or BUTTERFLIES AND
Morus.
(1) Historical Account of previous Researches ‘ ‘ : ; : 181
(2) Laws of Color Patterns : í N h à ; : à 183
(3) Detailed Diseussion of the Laws of Color-Patterns . à f ; 185
(4) Origin of Color-Variations x i : : : : : i 189
Bibliography of Color-Aberrations : : i í : 190
(5) Climate and Melanism : : : j ; ; : : : 190
(6) Relation between Climate and Colors of Papilios . à à 1 191

V, Tre Cacses WHICH HAVE LED TO THE DEVELOPMENT AND PRESERVATION
OF THE SCALES OF THE LEPIDOPTERA.

1) Experiments and Theory . ^ ! í : i k ; : 192
Summary of Conelusions . 1 1 i ; i : i 197

1
—

PART B.
COLOR-VARLATIONS IN THE. HIELICONIDAE,
I. GENERAL Causes WHICH DETERMINE COLORATION IN THE HELICONIDAE.,
IIl. METHODS PURSUED IN STUDYING THE Coron-PATTERNS OF THE
HELICONIDA I.

(1) The Two Types of Coloration in the Danaoid Heliconidae 204
(2) Detailed Description of the Melinaea Type of Coloration 205

256 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

III. GENERAL Discussion OF THE CoLoR-PATTERNS AND OF MIMICRY IN
THE GENERA HELICONIUS AND EUEIDES.

(1) The Four Color-Types in the Genus Heliconius . x , 208
(2) Mimiery between the genus Heliconius and the Danaoid

Group f , i ; : ; i ý j i 209
(3) The Three Color-Types in the Genus Eueides — . : i 210
(4) Detailed Discussion of Plates 5-8 í : ; à ; 210

IV. Generar Discussion or THE CorLon-PATTERNS AND OF MIMICRY
AMONG THE DaANAOID HELICONIDAE.

(1) The Origin of the two Types of Coloration : x 3 214
(2) Mimicry among the Danaoid Heliconidae . i à i 216

V. QUANTITATIVE DETERMINATION OF THE VARIATIONS OF THE CHAR-
ACTERISTIC WiNG-MARKINGS IN THE ACRAEOID AND DANAOID
HELICONIDAE.

(1) Variations of the ** Inner Rufous" Areas of the Fore and

Hind Wings à , ; : j $ : P ; 217
2) The “ Inner Black” Spo , : : : i ‘ : 218
(3) Variations of the “Inner Yellow " and “Middle Yellow "

Areas + i ; ; ^ y $ : : i 219
(4) Variationsof the “ Middle Black ” Mark of the Fore Wing 221
(5) Variations of the * Outer Yellow " Area of the Fore Wing 22]
(6) The Relative Permanency of the Black Areas upon the Fore

and Hind Wings : : ; ; à 3 t s 222
(7) The “Middle Black Stripe" of the Hind Wing ; ; 222
(8) Variations of the Marginal Spots of the Fore Wing i 223
(9) 'The Marginal Spots of the Hind Wing i i i i 223

VI. Comparison OF THE COLOR-VARIATIONS OF THE PAPILIOS OF
SOUTH AMERICA WITH THOSE OF THE HELICONIDAE.

PART C.

GENERAL SUMMARY OF RESULTS BELIEVED TO BE NEW TO SCIENCE,

Tables : ; : à : s , : : : x ; i 230
Bibliography 249

Explanation of Plates

MAYER, — Color and Color-Patterns,

PLATE I.

ABBREVIATIONS,

B. Back surface covered with wings. O. Orifice for admission of light.
F. Front surface covered with wings. 8. Spectroscope.
Arrow indicates directions of rays of light.

Fig. 1. Perspective view of spectroscopic apparatus used in determining the
composition of the colors of Lepidoptera.

Fig. 2. Horizontal section of same. See p. 175.

Fig. 9. Pendulum used in determination of the frictional resistance between
the air and the wings of Lepidoptera. See p. 193.

Figs. 4,5. Diagrams to illustrate Keeler's method of projeetion, as applied to

Lepidoptera. See p. 207.

Mayer, — Color and Color-Patterns,

PLATE 2.

Diagrams to illustrate the laws which govern the Color-Patterns of Lepidoptera.

Fig. 6. Euthalia bellata (W. L. Distant, '82-'80, Plate 43, Fig. 12). Illustrates
the law of bilaterality of spots. See p. 183.

Fig. 7. Zethera musa (G. Semper, '86—92, Taf. 7, Fig. 10). Bilaterality of
double spots. See p. 183.

Fig.8. Eye-spotsin Morpho. See p. 182, 183.

Fig. 9. Parthenos gambrisius (W. L. Distant, '82-'86, Plate 11, Fig.
of complex spots, each one being similar to the rest, and bilaterally

"

7). A series

symmetrical.

Figs. 10, 11. Ornithoptera urvillana and 0. priamus (R. H. F. Ripon, 89-293).
Spots within spots, all being bilaterally symmetrical.

Figs. 12, 13. Hestia jasonia and H. leuconoe. Axis of lateral symmetry
(H, H), for spots passes through center of interspace. H. jasonia,
(F. Moore, 90-96, Plate 3, Fig. 1). H. leuconoe (G. Semper,
86-92, Taf. 1, Fig. 3).

Fig. 14. Papilio emalthion, to illustrate fusion of two rows of spots.

Fig. 15. Ornithoptera trojana, an apparent exception to the law of bilaterality.
See p. 187.

Fig. 16. Limenitis proserpina (S. H. Scudder, '88—89, Plate 2, Fig. 9), showing
fusion of two rows of differently colored spots. See p. 187.

Fig. 17. Saturnia spini, false eye-spot. See p. 187.

Fig. 18. Cases of degeneration of bands of color. See p. 184.

Fig. 19. Missanga patina (F. Moore, 90-296, Plate 72, Fig. 2°). Exceptional
form of eye-spot. See foot note p. 186.

Figs. 20-23. Hypothetical conditions of coloration, not found in nature, being

contrary to the laws of color-pattern. See p. 188.

MAYER - COLOR AND COLOR PATTERNS.

BUurL.Mt

COM ZOOL.VOL. XXX.

PLATE.2.

B. Morsel ith

Mayer. — Color and Color-Patterns.

PLATE 3.

To illustrate color-development in Callosamia promethea and Danais plexippus.

Fig, 24. Enlarged view of pupal wing of C. promethea in the “ white stage.’
See p. 178.
Pig. 25. Scale from wing of C. promethea in white stage of color-development,
showing the total absence of pigment in the scale. See p. 178.
26. Scale from light drab-colored area of mature wing of C. promethea.

Fi
Successive stages in the formation of color in pupal hind

g

Figs. 27, 36, and 25.
wing of C. promethea.

28, 37-40. Successive stages in the formation of color in the pupal fore
wing of 9 C. promethea. See p. 179, 180.

29, 30-35. Successive stages in the formation of color in the pupal wings
of d C. promethea. (Figs. 20, 30-33, 35 fore wing; Fig. 34 hind

Figs.

Figs.
,
wing.) See p. 179, 180.
and 41. Pupal hind wings of C. promethea, respectively mature d
and 9.
42-45. Successive stages in the color-development of D. plexippus.
p. 180—181.

Figs. 3.

See

Figs

MAYER. — Color and Color-Patterns.

PLATE 4.
Systematic analysis of the characteristic markings upon the wings of the Heli-
conidae. Homologous markings are designated by the same numerals, I, II,
III, etc.

”

Fig. 46. Lycorea ceres; an example of the “ Melinaea type" of coloration.
See p. 205.

Fig. 47. Thyridia psidii; an example of the “Ithomia type
See p. 206 and Plate 7, Fig. 79.

Fig. 48. Melinaea paraiya.

Fig. 49. Ceratinia ninonia.

Fig. 50. Heliconius antiochus.

Fig. 51. Napeogenes duessa.

Fig. 52. Ithomia sao.

Fig. 53. Melinaea gazoria.

Fig. 54. Ithomia nise.

Fig. 55. Mechanitis polymnia.

56. Eueides cleobaea.

Fig. 57. ‘Tithorea furia var.

Fig. 58. Heliconius eucrate.

Fig. 50. Heliconius melpomene.

" of coloration.

Heliconius erato.

LOR AND COLOR PATTERNS PLATE. 4.

MAYER - C

AG. de B Meise

r pany
v

> » T 1 " oM. Da X
BurL.Mus. Comp Z OGL Vor. XXX.

MAYER. — Color and Color-Patterns.

PLATE 5.

Color-patterns of the Antiochus and Erato groups of the Heliconidae projected
by Keeler's method. See p. 207.

Figs, 61, 02. Heliconius sara and H. antiochus, to show variation of yellow to
white. See p. 210, and Fig. 50, Plate 4.

g. 63. II. galanthus, showing development of white.

g. 64. II. charitonia, rows of double spots.

Fig. 65. H. phyllis, close relation between yellow and red.

Fig. 66. II. ricini.

Figs. 67, 68. II. erato; two color-ty pes.

Fig. 69. II. claudia; an example of the Sylvanus group.

HIND WING.
M pL

ls
U
«|

Ste» Ces at

( Pbea en t$ n

AGM del B Meise: Ah Bo

Buu. M

Heliconius sara

Godt.

, antiochius

Linn,

galanthus
Bides.

Hcharthonte

Linn.

IL phidls Febr,

IL reu Lint ,

Herato Linn

Herato Lint

Helandıu

hom A Sah

a
Fig.

MAYER, — Color and Color-Patterns.

Kueides.
Fig. 70.
Fig. 7T
Fig. 72,
Fig. 73
Fig. 74.
Fig. 75.
Figs. 76

78.

PLATE 6.

Color-patterns of the Melpomene group of Heliconius and of the genus

H. melpomene, the type of the Melpomene group. See p. 212, and
Fig. 59, Plate 4.

H. melpomene var. callicopis, showing the breaking up of the red
area of the primaries.

H. melpomene var. cybele ; the fore wing has assumed a color-pattern
which recalls the Melinaea type" of coloration found in the
Danaoid Heliconidae.

H. thelxiope, derived phylogenetically from H. melpomene, and
showing a rather elose approach to the * Melinaea type"
ation. See p. 212.

H. vesta.

Eueides thales ; represented to show the close resemblance of its
color-pattern to H. vesta. See p. 212.

77. E. mereaui and E. aliphera. E. mereaui is intermediate in color-
pattern between E.thales and E. aliphera.

Eueides cleobaea, to show the close approach of this insect to the

“ Melinaea type" of coloration.

of eolor-

MAYER - COLOR AND COLOR PATTERNS PLATE 55

HIND WING. FORE WING
N va Vill Be ID ID I o

£9, hs : : Heliconins

metpomene Linn

A melpomene var:

cally copus (ram

IL metpomene var:

hele (ram.

4 Thelciope
libn. d

Hvesta (ram.

Luetdes thates

(rams

E mereant

Hein

EL. adiphera
(rod t

E.eleobaea

hbn

MAYER. — Color and Color-Patterns.

PLATE 7.

Intended to show some types of coloration which are found in the Danaoid
Heliconidae, and also the remarkable individual variation in Mechanitis isthmia.
Fig. 79. Thyridia psidii, an example of the “Ithomia” type of coloration.

See p. 213, and Plate 4, Fig. 47.

Fig. 80. Napeogenes cyrianassa, showing semi-translucent condition of wings.
See p. 213.

Fig. 81. Ceratinia vallonia.

Fig. 82. Melinaea parallelis ; albinism of spots on primaries ; black band of hind
wing broken in the middle. See p. 188, 215.

Fig. 83. Ceratinia leucania, which probably mimics M. parallelis.

Figs. 84-87. Mechanitis isthmia, showing remarkable individual variation in
the black stripe of the hind wings, and also in the * inner yellow "

spot of the fore wings. See p. 184, 213.

y Dramy
MAYER > COLOR. AND GOluk PATTERNS POAT hee
HIND WING FORE WING
te D e qoem VENEN Me Vi oth ae qwe

li T6
Lodlg
Thiridta

psidii Linn

"n Mapeogenes

C

* Ceraiinia
vallonta Hew.

Nelınaca parallelis

Buil.

Coradinia lucanta

6.4
Bates,
Mechanitis
YE s M
isthinta Dates.
Mechanitis
D isthmin Bates.
Mechanitis
Sb

isthinia Baies.

Meehiettilus

esthinia Bades

MAYER,— Color and Color-Patterns.

PLATE 8.

Illustrates the mimicry between members of the “ Sylvanus” group of the genus
Heliconius and various Melinaeas, etc.

Fig. 88. Heliconius eucoma, an example of the“ Sylvanus” type of coloration
in genus Heliconius. See p. 214.

Figs. 89,90. Heliconius dryalus and Melinaea paraiya; close resemblance of
their color-patterns. See p. 214.

Figs. 91, 92, 98, 94. Respectively Heliconius eucrate, Melinaea, thera, Eueides
dianasa, and Mechanitis polymnia; showing close resemblance
between color-patterns. See p. 214.

Figs. 95, 96. Heliconius sylvana and Melinaea egina; these two forms are said
by Bates to mimic each other. See p. 214.

MAYER - COLOR AND COLOR. PATTERNS Prats. 8.

HIND WING FORE WING
vul E

S Heliconius
dd.
auoma Hübn.
Mor» y +
"m Heliconius
drvadus Hf.
Melinaea
40. : )
paraiva Weak.
Helrcontius
enerale bn
I.
Velinaca
9? ;
P^ hera teid .
| Anerdes dinasa
YS Lhilne.
Mechanitis
9% volvrrenia Linn.
Heliconius
95 sila Cane,
" leliru CCU
7 6

B ever (ram

B verae hih Boston

Burr Mus Comp Zoón VOL: XXX,

MAvER.— Color and Color-Patterns.

PLATE 9.

Diagrams to illustrate color-variations.

The various colors are laid off at definite intervals along the axis of abscissae,
and the ordinates represent the number of species which exhibit the various
colors.

Fig. 07. Represents the color-variations of the “inner rufous” area of the fore
and hind wings in the Danaoid Heliconidae. The full line represents
the variations of the fore wing. The dotted line those of the hind
wing. The closeness of these two lines shows the intimate relations

,

between the color-variations of the ** inner rufous" areas upon fore
and hind wing. See p. 218.

Fig. 08, "Phe full line represents color-variations of “inner yellow " spot of fore
wings in Danaoid Heliconidae. The dotted line represents same for
“middle yellow." It is apparent that the “inner yellow“
variable than the outer yellow," and also that the variations of

is more

both are quite similar to those of the “inner rufous.” See p. 219.
Fig. 99. Color-variations of “inner rufous” areas of Acraeoid Heliconidae.
The full line represents the fore wing and the dotted line the hind

wing. See p. 218.

"TI

32v18

100

un
e

150

MAYER, — Color and Color-Patterns.
.

PLATE 10.

Fig. 100. Color-variations of “inner yellow " spots on fore wings of Acraeoid
Heliconidae. The full line represents the * inner yellow,” the dotted
line the “ middle yellow." See p. 221.

Fig. 101. Variations of marginal spots upon fore wing in Danaoid Heliconidae.
'These spots tend to appear either as 2 or 3, or as 6 or 7 spots. See
p. 223.

Fig. 102. Variations of marginal spots of hind wings in Danaoid Heliconidae.

These spots tend to appear either as 4 or as 5 spots. See p. 223.

/00

3903953018)

3n198

„9vy718

3101HM

38H20

MOI13A

SOA

fol

102

B. Meisel lith Boston

A G.M. dei

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.
Vor. XXX. No. 5.

THE MESENTERIES AND SIPHONOGLYPHS IN METRIDIUM
MARGINATUM MILNE-EDWARDS.

Bv G. H. PARKER.

Wira ONE PLATE.

CAMBRIDGE, MASS., U. S. A.:
PRINTED FOR THE MUSEUM.
March, 1897.

No. 5. — The Mesenteries and Siphonoglyphs in Metridium mar-
ginatum Milne- Edwards. By G. H. PARKER.!

Introduction. — Since the publication of the Hertwigs' (79) paper on
the anatomy of the actinians, the attention of investigators has been
more and more directed toward the details of the internal structure of
these organisms. ‘This new departure has been conducted in the main on
the lines of systematic zoölogy, and, though its advocates in the begin-
ning may have been somewhat Utopian in their expectations, it has
certainly carried our understanding of the natural relations of this group
of animals a long step forward. The new features thus introduced into
the classification have, however, been subject to frequent modification,
and every actinian newly investigated may be expected to exert some
influence on the classification finally adopted. It is to be regretted that
much of this kind of investigation has been of necessity carried out on a
limited, often a very limited, number of specimens, so that the possible
error of regarding individual variations as characteristic of large groups
is not always eliminated.

The following pages contain a record of certain structural peculi-
arities in a single species of actinian, the common Metridium marginatum
Milne-Edwards of our coast, as represented by a considerable number
of specimens. As this record shows, uniformity of structure is by no
means a general characteristic of this species; hence these observa-
tions are to some extent a contribution to the study of the variability
of this animal.

The material on which the following observations were made consisted
of 131 adult specimens of Metridium marginatum. These were collected
in part by myself and in part by my laboratory assistant, Mr. J. I.
Hamaker, to whom I am under obligations for this kindness. All the
specimens came from the neighborhood of Newport, R. I., and were
prepared and to some extent studied in Mr. Alexander Agassiz's Labora-
tory at that place. I here wish to express my thanks to Mr. Agassiz for
the privilege of carrying on this work at the Newport Laboratory.

1 Contributions from the Zoölogical Laboratory of the Museum of Comparative
Zoölogy at Harvard College, E. L. Mark, Director, No. LXXV.
VOL. XXX. — NO. 5. 1

260 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

The specimens were prepared by the Tullberg (91) method, which
consists in stupefaction by the gradual introduction of magnesic sulphate
into the water containing the actinians, and in subsequent hardening by
means of chromic acid. This method, when properly employed, yields
beautifully expanded and thoroughly hardened specimens, and my expe-
rience with it has been such that I can fully indorse the recommendations
given it by Tullberg (91), Carlgren (93, p. 7), and others. Speci-
mens prepared in this way were cut transversely with a common razor,
and the number and arrangement of the mesenteries and siphonoglyphs !
were recorded. Owing to the large size of the specimens, this could be
easily done under the magnification of an ordinary hand lens.

Siphonoglyphs. — The Hexactinia, to which Metridium belongs, were
until recently supposed to possess always two siphonoglyphs; but this
surmise has been shown to be not well grounded, and, in, the species
under consideration, as MeMurrich (91, p. 131) has already pointed
out, either one or two siphonoglyphs may be present. Iu the 131
specimens that I examined, 77 (or about 59 per cont) had only one
siphonoglyph (Fig. 3), 53 (or about 41 per cent) had two siphonoglyphs
(Fig. 1), and a single specimen possessed three such organs (Fig. 6). In
no instance was a specimen found without a siphonoglyph. ‘The smooth
surface of the siphonoglyph is so strongly contrasted with the longitu-
dinally ribbed surface of the rest of the oesophagus that in none of the
specimens examined was there any uncertainty as to the number of
siphonoglyphs present. The striking difference between these two kinds
of surface cannot be made to appear so clearly in tho figures as it did
in the actual specimen, where, in addition to tho cut face, the natural face
of the esophagus could also be inspected. MeMurrich (QUIS De lol)
remarks that in the individuals examined by him, those with one sipho
noglyph were almost, if not quite, as frequent in oceurrence as tho
with two, but in my enumeration it will be seen that they were really
somewhat more numerous.

Since only one of the 131 specimens possessed three siphonoglyphs, it

1 The term “ siphonoglyphe " was first introduced into zoólogical nomenclature
by Hickson ('84, p. 694), and has since been widely accepted. Professor Hickson
kindly informs me that the last syllable of this term is derived from the Greck
word yAvgts, which in the plural form, yAvoí8es, has been used to signify the grooves
on an arrow for the insertion of the feathers. The root of this word appears to
call for no final e, and since in making English words it is best, as Professor
Hickson remarks, to use only roots, I therefore propose to change the spelling
of the term in question by omitting the final e, and to this Professor Hickson
assents.

PARKER: METRIDIUM MARGINATUM. 261

is obvious that this condition may be set aside as distinctly exceptional,
and, further, since the other specimens were almost equally divided be-
tween those with one and those with two siphonoglyphs, these conditions
may fairly be considered typical. It will be convenient in the subse-
quent discussion to designate these two types by special names, and I
shall call that characterized by one siphonoglyph the monoglyphic type,
and that by two the diglyphie type.

Variations in the number of siphonoglyphs have already been recorded
in other actinians. ‘Thus, besides the observations of McMurrich already
alluded to, Thorell (59, Tab. I. Figs. I and 2) figured and described
specimens of Metridium dianthus either with one or with two siphono-
glyphs. The monoglyphic condition was also recognized for this species
by Gosse (760, p. 12), who, in ignorance of Thorell's observations, sup-
posed this condition to be characteristic of the species, a mistake after-
wards corrected by Foot (63, p. 64). The presence in some specimens
of one, and in others of two siphonoglyphs in M. dianthus, as first as-
serted by Thorell, has recently been confirmed by G. Y. and A. F.
Dixon (91, p. 19), and by Carlgren (793, p. 104). Furthermore, the
Dixons and Carlgren agree in stating that, though two siphonoglyphs
may be present in this species, one is the rule. G. Y. and A. F. Dixon
(91, p. 20), moreover, have recorded one specimen of M. dianthus with
three siphonoglyphs.

Representatives of the genus Sagartia also show variations in the
number of their siphonoglyphs; thus G. Y. Dixon (788, p. 120) ob-
served that in Sagartia venusta, S. nivea, and S. mineata, either one or
two siphonoglyphs may be present. The same is probably true of
S. rosea (cf. F. Dixon, '88, p. 139), and of S. lactea (cf. McMurrich,
94, p. 177). In specimens of Bunodes thallia, studied by G. Y. and
A. F. Dixon (89, p. 318), one, two, three, and even four siphonoglyphs
were observed, although in each of twenty-three adult specimens of
B. verrucosa the same authors (89, p. 322) found regularly two siphono-
glyphs. Finally Blochmann and Hilger (88, p. 391) described a specimen
of Gonactinia in which traces of a third siphonoglyph seem to have
been present.

It is evident from the foregoing account that in several actinians besides
Metridium a variation in the number of siphonoglyphs is not unusual,
though this variation may not be so pronounced as to constitute a struo-
tural type. The importance of these peculiarities from a systematic
standpoint has already been appreciated, and in the more recent defini-
tions of the TTexactinia the statement is made that these actinians possess

262 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

two siphonoglyphs (occasionally one), thereby recognizing the mono-
glyphic and diglyphic types as normal.

Mesenteries. — In Metridium marginatum the pairs of mesenteries are
attached lengthwise to the wall of the column, and either reach the
cesophagus and unite with it (complete mesenteries) or fall short of
that structure (incomplete mesenteries) Of the pairs of complete
mesenteries the two usual kinds can be distinguished: those whose
longitudinal muscles face the exocols (directive mesenteries) and
those whose longitudinal muscles face the endocods (non-directive
mesenteries).

The dérective mesenteries are remarkable for the constancy of their
relations to the siphonoglyphs. To each siphonoglyph is attached a
single pair of directives, and in no instance among the 131 specimens
examined was an exception to this rule found. In the monoglyphic
type (Figs. 3, 4, 5, 7, and 8) one pair, and only one pair, of directives
was present; in the diglyphic type (Figs. l and 2) two pairs were
invariably observed ; and even the single specimen with three siphono-
glyphs (Fig. 6) formed no exception, but exhibited three pairs of
directives.

This exact correlation between the number of siphonoglyphs and of
directives, which probably also obtains in other species of Metridium
(cf. Carlgren, 93, p. 106), as well as in the allied genus Sagartia (cf.
F. Dixon, ’88, p. 136), is rather striking, because .the two sets of
structures concerned are not invariably thus associated in all actinians.
For instance, in Peachia and Oractis (ef. MeMurrich, ’91, pp. 135, 137)»
though two pairs of directives are present, only one siphonoglyph
occurs ; and in Ptychodactis (cf. Appellóf, 94, pp. 5, 7), though two
pairs of directives can be seen, no siphonoglyphs are observable. These
show that in some actinians directive mesenteries
and thus they render more striking
s of the directives and of the

instances serve to
may occur without siphonoglyphs,
the correlation between the variation
siphonoglyphs in Metridium marginatum.

The non-directive mesenteries vary so much in their number and ar-
rangement that they can best be considered in connection with the par-
ticular types with which they occur. In the diglyphie type (53 speci-
in addition to the two pairs of directives, there may be from

mens),
The frequency of the occurrence

four to ten pairs of non-direocti ves.
of the different numbers of pairs is indicated in the following table: —

PARKER: METRIDIUM MARGINATUM. 263

Diernvrnio Tyre (two siphonoglyphs and two pairs of directives).

Pairs of Non-directives . . . . 4 5 6 7 8 9 10

Number of Cases observed . . 40 4 9 3 2 0 1

In this type (Fig. 1) the two pairs of directives of course divide the
non-directives into two groups. I regret that, before it occurred to me
to determine the number of mesenteries in each of these two groups for
the 53 specimens of this type, some of the specimens were so far dis-
sected as to render them no longer serviceable for this enumeration. I
can therefore make a statement concerning this division in the cases of

only twenty specimens.

Groups of Non-directives. .|1+7/2+2/2+3/2+4 27 5 27 6 44-6

Number of Cases observed . 1 10 2 2 3 1 1

In comparing these results, it will be observed, first, that the great
majority of individuals (40 in 53) possess four pairs of non-directives,
and, next, that the arrangement of these non-direotives in ten cases out
of twenty is in two groups of two pairs each. This symmetrical
arrangement of the four pairs of non-directives in the diglyphic type
reproduces the assumed typical Hexactinian arrangement, and, since
the representatives of the other variations are comparatively so few in
numbers, this may be taken to be the only characteristic condition of
the diglyphie type.

In the monoglyphie type (77 specimens), in addition to the one pair
of directives, there were from three to fourteen pairs of non-directives.
The frequency of their occurrence is shown in the following table: —

Mowoarvyruro Typer (one siphonoglyph and one pair of directives).

Pairs of Non-directives . . 3 465 6 7 | 8 | 9 | 10} 11) 12] 18 | 14
|
|
|

Number of Cases observed .| 1 | 4 a 19 21) 5 JE 1:590 106€ T

Admitting the monoglyphie type to be derived from the diglyphie by
the conversion of a pair of directives into a pair of non-directives,

264 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

one would expect the monoglyphic type with five pairs of directives to
be most often met with. Such, however, is not the case, for specimens
with six or seven pairs of non-directives are about as numerous as those
with five. Since any of the three groups with five, six, or seven pairs
of non-directives is represented by a greater number of individuals than
all the other minor groups of variations taken collectively (of. Table of
Monoglyphic Type), it is clear that in the monoglyphic type there are
three structural subtypes characterized respeotively by fivo, six, and
seven pairs of non-directives, instead of only a single such subtype, as in
the diglyphic condition. These relations indicate a certain degree of
distinctness between the diglyphie and the monoglyphie type; for the
monoglyphic has obviously a greater range in variation, as shown in
its three subtypes, than the diglyphie with only a single one. It is an
interesting,fact in this connection, that the monoglyphie subtype with
six pairs of non-directives often repeats (Fig. 4), so far as its com-
plete mesenteries are concerned, the arrangement of mesenteries found
in Seytophorus, for which R. Hertwig ('82, p. 104) constructed a sep-
arate family, the Monaulec.

It might at first be suspected that the three monoglyphie subtypes
pointed out above, and in fact all the variations in the number of com-
plete mesenteries, could be explained on the assumption that certain
incomplete mesenteries by excessive growth had become complete, or
that complete ones had become incomplete, thus introducing a varia-
tion in the number of complete mesenteries, without, however, altering
the total number of all kinds of mesenteries ; but in the individuals ex-
amined the relative development of the incomplete mesenteries was
found to be subject to so much variation that the satisfactory deter-
mination of the total number of mesenteries as a basis of comparison
was practically impossible, and all attempts to carry through interpre-
tations such as that suggested above resulted in such ambiguous and
strained results that the unnaturalness of the method condemned it.
Moreover, in the monoglyphie type with sé pairs of non-direetives
(Fig. 4), incomplete as well as complete mesenteries are sometimes so
symmetrically placed that no attempt to readjust them is warranted.
What may be said of such cases is, that, in place of the usual five pairs
of non-directives, six pairs are present, and this increase cannot be
ascribed to reinforcement from the ranks of incomplete mesenterics.
Such cases as these are so frequent, and instances that may be inter-
preted as the conversion of complete into incomplete mesenteries or
the reverse are so few, that it must be admitted, I believe, that these

PARKER: METRIDIUM MARGINATUM. 265

differences are due much more frequently to fundamental differences in
the plans on which the mesenteries of different individuals are laid
down than to the more easily conceived relation between complete and
incomplete mesenteries.

The incomplete mesenteries have not been exhaustively investigated.
Their great number, variability in size, and the frequent difficulty met
with in attempting to classify them, render such a task nearly impos-
sible. In what are generally assumed to be the more typical specimens
of Metridium (Fig. 1), an exocol may contain one pair of secondary
mesenteries, two pairs of tertiaries, four pairs of quarternaries, and
evidences (ridges) of eight pairs of quinaries. Though this condition
was occasionally realized, in the great majority of cases irregularities in
what are presumably secondaries and tertiaries, not to mention higher
orders, were so numerous that consistent tabulation was out of the
question. So far as size and position were concerned, what seemed to
be secondaries showed such variations that no two specimens in which
the arrangement of the complete mesenteries agreed, had similar arrange-
ments of the secondaries, except in six instances of the 40 typical
diglyphic specimens; and each of these six instances showed variations
in the tertiaries characteristic of it as an individual. So far, then, as
the incomplete mesenteries are concerned, we soon reach groups of vari-
ations by which individuals may be characterized ; in other words, if the
variations of the primaries (complete mesenteries), secondaries, and
tertiaries be considered together, it will be seen that no two of the
131 specimens examined were alike, each one having a combination
of variations peculiar to itself. This is, perhaps, the most important
feature in the variations of the incomplete mesenteries.

That variations in the number of mesenteries, such as have been
pointed out in the preceding paragraphs, occur in other actinians: is
well known. Thus Carlgren (793, p. 106) states that in Metridium
dianthus, in addition to a single pair of directives, six, seven, on even
nine pairs of non-directives may occur, and F. Dixon (88) has shown
that in several species of Sagartia the number of non-direotives may
reach twelve or even sixteen pairs. Further, in four speeimens of
Bunodes thallia, G. Y. amd A. F. Dixon (89, pp. 317, 318) found
respectively 15, 19, 21, and 26 pairs of non-directives, These citations
suffice to show that extensive variations in the mesenteries may occur
in other actinians than Metridium marginatum, but the cases recorded
for any one species are so few that generalizations cannot be drawn

from them.

266 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

As a rule, variations in the mesenteries occur in both members of a
pair in the same way, but not infrequently one finds pairs in which the
two members are not equally developed. When this occurs amongst
the complete or nearly complete mesenteries, it may result in the forma-
tion of à pair one member of which is complete and the other incom-
plete (Fig. 2). The 131 specimens of Metridium examined possessed
in all 739 pairs of non-directives, and, of these, 17 pairs (or about 2.4
per cent), distributed through thirteen individuals, possessed each an
incomplete member. Of the thirteen individuals exhibiting this vari-
ation, ten were of the monoglyphic type, and three of the diglyphic
type. In the monoglyphic type it is customary to assume that the
single siphonoglyph present corresponds to the so-called ventral one
of the diglyphie condition. This assumption is at least convenient,
for it allows ys to distinguish in each pair of lateral non-directives a
dorsaland a ventral member. Admitting this distinction for the sake
of description, it may be said that seven of the ten monoglyphic
specimens had each a single pair of non-directives in which one mem-
ber was incomplete, and of these incomplete mesenteries four were
dorsal, two were ventral, and one was indeterminable (Fig. 8); and
that the three remaining monoglyphic specimens had each two such
pairs, of which in one instance both the incomplete mesenteries were
dorsal, and in two instances one was a dorsal and the other a ventral
mesentery (Fig. 7). Thus in the ten monoglyphic specimens, this
variation was observed in thirteen pairs of mesenteries, of which eight
presented incomplete dorsal members, four incomplete ventral members,
and one was indeterminable. It is evident that this variation is not
limited to either dorsal or ventral members, and is not correlated with
the fact that in many actinians ventral members, as a rule, develop later
than dorsal ones; in other words, this variation is probably not to be
regarded as atavistic.

In the adult condition of the diglyphie type, I see no way of distin-
guishing dorsal from ventral, and the most that can be said of the three
cases of variation met with under this type is that in two of them only
one mesentery each was incompleto (Fig. 2), while in the third two were
incomplete. In the latter case the two mesenteries (as in Fig. 7) were
not on corresponding sides; hence one of them must have been dorsal
and the other ventral, but exact determination could not be made. The
variations in the diglyphic type, then, present no essential features
not already mot with in the monoglyphic type.

Many pairs of mesenteries in which both members are incomplete

PARKER: METRIDIUM MARGINATUM. 267

show variation of the kind indicated above, in that one member is
larger than the other (Fig. 4), but because of the extreme variability of
these parts no record has been kept of such variations.

In a few cases single mesenteries have been observed (Fig. 2). These,
as the arrangement of the longitudinal muscles of their neighbors shows,
have absolutely no trace of a mate. In the instance figured, it is diffi-
cult to decide which of the two mesenteries, the complete (y) or the
incomplete (x), is the single one. One or other must be. Single mesen-
teries as exceptions have already been recorded by F. Dixon (88, p. 138)
in Sagartia, and by Carlgren (793, p. 106) in Metridium.

Among the complete mesenteries, two cases of union by what would
have been the median margins of the participants have been observed
(Fig.8). An instance of this kind has already been recorded by R.
Hertwig (82, p. 37) in Tealia, and in this, as in Metridium, the united
mesenteries were not members of the same pair, but of adjacent pairs.

No instances of the occurrence of longitudinal muscles on both the
exoool and the endoccel face of the same mesentery, as observed by
McMurrich (89, p. 30) in Aulactinia, have been noticed.

So far as the mutual arrangement of complete and incomplete
mesenteries is concerned, the monoglyphie and diglyphic types show
rather characteristic differences. In the diglyphie type the complete
mesenteries usually show no special tendency to collect at one pole or
the other of the animal (cf. Fig. 1). In the monoglyphic type there is
often a marked tendency for all but two pairs of the non-directives
to collect opposite the directives (cf. Fig. 5) ; consequently the half of
the animal centering about the directives has an arrangement of parts
like that found in the corresponding half of a diglyphic animal, while
the other half contains a more or less orowded group of non-directives.
In this respect Metridium seems to differ from Sagartia, in which,
according to the figures given by F. Dixon ('88, Plate I.), such a
crowding of non-directives is not noticeable, This condition recalls in
a superficial way that found in Cerianthus, in which an active growth
of mesenteries takes place opposite the siphonoglyph.

The characteristic arrangements of the mesenteries in connection with
the monoglyphic and diglyphic types probably recur under similar con-
ditions in M. dianthus; for such arrangements have been figured by
Thorell (59, Tab, I. Figs. 1 and 2) and briefly described by Carlgren
(793, p. 106).

While the crowding of the mesenteries occurs as a rule only in mono-
glyphic specimens of Metridium, one instance of it has been observed

268 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

in a diglyphie specimen (Fig. 2), and here vh. general resemblance to
the specimen with three siphonoglyphs (Fig. 6) is so striking that
I have felt almost justified in interpreting this specimen as a triglyphic
animal, at one pole of which the directives, with the loss of the siphono-
glyph, had given place to a group of non-directives.

In tne preceding account I have intentionally avoided, as far as pos-
sible, the use of the terms dorsal and ventral as applied to the two poles
of the actinian’s body. This has not been because of objections that
might well have been raised against these terms in themselves, as
Haddon (789, p. 300) has done, but because of the more fundamental
question of whether dorsal and ventral can really be distinguished in an
adult Metridium. These terms, as is well known, may be applied with
perfect precision to the adults of forms like Edwardsia, where the longi-
tudinal muscles bear very unlike relations to the two poles of the animal ;
but in forms like the diglyphic type of Metridium (Fig. 1), where the
muscles of the pairs of non-directives are similarly related to both poles,
this means of distinguishing dorsal and ventral is lost. It has been
suggested that even in cases of this kind dorsal and ventral may still be
distinguished, either by the conditions of the siphonoglyphs, — the ven-
tral being better developed than the dorsal (Faurot, '05, p. 62), — or
by the condition of the subsidiary mesenteries, — the more dorsal pairs,
because of their earlier development, remaining larger than the ventral
ones (Carlgren, ’93, p. 100). Unfortunately, these criteria, even sup-
posing them to be true, which is by no means certain, cannot be em-
ployed on the diglyphie type of Metridium because of the similarity of
its two poles. So far as the adult diglyphie Metridium is concerned, I
am obliged to confess that I can find no satisfactory criteria for the
determination of dorsal and ventral relations.

With the monoglyphie type the case seems simpler. It is generally
stated that, when only one siphonoglyph is present, it is the ventral one ;
but, as Carlgren (93, p. 100) remarks, so far as Sagartia is concerned,
this statement has never been accompanied with any direct proof; nor,
I may also add, has it been proved for Metridium. The argument used
by MeMurrich (91, p. 133) to show that the single siphonoglyph in the
monoglyphie Metridium is the ventral one may be used with equal accu-
racy to show that this siphonoglyph is the dorsal one, for the argument
advanced rests upon the sequence of the development of the mesen-
teries, which, being unknown in Metridium, has simply been assumed by
McMurrich. The case of Metridium seems to be precisely like that of

PARKER: METRIDIUM MARGINATUM. 269

Sagartia, in which, as Haddon (789, p. 300) remarks, it seems impossible,
in our present state of knowledge, to determine dorsal and ventral rela-
tions. It is probable that this determination can be made only after
the sequence of development of these mesenteries has been discovered.
In the four types of sequence thus far known (cf. Fowler, 94, p. 470),
the ventral directives are always the third pair of mesenteries to form,
and the dorsal directives either the second or fourth. It is probable
that, when the developmental sequence of the mesenteries is discovered
for the two types of Metridium, the determination of dorsal and ventral
in this actinian will he made with as much certainty as in any other,
and we shall probably then know whether in the monoglyphie type tho
single siphonoglyph is a dorsal one, a ventral one, or in some specimens
one and in others the other.

Before concluding this account of the mesenteries in Metridium, I
wish to consider briefly some other aspects of the monoglyphio and
diglyphie types. When I first perceived that there were two structural
types in Metridium, I suspected that they might be correlated with
sexual differences. To test this question, I determined the sexes of a
number of individuals of each type. In ten monoglyphie specimens,
five were females and five were males; in twenty-seven diglyphie speci-
mens, fifteen were females and twelve were males. Evidently the two
types are not correlated with difference in sex.

The fact that the two sexes occur in about equal numbers under both
types suggests that these two types may in reality be two varieties of
the species Metridium marginatum. In support of this opinion, it may
be mentioned that the two types show a difference in the degree of their
variability, tho diglyphic type having only one subtype, the monoglyphic
three; and, further, that, while the diglyphic type presents usually a
rather typical Hexactinian arrangement of mesenteries, the monoglyphie
type shows a general tendency to crowd the non-directive mesenteries to
the region opposite the one siphonoglyph.

These differences, however, fairly marked as they are, are insufficient
in my opinion to warrant the assumption that the two types are really
varieties, and tho determination of this question must wait, I believe,
till more is known of the breeding habits of Metridium. If it can be
shown that in the offspring of one animal representatives of both typos
occur, the idea that we are dealing with varieties could not be main-
tained, and the species could at most be said to be dimorphic. If, how-
ever, tho types could be shown to breed true, they might with justice
be described as varieties.

270 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Should these types prove not to be of the value of varieties, they may
still possibly be correlated with the methods of reproduction. Besides
the sexual method, Metridium marginatum reproduces non-sexually by
small buds cut off from the margin of the animal between its aboral
disk and its column. This method of reproduction, long ago hinted at
by Verrill (69, p. 257), can usually be seen taking place in any large
specimen.. Similar conditions have been observed in Metridium dianthus
by G. Y. and A. F. Dixon (91, p. 20), and by Carlgren (93, p. 108).
Possibly the two types here described are the products, one of the
sexual, the other of the non-sexual, method of reproduction. The solu-
tion of this question, however, must be left to future investigation.

CAMBRIDGE, January 9, 1897.

PARKER: METRIDIUM MARGINATUM. 271

PAPERS CITED.

Appellöf, A.
94. Ptychodactis patula, n. g. & sp. der Repräsentant einer neuen Hex-
actinien-Pamilie. Bergens Museums Aarbog for 1893, No. LV. pp. 1-22,
Pls. I-III.
Blochmann, F., und Hilger, C.
88. Ueber Gonactinia prolifera, Sars. Morph. Jahrb., Bd. XIII. pp. 385-
401, Taf. XIV., XV.
Carlgren, O.
'93. Studien über nordische Actinien. Kongl. svenska Vet-Akad. Hdlgr.
N. F., Bd. XXV. No. 10, pp. 1-148, Taf. I.-X.
Dixon, F.
88. On the Arrangement of the Mesenteries in the Genus Sagartia, Gosse.
Sci. Proc. Roy. Dublin Soc., N. S., Vol. VI. pp. 136-149, Pls. I., II.
Dixon, G. Y.
88. Remarks on Sagartia venusta and Sagartianivea. Sei. Proc. Roy. Dub.
lin Soc., N. S., Vol. VI. pp. 111-197.

Dixon, G. Y., and Dixon, A. F.

'89. Notes on Bunodes thallia, Bunodes verrucosa, and Tealia erassicornis.

Sci. Proc. Roy. Dublin Soc., N. S., Vol. VI. pp. 810-326, Pls. IV., V.
Dixon, G. Y., and Dixon, A. F.

91. Report on the Marine Invertebrate Fauna near Dublin. Proc. Roy.

Irish Acad., 3d Ser., Vol. II. pp. 19-33.
Faurot, L.

95. Etudes sur l'Anatomie, l'Histologie et le Développement des Actinies.
Arch. Zool expór. et gén., 3e Ser, Tome IIL pp. 43-262, Pls.
I.-XII.

Foot, F. J.

'63. Notes on the Astreacea of the Coast of Clare. Proc. Nat. Hist. Soc.

Dublin, Vol. ILI. pp. 63-69.

212 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Fowler, G. H.
'94. Octineon Lindahli (W. B. Carpenter): an Undeseribed Anthozoon of
Novel Structure. Quart. Journ. Mier. Sci., Vol. XXXV. pp. 461-480,
Pls. 29, 30.
Gosse, P. H.
'60. A llistory of the British Sea-anemones and Corals. London, 1860.
viii + 362 pp., Pls. I.-XII.
Haddon, A. C.
89. A Revision of the British Actinix. Part I. Sci Trans. Roy. Dublin
Soc., 2d Ser., Vol. IV. pp. 297-361, Pls. XXXI.-XXXVII.
Hertwig, O. und R.
79. Die Actinien anatomisch und histologisch mit besonderer Berücksichti—
gung des Nervensystems untersucht. Jena. Zeitschr., Bd. XIII. pp. 457-
640, Taf. XVII.-XXVI.
Hertwig, R.
82. Iteport on the Actiniaria dredged by II. M. S. Challenger during the
years 1873-1876. The Voyage of H. M. S. Challenger, Zoology, Vol.
VI. pp. 1-136, Pls. L-XIV.
Hickson, S. J.
84. On the Ciliated Groove (Siphonoglyphe) in tho Stomodeeum of the
Aleyonarians. Phil. Trans. Roy. Soc., London, Vol. 174, pp. 693-705,
Pls. 50, 51.
McMurrich, J. P.
89. The Actiniaria of the Bahama Islands, W. I. Journ. Morph., Vol. III.
pp- 1-74, Pls. I.-IV.
McMurrich, J. P.
91. Contributions on the Morphology of the Actinozoa. III. The Phylo-
geny of the Actinozoa. Journ. Morph., Vol. V. pp. 125-164, Pl. IX.
McMurrich, J. P.
94. Report on the Actiniæ collected by the United States Fish Commission
Steamer Albatross during the Winter of 1887-1888. Proc. U. 8. Nat.
Mus., Vol. XVI. pp. 119-216, Pls. XIX.-XXXV.
Thorell, T. i
'59. Om den inre bygenaden af Actinia plumosa Müll. Ofversigt kongl.
Vet-Akad. Fórhandl, Vol. XV. pp. 7-25, Tab. I.
Tullberg, T.
'91. Ueber Konservierung von Evertebraten in ausgedehntem Zustand. Biol.
Fören, Fórhdlgr., Bd. IV. pp. 4-9.
Verrill, A. E.
'69. Our Sea-Anemones. Amer. Naturalist, Vol. IT. pp. 251-262.

|
|
|
|

Parker. — Metridium marginatum,

EXPLANATION OF PLATE.

All figures represent transverse sections of the column of Metridium marginatum
Milne-Edwards. In each case all the complete mesenteries have been drawn, and
in some instances a few of the incomplete ones. All figures are magnified about

1.5 diameters.

Fig.

Fig.

1.

2.

Diglyphie specimen, showing typical hexamerous arrangement of the
complete mesenteries, and a regular group of subordinate mesen-
teries in a primary exocal.

Diglyphie specimen, showing a very irregular arrangement of the com-
plete mesenteries, one group of which resembles the crowded group
of mesenteries found opposite the directive pole in the monoglyphie
type. In this crowded group occurs an unpaired mesentery (x or y).

Monoglyphie specimen, with five pairs of non-directives, showing the
regular arrangement of the subordinate mesenteries about tie pair of
non-directives opposite the directive pole.

Monoglyphie specimen, with six pairs of non-directives, showing the
regular arrangement of the subordinate mesenteries about the two
pairs of non-directives opposite the directive pole.

Monoglyphie specimen, with ten pairs of non-direetives, showing the
crowding of the non-directives in the region opposite the directive
pole.

'Triglyphie specimen.

Monoglyphic specimen, showing pairs of mesenteries in which dorsal as
well as ventral components are incomplete.

Monoglyphie specimen, showing the union of two primary mesenteries.
UT I , g

Bulletin of the Museum of Comparative Zoölogy
AT HARVARD COLLEGE
Vor. XXX. No. 6.

PHOTOMECHANICAL CHANGES IN THE RETINAL PIGMENT
CELLS OF PALZXMONETES, AND THEIR RELATION
TO THE CENTRAL NERVOUS SYSTEM.

Bv G. H. PARKER.

Wıru ONE PLATE.

CAMBRIDGE, MASS., U. S. A.:
PRINTED FOR THE MUSEUM.
APRIL, 1897.

No. 6. — Photomechanical Changes in the Retinal Pigment Cells of
Palemonetes, and their Relation to the Central Nervous System.
By G. H. PARKER.!

CoNTENTS.
Page Page
Introduction . . 275) 6. Summary of Changes in Nor-
Structure of the Eye in Palamonetes 276 THe Rinn 4. Se, » 05
Photomechanical Changes in Nor- Sympathetic Photomechanical
wal Net“ 28 CHAR 9S, ENS. 280
er las heey 278 | Localized Photomechanical Changes 290
2. General Changes in Retina . 279| Photomechanical Changes in Ex-
3. Changes in Proximal Retinu- cised Eyes and Retinas . . . 291
lär Golls s- . * . e . . 210 | General Summary, «. vo» 0. r 200
4 Changes in Accessory Celle”. 281 Note 29097
5. Changes in Distal Retinular Papers Cited 771 de
Cells. . . . 284| Explanation of Plate.

INTRODUCTION.

THE present paper is a record of a series of experiments on the photo-
mechanical changes in the pigment cells of the retina in Palemonetes
vulgaris Stimp. This species is especially favorable for such work,
since its retina exhibits in a marked degree all the kinds of pigment
changes that have thus far been observed in the eyes of crustaceans.

Those that have worked upon this subject have, in the main, fol-
lowed in the lines laid down by Boll, Engelmann, and others in their
studies on the eyes of vertebrates. Although the pigment changes in
vertebrates are relativoly simple, they are, even now, far from being
satisfactorily understood, and it is therefore not surprising that in
the arthropods, where the pigment changes of the retina are probably
more complex than in any other group of animals, much still remains
to be done. There has been a tendency, moreover, among some of
those that have studied such phenomena, to generalize on observations
taken from eyes of totally different types, such as the compound eyes
of insects and the simple eyes of arachnids; and this tendency, though

1 Contributions from the Zoólogical Laboratory of the Museum of Comparative

Zoology at Harvard College, E. L. Mark, Director, No. LXXVI.
VOL, XXX. — NO. 6. 1

276 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

in a measure justifiable, has led, I believe, to a want of attention to
the characteristic differences of the pigment changes in each given type
of eye, a matter that, in my opinion, lies at the foundation of any
satisfactory understanding of these changes. What is most needed at
present, therefore, seems to be a thorough and exhaustive study of the
pigment changes of each of the more important types, rather than an
inspection, necessarily more or less superficial, of the various arthropod
eyes that have not as yet been examined. The following studies have
been made with the hope that they would contribute in this respect to
a more complete understanding of the pigment changes in the com-
pound eye, especially in crustaceans.

The earliest paper on the pigment changes in compound eyes, so far
as Iam aware, was published in 1889 by Exner (789), and contains in
a condensed form the essential peculiarities of the pigment changes
in the compound eyes of certain insects. In 1890 Stefanowska (90)
published an account in which this subject was again considered, but
with a wider range of material In the next year three contributions
appeared: Exner's (’91) brilliant and important essay on the physi-
ology of compound eyes, of which his former publication had been in
the nature of a partial preliminary notice ; Szczawinska’s (’91) article
on the pigment changes in the eyes of crustaceans and arachnids; and
Herrick’s (91) account of similar changes in the eyes of Palamonetes,
contained in his monograph on the development of Alpheus. Three
years later Kiesel (94) described some very noteworthy observations on
the pigment changes in the eyes of insects. The following year the writer
(Parker, '95) published, in connection with other matters, an account
of the retinal pigment changes in Astacus, and a preliminary state-
ment of the results given in full in this paper was published last year
(Parker, 96). These, I believe, are all the publications in which the
questions here raised have been considered. Critical comments on their
contents will be found in the following pages.

ÖTRUOTURE OF THE Eym IN PALZXMONETES.

Before describing the pigment changes in the retina of Paloemonetes,
it will be necessary to outline briefly the structure of the eye in this
animal. The eye may be said to be that portion of the optic apparatus
contained in the optic stalk. It consists of a retina, at the distal end
of the stalk, and a series of four optic ganglia, which extend through
the axial portion of the stalk. The retina is connected with the first

PARKER: RETINAL PIGMENT CELLS OF PALAIMONETES. 277

optic ganglion by the retinal nerve fibres. Nerve fibres connect the
first optic ganglion with the second, the second with the third, and the
third with thé fourth. From the fourth optic ganglion, which is sit-
uated near the proximal end of the stalk, the optic nerve extends to the
brain. The finer structure of the optic ganglia in Palaamonetes is in all
probability essentially the same as in Astacus, where, as I have already
shown (Parker, '95, Taf. 3, Fig. 59), each optic ganglion represents a
region of interruption for the great majority of the nerve fibres that
intervene between the retina and the brain.

The retina in Palemonetes is composed of ommatidia, the structure of
which has already been described at length (Parker, '91, p. 108, Pl. IX.).
For convenience I add a brief summary of this description. Each omma-
tidium is composed of five kinds of cells. Immediately under the cor-
neal facet (Fig. 1, ern.) are two corneal hypodermal cells (nl. ern.). The
distal portion of tho axis of the ommatidium is occupied by the cone (con.),
which, as seen in transverse sections (Fig. 3, cl. con.), is composed of four
parts. Each part contains near its distal end a nucleus (Fig. 1, nl. con.)
and represents a cell The four cone cells are closely applied to one an-
other in the region of the cone proper (Fig. 3). Proximally they taper
off as thick, more or less independent fibres. (Compare Figs. 1, 4, and 5.)
These fibres separate and apparently terminate near the distal end of
the rhabdome (Fig. I, r4b.). I have been unable to trace them further,
though I suspected that they might end, as in Homarus (Parker, '90,
p. 14), on the basement membrane. The distal retinular cells either
apply themselves to the lower portion of the sides of the cone (Fig. 2,
cl. dst.), in which case they are so closely packed that their outlines
cannot be distinguished (Fig. 3, el. dst.), or they oceupy a more prox-
imal position (Fig. 1, el. dst.), forming a ring around the attenuated ends
of the cone cells (Fig. 5). There is, of course, one ring for each om-
matidium. Each ring contains six distal retinular cells, but these rings
are so constituted that each cell is at the same time a member of three
rings; hence there are in reality only twice as many distal retinular
cells as there are ommatidia. The proximal portion of the axis of the
ommatidium is occupied by the rhabdome (Fig. 1, /.), which is sur-
rounded by seven functional proximal retinular cells (Wig. 6, el. px.),
in addition to which an eighth rudimentary one is present (Parker, ’91,
p. 111). Each funotional cell ends distally in a somewbat swollen knob
containing its nucleus (Fig. 1, nd. px.). From this swollen end the
cell extends proximally over the rhabdome, beyond whieh it becomes
slightly attenuated, and, as a retinal nerve fibre (Figs. 1 and 7, for. r.),

278 BULLETIN: MUSEUM OF COMPARATIVE 790LOGY.

pierces the basement membrane (mb. ba.) and extends to the first optic
ganglion. Here it probably terminates in a fibrillation, as has already
been shown to be the case in Astacus (Parker, ’95, p. 41). The acces-
sory pigment cells (Fig. 1, el. sn.) occupy the space in the deeper part
of the retina. The number of these cells is not constant, but, judging
from their nuclei, it is not more than one or two for each ommatidium.
Proximal processes extend from these cells through the apertures in the
basement membrane to the distal surface of the first optic ganglion, and
distal processes may extend forward: to the front faces of the distal
retinular cells. Each ommatidium in Palsemonetes, then, is composed
of the following cells: two corneal hypodermal cells, four cone cells, two
distal retinular cells, eight proximal retinular cells (one of which is
rudimentary), and a variable but small number of accessory pigment
cells. Black pigment granules are contained in both the distal and the
proximal retinular cells, and are limited to these cells ; the whitish pig-
ment lies exclusively in the accessory pigment cells. The seven func-
tional proximal retinular cells are the only elements of the ommatidium
that are known to have nervous connections. These brief anatomical
statements may suffice as an introduction to the consideration of the
pigment changes in the retina.

PHOTOMECHANICAL CHANGES IN NORMAL RETINA.

The general method by which the normal photomechanical action of
the retinal pigment cells in Palamonetes was determined consisted in
the examination of eyes that had been kept in the light or in the dark
known periods of time. For a dark chamber T used a box with a tight-
fitting cover. From time to time during the course of the experiments
this box was tested for its light-proof qualities by exposing in it a very
sensitive bromide paper, such as is used by photographers. In all my
experiments this showed complete absence of light. The top of the
box was pierced by a hole, through which a piece of rubber tubing was
introduced so that fluids could be poured into the box without exposing
its contents to light. Two or three turns in this tube were found suffi-
cient to prevent such light as entered the outer end of the tube from
reaching the interior of the box. Living shrimps in a vessel of water
were placed in the box, and the cover was carefully closed. After the
expiration of the required interval, hot water was run in through the
tube, and the animals were thus killed in the dark. Other killing re-

agents, such as corrosive sublimato, picric acid, ete., were tried, but

PARKER: RETINAL PIGMENT CELLS OF PALAIMONETES. 279
none proved so satisfactory as water at about 80° C. The periods of
exposure to dark in the first set of experiments were as follows: 1 min.,
5 min., 10 min., 15 min., 30 min., 45 min., 60 min., and then at intervals
of an hour up to 8 hours. It was found subsequently that the experi-
ments need not have extended over a maximum period of more than
two hours, and that intervals of about fifteen minutes were all that
were needed to observe the steps of the change. From each lot of ani-
mals prepared in this way, the optic stalks were out into sections for
examination under the microscope. In a similar way, the eyes of ani-
mals that had been kept some four hours in the dark were exposed to
the light for given intervals, killed, cut, and examined. In cases where
it was necessary to make very accurate comparisons, the eyes of the
same animal were used for the two conditions; thus, after keeping
the animal a given time in the light, one optic stalk was removed, and the
animal kept in the dark. At the expiration of the second interval, the
second optie stalk was removed and prepared. To guard against indi-
vidual varietions, in every experiment the eyes of at least, three animals
were examined.

The only general changes shown by retinas subjected to light or dark
were changes in the arrangement of the pigment. In other respects
they were not noticeably altered. Thus, no change in thickness was
observable; in one case, a left retina that had been kept in the dark
measured in its middle region from the corneal cuticula to the basement
membrane 263 % while the right retina from the same animal exposed
to light measured 270 u. In à second ease, a dark left retina measured
240 % tho light right one measuring 233 u. The cones likewise showed
ao significant differences. By analogy with the perceptive elements in
the vertebrate eye, one might have expected the rhabdomes, the termi-
nal nervous organs of the crustacean eye, to shorten in the light and
lengthen in-the dark. I was unable to obtain evidence of such a change
in Palemonetes, and yet the conditions for the exact measurement of
the rhabdomes are so unfavorable in this animal that I am by no means
certain that these changes may not occur If, however, they do take
place, they must be relatively small. The observable changes induced
in the retina by the absence or presence of light affect the three kinds
of pigment cells, — the proximal retinular cells, the accessory cells, and
the distal retinular cells. These will be considered in the order given.

The pigment in the proximal retinular cells forms at the base of the
retina a band, called by Exner (91, p. 62) the retinal pigment. The
photomechanical changes that this pigment undergoes have already been

280 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

observed in various crustaceans by Exner (91), Szczawinska (91), and
myself (Parker, 95); and, so far as the chief facts of these changes
are concerned, the accounts given by these writers are in substantial
agreement. In no case, however, has the precise character of these
changes been followed, nor the time needed for their completion been
recorded,

In a proximal retinular cell that shows the full effect of light (Fig. 1),
the black pigment granules are almost uniformly scattered from the
distal end of the cell backward through its whole length, including the
retinal nerve fibre, to the region of the first optic ganglion. In the body
of the cell proper (Fig. 6), as well as in the retinal nerve fibre (Fig. 7),
it will be observed that the pigment granules lie entirely within the
limits of these structures; in other words, the black pigment of this
portion of the retina is contained entirely within the proximal retinular
cells. "This pigment, though in the main uniformly distributed through
the cell, shows regularly two slight concentrations, — one at the swollen
distal end of the cell (Fig. 1), and another on the sides of the rhabdome.
Small irregular concentrations may also occur in the body of the cell.
In the eye subjocted to light, the only part of the cell except the nu-
cleus that is entirely free from pigment granules is a transparent axis
that can be traced from the region of the rhabdome down through the
‘body of the cell, and through the whole length of the retinal nerve
fibre. This is undoubtedly the axis cylinder of the nerve fibre, which, in
its passage to the rhabdome, extends through the body of the cell.

In erustaceans like Cancer (Parker, '91, p. 116, Plate X. Fig. 131),
in which the proximal retinular cells are more fully provided with pig-
ment granules than in Palemonetes, this axis is more conspicuous.

In an eye that has been kept in the dark for several hours, the bodies
of the proximal retinular cells are without trace of pigment (Fig. 2),
the whole mass of black pigment being concentrated in the retinal nerve
fibres, i. e. proximal to the basement membrane. Here, as in the for-
mor case, the pigment lies entirely within the limits of the retiuular
cell.

The transition from the dark condition to the light condition of the
eye was accomplished by the following steps. In an eye that had been
kept some four hours in the dark and then exposed for five minutes to
the light, the arrangement of the pigment in the proximal retinular
cells was indistinguishable from that characteristic for full darkness.
After ten minutes’ exposure to light, the pigment was found to have
moved forwards to the level of the basement membrane. After fifteen

PARKER: RETINAL PIGMENT CELLS OF PALZEMONETES. 281

minutes, it was found throughout the bodies of the cells; and, at thirty
minutes, well marked concentrations had appeared about the rhabdome
and at the distal end of each cell. At forty-five minutes, these concen-
trations were somewhat more pronounced, but after that time no further
changes were observable.

The revorse change, which takes place in the dark, is accomplished in
the following manner. After the animal has been in the dark fifteen
minutes, the concentrations of pigment about the rhabdomes and at
the distal ends of the cells have almost disappeared, though the bodies
of the cells still contain an almost uniform amount of pigment through-
out their whole length, After thirty minutes, much more pigment is to
be found proximal to the basement membrane than distal to it, and after
forty-five minutes almost all the pigment is proximal in position. At the
end of an hour, the condition characteristic of darkness is fully realized.

The changes just recorded occur entirely within the limits of each
proximal retinular cell. There is no reason for believing that the
changes are the results of a process of pigment production in one part
of the cell, and of pigment destruction in another. The observed facts,
on the contrary, suggest that the pigment granules of one region in the
cell are moved to another. The movement, however, is not accompanied
by any noticeable change in the position or even the form of the contain-
ing cell. The pigment granules seem to be carried up and down through
the cell as though by a streaming of the cell protoplasm. A similar
stability of form, accompanied with an internal movement of pigment,
has been described by Ballowitz ('93, Taf. XXXVI. Fig. 12, and '93*,
p. 629) in the pigment cells of the skin of fishes.

Through the kindness of Professor F. H. Herrick, I have had the priv-
ilege of examining an interesting series of eyes taken from specimens of
Palwemonetes that had been kept living in a dark chamber thirty-eight
days. The pigment in the proximal retinular cells of such animals
showed the condition characteristic of the dark. In an animal that had
been kept in the dark for this period and then exposed to light for four
hours and three quarters, the pigment returned partially to the position
characteristic of the light. The greater part of it remained proximal to
the basement membrane, and from that which moved into the bodies of
the cells no marked concentrations were formed, either about the rhab-
domes or at the distal ends of the cells. Long confinement in the dark,
then, seems to interfere somewhat with the mechanism by which the
pigment of these cells is normally moved.

The accessory pigment cells are located in the base of the retina, and

282 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

send a few processes distad to the outer surfaces of the dista] retinular
cells, and many proximad through the apertures in the basement, mem-
brane to the distal surface of the first optic ganglion. The pigment
with which these cells seem to be almost entirely filled is yellowish by
transmitted light, and white by reflected light. It is especially remark-
able for its powers of reflecting light, and this quality led Exner (91,
p. 97) to designate the layer formed from it by the name of the tape-
tum. Whether this pigment is influenced by.the presence or absence
of light is a matter of some uncertainty. Szezawinska (91, p. 552)
states that in Astacus, under the influence of light, the cells containing
it enlarge slightly. Exner (91, p. 105), though at first inclined to re-
gard the accessory cells as influenced by the light, was finally led to
abandon this view, and to explain their two apparent conditions by the
greater or less degree with which they were covered by the migrating
pigment of the proximal retinular cells. Ina preparation from an eye
kept in the dark, the retinular pigment, as already mentioned, is entirely
below the basement membrane, and the accessory pigment is almost en-
tirely exposed, and consequently conspicuous. In the light it is some-
what covered by the black pigment, which under these circumstances
fills the bodies of the proximal retinular cells, and it thus becomes less
noticeable than before. My own studies on the retina of Astacus (Par-
ker '95, p. 25) led me to agree with Exner that the accessory pigment
showed only an apparent change. If, however, any change did occur, it
was certainly not an increase in the size (conspicuousness ?) of the acces-
sory cells under the action of light, as maintained by Szezawinska, but
rather the reverse.

Although in respect to Astacus I am still in doubt as to whether or not
the accessory cells show any photomechanical changes, I have not the
least hesitancy in stating that in Palemonetes such a change does occur.
The principal difficulty in demonstrating this change comes from the dis-
turbing influence produced by the migration of the pigment in the proxi-
mal retinular cells. This difficulty, however, can be overcome by the
employment of a depigmenting reagent that will remove the retinular
pigment without affecting the accessory pigment, Such a reagent is
the depigmenting fluid recommended by Grenacher (’86, p. 214). In
preparations representing the dark and the light condition, and depig-
mented by this means, the differences in the distribution of the pigment
in the accessory pigment cells is so striking that no one would question
for a moment the photomechanical activities of these cells. In the light
(Fig. 1) the accessory pigment forms two concentrations, one in the base of

PARKER: RETINAL PIGMENT CELLS OF PALJEMONETES. 283

the retina, and the other near the distal surface of the first optie ganglion.
These two concentrations are connected by irregular bands of pigment.
In the dark (Fig. 2) almost all the accessory pigment is in the base of
the retina, the concentration near the ganglion as well as the interme-
diate pigmented bands being represented by only a few small pigmented
patches.

The change from the condition produced by the light to that produced
in the dark is indicated in the following steps. After the animal has
been about thirty minutes in the dark, the concentration of pigment for-
merly near the optic ganglion is appreciably nearer the retina, After
forty-five minutes, this concentration as such has disappeared, and that
in the retina has considerably increased. Finally, after two hours, almost
all the accessory pigment lies in the base of the retina, there being only
a few small strands proximal to the baserhent membrane.

In the reverse change under the influence of light, the intermediate
pigment strands show a perceptible thickening between ten and fifteen
minutes after the eye has been placed in the light, and the full concen-
tration at the level of the ganglion is completed within the period extend-
ing from forty-five minutes to an hour after that event.

I have never been able to discover any outlines to the accessory pig-
ment cells except those indicated by the pigment mass itself. Judging
from these, the photomechanieal changes in the accessory cells involve so
radical an alteration in the forms of the cells that the latter may be said
to have assumed a different position. In this respect, then, the pigment
changes in these cells involve much more active movements than in the
case of the proximal retinular cells, and possess something of a locomotor
character. So far as I have observed them, they may be compared with
perfect propriety to the more or less circumscribed movements of an
amoba. When the retina is placed in the light, the cells with their
contained pigment creep slowly backward through the apertures in the
basement membrane toward the optic ganglion. When the retina is in
the dark, they reverse this movement and creep out into the base of the
retina. The one particular in which this movement differs from that of
an amooba is that of its limitations in direction. Thus the cells always
creep either outward or inward. Moreover, in darkness they do not
creep indefinitely outward, but after about two hours reach a maximum
limit ; the same is true of their inward course. These limitations may
be due either to the structure of the regions into which the cells creep,
or to the intrinsic qualities of the cells themselves; but I have been
unable to get conclusive evidence as to which it is.

284 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY,

In the interesting series of eyes of Palzemonetes loaned me by Professor
Herrick, the accessory pigment cells of the eyes that had been kept in
the dark thirty-eight days presented a condition normal for exposure to
the dark. In those eyes that had afterwards been exposed to the light
for four hours and three quarters, this pigment had apparently resumed
the position normal for exposure to light. The mechanism by which the
accessory pigment changes are brought about, unlike that for the proxi-
mal retinular pigment changes, is therefore apparently not interfered
with by prolonged retention in the dark.

The distal retinular cells present photomechanical changes more com-
plex than those in the two kinds of cells already considered. These
changes have been described by Exner (’89 and '91), Szezawinska (’91),
Herrick (91), and myself (Parker, 95). All investigators are agreed,
I believe, in stating that in the dark these cells occupy a more distal
position than in the light. Their probable influence on the amount of
effective light that enters the retina led Exner (91, p. 63) to call them
the iris pigment. In Palaemonetes, as I have already shown, there are
two distal retinular cells for each ommatidium.

In an animal that has been subjected to the full action of light, the
distal retinular cells (Fig. 1, c/. dst.) are plump ovoid bodies in contact
with the outer ends of the proximal retinular cells. The body of each
distal cell has the length of about 30,4. From its outer end a single
process usually extends to, or at least toward, the corneal hypodermis.
The whole distal retinular cell, excepting its nucleus and sometimes a
portion of its distal process, is filled with black pigment. The whitish
pigment that often occurs on the outer surface of these cells repre-
sents, as already mentioned, a distal process from the accessory pigment
cells.

In animals kept a sufficient time in the dark, the bodies of the distal
retinular cells (Figs. 2 and 8, ed. dst.) are flattened, and applied to
the sides of the cones. They measure about 70 u in length and pos-
sess, in addition to their distal processes, shorter proximal ones, which
extend backward to the outer ends of the proximal retinular cells. As
before, the cytoplasm is largely filled with black pigment granules,
which, however, are often more concentrated in the body of the cell
than elsewhere.

It must be obvious from this brief description that in considering the
photomechanieal changes of the distal retinular cells two factors are to
be kept distinct: first, the lengthening and the shortening of the cell
body, and, secondly, the distal and the proximal migration of the cell as

PARKER: RETINAL PIGMENT CELLS OF PALZEMONETES. 285

a whole. That these two elements are distinct can be seen from the fact
that in certain insects they are related to each other in a way just the
reverse of that which occurs in Palemonetes; thus, in Lasiocampa, as
figured by Exner (91, Taf. IV. Figs. 28, 29), the distal cells are short
in their distal position, and long in their proximal one.

The average length of the bodies of the distal retinular cells from the
left eye of a given animal prepared in light was about 30 h. The aver-
age length of the corresponding cells from the right eye of the same
animal prepared in the dark was about 70 u. In a series of preparations
taken from animals of approximately the same size as that just described,
the lengthening of the distal cells in the dark took place at a rate indi-
cated in the first of the following tables.

In a second series the shortening of the cells under the influence of
light was shown to take place as indicated in the second table.

The distal and the proximal migration of the cells are difficult to define,
because they are accompanied by the lengthening and the shortening of
the cells. Taking the nucleus as a fixed point in the cell, the maximum
distance of migration is about 50 fl. In the migration from the proxi-
mal to the distal extreme, made in the dark, the cell traverses this
distance in about two hours. The migration in the reverse direction,
under the influence of light, is completed in about one hour and three
quarters.

As the cells move outward, their distal processes shorten and their
proximal ones form and lengthen. As they move inward, their distal
ones elongate and their proximal ones shorten and finally disappear.
'The rates of these changes, as wellas of those given in the preceding
paragraphs, are indicated in the following tables of summaries, in which
the varying lengths of the parts of the cells are given for successive
periods.

Migration in the Distal Direction (in the Dark).

Time. Ohr. I hr. | hr. | hr. 1 hr. 14 hr. 13 hr. 14 hr.|2 hr.
Length of distal process . | 1301| 125 | 110 | 95 | 85 | 65 | 56 40 | 30
Length of cell body . . . 80 | 85 | 50 | 55 | 60 | 70 | 70 | 70 | 70
Length of proximal process 0 0 0 10] 15 | 25 | 35 | 50 | 60

1 All measurements of length are expressed in mikra (thousandths of a milli-
meter).

286 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

Migration in the Proximal Direction (in the Light).

Time. Ohr. | 4 hr. | hr. | $ hr. 1 hr. | 14 hr. 14 hr.| 13 hr.| 2 hr.
Length of distal process . | 30 | 85 | 55 | 80 | 100 | 115 | 125 | 130 | 180
Length of cellbody . . . | 70 | 70 | 65 | 60 50| 40| 365 30, 30

Length of proximal process | 60 | 55 | 40 | 20 | 10 5 0 0 0

The changes induced in the distal retinular cells by the light are
completed, then, in a period between an hour and a halfand an hour
and three quarters long. The changes that take place in the dark require
for their completion from an hour and three quarters to two hours.

Rough estimates of the time necessary for the completion of these
changes in different arthropods have been made by various investigators.
Szezawinska (91, p. 552) states that in Astacus the condition charac-
teristic for the dark is reached in six hours, that for the light in two
hours. Exner (791, p. 70) states that in an insect, Lasiocampa, the
changes require about half an hour, and Kiesel (’94, p. 105) gives the
same time for Plusia. Herrick (01, p. 455) believes that in Palamo-
netes the changes are accomplished in about twenty-five minutes, an
estimate that I should regard as rather too low.

Exner (91, p. 70) has suggested that muscle fibres might be con-
cerned in the migration of the distal retinular cells, an idea that gains
some support from the fact that in the eyes of some insects structures
like muscle fibres have been seen and described. In the crustacean
retina, however, Exner was unable to find anything like muscles. At
first sight it might seem probable that what I have described as the
proximal and distal processes of the distal retinular cells might be mus-
cular in nature. But the facts that the proximal process disappears
entirely during the proximal migration of the cell, and that the distal
one seems never to be firmly attached near the periphery of the retina,
are opposed to this view. Moreover, in the distal process, which, on the
whole, is the more muscle-like of the two, I have been unable to discover
any evidence of transverse enlargement in the shortened condition, such
as a contracted muscle exhibits. The cell in its distal migration seems to
move over the fibre rather than to be drawn onward by a contraction of
the fibre. Further evidence against the muscular nature of the motor
mechanism of these cells is to be found in the rate at which the move-
ment takes place; 50 u in two hours is exceptionally slow for the action

PARKER: RETINAL PIGMENT CELLS OF PALZEMONETES. 287

of any kind of muscle. These observations have led me to conclude
that muscular action, as ordinarily understood, has nothing to do with
the migration of the distal retinular cells. Obviously, ciliary action is
in no way connected with the movements of these cells, and there is left
then only amoeboid movement as a means of explaining these changes.
Each distal cell might be compared to an amoeba, which in its migrations
outward and inward uses its processes to guide its general motion. The
rate and general character of the movement agree well with this expla-
nation. In one respect, however, there is disagreement. Herrick (91,
p. 455), in his account of the action of the distal cells in Palamonetes,
states that, on contracting, these cells fold together somewhat as a ribbon
might be folded transversely to its length (cf. Fig. 10), and he believes
that, on expanding, they unfold again. "This condition is one not easily
reconciled with amoeboid movement.

Through the kindness of Professor Herrick I have had the privilege
of studying his preparations, and I can confirm his statement that in the
contracted condition (Figs. 9 and 10) the cells exhibit a series of trans-
verse folds, which are entirely absent from the expanded form. These
folds, however, occur, so far as I am aware, only in eyes which have been
kept an exceptionally long time in the dark (thirty-eight days in the
case of Professor Herrick's specimens), and are then exposed to the light.
In my own preparations, none of which had been kept in the dark more
than twelve hours, no trace of such folding could be discovered, and I
have therefore been led to regard these folds as abnormalities induced
by protracted retention in the dark. Notwithstanding this interpreta-
tion of the folds, they throw important light, I believe, upon the normal
action of the distal retinular cells.

The exact form of these folds is not so simple as might at first be
supposed. The body of each cell in its contracted condition consists of
an elongated thickened axial portion and two later»! wing-like expan-
sions, each of which terminates in a rather sharp edge. In other
words, these distal cells, when contracted, instead of assuming the
usual ovoid form, retain more or less the shape that they had when
expanded (Fig. 8). In a longitudinal section through the axial portion
of the body of one of these cells (Fig. 9) slight folds are observable.
In similar sections through the edge of the lateral wings (Fig. 10) the
folding is seen to be much more pronounced. The folds are most con-
spicuous at the edges of the wings, and lose in prominence toward
the axial part of the cell. Another peculiarity of these folded cells,
as compared with those kept a shorter time in the dark, is that

288 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

they shorten only to about five sevenths of their original length in-
stead of to three sevenths. Thus their long retention in the dark
seems to have prevented a return to the more completely contracted
condition normal for the light. The fact that the more peripheral parts
of the cells in the contracted condition are the more wrinkled indicates
that the axial part has retained its contractile nature more completely
than the peripheral parts, and suggests the idea that, since this axis
contracts in a definite direction, it must possess something of the nature
of a muscular core. It seems to me probable that, whilst the periphery
of these cells may be characterized by amoeboid movements, the core
acts in a more circumscribed way, much as a muscle would. If this is
so, the distal and proximal migrations, as well as the expansion of the
cell body, are probably manifestations of its amaboid movements,
while its shortening is probably due in the main to the muscle-like con-
traction of its central core. Objection might be raised to this com-
bination of different modes of motion, were it not generally admitted
that muscular action is, after all, only a more circumscribed form of
amoeboid movement.

The presence of a contractile axis in the distal retinular cells is further
rendered probable by the fact that in Mysis (Parker, '91, p. 120) an
axial core free from pigment has been observed in each distal retinular
cell. At the time I first noticed these cores I suspected that they might
be the remains of nervous axes, but I now believe there are stronger
reasons for suspecting them of being contractile bodies.

The following table gives by way of summary the periods required
for the completion of the various photomechanical changes in the retina
of Palemonetes.

From dark to light. From light to dark.

Proximalretinularcells . . . . . jhr.to $hr. 4 hr. to 1 hr.
Distal retinular cells 1} hr. to 13 hr. 1} hr. to 2 hr.
Accessory pigment cells 4 hr. to 1 hr. 14 hr. to 2 hr.

It is a noteworthy fact, that of these changes those that take place
in the light (positive stimulus) are always accomplished more rapidly
than the corresponding reversals in the dark. To this statement an
apparent exception may be found in the tables on pages 285 and 286, in
which are recorded the periods for the contraction and expansion of the

PARKER: RETINAL PIGMENT CELLS OF PALAMONETES. 289

distal retinular cells. The body of the distal cell contracts in the light
in an interval between 90 and 105 minutes, and expands in the dark
in between 60 and 75 minutes, thus apparently accomplishing a change
more rapidly in the dark than in the light. The expansion and con-
traction are, however, not simple operations, but are complicated by the
simultaneous production or absorption of the large proximal processes,
and itis possible that the discrepancy just pointed out is to be accounted
for by this complication.

Before leaving this subject I wish to call attention to the compara-
tive slowness with which all the photomechanical changes of the retina,
but particularly those of the distal retinular cells, take place. Exner
has shown that the amount of effective light that enters the eye is, in
all probability, largely controlled by the action of the distal cells, and
has therefore called them the iris pigment. The slowness with which
they respond, however, shows clearly that in their action they have
little resemblance to the iris of the vertebrate eye, and that their changes
correspond only to the more general changes in the amount of light in
their surroundings. The name iris pigment seems to me, therefore,
somewhat misleading, and hence I prefer to retain the name of distal
retinular cells, which indicates at once the present position and the
probable origin of these cells, namely, from cells that once formed a
part of the retinula itself (Parker, 95, p. 64).

SYMPATHETIO PHOTOMECHANICAL JHANGES.

To ascertain whether the retinas in the two eyes of Palwemonetes
were sympathetic toward each other in the same sense that Engelmann
believed the retinas in the eyes of vertebrates were, I carried out two
sots of experiments, in both of which animals were so placed that one
eye was in the dark while the other was exposed to the light. After a
sufficient, period both eyes were prepared and examined. The two sets
of experiments differed only in that I used different means to accom-
plish the exposure. In one set I tied a living shrimp to the inside of a
light-proof box, in which a small hole was made so as to allow one optic
stalk of the animal to project into the lighted exterior. Jaro was taken
that the small space between the optic stalk and the edge of the hole
should be filled with an opaque material (a mixture of thick Janada
balsam and lampblack). After several hours the animal was killed, and
its eyes prepared. In the other set of experiments one optie stalk of a
living animal was covered with a considerable quantity of the mixture

VOL. XXX. — NO. 6. 2

290 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

of balsam and lampblack, and, after allowing the animal to swim in a
brightly illuminated dish for several hours, it was killed, and both its
eyes prepared and examined.

The two sets of experiments yielded essentially the same results,
namely, the eyes exposed to the light always presented the condition
normal for the light, and those kept in the dark always showed an ap-
proach, more or less incomplete, to the condition characteristio for the
dark. This incompleteness might be taken as evidence of a partial
sympathetic relation between the two retinas; but I believe it is to be
explained otherwise. In both sets of experiments the eyes supposed to
be blinded were in reality only partially cut off from the light. In the
experiment with the light-proof box, I know by actual observation that
more or less light made its way through the optic stalk that projected out-
ward to the exterior, and thus gained access to the interior of the box.
If this is true of the experiment with the box, it is very probable that in
the second experiment light passed up through the base of the blinded
stalk, and thus reached at least the proximal part of the retina.

These experiments, then, are not wholly conclusive, but, so far as
they go, indicate considerable independence in the relations of the
two optie stalks. For reasons to be given later, in connection with the
experiments on excised stalks, I believe I am justified in concluding that
the two retinas are, in reality, wholly independent of each other.

Locauızen PHOTOMECHANICAL CHANGES.

Another question that naturally presents itself is, whether different
parts of the same retina are sympathetic toward one another, or whether
they are entirely independent, i. e. whether or not a retina responds
locally to stimulus.

To test this matter, I put minute drops of the mixture of balsam and
lampblack on the corneal cuticula of the eyes of several shrimps, and
let them swim for a few hours in well illuminated basins. On examin-
ing sections of their eyes later, it was found that under each mass of
applied pigment the retinal cells showed a condition characteristic for the
dark. This was most pronounced in the distal retinular cells, but was
also observable in the proximal retinular cells, as well as in the acces-
sory pigment cells. This experiment shows beyond a doubt that the
elements of the retina act locally, and respond to differences of light
and dark independently of one another. "This independence furthermore
explains what is not infrequently seen in sections of otherwise normal

PARKER: RETINAL PIGMENT CELLS OF PALZEMONETES. 291

eyes that have been kept in the dark, namely, occasional single proxi-
mal retinular cells which, instead of having their pigment granules
transported to the retinal fibres, still hold them in their bodies, Such
cells have probably suffered some pathological change by which their
individual photomechanical functions have been interfered with. This
independence in the action of parts of the retina has already been
affirmed by Exner (91, p. 66) for the compound eyes of insects.

PHOTOMECHANICAL CHANGES iN Exotsep EYES AND RETINAS.

The extent to which the photomechanical changes in the retina are
influenced by the central nervous organs has never been determined, I
believe, for any arthropod. That some such influence is exerted is im-
plied by several investigators ; thus Stefanowska (90, p. 156) states
that, in preparing insects’ eyes, she cut the heads of the animals in two
so as to prevent the nervous centres from affecting the retinal pigment
cells, and Szczawinska ('91, p. 531) recommends as a fixing reagent a
hot solution of eorrosive sublimate, because the action is so rapid that
it is not necessary to use other means of intercepting the central ner-
vous influences. This belief, that the central nervous organs can exert
an influence on the retinal pigment cells, is not to my knowledge the
result of direct oxperiment, but is the application to other groups of
animals of a generalization firs& made by Engelmann and his followers for
vertebrates. As is well known, Engelmann showed that, when one eye
of a frog was protected from the light, the illumination of the other
eye, or even of a portion of the surface of the body, sufficed to produce
in the pigment cells of the protected eyo a condition characteristic for
the light. This observation naturally led to the conclusion that the
pigment colls of the retina were controlled in their movements by the
central nervous organs, and that the optic nerve transmitted impulses
centrifugally as well as centripetally., Fick (95, pp. 77 and 81), how-
ever, has recently demonstrated that the same changes occur in a frog’s
eye even after the optic nerve and sympathetic nerves have been cut,
and that therefore the central nervous organs take no part in these
changes.

Before turning to the experimental evidence obtained from Palemone-
tes, it will be well to consider some of the consequences of this question.
In order that the central nervous organs should have any influence on
the retinal pigment cells, the two sets of structures must be in nervous
connection. So far as is known, the only structures in the retina of

292 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Palsemonetes that have nervous connections are the proximal retinular
cells, the accessory cells and the distal cells not being supplied with
nerves. Since photomechanical changes occur in both the accessory and
the distal cells, the inference might be drawn that in these instances the
changes were necessarily independent of the central nervous organs.
But it might also be argued that these very changes indicate nervous
connections that have escaped the eye of the anatomist. To this it
might be replied that, as the nervous connections of the proximal cells
are so very obvious, it is highly improbable that the distal and accessory
cells have a hitherto undescribed nerve supply. So far, then, as the
purely anatomical relations are concerned, they indicate. that the photo-
mechanical changes, in the accessory and distal cells at least, are inde-
pendent of central nervous influences.

In the case of the proximal retinular cells, where each cell possesses a
single nerve fibre, the central nervous organs might control the pigment
changes. However, if they do, the retirial fibres afford, so far as I know,
the first good instance of normal double conduction, Since each retinal
nerve fibre is the one nervous process from some proximal retinular cell,
and since all these cells show photomechanical changes, it follows that,
if these changes are controlled by the central nervous organs, all retinal
fibres must transmit central impulses peripherad. As these same fibres
are the only nervous connections between the retina and the central ner-
vous organs, some at least must also transmit retinal impulses centrad.
Therefore, if it can be shown that the central nervous organs influence
the photomechanical changes in the proximal retinular cells, it is like-
wise demonstrated that double conduction is a natural occurrence. As
Fick (95, p. 73) justly remarks, the solution of this problem involves
one of the most fundamental principles concerning the transmission of
nervous impulses.

The method by which I proceeded to test this matter in Paleemonetes
consisted in examining the changes that went on in eyes after their con-
nection with the central nervous organs had been severed. These con-
nections were conveniently cut in one of two places; either the whole
optic stalk was excised, in which case the optic nerve was cut between
the brain and the optic ganglia, leaving the latter in normal connection
with the retina, or the retinal end of the stalk was cut off, thus separat-
ing the retina in the region of the retinal nerve fibres from the optic
ganglia as well as from the brain.

To ascertain whether the brain had any influence over the retinal pig-
ment, the following experiments were tried. Four live shrimps, whose

PARKER: RETINAL PIGMENT CELLS OF PALZEMONETES. 293

eyes were in the condition characteristic for the dark, were decapitated ;
their right optic stalks were cut off and put in a light moist chamber,
and their left stalks, likewise cut off, were placed, as a check on the
results of the experiment, in a dark moist chamber. After an interval
of two hours both sets of stalks were hardened and afterwards cut and
examined. In a corresponding way optic stalks in the condition charac-
teristic for the light were cut off and subjected to the dark.
The results of these experiments are shown in the following tables.

Four Right Optio Stalks in Dark Condition cut off and placed in the
Light two Hours.

Complete change. | Partial change. No change.
Proximal retinular cells . . . 0 3 1
Distal retinular cells . 0 3 1
Accessory cells 0 2 2

The four left optic stalks cut from the same animals and retained, as
a check, in a dark chamber, all presented on examination the condition
typical for the dark.

Four Right Optic Stalks in Light Condition cut off and placed in the
Dark two Hours.

Complete change. Partial change. No change.
Proximalretinular cells . . . 1 3 0
Distal retinular cells . . . . 1 3 0
Accessory cells 1 1 2

The four left optic stalks cut from the same animals and retained, as
a check, in a light chamber, all presented on examination the condition
typical for the light.

It is obvious from these observations that, after the excision of an
optio stalk, the photomechanical changes may still take place, if not com-
pletely, at least partially, and it might be inferred from this that the
brain exerted at least a partial influence over these changes. This con-
clusion, however, is invalidated by the fact that in one case recorded in the

294 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

second table the photomechanical changes were carried out completely, thus
demonstrating that the brain is not in any way essential to these changes.
Why this completeness was not seen in other cases I am unable to state
positively, though I believe it was owing to the changes that gradually
appear in the tissue of the stalk after its severance from the body of the
animal. When an optic stalk is excised, the blood in it soon coagulates
and other alterations doubtless start up, which finally result in the com-
plete death of the tissues of the stalk. It is these alterations, I believe,
that overtake and bring to a standstill the slowly progressing photome-
chanical movements. But, whatever may be the true explanation of the
incompleteness of the changes in excised stalks, the general conclusion
remains unaffected, that in Palaemonetes the brain is not essential to the
photomechanical changes in the retina.

This conclusion has an important bearing on the question of the sym-
pathetie relations of the two retinas in a given animal. Since the two
retinas are nervously connected only through the brain, and since the
retinas are not influenced from the brain, it follows that the two retinas
cannot be sympathetically related, a conclusion to which observations
already recorded have likewise pointed.

If the photomechanical changes are not dependent in any degree on
the brain, it may still be asked whether they are not influenced by the
optic ganglia. To answer this question, I carried out on excised retinas
a series of experiments similar to those just described for the optic
stalks. It is much more difficult to separate the retina from the optic
ganglia than it is to separate the optic stalk from the brain, but with
careful manipulation it can be done, and the following tables give the
results of experiments carried out upon such retinas.

Four Right Retinas in Dark Condition cut off and placed in the
Light about two Hours.

Cqmplete change. Partial change. No change.

Proximal retinular cells. 0 4 0
Distal retinular cells . . . - 0 3 1
Accessory cells 0 4 0

The four left retinas kept in the dark as checks on the experiment
exhibited the normal condition for the dark.

PARKER: RETINAL PIGMENT CELLS OF PALASMONETES. 295

Four Right Retinas in Light Condition cut off and placed in the
Dark about two Hours.

Complete change. Partial change. No change.
Proximal retinular cells 0 2 2
Distal retinular cells 0 4 0
Accessory cells . . . . .. 0 3 1

The four right retinas kept in the light as checks on the experiment
exhibited the normal condition for the light.

Although, as the tables show, no case of excised retina with complete
photomechanical changes has been observed, several of the cases were
so nearly complete that I have no hesitancy in stating that, in my
opinion, the photomechanical changes in the retina are as little influ-
enced'by the optic ganglia as by the brain.

These experiments, then, lead to two conclusions : first, the brain of
Palæmonetes is not essential to the complete photomechanical changes
of the retinal pigment cells; and, secondly, the optic ganglia are like-
wise unessential to these changes. In the latter case, however, the pos-
sibility of a slight influence must be admitted. The photomechanical
changes of the retinal pigment cells are, in my opinion, induced by the
direct influence of the presence or absence of light on these cells. Each
cell, then, so far as its mode of action is concerned, is not comparable
to a muscle controlled by an efferent nerve, but to a more or less inde-
pendent organism, which receives a direct stimulus from the exterior,
and responds appropriately. The uniformity usually shown by the
photomechanical movements in the retina as a whole is to be under-
stood as an individual but uniform reaction of many separate elements
wo a uniform stimulus. There is nothing in the action of the retinal
pigment cells of Palemonetes that supports the idea of normal double
conduction of nervous impulses

GENERAL SUMMARY.
1. The only parts of the retina in Palemonetes that exhibit photo-

mechanical changes are the three kinds of pigment cells.

2. The proximal retinular cells contain black pigment granules. Jn
the light these are scattered more or less uniformly throughout the whole

296 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

length of the cell, including the retinal nerve fibre. There are slight
concentrations of pigment at the distal end of the cell and around the
rhabdome. In the dark the pigment is limited to the retinal nerve
fibres.

3. The change from the dark condition to the light one is accom-
plished in from 30 to 45 minutes. The reverse change requires 45 to
60 minutes.

4. These changes are probably due to internal protoplasmie move-

ments, by which the pigment granules in the cells are moved in one

or other direction.

5. The accesssory pigment cells contain a yellowish white pigment.
In the light this is massed partly in the base of the retina, and partly
near the distal surface of the first optic ganglion. The two pigment
masses are connected by pigmented strands. Jn the dark the pigment is
almost entirely in the base of tho retina.

6. The change from the dark condition to the light one is accom-
plished in from 45 to 60 minutes; the reverse change, in from 105 to
120 minutes.

7. These changes are probably produced by amoboid movements of

the cells.

8. The distal retinular cells contain black pigment granules. Jn the
light they are contracted, and occupy a proximal position in the retina
f the ommatidium near the outer ends of the
In the dark they are expanded (flattened),
n in the retina, surrounding more or less

surrounding the axis o
proximal retinular cells.
and occupy a distal positio
completely the sides of the cone.

9, The change, from the dark condition to the light one is accom-
plished in from 90 to 105 minutes; the reverse change requires from
105 to 120 minutes.

10. These changes are produced in part by an amoeboid movement of
the cell, and probably in part by a muscle-like contraction of its axial

portion.

11. Each set of photomechanical changes carried out in the light is
completed in less time than the corresponding set of reverse changes
carried out in the dark.

12. The photomechanical condition of the retina in one eye has no
effect upon that in the othér eye; i. e. the retinas are not sympathetic.

PARKER: RETINAL PIGMENT CELLS OF PALZEMONETES. 297

13. The photomechanical action within the retina is localized, small
groups of pigment cells responding to local stimulation.

14. In excised eyes (optic nerve cut), complete photomechanical
changes may occur, thus proving that the brain is not essential to these
changes.

15. In excised retinas (retinal nerve fibres cut), nearly complete
photomechanical changes may occur, thus showing that the optic
ganglia are probably not essential to these changes.

16. The incompleteness of the changes in either the exeised eyes or
excised retinas is probably due to the death of the retinal tissues before
the photomechanical changes have been completed.

17. The three kinds of retinal pigment cells probably respond to
direct stimulation from without, and are not influenced by nervous im-
pulses from within. 'l'here is no good evidence in favor of normal double
conduction of nervous impulses.

Nor,

Since the preceding pages were written, Rosenstadt's (96) paper on
the structure of the compound eyes in Decapods has been published.
This contains a brief general account of the migration of the retinal
pigment in these crustaceans, and calls for a word of comment. In his
description of the directions of motion shown by the pigment under
various conditions of light, Rosenstadt agrees with Exner and later
investigators, but in his account of how this movement is accomplished
he stands entirely alone. His conception of the process can best be put
in his own words (Rosenstadt, 96, p. 759): “ Beim Uebergange des
Lichtauges in ein Dunkelauge gehen mit dem Pigmente volgende Verän-
derungen vor sich: Das Pigment tritt aus dem vorderen Ende. der
Retinulazellen [= proximal retinular cells] und wohl auch aus den
Retinapigmentzellen [= rudimentary retinular cells] aus. Dasselbe
wird von den Fortsätzen der Irispigmentzellen [— distal retinular cells]
aufgenommen, die, wie wir gesehen haben, mit dem im Vorderende der
Retinulae angesammelten Pigmente im Contact stehen. An diesen
Fortsätzen kriecht nun das Pigment hinauf; es findet eine Art Pig-
mentinfiltration der Irispigmentzellen statt. Gleichzeitig wandert das
Pigment nach hinten zu aus den Retinulazellen aus und gelangt hinter
die Membrana fenestrata [= basement membrane], wo es von den mit
Ausläufern versehenen Zellen aufgenommen wird.” This idea that the

298 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

pigment migrates from one cell to another is, so far as Rosenstadt’s ac-
count goes, entirely unsupported by direct evidence, and seems to me
an unwarranted assumption. The proximal movement of the pigment
from the distal end of the retinula to the opposite side of the basement
membrane is certainly accomplished within the limits of one set of cells,
for, a8 I have shown in this paper, the pigment even when entirely
proximal to the basement membrane lies in the thick retinal nerve
fibres, which are merely processes from the proximal retinular cells.
Although it cannot be stated with certainty that there is no exchange of
pigment between the distal and the proximal retinular cells in Palæ-
monetes, for in this crustacean in bright light these two kinds of cells
are closely applied to each other, it is perfectly certain that in other
decapods, as for instance Palemon, no such exchange is possible; for,
as Exner (’91, Taf. V. Fig. 51) has shown, and I can confirm his obser-
vations, the pigmented parts of the distal and the proximal retinular
cells never touch, even under full light. "These reasons, togother with
the facts set down in the present paper, confirm me in the belief that
Rosenstadt’s explanation of the migration of the pigment is erroneous,
and that the one presented in the foregoing account is correct.

PARKER: RETINAL PIGMENT CELLS OF PALZEMONETES. 299

PAPERS CITED.

Ballowitz, E.
'93. Die Nervenendigungen der Pigmentzellen. Zeit. wiss. Zool, Bd. LVI
pp- 671-706, Taf. XXXV.-XXXIX.

Ballowitz, E.
'93*. Ueber die Bewegungserscheinungen der Pirmentzellen. Biol. Centralb.,
Bd. XIII. pp. 625-630.

Exner, S.
89. Durch Licht bedingte Verschiebungen des Pigmentes im Insectenauge
und deren physiologische Bedeutung. Sitzungsb. Acad. Wiss. math.

naturw. Cl, Bd. XOVIII. Abt. III. pp. 143-151, 1 Taf.

Exner, S.

91. Die Physiologie der facettriten Augen von Krebsen und Insecten.

Deuticke, Leipzig und Wien. vi ＋ 206 pp., 7 Taf.
Fick, E. A.

95. Ueber die Frage, ob zwischen den Netzhauten eines Augenpaares ein
sympathischer Zusammenhang besteht. Vierteljahrsschr. naturf. Gesell.
Zürich, Jahrg. XL. pp. 71-83.

Grenacher, H.

86. Abhandlungen zur vergleichenden Anatomie des Auges. T. Die Retina
der Cephalopoden. Abhandl. naturf. Gesell. Halle, Bd. XVI. pp. 207-
256, 1 Taf.

Herrick, F. H.

91. Alpheus: A Study in the Development of Crustarea. Memoirs Nat.

Acad. Bci., Vol. V. pp. 370-401, 38 Pls.
Kiesel, A.
94. Untersuchungen zur Physiologie des facettirten Auges. Sitzungsb.

Acad. Wiss. math.-naturw. Cl., Bd. CIII. Abt. III. pp. 97-189, 1 Taf.

Parker, G. H.
90. The Histology and Development ot the Eye in the Lobster. Bull. Mus.
Comp. Zoól. Harvard Coll., Vol. XX. pp. 1-60, Pls. I.-IV.

300 BULLETIN: MUSEUM OF COMPARATIVE ZOÖLOGY.

Parker, G.H.
91. The Compound Eyes in Crustaceans. Bull. Mus. Comp. Zoöl. Harvard

Coll., Vol. XXI. pp. 45-140. Pls. 1.-X.
Parker, G. H.
'95. The Retina and Optic Ganglia in Decapods, especially in Astacus. Mitt.
Zool. Stat. Neapel, Bd. XII. pp. 1-73, Taf. 1-3.
Parker, G. H.
96. Pigment Migration in the Eyes of Palemonetes. A Preliminary Notice.
Zool. Anzeiger, Jahrg. XIX. pp. 281-284.
Rosenstadt, B.

'96. Beiträge zur Kenntniss des Baues der zusammengesetzten Augen bei den
Dekapoden. Arch. mikr. Anat., Bd. XLVII. pp. 748-770, Taf. XXXIX.,
XXXX.

Stefanowska, M.

'90. La Disposition histologique du Pigment dans les Yeux des Arthropodes
sous l'Influenoe de la Lumière directe et de l'Obscurité complete. Recueil
Zool. Suisse, Tome V. pp. 151-200, Pls. VIII., IX.

Szczawinska, W.

'91. Contribution à l'étude des yeux de quelques Crustacés et Recherches ex-

périmentales.sur les mouvements du pigment granuleux et des cellules
pigmentaires sous l'influence de la lumière et de l'obscurité dans les yeux
des Crustacés et des Arachnides. Arch. de Biol., Tome X. pp. 593-560,
Pls. XVL, XVII.

PARKER, — Retinal Pigment,

EXPLANATION OF PLATE.

All the figures were taken from preparations of the eyes of Palemonetes vul-
garis Stimp. They were drawn with the aid of an Abbé camera, and are all mag-
nified 335 diameters.

el, con.
cl. dst.
(EDO,
cl. sn.
con.
ern,
Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig: T
Fig. 8
Fig. 9.
Fig. 10.

ABBREVIATIONS.
Cone cell. br. r. Retinal nerve fibre.
Distal retinular cell. mb. ba. Basement membrane.
Proximal retinular cell. nl. con. Nucleus of cone cell.
Accessory pigment cell. nl. ern. Nucleus of corneal hypodermis cell.
Cone. nl.px. Nucleus of proximal retinular cell.
Corneal cuticula. rhb, Rhabdome.

Longitudinal section of an ommatidium, showing the arrangement of
pigment characteristic for the light.

Longitudinal section of an ommatidium, showing the arrangement of
pigment characteristic for the dark.

Transverse section of a cone from an ommatidium, such as is shown in
Fig. 2 (dark).

Transverse section through the proximal processes of the distal retinular
cells in an ommatidium such as that shown in Fig. 2 (dark).

Transverse section through the distal retinular cells of an ommatidium
such as that shown in Fig. 1 (light).

Transverse section through the retinula (rhabdome and proximal retinular
cells) of an ommatidium like that shown in Fig. 1 (light).

Transverse section through two groups of retinal nerve fibres.

Lateral view of a cone with one of its two distal retinular cells still
attached. The distal retinular cell shows the condition characteristic
for the dark. The preparation was isolated from a retina macerated
in Müller's fluid.

Longitudinal section through the bodies of two distal retinular cells,
which show slight foldings accompanying their shortening. The
preparation was made from an animal which had been kept in the
dark thirty-eight days and then exposed to light for four hours and
three quarters. The figure was drawn from preparations made by
Professor F. H. Herrick, who kindly granted the author the privilege
of studying them and making drawings from them.

Longitudinal section through the edges of two distal retinular cells (see
p. 287), from the same set of sections as that from which Fig. 9 was

drawn.

PARKER 7 RETINAL, PIGMENT.

— ͤDMD— is
= e ern 3
nen
Tog, nd. cnn
nl.con cl.con.
cl dst.

ct. dst
e "d U A.
= : eo cl.con
7. k O¢
e à
* €
cl. d st
f:
nl py 5.
rhb

cl sn s el PA
6.
mb ba
f hr: à
? fh.
!
i el.con
JO 9
cl dsl

wii i

Date Due

FEBS +1069

AR. PcG
NOV 3 0 1988

he oe