LIBRARY

UNIVERSITY OF CALIFORNIA

DAVIS

V

^

Digitized by the Internet Archive

in 2007 with funding from

IVIicrosoft Corporation

http://www.archive.org/details/biologicallectur06maririch

BIOLOGICAL LECTURES

FROM

THE MARINE BIOLOGICAL LABORATORY
WOOD'S HOLL, MASS.

1898

BOSTON, U.S.A.
GINN & COMPANY, PUBLISHERS

LIBRARY

Copyright, 1899
By GINN & COMPANY

ALL RIGHTS RESERVED

CONTENTS,

LECTURE PAGE

— I. TJic StnicULrc of Protoplasm. E. B. Wilson . . i

^ II. Cell-Lineage and Ancestral Reinijiiscence. E. B.

Wilson 21

^ III. Adaptation in Cleavage. F. R. Lillie . . 43 .,/'

— IV. Protoplasmic Movement as a Factor of Differen-

tiation. E. G. CoNKLiN 69

V. Eqnal and Unequal Cleavage in Annelids. A. L. -

Treadwell 93 v^

VI. TJie Cell Origin of the Prototroch. A. D. Mead .113 _,

VI 1. Relation of the Axis of the Embiyo to the First

Cleavage Plane. C. M. Clapp 139 ^

VI 11. Observations on Various Nucleolar Strnctures of

the Cell. T. H. Montgomery, Jr 153

IX. Protoplasmic Contractility and PhospJiorescence.

S. Watase 177

"* X. Some Problems of Regeneration. T. H. Morgan 193

— XI. The Elimination of the Unfit as Illnstrated by

the Intivdiiced Sparroiv, Passer Domesticns.

H. C. BuMPUS 209

XII. On the Heredity of the Marking in Fish Embryos.

Jacques Loeb . , 227

XIII. Do the Reactions of Loiver Animals Dne to Injury

Indicate Pain-Sensations f W. W. Norman . 235

XIV. North American Ruminant-Like Mammals. W. B.

Scott 243

XV. Caspar Friedrich Wolff and the Theoria Gene-

rationis. W. M. Wheeler 265

— XVI. Animal Behavior. C. O. Whitman ...".. 285

FIRST LECTURE.

THE STRUCTURE OF PROTOPLASM.^

EDMUND B. WILSON.

It would be superfluous to dwell in this place on the deep
and enduring interest that attaches to the microscopical study
of protoplasm. Since the time when the studies of Cohn and
Schultze led to the general recognition of protoplasm as the
material substratum of vital activity, — a conclusion so elo-
quently set forth by Huxley in his celebrated essay on the
physical basis of life, — this interest has continually increased,
as we have come to see even more clearly that all biological
phenomena are directly or indirectly traceable to the effects af
protoplasmic activity, for we have thus been impelled to seek
for an understanding of that activity in the morphological struc-
ture of protoplasm, as revealed by the microscope. It is small
wonder that to this quest some of the ablest of modern biolo-
gists have devoted their best energies. And yet, if we take
account of the actual knowledge gained, we cannot repress
a certain sense of disappointment, partly that microscopical
research should have fallen so far short of giving the insight
for which we had hoped, but still more because of the failure
of the best observers to reach any unanimity in the interpreta-
tion of what is actually visible under the microscope. In any
consideration of the general subject, therefore, it is well to keep
clearly in view the fact that such disagreement exists, and that

1 A more adequately illustrated special paper on this subject, containing more
specific references to the literature, is now in press. It should be borne in mind
that such delicate textures as those seen in the protoplasm of living cells cannot
be properly illustrated by black and white figures. The accompanying text figures,
though copied as accurately as possible from the original drawings, are of necessity
relatively rude and schematic.

I

2 BIOLOGICAL LECTURES.

we are not yet in a position to justify any very certain or far-
reaching conclusions.

I would like, at the outset, to express the opinion that, if we
except certain highly specialized structures, the hope of find-
ing in visible protoplasmic structure any approach to an un-
derstanding of its physiological activity is growing more, instead
of less, remote, and is giving way to a conviction that the way
of progress lies rather in an appeal to the ultra-microscopical
protoplasmic organization and to the chemical processes through
which this is expressed. Nevertheless, it is of very great im-
portance to arrive at definite conclusions regarding the visible
morphology of protoplasm, not only because of its intimate
connection with all the problems of cell-morphology, but also
in order to find the right framework, as it were, for our physio-
logical conceptions, and thus to gain suggestion for further
physiological and chemical inquiry. And this must be my
excuse for reviewing a subject which is still so largely obscured
by doubt, and of which the outcome gives, after all, so little
satisfaction.

It is especially important in this field of biological inquiry
to distinguish clearly between theory and observed fact ; for
theories of protoplasmic structure have always far outrun the
actual achievements of observation. From the time of Briicke
(one of the first to insist that protoplasm must possess a far
more complicated organization than that visible under the
microscope), speculation has gone steadily forward, to reach, per-
haps, its most elaborate expression in Weismann's interesting,
but unconvincing, work on the germ-plasm — an elaborate specu-
lative system built out of hypotheses which for the most part
float in the air without visible means of support. We need not
consider this side of the subject in extenso, but I will ask atten-
tion, for a moment, to what is the most characteristic and, to the
morphologist, the most interesting point in these speculations,
namely, the doctrine of genetic continuity as applied to the
corpuscular, or micellar, theory of protoplasm. We are all
familiar with the successive steps by which that doctrine gradu-
ally developed. Harvey's celebrated formula, ex ovo omnia^ —
or, as usually quoted, ornne vivuni ex ovOy — took with Redi the

THE STRUCTURE OF PROTOPLASM. 3

far more philosophical form, omiie vivum e vivo, thus expressing
a truth which forms the very foundation of all biological teach-
ing at the present day. The development of the cell-theory,
long afterwards, enabled Virchow to pronounce the more specific
aphorism, oinnis cellula e celhila (1855), — a statement involving
the highly interesting conclusion that protoplasm is never formed
de novo, but always arises from or through the activity of pre-
existing protoplasm differentiated into the form of a cell. Still
later a like conclusion was reached with respect to at least one
of the structural components of the cell, namely, the nucleus,
and the work especially of Flemming and Strasburger justified
the saying, oninis nucleus e niicleo. Not long afterwards, the
researches of Schmitz, Schimper, and others showed that in
plant cells some, if not all, forms of plastids (for example, the
chlorophyll-bodies) likewise arise by the division of preexisting
bodies of the same kind. Thus the law of genetic continuity
was gradually extended downwards from the grosser and more
obvious characters of the organism into the finer details of its
structural elements. Genetic continuity, the origin of like from
like, may now safely be regarded as a demonstrated fact in the
case of all existing organisms and of all cells ; it hardly falls
short of the same degree of certainty as applied to the nucleus ;
it is probable in the case of various forms of plastids in plant
cells; while the centrosome is now being weighed in the balance
with the evidence for the moment apparently accumulating on
the negative side.

Up to this point we have been dealing with matters of
observed fact. The next and final step was, however, taken in
the region of pure speculation, which had in the mean time been
at work building upwards from hypotheses regarding the basic
composition of protoplasm. Briicke's suggestion, that the cell
might be a congeries of bodies more elementary than itself,
found a much fuller expression in Herbert Spencer's theory of
physiological units ; but it was Darwin's theory of pangenesis
that laid the real basis for what followed in the works of
De Vries, Wiesner, Weismann, and Hertwig. The common
feature in all these later views is the conception of protoplasm,
not as a homogeneous substance or mixture of substances, but

4

BIOLOGICAL LECTURES.

as made up of a host of elementary ultra-microscopical corpus-
cles ("pangens," '* biophores," etc.), specifically different, capable
of assimilation, growth, and multiplication, and arising by divi-
sion of preexisting bodies of like kind. Developed as a purely
theoretical hypothesis, and within somewhat narrower limits, by
Darwin, this conception was expanded, and brought into more
direct relation with observed fact, especially by De Vries and
Wiesner, who showed how the assumption of such elementary
self-propagating corpuscles at the basis of living matter enabled
us to bring all the observed phenomena of genetic continuity
under a common point of view. The fundamental hypothesis
itself — i.e.^ the genetic continuity of the ultimate morphological
units — has, however, always remained, and still remains, a pure
assumption, incapable of direct proof or disproof ; for, with the
exception of Altmann and a few of his followers, all are agreed
that such elementary corpuscles, if they exist, must lie be-
yond the limits of microscopical vision. Altmann, however,
has sought to identify the elementary units, or "bioblasts,"
with the visible protoplasmic granules; and, in his writings, the
series of Latin aphorisms initiated by Redi culminates in the
saying, omfie gramiliim e granulo (!), but this conclusion has not
been taken very seriously by most other investigators.

I have given this very brief sketch of the theoretical side
of the question merely as an introduction, and shall dwell no
farther on it at this point, since my main purpose is to ask
attention to the visible, as opposed to the hypothetical invisible,
structure of protoplasm. A subject so vast, displaying so great
a conflict of opinion, must be very briefly treated within the
limits of a single lecture ; and I shall, therefore, confine the
discussion in the main to the protoplasm of the echinoderm-
^g%, which is accessible to every one, has been made a classical
object through the studies of such leaders of research as Flem-
ming, Biitschli, and Hertwig, and illustrates as clearly, perhaps,
as any other the various interpretations of protoplasmic struc-
ture that have been given.

In thin sections of well-preserved material, the protoplasm
of a star-fish or sea-urchin ^gg gives the appearance, under a
high power, of a fine meshwork or framework composed of innu-

THE STRUCTURE OF PROTOPLASM.

merable minute granules, or miavsomeSy suspended in a clearer,
less deeply staining, continuous substance (Figs, i, ^, and 4).
The spaces of the meshwork, which measure from one to nearly
two microns, are occupied by a third substance, clear, homo-
geneous, and of only slight staining capacity, which has often
been called the ground-substance. During cell-division, the
meshwork in the neighborhood of the dividing nucleus assumes
a radiating appearance, giving rise to the so-called asters, or
astral systems which are typically double, form ins: the amphi-

0..- OoQ. P. .vAv^rPo-^' <:^:'6o-

Fig. !. — («) Protoplasm of the egg of the sea-urchin {ToxoPfieustes) in section ; (b) protoplasm
from a living star-fish egg {Asierias)\ (c) the same in a dying condition after crushing
the egg; (<f) protoplasm from a young ovarian egg of the same. (All the figures mag-
nified 1200 diameters.)

aster (Fig. 3, b). We may define the problems suggested by
these appearances by a series of questions as follows :

1 . What is the actual structure that gives the appearance of
a meshwork .^

2. How faithfully does the preserved structure, as seen in
sections, reproduce that existing in life }

3. What is the relation of the astral systems to it.?

4. What is the finer structure and origin of the meshwork }

5. Can this structure be taken as typical of all protoplasm;
and if not, what is its relation to other forms of protoplasmic
structure .?

After seeking for answers to these queries, we may finally
inquire how they bear on the theoretical views briefly reviewed

6 BIOLOGICAL LECTURES.

above. Incidentally, still another interesting question arises,
namely : Is it possible to identify any one of the three elements
in question — granules, continuous substance, ground-substance
— as the /m>/^ substance or protoplasm proper, as distinguished
from a lifeless metaplasiriy and, if so, what are its structural
relations ?

Could we positively answer all these questions, we should
have taken a long step forwards in the study of the cell. Far
from this, however, in point of fact hardly any two observers
have given exactly the same answers to them. Leaving aside
the earlier views, we find in the recent literature of the subject
two principal general views with a number of modifications of
each.

The first of these agrees with the early view of Klein and
Van Beneden, that the protoplasm forms a network, reticulum^
or thread-work, composed of branching fibres embedded in a
homogeneous ground-substance which fills the interstices of
the network, and with granules or microsomes lying along the
course of the threads, or at the nodes of the network. Many
of those who adopt this interpretation further agree with their
predecessors, that the astral systems formed during cell-division
arise directly through a rearrangement of the preexisting net-
work about active centres of attractive or other forces, some-
what as iron-filings arrange themselves along the radiating
lines of force in a magnetic field, — an arrangement which bears
a remarkably close though only superficial resemblance to the
protoplasmic amphiaster. Boveri, and some others, however,
regard the astral system as having no direct relation to the pre-
existing network, believing that the rays either arise from a
specific substance (" archoplasm "), distinct both from the gen-
eral network and from the ground-substance, or are wholly new
formations which, as it were, crystallize afresh out of the proto-
plasmic substance.

The second view is that of Biitschli, who believes it to be
applicable to all forms of protoplasm, and who has been followed
by a considerable number of recent investigators. Biitschli's
interpretation differs entirely from the foregoing, the meshwork
being regarded not as a network, but as an appearance resulting

THE STRUCTURE OF PROTOPLASM. 7

from the optical section of ''alveolar" or emulsion-structure.
The spaces of the meshwork are drops of liquid occupying
spherical spaces or '' alveoli " ; the " fibres " are optical sections
of the thin layers, or lamellae, by which the drops, or alveoli,
are surrounded. Even the astral systems receive the same
interpretation, the astral "rays" and ''spindle-fibres" being
an optical illusion resulting from the radial arrangement of
the alveoli, and hence of the inter-alveolar septa by which they
are separated.

The greater number of observers of protoplasm have given
their adherence to one or the other of the two widely dissimilar
views just outlined, though there are others to which we shall
return later. Some investigators have taken a position inter-
mediate between these two extremes. Thus Reinke has main-
tained that the cytoplasm of the echinoderm-egg is alveolar,
as described by Biitschli (though, as will appear beyond, he
ascribes to this structure a different physiological interpreta-
tion), while the astral systems are fibrillar, as held by Van
Beneden, and arise as new formations at the cost of the alveolar
walls. More recently, Strasburger has developed the related,
but still different, view that the cytoplasm of the cell at large
consists of two distinct substances, namely, the trophoplasm, or
general nutritive plasma, which is alveolar, and the kinoplasm,
or the substance active in division, which is fibrillar and gives
rise to astral systems consisting of true rays and fibres.

It is remarkable that the best observers, working in many
cases at the same object, should have reached conclusions so
diverse. It is obvious, further, that in the face of such contra-
dictions it is impossible to give any discussion of the subject
that is not more or less strongly tinged with the personal views
of the writer. Such views, by whomsoever expressed, can at
present have no more than a provisional value ; and this is the
last subject on which dogmatism should be allowed. It is with
full recognition of these difficulties that I venture to state some
of my own conclusions, partly because they may serve to explain,
in some measure, to those who have not specialized in this field,
how the existing diversity of opinion has arisen, partly because
they have perhaps some bearing on the more general questions

8 BIOLOGICAL LECTURES.

that were referred to at the outset. I shall take up in order
the questions raised at page 5.

The Nature of the Meshwork. — Although in earlier papers
I was inclined to regard the meshwork of the echinoderm-egg
as a reticulum, further studies have left no doubt whatever, in
my opinion, that in the resting cell it is in reality an alveolar
structure — or, as I do not hesitate to call it, an emzdsion —
such as Butschli has described. I was first led to this conclu-
sion through the study of sections of the eggs of sea-urchins
{Toxopneiistes) and star-fish (Asterias)\ but whatever doubt may
have remained was completely dissipated by the study of the
living eggs of Asterias (Fig. i, b)^ Echinarachnius , Arbacia^
Ophiura (Fig. 2, a), under high powers. All of these eggs give
in life essentially the same appearance, though no two are
exactly alike. In all, the protoplasm consists of innumerable
closely crowded minute spheres suspended in a clear basis.
The spheres may be called the alveolar spheres, or more briefly
the alveoli, though strictly speaking the latter term should
designate the cavities which the spheres fill. The clear basis
in which they lie, and which forms the inter-alveolar walls, may,
with Mrs. Andrews, be called the continuous stcbstance. Scat-
tered about in these walls are numerous granules, or microso^nes,
far smaller than the alveoli, which often give the appearance of
an irregular network. If now we compare these appearances
of the living protoplasm with those seen in the sections mounted
in balsam, we find at first sight very considerable differences.
More critical study shows, however, that the differences are
almost wholly due to the effect of differential staining and to
the difference of refractive index in the mounting media in the
two cases. The alveoli of the living protoplasm form the spaces
of the meshwork. The latter consists of the continuous sub-
stance with the granules suspended in it. In the section, what
especially strikes the eye is the meshwork ; for the alveolar
spheres do not stain, and their contours become indistinct in
the highly refracting balsam, while the continuous substance
stains slightly, and the granules intensely, thus giving the
appearance of a conspicuous granular meshwork. We thus
arrive at a definite answer to two of the questions propounded

THE STRUCTURE OF PROTOPLASM.

on page 5, namely : (i) the meshwork shown in sections is not
a network, but the expression of an alveolar or emulsion-struc-
ture, and (2) proper fixation does not produce a mass of coagu-
lation-artifacts, but preserves the visible structure very nearly
as it exists in life.

The above conclusions are based mainly on the study of star-
fish eggs, but are confirmed by the facts observed in other forms.
In Arbacia the emulsion is considerably finer, the alveoli meas-
uring on an average no ,^^,
more than i.o micron, ^'^^■ik>J'':^
while the finer granules
are relatively less nu-
merous. The pigment-
granules characteristic
of this form appear to be
nothing other than mod-
ified alveolar spheres.
In Toxopneiistes the al-
veoli measure approxi-
mately from 1.0 to 1.3
microns, while the gran-
ules are more numerous
than in Asterias. In
Echinarachnius the al-
veoli are less uniform
in size than in Asterias,
the largest measuring
up to about 1 .7 microns,
while the granules are
less numerous. The
^gg of Ophiicm, finally, has an extremely coarse structure, the
alveolar spheres measuring on an average 3.0 to 4.0 microns,
while the granules or microsomes are also very large and, in the
superficial layers of the protoplasm, even more numerous than
in Toxopneustes. The protoplasm of Ophiura (Fig. 2) is highly
favorable for study, not only on account of the great size of its
elements, but also by reason of the remarkable fact that these
elements are colored in life, the alveolar spheres being in most

Fig. 2. — (a) Protoplasm from a living ophiuran egg {Ophi-
ura), slightly compressed, so as to spread the yolk-
spheres somewhat apart ; {b) the same as seen in a
section (sublimate-acetic, iron-haematoxylin ; 1200 diam-
eters).

lO BIOLOGICAL LECTURES.

individuals distinctly of an olivaceous or pinkish-brown color,
while the larger granules or microsomes are lemon yellow.
This circumstance makes possible an observation of great im-
portance, namely, that all the elements of the protoplasm are
liquid or viscid. If the eggs of Ophiura be crushed by pres-
sure on the cover-glass, the protoplasm flows out, most of the
alveolar spheres going in advance, while the granules and con-
tinuous substance lag behind. Meanwhile, the alveolar spheres
often run together to form larger drops of all sizes, the origin
of which is placed beyond question by their color. The same
is true of the yellow microsomes, though this takes place less
readily, and only under somewhat rough treatment. This
demonstrates the liquid, or at least viscid, nature of both the
spheres and the microsomes, and no less certainly that of the
continuous substance in which both lie. As far as the alveolar
spheres are concerned, the same observation may readily be
made in the colorless protoplasm of Asterias (Fig. i, c), Echin-
arachnius^ or Arbacia^ but I could never satisfy myself of the
liquid nature of the microsomes in these forms. The case of
Ophiura renders it highly probable, however, that the granules
are liquid in these forms also, — a conclusion which I confess
was a surprising result to me ; for we are so accustomed from
our studies on sections to regard the granules as solid bodies
that we are apt to forget that sections show us only coagulated
material.

To sum up, a critical study of the living protoplasm of
echinoderm-eggs shows that it is a liquid, or rather a mixture
of liquids, in the form of a fine emulsion consisting of a con-
tinuous substance in which are suspended drops of two general
orders of magnitude and of different chemical nature, as indi-
cated by their staining reactions. The larger drops, forming
the alveolar spheres, stain only slightly in haematoxylin, and
constitute the so-called ''ground-substance" in the spaces of
the meshwork ; these have an average size, ranging in the vari-
ous forms studied from i.o micron or less {Arbacia) up to 4.0
microns {Ophiura). The smaller drops, forming the granules
or microsomes, are very much more minute, and stain intensely
with iron-haematoxylin. The presence of the larger drops deter-

THE STRUCTURE OF PROTOPLASM. \\

mines the primary alveolar structure as described by Biitschli.
The smaller drops ("granules") lying between these gives rise
to the "secondary," or finer alveolar structure as described by
Reinke, and subsequently by Mrs. Andrews, as I understand
these authors.

Relations of the Astral Rays to the Meshivork. — We may
now make a brief digression to consider the third question pro-
pounded on page 5, namely : What is the relation of the astral
rays and spindle-fibres to the alveolar substance } It is easy to
see, both in sections and in living material, that in a well-devel-
oped aster the alveoli are arranged in radiating lines between
the astral rays (Fig. 4), precisely as Biitschli and so many others
have described. The rays themselves are, however, something
more than the radially arranged inter-alveolar septa, for, in the
first place, they are often much thicker than these septa, and,
in the second place, they stain more intensely than the continu-
ous substance. A careful study of the rays in the echinoderms,
and in many other forms (especially in Nereis^ Thalassema,
Lamellidoris, and Ascarts), leaves, I think, no room for doubt
that, in sections at least, the rays are actual branching fibrillae,
as described by so many observers since the time of Van
Beneden, that thread their way through the continuous sub-
stance between the alveoli, often in a zigzag course. The
strongest evidence that they are fibrillae is given by the appear-
ance-of the cut ends of the rays as they appear in somewhat
excentric or rather thick sections of the asters. In such sec-
tions, particularly in the case of large and coarse asters like
those of Nereis (Fig. 3, b), the rays may be seen in the clearest
manner to terminate as they pass upwards towards the eye in
well-defined cut ends, and I think no one who studies these
preparations can doubt that in them the asters are true fibrillar
structures.

We may now inquire in what manner the rays arise and grow,
and what is the origin of their substance. In the growing aster
the rays progressively extend themselves from the centre out-
wards, gradually losing themselves in the general meshwork.
It has been maintained by some writers that the rays grow out-
wards from their bases like the roots of a plant, and in a certain

12

BIOLOGICAL LECTURES.

sense this is undoubtedly true. But it is difficult to believe that
all of the material of the rays comes from the base {i.e.^ from
the nucleus or the centrosome), for they often extend themselves
throughout the entire cytoplasm, even in cases where, as in the

Fig. 3. — («) Protoplasm and yolk-spheres from the egg of Thalassema in section. The upper
part of the section shows the result of prolonged extraction of the dye (iron-hjematoxylin),
the lower half represents varying degrees of extraction (1200 diameters); {b) egg of Nereis
in section showing yolk-spheres and the first polar amphiaster above (600 diameters).

sperm-aster of echinoderms, the centre of the aster remains very
small, and the nucleus still consists of a compact mass of chro-
matin (Fig. 4). It is more probable that they grow at the tip,
continually extending themselves at the cost of the material
lying in the meshwork. When the rays are followed out periph-

THE STRUCTURE OF PROTOPLASM.

13

erally they may often be seen to run out into rows of granules
like beads on a string. Van Beneden, .who has been followed
by many later writers, was inclined to regard the rays as essen-
tially rows of microsomes strung together by a homogeneous
clear substance, — i.e.^ by the continuous substance, — and I was
led to the same conclusion in the case of sea-urchin eggs. A
study of the asters in Ophiiira throws doubt upon this conclu-

v'-^'^.rr':"*

Fig. 4. — Section of sea-urchin &gg {Toxo/>neustes), i^ minutes after entrance of the sperma-
tozoon, showing sperm-nucleus, middle piece, and aster (about 2000 diameters).

sion, for it is here certain that the larger and deeply staining
microsomes do not build up the ray, but are quite irregularly
scattered along its course. The rays here mainly arise, I believe,
in, and at the expense of, the continuous substance, and the
linear arrangement of the microsomes is incidental to the dif-
ferentiation of this substance along a definite tract which more
or less involves the microsomes as it progresses. This conclu-
sion probably also applies to other forms. The material active
in the ray-formation appears to be the continuous substance,

14 BIOLOGICAL LECTURES.

and while the microsomes may, and probably in many cases do,
contribute to the ray, they probably play the part of reserve-
material rather than of active elements.^

To sum up, the general result indicates that the opinions
regarding the aster-formation referred to on page 6 can in a
measure be reconciled. In the case of echinoderm-eggs Biitschli
and Erlanger correctly describe the aster as involving a radial
arrangement of the alveoli, but they have failed to recognize
the fibrillae that lie between them, and Boveri is therefore thor-
oughly justified in the contention that the astral systems cannot
be regarded as merely a radial configuration of the preexisting
meshwork. I nevertheless think that Hertwig, Reinke, and
myself were right in the contention, which has been made
also by many others, that the rays grow by progressive differ-
entiation out of the general cytoplasmic meshwork, and that
there is no ground, in the echinoderm-egg at least, for the
recognition of a specific "archoplasm" or *'kinoplasm" from
which they arise.

Finer Strucitire and Oiigm of the Meshwork. — We may
now consider what is, I think, the most suggestive of the ques-
tions propounded, namely, that relating to the finer structure
and origin of the meshwork. We have thus far distinguished
sharply between alveolar spheres, granules, or microsomes, and
continuous substance. Morphologically considered, however,
there is good reason for the view that all these are but different
gradations of one structure. In the first place, a nearly or
quite complete series of size-gradations exists between the
largest alveoli and the microsomes (Fig. i, b, c). Although
most of the alveoli vary but slightly in size from the mean, a
little search shows the presence of many smaller ones, and here
and there they may seem almost, if not quite, as small as the
larger microsomes. In the second place, careful study of the
'^ continuous " substance in life, especially in the crushed proto-
plasm, shows that the larger microsomes in turn graduate down
to granules so small as to lie near or at the limit of microscopical
vision. The ** continuous " substance is, in other words, filled

1 As already pointed out, we cannot assume that the ray is merely an accumula-
tion of the continuous substance on account of its different staining capacity.

THE STRUCTURE OF PROTOPLASM. 15

with granules, i.e.^ drops, of all sizes, ranging from the smallest
visible ones up to the largest alveoli. It is this fact which Mrs.
Andrews, as I understand her statements, has in view in main-
taining that the coarser alveolar structure '' is not indeed the
final structure of the living substance, but is part only of an
infinitely graded series of vesiculations of the protoplasmic
foam" ('97, p. 12), and with this statement I entirely agree.
But we cannot stop here. Irresistibly the further question
suggests itself : Why should we place the end of this series at
the end of microscopical vision under a 1.5 mm, immersion
objective — which is of course a perfectly arbitrary and arti-
ficial limit } It is impossible to doubt that powers still higher
than any at our command would reveal the existence of granules
still smaller, and that what appears as "■ continuous " or "homo-
geneous " substance is itself an emulsion beyond the range of
vision.

We may now inquire whether the coarser visible alveolar
structure is characteristic of all protoplasm. This question
has in a measure already been answered, for in these very eggs
we have seen the alveolar structure giving rise to a fibrillar one
in the aster-formation — in other words, the protoplasm of the
same cell may in different phases pass back and forth from one
state into another. This fact appears in its clearest form when
we study the growth of the ovarian ova, which gives us many
additional suggestions of high interest. The entire coarser
alveolar structure^ as described above, — i.e., the foam-structure
of Biitschli, — is in these eggs of secondary origin. The very
young living ovarian eggs consist of "homogeneous" proto-
plasm, such as has been described by many botanists in the
embryonic tissue-cells, through which are irregularly scattered
a few small spheres and many excessively small granules. As
growth proceeds, both the spheres and the granules increase in
size, the latter enlarging to form new spheres, while new gran-
ules continually emerge from the protoplasmic background into
the limits of vision. In the middle stages of growth, the proto-
plasm is thus converted into an emulsion, being filled with
spheres of all sizes, ranging downwards from i.o micron to
the smallest granules, but still showing no regular arrangement

1 6 BIOLOGICAL LECTURES.

(Fig. \, d). As the ^gg approaches maturity, the spheres be-
come differentiated into two groups, the larger ones becoming
approximately of the same size (cf. p. lo) to form the alveolar
spheres and crowding together, while the smaller ones remain
as the microsomes and finer granules embedded in the remains
of the continuous substance which forms the basis of the mesh-
work. In one sense, therefore, the alveolar spheres and the
microsomes are only different stages in the same morphological
series, — though it should be remembered that they differ chem-
ically as well as in size, and I do not mean to imply that the
one may develop into the other, — and both the alveolar and
the fibrillar or reticular structures in these eggs are of second-
ary origin. If this be the case, neither of these types of struc-
ture can be of fundamental importance ; and I fully agree with
the opinion of Kolliker, which has been adopted by an increas-
ing number of later observers, that no universal or even general
formula for protoplasmic structure can be given. The evidence
indicates that alveolar, granular, fibrillar, and reticular structures
are all of secondary origin and importance, and that the ultimate
backgronnd of protoplasmic activity is the seitsibly homogeneous
matrix or contimious substance in which those structures appear.
I do not mean to say that this is the only " living " element
in the cell. The distinction between ''living" and "lifeless,"
between ''protoplasmic" and " metaplasmic," substances is
exceedingly difficult to define, — largely on account of our vague
and inconsistent use of terms, for in practice we continually use
the word "living" to denote various degrees of vital activity.
Protoplasm deprived of nuclear matter has lost, wholly or in
part, one of the most characteristic vital properties, namely,
the power of synthetic metabolism; yet we still speak of it as
"living," because it may for a long time perform some of the
other functions, manifesting irritability and contractility, and
showing also definite coordinations of movements (as in the
enucleated protozoan) ; and in like manner various structural
elements of the cell may be termed living in a still more
restricted sense. In its fullest meaning, however, the word
"living" implies the existence of a group of cooperating factors
more complex than those manifested by any one substance or

THE STRUCTURE OF PROTOPLASM. 17

Structural element in the cell, and I am therefore thoroughly in
accord with those who have insisted that life in its full sense is
the property of the cell-system as a whole rather than of any
one of its separate elements. Nevertheless, we are perhaps
justified in maintaining that the continuous substance is the
most constant and active element, and that which forms the
fundamental basis of the system, transforming itself into gran-
ules, drops, fibrillae or networks in accordance with varying
physiological needs. ^

Whether any or all of these elements are ''living" or "life-
less " depends largely on the sense in which these words are
used ; and it is well, therefore, to follow the example of Sachs,
in substituting for these words, as applied to special structural
elements of the cell, the terms "active" and "passive," which
properly admit of degrees of comparison. The distinction
between "protoplasmic" (active) and " metaplasmic " or "para-
plasmic" (passive) elements, though a real and necessary one,
thus becomes, after all, one of degree only.

We are thus brought to consider another point of some inter-
est suggested by the comparative study of the facts described
above. Biitschli states that in the true or finer alveolar struc-
ture, characteristic of protoplasm in general, the alveoli do not
measure more than 2.0 microns, and as a rule are considerably
smaller. This, he insists, is not to be confounded with a
"coarser vacuolization," characterized by larger drops or spheres,
which may secondarily arise in the finer structure. Again,
Reinke and Waldeyer in a somewhat similar manner character-
ize as " pseudo-alveolar " a structure arising secondarily through
the deposit of passive metaplasmic products of metabolism, such
as yolk-spheres, fat-drops, and the like, in the living proto-
plasmic basis. Both distinctions break down, I think, in the
light of the foregoing facts. In most of the forms considered,
— Arbacia^ Toxopneustes^ Echmarachnins^ Asterias, — the alve-
olar spheres are considerably less than 2.0 microns (i.o to 1.7),

^ It is hardly necessary to state that this view is not original, except in so far
as it has been directly suggested by the observations described above ; for it has
been more or less definitely maintained by many others, and I am only expressing
what seems to be a growing opinion among workers in this field.

1 8 BIOLOGICAL LECTURES.

and the structure is therefore a true alveolar one in Biitschli's
sense ; indeed, Butschli himself describes and figures the proto-
plasm of the SpJiaerechiniis ^g^ as an example of that structure.
In Ophiura, however, the spheres measure up to 3.0 or 4.0
microns, and are undoubtedly '' yolk-spheres " in the usual sense.
It is, however, quite certain from the ovarian development of
these eggs that they differ from the others only in degree, and
that Biitschli's criterion of size gives no satisfactory ground for
any real distinction. The alternative is to regard all the forms
as pseudo-alveolar, irrespective of the size of the alveolar
spheres, which are in all cases to be regarded as metaplasmic
bodies ; and this is the view which Reinke specifically applies
to Sphaerechinus. But if this view be adopted, we seek in vain
for any ground of distinction between such a fine "pseudo-
alveolar " structure as that of Arbacia^ and the " true " alveolar
structure of tissue cells, and are forced to the conclusion that
in the latter case also the alveolar substance consists of passive
or metaplasmic material, — a view which has in fact been adopted
by some writers. For my part, I am convinced that the entire
distinction is without adequate basis, and that no definite bound-
ary-line can be drawn between even the largest deutoplasm-
spheres, vacuoles, or other metaplasmic deposits, the alveolar
spheres of Arbacia or ToxopneusteSy and those occurring in tissue-
cells; and probably all are, in the sense indicated above, to be
classed among the relatively passive or metaplasmic material.

How generally the alveolar, reticular, or fibrillar formations
may occur is a matter still to be determined by observation. It
is probable that the alveolar structure will be found to be of
more general occurrence than has been supposed; and judging
by the appearance observed in echinoderm and other eggs, and
in coagulated albumen and other structureless proteids, I sus-
pect that some cases of so-called "reticular" formations will be
found to arise through the more or less imperfect fixation of
the alveolar, leading to the coagulation, contraction, and break-
ing down of the alveolar walls,^ though I do not for a moment
mean to imply that such is the case with all reticula.

1 It may be well to point out that Rhumbler has produced true fibrillar and
reticular formations in coagulated artificial gelatine-emulsions.

THE STRUCTURE OF PROTOPLASM. 19

What light, if any, do the foregoing general conclusions
throw on the theoretical views outlined at the beginning of this
lecture ? The answer must be : None that is clear and satis-
factory, for the background of all the phenomena appears to
lie in the invisible organization of a substance which seems to
the eye homogeneous. Yet there is, I think, much in these
conclusions to suggest, and nothing to contradict, the hypothesis
that the ''homogeneous" or "continuous" substance may be
composed of ultra-microscopical bodies by the growth and dif-
ferentiation of which the visible elements arise, and which differ
among themselves chemically and otherwise, as is the case with
the larger masses to which they give rise. I will not enter
upon a discussion of the question whether these bodies are
merely molecules, more or less complex, or groups of molecules
forming protoplasmic units or micellae, but will only make three
suggestions. First, if such units exist, they cannot be identi-
fied with the visible granules or " bioblasts " of Altmann, but
are bodies far smaller. Second, if there be any truth in what
has been said above regarding the localization of *' living "
matter in the cell, such protoplasmic units, if they exist, cannot
properly be called '* biophores," since life is a manifestation of
the system which they form and not of the individual units.
The corpuscular or micellar theory of protoplasm, as an hypoth-
esis of morphological organization, should not be confounded
with the physiological theory that biophores or pangens are
"elementary living units." Third, by ascribing to these hypo-
thetical units the power of growth and division, in accordance
with the pangen theory, we are enabled to get a certain amount
of light upon some of the most puzzling questions of cytology,
such, for example, as the ultimate nature and origin of dividing
cell-organs like the nucleus or the plastids, and especially such
a contradiction as that presented by the centrosome which may
apparently arise either de novo or by division of a preexisting
body of the same kind. As De Vries and Wiesner have so
suggestively urged, the power of division on which the law of
genetic continuity rests and which is manifested by morphologi-
cal aggregates of so many different degrees, may have its root
in a like power of the primary units at the bottom of the series.

20 BIOLOGICAL LECTURES.

out of which all the higher members are built. But while giving
due weight to this suggestive hypothesis, we may question
whether its acceptance does not introduce as many new special
difficulties as those which it sets aside ; while we must admit
that it leaves untouched the fundamental problem of division.
The solution of this problem may perhaps have to be sought in
a quite different direction from the pangen hypothesis. Whether
we shall succeed in finding it is another question.

SECOND LECTURE.

CELL-LINEAGE AND ANCESTRAL
REMINISCENCE.i

EDMUND B. WILSON.

Every living being, at every period of its existence, presents
us with a double problem. First, it is a complicated piece of
mechanism, which so operates as to maintain, actively or pas-
sively, a moving equilibrium between its own parts and with its
environment. It thus exhibits an adaptation of means to ends,
to determine the nature of which, as it now exists, is the first
task of the biologist. But, in the second place, the particular
character of this adaptation cannot be explained by reference to
existing conditions alone, since the organism is a product of the
past as well as of the present, and its existing characteristics
give in some manner a record of its past history. Our second
task in the investigation of any problem of morphology or phys-
iology must accordingly be to look into the historical back-
ground of the phenomena ; and in the course of this inquiry we
must make the attempt, by means of comparisons with related

1 This lecture is based on a paper entitled "Considerations on Cell-Lineage and
Ancestral Reminiscence, Based on a Reexamination of Some Points in the Early
Development of Annelids and Polyclades," in Ann. N. V. Acad. Set., 1898.
In some passages the wording of that paper has been reproduced with only
slight change. With the exception of Fig. 4, the figures are entirely schematic and
are designed to show only the broadest and most essential topographical features.
For this purpose the subdivisions of the micromeres have been omitted, and,
except in Fig. 4, none of the figures represent the actual condition of the embryo at
any given period. While, therefore, very misleading in matters of detail, they are,
I think, true to the essential phenomena ; and through the simplification thus
effected the reader is spared a mass of confusing descriptive detail in no way essen-
tial to the broad relation on which it is desired to focus the attention.'

2 2 BIOLOGICAL LECTURES.

phenomena, to sift out adaptations to existing conditions from
those which can only be comprehended by reference to former
conditions. Phenomena of the latter class may, for the sake of
brevity, conveniently be termed "ancestral reminiscences," —
though it may not be superfluous to remark that every char-
acteristic of the organism is in a broad sense reminiscent of
the past.

It is in embryological development that ancestral reminis-
cence is most familiar and most striking. We all know that
development rarely takes the shortest and most direct path,
but makes various detours and sometimes even moves backward
so that the adult may actually be simpler than the embryo.
Such vagaries of development are in many cases only intelli-
gible when regarded as reminiscences of bygone conditions,
either of the adult or of the embryo. Sometimes these records
of the past are so consecutive and complete that the individual
development, or ontogeny, may be said to repeat or recapitulate
the ancestral development, or phylogeny. The development of
the toad's ^g^, for example, probably gives in its main outlines
a fairly true picture of the ancestral history of the toad race,
which arose from fish-like ancestors, developed into aquatic air-
breathing tailed forms, and finally in their last estate became
tailless terrestrial forms. It was such facts as these that led
Haeckel, building on the basis laid by Darwin and Fritz Miiller,
to the enunciation of the famous so-called "biogenetic" law,
that the ontogeny, or history, of the individual tends to repeat in
an abbreviated and more or less modified form the phylogeny, or
history, of the race. The event has shown that actual recapitu-
lation or repetition of this kind is of relatively rare occurrence.
Development more often shows, not a definite record of the
ancestral history, but a more or less vague and disconnected
series of reminiscences, and these may relate either to the adult
or to the embryonic stages of the ancestral type. Thus the
embryo mammal shows in its gill-slits and aortic arches what
must probably b6 regarded as reminiscences of a fish-like adult
ancestor, while in the primitive streak it gives a reminiscence
not of an adult form but of an ancestral mode of development
from a heavily yolk-laden q^% like that of the reptiles.

CELL-LINEAGE. 23

If we survey the general field of embryology, we find that
ancestral reminiscence in development is most conspicuously
shown and has been longest known in the later stages, and
many of the most interesting and hotly contested controversies
of modern embryology have been waged in the discussion of
the possible ancestral significance of larval forms, such as the
trochophore, the Nauplitis, the ascidian tadpole, and many
others. It is generally admitted, too, that ancestral remi-
niscences may occur in earlier embryonic stages. While few
naturalists would to-day accept Haeckel's celebrated Gastraea
theory in its original form, probably still fewer would deny that
the diblastic embryo (gastrula, planula, etc.) of higher forms
is in a certain sense reminiscent of the origin of these forms
from diblastic ancestors having something in common with
existing ccelenterates.

It is in respect to still earlier stages, namely, those including
the cleavage of the Qgg, that the greatest doubt now exists ;
and there is hardly a question in embryology more interesting
or more momentous than whether these stages may exhibit
ancestral reminiscence, and whether they, like the later stages,
exhibit definite homologies, and thus afford in some measure a
guide to relationship. None of the earlier embryologists were
disposed to answer this question in the affirmative. To them,
and it should be added to some of our contemporaries as well,
the cleavage of the ovum was '' a mere vegetative repetition of
parts," the details of which had no ancestral significance, and
the ontogeny first acquired a definite phyletic meaning and in-
terest with the differentiation of the embryonic tissues and
organs. To these observers the cleavage of the ovum pre-
sented merely a series of problems in the mechanics of cell-
division, and its accurate study was almost wholly neglected as
having no interest for the historical study of descent. And
yet it was long ago shown that the blastomeres of the cleaving
ovum have in some cases as definite a morphological value as
the organs that appear in later stages. Kowalevsky and Rabl
traced the mesoblast-bands in annelids and gasteropods back to
a single cell, which still later research has shown to have the
same origin and fate, and hence to be homologous in the two

*24 BIOLOGICAL LECTURES.

cases by every criterion at our command. A long series of later
researches, beginning with Whitman's epoch-making studies
on the cleavage of Clepsine, has demonstrated analogous facts
in the case of many other cells of the cleaving ovum, and has
finally shown that in many groups of animals (though appar-
ently not in all) the origin of the adult organs may be determined
cell by cell in the cleavage stages ; that the cell-lineage thus
determined is not the vague and variable process it was once
supposed to be, but is in many cases as definitely ordered a
process as any other series of events in the ontogeny ; and that
it may accurately be compared with the cell-lineage of other
groups with a view to the determination of relationships.

The study of cell-lineage has thus given us what is practi-
cally a new method of embryological research. The value and
limitations of this method are, however, still under discussion,
and among special workers in this field opinion as to its mor-
phological value is still so widely divided that most of its results
should be taken as suggestive rather than demonstrative. Like
other embryological methods, it has already encountered con-
tradictions and difficulties so serious as to show that it is no
open sesame. In some cases closely related forms {e.g., gastero-
pods and cephalopods) have been shown to differ very widely,
apparently irreconcilably, in cell-Hneage. In other cases (echi-
noderms, annelids) the normal form of cleavage has been
artificially changed without altering the outcome of the devel-
opment. In still other cases (e.g., in teleost fishes) the form
of cleavage has been shown to be variable in many of its most
conspicuous features, so that apparently no definite cell-lineage
exists. These and many other facts, less striking but no less
puzzling, can be built into a strong case against the cell-lineage
program, and I wish to acknowledge its full force. Admitting
all the difficulties, I am nevertheless on the side of those who as
morphologists believe that the study of cell-lineage has demon-
strated its value, and that it promises to yield more valuable
results in the future. In this lecture I propose to illustrate
some of the more interesting results already attained, and some
of the suggestions that they give for future work, by a broad
consideration of the cell-lineage of three related groups of

CELL-LINE A GE. 2 5

animals which on the one hand have been very carefully exam-
ined as regards their anatomical and general embryological
relationships, while on the other hand their cell-lineage has
been more exhaustively studied than that of any other forms.
These groups are the platodes (more especially the Turbellaria),
the mollusks, and the annelids.

That these three groups belong in the same morphological
series will probably be admitted by all zoologists, and most
will no doubt further agree with the view of Lang, that in the
essential features of their organization the platodes are not very
far removed from the ancestral type from which the two higher
groups have sprung, the former having remained non-metameric
like the platodes, while the latter have acquired metamerism.
Accepting this view we should expect, if there be any evidence
of race-lineage in cell-lineage, to find in the annelids and mol-
lusks a common type of cleavage, and one which in its main
features may be derived from that of the platode. Recent
studies in cell-lineage have, on the whole, justified this expec-
tation, and have brought to light some cases of vestigial proc-
esses in cleavage which are, I believe, to be reckoned among
the most striking and beautiful examples of reminiscence in de-
velopment. It is especially to these cases that I wish to direct
attention.

The cleavage of a number of Ttirbellaria and nemerteans, and
of many annelids, gasteropods, and lamellibranchs, has now been
shown to conform to a common type which, though complex in
detail, is exceedingly simple in its essential plan. A few excep-
tions there certainly are ; but some of these are apparent only
(for example, in the acoelous Ttirbellaria), and are readily re-
ducible to the type, while others are undoubtedly correlated
with bygone changes in the mode of nutrition of the ovum (as
in some of the earthworms and leeches). The most conspicu-
ous exception is afforded by the cephalopods, which have a mode
of cleavage entirely unrelated to that of the other mollusks; but
the entire development of this group is of a highly modified
character. Fully recognizing the real exceptions, we never-
theless cannot fail to wonder at the marvellous constancy with
which the cleavage of the polyclades, nemertines, annelids, gas-

26 BIOLOGICAL LECTURES.

teropods, and lamellibranchs conforms to the typical mode of
development. In all these forms the ^^^ first divides into four
quadrants. From these at least three and sometimes four or
five quartets of cells — usually smaller, and hence designated
as mici'o^nei'es — are successively produced by more or less
unequal cleavages towards the upper pole. The arrangement
of these micromeres (Fig. i) is constant and highly character-
istic, the first quartet being more or less displaced, or, as it
were, rotated in a direction corresponding with the hands of a
watch (clockwise), the second in the opposite direction (anti-
clockwise), the third clockwise again, and so on, the spindles of
each division being at right angles to those of the preceding
and following. In the later subdivisions of the micromeres,
also, a most remarkable agreement has been observed ; but I
shall pass this over entirely in order to focus attention on the
broader features of the development.

A large part of the work in cell-lineage during the past ten
years has been devoted to a comparison of the morphological
value of these quartets of cells in the annelids, mollusks, and
platodes ; and the remarkable and interesting fact is now be-
coming apparent that while they do not have exactly the same
value in all the forms, they nevertheless show so close a cor-
respondence both in origin and in fate that it seems impossible
to explain the likeness save as a result of community of descent.
The very differences, as we shall see, give some of the most
interesting and convincing evidence of genetic affinity ; for
processes which in the lower forms play a leading role in the
development are in the higher forms so reduced as to be no
more than vestiges or reminiscences of what they once were,
and in some cases seem to have disappeared as completely as
the teeth of birds or the limbs of snakes. The processes in
question relate to the formation of the mesoblast in its relation
to the micromere-quartets, and on them the whole discussion
may be made to turn.

The higher types — i.e.^ the annelids, gasteropods, and lamel-
libranchs— have for some time been known to agree closely in
the general value of the quartets. Rabl first demonstrated
that in Planorbis the entire ectoblast is formed from the first

CELL-LINEAGE.

27

three quartets, while the mesoblast-bands arise from the pos-
terior cell of the fourth quartet, the other three, with the
remains of the primary quadrants, giving rise to the entoblast
(Fig. i). The same general result has been reached by sub-

FiG. I. — Diagram of the typical quartet-formation in an annelid or gasteropod; the quartets
numbered in the order of their formation ; A, B, C, D, the basal quadrants. Ectoblast
unshaded, mesoblast dotted, entoblast ruled in parallel lines. In many forms {e.g.,
Aricid) a fifth quartet (en toblastic) is formed; in others (^.^., iVier^w) only three com-
plete quartets and the posterior member of the fourth quartet {d^ or M).

sequent investigators of molluscan cell-lineage, though there
are one or two apparent exceptions {e.g.^ Teredo, according to
Hatschek) that demand reinvestigation. The same remarkable
fact holds true throughout the annelids,^ the well-determined

1 See footnote at p. 36 for reference to Eisig's widely divergent account *of
the development of Capitella.

28 BIOLOGICAL LECTURES.

exceptions being some of the earthworms and leeches referred
to above, in which the typical relations seem to have been dis-
turbed through changes in the nutrition of the embryo. Wher-
ever the typical quartet formation takes place — and this is the
case in nearly all the forms that have been adequately examined
— the general value of the quartets is the same, the first three
giving rise to the entire ectoblast, the fourth giving rise, one
cell to the mesoblast-bands and the other three to entoblast,
while the remnants of the primary quadrants, including the fifth
quartet if one is formed, give rise to the entoblast. This result
seems almost too simple and produces an impression of artifi-
ciality which may probably account for the reluctance with which
it has been accepted in some quarters ; but I think it is not too
much to say that few facts in embryology have been more
patiently studied or more accurately determined. The above
statement does not, however, contain the whole truth ; but
before completing it we may advantageously turn to the
development of the Tiirbellaria.

It was long since shown by researches, beginning with Hallez
and Gotte and culminating in those of Lang, that the cleavage
of polyclades shows an extraordinarily precise resemblance to
that of the annelids and mollusks. Taking Lang's work on
Discoccelis as a type, we find four quartets of cells successively
produced from the primary or basal quadrants, following exactly
the same law of displacement as in the higher types, assuming
the same arrangement, and in their subsequent subdivision up
to a relatively late stage following so exactly the plan of the
annelid o^gg that even a skilled observer might easily mistake
one for the other (Fig. 4, A). Despite this accurate agreement
in the form of cleavage, Lang's observations seemed to show
that the cell-quartets had a totally different value from those
of the higher forms; for he believed the first quartet to pro-
duce the entire ectoblast, the second and third to give rise to
the mesoblast, while the fourth quartet, with the basal cells,
formed the entoblast (Fig. 2, B). Such a result was more than
a stumbling-block in the way of the comparison. It was sub-
versive of the whole cell-lineage program ; for it seemed to
show that the cell-lineage of derivative animals {i.e.y annelids

CELL-LINEAGE.

29

and gasteropods), while exactly conforming to the 3.ncestr 3.1 form
of cleavage (i.e., that of the Tiirbellaria), differed toto ccelo from
it in morphological significance. When, some years ago, I first
called attention to this difficulty, I felt constrained to the ad-
mission that, in the face of such a contradiction, the study of
cell-lineage could only be regarded as of very restricted value
in morphological investigation ; indeed, in a lecture delivered
here four years ago on the inadequacy of the embryological

^'^v_:I_^

M

Fig. 2. — Diagrams contrasting the value of the quartets in an annelid or gasteropod {A) with
those of a polyclade according to Lang's original account {B). Lettering and shading as
in Fig. I. (The true proportions of the basal quadrants and the fourth quartet, which are
here misrepresented, are shown in Fig. 4. It is characteristic of the polyclades that the
fourth quartet-cells are greatly enlarged at the expense of the basal quadrants.)

criterion of homology,^ I cited this very case as representing a
climax in the contradictions of comparative embryology.

It is not rare in the history of science to find that fuller
knowledge may so change the point of view as to transform a
seeming difficulty into a pillar of support ; and it seems not
unlikely that such may be the case with the present one, though
some new difficulties have arisen which still await solution.
The new evidence relates, on the one hand, to the annelids and
mollusks, on the other hand to the polyclades ; and since on
both sides it tends to bridge a gap which once seemed hope-
lessly wide, I shall consider it in some detail. In approaching

1 " The Embryological Criterion of Homology.
Lectures, 1894, p. 113.

Wood's Holl Biological

30 BIOLOGICAL LECTURES.

this evidence the two principal difficulties should be clearly
borne in mind. The first lies in the fact that the mesoblast-
bands of the annelids and mollusks arise from one cell of the
fourth quartet, while in the polyclade the mesoblast was stated
to arise from all of the eight cells of the second and thii'd quar-
tets. The second difficulty relates to the ectoblast, which in
the annelid and mollusk arises from the twelve cells of the
first, second, and third quartets ; while in the polyclade it was
believed to arise solely from the first quartet (Fig. 2). We
may consider these two difficulties in order.

As regards the first point, a series of researches during the
past three years have shown that in some of the mollusks and
annelids the mesoblast has a double origin, a part — and usually
the major part — arising from the posterior cell of the fourth
quartet, as stated above, while a part arises from cells of the
second or third quartet, as in the polyclade (Fig. 3). The major
part — which, for reasons that will appear beyond, I propose
to call the entomesoblast — gives rise to the so-called mesoblast-
bands. The minor part, or ectomesoblast {" secondary meso-
blast," "larval mesoblast," of various authors), apparently does
not contribute to the formation of the mesoblast-bands, and
in at least one case — namely, that of Umo, as described by
Lillie — it gives rise to cells of a purely larval character and
designated as "larval mesenchyme." The first step in this
direction was that of Lillie, just referred to, who in 1895 an-
nounced the discovery that in a lamellibranch, Unio, one cell
of the second quartet {a'^ on the left side) gives rise not only to
ectoblast, but also to a single mesoblast-cell which passes into
the interior, divides, and gives rise to some of the larval mus-
cles ("larval mesenchyme," Fig. 3, C). Lillie's discovery was
quickly followed by the no less interesting one of Conklin that
in another mollusk, the gasteropod Crepidula, three cells of the
second quartet, median anterior, right, and left (<^% ^% d^), like-
wise give rise to mesoblastic as well as to ectoblastic elements
(Fig. 3, B), — a process still more forcibly recalling the origin
of the mesoblast in the polyclade.

Two years later mesoblastic cells were found, both in the
mollusks and in the annelids, to arise from members of the

CELL-LINEAGE.

31

third quartet. The first of these cases was observed by Wier-
zejski (1897) in the case of PJiysa, where the two anterior cells
of this quartet {C^, b^) give rise to mesoblastic as well as to
ectoblastic cells, and exactly similar facts were soon afterwards

Fig. 3. — Diagrams illustrating the value of the quartets in a polyclade {Leptoplana), a lamelli-
branch (f/wzV?), and a gasteropod (Cr^//</7^/a). Lettering and shading as in Fig. i. (For
comment on these figures see footnote at first page.) A, Leptoplana, showing mesoblast-
formation in the second quartet. (Cf. Fig. 4.) B, Crepidula, showing source of ecto-
mesoblast (from a^, b^, c^) and entomesoblast (from quadrant D). C, Unio, ectomesoblast
formed only from a^.

observed by Holmes in Planorbis. Simultaneously with these
researches I independently discovered in the annelid Aricia
two mesoblast cells arising from the two posterior cells of
either the second or the third quartet {i.e., from c^ and d^ or

32

BIOLOGICAL LECTURES.

from d"" and c^), though I could not positively determine which.
This was immediately followed by Treadwell's discovery that
in the annelid Podarke mesoblast cells are formed from three
cells of the third quartet, namely, the anterior median and the
two lateral cells {cfi, c^, d^). It was thus shown that in at least
four genera of mollusks and two of annelids a part of the meso-
blast has an origin which recalls that of the polyclades, and the
view is irresistibly suggested that the formation of this ecto-
mesoblast in one, two, or three quadrants in the higher types
is a vestigial process or ancestral reminiscence of what occurred
in all four quadrants in the ancestral prototype and still persists
in the polyclade.

The second difficulty — i.e.^ the origin of the ectoblast — has
entirely disappeared upon a reexamination of the cell-lineage of
a polyclade {Leptoplana) which I was enabled to make in the
summer of 1897. In this form careful study shows in the
clearest manner that the formation of ectoblast is not confined
to the first quartet, but that all of the twelve cells of the first
three quartets contribute to the ectoblast, precisely as is the
case in the annelids and mollusks (Fig. 3, A; Fig. 4, for details).
A comparison with Lang's figures gives every reason to believe
that the same is true in Discocoelis and the other forms studied
by him, and that on this point he fell into an error which was
certainly very pardonable at the time. The quartet-cells from
which in the polyclade the mesoblast arises are, therefore, not
pure mesoblasts, as Lang supposed, but are mesectoblasts, pre-
cisely like the cells from which the "larval mesoblast " arises
in Crepidiila or Unio}

The researches reviewed up to this point have cleared up the
contradiction relating to the second quartet. Passing now to
the third and fourth quartets, we find that the newer re-
searches have introduced a new difficulty with respect to each
of these quartets ; but the new difficulties differ from the old in

^ In Leptoplana each cell of the second quartet divides off in succession three
ectoblast cells before the delamination of mesoblast into the interior occurs at the
fourth division (Fig. 4). In Unio, according to Lillie, the larval mesoblast is
definitely separated at the third division of the micromere {a^). Professor Conklin
informs me that in Crepidula the ectomesoblast is formed at about the fourth or
fifth division of the micromeres {a^, b^, c^).

CELL-LINEAGE.

B

Fig. 4, Leptoplana. — (Camera drawings from the transparent living embryos. In these figures
the subdivisions of the micromeres are accurately shown .)

A, 32-cell stage, from the upper pole ; B, 36-cell stage, from the side, showing second division
of 2 ; C, side view, approximately 60 cells, showing the third ectoblast cell (23) derived from
2, the fourth quartet (4), and the basal entoblasts {D, C). D, delamination of mesoblast in
the fourth division of 2 (shaded), from the lower pole, showing the basal quartet of entomeres
(A-D), and the two somewhat unequal cells (4*/', 4^^) formed by the vertical division of the
posterior cell of the fourth quartet. £, posterior view of ensuing stage, showing the two pos-
terior mesoblast cells (shaded) lying in the interior, and a marked inequality between (4^^
and 4^/2). f, later stage ; multiplication of the mesoblast-cells (shaded), equality of 4^^ and
4d^, as in Discocoelis.

34 BIOLOGICAL LECTURES.

that they suggest a number of highly interesting problems for
future research. As regards the third quartet I was unable to
find in Leptoplana any evidence that it gives rise to mesoblastic
elements such as we should expect to find in the Turbellaria in
view of the formation of ectomesoblast from this quartet in
Physa, PlanorbiSy Podarke, and probably in Aricia. As far
as I could find, the third quartet gives rise only to ectoblast
cells at the lip of the blastopore (Fig. 4), and Lang's results seem
to me inconclusive on this point. Only renewed researches
can determine whether this difficulty be real or only apparent.
In the mean time it would be well not to lose sight of the fact
that the polyclades cannot, of course, be the actual ancestors
of the annelids and mollusks, and that the cleavage in the
former may differ very considerably from the common ancestral
type. A natural hypothesis is that in the ancestral mode of
development all of the first three quartets gave rise both to
ectoblast and to mesoblast, and that in all the existing forms
the mesoblast formation has been lost in the first quartet and
variously reduced or entirely suppressed in one or both of the
two succeeding quartets. I think, therefore, that we need not
hereafter be surprised to find the formation of ectomesoblast
from more than one of the first three quartets, whether in the
Turbellaria or in the higher forms.

It is when we attempt to bring the foregoing considerations
into relation with the history of the fourth quartet in annelids
and mollusks that we arrive at a far more serious difficulty ; but
we can hardly regret a difficulty that is so suggestive of further
research. In the polyclade the fourth quartet is relatively very
large, the basal quadrants being correspondingly reduced (Fig. 4).
All of the eight cells formed give rise, as far as known, to ento-
blast only. In the annelids and mollusks, on the other hand,
only three cells of this quartet — anterior, right, and left — are
purely entoblastic, while the fourth, or posterior, cell ("d^*")
divides into symmetrical halves to form the "primary meso-
blasts," or pole-cells, from which arise the two mesoblast-bands
characteristic of these groups (Fig. 2, A). Now, in comparing
this mode of development with that of the polyclade, we must
choose between the following alternatives. Either the meso-

CELL-LINEAGE. 35

blast of the annelid or mollusk, as a whole, corresponds with
that of the polyclade — in which case we must assume that in
the course of the phylogeny the posterior cell of the fourth
quartet has gradually taken upon itself more or less completely
the mesoblast formation formerly occurring in the second or
third quartet ; or the mesoblast of the polyclade has dwindled
away, perhaps has even disappeared, in the higher forms, where
it is represent-ed only by the ectomesoblast, its place having
been taken, through a process of substitution, by the mesoblast-
bands derived from the fourth quartet. To vary the statement
we must assume that a substitution has taken place either in the
cell-mechanism by which the mesoblast is formed or in the meso-
blast itself, and upon our choice between these alternatives de-
pends the entire point of view from which we regard cell-lineage.

Now, it must be admitted, forthwith, that we have not at
command sufficient data to give any certain answer to this
question, and we should be careful not to draw premature
conclusions in a matter which involves further consequences
of such importance. But there are a number of well-ascertained
facts drawn from widely diverse sources that point towards the
second of the above alternatives ; i.e., the view that the mesoblast-
bands of the annelid or gasteropod are not as such represented at
all in the polyclade, but, phyletically considered, are neomorphs
which have more or less completely replaced the ancestral meso-
blast. This evidence may be arranged in three lines : —

I. As a result of exact and thorough studies upon the his-
tology and larval development of the annelids, Eduard Meyer
was several years ago led to the conclusion that the mesoblast-
bands, both in origin and in fate, differed widely from the scat-
tered larval mesenchyme-cells, though the lineage of the latter
was then unknown. Developing this idea, Meyer was led to
the remarkable conclusion that the mesoblast-bands of the
higher types represent the paired gonads of the ancestral form
— a view nearly related with the earlier one of Hatschek, that
the primary mesoblasts were originally eggs, which, in the
course of the phylogeny, became in part transformed into peri-
toneal and other somatic cells, and in part remained as germ-cells.
Thus the original mesoblast — which Meyer definitely compared

36 BIOLOGICAL LECTURES.

with that of the Ttirbellaria — was gradually replaced, though
still persisting in a reduced form as the larval mesenchyme.

I would not at present urge the acceptance of this daring
hypothesis ; but in the light of later research it has become
highly significant, and whether true or false is of great interest
as giving a clear picture of how such a process of substitution
may have been possible.

2. In the second line of evidence lies Lillie's discovery that
the ectomesoblast of Unio (derived from a-) gives rise to purely
larval transitory structures; namely, to the adductor muscle
and the scattered contractile myocytes of the Glochidhim larva.
In the annelids, too, the same conclusion seems probable, and
my friend Professor Tread well informs me that in Podarke
there is every reason to believe that the ectomesoblast (derived
from ^3j ^3^ </3) is entirely devoted to the formation of the ring-
muscle and myocytes of the trochophore, which apparently take
but an insignificant part, if any, in the building of the adult body.
This result tallies with the view that the ectomesoblast forma-
tion in the higher types is a reminiscence of the ancestral process
still existing in the polyclade, but in the higher forms relegated
to the early stages, and even in them is more or less reduced.^

1 Eisig has very recently {Mitth. Zool. Station, Neapel, xiii, i, 2, 1898) published
the results of a study of the cell-lineage and later development of Capitella, which
are totally at variance with the view here suggested, and the facts on which it is
based. Broadly speaking, his results exactly reverse those of all the authors cited
above, the mesoblast-bands (" Coelomesoblast ") being derived from the third quartet
(r^-* and ^^•^), while the larval mesoblast ("Paedomesoblast ") arises from a portion
of M (d^), the remaining portion giving rise to ectoblast. If well founded, this
result is not only fatal to the view I suggest, but is, I believe, nothing less than a
reductio ad absurdum of the whole cell-lineage program, regarded as a method of
morphological research. No one will lightly call in question the results of so
conscientious and eminent an observer ; and they must be regarded as by far the
most serious obstacle that the morphological study of cell-lineage has thus far
encountered. I will not attempt to explain away this adverse evidence, based on
so prolonged and thorough a research. It should not be forgotten, however, that,
as Professor Eisig is himself careful to point out, the nature of the material has
forced him to contend wath great difficulties, since the eggs are normally distorted
by pressure (a factor which, as I have experimentally shown in Nereis, may greatly
alter the form of cleavage) between the membranes of the tube ; and, further, the
development cannot be continuously followed in life. A result, based on this
material, which stands in such flat contradiction to what is known in other and
more favorable forms, must await the test of further research.

CELL-LINEA GE.

?>7

3. In the third line lies the evidence, recently obtained, that
the pole-cells or teloblasts of the mesoblast-bands of the anne-
lids and mollusks are to be regarded as derivatives of the
archenteron, and hence differ wholly from the ectomesoblast in
their relation to the primary germ-layers. Kowalevsky, the dis-
coverer of these teloblasts, expressed the opinion, more than
twenty-five years ago (1871), that they were to be regarded as
derivatives of the archenteron ; and a large number of later

Fig. 5. — Diagrams comparing the early divisions of the posterior cell of the fourth quartet
{d* or M) in Crepidula («), Nereis {b), Clymenella (c), Unio {d ), and Aricia {e). The
numerals show the order of division. Cells destined to form entoblast (their fate as actu-
ally observed in Crepidula and Nereis, but only inferred in the other cases) ruled in
parallel lines, mesoblast dotted. After the divisions here shown, the symmetrical mesoblast-
bands are formed from the dotted cells.

workers from the time of Rabl (1876) have accepted his view,
though only very recently has the full strength of the evidence
been developed. In the first place, it was shown through the
studies of Rabl, Blochmann, and later workers, that while the
posterior cell of the fourth quartet gives rise to the mesoblastic
pole-cells the other cells are purely entoblastic. In the second
place, the recent studies of Conklin and myself have shown that
even the posterior cell of the fourth quartet (^4) may contain
entoblastic as well as mesoblastic material. I showed several

38 BIOLOGICAL LECTURES.

years ago that in Nereis each of the cells into which d^ divides
buds forth several small cells (Fig. 5, b), which do not enter
into the mesoblast-bands, though I did not correctly determine
their fate. More recently Conklin was able to show that a
similar process occurs in Crepidula (Fig. 5, a), and that the
cells thus formed are entoblast-cells which enter into the forma-
tion of the archenteron. On reexamining the matter in Nereis
I found the clearest evidence that the same was true here. In
both these cases, therefore, the posterior cell of the fourth
quartet is of mixed character, and divides into two mesento-
blasts, each of which first gives rise to a number of entoblast
cells (two in Crepidtila, four or five in Nereis), the residue con-
stituting the mesoblast. In both these forms, therefore, the
ectoblast (and in Crepidula the ectomesoblast) are first com-
pletely segregated, and the archenteron which remains gives
rise to the mesoblastic pole-cells. The latter are, therefore, of
entoblastic rather than ectoblastic origin, and may be designated
as the entomesoblast.

Further examination of these phenomena brings out some
highly interesting facts which seem to constitute a striking
case of ancestral reminiscence in cleavage. Several years ago
I found in two genera of annelids, Aricia and Spio, that the
small entoblast-cells of Nereis and Crepidula {i.e., those budded
forth from the two mesoblasts derived from the division of dA
or M) are represented by a single pair of quite rudimentary
cells, scarcely larger than polar bodies (Fig. 5, e), which ap-
parently take no part in the building of the archenteron, and
can only be explained as vestiges or reminiscences of such a
process as occurs in Crepidula or Nereis. Later researches
have revealed the presence of these vestigial entoblasts in
several other forms, and have shown further that they are con-
nected by several intermediate steps with the larger functional
cells found in Crepidula. Thus in Amphitrite (Mead) and
Planorbis (Holmes) they are quite vestigial, agreeing essen-
tially in size and origin with those of Aricia. In Unio (Lillie)
they are considerably larger (Fig. 5, d), in Clymenella (Fig. 5, c)
they are as large as the mesoblastic moiety (Mead) ; while in
Crepidula (Fig. 5, a) their bulk surpasses that of the meso-

CELL-LINEAGE. 39

blastic part.i Such a series creates a strong probability that we
have before us a vanishing series like those so well known in
adult organs, such as the limbs, the tail, or the teeth. Further,
just as the lateral toes of the horse seem to have wholly vanished,
even from the ontogeny, so the vestigial entoblasts would seem to
have disappeared in some annelids and mollusks, leaving the
posterior cell of the fourth quartet purely mesoblastic.

These considerations invest with a special interest the cor-
responding cell in the Ticrbellaria {i.e., the posterior member
of the fourth quartet, 4^) ; and this interest is heightened by
Lang's discovery that in Discoccelis this cell divides earlier
than the other cells of the quartet, and into equal halves which
lie symmetrically at the posterior end of the embryo. These
two cells thus correspond exactly in origin and position with
the paired mesentoblasts of the annelids and gasteropods, and
the facts naturally led to the suggestion, made by Mead, that
they would perhaps be found to give rise to paired mesoblast-
bands, as in the higher types. In Leptoplana (Fig. 4, D, E, F)
a similar division occurs, but as far as their fate is concerned
my own observations do not sustain Mead's suggestion, on the
one hand giving no evidence that these cells give rise to any-
thing other than the posterior cells of the archenteron, on the
other showing that they are often unequal or asymmetrically
placed (Fig. 4, D, E) and only rarely conform to Lang's scheme
(Fig. 4, F). If, therefore, the polyclades represent the ances-
tral type in this respect, we must conclude that the entomeso-
blast was a later development. The remarkable fact is that, if
such has been the case, the new mesoblast-formation has been
fitted, as it were, upon an old form of cleavage occurring regu-
larly in Discoccelis and occasionally in Leptoplana. The two
symmetrical posterior entoblast-cells of the polyclade might
thus be conceived as the prototypes of the primary mesoblasts
or mesentoblasts of the higher forms, which in the course of
the phylogeny undertook the formation of mesoblastic as well
as of entoblastic elements.^ The old building pattern was still

1 1 am here placing my own interpretation on Mead's and Lillie's observations.

2 Lang has pointed out a motive for this form of cleavage in the polyclade,
correlating the early and symmetrical division of ^^^ with the posterior bifurcation
of the gut.

40 BIOLOGICAL LECTURES.

retained but adapted to a new use, precisely as has been the case
with the evolution of larval or adult organs, such as the branchial
or aortic arches and the limbs. As the change progressed the
posterior cell of the fourth quartet became more and more
strictly given over to the formation of mesoblast, its entoblastic
elements becoming correspondingly reduced to truly rudiment-
ary or vestigial cells {Aricia^ etc.), or finally, perhaps, disap-
pearing wholly.

I have endeavored to place these special conclusions in
strong relief, not because they can yet be accepted as demon-
strated,— and it is quite possible that some other interpreta-
tion may yet be placed upon some of the facts, — but because
they seem to me highly suggestive of further research in the
field of cell-lineage. There are among them two general con-
siderations on which I would lay emphasis.

First, the study of cleavage or cell-lineage in the case of
these groups raises a number of highly interesting and sug-
gestive questions in pure morphology. If the mesoblast-bands
are a new formation, what is the motive, so to speak, for their
origin.? Did they perhaps arise through the development of- a
new body-region, or a new growth-zone, or budding-region from
the posterior part of the ancestral body, as has been assumed
by Leuckart, Haeckel, Hatschek, and Whitman in explanation
of metamerism.? Is the body of the turbellarian homologous
to the entire body of an annelid or mollusk, or does it repre-
sent only the head or the larval body, to which a trunk-region
is afterwards added.? What is the relation of the entomeso-
blast to the archenteric pouches of the enterocoelous types.?
How do the above results harmonize with the general doctrine
of development by substitution.? These are examples of some
of the morphological questions suggested by the general in-
quiry. They are admittedly of a highly speculative character,
and I, for one, am not prepared to give a positive answer to
any of them. But the mere fact that morphological questions
of such character and scope are inevitably suggested by studies
in pure cell-lineage shows that such studies must not be passed
over by the morphologist as having no interest or value for his
own researches.

CELL-LINEAGE. 4 1

Second, the phenomena we have considered seem to leave
no escape from the acceptance of ancestral reminiscence in
cleavage, with all that that implies. That the rudimentary
entoblasts of Aricia or Spio are such reminiscences of former
conditions seems almost as clear as that the mammalian yolk-
sac or the avian primitive streak are such. The formation of
the ectomesoblast in annelids and mollusks is nearly if not
quite as strong a case. Both these are processes that appear
to be vestigial, or, at any rate, approach that character. But the
evidence of genetic affinity is no less clearly shown in proc-
esses that are not vestigial, such as the formation of the ecto-
blast in Tiirbellaria^ annelids, and gasteropods or lamellibranchs,
from neither more nor less than three quartets of micromeres,
or in the origin of the archenteron from the fourth quartet
with the remains of the basal quadrants. Between the anne-
lids, gasteropods, and lamellibranchs a far more precise and ex-
tended series of resemblances exists. The question has been
much discussed of late whether such resemblances can be called
homologies. Probably no one will deny that the ectoblast-cap,
arising from twelve cells, is as a whole homologous in the an-
nelid and the gasteropod embryo. Are the individual micro-
meres respectively homologous.-* In the present state of our
knowledge this is a question of name rather than of fact ; for
homologies only gradually emerge during development from
their unknown background in the ^g'g. It is for this reason
that, as I have urged in a preceding lecture, the tdtimate court
of appeal hi this question lies i7t the fate of the cells. If the
structures to which they give rise are homologous, I can find
no logical ground for refusing the claim to the cells from
which they arise. Furthermore, this homology must be irre-
spective of the origin of the cells, just as the ganglion of a bud-
embryo of Botrylhcs is homologous with that of an egg-embryo
in the same form, despite the total difference of origin in the
two cases. When, however, we find that the homologous pro-
toblasts or parent-cells have the same origin as well as the
same fate, the homology becomes the more striking ; and it is
in the determination of common origin as well as common fate,
as has been done in so many cases, that the principal signifi-

42 BIOLOGICAL LECTURES.

cance of recent work in cell-lineage seems to me to lie. Some
of the objections urged against the reality of cell-homology
have, I think, arisen through a failure to recognize among cell-
homologies the same distinction between complete and incom-
plete homology that was long ago urged by Gegenbaur in the
case of organ-homologies. The posterior member of the fourth
quartet in annelids, for example, is in a broad sense homolo-
gous throughout the group ; but the homology is probably not
an absolute or complete one, since this cell may contain func-
tional entoblast {Nereis), rudimentary or vestigial entoblast
(Aricia), or apparently in some cases no entoblast, as I have
described in Polymnia. Again, the acceptance of cell-homology
does not, I think, carry with it the necessity of finding a homo-
logue for every individual cell throughout the ontogeny ; for
in the case of later structures no one demands or expects that,
in the comparison of related forms, an exact equivalent shall be
found for every subdivision of homologous nerves or blood-
vessels or sense organs. Finally, the fact that cleavage may
show no constant or definite relation to the adult parts — as is
the case in the teleost fishes — does not alter the equally in-
dubitable fact that cleavage often does show such a constant
relation. The probability that the Nauplitis larva is not a true
ancestral form does not come into collision with the probability
that the ascidian tadpole is such a form. How far in the
course of phylogeny the ontogeny has adhered to its original
type and retained the same relation to the adult parts is a ques-
tion which stands, as far as I can see, both a priori and a pos-
teriori on essentially the same basis, whether it be applied to
the cleavage or to the later stages. Let us not forget the
difficulties that still beset us in the application of the biogenetic
law to the larval stages and to general organogeny,- and let us
not make a greater demand in this regard upon cell-lineage
than on other lines of embryological research. The time has
not yet come for a last word on this subject, and we shall
probably have to await the result of much more extended re-
search before a satisfactory point of view can be attained.

THIRD LECTURE.

ADAPTATION IN CLEAVAGE.

FRANK R. LILLIE.

I. Introduction.

It has happened, very naturally, that writers on the subject
of cell-lineage have laid special emphasis on the resemblances,
which are nothing short of marvellous from the older points
of view, between the cleavage of the eggs of even widely
separated forms. Over and over again it has been demon-
strated that in gasteropods, lamellibranchs, annelids, and tur-
bellarians, the ectoderm is made up of three quartets of cells,
formed from the first four cells ; that the fourth product of the
left posterior macromere contains the mesoblast, excepting in
the turbellarians ; that the turbellarian mesoblast is represented
in the other groups ; that the four basal cells, after the separa-
tion of the ectoderm and mesoblast, form endoderm ; that the
ectoderm of the entire trunk is derived from d^ (speaking
technically), and that there is a wide-reaching sameness in the
cell-lineage of the prototroch, cross, and of other larval organs.
Ancestral reminiscences even have been discovered in the
cleavage (E. B. Wilson, 14). I do not in the least underesti-
mate the immense value of these results, which form one of
the most brilliant and interesting chapters in modern embryol-
ogy. But the tendency to schematize has naturally arisen, and
one of the most instructive aspects of cell-lineage is thus lost
sight of, namely, the special features of the cleavage in each
species, which are, I believe, as definitely adapted to the needs of
the fntnre larva as is the latter to the actual conditions of its
environment.

43

44

BIOLOGICAL LECTURES,

This principle is meant to apply only to the '' determinate type
of cleavage " (Conklin), in which larval, and hence adult struc-
tures, can be shown to have a definite cell-lineage extending
back to the unsegmented ovum. The illustrations and argu-
ments will be drawn, first, from the cleavage of the ^g<g oJ[ a
fresh-water lamellibranch, UniOy which I have specially studied,
and then, to secure a broader foundation in fact, some signifi-
cant variations in the cleavage of the eggs of annelids will
be considered. This theory was first used, I believe, for the
explanation of determinate cleavage in my papers on the
" Embryology of the Uniofiidce " (6 and 7), where I pointed
out that the size of the cells and the rate and direction of
their cleavages **are ruled by the needs of the embryo," and
concluded: "The peculiarities of the cleavage in Unio are but
a reflection of the structure of the glochidium, the organiza-
tion of which controls and moulds the nascent material." In
his lecture on " Cleavage and Differentiation " {Biological Lec-
tiireSy 1896-97), Conklin has maintained a similar view, viz.^
"that all differential cleavages . . . are directly and causally
related to the uses to which these cells are put." In the
second part of this paper the retrospective significance of
this principle will be considered in the light of some observa-
tions on the maturation, fertilization, and first two cleavages
of the ^g'g of Unio.

If we are to understand the special features of the cleavage
in Unio, which are, I believe, purely adaptive, it will be neces-
sary, first of all, to describe the peculiar larva. The glochidium
(Fig. i) possesses two shell-valves, shaped something like the
bowls of two spoons cut across near their bases, and united by
a central hinge. The curvature of the anterior margin of each
valve is somewhat greater than that of the posterior margin,
and each lateral angle is provided with a strong hinged hook,
which is secondarily toothed in Afiodonta on its outer face.
The valves may be closed with great force by a single strong
adductor muscle, the main bulk of which is somewhat anterior
to the middle of the body. The interior of the spoon-shaped
valves is lined by the larval mantle, composed of very large five
or six sided cells, in the exact centre of which is the opening

ADAPTATION IN CLEAVAGE. 45

of a long tubular gland, the thread-gland, coiled several times
round in the right valve of the shell. The secretion of this
gland takes the form of a long thread, which can, apparently,
be respun when lost, and which floats freely in the water. There
are four tufts of sense-hairs on each side of the larval mantle,
three arranged in the form of a right-angled triangle just within
the outer angle of each valve, and one on either side of the
middle line, a trifle behind the opening of the thread-gland.

Now all of the structures mentioned thus far, the shell, the
larval mantle, the adductor muscle, the thread-gland, and the

l.m.

Fig. I . — Glochidium of Anodonta, one of the Unionidce,\y'\n%ox\. its dorsal surface, with the
shells expanded so as to show the soft parts. A, anterior; P, posterior; R, right; L,
left ; S, shell ; H, hook ; /. m., larval mantle ; /. a., larval adductor muscle ; t.g., thread-
gland ; t., thread ; s.h., sensory hairs ; tn., mouth ; E. A., embryonic area.

tufts of sense-hairs, are purely larval organs, destined to form
no part of the adult, but to degenerate. They are already old
when the life of the individual has hardly begun, and they form
by far the greater bulk of the larval body. The embryonic
material, from which the body of the adult is to develop, is all
included within a small area {E. A.) just in front of the poste-
rior angle of the valves of the shell. This area is only about
one-third of the length of the glochidium, and less than a third
of its width, when the valves are expanded ; yet within it we
find the rudiments of the intestine with minute liver-diverticula
and end-gut well marked but no opening to the exterior, of
the mesoblastic structures, and of the foot, mantle, and gills.

46 BIOLOGICAL LECTURES.

All these parts are entirely functionless, and great only in
potentialities.

The glochidium is a larva highly specialized for a peculiar
mode of development. When the larva is extruded into the
water by the mother-clam, it falls to the bottom and lies there
with its thread floating in the water, violently snapping its
valves together at the least disturbance. It is really a most
sensitive little creature, and so readily perceives any fish swim-
ming by. If it be so fortunate as to have its thread caught on
the fish, it at once attacks it with its hooks. It is, of course,
only on the unprotected parts of the fish, such as the fins or
the gills, that the larva is able to make any impression ; and it
is stated that the larvae of Anodonta usually become parasitic on
the fins, and those of Unio on the gills. The irritation caused
by the presence of the larva sets up a proliferation of the epi-
dermis, which speedily encloses the little parasite in a sort of
cyst, within which the later development takes place.

II. Cleavage of the Egg of Unio.

Adaptation in cleavage can manifest itself only in the three
possible modes of cleavage variation, which are, as has been
pointed out by Mead, Jennings, zur Strassen, and others, first,
differences in the rate of cleavage of cells ; second, differences
in size ; and third, differences in the direction of the cleavage
or position of the resulting parts. Let us now examine the cleav-
age of the Qgg of Unio with these general principles and the struc-
ture of the larva in mind. I may as well state at once that I
regard the general form of the cleavage as inherited from a long
series of ancestors, extending back, probably, to the Tnrbcllaria.

The general plan of the cleavage is the same as in annelids,
Turbellaria, and most mollusks, described in other lectures con-
tained in this volume. From the first four cells, separated by
meridional planes, the ectoderm is separated by three slightly
oblique cleavages horizontal in their general direction, the first
being dexiotropic, the second laeotropic, and the third dexiotropic
again. A fourth division of the left posterior macromere (basal
cell) forms the mesoblast cell. (See Figs. 2-5.)

ADAPTATION IN CLEAVAGE,

47

The first cleavage is meridional and very unequal, although the
distribution of the yolk throughout the cytoplasm is perfectly
uniform in the form of small granules (Fig. 20). The larger
cell marks the position of the posterior end of the embryo. This
cell divides before the smaller one, and unequally, the larger
product again being posterior in position. The smaller cell of
the two-celled stage then divides somewhat unequally, so that its
larger product lies on the left side of the embryo. Thus we reach
a four-celled stage (Fig. 2)
composed of one very large
cell, D, posterior in position,
and three smaller cells, of
which the left one, A, is the
largest. The anterior and
posterior cells, B and D, meet
in a broad cross-furrow at the
animal pole (Fig. 3), the cross-
furrow at the vegetative pole
being' much smaller, contrary
to the general rule in eggs of
this type of cleavage. Thus
the first two cleavages illustrate the three possible modes of
variation, viz.^ inequality of the cleavage products, difference in
the rate of cleavage, and difference in the position of the cells,
indicated by the broad cross-furrow at the animal pole. What
is the meaning of these differences.'*

a. Relative Rates of Cleavage of the Constituent Cells. — Let
us place side by side a table of the constitution of the thirty-
two-celled stage of the ^gg of Unio and of an ideal ovutn fol-
lowing, in direction of cleavage, the ordinary spiral type, with
equal cleavage throughout, and the same rate of division in all
the cells.

First generation of ectomeres
Second generation of ectomeres
Third generation of ectomeres
Fourth division of D (mesoblast cell)
Entomeres

Fig. 2. — Four-celled stage of Unio complanata
seen obliquely from the left side and above.

DEAL Ovum.

Unio.

16

ID

8

13

4

4

I

32

32

48 BIOLOGICAL LECTURES.

It is plain that as regards the relative rate of cleavage of the
constituent cells, the ^^^ of Unio departs very far from the ideal
condition. Let us attempt to analyze this and see if these vari-
ations in the rate of cleavage are adaptive or not. The first
thing we notice is that the number of the first generation of
ectomeres is ten instead of sixteen ; this means that these cells
have divided more slowly than the average rate. Now in most
mollusks and annelids they form the prototroch and the entire
region in front of it, with the apical plate in the centre. But
in Unio this region is degenerate, and I have never been able
to find any trace of prototroch or apical sense-organ. It is
plain, then, that the slow rate of cleavage is anticipatory of
this, and is, in this sense, an adaptive modification. But why
ten cells in place of eight .-* This means that two of the cells,
c^ and ^S of this group have divided once oftener than their
fellows. These cells occupy the posterior portion of the upper
hemisphere, and later enter into the formation of the head-
vesicle, and I believe that their more rapid cleavage is an early
indication of their subsequent fate.

In the second generation of ectomeres we have a very dif-
ferent set of conditions as to number ; the ideal number of cells
of this generation in the thirty-two-celled stage is eight, but in
Unio thirteen. This increase above the ideal number is due
almost entirely to the rapidly succeeding divisions of one cell
of this generation, d'^, but in part also to another, a'^. The
other two cells, the anterior and the right members of the
quartet, fall in with the ideal scheme, each having divided
once. In this stage d^ has broken up into six cells, a- into
three. Is there anything adaptive in this hastening of the
process of cleavage } d^ is the cell from which the entire ecto-
derm of the trunk region is formed. The most characteristic
thing about this region in the embryo of Unio is the relatively
enormous size and very early formation of the shell-gland, and
we have seen that the shell of the glochidium is of very special
importance to it, not only as an organ of defence, but also of
offence. Is it not wonderful that this important part should
be represented, not only by the largest, but also by the most
rapidly segmenting cell in the embryo, as though the ^gg knew

ADAPTATION IN CLEAVAGE.

49

what was required of it, and acted accordingly ? Can we deny-
that the rate of cleavage of this cell is an adaptive feature of
the cleavage ? The other cell of the second generation that
divides more rapidly than the average rate is ^2-2^ or Y, the lower
product of the first division of a~. In the thirty-two-celled
stage (Fig. 5) this cell already projects some distance into the
segmentation cavity ; after budding off two more small cells at
the surface, it withdraws entirely within the segmentation cav-
ity, and divides equally into two cells, which arrange themselves
symmetrically on either side of the middle line. From these
cells of the larval mesoblast the adductor muscle of the glo-
chidium and certain isolated myocytes of the cleavage cavity are
derived. Now the adductor muscle is an extremely large organ,
of the utmost functional importance to the larva, and the varia-
tion in the rate of cleavage of the cells which are to form it
appears to be adaptive.

The third generation of ectomeres possesses the normal num-
ber of cells for this stage, but it is interesting to notice that
the posterior member, ^3, is formed much before the others,
owing to the tendency to more rapid division of the basal cell
from which it arises. This cell, D, has also undergone another
division, in excess of the divisions of the other basal cells, giv-
ing rise to the mesoderm proteloblast, d^ or M.

To summarize : groups of cells and even single cells vary
greatly in their rates of segmentation. Each acts as though
animated by an independent force. Moreover, these differ-
ences possess prospective significance, looking forward to the
final outcome ; and this phenomenon may fairly be called adap-
tation m the rate of cleavage.

But the objection may be raised : granted that the rate of
division varies in these cells, and that those segmenting more
rapidly form the earliest functioning parts, is it not possible
that the variation is induced from without, and that this deter-
mines the subsequent fate } So far as we know, the only
external factors which influence the rate of cleavage are tem-
perature, chemical constitution of the medium, and perhaps
some other general conditions of the environment, which must
act uniformly on the whole Qggy and hence cannot be held

50 BIOLOGICAL LECTURES,

accountable for differences in the rate of cleavage of constitu-
ent elements of the same ^g^. In addition to these factors,
the presence of a larger proportion of yolk in certain cells may-
determine a slower rate of cleavage. But in the ^g^g of Unio
the yolk is distributed with perfect uniformity among the cells,
so that each possesses practically the same relative quantity.
Is it not, then, the absolute amount of yolk that determines
the rate t No, unless one is willing to adopt the hypothesis
that yolk tends to hasten the cleavage, for it is the largest cell
in the ^gg that segments most rapidly. However, there is no
constant relation between size and rate of cleavage, so even
such an inverted hypothesis will not avail. There is, in fact,
no escape from the conclusion that the factors determining
differences in the rate of cleavage of the separate cells or defi-
nite cell-groups are of the same nature as those that deter-
mine differences between ova of different species, viz., purely
constitutional causes. In a very real sense, each cell in the
segmenting ^gg of Unio acts, in respect to rate of cleavage,
like an ovum in ovo}

b. Variations in Size. — Although difference in size of
daughter-cells in normal cleavage generally indicates difference
in quality or constitution, it is not always safe to conclude that
daughter-cells of the same size necessarily possess the same
constitution. In other words, a division may be differential,
even though the resulting cells are of the same bulk. The
adaptiveness of determinate cleavage is illustrated in a man-
ner even more striking than in differences of rate, by the
inequality of certain divisions. I believe that it can be shown
that the relative sizes of the cells in the early cleavages of the
^gg of Ufiio is adapted to the size and time of development of
the larval organs. The first two cleavages are unequal with the
result that two cells of the four-celled stage are larger than the
two other equal ones. The largest of all contains the material
of the two somatoblasts, and the next largest the material of
the larval mesoblast.

1 I believe that the entire egg is an undivided organism, and not a heap of equiva-
lent cells, each of which may occupy any position in the whole, as Hertwig and
Driesch have maintained. The above statement must be understood in this sense.

ADAPTATION IN CLEAVAGE.

51

Let us now look at the first four divisions of D, the largest
cell of the four-celled stage (Fig. 2). Its first division (Fig, 3)

Fig. 3. Fig. 4.

Fig. 3. — Eight-celled stage of Unio complanata from the animal pole.
Fig. 4. — Eighteen-celled stage of Unio complanata from the vegetative pole. The two cells

ruled with horizontal lines are products, the first somatoblast, d^ = X, which forms the

trunk, including the shell-gland and foot of the larva.

is very unequal, and the smaller product, perhaps not one-
tenth of the whole, lies nearer the animal pole ; it is one of the
first generation of ectomeres. The second division is likewise
unequal (Fig. 4), but here the
relations are reversed, for two-
thirds, at least, of the substance
of the cell passes into the upper
product, the first somatoblast, d~
or X] the third division is likewise
unequal {D and ^3, Fig. 4), and
this time again the smaller prod-
uct is uppermost, forming one of
the relatively unimportant third
generation of ectomeres. The
fourth division, finally, is extremely
unequal (Fig. 5), only a minute
portion of the cell remaining at
the lower pole, while the remain-
der forms the second somatoblast,
</4 or M, the proteloblast of the mesoderm. In each of
these unequal divisions there is a manifest adaptation, the

Fig. 5. — Thirtytwo-celled stage of Unio com-
planata from the vegetative pole. The
separation of the germ-layers is practically
complete in this stage. Products of first
somatoblast, d'^ = X, with horizontal lines ;
second somatoblast, d* = Af (mesoblast
cell), with vertical lines ; endoderm cells
with both vertical and horizontal lines.
F=:«2-2 larval mesoblast.

52 BIOLOGICAL LECTURES.

great bulk of the material passing into the two somatoblasts.
In the thirty-two-celled stage (Fig. 5) the largest cells are
those of the second generation of ectomeres, and this finds a
sufficient explanation in the fact that the material of the first
somatoblast, of the larval mesoblast, and of the larval mantle, is
found in these cells. The relatively small size of the first and
third generations of ectomeres and of the endoderm-cells is
correlated with the relatively small part that they play in the
upbuilding of the larva. The mesoderm proteloblast d^ is
actually smaller than the larval mesoblast Y (Fig. 5) because
of its deferred functional activity.

If there were any other explanation of these differences in
size of the cleavage products than this seemingly teleological
one, it might seem to simplify the problem of development.
However, all theories that seek to explain inequalities of cleav-
age by the presence of yolk or the action of any known external
force seem inadequate to explain the unequal cleavages in the
^gg of JJjtio.

c. Variations in the Direction of the Cleavage-Planes. — The
early divisions follow the law of alternating spiral cleavage
(Kofoid) for some time ; and I shall not stop to consider these,
although we are far from an adequate mechanical explanation
of this law. I shall take up, instead, certain special divisions,
which illustrate in a more striking manner the adaptiveness of
variations in the direction of cleavage. The first somatoblast
is established in the nine-celled stage by the second division of
D\ it is much the largest cell in the embryo, as we have seen,
and it goes through with an exceedingly characteristic series
of cleavages. It first buds off a small cell, x^, on the right
side of the vegetative pole, then another small cell, x-^ symmet-
rically placed on the left side, then a third small cell, ;ir3, in the
middle line, but towards the animal pole, then a fourth small cell,
x\^ in the middle line towards the vegetative pole (see Fig. 5); it
then divides equally and bilaterally, and each half buds forth
another small cell, xS, behind xA on the vegetative pole. I do
not know the fate of the cell xZ, budded forth towards the ani-
mal pole. However, the other small products form the rudi-
ment of the foot, which is an exceedingly small organ in the

ADAPTATION IN CLEAVAGE. 53

larva ; and the large cells, which henceforward divide equally,
the immense shell-gland. Can one imagine a more manifest
adaptation? The small cells are placed just where they are
needed, and are of the proper dimensions. Is there any other
reason why the somatoblast should not have divided equally
from the first, and the cells lying in one position have gone
into the foot, and the rest into the shell-gland .-* If this were to
happen, there would be no indication of adaptiveness in the
cleavage, for then it might seem possible that, simply because
of their different positions and conditions, part of the cells
formed the foot and part the shell-gland.

d. To summarize. — In the cleavage of the o^gg of Unio there
are marked variations in the size of the cells and in the rate
and direction of their cleavages ; in every case, these possess
prospective significance, and by means of them the organism
is able, so to speak, to realize, in the most direct manner
possible, on its available capital, the substance of the ^g'g. To
this principle I have given the name of adaptation in cleavage.

III. Adaptation in Cleavage in the Annelids.

In the annelids, as in the mollusks, the entire ectoblast of
the trunk is segregated in a single cell, d'^ or Xy known as the
first somatoblast, and the mesoblastic germ-bands in the second
somatoblast, dA or M. Now the cell-lineage of a great many
annelids is known through the studies of Whitman (11), Wilson
(13 and 14), Mead (9), Eisig (5), Treadwell (10), Child (i), and
others. It is possible to arrange these annelids in a series
according to the relative size of their somatoblasts. First there
are forms like Polygordius, Lepidonottis, Podarke, Hydroides, and
EiipomatziSy with equal cleavage, in which the somatoblasts do
not differ in size from the other cells of the same quartets (Fig.
6). Then in the order of increasing relative size of the somato-
blasts come Amphitrite (Fig. 7) and ChcetopteruSy Arenicola,
'Nereis limbata (Fig. 8), Clymenella (Fig. 9), Capitella, Aricia
(Fig. 10), Scolecolepis and Spio, Nereis dtimeriliiy and finally,
Clepsi7ie (Fig. 11). Now all of the forms with equal cleavage,
although most widely separated in relationships, possess trocho-

54

BIOLOGICAL LECTURES.

phores characterized by an almost equatorial prototroch, a very
large exumbrellar region, and relatively extremely slow devel-
opment of the trunk. I merely allude to this here, because
in another lecture of this series the causes of equal cleavage
in the annelids are considered in detail by Mr. Treadwell. At
the other end of the series comes Clepsine with its huge somat-
oblasts, in which the parts characteristic of the trochophore
are reduced to a mere rudiment, and the entire trunk develops
on the surface of the ^g^. Between these two extremes come
the other forms mentioned, and it may be said in general of
them, that there is a gradual acceleration in the time of devel-
opment of the trunk very nearly in proportion to the increase
in relative size of the somatoblasts. Treadwell has elsewhere
called attention to part of this series, and has come to the con-
clusion that '' the extra amount of material stored in cell D of
Nereis^ Amphitriie^ etc., is in some way related to the need for
an extra amount of somatic and mesoblastic material in the
young larva."

IV.

But in order to show that the adaptation need not run
in the one direction of the trunk, let me cite one more case,
which I am able to add through the kindness of Dr. E. B.
Wilson, from some unpublished observations of his on the cell-
lineage of a Nemertean. Here the four upper cells of the
eight-celled stage are larger than the four surrounding the
vegetative pole, the only case of this sort known, I believe.
We should expect on a priori grounds that the resulting larva
would possess a large helmet in front of the ciliated band,
which marks the posterior boundary of the products of the
first quartet of ectomeres, with a hardly developed trunk region
and a rudimentary archenteron. Such is, in fact, the case. The
enormous development in the pilidium of the exumbrella is pre-
delineated in the unique third cleavage.

I do not believe that adaptation is one whit less far-reaching
in the cleavage than in the larva. In fact, if the principle
which I am defending be correct, the two must be coextensive ;
that is to say, adaptation in cleavage is no more caused by

ADAPTATION IN CLEAVAGE.

55

Figs. 6-ii. — Cleavage of a number of annelids, showing progressive increase in the size of the
trunk-forming cells, viz., the first and second somatoblasts, <^2 := A'and d*=^M. The first
somatoblast, which forms the ectoderm of the trunk, is indicated throughout by horizontal
lines, and the second somatoblast, which forms the mesoderm of the trunk, by vertical
lines. The endoderm-cells are ruled with lines crossing at right angles. The stippled
cells are the primary trochoblasts, destined to form the prototroch. i. First generation
of ectomeres ; 2. Second generation of ectomeres; 3. Third generation of ectomeres.

Fig. 6. — Thirtv-two-celled stage of Podarke from the left side. Taken, with the kind permis-
sion of Mr. Treadwell. from one of his unpublished drawings.

Fig. 7. — A viphitrite, after Mead. From the left side and below.

Fig. %. — Nereis limbnta, after E. B. Wilson, from the left side.

Fig. 9. — Clymeneila torqitata. after Mead, from the left side.

Fig. id. — Aricia, optical longitudinal section, after E. B. Wilson.

Fig. II. — Clepsine, after C. O. Whitman. From the animal pole. The great size of the eudo-
derm-cells is due to yolk. The somatoblasts far exceed in bulk all the other cells.

56 BIOLOGICAL LECTURES.

adaptation in the larva than vice versa, but the adaptation in
all stages of the development is due to some common cause.
It is, of course, impossible to be blind to the fact that this
'* explains " determinate cleavage only in a very limited sense.
But if the principle serves to restrict the too universal applica-
tion in the cleavage of the Qg^ of the so-called "■ mechanical
laws of cell-division," much will have been accomplished.

V. Protoplasmic Basis of Adaptation in Cleavage.

It is possible to explain variations in size, position, and rate
of cleavage of the cells, by the hypothesis of differentiation of
the cytoplasm into materials of different qualities and positions
(His, Whitman et al.)} A much more difficult question is imme-
diately suggested, namely. How do these different substances
arise within the cell, and how are they distributed to definite
regions } Are all of the differences that exist in the thirty-two-
celled stage, for instance, also to be found in different sub-
stances of the unsegmented ovum, or are they formed, wholly
or in part, during the cleavage, and, if so, how }

The following observations, selected from my work on the
fertilization and first two cleavages of the Qgg of Unio, seem to
throw some light on this difficult problem. The spermatozoon
enters always at the centre of the vegetative pole, and describes
a penetration-path that carries it to a point between the centre
and the periphery of the Qgg near the lower boundary of the
upper (animal) half. The penetration-path is marked by a
portion of the cytoplasm cleared of yolk-granules by the activity
of the sperm-amphiaster (Fig. 12). During its penetration the
sperm-head has formed some caryolymph, and hence is some-
what vesicular ; but, arrived at its region of rest, it contracts
into a clump of small chromatic elements. All this has taken
place during the formation of the first maturation-spindle, and
is completed before the cytoplasm of the animal pole begins to
protrude to form the first polar globule. The sperm-amphiaster
and visible penetration-path have disappeared by the time of

1 A theory of polarization of the ultimate morphological units would afford
only a partial explanation of the differences in rate and direction of cleavage and
in the size of the cleavage products.

ADAPTATION IN CLEAVAGE.

57

the metaphase of the first maturation-spindle, and the sperm-
nucleus remains absolutely quiescent during the remainder of

"--°<^<5

Fig. 12. — Early fertilization of the egg of l/m'o complanatava. actual section. The space be-
tween the membrane and the egg proper has been reduced to economize space ; the micro-
pyle, penetration-path of the spermatozoon, sperm-amphiaster, and first maturation-spindle
are shown . A large proportion of the yolk -granules is temporarily forced to the periphery.

o 0
0° o
0 o

OoOo o

o O
O o

'Po 2=0^0 o

Oo =00^0

O o O CO 0 Oo9fl

o O o o o 0/

0 o 00 o o '

OoOo°ooq^

Fig. 13. — Formation of the second polar globule in [/nt'o. Sperm-nucleus to the left.

the maturation processes, until the egg-nucleus begins to grow.
Let us now study the egg during the final constriction off of

58

BIOLOGICAL LECTURES.

Fig. 14. — Unto. Small portion of the
upper pole. Egg-nucleus reconstitut-
ing ; sphere-substance forming.

the second polar globule. At this time (Fig. 13) the sixteen
chromosomes that are to remain in the egg are closely packed
together near the animal pole, and are almost in contact with
the inner sphere of the aster. The inner sphere is derived from
the centrosome of the first maturation-spindle (see No. 8) and

has the following structure : it is a
spherical body with a perfectly defi-
nite wall, containing a non-staining
substance in which is excentrically
placed a single centrosome united
to the wall by a few strands. The
cytoplasm in the immediate vicinity
is arranged in the form of an aster,
the fibres of which are centrally
attached to the inner sphere ; the
first row of microsomes on the radi-
ations of the aster bound a fairly
definite outer sphere. The sperm-nucleus occupies the same
position as before, and has undergone no change of structure ;
in Fig. 13 it is represented lying to the left. During the final
stages of formation of the second polar globule, the inner
sphere begins to enlarge, and the interior is occupied by a
reticulum, in the nodes of which are a number of centrosome-
like bodies. After the second polar globule is fully formed,
the sphere enlarges with great rapidity (Fig. 14), and, as its
boundary spreads out, it becomes less and less definite, until
its substance merges with the general cytoplasm. During its
enlargement, the interior is occupied by a vesicular substance,
at the nodes of which are deeply staining granules, in no wise
distinguishable from microsomes. The substance of the inner
sphere is now, in fact, part of the general cytoplasm. Yolk-
gra:nules are entirely absent in it (Fig. 15), and this enables one
to follow its subsequent fate for a considerable period of time.
We shall call this substance, provisionally, sphei^e-stibstance^
following Conklin (4); it is important to remember that it is
derived entirely from the inner sphere of the second maturation-
spindle. Either this is the case, or else the apparent enlarge-
ment of the inner sphere at the close of the maturation is due

ADAPTATION IN CLEAVAGE. 59

to a wave of condensation in the cytoplasm, travelling outwards,
sweeping away the yolk-granules, but leaving the cytoplasm
behind. The latter theory would explain the appearance of
vesicular cytoplasm within the boundary during the phenome-
non, and so avoid the difficulty of explaining the extremely
rapid growth of substance within the sphere ; it could be com-
pared to the wave that spreads out from a pebble dropped in
the water. But it is difficult to understand the nature of such
a disturbance. On the whole the expansion theory seems to
me much more probable than the wave theory, and the latter is
certainly untenable in the first slow stages of the process.

Another important thing to notice is that just before the
expansion begins the sphere is three-quarters surrounded by
the chromosomes, and I would like to hazard the conjecture
that, at this time, there may be a diffusion of some chromatin
substance within the sphere, the interior of which tends to
stain more darkly than before. If this be actually the case, it
has an extremely interesting bearing on subsequent events.

Let us now trace the further actions of the germ-nuclei and
sphere-substance during the first cleavage. In general, what I
propose to show is that the sphere-substance moves and elon-
gates so as to mark out a definite horizontal plane in the ^gg^
and that the first cleavage-spindle places itself in conformity
with this predetermined arrangement.

Each of the germ-nuclei, at the close of the maturation, is a
small, dense clump of chromosomes, and both begin to swell up
into the vesicular form of the ordinary resting nucleus at the
same time, and keep step throughout the process. This indi-
cates a new general condition of the egg-cytoplasm; for, while
we might explain the enlargement of the egg-nucleus alone as
part of the usual sequence of mitosis, due to purely localized
conditions of the egg-substance, we can only explain the effect
simultaneously produced on both nuclei by the assumption
that the entire egg-cytoplasm is entering on a new phase of its
development.

The first movements of the germ-nuclei begin after each has
enlarged considerably. In its early movements the egg-nucleus
is preceded by the sphere-substance, which moves towards the

6o

BIOLOGICAL LECTURES.

centre of the egg, and always towards the side opposite to that
in which the sperm -nucleus lies (Fig. 15). Thus the egg-nucleus
first moves away from the sperm-nucleus, which immediately
takes up its march in an amoeboid manner towards the centre
of the Q.^^, and in the general direction of the egg-nucleus. I
may finish at once with the movements of the germ-nuclei by

saying that they ultimately
come together in the cen-
tre of the Q^^. Whether
or not the first movement
of the egg-nucleus away
from the sperm-nucleus is
passively brought about by
the sphere-substance, the
behavior of the germ-nuclei
shows that their meeting
is not entirely due to mu-
tual attraction, but partly,
at least, to a common tend-
ency to seek a dynamic cen-
tre of the ^^g.

To. return to the sphere-
substance : its invariable migration towards the side of the
^gg opposite to that in which the sperm-nucleus lies indicates
one of two things, — either that the sperm-nucleus has driven
it away, or else that it is moving along lines of orientation of
the egg-substance. To suppose that the minute sperm-nu-
cleus could exercise such an effect seems impossible, and the
corollary of the second alternative is that the sperm-nucleus
has occupied throughout its entire resting period a definite
position within the ^g^, which can only be explained on the
assumption of a definite orientation of the ^g^ at the time of
fertilization, not only polar, but also corresponding to one of
the other chief axes of the embryo. The first cleavage-plane
passes very nearly through the point at which the sperm-
nucleus has been resting during the maturation.

During its migration towards one side of the egg, the sphere-
substance has undergone a remarkable change of form; it has

Fig. 15. — Unio. Early movements of the germ-
nuclei and of the sphere-substance .

ADAPTATION IN CLEAVAGE.

6i

Fig. i6. — Unio. Outline of section in the direction
of the line ruled across Fig. 15 ; to show the elon-
gation of the sphere -substance.

elongated greatly ijt a horizontal direction, at right angles to
the line uniting the germ-niiclei (Fig. i6), and has thus ^narked
ont the directio7i of tJie first
cleavage-spindle . Now this
elongation has begttn, and
the plane of division is indi-
cated before the germ-nuclei
have met.

It might very readily be
assumed that the plane of
the first cleavage is deter-
mined by the copulation-
path of the germ-nuclei, as
is stated to be the case in
the ova of some other ani-
mals (frog, Roux ; Toxo-
pneustes, E. B. Wilson).
But whoever should take
this position for Unio would have to explain how it happens
that the sphere-substance elongates in the plane of the first

cleavage-spindle before the
germ-nuclei come together.
It would be necessary, I
believe, to assume that the
distant sperm-nucleus ex-
ercises an influence on the
direction of migration and
of elongation of the sphere-
substance in the first cleav-
age, but that, in the next
division, the sphere-sub-
stance acts independently;
and this assumption is ab-
surd on the face of it.

The first cleavage-spindle
forms in the centre of the
Qgg (Fig. 17), in the plane already indicated by the elongation
of the sphere-substance. At first there is a single very minute

Fig. 17. — Unio. First cleavage-spindle forming. Basi-
chromatin granules near the ends of the spindle ; oxy-
chromatin granules on the spindle-fibres.

62

BIOLOGICAL LECTURES.

centrosome at each pole, almost in contact with the nuclear
membrane, then a group of centrosome granules imbedded in
a ground substance ; then comes a clearing of the centre, accom-
panied by peripheral arrangement and subsequent fusion of all
the centrosome granules, excepting one, which remains in the
centre. Thus is established a hollow sphere (" centrosome ")
with an included centrosome (''centriole") (Fig. i8). During
the early stages of the metamorphosis of the centrosomes,
chromatin granules from the nucleus are found in their im-
mediate neighborhood, and, as the spindle forms, they become

closely pressed to the
centrosomes (Fig. 17).
The entire spindle then
moves directly along the
prolongation of its axis,
thus parallel to the di-
rection of elongation of
the sphere-substance,
to one side of the ^^'g
(Fig. 18), until the cen-
trosome of one end
comes almost in con-
tact with the peripheral
layer of protoplasm.
Then comes the meta-
phase (Fig. 18), and, in
a late stage of the anaphase, the sphere begins to enlarge,
and a reticulum (or vesicular substance) with nodal micro-
somes (Fig. 19) develops in it in place of the single centro-
some.

The egg now begins to elongate at right angles to the plane
of division, and the entire spindle, including the enlarged
spheres, shifts towards the centre of the ^gg a short distance,
and finally comes to rest (Fig. 19). The spindle acts as though
oscillating through a point of equilibrium, at first shifting too
far in one direction (Fig. 18), then swinging back (Fig. 19), and
possibly undergoing lesser shiftings before coming to rest for
the first cleavage.

Fig. 18. — Migration of the spindle to one side of the egg ;
metaphase ; for the sake of clearness only a few of the
chromosomes were drawn in.

ADAPTATION IN CLEAVAGE.

63

To what extent this centre of equilibrium is determined by
the primary orientation of the cytoplasm, and how much by the
secondary distribution of the sphere-substance, it is impossible
to say ; but the orientation of the cytoplasm is, in either event,
the primary factor, inasmuch as it determines the distribution
of the sphere-substance.

The first cleavage-furrow now forms rapidly, and, as it forms,
the spheres undergo an enlargement, migration, and change
of shape analogous to that preceding the first cleavage. We
shall follow it only in the larger cell. The sphere-substance

Fig. 19.— Unio. Secondary shifting of the first cleavage-spindle. Beginning of
growth of the sphere-substance.

in this cell elongates during the early stages of reconstitution of
the nucleus in a horizontal direction parallel to the axis later
taken by the second cleavage-spindle in this cell (Fig. 20). At
the same time the entire cell elongates in the same direction
and becomes slightly constricted in a plane and position corre-
sponding precisely with the next cleavage-plane ; this, before
the nucleus is reconstituted, or has moved away from the neigh-
borhood of the first cleavage-wall. It is as though the cytoplasm
were making an attempt at division, which is rendered abortive
by the stage of development of the nucleus. Later both cells
round off and then become applied together, and the second
cleavage-spindle forms and moves into the position of the division
already indicated.

64

BIOLOGICAL LECTURES.

Conklin (4) was the first to call attention to the possible
importance of the sphere-substance in the cleavage of the ^g'g.
He summarizes his results on the ^g'g of Crepidtda thus : "After
the first two cleavages the sphere-substance is differently dis-
tributed to the different cells, the entire sphere-substance of
one generation always going into those cells of the next genera-
tion which lie nearest the animal pole. This differential distri-
bution of the spheres has been followed through every cleavage
up to the twenty-four-cell stage. As the form of the cleavage

is perfectly constant, it

follows that the sphere-
substance of any genera-
tion goes into certain
definite cells which have a
perfectly constant origin
and destiny. This differ-
ential distribution of the
spheres is not caused by
their specific weight, since
their movements are the
same in whatever position
the cell may be placed. It
seems to be the result of
a form of polarity which, like that of the ^gg itself, is not
the result of gravity.

'* The centrosomes do not, apparently, arise from the sphere-
substance of the previous division, but some distance from it,
and the sphere-substance never divides, but each sphere ulti-
mately grows ragged at its periphery and gradually fades out
into the general cytoplasm.

"The differential distribution of the spheres and their
subsequent conversion into cytoplasm suggest that they
may be important factors in the differentiation of cleav-
age cells, and if further investigation should establish the
fact that they are, in part, composed of the oxychro-
matin of the nucleus, it would furnish a basis in fact for
certain well-known speculations of de Vries, Weismann,
and Roux."

Fig. 20. — Unio. tirst cleavage ; the elongation of the
sphere-substance in the larger cell and of the cell
itself marks the plane of the second cleavage. In a
slightly later stage there is a well-marked constriction
across the cell in the position of the future second
cleavage-spindle in this cell.

ADAPTATION IN CLEAVAGE. 65

Thus, putting Conklin's results ^ on the movements of the
sphere-substance and my own together, it would appear that,
whereas in the first two cleavages this substance is divided
between the cells in proportion to their size, in the formation
of the generations of ectomeres the substance enters special
cells. This would coincide very closely with the differential
value of the cleavages, the first four cells possessing ectoblastic,
entoblastic, and, in part, mesoblastic materials, while the three
subsequent divisions of the macromeres separate ectoblastic
portions. This tends, it seems to me, to strengthen Conklin's
conclusion that the sphere-substance may be an important
factor in the differentiation of cells.

Finally, I do not believe that the process of nuclear or cell-
division is ever in itself an act of differentiation. That it is
not, in certain cases at any rate, is shown beyond the possibility
of any doubt by examples of non-determinate cleavage, such as
that of the fish-egg, in which the cleavage-planes bear no con-
stant relation to each other or to the embryonic parts, and,
still more strikingly, in the case of ciliate Iiificsoriay where the
entire process of development takes place without any cell-
division. If my observations are correctly interpreted in
what has preceded, the essential process in early embryo-
formation proceeds on the basis of a definite orientation and
organization of the egg-substance, carried forward and elab-
orated by certain intercellular processes, in which the produc-
tion of special substances which have been acted on by the
chromatin may play an important role. Now the distribution
of these substances is not dependent on cell-division, though by
this they may be isolated in separate cells ; but it is conceiv-
able that the cleavage-planes may, so to speak, ignore the lines
of orientation of the ^g'g and of distribution of specific parts of
it ; thus it may be that determinism in the cleavage is no
measure of the degree of organization of the ^gg, as Whitman
has so ably argued.

It is quite possible that there is no sharp distinction between

1 In Crepidtcla, apparently, the substance of the spheres is not divided in the
second cleavage, but passes into special cells. See lecture by Dr. Conklin in this
volume.

66 BIOLOGICAL LECTURES.

determinate, and indeterminate cleavage, and that one grades
into the other, the apparent difference being due to insufficient
knowledge, as Conklin and Eisig (5) have suggested. But there
can be no doubt in the mind of any one as to the existence of
a very real difference between determinate and indeterminate
types of cleavage, who has compared, for instance, the cleavage
of the Q.gg of an annelid, possessing a perfectly definite and
unvarying mode of cleavage and cell-lineage of organs, with
that of a fish, in which slight alterations of the external condi-
tions cause the very greatest variations in cleavage, so that
often the cells of two eggs of the same species cannot be
homologized, and no definite cell-lineage of organs exists. The
explanation of this difference, it seems to me, is a prospective
one. It is dependent, I believe, on the actual number of cells
composing the embryo at the time that the first larval or
embryonic organs come into service. In other words, I would
think of determinate cleavage as an adaptation to a condition
in which the functional activity of organs begins with a rela-
tively small number of cells, and in which, therefore, each cell
is of special importance.

Marine Biological Laboratory,
August, 1898.

ADAPTATION IN CLEAVAGE. 67

LITERATURE.

1. Child, C. M. A Preliminary Account of the Cleavage of Arenicola

cristata, with Remarks on the Mosaic Theory. Zool. Bulletin. Vol. i.
1897.

2. CoNKLiN, E. G. Cleavage and Differentiation. Biological Lectures.

DeHvered at the Marine Biological Laboratory. 1896, 1897.

3. CoxKLix, E. G. The Embryology of Crepidula. Journ. of Morph.

Vol. xiii. 1897.

4. CoxKLix, E. G. See summary of a paper delivered before the Ameri-

can Society of Morphologists at the Ithaca meeting, December, 1897.
Science. March, 1898.

5. EisiG, Hugo. Zur Entwicklungsgeschichte der CapiteUiden. Mitth.

aus der Zool. Station zu Neapel. Bd. xiii. 1898.

6. LiLLiE, Frank R. Preliminary Account of the Embryology of Unio

complanata. Journ. of Morph. Vol. viii. 1893.

7. LiLLiE, Fraxk R. Embryology of the Unionidae. Journ. of Morph.

Vol. X. 1895.

8. LiLLiE, Frank R. Centrosome and Sphere in the Egg of Unio. Zool.

Bulletin. Vol. i. 1898.

9. Mead, A. D. The Early Development of Marine Annelids. Journ.

of Morph. Vol. xiii. 1897.

10. Treadwell, a. D. The Cell-Lineage of Podarke obscura. Prelimi-

nary Communication. Zool. Bulletin. Vol. i. No. 4. 1897.

11. Whitmax, C. O. The Embryology of Clepsine. Quar. Journ. of

Micr. Science. Vol. xviii. 1878.

12. Whitmax, C. O. The Inadequacy of the Cell-Theory of Development.

Biological Lectures. Delivered at the Marine Biological Laboratory.
1893.

13. WiLSOX, E. B. The Cell- Lineage of Nereis. Journ. of Morph.

Vol. vi. 1892.

14. WiLSOX, E. B. Considerations on Cell-Lineage and Ancestral Remi-

niscence. Annals of the New York Academy of Sciences. Vol. xi,
No. I. i8q8.

FOURTH LECTURE.

PROTOPLASMIC MOVEMENT AS A FACTOR OF
DIFFERENTIATION.

EDWIN G. CONKLIN.

The fundamental problems of development and inheritance
are in the last analysis questions of differentiation. Develop-
ment is progressive differentiation coordinated as to time and
place ; hereditary likeness consists in the repetition by the off-
spring, at certain stages of its life cycle, of definite differentia-
tions of the parent ; and hereditary unlikeness, or variation, is
a modification of these differentiations either as to their char-
acter or as to the time of their appearance. The phenomena of
differentiation are therefore of the greatest interest, and their
causes one of the most important problems of biology.

In many respects the simplest and yet most important phe-
nomena of differentiation occur in the early stages of develop-
ment, while the later differentiations of tissues and organs are
more complicated and less general in character. The polarity
of the egg is one of the earliest differentiations of the de-
veloping organism ; it consists not only in the aggregation of
yolk at one pole and of protoplasm at the other, but also in
the establishment of certain structural peculiarities which in
most cases determine the position and direction of the two
maturation spindles and cause the first two cleavage planes to
pass through the polar axis of the egg. Further, it has a defi-
nite prospective significance, since it is probable that in all ani-
mals it determines the ectodermal and endodermal poles of the
embryo, while in most cases the animal and vegetal poles of
the egg give rise, respectively, to the apical and oral poles of the

69

70 BIOLOGICAL LECTURES.

larva or adult. This polar differentiation may appear at an
early stage in the ovarian egg, or it may be delayed until after
fertilization. In some cases (insects and cephalopods) not only
the primary axis of the ^gg^ but all the axes and regions of the
future animal are marked out in the ovarian ^gg ; in other cases
these axes, except the primary one, are not apparent until the
end of cleavage or even after gastrulation.

In the cleavage of the ^gg^ differentiations occur in a
remarkable degree in certain cases, while they appear to be
absent in others. Typically, cell divisions are rhythmical, alter-
nating, quantitatively and qualitatively equal, and consequently
non-differential. The differentiations of cleavage cells are due
to departures from this typical condition in one or more partic-
ulars. In certain animals these departures are very notable, the
cleavages being from the first non-alternating, non-rythmical,
unequal, and qualitatively dissimilar. These differentiations of
cleavage have also a far-reaching prospective significance, since
in certain cases (polyclades, nematodes, rotifers, annelids, mol-
lusks) the principal axes and body regions of the future animal
are marked out by the cleavage planes, and the building mate-
rial of entire organs is segregated into a single cell or group
of cells.

I have repeatedly observed these unequal, non-alternating,
and non-rhythmical cleavages with the feeling that the causes
of such differentiation were almost within sight, and with the
conviction that continued study could not fail to reveal them ;
and yet it must be said that these causes, which seem so near
at hand, generally elude one's grasp. Unequal cleavage is due
to the eccentricity of the mitotic spindle, but why is the spindle
eccentric.-* Non-alternating cleavage is due to the spindles hav-
ing approximately the same direction during successive cleav-
ages; but why do the spindles take this peculiar position.!*
Non-rhythmical cleavage can be referred only to differences in
the substance of cells, but how these differences operate can-
not in most cases be explained. It is hopeless to look for an
answer to the last question that may be asked concerning the
cause of these or of any other phenomena; all that can reason-
ably be expected is that the many different phenomena and

PROTOPLASMIC MOVEMENT. 71

factors of development may some time be harmonized and uni-
fied by the discovery of some common causal principles.

I. Mechanical Factors of Differentiation.

Of the many factors of differentiation which have been pro-
posed within recent years, the majority are in the nature of
simple physical causes. Thus the polar differentiation of the
^gg has been attributed to differences in the specific weight of
protoplasm and yolk ; e.g., Hertwig ^ says : " Polar differentiation
consists in this, that the lighter protoplasm collects at one pole
and the heavier yolk substance at the other." Experiments on
the frog's ^gg led Pfliiger, Born, and Schultze to essentially the
same conclusion.

The differentiations of cleavage are commonly attributed to
factors of a similar character ; thus the direction of cell division
is said to be due to the fact that the mitotic spindle lies in the
direction of least resistance (Pfliiger) or in the longest axis of
the protoplasmic mass (Hertwig) ; the shape and position of
cells, and consequently to a certain extent the direction of divi-
sion, are said to be due to the rectangular intersection of cleav-
age planes (Sachs), or to the principle of smallest surfaces
(Berthold) ; the rate of division and the relative size of daughter-
cells are commonly attributed to the mechanical influences of
inert yolk (Balfour, Hertwig). Finally, the differences in the
quality of cells have been referred to intercellular reactions
(Hertwig, Wilson),' which, in some cases at least, are regarded
as of a physical rather than of a physiological character.

It has been repeatedly shown that none of these principles
are of universal application, and it seems doubtful whether in
any case they are the real causes of the phenomena in ques-
tion. How little gravity has to do with polar differentiation
is well shown in the eggs of many gasteropods where the eggs
lie in all possible positions in the ^gg capsules with their pri-
mary axes turned in all possible directions, and yet the polar
differentiation occurs as perfectly and as rapidly in one position
as in another. That gravity can have nothing directly to do

1 The Cell, p. 215.

72

BIOLOGICAL LECTURES.

with this differentiation is further shown by the fact that the
protoplasm which Ues at the animal pole in early stages of
cleavage lies at the vegetal pole of the macromeres in later
stages ; while the yolk, which originally lay at the vegetal pole,
comes to lie at the animal pole of the macromeres ; in short,
the polarity of the macromeres is reversed during cleavage, and
this always happens in the same way, irrespective of the posi-
tions in which the eggs may lie.

How far these mechanical principles fall short of explaining
the differentiations of cleavage has been pointed out by Mead
('94), Kofoid ('94), Lillie ('95), McMurrich ('95), Wheeler ('95),
Zur Strassen ('95), Castle ('96), Jennings ('96), Conklin ('92, '94,
'97), and many others. Much has been said, and very justly, of
the difficulty of explaining by simple mechanical causes non-
alternating and unequal cleavages, and yet it ought not to be
forgotten that the causes of alternating and equal cleavages
have not been given. To say merely that cell divisions are typi-
cally alternating and equal affords no causal explanation. If we
knew why cleavages are usually alternating and equal, we should
probably be able to explain why they are sometimes neither.
How little alternating cleavage has to do with the mere diver-
gence of centrosomes in planes successively at right angles to
one another will be apparent further on, where we shall see
that the centrosomes do not preserve their original positions
in the daughter-cells, and that the direction of their divergence
bears no constant relation to the direction of the cell division ;
and how little yolk has to do with the ineqiiality of division is
shown by the fact that in the formation of the polar bodies we
have two very unequal divisions of the ^gg^ while the first and
second cleavages, are frequently equal ; the first, second, and
third divisions of the macromeres thus formed are usually very
unequal, while the fourth and fifth divisions of these cells are
again nearly equal ; finally, among the micromeres, which fre-
quently contain no yolk, the early divisions are more frequently
unequal than equal. These same considerations apply in the
main to the rate of division ; in most cases of determinate
cleavage it bears no constant relation to the presence or absence
of yolk in the cells. In every case of determinate cleavage

PROTOPLASMIC MOVEMENT. 73

these so-called ''laws" are repeatedly set at naught, and this
fact has led those who have studied such cleavage to the con-
clusion that no simple physical explanation of these processes
of differentiation is possible, and that their cause must be found
in the structure of the protoplasm or in some physiological
factor.

II. Physiological Factors.

The causes of differentiation are frequently referred to the
structure of the germinal protoplasm, as if this were a sat-
isfactory explanation. But to say that polarity and differen-
tiations of cleavage are due to the constitution of the ^gg is
merely a form of words which means little or nothing. In the
same way it might be said that all the multifarious aspects of
the universe are the results of the constitution of matter. To
refer vital phenomena to the constitution of protoplasm and
there to rest is merely to juggle with words. The phenomena
in question must be analyzed and their immediate causes deter-
mined step by step before any '* explanation " can be thought of.
If, then, those who attempt to explain differentiation as the re-
sult of simple physical factors have taken too narrow a view
of the problem, and have unduly simplified it, those who suggest
the structure of protoplasm as an explanation propose a factor
so vague and remote that it has no real value. The immediate
causes of differential cleavage have been located, by different
authors, in various parts of cells.

a. Nucleus. — According to the views of Weismann, Roux,
De Vries, and many others, the nucleus is the prime mover in
all processes of differentiation ; and hence we find that they
attribute the direction, size, and rate of division, and the quality
of cells, to the action of the nucleus. According to Roux ('95)
there is immanent in the nucleus a direction of division which
may be independent of the chief dimensions of the protoplasmic
body. Almost all observers agree that the direction and size
of division depend upon the direction and position of the mitotic
spindle, and many consider that the position of the spindle is
not determined by the cell body ; thus Jennings says of As-
planchna ('96, p. 71), "The form of the cell is greatly influenced

74 BIOLOGICAL LECTURES.

by the direction of the contained spindle," but ''the form of
the cell has no effect upon the position of the spindle " (p. 72).
Experiment has shown that the position of the spindle and
the direction and size of division may be modified by pressure,
and this again has lent support to the view that the position
of the spindle is the determining feature in cell division.

b. Centrosome. — Others consider the centrosome the prime
mover in cell division, " the dynamic centre of the cell." Rauber
(■83) holds that the position of the spindles is the result of the
mutual attractions of neighboring asters. Heidenhain ('95) re-
fers the direction of division to a definite angle of rotation of
the centrosomes.

c. Cytoplasm. — Boveri ('97) has shown that neither nucleus
nor centrosome can be said to take the initiative in cell division,
since the mitotic processes may go on independently in each,
and both he and Driesch ('98) have shown that the rhythm of
cleavage depends upon the ^g'g cell and not upon the sperm, and
therefore, in all probabilities, upon the cytoplasm and not upon
the nucleus or centrosome.

III. Protoplasmic Movement as a Factor.

Almost all views as to the factors of differentiation regard
the cell as in a static condition ; very few consider it from the
kinetic standpoint, though the very names karyokinesis and
cytokinesis suggest movement as the fundamental fact of cell
division.

In the maturation, fertilization, and cleavage of certain, gas-
teropods I have observed successive stages of a complex and
orderly movement of the entire cell body by which the positions
of the cytoplasm, yolk, nuclei, centrosomes, and mid-bodies
(Zwischenkorpern) are changed in a definite and orderly way.
Unfortunately, I have been unable to actually observe these
movements in the living ^gg, since the eggs studied contain a
large amount of yolk and are therefore opaque, and since the
movements described are very slow. However, the evidences
of these movements are so abundant and unmistakable that one
could not be more certain of them if he had seen the actual

PROTOPLASMIC MOVEMENT. 75

flowing of the cell substance. These observations are based in
the first instance upon the eggs of Crepidula plana^ though
many other species and genera of gasteropods have been studied
with the same results. The phenomena are, therefore, not iso-
lated, and some of them are probably of general occurrence.

a. Movements m the Unsegmented Egg.

The cell movements during maturation result in the segrega-
tion of yolk and cytoplasm at opposite poles of the ^gg^ and in
the transportation of the mitotic figure to the animal pole.
While the germinal vesicle is still intact it is closely surrounded
by yolk spherules, and there is a very incomplete separation of
yolk and cytoplasm throughout the entire ^g'g. As soon as the
mitotic spindle is formed and the nuclear membrane is broken,
there is an area immediately surrounding the spindle and asters
free from yolk, but elsewhere in the ^gg there is an intimate
mingling of yolk and cytoplasm. Gradually the spindle, which
at first has a position nearly at right angles to the ^gg axis and
some distance from the surface of the ^gg, turns until its axis
nearly coincides with that of the ^ggy and at the same time the
whole spindle is moved out toward the surface, until finally the
outer end of the spindle comes in contact with the cell mem-
brane, and the surface of the ^gg is elevated into a papilla at
this point. This movement is in part due to the mere length-
ening of the nuclear spindle which doubles in length during the
process, but in part also to a general movement of the cell body
by which the spindle is turned and carried bodily toward the
surface of the ^gg. At the same time there are movements
within the ^gg which lead to an accumulation of cytoplasm at
the animal pole and a movement of the yolk spherules toward
the opposite pole. There is no evidence that this movement is
due to activity on the part of the nucleus or centrosomes. The
initial position of the centrosomes and the direction of the cen-
tral spindle are not the same in different eggs, and yet the final
position of the mitotic figure is the same in all cases; the cen-
trosomes and asters at the two poles are identical in form, size,
and staining reactions until the outer pole of the spindle comes

76 BIOLOGICAL LECTURES.

into contact with the surface of the ^gg. Under these circum-
stances it is difficult to conceive what causes the spindle to
rotate into a radial position and to move bodily toward the
surface, unless it be a general movement of the cytoplasm, and
the fact that throughout the ^^'g the separation of yolk and
cytoplasm is going on at the same time is additional evidence
in favor of such a general movement of the cell body.

During the fertilization similar movements of the ^g'g con-
tents are apparently taking place ; the polar segregation of yolk
and cytoplasm goes on during the approach of the germ nuclei,
and, as during maturation, appears to be due to movements of
the cytoplasm. The spermatozoon usually enters near the vege-
tal pole, and is carried through almost the whole diameter of the
^gg to the animal pole, but it may enter at any place except
the protoplasmic area immediately around the animal pole. If
the sperm enters at the vegetal pole, its course toward the
animal pole is nearly straight ; if it enters elsewhere, its course
is curved, and the nearer the point of entry to the animal pole
the greater the curvature. An aster is formed in advance of
the sperm nucleus and precedes it in its course through the
^gg. The ^gg nucleus and aster lie very near the animal pole
and do not move from this position ; they are surrounded by an
area of protoplasm free from yolk. The sperm nucleus and
aster in their advance through the yolk leave no path behind
them ; either they are carried along by a general movement of
the cell contents, or the yolk is pushed out of their way, to close
in again behind them immediately after they have passed. The
germ nuclei and asters approach each other as I described in a
former lecture at this laboratory ('94), and when the two are close
together they lie in an area entirely free from yolk, except that
a few spherules are usually found between the two nuclei or
asters. These spherules, which are separated from all the rest of
the yolk, appear to have been carried before the sperm elements
in their advance. After the origin of the cleavage centrosomes
the remnants of the asters are carried to a point above the nuclei
and immediately under the polar bodies, where they disintegrate
and are scattered as coarse granules, — a process which will be
described more fully when we come to consider the cleavage.

PROTOPLASMIC MOVEMENT. J J

What brings the germ nuclei and asters together ? In a
former lecture I suggested that the nuclei were passively drawn
together by the formation, attachment, and contraction of
astral rays. Wilson ('96) regards this view as untenable, and
concludes that ''the nuclei are drawn together by an actual
attraction which is neutralized by union, and their movements
are not improbably of a chemotactic character." Unless the
nuclei have organs of locomotion it must still be true that they
are brought together by something outside themselves. This
something must of necessity be found in the cytoplasm (includ-
ing the aster), unless the nuclei are able of themselves to move
actively. I presume that amoeboid movement of the nucleus
would be the only kind possible, and yet I have never seen a
single case in which either of the germ nuclei was amoeboid.
There is every evidence that the nuclei in this, as in most other
cases of movement, are passive, and that their movements are
brought about by the activity of the cytoplasm.

The migration of the sperm nucleus, like that of the matura-
tion spindles, is accompanied by progressive separation of yolk
and cytoplasm, and it is probable that these coincident phe-
nomena have a common cause in general movements of the
cytoplasm.

Furthermore, there are certain elements of constancy in the
polar differentiation and in the plane of the first cleavage which
cannot be attributed to the nuclei, and, so far as I can see, can
be due only to definite characters of the cell body. It is the
tgg cell and not the nuclei which shows polar differentiation.
The sperm nucleus and aster approach the animal pole from
various positions; there is great variation in all the positions of
the nuclei and asters relative to each other, and yet there is no
variation in the plane of the first cleavage which always passes
through the point of extrusion of the polar bodies, and in cases
where the first cleavage is unequal the mitotic figure is always
eccentric to the same degree. Now the first cleavage, as we
shall see, is accompanied by extensive rotary movements of the
cell contents, and this fact, joined to the evidences of cyto-
plasmic movement during maturation and fertilization, leads me
to believe that definite movements of cell substance exist in the

78

BIOLOGICAL LECTURES.

unsegmented ^^%. The constancy of cleavage in later stages
is associated with constancy of movements in the cytoplasm,
and it is probable that the same is true of the constancy which
precedes cleavage.

That the movements within the cell substance of the unseg-
mented Q^^g are, in certain cases at least, of a vortical character

•^o;,o° o;-a;^

Fig. 2.

Figs, i and 2. — Sections of eggs of Crepidula plana, showing prophase and metaphase

of first cleavage.

is indicated by spiral asters, first described by Mark for Limax,
and since observed by several investigators in other animals,
and also by my observation that the first cleavage in Crepi-
dula is a spiral one, being oblique to the right, or dexiotropic.

PROTOPLASMIC MOVEMENT.

79

b. Movements during Cleavage.

It is, however, in the cleavage cells that I have found the
most unmistakable evidences of the definite and orderly move-
ments of cell contents. In the first cleavage the spindle lies at

Figs. 3 and 4.

.0 0 O
Q' 000

Fig. 4.

Sections of eggs of C. plana, showing anaphase and telephase
of first cleavage.

right angles to the chief axis of the ^gg, and the centrosomes
are no nearer the surface of the ^gg than is the middle of the
spindle (Figs. 1-3). As the daughter-nuclei pass into the rest-
ing phase, however, the centrosomes are transformed into cen-

8o BIOLOGICAL LECTURES.

trospheres, which are carried up toward the surface of the ^gg,
and the mid-body, which marks the middle point of the spindle,
is carried down into the ^ggy until finally the centrospheres
come to lie immediately under the cell membrane and above
the nuclei, while the mid-body is located near the centre of the
^gg (Figs. 4-6). The currents which bring about this affect
the whole cell body, for not only are the centrospheres, nuclei,
and mid-bodies moved, but the yolk and protoplasmic areas
shift so that the protoplasm, which was spread out as a rather
broad area above the yolk, becomes narrower and deeper, the
yolk moving up at the periphery and down in the middle of the
^ZZ (Figs. 3-5). Such a change of position could be brought
about only by a general movement of the entire cell body in the
direction of the arrows in Fig. 5. The movements in the two
cells are not, however, directly at right angles to the plane of
the first cleavage, but viewed from the animal pole they are
slightly dexiotropic, as is shown by the fact that the nuclei,
spheres, and protoplasmic areas all move in a dexiotropic direc-
tion (Fig. 6). The remains of the centrospheres of the first
cleavage can be seen, until the anaphase of the second cleavage,
lying near the upper surface of the two blastomeres and close
to the wall between them; in this position they gradually fade
out into the cytoplasm, until at the close of the second cleavage
no trace of them can be seen. The origin of the centrosomes
of the second cleavage has not been traced in detail, but judging
by analogy with other cleavages it is almost certain that they
come from the inner portion of the centrospheres of the first
cleavage. The plane of divergence of the centrosomes is not
constant; in extreme cases the spindles may vary as much as
90° from the usual direction, and yet in all cases the spindle is
ultimately turned into a laeotropic position. In this case, there-
fore, as in the formation of the polar bodies and in many of the
later cleavages, the initial position of the mitotic spindle does
not determine the direction of the cell division. The flow of
cell substance indicated in Fig. 6 continues until the superficial
extent of the protoplasmic area is smaller and its depth greater
than is indicated in that figure, and until the new centrosomes
have taken their positions at the poles of the greatly inflated

PRO TOP LA SMIC MO VEMENT.

8l

nuclei. During the mitosis the surface extent of the proto-
plasmic area becomes greater, and the blastomeres, which had
become so flattened against each other that they were nearly
perfect hemispheres, again become more and more spherical in
shape ; at the same time currents seem to move outward from

Fig. 5-

Fig. 6.

Figs. 5 and 6. — Entire eggs of C. plana, lateral and apical views. The centrospheres are shown
in light shading. The arrows indicate the directions of the protoplasmic movements.

the centre of the ^gg, and this time in such direction as to
bring about a laeotropic shifting of the nuclei and protoplasmic
areas (Fig. 7). This movement continues until the anaphase
of the second cleavage, when again the cell currents move toward
the centre of the Q,gg (Fig. 8). The direction of these currents,
though nearly radial, is still slightly laeotropic in the two blasto-

82

BIOLOGICAL LECTURES.

meres which lie at the higher level and dexiotropic in the two
lower ones. This movement continues until the spheres are
carried into the inner angles of the cells, immediately below the

Fig. 7.

Fig. 8.
Figs. 7 and 8. — Entire eggs of C. plana ; second cleavage.

polar bodies, and until the daughter-nuclei, which at first lie
very near each other on opposite sides of the second cleavage
plane (Fig. 8), are swung out into the centre of the protoplasmic
area. The centrosomes and spindles then appear as in the

PRO TO FLA SMIC MO VEMENT.

83

Fig. g.

Fig. 10.

Figs. 9 and 10. — Entire eggs of C. plana ; third cleavage. Sphere substance of second cleavage
at central and upper angles of cells. Initial direction of the spindles not uniform.

preceding cleavage. The protoplasmic area increases with the
breaking of the nuclear membrane, and the spindles do not
occupy similar positions in the four cells but are often dexio-

84 BIOLOGICAL LECTURES.

tropic in the two lower blastomeres (the ones meeting in the
polar furrow), and frequently laeotropic in the upper ones (Fig. 9).
As the cleavage advances the spindles are all turned in a dexio-
tropic direction. All this time the remains of the spheres of
the preceding cleavage occupy the central angles of the cells,
where they are plainly undergoing disintegration (Figs. 9-1 1).
During and after this cleavage the cell currents rotate in a
dexiotropic direction in the upper cells (micromeres), and in a
laeotropic direction in the lower cells (macromeres) ; this move-
ment is manifest during cleavage in the dexiotropic lobing of
the cytoplasm in each cell preparatory to the formation of the
micromeres, and in the obliquity of the cell plate to the axis of
the spindle (Fig. 11); after cleavage it appears in the rotation
of the centrospheres and nuclei. Even after the third cleavage
is completed the sphere substance of the second cleavage can
be seen in the inner angles of the micromeres, where it grad-
ually disintegrates and passes into the cytoplasm. The dexio-
tropic rotation of the protgplasm in the micromeres continues
until the daughter-nuclei are carried from the left to the right
side of the cells; at the same time the substance of the macro-
meres rotates to the left, until the nuclei and spheres are
brought into contact with the cell membranes at the left of
these cells (Fig. 12). Here the centrosomes and central spindles
arise from the inner spheres of the centrospheres, and the laeo-
tropic movement of the cytoplasm continues until the spindles,
which at first are equatorial, become decidedly laeotropic in
position, and the second quartette of micromeres is given off in
a laeotropic direction. As the whole of the sphere substance
of the second cleavage goes into the first quartette of micro-
meres, so all the sphere substance of the third cleavage remain-
ing in the macromeres goes into the second quartette, where it
rapidly disintegrates and disappears. The contents of the second
quartette move in a laeotropic direction until the centrosphere
is carried from the extreme left to the extreme right of each
cell; at the same time the substance of the macromeres moves
in a dexiotropic direction until the centrospheres and the nuclei
are carried to the right side of each cell, and in this position
the third quartette is given off in a dexiotropic direction (Figs.

PROTOPLASMIC MOVEMENT.

8=;

Fig.

Fig. 12.

Figs, ii and 12. — Entire eggs of C. plana. Anaphase and telephase of third cleavage. Sphere
substance of second cleavage in inner angles of micromeres (Fig. 11). The two figures
show the beginning and the end of the rotations indicated by the arrows.

13 and 14). Here, as in the preceding cleavages, the whole of
the sphere substance left in the macromeres goes into the upper
cell products, i.e.^ in this case, the third quartette.

86

BIOLOGICAL LECTURES.

Finally, I shall describe a few subdivisions of the first and
second quartettes with reference to these cell movements.
During and after the formation of the first quartette, the rota-
tions in these cells are dexiotropic ; when the nuclei have been

Fig. 13.

Fig. 14.

Figs, 13 and 14. — Entire eggs of C. plana. Fig. 13, Unequal division of first quartette cells.
Fig. 14, Subdivision of second quartette cells and formation of third quartette.

PROTOPLASMIC MOVEMENT. ^']

carried from the left to the right side of each cell, the spindles
for their first subdivision appear, and the cleavage takes place
in a laeotropic direction (Fig, 13). This cleavage is a very
unequal one, the outer products being the small trochoblasts.
The inequality is preceded by a marked eccentricity of the
nuclear spindle (Fig. 13); this eccentricity is not indicated in
the position of the nuclei before the spindle appears, for they
lie near the inner angles of the cells (see the cell not dividing
in Fig. 13), but during the prophase the nuclei and spindles
move outward in the cells until the outer (lower) centrosome
comes almost into contact with the cell membrane, while the
inner (upper) one lies near the middle of the cell. In this
position the division takes place in a laeotropic direction (Fig.
13), and immediately the currents in the upper cell moiety
begin to rotate in a laeotropic direction, while those in the
lower moiety usually rotate in a dexiotropic direction. In this
case the eccentricity of the spindle is the immediate cause of
the unequal cleavage, and this eccentricity is the result of vor-
tical movements in the cell, which begin coincidently with the
breaking of the nuclear membrane.

The first division of the second quartette is a nearly equal
one, and is slightly dexiotropic in direction. These cells were
formed by laeotropic cleavage, and consequently the rotation
within them is a laeotropic one. When this laeotropic move-
ment has carried the centrospheres from the extreme left to
the extreme right of the cell, the cleavage begins, the whole of
the sphere substance going into the right product of division
(Fig. 14).

Especial interest attaches to the cell movements in reversed
cleavage, or cases in which two successive divisions are in the
same direction. Such an instance occurs in the first division
of the basal cells of the cross. (See Conklin, '97 and '98.) These
cells were formed by dexiotropic division of the apical cells, and
to preserve the law of alternation they should divide in a laeo-
tropic direction, but they all divide dexiotropically. Immedi-
ately after the apicals have given rise to these basal cells, the
contents of the former rotate in a dexiotropic direction until
the centrospheres are carried through an angle of about 90° ;

88 BIOLOGICAL LECTURES,

at the same time the centrospheres of the basal cells remain
almost exactly in their first position, though they move slightly
toward the surface and sometimes a little to the left. The re-
versed cleavage of these cells seems not to be due to reversed
currents in the cells, but rather to the absence of any currents.
Further divisions have been followed in detail up to a late
stage in the cleavage, but as they illustrate merely the princi-
ples which have been already described, no account is given of
them here.

IV. Summary and Comparisons.

The principal facts as to the rotation of the cell contents, or
what might be called the cytokinesis of cleavage, are the fol-
lowing: I. With the escape of nuclear sap into the cell body
at the beginning of mitosis, vortical movements are set up in
the cytoplasm, the poles of the spindles being the centres of
such vortices. 2. The vortices in daughter-cells are in reverse
directions, the movement in the upper cell being in the direc-
tion of the cleavage by which that cell was formed. 3. Cell
walls form where vortices meet. 4. Successive divisions alter-
nate in direction when the rotations carry the nucleus to the
side of the cell opposite to that in which it first lay ; non-
alternation is due to the absence of such rotation. 5. When
the cell movements carry the mitotic spindle out of the middle
of the cell, unequal cleavage results. 6. After the first two
cleavages, every cell division is qualitatively differential, since
the sphere substance of the preceding cleavage is carried into
one only of the two daughter-cells.

Such movements of the cell substance as are here described
have long been known in certain plant cells and among protozoa.
Ryder ('94) in particular has discussed the movements of
amoeboid organisms and finds the cause of such movement in
a vortical flow of protoplasm. A few observations have also
been made on the movements of cell substance in certain ova,
but in no case have these movements been followed in detail,
and in no case have they been connected with processes of
differentiation. Whitman ('8?) emphasized the importance of
the cytoplasm in the movements of the germinal vesicle and

PROTOPLASMIC MOVEMENT. 89

of the pronuclei during maturation and fertilization. For
these active changes in the cytoplasm he proposed the name
cytokinesis. Morgan ('93) observed that the reddish pigment
granules found over the surface of the eggs of Arbacia move
entirely away from the micromere pole of the ^gg before the
micromeres are formed. In some eggs this movement begins
in the two-cell stage, and is carried on until the micromeres
are formed at the sixteen-cell stage. Nussbaum i^z) observed
in the division of entoderm and mesoderm cells of young em-
bryos of Rana temporaria that the brown-black pigment col-
lected in a ring around the equator of the dividing cell, and as
the division advanced the ring became narrower and deeper
until it formed a true cell plate between the daughter-cells.
Van Bambeke ('96) has observed a similar phenomenon in the
cleavage of the toad's ^gg. Gardiner ('95) observed in the eggs
of Polychoeriis and Aphanosto^na a reddish-yellow pigment
which, because of its form and peculiar movements, he sup-
posed might be some form of alga. After the ^gg is laid it
migrates from the centre toward the periphery, and forms a
girdle around the ovum in the plane of the first cleavage. A
similar line of pigment marks out the division plane of every
succeeding cleavage up to the ten-cell stage. He also observed
that these pigment granules migrated from one pole of the ^gg
to the other, though they never passed from one cell to the other.
These movements greatly impressed Gardiner with the won-
derfully active and powerful forces within the ^gg. When the
living ^gg is seen under an immersion lens he says "the sur-
face fairly scintillates with the movements of the protoplasm
and these pigment granules."

About the same time Loeb ('95^) suggested that a mechan-
ical explanation of the division of the ^gg or embryo was to be
found in diffusion and vortex movements of the protoplasm,
similar to those observed by Quincke in an emulsion of oil and
soda solution. '* I conceive," says he, "that on the surface of
the ^gg, possibly in the meridian or circle whose plane sepa-
rates from one another the two radiating systems of the cen-
trosome, diffusion phenomena occur as soon as the nuclear
division has physically ended. These lead to the formation of

90 BIOLOGICAL LECTURES.

vortical movements, symmetrical in relation to this plane." If
these movements are violent, they lead to the complete sepa-
ration of the daughter-cells ; if not, ordinary cleavage results.
Later ('952) j^g observed in the segmenting eggs of Ctenola-
brus, droplets over the surface of the Qgg which collected in the
plane of the next succeeding cleavage ; this phenomenon he
considered a confirmation of his theory.

The cell movements which occur during the maturation, fer-
tilization, and cleavage of Crepidula not only confirm Loeb's
view as to the mechanics of cleavage, but they throw light
upon the mechanics of differentiation. Of the four features of
differential cleavage, viz.^ differentiations in the (i) direction,
(2) size, (3) quality, and (4) rate of division, the first three are
due, in part at least, to these movements of the cytoplasm. Evi-
dently, varying rates of division must be attributed to some
other cause.

If we go further and inquire what causes and directs these
movements, we cannot at present find any positive answer. It
seems probable that they are due to the appearance of unlike
substances in different parts of the cell, the movements being
of a chemotropic character. Further, these movements are
correlated with the growth and collapse of the nucleus and
centrosome,^ and especially with the escape of nuclear sap into
the cell. It is a priori improbable that any one cell constituent,
as distinguished from the whole, is responsible for these move-
ments. The centrosomes are not the sole dynamic or kinetic
centres of the cell, since the movements of the cytoplasm carry
the centrosomes where they will, and control the direction of
division and the relative size and quality of the daughter-cells.
In this, as in other phenomena, the cell acts as a whole, and in
the interaction of its various parts are to be found the causes
of all vital phenomena.

1 My complete paper on the history of these structures during maturation, fer-
tilization, and cleavage, will appear shortly.

PROTOPLASMIC MOVEMENT. 91

BIBLIOGRAPHY.

'96 Bambeke, van. Sur un groupement de granules pigmentaires dans

I'ceuf en segmentation d'Amphibieus. Bull. Acad. Belg. T. 31.
'97 Boveri. Zur Physiologie der Kern- und Zellteilung. Sitz.-Ber. Phys.

Died. Ges. z. Wiirzbiirg.
'96 Castle. The Early Embryology of Clone intestinalis. Bull. Mus.

Comp. Zool. (Harvard University). Vol. xxvii.
•92 Con KLIN. The Cleavage of the Ovum in Crepidula fornicata. Zool.

Atizeiger. Bd. xv.
'94 CoNKLiN. The Fertilization of the Ovum. Wood's Holl Lectures.
'97 CoNKLiN. The Embryology of Crepidula. /^?/r//. ^i^^r/^. Vol. xiii.
'98 CoNKLLX. Cleavage and Differentiation. Wood''s Holl Lectu7-es.
'98 Driesch. Ueber rein-miitterliche Charactere an Bastardlarven von

Echiniden. Arch. f. Entw. d. Mech. Bd. vii.
'95 Gardiner. Early Development of Polychoerus caudatus. Journ. of

Morph. Vol. xi.
'95 Heidenhain. Cytomechanische Studien. Arch. f. Entw. d. Mech.

Bd. i.
'96 Jennings. The Early Development of Asplanchna Hefickii. Bull.

Mus. Co7iip. Zool. (Harvard University). Vol. xxx.
'94 KoFOiD. On Some Laws of Cleavage in Limax. Proc. A7ner. Acad.

Arts and Set. Vol. xxix.
'95 LiLLiE. The Embryology of the Unionidae. Journ. of Morph.

Vol. X.
'95 LoEB. Beitrage zur Entwicklungsmechanik der aus einem Ei entste-

henden Doppelbildung. Arch, f Entw. d. Mech. Bd. i.
'95 LoEB. Untersuchungen iiber die physiologischen Wirkungen des

Sauerstoff mangels. Pfluger''s Archiv. Bd. Ixii.
'95 McAIuRRiCH. Cell Division and Development. Wood's Holl Lec-
tures.
'94 Mead. Preliminary Account of the Cell Lineage of Amphitrite and

other Annelids. Journ. of Morph. Vol. ix.
'93 Morgan. Experimental Studies on Echinoderm Eggs. Anat. An-

zeiger. Bd. ix.
'93 NussBAUM. Ueber die Verteilung der Pigmentkornchen bei Karyoki-

nese. Anat. Anzeiger. Jahrg. viii.
'83 Rauber. Neue Grundlegungen zur Kentniss der Zelle. Morph.

Jahrb. Bd. viii.
'94 Roux. Discussion of Ziegler's paper, " Furchung unter Pressung."

Verhandl. Anat. Gesellsch., Achte Versafnml.

92

BIOLOGICAL LECTURES.

'94 Ry"der. Dynamics in Evolution. Wood's Holl Lectures.

95 Wheeler. The Behavior of the Centrosomes in the Fertilized Egg of

Myzostoma glabrum. Journ. of Morph. Vol. x.
'87 Whitman. Ookinesis. Journ. of Morph. Vol. i.
'96 Wilson. The Cell in Development and Inheritance.
'96 ZuR Strassen. Embryonal Entwicklung der Ascaris megalocephala.

Arch.f Entw, d. Mech. Bd. iii.

FIFTH LECTURE.

EQUAL AND UNEQUAL CLEAVAGE IN
ANNELIDS.

A. L. TREADWELL.

Investigations in cell lineage, the line of work contemptu-
ously regarded by some morphologists as a process of *' cell
counting," have given us in the last few years some important
data concerning the early cleavage stages of animals belonging
to a number of different classes, the best-known groups being
the annelids, molluscs, and platodes. We are now in possession
of a sufficient number of facts to draw some conclusions con-
cerning the cleavage of the ovum and the formation of the
embryo. Some of the results of these investigations, bearing
upon questions of cell homology, differentiation, etc., have
been discussed by other lecturers in this course, and I shall
therefore limit myself to problems connected with the early
cleavages.

Concerning cleavage in general, its origin and meaning, very
diverse views are held by different writers. On the one hand,
it is maintained that there is from the beginning a definite
organization of the ovum, and that cleavage merely brings this
organization to light by isolating certain areas by means of cell
boundaries. On the other hand, it is argued that the ovum is
at the beginning isotropic, and that differentiation only begins
after cells are formed, and as a result of their mutual action
upon one another. It is maintained by some that the similari-
ties which appear, for example, between the cleavage stages of
an annelid and a platode are of as great value in establishing
relationships as are the resemblances which exist between the
so-called homologous parts of the adult animals ; in other words,

93

94

BIOLOGICAL LECTURES.

that cleavage has a phylogenetic significance. Others regard
these resemblances as superficial and accidental, and cleavage
as having no phylogenetic significance whatever, except as a
possible indication of the mode of origin of the primitive
metazoon from its protozoon ancestor.

It is noteworthy in this connection that the theory which one
adopts is apt to be dependent upon the animal which he has
studied. In some groups, for example, in many annelids, mol-
luscs, platodes, ascidians, and nematodes, the cleavage always
follows a definite form or pattern, and, apparently, is so corre-
lated with differentiation that Conklin has proposed to designate
it as determinate^ as opposed to the indeterinutate type found
in many vertebrates, echinoderms, and Cnidariay where the
cleavage seems to be very indefinite and inconstant. Investi-
gators who have worked on the former groups are generally
inclined to give greater importance to cleavage than those who
have studied the latter. No one who" has followed the cell
lineage of an annelid or a mollusc can deny an apparent differ-
entiation, manifest even in the earliest cleavages, and the notion
of such differentiation seems not disproved by the results of
the experimental embryologists. The argument in this line has
been so ably presented in another place ^ that I will not attempt
to repeat it here, but pass to the more especial discussion of
equal and unequal cleavages.

It is a fact, familiar to all who have ever seen a segmenting
ovum, that the first cleavage in all holoblastic eggs divides the
cell into two equal or unequal parts, according to the particular
animal studied, the widest possible variations appearing among
closely related animals. Inasmuch as unequal cleavage is usu-
ally associated with the presence of a considerable amount of
food yolk, it was formerly supposed that yolk alone is respon-
sible for this form of division. A careful study of the develop-
ing ovum shows, however, that this explanation will not apply
in all, or even in the majority of cases, and I shall offer some
suggestions, based on a study of the annelids, which may,
perhaps, throw some light on the subject. For this purpose

1 Conklin, " Embryology of Crepidula," Biol. Lectures (1896, 1897), zxid fottrn.
of Morph., vol. xiii.

CLEAVAGE IN ANNELIDS.

95

I shall make a somewhat detailed comparison of Ajnphitrite}
which has been studied by Dr. Mead, as an example of unequal,
and Podarke, on which I have been working for some time, as
an example of equal cleavage. In the unequal type, to which
most of the annelids and all of the gasteropods and lamelli-
branchs which have been studied belong, frequently at the two-,
and always at the eight-cell stage, differences in size appear
among the segments ; and, supposing the first and second cleav-
age planes to divide the body into four parts, the quadrmtts are
never exactly symmetrical ; so that, from the very beginning,
orientation is comparatively a simple matter. In Amphitrite

Fig. I. —Amphitrite, four-cell stage.

Fig. 2. — Podarke, four-cell stage.

the first division is unequal, and this difference in size is con-
tinued into the four-cell stage (Fig. i).^ Here one cell is
very much larger than either of the other three, and lies at
what will be the posterior end of the future embryo. Notice
here that the second plane of cleavage is not a straight line,
but it cuts the first in two points some distance apart. The
portion of the second furrow between these two points is the
so-called " cross furrow," or '' Brechungslinie." I shall follow
Conklin in calling it the '' polar furrow." It is" obviously
formed by the rotation of cells A and C to the left, so that
they lie partly over ^ and D. In Amphitrite, as in other

1 Mead, " Development of Marine Annelids," ybwrw. of Morpk., vol. xiii.

2 I have modified slightly the lettering in the figures taken from Mead to corre-
spond with a nomenclature proposed for Podarke.

96

BIOLOGICAL LECTURES.

annelids and molluscs, this "polar furrow" retains its original
direction until a late stage of the development, and is of great
service in orientation.

At the next division each cell divides equatorially, so that
there results an eight-cell stage, having four cells above and
four below, the four upper being slightly smaller than the
lower. (See Fig. 3 of an eight-cell stage from below.) Note
especially the large size of D and the difference in size between
the upper and lower quartettes. Because of this difference in
size, we may speak of the upper quartette as micromeres, and

Fig. 3. — Amphitrite, eight cells from below.

Fig. 4. — Podarke, eight cells from the side.

the lower quartette as macromeres, the distinction between the
two quartettes being much less noticeable here than in some
other annelids.

This process is repeated during the formation of two more
groups of micromeres, which form the body ectoderm, and a
fourth group, one member of which forms the mesoderm, while
the other three unite with the remaining macromeres to form
the entoderm. Meanwhile cells earlier formed have divided,
so that the separation of the mesoblast cell coincides with the
completion" of the ideal sixty-four-cell stage, a stage which,
owing to delay in the division of some cells and a hastening
of the division of others, may never actually be attained.

A marked feature of this formation of micromeres is the
division of each group in an alternating spiral direction. The
eight-cell stage (Fig. 3) arises from the four-cell by a division

CLEAVAGE IN ANNELIDS. 97

of each cell and a rotation of the upper portion in a clockwise
direction. The next division to the sixteen-cell stage is anti-
clockwise ; so that we can speak of an alternating direction of
cleavage as characteristic. This alternating direction is often
obscured in cases where the segmentation does not follow a
regular geometrical progression ; e.g.. Nereis, Crepidula, and
Unio, and beyond the sixty-four-cell stage it is lost more or
less rapidly in all. Why cleavages should always take such
definite directions we do not know. Mechanical conditions
undoubtedly play a large part in the arrangement, but mechan-
ical explanations are not sufficient to account for the definite
direction assumed by each cleavage. The fact that in molluscs
with a dextral shell the first rotation is to the right, while in
those with a sinistral shell it turns to the left, may indicate
some relation between the form of the first divisions and the
symmetry of the future body ; but in the present state of our
knowledge such an assumption would be premature.

The important points in the development thus far are these :
I. At the sixteen-cell stage a separation of certain cells form
the first generation of ectomeres, which go to form the primary
prototroch, and are, therefore, trochoblasts. These divide twice,
and twice only ; so that in the sixty-four-cell stage there are
sixteen of these cells, which become ciliated and form the larval
locomotor apparatus. 2. The second micromere, separated from
the large macromere D, is larger than any of the other members
of the same quartette, and, as further study shows, forms a
large part of the ectoderm of the future trunk of the animal.
According to Dr. Mead, all of the ectoderm of the body behind
the first septum arises from this cell. 3. The separation of the
mesoderm, in large part at least, in the fourth micromere in
quadrant D. 4. The separation of four rosette cells in the
upper pole, which later divide once, and carry the apical tuft of
cilia, which is such a marked feature of the larva. 5. The
segregation of the entoderm in the macromeres, which invagi-
nate to form the alimentary canal.

Evidently, then, the ^g^ of such an animal as Amphitriie
possesses from the beginning a definite orientation. This is
indicated by the ** polar furrow/' which always has a definite

98 BIOLOGICAL LECTURES.

relation to the future body axes, and by the fact that the large
cell D always lies at the posterior end of the animal. Compari-
son with other annelids shows that wherever this difference in
size appears, the large cell always lies at the posterior end of
the body, and from it are given off the cells 2d and ^^d, which
are larger than the corresponding members of the same quar-
tette. In Ufiio, among the molluscs, a similar relation appears.
In other molluscs apparent exceptions occur, of which I will
speak later.

Let us now turn to a consideration of equal cleavage among
annelids. For an explanation of this type I will take Podarke^
as the one whose development has been most completely stud-
ied. The four-cell stage of this form is shown in Fig. 2. All
the cells are equal in size, and the only orientation point which
presents itself is the " polar furrow." Observation of the living
Qgg shows that this " polar furrow " remains constant in direc-
tion up to a comparatively late stage in the cleavage, when it
becomes possible to orient the Qgg by means of other landmarks.

In the next figure (Fig. 4) is shown the eight-cell stage, seen
from the side. Instead of a set of smaller micromeres, lying
on top of four larger macromeres, we have eight cells, all
exactly alike in size ; and only by means of the polar globules
are we able to distinguish the animal from the vegetative pole
of the tgg. The "polar furrow " enables us to distinguish the
second from the first plane of cleavage ; but there is no differ-
ence in the size of the cells, as in Ampkitrite, by means of which
one end of this furrow can be distinguished from the other.

At the next division a second group of micromeres is formed,
and, by a division of the first group, four smaller cells, which
are the parent cells of the primary prototroch. These, exactly
as in Ampkitrite, divide twice, and twice only, and form the
beginning of the larval locomotor organ, which becomes func-
tional at the completion of the sixty-four-cell stage.

The cells of the second quartette, given off at the sixteen-cell
stage, are all of the same size, in contrast to what occurs in
Ampkitrite, where 2d is so much larger than the other members
of the same quartette.

As in Ampkitrite, a third quartette of ectomeres is given off,

CLEAVAGE IN ANNELIDS.

99

and these are followed by the formation of a fourth quartette,
of which three of the members become mainly entoderm, while
a fourth contains the mesoderm. This fourth micromere, aris-
ing from the posterior cell jD^ is of the same size as the other
members of the quartette, and at the sixty-four-cell stage is only
distinguishable from them by its position. But, while the others
enter into the formation of the entoderm, this divides bilaterally
at the surface, and, after dividing off small entoderm cells (the
" secondary mesoblast " of Wilson), each divides to form the
mesoblast of the corresponding side. This cell, ^, is separated
from jD at the forty-cell stage, and the sixty-four-cell stage,
which is actual rather than ideal in this animal, is completed
by the division of the primary trochoblasts, the division of the
intermediate cells lying between the arms of the cross, and by
the division of the second quartette of micromeres, now two in
number in each quadrant. The latter division presents some
features of importance. In each of three quadrants the ecto-
meres divide in such a way that, of the resulting four cells, two
smaller are above, and two larger are below. In the remaining
quadrant, on the other hand, while one of the cells divides like
those of the other quadrants, the other divides in a different
way, and there is given off a large cell above, and a very small
cell below, which has a deeply staining nucleus, and is an
important landmark in the further study of the embryo. This
small cell, as later study shows, lies directly over the micro-
mere ^d, and since the descendants of 2d are conventionally
distinguished by the term X, this cell {2d2, 2) would be in this
system of nomenclature Xi, 2. Exactly this cell has been found
in a number of other annelids, including those with the most
unequal cleavage, and in Uriio^ and Crepidida, among the mol-
luscs. I have followed the living ^g^ up to the stage where this
cell appears, and find that the '* polar furrow " retains its posi-
tion, at least, until then — a position which agrees with the
position assumed in Amphitrite and Arenicola? By the aid of
this cell, X\, 2, and the mesoblast cell, which lies immediately
under it, orientation of the embryo is complete. t

1 Lillie, "Embryology of the Unionidae,"y<?Mr«. of Morph., vol. x.
'^ Child, " Cleavage of Arenicola cristata," Zo'dl. Bulletin, vol. i, No. 2.

lOO

BIOLOGICAL LECTURES.

A comparison of the embryo of Podarke with that of Amphi-
trite at the sixty-four-cell stage will be instructive. Fig. 6
represents the completion of the sixty-four-cell stage in Podarke
by the division of the descendants of the second group of micro-
meres. Note especially the small size of the macromeres,
^^A, 4B, 4C, 4D, the equality in size between the members of
the fourth quartette of micromeres, and the arrangement of
spindles in the cells undergoing karyokinesis. At this time
the embryo of Podarke is apparently radially symmetrical, a
symmetry which is only apparent, as is indicated by the position
of the "polar furrow." This apparent symmetry is destroyed

Fig. 5. — Amphitrite, sixty-four cells from
below ; cp-dp, primary trochoblasts.

Fig. 6.

Podarke, fifty-six to sixty-four
cells from below.

by the completion of the division shown in the figure, separating
the small cell Xi, 2, which comes to lie just over^</, the mesoderm.

From this time on, bilateral divisions appear among the cells
of both hemispheres, the most important being those which
lead to the formation of the bilaterally symmetrical cross at the
upper pole, and the bilateral division of ^d with the cells of the
third quartette of micromeres, which immediately adjoin it at
the lower pole. The last division is of especial interest, as we
shall see, as it leads to the formation of a larval mesoblast.

Let us now compare this with Amphitrite. Fig. 5 represents
an embryo of Amphitrite at approximately the sixty-four-cell
stage. Note the large macromeres, and the large size of
2d^=X and 4d=.M when compared with other members of

CLEAVAGE IN ANNELIDS.

OI

the corresponding quartette, while in Podarke these cells are
no larger than corresponding celb in the other quadrants.

Fig. 7, of Amphitrite, and Fig. 8, of Podarke, show the upper
poles of embryos in practically the same stages as Figs. 5 and 6.
Note here the comparatively large size of the cross-cells in
Podarke, composing a larger portion of the umbrella than do
the corresponding cells in Amphitrite {cf. Figs. 7 and 8).

Leaving the question of strictly cell development, let us pass
to the later embryonic history and note how the trochophore
arises from the embryo of sixty-four cells. The differentiation
of the prototroch cells enables us to distinguish an upper

Fig. 7. — Amphitrite, about sixty-four cells
from above ; ap-dp, primary trochoblasts.

Fig. 8. — Podarke, forty cells from above ;
ap'-dp' , trochoblasts before division .

umbrella from a lower subumbrella region. Later, the proto-
troch ring is partially completed in Amphitrite by the addition
of three cells in each of three quadrants from the lower hemi-
sphere, the fourth, that of D, remaining open for some time.
Through this opening a number of cells from the upper hemi-
sphere invaginate and form a portion of the subumbrella ecto-
derm. The descendants of 2d, or X, have meanwhile multiplied
rapidly, and compose the greater part of the trunk ectoderm.
As I said before. Mead believes that all of the trunk ectoderm
behind the first segment is composed of the descendants of
these cells. Meanwhile certain of them become ciliated and
form the paratroch, a band of cilia incompletely surrounding
the body just in front of the anus. At this time, when the

I02 BIOLOGICAL LECTURES.

paratroch is differentiated, the embryo of AmpJiitrite contains
about two hundred cells, and the mesoblast bands are composed
of four cells each. At this stage the large invaginated ento-
derm cells are still solid, and have not formed the cavity of the
alimentary canal, and the large prototroch cells are very promi-
nent on the surface of the larva.

In general, the development of Podarke through these
later stages agrees with AmpJiitrite^ but from the standpoint
of strictly cellular development important differences appear.
The prototroch is at first completed by one "intermediate"
cell {i.e.^ one of the cells which originally lay between the arms
of the cross), and later two more cells come into it from the
lower hemisphere. These latter acquire their cilia very late,
after the trochophore is fully formed. Cells from the upper
hemisphere also migrate through the dorsal break in the proto-
troch, but instead of giving rise to a very small portion of the
subumbrella ectoderm fully three-fourths of this latter arises
from them. The descendants of 2d form a narrow band of
ectoderm around the proctodoeum, and some of them pass into
the wall of the latter. In Fig. 12 the portion of the trocho-
phore behind the dotted line, x, ;r, is the only part whose ecto-
derm is formed by the X cells. I have been unable to discover
any trace of a paratroch.

In Podarke, also, immediately after the formation of the fifth
quartette of micromeres, which happens about ten hours after
fertilization, the entoderm plate invaginates and rapidly gives
rise to the alimentary canal, which is pretty definitely formed
by the twentieth hour, and the trochophore must begin to feed
very soon after this time. In AmpJiitrite, on the other hand, if
I understand Dr. Mead correctly, the alimentary canal is formed
much later, — about sixty hours, — and considerably later the
embryos settle down to the bottom and begin to feed. While,
too, there is a similarity in the method of formation of certain
organs, there is a marked difference, becoming more and more
noticeable as the development progresses, in the relative rate
of development of different parts. The early stages of PodarJ^e
and of AmpJiitrite are passed through in about equal spaces of
time, and both begin to swim immediately after reaching the

CLEAVAGE IN ANNELIDS.

103

sixty-four-cell stage, about five hours after fertilization ; but,
from this time on, important differences are to be noticed.
Briefly stated, these differences may be said to be connected
with the more rapid development of the descendants of 2d and
//^ in AmpJiitrite than in Podarke. Figs. 5 and 6 show the rela-
tive development of these two sets of cells in early stages, and
Figs. 9 and 10 show the proportional size of the descendants
of these cells in a later stage. In the figure of AmpJiitrite the
mesoblast consists of three cells on a side, while in Podarke
there is only one, the third division of ^d resulting in the sepa-
ration of two small cells, which enter into the formation of the

Fig. g.

Fig.

Figs. 9 and

A mphitrite and Podarke, respectively, in a late stage of segmentation ;
BL, blastopore.

wall of the archenteron. Similar cells have been described for
Nereis and Aricia by Wilson,^ and for Crepidula by Conklin.
Note in the figures the relative development of 2d and its
descendants (indicated by the letter x), when compared with
one another and with corresponding cells of the other quadrants
(indicated by Fig. 2). Evidently in Amphitrite the descendants
of 2d form a large part of the subumbrella, and the difference
between the development of these cells here and in Podarke^ a
difference which was noticeable in earlier stages, has now become
much more pronounced.

1 Wilson, {a) "Cell Lineage of Nereis," y<?«r«. of Morph., vol. vi; {b) "Consid-
erations on Cell Lineage," etc., Annals of the N. Y. Academy of Sciences, vol. xi,
No. I.

I04

BIOLOGICAL LECTURES.

Extending the comparison into the later larval stages, we see
that this difference still persists, and is even more noticeable

than in the earlier. In the larva of
AmpJiitrite of forty hours there is a
very marked development of meso-
dermal and trunk ectodermal tissue,
or, in other words, a very large
amount of material derived from
the descendants of 2d and /f-d (Fig.
II, Amphitriie of forty-four hours,
from Mead). Amphitriie has four
well-marked body segments, with
strong setae on two of them (ecto-
dermal structures which have arisen
from 2d)y and, according to Dr.
Mead, well-developed mesodermal
tissue lining the gut and body wall.
To be noted, also, is the relatively
small development of the umbrella
portion (the portion in front of the
prototroch) and the extreme width
of the prototroch itself. The large
mucous glands and the "proble-
matic bodies " of Mead occupy a
large part of the surface of the um-
brella, the former extending back
into the subumbrella portion.
Turning now to Podarke^ we find a very different condition
of affairs. The most noticeable feature is the complete lack of
any trace of segmentation. The subumbrella has elongated con-
siderably, but shows no trace whatever of a division into somites.
The umbrella is relatively much larger than in Amphitriie, a fact
which is undoubtedly correlated with the unusual development
of the cells comprising the cross in the former animal. I have
not succeeded in finding any mucous glands, but what are prob-
ably identical with Mead's ''problematic bodies" and Wilson's
"frontal bodies" are found lying just in front of the small eye
spot on either side {prob., Fig. 12, Podarke of forty-two hours).

Fig. II. — Amp}iitrite{zi\.t.r
Mead), forty-four hours.

CLEAVAGE IN ANNELIDS.

05

Fig. 12. — Podarke, forty-two hours; prob.,
problematic bodies.

From their staining reactions I infer that in Podarke, as in
Nereis, these organs have a glandular function. A ventral
band of cilia is
present, but I
have not been
able to discover
any paratroch.

The enteron
has grown and
expanded so as
to completely ob-
literate the body
cavity. Owing to
the slow develop-
ment of muscular
tissue on the dor-
sal surface of the
larva, the body
wall is composed
of two very thin layers, ectoderm and endoderm, while the
body wall on the ventral side is much thicker, and the amount
of mesodermal tissue is evidently much greater. At present I
am unable to state just what are the relations of the meso-
dermal portion of the larva of this age, but an investigation
now in progress will undoubtedly clear them up. Certain
it is that the amount of definite mesoderm is very small when
compared with such a form as Amphitrite, for in an earlier stage
of approximately twenty hours, when Ainphitrite has already
begun to elongate and the germ bands are composed of "a great
many " cells (Mead does not state how many), Podai-ke has no
more than two mesoblast cells on a side. The young larva is
very muscular, undergoing all sorts of distortions when irritated;
but this activity is produced by the larval mesoblast cells de-
rived from J</, J^, and ^a. These cells wander apart from one
another, elongate, and form the larval musculature. As said
before, I am not yet sure about the relations between these two
sorts of mesoderm and the other body layers during the later
larval stages ; but I believe that it will be found that by far the

I06 BIOLOGICAL LECTURES.

larger part, if not the whole of the musculature of the trocho-
phore, comes from this larval mesoderm, and that it is only
as the larva elongates that the definitive mesoderm takes any
important part in body formation.

For this comparison I have used Amphitrite as being the most
convenient for the purpose, but exactly similar results would
have been attained had I used any of the other annelids with
unequal cleavage, whose cell lineage is at all accurately known.
Of the annelids with equal cleavage, Podarke happens to be the
only one which has yet been studied with any completeness ;
but, so far as one can judge from the figures given by Hatschek^
for Eupomattts, Van Drasche^ for Pomatoceros^ and Mead for
Lepidonotiis, similar results would be obtained there, were the
development fully studied. The trochophores of Eiipomatus
and Pomatoceros are thin walled, and with exactly the feeble
development of ectodermal and mesodermal tissue that has been
noted for Podarke. The same is true of Hydroides^ as I am
informed by Dr. Wilson, and of Lepidonottis, according to Dr.
Mead. Connected with the absence of a large amount of yolk in
these forms is the rapid development of the archenteron, so that
the embryo begins to feed much earlier than in the other case.

How now are we to explain the difference in manner of
cleavage which has been observed in these forms.? Why does
one Qgg divide with absolute equality, while others, little or no
larger, divide very unequally } Within certain limits, undoubt-
edly, presence or absence of yolk does affect the form of
cleavage. It can hardly be .disputed that the peculiar form
of cleavage found, for example, in the egg of the teleost is
caused by an accumulation of yolk at the lower pole, or that a
similar cause must be invoked to explain the form of segmenta-
tion of the Qgg of the bird ; and earlier writers assumed almost
universally that to the effects of yolk, either actually present,
or exerting its influence through the action of heredity, must
be ascribed the differences in the manner of cleavage. Rabl ('79),
in his work on Planorbis,^ finds a series of molluscs in which,

1 " Entwick. von Eupomatus uncinatus," Arbeiten a. d. htst zu Wien, 1885.

2 Beitrage zu Entw. d. Polychaeten. Wien, 1884.

3 "Entwick. der Tellerschnecke," Morph. Jahrb., 1879.

CLEAVAGE IN ANNELIDS. 107

at the four-cell stage, are found all gradations from four equal
cells, through others with an accumulation of yolk in three,
then in two, and finally in one cell, the yolk-laden cells being
larger than the others. From these he concludes that unequal
cleavage is caused first by an accumulation of yolk, and second
by an unequal distribution of this yolk ; e.g., if there is a
difference between " micromeres " above and " macromeres "
below, it is because the yolk is collected at the lower pole of
the Qgg ; if a difference in size appears at the four-cell stage,
it is because of an unequal distribution of the yolk among the
blastomeres, etc. As far as I can discover, however, Rabl
makes no attempt at explaining why the yolk should have this
unequal distribution. While, as far as the first three cleavages
are concerned, this explanation might suffice, the most super-
ficial observation will show that it does not apply at all in later
stages. There cells of the same size divide equally or unequally,
and that, whether there is any yolk present or not. A glance
at any plate of illustrations of a cell-lineage paper would show
plenty of examples of this. In a discussion of this subject, Lillie
{loc. cit.j p. 45) comes to the following conclusion : '' Unequal
cleavage is conditioned by the constitution of the segmenting
ovum, and always means the precocious localization of an organ
or set of organs in the larger cell. This organ may be the
entoderm, in which case it is usually accompanied by yolk ; but
the inequality of the first two cells in the annelids and molluscs
is the earliest visible indication of another differentiation, the
larger cell containing the two somatoblasts." This explanation
seems to me in perfect accord with the facts.

Turning now to equal cleavage, we might suppose it to be
due to a lack of differentiation in -the early stages, and, in fact,
this explanation has been offered. Rabl ^ ('76) tried to explain
the difference between the unequal cleavage of a mollusc and
the equal cleavage of such a form as an ascidian by the assump-
tion that in the former there is an early differentiation of parts
manifesting itself in cells of different sizes, which does not
appear in the latter. This early differentiation, being of distinct
advantage in the struggle for existence, probably has developed

1 "Entwick. der M3.\Qxmnsche\" Jenaische Zeitschrift, 1876.

I08 BIOLOGICAL LECTURES.

for that reason, while in the ascidian the equal cleavage indicates
that differentiation begins much later.

From the facts which I have given above there can be no
question that in Podarke there is as great a differentiation as
in any annelid of the unequal type. The position of the «* polar
furrow," the appearance of the cell Xi, 2, and the perfectly
regular formation of definite cells at definite times, all indicate
this beyond the shadow of a doubt.

In a note published about a year ago ^ I ventured a suggestion
concerning the meaning of this equal cleavage, and what I have
seen since then confirms me in my original opinion. Accepting
the principle laid down by Lillie, and further developed by
Conklin {Biol. Lectures, 1896, 1897), that the initial size of a
blastomere stands in direct relation to the size and the time of
formation of the part to which it gives rise, I pointed out that
in all probability the small size of 2d and ^d is connected with
the slow development of the parts of the embryo arising from
these cells. The facts upon which this conclusion is based will
be evident from the comparison which I have made between
Podarke and Amphitrite, where the slow development of ecto-
derm and mesoderm in the former (the descendants of 2d and
4d), and the rapid development of the corresponding portions
of the embryo in the latter, are noticeable features of the ontog-
eny. In Podarkey also, a large part of the embryo, which is
composed of definitive mesoderm in other annelids, is made up
of the larval mesoderm mentioned above, — an additional reason
for the slow development of ^d. I believe further that absence
of yolk is not alone responsible for the equality of the first
three cleavages ; but that the fact that the third division (form-
ing the eight-cell stage) is equal is connected with the large
size of the umbrella, relative to the subumbrella of Podarke,
when compared with other annelids, and with the fact that a
large part of the subumbrella ectoderm really comes from the
umbrella, having migrated through the break in the proto-
troch.

We have already seen that at the sixteen-cell stage, and later,
the largest cells in the embryo are at the upper pole.

1 Zobl. Bulletin, No. 4.

CLEAVAGE IN ANNELIDS. IO9

To avoid misunderstanding, let me say here that in accepting
this principle of Lillie's, and its application by Conklin, I can
hardly follow the latter author in assuming as complete a cellu-
lar differentiation as he does. Inasmuch as we know that pro-
toplasmic continuity undoubtedly exists between blastomeres,
we are hardly at liberty to suppose that differentiation is com-
plete at the time of cell formation. In using the term " differ-
entiation " I mean that a certain amount of material destined
to form a certain organ is accumulated at a definite place, even
though this material at the time of its separation may be in an
undifferentiated condition. For example, I do not believe that
at the time of formatio7t 4d is necessarily mesoderm, but that it
is destined to develop into mesoderm ; and, according as it con-
tains a larger or a small amount of material, the mesoderm of
the body will be large or small.

This evidently is but a limited application of the principle
stated by Lillie, and it would be possible in the ontogeny of
Podarke to find many illustrations of the law which might,
perhaps, strengthen the position I have taken here. Inasmuch
as differentiation in cleavage forms, the subject of another
lecture in this course, I shall not attempt to describe them,
but will refer to Dr. Lillie's lecture for additional evidence in
favor of the theory.

Quite recently, as said before, it has been shown that the
cell 4d is not, as was originally supposed, purely mesodermal,
but that in some animals it constitutes in addition a certain
amount of entoderm. It might be supposed that the large size
of /j-d in these 'forms is due to this double function. Apart
from the fact that in certain annelids, Amphitrite} Polymnia,
Arenicola, no such entodermal portion has been found, — and
yet here the cleavage is very unequal, — is the further consid-
eration that in no case is the amount of entodermal tissue there
present sufficient to account for the difference in size. In
Alicia there is but one small entodermal cell budded off from
^d on either side ; while in Nereis, though some three or four

1 I can hardly agree with Wilson that the small cell described by Mead as
budded off from 4d during its migration inward is entoderm, inasmuch as its
permanent position is at the anterior end of the germ band.

no BIOLOGICAL LECTURES.

are formed from each product of ^d, they are very small,
"scarcely larger than polar globules," and make up but an
insignificant portion of ^d.

Among the molluscs the same principle applies. We have
already seen that it was the study of Unio which led Lillie to
an enunciation of the law ; and, from the results of his work
on CrepidtdUi Conklin ^ has stated that " the initial size of the
blastomere stands in direct relation to the size and time of
formation of the part to which it gives rise." Crepidttla^ indeed,
shows apparent exceptions, for the entodermal portion of /j-d
is as large as the mesodermal. Here, however, there is a
larval mesoblast, and the mesoderm derived from ^.d is, I
infer from Conklin's figures, comparatively small in amount.
In the earlier stages, also, 2d is no larger than its sister cells
of the same quartette ; but from this cell arises the shell gland,
which, according to Conklin, "appears late in the development."
Other exceptions are Umbrella? according to Heymons, and
Cymbulia^ (Fol), where D is the smallest cell of the four-cell
stage; but I believe that careful study would show that here,
as in the other molluscs, this small size is correlated with a
slow development of organs. Certain it is that in Umbrella^ at
least, where, if we can trust analogy at all, the shell gland should
arise from 2dj this organ develops very slowly when compared
with such a form as Unio. Unfortunately, too little is known
of the later stages to enable us to draw any very certain
conclusions.

Even in molluscs, where yolk appears to play an important
part in determining the size of cells, the constancy in the
arrangement of this yolk seems to me an indication that it
is distributed according to the needs of the embryo, and is
related to the greatest amount of protoplasm. If, for example,
all the yolk could be removed from the ^gg of Umbrella without
interfering with the normal development, I venture to suggest
that, while the ^g^ as a whole would be smaller, nearly the
same size differences would appear as in the normal ^gg. In

1 Biol. Lectures, 1896, 1897.

2 " Zur Entwick. von Umbrella," Zeit. Wiss. ZooL, Bd. Ivi.

8 " Sur le Developpement des Pteropodes," Arch, de Zool. Exp. et Gen., 1885. ^

CLEAVAGE IN ANNELIDS. Ill

other words, the greatest amount of yolk accumulates in the
largest cell, and is not in itself the cause of that larger size.

To summarize : equality of cleavage is not an indication of
lack of differentiation in the ovum, for definite cells appear
at definite places and at definite times, just as accurately as in
unequal cleavages.

Inequality of cleavage may be due to a limited extent to the
accumulation of yolk ; but a more important factor is the method
of arrangement of the formative material. If an organ is to
appear early in the ontogeny, or is very large, it will be repre-
sented in the cleavage by a larger amount of material than if it
appears late, or is small when formed ; and it is this accumula-
tion of material within what we call a "cell " that leads to the
"unequal," as opposed to the "equal " type. It is not merely
the segregation of material in a single cell, but the segregation
of a large amount of material, which produces the differences
between the two types ; i.e.^ this difference is not qualitative,
but quantitative.

SIXTH LECTURE.

THE CELL ORIGIN OF THE PROTOTROCH.

A. D. MEAD,
Brown University.

Notwithstanding the wholesome reaction against the fre-
quently reckless application of Baer's biogenetic law, larval and
embryonic characters still hold an important place as criteria of
the relationship of organic forms. The larval characters of the
tunicates are significant indices of the relationship of these
animals to the vertebrates, but perhaps no larval form has
rivalled the trochophore or trochosphere in provoking phylo-
genetic speculation. The splendid work of Roule, V Anatomie
CompareCy which has appeared within the last year, bears wit-
ness to the importance which is at present attached to this
larval type in establishing the relationship of such various
forms as the annelids, rotifers, bryozoans, mollusks, etc. All
these forms are grouped together by him into one branch of
the animal kingdom, the Trochozoa ; for, although the bodies
of the different Trochozoa are highly variable in form and struc-
ture, they all may be derived in the normal embryonic develop-
ment from the trochophore larva, which is constructed on one
constant organic plan.

The features characteristic of the trochophore arise in
various Trochozoa under very different conditions of embryonic
development, and are possessed by some adult animals, for
example, TrochospJicBra (Semper). These facts give deep
interest and wide bearing to the comparative study of the
embryonic origin of the trochophore itself.

One of the essential structural characters of this larval form

113

114

BIOLOGICAL LECTURES.

Fig. I. — Left side of fertilized unsegmented egg of
A mphitrite. Polar globules indicate the position
of the animal pole.

is the transverse ciliated fringe, the prototivcJi (Figs. 13-16),
and it is the purpose of this paper to bring together and

compare the observations of
several workers who have
employed the method of cell-
lineage to determine the exact
origin of this organ.

In order to establish a con-
crete example for comparison,
let us proceed at once to the
account of the origin of a
simple and fairly typical pro-
totroch, namely, that of the
marine annelid AmpJiitrite.

The spherical ^g^ is of
medium size, about one-tenth
of a millimeter in diameter,
without much yolk, and is
unprotected save by a thin
wrinkled ^^g membrane.
Fig. I represents the egg
seen from the left side.
Fertilization has taken place,
and the polar globules remain
attached. The succeeding
figures represent the ^gg in
the same position, unless
otherwise stated. Fertiliza-
tion takes place in the sea
water, and the cleavage pro-
gresses so rapidly that the
larva swims in a few hours.
The first cleavage being un-
equal, the 2-cell stage is char-
acterized by the possession
of a larger and a smaller
blastomere, which are represented in side view in Fig. 2. After
this the cells divide with some irrregularity in time and in the

Fig. 2.

— Left side of A mphitrite egg, 2-cell
stage, oriented as in Fig. i.

THE CELL ORLGLN OF THE PROTOTROCH.

Fig. 3. — Loxodromic curve (" left-handed spiral ") described
about a sphere, showing the general direction of the
cleavage furrows which cut the cells of the 2-, 8-, and
32-cell stages.

size of the resulting blastomeres, yet the irregularity is not so
great as to obscure the rhythm of cleavage, which expresses
itself in the following
manner :

Each of the two cells
divides, so that we have
a 4-cell stage ; then each
of these cells divides,
and so on, the number
of cells increasing in
geometrical progression
until the Q,g^ consists
of sixty-four cells. A
distinct rhythm is ob-
served also in the direc-
tion of cleavage, for the
cleavage planes lie ob-
liquely across the meridians of the ^g^. The cells in the 2-, 8-,
and 32-cell stages are divided by cleavage furrows which take
the general direction of the loxodromic curve in Fig. 3,^ which
may be rather loosely but conveniently defined as a left-handed

spiral described around a
sphere from pole to pole.
The cells of alternate gen-
erations, namely, of the 4-
and i6-cell stages, are cut
by furrows which have an
opposite direction, as rep-
resented in Fig. 4. These
curves may be called, for
convenience, left and right
respectively, following
the conventional nomen-

FiG. 4 —Loxodromic curve ("right-handed spiral"), clature for thc thrCads On
having the general direction of the furrows which cut
the 4- and i6-cell stages and many of those of the 64-
cell stage.

a screw.

The rhythm of division

The curve

1 In the diagram this curve cuts the equator at one point only,
will, of course, cut the equator at any point, if rotated.

ii6

BIOLOGICAL LECTURES.

Fig. s.

Atnphitrite, left side, four cells getting
ready to divide into eight.

ceases abruptly at the 64-cell stage. The cells of this stage
are, of course, all of the same generation, the seventh, counting
the ovum as one ; yet, as a consequence of the obliquity of the

cleavage furrows, the cells are
not directly superimposed, but
alternate with one another.

For many reasons it is help-
ful to conceive of the ^^g as
consisting of four quadrants,
each of which is composed of
the derivatives of one of the
cells of the 4-cell stage. Any
cell in one quadrant corre-
sponds in origin and in posi-
tion to a cell in each of the
other quadrants.

Let us consider the ^g'g in
the 4-cell stage (Fig. 5) as
composed of one quartette of cells about to divide. Four of the
descendants of each of these cells will belong to the primary
prototroch. In a similar way the ^gg in the 8-cell stage con-
sists of two quartettes of cells,
those of one quadrant meet-
ing at the anterior pole, the
position of which is indicated
in Fig. 6 by the polar glob-
ules, those of the other meet-
ing at the posterior pole. The
priffiary prototrocJial cells are
all descended from the cells
of the anterior quartette^ and
from that portion of the cells
which is stippled in Fig. 6.
The cells of this quartette, or,
in the latter stages their descendants, constitute the anterior
hem^isphere^ or umbrella, and. the cells of the posterior quartette
constitute tho, posterior hemisphere, ox subiim^brella. The hemi-
spheres are separated in the drawings by a dotted line.

Fig. 6.

• A inphitrite, left side, last phase
of 8-ceIl stage.

THE CELL ORIGIN OF THE PROTOTROCH.

117

After the next cleavage the i6-cell stage is formed, in which
the cells are disposed in four quartettes. Fig. 7 represents the
Qgg in this stage seen from the side, but slightly rolled to
show all the eight cells of the
anterior hemisphere. The
polar globules lie in the mid-
dle of the hemisphere. The
derivation of the cells from
those of the 8-cell stage is in-
dicated by the arrows, and the
furrow between each pair of
cells has the general direction
of the loxodromic curve in
Fig. 3. The dotted line in-
dicates the boundary between
the two hemispheres, and the
heavy lines separate the quad-
rants. The four cells which
are stippled in Fig. 7, one in
each quadrant, are the pri-
mary trochoblasts in the strict
sense, for they co7itain all the
material of the primary pi^o-
totroch and no other tnaterial.

By the division of the six-
teen cells represented in Fig.
7 the 32-cell stage is attained
(Fig. 8). It consists of eight
quartettes (four cells in each
quadrant), two of them, the
third and fourth, from the
animal pole, being derivatives
of the primary trochoblast
and stippled as in Fig. 7.
(Fig. 8 really represents the
beginning of the transition between the 32- and 64-ceir
stage. Two of the cells of the subumbrella hemisphere have
divided. These cells, however, do not concern us especially.)

Fig. 7. — Amphitrite, from left side and top, i6-cell
stage. Heavy solid lines divide the quadrants;
dotted lines divide the anterior and posterior
hemispheres.

Fig. 8. — Amphitrite, left side, 32-cell stage.

ii8

BIOLOGICAL LECTURES.

The spindles in the trochoblast cells indicate the direction of
their next division, which results in the 64-cell stage (Fig. 9).
Fig. 10 is an apical view of the anterior hemisphere in the

32-cell stage, showing the
four alternating quartettes
of umbrella cells, and the
uppermost quartette of sub-
umbrella cells. The cells
of the first quartette have
the nuclei figured ; those of
the second are left clear ;
those of the third and fourth
quartettes (primary trocho-
blast cells) are stippled. All
of the thirty-two cells divide
again, so that the ^g^ is
composed of sixty-four cells,
arranged in sixteen quar-
tettes, making sixteen cells
in each quadrant. Four of
these quartettes, the fifth, sixth, seventh, and eighth, stippled
in Fig. II, comprise the cells of th.Q primary prototroch.
It is obvious that, in consequence of the regularity of cleav-

FiG. 9. — A mpkitrite, 32-cell stage, from anterior end,
View of umbrella or anterior hemisphere.

■9'

Fig. id. — A mpkitrite, from left side. Primary
and secondary, shaded, respectively, with
dots and lines.

Dorsal

Ventral

Fig. II. — A mpkitrite, beyond 32-cell stage,
from anterior poles, slightly distorted to
show all 25 prototrochal cells.

THE CELL ORIGIN OF THE PROTOTROCH.

19

age, one-half of all the cells are descendants of the four upper
cells of the 8-cell stage (the anterior hemisphere or umbrella),
and that the sixteen primary prototroch cells all belong to this
hemisphere.

By means of the primary prototroch, which consists of four
isolated groups of cells, the ^g'g, or trochophore, as we may now
call it, swims about for a considerable time. In the course of
a few hours, however, nine more cells, secondary prototroch,
become differentiated and develop cilia, and these, together
with the sixteen cells already described, constitute the defini-
tive prototroch. As is shown in Fig,
9, these additional cells, which are
shaded with lines, are disposed in
such a manner that they fill the
spaces between the isolated portions
of the primary prototroch except in
one quadrant. Thus the definitive
prototrocJi consists of twenty-five cili-
ated cells wJiicJi lie in a nearly com-
plete belt around the trochophore
(Fig. 13, p. 120).

The primary and the secondary
prototroch cells, although together
they form a single larval organ, have
different modes of origin, for the
cells of the secondary prototroch are
derived from the lower (posterior)
quartette in the 8-cell stage, that is, from the posterior hemi-
sphere, or subumbrella.

For convenience of description, we may assume the existence
of a specific material which will become the secondary proto-
troch, and may locate this material in the subumbrella during
the successive stages of cleavage. In the i6-cell stage the
material occupies a portion of three of the cells of the third
quartette, counting from the upper pole, which alternate with
the primary trochoblasts. In Fig. 7, p. 117, the cell marked a^
and^ partially shaded with lines is one of the three cells in
question. The cell marked d"^ belongs to the quadrant which

Fig. 12. — Diagram of prototrochal
cells, seen from left side, \n Amphi-
trite, Clymenella, and Arenicola.
Two groups of primary and one of
secondary prototrochal cells.

I20

BIOLOGICAL LECTURES.

does not contribute any secondary prototroch cells. In the
32-cell stage the material is still in these cells of the highest
quartette of the subumbrella. One of them is shown in the side

vw

Fig. xz. — Amphitrite trochophore, showing apical tuft, prototroch, paratroch, stomadsma,
and one arm of mesoblast consisting of six cells.

view, Fig. 8, and all three in Fig. 9, shaded with lines. These
three cells may be called secondary trochoblasts, although they
contain a small amount of material (see unshaded portion of

Fig. 14.

Amphitrite,\&i\.%\&<t\ prototroch, paratroch; mouth at arrow ; outline of
archenteron stippled ; mesoderm stippled.

cell in Fig. 8) which does not enter the prototroch. With the
next cleavage each of these cells divides so that the six cells,
two in each of the three quadrants, contain the material of the

THE CELL ORIGIN OF THE PROTOTROCH.

121

mvi'^

mtuc

secondary prototroch. These finally divide again in the direc-
tion indicated by the spindles in Fig. lo. Thus each of the
three secondary trochoblasts is at last represented by a group
of four cells, q, r, j, and t, in Fig. 12". Three of these, ^, r, and
s, are about equal in size, and constitute the secondary prototroch.

The relative position of all
the prototroch cells is shown in
the apical view of the umbrella
hemisphere in Fig. 11, which
is slightly distorted and brings
the secondary prototroch cells
into sight. The shading accords
with that of the previous figures.
The primary prototroch cells are
stippled, the secondary shaded
with lines, and the minute prod-
ucts of the secondary trocho-
blasts are drawn with dotted
lines. The unshaded cells in
the upper margin between two
groups of primary prototrochal
cells interrupt the continuity of
the prototroch. They corre-
spond in origin, however, to the
secondary prototroch cells of the
three other quadrants. The re-
maining cell, t, Fig. 12 (dotted
in Fig. 11), is small, divides
again much later, and has an
unknown destiny. The inter-
ruption in the continuity of the
ciliated band in the mid-dorsal region is later obliterated by the
approach and concrescence of the prototroch cells from either
side. The completed prototroch, consisting of twenty-five
cells, persists for five or six days, until the larva has developed
several seta-bearing metameres.

Some of the later stages of the trochophore are represented
in Figs. 13, 14, 15, and 16. For a considerable time the outlines

Fig. 15. — Amphitrite. Prototroch, paratroch
closed ventrally. muc. gl., mucous glands;
fl., flagella; v.c, ventral band of cilia;
prob., problematic bodies which persist only
as long as the prototroch.

122

BIOLOGICAL LECTURES.

of the component cells of the prototroch can be distinguished,
but in late stages only a few of them can be made out in sur-
face view, though tangential sections prove that they still exist.
As the larva elongates, the prototroch extends over more surface,
but the cells become thinner. The cilia are fine, not very long,
and are thickly and evenly distributed. In the figures they are
represented in a purely diagrammatic manner, being much too
coarse and not nearly numerous enough to reproduce the actual
appearance. When the larva begins to elongate, a second belt
of cilia, the paratroch, develops around its posterior end. The
paratroch consists of four ciliated cells which are derived from
the cell marked d'^ in Fig. 7, p. 117, and, therefore, from that

med tent

Fig. 16. — Lateral view of Amphitrite just before the disintegration of prototroch, paratroch,
and mucous glands ; med. tent., median tentacle.

quadrant which did not contribute to the secondary prototroch.
At first the four cells lie in a nearly straight row, but the end
(lateral) cells gradually approach each other, and finally meet
on the ventral side, as may be seen by comparing Figs. 13, 14,
15, and 16. Other cilia and flagella appear in tufts or bands
over various parts of the body, while the larva is leading a
free-swimming existence ; but all these, including prototroch
and paratroch, suddenly disappear when the larva, about five or
six days old, settles to the bottom.

Figs. 13-16, when compared with the earlier drawings, illus-
trate the orientation of the body and the elongation of the
trunk. The anterior end is at the left of the picture, the pos-
terior end at the right ; the dorsal side at top, and the ventral
side at bottom. The anterior end is marked by the polar

THE CELL ORIGIN OF THE PROTOTROCH. 1 23

globules in the earlier drawings, and by the apical tuft of
flagella in Figs. 14-16. The median tentacle grows out just
ventral to the apical tuft. The posterior end is the middle of
the area which becomes enclosed by the paratroch cells, and
here the proctodaeum is formed (Fig. 16). The ''budding
zone " lies just in front of the paratroch, and it is by the rapid
growth in this region that the trunk elongates. The distance
between the prototroch and paratroch, therefore, continually
increases.

We stated at the beginning of this paper that, inasmuch as
the trochophore is a highly representative larval form, the
embryonic origin of one of its essential organs might prove to
be of especial value, and the history of the origin of the proto-
troch in the one species which we have reviewed strengthens
this conjecture and brings out some additional points of in-
terest.

The prototroch is differentiated as a definite functional
organ at a very early stage in development, when the whole
body consists of less than 1 50 cells. It is composed of com-
paratively few cells (twenty-five), which are constant in num-
ber, in size, in relative position, and in mode of origin, in all
individuals of the species. The portion of the larva differen-
tiated as the prototroch is sharply marked off from the surround-
ing areas by the cell boundaries, and the cells in the immediate
vicinity of the prototroch do not especially resemble the latter
either in behavior during cleavage or in ultimate function.

Having thus taken our bearings, we find ourselves in this
position. The trochophore is a larval form characteristic of a
large group of animals, annelids, gasteropods, rotifers, etc., and
the prototroch in one species, at least, arises always in the
same manner from certain cells through a definite process of
cleavage. Although the origin of the prototroch is compara-
tively simple, yet it is sufficiently complex to offer ample
opportunity for variation, and by comparing its cell-origin in
other animals we may hope to ascertain whether it always
arises from the same cleavage cells. If it does so arise, then
we may consider that its component cells are homologous, and,
of course, the homology of the prototroch itself will thereby be

124 BIOLOGICAL LECTURES.

Strengthened. If it arises sometimes from one group of cells
and sometimes from another, then its origin may be considered
either an accident of the cleavage processes which are without
morphological significance, or as indicating that the prototroch
is not homologous in the forms examined. To illustrate : let
us suppose that we were ignorant of the internal structure of
the limb in the vertebrates. This organ, being characteristic
of the higher vertebrate body, and having a general external
resemblance in different forms, would seem to be homologous
throughout the series. Now let us suppose that we have
ascertained the origin and arrangement of the component struc-
tures of this organ in one species man. If, on comparison
with other species, we find that the component structures of
the limb have the same arrangement and arise in the same way,
we may then consider these parts also to be homologous struc-
tures, and the homology of the entire limb is sustained thereby.
If, however, we find that the origin of the component struc-
tures is essentially different in the various forms, we may then
conclude either that the embryonic origin is without morpho-
logical significance, or that the limbs themselves are not
homologous.

We will begin the comparison of the trochophore of Amphi-
trite with those of other annelids, assuming that the prototroch
cells are homologous structures, and, in keeping with the illus-
tration we have just given, we may expect to find :

First, that the particular features of the cell-origin of the
prototroch in Amphitrite are not the result of the environmental
conditions which are peculiar to this ^g^, but that the same
origin obtains in other well-developed trochophores whose envi-
ronment differs considerably from that of Amphitrite.

Second, that in those annelids in which the prototroch is
more or less suppressed and develops at a late stage, the cells
which we have identified as trochoblasts are small and slow in
dividing.

Third, that in distantly related Trochozoa which possess a
larval structure similar to the annelid prototroch some similari-
ties obtain in the cell-origin of this structure.

Lastly, we might expect that annelid trochophores would be

THE CELL ORIGIN OF THE PROTOTROCH. 125

found in which the mode of origin of the prototroch is a modi-
fication of that in the assumed type, a modification analogous
to that in the skeleton of the vertebrates, due either to the
fusion or reduplication of parts.

Fortunately the few annelids in which the cell-origin of the
prototroch has been worked out represent different families.
Clymeftellay for example, belongs to the Maldaiiidce, a family
not closely related to the Terebelidce to which Amphitrite
belongs.

The ^g'g of Clymenella is nearly twice the diameter of that
of AmpJiitrite^ contains more yolk, and develops more slowly.
The upper hemisphere is relatively smaller, and there are many
minor differences in the rhythm of the cleavage and in the size
of the cells. The trochophore is similar in many respects, but
is less active. The definitive prototroch, however, consists of
twenty-five cells, forming a nearly complete band around the
larva. In the i6-cell stage we find the four primary trocho-
blasts which give rise, by two subsequent cell divisions, to
sixteen cells, which constitute the primary prototroch. Each
of the three secondary trochoblasts (32-cell stage) divides
into a group of four cells, of which three are larger and sub-
equal, while a fourth is small. The three larger cells in each
group become ciliated and complete the prototroch, while the
small one does not enter into it. In brief, the origin of the
definitive prototroch of Clymenella is identical, cell for cell, with
that of Amphitrite.

In Arenicola cristata we have a representative of another'
family of annelids, the Arenicolidce. The eggs of this worm
undergo their early development while encased in a gelatinous
capsule, and the cleavage differs somewhat from that of Amphi-
trite and Clymenella in rhythm and in the relative size of the
blastomeres ; but Dr. Child has shown that the primary proto-
troch and, later, the definitive prototroch originate in precisely
the same manner, cell for cell, as in Amphitrite and in Clyme-
nella.

Though, I think, the cell-origin of the definitive prototroch
has not been traced in other annelids, there are records of the
mode of origin of the primary prototroch in several.

126 BIOLOGICAL LECTURES,

In Lepidonottis it is composed of sixteen cells, which arise
exactly as in the annelids just described (see Figs, 6-11,
stippled cells). These cells put out cilia very soon after they
are formed, as they do in Amphitrite (Fig. 10). Later, three
of the gaps between the separated portions of the primary pro-
totroch are filled by other ciliated cells, while the mid-dorsal
interruption persists just as in Fig. 11 of Amphitrite. It is not
improbable, therefore, that the definitive prototroch of this
worm has the same origin as that of other forms. Lepidonotus
belongs to the Aphroditidce ; the eggs are comparatively small,
and the cleavage is equal, so that up to the 64-cell stage
it is not possible to distinguish one quadrant from another.
This fact makes it of especial interest that the gaps in the
primary prototroch are filled in only three quadrants instead of
in all four, for, if the destiny of the secondary trochoblasts were
dependent only upon \}!\^\x position in the cell complex, we might
expect that the destiny of the corresponding cells would be the
same in all four quadrants.

Podarke, which has been described by Treadwell, has a func-
tional primary prototroch of sixteen cells of the same origin
as in the other forms just described. (Mr. Treadwell has not
yet published the history of the secondary prototrochal cells.)
Podarke belongs to the family Nereidce. The ^gg is even
smaller than that of Lepidonotus^ and the cleavage is equal.

SthenelaiSy family AphroditidcBy according to Treadwell, also
has an equal cleavage, and a functional primary prototroch of
sixteen cells is derived exactly as in the five annelids which we
have already described.

We may now proceed to those annelids in which the mode of
origin of the primary prototroch appears to be a modification
of the type. Among these forms Hydroides is of special
interest.

Wilson and Treadwell have shown that the functional proto-
troch of Hydroides consists of eight cells, two in each quadrant ;
but the origin of the primary prototroch in Hydroides is different
from that of Amphitrite only in this respect, that the primary
trochoblasts divide only once instead of twice. In the seven
annelids which we have considered, the primary trochoblasts

THE CELL ORIGIN OF THE PROTOTROCH. 12J

of the i6-cell stage, stippled in Fig. 8, p. 117, contain material
for the prototroch only, and the phenomenon in Hydroides,
as compared with that of other species, seems to indicate that
a cell in which material of a specific structure only is segregated
may divide once, twice, or perhaps more times. As long as all
the descendants of the trochoblasts enter into the formation of
the structure in question, it does not matter how many divisions
take place. This view is further supported by what has been
found in Capitella. According to Eisig's recent account, we
find in this annelid a modification of the type in the direction
of multiplication rather than reduction. Though sixteen cells
are produced from the four trochoblasts, and all contribute to
the prototroch, they undergo still further division before they
become functional.

The first attempt to determine the exact cell-origin of the
prototroch was made by E. B. Wilson on the ^g'g of Nereis.
He described what now seems to be a peculiar variation from
the type. The four trochoblasts divide as in Amphitrite into
sixteen cells, but four of these cells, the upper one in each
quadrant, do not enter into the prototroch, but the other twelve
cells do so. More cells are subsequently added, but their
origin is uncertain. If this account is correct. Nereis is the
only annelid in which the primary trochoblast has been shown
to give rise to other than prototrochal cells.

We may now pass from a comparison of the prototroch and
umbrella hemispheres in those annelids which represent the
indirect type of development, and examine these features in
annelids of more direct development, in which the trochophore
is more or less suppressed.

The partially suppressed trochophore of Scolecolepis is de-
veloped from an ^^^ which contains an abundance of yolk and
is protected during its early development within the sand-tube
dwelling of the worm. In the 8-cell stage the four anterior cells
(upper quartette) are very small in comparison with the four
posterior ones. All eight cells divide so nearly at the same
time that a i6-cell stage results. Figs. 17 and 18 show that
the four stippled cells, which correspond to the primary trocho-
blasts of other forms, are far smaller than any of the other cells ;

128

BIOLOGICAL LECTURES.

not only are they smaller, but they are the most tardy in divid-
ing. After all the other cells have undergone another cleavage,
these trochoblasts remain undivided. For this reason the
next stage, which is shown in Figs. 19 and 20, consists of but
twenty-eight cells instead of thirty-two. Three of these cells
correspond exactly, of course, to the secondary trochoblast
of Amphitrite, Clymenella, and Arenicola, and are therefore
shaded as in Figs. 8 and 9, though they are very small in
Scolecolepis. Thus in Scolecolepis the smallness of the anterior
quartette of cells in the 8-cell stage, and the conspicuous

Fig. 17. — Scolecolepis, i6-cell stage, left side.
Primary trochoblasts stippled ; anterior hem-
isphere (umbrella) bounded by dotted line.

Fig. 18. — Scolecolepis, i6-cell stage, seen
from anterior end.

minuteness of the trochoblasts, as well as their tardiness in
dividing, correspond exactly with the small size of the umbrella
of the trochophore and the small size and late development of
the prototroch. This annelid, then, in its mode of development
stands intermediate between such forms as Amphitrite or Lepi-
donotuSy which possess a typical trochophore, and Rhynchelmis
and Clepsine (Fig. 22), in which the development is direct and
the trochophore undeveloped.

Nereis Dummerillii (von Wistinghausen) approaches even
nearer than Scolecolepis to the direct type. In the former the
anterior quartette (umbrella cells) in the 8-cell stage (Fig. 21)
are smaller still in proportion to the posterior quartette. With

THE CELL ORIGIN OF THE PROTOTROCH.

129

Fig. 19. — Scolecolepis, 28-cell stage, from left side.
Shaded like figures of A mphitrite.

regard to the origin of the prototroch, however, it must be
admitted that, according to von Wistinghausen's account, this
organ is formed from cells
which do not correspond with
any of the trochoblasts of
other annelids. His account
places us in the same predi-
cament with regard to the ori-
gin of the prototroch as with
regard to the origin of the
primary mesoblast cell. In
respect to both, if the account
is correct. Nereis Diimmerillii
is unique. In order to bring
this annelid into line with the
others, most writers on cell-
lineage have expressed doubt,
and rightly I think, as to the
accuracy of the account in
the case of the mesoblast. The origin of the mesoblast cell,
however, is definitely described by von Wistinghausen, while

he did not even pretend to
work out the prototroch cell
by cell. So we may, perhaps,
be permitted at least to sus-
pend judgment in regard to
the discrepancy in the origin of
the prototroch in this species.
In those annelids in which
the trochophore is still farther
suppressed, and which undergo
a direct development, e.g.^ in
Rhynchelmis and Clepsine, the
anterior quartette in the 8-cell
stage is still farther reduced.
Fig. 22 is from an unpublished
drawing of Clepsine, 8-cell stage, given me by Dr. Whitman.
We may say in general of all the forms which we have

Fig. 20. — Scolecolepis, 28-cell stage, from
anterior end.

I30

BIOLOGICAL LECTURES.

Fig. 21. — Nereis Dujtiineriliii {nh&T von Wisting-
hausen), 8-cell stage, from anterior end (" upper
pole").

reviewed, that when the cells of the anterior quartette are com-
paratively large and divide readily, the umbrella is large and
the trochophore active. When, on the other hand, the cells

of the anterior quartette are
smaller and divide less rap-
idly, the umbrella is smaller
and the trochophore less
active. Lepidonotiis, Amphi-
tritc, Nereis liinbata^ Cly-
nienella, Nereis Diimmerillii^
RJiyncJiehiiis^ and Clepsi7te
form a series of annelids in
which there is a gradual de-
crease in the relative size and
karyokinetic activity of the
anterior four cells and a cor-
responding decrease in the
size of the umbrella and the
activity of the trochophore.

To sum up : the cell-
origin of the prototroch has
been completely worked out
in three annelids, viz., in
Amphitrite, Clymenella, and
Arenicolay which represent
three distinct families. I71
all three the origin of the
prototi^och is identical, cell
for cell. In eight genera of
annelids which represent seven
families, the whole primary
prototroch is formed from
identical cleavage cells, the
four so-called primary trochoblasts of the l6-cell stage (Fig. 7),
and all the derivatives of these four cells enter into its forma-
tion. These genera are Lepidonotiis, Podarke, Sthenelais,
Hydroides, Amphitrite, Clymenella, Arefiicola, and Capitella.
In all but two the primary prototroch consists of sixteen

Fig. 22. — Eight-cell stage of Clepshie . Apical

view from drawing furnished by Dr. Whitman.

THE CELL ORIGIN OF THE PROTOTROCH. 131

cells — in Hydroides of eight, and in Capitella of more than
sixteen. Nereis limbata differs from these forms in that
four of the sixteen descendants of the trochoblast do not
enter the prototroch, and Nereis Dummerillii is the only
annelid, so far as I am aware, in which the primary proto-
troch has been thought to arise from other than the primary
trochoblasts.

Passing now to other representatives of the branch Trocho-
zoa^ we have at present among mollusks more or less complete
accounts of the cell-origin of the velum, in Neritina by Bloch-
mann, in Crepidtda by Conklin, in Planorbis by Holmes (prelimi-
nary), and in Ischnochiton by Heath (as yet unpublished). The
most elaborate published account is given by Conklin in his
beautiful work on the gas-
teropod Crepidtda^ and for
this reason we may con-
sider this form first.

Crepidida lays its eggs
in capsules, which are pro-
tected under the shell of
the animal. In two species
*' it is about four weeks
from the time the ova are
laid until the fully formed
escape from the ^gg cap-
sules," and probably much
longer in other species.
Hence it is not surprising

that the velum becomes Fig. 23. — Cr^/?V/?//«, i6-cell stage, shaded Uke previous
_ . , , . , figures, viewed from anterior end.

functional at a late period

compared with the annelid prototroch, for the annelid begins

to swim usually in a few hours after fertilization.

The general type of cleavage in Crepidtda is the same as
that in the annelids. Fig. 23 of the i6-cell stage of Crepidida
recalls in particular the same stage in Scolecolepisy whose eggs
are protected like those of Crepidula. In both, the cells of the
subumbrella, or posterior hemisphere, are very large and full of
yolk, while those of the anterior hemisphere are very small and

132

BIOLOGICAL LECTURES.

clear. The cells which correspond to the primary trochoblasts
are the smallest of all, and the slowest to divide. In Crepidula
and in other gasteropods these four cells are known as *' turret-
cells," but for the sake of uniformity in this paper I will call
them primary trochoblasts. They remain undivided, or at most
divide only once up to a very late cleavage stage. During
these stages, however, they undergo a remarkable growth, so
that finally, instead of being the smallest, they are the largest
cells in the umbrella. It is the destiny of the primary trocho-
blast, however, which is of special interest to us. On this

Fig. 24. — Three successive stages, representing the relative position of derivatives of the secondary
trochoblast q r s t'va. several forms. Diagram approaches most nearly the condition seen in
Amphitrite.

point Conklin says : " In four species of Crepidula^ at least the
two anterior ones form a portion of the preoral velum (proto-
troch), and this is probably true of the two posterior ones also."
In other words, the primary trochoblasts in Crepidula have
probably the same destiny as those of the eight genera of
annelids mentioned on page 130.

We may now consider the behavior of the secondary tro-
choblasts in Crepidtda — those three cells of the ideal 32-cell
stage, most of whose descendants in Amphitrite^ Clymenella^
and Arenicola complete the prototroch by filling three of the
gaps between its separate parts. In Crepidula tJie secondary

THE CELL ORIGLV OF THE PROTOTROCH. 133

trochoblasts arise in exactly the same mangier as in the annelids
named^ and tfiost, if not all, of their descendants, enter into the
prototroch. This general statement should be qualified by an
account of at least some of the details of the later cleavage,
in order that it may give a fair idea of the degree of resem-
blance- in question. For this purpose, and to avoid technical
expressions, we will make free use of diagrams and arbitrary
symbols.

The secondary trochoblast in each of the three quadrants
lies between the primary prototroch cells, as in Diagram 24,
which is from Amphitrite (compare Figs. 8, 10, and 12). It is
represented by the compound symbol qrst. The phenomena
are the same in three quadrants, but the corresponding cells
in the fourth quadrant never enter the prototroch. In both
Amphitrite and Crepidnla each trochoblast q, r, j, and t divides
obliquely in the same manner into two cells, qr and st, the
former lying a little anterior to the latter (Fig. 24 B). Later,
in both animals, each of these two cells divides in the reverse
oblique direction, making a group of four cells, q, r, s, t (Fig.
24 C). In the annelid q, r, and s, soon after their formation,
become ciliated and form a part of the prototroch in the manner
already described, while t, a very minute cell, does not do so.
In Crepidnla q and r enter the prototroch ; s divides also, its
upper product probably entering the prototroch, while the
lower product may enter it or may not. The remaining cell,
t, is very much larger than in Amphitrite, and at least part of
it probably enters the prototroch, thus giving rise to the only
probable discrepancy in the destiny of all the descendants of
the primary and secondary trochoblasts in the two forms.
What the significance of this discrepancy may be, I will not
venture to say, but I wish to call attention again to the fact that
the cell / in Amphitrite is exceedingly small, does not divide
for a long time, and has the appearance of a vestigial cell.

It must be said here that Conklin thinks it possible, although
by no means certain, that cells other than the trochoblasts enter
into the formation of the velum, and in respect to many of
the cells which we have already described there is also an
element of uncertainty, since the velum of Crepidula does not

134 BIOLOGICAL LECTURES.

become functional until a very late period of cleavage, so that
it is almost impossible to trace its cell-origin accurately.

Blochmann, in his work on Neritina, was able to trace the
descendants of two cells in the early cleavage into the velum.
They could be traced the more readily because of a peculiar
texture which distinguished them from the surrounding cells.
These two cells in Neritina correspond precisely to the two
lateral secondary trochoblasts of Amphitrite and the other
annelids and Crepidula (the two lateral cells shaded with lines
in Fig. 9). The primary trochoblasts of Neritina are also dif-
ferent from the surrounding cells. They are very minute, as
in Scolecolepisy and were never observed by Blochmann to
divide. As far as observations go, therefore, Neritina falls in
line with the other forms which we have studied.

Conklin has called attention to the altogether characteristic
appearance of the primary trochoblast, ' not only in Neritina
and in four species of Crepidula which he studied, but in
Umbrella (Heymons), Urosalpynx, and Fulgar. In all it is
conspicuous for its small size, and does not divide for a
comparatively long time.

It will appear from Holmes's preliminary paper that the
origin of the velum in Plariorbis bears a remarkable resem-
blance to that of the annelids. Planorbis resembles Amphitrite
even more than it does Crepidula in this respect, viz.^ the primary
trochoblasts are large when formed, and the whole prototroch
becomes functional at an early period. With regard to the
cell-origin of the prototroch Holmes says: "The prototroch is
formed from the [primary] trochoblasts previously mentioned,
and the uppermost cells of the second quartette which form the
tips of the arms of the cross [secondary trochoblasts, q, r, s, t,
in the three quadrants]. Possibly other cells of the second
quartette may take part in this formation ; it is certain that the
third quartette has no share in the process."

I am greatly indebted to Mr. Harold Heath for a very com-
plete account of the origin of the prototroch in Ischnochitony
and for his courtesy in allowing me to use his discoveries
in this paper. The four primary trochoblasts are formed at
the i6-cell stage, exactly as in Amphitrite, etc., etc. (Fig. 7,

THE CELL ORIGIN OF THE PROTOTROCH.

135

Stippled cells). These four cells divide into sixteen, all of
which enter the prototroch ; so far, therefore, the resemblance
is complete. The secondary trochoblast also arises in the
32-cell stage, precisely as it does in AmpJiitrite^ Clymenella, and
Arenicola (Fig. 8, cell shaded with lines). All of the deriva-
tives of this cell enter the prototroch in Isc/tnochiton, and all
but one very minute cell product do so in the annelids named
(Figs. 10-12).

The essential features of Heath's account can be most effect-
ively presented by means of diagrams. Fig. 25 represents a
lateral view of the prototrochal cells of one side in three stages

Fig. 25. — Represents trochoblast as seen on left side in Ischnochiton (Heath). Three stages shaded
like previous drawings. Unshaded cells are not part of the prototroch in the other forms
examined. See p. 132.

of development, and the various cells are shaded like those in
Fig. 24 {AmpJiitrite). Diagram A, Fig. 25, shows the condition
of the trochoblasts at about the 32-cell stage. It shows two
of the four groups of primary trochoblasts (stippled), each
group consisting at this period of two cells, as in Ainphitrite,
etc.. Figs. 8 and 24, A. Between these cells lies the trocho-
blast marked q, r, s, t. The clear cell marked ;r is a trochoblast,
and its two derivatives in each quadrant enter the prototroch,

136 BIOLOGICAL LECTURES.

as may be seen from Diagram C. To show the position of this
cell in Ampkitrite, I have marked it x in Fig. 10 in one quad-
rant, and in each of the four quadrants in Fig. 11. If my
observations on Amphitrite and Clymenella and those of Child
on Arenicola are correct, this cell does not enter the prototroch
in these annelids. However, I have never followed its career
in either form, except in the posterior quadrant, and here
Heath has called my attention to the fact that one of the
derivatives of this cell in Amphitrite is excessively minute (not
as large as a polar globule), and has the appearance, therefore,
of a ''vestigial cell." As the matter stands at present, this is,
I believe, the first important well-sustained discrepancy in the
origin of the prototroch, in all the Trochozoa Vhich have been
studied, and for this reason I would recommend the investiga-
tion of the destiny of this cell in the annelids as a fruitful
problem in cell-lineage.

Diagrams B and C in Fig. 25 show that all four derivatives
of the secondary trochoblast ^, r, s, and t enter the prototroch.
This is true in three quadrants. It will be recalled that in the
three annelids in which these cells are known, t is not a proto-
trochal cell, but is minute and of a vestigial appearance. It is
shown in Figs. 11, 12, and 24.

The cell y in Fig. 25, Diagram B, remains to be discussed.
It divides once, and the two daughter-cells take their position
in the prototroch, as seen in C. This cell y^ like the cell x,
does not enter the prototroch in Amphitrite (see Fig. 11),
Clymenella^ and Arenicola. It is small in Amphitrite, but not
minute enough to suggest that it is a vestigial cell, though one
of its products is distinctly of this appearance. In respect to
the cells y, y', t, and s, however, there is an important fact to
mention. They are, as far as could be observed, not ciliated,
and probably serve. Heath suggests, as a sort of supporting
layer. For this reason it may, perhaps, be a question
whether they should be considered part of the prototroch.
If we throw them out, however, there arises another dis-
crepancy between this prototroch and that of the annelid,
for in the latter the cell s is undoubtedly a part of the
prototroch and is ciliated.

THE CELL ORIGIN OF THE PROTOTROCH. 137

For the sake of convenience in comparing the cell-origin of
the prototroch in the various annelids and gasteropods in which
it has been investigated, we assumed the annelid Amphitrite as
the type. In this form the prototroch arises from two sources :
{a) Four cells, primary trochoblasts, one belonging to each
quadrant of the umbrella hemisphere, give rise to a functional
"primary prototroch," which is for a time the only ciliated
portion of the larva. The four cells in question are formed
at the i6-cell stage, {b) Three cells, secondary trochoblasts,
belonging to the right, left, and ventral quadrants of the subum-
brella hemisphere, give rise, later, to the rest of the prototroch.
These three cells are formed at the 32-cell stage. The whole
prototroch is at first interrupted in the mid-dorsal region, but
ultimately grows together.

Almost all of the positive evidence furnished by the compari-
sons which we have drawn points, I beUeve, to the conclusion,
first, that the component cells of the prototroch are homologous
in the various types studied, in the same sense as are the com-
ponent bones in the skeleton of the vertebrate limb. If this
conclusion is correct, the homology of the whole prototroch is
placed on a firmer basis, and the evidence of the relationship
of the two diverse groups of the Trochozoa — annelids and
mollusks — is likewise strengthened. As a corollary of this
conclusion, it would appear that, since certain cells form the
same organ under environmental conditions which are unlike, the
environmental factors do not have the importance in directly
shaping development which is sometimes claimed for them.

Some of the evidence, however, seems to controvert this
conclusion. I refer particularly to the discrepancies between
Wilson's account of Nereis limbata and all other accounts in
regard to the destiny of four cells, and to that between Heath's
account of IscJmocJiito7i and all the other forms in regard to the
destiny of four or more cells.

Considering the practical difficulties, we may say with regard
to the value of different observations that positive evidence that
a certain cell enters the prototroch should have more weight
than the negative evidence that it does not, unless a different
destiny for the cell is established in the latter instance.

138 BIOLOGICAL LECTURES.

A re-investigation of the prototroch focussed upon these dis-
crepancies would be of great value in determining the trend of
the whole mass of evidence, whether it is for or against the
doctrines of the homology of cleavage cells, of the significance
of cleavage as a factor in differentiation, and of the preorgani-
zation of the ^g'g itself.

SEVENTH LECTURE.

RELATION OF THE AXIS OF THE EMBRYO TO
THE FIRST CLEAVAGE PLANE.

CORNELIA M. CLAPP,

Mt. Holyoke College, South Hadley, Mass,

What is the meaning of cleavage ? is the constantly recur-
ring question of the embryologist. Is cleavage a differentiating
process in development, or is it a process which may run on
independently of differentiation, sometimes coinciding with it,
sometimes not ?

Is the egg a mosaic, as Roux maintains? Does the first
plane of cleavage determine the axial position of the embryo,
or is the Qgg already definitely oriented with respect to the
future embryo before cleavage begins ? Is the egg an isotropic
body to be converted gradually into a definite mosaic as cleav-
age advances, or are the blastomeres all endowed with the same
potentialities, the fate of each being settled by the position it
happens to hold ?

Does the division of the egg by the first line of cleavage
mean the separation of the parts destined to become the right
and left sides of the bilaterally symmetrical animal, or may this
first cleavage run in any direction, and have no fixed and
necessary relation to the future embryo ?

The well-known discoveries of Newport ('54) and of Pfliiger
and Roux in '87 and '88 seem to point toward some general
law controlling the phenomena of cleavage.

In 1 89 1, while studying the development of BatracJms tan
at the Marine Biological Laboratory at Wood's Holl, Professor
Whitman directed my attention to the fact that the egg

139

I40 BIOLOGICAL LECTURES.

was a favorable one for testing the conclusions of Roux. An
^gg provided with an adhesive disk, by means of which it is
held constantly in a fixed position, seemed to offer just the con-
dition required for testing the mosaic theory of cleavage.

The results have been reported in \.\i^ Joiumal of Morphology ^
vol. V, p. 498. In only tJiree cases out of twenty-three was
there coincidence of the first line of cleavage with .the median
plane of the embryo.

In 1892 Born^ called attention to my paper and pointed out
what he regarded as the " sources of error " in the case. His
criticism was accepted by Roux,^ who adds : *' Only slight errors
of experiment are required in order to obtain almost equally
incorrect figures in the frog's ^gg''

Born doubts the validity of the results, because of the long
period of time that elapses between the appearance of the first
furrow and the outline of the embryo. During this period of
six days, slight changes may occur to account for the ** abwei-
chende Resultat." Born also finds difficulty in accepting the
assertion that " the adhesion of the yolk to the egg-membrane
prevented rotation," and says that "such adhesion would cer-
tainly hold good only for the first stages."

Weysse,^ in 1894, believing that if there was adhesion of the
yolk to the membrane it could be demonstrated in sections of
the ^gg^ made a careful examination, but found no trace of any
such attachment. He concludes, with Born, that " during the
six days mentioned abundant opportunity is furnished for a
rotation of the yolk within the egg-membrane."

These experiments have been twice repeated since '91, with
essentially the same result. In order that the possible sources
of error may be fairly estimated, it seems desirable, in giving
the results of my later experiments, to repeat and perhaps
expand the brief description of the toadfish ^gg given in '91.

The ^gg is large, — 5 mm. in diameter, — and as the blasto-
disk always develops at the free pole of the ^gg, the early
cleavage lines can be observed with a hand-lens without dis-

1 Merkel and Bonnet, Erg., Bd. I, p. 502.

2"Beitrag zur Entwickelungsmechanik des Embryo," Anat., Heft 7, p. 313.

^ Proc. of Amer. Acad., vol. xxx, p. 308.

THE AXIS OF THE EMBRYO. 1 41

turbing the ^g^. The ^gg is hardy, and develops well in
shallow dishes holding sea-water about two inches deep.

The membrane of the ^gg has a peculiar adhesive disk, about
3 mm. in diameter, which has a constant position, with the
centre of the disk at the vegetative pole, directly opposite the
micropyle. By means of this disk the ^gg is firmly glued to
the supporting surface, usually the underside of a rock, the
inside of a broken jug, tin can, a piece of stovepipe, or even
an old boot-leg. The disk consists of a transparent secretion,
which becomes opaque and gluey on contact with water. It is
of nearly uniform thickness, and is closely applied to the egg-
membrane everywhere except for a narrow margin which pro-

FiG. I. Fig. II.

jects all around as a thin rim (Fig. I). The disk is saucer-
shaped, and only a little thicker than the egg-membrane itself.
I have been able to separate it from the membrane in the case
of eggs hardened before attachment. As the ^gg is generally
fastened to miore or less plane surfaces, it appears strongly
flattened on the side of attachment, as described by Dr. Ryder
and as shown in Fig. II.

That the egg-membrane, with its adhesive disk, does not
change position after attachment is, then, very certain ; but is
the egg itself also fixed within the membrane, and does it main-
tain a fixed position for the time required to reach an early
stage in the axial differentiation — six or seven days after fer-
tilization } or is it liable to rotate and get displaced, and so
invalidate the results obtained }

In the experiment of '91 the eggs were allowed to flow from

142 BIOLOGICAL LECTURES.

the opened ovary of the fish into a glass dish containing just
water enough to cover them. After becoming fixed to the
bottom they were fertilized. The side of the ^gg resting on
the glass was much flattened, and the fact that the blastodisk
retained its position at the free pole of the ^gg when that pole
was directed Upward, led to the statement that '' the adhesion
of the yolk to the egg-membrane, as it rested on the disk area,
prevented rotation."

In the experiments of '96 and '98 the eggs were allowed to
fasten themselves on pieces of glass which were inverted and
which rested on supports within a large dish of sea-water. Con-
sequently the cleavage and development had to be observed from
below, by looking up through the bottom of the glass dish.

In this case the eggs were in their natural position, that is,
in the position in which they are usually deposited by the fish
when making her nest under a flat stone.

It makes no difference, then, how the ^gg is attached,
whether on the roof of the nest, or on the side or on the floor ;
the axis of the ^gg is ^iwdiys perpendiailar to the plane of attach-
ment. The fish usually fixes its eggs so that they hang from
the surface above ; but if they are deposited in a piece of stove-
pipe, so as to cover the upper half of the concave inner surface,
the position of the axis of the Qgg will vary at all angles
between a vertical and horizontal position.

If the eggs are dropped loosely into a glass dish containing
very little water, some of them will fall so as to become quickly
fixed to the bottom of the dish, while others may fall so that
the adhesive surface fails to come into contact with the glass,
and so soon loses its adhesive property, leaving the &gg free to
roll. The axis of the fixed eggs will be vertical, but the free
pole will face upward instead of downward, as it does when
the Qgg is placed on the roof of the nest, or on the underside
of a glass plate. It is evident, therefore, that gravity has no
decisive influence in determining the direction of the egg axis.
In other words, the egg has a definite and constant position in
its membrane, the vegetative pole always lying at the centre of
the adhesive disk, while the animal pole and micropyle coincide
on the opposite side.

THE AXIS OF THE EMBRYO. 143

This is not saying that the ^gg cannot be forced to change its
position within the membrane. The eggs that fail to become
attached to the bottom of the dish are liable to be rolled about,
and the movement may be sufficient to cause some shifting in
position. The important fact here is, that, whatever be the
place of attachment, whether above, below, or on the side, the
axis of the ^gg always maintains the constant relation of being
perpendicular to that point, and there is no inherent tendency
to assume a vertical position, as is the case in most pelagic fish
eggs.

There is, then, no cause for rotation within the membrane
unless it be supplied artificially from without. The ^gg remains
in the position in which it becomes fixed, neither rotating nor
becoming displaced except as the result of rough treatment.

In the experiment of '98 great care was taken to guard
against any possible disturbance of the eggs. The pieces of
glass with the adherent eggs were arranged conveniently for
observation before they were fertilized, and the water changed
from time to time by means of a siphon, so that no disturbance
seemed possible.

If the ^gg does not rotate during the early stages of develop-
ment, it may be asked. What is the nature of the adhesion that
prevents rotation }

In regard to this matter Professor Whitman assures me that
adhesion of the yolk to the membrane is a very general phe-
nomenon which he has noticed in the ^gg of Necturus^ Am-
blystoniay frog, newt, pelagic fish eggs, and even in the small
eggs of many annelids. Speaking of pelagic fish eggs, he
says : " It is only necessary for an egg to be left at rest for a
few minutes in order to see how readily it adheres to its mem-
brane. I have often taken advantage of this adhesion to roll
over and hold them some moments by the aid of needles, with
the lower pole uppermost. Others, I am sure, have done the
same thing while studying the cleavage." He also adds that
in the case of the toadfish "the strength of the adhesion is
probably due to the relatively large weight of the ^gg and the
large surface of contact." The perivitelline space in this ^gg
is so slight as to be difficult of recognition, and the extent of

144 BIOLOGICAL LECTURES.

the surface in actual contact with the membrane is 'correspond-
ingly very great as compared with small eggs with ample space
between the yolk and the membrane. The conditions, then,
are most favorable for strong adhesion.

Born asserts that '' such an adhesion cannot hold for more
than the first stages." As this statement does not rest upon ob-
servation of the toadfish ^g^^ it can have no value beyond that
of conjecture suggested by experience with other eggs. It is
quite true that six days gives ample tune for any number of
rotations, but it is not a question whether there is time for rota-
tion, but zvhe titer rotation actually occurs ^ and that can only be
determined by close study of the developing ^gg, not by sec-
tioning of hardened and imbedded eggs, nor by any amount of
experiments on frogs' eggs.

In one very important respect the toadfish ^gg is far superior
to the frog's ^gg for the study of the question here considered,
since it can be observed under perfectly normal conditions,
without resort to those artificial means of fixation or marking
which are necessary in a frog's ^gg, and which must always
cast some doubt on the reliability of the results.

In the experiment of '91 the method of determining the
relation of the first cleavage plane to the axis of the embryo
was as follows : After the eggs had become fixed by the adhe-
sive disk to the bottom of a glass dish they were artificially
fertilized. The blastodisk appeared on the free pole of the
Q,gg^ where it was easily watched by means of a lens. The
eggs were plotted on paper, each ^gg being represented by a
circle (Fig. Ill), and the paper and the dish containing the eggs
oriented by fastening a label on each in the same relative posi-
tion. When the first line of cleavage appeared the direction
was indicated in the circle representing the ^gg, by the diame-
ter, and when the axis of the embryo became visible, that was
indicated by an arrow drawn across the same circle.

The first furrow appeared seven hours after fertilization, and
on the seventh day the axis of the embryo could be distinctly
seen as a light streak in the blastoderm.

The result of this experiment is seen by an examination of
the circles given in Fig. III. Of the twenty-three developing

THE AXIS OF THE EMBRYO.

145

embryos, three show coincidence of the axis of the embryo with
the first cleavage plane. There is no case of exact coincidence
with the second cleavage plane. Fourteen of the embryos have
the head directed towards the right of the first line of cleavage,
the axis of the body being at an angle with the first cleavage
plane of from 30° to 70°. In the remaining six the head was to
the left of the first cleavage plane, the angle varying as before.
In the experiments of '96 the main purpose in view was to
settle the question of rotation of the Qgg within its membrane

n

/f

/5

16

jr

Fig. III.

and to find out, if possible, whether gravity had anything to
do with the position of the blastodisk. The eggs attached to
pieces of glass were fertilized and the glass plates placed in
various positions, so that the axis of the Qgg in different cases
was vertical, horizontal, and inclined at various angles.

It was observed that the small oil globules within the Qgg
were always found uppermost in whatever position the Qgg was
placed, in one case being directly under the blastodisk at the
animal pole, and again directly underneath the adhesive disk at

146

BIOLOGICAL LECTURES.

the vegetative pole, or halfway between these poles on the side
that was uppermost.

As for the blastodisk itself, that retained its position at the
free pole of the egg, near the micropyle, in whatever position the
eggs were placed.

The position of the eggs was plotted in the case of ten eggs
attached in the inverted position.

The coincidence of the first furrow with the embryonic axis
was observed in three out of ten cases (Fig. IV).

The third experiment was made during the summer of '98.
A broken bottle from the eel pond was brought into the labora-
tory containing a toadfish and a few eggs which had been
already deposited by the fish.

On the morning of July 6 the fish was opened and the
mature eggs allowed to flow out of the ovaries and fasten them-
selves on pieces of glass.

About one-third of the eggs set free from the ovaries became
fixed to three glass plates. The rest failed to adhere and so
were of no use in the experiment. Two minutes was sufficient
time to allow for the eggs to become fixed. The plates were
then inverted and placed on supports in a large shallow dish
with sufficient sea-water to cover them.

The eggs were then fertilized, and soon after the water was
siphoned off by means of a rubber tube and fresh sea-water

Fig. IV.

quickly supplied. The dish was so placed that the eggs could
be observed from below.

At 4.30 P.M. — eight hours after fertilization — the first line
of cleavage was noted in fourteen eggs, and the direction of the

THE AXIS OF THE EMBRYO.

147

furrow indicated on the paper oriented as in the earlier experi-
ments.

On July 12, six days afterward, during which time the dish
containing the eggs was not moved, the axis of the embryo was
observed, and its direction indicated as before. The result is

Fig v.

seen in Fig. V. Of the fourteen eggs only two show coinci-
dence of the first cleavage furrow with the embryonic axis
(Fig. V, 5, II).

Similar experiments with the eggs of Amia have been made
by Eycleshymer, '96.i Observations were made on three sets
of eggs and the results are given in Fig. VI.

Twenty eggs developed normal embryos, and two showed
coincidence of the first plane of cleavage with the axis of the
embryo (Fig. VI, 2 and 6). The following is quoted: *'The
^^'g is oval in profile view, measuring in its longest diameter,
including the membrane, 2.5 to 3 mm. ; in its shortest, 2 to
2.5 mm. The freshly deposited ^gg is firmly fixed to the
object with which it first comes in contact by means of the
threads of the villous layer, which are elongated over the lower
hemisphere of the egg-membrane.

The eggs still attached to blades of grass or rootlets are
placed in shallow watch-glasses and held in position by weight-
ing with small pebbles. The watch-glasses are then placed

'^ Journ. of Morpk., vol. xii, p. 344.

148

BIOLOGICAL LECTURES.
3

Fig. VI.

upon a mirror fastened to the stage of a dissecting microscope.
We could thus observe the changes occurring on opposite sides
of the Q^^ without disturbing its position.

The elongated form of the ^gg of Amia, in a closely applied
envelope, prevents rotation about its minor axes. It is, there-
fore, a favorable Qgg for ascertaining what effects, if any,
gravity may have on the direction of cleavage, and for deter-

THE AXIS OF THE EMBRYO. 149

mining the relation of the early cleavage planes to the median
plane of the embryo. Dean found " in the early stages of seg-
mentation that the cleavage planes occurred in the normal way
when the position of the ^gg was reversed." This is true
not only for inverted eggs, but for eggs placed in any posi-
tion whatsoever. It seems to follow that gravity can have
no directing influence on the cleavage. In order to ascertain
whether there is any constant relation of the embryonic axis to
either of the first two cleavage planes, eggs were fixed in given
position by weighting, as before mentioned, and a sft:etch of the
early grooves was carefully made in each case. These grooves
are easily identified for a long time in the lower hemisphere of
the ^gg, even as late, in some cases, as the early stages of
gastrulation. As the sketches made at successive intervals
showed no movement of the ^gg during all this time, it seemed
probable that the position of the ^gg remained practically
unchanged up to the time when the median plane of the
embryo was ascertainable. In some cases accidental markings
on the surface of the ^gg remained in a fixed position until the
embryo was well defined.

In 1893 Morgan^ experimented with the eggs of teleosts.
The pelagic eggs of Ctenolabriis and Serraniis were selected.
Morgan says : *' My results show that there is no relation
whatsoever between the cleavage planes of the ^gg and the
median plane of the adult body. I have definite records of
twenty-two eggs carefully marked. The results, expressed as
nearly as possible, show for Ctenolabriis the plane of first
cleavage agrees approximately with the median plane of the
body vcvfive cases, the plane of second cleavage with the median
plane in ten cases. The median plane of the adult body lay
between the first and second cleavage planes in two cases.
For Serramcs the first cleavage agreed in one case, the second
agreed in two cases, and neither in two cases."

Two methods were employed. In one case the ^gg was
watched continuously, and the median plane of the embryo
corresponded to a plane midway between the first and second
cleavage planes.

1 An. Afiz., viii. Jahrg., p. 803.

150 BIOLOGICAL LECTURES.

The second method consisted in marking the membrane
above the blastoderm and using such marks as points of orien-
tation. *' The success of the second method depended upon the
close adherence between the ^gg and its membrane. After a
very careful test it was found that even after rough treatment
the Qgg retained a fixed position relatively to the surface mark-
ings. And again by watching individual eggs for some hours
it was seen that the ^g<g did not rotate within the membrane in
the early stages or change its orientation with respect to the
marks on the egg-membrane. In order to mark the eggs they
were removed from the water and partially dried. A needle
covered with finely divided carmine was drawn horizontally over
the eggs. Small particles of carmine stuck to the membrane
in many cases. The eggs were returned to the water and the
best marked chosen."

Jordan and Eycleshymer^ published a joint paper in 1893,
from which it becomes plain that in the case of two species
of Urodeles and two Anura there exists no constant relation
between the early cleavage planes and the adult axes. " The
first and second cleavage planes undergo extensive torsion, and
the cells originally on one side of the mid-line come to lie on
the opposite side."

Jordan 2 states, in regard to the cleavage in the newt, that
" the total absence of any regularity in the arrangement of the
cells is the most conspicuous feature."

Experiments of a somewhat different nature by Driesch,
Wilson, and others have also shown conclusively that the first
line of cleavage cannot have the significance which the earlier
experiments of Roux would indicate. Recent studies in cell-
lineage, together with the work of experimental embryologists,
have thrown a great deal of light on the subject here considered,
but without furnishing a decisive answer as to the meaning of
cleavage.

The opinion expressed by van Beneden,^ as early as 1883,
that all bilateral animals would be found to agree in having the

1 Journ. of Morpk., vol. ix, p. 407.

2 Journ. of Morph., vol. viii, p. 269.
2 Archives de Biologic, IV, 1883.

THE AXIS OF THE EMBRYO. 151

axis of the body located and pre-marked by the first plane of
cleavage, has steadily lost ground.

The study of cleavage has not yet given us any such funda-
mental law of development as the mosaic theory claims. The
number of animals whose development does not conform to the
supposed law of coincidence has increased so rapidly of late,
that the only reasonable conclusion seems to be that while the
first cleavage plane may coincide with the median axis of the
embryo, as Roux and others have shown, it is not a constant
rule in any single case, much less a universal law.

The opinion is gaining ground that the phenomena of cleav-
age are to be regarded as the expression of non-differential
rather than qualitative divisions of the germinal material. The
pressure experiments have proved that there can be extreme
variation in cleavage produced without in any way interfering
with the normal development of the embryo.

Note. — Batrachus tau = Opsanus tau. See Bulletin of U. S. National
Museum. No. 47, Part 3, p. 2313. 1898.

EIGHTH LECTURE.

OBSERVATIONS ON VARIOUS NUCLEOLAR
STRUCTURES OF THE CELL.

THOS. H. MONTGOMERY, Jr.

{^Lecture given August 3, i8g8.)

Within cells occur structures of various kinds which have
been loosely termed '' Nucleoli." In the strict sense of the
term, however, the word '* nucleolus " should be applied solely
to those elements of the nucleus which, with suitable double
stains, such as haematoxylin-eosin, or the Ehrlich-Biondi-Hei-
denhain mixture, stain differently from the chromatin, at least
when the latter substance is in the chromosome stage. Thus
the term ** nucleolus," introduced by Valentin in 1839, corre-
sponds to the " plasmosoma" of Ogata, 1883.

Besides the true nucleolus, as just defined, numerous other
so-called " nucleolar " structures occur in various cells. Among
them are the katyosomes, which are merely thickened nodal-
points of the chromatin reticulum ; further, what I shall term
the *< chromatin-nucleus," which is found in certain spermato-
cytes ; and then various structures, which stain neither like the
true nucleolus nor the chromatin, and to which such terms
as " Paranucleoli," *' Nebennucleoli," and " Pseudonucleoli "
have been applied. It is one of the most difficult questions to
determine the nature and correspondence of the latter struc-
tures, and very extensive comparative studies are necessary for
their elucidation. But from the cases studied by me, it would
appear that some of these structures in Metazoa probably must
be placed within the category of true nucleoli, and be regarded
as true nucleoli of a different chemical nature. That is to say,

153

154 BIOLOGICAL LECTURES.

all segregations of formed substance within the nucleus, not
including the chromatin, linin, caryolymph (and the problemat-
ical lanthanin granules?), probably should be regarded as true
nucleoli. Or, in other words, our criterion of nucleoli probably
should not be based as much upon chemical as upon morpho-
logical facts : the criterion should be rather its mode of forma-
tion and deposition than its chemical nature, for there is reason
to believe, as I think the facts here presented will show, that
the nucleolus is a substance formed under the action of nutri-
tive metabolism, and hence that its chemical nature may vary
in different cells, and at different stages of. the same cell, just
as the kind of nutrition varies in different cells and at different
stages. In estimating the valence of a nucleolar structure, its
morphology must be considered as well as its chemistry, and
not its chemical nature alone.

The present lecture gives a short review of a few of the
observations made by me upon different cells, while the detailed
observations, accompanied by figures and literature reviews, are
contained in two papers, one (not yet published) in th.Q Journal
of Morphology J and the other in Spengel's Zoologische Jahrbiicher
for 1898 ; the former paper, dealing mainly with the true nucleo-
lus, was completed in February, 1897; and the second, on the
spermatogenesis of Euchistus {Pentatoma), in April, 1898. The
objects studied in the former paper were the germinal vesicles
of nudibranch Mollusca {Montagiia^ Doto), of Nemerteans {Tet-
rastemma, Amphiporzis, Zygonemertes, Lineus), of the polychaete
annelid Polydora^ of the leech Piscicola, and of the siphonophore
Rodalia; further, nuclei of two Gregarines, and of various
ganglion, gland, muscle, and blood cells of the forms mentioned
above, and the hypodermis of the larva of Carpocapsa.

I. The True Nucleolus.

With Delafield's or Ehrlich's haematoxylin, followed by eosin,
the nucleolus stains red ; while with the Ehrlich-Biondi-Heiden-
hain mixture as usually employed it is stained red or orange.

I. Number and Size. — The usual number is about 1-5 to a
nucleus, but in many cells a much larger number is present.

NUCLEOLAR STRUCTURES OF THE CELL. 1 55

Thus, in germinal vesicles of Amphibia, SelacJiiiy Sagitta, Meta-
nemertini, most Teleostei, and mdiny Arthropo da, frequently hun-
dreds are present, and an equally large number is found in
clitellar gland cells of Piscicola and in various glands of Arthro-
poda. The number in a germinal vesicle is not dependent upon
the amount of yolk present in the cytoplasm, since it is the rule
in birds and many arthropods that only one or a few occur.
There are very few cells in which no nucleoli are present, but
that may be the case occasionally in spermatozoa, and in certain
connective-tissue elements ; so it would appear that the absence
of nucleoli is to be noted mainly in such cells where the nutritive
metabolism has ceased or is reduced to a minimum. The num-
ber is not constant for all the cells of a given organism, as was
claimed by Auerbach. The volume of a nucleolus stands, as
a general rule, in inverse ratio to the number ; i.e., at a given
stage of a given species of cell the amount of nucleolar sub-
stance remains more or less constant. The greatest relative
amount of this substance is found in cells which are undergoing
rapid growth, and in which, consequently, the nutritive metab-
olism is intense ; it is particularly in germinal vesicles, during
the growth period, that a large amount of nucleolar substance
is found.

2. General Striictttre and Position. — The ground substance
is usually homogeneous, but not infrequently granular (e.g.,
Gregarines, ova of insects). In most cases it has no limit-
ing membrane, the supposed membranes described by various
writers being probably in most objects only optical illusions due
to light refraction by the curved surface of the nucleolus, and in
other cases to the presence of a chromatin reticulum around
the nucleolus. Since the nucleolus lies in the meshes of the
chromatin reticulum, and is rarely, if ever, penetrated by chro-
matin strands, it follows that when a nucleolus is increasing in
size it pushes aside the chromatin so that the latter may form
a latticework around it. But in a few cases a true membrane,
a differentiated portion of the ground substance, is present, as
in the germinal vesicle of Polydora.

The position of the nucleolus is usually excentric, rarely cen-
tral. But the position often varies at different stages of the

156 BIOLOGICAL LECTURES.

nuclear history, as is shown very clearly in the germinal vesicles
of Nemerteans, where the nucleoli are formed at the periphery,
wander towards the centre, and finally pass to the periphery again.
Similar '' migrations " have been described also for Amphibia.

Within nucleoli vacuoles are usually present, which are
probably, in most cases at least, fluid globules ; they are nor-
mal structures, since they may frequently be seen in the liv-
ing cell. The origin of these vacuoles or globules was traced
by me in the germinal vesicles of nudibranch mollusks. In
these objects the nucleolus is at first homogeneous, but with
its increase in size, which keeps pace with that of the nu-
cleus, lighter staining globules make their appearance within
the nucleolus. Just at this time, fluid drops of similar chem-
ical nature are seen within the caryolymph, and these drops, on
the other hand, resemble chemically and structurally the early
yolk globules, which are simultaneously appearing in the cyto-
plasm. Consequently, we may conclude that globules which
are probably closely related to the earliest yolk globules are
taken up into the nucleus, probably as nourishment, and that
some of them are deposited within the nucleolus in the form of
intranucleolar globules or vacuoles. Thus, at least in these
objects, there are facts to show that the nucleolar vacuoles are
extranucleolar in origin. In these ovarial ova the vacuoles
gradually fuse together until a single large, excentric vacuole
is formed, bounded by a thin layer of nucleolar ground sub-
stance. Similar fusions of vacuoles to produce one large one
are found also in nucleoli of many other germinal vesicles, as,
e.g., in Polydora and Piscicola. In the objects studied by me
the nucleolar vacuoles would appear to be in all cases fluid
globules and not empty spaces, which is proved by their be-
coming stained with suitable reagents. In some cases these
vacuoles are not spherical (the usual case), but irregular in out-
line ; and the latter form may be easily explained by the as-
sumption that their consistency is occasionally viscid. The
same assumption serves to explain likewise the irregular forms
of many nucleoli. That is to say, the form of the nucleolus
and of its contained vacuoles will be spherical, only provided
that they have a thinly fluid consistency.

NUCLEOLAR STRUCTURES OF THE CELL. 157

In some cases, as in the nucleolus of the ^gg of Polydora,
the vacuolar substance is distributed in the nucleolar ground
substance in the form of irregular canals. I believe that the
study of similar nucleoli misled Carnoy in his idea of a " nucleole-
noyau," i.e., of the assumption of a particular form of nucleolus
which contains all the chromatin of the nucleus in the form of
a looped skein, and so represents a nucleus within the nucleus.
That is to say, in such a nucleolus as the one jusc mentioned,
he mistook the network of true nucleolar ground substance
for chromatin, and the true vacuolar substance for the nucleo-
lar ground substance. Certainly in no true nucleolus of any
Metazoan have I seen any trace of chromatin, much less a loop
of chromatin ; but it is possible, though I have been unable to
decide the point positively, that in the Gregarines the chroma-
tin may be contained in the nucleoli.

A few observers have remarked contractile vacuoles in the
nucleolus.

Very frequently there may be observed, within a nucleolar
vacuole, one or several small granules, the nucleolini ; but as
far as my observations extend, such nucleolini appear to be
inconstant in number and occurrence, and to be merely detached
portions of the ground substance of the nucleolus. Accord-
ingly there is no good evidence for believing that the nucleolinus
represents the dynamic centre of the cell.

3. Divisions y Fusions ^ and Amoeboid Movements in the Rest-
i7ig Cell. — Divisions and fusions are normal phenomena of
many nucleoli. In fact, the general mode of increase in volume
of a nucleolus, apart from accessions to it of vacuolar substance,
appears to be by the addition to the surface of secondary nucle-
olar masses. This may be observed very clearly in the germinal
vesicle of Linens. In the case of the other Nemerteans ex-
amined by me, the nucleoli formed at the periphery of the
nucleus remain there for a while and increase in size, then
pass towards the centre, where they may either fuse first to
form a single large one (Stichostemmd), or else immediately
subdivide and so produce numerous smaller ones, all of which
finally pass again to the nuclear membrane. In fusions of
nucleoli which have been formed at successive stages, and

158 BIOLOGICAL LECTURES.

which accordingly may differ chemically, nucleoli are produced
which are not chemically homogeneous throughout ; this seems
to be the case in the ova of numerous arthropods (e.g.^ Grylhis,
Etickistus, PhalangopsiSy and other insects).

Amoeboid movements have been observed in life by numerous
authors. But from the probable nature of the nucleolus, which
will be discussed below, the conclusion seems probable that all
movements, fusions, and divisions are passive in regard to the
nucleolus ; that they are not spontaneous movements of the
nucleolus, but due rather to chemical dissolutions and changes.
This is in agreement with Rhumbler's conclusions.

Besides the movements due to change of form, in which
process divisions and fusions are frequently concerned, there
remain to be considered those of change of location. The
migration of nucleoli in the germinal vesicles of Nemerteans
has been already considered ; and such migrations are possibly
of general occurrence in all cases where the number of nucleoli
is large. Besides such phenomena, migrations of nucleoli into
the cytoplasm in the resting cell have been frequently described.
Among others may be mentioned the conclusions of Roule
in Ascidian ova, to the effect that such nucleoli eventually
become nuclei of new cells, and similar observations on endo-
genetic cell formation in molluscan ganglion cells, as described
in a recent paper by Rohde. But it is hardly necessary to
remark that all recent cytological studies render it very prob-
able that no such endogenetic cell formation occurs in the
Metazoa. Among the numerous objects studied by me, only
one was found showing in the resting cell a migration of
nucleoli into the cytoplasm, but that one, now to be described,
proved to be of a unique character.

The clitellar gland cells of Piscicola are cells of enormous
size, to be readily seen with the naked eye, which occur in the
body cavity throughout the whole length of the worm, except in
the cephalic region, and the ducts of which open in the region of
the genital pores. During the process of formation of their
secretion two main stages maybe distinguished : a /r^///^j-^,
from the stage of the young cell until all the secretion has
been formed and the cell has attained its maximum dimensions ;

NUCLEOLAR STRUCTURES OF THE CELL. 159

and a metaphase, from the time of the discharge of the secre-
tion until the latter has been completely discharged. This
includes the cycle of activity of the cell, and it is probable that
this cycle is passed through once during each season of sexual
maturity of the animal. My deductions as to the sequence of
the stages were based upon an examination of a number of
individuals of various sizes.

At the commencement of the prophase the spherical nucleus
contains a single, more or less rounded nucleolus. The cell
body increases gradually in size, and secretion globules, which
stain much in the same way as the nucleoli, make their appear-
ance at some distance from the nucleus. The cell steadily in-
creases in size as do the nucleus and nucleolus, and then the
latter elongates and passes to the nuclear membrane. There
it continues to grow in volume, at the same time becoming so
irregular in form that in no two nuclei does it present the
same outlines ; ^-shaped, 5-shaped, and all kinds of irregularly
lobular forms are gradually produced. Then the nucleolus
commences to break up into pieces, and these fragment further,
until finally several hundred minute nucleoli are scattered
through the nucleus, in the place of the original large one.
During this time the cell body has become greatly dilated,
owing to the formation of secretion globules which are now
crowded throughout the cytoplasm. Simultaneously the nucleus
has increased in volume, and to a greater relative extent than
the cell body ; and it has gradually become excessively amoe-
boid in form, sending out long, branching, and anastomosing
processes, on which a nuclear membrane does not appear to
be present. These nuclear processes extend nearly to the
proximal end of the cell duct, and so penetrate between and
come into direct contact with the secretion globules.

The metaphase of the gland cell commences with the dis-
charge of the secretion corpuscles ; and here it may be noted
that the secretion appears to have been formed in the cyto-
plasm of the duct as well as of the cell body. During the
process of discharge the duct becomes swollen. After a con-
siderable portion of the secretion has been extruded, approxi-
mately a third of that previously present in the cell body, the

l6o BIOLOGICAL LECTURES.

nucleus suddenly commences to contract, and by the withdrawal
of its pseudopodia gradually to assume the spherical form again.
Now the point of chief interest to us is, that during the process
of nuclear contraction the nucleoli are discharged into the
cytoplasm. And the appearance of concentric rows of dis-
charged nucleoli within the cytoplasm leads to the conclusion
that the nuclear contraction is not gradual, but sudden and
intermittent. The case is as if one holds in his hand a sponge
filled with water, and by successive pressures of the hand forces
out the water in jets ; so by the contraction of the nucleus,
which is automatic, as shown by the retraction of its pseudo-
podia, the nucleoli are forced out through the nuclear mem-
brane into the cytoplasm, where the pressure is less. And the
nucleolar substance in the cytoplasm is found in the form of
rods, radial to the surface of the nucleus, which proves that it is
forced out rapidly in streams. But a remarkable point remains
to be noted : one nucleolus, and one only, appears in every case
to be retained within the nucleus ; for this phenomenon I can
give no explanation, since this one does not appear to differ
chemically from the others. Finally the nucleus becomes as
small as it was at the commencement of the prophase and
gradually rounds off; no longer are secretion globules present
in the cytoplasm ; and the extranuclear substance gradually
disappears, — howy it is impossible to determine, though since
it is not discharged through the cell duct, and there is no evi-
dence of any migration through the cell membrane, it would
seem probable that it becomes dissolved within the cytoplasm.
Such is the history of a cycle of functional activity.

In these remarkable stages certain salient points are to be
noted : the increase, first in size and then in number, of the
nucleoli, accompanying the formation of the secretion ; and the
discharge of nucleolar substance following upon the extrusion
of the secretion. These phenomena would point to some con-
nection between formation of nucleolar and secretory substances,
which is rendered the more probable, since by its extreme
amoeboid form the nucleus appears to be in close physiological
connection with the cytoplasmic processes. The fact that the
secretion globules appear first in the cytoplasm at some distance

NUCLEOLAR STRUCTURES OF THE CELL, l6l

from the nucleus would speak against any direct connection of
nucleolar and secretion substances. The nucleus evinces great
metabolic activity, and so would appear to play an important
part in the formation of the secretion ; but for the production
of suc^h activity an intense process of nutrition is necessary,
and it would seem to me probable that the nucleoli are in
some sense by-products of this nutritive metabolism. There
are few other observations on nucleolar changes accompanying
secretion stages ; but reference may be made to those of Nuss-
baum on gland cells of Argulus, to the effect that increase in
the number of nucleoli is produced by starvation.

4. Ontogenetic Origin. — How and where does a nucleolus
make its first appearance within the nucleus } Obviously it
would be incorrect to attempt to decide this point by a study
of the reformation of nucleoli within a daughter-nucleus after
mitosis, since the objection might well be raised that in such
a case nucleolar material discharged into the cytoplasm during
mitosis might be taken up again after mitosis, and accordingly
that the first origin of this substance is not explained.

In studying this point I was so fortunate as to find a fit-
ting object in the germinal vesicles of Nemerteans. In all the
Metanemerteans examined, but not in Linens, the germinal
vesicle appears to arise by the increase in size of a nucleus of a
connective tissue — a tissue of branching cells of an embryonal
character, which has been termed by me ** mesenchym." These
nuclei of the undifferentiated tissue show no trace of nucleoli.
By increase in size of a nucleus and its cell body, and by the
rounding off of the latter, an ovarial ovum (first ovocyte) is
formed. Now, when the latter has attained a certain size, nu-
cleoli make their first appearance, and at first are always closely
applied to the inner surface of the nuclear membrane. Simul-
taneously the yolk globules begin to segregate within the cyto-
plasm ; yolk and nucleoli show nearly the same chemical reaction.
But one essential point remains to be emphasized : the yolk sub-
stance, when it can first be differentiated by suitable stains, is
found to be at some distance from the nucleus. These facts
would seem to lead to the conclusion that in these cases the
nucleolar substance is extranuclear in origin, and that it may

1 62 BIOLOGICAL LECTURES.

Stand in a genetic connection with that early stage of the yolk
substance, when the latter cannot as yet be differentiated by
staining methods. In other words, it would seem that the yolk
is at first present in the cytoplasm in the form of a diffused,
unstainable fluid ; that a portion of it, that remaining in the
cell body, later becomes segregated as, or chemically changed
into, yolk globules ; and that another portion of it is taken into
the nucleus and, after passing the nuclear membrane, changed
into nucleolar substance.

The objection might be raised by those who assume the
nuclear origin of yolk, that the nucleoli are yolk globules formed
within the nucleus and destined to be expelled into the cell
body. But two answers may be given to such a criticism.
First, we have seen that the first discernible yolk globules
within the cytoplasm lie near the periphery of the cell, conse-
quently at some distance from the nucleus ; and, second, the
broader reason, that the ovarial ovum is essentially charac-
terized by its rapid growth, and that this growth is maintained
necessarily by accretion of nutriment from without ; the cyto-
plasm derives its nourishment from the body cavity or blood
vessels, and the nucleus from the cytoplasm. This being the
case, and the fact recalled that increase in nucleolar substance
appears to always accompany rapid growth of the nucleus, leads
to the conclusion that nucleolar substance stands in a pretty
close relation to nutritive metabolism ; while the fact that
nucleoli first appear at the periphery of the nucleus speaks for
their extranuclear origin.

There can be no doubt that the physiological processes of
nucleus and cytoplasm stand in a most intimate relation, and
possibly even that the nucleus directs and controls the cyto-
plasmic processes. Accordingly, there are good reasons for
concluding that the formation of yolk globules may be directed
by the nucleus. But the point for which I am contending is
that the yolk substance, which is undoubtedly nutritive and
hence in the first place extracellular in origin, is not produced
but only chemically changed and segregated by the action of
the nucleus. And since, in the germinal vesicles of Nemerteans
and nudibranch mollusks at least, the nucleolar substance shows

NUCLEOLAR STRUCTURES OF THE CELL. 163

a close chemical resemblance to yolk globules, we have every
reason for supposing that in these objects, if not generally, the
nucleolar substance is nutritive and extranuclear in origin.

5. Behavior in Nuclear Divisions. — In cases of amitosis it
is the general rule, as I conclude after a thorough review of the
literature on the subject, that division of the nucleolus accom-
panies or precedes that of the nucleus. I have seen prepara-
tions showing this very clearly in cells of the ovarial follicle of
Grylhis ; Dr. Conklin kindly demonstrated these preparations
to me, and allowed me to mention here this result of his as yet
unpublished observations.

In mitosis, on the other hand, it is the general rule that
the nucleolus disappears by dissolution during the prophases,
dissolving either in the caryolymph or in the cytoplasm after
the disappearance of the nuclear membrane. Such a dissolu-
tion may be either gradual, or the nucleolus may first fragment
into pieces, and then these dissolve. In the first maturation
division of the ovum of Piscicola the dissolved nucleolar sub-
stance is easily demonstrable by the aid of suitable stains in
the caryolymph, and after the disappearance of the nuclear
membrane seems to commingle with the cytoplasm.

In some few cases, relatively speaking, the nucleolus may
not at all or only partially become dissolved in mitosis, and so
be .cast out into the cytoplasm, where it may be discerned
through the metaphase. This has been observed, for instance,
in the ova of Aequorea (Hacker), Chaetopterus (Mead), Allolo-
bophoi-a (Foot, whose preparations I have had the pleasure of
seeing), and Myzostoma (Wheeler), though in plant cells this
phenomenon would appear to be more usual than in animal
cells.

In still fewer cases the nucleolus divides in the equatorial
plate along with the chromosomes. Such cases, which could
not, however, be proved with certainty, were observed by me
in the ovogonia of Linens and Polydora, and more or less
similar cases have been noted by a few other writers. But
such a phenomenon is very unusual, and it is very doubtful in
such cases whether the nucleolus always divides regularly. It
might be regarded as a case of a nucleolus extruded whole into

164 BIOLOGICAL LECTURES.

the cytoplasm, which, becoming caught by mantle fibres of the
spindle, would be passively divided. Or, to state the case differ-
ently, it would appear that in mitosis equal division of the
nucleolar substance does not appear to be effected.

Judging from my observations on the reduction divisions in
Anasa and Euchistiis (Hemipterous Insects), and in Piscicola^
as well as on the ovogonic divisions of Linens^ there appears to
be no good evidence that the nucleolar substance takes part in
the production of the spindle fibres. Yet such a view is held
by Strasburger and others, based upon the generally simul-
taneous disappearance of nucleolus and appearance of spindle
fibres. Thus in Etichistus, the object in which this question
has been most thoroughly studied by me, in the spermatogonia
the central spindle connecting the centrosomes is formed at
some distance from the nucleus, and the pole rays, also of cyto-
plasmic origin, may be seen before the nuclear membrane has
disappeared ; while the mantle fibres are undoubtedly of nuclear
origin, as are the interchromosomal threads, both being pro-
duced by a gradual change of position of the linin threads
before the disappearance of the nuclear membrane. However,
it may be going too far to conclude that in no objects are the
spindle fibres formed of nucleolar substance. But in view of
the evidence showing that nucleoli are masses of passive mate-
rial, and that the spindle fibres are (temporary.?) states of pre-
existing living matter, one would be justified in concluding
that in all probability there is no genetic connection between
spindle fibres and nucleoli.

Another point may be briefly referred to. The central
spindle corpuscles (" Zwischenkorperchen " of Flemming, by
whom they were first described) may by certain methods be
caused to stain like nucleoli. The general consensus of opinion
is that these corpuscles are thickenings of central spindle fibres
or of connecting fibres. But might not in some cases undis-
solved portions of nucleolar substance come to lie in the equator
of the spindle, and there be confused with true '* Zwischen-
korper".-* This question is of interest with regard to the older
view of Strasburger, that the nucleolar substance forms the
cell plate in plant cells, and consequently the cell membrane.

NUCLEOLAR STRUCTURES OF THE CELL. 165

Metzner (1894) considers that in the mitoses of the salamander
testicle the nucleoli play an important mechanical part, since he
considers some of them serve to attach the spindle fibres to the
chromosomes.

In regard to the reappearance of nucleoli in the daughter-
nuclei, after mitosis, very little has been ascertained. Some
authors consider that the nucleolar substance dissolved within
the nucleus may be carried over into the daughter-nuclei by
the chromosomes. Others, as, e.g., Zimmermann (with the con-
clusion *' omnis nucleolus e nucleolo "), consider that in the
anaphase all the nucleolar substance in the cytoplasm is taken
up again by the daughter-nuclei. But the majority of observers
leave the question undecided. My own standpoint inclines to
the assumption that the nucleolar substance discharged at the
stage of the equatorial plate into the cytoplasm is not again
taken into the nuclei. This view is difficult to prove. But
from a number of comparative observations it seems to me evi-
dent that the nucleolar substance is of no special morphological
value, and that it is in all probability not carried over in consecu-
tive cell generations. If this assumption be proved correct,
it leads to the further conclusion that one reason for the dis-
appearance of the nuclear membrane in mitosis may be the
necessity for a discharge of those nuclear waste products which
are in such a physical state that they cannot pass through the
nuclear membrane. In the final stages of their existence the
nucleoli may well be considered waste products, and the dis-
appearance of the nuclear membrane offers an opportunity for
their discharge from the nucleus. Thus in mitosis, while
nutritive processes are arrested, free play is given to nuclear
excretion, and so the amassing of waste products in the nucleus
is prevented.

6. Relatio7i to Centrosomes. — The following writers have
assumed a genetic connection between nucleoli and centro-
somes : Karsten {Psilotiim, the observations combatted by
Humphrey) ; Wasielevsky, Sala, Carnoy and Lebrun {As-
carts) ; Lavdowsky (various cells of animals and plants) ; Wil-
cox (spermatocytes of Caloptemis) ; Balbiani {Spirochond) ;
Bremer (blood corpuscles of Reptiles) ; Toyama (spermato-

1 66 BIOLOGICAL LECTURES.

cytes of Bombyx) ; and Julin. Keuten finds in Ceratium a
<* Nucleolo-Centrosoma," which is supposed to unite the char-
acteristics of nucleolus and centrosome, but the nature of this
body is still problematical, and Lauterborn found a nucleolus
but no centrosome.

The great mass of evidence goes to show that there is no
relation between nucleoli and centrosomes, and the conclusion
that such a connection exists is probably incorrect. With cer-
tain stains it is possible to color the nucleolus like the centro-
some, but there are other methods which in most cases serve
to differentiate the two. The centrosomes are frequently ap-
parent at an early stage of the prophase, while the nucleolus
is still present within the nucleus (as I have observed in Pisci-
cola and Euchistiis)^ so that in such cases no genetic relation of
the two is possible ; and it would be contrary to all well-founded
ideas to conclude that an inert mass like the nucleolus is capable
of becoming a dynamic centre of the cell.

7. Nature and Function of the Nucleolus. — From a careful
and unbiased study of the subject, one is led to conclude that
the true nucleolus is an inert mass and not an organ. Its
substance appears to be extranuclear in origin, as already
discussed, and at the time of its first appearance would not
seem to be a derivative of the chromatin, as maintained by
Hacker. Since nucleolar substance is found in the greatest
abundance in cells w^hich are undergoing a rapid process of
growth, as is well shown in the growth period of ovarian ova,
its amount would seem to stand in a pretty direct relation to
the intensity of nutritive metabolism ; and this is the reason
why its substance may well be considered a product of nutri-
tive metabolism. Exactly what kind of a product, we are not
yet in position to decide ; it may represent accumulated nutri-
tive material, or waste products of such material, or both.
Perhaps it is at an earlier stage a mass of mixed nutritive mat-
ter, from which the other elements of the nucleus gradually
extract the nourishment, until finally only indigestible sub-
stances are left in it. Possibly chromatin debris may become
mingled with the nucleolar substance, but there is good reason
for supposing that the nucleolar substance is not wholly a

NUCLEOLAR STRUCTURES OF THE CELL. 167

secretion of the chromatin, and that little, if any, chromatin
substance enters into its composition.

That the chemical nature of nucleoli varies in different cells
is very probable; and in such objects as germinal vesicles of
insects the component granules of a nucleolus appear quite
heterogeneous in nature. These cases we may explain on
Rhumbler's theory, that compound nucleoli are formed by
consecutive fusions of smaller nucleoli, the latter varying
physically and chemically, since they are produced at dif-
ferent times.

Such appears to me to be the genesis of nucleolar substance.
But is there any probability that a nucleolus at some stage in its
history may become an organ, i.e.^ actively fulfil a particular func-
tion } It has been considered by Hacker as an excretory organ,
contractile vacuoles in it having been observed by him and Bal-
biani ; while van Beneden has described rhythmic expansion and
contraction of a nucleolus. It seems to me probable, but by no
means definitely settled, that in most cases the nucleolus can-
not be considered more than an inert mass ; and that if it should
in any case prove to become differentiated, or vitalized, into an
organ, it could be regarded as little more than a receptacle for
waste products of the nucleus, possibly exerting a force to attract
such substances to itself. But the mass of evidence would show
the nucleolus to be a mass of non-living substance. The usually
globular form can well be explained upon its usual fluid consist-
ency ; and the regularity in number, which is a striking phe-
nomenon in certain nuclei, as, e.g., the pronuclei of fecundation,
may well be explained by the assumption that there are partic-
ular spaces within the nucleus specially fitted for the deposition
of nucleolar substance.

The hypothesis of Julin and Henneguy, that the nucleolus
and Balbianian corpuscle of metazoan ova together correspond
to the macronucleus of Infusoria is interesting, but an interest-
ing speculation only. It is of importance to note that in this
group of Protozoa only the vegetative macronuclei contain true
nucleoli, while they are absent in the animal micronuclei.

Here may be mentioned briefly the conclusions of certain
authors in regard to the nature of the nucleolus, in order to

1 68 BIOLOGICAL LECTURES.

demonstrate how little unanimity exists in regard to it. Many
{e.g.y Leydig, Klein, Retzius) regard it as merely a portion of
the chromatin reticulum, confusing it with the karyosome.
Macfarlane considers it the dynamic centre of the cell. Car-
noy distinguishes '' nucleoles nucleiniens " {i.e., karyosomes),
*'n. plasmatiques " {i.e., true nucleoli), "n. mixtes " (formed
by the union of the preceding two), and ** n.-noyaux," which he
regards as nuclei within nuclei. Strasburger and others regard
it as a reserve supply of chromatin ; Hacker, as a secretion of
the chromatin, with possible excretory function. Flemming
considers it a nuclear organ, and as a *' chemische Modification,
Vorstufe oder Doppelverbindung," of chromatin. Korschelt's
view is approximately the same as that reached by myself ;
namely, that nucleoli are formed as depositions of nutritive
substance. Rhumbler considers them to represent reserve ma-
terial, standing in some connection with chromatin ; Watase,
that they are metabolic products of the cell. Born concludes
that " die Nucleolen stehen in Beziehung zum individuellen
Zell-Leben, nicht zur Fortpflanzung." Further, there are nu-
merous modifications of these several views.

As a matter of fact, the nucleolar phenomena are so multi-
form and complex, on a broad comparative view, that it is ex-
tremely difficult to decide the formation and nature of nucleoli.
The nucleolus is a structure upon which careful metabolic stud-
ies can be well pursued, but such investigation must be by the
comparative anatomical method as well as by the minute re-
search of a particular object. Neither method should be disre-
garded in morphological work, but the comparative would seem
to offer safer results.

II. Other Nucleolar Structures.

Besides the true nucleoli, which have just been discussed,
and the karyosomes, which in the strict sense of the term are
merely thickenings of the chromatin reticulum, there are found
in some cells structures which appear to resemble the chromatin
chemically, but which are not integral portions of the nuclear
network. Such, for instance, are the ''double nucleoli" of

NUCLEOLAR STRUCTURES OE THE CELL. 169

many spermatocytes and of some germinal vesicles (particu-
larly of polychaete annelids), which are less frequent in somatic
cells. These double nucleoli are usually formed of two apposed
spheres, of which the one is a true nucleolus, while the other
consists of a substance staining like chromatin. Not only may
these bodies occur in pairs, but in chains of three or more (as
first noted by Hermann in follicle cells of the testicle of the
mouse) ; in other cases the true nucleolus may be separated
from the other sphere, or the two may be at one time sepa-
rated and at another united. Many of the older observers,
who did not make use of sections, have described as double
nucleoli, true nucleoli which enclose a large excentric vacuole,
considering this vacuole to be a second nucleolus applied to the
first ; such cases should not be regarded as double nucleoli.
These remarkable " nucleolar " structures, which stain like
chromatin, have been observed by numerous writers, but as
yet no satisfactory description has been given of their mode of
origin. They have been observed by me in spermatocytes of
various insects, in hypodermal and other cells of Carpocapsa^
and in follicle cells of the testicles of Plethodon and Miis.
Whether structures comparable to them occur in germinal vesi-
cles, I have not as yet been able to determine. In two cases I
have followed their mode of formation, and these cases may be
described here.

One point may be emphasized : the term " chromatin nucle-
olus " should be applied only to those structures which can be
proved to be derived from chromatin ; in those cases where
their mode of formation has not been followed it seems best to
avoid the use of a special term, since it is by no means certain
that all these structures should be classed in the same category.

I. The Spermatocytes of Euchistiis {Petttatoma). — The sper-
matogonia contain only a true nucleolus, which disappears in
the prophase of mitosis. The fourteen chromosomes of the
last spermatogonic division are halved in metakinesis, so that
each daughter-cell (first spermatocyte) receives fourteen daugh-
ter-chromosomes. The stain necessary for demonstrating the
consequent phenomena is one which will stain the dividing
chromatin a different color from the resting chromatin : the

170 BIOLOGICAL LECTURES,

only method known to me which does so satisfactorily is the
excellent safranine-gentian violet-iodine stain of Hermann (pub-
lished in 1889 in the Arch. f. mikr. Anatomie)^ by which the
resting chromatin is colored violet, and the dividing chromatin
a bright red. By the term " dividing chromatin " is understood
the chromatin from about the commencement of the prophasic
loose spirem stage until about the commencement of the ana-
phase. This method of Hermann cannot be too highly rec-
ommended for cytological work, and in many respects is far
superior to the favorite method of the day, the iron haematox-
ylin stain, with which it is difficult to secure satisfactory differ-
entiation of the various cell constituents.

To return to our first spermatocyte, at the stage of the last
spermatogonic metakinesis. The fourteen short chromosomes
are stained red until the nuclear membrane appears ; they
seem to be all alike in form and color. At the commence-
ment of the anaphase the chromosomes become irregular in
outline and elongated, and now lose the red and commence to
take on the violet stain, passing first through a purple stage.
But now a most remarkable phenomenon is to be noted : one
of the fourteen chromosomes does not become violet but re-
mains red, the color characteristic for the dividing chromatin.
As the anaphase proceeds, thirteen chromosomes gradually
become intensely violet, while the fourteenth remains red.
The latter is the " chromatin nucleolus," as it has been
termed by me ; genetically it is a chromosome, but destined to
have an entirely different metamorphosis from the others. The
chromosomes and the chromatin nucleolus continue to elon-
gate until their length nearly equals that of the diameter of
the nucleus. Then the chromatin nucleolus ceases to elongate,
passes to the periphery of the nucleus, and gradually shortens
up into a more or less spherical form ; during this process it
may divide into two or three parts, in which case one is usually
much larger than the others. At this time the true nucleolus,
which does not stain at all, or only a faint lilac, is forming at
the periphery of the nucleus ; it may at every stage be sharply
distinguished from the chromatin nucleolus. The thirteen true
chromosomes elongate to a length of about double the diameter

NUCLEOLAR STRUCTURES OF THE CELL. lyi

of the nucleus, and then form an inextricable coil at the centre
of the nucleus, this being the stage of synapsis.

In EiicJiistiis there is a well-marked stage following the
synapsis, which may be termed the post-synapsis ; it is charac-
terized by the loosening up of the long chromosomal loops, so
that the latter may now be counted, which was impossible dur-
ing the synopsis. A careful count of the loop in a number of
nuclei shows that a reduction of the number of chromosomes
had taken place in the synapsis ; in the post-synapsis the chro-
matin loops (chromosomes) are of unequal length, and in differ-
ent nuclei vary in number from three to six, while four is the
usual number. It is specially to be noted here that in the post-
synapsis there are never exactly seven chromosomes, which
would be exactly half the normal number, but less than seven,
which fact must be brought into connection with the fact that
one of the earlier chromosomes had been metamorphosed into
the chromatin nucleolus. Interesting as the relations are from
the standpoint of chromatin reduction, they are not exactly in
line with the subject in hand, and so cannot be further described
here.

After the post-synapsis follows the telophase, and then the
rest stage of the spermatocyte. In the latter period the chro-
matin nucleolus, still red in color, lies against the nuclear mem-
brane and encloses a large globule of lighter staining substance.
The true nucleolus, larger than the former, is usually unstained
(only browned by the action of Hermann's fixative) and lies
near the centre of the nucleus. The chromatin reticulum is
deep violet, and special chromosomes are not distinguishable.

In the prophase of the first reduction division no continuous
spirem is formed, but at the stage of the dense spirem the chro-
matin reticulum segregates into from three to six long loops.
The latter are of unequal lengths, and, according to the num-
ber of segments present in the dense spirem stage, one or more
of these loops divide transversely, until in the loose spirem
stage always exactly seven chromosomes (exactly half the nor-
mal number) are evolved. This remarkable power of self-reg-
ulation of the number of the chromosomes has been discussed
more fully in another publication. In the loose spirem stage

172 BIOLOGICAL LECTURES,

each of these seven chromosomes shortens up into a regular
dumb-bell shape.

During these prophases the chromatin nucleolus decreases in
size, partly by the loss of the substance of the contained vacu-
ole, partly by a dissolution of a portion of its own substance ;
but a portion of it always remains, and this portion gradually
becomes dumb-bell shaped before the disappearance of the
nuclear membrane. The true nucleolus disappears completely
during the prophases.

In the equatorial plate of the first reduction spindle lie the
seven chromosomes, and by their side the chromatin nucle-
olus, in every case easily recognizable by its much smaller size.
The seven chromosomes and the chromatin nucleolus are then
divided transversely in metakinesis, so that each daughter-cell
(second spermatocyte) receives seven chromosomes and one chro-
matin nucleolus. But in the second reduction division, unlike
the first, the chromatin nucleolus does not divide as a rule, but
is carried undivided into the one or the other spermatid.

Thus it is that in the spermatocytes of Eiichistiis a whole
chromosome becomes metamorphosed into a chromatin nucle-
olus ; and from observations made by me upon another genus
of the same group, Anasa^ it would appear that the chromatin
nucleolus has a similar origin here also. In the spermatocytes
of some other insects {Phalangopsisy Gryllus^ and Harpahis) I
have found similarly staining structures (which in Harpahis
are attached to the true nucleolus, thus constituting a double
nucleolus), but in these objects have not yet traced their origin.
May it not be that chromatin nucleoli will be found to be of
general occurrence in the spermatocytes of insects, if not indeed
of other groups also.^* If this be proved to be the case, and
the chromatin nucleolus in all objects be found to be a meta-
morphosed chromosome, then the chromatin reduction in such
spermatocytes would be of a different character from that in
ovocytes ^nd the whole problem be still further complicated.

2. Hypodermal Cells of Cajpocapsa. — The budding extremi-
ties of the apple-worm, which is the larva of the moth Carpo-
capsa, offer a beautiful object for the study of certain nucleolar
structures. The hypodermis covering the body consists of

NUCLEOLAR STRUCTURES OF THE CELL. 173

somewhat flattened epithelial cells, in the nuclei of which only
true nucleoli are present. But in the hypodermal cells of the
larval extremities both nucleoli and chromatin nucleoli occur
within the same nucleus ; and at the place where the extremity
joins the body proper the hypodermal nuclei offer a perfect
succession of gradations, showing all stages in the formation
of the chromatin nucleolus. The same gradations of nuclear
structure are also to be found where the hypodermis is invagi-
nated to form the stomodaeum and proctodaeum.

At the proximal end of a larval extremity the nuclei contain,
besides the chromatin reticulum, one or two true nucleoli. As
we proceed, distad nuclei are found, which, in addition to the
true nucleoli, contain each one small granule, which stains ex-
actly like the chromatin, and for the designation of which the
name ** chromatin nucleolus " is proposed. At the time when it
may first be clearly distinguished it is simply a small granule,
a little larger than the nodal points of the chromatin network.
While it is difficult to be positive of the conclusion, it would
appear to be a portion of this network, separated from the rest
of the chromatin. It is at the start usually in contact with a
true nucleolus, so that double nucleoli are thus produced. Now,
as the succession of stages is studied, it is found that the chro-
matin nucleolus steadily increases in size, while the true nucle-
olus attached to it either remains constant in volume or grows
much more slowly. A vacuolar substance appears within the
chromatin nucleolus and stains like the true nucleolus ; the nu-
cleus itself is increasing in size and gradually becoming amoe-
boid. Putting these facts together, — though unfortunately
they can be mentioned here only briefly, — it appears that
to the true nucleolus new substance is continually being added
from without, while the chromatin nucleolus is absorbing mate-
rial from the true nucleolus. In no other way can be explained
the apposition of the two, accompanied by the rapid growth
of the chromatin nucleolus and the deposition of nucleolar-like
substance within the latter. Finally, in the large amoeboid
nuclei, which are found at the distal end of the extremity,
where these cells produce cuticular spines, a large chromatin
nucleolus is seen, but no longer any trace of true nucleoli.

174 BIOLOGICAL LECTURES.

In this interesting series of changes we observe a granule
which appears to be a derivative of the chromatin reticulum,
applying itself to a true nucleolus, and by gradually absorbing
the substance of the latter, becoming a large chromatin nucle-
olus. Is it a specially modified portion of the chromatin retic-
ulum which is predestined to undergo this change? It would
seem more probable that that portion of the chromatin reticu-
lum lying closest to the true nucleolus would be the portion
selected to grow.

The cells of the extremities are the producers of the cutic-
ular structures of the latter, and their large size, amoeboid
nuclei, and chromatin nucleoli are all evidently occasioned in
the course of this production ; and beneath each of the cuticu-
lar spines, which occur elsewhere on the body surface, lies a
mass of cells much larger than those of the surrounding hypo-
dermis, the relatively enormous amoeboid nuclei of which con-
tain each a chromatin nucleolus. There can be no doubt that
these chromatin nucleoli have been formed in the same manner
as those in the cells of the extremities, though I have not seen
the various stages of their formation. And in the body cavity
are enormous cells, each with an amoeboid nucleus and a chro-
matin nucleolus, which are, undoubtedly, cells which at one
time were situated within the hypodermis and functionated as
spine producers, and later lost this connection, falling into the
body cavity, where they increase greatly in size and probably
fulfil some new function.

In cells of other tissues of the Carpocapsa larva chromatin nu-
cleoli are also to be found, as in the intestine and salivary gland
cells ; but their genesis could not be traced by me, as it was
for the hypodermal cells of the extremities and of the stomo-
daeum and proctodaeum ; earlier ontogenetic stages might give
the solution of the question.

What is the meaning of the evolution of such chromatin
nucleoli — structures which have not frequently been observed
in cells, and the genesis of which has not heretofore been
followed.? The chromatin is a living, and, from the morpho-
logical standpoint, most important substance. So that, if my
observations prove the chromatin origin of the chromatin nu-

NUCLEOLAR STRUCTURES OF THE CELL. 175

cleoli in Carpocapsa and Euchistiis, these nucleolar structures
might be considered organs of the cell, e.g., centres of nutri-
tion, and so be of a nature essentially different from that of
the true nucleoli.

In conclusion, the hope may be expressed that the study of
nucleolar structures will receive more attention in the future
than it has in the past. Whoever has had the patience to read
this brief summary of observations will grant that the nucleoli
are structures of manifold complexity, the study of the genesis
and fate of which may throw needed light upon questions of
cellular metabolism ; and there are three groups of animals
which bid fair to offer the best opportunity for work : the Pro-
tozoa, and particularly the MastigopJiora and Foraminifera, the
study of which must eventually give the best basis for cyto-
logical conclusions ; the Arthropoda, which surpass all other
forms in the diversity of cellular structure ; and the Mollusca,
which in this regard are close rivals of the arthropods.

NINTH LFXTURE.

PROTOPLASMIC CONTRACTILITY AND
PHOSPHORESCENCE.

S. WATASfi.
I.

It is now generally admitted that the emission of light from
an animal organism is due to oxidation of a certain substance
produced by the metabolism of the cell. This is easily proved
in the case of a glowworm or a firefly, where the photogenic
tissue can be separated from the organism artificially and sub-
jected to various physical and chemical tests. We learn from
such experiments that anything that helps the oxidation in
general, such as alkaline media, moisture, agitation, heat, etc.,
also facilitates the phosphorescence of the material; while, on
the other hand, exposure to carbon dioxide stops it. The pho-
togenic material is formed as the result of metabolism in the
cell, and in the process of its formation is identical with that
known as secretion.

Before we go any further into our subject a few words are
needed on the use of the term " secretion," with special refer-
ence to the formation of photogenic material in phosphores-
cent organisms, because, as I shall point out more fully later,
it was the narrow, limited use of the term that impeded the
better understanding of the phenomena of phosphorescence as
a whole.

Secretion, as popularly understood, suggests the existence of
a gland, secreting a material which is eventually discharged
from the organism, as is familiarly seen in the secretion of
saliva, mucus, or poison of various kinds in different organ-

177

178 BIOLOGICAL LECTURES.

isms. But this is a rather limited use of the term "secretion."
As we know, there are several glands in the organism whose
products of secretion are never emitted to the external world,
but put to some definite use within the organism. Discharge
of the material from the organism that produced it is not the
essential characteristic of the secretion process.

The process of secretion does not necessarily imply the ex-
istence of a gland. When one examines the cytological aspect
of the phenomena of secretion, it resolves itself into this : dis-
integration of the protoplasm and the subsequent conversion
of the material thus produced into the secretion granules, or
the separation of the specific substance from the more or less
crude material incorporated within, or brought in contact with
the body of the cell, under the influence of the protoplasm.
So far as the fundamental process of secretion is concerned,
it may be carried out by an isolated single cell just as well as
by the hundreds and thousands of similarly constituted gland
cells forming a sac-like recess, opening on the free surface
of the organ or of the organism.

In the case of ordinary, secretion, such' as saliva, the granules
formed by the metabolic activity of the protoplasm undergo
still further changes, and eventually escape from the body of
the cell. In other cases these granules remain as such, and
subsequently become consumed by the activity of the proto-
plasm that produced them, as is well seen in the history of the
yolk-granules in the egg-cell. The granules produced by the
process of secretion at one period become utilized by the
growing egg-cell in the course of its own development. ^

Just as the liver or the thyroid is an organ of " internal

1 " The egg-cell of most animals, at any rate daring the period of growth, is by
no means an indifferent cell of the most primitive type. At such a period its cell-
body has to perform quite particular and specific functions ; it has to secrete nutri-
tive substances of a certain chemical nature and physical constitution, and to store
up this food material in such a manner that it may be at the disposal of the embryo
during its development. In most cases the egg-cell also forms membranes which
are often characteristic of particular species of animals. The growing egg-cell is,
therefore, histologically differentiated, and in this respect resembles a somatic cell.
It may perhaps be compared to a gland cell, which does not expel its secretion, but
deposits it within its own substance?^ Weismann, The Continuity of the Germ-
plasm as the Foundation of a Theory of Heredity, 1885.

CONTRACTILITY AND PHOSPHORESCENCE. 179

secretion " for the whole organism, so the yolk-granules in the
egg-cell are the products of " internal secretion " for the uni-
cellular stage of the same organism.

The cytological significance of secretion, then, whether in a
single cell or in a multicellular gland, is essentially the same,
and the method of the secretory process is wholly independent
of the ultimate disposal or use of the products. Secretion,
then, consists essentially in the production of specific sub-
stances, with the help of, or at the expense of, the protoplasm ;
and whether that material be expelled from, or consumed in,
the organism is a question of secondary importance.

Bearing in mind, then, the general sense in which the term
''secretion " has been defined, let us examine luminous organs in
different organisms and see how the light-giving material is
disposed of. According to the way in which it is used, lumi-
nous organisms may be divided into three groups. I am here
speaking of the true phosphorescent organisms which produce
light by the activity of their own tissues, and not of those which
assume a luminous appearance by the habitual or accidental
incorporation of other luminous organisms in their own body.

{a) The first group of luminous organisms includes those in
which photogenic substance is thrown out of the body, from a
special gland, in the form of liquid, as Pholas, Copepods, etc.;
or from the external surface of the body epithelium in the form
of fine refringent granules, as in ChcBtopterus^ luminous earth-
worms, myriapods, etc. The secreted material assumes its
luminous appearance when it comes in contact with the sur-
rounding media, the air or sea-water, as the case may be.

{b) The second group embraces those in which photogenic
material produced by the secretory activity of the cell is never
thrown out to the external world. Oxidation of the material is
accomplished by the oxygen taken with the air, or with the
oxygen dissolved in the body fluid. In many cases elaborate
mechanisms are developed in connection with the luminous
tissue to facilitate the oxidation of the photogenic material.
Various devices to intensify the luminous effect may be su-
peradded, making them wonderfully like the visual organs,
with which, indeed, they have often been confounded. This

i8o

BIOLOGICAL LECTURES.

type of luminous organ, in which photogenic substance never
leaves the body, is illustrated by the more familiar phosphor-
escent organisms, such as fireflies (Fig. i), glowworms, Pyj'o-
phoruSf Euphaiisiay Nyctiphanes (Fig. 2), Phengodes, Scopehcs
(Fig. 3), etc.i

The phosphorescent organism of this group, as a rule, has
good control over its luminous organ, and may often be able to

Fig. I. — Photinus consanffUine7{s,nghifig[xrQ, female ; lait, ma/e. Luminous
organs are shaded.

inhibit the exhibition of light for a considerable length of time.
The organism may be endowed with a wealth of structures sug-
gestively like the luminous organs, yet they may never satisfy
the naturalist that they are such if the exhibition of their lumi-

^ It is impossible to give any adequate idea of the beauty and complexity of
the luminous mechanism of this type without the use of a large number of illus-
trations. Some ideas in regard to the size and distribution of the luminous organs
that obtain in some forms may be gathered from a few figures that accompany the
present paper.

CONTRACTILITY AND PHOSPHORESCENCE.

8i

nous property is left to the choice of the organism itself. The
best way to deal with an organism of this nature is to expose it
to the fumes of ammonia, or add a few drops of strong ammonia
to the sea-water in which it lives. Even the most obstinate of
phosphorescent organisms may be made to reveal their luminous
property by this method. Another and perhaps a simpler

Fig. 2. — Nyctiph'anes norvegica. Luminous organs are shown by dark dots.

method is to extract the tissue of the suspected organ, and
crush it or tease it and expose the contents freely to the air,
in the presence of moisture. If the tissue is really photogenic,
it will shine when thus separated from the inhibitory control of
the organism, and the direct access of oxygen to the tissue is
thus artificially made possible. A small and slender phosphor-

0 t W 01

Fig. 3. — Scopelus Rafinesqiiii. (After Liitken.) Luminous organs are shown by dark dots.

escent earthworm which inhibited the emission of light in a
most obstinate way, when gently crushed between two slides
and exposed to the air or thrown into water, shone like the
filament of an incandescent lamp.^

1 The presence of water seems to exercise quite a marked influence upon the
luminous oxidation of the substance, although water does not appear to take any

1 82 BIOLOGICAL LECTURES.

It is, however, possible that some tissues in a non-luminous
organism, when thus treated and examined in the dark, may
prove to be phosphorescent.

It is highly desirable that those who have access to various
organisms whose photogenic property is suspected from ana-
tomical grounds, but not yet definitely proved from the func-
tional standpoint, subject such organisms or their suspected
tissues to these methods of inquiry. The long-standing con-
troversy whether the lantern-fly {Fulgora) is phosphorescent or
not, might be settled at once by taking the tissues from the
suspected region in the living insect and crushing them. The
question whether the eye-like organs in the different parts of
the body in bony fishes, Cephalopods, and other organisms, are
phosphorescent or not, ought not to be difficult to settle for
those who have access to living individuals, or even to dead
ones, if the decomposition is not very far advanced ; for the
photogenic substance retains the power of phosphorescence
long after the death of the organism.

{c) The third group consists of those in which the definite
photogenic organ as such does not exist, but the light-giving
material formed by the secretory process of the protoplasm
accumulates along the course of muscle-fibres or other contrac-
tile protoplasmic material, the light manifesting itself in sparks
or scintillations along the course of the fibres at the time of
their contraction. This species of phosphorescence is the most
interesting one, complicated as it is with the fundamental prop-
erty of the protoplasm, namely, contraction. It was this kind
of luminous phenomena which retarded the acceptance of the
theory that phosphorescence is due primarily to the secretion
of a certain material which shines in the process of oxidation.
Thus, Alexander Agassiz,^ in his interesting observations on
the phosphorescence of Ctenophores, says: "Although we
know now something of the nature of the phosphorescence of
a few marine invertebrates from the observations of Panceri,

part in the process itself. It is probable that water acts as a catalytic agent
through which the process of oxidation is accelerated.

1 Agassiz, Alexander, "Embryology of the Ctenophorae," Mem. Amer. Acad,
of Arts and Sci., vol. x, No. 3, 1874.

CONTRACTILITY AND PHOSPHORESCENCE. 1 83

who has plainly traced it to the secretion of special glands, yet
when we find the same phosphorescence equally as brilliant in
eggs of Ctenophorae as in the adults, even in the stages in
which the masses of segmentation can still be counted, we have
evidently not yet reached the solution of the true nature of this
phosphorescence. The whole embryonic mass becomes bril-
liantly phosphorescent when the least shock is given to the jar
in which the eggs are kept."

The probable explanation of this remarkable phenomenon is
that the mechanical shock given to the jar causes the proto-
plasm of the ^^g to contract, and this contraction is accom-
panied by the manifestation of light. The well-known case of
phosphorescence of Noctiltica, to which I shall refer in detail
later, belongs to the- same category. Any stimulus that causes
the contraction of the protoplasm also causes the phosphores-
cence of the organism.

One of the most eminent students of phosphorescence, Qua-
trefages,^ while freely admitting the efficiency of the oxida-
tion theory of phosphorescence, sharply draws a line at the
last class, where phosphorescence is associated with contrac-
tion, and falls back on the vitalistic theory of luminous phe-
nomena in vogue at the beginning and in the middle of the
present century. But I think it is especially this sort of phos-
phorescence which is of peculiar interest to us. Intimately
connected as it is with the phenomena of contraction, we may
hope that the careful consideration of the subject may throw
some light on the understanding of protoplasmic contraction
itself.

Let us examine somewhat in detail the facts of the case,
and then consider them in connection with the phenomena of
muscular contraction. Noctiluca is a little microscopic animal,
sometimes attaining the diameter of i mm., bearing a general
resemblance to little melons, deeply indented at one end (Fig. 4).
Near this depression there is an appendage by which the ani-
mal moves slowly to and fro, swaying it from right to left.

1 Quatrefages, A. de, " Memoire sur la phosphorescence de quelques in verte-
bras marins," ^««. des Sci. Nat., 1850, vol. xiv, pp. 236-287; "The Phosphor-
escence of the Sea," Popular Science Review, vol. i, 1862, pp. 275-298.

184 BIOLOGICAL LECTURES.

The body is so transparent as to admit of its structure being
studied in minutest detail. The protoplasm of the body shows
a fine ramification from the centre outward to the periphery.

The ultimate branches of the protoplasm spread themselves
over the whole inner surface of the animal. The central mass
of the protoplasm and its ramifications are constantly moving
and changing ; larger branches becoming thinner ; thinner
branches becoming thicker by borrowing material from their

Fig. 4. — Noctiluca miliaris. (After Leuckart and Nitsche.)

neighbors. The protoplasm is highly contractile, so much so
that when the organism is stimulated energetically, the minute
branches along the periphery contract so violently as to tear
themselves completely from the external membrane.

The most interesting phenomenon in this connection is the
close correlation of the phosphorescence with contraction, as I
have already said. This Quatrefages proved by two sets of
experiments, the one tried in the daytime, and the other at
night. Thus, the action of heat, electricity, compression, and

CONTRACTILITY AND PHOSPHORESCENCE. 185

of different chemical substances — alkaline, acid, and neutral —
upon Noctiluca have been investigated, the general result being
that any agent whatsoever capable of causing the protoplasmic
substance to contract was efficacious to precisely the same degree
and at the same time in producing phosphorescence. An isolated
shock caused during the daytime a violent but momentary con-
traction ; at night it gave rise to a bright but transient exhibi-
tion of light.

Under the influence of continued or too violent irritation,
the whole substance of the body is seen in daytime to contract
to such an extent as to become free from the external envelope,
and gradually become disorganized or lose its vital properties.
At night the same operation leads to a continual phosphores-
cence, apparently pervading the whole animal. The steady
and permanent brilliancy always denotes approaching dissolu-
tion. Thus contraction and phosphorescence are invariably
met side by side, and, as it were, hand in hand. The contrac-
tion of the protoplasm may be either partial or general ; it may
manifest itself in different places at the same time, and so also
with the luminosity. The important point to bear in mind in
this connection is Quatrefages' discovery that the luminous
area of the Noctiluca is composed of an enormous number of
hmiinous points.

Examined with a power of twenty or thirty diameters, the
illuminated portion is uniformly bright. With sixty diameters,
a number of small but brilliant scintillations become visible,
and these come and go with the rapidity of lightning. An
enlargement of one hiindred and fifty diameters reveals the
true character of the phenomenon. It then becomes obvious
that the light e7nitted fro7n the whole body, or any of its pai'ts,
is composed of a vast number of instantaneotis scintillations ^
closely approximating to one another at tJie centre of the '* phos-
phorescent'' portions^ but clearly distinguishable at the edges

(Fig. 5)-

Occasionally there may be seen isolated sparks at the extreme
limit of the luminous areas, or even beyond.

Quatrefages extended his observations to higher forms,
where the phosphorescence is intimately associated with con-

1 86 BIOLOGICAL LECTURES.

tractility. He succeeded in seeing the motor muscles of the
luminous Ophmra suddenly illuminated under the microscope.
The muscle alone emitted light, the other portions of the foot
remaining perfectly obscure. These muscles were not perfectly
luminous throughout their entire length, but there appeared
disseminated over them a vast number of exceedingly minute,
but at the same time remarkably brilliant points, which appeared
and vanished with lightning-like rapidity. The outlines which

Fig. 5. — Noctiluca miliaris, representing a portion of the body, with a large number
of scintillating points. (After Quatrefages.)

they presented consisted not of an uninterrupted track of light,
but of a line formed by a succession of scintillations.

It would be easy to multiply examples, but the preceding,
quoted freely from Quatrefages, will suffice to show some deep-
rooted physiological connection between the production of light
and the contractility of protoplasm.

I shall now pass on to the second section of my paper,
concerning the nature of protoplasmic contraction ; more par-
ticularly the contraction of protoplasm in one of its most differ-
entiated aspects — the striated muscle cell.

CONTRACTILITY AND PHOSPHORESCENCE. 187

II.

Among many theories that have been proposed by physi-
ologists, that of Engelmann ^ is most complete. This theory
Engelmann calls the *' thermodynamic theory of muscular con-
traction," in which a muscle cell, or a bundle of muscle fibres,
is compared to a thermodynamic machine, not unlike our steam
engine. The theory was proposed by Julius Robert Mayer and
elaborated by Engelmann, from the careful study of the ana-
tomical, physiological, chemical aspects of the muscle cell,., cilia,
unicellular forms, in short, in all organisms in which truly con-
tractile phenomena are exhibited. According to this theory,
the work the miLscle accomplishfs derives its energy from the heat
developed iii the tissue, just as the work an engine accomplishes
is developed by the combustion of the fuel. The primary form
of contractile energy in the muscle is heat, which becomes
transformed into the mechanical work of contraction.

The fact that the filamentous structures of the muscle are
really contractile elements had been demonstrated by Engel-
mann by separating the fibres from the muscle cell. The
isolated fibrils contract under a definite stimulus, thus refuting
an idea entertained by some histologists, that the liquid portion
of the muscle cell is contractile. According to Engelmann, the
impulse given to the muscle cell causes the chemical change
in the non-contractile substance surrounding the contractile
filaments ; the latter are metamorphosed to such a high degree
that they do not readily undergo chemical changes. The chem-
ical disturbance started in the neighborhood and in the space
surrounding the contractile filaments takes the form of oxida-
tion chiefly, as shown by the nature of by-products. The heat
developed by the oxidation of thermogenic particles in the inter-
stices of contractile fibrils enables the latter to absorb more
water. And when an organized substance absorbs water, it
becomes thicker and shorter. This thickening and shortening
of the component fibres is the contraction of the entire cell.

1 Engelmann, Th. W. Uber den Ursprung der Muskelkraft. Leipzig, 1893.
♦* On the Nature of Muscular Contraction," Croonian Lecture before the Royal
Society of London, 1895.

1 88

BIOLOGICAL LECTURES.

Engelmann illustrates this by an ingenious model (Fig. 6).
A piece of an E string of a violin about 5 cm. long, previously
soaked in water, is fastened to the rigid end of the rod a^ of

Fig. 6. — Diagrammatic Model of the Muscle. (After Engelmann.)

ab^ while the upper end of the string is fixed on the shorter
arm of the lever H. To this string different tensions may be
imparted by weights d, di.

Round the string, but without touching it, runs for a length
of about 20 mm. and in about 20 curves a spiral of thin plati-
num wire. The poles of the wire are connected with two wires
coming from the poles of a Grove or Bunsen battery of three
or more cells. The whole is immersed in a glass of about 50
cc. contents, and filled with water of about 55°-6o° C. A ther-
mometer is placed through the aperture of the ebonite cover.

When no change in the position of the lever occurs, close for
some seconds the circuit of the battery through the spiral ; the
lever rises. Upon opening the circuit it falls. The thermometer
in the glass indicates scarcely any rise in temperature.

CONTRACTILITY AND PHOSPHORESCENCE. 189

The doubly refracting violin string corresponds to the doubly
refracting particles in muscle fibrils ; the water in the glass
represents the watery isotropic substance round the contractile
fibril, doing duty as refrigerant ; the spiral wire supplies the
place of the chemically active thermogenic molecules ; the clos-
ure of the galvanic circuit corresponds with the process of the
stimulation of the muscular element.

Some physiologists object that the body of an organism or a
muscle ''is sensibly uniform in temperature throughout, and
that the more work done, the more rapid the circulation ; and
the more certain this uniformity of temperature, the more it
seems to be impossible that such thermodynamic processes are
carried on in the animal system as are familiar to us in our
heat-engines."^

To the objection that the body of an organism, or muscle
cell, does not satisfy the conditions of a thermodynamic
machine because the temperature is uniform, Engelmann re-
plies that '* we must assume, on the contrary, exceedingly large
differences of temperature in the stimulated muscle. What
holds good of the whole body holds good of the muscle also."
As Pfliiger observes, the temperature measured with our instru-
ments is but an arithmetical average^ ^^ comprising an infinite
number of different temperatures^ pertaining to an infi7iite num-
ber of diffei'ent points.

" From the fact that at the contraction of the muscle an
infinitesimal part only of the muscular mass is chemically
active, we infer that the temperature of their particles must, at
the momejit of combustion, be art uncommonly high one. Great
as the specific heat of muscular substance is, it would be im-
possible otherwise to account for a rise in temperature of the
whole mass even of 0.001° C, only. Without any exaggera-
tion, we must assume that the temperature of the chemically
active particles may, at the moment of combination, exceed the
average muscular temperature by hundreds of degrees.

*' Since each thermogenic particle is surrounded by a rela-
tively enormous cool mass, conducting heat and diathermous,

^ Thurston, R. H. The Animal as a Machine and a Prime Motor, and the
Laws of Energetics. New York, 1894.

190 BIOLOGICAL LECTURES.

the principal condition for the transformation of heat into
mechanical work is satisfied, and, on account of the enormous
difference in temperature which we have to assume in such a
high degree that even an economic coefficient of 30 per cent, nay,
50 per cent, and even more, seems to be theoretically possible."

The objection has been raised ''that those high tempera-
tures must necessarily destroy the life of the muscle, since the
latter becomes rigid and dies, even at 50° C." But Engelmann
considers this objection is of small value. " For it is ever an
infinitesimal part only of the muscular mass that is exposed to
the high temperatures. At a small distance from these fur-
naces of heat the temperature must have fallen so low as to
be harmless. The muscle will no more be destroyed by stimu-
lation than a steamer will be destroyed by heating the furnaces.
The material of combustion only will be destroyed ; the whole
steamer as such remaining unharmed."

How the heat thus generated by the thermogenic particles
of sarcoplasm becomes converted into mechanical energy of
contraction may be stated in a few words. It acts as a stimu-
lus (Reiz) to the contractile fibrils, which, under the influence
of a higher temperature, absorb watery liquids with avidity.
Production of heat, then, is the essential antecedent to the
contraction of the contractile fibrils, because it is through in-
creased temperature that contractile filaments are made to
absorb more watery substance, and therefore become shorter
and thicker.

There is one point which ought to be kept in mind in apply-
ing this theory to the explanation of contractility in general,
and that is, we must not confound the conception of heat with
that of temperature. Heat is a definite thing, something objec-
tive ; temperature is a state, a mere condition of the body.
Burn ten pounds of coal, and warm ten gallons of water, and
we get a certain temperature of the water. Burn the same
amount of coal, and warm twenty gallons of water instead, and
we get a temperature of the water differing from the first. The
amount of heat developed in each case is the same, but the
temperatures of the water are different. We must not con-
clude, therefore, because a certain body is cold to our touch it

CONTRACTILITY AND PHOSPHORESCENCE. 191

is devoid of heat. The organism may live in water of freezing
temperature, yet it is perfectly possible that there is enough
difference of temperatures in different parts of the living cell
to serve as a stimulus for the contraction of the protoplasm.

III.

Now, coming to the third and concluding part of the paper,
vi::., the relationship of protoplasmic contractility and phos-
phorescence, we may note at the outset that the distinction
drawn between heat and light is purely of an artificial charac-
ter. They are simply variations of the same radiant energy.
Their difference objectively considered is simply that of degree
and not of kind. The wave-length of the one is a little longer
than that of the other. If they appear to be so different to us
as heat and light, the trouble lies in our organization, and not
in the nature of original radiant energy. The difference is
subjective and not objective.^ The heat-producing particles
and the light-producing particles, objectively considered, may
not be very different from each other. TJiey may be variations
of similar chemical substances^ as the resulting energies^ the prod-
ucts of tJieir oxidation, are the variatioiis of tJie same radiant
energy. They may be very nearly related in chemical struc-
ture as well as in anatomical origin ; they may exist in one and
the same cell, as the products of "secretion " or of the metab-
olism of the cell.

The stimuli, therefore, such as mechanical jars, shocks, agi-
tation, nervous impulse, etc., which induce combustion of the
thermogenic molecules, may also be presumed to incite combus-
tion of the photogenic molecules as well, which exist side by
side.

When there is a muscle in which these two varieties of oxi-
dizable substance exist side by side, then we have contraction
as well as phosphorescence at the same time. From one vari-
ety of granules, heat is generated, which becomes converted
into the mechanical work of contraction ; from another, light,

1 See in this connection my previous paper, " On the Physical Basis of Animal
Phosphorescence," Biological lectures, 1895.

192 BIOLOGICAL LECTURES.

which affects our eyesight. The cause which lies at the basis
of muscular contraction is essentially of the same nature as
that which produces the phosphorescence in several organisms.

Every time our muscles are put in a state of contraction
there takes place a phenomenon which is analogous to the
emission of light. ^ If we do not see the flash of light in the
contraction, the explanation perhaps lies in the nature of our
visual organs with a comparatively limited range of visual re-
action.

It is theoretically possible that all living protoplasm is phos-
phorescent, and that there may exist some organisms endowed
with different kinds of visual organs, to which organisms like
ourselves may appear luminous in the dark, especially when our
muscles are in motion. And that only when the photogenic
property becomes grossly exaggerated, as in comparatively few
representatives in each of different phyla, of living organisms,
does our eye recognize them as such. The true physical basis
of phosphorescence finds its closest analogue in the comrhon
phenomena of heat-production, and is as extensive as life itself.

1 Engelmann (Ursprung der Muskelkraft, 1893, p. 10), for example, stimulated
the muscle in the dark, until the state of tetanus was induced, thereby hoping to
see the emission of light from those heat-producing particles, which may become
incandescent at the height of their combustive process. His result was negative,
however; but, as he says elsewhere ('* Die Purpurbacterien and ihre Beziehungen
zum Licht," Bot. Zeit., 46 Jahrg. 1888, Reprint, p. 17), if it were not for the fact
that such thermogenic particles are so very minute and comparatively scarce, one
would be able to see the scintillations emanating from the inner recesses of cell-
substance, on account of the enormous heat developed in the combustion of such
particles.

The position assumed in the present paper is, that when contraction is accom-
panied by phosphorescence, which is the emission of light without any sensible
heat, there may be two kinds of oxidizable substances very nearly related to each
other genetically, one of which gives rise to heat, and the other to light. The
assumption that there exist particles which give rise practically to light, and to
light alone, is supported by the fact that such particles form the physical or ana-
tomical basis of phosphorescence in fireflies and glowworms.

If the facts and inferences presented in the present paper are somewhat for-
eign to Engelmann's theoretical conception of the relation of protoplasmic con-
traction and the emission of light, they nevertheless seem to me to throw a strong
side-light on the correctness of the fundamental assumptions in Engelmann's
theory of muscular contraction.

TENTH LECTURE.

SOME PROBLEMS OF REGENERATION.

T. H. MORGAN.

The experiments made by Trembley, in 1740, on the regen-
eration of Hydra excited amongst his contemporaries the live-
liest interest. Curiously enough, Trembley's experiments were
made in order to discover whether the new creature that he
had found was an animal or a plant, for at that time it was
well known that pieces cut from a plant would give rise to new
plants. Trembley says : " I thought once more that perhaps
these organized bodies that I had observed might be plants, and
I had the good fortune not to reject this idea. I say that I had
the good fortune not to reject this idea because, although the
less natural one, it led me to cut up some of the polyps. I
judged that if the two parts of one polyp should live after being
separated and become each a perfect polyp, it would be evident
that this organized body was a plant. Yet, since I was much
more inclined to believe that it was an animal, I did not count
much on this experiment. I expected to see the polyps die."

Reaumur had some time before shown that plant-lice could,
if they were isolated, multiply without sexual union, and Bonnet,
in 1740, convinced himself of the same fact, and Lyonet also
in the same year. These discoveries, Trembley said, sufficed
to show that in regard to the property of multiplying without
sexual union there is no difference between plants and animals.^

Trembley's results on Hydra led to the cutting up of count-

^ Je sentois vivement que la Nature etoit trop vaste et trop peu connue pour
qu'on put decider sans temerite que telle ou telle propriete ne se trouvoit pas dans
telle ou telle classe de corps organises.

193

194 BIOLOGICAL LECTURES.

less animals of all classes in the endeavor to discover whether
the same power of regeneration existed in other forms. Bonnet
(1745) and Spallanzani (1768) were most successful with their
experiments made on fresh-water worms, earthworms, the limbs
of newts, and the heads of snails {Helix). The results of the
experiments on snails, made by Spallanzani, called forth the
most vehement criticism from his contemporaries ; for many
of them seemed to consider the discovery that a snail could
regenerate its head and all the contained organs was beyond
credence. No fewer than eleven writers who had themselves
repeated this experiment (in one case on 1400 snails) affirmed
that when the head was completely cut off no regeneration took
place. On the other hand, as many as ten experimenters con-
firmed Spallanzani's results. Bonnet championed Spallanzani's
cause, and the weight of his authority did much toward the
acceptation of the latter's results.

Similar discussions have arisen from time to time in regard
to the power of regeneration of many other forms, and the
fact is of interest as showing that the conditions of the ex-
periment, as well as the condition of the animal, are of im-
portance in deciding whether or not regeneration will take
place.

The majority of naturalists, at the time of which I have just
spoken, were content to wonder at the results ; a few, however,
took a philosophic interest in the phenomena, and attempted
to bring them into line with the current theories of genera-
tion. The speculations of Spallanzani and of Bonnet are of
great interest, and not dissimilar from certain modern attempts
to explain the process of regeneration by the preformation
theory.

With the development of our modern technique a new interest
was awakened in the phenomenon of regeneration, particularly
in connection with the cellular and germ-layer theories. Nu-
merous attempts have been made to discover the origin of the
new cells in the restored part, and to determine whether the
newly formed organs arise from the same germ-layers that give
rise to similar organs of the embryo. It is instructive to ob-
serve how often the facts have been interpreted so as to bring

SOME PROBLEMS OF REGENERATION. 195

them into line with the results of embryology because of an
assumed necessity of the so-called germ-layer theory that ecto-
derm, endoderm, and mesoderm always give rise to the same
layers and organs. Recent years have witnessed the decline
of the germ-layer theory, and numerous discoveries in budding
and dividing animals as well as in regeneration of new parts
have shown that the germ-layers do not in some cases play the
part assigned to them. The overthrow of the germ-layer dogma
has done much to prepare the way for a broader conception of
development and cell-differentiation. If I mistake not, there is
a tendency at present, that is slowly gaining ground, to give
up as unprofitable the interpretation of regenerative and even
embryological phenomena in terms of speculative phylogeny.

A large body of information has now been gathered in regard
to the power of regeneration in the plant and animal kingdoms,
and the time has come, I believe, when we can profitably cease
from indiscriminate mutilations, and when, after an examination
of the known facts, we can ask ourselves what are the real
problems that are presented to us in the regeneration of living
things. We can then hope to direct our experiments to some
definite purpose — toward the solution, if solution be possible,
of the main problem. It must be admitted that well-directed
attempts often prove futile, while accidental discoveries are
sometimes of the greatest importance. Nevertheless, the solu-
tion of one well-propounded problem may fully repay the
investigator for a dozen failures. The failures are due, no
doubt, not to the capriciousness of nature, but to our inability
to grasp all of the conditions within and without the animal or
plant, and also in no small part to our failure to appreciate what
questions we can expect to be answered and what not ; for
lacking, as we do, all fundamental knowledge of the phenom-
ena that we call life, it is not surprising that we find ourselves
so often foiled.

My present object is to examine some of the problems of
regeneration that are at present before us, and then to see if
we can get any clearer idea of the main problem of regenera-
tion. Under the following headings I have arranged as many
problems.

196 BIOLOGICAL LECTURES.

I.

Numerous experiments on Protozoa and Protophyta have
shown that a nucleated part is capable of forming a new indi-
vidual. So far as we can see there is not in most cases, perhaps
in none, the formation of new indifferent protoplasm in which
the new parts are developed, but the entire piece is changed
over into a complete animal or plant of smaller size. At first
sight there seems to be here a marked difference between the
regeneration of unicellular and multicellular forms, for in the
latter it is usual for a knob of new tissue to appear, and out
of this the new part develops. However, Trembley saw that
when a Hydra is cut longitudinally the cut edges bend in and
fuse, forming a new tube of smaller diameter. Nussbaum has
also observed in Hydra the rolling in and fusion of the cut
edges. In both cases the new form develops without the
previous formation of new tissue. In a tubularian hydroid
Bickford has found that when a piece is cut from the stem the
new tentacles appear in the old tissue, and I shall describe
more fully below the results of some experiments on planarians
which show that the old part plays an important r61e in the
formation of a new individual. We see then that the difference
between unicellular and multicellular forms is not so great as
appears at first sight. The special problem that we meet with
here is this : How far are the processes of regeneration in uni-
cellular and multicellular forms the same } Can we compare
with the Protozoa only those multicellular forms in which
regeneration takes place directly in the old parts } If so, is the
relation between the cells only one of correlation, as Hertwig
has recently urged for all multicellular forms ; or are we still
prejudiced by the cell-dogma t Do not these cases of regenera-
tion in the multicellular form indicate that the individual is a
whole in the same sense that the unicellular form is a whole }

II.

In the regeneration of the lower animals and lower plants we
do not find any differences that need detain us, but a curious

SOME PROBLEMS OF REGENERATION. 197

difference is found between the higher animals and plants ; for
instance, in vertebrates and flowering plants. In the former,
regeneration takes place by the formation of a knob of indiffer-
ent cells and subsequent differentiation. In the higher plants,
on the contrary, regeneration of this kind is extremely rare
(root-tips). The new part does not appear at the cut end, but
the injury acts as a stimulus to the development of buds more
or less distant from the exposed surface. The buds that, develop
may be in some cases already formed, but in other cases new
buds appear after the injury, and in places where there were
no buds before the injury. Why do we find this difference in
the renewal of organs between the higher animals and plants .-*
The usual reply is that plants do not regenerate at the cut end,
because they have acquired a new method of replacing lost
parts. The answer is not only teleological, but a quibble, so
long as it cannot be shown that there is a connection between
the development of buds and regeneration at the cut ends. If,
for instance, the new buds be continually destroyed as soon as
they appear, will the plant then regenerate at the cut surface }
Can we imagine that, were it possible to destroy or injure the
buds for many generations, the plant would then "acquire" a
power to regenerate at the cut end t If, on the other hand, it
is assumed that the plant has lost its power of regeneration at
the cut end, because it acquired a new method of renewal of
lost parts, the argument may be met by the fact that certain
hydroids regenerate both at the cut end of a stem, as Loeb has
shown, and also along the stem itself. There is, so far as we
can see, no contradiction between development from the cut
surface and development by means of buds at other points.
Our immediate problem is to examine the conditions of regener-
ation in the plant itself, to see if we can discover why regeneration
takes place at some distance from the exposed surface.

III.

It is not unusual to find that development will take place
in one direction and not in another. A newt that has had a
leg cut off will regenerate the lost limb (Spallanzani, Bonnet,

198 BIOLOGICAL LECTURES.

Barfurth, etc.), but the limb does not make a new newt ;
because, one may say, it dies before it can bring about the
development. But if it were grafted on to the body of the
same or of another newt by its distal end, so that its proximal
end is exposed, would a new newt bud out from that end }
We should certainly be surprised if this happened. The experi-
ment has not been tried with the newt, but a not dissimilar
experiment has been made with the rat by Bert. The tail was
bent over and fixed beneath the skin of the back of the animal,
and after it had grown to that part it was cut off at its root.
The exposed end of the transplanted tail did not — so far as the
experimenter has recorded — give rise to a new rat.^

The same problem is met with in the regeneration of the
earthworm. If a few of the anterior segments (one to five) be
cut off, the same number come back ; if more are cut off, the
process of regeneration begins only after a longer interval, and
only four or five segments come back, as a rule ; if the cut be
posterior to the middle, the time before regeneration begins is
still longer, and fewer worms succeed in regenerating at all;
and still nearer to the posterior end there is no regeneration
anteriorly. In the same worms we find that the posterior end
will not regenerate from a small anterior piece, although the
piece will regenerate anteriorly if a few of its anterior segments
are also cut off. The latter result shows most strikingly that
it is not due to the size of the piece that regeneration does not
take place posteriorly. We see, therefore, that the posterior
end of the body can regenerate in one direction only, namely,
posteriorly, and conversely the anterior end can only regenerate
anteriorly. There is no structural peculiarity, such as differ-
entiation of the cells, that can account for the facts. What,
then, is the explanation of this curious phenomenon } Is there
not here a problem that we can hope to solve by means of
further experiment }

1 The experiment is not, perhaps, altogether to the point, since the rat does not
regenerate a new tail.

SOME PROBLEMS OF REGENERATION. 199

IV.

An examination of the literature of regeneration will show
that there is a widespread dogma that the power of regenera-
tion of a part is commensurate with its Hability to injury. I
have been able to trace this belief back to Bonnet, who says
that it is possible that the number of times a worm can regen-
erate a new part is in proportion to the number of accidents to
which it is exposed during its life. Bonnet found it convenient
to propound this statement in order that the results of the
experiments on regeneration might be fitted into the inelastic
preformation theory.

The modern preformationists avoid the difficulty in other
ways, as rtiay be seen in Weismann's book on the Germ-
plasm.

What is even more surprising is the fact that modern
Darwinians pretend to be able to "explain" the process of
regeneration as the result of natural selection. None of the
writers have, so far as I know, taken the trouble to tell us
how an indefinite series of injuries to a part could at last make
that part "acquire" a power to regenerate in succeeding gen-
erations. Darwin does not discuss the point, and his more
enthusiastic followers can only state that those animals are
" fittest " that regenerate. They fail to see that even on their
own assumption regeneration often takes place without relation
to the survival or non-survival of the animal, and at best the
only possible statement that can be made is that all forms have
died that have not had the power to regenerate, and this admis-
sion would show only too clearly that they cannot pretend to
offer any "explanation" of the process itself. Such a state-
ment is, moreover, in direct contradiction to a large number of
known facts, for all animals alive at the present time do not
possess this power of regeneration. It would surely be as use-
ful to man to be able to reproduce a new arm or leg as it is for
a salamander to regenerate its limb or for an earthworm its
tail ; and I cannot but feel assured that mankind will never
acquire the property of replacing his arms and legs by a minutely
graduated series of injuries combined with an inbreeding of the

200 BIOLOGICAL LECTURES.

unfortunate individuals subjected to the experiment. I need
not here recall numerous facts in regard to regeneration of
internal organs that can seldom or never (except in artificial
operations) be exposed to injury.

Last year I undertook a series of experiments on the regen-
eration of the appendages of the hermit crab, in order to deter-
mine whether regeneration takes place more readily in the
appendages exposed to injury than in those protected by the
shell. The results show that although the first three legs are
often lost by the animals, yet they do not regenerate more
readily than do the other appendages protected by the shell, and
some of which at least can scarcely ever be lost or injured.

The question may be asked : Do we find regeneration more
common in animals and plants that are most exposed "to danger }
If, as several naturalists believe, this is true, what explanation
may be offered } The most frequent " explanation " is that in
some way this is the result of natural selection, but it may be also
"explained" in another way ; namely, only those animals that had
from the beginning the power of complete regeneration of a lost
part can exist where the danger exists. In other words, only
these forms can extend their area of distribution over places
dangerous to their existence. At present, however, we lack
the data to show that regeneration is really more frequent in
animals and plants that are most exposed.

In both unicellular and multicellular forms it has been found
that below a certain limit of sijze a part fails to regenerate,
although in some cases the small piece may live to a time when
larger pieces have begun or completed their regeneration. The
same result has been obtained for nucleated fragments of the
unsegmented ^g^. Do we meet here with the same problem
mentioned above, where a leg does not give rise to a newt,
although a newt gives rise to a leg, or are the two problems
quite different t I am inclined to think that they are different,
because the small pieces may come from regions of the body
that are capable of regeneration, and if the small pieces are

SOME PROBLEMS OF REGENERATION. 20I

grafted on to larger ones they may then regenerate, as Joest
has shown for the earthworm.

There is a curious fact in regard to the development of small
pieces, namely, that they do not give rise, in some cases at
least, to the complete number of organs characteristic of the
form. For instance, Miss Peebles has found that a small piece
of Hydra forms only one or two tentacles. Why does this
reduction in number of organs take place .-*

VI.

We come now to some of the most important problems of
regeneration — the external and internal conditions that deter-
mine whether or not a part will regenerate. The little that
we know in regard to the effect produced by external factors
is almost entirely due to Loeb's important experiments. A
hydroid Eudendriwn failed to develop new heads when kept in
the dark, but when brought into the light the new heads quickly
appeared. Here it seems that light, in some unknown way, acts
as a stimulus necessary for the growth of the new head. In
another hydroid — Tzibularia — Loeb has shown that instead
of a head a root will develop at the distal end of a piece if that
end be brought into contact with some fixed object, and con-
versely a new head will appear at either the proximal or distal
end if the end be freely surrounded by water. Here also we must
look upon the external agent as a stimulus that determines the
differentiation of the part. At present we can form scarcely
any idea of what this relation is between the external agent and
the organism. The vital question is, whether we can extend
our experiment so as to determine what this relation is, or have
we here a fact beyond which we cannot hope to go .-*

Loeb has also described some further experiments on an-
other hydroid, Antenmilaria. A piece cut from the stem and
suspended vertically in the water will develop a new stem at
the upper end and roots at the lower end, regardless of whether
the upper end corresponds to the distal or proximal end of the
' original form. Here gravity alone determines that the upper
end shall grow into a new stem and the lower end into a new

202 BIOLOGICAL LECTURES.

root. In the one case, where the upper end corresponds to the
proximal end of the original hydroid, the new part (a stem)
replaces the lost root end, and this change Loeb calls heteromor-
phosis. In the other case, where the upper end corresponds
to the distal end of the original hydroid, the new part (a stem)
replaces the lost part of the stem. This is what is usually
meant by regeneration. It would be most instructive to find
out what pieces of this hydroid would do were they fixed to a
revolving wheel so arranged that the direction in which gravity
acts on the piece would change at every instant.

It is of the greatest importance to note that in all these cases
described by Loeb, in which regeneration is determined by
external circumstances, the same influences affect the growth
of the uninjured hydroids in the same way. Only those forms
whose normal growth is influenced by light, or by gravity, or
by contact, could be induced to respond to the same influences
during the time of regeneration.

In review of the few facts that we possess in regard to the
effect of external factors, we must bear in mind that in some
cases the external agent, gravity for instance, does not seem in
itself necessary for the growth of the animal, but it acts in such
a way that it determines what sort of differentiation may take
place. Other factors, temperature for example, determine only
whether or not growth can take place at all, and if so, the
rapidity of the growth. The salts contained in the water act
in the same way, in so far as they are necessary for the metab-
olism of the cells. In so far as the salts affect the osmotic
condition of the cells they seem to determine — as Loeb has
shown for these hydroids — the amount of growth, but not its
kind. It may be that other factors, light for instance, may
affect not only the amount, but also the kind of growth.

In this same connection I should like to describe an ingenious
experiment made by Herbst. He found that when the eyes
of a prawn were cut off, sometimes an eye and sometimes an
antenna regenerated at the cut surface. He then tried the
experiment of placing an equal number of new individuals —
after cutting off the eyes — in the light and in the dark, to see
if the light would act as a stimulus on the forms in the light, and

SOME PROBLEMS OF REGENERATION. 203

thus cause an eye to appear. It was also conceivable that in
the absence of light an antenna might develop, since an eye
would be of no value to an animal in the dark, while an antenna
would be a useful organ. But these results did not follow, since
eyes and antennae appeared in equal numbers in the animals in
the dark and in the light. In the hermit crab also I have found
that sometimes eyes and sometimes antennae appear after the
eye-stalk is cut off, and I have discovered that when only the tip
of the eye-stalk is cut off a new eye reappears, but when the
eye-stalk is cut off near its base an antenna-like organ develops.
In this case it seems that the factor that determines whether an
eye or an antenna develops comes from within and not from
without. We find here a process of heteromorphosis depend-
ing, so far as we can see, on internal factors. Does this kind of
heteromorphosis belong to the same class of phenomena shown
by the hydroids .? In the latter the stimulus that determines
the kind of regeneration comes from outside the animal.

VII.

It is of some theoretical interest to know whether the old
cells form directly the new tissue, or whether reserve cells are
present that bring about the result. From the work already
done it seems probable that this may vary with different forms.
In some cases, as Randolph has shown for Lumbriculus, reserve
cells are present that assist in the regeneration of the meso-
derm ; in other cases, as Hescheler has shown for the earth-
worm, the old cells seem to give rise to the new ones. In
Ttibtdaria, as Bickford has shown, the old cells go directly
over into the new tissue.

Not only do the old cells give rise to new ones, but in the case
of Planaria maailata the old part may itself change its form as
a whole, and in this way play an important part in the develop-
ment of the new worm. There is an extensive remoulding of
the old tissue, so that the new parts form a miniature copy of
the adult worm. It is difficult, even impossible, to form any
idea of what internal changes are taking place during this
period of change, and our ideas of reorganization and of correla-

204

BIOLOGICAL LECTURES.

tion of parts are entirely inadequate at present to make clear
the process. More elaborate series of experiments must be
undertaken before we can even attempt to make an analysis of
these results.

VIII.

There are certain problems in connection with the relation
between the old parts and the new that need careful study.
I have already touched on some of these points in the preceding
pages. The limb of a newt, for instance, is replaced by a new
one precisely like the old. If only the hand is cut off, just this
much is reproduced ; if the forearm and the hand are both cut
off, then a new forearm and hand reappear ; and this is true at
whatever level the cut is made. In other forms less (or even
more) reappears, under certain circumstances, than was removed;
but as yet our experiments have not been carried far enough to
make clear how this result is related to the renewal of only as
much as removed.

In the earthworm when one, two, three, four, or five anterior
segments are cut off, the same number come back ; but when
more than five are cut off, only five are regenerated. In
Liimbricidiis} also, the number of new head segments is never
more than seven or eight, even when a much greater number
have been cut off. In Planaria a new head forms on the
anterior end of a posterior half of the worm, but the entire
anterior half is never replaced by new tissue. In this same
planarian Randolph has discovered a most important relation
existing between the old and new parts. If a planarian be
cut in two longitudinally in the median plane, the right half
regenerates a new left half of the same size as the part removed,
and the left half also develops a new right half of correspond-
ing size. If, however, the worm be cut longitudinally into a
larger and a smaller strip, the former replaces as much as con-
tained in the smaller part that was removed, but the smaller
part does not develop the lost larger part, but forms only as
much new tissue at its cut side as is about equivalent to its own
breadth.

1 An American species.

SOME PROBLEMS OF REGENERATION. 205

When the posterior end of an earthworm is cut off, there
appears at first at the posterior end of the anterior part a small
knob, out of which a few segments and a growing end develop.
The latter continues to extend backward and add new segments.
After a time the posterior growth comes to an end, i.e., when the
full complement of segments has been reached. If, however,
instead of the entire anterior end we take a small piece from
the middle of the worm, we find that even this piece only forms
as many new posterior segments as lay originally behind that
piece in the worm. In other words, the growing region ceases
to give rise to new segments after those that were lost have
been replaced, so that the entire number of segments in the
! new worm may be far below the normal number. It seems to
me that we see here a factor at work that we cannot in the
least grasp. To call it correlation is of no avail unless we
can define what we mean by the term. The usual definition of
correlation is only a restatement of the facts in different words,
and not a causal explanation of the facts.

It is not difficult to show how much we must include in our
definition of correlation if we use the word to cover all the
known facts. According to Blumenbach (1784), the eye of
I Triton will regenerate if only a small part of the original eye be
' left. If the anterior end of the earthworm be cut off very
obliquely, the pointed end will bend over the cut surface and
fuse with it, and slowly a new anterior end will develop between
the bent-over portion and the anterior cut surface. In both
these cases the relation of the new part to the old must be an
extraordinarily complex one, and yet a new organ develops like
the one lost. Again, pieces may be cut from the Qggy and it
will still, in some cases, produce a whole animal of smaller
size — a part developing a new whole much in the same way
that a part of a protozoon may form a new animal of smaller
size. Something more is included in these phenomena, I think,
than can be explained by simple physical interaction or by
chemical influences. The process that takes place suggests
that something like an intelligent process must be at work — I
mean that what we call correlation of the parts seems here
to belong rather to the category of phenomena that we call

2o6 BIOLOGICAL LECTURES.

intelligent, than to physical and chemical processes as known
in the physical sciences. The action seems, however, to be
intelligent only so far as concerns the internal relations of the
part, i.e., it acts rather as a *' perfecting principle" than as a
process of adaptation to external needs (adaptation).

IX.

Finally, I wish to touch briefly on a few points that seem to
me to bring us nearer to the heart of the problem. In much of
our biological work we have been guided by methods derived
from the physical sciences, and most fortunately so, for per-
haps only in this way can we hope to reduce living phenomena
to simpler terms. But sooner or later we meet with a factor
that defies further physical analysis, and this factor seems to
be present in all biological phenomena. We gain nothing by
calling it a vital force, unless we can define what we mean by
vitality. Whether or not this factor ^ is only a complex of
physical forces that we cannot unravel, or whether there exists
something that cannot be expressed in terms of physics and
chemistry — that is the question!

We err, I think, in going at present to either extreme, i.e.,
either in ignoring this something that has been called a vital
force and pretending that physics and chemistry will soon
make everything clear, or, on the other hand, in calling the
unknown a vital force and pretending to explain results as the
outcome of its action.

In our studies of the development of form we meet most often
with this factor. Are we at' bottom trying to give a causal
explanation of form itself, and, if so, is not our problem insol-
uble } Can we hope to do more than determine under what
internal and external conditions a given form appears } If we
limit our researches to this problem we can hope to succeed.
But can we go back of this and explain the reaction itself t At
present we have not succeeded in doing so, any more than has
the mineralogist explained the form of a crystal. It may be

1 It is simpler to speak of it as one factor, but it may equally well be true that
there are many factors.

SOME PROBLEMS OF REGENERATION. 207

that what we call a formative force or a vital force is the prop-
erty of living things to assume a given form under certain con-
ditions. If so, is there here legitimate ground for investigation,
or rather let me ask, can we hope to extend our investigations
beyond the knowledge of the internal and external conditions
within which new forms arise. It is this uncertainty in regard
to the problem of vitality that we need first to clear up, and it
seems to me that this is the cardinal point for us to examine at
present. It is possible, I think, by means of experiment alone,
to determine how far and in what sense we can pursue the
investigation of the causes of form. In this regard experi-
mental studies on the regeneration of animals and plants offer
a most admirable field for future work.

ELEVENTH LECTURE.

THE ELIMINATION OF THE UNFIT AS ILLUS-
TRATED BY THE INTRODUCED SPARROW,
PASSER DOMESTICUS.

{A FOURTH CONTRIBUTION TO THE STUDY OF VARIATION.)
HERMON C. BUMPUS.

We are so in the habit of referring carelessly to the process
of natural selection, and of invoking its aid whenever some pet
theory seems a little feeble, that we forget we are really using
a hypothesis that still remains unproved, and that specific
examples of the destruction of animals of known physical dis-
ability are very infrequent. Even if the theory of natural
selection were as firmly established as Newton's theory of the
attraction of gravity, scientific method would still require fre-
quent examination of its claims, and scientific honesty should
welcome such examination and insist on its thoroughness.

A possible instance of the operation of natural selection,
through the process of the elimination of the unfit, was brought
to our notice on February i of the present year (1898), when,
after an uncommonly severe storm of snow, rain, and sleet, a
number of English sparrows were brought to the Anatomical
Laboratory of Brown University. Seventy-two of these birds
revived ; sixty-four perished ; and it is the purpose of this
lecture to show that the birds which perished, perished not
through accident, but because they were physically disquali-
fied, and that the birds which survived, survived because
they possessed certain physical characters. These characters
enabled them to withstand the intensity of this particular
phase of selective elimination, and distinguish them from

209

2IO BIOLOGICAL LECTURES.

their more unfortunate companions. It will be convenient
for us to arrange our material in the form of tests, as follows.

Test I : Sex. — It will be noted by reference to the tables that
of the surviving birds the males are much more numerous than
the females. Of the former there are fifty-one (thirty-five adults
and sixteen young), while of the latter there are only twenty-
one. Among the birds which perished, the females are abso-
lutely and relatively more numerous than they are among the
birds which survived, although more than one-half (thirty-six
out of sixty-four) of the unfortunate birds are males. Of course
it may be that male birds are naturally more abundant than
females, but the present question is not one of distribution of
sex, but rather of distribution of fitness, and the inference is that
the females are less competent to resist severe winter weather
than are the males, for, while only 28^ of the survivors are
females, they constitute 43^ of those that perished.

Test 2 : Length. — The first column of figures on the several
tables gives, in millimeters, the lengths of the birds from the
tip of the beak to the tip of the tail. An examination of the
averages, printed at the bottom of each column, will prove par-
ticularly instructive. It will be noted on Tables I and I^ that
the average length of the adult males which survived (i 59 mm.)
is really less than that of the adult males which perished
(162 mm.).i Similar figures, 159 mm. and 162 mm. on Tables
II and 11^ indicate the same relative lengths of the young
males of the two groups. The average lengths of the females
of the two groups, 157 mm. and 158 mm., Tables III and 111%
also indicate an excess in the average length of the birds
which perished. The birds which perished, then, males or
females, adult or young, are longer than those which endured,
and we are justified in concluding that when nature selects,
through the agency of winter storms of this particular kind of
severity, those sparrows which are relatively short stand a
better chance of surviving.

Test J: Alar Extent. — AvQVdigQS, based on measurements
from tip to tip of the extended wings fail to bring out any

1 The numbers printed in light type, both in the text and in the tables, refer to
birds which survived ; those printed in heavy type refer to birds which perished.

THE ELIMINATION OF THE UNFIT. 211

striking difference between the two classes of birds. Both
have an indicated average of 2.45 mm., although, to be more
exact, the birds which perished averaged 2.449, while those
that survived averaged 2.455, 2. difference too slight to be of
material significance. This similarity of the two groups is
not to be wondered at, since it is not to be expected that one
eliminative agent will express itself in all possible anatomical
features. Were the eliminative agent, for example, a severe
northerly wind of protracted duration, the alar extent might
then enter in as a factor of considerable selective value, and
survivors would then have an alar extent materially different
from that of the birds eliminated.

The alar extent of the females, corresponding with their
smaller size, is less than that of the males.

Test 4. : Weight. — Had I been called upon to express an
opinion as to whether heavy or light birds would be more suc-
cessful in resisting the severity of the February storm, I should
have declared unhesitatingly in favor of the heavy birds. An
examination of the third column of measurements, however, will
show that the birds which survived invariably average less in
weight than those which perished, and that among the males
this difference amounts to more than a gram ; that is, to about
one twenty-fifth of the weight. The surviving birds of both
sexes had an average weight of 25.2 grams, and those which
succumbed had an average weight of 25.8 grams.

It may not be out of place to call attention here to certain
objections which may be raised to the method which I have
adopted, and to the conclusions thus far derived therefrom.
One may claim that the greater relative number of females in
the group of birds which perished vitiates the numerical result,
since the females are of less stature than the males. But it
will be noted that this objection answers itself, for the birds
which perished are not shorter, but longer, than those which
survived ; and again, that the birds which perished, though
having a disproportionate number of the lighter sex, never-
theless have an average weight considerably greater than that
of the birds which survived. Moreover, comparing, in the
two groups, adult males with adult males, young males with

212 BIOLOGICAL LECTURES.

young males, and females with females, we find that the
differences between the two classes of birds are expressed
in these three smaller divisions, and I think we are justified in
concluding that the differences are really significant.

The explanation that the birds which lived were those which
sought, or at least enjoyed, better shelter cannot be entertained,
for the storm was of long duration, and the birds were picked
up, not in one locality, but in several localities ; and, moreover,
it is a fact that the survivors are structurally differeiit from
those which perished. If to these structural characters one
desires to add also the intellectual character that the birds knew
enough to go in out of the storm, the difference between the
two groups becomes so much the greater.

Test 5.- Length of Head. — A comparison of the average
lengths of head, from the tip of the beak to the occiput, shows
only a similarity between the survivors and those which per-
ished, and indicates that under the present environmental
conditions this feature is not sufficiently prominent to be
expressed by this method of computation.

Test 6: Length of Humerus. — An examination of the fifth
column of figures will show that the length of the arm bones
of the birds which perished always averages less than that of
the survivors. This difference is most conspicuous in the adult
males, where the surviving birds have an average length of
humerus of .738 of an inch, considerably more than that of
their unfortunate companions, .727.

Here again I wish to emphasize the fact that these differ-
ences cannot be merely accidental, because they so often tend
in the same direction. If among the survivors it is the proper
thing for adult males to have a long humerus, then the young
males have a long humerus, and the females follow the prevail-
ing fashion with characteristic servitude. If a short humerus
is an index of inferiority, all three groups of eliminated birds
(adult males, young males, and females) bear this same mark of
inferiority. This fact is the more striking since the averages
are established on a relatively small number of birds, while
usually in the statistical methods of the study of variation an
abundance of material is necessary.

THE ELIMINATION OF THE UNFIT. 213

Test J : Length of Femin^. — An examination of the general
averages on Tables III and III^ shows that the survivors pos-
sess longer thigh bones than do the birds which succumbed.
The average length of femur in the former is .716 inch ; in the
latter .709. This difference in the averages cannot be ascribed
to the large number of dead females, since the difference pre-
vails also for both the adult and young males.

Test 8 : Length of Tibio-Tarsiis. — Measurement of the tibio-
tarsus yields practically the same comparative data as the
measurement of the upper bone of the leg, although in both
groups of birds this bone in the females is considerably longer
than in the adult males, notwithstanding that the females are
smaller. This series of measurements agrees with the sixth, in
that the young males have longer legs than the adult males.

Test g. — Measurements across the skull, from the postorbital
bone of one side to the postorbital bone of the other, are given
in the eighth column, and are less satisfactory, perhaps, than
those of other portions of the skeleton. The breadth of the
cranium, as thus indicated, is somewhat less in the females than
in the males. The averages denote that the birds which sur-
vived had wider heads than those which perished, but these
averages are considerably influenced by data furnished by the
young males. The irregularities in the subordinate groups
induce me to place less confidence in these numerical results
than in the results from measurements of other structures.

Test 10 : Lejtgth of Sternum. — This test differs from other
tests in that it relates to measurements in the longitudinal
axis of the body. In the males the sternum is long, and in
the females it is short. In the birds which survived it has a
general average length of .845 inch ; in those which perished
it has a general average length of only .834.

I think these tests prove that there are fundamental dif-
ferences between the birds which survived and those which
perished. While the former are shorter and weigh less {i.e.,
are of smaller body), they have longer wing bones, longer legs,
longer sternums, and greater brain capacity. These characters
are in accordance with our ideas of physical fitness ; their defec-

214 BIOLOGICAL LECTURES.

tive development is evidently a mark of inferiority, and we are
justified in concluding that the birds so handicapped failed to
pass one of Nature's rigorous tests and perished.

In an earlier lecture, on the "• Variations and Mutations of
the Introduced Sparrow," facts were adduced which, it was
claimed, were sufficient to show that the English sparrow, since
its introduction into this country, has found life so easy that the
operation of natural selection has been practically suspended,
and that the American type consequently has become degen-
erate. No active agent had eliminated anomalies, and certain,
"freaks " had increased in number, until they had become over
four times as numerous as in England.

When calling attention to the occurrence of these variations,
and to the fact that they were an indication of the absence of
an active eliminative factor, I little thought that within a few
months I might witness the action of an eliminating factor that
would test the structural qualifications of all the birds : destroy
those which had departed unduly from the ideal type, and thus
raise the general standard of excellence.

It will be recalled that, after the storm of February i, one
hundred and thirty-six birds were taken, and that, of these,
seventy-two revived, while sixty-four failed to recover. But
the fact that the birds which perished had in the average
longer and larger bodies, and shorter head, wing, and leg
bones, does not tell all the story of selective elimination.

Reference to the tables will show, not only that the longest
bird perished, but also that the shortest bird perished. The
longest bird was No. 33, the shortest No. 40. (In these and
other cases of extreme departure from the mean, the exponent i
is placed in the table beside the number of the bird.)

Again, if we examine the columns of figures which indicate
the alar extent of the different birds, we find that both the bird
with greatest spread of wings, No. 32, and the one with least
spread of wings. No. 52, perished.

The heaviest bird. No. 23, weighed 31 grams; it perished.
The honors for lightness are evenly divided; No. 53, among
the survivors, and No. 60, among the eliminated, have the
same weight, viz.y 22.6 grams.

THE ELIMINATION OF THE UNFIT. 215

The bird (No, 55) whose head was longest (measured from
the tip of the beak to the occiput) suffered elimination. The
extreme variant in the opposite direction (No. 9) survived.

The honors for the longest humerus, .780 mm., are divided,
Nos. 6 and 44. The bird with the shortest humerus, No. 21,
perished.

The longest femur was possessed by bird No. 55, the shortest
by No. 5 1 .

The surviving birds represent both extremes of variation of
the tibio-tarsus (Nos. 18 and 41). In respect to all other col-
umns of measurements the survivors possess exclusively never
more than one of the extreme forms.

Both extremes of variation in width of cranium (Nos. 55
and 52) are found among the eliminated birds.

The longest sternum is found in one of the surviving birds
(No. 15), and it will be remembered that a long sternum was
considered a mark of excellence. The shortest sternum (No. 52)
is found among the eliminated birds where the standard is low.

These extremes of variation are represented on Table IV,
and by counting the dark numbers we find that eleven extreme
positions (maximum or minimum) are occupied exclusively by
the birds which perished, whereas the light numbers show that
only five extreme positions are occupied by those which sur-
vived. In two cases (the minimum weight and the maximum
length of humerus) the extreme positions are occupied alike
by birds of both groups, and consequently I have left the
spaces blank. In three cases two birds of the same group
occupy the same extreme position, but the table is designed to
indicate only the extreme positions and not the number of birds
occupying them. The munber of birds occupying these extreme
positions is represented on the previous tables by the exponent
I, and if we count up these exponents, we shall find that among
the surviving birds there are nine cases of this extreme type,
whereas among the birds which perished there are fourteen
cases. These numbers are the more impressive when one con-
siders that, inasmuch as there are seventy-two of the former
birds and only sixty-four of the latter, the chances for the
occurrence of extreme variation are not equal in the two groups.

2 10 BIOLOGICAL LECTURES.

The birds which perished are at a decided disadvantage because
of their smaller representation, yet there are many more ''freaks"
among them than among the surviving birds.

If it is thought that the association of the larger number of
extreme variants with the eliminated birds is merely a matter
of accident, we will not stop to argue the matter, but will apply
the same test to the birds that remain after these extreme
examples have been removed. We find even after the removal
of these twenty-three examples, that extreme examples of the
second order, indicated by the exponent 2, show the same tend-
ency to occur more frequently among the eliminated birds.

The longest birds now are i66 mm.; the shortest, 153 mm.
Of the former, Nos. 22, 24, 28, 32, 35, perished, and No. 18 sur-
vived ; of the latter, Nos. 45 and 62 perished, and Nos. 35, 54,
and 55 survived.

If we count the times that the exponent 2 occurs in the
tables, we shall find that there are ten birds of extreme abnor-
mality of this second grade which survived, while there are
twenty of the same grade which perished.

These figures indicate that the amplitude of variation of the
surviving birds is less than that of the birds which perished.
Were we to attempt the arrangement of the data into curves
of distribution, the curve representing the distribution before
the storm would be found to have a broad base, whereas the
curve representing the distribution after the storm would be
found to have a narrow base, for the eliminative process con-
centrated its energy on the individuals which occupied extreme
positions.

Lest there remain some doubt as to the importance of this
eliminative process, and of its efficiency in exterminating ex-
treme variants, let us examine our figures again and see whether
the group of birds which has already contributed thirty-four of
the extremes of variation has still an excess of variability.

If we count up the exponents (3) of this third order of
variable individuals, we find that the birds which survive give
eleven examples, whereas those which perished give twenty-one.
— It appears unnecessary to carry our investigations further
along this line, for our results point always in one direction.

THE ELIMINATION OE THE UNFIT. 21 7

Natural selection is most destructive of those birds which have
depai'ted most from the ideal type, and its activity raises tJie
general standard of excellence by favoring those birds wJiich
approach the struct U7^a I ideal.

Inasmuch as the variation in structure in the birds which
perished tends to centre about certain individuals, as, for
example, Nos. 45, 52, and 55, it might be claimed that the
accidental presence of a few of these extremely abnormal
individuals in this group is what really makes all the differ-
ence. Let us see.

There are twenty-three birds among the seventy-two sur-
vivors whose measurements bear exponents of extreme vari-
ation, and there are twenty-four birds similarly distinguished
among the sixty-four which perished. But none of the birds
in the first group has more than three exceptional features,
whereas several of the birds which perished have a consider-
ably larger number of exceptional features : four, five, and in
one case. No. 52, even six.

Of the twenty-three survivors which bear exponents, nine-
teen have only one exceptional character, and it is not surprising,
considering the high standard of excellence possessed by these
birds as a whole, that a single unfavorable feature does not
prove fatal. There are but ten of the eliminated birds which
have only one exceptional character, and the fact that some
are burdened with more than one is apparently the reason for
their mortality.

In an earlier contribution to the Study of Variation I called
attention to a coincidence which may have considerable signifi-
cance. When specimens of Nectnrns varied in respect to any
one feature, there was a tendency for such specimens to' pre-
sent other and not necessarily correlated variations. Stated in
another way, instability in respect to any one feature is an
index of general organic instability. A similar coincidence of
variations occurs among the sparrows.

Of the one hundred and thirty-six birds, five (Nos. 3, 47, 70,
21, 52) had albino feathers. Like other abnormalities endured
by the surviving birds, albinism in two out of the three cases
is the only affliction. But among those that were eliminated,

2l8 BIOLOGICAL LECTURES.

where albinism twice occurs, it affects in one case a bird
marked by four other abnormalities (No. 21), and in the other
a bird (No. 52) already cursed by six abnormalities, the most
miserable individual in the entire collection.

While we have shown that the birds which perished have
certain average structural peculiarities which distinguish them
from the survivors, and that the intensity of selective elimina-
tion has been felt most by birds of extreme structure, it remains
to be shown that a general instability of structure is as char-
acteristic of the birds which perished as a general stability of
structure is characteristic of those which survived.

If we had sufficient data, this fundamental difference in the
two groups of birds might be indicated by curves of distribu-
tion, one curve narrow and elevated, showing that its compo-
nents are closely crowded around an ideal mean, the other
broad and low, showing that its components are relatively
indifferent to any ideal. But in the absence of sufficient data
to illustrate the differences in this manner, we can arrive at a
numerical result equally instructive by another method.

Having determined the ideal means for the several characters
in each group of birds, we can then find the distance that each
individual departs from this ideal. By adding these degrees
of departure in respect to the several characters, and dividing
by the number of individuals, we shall have numbers which
represent the average departures from the ideal means. These
numbers will be large if the members of a group of birds show
a general tendency towards disregard of the ideals, and they will
be small if the birds tend to crowd around the ideals. If all
the birds actually attain the ideals, the number will be zero.
— This is simply following out the principle that one man at
the end of a ten-foot lever can do as much work as ten men
at the end of a one-foot lever. A bird removed ten units from
the mean exerts the same divergent influence upon its group
that ten birds would exercise if removed one unit.

The results of this test, numerically expressed in Table V,
are most instructive. In every case but one the numbers indi-
cating the average departure from the ideal mean are smaller
for the birds which survived, and thus indicate a general tend-

THE ELIMINATION OF THE UNFIT. 219

ency toward conservatism on the part of the survivors. In
the single exceptional case the numbers are not very different,
32 and 31. Granting this exception to the uniformity in the
figures, it is exceedingly interesting to examine the series. In
respect to length, the birds which perished had an average
departure from the ideal mean expressed by the number 3.48,
while the average departure of the birds which survived was
only 2.51, or, expressed in tabular form:

In respect

to length,

3.48 is

greater

than

2.51.

" alar extent,

4.60 "

4.20.

" weight,

12.6 "

10.9.

" length of head,

5.64 "

S.51.

" " " humerus,

20.1 "

16.0.

" " " femur,

20.0 "

14.0.

" " " tibio-tarsus,

33.8 "

29.4.

" width of head.

12. "

10., but

" length of keel,

31. "

less

32.

A series of eight consecutive cases like the above, all point-
ing in the same direction, can hardly be considered accidental.

To summarize :

(i) We have found that there are fundamental differences
between the surviving birds and those eliminated, and we
conclude that the birds which survived survived because they
possessed certain structural characters, and that the birds which
perished perished not through accident, but because they did
not possess certain structural characters which would have
enabled them to withstand the severity of the test imposed
by nature ; they were eliminated because they were unfit.

(2) The process of selective elimination is most severe with
extremely variable individuals, no matter in what direction the
variations may occur. It is quite as dangerous to be conspicu-
ously above a certain standard of organic excellence as it is to
be conspicuously below the standard. It is the type that nature
favors.

(3) Disregard of structural qualifications finally produces a
throng of degenerates, whose destruction will follow the arrival
of adversity.

220

BIOLOGICAL LECTURES.

Table I.

Measurements of Thirty-five Males which Survived.

ti

1^

a
0

1

Length of

Beak
AND Head.

0 t/i

II

0 .

Length of
Ti bio-
tarsus.

Ex.

length of
Keel of
Sternum.

I ^

154=^

241

24.5

31.2

.687

.668

1.0222

•587

-830

2 ^

i6o

252

26.9

30.8

736

.709

1. 180

.602

.841

3 ^

155

243

26.9

30.6

733

.704

1. 151

.602

.846

4 ^

154^

245

24-3

317

741

.688

1. 146

.584

•839

5 ^

156

247

24.1

31-5

715

.706

1. 129

•575

.821

6 ^

161

253

26.5

31-8

.780I

743

1. 144

.607

•893

7 ^

157

251

24.6

3I-I

.741

736

I-I53

.610

.862

8 Z

159

247

24.2

314

.728

.718

1. 126

.609

•793

9 $

158

247

23.6

29.81

703

•673

1.079

.602

.820

lO $

158

252

26.2

32.

749

739

I -1 53

.614

-857

II ^

160

252

26.2

32-

741

723

I.I 29

.624

.892 .

12 ^

162

253

24.8

32-3

.766

752

I-I34

•633

.923'

13 $

161

243

25-4

31-8

.721

.722

1. 126

•597

.891

14 ^

160

250

237

29.81

730

703

1-103

•590

.820

i^ $

159

247

257

314

729

.717

1. 141

•592

.927I

i6 ^

158

253

257

31-9

743

.699

1-150

.600

.860

17 ^

159

247

26.5

31-6

733

714

1155

.611

•923^

i8 $

1 662

253

26.7

32.5

767

.7652

1.230I

.600

.878

19 ^

159

247

23-9

314

752

723

1. 113

.602

.825

20 $

160

248

247

3^-3

752

737

1. 176

.603

-803

21 ^

161

252

28.

31-8

.7702

731

1. 190

•590

.885

22 ^

163

251

27.9

31-9

769^

745

1. 168

.622

.860

23 $

156

242

25-9

32-

723

711

1. 116

.609

.886

24 ^

1653

251

257

32.2

751

742

1. 161

.613

-865

25 $

160

247

26.6

324

.728

.707

1. 108

•590

.836

26 ^

158

244

23.2^

31-6

730

713

1. 142

-585

.888

27 ^

160

242

25.7

31.6

709

705

1. 124

.620

.788

28 $

157

245

26.3

32.2

741

.726

1143

-595

.850

29 ^

159

244

24-3

31-5

723

.698

1. 107

.615

.847

30 $

160

253

26.7

32.1

739

714

1. 117

•592

.864

31 ^

158

245

24.9

314

.726

703

1. 119

•580

-854

32 ^

161

247

23.8

314

735

.694

I.IOI

.602

.789

33 $

160

247

25.6

32.3

756

745

1-135

.607

.902

34 ^

160

247

27.

32-

755

736

1. 174

.631

.873

35 $

153^

241

247

32.2

.728

.680

1.092

•592

.884

Average . .

159

247

254

31.6

738

.716

1-135

.602

•857

THE ELIMINATION OF THE UNFIT

21

Table I^

Measurements of Twenty-four Adult Males which Perished.

n

n

< X

X

o

Length of

Beak
AND Head.

o (A

II

u.
o .

Length of
Ti bio-
tarsus.

Length of
Keel of
Sternum.

1

t

1663

249

26.5

31.

.738

.704

1.095

.606

.847

2

$

160

245

26.1

32.

.736

.709

1.109

.611

.842

3

$

161

249

25.6

32.3

.743

.718

1.128

.602

.828

4

$

162

246

26.9

32.3

.738

.709

1.136

.607

.869

o

$

163

250

26.5

32.6

.752

.731

1.197

.623

.888

6

$

162

247

27.6

31.8

.731

.719

1.113

.597

.869

1

$

163

246

26.8

31.4

.689

.6623

1.073

.604

.836

8

$

161

246

24.9

30.6

.739

.726

1.138

.580

.803

9

$

160

242

26.

31.

.745

.713

1.106

.600

.803

10

$

162

246

26.5

31.5

.720

.696

1.092

.606

.809

11

$

160

249

26.

31.4

.726

.689

1.097

.602

.850

12

$

161

260

27.1

31.6

.737

.711

1.120

.631

.862

13

$

162

248

26.1

31.9

.744

.722

1.154

.591

.839

14

$

1653

252

26.

32.3

.726

.710

1.146

.609

.887

15

$

161

243

26.6

32.5

.709

.707

1.122

.607

.832

16

$

161

244

25.

31.3

.702

.686

1.082

.595

.874

17

$

162

248

24.6

31.

.713

.700

1.086

.690

.837

18

$

164

244

25.

31.2

.703

.690

1.074

.608

.795

19

$

158

247

26.

32.

.729

.710

1.146

.607

.803

20

$

162

263

28.3

31.8

.752

.718

1.152

.600

.857

21

$

156

239

24.6

30.5

.6591

.6582

1.0423

.6703

.810

22

$

166

261

27.5

31.5

.720

.691

1.118

.612

.847

23

$

1653

253

31. 1

32.4

.766

.750

1.183

.613

.906

24

$

1662

250

28.3

32.4

.754

.718

1.179

.607

.9163

Average . .

162

247

26.2

31.6

.727

.706

1.120

.603

.845

222

BIOLOGICAL LECTURES,

Table II.

Measurements of Sixteen Young Males which Survived.

1!

< H

•J h

W

X

Length of

Beak
AND Head.

b
0 t«

11

Length of

Tl BIO-
TARSUS.

0 .

Length of
Keei. of
Sternum.

36 juv. Z
31 juv. t

38 juv. Z

39 juv. Z

40 juv. Z

41 juv. Z

42 juv. Z

43 juv. Z

44 juv. Z

45 juv. Z

46 juv. ^

47 juv. ^

48 juv. Z

49 juv. ^

50 juv. Z

51 juv. ^

156
156
163
163

irx)
156
162

163
164

163
160
160
158
158
158
155

246

245

248

248

250

237

253

254^

251

244

247
250

247
249
243
237

24.6

25-5
24.8

26.3
24.4

23-3
26.7
26.4
26.9

24.3

27.

26.8

24.9

26.1

26.6

23-3

32.

32.1

32.2

33-^
31-5
30.6

32.

32.

32-

31-3

31-5

32.5

324

32.2

324
30.2*

■7A\
.761
.742
•736
.746
.692

•759
.766

•755
.718
.764
.764
•745
•742
•747
.685

•735
.717

■73Z
.704

•715
.664
•734
•750
.742
.680

•732
.729
.724

•736
.711

•653^

1.167

1. 147

I.I65

1. 148
I.I73

I.OII^

1. 197

1.16s

1. 171
1.082
1. 177
1. 123

i^i39

1. 148

1.163

i.oiii

•592

.620

.606

.609

.604

.588

.630

.605

.620
.610

.617

•635^

.588

.602

.612

•587

.849
.816
•854
•839
•893
•774
.878
.886
.886
.892
.846
.842
.865
.817
.891
•794

Average . .

159

246

254

3.-8

.741

.716

I.I36

.607

.851

THE ELIMINATION OF THE UNFIT.

223

Table. II ».

Measurements of Twelve Young Males which Perished.

26 juv. Z
26JUV. Z

27 juT. $

28 juv. $

29 juv. $
80 juv. $
31 juv. $
82 juv. $

33 juv. $

34 juv. ^

35 juv. $

36 juv. ^

Average

J s

< H

h O

O Z

►J

160

156

158

1662

1653

157

164

1662

1671

161

1662

161

162

< X

249

236

240

245

2552

238

250

2561

2552

246

2543

251

24.2

26.8

23.5

26.9

28.6

24.7

27.3

25.7

29. 3

25.

27.5

26.

248 26.2 31.5

< X
w ^

P3 Q

30.4

80. 2

31.

31.7

31.5

31.2

31.8

31.7

32.2

31.5

31.4

31.5

.740

.690

.715

.715

.766

.6803

.764

.752

.765

.739

.760

.731

.734

.717
.671
.702
.695
.744
.677
.726
.751
.745
.707
.742
.707

.715

s 6 S
S 2 S
I ^ ^

1.130
1.067
1.113
1.107
1.175
1.156
1.171
1.187
1.197
1.123
1.124
1.122

1.141

.620

.5632

.595

.601

.613

.599

.588

.595

.6382

.587

.604

.589

.599

.840
.832
.805

.847
.854
.769
.860

.858
.855
.850
.914

.828

.842

224

BIOLOGICAL LECTURES.

Table III.

Measurements of Twenty-one Adult and Young Females which Survived.

X

Length of

Beak
AND Head.

si

ii

Length of

Ti BIO-
TARSUS.

X J

i. i

Length of
Keel of

Sternum.

52 9

53 9

54 9

55 9

56 9

57 9

58 9

59 9

60 9

61 9

62 9

63 9

64 9

65 9

66 9

67 9

68 9

69 9

70 9

71 9

72 9

156

1543

153^

153^

155

163

157

155

164

158

158

160

161

157

157

156

158

153^

155

163

159

245
240
240
236
243
247
238

239

248

238
240
244
246
245
235
237
244
238
236
246
236

25-3

22.61

25.1

23.23

24.4

25.1

24.6

24.

24.2

24.9

24.1

24.

26.

24.9

25-5
234
259
24.2
24.2
27.4
24.

31.6

304

31-

30-9

31-5

32.

30-9
32.8

32.7
31-

31-1

32.3

32.

31-5

30-9

314

30.5

30-3

32-5

31-5

.729

•705
.724
.698
•734
.748
.726
•732
•752
.741

•733

•758
•752
.712
.708
.729

•715
.727

•732
.709

.710

.686

•713
.678

.736
.734

•727
.742

•752
.689
.706
•730
•732
.740
.704
.691
.705
.707
.705
.711
•713

1. 152
I^I03
I^I23

1-132
1. 170
1. 166

1. 175
I.I75

1. 201
1. 09 1
1. 107
1-152
1-154
1. 186
1. 132
1. 123
1. 146
1. 116
1. 1 20
I.163
I.I 29

620
584
585
596
596
602
588
601
604
592
591
589
623

593
611

613
597
595
585
630
607

.809
.770
.812

•795
.801
.821
797
•835
.830
.866
.867
.808

•859
.787
.781
•798
.851
.821
.790
.862
•845

Average . .

157

241

24.6

314

.728

.714

I-I43

600

.819

General average
for 72 birds . . .

158

245

25.2

31.6

•736

.716

1. 138

.603

.845

THE ELIMINATION OF THE UNFIT

225

Table III^
Measurements of Twenty-eight Adult and Young Females which Perished.

V

X

Length of

Beak
AND Head.

(I. .
0 g

ii

Z 3

[I.
0 .

il

Length of

Ti BIO-
TARSUS.

X J

Length of
Keel of
Sternum.

37 9

165

240

26.3

31.4

.709

.710

1.126

.614

.815

38 9

166

240

26.8

31.5

.715

.678

1.127

.597

.812

39 9

160

242

26.

32.6

.740

.732

1.157

.597

.854

40 9

1521

2323

23.23

30.3

.6762

.683

1.048

.590

.780

41 9

160

260

26.5

31.7

.741

.731

1.187

.615

.886

42 9

155

237

24.2

31.

.727

.723

1.118

.610

.787

43 9

157

245

26.9

32.2

.766

.751

1.2272

.620

.841

44 9

1653

245

27.7

33.12

.7801

.7573

1.195

.633

.896

45 9

1532

2312

23.9

30.1

.6803

.6623

1.0423

.692

.781

46 9

162

239

26.1

30.3

.7«9

.685

1.092

.587

.911

47 9

162

243

24.6

31.6

.741

.729

1.162

.605

.840

48 9

159

245

23.6

31.8

.727

.700

1.129

.610

.855

49 9

169

247

26.

30.9

.711

.666

1.098

.580

.7492

60 9

155

243

25.

30.9

.730

.711

1.127

.698

.839

61 9

162

252

24.8

31.9

.762

.738

1.180

.615

.875

62 9

1621

2301

22.82

30.4

.682

.664

1.0423

.6511

.7341

53 9

159

242

24.8

30.8

.717

.667

1.090

.676

.809

54 9

165

238

24.6

31.2

.706

.702

1.102

.588

.7683

55 9

163

249

30.52

33.41

.767

.7671

1.2073

.6401

.896

56 9

163

242

24.8

31.

.713

.713

1.128

.607

.813

67 9

156

237

23.9

31.7

.718

.716

' 1.090

.611

.800

58 9

169

238

24.7

31.6

.726

.701

1.145

.600

.800

59 9

161

245

26.9

32.1

.751

.704

1.142

.607

.819

60 9

165

235

22.61

30.7

.695

.692

1.119

.684

.771

61 9

162

247

26.1

31.9

.751

.735

1.167

.618

.802

62 9

1532

237

24.8

30.6

.732

.718

1.172

.694

.802

63 9

162

245

26.2

32.6

.728

.731

1.102

.614

.832

64 9

164

248

26.1

32.3

.739

.707

1.159

.592

.823

Average . .

158

241

25.3

31.4

.726

.709

1.131

.601

.820

General average
for 64 birds . . .

160

245

25.8

31.5

.728

.709

1.128

.601

.834

226

BIOLOGICAL LECTURES.

Table IV.

The Maximum and Minimum Measurements.

it

X

Length of

Beak
AND Head.

0 Ji

X g
H W

0

11

U' (3L4

Length of

Tl BIO-
TARSUS.

0 .

X i

Length of
Keel of
Sternum.

Maximum . .

167

256

31

33.4

ict)

.767

1.230

.640

.927

Minimum . .

{a)
152

230

kh

29.8

.659

•653

{c)
I.O[I

.551

.734

(a) The minimum length (152 mm.) occurs twice among the birds which perished : Nos. 40 and 62.
(ib) The minimum weight (22.6 grams) occurs in each group : Nos. 53 and 60, and therefore is not

entered.
(c) The minimum length of head (29.8 mm.) and of tibio-tarsus (i.oii inch) occurs twice among the

surviving" birds : Nos. 9 and 14, 41 and 51.
{d) The maximum length of humerus (.780 inch) occurs in each group : Nos. 6 and 44, and therefore

is not entered.

Table V.

Average Departures from Ideal Mean.

1!

H

X

Length of

Beak
-and Head.

5 1

ii

Length of

Tl BIO-
TARSUS.

It

Length of
Keel of
Sternum.

Seventy-two
which survived

2.51

4.20

10.9

551

16.

14.

29.4

10.

32-

Sixty-four
which perished

3.48

4.60

12.6

5.64

20.1

20.

33.8

12.

31.

TWELFTH LECTURE.

ON THE HEREDITY OF THE MARKING IN
FISH EMBRYOS.

JACQUES LOEB.

' I.

Until recently heredity has been treated chiefly as a problem
for whose solution one single theory or one single principle
was considered possible and sufficient. The theories of hered-
ity by rjarwin, Weismann, Spencer, and Jaeger are proofs of
this. Heredity is explained by Darwin on the assumption
that the organs of the parents give off particles from which the
organs of the offspring originate. According to Weismann,
the egg contains determinants for every organ contained in the
later organism and which have a definite arrangement. Accord-
ing to Jaeger, the offspring resembles the parents because both
consist of the same chemical constituents. None of these theo-
ries of heredity have been generally accepted, nor do I believe
they ever will be. They overlook the fact that heredity is a
collective term for a series of heterogeneous circumstances
which cannot possibly be explained by one principle. Thus it
happens that while each of these theories is adapted to some
of the facts of heredity, other facts can be better explained by
some other theory.

. In contradistinction to these attempts to explain heredity by
one single principle and by means of one single theory, a more
analytical study of the subject has been undertaken. This has
led to the conception that very different circumstances deter-
mine the various details in heredity. We thus find ourselves
face to face with the task of investigating in detail which cir-

227

2 28 BIOLOGICAL LECTURES,

cumstances determine the single facts of heredity. The change
in our attitude toward this problem is similar to that which has
taken place in psychology. Psychologists no longer try to give
a theory of the soul or of consciousness, but try to analyze the
various groups of psychical phenomena more or less independ-
ently of each other.

It may be desirable to illustrate what we mean by the analyt-
ical study of heredity. From the ^gg of the sea urchin only
one embryo originates. If we isolate the first two cleavage
cells, we get two embryos. This agrees very well with Jaeger's
chemical theory, but it does not agree with Weismann's. I
have made an experiment which seems to show that neither
Jaeger's nor Weismann's theory is correct. If the contents
of the ^gg be transformed into a double sphere before segmen-
tation sets in, either a dumb-bell-shaped blastula or two blastulae
are formed which are at first grown together, but as a rule are
separated later. It seems that a blastula originates by the
cleavage cells being forced to migi'ate to the limit between sea
water and egg or to the surface of the egg. Probably tropisms of
the cleavage cells, chemotropism, or stereotropism are involved
in this. It is easy to observe that the cleavage cells of the
sea urchin's ^gg can execute amoeboid motions. Hence we
have to deal with a circumstance which neither the theory of
Jaeger nor of Sachs nor of Weismann nor of any other author
could anticipate ; namely, the necessity of the cleavage cells
creeping to the periphery of the ^g'g.

We may select another circumstance, for instance, the
heredity of the arrangement of organs in an organism. From
the germ of Hydroids an organism develops which has stolons
at one end of its body and polyps at the other. This definite
arrangement is in many cases due to a fact which none of the
general theories of heredity could have predicted. We find in
many Hydroids that an excised piece of the stem forms stolons
if the end comes in contact with solid bodies, while the same
end forms polyps if it is surrounded by sea water. When
a germ or a larva settles down at the close of the swimming
stage, that end of the germ which is in contact with a solid
body will produce stolons, while the other end, that is in con-

THE MARKING IN FISH EMBRYOS. 229

tact with sea water, must give rise to polyps. Hence the
hereditary arrangement of organs is in such cases due to the
fact that while one side of the germ comes in contact with
solid bodies, the other end remains in contact with sea water.
Let us next consider the hereditary power of assimilation.
We know that the corresponding proteids of a horse, a dog,
and a cat are different. If we feed a cat with horse flesh, it
transforms it into the specific proteids of a cat, while in a dog
the same hoise flesh is transformed into the specific proteids of
a dog. For this transformation specific enzymes are necessary
which are different in both forms. As the fertilized ^gg is
able to carry on processes of assimilation, we must assume that
such enzymes are transmitted through the ^g'g to the offspring.
The presence of enzymes or zymogens, from which the fer-
ments originate, is another circumstance in heredity which
none of the general theories has considered. We see thus that
what' we call heredity is composed of very heterogeneous con-
stituents which cannot be explained by one single principle or
theory, unless the theory should become so indefinite as to
be almost meaningless. But such a theory would cease to be
of value. This paper will deal especially with the heredity of
the markings in fish embryos, which will offer another example
of the analytical study of heredity.

II.

Seven years ago I published a series of observations on the
origin of the markings in the yolk sac of the embryo of Fundu-
lus. The yolk sac of Fundulus possesses a tiger-like coloration.
Its origin is as follows : black and red chromatophores are
found on the surface of the yolk sac. They gradually creep
upon the blood vessels and form a sheath around them. This
creeping of the chromatophores upon the blood vessels is due
to a tropism of the former. I then believed it was chemotrop-
ism and assumed that the chromatophores were oriented by a
chemical constituent of the blood, especially by oxygen. But it
may be that it is stereotropism, or that both tropisms cooperate.
The heredity of the marking is, therefore, in this case deter-

230 BIOLOGICAL LECTURES.

mined by a stimulus which the blood vessels exercise upon
another tissue, namely, the chromatophores. Both tissues are
formed rather independently of each other, but from the fact
that the chromatophores must creep upon the blood vessels,
and that the latter have a hereditary arrangement, the marking
becomes hereditary too. This contradicts those theories of
heredity which try to derive all the peculiarities of the animal
from corresponding peculiarities of the sexual cell, for instance,
Weismann's theory.

The observations I made on Fundulus have since been con-
firmed for other classes of animals too. Dr. Arnold Graf made
similar observations in leeches, and Zennek has found that the
longitudinal stripes in the Ringelnatter are determined by
blood vessels. But as he has overlooked my papers on the
subject he did not realize that the marking is produced by
the chromatophores being bound to creep upon the blood vessels.
I have continued my observations on the yolk sac and the em-
bryo of Fundulus, and will try to explain the history of the
origin of the marking in connection with some drawings.

Fundulus has two kinds of pigment cells: (i) large black
cells with almost no ramifications and processes, and (2) small
red pigment cells with an enormous amount of ramification.
Both differ somewhat in their reactions. The black cells react

and move quicker than the red
cells. In Fig. 3 the large black
pigment cells can easily be dis-
criminated from the star-shaped
red pigment cells. The latter
are characterized in the drawings
by fine lines.

At first the yolk sac is free
from pigment cells. As soon as
the vessels and the blood corpus-
cles develop, the first pigment
cells make their appearance on
the yolk sac. Fig. i shows this stage. Some of the large
black pigment cells are situated on blood vessels, others are
scattered in the gaps between the capillaries, but there is no

THE MARKING IN FISH EMBRYOS.

231

definite relation between the position of the pigment cells and
the blood vessels, and there is no definite marking. The same
is true in regard to the red pig-
ment cells. (The little spheres
are oil globules.)

A few days later the number
of capillaries as well as that of
the pigment cells has increased
(Fig. 2), but a further change
is also noticeable ; each pigment
cell has crept upon a blood ves-
sel, and the shape of the black
pigment cells has changed some-
what. Now they are all rather
long ; that is, they follow the
direction of the blood vessels.
In the small ramified red pig-
ment cells no change of form is
as yet perceptible. Fig. 3 shows
a very characteristic condition,
such as is usually found toward
the end of the first week. The
black pigment cells follow the
blood vessels entirely. Where
a blood vessel branches off into
two parts the pigment cells show
the same ramification. The
black pigment cells now form a
sheath around the blood vessels.

Fig. 4 shows the yolk sac
v/hen the marking is almost com-
pleted. The black and red pig-
ment cells surround the blood
vessels so completely that no
one who sees the eggs for the
first time at this period could
realize that the black and red stripes in the yolk sac could
have originated from chromatophores. That part of the blood

Fig. 3.

Fig.

232

BIOLOGICAL LECTURES.

Fig. 5.

vessels which is marked by fine lines is covered by red pigment
cells, the other part by black pigment cells.

In my first publication on the subject I confined myself to
a description of the origin of the marking in the yolk sac. I
mentioned, however, that in the embryo itself the marking

seemed to originate in the same way.
I have since examined the origin
of the marking in the embryo more
closely, and have found it the same
as in the yolk sac. Fig. 5 shows
a comparatively young embryo, in
which the pigment cells are still
scattered irregularly. Later on,
however, the pigment cells here,
too, creep upon the blood vessels.
Fig. 6 shows the blood vessels and
pigment cells in the tail of an older
embryo in which the marking has
taken place. In some of the blood vessels the direction of the
blood flow is indicated by arrows. This drawing, like all the
others, was taken from life.^ We see that in the embryo, as in
the yolk sac, all the pigment cells have crept upon blood ves-
sels. I should like
to direct the atten-
tion of the reader
to one special
point : in the mid-
dle of the body two
large blood vessels,
an artery and a
vein, run side by
side. They are
marked by two ar-
rows. Here all the chromatophores have crept upon the artery
while the vein has remained free. It is obvious that the artery
contains more oxygen than the vein, and this seems to suggest

1 These drawings were made by Mr. Bridgham, of Providence, R. I., under my
direction.

Fig. 6.

THE MARKING IN FISH EMBRYOS. 233

that the oxygen of the blood may be one of the causes that
force the chromatophores to creep upon the blood vessels. I
have expressed a similar belief in my former papers, but the
oxygen in the arterial blood is possibly not the only reason
why the chromatophores creep upon the blood vessels, since
wherever a vein is isolated they creep upon the vein too. Thus
it seems that positive chemotropism for oxygen is one stimulus,
but possibly not the only one that causes the migration of the
chromatophores upon the blood vessels.

III.

The further examination of the marking of the Fundulus
embryo has shown that the blood vessels are not the only
factors that determine the marking. A second factor is the
central nervous system. The back of the embryo is colored
black by pigment cells which follow the brain and the spinal
column also. They follow the outline of these organs almost
as closely as they follow the blood vessels. The details of the
origin of this marking must be reserved for future investigation.

Everything we have said thus far about the marking of the
embryo refers to chromatophores, but in addition to chromato-
phores two other elements aid in producing the markings in
fish embryos. In certain places a diffused yellowish pigment
is deposited. It is a kind of excretion. The chemical char-
acter of this pigment and its origin have yet to be determined.
This pigment appears at a rather early stage in the develop-
ment. The second element is the structural colors. They do
not play as much of a r61e in the early stages of the embryo.
I intend to continue my study of the development of the mark-
ing in these embryos, especially in regard to the influence of
hybridization. I have already begun experiments on the latter
topic, but they are not far enough advanced to be published.

Finally, may I be pardoned for adding a few remarks con-
cerning the theories of Eimer. I am no believer in Weismann's
theories of heredity, but it seems to me that the so-called theo-
ries of Eimer on the origin of the longitudinal striation in
animals are nothing but a play on words. One of his general

2 34 BIOLOGICAL LECTURES.

"laws" maintains that every marking is at first longitudinal.
The truth of the matter is that most animals are not spherical,
and that most organs, especially in segmented animals, run in
a longitudinal direction through the animal ; for instance, the
spinal cord, vertebral column, blood vessels, intestine, etc. In
the embryo it is especially true for the large blood vessels.
The first striation in such embryos is, of course, longitudinal,
simply because it is determined by the blood vessels ; but wher-
ever we have to deal with the coloration in a spherical organ,
as, for instance, in the yolk sac of the Fundulus embryo, the
absurdity of Elmer's law becomes evident. Here, on the sur-
face of the sphere, there is no longitudinal direction, and yet
here, too, the chromatophores follow the blood vessels. In
the same way the so-called law of Eimer appears absurd if we
consider the marking, which is determined by the pigment cells
that cover the brain.

Although it seems to me that at present we cannot avoid
dealing with the problem of heredity from an analytical point
of view, it is clear that this is only a temporary necessity.
Sooner or later the facts found by the analytical method must
be utilized for the establishment of more general laws.

THIRTEENTH LECTURE.

DO THE REACTIONS OF LOWER ANIMALS DUE
TO INJURY INDICATE PAIN-SENSATIONS?

W. W. NORMAN.

Associate Professor of Biology, University of Texas.

A BRIEF note by me concerning pain-sensations in the lower
animals appeared in Pfliiger s Archiv for 1896.^ It was my aim
to give experimental proof that the reactions of lower animals
upon injury furnish no safe evidence of pain-sensations. For
these experiments chiefly the earthworm (Allolobophora foetidd)
was used.

It had already been observed by Friedlander^ that when an
earthworm is cut in two in the middle, the two halves react
differently — the anterior half crawling away and burying itself
like a normal worm, the posterior end making strong winding
and jerking motions. [Unmittelbar nach der Operation machen
sie (die gekopften Wiirmer) heftige schlagende und windende
Bewegungen.] Loeb ^ had observed similar motions in other
worms similarly treated.

If Friedlander had carried his experiment further, namely, if
he had cut each half of the already halved worm in two in the
middle, the results would have become, indeed, still more strik-
ing. It is here that my experiment presents an essentially new
point.

1 " Diirfen wir aus den Reactionen niederer Thiere auf das Vorhandensein von
Schmerzempfindungen schliessen ?" Pfluger's Archiv, Bd. Ixvii, p. 137.

2 Friedlander, " Ueber das Kriechen der Regen wiirmer," Biol. Centralblatt,
Bd. viii.

3 Loeb, J., " Beitrage der Gehirnphysiologie der Wiirmer," Pfliiger^s Archiv,
Bd. Ivi.

23s

236 BIOLOGICAL LECTURES.

For convenience let us call the anterior half of the worm a
and the posterior half b. Now if a be cut in two in the mid-
dle, the anterior piece a\ elongates and begins to crawl, while
the posterior piece az acts like b\ i.e.^ executes jerking and
squirming motions. Thus far in the experiment each piece
that began the progressive movements immediately after the
injury was the anterior piece, containing the so-called brain.
Now if we cut in two the posterior half b, we get just the same
kind of reactions as in case a — namely, the anterior part b\
elongates and begins progressive motion, while the posterior
piece bz acts as ^2, carrying out a series of jerking and squirm-
ing motions. We observe, then, that each time after the halving,
the anterior piece elongates without any irregular muscular
contractions and begins to crawl forwards, while the posterior
piece always executes the same kind of squirming motions.

The experiment may be carried still further, and each of the
four pieces cut in two. In each case the posterior half (one-
eighth of the whole worm) immediately begins the jerking and
squirming motions, while the corresponding anterior half elon-
gates and, if not too short, begins to crawl. Indeed, if small
vigorous worms be used, the characteristic reactions of the two
halves may be observed in pieces that are only about five
millimeters in length.

As to the significance of the reactions of the lower animals,
there are two opposite views. Those supporting the one view
assume that the reactions to external stimuli are the outward
expression of psychical processes resulting therefrom. The
larvae of crustaceans, for example, may prefer blue light to
red, if there be opportunity for choice. ** The flesh-fly, Mtisca
carnaria, deposits its eggs in the flowers of the carrion plant,
the smell of which resembles that of putrid meat, and so
deceives (!) the fly." ^ Again, according to the same way of
reasoning, the moth flies into the flame, prompted by curiosity
to get acquainted with the strange object.

The supporters of this view ignore the difliculty that con-
fronts them — namely, that the only means of determining
the presence of consciousness in the lower animals are the

1 Romanes, Mental Evolution in Animals, p. 167.

THE REACTIONS OF LOWER ANIMALS. 237

movements themselves, which in turn are to be explained
through psychical processes. Under the title '* Mechanik,"
Verworn,^ for example, expresses himself as follows: "The
question might be raised as to the possibility of reaching a
conclusion concerning the subjective conditions of another
organism, since they lie outside the investigator. The following
consideration, however, proves the possibility of gaining defi-
nite knowledge in this direction. We know from our own
subjective experiences our own subjective states, and, on the
other hand, the objective expressions which characterize them.
Here we have two known quantities. A third known quantity
we can get through the objective expressions of psychical
processes (!) of the organisms under observation. If now we
make a proportion between the processes of man and of the
organism to be studied, i.e.^ if we compare the two with each
other, we can reach a conclusion as to the subjective processes
of the organism under consideration according to the principle

X c c

- = - ^ X = a • -, in which a represents the subjective processes

a b b

of man, c the objective expression of the same, b the objective
expression of the subjective processes of the object under investi-
gation, and ;i:its unknown subjective processes to-be-found-out."

Man konnte bestreiten, dass es iiberhaupt moglich sei, irgend
welchen Aufschluss iiber die subjectiven Zustande eines anderen
Organismus zu erlangen, da dieselben ausserhalb des Untersuchers
liegen. Indessen beweist die folgende Ueberlegung trotzdem die
Moglichkeit, wenigstens durch Schliisse sichere Erfahrungen in
dieser Richtung zu gewinnen. Wir kennen aus unserer subjectiven
Erfahrung einerseits unsere eigenen subjectiven Zustande und ande-
rerseits die objectiven Aeusserungen welche dieselben characteri-
siren. Hierin besitzen wir aber zwei bekannte Grossen. Eine dritte
bekannte Grosse konnen wir uns schaffen durch die Beobachtung
der objectiven Aeusserungen psychischer Vorgange bei den zu
untersuchenden Organismen. Wenn nun zwischen den betreffenden
Vorgangen beim Menschen und bei den Untersuchungsobjecten eine
Proportion gesetzt wird, d. h. wenn beide mit einander verglichen

werden, so kann man nach dem Prinzip -^-"i x=^ a- -•> wobei a

a b c

1 Verworn, Psycho-Physiologische Protisten-Studien, p. 18. Jena, 1889.

238

BIOLOGICAL LECTURES.

die subjectiven Vorgange beim Menschen, c deren objective Aeus-
serungen, b die objectiven Aeusserungen der subjectiven Vorgange
bei den Versuchsobjecten und x die zu erforschenden subjectiven
Zustande dieser selbst sind, Aufschluss iiber die subjectiven Vor-
gange der betreffenden Organismen erlangen.

The striking feature in this method is the assumption of the
fact to be proved — namely, the existence of psychical processes
in the organisms to be investigated. According to the equa-
tion given, not only the whole earthworm but any isolated piece
of the same would be capable of psychical processes, since upon
injury it reacts in the same manner as the whole worm.

Those holding the second view seek to analyze the reactions
of animals on purely mechanical grounds. The most pro-
nounced defender of this view is Loeb.^ He has by this
method shown that the phenomena of orientation of animals
towards light agree in every particular with the phenomena of
orientation of plants towards the same source of stimulus.
Hence the heliotropic reactions of plants must be referred to
*' curiosity," or to some other anthropomorphic process, or else
it must be admitted that the phenomena of orientation of ani-
mals are to be explained just as mechanically as in the case of
plants.

That consciousness is a function of associative memory
(associatives Gedachtniss) has been emphasized by Loeb.*-^ But
memory has thus far been proved with certainty only in such
forms as have a well-developed cerebrum. The facts of brain
physiology speak decidedly against the view that phenomena
of consciousness are everywhere present in the animal
kingdom.^

It cannot, however, be denied that certain reactions of lower
animals against injury, which in man cause pain, lead the inex-
perienced person easily to the conclusion that these animals
really suffer pain. An earthworm, for example, touched with

1 Loeb, J., Der Heliotropismus der Thiere und seine Uebereinstimmung mit
dem Heliotropismus der Pflanzen. Wiirzburg, 1890.

*^ Loeb, J., " Beitrage der Gehirnphysiologie der Wiirmer," Pfluger's Archiv^
Bd. Ivi.

3 Loeb, J., "... Zur Physiologie und Psychologie der Actinien," Pflilger^s Archiv,
Bd. lix.

THE REACTIONS OF LOWER ANIMALS. 239

a needle, or otherwise slightly stimulated, may be thrown into
violent jerking and squirming motions. According to Loeb we
have no more right in this case to conclude that these motions
are due to pain-sensations than we have to make a similar con-
clusion upon the contractions of an isolated frog's muscle when
put into a too concentrated salt solution. My experiments
upon the earthworm clearly prove that this view in contrast to
that of the pain-sensations is the correct one.

It has already been observed that if we attribute feeling to
the earthworm because of certain reactions, we must do like-
wise for any isolated piece of the worm ; and, further, that the
front half of any such piece must have a state of feeling differ-
ent from that of the posterior half. But since the piece to be
experimented upon may be taken from any part of the worm,
the front half could as easily have been the posterior half, and
vice versa. Hence it follows that the nature of the reaction of
any given piece of the earthworm is a function of the direction
of the impulse. If the impulse travels from the injured spot
anteriorly, the piece of worm elongates ; this initiates the nor-
mal progressive movements ; if the impulse travels from the
injured spot posteriorly y the piece of worm executes the charac-
teristic jerking and squirming motions. In the first case the
circular muscles are the first to contract, and this causes the
elongation of the piece ; in the second case the longitudinal
muscles are the first to contract. This causes, of course, the
jerking and winding motions. Why the impulses traveling for-
wards should always reach first the circular, and those traveling
backwards first the longitudinal muscles, I have not yet been
able to explain.

These reactions are not peculiar to this special mode of
application of the stimulus, but are the consequence of locally
applied stimuli. If an earthworm, for example, while crawling,
be struck somewhat posterior to its middle point with the dull
edge of a scalpel, the part of the worm in front of the offended
spot continues to crawl, but at a more rapid gait ; while the part
beyond the offended spot at once begins to squirm, at the same
time being dragged along by the anterior part. Or, again, if
the entire normal worm be irritated at its anterior end, it is

240 BIOLOGICAL LECTURES.

thrown into violent squirmings ; while the same stimulus applied
at the posterior end only hastens the normal progressive move-
ments. After a full study of the reactions of the earthworm
against injury ^ including also heat and electricity^ it was found
that any stimulus having a definite local action would call forth
the two characteristic sets of reactions.

Are there other animals whose reactions to injury are as
difficult to explain according to the pain-sensation theory as
those of the earthworm ? This question may be answered
positively in the affirmative. Let us consider the reactions
of animals chosen from four different groups. A live starfish,
as every one knows, may be cut into pieces with a pair of
scissors, or otherwise mutilated, without so much as calling
forth a responsive reflex action other than a temporary with-
drawal of the tube-feet. An arm thus cut off from the body
and left free in a little water soon begins to move about in
the same easy way that characterizes the motions of the whole
animal.

The leech is perhaps the most highly differentiated of the
worms, and it is here that we should expect definite expression
of pain-sensations, if they are at all to be expected among such
forms of life. Yet what do we actually find t The normal pro-
gressive movement of the leech in water is a rapid sinuous
swimming motion, carried out by the entire animal, but most
effectively by the flattened posterior portion. If now while
the animal is gliding through the water it be cut in two in
the middle with a sharp pair of scissors, both pieces without the
slightest reaction to the injury (or at most the slightest momen-
tary pause) continue to swim, with,. however, a modified rate of
speed, due to their modified length and shape.

The peculiar habit of crabs^ of breaking off their thoracic
appendages near the base, due to strong stimulation, either
directly or indirectly, has been proved to be a purely reflex
action, and may take place even after removal of the brain.

Among insects, the honey-bee has held an undisputed place
for intelligence by all who put the lower animals in the list of

1 Fredericq, Leon, " Nolivelles Recherches sur L'Autotomie chez le Crabe,"
Archives de Biologic^ XII, p. 169.

THE REACTIONS OF LOWER ANIMALS. 24 1

conscious beings. Bethe,i however, says : "■ I have cut away the
entire abdomen from bees and have seen them live over an
hour, during which time, from the moment of the operation on,
if I placed them at some honey, they sucked at it unceasingly.
Indeed, while a bee sat on my hand and sucked honey I have
suddenly cut off its abdomen. It raised up for a moment and
then quietly sucked on at the honey." I have cut off the
abdomen, piece by piece, of Libellida without the animal's
making any reactions whatever.

In conclusion it may be here remarked that of course the
absence of reactions on the part of the lower animals to injury
does not directly prove the absence of pain. It may, however,
be as strongly asserted that the reactions of these animals to
injury furnish no safe evidence that they are due to the pres-
ence of pain-sensations ; and, further, if such reactions do indi
cate pain, then by the same criterion we must attribute pain
sensations to pieces of an animal which likewise react to injury,
— a view which to me seems entirely unbiological.

1 Bethe, Albrecht, " Vergleichende Untersuchungen iiber die Functionen des
Centralnervensystems der Arthropoden," Pfluger's Archiv, Bd. Ixviii, p. 509.

FOURTEENTH LECTURE.

NORTH AMERICAN RUMINANT-LIKE MAMMALS.

W. B. SCOTT.

In palaeontology, as in other lines of inquiry, the progress
of discovery calls for frequent revision of opinion and changes
of standpoint. In all kinds of investigation, the steps of which
cannot be rigidly demonstrated, but depend upon the accumu-
lation of evidence and the balancing of probabilities, it some-
times happens that obscure and difficult problems are suddenly
lighted up by a new discovery, which gives a totally new and
unexpected aspect to the subject. I have come here this even-
ing to make public recantation of some of the errors which I
lately upheld, having had my opinion completely changed by
the force of new evidence, which has led to some very surpris-
ing results. This new evidence consists in the fossil mammals
lately collected in the Uinta and White River formations, by
Mr. Hatcher for the museum of Princeton University, and by
Messrs. Wortman and Peterson for the American Museum of
Natural History in New York.

The Uinta fauna is a most interesting transitional one, which
is still very imperfectly known, but every new collection made
of it increases our appreciation of its supreme importance, for
by its aid one phyletic series after another is being completed.
These series have a far-reaching significance for all departments
of morphology, for they bring before us what we have every
reason to regard as the actual steps of descent, and thus enable
us to learn exactly what kinds of evolutionary changes do take
place. The Uinta beds have a comparatively limited geographi-
cal extent, and have been found only in the basin south of the
Uinta Mountains in northeastern Utah and northwestern Colo-

243

244 BIOLOGICAL LECTURES.

rado. They overlie the strata of the upper Bridger stage
(Washakie substage) and clearly precede the White River in
time, though not separated from the latter by any great interval.
Ordinarily the Uinta beds have been called Upper Eocene, but
it would perhaps be better to refer them to the Lower Oligo-
cene, for they are to be correlated with the Paris Gypsum,
which the best French opinion now refers to the base of the
Oligocene. It will be convenient to repeat here a part of the
table of American fresh-water Tertiary formations given in a
previous volume of these lectures.

r John Day.
Oligocene ^ White River.
1^ Uinta.

Eocene

f Bridger.
J Wasatch.
I Torrejon.
I Puerco.

The correlation of the Uinta with the Paris Gypsum is made in
spite of very marked differences in the mammals which are
found in the two regions, for the differences are geographical
rather than geological, and are doubtless to be explained by the
existence of barriers which made migration between the two
continents difficult, though not impossible. In the Wasatch
the mammalian faunas of Europe and North America were
remarkably similar and clearly indicate that the two continents
were connected by land areas, which allowed the freest migra-
tion between one region and the other. In the Bridger there
appear to have been some obstacles raised in , the way of this
free interchange of terrestrial mammals between the northern
continents, though the correlative fauna of Europe is still too
incompletely known to show the exact amount of difference. In
the Uinta, however, the materials for comparison are abundant
and prove that intermigration was opposed by such difficulties
that only a few forms were able to overcome them. By White
River times these obstacles, of whatever nature they may have
been, were removed and once more the mammalian faunas of
the two continents contained a large number of genera common

AMERICAN RUMINANT-LIKE MAMMALS. 245

to both. In the long interval between the Wasatch and the
White River, North America had ample opportunity to develop
a peculiar mammalian fauna, and one which was composed of
types specially adapted to the conditions of life on this conti-
nent, and well fitted to resist the invasion of more or less similar
forms from the other continents, when easy intercommunica-
tion was reestablished.

Any fauna may, for our present purpose, be conveniently
divided into two somewhat heterogeneous assemblies, one of
which is composed of immigrants from other regions, and the
other is indigenous. By the latter term are meant those forms
which have long been established in a continent and whose
ancestry may be traced through several geological horizons.
Using this convenient mode of discrimination, the ruminants
which from time to time have inhabited North America may be
distinguished as either indigenous or immigrant types. The
indigenous ruminants (it will be more accurate to call them selen-
odonts) predominated for a very long period of time and only
toward the end of the Miocene did Old World types of seleno-
donts obtain a permanent foothold here. It is a remarkable
fact that of these indigenous types which so long held sway in
this continent, not one is left here at the present time, and,
except the camels of the Old World and the llamas of South
America, they have become altogether extinct and have left no
descendants among recent mammals.

The indigenous selenodonts of North America have long puz-
zled the students who have attempted to work out their system-
atic position and their relationships to the selenodonts of other
continents. Even the complete skeletons of various genera,
recovered from time to time, seemed to give little help in solv-
ing the problem. The great obstacle to progress has been the
absence of well-defined phylogenetic series, which would enable
the observer to trace out the history of the various families and
groups, step by step, through their numerous ramifications to
final extinction, or to their modern representatives. An isolated
genus, standing by itself, the predecessors and successors of
which are unknown, offers almost insuperable difficulties to the
determination of its proper taxonomic position. Such a genus

246 BIOLOGICAL LECTURES.

always displays resemblances to a number of widely separated
groups, not all of which resemblances can possibly be due to
relationship, and many of which must have been independently
acquired. It is easy to recognize this general principle, but it
is by no means easy to apply it in a given case by estimating
the taxonomic value of the various resemblances and differences,
and to distinguish the characters which are due to real affinity
from those which have resulted from a convergent or parallel
course of development. Hence it can hardly be a matter for
surprise that, even among competent observers, great differ-
ences of opinion should arise concerning the systematic posi-
tion of isolated groups, one writer giving special emphasis to
one set of characters, and another to another set. Only when
the ancestry of the group in question has been made out does
a satisfactory solution of the problem become possible.

The indigenous selenodonts of North America first began to
be important in the Uinta formation, for in the strictest sense
of the term the Bridger has as yet yielded none, though there
are two or three genera which obviously represent the incipient
stages of the group. Their culmination, so far as numbers,
variety, and relative faunal importance are concerned, may be
regarded as falling within the White River age, though they
nearly held their own in the succeeding John Day. In the
White River we find mingled with the indigenous selenodonts
certain genera, like Aiicodus and Anthracotheriiim, which had
evidently migrated from the Old World, but did not secure a
lasting foothold here, for no trace of them has been found in
the John Day. In the Loup Fork the North American type
of selenodonts underwent a very marked reduction, while
migrants from Eurasia assumed a more and more important
r61e, until, at the present, all our representatives of the group —
deer, antelope, sheep, bison, etc. — are descendants of Old World
ancestors, which, in some instances, reached North America at
a very recent geological date. The only survivors of the Amer-
ican type, the camels and llamas, are no longer found in their
original home, but are confined to two widely separated regions,
the camels to the Old World and the llamas to South America.
The explanation of this curious case of discontinuous distribu-

AMERICAN RUMINANT-LIKE MAMMALS. 247

tion is that the group originated in North America, and after
sending out migrants to the other regions, became extinct
here.

The study of the peculiarly American selenodonts can most
conveniently begin with those of White River times, because
they are the best known, nearly every genus being represented
in the collections by complete skeletons, and because they were
then perhaps in the most characteristic stage of their develop-
ment. As to the number of families represented in the forma-
tion, it is difficult to reach a conclusion, and almost as good
reasons may be given for grouping them into seven families as
into four, the arrangement here adopted. The position of only
one of these groups has long been understood and very gener-
ally agreed upon, and this is the family represented in White
River times by Poebrotherhim, a genus which, there can be
little doubt, is the ancestor of the modern Tylopoda, or, at the
very least, of the llamas. The whole appearance of the skele-
ton, with its small tapering head, long neck, and elongate
slender limbs and feet, is like that of a small llama, but of
course it is very much more primitive and less specialized than
the existing members of the group. The teeth are still undi-
minished in number ; the canines are hardly larger than the
incisors and only beginnings of diastemata are visible ; the pre-
molars are much elongated in the antero-posterior direction, and
the molars are commencing to take on the prismatic or hypso-
dont shape. The skull is unmistakably tylopodan in character,
with its triangular form and slender tapering muzzle, and while
primitive features are retained, the enlarged tympanic bulla
filled with cancellous bone has already been formed. The
cervical vertebrae display the tylopodan peculiarity of a con-
cealed vertebrarterial canal, perforating the neural arch, and
only in the sixth vertebra does the canal occupy its normal
position. The limbs and feet are already very elongate ; the
ulna and radius are coossified, and the fibula is completely
reduced ; the feet are didactyl, the lateral digits being reduced
to mere nodules. The phalanges are slender and the unguals
long and pointed, the shape of the latter showing that the
hoofs were like those of the deer and antelope, and that the

248 BIOLOGICAL LECTURES.

characteristic pad of the modern tylopodan foot was not de-
veloped till a later time.

While there has been no dispute that P oebrotherium is the
White River representative of the main line of tylopodan
descent, there has been no such agreement concerning the
other selenodonts of that age, and this for the reason that
nothing has been known of their ancestry, and that among
existing mammals none could be selected as being a descendant
of any of these genera ; hence almost every conceivable opinion
as to their systematic position has been held and defended. It
is upon this very problem that the newly discovered Uinta forms
shed such welcome light, and they render it exceedingly prob-
able that all the strictly indigenous North American selenodonts
are branches of the great tylopodan stan.

Paradoxical as this conclusion may appear to be, I believe
that it is fully justified by the evidence, though it is, unfortu-
nately, not possible to bring before you any adequate digest of
that evidence, for it consists of a great mass of tediously minute
dental and osteological comparisons of many genera and species.
However, some of the more striking parts of the testimony
may be made intelligible without an undue amount of detail.
The Tylopoda are thus seen to be a very ancient and highly
diversified group, comparable in these respects to the Pecora,
or true ruminants, which in so many features they closely
resemble, though the resemblances have, for the most part,
been independently acquired in the two suborders. The Pecora
are an Old World group, which underwent a great expansion
and diversification in Eurasia, but did not reach this continent
till late Miocene times, and they never attained the importance
here that they have so long had in the eastern hemisphere.
Even at the present time, when they have completely sup-
planted and driven out the Tylopoda from North America, they
are far less numerous and varied here than in the Old World.
Several of their American representatives, such as the bison,
sheep, and musk-ox, are very recent immigrants, not occurring
in beds older than the Pleistocene. In America the place of
the Pecora was taken, to a very great extent, by the Tylopoda,
which ran a course of development, in many respects, parallel

AMERICAN RUMINANT-LIKE MAMMALS. 249

to that followed by the Pecora and Tragulina — the latter a
group which never reached America at all.

It is this very parallelism of the Tylopoda with the Pecora
and Tragulina which has led astray so many students of the
peculiar North American selenodonts, myself among the num-
ber. We have continually been endeavoring to detect relation-
ships between these forms and the European ruminants and
chevrotains, where no such relationships existed, but only analo-
gies, parallelisms, or convergences. The truth appears to be
that the indigenous American selenodonts make up a natural
assemblage of forms which, with a remarkable degree of diversity
in size and structure, are yet all quite closely related among
themselves, but only distantly with the Old World types which
more or less resemble them. The group formed by the asso-
ciation of these American families is, in fact, so diversified
that a definition can hardly be framed for it ; but that is, of
course, no valid reason for refusing to recognize it as a natural
group. Just as the Pecora are typically Old World both in
origin and development, so the Tylopoda are typically North
American, and did not reach the eastern hemisphere till the
end of the Miocene or beginning of the Pliocene, and then in
very limited numbers, Cameliis and its immediate forerunners
being the only known Eurasian representatives of the group.

The late Professor Riitimeyer of Basle, one of the greatest
of palaeontologists, reached practically the same conclusion long
ago, though no American agreed with him. It is an excellent
example of his wonderful power of insight into a tangled prob-
lem of phylogeny that he should have discerned this fact, as I
believe it to be, at a time when the fauna of the White River was
but very imperfectly known and that of the Uinta had not yet
been discovered. Dr. Wortman, in his latest paper, seems to
have adopted Riitimeyer's views, though he does not explicitly
say so, and for the same reasons that have led me to change
my previous opinions, namely, the convincing power of the
Uinta genera, all of which seem to be converging into a com-
mon term with the primeval member of the main tylopodan
phylum.

It remains to bring forward the evidence upon which the

250

BIOLOGICAL LECTURES.

conclusion here advocated is founded ; and for that purpose we
must first return to the White River selenodonts. One of the
most largely represented families in that fauna is the Lepto-
nierycidae, though the four genera which are associated in it —
Leptomeryx^ Hypertragulus^ HypisodtiSy and Protoceras — are so
different from one another that much might be said in favor of
referring each of them to a separate family. Leptomeryx was
a very small animal, the skeleton of which is extraordinarily
like that of a traguline, in which group most students of the
subject have placed it ; but the resemblance is almost certainly
a deceptive one, and the real affinities are with the Tylopoda
— a conclusion in which I am glad to find myself in complete
accord with Dr. Wortman. Of the dentition only the upper
incisors are unknown, and at least one of these is present, but the
canine has been lost. The lower canine has become an incisor
in form and function, while the first lower premolar, though
minute, is caniniform, and its shape strongly suggests that in
the immediate ancestors of the genus this tooth functioned as
a canine. The other premolars are sharp and trenchant, and
the molars, as Rutimeyer pointed out, are singularly cameloid
in character, though with traguline features also. The skulls
hitherto figured and described have all been broken across the
very delicate and fragile muzzle, but newly collected specimens
show that the skull has a very llama-like aspect and much
more nearly resembles that of Poebrothermm than had been
supposed. One important difference from the latter should,
however, be noted ; namely, that the auditory bulla is small
and free from cancellous bone. The neck is short, and the
cervical vertebrae have none of the tylopodan peculiarities.
The fore-limb is much shorter than the hind, as in the tragu-
lines, but the individual limb-bones are very like those of
Poebrotheriuniy though the ulna and radius are separate. The
forefoot has four digits, the lateral pair very much reduced ;
the trapezoid and magnum are coossified, as are the cuboid
and navicular in the tarsus — both very exceptional features in
the family and suborder. The hindfoot has a cannon-bone, of
which the distal end is split in the characteristic tylopodan
way. The hoofs are slender and pointed.

AMERICAN RUMINANT-LIKE MAMMALS. 25 I

Hypertragtilus is very much like Leptomeryx^ and also much
like Poebrotheritim. It has kept the original shape and func-
tion of the canines, which are long and slender. The skull is
thoroughly tylopodan in form and proportions and, except for
the small and hollow tympanic bulla, greatly resembles that of
Poebrotheritim. The limbs and feet differ only in a few details
from those of Leptomeryx ; for example, the ulna and radius are
coossified, but there is no cannon-bone in the hindfoot.

Hypisodiis is the smallest member of the family and was a
minute animal. It is remarkable as the most ancient American
type with hypsodont molars growing from persistent pulps, and
for having apparently ten lower incisors, the canine and first
premolar having gone over to that series. So far as it is known,
the skeleton is like that of Leptomeryx^ though no cannon-bone
is formed in the pes.

ProtoceraSj the largest genus of the group, equaling in stature
the modern musk-deer {Moschics)y is also the most curious —
indeed, one of the most peculiar and bizarre looking of known
mammals. The upper incisors have disappeared, but the
upper canine, which in the female is small, is in the male a
formidable tusk, opposed by the caniniform first lower premolar.
The other teeth have resemblances partly to those of Lepto-
meryx and partly to those of Poebrotheritim. The skull is
extraordinary, especially in the male, in which sex there are
horn-like protuberances on the parietals and great thickened
plates arising from the upper edges of the maxillaries. In
both sexes the nasals are extremely short and the narial open-
ing exceedingly large, much as in the saiga antelope. The
skeleton, limbs, and feet are those of an enlarged Hypertrag-
tilus. It may seem to involve a great strain upon credulity to
refer Protoceras to the Tylopoda ; yet its relationship to Lepto-
meryx and Hypertragidtis is perfectly clear, and wherever the
latter are placed, the former must accompany them. Wortman
is of the same opinion, and in the paper so frequently cited he
speaks of "the early Cameloids, Protoceras and Leptomeryx!' ^

One of the most characteristic of American families is the

1 " The Extinct Camelidae of North America," etc., Bull. Amer. Mus. Nat.
Hist., vol. X, p. 100.

252 BIOLOGICAL LECTURES.

long-lived and diversified group of the Oreodontidae. The type
genus is Oreodon, by far the commonest of White River fossils,
at which time great herds of O. culbertsoni roamed over the
western plains. In this genus the dentition is closed, without
diastemata ; the upper canine is large, but the lower is incisi-
form, its place being taken by the caniniform first premolar.
The other premolars are simple and trenchant, and the molars
very like those of the deer. The skull has a rather short face
and very long, narrow cranium ; the orbits are completely encir-
cled with bone, and a deep pit impresses the surface of the
lachrymal. The neck is short, the trunk long, and the tail
very long and heavy. The limbs and feet are short and thick ;
a rudimentary pollex is retained in the forefoot (this is the first
artiodactyl in which that structure was demonstrated) ; and the
hoofs are curiously rounded and flattened. In appearance and
general proportions the skeleton of this genus recalls that of
the modern peccary.

The taxonomic position of the oreodonts has been the subject
of a great deal of discussion, and almost every possible opinion
has been expressed ; but the evidence of the Uinta fauna is
very strongly in favor of the view held by Riitimeyer and
Schlosser; namely, that they are aberrant members of the
Tylopoda. In time they ranged from the Uinta through the
Loup Fork, each successive horizon yielding peculiar genera ;
but with their later modifications we need not concern our-
selves. Geographically they were entirely restricted to North
America.

The most extraordinary and puzzling of White River mam-
mals, not even excepting ProtoceraSy is Agriochcenis. Before
complete skeletons of this creature had been collected, the
skull and feet had been found separately, and were referred to
no less than three distinct and widely separated mammalian
orders. The dentition resembles that of Oreodon in the char-
acter of the canines, but the molars are much less distinctly
selenodont and have a decided resemblance to those of the
European genus Ancodus. The skull is oreodont except in a
few details, such as the absence of the lachrymal pit, the
incomplete closure of the orbit, etc. The neck is short, the

AMERICAN RUMINANT-LIKE MAMMALS. 253

backbone heavy, and the tail exceedingly long and stout, like
that of the great cats. The limb-bones are so peculiar that
they cannot be described in a few words ; suffice it to say that
they have a great many points of resemblance to those of the
carnivorous groups, and the feet are provided with great claws,
instead of hoofs, giving them a very sloth-like appearance.

If it seemed too great a demand upon our imagination to
refer Protoceras to the tylopodans, it will appear obviously
absurd to call Agriochoerus a camel ; and yet that is the direc-
tion in which the evidence, as yet unfortunately incomplete,
distinctly points. The Uinta forerunner of Agriochcerus (or
what I regard as such) and that of Oreodon have drawn so close
together as to indicate the origin of both of these families from
some common Bridger ancestor.

The selenodont fauna of the Uinta, as a whole, is obviously
ancestral to that of the White River, with the exception of
certain forms, like the anthracotheres, which had immigrated
from the Old World in the interval between the two epochs.
But we may go farther than this, and may in several instances
quite confidently point out the Uinta ancestor of a given
White River genus. In other cases there is more uncertainty,
because of less complete information, but even in these the
Uinta has yielded forms which, if not directly ancestral, are
yet very near to the ancestors sought for, and may be taken as
representing them for all practical purposes of comparison and
study.

Of these newly discovered Uinta genera one of the most
interesting is the genus described by Wortman under the
name of Protyloptis, which is unmistakably the ancestor of
Poebrotheriicm. It closely resembles the White River genus,
but, as we should naturally expect, is smaller in size and more
primitive • in structure. For example, the dentition is closed
and without diastemata ; the premolars are not much extended
antero-posteriorly ; and the molars are very short-crowned. The
canines, as in the descendant, are small — hardly larger than the
incisors. The skull, while recalling that of Poebrotherium at
the first glance, has a shorter muzzle, a widely open orbit, and a
small, hollow tympanic bulla. The neck is, most unfortunately,

254 BIOLOGICAL LECTURES.

not known. It would be of great interest to learn to what
extent the peculiar specialization of the cervical vertebrae had
proceeded. The vertebrae of the trunk and tail are like those
of the White River genus on a small scale. The limbs are
similar to those of the latter, but more primitive ; they, and
especially the feet, are less elongate ; the ulna and radius are
separate, except in old individuals ; the fibula, though extremely
slender, is still uninterrupted. In the manus are four func-
tional digits, and in the pes two, with the lateral pair reduced
to long filiform splints. Protylopus carries the main line of
tylopodan descent one stage farther back than had previously
been known, and, what is of even wider interest, it approxi-
mates this main line very distinctly to the other selenodont
groups above described, and which we have already allotted to
the tylopodan suborder.

Another genus recently named by Wortman — Leptoreodon —
is very similar to Protylopus, but with most interesting and sig-
nificant differences. It has diastemata in the dentition, large
upper canines, and caniniform first lower premolar, as in the
oreodonts and Protoceras, and also, though in rudimentary fash-
ion, in Lepto7neryx ; the limbs are rather short and the feet
tetradactyl. This genus nearly, but not quite, represents the
meeting point of the main tylopodan phylum, the Leptomery-
cidae, and the oreodonts, and greatly diminishes the gaps which
in White River times separated the three families. I regard
Leptoreodon as the probable ancestor of Protoceras, and I do not
know of any objection to such an arrangement, which would
explain the oreodont characters of the descendant genus.
These oreodont characters are so distinct that they led the late
Professor Cope to the conclusion that Protoceras was merely
an aberrant oreodont. While probable enough, the descent of
Protoceras from Leptoreodon cannot yet be proven, for the gap
between the two genera in structure and in time is still too
wide. The former has been found only in the uppermost divi-
sion of the White River stage {Protoceras beds), and we have as
yet found no ancestors for it in the middle and lower divisions.
Until these missing ancestors have been recovered, the relation
of Leptoreodon and Protoceras must remain somewhat uncertain.

AMERICAN RUMINANT-LIKE MAMMALS. 255

Cai7telo7ne7yx is another genus which differs only in minor
details from Leptoreodon and which seems to be either the
ancestor of Leptoyneryx or very near to that ancestor. In the
dentition of the White River genus we found reason to think
that it had been derived from some form in which the first
lower premolar functioned as a canine, and this condition, with
many others, is fulfilled in Camelomeryx, which is quite a small
and delicately built animal.

The Uinta ancestor of Hypertragulus cannot yet be deter-
mined, because of the possible genera none are sufficiently
well known. The genus I^eptotragiilus seems, at the present
writing, to be the most likely candidate, but Oromeryx, or even
BimomeryXy may prove to be the chosen one.

The forerunner of the White River Oreodon has long been
known and has received the suggestive name of Protoreodon,
which expresses the ancestral relation. Protoreodon is, to all
intents and purposes, an oreodont, but it has several most
interesting resemblances to the agriochoeres, such as the open
orbits, the absence of the lachrymal pit, the elongate cranium,
and the pattern of the lower molars. The upper molars have
the fifth, or unpaired, cusp in the anterior half of the crown,
as Schlosser predicted would be found — a prediction made
before the present genus was known. The skeleton is not
sufficiently different from that of its White River successor
to require any description. This genus shows us that the
oreodont family had become segregated as a distinct group in the
Uinta, but the very many likenesses of Leptoreodon to Proto-
reo doily on the one hand, and to Protylopus on the other, afford
the strongest confirmation to the opinion of Riitimeyer and
Schlosser that the oreodonts are a branch of the Tylopoda.

What I believe to be the ancestor of Agriochcerus is a Uinta
genus as yet undescribed, which I propose to name Protagrio-
ckcertis. The only known specimen of this most interesting
form belongs to the American Museum of Natural History,
and for the opportunity of making a study of it I am indebted
to the kindness of Morris K. Jesup, Esq., President. It is,
unfortunately, very fragmentary, and hence indecisive upon
certain important points, but it is, nevertheless, exceedingly

256 BIOLOGICAL LECTURES.

suggestive. The upper premolars are like those of Protoreodofty
and, as in the latter genus, the upper molars have an unpaired
cusp, but the most cursory glance at the pattern of these
molars is sufficient to recognize their essential likeness to
those of Agriochoerus . The tarsus is also sti-ongly suggestive
of the ancestral position here assigned to the genus, but it is
provoking that the ungual phalanges are all missing. Whether
or not the ancestral position rightfully belongs to Protagrio-
chcerus, it is certain that this genus and Protoreodon bring the
two families very close together and make it altogether prob-
able that both groups lead back to a common ancestor in
Bridger times.

It may well prove to be the case that some of the relations
between Uinta and White River genera here suggested will be
shown by future discoveries to be erroneous. It matters little,
however, whether we have chosen precisely the proper ances-
tors for the later forms ; there still remains the highly impor-
tant and significant fact that in this Uinta fauna all these
different families are seen to be obviously converging back to
some common term, and that they are much nearer together
than they afterwards became in White River times. It is this
fact which justifies us in maintaining the essential unity of all
the indigenous American selenodonts, diversified and highly,
specialized as many of them eventually became. Having the
great continent practically to themselves, they adopted many
rdles, which naturally resulted in a greater or less likeness to
the forms among the ancient Pecora and Tragulina which were
playing similar parts in the eastern hemisphere. This explains
the tantalising and elusive likenesses to European genera,
which have so long misled us, and shows why it was impos-
sible to make any satisfactory arrangement of these AmeHcan
genera in European families or even suborders.

When we attempt to trace these various lines back of the
Uinta, we find ourselves very much in the dark, because of our
ignorance of the more ancient forms. The little Bridger genus
Homacodon comes very near to filling the requirements of the
common ancestor of all these groups, and it is exceedingly
probable that the family of which it is the representative will

AMERICAN RUMINANT-LIKE MAMMALS. 257

eventually prove to be the fountain head whence these diverg-
ing streams were derived. Homacodon is of a very generalized
type and seems to approximate the European genus DicJiobmiey
which Schlosser regards as the remote ancestor of the Pecora.
But the European genus is not known with sufficient complete-
ness to make it plain whether it should be included in the
same family as Homacodon or not. The latter is very primi-
tive in structure, and has an unreduced dentition with rela-
tively large canines, simple premolars, and sexituberculate upper
molars which are just beginning to assume the selenodont pat-
tern ; the feet are probably pentadactyl. Nothing is known
of Homacodon which can militate against the view that it
represents a family whence were derived the various seleno-
dont lines that have become distinctly segregated in the Uinta
and widely diversified in the White River.

The probable ancestor of Homacodon is Trigonolestes {Pan-
tolestes) of the Wasatch, a little creature which, having typi-
cally artiodactyl feet, possesses an extremely primitive type of
dentition, so much so that, when only the teeth were known,
the genus was supposed to be a lemuroid. In Trigonolestes we
perhaps have the ancestor of all those selenodonts which I have
described as indigenous to North America. It is of interest
that one little character which persists throughout all the later
genera of the group is already present in the Wasatch type;
namely, the coossification of the meso- and ecto-cuneiforms in
the tarsus.

The table on the following page will show conveniently the
mutual relationships of the various selenodont genera as
presented in what has been said.

If these results are well founded, we shall have to regard
the Tylopoda as a highly important and eminently character-
istic group in the history of mammalian life on this continent,
a group which was very ancient, very peculiar, very long-lived,
and greatly diversified, and one which in these respects may
sustain a comparison with the Pecora. Admitting these facts,
it becomes well-nigh impossible to define the suborder. As
originally limited and defined by Flower, the demarcation is
easy enough, because the definition is taken from the existing

258

BIOLOGICAL LECTURES.

White River ,

Uinta

o

I

(0

1

5
•n

<LI

in

(LI

1

s,

>>

s

>>

8

35

^

w

w

Ph

Bridger

Wasatch

?Hotnacodonts

•Trigonolestes

members of the suborder, which are in many ways exceedingly
peculiar. When, however, the extinct members of the group
are taken into account, we find that every item of almost any
definition that can be framed will be transgressed by one or
other genus. This difficulty of definition is, of course, no objec-
tion to the classification proposed ; on the contrary, it is the
inevitable result of greater completeness in phylogerietic his-
tory. Definition is easy in exact proportion to the isolation of
the group defined, but with a knowledge of the history and
ancestry of a group its isolation disappears. As the bounda-
ries between connected groups grow hazy, definition grows
more and more difficult, until at last it becomes impossible.

Another necessary consequence of these conclusions is that
the Tylopoda and Pecora are but very remotely connected and
can have no common ancestor later than the middle or lower
Eocene. Hence the many characters which these two sub-

AMERICAN RUMINANT-LIKE MAMMALS. 259

orders have in common must have been independently acquired
by each. Startling as such a statement may seem to many,
there is no escape from it, even though the position which has
here been assigned to such genera as Leptomeryx^ Protoceras,
the oreodonts, etc., be an altogether mistaken one. We can
now trace the history of the main tylopodan phylum, step by
step, back to the Uinta epoch, and the Uinta representative of
the series, Protyloptts^ has no pecoran features, save only the
selenodont molars. As we may feel perfectly confident that
Protylopus is- not ancestral to the Pecora or any part of them,
it follows that characters common to the two suborders, but
not found in the Uinta genus, must have been separately de-
veloped. Riitimeyer long ago pointed out, from* a comparative
study of the recent forms, that the camels and llamas were
but distantly connected with the true ruminants, and his masterly
work on these questions is abundantly confirmed by the Ameri-
can fossils, which were very imperfectly or not at all known
when he wrote.

The conclusions which we have reached silggest the impor-
tance of giving due weight to geographical considerations in
dealing with phylogenetic and taxonomic problems. A little
study and experience enable the observer to detect the- foreign
and migrant elements in a fauna, even though they firmly
establish themselves in their new home and there give rise to
new branches. When isolated genera are found, one of the
most important questions which arises concerning them is, are
they most like the types of this or of some other continent .!*
This question rightly answered will serve as a most valuable
clue in following out the history of the genus. Riitimeyer
seems to have been guided largely by geographical consid-
erations in the statement that I have mentioned as to the
importance of the Tylopoda in North American life, and I am
thoroughly convinced of the value of such criteria. Hitherto
we have ignored them entirely too much.

I am perfectly conscious that I have been asking you to
accept a great deal upon my ex-cathedra statements ; but the
difficulty lies, not in the absence of evidence, but in the impos-
sibility of producing that evidence in court. It can be gained

26o BIOLOGICAL LECTURES.

only by a minute study of the fossils themselves. Let us
assume, however, that the conclusions so far reached are well
established and consider some of the consequences which have
an important bearing upon evolutionary philosophy in general,
(i) We see, in the first place, that parallelism and conver-
gence of development are very real phenomena, and that they
have played a highly important part in the course of evolution.
Many morphologists now accept this mode of development
unreservedly, but some still reject it altogether, or regard it
as something unusual and exceptional. However that may
be, the well-defined and established phyla of extinct mammals
demonstrate the actuality of these modes of development be-
yond peradventure. Shufifie the cards as we may, we cannot
arrange them so as to bring out any other result. One is
sometimes tempted to believe that the number of possible
tooth-patterns or osteological structures must be limited, so
often are the same ones repeated in different phyla. In many
instances we are able to follow out the history of a dental or a
skeletal structure, step by step, from its point of origin to its
final completion, and to show that the same structure has been
independently attained over and over again. So far as single
structures are concerned, this is now an old and familiar story ;
the spout-shaped odontoid process of the axis, which is so
common among hoofed animals, the double bicipital groove on
the humerus of horses, camels and giraffes, have often been
pointed out as examples of this mode of development. Only
of late, however, have we been in a position to prove that the
entire structure may be so modified in two different groups as
to keep parallel courses through long periods of time. One of
the most striking instances of this is offered by the history of
the cats. The family comprises two subfamilies, the FelincBj
or true cats, and the Machairodontince, or sabre-tooth cats ;
these two groups have been separate, at least since early
Oligocene times, when the principal family characteristics had
not yet appeared, and through the Miocene and Pliocene and
into the Pleistocene they follow parallel courses, the final genus
of one series, Smilodon, being almost identical with the terminal
genus of the other, i.e.y Felis.

AMERICAN RUMINANT-LIKE MAMMALS. 26 1

Another case of similar kind, and perhaps even more re-
markable, has been brought to light by Ameghino. Until the
junction of the two Americas, which was effected- at the close
of the Miocene, the southern continent was an extremely
isolated region and had an altogether peculiar fauna, which
differs from that of the northern hemisphere not merely in the
genera and families of its mammals, but in their orders. So
far as I am at present able to judge, the only mammalian order
common to North and South America in the late Oligocene or
early Miocene is the Rodentia. Among the peculiar Patago-
nian mammals is one which, upon a superficial examination,
seems to be an undoubted horse. The aspect of the teeth,
skull, trunk, and limb-bones is strikingly equine, and of the
feet even more so ; the feet are functionally monodactyl, a
large median toe bearing almost the entire weight, and two
much reduced lateral toes forming dew-claws, quite as in Proto-
hippiis or Hipparion. Yet, when we come to make any careful
examination of this skeleton, we soon learn that not only is it not
a horse, but that it is not even a perissodactyl. It forms a most
interesting and striking example of convergent development.

Between the Tylopoda and the Pecora we may observe
another example of curiously complete parallelism. In this
case we see how the former group firmly established them-
selves in North America at a time when communication with
the Old World had, by some means, been rendered difficult,
and how they ramified in many directions, taking here the
roles which in the eastern hemisphere were filled by the Pecora
and Tragulina. In adapting themselves to these various parts
they came to resemble the Eurasian groups in many important
respects, but if we attempt to interpret these resemblances as
having arisen from relationship and genetic affinity, we are at
once landed in hopeless confusion. The conception of the
indigenous American selenodont fauna as composing one diver-
sified suborder removes these difficulties and marshals the
families and genera in orderly array, but it involves such a
degree of parallelism in development as may stagger the belief
of those who have not made themselves familiar with the
actual steps of descent in the mammalian phyla.

262 BIOLOGICAL LECTURES.

(2) A problem as to the modes in which evolution operates,
concerning which there has been much discussion, is whether
development is always by a series of direct and unswerving
changes, each successive step in a given phylum coming, in
every detail of structure, just so much nearer to the final
result. Some years ago I had occasion to make a careful com-
parison of several successive genera in the equine phylum, and
this study led me to the following conclusion. As a rule,
development is remarkably unswerving and direct, in a large
sense, yet in minor details a certain latitude is permitted, and
in these evolution may pursue a more or less zigzag course,
with many ups and downs. This conclusion is considerably
strengthened by what we have learned concerning the main
line of descent in the tylopodan phylum. Comparing the more
ancient genera of this series with its modern representatives,
we are at once struck by the remarkable difference in their
canine teeth. In Protylopus and Poebrotheriiim the canines are
very small and may almost be called incisiform, but ih Gompho-
therium of the John Day the canines begin to enlarge, and
from that time onward these teeth become larger and larger,
until the formidable lacerating apparatus of the modern type is
attained. In the probable ancestors of the Uinta camel, Homa-
codon and Trigonolestes^ the canines are relatively long and
pointed, but even though we should exclude those genera from
the series, the analogy of all the ungulate groups would justify
the assumption that the ancestors of Protylopus had canines
which were of fairly large size and formed effective weapons.
No one can imagine that in the Uinta genus these teeth are in
their primitive condition. We are forced to infer, then, that
the canines first dwindled to very small proportions, only to
enlarge again and become formidable.

Another instance of much the same kind is afforded by the
history of the premolars. In Protylopus these teeth resemble
in general form those of the other contemporary selenodont
genera and of the White River Leptomeryx, but they are dis-
tinguished by an incipient elongation in the antero-posterior
direction. In Poebrotherium the premolars are greatly elon-
gate, carrying much farther the process which was begun in the

AMERICAN RUMINANT-LIKE MAMMALS. 263

Uinta type. This elongation of the teeth accompanies the
extension of the muzzle and, as it were, prevents the formation
of diastemata, though these appear in the later species, P.
labiaUtniy in which the growth of the rostrum outstrips that of
the premolars. In the John Day the tendency is changed, and
the premolars of Gomphotherium revert almost to the Uinta
type in form, while Procaniehis and the subsequent genera of
the phylum are remarkable for the reduction of their premolars
both in size and number.

Wortman has called attention to a third example of this fluc-
tuation in the tylopodan phylum, affecting the form of the
tympanic bulla. In Protyloptis the bulla is small and hollow,
but in Poebrotherium it has become greatly inflated and filled
with cancellous bone, the inflation especially affecting the
medial portion of the bulla. In Gomphotherium^ once more,
the direction of development is changed and the outer portion
of the bulla begins to enlarge at the expense of the inner, a
change which reaches its culmination in the existing genera.

So general al-e these minor fluctuations, that it would be
difficult to point out a single genus which in every minute
detail is exactly fitted to be the ancestor of a later genus, assum-
ing that these fluctuations do not occur. Just how great they
may be in degree we have at present no means of determin-
ing. It seems, a priori, improbable that, after a structure has
been lost or reduced to a rudimentary condition, it can ever be
regained, or become functional once more, and yet certain
cases do suggest that even such regeneration may occasionally
take place. At all events, it would be premature to deny the
possibility of changes of this character.

The features of alternating, up and down, or zigzag develop-
ment, to which attention has been called, are, after all, of a very
trifling nature. When we survey the successive and closely
connected genera of a long and crowded phylum, we cannot
fail to be impressed with the steady, orderly, unswerving
advance in all important structural features. This advance is
not always, perhaps not even usually, uniform in all parts of
the structure. One part may be accelerated and another
retarded, and what was retarded in one genus may be accel-

264 BIOLOGICAL LECTURES.

erated in a succeeding one. Thus, in the reduction of digits
the hindfoot is more rapidly modernized than the forefoot.
In ProtylopiLs^ for example, the hand has still four functional
digits, while the pes already has but two ; in the succeeding
Poebrotherium both manus and pes are in the same stage of
simplification, with two functional digits, and two small nodules
representing the rudiments of the lateral pair. But while one
structure may thus be retarded in its development and another
accelerated, the differentiation of the organism, as a whole,
keeps steadily advancing. It almost seems as if the animals
were consciously striving for a goal, though, of course, this is
only an impression given by the direct and unswerving course
their evolution takes.

There are many other ** morals " which might be drawn from
the history of the American selenodonts, but I shall lay no
further tax upon your patience. If it seems to you that I am
attributing too much importance to palaeontology and ignoring
other means of investigation, this is simply because to praise
morphology and physiology in Wood's HoU would be carrying
coals to Newcastle. The fair structure of our science must be
reared upon a broad and deep foundation, not of a single
department of inquiry, but of all of them combined and ** fitly
joined together."

FIFTEENTH LECTURE.

CASPAR FRIEDRICH WOLFF AND THE THEORIA
GENERA TIONIS.

WILLIAM MORTON WHEELER.

Mag 's die Welt zur Seite weisen,
Edle Schiller werden 's preisen,
Die an deinem Sinn entbrannt,
Wenn die Vielen dich verkannt.

Goethe, Morphologic, p. 256.

The universe which we apprehend — reducible in last analy-
sis to various sequences and coexistences in time and space —
seems to have a twofold aspect to the contemplative mind.
The minds of some men are vividly affected by the succession
of phenomena, the ceaseless current of events, the changes
that alter the complexion of the world, the great qualitative
and quantitative differences produced by these changes in that
which we call matter. These observers may note the rhythm
that is forever recurring in nature, the alternate repetition of
day and night, the return of the seasons, the cyclical recur-
rence of stages in the development of living organisms — in
short, the regular emergence from time to time of typical
forms and conditions from the flowing current of events. This
rhythm and repetition does not, however, produce the same
deep impression on these observers as the successive and
multiform changes themselves.

The other class of observer, although he may note the on-
rushing current of events, is more vividly impressed with the
similarity of the forms and conditions that recur from time to
time and from place to place. The attention is fixed on these
recurring objects and conditions, and gradually builds them
into general concepts that ultimately acquire a stability which

265

266 BIOLOGICAL LECTURES.

nothing can shake. The movement of the stream of phenomena
takes a subordinate position in consciousness, and the mental
activities attach themselves by preference to stable, island-like
forms and principles.

Thinkers from the earliest times to the present day seem to
be referable to one or the other of these two classes. The dif-
ferentiation begins in early Greek philosophy with men like
Heraclitus and Parmenides. To Heraclitus the world was an
unceasing flux — iravra pet, ovhev fievet, all things are flowing j
nothing is standing still. All things are forever becomings noth-
ing ever is. Parmenides, who fixed the trend of the Eleatic
school, belonged to the other class. He is the philosopher of
rest. The chaotic, multiform world of Heraclitus, forever in
motion, becomes for him merely a world of nonexistent appear-
ances, a shifting phantasmagoria, and only being is — the abso-
lute— the oney forever at rest.

The contrast in these two views reappears between Aristotle
and Plato. This difference is seen in the all-pervading move-
ment as conceived by Aristotle in his Physics, in contrast
with the " ideas " of Plato. Movement to Aristotle is " some-
thing very analogous to our modern biological conception of
transformation in development, for he analyzes * movement '
as every change, as every realization of what is possible." ^
Plato, on the other hand, under the influence of Parmenides
and the philosophy of rest, emphasizes the forms and qualities
that keep recurring to our minds in time and space, generaliz-
ing them into his "■ ideas " and endowing them with all the
attributes of reality.^ He would say, e.g., of a living animal as
it stands before us : " This animal as we see it does not exist in
reality, but is only an apparition, a continual becoming, a rela-
tive existence, which can as well be called nonexistent as
existent. The idea alone actually exists which is represented
in this animal, or the animal itself {avro to Orjpiov). This idea
is independent of everything; it exists by itself; it has not
become; it does not decay, but exists always in the same

1 Osborn, H. F. From the Greeks to Darwin. New York, Macmillan & Co.,

1894. p. 50-

2 See Pater, Walter. Plato and Platonism, Chaps. I, II.

THE THEORIA GENERATIONIS. 267

manner (aei 6v, kul /jLrjSeTrore ovre fytyvo/nevov, ovre airoWvfjievov).
If we can recognize the idea in this animal, it is immaterial
and unimportant whether we are looking at the animal now
before us or its ancestor that lived a thousand years ago, or
whether the animal is here or in a distant land, or whether it
appears in this or that manner, position or action, whether,
finally, it be this or another individual of the same species : 4II
this is unessential and appertains only to appearances : the idea
of the animal alone really is, and really is an object of the
understanding." ^

It would not be difficult to trace the Heraclitean conception
of the flux through Aristotle down to such modern philosophers
as Hegel and Herbert Spencer, and to trace the Platonic idea,
through the \0709 o-Tre/o/xart/co? of the Stoics, th.Q forma siibstan-
tialis and the causae primordiales of the scholastics, to Kant's
Ding-an-sich, Schelling's Absolute, and the Platonic idea as
adopted by Schopenhauer. But the tracing of these concep-
tions in detail would lead us far afield in metaphysics. I should
beg your indulgence for mentioning these matters did they not
seem to me to be, in some measure, necessary to a proper
understanding of the two great views of embryonic development
that have been and still are held by thinking students of nature
— preformation and epigenesis.

The development of the living organism is the most striking
special case of development we know. The development of
what appears to be a simple ^^g, within a comparatively short
time, and beneath our very eyes, into a complex living animal, is
development par excellence — the very perfection of that devel-
opment which is more dimly apprehended in the much slower
growth of worlds, of human societies and human institutions.
Hence we do not wonder that the development of the individ-
ual organism has become one of the main tests of two alterna-
tive views which, with a more general application, have from
the earliest times vexed philosophic thinkers.

Under the influence of the Christian church the Platonic
conception seems to have led to the notion of the special crea-

1 Schopenhauer, A. Die Welt als Wille und Vorstellung. Leipzig, Brock-
haus, 1888. Bd. i, p. 203.

268 BIOLOGICAL LECTURES.

tion of fixed types or forms. It culminated in that finished
theory of predelineation in embryonic development known as
emboitement} This was, in reality, the very negation of all
development, since the theory held that all the individuals of
a species had been created simultaneously for all time.^ In the
forcible language of the last century, Eve's ovary contained
the compressed and diminutive germs of all coming human
beings incapsulated one within the other. Such a theory
could arise only from overestimation of the definitive form
attained through development, and an underestimation of the
changes undergone by the ^gg during its development. The
typical adult form usurped the theorist's attention, and the
elaborate process whereby the type was gradually realized
shrunk to a mere unshelling and subsequent growth in size
of the next individual in order in the incapsulated series.

For the theory of emboitefnent the creation not 'only of every
species, but of every individual organism on our planet, by a
single preadamite fiat, was a necessary postulate. The rival
theory, epigenesis, implied in the cosmology of Heraclitus and
easily traceable to Aristotle, starts with a simple form of
unorganized matter, which through the agency of certain forces
undergoes the complicated changes that finally result in the
adult living organism. The homogeneous becomes the hetero-
geneous. The creation of new organisms is no longer . con-
ceived as having taken place once for all in a remote and
inscrutable past, but as taking place everywhere and at all
times. An exaggeration of epigenesis is spontaneous genera-

' 1 Passages which show the dose genetic relationship of Neo-Platonic and
Christian thought on the subject of creation are not infrequent in the writings of
the Church Fathers. The following quotations from Augustine clearly express the
idea of emboitement : " Sicut autem in ipso grano invisibiliter erant omnia simul,
quae per tempora in arborem surgerent, ita ipse mundus cogitandus est, cum Deus
simul omnia creavit, habuisse simul omnia, quae in illo et cum illo facta sunt,
quando factus est dies : non solum coelum cum sole et luna et sideribus . . . sed
etiam ilia quae aqua et terra produxit, potentialiter atque causaliter priusquam per
temporum moras ita exorentur, quomodo nobis jam nota sunt in eis operibus,
quae Deus usque nunc operatur." De Genesi ad lit., v, 45. " Omnium quippe
rerum quae corporaliter visibiliterque nascuntur, occulta quaedam semina in istis
corporis mundi hujus elementis latent." De Trinitate, iii, 8.

2 " Qui igitur systemata praedelineationis tradunt, generationem non explicant,
sed, eam non dari, affirmant." C. F. Wolff, Theoria Generationis, 1759, p. 5.

THE THEORIA GENERATIONIS. 269

tion. Aristotle even believed that mud could become earth-
worms and earthworms become eels.^

Before the end of the past century these two views of devel-
opment which I have attempted to trace back respectively to
Aristotle and Plato had assumed definite and contrasting
forms. Bonnet, Haller, and Leibnitz, following a Platonizing
tendency in dealing with natural phenomena, had elaborated
and accepted the theory of e7nboitementy or *' evolution," as
the word was then understood. Bonnet's contributions to this
view have been adequately presented by Professor Whitman, in
his lectures to the members of the Marine Biological Laboratory
during the summer of 1894.^ Haller, justly styled by his con-
temporaries an *' abyss of learning," though devoted to emboite-
ment, had too great a store of mental riches to give himself up
year after year, like Bonnet, to exhaustive rumination on a
single theory. The opinion of Leibnitz on e^nboitemetit is
not so generally known, and may be considered briefly. The
philosopher of a preestablished harmony could hardly overlook
a theory like that of predelineation. Like many philosophers
of the present day, Leibnitz was glad to accept the theories of
contemporary scientists, weave them into his general scheme,
and, without adding anything really new, again present them to
the public, heavier with the weight of his name and authority.
In his '* Monadologie," he says^: ** Philosophers have had
much difficulty in dealing with the origin of forms, entelechies,
and souls. Of late, however, careful investigations on plants,
insects and animals, have led to the conclusion that in nature
organic bodies never arise from chaos or decomposing matter,
but always from germs (semen ces), in which, without a doubty
they are already preformed. Hence we may conclude that in
this Anlage not only do organic bodies exist before generation,
but that there is a soul in these bodies, in a word, the indi-

1 Aristoteles. ^Ycropiai irepi ^ojcov. Ed. Aubert u. Wimmer. Leipzig, 1868.
ii, 6. 16. pp. 56 and 58. — J. Bona Meyer. Aristoteles Thierkunde. Berlin, 1855.
pp. 97, 98.

2 Whitman, CO. (i) " Bonnet's Theory of Evolution a System of Negations."
(2) "The Palingenesis and the Germ Doctrine of Bonnet," Biological Lectures
(1894). Boston, Ginn & Co., 1895.

3 Leibnitz, Op. Phil., p. 711.

270 BIOLOGICAL LECTURES.

vidual itself, and that reproduction is merely a means of enabling
this individual to undergo a greater change in form, to become
an individual of a different kind. Something similar to gener-
ation is seen when maggots become flies and caterpillars butter-
flies." At another place, in the *' Theodicee," he says,i after
referring to the microscopic observations of Leeuwenhoek :
" Thus I would contend, that the souls, which are some day to
become human souls, were already present in the germ like
the soiils of other species, that they have always existed in our
fore-fathers as far back as Adam, i.e., since the beginning
of things, in the form of organized bodies." These remarks
of Leibnitz are the ne phis ultra formulation of the theory of
embottement — its extension to embrace not only the physical
but also the psychical and spiritual aspect of living things.

It is, perhaps, easy to understand how philosophical and
religious preconceptions could give this final form to the theory
of embottement. Other considerations, however, of a more
real and scientific character seem to have led men's minds in
the same directions. The microscope, invented in the six-
teenth and bequeathed to the seventeenth century, had pro-
foundly influenced speculation. Magnification had revealed,
as if by magic, the existence of a great world of structures
undreamed of by the greatest intellects the race had hitherto
produced. The authority of the ancients weakened perceptibly,
for little value could thenceforth be attached to their opinions
on the nature of the great world that stretched out beyond the
confines of unaided vision. The mind, full of the great micro-
scopic discoveries of the time, was carried away by its own
inertia, and, outrunning the instrument, first dreamed of and
then believed in the existence of structures too minute to be
revealed by the available lenses. This speculation was, per-
haps, justifiable, except when it undertook to define the pre-
cise nature of what was at that time an ultra-microscopic region.
It was natural but erroneous to conceive unseen structures as
diminutive duplicates of the seen. The verisimilitude of this
error increased when it became apparent that the microscope
was unable to resolve perfectly transparent structures even of

1 Op. Phil., p. 527.

THE THE OKI A GENERA TIO ATS.

271

considerable size. And here the theorist triumphed over the
empirical observer, for he could assert, what was not easily
disproved, that owing to their transparency the microscope
must ever fail to reveal the germs incased one within the
other.

The Siegfried destined to overcome this monstrous theory of
emboitement, a theory not only false in itself, but one jealously
guarding the problem of development, and preventing all access
to it, as the dragon guarded the treasure of the Niebelungen,
was Caspar Friedrich Wolff. Wolff, one of the many great
intellects that northern Germany has produced, was born in
Berlin in 1733. You will find nearly all that is known of his
life in a letter by his amanuensis Mursinna to Goethe, pub-
lished by the great German poet in his Morphologie} The
scant facts of this letter, with Wolff's own writings, in which
his personality is studiously kept in reserve after the manner
of scientific men, leaves us with a sense of uncertainty not
entirely free from sadness. We long to know more of this
sweet-natured student who, at the early age of six-and-twenty,
was an intellectual giant, defending an epoch-making thesis, the
theoria generationis, simply pro gradu doctoris medicinae.

Before giving a brief account of this Theoria it may be
well to try to form some idea of its author's genera^ mental
characteristics. Wolff was a disciple of Aristotle. The training
of the schoolman is only too apparent in all his scientific writ-
ings, apart from Mursinna's statement ^ to the effect that when
Wolff was lecturing on medicine in Berlin " he taught logic prob-
ably better than it had ever been taught before, and applied it
in particular to medicine, thereby creating, so to speak, a new
spirit in his hearers, so that they were enabled to understand
and assimilate his other teachings more easily." His skill in
deductive logic seems to have been noticed by Sachs,^ who
claims that some of Wolff's observations on plant structure
" are highly inexact, and influenced by preconceived opinions,
and his account of them is rendered obscure and often quite

1 Goethe. Morphologie, 1820, pp. 252-256.

2 Goethe. Morphologie, p. 254.

3 Sachs. History of Botany, p. 251.

272 BIOLOGICAL LECTURES.

intolerable by his eagerness to give an immediate philosophic
explanation of objects which he had only imperfectly examined."
The same statement may be extended to many of his zoological
observations, but this is far from convincing us that his method
of investigation was at fault. Wolff's method, which did not
differ from that of the scientist of to-day, was, if anything, more
admirable than his observations. The very fact that he was
full of his Aristotelean hypothesis of epigenesis places him head
and shoulders above the investigators both of his day and of
to-day, who naively believe that they are starting their investi-
gations on a solid foundation of facts divorced from all theory.
Even Sachs admits that Wolff's phytotomical work, though
poor from the standpoint of observation, was the most impor-
tant that appeared in the period between Grew {1682) and
Mirbel (1802), " because its author was able to make some use
of what he saw, and to found a theory upon it." ^

Apart from this preconceived hypothesis of epigenesis it is
surprising with what perfect naivete Wolff approaches the phe-
nomena to be observed. Armed with his microscope, which it
does not require a Sachs to tell us " was of insufficient power
and its definition imperfect," he entered what was practically
an unknown domain, peopled only with the figments of the
predelineationists. The fascination of the growing plant and
developing embryo soon possessed him and never afterwards
left him. During his long life he returned again and again to
the study of the chick. Those who teach embryology year
after year cannot fail to appreciate Wolff's power and great-
ness when they observe the superficial impression left on
nine-tenths of the students who study the developing chick
in the well-equipped laboratories of to-day.

Wolff's instruments, poor as they were, enabled him, never-
theless, to traverse a considerable and very significant portion
of the region that lies beyond the boundary of our unaided
vision. What he saw there at once convinced him that em-
bottement was a myth. We should expect so young a man as
Wolff was when he wrote the Theoria to do two things — to
repeat his main thesis ad nauseam, and to be rather unsparing

1 Sachs. History of Botany, p. 251.

THE THEORIA GENERATIONIS.

^n

of his opponents. He did neither. His main contention is
clear enough, although not often expressly stated. He rarely
refers directly to the theory of predelineation and when he
does there is no sting in his refutation.

Stripped of many details that are somewhat wearisome to
the modern reader, the result of Wolff's observations may be
expressed in his own words taken from the very middle of the
Theoria} I translate : " In general we cannot say that what
cannot be perceived by the senses does not therefore exist.
This principle is more facetious than true when applied to
these observations. The particles which constitute all animal
organs in their earliest inception are little globules, which may
always be distinguished under a microscope of moderate mag-
nification. How, then, can it be maintained that a body is
invisible because it is too small, when t\\^ parts of which it is
composed are easily distinguishable .'* " If we of to-day read
in the place of ** globules" the word "cells," which are what
Wolff actually saw and distinguished in both plants and animals,
we shall have no difficulty in understanding how his observa-
tions disproved embottementy at least for any one who would
take the trouble to repeat them. Wolff had looked further
than the adult form and had found not a series of similar,
incapsulated embryos, but a single embryo made up of a vast
number of minute particles, the cells, closely resembling one
another, but placed side by side. There was no expanding of
a preexisting organism till it entered the field of vision, but a
host of minute and always visible elements that assimilated
food, grew and multiplied, and thus gradually in associated
masses produced the stem, leaves, stamens — in short, every
organ of the plant. This he shows in the first part of the
Theoria. In the second part, carrying on the same method, he
shows how in the animal body the heart, blood vessels, limbs,
alimentary canal, kidneys, etc., arise in a similar manner.

The third part of the work is devoted to theoretical con-
siderations. Wolff conceives living things to be constructed
like a plant, to consist of a main stem, or trunk with roots and
branches. In the embryo of the bird the umbilical duct cor-

1 Theoria Generationis, p. 72.

274 BIOLOGICAL LECTURES.

responds to the stem of the plant ; the blood vessels of the
vascular area that bring the nutriment from the yolk to the
embryo are the roots ; the organs and appendages of the embryo
correspond to the branches of the plant. ^ The organism starts
out on its development with a stem which is to connect it
with the source of nutrition on the one hand and its branches
on the other. All the substance of the embryo is originally
unorganized, inorganic. Organization sets in at one point in
the stem and thence gradually spreads to the tips of the
branches. A branch is first formed as a little bud of unor-
ganized substance, then the sap (in plants) or the blood (in
animals) flows into it from the adjacent organized part ; thus
it becomes organized, and the process continues till the organism
has acquired its definitive size and development. The blood or
sap is propelled into the unorganized substance, consisting of
globules, by a peculiar force — Wolff's vis essentialis, which is
defined in the opening chapter of the Theoria^ and was made
the special subject of Wolff's last work, written thirty years
later.2 Organization of the unorganized substance is the com-
bined result of this vis essentialis and a property which Wolff
calls solidescibilitas^ a tendency to solidify, shown most clearly
in the formation of the walls of plant cells. The vis essentialis
propels liquid nutriment into the dense unorganized matter
already present. The paths along which it flows become the
cavities of the blood vessels or plant vessels not before existent.
The liquid nutriment solidifies to form more unorganized sub-
stance, by intussusception with that already present, and the
part grows. Wolff explains the origin of the kidney which he
discovered in the chick — the Wolffian body, or mesonephros,
as we now call it — in a similar manner. Here it is the urine
that is impelled by the vis essentialis into a mass of preexist-

1 Wolff (Theorie von der Generation, 1764) compares four-footed animals to
pinnatifid leaves and " the bat is a perfect leaf — a startling statement, but, as
I have shown, the analogy is not chimerical, for the mode of origin of the two is
the sained Quoted by Huxley ("The Cell TYieoxy," Brit, and Foreign Medico-
Chir. Review, vol. xii, 1853).

2 Wolff, C. F. Von der eigenthiimlichen und wesentlichen Kraft der vege-
tabilischen sowohl als auch der animalischen Substanz. St. Petersburg,
1789.

THE THEORIA GENERATIONIS. 275

ing, unorganized substance, and the paths along which it flows
become the lumina of the uriniferous tubules and the ureter.

We know that Wolff's main error lay in grossly underestimat-
ing the complexity of the problem he attempted to solve. This
has always been a great pitfall in attempting an explanation of
life. Perhaps it is well that it is so, for Wolff would hardly have
had the heart to attempt it if he could have seen the problem with
our eyes. And may not we, too, daily commit the same blunder
when we lend a willing ear to those who regard living protoplasm
as nothing more than a " complex chemical compound " ?

Wolff accepted a simple substance as the basis of life because
he was unable to detect structure in the embryo beyond a cer-
tain limit which happened to coincide with the limits of magni-
fication of his lenses. We should suppose that Wolff would
have longed for a better lens and have at least suspected the
possible existence of some kind of structure beyond that which
he could detect. Instead of doing this, however, he writes the
following remarkable sentences which will draw a smile from
the modern searchers after centrosomes ^ : "No one has ever
yet, with the aid of a stronger lens, detected parts, which he
could not perceive by means of a weaker magnification. These
parts either have not been seen at all, or they have appeared
of sufficient size. That parts may remain concealed on account
of their infinitely small size and then gradually emerge, is a
fable." There it is in cold Latin ! Was Wolff merely nodding
when he wrote this, or was he trying to hoodwink the pre-
delineationists into believing that he had seen everything that
was worth seeing in the embryo }

In the closing paragraph of his great work on the develop-
ment of the intestinal tract, a work which appeared in 1768,
some nine years after the Theoria^ Wolff seems to rise to a
clearer perception of the complexity of the problem. He
appears to be far more doubtful concerning the way in which
simple matter becomes organized. Referring to the develop-
ment of the anterior body wall, he says^: <'This is one of the

1 Theoria, Sect. 166.

2 Wolff, C. F. Ueber die Bildung des Darmkanals im bebriiteten Hiihnchen.
Uebersetzt von J. F. Meckel. Halle, 181 2, p. 245.

276 BIOLOGICAL LECTURES.

most important proofs of epigenesis. We may conclude from
it that the organs of the body have not always existed, but have
been formed successively : no matter how this formation has
been brought about. I do not say that it has been brought
about by a combination of particles, by a kind of fermentation,
through mechanical causes, through the activity of the soul,
but only that it has been brought about."

Remaining within the province of observation which he
staked out for himself, and pursuing his excellent method,
Wolff was not only able to undermine the theoretical edifice of
the predelineationists, but also to lay the foundations for future
structures of great promise. Thus all conscientious investiga-
tion with good methods leads to subordinate facts of value
besides the main line of facts accumulated in support of the
theory in hand. Wolff was a biologist in the true sense of
the word. He regarded plant and animal life as but slightly
different aspects of a single set of phenomena. It can be
shown that he anticipated to some extent the modern theories
of protoplasm and the cell.^ According to Sachs "it was
Wolff's doctrine of the formation of cellular structure in plants
which was in the main adopted by Mirbel at the beginning of
the present century," and "the opposition which it encountered
contributed essentially to the further advance of phytotomy." ^

The theory of the metamorphosis of plants, usually attrib-
uted to Goethe, was clearly expressed by Wolff. In fact, Wolff
seems to have had clearer notions on the subject than Goethe,
according to Schleiden's statement.

To embryology Wolff made many valuable contributions,
not the least of which was his description of the formation of
the intestinal tract of the chick. This work was styled by
Carl Ernst von Baer "die grosste Meisterarbeit, die wir aus
dem Felde der beobachtenden Naturwissenschaften kennen."
It was published in Latin in the twelfth and thirteenth volumes
of the St. Petersburg Commentaries ^ where it lay buried and
forgotten till it was unearthed and translated into German by

1 Cf. Huxley.

2 Sachs. Hist, of Botany, p. 250. For the relations of Wolff's views to those
of Schleiden and Schwarm, see Huxley, The Cell Theory.

THE THEORIA GENERATIONIS. 277

the younger Meckel in 1812 and used for the purpose of refut-
ing some of Oken's erroneous views on the development of
the alimentary tract.

In general it may be said that the effect of Wolff's work on
his contemporaries was anything but immediate.^ There are
writers who even doubt the truth of the oft-repeated statement
that Wolff refuted the theory of predelineation. Sachs, e.g.y
speaking of Wolff's Theoria, says that the *' weight of his argu-
ments was not great" and that <*the hybridization in plants
which was discovered at about the same time by Koelreuter
supplied much more convincing proof against every form of
evolution." ^ We cannot lay much stress on this statement,
which seems to imply, what some physiologists seem never
to tire of implying, that evidence derived from experiment is
eo ipso more convincing than evidence derived from observa-
tion. It is certain that the predelineationists had considered
the case of hybrids, for did not the ever-watchful Bonnet
endeavor to explain the origin of the mule on the assumption
of embottement f And why should Koelreuter's plant hybrids
have more value in refuting embottement than that commonest
of all hybrids, the mule t If Sachs wishes to imply that at the
present day we should regard the evidence from hybrids as a
complete and satisfactory refutation of the theory of emboite-
ment, we may assent ; but this is not tantamount to saying that
in the latter half of the eighteenth century it was Koelreuter
and not Wolff who refuted the theory of evolution. Perhaps
it would be better to leave this question of the relative merits
of Wolff and Koelreuter to the student who has the time and
the opportunity to study all the relevant literature of the clos-
ing decades of the eighteenth century.

Wolff's position in the history of thought on the subject of
organic development becomes somewhat clearer when we com-
pare him with Darwin, for whose coming he helped to prepare

1 " Though every reader of the Theoria Generationis must see that Wolff
triumphantly establishes his position, yet, seventy years afterwards, we find even
Cuvier (Histoire des Sciences Naturelles) still accrediting the doctrine of his
opponents." — Huxley, The Cell Theory.

2 Sachs. History of Botany, p. 405.

278 BIOLOGICAL LECTURES,

men's minds. Wolff's Theoria vf^s, published in 1759; Dar-
win's Origin of Species in 1859. Wolff had been preceded by
Harvey in much the same way as Darwin was preceded by
Lamarck. Both Wolff and Darwin were ideal investigators,
patterns for all time. Darwin's love of truth, his perfect fair-
ness and modesty withal, seem to have been Wolff's possession
also. This is shown in a letter to Haller,^ thanking the great
champion of emboitement for his kindly notice of the Theoria in
his " Elementa'' : '* I thank you for wishing me well, for loving
me, sublime man, although you have never seen me, and know
me and my character only from my letters. May God reward
you for this, since I can never hope in all my life to attain
to such distinction, that I may show you worthy acknowledg-
ment of your goodness, if you will not receive in lieu of it my
everlasting veneration of your intellect. And as to the matter
of contention between us, I think thus : For me, no more
than for you, glorious man, is truth of the very greatest con-
cern. Whether it chance that organic bodies emerge from an
invisible into a visible condition, or form themselves out of the
air, there is no reason why I should wish that the one were
truer than the other, or wish the one and not the other. And
this is your view, also, glorious man. We are investigating
for truth only ; we seek that which is true. Why, then, should
I contend with you t Why should I withstand you, when you
are pressing towards the same goal as myself "i I would rather
confide my epigenesis to your protection, for you to defend
and elaborate, if it is true; but if it is false, it shall be a
detestable monster to me also. I will admire evolution, if it is
true, and worship the adorable Author of Nature as a divinity
past human comprehension ; but if it is false, you, too, even if
I remain silent, will cast it from you without hesitation."

Both Wolff and Darwin devoted their lives to the investiga-
tion of the same great problem — the development of life on

1 Epistolae ad Hallerum, October, 1766. Quoted from Alf. Kirchhoff, " Caspar
Friedrich Wolff. Sein Leben und seine Bedeutung fiir die Lehre von der orga-
nischen Entwickelung," /<?«. Zeit. f. Med. u. Naturwiss., Bd. iv, Heft i, 1868,
pp. 193-220. This valuable essay has been of great assistance in the preparation
of my lecture.

THE THE OKI A GENERATIONIS. 279

our planet. Both found answers to their respective parts of
this problem. Wolff published his answer when he was very
young ; Darwin waited till he was well along in years. Each
was confronted by a formidable, clearly formulated theory of
special creation. The theory that confronted Wolff was spe-
cial creation of all individual organisms by a preadamite fiat.
Darwin was confronted by a theory of the special creation of
all the species at the same inscrutable time. This view had
found favor with such eminent men as Linne, Cuvier, L.
Agassiz, Owen, and the numerous systematists who followed
in their footsteps, making species Platonic ideas, just as indi-
vidual organisms had been made Platonic ideas in Wolff's day.
During the closing half of the eighteenth century it became
clear to thinking men that individual organisms always have
an epigenetic origin from preexisting individuals. The closing
half of the present century has been consumed in demonstrat-
ing that species always arise from preexisting species.

Both Wolff and Darwin collided with prevailing theological
views. Darwin's experience in this matter is well known, and
perhaps the less said about it the better. It is not so generally
known that Wolff's failure to establish himself as a professor
in Germany, and his departure in 1769 for St. Petersburg,
where he spent the remainder of his life, was probably due not
only to professional jealousy, but also to a certain antagonism
on the part of religious contemporaries. In his letters Haller
often warned Wolff of the dangers of his views to religious
dogma, and endeavored to persuade him to abandon them "on
grounds of utility." ^ Just before leaving for St. Petersburg
Wolff wrote the following to Haller : " There is, of course, no
reason why a divine being should not exist, even if organic
bodies are formed by natural forces and through natural causes ;
for these very forces and causes, yes. Nature herself, has as
much need of an Originator as organic bodies; still the evi-
dence would be far more cogent and apparent, if we should find
from contemplation of Nature that her individual products, the
organic bodies, required a Creator, and that nothing organic
could be produced through natural causes." Does not this

1 Kirchhoff, A., I.e.

28o BIOLOGICAL LECTURES.

remind us of the following passage in the last chapter of
Darwin's Descent of Man ? " I am aware that the conclusions
arrived at in this work will be denounced by some as highly
irreligious ; but he who denounces them is bound to shew why
it is more irreligious to explain the origin of man as a distinct
species by descent from some lower form, through the laws of
variation and natural selection, than to explain the birth of the
individual through the laws of ordinary reproduction. The
birth both of the species and of the individual are equally parts
of that grand sequence of events, which our minds refuse to
accept as the result of blind chance. The understanding
revolts at such a conclusion, whether or not we are able to
believe that every slight variation of structure, — the union of
each pair in marriage, — the dissemination of each seed, — and
other such events, have all been ordained for some special
purpose."

Both Wolff and Darwin left their theories unfinished. They
maintained a transformation of simpler into more complex mat-
ter, but they did not succeed in demonstrating how this trans-
formation is accomplished. I have already quoted Wolff's con-
fession of ignorance of the way in which epigenetic develop-
ment is brought about. The doubts entertained by Darwin and
his successors concerning the adequacy of natural selection
as a complete explanation of descent are familiar to us all.
The absolute completeness of the old embottement and special
creation hypotheses doomed them to a speedy death. Wolff's
and Darwin's hypotheses have lived because they represented
only parts of a great truth. On this account, also, they have
supplied and will continue to supply powerful incentives to
investigation.

Both the theory of epigenesis and the modern theory of
descent are manifestly imbued with the old Heraclitean and
Aristotelean conception of heterogeneity arising from homo-
geneity in a continual flux of events. We have come to
regard this as the essence of evolution. We still repeat Her-
bert Spencer's definition^: ''Evolution is an integration of

1 Spencer, Herbert. First Principles. New York, Appleton & Co., 1886.
p. 396.

THE THEORIA GENERATIONIS. 28 1

matter and a concomitant dissipation of motion ; during which
the matter passes from an indefinite, incoherent homogeneity
to a definite, coherent heterogeneity ; and during which the
retained motion undergoes a parallel transformation." If
Wolff could have read this sonorous definition he would prob-
ably have accepted it as the expression of a general truth, as it
is accepted to-day. Still when we compare our views on the
development of living organisms with those of Wolff, we observe
a vast difference, to which Professor Whitman calls attention
when he says^: ''The indubitable fact on which we now build
is no bit of inorganic homogeneity, into which organization is to
be sprung by a coagulating principle, or cooked in by a calidum
innattim, or wrought out by a spinning archaeus, but the ready-
fonnedy livmg germ^ with an organization cut directly from a
preexistings parental organization of the same kind.

" The essential thing here is, not simply continuity of germ-
substance of the same chemico-physical constitution, but actual
identity of germ-organization with stirp-organization^

To-day we recognize three conditions of matter : dead
matter, undifferentiated living protoplasm, and differentiated
living protoplasm. The germ is cut from its parent as organ-
ized but undifferentiated living protoplasm ; during ontogeny it
is converted into differentiated living protoplasm. Wolff saw
no difficulty in leaping with a bound from dead matter to
highly differentiated living protoplasm. There are scientific
acrobats still living who are not at all afraid of taking a like
hop, skip, and jump. The majority of biologists, however, are
too heavy with the past century's observations on living matter
to be able to leap so fast and so far. They find the distance
between dead matter and undifferentiated protoplasm enor-
mous — a chasm so wide and deep that they prefer making a
long detour to attempting a perilous leap across it. They are
more intrepid in passing over the gap, great as it undoubtedly
is, which separates the undifferentiated from the differentiated
phase of organized matter, the germ from the adult. Current
discussion is, therefore, mainly limited to the valuation of the

1 Whitman, C. O., " Evolution and Epigenesis," Biological Lectures (1894).
Boston, Ginn & Co., 1895. p. 212.

282 BIOLOGICAL LECTURES,

extent of complexity in the germ as compared with the com-
plexity of the fully developed organism.

He who finds little difficulty in passing from the simple to
the complex, from the homogeneous to the heterogeneous, will
take an epigenetic view of development. The physiologist, who
deals with processes, who is ever mindful of the Heraclitean
flux, inclines naturally to this view. On the other hand, he
who readily idealizes and schematizes, whose mind is endowed
with a certain artistic keenness, an appetite for forms and
structures, and a tendency to make these forms final patterns,
eternal molds, more permanent than the substance that is
poured into them — such a one will find more difficulty in
understanding how the homogeneous can become the hetero-
geneous. Of this type is the modern morphologist who is con-
tinually diagrammatizing, who has his eye fixed on complex
static structures and conceives the continually changing form
of the developing ^g'g as a series of kinematograph pictures in
three dimensions of space. He is as much inclined to Plato-
nize as is the modern physiologist to reason along lines sug-
gested by Aristotle. He is by nature a preformation ist.

Just as Wolff's followers have split into two schools — one
believing in little, the other in much preformation in the germ
— so Darwin's followers have split into two schools, the Neo-
Lamarckians and the Neo-Darwinians, in obedience to the two
psychological tendencies to which I have called your attention.
The Neo-Darwinians, in laying great stress on the segregation,
stable and complex intrinsic structure of the germ plasma
and its importance as a vehicle of hereditary characters, and in
attributing less value to the extrinsic factors, like food and
environment, are allying themselves with theorists of the
type of Parmenides and Plato. On the other hand, the Neo-
Lamarckians who believe in the permanent change-produ(!ing
effects of the extrinsic factors (environment, etc.) on structure,
and attribute less value to the architecture of an intrinsic vehi-
cle of heredity (germ plasma), range themselves with Heraclitus
and Aristotle.

The amount of differentiation displayed during the ontogeny
of an organism or during its phylogenetic history will be

THE THEORIA GENERATIONIS. 283

differently estimated by different workers, till we are in posses-
sion of some means of mathematically measuring differentia-
tion and variation. The demand for mathematical measurement
is already being made in certain quarters, and this demand
progressing science will undoubtedly supply. At present we
are quite adrift in our discussions, so long as we ignore the
more general and philosophical aspect of the question and
thereby overestimate its simplicity. Even if we accept differ-
entiation and the interaction of differentiated products as the
root ideas of the evolution of the individual and of the race,
we are still at a loss to understand how the initial differentia-
tion arose — how the homogeneous first became the heteroge-
neous. Lloyd Morgan has recently expressed this hopeless
and baffling search for initial differentiation as follows ^ : —

" Assuming, with the nebular hypothesis, a primitive fire-mist, we
must assume also an environment from which it is already differen-
tiated and to which its heat energy is communicated by radiation.
Or if we accept the meteoric hypothesis, we must grant the existence
of already differentiated cosmic dust and the interaction of its con-
stituent meteors. If we give yet freer rein to the speculative tend-
ency, which, chastened or running riot, is man's blessing or curse,
and, straining our mental vision, search deeper still into the begin-
nings of our universe, to find in the homogeneous substance that
Sir William Crookes calls protyle, the stuff from which the chemical
elements were differentiated ; even in this dim and wholly hypothet-
ical region we are forced to assume, as the antecedent conditions of
differentiation, transformations and redistributions of energy, imply-
ing a prior differentiation to render such interaction conceivable.
Or if, once more, we conceive the elemental atoms as vortex rings,
differentiated from the ether and thenceforth interacting, even here
at the very threshold of differentiation, we seek for an answer to the
question : Under what physical conditions did such vortex motion
originate ? "

1 Morgan, Lloyd, "The Philosophy of Evolution," Monist, July, 1898,
p. 489. Weismann, too, expresses this difficulty in his " Germinal Selection " :
" Die sog. ' epigenetische ' Theorie mit gleichen Keimeseinheiten ist deshalb
eigentlich nichts Anderes, als eine Evolutionstheorie mit unbewusster Zuriick-
verlegung der Anlagen in die Molekiile und Atome, eine, wie mir scheint, unstatt-
hafte Vorstellung. Eine wirkliche Epigenese aus vollig gleichartigen, nicht bloss
aus untereinander ^/<?/V./^^« Einheiten ist nicht denkbar."

284 BIOLOGICAL LECTURES.

The pronounced '* epigenesist " of to-day who postulates
little or no predetermination in the germ must gird himself to
perform Herculean labors in explaining how the complex
heterogeneity of the adult organism can arise from chemical
enzymes,^ while the pronounced ** preformationist " of to-day
is bound to elucidate the elaborate morphological structure
which he insists must be present in the germ. Both tenden-
cies will find their correctives in investigation.

1 See, e.g., Loeb, J., "Assimilation and Heredity," Monist, July, 1898, p. 555.

SIXTEENTH LECTURE.

ANIMAL BEHAVIOR.

C. O. WHITMAN.

" Natura non facit saltum, is applicable to instincts as well as to corporeal structure."
Darwin, Origin of Species, p. 231.

CONTENTS. ^^^^

Introductory 286

Behavior of Clepsine 287

a. Deceptive Quiet 287

b. Rolling into a Ball 289

1. Provoked by Exposure 289

2. Forced by Attack 290

3. Induced by Gorging Blood 290

4. Origin and Utility 290

c. Sensitiveness to Light 293

Instinct of Rolling into a Ball among Insects 293

Behavior of Necturus 295

a. Refusal of Food from Fear 295

b. Behavior of Young in Taking Food 296

c. Influence of Innate Timidity 298

d. Organization Shapes Behavior 298

e. Origin and Meaning of Behavior 299

f. Sensibility — Sources of Error 300

g. Orientation through the Dermal Sensillae 302

k. Origin and Nature of the Behavior in Taking Food 304

1. Some Intelligence Implied 304

2. Orientation Learned by Experience 305

3. Deferred Instinct 305

4. Pause before the Bait 306

5. Meaning and General Occurrence 308

6. Part Played by Fear 309

285

286 BIOLOGICAL LECTURES.

PAGE

General Considerations 310

a. Instinct Precedes Intelligence 310

b. Theories of Instinct 311

1. Pure Instinct the Point of Departure 311

2. Embryology and the Lamarckian Theory 312

3. Darwin's Refutation of Lamarck's Theory. 313

4. Weak Points in the Habit Theory 314

c. Two Demonstrations of the Habit Theory Claimed by Romanes 314

1. The Instinct of Pouting 316

2. The Instinct of Tumbling 316

d. The Habit Theory Losing Ground 318

e. Hyatt on Acquired Characters 320

f. Preformation the Essence of the Doctrine 321

The Genetic Standpoint in the Study of Instinct 322

a. Genealogical History Neglected 322

b. The Incubation Instinct 323

1. Meaning to be Sought in Phyletic Roots 323

2. Means Rest and Incidental Protection to Offspring 324

3. Essential Elements of the Instinct 324

(i) Disposition to Remain over the Eggs 325

(2) Disposition to Resist Enemies 325

(3) Periodicity 327

A Few General Statements 328

Instinct and Intelligence 331

a. Experiment with Pigeons 332

b. The Step from Instinct to Intelligence -t^t^t^

1. The Passenger Pigeon 334

2. The Ring-neck Pigeon 334

3. The Dovecot Pigeon 335

4. Results Considered 335

Animal behavior, long an attractive theme with students
of natural history, has in recent times become the centre of
interest to investigators in the field of pyschogenesis. The
study of habits, instincts, and intelligence in the lower ani-
mals was not for a long time considered to have any funda-
mental relation to the study of man's mental development.
Biologists were left to cultivate the field alone, and psycholo-
gists only recently discovered how vast and essential were the
interests to which their science could lay claim.

ANIMAL BEHAVIOR. 287

The contribution which I have to offer aims at no extensive
exposition of the subject, but rather to call attention to some
phenomena which I have observed, and to connect therewith
such interpretations and theoretical considerations as may come
within the sphere of general biology.

In animal life there are many interesting modes of keeping
quiet, which are instinctive and adapted to special purposes.
It is a very general means of concealment and escape from
enemies. For illustration, we may take first the leeches, ani-
mals relatively low in the scale. One of the lower and least
active forms, occurring everywhere in ponds, lakes, and streams,
is known under the generic name Clepsine. There are many
species, varying from one-quarter inch to one inch or more in
length. They are found on their regular hosts, turtles, frogs,
fishes, molluscs, etc., or on the under side of stones, boards,
branches, or other submerged objects near the shore. One of
the larger species, found often in large numbers on turtles, will
be favorable for observation.

Behavior of Clepsine.

a. Deceptive Quiet.

Place the animal in a shallow, flat-bottomed dish and leave it
for a few hours or a day, in order to give it time to get accus-
tomed to the place and come to rest on the bottom. Then,
taking the utmost care not to jar the dish or breathe upon the
surface of the water, look at the Clepsine through a low magni-
fying lens and see what happens when the surface of the water
is touched with the point of a needle held vertically above the
animal's back. If the experiment is properly carried out, it
will be seen that the respiratory undulations (if such move-
ments happen to be going on) suddenly cease and that the
animal slightly expands its body and hugs the glass. Wait
a few moments until the animal, recovering its normal compo-
sure, again resumes its respiratory movements. Then let the
needle descend through the water until the point rests on
the bottom of the dish at a little distance from the edge of
the body. Again the movements will cease and the animal

288 BIOLOGICAL LECTURES.

will hug the glass with its body somewhat expanded. Now
push the needle slowly along towards the leech, and notice, as
the needle comes almost in contact with the thin margin of the
body, that the part nearest the needle begins to retreat slowly
before it. This behavior shows a surprising keenness of tactile
sensibility, the least touch of the water with a needle-point
being felt at once. This delicate sensitiveness is manifested
in such a quiet way that it would be generally overlooked, and
an observer unfamiliar with the habits of Clepsine, and not
realizing the necessity of extreme quiet in his own movements,
would almost certainly draw false conclusions. If the dish
were moved or the water carelessly disturbed in any way, the
Clepsine would assume its motionless attitude and appear to be
wholly indifferent to the disturbance. If its back were rubbed
with a brush or the handle of a dissecting needle, in order
to test its sensitiveness to touch, the appearance would prob-
ably still be that of insensibility and indifference to the treat-
ment. Closer examination, however, would show that the flesh
of the animal was more rigid than usual, and that the surface
was covered with numerous small, stiff, conical elevations, the
dermal papillae or warts, which are so low and blunt in the
normal state of rest as to be scarcely visible. It would be
seen that the animal, although motionless, was in a state of
active resistance to attack. Every muscle would be strained ;
the whole skin would be tense and rough with the stiff, pointed
papillae ; and at the same time the body would be found exces-
sively slippery ^nd difficult to lay hold of, owing to the mucous
secretion poured forth from the dermal glands. To guard still
further against dislodgment, the body would be flattened out as
much as possible and tightly applied to the glass. The activity
of the resistance offered by this passive-looking creature would
be very forcibly realized if the observer attempted to circum-
vent it by slipping a thin blade or spatula beneath it with a
view to forcing its hold. If overcome in one part it would
stick by another, and skillful manipulation would be necessary
to get both ends free at the same time. With one end detached,
the other will often hold against a pull strong enough to snap
the body in two.

ANIMAL BEHAVIOR. 289

b. Rolling into a Ball.

Clepsine has another and entirely different method of keeping
quiet. The behavior bears striking analogy to that which has
been described as "feigning death" in some insects. The
animal rolls itself up (head first and ventral side innermost)
into a hard ball, outwardly passive, and free to roll or fall
whithersoever gravity and currents in the water may direct it.
The ball will bear considerable pressure and rough handling
without unfolding or exhibiting any marked movements. Left
in quiet for a few seconds, the animal slowly unrolls itself and
creeps off. This instinct has many advantages for a slow-
moving creature like Clepsine^ as will presently be seen.

I . Provoked by Exposure. — The ball-like attitude is assumed
under various circumstances. If a stone or board with Clepsine
attached to its under surface be quietly turned upside down,
thus bringing the leech from shade or darkness into light and
exposure, it may sometimes maintain its position of rest un-
changed, only hugging the stone a little more closely and not
moving until all is quiet. More generally, however, it rolls
itself up, and by the time the stone is turned, or before, it falls
to the bottom, where it can unfold and escape without danger
of discovery. If, by chance, the animal has eggs, it will not
desert them to escape in this way. As soon as the eggs hatch
and the young become attached to the ventral side of the
parent, the latter may roll itself up with its brood inside,
fall to the bottom as before, and thus escape with all its
progeny.

This species, then, has two quite distinct and peculiar ways
of keeping quiet and thus avoiding its enemies. If the animal
has no eggs, or if it has young, it may adopt either mode of
escape, while if it has eggs it has no choice but to remain
quiet over them. In the species here considered, the eggs are
held together in a thin, gelatinous sheet, secreted at the time
of ovipositing, and of a size and form to be entirely covered
by the expanded body of the parent. In some species of Clep-
sine the eggs are laid in thin membrane-like sacs, which are
fastened to the under side of the parent, and in this case the

290 BIOLOGICAL LECTURES.

rolling up into the form of a ball is the safest course of behavior
and the one generally adopted.

2. Forced by Attack, — The same behavior will almost inva-
riably follow when any species of Clepsine is closely pursued
and finds itself unable to fix itself by either end, as when a
spatula is repeatedly thrust under it in such a way as to break
its hold and defeat its efforts to regain a footing.

3. Induced by Gorging Blood. — The provocations to such
behavior thus far considered have all been such as might, and
probably do, cause more or less alarm. It is important to note,
however, that the instinct may manifest itself frequently under
conditions that seem to exclude the influence of fear. I have
often seen these leeches fold themselves into balls at the end
of a good meal, and so roll to the edge of the shell of their
host, and fall to the bottom. This mode of concluding a quiet
repast, with no assignable cause for alarm, and with every
reason for satisfaction and contentment, except for the desire
to get out of light into darkness, under cover of a stone or
some other object, will hardly pass as feigning; and cataplexy
and the tropisms are equally out of question. We could not
assume, for example, that Clepsine is positively heliotropic
when hungry, and negatively heliotropic after feeding. Shade
is preferred at all times in both conditions. If hungry, Clep-
sine leaves the shade, not because it prefers light, but because
it prefers its host more than it prefers shade. If the host is
not found it will again return to a shady retreat, if one is to
be found, however hungry it may be. The rolling up cannot
be attributed to light, as the animal takes the extended posi-
tion when at rest, even if compelled to remain in the light.
What, then, shall we conclude }

4. Origin and Utility. — Observation and inference may be
stated as follows :

I. The act of rolling up into a passive ball may be performed
{a) nnder compulsion, as when it is the last resort in self-
defence ; {b) nnder a milder provocation, as 07ie of three courses
of behavior, as when the resting place is turned up to light,
and the choice is offered between remaining quiet in place,
creeping away at leisure, or rolling into a ball and dropping to

ANIMAL BEHAVIOR. 29 1

the bottom ; {c) or, finally, under no special external stimulus,
but rather from internal motive, the normal demand for rest
and shady seclusion, presumably very strong in Clepsine after
gorging itself with the blood of its turtle host.

2. This mode of taking leave of the turtle, after a full meal,
is the easiest, the quickest, and the safest way available. To
drop off fully extended, as Clepsine sometimes does, would
retard descent and increase the chances of capture by fish.
To creep about on the back of the host, waiting for an oppor-
tunity to grasp a plant or stone, would be decidedly hazardous,
for if it came within the stretch of its host's neck, annihilation
would be almost certain, while if lucky enough to keep out of
reach of its host, it would still be in danger of the same fate
from other turtles.

3. This behavior is instinctive, since it is performed by the
young after the first meal as perfectly as by the adult.

4. Looking more closely at the nature and origin of this
instinct, it will be seen to be quite a natural performance, in
keeping with the most fundamental features of the animal's
organization, and only a special application of a more general
act that is primary and organic as much as tasting, seeing, or
sleeping.

The more general act consists simply in tucking or rolling the
head under, as often happens when the animal is resting. The
habit may be observed in the young as soon as they are suffi-
ciently developed to be capable of bending the tip of the head
under. The same act, carried a little further, gives the half-
rolled condition, in which only the anterior of the animal is
folded, while the posterior portion remains unrolled and attached
by the sucker. This attitude is often assumed if the leech is
sick or has been injured. It is only a step further to release the
sucker and fold it over the part already rolled up, thus com-
pleting the part ball to a whole ball, which can move passively
more rapidly and safely than is possible by active creeping.
From beginning to end we have only one act, in different stages
of completion, simply different degrees of one and the same
process.

5. Having the general act to start with, it is easy to see

292 BIOLOGICAL LECTURES.

how it might be made of use for particular purposes ; in other
words, how special adaptations of a useful kind might arise. If
the act is a natural concomitant of the resting condition, and is
associated with a feeling of ease and security, we see how sick-
ness, injury, fear, a heavy meal, etc., might prompt it, and in
higher or lower degree, according to the nature and intensity
of the inciting cause. Full and prompt action under exposure,
pressure, injury, and in the event of a good meal, would carry
decisive advantages, so that individuals reacting in the more
favorable degrees would stand the best chance of escape and
survival. Natural selection would steadily improve upon the
results, and the special adaptation, in different stages of devel-
opment in different species, as we find it to-day in different
ClepsineSy would lie in the direct line of progress. This view
does not of course presuppose intelligence as a guiding factor,
and therefore lends no support to the theory of instinct as
"lapsed intelligence," or "inherited habit."

6. An instinct of the kind here considered does not depend
for its development upon effort and the transmission of func-
tionally acquired improvements in organization, but upon the
natural selection of the best qualified germs, for that is what
the survival of the fittest individuals always means. Many
species of Clepsine require but one full meal a year, and as they
seldom live more than two or three years, the number of meals
is very limited. There is little room, then, for repeating the
experiment often enough to affect the organization. Indeed,
such a supposition would here appear to be absurd in the last
degree. On the other hand, the selection of the fittest germs,
provided for in the survival of the best-adapted individuals,
would inevitably advance the species along the line leading to
the special instinct.

If the view here taken be correct, the instinct of rolling into
a ball is not a matter of deliberation at all, but merely the
action of an organization more or less nicely adjusted to special
conditions and stimuli. Intelligence does not precede and
direct, but the indifferent organic foundation with its general
activities is primary ; the special behavior or instinct is built
up by slowly modifying the organic basis.

ANIMAL BEHAVIOR. 293

c. Sensitiveness to Light.

The question as to how much intelligence, if any, Clepsine
may have, I do not here undertake to settle or discuss. That
the animal is endowed with keen sensibilities is evident from the
behavior before described. The following simple experiment
affords a striking demonstration : Pass the hand over a dish in
which a number of Clepsines are resting quietly on the bottom,
at a distance of a few inches above the animals, taking care not
to make the least jar or other disturbance. If the animals are
quite hungry, the slight shadow of the hand, imperceptible
though it be to our eyes, will be instantly recognized by them,
and a lively scene will follow, every leech rising up, sup-
ported on its posterior sucker, and swinging at full length
back and forth, from side to side, round and round, as if in-
tensely eager to reach something. Put a turtle in the dish and
see what a scramble there will be for a bloody feast. The
shadow of the hand was to these creatures like the shadow of
a turtle swimming or floating over them in their natural haunts,
and hence their quick and characteristic response. A piece of
board floating over them would have the same effect. Although
so sensitive to a small difference in light, the Clepsine eyes can
give no pictures, and hence there is no power of visual discrimi-
nation between objects. They probably recognize their right
host by the aid of organs of taste ; at any rate they are often
able to distinguish their host from closely allied species.

Instinct of Rolling into a Ball among Insects.

The following examples of the instinct of rolling into a ball
among insects are from Kirby and Spence.^

'* I possess a diminutive rove-beetle {Aleochara cotnplicans
K. Ms.), to which my attention was attracted as a very minute,
shining, round, black pebble. This successful imitation was
produced by folding its head under its breast, and turning up its
abdomen over its elytra, so that the most piercing and discrimi-
nating eye would never have discovered it to be an insect. I

1 Entomology, pp. 411, 412.

294 BIOLOGICAL LECTURES.

have observed that a carrion beetle {Silpha thoracica) when
alarmed has recourse to a similar manoeuvre. Its orange-
colored thorax, the rest of the body being black, renders
it particularly conspicuous. To obviate this inconvenience, it
turns it head and tail inwards till they are parallel with the
trunk and abdomen, and gives its thorax a vertical direction,
when it resembles a rough stone. The species of another
genus of beetles {Agathidiiim) will also bend both head and
thorax under the elytra, and so assume the appearance of
shining, globular pebbles.

" Related to the defensive attitude of the two last-mentioned
insects, and precisely the same with that of the Armadillo
{Dasypics) amongst quadrupeds, is that of one of the species of
wood-louse {Annadillo vulgaris). The insect, when alarmed,
rolls itself up into a little ball. In this attitude its legs and
the underside of the body, which are soft, are entirely covered
and defended by the hard crust that forms the upper surface
of the animal. These balls are perfectly spherical, black, and
shining, and belted with narrow white bands, so as to resemble
beautiful beads ; and could they be preserved in this form and
strung, would make very ornamental necklaces and bracelets.
At least so thought Swammerdam's maid, who, finding a num-
ber of these insects thus rolled up in her master's garden, mis-
taking them for beads, employed herself in stringing them on a
thread ; when, to her great surprise, the poor animals beginning
to move and struggle for their liberty, crying out and running
away in the utmost alarm, she threw down her prize. The
golden wasp tribe also {Chrysididce), all of which I suspect to
be parasitic insects, roll themselves up, as I have often observed,
into a little ball when alarmed, and can thus secure themselves
— the upper surface of the body being remarkably hard, and
impenetrable to their weapons — from the stings of thosQ Hymen-
optera whose nests they enter with the view of depositing
their eggs in their offspring. Latreille noticed this attitude
in Parnopes carnea^ which, he tells us, Bembex rostrata pursues,
though it attacks no other similar insect, with great fury ; and,
seizing it with its feet, attempts to dispatch it with its sting,
from which it thus secures itself. M. Lepelletier de Saint-

ANIMAL BEHAVIOR. 295

Fargeau, to whom entomology is indebted for so many new
facts relative to the manners of hymenopterous insects, has
given us a striking account of a contest between the art of one
of these parasites (Hedychrum regiiim) and the courage of one
of the mason-bees in endeavoring to defend its nest from its
attack. The mason-bee had partly finished one of her cells,
and flown away to collect a store of pollen and honey. During
her absence the female parasitic Hedychrum, after having exam-
ined this cell by entering it head foremost, came out again, and
walking backwards, had begun to introduce the posterior part
of her body into it, preparatory to depositing an ^gg, when the
mason-bee arriving laden with her pollen-paste threw herself
upon her enemy, which, availing herself of the means of defence
above adverted to, rolled herself up into a compact ball, with
nothing but the wings exposed, and equally invulnerable to the
stings or the mandibles of her assailant. In one point, however,
our little defender of her domicile saw that her insidious foe was
accessible ; and, accordingly, with her mandibles cut off her
four wings, and let her fall to the ground, and then entering
her cell with a sort of inquietude, deposited her store of food,
and flew to the fields for a fresh supply ; but scarcely was she
gone before the Hedychrum^ unrolling herself, and, faithful to
her instinct and her object, though deprived of her wings, crept
up the wall directly to the cell from whence she had been pre-
cipitated, and quietly placed her ^gg in it against the side below
the level of the pollen-paste, so as to prevent the mason-bee
from seeing it on her return."

Behavior of Necturus.

a. Refusal of Food from Fear.

Our large fresh-water salamander, popularly called mud-
puppy, water-dog, hellbender, etc., is another animal that
may be profitably studied with reference to its modes of quiet.
The first adults which I kept in captivity in a large aquarium
refused to eat pieces of raw beef or small fish, whether dead
or alive. For months they went on, seemingly entirely indif-

296 BIOLOGICAL LECTURES.

ferent to any proffered food, not paying the least attention, so
far as I noticed, to tempting morsels dropped quietly in front
of them or held in suspension before them. Living earth-
worms and insect larvae were presented to them, all of which
were known to be palatable to the creature in its natural
habitat ; but nothing availed to draw attention or elicit any
evidence of hunger. Quiet and wholly indifferent in outward
behavior, yet the animals were actually starving and wasting
away. Were the creatures feigning quiet and indifference.^
Or was the behavior merely the expression of timidity, the
animal not having the courage to perform the acts necessary
to secure the food which it must have craved .-* I confess that
I did not for a long time understand the cause of this refusal
of food.

Further acquaintance with the adults, supplemented by an
experience of two seasons in rearing the young, opened my
eyes to the extreme timidity of these animals, which is so
deep-seated and persistent that one can form only a poor idea
of it without considerable actual contact with it. The outward
behavior is very quiet and mild and gives little indication of
fear. The animal will often submit to gentle handling without
making any violent effort to escape. In short, the behavior
is misleading, and one stands no chance of understanding it
until he learns to keep quiet himself while observing, and
discovers how to get into confidential relations with the crea-
ture. This can be done with the adults, but to better advan-
tage with the young.

b. Behavior of Young in Taking Food.

The eggs may be readily hatched in a shallow dish and
young thus obtained which have never learned anything from
the parents. I had about fifty young hatched in this way
towards the end of July. When first hatched they were loaded
with food-yolk sufficient to meet their needs for about two
months. By the end of September I began to get intima-
tions of a desire for food. The method of feeding was as
follows : The dish containing the young was kept on a table,

ANIMAL BEHAVIOR. 297

where, without being moved, food could be offered in perfect
quiet. I used the tiniest bits of raw beef and offered only one
piece at a time, which I held in small forceps or on the point
of a needle a little in front of the animal to be tested. If the
meat is held closely enough to touch the head, the animal is
frightened and may retreat with such haste as to alarm all its
companions. If the bait is held a little to one side, an inch or
so away, and very quietly for a minute or more, a slight turn-
ing of the head in that direction may be noticed, in case the
animal is ready to eat and feels confidence enough to try to
reach it. The turning of Ihe head is done very cautiously and
almost as slowly as the minute-hand of a clock moves, so that
one may become aware of it, not by seeing the movement, but
by noticing the inclination of the head to the axis of the body.
If there be a decided turn of the head of this kind, the case is
hopeful, as it shows an interest which may be encouraged to
action by bringing the bait a little nearer, but very slowly and
without any jerky movement. Halting about half an inch
away, wait for further movement on the part of the animal, if
you are fortunate enough not to have frightened it away. If
the animal's courage holds out — in most cases it does not in
the first trials — it will soon begin to move, but with a slow-
ness that tries the observer's patience. The head at length
comes up to a point a quarter of an inch away, more or less,
and after making sure of the position of the bait, which seems
to be done less by the aid of the eyes than by the sense of
touch, the animal tries to seize it by a quick side movement of
the head and a snap of the jaws. The first attempt to take the
bait corresponds in all essential points with the behavior of the
adult when trying to capture a fish, a worm, or an insect larva,
although the aim may not be quite so sure.

If one is successful in getting one or more to feed, the
more timid ones may be brought forward in the same way by
patiently alluring them from day to day, until they are tempted
to an effort. Once made, the effort becomes easier at the
next trial, and in the course of a month or six weeks the
bolder ones will respond fairly promptly. A few of my speci-
mens became very familiar with me, and would come towards

298 BIOLOGICAL LECTURES.

me when I approached the dish, looking up at me wistfully,
as if knowing well the meaning of my visits.

c. Influence of Innate Timidity.

In the behavior above described, we see an instinctive mode
of capturing prey held in check, and probably directed to some
extent, by innate timidity. Fear seems to be the main factor
in control at the start, holding the animal in a trance-like
quiet, undecided as to what to do, waiting for confidence to
attack, or for a stronger motive to flee. As fear subsides a
little, the preparatory movement of attack begins, but the sly
behavior is due to fear rather than the slyness of stratagem.
The slow and cautious method of approach is certainly not all
finesse, for the deportment bears still the stamp of hesitating
timidity, and this part of the act may become much freer as
the animals become tamer and more fearless. The final part
of the act, that of snapping the bait, was always performed
in the same characteristic way. The piece of meat seemed
always to be regarded as a living prey, which was to be seized
quickly, held firmly for a moment or two, and then swallowed.
Unfortunately I did not experiment to see what could be done
in modifying this part of the act.

Instinctive fear is evidently a very important element in the
conduct of the lower as well as the higher animals. In Nec-
turus we see how it may be just as effective as intelligence
in securing a sly mode of attack. So strong is its influence
that I doubt whether there is any finesse in the movement.
The adaptation of acts to purposeful ends must not be accepted
too quickly as proof of intelligence in the doer. Such acts are
common enough in plants, and there we are under the neces-
sity of finding some other explanation.

d. Organization Shapes Behavior.

Necturus appears to understand well the act of capturing its
prey, and the nice adaptation of each act to the end in view
naturally enough suggests forethought and refined experience.

ANIMAL BEHAVIOR. 299

But we see the performance executed by the young, which have
never had any experience of that kind, nor any opportunity to
copy from others. We cannot therefore suppose that they
perform these acts understandingly. The young Necturus,
hatched in a dish where it has never met any living thing
except its companions, has nothing to guide its first effort
to capture food except its organization and its simple expe-
rience in walking and swimming, which acts are again like
those of the adult, not because directed by intelligence or
example, but because they are performed with the same organs
under similar conditions. The young has the same sensory
and motor apparatus as the adult, but it has never before
known the feeling of hunger, it has never experienced pain
from contact with an enemy, it has never learned that a prey
may escape if not approached slyly. Its movements in ap-
proaching and snatching a piece of meat, as if it were a
living object, are, then, those characteristic of the species,
not because they are measured and adapted to a definite end
by intelligent experience, but because they are organically
determined ; in other words, depend essentially upon a specific
organization.

The timidity of young hatched in a dish is the same as that
of specimens hatched in the lake, and therefore it cannot be
charged to individual experience or to parental influence. It,
too, inheres in the special brand of organization, and has noth-
ing to do with memory of pain sensations.

e. Origin a7id Meaning of Behavior.

We have taken a very important step in our study when we
have ascertained that behavior, which at first sight appeared
to owe its purposive character to inteUigence, cannot possibly
be so explained, but must depend largely, at least, upon the
mechanism of organization. The origin and meaning of the
behavior antedate all individual acquisitions and form part of
the problem of the origin and history of the organization itself.
It is the first and indispensable step, without which it would be
impossible to reach sound views, either as regards the particular

300

BIOLOGICAL LECTURES.

behavior or the difficult question of the relation of instinct to
habit and intelligence. If the problem is not simplified, its
nature is better defined and its perspective is relieved of many
a myth that might otherwise obscure our vision. We see at
once that the behavior does not stand for a simple and pri-
mary adaptation of a preexisting mechanism to a special need.
As the necessity for food did not arise for the first time in
NecturuSy the organization adapted to securing it must be traced
back to foundations evolved long in advance of the species.
The retrospect stretches back to the origin of the vertebrate
phylum, and, indeed, to the very beginning of genealogical
lines in protozoan forms. The point of special emphasis here
is that instincts are evolved, not improvised, and that their
genealogy may be as complex and far-reaching as the history
of their organic bases.

f . Sensibility — Sources of Error.

Another important factor in animal behavior, namely, sensi-
bility, is very generally underestimated and often sadly mis-
understood. We are apt to gauge sensibility according to the
intensity of the overt response to stimulus, forgetting that
the animal has the power to inhibit such manifestations or
to moderate them in a way to mislead the observer. In the
struggle for existence a high premium has been placed on
this power, with the result that it is well-nigh a universal
attribute. The best proof of its high value to the possessor
is our own readiness to accept the disguise it affords as an
evidence of lack of sensibility. We are so prone to think that
the exercise of such power depends upon considerable intelli-
gence that we are incredulous of its existence in forms that
give only doubtful signs of intelligence. The power is pos-
sessed to a very marked degree by Clepsine, and it is only
when we become aware of this fact and take all necessary
precautions that we can get any reliable tests of the ani-
mal's keenness of sensibility. Nectiirus is even more difficult
to manage, for not until after we have won its confidence by
slow degrees can we expect free responses.

ANIMAL BEHAVIOR. 3OI

Besides the great danger of being deceived by the response,
or the lack of response, to stimulus, there are two other
insidious sources of error to be guarded against. We habitu-
ally assume that intelligence and sensibility rise and fall
together. This idea may lead to false conclusions in two
directions — to overestimating sensibility at the upper end of
the scale and underestimating it at the lower. That high
sensibility does not imply high intelligence is clear in the case
of Clepsine and equally so in almost any other case that might
be selected among the lowest segmented animals. That high
intelligence does not necessarily imply correspondingly high
sensibility is shown by the well-known fact that many animals
greatly surpass man in their sense powers. It can be shown,
I believe, that the difference in sensibility between higher and
lower animals is very much less than is generally supposed.

The second source of error is the common assumption that
the grade of sensibility rises with the structural complexity of
the sense organs. This view is likewise untenable. It is true
that the sense organs as a rule become more complex in struc-
ture as we go up the scale, but this advance in structure is
mainly confined to accessory and non-sensory parts, which are
either of a protective nature or else concerned in some sub-
sidiary function, such as muscular adjustments and regulation
of the stimuli. Such improvements in the non-sensory parts
may be carried to a high state of perfection and greatly raise
the general efficiency of the organ {e.g..^ the vertebrate eye),
without adding much, if anything, to the sensitiveness of the
individual sense cells. The sense cells may be multiplied in
number and placed in a position of safety and advantage for
receiving stimuli, and the stimuli may be strengthened,
directed, and otherwise regulated so as to secure the best
results ; but all that may obviously not affect the functional
power of the cells themselves. We do not know the range of
variability in this power, but we do know that the sense cells
often vary relatively little in structure, sometimes retaining in
the higher forms the same typical features that characterize
them in the lower forms. There is no known difference of
structure that would warrant the assumption that the dermal

302 BIOLOGICAL LECTURES.

sensillse in annelids are less sensitive than those in aquatic
vertebrates. We have, then, no reliable test of sensibility
either in the structure of the sense organ, in the rank of the
organism, or in its intelligence. We have to depend upon the
response to stimuli, and, remembering that this may be decep-
tive, observe and experiment under conditions that insure free
behavior.

No one who has never come into close communion with the
lower animals can begin to appreciate the delicacy and effi-
ciency of their sensory apparatus. We take up the earthworm
and, as we see no eyes, we conclude that it cannot see. A
little experiment shows that it is extremely sensitive to light,
and further study of its structure reveals unpigmented eyes
lying beneath the skin, and the whole surface thickly set with
minute delicate tactile sensillae. Even Amphioxus, so long
reputed to have no visual organs, turns out to have such
organs from end to end, imbedded in its spinal cord. I have
before called attention to the highly sensitive organization of
Clepsine and its allies. In the very lowest organisms, plant
and animal alike, where special visual organs do not exist, the
living protoplasm has, as has been demonstrated in many ways,
a keen sensibility to light, so that one might look upon the
whole organism as fulfilling the light-perceiving function.

g. Orientation through the Dermal Sensillce.

NecturuSy as before remarked, has a very keenly sensitive
organization. The skin is richly provided with sense organs,
which terminate at the surface in very short, fine hairs, invisible
to the naked eye. These organs, which are of the same nature
and function as the dermal sensillae in Clepsine and in so
many other aquatic animals, are sensitive to slight vibrations
in the water that are far beyond the reach of any of our sense
organs, and they are the main reliance, both in avoiding enemies
and seeking prey.

It is interesting to see how little the eyes are depended upon
in finding a piece of meat. A bit dropped in front of a young
Necturus receives no attention after it reaches the bottom.

ANIMAL BEHAVIOR. 303

An object must be in motion in order to excite attention, and
it is not generally the moving form that is directly perceived,
but the movements of the water, travelling from the object to
the sensory hairs, are felt, and in such a way as to give the
direction of the disturbing centre with most surprising accu-
racy. If a bit of beef is taken up adhering to the point of a
needle, and held in the water, the vibrations imparted to the
needle by the most steady hand will be sufficient to give the
animal the direction. If the meat falls to the bottom, and
the needle is held in place, the animal approaches the needle
and tries to capture it, without paying the slightest attention to
the meat lying directly below. If, after the meat has fallen, the
needle is withdrawn and touched to the surface of the water
behind or at one side of Nectunis, it turns instantly in the
direction of the needle, not because it sees, but because it feels
wave motions coming from that direction. Long experience
with Nectiiriis and with many of its nearer allies enables me to
speak very positively on this point. When it is remembered
that in the higher animals the direction of sound waves is given
by the auditory sense organs, which are primarily surface sen-
sillae homologous with those in the skin of Necturus, it may
not seem so strange that the animal directs its movements in
the way described. Necttirus can see, but it can feel (perhaps
we should say hear) so much more efficiently that its small
eyes seem almost superfluous. ^

1 Professor Eigenmann has kindly written the following note on the use of the
tactile organs in the blind fishes :

Chologaster papilliiferus^ a relative of the blind fishes living in springs, de-
tects its prey by its tactile organs, not by its eyes. A crustacean may be crawl-
ing in plain view without exciting any interest unless it comes in close proximity
to the head of the fish, when it is located with precision and secured. The action
is in very strong contrast to that of a sunfish, which depends on its eyes to
locate its prey. A Gammarus seen swimming rapidly through the water and
approaching a Chologaster from behind and below was captured by an instanta-
neous movement of the Chologaster^ when it came in contact with its head. The
motion brought the head of the Chologaster in contact with the stem of a leaf,
and instantly it tried to capture this also. Since the aquarium was well lighted,
the leaf in plain sight, it must have been seen and avoided if the sense of sight
and not that of touch were depended upon.

In Amblyopsis, the largest of the blind fishes of the American caves, the bat-
teries of tactile organs form ridges projecting beyond the general surface of the

304 BIOLOGICAL LECTURES.

h. Origin and Nature of the Behavior in Taking Food.

I. Some Intelligence Implied. — Let us now return to the
question of the origin and nature of the behavior of Necturus
in capturing its food ; not, however, with the expectation of
reaching a complete solution, but rather in the hope of coming
nearer to the problem and to the guiding principles in dealing
with it. It is obvious, first of all, that automatism will not
suffice to account for the whole behavior. That there is
organic coordination of movements no one will dispute. But
these movements must be steered in the direction of the
object, and this orientation does not seem to be a purely
automatic arrangement. The dermal sensillae ("lateral-line"
organs) give the impressions which enable the animal to
steer its course ; but action and sense impression are evidently
not linked in a way to be independent of inhibitory influ-
ences. I assume that the creature is conscious, and that it
has a certain intelligent appreciation of the sense impressions
received. This is not saying that the young Nectnrus is a born
philosopher ; I assume nothing more than that it has already
learned by experience how to direct its movements. That does
not imply much, but certainly some, intelligence. I cannot
otherwise understand why the same stimulus should not always
evoke the same response under the same conditions. But we
see that there is hesitation about starting, and this hesitation
may be prolonged to any length, showing conclusively that
sensation and response are not so connected as to exclude

skin. Its prey, since it lives in the dark and its eyes are mere vestiges, is located
entirely by its tactile organs. This is done with as great accuracy as could be
done with the best of eyes in the light, but only when the prey is in close prox-
imity to the head. Coarser vibrations in the water are not perceived or are
ignored, and apparently stationary objects are not perceived when the fish
approaches them. If a rod is held in the hand, the fish always perceives it when
within about half an inch of it, and backs water with its pectorals. If the head
of a fish is approached with a rod, the direction from whence it comes is always
perceived and the correct motion made to avoid it. This reaction is much more
intense in the more active young than in the adult. One young, about 10 mm.
long, determined with as great precision the direction from which a needle was
coming as any fish with perfect eyes could possibly have done. It reacted
properly to avoid the needle, and this without getting excited about it.

ANIMAL BEHAVIOR. 305

inhibitory influences. There is unmistakably a power of inhi-
bition strong enough to counteract the strongest motive to act
— the hunger of a starving animal in the presence of food.

2. Orientation Learned by Experience. — In assuming that
the young Necttirus, at the time of its first attempt to capture
a piece of beef, has already learned to orient itself with refer-
ence to external objects, I have not gone beyond the possibili-
ties. The animal has been out of the ^g'g envelopes for about
two months. It has been confined in a glass dish about ten
inches in diameter, holding water about one inch in depth.
Its life has been about as exclusively vegetative as if it had
been all the while within the ^gg membrane, the only differ-
ence being that it has had room to straighten itself and to
move about to some extent among its fellows. It has been
heavily laden with food-yolk and has maintained a quiet atti-
tude except when disturbed by change of water. Simple as
the life has been, the animal has had some experience in swim-
ming and walking, and opportunity to use to some extent its
organs of orientation. The bait offered to it is something
totally new in its experience, but we cannot, of course, claim
that its behavior towards the bait is wholly uninfluenced by its
previous experience.

3. Defended Instinct. — We have to do with what Mr. Lloyd
Morgan has termed a " deferred " instinct, i.e., an action per-
formed for the first time, but not until some time after birth.
Mr. Morgan's remarks on the first dive of a young moor hen ^
bring out very clearly the possible influence of experience in
the case of such deferred instincts.

Mr. Morgan says :

In the case of such an instinctive procedure of the deferred type
as that presented by the diving of a young moor hen, though, on the
first occasion of its performance, the congenital automatism predomi-
nates, yet it is difficult to believe and is in itself improbable that the itidi-
vidual experience of the young bird does not, even on the first occasion,
exercise some influence on the way in which the dive is performed. If we
desire to reach a true interpretation of the facts, we must realize the
fact that an activity may be of mixed origin. And if we distinguish
1 Habit and Instinct, pp. 136, 137.

306 BIOLOGICAL LECTURES.

— as we have endeavored clearly to distinguish — between instinct
as congenital and habit as acquired, we must not lose sight of the
fact that there is much interaction between instinct and habit, so
that the first exhibition of a deferred instinct may well be carried out
in close and inextricable association with the habits which, at the
period of life in question, have already been acquired.

Although Mr. Morgan's young moor hen had undoubtedly
learned far more by experience before its first dive than my
young Nectnrus could have learned before its first effort to
capture food, we are nevertheless well admonished to keep in
mind the fact that the activity here considered may not be pure
instinct. Allowing for the small though important part played
by intelligence, there remains a purposive sequence of coordi-
nated acts, which are always performed in essentially the same
manner by young and old, and by the young without instruction
or example or previous experience of like motive and stimulus.
In so far, then, as intelligence cannot possibly be a regulating
factor we must refer the activity to organization.

4. Pause before the Bait. — In order to exclude as much as
possible the influence of experience it will be well to confine
attention to the least variable part of the behavior. The con-
cluding phase of the performance is so typical and character-
istic, and so far removed from anything previously experienced,
that it may be regarded as a very near approach to pure instinct.
I have in mind the pause before the bait and then the quick
side-movement of the head as the jaws are opened to seize.

If this series of acts represents an organic sequence, and if
the behavior as a whole takes the form determined by the
organization, as seems to me beyond reasonable doubt, we
have an instinct the history of which may be coextensive with
the evolution of the animal. We stand at the end of an inter-
minable vista. The specific peculiarities of organization in
Necturus form but an infinitesimal element of the problem.
Scarcely a feature of the instinct belongs exclusively to Nectu-
rus. It is at least widely diffused among vertebrates, especially
among fishes. The differences in the manner of execution
among different forms, so far as I have observed, are of quite
a superficial nature. The instinct evidently has its root in the

ANIMAL BEHAVIOR. 307

general instinct of preying, which is doubtless coeval with ani-
mal organization. The cannibalism of our protozoan ancestors
was the starting-point, and their carnal propensities were not
acquired by the aid of intelligence, but given in the fundamental
properties of protoplasm. The stronger ate the weaker, and
made themselves stronger and more prolific by so doing. The
promise of the whole animal world was contained in the act.
The constitutional disposition to feed, with variable foods avail-
able, would give occasion for different appetites and various
modes of getting outside of palatable victims. In primitive
organisms multiplying by simple fission, structural modifica-
tions acquired during the lifetime of the individual would be
carried right on from generation to generation, and hence the
structural foundations for a whole animal world such as we now
see could be laid in a relatively short period as compared with
the time necessary to advance organization in forms limited to
reproduction by germs. In fact, these fundamentals could all
be established within the realm of the unicellular protozoa.
Nucleus and cell-body, inner and. outer layers, nerve-muscle
elements, sensory and locomotor organs, mouth and stomach,
respiratory and excretory mechanisms, reproductive elements,
anticipating embryological development from germs — all these
essentials of higher organization are presented in the pro-
tozoan.

The organic bases furnished in the protozoan world might be
passed directly on to the first metazoa, or they might be reac-
quired in essentially the same manner as before, and in a not
much longer period, as reproduction by fission would still be a
condition favoring rapid organo-genesis.

To try to fill up the gaps between the protozoan and Nectu-
rus would lead us too far into the field of speculation, and
would not contribute much to a grasp of the problem. We
have to content ourselves with general facts and principles and
probabilities drawn therefrom. It is enough for present pur-
poses to know that the roots of the instinct organization we are
considering run clear back to the beginnings of organo-genesis,
and that they are natural products of the properties of living
protoplasm. We start with known properties and get to known

308 BIOLOGICAL LECTURES.

rudiments of organization without invoking the aid of intelli-
gence, or finding any way in which it could be supposed to
have been a guiding factor in development.

The organic basis of the preying instinct may have grown
and multiplied in different phyla a long time before receiving
much aid from intelligence. The rapidity and freedom of modi-
fication would be very much limited when fission ceased and
reproduction by germs became the sole mode of generation.
Very early in the vertebrate phylum, possibly at its dawn, the
chief characters of the instinct, as we now find it, were proba-
bly fixed in structural elements differing from those in Necturus
only in superficial details. The strikingly fish-like character of
the behavior certainly suggests as much.

5. Meaning and General Occurrence. — If now we look more
closely at the purposive character of the behavior, it will
become clearer that the instinct is shared, not only by animals
below Necturus^ but also by some far above it. The pause
before the final act of seizing is a well-marked feature, which
means locating the prey and fixing the aim. The same action
with the same meaning runs all through the different branches
of the vertebrate phylum. It is, as I have already said, espe-
cially characteristic of the fishes and amphibia, and it is not
rare among the higher branches, the reptiles, birds, and
mammals. It may be seen to good advantage in the turtles,
and even the common fowl halts on coming up to the insect it
is pursuing in order to make sure its aim. I believe the same
instinct underlies the act of pointing in the dog. The origin
of this behavior in pointers cannot be referred to training, as
was clearly seen by Darwin.^

It may be doubted [says Darwin] whether any one would have
thought of training a dog to point had not some one dog naturally
shown a tendency in this line, and this is known occasionally to
happen, as I once saw in a pure terrier ; the act of pointing is prob-
ably, as many have thought, only the exaggerated pause of an animal
preparing to spring on its prey. When the first tendency to point was
once displayed, methodical selection and the inherited effects of com-
pulsory training in each successive generation would soon complete

the work.

1 Origin of Species, p. 207.

ANIMAL BEHAVIOR. 309

The ''tendency" manifested in some one dog was regarded
by Darwin as an accidental variation, the cause being unknown.
May not many of the variations appearing in domestic animals,
which we call "accidental," be manifestations of instinct roots
of more or less remote origin ?

6. Part Played by Fear, — We may now glance once more at
the behavior as a whole, for the purpose of pointing out the
part played by instinctive timidity. Gentle movements in the
water, kept up with steadiness, such as are imparted by a
needle in feeding as before described, may induce an attack,
while less gentle or unsteady movements may lead the animal
to remain quiet or to take flight. The same stimulus, accord-
ing to amplitude and evenness, may then be followed either by
advance, by quiet, or by retreat. In retreat, fear is manifest ;
in quiet it is concealed ; in advance it is less concealed. There
can be no doubt that fear predominates in flight and in quiet,
while it certainly tempers the advance, giving the appearance
of slyness deliberately acted in order to take the prey by sur-
prise. This sly manner of advance, whatever it be due to, has
a double advantage, for it is concealment against a possible foe
and prevents alarming a harmless prey. If I could suppose
that fear did not strongly influence the advance, I should
certainly incline to think that the animal really appreciated
the great advantages in quietly surprising its prey; but for
reasons before given I believe the animal is quite blind to any
such bearing of its action. The advantages of this manner of
action, however, are just the same as if it were deliberately
assumed, and the Necturus conducting itself in this way would
certainly fare better than one reacting in a contrary way.
The instincts of Necturus in this case cooperate to secure
its welfare, while if the creature depended upon its intelligence
it is difficult to see how it could escape immediate extinction.

3IO BIOLOGICAL LECTURES.

General Considerations.

a. Instinct Precedes Intelligence.

The view here taken places the primary roots of instinct in
the constitutional activities of protoplasm ^ and regards instinct
in every stage of its evolution as action depending essentially
upon organization. It places instinct before intelligence in
order of development, and is thus in accord with the broad
facts of the present distribution and relations of instinct and
intelligence, instinct becoming more general as we descend the
scale, while intelligence emerges to view more and more as
we ascend to the higher orders of animal life. It relieves us of
the great inconsistencies involved in the theory of instinct as
" lapsed intelligence." Instincts are universal among animals,
and that cannot be said of intelligence. It ill accords with any
theory of evolution, or with known facts, to make instinct de-
pend upon intelligence for its origin; for if that were so, we
should expect to find the lowest animals free from instinct and
possessed of pure intelligence. In the higher forms we should
expect to see intelligence lapsing more and more into pure
instinct. As a matter of fact, we see nothing of the kind.
The lowest forms act by instinct so exclusively that we fail to
get decided evidence of intelligence. In higher forms not a
single case of intelligence lapsing into instinct is known. In
forms that give indubitable evidence of intelligence we do not
see conscious reflection crystallizing into instinct, but we do
find instinct coming more and more under the sway of intelli-

1 Professor Loeb* refers instinct back to " (i) polar differences in the chemical
constitution in the egg substance, and (2) the presence of such substances in the egg
as determine heliotropic, chemotropic, stereotropic, and similar phenomena of irri-
tability." According to this view, the power to respond to stimuli lies in unorgan-
ized chemical substances, and the same powers exist in the adult as in the egg,
because the same chemical substances are present. Organization serves at all
stages merely as a mechanical means of giving definite directions to responses.

The view I have taken regards instinctive action as organic action, whatever
be the stage of manifestation. The egg differs from the adult in having an organ-
ization of a very simple primary order, and correspondingly simple powers of
response. Instinct and organization are, to me, two aspects of one and the same
thing, hence both have ontogenetic and phylogenetic development.

*" Egg Structure and the Heredity of Instincts," The Monist, vol. vii, July, 1897.

ANIMAL BEHAVIOR. 3II

gence. In the human race instmctive actions characterize the
life of the savage, while they fall more and more into the back-
ground in the more intellectual races.

Every hypothesis that would derive instinctive action from
teleological reflection is open to the same objections. In many
cases it would be necessary to postulate an amount of prevision
on the part of the animals in which the instincts arose that
would simply be psychologically impossible. Conscious pre-
vision without a possible basis in the experience of the
individual, or any means of learning from others, is simply a
self-contradiction. The frequently cited instance of the emperor
moth puts this point in strong light. The caterpillar of this
moth so constructs the upper part of its cocoon that it will
resist strong pressure from without and yield to slight pressure
from within. Easy egress for the imago and security against
attacks from outside enemies are thus provided for. As the
spinning of the cocoon happens but once in a lifetime, the
caterpillar could not anticipate such needs from its own experi-
ence, nor could it learn from its parents, which were dead long
before it hatched. The possibility of imitation is also excluded,
as the species is not a social one.

b. Theories of Instinct.

I . Pure Instinct the Point of Departure. — The first criterion
of instinct is, that it can be performed by the animal without
learning by experience, instruction, or imitation. The first
performance is therefore the crucial one. It is of the utmost
importance in all discussion of the origin of instinct to make
sure of this point, and keep clear of all ambiguous activities
such as have been designated " instinct habits " (Lloyd Mor-
gan), ''acquired instincts" (Wundt), ** secondary instincts"
(Romanes), etc. We must not allow the question as to the
relation of instinct to habit and intelligence to be obscured
by confusing terminology. There may be "mixtures" and all
sorts of " interactions " between habit and instinct, and these
may have a far-reaching theoretical import, but they lack definite-
ness, and are therefore dangerous foundations for theories. A

312 BIOLOGICAL LECTURES.

theory of instinct must obviously make pure instinct its first con-
cern, and keep the general course of evolution always in view.

It is not my purpose to engage in a critical examination of
theories, but to indicate briefly which of the two rival theories
now most in favor accords best with facts and general prin-
ciples as I understand them. These two theories are the habit
theory of Lamarck and the selection theory of Darwin, Wallace,
and Weismann.

2. Embryology and the Lamarckian Theory. — The habit
theory is a part of the more general theory of the transmission
of acquired characters. This doctrine has never been recon-
ciled with the teachings of embryology, the science which
deals directly with the phenomena of heredity, and which is,
therefore, the touchstone of every theory of inheritance. It
is a fundamental tenet in embryology that all organisms repro-
ducing exclusively by germs owe their inherited characters
to the germs from which they arose, and that germs carry the
primordials of adult structure, not by virtue of any mysterious
transference of parental features, but by virtue of the consti-
tution they bring with them when they arise by division of
preexisting germs. That is, I believe, a fair statement of the
embryological doctrine of inheritance, which must be the final
test of our theories.

The selection theory propounded by Darwin and Wallace, and
further developed by Weismann, starts from the embryological
law of germ continuity (Weismann), or, otherwise expressed,
germ lineage, and interprets the phenomena of variation, hered-
ity, and development, in harmony with this law and the prin-
ciple of selection. This theory is incompatible with the idea
that instinct is inherited habit. We could not, for example,
say with Professor Wundt^ :

"We have supposed that father can transmit to son the physio-
logical dispositions that he has acquired by practice during his own
life, and that in the course of generations these inherited dispositions
are strengthened and definitized by summation."

"The occurrence [p. 405] of connate instincts renders a subsid-
iary hypothesis necessary. We must suppose that the physical
1 Lectures on Human and Animal Psychology, p. 408.

ANIMAL BEHAVIOR.

313

changes which the nervous elements undergo can be transmitted
from father to son. . . . The assumption of the inheritance of
acquired dispositions or tendencies is inevitable if there is to be
any continuity of evolution at all."

3. Darwin s Refutation of Lamarck's Theory. — Although
Darv^in dwelt at some length on the points of resemblance
between habits and instincts, and although he thought it
possible that habits could sometimes be inherited, it should
be remembered that he was the first to show conclusively
that "the most wonderful instincts with which we are ac-
quainted, namely, those of the bee hive and of many ants, could
not possibly have been acquired by habit " {Origin of Species,
p. 202). Indeed, it was he who first found in the case of neuter
insects a refutation of Lamarck's doctrine of inherited habit,
and at the same time a demonstration of the high efficiency of
the principle of natural selection. Darwin concludes his chapter
on instinct with these memorable words :

" The case of neuter insects, also, is very interesting, as it proves
that with animals, as with plants, any amount of modification may be
effected by the accumulation of numerous slight, spontaneous varia-
tions, which are in any way profitable, without exercise or habit hav-
ifig been brought into play. For peculiar habits confined to the workers
or sterile females, however long they might be followed, could not possibly
affect the males and fertile female, which alone leave descendants. lam
surprised that ?to one has hitherto advanced this demonstrative case of
neuter insects against the well-known doctrines of inherited habit, as
advanced by Lamar ck.^^

What could more forcibly illustrate the importance of crucial
cases than just this work of Darwin's on the instincts of neuter
insects .? Here a conclusive test is reached, and no theory of
the origin of instinct can stand that disregards it. If habit
cannot possibly have had anything to do with the origin of such
typical instincts, then we should at least be very cautious in
appealing to it in any case. We certainly do not want two
theories to account for the same phenomenon, if one will suffice.
If the theory of inherited habit is certainly false in a single
case, it must be deemed false in every case, until at least it has
been shown that some cases cannot be explained without it. Is

314 BIOLOGICAL LECTURES.

there any case where it can be clearly shown that an undoubted
instinct arose from inherited habit, or any case in which it can
be made clear that the theory adopted for neuter insects can-
not possibly hold ? Both questions, it seems to me, must be
answered in the negative.

4. Weak Points in the Habit Theory. — The habit theory has
many adherents still, and Darwin himself often found it a con-
venient hypothesis. But neither Darwin nor anybody else has
given us a crucial test that would stand beside that furnished
in neuter insects. The failure to find such a test is certainly
not due to any lack of zeal or effort on the part of the advocates
of the theory. The tests claimed are numerous enough, but
they always fall short of the requirement. The weak points in
the theory are :

1 . It starts on a disputed, if not refuted assumption ; namely,
that habits wholly new to the individual and the species, having
no hereditary basis predisposing to them, may, as the result of
exercise frequently repeated, and continued in successive gen-
erations, eventually become hereditary instincts.

2. It appeals to the less typical rather than to the more
typical cases — to cases in which the critical points are unde-
termined or doubtful, or open to a different interpretation.

3. Its definition of habit and instinct verges towards a
petitio principii. Two or more classes of instincts are set up
so as to facilitate a nearer approach to habit, e.g., acquired and
connate (Wundt) ; primary and secondary (Romanes). Habit
is used indiscriminately for an action originating in some con-
genital variation and an action forced upon the individual by
special circumstances. A fundamental distinction, on which the
validity of the theory must be tested, is thus ignored.

c. Two Demonstrations of the Habit Theory Claimed
by Romanes.

The evidence adduced to show that habit may pass into
instinct cannot here be examined in detail. Romanes brings
forward two cases — the instincts of tumbling and pouting in
pigeons — which he declares are alone sufficient to demonstrate

ANIMAL BEHAVIOR. 315

the theory. We may, therefore, take these as fair samples of
the arguments generally appealed to.

After quoting Darwin's remarks on this subject, Romanes
adds :

" This case of the tumblers and pouters is singularly interesting
and very apposite to the proposition before us ; for not only are the
actions utterly useless to the animals themselves, but they have now-
become so ingrained into their psychology as to have become sever-
ally distinctive of different breeds, and so not distinguishable from
true instincts. This extension of an hereditary and useless habit
into a distinction of race or type is most important in the present
connection. If these cases stood alofie, they would be enough to shoiv
that useless habits may become hereditary, and this to an extent which
renders them indistinguishable from true instincts." ^

Granting that we have here true instincts, — and I do not
doubt that, — what proof have we that they originated in hab-
its } Did there preexist in the ancestors of these breeds organ-
ized instinct bases, which, through the fancier's art of selective
breeding, were gradually strengthened until they attained the
development which now characterizes the tumblers and pouters ?
Or was there no such basis to start with, but only a new mode
of behavior, accidentally acquired by some one or more individ-
uals, and then perpetuated by transmission to their offspring,
and further developed by artificial selection 1 The original
action in either species is called a " habit," and this so-called
habit must have been inherited ; ergo, habit can become
instinct. Obviously, argument of that kind can have weight
only with those who overlook the test-point, namely, the real
nature and origin of the initial action.

If the instinct had its inception in a true habit, i.e., in an
action reduced to habit by repetition in the individual, and not
determined in any already existing hereditary activity, is it at
all credible that it could have been transmitted from parent to
progeny ? Does not our general experience contradict such an
assumption in the most positive manner } But may not the
habit have originated a great many times, and by repetition in
successive generations, gradually have become " stereotyped

1 Mental Evolution in Animals, p. 189.

3l6 BIOLOGICAL LECTURES.

into a permanent instinct"? To suppose that such iitterly
useless action originated a great many times without compelling
conditions or any organic predisposition is not at all admissible.
Darwin saw at once from the nature of the actions that they
could not have been taught, but '^ 7mist have appeared 7iahirallyy
though probably afterwards vastly improved by the continued
selection of those birds which showed the strongest propensity !'
Darwin, then, postulates as the foundation of each instinct a
"propensity" — something given in the constitution. That
view of the matter is in entire accord with the theory adopted
in the case of neuter insects and quite incompatible with the
habit theory.

1 . The Instinct of Pouting. — I believe the case is much
stronger than Darwin suspected, and that it shows, not the
genesis of instinct from habit, but from a preexisting congen-
ital basis. Such a basis of the pouting instinct exists in every
dovecot pigeon, and is already an organized instinct, differing
from the instinct displayed in the typical pouter only in degree.
I could show that the same instinct is widely spread, if not
universal, among pigeons. It will suffice here to call attention
to the instinct as exhibited in the common pigeon. Observe
a male pigeon while cooing to his mate or his neighbors.
Notice that he inflates his throat and crop, and that this
feature is an invariable feature in the act, often continued for
some moments after the cooing ceases. Compare the pouter
and notice how he increases the inflation whenever he begins
cooing. The pouter's behavior is nothing but the universal
instinct enormously exaggerated, as any attentive observer may
readily see under favorable circumstances.

2. The Instinct of Tumbling. — The origin of the tumbling
instinct cannot be fixed by the same direct mode of identi-
fication ; but I believe that here also it is possible to point
to a more general action, instinctively performed by the dove-
cot pigeon, as the probable source of origin. I have noticed
a great many times that common pigeons, when on the point
of being overtaken and seized by a hawk, suddenly flirt them-
selves directly downward in a manner suggestive of tumbling,
and thus elude the hawk's swoop. The hawk is carried on by

ANIMAL BEHAVIOR. 317

its momentum, and often gives up the chase on the first failure.
In one case I saw the chase renewed three times, and eluded
with success each time. The pigeon was a white dovecot
pigeon with a trace of fantail blood. I saw this same pigeon
repeatedly pursued by a swift hawk during one winter, and
invariably escaping in the same way. I have seen the same per-
formance in other dovecot pigeons under similar circumstances.

But this is not all. It is well known that dovecot pigeons
delight in quite extended daily flights, circling about their home.
I once raised two pairs of these birds by hand, in a place sev-
eral miles from any other pigeons. Soon after they were able
to fly about they began these flights, usually in the morning.
I frequently saw one or more of the flock, while in the middle
of a high flight, and sweeping along swiftly, suddenly plunge
downwards, often zigzagging with a quick, helter-skelter flirting
of the wings. The behavior often looked like play, and prob-
ably it was that in most cases. I incline to think, however,
that it was sometimes prompted by some degree of alarm. In
such flights the birds would frequently get separated, and one
thus falling behind would hasten its flight to the utmost speed
in order to overtake its companions. Under such circumstances
the stray bird coming from the rear might be mistaken for the
moment for a hawk in pursuit, and one or more of the birds
about to be overtaken be thus induced to resort to this method
of throwing themselves out of reach of danger.

The same act is often performed at the very start, as the
pigeon leaves its stand. The movement is so quick and crazy
in its aimlessness that the bird often seems to be in danger of
dashing against the ground, but it always clears every object.

As this act is performed by young and old alike, and by
young that have never learned it by example, it must be
regarded as instinctive, and I venture to suggest that it prob-
ably represents the foundation of the more highly developed
tumbling instinct.

The behavior of the Abyssinian pigeon, which, when " fired
at, plunges downwards so as to almost touch the sportsman,
and then mounts to an immoderate height," may well be due
to the same instinct. The noise of the gun, even if the bird

3l8 BIOLOGICAL LECTURES.

were not hit, would surprise and alarm it, and the impulse to
save itself from danger would naturally take the form deter-
mined by the instinct, if the instinct existed. This seems to
me more probable than Darwin's suggestion of a mere trick or
play.

d. The Habit Theory Losing Ground.

The two instincts of pouting and tumbling, claimed as dem-
onstrations of the habit theory, thus turn out to be explicable
only on the selection theory. It is significant that this theory
is fast losing ground even among the psychologists. A. Forel's
conversion illustrates the trend of opinion. *' I, too," he says,
" used to believe that instincts were hereditary habits, but I am
now convinced that this is an error, and have adopted Weis-
mann's view. It is really impossible to suppose that acquired
habits, like piano playing and bicycle riding, for instance (these
are certainly acquired), could hand over their mechanism to
the germ-plasm of the offspring." ^

In his latest work. Habit and Instinct, Lloyd Morgan has
also abandoned the theory. On the same side stand James,
Baldwin, Ziehn, Fliigel, and others. The following, from Karl
Groos, pp. 60, 61, will show how the difficulties with the theory
are multiplying.

"As regards instinct," says Groos, *' there is, further, the
a priori argument that it is inconceivable how acquired connec-
tions among the brain cells could so affect the inner structure
of the reproductive substance as to produce inherited brain
tracts in later generations. And, finally, there is this consid-
eration mentioned by Ziegler as a suggestion of Meynert's :
* It is well known that in the higher vertebrates acquired asso-
ciations are located in the cortex of the hemispheres. As an
acquired act becomes habitual, it may be assumed that the cor-
responding combination of nervous elements will become more
dense and strong, and the tract proportionally more fixed. This
being the case, it follows that the tracts of acquired and habit-
ual association, as well as those of acquired movement, pass

1 Gehirn und Seele, 1894, p. 21. Taken from The Play of Animals, by
Groos, p. 56. (Translated by Elizabeth L. Baldwin.)

ANIMAL BEHAVIOR. 319

through the cerebrum. Instincts and reflexes, however, have
their seat for the most part elsewhere. The tracts of very few
of them are found in the cortex of the hemispheres. It is
chiefly in the lower parts of the brain and spinal cord that the
associations and coordinations corresponding to instincts and
reflexes have their seat. When the comparative anatomist
investigates the relative size of the hemispheres in verte-
brates (especially in amphibians, reptiles, birds, and mammals),
a very evident increase in size is observed which apparently
goes hand in hand with the gradual gain in intelligence. In
the course of long phylogenetic development, during which the
hemispheres have gradually attained their greatest dimensions,
they have constantly been the organ of reason and the seat
of acquired association. If, then, habit could become instinct
through heredity, it is probable that the cerebrum would, 'in
much greater degree than is the fact, be the seat of
instinct.' "

The stronghold of the Lamarckian view is Paleontology.
It is here that the doctrine of acquired characters, or ctetology
as Professor Hyatt calls it, has been nearly as unyielding as the
fossils to which it adheres. But a new light seems to be pene-
trating even here under the name of "organic selection."
This idea, first formulated by Professor Baldwin, but almost
simultaneously and independently reached by Lloyd Morgan
and Professor Osborn, is, that adaptive modifications are not
transmissible, but that they have, nevertheless, acted as the
fostering nurses of congenital variations^ since organisms sur-
viving through them would carry forward to the next genera-
tion such congenital variations as happened to be coincident
with them. It may be, perhaps, a fine question to determine
whether so-called "adaptative modifications," which really have
selection value, are not themselves the coincidents of con-
genital bases. Be that as it may, the conversion of so eminent
a paleontologist as Professor Osborn to the selection theory is
all the more significant on account of the prominent part he
has taken in defending the Lamarckian idea.

320 BIOLOGICAL LECTURES,

e. Hyatt on Acquired Characters.

Professor Hyatt was the first to demonstrate a wonderfully
complete parallelism between the ontogenetic and the phylo-
genetic series, and he has presented the paleontological argu-
ment in terms that seem, to many at least, to be beyond
controversy. With all respect to Professor Hyatt's monu-
mental work, I must say that I find nothing in the evidence
that compels one to take his view of acquired characters.

" We have been unable " [says Professor Hyatt] " to find any char-
acters which were not inheritable in some series. The behavior
of all characteristics which have been introduced into any series of
species shows them to be subject to the law of acceleration, in what-
ever way they have originated, whether primarily as adaptive char-
acters, according to our hypothesis, or by natural selection and
through the combination of the sexual variations, as supposed by
Weismann."^

This is a very sweeping statement, at least in implication.
I can hardly believe that the author would have us understand
that acquired characters are just as readily and invariably trans-
mitted as congenital characters ; and yet, if that is not the
argument, there is no argument there. Nothing is more cer-
tain than that, in living forms accessible to direct experimental
test, acquired characters are not invariably, if at all, trans-
missible. Demonstrations have been sought for, but so far
without avail. Unless the Arietidce are a wholly exceptional
group, we must conclude from the above statement that all the
characters found were of congenital origin, and that no acquired
characters were recognized. It is easier to believe that such
characters were overlooked than to believe a miracle.

The laiv of acceleration established by Professor Hyatt is
complemental to the biogenetic law formulated by Fr. Miiller
and Haeckel, and both laws rest on the theory of germ
continuity, as formulated by Weismann. Logically, neither of
these laws implies the transmission of acquired characters.
That is an assumption which has never been reconciled with
the fundamental law of the genetic continuity of germs. The

1 " Genesis of the Arietidse/'p. 43, Mem. Mus. Comp. ZooL, Cambridge, 1889.

ANIMAL BEHAVIOR. 32 I

pangenesis theory of Darwin was an attempt in this direction,
but that theory has no scientific basis and it stands as a theo-
retical failure, rejected because it could not possibly be recon-
ciled with what we know about the genesis of germs. That
is the inevitable fate of every view which fails to adjust itself
to the primary law of germ continuity.

Sense impressions and physical impressions or modifications
stand on the same footing. Repetition may become habit and
produce marked effects on the nervous mechanism or other
organs ; but the individual structure so affected is not con-
tinued from generation to generation, so that the effects are
cancelled with each term of life, and there is no conceivable
way by which they could be stamped upon the germs and so
carried on cumulatively. If they reappear in the offspring, it
cannot be because they were inherited, but because they are
reproduced in the same way as they were acquired in the
parent.

f. Preformation the Essence of the Doctrine.

This doctrine of the transmission of acquired characters is a
species of preformation that eclipses the old creatio7i hypothesis y
for the m,iracle of stamping the germ with the form it is to
present in the adult has to be repeated at each gene^^ation.

It may be objected that "stamping" is not the method by
which parental characters are given to the germ. They are
commonly said to be inherited. But it is too late to juggle with
the term "heredity." That term either means something or
nothing. If it means that characters acquired by the parent
can be transmitted to the offspring, then the transmitted
characters, which ex hypothesi are not originally determined in
the germ, must in some way be determined for it by the parent.
What better term than "stamp" or " impress" can be suggested.?
Whatever the modus operandi, the determining influence or
impress must be imparted, at least in the great majority of
cases, before development begins. Is it conceivable that per-
fectly definite form features can be in any way reflected back
upon the ovum t Can we think of the germ as vibrating
sympathetically with each acquired peculiarity of the parent

32 2 BIOLOGICAL LECTURES.

organism ? What vibration could there be between germ and
passive structures, such as shell configurations ? Could an
indentation, groove, ridge, or protuberance forced upon a shell
by environmental action be at the same time wrought into the
germs in such a definite way as to reappear in the offspring
without the aid of the same environmental causes ? Or could
the repetition of the same environmental action on a long line
of parents gradually modify the germs in the same direction ?
In whatever way we turn the question, we are confronted with
the same miracle of preformation. The character arises in the
parent organism by epigenesis, but it is throwrt back on the germ^
nobody knozvs or can conceive how, in such a way that it becomes
a preformation capable of unfolding without the aid of its epi-
genetic causes.

On the other hand, the hypothesis that all hereditary
characters in organisms exclusively gamogenetic originate in
spontaneous or induced (by direct action of environment) germ
variations, appeals only to known facts and principles, and
provides for the same amount of preformation as before with-
out any miraculous transfer of characters from one organism
to another. We know that germ variations are transmissible;
we do not know that individually acquired modifications can be
transferred to germs ; we know the principle of selection to be
rational and verifiable ; we know of no substitute for it.

The Genetic Standpoint in the Study of Instinct.

a. The Genealogical History Neglected,

The problem of psychogenesis requires a more definite
genetic standpoint than that of general evolution. It is not
enough to recognize that instincts have had a natural origin ;
for the fact of their connected genealogical history is of para-
mount importance. From the standpoint of evolution as held
by Romanes and others, instincts are too often viewed as dis-
connected phenomena of independent origin. The special and
more superficial characteristics have been emphasized to the
exclusion of the more fundamental phylogenetic characters.

ANIMAL BEHAVIOR. 323

Biologists and psychologists alike have very generally clung
tenaciously to the idea that instincts, in part at least, have been
derived from habits and intelligence; and the main effort has
been to discover how an instinct- could become gradually
stamped into organization by long-continued uniform reactions
to environmental influences. The central question has been :
How can intelligence and natural selection, or natural selection
alone, initiate action and convert it successively into habit,
automatism, and congenital instinct ? In other words, the
genealogical history of the structural basis being completely
ignored, how can the instinct be mechanically rubbed into the
ready-made organism ? Involution instead of evolution ; mech-
anization instead of organization ; improvisation rather than
organic growth ; specific versus phyletic origin.

This inversion, or rather perversion, of the genealogical
order leads to a very short-focussed vision. The pouting
instinct is supposed to have arisen de novo, as an anomalous
behavior, and with it a new race of pigeons. The tumbling
instinct was a sort of lusus natm-cBy with which came the fan-
cier's opportunity for another race. The pointing instinct was
another accident that had no meaning except as an individual
idiosyncrasy. The incubation instinct was supposed to have
arisen after the birds had arrived and laid their eggs, which
would have been left to rot had not some birds just blundered
into " cuddling " over them and thus rescued the line from
sudden extinction. How long this blunder-miracle had to be
repeated before it happened all the time does not matter.
Purely imaginary things can happen on demand.

b. The Incubation Instinct.

I . Meaning to be Sought in Phyletic Roots. — It seems
quite natural to think of incubation merely as a means of
providing the heat needed for the development of the egg,
and to assume that the need was felt before the means was
found to meet it. Birds and eggs are thus presupposed, and
as the birds could not have foreseen the need, they could not
have hit upon the means except by accident. Then, what an

324 BIOLOGICAL LECTURES.

infinite amount of chancing must have followed before the first
"cuddling" became a habit, and the habit a perfect instinct!
We are driven to such preposterous extremities as the result
of taking a purely casual feature to start with. Incubation
supplies the needed heat, but that is an incidental utility that
has nothing to do with the nature and origin of the instinct.
It enables us to see how natural selection has added some
minor adjustments, but explains nothing more. For the
real meaning of the instinct we must look to its phyletic
roots.

If we go back to animals standing near the remote ancestors
of birds, to the amphibia and fishes, we find the same instinct
stripped of its later disguises. Here one or both parents sim-
ply remain over or near the eggs and keep a watchful guard,
against enemies. Sometimes the movements of the parent
serve to keep the eggs supplied with fresh water, but aeration
is not the purpose for which the instinct exists.

2. Means Rest and Incidental Protection to Offspring. —
The instinct is a part of the reproductive cycle of activities,
and always holds the same relation in all forms that exhibit it,
whether high or low. It follows the production of eggs or
young, and means primarily, as I believe, rest with incidental
protection to offspring. That meaning is always manifest, no
less in worms, molluscs, Crustacea, spiders, and insects, than
in fishes, amphibia, reptiles, and birds. The instinct makes no
distinction between eggs and young, and that is true all along
the line up to birds which extend the same blind interest to
one as to the other.

3. Essential Elements of the Instinct, — Every essential ele-
ment in the instinct of incubation was present long before the
birds and eggs arrived. These elements are : (i) the disposi-
tion to remain with or over the eggs ; (2) the disposition to
resist and to drive away enemies ; and {3) periodicity. The
birds brought all these elements along in their congenital equip-
ment, and added a few minor adaptations, such as cutting the
period of incubation to the need of normal development, and
thus avoiding indefinite waste of time in case of sterile or
abortive eggs.

ANIMAL BEHAVIOR.

325

(1) Disposition to Remain over the Eggs. — The disposition
to remain over the eggs is certainly very old, and is probably
bound up with the physiological necessity for rest after a
series of activities tending to exhaust the whole system. If
this suggestion seems far-fetched, when thinking of birds, it
will seem less so as we go back to simpler conditions, as we
find them among some of the lower invertebrate forms, which
are relatively very inactive and predisposed to remain quiet
until impelled by hunger to move. Here we find animals
remaining over their eggs, and thus shielding them from
harm, from sheer inability or indisposition to move. That
is the case with certain molluscs {Crepidula)^ the habits and
development of which have been recently studied by Pro-
fessor Conklin.^ Here full protection to offspring is afforded
without any exertion on the part of the parent, in a strictly
passive way that excludes even any instinctive care. In Clep-
sine there is a manifest unwillingness to leave the eggs, show-
ing that the disposition to remain over them is instinctive. If
we start with forms of similar sedentary mode of life, it is
easy to see that remaining over the eggs would be the most
likely thing to happen, even if no instinctive regard for them
existed. The protection afforded would, however, be quite
sufficient to insure the development of the instinct, natural
selection favoring those individuals which kept their position
unchanged long enough for the eggs to hatch.

(2) Disposition to Resist Enemies. — The disposition to keep
intruders from the vicinity of the nest I have spoken of
as an element of the instinct of incubation. At first sight
it seems to be inseparably connected with the act of covering
the eggs, but there are good reasons for regarding it as a
distinct element of behavior. In birds this element manifests
itself before the eggs are laid, and even before the nest is
built ; and in the lower animals the disposition to cover the
eggs is not always accompanied by an aggressive attitude.
This attitude is one of many forms and degrees. A mild
self-defensive state, in which the animal merely strives to hold
its position without trying to rout intruders, would perhaps

^ Journ. of Morpk., vol. xiii, No. i, 1897.

326 BIOLOGICAL LECTURES.

be the first stage of development. In some of the lower ver-
tebrates the attitude remains defensive and is aggressive only
in a very low degree, while in others pugnacity is more or
less strongly manifested. Among fishes the little Stickleback
is especially noted for its fiery pugnacity, which seems to
develop suddenly and simultaneously with the appearance of
the dark color of the male at the spawning season.

In pigeons, as in many other birds, this disposition shows
itself as soon as a place for a nest is found. While showing
a passionate fondness for each other, both male and female
become very quarrelsome towards their neighbors. The white-
winged pigeon {Melopelia leucoptera) of the West Indies and
the southern border of the United States is one of the most
interesting pigeons I have observed in this respect. At the
approach of an intruder the birds show their displeasure in
both tone and behavior. The tail is jerked up and down
spitefully, the feathers of the back are raised as a threatening
dog "bristles up," the neck is shortened, drawing the head
somewhat below the level of the raised feathers, and the whole
figure and action are as fierce as the bird can make them. To
the fierce look, the erect feathers, the ill-tempered jerks of the
tail, is added a decidedly spiteful note of warning. If these
manifestations are not sufficient, the birds jump toward the
offender, and if that fails to cause retreat, wings are raised and
the matter settled by vigorous blows.

This pugnacious mood is periodical, recurring with each
reproductive cycle, and subsiding like a fever when its course
is run. The birds behave as if from intelligent motive, but
every need is anticipated blindly ; for the young pair, without
experience, example, or tradition, behave like the parents.

It seems to me that this mood or disposition, although in
some ways appearing to be independent of the disposition to
cover the eggs, can best be understood as having developed in
connection with the latter. It has primarily the same mean-
ing, — protection to the eggs, — but the safety of the eggs
and young depends upon the safety of the nest, and this
accounts for the extension of its period to cover all three
stages — building, sitting, and rearing.

ANIMAL BEHAVIOR. 327

(3) Periodicity. — The periodicity of the disposition to sit
coincides in the main with that of the recuperative stage. Its
length, however, at least in birds, is nicety correlated with,
though not exactly coinciding with, the time required for hatch-
ing. It may exceed or fall short of the time between laying
and hatching. The wild passenger pigeon {Ectopistes) begins to
incubate a day or two in advance of laying, and the male takes
his turn on the nest just as if the eggs were already there. In
the common pigeon the sitting usually begins with the first ^gg^
but the birds do not sit steadily or closely until the second ^^^
is laid. The birds do not, in fact, really sit on the first ^gg, but
merely stand over it, stooping just enough to touch the ^gg with
the feathers. This peculiarity has an advantage in that the
development of the first o-gg is delayed so that both eggs may
hatch more nearly together. The bird acts just as blindly to
this advantage as Ectopistes does to the mistake of sitting
before an ^gg is laid. Ectopistes is very accurate in closing
the period, for if the ^gg fails to hatch within twelve to twenty
hours of its normal time, it is deserted, and that too if, as may
sometimes happen, the ^gg contains a perfect young, about
ready to hatch. Pigeons, like fowls, will often sit on empty
nests, filling up the period prescribed in instinct, leaving the
nest only as the impulse to sit runs down. It happens not
infrequently that pigeons will go right on with the regular
sequence of activities, even though nature fails in the most
important stage. Mating is followed by nest-making, and at
the appointed time the bird goes to the nest to lay, and after
going through the usual preliminaries, brings forth no ^gg.
But the impulse to sit comes on as if everything in the normal
course had been fulfilled, and the bird incubates the empty
nest, and exchanges with her mate as punctiliously as if she
actually expected to hatch something out of nothing. This
may happen in any species under the most favorable condi-
tions. It is possible by giving an abundance of rich food to
wind up the instinctive machinery more rapidly than would nor-
mally happen, so that recuperation may end in about a week's
time, when incubation will stop and a new cycle begin, leading
to the production of a second set of eggs in the same nest.

328 BIOLOGICAL LECTURES.

This has happened several times with the crested pigeon of
Australia (yOcyphaps lophotes).

Schneider ^ says : '' The impulse to sit arises, as a rule, when
a bird sees a certain number of eggs in her nest." Although
recognizing a bodily disposition as present in some cases, sitting
is regarded as a pure perception impulse. I hold, on the con-
trary, that the bodily disposition is the universal and essential
element, and that sight of the eggs has nothing to do primarily
with sitting. It comes in only secondarily and as an adapta-
tion in correlation with the inability in some species to rear
more than one or two broods in a season. In such species the
advantage would lie with birds beginning to incubate with a
full nest.

The suggestions here offered on the origin of the incubation
instinct, incomplete and doubtful as they may appear, may
suffice to indicate roughly the general direction in which we
are to look for light on the genesis of instincts. The incuba-
tion instinct, as we now find it perfected in birds, is a nicely
timed and adjusted part of a periodical sequence of acts. If
we try to explain it without reference to its physiological con-
nections in the individual, and independently of its develop-
mental phases in animals below birds, we miss the more
interesting relations, and build on a purely conjectural chance
act that calls for a further and incredible concatenation of the
right acts at the right time and place, and is not even then
completed until its perpetuation is secured by a miracle of
transmission.

A Few General Statements.

I. Instinct and structure are to be studied from the common
standpoint of phyletic descent, and that not the less because
we may seldom, if ever, be able to trace the whole development
of an instinct. Instincts are evolved rather than involved
(stereotyped by repetition and transmission), and the key to
their genetic history is to be sought in their more general
rather than in their later and incidental uses.

1 Der Thierische Wille, pp. 282, 283. As cited in Professor James's Psychol-
ogy, P- 388.

ANIMAL BEHAVIOR. 329

2. The primary roots of instincts reach back to the consti-
tutional properties of protoplasm, and their evolution runs, in
general, parallel with organogeny. As the genesis of organs
takes its departure from the elementary structure of protoplasm,
so does the genesis of instincts proceed from the fundamental
functions of protoplasm. Primordial organs and instincts are
alike few in number and generally persistent. As an instinct
may sometimes run through a whole group of organisms with
little or no modification, so may an organ sometimes be carried
on through one or more phyla without undergoing much
change. The dermal sensillae of annelids and aquatic verte-
brates are an example.

3. Remembering that structural bases are relatively few
and permanent as compared with external morphological char-
acters, we can readily understand why, for example, five hun-
dred different species of wild pigeons should all have a few
common undifferentiated instincts, such as, drinking without
raising the head, the cock's time of incubating from about
10 A.M. to about 4 P.M., etc.

4. Although instincts, like corporeal structures, may be said
to have a phylogeny, their manifestation depends upon differ-
entiated organs. We could not, therefore, expect to see phyletic
stages repeated in direct ontogenetic development, as are the
more fundamental morphological features, according to the
biogenetic law. The main reliance in getting at the phyletic
history must be comparative study.

5. Instinct precedes intelligence both in ontogeny and phy-
logeny, and it has furnished all the structural foundations
employed by intelligence. In social development also instinct
predominates in the earlier, intelligence in the later stages.

6. Since instinct supplied at least the earlier rudiments of
brain and nerve, since instinct and mind work with the same
mechanisms and in the same channels, and since instinctive
action is gradually superseded by intelligent action, we are
compelled to regard instinct as the actual germ of mind.

7. The automatism, into which habit and intelligence may
lapse, seems explicable, in a general way, as due more to the
preorganization of instinct than to mechanical repetition. The

330 BIOLOGICAL LECTURES.

habit that becomes automatic, from this point of view, is not
an action on the way to becoming an instinct, but action pre-
ceded and rendered possible by mstinct. Habits appear as
the uses of instinct organization which have been learned by
experience.

8. The suggestion that intelligence emerges from blind
instinct, although nothing new, will appear to some as a com-
plete rediictio ad abstirdtun. But evolution points unmistakably
to instinct as nascent mind, and we discover no other source of
psychogenetic continuity. As far back as we can go in the
history of organisms, in the simplest forms of living proto-
plasm, we find the sensory element along with the other funda-
mental properties, and this element is the central factor in the
evolution of instinct, and it remains the central factor in all
higher psychic development. It would be strange if, with this
factor remaining one and the same throughout, organizing
itself in sense organs of the keenest powers and in the most
complex nerve mechanisms known in the animal world — it
would be strange, I say, if, with such continuity on the side of
structure, there should be discontinuity in the psychic activi-
ties. Such discontinuity would be nothing less than the nega-
tion of evolution.

9. We are apt to contrast the extremes of instinct and intel-
ligence— to emphasize the blindness and inflexibility of the one
and the consciousness and freedom of the other. It is like
contrasting the extremes of light 'and dark and forgetting all
the transitional degrees of twilight. In so doing we make the
hiatus so wide that derivation of one extreme from the other
seems about as hopeless as the evolution of something from
nothing. That is the last pit of self-confounding philosophy.

Instinct is blind ; so is the highest human wisdom blind.
The distinction is one of degree. There is no absolute blind-
ness on the one side, and no absolute wisdom on the other.
Instinct is a dim sphere of light, but its dimness and outer
boundary are certainly variable ; intelligence is only the same
dimness improved in various degrees.

When we say instinct is blind, we really mean nothing more
than that it is blind to certain utilities which we can see. But

ANIMAL BEHAVIOR.

331

we ourselves are born blind to these utilities, and only discover
them after a period of experience and education. The dis-
covery may seem to be instantaneous, but really it is a matter of
growth and development, the earlier stages of which conscious-
ness does not reveal.

Blindness to the utilities of action no more implies uncon-
sciousness in animals than in man. It is the worst form of
anthropomorphism to claim that animal automatism is devoid
of consciousness, for the claim rests on nothing but the assump-
tion that there are no degrees of consciousness below the
human. If human organization is of animal origin, then the
presumption would be in favor of the same origin for conscious-
ness and intelligence. Automatism could not exclude every
degree of consciousness without excluding every form of
organic adaptation.

10. The clock-like regularity and inflexibility of instinct, like
the once popular notion of the '' fixity " of species, have been
greatly exaggerated. They imply nothing more than a low
degree of variability under normal conditions. Discrimination
and choice cannot be wholly excluded in every degree, even in
the most rigid uniformity of instinctive action. Close study
and experiment with the most machine-like instincts always
reveal some degree of adaptability to new conditions. This
was made clear by Darwin's studies on instincts, and it has
been demonstrated over and over again by later investigators,
and by none more thoroughly than by the Peckhams in the
case of spiders and wasps. ^ Intelligence implies varying
degrees of freedom of choice, but never complete emancipa-
tion from automatism. The fundamental identity of instincts
and intelligence is shown in their dependence upon the same
structural mechanisms and in their responsive adaptability.

Instinct and Intelligence.

In order to see how instinctive action may graduate into
intelligent action it is well to study closely animals in which
the instincts have attained a high degree of complexity, and

1 Wisconsin Geological and Natural History Survey, Bulletin No. 2, 1898.

332 BIOLOGICAL LECTURES.

in which there can be no doubt about the automatic character
of the activities. These conditions are perfectly fulfilled in
the pigeons, a group in which we have the further advantage
that wild and domestic species can be studied comparatively.

It is quite certain that pigeons are totally blind to the
meanings which we discover in incubation. They follow the
impulse to sit without a thought of consequences ; and no
matter how many times the act has been performed, no idea
of young pigeons ever enters into the act.^ They sit because
they feel like it, begin when they feel impelled to do so, and
stop when the feeling is satisfied. Their time is generally
correct, but they measure it as blindly as a child measures its
hours of sleep. A bird that sits after failing to lay an Q.g^y or
after its eggs have been removed, is not acting from '* expecta-
tion," but because she finds it agreeable to do so and disagree-
able not to do so. The same holds true of the feeding instinct.
The young are not fed from any desire to do them any good,
but solely for the relief of the parent. The evidence on this
point cannot be given here, but I believe it is conclusive.

But if all this be true, where does the graduation towards
intelligence manifest itself. Certainly not in a comprehension
of utilities which are discoverable only by human intelligence.
Whatever the pigeon instinct-mind contains, it is safe to say
that the intelligence is hardly more than a grain hidden in
bushels of instinct, and one may search more than a day and
not find it.

a. Experiment with Pigeons.

Among many tests, take the simple one of removing the
eggs to one side of the nest, leaving them in full sight and
within a few inches of the bird on the nest. The bird sees
the uncovered eggs, but shows no interest in them; she keeps

1 Professor James, Psychology, II, p. 390, thinks such an idea may arise and
that it may encourage the bird to sit. " Every instinctive act in an a?iimal with
memory" says James, '■^must cease to be '■blind'' after being once repeated." That
must depend on the kind of memory the animal has. It is possible to have
memory of a certain kind in some things, while having absolutely none of any
kind in other things. That is the case in pigeons, as I feel very. sure.

ANIMAL BEHAVIOR.

m

her position, if she is a tame bird, and after some moments
begins to act as if the current of her feelings had been sHghtly
disturbed. At the most she only acts as if a little puzzled, as
if she realized dimly a change in feeling. She is accustomed
to the eggs, and now misses them, or, rather, misses some-
thing, she knows not what. Although she does not know or
show any care for the eggs out of the nest, she does appear to
sense a difference between having and not having.

There is, then, something akin to memory and discrimina-
tion, and little as this implies it cannot mean less than some
faint adumbration of intelligence. Now this inkling of intelli-
gence, or, if you prefer, this nadir of stupidity, so remote from
the zenith of intelligence, is not something independent of
and foreign to instinct. It is instinct itself just moved by a
ripple of change in the environment. The usual adjustment is
slightly disturbed, and a little confusion in the currents of feel-
ing arises, which manifests itself in quasi-mental perplexity.
That is about as near as I can get to the contents of the
pigeon mind without being able, by a sort of metempsychosis
suggested by Bonnet, to live some time in the head of the
bird.

In this feeble perplexity of the pigeon's instinct-mind, in
this ''nethermost abyss" of stupidity, there is a glimmer of
light, and nature's least is always suggestive of more. The
pigeon has no hope of graduating into a homo sapiens^ but
her little light may flicker a little higher, and all we need to
know is, how instinct behavior can take one step toward mind
behavior. This is the dark point on which I have nothing
really new to offer, although I hope not to make it darker.

b. TJie Step from Instinct to Intelligence.

Some notion of what is involved in the step may be gathered
by comparing wild with semi-domesticated and fully domesti-
cated species. These grades differ from each other in respects
that are highly suggestive. In the wild species the instincts
are kept up to the higher degrees of rigid invariability, while
in species under domestication they are reduced to various

334 BIOLOGICAL LECTURES.

degrees of flexibility, and there is a correspondingly greater
freedom of action, with, of course, greater liability to irregu-
larities and so-called ** faults." These faults of instinct, so far
from indicating psychical retrogression, are, I believe, the first
signs of greater plasticity in the congenital coordinations and,
consequently, of greater facility in forming those new combina-
tions implied in choice of action.

If we place the three grades of pigeons under the same con-
ditions and test each in turn in precisely the same way, we can
best see how domestication lets down the bars to choice and
at the same time gives more opportunities for free action.
The simplest experiment is always the best. Let us take
three species at the time of incubation and repeat with each
the experiment of removing the eggs to a distance of two
inches outside the edge of the nest. The three grades are
well represented in the wild passenger pigeon {Ectopistes)^
the little ring-neck {Turtur risorms)^ and the common dove-
cot pigeon {Columba livia domesticd). The results will not, of
course, always be the same, but the average will be about as
follows :

1 . The Passenger Pigeon. — The passenger pigeon leaves
the nest when approached, but returns soon after you leave.
On returning she looks at the nest, steps into it, and sits
down as if nothing had happened. She soon finds out, not
by sight, but by feeling, that something is missing. Her
instinct is keenly attuned and she acts quite promptly, leaving
the nest after a few minutes without heeding the ^g^. The
conduct varies relatively little in different individuals.

2. The Ring-neck Pigeon. — The ring-neck is tame and sits
on while you remove the eggs. After a few moments she
moves a little and perhaps puts her head down, as if to feel
the missing eggs with her beak. Then she may glance at the
eggs and appear as if half consciously recognizing them, but
make no move to replace them, and after ten to twenty
minutes or more leave the nest with a contented air, as if
her duty were done ; or, she may stretch her neck toward the
eggs and try to roll one back into the nest. If she succeeds
in recovering one, she is satisfied and again sinks into her

ANIMAL BEHAVIOR, 335

usual restful state, with no further concern for the second Qgg.
The conduct varies considerably with different individuals.

3. The Dovecot Pigeon. — The dovecot pigeon behaves in a
similar way, but will generally try to get both eggs back ; and,
failing in this, she resigns the nest with more hesitation than
does the ring-neck.

4. Results Considered. — The passenger pigeon's instinct is
wound up to a high point of uniformity and promptness, and
her conduct is almost too blindly regular to be credited even
with that stupidity which implies a grain of intelligence. The
ring-neck's stupidity is satisfied with one ^%^. The dovecot
pigeon's stupidity may claim both eggs, but it is not always
up to that mark.

In these three grades the advance is from extreme blind
uniformity of action, with little or no choice, to a stage of less
rigid uniformity, with the least bit of perplexity and a very
feeble, uncertain, dreamy sense of sameness between eggs in
and eggs ojU of the nest, which prompts the action of rolling
the eggs back into the nest. That is the instinctive way of
placing the eggs when in the nest, and the neck is only a
little further extended in drawing the eggs in from the out-
side. How very narrow is the difference between the ordinary
and the extraordinary act ! How little does the pendulum of
normal action have to swing beyond its usual limit ! ^

But this little is in a forward direction, and we are in no
doubt as to the general character of the changes and the modi-
fying influences through which it has been made possible.
Under conditions of domestication the action of natural selec-
tion has been relaxed, with the result that the rigor of instinc-
tive coordinations which bars alternative action is more or less

1 We come to equally surprising results in many different ways. Change the
position of the nest-box of the ring-neck, without otherwise disturbing bird, nest,
or contents, and the birds will have great difficulty in recognizing their nest, for
they know it only as something in a definite position in a fixed environment. If
a pair of these birds have a nest in a cage, and the cage be moved from one room
to another, or even a few feet from its original position in the same room, the
nest ceases to be the same thing to them, and they walk over the eggs or young
as if completely devoid of any acquaintance with or interest in them. Return
the cage to its original place and the birds know the nest and return to it at
once.

2,S6 BIOLOGICAL LECTURES.

reduced. Not only is the door to choice thus unlocked, but
more varied opportunities and provocations arise, and thus the
internal mechanism and the external conditions and stimuli
work both in the same direction to favor greater freedom of
action.

When choice thus enters, no new factor is introduced.
There is greater plasticity within and more provocation with-
out, and hence the same bird, without the addition or loss of a
single nerve-cell, becomes capable of higher action and is
encouraged and even constrained by circumstances to learn
to use its privilege of choice.

Choice, as I conceive, is not introduced as a little deity,
encapsuled in the brain. Instinct has supplied the teleological
mechanism, and stimulus must continue to set it in motion.
But increased plasticity invites greater interaction of stimuli
and gives more even chances for conflicting impulses. Choice
runs on blindly at first, and ceases to be blind only in propor-
tion as the animal learns through nature's system of compul-
sory education. The teleological alternatives are organically
provided ; one is taken and fails to give satisfaction ; another
is tried and gives contentment. This little freedom is the
dawning grace of a new dispensation, in which education by
experience comes in as an amelioration of the law of elimina-
tion. This slight amenability to natural educational influences
cannot, of course, work any great miracles of transformation
in a pigeon's brain ; but it shows the way to the open door
of a freer commerce with the eternal world, through which a
brain with richer instinctive endowments might rise to higher
achievement.

The conditions of amelioration under domestication do not
differ in kind from those presented in nature. Domestication
merely bunches nature's opportunities and thus concentrates
results in forms accessible to observation. Natural conditions
are certainly working in the same direction, only more slowly.
The direction and the method of progress must, in the nature
of things, remain essentially the same.

Nature works to the same ends as intelligence, and to the
natural course of events I should look for just such results as

ANIMAL BEHAVIOR. 337

Lloyd Morgan ^ so clearly pictures and ascribes to intelligence.
"Suppose," says Mr. Morgan, ''the modifications are of vari-
ous kinds and in various directions, and that, associated with the
instinctive activity, a tendency to modify it indefinitely be inher-
ited. Under such circumstances intelligence would have a tend-
ency to break np and render plastic a previously stereotyped
itistinct. For the instinctive character of the activities is main-
tained through the constancy and uniformity of their perform-
ance. But if the normal activities were thus caused to vary in
different directions in different individuals, the offspring arising
from the union of these differing individuals would not inherit the
instinct in the same purity. The instincts would be imperfect,
and there would be an inherited tendency to vary. And tkisj
if continued^ would tend to convert what had been a stereotyped
instinct into inflate capacity ; that isj a general tendency to cer-
tain activities {mental or bodily), the exact form and directioji of
which are not fixed, until by training, from, imitation or through
the guidance of individual intelligence, it became habitual. Thus
it may be that it has come about that man, zuith his enormous
store of iiinate capacity, has so small a number of stereotyped
instincts!'

The following from Professor James ^ is suggestive :
" Nature implants contrary impulses to act on many classes
of things, and leaves it to slight alterations in the conditions of
the individual case to decide which impulse shall carry the day.
Thus, greediness and suspicion, curiosity and timidity, coyness
and desire, bashfulness and vanity, sociability and pugnacity
seem to shoot over into each other as quickly, and to remain
in as unstable equilibrium, in the higher birds and mammals as
in man. They are all impulses, congenital, blind at first, and
productive of motor reactions of a rigorously determinate sort.
Each one of them, then, is an instinct, as instinct is commonly
defined. But they contradict each other; experience, in each par-
ticular opportunity of application, usually deciding the issue.
The animal that exhibits them loses the ' instinctive ' demeanor
and appears to lead a life of hesitation and choice, an intellec-

1 Animal Life and Intelligence, pp. 452, 453.

2 Psychology, II, pp. 392, 393.

338 BIOLOGICAL LECTURES.

Uial life ; not, however, because he has no instincts — rather
because he has so many that they block each other s path!'

Looking only to the more salient points of direction and
method in nature's advance towards intelligence, the general
course of events may be briefly adumbrated. Organic mechan-
isms capable of doing teleological work through blindly deter-
mined adjustments, reproduced congenitally, and carried to
various degrees of complexity and inflexibility of action, were
first evolved. With the organization of instinctive propensi-
ties, liable to antagonistic stimulation, came both the possibil-
ity and the provocation to choice. In the absence of intelligent
motive, choice would stand for the outcome of conflicting
impulses. The power of blind choice could be transmitted,
and that is what man himself begins with.

Superiority in instinct endowments and concurring advan-
tages of environment would tend to liberate the possessors from
the severities of natural selection ; and thus nature, like domes-
tication, would furnish conditions inviting to greater freedom
of action, and with the same result, namely, that the instincts
would become more plastic and tractable. Plasticity of instinct
is not intelligence, but it is the open door through which the
great educator, experience, comes in and works every wonder
of intelligence.

Spencer 1 has shown clearly that this plasticity must inevi-
tably result from the progressive complication of the instincts.

" That progressive complication of the instincts,'' he says,
" which, as we have found, involves a progressive diminution
of their purely automatic character, likewise involves a simulta-
neous co'mm,encement of 7nemory and reason."

1 Psychology, I, pp. 443 and 454, 455.

IN MEMORIAM.

JAMES INGRAHAM PECK.

Professor Peck died of pneumonia on Friday, Nov. 4, 1898,
after a brief illness.

He had lived thirty-five years, and had accomplished so much in
making his character fine and strong, in serving others as a rarely
successful teacher, in winning results as an investigator, and in
brightening and encouraging the lives of many others by his faith
and his loyal friendship, that the sense of loss is deeply and widely
felt.

He is remembered in his college days as one of the earnest men
who was seeking knowledge, not prizes nor marks. He was an
untiring and eager worker, but withal so interested in all the doings
of his class and the various general interests of the student, that he
was never looked upon as a "dig." He was too genial, too good a
comrade, too sympathetic with all that was bright and wholesome,
for that.

While I recall long evenings of work with him in the laboratory,
frequently lasting until midnight and after, others remember him for
his work in athletics, as artist of his class, and as generally one of
the best and truest fellows in 1887.

He was graduated from Williams College with high rank and spent
the following year there as Assistant in Biology. He began his
original work that year by a study of the "Variation of the Spinal
Nerves in the Caudal Region of the Domestic Pigeon." He showed
at once the true scientific spirit. He was so eager and determined
to know the truth that no perplexing details could overcome his
patience, nothing could discourage his perseverance and enthusiasm.
He must have the truth, and no price was too great to give for it.

In the following year, 1888-89, at the Johns Hopkins University
as a graduate student, he came under the influence and special
guidance of Professor Brooks. The opportunities for development
that were there provided him were met in the, same eager way so
characteristic of him. He appreciated this training keenly, and

339

340 BIOLOGICAL LECTURES.

often said that nothing could have been more enjoyable and more
successful in developing his powers than the wisely ordered work so
sympathetically directed, which he carried on at the Johns Hopkins
in the two years of his life there.

Among the results of his university work are two published papers,
one a report on Pteropods and Heteropods, and one on the "Anatomy
and Histology of Cymbuliopsis calceola."

In 1889 he was appointed Biologist to the Boston Water Works
and had begun a careful system of investigation of the water supplied
to the city of Boston, when failing health made it necessary for him
to give up his work for a long rest. With health recovered after two
years of farm life, he accepted a call to become Assistant in Biology
at Williams College. He received the degree of Ph.D. there in 1893,
and in the following year he was made Assistant Professor.

For several years Professor Peck was connected with the United
States Fish Commission, at first as Assistant and later in charge of
the laboratory of the Commission at Wood's Holl. He developed a
line of work there concerning the food of the Menhaden, and the
distribution of the food of certain marine fishes, which led him to
results quite unexpected and of far-reaching importance. Further
opportunity for work of this nature was offered him by observations
and collections made by Mr. N. R. Harrington in Puget's Sound.
The results of this work were studied and a report upon them by
J. I. Peck and N. R. Harrington was published in the Transactions
of the New York Academy of Science in 1898.

It was Professor Peck's hope that he might sometime have a
chance to extend his observations on the Atlantic coast outward to
and across the Gulf Stream and also into greater depth.

In 1896 the position of Assistant Director of the Marine Biological
Laboratory at Wood's Holl was offered him and he accepted it. Dur-
ing the three sessions of 1896, 1897, and 1898, Professor Peck devoted
himself to the interests and the development of the laboratory,
taking the supervision of the class in Animal Morphology. The
course of study had to be revised and a new staff of teachers brought
in and trained to the work. Professor Peck met all the difficulties of
the new situation with remarkable energy, perseverance, and tact.
Teachers and students alike caught the spirit of his enthusiasm and
zeal, and every day's work was carefully planned in a teachers' con-
ference the evening before. The example of such whole-souled devo-
tion to the work will not be soon forgotten.

I have already referred to the scientific character of his mind, and

IN MEMORIAM.

34

there was no lack of what that should include — the unprejudiced,
reasonable position, at all times, toward all views. He was a genial,
loyal friend, a man of high ideals, of great courage and strong faith,
and especially a man of full and ready sympathy. I never knew a
teacher who gave himself so generously, without measure, at all times
to all his students, as he did. They honored and loved him in an
unusual degree, and gladly did for him their best. It is said in the
college that very few have ever obtained so deep a hold upon the
students and have been such a power for good here as Professor
Peck.

We shall sadly miss him in the world of science, but the memory
of his true, brave, eager spirit, with his great love for truth and his
ready sympathy toward all, will be an uplifting and an abiding force
to all who knew him.

Samuel F. Clarke.

List of Dr. Peck's publications :

1888: A Report on the Pteropods and Heteropods Collected by the Steamer

Albatross on a Voyage from New York to San Francisco.
1889: Variation in the Nerves of the Caudal Region of the Domestic Pigeon.
1890 : Anatomy and Histology of Cymbuliopsis calceola.
1893: On the Food of the Menhaden.
1895 • Sources of Marine Food.

J. I. Peck and N. R. Harrington.
1898 : Observations on the Plankton of Puget Sound.

JAMES ELLIS HUMPHREY

A SKETCH of Dr. Humphrey's life and the minutes of a memorial
meeting are given in the Johns Hopkins University Circular of
November, 1897. From this source the following extract embodies
the sentiments of Professor Humphrey's colleagues at the Marine
Biological Laboratory :

" Professor Humphrey's value in the department of instruction under his
charge in this university increased rapidly and steadily during his three
years of service, until he became the essential factor in its success. Pupils
and colleagues soon learned to draw upon his well-ordered store of knowl-
edge and to trust his judgment, which ever proved sound and accurate.
His boundless patience and care as teacher gave assurance and encourage-
ment to his pupils, who soon learned to seek his advice in well-founded

342 BIOLOGICAL LECTURES.

confidence of ready and willing response. Nothing which promised to be
helpful to others was a trouble. Few have exemplified better than he the
beauty of orderliness. Books, papers, instruments, and his store of knowl-
edge were all at his ready command, and nothing with which one had to
do, either of facts or things tangible, was ever out of place.

" He was a ready writer and a well-known contributor to botanical litera-
ture, and all his work is marked by minute care, as well as by broad culture
and great power for original research and reflection, for he was broad and
liberal in all his thoughts and in all his deaUngs, and nothing that he
undertook was left without new additions with permanent scientific
value.

" To overwork, and a consequent lowered state of health, may probably be
attributed the lamentable catastrophe of his death. In spite of warning
and against the wishes of one whose pleasure and comfort were always of
supremest moment, he went because the duty of going seemed too plain
to resist, and then he fell ; — a loss to the university which the memory
of his character and example lightens but little ; a loss to his more intimate
friends which nothing but the hand of unfeeling time will remove ; a
bereavement to those yet nearer to him upon which we must not intrude,
except in sympathy."

The following is a list of Dr. Humphrey's published works :

1886 : On the Anatomy and Development of Agarum Turneri. Proc. Amer. Acad.

A. ^ S., June, 1886, p. 195, i pi.

1887 : The Preparation of Agarics for the Herbarium. Bot. Gaz., vol. xii, 1SS7,

p. 271.
1888: Potato Scab. Rep. Mass. Ag. Expt. Sta., 1888, p. 131, i pi.
1889: Mildews. Mass. Hort. Soc, 1889, pp. i-ii.

Fungous Diseases of Plants. Bull. No. 6, Mass. Ag. Expt. Sta., Oct. 1889,

p. 9.
A General Account of the Fungi. Mass. Ag. Expt. Sta. Rep., 1889, pp.

195-230.
1890: Notes on Technique. Bot. Gaz., vol. xv, 1890, p. 168.

Black Knot of Plum, Mildew of Cucumber, etc. Rep. Mass. Ag. Expt. Sta.,

1890, p. 200, 2 pi.

1891 : Treatment of Fungous Diseases. Bull. No.jg, Mass. Ag. Expt. Sta., April,

1891, pp. 1-12, 5 figs.

— — Protoplasmic Physics. Amer. Nat., vol. xxv, 1891, p. 376.

Comparative Morphology of the Fungi. Amer. Nat.,\o\. xxv, 1S91, p. 1055.

Notes on Technique. Bot. Gaz., vol. xvi, 1891, pp. 71-73.

Some Diseases of Lettuce, etc. Btdl. No. 40, Mass. Ag. Expt. Sta., 1891.

The Rotting of Lettuce. Rep. Mass. Ag. Expt. Sta., 1891, pp. 218-248, i pi.

1892 : Amherst Trees. Amherst, Mass., 1892, pp. 1-78.

Fungous Diseases and Their Remedies. Mass. Hort. Soc, Jan. 1892, p. i.

The Saprolegniaceae of the United States. Amer. Phil. Soc, Nov. 1892,

p. 63, 7 pi.

Rep. of Mycologist. Alass. Ag. Expt. Sta., 1892, p. i, 5 pi.

IN MEMORIAM.

343

1893: On Monilia Fructigena. Bot. Gaz., vol. xviii, 1893, p. 85, i pi.

Botanical Microtechnique (from the German by Zimmermann). New York,

1893, pp. 1-296.

1894 : Where Bananas Grow. Pop. Set. Month., vol. xliv, 1894, p. 486.

- — - Neucleolen und Centrosomen. Berich. der Deutschen Bot. Gesell., Bd. xii,

1894, p. 108, I pi.

Nucleoli and Centrosomes. Ann. of Bot., vol. viii, 1894, p. 373.

Eduard Strasburger. Bot. Gaz., vol. xix, 1894, p. 401, i pi.

1895: Some Recent Cell Literature. Bot. Gaz., vol. xx, 1895, P- 222.

Some Constituents of the Cell. Ann. of Bot., vol. ix, 1895, P- S^^' ' P^-

1896: Botany and Botanists in New England. N'ew England Mag., vol. xiv,

1896, p. 27, 2 pi.

On the Development of the Seed in the Scitamineae. Ann. of Bot., vol. x,

1896, p. I, 4 pi.

Some Modern Views of the Cell. Pop. Set. Month., vol. xlix, 1896, p. 603.

A University Celebration in Germany. Amer. Univ. Mag., Nov. 1896.

Dr. Humphrey was an associate editor of the Zeitschrift fiir
PJianzenkrankheiteTi^ a frequent contributor of notes on American
botany to the Botanisches Centralblatt^ and of specimens to the
Phycotheca Boreali Americani of Collins, Holden, and Setchell.

WESLEY WALKER NORMAN.

As this volume of Biological Lectures was going to print, Dr. W.
W, Norman, one of the contributors, died of typhoid fever.
He had been for a longtime a firm friend of, and zealous laborer in,
the Marine Biological Laboratory. Although already suffering, he
had come from Texas to take up his work as an instructor in the
laboratory during the summer of 1899. He died at the Massachu-
setts General Hospital, July 2, in his thirty-fourth year. A bio-
graphical notice will appear in the next volume of these Lectures.

Notices of Drs. Arnold Graf and Nathaniel Russel Harrington
must also be deferred.

RETURN TO the circulation desk ot any
University of California Library
or to the
NORTHERN REGIONAL LIBRARY FACILITY
BIdg. 400, Richmond Field Station
University of California
Richmond, CA 94804-4698

ALL BOOKS MAY BE RECALLED AFTER 7 DAYS

• 2-month loans may be renewed by calling
(510)642-6753

• 1-year loans may be recharged by bringing
books to NRLF

• Renewals and recharges may be made 4
days prior to due date.

DUE AS STAMPED BELOW

MAY Oft mn

12,000(11/95)

N9 518663

Woods Hole, l^as»
Marine Biological
Laboratory*

Biological lectures
delivered at the Ma-
rine biological labo-
ratory of Wood's Hole,

QH301

M3

V.6

LIBRARY

UNIVERSITY OF CALIFORNIA

DAVIS

3 1175 0108i|..4I22,