BIOLOGICAL BULLETIN

OF THE

flDarine Biological laboratory

WOODS HOLE, MASS.

Editorial Staff

E. G. CONKLIN — Princeton University.

JACQUES LOEB — The Rockefeller Institute for Medical Research.

GEORGE T. MOORE — The Missouri Botanical Garden.

T. H. MORGAN — Columbia University.

W. M. WHEELER — Harvard University.

E. B. WILSON — Columbia University.

EMtor

FRANK R. LILLIE — The University of Chicago.

VOLUME XXIX.

WOODS HOLE, MASS.
JULY TO DECEMBER 1915

PRESS OF

THE NEW ERA PRINTING COMPANY
LANCASTER. PA.

lirt 3S

CONTENTS OF VOLUME XXIX

No. i. JULY, 1915.

NEWMAN, H. H. Heredity and Organic Symmetry in Armadillo

Quadruplets I

VAN CLEAVE, H. J. Factors Concerned in the Production of

Mitosis in Organisms Displaying Cell Constancy 33

WHITNEY, DAVID D. The Production of Males and Females

Controlled by Food Conditions in the English Hydatina senta. 41
CALKINS, GARY N. Microt&niella clymenellce, a New Genus and

New Species of Colonial Gregarines 46

LOEB, JACQUES. The Blindness of the Cave Fauna and the

Artificial Production of Blind Fish Embryos by Heterogeneous

Hybridization and by Low Temperatures 50

CHURCHILL, E. P., JR. The Absorption of Fat by Freshwater

Mussels 68

No. 2. AUGUST, 1915.

BLAKESLEE, ALBERT F. Sexual Reactions between Hermaphro-
ditic and Dioecious Mucors 87

LOEB, JACQUES. Reversible Activation and Incomplete Membrane

Formation of the Unfertilized Eggs of the Sea Urchin 103

ALLEN, GEORGE D. Reversibility of the Reactions of Planaria

dorotocephala to a Current of Water ill

CHAPIN, CATHARINE L. A Case~qf .Hermaphroditismin Spelerpes

bilineatus 129

WOODWARD, ALVALYN E. Note on the Nature and Source of

" Purple X" 135

UHLENHUTH, EDUARD. Are Function and Functional Stimulus

Factors in Producing and Preserving Morphological Structure? 138

No. 3. SEPTEMBER, 1915.

HEILBRUN, L. V. Studies in Artificial Parthenogenesis 149

GATES, R. RUGGLES. On Successive Duplicate Mutations 204

No. 4. OCTOBER, 1915.
WELLS, MORRIS M. Reactions and Resistance of Fishes in their

Natural Environment to Acidity, Alkalinity and Neutrality. . 221

iii

iv CONTENTS.

SHUMWAY, WALDO. A Process of Temporary Chain Formation

by Fronionia 258

HARMAN, MARY T. Spermatogenesis in Paratteiix 262

Xo. 5. NOVEMBER, 1915.

TURNER, C. H. Notes on the Behavior of the Ant-lion with

Emphasis on the Feeding Activities and Letisimulation 277

HARVEY, E. NEWTON. The Effect of Certain Organic and In-
organic Substances upon. Light Production by Luminous
Bacteria 308

BORING, A. M., AND POOLER, R. H. Further Notes on the Chromo-
somes of the Cercopidce 312

BARROWS, WILLIAM M. The Reaction of an Orb-weaving Spider,

Epeira sclopetaria Clerck, to Rhythmic Vibrations of its Web. 316

No. 6. DECEMBER, 1915.
PATTERSON, J. T. Observations on the Development of Copidosoma

gelechice 333

ANDREWS, E. A. Distribution of Folliculina in 1914 373

KRECKER, F. H. Phenomena of Orientation Exhibited by

Ephemeridce 381

PAXMAN, DALTON G. Cell Multiplication in the Sub-cuticula of

Dilepis scolacina 389

Vol. XXIX. July, 1915. No. i.

BIOLOGICAL BULLETIN

HEREDITY AND ORGANIC SYMMETRY IN ARMA-
DILLO QUADRUPLETS.

I. MODES OF INHERITANCE OF BAND ANOMALIES.

H. H. NEWMAN.

INTRODUCTION.

Since the winter of 1909, when I first secured and began the
study of advanced polyembryonic fetuses of the nine-banded
armadillo, I have been struck not only by the striking resem-
blances among the individuals of a set of quadruplets but also
by certain equally striking differences. It was early noted that
well defined anomalous arrangements of scutes in the armor
bands were sometimes repeated with closely similar detail in
twTo or more fetuses, and were totally absent in others. Occa-
sionally all four fetuses of a set showed a highly localized anomaly,
but differed materially in the extent of the irregularity and in
its symmetrical relations.

The full significance of these conditions did not dawn upon me,
however, until some years later, when a set of quadruplets was
obtained in which the band anomaly in question was found to be
present not only in all of the four fetuses but in the mother as
well. Previous to this time only a few anomalous sets had been
studied and these happened to have normal mothers, a circum-
stance which led to the belief that these characters were not
strictly inherited from parents but were merely predetermined
early in embryonic life before the separation of the four embryos
from the originally single embryonic vesicle. The finding of
one unequivocal case of the direct inheritance from the mother
of the anomalous character stimulated a new interest in the
problem and made it necessary to collect a large amount of new

i

2 H. H. NEWMAN.

data. Accordingly in the winter of 1911 I journeyed from
Chicago to Texas and spent about a month in the armadillo
country collecting material, obtaining nearly two hundred
advanced polyembryonic sets together with the armor of the
mothers.

A study of this material by my students and myself has brought
to light a situation so interesting, but withal so intricately com-
plex, that I almost despair of being able to render an intelligible
account of it. Yet I cannot but be impressed with the unique-
ness of the data and with the fact that it contains clews, afforded
by no other material, as to methods of attacking some of the
problems of the mechanics of hereditary transmission and of the
development of organic symmetry.

In presenting the results of the study of the heredity of band
and scute anomalies I have found that the problems of heredity,
and that of the symmetrical distribution of these inherited
characters among the quadruplet embryos, are inseparably
parts of a larger problem that has to do with the mechanics
of organic symmetry. When, for example, it is found that a
bilateral anomaly in a mother is inherited unilaterally in the
various offspring and appears in reversed symmetrical relations
in twins, we have something more than a case of simple inherit-
ance. When again we find a unilateral anomaly in one pair
appearing bilaterally without reversed symmetry in the opposite
pair we see the problem in a more complex form. When, still
further, we note that an anomaly of a single scute in the mother
is inherited as such in some of the offspring and as a whole series
of anomalous scutes (a band anomaly) in others, we begin to
suspect that the problem is too intricate for any simple solution
and that only in certain of its aspects is it likely to yield to
analysis. Yet it is our duty to carry the analysis as far as our
data allows.

That the circumstance of polyembryony vastly complicates
the already sufficiently complex problems of heredity can scarcely
be denied, but it is confidently hoped that the new relations
introduced by this unique type of development may throw
light from a different angle upon certain phases of heredity
that are now quite obscure.

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO. 3

Before the reader can appreciate the significance of the data
on heredity and organic symmetry, it will be necessary for him
to be made familiar with the character of the material and its
specific frequency and distribution.

MATERIAL AND METHODS OF STUDY AND PRESENTATION.

The material for this investigation consists of nearly two
hundred sets of quadruplet fetuses in advanced stages, showing
the definitive arrangement and number of scutes in the armor.
In every case the shell of the mother was obtained and preserved
so that there is on hand complete data on uniparental inheritance
of armor characters. The fetuses and the respective mothers
make up a collection of nearly one thousand individuals, which,
is unquestionably a representative sample of the species.

Counts of scutes and records of anomalies have been made-
independently by at least two workers and all differences or dis-
crepancies have thus been checked and corrected, so that the
probability of error and personal bias have been eliminated.

Since the anomalies here dealt with are rare, occurring in
only about three per cent, of individuals, I have made an examina-
tion of the preserved armor of 1,800 adults that formed the stock
of a single dealer. This was done in order to determine the
limits of diversity and the specific distribution of the anomalies.,
A complete record of all these anomalies has been made in the
form of a pictorial diagram in Table A, I, 2 and 3.

In order that the reader may understand the nature of the
anomalies it will be necessary to give a brief statement of the
normal relations of armor units. The primary unit of the armor
is a threefold complex consisting of a bony plate, covered by a
primary epidermal scale of scute, and a group of hairs that are
imbedded in the bony plate and perforate the scute. For con-
venience this complex is called a "scute." In young fetuses the
outline of the epidermal scute is all that can be seen ; but this
visible unit stands for the whole complex. These scutes or
armor units of the carapace are arranged differently in the three
main subdivisions of the armor, the pectoral, banded and pelvic
shields. In the pectoral and pelvic regions the bony plates are
arranged like tiles and form a rigid immovable shield. Naturally

4 H. H. NEWMAN.

the primary scutes are arranged in more or less regular transverse
rows, but these rows seldom run straight across the shield.
Instead it is commoner for a row that starts single on the margin
to become double for some distance where the bulge of the shell
is greatest, then to be single again toward the middle. Again a
row may start double laterally and be single for a short distance
toward the middle. Irregularities in scute rows are commonest
near the posterior border of the pectoral shield next to the
banded region.

In the banded region the arrangement of scutes is quite
different. The bony plates are elongated antero-posteriorly to
form rows of units, in general appearance somewhat like the
keyboard of a piano. The posterior margin of the first band is
free from and slips over the anterior margin of the second, an
.arrangement that lends flexibility to the armor. The second
band bears a similar relation to the third and so throughout
the nine movable bands. Each bony plate is accompanied by
and partially covered by an elongated, wedge-shaped primary
scute. Figs, i and 3 represent two bands removed whole from
the armor. The typical band, however, is composed of an
even array of scutes arranged in a single row from margin to
margin. Only one third of one per cent, of the 16,200 bands
•examined are in any way irregular in the arrangement of the
armor units. The rare exceptions, however, are of especial
interest and constitute the anomalies that are the subject of the
present study. These peculiar bands belong for the most part
to the types, the details of which are shown in Figs. I, 3 and 5.
They consist of bands that are partly single and partly double.
In Fig. I is seen a very common type of bilateral anomaly in
which a few scutes on each margin constitute the single part of
the band, while the main central region is double. In Fig. 3
we see an equally common condition in which the double part
of the band is confined to a number of scutes starting at some
distance from the margins but stopping short of the middle
portions of the band. Similar conditions of greater or less extent
are found unilaterally as frequently as bilaterally. Now these
irregular bands are in no sense abnormalities or products of
injury, but are of the same nature exactly as are the irregulari-

HEREDITY AND ORGANIC 'SYMMETRY IN ARMADILLO.

7

60

4

60

IZ

29

8

8

8. -y/

3

/v

//

4

9- Co

4-

12.

O o

6'

6 H. H. NEWMAN.

ties of scute rows in the adjacent pectoral shield, where regularity
is rare and irregularity the typical condition. In about three
per cent, of individuals the irregular conditions typical for
the posterior part of the pectoral shield invade the adjacent
parts of the banded region. That this is the correct interpreta-
tion of the anomalies in the banded region is evidenced by the
fact that out of 84 band anomalies recorded for this study 73
occur in the first band, which is contiguous to the pectoral shield.
Of the remaining eleven cases 7 occur in band 2, I in band 3 and
I in band 8, which is near the other irregular region, viz., the
pelvic shield. Irregularities never occur in the middle bands of
the carapace, which are farthest from the irregular pectoral and
pelvic shields. Another way of interpreting the facts is to look
upon the arrangement of scutes in the banded region as due to
mechanical adjustments of the primordia of armor elements to
flexures during embryonic development. Typically the banding
confines itself to the abdominal parts of the armor but occasionally
comes in a little farther forward or a little farther back and thus
includes parts of the pectoral and pelvic carapace with the
typical peculiarities of scute arrangements of these regions.
As a result we have these anomalous cases in which the first or
second band has the characteristics, doubling, etc., of the
pectoral region, and the posterior bands have similarly the char-
acteristics of the pelvic region.

An irregularity in a given band of the banded region may
involve as few as two or even one scute. Such an irregularity
may consist of a single scute or two in an otherwise double band
or a double scute or two in a band otherwise single. The latter
situation is much the commonest of all anomalies studied.
Sometimes the anomaly manifests itself by a more or less com-
plete longitudinal or diagonal splitting of a single acute. Such
conditions are to be dealt with separately as there is a consider-
able mass of interesting data on the inheritance and distribution
of these anomalous double scutes. It is impossible, however,
to deal with the inheritance of anomalous band conditions with-
out discovering the intimate genetic connection that exists
between band and scute anomalies, for sometimes a scute anomaly
in a parent reappears as a band anomaly in offspring and vice
versa.

HEREDITY AND ORGANIC SYMMETRY IX ARMADILLO. ~

After many experiments in tabulating the occurrence of band
and scute anomalies I have adopted a simple pictorial scheme
which will readily explain itself on examination of the diagram-
matic Figs. 2, 4 and 7 which are simplified representations of the
scute conditions shown in Figs, i, 3 and 5 respectively. The
numbers of scutes in both double and single regions are repre-
sented by arabic numerals and the number of the band is indi-
cated sometimes, as in Table A, i, 2 and 3, by the abbreviation
Bd. i or Bd. 2 just above the margin of each band, and sometimes
as in table B by a single arabic numeral followed by a colon and
the total number of scutes in the band. When the quadruplet
fetuses and the mother are dealt with, the number of the set is
indicated as A.ioi 9 or K^Oc?, indicating at the same time
the sex of the litter; the bands of the mother are labeled M and

those of the fetuses I., II., III. and IV. When the mother or any
fetus is not tabulated the inference is that no anomaly is present
in the omitted individuals. Other schemes of tabulation, such
as the circular figures and those showing double scutes will be
explained in the proper place.

H. H. NEWMAN.

SPECIFIC DISTRIBUTION AND FREQUENCY OF BAND ANOMALIES.

Little need be added to the data given in Table A, 1,2, 3,
which shows the anomalies found in 1,800 adult specimens.
On previous occasions I have made records of over 1,000 other
specimens and have failed to find any other types of diversity
than those shown in this table. So it may be considered as
established that we have before us in this collection an adequate
representation of the diversity of band anomalies and their
distribution among the various bands. Although no two ir-
regular bands in unrelated individuals are just alike there are
certain well-defined classes of anomaly such as the bilateral
(symmetrical or asymmetrical) and the unilateral. There are
types single on the margins and double in the middle; there are
types double at the margin and single in the middle; and there
are mixed types.

Specimens 1-14 show various types of anomaly in which the
bands are single at the margin with bilateral regions of doubling.
Specimens 15 to 35 show unilateral expressions of the same types
of anomaly. Specimens 36-42 show various mixed types, which
are perhaps reversals of anomalies of the two sides of the indi-
viduals. Fig. 42 is an especially interesting case of such a
reversal of symmetry, for the right and left halves of the band
are duplicates but are not mirror-image effects. They bear the
same relation to each other as the reversed finger prints found
occasionally on the right and left index fingers of human duplicate
twins, as shown by Wilder.

Some of the cases of exact bilateral symmetry, such as those
in specimens 5, 7 and 36 are very significant, and there are all
degrees of inexact bilateral duplication of anomalies ranging from
specimens 12, 4, 8, n, 14 down to specimens 13, 10, 9, 6, etc.

It must not be forgotten that there are no two anomalies alike
in nearly three thousand specimens taken at random. When,
therefore, we find, as we soon shall, exact duplicate anomalies
in two or more fetuses in a set we shall not be able to explain
them as coincidences.

Having made clear the nature of the anomalies and their
diversity and distribution in the species, we are now in a position
to examine the data on the inheritance of these characters.

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO.

TABLE Ai.

Bd. i

32,

/ft

10

/e

/e

3

BJ. I

3J 1

JLL

5

?8

8

30

30

r

7

/3c/.

42-

6

22

30

/o

2,7

//

7

6

3Z

4-9

5-0

IO

H. H. NEWMAN.

TABLE A2.

f

if

16

/

ZO

/3d < Z

6

. r

/3V./

8

16

27

QJ.f

36

6 J.I

23
23

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO. II

TABLE AS.

(3d I

G

10

IO

4-0

9

37

&JZ

60

6d I

6'G

18

/8

34-

33

37

8

2,0

56'

60 l

GJ2

3/

/6

/O

39

39

TT

8

27

AZ.

Z7

~

12 H. H. NEWMAN.

THE INHERITANCE AND DISTRIBUTION OF BAND ANOMALIES.

After a study of over 150 of the most advanced sets of quad-
ruplets in my collection I am convinced that both band and
scute anomalies are strongly inherited. In every case in which
a mother exhibits a band or a scute anomaly, a related anomaly
is found in one or more of the offspring. If the character were
not inherited as a dominant, one would expect some exceptions
to this rule, but none have been found. When, therefore, we
find an equal number of offspring of normal mothers exhibiting
anomalies of the same sort we are justified in concluding that
the characters represent a heritage from the unknown fathers.
This assumption is farther justified by the finding that the char-
acters in question are neither sex-limited nor sex-linked.

For our purposes, then, the data here published are adequate
in that a study of uniparental inheritance reveals fully the modes
of inheritance that obtain for band and scute anomalies. What-
ever genetic relations are found to hold between mothers and
offspring would doubtless hold for fathers and offspring. The
only unfortunate complication that is encountered is in connec-
tion with a small per cent, of cases in which both fathers and
mothers possess anomalies. A few sets of fetuses are obviously
of this dual anomalous parentage, and we can, by knowing the
maternal anomaly, make a well-founded conjecture as to the
probable nature of the anomaly in the unknown father.

Whether or not I am justified in assuming that the study of
maternal inheritance reveals the essential facts concerning the
inheritance of the characters in question, can be settled only by
breeding and, as has been pointed out in extenso in an' earlier
paper (Newman, '13), breeding experiments with the armadillo
are at present totally impracticable. Consequently, we are
forced to rely upon a study of inheritance from one parent,
the mother.

The nature of the anomalies is such that I have been unable to
devise any really convenient method for tabulating the facts that
must be known about them. It seems necessary to consider
each case of inheritance separately, and this may be done without
undue prolixity because the cases are not numerous. The
pictorial method appears to be well adapted for the data, but it

HEREDITY AND ORGANIC SYMMETRY IX ARMADILLO. 13

would be too tedious a task to draw each anomaly in detail after
the fashion of Figs. I and 3. Instead I shall use the diagrammatic
form seen in Figs. 2 and 4, which give all the necessary informa-
tion. The best method I have been able to devise for showing
the symmetrical or asymmetrical relations of the anomalies and
their distribution among the quadruplet fetuses of a set is that
shown in Figs. 9-16, in which the anomalous bands are placed
as though within the embryonic vesicle. The reader must
remember that the structures studied are integumentary units,
that the embryonic vesicle is so inverted that the ectoderm forms
the inner lining and the endoderm is on the exterior.

The diagram (Fig. 8) shows the relative positions of the band
anomalies of the fetuses in set 87 (Fig. 9). Here we have an
equatorial cross section of an advanced vesicle with the body
of each fetus severed at the first armor band where the anomaly
occurs. The ventral surfaces face outward and are attached
by their respective umbilical cords to the placental portion of
the vesicle. The dorsal aspects are turned inward. Right and
left sides of each fetus are indicated. The reader must imagine
himself in the center of the vesicle facing outward toward each
fetus. In order to simplify this type of diagram, I have found
it best to ignore all structures but the band in question and to
represent the latter as though straightened out against the
periphery. The reader will understand the adopted form of
diagram after a comparison of Fig. 8, which represents nearly
the actual relations present in set 87, and Fig. 9, which is a
diagram of the same set showing only the bands that are of
interest in this study. In both figures the solid quadrant line
divides the twins I. and II. from twins III. and IV. and the
dotted line separates twin individuals, I. from II. and III. from
IV. In the actual vesicle these quadrant lines are occupied by
amniotic partitions that hermetically isolate each fetus from
the others. When the anomaly is found in the mother also the
material condition is indicated in a straight band diagram beside
or beneath the circle showing the offspring. When any individual
shows an anomaly in more than one band, the more anterior
anomaly is. shown in the circle and the more posterior anomaly
in a concentric sector outside of the circle, as in Fig. 9.

14 H. H. NEWMAN.

Of the twenty-six sets of fetuses one or more individuals of
each of which show band anomalies, twelve came from mothers
that also showed either band or acute anomalies, and fourteen
came from mothers that showed no anomaly. Presumably the
anomaly has been inherited from the father in these fourteen
cases.

Of the twenty-six sets showing band anomalies exactly half
are male and half female, which demonstrates that the characters
dealt with have no sex-limitation or sex-linkage. Males and
females inherit equally strongly from the mothers, for of the
twelve sets of offspring in which the mother also shows an
anomaly, five are of the female and seven of the male sex. In
the case of four sets with anomalous mothers and similarly in
four sets with normal mothers, I shall follow the plan of placing
a circular diagram and the verbal description and analyses
close together in the text. The remaining sets are tabulated in
somewhat more compact form and placed at the end of the paper.
Those in Table B are the eight remaining sets that have the
anomaly in both mother and one or more offspring; those in
Table C are the remaining ten sets with band anomaly in one or
more fetuses but with normal mothers. Each set is placed in a
separate block and labeled at the upper left-hand corner, with
the number of the set and the sex of the young, as C 29 cf . The
number of the band showing the anomaly is indicated in two
ways, as bd. I or bd. 2 placed next to or on the band, or by a
number placed on the band followed by a colon and the total
number of scutes in the band, as 7: 65, meaning band 7 having a
total of 65 scutes. In bands more or less broken by doubling
the number of scutes in each part is indicated. In the case of
short series of double scutes or one double scute, the number
placed on the element, together with the arrow, indicates the
numerical position of the element from either margin, or from
the middle, which is indicated by a dotted line.

Set K. 87, 9 (Figs. 8 and Q).

This set shows more certainly than any other the direct
inheritance of a material band anomaly by all of the offspring.
In the mother the anomaly consists of a unilateral local doubling

HEREDITY AND ORGANIC_SYMMETRY IX ARMADILLO. 15

8

3

o

H
X

PI

7)

1 6 H. H. NEWMAN.

of band I, confined to the left half of the band and almost
identical with the left half of the band pictured in Fig. 3. The
band begins on the left lateral margin with six single scutes, is
then double for 14-13 scutes and single for the remainder of the
series. No other anomalies are exhibited by the mother.
Fetuses I., II. and IV. show similar anomalies of band I, but
they are unequally bilateral in their expression (like Fig. 3), all
three showing more scutes double on the left than on the right.
Fetuses I. and II. are strikingly similar in the distribution of
single and double scutes and show no reversal of symmetry.
Fetus IV. is the most nearly bilaterally symmetrical in the dis-
position of its scutes. Fetus III. is unilateral like the mother,
but shows a reversal of symmetry in that the doubling is confined
to the right side. The symmetry of fetus III. is also a reversal
of that of fetuses I., II. and IV., which show a preponderance of
doubling on the left side. In addition to the anomalies referred
to, fetuses I. and II. show scute anomalies of bands 4 and 5
respectively that do not appear to be traceable to the mother
and have presumably come from the father. This is one of the
rare sets in which it is probable that anomalies came from both
sides of the family. The double scutes of I. and II. are in a
reversed symmetry to each other, being near the right in fetus I.
and near the left in fetus II.

Set K.JO, 9 (Fig. 10} .

This set is of importance in that it deals with short rows of
double scutes ranging from 2 to 7 elements and serves to empha-
size the genetic connection between plate and scute anomalies.

The mother has in band I in the positions 7 and 8 from the
left two adjacent double scutes forming a band anomaly of
minimal size but of the same character as the short double
regions of the band pictured in Fig. 5. If only one scute had
been double-tiered, we would have called it a "scute" anomaly
instead of a "band" anomaly. There is also in the mother an
independent double scute situated 10 places to the left of the
middle of band 2.

Fetus I. has a double scute in band I in exactly the position
of the first of the two double scutes of the mother. This seems

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO. 17

to indicate that a double scute and an incipient double band are
genetically equivalent.

Fetus II. has in the exact middle of band I a short series of
three double scutes like the two which are near the left margin
of the mother; and in the same band, midway between the
middle and left band margin, one double scute. In band 6 of the
same fetus there occurs at a point three scutes to the left of the
middle another double scute like that in band i. Note that all
asymmetrical anomalies are on the left side as in the mother.

54

MOTH e '

Bindl

7?.

10

Fetus III. has one double scute in the exact middle of band I,
or in a position identical with the short double band in fetus II.

Fetus IV. has a double region of 7-7 scutes in the exact middle
of band I. This doubling is simply a more pronounced case of
the doubling shown in the mother as in fetuses II. and III.

The points of interest illustrated by this set are:

I. In mother and offspring the unilateral symmetry is the
same throughout except that there is a tendency for the anomaly
to shift to the center of the band. This shifting of lateral
anomalies to the center is a peculiar type of symmetry reversal
in which the central and lateral portions of each half band become

i8

H. H. NEWMAN.

reversed. That this is the correct interpretation of the situation
is amply shown by many cases shown subsequently.

2. There occur several examples of the reduplication of a single
inherited unit, sometimes involving a single band, sometimes
two or more. Doubtless the anomalies of band I and band 2 in
the mother represent a merely reversed reduplication of the same
inheritance unit. Similarly in fetus II. the series of three double
scutes in the middle of band I, the double scute near the middle
of band 6 and the scute in the middle of the left half of band I
are varied manifestations of the same anomaly.

3. There is more extensive doubling in the two primary bud
individuals II. and IV. than in the two secondary bud individuals
I. and HI.

S-3-/0

II

Set K. 4, <? (Fig. //).

The mother has a short series of 3 double scutes in positions
8, 9 and 10 from the left.

Fetus I. has a series of 5 double scutes occupying positions 5, 6,
7, 8 and 9 from the middle of the band.

Fetus II. has on the left 4 single scutes, and the rest of the

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO.

hand, beginning in position 5, is double. The symmetry is
that of the mother, but the doubling involves most of the band.
The symmetry is a half-band reversal of that in fetus I. in that
the doubling begins 5 scutes from the left margin instead of 5
scutes from the middle.

Fetus III. has either no anomaly at all or else has the entire
band doubled. It is difficult to decide as to these alternatives.

Fetus IV. has the anomaly bilaterally symmetrically. One
might claim that the anomaly that should have appeared on
fetus III. has been transferred to fetus IV. so that the latter has
a double dose of it.

Set C. i, 9 (Fig. 12).

The mother has in band 3 in the second position from the left
a double scute.

Fetus I. has an extensive doubling beginning 5 scutes from the

12

right-hand margin of band I. Thus the symmetry of the mother
is reversed.

Fetus II. has an extensive unequal bilateral doubling beginning
after two scutes on the right and three on the left, continuing to
the middle on the left and for 13-14 scutes on the right.

2O H. H. NEWMAN.

Fetus III. has the doubling of band i, beginning 4 scutes from
the left. The symmetry is the same as that of the mother, but
a reversal of that of the opposite individual, fetus I. Fetus III.
is therefore a mirror-image of fetus I.

Fetus IV. has the anomaly beginning after 3 single scutes from
the right margin of band I. The symmetry is a reversal of that
seen in its twin partner, fetus III., and of that of the mother, so
that III. and IV. are mirror-images of each other.

This set shows clearly that a double scute is genetically equiva-
lent to a double band. It is seldom that one finds a set showing
such extensive reversals of symmetry.

Set C. 29, rf1 (Table Bi}.

This set is the complement of set C. i, in that the mother has
an extensive band doubling and the offspring inherit it in the
form of scute doubling.

The mother has in band I beginning 4 scutes from the left
a doubling extending to the right margin.

Fetus II. has in band 7 the second scute from the right double.

Fetus III. has in band i and in band 8 just to the left of the
middle a double scute, and in addition a double scute 8 scutes
from the left in band 8.

Fetus IV. has in band 3 a double scute in the ninth position
from the left.

This set illustrates an extreme case of positional shifting of a
genetic character. There can be no question of the relationship
between the two scutes of fetus III., occurring in band i and
band 8 just to the left of the middle. This is a case of reduplica-
tion of inheritance unit along the antero-posterior axis. Similarly
scute 8 in band 8 of fetus III. and scute 9 in band 3 of fetus IV.,
are expressions at different levels of the antero-posterian axis
of the same genetic unit.

The double element next to the right-hand margin of band 7
in fetus II. is certainly the same genetic unit as the element next
to the middle of band 8 of fetus III., but with a reversal of
symmetry. Thus by the application of the principle of sym-
metry reversals and reduplications down the primary axis we can
connect up even such apparently unrelated units as these.

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO.

21

C.29&

TABLE Bi.

M
1C
3E
IE
IE

4

59 «<*/.

7-64 2

i:Gl

1

*-

\/

8

B-.ei

B:C1

+

V

0

3-6Z \

C. 4-0

17 "R "7ft

R

3 Q

17 ^-8 /V

\\

o

;

8

-OO-^o

9:6-6

^

->

«-

8 g:<T5"

K

M

„

^"5

' 1

£

r

X

<-

sy

/o

ff:£V

JT

^

V

3 • 6^2

M

/o

V

S 63"

2.: 67

8:6-6

X

or

C. 40 c^ (Table

Here the same type of inheritance is seen as in set C. 29, but
the genetic relations are much clearer, and involve no strain
upon one's credulity.

The mother has in band i a bilateral doubling, beginning on
the left after four single scutes and extending in a series of 17-17
double scutes and beginning on the right after 8 single scutes and
continuing for a short series of 3-3 double scutes. In addition
we find in band 8 a doubling strikingly like that of the right side

H. H. NEWMAN.

of band i, but reversed to the left and involving only 2 scutes.
This is another case of reduplication in anterior and posterior
parts of the banded region, but there is also in this case a sym-
metric reversal.

Fetus I. has in band 9 in position 8 from the right a double
scute. This is evidently the genetic equivalent of the two double
scutes in band 8 of the mother, but the symmetry is reversed.

Fetus II. has in band 8, 8 scutes from the right, a double scute
which is doubtless the reversal or mirror-image of that in its
twin partner, fetus I.

Set K. 34, tf (Table Bi).

In this set the genetic relations are quite obscure.

The mother has in band I, beginning after 3 single scutes on
the left a doubling of the rest of the band.

Fetus I. has in position 10 from the left margin of band 6 a
double scute.

Fetus IV. has in position 5 to the right of the middle of band
3 a double scute.

Set K. 16 cf (Table Bi}.

The mother shows a right lateral and a median doubling of
band i.

Fetus II. has one double scute in band i, situated in the middle
of the left side of band i, another case of centre-lateral sym-
metry reversal.

Set K. 35, 9 (Table Bi}.

The mother has a rather unusual anomaly in the form of two
"'split" scutes separated by several single scutes, occurring in
band I, in places 10 and 15 respectively from the left-hand
margin. In band 2 we find the bilateral equivalent of these
-elements in the form of one double scute situated in position 13
from the right.

Fetus I. is the only one of the offspring to inherit the maternal
anomaly. There is in band i a double scute situated 13 places
from the left in the reverse position of that seen in band 2 of
the mother. The genetic relations here are quite clear, but
why should only one out of four offspring show an inherited
character?

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO.

TABLE B2.

f

r-G3

^

In

*

5 2: 6V

X

*

^" 2:ff/

JL

^4)

^6- 8 :65-

C4£<T

r

*:^ a

(M

C:« |

\±

$ Z-C3

u

Z4 '< z

'

Od / * ~ i>

o ' '8

y'

9

1 1

r? 9 o

L

>

9

1 7

y 9 ^

TT

(

6

/

SJ '

_LL

^

K

P^

>

Bd 1

5. 6 _,<

I-1-

32

p

2

^6 ±

ll

<-

-'g

5^(58

P

1

:64

^7

F

3

.63

7 ^

</ r- V

8 5^/4

U

7

:£f

rtR

IE

'

:G3

S «--

/

1

HE

3

:C/

^

7

r

JE

5

-:64-

V

\y^

JJL

v

4J

/

• 68

TJT

k ~~

€

ei-

V

pf

V
12

~->

r.s

3

tn

f

.65

V

9-^

. 14 cf

The mother has an incipient double band of two scutes in
positions 5 and 6 to the left of the middle of band 5.

Fetus I. has a double scute in position 4 to the right of the
middle of band 8. This shows reversed symmetry and a shifting
down the primary axis of the genetic unit inherited from the
mother.

24 H. H. NEWMAN.

Fetus II. has a double scute in position 2 from the left-hand
margin of band 3. This is evidently a reversal of the conditions
seen in the mother and is also a complex reversal, with shifted
position on the primary axis, of the condition seen in its twin
partner, fetus I.

Fetus IV. has a double scute in band 6, situated with refer-
ence to the middle of the left half band in about the same relation
as is the peculiarity of the mother with reference to the middle of
the whole band. In other words it is 5 scutes to the left of the
middle of the left half of band 6, while that of the mother is
situated 5 scutes to the left of the middle of band 5. Considering
the rarity of anomalies of any sort in the bands 3, 4, 5 and 6,
there can be no doubt of the genetic identity of these elements.
Granting this the interpretation offered is not forced.

Set K. 75, 9 (Table B 2}.

Here the mother has two double scutes, one near the middle of
the right half of band I and another to the right of the middle of
band 2.

Fetus I. has a double scute in band 2 identical in position with
one of those of the mother.

Fetus IV. has in band 8, in a position with reference to the
middle identical with one of those of the mother and that of
fetus I., an incipient band doubling of 2 scutes. This is another
example of the shifting of an anterior element to a position
further down the primary axis. It also emphasizes the genetic
identity of band and scute doublings.

Set €.46, rf1 (Table B 2}.

The mother shows a reduplicated double scute anomaly in
bands 2 and 6 in positions 2 and 3 respectively from the right-
hand margin.

Fetus I. shows in band 2 an exactly reversed mirror-image
reversal of the anomaly in band 2 of the mother and in addition
an extensive doubling in the middle of band I. This illustrates
three points; (a) reduplication down the primary axis, (b) the
genetic relation between scute and band doubling, (c) the half-
band type of reversal of symmetry.

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO.

TABLE

I

3o

/

C

3 1

9

\
z

^
/-

>

/

TT

/£

->-

'jd

i

Z. /i

/Z

~ i

I

3C

I ^

61

>

7

V
23

->

5

I

3

7-

G

i

i

9

V

ri

t

i-

>

7

fur

3o

tfb

30

I;

id -t

2

!nr

32 !

^ •.

G

23

6

->

8

C.90?

f*

3-

J i

*

S(

>

;s

17

T>

;

JL

3

f C

GG

2C

>

/2

-j>

6

nr

Eg

T

Q

i

G

3

3

7

"TO"

8

7

c

•}

2

3

3d /

-UL

3

6'

c

>

a

^

C 63 £

I

3:€

'3

*

-^

'JIT

<

24

3

•d

-2

'HE

• Q

«5~3

k/

i

O

J-S

X

HL

G : 6'3

6-3,

3L

C.729

3d /

30

Al.

/3

/3c//

4 G4

C 32

8

56

It should be noted by way of summary that in all cases in which
the mother has a band anomaly, one or more offspring of a set
have usually a similar band anomaly; but sometimes a scute
anomaly is found to be the expression in the offspring of a mater-
nal band anomaly. In several cases where there are band
anomalies in the offspring a scute anomaly is found in the mother.

26 H. H. NEWMAN.

It has been shown clearly in several sets of offspring that there
is an extremely close genetic relation between band anomalies
and anomalies of single scutes. The two types of anomaly are
merely more or less extensive expressions of the same genetic
factor. Whenever a mother shows either type of anomaly, one
or more offspring show one or the other type. The doubling
factor is therefore strikingly dominant in the Mendelian sense.
When we find sets of fetuses that show anomalies and the mothers
of these show no anomalies, we must conclude that the anomaly
is paternal, for a dominant factor in the mother, would appear
phenotypically if present genotypically. It is interesting to

note that almost the same number of sets have normal as have
anomalous mothers.

ANOMALOUS OFFSPRING OF NORMAL MOTHERS.

There are in the present collection 14 sets of fetuses showing
band anomalies, the mothers of which showed neither band nor
scute anomalies. In some ways these sets are better for our
purposes than are those derived from anomalous mothers, be-
cause we can be sure of the uniparental character of the inheri-
tance. Since the mothers show no anomalies, those in the off-
spring, being unquestionably inherited, must have come from the
father. It would be highly interesting to know the conditions
in the fathers, but the latter are inaccessible. Since there is no
sex difference with regard to the anomalies, we may assume, how-
ever, that the same state of affairs would be revealed for the
paternal as that made out for the maternal relation.

Some of the most interesting cases illustrating the symmetrical
distribution of anomalies, intra- and inter-individually, are found
among the offspring from normal mothers, and I shall give a
complete tabulation of these sets, calling attention in the text
only to certain of the more striking conditions, or to situations
differing from those dealt with in the sets derived from anomalous
mothers.

SetK.2, 9 (Fig. 13).

This set perhaps better than any other shows the phenomenon

of symmetry reversal, or mirror-imaging among the quadruplets.

Fetuses I. and III., which face one another across the vesicle,

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO. 2"J

and are the secondary bud derivatives of II. and IV. respectively,
are exact symmetrical reversals of each other, I. having the
anomaly only on the left and III. only on the right.

Fetuses II. and IV., the two primary bud individuals, both
show the anomaly bilaterally, but are also reversals, in that II.
has the more extensive doubling 9-9 on the right and IV. has the

more extensive doubling 9-8 on the left. It is interesting to

note that the primary individuals II. and IV. are bilateral in

their expression of the anomaly, and that the secondary indi-
viduals I. and III. are unilateral.

Set A. 64, cf (Fig. 14).

This set is remarkable for the striking identity of detail
exhibited by the anomalies of the different fetuses.

Fetus I. has both halves of band I completely double.

Fetus II. has both halves partially double and exactly bi-
laterally symmetrical, the doubling starting 3 scutes from the
margin on each side and extending for 15-15 scutes.

Fetuses III. and IV are both exactly like II. on the left side
and exactly like I. on the right.

There is exactly the same amount of doubling in I. plus II.

28

H. H. NEWMAN.

as there is in III. plus IV., but the distribution of the fully
doubled and incompletely doubled half bands is different. In
the case of the left-hand pair I. and II., I. gets both completely
doubled halves and II. gets both incompletely doubled halves;

14

while in the case of III. and IV. both get a completely and an
incompletely doubled half. There is no symmetry reversal in
this set.

Set A. 96, d" (Fig. 15}.

Fetus I. has band I double except for 5 single scutes on the
right margin.

Fetus II. has a similar anomaly bilaterally.

Fetuses III. and IV. are very much alike but differ from I.
and II. in the extent of the single part of the band, which is much
longer than in the other pair. The pairing of fetuses is very
clear, but no symmetry reversal is shown. One of four fetuses
(II.) is bilateral.

Set K. 80, cf (Fig. 16).

Fetuses I. and II. are practically identical. In both band I is
doubled, except for four single scutes on the left.

Fetuses III. and IV. show the incomplete doubling bilaterally

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO. 2 9

16

3O H. H. NEWMAN.

with a larger and smaller series of double scutes arranged on
opposite sides. Fetus III. has the long series 28-29 six scutes
from the right, while fetus IV. has a series of 29-29 scutes 15
scutes from the left and extending past the middle. This is an
imperfect reversal of symmetry. There is a better reversal in
connection with the short series of double scutes, for in III. the
series of 10-12 scutes is situated 5 scutes from the left, and in IV.
the series of 9-10 scutes is situated 6 scutes from the right.

Set A. 101 (Table 62) has the anomaly confined to one pair of
fetuses (I. and II.) and the expression is very different in the two.
It is well to note that although the doubling is nearly complete
in II. it starts after 6 single scutes at the margin, as is the case
on both sides of I., which shows a remarkably exact bilateral
symmetry.

In the remaining sets, K. 8, K. 12, C. 90, C. 63, C. 69, C. 72,
C. 88, C. 32 and C. 41 (Table B2 and 3), band doubling is found in
only one fetus of a set. In most sets however scute anomalies,
which are usually rather definitely related to the band anomaly,
are present in one or more fetuses. Set K. 8 is remarkable for the
large amount of reduplication down the axis of scute anomalies,
for example fetus III. has double scutes in bands I, 3, 5, 7, and 8,
those on bands 5 and 7 being reversals of each other. It is
clear that all of the double scutes are in the same region of the
band as is the double series in fetus I., and the majority are on
the same side of the body.

The remaining sets show nothing that has not already been
seen in sets previously described.

SUMMARY AND CONCLUSIONS.

Band anomalies are of the nature of irregularities in the normal
regular rows of scutes that make up the typical band, and consist
of more or less extensive regional doubling of rows of scutes in
a band. This condition is typical for the non-banded parts of
the armor but quite rare in the banded regions. It is practically
confined to the band or bands nearest the non-banded region.
Band anomalies occur in only about three per cent, of individuals
and an examination of over two thousand adult individuals
taken at random shows no duplicate anomalies.

HEREDITY AND ORGANIC SYMMETRY IN ARMADILLO. 3!

These anomalies are strongly inherited but are subject to more
or less modification. Sometimes a band anomaly unilaterally
placed may be inherited unilaterally but with a reversed sym-
metry to that of the mother, or it may be bilateral. Sometimes
an extensive doubling is inherited both as a similar band doubling,
and as a single double scute in the individuals of a single poly-
embryonic set of offspring. Contrariwise a double scute in the
mother may be inherited as a more or less extensive unilateral
or bilateral band doubling. The peculiarity may also be re-
duplicated down the primary axis in two or more bands.

Frequently in unilateral anomalies the different fetuses of a
set show reversed symmetry or mirror-imaging, but it is even
more common to find the unilateral anomaly on the same side
of most of the individuals, or bilaterally in one or more of them.

In a number of cases the anomalies in different fetuses of a
set are so strikingly identical as to indicate a rigid predetermina-
tion of the details of the character, but in other cases there
appears to be only a predetermination of a generalized anomaly
that expresses itself to a greater or less extent in the various
embryos. In terms of Mendelian inheritance we may say that
an anomaly factor is inherited as a dominant character, but its
distribution among the fetuses of a set and its location and
extent are due to varying ontogenetic or epigenetic factors.

The distribution of these inherited anomalies among the
various fetuses of the sets furnishes much interesting data, which,
together with data on the inheritance of double scutes, to be
presented subsequently, furnishes the basis of a discussion of
several general questions and especially those involved in the
concept of organic symmetry.

This subject is reserved for a subsequent paper of the present
series.

BIBLIOGRAPHY.

Newman, H. H. and Patterson, J. T.

'09 A Case of Normal Identical Quadruplets in the Nine-banded Armadillo
and Its Bearings on the Problems of Identical Twins and of Sex Deter-
mination. BIOL. BULLETIN, Vol. 17, No. 3.

'10 The Development of the Nine-banded Armadillo from the Primitive Streak
Stage to Birth; with Especial Reference to the Question of Specific Poly-
embryony. Journal Morphology, Vol. 21, No. 3.

32 H. H. NEWMAN.

'n The Limits of Hereditary Control in Armadillo Quadruplets: A Study of

Blastogenic Variation. Journal^Morphology, Vol. 22, No. 4.
Newman, H. H.

'i3a The Natural History of the'Nine-banded Armadillo of Texas. American

Nat., Vol. 47, Sept., 1913.
'i3b The Modes of Inheritance of Aggregates of Meristic (Integral) Variates in

the Polyembryonic Offspring of the Nine-banded Armadillo. Journal

Exper. Zool., Vol. 15, No. 2.

FACTORS CONCERNED IN THE PRODUCTION OF

MITOSIS IN ORGANISMS DISPLAYING

CELL CONSTANCY.1

H. J. VAN CLEAVE.

The condition of absolute identity in cellular structure found
in individuals of many species of Metazoa has led the writer to
undertake an analysis of the factors governing mitosis in these
forms. Loeb ('12: 4) has called attention to the fact that the
first attempt to reduce the phenomena characteristic of life to
purely physico-chemical terms is found in the works of Lavoisier
and Laplace ('80) which indicated that the heat produced in the
body of a warm-blooded animal equalled that given off by a
burning candle when the amounts of carbon dioxide produced
in the two instances were equal. These results have been so
much amplified by later workers that today no one doubts that
the general processes having to do with the phenomena of meta-
bolism are identical with the so-called purely physical and
chemical reactions occurring outside the living body. In fact
while metabolism, movement, and irritability are generally
granted as probably due to the workings of the same principles
that govern the inanimate realm the fourth of the vital properties,
that of reproduction, is in great part still unexplained. No
problem has offered more secure harbor for a final possibility
of the presence of some supernatural force than has this one
dealing with the processes of reproduction. The statement of
Davenport ('08: i); "The vital processes are chemical processes,
taking place in a highly complex, very unstable, constantly
changing substance, whose activities we call life," embodies the
modern conception of the nature of life phenomena in general, but
no attempt has been made to show the actual possibility of
explaining all life processes on this basis.

An analysis of the possible factors determining or directing

1 Contributions from the Zoological Laboratory of the University of Illinois,
No. 43.

34 H. J. VAN CLEAVE.

the course of nuclear and cell division is extremely difficult on
account of problems involved in eliminating factors until but
a single causal agent is operative. However, of the possible
factors involved in the production of mitosis and of cleavage
two groups are clearly recognizable, namely: (i) Environmental
factors, and (2) Internal factors. Any change which occurs
within the living organism must find explanation upon the basis
of one of these two groups of factors or upon the basis of a com-
bination of them. It is the purpose of this paper to show the
relationship of these two groups of factors in the production
of mitosis among those organisms which are made up of a fixed

number of cells.

i. ENVIRONMENTAL FACTORS.

If environmental factors either caused or directly controlled
the phenomena of mitosis the fact of cell constancy could not
exist, for no two individuals, and more strongly no two groups of
individuals even of the same species, develop under absolutely
the same environmental conditions. Consequently if environ-
mental factors were the limiting factors in mitosis no two indi-
viduals would of necessity contain the same number of cells.
It is generally granted that the mitotic process may be accelerated
or retarded through the application of purely external stimuli.
At least in those organisms made up of a fixed number of somatic
cells such an acceleration or retardation of the mitotic process
could result in nothing beyond a modification of the normal
rate of the process and could in no wise be considered as a direct
factor in the determination of the extent of the series of mitotic
divisions during the cleavage stages and in subsequent develop-
ment. If the role of temperature, for instance, be examined as
a possible controlling factor in the mitotic process it becomes
apparent that if this alone and directly controlled the number of
mitotic divisions through which the developing organism should
pass there could be but slight possibility of any two individuals
having identical numbers of somatic cells. For under conditions
of nature no two individuals are at all times during their develop-
ment under absolutely identical temperature relations. A similar
variability of conditions for the development of different indi-
viduals exists in the case of almost all of the other environmental

FACTORS CONCERNED IN THE PRODUCTION OF MITOSIS. 35

factors which might be considered as directly influencing tin-
process of mitosis.

Furthermore if environmental conditions could directly control
and determine the mitotic process, cell division would continue
indefinitely through the life of the organism as long as the
external conditions remained favorable for it. Again the data
brought from the field of cell-constancy show the impossibility
of such factors being operative in a controlling manner. In the
development of any organism or part of organism the powers
of reproduction of most of the somatic cells are restricted to
definite periods in the early stages of development, usually
preceding the introduction of histological differentiation. Thus
in the genus Eorhynclius the writer ('14: 280) has shown that in
no instance did he find any adult worm displaying abnormal
numbers of somatic nuclei or presenting any evidence whatever
of further division of the somatic nuclei after the adult body form
had been attained. These observations were upon over two
hundred individuals collected during a period of four years at
various localities and at various seasons of the year.

In conclusion, if the determination of the mitotic process were
to find its explanation in terms of environmental factors, or to
a continuous production of combinations of environmental
factors during the process of development, then in every instance
the number of mitotic divisions a fertilized egg would undergo
would be a direct resultant of the complex of environmental
factors operative during its process of development. At least
in those forms with a high degree of cell constancy it seems
obvious that purely environmental factors have but one relation-
ship to the process of mitosis and that consists in the modifiability
of the rate, either as an acceleration or as a retardation. In this
respect purely environmental factors have the same general
effect upon the process of mitosis as they exert upon purely
physical and chemical reactions.

2. INTERNAL FACTORS.

Our incomplete knowledge of the finer structure of the cellr
and of the nucleus in particular, make it impossible to associate
the control of the mitotic process directly with any structure or

36 H. J. VAN CLEAVE.

chemical bodies found in the cell. However on the basis of
facts pointed out earlier in this paper it seems certain that since
there is no possibility of environmental factors acting as the
controlling element in mitosis the ultimate cause of the process
must be sought within the cell. In this connection invaluable
support is found in the field of experimental embryology.
Morgan ('95, '01, and '03) and Driesch ('98 and 'oo) have both
shown that in the embryos of echinoderms developed from
isolated blastomeres of the two cell stage the number of cells
present at any point in the development is approximately half
of the number present in a normal embryo. Similarly from one
of the blastomeres of the eight cell stage the gastrula is composed
of only one eighth the number of cells found in the normal
gastrula. Loeb ('06: 59) interprets these results as supporting
the hypothesis of Sachs ('93 and '95) which regards the factors
determining cleavage controlled by the ability of each nucleus
to gather around itself and control a definite amount of proto-
plasm. Yet what determines the amount of protoplasm present
in the developing individual? The cytoplasm is constantly
being replaced through the processes of anabolism which experi-
ments with enucleated cells have shown to be under the control
of the nucleus. Consequently it seems that the view just stated
comes not much short of being an argument in a circle. Does
the amount of cytoplasm determine the number of nuclei that
are to be formed or is the numerical relation of nuclei to cyto-
plasm a mutual one brought about not through the influence of
•either cytoplasm or of nucleus but through some fundamental
factor which determines the number of nuclei and at the same
time indirectly the amount of cytoplasm that is to be formed?
The writer interprets the data of Morgan and of Driesch in an
entirely different manner. If by the first division of the egg
there are set apart two units, each of which has the possibility
of developing into a given number of cells by the process of
mitosis and this tendency is retained, even though the two units
become separated, it seems logical to conclude that within the
fertilized egg there are resident potencies which through the
process of mitosis become divided between the two daughter
nuclei of the first and then of each succeeding generation of cells.

FACTORS CONCERNED IN THE PRODUCTION OF MITOSIS. 37

As to the nature of this partition with each mitotic division two
explanations present themselves. According to the simplest of
these mitosis may be the result of the direct chemical activity of
certain substances, present in the fertilized egg, which become
used up in the mitotic process so that each cell of the two cell
stage receives an equal amount of the substance present in the
fertilized egg after the amount necessary for the first mitotic
division has been eliminated. On this hypothesis with each
succeeding mitotic division the materials resident in the de-
veloping embryo become partitioned and dissipated in the process
of development. Thus if x equals the entire amount of the
material for the execution of the mitotic process present in the
fertilized egg, and a the amount required for the realization of
the first mitotic division then each cell of the first cleavage would
receive (x -- a)/2. In this manner each succeeding division
would reduce the amount of the material present until the
amount apportioned to each cell would be less than the amount
required for the execution of the mitotic process, thus bringing
about an automatic check upon the course of the series of nuclear
divisions. The following objection to such a hypothesis shows
its weak point. Upon this basis each blastomere of the early
stages would be required to produce the same number of cells, a
supposition which the facts of cell lineage do not support.

The second explanation of the method of control over the
number of divisions of the nucleus seems more natural and does
not convey so much of the idea of predestination, in that it
requires less emphasis upon the inherent qualities of the fertilized
egg. According to this hypothesis mitosis is just as markedly
a result of chemical processes going on within the egg as indicated
above but the individual cells may in varying degrees retain the
power of synthesizing the materials necessary to initiate and
carry out the mitotic division of the nucleus. In the application
of this explanation the check to the course of the nuclear divisions
may come as the result of an accumulation of materials within
the cell, probably as metabolic by products, which serve to retard
and finally to prohibit the chemical activity incident to nuclear
division. Thus at the end of any definite period of physio-
logical activity the organism will be composed of a definite

3§ H. J. VAN CLEAVE.

number of cells and so upon attaining maturity will have a
fixed number of somatic cells which are unable to divide farther
on account of the presence of the inhibiting materials. It must
be remembered that throughout this discussion principles are
being laid down for the determination of the mitotic process in
organisms or parts of organisms displaying cell constancy. The
objection involving the variability in the number of cells arising
from the early blastomeres, cited in connection with the dis-
cussion of the influence of environmental factors, finds no
grounds here. In fact it lends its support to this second hypothe-
sis. If the ultimate check to the mitotic process comes as the
result of an accumulation of metabolic by products, then the
rate of the nuclear division, which unquestionably may be in-
fluenced by environmental factors, would tend but to be directly
proportional to the rate of the accumulation of the inhibiting
elements, though the actual amount of the inhibitors necessary
to terminate the series of mitotic divisions would remain the
same. Evidently the accumulation of the inhibitors proceeds
unequally in various tissues of the metazoan body. The degree
of the differentiation of the cell and, in all probability connected
with this, the nature of the cell membranes, determines the rate
of the accumulation of the inhibiting factors.

The impossibility of explaining the phenomena of cell con-
stancy on the basis of mitotic control by environmental factors
and the facts from experimental embryology in their bearing
upon the control of mitotic division by internal factors, lead to
the evident conclusion that during the development of cell
constant forms the mitotic process is controlled and determined
from within the cell but its rate may be regulated by factors
which are considered purely environmental.

CONCLUSIONS.

1. At least in those forms displaying a high degree of uni-
formity in the number and arrangement of their component cells
and nuclei, environmental factors such as ordinarily influence
physical and chemical changes play no direct part in the deter-
mination of the number of mitotic divisions of the somatic nuclei.

2. The only influence of normal environmental factors in these

(••ACTORS CONCERNED IN* THE PRODUCTION OF MITOSIS. 39

forms consists in the modification of the rate of the process of
mitosis.

3. The direct control of the mitotic process must be sought in
the chemical activity going on within the cells of the developing
individual.

4. Experimental evidence indicates that with each cleavage of
an egg with determinate cleavage there is retained a definite
relationship between the number of any given cleavage and the
total number of cleavages that the embryo would undergo even
in those cases where the blastomeres become isolated in t he-
early stages of development.

5. The foregoing would indicate clearly that within the cells
derived from the fertilized egg there are present factors or
potencies which exert direct control over the number of mitotic
divisions which shall ensue.

6. The fact that not all of the blastomeres of the early cleavages
produce the same number of cells indicates that the number of
cells produced must be controlled by conditions developing as
the process of development progresses rather than by the parti-
tion and distribution of some definite materials present in the
egg at a time prior to the first cleavage.

7. In tissues which retain the power of continued mitotic
division, as for example in the formation of the germ cells and
in tissues which have widely varying numbers of cells, the expla-
nation of the inconstant nature of the numbers of cells produced
may be sought in the acquisition of the power of eliminating
from the cell those materials which in the course of the process
of metabolism tend to accumulate and serve as inhibitors to the
mitotic process.

LITERATURE CITED.

Davenport, C. B.

'08 Experimental Morphology. New York.
Driesch, H.

'98 Von der Beendigung morphogener Elementarprocesse. Arch. f. Entw.-
mech., 6: 198-227.

'oo Die isolirten Blastomeren des Echinidenkeimes. Arch. f. Entw.-mech.,

10: 361-410.
Loeb, J.

'06 The Dynamics of Living Matter. Columbia U. Biol., Ser. VI 1 1.

'12 The Mechanistic Conception of Life; Biological Essays. Chicago.

40 H. J. VAN CLEAVE.

Morgan, T. H.

'95 Studies on the Partial Larvae of Sphserechinus. Arch. f. Entw.-mech., 2:
81-126.

'01 The Proportionate Development of Partial Embryos. Arch. f. Entw.-
mech., 13: 416-435.

'03 The Gastrulation of Partial Embryos of Sphserechinus. Arch. f. Entw.-
mech., 16: 117-124.
Sachs, J.

'93 Physiologische Notizen VI. Ueber einige Beziehungen der specifischen
Grosse der Pflanzen zu ihrer Organisation. Flora, 77: 49-81.

'95 Physiologische Notizen IX. Weitere Betrachtungen iiber Energiden und

Zellen. Flora, 81: 405-434.
Van Cleave, H. J.

'14 Studies on Cell Constancy in the Genus Eorhynchus. Jour. Morph., 25:
253-299.

THE PRODUCTION OF MALES AND FEMALES CON-
TROLLED BY FOOD CONDITIONS IN THE
ENGLISH HYDATINA SENTA.

DAVID D. WHITNEY.

It has been found recently in an American parthenogenetic
strain of the rotifer, Hydatina senta, from New Jersey that a
continuous diet of the colorless protozoan flagellate, Polytoma,
causes practically all females to be produced.1 This production
of only females can be maintained in this manner through
generation after generation for several years. If, however, the
diet is suddenly changed to a green protozoan flagellate, Chlamy-
domonas, that is in an active state, males can be produced in
great numbers.

It has been suggested that perhaps this phenomenon of the
regulation of the two sexes by food conditions is peculiar to this
particular strain of New Jersey Hydatina senta and is not an
universal characteristic of the species. Fortunately it has been
possible to test this hypothesis on an English strain of Hydatina
senta and some very clear and conclusive results have been
obtained.

I am greatly indebted to A. F. Shull for the stock of English
rotifers with the accompanying data. 'The English line was
received from Mr. C. F. Rousselet, who collected them as resting
eggs in mud at the bottom of a duck pond in England in August,
1912. They were sent to me about November I, 1912, in dry
dirt. The first ones hatched a few days later and from them the
line sent to you has been reared by parthenogenesis ever since.
As they were reared there was a generation about every three
days and so you have about the 27Oth generation." The stock
of this English line was received from Dr. Shull on January 7,

I9I5-
The females of this strain produce fewer offspring in the same

period of time than the females of the New Jersey strain. This

1 Jour. Exper. Zoo/., Vol. 17, November, 1914.

41

DAVID D. WHITNEY.

is probably due to the fact that the former strain had been
reproducing parthenogenetically for several years and had
become weakened whereas the latter strain was developed from
a resting egg in November, 1914, and at the present time is as
vigorous as at first.

The method of making the media and rearing the two kinds
of protozoan food cultures has been given in detail in a former
paper and will not be repeated here, excepting a few additional
words about the green food. When the Chlamydomonas had
been 2-3 weeks in the same media the individuals were large
and more or less quiescent. In this condition they were only
moderately effective although in one or two instances some
that had been two to three months in the same media were very
effective. When they had been in new bouillon media for 2-7
days with considerable sunshine at a temperature of about
28° C. they were of various sizes and seemed to be the most
effective. However, there seems to be a considerable amount of
chance in getting the Chlamydomonas in the optimum condition
for every experiment. This probably explains the varying per-
centages of the male-producing females obtained in the different
experiments.

As it seems that the effect of a uniform diet on this rotifer has
not been sufficiently emphasized the following Table (I.) con-

TABLE I.

SHOWING THAT A CONSTANT AND UNIFORM DIET OF THE COLORLESS PROTOZOA,

Polytoma, REPRESSES THE PRODUCTION OF MALE-PRODUCING FEMALES

AND CAUSES THE PRODUCTION OF FEMALE-PRODUCING FEMALES.

Kepi on a Uniform Polytoma Diet.

Race.

Generation.

c?9

9 9

<ic 9 9

Time.

A

1-288

02

2,s6=;

06 +

22 months

B. . . .

I-lSl

8

I 7^1

oo +

14 months

C. .

1-288

o

2.7/LO

IOO

22 months

cerning this point in three pedigreed races has been compiled
from the records in an earlier paper.1 This shows that a con-
stant and uniform diet of Polytoma, has repressed the production
of males for nearly 300 generations. The long period of time
1 BIOL. BULL., Vol. 22, March, 1912.

PRODUCTION OF MALES AND FK MALES.

43

and the large number of generations through which these races
were reared and observed would seem to warrant the reliability
of the results.

Table II. contains the results of some new observations made
in November and December of 1914. The main point of interest
in it is that it shows at what place or stage after the diet has been
changed females may be isolated that will produce a high per-
centage of male-producing daughters. Adult females (mothers)
that were put into the new diet of green food produced in it,

TABLE II.

As A CONTROL FOR TABLE III. AND ALSO SHOWING THE HIGH PERCENTAGE OF

MALE-PRODUCING DAUGHTERS PRODUCED AFTER THEIR MOTHERS HAD

BEEN TRANSFERRED FROM THE Polyloma DIET TO THE GREEN

Chlamydomonas DIET FOR 12 HOURS.

Experiments During November and December 1914 on a New Jersey Strain of

Hydatina senta.

Control Reared and Continued Adult 99s from the Control Reared on a Polvtotna

on a Polytoiim Diet.

Diet Transferred to a Chlamydontonas Diet.

Ex peri-

Daughters Produced

Daughters Produced

9 9
Mothers.

Daughters.

9 9
Mothers.

During First
12 Hours.

During Hours
12-24.

9 9

c? 9

','«? 9

9 9

<79

* cf 9

9 9

d"9

jtd-9

I

10

IO

0

o

6 14

20

59 +

O 22 IOO

2

IO

IO

o o

5 33

6

15 +

14 12

46 +

3

10

10

o

0

5 24

6

20

8

32

80

4

10

10

0

0

5 No record

7

17

70

5

10

10

o

o

IO

31

61

66*

6

10

IO

o

o

5 7

13

61 +

7

10

IO

0 0

40 No record

9

104

92 +

8

10

10

0 0

40

"

24

126

84

9

IO

10

0 0

40

" "

30

129

81 +

10 10

IO

O O IO

ii ii

I

39

97 +

Control continued through

30 additional generations

300

300

o

o

Total ....

400

400

o

o

166 78

45

36 +

124 542

81 +

during 1-12 hours, 15 per cent.-6o per cent, of male-producing
daughters, but the same adult females in the second 12 hours
produced a much higher proportion of male-producing daughters,
usually from 80 per cent.-ioo per cent. In most of the experi-
ments the adult females were taken out of the green food after
they had been in it 12 hours, in more or less of sunlight, and
placed in filtered culture water from a general stock battery jar

44

DAVID D. WHITNEY.

in which various protozoa and rotifers were living. They were
left in this filtered water for about 12 hours, during the night,
without food. During this time each female laid 3-4 eggs and
the next morning the old females were taken out, Polytoma food
was added, and the eggs allowed to hatch. Several hours later
the young females that hatched from these eggs were isolated in
separate watch glasses and fed Polytoma and a species of a small
Euglena.

In Table III. this same fact is shown again to be true for the
English species. However, this is a relatively minor point and

TABLE III.

SHOWING THAT WHEN THE ENGLISH STRAIN WAS SUBJECTED TO A UNIFORM AND
CONSTANT DIET OF THE COLORLESS PROTOZOA, Polytoma, ONLY FEMALE-
PRODUCING FEMALES WERE PRODUCED BUT THAT WHEN IT WAS
SUBJECTED TO A SUDDEN CHANGE OF DIET FROM THE
Polytoma TO A GREEN PROTOZOA, Chlamydomonas,
AS HIGH AS 85 PER CENT. OF MALE-PRODUCING
FEMALES WAS PRODUCED.

Experiments During January and February 1015 on an English Strain of Hydatina

senta.

Control Reared and Continued

Adult 99s from the Control Reared on a Polytoma

on a Polytoma Diet.

Diet Transferred to a Chlamydomonas Diet.

Experi-

Daughters Produced

Daughters Produced

ment.

9,-.

Daughters.

During First

During Hours

9
Mothers.

,, ". ° 12 Hours.
Mothers.

12-24.

9 9

c?9

Jtd-9

9 9

c?9

**9

9 9

c? 9

£ c? 9

I

10

10

0

o

5 7

4

36 +

4

14

77 +

2

10

10

o

0

5 8

2

20

5

5

50

3

10

IO

o

0

5

4

3

42 +

9

3

25

4

IO

IO

0

0

5

14

4 i 22 +

8

4

33 +

5

IO

10

0

0

20

29

I I

27 +

16

27

62 +

6

10

10

0

o

2O

38

o

17 +

7

40

85 +

7

10

10

o

o

15

No record

15

13

46 +

8

IO

IO

o

0

20

30

18

37 +

No record

9

IO

10

0

0

90

No record

50

76

60 +

Control continued through

20 additional generations

200

2OO

o

o

Total ....

290

290

0

0

185

130

50

27 +

114 182

61 +

may only be of importance to any one wishing to obtain cyto-
logical material.

If Table III. is compared further with Table II. it is evident
that they are very similar in all points with the exception that

PRODUCTION OF MALES AND FEMALES. 45

the percentages of male-producing females in Table III. are
somewhat lower than those in Table II. However, both tables
show that a continuous and uniform diet of Polytoma causes only
female-producing daughters to be produced while a sudden
change of the food from Polytoma to Chlamydomonas causes
male-producing daughters to be produced in great numbers.

SUMMARY.

1. A uniform diet of the colorless protozoan flagellate, Poly-
toma, continued for 22 months through 288 generations prac-
tically suppressed the production of males and caused only
females to be produced in the rotifer, Hydatina senta.

2. After the adult females had been transferred from the diet
of Polytoma to the diet of Chlamydomonas for 12 hours they
produced a higher percentage of male-producing daughters than
they did during these 12 hours.

3. The sudden change of food from a constant diet of Polytoma
to a diet of the green Chlamydomonas caused male-producing
daughters to be produced in the English Hydatina senta about
as readily as in the New Jersey Hydatina senta.

BIOLOGICAL LABORATORY,
WESLEYAN UNIVERSITY,

MlDDLETOWN, CONN.

MICROT^NIELLA CLYMENELIwE, A NEW GENUS
AND NEW SPECIES OF COLONIAL GREGARINES.

GARY N. CALKINS.

While examining the gut contents of marine annelids at Woods
Hole during the month of June, 1914, a number of new gregarines
were discovered. The majority of these were ordinary types
which could not be placed systematically without knowledge
of the sporulation stages. One form, however, found in the
digestive tract of Clymenella torquata, deserves mention because
of its remarkable novelty.

To obtain the material, the worms were opened along the mid-
dorsal line; sections of the digestive tract about half an inch
long, were removed and teased in salt solution on cover glasses.
The material thus prepared was examined while fresh, and, if
interesting, was then fixed in sublimate acetic and stained. In
this way I have obtained more than a hundred specimens of the
curious organism described here.

The name Microtceniella is given because of the Taenia-like
structure of the organism (Figs. 1-5). • The initial single cell
resembles a scolex and the chain of cells, formed partly from the
initial cell and partly by cell division of the daughter individuals,
resembles a chain of proglottids. In sporozoan terminology
these may be termed the primite and satellites respectively,
although this involves some elasticity in the use of the terms
primite and satellite which are usually employed to designate the
primary and secondary individuals in a chain of gregarines formed
by secondary association.

In the living state, the organism is colorless but with the char-
acteristic dense protoplasmic structure of the gregarines. The
average dimensions of the chain are 84 /JL in length and 15 n in
width, the size of mature chains varying but little from the
average.

The earliest stage found was an initial cell with two nuclei

and the beginnings of a septum cutting off a first satellite (Fig. i).

46

MICROTJENIELLA CLYMENELL^E.

47

All intermediate stages between this young form and stages
pictured in Figs. 4 and 5 were found. Each satellite contains a
single nucleus except in the division stages and in the terminal
cells of the oldest satellites.

New cells are continuously given off by the primitc; these
reproduce by division, the plane of division being at right angles
to the long axis of the aggregate for the first one or two divisions.
In some chains the successive individuals furnish almost every
stage for an ideal diagram of cell division (Fig. 3). In the older

m&

&&(-' —
p<M

^"Sk -&

satellites the plane of division changes from transverse to longi-
tudinal in respect to the long axis of the chain. In this \vay
two terminal chains may originate, but such branch chains never
become very long owing to further development and to separation
of the satellites from the parent colony.

After from eight to ten satellites have been produced thus by
division, the terminal cells become more spherical, their nuclei
divide twice, and without cytoplasmic division, giving rise to a
terminal group of cells each with four nuclei (Fig. 4). This stage
must persist for some time for several specimens were found

48

GARY N. CALKINS.

without evidence of further advance. In one specimen (Fig. 5),
a group of terminal satellites, evidently the result of this phase
of activity in two branched terminal chains, were present. Ulti-
mately, division planes appear in the cytoplasm around each of
the four nuclei, and four naked cells are produced. The fate
of these cells is not known; possibly theyjbecome new primites

and repeat the cycle, but more probably, they are gametes
which unite to form sporoblasts characteristic of the gregarines.
I shall try to work out their further history during the coming
June.

This rather paradoxical organism has given me a great deal of
trouble. Is it a metazoon or a protozoon? Is the "individual"
the entire chain or is it a chain of individuals? Although the
aggregate has a definite morph and a definite ontogeny I am
inclined to regard it as a chain of individuals in which original
reproduction by division has become modified to asymmetrical
division or budding in the primite, but is retained by the satel-
lites. Originally, it may be assumed, the daughter individuals
became separated and after a few divisions finally gave rise to

MICROT/ENIELLA CLYMENELL/E. 49

sporonts. If the entire chain is regarded as the individual we
are met by the same difficulty which confronts us in connection
with the colonial flagellates.

The possibility of this being a plant form has not been over-
looked. The absence of definite walls together with the method
of cell division and terminal cell changes are opposed to what is
known of types in the group of fungi. To make sure, however, I
showed the specimens to my colleague Professor R. A. Harper
who gives me permission to quote him to the effect that he finds
nothing in the structure or the history of this organism that would
justify him in placing it with plant forms.

Amongst protozoa, nothing, so far as I am aware, has been
described of like nature. Leger's Tceniocystis is a gregarine with
numerous septa and with external annulations which give it the
appearance of a T&nia, but it is a single cell and has a single
nucleus. The absence of motile organs in the present form, and
its mode of reproduction, leave no grounds for placing it otherwise
than with the sporozoa, and here the only possible place for it is
with the gregarines. There is some evidence that the often
sharply pointed end of the primite is the attaching portion, but
no stage showing such attachment has been found in the smears.

I would classify Microtaniella as a colonial parasite belonging
to the order Gregarinida, suborder Schizogregarina, of which it
should form a new subdivision.

COLUMBIA UNIVERSITY,
March, 1915.

THE BLINDNESS OF THE CAVE FAUNA AND THE
ARTIFICIAL PRODUCTION OF BLIND FISH
EMBRYOS BY HETEROGENEOUS HY-
BRIDIZATION AND BY LOW
TEMPERATURES.1

(WITH 13 FIGURES.)
JACQUES LOEB.

I.

While many of the animals inhabiting caves are blind or have
degenerated eyes, the same phenomenon is rarely found among
animals that live in the open. At first sight this seems to suggest
that the disuse of the eyes in the complete darkness of the cave
has gradually led to the degeneration of the eyes and this
idea seems at one time to have been widely accepted. In forming
a judgment of the connection between the darkness of the caves
and the blindness of cave dwellers we must remember that some
of the cave dwellers have perfectly normal eyes. Thus Eigen-
mann, to wThom we owe the most thorough study of this subject,
points out that of the four species of salamanders living habitually
in North American caves two have apparently quite normal eyes.
They are Spelerpes maculicauda and Spelerpes stejnegeri. Two
others living in caves have quite degenerate eyes, Typhlotriton
spelcEus and Typhlomolge rathbuni.2 If disuse is the direct cause
of blindness we must inquire why Spelerpes is not blind.

Another difficulty arises from the fact that a blind fish,
Typhlogobius, is found in the open (on the coast of southern
California) in shallow water, where it lives under rocks in holes
occupied by shrimps. One must again ask the question: How
can it happen that in spite of the exposure to light Typhlogobius
is blind?

The most important fact is perhaps the one found by Eigen-

1 From the Rockefeller Institute for Medical Research, New York.

2 Eigenmann, "Cave Vertebrates of America," Carnegie Institution Publica-
tions, Washington, 1909.

50

THE BLINDNESS OF THE CAVE FAUNA. 51

mann in the fishes of the family of Amblyopsidae. Six species
of this group live permanently in caves and are not found in the
open, while one lives permanently in the open and is never
found in caves and one comes from subterranean springs. The
one form which is only found in the open, Chologaster cornutiis,
has a simplified retina aside from having a comparatively small
eye, in other words, its eye is not normal. This indicates the
possibility that the other representatives which are only found
in caves also might have abnormal eyes even if they never had
lived in caves.

Through these facts the old idea becomes doubtful, namely,
that the cave animals had originally been animals with normal
eyes in which the disuse had led to a gradual hereditary de-
generation of the eyes. Instead we must consider the possi-
bility that in the blind cave animals as well as in the blind animals-,
which live in the open the tendency towards blindness developed
independently of presence or absence of light. From this point
of view the tendency towrard degeneration of the eye appears
as a hereditary mutation comparable to the inherited glaucoma
which is known in the human. Glaucoma is a form of blindness
caused by the atrophy of the optic nerve in consequence of an
increased intraocular pressure and this high pressure seems to be
caused by a certain disturbance in the circulation in the eye..
The fact that these patients are born with normal eyes and do-
not become blind until later in life shows that lack of light has
no share in the development of this hereditary disease or muta-
tion. The question then remains, how can we account under
such an assumption for the fact that blind animals are so preva-
lent in caves and so rare in the open? We shall return to this,
question later on.

II.

We will now review briefly the literature dealing with observa-
tions and experiments on the influence of light on the formation
of eyes. The fact that the eyes of mammals are formed in
complete darkness (in the uterus) may serve at the outset as a
warning against overestimating the effect of light on the forma-
tion of eyes. F. Payne1 has raised sixty-nine generations of a fly

1 F. Payne, BIOL. BULL., XXI., p. 297, 1911.

52 JACQUES LOEB. «

(Drosophila) in complete darkness without noticing any changes
in their eyes or in their sensitiveness to light.

Recently Uhlenhuth1 has demonstrated in a very striking
way that the development of the eye does not depend upon the
influence of the light or upon its functioning. He transplanted
the eyes of young salamanders into different parts of their
bodies where they were no longer connected with the optic
nerves. The eyes after transplantation underwent a degenera-
tion which was followed by a complete regeneration. Uhlen-

BEATING •
HEART

FIG. i.

huth showed that this regeneration took place in complete dark-
ness and that in salamanders, kept in the dark for 15 months,
the transplanted eyes remained normal. Here the eyes were no
longer in connection with the central nervous system, received
no light, and could not possibly have functioned and yet they
regenerated and kept normal. The degeneration which took
place in the eyes immediately after they were transplanted was

1 Read at the meeting of Anatomists at St. Louis, December 1914.

THE BLINDNESS OF THE CAM. I \1 \A.

53

apparently due to the interruption of the circulation in the eye
and the regeneration probably set in with the reestablishment of
the circulation in the transplanted organ.

The older literature had many observations which were as-
sumed to prove an influence of light on development but the
writer has shown in a former paper that these conclusions were
not supported by the facts on which they rested.

The writer has for years paid special attention to the possi-
bility that light might influence the formation of organs in

BE.ATING HEART

FIG. 2.

animals, but he has succeeded in rinding only one form in which
such an influence can with certainty be demonstrated, namely
the hydroid Eudendrium. In his experiments in Naples he had
noticed that the polyps of this form regenerate better in the
light than in the dark and in Woods Hole he could convince
himself that the stems of Eudendrium cannot regenerate their
polyps if kept permanently in the dark.2

2 Loeb, " Uber den Einfluss des Lichtes auf die Organbildung bei Tieren," Arch.f.
d. ges. Physiol., LXIII., p. 273, 1896; Goldfarb, Journal Experimental Zoology, 3,
129, 1906; 8, 133, 1910.

54

JACQUES LOEB.

Kammerer1 seems to be the only author who still takes it for
granted that cave animals owe the degeneration of their eyes to
lack of light, and his support for this view consists in the state-
ment that five young cave salamanders (Proteus) developed
larger eyes under certain (somewhat puzzling) conditions of
illumination. In other cases the eyes of Proteus remained
unaltered in the light. It would be advisable to make certain
whether or not there are two varieties of Proteus and moreover
it would be desirable to repeat these experiments on a larger
scale.

The wrriter wishes to publish in this note the results of some
experiments made in 1912, which prove that in the fish Fundulus
it is comparatively easy to produce embryos with degenerated
eyes by various means except by the one Kammerer holds

FIG. 3.

responsible for this phenomenon, namely lack of light. The
reader may in passing be reminded of the fact that in this form
Stockard and later McClendon induced the formation of cyclo-

1 Kammerer, Arch. f. Rntwicklngsmechn. d. Organ., XXXIII., p. 350, 1912.

THE BLINDNESS OF THE CAVE FAUNA.

55

pean eyes by altering the constitution of the sea water (addition
of magnesium salts or of alcohols and other narcotics.)1 The
writer's attention to an abnormal development of the eyes in
this embryo was first attracted in his experiments on hetero-
geneous hybridization.2 He noticed that among the anomalies
noticeable in heterogeneous fish hybrids the lack of circulation
was the most common, but that incomplete or abnormal develop-
ment of the eyes was not infrequent. Such embryos often give
the impression that they have no eyes at all, though in reality a
histological analysis would probably show that some of the
organs of the eye are present. But it is safe to say that the

BEATING HEWVT

FIG. 4.

condition of the eyes is such as to make them unfit to form a
retina image and for this reason we may call these embryos with
abnormal (or apparently lacking) eyes blind; in the same sense
in which this term is used in the case of the blind salamanders or

1 Stockard, Am. Jour, of Anatomy, 10, 369, 1910; Jour. Ex p. Zoo/., 4, 165, 1907;
6, 286, 1909; McClendon, Am. Journ. Physiol., 29, 289, 1912.

- Loeb, "Heredity in Heterogeneous Hybrids," Jour. Alorphol., XXIII., p. i,
1912.

56 JACQUES LOEB.

fish found in caves wThich also have some of the elements of an
eye but in a condition unfit for use. The term "blind fish" is
therefore used in the following as synonym with imperfect or

degenerated eyes.

III.

BLINDNESS PRODUCED THROUGH HETEROGENEOUS CROSSINGS.
Figs, i and 2 are camera drawings of hybrids between Fundulus
heteroclitus 9 and Menidia d" . The first embryo is eight, the
second fourteen days old. In neither of these two embryos are
eyes noticeable, though of course a histological examination
might have divulged rudimentary eyes. The embryos are in this

BEATING
HEART

FIG. 5.

respect to all appearances comparable to the blind cave fish or
salamanders.

It is a known fact that certain blind animals have in an early
stage a comparatively better developed eye than in later stages,
i. e., that the eye either stops developing early while the develop-
ment of the rest of the body continues or that the eyes degenerate

THE BLINDNESS OF THE CAVE I- A I NA.

57

directly. The writer noticed not infrequently that the eyes of
the cross between Fundiiliis and Menidia were normal in the
beginning but appeared more abnormal the further the develop-
ment progressed. We will illustrate this statement by two series
of drawings. The first case is illustrated by Figs. 3 to 6. Fig. 3
shows the embryo four days old. The embryo has normally
developed eyes writh lenses. The embryo's heart, which is
visible in front, was not yet beating and no pigment wras formed.
The same embryo is shown two days later in Fig. 4. The eyes
instead of developing any further are less distinct. Fig. 5 is the
same embryo two days later (eight days old). The eyes are

/BEATING HEART

FIG. 6.

hardly recognizable, it is difficult to decide without sectioning
whether the lens still exists. The internal ear is distinguish-
able. The heart is pulsating.

Fig. 6 shows the same embryo at the age of two \veeks.
Apparently the pigment is the only organ of the eye which is
developed.

58 JACQUES LOEB.

Figs. 7 and 8 illustrate the same fact, that a rather normal eye
with a lens was formed while later on only a very imperfect eye
remained. Fig. 7 shows the embryo when four days old with a
lens and an apparently normal eye, while five days later (Fig. 8)
no eye was perceptible.

IV.

BLIND FISH PRODUCED FROM PURE BREEDS OF FUNDULUS

HETEROCLITUS.

The writer was under the impression that the abnormalities in
the development of the heterogeneous hybrids in fish were due
to the influence of the foreign sperm as in the similar case of the

FIG. 7.

fertilization of the egg of the sea urchin with the sperm of the
star fish, where the mortality is also enormous. In order to
prove this assumption he tried to find a method which would
allow him to produce embryos with abnormal eyes in the pure
breeds of heteroclitus . He has already mentioned in a former
publication that such abnormal embryos were obtained when

THE BLINDNESS OF THE CAVE FAUNA.

59

their development was retarded by putting them into sea water
to which some KCN was added. Fig. 9 shows a blind embryo
obtained in this way. The eggs were put immediately after
fertilization into a weaker solution of KCN in sea water where
they remained sixteen days. They developed slowly. The
embryo was drawn two days after it had been put back into

FIG. 8.

normal sea water. The heart-beat and circulation were estab-
lished. This method of producing blind embryos is not reliable
and need not be discussed any further.

A better method was found by exposing the newly fertilized
eggs to a low temperature (from between o to 2° C.) for some
time. The writer found that the egg of Fundulus can be put
for weeks into such a low temperature after the embryo is once
formed, without any injury to the latter. As soon as it is put
back to room temperature it continues to develop. This corre-
sponds with the idea that the low temperature only retards the
chemical reactions underlying development. If, however, the

6o

JACQUES LOEB.

egg of the same fish is exposed to such a low temperature imme-
diately after fertilization or when the eggs are in the process of
early segmentation they suffer severely.

Thus in one experiment the eggs of Fundulus (fertilized with
sperm of their own species) were distributed into a number of
Erlenmeyer flasks and put on ice immediately after fertilization
and kept at a temperature of between o and 2° C. After four

BEATING HEART

FIG. 9.

hours, seven and a half hours, twenty-three hours, thirty-two
hours, forty-eight hours, sixty-four, and ninety-six hours one lot
was returned to room temperature. All the eggs taken out after
forty-eight hours or later were dead; few of those taken out
after thirty-two and twenty-three hours survived. A consider-
able percentage were abnormal and resembled the heterogeneous
hybrids between heteroditus 9 and Menidia cf . Most of those
taken out after four and seven and a half hours survived, but
about 20 per cent, to 30 per cent, were abnormal. Among the
abnormal embryos a number of blind fish were found.

THE BLINDNESS OF THE CAVE FA I \ A.

61

A second lot of eggs were put on ice four and a half hours after
fertilization when they were in the 4-cell stage or beyond. These
eggs were just as sensitive to the effect of the low temperature
as those put on ice immediately after fertilization.

This experiment was repeated four times with the same result.
In another experiment the eggs were allowed to develop to
about the 128-cell stage at normal temperature before they
were put into a temperature of o° (about fifteen hours after
fertilization). Eggs that were kept at that low temperature for

BEATING
HEftRT

FIG. 10.

two days were still able to develop into normal embryos but
those that had been kept three days at the temperature of from
o to 2° C. were practically all dead.

When the eggs were put into a temperature of from o to 2° C.,
after the embryo was formed and the circulation established,
they could resist the low temperature for weeks. When put
back to normal temperature they recovered and developed
normally.

Further experiments are required to ascertain more accurately
when the eggs become immune to the temperature of from o° to 2°.

62

JACQUES LOEB.

It is only this very low temperature at or very near the freezing
point which is injurious for the newly fertilized eggs. If the
temperature is a little higher, e. g., 7° C., the newly fertilized
eggs can live for weeks in it without being injured. Thus in the
experiment just mentioned some of the eggs were put immediately
after fertilization into a temperature of about 7° C. These eggs

FIG. ii.

developed very slowly but no abnormal embryos were observed
although some of the eggs were kept at a temperature of 7° C.
for four weeks. Those that were kept still longer at that tem-
perature suffered but probably not from the low temperature
but from the fact that the flasks in which they were kept were
closed.

These experiments thus establish the interesting fact that
immediately after fertilization the eggs of Fundulus are rapidly
injured or killed when exposed to a temperature of o° or a little
above, while after the embryo is formed they can be exposed to

THE BLINDNESS OF THE CAVE FAUNA. 63

such a temperature for a month without suffering. The newly
fertilized eggs can, however, be exposed without injury to a
temperature of 7° for weeks and possibly longer.

As already stated, among those eggs which were put into a
temperature of from o to 2° C. a certain number had defective
eyes. Thus the egg Fig. 10 wras put immediately after fertiliza-
tion into a temperature of between o and 2° for twenty-four
hours and then put back to room temperature. The egg de-
veloped like a typical heterogeneous hybrid. Fig. 10 represents
the embryo when two weeks old; it had a beating heart but no
circulation. No eyes are noticeable. Fig. n represents an
embryo from another experiment; in this too the egg had been

^-REMNANT Of EMBRYO

BEATING HEftRT

FIG. 12.

kept for four hours (immediately after fertilization) on ice
(o to 2°). Apparently no eyes are formed but the circulation is
established.

Figs. 12 and 13 are added to show to what extent the abnormal
embryos, produced by a short exposure of the egg to a very
low temperature (o°) immediately after fertilization, resemble
the monstrosities which are formed in the case of heterogeneous

64

JACQUES LOEB.

hybridization. The embryo in Fig. 12 consisted practically only
in a pulsating heart, what was left of the embryo was much less
conspicuous than it is in the drawing. The embryo was ten days
old. The egg had been put into the temperature of between o
and 2° when it was in the 2-cell stage and had remained there
for twenty-four hours.

The remnants of the embryo Fig. 13 had the same history. In

QGCASSIONAIJ
CONTRACTION
OF TAIL

REGION OF
BEATING HEART

FIG. 13.

this case, curiously enough, one eye with a lens and the tail was
all that could be recognized. The monstrosity when drawn was
eleven days old. An inexperienced observer might easily have
concluded that in this case only the eye had been developed,
while in reality the whole embryo had developed but the eye
had survived while other parts perished.
The analogy with teratomata is obvious.

THE BLINDNESS OF THE CAVE FAUNA. 65

V.

The writer was anxious to know whether it was possible to
produce deficiencies in the eyes of Fundulus by raising them in
the dark. It was necessary to carry on such an experiment with
a large number of embryos. Since it was possible that a short
exposure of the unfertilized egg or of the sperm to the light might
already have some effect, the females and males of Fundulus
heteroclitus to be used for the experiment were put into an abso-
lutely dark room where everything was prepared for the experi-
ment. The females and males were stripped of their sexual cells
in the dark and the jars containing both sperm and eggs were
put into dark boxes and kept in the dark room for four weeks.
Although many embryos had died, hundreds had survived. All
had perfectly normal eyes. This experiment confirms similar
experiments made by the writer in previous years.

All these experiments show that while it is comparatively easy
to produce blind Fundulus embryos, or, more correctly speaking,
fish with degenerated eyes, by heterogeneous hybridization or by
low temperature or by lack of oxygen (or by an excess of mag-
nesium salts or by alcohol, as shown by Stockard and McClendon)
no such result can be produced by lack of light.

VI.

If we consider all the facts in the case there is nothing at
present to warrant the assumption that the blind cave animals
owe the deficient development of their eyes to the lack of light,
since lack of light is according to our present knowledge a less
efficient agency in the causation of an abnormal development
of the eye than a number of other injurious influences. Under
these conditions we must be prepared to consider the possibility
that many if not all the blind species found in caves owe their
blindness to other influences than those of the cave. Eigenmann
states that no blood vessels enter the eye of the blind cave
salamander Typhlotriton. Since the experiments of Uhlenhuth
as well as those of Stockard and of the writer reported in this
paper indicate the importance of the circulation and of chemical
factors in the development of the eyes it is not impossible that
in the blind fish (Amblyopsidae) as well as in the blind sala-

66 JACQUES LOEB.

manders a hereditary disturbance in the circulation or nutrition
of the eye or its surroundings is the cause of the degeneration of
the eyes. There remains then the one difficulty mentioned in
the beginning, namely to account for the fact that the relative
number of blind species is greater in caves than in the open.

Eigenmann has shown that all those forms which live in caves
were adapted to life in the dark before they entered the cave.
These animals are all negatively heliotropic and positively
stereotropic and with these tropisms they would be forced to
enter the cave whenever they are put at the entrance to the cave.
Even those among the Amblyopsidae which live in the open have
those tropisms of the cave dweller. This eliminates the idea
that the cave adapted the animals for the life in the dark.

Eigenmann 's observation makes it also clear that blind animals
are comparatively rare in the open and that animals with normal
eyes are not in the majority in the caves. Only those animals
can thrive in the caves which for their feeding and mating do not
depend upon visual mechanisms; and conversely animals which
are not provided with visual mechanisms can only under excep-
tional conditions hold their own in the open where they meet
the competition of animals which can see. This would account
for the fact that in caves blind species are comparatively more
prevalent than in the open.

In spite of all this, Eigenmann is inclined to assume that the
darkness of the caves was a factor in promoting the blindness of
the cave fauna. While those Amblyopsidae which live in the
open have already abnormal eyes those species of the family
which are found in the caves have more degenerated eyes than
those in the open and Eigenmann is inclined to ascribe this
fact to an accumulated influence of the darkness. This would,
however, compel us to account in a similar way for the incipient
degeneracy of the eyes of those Amblyopsidae that have never
been found in caves and for the complete blindness of Typhlo-
gobius; and would leave us at a loss to account for the presence
of salamanders with perfectly normal eyes in the caves. It seems
to the writer that consistency would demand to consider a com-
mon mode of origin of all these blind forms, namely as mutations,
i. e., as the result of some factorial change in the germ the cause
and nature of which we are not yet able to define. Among the

THE BLINDNESS OF THE CAVE FAUNA. 67

Amblyopsidae various mutations in the state of the eye may be
constantly arising (as in the case of Drosophlla) ; and some of
these mutants, especially those that are perfectly blind, can not
hold their own in the open while in the caves they can preserve
and perpetuate themselves. This assumption would not exclude
the possibility that Kammerer's observation may have been
correctly interpreted by him, it would only provide for the
possibility that his conclusion can not be generalized — a possi-
bility which from the experiments mentioned in this paper
must be seriously considered.

SUMMARY.

1. It is shown that blind embryos or more correctly embryos
with degenerated eyes can be produced by heterogeneous hybridi-
zation in fish embryos (e. g., Fundulus heteroclitus 9 and Menidia
cf). Since in these cases as a rule no circulation exists the
inference is possible that the anomalous condition of the eye
may be due to lack of circulation.

2. Blind embryos of the pure breed of Fnndulus heteroclitus
may be produced by the addition of KCN to the sea water.

3. It is shown that immediately after fertilization (by sperm of
their own species) and during the early stages of segmentation
the egg of Fundnlns heteroclitus is rapidly killed or injured if it is.
exposed to a constant temperature of about o° (or slightly
above) ; while it may be exposed to a slightly higher temperature
(e. g., 7° C.) for weeks without being injured. If the egg is
exposed to the low temperature after the embryo is once formed
it can resist the low temperature of from o to 2° C. for weeks
without permanent injurious effects.

4. If eggs of Fundulus heteroclitus are fertilized with the sperm
of the same species and exposed immediately after fertilization
for a number of hours or a day to a temperature of between o
and 2° C. abnormal embryos can be produced a certain percentage
of which may show degenerated eyes.

5. Lack of light does not influence the development of the
eyes of Fundulus.

6. It is pointed out that internal mutational changes and not
lack of light may account for the blindness of certain cave fish
and salamanders.

THE ABSORPTION OF FAT BY FRESHWATER

MUSSELS.1

E. P. CHURCHILL, JR.

INTRODUCTION.

The present paper embodies the results of the first of a series
of investigations designed to ascertain whether or not animals
living in the water use food which is in solution in the surrounding
medium in addition to "formed" food such as plankton, etc.
If it is found that such food in solution is used, it is of importance
to learn whether it is absorbed by the alimentary tract alone or
to some extent by the epithelium of the outer body walls, espe-
cially that of the gills in the case of mussels.

Putter ('08) was the first to advance the theory that food
could be so taken. He based his conclusions in part on a com-
parison of the amount of carbon necessary for the maintenance
of the organism with the amount furnished by the plankton.
Putter considered the latter too small for the needs of the animal
and argued that it must use some carbon which is in solution
in the water, resulting from the decay and disintegration of
organic life. He further stated that the amount of material
found in the alimentary canal is never large enough to supply
the requisite quantity of carbon. Besides the alimentary canal,
the uncutinized epithelium of the outer surface of the body,
-especially that of the gills, was thought to function in absorbing-
dissolved food. This process went on in addition of course to
.the digestion of "formed" food by the alimentary canal. Putter
'concluded that the above conceptions applied to Protozoa,
Porifera, Echinoderms, Crustaceans, Mollusks and Fishes. He
tested the matter experimentally in two ways: first by noting
that goldfish and perch lived much longer in solutions of aspara-
gin, somatose and glycerin than in tap water: secondly by com-

1 Contribution from the United States Biological Station, Fairport, Iowa.
(Two of the experiments were carried out at the zoological laboratory of the Johns
Hopkins University.) Published by permission of the Commissioner of Fisheries.

68

THE ABSORPTION OF FAT BY FRESH-WATER MUSSELS. 69

paring the amount of oxygen needed to oxidize the lost weight of
tissues of actinians, tunicates and fish while kept in their natural
medium with the amount of oxygen actually used. As the
latter was found to be greater than the estimated quantity
needed he concluded that the extra oxygen was used in oxidizing
some food that had been taken from the water where it was
present in the form of a solute.

Knorrich (10), working with Daphnia, which lived 14 days in
sterilized hay solution, concluded that nutriment was absorbed
from the solution.

Kerb (10) kept eels in sugar solution and noted no diminution
of the amount of sugar from day to day. He obtained similar
results while working with Corethra larvae in sugar solutions.
Also he found that Daphnia lost in dry body weight as rapidly
in solutions of peptones as in tap water.

Wolff (10), working with Simocephalus, found that it lived
twice as long in bacteria-free water, which contained some dis-
solved carbon compounds, as it did in tap water. He made no
observations as to body weight lost or gained.

Lipschiitz (13) reviewed the entire subject including his own
previously published experiments along that line and offered
criticism of Putter's work. Lipschiitz noted that fish and eels
when kept in nutrient solutions lost as much weight as in tap
water. He also thought that Putter overestimated the amount
of material in solution in the water and underestimated the
carbon content of the plankton. His general conclusions are
the opposite of those of Putter.

Lund1 found that if Protozoa are kept in a weak soap solution
they will absorb fat from such solution through their body walls.
At his suggestion the fats used in the following experiments were
rendered soluble in water by saponification.

I am especially indebted to Dr. Caswell Grave, at whose sug-
gestion the work was undertaken, for his advice and aid on
many occasions and his supervision of the preparation of the
manuscript. I also desire to express my obligations to Dr. E. J.
Lund for suggestions concerning some of the chemical reactions

1 Dr. Lund's paper is not yet published, his results having been communicated
to me verbally.

7O E. P. CHURCHILL, JR.

discussed in the latter part of this paper; to Dr. G. L. Houser
for the use of several books and a microscope from the laboratories
of the University of Iowa; and to Dr. R. E. Coker for aid ren-
dered while the work was being carried on at the Fairport
Laboratory.

MATERIALS AND METHODS.

The mussels upon which the investigations were carried out
were individuals selected from the more common species found
in the Mississippi River near Fairport, Iowa. Both adult and
juvenile specimens were employed. Care was exercised to
choose for the experiments non-gravid mussels which were in a
seemingly healthy condition, "shoulder-raked" or hand collected
individuals being generally used in preference to those dragged
out of the water by "crow-foot" hooks.

The water used in the experiments wTas that of the general
supply for the Fairport Laboratory which had come from the
Mississippi River through a filter which removed the mud and
at least the greater part of the living animal and plant organisms.

The fats employed were those of olive oil and cottonseed oil.
The results were the same in both cases as far as could be observed.
The oils were saponified by boiling with sodium hydroxide, the
resulting solution being diluted to strengths varying from .001
per cent, to .005 per cent, in different cases. This method can
be used only with freshwater animals as the soap is precipitated
from the solution in salt water.

The adult mussels experimented with were kept in glass aquaria
containing about 5,000 c.c. of the fat solution of the desired
strength. The bottoms of the aquaria were covered to a depth
of an inch or two with coarse sand. Control individuals were
run side by side with the experiments in similar aquaria contain-
ing filtered water only. The juvenile mussels were generally kept
in smaller aquaria. In the cases of experiments extending over
more than twenty-four hours, the solutions were changed daily,
no effort being made to balance the aquaria. As the mussels
from the river are accustomed to a current of water this seemed
to be the closest approximation to natural conditions which it
was possible to obtain in the laboratory.

The tissues of various portions of the adult animals experi-

THE ABSORPTION OF FAT BY FRESH-WATER Ml -^KI.S. JI

men ted with and of the control mussels were sectioned by the
freezing method, stained with Sudan III. and mounted in
glycerin. Also sections of the tissues of the adult and of the
entire bodies of the juvenile mussels were made after fixation
in osmic acid and Muller's fixing fluid. In this method the higher
alcohols were avoided by clearing the sections in clove oil from
85 per cent, or 90 per cent, alcohol and infiltrating in paraffin from
the clove oil. The paraffin was removed with xylol and the
sections were mounted in xylol balsam. Such sections were
studied very shortly after mounting, as the fat is gradually
dissolved out by the xylol.

In some cases the fat solution was stained by dissolving Sudan
III. in the same to the point of saturation. The external appear-
ance of various parts of mussels which had been kept in such
stained solutions was noted. Also frozen sections of such mussels
were examined to ascertain whether or not stained fat had been
absorbed from the solution. In such cases the sections were
mounted in glycerin directly after cutting.

OBSERVATIONS.

The mussels were found to live readily in .001 per cent, to
.005 per cent, fat solutions. The only abnormal manifestation
to be observed was that a considerable quantity of mucus was
thrown off. Six adult individuals which were kept in fat solution
for one month wTere living at the close of that period. At the
expiration of the same length of time one of the control mussels
of this experiment had died. After the adoption of the method
of manipulating the aquaria outlined above, no mussels died
while kept in fat solutions varying in strength from .001 per
cent, to .005 per cent. However, beyond the fact that mussels
can live for at least one month in fat solutions of these strengths,
no evidence was secured concerning the relative longevity of
those kept in fat solutions and of the control individuals. No
weights were taken. Experiments calculated to ascertain
whether or not the absorption of fat from the solutions employed
is of advantage to mussels were deferred until the actual entrance
of the fat into the tissues of the animal was proved. Emphasis
was therefore laid upon the histological evidence of the absorption

72 E. P. CHURCHILL, JR.

of fat and it is such evidence that this paper is designed to set
forth.

Mussels which had been kept in unstained fat solution will be
considered first. Camera lucida drawings were made of typical
portions of the sections, care being taken to indicate as far as
possible the number and position of the fat droplets.1 In some
cases the fat was so massed together that individual droplets
were indistinguishable. Fat is represented in all the drawings
by heavy black dots or areas.

The case of an adult specimen of Quadrula ebena which had
been kept in .001 per cent, fat solution for 15 days will first be
discussed. Sections of the gill filaments of this mussel were
prepared after fixation in a solution of osmic acid and Muller's
fixative. In such sections an abundance of fat droplets was to
be seen in the epithelial cells and a number in corpuscles in the
blood vessel in the interior of each filament (Fig. i, PI. I.).
Very few fat droplets were found in sections of the control
individual prepared by the same method (Fig. 2, PL I.). Frozen
sections, after having been stained with Sudan III., gave results
parallel to those just stated — a large number of fat droplets in
the gill filaments of the mussels which had been kept in fat
solution for 15 days (Fig. 3, PI. I.), but practically nothing that
took the Sudan III. stain in the control individuals. The
epithelial lining of the water tubes of the gills of the individuals
which had been kept in fat solution was also quite crowded with
fat. In sections of the intestine of mussels which had been kept
in fat solution, fat droplets appeared in the epithelial lining.
Also in sections of palps and mantle fat droplets were revealed
but not in such abundance as was true in the case of the gills.
Fat droplets were quite numerous in the cells of the side of the
mantle next to the body but there were practically none in the
cells of the side lining the shell. The tissues of three adult
mussels which had been kept in fat solution for 15 days and of
two control individuals which had remained in filtered water for
the same period were examined by the two methods of preparing

1 The term "droplets" will in this paper be applied to the spherules of fat found
in the tissues of the mussels which were studied. These droplets usually were of
diameters varying from 5 to 10 microns, though in frozen sections of the gills in-
stances were found in which the diameter was as great as 20 microns.

THE ABSORPTION OF FAT BY FRESH-WAT]-: R Ml »KLS. 73

sections. The former all presented abundant evidence of the
presence of fat. The difference between those which had been
kept in the solutions and the control mussels was as striking as
that shown in Figs, i and 2.

The best results were obtained from the study of juvenile
mussels as in these cases the entire animal could be sectioned
serially and the distribution of the absorbed fat over the entire
body observed. A specimen of Anodonta imbecillis which had
been kept in .002 per cent, fat solution for 10 days and a control
specimen of the same species which had been kept in filtered
water for the same length of time were selected as typical cases
for study.

Figs. 4 and 5, PI. II., showing corresponding parts of the foot,
give an idea of the abundance and distribution of fat in the mussel
which had been kept in the fat solution and of the almost entire
absence of it in the case of the control specimen. (In this case
as well as in that of several of the other figures, no effort has been
made to draw the outlines of the epithelial cells, the intention
being merely to represent the relative amount and distribution
of the fat.) It will be noted that the fat is very abundant in
the epithelium. It is present in some quantity in the interior
of the folds of the foot and also within the deeper parts of the
foot, apparently adhering to muscle fibers. The fat droplets
exhibit some tendency here, and more markedly elsewhere, to
gather together in clumps or to form chains. Several blood
corpuscles which contain fat may be seen.

Fig. 6, PI. II., shows a portion of the foot of the same indi-
vidual, extending from the base of a fold to the blood vessel in
the center of the foot. An abundance of fat was observable in
the epithelium and quite numerous droplets appeared clinging to
muscle fibers or enclosed within corpuscles in the central part
of the foot.

Fig. 7, PI. II., representing a portion of a fold of the foot, was
drawn under the 1/12 oil immersion lens, an effort being made
to depict the exact number and position and also the relative
size of the fat droplets, muscle fibers and cells.

Fig. 8, PI. I., shows in (a) the cross section of a liver tubule of
the mussel used as the control, the few coarse stipples represent-

74 E. P. CHURCHILL, JR.

ing the fat present; and in (&) the cross section of the corre-
sponding part of the mussel which had been kept in the fat
solution. In the latter the liver cells are seen to be heavily
loaded with fat, in some cases practically filled with it. A very
great abundance of fat was found throughout the liver cells in all
sections of this mussel.

Figs. 9, 10, and n, PI. I., are drawings made from the sections
of the epithelial cells, respectively, of the intestine, of a gill fila-
ment and of the side of the mantle next to the body, of the
specimen of A . imbecillis which had been kept in the fat solution
for 10 days. They represent as accurately as could be drawn
with the camera lucida the relative amount and arrangement of
the fat droplets. Many corpuscles containing fat were found
in the blood lacunse immediately outside of the cells of the intes-
tine. Some of these corpuscles were lying at the bases of the
cells as shown in Fig. 9. Corpuscles containing fat were also
found in nearly all blood vessels of the gill filaments. Many
such corpuscles were in the position shown in Fig. 10, that is,
close against the bases of the cells of the filament. In the mantle
fewer corpuscles were seen. Here the fat was found adhering to
muscle fibers and strands of connective tissue.

Figs. 12 and 13, PI. III., represent mesenchyme cells and blood
corpuscles of the mussel which had been kept in the fat solution
and of the control individual respectively. In these sections may
be observed the relation of the fat droplets to the tissues of the
areas occupied by mesenchyme cells, which are especially
numerous in the dorsal part of the body of the mussel. Fat was
found scattered quite thickly throughout such regions in the
case of the animal which had been in the solution.

The gill of an adult specimen of Quadrula ebena was sectioned
immediately after it had been taken from the river. Fig. 14,
PL I., represents the relative amount of fat found in its gill
filaments. Somewhat more fat appeared in these sections than
in those prepared from the tissues of the control mussels which
had been kept in filtered water for 10 or 15 days but a very much
smaller quantity was found than in the mussels which had been
kept in fat solution.

The mussels which had been kept in the fat solution that had

THE ABSORPTION OF FAT BY FKESEI-WATER MUSSELS. 75

been stained with Sudan III. will now be considered. Before
beginning the experiment the valves of the shell were opened
sufficiently to permit the observation to be made that the tissues
of the animal to be experimented with were of normal color and
condition. The mussels were then kept in stained fat solution
for periods varying from 5 to 15 days. At the expiration of
these periods the gills of the adult and the gills, mantles and foot
of the juvenile animals were found to be red or pink in color.
Sections of the gills of such adult mussels were prepared by the
freezing method. In these sections pink colored fat droplets
were found within the epithelial cells of the gill filaments and of
the water tubes (Fig. 15, PI. I.). This fact furnishes additional
proof of the absorption of fat from the solution. It renders
negligible the consideration that the heavy loading of fat found
in each case in the mussels which had been kept in the fat solu-
tions might have been due to the chance use of an extraordinarily
fat individual. Some of the juvenile mussels which had taken
on a red color while they had been kept in the stained fat solution
were then transferred to filtered water. The color remained
visible in the foot and gills for more than a week in some cases.
This would tend to show that the red colored fat was absorbed
and oxidized within the organism and that the red color was not
due merely to the adherence of the stained solution to the
surface of the gill or foot.

In all, including those mussels which were kept in stained and
unstained fat solutions and the control individuals, the tissues of
twelve mussels were examined. The sections all revealed much
fat within the tissues of the mussels which had been kept in the
fat solutions and very little or none in those of the control
individuals. Consideration of the above results leaves no doubt
of the fact that fat may pass from a fat solution into the body
of the mussel. Now there remains the question of whether it is
all absorbed through the intestine or partly through the tissues
of the outer body walls. Experiments intended to throw light
upon this question were undertaken with juvenile mussels.
These had been kept in water containing little food for several
weeks so that their tissues contained little stored fat at the
beginning of the experiment.

76 E. P. CHURCHILL, JR.

Four mussels were kept in a .005 per cent, fat solution for the
several periods of 4, 7, 18 and 24 hours. A small amount of fat
was found in the epithelium of the intestine, mantle, gills and
foot of the individual of the 4-hour period. None was found in
the other parts of the body. About the same amount and dis-
tribution of fat was observable in the case of the mussel which
had been in the solution for 7 hours. In the animal which was
kept in the fat solution for 18 hours a very appreciable quantity
of fat was found in regions corresponding to those in which it
had been found in the mussels which had remained in the solution
for 4 and 7 hours. A small amount was noted in the tissues
immediately beneath the epithelium. None was found in the
deeper body tissues, a fact which is in striking contrast to the
case of the mussel which had been kept in fat solution for 10
days. The mussel which had remained in fat solution for 24
hours contained a considerable amount of fat in the same parts
of the body in which it was found in the cases of the individuals
which had remained in the solution for 4, 7, and 18 hours.
Even more fat was found in the epithelium of the gills, mantle
and foot than in that of the intestine (Figs. 16, 17, 18, 19 and 20,
PI. III.). A very small quantity of fat was observable in the
liver cells. In certain parts of the foot a moderate amount of
fat appeared immediately beneath the epithelium, some of which
fat was adhering to the muscle fibers. The fat droplets became
progressively less numerous as the tissues were studied to a
greater depth within the foot and none were found among the
muscle fibers or in the other tissues making up the main central
mass of the foot (Fig. 17, PI. III.). The corpuscles, especially
those found in the gills, contained fat in many cases. In the
gills the crowding of the corpuscles against the bases of the cells
of the filaments was marked (Figs. 18 and 19). Every appear-
ance was given that the fat was absorbed from the solution by
the epithelial cells of the gills and was taken up from them by
the corpuscles in some cases and in others thrown directly into
the plasma of the blood. Fig. 20, PI. III., shows a portion of the
epithelium of the mantle. Fat was seen in the epithelial cells
of the side next to the body and for a short distance beneath
them but it was not found scattered throughout the deeper parts

THE ABSORPTION OF FAT BY FRESH-WATKK Ml'SSELS. 77

of the mantle. No fat was found in the cells of the side of the
mantle next to the shell.

The observations made on the mussels which were kept in fat
solution for the periods of 4, 7, 18 and 24 hours may be sum-
marized as follows:

1. More fat was found in the epithelium of the mantle, gills
and foot than in that of the intestine.

2. In the case of the mussels which were kept in fat solution
for 4 and 7 hours, fat was found only in the various epithelia.

3. In the mussel which remained in fat solution for 18 hours,
much more fat than in the two above cases was found in the
epithelium and a small amount was observable at various points
beneath the epithelium, but none could be discerned distributed
throughout the deeper body tissues.

4. In the mussel which had remained in fat solution for 24
hours, a very marked quantity of fat was found in the epithelium
and a considerable amount appeared beneath the epithelium,
diminishing in quantity, however, toward the interior of the
body so that there was none distributed throughout the deeper
body tissues.

5. No fat could be seen in the epithelium of the side of the
mantle next to the shell but an abundance was present on the
side which had been in contact w'th the solution.

6. Scarcely any fat was found in the cells of the liver.

From these facts it is probable that fat is absorbed by the
, outer epithelium of the body as well as by that of the intestine.
Unless such were the case it would be very difficult to account
for the presence of as much fat in the epithelium of the gills,
mantle and foot as in that of the intestine, taking into considera-
tion the fact that the mussels had been kept in the fat solution
for such short periods, viz., intervals varying from 4 to 24 hours.
It would also be difficult to account for the fact that such a
very small amount of fat appeared within the deeper body tissues
of these mussels while a very marked quantity was found in the
epithelium of the outer body walls. It is very unlikely that the
blood carried such a large amount of fat from the intestine to
the outer epithelium in such a short time and during the same
time carried none to the muscle fibers, mesenchyme cells or

78 E. P. CHURCHILL, JR.

connective tissue or conveyed none to the liver, which is the
normal fat storing organ. If all the fat were absorbed by the
intestine it seems certain that the sections of one at least of the
mussels which had been kept in the fat solution for the short
periods would have revealed more fat within the deeper body
tissues than was found there. Wherever fat was found beneath
the outer epithelium of these mussels which had been in the
solutions for the short periods, much the larger quantity lay in
close proximity to that epithelium and the amount lessened to
zero toward the interior of the body. While the quantity of
fat beneath the epithelium increased with the length of the time
the mussel was kept in the solution, in all cases it was greatest
nearer the epithelium. Apparently the fat found lying closely
beneath the outer epithelium of the body had been absorbed
and passed into the deeper tissues by that epithelium and had
not come from the intestine. Also, if all the fat were absorbed
by the intestine and the blood did transport it almost entirely
to the outer epithelium it would seem most likely that the cells
of the side of the mantle next to the shell would receive at least
an observable amount.

In order further to test the question of whether or not fat
may be absorbed from a fat solution by the epithelial cells of
the outer body walls of the mussel the following experiment wras
performed. The valves of three adult mussels were wedged open
with bits of wood and the animals suspended on a wire rack
over the fat solution so that only the ventral parts of the mantle
and foot were immersed. The mouth and siphons were above
the solution so that it was not likely that any of the solution
could enter the intestine by way of the oral or anal openings.
One mussel was thus treated for 6 hours, the other two for 22
hours. They were compared with a fourth individual which
had not been in the solution, all four mussels having been removed
from the river only four or five days previous to the experiment.
One of the animals which had been treated as above for 22
hours did not have much fat throughout the deeper body tissues
but a moderate amount appeared in the epithelium of the part
of the mantle which had been immersed in the solution. The
other mussel which had been treated for 22 hours contained some

THE ABSORPTION OF FAT BY FRESII-WATKR MUSSELS. 79

fat throughout the body and a heavy amount in the epithelium
of the part of the mantle which had been exposed to the solution
(Fig. 2ia, PI. III.). The control individual for this experiment
contained much more fat throughout the deeper body tissues
than the other three individuals but very little in the epithelium
of the ventral part of the mantle (Fig. 216, PI. III.). The
ventral part of the foot of the mussel which had been so treated
for 6 hours revealed a moderate amount of fat in the epithelium
but scarcely any in other parts of the foot. The fat found
throughout the deeper body tissues of these individuals was of
course previously stored fat, the mussels, as stated above,
having been removed from the river but a short time. The
point intended to be made is that a heavier loading of fat is
found in the epithelium which had been in contact with the fat
solution than in that which had not. While there is a possi-
bility that some solution may have been carried by ciliary action
up the mantle to the mouth, this probably did not occur as, in
sections of the intestine, absolutely no fat was found in the
epithelial cells. It seems highly probable that the epithelium
of the mantle and foot absorbed the fat directly from the solution.
In regard to the mechanism of absorption of the fat by either
the intestine or body walls little can be offered. The matter
is quite complicated, some of the same possibilities entering here
that serve to render the method of absorption of fat by the
mammalian intestine as yet obscure. In some of the sections,
especially in those cut by the freezing method, numerous drop-
lets which took the stains used were found closely attached to
the outer ends of the epithelial cells of the gills or mantle. These
droplets were probably the fatty acids, a small part of which were
present in the soap solution due to hydrolysis, by which process
sodium hydroxide and the fatty acids would be formed; the
remaining droplets were no doubt due to a slight acidity of the
surface of the living cells resulting from the union of carbon
dioxide from the cells with the water, forming carbonic acid.
The fat may have been taken into the cells from these droplets
upon the surface by phagocytic action or by solution in the
plasma membrane, or it may have come directly from the sodium
soap, the collection of the droplets upon the surface of the cells

8O E. P. CHURCHILL, JR.

being a phenomenon extraneous to the process of absorption.
The latter interpretation seems the more probable as the droplets
were not found upon the surface of the cells in nearly all the
sections of tissues which had absorbed fat. The sodium may
enter the cells with the acid radical and later be separated, or the
fatty acid may be split off outside and enter the cell alone. As
the fatty acid radical is the constituent of the fat molecule which
takes the stain, only that radical could be followed in my prepa-
rations. In nearly all cases a fairly definite, though narrow, area
lies between the outer ends of the cells and the fat clusters
within. This band is the plasma membrane. The fact that no
fat droplets could be observed actually in this membrane points
to the conclusion that absorption of fat is effected by its solution
in the plasma membrane and precipitation within the cell.

The manner of transportation of the fat within the body
deserves mention. The bases of the cells of the intestine are in
contact with blood lacunae. Blood vessels traverse the gill
filaments. A meshwork of blood lacunae lies in the connective
tissue beneath the epithelium of the mantle and foot. A con-
siderable amount of fat is taken up and transported by the
corpuscles to various parts of the body. Evidence of this is
clearest in sections of the intestine and gills (Figs. 1,9, 10, 18
and 19), in which corpuscles containing fat are found in close
contact with the bases of the epithelial cells. Fat-loaded
corpuscles are found in the blood spaces throughout the body
of the mussel which had been kept in fat solution for several
days. However, sections of the foot and mantle reveal many fat
droplets lying immediately beneath the epithelium and clinging
to muscle fibers or to the connective tissue instead of being
contained in corpuscles (Figs. 4, 7, u and 17). No doubt,
besides being transported by corpuscles, fat is thrown directly
into the blood stream and carried thus to various parts of the
body. In some cases the fat droplets may have become attached
to muscle fibers, etc., as was observed in the sections, during the
preparation of the tissues. In other cases the attachment may
have occurred before the mussel was killed, the tissues to which
the fat droplets were adhering being those which were later to
oxidize the fat.

THE ABSORPTION OF FAT BY FRESH-WATER Ml-^I.LS. 8 1

SUMMARY.

1. Fat which is in solution in water can be absorbed by fresh-
water mussels.

2. Such absorption is accomplished by the epithelium of the
intestine and also most probably by that of the gills, mantle and
foot.

3. Fat is transported both by the blood corpuscles and by the
plasma directly.

LITERATURE CITED.
Kerb.

'10 liber den Nahrwert der im Wasser gelosten Stoffe. Revue d. ges. Hydro-

biologie u. Hydrog., Bd. 3, 1910.
Knorrich.

'10 Studien iiber die Ernahrungsbedingungen der Daphnien. Ploner Berichte,

1910.
Lipschiitz.

'10 Zur Frage iiber die Ernahrung der Fische. Z. f. allg. Phys., Bd. 12, 1910.

'10 Uber den Hungerstoffwechsel der Fische. Ebenda.

'13 Die Ernahrung der Wasserthiere durch die gelosten organischen Verbindun-

gen der Gewasser. (Eine Kritik.) Ergebnisse der Physiologie, Bd. 13,

1913-
Putter.

'08 Die Ernahrung der Wassertiere. Z. f. allg. Phys., Bd. 7, 1908.

'09 Die Ernahrung der Fische. Ebenda, Bd. 9, 1909.

'n Die Ernahrung der Wassertiere durch geloste organische Yerbindungen.

Pfliiger's Archiv, Bd. 137, 1911.

'n Der Stoffwechsel der Aktinien. Z. f. allg. Phys., Bd. 12, 1911.
Wolff.

'10 Ein einfacher Versuch zur Putterschen Theorie von der Ernahrung der

Wasserbewohner. Intern. Revue d. ges. Hydrob. u. Hydrogr., Bd. 2, 1910.

82 E. P. CHURCHILL. JR.

EXPLANATIONS OF PLATES.

All drawings were made with camera lucida; Leitz Oc. No. 4, Obj. No. 6, unless
otherwise indicated. The drawings are of sections prepared by the osmic acid
method unless otherwise stated. Fat is represented by heavy black dots or areas.

PLATE I.

FIG. i. Gill filament of adult Quadrula ebena after remaining in fat solution
1 5 days, a, epithelial cells, b, blood vessel, c, blood corpuscles, d, chitinousrods.

FIG. 2. Gill filament of the control for above figure: mussel had remained in
filtered water 15 days, a, b, c, d, same as in Fig. i.

FIG. 3. Gill filament of adult Quadrula ebena after remaining in fat solution
15 days. Frozen section, Sudan III.

FIG. 8. a, Cross section of liver tubule of juvenile A. imbecillis after remaining
in filtered water 10 days, the control, b, Cross section of liver tubule of juvenile
A. imbecillis after remaining in fat solution 10 days.

FIG. 9. Epithelial cells of anterior part of intestine of juv. A. imbecillis after
remaining in fat solution 10 days, a, epithelial cells, b, blood corpuscles.

FIG. 10. Gill filament of juv. A. imbecillis after remaining in fat solution 10
days, a, blood corpuscles.

FIG. ii. Portion of epithelium on side of mantle next to the body of juv. A. im-
becillis after remaining in fat solution 10 days.

FIG. 14. Gill filaments of adult Quadrula ebena sectioned immediately after
taken from the river.

FIG. 15. Portion of gill filament of adult Quadrula ebena after remaining in
Sudan stained fat solution 15 days. Black areas represent droplets of pink colored
fat. Frozen section.

BIOLOGICAL BULLETIN, VOL. XXIX.

E. ". CHURCHIIL, JR.

84 E. P. CHURCHILL, JR.

PLATE II.

FIG. 4. Portion of foot of juv. A. imbecillis after remaining in fat solution 10
days. Epithelial cells not indicated. Fat shown heavily massed in epithelium
and adhering to muscle fibers and in corpuscles, a, muscle fibers in cross section.
b, blood corpuscles, c, muscle fibers in longitudinal section.

FIG. 5. Portion of juv. A. imbecillis after remaining in filtered water 10 days.
Control for Fig. 4.

FIG. 6. Portion of foot of same individual as in Fig. 4. a, epithelium, b,
muscle fibers in longitudinal section, c, muscle fibers in cross section, d, blood
corpuscles.

FIG. 7. Portion of fold of foot of same individual as in Fig. 4. 1/12 oil im-
mersion, a, epithelial cells, b, muscle fibers.

BIOLOGICAL BULLETIN, VOL. XXIX.

PIATE II.

E. P. CHURCHILL, JR

86 E. P. CHURCHILL, JR.

PLATE III.

FIG. 12. a, mesenchyme cells and b, blood corpuscles of juv. A. imbecillis after
remaining in fat solution 10 days.

FIG. 13 a and b. Mesenchyme cells and corpuscles of the control kept in filtered
water.

FIG. 1 6. Epithelial cells of intestine of juv. Q. pustulosa after remaining in fat
solution 24 hours.

FIG. 17. Portion of foot of same individual as in Fig. 16.

FIGS. 18 AND 19. Gill filaments of same individual as in Fig. 16.

FIG. 20. Portion of side of mantle next to body of same individual as in Fig. 16.

FIG. 21. a. Epithelial cells of that portion of mantle immersed in fat solution
22 hours. The mouth of this mussel was kept above the solution, b. Epithelial
cells of corresponding part of mantle of mussel used as control for Fig. 210.

BIOLOGICAL BULLETIN, VOL. xxix.

PLATE

Fls.nb.

Fig, 21. a

E P. CHURCHILL, JR.

Vol. XXIX August, 1915. No. 2

BIOLOGICAL BULLETIN

SEXUAL REACTIONS BETWEEN HERMAPHRODITIC
AND DKECIOUS MUCORS.1

ALBERT FRANCIS BLAKESLEE.
CONNECTICUT AGRICULTURAL COLLEGE, STORRS, Coxx.

INTRODUCTION.

Conjugation among the Mucors has been assumed to repre-
sent the simplest type of reproduction acknowledgedly sexual
in character. Some even have denied the term sexual to the
union of morphologically similar gametes such as occur in this
group. The present article will present further evidence in
favor of the author's contention that although the process is
morphologically a simple one, conjugation in the Mucors is as
definitely a sexual process as the morphologically more complex
type of reproduction in higher forms and that the sexes seem
even more sharply distinct.

It has been shown by the writer ('04, '09) that the majority
of the forms among the Mucors are dioecious, with the sexes
separated in male and female races which are capable of being
propagated apparently to an indefinite number of vegetative
generations by means of nonsexual spores formed in sporangia.
In all the dioecious species carefully investigated the opposite
gametes, which are produced and unite to form zygospores
when the two sexual races of a given form are grown together,
do not appear to differ morphologically. Lacking a definite
criterion which an inequality of the gametes would have afforded,
the writer has provisionally designated the opposite sexes in
these forms by the signs ( + ) and ( — ) on account of a generally
greater vegetative luxuriance of one sex over the other. That

1 Report of investigation carried on, 1912-13, at the Carnegie Station for
Experimental Evolution, Cold Spring Harbor, N. Y.

87

88

ALBERT FRANCIS BLAKESLEE.

in reality the two sexes are represented in the (+) and ( — )
groups is shown by the sexual reaction which may occur not
only when the ( + ) and ( — ) races of the same species are grown
together and perfect zygospores are produced, but also by the
sexual reaction which may occur when ( + ) and ( — ) races be-
longing to different species are grown together. This reaction
between the opposite races of different species has been called
"imperfect hybridization" since it does not lead to the formation
of perfect hybrid zygospores, but usually stops short with the
formation of progametes, though occasionally gametes are pro-
duced which, however, never unite.

A sexual race of a dioecious species if grown between the ( + )
and ( — ) races of another test species used as a standard, will
show a line of imperfect hybridization on one side only. Some

Petri dish culture showing dark line of zygospores between the (+) and (— )
strains of a species of Choanephora (MC) and white lines of "imperfect hybridiza-
tion" between the strains of this species and the opposite strains of Mucor V. (MV).

of the hermaphroditic species, on the other hand, when similarly
grown, show a response to both ( + ) and ( — ) test races and
produce therefore two lines of imperfect hybridization. When

SEXUAL REACTIONS BETWEEN MUCORS. 89

the sexual reaction is very strong it may become evident to
the naked eye by the formation of a distinct white line as is
shown in the adjoining photograph.

Some few species in the hermaphroditic group are distinctly
heterogamic with a constant difference in size between the conju-
gating gametes. It seems reasonable to consider the larger
gamete female and the smaller male. Upon this basis, if a sexual
reaction could be established between these unequal gametes
and the ( + ) and ( — ) races, the race reacting with the larger
female gamete must be considered male, while the race reacting
with the smaller male gamete must be considered female. This
was pointed out in 1906 ('066) and in a recent article ('136) it
has been shown in brief that sexual reactions have been induced
between unisexual races and heterogamic species. The conclusion
is reached that the ( — ) race is to be considered male and the
(+) female. The present article will give the detailed results
which have formed the basis for this conclusion.

HETEROGAMIC HERMAPHRODITES.

Since inequality in the size of the gametes in heterogamic
species is used as the criterion of sex, it is essential that there
be no doubt as to the process of sexual reproduction in these
forms. With the exception of Syncephalis, which is difficult to
cultivate, the heterogamic forms at present known are Dicra-
nophora fulva and an undescribed American Dicranophora,.
neither of which are now in cultivation, Absidia spinosa,
Zygorhynchus heterogamus, Z. Moelleri and a number of forms
recently described (including Z. Vuillemini) which are perhaps
too closely allied to Z. Moelleri to be deserving of separate names.

The writer in a recent publication ('i3#) has made a restudy
of the genus Zygorhynchus and attempted to correct certain
misinterpretations of zygospore formation which had been
advanced for this genus. Moreau ('12, '13) also, both before
and since the publication mentioned, has insisted upon the
correct interpretation. So far as concerns the formation and
union of unequal gametes, the process is essentially similar to
that in Absidia spinosa which is the heterogamic species most
extensively used in the tests discussed in the present paper.

9O ALBERT FRANCIS BLAKESLEE.

Absidia spinosa was first found by Lendner and is correctly
described by him (Lendner '08). Its zygospores agree with
other Absidias in possessing circinate outgrowths which develop
from the suspensor soon after the formation of gametes. It
differs from other Absidias, which it resembles in its method of
nonsexual multiplication, by being hermaphroditic (cf. right-
hand side of Fig. 8, Plate I.) and heterogamic (Fig. 4). Figs. 1-7
represent stages in development in living material drawn at the
times indicated in the legend. A considerable number of conju-
gations have been followed in living material and stages have
been studied in stained and mounted material. Unequal
gametes, such as are shown in Fig. 4, unite, as in Fig. 6, and the
zygote thus formed grows into a mature zygospore such as is
shown in Fig. 7. In moist chamber cultures, drops of fluid
generally accumulate around the uniting gametes (Fig. 6), often
causing appearance of transverse lines that suggest additional
cross walls. They can be distinguished from cross walls, how-
ever, by forcing the conjugating filaments against condensed
moisture on the under side of the cover-glass. Perhaps the most
typical form of conjugation is that shown in Fig. 8, where a
rather stout branch on the right has applied itself to the more
delicate termination of the main axis causing the production
from the latter of a small male gamete, while it produces from
its own enlarged tip a large female gamete. Curved outgrowths
arise from the swollen suspensor but are not produced from the
side of the smaller gamete. In Zygorhynchus a septum is regu-
larly formed across the main axis just above the origin of the
stout conjugative branch. This septum is as regularly absent
in Absidia spinosa.

When for any reason the process of conjugation is arrested
after the formation but before the union of gametes, these sexual
cells may round themselves off, form thick walls and become
azygospores. This is brought about in greater or less degree
when the conjugating filaments in a moist chamber culture
become immersed in a drop of condensed water on the underside
of the cover-glass. Whatever the cause of the check to normal
conjugation, the size of the azygospores depends upon the size
of the gametes from which they develop. Thus in Fig. 33 it

SEXUAL REACTIONS BETWEEN MUCORS. QI

would be possible to label the azygospores of A. spinosa male
and female, from their relative size alone, even if one were not
subtended by a larger suspensor and surrounded by the out-
growths peculiar to the side of the female gamete. Similarly
Zygorhynchus may produce male and female azygospores from
adjacent gametes that have failed to fuse (Fig. 13). As will be
shown later, azygospores may be produced from gametes that
have formed as the result of a sexual reaction with a different
species. If the opposite gametes of hermaphroditic forms are
strictly male and female, it would theoretically be possible to
transform a hermaphroditic species into male and female races
by germinating its male and female azygospores. This has
been attempted ('o6a) and germinations have been obtained of
the azygospores of Sporodinia. The mycelia that have thus been
obtained produced only a very scanty growth and soon died
out before it was possible to test their sexual reaction. Germi-
nations of other azygospores, so far as the writer is aware, have
never been critically investigated.

HOMOGAMIC HERMAPHRODITES.

Spinellus fusiger, Sporodinia grandis — two forms found upon
fleshy fungi, — Mortierella polycephala and Mucor Genevensis—
are the only homogamic forms definitely known among herma-
phroditic Mucors. In none of these is there apparently a con-
stant difference in the size of the gametes. Neither Spinellus
(which is difficult to cultivate on artificial media) nor Sporo-
dinia shows sexual reactions with other species. Mortierella
has not been available for study. M. Genevensis, of which the
writer has races from four distinct sources as well as derived
mutants, shows a strong sexual reaction with both ( + ) and ( — )
sexes of dioecious species. The reaction has already been noted
('046, p. 311, pi. IV., fig. 56). Observation on living material
shows the reaction to be similar to imperfect hybridization
between different dioecious species. The formation of azygo-
spores, however, has been observed in different races of this
hemaphroditic species as a result of the stimulus of contact with
(+) and ( — ) strains of Mucor V. Azygospore formation will
be further discussed under Absidia spinosa below.

92 ALBERT FRANCIS BLAKESLEE.

DICECIOUS SPECIES.

No dioecious form of the Mucors is known to be regularly
heterogamic. In Rhizopus ('04^, Fig. 15) and in other dioecious
species the writer has found that the inequality in size between
conjugating gametes has no necessary relation to sex, since the
larger gamete is formed sometimes by one sex, and sometimes
by the other. In many species, however, the ( + ) race regularly
shows a decidedly greater vegetative vigor than its ( — ) mate.
This greater vigor of growth may be responsible for the fact that
the conjugative filaments arising from the (+) mycelium average
stouter than those from the ( — ), and this difference may be
accompanied by a similar difference in the gametes and suspen-
sors. One was generally able to recognize the ( + ) from the ( — )
sides of the line between the two sexes of Mucor V. by this
greater vigor of the (+) conjugative filaments but even in this
form where the distinction is most marked, the stouter of the
conjugating filaments is not invariably on the ( + ) side. Mucor
V. is a form found by the writer ('046) in 1904. It is apparently
not specifically distinct from M. hiemalis Wehmer, the sexual
races of which were isolated by Hagem ('08) and forms zygo-
spores, though in no great abundance, with the strains of this
name sent from the "Centralstelle" at Amsterdam. It differed
from the Amsterdam material, however, in its much greater
sexual vigor. In 1912 when the reactions described in the
present paper were investigated, the Amsterdam M. hiemalis
failed to show any reactions with other species tested, while
Mucor V. was sexually the most active of all the Mucors known.
At the present writing, however, Mucor V. has apparently
become weakened in sexual activity. Hagem ('08) reports a
similar loss of sexual activity in one of his strains of M. hiemalis.

REACTIONS BETWEEN DICECIOUS SPECIES AND ABSIDIA

SPINOSA.

The difficulties in technique involved in following the sexual
reactions in a thicket of filaments have been overcome by growing
the heterogamic hermaphrodite ( 8 ) in a Petri dish between the
(+) and ( — ) test strains and cutting channels in the nutrient
agar between the different growths. If the Petri dish be

SEXUAL REACTIONS BETWEEN MUCORS.

93

then inverted, the growth of the reacting filaments may be
followed in mid air in the channels, under the low power ob-
jective.

Absidia spinosa is the only heterogamic species which has
been found to give reactions with both ( + ) and ( — ) races and
Mucor V. is the only dioecious form that will react with both
male and female gametes of A. spinosa. These have accordingly
been most extensively employed in making the tests shown in
the accompanying diagram.

In the left-hand channel at A in the diagram, a filament from
the ( — ) race is shown giving a sexual reaction with the larger
9 gamete of the hermaphrodite, while in the right-hand channel
at B a filament of the ( + ) race is figured, showing asexual re-
action with the smaller cf gamete of the hermaphrodite. The

Diagrammatic representation of a Petri dish culture showing a heterogamic
hermaphroditic mucor ( S ) in the center separated by channels on either side from
the (+) and ( — ) races, respectively, of a dioecious species.

Sp., Sporangia containing spores by means of which the plant may be repro-
duced nonsexually.

i—6, stages in development of a hermaphroditic zygospore from unequal male
and female gametes.

A, sexual reaction between a (— ) filament and female gamete.

B, sexual reaction between a (+) filament and male gamete.

C, a male azygospore formed at stimulus of contact with a (+) filament.

male gamete, which has been cut off from a filament of the
hermaphrodite at the stimulus of contact with a (-f) hypha,
frequently surrounds itself with a thick wall and assumes the
appearance of a resting azygospore, as is shown at C.

Figs. 20-24, Plate II., are drawings from living material.
The female gamete of A. spinosa is recognized not only by its

94 ALBERT FRANCIS BLAKESLEE.

larger size but also by the curved outgrowths that arise from its
suspensor. It will be observed further from these figures that
the female gamete is formed from a rather stout branch which
appears to have been attracted by Mucor V. ( — ) and drawn
away from the delicate termination of the main axis with which
it would normally have conjugated (cf. Fig. 8). In order to
obtain such reactions as shown in Figs. 20-24, some care had
to be exercised in regard to the condition of the culture. If the
nutrient is too favorable for growth the filaments will choke
the channels and cannot be followed. If the nutrient is de-
ficient, conjugations will not occur. The depth and width of
the channels and the time at which they are cut must be care-
fully regulated or no reactions can be observed. They have
been obtained and studied in a sufficient number of cases, how-
ever, to leave no doubt as to their occurrence. Mucor V. is
the only form whose ( — ) race was found to give a sexual reaction
with A. spinosa.

The sexual reaction of A. spinosa wTith the ( + ) race of Mucor
V. is entirely different from what has just been described. At
the line of contact between the two mycelia, the filaments are
much branched and imperfect hybridizations are abundant.
Figs. 1 8 and 19 show simplifications of this condition from living
material. Fig. 17 shows gametes cut off from A. spinosa,
while Figs. 14-16 show such gametes transformed into dark
thick- walled azygospores. Fig. 15, which shows the progamete
of Mucor V. ( + ) rounded off and in bare contact with the
azygospore of A. spinosa, is a typical condition.

The (+) race of Absidia car idea (A ( + ), Figs. 8-10) shows
a less active sexual reaction with A. spinosa but similar to that
shown by Mucor V. ( + ), and the same is true of A. cylindrospora
and Mucor N. The ( + ) strains of all four dioecious species
mentioned are capable of stimulating the production of small
male gametes from filaments of A. spinosa and their transfor-
mance into dark thick-walled azygospores. Unfortunately
it has not yet been possible to bring these azygospores to ger-
mination. It would be interesting to discover if their germina-
tions would give rise to unisexual male mycelia.

Judging from the reactions just described between Absidia

SEXUAL REACTIONS BETWEEN MUCORS. 95

spinosa and the sexual races of Mucor V., one would seem
justified in considering the vegetatively more vigorous ( + )
race as female, and the less vigorous ( — ) race as male. The
terms male and female therefore may be used hereafter in dis-
tinguishing the sexual races of dioecious Mucors.

REACTIONS BETWEEN DIOECIOUS SPECIES AND ZYGORHYNCHUS.

Zygorhynchus Moelleri shows no sexual reaction with the ( — )
race of Mucor V. but contact with its ( + ) race causes the pro-
duction of delicate, contorted, warty filaments which take part
in imperfect hybridizations as shown in Fig. 12. The gametes
formed by Z. Moelleri in contact with Mucor V. ( + ), fre-
quently are transformed into azygospores shown in Fig. n. Z.
Vuillemini apparently is identical with Z. Moelleri, at least so
far as its sexual reactions with Mucor V. are concerned.

In contrast with these two Zygorhynchus forms just discussed
which show only a ( — ) or male tendency in contact with Mucor
V., Z. heterogamus gives evidence of a female tendency. Its
reactions with the ( — ) race of Mucor V. are shown in Figs. 25-28.
It does not seem to react with ( + ) races and even with the
( — ) race it is not easy to induce reactions and those that result
do not lead to azygospore formation. If Mucor V. ( — ) in
Fig. 25 is actually showing a sexual reaction with the male gamete
of Z. heterogamus as might appear from its position on the termi-
nal filament of the latter, the evidence would be in conflict with
that to be drawn from the reactions with A. spinosa, i. e., that
the ( — ) race is male and the ( + ) race female. It will not be
possible to attach much significance to the individual reactions
in Zygorhynchus, however, since so far no form of this genus
has been induced to react with more than a single sexual strain
of the test dioecious species.

Z. Vuillemini var. agamus is a zygosporeless strain which Xamy-
slowski (10) separated from a culture of Z. Vuillemini. It was
found impossible to induce zygospore formation from cultures
of this species supplied by the Centralstelle. In contact with
the ( + ) race of Mucor V. (Figs. 30-32) however, it reacted
like the sexually active strain of Z. Vuillemini and even pro-
duced azygospores (Fig. 32). This form therefore has a male
tendency.

96 ALBERT FRANCIS BLAKESLEE.

Judged by their reaction with the (+) and ( — ) strains of
Mucor V., the forms Z. Moelleri, Z. Vuillemini and Z. Vuillemini
agamus have a male tendency, while Z. heterogamus has a female
tendency. It is in harmony with this classification that Z.
Vuillemini agamus produces azygospores in contact with Z.
heterogamus (Fig. 29).

SUMMARY

It has been shown that sexual reactions called "imperfect
hybridization " may occur when hermaphroditic species of Mucors
are grown in contact with the sexual races of dioecious forms.
Certain of these hermaphrodites are heterogamic, showing a
constant difference in the size of their gametes. It is assumed
that the larger gamete is female and the smaller one male. The
race of dioecious Mucors provisionally called (+) shows a sexual
reaction with the smaller or male gamete while the ( — ) or
vegetatively less vigorous race shows a reaction with the larger
or female gamete. It. is concluded therefore that the ( + ) race
of dioecious Mocurs is female and the ( — ) race, male.

BIBLIOGRAPHY.
Blakeslee, A. F.

'043 Zygospore Formation a Sexual Process. Science, N.S., 19: 864-866.
'O4b Sexual Reproduction in the Mucorineae. Proc. Am. Acad., 40: 205-319.
'o6a Zygospore Germinations in the Mucorineae. Annales Mycologici, 4, 1-28.
'o6b Differentiation of Sex in Thallus, Gametophyte and Sporophyte. Bot.

Gaz., 42: 161-178.

'09 Papers on Mucors. Bot. Gaz. 47: 418-423.
'133 Conjugation in the Heterogamic Genus Zygorhynchus. Mycologisches

Centralblatt, 2: 241-244.
'i3b A Possible Means of Identifying the Sex of (+) and ( -) Races in the

Mucors. Science, N.S., 37: 880-881.
Hagem, Oscar.

'08 Untersuchungen iiber Norwegische Mucorineen. Vidensk.-Selsk. Skrif-

ter. I., Math.-Nat. Kl., pp. 50.
Lendner, Alf.

'08 Les Mucorinees de la Suisse. Materiaux pour la flore cryptogamique Suisse,

Vol. III., Fasc. I., pp. 180.
Moreau, F.

'12 Sur la reproduction sexuee de Zygorhynchus Moelleri. C. R. Soc. Biol.,

73. P- 14-
'13 Recherches sur la reproduction des Mucorinees et quelques autres Thallo-

phytes. Poitiers, pp. 136.
Namyslowski, B.
"10 Studien iiber Mucorineen. Bull. Acad. Cracovie, 477-520.

98 ALBERT FRANCIS BLAKESLEE.

EXPLANATION OF PLATES.

All figures were outlined with the aid of a camera lucida. Those in outline were
taken from stained and mounted material. Those stippled show development in
living material from moist chamber cultures. Figs. 8-12, 18-24 and 31-32 were
viewed with a low power objective and were magnified by the use of increased
tube length or higher eye-pieces. Other figures represent high power drawings.

PLATE I.

FIGS. 1-7. Absidia spinosa. Consecutive stages in zygospore development
from living material. Only those outgrowths in median focus are represented.
Drawings were made at following hours: Fig. i at 4.15 P.M., Dec. 4; Fig. 2 at
4.52 P.M.; Fig. 3 at 5.50 P.M.; Fig. 4 at 7.10 P.M.; Fig. 5 at 12.30 A.M., Dec. 5;
Fig. 6 at 2.15 A.M.; Fig. 7 at 3.00 P.M. The dotted circle in Fig. 6 represents
the outline of a drop of fluid around the young zygote.

FIGS. 8-10. "Imperfect hybridization" between Absidia spinosa (8) and
the (+) race of Absidia ccerulea (A(+)). Fig. 8 on right, typical conjugation of
Absidia spinosa; on left, "imperfect hybridization." Fig. 9 drawn at 8.30 A.M.,
Sept. 23. Fig. 10 drawn at 4.00 P.M., Sept. 24. Note azygospore formed from
the male gamete of the $ hypha.

FIGS, it AND 12. "Imperfect hybridization" between Zygorhynchus Moelleri
( 8 ) and Mucor V. (+). Fig. n, male azygospore from 8 hypha. Fig. 12, three
"imperfect hybridizations."

FIG. 13. Azygospores formed from male and female gametes of Zygorhynchus.

FIGS. 14-17. "Imperfect hybridizations" between Mucor V. (+) and A.
spinosa ( § )• Figs. 14-16, azygospores developed from male gametes of the Q .
Fig. 17, gametes on the $ .

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE I

ALBERT FRANCIS BLAKESLEE.

IOO ALBERT FRANCIS BLAKESLEE.

PLATE II.

FIGS. 18 AND 19. "Imperfect hybridizations" between Mucor V. (+) and
A. spinosa ( U ) drawn from living material.

FIGS. 20-24. "Imperfect hybridizations" between Mucor V. (— ) and A.
spinosa ( S ). Fig. 20 drawn at 7.30 A.M. Fig. 21, the same at 3.30 P.M.,
Sept. 12. Fig. 23 drawn P.M., Sept. 12. Fig. 24, the same, A.M., Sept. 13.

BIOLOGICAL BULLETIN, VOL. XXIX

•n

PLATE II.

V

ALBERT FRANCIS BLAKE8LEE.

102 ALBERT FRANCIS BLAKESLEE.

PLATE III.

FIGS. 25-28. "Imperfect hybridizations" between Mucor V. (— ) and Zygo-
rhynchus heterogamus ( § ). Fig. 25 from living material. Figs. 26-28, gamete
formation.

FIG. 29. "Imperfect hybridization" between Zygorhynchus Vuillemini var.
agamus and Zygorhynchus helerogamus showing formation of azygospore from
gamete of former species.

FIGS. 30-32. "Imperfect hybridizations" between Mucor V. (+) and Zygo-
rhynchus Vuillemini var. agamus. Fig. 30, progamete formation. Fig. 31 drawn
at 9.55 P.M., Dec. 20. Fig. 32 drawn at 10.45 A.M., Dec. 21.

FIG. 33. Apposed azygospores of Absidia spinosa developed from male and
female gametes, m and / apparently represent respectively male and female
progametes which have been pulled apart in mounting.

BIOLOGICAL BULLETIN, VOt. XXIX.

PLATE in.

ALBERT FRANCIS BLAKE8LEE.

REVERSIBLE ACTIVATION AND INCOMPLETE MEM-
BRANE FORMATION OF THE UNFERTILIZED
EGGS OF THE SEA URCHIN.1

JACQUES LOEB.

I. The writer showed in 1913 that the artificial activation of the
egg can be reversed under certain conditions.2 When the egg
of Arbacia is treated with a base in the proper concentration
and for the proper length of time the effect is somewhat similar
to that which follows the treatment of the egg with a fatty acid.
The egg is induced to develop, it may segment (as a rule abnor-
mally) once or twice, and if nothing else happens it will perish
rather rapidly (unless the developmental processes are prevented).
If the eggs after the treatment with the acid or with the base
receive a treatment with a hypertonic solution of the proper
concentration and for the proper time they may develop into
larvae.

The writer found that if the eggs are treated with alkali in the
proper way to induce development and if they are immediately
afterwards put for several hours into a solution which prevents
their development (sea-water with chloral hydrate or NaCN)
the eggs when taken out behave as if nothing had been done to
them. They neither segment nor do they disintegrate and they
can again be induced to develop by fertilization (or probably
with the methods of artificial parthenogenesis though the writer
has not yet tried this). The activating effect of the alkali is
therefore reversible.

This reversibility is, however, only possible if the eggs have
not been exposed to the alkali too long. After too long an ex-
posure the effect of the alkali is no longer reversible by a tem-
porary suppression of the developmental processes. A second
condition for the reversibility is that the eggs are immediately
transferred from the alkaline solution into the chloral hydrate

1 From the Rockefeller Institute for Medical Research, New York.
2Loeb, Science, N. S.f XXXVIII., 749. iQU; Arch. f. Entwcklngsmech.,
XXXVIII., 277, 1914.

103

IO4 JACQUES LOEB.

or NaCN solution. If they are first transferred into normal
sea-water and then after fifteen or thirty minutes later are
transferred to the cyanide solution nothing may happen to them
as long as they remain in the cyanide solution (unless the HCN
evaporates) but they will disintegrate when put back into normal
sea-water. From this we must conclude that in the sea-water
the processes started in the alkaline solution will continue and
the egg behaves as if it had been overexposed to the hyperal-
kaline solution.

Somewhat similar experiments were made with the eggs of
Arbacia which had been treated with butyric acid. When such
eggs were put over night into the cyanide-sea-water immediately
after the artificial membrane formation they did not as a rule
disintegrate when put back into normal sea-water but could be
fertilized afterwards. The writer is, however, not certain that
the phenomenon of reversibility is as constant here as in the
case of the alkali treatment.

The writer pointed out that the explanation for this phenome-
non could probably be found if we compare the behavior of the
eggs of Arbacia with that of eggs of purpuratus under similar con-
ditions. If we cause the production of a butyric acid membrane
in the egg of purpuratus the activation of the egg is usually irrever-
sible. In the egg of purpuratus the membrane which is formed
under the butyric acid treatment is very tough and is separated by
a wide area from the protoplasm, while in the egg of Arbacia the
membrane consists often only of a fine gelatinous film which lies
tightly around the egg. The writer concluded from this that there
may be a quantitative if not a qualitative difference in the
amount of change produced by the butyric acid treatment in
the cortical layer of the two kinds of eggs; in the egg of Arbacia
where this change is quantitatively smaller the activation is
reversible, while in the egg of purpuratus where the change is
quantitatively larger (and possibly also qualitatively different)
it is irreversible (Loeb, "Artificial Parthenogenesis and Fertiliza-
tion," Chicago, 1913, p. 286).

2. This idea is further supported by the curious phenomena
of reversibility in the eggs of Strongylocentrotus purpuratus.
When one tries to induce artificial parthenogenesis in the eggs

ACTIVATION OF EGGS OF SEA URCHIN. 1 05

of purpuratus by a treatment with a hypertonic solution (without
artificial membrane formation) it is found that this is possible
only with the eggs of certain females. Such eggs develop into
plutei. In the eggs of other females a small percentage of eggs
will begin to segment and they may go to the 2, 4, 8, or
i6-cell stage or still a little further but then stop developing.
Such eggs will go into the resting stage again and are normal to
all external appearances. If fertilized with sperm a number of
hours or a day later the individual blastomeres will each form a
special fertilization membrane and develop into small but appa-
rently, perfect larvae.1 This experiment shows that the activating
effect which the hypertonic solution had was reversed, since
all these eggs were at first in the state of active development.2

3. The writer wishes to report in this paper the fact that in
the egg of purpuratus under definite conditions a membrane
formation can be produced by butyric acid which leads neither
to any development if followed by the usual treatment with a
hypertonic solution nor to a rapid disintegration when not fol-
lowed by any "corrective" treatment. Since in all previous
observations on the effects of membrane formation by butyric
acid the reverse was found it seemed of theoretical importance
to study this exception.

The main condition for the experiment is that the eggs are
put after the treatment with butyric acid into a w/2 solution of
NaCl + KC1 + CaCl2 instead of into sea- water; the various salts
are always used in the proportion in which they are contained
in sea-water. The eggs were first washed three times in a mix-
ture of m/2 NaCl + KC1 + CaCl2 and were then put for from
one and a half to two and a half minutes into a mixture of 50 c.c.
m/2 NaCl + KC1 + CaCl2 + I c.c. Nfio butyric acid; from
here they were transferred into a neutral or a slightly alkaline
solution of m/2 NaCl + KC1 + CaCl2. In this solution all

1 Loeb, Arch. f. Entwcklngsmech., XXIII., 479, 1907; "Artificial Partheno-
genesis and Fertilization," Chicago, 1913, p. 237.

2 Only the activating effect of the hypertonic solution is reversible, the second,
corrective effect, which the hypertonic solution imparts to the egg was not reversed
since these blastomeres develop also if only the artificial membrane formation is
induced in them by butyric acid, while without the previous treatment with a
hypertonic solution such eggs would soon perish. Loeb, Jour. Exper. Zoo/., XV.,
201, 1913; "Artificial Parthenogenesis and Fertilization," p. 238.

io6

JACQUES LOEB.

the eggs form a membrane which is very thin and which is at
first at a great distance from the egg. Figs, i and 2 give an idea
of the first appearance of this membrane; the membrane in
Fig. i was formed in a neutral, that of Fig. 2 in a faintly alkaline
solution. Later on, the membrane collapses and is found lying
rather close to the egg. Such eggs can be readily fertilized with
sperm in spite of the existence of the membrane. The latter
is either naturally permeable for sperm or it is easily torn and
thus allows the sperm to reach the egg or it is not entirely con-
tinuous around the eggs.

FIG. i.

FIG. 2.

Eggs were washed three times in a mixture of m/2 NaCl +
CaCl2 + KC1 and then put for from one and a half to two
and a half minutes into 50 c.c. m/2 NaCl + KC1 + CaCl2 + i.o
c.c. N/io butyric acid. From this solution they were transferred :

(a) Into normal sea-water.1

(6) Into a neutral mixture of m/2 NaCl + KC1 + CaCl2.

Lot a formed normal fertilization membranes and disinte-
grated in a few hours. The eggs of lot b formed the fine fertili-
zation membrane which was at first very distant from the egg.
Some of these eggs were transferred into normal sea-water after
one hour. These eggs did not form a better membrane but per-
ished in the same way as the eggs of lot a.

A second portion of lot b was transferred into normal sea-
water after having been for seven hours in the mixture of m/2

1 Such eggs have at first a tendency to stick to the glass when transferred to
sea-water. It is necessary to keep them in slight agitation for some time until they
have lost their tendency to stick.

ACTIVATION OF EGGS OF SEA URCHIN. IOJ

NaCl -f- KC1 + CaCl2. These eggs did not disintegrate during
the following night although they did not appear quite normal the
next day. The rest of the eggs remained in the m/2 NaCl -f
KC1 + CaCl2 solution for twenty-four hours. When these
eggs were put into sea-water they did not disintegrate. To some
of these eggs sperm was added the next morning and about
20 per cent, developed.

This experiment indicates that the nature of the membrane
and the cortical changes which activate the egg of purpuratus
are to a large extent determined by the solution into which the
eggs are put after the butyric acid treatment. If the eggs are
put directly into sea-water after the butyric acid treatment
they form the typical membrane and rapidly perish if nothing
else happens to them. The same may be the case if they are
transferred from the butyric acid solution into a neutral solution
of m/2 NaCl + KC1 -f- CaCl2 for a short time (less than one
hour) and if they are afterwards transferred into normal sea-
water. But if they remain in the NaCl + KC1 + CaCl2
solution for a number of hours they will as a rule not disintegrate
if put back into sea-water.

A second experiment may be quoted. Eggs (that had been
washed three times in a neutral m/2 NaCl + KC1 + CaCl2
solution) were put for one and a half to two and a half minutes
at a temperature of about 14° into 50 c.c. of the same solution -f-
I c.c. N/io butyric acid. From here the majority of the eggs
were transferred into a neutral solution of m/2 NaCl + KC1
CaCl2 and the rest into sea-water. The latter formed normal
fertilization membranes and disintegrated the same day. Those
transferred into the neutral NaCl + KC1 + CaCl2 solution
had all formed membranes which were at first separated from
the egg by a great distance. They were kept in this sol-
ution over night, and the next morning they were transferred
into normal sea-water. All looked normal except that the
membrane now formed a fine veil lying closely to the egg.
Some of these eggs were fertilized with sperm. They all seg-
mented and developed into swimming larvae. The others were
kept in sea-water without sperm to find out whether they would
now disintegrate or whether the effect of the butyric acid treat-

108 JACQUES LOEB.

ment had worn off. This was the case. The eggs no longer
disintegrated in sea-water. A fraction of these eggs was fer-
tilized later and segmented.

When the eggs are transferred into normal sea-water imme-
diately after the butyric acid treatment they perish rapidly if
no second treatment is given to them. If they are treated with
hypertonic sea-water (50 c.c. sea-water + 8 c.c. 2 1/2 m NaCl
or NaCl + KC1 + CaCl2) for from thirty to sixty minutes they
will develop into larvae. If they are treated longer they will
begin to develop but form very abnormal larvae.

If the eggs, however, after the butyric acid treatment are not
put into sea-water but into a mixture of m/2 NaCl + KC1
+ CaCl2 in which they form the thin membrane and if they
are kept here until they no longer disintegrate in normal sea-
water, they cannot be induced to develop when treated for from
thirty to sixty minutes with a hypertonic solution. Such eggs
have, therefore, in spite of their membrane formation gone back
into the condition of a resting egg.

4. It was obvious from these experiments that the constit-
uents of the sea-water determine the nature of the membrane
after butyric acid treatment. The writer expected to find that
Ca was an absolute prerequisite for the membrane formation
but this was not the case, although the presence of CaCl2 im-
proves the character of the membrane. Eggs that had been
taken out of a m/2 NaCl solution were treated with 50 c.c. m/2
NaCl + i c.c. N/io butyric acid for one and a half to two and a
half minutes. They were then transferred into the following
three sets of solutions:

(i). NaCl + KC1,
(2). NaCl + CaCl2,
(3). NaCl + KC1 + CaClo,

which were made in duplicates, the one set remaining neutral,
the second set faintly alkaline (0.5 c.c. N/ioo NaOH were added
to 100 c.c. of the solution). The result was not very different
in the neutral and alkaline solutions. Many but not all the
eggs put into NaCl + KC1 formed very thin membranes often
only in part of the egg. The membrane formation does there-
fore not require the addition of CaCl2, but the eggs transferred

ACTIVATION OF EGGS OF SEA URCHIN.

into NaCl + CaCl2 or into NaCl + KC1 + CaCl2 had formed
a better and apparently stronger membrane than those put into
NaCl + KC1.

In another experiment it was found that in a mixture of NaCl
+ MgCl2 in the proportions in which these two salts occur in
the sea-water a membrane formation was also possible, but the
membrane had the same fine and thin character. In all these
cases the eggs could be fertilized with sperm in spite of the mem-
branes. The writer is under the impression that no second mem-
brane was formed upon the addition of sperm and only the hya-
line membrane developed later. That the eggs after this
inefficient membrane formation were not in the same condition
as eggs which are put into normal sea-water after artificial mem-
brane formation was shown by the fact that when put into a
hypertonic solution for from thirty to sixty minutes not a single
egg could be induced to segment.

The fact that the membrane always collapsed after some
time seems to indicate that it soon became permeable for the
colloids liquefied at the surface of the egg in the process of mem-
brane formation.

5. The explanation of the phenomenon of ineffective mem-
brane formation will probably be the same as that for the cases
of reversible activation of the egg. It is an established fact
that the rate of oxidations in the egg of purpuratus is raised about
600 per cent, by the act of fertilization. This rise is entirely
due to the alteration of the surface layer of the egg which results
in the membrane formation since artificial membrane formation
alone has the same effect on oxidation. Even if the egg is
killed in the process of membrane formation, e. g., if we treat the
unfertilized egg with saponin and leave it in this solution, the
rate of oxidation is raised to the same amount.1 The amount
to which the artificial membrane formation raises the rate of
oxidations in a sea-urchin egg seems to be a constant for each
species and not to depend upon the nature of the agencies em-
ployed for this purpose.

When we treat the unfertilized eggs of purpuratus with a hy-
pertonic solution (without inducing a membrane formation) the

1 Loeb and Wasteneys, Jour. Biol. Chem., XIV., 479, 1913-

110 JACQUES LOEB.

eggs of some females will develop into larvae, while the eggs of
others will not. The writer assumed that the hypertonic solution
if applied to the unfertilized egg has two effects, one consisting
in the alteration of the surface layer (comparable to the effect
of butyric acid) and secondly the corrective effect. Wasteneys
and the writer1 published a series of measurements on the in-
fluence of a hypertonic solution on the rate of oxidations in
unfertilized eggs of purpuratus, and found that the effect varied
enormously with the eggs of various females. The eggs were
in the hypertonic solution for one hour and the rate of increase
in oxidation varied for the eggs of various females between 40
and 400 per cent. As we stated above, the eggs of only a limited
number of purpuratus can be induced to develop into larvae
by a mere treatment with a hypertonic solution and it is probable
that the eggs in which the oxidations were raised efficiently were
the ones that could be induced to develop into larvae in this way,
while those in which the rate of oxidations was not raised to the
same height did not segment. We also noticed that in the latter
type of eggs the rate of oxidations diminishes after some time.
It is possible that the reversion of development in such cases is
due to a decline in the rate of oxidations below the height re-
quired for development.

The same possibility holds for the lack of development of the
eggs after artificial membrane formation through butyric acid
when they are afterwards put into a solution of NaCl + KC1
+ CaCl2 (instead of into sea- water). The fact that no correct
membrane is formed and that the eggs neither develop nor dis-
integrate, and behave towards a treatment with the hypertonic
solution like eggs without a membrane, arouses the suspicion
that in this case the butyric acid treatment did not lead to the
proper increase in the rate of oxidations. It is intended to in-
vestigate this possibility next summer, since it may also furnish
the explanation of the phenomena of reversibility in development.

1 Loeb and Wasteneys, Jour. Biol. Chem., XIV., 474, 1913.

REVERSIBILITY OF THE REACTIONS OF PLANARIA
DOROTOCEPHALA TO A CURRENT OF WATER.

GEORGE DELWIN ALLEN,

DEPARTMENT OF ANIMAL BIOLOGY, UNIVERSITY OF MINNESOTA.

CONTENTS.

I. Introduction in

II. Material and Methods 112

III. Difference of Reaction to Currents of Different Velocities 115

IV. Reversal of Reaction Induced by Chemical Changes 118

V. Reversal of Reaction Induced by Changes of Temperature 121

VI. Variations in Rheotropic Reactions 122

VII. Rheotropic Reactions in Nature 123

VIII. Summary 124

IX. Bibliography 125

I. INTRODUCTION.

A reaction of planarians to currents of water has been de-
scribed by Pearl ('03) in his careful study of the general features
of the behavior of planarians. He has described the reaction
and his method of obtaining it as follows: "In the course of
the experiments to localize chemical stimuli by the capillary
tube method, it was discovered that by means of a tube with
a relatively large opening (from .25 to .50 mm. in diameter)
and letting the ordinary tap-water in which the animals were
flow out of it by its own weight, a current of just the right in-
tensity to cause a positive reaction could be produced. The
animals would turn very sharply toward the source of such a
current, the reaction being evidently the same as that to other
weak stimuli (chemical and mechanical). This reaction is
localized in the same way as the usual positive reaction. It
is given only when the current is directed against the head or
anterior part of the body" (p. 698). He states that earlier in
his work a large number of experiments were performed with
various devices to determine whether these animals would show
such a reaction, but without success. Streams of water from a

in

112 GEORGE DELWIN ALLEN.

pipette and similar devices caused only a stopping, longitudinal
contraction, and gripping of the bottom without any turning
either toward or away from the source of the stimulus.

That the rheotropic reactions of planarians were found so
difficult to demonstrate, it seems probable, must have been
due to the experimental methods employed. Using the methods
described below, it has been found that rheotropic reactions
of these animals can be demonstrated very easily, not only in a
"current of just the right intensity" but in currents of a large
range of intensities. Under the conditions of these methods a
worm is entirely surrounded by the flowing water on all surfaces
except the ventral surface which is attached to the substratum,
and the conditions of stimulation are more typically rheotropic
than when the stream of water is directed as a small jet against
a localized part of the body. A negative reaction, i. e., a turning
away from the side stimulated, was not described by Pearl but
Dr. C. M. Child, who suggested this study, has observed a
negative reaction as well as a positive reaction in currents of
water in his laboratory stocks of worms used in studies in re-
generation. It has been found that these reactions are rever-
sible experimentally. The study of their reversibility which is
reported in the present paper was preliminary to a more detailed
study of the rheotropism of these animals which is in progress
at the present time.

The work reported in the present paper was done some time
ago in the zoological laboratory of the University of Chicago.
My acknowledgments are due Dr. C. M. Child and Dr. V. E.
Shelford for helpful suggestions.

II. MATERIAL AND METHODS.

Planaria dorotocephala has been used exclusively for this
study. Specimens were collected from a spring-fed marsh at
the margin of the Fox River near Chicago, Illinois. These
animals are very easy to keep in the laboratory without special
care. Since this study was in the nature of a testing of the re-
action possibilities of the worms toward currents of water, rather
than an effort to determine the normal habits, no efforts
were made to duplicate the normal conditions of existence in

REACTIONS OF PLANARIA TO WATER. 113

nature. Specimens were kept in large numbers in glass and
galvanized iron containers which were emptied of water, rinsed
and filled with fresh tap-water from time to time without re-
moving the worms which cling to the surface of the vessel.
They were fed two or three times a week with fresh beef cut
into small bits and left in the dish for several hours or during
the night. These worms collect on fresh meat and secure blood
and juices but it is very improbable that they are able to make
use of the solid portions.

Rheotropic reactions were tested in a "circular current" as
follows: A considerable number of specimens, sometimes several
hundred, were placed in a circular shaped vessel such as a glass
battery jar or a kitchen pan and the water was stirred vigorously
around the dish several times with a stirring rod. Vigorous
stirring usually dislodged the most of the specimens which were
then swept into a bunch at the center of the current. If they
were not dislodged by the stirring, they were loosened by means
of more vigorous streams of water from a large bulbed pipette.
The movements of the animals on the bottom of the vessel only
were studied since those on the sides were in a different relation
to gravity which introduced a geotropic factor into the reaction.
Rheotropism has been observed, however, in worms gliding on
vertical surfaces.

It was found convenient to have all the specimens, or the
larger part of them, enter the experiment at the center of the
"circular current" since they were then all placed under similar
conditions. In whatever direction any worm started out from
the center, it would receive the current against the same side of
the body as all the other worms starting from the center. If
the current were stirred in the clockwise direction, for example,
all the worms starting from the center of the dish would receive
the current against the left side of the body. If they gave a
positive reaction, they turned toward the left side; that is,
toward the side stimulated, or up-stream, and if they gave a
negative reaction, they turned toward the right side; that is,
away from the side stimulated, or down-stream. When a large
number of worms were gliding on the bottom of the dish, if their
reactions were uniform, a very striking figure was produced.

114 GEORGE DELWIN ALLEN.

This is illustrated in Figs. I and 2. Fig. i is a photograph
of a lot of worms in a dish pan, reacting positively to a current
stirred in the clockwise direction. The conspicuous spiral
form of the figure is characteristic of the positive reaction. Fig.
2 is a photograph showing the characteristic negative reaction
near the center of the pan.

These peculiar and characteristic spiral figures in a "circular
current" called for a more careful examination of the physical
conditions in the experiment, which revealed the fact that the
worms were subjected to a system of spiral currents instead of a
circular current. If a drop of a water suspension of carmine
is placed on the bottom of a dish in which the water has been
stirred in this way, the carmine in the lower layers of water close
to the bottom will be seen to stream inward along a spiral course
toward the center. Fig. 3 is a photograph of a dish pan in which
the water was stirred in the clockwise direction and then drops
of carmine were placed on the bottom at eleven points around
its circumference. The carmine was dragged along by the
currents and left streaks on the bottom of the pan which show
as spiral lines in the photograph. The worms were subjected
to this spiral system of currents. Carmine in the upper layers
of water was swept about the dish in a fairly circular direction,
and the water which flowed along the spiral lines on the bottom
toward the center, rose, as it neared the center, and spread out-
ward above.

A comparison of Fig. 3 with Fig. I shows that the spiral lines
of the currents and the spiral lines of the positive reaction are
alike. This explains, therefore, the spiral character of the re-
action. The worms orient themselves to a spiral current. A
comparison of Fig. 2 with Fig. 3 shows that in the usual negative
reaction the direction taken is away from the side stimulated,
but rather diagonally than directly down-stream. In some cases,
however, when worms were distributed over the bottom and
were not dislodged by the stirring, they were observed to turn
inward and follow along the lines of the spiral current in as
precise a negative orientation as was often characteristic of the
positive reaction.

A mixture of definitely positive and definitely negative re-

REACTIONS OF PLANARIA TO WATER. 115

actions at the same time, that is, an intermingling of the two
systems of spirals, was not common. The reactions varied,
however, in uniformity at different times; that is, in the pre-
cision of orientation, and in some cases the worms scattered from
the center without forming any very definite spiral in either direc-
tion. Such "reactions" were designated "indefinite."

III. DIFFERENCE OF REACTION TO CURRENTS OF
DIFFERENT VELOCITIES.

In studying the reversal of rheotropic reactions induced by
changing the chemical and thermal conditions, it was found
that it is important to consider the velocity of the current. A
change in the velocity of the current, or, in other words, in the
strength of the rheotropic stimulus, may itself cause reversal of
the sign of the reaction.

Since a current in a small vessel diminishes rapidly in velocity
after the stirring is discontinued, the different worms in a
single trial are subject to different velocities of current, the
first worms to leave the center entering the strongest current
and later worms finding successively weaker current. It was
frequently observed that the positive reaction was given by
the earlier worms; that is, those in the strongest current, and
the negative reaction by the later worms; that is, those in the
weaker currrent. The proportion of individuals giving each
reaction varied in different observations while in some cases
no negative reaction was given and in other cases no positive
reaction was given.

A positive reaction in stronger current and a negative reaction
in weaker current can be seen at the same time in a broad-
bottomed pan where the negative spiral may begin about the
center before the last individuals belonging to the positive spiral
have reached the sides. This is possible because the current
is swifter about the outside where the positive reaction is still
being given than about the center where the negative reaction
is beginning. Fig. 2 shows such a combination of positive
spiral about the outside and negative spiral about the center.
In the case which is photographed there is considerable irregu-
larity in the positive spiral. In many cases the positive spiral
about the outside was as regular as that in Fig. I.

u6

GEORGE DELWIN ALLEX.

Table I. shows the total number of trials in which the different
types of reaction were given during all the experiments.

TABLE I.

SUMMARY OF ALL REACTIONS OBSERVED (EXCEPT THOSE GIVEN UPON INCREASING
THE VELOCITY OF THE CURRENT, FOR WHICH SEE TABLE II).

The sign + indicates a positive spiral, — , a negative spiral and ?, no dominant
spiral of either sign. The sign + — indicates positive reaction in stronger current
and negative reaction in weaker current; (+) — indicates only a few worms in
the positive spiral, and + ( — ) only a few worms in the negative spiral.

Total Number of Trials of

Worms in Normal Conditions

Total Number of Trials During

Reactions.

Preceding Experiments, and

Experiments, in all Conditions,

Totals.

in Normal " Controls" During

Excepting the " Controls."

Experiments.

+

324 lip

443

—

31

130

161

(+) -

45]

48)

93)

+ -

46 MI3

36 y 104

82 f 217

+ (-)

22 J

20 )

42]

p

32

29

61

Totals

500

382

882

In the observations described above, the same individuals
were not observed to give both reactions because under the con-
ditions of the tests the same individuals are not subjected to
both extremes of current velocity. The earlier worms receive
the stronger current and the later worms receive weaker current.
It was readily shown, however, that the same individuals which
responded negatively in a weaker current would respond posi-
tively in a stronger current. Upon second stirring of the water
over a negative spiral without dislodging the worms, it was gen-
erally found that the worms composing the negative spiral re-
versed their direction of movement from negative to positive,
so that the negative spiral became converted into a positive
spiral composed of the same individuals. Out of 100 cases in
which second stirring of the water over a negative spiral was
tried, 80 cases showed reversal to positive. In some cases all
the individuals gave this reversal while in other cases few in-
dividuals, that is, only those forming the outside of the spiral
gave the reversal. In some cases reversal was readily induced
by currents of moderate velocity while in other cases reversal
was given only in very strong currents.

REACTIONS OF PLANARIA TO WATER.

TABLE II.

117

REVERSAL OF REACTION INDUCED BY INCREASING THE VELOCITY OF THE CUR-
RENT. SUMMARY OF ALL REACTIONS IN EXPERIMENTS IN WHICH THE VELOCITY OF
THE CURRENT WAS INCREASED WHILE A REACTION WAS BEING GIVEN.

Velocity Increased During a Negative Reaction.

Complete
Reaction
Before In-
crease in
Velocity.

Reactions After Increase in Velocity.

Reversal to + .

More Precise — .

No Change — .

Worms Dis-
lodged.

Reaction Be-
came ?.

(+T-

21
18

5

I

5

2

I

I

+

28

3

2

+ (— )

13

Totals

80

6

10

3

I

Velocity Increased During a Positive Reaction.

Complete

Reactions After Increase in Velocity.

Before In-
crease in
Velocity.

More Precise -(-.

No Change + .

? _|_

8

i

4

Velocity

Increased During an Indefinite Reaction.

Complete

Reactions After Increase in Velocity.

Before In-
crease in

Velocity.

Reaction Be-
came + .

?

(+) ?
+ ?
+ (?)

4
4
15

I

No evidence has been obtained that a stronger rheotropic
stimulus will cause a negative reaction while a weaker stimulus
causes a positive reaction. No observations showed a sequence
of reactions during the diminution of the velocity of the current;
of negative in stronger current and positive in weaker current
(cf. Table I.). And second stirring of the current while the worms
were distributed over the bottom was never observed to reverse
the reaction from positive to negative. If the worms were giving
the positive reaction in weaker current they continued to give
the same reaction in stronger current, and usually with greater
precision of orientation (Table II.). When the reaction was

Il8 GEORGE DELWIN ALLEX.

"indefinite" in weaker current, the stronger current usually
called forth the positive reaction.

IV. REVERSAL OF REACTION* INDUCED BY CHEMICAL

CHANGES.

Reversal of reaction from positive to negative can usually be
induced by pouring off the water from the worms and replacing
it with fresh tap-water. Thus in 23 out of 24 observations
this treatment resulted in reversal to the negative reaction. Two
factors may be responsible for this reversal, change in temperature
and change in the chemical composition of the water. Each
of these factors was investigated by modifying the condition in
question while preserving other conditions unmodified. The
velocity of the current could not be preserved unmodified with
the methods employed, but this factor was controlled indirectly.

The effect of changing the chemical character of the water
was tested as follows: A lot of worms were tested in the water
in which they were living at the time. Then an equal volume
of fresh tap-water, or modified water, was brought to the same
temperature as the aquarium water, and the worms were tested
alternately in these two kinds of water.

A number of experiments were performed in which aquarium
water was replaced by fresh tap-water of the same temperature.
Since the rheotropic reaction in many cases may be positive in
stronger current and negative in weaker current, it must be
made certain that the reversal of reaction after a chemical change
in the environment is a reversal of reaction in the same velocity
of current. Although it seemed fairly certain to the operator,
during the earlier experiments, that there was such a reversal in
the same velocities of current, the later experiments were per-
formed with a "control" to make them more conclusive. In
these experiments two dishes of worms living under the same
conditions, frequently in the same aquarium tank, were tected
side by side. One lot of worms was tested in waters of different
composition while the other lot was kept as the "control" and
was tested always in the original aquarium water. The ex-
perimental and control dishes were placed side by side, and the
control worms were tested at each trial simultaneously with

REACTION'S OF PLAXARIA TO WATER.

119

the experimental worms so that the conditions as regards ve-
locity of current were practically the same in the two dishes.
Progressive diminutions in velocity were practically the same
in the two dishes. Differences of reaction given by the worms
in the two dishes at the same time of observation must then be
traced to other factors than the velocity of the current.

The following experiment will illustrate the method of the
controlled experiments and the characteristic reversal of reaction
induced by change from aquarium water to fresh water. Many
hundreds of worms of miscellaneous sizes were tested in two
large galvanized iron tanks. A current was produced by shaking

TABLE III.

EXPERIMENT SHOWING REVERSAL OF REACTION FROM POSITIVE TO NEGATIVE
INDUCED BY CHANGE FROM AQUARIUM TO FRESH WATER AT THE

SAME TEMPERATURE.

Trial.

Experiment.

Control.

Water.

Reactions.

Reactions.

Aquarium + good general response

Fresh

4

5

Aquarium

Fresh
Aquarium

+ on second stirring, even in
weak current

— even in fairly strong current.

— on second stirring, even in

fairly strong current, a
few + in strongest cur-
rent

— slight; worms slow in dis-

tributing on the bottom.
+ on second stirring, even in
only moderately strong
current.

— even in fairly strong current.

— on second stirring, no posi-

tive reaction.

movement very slow, cur-
rent reduced before any
reaction was given.
+ on second stirring only
moderately strong cur-
rent.

+ in strong current
— in weak current

+ on second stirring, uni-
versal.

+ even in very reduced cur-
rent.

+ even in very weak current.

+ even in very weak current.

+ even in very weak current.

the tanks in the hands, which swept the worms to the center. In
the experimental tank aquarium water was used in trials i,
3 and 5 ; and this was replaced by fresh water in the alternate

I2O GEORGE DELWIN ALLEN.

trials 2 and 4. In the control tank the worms remained in the
aquarium water in all the trials. The temperature was 21° C.
The reactions in each trial are described in Table III.

From the table it can be seen that placing the worms in fresh
water reversed the reaction from positive to negative while
returning them to their former aquarium water brought back
the former positive reaction. These reactions were reversed
back and forth by changing the water. The reversal of reaction
cannot be attributed to a difference in the velocity of the current.
The negative reactions in the fresh water were given in fairly
strong current while the positive reactions in aquarium water
were given in very weak as well as stronger current; and in the
control, in which the velocity of the current was approximately
the same at each moment as that in the experiment, a positive
reaction was given at the same time that a negative reaction was
observed in the experiment. In the two tests in this experiment
with fresh water (trials 2 and 4 ), therefore, a change to fresh
water induced a reversal of reaction from positive to negative
while a return to the normal environment brought back the
positive reaction. In 40 such tests in experiments at different
times, a reversal from positive to negative was given in 21 tests
and a reversal from negative to positive in 5 tests. In 7 tests
a positive reaction remained unchanged while in 5 tests a nega-
tive reaction remained unchanged. In 2 tests the reaction was
"indefinite."

Fresh water differs from aquarium water in oxygen content
and in the absence of the metabolic products of the worms and
the decomposition products that accumulate in the aquarium.
The aquarium water in experiments at different times varied
greatly in composition, while the fresh drawn tap-water probably
varied little. It would seem that fresh water would make a
more suitable environment than stale aquarium water where
many worms have lived together without green vegetation.
In the experiment described above the aquarium water was
quite foul from the decomposition products of juices of meat
that had stood in the aquarium for nearly 24 hours. When
modified aquarium water was used instead of fresh water, re-
versal of the same sort was induced. In one experiment part of

REACTIONS OF PLANARIA TO WATER. 121

the aquarium water was boiled, and in the two tests made with
this boiled aquarium water reversal of reaction from positive to
negative was obtained. In another experiment cane sugar was
dissolved in part of the aquarium water to make a I per cent-
solution which induced reversal from positive to negative in the
6 tests made. Similarly a well-mixed solution of juices of raw
beef in aquarium water induced the same reversal in the 6 tests
made. So far as these experiments indicate, therefore, the re-
versal of the rheotropic reaction in the new environment seems
to be due more to the strangeness of the new conditions than to
their fitness or unfitness. In a total of 54 tests with modified
aquarium water and fresh water, reversal of reaction from posi-
tive to negative was induced in 35 tests, while in 5 tests the re-
action was already negative from other causes. These figures
may not necessarily represent the proportion of individual worm
reactions which can be reversed in this way, since the conditions
in different experiments were not entirely identical, but it may
be repeated that in each of these tests a large number of worms
were tested, in some cases many hundreds of individuals.

V. REVERSAL OF REACTION INDUCED BY CHANGES
OF TEMPERATURE.

The effect of changes of temperature upon the rheotropism
was tested by dividing the water in which worms were living
into two portions, preserving one portion at the temperature at
which the worms were found at the beginning of the experiment,
and cooling or warming the other portion to the desired tempera-
ture. The worms were then tested alternately in these two
portions of water. It was found that the rheotropism can be
reversed from positive to negative by lowering the temperature in
the same manner as by changing the composition of the water.

In 34 individual tests, lowering the temperature by 4°-io° C.
was found to cause a reversal in 17 tests from positive to negative,
and reversal from negative to positive in 2 tests. In 4 tests
this change in temperature made an indefinite reaction become
positive. In the remaining n tests, lowering the temperature
produced no observed change in reaction. This remained as
before the temperature change; that is, as follows: positive in 7

122 GEORGE DEL WIN ALLEN.

tests, negative in 3 tests, positive in stronger current and nega-
tive in weaker current in one test.

The characteristic rheotropic effect, therefore, of lowering
the temperature below that to which the worms are accustomed
is a reversal of reaction from positive to negative. Return to
the normal temperature brings back the positive reaction so
that the reactions can be reversed back and forth by alternately
lowering the temperature and raising it again to the normal.
The reversal from negative to positive on raising the temperature
seems to be due to the return to the temperature to which the
the worms are accustomed, rather than to the fact that the tem-
perature is raised, since in 13 tests of raising the temperature
above the normal by 4°-io° C. reversal from negative to posi-
tive was obtained in only one case. In 4 tests the reversal was
in the opposite direction ; that is, from positive to negative. In
5 tests the negative reaction remained unchanged and in 2 tests
the positive reaction remained unchanged, while in one test the
reaction became "indefinite."

VI. VARIATIONS IN RHEOTROPIC REACTIONS.

The general features of the rheotropic reactions of planarians
and their reversibility described above were observed in mass
tests of large numbers of individuals. These depend upon con-
siderable uniformity in the behavior of the different individuals
in the trial, much as do the behavior experiments with cultures
of protozoa. The uniformity of the reactions of the different
individuals in a single trial in the experiments with planarians
was generally very striking, at least with respect to the sign of
the reaction in the same velocity of current. The precision of
the reactions of different individuals varied greatly in different
trials, although the sign of the reaction was definite. When
the reaction is definitely either positive or negative, the char-
acteristic spiral figure is very striking, and when a spiral could
not be seen, the reaction was called indefinite. Only a small
number of trials, however, failed to show a general uniformity
in the sign of the reaction in the same velocity of current (Table I.).

The different tests of reversal in the same experiment gener-
ally gave fairly consistent results although with some variations

REACTIONS OF PLANARIA TO WATER. 123

among the different tests. But in some experiments the lot of
worms gave consistent results directly opposite to those generally
obtained. Thus in the experiments of lowering the temperature,
five of the seven tests in which the positive reaction remained
unchanged were the successive tests in a single experiment.
In this case the temperature at the beginning of the experiment,
and the amount of the depression of the temperature, were the
same as in other experiments in which reversal of reaction was
produced in the characteristic manner. Such differences in
behavior must be attributed to that complex of internal pro-
cesses which has been characterized the "physiological state"
of the organism.

VII. REACTIONS TO CURRENTS OF WATER IN NATURE.

That rheotropic reactions are normal to the life of planarians
in nature is shown by field observations. The specimens studied
in the laboratory were collected in a spring-fed marsh at the
edge of the Fox River near Chicago, called "Station 10 Cary
Spring" in Shelford's description of the animal communities of
the Chicago Region (Shelford, '13, pages 52, 93 and 118). The
marsh bottom is composed of black humus without stones under
which the worms could collect in the manner that Pearl found
characteristic of this species in the Huron River, but an abundance
of leaves serves the same purpose. The water is very shallow
and well shaded by marsh vegetation. There is a sluggish current
fed by springs at the upper edge of the marsh, where water cress
is found growing.

A small channel about five inches wide and fifteen feet long
was discovered at the edge of the marsh through which a moder-
ate current of water was flowing. A large number of worms were
found in this stream, some moving on the bottom, generally
up-stream, but most of them at rest on the under sides of leaves
or in sheltered places on the bottom. Removing the shading
vegetation for better observation stimulated many of the worms
at rest to become active, and half an hour later 60 worms were
counted going up-stream and 5 going down-stream and 200 were
counted at rest on the under side of leaves. At that time the
sun was shining somewhat from the side, and down-stream, but

124 GEORGE DELWIN ALLEN.

it was not far from vertical as it was 1 1 130 in the morning. At
12:20 only 4 worms could be found in sight, 3 going up-stream
and one going down-stream. At 1 :3O, when it was cloudy, 40
worms were found going up-stream and 4 going down-stream.
At 3:10 a piece of raw beef was placed in the upper part of the
stream and fifteen minutes later 176 worms were counted going
up-stream and 34 going down-stream. It has frequently been
noted that worms in a moderate stream of water will respond
in this way and collect upon a piece of fresh beef and this be-
havior has been put to practical advantage in collecting speci-
mens. Worms removed from this stream and tested in a cir-
cular pan by the method employed in the laboratory showed
both the positive and the negative reactions very clearly, the
positive in stronger current and the negative in weaker current,
as described above.

These observations show that planarians may react to currents
of water in nature. Pearl considered it "very improbable that
this reaction is of any importance in the normal activity of the
animal." This opinion may have been due to the difficulty which
he experienced in demonstrating the reaction experimentally
which might lead one to think that unusual conditions are neces-
sary for its production in nature; whereas it has been shown that
the reaction is given with readiness under experimental conditions
which are similar to those in a natural stream of water, and also
that it can be observed in nature. It would seem that a reaction
which is so characteristic of the animals in the laboratory would
play a part in their daily life in nature, when they live in an
environment which offers the appropriate stimuli for the reac-
tion. These animals seem to belong characteristically to stream
communities where these conditions are fulfilled.

VIII. SUMMARY.

1. Planaria dorotocephala show both positive and negative
reactions in a stream of water.

2. The sign of the reaction may differ, depending upon the
velocity of the current. The positive reaction is then given in
stronger current and the negative reaction in weaker current.

3. A negative reaction in weaker current can often be reversed
to the positive by increasing the velocity of the current.

REACTIONS OF PLANARIA TO WATER. 125

4. Reversal of reaction from positive to negative can be in-
duced by changing the composition of the water; and from nega-
tive to positive by return to the former conditions.

5. Similar reversals of reaction can be induced by sudden
changes in temperature.

6. Rheotropic reactions are given by Planaria dorotocephala
in nature.

7. Rheotropic reactions are very characteristic of these animals
in the laboratory, rather than unusual as has been supposed.

IX. BIBLIOGRAPHY.
Alice, W. C.

'12 An Experimental Analysis of the Relation between Physiological States

and Rheotaxis in Isopoda. Jour. Exp. Zool., Vol. 13, pp. 269-344.
Bardeen, C. R.

'01 The Function of the Brain in Planaria Maculata. Amer. Jour. Physiol.,

Vol. V., pp. 175-179.
Boring, E. G.

'12 Note on the Negative Reaction under Light Adaptation in the Planaria.

Jour. Animal Behav., Vol. 2, pp. 229-248.
Jennings, H. S.

'06 Behavior of the Lower Organisms, Columbia University Press, New York.
Mast, S. O.

'03 Reactions to Temperature Changes in Spirillum, Hydra, and Fresh-water

Planarians. Amer. Jour. Physiol., Vol. X., pp. 165-190.
Parker, G. H. and Burnett, F. L.

'oo The Reactions of Planarians, with and without Eyes, to Light. Amer.

Jour. Physiol., Vol. IV., pp. 373-385.
Pearl, Raymond.

'03 The Movements and Reactions of Fresh-water Planarians. A Study in

Animal Behavior. Quart. Jour. Micro. Sci., N.S., Vol. 46, pp. 509-714.
Shelford, Victor E.

'13 Animal Communities in Temperate America as Illustrated in the Chicago
Region. Bull. No. 5, Geographical Society of Chicago. The University
of Chicago Press, Chicago, 111.
Walter, H. E.

'07 The Reactions of Planarians to Light. Jour. Exp. Zool., Vol. V., pp. 35-162.
Woodworth, W. McM.

'97 Contributions to the Morphology of the Turbellaria. II. On Some
Turbellaria from Illinois. Bull. Mus. Comp. Zool. Harvard Coll., Vol.
XXXI., pp. 1-16.

126 GEORGE DELWIX ALLEN.

EXPLANATION OF PLATES.
PLATE I.

FIG. i. The positive spiral. Photograph of worms giving a positive reaction
to a current of water stirred in the clockwise direction.

FIG. 2. The negative spiral. Photograph of worms giving a negative reaction
to a current of water stirred in the clockwise direction. The worms near the center
are reacting negatively while those about the outer part of the pan, in the stronger
current, are reacting positively.

B>OLOGICAL BULLETIN, VOL. XXIX.

PLATE 1.

-

\

-v

N \
v

N

V' Ji

-^^VV'-M;

-^^m^lL J * ' -/;

• / ,/*•/'*' *&** J \r/ r i y /' ' '

*'*%fmi&tJf, '<

''"?/

VV- — - _ "X

r- v'^A

• x V >/ » k ' '

-i*' ' / JT ^ ^.^ \\jLf'

]\';r^.f^^-

i i ' / ' >• >^Tf~' ^-^^ ' '•

\i if ( '[,-: V^^^

\ ' • ' ,'. Jv\ ^--i?
\ »i.- • "v J^v*^ ^-

/

\

v\

FIG. i.

GEORGE DELWIN ALLEN

FIG. 2

128 GEORGE DELWIN ALLEN.

PLATE II.

FIG. 3. The system of spiral currents. Photograph of pan in which the water
has been stirred in the clockwise direction and then drops of carmine have been
placed on the bottom at various points about its circumference. The dark lines
are streaks of carmine produced by the currents and showing the direction of the
currents upon the bottom.

FIG. 4. A positive reaction showing individual variation in the precision of
orientation.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE II.

FIG. 3.

GEORGE DELWIN ALLEN

FIG. 4.

A CASE OF HERMAPHRODITISM IN SPELERPES

BISLINEATUS.

CATHARINE LINES CHAPIN, A.M.
DEMONSTRATOR IN ZOOLOGY, SMITH COLLEGE, NORTHAMPTON, MASS.

Hermaphroditism among Anura seems to be comparatively
common. Ecker and Wiedersheim's "Anatomic des Frosches,"
as revised by Gaupp, devotes several pages to the subject,
dividing it into various types, the most common being that in
which the glands are essentially male with female elements.

Miss King ('10) in her paper on anomalies in the genital
organs of toads, says: "Evidently hermaphroditism occurs
much less frequently among the Urodela than among the Anura,
as only two cases have as yet been reported for this group of
Amphibians. La Valette St. George ('95) has given a brief
description of a case of hermaphroditism in Triton tceniatus
and Knappe ('86) has noted the presence of a Bidder's Organ
in a young salamander; neither investigator gives any details
regarding the structure of the ovo-testes in these forms."

Knappe mentions Triton also in his investigations but seems
to have found a Bidder's Organ only in a two-year-old Sala-
mandra maculata. Concerning the frequency of occurrence of
hermaphroditic salamanders, La Valette St. George says: "Fur
die Urodelen liegen, soviel mir bekannt, noch keine Angaben
iiber Zwitterbildung vor.

"Spengel will solche niemals angetroffen haben, obgleich
er zahlreiche Salamander und Tritonen zerlegt und allein von
Triton cristatus iiber 100 mannliche Individuen untersucht hat."

While collecting data concerning the spermatogenesis and
age of attainment of sexual maturity of Spelerpes bislineatiis l
I found that a series of sections through the testes of a 46 mm.

1 Although this species has long been called bilineatus, Dr. Green in his origi-
nal description of it (Jour. Acad. Nat. Sc. Phila. Vol. I, Pt. II, 1818) gave it the
name of Salamandra bislineata. With the change in generic name the masculine
ending was adopted and the species is correctly Spelerpes bislineatus.

129

130

CATHARINE LINES CHAPIN.

larva, collected in September, 1913, contained several eggs.
The larva was nearly of adult proportions and presumably
would have undergone metamorphosis the following summer.
Fig. i, B, is from a drawing of the pair of gonads and kidneys,
made with an Abbe camera before imbedding. It reveals the
peculiarities of the glands when compared with Fig. I, A, which
is a drawing, at the same magnification, of a normal pair of

!o

FIG. i. A. Testes and kidneys of normal 48 mm. larva. B. Gonads and
kidneys of hermaphroditic 46 mm. larva. C. Ovaries and kidneys of normal 46
mm. larva. These gonads and kidneys are from larvae collected in September.
Ventral view, X 75. k = kidney, lo = largest ova, U = left testis, o = ova,
ov = ovary, rt = right testis, so = smaller ova, t = testis.

testes and kidneys of a 48 mm. larva. The original of Fig. I,
B, being merely a hastily made outline drawing and the testes
having been very small, I noticed the hermaphroditic condition
only on careful study of the sections.

A comparison of Fig. i, A and B, shows that, in the abnormal
specimen, neither testis is fully developed. The hermaphroditism

HERMAPHRODITISM IN SPELERPES BISLIXEATUS.

is of two types. Macroscopically, the anterior part of the left
gonad, which is much reduced in size, resembles the normal testis
in texture, though not in shape, while the posterior region is
distinctly like an ovary. The larger ova in this latter region,
indicated in outline in Fig. I, B, are six in number.

Fig. i, C, is a camera drawing of the ovaries and kidneys of
a normal 46 mm. larva, collected in September, showing the
arrangement of eggs, for comparison with'their arrangement
in the posterior region of the left gonad of the hermaphrodite.
In each normal ovary the larger ova lie approximately in two
rows, one lateral and one medial. A dozen or fifteen on each
side are appreciably larger than the others. These larger ones,
as may be deduced from a comparison with the ova in specimens
of adult female Spelerpes, collected in September, are of the
size and state of development which indicate that they would
normally have been deposited a year from the following spring
as the first brood of this individual, while the next smaller ones
would have been deposited two years from the following spring.

The larger ova in the left gonad of the hermaphrodite are
about the size of the smaller ova in the normal ovaries, though
their irregularity in shape, especially where large numbers of
eggs are packed together, makes exact measurements impossible.
The ova of both the normal specimen and the hermaphrodite
have their long axes parallel to the longitudinal axis of the body.
This diameter, as measured from sections, is .15 to .20 mm.
while the shorter diameter, at right angles to the long axis,
averages about .12 mm. The largest ova found in the herma-
phrodite are, in development, a year behind the largest ova in
a normal female having the same total length, collected at the
same time, and therefore, presumably, of the same age. Beside
these ova, there are oogonia in the left gonad of the herma-
phrodite, similar to those in the normal ovary which have not
yet begun to elaborate yolk (Fig. 2). The relation of ova to
follicle cells and to peritoneum in the hermaphrodite is like that
in the normal ovary.

Although smaller than the right testis, the anterior portion
of the left gonad has somewhat the structure of the normal male
gland. The elongated mass is composed of too few loulebs to

132

CATHARINE LINES CHAPIN.

have the usual arrangement about a collecting duct. The cells,
showing no signs of division to form cysts, which is the im-
mediate preparation for spermatogenesis, are not as far advanced
in their development as the germ cells of a normal male larva
several months younger than the hermaphroditic one. The left
gonad shows throughout a retarded development.

The right gonad, on the other hand, although smaller than
the normal testis of an individual of the same size, seems to

0

FIG. 2. Horizontal section through posterior region of left gonad of hermaphro-
dite, X 280. / = follicle cell, o = oogonium, ov = ovum, p = peritoneum.

be, in general, normal in structure and in its cellular development.
However, it shows another sort of hermaphroditism. Two ova
are, in this case, found in the otherwise apparently normal
testis, each one completely filling one lobule, which would nor-
mally contain a large number of male cells (see Fig. 3). These
ova are about two thirds the size of those found in the right
gonad. All the lobules are divided into cysts, with the excep-

HERMAPHRODITISM IN SPELERPES BISLIXKATUS.

133

tion, of course, of the two which contain a large ovum apiece.
Most of these cysts are composed of spermatogonia of the last
generation, having round or oval nuclei with the chromatin
arranged in an irregular network. Nucleoli appear in a few
cells of one section. The spermatogonia resemble Kingsbury's
Figs, i and 2 ('02). In a few cysts, all the cells are undergoing
mitosis. These do not resemble the maturation mitoses described
for Desmognathus, and are probably spermatogonial divisions.

P

FIG. 3. Horizontal section through right gonad of hermaphrodite, < 280.
/ = follicle cell, / = lobule, subdivided into six cysts of spermatogonia, o = ovum,
p = peritoneum, spg = spermatogonium, spin = spermatogonial mitosis.

In other cysts the nuclei are in the "contracted" condition,
described by Kingsbury (Fig. 18), occasionally occurring in
Desmognathus fusca at the beginning of the period of growth of
spermatogonia into spermatocytes and more commonly, in (hat
species, in the last generation of spermatocytes. Fig. 3 shows
an egg filling one lobule and surrounded by follicle cells. In
this case, as well as in the left gonad, the relation to follicle cells
and to peritoneum is normal. The adjacent lobules show sper-

134 CATHARINE LINES CHAPIN.

matogonia, grouped in cysts, surrounded by long follicular cells.
The cells of one cyst are undergoing spermatogonial mitosis.
A great deal of pigment from the peritoneum appears in this
section which was cut near the surface of the testis.

The germ cells of the right gonad seem to have reached a
normal degree of development. The individual, collected in
September, with a total length of 46 mm., would have undergone
metamorphosis during the following summer. It is probable
that the male Spelerpes attains sexual maturity in the fall after
metamorphosis. Then this individual : hould have reached
maturity in a little over a year. The spermatogonia are just
ready to commence the process of maturation. This process in
Desmognathus takes about a year, while Kingsbury's obser-
vation that, in Spelerpes, regeneration of the lobule begins
before the spermatozoa leave it, indicates that possibly the proc-
ess of spermatogenesis requires even more time in Spelerpes.
Then it may be supposed that these male germ cells of the right
gonad would have become ripe spermatozoa at the normal time.
The two ova in this gonad, however, are at least a year behind
the normal development. In the left gonad, in which there is a
greater abnormality, the development of both kinds of germ
cells is retarded.

BIBLIOGRAPHY.

Ecker, A., & Wiedersheim, R.

'96 Anatomic des Frosches, auf grund eigener Untersuchungen durchaus neu

bearbeitet von Dr. Ernst Gaupp. Dritte Abtheilung, pp. 347-355-
King, Helen Dean.

'10 Some Anomalies in the Genital Organs of Bufo lentigenosus and their
Probable Significance. Amer. Journ. of Anal., Vol. 10, No. i, January,
1910, pp. I59-I75-
Kingsbury, B. F.

'02 Spermatogenesis of Desmognathus fusca. Amer. Journ. of Anal., Vol. r.

No. 2, February, 1902, pp. 99-135-
Knappe, Emil.

'86 Das Bidder'sche Organ. Morph. Jahrb., Bd. XL, pp. 489-552.
St. George, La Valette.

'95 Zwitterbildung beim kleinen Wassermolch (Triton taeniatus). Arch.
Mikr. Anat., Bd. XLV., pp. 1-14-

NOTE ON THE NATURE AND SOURCE OF

"PURPLE X."1

ALVALYN E. WOODWARD.

In the summer of 1913, Dr. O. C. Glaser ('14) found (i) that
if a suspension of Arbacia sperm be boiled it turns purple, (2)
that this purple color disappears if the boiled suspension be allowed
to stand over night, and (3) that initiation of development in
Arbacia eggs, either by sperm or by egg secretion, can be in-
hibited by the addition to the eggs of this boiled suspension
so long as it is purple. This purple substance, he provisionally
designated as "purple x."

I. SOURCE OF "PURPLE X."

In 1914, while working with various inhibitors of fertilization
in Arbacia, I wished to use "purple x" and found that it did
not always appear when Arbacia sperm was boiled. This
naturally led to a search for the source of the substance. Sperm
was obtained from a large number of males. Each one was
rinsed in fresh water, and after the peristome was cut, placed
aboral side down in a clean dry watch-glass, until the seminal
fluid had been freely shed. A portion of the sperm from each
animal was then mixed with filtered sea-water and examined under
the microscope and a second, larger portion of each was boiled.
It was found that, as a rule, the suspensions which showed fewest
foreign cells — blood-cells or fragments of organs — were least
likely to form "purple x." Suspensions which appeared, under
the microscope, fairly free from foreign cells, could be depended
upon to give a colorless filtrate when boiled.

What, then, is the impurity which produces "purple x"?
By a series of experiments properly checked and repeated, I was
able to eliminate successively, the filtered perivisceral fluid
(serum), the blood clot which had been washed in sea-water

1 From the Marine Biological Laboratory at Woods Hole and the Zoological
Laboratory of the University of Michigan.

135

136 ALVALYN E. WOODWARD.

to free it from serum, fresh fluid containing both serum and blood
cells, pieces of the alimentary tract, and mesentery. If a piece
of fresh mature testis be boiled in sea-water, in sea-water plus
sperm, or in distilled water, a purple compound is formed. This
color can also be obtained by treating the testis with strong
alcohol, which, as observed under the microscope, turns testis
cells from greenish brown to purplish. Whether or not this
would happen with immature testes could not be determined,
since the experiments were carried on during the breeding season.

The sperm from one male, when examined under the micro-
scope, showed the presence of a few small pieces of testis, and,
when boiled, or treated with alcohol, turned purple. The testes
were then washed in running sea-water for about three hours,
during which they were occasionally pinched with forceps to
help free them from sperm. When this was apparently all washed
out, a portion of the testis was boiled in sea-water, without
giving a purple color. Treating with absolute alcohol, boiling in
distilled water, and boiling in sea-water plus fresh sperm also
failed to bring out any purple tinge from pieces of the washed
testes. The rest were then left standing over night in a finger-
bowl of sea-water. After twenty-four hours, it was observed
that they tinged the water slightly but distinctly purple.

It seejns clear, then, that "purple x" arises from the fresh
testis, or from a reaction between testis and sperm. It still
remains to be learned whether or not it may be obtained from
the ovaries also.

II. CHEMICAL NATURE OF "PURPLE X."
The best known pigment of sea-urchins is echinochrome,
whose chemical reactions are well established. It may be ob-
tained from the red blood cells by laking with distilled water,
by extracting with alcohol, chloroform or ether. The neutral
extracts are a cherry red, the addition of a small amount of NaOH
turns them yellowish, and acidulating with HC1 produces a red-"
yellow color (MacMunn, '85).

Echinochrome was extracted from the washed blood clot
by laking with one volume of distilled water. An equal amount
of "double sea-water" (sea-water boiled down to one-half its

NOTE ON THE NATURE AND SOURCE OF " PURPLE X.'

137

original volume) was then added so that the salt content might
equal that of the solution with which it was to be compared.
A series of experiments was then run in duplicate, using this
echinochrome on the one hand, and "purple x" on the other.
The addition of NaOH and HC1 to echinochrome gave the results
described by MacMunn. No visible change could be obtained
in the "purple x," however, by the addition of either reagent.
Thus it was established that, whatever the chemical nature of
"purple x" may be, it is not identical with echinochrome.

III. PHYSIOLOGICAL EFFECT OF BOILED SPERM.

The addition of boiled sperm suspension to eggs, in Arbacia,
causes the jelly surrounding the egg to swell, as can be demon-
strated by putting them into India ink. It also becomes more
sticky. As a result of this, the eggs adhere to each other and
to the bottom of the dish. When fresh spermatozoa are added,
many get caught in the jelly and form "halos," but some are
able to penetrate to the eggs, so that fertilization is not inhibited.

However, when "purple x" is present in the boiled sperm,
fertilization by fresh sperm and auto-initiation by egg-secretion
is inhibited, as shown in the following table. The percentages
were all obtained by counting 200 or more eggs, and a vertical
column represents eggs from the same female, treated in dif-
ferent ways.

J

Per Cent, of Kggs Divided.

A

B

c

D

£

F 1 <;

Eggs without sperm . .

O

84

O

O

72

68

0

73

55

0

72.3
76.7
26.6

O
P<

82
39

o
jlys

43
o

o
Dermy

56.9

i±

Eggs + sperm

Eggs + boiled sperm (colorless) + fresh sperm .
Eggs + boiled sperm (purple) + fresh sperm . .
Eggs + egg secretion + hypertonic after-treat-
ment

Eggs + egg secretion + boiled sperm (purple)
+ hypertonic after-treatment .

UNIVERSITY OF MICHIGAN.
March 17, 1915.

LITERATURE.
Glaser, Otto.

'14 On Auto-parthenogenesis in Arbacia and Asterias. BIOL. BULL., Vnl.

XXVI.
MacMunn.

'85 On the Chromatology of the Blood of Some Invertebrates. Quart. Jour.

Mic. Sci., Vol. 25.

ARE FUNCTION AND FUNCTIONAL STIMULUS FAC-
TORS IN PRODUCING AND PRESERVING
MORPHOLOGICAL STRUCTURE?

EDUARD UHLENHUTH, PH.D.,
ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH, NEW YORK.

Since the days of Lamarck the attempt has often been made
of explaining the genesis of the morphological structure of or-
ganisms through the theory of adaptation. A special form of
this theory is that of "functional adaptation" which was for-
mulated under this name byWilhelm Roux about 1880, and was
later elaborated by that investigator in an extremely extensive
and thorough manner.

The most striking organic structures are those which like the
bones seem to be constructed on a definitely purposeful plan,
offering the largest amount of strength with the smallest amount
of material. Other organs, such as the muscles, increase in
size as a result of increased function. Roux named this phe-
nomenon "functional adaptation," while the structures under-
lying this principle he described as "functional structures."
He made a number of exceedingly careful anatomical studies
of such "functional structures." Endeavoring to explain the
genesis of such seemingly purposeful structures from a purely
mechanical standpoint, he found that they possessed exactly
that construction which was to be expected from a mathematical
calculation based on the principle of functional adaptation.

In order to make clear the development of functional adapta-
tion, Roux fell back upon the most primitive particles of living
matter. In his opinion some of these particles have been adapted
to respond to functional stimuli, that is, they show a greater
tendency to proliferate in the presence of functional stimuli
than in their absence. Thus, those elements which were sub-
jected to stimuli soon predominated over those which were not
thus exposed.

If functional structures consisted of such particles, they would

FUNCTION AND FUNCTIONAL STIMULUS. 139

show the following characteristics: first, enlargement (functional
hypertrophy) with increased stimulus; second, atrophy from
inactivity upon cessation of the stimulus; third, regeneration
of organs of functional structure to the condition of the original
structure would only be possible in the presence of function and
functional stimuli (functional regeneration); and fourth, the
successful transplantation of functional structures would only
be conceivable if after transplantation the organs involved were
supplied with an appropriate amount of functional stimulus.
The principles, therefore, necessary for an acceptance of the
theory of functional adaptation are: functional hypertrophy,
atrophy from inactivity, functional regeneration, and functional
transplantation. (Various other factors which are similarly
instrumental in this connection will be dwelt upon more exten-
sively in a subsequent paper.) Should it now appear that these
essential principles in so-called functional structures are non-
existent, the theory of adaptation would then fail to adequately
account for so-called functional structures.

Before giving the results of my experiments I wish to say a
few words about the general value of Roux's theory of functional
adaptation.

This theory at a cursory glance has many merits, and at the
time when first introduced by Roux it marked a substantial
advance, inasmuch as it showed that it was possible to produce
highly purposeful structures through the influence of purely
mechanical principles. On this basis the theory has become
a material factor in connection with the investigations and studies
of many workers in the realms of pathology and physiology.

Roux's theory, however, has great disadvantages, one of
which is its extremely complicated and extensive terminology.
It is probably owing to the extreme obscurity of the doctrine
that comparatively few investigators, apart from the originator
of the theory, have familiarized themselves with it and fully
understood its principles. As a result of this practically the
only valuable work along this line has been carried on under the
direct control of Roux himself, while in contrast with the com-
paratively scanty publications emanating from Roux's laboratory
an amazingly voluminous mass of literature has been supplied

140 EDUARD UHLENHUTH.

by outside writers, all of whom were under the impression that
they were contributing something to Roux's theory, but who
in reality had hardly grasped more than a few but imperfectly
comprehended terms connected with the essential problem.
These for the most part misleading publications have caused
more error than progress in experimental work. I shall endeavor
to prove this point in a subsequent and more extensive article.

A second disadvantage of Roux's theory is the fact that the
extent of the field in which it is applicable becomes more and
more restricted with the increasing number of "experimental"
investigations bearing directly upon this problem. Conse-
quently, in course of time the phenomena which are not in
harmony with this principle of functional adaptation increase
in number, although they may be explained, together with other
phenomena, from another point of view.

However, by far the weightiest objection to the theory is that
it threatens to become more and more of a stumbling block to
workers who are setting out to investigate the problems of or-
ganic and inorganic life from a common viewpoint. Nowa-
days theories such as that of function and functional stimulus
can hardly be reconciled with a chemico-physical view of the
life processes; for the whole underlying principle of the theory
of adaptation does not lend itself to methods of measurement.
For this reason a detailed revision of the theory of functional
adaptation has become necessary and will be published in a later
communication.

In the present paper I shall report on the results of a few
experiments which I began five years ago with the above men-
tioned object in view. As they are at the present moment
sufficiently advanced to allow of a survey of the whole point
at issue and to show that they are qualified to throw some light
on the problem of functional adaptation, a preliminary discussion
of these experiments may be warranted.

The experiments in question were performed on the trans-
planted eye of Salamandra maculosa. The eye of a larva was
transplanted into the neck of another larva, where after a few
days' partial or complete disappearance of the retina resulted,
ending finally in complete degeneration. The remarkable fact

FUNCTION AND FUNCTIONAL STIMULUS. 14!

was that this degeneration was followed later by a complete regen-
eration of the transplanted eye, even in the dark.

My experiments were conducted with the following considera-
tions in view. The retina is one of those structures which ac-
cording to Roux's definition should be called a functional struc-
ture. But in accepting this definition one need not necessarily
assume that these structures are the result of functional stimulus,
for such structures may have been predetermined in the embry-
onic stage. They develop independently of functional stimulus
until they reach the third period — the so-called functional
period.1 Functional structures can only exist during this period
if they are supplied with a sufficient amount of functional
stimulus; otherwise they undergo atrophy through inactivity.
From Roux's point of view it would appear probable that the
morphogenesis of the eye of Urodela is only partially determined
by inheritance, and in accordance with this determination it
would reach the same stage of development as that attained by
Proteus anguineus. The further development of the eye, com-
prising the formation of the rods and cones, that is the true
functional parts, would be brought about by the penetration
of the rays of light through the skin of the salamander, to its
ovaries, and would therefore be the outcome of functional
stimulus. Secerov2 asserts that the skin of Salamandra macu-
losa permits of the penetration of approximately 1/173 of the
quantity of light in which the animal lives.

The eyes of Proteus, which inhabits dark caves, must there-
fore remain in the primitive Proteus stage, according to the state-
ment of Kammerer,3 who believes he has demonstrated that
further differentiation can only occur if Proteus be kept in
the light. Thus, he ascribes the process of full differentiation

1 See, for example, Roux, "Die Entwicklungsmechanik, ein neuer Zweig der
biologischen Wissenschaft," Vortrage und Aufs. iiber Entwickl.-Mech. d. Organ-
ismen. 1905, Heft I., p. 94, note n. Also Roux, "Die vier Hauptperioden der
Ontogenese, sovvie das doppelte Bestimmtsein der organischen Gestaltungen,"
Mitteilung der naturforschenden Gesdlschaft, Halle a.d.S., 1911, I., p. i.

2 Secerov, S., "Die Umwelt des Keimplasmas. II. Der Lichtgenus im Sala-
mandra-Korper," Arch. f. Enlwcklungsm., 1912, XXXIII. , 682.

3 Kammerer, P., "Experimente iiber Fortpflanzung, Farbe, Augen und Korper-
reduktion bei Proteus anguineus Laur. Zgl. Vererbung erzwungener Farbverander-
ungen." III. Mitteil., Arch. f. Entwicklungsm., 1912, XXXIII., 349.

142 EDUARD UHLENHUTH.

of the eye in the case of several Proteus which were kept in the
light, directly to the influence of function, and Roux is apparently
of the same opinion. This point will be discussed by the writer
in a later paper.

My own observations on the transplanted eye of Salamandra
soon convinced me that this case lends itself very well for the
test of the theory of functional adaptation.

First of all I severed the optic nerve, a procedure which ac-
cording to general opinion should induce permanent degeneration
of the retina as a result of the eye becoming isolated from the
brain. In all previous operations of this nature, where, however,
the eye remained in its normal position, a reunion of the ampu-
tated stumps of the nerve took place, so that it was natural to
suppose that the re-connection of the eye with the brain brought
about regeneration of the retina. In my own experiments I
obviated the possibility of such subsequent reunion of the eye
with the brain by transplanting the eye to an abnormal position
(in the neck of the salamander). But regeneration took place
in spite of this fact. The point of chief interest, however, is
the fact that by means of this operative measure, which, as
has been demonstrated in 95 per cent, of the cases, excludes
reunion of the eye with the central nervous system, the eye is
permanently deprived of functional power.

It is thus obvious that in these eyes no function was pos-
sible and the experiment therefore shows that a whole series
of phenomena, hitherto designated as cases of functional adapta-
tion, require a different explanation. We will now discuss these
phenomena in greater detail.

I. In about a week after transplantation of the eye into the
neck of the salamander the retina had degenerated to such an
extent that in many cases only the peripheral part of the retina,
which was not differentiated in layers, had survived.1 But
in spite of its permanent isolation from the brain and despite
the fact that the eye was permanently deprived of function,
the retina from this time on began to show signs of regeneration
and the transplanted eye began to receive a progressively im-

1 E. Uhlenhuth, "Die Transplantation des Amphibienauges," Arch. f. Ent-
ivcklungsm., 1912, XXXIII.

FUNCTION AND FUNCTIONAL STIMULUS. 143

proved supply of blood, so that in a comparativey short time
(from 4 to 6 weeks) it had regained a perfectly normal structure.

II. This regeneration of the transplanted eye even takes
place when the organ is deprived of functional stimulus by
light. A series of salamanders operated on in the above-de-
scribed manner, were placed in a dark room where neither red
nor white light could penetrate to their eyes; but in spite of this
fact the transplanted eyes regenerated and developed a normal
retina.

These experiments show that the "quality" of this process,
namely, regeneration as such, is independent of any sort of func-
tional influence. We are here dealing with a case of simple
regeneration, such as is found in many organs, not with func-
tional regeneration, such as we might expect to find in so-called
functional structures. Of course this fact does not warrant us
in entirely rejecting the theory of functional adaptation, for
the possibility must not be ignored that although regeneration
occurs in eyes treated in this way as a result of the agency of
certain other factors, nevertheless degeneration brought about
by atrophy through inactivity might follow later, as a result of
the permanent lack of function and functional stimulus, a possi-
bility which would be expected to arise according to Roux's
theory.

III. But secondary degeneration as a result of atrophy from
inactivity failed to occur in my experiments, even when the
eyes were permanently deprived of function, as occurred in the
"light" series.

IV. Degeneration similarly failed to occur in the transplanted
eyes which were permanently deprived of both function and
functional stimulus, namely in the "dark" series. These eyes,
although severed from the brain and in permanent darkness,
grew and metamorphosed simultaneously with the normal eyes
of the hosts.1

Up to the present time I have had at my disposal preparations
of eyes of the "dark" series which were preserved 15^ months
after transplantation; at that time the hosts were about 21

1 E. Uhlenhuth, "Die synchrone Metamorphose transplantierter Salamander-
augen (Zugleich, Die Transplantation des Amphibienauges II. Mitteil.) Arch,
f. Entwcklungsm., April, 1913, XXXVI., 211.

144 EDUARD UHLENHUTH.

months old, and all the structures, with the exception of the
sex organs, were perfectly developed. The retinas of these
transplanted eyes were found to be normal in every detail.

In addition to the above I have a preparation of an eye of the
"light" series, which was preserved 3 1/4 years after trans-
plantation, at which time the sex organs were also fully developed.
According to Roux the eye of an amphibian should by that time
already have entered the functional period, as is believed to
have been proved in the case of the eye of the Proteus. Never-
theless, the old transplanted eyes were also found to be normal,
and the functional parts of the retina, viz., the rods and cones,
were present and well developed.

The above results, namely, the permanent preservation in a
normal condition of transplanted eyes, prove beyond any doubt
that the so-called functional structures of the eye do not undergo
atrophy through inactivity, even if they are kept under extremely
unfavorable conditions and are deprived of all function and func-
tional stimulus.

This fact alone is sufficient to show that atrophy from inac-
tivity, which is one of the fundamental postulates of Roux's
theory, is by no means a phenomenon of general occurrence
which takes place in all so-called functional structures perman-
ently deprived of functional stimulus, as was supposed to be
the case.

As far as regeneration is concerned, the experiments mentioned
above only serve to show that regeneration in itself is indepen-
dent of function and functional stimulus.

V. I am, moreover, able to demonstrate that the "quantity"
of the regenerative process in the eyes, viz., the rapidity of this
regeneration are not influenced by functional stimulus, viz., light.

A certain number of animals (260 in all) of both the light
and the dark series were preserved at certain intervals of time
and the transplanted eyes were cut in sections. Eyes preserved
at equal intervals of time were then compared with each other.
It was found that the transplanted eyes of the same series which
had lived on the host for the same period of time may show
considerable differences in the rapidity with which they under-
go regeneration, even if they are subjected to equal conditions

FUNCTION AND FUNCTIONAL STIMULUS. 145

as far as functional stimulus is concerned. These differences
must therefore be caused by other non-specific factors. In
order to ascertain how far-reaching is the influence of light, it
was necessary to determine the average rapidity of regeneration
in every group of eyes of the same series and make a curve for
each series. Although these curves are not yet completed, the
results thus far obtained show no differences between light
and dark series at all. Even the quantity of the regeneration is
therefore uninfluenced by light.

From the above data we must draw the following conclusions:
Functional adaptation plays no part either in transplantation
or in regeneration of the retina; nor is it a factor which determines
either the quality or the quantity of these processes.

This, of course, does not mean that regeneration or trans-
plantation of the eye cannot be influenced at all by chemical
or physical factors. On the contrary, as is shown by the dif-
ferences between transplanted eyes of the same series, examined
at equal intervals of time after transplantation, the quantity
of regenerative processes, viz., the rapidity of regeneration are
subject to variation by one or more factors. The point of im-
portance is the fact that these factors are not connected with
the specific functional stimulus, viz., light. Apparently they
are the same factors which also affect the rapidity of the re-
generative process of other organs. The factor concerned is
probably the length of time since circulation in the transplanted
eye was reestablished.

Hitherto the regeneration of the retina has been considered
as being different from the regeneration of the bones. It was
supposed that a bone could only regenerate its architectural
structure if in use, otherwise the result would not be a normal
bone but a disorderly bony mass.

Aug. Bier,1 however, has shown that even the bones do not
regenerate as a mere indefinite mass of bone if kept without
functional stimulus, but that on the contrary, in the absence
of any such stimulus, they resume their original functional struc-
ture to the minutest details. In certain experiments a con-
siderable part of a human tibia was removed. Although these

1 Aug. Bier, " Beobachtungen iiber Knochenregeneration," Arch. f. klin. Cliir.,
Dec. 1912, c, 91.

146 EDUARD UHLENHUTH.

bones were not exposed to any function, the tibiae after a certain
length of time assumed their normal shape and structure, but
only if they were supplied with nourishment in a proper way,
and if sufficient space was left for them to regenerate the missing
part.

The same is true of the tendons. H. Triepel1 showed that
the tendo Achillis of a cat can regenerate only tendon tissue,
irrespective of the presence or absence of functional stimuli.

Our own experiments now prove the same principle also in
the case of the eye of Salamandra maculosa; this organ regener-
ates its functional structures in the absence of functional stimulus,
and furthermore it retains its structure permanently, despite
the permanent absence of functional stimulus. For a long time
it was believed that a bone only regenerated a structureless
mass of callus in the absence of function, and according to this
theory it would be assumed that an eye if once degenerated
would in the absence of light regenerate undifferentiated
retina cells, such as we find in the normal Proteus eye. Never-
theless, both eye and bone regenerate the normal and fully
differentiated structure, even in the absence of functional
stimulus.

We have seen that in my experiments the velocity of the proc-
cess of regeneration was not influenced by function; and even
if this had been the case it could not be used as a proof in favor
of functional adaptation. There are a number of well known
morphogenetic processes, the rapidity of which can be acceler-
ated by light, although light bears no relation to the function
of the developing organ, that is to say, is not a functional stimu-
lus. The most striking experiments made in this connection
are those of J. Loeb,2 in which he showed that the regeneration
of the hydrants of Eudendrium is impossible in the absence of
light. Nevertheless, we cannot call this a case of functional
adaptation, because here light is obviously not a functional
stimulus. On the other hand, development of the eyes of
fish embryos cannot be prevented by the exclusion of the func-

1 Triepel, H., "Selbststandige Neubildung einer Achillessehne," Arch. f. Ent-
•wcklungsm., Aug., 1913, XXXVII., 278.

2 Loeb, J., "Uber den Einfluss des Lichtes auf die Organbildung bei Tieren,"
Pflilgers Arch., April, 1896, LXIII., 273.

FUNCTION AND FUNCTIONAL STIMULUS. 147

tional stimulus. But it is incorrect to ascribe this fact to the
assumption that heredity may have fixed this character so far
that development of an eye now occurs in the absence of light.
For Loeb1 has shown that it is very easy to prevent the develop-
ment of these eyes by a number of different means, such as
lack of oxygen, which is a non-specific, non-functional factor.
Moreover, it is not necessary to point out that the influence of
light on a photographic plate has never been considered to be a
case of functional adaptation, although the sensitiveness of the
plate to light is just as much response to a physical factor as is
the regeneration of Eudendrium or as might be the variation in
the rapidity of eye-regeneration which although not found in our
experiments, might possibly have occurred.

The theory of functional adaptation complicates instead of
simplifying the problem. What we should emphasize is not
the fact that the result of the response to light is different in
the case of Eudendrium from what it is in the case of the regener-
ated eye or of the photographic plate, but the fact that these
three phenomena all possess a common basis. It is obvious
that all three are governed by the same laws, with which we
are familiar from our knowledge of physics and chemistry; but
these laws are free from such terms as function, functional
stimulus, or any other stimulus, or the principle of adaptation.
In order, therefore, to obtain a fertile method for attacking
the problems confronting us we must constantly bear in mind
the fact that the same laws expressed in the same terms can
explain both organic and inorganic phenomena.

1 Loeb, J., "Heredity in Heterogeneous Hybrids," Jour, of Morphol., March,
1912, XXIII., i.

Vol. XXIX. September, 1915. No. j

BIOLOGICAL BULLETIN

STUDIES IN ARTIFICIAL PARTHENOGENESIS.
II. PHYSICAL CHANGES IN THE EGG OF ARBACIA.

L. V. HEILBRUNN.

CONTENTS.

I. Physical Organization of the Egg 149

II. Cortical Changes 158

A. Membrane Elevation and Membrane Swelling 158

B. Permeability Changes in the Vitelline Membrane 160

C. Theories of Membrane Elevation 163

D. Cortical Action of Heat and of Various Chemicals 169

E. Inhibitors to Membrane Swelling 174

F. Cortical Changes at Fertilization 178

G. The Significance of Cortical Change 183

III. Internal Changes. The Problem of Segmentation 184

A. Theories of Segmentation 184

B. The Action of Hypertonic Sea-water in the presence of KCN 185

C. An Analysis of the Methods of Producing Segmentation in the

Unfertilized Arbacia Egg 188

IV. Summary 200

V. References 201

I. PHYSICAL ORGANIZATION OF THE EGG.

In spite of the many researches on the sea-urchin egg very
little is known of its physical make-up. Year after year the
egg is used in attempts to analyze fundamental problems and
various theories have been based upon experiments with it.
And yet, but little is known definitely about the type of physical
organization which it possesses. Is the egg essentially fluid
or is it a more or less rigid jelly, is the unfertilized egg surrounded
by a microscopically visible membrane or by a hypothetical
film beyond the limits of vision, what indeed is the nature of
the membrane which controls osmotic interchange? These

149

150 L. V. HEILBRUNN.

are some of the questions which must be considered if our theories
are to be more than mere generalizations.

The first point to be decided upon is the physical nature of
the egg contents. Embryologists in the past have often ob-
served flowings in egg cytoplasm and such observations indicate
a fluid composition. This would accord with Rhumbler's
demonstration of fluidity in other types of protoplasm. We
might accept these results without further comment if it were
not for the fact that recently there has been a tendency to regard
protoplasm as a gel. Kite ('13) has in fact concluded that the
protoplasm of the starfish egg is of this nature. The protoplasm
of the normal unfertilized sea-urchin egg is undoubtedly in a
typically fluid condition. If pressure is applied to eggs under-
neath a coverslip, the egg contents flow out, indeed, if the pres-
sure is vigorous enough, the protoplasm is shot out in a long
stream as from a pipette. Facts such as these are probably
known to many embryologists, similar observations have been
especially described by Reinke ('95) and Albrecht ('98).

If the egg is essentially a fluid mass, what is there to prevent
it from diffusing through the sea-water? Two possibilities
exist, either (i) The substance of the egg is as a whole insoluble
in sea-water or (2) it is surrounded by a membrane insoluble
in sea-water.1 The first possibility is excluded by the fact that
we know protoplasm to be in aqueous solution. We must there-
fore conclude that the egg is surrounded by a membrane insoluble
in sea-water. Careful observation reveals the existence of a
membrane around the unfertilized egg; just outside of the darker
substance of the egg cytoplasm, a dim outer line can usually be
made out. The faintness with which the outer margin of this
vitelline membrane appears is apparently due to the fact that
its refractive index is very close to that of sea-water. That the
appearance of a membrane is not the result of a diffraction
illusion is shown by the fact that the membrane may be isolated
by pressing out the egg contents, as Herbst first found. At
fertilization, as will be pointed out more fully later, it is this

1 The oft-made assumption of a surface film like that found at the surface of
peptone solutions, etc., is really a special case of the first alternative. For such a
film could only exist at a surface of discontinuity, and this could only occur at
the junction of 2 immiscible fluids.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 15!

membrane which becomes lifted from the surface of the egg as
the fertilization membrane.

A study of the chemical behavior of the membrane gave
results of interest. Dilute acids cause it to swell. When one
part of normal HC1 is diluted with 9 parts of sea-water the
resultant solution causes a marked swelling. The membrane
becomes sticky and agglutination follows.1 Dilute solutions
of nitric, butyric, and valerianic acids give similar results. As
far as can be ascertained however, this acid swelling of the mem-
brane does not result in complete solution. On the other hand
dilute alkaline solutions, although they cause little if any swelling
of the membrane, are quite effective in dissolving it away. In
order to study the effect of alkali on the membrane, it is best
to shake the jelly off the eggs first, as this often becomes
saturated with the Mg (OH)2 precipitated by the alkali, and
obscures the result. Eggs in 50 c.c. of sea- water plus 2.5 c.c.
n/io NaOH soon become sticky; they cling to the bottom
of the dish and to each other. Soon the exterior surface of
the egg becomes rough, it is evidently no longer surrounded
by a membrane (and it is only prevented from diffusing through
the sea-water by coagulation). Many salts exert a swelling
effect on the membrane, and in some cases this is accompanied
by a complete solution of the membrane. When eggs are
dropped into 0.55 N Nal, the swollen membranes are very ap-
parent after 10 or 15 minutes have elapsed. In 40 minutes
almost all of the eggs are no longer surrounded by a membrane,
the periphery of these eggs instead of being smooth, now pre-
sents a roughened appearance. It is quite evident that the
protoplasm is naked and that the membrane has been completely
dissolved away by the sodium iodide. This behavior of the
membrane towards acids, alkalis, and salts, indicates its protein
nature.

That it is not a lipoid is evident from the fact that it is insoluble
in any of the ordinary lipoid solvents. It might however contain
an admixture of lipoids. This is rendered improbable by the
following line of evidence. As was pointed out above, the almost
invisible character of the membrane indicates that its refractive

1 Loeb ('08) described agglutination in HC1.

152 L. V. HEILBRUNX.

index is very close to that of sea-water. Now it is well known
that many proteins have a refractive index close to that of sea-
water; on the other hand, lipoids have a considerably higher
refractive index. As the refractive index is an additive function
of the constituents of a mixture, the presence of lipoids in any
abundance would make it impossible for the refractive index of
the membrane even to approach that of sea-water. Hence no
great admixture of lipoids can be present. More direct evidence
of the absence of any appreciable quantity of lipoids is also avail-
able. The membrane was tested with a Scharlach R solution
such as recommended by Herxheimer. In order to render them
more visible, the membranes were made to swell by placing the
eggs in 15 c.c. sea-water plus 10 c.c. 2.57VNaCl. To a drop of
Scharlach R solution on a slide was added a drop of egg suspen-
sion. The vitelline membrane remained hyaline and unstained,
whereas the egg cytoplasm itself was colored red. The Scharlach
R solution used in this test was a saturated solution of the dye
in equal parts of acetone and 70 per cent, alcohol.

If the egg contents be made to flow out from the membrane or
if the egg be cut or shaken into fragments, a new membrane
immediately forms around the momentarily naked protoplasm.1
Such a membrane has the same chemical properties as the vitel-
line membrane. The acids and salts which produce swelling
in the latter, cause it to swell also. The immediate production
of a protein membrane about these egg fragments must be re-
garded as similar to the formation of precipitation membranes
like those studied by Traube and Quincke. Evidently, some
protein contained in the egg is precipitated on contact with the
outer sea-water. It is probable that this protein is, in the in-
terior of the egg, prevented from coagulation by the presence of
a protecting colloid. At the outer surface of the egg, the mem-
brane-protein is coagulated by direct contact with sea- water.
Support for this view is found in the fact that the presence of
some colloids (e. g., egg albumen) in the sea-water, causes the
membrane to take up water and swell. Since the vitelline mem-

1 O. and R. Hertwig in their " Untersuchungen zur Morphologic und Physiologic
der Zelle," Heft 5, 1887, first observed the elevation of membranes on egg frag-
ments. (This observation was also repeated by Ziegler, Arch. f. Ent. Mech., VI.,
249 (1898), by Moore, Univ. of California Publications in Physiology, IV., 89 (1912).

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 153

brane behaves toward reagents in the same manner as does the
membrane on egg fragments, we are justified in regarding it too
as a precipitation membrane. The most noteworthy feature of
the precipitation membranes of Traube and Quincke is their
semipermeable (or partially semi-permeable)1 character. \Ye
should therefore expect the vitelline membrane to exhibit semi-
permeable properties, and in this way to govern osmotic inter-
change. When the egg is caused to shrink in a hypertonic
solution, the vitelline membrane shrinks with it. This has been
observed in a great variety of hypertonic solutions. For such
a shrinkage, there are 3 possible types of explanation. The
inward tension may be the result of a force arising from (i) the
substances within the membrane, (2) the membrane itself, or
(3) something immediately outside of the membrane.

Physico-chemically the possibilities, as I see them, are (i)
The fluid interior of the egg is coagulated by the various hyper-
tonic solutions, shrinks as a result and pulls the membrane with
it. (2) The vitelline membrane as a result of its semipermea-
bility2 is responsible for the shrinkage. (3) There is an invisible
semipermeable membrane outside of the vitelline membrane.
As there is nothing to warrant the assumption of any semiper-
meable membrane outside of the vitelline membrane, it seems
scarcely necessary to discuss the third possibility. The first
possibility is essentially the position maintained by M. Fischer
('io). He regards the passage of water into and out of cells as
due primarily to the attraction of the interior colloids for water.
According to this view endosmosis, such as occurs in hypotonic
solutions, would be the result of the taking up of water by col-
loids within the cell. That the process of water absorption is
by no means dependent on the affinity of the egg colloids for
water is conclusively proven by the fact that influx of water

1 The existence of an absolutely semipermeable precipitation membrane (i. e.,
one which prevents the passage of all substances in solution, but permits the
passage of solvent) is extremely doubtful; cf. for example Quincke '02. The fact
that a substance may penetrate a membrane and yet exert considerable osmotic
pressure against it, has often been neglected by biologists. To assume that
because a substance passes through a membrane it can exert no osmotic pressure
against it, is just as foolish as to assume that the air in an air-bubble exerts no
pressure on the film of water surrounding it.

- Using the term in its broadest sense.

154 L- v- HEILBRUNN.

leads not to transformation of gel to sol, as we would have to sup-
pose, but on the contrary such an influx transforms the egg pro-
toplasm from the sol to the gel condition. The experimental
evidence in support of this fact is given on p. 195. Water absorp-
tion by the egg is indeed correlated with a loss of water on the part of
the egg proteins. The passage of water into and out of the cell
can not therefore be due to changes in the water content or
aggregation state of the egg proteins, and the first possibility is
unable to account for the facts. Only the second possibility
remains, i. e., that the vitelline membrane acts as a semiper-
meable membrane and controls osmotic intercourse. On the
basis of this view many facts are understandable, which can be
explained in no other way. This will, I think, be demonstrated
in the course of the argument.

The plasma-membrane of a cell is defined as the membrane
which governs its osmotic intercourse. According to this defi-
nition, the vitelline membrane is the plasma-membrane of the
Arbacia egg cell. Hitherto no one has either described a plasma
membrane, or studied directly the properties of one. Although
he often uses the concept of such a membrane, Lepeschkin ('n)
admits that the actual structure is "zurzeit unbekannt." Loewe,
('13) in referring to the plasma-membrane, says: '"Allein weit
davon entfernt. dasz auch nur ihre Existenz irgendwie sicher-
gestellt ware, sind auch iiber die Beschaffenheit dieses hy-
pothetischen Gebildes die Meinungen so zahlreich wie die
Moglichkeiten." Fischer ('12) is indeed of the opinion that
"the entire conception of an osmotic membrane about cells is an
impossibility."

The plasma-membrane of the Arbacia egg is a protein gel.
As such, it possesses a certain degree of rigidity. Suppose a
hypothetical system completely surrounded by an extremely
rigid semipermeable membrane.1 If such a system were placed
in a concentrated solution no exosmosis could take place, for if
the membrane were perfectly rigid, there could be no removal
of solvent from the system without the production of a vacuum.
But the membrane would be subjected to a considerable pressure,
which would tend to make it rearrange its particles in such a

1 Possibly this is the case in the Fundulus egg.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 155

fashion that the volume enclosed within it might be lessened.
Whereas an extremely rigid membrane would resist such forces,
one with only a certain degree of rigidity would yield (in the case
of sufficient pressure), and exosmosis would be possible. Thus
osmosis in an enclosed system depends, to some extent at least,
on the rigidity of the confining membrane. These conclusions
apply in some measure to the sea-urchin egg, for the vitelline
membrane possesses a slight degree of rigidity. Most salts in
hypertonic solution cause the membrane to absorb water and
swell, the gel becomes less stiff, and the particles of the mem-
brane can more readily rearrange themselves. Thus shrinkage
of the egg is favored. We should in fact expect that hyper-
tonic solutions of salts which cause membrane swelling would
be more effective in causing shrinkage than those which do not.
Direct evidence of this fact is difficult to obtain without in-
troducing complications. One might compare the shrinkage of
the egg in two solutions of equal osmotic strength, one of
which causes membrane swelling and the other of which does
not. But we have no means of knowing when two solutions
are of equal osmotic power. Vant Hoff's law does not apply
with sufficient accuracy to warrant its use,1 and even if two
solutions could be obtained which were isosmotic towards a
given membrane, they would not necessarily be isosmotic
towards the plasma-membrane of the Arbacia egg, which is very
probably only partially semipermeable. Fortunately theie is a
way out of the difficulty. I found that when 2.5 N NaCl was
added to sea-water in the proportion of 8 parts by volume of
the former, to 50 of the latter, the resulting hypertonic solution
would usually cause membrane swelling when freshly prepared,
but would in large measure lose this power after it had stood for
some time. Thus one can obtain two solutions of identical
strength, one of which produces a softening effect on the mem-
brane, the other lacking this effect. A number of experiments
showed that in every case the eggs shrank more in the solution
which caused membrane swelling than in the solution which
left the membrane with its original rigidity. Two sample
experiments are recorded here. To save time the term "NaCl

1 Cf. Findlay, "Osmotic Pressure," London, 1913 (Chapt. IV.).

156 L. V. HEILBRUNN.

hypertonic sea- water" is used to designate 50 parts (by volume)
of sea-water plus 8 parts of 2.5 N NaCl.

August 21. Two small stender dishes, A and B, were used.
A contained 29 c.c. of NaCl hypertonic sea-water made up on
August 1 8, B contained 29 c.c. of NaCl hypertonic sea-water
freshly prepared. At 10:04 A.M., 5 drops of egg suspension
were added to A, and a similar amount to B.

The following measurements of egg diameters were made at
the times indicated. In making these measurements a Spencer
movable scale micrometer was used. It was not found possible
to obtain any very great accuracy in the use of this instrument.
If the eggs are not subjected to pressure of the coverslip, they
tend to move their position slightly. After various attempts
I reached the conclusion that the use of the movable scale was
inadvisable, especially as the usual difficulties of focusing made
very accurate measurements out of the question. Accordingly
the measurements were made with the scale stationary. No
great claim for accuracy is therefore made, but the error is not
over one micron. Fortunately the difference between the
diameters of the two sets of eggs is markedly greater than the
experimental error of the method. The measurements were
made at a magnification of about 650 diameters.

DIAMETERS OF EGGS IN A.

70 /j. 70.5 M at 10.25 A.M.

69 71 at 10.28 A.M.

69 70 at 10.29 A.M.

67.5 69 at 10.30 A.M.

69 69 at 10.51 A.M.

68 71 at 10.53 A.M.

68 71 at 10.55 A.M.

69 69 at 10.57 A.M.
68 69 at 10.58 A.M.

Average 69.2^.

DIAMETERS OF EGGS IN B.

66 fj. 66 M at 10.32 A.M.

67 68 at about 10.35 A.M.
66 67 at about 10.35 A.M.
66 66 at about 10.35 A.M.
66 66 at ii. 15 A.M.

70 70.5 at 11.17 A.M.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 157

The above egg showed a completely unswollen membrane, and so should be ex-
cluded from the experiment.

63.5 65 at about 11.20 A.M.

67 67.5 at ii. 24 A.M.

Average 66.2 /j. (egg with unswollen membrane excluded).

The eggs used in the above experiment were all from a single
female. The normal untreated eggs of this female were practi-
cally spherical and measured 74. 5M, 75M, 75/*i 74M, 75M, 75M, 75/".
75Mi 75-5M> 75M- They thus possessed an average diameter of
75/x and were practically all of the same size. The experiment
shows that the eggs in B with swollen membranes shrank more
than did the eggs in A with unswollen membranes.1 The average
decrease in diameter of the former was about 9^1, of the latter
about 6^. The difference was also constant, for the largest egg
with swollen membrane was smaller than the smallest egg with
unswollen membrane.

August 23. Two small stender dishes were used. A con-
tained 29 c.c. of "NaCl hypertonic sea-water," which had been
made up on August 18 (at n 130 'P.M.). B contained 25 c.c.
of sea-water. At 9.55 A.M., 4 drops of an egg suspension were
dropped into A and the same amount in B, and then 4 c.c.
2.5 TV NaCl were added to B, so that this solution became
"NaCl hypertonic sea-water." The diameter of the eggs was
determined as in the previous experiment. At 10:25 it was
noticed that some of the eggs in A were beginning to acquire
swollen membranes, so that only a few measurements were
made after this time, and only those eggs with unswollen mem-
branes were selected.

DIAMETER OF EGGS IN A.

65 fj. X 65 /j. at 10.12 A.M.

65 66 at 10.13 A.M.

67.5 70 at 10.14 A.M.

65 70 at 10.15 A.M.

65 68 at 10. 16 A.M.

63.5 X 70.5 at 10.17 A.M.

65 66 at 10.18 A.M.

1 This result was obtained in spite of the fact that membrane swelling always
tends to produce an increase in egg volume. Whenever membrane swelling occurs
in isotonic solutions, the egg rapidly increases its volume, it cytolyzes. The cause
of this cytolysis resulting from membrane swelling will be considered later.

X

68

at

10.19 A.M

X

65

at

10.20 A.M

X

66

at

10.26 A.M

X

69

at

10.30 A.M

X

68

at

10.31 A.M

X

67

at

10.32 A.M

Average 66.5 ,

u..

158 L. V. HEILBRUNN.

DIAMETER OF EGGS IN A.
65
65
66
66
66
66

DIAMETER OF EGGS IN B.

63 n X 63.5 M at 10.04 A.M.

63 63 at 10.05 A.M.

64 65 at 10. 06 A.M.

63 64 at 10. 06 A.M.

64 65 at 10.07 A.M.

62 65 at 10.08 A.M.

63 63.5 at 10.08 A.M.

62 62 at 10.09 A.M.

63 63.5 at 10.10 A.M.
63.5 64 at 10. ii A.M.

64 65 at 10.34 A.M.
63.5 63.5 at 10.35 A.M.
63.5 66 at 10.36 A.M.
63.5 65 at 10.37 A.M.

63 63.5 at 10.38 A.M.
62 66 at 10.39 A.M.

64 64 at 10.40 A.M.
63.5 65 at 10.41 A.M.
63.5 64 at 10.41 A.M.
63.5 65 at 10.42 A.M.

Average 63.7 /j..

In this experiment also only eggs from a single female were
used. The untreated eggs measured 72/4 x 73^1, 72^1 x 73, 7iju x
73M». 7I-5M x 74/z, 7i.5/zx73/z. Considering the eggs as spher-
ical, their average diameter was 72.4^. Thus on the average,
the diameter of the eggs in A decreased 5-9/i, whereas the diame-
ter of the eggs in B decreased 8-7/i. Thus exosmosis was much
more pronounced in the solution which caused membrane swelling
than in a solution of equivalent concentration in which this
effect was lacking.

II. CORTICAL CHANGES.
A. Membrane Elevation and Membrane Swelling.

When the sea-urchin sperm comes in contact with the egg,
almost immediately the vitelline membrane is lifted away from

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 159

the egg surface. This is the well-known process of membrane
elevation, or, as it is usually spoken of in this country, membrane
formation. During elevation, the membrane under normal
conditions does not undergo any evident increase in thickness.
The inner border of the vitelline membrane is often not very
plainly visible. In order therefore to estimate the thickness of
the membrane after elevation, it is convenient to compress the
eggs gently. The egg then becomes pushed out against the
vitelline membrane, in some directions at least, and the distance
between its outer border (i. e., the hyaline layer, see below)
and the outer boundary of the vitelline membrane is a measure
of the greatest possible thickness of the membrane. Under such
conditions, high power examination showed that the thickness
of the elevated membrane is approximately the same as that
of the unelevated membrane. Quantitative measurements were
not found to be practicable.

Moreover, after elevation the membrane still retains the
same chemical properties that distinguished it before fertiliza-
tion. In dilute HC1 it swells rapidly, and soon becomes sticky.
Nal also induces the elevated membrane to swell, much as it
did before fertilization.

After the vitelline membrane has been lifted from the egg
surface, a new membrane appears around the cytoplasm. This
structure has received an unusually large number of names;
of these I shall use the term "hyaline layer." It seems reason-
able to conclude that its formation depends on the same pre-
cipitation reaction which produces the membrane about egg
fragments, and which is probably also responsible for the vitel-
line membrane. This conclusion is supported by the fact that
it shows semipermeable properties.

The process of membrane swelling has often been confused
with that of membrane elevation. This is perhaps due to the
ambiguity of the term membrane formation. It has been the
custom to apply the term whenever the observer notices some-
thing at the egg surface which he did not see at the beginning
of the experiment. But membrane swelling and membrane
elevation are two very different processes and are usually
easily distinguishable under the microscope. The elevated

160 L. V. HEILBRUNN.

membrane appears as a thin membrane at some distance from
the egg surface, which is now bounded by the above-mentioned
hyaline layer. On the other hand, the swollen membrane
appears as a homogeneous layer surroundng the egg, a layer in
which neither the inner boundary of the vitelline membrane, nor
the outer boundary of the hyaline layer, make their appearance.
In addition to the purely morphological differences, there are
other distinguishing features. Among these may be mentioned
the fact that swollen membranes are always sticky, and as a
result, eggs with such membranes tend to agglutinate. Normal
elevated membranes are never sticky. Another criterion depends
on the fact that elevated membranes collapse when placed in a
solution of egg albumen (or other colloid).1 Swollen membranes
are of course unable to collapse.

B. Permeability Changes in the Vitelline Membrane.

The elevated membrane is known to be readily permeable
to electrolytes.2 Hence, since it offers considerable resistance
to their passage before elevation, it must undergo a change in
permeability at some stage in the process. An attempt was
made to determine if this increase in permeability took place
before or after membrane elevation. In the first case, it might
be considered as causally related to the process, and R. Lillie
has in fact suggested that the cause of membrane elevation is
an increased permeability of the "plasma-membrane." Ex-
periments, however, have shown that the increase in permea-
bility follows rather than precedes membrane elevation. Im-
mediately after elevation, the membrane is still more or less
impermeable to electrolytes. The following experiments show
this to be the case:

August 21, 1913. Eggs from a single female were washed
twice and then gathered into about 10 c.c. of sea- water at the
bottom of a small beaker. Four or five drops of diluted sperm
were then added, and the beaker shaken. At intervals of
i, 2, 3, 4, 5 minutes after insemination, the eggs were re-

1 It is only some few minutes after elevation that this collapse can be produced
by a colloid. Cf. p. 162.

2 Cf. Loeb, "Artificial Parthenogenesis and Fertilization," p. 208.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. l6l

moved with a pipette and dropped into Syracuse dishes filled
with 2 M MgCl2. Just two minutes after insemination, a
necessarily hasty examination of the eggs in the beaker showed
that all the eggs had well-elevated membranes at this time.
On the other hand, eggs removed from the beaker at the same
time and placed in 2 M MgCU, upon later examination, showed
no signs of ajt elevated membrane, and the membrane had
evidently been pushed back against the egg. In the following
table, the fractions in the second column indicate the proportion
of eggs which showed the membrane elevated in the various
Syracuse dishes. In each case, the numerator denotes the
number of eggs with membranes elevated, the denominator
the total number of eggs counted.

Minutes after Insemination Membranes Elevated
Before Transfer to 2 M MgCk. (Free from Egg).

1 0/50

2 0/50

3 1/50

4 30/50

5 49/50

In the case of the eggs transferred to the MgClo solution
three minutes after insemination, some of the membranes were
not completely collapsed, but one, two, or even several small
globular expansions could be detected at the egg surface.

This experiment was repeated on August 22, 1913, with almost
identical results. In this case the eggs were transferred to
2 M MgClo at intervals of 2, 3, 4 minutes after insemination.
At i^ minutes after insemination hasty observation showed all
the eggs to have well-elevated membranes. Thus the eggs
were placed in the magnesium chloride solution after membrane
elevation had occurred. Nevertheless, as the following table
shows, the eggs removed to MgCl2 two and three minutes after
insemination, showed no membranes free from the egg.

After 2 minutes, 0/50 with membranes free from egg.

3 i/50

4 13/50

These experiments were also confirmed in the summer of
1914. I have interpreted the results as indicating that after
elevation the membrane still remains impermeable to MgClo

l62 L. V. HEILBRUNN.

(or as would no doubt be a more exact expression of fact, the
membrane still offers sufficient resistance to the passage of
MgClo so that this [salt exerts osmotic pressure against it).
Indeed after being collapsed by the MgCl2, the previously
elevated membrane behaves as a semipermeable membrane
and shrinks with the egg.

That the membrane is still "semipermeable" shortly after
elevation is also shown by another series of experiments. It
was found that the well-known collapse of the elevated membrane
in solutions of egg albumen, did not take place immediately
after elevation.

In an experiment of July i, 1914, eggs were inseminated in a
Syracuse dish at 9:43 A.M. Two drops of egg suspension were
then placed 'in various Syracuse dishes containing 10 c.c. of
i per cent, egg albumen1 in sea-water, at intervals of 1,3, 6,
12^ minutes after insemination. In the case of the eggs placed
in the albumen solution i and 3 minutes after insemination,
the membrane was well elevated and had suffered no collapse.
On the other hand, when the eggs which had been placed in the
albumen solution 6 and \2\ minutes after insemination were
examined, their membranes were found to have been bent back
and collapsed. This experiment was repeated a number of
times, similar results being obtained in each case.

These surprising results find an easy explanation on the
ground that the membrane is still partially permeable shortly
after elevation. Because of this semipermeability it is subjected
to the pressure of the electrolytes contained within it. Only
when the membrane loses its semipermeable properties does
this pressure cease and only then can the albumen cause its
collapse. The albumen seems to exert a protective effect
on the membrane and to prevent increase of permeability. Thus,
when eggs were placed in albumen i minute after insemination^
their membranes were found to retain their semipermeability
and they collapsed upon being placed into 2 M MgClo 14
minutes after insemination.

From these experiments we can, I think, conclude that in-
creased permeability follows upon, rather than precedes, mem-

1 Kahlbaum's crystallized egg albumen was used. The solution was filtered.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 163

brane elevation. At a later point, in discussing what I believe
to be the true explanation of membrane elevation, I shall en-
deavor to point out a reason for this sequence.

C. Theories of Membrane Elevation.

The problem of membrane elevation is quite distinct from
that of segmentation, which is the central theme of the study
of artificial parthenogenesis. No method of producing mem-
brane elevation ordinarily results in more than one or two per
cent, of segmentations, and many methods usually result in
no segmentations whatsoever (e. g., chloroform, urethane).
Moreover, some of the best methods for producing segmentation
do not involve membrane elevation: thus hypertonic sea-water
never results in an elevation of the membrane.

Of late Loeb has been of the opinion that membrane "forma-
tion," involves the swelling of some substance at or near the egg
surface. To quote from his book1 "a colloidal substance which
lies below the surface layer of the unfertilized egg or is secreted
from the egg, suddenly swells by absorption of sea-water. In
the typical case of membrane formation this swelling results finally
in a complete liquefaction of the colloid. In other cases the
swelling is less complete and the formation of a gelatinous film
results." Loeb thus regards the process of membrane formation
as due to the swelling of a colloid, concerning the position of
which he is not quite clear. On page 213 of the same book he
says it is in the cortical layer. Finally, if the egg is left too long
a time in a solution which causes membrane formaton, not only
a colloid at the periphery, but colloids throughout the egg swell,
and cytolysis results.

In the discussion of Loeb's theory, I shall assume that he
uses the word "swell" in its colloid-chemical sense, i. e., to de-
note the absorption of a liquid (in this case water) by a gel.
Of course any less specific use of the term would rob the theory
of all theoretical significance. Briefly, the objections to the
swelling theory, as upheld by Loeb, are:

I. All the reagents which cause membrane elevation can
scarcely induce the swelling of any one colloid. Loeb claims

1 Loc. cit., p. 210.

164 L. V. HEILBRUNN.

that some of the "membrane-forming substances" cause the
egg membrane of the Mollusk Lottia to swell, but he admits
that this is not true of all of them. Some seem to cause swelling
and final liquefaction of the chorion or jelly of the sea-urchin
egg, but here again all are not effective, and I can suggest
chloroform as an exception. In fact it must be a strange colloid
which can be made to absorb water by such reagents as distilled
water, alcohol, chloroform, toluol, picric acid.

2. It is difficult to conceive of the location of the swelling
colloid. The egg has been shown to consist essentially of a more
or less rigid membrane surrounding a mass of fluid contents.
Evidently the latter can not swell, as only gels possess this prop-
erty. It is also demonstrable that the outer vitelline membrane
itself does not swell in the case of true membrane elevation,
for it can scarcely be doubted that the vitelline membrane
undergoes an increase rather than a decrease of rigidity after
elevation. And swelling is always correlated with a decrease
in rigidity on the part of the gel. It is difficult to understand
how Loeb seeks to explain by swelling what he regards as the
formation of a more or less rigid membrane.

3. Cytolysis results not in colloidal swelling and liquefaction,
but in coagulation. Loeb considers membrane elevation and
cytolysis as due to the same processes. He says:1 "Substances
like benzol, saponin, etc., can cause both membrane formation
and cytolysis. The first of the two is produced when they
have time to affect only the surface of the egg; cytolysis is
produced when their effect extends to the deeper layers of the
egg . . . the greater the fraction of the egg which conies under
the effect of the membrane-forming reagents, the greater the
amount of colloid that must be liquefied." Loeb thus states
explicitly that the membrane-forming reagents "liquefy" the
colloids of the eggs, and that this effect in the case of cytolysis
extends into the interior. Now it is a fact that the membrane-
forming reagents, far from producing liquefaction, have an ex-
actly opposite effect upon the colloids of the Arbacia egg.
Instead of transforming a solid mass to a more fluid state, what
they really do is exert a solidifying or rather a coagulating effect

1 Loc. cit., p. 213.

STUDIES IN ARTIFICIAL PARTHENOGENESIS.

165

upon the egg colloids. The actual evidence in support of this
statement will be given at a later point (p. 196).

In the first paper of this series, after showing that all sub-
stances which produce membrane elevation cause a lowering
of surface tension, I proposed a theory of membrane-elevation
based on a simple consideration of the forces in equilibrium
at the vitelline membrane. At the time I was not yet aware
of the semipermeable nature of this membrane, and hence in
discussing the various forces acting upon it, I did not consider
osmotic forces as being involved. The theory therefore re-
quires slight modification in the light of this new fact. Let

FIG. i.

us suppose in Fig. I that the shaded area is the vitelline mem-
brane. The arrows represent the forces acting upon it. Arrow
A directed outward indicates the outward force due to the
osmotic pressure of the dissolved substances within the membrane.
Were gels present in the egg we should also have to add the
swelling pressure of these. Directed inward are two arrows,
arrow B is the force due to the osmotic pressure of the salts
outside the membrane, arrow C is the radial component of the
surface tension of the membrane. As the membrane is slightly
thicker than twice the range of molecular action, it is really
composed of two surfaces of surface tension, i. e., an outer
surface of contact with the sea-water and an inner surface of

166 L. V. HEILBRUNN.

contact with the egg contents. Both surfaces are solid-liquid
surfaces, and as such no doubt possess a high surface tension.1
The arrow C represents the sum of the radial components of
the tension of both inner and outer surfaces of the membrane.
Suppose now that a substance which lowers the surface tension
of water, is added to some eggs in sea-water. By what is known
as the Gibbs-Thomson Law it will tend to accumulate at the
surfaces of the vitelline membrane. This will result in a lowering
of the surface tension of the membrane.2 As a result equili-
brium no longer exists, the force pushing outward is now stronger
and the membrane is lifted away from the egg surface. Most
of the egg proteins do not follow the membrane as it is lifted
away, but remain in their original position. This is due to the
fact that they diffuse much less readily than the salts, which
can be thought of as pushing out the membrane. Around the
mass of egg proteins a membrane is rapidly formed. As prev-
iously pointed out, this "hyaline layer" is to be regarded as
similar to the precipitation membranes formed on egg fragments.
As a result of elevation, the vitelline membrane is no longer

1 The surface tension of a solid-liquid surface has never been accurately deter-
mined, but there are various reasons for considering it the seat of an exceptionally
high tension. For the similar case of a solid-gaseous surface Freundlich ("Kapil-
larchemie," p. 90) following Quincke states that "die Oberflachenspannung fliissig-
gasformig steigt allgemein mit sinkender Temperatur. Wenn nun die Flussigkeit
stetig in einen amorph-festen Korper — eine Flussigkeit mit sehr groszer innerer
Reibung— iibergeht, musz man auch annehmen, das die Oberflachenspannung
bestehen bleibt, ja das sie zunehmend groszere Werte erhalt wenn sie sich auch
wegen der groszen Zahigkeit nicht auszern kann." With the aid of a formula
derived by Wilh. Ostwald, Freundlich calculates the tension of solid-liquid
surface BaSC>4- water to be several thousand dynes per cm., an enormous value.
In his wonderful treatment of capillarity Gibbs devotes a long section to the dis-
cussion of solid-liquid surfaces (Gibbs, Collected Papers, I., 314-331), and derives
various equations for them. In discussing the surfaces of the membranes, I have
for the sake of simplicity not considered the presence of the jelly of chorion, which
surrounds the egg. This is so diffuse a gel that it no doubt [has little if any effect.
In fact, after its removal (by shaking), the eggs behave just as they did before.

2 Gibbs, loc. cit., p. 274: "Now the potential of a substance which forms a very
small part of a homogeneous mass certainly increases, and probably very rapidly,
as the proportion of that component is increased. (See (171) and (217).) The
pressure, temperature, and the other potentials, will not be sensibly affected (see
(98)). But the effect on the tension of this increase on the potential will be pro-
portional to the surface-density, and will be to diminish the tension when the
surface-density is positive (see (508))." The numbers refer to equations. When
a substance accumulates at a surface, its surface density is by definition positive.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 167

in contact with the egg cytoplasm. It is, accordingly, now ex-
posed on both of its faces to a solution which has the power of
coagulating it. It might be expected that the process of gelatini-
zation or coagulation be carried a step farther, and that the
membrane become more rigid. This increased coagulation,
if I may so speak of the process,1 results, as is typical for such
cases, in an increase of refractive index, and the optical image
of the membrane becomes more clearly defined. It may also
result in an increased permeability of the membrane, for Quincke
('77) claims that precipitation membranes lose their semiper-
meable properties on becoming rigid.

On the basis of the above theory, it is evident that when the
surface tension of the membrane drops below a certain limiting
value, elevation -occurs. The determination of this value is
at present impossible. It might be thought that the determi-
nation of the surface tension of the various membrane-elevating
solutions at their surface of contact with air, would throw some
light upon the matter.2 But the lowering effect of a substance
on a solid-liquid (or liquid-liquid) surface tension can not be
measured by the effect the same substace produces on a gas-
liquid surface. The chief reasons for this are: (i) Adsorption
of added substances plays a very decided role in the case of
the solid-liquid surface, different solids of course showing dif-
ferent degrees of selective adsorption, (2) A solid-liquid surface
is not subject to the action of evaporation, as is a gas-liquid
surface. This is of especial importance if the substance which
lowers surface tension is volatile. In discussing films like
those of bubbles, Gibbs3 says: "But when a component which
greatly diminishes the tension of the film although forming but
a small fraction of its mass (therefore existing in excess at the
surface), is volatile, the effect of evaporation and condensation
may be considerable, even when the mean value of the potential
for that component is the same in the film as in the surrounding
atmosphere."4 Thus chloroform in water is quite effective

1 It might also be designated as "loss of water."

2 Czapek ('n) has thus endeavored to draw conclusions as to the surface tension
of the plasma membrane of plant cells by measurements of air-liquid surface tensions.

3 L. c., p. 310.

4 A condition which would at least be approximated if the film were in equilibrium
with its vapor in an enclosed chamber.

1 68 L. V. HEILBRUXX.

for membrane elevation, but such a solution of chloroform has
a surface tension against air just slightly below that of pure
water.1 The great volatility of chloroform explains its inability
to lower markedly the tension at the water-air surface; it is
unable to accumulate in the surface film. But it can readily
accumulate at a solid-liquid surface and here it no doubt pro-
duces a marked lowering of surface tension. It follows that,
although by a measurement of the air-water surface tension we
can determine qualitatively whether a substance lowers the
surface tension of the membrane or not, any truly quantitative
measurements are impossible.

In order to show qualitatively that a substance lowers the
surface tension of the membrane, it is only necessary to make
certain that its solution in water has a lower surface tension
than the pure solvent. The theoretical basis for this assertion
has already been pointed out. There is also another case to
be considered. Some substances exert a liquefying effect
upon the vitelline membrane. The liquefaction of a gel no
doubt results in a considerable lowering of its surface tension.2
Ordinarily, however, the liquefying effect is a slow one, and
in such cases, membrane elevation does not occur. For if the
surface tension of the membrane is lowered only very slowly,
the egg proteins have time to follow the membrane as it is pulled
outwards. The result is an increase in egg diameter; the egg
cytolyzes. Thus when the surface tension of the vitelline
membrane is slowly lowered, cytolysis follows directly.

On the other hand, when it is rapidly lowered membrane
elevation takes place first and cytolysis follows after a short
interval. After the vitelline membrane is elevated, it loses
its semipermeability and the hyaline layer takes its place as
the plasma-membrane of the egg cell (cf. p. 159). The same
forces act on this membrane as were found to act on the vitelline
membrane prior to its elevation. Hence when its surface tension
is lowered, since the osmotic pressure within the egg is no longer
completely counterbalanced, the hyaline layer tends to become
pushed outward. But the hyaline layer appears to be more
closely adherent to the egg proteins3 so that they are pulled out

1 Czapek ('n), p. 35.

2 Cf. note p. 166.

3 Probably because of a slight cortical coagulation.

STUDIES IN ARTIFICIAL PARTHKM >< ,|..\ KM-,. 169

with it and the entire egg increases in diameter. According
to this view cytolysis (when not directly an osmotic phenomenon)
is in every case due to a lowering of the surface tension of the
plasma membrane. This produces an influx of water into the
cell.

Various other explanations of cytolysis have been attempted. It has been
regarded as due to the swelling of the egg proteins. This is an incorrect assump-
tion, as I have already pointed out that the egg proteins, instead of becoming
swollen, are coagulated in the cytolyzed egg (see pp. 195-198). It might also be
considered as the result of a change in the plasma-membrane, of such a sort that
this membrane would become readily permeable for salts though still impermeable
for colloids. The unopposed osmotic pressure of the egg colloids would then push
out the plasma membrane and the egg would increase its diameter. But such an
explanation can not hold generally, for membrane-swelling although it produces
cytolysis, does not render the plasma-membrane readily permeable to salts. In-
deed in hypertonic salt solutions, eggs with swollen membranes shrink even more
than those with unswollen membranes (see pp. 155-158). This is sufficient proof
that the swollen membrane retains its semipermeability.

In an earlier paper it was pointed out that all substances
which had been described as producing membrane elevation
(as well as some other newly discovered ones) did actually lower
surface tension in aqueous solution. In a few substances, how-
evei, this lowering could apparently only be a very slight one.
These cases have been reinvestigated, and they have been found
to involve membrane swelling rather than membrane elevation.

When membrane swelling is rapid, it is possible that the lowered
surface tension so produced should be followed by membrane
elevation if the eggs are immediately returned to sea-water.
One such case was indeed observed, but the observation was
made at the close of the season, and further study of this point
is necessary.

D. Cortical Action of Heat and of Various Chemicals.

Heat. — Although heat has been described as producing arti-
ficial parthenogenesis in the sea-urchin egg (McClendon, !io),
I know of no description of membrane elevation as a result of
heat treatment. In several experiments, I was unsuccessful
in producing membrane elevation in this fashion. In experiment
a, eggs were dropped into sea-water which had been heated to
36.5 degrees, and were exposed I, 2, 3, 4 minutes. In experi-
ment b, the eggs were dropped into sea-wrater heated to 35 de-

17° L. V. HEILBRUXX.

grees, and were exposed 2, 4, 6, 8 minutes. In experiment c,
eggs were dropped into sea-water heated to 37.5 degrees and
kept at this temperature on a waterbath; the eggs were ex-
posed 3^ minutes. In no case could any membrane elevation
be observed, but the vitelline membrane did appear to be slightly
swollen after the heat treatment.

Alkalis. — If alkali is added to sea-water, the resultant pre-
cipitation of magnesium salts tend to convert the sea-water
into a solution containing little else than sodium salts. No
doubt this fact alone is of significance. It is probable, however,
that there is a more specific action of the alkali. As was shown
in a previous discussion (p. 151), alkali apparently is able to
liquefy the vitelline membrane. It is doubtful if any observer
has actually obtained membrane elevation with NaOH or KOH.
Certainly it does not occur when the eggs are left in the alkaline
solution.

Butyric Acid. — This famous method is one of the best for
artificial membrane elevation. Loeb used it originally on
Strongylocentrotus eggs, employing a concentration of 50 c.c.
sea- water plus 2.8 c.c. N/io butyric acid and exposing the eggs
1^-3 minutes. In his recent book, he applies the method to
Arbacia eggs reducing the concentration of acid slightly (50 c.c.
sea-water plus 2.0 c.c. n/io butyric), but retaining the same
time of exposure. He says that the eggs do not form a conspic-
uous fertilization membrane, but only a "fine gelatinous layer
which was not easily visible." This is evidently a membrane
swelling and not an elevated membrane. For the benefit of
workers in the field, I mention a slight modification in the method,
by which true elevated membranes may be obtained for Arbacia
eggs. Instead of 1^-3 minute exposures, a \ minute exposure
was found very effective, and over ninety per cent, of elevated
membrane could regularly be obtained if the eggs were well
washed and in good condition. The concentration used was
the original one of Loeb's, 50 c.c. sea-water plus 2.8 c.c. N/io
butyric; slight variations are however immaterial. Although
eggs exposed I minute often showed membrane elevation, no
elevation was found to occur on longer exposure. Instead the
membrane in every case appeared swollen, the amount of swelling
increasing with the length of exposure.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. IJI

Inorganic Salts. — That isotonic solutions of various salts
cause membrane "formation" was discovered by R. Lillie ('10).
The effectiveness of the various salts was found to correspond
with their order in the lyotropic series, i. e., those salts which
are most effective in producing swelling of protein gels (in
alkaline solution) were also most effective in inducing membrane
"formation." The reason for this correspondence becomes
apparent, when we consider that the membrane "formation"
produced by isotonic salt solutions is nothing other than a mem-
brane swelling. The membranes resulting from salt treatment
are sticky, they do not collapse in the presence of a colloid (cf.
p. i 60).

The concentration of a salt is an important factor in deter-
mining whether swelling shall take place, and how pronounced
the swelling shall be. Let us consider as an example the case
of NaCl. Seven solutions of NaCl were prepared by adding
distilled water to a 2.5 normal solution.

Thus Solution A contained 50 c.c. 2\ M NaCl plus o c.c. distilled water.

B 40 10

C 30 20

D 20 30

E 10 40

F 5 45

G 2\ 47-5

Eggs were placed in all these solutions at 3:30 P.M. By
3:39 eggs in A showed a membrane swollen to 2.51* in diameter.
By 4:50 it had reached approximately $/j..

As for the eggs placed in B, the membrane swelled at least
as rapidly as in the first case.

Membranes on eggs in C and D also swelled rapidly, but in
no case did eggs in E, F, G show any membrane swelling. Thus
a certain concentration is necessary for swelling to occur.

There appears to be at least two factors involved in this effect
of concentration. In the first place, the swelling action of the
salt no doubt increases with the concentration. But it is prob-
able that the hypertonicity of the solution may also play a
role.1

1 The explanation of such an effect might be as follows : When water is extracted
from the cell, there is a tendency for vacuum production and consequent negative
pressure (release of pressure) on the particles of the membrane. But we know

L. V. HEILBRUNN.

Sensitization with Strontium Chloride. — The process of sen-
sitization was originally used by Loeb, later by Robertson, to
induce the "formation" of a membrane in eggs exposed either
to blood sera or to various tissue extracts, all dissoved in what
was essentially a NaCl solution isotonic with sea-water. The
method employed consists either of adding 3/8 M SrCl2 directly
to the serum, or of placing the eggs first into the 3/8 M SrCl2
for several minutes, then into the serum. In the first paper
of this series I urged that SrClo caused a precipitation of sul-
phates and that much of the effect of sensitization was no
doubt due to this action. In the summer of 1913, I was able
to prove the truth of this statement. On July 31, 10 c.c. of 3/8
M SrCl2 was added to 90 c.c. of sea-water. The voluminous
precipitate of SrSO4 was allowed to settle and filtered off, but
the precipitate still continued to settle from the filtrate. When
eggs were placed into this filtrate membrane swelling occurred,
and this was followed by cytolysis. In another experiment,
the eggs were placed first into 3/8 M SrCl2 and then into a
0.55 M NaCl solution. Used alone, the NaCl solution did
not cause membrane swelling when two drops of a thick egg
suspension were placed into 75 c.c. of it. But when several
drops of egg suspension were placed into 3/8 M SrCl2, and then
after five minutes, 2 drops of liquid containing eggs were taken
from the SrCl2 solution and placed into the NaCl solution,
membrane swelling could be observed to take place in the eggs
so transferred. The vitelline membrane could be seen slowly
to increase in thickness, so that after about half an hour a good
proportion of the eggs were surrounded by a transparent outer
layer which bore a close resemblance to an elevated membrane,
except for the fact that the inner boundary of such a membrane
was absent. The membranes produced as a result of the "sen-
sitization" process did not collapse in the presence of a colloid.
They were also sticky, and as a result the eggs tended to aggluti-
nate. Such an agglutination of eggs with swollen membranes '
has no doubt led Robertson ('12) to the view that "fertilization
and agglutination are similar phenomena."

from Le Chatelier's theorem that whenever the pressure on a system in equilibrium
is diminished, a change or reaction ensues which is accompanied by increase of
volume. Hence the swelling.

STUDIES IX ARTIFICIAL PARTHENOGENESIS. 173

These facts enable us to interpret many of the results of Loeb
and Robertson on a truly chemical basis, and without the aid
of a hypothetical lysin. But there is also another factor to be
considered in the analysis of the action of sera, and tissue ex-
tracts. It is the fact that many proteins are probably able
to effect a swelling of the vitelline membrane. Such an action
is at least true for egg albumen. The eggs of 2 females were
washed twice and pipetted into 100 c.c. of I per cent, egg al-
bumen in sea-water. Before ten minutes had elapsed the mem-
branes on most but not all of the eggs had swollen considerably.
The swelling action of albumen seems to vary considerably,
sometimes being almost negligible. It is, I think, to be regarded
as a phenomenon akin to peptonization, the albumen playing
the part of a protective colloid.

KCN. — The action of potassium cyanide is rather difficult to
explain on basis of the surface-tension theory. Although KCN
alone, in the dilutions used, never caused elevation of the mem-
brane, if the solution was made hypertonic the membrane did
separate from the egg. Thus on July 31, 1914, membrane ele-
vation did not occur in 25 c.c. of sea-water plus I c.c. of 1/20 per
cent. KCN, but in a similar solution plus 4 c.c. 2.5 M NaCl, the
vitelline membrane in practically every case became lifted away
from the egg surface. How can this observation, which was
repeated several times, be interpreted on the basis of the surface-
tension theory? As is well known, KCN hydrolizes readily,
so that KOH and HCN are always present in a solution of the
cyanide. Probably HCN plays the most important part, for
it is a gas, and hence its surface tension is practically zero.
Thus it no doubt lowers the surface tension of the vitelline mem-
brane (cf. p. 166). But membrane elevation does not result,
because of another effect of the cyanide. In the first paper
of this series, it was pointed out that even if a substance lowered
surface tension, it would not produce membrane elevation if it
increased too greatly the modulus of elasticity of the vitelline
membrane, for in that case the toughened membrane would be
incapable of stretching. There is evidence that KCN actually
does increase the modulus of elasticity. In the next section it will
be shown that the presence of KCN retards membrane swelling,
and such anti-swelling action is a general characteristic of all

174 L- V- HEILBRUNN.

substances which increase the modulus of elasticity.1 As a result
of the stiffening action of KCN, the membrane resists elevation,
for the elevating force is not sufficient to stretch it. But in a
hypertonic solution, as the egg shrinks, no stretching is necessary
to separate the vitelline membrane from the egg, and the mem-
brane is accordingly pulled away from the egg surface.

E. Inhibitors to Membrane Swelling.

It has been found that swelling of the vitelline membrane
may be retarded or even inhibited in the presence of certain
substances. Up to the present only two such inhibiting sub-
stances have been discovered. Of these, KCN inhibits the
swelling of the membrane by salts, sea-urchin blood on the other
hand inhibits acid swelling. But the inhibitor of salt action
has no such effect on acid swelling, and sea-urchin blood in-
stead of inhibiting swelling produced by salts, actually seems
to favor it. These results, paradoxical as they appear at first
sight, are really in direct line with recent findings in the field of
colloid chemistry. There it has been shown that the salt and
the acid swellings of gelatine must be essentially different
processes, for the very salts which, by themselves, favor or even
cause swelling, retard the swelling effect of an acid.

Although membrane swelling usually occurs in "NaCl hyper-
tonic sea-water,"2 in the presence of traces of KCN no such
swelling will occur. This inhibiting effect of KCN has as yet
not been found to be shared by any other substance, but so far
only a few experiments in this direction have been made. Some
of the experiments made during the summers of 1913 and 1914
are recorded below.

August 29, 1913. To 200 c.c. of sea-water was added 32 c.c.
of 2.5 M NaCl and the resultant "NaCl hypertonic sea-water"
was divided into four 50 c.c. portions, A, B, C, D. To A was
added 0.5 c.c. I per cent. KCN, to B 0.5 c.c. N/io NaOH, to
C, 0.5 c.c. ether, and nothing was added to D, which several
as a control.

Eggs were then placed in A, B, C, D, and it was found that
although membrane swelling very evidently took place in D,

1 Freundlich, "Kapillarchemie." p. 512.

2 As previously stated, I use this term to indicate a solution of 50 parts (by
volume) of sea-water plus 8 parts of z\ M NaCl.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. IJ5

no such swelling occurred in A. The effect was not due to the
alkaline reaction of KCN, for the swelling of egg membranes in
B showed that NaOH had no retarding effect on the process.
Likewise the ether present in C appeared to have no inhibiting
effect.

In other experiments KCN behaved similarly. In general
it was found advisable to first place the eggs into sea-water
which contained KCN and to add the 2.5 M NaCl later. Thus
on August 29, 1914, it was found that some membrane swelling
occurred when eggs were dropped into hypertonic sea-water to
which KCN had already been added, but that this was entirely
prevented when the eggs were first exposed to the KCN solution
in sea-water, the 2.5 M NaCl being added later.

August 2Q, 1914. Fingerbowls A, C, D, were used. Into
A were placed 50 c.c. sea-water plus 8 drops of egg suspension,
and to this were added 8 c.c. of 2.5 M NaCl. Fingerbowl
C contained 50 c.c. sea-water plus i c.c. 1/5 per cent. KCN,
to this were added at 10:30^ A.M. 8 c.c. 2.5 M NaCl, and at
10:31 A.M. 8 drops of egg suspension were dropped into the
cyanide containing "NaCl hypertonic sea-water." Fingerbowl
D contained 49 c.c. of sea-water plus i c.c. 1/5 per cent. KCN.
Several drops of egg suspension were added to this solution
of cyanide in sea-water, at 10:40 A.M. At 11:28 A.M. (48
minutes later), 8 c.c. of 2.5 M NaCl were added to D.

On microscopical examination, it was found that pronounced
membrane swelling had taken place in A (in the absence of
KCN), swelling also occurred in C, but did not appear to be
pronounced as that occurring in A. No membrane swelling
at all could be observed in D, in which the eggs had been treated
first with KCN before the concentrated NaCl solution had been
added.

The above experiment indicates that the action of KCN in
inhibiting membrane swelling produced by NaCl, is not the
result of a reaction between the cyanide and the salt, but is
due to an effect of the cyanide on the membrane. For if only
a reaction between salt and cyanide was involved, there could
be no advantage in first subjecting the eggs to the action of the
cyanide alone.

Although KCN inhibits the membrane swelling effect of

176 L. V. HEILBRUNN.

NaCl, it does not appear to have the slightest retarding effect
on membrane swelling when this is produced by an acid.

August 28, 1914. Fingerbowl A contained 50 c.c. sea-water
plus 3 c.c. TV/io butyric acid. Fingerbowl B contained 49 c.c.
sea-water plus 3 c.c. N/io butyric acid plus I c.c. i/io per cent.
KCN. When eggs were added to A and B, membrane swelling
occurred in both.

September j, 1914. Stender dish A contained some eggs
in 25 c.c. of sea-water. To this was added I c.c. 1/5 per cent.
KCN at 10:52 A.M. Then 2 c.c. N/io butyric acid were added
29 minutes later (at 11:21).

Stender dish B contained eggs in 25 c.c. of sea-water, 2 c.c.
of N/io butyric acid were added at n :22 A.M.

In both A and B, membrane swelling occurred and as a result
the eggs in both cases stuck to each other and to the bottom
of the dish. No difference could be observed between the
two sets of eggs, and apparently the KCN has no effect on
acid swelling of the membrane.

KCN is thus capable of inhibiting membrane swelling by
NaCl, but it has apparently no effect when the swelling is
produced by butyric acid. On the other hand, sea-urchin blood
was found to retard or inhibit acid swelling.

June 22, 1914. A solution of butyric acid in sea-water was
prepared by adding to 50 c.c. of sea-water, 2.5 c.c. of N/io
butyric acid. Approximately 5 c.c. of the resulting solution
were placed in each of two Syracuse watch-crystals (A and B}.
To watch-crystal A were added 3 c.c. of filtered sea-urchin blood.
(The blood was filtered after it had been allowed to "clot" by
standing.) To watch-crystal B, 3 c.c. of sea-water were added.
In B, membrane swelling and agglutination occurred, in A very
little, if any, membrane swelling occurred, and there was no
agglutination. The jelly was dissolved away from the eggs in
A, but although the eggs were thus able to come into close con-
tact, they would separate again, showing that they were not
sticky, and that no membrane swelling had occurred. After
a few hours, eggs in B had completely lost their color and ap-
peared white to the naked eye, those in A appeared normal.

July 7, 1914. The above experiment was repeated. In this
case an acid solution was made up by adding 5 c.c. N/io butyric

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 177

acid to 50 c.c. of sea- water. Of this solution, 5 c.c. \vas put
into each of t\vo Syracuse watch-crystals A and B.

To A were added 5 c.c. of sea- water and five drops of egg
suspension.

To B were added 5 c.c. of filtered blood (from several cf s and
9 s) and then 5 drops of egg suspension.

The result of the experiment was that in A the vitelline mem-
branes swelled in practically every case. A count of a hundred
eggs gave ninety-nine eggs with swollen membranes and the
single exception was doubtful. On the contrary, there was
practically no membrane swelling in B, which contained blood
in addition to the acid. A count gave, of a hundred eggs ob-
served, only three with swollen membranes.

The inhibiting effect of blood upon acid swelling, unlike the
effect of cyanide on salt swelling, may perhaps be the result
of a direct action of the blood upon the acid. This is barely
suggested by the fact that N/io HC1 produces a flocculent white
precipitate when added to filtered sea-urchin blood. However
in the above recorded experiments with butyric acid, no such
precipitation could be observed.

Instead of exhibiting a retarding effect upon membrane
swelling by salts, sea-urchin blood seemed to favor the process.
This favorable effect was much more pronounced in some cases
than in others, and in several experiments it was not readily
observed. On July 28, 1914, membrane swelling was found to
be much more rapid and pronounced in a solution of 20 c.c.
sea-water plus 5 c.c. filtered blood (from 9 s) plus 4 c.c. 2.5
M NaCl, than in a similar solution without blood, i. e., 25 c.c.
sea- water plus 4 c.c. 2.5 M NaCl. It sometimes happens, and
this appeared to be more frequent in 1914 than in 1913, that
membrane swelling does not occur in "NaCl hyper tonic sea-
water." In such cases, it was found possible in several instances
to produce a membrane swelling by the addition of blood. Thus
on July 30, 1914, although there was no swelling in "XaCl
hypertonic sea- water," when eggs of the same lot were placed
into 5 c.c. blood plus 20 c.c. sea- water plus 4 c.c. 2.5 M NaCl
membrane swelling did occur.

It might be reasoned that this action of blood is analogous
to the cytolytic effect of sera foreign to the individual. But

1/8 L. V. HE1LBRUNN.

in one case at least, blood of the same individual from which
the eggs had been taken, was found to exert an accelerating
effect upon membrane swelling in "NaCl hypertonic sea-water."
If it be true that blood retards- acid swelling and favors salt
swelling, this fact can be used in cases of doubt, to determine
if a given type of membrane swelling is the result of the action
of a salt or an acid.

F. Cortical Changes at Fertilization.

The central object of studies in artificial parthenogensis is
to find an explanation of the processes occurring in normal ferti-
lization. The fact that artificial membrane elevation is apparent-
ly always the result of a lowered tension of the vitelline mem-
brane has of course led to the view that the spermatozoon also
produces a lowered surface tension. There are two conceivable
ways in which this could happen. In the first place the sperm
might carry a substance which lowers surface tension directly.
This is improbable, in view of the fact that it has not been pos-
sible to extract from sperm a membrane-elevating substance.
It is more logical to suppose that the very act of penetrating
on the part of the sperm lowers the tension.

If the tension of a stretched thread be lowered at one point,
instantaneously the tension throughout the thread is lowered.
Similarly if the tension of a spherical stretched film or membrane
be lowered at one point, there will be a lowering of tension in
every point of the film, for in order that equilibrium be estab-
lished, the tension in every part of the spherical film must be
equal. This equalization of tension is probably a rapid process,
especially when not merely a point, but an appreciable area of
the surface has its tension lowered.1

Thus the penetration of the sperm almost immediately pro-
duces a lowered tension in all parts of the vitelline membrane.
In the sea-urchin egg, the sperm can not bore its way through
the membrane mechanically, as it is not provided with a per-
feratorium. It is therefore probable that the sperm has a
solvent action on the membrane upon coming in contact with

1 However, in cases where a very thin film surrounds a large spherical mass as in
air-bubbles the attainment of equilibrium between the parts of the film is much
slower than the attainment of equilibrium between the film and the contiguous
fluids. Cf. Gibbs, 1. c., p. 300.

STUDIES IN ARTIFICIAL PARTHENOGENESI-. 179

it. The partial liquefaction or swelling of the vitelline mem-
brane at the point of sperm entrance can be conceived of as
serving two functions: (i) it enables the sperm to enter, (2) it
lowers the surface tension of the membrane and thus produces
membrane elevation.1

That the sperm actually does produce a substance capable
of causing membrane swelling, can be demonstrated. It is not
possible to observe the swelling produced by a single sperm.
But if the eggs are placed into very concentrated sperm suspen-
sions, the vitelline membrane can be seen to swell all around
the egg. Such concentrated suspensions are obtained by al-
lowing the sea-urchins to shed their sperm. As is well known,
the shedding reaction is aroused when the oral part of the shell
is cut away. If the "dry" sperm be diluted only very slightly,
an enormous sperm concentration can be obtained. When
eggs are mixed with sperm suspensions of such high concentra-
tion, each egg immediately becomes surrounded by a halo of
wriggling sperm. Soon the vitelline membrane can be seen
slowly to increase in thickness, it swells until it may reach a
thickness of about 3 microns. That normal membrane elevation
has not taken place can be shown by the fact that the swollen
membranes thus produced do not collapse when the eggs are
placed in a i per cent, or a 2 per cent, albumen solution (in sea-
water). In a concentrated' sperm suspension, each point on
the vitelline membrane is a point of attack on the part of the
spermatozoa, and the entire membrane becomes swollen.2

The concentration of sperm necessary to produce a complete

1 It might be thought that puncture of the vitelline membrane, e. g., by a needle,
should produce elevation. But this is not necessarily the case, for the hole pro-
duced by a mechanical puncture of the membrane, if not immediately closed,
would involve a loss of its semipermeable properties, and these on the basis of the
theory (see pp. 165-166) are necessary for membrane elevation. A deep prick
would also produce coagulation of the underlying cytoplasm, which would tend to
prevent elevation.

2 It might be asked why elevation of the membrane does not follow swelling
produced by concentrated sperm suspensions, since this no doubt results in a rapid
lowering of surface tension. The answer is clear. In a previous paper (Heilbrunn,
'13) it was pointed out that membrane elevation never occurred when the egg or
its cortex was coagulated. Now in concentrated sperm suspensions, it can be
shown that a profound coagulation does take place. This was demonstrated by
a method which has been developed for revealing the presence of coagulation
within the egg. For details of this method see p. 192.

ISO L. V. HEILBRUNN.

swelling of the membrane varies with the season. In the height
of the season only quite concentrated sperm suspensions produce
the phenomenon. But towards the end of August, as the season
begins to wane, more and more dilute suspensions become ef-
fective, until at the very close of the season, it is difficult to
secure true membrane elevation at all. Probably the fluid
emitted with the sperm is then charged with the substance which
causes the swelling.

In the preceding section it was shown that sea-urchin blood
retards acid membrane swelling, but favors salt membrane swel-
ling. This fact makes it possible to determine the general
nature of the membrane swelling produced by sperm. One
has only to observe the action of concentrated sperm suspensions
in the presence of blood. If blood favors the swelling, this can
be taken as evidence that the sperm action is similar to that of
salts, if on the other hand it retards swelling, an acid is probably
responsible. On July 7, 1914, 5 c.c. of sea-water plus 9 drops
of egg suspension plus 5 drops of "dry" sperm were mixed in
Syracuse watch-crystal A. Watch-crystal B contained 4 c.c.
of blood (filtered from cfs and 9s) plus 9 drops of egg sus-
pension plus approximately the same amount of sperm as did
A. Both watch-crystals were shaken slightly to insure mixing.
In A, the membranes swelled gradually; after 43 minutes they
measured approximately 2p. In B, on the contrary, no membrane
swelling could be observed.

It might be objected that in the above experiment the blood
had some effect which prevented intimate contact of sperm
with egg. This objection is obviated by the following experi-
ment. It was performed at the very close of the season, at a
time when as previously pointed out, sperm suspensions have a
much greater tendency to produce swelling. On September I,
1914, considerable difficulty was experienced in procuring sperm.
Ten males were cut open and allowed to shed (several others had
been rejected as being totally incapable). Of these ten, only
two shed any sperm at all, and this was rather watery. Pre-
liminary experiments showed that when 3 drops of this watery
"dry" sperm were added to about 10 c.c. of sea-water and about
8 drops of the resulting suspension were again diluted with 10
c.c. of sea-water, a suspension was obtained which caused mem-

STUDIES IN ARTIFICIAL PARTHENOGENESIS. l8l

brane swelling when 5 drops were added to 25 c.c. of sea-water,
but only a moderate per cent, of fertilization when i drop was
added to 25 c.c. of sea-water containing eggs. In the main
experiment 3 drops of watery "dry" sperm were added to approx-
imately 10 c.c. of sea-water, and the resulting suspension was
the one used. Two Syracuse watch-crystals were employed.
\Yntch-crystal A contained 4 c.c. of sea-water plus I c.c. of
filtered blood (from 9s). Watch-crystal B contained 5 c.c.
of sea-water and no blood. 8 drops of a dilute egg suspension
were then added to A and to B (at n 115^ A.M.) and two min-
utes later, 2 drops of the sterm suspension just mentioned,
were added. At II :2O, eggs in B (i. e., in the absence of blood)
all had membranes widely swollen all around. At 11:22, the
eggs in A were examined. Most of the eggs showed not a trace
of cortical change, but some showed membrane elevation. Short-
ly after, a count was made of eggs in A. It was found that
23 showed no cortical change, I was doubtful, it may have had
a swollen membrane, 9 evidently possessed elevated membranes.

Thus the presence of blood prevents membrane swelling,
and as one result of this prevention of excessive swelling, mem-
brane elevation is possible, although without blood it could
not have been produced. In this experiment, the anti-swelling
effect of blood towards sperm suspension is conclusively demon-
strated, for the blood evidently does not prevent access of sperm.1

The fact that the presence of blood inhibits membrane swelling
in concentrated sperm suspensions, indicates that the swelling
is produced by an acid. As is well known, all spermatozoa are
abundantly provided with nucleic acid, and 'it is very probable
that a nucleic acid or its derivative is responsible for the swelling.

I have tried a number of times to watch the process of normal
membrane elevation under high power, but with only scant
success. The presence of a coverslip produces difficulties. If a
drop of egg suspension is placed on each of two slides, and one

1 The further history of the eggs in this experiment is very interesting. On
September 2, there are considerably fewer larvae in A (where blood was present)
than in B. But whereas all of the larvae in A have well-marked arms, none of the
larvae in B possess even the suggestion of arms. On September 3, the larvae in B
are still perfectly armless, whereas those in A have the usual long arms of the
pluteus stage

l82 L. V. HEILBRUNN.

drop is then covered by a coverslip with a drop of sperm sus-
pension on its lower surface, and to the other drop is added a
drop of the same sperm suspension, but no coverslip, a marked
difference between the two preparations can be noted. In
the absence of a coverslip, a much greater per cent, of eggs
undergo membrane elevation. When the drop of sperm sus-
pension is added to the eggs after the coverslip has been placed
in position, practically no membrane elevation occurs. In
these experiments, the coverslip was always supported by strips
of paper or thin glass tubes, so that there could be no question
of compressing the eggs.

The effect of the coverslip is in part due to the action of the
glass (or of substances diffusing out of it)1 on the spermatozoa.
The sperm apparently congregate at the surface of the coverslip.
But this is not believed to be the only effect, and some evidence
that I possess, although not absolutely unimpeachable, tends to
show that the pressure of the coverslip is also partly responsible
for preventing elevation. However, further experiments on
this point are necessary; I merely bring up the matter here in
order to emphasize the difficulties in the way of direct obser-
vation. Fol ('79) speaks of the great difficulty in observing
fertilization in the sea-urchin egg. Pictet ('91) found no such
difficulty, but I think that the cortical effect that he describes
was not membrane elevation, but membrane swelling, which
is not retarded by the presence of the coverslip.2

My observations, though admittedly incomplete, tend to show
that the membrane is elevated from all sides of the egg at the
same moment. It is possible that elevation does start at the
point of sperm entrance as Ries ('09) for example claims, but
if this part of the membrane does show any priority, it is only
an exceedingly brief one.

On the basis of Loeb's view that the sperm contains a lysin
which produces membrane "formation" directly, we would have
to suppose that this lysin diffuses around the egg surface in
incredibly fast time. The morphology of sperm membrane
elevation offers a severe obstacle to most theories which attempt

1 Possibly dissolved out by the alkali of sea-water.

2 Both Pictet and Fol subjected the eggs to slight compression.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 183

to account for the process, but it is readily understandable
on the basis of the surface tension theory. The membrane
swelling which occurs at the point of sperm entrance, causes
a lowering of surface tension not only in the immediate vicinity,
but everywhere on the egg surface. Hence the vitelline mem-
brane is lifted from all parts of the egg at practically the same
moment.

G. The Significance of Cortical Change.

Although it has never been proven that cortical change is
absolutely essential before development can take place, it is
generally admitted that membrane "formation" must precede
any normal development. We- have seen that cortical change
or membrane "formation" may involve either of two quite
different processes. What then is the fundamental significance
of cortical change, why is it necessary for normal development?
I think that at least part of the answer is forthcoming. If an
egg be watched under a micrometer scale, it will be noticed that
shortly before the first cleavage the egg rapidly lengthens in the
direction of its polar axes. In three minutes, an egg was in one
case observed to increase its polar axis by 6.5/z, in another case
by S/JL. (This fact is by no means new, all the pictures of cleavage
show an increase in polar axis.) If the egg is surrounded by a
stiff membrane, such a rapid change of form could scarcely be
possible. But cortical change, whether it be membrane swelling
or elevation, always results in the removal of this obstacle.
The vitelline membrane is either rendered soft by swelling, or
it is lifted away from the egg surface and its place taken by the
no doubt less rigid hyaline layer. As a result, rapid changes irt
egg form can occur. Moreover, as is well known, the hyaline
layer is normally pulled in between the first two blastomeres
during the cleavage process. The stiff vitelline membrane
could scarcely act in this way. But either membrane swelling
or membrane elevation would result in the egg being invested
by a membrane which was not too rigid to be pulled in. Thus
at least two processes which play a part in normal development
would be greatly hindered if some kind of cortical change did
not occur.

184 L. V. HEILBRUNN.

Moreover when hypertonic solutions are used as reagents,
the degree of rigidity of the plasma membrane becomes a factor
of importance. As was shown previously (p. 158), exosmosis is
favored by a cortical change which tends to soften the vitelline
membrane.

III. INTERNAL CHANGE. — THE PROBLEM OF SEGMENTATION
A. Theories of Segmentation.

The real problem in the study of artificial parthenogenesis
is not the problem of cortical change, but is rather the analysis
of the factors which produce initiation of development. How
can we define initiation of development? Although it has been
shown that after fertilization there is a quickening of various
energetic processes in the sea-urchin egg, plausible as such a
view no doubt is, no one has ever shown that all the recorded
instances of artificial parthenogenesis involve an increased
metabolism. ^ In practically every case, segmentation or mitosis
has been regarded as the sole criterion of initiation of development.
This is, I believe, wholly justifiable, as the act of segmentation
is in itself a beginning of development.

Of the various theories of artificial parthenogenesis, those
of R. Lillie and of Loeb have in recent years met with the most
favor. The former believes that the initial change is an increase
in permeability of the plasma membrane. This is the direct
cause of (i) mitosis, and (2) increased metabolism. In his
theory of mitosis (R. Lillie 'u), now widely accepted as the
least objectionable of any of the numerous theories of mitosis,
he considers the hypothetical plasma membrane as charged by
an electrical double layer; this charge is neutralized upon an
increase in permeability. Such a neutralization of charge pro-
duces, according to Lillie, a lower potential at the surface as
compared with the interior, and the difference in potential
thus produced brings about a number of internal changes, which
lead finally to the formation of a mitotic spindle. I confess
that I am unable to understand the fundamental basis of the
theory. It appears to me that any drop in the potential at the
surface of a body involves simultaneously an exactly equivalent
drop of the potential at every point in the interior. This seems

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 185

a direct consequence of the textbook definition of potential at a
point, as the amount of work done by the electric field on unit
charge, when it is brought to that point from infinity.

Moreover, even if we could accept some of the extiemely
dubious evidence that has been offered in favor of increased
permeability, no one has ever tried to show that all parthenogene-
tic agents do cause increased permeability. In fact R. Lillie
has at times assumed that hypertonic sea-water and that mag-
nesium salts cause a decrease rather than an increase in per-
meability, and yet hypertonic sea- water, even in the presence of
magnesium salts, produces initiation of development.

Of Loeb's ideas concerning the initiation of development only
a bare outline is possible here. Recently, he is "inclined to be-
lieve that in all cases in which an unfertilized egg has been caused
to develop a typical or atypical membrane had been formed."1
This membrane "formation," which, as we have seen, Loeb
regards as a swelling process, is the important initiative factor.
It causes directly an increase of oxidations, but leaves the egg in
a sickly condition, hence it is necessary to provide a corrective
agent which may be either oxygen-containing hypertonic sea-
water or the absence of oxygen. The theory of the necessity
of the corrective factor need not concern us here, for we are
primarily interested, not in the best method of obtaining arti-
ficial parthenogenesis, but rather in the nature of the physical
or chemical change which causes any initiation of development.
Loeb believes this change to be an increase in oxidations. In
support of this view he has accumulated a mass of data. He
first showed that in his original method of using a hypertonic
solution alone, artificial parthenogenesis could only be produced
in the presence of oxygen. More recently he has measured egg
oxidations and has confirmed Warburg's observation that mem-
brane "formation" produces a great increase in oxidations.

According to Loeb's measurements, any cytolytic change re-
sults in a great increase of egg oxidations.

B. The Action of Hypertonic Sea-water in the Presence of KCN.

Loeb was originally led to adopt the oxidation theory, by
the fact that either the addition of KCN, or the removal of oxy-
1 Loeb, '130, p. 223.

1 86 L. V. HEILBRUNN.

gen by the passage of hydrogen, seemed to prevent the action
of hypertonic solutions upon sea-urchin eggs. He reasoned
from this that the hypertonic solutions produced an increase in
oxidations. In 1913 and 1914, I performed a number of ex-
periments with hypertonic solutions in the presence of KCN,
or in an atmosphere of hydrogen. In this paper I shall, however,
report only on the KCN experiments. The results obtained in
an atmosphere of hydrogen are perhaps more interesting, but
I do not feel as yet that every possible source of error has been
eliminated.

At present it is generally accepted by most physiologists
that KCN suppresses cell oxidations.1 Loeb states that 2 c.c.
1/20 per cent. KCN to 50 c.c. of sea-water suffice for the purpose.2
In my experiments, a much higher concentration of the reagent
was employed.

On August 19, 1913, I added 0.5 c.c. of I per cent. KCN to
every 58 c.c. of "NaCl hypertonic sea-water." (This addition
of only 0.5 c.c. of more concentrated reagent had the advantage
of not incurring as much dilution as the addition of 2 c.c. of
almost pure solvent.) Eggs were exposed to this cyanide-con-
taining "NaCl hypertonic sea- water," as well as to the normal
"NaCl" hypertonic sea-water. The results appear in the fol-
fowing table, in which the numerators of the fractions indicate
the number of eggs observed to segment, the denominators the
total number of eggs counted. Before return to sea-water,
the eggs which had been exposed to KCN were washed twice by
Lyon's test-tube method (Lyon '02), so that all trace of the poison
might be removed.

" NaCl Hypertonic Sea-
Minutes Exposed. "NaCl Hypertonic Sea-water." water" plus KCN.
25 127/209 10/102
30 83/100 9/108
35 62/100 22/247

Of the eggs exposed 25 or 30 minutes to the cyanide-containing
hypertonic solution, none went beyond the 2 (or 3)-celled stage,
but of the 22 eggs which segmented after an exposure of 35
minutes, two became many-celled blastulae and may have gone

1 Personally I do not believe that the evidence in support of this view is con-
clusive.

2 Loeb, '09, p. 55.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 187

farther, one reached a few-celled blastula stage. The control
of unfertilized eggs in sea-water showed practically no seg-
mentation, a count showed three out of more than five thousand.

In the above experiment only about 30-40 per cent, of the
eggs underwent membrane elevation. In other experiments,
in which less concentrated solutions of KCN were employed,
all cortical change was inhibited and the eggs did not segment,
although nuclear division sometimes occurred. In order to
obtain the best results, widely elevated membranes must be
obtained. If the membranes were only lifted a short distance
from the egg in the cyanide-containing hypertonic sea-water,
on a return to normal sea-water the egg expanded until it again
came into contact with the membrane (often it ruptured the
membrane and an exovate was produced). The best results
were obtained in an experiment of September 2, 1914. In this
case, I c.c. of 1/5 per cent. KCN was added to 24.5 c.c. of sea-
water containing eggs, at 11:45 A.M. At 11:45 !/6 A. M.,
4 c.c. of 2.5 M NaCl were added, so that the eggs were then in
"NaCl hypertonic sea-water" plus KCN. They were allowed
to remain in this solution for 33 minutes, during which time
it was noticed that ninety-seven per cent, underwent membrane
elevation. At 12:18 P.M., some eggs in 4 c.c. of the solution
were transferred to about 600 c.c. of sea-water in a tall graduate.
At about 12:45 P. M., most of the sea-water was siphoned off
from the graduate and fresh sea-water added. At 2:30 P.M.,
out of 233 eggs examined 22 were found to have segmented, but
the count was evidently too low, as doubtful cases were rejected.
In some instances, exovates simulated cleavage. Only those
eggs in which a nucleus could be observed in each cell were
counted. At 4:55 P.M., 50 eggs, out of 110 examined, were
found to have segmented, but this count was likewise probably
too low. At 9 P.M., hundreds of motile blastulae could be
observed. (Of 337 eggs noted, 28 were motile larvae, of these
18 swam on the bottom, 10 on the top.)

These experiments indicate that the action of the hypertonic
solutions in initiating development is not due to an effect on
the oxidative processes, for the hypertonic solution has the same
action in the presence of a concentration of KCN, which ac-

188 L. V. HEILBRUNN.

cording to Loeb is considerably above that sufficient to check
oxidations.

Loeb also found that the presence of KCN prevented degener-
ative change in eggs exposed to hypertonic sea-water, and
assumed that the KCN acted by retarding excessive oxidations.
However on the basis of our knowledge concerning the anti-
swelling effect of KCN (cf. p. 174), it is simpler to assume that
the inhibition of membrane swelling is the prime cause in pre-
venting disintegration by the hypertonic solution.

The main support of the oxidation theory no doubt lies in
the actual measurements of oxidations made by Warburg and
Loeb. The method by which these measurements were made
has been criticized in a note in Science (Heilbrunn '15).

C. An Analysis of the Methods of Producing Segmentation in
the Unfertilized Arbacia Egg.

Hitherto in the study of artificial parthenogenesis, the general
tendency has been to find new and different methods, rather
than to find points of resemblance in the various means employed
to produce segmentation in any one egg. The result has been
that too much emphasis has been placed on the diversity and
the unrelated character of the numerous parthenogenetic agents.

As a matter of fact there are in general only two ways of pro-
ducing segmentation in the Arbacia egg, the endosmotic method
and the exosmotic method. Whenever a reagent lowers the
surface tension of the plasma-membrane, endosmosis is the re-
sult. The theoretical reason for this has already been considered
(cf. p. 1 68). All those reagents which induce either true membrane
elevation or membrane swelling are thus included in this cate-
gory. In the first case the reagent produces a lowered surface-
tension directly, in the second case, the swelling of the membrane
results in a lowering of tension. Practically all of the reagents
which have been used in artificial parthenogenesis produce either
the one type of cortical change or the other. As a result, endos-
mosis follows, unless the eggs are in a hypertonic solution. In
the latter case, water is extracted from the cell, and exosmosis
occurs.

An apparent exception is found in the action of cold, which

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 189

has been stated to produce parthenogenesis in the sea-urchin
egg (McClendon '10). Lowered temperatures should result
in an increase, rather than a decrease of surface tension, nor
can they be thought of as causing exosmosis.

In the only experiment tried, I found that instead of producing artificial par-
thenogenesis, lowered temperatures tended to retard the natural parthenogenesis
which is usually manifested by the Arbacia egg. On June 25, the temperature of
the aquarium in which the sea-urchins were kept, was 19.5°, the temperature of the
sea-water as it emerged from the tap was 19°; the room temperature (at 4.45 P.M.)
was 24.5°. At 10.40 A.M., a beaker containing eggs in a small amount of sea-
water was placed in a large beaker, and the space intervening between the two
beakers was filled with cracked ice. The beakers were then placed in the ice-box.
At ii. oo A.M., the temperature of the sea-water surrounding the eggs was 3.5°,
at 11.35 A.M. it was 1.5°, at 2.15 P.M. it was 0.5°, and at 4.15 P.M. it was 1.5°.
Eggs were transferred from cold to normal sea-water after 3! hours (2.15 P.M.),
5 hours (3.40 P.M.), and 6\ hours (5.05 P.M.). We can refer to these three lots
of eggs as lot A, lot B, lot C, respectively.

Lot A when counted at 4.30 P.M. showed one doubtful case of segmentation,
out of 100 eggs observed.

Lot B at 5.30 P.M., out of 1,000 eggs counted, showed 3 eggs apparently cleaving
irregularly, and i doubtful case. At best 4/1000.

Lot C at 8.55 P.M. showed 9/1000 cleavages.

The control of untreated eggs (from the same female) at 4.40 P.M. showed,
of 1,100 eggs counted, 4 eggs cleaving irregularly (i a 7-celled stage) and 5 with
attempted or incomplete cleavages. At 9.05 P.M. however, the control showed
a much higher count. Of 526 eggs counted, 36 showed irregular divisions. Thus
of the eggs counted at about 9.00 o'clock, those which had been exposed to cold for
6| hours showed less than i per cent, of segmenting eggs, whereas the control
showed over 6 per cent. Thus the cold evidently retarded the process of natural
parthenogenesis.

Of the two general methods of obtaining parthenogenesis,
the exosmotic method yields the better results. Usually the
endosmotic method produces only a small per cent, of segmenting
eggs. But this is probably due to the specific poisoning action
of the substances used in lowering surface tension, for when
the sea-water is simply diluted, or when a harmless substance
like egg albumen is used, much higher per cents, of dividing
eggs are obtained. For example, when eggs were subjected to
the action of a i per cent, solution of egg albumen (Kahlbaum)
in sea-water, more than half of them divided, as shown in the
accompanying table. In such a solution, the eggs were observed
to undergo membrane swelling (see p. 173) and this was followed
bv endosmosis as shown bv the increased diameter.

L. V. HEILBRUNN.

Length of Exposure to i Per

Cent. Egg Albumen. Segmentations.

30 minutes 55/ioo

66 48/100

118 SO/TOO

233 6/100

The results gained with the endosmotic method used alone,
are never as good as those which can be obtained with the ex-
osmotic method. Neither are the per cents, of segmentation as
high, nor is the degree of development attained as great. In
spite of the fact that Loeb ascribes to hypertonic solutions a
mere correcting effect, it is I think a noteworthy fact that no
method of artificial parthenogenesis yet tried on the Arbacia
egg, is truly effective, unless at some stage of the process it
requires a hypertonic solution. In connection with this point,
I made a number of experiments to test the butyric acid-KCN
method which Loeb found so effective for Stronglocentrotus.
I was at the time convinced that the method was essentially
the same as Delage's acid and alkali method (Delage '07), and
I tried to find if NaOH could not be substituted for the KCN,
in other words if the correcting action of the latter was not due
solely to its alkalinity. In no case, however, was I able to get
any results either with KCN or NaOH as a "correcting agent"
after butyric acid treatment. Indeed recently Loeb ('136)
has pointed out that the method is not suited for the Arbacia egg.

The question now arises if the two methods of producing
segmentation, the one endosmotic, the other exosmotic, have
anything in common. As a result of my experiments, I find
that both methods produce a gelatinization or coagulation
within the egg.1

Most biologists believe that the mitotic spindle is a con-
densation or coagulation product, and there is excellent support
for this view. Hence any initiation of development must soon

1 In the test-tube, gelatinization and coagulation of proteins are apparently
quite different phenomena, the former converting the entire mass into a jelly,
the latter resulting in a separation of a precipitate. But in the small field of
action of a sea-urchin egg, it would not be so easy to distinguish between the two,
for the entire egg is smaller than a single flake of the usual precipitate. Moreover,
even in the test-tube coagulation is usually preceded by a stage strictly comparable
to gelatinization; the entire mass becomes opalescent and assumes a greatly
increased viscosity; only later does the precipitate appear.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 19 1

lead to some coagulation before mitosis can be accomplished.
At present all students of artificial parthenogenesis, if they con-
sider this coagulation at all, regard it as a secondary result.
Thus, some think an increase of oxidations is first produced by
all parthenogenetic agents, and that the increased oxidation
involves changes which produce coagulation. Others think of
the prime cause as an increase of permeability. But it is also
possible to believe that the primary effect of all the partheno-
genetic reagents is a coagulation effect. This view has had as
its adherents, at one time or another, some of the foremost
students of artificial parthenogenesis. In his earliest papers
on the subject, Loeb occasionally leaned toward coagulation as a
possible primary effect, but he soon abandoned the idea to
become its vigorous opponent. Delage for many years main-
tained that artificial parthenogenesis is the result of a coagu-
lation followed by a liquefaction, he considered membrane
elevation as one evidence of such a coagulation. At present,
however, (Delage and Goldschmidt '13), he favors the Lillie
theory of increased permeability as affording a more probable
explanation of the facts. Possibly the most vigorous and
scientific attempt to support the coagulation theory was that of
Fischer and Ostwald ('05). These workers argued from a theo-
retical standpoint that all parthenogenetic agents are of such a
nature as to produce coagulation. Later Ostwald ('07) retreated
from this view and admitted that coagulation might be only
secondarily produced as a result of increased oxidation. That
all parthenogenetic agents cause coagulation, was denied by
Loeb.1 He pointed out that benzol, toluol, and saponin are
not protein coagulants. Since then the theory of Fischer and
Ostwald had had no one to defend it.

The view that I would maintain is that the only physico-
chemical effect which all parthenogenetic agents possess in
common is the production of a gelatinization or coagulation
within the egg. Hence I regard this gelatinization (or coagu-
lation) as the direct cause of the initiation of development.

Leaving theory aside, it is possible to demonstrate that all
parthenogenetic agents actually do produce gelatinization or

1 Loeb, '09, p. 217.

192 L. V. HEILBRUNN.

coagulation within the egg. The distinguishing feature of a gel
as opposed to a sol is its greater viscosity. All other macroscopic
differences depend upon this. Indeed, according to Freund-
lich, the viscosity of a colloidal solution may be taken as a measure
of its tendency to gelatinize ("Gelatinierungsbestreben").1
The viscosity of the Arbacia egg protoplasm was regarded as an
index of the state of aggregation of its constituents.

There are two general methods of measuring viscosity. One
can either measure the rate of flow of a fluid, or one can study
the movement of particles through the fluid. Both of these
methods were used in studying the viscosity of the Arbacia
cytoplasm. It is very easy to observe the rate of flow of the
egg cytoplasm. One has only to exert pressure on the coverslip
and the vitelline membrane soon bursts, allowing the egg con-
tents to flow out. With this method, only great changes in
viscosity can be noted, but gelatinization always does produce
a very great increase in viscosity, so that the method usually
suffices. The pressure can be applied by pushing on the cover-
slip with the point of a dissecting needle, or if greater accuracy
is desired, a square piece of glass (broken from a slide) can be
dropped from a known distance. This method of observing
changes in viscosity is not entirely above criticism; for cortical
changes, produced by the reagents used, might have an effect
on the size of the aperture through which the cytoplasm flows.

The second method is much more exact and reliable. It
involves a study of the movement of granules through the
cytoplasm. The force which must be exerted to push the
granules through the cytoplasm is a measu e of the viscosity of
the latter. Centrifugal force is of necessity used, and a hand
centrifuge does very nicely.2 Into one tube of the centrifuge
are placed eggs which have been treated in various ways; into
the other, normal eggs. After the centrifuging has been accom-
plished, a microscopic examination reveals any differences which

1 Freundlich, '09, p. 416. Freundlich and Ishizaka, '13.

2 A Bausch and Lomb instrument was used in my experiments, and the eggs
were placed into the small glass tubes of the haematocrit attachment. New tubes
were used in each experiment so that all danger of contamination was avoided.
One turn of the high-speed handle involved 130 revolutions of the tubes. The
distance between the end of each tube and the axis was approximately 7^ cm.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 1 93

may exist between norma' and treated eggs. When normal
unfertilized Arbacia eggs are centrifuged vigorously, the pro-
toplasmic materials rapidly become separated into four layers
or zones. Lyon ('07) has given an excellent description of the
appearance of these zones and the reader is referred to his paper
for detai s. The pigment-bearing granules come to be all massed
at one pole, in the pigment zone; next to this is a zone of granular
material, then a hyaline zone, and at the pole opposite the pig-
ment zone is a small dense accumulation of substance which
because of its color is known as the gray cap. The egg nucleus
lies in the hyaline zone, directly beneath the gray cap. When
an egg shows all these zones, I shall refer to it as "stratified."
As the viscosity of the protoplasm increases, stratification be-
comes more and more difficult; in a thoroughly coagulated egg
no stratification is possible.

Hypertonic solutions produce a very noticeable coagulative
change in the sea-urchin egg. In 1913 a few preliminary ex-
periments were performed in which the eggs were pressed out
of shape by pushing down on the coverslip with a dissecting
needle. It was found that "Nad hypertonic sea-water >:
produced a marked increase in the viscosity of the cytoplasm.
Such a solution causes swelling of the vitelline membrane. This
swelling is apparently absent if a freshly prepared 0.49 M MgCl2
solution is used. Eggs treated with this solution became very
much more viscous than they had been previously. Whereas
the normal unfertilized eggs shot out their contents rapidly if
subjected to a slight pressure, eggs which had been immersed in
MgCl2 for 85 minutes could be subjected to considerable pres-
sure without losing their circular outline. They could indeed
be flattened out into a thin "pancake."

More accurate data were obtained with the centrifuging
method. On July 23, 1914, some eggs were placed into 50 c.c.
of sea-water plus 8 c.c. of 2.5 M NaCl at 2:2 if P.M. At
2:44 P.M., these eggs were placed into one tube of the centrifuge,
and into the other tube were placed some normal untreated eggs
of the same female. At 2:45 P.M., after a few preliminary
turns, the tubes were revolved for 28 seconds at a rate of 162
revolutions per second. The eggs were then examined. The

L- v- HEILBRUNN.

normal eggs showed the typical zones of centrifuged eggs. The
pigment zone extended for about one fourth of the egg diameter
along the axis of stratification. The contrast between normal
eggs and those which had been treated with the hypertonic
solution was very marked. The latter as a rule showed no
stratification whatsoever, although some eggs showed just the
beginnings of such a process, the pigment granules being slightly
more abundant at one pole of the egg than at the pole opposite
to it.

The experiment was repeated on July 24. In this case un-
fertilized eggs were subjected to "NaCl hypertonic sea-water"
at 10:29 A.M., and about 25 minutes later they were placed
into one tube of the centrifuge. Into the other tube, normal
eggs were placed. Then (at 10:55^ A.M.) the tubes were
revolved for 19 seconds at the rate of 171 revolutions per second.
Upon examination, the normal eggs showed complete strati-
fication under low power of the microscope. High power examina-
tion showed that a few pigment granules had not reached the
pigment zone, but had lingered near the equator of the eggs.
The eggs treated with hypertonic sea-water showed no strati-
fication. In some eggs one pole was slightly paler than the
other, and fewer pigment-bearing granules could be found at
this pole. But even in these eggs, the pigment was scattered
throughout every part of the cytoplasm. At 11:29 the eggs
in hypertonic sea-water were centrifuged again and compared
with normal eggs given the same treatment. Longer and more
vigorous centrifuging was resorted to, and the tubes of the centri-
fuge were revolved 186 times per second for 35 seconds. The
normal eggs now appeared perfectly stratified; they were often
elongated as a result of the treatment. The eggs in hypertonic
sea- water usually showed only the beginnings of stratification,
a tendency for the pigment to be massed toward one pole. In
a few eggs, however, the pigment was practically limited to one
half of the egg.

These experiments show a striking increase in cytoplasmic
viscosity after eggs have been exposed to hypertonic solutions.
This could be noted both by observation of the flow of the cyto-
plasm itself, as well as by more careful observations of the

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 1 95

movements of granules through the cytoplasm. The only
explanation of such a marked increase in viscosity is that gela-
tinization or coagulation has taken place within the egg. Un-
doubtedly exosmosis causes some constituent of the cytoplasm
to change from sol to gel.

Endosmosis was also found to cause coagulation, either when
it resulted from a dilution of the outer medium, or when it was
the result of a lowered surface tension of the vitelline or plasma
membrane. This could be demonstrated either by observing
the cytoplasmic flow from compressed eggs, or by noting the
granular movements in centrifuged eggs. Numerous experi-
ments were made with distilled water. In an experiment of
June 30, 1914, a square piece of glass which had been broken
from a slide was used to compress the eggs. It weighed 1.31
grams. The piece of glass was held between the thumb and
forefinger, so that its one edge rested on the slide just to one
side of the coverslip which covered the eggs under observation.
At the desired time, the glass piece was dropped and the com-
pressed eggs could then be immediately observed. Normal
eggs were deprived of their jelly by shaking them 6 or 8 times in
a test-tube, and were then subjected to the pressure of the piece
of glass as just described. The eggs all ruptured, the contents
flowing out for a distance of about 60/1. Eggs (from the same
female) were then dropped into distilled water at 10:28 A.M.
At io:28f any jelly which may have remained around them
was removed by shaking the eggs. At 10:29^ the eggs were
subjected to the pressure of the same piece of glass which had been
used in the case of the normal eggs. They resisted the pressure
and remained circular or nearly circular in outline. In the eggs
observed, membrane elevation had not taken place. The
greater resistance to pressure of the eggs treated writh distilled
water is an indication of their increased viscosity. The eggs
tended to remain spherical, they showed a definite elasticity.
Similar observations showing increased viscosity after treatment
with distilled water, were made a number of times both in 1913
and 1914.

On July 25, 1914, at 3:59 P.M., 7 drops of an egg suspension
were dropped into 15 c.c. of distilled water. The eggs in the

196 L. V. HEILBRUNX.

distilled water were then hastily pipetted into one tube of the
centrifuge, the other tube contained untreated eggs. The
centrifuge was started at 4:00^ P.M., and for 25 seconds the
tubes were revolved at a rate of 156 revolutions per second.
The eggs in distilled water were examined as soon as possible
(about half a minute later). They showed not a sign of strati-
fication. Most of the pigment had been lost as a result of the
action of the distilled water, but high power examination showed
that the pigment-bearing granules still retained some pigment.
These granules were scattered all through the cell. On the
other hand the normal eggs showed the typical stratification.
The viscosity of the cytoplasm had therefore increased enor-
mously in the eggs treated with distilled wrater, so that the
granules were prevented from wandering through it. In another
experiment (July 22, 1914) the eggs wrere transferred back to
sea-water before being centrifuged. At 3:51 P.M., 5 drops of
egg suspension were added to 15 c.c. of distilled water. At
3:55 P.M., some of these eggs were removed from the distilled
water and placed into sea-water again. The eggs thus trans-
ferred were the ones studied; they were placed into one tube of
the centrifuge, normal eggs occupying the other. Beginning at
3:59^ P.M., the tubes were revolved 165 times per second for
30 seconds. Upon examination it was found that whereas the
normal eggs were completely stratified, not a sign of stratification
could be observed in the eggs which had been exposed to distilled
water. Evidently endosmosis, following immersion in distilled
water, leads to a gelatinization or coagulation in the cytoplasm.
This effect is also produced when endosmosis follows a lowered
surface tension of the plasma membrane. A drop of egg sus-
pension was stirred up with a drop of chloroform, and the eggs
rapidly increased in diameter. When subjected to the pressure
of a piece of glass broken from a slide, the eggs flattened out but
remained perfectly circular in outline. The same result was
obtained if toluol was used instead of chloroform. The eggs
under pressure behave like eggs which have been coagulated by
some typical coagulant, e. g., HgCl2. On the other hand, when
the normal eggs were subjected to the pressure of the same piece
of glass, the cytoplasm flowed out a considerable distance. The

STUDIES IN ARTIFICIAL PARTHENOGENESIS.

results obtained with the centrifuging method are still more
convincing.

At 10:02^ A.M. (July 23, 1914), 2.5 c.c. of toluol were added
to 2.5 c.c. of sea- water containing eggs, and the mixture was
stirred thoroughly with a glass rod (in a Syracuse watch-crystal).
At 10:07^ A.M., 7 drops of the egg-containing liquid were
transferred back to normal sea-water. Into one tube of the
centrifuge were placed normal eggs, the other contained eggs
which had been treated with toluol for five minutes. The eggs
were then centrifuged for 30 seconds at a rate of 152 revolutions
per second. Upon examination, the normal eggs appeared
perfectly stratified, whereas those which had been exposed to
toluol showed not a trace of stratification. As a result of the
toluol treatment the eggs increased their diameter about 2/4.
The experiment was repeated with identical results on July 25,
I c.c. of toluol being added to 3 c.c. of sea-water. In this case
the eggs were exposed 5 minutes, and were then centrifuged
23 seconds at an average of 158 revolutions per second.

Saponin was also found to produce a coagulation within the
egg. At 5:35 P.M. (July 23), 7 drops of an egg suspension
were placed in about 15 c.c. of 0.2 per cent, saponin solution
(in sea-water). -At 5 140 P.M., the eggs were returned to normal
sea- water. The eggs were then subjected to an exceptionally
long centrifuging process. Beginning at 5:47 P.M., for 50
seconds they were revolved at an average of 166.5 revolutions
per second. Normal eggs subjected to this centrifuging process
were of course completely stratified. Of the eggs treated with
saponin, only about 15-20 per cent, showed any signs of strati-
fication, the remaining eggs were all totally unstratified.

In the experiments with saponin and toluol, the reagent pro-
duces a lowering of surface tension directly. With some re-
agents, membrane swelling occurs first, and the lowered surface
tension thus produced results in endosmosis. In this case also,
coagulation follows an increase in egg volume. On August 4,
1914, eggs were washed in 0.55 M NaCl solution and were then
dropped into 0.55 M NaCl plus 1.5 c.c. N/io NaOH at 10:59^
A.M. These eggs were then placed into one tube of the centri-
fuge, and into the other were placed eggs which had been im-

198 L. V. HEILBRUNN.

mersed in 25 c.c. of 0.55 M NaCl at 10:58 A.M. Beginning at
11:165 A.M., the eggs were centrifuged for 17 seconds at the
rate of 153 revolutions per second. When examined, the eggs
in NaCl alone showed stratification, although the boundary
between the various zones was not a sharp one. Gray cap,
hyaline zone, granular zone, were distinguishable, but the pig-
ment was not entirely restricted to^the pigment zone. On the
other hand, the eggs in the alkaline NaCl solution showed not
a vestige of stratification, in every case the pigment-bearing
granules were evenly distributed throughout the cell. In these
eggs exposed to an alkaline NaCl solution, membrane swelling
had taken place. Similarly when membrane swelling was in-
duced by i per cent, albumen, coagulation followed the endos-
mosis thus produced. On August 4, 1914, at 10:40 A.M., some
eggs were placed in a filtered I per cent, solution of egg albumen
solution in sea-water. After about 50 minutes, some eggs in
the albumen solution were put into one tube of the centrifuge,
normal eggs were placed in the other. The eggs were then
centrifuged at the rate of 143/evolutions persecond for2O seconds.
The normal eggs all showed stratification, although the various
zones were not sharply marked off from each other. Of the
eggs in the albumen solution only a few were examined. Of
these ten showed not a trace of stratification, one egg showed
stratification, but on examination it was found that its membrane
had remained unswollen. Pressure experiments also showed a
great increase in viscosity after membrane swelling. Sodium
iodide and I per cent, albumen were used in these experiments.
It is evident that botfy Jhe reagents which cause exosmosis
and those which produce endosmosis, cause gelatinization or
coagulation of some substance fn ttye Arbacia egg. Thus all
parthenogenetic agents produce such ^coagulative change.1
The experiments indicate that the coagulation occurs just as

1 I have omitted mention of radium and ultra-violet rays as parthenogenetic
agents. I have not worked with either of these methods, and so can only offer a
theoretical interpretation. Judging from Loeb's description (Science, N.S., XL.,
680 (1914)), ultra-violet rays produce membrane swelling and thus they no doubt
induce endosmosis. Moreover both radium and ultra-violet rays are protein
coagulants, according to Dreyer and Hanssen, C. R. Acad. Sci., CXLV., 234
(1907). Hardy has also decribed gelatinization of a globulin solution as a result
of radium radiation. Proc. Camb. Phil. Soc., XII., 201 (1903).

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 199

soon as water enters or leaves the cell. It is probable therefore
that the effect is primary, and not the result of intermediate
changes.

In normal fertilization, the sperm likewise produces coagula-
tive changes in the cytoplasm. Albrecht in 1898 showed that
there was an increase in viscosity after fertilization. He drew
his conclusions from observations on compressed eggs. I have
repeated Albrecht's observations a number of times. Much
more striking demonstration of gelatinization or coagulation
after fertilization is afforded by the centrifuge method. On
July 22, 1914, eggs from a single female were washed twice and
divided into two lots. At 10:00 P.M., half of the eggs were
fertilized. One tube of the centrifuge was then filled with fer-
tilized eggs, the other contained unfertilized eggs. At 10:10
P.M., the tubes were revolved for 18 seconds at an average
speed of 180.5 revolutions per second. The eggs were then
examined immediately. The unfertilized eggs were typically
stratified. On the other hand, the eggs which had been ferti-
lized were not at all stratified. Out of hundreds examined,
only one showed stratification and that was peculiar in lacking
an elevated membrane. Probably it had escaped fertilization.
The coagulative change begins to be apparent very soon after
fertilization. In one case the beginnings of the process could
be observed 2\ minutes after insemination. At 8:24 P.M.
(July 26) some eggs wrere fertilized in a small volume of sea-water
and were shaken about to insure rapid sperm contact. At
8:26^ P.M., fertilized and unfertilized eggs were centrifuged
at the same time (in separate tubes). For 23 seconds an average
speed of 141 revolutions per second was maintained. Upon
examination the fertilized eggs showed some evidences of strati-
fication. The gray cap was becoming evident in some. But
most of the fertilized eggs showed little more than a tendency
for the pigment to mass at one pole. The unfertilized eggs were
clearly more stratified. They showed the gray cap plainly
in all cases, and a hyaline zone was also recognizable in them.
Thus only 2\ minutes after insemination, the sperm had already
begun to produce coagulative changes in the cytoplasm.

From these experiments, I have been led to conclude that

2OO L. V. HEILBRUNN.

initiation of development, either artificially produced or the
result of fertilization, always involves a gelatinization or coagu-
lation with the egg. This coagulative change is evoked before
the egg interior shows any other signs of the approach of develop-
ment. The coagulating substance (or substances) is evidently
very ready to coagulate. The ensemble of conditions within
the egg is no doubt responsible for this unstable state. Only
a slight change in salt concentration is then sufficient to bring
about coagulation. The fact that both increase and decrease
of salt concentration are effective, suggests that the protein
involved is of the globulin type.1

The mitotic spindle probably arises as a direct result of the
coagulative change. The actual explanation of how this occurs
is not truly a problem of biology but a problem of colloid chemis-
try, for it has been shown (Fischer, '99) that coagulation of
inanimate proteins can produce structures identical in appear-
ance with the mitotic spindle.

IV. SUMMARY.

1. The unfertilized Arbacia egg consists essentially of fluid
proteins, surrounded by a stiff membrane (the vitelline mem-
brane), which is the plasma membrane of the egg cell. This
membrane is a protein gel with little or no admixture of lipoids.

2. There are two types of cortical change, membrane swelling
and membrane elevation. In the former, the membrane ab-
sorbs water and increases in thickness; in the latter, the normal
process, it becomes lifted away from the egg surface.

3. The vitelline membrane only loses its semipermeable proper-
ties several minutes after elevation, and the increase of permea-
bility that then ensues is best regarded as a result rather than
as a cause of the process.

4. Artificial membrane elevation is produced only by sub-
stances which lower the surface tension of the vitelline mem-
brane. This is explainable on the basis of a theory that considers
the various forces at play on the membrane. After its surface
tension is lowered, the forces exerted outward are stronger than

1 If the protein referred to here is, as seems most likely, the substance which
forms the spindle fibers, the fact that it shows the properties of a globulin becomes
of significance. For globulins are the proteins most closely associated with con-
tractile processes, muscles consisting almost entirely of them.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 2OI

those balancing them, and the membrane is as a result pushed
away from the egg.

5. Cytolysis is not due to a simple swelling of egg proteins
for these can be shown to coagulate rather than to swell, it is
due to a continuation of the same process as that which results
first in membrane elevation.

6. The sperm produces membrane elevation by lowering the
surface tension of the vitelline membrane. This it accomplishes
by causing the membrane to swell at the point where it strikes
the egg.

7. Experimental evidence is presented to show that this
membrane swelling produced by the sperm is due to the action
of an acid.

8. The fact that potassium cyanide does not prevent initiation
of development by hypertonic sea-water, is taken as evidence
against the oxidation theory of artificial parthenogenesis.

9. All initiation of development involves the gelatinization
or coagulation of some substance within the egg. This coagu-
lative change can be demonstrated to take place before the egg
interior shows any other signs of the approach of development.

The above summary covers only the more important points
of the paper.

I desire to thank Professor F. R. Lillie, and also Professors
Child and Tower, for their kind criticism of the manuscript.
To Professor Lillie I am also indebted for the use of a table at
the Woods Hole Marine Biological Laboratory, where all of the
experimental work was done.

V. REFERENCES.

Albrecht, E.

'98 Untersuchungen zur Struktur des Seeigeleies. Sitzungsber. d. Ges. f.

Morph. u. Physiol. in Miinchen, 1898, pp. 133-141.
Czapek.

'n Ueber cine Methode zur direkten Bestimmung der Oberflachenspannung

der Plasmahaut von Pflanzenzellen, Jena, 1911, G. Fischer.
Delage, Y.

'07 L'oxygene, la pression osmotique, Jes acides et les alcalis dans la partheno-

genese experimentale. C. R. Acad. Sci. Paris, T. 145, pp. 218-224.
Delage, Y. and Goldschmidt, M.

'13 La parthenogenese naturelle et experimentale, Paris, 1913.

202 L. V. HEILBRUNN.

Fischer, A.

'99 Fixirung, Farbung und Bau des Protoplasmas, Jena, 1899, G. Fischer.
Fischer, M. H.

'10 Oedema. New York, Wiley.

'12 A Further Response to Some Criticisms of the Colloid-chemical Theory of
Water Absorption by Protoplasm. Journ. Amer. Med. Assn., Vol. 59, pp.
1429-1433.
Fischer, M. H. and Ostwald, Wolfgang.

'05 Zur physikalisch-chemischen Theorie der Befruchtung. Arch. f. d. ges.

Physiol., Bd. 106, pp. 229-266.
Fol, H.

'79 Recherches sur la fecondation et le commencement de 1'henogenie chez
divers animaux. Mem. d. 1. Soc. de Phys. et d'Hist. Nat. de Geneve, T. 26.
Freundlich, H.

'09 Kapillarchemie, Leipzig, 1909.
Freundlich, H., and Ishizaka, N.

'13 The Velocity of Coagulation of Aluminium Hydroxide Sols as Measured by

Changes in the Viscosity. Trans. Faraday Society, Vol. 9, pp. 66-79.
Gibbs, J. Willard.

'7S-'78 On the Equilibrium of Heterogeneous Substances. Republished in the

Scientific Papers of J. Willard Gibbs, London and New York, 1906.
Heilbrunn, L. V.

'13 Studies in Artificial Parthenogenesis — I. Membrane Elevation in the

Arbacia Egg. BIOL. BULL., Vol. 24, pp. 345-361.
'15 The Measurement of Oxidations in the Sea-urchin Egg.

Science (in press).
Herxheimer, G.

Mikroskopische Technik, in Abderhalden's Handbuch der biochemischen
Arbeitsmethoden, Bd. VII., pp. 632-714 (see pp. 679-680).
Kite, G. L.

'13 Studies on the Physical Properties of Protoplasm. I. The Physical Prop-
erties of the Protoplasm of Certain Animal and Plant Cells. Amer. Journ.
Physiol., Vol. 32, pp. 146-164.
Lepeschkin, W. W.

'n Zur Kenntnis der chemischen Zusammensetzung der Plasmamembran.

Ber. d. deutsch. bot. Ges., Bd. 29, pp. 247-261.
Lillie, Ralph S.

'10 The Physiology of Cell Division — II. The Action of Isotonic Solutions of
Neutral Salts on Unfertilized Eggs of Asterias and Arbacia. Amer. Journ.
Physiol., Vol. 26, pp. 106-133.

'n The Physiology of Cell Division — IV. The Action of Salt Solutions
Followed by Hypertonic Sea-water on Unfertilized Sea-urchin Eggs and the
Role of Membranes in Mitosis. Journ. Morph., Vol. 22, pp. 695-730.
Loeb, Jacques.

'08 Ueber den Mechanismus der Agglutination. Kolloid-Zeitsch., Bd. 3, pp.

H3-II4-
'09 Die chemische Entwicklungserregung des tierischen Eies, Berlin, 1909,

Springer.
'i3a Artificial Parthenogenesis and Fertilization. Chicago, 1913, University

of Chicago Press.

STUDIES IN ARTIFICIAL PARTHENOGENESIS. 2O3

'i3b Reversibility in Artificial Parthenogenesis. Science, N.S., Vol. 38, pp.

749-751.
Loewe, S.

'13 Membran and Narkose. VVeitere Beitrage zu einer Kolloidchemischen

Theorie der Narkose. Biochem. Zeitschr., Bd. 57, pp. 161-260.
Lyon, E. P.

'02 Effects of Potassium Cyanide and of Lack of Oxygen upon the Fertilized
Eggs and the Embryos of the Sea-urchin (Arbacia punctulala). Amer.
Journ. Physiol., Vol. 7, pp. 56-75.
'07 Results of Centrifugalizing Eggs. Arch. f. Entwicklungsmech., Bd. 23,

pp. 151-173-
McClendon, J. Q.

'10 On the Dynamics of Cell Division — II. Changes in Permeability of De-
veloping Eggs to Electrolytes. Amer. Journ. Physiol., Vol. 27, pp. 240-275.
Ostwald, Wolfgang.

'07 Uber das Vorkommen von oxydativen Fermenten in den reifen Geschlechts-
zellen von Amphibien und fiber die Rolle dieser Fermente bei den Vorgangen
der Entwicklungserregung. Biochem. Zeitsch., Bd. 7, pp. 409-472.
Pictet, C.

'91 Recherches sur la spermatogenese chez quelques Invertebres de la Mediter-
ranee. Mitth. d. zool. Stat. z. Neapel, Bd. 10, pp. 75-152 (see pp. 100-105).
Quincke, G.

'77 Ueber Diffusion und die Frage, ob Glas fur Case undurchdringlich ist.

Poggendorff's Annalen, Bd. 160, pp. 118-123.

'02 Ueber unsichtbare Fltissigkeitsschichten und die Oberflachenspannung
fliissiger Niederschlage bei Niederschlagsmembranen, Zellen, Colloiden und
Gallerten, Ann. d. Physik. 4te Folge, Bd. 7, pp. 631-682, 701-744.
Reinke, F.

'95 Untersuchung tiber Befruchtung und Furchung des Eies der Echinodermen.

Sitzungsber. d. Akad. d. Wiss. z. Berlin, 1895, pp. 625-637.
Ries, J.

'09 Kinematographie der Befruchtung und Zellteilung. Arch. f. mik. Anat.,

Bd. 74, pp. 1-31.
Robertson, T. B.

'12 Studies in the Fertilization of the Eggs of a Sea-Urchin (Slrongylocentrotus
purpuratus) by Blood-Sera, Sperm-Extract and Other Fertilizing Agents.
Arch. f. Entwicklungsmech., Bd. 35, pp. 64-130.

ON SUCCESSIVE DUPLICATE MUTATIONS.1

R. RUGGLES GATES,
UNIVERSITY OF LONDON.

Nilsson-Ehle2 was the first to formulate the hypothesis of
duplicate factors or representatives for the same character.
He brought forward evidence from crosses of red and white
varieties in certain Swedish strains of wheat, tending to show
that in different F2 families plants with red and white grains
occurred in the ratios respectively 3:1, 15:1 and 63 :i ; from which
he concluded that three independent units for red were present,
each of which could produce the color alone. Although his
conclusions were criticized by Kajanus,3 yet there remains a strong
presumption in their favor, and several other cases of supposed
duplicate factors have been described, though these have for
the most part rested upon more insecure data than the original
instances of Nilsson-Ehle.

Certain suggestions have been made concerning the origin of
this duplicate or triplicate condition. Emerson and East4
pointed out in general terms that if a factor should become
located in a different chromosome or should be affected in
any way so as not to be always allelomorphic to itself, then a
duplication of determiners would result. Shull5 has listed the
supposed cases of duplicate determiners and remarks that such
a condition of duplication might also result from "repeated pro-
gressive mutations." In the same paper, Shull endeavors to
account for the origin simultaneously of a duplicate "gene"

1 Presented before the American Genetic Association, San Francisco meeting,
August 3, 1915.

2 Nilsson-Ehle, H., 1909, " Kreuzungsuntersuchungen an Hafer und Weizen,"
I., Lunds Univ. Arsskrift., N.F., Afd. 2, Bd. 5, Nr. 2, pp. 122.

3 Kajanus, B., 1914, "Zur Kritik des Mendelismus," Zeitschr. f. Abst. u. Vererb.,
12: 206-224.

4 Emerson, R. A., and East, E. M., 1913, "The Inheritance of Quantitative
Characters in Maize," Agric. Exp. Sta. Nebraska, Research Bull. 2, pp. 120.

6 Shull, Geo. H., 1914, "Duplicate Genes for Capsule-form in Bursa bursa-
pastoris," Zeitschr. f. Abst. u. Vererb., 12; 97-149, Figs. 7.

204

ON SUCCESSIVE DUPLICATE MUTATIONS. 2O5

for capsule form in Bursa bur sa- pastor is and, at the same time,
of the mutant B. Heegeri. That hypothesis will not, however,
apply to the probably more frequent cases in which duplicate
factors for a particular character are found without any other
mutation having taken place. An explanation will therefore
have to be found for the duplicate or triplicate condition in wheat
or in any other organisms in which it occurs.

It is the purpose of the present paper to discuss more pre-
cisely the manner in which such monohybrid characters originate
and particularly the way in which they may afterward become
duplicate or triplicate. (Enothera rubricalyx affords a typical
case of a mutant originating as a monohybrid, probably through
a transformation in one chromosome or one pair of chromosomes.1
I have pointed out elsewhere2 that when the duplicate or tripli-
cate condition occurs it might be reasonably supposed to have
arisen through the same general change having taken place
independently in two or three different chromosomes of the
x series.

In an original mutation of this kind the new character of
course forms a pair by contrast with the old unaltered character.
If a single chromosome has undergone this change and the new
condition is dominant, then a heterozygous mutant Aa will be
produced having the new character but splitting in its offspring
in a 3:1 ratio. This is the way in which (Enothera rubricalyx
originated from (E. rubrinervis, as I have shown elsewhere.1

If now in the mutant race one or both members of a second
pair, a' a', of chromosomes undergoes a corresponding change,
to A'a', or A' A' then we shall have duplicate factors AA' for
the same character, and in the offspring of such individuals
the new type and the original type would appear in the ratio of
15:1. A similar mutation in a third pair would give the triplicate
condition with a ratio of 63:1.

It may be pointed out that this assumption of similar changes
in different members of the x series of chromosomes is by no

1 See Gates, R. Ruggles, 1915. "On the Origin and Behaviour of Oenothera
rubricalyx," Journ. of Genetics, 4: 353—360.

2 Gates, "The Mutation Factor in Evolution," p. 317, Macmillan, London,
IQI5-

206 R. RUGGLES GATES.

means an improbable one. It does not assume that the chromo-
somes which underwent the change were alike, but merely that
they were more nearly alike than the others of the series. That
the chromosomes of the x series are actually differentiated
there are many lines of evidence to show. One of the most
recent is the fact, ascertained by Doncaster and Gray,1 that in
certain echinodern crosses particular chromosomes swell up and
form vesicles in the strange cytoplasm of the egg or of another
species while other chromosomes exhibit no such effect. On
the other hand, the materials of the chromosomes obviously
possess many similarities which probably outweigh their chemi-
cal differences.

It may further be pointed out that if both members instead
of one member of a pair of chromosomes underwent a change,
say in a pollen mother cell, the only difference would be that
four instead of two mutated germ cells would result, each of
which might function in producing a mutant. It is almost
impossible to determine whether one or both members of a pair
of chromosomes underwent the change in any particular in-
stance, but in either case the original mutant would be hetero-
zygous, though continued inbreeding would produce ultimately
a homozygous race, as in the case of (E. rubricalyx. This is
probably the history of Nilsson-Ehle's wheats which are dup-
licate or triplicate for the red color factor in their grains.

From the evidence obtained in F2 and F3 in crosses of various
Swedish wheats having red kernels, with white-grained varieties,
Nilsson-Ehle concludes2 that while the varieties known as
Sammet and Grenadier have 3 independent units for red, Extra-
Squarehead has only one, since it gave (p. 67) only ratios approxi-
mating 3:1. In a later paper3 continuing this work the same
author finds (p. 22) that Swedish Binkel wheat contains two
factors for red. From one FZ family he grew 94 FS families,
with results which may be tabulated as follows:

1 Doncaster, L., and Gray, J., 1913, "Cytological Observations on the Early
Stages of Segmentation of Echinus Hybrids," Quart. Journ. Micr. Sci., 58: 483-
510, pis. 28-29.

2 Nilsson-Ehle, H., 1909, Kreuzungsunterschungen an Hafer und Weizen. I.
Lunds Univ. Arsskrist., N.F., Afd. 2, Bd. 5, Nr. 2, pp. 122.

3 Nilsson-Ehle, H., 1911, " Kreuzungsuntersuchungen an Hafer und Weizen,"
II., Lunds Univ. Arsskrift., N.F., Afd. 2, Bd. 7, Nr. 6, pp. 82.

OX SUCCESSIVE DUPLICATE MUTATIONS.

TABLE I.

207

Expected Ratio
of Families.

Totals.

Expectation.

7
4

4

I

40 families were constant red
23 families split in the ratio 3 : i
25 families split in the ratio 15:1
6 families were constant white

483 : 142
789 : 47

468.75 : 156.25
783-75 : 52.25

It will be seen that the frequency of families is very close to
expectation. The totals of the families containing a 15:1 ratio
are also very close to expectation, but for the 3:1 families the
agreement is not so good. The evidence seems sufficient, how-
ever, to justify the conclusion that two factors and two only
are here concerned.

In the same way evidence is obtained (p. 25) to show that a
certain pure line (0406) has in one case a single factor for red
and in another case two factors. To use the terminology of
Lang, the race is monomerousin one case and dimerous in another.
In crosses between the 0406 race and 0234, which was also red,1
ratios of 15:1 and 3:1 were obtained showing that two factors
were present, one of which must have been derived from each
parent. Hence the 0406 race must in this case have been
monomerous. In crosses between 0406 and a white race, 15:1
ratios were again obtained, showing that the 0406 race is now
dimerous. The genetic relationships of the strains used in these
two crosses is not stated, but a simple explanation is that in the
meantime the strain had undergone a second (invisible) mutation.

No explanation of the origin of this condition was offered.
But there are at least two ways in which the dimerous condition
may have been derived from the monomerous: (i) Through a
mutation on the part of a second pair of chromosomes, (2)
through a re-mating of the chromosome pairs. Later we shall
compare the consequence of each of these methods of deriving
the duplicate condition. In the first case the duplicate mutation
is produced by a change very similar to that which produced the
original mutant. In the second case the secondary change is a
mechanical one, very different from the primary change which
was probably chemical in nature.

1 The results are given in Ber. dent. hot. Gesetts., 29: 65-69, 1911.

2O8 R. RUGGLES GATES.

Another pure line of wheat (0290) was also found to be dimer-
ous for the red factor in one case and monomerous in another.
A race called 0501 was found to be probably trimerous like
Swedish Sammet. Nilsson-Ehle considers it scarcely probable
that in these two wheats the same three factors for red are present,
and thinks that perhaps many more than three independent
factors have to be reckoned with. There seems, however, no
reason for such an assumption. It appears more probable that
corresponding chromosomes undergo the same change in each
case so that the factors are all homologous with each other,
though of independent origin in the different races.

We may, therefore, account for the origin of the duplicate
and triplicate "factors" for red in the Swedish wheats by assum-
ing that successive mutations have occurred and that in each
case the duplicate or triplicate condition has afterward become
homozygous and stable through the repeated self-fertilization
occurring in later generations.

Turning now to the history of (Enothera rubricalyx, it appears
that the original monomerous condition has become dimerous
in subsequent generations of culture. And it will be seen from
the culture records that this has happened independently several
times in different lines of descent.

It may be worth while first to recapitulate in briefest form the
evidence for the original monomeric character of (E. rubricalyx.
The original mutant gave an FI offspring of 12 plants, n of
which had red buds (R) and one green buds (r). Three of the
former selfed produced ¥2 families in which the ratios R :r were
respectively 10:5, 14:6, and 33:11. The sum of these three
families is 57:22 which is close to a 3 :i ratio and could not reason-
ably represent a 15:1 ratio, nor could either of the three ratios
individually. Two plants descended from the F* family which
yielded 33:11, a perfect 3:1 ratio were used to cross reciprocally
with (E. grandiflora, a green budded species of diverse habit.
Since these plants were from a family which was obviously mon-
omerous, the FI from the cross would either be all R (if the parent
was homozygous) or R:r in equal numbers (if the parent was
heterozygous). The numbers obtained were 30 R:28r in one
cross and 79 R: 71 r in the other. Hence the family which gave

ON SUCCESSIVE DUPLICATE MUTATIONS. 2OQ

the ratio 33:11 was undoubtedly monomerous and up till that
time a single mutation had taken place involving only one pair
of chromosomes.

It was anticipated that the F2 from rubricalyx X grandiflora
and its reciprocal would again yield 3:1 ratios but it was found
that in fact there were other ratios as well, the chief of which
now appear to be 2:1, 4:1, 5:1 and 15:1. In my discussion of these
extensive results1 I was at first inclined to attribute them to an
effect of the g andiflora parent in modifying the frequency of
inheritance of the R character, and to conclude that since the
cross with grandiflora had obviously modified the red-bud char-
acter R by dilution in many cases, it must also have modified
the frequency with which R would appear. I have since grown
a large series of Fg families, the results of which are published
in detail elsewhere.2 In the present communication a further
analysis of these F2 and F3 ratios will be made, from which it
appears that the unexpected ratios obtained in these generations
are probably not an effect of the cross with (E. grandiflora,
but they result in part from the subsequent occurrence of dupli-
cate mutations in rubricalyx. Other ratios, such as 5:1 cannot,
however, be fully explained in this way.

In this connection it seems desirable to point out that in the
inheritance of any character there are two features to be taken
into consideration: (i) The nature of the character, and (2)
the mechanism of its distribution in the germ cells. Mendelian
writers frequently ignore the former, and biometrical writers
vitiate their case when they take no account of the latter; but
in a complete account of the inheritance of any character both
must be considered. As a matter of fact, although crossing with
grandiflora probably does not modify the mechanism of trans-
mission of R, yet it does seriously and permanently modify the
character itself in some cases, as I have shown in previous
publications.

We may now consider the ratios R:r in the F2 and F$ of (E.
rubricalyx X grandiflora and the reciprocal. A further study of

'Gates, R. R., 1914, "Breeding Experiments which Show that Hybridization
and Mutation are Independent Phenomena," Zeitschrift /. Abst. u. Vererb., n:
209-279, Figs. 25.

2 "The Mutation Factor in Evolution," pp. 254 ff.

210

R. RUGGLES GATES.

these ratios makes it evident that they nearly all fall remarkably
close to three or four ratios. So close is the fit that it seems
probable that several ratios, such as 5:1, are significant as such,
though at present no complete explanation of them can be of-
fered. I was formerly inclined to regard some of these ratios
as the expression of merely quantitative differences without

TABLE II.

Fz (Oe. grandi flora X rubricalyx}.

Ratios.

Expectation.

Agreement.

Conclusion.

68 :

16

f 63.00 : 21
\ 67.20 : 16.8

3 : i possible
4 : i very near

4 : I

142 :
133 :

IS
4

147.00 : 10
128.40 : 8.6

15 : i
IS : i

15 = I
15 : I

Total 275 :

19

275.60 : 18.4

15 : i perfect
Hence 2 families 15 : i
i family 4 : i

¥2 (Oe. rubricalyx X grandiflora).

(a) 66 : 13

f 59.25 19.75
\6s.84 13.16

3 : I
5 : i perfect

5 : I

(b) 45 = 14

44.25 14.75

3 : i nearly perfect

3 : i

(c) 47 : 3

Incomplete. In addition

9 dwarfs, i intermedi-

ate.

(b) 134 : 44

133.50 44-5

3 i perfect

3 : i

(a) 67 : 13

f 66.70 13.3
\ 60.00 20

5 i perfect
3 i unlikely

5 : i

r 79.20 15.8

5 i very near

(a) 82 : 13

J 71.25 23.75

3 i very unlikely

5 : i

1 89.00 6

15 i "

r 76.70 15.3

5 i perfect

(a) 77 IS

^ 69.00 23

3 I unlikely

5 : i

I 86.25 5.75

15 i very unlikely

/ 45 14

I 134 44

(b)i?9 58

177-75 = 59-25

3 : i wry close

3 : i

66 13

67 13

82 13

77 15

("288.30 57.7

5 i very close

5 : i

(a) 292 54

^ 259.50 86.5
I 324.40 21.6

3 i unlikely
15 I impossible1

Hence in Fz

2 families 3 : i

4 5:i

o constant

1 By "impossible" is meant that the chances against this interpretation, taken
in connection with the other results, are so great that for practical purposes it
need not be considered.

ON SUCCESSIVE DUPLICATE .MUTATIONS.

211

Fa (Oe. grandiflora X rubricalyx).

Ratios.

Expectation. Agreement.

Conclusion.

f 229.60 57.4 4 i very near

4 : I

231 : 56

J 239.20 47.8 5 i ?
191.30 95.7 3 i impossible

L268.IO 18.9 15 i

("234.40 58.6 4 i very near

4 : i

237 : 56

I 244.20 48.8 5 i ?
I 219.75 73-25 3 i improbable

^274.70. 18.3 15 i impossible

r 464. oo 116 4 i very near

4 : i

Total 468 : 112 •{ 483'3° 9°'7 5 I ""Probable
| 435.00 145 3 i improbable

^543-75 36.25 15 i impossible

Hence Fs (grandiflora X

rubricalyx)

2 families 4 : i

Also

4 families constant R

3 constant r

2 intermediate

in pigmentation of

buds.

Fs (Oe. rubricalyx X grandiflora).

r 58.70

29-3

2

i very near

2 : i

57

3i

•< 66.00

22

3

i ?

(.82.50

5-5

IS

i impossible

2

3

?

22

6

21. OO

7

3

i very near

3 : i

112

69

r 120.70

60.3

3

i impossible

\I3S-7S

45-25

2

i near?

2:1?

12

2

?

Ci76.oo

35

5

i near

182

29

4 168.80

42.2

4

i impossible

5 : i

1 197. 80

13.2

15

i impossible

4

2

?

55

2

53-40

3-6

15

i close

15 i

37

2

36.60

2.4

J5

i perfect

15 i

72

6

73-10

4-9

15

i very near

15 i

59

I

56-25

3-75

15

i not very close

15 i

Total 223

II

219.40

14.6

15

i very close

15 i

r 77-30

38.7

2

i very near

2 I ?

79

37

-s 87.00

29

3

i ?

1 108.75

7-25

15

i impossible

5

3

•)

80

34

(85.50
\ 76.00

28.5
38

3

2

i near
i nearer

2:i?

Hence in Fs rubricalyx X

grandiflora

Probably

4 families 2 : i ratio

I family 3 : i

i 5 = 1"

4 families 15 : I

Also

2 families constant for R

8 constant for r

i family intermediate

186 plants.

212 R. RUGGLES GATES.

more precise significance, but particularly the ratios 5:1 and
15:1 in addition to 3:1 fall so closely into definite categories
that the probability of there all being significant is great. The
foregoing table contains the F2 and F3 ratios for the various
families.

Considering these data as a whole, the ratios nearly all fall
remarkably close to whole numbers. In the first column of
Table II. are given the actual ratios obtained, in the second
column the expectation for different ratios, and in the third
column the conclusion as to the ratios probably represented in
each case. Many of the families are larger than these of Nilsson-
Ehle, and in general they appear to fit the various ratios more
closely.

Considering first the F2 families, it will be seen that those
whose ratios are 3:1 or 15:1 are in perfect or almost perfect
accord with expectation. This being the case, it seems probable
that the 5:1 and 4:1 ratios obtained are significant as such, and
in any case they cannot be considered merely wide departures
from 3:1. Of the 5:1 ratios the first, second, and fourth are in
perfect agreement with 5:1 while the other one is very close, as
is also the sum of these four families (292 154) . The significance of
these facts is further heightened by the fact that three of these
four families (the first, third and fourth) are derived from selfing
different flowers of the same FI plant. This is shown in my
original record of these experiments,1 and it almost forces the
conclusion that in this particular plant as well as others giving
similar ratios, R gametes were being produced with greater
frequency than r gametes in the ratio 5:3. There is, however,
another explanation which will be considered later.

In the results of Nilsson-Ehle, on the other hand, the ratios
do not fall clearly into such intermediate categories but tend to
form a continuous series of ratios as Kajanus pointed out. Thus
in one series of crosses2 between black and white glumes involving
only monohybrid ratios, the F2 ratios in the 13 families actually
range from 2.2:1 (323:144) to 4.1:1 (230:56), yet the total (2468:
795) is fairly close to 3:1. One of these families contained 86

1 Zeitschr.f. Abst. u. Vererb., n, p. 236.

2 " Kreuzungsuntersuchungen," I, p. 18.

OX SUCCESSIVE DUPLICATE MUTATIONS. 213

black (B): 22 white (B). F3 offspring were grown from each of
these 1 08 plants and the results showed their composition to
have been as follows: 36 BB:5oBb:22 bb. From this result
Nilsson-Ehle concludes that there was a preponderance of
"black" gametes over "white" ones. But a series of F3 families
in another cross gave the reverse condition, 26 BB:6oBb: 33 bb,
from wrhich the conclusion is drawn that wrhite gametes were
here more numerous than black ones. Even though these re-
sults offset each other yet they cannot be referred merely
to chance fluctuations in ratios. But no further explanation
of them was offered. It will be shown later that in my crosses
of (E. rubricalyx and (E. grandiflora these deviating ratios do not
offset each other but are all consistent with the hypothesis
that R gametes are being produced with greater frequency
than r gametes.

Returning now to Table II. the first ratio (68:16) is not a very
bad fit for 3:1, although exceedingly close to 4:1. It might
easily pass for 3:1 without further comment were it not for the
fact that two other ratios in this table are in very close agree-
ment with 4:1 while they depart very widely from 3:1. The
sum of these two ratios (468:112) is very close to 4:1 while it is
highly improbable as a 3:1 ratio, the more so since the actual
3:1 ratios are almost in precise agreement with expectation.

Among the three ¥% families from (E. grandiflora X rubricalyx,
two show a 15:1 ratio and one a 4:1 or perhaps a 3:1 ratio. Hence
it might be supposed that the rubricalyx plant which was used
as pollen parent, already possessed duplicate factors for red.
But this was not the case. That plant was in fact heterozygous
for a single factor, since when crossed with grandiflora it gave
an FI of 79R : yir, which is as near to equality as could be ex-
pected. The full history of the rubricalyx individuals used for
this and the reciprocal cross has been given in pedigree form in
another paper,1 to which reference should be made. It may be
said that in both cases they are descended from the family which
contained 33R : nr. One member of this family was pollinated
by nanella and produced a family of 42 plants. One of the latter
(No. IV., 2), which was a perfect rubricalyx in appearance but

1 Gates, Zeitschr. f. Abst. u. Vererb., u : opp. p. 216 and on p. 217.

214 R- RUGGLES GATES.

carried dwarfing latent, was used to pollinate (E. grandiflora.
Since the F% offspring of this cross gave 15:1 ratios in two
families, while the FI was a 1:1 ratio, duplicate mutations must
have intervened between these two generations. The two
plants which were the parents of the families containing 142 R:
15 r and 1 13 R: 4 r respectively must have possessed the duplicate
factor in all their germ cells, so that they were heterozygous for
R and R'. Their composition might then be written RrRV.

As pointed out earlier in this paper, such a condition might
have arisen (a) through the transformation of a chromosome
belonging to a second pair, (b) through an exchange of mates
on the part of two pairs of chromosomes. We may now examine
the comparative credibility of these two alternatives. There
are certain difficulties with either hypothesis, one of which is
that the transformation from the monomerous to the dimerous
condition, whether effected by chemical or mechanical means,
must apparently have taken place early in the ontogeny, before
definitive germ cells are formed. The alternative hypothesis
would be that all the germ cells had undergone the transformation
simultaneously and independently, which one cannot believe
possible.

There is, however, one consideration which makes it appear
probable that the duplicate condition for R is not usually arrived
at through a transformation of a new chromosome, but rather
through a redistribution of the chromosomes. The 15:1 ratio
can only be obtained from an RrRV parent, in which both
duplicate factors are heterozygous. It would therefore be
necessary to assume when a 15:1 family is derived from a 3:1
family, that a chromosome belonging to a new pair had under-
gone a chemical transformation while its mate and the mate of
the original modified chromosome were unaffected, i. e., that
the condition RrrV became altered directly to RrRV. This
is very unlikely. On the other hand, as I showed long ago,1
the chromosomes in (Enothera are very loosely paired during the
reduction division, and moreover irregular chromosome distri-
butions have been shown to occur at this time (as in the pro-
duction of (E. mut. /a/a). I also (/. c.) pointed out the probability

1 "A Study of Reduction in QLnothera rubrinervis," Bot. Gazette, 46: 1-34, pis. 3,
1908.

ON SUCCESSIVE DUPLICATE MUTATIONS. 215

that exchanges of chromosomes of different pairs but without
change in number would take place under these conditions.
By such a mismating or exchange of mates on the part of two
chromosome pairs, a plant which was homozygous (RR) for one
factor would give rise to plants which were heterozygous for
duplicate factors (RrRV). This is then what has probably
occurred in the cases where plants from a 3:1 family have given
rise to 15:1 ratios. The frequency with which such mismating
occurs in CEnothera may thus be estimated.

It is known that the chromosomes of CEnothera are in pairs
(doubtless of paternal and maternal origin) throughout the
somatic divisions, and the paired arrangement is probably a
feature of the first mitosis after fertilization. If, then, a plant
which would have been homozygous for a single factor (RR)
becomes transformed into one which is heterozygous for dupli-
cate factors (RrRV) and so gives a ratio 15:1 in its offspring,
the most likely assumption is that at the time of fertilization the
two R chromosomes, instead of becoming paired with each other,
each paired with another (r) chromosome. Hence in this case
the regrouping of chromosomes probably occurred not during
meiosis where it would have to occur simultaneously in all the
germ cells, but as a feature of fertilization or the first mitosis of
the embryo. It will be shown later, however, that mismatings
of the chromosome pairs probably also occur during meiosis and
so modify the 3:1 ratio. To sum up, it appears that when a
15:1 family is derived directly from a plant in a 3:1 family, the
remating of the chromosomes must have occurred at fertilization
or soon afterwards; but when, for example, a 4:1 or a 5:1 family
is derived from a 3:1 family, this may be accounted for by a cer-
tain amount of remating of chromosomes during meiosis.

The method above described will also apply to the origin of
duplicate and triplicate factors in wheat and is perhaps more
probable than the successive chemical transformation of dif-
ferent chromosomes. There is, however, a method of testing
between these two possibilities. If the duplicate condition
arises through a regrouping of the chromosome pairs, then, as
has been mentioned, a race or a plant homozygous (RR) for one
factor will give rise to a plant heterozygous for two factors

216 R. RUGGLES GATES.

(RrRV). On the other hand, if the chemical transformation
of a fresh chromosome takes place in a homozygous monomeric
plant (RR), then the dimerous individual derived from such a
monomerous plant should have the constitution RRRV.1 It
would be possible to determine between these two alternatives
by breeding tests. If the constitution of the plant is RrRV its
offspring should give a 15:1 ratio. If it is RRR'r' they would
all be red in FI and F2. But plants having the former formula
could also be produced by the mismating of chromosome-pairs
during meiosis in RR plants.

If we now return to the table (p. 210) and examine the F2 from
the reciprocal cross (rubricalyx X grandiflora) we find a total
absence of 15:1 ratios, showing that not only was the rubricalyx
parent of this cross monomerous but its offspring remained so.
The parent of this cross was a member (No. IV., 8) of the monom-
erous family 33:11. As will be seen from the table, two of the
F2 families from rubricalyx X grandiflora gave perfect or almost
perfect 3:1 ratios. Four others gave 5:1 ratios, three of which
were perfect and the other very close to expectation as already
pointed out. I have at present no further explanation of these
5:1 ratios to offer, but it seems probable that their significance
will later become apparent.

Ratios more or less in excess of 3:1 could be obtained from
plants homozygous for one factor, if there was a tendency for
mismating of the chromosomes in meiosis. But this will not
account for the definiteness of the 5:1 ratios obtained.

Turning to the F3 of grandiflora X rubricalyx the full data are
given in my book (p. 255). Four families were constant for
R, 3 constant for r, 2 families numbering respectively 283 and 20
plants bred true to an intermediate condition, and 2 families
split in the ratio 4:1, as shown in the table (p. 21 1). The excess
of R's in the last two families is a significant excess over 3:1,
however it is brought about.

In the F3 of rubricalyx X grandiflora, four families give ratios
nearest 2:1, one family near 3:1, one near 5:1 and four very close
to 15:1. Whatever the significance of the 2:1 and 5:1 ratios in

1 We have already found it highly improbable that a plant Rr could be directly
transformed chemically into RrR'r', since we should anticipate that the chromo-
some r would undergo a mutation before the chromosome r'.

ON SUCCESSIVE DUPLICATE MUTATIONS. 2IJ

these families, the appearance of 15:1 ratios in the Fs of this
cross is of much interest, since the F% contained no families which
could reasonably be construed as containing duplicate factors,
except the one having the incomplete ratio 47:3. Reference to
the pedigree numbers1 shows that the first two are derived
from the FZ family No. 60 in which the ratio is doubtful, the
third is derived from selfing a plant in the Fg family No. 62,
and the fourth from selfing one in family No. 63. In these two
families the ratios were respectively 67:13 and 82:13, both of
which are shown (p. 210) to be very near 5 :i ratios. The appear-
ance of these 15:1 ratios in F^ from 5:1 families can be explained
if we assume that independent duplicate mutations have occurred
in the F% families 60, 62, and 63. This must happen as pre-
viously outlined, through a plant which is homozygous for one
factor giving rise to a plant which is heterozygous for two; or in
other words, through the rearrangement of a pair of homologous
chromosomes so that they belong to different pairs.

Another point which will be explained by the present hypoth-
esis is the difference in the depth of color in homozygous red-
budded races. Thus in the FS families 93 and 95, - containing
respectively 280 and 312 plants, the latter were constantly
darker red than the former. The latter family was doubtless
homozygous for duplicate factors (RRR'R/), or at least RRRV,
since the family from which it was derived yielded 15:1 ratios.
The former family was on the other hand probably homozygous
for a single factor (RR) and hence not so densely red-pigmented.

It will thus be seen that in several instances 15:1 families
have been obtained from the offspring of 3:1 or 5:1 families.
All such cases can be explained by assuming that a duplicate
mutation has intervened. The original mutation by which deep
red buds in (Enothera first appeared is an extremely rare occur-
rence, having occurred but once in all cultures of (Enothera.
When, however, a chromosome has once undergone this change
it is reasonable to suppose that other chromosomes in the same
nucleus could without difficulty take on an analogous trans-
formation. The whole mechanism is, however, at hand in the

1 See "The Mutation Factor in Evolution," p. 256.
"-L. c., p. 255.

2l8 R. RUGGLES GATES.

meiotic divisions, for transforming the original 3:1 ratio into a
15:1 by merely redistributing the chromosome pairs.

In concluding this paper it is desirable to compare the related
but different results recently obtained by Honing,1 with two
varieties of Canna indica which are naturalized in Sumatra.
One variety has green leaves while in the other the leaves have a
broad red margin.

From the offspring of plants of the latter variety Honing
obtained ratios red : green of 3 :i , 9 7 and 27 137. The same ratios
were obtained in crossing the two varieties. These ratios are
accounted for by the hypothesis that the cooperation of three
"factors" is necessary to produce the red margin. If these are
located in chromosomes belonging to three different pairs,
then the resulting ratio should be 27 red: 37 green, since the
character can only appear in the presence of all three factors
A, B, C. On the other hand, if all three factors are located in
the same chromosome a 3:1 ratio would be obtained, while if
two of them were in one chromosome and the third in a chromo-
some of a second pair, the ratio would be 9:7.

It was found that in certain cases plants in families having a
3:1 ratio gave rise in the next generation to a 9:7 or 27:37 family.
In such cases one may assume that a mutation has taken place
resulting in a redistribution of the determiners, the three which
were present in one chromosome being rearranged so that they
are in chromosomes belonging to two or three different pairs.
So far as I am aware, this is the first experimental evidence that
an actual rearrangement of the chromomeres in the chromo-
somes is one of the kinds of change which the nucleus may
undergo, the case being somewhat different from Morgan's
well-known phenomena of "crossing over" in Drosophila.
Further experiments are necessary to test the nature of this
evidence for the occurrence of mutations in which such a re-
arrangement of the nuclear material can take place.

SUMMARY.

Nilsson-Ehle was the first to show that duplicate and triplicate
factors for red are present in certain strains of wheat. He

1 Honing, J. A., 1915, " Kreuzungsversuche mit Canna- Varietaten," Rec. Trav.
bot. Neerlandais, Vol. 12: Livr. i, pp. 26.

ON SUCCESSIVE DUPLICATE MUTATIONS. 219

found, moreover, that the same strain may be in one case monom-
erous and in another case dimerous for this character; and that
while, for example, Grenadier wheat possessed three independent
units for red, Extra-Squarehead possessed only one. The origin
of the original ' ' factor ' 'for red may be accounted for in the wheats
as in (Enothera rubricalyx, through the chemical transformation
of one chromosome or a pair of homologous chromosomes. The
duplicate condition for the character R may have arisen (i)
through a chemical mutation in a second pair of chromosomes, (2)
through a re-mating of the chromosomes (RR) forming a homo-
zygous pair. The latter method is for various reasons the more
probable.

Although the original (Enothera rubricalyx was a monohybrid
and continued so for at least two generations, yet in subsequent
generations involved in crosses with (E. grandiflora, 15:1 or
di-hybrid ratios were derived from the offspring of members of
3:1 families. This can best be accounted for by supposing that
in a plant (RR) homozygous for one factor, a re-grouping of the
chromosome pairs occurred. This re-grouping involves merely
an exchange of mates on the part of the chromosomes RR so
that they now belong to different pairs. The formula for the
plant may now be written RrRV, i. e., the plant is heterozygous
for two independent units for red and its offspring will give a
15:1 ratio.

The second mutation, producing the duplicate condition for
R, is thus probably a purely mechanical process, while the origi-
nal mutation which produced the "factor' R is a chemical
change of wholly different nature. It is possible that in some
cases the duplicate and triplicate conditions also arise through
the chemical transformation of additional chromosomes.

When a 15:1 family arises from a 3:1 or 5:1 family, as has
happened several times in (E. rubricalyx hybrids, it is necessary
to assume that the regrouping or remating of chromosome pairs
which led from the monohybrid to the dihybrid condition, took
place at fertilization, or at any rate early in the ontogeny, and
is then handed down to the germ cells by mitosis. The chro-
mosomes are known to be paired in the somatic divisions, and
it seems probable that the manner of pairing set up in fer-

220 R. RUGGLES GATES.

tilization continues in this case throughout the ontogeny, though
this is not true for all organisms. Otherwise it would be necessary
to assume that when a plant in a 3:1 family gives rise to a 15:1
family all its germ cells have simultaneously undergone a mis-
mating of the chromosome pairs during meiosis, a highly im-
probable event.

In the F2 and F3 hybrids of CE. rubricalyx and CE. grandiflora,
in addition to 3:1 and 15:1 ratios, 2:1, 4:1 and 5:1 ratios occur.
The 5:1 ratios at least seem to be significant, indicating that
R and r gametes are regularly being produced in the ratio 5:3,
or that a certain amount of re-grouping of the R chromosomes
is regularly occurring during meiosis.

Vol. XXIX. October, 1915. No. 4

BIOLOGICAL BULLETIN

REACTIONS AND RESISTANCE OF FISHES IN THEIR

NATURAL ENVIRONMENT TO ACIDITY,

ALKALINITY AND NEUTRALITY.

MORRIS M. WELLS.

PAGE.

I. Introduction 221

II. Apparatus and Methods 222

A. Reaction Experiments 222

B. Resistance Experiments 225

III. The Water 225

IV. Presentation of Data 231

A . Reaction Experiments 231

1. Reaction to Acids 231

(a) To Carbonic Acid 231

(b) To Sulfuric Acid 232

(c) To Acidity in Distilled Water 234

2. Reaction to Alkalies 235

(a) In Neutral Water 235

(b) In Strongly Acid Water 235

3. Conclusions Drawn from the Reaction Experiments 239

B. Resistance Experiments 239

1. Resistance to Acids 239

2. Resistance to Alkalies 240

3. Resistance to Neutrality 240

V. General Discussion 243

VI. General Conclusions 254

VII. Acknowledgments and Bibliography 254

I. INTRODUCTION.

The present paper is the first of a series that is to deal with
the relation of fishes to ions in the natural environments. It
is purposed to point out some of the close correlations which
exist between the physiology of fishes and their behavior, and
to present evidence concerning the importance of such corre-
lations in biological investigation in general. The data pre-
sented in the following pages deal with the reactions and re-
sistance of fresh water fishes to acidity, neutrality and alkalinity;

221

222 MORRIS M. WELLS.

the discussion of the data shows that the phenomena outlined
receive much support from the work of other investigators and
that the environmental factors which are important to fresh
water fishes are probably of importance to many, if not all,
other organisms as well.

The investigation has been carried on at the University of
Illinois, in Professor V. E. Shelford's laboratory. The work
has been done in connection with another line of inquiry regarding
the reactions and resistance of fishes to salts. The results of
this second investigation will appear as the second paper of
the series.

II. APPARATUS AND METHODS.

Two general types of experiments have been run, namely,
reaction experiments, and resistance experiments.

A. REACTION EXPERIMENTS.

This method of experimentation was devised by Shelford
and Allee ('13) and may be designated as the "gradient method."
In brief the procedure is as follows: A solution gradient is es-
tablished in an observation tank, the fish introduced, and its
movements graphed. The graph, together with notes taken at
the time, makes up the experimental record. Fig. i shows
the type of tank used. A similar tank was used by Shelford
and Powers ('15) in their experiments with marine fishes. A
black hood screens the tank, the movements of the fishes being
viewed through slits in the front of the hood.

The tank has a plate glass front and is lighted by symmetrical
lights placed above. A plate-glass cover fits into the top and
rests against the surface of the water. This cover is useful in
experiments with gaseous gradients as it lessens the vertical
gradient due to escape of gas at the surface.

The water flows into the tank through the openings (inlets)
in the ends, then toward the middle; at the middle the water from
the two ends mixes, the water from each end drifting somewhat
past the middle, thus forming the gradient. The water flows
out through the exits (outlets) in the bottom and at the top of
the tank. An experiment consists of first establishing the grad-
ient, and then introducing the fish and graphing its movements.

REACTION AND RESISTANCE OF FISHES.

223

In establishing the gradient the flow at each end of the tank was
fixed at 500 c.c. per minute in practically all the experiments.
The flow of tap water was regulated to 500 c.c. per min. at one

*j

-/

)-i

01

4J

IU

M

O

•4-t

<u

n

—•

£
TJ

+->

o

0>

rn

— '

^_J

o
J2

CJ

0

<=^i i — ^=

ET^ s

•o

to

4_i

T3

"w

°

•o

\

«

OJ

"

*G

g

>

o

0

'' '' !

!|i|

i|i

II ,"l

II ''

I ' ,
1 ' 0,

1 1 ' fl

1 '

n' '

n |,i

0

' ' 1

''1 Ij

'V1

ilsk

Sty «— ^ \x /p — -> /«
SC Sc OulUls sc

rte>v

in which the reaction experiments were performed.

~
~

a

SJ
4J

o

4->

j.-.
_o

"w

1

n

•o

g

<u

0)
-t->

to
to
cS

5

O

-4->
>»

ja
•o

V

—

0

o

-t-»
OJ
— •

14-I

o

o

cd

<L>

J3
M

3
O

—
-u

O

•o

<u

o
q3

tj
ui

3
<S
0)

6

the bottom. The experimental factor was introdu<

led with a solution five times as concentrated as that

c
rt

c

•4-1

C

r-

'e
1-1
u

a
6

U

o
o

M

*0

OJ
-*-)
03

0)

r«

~

^j
ri

S

IH

.^

o;

fj
'---1

0

•4->

a
-w

OJ

4J

O

-4_>
S

0

•r
ri
<u
a

M
w

15
H

^

S

^

oj
D,

u
cj

o

c

Tj-
U-i

O

c;

*j

Fap water, or water modified to a different degree, floi

testing were withdrawn through the stopcocks (5C)

I

I®K

£ x

s

^

rt

53

D

O

}-.

%

\

M^

" 5

—

^2

w

"""'

r^

.

'-*+

i 3

XN

"^

^

cd

<u

'

C

. r^K=^

In

2J

-;

£

^

S

r - 1
!• -a "3 -°

•C <u 3 bo O o *i in

1 .S ° § S 1 -S

5 1

+» 0) "

S •

.S S § 2 S. .5 S -a S

end and to 400 c.c. per min. at the other. Then 100 c.c. per min.
of a solution of treated water was introduced into the 400 c.c.
flow from a mixing bottle (MB). The 100 c.c. flow was kept

224 MORRIS M. WELLS.

constant by using an aspirator bottle as in Fig. I (A B),
This bottle was filled with treated water, corked at the top and
placed in the jar (J). The water was siphoned from J and the
pressure was constant as the solution escaped from A B, when
the level began to fall in /. The strength of the solution in the
aspirator bottle was always five times that desired in the treated
water end of the tank.

The following variations of the simple graphing method of
recording were used by Shelford and Allee and have been intro-
duced here, (i) In experiments where the fish was decidedly
inactive and remained in one end of the tank, it was driven into
the opposite end with a rubber-tipped glass rod. The driving
was done at regular intervals and was repeated at similar in-
tervals in the controls. Dotted lines in the graphs indicate that
the fish was driven. Experiments of this sort were few in num-
ber and have for the most part been thrown out. In some ex-
periments, however, the fishes were active and yet remained
constantly in one end of the tank. Driving was again resorted
to, in some cases, to make sure that the selection of the given
end was a reaction to the gradient. A return to the original
end would indicate this to be the case. (2) A number of ex-
periments was performed with 4—10 small fishes in the tank at
the same time. These experiments were recorded by readings
taken 30 seconds apart. The readings indicate per cent, of
fishes in each third of the tank at the time of reading. (3)
•Usually the fishes were not placed in the tank until the flow at
the ends had been going for some time. Thus a gradient was
established before the fishes were introduced. In some cases,
however, the fishes were either left in the tank when the ends
were reversed, or introduced before a gradient had formed. The
results of these experiments do not differ from the others ex-
•cept in per cent, of time spent in the thirds of the tank.

The controls were blank experiments, run with untreated
water flowing in at both ends, or with no flow at either end.
Experiments with the treated water first at one end and then
at the other, also served as controls.

The gradient was determined by simple titration with stan-
dard acid or alkali, using phenolphthalein or methyl orange

REACTION AND RESISTANCE OF FISHES. 225

(or both) as indicator. The samples were collected by means
of a pipette inserted to a given level below the surface, and the
titrations were always made at once, with as much care as seemed
necessary. For instance samples containing a high concen-
tration of CO2 need to be titrated with rapidity, while samples
containing H2SO4 or KOH may be titrated without haste.

B. RESISTANCE EXPERIMENTS.

The procedure in the resistance experiments was very simple in
most cases. In general the desired solutions were made up from
standard solutions of the acid or alkali (measured from burettes)
and the fishes introduced. Temperature was controlled by
setting the jars containing the solutions, in running tap water.
As the experiment proceeded, samples for testing were with-
drawn when necessary, and the same amount of water was re-
moved from the control. General controls were kept running
throughout the entire time, while numerous temporary controls
were set up as demanded by individual experiments.

The species of fishes used principally, have been the blue-gill
(Lepomis pallidus), white crappie (Pomoxis annularis), green
spotted sun-fish (Lepomis cyanellus), and bull-head (Ameiurus
melas). Most of the fishes were caught (by seining) in the small
streams in the vicinity of the university (the crappie came from
a small artificial lake) ; all were brought into the laboratory
with as little handling as possible, and placed at once in large
aquaria. They were fed from day to day but fishes do not
always eat well in confinement and as time went by they became
more or less starved. The changes in the reactions of fishes,
which accompany starvation have been investigated and will
be discussed in another paper.

The chemicals used have been the chemically pure prepara-
tions of Kahlbaum or the analyzed preparations of Baker.

III. THE WATER.

An investigation of the reactions of fishes to salts in solution
was begun at Chicago in 1912, at the suggestion of Dr. Shelford.
In the fall of 1914 Dr. Shelford left Chicago to take a position at
the University of Illinois and the writer accompanied him to

226 MORRIS M. WELLS.

continue the work in his laboratory at that place. The differ-
ences in the water supply of the two institutions brought up a
number of new questions regarding the reactions of the fishes,
and it was decided that to continue the investigation satisfac-
torily, more must be known of the effects of acid and alkaline
water upon fishes and their reactions. This second investigation
was therefore taken up and the results are published in advance
of those with the salts, since they bear directly upon the inter-
pretation of the latter. A brief comparison of the water of the
two institutions will be profitable at this point.

The water at Chicago is pumped from Lake Michigan and
analyses show it to be considerably different in gaseous and
solid content from the water at Illinois, which comes from deep
wells. In 1912 Alice, who had been working at Chicago on
rheotaxis in isopods, came to Illinois, bringing with him a stock
of animals. In his paper ('13) he compares the waters of the
two institutions, and gives a table showing the differences in
the dissolved content. This table is inserted here.

TABLE I.

A COMPARISON OF CHEMICAL ANALYSES OF CHICAGO AND UNIVERSITY OF ILLINOIS

TAP WATER.

Analysis of solids in parts per million, and gases in cubic centimeters per liter.

Chicago Tap. U. of I. Tap.

Potassium, K 6.0 2.6

Sodium, Na 42.1 29.0

Ammonium, NH4 0.04 2.3

Magnesium, Mg 11.3 34.9

Calcium, Ca 34-6 70.1

Iron, Fe O;I5 i.o

Aluminum, Al o.oo 1.3

Silica, Si 3.3 18.9

Nitrate, NOs 1.7 0.7

Chlorine, Cl 12.0 3.5

Lead, Pb o.oi o.oo

Sulphuric acid, SOs 0.04 2.3

Oxygen, 0 10.46 0.12

Free carbon dioxide, COa 2.5 18.0

Half bound CO2 (bicarbonates) 32.5 101.12

Alice states that the change of water did not greatly affect
the rheotactic response of the isopods. At Illinois he kept the

REACTION AND RESISTANCE OF FISHES. 22J

animals in aerated water, which was thus saturated with oxygen
(5-5 to 7 c.c. per liter) while the free carbon dioxide was removed.
If aerated sufficiently, the Illinois tap water becomes alkaline
to phenolphthalein, and upon writing to Allee with regard to
this matter he gives me permission to state that he kept his
stock of isopods in such alkaline water for a period of 22 weeks,
without increased mortality. He points out that the per cent,
of rheotactic response, after a large number of trials, was 8 per
cent, less than at Chicago but does not know whether or not to
regard this as significant.

Table I. shows that Illinois water contains 18 c.c. of CO2
per liter and practically no oxygen. Either of these conditions
would alone prove fatal to fishes, while the combination would be
doubly fatal (Wells, '13). Since aeration removes the CO2 and
at the same time saturates with oxygen, it was thought that this
would fit the water for supplying the fish aquaria. The water as
it came from the tap was therefore run through the aerating pans
which form a part of the apparatus described by Shelford and
Allee ('13). The device consists of a series of galvanized pans,
placed one beneath the other. The water runs into the upper
pan and trickles down through successive pans into a galvanized
tank. From the tank, pipes lead to the aquaria. The flow
into two large aquaria was regulated to 500 c.c. per minute for
each and the fishes were now brought in from the nearby streams
and placed in the aquaria in rather large numbers. The aquaria
were 8 ft. x 2 ft.; about 300 small fishes were placed in each.
This was overcrowding, but fishes have been kept successfully
for some time, in closer quarters at Chicago.

The immediate mortality of the stock was not great. It was
noted that the darters and other more sensitive fishes did not
live well but the sunfishes, bullheads and minnows seemed to be
normal. In a few days, however, these fishes began to die.
It was thought that the water contained too large an amount of
carbonates and an arrangement was made to introduce sulfuric
acid into the aerated water at the galvanized supply tank.
Enough acid was added to convert about one third of the carbon-
ates into sulfates and with some benefit. It had been noted in the
experiments that the fishes did not swim about as actively as

228 MORRIS M. WELLS.

usual and that their sensitivity seemed to be lessened, as they
would swim into factors to which they are normally very nega-
tive. It was at this point that the study of the effects of acidity,
etc., was decided upon.

Tests showed that the water entering the aquaria was practi-
cally neutral to phenolphthalein, varying a little from day to
day. To determine the effect of the neutral water upon the
fishes, a number was taken from the aquaria and placed in tubs
of partially aerated tap water (water contained 6-10 c.c. COa
per 1.). After a day or so in this water, they began to behave
normally in the gradient again. The flow of water into the
aquaria was now modified by diminishing the amount of aeration.
The tap water was run down a wooden trough 12 ft. long, into
the aquaria. This saturated it with oxygen but left it decidedly
acid with CO2. From now on the mortality of the stock of
fishes was very low. The aquaria were not so crowded as at
first, but that the decrease in the number of fishes does not ex-
plain the low mortality, will be brought out in experiments to
be presented later.

The importance of the chemical reaction of the water to fishes
had been foreseen (Wells, '13, p. 337) and it was decided that
the peculiar properties of the Illinois water offered an excellent
opportunity for continuing this investigation. At the same
time it was thought that the work might perhaps throw some
light upon a number of the reactions of fishes to salts, which
seemed difficult to explain.

The advantages of the Illinois water are due to the following
chemical properties: As it flows from the tap it is acid to phenol-
phthalein from the excess (18 c.c. per liter) of COo, and alkaline
to methyl orange because it contains a large quantity (101 c.c.
per liter) of bicarbonates in solution. The bicarbonates have
been formed from carbonates according to the equation CaCOa
+ HoCOa t^ Ca(HCO3)2 and when carbonates are dissolved
under the influence of excess of carbonic acid they are practically
all converted into bicarbonate, the quantity of unconverted
carbonate being negligible (Stieglitz, '09, p. 246, Seyler, '94, p.
105). Under the pressure in the water pipes, there exists an
equilibrium between the carbonic acid and the bicarbonates,

REACTION AND RESISTANCE OF FISHES. 229

hut when the water flows out of the tap, the pressure is removed
and the carbonic acid at once begins to dissociate into CO2 and
water. The CO2 passes off into the air and the dissociation
of the acid continues until equilibrium with the CO2 in the
atmosphere is established. Parallel with the dissociation of
the carbonic acid there goes an increasing tendency for the bicar-
bonates to break up to form the normal carbonate, and by the
time the acidity from the carbonic acid has diminished to ap-
proximate neutrality, the bicarbonates are producing a sufficient
quantity of the normal carbonate to give the water an alkaline
reaction to phenolphthalein. Thus by regulating the amount
of aeration, the water can be left acid, made neutral or even al-
kaline.

Biologists speak of the carbonates, bicarbonates, and car-
bonic acid, as fixed, half bound and free CO2, respectively. The
fixed is that existing as simple carbonates, the half bound that
necessary to convert the carbonates into bicarbonates, and the
free that remaining in excess (Seyler, '94, p. 104). It, will be
seen that the bicarbonates contain both fixed and half bound
CO2, i. e., CO2 which is to become half bound is added to CO2,
that is already fixed, to form the bicarbonates. Failure to
recognize this fact often leads to confusion when these terms
are used.

The amounts of the three kinds of CO2 can be determined
accurately by titration, using two indicators, phenolphthalein
and methyl orange. Methyl orange is unaffected by H2COs
and hence the bases present as carbonates or bicarbonates can
at once be titrated with acid. Carbonates are alkaline to phe-
nolphthalein, bicarbonates are neutral, and free CO2 is acid.
A carbonate titrated with acid, therefore, becomes neutral to
phenolphthalein (if titrated under conditions which prevent
loss of CO2) when the carbonates have all been converted into
bicarbonates.

Methyl orange is not affected by H2CO3 because this acid
does not produce a high enough concentration of H ion. The
indicator is however very sensitive to OH ion and reacts to the
minute amounts that are present in a bicarbonate solution.
Phenolphthalein, on the other hand, is very sensitive to H ion

230 MORRIS M. WELLS.

but not to OH ion. It therefore gives an acid reaction with
CO2 but is unaffected by the minute amount of OH ion which
is present in solutions of bicarbonates.1 Methyl orange will
give an alkaline reaction in water in which the concentration
of H ion is considerably greater than that of OH ion. Thus,
in the presence of CO2, titration with this indicator is not a
determination of true alkalinity for the water is as a matter
of fact acid, since it contains a higher concentration of H than
OH ions. Marsh ('07) makes this error when he states (p. 337)
that "the reaction of water which will support fish life must be
slightly alkaline." His determinations were made with sulfuric
acid, using methyl orange as indicator. The water to which he
refers ("Potomac service water") was in all probability acid to
phenolphthalein. Marsh also states that "when the water
becomes even slightly acid, fishes cannot live in it." This
would mean that fishes can not live in water which has been
made slightly acid to methyl orange by the addition of an acid.
I have added sulfuric acid to tap water until it gave an acid
reaction to methyl orange, and find that fishes live in it as well
as in the original tap water, i. e., normally. The fishes should
not be placed in such water until some little time after the adding
of the sulfuric acid, however, for in the process of changing the
carbonates to sulfates, a large amount of carbonic acid is liberated
(CaCO3 + H2SO4 5 CaSO4 + H2CO3) and until this carbonic
acid has dissociated and the CO2 passed off into the atmosphere
to a large degree, its presence will kill fishes which may be in-
troduced. The amount of carbonic acid formed will depend
upon the amount of carbonates in the water. (The reaction
with bicarbonates will give the same result.)

In the following pages I shall introduce experimental data to
show that fresh water fishes cannot live normally in water that
is alkaline but that they require a certain degree of acidity to
carry on their normal activities.

1 Noyes ('13) gives a table (p. 388) showing the acidity or alkalinity of solutions
at the change of color for various indicators. Methyl orange gives an alkaline
reaction when the OH concentration is only io~9 while phenolphthalein is not
affected until the concentration of OH ion reaches io~5. Methyl orange gives an
acid reaction when the H ion concentration is io~4 while phenolphthalein reacts
when the H ion concentration is only a little more than io~8.

REACTION AND RESISTANCE OF FISHES. 23!

IV. PRESENTATION OF DATA.

The following experiments show the effect of different degrees
of acidity and alkalinity, upon the reactions and longevity of
fresh water fishes.

A. REACTIONS OF FISHES TO ACIDITY AND ALKALINITY.
i. Reaction to Acids.

(a) To Carbonic Acid. — A number of experiments was run
to determine the reactions of the fishes to this acid. Three
degrees of acidity were used for the most part: (i) Neutral, to
very faintly acid (aerated water) (2) moderately acid (8-10 c.c.
CO2 per liter; obtained by using half and half mixture of I and
3); (3) strongly acid water (unaerated tap; 18 c.c. CO2 per liter).

(1) Moderately Acid Water vs. Strongly Acid Water (Graph i,
Chart /.). — The fishes selected the lower acidity with much
precision. They also spent much time at the surface, as is
characteristic when the concentration of CO2 is high.

(2) Slightly Acid vs. Moderately Acid Water (Graph 2, Chart /.).
-The fishes were left in the tank, and the flow altered so that the

moderately acid water ran into the end that had previously been
strongly acid and neutral water was run into the opposite end.
The fishes were graphed after five minutes. They definitely se-
lected the end of the tank into which the neutral water was
flowing. Test showed this end to contain 3 c.c. CO2 per liter.
Seven experiments with this combination were run and all gave
similar results. Variations were due to specific and size dif-
ferences. The larger fishes and the crappies and green spotted
sunfishes selected a somewhat higher acidity than the smaller
fishes, especially the blue-gills.

(3) 6 c.c. COz per Liter vs. 4 c.c. per Liter (Graph j, Chart /).—
The concentrations of CO2 were obtained by regulating the
amounts of aerated and unaerated water. Six experiments were
run. The bullheads and blue-gills selected the lower concen-
tration with precision, wrhile the sunfishes and crappies chose
the higher end with as much definiteness. Thus, as was seen
in (2), the species differ in the optimum CO2 concentration
which they select at this time of year.1 The difference in

1 Because of the fact that the resistance of fishes varies with the season (Wells,
'14), it is very probable that the COs concentration selected by a given species will
show seasonal variations also. This point is yet to be investigated.

232 MORRIS M. WELLS.

specific reaction may be in part a matter of size, as the crappies
and sunfishes averaged larger than the blue-gills. Small sun-
fishes were, however, found to be less sensitive to CO2 than were
blue-gills of the same size. In this case the reaction is correlated
with resistance as the sunfishes are more resistant than the blue-
gills. The bullheads, however, are perhaps the most resistant
of our fresh water fishes yet they are very sensitive to CO2.
The sensitiveness of the bullheads is probably related to the
peculiarity of their integument which (Herrick, '02) has chemical
perceptors "taste buds" scattered over its entire surface.

So far the fishes had for the most part selected the end of the
tank containing the largest proportion of neutral water, i. e.,
the lowest acidity. To ascertain definitely the reactions to the
neutral water, the following experiments were performed.

(4) Slightly Add (j c.c. C02 per Liter) vs. Neutral Water.} —
The slightly acid water was obtained by partially aerating the
water which flowed into one end. This was done by running it
through a galvanized tank which was a part of another piece
of apparatus. The gradient tested practical neutrality at one
end, and slight acidity at the other. The fishes definitely se-
lected the end containing the CO2 and were thus negative to
the neutral water.

(b] Reactions to Sulfuric Acid. — It should be pointed out
that experiments where other acids than H2CO3 are added to
water, which contains bicarbonates, are open to misinterpre-
tation if an attempt is made to compare the reactions of the
animals to the acids, in this way. The addition of a strong
mineral acid to such water does not result in the presence of a
hydrogen ion concentration from the mineral acid itself, until
all the bicarbonates have been decomposed (i. e., changed to
sulfates, etc.) or, in other words, not until the water has become
acid to methyl orange. The reaction is a double decomposition,
H2SO4 + Ca(HCO3)2l;CaSO4 -f 2H,CO3. Until this reaction
is completed, the chief result of adding the mineral acid
will be to increase temporarily the concentration of carbonic
acid by liberating the fixed and half bound CO2. The concen-
tration of H ion from this weakly ionized acid will be but a small
per cent, of that which would have been obtained from the min-

REACTION AND RESISTANCE OF FISHES. 233

eral acid itself; yet this will be the only hydrogen ion supply
until the bicarbonates have all been changed to sulfate (if
sulfuric acid is the mineral acid used). The tap water at Illinois
requires approximately 90 c.c. of .1 TV H2SO4 to neutralize the
bicarbonates in one liter of water while the Chicago water re-
quires about a third as much. According to the above equation
one molecule of the acid liberates 2 molecules of COo. There-
fore 90 c.c. of .1 TV acid will liberate 210 c.c. of COz. Powers
('13) did not take this reaction into account and speaks of com-
paring the reactions of crayfishes in gradients of HC1 and CCV
As a matter of fact the amounts of HC1 which he added to the
Chicago tap water were probably used up in neutralizing the
bicarbonates. Thus all his gradients were with carbonic acid,
and the differences which he gets are due to the liberation of
this acid in excess, in the reaction of the HC1 and the bicar-
bonates.

(1) H%SOi to Neutralize all Bicarbonates vs. Neutral Water.—
The CC>2 liberated in this gradient made the water so acid that
the fishes were soon overcome, and died if not removed from the
tank. However at first they gave a decidedly negative reaction
to the acid end.

(2) H2SOt to Liberate 40 c.c. COZ per Liter vs. Neutral Water.—
The fishes reacted much as they did in gradients of aerated
vs. unaerated water. They were very negative to the acid
end.

(3) H%SOi to Liberate 4 c.c. C02 per Liter vs. Neutral Water
(Graphs 4 and 5, Chart /.). — Eighteen experiments of this sort
were run. Of the 18 graphs, 14 show that the fishes spent 90
per cent, of the time in the acid half of the tank; 3 show more
than 50 per cent, of the time in this half and I (small blue-gill)
shows an 80 per cent, preference for the neutral end. That the
fishes are negative to neutral water is thus confirmed. To
ascertain the chemical reaction of the water at the point in the
tank where the fishes turned back from the neutral end, numerous
samples were titrated from this region, during the experiment.
They showed that the water at this point contained about I c.c.
of CC>2 per liter. The graphs shown in Chart 1. are typical for
all.

234

MORRIS M. WELLS.

Bluegill.

Bull-head.

Crappie.

Bluegill.

Bluegill and
Bull-heads.

Bull-head.

be"

I", ^

t/2

TtU

.
be "

55 «

Control.

CHART I.

Showing the reactions of the fishes to different degrees of acidity. The gradient
is between the two kinds of water, indicated at the top of each graph. Numbers
at the left of the chart indicate time in minutes. Strongly acid = 18 c.c. CCh per
liter; moderately acid = 8-10 c.c. per liter; slightly acid = 2-3 c.c. per liter;
and neutral water = actual neutrality to i c.c. per liter. Dotted lines indicate
that fish was driven.

(c) Reaction to Acidity in Distilled Water. — The distilled water,
which was available in quantity from the chemistry department,
was not rapidly toxic to the fishes and since the foregoing results
are of some general biological importance, it was decided to repeat
the experiments in distilled water. This water was faintly acid
with CO2 containing 2-3 c.c. per liter. It contained no salts ; so the
addition of a strong acid resulted in no complications such as those
discussed in the case of the tap water. A number of experiments
was performed with various strengths of acid and alkali. The
neutral portion of the tank was kept track of by means of titra-
tions and the reactions of the fishes to this neutral region es-

REACTION AND RESISTANCE OF FISHES. 235

pecially noted. The results are presented in Table 1 1.; in brief
they are as follows. The fishes spent practically all the time in
the acid portion of the tank, turning back from the alkaline end
at a point just on the acid side of neutrality. They did not, how-
ever, select the highest acidity available, but swam back and
forth in the tank betAveen neutrality on the one hand and about
.OOO2N HoSO4 on the other. The small amount of COa present
in the distilled water may be neglected in the presence of the
much more ionized acid. At the range of dilution used in these
experiments, carbonic acid would have to be about 1,400 times
as concentrated as sulfuric acid, to give an equal concentration
of H ion.

TABLE II.

SHOWING THE REACTIONS OF FISHES TO ACIDITY AND ALKALINITY IN DISTILLED

WATER.
Acid
Used. Concentration. Reaction.

.00004 N Negative: choose higher acidity.

.0001 N Positive: some fishes choose this concentra-

tration in preference to either higher or
lower acidity.

H2SO4 .0005 N Very negative.

H2SO4 .0002 N Still very negative.

H2SO4 .00005 N Positive when neutral water is the other

choice.

The fishes used did not select alkaline water in any case except
when the only other choice was neutrality. Then they spent
most of the time on the alkaline side, rather than at the neutral
point.

2. Reactions to Alkalies.

(a) Alkalies in Neutral Water, (i) Na2COs (.01 N) in Neutral
Water vs. Neutral Water. — Six experiments were run with this
combination. The results were rather indefinite. However, the
graphs as a whole show a slight preference for the alkaline half of
the tank. As has been pointed out already, the fishes are nega-
tive to the neutral water, and these experiments confirm this
reaction, even though the only other choice is alkalinity.

(b~) Alkalies in Strongly Acid Water. — In this water which is
acid with CO2 (18 c.c. per liter), the first action of the alkali
will be to neutralize the acid. Thus a small amount of alkali
introduced at one end will simply produce an acid gradient by

236 MORRIS M. WELLS.

lessening the acidity at this end. Eighteen c.c. of CO2 equals an
.0008 N solution. In most cases, the concentrations of alkali
used have been much greater than this and the amount used up
in neutralizing the acid may be looked upon as negligible. In
some experiments, to be cited, the acid factor is of much impor-
tance.

(1) NazCOz (.01 N) in Strongly Acid Water vs. Strongly
Acid Water (Graph I, Chart II.}. — The fishes stayed in the middle
of the tank, coming to the surface very little. The gradient
was acid at one end and alkaline at the other. Titrations showed
that the fishes spent most of the time on the acid side of neutrality.

(2) NaiCOz (.002 N} in Strongly Acid Water vs. Strongly
Acid Water (Graph 2, Chart II.}. Fifteen experiments were run
with this combination. The graphs show that the fishes spent
most of the time nearer the alkaline end than before, but titra-
tion showed that they were merely following the neutral point,
remaining on the acid side most of the time.

(3) Na^COz (.0005 N) in Strongly Acid Water vs. Strongly
Acid Water (Graph 3, Chart II.). This concentration of alkali
was just a little more than enough to neutralize the acid in the
water of the alkaline end. The end was really slightly acid,
however, from the diffusion of more acid from the acid end of
the gradient. The fishes moved into the so-called alkaline
(really slightly acid) end and remained there during the experi-
ment. This was true for all the fishes used.

(4) NaHC03 (.oiN) in Strongly Acid Water vs. Strongly
Acid Water. — This salt is neutral to phenolphthalein as has been
pointed out in the preceding discussion. A number of experi-
ments, recorded both by graphs and readings at short intervals,
were run with it. The results were not at all definite. The
fishes seemed to be indifferent to this bicarbonate in acid water,
or else they were not at all stimulated by its presence.

(5) NHiOH in Moderately Acid Water (Made it Faintly Alka-
line} vs. Moderately Acid Water (Graphs 4 and 5, Chart II.}. Ten
experiments with this alkali were run, to check up Shelford and
Alice's work ('13) with it. They say (p. 252) that the fishes
(Abramis) did not react to ammonia in a concentration which
•caused them to turn on their sides after an hour or more. In

REACTION AND RESISTANCE OF FISHES.

237

my experiments, I found also that the fishes do not react to this
alkali with the precision found for the other alkalies used.

liull-head.

Sun-fish.

Crappie.

Bluegill.

Bluegill.

liluegill.

H

||

|8

g'D
W

U g>

Q

M-c

5 •

•r" rt *^

5. -

WS
z '

S >-T3

"C ^ •-

Control.

CHART II.

Showing the reactions of the fishes to alkalies. The gradient was between the
two kinds of water indicated at the top of each graph. Strongly acid water = 18
c.c. CO2 per liter; and moderately acid = 8-10 c.c. per liter. Numbers to left =
time in minutes.

In the first experiments a .005 N solution was run into one
end of the tank. The fishes selected the middle of the tank for
the most part, though one blue-gill was positive to the ammonia
end. The concentration of OH ion was of course very low,
with so small a concentration of so weakly ionized a base, and
since other experiments have shown that blue-gills are less
negative to neutrality than are other fishes, this reaction is not
surprising. The ammonia concentration was raised to .01 N
and the fishes, blue-gill included, moved toward the tap-water
end of the tank. Later the concentration of the alkali was

238 MORRIS M. WELLS.

raised to .02 N, but even now the avoidance of the alkali end was
not nearly so definite as in the experiments with the other alkalies.
Graphs 4 and 5 (Chart II.) show this indefinite reaction very
clearly. In the .02 TV gradient, the fishes were soon overcome
by the toxicity of the water, which they selected, and they died
there if not removed.

The fact that fishes fail to recognize ammonia in solution is of
considerable importance, for this substance is being introduced
into fish waters in many kinds of sewage. Furthermore it will
be shown in the second paper of this series, that the gas
has not lost its toxicity even when it has been converted into
its various salts. The chemical explanation of the failure of
the fishes to recognize and react to the presence of fatal concen-
trations of the hydrate in solution is probably due to the fact
that the concentration of ammonia as gas, reaches a fatal con-
centration before the concentration of OH ion stimulates the
fishes sufficiently to cause them to react negatively. They do
not appear to react to the gas itself. Noyes ('13, pp. 203-4)
states that ammonia dissolves in water, in part, without chemical
change and that it is probable that a large part of the ammonia
exists, as such, in the solution. He quotes Moore ('07) as cal-
culating that only 30-40 per cent, of the ammonia exists as am-
monium hydroxide, NH4OH, at 20° C. Noyes thinks that the
per cent, may be even less than this.

Again, the solution of ammonia diffuses through the water
with great rapidity; much more rapidly than do most other
substances. To determine the rate of diffusion, a little phe-
nolphthalein was added to the aspirator bottle (AB, Fig. i)
containing the ammonia solution. The pink solution could be
seen as it moved through the tank, and in less than a minute
it had spread over the entire surface, and to a lesser extent, had
penetrated the deeper water. Because of this rapid diffusion,
no perfect gradient could be established with this substance.
It may also be noted that ammonia behaves just opposite from
the salts, the latter spreading along the bottom. In the am-
monia experiments, the fishes seldom approach the surface,
while in strong carbon dioxide gradients, they spend much time
gulping the surface film. Shelford and Alice ('13, p. 231) state

REACTION AND RESISTANCE OF FISHES. 239

that in open tanks the amount of CQz at the surface is markedly
less than at deeper levels.

3. Conclusions Drawn from the Reaction Experiments.

The reaction experiments recorded in the previous pages sug-
gest the following conclusions. (i) Fresh-water fishes are
negative to neutrality in favor of either slight acidity or slight
alkalinity. Their normal choice is slight acidity' (about .00005 N
H2SO4 or .0001 N CO2). (2) Species of fishes differ in the de-
gree of acidity selected. Blue-gills select water that is but
very slightly acid (1-2 c.c. CC>2 per liter, i. e., .0001 N carbonic
acid) while crappies select a concentration of from 4-6 c.c. CO?
per liter. (3) The principal stimuli to which fishes react are
H and OH ions. They do not react to ammonia, as a gas in
solution.

B. RESISTANCE EXPERIMENTS.

It has been pointed out that the stock of fishes did not live
well in the aquaria when these were supplied with water, which
was neutral, or nearly so; to determine more exactly the reasons
for the high mortality, between 50 and 60 experiments were
performed. Some of these experiments lasted through a number
of weeks, while others were finished in a few hours. The fishes
were placed in different concentrations of acid and alkali in
partly aerated wrater (from the aquaria) and in distilled water.

i. Resistance to Acids.

The resistance of fishes to carbonic acid has been worked out
(Wells, '13) and it was decided to try the effects of other acids.
Ten experiments with sulfuric acid in distilled water are sum-
marized in Table III. The table shows that there is a concentra-

TABLE III.

SHOWING THE RESISTANCE OF FISHES (3 GRAM BLUE-GILLS) TO HiSO4 IN DISTILLED

WATER.

Concentration of Acid. Dying Time in Hours.

.001 N 3.5

.0005 N 7.0

.0002 N 42.0

.00015 N 60.0

.000075 ^ Alive and vigorous at end of a month.

240 MORRIS M. WELLS.

tion of this acid in distilled water, at which the fishes in question
live as well as though in tap water. Higher concentrations of
acid are fatal, the time required to kill the fishes being propor-
tional to the hydrogen ion concentration.

2. Resistance to Alkalies.

In a .001 N solution of KOH in distilled water, a 3 gram blue-
gill lived 4 hrs. and 25 min. In a .0005 N solution, a fish of the
same size was alive at the end of 10 days. Titration at this time
showed that the water had become acid to phenolphthalein from
the CO2 given off in the metabolism of the fish. The experi-
ment was discontinued. To make sure that the fish in the first
experiment had not been killed by the toxic potassium ion,
another 3 gram blue-gill was placed in a .01 N solution of
NaHCOs in distilled water. At the beginning, this solution was
neutral, but it was expected that the bicarbonate would dissociate
and the solution would become slightly alkaline from the
carbonate thus formed. A blank control, containing the same
amount of bicarbonate, but no fish, was run. The fish in the
experiment died on the third day. Titration showed that the
water had become .0009 N alkaline. The control was .001 N
alkaline. Blue-gills therefore do not live well in water which is
even very slightly alkaline.

3. Resistance to Neutrality.

The foregoing experiments, together with many facts recorded
in the literature, suggested the possibility that the fact that it is
neutral may have something to do with the toxicity of distilled
water. Thirteen experiments were performed to test this pos-
sibility in a preliminary way. The facilities available did not
make it possible to experiment with absolutely neutral water,
but the results obtained are suggestive, as neutrality was ap-
proached very closely in some cases. Most of the experiments
were performed with water that came from a copper still and
will be referred to as once-distilled water. A few experiments
were performed with a much purer water which was the once-
distilled water redistilled in a better still and coming in contact
with little copper. In neither kind of water could the amount

REACTION AND RESISTANCE OF FISHES. 24!

of copper have been especially large, however, for small blue-gills
lived in both kinds as well as in tap-water, so long as the water
was slightly acid.

A comparison of the conductivities of the two kinds of water
showed that the once-distilled had a conductivity1 of 600 X io~r
while the conductivity of the twice-distilled was only 10 X io~7.
These conductivities are for 25° C. The conductivity of the
water probably does not indicate the amount of copper present
however for the metal is in all likelihood present in the col-
loidal state. Mengarini and Scala ('12) have shown that a num-
ber of metals, including copper, form a colloidal solution with
distilled water even at room temperature, and especially in
the absence of air. The conditions in a still would be especially
favorable for the reaction, since the temperature is high and
air excluded.

The addition of an acid to a colloidal solution would tend to
precipitate the colloid, and this undoubtedly explains in part
the effect of addition of acid in making distilled water less
toxic,2 as it will be shown that it does. Since, however, it has
been shown (Bullot, '04) that distilled water which contains
no copper is still toxic to organisms, other factors must be con-
cerned. The evidence of the experiments presented in the present
paper, indicates that the neutrality of the water is one of these
factors.

It has been suggested in the preceding pages that the blue-
gills and crappies differ in respect to the hydrogen ion concen-
tration which they select and their resistance to the distilled
water bears out this point as the crappies die in it in a day or
so, while the blue-gills live indefinitely.

(i) Experiments with Once-distilled Water. — This water was
slightly acid to phenolphthalein and was neutral to methyl
orange. Its toxicity was tested by placing fishes in jars
containing a liter of the water. A 1 2-gram crappie died in this

1 The conductivity of pure water is I X io~7.

2 Locke ('95) calls attention to the fact that poisonous distilled water may
lose its poisonous properties (if due to copper) by long boiling, and especially when
brought into contact with sulphur, carbon, manganic oxide, cotton wool, silk,
and other substances. The effect is very probably again due to the precipitation
of the colloidal copper.

242 MORRIS M. WELLS.

water in 2 days, but when this same liter of water was divided
into two parts and a 3-gm. blue-gill placed in each part, both
fishes were normal at the end of a month. In Table IV. is given
a summary of a number of experiments performed with blue-gills
in once-distilled water.

TABLE IV.

SHOWING THE RESISTANCE OF SMALL BLUE-GILLS (3-5 GRAMS) TO DISTILLED

WATER THAT is BARELY ACID WITH COz.
Conditions of Expt. Fish Placed in, Resistance of Fishes.

1. Freshly distilled water Normal after 5 days; expt. discont.

2. Boiled distilled water Normal after 5 days; expt. discont.

3. Distilled water plus Na2COs to make

neutral Normal after 2 days. Water acid

again.
Added a little NaHCO3 to (3) to

keep neutral Dead on loth day.

4. In dist. water as in (i) Normal after 30 days.

Table IV. shows that the once-distilled water is not greatly
if at all toxic to the blue-gills, but experiment 3 shows that these
fishes cannot live in the water if it is slightly alkaline. This
same distilled water is rapidly toxic to the crappies and sun-
fishes, however, as was shown in an experiment already described
and in those which follow. This lack of resistance of the sun-
fishes in particular is a complete reversal of the ordinary specific
resistance of these fishes as compared with the blue-gills, for
in carbon monoxide, ethylene, sulphur dioxide, etc., the sun-
fishes are much more resistant than are the blue-gills.

On January 30, a liter of water (once-distilled) was made
.00005 N acid with H2SO4 and another liter left as it came from
the still. An 8-gram crappie was placed in each jar. The
fish in the pure distilled water was dead in 12 hrs. while the one
in the distilled water made acid, lived for 65 hrs. Several other
•experiments of this sort gave similar results, showing that the
crappies cannot live in the neutral distilled water when it is
pure, as well as they can when it is made slightly acid. It is
very probable that sightly higher concentrations of acid than
those used would have prolonged the lives of these fishes even
more successfully than the .00005 N DUt as the stock of fishes
was running low, these experiments were reserved for another
time.

REACTION AND RESISTANCE OF FISHES. 243

An experiment with small bullheads is very interesting.
Normally the bullheads are perhaps the most resistant fresh-
water fishes known. In the reaction experiments they selected
a rather low concentration of hydrogen ion but were decidedly
on the acid side of neutrality. In the pure distilled water a
bullhead (4 in. long) lived 8 days; another in distilled water
made .00005 N acid, lived for 20 days.

(2) Resistance to Doubly Distilled Water. — This water was less
toxic to the crappies than was the once distilled water, as it
contained less colloidal copper. It has been pointed out that
the toxicity of the once-distilled water was lessened by the
addition of acid, partly because the acid precipitated the col-
loidal copper. The experiments indicate further, however, that
the neutrality of the water must be reckoned with also. This is
again brought out, and more definitely, by a few experiments
with the twice-distilled water. A quantity of this water was
placed in a large bottle and a solution of barium hydroxid was
suspended over it. At the end of a week the water was practically
neutral. Two portions were taken in 500 c.c. Erlenmeyer
flasks and a small bullhead (2.5 in.) placed in each. One portion
was left neutral and the other made slightly acid with HoSO4.
The fish in the neutral water lived 16 days and the one in the
acid water 19 days. A few other experiments were performed
with the twice-distilled water and all gave similar results. The
stock of fishes was about exhausted, however, and further ex-
periments were delayed until another time.

V. GENERAL DISCUSSION.

The fact that in natural bodies of water the chemical reaction
of the water may vary from alkalinity through neutrality to
acidity or the reverse, makes the practical importance of a knowl-
edge of the reactions and resistance of fishes and other organisms
to such chemical conditions an obvious one. From the experi-
ments and discussion which have gone before, it is clear that
water which gives an alkaline reaction to phenolphthalein for
any length of time during the year, is undesirable as a home
for most fresh-water fishes. On the other hand, marine fishes
(Shelford and Powers, '15) with the exception of the anadro-

244 MORRIS M. WELLS.

mous species, probably would not survive in water which was
even faintly acid. Since algae and other phytoplankton forms
(Birge and Juday, 'n and '12) may cause a body of water to be-
come partially or wholly alkaline, through their ability to dis-
sociate the bicarbonates, vegetation in fish waters assumes a
line of importance heretofore little considered. The effects
of sewage upon the acidity or alkalinity of natural bodies of
water must also be reconsidered in the light of its possible in-
jurious or beneficial effects due to its chemical reaction. Thus
a large number of interesting and important questions suggest
themselves.

The effect of the chemical reaction of the water upon the
distribution of organisms promises much room for investigation.
There is no doubt but that fishes recognize the difference be-
tween very faintly acid or very faintly alkaline, and neutral
water. Henderson's work ('13), upon the mechanism which
maintains a constant proportion of H and OH ions in the blood
of animals, suggests the physiological reason for this extreme
sensitiveness of the fishes. It is clear that even very small
variations in the proportions of these two ions in the blood of
the organism, are of grave importance, and we find in the blood
a combination of gases and salts that makes such variations
impossible as long as the animal is normal. The blood will main-
tain its normal chemical reaction (just on the alkaline side of
neutrality) in the face of relatively large changes in the environ-
ment, yet we know that the mechanism breaks down when the
change is either too great or too long continued (acclimatization
is not considered at this time). The hyper-sensitiveness of the
animals to the chemical reaction of the water, in the case of aqua-
tic organisms, is another important factor in preserving the
normal reaction of the blood, as the reactions of the organisms
work in a way that causes them to turn back from concentrations
of H or OH ion that would be detrimental. The delicacy and
accuracy of these reactions are evidenced in the reaction ex-
periments which have been discussed in the preceding pages.

The physiological effect of the acid, neutral, and alkaline
water upon the organism very probably has to do with decrease
or increase in the permeability of the exposed tissue cells (es-

REACTION AND RESISTANCE OF FISHES. 245

pecially gills in case of fishes). Osterhout ('14) has shown that
in plant cells alkalies increase the permeability up to death;
acids however at first produce a rapid decrease in permeability,
followed later by an increase which continues up to death. The
concentrations of acid used by Osterhout were .001 .V to .03 N.
Very low concentrations such as those used in the experiments
discussed here would very likely maintain a permanent decrease
in the permeability of the cells, and the concentrations of acid
in which the fresh-water fishes normally live, may thus protect
the fishes by decreasing the permeability of their gills and pre-
serving the normal reaction of the blood. Alkaline water, on
the other hand, does not do this for fresh-water fishes, and they
soon succumb in it. The results of Shelford and Powers ('15)
indicate that the action of alkaline water upon marine fishes is
to produce a normal permeability of the membranes and it may
be that acid water would kill these fishes by decreasing the per-
meability below normal.

The effect of neutrality upon the permeability of tissues has
not been worked out, so far as I am aware, but since fresh-water
fishes, and probably marine fishes also, are negative to neutral
water, it must be that such water exerts a marked effect upon
the permeability, or some other physiological condition, in the
gill membranes. The negativeness of organisms to neutral
water indicates that they are either over-stimulated in such
water, or under-stimulation sets up internal disturbances. Thus
they may avoid it because of its non-stimulating character. It
may well be that in neutral water, the normal chemical reactions
do not go on, for acidity and alkalinity surpass all other con-
ditions, even temperature and concentration of reacting sub-
stance, in their influence upon many chemical processes. Of
all the catalytic agents, H and OH ions are by far the most im-
portant, and in their influence upon the stability of colloidal
systems they are unapproached by other substances (Henderson,

'13).

Birge and Juday ('n and '12) attempt to explain the vertical
distribution of the plankton in the lakes of Wisconsin and New
York, upon the basis of relation to oxygen and food. This
attempt has, it seems to me, met with little success, and they

246 MORRIS M. WELLS.

themselves point out many contradictions. According to their
idea, the plankton forms must in many instances be reacting
positively to concentrations of oxygen which are as small as
.1 c.c. per liter, or even less. This supposition is contrary to
all the experimental evidence regarding the reactions of aerobic
fresh-water organisms to this gas. In an attempt to correlate
the distribution of the zooplankton with the chemical reaction
of the water, I have gone over Birge and Juday's tables and
figures, and have come to the conclusion that such correlation
exists. Their data indicate in practically all of the lakes (in
the summer condition) a point at some depth below the surface
of the lake, where the organisms are more numerous than at
any other depth. In many cases this rise is proportionately
very high and is usually of small amplitude. Thus the large
number of forms occurs in a rather limited region vertically.
After the rise, there is a marked diminution in the number of
forms and then again at a little greater depth there is another
increase, smaller than the first, but still very noticeable in their
curves. This increase is followed by a second diminution.
The first diminution usually occurs in or near the thermocline
where the temperature often shows a very sudden lowering.
The oxygen supply sometimes falls off here also, but not always,
and in the lakes to which I refer particularly, the oxygen supply
is practically the same at all depths. A very important fact,
however, is that the water in the region of the thermocline, i. e.,
at the region of smallest numbers of plankton, is often neutral
or very nearly so (summer condition). Above this region the
water is alkaline, and below acid. From the data given in
Birge and Juday's Tables XVIII. and XIX. ('12, pp. 602-608),
I have compiled the following table (Table IV.) to show the re-
lation of the zooplankton to this neutral region. Birge and
Juday's Table XVIII. is a record of temperatures and gas con-
tents at the different depths; Table XIX. is an analysis of the
plankton catches made in ten lakes. The records for a given
lake were all made on the same day. Table XVI 1 1 . gives titration
records which show that in three of the lakes at a definite depth,
the water was neutral. Table XIX. gives the plankton col-
lections at different depths in these three lakes, on the same

REACTION AND RESISTANCE OF FISHES.

247

day. Table IV. inserted below, is made up from a combination
of the data found in the two tables; most of the data in Table
IV. refer to the three lakes in question. In the instance of
Triarthra, however, the data come from two other lakes, as this
form does not occur in the three lakes from which the other
data are taken.

TABLE IV.

SHOWING THE RELATION OF ZOOPLANKTON TO NEUTRALITY.
The table is compiled from Birge and Juday's ('12) Tables XVIII and XIX,
Most of the data are taken from their records for the three lakes, Canandaigua,
Seneca, and Skaneateles. In these lakes the neutral depth was accurately located
by titration of samples. The titrations and plankton collections were made on
the same say. The data for the rotifer triarthra are taken from lakes Hemlock
and Keuka as these are the only lakes in which this form occurs in the records,
n.c. = no collection at this degree of acidity or alkalinity. The figures in the
columns indicate the number of forms per cubic meter of water.

Name of Animal.

Alkalinity in C.c. per Liter
of CO2 to Make Neutral.

Neutral-
ity.

Acidity Ln C c. of CO2 per
Liter.

3-2-

1.5-1.

•5-.2S.

o.

.25-5-

•75-1-

.1-1.5.

Pleosoma (R)

3.925
12,250
11,320
43-700
2,885
28,150
7.850
130
4.0OO

13-775
625
1,260
52,330
12,350
0
0

o
o
400

19.130
2,750

28,050

6,660

290
1,250
7,620

685
650
104,500
1,620
n.c.
n.c.

0
O
O
42,8OO

n.c.
13.250
17.350
250

200
7,62O

65
4OO
85,160
2,350
77O,40O
O

O
0
0
0
260
570
2,220
250
30

25
O
130
2,025

160
95,600
n.c.

O
0

o
o
o
140

1,440
30

20

30

65

1.145

11,760
1,190
1,900

I.O=;O

O
O
O
O
O

40
390

65
20
o

0

25
5.750

1,240
o

I. IIO

o
o
o

0

o

205

IOO

30

20

5
o
o
1,670

40

1,900
2.4.2=;

Vorticella (P)

Asplanchna (R) ......

Dinobryon (P)

Diaphanosoma (C). . .
Nauplei

Diaptomus (Co)

Conochilus (R)

Anuroea (P)
Cyclops (Co)

Notholca (R)

Daphnia (C)

Ceratium (P)

Polyarthra (R)

Malamonas (P)

Triarthra (R) .

Letters in parenthesis after name of animal indicate the following.
Rotifer; (P) = Protozoan; (C) = Cladoceran; (Co) = Copepod.

(R) =

Table IV. shows (i) that all the zooplankton forms are more
numerous on either the acid or the alkaline side of neutrality,
than they are at neutrality itself, i. e., they are negative to neu-
trality; (2) some forms as Pleosoma and Vorticella, are found
only in the alkaline water; (3) others range between slight alka-
linity and slight acidity but are never very numerous at neu-
trality and often (Daphnia, Ceratium, etc.) show an increase
•on either side; (4) a few forms (Triarthra) occur wholly on the
acid side of neutrality.

248 MORRIS M. WELLS.

The factors that regulate the distribution of the plankton
in the lakes are undoubtedly numerous. The only certain way
to determine them is to investigate experimentally the reactions
of the animals to the factors concerned, both singly and in com-
bination. To do this would be tedious but not especially dim-
cult. As an index to the distribution of these forms, I believe
that the presence and position of a neutral layer of water will be
found to be important.

Besides the experimental data presented in the papers by
Birge and Juday, the literature contains much other experimental
evidence which bears directly upon the question of the toxicity
of neutrality to organisms. Much of this evidence is found in
connection with experiments upon the toxic effects of distilled
water, and the action of salts in antagonizing this toxicity. In a
series of papers published by Ringer and his students between
the years 1883 and 1893 the question of the toxicity of distilled
water was investigated and its reality apparently demonstrated.
It was also shown that various salts are effective in neutralizing
this toxicity, some being much more efficacious than others.
In 1893 Naegeli showed that for Algae (Spirogyra) at least,
the toxicity of distilled water was due to contamination from
the copper stills in which it was prepared. Locke ('95) con-
firmed Naegeli's results by showing the effect upon certain
fresh-water animals to be due also to the minute amounts of
copper present, and Ringer ('97) again taking up the subject
reversed his former conclusions and confirmed those of Locke.
Jennings ('97) found that Paramcecia live for weeks in distilled
water. Moore ('oo) says that young trout and tadpoles (unfed)
live as long in distilled as in tap water, i. e., several weeks.
Lillie ('oo) says that Planaria maculata will live in distilled water.
Pure distilled water seemed then not to be toxic to fresh-water
animals though apparently toxic to most marine animals. Fun-
dulus eggs seem to be an exception among marine animals
(Loeb, '99), as they can live in distilled water for weeks and
still produce normal embryos. In 1903 Bullot after testing the
effects of distilled water upon the fresh-water amphipod,
Gammarus concluded that pure distilled water was toxic to this
crustacean. Bullet's experiments were performed with great

REACTION AND RESISTANCE OF FISHES. 249

care; he considered and seemed to have eliminated the following
possible toxic factors: copper, impurities from the glass, low
oxygen, ammonia, and carbon dioxide. He found also that
NaCl in small concentrations would neutralize the toxicity of
the pure water to such an extent that the animals lived almost
as well in an .00008 N solution of this salt as in the natural
fresh water. The toxicity of pure distilled water, he concluded,
is due to the lack of salts in solution. Peters in 1908 performed
some very careful experiments to test the effect of pure distilled
water upon protozoa. He came to the conclusion that distilled
water which contains no salts, and which is changed often enough
to prevent their accumulation from the metabolism of the ani-
mals, is rapidly toxic to these forms.

I have gone over the above papers and have found many
statements which indicate that the presence of a certain con-
centration of hydrogen ion, was beneficial to the animals ex-
perimented upon. For instance, Ringer ('83) states that the
distilled water which he used killed minnows, on an average,
in 4.5 hours. He also says that the distilled water was very
faintly acid; so faint was the acidity that he did not rely upon
his own judgment but had others make the test also. However,
he says, to prove that the acidity was not the cause of the death
of the minnows, he took three liters of water and to one added
6 drops of 10 per cent, acetic acid, to the next 12 drops and to
the third 20 drops. He then placed three minnows in each liter
of acid solution. After 24 hours, the minnowrs were "quite
natural" and he concluded, therefore, that the acidity could
not have been harmful in the case of the distilled water. This
conclusion of Ringer's illustrates the attitude taken by most
authors with regard to the presence of acid in the water, that is,
the acid is looked upon as a detrimental factor, to be considered
negatively. So far as I have been able to read, the authors
quoted above have taken little consideration of the possibility
that the presence of a certain concentration of H or OH ion is
essential to the welfare of animals.

This Ringer does not suggest, even though the minnows in
the acid water were well on the way to live as long as any of the
animals kept in salt solutions. In this same paper, Ringer notes

250 MORRIS M. WELLS.

that when he put a large number of fishes (up to a maximum)
into a given volume of distilled water, they lived longer than
one or two fishes placed in the same volume of water. He at-
tributes this to the excretion by the fishes of inorganic salts,
and does not take into consideration the carbon dioxide factor
which would have increased the acidity of the water to many
times that of the almost neutral distilled water. Again, in
speaking of the salts which are best for preserving life in distilled
water, Ringer states that the calcium salts are better than those
of sodium and potassium, that CaSo4 is better than CaCb
and that the phosphate of lime (Ca3PO4)2 is much superior to
all the other salts. This latter salt, he states, is decidedly acid,
and he says ('86) "it is interesting to observe that though the
circulating fluid with phosphate of lime gives a slight acid reaction
to delicate blue litmus paper it will sustain contractility of
muscle for hours." Thus a small hydrogen ion concentration
seems to be beneficial, if not essential, to the continued life and
activity of the organisms and tissues in question.

The question of the existence of a carbon dioxide optimum
for animals has received considerable investigation with varying
results. Ringer in 1893 investigated the influence of carbonic
acid upon the frog's heart and concluded that free CO^ in saline
solution arrests cardiac contractility. He does not state what
concentrations of CO2 were used, but since he speaks of passing
carbonic acid through the solution "for some time," his solutions
were probably very acid. In a few experiments he neutralized
the slightly acid distilled water which was used to make up the
saline solutions, with NaOH, and noted that in this neutral
solution, the contractions of the heart very soon became ab-
normal. Jerusalem and Starling ('10) review the literature
regarding the importance of carbon dioxide for the ordinary
functions of the body, and report a series of experiments to
determine its influence upon the beat of the heart of the frog,
tortoise and mammal (cat). They conclude that the CO2 tension
in the blood must be maintained at a certain height, if the pump-
ing action of the heart is to be normally carried out. In their
review of the literature they point out that their conclusions
are in accord with those of Miescher, Haldane, Mosso, Hender-

REACTION AND RESISTANCE OF FISHES. 25!

son, and Bottazzi (see pp. 279-280). The lowest concentration
which Jerusalem and Starling used was 2 per cent, of an atmos-
phere or about 20 c.c. CO2 per liter. Their highest ran up to
200 c.c. per liter. Hooker ('12) tested the effect of carbon
dioxide upon muscular tone and, in opposition to Jerusalem and
Starling, concluded that this gas does not appear to be directly
beneficial to tissues, except in case of intestinal muscle rhythm.
He thinks it may be indirectly beneficial. Like most other
workers upon this problem, Hooker used very high concentra-
tions of the gas. His concentrations varied from 5 per cent, to
20 per cent, of the gas, in the atmosphere to which the solution
bathing the tissue was exposed. Water will dissolve nearly
its own volume of CQ2 and thus the concentration of carbonic
acid varied from 50 to 200 c.c. of CO2 per liter. The smallest
concentration used would kill most fresh-water fishes in a short
time.

Reuss ('10) worked upon the effect of CO2 upon the respira-
tion of fishes and concluded that it is an important one. The
regulation according to him is through the respiratory center
and not peripheral as Bethe ('03) believed. Shelford and
Allee ('130) note the extreme sensitiveness of fishes to COo in
gradients, and think the production of the gas as a product of
the metabolism of the organism may tend to increase its ex-
ternal effect when the fishes come in contact with water con-
taining it.

Bullot ('04) in his work with the fresh-water amphipod (Gam-
marus) noted, as did Ringer in the case of fishes, that the animals
lived longer in distilled water when a number was present in a
given volume, or in other words, when the volume of water per
individual was small. He says: "If the amount of water
falls below a certain limit, the animals will live the longer, the
smaller the amount of water, provided the quantity does not
fall below a certain minimum." In Table V. I have collected
Bullet's data showing this point. The table shows that the
relation holds for both redistilled water and water distilled in copper
alone. The length of life in the water from the copper still is
proportionately shorter throughout.

252

MORRIS M. WELLS.

TABLE V.

SHOWING THE RESISTANCE OF AMPHIPODS IN DISTILLED WATER, WHEN EQUAL
NUMBERS OF ANIMALS ARE PLACED IN DIFFERENT VOLUMES OF WATER, OR
WHEN DIFFERENT NUMBERS OF ANIMALS ARE PLACED IN EQUAL VOLUMES OF
WATER (COMPILED FROM BULLOT, '04, PP. 204-5).

Per Cent, of Animals Alive After 2 Days.

No. Animals.

Volume of Water.

Water.

Redistilled in Glass.

Distilled in Copper.

I
5

10

Same throughout

45%
80%

90%

Proportionately
less throughout.

Time Required to Kill One Half the Animals.

Same throughout

5 c.c.

10 days

8 days

20 c.c.

2.5 days

1.5 days

50 c.c.

1.5 days

i day

IOO C.C.

same

same.

In considering the possible importance of CO* as a factor in
the toxicity of distilled water, Bullot states that the water which
he used was very faintly acid to phenolphthalein, but not enough
to injure the animals. He says: "We know from the works of
P. Bert that, for cold-blooded animals like the frog, for instance,
CC>2 is toxic only when its concentration in the air reaches 15 per
cent. This corresponds to a solution of 15 per cent, of this gas
by volume, as the water dissolves its own volume of COo at
ordinary temperature and normal pressure. This quantity is
350 times larger than the one which could be found in the dis-
tilled water." A 15 per cent, solution of CO2 means 150 c.c.
per liter of water. 1/350 of this is .42 c.c. In other words the
distilled water used by Bullot was practically neutral, since the
amount of hydrogen ion to be obtained from so small a quantity
of so little ionized an acid as carbonic acid, would be almost
negligible. In the gradient experiments cited in this paper, I
have shown that certain fishes are negative to so small a con-
centration of CC>2 as i c.c. per liter, in preference for slightly
higher concentrations. I have further shown that these fishes
do not live as well in distilled water that is practically neutral,
as they do in the same water made slightly (.00005 N) acid.
Thus the existence of a hydrogen ion concentration optimum
for these forms seems to be clearly demonstrated.

REACTION AND RESISTANCE OF FISHES. 253

Peters ('08) makes no mention of the possibility of the neu-
trality of the distilled water which he used, having something to
do with its toxicity, yet in a previous paper ('07) he recognizes
the importance of the presence of a certain concentration of
hydrogen ion for the existence of certain protozoa in hay in-
fusions. On page 346, he says: 'The data obtained indicate
that, of the chemical conditions, the concentration of acid ... is
one of the chief factors determining the biological content and
history of a culture."

From the data and discussion that have gone before, it seems
certain that the chemical reaction of the water is a factor of
marked importance in the life history of fresh-water animals.
Some fresh-water forms are apparently positive to alkalinity as
seen in the fresh-water lakes (Birge and Juday, loc. cit.} and
others, that normally live in water that is acid with CO2 are not
killed by living in alkaline water (isopods). On the other hand,
many forms, and probably most of the fresh-water fishes belong
here, are always found in acid water if such be available, and
these forms cannot live normally in neutral to alkaline water.
Shelford and Powers ('15) have shown that marine fishes select
the alkaline side of neutrality in a gradient, and in this dif-
ference in the behavior of the fishes, lies a key to the fundamental
physiological difference in the organisms of these two habitats.
Fresh-water fishes must live in the presence of an excess of hy-
drogen ion if their life processes are to be carried on in normal
fashion. Shelford ('14) states that the carbon dioxide content
of the water over the breeding grounds of fresh-water fishes
should not average more than I c.c. per liter, nor exceed 5 c.c.
during the summer months. This statement is probably wrong
in limiting the average to I c.c. per liter for some fishes, as the
green spotted sunfish and the crappies are negative to this
small concentration of CO2 showing a preference for slightly
higher concentrations. Blue-gills, on the other hand, select a
degree of acidity that is very little above neutrality. The CO2
concentration selected by fishes probably varies with the season,
and certainly with the salt content of the water in which they
live. The variations of the CO2 optimum in salt concentration
will be discussed in the second paper of the series.

254 MORRIS M. WELLS.

One thing is clear; the chemical reaction of the water should
be known with accuracy, in all experiments with salts and gases
in solution. A recognition of this fact will help to clear up some
of the many contradictory results which have been obtained by
various workers. It seems to be demonstrated beyond a doubt
that the toxicity of distilled water is in part due to the absence
or scarcity of inorganic salts, but it is also evident that the
neutrality of such water may be an important factor in its
toxicity.

VI. GENERAL CONCLUSIONS.

1. The chemical reaction of the water is an important factor
in the reactions and resistance of organisms.

2. Fresh-water fishes select slight acidity in a gradient, when
the other choices are neutrality and alkalinity. They choose
slight alkalinity in preference to neutrality.

3. The CO2 optimum for the different species of fishes ex-
perimented upon, varies from very close to neutrality for the
blue-gill, to 6 c.c. per liter for the sunfishes and crappies. This
is for the November to January condition. At other seasons
and in other waters, the optimum probably varies somewhat.
The optimum acid concentration for fresh-water fishes in dis-
tilled water is about .00005 N H2SO4.

4. The distribution of the plankton in the lakes of Wisconsin
and New York (Birge and Juday, 'n and '12) shows a very in-
teresting correlation with the chemical reaction of the water.
There are fewer animals at neutrality than in the slightly alkaline
and slightly acid water just above and below the neutral layer.
Thus the forms are negative to neutrality.

5. The neutrality of distilled water is a factor to be considered
in its toxicity.

VII. ACKNOWLEDGMENTS AND BIBLIOGRAPHY.

I am indebted to Professor V. E. Shelford for many helpful
suggestions and numerous courtesies during the working out of
the experimental data and preparation of the manuscript.

REACTION AND RESISTANCE OF FISHES. 255

BIBLIOGRAPHY.
Alice, W. C.

'13 Further Studies on Physiological States and Rheotaxis in Isopoda. Jour.

Expt. Zool., Vol. 15, No. 3.
Birge, E. A. and Juday, C.

'u The Inland Lakes of Wisconsin. The Dissolved Gases of the Water and
their Biological Significance. Wis. Geol. Nat. Hist. Surv., Bull. No. 22.
Sc. Series, Vol. 7.
'12 A Limnological Study of the Finger Lakes of New York. Bull. Bur.

Fisheries, Vol. 32, Doc. No. 791.
Bullot, G.

'04 On the Toxicity of Distilled Water for Fresh-water Gammanis. Suppres-
sion of This Toxicity by the Addition of Small Quantities of NaCl. Univ.
Calif. Publ. Physiol., Berkeley, Vol. i, pp. 199-217.
Henderson, Lawrence J.

'13 The Fitness of the Environment. Macmillan Co., New York.
Herrick, C. Judson.

'02 The Organ and Sense of Taste in Fishes. Bull. U. S. F. C., Vol. 22, p. 239-

271.
Hooker, D. R.

'12 The Effect of Carbon Dioxide and Oxygen upon Muscular Tone in the
Blood-vessels and Alimentary-canal. Am. Jour. Physiol., Vol. 31, p. 47.
Jennings, H.

'97 Studies on Reactions to Stimuli in Unicellular Organisms. I. Reactions
to Chemical, Osmotic, and Mechanical Stimuli in the Ciliate Infusoria.
Jour, of Physiol., Vol. 21, p. 258, 64 p.
Jerusalem, R. and Starling, E. H.
'10 On the Significance of Carbon Dioxide for the Heart Beat. Jour. Phys.

Lond., Vol. XL., No. 4.
Lillie, F. R.

'oo Some Notes on Regeneration and Regulation in Planarians. The American

Naturalist, Vol. 34, 5 p.
Locke, S.

'95 On a Supposed Action of Distilled Water as Such, on Certain Animal

Organisms. Jour. Physiol., Vol. 18, p. 318.
Loeb, J.

'99 On lon-proteid Compounds and their Role in the Mechanics of Life Phe-
nomena. I. The Poisonous Character of a Pure NaCl Solution. Am.
Jour. Physiol., Vol. 3, p. 327, n p.
Marsh, M. C.

'06 The Effect of some Industrial Wastes on Fishes. House Documents,

Vol. 64, Water Supply and Irrigation Papers.
Moore, A.

'oo Further Evidence of the Poisonous Effects of a Pure NaCl Solution. Am.

Jour. Physiol., Vol. 4, p. 386.
Naegeli, C.

'93 Ueber oligodynamische Erscheinungen in lebenden Zellen mit einem
Vorwort von S. Schwendener und einem Nachtrag von C. Cramer. Denk-
schriften d. Schweiz. Naturforsch. Gesellsch., Bd. 33, 1893.

256 MORRIS M. WELLS.

Noyes, W. A.

'13 A Text-book of Chemistry. Henry Holt Company, New York.
Osterhout, W. J. V.

'14 The Effect of Alkali on Permeability. Jour. Biol. Chemistry, Vol. 19, p.

335-
'14 The Effect of Acid on Permeability. Jour. Biol. Chemistry, Vol. 19, p.

493-
Peters, A. W.

'07 Chemical Studies on the Cell and Its Medium. II. Some Chemico-
biological Relations in Liquid Culture Media. Am. Jour. Physiol., Vol. 18,
p. 321.

'08 Chemical Studies on the Cell and Its Medium. III. The Function of
the Inorganic Salts of the Protozoan Cell and Its Medium. Am. Jour.
Physiol., Vol. 21, pp. 105-124.
Powers, E. B.

'14 The Reactions of Cray-fishes to Gradients of Dissolved Carbon Dioxide
and Acetic and Hydrochloric Acids. BIOL. BULL., Vol. 27, No. 4, pp.
177-199.
Reuss, H.

'10 Die Wirkung der Kohlensaure auf Atmung der Niederen Wirbeltiere in

besonderen der Fische. Zeit. fur Biol., Vol. 53, pp. 555-587.
Ringer, Sidney.

'83 Concerning the Influence of Saline Media on Fish. Jour. Physiol., Vol. 5,

p. 98.
'84 Regarding the Influence of the Organic Constituents of the Blood on the

Contractility of the Ventricle. Jour. Physiol., Vol. 6, p. 361.
'86 Further Experiments Regarding the Influence of Lime, Potassium and

Other Salts on Muscular Tissue. Jour. Physiol., Vol. 7, p. 290.
'93 The Influence of Carbonic Acid Dissolved in Saline Solutions, on the

Ventricle of the Frog's Heart. Jour. Physiol., Vol. 14, p. 125.
'97 The Action of Distilled Water on Tubifex. Jour. Physiol., Vol. 22, Proc.

of Physiol. Soc.
Ringer, S. and Buxton, D. W.

'84 Concerning the Action of Small Quantities of Ca, Na, and K Salts upon the
Vitality and Function of Contractile Tissues and the Cuticuler Cells of
Fishes. Jour. Physiol., Vol. 6, p. 154.
Seyler, C. A.

'94 Notes on Water Analysis. Chemical News, Vol. 70, p. 104.
Shelford, V. E.

'14 Suggestions as to Indices of the Suitability of Bodies of Water for Fishes.

Transactions Am. Fisheries Soc., Dec., 1914.
Shelford, V. E. and Alice, W. C.

'13 The Reactions of Fishes to Gradients of Dissolved Atmospheric Gases.

Jour. Expt. Zool., Vol. 14, No. 2, Feb., 1913.
'14 Rapid Modification of Behavior of Fishes by Contact with Modified Water.

Jour. Animal Behavior, Vol. 4, No. i, pp. 1-30.
Shelford, V. E. and Powers, E. B.

'15 An Experimental Study of the Movements of Herring and Other Marine
Fishes. Biol. Bull., Vol. 28, No. 5, pp. 315-334.

REACTION AND RESISTANCE OF FISHES. 257

Stieglitz, Julius.

'09 The Relation of Equilibrium between the Carbon Dioxide of the Atmosphere
and the Calcium Sulphate, Calcium Carbonate, and Calcium Bicarbonate
of Water Solutions in Contact with It. Carnegie Inst. of Wash. Pub. No.
107, p. 235.
Traube-Mengarini, M. and Scala, A.

'12 Die Wirkung des reinen und des elektrolythaltigen destillierten Wassers
auf Metalle. Zeitschrift fur Chemie und Industrie der Kolloide, Bd. 10,
p. 114-119.
Wells, M. M.

'13 The Resistance of Fishes to Different Concentrations and Combinations

of Carbon Dioxide and Oxygen. BIOL. BULL..VO!. 25, No. 6.
'14 The Reactions and Resistance of Fishes to Temperature. Transactions
111. Acad. of Science, Vol. VII., p. .

A PROCESS OF TEMPORARY CHAIN FORMATION

BY FRONTONIA.1

WALDO SHUMWAY.

In May of this year, while studying the protozoan fauna of a
small pond in the New York Botanical Gardens, the writer
observed what appeared to be a chain of four holophrya-like
ciliates. A careful search of the culture made during the fol-
lowing two weeks revealed about a dozen more such chains before
the species disappeared. During this time the writer attempted
to obtain a pure culture of these interesting forms, with an idea
to studying the nuclear changes involved. Although these
attempts were all, ultimately, unsuccessful, it has seemed best
to put on record what few facts have been observed. The
data here given has been obtained from the study of the living
material and some few preparations both in toto and sectioned.

The following stages in the formation of these chains have
been observed; single individuals, two-cell chains and four-cell
chains. No chains of three or more than four cells have been
observed among the large number (over one hundred and fifty)
observed.

The solitary individual is an exceedingly large (circa 300
micra) ovoid holotrichous ciliate, densely pigmented and filled
with large alveoli, of some alloplasmic substance which stains
deeply with nuclear dyes. The color is dark brown to black
with transmitted light and light brown to white with reflected
light. The mouth is anterior and lateral with rows of cilia which
simulate two undulating membranes. There is a large lateral
contracting vacuole. The cortical layer contains trichocysts.
The macronucleus is large and oval, the micronuclei have not
been observed in my material. The form is an active swimmer
in the surface film of a large culture, but when isolated in Syra-
cuse watch glasses sinks to the bottom and encysts. Details
of this process are given below.

In cases where the individual neither dies nor encysts after

1 From the Zoological Laboratory, Columbia University.

258

CHAIN FORMATION BY FRONTONIA.

259

isolation, it becomes transformed by clearing of the pigment,
lengthening and flattening the general shape of the body until
it resembles the well-known species Frontonia leucas. To this
species, therefore, I have assigned my material, although I am
not positive whether what I have observed is a new stage in the
life-cycle of Frontonia leucas or a new species.

The process of chain formation is inaugurated by the formation
of a large transparent cyst through the exudation of some gelati-
nous material in which the surrounding zooglcea becomes en-
tangled in large quantities. Within this cyst the individual
slowly rotates, all the cilia beating and the contractile vacuole
pulsating regularly. In one individual followed under the mic-
roscope a single transverse division occured about forty-five
minutes after the beginning of encystment, and half an hour
from its completion. Nine hours after, the daughter cells were
dead, still connected and inside the cyst wall.

FIG. i. Chain of four cells, Frontonia leucas. 200 X. Outline drawn with
camera from preparation fixed with sublimate-acetic and stained with picro-
carmine. Details restored from free-hand drawings of living material.

Another encysted individual which had been isolated at
the same time but which unfortunately had not been followed
under the microscope had formed a chain of four cells. The
writer has at different times observed in these cysts, single,
double, and quadruple forms, as well as cysts containing two or
four separate cells. One chain of four individuals was removed
from its cyst for observation: it swam about for a time resembling
an animated chain of beads, but finally broke up into its four

26O WALDO SHUMWAY.

constituent parts, which shortly assumed the typical Fron-
tonia leucas appearance.

While the writer has not actually observed the process of
transformation from the two-cell to the four-cell stage assumed
above, he feels convinced from the facts cited as well as from
sectioned material in which the macronuclei of the two anterior
and the two posterior cells seem to have just undergone division,
that these chains are formed by two successive transverse divis-
sions without separation of the daughter cells and not by a single
quadrupartite or by three successive terminal divisions as
might be argued from a priori grounds.

Division preceded by encystment is not unknown among
the free-swimming ciliates. In Otostoma carteri and Ampliilep-
tus meleagris according to Saville-Kent the divisions are some-
times multiple, but the daughter cells separate immediately
after each division.

Chain formation too is not unknown among the ciliates,
although heretofore it has been observed only among the asto-
matous forms parasitic in the digestive tracts of vertebrates.
In these parasites, however, the chains are often of great length
and are formed, so the weight of evidence appears, by a process
of terminal budding. For an excellent discussion of these forms
see Cepede (1910).

Jennings ('08) records a strain of Paramcecia appearing
in his cultures which had apparently lost the power of separation
after division. The general appearance and weak vitality of
these forms, however point to the conclusion that Jennings was
dealing with a weakened pathological race and not a genetic
mutation.

The frequency with which these chains appeared, nearly
ten per cent, of all the Frontonia observed, leads the writer
to conclude that the phenomena described in this paper form
part of a normal method of reproduction.

I have made attempts to raise these forms on hay infusion,
rotten lettuce infusion (recommended by Popoff, '08), thyroid
extract and the filtered water of the medium in which they
were discovered. All were unsuccessful. I have been unable
to maintain isolated individuals in Syracuse glasses of their

CHAIN FORMATION BY FRONTONIA. 26 1

normal medium longer than two days. For this reason I am
unable to give a fuller account of the process under discussion
nor carry out the experimental work suggested by it. The case
however appears to be a unique instance of normal temporary
chain formation among the free-living ciliates.

LITERATURE CITED.
Cepede, C.

'10 Recherches sur les Infusoires astomes. Arch, de Zool. Exp., 5 ser., III.,

4, 1910.
Jennings, H. S.

'08 Heredity, Variation and Evolution in Protozoa. Jour. Exp. Zool., V., 4,

1908.
Popoff, M.

'08 Experimented Zellstudien. Arch. f. Zellforsch., I., 2, 1908.
Saville Kent, W.

'81 Manual of the Infusoria, 1881-2.

SPERMATOGENESIS IN PARATETTIX.1

MARY T. HARMAN.

Wilson has said that "heredrity is a consequence of the genetic
continuity of cells by division, and the germ cells form the
vehicle of transmission from one generation to another."

If this be true we should look to the structure of the germ-cells
fo • an explanation of the phenomena that have been and are
being found out in heredity. Cytologists have discovered much
concerning the structure of the germ-cells and the behavior of
the chromosomes during the processes of maturation and division.
The combined knowledge of sex and sex ratio, and the cytologi-
cal constitution of germ-cells has shown in many forms, at least,
a correlation between the inheritance of sex and the dimorphism
of spermatozoa or eggs, or both. However, the vast amount
of cytological work has been done with forms the behavior of
whose characteristics in heredity is unknown. On the other
hand, much of the work in heredity has been done with forms
of which little or nothing is known of the structure of the germ-
cells. It is the writer's good fortune to have access to material
of which some of the ancestry is known for eighteen generations,
covering a period of five years.

For a number of years Dr. R. K. Nabours2 has been conducting
experiments with regard to inheritance in Paratettix, a genus of
the short-horned grasshoppers. The characteristics used in
his investigations are the color patterns of the pronotum and
femora of the jumping legs, and the lengths of the pronotum and
wings. The data show that the inheritance of the color patterns
is Mendelian in its behavior. In the FI hybrid no part of the
color pattern of one parent species is ever replaced by the color
pattern of the other parent, but the color patterns of both parents
are present. Reciprocal crosses give identical results. The

1 Contribution from the Zoological Laboratory, Kansas State Agricultural
College, No. 7.

2 The writer wishes to thank Dr. R. K. Nabours for the grasshoppers which
have furnished the material for this paper.

262

SPERMATOGENESIS IN PARATETTIX. 263

lengths of the wings and pronotum are not inherited but are
closely correlated with the length of time required for the animal
to reach maturity. These grasshoppers have furnished the
material for the present paper.

The work on the cytological constitution of the germ-cells of
Paratettix has been undertaken for the purpose of discovering
whether or not the microscope will reveal any differences in
the germ-cells of very closely related forms which may be corre-
lated with the differences in the color pattern. The spermato-
genesis of only one form (Paratettix leuconotus-leucothorax}
is given here. Paratettix leuconotus-leucothorax is a hybrid,
obtained by crossing P. leuconottis with P. leucothorax (or by
the interbreeding of two hybrids, one being a hybrid of leuconotus
with some other form and the other a hybrid of leucothorax
with some other form). No attempt has been made to show any
relation between the structure of the germ-cells and the somatic
structures. This will be discussed in a later paper.

The chromosomal complex of the spermatogonia of Para-
tettix leuconotus-leucothorax consists of thirteen rod-shaped
bodies which may be divided into two groups — one group con-
sisting of four larger chromosomes and the other of nine smaller
ones. Neither the larger nor the smaller chromosomes form
equal sized pairs as Sutton has found in Brachystola magna and
which is so frequently described for the Hemiptera and is ap-
parently characteristic of all Diptera. All of the large chromo-
somes and one of the small ones are bent rods or slightly U-
shaped, but the other eight are almost straight. No one of
these chromosomes has been surely identified as the accessory.
However, in the early prophases there is always present a mass
of chromatin which has a more compact consistency and stains
more intensely than the remainder of the chromatin (A, Fig. i).
This mass has not been identified with any chromosome nor is
it associated with a vesicle as described by Carothers for Arphia
simplex. There appears to be no difference in the staining
capacity nor in the compactness of the chromosomes in the late
prophases (Fig. 2).

The spermatogonial spindle is long and slender, and has fine
but distinct fibers which converge at the poles. The centrosome

264 MARY T. HARM AN.

which is very distinctly visible in the metaphase stage is small
and spherical. It stains almost as intensely as the chromosomes.
The astral rays are short and indistinct. The chromosomes are
at right angles to the spindle fibers in the metaphase stage
(Fig. 7). A metaphase plate always shows one chromosome
nearer the center of the spindle than any other chromosome.
Sometimes it is completely surrounded by the others (Figs. 4 and
6) and sometimes merely one end is at the center of the spindle
(Figs. 3 and 5). This chromosome is always one of the larger
of the group of smaller ones but it is never the bent one. Few
anaphase and no telophaase stages have been observed. Fig. 8
shows an anaphase with rather indistinct spindle fibers, which
is characteristic of all the anaphase stages observed. The cen-
trosome, which shows distinctly in the metaphase stage (Fig. 7),
is now invisible. The chromasomes are no longer at right angles
to the spindle fibers but are nearly parallel with them.

Fig. 9 illustrates the condition of the cell at the beginning of
the growth period. The nucleus is large and comparatively
clear. Some of the chromatin is in a finely reticular condition
and stains faintly with iron-hsematoxylin. However, a mass of
the chromatin retains the compact consistency and the density
of stain of the chromosomes (A, Fig. 9). It has a rounded form
like a nucleolus. The boundary between the nucleus and the
cytoplasm is quite evident. The nucleus continues to increase
in size and the reticular chromatin, which now has a greater
staining capacity, forms a thread or threads having a woolly
appearance. There is no polarization of the loops of the spireme
but they occupy almost all of the space of the nucleus and form
an irregular tangled mass. The nucleolus remains at one side
of the nucleus and does not have the woolly appearance that
the spireme has (Fig. 10).

In the synezesis or contraction stage, the spireme seems to
shrink away from the nuclear wall, leaving a clear space between
the cytoplasm and the chromatin material. There is little
difference between the character of the chromatin and its
staining capacity in this stage and the preceding one. The
compact mass of chromatin never loses its identity and always
remains at one side of the nucleus (Fig. n). The boundary of

SPERMATOGENESIS IN PARATETTIX. 265

the nucleus soon becomes irregular, and the chromosomes of
the primary spermatocyte is formed by a breaking up of the
spireme thread into segments. The compact intensely staining
mass which has been traced through the growth period is shown
in Fig. 12 as a chromosome which differs in shape from the
other chromosomes. It is ovoid and without a constriction in
the middle, while all the other chromosomes are dumb-bell
shaped. Not all of the chromosomes are formed at the same
time. The chromatin retains its loose woolly appearance, until
after it has broken up into parts, then it gradually becomes
more compact, takes the stain more readily and each part assumes
the characteristic dumb-bell shape. While this is taking place
the boundary between the nucleus and the cytoplasm becomes
more irregular and by the time the chromosomes are completely
formed the cytoplasm has formed a vesicle around each of them
(Figs. 13 and 14).

The chromosomal complex of the primary spermatocyte con-
sists of six dumb-bell shaped chromosomes and one ovoid chro-
mosome. Of the six dumb-bell chromosomes two are decidedly
larger than the others and one of these is much larger than the
other one, as is shown in Figs. 13 to 16 inclusive. The ovoid or
accessory chromosome is never among the other chromosomes
but always lies near the periphery of the nucleus as it did in the
growth period. When the chromosomes have become arranged
on the spindle the dumb-bell chromosomes are well toward
the center of the spindle, while the accessory is always near the
periphery. It does not remain long in the metaphase plate
but soon passes toward one pole undivided much in advance of
the other chromosomes. For this reason many sections of
metaphase plates show only six chromosomes and those which
show seven are often cut obliquely. Not all of the chromosomes
in the primary spermatocyte divide synchronously. Fig. 20
shows five of the dumb-bell chromosomes divided while the
largest one shows little constriction. This division is transverse
as is shown in Figs. 16 and 20. There are no loops, rings, or
U's which would give the least indication of a longitudinal
division. In the metaphase or early anaphases the centrosome
is a small spherical body and takes the stain readily. The

266 MARY T. HARMAN.

spindle fibers are fine but distinct. The astral rays are similar
to those of the spermatogonial divisions. In the late anaphases
the centrosome is no longer visible and the spindle fibers are
indistinct. There seems to be no resting stage between the
first maturation division and the second maturation division.

The chromosomes of the second spermatocyte are ovoid.
Metaphase plates show six and seven chromosomes (Figs. 23
and 24. The accessory cannot always be distinguished from
the other chromosomes. It is either the second or the third
largest. It divides in this division and passes to the poles
in advance of the others (Fig. 25). In the late anaphases all
the chromosomes have coalesced, although the number may yet
be distinguished (Fig. 26). By the time the chromosomes have
reached the poles they form a diffuse mass of chromatin at
each pole, and the cell has begun to constrict in the middle.

The centrosome which is similar to the centrosome of the
primary spermatocyte has disappeared. The spindle fibers have
become indistinct. As the constriction of the cell is completed
the chromatin has migrated to the center of each daughter cell.

In the changing of the spermatid into the spermatozoon two
things are conspicuous from the beginning, the changes in the
character of the chromatin and the elongation of the cell body.
From the diffuse irregular mass there is formed an ovate body
with the chromatin in a coarsely reticular condition largely
around the periphery of it. The cell becomes elongate and larger
at one end than at the other. The cytoplasm has changed from
the tangled network to fibrillar strands of granules extending
longitudinally across the cell. The cell wall is quite distinct.
This condition is illustrated in Fig. 29. The spermatid continues
to elongate. The nucleus becomes spherical and remains at
one end of the spermatid. The more granular cytoplasm lies
toward the periphery of the tail-like elongation. There is a
portion of the cytoplasm extending from the nucleus through
the center of the tail wilich is more finely granular than the
remainder and takes the stain less readily. Part of the chromatin
has become massed together, forming a sphere situated to the
side of the nucleus near the end of the lightly staining area of
the tail. The greater part of the remaining chromatin is dis-

SPERMATOGENESIS IN PARATETTIX. 267

tributed around the periphery of the nucleus (Fig. 30). As
the tail becomes longer it becomes thinner and a filament extends
from the nodule of the head through the entire length of the
tail. Very little cytoplasm now surrounds the head (Fig. 31).
Finally the head becomes arrow shaped and stains very intensely.
The tail is long and filamentous and stains a little less intensely
than the head. The head and a portion of the tail are illustrated
by Fig. 32. The tail is more than four times as long as is shown
in the figure.

McClung ('14) says: "It seems very evident that in the
spermatogonia of the Acrididae we are dealing with a chromo-
some complex of a very definite and precise organization which,
in the great majority, presents itself without essential variation
in number, size and form, fiber attachment, arrangement in the
metaphase and behavior during division of its elements. Steno-
bothrus and Pamphagns seem to be definite exceptions in some
of these respects. . . . And again he says:

"With the exception of the Stenobothrus-like species, and
Pamphagns, the students of the Acrididae have reported a reduc-
tion of the 23 spermatogonial chromosomes to 12."

All of the genera of the family Acrididae1 discussed in McClung's
paper belong to the three subfamilies, Tryxalinae, (Edipodinae,
and Acridinee, and none belong to the subfamily Tettiginae.
It would seem that with the general agreement of the great
numbers of genera of this family which he and his students have
studied as well as those of other independent workers that he
would be justified in making the general statements concerning
the family. However, Paratettix leuconotus-leucotliorax of the
subfamily Tettiginae, show some exceptions. The spermato-
gonial number of chromosomes are thirteen instead of twenty-
three. The number of chromosomes in the primary spermato-
cyte is seven. This the writer has found to be true also for
both the parent forms of the hybrid as well as for others belonging
to the genus Paratettix.

1 The writer is aware of the fact that there has been much shifting about of
names of the short-horned grasshoppers and that some taxonomists consider the
the grouse locusts of family value. If this should be the position which McClung
takes, then he would not consider Paratettix as belonging to the family Acrididae
and it would follow that the observations recorded in this paper would not be excep-
tions to his statements concerning Acrididae.

268 MARY T. HARMAN.

Robertson ('15) says: "In the Tettigidae (Tettiginse) a sub-
family of the short-horned grasshopper family Acrididse, I have
found for all the specimens of at least four different genera which
I have examined the number of chromosomes to be uniformly
14 in the female and 13 in the male."

From the above data one would scarcely be justified in saying
that the characteristic number of spermatogonial chromosomes
of the subfamily Tettiginae is thirteen but the writer feels justified
in saying that this is an essential variation in the number of
chromosomes given in the above quotation from McClung as
the number characteristic for the family Acrididae. The writer
has found no indication of multiple chromosomes.

In the prophase tetrad six of the chromosomes are always
dumb-bell-shaped and one ovoid. There are none of the irregu-
lar shaped chromosomes as described by McClung and no in-
dications of the annular chromosomes which he says "that
practically without exception every investigator of recent years
who has made a careful study of the maturation stages in the
Orthoptera has seen and figured." If the dumb-bell-shaped
chromosomes are similar to his I-shaped chromosomes they
differ in that they do not have an enlargement in the middle,
but rather they have the appearance of a constriction. This
constriction is not due to the initiation of the division, for it is
present before the chromosomes are arranged on the spindle;
in fact, they have their characteristic shape before the spindle
is visible.

The presence of a mass of chromatin in the resting stage of
the spermatogonial divisions which is of a different form and
different staining capacity and also the presence of a similar mass
in the growth period, which can be identified as the accessory
chromosome of the spermatocyte, gives added evidence for the
continuity of chromosomes as definite entities.

The spermatozoon of Paratettix leuconotus-leucothorax is very
different from the spermatozoon of Paratettix cuculatus as
described by Hancock. He describes and figures the head of
P. cuculatus as being small, thin, and acutely pointed. In fact,
from his figure one would think that the head is very little thicker
than the tail. He says that the middle piece is formed into a

SPERMATOGENESIS IN PARATETTIX. 269

high, rather short, protoplasmic keel. No keel has been observed
as forming a part of the middle piece of P. leuconotiis-leucothorax.
The head is many times thicker then the tail and is decidedly
arrow shaped. The middle piece seems to continue from the
head to the long filamentous tail without a definite dividing line
between them.

SUMMARY.

1. Paratettix huconotus-leucothorax has thirteen spermatogo-
nial rod-shaped chromosomes, four larger and nine smaller ones.

2. Neither the larger nor the smaller chromosomes form equal
sized pairs.

3. The four larger chromosomes and one of the smaller ones
are bent rods, the others are almost straight.

4. Neither spermatogonial chromosome has been surely
identified as the accessory chromosome.

5. In the growth period is a mass of chromatin which is more
compact and stains more intensely than the remainder of the
chromatin. This is the accessory chromosome of the first
spermatocyte.

6. The first spermatocyte chromosomes are formed in vesicles.

7. There are six dumb-bell-shaped bivalent chromosomes
and one ovoid univalent chromosome in the primary spermato-
cyte.

8. The accessory is near the periphery of the nucleus and
passes to one pole undivided slightly in advance of the other
chromosomes in the first spermatocyte division.

9. The bivalent chromosomes divide crosswise.

10. The accessory chromosome divides in the second division.

11. The spermatozoon has an arrow-shaped head and a long
filamentous tail.

BIBLIOGRAPHY.
Artom, C.

'09 Cromosomi ed eterocromosomi nelli cinesi spermatogenetiche di Stauronotus-

maroccanus Thumb. Biologica, Vol. 2.
Buchner, P.

'09 Das accessorische Chromosom in Spermatogenese und Oogenese der Or-
thopteren, Zugleich ein Beitrag zur Kenntnis der Reduktion. Archiv fur
Zellforschung, Bd. 3, Hft. 3.
Brunelli, Gustavo.

'10 La Spermatogenesi della Tryxalis. Division! Spermatogoniali. Memorie
della Societa Italiana Scienze, Tome 16, serie 30.

27O MARY T. HARMAN.

'n La Spermatogenesi della Tryxalis. Division! maturative. Reale Academia

dei Lincei, Tome 8, serie 5a.
Carothers, Eleanor.

'13 The Mendelian Ratio in Relation to Certain Orthopteran Chromosomes.

Journ. Morph., Vol. 24, No. 4.
Davis, H. S.

'08 Spermatogenesis in Acrididse and Locustodae. Bull. Museum Comp. Zool.,

Harvard College, Vol. 43, No. 2.
Gerard, Pol.

'09 Recherches sur la Reduction Karyogamique dans la Spermatogenese de
Stenobothrus biguttulus (L.). Bull. Societe Roy. Sciences med. nat. de
Bruxelles, No. i.
'09 Recherches sur la Spermatogenese chez Stenobothrus biguttulus (Linn.).

Archives de Biologic, Tome 24.
Granata, Leopoldo.

'10 La cinesi Spermatogenetische di Pamphagus marmoratus (Brum). Archiv

fur Zellforschung, Bd. 5, Hft. 2.
Hancock, J. L.

'02 The Tettigidae of North America. The Lakeside Press. R. R. Donneley

& Sons Company, Chicago.
Hartman, Frank.

'13 Variations in the Size of Chromosomes. BIOL. BULL., Vol. 24, No. 4.
McClung, C. E.

'oo The Spermatocyte Divisions of the Acrididae. Kansas U. Quarterly, Vol.

9, No. i.
'02 The Accessory Chromosome — Sex Determinant? BIOL. BULL., Vol. 3,

Nos. 1-2.
'05 The Chromosome Complex of Orthopteran Spermatocytes. BIOL.

BULL., Vol. 9, No. 5.

'08 Cytology and Taxonomy. Kansas U. Science Bull., Vol. 4, No. 7.
' '14 A Comparative Study of the Chromosomes in Orthopteran Spermatogenesis.

Journ. Morph., Vol. 25, No. 4.
Meek, C. F. Y.

'n The Spermatogenesis of Stenobothrus viridulus, with Special Reference to
the Heterotropic Chromosome as a Sex Determinant in Grasshoppers.
Linnean Society's Journ. Zool., Vol. 32, No. 208.

'i2& A Metrical Analysis of Chromosome Complexes, showing correlation of
Evolutionary Development and Chromatin Thread-width Throughout the
Animal Kingdom. Philosoph. Transactions Roy. Soc. of London, Series
B., Vol. 203.

7i2b The Correlation of Somatic Characters and Chromatin Rod-lengths,
being a further study of Chromosome Division. Linnean Society's Journ.
Zool., Vol. 32.
Metz, Charles W.

'14 Chromosome Studies in the Diptera. I. A Preliminary Survey of Five
Different Types of Chromosome Groups in the Genus Drosophila. Journ.
Exper. Zool., Vol. 17, No. i.
Montgomery, T. H., Jr.

'05 The Spermatogenesis of Syrbula and Lycosa, with general Considerations

SPERMATOGKXKSIS IN PARATJ.TIIX. 271

upon Chromosome Reduction and the Heterochromosomes. Proc. Academy
Nat. Sci. Philadelphia, February.
Nabours, R. K.

'14 Studies of Inheritance and Evolution in Orthoptera, I. Journ. of Genetics,

Vol. 3- No. 3.
Nowlin, Nadine.
'08 The Chromosome Complex of Melanoplus bivitattus Say. Kansas U. Sci.

Bull., Vol. 4, No. 12.
Pinney, Edith.

'08 Organization of the Chromosomes in Phrynotettix magnus. Kansas U.

Sci. Bull., Vol. 4, No. 14.
Robertson, W. R. B.

'08 The Chromosome Complex of Syrbula admirabilis. Kansas U. Sci. Bull..

Vol. 4, No. 13.

'15 Chromosome Studies. III. Inequalities and Deficiencies in Homologous
Chromosomes; their Bearing upon Synapsis and the Loss of Unit Char-
acters. Journ. Morph., Vol. 26, No. i.
Wilson, E. B.

"13 Heredity and Microscopical Research. Science, N.S., Vol. 37, No. 961.

272 MARY T. HARMAN.

EXPLANATION OF PLATES.

All figures were made at table level by means of a Zeiss compensating ocular
No. 6 and a 1.5 mm. objective with the aid of a camera lucida. The drawings
were enlarged two diameters and then reduced one third.

PLATE I.

FIG. i. Early prophase of spermatogonial division. A, a mass of chromatin
which does not become reticular but remains more or less compact.

FIG. 2. Formation of spermatogonial chromosomes.

FIGS. 3 TO 6. Metaphase plates of spermatogonial chromosomes.

FIG. 7. Metaphase, lateral view, spermatogonial division showing position of
the chromosomes on the spindle.

FIG. 8. Anaphase of spermatogonial division.

FIG. 9. Beginning of the growth period. A, a mass of chromatin which does
not pass into a reticular condition and forms the accessory chromosome.

FIG. 10. Formation of chromatin thread.

FIG. ii. Synizesis.

FIG. 12. Beginning of the formation of the primary spermatocyte chromosomes.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE I.

g

,/.

..v-.>vY ,'</•£• -

&W

»

;

MARY T. HARMAN

274 MARY T. HARMAN.

PLATE II.

FIG. 13. Primary spermatocyte chromosomes showing the beginning of the
formation of the vesicles.

FIG. 14. Primary spermatocyte chromosomes in vesicles.

FIGS. 15 AND 16. Metaphase, lateral view, of first spermatocyte division. The
accessory chromosome is at the periphery of the spindle.

FIGS. 17 AND 19. Metaphase plates of first spermatocyte division showing
seven chromosomes.

FIG. 18. Metaphase plate of first spermatocyte division showing six chromo-
somes. The accessory chromosome is not in the metaphase plate.

FIG. 20. Beginning anaphase of first spermatocyte division.

FIG. 21. Late anaphase of first spermatocyte division showing seven chromo-
somes.

FIG. 22. Metaphase, lateral view, of second spermatocyte division showing
seven chromosomes.

FIG. 23. Metaphase plate of second spermatocyte division showing seven
chromosomes.

FIG. 24. Metaphase plate of second spermatocyte division showing six chromo-
somes.

BIOLOGICAL BULLETIN VOL. XXIX.

PIATE II.

/^[^yTxptK

' ' ~ '^^ si-%

9! './•; .

•'-? •-

m

• ' *-rrv. ' ,

"^•^u

.

23

A «

.- .:.--.;-x-.

WARY T. HERMAN

276 MARY T. HARM AN.

PLATE III.

FIG. 25. Anaphase of second spermatocyte division showing seven chromo-
somes going to each pole.

FIG. 26. Late anaphase of second spermatocyte division.

FIG. 27. Telophase of second spermatocyte division.

FIG. 28. Spermatids.

FIGS. 29 TO 31. Stages in the development of the spermatozoon.

FIG. 32. Head and part of the tail of a spermatozoon.

BIOLOGICAL BULLETIN, VOL. xxix.

, " X

f ' .. . • X

'^""'"

~--

' ^A.

31

32

MARY T. HARMAN

Vol. XXIX. November, 1915. No. j

BIOLOGICAL BULLETIN

NOTES ON THE BEHAVIOR OF THE ANT-LION WITH

EMPHASIS ON THE FEEDING ACTIVITIES

AND LETISIMULATION.

C. H. TURNER,
SUMNER HIGH SCHOOL, ST. Louis, Mo.

INTRODUCTION.

The ant-lion is one of the marvels of the insect world and is
discussed in practically every text-book on entomology and in
almost every popular book on insects. With the exception of
results derived from attempts to analyze the behavior of these
insects into tropisms (4), European papers may be epitomized
as follows: (i) the pits are formed in sand that is protected from
the weather; (2) the larva excavates this pit by moving backward
in a constantly narrowing spiral and using its abdomen as a
plowshare and its head for a shovel; (3) with one of its forelegs,
the ant-lion scrapes the sand on to its head from the inner side
of the spiral; (4) with its body entirely concealed, the larva lies
in ambush, with its open jaws resting in the bottom of the finished
pit; (5) by tossing up sand at random, the ant-lion forces insects
that alight on the side of the pit to tumble to the bottom; (6) any
small terrestrial invertebrate may become its prey; (7) there is
no mouth opening, the food being imbibed through tubes formed
by each mandible and another mouth-part.

In American scientific journals, I have been able to find only
four articles treating of our ant-lion. The first and the longest
of these is by Emerton (6). In the fall of 1870, he found a pit
of Myrmeleon immaculatiis De Geer, under the shade of a boulder,
at Danvers, Mass. The larva was carried home and placed in a
bowl of sand. Immediately it buried itself. After remaining
beneath the surface for several days, it excavated a pit. No

277

278 C. H. TURNER.

mention is made of the larva using the foreleg to shovel sand on
the flat head in the manner described by European and American
popular writers. The larva was fed with flies. When given
more than one at a time, it would kill all before eating any. It
was kept over winter in a warm room. In May it spun its cocoon
beneath the sand. In June the adult emerged, leaving half its
pupa skin in the cocoon.

In August, 1871, Birge (2) found a colony of 600 ant-lions,
under an overhanging cliff, in Albany County, N. Y. These
pits were in a soil composed of fine disintegrated limestone
commingled with pebbles and minute fragments of stone.
Whenever an insect alighted upon the sides of the pit the ant-
lion began to toss up the soil in all directions.

Moody (14) states that the ant-lion observed by him rests at
the bottom of the pit with its jaws only showing, and that it
throws up sand at escaping prey. It formed its cocoon June 4
and emerged July 8.

Moffat (13) writes: "Fine loose sand is evidently a necessity
of their existence in any locality." He mentions the throwing
up of sand when an ant steps on the side of the pit.

THE PIT.

Most accounts give the erroneous impression that the pits of
ant-lions are formed only in sand. Even in a scientific magazine,
Moffat (13) writes, "Fine loose sand is evidently a necessity of
their existence in any locality." A loose friable soil protected,
more or less, from rain and shielded from chickens and similar
insect feeders is all that is needed. It may be anything from the
finest dust to coarse sand. In open sheds with dirt floors, under
porches where the place is not too dark, beneath low railroad
bridges that span sandy, dusty, or cindery ground, under ledges
of rock, and beneath the shelter of logs that do not touch the
ground at all points are good places to look for them. From
time to time, during the past three years, I have had, in my
insectary, more than 500 ant-lions. Many of these were ob-
tained in Kansas by my friend, Mr. Phil Rau; the remainder
were collected in and about St. Louis. The majority of these
were found in the loamy clay (loess) that forms much of the top

BEHAVIOR OF THE ANT-LION. 279

soil of Missouri and Kansas; some were found in cinders in shel-
tered places along railroad tracks; some, in disintegrated mortar
along the walls of buildings; some, in the rotton-wood dust of
hollow logs; none of these was obtained from sand. In the wide
jelly glasses of my insectary, where each larva was kept in solitary
confinement, some were placed in loamy clay, some in sifted coal
ashes, some in coal ashes that had been weathered for a year,
some in fine sand and others in coarse. The larvae seemed to
flourish as well in one medium as in the other.

The ant-lion usually begins the construction of its pit by striking
out a circle in the friable earth. Using its abdomen as a plow-
share and its head as a shovel, the larva burrows backward, in a
circular path, just beneath the surface of the soil, tossing upward
and outward the dirt that falls upon its head. Almost all of
the articles that I have read state that this initial circle marks
the outer boundary of the finished pit. With the American
ant-lion of the Middle West this is not always so. In most
cases observed by me the finished pit is wider than the diameter
of this circle. In the first place, the falling inward of the soil
as the excavation progresses enlarges the diameter. Then, too,
the ant-lion sometimes enlarges the partly or apparently entirely
completed pit. After this first circle is completed, within this
the ant constructs, in a similar manner, a deeper adjacent circle
and so on until the center is reached. Then, with the major
portion of its body hidden in the walls of the pit and using its
head and mandibles as a shovel, it tosses out the material from
the bottom of the pit, until the dirt no longer runs down the
sides.

European writers state that the ant-lion shovels sand on to its
head by means of one of its forefeet, and Kirby and Spence (8)
insist that, in excavating its burrow, the ant-lion reverses the
direction it is going at the completion of each circle, so as to alter-
nately exercise each foreleg. In our American ant-lion this
pair of forelegs functions, not as a scrape, but as a brace to the
body when the ant-lion is shovelling dirt or turning. Patient
watching with a magnifying glass has failed to detect the fore-
foot loading dirt upon the head; and certainly the ant-lion does
not reverse the direction it is going every time it completes a

28O C. H. TURNER.

circle. The dirt gets upon the head by falling from above and
from the sides, as the larva burrows backward through the soil.
Some of this material comes from the outer edge as well as from
the inner. While constructing its pit, the larva often pauses.
After each rest it usually continues in the direction that it was
going. On rare occasions, it does turn about and go in the
opposite direction. This is usually when it has met some ob-
struction.

To test the matter more thoroughly, a sufficient portion of each
of the forelegs of an ant-lion was amputated to render them
much too short to be of value in shovelling soil on to the head.
As soon as it was returned to its glass, the larva burrowed back-
ward into the soil. For four days it remained beneath the surface.
On that day it excavated a small pit. The next day the pit had
been enlarged. On this day it was fed with ants (Formica
subsericea). The ant-lion was then removed from the soil and
examined under a simple microscope. The legs had not re-
generated; each stump was covered with a ball of soil. This ant-
lion had constructed its burrow without using its front feet as
scrapers with which to load dirt on the head.

The force with which an ant-lion tosses the materials from its
pit is astonishing. Often they are cast several inches beyond the
rim. Sometimes the larva encounters particles which cannot
be disposed of with a toss of the head. When these are not too
heavy the insect has an unique method of disposing of them.
The insect backs up the side of the pit with the obstacle poised
on the posterior portion of its abdomen and deposits it beyond
the edge of the pit. Although this behavior is described by
Bingley (19), most writers do not mention it. Perhaps it sounds
too much like a fairy tale; yet it is comparatively easy to induce
an ant-lion to behave in this manner. I frequently induced it
either by placing a small stone in the center of the ring of a pit
that was being constructed; or, by depositing a similar object in
the bottom of a completed pit. When the stone is placed in the
center of the ring, as the ant burrows spirally inward, there is
sure to come a time when the stone will fall into the furrow.
When the ant-lion returns to that point it encounters the obstacle.
Usually it burrows under the object and continues on part of

BEHAVIOR OF THE ANT-LION.

28l

the way around the circle. Then, turning, it backs through the
furrow thus made until it has inserted the tip of its abdomen
under the impediment. It then backs slowly up the slope with
the burden poised upon the tip of its abdomen. The edges of
the abdominal somites and the stiff bristles thereon prevent the
stone from slipping forward; while the dirt on each side prevents
it from falling sidewise. Throughout this entire upward journey
the whole body of the ant-lion is above the ground. It is an
astonishing sight to see the insect backing, in almost a straight

Fig.l 1

Larval ant-lion. Dorsal view.

Larval ant-lion. Ventral view.

line, up the steep slope, with the burden poised on its back.
When the burden has been disposed of, usually at the edge of
the pit, the ant-lion turns about and returns to the bottom of the
pit, usually in the furrow made by the upward struggle, and con-
tinues her digging. The furrows made before my eyes have always
been straight or nearly so; but, one made in my absence was
quite curved. When the object was placed in the bottom of a
finished pit, sometimes the object was allowed to remain; but,
in most cases, sooner or later, it would be removed, in the follow-

282 C. H. TURNER.

ing manner. When it had tossed up a few loads of dirt, the larva
would back away from the obstacle in a straight or a curved line;
then turning, it would back through the furrow thus made and
proceed as described above. When the stone is too heavy for
the insect to handle in the manner mentioned above, it either
deepens the pit on one side of the obstacle, or buries the obstacle
by mining under it, or else abandons the pit. In several impor-
tant respects the behavior observed by me differs from that
described by Bingley; (i) never once did I see the stone fall from
the insect's back and roll to the bottom of the pit, (2) obstacles
encountered in constructing the pit were usually removed at
once, (3) such bodies were usually deposited just beyond the rim
of the pit, (4) occasionally they were left on the side of the pit.

On rare occasions I have seen pits constructed in a different
manner. Instead of beginning by striking out a circle, the ant-
lion burrowed downward into the ground and began at once
casting out the soil, thus making a pit of small diameter. Usually
such pits were afterwards enlarged by burrowing into the walls
and proceeding about as described above.

Thus my experience with the pit-building behavior of the ant-
lion harmonizes with McCook's account (10) ; but is not in accord
with that of Mrs. Comstock (5) who writes: "Wonderful stories
are told about the way ant-lions dig their pits, marking out the
outer margin in a circle and working inward. However, our
common ant-lion of the East simply digs down into the sand and
flips the sand out until it makes a pit."

The magnitude of the pit and the slope of the sides depend
upon the size of the larva and the nature of the soil ; the coarser
the individual particles and the greater their specific gravity the
more gentle the slope. In the loess about here the pits vary in
diameter from less than half an inch to about two inches; the
latter being the more abundant. Often the depth of the pit is
almost as great as the diameter. Although a small ant-lion
usually excavates a small pit, a small pit does not necessarily
contain a small larva; for large larvae sometimes construct small
pits which they afterwards enlarge.

Occasionally one finds an isolated ant-lion pit; usually they
occur in groups (Figs. 10, n). In the same cluster the ant-lions

BEHAVIOR OF THE ANT-LION.

283

differ much in size, and this is true even in the early spring.
Certain writers attribute these differences in size to the fact that
some obtain more food than others. The following simple
experiment lends support to this view. From a certain well-
circumscribed area, containing about fifty ant-lion pits, a dozen
larvse were removed, on June 22, and placed in my insectary.
A portion of these were well fed daily, the remainder were fed
only occasionally. A few were lost by accidents. By August
8 all of the survivors of the well-fed lot had formed cocoons and a
few imagoes had emerged. The poorly fed individuals were
still larvae. The majority of those left in the field were still
larvse.

Morphologically the ant-lion (Figs. 1-3) is well adapted to
this pit-building behavior. The flat head, which, with the stout
mandibles, forms an excellent shovel, is so articulated to the rest

Fig. 3

Larval ant-lion. Lateral view.

of the body that it is possible to give it a powerful upward jerk.
The abdomen is flat on the ventral and convex on the dorsal
side, the whole tapering toward the tip in such a manner as to
form an excellent burrowing instrument. From the sides of the
body clusters of stiff bristles project outward and forward in
such a manner that the body is prevented from slipping forward
after it once has penetrated the earth. Then, too, the terminal
claws on the legs (Fig. 12) make efficient anchors. The front of
the dorsal portion of the prothorax is so rounded that dirt easily

284 C. H. TURNER.

falls forward and loads the shovel-like head. There is no
functional anal opening; hence there is no danger of vigorous
thrusts of the abdomen clogging the intestine with dirt.

FEEDING BEHAVIOR.

The finished pit is an inverted hollow cone, at the apex of which
the wide-open mandibles of the larva, with their sharp teeth,
await to grasp any unfortunate that happens to fall therein.
What an efficient trap for small creeping invertebrates ! The steep
and unstable sides often cause the animals to fall to the bottom.
If the intruder does not at once slide to the bottom, its struggles
to escape tumble the soil upon the mandibles of the waiting ant-
lion. Immediately the ant-lion begins to toss the soil upwards.
The claim that the dirt is cast at the struggling creature is
erroneous. Digging its mandibles edgewise into the bottom of
the side of the pit, the ant-lion shovels out head-load after head-
load of soil. It is not thrown at something; it is simply tossed
upward and outward. Some of these random shots may take
effect; and the constant undermining of the walls of the pit
produces miniature landslides which, usually, drag the prey to the
bottom of the trap.

Until something falls into the pit or alights on its treacherous
sides, these mandibles of the larva usually rest horizontally in
line with the body, which is hidden in the wall of the pit. Or-
dinarily the pits appear to be empty, for the mandibles are often
covered with fine dirt. Even when the whole head is uncovered,
its color harmonizes so perfectly with that of the soil as to render
it invisible. As soon as its jaws close upon a creature the ant-
lion backs deeper into the walls of the pit and, by interring its
victim, subdues it. Thus the ant-lion is enabled to conquer
creatures that are much larger and apparently stronger than it.
Unless its first few struggles free it, the captive is doomed; for
the ant-lion slowly but surely drags it deeper and deeper into
the soil, while it feasts on its body juices. To assist in holding
the prey while its body contents are being imbibed through the
hollow mandibles, each mandible is provided, on its inner
surface, with three stout teeth (Fig. 13.4).

MacLachlan (n), in discussing the feeding of the European

BEHAVIOR OF THE ANT-LION.

285

ant-lion (Myrmeleon formicarius) says: "The house flies and other
small insects were usually dragged partially or wholly under the
sand, whilst blue-bottles and similar bulky creatures were feasted
upon on the surface." In his account of the feeding of the
common ant-lion of the east, McCook remarks: "The ants were
held off at 'arm's-length,' so to speak, and were thrashed and
jerked about until they were exhausted. Meanwhile efforts at
defence were made futile by the captor, who held its victim out
of reach of any vital part." Neither of these accounts tallies
exactly with my experience, although MacLachlan's can be har-
monized with it. I have watched ant-lions feed thousands of
times and have fed them with a variety of invertebrates. In
every case the larva has attempted to drag its captive beneath
the ground. In no case was the insect held off at arm's length

Empty cocoon of ant-lion.

as described by McCook. Often, a few moments after its capture,
all that would be visible of an ant were the 'tips of its waving
antennae, or the extremity of its wriggling abdomen, or both.
Naturally the captive struggled and squirmed; but there was no
attempt on the part of the ant-lion to hold its prey at arm's length
above the ground, while it thrashed it and jerked it. If the first
closing of the mandibles does not capture the creature that

286

C. H. TURNER.

happens to fall into the pit, remaining at its post, the ant-lion
elevates its head and makes repeated snaps at the creature as
long as it remains near. It may be that the ocelli located at the
base of the mandibles, on the dorsal side of the head, aid in this.
The name ant-lion is a misnomer; for it creates the impression
that this insect feeds exclusively, or almost exclusively, upon ants.
Such is not the case. Any small creeping invertebrate — be it
insect, crustacean, or arachnid — is acceptable. Several of the most
flourishing colonies of ant-lions found near St. Louis are located in
the dirt floor of a dilapidated stone-crusher of an abandoned quarry.
The diet of the inmates of those pits is composed largely of sow-
bugs (Porcellio). Emerton (6) and MacLachlan (i i) fed their ant-
lions on living flies that had been disabled; Berce (i) reared his
on living flies, wood-lice and earwigs. I supplied mine with

Fig. 5

FIG. 5. Chrysalis of ant-lion that died on way to surface.

living specimens of the following invertebrates; caterpillars (even
hairy ones), wood-lice, small roaches, small moths (held by the
wings until the ant-lion had secured a hold), spiders, nymphal
squash bugs, ants, small beetle larvae, soft-bodied beetles, and
bed-bugs. All of these were accepted and, after the juice had
been sucked from each body, the dried remains were cast out of
the pit.

The ant-lion has no mouth opening in the true sense of the
word. The strong curved mandibles are perforated at the tip,

BEHAVIOR OF THE ANT-LION.

287

and along the ventral surface of each there runs a prominent
tube through which the juices of the victim are sucked. This
tube is composed of two parts. Along the whole of the ventral
surface of each mandible [Fig. 13-4] there is a deep groove with
incurved edges. Another mouth part [Fig. 13.8], probably the
maxilla, fits so tightly into this groove of the mandible that, even
when viewed with a 2/3-inch objective, the two seem to form a
single structure. With that power, on the underside of each
mandible one sees two ridges. These mark the junctions of the
two pieces; but, unless you had previously dissected a mandible,
you would not suspect that there were two pieces and that they
were not rigidly united. Turn the insect on its ventral surface,

FIG. 6. Shed chrysalis skin of ant-lion.

carefully disarticulate the mandible, and, with a pair of dissecting
needles, gently push it forward. Thus the other mouth part will
be gradually drawn out of the mandible and left attached to the
ventral part of the front of the head.

The ant-lion preys upon living invertebrates. How does it
distinguish the living from the not-living? There may be several
factors which help it solve this problem. The following experi-
ments show that one attribute by means of which the ant-lion
differentiates between desirable and undesirable prey is the
exhibition of restlessness:

Experiment i. — I fastened a bit of straw to the end of a silk thread.

288 C. H. TURNER.

Twirling the other end of the thread between my thumb and fore-
finger, I gently lowered it to the bottom of the pit. Three times in
succession the ant-lion caught hold of the straw with its man-
dibles. Each time I jerked the string and thus removed the
straw from its grasp.

Experiment 2. — I fastened a dead chinch bug to one end of a silk
thread. Twirling the other end of the thread between my thumb and
fore-finger, I gently lowered the bug to the bottom of the pit. Im-
mediately the ant-lion seized it with its mandibles and held on
until, by pulling on the thread, I had drawn the insect partly
out of the pit. This experiment was tried with five different
individuals. Four responded in the manner just described; the
fifth made no response.

Experiment j . — / fastened a piece of cotton to one end of a silk
thread. Twirling the other end between my thumb and fore-finger,
I gently lowered it to the bottom of the pit. The ant-lion gripped
the bit of cotton with its mandibles and held on until I, by pulling
on the thread, had dragged the larva partly out of the pit. This
experiment was tried with five different individuals. The result
was always the same.

MacLachlan (u), in 1864, placed between two and three dozen
ant-lions in a small box of sand and carried them from Fontain-
bleu, France, to London. When he arrived, about half of the
larvae had been killed and the juices of their bodies extracted by
the others. In shipping ant-lions to me from Kansas, Mr. Rau
placed a hundred or more in the same small box of dirt. In
sorting over the material to place each one in an individual
retainer, I always found several dead specimens that looked as
though the juices of their bodies had been extracted. Are these
deaths caused by cannibalism? To test the matter, an ant-lion
was dropped into the pit of another individual. This experiment
was repeated over and over again throughout a summer devoted
largely to the field study of this creature. In the majority of
cases the intruder escaped either by burrowing into the wall of
the pit or else by backing out of it. In several instances, how-
ever, it became the prey of the rightful owner of the pit. Evi-
dently, when opportunity permits, this creature is a cannibal.

BEHAVIOR OF THE ANT-LION.

289

LOCOMOTION.

The forms of locomotion used in excavating pits and in re-
moving obstacles therefrom have been described in the section
on "The Pit" and will not be repeated here.

When placed on loose dry soil, the ant-lion may letisimulate.
As soon as it begins to move, it burrows backward into the ground.
If an ant-lion is placed in an open rectangular pasteboard box, it
backs along, sometimes in a straight line and sometimes in a
curved line, until it comes in contact with one of the sides. It
then backs along that side until it comes to a corner, turning the
corner it continues along to the next corner. It may continue
thus for a long time, or it may vary it by creeping backwards up

Fig. 7

FIG. 7. Cocoon of ant-lion, with chrysalis partly emerged. This cocoon is
from a form that was raised in shifted coal ashes.

one of the angles until it reaches the top of the box and then pass
downward to the ground. After it has once reached the side of
the box, no matter how long it remains within the box, it almost
never moves out into the open. These two simple experiments
indicate that this insect is positively thigmotactic. With this
statement must be coupled the reservation that, at times, the
creature moves about with all of the upper portion of the body

290

C. H. TURNER.

exposed. This is the case when it is removing an obstacle from
its pit.

Since this larva burrows downward into the earth, it may be
considered positively geo tactic; but, it must be remembered that
it does not always pass downward. When disturbed in its pit,
it usually backs upward, just beneath the surface, until the rim
is reached; sometimes, it continues onward, in a horizontal
direction, beneath the surface. MacLachlan (n) observed that,
at night, they frequently make perigrinations over the surface
of the ground. Then, too, they sometimes ascend vertical
surfaces.

When placed on a horizontal surface [I used sheets of brass,

Ant-lion pit in one of my tumblers.

glass, wood, and cardboard], the larva backs away from the
light. If placed with the tip of its abdomen toward the source
of light, usually, it will move a short distance toward the light
then turn, to either the right or the left, and back away in a
straight line. This, coupled with the fact that, when placed on
its back, the ant-lion invariably rights itself by turning away from
the source of light, induces the conclusion that this creature is
negatively phototactic; but, it must be remembered that in

BEHAVIOR OF THE ANT-LION.

291

constructing its burrow the larva crosses the light at every possible
angle, and that, at times, it moves toward the light.

This insect invariably moves backward; never under any
conditions does it move forward; yet it is capable of performing
all the ambulatory feats possible to an insect that progresses in
the orthodox way. It can move in a straight line, it can describe
simple or s-shaped curves to either the right or the left, and it
can ascend or descend rough surfaces inclined at any angle from
zero to ninety degrees. By means of wrhat structures does it
perform these movements? Are they produced entirely by
flexing the abdomen and trusting to the body bristles to prevent

Fig. 9

Photograph showing the comparative size of cocoon, chrysalis and imago.
They belong to the same ant-lion and are photographed to the same scale. A,
cocoon; B, chrysalis skin; C, imago.

forward movements? Do the legs take any part in the move-
ments? Do the mandibles assist? To answer these questions
the following experiments were devised.

Experiment I. — An ant-lion was placed on a glass plate arranged
horizontally. By means of a hand magnifying glass every movement

292 C. H. TURNER.

was watched. It moved backward in jerks. The hind legs, which
were doubled back under its abdomen, made jerky pulls. The
middle pair of legs was directed outwards in almost a straight
line. The anterior pair of legs was stretched forward. The
tips of both the first and second pair of legs touched the glass.
The mandibles took no part in the movement.

Experiment 2. — The ant-lion was placed on the glass plate and
held, in a horizontal position, above my head; so that I could look
up. at it with a magnifying glass. The results were the same as in
experiment I ; but it was easier to observe that the tips of all
the legs touched the glass. The third pair of legs was the only
pair making vigorous movements.

Experiment j. — / tilted the glass plate so that the posterior portion
of the ant-lion was uphill. When the angle became steep the ant-
lion fell.

Experiment 4. — Repeated number i, substituting a pasteboard
rectangle for the glass plate. The result was the same as in ex-
periment i ; but the insect moved faster.

Experiment 5. — / tilted the paste-board rectangle so as to have the
posterior portion of the insect up-hill. Even when it had reached
an angle of 90 degrees, the insect retained its hold. It moved
upward, sidewise and downward.

Experiment 6. — While the cardboard rectangle was inclined at a
steep angle and the ant-lion was resting head downward, with a
dissecting needle, I raised the tip of the abdomen from the support.
The ant-lion retained its hold.

These experiments show conclusively that the mandibles do
not assist in locomotion ; at the same time, they indicate that the
hind pair of legs play an important role. Yet, so far as these
experiments go, the hind legs might be mere grappling hooks to
prevent the creature from slipping forward and the real locomotion
be due entirely to the flexing and stretching of the abdomen, all
forward motion being prevented by the stiff bristles on the sides
of the body and the grip of the legs.

Experiment 7. — A layer of dirt eqiial to the height of the greatest
height of the ant-lion was spread on a glass plate. The ant-lion
was placed on this pile of dirt. The larva began to burrow back-
ward into the dirt; but made practically no progress. By the

BEHAVIOR OF THE ANT-LION.

293

behavior of the body I could see that it was making vigorous
movements with its third pair of legs; but it made practically no
progress.

Experiment 8. — A layer of dirt equal in elevation to the greatest
height of the ant-lion was placed on a pasteboard rectangle. An
ant-lion was placed on this pile of dirt. Immediately it began
to burrow backward and continued to progress at a rapid rate.

(In all of the experiments from 1-8 the same individual was
used. The series was repeated many times with different in-

FIG. 10. A cluster of ant-lion pits (average cluster).

dividuals; but, in each case, the same individual was put through
the eight experiments.)

In experiments 7 and 8 the presence of the dirt makes it neces-
sary for more work to be performed in making progress backward.
Since the height of the pile is the same in each case, the amount of
work required is the same. Since the bristles are more numerous
on the sides of the body than on the ventral surface, the presence
of the dirt should give an added opportunity for them to function
in preventing forward slipping of the body; hence, if the progress
is due simply to a flexing and stretching of the body the ant-lion
should be able to move just as fast, if not faster, on a glass plate
with a layer of soil as on the naked glass. If, however, the hind
legs play an important role in dragging the body backward, then
the larva on the dirt-covered pasteboard should have a great
advantage over the one on the dirt-covered glass plate. These

294

C. H. TURNER.

experiments prove, I think, that the hind legs assist in dragging
the body backward. A microscopic examination of the legs
reveals two terminal claws which function in this work (Fig. 12).

EMERGENCE OF THE IMAGO.

This section does not pretend to be a life history of the ant-lion.
That the author hopes to make the subject of a future paper.
This is simply an attempt to state some interesting facts about
the last stages of the metamorphosis.

At the close of its larval period, the ant-lion constructs a
subterranean spherical cocoon of silk and soil. In my insectary,
most of the cocoons have been formed quite near the surface;
sometimes projecting slightly above the soil. In one case,
however, I found a cocoon on the bottom of the jelly glass,

FIG. ii. A cluster of ant-lion pits (small cluster).

fully two inches below the surface. In my insectary, the cocoons
have been formed in July and in August and the imagoes have
emerged in from 9-20 days thereafter; but I do not consider that
I have sufficient data to warrant the statement that they are
always formed at those times.

Inside of this cocoon the insect sheds its last larval skin and
becomes a chrysalis. At the end of a certain period of time the
anterior portion of the thorax protrudes from the cocoon (Fig. 7).

All of the accounts that I have read state that the chrysalis
comes about half way out of the cocoon and from its dorsal

BEHAVIOR OF THE ANT-LION. 295

surface the imago emerges. In my limited experience I have
noticed three methods of emergence. In one case the chrysalis
protruded about half way out of the cocoon and the imago emerged
from its back. In another case the chrysalis had left the cocoon
entirely and protruded about half way out of the soil. In the
third case both the head end and the tail end of the chrysalis
remained within the cocoon and from its back the imago emerged.
I am inclined to think the third case abnormal, caused by the
head of the chrysalis becoming entangled in some strands of the
cocoon. Fig. 6 is a photograph of the cast skin of that chrysalis,
made just after I had removed it from its cocoon. It seems to
me that the other two cases may be explained as follows: when

Fig. 1 2

FIG. 12. One of the third pair of legs of an ant-lion larva.

the cocoon is near enough to the surface for the chrysalis to expose
the upper portion of its body without coming entirely out of the
cocoon it does so; when the cocoon is a little deeper then the
chrysalis leaves the cocoon entirely and continues upward until
the anterior portion of its body is above the surface.

When the cocoon is too far beneath the surface, the chrysalis
dies on its upward journey. Fig. 5 is the photograph of such a
chrysalis. It was found dead about half an inch below the sur-
face. Attached to the bottom of the jelly-glass — about an inch
below — the empty cocoon was found.

Soon after emerging the imago undergoes an enormous increase
in size. It soon becomes more than twice as large as the chrysalis
from which it came, and this without partaking of food. Fig. 9
illustrates this. The jelly glass containing the cocoon had been
tightly closed to prevent the possible escape of the imago when
it emerged. It emerged at an unexpected time and when dis-
covered it was dead. It had lost one antenna and its body was

296

C. H. TURNER.

slightly damaged. Under the conditions it could not possibly
have obtained food. Half exposed above the soil was the shed
chrysalis skin, and a short distance below the surface the empty
cocoon (Fig. 4) was found.

MISCELLANEOUS ACTIVITIES.

Experiment I. — A ring oj water eight inches in diameter was made
on a glass plate and an ant-lion placed in the center of the dry space
that the ring surrounded. The ring was one half of an inch in width.
When the ant-lion reached the ring of water, it would usually
turn and move away from it. Often, in turning, its mandibles
would get into the water. In that case the mandibles would

A

11 I i i i i I i i i i I

i i

FIG. 13. The parts that form the sucking tube of the ant-lion larva,
mandible. B, the part that fits into the groove of the mandible.

A,

leave a broad water band wherever the creature went. After
its mandibles had become wet, on its next approach to the
water, it was apt to get some other part of its body wet. After
that it was apt to move, away from the light, on through the
water.

Experiment 2. — To see if there was anything about dirt as such
that would direct an ant-lion to it, a pile of dirt three inches in di-
ameter was placed in the center of a glass plate that was twelve inches

BEHAVIOR OF THE ANT-LION 2Q7

square. The ant-lion was placed at various places near the peri-
phery of the plate. Unless the ant-lion was placed in such a
position that in going directly away from the light it would en-
counter the dirt, never once did an ant-lion discover it. Some-
times the larva passed within less than a centimeter of the dirt
without being attracted by it.

LETISIMULATION.

Letisimulation (from letum, death, and simulare, to feign) is a
term coined by Weir1 in 1899, to designate the death-like attitude
assumed by individuals of many different groups of the animal
kingdom, when roughly handled. While citing examples from
among the worms, insects, reptiles, birds, and certain mammals,
he leaves the impression that the most remarkable examples of
death-feigning are to be found among the reptiles and certain
mammals. Since that time much careful attention has been
given to the letisimulation of insects. Barret (22) has studied
it in the mole-cricket; Gee and Lathrop (26), and Johnson and
Girault (32), in the plum curculio; Girault (27), in trox; Holmes
(30), in the water scorpions; Newell (34) and Weiss (40), in the
rice weevil; Riley (36), in dragon-fly nymphs; the Severins (37),
in the giant water bugs, and Wodsedalek (41, 42), in May-fly
nymphs and in a dermestid larva. In the light of the remarkable
traits revealed by these investigators, were he writing his article
today, Weir, no doubt, would agree with Homes that "it is among
the insects that the death-feigning instinct reaches its highest
development, occurring in a greater or less extent, in most of the
orders. It is especially common in beetles and not unusual among
bugs, but it is quite rare in the highest orders such as the Diptera,
or flies, and the Hymenoptera, or ants, bees and their allies. It
occurs in a few cases among the butterflies and moths, both in
the imago as well as the larval state. The instinct is exhibited
in different species in all stages of development from a momentary
feint to the condition of intense vigor lasting for over an hour.
Some insects may be severely mutilated, or, according to De
Geer, even roasted over a fire before they cease feigning."

Although the activities of ant-lions have interested many

1 Weir, "Dawn of Reason," pp. 202-214.

298 C. H. TURNER.

naturalists, very little attention has been paid to their death-
feigning behavior. Emerton (6) asserts that rough handling
caused his specimens to remain inactive for a time, and Mac-
Lachlan (n) states that the form observed by him letisimulates.
Each of these students dismissed the matter with a single
sentence.

The results recorded in this article are based on a careful
laboratory study of 100 individuals selected at random; supple-
mented by observations made in the field. About 60 per cent,
of these came from Kansas; the remainder from the vicinity of
St. Louis, Mo. Some were quite small and others were almost
large enough to form their cocoons. They were isolated in
numbered jelly-glasses, partly filled with loamy loess, and were
kept in an out-of-doors insectary, the north wall of which was
constructed of wire netting. The other walls were opaque.

Any stimulus which produces a shock will usually cause an
ant-lion to letisimulate. Rough handling, roughly turning it
upon its back, dropping it from a slight elevation, all have a
similar effect. I usually induced the feint by roughly turning
the larva upon its back, or by dropping it from a slight elevation.
Occasionally I found an individual that I could not induce to
letisimulate; but this was a very rare occurrence.

Several investigators have thought it important to determine
if the poses of letisimulating individuals are death attitudes.
Based on a consideration of seventeen species of invertebrates,
Holmes (31) concludes that the poses assumed were usually quite
different from death attitudes; although there were some species
in which they were always identical. I find that the ant-lion
has not one, but several death attitudes; likewise it possesses a
number of death-feigning postures, some of which resemble a
death pose and some of which do not. The insect suddenly
becomes rigidly immobile in whatever attitude it may be when
it receives the shock. Absolute immobility is the character that
is common to all cases; when the feint follows a long period of
fasting, this inactivity often simulates death. The rigidity,
however, is not so great as that described for certain insects.
In some species of insects the rigidity of the parts during a death
feint is so great that the insect may be picked up by a tarsus and

BEHAVIOR OF THE ANT-LION. 2Q9

held out at right-angles without the leg bending in the least.
That is not the case with the ant-lion. When an attempt is
made to lift it, in that manner, by a tarsus, the leg bends and the
insect awakes from its feint.

When an ant-lion is recovering from a feint, usually, although
not always, there is a preliminary waving of the antennae and a
twitching of the legs and, sometimes, a movement of the head.
Then the larva suddenly turns over. Throughout the whole
series of experiments, a careful record was kept as to whether the
insect turned towards the right or towards the left; towards the
light or away from it. It was found that whether the insect
turned toward the right or toward the left depended upon the
location of the strongest light; for the ant-lion invariably turns
away from the light.

Pinching the legs of a letisimulating individual almost always
caused it to come out of its feint. Blowing upon a death-feigning
larva would sometimes bring no response; at others it would
induce a twitching of the legs; at yet others it would cause a
complete recovery. Since the pinching of a leg, and even attempts
to lift the letisimulating ant-lion by a leg, usually terminates
the feint, I was surprised at the results produced by the following
mutilations.

Experiment I. — With a pair of small, but sharp, dissecting scissors,
I cut off the tip of a fore-leg of a letisimulating ant-lion. The insect
did not respond.

Experiment 2. — With a pair of small, but sharp, dissecting scis-
sors, I clipped off the tip of a mandible of a letisimulating ant-lion.
The insect did not respond.

Experiment j. — With a pair of small, but sharp, dissecting
scissors, in rapid succession, I cut off the tips of both fore-legs and
of a mandible of a letisimulating ant-lion. The insect did not
respond.

Experiment 4. — With a pair of small, but sharp, scissors I
severed the head from the body of a letisimulating ant-lion. The
insect did not respond, nor did it recover from the operation.

Experiment 5. — A pair of small dissecting scissors, identical with
those with which the above experiments were performed, was heated
red hot, in a Bunsen flame, and allowed to cool. This softened the

300 C. H. TURNER.

steel and made the edges quite dull. With these scissors an attempt
was made to remove the tip of a leg of a letisimulating ant-lion. The
scissors were too dull to cut through the chitin ; instead of being
severed, the leg became wedged in between the blades. The
feint was terminated immediately.

With the exception of experiment number four, these ampu-
tations were performed several times. On one occasion, a larva
recovered the moment I cut a leg; on another day, the same thing
happened when I severed a mandible. With these two excep-
tions, the results were always as stated above. How shall we
harmonize the fact that the pinching of a leg or, sometimes, the
blowing of the breath on the larva terminates the feint, while the
severing of a leg, or a mandible, or both invokes no response?
Shall we decide that such a pinch produces a greater physiological
shock than a sudden cut with a pair of sharp scissors? Is it
possible that a breath of air produces a greater shock than an
amputation with sharp instruments?

Relative Duration of Successive Death Feints. — In his study of the
beetle Scarites gigas Fabre (24) found that the duration of the
first five successive feints gradually increased from the first to
the last. The Severins (37), in their study of the giant water
bugs Belostoma and Nepa, and Gee and Lathrop (26), in their
study of the plum curculio (Conotrachelus nemuphar), find great
irregularity in the lengths of the successive feints.

To test the matter, an ant-lion was removed from its pit,
placed on a board, and made to letisimulate by roughly turning
it on its back. As soon as it recovered from one feint, it was
roughly turned on its back and induced to letisimulate again.
This was repeated until it had had an opportunity to letisimulate
twenty times. By means of a stop-watch, the duration of each
feint was obtained. One hundred individuals were thus experi-
mented with and the results recorded in a table. Criticially
examined, the table revealed a number of interesting things,
(i) There are marked individual variations. (2) In twenty
opportunities the individual usually letisimulates less than twenty
times. (3) The total time consumed in twenty opportunities
to letisimulate varied from one minute to two hours and twenty-
three minutes. The average for the 100 individuals was nineteen

BEHAVIOR OF THE ANT-LION.

301

and six-tenth minutes. (4) This death-feigning cannot be in-
definitely prolonged. (5) The duration of the feints near the
end of a long series of trials is always shorter than that of the
earlier ones. (6) A curve representing the relative lengths of a
series of letisimulations always contains two or more crests.
(7) The longest feints usually occur somewhere near the beginning
of the series. Of the 100 cases recorded, 33 letisimulated longest
on the first trial, ti on the second, 15 on the third, 2 on the
fourth, 5 on the fifth, 7 on the sixth, 3 on the seventh, 5 on the
eighth, 6 on the ninth, 5 on the tenth, I on the eleventh, 4 on the
twelfth, 2 on the thirteenth, and I on the sixteenth. We have
here, in a pronounced manner, the irregularity noticed by the
Severins and Gee and Lathrop in the forms studied by them.

Effects of Temperature upon the Duration of Letisimulations. —
To test this matter the TOO individuals mentioned were grouped
according to the temperatures at which the experiments had been
performed, and the tesults recorded in six tables. From the
averages of those tables the following table was compiled.

TABLE SHOWING THE AVERAGES OF THE EFFECT OF TEMPERATURE ON THE

LETISIMULATIONS.

Temperature in
F. Degrees.

Number of
Individuals
Used.

Number of
Letisimulations
in Twenty
Trials.

Maximum Time
in Min. of a
Single Letisimu-
lation.

Number of the
Trial on which
Max. Let. was
Made.

Total Time
Consumed in
Letisimulations.

60-65

7

IS

3-35

4

14.19

65-70

12

16

6.86

2

26.11

70-75

I?

17

7-25

5

35-n

75-80

23

13

3-97

5

13.06

80-90

17

12

5-91

6

13-30

85-90

21

9

3-83

6

12.15

If we were to rely upon these averages, we would conclude that
up to 75° F. both the length of the maximum feint and the total
duration of twenty feints vary directly with the temperature;
and that beyond that point there is no definite relation between
temperature and the feints. This conclusion, however, is not
supported by a critical study of the individual records from which
the averages were compiled. To test the matter further, four
individuals were selected and each put through five series of
twenty letisimulations, each series being conducted at a different
temperature. The results were recorded in four tables. There

302 C. H. TURNER.

was no obvious relation between temperature and the duration
of the feints.

Effects of the Strength of Stimulus upon Letisimulation. — To get
the stimulus as nearly uniform as possible, the ant-lion was
gently shoved from a glass ledge and caused to fall three inches.
To secure a strong stimulus the ant-lion was permitted to fall
upon a glass plate; to secure a weak one it was allowed to drop
on a layer of cotton batting. The results of experimenting with
100 ant-lions was tabulated. In 36 cases the first letisimulation
following a strong stimulus was the longer and in 58 cases the
first feint following a weak stimulus was the longer. In six
cases the duration of the feint was the same for each stimulus.
The average of 100 individuals gave the duration of the first
letisimulation following a weak stimulus as of longer duration
than the first following a strong stimulus. These data do not
seem to warrant a conclusion.

Effects of Hunger upon Letisimulation. — Certain selected in-
dividuals were well fed and others were forced to fast for a long
time before they were used for experiments identical with those
mentioned above. The results were carefully tabulated. No
relation could be detected between hunger and the length of the
letisimulation.

Apparently the reason for the longest letisimulation being
located sometimes at one place and sometimes at another in the
series is due to some internal (physiological) factor not revealed
by these experiments.

Weir1 considers the letisimulation of animals "one of the
greatest evidences of intellectual action, on their part." Hamilton
(29), Webster (39) and a few others feel that the creatures con-
sciously fear death and take this means to avoid it. Dr. Lindsley,
in "Mind in Animals," thinks "this must require great command
in those that practice it." However, the majority of modern
students of the subject look upon it as merely a remarkable
instinct.

No one who is acquainted with how slowly the ant-lion recovers
from injuries could, for a moment, consider anything intellectual,
which induces it to passively submit to portions of its legs and of

1Weir, "Dawn of Reason," 1889, p. 202.

BEHAVIOR OF THE ANT-LION. 303

its mandibles being amputated. The tonic contraction of the
muscles and the diminished reflex irritability suggest hypnotic
phenomena and lead one to agree with Holmes (31) that "the
instinct of feigning death is doubtless connected with much of
what has been called hypnotism in the lower animals." It is
well known that most animals pause momentarily when con-
fronted with an unexpected or violent stimulus. To me the
letisimulation of the ant-lion appears to be such a pause pro-
longed and exaggerated. The more I ponder over the results of
my experiments upon the death-feigning of the ant-lion, the more
I feel inclined to exclaim with James: "It really is no feigning of
death at all and requires no self-command. It is simply terror
paralysis which has been so useful as to become hereditary."

CONCLUSIONS.

1 . The pits of the ant-lion are not confined to sand ; they may
be found in any kind of dry friable soil, that is protected from
the rain and from insect-eating creatures. They are usually in
clusters; but, occasionally, a solitary pit is found. On yet rarer
occasions, pits may be found that are not under a shelter.

2. The ant-lion of the Middle West has two methods of ex-
cavating its pits. Usually it furrows backward, excavating a
series of concentric, adjacent circles, each deeper than the last,
and shovelling out the soil with its head. The front of the body
is so curved as to make it easy for the dirt to fall forward on the
head. In the second method, the larva simply burrows downward
into the ground and tosses out the soil with its head until the
sides of the pit become approximately stable. Pits formed by
the second method are usually subsequently enlarged.

3. The ant-lion removes a medium sized obstacle from its pit
by inserting the tip of its abdomen under it and, with the burden
poised on its abdomen, backing slowly up the slope.

4. After its trap has been completed, the ant-lion rests quietly,
in practically a horizontal position, with its body beneath the
soil and its open mandibles in the bottom of the pit.

5. Any invertebrate, be it insect, arachnid, or crustacean, that
happens to fall into the trap is acceptable food. Some escape,
but the larva attempts to capture all. When captured, the

3O4 C. II. TURNER.

victim is dragged partly or wholly beneath the soil, and the
juices imbibed through the hollow mandibles. Later the dried
carcass is tossed away.

6. The ant-lion may be considered positively geotactic, posi-
tively thigmo tactic and negatively photo tactic; with the res-
ervation that all of its movements cannot be explained as
tropisms in the Loebian sense.

7. It is impossible for the ant-lion to move forward; but, in its
backward movements, it can move in straight lines or curves,
and can scale vertical surfaces that are not too smooth. The
hind legs assist in producing this backward movement, and the
other legs brace the body.

8. Sometimes it avoids water and at others it backs into it.

9. If the spherical cocoon of this insect is near the surface of
the ground, the chrysalis comes only part of the way out and the
imago emerges from its back. If the cocoon is at a slightly
greater depth, the chrysalis comes entirely out of the cocoon and
part of the way out of the ground. If the cocoon is at a greater
depth, the chrysalis emerges entirely from the cocoon and
perishes on the way to the surface of the ground.

10. Rough handling or dropping from a slight elevation will
usually cause an ant-lion to letisimulate. The length of a feint
and the position of the longest feint in a series of successive feints
varies in different individuals and in the same individual at
different times.

1 1 . There is no obvious relation between the temperature, the
strength of the stimulus, or fasting and the duration of a letisimu-
lation.

12. If the relative durations of the successive feints of a long
series of letisimulations are plotted, the curve will have two or
more crests.

13. In the ant-lion all letisimulation poses are not death at-
titudes. The ant-lion has no characteristic death-feigning
posture. It is to be grouped with those insects in which the
letisumation pose varies with the attitude of the individual
at the time when the stimulating shock is received.

14. Although pinching a leg and, sometimes, even blowing on
the body, will usually cause a letisimulating ant-lion to come out

BEHAVIOR OF THE ANT-LION. 305

of its feint, in the majority of the cases, it will submit to the
clipping off of the tips of its legs and of its mandibles without
responding in any visible manner.

15. In the ant-lion letisimulation seems to be but an exagger-
ated prolongation of the pause made by most animals when they
are startled. The total behavior of a death-feigning ant-lion
supports Holmes's contention that "the instinct of feigning death
is connected with much of what is called hypnotism in the lower
animals"; and endorses James, when he says: "It is really no
feigning of death at all and requires no self-command. It is
simply terror paralysis which has become so useful as to become
hereditary."

REFERENCES
BEHAVIOR OF ANT-LioNS1
i. Berce.

'65 [Records his experience in rearing Myrmeleon formicariiis from the
larva.] Bull. Soc. Entom. Fr., p. xlvi.

2. Birge, E. A.

'73 The Ant-lion. Amer. Nat., Vol. VII., p. 432.

3. Brischke, C.

'79 Ueber das Eierlagen von Myrmeleon. Ent. Nachr., pp. 29-30.

4. Comes, Salvatore.

'09 Stereotropismo, Geotropismo, e Termotropismo nella Larva de
Myrmeleon formicarius L. Atti della Accademia Gioenia di Scienze
Naturali in Catania. Anno LXXXVI, memoria IV, pp. 1-14.

5. Comstock, Mrs. A. B.

'13 Handbook of Nature Study, pp. 395-397.

6. Emerton, J. H. The Ant Lion. Amer. Nat., 1871, Vol. IV., pp. 705-708.

7. Ferrari.

'65 Notes on the Habits of the Larva of Acanthaclisis. Wien Ent. Mon.,
Bd. VIII., p. 107.

8. Kirby and Spence.

'56 Introduction to Entomology, 7th edition, 242-244.

9. Lydekker, R.

New Natural History, Vol. VI., pp. 165-166.

10. McCook, H. C.

'07 Nature's Craftsmen, pp. 129-139.

11. MacLachlan.

'65 Habits of Myrmeleon formicarius L. Ent. Mag., Vol. II, pp. 73-75.

12. MacLachlan.

'66 [Note on eggs.] Bull. Soc. Entom. Fr., xvi.

13. Moffat, J. A.

'84 Notes on Ant Lions. Canad. Entom., Vol. XVI., pp. 121-122.
1 Practically every text-book on entomology and many popular books on insects
treat of ant-lions. With the exception of those to which special mention is made,
they are omitted from this list.

3O6 C. H. TURNER.

14. Moody, H. L.

'73 Habits of Myrmeleon immaculatus. Canad. Entom., Vol. V., pp.

63-65-

15. Navas, P.

'oo Algunas Costumbres de las Hormigas y Hormigaleoiies. Act. Soc.
Espan., pp. 219-222.

16. Navas, L.

'13 Biologische Beobachtungen. Ent. Mitt. Berlin, Bd. II., s. 81-87.

17. Navas, P. J.

'12 Notes Sobre Mirmeleonidos. Broteria, Vol. X., pp. 29-75.

18. Redtenbacher, J.

'84 Uebersicht der Myrmeleoniden Larven. Denk. Ak. Wien, Bd.
XLVIII., s. 335-368, pis. I-VII. [Gives a bibliography up to 1883.]

19. Romanes.

'83 Animal Intelligence, pp. 235-236.

20. Rudow, F.

'78 Myrmeleon formicalynx Pupating without Forming a Cocoon. Ent.
Nachr., Bd. IV., s. 272.

21. Step, Edward. Marvels of Insect Life, pp. 108-113.

LETISIMULATION OF INSECTS.

22. Barret, O. W.

'02 The Changa, or Mole Cricket (Scaptericus didactylus Latrl.) in Porto
Rico. Porto Rico Agri. Exp. Sta. Bull.

23. DeGeer, C.

'74 Memoirs pour Servir a 1'Histoire des Insectes, Vol. IV., pp. 229.

24. Fabre, J. H.

Souvenirs Entomologiques, 7th series, pp. 14-27.

25. Fabre, J. H.

'98 La Simulation de la Mort. Rev. Quest. Sci., Vol. XIV., pp. 111-139.

26. Gee and Lathrop.

'12 Death Feigning of Conotrachleus nemuphar Herbst. Ann. Ent. Soc.
of Amer., Vol. V., pp. 391-399.

27. Girault, A. A.

'13 Fragments of North American Insects. Entom. News, 1913, Vol.
XXIV., pp. 338-344.

28. Grote.

'88 Insects Feigning Death. Canad. Entom., Vol. XX., p. 120.

29. Hamilton.

'88 Insects Feigning Death. Canad. Entom., Vol. XX., p. 179.

30. Holmes, S. J.

'06 Death Feigning in Ranatra. Journ. Comp. Neur. and Psychol.,
Vol. XV., pp. 305-349-

31. Holmes, S. J.

'08 The Instinct of Feigning Death. Pop. Sci. Monthly, 1908, Vol.
LXXIL, pp. 179-185.

32. Johnson and Girault.

'06 The Plum Curculio. Circ. 73 Bu. of Entom. U. S. Dept. of Agri.

33. Kirby and Spence.

'56 An Introduction to Entomology, 7th edition, p. 412-413.

BEHAVIOR OF THE ANT-LION. 307

34. Newell, W.

'13 Notes on the Rice Weevil and Its Control. Jour. EC. Entom., Vol.
VI., pp. S5-6i.

35. Pieron, H.

'04 L' Immobilite Protectricechezles Animaux. Revue Sci., Vol. LXXIIL,

PP- 523-527-

36. Riley, C. F. Curtis.

'12 Observations on the Ecology of Dragon-Fly Nymphs. Ann. Entom.
Soc. of Amer., Vol. V., pp. 273-292.

37. Severin, H. H. P. & H. C.

'n Habits of Belostoma (Zaitha) flumineum Say and Nepa apiculata
Uhler, with Observations on Other Closely Related Aquatic Hemiptera.
Jour. N. Y. Entom. Soc., Vol. XIX., pp. 99-108.

38. Slater, C. R.

'79 Simulation of Death by Coleoptera and Arachnida. Sci. Cos., Vol.
XV., pp. 160-161.

39. Webster.

Insects Feigning Death. Canad. Entom., 1888, Vol. XX., p. 199.

40. Weiss, H. B.

'13 Notes on the Death Feint of Calandra orysae. Canad. Entom., pp. 135-

137-

41. Wodsedalek, J. E.

'12 Natural History and General Behavior of the Ephemeridae Nymphs
Heptagenia interpunctata Say. Ann. Entom. Soc. of Amer., Vol. V.,
pp. 31-40.

42. Wodsedalek, J. F.

'12 Life History of Habits of Trogoderma tarsale (Melsh). Ann. Entom.

Soc. of Amer., Vol. V., pp. 367-381.
August 10, 1915.

THE EFFECT OF CERTAIN ORGANIC'AND INORGANIC

SUBSTANCES UPON LIGHT PRODUCTION

BY LUMINOUS BACTERIA.

E. NEWTON HARVEY.
PRINCETON UNIVERSITY.

While engaged in a study of the chemistry of light production
by luminous bacteria I had occasion to investigate the effect of
diluting the sea water with distilled water and with isotonic
sugar solution and the influence of the various salts of sea water,
of acids and alkalies, and of certain anaesthetics upon the emission
of light. The results are of interest for comparison with the
known effects of these substances on other organisms and with
other vital manifestations of life.

In all experiments, except where otherwise noted, one drop of
the dense emulsion of luminous bacteria (a form isolated from
squid at Woods Hole, Mass.) was added to 30 c.c. of solution in
an uncorked Erlenmeyer flask and the whole thoroughly mixed.
For comparative observations it is essential that the eye be thor-
oughly adapted to the dark and that each flask be oxygenated
by shaking, before judging as to the emission or absence of light.
Observations were made after 10 minutes, one hour and 24 hours.

TABLE I.

EFFECT OF DILUTION OF SEA WATER WITH WATER AND WITH m CANE SUGAR

SOLUTION.

Dilution with Water.

Dilution with m Cane Sugar.

Parts
Sea
Water.

Parts
Water.

Light after

Parts
Sea
Water.

Parts
m
Sugar.

Light after

10 Min.

i Hr.

24 Hrs.

10 Min.

iHr.

24 Hrs.

2

I

+

+

+

2

I

+

+

+

I

I

faint

faint

I

I

-f

-\-

+

I
I
I
I

2

4
6

10

faint
faint
very faint

very faint

—

I

I
I

I

2

4
6

10

!

|

faint
faint
faint

I

14

—

—

—

I

14

+

+

—

I

20

—

—

—

I

20

+

+

—

Sea water

Water

±

1

1

m cane sugar

+

+

—

308

LIGHT PRODUCTION BY LUMINOUS BACTERIA.

309

It will be noted from the above Table I. that the bacteria cease
to give off light and experiment shows that they are killed by too
great dilution with water. That this effect is not entirely due to
the absence of salt but is chiefly due to a cytolysis through
lowered osmotic pressure is shown by diluting the sea water with
an inert iso tonic solution, cane sugar. Some salt is necessary
for the continued production of light as the bacteria no longer
glow after twenty four hours' emersion in w-sugar, a fact of no
great surprise as unicellular freshwater luminous animals are
unknown.

TABLE II.

EFFECT OF ACID AND ALKALI.

*
Cone, of Acid and Alkali Added to Mg-free

Light after

sea \\ater, ?«/2 (iooNaCl+2.2 K(_l + 2t_,ad2)
in Syracuse Watch Glasses.

10 Min.

i Hr.

24 Hrs.

M/2OOO HC1

w/4000 HC1

faint

w/Sooo HC1

+

faint

«/i6ooo HC1

+

-f

w/320oo HC1

+

+

faint

M/SOO valerianic acid

tt/IOOO " .

faint

M/2OOO "

faint

W/4OOO "

+

faint

M/800O "

+

+

M/I60OO "

+

+

faint

M/SOO NaOH

w/iooo NaOH

_

faint1

W/2OOO NaOH

+

+

+

«/25o methyl amine

n/soo " "

faint

_

faint1

M/IOOO " "

faint

+1

faint

M/2000 " "

+

+

+

Mg-free sea water

+

+

faint

Sea water. .

+

+

4-

1 Probably due to neutralization of alkali through absorption of COa.

As was to be expected acids and alkalies prevent light emission
in very weak concentration, the acids in much weaker concentra-
tion than the alkalies. In fact the bacteria are very sensitive to
acid and will not even phosphoresce with any brilliancy in a
neutral medium.

The organic acid (valerianic) and alkali (methyl amine) have
less effect than the inorganic, a result at variance with my results
for other organisms which are usually affected more readily by
the weak than by the strong acids and alkalies.1

1 Harvey, E. N., "Studies on Acids," in Carnegie Institution Publications
No. 212, p. 143, 1915; on alkalies, id., No. 183, p. 131, 1914.

3io

E. NEWTON HARVEY.

TABLE III.

EFFECT OF VARIOUS COMBINATIONS OF THE SALTS OF SEA WATER.

Light after

Salt Combinations.

10 Min.

i Hr.

24 Hrs.

Sea water

+

+

+

Artificial Sea Water, m/2 (100 NaCl+2.2 KC1 +
2 CaCh + io MgCl-) +M/4000 NaOH

+

+

+

Neutral artificial sea water

+

+

faint

m[2 NaCl .

+

+

faint

m/2 KC1

+

faint

m/3 CaCh

m/3 MgCU

_

m/2 (100 NaCl +2.2 KC1)

+

. +

faint

m/2 (100 NaCl+2 CaCU)

+

+

faint

m/2 (100 NaCl+io MgCh) . ...

+

+

faint

m/2 (100 NaCl +2. 2 KC1+2 CaCh)

+

r +

faint

m/2 (100 NaCl +2. 2 KCl+io MgCh)
m/2 Cioo NaCl +2 CaOU + io MeCU) .

+
+

' +
+

faint
faint

The most interesting point brought out in the above table is
the independence of these bacteria of a balanced medium. The
bacteria live and phosphoresce in pure NaCl without the addition
of any bivalent ions. This is true even when the solution is
changed three times to remove the last traces of Ca in the bacteria.
KC1 is also relatively non-toxic, although more so that NaCl.
CaCl2 and MgCl2 are very toxic when alone. All combinations
of NaCl with the other ions of sea water sustain the bacteria well
except that they are neutral media and hence the phosphorescence
is dimmed after 24 hours. That pure NaCl should have so little
effect on light production is astonishing when we consider its
poisonous effect on other marine organisms and tissues, partic-
ularly on ciliated cells.

The effect of the alcohols (Table IV.) on light production is very
similar to their effect on other life processes: they exert an inhibit-
ing or anaesthetic action which is perfectly reversible. If alcohol
solutions containing bacteria which have stopped emitting light
are diluted with sea water, light production again begins. As
with other tissues the higher the alcohol in the series the greater
anaesthetic power it has.

The effect of a number of other substances was studied in a
very rough way — namely, by adding a small quantity of the
substance to a sea water emulsion of the bacteria in test tubes
and then shaking the tubes. With toluol, benzol, ether, chloro-
form, carbon disulphide, carbon tetrachloride and ethyl butyrate

LIGHT PRODUCTION BY LUMINOUS BACTERIA.

TABLE IV.

EFFECT OF ALCOHOLS.

Light after

Cone, of Alcohol Added to Sea Water.

10 Min.

i Hr.

24 Hrs.

Methyl alcohol, 2m

HCH2OH, 1.5 m

4-

very faint

HCH'OH, m . . .

4-

faint

HCH2OH, m/2

+

4-

4.

HCH2OH, m/3

+

+

4-

Ethyl alcohol, m

CHsCH-OH, OT/i.5

very faint

CH3CH2OH, m/2

4.

faint

faint

CH3CH2OH, m/3. . . .

4-

+

4-

Propyl alcohol, wz/3

CH3CH2CH2OH, w/4

very faint

CH3CHoCH-OH, w/6

faint

very faint

CH3CH2CHiOH, m/S

4-

faint

CHsCHaCHoOH, m/i5

+

-4-

4.

Isosbutyl alcohol, m/io

(CH3)2CHCH->OH, w/12

very faint

(CH3),CHCH2OH, 7w/i6

+

very faint

(CH3)2CHCH2OH, m/20

+

+

(CH3)2CHCH2OH, w/24

+

4-

4-

Amyl alcohol, m/2o

C2H5CH3CHCH2OH, w/40.

very faint1

+1

C2H5CH3CHCH2OH, wz/8o

very faint

very faint

4-1

C2HSCH3CHCH2OH, m/i6o

faint

4-1

4-

C"H5CH3CHCH2OH, ?w/32O

+

4-

4-

Capryl alcohol, 7^/400

CH3(CH2)6CH2OH, TW/Soo

very faint

very faint

CH3(CH2)6CH2OH, ra/i6oo

faint

faint

CH3(CH2)6CH2OH, wz/3200. . . .

faint

faint

4-1

CHsCCH-OeCHzOH, 771/6400. .

+

4-

4-

Sea water

+

4-

4-

1 Probably due to evaporation of alcohol.

the light was found to disappear almost immediately; with
tannin, chloral hydrate, vanillin and sodium glycocholate the
light had disappeared in the course of one hour while saponine,
amygdalin, and sodium taurocholate had no effect. It is sur-
prising that saponin has no effect on luminous bacteria when we
consider its great cytolytic power 'on other forms in very small
concentration.

SUMMARY.

The effects on luminous bacteria of dilution of sea water with
water and m sugar solution; of HC1 and valerianic acid; of NaOH
and methyl amine; of the salts of sea water in different combina-
tions; and of methyl, ethyl, propyl, butyl, amyl and capryl
alcohol were studied. The points of interest in the results are
indicated after each table.

FURTHER NOTES ON THE CHROMOSOMES OF THE

CERCOPIM:.

ALICE M. BORING AND RAYMOND H. FOOLER.

The chromosomes in the spermatogenesis of five species of this
family of Hemiptera have already been studied by Stevens1
and Boring.2 Three more species have now been studied in
comparison with those previously studied. They are Philaenus
lineatus, Aphrophora paralJela and Clastoptera proteus. Each
of these three species belongs to a genus in which one or more
species has already been studied, so this gives a chance to compare
the spermatogenesis in closely related species. This has been
done very carefully by McClung3 for some families of Orthoptera.
The entire family Acrididae has the same spermatogonial chro-
mosome number, 23, and the Locustidae has 33, but within each
family there are generic and specific cytological differences.
The family Cercopidae of the Hemiptera does not show as closely
graded a series of cytological differences as the orthopteran families
studied by McClung. The facts as found are here recorded.

The material was collected at Woods Hole4 and in Orono;
Philanus lineatus from grasses, Aphrophora parallela from Scotch
pines, and Clastoptera proteus from alders. Dr. Herbert Osborn
has very kindly identified the species of the material. Flem-
ming's and Gilson's solutions were used for fixation, and iron
hasmatoxylin for staining.

Philcenus lineatus has 29 chromosomes as spermatogonial
number (Fig. i), two of which are larger than the others. The
odd chromosome is round or oval in the early (Fig. 2) as well as

1 N. M. Stevens, '06, "Studies in Spermatogenesis," Pt. II., Carnegie Institute,
Washington.

2 A. M. Boring, '07, "A Study of the Spermatogenesis in the Membracidae," etc.,
Jour. Exp. Zool., 4, p. 469. '13, "The Chromosomes of the Cercopidae," BIOL.
BULL., 24, p. 133.

3 C. E. McClung, '08, "Cytology and Taxonomy," Kans. Univ. Bull. 4.

4 We wish to thank the Director of the Marine Biological Laboratory for the
privileges of the laboratory during the summers of 1913 and 1915, at which time this

material was collected.

312

CHROMOSOMES OF THE CERCOPID.E.

313

the late (Fig. 3) spireme stages. The reduced number of chro-
mosomes is 15 in the first spermatocytes, one of which is larger
than the others (Fig. 4). The odd chromosome is univalent

6

8

9

10

11

12

14

15

16

17

- 5) and does not divide in the first spermatocyte division
(Fig. 6). The second spermatocytes have partly 15 (Fig. 7)

314 ALICE M. BORING AND RAYMOND H. FOGLER.

and partly 14 (Fig. 8) chromosomes. The chromosome number
is specific, as the reduced number is 15, while only 12 are found
in Phil&nus spumarius. But the roundness of the odd chro-
mosome throughout the spireme stages is a feature common to
both species of this genus, and distinguishing it from the species
of the genus Aphrophora.

Aphrophora parallela has 15 chromosomes as reduced number,
with one largest chromosome (Fig. n). The odd chromosome
is elongated in the early spireme stages (Fig. 9) and becomes more
nearly round in the later stages (Fig. 10). The odd chromosome
is, as usual, univalent (Fig. 12) and does not divide in the first
spermatocyte division (Fig. 13). The chromosome number in
the second spermatocytes is 14 and 15 (Figs. 14 and 15). Again
in this species, the chromosome number is different from that in
the other species of the same genus, that is, 15, in comparison
with 14 in Aphrophora quadrinotata and 12 in Aphrophora spu-
maria. The long odd chromosome in the early spireme stages
is a common feature of both A. spumarius and A. parallela, and
distinguishes them from the genera Philcemis and Clastoptera.
The early spireme stages of A. quadrinotata were not studied.
The species formerly classified as A . quadrangularis has since been
put into the genus Lepyronia. This species does not possess
the long odd chromosome characteristic of the genus Aphrophora.

Clastoptera proteus has 7 as reduced chromosome number (Fig.
1 6), one less than the reduced number in Clastoptera obtusa.
Unfortunately only a few stages were found in this material, so

TABLE I.

Genus. Species. Reduced Chromosome

Number.

Philaenus spumarius 12

" lineatus i S

Aphrophora spumaria 12

quadrinotata 14

parallela 15

Lepyronia quadrangularis 1 1

Clastoptera proteus 7

obtusa 8

that the only other significant point that was observed was that
the first spermatocyte division is the one in which the odd chro-
mosome does not divide (Fig. 17) as in all the species of this family.

CHROMOSOMES OF THE CERCOPID^E. 315

The eight species of Cercopida in which the spermatogenesis
has so far been studied belong to four genera. The chromosome
number (reduced) varies from 7 to 15. The chromosome number
seems to have no significance for family or genus. The specific
numbers are shown in Table I.

The odd chromosome in the spireme stages differs in its shape
in the genus Aphrophora from that in the other genera as far as
studied. It is a much elongated structure early in its appearance
in Aphrophora, while it first appears as an oval or round body in
the others.

All eight species of the Cercopidae studied show a typical odd
chromosome, which divides only in the second spermatocyte
division. In all of the species except Aphrophora quadrinotata and
Clastoptera proteus, in which the material was limited and the
equatorial plates consequently not studied in favorable positions,
there is one chromosome among the reduced number which is
distinctly larger than the others. In no case is the odd chromo-
some the largest one.

WOODS HOLE,

July 30, 1915-

THE REACTIONS OF AN ORB-WEAVING SPIDER,

EPEIRA SCLOPETARIA CLERCK, TO RHYTHMIC

VIBRATIONS OF ITS WEB.1

WILLIAM MORTON BARROWS.

WITH THREE PLATES.

I. Introduction 316

II. Materials and Methods 316

II. Experiments 3*9

1. Experiments Using Rhythmic Vibrations 319

2. Experiments Using a Y-shaped Vibrator 321

3. Response in the Dark 322

4. Distribution of Vibrations Through the Web 322

5. Mutilation Experiments 323

IV. Discussion and Summary 325

I. INTRODUCTION.

The work reported in this paper was suggested by a chance
observation2 made in the summer of 1911. A fly was held close
to one of the spiders without eliciting any response; when the
fly's vibrating wing was allowed to touch a strand of the web,
however, the response was instantaneous and positive. The spider
ran to the fly and seized it. A vibrating rubber band held against
a strand of the web caused a very similar response. During
the summer of 1913 these spiders were studied more carefully in
an attempt to determine: first whether the stimulus was vibratory
in nature or must be considered to be due to some other force and
second whether the response could be identified as a "tropism"

or taxis.

II. MATERIALS AND METHODS.

At the Lake Laboratory maintained by the Ohio State Uni-
versity at Cedar Point, Ohio, the large orb weaving spider,
Epeira sclopetaria, is very abundant, building its webs on the front
porch and in the angles of the building and roof. The habit

1 Contribution No. 42 from the Department of Zoology and Entomology, Ohio
State University.

2 The note of Boys (80) was not known to the writer until the larger part of
these experiments had been carried out.

REACTIONS OF ORB-WEAVING SPIDER TO VIBRATIONS OF WEB. 317

of the female of tfcis species of remaining at the center of her web
for long periods of time makes it a very convenient form to study
in its normal surroundings.

This species builds its web in dead branches or in the angles of
buildings where there is an abundance of small or medium sized
insects. The web usually consists of 17. 18 or 19 relatively
inelastic strands which radiate from a center like the spokes of a
wheel. These radiating strands are attached at their outer ends
to twigs or boards or to guys or stays which anchor several radii
to the support. Surrounding the center of the web is an irregular
network known as the hub and notched zone which serves as a
resting place for the inhabitant of the web. For a short space
outside the hub the radii are bare (the free zone) but beyond this
is found the viscid spiral consisting of finer strands which are
extremely elastic and are beaded with microscopic sticky drops
which serve to hold and entangle the insect prey. It is probably
the extreme elasticity of these spiral strands which allows them
to detain a strong insect without being snapped, thus giving the
spider time to reach the detained insect and complete its en-
snarement by the addition of fresh silk from the spinneretts.
The normal resting position of the female spider is with the head
directed downward and the legs spread outward on the notched
zone as is shown in Fig. i . The method used to obtain this pho-
tograph and the others following is that used by Comstock
('12, p. 181). A female spider was placed on a dead branch held
in the neck of a bottle which was set in a tray of water. During
the first or second night the spider usually built a perfect web.
The branch was then moved to the photographing table with as
little disturbance as possible and placed in front of a soap box
painted a dull black on the inside. Arranged in this way before
the camera it was possible to take pictures showing the spider,
web, and vibrator straw, and the various positions taken by the
spider in the act of responding to the vibrator.

The size of the web varies from two inches in diameter, or even
smaller when built by very young spiders, to eighteen inches or
more when built by mature females. The male builds a web
very much like that made by the female but as he has a roving
disposition one is never sure that the same individual can be

3l8 WILLIAM MORTON BARROWS.

located twenty-four hours later while the females often live for
weeks in the same place, repairing the web every evening but not
altering it materially.

In crawling across the web the spider always follows a radiating
strand or at the edge of the web, one of the guy strands, and
places its feet on the radii or on the junctions of the radii and
spiral threads where the latter hold no sticky materials. The
front feet are usually placed on the same radiating strand but
the second and third pairs may be spread out on the two ad-
joining strands. It is possible for the spider to crawl rather
swiftly along a single strand for a considerable distance, all eight
feet using the same thread. In crossing the web the spider usually
leaves behind a dragline which may remain across the web,
adhering to it after the spider has returned to the center Some
individuals on the other hand when they reach the edge of the
web swing free, held only by the drag line up which they climb
in returning to the center. Occasionally one finds both methods
employed by the same individual. Most spiders are not skilful
enough to cross the web several times without tearing out or
snagging several of the segments of the spiral thread. When the
web is violently disturbed the spider usually retreats to a niche
or corner (the retreat) and remains there motionless unless again
disturbed. Some individuals remain in the retreat instead of at
the center of the web. When this is done one forefoot is placed
on the trapeline leading to the hub and any activity of the web
such as that produced by an entangled insect sends the spider
like a flash down to the web. In this connection another fact
may be noted; a spider outside the center of the orb always
returns to the center, takes the normal position and then orients
before it finds an entangled insect. This might be explained as
due to the difficulty of crossing the web by any other path than
by the radii. However, the inability to orient accurately in any
other position than the center gives a clue to a more probable
explanation. Individuals of the species Epiera sclopetaria will
eat nearly any insects which happen to become entangled in the
web. The food of those studied inside the screened porch con-
sisted almost entirely of rather large flies of the genera Musca,
Sarcophaga, and Lucilia. It is in the snaring of these flies that

REACTIONS OF ORB-WEAVING SPIDER TO VIBRATIONS OF WEB. 319

this Epeira seems to be. especially expert. When a fly strikes a
web it often goes through, breaking out one or two spiral segments.
If, however, it does not break through it hangs for a second, buzz-
ing, then breaks one or two of the sticky strands and flies away.
A fly seldom entangles itself to such an extent that it cannot get

•

free inside of five seconds. A successful spider then must reach
the fly in less than two or three seconds after it strikes the web.
The actual capture of the fly is accomplished usually either by
biting the fly and stunning it or by winding it with web. The
entangled fly may be left where it struck or may be torn from the
web, and carried attached to one of the spider's hind feet to the
center of the web where it is thoroughly chewed and its liquid
parts swallowed.

The apparatus used to produce rhythmic vibrations consisted
of three tuning forks and an electric vibrator. One fork had
a vibration rate of 100 double vibrations per second, another a
rate of about 487 and the third was an adjustable fork with a
large range of vibration rates but with very limited amplitude.
The electric vibrator was a modified electric door bell in which the
clapper was replaced by a long grass straw. The number of
vibrations produced by this instrument could be varied to some
extent by changing the tension of a spring and a regulator screw,
while the amplitude of the vibration varied with the length of
straw used. The vibration rate of the vibrator was obtained
by comparing a tracing made by it on a sheet of blackened
paper on a revolving drum with a simultaneous record made by
the tuning fork giving 100 double vibrations per second. The
electric vibrator was found to be more effective than a fork because
it gave vibrations of equal intensity, i, e., it did not run down.
It had also another advantage in that it could be controlled by a
switch held in the hand and could be operated at a distance from
the operator. A stop watch was used to measure the time elaps-
ing between the beginning of the stimulus and the arrival of the
spider at the place where the straw touched the web.

III. EXPERIMENTS.

i. Experiments Using Rhythmic Vibrations.
When the vibrator straw is placed against one of the spiral
strands or against one of the radii and caused to vibrate the spider

32O WILLIAM MORTON BARROWS.

orients instantly and advances along the nearest radius to the
straw, seizes the straw with its mandibles and may spread web
on the straw with the hind pair of feet (Fig. 3). This reaction
is carried out in essentially this manner no matter where the
straw may strike the web.

The orientation is so rapidly executed and is followed so closely
by the forward locomotion that it is difficult to separate the two
parts of the response. If, however, the vibrator is set in motion
for a fraction of a second only the orienting is accomplished but
the forward locomotion toward the vibrator does not follow.
A second vibration while the spider is oriented calls forth the
forward response and an attack on the vibrator (Fig. 3). The
photograph reproduced in Fig. 2 shows such an orientation. If
the first vibratory stimulus is not too long or is not followed by a
second stimulus the spider usually returns to the resting position
at the end of a few seconds. Some individuals, however, follow
the orienting response by an interesting series of activities.
The fore feet are placed on neighboring radii, drawn toward the
animal's body and released suddenly. This release sets the web
vibrating parallel to the spider's longitudinal axis. The spider
then turns one space to the right or left and repeats the process
until she has oriented through a complete circle and set every
pair of radii in motion. The use of this activity is seen if there
happens to be a captured fly or a piece of dirt in the web. When
the two radii which pass on either side of the object are set vibrat-
ing the object is also set in motion but its motion is not of the same
rate as that of the rest of the web and it sets up an echo or return
vibration. To this the spider responds. A dead fly may be
rediscovered in this way or a piece of dirt may be located and
removed.

Responses to different frequencies show considerable variations
and it is not possible to predict that a certain individual will
respond in a definite way to a given stimulus. This variation in
response ranges from instantaneous orientation and forward
locomotion to a slow orientation and slow approach toward
the vibrating point or it may happen that no sign will be given
that the stimulus has been perceived. Roughly speaking a large
spider responds most quickly to a vibration of considerable am-

REACTIONS OF ORB-WEAVING SPIDER TO VIBRATIONS OF WEB. 321

plitude with a vibration rate of 24 to 300 per second. It was
impossible with the materials at hand to construct a vibrator
giving a high rate and having also a considerable amplitude, so
recourse to steel wires and small forks was necessary. The large
spiders did not respond well to wires and forks with high vibra-
tion rate and small amplitude but they did respond instantly
to the vibrating wings of Chrysops (127 per sec.), Microbembex
(208 per sec.), Musca (284 per sec.), where the amplitude ranged
from 4 mm. to 10 mm. Small spiders responded quickly to
vibrations ranging from 100 per sec. to 487 per sec. and even
higher although the amplitude was very small. This difference
in responsiveness between the young and old spiders is probably
correlated with differences in size and rate of wing vibration of
the insects which are ensnared and used as food by young and
old. In general small insects have high wing vibration rates
while the larger insects have lower rates of wing vibration
(Packard, '03. p. 150). The smaller spiders eat small insects
and the large spiders eat larger insects. The following species
of insects were caught and eaten by E. sclopetaria: Chrysops
vitatus (127 vibr. per sec.); Calliphora vomitaria (130 vibr. per
sec.) ; Microbembex monodonta (208 vibr. per sec.) ; Musca domes-
tica (284 vibr. per sec.). Many small midges (Chironomus and
others) were eaten by the young spiders and occasionally by the
adults. The vibration rate of these small midges is probably
very high, judged by the high pitched note which they give out,
but it was impossible at the time to determine its rate.

2. Experiments Using a Y-shaped Vibrator.

In order to determine whether the spider reacted to a single
vibrating strand or to the center of a vibrating area of the web,
a Y-shaped vibrator made up of insulated magnet wire was
adjusted to the vibrator and arranged in such a manner that its
ends touched the web at two places, 2 or 3 cm. apart. When the
vibrator so adjusted was operated the spider responded readily,
going to a point on the edge of the web midway between the two
vibrating points and then after some slight hesitation going toward
one or the other of the vibrator wires (see Fig. 4). If, however,
these points of wire were more than 3 cm. apart the spider at the

322 WILLIAM MORTON BARROWS.

center of the web usually hesitated, turning first toward one, then
toward the other, finally orienting to one and attacking this by

itself.

3. Response in the Dark.

In order to test the ability to respond in the dark the vibrator
was set up late in the afternoon, the straw touching one of the
radial strands of a web which was built in the frame of a window.
The window was shaded on the outside by a heavy thicket. At
9:30 P.M. the room was so dark that a person standing inside
could discern the outline of the window with the utmost difficulty.
A flash of light from a pocket electric lamp showed that the female
occupying the web was at the center of her web. The vibrator
switch was closed and at the end of about four seconds the electric
flash light showed the spider biting the vibrator straw in the
same manner as that shown in Fig. 3. This experiment indicates
that unless these spiders use rays of light which our eyes do not
perceive, sight plays no essential part in the orientation to and
the ensnaring of the prey.

4. The Distribution of Vibrations through the Web.
The distribution of vibrations as they travel across the web is
of some theoretical interest. The following method for recording
these vibrations was used with considerable success. A spider
in its web was placed before the camera and made to respond to
the vibrator repeatedly until it would respond no more. A
photograph (Fig. 5) of 15 seconds' exposure was then made while
the vibrator was in motion. The web was somewhat torn by
the spider before it ceased to respond, but the photograph reveals
by the thickening of the lines the distribution and amplitude of
the vibrations in all parts of the web. The amplitude of the
vibrations decreases rapidly from the periphery toward the
center. The radial strand connected with the vibrator shows
the greatest lateral displacement while the strands on either side
of this show less and less disturbance as the distance away from
the vibrator increases. A slight thickening of the spiral strands
in a direction at right angles to the direction of the primary vi-
bration can be noted on the segments directly across the center
from the vibrator. The center of the web seems to be the part

REACTIONS OF ORB-WEAVING SPIDER TO VIBRATIONS OF WEB. 323

least affected. If there is any motion here it is probably at right
angles to the original vibration, that is, it is probably parallel
to the spiders' long axis after orientation.

5. Mutilation Experiments.

The foregoing experiments coupled with careful observations
on the spinning behavior of the orb-weaver lead to the conviction
that the organs used in detecting the movements of the web
are proably tactile, at least there are no other organs described
which would seem to serve the purpose as well. There can be
little doubt that sense hairs are very abundant on the legs, par-
ticularly on the tarsi of these spiders. These hairs have been
described by Dahl (83), Wagner (88), McCook (90), and recently
by Mclndoo (n). The functions of these hairs have been inter-
preted in various ways, but little or no experimental work has
been accomplished other than attempts to show that some spiders
hear. Responses to sounds seem to have been observed only in
those forms which build webs. It seems likely that responses
in the web building forms are due to the vibrations of the air
being picked up by the strands of the web (Mclndoo, 'n, p. 412).
It was thought desirable to determine if possible the location of
the sense-organs used in detecting vibrations. By careful manip-
ulation with a pair of fine dissecting scissors it was possible to
snip off one or more of a spider's legs without causing the spider
to leave the web. It is necessary to use great care not to shake
the web because an irregular shaking gives rise to the negative
response, the spider running away to the retreat. The contrast
between this insensibility to the amputation of legs and extreme
sensitiveness to irregular vibrations of the web emphasizes the
fact that these spiders receive most if not all of their mechanical
stimuli through the web. These operations caused the spider to
lose considerable blood but two or three hours usually sufficed
to heal the wound. The stumps of the legs were always held up
so that they did not touch the web.

Experiment i. — After testing a spider to be assured that its
responses were normal the two forelegs were cut off as near the
middle of the metatarsus as possible. This spider immediately
put the stumps of the forelegs into its mouth. The next morning

324 WILLIAM MORTON BARROWS.

this spider was in its web. During the night the web had been
repaired and a new spiral thread put on.

In recording the test made on this spider and those following,
IX o'clock, XII o'clock, etc., refers to the position at the edge of
the web which corresponds to the same hour on the clock face.
Thus VI o'clock is used to designate the edge of the web which
the spider normally faces when at rest, i. e., directly downward.

Experiment i. — Spider with both forefeet cut off. Fork giving
100 vibrations per second touching web in
III o'clock position spider reached fork 8 inches from center in

3 seconds.

IX o'clock position reached fork (8 in.) in 2^2 sec.
VII o'clock position reached fork (8 in.) in 2^4 sec.

Experiment 2. — Spider with third legs cut beween femur and
patella.

IX o'clock position reacted 8 in. in 2% sec.
Ill o'clock position reacted 8 in. in 2>£ sec.
XII o'clock position reacted 8 in. in 7^4 sec.

This individual showed some difficulty in climbing, but
oriented accurately.

Experiment 3. — Spider with second legs cut off at patella.
Ill o'clock position reacted 7 in. in i^ sec.
XII o'clock position reacted 7 in. in i^ sec.

X o'clock position reacted 7 in. in i>£ sec.

Experiment 4. — Spider with fourth legs cut off at patella.
Reactions entirely normal as given above.

Another set of experiments which need not be detailed were
carried out. In these the right first leg and left fourth leg were
cut off and other similar combinations were made. In all cases
orientations and the locomotion following were entirely normal
except for the slight difficulties in locomotion which might be
expected. These experiments indicate that the sense organs used
in reacting to the vibratory stimuli are not restricted to any one
pair of legs below the metatarsus. There are two possible dis-
tributions of sense hairs which would seem to make possible the
reactions detailed above ; the sense organs may be confined to the
feet, where they come in contact with the web or they may be
located on the le.gs or body in such a manner that they pick up

REACTIONS OF ORB-WEAVING SPIDER TO VIBRATIONS OF WEB. 325

the vibration of the whole leg or whole body. Hinged sensitive
hairs uniformly scattered over the body might answer this pur-
pose. It seems most likely, everything considered, that the
particular sense organs used are on the tarsi of each leg and come
in contact with the web. It is difficult to conceive that an animal
whose feet are not extremely sensitive could travel on or manipu-
late the delicate strands of these orb-webs.

IV. DISCUSSION AND SUMMARY.

It is maintained in this paper as in a previous one (Barrows, '07)
that an animal exhibits a "tropism" or better a taxis, "when
under the influence of [chemical] stimuli acting unilaterally they
move toward or away from the source of the stimulus" (Verwrorn,
'99, p. 249). It has been shown above that Epeira sclopetaria
orients in its web and moves toward the source of a vibratory
mechanical stimulus when this is of an appropriate rate and
amplitude. Thus this method of response to a vibratory stimulus
identifies the reaction as a positive taxis. The term tonotaxis
would naturally be used in this connection, but since tonotaxis
has been used in another way it seems advisable that the terms
positive vibrotaxis should be applied if a short descriptive term
is desired.

The foregoing may be summarized as follows:

1. Epeira sclopetaria, an orb-weaving spider, starting from the
center of its web is able to orient, charge and seize flies which
strike and are detained in the web. This process is carried out
with extreme rapidity.

2. With the aid of a mechanical vibrator it is possible to show
that the stimulus is vibratory, the spider orienting to and
attacking the vibrator even in the dark.

3. The response can be analyzed into, (a the orientation,
(b) the forward response, and (c) the attack on the vibrating
object. The response is in essence a positive vibrotaxis.

4. The vibrations are transmitted through the web in all
directions from the vibrating point but the intensity (amplitude)
decreases toward the center of the web and on either side. The
lines of equal intensity of the vibration form roughly a series of
circles the centers of which are at the vibrating point.

326 WILLIAM MORTON BARROWS.

5. The sense organs used in detecting the stimulus are probably
sense hairs on the tarsi.

6. This orb-weaving spider provides itself with a temporary
extension of its tactile sense organs which makes its tactile sense
in reality a distance receptor, much like an auditory or an ol-
factory organ.

BIBLIOGRAPHY.
Barrows, W. M.

'07 The Reactions of the Pomace Fly; Drosophila ampelophila Loew, to

Odorous Substances. Jour. Exp. Zool., Vol. IV., No. 4, p. 515-53?-
Boys, C. V.

'80 The influence of a Tuning Fork on the Garden Spider. Nature, Vol. 23

p. 149.
Comstock, J. H.

'12 The Spider Book. Garden City, N. Y.
Dahl, T.

'83 Ueber die Horharre bei den Arachniden. Zool. Anz., VI.
Davenport, C. B.

'08 Experimental Morphology, New York.
McCook, H. C.

'90 American Spiders and their Spinning Work. Vol. II. Philadelphia.
Mclndoo, N. E.

'n The Lyriform Organs and Tactile Hairs of Araneads. Proc. Acad. Nat. Sci.

Phil., May, 1911.
Packard, A. S.

'03 A Text-book of Entomology, New York.
Peckham, G. W. and E. G.

'87 Some Observations on the Mental Powers of Spiders. Jr. Morph., I., pp

383-4I9-
Verworn, M.

'99 General Physiology. Trans, by F. S. Lee. London, 8vo. xvi+6ispp.

328 WILLIAM MORTON BARROWS.

EXPLANATION OF PLATE I.

FIG. i. Showing a female Epeira sclopetaria in the normal resting position in
the web. The arrow indicates the place where the vibrator straw touches a radial
strand of the web.

FIG. 2. The same individual, shown in Fig. i, orienting to the vibrator which
had been in motion for a fraction of a second just before the photograph was taken.

BIOLOGICAL BULLETIN, VOL. XXIX

FIG. i.

W. M. BARROWS.

FIG. 2.

330 WILLIAM MORTON BARROWS.

EXPLANATION OF PLATE II.

FIG. 3. A spider attacking the vibrator straw while it is in motion.
FIG. 4. A spider in the act of responding to the Y-shaped vibrator. One prong
of the vibrator appears in front, the other behind the spider.

BIOLOGICAL BULLETIN, VOL. XXIX.

FIG. 3.

W. M. BARROWS.

FIG. 4.

332 WILLIAM MORTON BARROWS.

EXPLANATION OF PLATE III.

FIG. 5. A photograph showing the spider in the normal resting position in the
web, while the vibrator is in motion. The arrow indicates the place where the
vibrator straw touches a radial strand. The doubling or blurring of the lines of the
web shows the distribution of the vibrations.

BIOLOGICAL BULLETIN, VOL. XXIX.

W. M. BARROWS.

FIG. 5.

Vol. XXIX December, 1915. No. 6

BIOLOGICAL BULLETIN

OBSERVATIONS ON THE DEVELOPMENT OF
COPIDOSOMA GELECHI^E.

J. T. PATTERSON.

From the Marine Biological Laboratory, and the Zoological Laboratory of the University

of Texas (Contribution No. 127).

CONTENTS.

I. Introduction 333

II. Note on the Life History of Gnorimoschema 336

III. Parasites of Gnorimoschema salinaris 340

IV. Development of Copidosoma gelechiae 341

1. Polygerm Stages 341

(a) Youngest Stages 341

(6) The Nucleated Membrane 343

(c) Precipitated Material 343

(d) The Granular Protoplasm and Embryonic Nuclei 343

(e~) Growth of the Polygerm and Formation of the Primary

Divisions or Masses 344

(/) Formation of the Secondary Masses 346

(g) The Inter-embryonal Substance 346

2. Dissociation of the Polygermal Mass 347

3. Pupation, and the Emergence of the Imagines 348

4. The Abortive Embryos 349

V. Number and Sex of Copidosoma Parasites found in Gnorimoschema 352

Summary 358

Literature 359

Plates 362

I. INTRODUCTION.

The discovery of polyembryonic development among certain
of the hymenopterous parasites has opened up an extremely
interesting field for investigation. Like most other important
biological discoveries, this one was foreshadowed by the obser-
vations of several different naturalists. In a paper of this nature
it is not necessary to give an extended account of the history of
this discovery. We shall therefore be content with a brief

333

334 J- T- PATTERSON.

statement on this point, limiting the account almost entirely to
the species with which the paper deals.

The general features of polyembryony in insects have been
given in the well-known papers of Marchal ('98, '04) and Sil-
verstri ('06, '08), but there are many points concerning the
details of this process which have not as yet been worked out.
It was with the view of studying certain of these details that led
the writer three years ago to seek an American species upon
which such studies could be made. Dr. L. O. Howard1 suggested
that Copidosoma gelechiiz, which parasitizes the larvae of the
Solidago gall moth, Gnorimo schema gallcesolidaginis, would be
a good form upon which to work, as it seemed to be an undoubted
case of polyembryony.

The Gnorimoschema moth makes the ellipsoidal galls on the
stems of several species of goldenrod. Baron Osten Sacken ('63)
seems to have been the first to have published a description of
the inflated carcass of the Gnorimoschema larva, caused by the
chalcis parasite; but apparently he was not acquainted with
the maker of the gall. In 1869 in connection with his account
on the life history of this moth, Riley states that the caterpillar
serves as a host for no less than six different species of hymenop-
terous parasites. One of these, which is shown in his Fig. 6,
Plate 2, is described as a "little fly of a dark metallic green color,
with reddish legs." This is clearly Copidosoma. Riley states
that the larvae of this species infests the caterpillar in great
numbers, more than 150 having been obtained from a single host.
He supposed that the "mother fly" gnawed a passage through
the gall and desposited her batch of eggs in the inmate. He
pointed out that the larval parasites cause the caterpillar to
swell to three or four times its natural size, and after having
absorbed all the juices of the victim, form very small brownish
cocoons, which are so packed together that they give to the worm
the puffed-up appearance which is typical of the mummified
carcasses of lepidopterous larvae that have been parasitized by a
polyembryonic species. It was this inflated condition of the
larval host that led Riley to call the parasite the " Inflating Chalcis

1 For this as well as for other suggestions received throughout the progress of
the work, the writer is greatly indebted to Dr. Howard.

DEVELOPMENT OF COPIDOSOMA GELECHI/E. 335

Fly." Howard ('85) later named this species Copidosoma
gdechicE.

Upon examining the various goldenrods about Woods Hole,
Mass., for galls of Gnorimo schema, it was found that Solidago
sempervirens furnished the best opportunity for obtaining
material. However, the common gall maker of this solidago
proved not to be Gnorimoschema gallcesolidaginis Riley, but a
closely related species, G. salinaris Busck. The parasites in-
festing these two moths are varieties of the same species, Copi-
dosoma gelechics.

The selection of this species has not proved altogether satis-
factory, because the gall-making habit of the host complicates
the life history and renders the collecting of material for early
stages of the parasite somewhat more difficult than from a host
which feeds openly. Furthermore, the moth, and likewise the
parasite, has but one generation a year. In addition to these
objections, there is the further one that the egg of Copidosoma
gives rise to a relatively large number of individuals (about 191
on the average). In attempting to obtain material for the studies
which the writer has in mind, it seems best to seek to find a
host which is an open or semi-open feeder, which has two or more
generations a year, and which harbors a parasitic egg giving rise
to but few individuals. During the past summer at least two
species have been found which in the main seem to fulfill these
conditions. It therefore seems best to publish the main facts
concerning the development of Copidosoma before giving it up
for more favorable material.

There is one feature in the development of Copidosoma which
makes further study desirable. We refer to the abortive em-
bryos (presently to be described), which at first were thought
to be comparable to the so-called asexual larvse of Litomastix
truncatdlus. It will be recalled that Silvestri ('06) described in
this species the development of both sexual and asexual larvae
from a single egg. In one instance he secured from a caterpillar
of Plusia gamma 1,700 sexual and 220 asexual larvae of Litomastix.
He believes that the asexual larvae play the role of raspers for
the normal larvae, tearing the tissues of the host so that the sexual
larvae may the more easily secure the necessary food. It may

336 J. T. PATTERSON.

be stated here that the abortive larvae of Copidosoma are in no
way comparable to the asexual larvae of Litomastix as described
by Silvestri.

II. NOTE ON THE LIFE HISTORY OF GNORIMOSCHEMA.

In order to collect polyembryonic material it is essential to
know something about the life history of the host; especially is
this true in cases like Gnorimo schema in which the larval host is
a gall maker. Considerable attention has therefore been given
to a study of the life history of G. salinaris.

The general habits of the Solidago gall moths were first made
known by Riley's ('69) studies on G. gallcBSolidaginis . According
to Riley this species winters over in the imago stage and may be
seen flying in the month of May. When the young plants
(Solidago nemoralis) are about six inches high the female moth
lays her egg either in the terminal bud or at the side of the stalk
immediately below the bud. The young caterpillar upon hatch-
ing burrows into the stalk and starts the development of the gall.
By the first of June the gall has just begun to form and contains
a larva about one-third grown. The larva and its ellipsoidal
gall reach their full size by the middle of July. The caterpillar
which now measures over half an inch in length prepares for the
change into the chrysalis state by first eating a round passage-
way through the wall well toward the upper end of the gall. The
orifice is then closed by a secretion of liquid silk, which hardens
to form a silken plug. After closing the orifice, the caterpillar
lines the passage-way and the walls with a delicate silk, and then
transforms into a shiny, mahogany-brown pupa, about a half
inch long. The moths begin to emerge about the middle of
August and continue to appear until the beginning of October.

Many phases of the life history of G. salinaris are similar to
those of G. gallaesolidaginis , but there are some important differ-
ences. The earliest date at which galls of the marsh goldenrod
have been secured was June 12, 1914, and at that time many of
the galls were well started. Between June 12 and 15, 63 galls of
various sizes were collected and examined. They varied in size
from 8 to 12 mm. in length and from 4 to 17 mm. in transverse
section. In shape the galls also vary greatly. Some are distinctly

DEVELOPMENT OF COPIDOSOMA GELECHI/E. 337

pear-shaped, while others are fusiform, with various gradations
between these two general types. The galls occur at different
heights on the stem, but the vast majority of them are located at
or near the base of the stalk (Fig. i). Their position is un-
doubtedly determined by the location of the point at which the
larva penetrates the young shoot. If this point is located toward
the base of the young stalk, the gall will naturally appear near the
base of the fully grown plant; but if it is located in or near the
terminal bud, the gall will appear some little distance up on the
stem. Occasionally two galls are found on the same plant (Fig.
8). A few cases have been observed in which the gall was located
at the tip of the terminal bud, producing a stunted plant without
a central, flower-bearing stalk. With these few exceptions, the
gall of G. salinaris does not seem materially to affect the growth
and vigor of the plant. It is true that many galls are found on
plants that are apparently stunted but such dwarfing is to be
attributed to the adverse conditions under which the plant
sometimes grows. In regions that are very much exposed to
the wind, like the banks along the coast, many of the goldenrods
are small and clearly dwarfed; but this condition applies as well
to the plants that are free from galls as to those that are infected.

The habits of gall making are similar in the three common
species of Gnorimoschema, although the following differences may
be pointed out. G. gallceasteriella produces a triangular gall at
the top of the dwarfed or stunted stems of Solidago ccesia, S. axil-
laris, S. latifolis, and Aster divaricatus -1 The form of the gall
differs somewhat with the plant. The gall of G. gallaesolidaginis
may occur toward the top of the stem, but usually it is located
just below the middle, especially is this true of the galls on S. can-
adensis. The galls of this moth do not dwarf the plant. The
condition of the galls of G. salinaris on the marsh goldenrod has
already been described. They occur nearer the base of the
stem than do those of last species, and like the latter there is
little or no tendency to dwarfing the plant.

The larvae secured from the galls collected between June 12
and 15 varied from 3 to 8 mm. in length. Beginning with the
middle of June, the young caterpillars grow rapidly, doubling

1 Part of these data were kindly furnished the writer by Dr. T. M. Forbes.

338

J. T. PATTERSON.

their size within a fortnight. By the middle of July they have
reached their full growth, and are beginning to show signs of
undergoing pupation, which is evidenced by the construction of
the passage-way. The passage-way and its orifice differ in two
respects from those of G. gallcesolidaginis as described by Riley
('69). The silk lining does not extend much beyond the lower
limits of the passage-way, and hence does not cover the inner
surface of the wall. The second difference is seen in the character
of the orifice and its silk plug. The caterpillar of G. salinaris
does not cut the passage-way quite through the wall, but leaves
the very thin epidermis of the stem, which is used as a back-
ground for the construction of the plug (Fig. 7).

TABLE I.

TABLE SHOWING DATES OF PUPATION AND EMERGENCE OF COPIDOSOMA AND

GNORIMOSCHEMA.

Pupation
(Beginning of)

Emergence •

Copidosoma

Gnorimoschema

Copidosoma

Gnorimoschema

Aug.

6,

IQI2.

Aug.

5.

1913.

July

3i.

1914.

.July

30,

I9I5-

Aug.

6,

1912.

July

23.

1913.

July

30,

1914.

.July

26,

I9I5-

Aug.

25

to Sept.

12,

1912.

Sept

3

to Sept.

13.

I9I3-

Aug.

30

to Sept.

18,

1914.

.Aug.

24

to Sept.

21,

I9I5-

Aug.

25

to Sept.

IO,

1912.

Aug.

25

to Sept.

10,

I9I3-

Aug.

22

to Sept.

II,

1914.

Aug.

24

to Sept.

14-

1915.

Pupation occurs during the last week of July and the first
week of August (Table I.). The imagines begin to emerge about
August 25, and continue to appear until September 10. The
moth has been seen flying in the open during this period.

Females kept in captivity often lay eggs. This they do within
ten days after emerging, and irrespective of their association with
males. As a rule the moths simply drop the eggs on the bottom
of the cage, or they may lay them on the leaves and flowers of
goldenrods placed in the cage. At first it was thought that
G. salinaris must differ from G. gattcBsolidaginis in respect to
its egg-laying habits, for Riley states that the latter species
although emerging in the fall, hibernates as an imago and lays

^H DEVELOPMENT OF COPIDOSOMA GELECHI^E. 339

its eggs in the following May. It has been discovered, however,
that G. gallasolidaginis from the galls of S. canadensis in western
Ohio likewise drops several eggs soon after emerging from the
pupa in September. This raises the question as to whether these
fall eggs develop into larvae, for if so it would be difficult to explain
how the young caterpillars could withstand the winter and succeed
in the spring in finding a young goldenrod bud or shoot in which
to start the new gall.

In reply to an inquiry, Mr. A. Busck of Washington kindly
informed the writer that the laying, of eggs by Gnorimoschema
was of no particular significance, as it is not uncommon for certain
Lepidoptera to drop their eggs prematurely, especially if kept in
captivity. In view of this fact an observation made in the fall
of 1913 is of special interest. During the first few days of Sep-
tember of that year a single female, confined in a cage with several
males, laid a dozen or more eggs on goldenrod leaves and flowers.
On the thirteenth of the month three larvae hatched from this
batch of eggs! There can be no possible doubt as to the correct-
ness of this observation, for the hatching of one of the little cater-
pillars was actually observed under a hand lens.

It is difficult to explain the development of these larvae from
fall eggs, except on the basis of parthenogenesis. It is true that
the female which laid the eggs from which the larvae developed
had been confined with males; but although males and females
have been kept together for several weeks during each of the last
three seasons, yet mating has never been observed. The sup-
position that the fall eggs of G. salinaris may develop by parthen-
ogenesis receives strong support from a study of sections of eggs
laid by a female not associated with males. In Fig. 20 is shown
a transverse section of one of her eggs and it can clearly be seen
that development is well started. Twelve eggs out of the batch
were sectioned, and it was found that eleven had started to
develop, although apparently not in a normal manner. It is
not improbable that some few eggs may develop normally and
eventually produce larvae. The question of parthenogenesis in
the Solidago moths is one needing further study.

It might be worth while to add that parthenogenetic develop-
ment among Lepidoptera is by no means unknown. DeGeer is

34° J- T. PATTERSON.

given credit for having first discovered long ago that certain
butterflies belonging to the genus Solenobia lay unfertilized eggs
which develop into normal imagines, and later von Siebold not
only confirmed this observation, but also discovered that Psyche
helix reproduced parthenogenetically. It has since been shown
by several workers that the silk moth, Bombyx mori, may under
certain conditions reproduce by parthenogenesis.

III. PARASITES OF GNORIMOSCHEMA SALINARIS.

Riley reports six hymenopterous parasites for Gnorimo schema
gallcesolidaginis , and in addition to these he found a beetle larva
and another lepidopterous larva which intrude as inquilines
within the cavity of the gall made by this species. At least five
hymenopterous parasites have been found associated with
G. salinaris. The most important of these is Copidosoma
gelechice, which is by far the most common parasite attacking
the moth. The other four species are CalUephialtes notanda
Cress, Epiurus sp., Eurytoma sp. (pupa), and Pseudacrias sex-
dentatus Girault. The first of these four occurs most frequently,
while the last has been observed but a few times. However, it
is of special interest, inasmuch as it is the only observed case of a
second parasite emerging along with Copidosoma, although the
larvae of other species have been found associated with the larvae
of Copidosoma. On September 3, 1914, six individuals, all
females, emerged together with a brood of about one hundred
Copidosomas from a single carcass. The small pupae of Pseu-
dacrias lying among those of Copidosoma were observed through
the transparent chitin of the carcass of the host some days prior
to their emergence. They were not grouped together but scat-
tered about in different parts of the carcass. Each pupa was
inclosed in a chamber very much smaller than, but exactly similar
to that containing a Copidosoma pupa.

Usually Pseudacrias larvae do not pupate until after the larval
host has undergone this process. About a dozen Gnorimoschema
pupae have been found containing Pseudacrias pupae, which later
emerged. It is not probable that Pseudacrias is polyembryonic.
First, because both male and female individuals usually emerge
from the same pupal host; and second, because the individuals

DEVELOPMENT OF COPIDOSOMA GELECHI/E. 34!

do not come out at the same time. The single instance of six
females ssuing simultaneously with the brood of Copidosoma
can be explained by assuming that a single female deposited six
fertilized eggs in the host at the same time. However, this case
is of special interest as it demonstrates the synchronous develop-
ment in a single host of the broods of two distinct parasites, and
thus supports Wheeler's ('10) suggested explanation of Silvestri's
so-called asexual larvae in Litomastix.

In addition to the five hymenopterous parasites, there are two
insect larvae associated with the larva of G. salinaris. They are
undoubtedly inquilines. One of these is a beetle and the other a
lepidopterous larva (Fig. 5). Judging from Riley's account,
these two species are very similar to if not identical with the cor-
responding inquilines reported by him for the galls of G. galltz-
solidaginis.

IV. DEVELOPMENT OF COPIDOSOMA GELECHI^:.
i. The Poly germ Stages.

(a) Youngest Stages. — We have not secured the cleavage
stages of Copidosoma, owing to the fact that they occur earlier in
the year than we have been able to reach Woods Hole. There-
fore, in describing the developmental processes which have their
inception in the cleavage stages, we must rely upon the work of
other investigators in this field for our interpretation of the sig-
nificance of these processes.

The youngest stages secured were found in a small larva of
Gnorimo schema, taken June 21, 1913. The series of sections of
this small caterpillar contains three young polygerms of Copi-
dosoma. Evidently the egg from which the caterpillar developed
had had three parasitic eggs deposited in it. Two of the poly-
germs, which lie close together, are situated in the first and
second body segments of the larva, just beneath the hypodermis;
while the third is found in sections 5 to 14 posterior to these, and
is also situated just beneath the hypodermis of the host.

The three polygerms are not of the same size, as is indicated
by the following measurements : Of the two specimens lying close
together, the larger measures 150^1 by 82 /JL and runs through
15 sections (150 /*), the smaller measures 103 n by 71 fj., and is

342 J- T. PATTERSON.

found in 12 sections; the single specimen measures 179 /x by 95 /JL
and occupies 8 sections only.

In structure the three polygerms are practically identical.
Each consists of two distinct zones: (i) An outer relatively
thick zone containing a large number of nuclei irregularly placed,
and (2) a central region containing the embryonic nuclei (Fig.
19). In the absence of the earlier stages, it is not an easy matter
to interpret the conditions seen in these polygerms. In the main
they correspond most nearly to the conditions in the egg of
Litomastix (Copidosoma} truncatellus , as described by Silvestri
('06). I therefore interpret the outer zone to be the product of
the "polar ooplasm" plus the "polar nuclei," while the central
region contains the true embryonic nuclei, derived from the
fertilized nucleus, or in the case of parthenogenetic development,
from the matured egg nucleus.

There is of course one essential difference in the corresponding
stages of Litomastix and Copidosoma. In the polygerm of the
former the central region is composed of a solid mass consisting
of distinct cells, while in the latter this region is on the point of
being broken into multi-nucleated masses, which form the pri-
mordia of the embryos (cf. Fig. 19 A with Silvestri's Fig. 33,
PI. III.). It may be that the embryonic nuclei are delimited
by cell walls in Copidosoma, although one can not make them
out with certainty, even under the highest powers of the micro-
scope. Judging from the work of other investigators, one would
expect to find the embryonic nuclei surrounded by cell walls.
In Ageniaspis, Marchal ('04) first reported that the early embry-
onal masses were pluri-nuclear in character, but Silvestri ('08)
and Martin ('14) have later demonstrated that the nuclei are
surrounded by cell walls. In Copidosoma the embryonic nuclei
are often so closely packed together that the demonstration of
cell walls would be extremely difficult.

The three polygerms mentioned above are of particular interest,
in that they show very clearly the manner in which the central
mass with its nuclei breaks up to form the primordia of the
multiple embryos. The central region itself consists of two
distinct substances, (i) A granular protoplasm in which the
embryonic nuclei lie, and (2) a more fluid-like material which

DEVELOPMENT OF COPIDOSOMA GELECHI/E. 343

becomes greatly shrunken during the process of fixation, and
which in sections appears as a precipitated substance (Fig. 19 A,
P.M.}. As to the origin of these different substances we know
nothing, but their subsequent history is clear. For the sake of
clearness in description we shall use the following terms: (i) l$u-
deated Membrane for the outer zone; (2) Granular Layer for the
protoplasm containing the embryonic nuclei ; and (3) Precipitated
Material for the shrunken fluid-like substance.

(b) The Nucleated Membrane. — In these young polygerms the
outer zone stains more deeply than the central mass. The
"polar nuclei" have no definite arrangement, but are irregularly
scattered throughout the protoplasm. The entire zone therefore
is in every sense of the word a syncytium. As the polygerm
grows in size the nuclei become arranged into a single layer,
and the protoplasm thins out, thus forming a true nucleated
membrane about the central or embryonic portion of the egg
(Fig. 21, N.M.). In the later history of the polygerm some of
the nuclei are clearly surrounded by cell walls, that is, there is a
tendency for the membrane to become cellular.

At first the young polygerms are naked, that is there are no
elements from the host tissue laid down on the outer surface
of the nucleated membrane. Later a few mesenchyme cells are
found on the surface of the membrane, and still later these cells
give rise to the adipose tissue (Fig. 22, A.T.}, which may com-
pletely surround the polygerm.

(c) Precipitated Material. — This material occupies the central
portion of the polygerm. Apparently it is formed through the
action of the fixing reagent upon the fluid-like protoplasm. In
sections it is very much shrunken, thus leaving an irregular clear
space (Fig. 21, C). As we shall see later, it persists throughout
the entire polygerm phase of development.

(d) The Granular Protoplasm and the Embryonic Nuclei. — In
Fig. 19 the condition of the embryonic nuclei and their surround-
ing granular protoplasm is especially clear. Most of the nuclei
are indifferently scattered in the protoplasm, but some of them
are collecting into groups. The number of nuclei in each group
is variable; some groups contain only two or three nuclei, while
others may have as many as ten or twelve. The granular pro-

344 J- T. PATTERSON.

toplasm surrounding a group of nuclei soon rounds off and the
primordial embryo with its surrounding layer lies free within the
more fluid contents of the central region of the egg (Fig. 19 A}.
The more usual condition is for the spherical mass to remain
attached at one side to the peripheral layer of the granular pro-
toplasm (Fig. 19 B, P.E.). Eventually all of the embryonic
nuclei thus become included in these spherical masses of pro-
toplasm, and thus become isolated as primordia of the embryos.

The condition at the close of the formation of the primordia
is shown in Fig. 21. This specimen was found in a series of
sections of a 3 mm. caterpillar, taken June 15, 1914. In the
median section it measures 113 M by 203 /i, and runs through 40
sections (280^1). It lies in the middle portion of the body
cavity, to one side of the intestine, which on account of the size
of the polygerm is pushed out of place. As compared with the
preceding polygerms this one is very much larger and shows a
number of important changes. The nucleated membrane has
become much thinner and its nuclei are arranged more or less
into a single layer. The adipose tissue is being laid down on the
outer surface of the membrane. The most important change,
however, has occurred in the embryonic masses themselves.
The protoplasm which surrounds a group of nuclei is differenti-
ated into two distinct regions. The central part, crowded with
nuclei, stains somewhat lighter than the peripheral zone, which
forms a relatively dense layer about the central core (Fig. 21,
P.E.). There are still a few nuclei which have not as yet been
surrounded by the dense layer, but this stage marks approxi-
mately the end of the division of the germ into separate embryos.

(e) Growth of the Polygerm and the Formation of the Primary
Divisions or Masses. — Upon the completion of the primitive
embryos, the polygerm grows very rapidly. It first extends in
the direction of its long axis, soon transforming into an elongated
cylindrical structure. One specimen showing this condition
measures in section 148 /j. by 430 yu. It never becomes an elon-
gated tube as does the polygerm of Ageniaspis. During this
growth the adipose tissue is laid down in the form of a thick
layer about the polygerm. One of the easiest ways in which
to find a polygerm of this and later stages is to examine the large

DEVELOPMENT OF COPIDOSOMA GELECHI/E. 345

fat bodies lying in the middle region of the body cavity of the
larval host. If the caterpillar is parasitized one of these bodies
is almost certain to contain the polygerm.

After the elongated condition is attained, the further growth
of the polygerm may take place in any direction. In some
cases the extension is mainly in one plane, thus transforming the
polygerm into a flat, plate-like structure (Fig. 13). In other
cases it forms a thick irregular mass (Fig. u), and when viewed
as a whole mount shows many elevations on its surface, due to
the breaking up of the entire polygerm into single masses, each
of which contains an embryo.

During the rapid expansion of the polygerm a very important
change takes place in its structure, whereby each embryo become
enclosed in a double involucre. The first step in this process
begins just prior to that period of development in which the
polygerm attains its elongated, cylindrical shape. It consists
in the formation of constrictions in the nucleated membrane
which break up the single polygerm into a series of primary
divisions or masses (Fig. 15). In the specimen shown in this
figure there are about twelve of these masses. Each primary mass
has the same general structure as the original single polygerm.
It is surrounded by a portion of the nucleated membrane, con-
tains precipitated material, and has a variable number of em-
bryos, from five to fifteen or more.

In Fig. 22 one end of a longitudinal section of a polygerm is
shown with the completed primary masses. Three of these
masses are seen in the figure, together with a portion of a fourth.
Attention should be called to the fact that the adipose tissue,
although in contact with the polygerm, is still a distinct structure.
In the process of forming the primary masses not all of the elements
of the nucleated membrane are taken into these structures.
Some of them are left behind and later lie in the inter-embryonal
spaces or interstices. In Fig. 22 a number of these elements
(cells and nuclei) are shown at the point marked " N," lying
between the primary masses and the adipose layer.

In another portion of the same polygerm a single primary
mass is being constricted off laterally. It appears as a bud ex-
tending from the main body of the polygerm. It is such cases

346 J. T. PATTERSON.

as this which give rise to the condition frequently seen in whole
mounts, in which the surface of the polygerm displays many
protuberances.

(/) Formation of the Secondary Masses. — The primary masses
soon become broken up into secondary masses. This is also
brought about by constrictions of the nucleated membrane (Fig.
23). These secondary masses may contain more than one
embryo, in which case they immediately form constrictions which
result in producing still smaller masses, each of which contains a
single embryo.

In the constrictions which lead to the cutting off of a single
embryo with its involucres, some of the precipitated material is
enclosed between that portion of the granular layer which is in
contact with the embryo and that part lying adjacent to the
inner surface of the nucleated membrane. These two parts of
the granular layer then fuse, forming a single involucre in which
are the spaces containing the precipitated material (Fig. 24).
The embryo is thus surrounded by two involucres, a granular
layer, and a nucleated membrane (Fig. 26). In some cases the
precipitated material may be so abundant as to form a solid zone
between the inner and outer parts of the granular layer; in others
it is small in amount and gives the appearance of much flattened
nuclei lying within this layer (Fig. 26, P.M.}.

(g) The Inter-embryonal Substance. — At the close of the for-
mation of the single embryonic masses and their involucres the
inter-embryonal interstices are already filled with a substance
derived from several different sources. It consists of a plasma-
like matrix in which are embedded cells and nuclei. We have
already noted that during the formation of the primary and
secondary masses some of the elements from the nucleated
membrane are not included in the outer involucre, but are left
in the inter-embryonal spaces. During the early history of the
inter-embryonal substance, it consists mainly of product from
this membrane. Later cells from two other sources enter into
its formation. First, leucocytes from the host are found
embedded in the matrix. They are especially abundant in those
regions of the polygerm exposed directly to the body cavity,
that is near a surface barren of adipose tissue. Second, fat cells

DEVELOPMENT OF COPIDOSOMA GELECHI^. 347

from the adipose layer invade the inter-embryonal spaces. The
fat cells are the last elements to enter the inter-embryonal sub-
stance. In Fig. 13 a wedge-shaped mass of fat tissue is seen
lying between the embryos in the middle region of the polygerm,
on the upper side. Perhaps it would be more correct to say that
the embryos bud out into the adipose tissue. Thus in Fig. 24
a single primary mass has been budded off into the adipose tissue.
The final condition of the polygerm at the end of the formation
of the inter-embryonal substance is shown in Fig. 16. The
adipose tissue has invaded the inter-embryonal substance from
all sides of the polygerm and has become an organic part of this
substance. The fat body and the included polygerm become
an extremely complex structure, which may be called the poly-
germal mass.

2. Dissociation of the Polygermal Mass.

The setting free of the larval parasites into the body cavity
of the host is brought about through the dissociation or disin-
tegration of the inter-embryonal substance. The fat brought
into close contact with the embryos by the invasion of the adipose
tissue is digested and absorbed by them. It is therefore the first
component of the inter-embryonal substance to disappear.
That the fat is digested and consumed by the embryos is evi-
denced by the fact that the numerous other fat bodies remain
intact during this period. The disappearance of the fat leaves
the embryos loosely held together by the plasmalike matrix,
which in turn soon disintegrates, freeing the larvae.

The first larvae to be set free are those situated at the periphery
of the polygermal mass. Such larvae are usually the largest
present in the mass. As the inter-embryonal substance slowly
disintegrates the remainder of the larvae are gradually set free
(Fig. 17). The earliest date at which free larvae have been found
was July 19; the latest, July 31. In the vast majority of cases
the mass dissociates during the last week of July.

The larvae retain the involucres for some time after being set
free (Fig. 18). Once free in the body cavity they proceed to
devour the contents of the host, first consuming the fat tissue,
and finally the various organs. The last internal organ to dis-
appear is the intestine.

34-8 J- T. PATTERSON.

3. Pupation, and the Emergence of the Imagines.

Pupation in Copidosoma occurs during the first ten days of
August. The pupa stage lasts twenty-eight days. As stated
above, the larvae destroy all of the internal organs of the host, and
consume such portions as are dissolved by the action of their
salivary secretions. The undissolved portion consists largely
of the chitinous parts of the trachese. The larvae also destroy
all of the body wall except the superficial layer of chitin. During
the process of pupation the non-digested content of the cater-
pillar hardens and forms the thin-walled, oval chambers in which
the parasitic larvae lie and in which they undergo their trans-
formation into pupae. The superficial layer is perfectly trans-
parent, and at first is very flexible. Later, as drying occurs, it
shrinks in on the walls of chambers and becomes hard and rigid,
the whole forming the typical mummified carcass (Figs. 2, 4, 6).
Practically all of the pupae are oriented in a definite fashion in
the carcass. Their heads are directed toward the anterior end
of the carcass. Just before becoming immobile, the Gnorimo-
schema larva almost invariably turns the head upward in the
gall chamber; likewise, the parasitic larvae, just before pupating,
orient themselves so that their heads are directed upward, in the
direction of the anterior end of the carcass.

The imagines come out during the last week of August and the
first week of September (Table I.). They escape by gnawing
holes through the walls of the chambers and the superficial
chitinous layer, both of which become very fragile. As a rule
they all emerge practically at the same time. Several cases
have been observed in which the entire brood has escaped within
a period of ten minutes.

Once free from the carcass, they immediately gnaw a hole
through the wall of the gall. Their escape is greatly facilitated
by the habit of the caterpillar, just before becoming immobile,
of eating out a passage-way to, or nearly to the epidermis of the
plant. But in no case does the parasitized caterpillar secrete
a silken plug. Hence, in order to escape to the exterior, the
parasites have only to cut through the remaining thin portion of
the wall.

The parasites must winter over in the imago state; otherwise

DEVELOPMENT OF COPIDOSOMA GELECHI^E.

349

they would not be able to parasitize the normal or spring eggs of
Gnorimo schema. Copulation, however, takes place immediately
after the adults emerge, but the females do not parasitize the

TABLE II.

TABLE SHOWING VARIATION IN LENGTH OF LARVAE IN THREE LOTS OF

COPIDOSOMA.

Length in Lines.

Lots.

Length in Lines.

Lots.

I

II

III

I

II

III

I

29

3

3

2

30

10

7

3

2

31

3

i

4

3

32

I

5

5

10

33

i

i

6

3

14

34

2

3

7

13

35

4

3

8

I

12

36

2

I

9

4

37

8

5

10

I

3

38

I

4

n

i

4

39

I

2

2

12

2

40

3

3

9

13

2

4i

4

4

4

14

2

42

2

7

5

15

I

6

43

2

3

7

16

I

3

44

I

3

7

17

4

4

12

45

3

5

2

18

i

7

46

I

8

4

19

i

6

16

47

I

5

3

20

4

17

48

I

10

i

21

9

49

2

13

2

22

9

7

50

10

2

23

5

6

5i

4

24

3

4

52

2

8

25

3

13

53

i

26

3

5

54

27

I

4

55

I

2

28

4

I

56

fall eggs of this moth. Only on one occasion has an attempt to
ovipost in such eggs been observed. In this instance the few
females which made the attempt were not able to penetrate the
shell of the egg with the ovipositer.

4. The Abortive Embryos.

One of the most interesting discoveries made in connection
with the study of Copidosoma is what we shall call the abortive
embryos or larvae, to which brief reference has already been
made. Abortive embryos occur in the development of many
different species of both invertebrates and vertebrates. They

35O J. T. PATTERSON.

are especially common in mammals. For example, my colleague,
Dr. C. G. Hartman, has found a great mortality of embryos in
the development of the opossum. Degenerating embryos are
found throughout the brief but entire period of gestation.
Abortive embryos have been found in at least three other species
which have a polyembryonic type of development. One of the
two embryos which develop from a single egg of the earthworm,
Lumbricus trapezoides, sometimes degenerates. Fernandez ('09)
has observed rudimentary embryos in the South American
armadillo, Tatusia hybrida, and I have on several occasions seen
them in the blastocyst of Tatusia novemcincta. But in no case
with which we are acquainted is their number and constancy of
occurrence so striking as in Copidosoma.

Our attention was first attracted to these abortive embryos
while dissecting out a batch of larvae from a large caterpillar.
Most of the larvae in the lot were large and about on the point of
undergoing pupation, but in addition to these large individuals,
there were a number of smaller ones. At first it was supposed
that two distinct species of parasitic larvae were present, or that
we had a condition similar to that described by Silvestri for
Litomastix, of sexual and asexual larvae. It was noted, however,
that the small larvae had the same general structure as the larger
individuals, except that they still possessed the two involucres
typical of all of the younger larvae of this species.

A study of serial sections of more than a hundred polygerms
has completely demonstrated beyond any possibility of doubt
that the small rudimentary embryos are derived from the same
egg as larger normal larvae, and consequently do not belong to a
different species. The sections show that degenerating embryos
are to be found in every stage of development of the polygerm,
from the time of the formation of single embryos until the larvae
are set free into the body cavity of the host. In Fig. 24 is shown
a degenerating embryo which has not yet been completely cut
off from its fellow by the constriction of the nucleated membrane.
Its nuclei have already completely disintegrated. In Fig. 26 is
another embryo well on the way to complete disintegration.
Finally Fig. 17, which is a portion of a polygermal mass about at
the close of dissociation, contains at least four or five rudimentary
embryos. They stain darker than the normal individuals.

DEVELOPMENT OF COPIDOSOMA GELECHI^E.

351

The degeneration of embryos or larvae does not cease immedi-
ately after the dissociation of the polygermal mass, but such
embryos are found up until the beginning of pupation. About
fifty lots of free larvae have been dissected out of caterpillars, and

TABLE III.

TABLE SHOWING THE NUMBER OF PARASITES IN FEMALE BROODS.

Brood.

No. of Individuals.

Brood.

No, of Individuals.

I

25

46

20O

2

42

47

2OI

3

49

48

2O7

4

52

49

2IO

5

73

50

210

6

89

Si

212

7

91

52

213

8

95

53

213

9

IOO

54

214

10

IOO

55

215

ii

1 06

56

215

12

108

57

216

13

H5

58

216

14

119

59

217

IS

I2O

60

229

16

121

61

234

I?

122

62

236

18

124

63

236

19

124

64

237

20

125

65

245

21

131

66

248

22

137

67

250

23

142

68

251

24

145

69

254

25

146

70

256

26

ISO

7i

257

27

151

72

260

28

153

73

261

29

154

74

264

30

156

75

272

31

161

76

275

32

163

77

280

33

164

78

284

34

167

79

286

35

174

80

292

36

174

Si

301

37

178

82

314

38

i?8

83

328

39

179

84

335

40

181

85

338

4i

183

86

340

42

189

87

347

43

192

88

378

44

194

89

385

45

195

90

395

Total = 17,864.
Average = 198.48.

352 J. T. PATTERSON.

»

almost without exception degenerating individuals were found.
During the early period of the free larval stage, any given lot will
show great variation in the size of the larvae. To show this, all
of the individuals of three lots have been measured in the terms
of lines on the eye-piece micrometer scale (Table II.). In Lot I.
there were only thirty-two larvae. All but six of these would
have reached maturity. Lot II. contained 176 larvae, but at
least twenty of these were degenerating. Lot III. contained
257 larvae, and probably more than a hundred of them would have
degenerated eventually.

A series of sketches of these larvae is shown in Fig. 25, A to H.
The first four or five of these types would have developed to
maturity, but such larvae as those illustrated in F to H degenerate.
The most common types of degenerating embryos are the small
spherical or oval-shaped masses (G, H). In one extreme case
the lot of embryos consisted of about thirty of these masses,
together with only a single normal larva. Doubtless many other
similar masses had already degenerated.

It is difficult to assign any definite cause to the degeneration
of these embryos, although it probably has something to do with
nutrition. In some cases it seems to be due to the fact that the
division of the egg has been carried too far. Some of the pri-
mordia receive but few embryonic nuclei, and these are invariably
the first to degenerate in the polygerm. In other cases the
degeneration is apparently due to the lack of proper nutrition.
Most of the polygerms are early surrounded by the thick layer
of adipose tissue, upon which the early development of the
embryos depends. But other polygerms are almost if not entirely
barren of adipose cells, and it is an observed fact that the mor-
tality of embryos in such cases is exceedingly high. In Fig. 14
one of these cases is shown. This polygerm, which is devoid of
fat tissue, contains more than a hundred embryos, not more than
thirty or thirty-five of which have developed normally.

V. NUMBER AND SEX OF COPIDOSOMA PARASITES FOUND IN

GNORIMOSCHEMA.

The number of matured parasites developing in the Gnorimo-
schema larva has been determined in 162 cases. This has been

DEVELOPMENT OF COPIDOSOMA GELECHLE.

353

done by removing the carcass from the gall chamber a short
time before the emergence of the parasites, and enclosing it in a
small vial. After all of the parasites have emerged they are
killed by filling the vial with 80 per cent, alcohol, and then
counted under a binocular microscope. This procedure has the
advantage of eliminating the possibility of contamination from
other polyembryonic broods. Furthermore, the use of the bi-
nocular in counting enables one to distinguish readily the two
sexes. The strong sexual dimorphism in Copidosoma makes this
task rather easy. The females have the enlarged club-shaped,

TABLE IV.

TABLE SHOWING THE NUMBER OF PARASITES IN MALE BROODS.

Brood.

No. of Individuals.

Brood.

No. of Individuals.

I

41

32

178

2

53

33

179

3

6l

34

180

4

67

35

180

5

90

36

180

6

93

37

182

7

96

38

190

8

IOO

39

190

9

101

40

192

10

106

4i

199

it

107

42

199

12

H3

43

202

13

118

44

2O4

14

119

45

214

15

124

46

215

16

124

47

218

17

124

48

223

18

127

49

225

19

128

50

232

20

136

5i

233

21

137

52

236

22

138

53

236

23

139

54

245

24

142

55

247

25

147

56

272

26

152

57

277

2?

168

58

278

28

171

59

323

2Q

172

60

324

30

177

61

328

31

177

62

345

Total = 19,874.
Average = 175.32.

terminal segment of the antenna, and bright yellow legs, while
the males do not have the enlarged segment and the legs are of a

354 J- T- PATTERSON.

dark, more or less mottled color. One can therefore readily
detect a mixed brood under the microscope.

The 162 broods studied were taken at random from the field,
and therefore in all probability the data on numbers and sex
yielded by them represent the approximate sex ratio for the
species. These 162 broods contained a total of 31 ,001 individuals,
or an average of over 191 to the brood. Ninety of these, or 55.56
per cent., contained only female parasites, 62, or 38.27 per cent.,
contained only male parasites, and 10, or 6.17 per cent., con-
tained mixed broods of males and females.

There are therefore not only a larger number of female broods
than male, but the average number of individuals in the former
exceed that of the latter. Female broods average a little over
198 individuals to the brood (Table III.), while male broods
average only about 175 (Table IV.). The range in the number of
individuals in these broods (from 25 to 395 in the female, and
from 41 to 345 in the male) makes it evident that these averages
are of little significance, except, perhaps, to show that the
fertilized egg is slightly more prolific than the unfertilized egg.

Of the total number of individuals (31,001), 63.41 per cent, are
females and 36.59 per cent, males; but obviously the true sex
ratio can not be based on these figures. It must be determined
from the number of male and female broods. It would not be a
difficult matter to determine this ratio were it not for the un-
certainty of the origin of some broods. There is always the pos-
sibility in these insects that more than one parasitic egg has
been laid in the egg of the host, and hence the parasites which
later emerge may not constitute a true polyembryonic brood, but
in fact represent two or even more such broods. Under the
circumstances, the best that one can do is to determine approxi-
mately the sex ratio for the species. This can be done in the
following manner. If we assume, as all previous workers have
done, that each of the mixed broods is the product of at least two
eggs, then, in accordance with the law of probability, we can
determine the number of unmixed male and female broods, each
of which must also have been produced from two eggs. Worked
out on this basis, it is found that the ratio of females to males
is 106/76 or a sex ratio of approximately 3 : 2

DEVELOPMENT OF COPIDOSOMA GELECHLE. 355

This leads to a discussion of mixed broods, and especially to a
consideration of the question as to how such broods have come
into existence. The obvious explanation of their origin is the
one given above, viz., that they arise from two eggs. Marchal
and Silvestri, who have studied the development of polyembry-
onic insects, both offer this explanation. They support the con-
clusion by citing the fact that two (or more) parasitic eggs are
sometimes laid in the egg of the host. According to Marchal,
such eggs develop independently, each producing a distinct
polygerm and consequently a distinct brood. If the two eggs
are of the same sex potentiality, the individuals developing from
them will be either all females or all males, according to whether
or not the eggs are fertilized or unfertilized. The dual origin of
these double broods naturally elude detection in lots that have
emerged. But if one of the two eggs is unfertilized and the other
fertilized, the result will be a mixed brood, consisting of males and
females. This conclusion of Marchal and Silvestri is strongly
supported by the facts of polyembryonic development in the
armadillos, in which it has been conclusively demonstrated
(Fernandez, '09, Patterson, '13) that all of the embryos of a given
pregnancy are the product of a single egg. As a result, mixed
litters are never found in these mammals.

That mixed broods may arise from two eggs in Copidosoma
is supported by the fact that two polygerms are sometimes
found in a single Gnorimoschema larva. However, certain facts
concerning the condition of mixed broods in this species, make
it doubtful whether the origin of all such broods can be explained
in this obvious way. Careful dissections of something over a
hundred parasitized Gnorimoschema larvae have revealed only
two cases in which a single larva contained more than one
polygerm. Since 6.17 per cent, of all broods are mixed, and
since a similar number of unmixed broods would have a dual
origin, we should expect to find over 12 per cent, of all parasitized
larvae containing two polygerms, but instead, less than 2 per cent,
are found.

Another line of evidence which is not in harmony with the
view that mixed broods are always the product of two or more
eggs, is the great preponderance of females in certain lots. Of

356

J. T. PATTERSON.

the nine complete lots (Broods 2 to 10) listed in Table V., the
number of females in each case is greater than the number of
males. In some cases (Broods 3, 4, 5, 7, 8), this difference is not
so great but that the origin of each lot can be explained on the
assumption that two eggs have been deposited in the egg of the
host. But in Broods 2, 6, 9, and 10 the number of females in
excess of males is indeed striking, making it difficult to explain
the origin of such broods on the basis of two eggs.

In view of these facts, the writer is convinced that some other
explanation must be offered for the origin of certain mixed broods;
in fact, one involving the idea that a single fertilized egg may give
rise to a few males as well as a relatively large number of females.
This would be possible on the basis of the following assumption.

TABLE V.

TABLE SHOWING THE NUMBER OF PARASITES IN MIXED BROODS.

Brood.

No. of Individuals.

Females.

Males.

I*

89

20

69

2

162

153

9

3

172

92

80

4

207

126

81

5

216

176

40

6

235

223

12

7

241

161

80

8

300

235

65

9

304

292

12

10

337

3i6

21

Totals

2,26"?

1,704

460

Average

226.^5

I7O.4

46.9

* This brood is not complete, owing to the fact that some of the larvae and pupae
had been destroyed by a large dipterous larva.

If Copidosoma conforms to the general scheme for sex determin-
ation in insects, the females must have the 2 X chromosomes, and
males the single X chromosome. Ordinarily, during the process
of cleavage, all of the chromosomes in the fertilized egg divide
equally, so that all of the nuclei entering into the formation of
the embryos will carry the XX chromosomes, thus producing a
brood of females. But if during the early development of the
egg it should happen that the two X chromosomes in one or more
cleavages should not divide but separate, one going to each pole
of the spindle, each daughter nucleus would then receive a single

DEVELOPMENT OF COPIDOSOMA GELECHI^.

357

X chromosome. If such nuclei later divided in the typical
manner and gave rise to embryos, such embryos would be males.
One is encouraged to make this suggested explanation in the light
of Bridges' ('13) discovery of the non-disjunction of the sex
chromosomes in Drosophila. In Copidosoma the separation of the
sex chromosomes during cleavage would be a case of "somatic" or
"cleavage disjunction," while in Drosophila these chromosomes
fail to separate or "disjoin" in the reduction division of the egg.
In conclusion attention should be directed to the frequency of
Copidosoma in nature. At Woods Hole about twenty per cent.
of all Gnorimoschema larvae are infected with this parasite

TABLE VI.

TABLE SHOWING PERCENTAGE OF PARASITIZED CATERPILLARS IN THE GALLS OF

SOLIDAGO SEMPERVIRENS.

Number of Galls.

Date.

Parasitized by
Copidosoma.

Normal
Galls

Empty.

Parasitized by
Other Parasites.

9

7-29-12

7

2 0

O

33

8- S-I2

5

15 0

13

33

8-I7-I2

9

16

5

3

56

8- 8-12

7

26

10

13

29

8-1 2-1 2

8

16

o

o

141

8-25-12

20

56

33

32

14

7- 7-13

i

13

o

o

16

7-14-13

o

13

o

o

39

7-15-13

8

31

0

0

38

7-19-13

4

*

*

*

23

7-23-13

2

20

o

I

38

7-26-13

6

*

*

*

27

8- 5-13

4

*

*

*

24

8-25-13

4

17

3

0

18

6-15-14

2

14

2

o

19

6-18-14

3

16

O

o

43

6-22-14

19

19

3

2

40

6-24-14

9

20

10

I

20

7-16-14

i

19

o

0

24

7-30-14

o

21

o

3

25

7-23-15

3

12

7

3

18

7-26-15

5

II

o

2

66

7-30-15

25

37

2

2

35

8- 4-15

14

35

I

6

Totals. . 828 | 166

* Record incomplete.
Copidosoma.

About 20 per cent, of the caterpillars are parasitized by

(Table VI.). As may be seen from the table, the extent of
infection varies greatly in the lots of galls taken from different
regions (those collected on a given date are all from a single
locality). Plants which grow in exposed places, as along the

358 J. T. PATTERSON.

roadside or barren spots, carry a higher percentage of galls than
do those which are located in protected regions. Likewise,
the moth larvae from the galls of the former are more highly
parasitized.

SUMMARY.

1. Copidosoma gelechiae, which is a parasite in the Solidago
Gall Moth, Gnorimo schema salinaris, has but one generation a
year.

2. The egg of this parasite is probably laid during the month
of May.

3. The type of development in Copidosoma is polyembryonic.
The number of individuals average about 191 per brood.

4. In the youngest stages secured the process of division of the
egg into embryonic primordia is already in progress. The young
polygerm consists of two distinct regions: (i) An outer zone, or
nucleated membrane, containing the free polar . nuclei; (2) a
central region, containing the true embryonic nuclei.

5. The embryonic nuclei segregate into groups, which become
surrounded by a dense layer of granular protoplasm and form
the primordia of the multiple embryos.

6. During early growth the polygerm elongates into a cylin-
drical-shaped structure, which becomes broken up into several
spherical, primary masses by the formation of constrictions in
the nucleated membrane. Each primary mass receives several
of the primitive embryos.

7. The primary masses become broken up into secondary
masses by further constrictions of the nucleated membrane. At
the end of these divisions, each embryo is separated from the
others and is surrounded by an inner and an outer involucre —
the former derived from the granular protoplasm and the latter
from a portion of the nucleated membrane.

8. The interstices between these masses become filled with an
inter-embryonal substance derived from at least three sources:
elements from the nucleated membrane, leucocytes, and cells
from the adipose tissue, which usually is laid down in the form
of a thick layer on the outer surface of the polygerm. The
entire structure thus becomes a complex, which may be called the
polygermal mass.

DEVELOPMENT OF COPIDOSOMA GELECHI^E. 359

9. The dissociation of the inter-embryonal substance sets the
larvse free in the body cavity of the host. This occurs during the
last week of July.

10. Abortive or degenerating embryos are found throughout
the entire period covered by the polygerm and free larval stages.

1 1 . The free larvae destroy the entire contents of the caterpillar,
except the chitinous parts of the trachae, and leave only the
superficial layer of chitin of the body wall intact. The detritus
left in the larval chitin hardens to form thin-walled, oval chambers
in which the larvae lie and undergo pupation The superficial
layer of chitin also hardens, and the larval skin thus becomes
transformed into the typical mummified carcass, filled with the
parasitic pupae.

13. Pupation takes place during the first ten days of August
and lasts about a month.

14. The number of adult parasites emerging from the carcasses
varies from 25 to 395. There is a preponderance of females,
about 55 per cent, of all broods being females. Furthermore, the
average number of females emerging from a single carcass is
198 as compared with 175 for the males. Ten mixed broods of
males and females have been obtained. Some of these have
doubtless arisen from two or more eggs; but it is suggested that
such broods may also arise from a single fertilized egg, by a
process of disjunction of the sex chromosomes during the early
cleavage stages.

WOODS HOLE, MASS.,
August 12, 1915.

LITERATURE REFERENCES.
Bridges, C. B.

'13 Non-disjunction of the Sex Chromosomes of Drosophila. Jour, of Exp.

Zool., Vol. XV., pp. 587-606.
Fernandez, M.

'09 Beitrage zur Embryologie der Gurteltiere. Morpholog. Jahrb., Bd. 39,

PP- 302-333.
Howard, L. O.

'85 Description of North American Chalcididae. Bureau of Entomology

Bulletin.
Marchal, P.

'98 Un exemple de dissociation de 1'oeuf. La cycle de 1'Encyrtus fuscicollis.

C. R. Soc. Biol. Paris. T. 5. pp. 238-240.
Marchal, P.

'04 Recherches sur la Biologic et le Developpement des Hymenopteres Parasites.

360 J. T. PATTERSON.

I. La Polyembryonie Specifique ou Germinogonie. Arch, de Zool. Exper.
et Gen., Vol. II., pp. 257-335.
Martin, F.

'14 Zur Entvvicklungsgeschichte des polyembryonalen Chalcidiers Ageniaspis

(Encyrtus) fuscicollis. Zeit. f. Wiss. Zool., Bd. no, pp. 419-479.
Patterson, J. T.

'13 Polyembryonie Development in Tatusia novemcincta. Jour. Morph.,

Vol. 24, pp. 559-684.
Riley, C. V.

'69 The Solidago Gall Moth. First Annual Report of the State Entomologist

of Missouri, pp. 172-178.
Sacken, Baron Osten.

'63 Lasioptera Reared from a Gall on the Goldenrod. Proceed, of the Ent.

Soc. of Phil. Vol. i, pp. 368-370.
Silvestri, F.

'06 Contribuzioni alia conoscenza biologica degli Immenotteri Parassiti.

I. Biologia del Litomastix truncatellus Ann. d. Regia Scuola Superiore di
Agricoltura di Portici, Vol. VI., pp. 1-51.

Silvestri, F.

'08 Contribuzioni alia conoscenza biologica degli Immenotteri Parassiti.

II. Sviluppo dell'Ageniaspis fuscicollis. Ibid., Vol. VIII., pp. 1-27.
Wheeler, W. M.

'10 The Effects of Parasitic and other Kinds of Castration in Insects. Jour,
of Exper. Zool., Vol. 8, pp. 377~438.

362 J. T. PATTERSON.

DESCRIPTION OF PLATES.

PLATE I.

FIG. i. A typical gall of Gnorimoschema salinaris, Busck, situated at the base
of the stalk of the swamp goldenrod, Solidago sempervirens. X /4-

FIG. 2. Gall cut open to show the position of the mummified carcass of Gnori-
moschema. Natural size.

FIG. 3. Gall cut open and carcass removed to show the shape of cavity. Note
that the walls of the cavity are smooth and that the excrement from the caterpillar
is packed in the bottom of the cavity. Natural size.

FIG. 4. Mummified carcass from gall shown in Fig. 3. Natural size.

FIG. 5. Lepidopterous larva which is an inquiline in the gall of Gnorimoschema.
Note the irregular shape of the cavity which contains scattered trash and excrement.
Natural size.

FIG. 6. This gall shows an incomplete passage-way, lying just above the head
of the carcass. Normal size.

FIG. 7. Side view of a gall showing the orifice of the passage-way, closed by
silken plug. Natural size.

FIG. 8. Stalk of swamp goldenrod containing two galls. X %•

FIG. 9. Gall containing a non-parasitized caterpillar. Natural size.

FIG. 10. Gall containing a parasitized caterpillar. Natural size.

BIOLOGICAL BULLETIN, VOL XXIX.

PLATE I.

J. T. PATTERSON.

364 J- T. PATTERSON.

PLATE II.

FIG. n. Photomicrograph of a section of an irregular polygermal mass. X40.

FIG. 12. Photomicrograph of a single embryo from mass shown in next figure.
Xi8o.

FIG. 13. Photomicrograph of a longitudinal section of a flat, plate-like poly-
germal mass. X 40.

FIG. 14. Photomicrograph of a spherical polygermal mass which is barren of
adipose tissue. X 40.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE II.

J. T. PATTERSON.

366 J. T. PATTERSON.

PLATE III.

FIG. 15. Photomicrograph of the middle portion of longitudinal section of a
small caterpillar. A fat body containing a polygerm lies just below the intestine.

X44-

FIG. 1 6. Photomicrograph of a portion of a section of a polygermal mass which

was about to begin disintegration. X 44.

FIG. 17. Photomicrograph of a section of a polygermal mass undergoing dis-
sociation. X 44.

FIG. 1 8. Photomicrograph of a mass of free larvae from the body cavity of the
caterpillar. Note that each embryo is still surrounded by the involucres. X 44.

Reference Letters Used in Plates IV. -VI.

A.E., Abortive Embryo. I.S., Inter-embryonal Substance.

A.T., Adipose Tissue. N.M., Nucleated Membrane.
C., Clear space left by contraction of O.I., Outer Involucre.

Precipitated Material. P.D., Primary Division of polygerm.

E.N., Embryonic Nuclei. P.E., Primitive Embryo.

G.L., Granular Layer. P.M., Precipitated Material.
/./., Inner Involucre.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE III.

. •_••.•»«.. „" .-•--.-._ .

' • '• •

'•

16

15

I

17

18

J T. PATTERSON.

368 J. T. PATTERSON.

PLATE IV.

FIG. 19. A and B longitudinal sections of two polygerms lying close together
in the same caterpillar. These polygerms show an early phase of the segregation
of the embryonic nuclei to form the separate embryos. X 489.

FIG. 20. Section of an egg of Gnorimoschema which has started to develop
parthenogenetically. X 173.

FIG. 21. Longitudinal section of a polygerm showing the end phase of embryo
formation. X 480.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE IV.

J. T. PATTERSON.

37° J- T. PATTERSON.

PLATE V.

FIG. 22. One end of a longitudinal section of a polygerm showing three of the
twelve primary divisions into which it has been divided by constrictions of the
nucleated membrane. X 373.

FIG. 23. Section of a primary mass showing the process by which it is further
divided up into secondary masses by constrictions of the nucleated membrane.
X 508.

FIG. 24. Section of a single isolated, primary mass about at the close of its
division into single embryonic masses. X 257.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE V.

\»

^ • ,jj »!»*• - -«*«j«" - r~, — k *. j-» , .

• J;

i. .1'. ' J*. '..L^*T\ ". -.-.\ ^v«_

KS^a^S^^B

&&. x- ,^-r ---'?; rL^vr*:^

..

WM^

jpg;^ ,^-vJ^
^(:^:^|^^^.

ft-a^^l*, f,:^^"1 , I

^*^1

*'- ^sS:^

A{

J. T. PATTERSON.

23

G. Z.

^

•.r--. -• ,i.{ tr2^©
24

• -

o

372 J. T. PATTERSON.

PLATE VI.

FIG. 25. A to H, Series of sketches from Lot III of the free larvae listed in
Table III. This figure shows the great variation in size of the larvae from a single
caterpillar. They are all drawn to the same scale.

FIG. 26. Detailed drawing of a section of one of the embryos seen in Fig. 13.
It shows the relation of the inter-embryonal substance and involucres to the
embryo. X 187.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE VI.

J. T. PATTERSON.

DISTRIBUTION OF FOLLICULINA IN 1914.

E. A. ANDREWS.

The finding of vast hordes of the Stentor-\ike infusorian Follicu-
lina both in 1912 and 1913 throughout the whole extent of the
Severn River which is a brackish side branch of the Chesapeake
Bay, led to further examination in 1914 to see if this were a
phenomenon to be repeated annually or only a rare inroad of an
outside fauna into new territory.

In I9I31 FolHculina'was found in inconceivable numbers living
upon the leaves of the fresh water plants Elodea and Potamogeton,
which have taken possession of definite zones of shallow brackish
water along some fifty and more miles of extent of the river and
its side creeks. It was also found on Elodea in Whitehall River,
just to the north of the Severn.

In 1914 it was taken on Elodea from the head of the Magothy
River, August 13, and on floating Elodea in the mouth of the
Magothy, August 23, when it was also found living upon stunted
Elodea growing in the narrow inlet canal to the nearly shut off
side branch known as the Little Magothy. It was taken also
at Deep Creek, a side branch of the Magothy.

As the Magothy opens into the Chesapeake some seven miles
from the Severn, the distribution of Folliculina is quite extensive.
Moreover, in 1880 Ryder2 found Folliculina in great numbers
upon oyster shells in shallow water on the west coast of the
Chesapeake, and as he seems to have then been at St. Jerome,
St. Mary's County, which is sixty miles down the Bay from the
Severn, the distribution of Folliculina is known for side branches
of the Bay opening into it seventy miles apart, approximately.

It is to be expected then that exceedingly large areas of the
side waters of the Chesapeake may be inhabited by this little-
known protozoan, which in the mid-summer season adds greatly

1 See BIOL. BUI.L., XXVI., No. 4, April, 1914.

2 Am. Nat., 14, 1880.

373

374 E- A- ANDREWS.

to the plankton, or swimming fauna, as well as to the microscopic
life attached to the summer vegetation of these waters.

Its advent and departure in Chase's Creek, a branch of the
Severn, showed in 1914 even more suddenness than in 1913, while
its time of abundance was noticeably less though actual numbers
present were even more vast.

Though searched for from the middle of June, every few days,
Folliculina was found first on July 19, 1914. It then appeared
only here and there, not on every plant of Elodea and on very few
plants of Potamogeton. On the sprays of Elodea the Folliculina
showed on comparatively few leaves, like black soot stuck on
the leaves; both isolated individuals and aggregates occurred

i

It
*#

FIG. i. Leaf of Potamogeton showing scattered colonies of Folliculina. X3
diam. Photograph of preserved specimen.

but there were very few large aggregates covering half the surface
of a single leaf. Most leaves had none, some leaves had many
scattered individuals. On the stems there were noticeable
numbers of the small form of sac. The occurrence on leaves
seemed entirely arbitrary as if from settlements of swimmers: the
Folliculina was not now crowded toward the tips of the sprays
but scattered along many inches of the spray.

At the date of this first appearance, jellyfish had been common
for two weeks but the other conspicuous summer visitor to these
waters, the young menhaden now for the first time came along
the shores over the Elodea, which may be correlated with the

DISTRIBUTION OF FOLLICULINA IN 1914. 375

feeding of the menhaden upon plankton in which the free swim-
ming Folliculina may be included as possible food for the men-
haden.

At this date the Elodea had grown up to a height of twenty
inches and formed some flower stalks and buds at the surface,
so that there had been a long period in which suitable attachment
base for Folliculina was present but the Folliculina had been
absent.

July 21 the water after long drought was turbid from the
presence of plankton and the Folliculina had increased but little,
appearing as black spots on one out of several hundred sprays of
Elodea and one out of many thousands of Potamogeton sprays.
Only a few of the leaves on each inhabited spray had dense
aggregates, so that the question arises: why do the Folliculina

FIG. 2. Tip of leaf of Elodea covered with a colony of Folliculina. X IS diam.
Photograph of preserved specimen.

crowd together in these rare, isolated aggregates? When sprays

L

of these dates were put into aquaria they gave rise to free swim-
ming forms, thus showing that these early settlers need not
remain fixed but might contribute to additional distributions.

On July 27 Folliculina had become much more abundant upon
sprays of Elodea and Potamogeton; some of the free-floating frag-
ments on the surface appeared black with the accumulated

376

E. A. ANDREWS.

Folliculina. In the water also some free-swimming Folliculina
could be seen near the surface swimming all through the water
as well as close to floating plants.

Out in the Severn River a two-quart jar of water taken up at
random at the surface showed several free-swimming Folliculina;

FIG. 3. Photograph of a preserved colony that had been formed on surface
of the water in aquarium ; showing form of case and tube spirals as well as animal
rectracted within case. Enlarged 30 diameters.

three days later these had settled down on the side of the jar and
were in two groups, two individuals in one and five in the other,
so that at least seven were in the two quarts of surface water,
which would make an immense number for the entire river.

DISTRIBUTION OF FOLLICULINA IN

377

By August i much of the Elodea growing in the Elodea zone
along shore was black with aggregates of Folliculina. Free
swimmers were in the water of the creek in vast numbers : a quart
dipped from the surface at random showed in a white bowl from
fourteen to one hundred, by actual count, for each quart of water
from the surface. By drawing the bowl along the surface, the
Folliculina swimming free were concentrated till thousands in a
quart made it dark as if sprinkled with black pepper. Though
these free-swimming Folliculinas easily escape notice in the
greenish water turbid with plankton and sediment, they are
readily observed in calm water by an eye near the surface; and
standing in water five feet deep one may see them swimming

FIG. 4. Photograph of two young colonies of free swimmers that have just
settled on surface of water in aquarium and formed sacs but no tubes: one individual
on extreme left is still in motile form. Preserved specimen, X 20 diam.

rapidly in all directions, individually in straight and in curved
paths. Many deep down in the water were seen best by holding
a white object below them, but most of them were near the surface
where they congregated especially about any floating object as
fallen leaf or floating chip, seemingly influenced by its presence
so that they swam toward it.

While at this time the Folliculina continued to colonize the
new growths at tip of the Elodea as fast as it grew so that the

378

E. A. ANDREWS.

black aggregates crowded on the young leaves nearly to the tip
where only the newest leaves were as yet unoccupied ; by August
1 8 the extension of the Folliculina hosts had ceased. The tips
of the growing Elodea were now bare or free from Folliculina
back some twenty leaves from the tip and many of the old dwel-
lings on the lower leaves were deserted. These dense black
colonies on old leaves contained in fact but few living Folliculinas.

FIG. 5. Photograph of natural size sprays of Elodea preserved to show successive
phases of colonization in 1914. Spray on left has grown enough to form flower but
as yet but a very few .isolated individual Folliculina have settled upon it. The next
spray shows scattered tubes all along its length. The third spray shows dense
aggregations of colonies even up to the tips of the rapidly unfolding new leaves.
The fourth spray illustrates the subsidence in colonization: the new colonies no
longer cover the leaves at the tip of the spray but these grow more rapidly than the
new colonists occupy them and are left more nearly free from any Folliculinas.

By August 26 this falling off in the colonization and rapid dis-
appearance of Folliculina was most pronounced : the Elodea sprays
showed an abrupt transition from the lower leaves black from
dense population of tubes, for the most part empty, to the upper
leaves only sparsely inhabited with scattered individuals. Evi-
dently some sudden change had operated not only to check the
previously rapid spread of the Folliculinas onto new leaves but to

DISTRIBUTION OF FOLLICULINA IN 1914. 379

almost exterminate them. Yet many remained alive here and
there so that when large quantities of the Elodea were put into
aquaria many free swimmers escaped. Yet these after forming
new tubes on the surface of the water did not remain alive but
had all vanished September 5, though in such apparently normal
environment others had been kept two weeks in captivity earlier
in the season.

Thus while appearing after the middle of July and being extra-
ordinarily abundant in August, the Folliculina were all gone about
the end of August and no way was found of keeping them longer.
Their period of existence in accessible regions of the river was
scarcely six weeks.

In 1913 they appeared before the end of June and a few lingered
on to the first of September in nature and were kept in aquaria
in a warm room till the 2/th and a few till November 11.

In 1912 no live ones were found after September 8. This
enormous crowding of the waters with free-swimming Folliculina
and dense settlements of the case-making Folliculinas during
about a month, the last weeks of July and the first of August,
coincides with very high temperatures and abundance of micro-
scopic plankton in these waters but it is not at all evident either
why the Folliculinas should not come earlier, as they did in 1913,
or remain later as they did in 1913 and 1912.

The great rapidity of their colonization of large areas suggests
either very great immigration or else very rapid multiplication,
or combination of both. As all material searched in the daytime
in 1913 failed to show more than a few cases of multiplication,
most all the free-swimming forms being merely the case-making
forms again freed, material was collected at all times of the night
in 1914, but here again but few cases of division were observed.

Hence it seems unlikely that fission of a few immigrants actually
produced the vast numbers found on the leaves of plants, and it
is probable that very large numbers came into the river suddenly
from some outside source and these settling down, migrating out
again, and in some cases increasing by fission, gave rise to the
succession of dwellings covering the leaves for some two months.

The causes leading to the immigration as well as the causes of
rather sudden diminution of numbers and utter disappearance
remain entirely unknown.

380 E. A. ANDREWS.

The food of the case-inhabiting Folliculina being bacteria and
some larger forms of plankton, the disappearance of Folliculina
may well be associated with changes in food supply, in turn
brought about in connection with such changes as those of tem-
perature and salinity.

The motile forms take no food and may be enabled to settle
and to continue migration and multiplication only when feeding
conditions allow the sessile form to accumulate enough energy.

SUMMARY.

1. The vast swarms of swimming protozoans of the genus
Folliculina that were found to settle down over the aquatic plants
along the shores of side branches of the Chesapeake Bay in 1912
and 1913, came in even greater numbers in 1914, and it is there-
fore probable that this immigration and colonization is a regular
annual phenomenon.

2. The incursions of swimming Folliculina do not take place
as soon as the plants have grown enough to supply places for
attachment, and the departure or disappearance of the living
Folliculinas antedates the cessation of growth and final dying down
of the plants upon which they settle.

3. As far as evidence is available the numbers that crowd the
leaves arise more from immigration from without the area than
from division of animals that have already settled in the area.

4. The times of appearance and disappearance differ in suc-
cessive years.

5. It is suggested that conditions of food possibilities are
determining factors in these inroads into the brackish fauna.

6. The great number of free swimming forms makes them, for
the time being, an important factor in the plankton.

7. The crowding of the dwellings or cases on the leaves all along
the shores is a considerable element in the transformation of
matter which, arising from decay of organic materials, is trans-
formed into bacteria and other plankton organisms, which in
turn are eaten by Folliculina and enable them to secrete resisting
tubes and sacs which finally settle to the bottom of the river.

PHENOMENA OF ORIENTATION EXHIBITED BY

EPHEMERID/E.1

F. H. KRECKER.

It is a well-known fact that in alighting Ephemeridsc orient
positively to a breeze. I became interested in this reaction and
the observations made naturally lead to others on reactions to
gravity and to light, and to the results of a conflict between any
of these three stimuli.

The observations were made during the summer of 1915 at the
Lake Laboratory of Ohio State University at Cedar Point on
Lake Erie. Ephemeridse appear here in almost incredible
numbers. When a brood is at its height it is a very common
occurrence to find piles of the insects three or four feet square
and six to eight inches deep undenelectric lights. At a neighboring
amusement resort several carts were required each morning to
haul away the dead insects. The species with which the following
observations are especially concerned is Hexagenia variabilis.
The number, variety and arrangement of lights at the resort
presented favorably conditions for observing the reactions to light
of great numbers of individuals in what may be termed natural
surroundings. The equipment used for experiments with air
currents and gravity was simple and largely improvised. Never-
theless, since it is not primarily my purpose to measure intensity
of stimuli or rapidity of reaction, I believe the results obtained
have some interest and value.

REACTIONS TO A CURRENT OF AIR.

There was a question in my mind as to whether the positive
orientation of the Ephemeridee to a breeze is a response to the
breeze per se or whether other factors are concerned. In order
to test this I took a piece of glass tubing several inches long
and sent through it a weak but steady current of air so directed

1 Contribution from the Department of Zoology and Entomology, Ohio State
University, No. 43.

381

382 F. H. KRECKER.

as to strike the insects on the side of the body. They were
resting on boards placed horizontally. A few of them flew
away but most of them eventually faced the current. Individuals
placed on a rough surface, such as a wire screen, which afforded
a better foothold frequently tried to walk away. When facing
the current of air an individual would raise its long, slender
front pair of legs and extend them forward and upward at an
angle of about 40 degrees. When held in this way the legs re-
semble antennae and it is possible they have a sensory function.
However, cutting them off had no apparent effect on the reactions
here in question. The time required for the turning reaction
varied from an almost instantaneous response to two minutes.
In the majority of cases the response was gradual and occupied
from 30 seconds to one minute. The rapidity of reaction de-
pended upon a correlation between the strength of the breeze
and the part of the body it struck.

The influence of the area stimulated is shown in experiments
with the wings. The latter are large in proportion to the body
and meet over the back in a perpendicular position. They,
therefore, present quite a broad surface. When a current of
air of an intensity sufficient to blow the wings slightly to one
side was directed against them individuals would react in fifteen
to thirty seconds, whereas when this current was directed against
the thorax or the abdomen the response was slower, if indeed
any occurred. A stronger current directed against any of these
parts brought about a correspondingly more rapid reaction.

In another series of experiments a current of air was directed
from the posterior lengthwise of the body along the dorsal
surface of a number of individuals. The response in these cir-
cumstances was also an eventual facing about to the current.
A current of air striking an individual longitudinally along the
mid-dorsal surface is neutral so far as lateral directions are con-
cerned. In the cases here in question the current blew the wings
to one side or the other and then as before the insects turned
around toward the side on which the strain was exerted.

The experiments were repeated on a group of individuals from
which the wings had been removed. The results from a current
of air striking the insects on the side of the body were the same

PHENOMENA OF ORIENTATION EXHIBITED BY EPHEMERIDyE. 383

as before; the insects faced the current. However, when a
current was directed from the rear longitudinally along the dorsal
surface of the body the previous results were not repeated. In
some cases the insects crawled with the current and away from
the point of origin. In other cases they remained stationary
and took an attitude similar to that assumed when facing the
current. If the current became very strong they either attempted
to crawl away or they retained the attitude until blown off their
feet. When the current veered sufficiently to strike them on the
side they began to turn toward it.

In these experiments with air currents the first noticeable
response from the insects was an attempt to hold on to the surface
upon which they were resting. This they did by fastening their
claws firmly and even changing the position of the legs. When
the current became so strong as to make it difficult to remain
attached and especially when the body was blown over to one
side the insects began to change position, rather hesitatingly it
appeared, and to face about toward the direction from which the
current came. When an insect reached a position where it did
not seem to have difficulty in maintaining its hold it came to rest.
This usually meant that it was directly facing the current, al-
though sometimes it stopped at a point between a half and a com-
plete about face. A half about face could generally be made
complete by increasing the strength of the current.

When directly facing a current of air an individual is in the
optimum position for resistance; it presents the least surface
and the claws because of their backward curve have the maximum
effect in holding the body. On the other hand when an individual
stands sidewise to the current a greater surface is presented,
the claws are not in a relatively favorable position and attach-
ment is clearly more difficult. WTith regard to the more rapid
reactions which result when the current strikes the wings it
may be said that the proportionately great expanse of the wings
above the body's center of gravity gives them such a leverage
that the body is more easily tipped over, a strain is more quickly
felt and attachment more quickly made difficult. In those
cases in which a current struck wingless individuals from the
posterior there was practically no obstruction to the current

384 F. H. KRECKER.

and it consequently did not so easily cause strain or seriously
disturb the attachment and there was therefore no turning
reaction.

It would appear from the foregoing experiments that the
Ephemeridae do not change position under the stimulus of a
breeze until a strain is exerted on the organs of attachment.
That this does not merely mean that the response was delayed,
until a breeze of a given intensity developed is shown by the fact
that a comparatively weak breeze directed against the wings
alone had the same effect as was caused by a somewhat stronger
breeze against the thorax. There is, therefore, evidence, I
believe, for concluding that Ephemeridae do not orient positively
to a breeze because of sensations derived from the breeze per se
but that they react positively to tension exerted on the muscles

of attachment.

REACTIONS TO GRAVITY.

The position of Ephemeridae when resting upon a perpendicular
surface is negative with regard to the earth's surface and usually
approximately vertical to it, although variations as great as
45 degrees occur. On comparatively smooth surfaces the orien-
tation is more generally an approximation to the vertical,
whereas on surfaces such as a wire screen, which affords a good
foothold at any angle, variations from the vertical may occur
in 50 per cent, of the individuals concerned. Individuals picked
up by the wings and replaced head downward, if they are not so
disturbed as to fly away, will struggle to gain a foothold. The
position of the claws, which are adapted to a vertical position,
make attachment rather difficult. This difficulty is increased
by the fact that the long abdomen is thrown forward and down-
ward and thus tends to destroy equilibrium. On comparatively
smooth surfaces such as a planed board the insects rarely suc-
ceeded in maintaining their equilibrium long enough to gain a
footing. On a wire screen they were more often successful and
once they gained a footing and their equilibrium they retained
the new position. The picking up process caused so many of the
insects to fly away that other methods were tried. Several in-
dividuals were placed in a vertical position on a straw hat held
perpendicularly and then the hat was slowly revolved until the

PHENOMENA OF ORIENTATION EXHIBITED BY EPHEMERID/E. 385

insects were upside down. The overhanging abdomen disturbed
the equilibrium of some of them sufficiently to cause them to
lose their hold and fly off. The others retained their footing, in
some cases by changing the position of the legs, and remained in
the inverted position for ten to fifteen minutes which was as
long as they were watched.

In explanation of the position normally assumed on an upright
surface the evidence derived from the experiments seems to
indicate that the position taken is not a negative reaction to
gravity per se but that it is largely, if not entirely, due to the
character of the insect's means of attachment.

Results obtained from experiments performed to test the in-
fluence of a breeze upon the position of the insects on a perpen-
dicular surface support this view. A current of air was directed
against the side of individuals resting in the normal upright
position on a perpendicular surface. As they turned the current
was so directed as to bring them still further around. During
the process some of them could not retain their foothold and
flew off. The others turned completely around and faced directly
downward. They maintained the inverted position at least as
long as they were under observation, ten to fifteen minutes,
which length of time, in view of a constant coming and going
among those normally situated, seemed sufficient.

REACTION TO LIGHT.

The conclusions with regard to reactions of the Ephemeridse
to light are largely the result of observations made in the amuse-
ment resort already mentioned. The observations have to do
mostly with artificial light. The insects react negatively to bright
sunlight and seek the shade. They are strongly attracted to the
lighter colors of artificial light. In the resort there are a great
many electric lights of sixteen candle power intensity with colorless
glass bulbs. Many of them are attached in a horizontal position
to the sides of buildings in such a way that there is a perpendicular
surface either above or below them and frequently on all sides.

The reaction to these lights seems to be satisfied if the insects
can come to rest within a zone which begins approximately six
inches from the light and covers a radius extending outward for

F. H. KRECKER.

about twenty-four to thirty inches. When individuals enter
this optimum zone they alight, if a surface is available, and orient
themselves in such a way that the body is parallel with a radius
projecting from the light. After alighting the insects usually
remain at rest, although there may be a certain amount of
crawling toward a position nearer the center. This is more often
done by those nearer the outer limits of the zone. When the
insects are numerous they become arranged in rows consisting
of individuals either directly behind one another or slightly to
one side and they thus form a striking pattern of radiating lines.

-;-vV;-',

^ s \ i
^ \

^ i

' , / •
/ / ;

FIG. i.

The accompanying figures illustrate the positions assumed with
regard to lights in different positions and combinations.

The first figure illustrates the position assumed when the
surface extends about a light in all directions whether the plane
be horizontal or vertical. When any portion of the surface is
absent the pattern is of course interrupted to a corresponding
extent. The clear zone immediately surrounding the light was
approximately six inches wide. I shall call it the excitement
zone. Individuals that entered this zone became greatly
excited and fluttered about the light in a confused state. There
was no evidence to show that individuals at rest deliberately
entered the excitement- zone. Those immediately bordering on
it were rather restless and occasionally in crawling about some
were pushed into it and others on taking wing came within the
influence of the light.

PHENOMENA OF ORIENTATION EXHIBITED BY EPHEMERID/E. 387

The second figure shows lights arranged along the lower edge
of a perpendicular surface at intervals of twelve to fifteen inches.
About each light was the usual excitement zone and upward from
this extended the radiating lines of insects in the optimum zone.
As shown in the diagram these lines were rarely at an angle of
less than 35 degrees. This was due to the fact that below this
point the lines from neighboring lights conflicted and caused such
confusion among the insects as to obliterate regular alignment.
The greatest confusion occurred in the comparatively short space

\ \

\

\

/ /
' / / /

'

'

<

',', '/ ^ NO ''/ / '.'

/ ,V I

V N^ XX '//

FIG. 2.

between the lights where insects attempting to arrange them-
selves about one light constantly came into conflict with others
attracted to the neighboring light.

When the insects rested on a horizontal plane about a light
they faced it. The most striking feature connected with the
arrangement of the insects on a perpendicular surface was that
the individuals on opposite sides of a horizontal plane passing
through the center of a light had opposite ends of the body directed
toward the light. The insects below the plane or parallel with
it faced the light, whereas those that were above the plane were
turned away from the light. In other words all the insects?
except those parallel with the horizontal plane, approximated a
vertical position with the anterior end uppermost. Those above
the plane and with the posterior end directed toward the light
were apparently as well content as those below the plane and
facing the light.

The position of the insects on a horizontal surface shows that
other things being equal they face the light. It is reasonable
to conclude that their normal reaction to light is positive. The
negative position assumed on a perpendicular surface above a
light can be explained, in view of the air current and the inversion
experiments, as being due to the difficulty experienced in main-
taining a foothold in the inverted position.

388 F. H. KRECKER.

Some observations were also made on the relative influence of
white and colored lights. On the sides of one of the buildings
in the resort there was a succession of alternating white, red and
blue lights. The slightly yellowish white bulb attracted the
insects in greatest numbers. There was the usual excitement
zone and the regular alignment of those at rest. The number of
insects about the red and the blue bulbs was decidedly small and
as between the two lights about the same. These lights appeared
to have a quieting effect on the insects. The alignment was
similar to that described for white lights but there was no well-
defined excitement zone, in fact the insects crawled about the
bulbs without exhibiting markedly abnormal reactions.

OHIO STATE UNIVERSITY,
COLUMBUS, OHIO.

CELL MULTIPLICATION IN THE SUB-CUTICULA OF

DILEPIS SCOLECINA.1

DALTON G. PAXMAN.

INTRODUCTION.

The process of cell division in cestodes as compared with that
in other Metazoa is apparently quite abnormal. An examination
of cestode material at once reveals the fact that mitotic figures
are very rare, and that an explanation of the process of cell
division analogous to any of the common types is apparently
impossible. The opinion of the various workers in cestode cy-
tology, as to how cell division is taking place, varies greatly.
Some state that it occurs by mitosis, others by amitosis, while
it has been asserted that nuclei arise ' de novo ' from the cytoplasm.

Child ('07) noted the apparent infrequency or total absence
of any evidence of mitosis in Moniezia, even in regions where
rapid growth was taking place. He says, "If my observations
are correct, amitosis is the more common method of division in
the generative cycle, except during the period of maturation and
early cleavage. And in the somatic cells of the adult body it
appears to be the usual method at all times."

Young ('08), working with Cysticercus pisiformis describes
what he calls the "de novo" formation of cells. He observed
irregular masses of coarsely granular cytoplasm lying in the
meshes of the parenchyma network. These masses contain
numerous small deep staining granules scattered haphazard
through the mass. Shortly succeeding the formation of these
granules, a nuclear membrane is formed around them; the
newly formed nucleus, together with a small mass of cytoplasm,
becomes partly constricted from the parent mass; and the
daughter cell has been formed."

Further, he says: "I believe that the nucleus in these forms
is not a morphological, but a physiological entity; that the

1 A thesis presented to the graduate faculty of the University of North Dakota
in partial fulfilment of the requirements for a master's degree.

389

39O DALTON G. PAXMAN.

nuclear granules are fundamentally the same as the remaining
protoplasm of the cell, but are differentiated therefrom under
physiological conditions which we do not at present understand;
that the granules are perhaps reserve material stored up in the
nucleus for future use, the entire cell body being thus occasionally
converted into a nucleus; and the nucleus varies in structure
from time to time in response to the varing physiological demands
made upon it. ... Further if my interpretation of my ob-
servations be correct, then distinction between germ and somatic
plasm is obviously impossible, a special vehicle for the trans-
ference of hereditary qualities is entirely wanting; such qualities
must be transmitted by the undifferentiated protoplasm; cell
lineage is manifestly lacking; a mosaic theory is plainly untenable;
and the fate of any given embryonic element — whether it shall
form parenchyma, muscle, nerve, etc. — must be determined by
physiological causes alone."

Richards (1911), working with Moniezia, does not agree with
Child. He says (p. 158): "I have after diligent search upon
carefully prepared material been unable to establish a series of
stages in the autoconstriction and subsequent division of the
nucleus and cell body by amitosis. Considering the evidence as
set forth, it seems to the writer that one is forced to the con-
clusion that mitosis is the method by which pre-oogonia and
cleavage divisions are accomplished."

Mary T. Harman ('13, p. 223) states: "My observations have
not shown that amitosis does not take place in Taenia or Moniezia,
but they have showrn no condition which cannot be as readily
explained as the result of mitotic as of amitotic division."

MATERIALS AND PROCEDURE.

The form I worked with was Dilepis scolecina parasitic in the
small intestine of the double-crested cormorant (Phalocrocorax
dilophus). These birds are found abundantly near the shores
and on the islands of Devils Lake, North Dakota.

Immediately after the bird was killed, the cestodes were
removed from the intestine and placed in fixing solution. Flem-
ming's solution and cestode mixture were the fixatives used.
Flemming's solution blackened the tissue so that the results

CELL MULTIPLICATION IN SUB-CUTICULA. 39!

from it were not satisfactory. The cestode mixture, however,
gave excellent results.

The stains used were the following: Heidenhain's iron-alum-
haematoxylin without counterstain ; safranin counterstained with
light green; thionin counterstained with acid fuchsin; methyl
green counterstained with acid fuchsin; and safranin counter-
stained with water blue.

OBSERVATIONS.

I began my study of cell multiplication in cestodes without any
previous knowledge of what had been done in the field of cestode
cytology. Moreover, I completed the study of my material
and drew my conclusions before I read any of the literature on
the subject.

I have confined my study of cell multiplication in Dilepis to
the sub-cuticula. In this tissue I have searched in vain for a
single clear case of mitosis or amitosis. Moreover, in order to
be certain I had not overlooked any, I counted 10,000 resting
nuclei in the sub-cuticula of the neck regions of ten worms with
the same result. Certainly active growth must have been taking
place in this region, but it could not be accounted for by mitotic
or amitotic division.

I have, however, observed numerous places in this region in
which active cell multiplication was apparently taking place.
Here multinucleate cells, such as shown in Fig. i, have been
observed. In addition to these, large protoplasmic masses were
present, which varied in size from that of a single cell to that of
perhaps fifty cells massed together. Fig. 2 shows a typical mass.
These masses stain rather deeply with nuclear stains, and contain
from one to five nuclei.

These masses are found abundantly in the neck region of every
worm I examined, and occur, although less frequently, in the
body region.

By reference to any of these figures it is seen at once that the
mass of cytoplasm is out of proportion to the mass of the nuclei.
Moreover, I have observed numerous lobes and occasionally
even entire masses in which I was unable to find any trace of a
distinct nucleus. Fig. 7 shows a lobe,1 i, and Fig. 6 a mass of

1 At focal levels other than that shown in the figure the lobe was seen to be
continuous with nucleate masses.

DALTON G. PAXMAN.

protoplasm, h, in which no well-defined nucleus is present. How-
ever, in this latter case the mass is so close to a nucleate mass that
I cannot say positively that it is not continuous with it.

By closely examining the nuclei present in these masses, I
find that the nuclear membranes are very indistinct in many
cases. Fig. 2 shows a mass in which the nuclei have indistinct
membranes. Also one of the nuclei, c, has a somewhat less
distinct membrane than the other, b. And this latter membrane
is in turn less distinct than the membranes of the nuclei in the
cell syncytium above it.

Moreover, a large number of nuclei have been seen which
lack membranes completely. The nucleus consisted of a
"nucleolus" or "karyosome" surrounded by a clear zone. Figs.

3, 4, and 5 show "karyosomes" which lack membranes. As
Child and Young have already suggested, I believe this " nu-
cleolus" represents the chromatin material of the nucleus.

By observing the protoplasm under high magnification (2,000
diameters) it is seen that the protoplasmic strands contain many
dark staining granules of various sizes and shapes. Some of
these granules were as large as the "nucleoli " of the complete
nuclei ; others, however, were so small as to be scarcely discernible.
Fig. 4 shows a mass which contains a number of varying-sized
granules. Fig. 5 shows a mass which contains a number of
varying-sized granules one of which, g, is becoming surrounded
by a clear zone.

The protoplasmic masses apparently arise by the outgrowth
of protoplasm from certain cells of the syncytium. Figs. 2, 3,

4, and 6, show masses of protoplasm continuous with the syncytial
cells around them. In Fig. 6, the developing mass is very small
and contains no definite nucleus. In Figs. 2, 3, and 4, the masses
are very large and contain from one to five complete nuclei.
A large number of masses have been observed varying in size
between these extremes. The nuclear membranes of the nuclei
in the cells from which these masses are developing, contain very
small, irregular granules which stain darkly like the granules
in the cytoplasm. I have insufficient evidence for or against
Young's view of the "de novo" origin of these granules. The
chromatin granules may arise "de novo" in the cytoplasm and

CELL MULTIPLICATION IN SUB-CUTICULA. 393

develop to complete nuclei in situ. Young bases his theory of
the independent origin of granules from a cytogenic protoplasmic
mass upon the following facts:

1. The occurrence of masses of granular protoplasm lacking
any evident nuclei.

2. The occurrence of isolated " nucleoli" of varying size from
34 to i micron in diameter, which are usually found in the above
mentioned masses of protoplasm but occasionally lie free in the
parenchyma strands.

I believe, however, that these facts may be equally well ac-
counted for by assuming the extrusion of chromidia from a
mother nucleus. Masses of granular protoplasm without any
evident nuclei, which occur but rarely may be explained as having
been severed from parent masses after impregnation \vith
chromidia. The occurrence of isolated "nucleoli" can be ac-
counted for just as well by assuming the migration of chromidia
from the nuclei along the strands of the cytoplasmic network,
as by the assumption of their development from the protoplasm
in situ.

Young, in a later paper ('13) dealing with gametogenesis, in
Tcenia pisiformis says (p. 375): "I believe that new nuclei arise
either from chromidial extrusions from old nuclei, or 'de novo' in
the cytoplasm. . . . The structure of the nucleus — a loose
collection of chromatin bodies without a membrane — renders
the extrusion of chromidia an easy matter. After their extrusion
new chromatin is added and that part of the cell containing them
is constricted off, to give rise in its turn to other cells. ... It is
obviously impossible to say, however, whether any chromatin
granule in the cytoplasm is a chromidial extrusion or a 'de novo'
formation."

Since I have seen these very small granules, all of about the
same size, present in the nuclear membrane as though impeded
by it in their exit, along the strands of the protoplasmic network,
from the nucleus to the cytoplasm, I believe that these granules
are extruded from the mother nucleus. Moreover, since I have
observed granules of various shapes and sizes, many of the
larger ones appearing to be composed of three or four smaller
ones partly united, and since I have often seen a number of

394 DALTON G. PAXMAN.

granules clustered together, I believe that the larger granules
are the result of the union of many smaller ones. Thus, I believe
that the small particles of chromatin or "chromidia" are ex-
truded from the mother nucleus. Then these "chromidia"
unite here and there throughout the protoplasm to form larger
granules or "karyosomes" which become surrounded by a clear
zone. Finally the nuclear membrane is formed, producing a
daughter nucleus. When a number of nuclei have been formed
multinucleate cells are the result. Since the tissue is always a
cell syncytium, constrictions of the cytoplasm around a nucleus
finish the production of a daughter cell. Thus one mother cell
may produce a large number of daughter cells.

COMPARISON WITH T^ENIA PISIFORMIS.

In order to compare the process of cell multiplication in Dilepis
with that in other cestodes, Dr. Young has permitted me to
examine his slides of Tcenia pisiformis, and Cysticercus pisiformis.
Here I have identified the protoplasmic masses in both the adult
and the larva. These also contain nuclei in the various stages
of formation from chromidia to complete nuclei. The young
larvae show large numbers of protoplasmic masses developing
in the cell syncytium. In the older larvae the masses often show
four or five nuclei developing membranes at the same time.

DISCUSSION.

Cell multiplication by means of protoplasmic masses and the
development of nuclei from chromidia, has, so far as I am aware,
never been observed heretofore in Metazoa by anyone except
Young. He has described the process as it occurs in Cysticercus
pisiformis (Young, '08) and has noted it in some other cestodes
(Young, '10) although his interpretation varies slightly from my
own. I have, in the present paper given an account of it as it
occurs in the sub-cuticula of Dilepis scolecina. It is true that
chromidia have been observed in certain Metazoa, but no account
of their functioning in the reproduction of the cell has ever been
given previous to Young's paper on the " Histogenesis of Cy-
sticercus pisiformis."

If cells are actually developing from protoplasmic masses in

CELL MULTIPLICATION IN SUB-CUTICULA. 395

the manner described, we have here an exceptional method of
cell multiplication, unlike anything previously described in
Metazoa.1 Moreover, if future research supports this view, the
present theories of the role of the nucleus in heredity will have to
be greatly modified at least with respect to cestodes.

As Young has previously suggested, the explanation of such a
method of cell multiplication as this may rest on the fact that the
cestode is highly degenerate in most characteristics due to its
long period of parasitism. In the development of cells from
protoplasmic masses the nucleus passes through a cycle in which
occur stages resembling nuclei of lower forms. The protoplasmic
mass with its diffused nuclei in the form of chromidia is com-
parable to a cell of the Bacteria or of the Myxophycese. In
certain Protozoa also, as noted by many observers, the nuclear
material at certain periods diffuses throughout the cytoplasm
in the form of chromidia which may give origin to secondary
nuclei, and these in turn to gametes. It is possible that the
cestode nucleus has lost the power of mitotic division, accom-
panying the somatic degeneration of the worm due to parasitism.
Richards, Harman, and others have shown, however, that we
still find cell division taking place by mitosis in the sex cells and

developing embryos.

CONCLUSIONS.

I have made the following conclusions in regard to cell multi-

•

plication in the sub-cuticula of Dilepis scolecina.

1. After a careful examination, and after counting 10,000 of
the nuclei in this region, I conclude that the growth of the sub-
cuticula cannot be accounted for by mitotic or amitotic division.

2. Tissue growth is taking place rapidly in this region by the
development of protoplasmic masses. My reasons for believing
this are the following:

A. The nuclei in the multinucleate cells are frequently seen
crowded together as if they had developed in protoplasmic masses.

B. In the protoplasmic masses the quantity of cytoplasm is
out of proportion to the number of complete nuclei present.

C. Developing nuclei have been actually observed in the cyto-
plasm. The different stages of nuclear formation are shown by
the following:

1 A similar process was suggested long ago by Schleiden and Schwann.

396 DALTON G. PAXMAN.

(a) The chromidia, or diffused nucleus.

(&) The irregular chromatin granules formed by the union of
numerous chromidia and surrounded by a clear zone.

(c) The nuclear membranes of the nuclei in the masses vary
considerably from delicate, scarcely discernible membranes to
heavy, well developed ones.

D. These masses appear to arise by the simultaneous growth
of cytoplasm and chromidial extrusions from the nuclei of certain
cells.

3. The degenerate character of the nucleus is perhaps the result
of the parasitic habit of the cestode.

I wish here to express my sincere thanks to Dr. R. T. Young,
whose valuable criticisms and suggestions made this work
possible. I also wish to express my indebtedness to Dr. B. H.
Ransom for identifying my material.

LITERATURE.
Child, C. M.

'<>7a Studies on the Relation Between Amitosis and Mitosis. BIOL. BULL.,

Vol. XII., pp. 89-114.
'o7b Ibid., Vol. XII., pp. 175-224.
'070 Ibid., Vol. XIII., pp. 138-160.
'O7d Ibid.. Vol. XIII., pp. 165-184.
'10 Ibid., Vol. XVIII. , pp. 109-119.
'01 Ibid., Vol. XXI., pp. 280-296.
Harman, Mary T.

'13 Method of Cell-Division in the Sex Cells of Taenia teniseformis. Journ.

Morphol., Vol. XXIV., pp. 205-242.
Richards, A.

'n The Method of Cell Division in the Development of the Female Sex Organs,

of Moniezia. BIOL. BULL., Vol. XX., pp. 123-179.
Young, R. T.

'08 The Histogenesis of Cysticercus pisiformis. Zool. Jahrb. (Anat. und

Ont.), Vol. XXVI., pp. 183-254.
'10 The Somatic Nuclei of Certain Cestodes. Archiv. fur Zellforschung, pp.

140-164.

'13 The Histogenesis of the Reproductive Organs of Taenia pisiformis. Zool.
Jahrb. (Anat. und Ont.), Vol. XXXV., pp. 355-419.

398

DALTON G. PAXMAN.

FIG. i.

FIG. 2.

FIG. 3.

FIG. 4.

FIG. 5.

FIG. 6.
present.

FIG. 7.

FIG. 8.
nucleus, j.

FIG. 9.

EXPLANATION OF PLATE.

Multinucleate cell, a.

Nuclei with indistinct membranes, b and c.

Nuclei, d and e, lacking nuclear membranes.

Chromatin granules, /, in the cytoplasm.

Large chromatin granule, g, in cytoplasm.

A developing protoplasmic mass, h, in which no definite nucleus is

A lobe, i, of a protoplasmic mass in which no definite nucleus is present.
A large protoplasmic mass in the body region which contains only one

Protoplasmic masses, k, developing in the body region.

BIOLOGICAL BULLETIN, VOL. XXIX.

PLATE I.

--d

Fiq.f,

tf. 5 .

D. G. PAXMAN

I (TO

LIBRARY