*••

THE

BIOLOGICAL BULLETIN

PUBLISHED BY

THE MARINE BIOLOGICAL LABORATORY

Editorial Board

GARY N. CALKINS, Columbia University FRANK R. LlLLIE, University of Chicago

E. G. CONKLIN, Princeton University CARL R. MOORE, University of Chicago

E. N. HARVEY, Princeton University GEORGE T. MOORE, Missouri Botanical Garden

SELIG HECHT, Columbia University T. H. MORGAN, California Institute of Technology

M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University

H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University

E. E. JUST, Howard University EDMUND B. WILSON, Columbia University

ALFRED C. REDFLELD, Harvard University
Managing Editor

VOLUME LX

FEBRUARY TO JUNE, 1931

Printed and Issued by

LANCASTER PRESS, Inc.

PRINCE & LEMON STS.

LANCASTER, PA.

11

THE BIOLOGICAL BULLETIN is issued six times a year. Single
numbers, $1.75. Subscription per volume (3 numbers), $4.50.

Subscriptions and other matter should be addressed to the
Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa.
Agent for Great Britain: \Yheldon & Wesley, Limited, 2, 3 and
4 Arthur Street, New Oxford Street, London, \V.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, 240 Longwood Avenue, Boston, Mass.

In tercel October 10, 1902, at Lancaster, Pa., as second-class matter under

Act of Congress of July 16, 1894.

LANCASTER PRESS, INC.
LANCASTER, PA.

CONTENTS

NO. 1. FKl'.klARY, 1931

PAGE

WASTENEYS, HARDOLPH

The Second Reynold A. Spaeth Memorial Lecture. Protein
Synthesis 1

HELFF, O. M.

Studies on Amphibian Metamorphosis. IX. Integumentary
Specificity and Dermal Plicae Formation in the Anuran, Rana
pipiens 11

WHITAKER, D. M.

On the Conduction of the Cortical Change at Fertilization in
the Starfish Egg 23

LYNCH, RUTH STOCKING, and HELEN BERENICE SMITH

A Study of the Effects of Modifications of the Culture Medium
upon Length of Life and Fecundity in a Rotifer, Proales
sordida, with special relation to their Heritability 30

NABRIT, S. MILTON

The Role of the Basal Plate of the Tail in Regeneration in the
Tail-Fins of Fishes (Fundulus and Carassius) 60

PHILPOTT, CHARLES H.

Relative Resistance of Fourteen Species of Protozoa to the
Action of Crotalus atrox and Cobra Venoms 64

HARVEY, E. NEWTON

A Determination of the Tension at the Surface of Eggs of the
Annelid, Chaetopterus 67

LUCRE, BALDUIN

The Effect of Certain Narcotics (Urethanes) on Permeability

of Living Cells to Water 72

SMITH, VERNON D. E., and C. M. JACKSON

The Changes during Desiccation and Rehydration in the Body
and Organs of the Leopard Frog (Rana pipiens) 80

No. 2. APRIL, 1931
JACOBS, M. H., and ARTHUR K. PARPART

Osmotic Properties of the Erythrocyte. II. The influence of
pH, temperature, and oxygen tension on hemolysis by hypo-
tonic solutions . 95

in

3783?

iv CONTEXTS

PAGE

KROPP, BENJAMIN

The Pigment of Velella spirans and Fiona marina 120

JULL, MORLEY A.

The Sex Ratio in the Domestic Fowl in Relation to Antecedent
Egg Production and Inbreeding 124

HOWARD, EVELYX

The Effect of Fatty Acid Buffer Systems on the Apparent
Viscosity of the Arbacia Egg, with Especial Reference to the
Question of Cell Permeability to Ions 132

STEWART, DOROTHY R.

The Permeability of the Arbacia Egg to Non-Electrolytes.. . 152

STEWART, DOROTHY R.

The Permeability of the Arbacia Egg to Ammonium Salts. . 171

STUXKARD, H. VY., and V. YY. DTXIHUE

Notes on Trematodes from a Long Island Durk with De-
scription of a Xew Species 1 7<>

TYLER, ALBERT

The Production of Normal Embryos by Artificial Partheno-
genesis in the Echiuroid, I rechis 187

•

No. 3. JUXE, 1931

GERARD, R. YY., and I. S. FALK

Observations on the Metabolism of Sarcina lutea. I.... 213

GERARD, R. \Y.

Observations on the Metabolism of Sarcina lutea. II 227

TANG, PEI-SUXG

The Oxygen Tension-Oxygen Consumption Curve of l"n-
fertilized Arbacia Eggs 242

GERARD, R. YY.

Oxygen Diffusion into Cells 245

SMITH, DIETRICH C.

The Effect of Temperature Changes upon tin- Pulsations of
Isolated Scale Melanophores of Fundulus heteroclitus. . . 269

LILLIE, RALIMI S.

Influence of Cyanide and Lack of Oxygen on the Activation of
Starfish Eggs by Acid, Heat and Hypertonic Sea-Water. .

BUCIIAXAX, J. \\II.I.I\M

Modification of the Rate of Oxygen Consumption by Changes
in Oxygen Concentration in Solutions of Different Osmotic
Pressure . • • 309

CONTENTS v

PAGE

HALL, S. R.

Observations on Euglena leucops, sp. MOV., a Parasite of
Stenostonium, with special reference to Nuclear Division 327

MILLER, HELEN MAR

Alternation of Generations in the Rotifer Lecane inermis
Bryce. I. Life histories of the sexual and non-sexual gener-
ations 345

GELLHORN, E.

Permeability and Fatigue in Muscle and its Bearing on the
Problem of Ion Antagonism 382

GELLHORN, E.

The Recovery Contracture in Muscle; a New General Salt
Effect 397

GATENBY, J. BRONTE

Note on Dr. Jan Hirschler's Paper on Spindle Bridges, and the
De Novo Origin of Moth Spermatocyte Golgi Bodies 409

INDEX . 416

Vol. LX, No. 1 February, 1931

THE

BIOLOGICAL BULLETIN

PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY

EDITOR'S NOTE:

Due to a printer's error, lines 30-31, p. 322 of the article by
E. N. Harvey, entitled, " Biological Aspects of Ultrasonic Waves,
A General Survey' in the December issue, were incorrectly
printed. The sentence should read as follows:

" . . . In one, for increasing pressure without dissolving more
gas, a pyrex test-tube is sealed by deKhotinsky cement to a brass
chamber connected with a pressure gage, a petcock and a screw-
piston for increasing the pressure (Fig. 6A)."

In the higher organisms which, unlike the plants, have not the power
to synthesize amino acids, material for the synthesis of the proteins
required for growth and maintenance must be ingested in the form of
proteins or amino acids. The proteins, before final absorption, must
be broken down in the digestive tract into the constituent amino acids
by means of enzymes, and these amino acids are built up in the body
into the infinite variety of proteins characteristic of the various organs
and tissues.

Now it is not difficult to conceive the possibility of building up this
immense variety of different proteins from the twenty or so different
amino acids presented to it. But the mechanism by which the differen-
tiation is achieved is more difficult to imagine, and provides a fascinating
field for speculation. It is possible, of course, but improbable, that a

1 Delivered at the Marine Biological Laboratory, Woods Hole on August 19,
1930.

1
1

Vol. LX, No. 1 February, 1931

THE

BIOLOGICAL BULLETIN

PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY

THE SECOND REYNOLD A. SPAETH MEMORIAL

LECTURE *

PROTEIN SYNTHESIS

HARDOLPH WASTENEYS
DEPARTMENT OF BIOCHEMISTRY, UNIVERSITY OF TORONTO, TORONTO, CANADA

It is not necessary to elaborate on the importance of an understand-
ing of the mechanisms involved in the enzymatic synthesis of proteins,
because it is perhaps the most fundamental characteristic of living mat-
ter, just as protein itself is the most characteristic constituent of proto-
plasm.

We know something of the intermediate and a good deal of the
final products of degradation of the protein molecule, though we know
little or nothing of the exact structure of the molecule itself.

In the higher organisms which, unlike the plants, have not the power
to synthesize amino acids, material for the synthesis of the proteins
required for growth and maintenance must be ingested in the form of
proteins or amino acids. The proteins, before final absorption, must
be broken down in the digestive tract into the constituent amino acids
by means of enzymes, and these amino acids are built up in the body
into the infinite variety of proteins characteristic of the various organs
and tissues.

Now it is not difficult to conceive the possibility of building up this
immense variety of different proteins from the twenty or so different
amino acids presented to it. But the mechanism by which the differen-
tiation is achieved is more difficult to imagine, and provides a fascinating
field for speculation. It is possible, of course, but improbable, that a

1 Delivered at the Marine Biological Laboratory, Woods Hole on August 19,
1930.

-
1

2 HARDOLPH WASTENEYS

specific enzyme performs each synthesis. Experiments which I shall
describe this evening provide at least a possible alternative.

The first part of the problem to be attacked was, of course, the
synthesis of protein from its degradation products by means of enzymes.
It was then necessary to prove that a real synthesis had been achieved.
The investigation and control of the conditions concerned in the syn-
thesis was next attempted, and finally the possibility was investigated
of producing from one substrate, with a single enzyme, proteins of
different composition.

Those of our own experiments to which I shall refer were performed
in collaboration chiefly with Dr. H. Borsook, now of the California
Institute of Technology, and I shall always look back with pleasure to
the many hours which we spent together in these investigations.

1 should explain, perhaps, afc the outset that we were unable to
obtain synthesis from anything less complicated than peptic digests of
various proteins, although it is possible to obtain synthesis from such
substrates with several enzymes. The experiments do not, therefore,
reproduce an actual biological synthetic process. Investigations into
the mechanism of peptic synthesis and its control are, however, at least
very suggestive as to the probable mechanism of the biological synthesis.

First let me say a few words about the theoretical aspects of en-
zymic synthesis. In applying the law of mass action to hydrolysis in
general, it is usual to write :

A + HZO ^products
e.g. ester -(- H.20 *=*• alcohol (-(- acid

and the equilibrium equation in this case becomes

(alcohol) (acid)

±

(ester) (water) K2

In such a system it is necessary, in order to obtain synthesis, to
diminish the concentration of water as far as possible, and it has been
assumed that the synthetic processes taking place in living cells include
a very efficient mechanism or mechanisms for accomplishing this, such
as surface condensation or imbibition by colloids. In our first experi-
ments with protein synthesis in vitro the attempt was made- to produce
such surface condensation by means of the artificial cells used by T. B.
Robertson and E. Newton Harvey. Robertson, you remember, used
them to illustrate the mechanics of cell division and Harvey to demon-
strate permeability. We were not successful in obtaining synthesis
by this means alone but we returned to a similar procedure in later
experiments, with interesting results which 1 will try to describe towards
the end of this lecture.

PROTEIN SYNTHESIS

To return to the question of concentration of water. In the alcohol-
acid-ester system the molecular concentrations of water and of the other
reacting components are of somewhat the same order of magnitude, but
in protein systems, at least /'/; vitro, the molecular concentration of
water, even in very concentrated solutions, must be enormously greater
than that of the other reacting components; and if we had to consider
concentrations in the same terms as for the alcohol-acid-ester system,
it would be impossible to attain syntheses in vitro because the relatively
large molecular concentration of water would always send the reaction
in the direction of hydrolysis. This is one reason why, under ordinary
conditions of protein hydrolysis, the reaction is always complete, i.e., the
equilibrium position appears to be, unlike the ester system, at 100 per
cent hydrolysis.

But in the hydrolysis of proteins and also, to a lesser extent, of
fats, we have another factor, which alters the picture and minimizes the
relative importance of the molecular concentration of water. This
factor is the formation from one molecule of substrate (protein or fat),
of a number of molecules of hydrolysate (amino acids or glycerine and
fatty acids, as the case may be).

Hydrolyses of this type are in a special category, because the forma-
tion of many molecules from one molecule leads to other consequences,
which were, as far as I know, first pointed out by Benjamin Moore.
I will very briefly review his suggestions.

Moore deduced on thermodynamical grounds that the conditions of
equilibrium for the reaction A ^± B may be expressed by the equation

P2 = KP, ,

where P2 and Px are terms proportional to the osmotic pressures of A
and B respectively.

When, however, one molecule decomposes into two, i.e., A ^ 25, as
in the hydrolysis of an ester, the equilibrium reaction may be expressed
by the equation

No term is inserted for the water which takes part in the reaction
because its molar concentration remains practically constant.

For such simple reactions as A ^ B the ratio of the concentrations
at equilibrium of A and B is a constant, and is independent of the
absolute concentration, as the expression P2 = =KP1 indicates. Conse-
quently dilution or concentration of the system does not alter the pro-
portions of the components at equilibrium, but where one molecule
decomposes into more than one molecule, the equilibrium ratio is not

4 HARDOLPH WASTENEYS

constant and is not independent of the absolute concentrations. For
example, if in the equilibrium reaction P.,-- =KPt, the concentration is
doubled,

then Pj becomes 2Pt
and Po" becomes 2P.,~,

but since P., is the -\/ of the left-hand term.

P2 becomes V 2P22 or \/ 2 X P2

/.r., while Px has been doubled in concentration, P2 has been only in-
creased \/2 or 1.4 times, so that under these conditions, when the con-
centration of substrate is doubled, the concentration of hydrolysate is
less than doubled at equilibrium. In other words, the degree of hy-
drolysis is less, or in still other words, the tendency towards synthesis
is greater. Conversely, on dilution, P, decreases in less ratio than
Pj, that is, the degree of hydrolysis increases. This effect, in the case
of substances like esters and disaccharides, is not very marked ; it is
more marked in the case of simple fats where, since the fat molecule
decomposes into four molecules, the equilibrium reaction becomes

p 4 _ - ]-p

r ., • - J\ r l ,

and it is of course very marked in the case of proteins, where the value
of the exponent is much higher.

The consequences of this relationship can be briefly stated as fol-
lows : In reactions in which one molecule is decomposed into a number
of molecules the position of the equilibrium point depends on the abso-
lute concentration of the solution. Tn concentrated solutions of such
systems these reactions are incomplete, — there is a tendency towards
reversion or synthesis. In dilute solutions these reactions tend to be
complete and irreversible, i.e., the tendency towards hydrolysis is great.

These effects of concentration and dilution apply with especial force
in the case of proteins and of polysaccharides. They are independent
of the effect of concentration on the molecular concentration of the
water itself (which, as we have seen, cannot be practically reduced for
proteins in vitro to proportions which would drive the reaction in the
direction of synthesis), and they have two consequences: First they
enable us to easily attain complete hydrolysis of proteins in dilute solu-
tions, and second, they enable us to attain synthesis with great ease in
concentrated solutions, eliminating the necessity for mechanisms for
reducing the molecular concentration of water to dimensions commen-
surate with the molecular concentration of other components. Since

PROTEIN SYNTHESIS 5

presumably the same considerations apply in vivo, we no longer need
to postulate special biological mechanisms for effecting syntheses.

It can also be deduced on thermodynamical grounds, as Moore
pointed out, that an increase in temperature should favour synthesis and
a decrease should favour hydrolysis. The method for the synthesis of
protein in a peptic digest of protein is then very simple: It consists
simply in concentrating the digest, raising the temperature and adding
pepsin. The [H+] must also be controlled. Synthesis begins imme-
diately and equilibrium is attained in a few hours. Under optimal con-
ditions a yield of 50 per cent of protein has been obtained.

Since 1886, when the first synthesis of protein-like material was
carried out by Danilewski, many workers have obtained synthesized
protein. Sawjalow in 1901 gave to the protein precipitate the name
plastein, and most of these protein precipitates have since been known
by that name.

The first unequivocal proof that the precipitate obtained is a syn-
thetic product was furnished in 1911 by Henriques and Gjaldbak, who
demonstrated that plastein formation is accompanied by a decrease in
free animo-nitrogen, the converse of the common method of following
protein hydrolysis by estimating the increase in free amino-nitrogen.
We confirmed these results and it can be shown that the free COOH
groups also decrease in a digest on plastein formation, and, as Rona has
shown, exactly parallel the reduction in free amino groups.

The plasteins may be hydrolyzed by pepsin in the same manner as
native proteins, yielding proteoses and peptones. Plastein, like native
protein, is precipitated from its solution by trichloroacetic acid, which
will not precipitate gelatin nor the proteoses, so that its complexity
must be of the order of native protein. It gives, of course, all the
usual protein reactions. Perhaps the best proof of synthesis is ob-
tained from a fractional analysis which we made of digests where
different amounts of plastein had been synthesized. We found that
all fractions of the digest contribute to the formation of plastein.

There is no doubt then that these plasteins represent synthesized
protein-like material. Their identity has been doubted by many because
they apparently differ in physical properties (more especially in their
insolubility) from the native proteins or, particularly, the proteins from
which the digest is prepared. As a matter of fact their insolubility
appears to be due to a denaturation which results from the conditions
in which they are formed. Albumin and other soluble proteins under
similar conditions, that is, in the presence of a concentrated digest at the
optimum pH of 4.0, are rendered as insoluble as plastein.

We have then the first part of the simple problem completed. We

6 HARDOLPH WASTENEYS

have synthesized protein and we have satisfied ourselves that the
product is indeed protein and that it is synthesized from all the principal
components, i.e., that the proteose, peptone and sub-peptone constituents
all contribute to the rebuilding of a protein or protein-like substance.

The next part of the problem is the investigation of the conditions
under which the synthesis is obtained. This has yielded results of the
greatest interest to us though I can discuss them only very briefly in
this lecture.

In studying the effects of concentration on synthesis, we found the
following relations. The limiting concentration below which no syn-
thesis could be made to occur corresponds to the equivalent of about
eight per cent of protein. Synthesis takes place at all concentrations
greater than 8 per cent. \Yith concentrations higher than 8 per cent
there is a straight line relationship between the amount of protein
synthesized and the concentration of material in solution : synthesis is
in simple inverse proportion to the dilution. At high concentrations,
however, that is, at concentrations which correspond to more than 40
per cent of protein, there is a rapid falling off, until at still higher
concentrations no synthesis is obtained. We have no explanation for
this falling off, but a similar phenomenon was observed by Armstrong
and Gosney in the enzymatic synthesis of fats.

These results on the effect of concentration confirm Moore's pre-
diction that the important factor in the enzymatic synthesis of proteins
is the concentration of material in solution, and not the relative con-
centration of water. It can also be shown that as he predicted, the same
holds true for hydrolysis : the tendency for hydrolysis diminishes as the
concentration of protein increases, and in concentrated solutions of
proteins, hydrolysis is incomplete no matter how much enzyme is used.

A study of the effect of the addition of proteins in each case shows
that we are dealing with true equilibria in these reactions. Our experi-
ments demonstrate that if previously synthesized protein is added to
the system, synthesis is inhibited to an extent directly proportional to
the amount of added protein, which also shows that the plastein takes
part in the equilibrium as if it were in solution. Conversely it can be
shown that addition of products to a digest in which hydrolysis of
protein is going on inhibits hydrolysis to an extent proportional to the
amount of products added.

1 mentioned earlier that Moore had deduced on thermodynamical
grounds that in reactions of the type of protein synthesis and hydrolysis,
the effect of temperature on the equilibrium position is such that in-
crease of temperature favours reversibility, i.e.. increase in temperature
favours synthesis and decrease favours hydrolysis.

PROTEIN SYNTHESIS /

We were able to confirm this deduction by a very striking experi-
ment. A solution in which hydrolysis was proceeding at a temperature
between 20°-24° was adjusted to the optimum pH for synthesis. It
was then divided into three portions. One was maintained at 20°
and showed the further change in the direction of hydrolysis to be
expected under the circumstances. The second portion was raised to
65° C. In ten minutes plastein began to form and in 30 minutes con-
siderable protein had been synthesized. A third portion was boiled
to destroy the enzyme and then maintained at 65° C. for 8 hours
without the occurrence of any change except the faint coagulation
produced on boiling.

In this experiment then we have a solution in which hydrolysis is
proceeding at 20° C., and simply by raising the temperature we are
able to induce synthesis. We find that the equilibrium amount of
protein synthesized increases with increasing temperature up to 72° C.,
at which temperature the enzyme is destroyed ; that is, the effect of
increase of temperature is to move the equilibrium point more and
more over to the protein side.

So far I have made hardly any reference to the effect of hydrogen
ion concentration on peptic synthesis. As you could guess, it is a most
important factor. Practically all the previous work on protein synthesis
was performed before the importance of the [H+] in enzyme reactions
had been realized, and before methods for its exact determination had
been developed. The only determination, that of Henriques and Gjald-
bak, had placed the optimum for synthesis with the optimum for hy-
drolysis at a pH of 1.5. On many occasions we determined the optimum
for synthesis to be at a much more alkaline reaction, viz., pH 4.0. It has
been shown by Oda that this optimum holds, regardless of the acid
employed, and that it is unaffected by the presence of a great variety of
salts.

We have tried to determine the mechanism by which the [H+] con-
trols the extent of peptic synthesis. The following interesting fact
was discovered: If we construct a curve, plotting the yields obtained
at different [H+], and plotting the yields obtained as per cent with the
yield at the optimum pH as 100 per cent, the curve resembles in form
and in slope either the primary dissociation curve of a dibasic acid or
the undissociated residue of an amphoteric electrolyte. It seems proba-
ble from this resemblance, or correspondence, — because the correspond-
ence is remarkably exact, — that the amount of protein synthesized at
any given [H+] is dependent on the degree of ionization of .some com-
ponent of the system. It can hardly be the ionization of the enzyme
which is concerned, because, as Northrop has shown, pepsin exists as

8 HARDOLPH WASTENEYS

a monovalent anion between pH 1-7, and changes in pH within that
range do not aflfect the degree of dissociation. It might be an effect
on the autodestruction of the enzyme, but at least within the range of
pH 1-4 there is no autodestruction under the experimental conditions.
One can only suppose that the result is to be interpreted as showing the
relation between synthesis and ionization of the digest itself ; but
titration curves of the digest show that the ionization of the digest as
a whole is not represented by these curves, and one is forced to the
conclusion that since there is this remarkable correspondence between
the experimental values and a theoretical dissociation curve, and since
we have ruled out other possibilities, the protein synthesized at any
given [H+] is dependent on the degree of ionization of one component of
the digest or rather, of the substrate.

Another respect in which peptic synthesis behaves rather differently
from other enzyme reactions is in the effect of varying enzyme con-
centration on the equilibrium position. It lias been found by other
observers, and confirmed by the writer, that the equilibrium amounts
of protein synthesized are increased with increasing amounts of pepsin.
This is, of course, an exception to the general law of enzyme action
and catalysis, which always assumes that the equilibrium position is not
affected by the concentration of the enzyme. The explanation of this
effect of varying concentrations of enzyme is suggested by another
series of experiments which we carried out on the effect of the presence
of emulsions in the system on peptic synthesis. These are the experi-
ments I referred to earlier as resembling the experiments where the
substrate was concentrated at the surface of artificial cells. We found
that certain emulsions, such as those formed when chloroform, benzene.
benzaldehyde, etc., are shaken up with the digest, accelerated the at-
tainment of equilibrium. Moreover, in the presence of some emul-
sions. Mich as bcn/ene, toluene, benzoic acid and benzaldehyde, synthesis
is effected even in the absence of enzyme.

Now we may consider that pepsin, being a colloid or in colloidal
association, owes its action, in part at least, to properties resembling
those possessed by these emulsifying agents. This action occurs by
virtue of the presence of the interphase surface. The action of emulsi-
fying agents further resembles that of pepsin in that increasing amounts,
under certain conditions, also augment the yield of protein synthesized
at equilibrium. This effect of concentration of enzyme or emulsifying
agent on the equilibrium is consistent with thermodynamical conceptions,
Miice the addition of an emulsifying agent means an increase in the
interphase surface, and this increase in surface means in turn an increase
in MII fan energy to the system because of the increased amount of

PROTEIN SYNTHESIS

surface energy available. Thus, increasing the concentration of pepsin
should have the same effect as that obtained when energy in any other
form is added to the system, as, for example, by increasing temperature.
We have already seen that increasing the temperature up to the limit
of destruction of the enzyme has a similar effect. On this basis we
can therefore explain the departure from the classical conception that
the concentration of enzyme should have no effect on the amount
synthesized at equilibrium.

There is another point of similarity in the action of pepsin and
of emulsions. T. B. Robertson, in 1926, discussed the possibility that
concentration of the reactive groups of amino acids by orientation at
lipoid surfaces might account for the completeness of enzymatic syn-
thesis in living tissues. As we have just seen, in certain cases synthesis
can be effected even without enzyme by the emulsifying agent alone if
it is an accelerator of synthesis. Some of these emulsifying agents,
e.g.. benzene and toluol, are chemically inert compounds. They can
hardly enter into chemical combinations with digest constituents, and
they are readily separable again from the synthesized protein. Their
effect is almost certainly due to a specific orientation of the components
of the digest which participate in the synthesis. That this action is
specific is shown by the fact that the chemical and physical properties
of the proteins synthesized vary with the means employed in synthesis.
The base-combining properties, the ratios of free amino and free car-
boxyl groups to total nitrogen, the hydrolysability by pepsin, all show
marked differences which are reproducible.

The solubilities also show very marked differences, e.g., proteins
synthesized by pepsin with or without benzaldehyde are soluble only at
alkalinities greater than pH 7.7; at 7.6 the proteins are precipitated
almost completely.

Proteins synthesized with benzaldehyde in the absence of pepsin are
not precipitated until the reaction reaches pH 6.0, and the benzene-syn-
thesized protein is not precipitated until the pH reaches 5.0. At more
acid reactions these proteins are also completely precipitated.

These results indicate that, presumably through variations in the
orientation of the participating components of the digest, proteins of
different chemical composition are synthesized. They suggest also a
possible explanation for the specificity of the action of the enzymes
themselves.

We have been able to work out other interesting suggestions from
studies of the conditions of peptic synthesis, but time will not permit
me to discuss these results.

It is unfortunate that we have had very little success with protein

1" HARDOLPH WASTENEYS

synthesis from digests less complex than peptic digests, but we still
hope to obtain synthesis from simpler substrates, and are continuing
our studies.

Peptic synthesis, therefore, is at best only a possible model of the
biological synthesis of proteins. However, it provides many inter-
esting suggestions as to the mechanism of the biological process, and
not the least important of these is the last discussed, namely, the sugges-
tion that specific orientation on the colloidal interphases provided by the
proteolytic enzymes and their associated colloids may best account for
the ability of the organism to synthesize its many varieties of proteins
from a common substrate.

STUDIES ON AMPHIBIAN METAMORPHOSIS

IX. INTEGUMENTARY SPECIFICITY AND DERMAL PLIC.E FORMATION

IN THE ANURAN, RANA PIPIENS

O. M. HELFF

(From the Department of Zoology, State University of loiva, and the Department
of Biology, University College, New York University)

INTRODUCTION

The dermal plicae or folds of newly-metamorphosed Rana pipiens, as
is well known, constitute a striking and characteristic development for
this species. Although minor folds have been described, the most dis-
tinct plicae are those of the dorso-lateral trunk and lateral upper jaw re-
gions. The former consists of two folds, one on each side of the trunk,
running from the dorsal-posterior border of the eye back to a point just
lateral to the urostyle prominence. The latter likewise consists of two
folds, one on each side of the upper jaw, running from a point just an-
terior and ventral to the external nares posterior on a level just ventral
to the tympanic membrane to a point just dorsal to the origin of the
fore-limb. The folds are usually about 0.5 mm. in width at this time
and are white in color in sharp contrast to the black or dark-greenish
coloration of the adjacent integument. Histologically, the folds are seen
to be composed of integument having a much thickened stratum spongi-
osum in which are found mucous and unusually large poison glands. It
is the concentration and abnormal size of these glands in this particular
region which result in the formation of an actual ridge or fold as viewed
externally.

The development of the dermal plicae during metamorphosis may be
roughly estimated by external observation. In the normal larva, just
prior to the onset of involution, no indications of plicae formation are
visible. During the early stages of transformation, however, broad
though very indistinct bands of integument are noticeable in the dermal
plicee regions. These are at first of a dull, grayish-white shade. Later,
however, towards the close of metamorphosis, they become lighter in
color, as the bands appear to narrow, and more sharply demarcated from
the adjacent integument. Following larval transformation, the plicae
become increasingly whiter and very distinct, after which they undergo
relatively little change during the early growth period of the young frog.

11

O. M. HELFF

The embryology of the mucous and poison glands as they occur in
amphibian integument has interested many workers and given rise to
numerous conflicting results and interpretations. Practically all work-
ers, however, agree that the glands have an ectodermal origin, being de-
rived from the Malpighian layer of the epidermis. The point of differ-
ence seems to lie chiefly in whether or not the two types of glands have a
common origin or arise from separate sources. Separate developments
have been described by Schultz (1889) and Nirenstein (1908) in Sala-
mandra, by Heidenhain (1893) and Nicoglu (1893) in Triton, by En-
gleman (1872) and Seeck (1891) in the frog, and by Dawson (1920)
in Nccturus. Common origins, which in some cases have regarded the
mucous and poison glands as simply different developmental stages of
one and the same gland, have been described by Ancel (1902) in Sala-
mandra, by Fano (1903) in Triton, by Esterly (1904) in Plcthodon, by
Wenig (1913) and Calmels (1883) in the toad, and by Szczesny (1867),
Junius (1896), and Weiss (1908 and 1915) in frogs.

Although the present work has indicated certain points regarding the
embryology of the acinous glands as they occur in Rana pipicns, the main
object in view was to determine to just what extent various integuments
of the larva are specific for glandular formation as found in dermal
plicae regions. Histologically. all integument of the larva, prior to the
onset of metamorphosis, is similar in that there is an absence of acinous
glands or of any indication of their future development. The question
naturally arises, therefore, whether or not the integument of the dermal
plicae regions is hereditarily specific for the development of the glandu-
lar picture as it is found during larval involution. Conversely, the pos-
sibility suggested itself that all integument might be so stimulated to
unusual glandular development, if located in the specific regions char-
acterized by dermal plicae formation during metamorphosis. The above
points were subjected to analysis by making suitable integumentary
transplantations. In general, these consisted of the autoplastic trans-
plantation of larval integument from the dorso-lateral and lateral dermal
plicae regions to the back or belly, and vice versa. The development or
non-development of dermal plicae tissue in such transplants during larval
involution could then be studied and information relating to integu-
mentary specificity in this respect obtained.

The writer wishes to express his appreciation for the facilities ten-
dered him during the summer of 1929 by the Iowa Lakeside Laboratory,
Mil ford, Iowa, where the operations described in the present paper were

made.

DERMAL PLIC/E FORMATION

MATERIALS AND METHODS

13

The stock used for all operations consisted of large Rana pipiens
tadpoles collected in the vicinity of Spirit Lake, Iowa, during the months
of June and July, 1929. Prior to the onset of metamorphosis the larvae
of this particular strain of Rana pipiens grow to the unusual size of
110-115 mm. total length. When collected and operated on, however,
they averaged about 80 nun. in length, with hind limbs 8 to 12 mm. long.
The selection of larvae in a stage at least three weeks prior to the earliest
signs of metamorphosis was essential, since preliminary sectioning of
integument had shown that the acinous glands begin their development
at about the beginning of the metamorphic period.

The various integumentary transplantations were made as follows:
The larva was first anaesthetized in a 0.05 per cent aqueous solution of
chloretone. The approximate regions of the integument in which der-
mal plicae development later occurs were next located and a rectangular
piece of skin, large enough to include a small amount of the adjacent
integument above and below the dermal plicae region, was cut and re-
moved. The fact that all integument, including that of the dermal
plicae regions, is of the same macroscopic appearance, necessitates a
knowledge of the approximate level of integument which will later
develop into dermal plicae, in order that the skin graft will include all
of the potential plicae integument. A similar sized piece of integument
was now removed from either the mid-dorsal or mid-ventral region of
the same larva and transplanted to the wound area left subsequent to

DP

LP

VI

TEXT FIG. A

the removal of the dermal plicae skin graft. The latter was now trans-
planted to the area exposed by the removal of the back or belly integu-
ment (text fig. A}. Following this, the larva was placed in shallow
water so as to expose the two grafted areas to the air, thereby hastening

14 O. M. HELFF

the adhesion of the transplants. Following adhesion of the transplants,
the larva was immersed in an individual aquarium and allowed to re-
cover from the effects of the anaesthetic. Subsequent treatment in-
cluded feeding with Spirogyra to induce growth and daily observations
to record the onset of metamorphosis and the development of macro-
scopic signs of dermal plicae formation in the various transplants during
larval involution. Representative skin transplants were removed at
various stages of metamorphosis in order to determine the development
or non-development of glandular structures.

RESULTS
Normal Histological Development of Dermal Plica: Structures

Although the normal development of the acinous glands has been pic-
tured and described by many authors in a variety of amphibians, rela-
tively few have dealt with the formation of these glands as they occur
in the dermal plica- and none, with the possible exception of Massie
(1894), has dealt with the conditions as found in Rana piplens. It is
assumed that Massie probably worked with Rana pipiens, although the
species of frog tadpole used was not specifically stated. The exact mode
of the various structural differentiations in relation to the results of skin
transplantation as described in the present paper was not so important
to determine as was a knowledge of the histological picture of the dermal
plicae typical of various stages of larval involution.

For convenience, three stages of glandular development in dermal
plicae regions corresponding with three stages of larval involution will
be briefly described. These descriptions apply especially to the dorso-
lateral dermal plicae.

Stage 1. Larrcc i^lthin One Week of Onset of Metamorphosis. — The
larvae are in all respects typical tadpoles. There are no macroscopic
signs of dermal plicae formation. Histologically. all layers of the in-
tegument appear normal as regards thickness and pigmentation. A
few, widely-scattered, oval-shaped, embryonic gland " nests " are to be
found partly embedded in the lower layers of the epidermis and in-
vaginated into the stratum spongiosum. The "nests" appear to be
composed of large, oval-shaped nuclei and cells having large nuclei with
a minimum of cytoplasm. Mitotic figures are present in some of the
cells. The "nests" all appear to resemble one another cytologically.
It is clear that the component cells have been derived from the Mal-
pighian layer of the epidermis, probably for the most part from the
nuclei.

DERMAL PLIC;E FORMATION is

Stage 2. Lame in Early Stage of Metamorphosis. — The visible ex-
ternal metamorphic changes of the larva include growth of the hind-
limbs and slight atrophy of the tail. Otherwise, the larva is still tad-
pole-like in most respects. Externally, the dermal plicae are indicated
by broad, indistinct bands of a grayish coloration. Histologically, the
epidermis is unchanged in thickness but appears to be less pigmented.
The stratum spongiosum is considerably thickened to accommodate the
glands. The poison glands are fairly well developed and of a good
size. The muscular and fibrous layers of the latter are present. Prac-
tically all have ducts developed. The inner epithelial layer is composed
of cuboidal-shaped cells, while large granules occupy the central cavity
of the gland. The glands are definitely poison glands. Small mucous
glands are also found, some of which have acquired ducts. In addition,
glandular " nests " are also found in the usual location, the inference
being that they are developing mucous glands. The adjacent back in-
tegument now possesses a few, widely scattered gland " nests," which
appear smaller than those described in Stage 1 dermal plicae.

Stage 3. Larva near End of Metamorphosis. — The animals are typi-
cally frog-like, although the tympanic membranes are not as yet fully
developed. The tail has atrophied to a small stump. The dermal plicae
appear narrower, distinctly marked off from the adjacent integument
and grayish-white in color. Histologically, the epidermis has thickened
and contains more layers than that of the adjacent back integument.
The stratum spongiosum is especially thickened, while the melanophores
in this layer and the epidermis appear to be considerably less evident.
The integument as seen in cross-section is markedly convex along its
epidermal surface. The glands are much larger than hitherto described.
The poison glands show considerable variation in size, while the epi-
thelial cells lining the base of the gland cavity are usually columnar in
shape. The mucous glands are more numerous and larger than in Stage
2, while glandular " nests " may still be found, especially towards the
edges of the dermal plicae.

Further developments, following Stage 3, include the general whiten-
ing of the plicae as viewed externally, while histologically the glands
(especially the poison glands) increase still further in size and tend to
crowd together to form the arrangement typical of the fully developed
dermal plicae.

Reciprocal Transplantation of Dermal Plica Integument with that of

the Back and Belly

Series A. — This series consisted of reciprocal, dorso-lateral dermal
plicae and back skin transplantations (DP and DI, text fig. A). Fifty-

16 O. M. HELFF

seven larvae were operated on in all. During metamorphosis, it was
soon evident that typical macroscopic signs of dermal plicae formation
were present in the grafts transplanted to the back. If the graft had
been transplanted in the same orientation as it had previously occupied
in relation to the anterio-posterior axis of the body, the dermal plicae
developed, parallelling the normal ones of the animal. If, however, the
graft had been rotated 90° during transplantation, then the dermal plicae
developed at right angles to the normal ones of the animal. The nar-
rowing and whitening of the dermal plicae band of integument progressed
at a uniform rate identical with the normal formation of the dorso-

TEXT FIG. B

lateral dermal plicae on the same animal (DC, text fig. B). It was also
observed that the general pigmentary changes occurring in the transplant
were characteristic of the region of integument from which the graft
had been secured. The back skin grafts transplanted to dermal plicae
regions underwent macroscopic changes in coloration, including the de-
velopment of spots, and corresponding in all details to the pigmentary
changes of the integument of the back. There were no external indica-
tions of dermal plicae formation, not even where the edges of the grafts
were in contact with the developing normal dorso-lateral dermal plicae
(BG, text fig. B). Following the complete resorption of the tail, the
metamorphosed animals generally lived for about three weeks, at which
time they died due, no doubt, to lack of proper food. Representative
skin transplants were removed at this time and sectioned for histological
study.

The microscopic appearance of dermal plica- integumentary grafts
transplanted to the back is represented in Fig. 1, Plate 1. It will be
noted that typical glandular development and tissue thickening has oc-
curred, corresponding in all details to the condition typical of normal
dorso-latcral dermal plica? at this stage of the animal's life. The various
histological features arc representative of a somewhat later stage, as
previously described for Stage 3 of dermal plicae development. The
epidermal and stratum spongiosum layers are especially thickened, giv-

DERMAL PLIC;E FORMATION 1 7

ing the external surface of the dermal plica* integument a decidedly
convex shape. The poison glands are larger than those described for
Stage 3 and more closely crowded together. Usually five or six are
found across the width of the plicae. The mucous glands are relatively
much smaller and fewer in number and occupy only the upper regions of
the stratum spongiosum. The microscopic appearance of the back skin
grafts transplanted to dorso-lateral dermal plicae regions is represented
in Fig. 2, Plate 1. Again the various histological structures correspond
to the conditions typical of normal back integument at this time. The
epidermis is normal in thickness, while the stratum spongiosum is but
moderately thickened to accommodate the glandular development. Re-
garding the latter, the mucous glands appear to predominate both in size
and number. In fact, the mucous glands are larger and more numerous
at this time than is true of normal dorso-lateral dermal plicae. Only a
few, small, widely-scattered poison glands are present, which have prob-
ably just begun their development (PG, Fig. 2, Plate 1). It should be
emphasized, however, that this same histological picture is typical of
normal back integument at this time.

Scries B. — Series B consisted of reciprocal, dorso-lateral dermal
plicae and belly skin transplantations (DP and VI, text fig. A). Ten
larvae were operated on in all. The dorso-lateral dermal plicae integu-
mentary grafts developed normal dermal plicae during metamorphosis
which, together with the characteristic pigmentation pattern of the ad-
jacent transplanted back skin, presented a striking contrast to the sur-
rounding white belly skin (Fig. 5, Plate 1). Histologically, the same
glandular and integumentary conditions were found as already described
for this type of graft when transplanted to the back (see Series A,
above). The belly skin transplants were typically pure white in color
when transplanted to the dorso-lateral dermal plicae regions. This lack
of melanophore pigmentation and spotting was maintained during sub-
sequent larval involution and likewise presented a strong contrast as
compared with the darkly pigmented adjacent skin of the back and side
(Fig. 6, Plate 1). The histological appearance of such grafts, at a time
when the normal dermal plicae were well formed, is represented in Fig.
3, Plate 1. Compared with back skin grafts transplanted to this region,
the belly skin transplants were somewhat thinner as a whole, while the
mucous glands were correspondingly smaller though more numerous.
The poison glands were very small and scarce. In general, however, it
may be said that the microscopic appearance was identical with the nor-
mal integument of the belly at this time.

Series C. — This series consisted of reciprocal, lateral (upper jaw)
dermal plicae and back skin transplantations (LP and DI, text fig. A}.

2

18 O. M. HELFF

In all, twenty larvae were operated on. The development of typical
macroscopic signs of dermal plicae formation were found to occur dur-
ing involution in the transplants made to the back. The appearance was
quite similar to the dorso-lateral transplants to the back as illustrated in
text fig. B, except that the whitening of the integument was slightly de-
layed and the width of the plicae bands somewhat narrower. This, how-
ever, is typical of normal lateral plicae development. The back skin
grafts transplanted to lateral dermal plicae regions, like the back skin
transplants of Series A, failed to exhibit visible macroscopic signs of
plicae formation during larval involution. Histologically, the appear-
ance of these transplants was also similar to that described for the back
skin transplants of Series A and pictured in Fig. 2, Plate 1. The micro-
scopic appearance of the lateral dermal plicae transplants on the back
presented several differences as compared with the dorso-lateral dermal
plicae transplants described in Series A and pictured in Fig. 1, Plate 1.
Figure 4, Plate 1. represents a typical section taken through one of these
transplants at the usual time, — some three weeks following the comple-
tion of larval involution. Perhaps the greatest difference observed was
in the relatively greater abundance and size of the mucous glands. The
poison glands, on the other hand, were smaller and less numerous than
in the dorso-lateral dermal plicae transplants, being evidently less de-
veloped. The fact that they were usually larger than the mucous glands,

Fins. 1 TO 6. BI, adjacent integument of back; CE, columnar epithelial cells;
DG, dorso-lateral dermal plicae skin graft transplanted to belly; DI, integument of
dorso-lateral trunk region ; DP , normal dorso-lateral dermal plicae ; DPG, dorso-
lateral dermal plicae developed in skin transplanted to the belly; 11, thickened epi-
dermis of dermal plicae; G, granules of poison gland; MG, mucous gland; /', pig-
ment masses; PG, poison gland; '/", thickened stratum spongiosum of dermal plicae;
VG, belly skin graft transplanted to dorso-lateral dermal plicae region; VI, integu-
ment of the belly.

FIG. 1. Histological section through dermal plicae developed in skin graft from
dorso-lateral dermal plicae region previously transplanted to the back. Transplant
sectioned three weeks following close of larval metamorphosis. ( 112.

Fi<;. 2. Histological section through back skin graft previously transplanted to
dorso-lateral dermal plica: region. Transplant sectioned three weeks following
close of larval metamorphosis. X 224.

FIG. 3. Histological section through belly skin graft previously transplanted
to dorso-lateral dermal plicae region. Transplant sectioned three weeks following
close of larval metamorphosis. < 336.

Fin. 4. Histological section through dermal plicae developed in skin graft from
lateral (jaw) dermal plicae region previously transplanted to the back. Transplant
sectioned three weeks following close of larval metamorphosis. ' 112.

FIG. 5. Macroscopic appearance of skin graft from dorso-lateral dermal plica;
region, previously transplanted to the belly, three weeks following the close of
larval metamorphosis. ( 10.

Fin. 6. Macroscopic appearance of belly skin graft, previously transplanted to
dorso-lateral dermal plicae region, three weeks following the close of larval meta-
morphosis. ( 10.

PLATE I

MQ

MG

PG

£vf?-:1 ". ''~"^-,¥-;~.-*''''i_~':^ ~- -'-:•

S3 ^ --• -

'.-•

MG PG E T

~^^

'^~T:-S^^^^-''-,'^,^^' '10t'

4

-DPG

^;;-'-'VVirivii^
|:,^P??3

^^^K^^f

;•"»'. v"v**7»*1.**"'-i;*'^"«

na»- ,tt'-i
^^

&m&&&

Jii

DP

DG VI

6

VG

DI

DERMAL PLIC^: FORMATION 19

however, probably accounted for the thickening of the stratum spongi-
osum present. The slight thickening of the epidermis and the moderate
increase in thickness of the stratum spongiosum resulted in a convexity
of the epidermal surface of the integument which, however, was not
nearly as pronounced as in the case of the dorso-lateral dermal plicae
transplants. In general, it can be said that the histological appearance
of the lateral dermal plicae transplants on the back corresponded very
closely to the microscopic picture of the normal lateral dermal plicae of
the same animal at this time.

DISCUSSION

The results of the present work confirm those of Massie (1894) for
the frog tadpole and of Wilder (1925) for Eurycea larvae in that no
anlagen of the acinous glands appear until just prior to the onset of
metamorphosis. Likewise, the evidence is strongly indicative that they
are derived from the Malpighian layer of the epidermis, separate devel-
opments being necessary for the two types of glands. There is no in-
dication in the histological sections that transformation ever occurs from
mucous to poison gland. In fact, as regards the dorso-lateral dermal
plicae regions especially, there is every reason to believe that the poison
glands develop first and contain typical granular inclusions in the lumen
at a time when the mucous glands are still in the anlage stage.

Study of the dermal plicae in various stages of development leads to
the conclusion that the progressive whitening of dermal plicae integument
is due to several factors. Of these the gradual reduction in the number
of melanophores both in the epidermis and stratum spongiosum is prob-
ably of primary importance. This reduction in pigmentation apparently
more than counterbalances the effects of increased thickening of the epi-
dermis and stratum spongiosum. The latter layer, it may be stated, is
apparently fully as dense in dermal plicae regions as it is elsewhere
throughout the integument. The extraordinary large size of the poison
glands in dermal plicae regions may also help to give the integument a
certain degree of transparency and so facilitate the development of the
characteristic white coloration of the plicae as viewed externally.

The results of reciprocal, dermal plicae skin transplantations with
integument of the back and belly are especially clear in that certain nar-
row bands of larval integument are evidently physiologically specific for
the development of mucous and large poison glands during larval involu-
tion. This specificity would also seem to include the ability to reduce
pigmentation and to thicken the epidermis and stratum spongiosum.
Apparently, therefore, neither the adjacent musculature or other local
factors have anv influence on the integument in the development of

20 O. M. HELFF

dermal plica' structures. This is borne out by the fact that back and
belly skin transplanted to dermal plicae regions failed to develop the
slightest tendency towards dermal plicae characteristics and that integu-
ment from dermal plicae regions was capable of forming typical dermal
plicae structures at the usual metamorphic stage when transplanted to
foreign locations. Logically, this specificity must be determined at
some stage in the development of larval integument, the probability be-
ing that it is completed at a relatively early growth period of the larva,
possibly at the same time that the integument of the embryo is formed.
It would be interesting, in this connection, to make suitable integu-
mentary transplantations on newly hatched tadpoles.

Granting that certain bands of integument are definitely specific for
dermal plicae formation during larval involution, the question naturally
arises regarding the possible causative influences which evoke this dif-
ferentiation and growth. It would seem probable that these influences
are hormonic in nature and hence transported through the blood stream
when released at a definite stage of metamorphosis. Possibly only an
initial hormonic stimulus is necessary to bring about development of
dermal plicae structure or it may be that continued stimulation of this
nature is essential for complete differentiation. Suitable homoplastic
transplantations of integument, including half-formed dermal plicae, to
non-metamorphosing larvae would undoubtedly answer this question.

The apparently direct influence of metamorphic hormones through
the blood stream has been shown to effect the differentiation and growth
or degeneration of other larval structures during anuran metamorphosis.
The transformations so effected include the differentiation and growth
of the fore-limbs (Helff. 1926), the hind-limbs (Schubert. 1926). the
tongue (Helff, 1929a). the columella (Helff, 1929&), the nictitating
membrane (Lindeman, 1929a) ; the histolysis of tail muscle (Helff and
Clausen, 1929, and Clausen, 1929), and of tail integument (Lindeman,
\929b, and Clausen, 1929) ; and the development of the adult pigmenta-
tion pattern (Lindeman, I929b). Preliminary work by the writer also
indicates that the shedding of larval teeth during metamorphosis falls
under the same category. Contrasted to this direct mode of hormonic
influence, indirect mechanisms, involving first the degeneration or dif-
ferentiation of other structures which in themselves then bring about
other degenerations and differentiations, have been shown to apply to
anuran metamorphosis. The transformations so initiated include the
histolysis of opercular integument in the formation of the fore-leg per-
forations (Helff, 1926) ; the differentiation of the integumentary por-
tion of the tympanic membrane (Helff, 1928) ; and the differentiation of

DERMAL PLIC;E FORMATION 21

the yellow fibrous region of tympanic membrane lamina propria (Helff,
1929/0-

SUMMARY AND CONCLUSIONS

1. The origin, time and separate development of the mucous and
poison glands of Rana pif>icns have been determined as they are found,
especially, in dermal plicae regions. The results confirm those of the
majority of workers on other anurans and on urodeles.

2. Larval skin transplantations designed to test the specificity of
various integuments for plicae formation were made. These included
autoplastic, reciprocal transplantation of skin known to normally develop
dorso-lateral or lateral dermal plicae with integumentary grafts from the
belly and back. In such cases no demonstrable histological differentia-
tion of glandular structures was observable at the time of transplanta-
tion. During metamorphosis, the various skin grafts developed the
typical macroscopic and microscopic pictures of dermal plicae character-
istic of the region from which transplantation was made. Back and
belly integumentary grafts transplanted to dermal plicae regions main-
tained their specific histological characteristics, failing in every case to
develop dermal plicae structure.

3. The results are discussed and the conclusion reached that the po-
tentiality to develop dermal plicae and their histological components is
determined and fixed in certain bands of integument at a very early
stage of larval growth. The differentiation of this integument to form
dermal plicae structure is probably the result of direct hormonic influence
through the blood stream at a definite stage of larval involution.

LITERATURE CITED

ANCEL, P., 1902. fitude du developpement des glandes de la peau des Batraciens et

en particulier de la Salamandre terrestre. Arch, dc Blol., 18: 257.
CALMELS, M. G., 1883. fitude histologique des glandes a venin du crapaud. Arch.

de PhysioL, 15: 321.
CLAUSEN, H. J., 1929. Rate of Histolysis of Anuran Tail Skin and Muscle during

Metamorphosis. Anat. Rcc., 44: 218.
DAWSON, A. B., 1920. The Integument of Necturus maculosus. Jour. Morph., 34:

487.
ENGLEMAN, TH. W., 1872. Die Hautdriisen des Froschcs. Eine physiologische

Studie. Arch. f. d. gcs. PhysioL, 5: 498.
ESTERLY, C. O., 1904. The Structure and Regeneration of the Poison Glands of

Plethodon. Univ. Calif. Pub. Zoo]., 1 : 227.
FANO, L., 1903. Sull' origine, lo sviluppo e la funzione delle ghiandole cutanee

degli Anfibi. Arch. Ital. Anat. Embr., Firenze, 2: 405.
HEIDENHAIN, M., 1893. Die Haudriisen der Amphibien. Sitzb. phys.-mcd. Gesell.,

Wiirzburg, Jahrg. 1893, p. 52.
HELFF, O. M., 1926. Studies on Amphibian Metamorphosis. I. Formation of the

Opercular Leg Perforation in Anuran Larvae during Metamorphosis.

Jour. Expcr. Zool., 45: 1.

O. M. HELFF

HELFF, O. M., 1928. Studies on Amphibian Metamorphosis. III. The Influence

of the Annular Tympanic Cartilage on the Formation of the Tympanic

Membrane. Physiol. Zool., 1: 463.
HELFF, O. M., 1929c. Studies on Amphibian Metamorphosis. IV. Growth and

Differentiation of Anuran Tongue during Metamorphosis. Ph\siol. Zool.,

2: 334.
HELFF, O. M., 1929fr. Formation of the Yellow Elastic Fibrous Region of the

Lamina Propria of Anuran Tympanic Membrane. Anat. Rcc., 44: 214.
HELFF, O. M., AND CLAUSEN, H. J., 1929. Studies on Amphibian Metamorphosis.

V. The Atrophy of Anuran Tail Muscle during Metamorphosis. Ph\s-

iol.Zodl. 2: 575.
JUNIUS, P., 1896. Ueber die Hautdriiscn des Frosches. Arch. f. mikr. Anat., 47:

136.
LINDEMAN, V. F., 1929a. Development of the Nictitating Membrane of the Frog

(Rana pipicns). Anat. Rcc.. 44: 217.
LIXDEMAX, Y. F.. 1929/?. Integumentary Pigmentation in the Frog, Rana pipicns,

during Metamorphosis, with Especial Reference to Tail-Skin Histolysis.

Physiol. Zool, 2: 255.

MASSIE, J. H., 1894. Studies from the Neurological Laboratory of Denison Uni-
versity. YIII. Glands and Nerve-endings in the Skin of the Tadpole.

Jour. Co in par. Neiir.. 4: 7.
NICOGLU, P., 1893. tvber die Hautdriisen der Amphibian. Zeitschr. f. wiss. Zool.,

56:409.
NIRENSTEIN, E., 1908. Uber den Ursprung und die Entwicklung der Giftdriisen

von Salamaudra maculosa nebst einem Beitrage zur Morphologic des Sek-

retes. Arch. f. mikr. Anat., 72: 47.
SCHUBERT, M., 1926. Untepsuchungen iiber die Wechsclbeziehungen zwischen

wachsendcn und reduktiven Geweben. Zeitschr. f. mikr. Anat. Forschung,

6: 162.
SCHULTZ, P., 1889. Ueber die Giftdriisen der Kroten und Salamander. Arch. f.

mikr. Anal., 34: 11.
SEECK, O., 1891. Uber die Hautdriisen einiger Amphibien. Inaug.-Dissert.,

Dorpat.

SZCZESNY, O., 1867. Beitnigc zur Kenntnis der Textur der Froschhaut. Inaug.-
Dissert., Dorpat.
WEISS, O., 1908. Ueber die Entwickelung der Giftdriisen in der Anurenhaut.

Anat. Anseig., 33: 124.

WEISS, O., 1915. Zur Histologie der Anurenhaut. Arch. f. mikr. Anat.. 87: 265.
WENIG, J., 1913. Der Albinismus bei den Anuren, nebst Bemerkungen iiber den

Ban des Amphibien-Integumcnts. Anat. Anseig., 43: 113.

WILDER, I. W., 1925. The Morphology of Amphibian Metamorphosis. Smith Col-
lege Fifteenth Anniversary Publications, Yol. 6. The Southworth Press,

Portland, Me.

ON THE CONDUCTION OF THE CORTICAL CHANGE AT
FERTILIZATION IN THE STARFISH EGG

D. M. WHITAKER

(From the Hopkins Marine Station, Stanford University, and
the Department of Zoology, Columbia University)

A number of radical changes take place in the echinoderm egg when
it is fertilized, one of which is a change in the cortex which prevents
the entrance of a second sperm. That the egg is not entirely passive at
the time of the entrance of the sperm is well known in a number of eggs
from the phenomenon of the entrance cone which flows out and partly
engulfs the sperm head.

For many years there has been a difference of opinion as to whether
the fertilization membrane pre-exists in the sea urchin egg before fer-
tilization. The more recent general opinion has been that some sort of
a membrane is present. A. R. Moore (1929), however, has recently
presented evidence which he regards as supporting Loeb's view that this
membrane does not pre-exist but is formed dc noi'o at the time of fer-
tilization. In the case of the starfish egg, however, there is no question
that a definite tough membrane pre-exists which normally lifts off fol-
lowing fertilization. Chambers (1921) has lifted this membrane from
the unfertilized Asterias egg with a micro-needle.- It has also been
shown (Whitaker, 1928) that the cytoplasm of the egg of the starfish
Patina mimata may be readily divided into two separate parts with a
micro-needle without severing the tough outermost membrane. If an
egg which is divided in this way is inseminated, both fragments can be
fertilized and will develop independently within the common fertiliza-
tion membrane. In this case there can be no doubt that a morphological
membrane is present, and that it lifts off as the fertilization membrane,
even though the properties of the membrane may perhaps change fol-
lowing fertilization.

As early as 1878, O. Hertwig expressed the view that it is the egg
plasma itself rather than the fertilization membrane which can prevent
the entrance of a second sperm. Just (1919) has pointed out that ob-
viously the fertilization membrane rises off the egg so late after the
contact of the fertilizing sperm that another and more rapid change of
the cortex must be postulated to explain rejection of the second sperm.1

1 In Arbacia the membrane begins to lift off about eighteen seconds after con-
tact of the fertilizing sperm, and in Echinarachnius after about twenty to forty
seconds (from Just, 1929).

23

-7-± D. M. WHITAKER

Just has further observed in Echinarachnius that this change, which he
calls a " wave of negativity," spreads rapidly from the point of contact
of the first sperm, just as the lifting of the fertilization membrane itself
spreads from this point, but spreading much more rapidly so that it has
covered the entire egg before the fertilization membrane has even started
to rise at the entrance point of the sperm. His evidence for this " wave
of negativity " is based on observations on living eggs. As soon as the
tip of the first sperm has penetrated the egg, other sperm are not en-
gulfed in the immediate vicinity of the first sperm, although they may
be taken in further around the egg. As the first sperm penetrates
still further, the " wave of negativity " progresses further around the
egg until at the moment the sperm head has entirely disappeared into
the egg, only the opposite pole of the egg can engulf another sperm.
Sperm which become attached but do not enter the egg are lifted off as
the fertilization membrane rises.

More recently Just (1929) and others have observed that in eggs
which are in excellent condition, the cortical reaction is so rapid as to
be practically instantaneous. It is practically impossible to produce
polyspermy by the use of concentrated sperm suspensions. It is evi-
dent that in such innoculations the time elapsing between the contact of
the first and the contact of the second sperm must be an exceedingly
small fraction of a second. The spread of the cortical change in eggs
in perfect condition is much too rapid to be traced by the eye. Just
(1930, page 337) says, " At least for the normally monospermic ova of
the marine forms which I have studied, it is certainly true that if they
are in optimum condition, polyspermy is difficult if not impossible.
Such ova in order to become polyspermic must undergo treatment which
impairs their cortices. . . . Doubtless the very instant that contact be-
tween spermatozoon and ova is made, polyspermy is blocked."

The response of the egg to the stimulus of the first sperm involves
the whole interior of the egg as well as the cortex. Flowing of the
interior protoplasm immediately succeeds fertilization, as well as changes
in viscosity, etc. (Heilbrunn, 1928). The interior changes may be sec-
nndary, however, — that is, consequences of changes in permeability and
•Mir face tension of the cortical region.

That the cortex performs a necessary function in admitting the first
sperm, as well as in preventing polyspermy, has been shown by Just
(1923) fo,- 1-.t-hmarachmus and for Arbacia and by Chambers (1921)
for the starfish Asterias. Just burst eggs by returning them from
hypertonic to normal sea water, and also by passing eggs through fine-
meshed bolting silk and through lens paper to obtain rndoplasmic buds
or fragments bearing none of the original cortex of the egg. Chambers

CONDUCTION OF CORTICAL CHANGE

obtained similar fragments by means of the microdissection needle. In
none of these eggs did the endoplasmic buds react to sperm or become
fertilized.

The present observations add little evidence as to the nature of the
cortical reaction. They are concerned with the locus or site of conduc-
tion of the spreading impulse. If it starts at a point on the cortex, it
must spread either in the form of the cortical change itself, or else it
must be an equilibrium shift transmitted either around the cortex, or
through the egg, giving rise to the cortical change.

In a previous paper on the development of fragments of Patiria eggs
(Whitaker, 1928), 69 cases are recorded in which two completely sep-
arated fragments lying within the same fertilization membrane-to-be
were inseminated and became fertilized independently, one sperm enter-
ing each fragment. This result is almost invariably obtained when the
two masses of cytoplasm are not in contact at the time of insemination.
This seems to show that the impulse which spreads through the egg and
results in the cortical change which prevents the entrance of a second
sperm does not pass through this outermost membrane. There is no
evidence of injury to the outer membrane as a result of pinching the
protoplasm. It lifts off in the normal way after fertilization, although
sometimes it is temporarily constricted where the needle has pressed.
Additional experiments have been made in which eggs have been pinched
into two cytoplasmic fragments lying within a common outer membrane,
and the fragments have been allowed to flow together before insemina-
tion. In this case the treatment of the outer membrane is practically
identical with that in the cases of fragments which remain apart, the
difference between the two types of cases being only in the fusion or
non-fusion of the protoplasm within the membrane. Only sets of eggs
were used in which control tests gave 98-100 per cent fertilization. A
moderately concentrated suspension of sperm was used. As a control,
to test for polyspermy, five or six normal eggs were placed by means of
a mouth pipette around the experimental egg, not more than several
egg diameters away. Half an hour after insemination the experimental
egg was transferred to a separate dish, since at the time of cleavage when
the eggs are changing shape there is otherwise some danger of mistaken
identity. For some time after insemination, however, the marks of
cutting are clearly discernible on the experimental egg. In all cases the
eggs were left in fresh sea water for about ten minutes after cutting
before insemination. It has been found in a number of echinoderm
eggs that this procedure permits the fragments to recover from the
operation and in the case of Arbacia eggs, for example, it avoids poly-
spermy which often results from immediate insemination.

26 D. M. WHITAKER

The degree of separation of two cytoplasmic fragments lying within
the same outer membrane can be regulated by the extent to which the
membrane is stretched with the needle, a very slight difference in the
amount of stretching determining whether the fragments will flow back
together or not. The eggs of Patiria are large, averaging about 190 or
200 microns in diameter, and they are comparatively clear, so that the
grosser aspects of the nuclear and astral phenomena can usually be
clearly seen in the living egg or egg fragment.

Eleven eggs were cut, after the fashion of the 69 cases already
quoted, so that the two fragments were separate at the time of insemina-
tion. The control eggs lying near the experimental egg showed no poly-
spermy. In 9 of the 11 cases, both fragments were fertilized inde-
pendently, each dividing with a single amphiaster into two normal cells
(Fig. 1, a,, b, c). In two cases one fragment only became fertilized
with one sperm, the other fragment failing to become fertilized.

Twenty-five eggs were cut into two fragments and allowed to flow
completely together before insemination. In 18 of these cases the frag-
ments remained separate for one to two seconds, in three cases for more
than five seconds, and in four cases for fifteen seconds to one minute.
If the fragments remain apart for several minutes, a solidification of the
cut surface usually prevents a complete refusion of the fragments, or
in cases of fusion results in an ectoplasmic wall along the surface of
fusion. In the 25 cases of complete fusion, 23 cleaved with one amphi-
aster into a normal two-cell stage. Apparently only one sperm had
entered. There were no cases of polyspermy in the control eggs. In
two cases division was by means of a triaster into three cells, apparently
due to the entrance of two sperm. One of these cases of polyspermy
was an egg separated for one to two seconds. The controls showed no
polyspermy. The other was an egg which had been apart for more than
fifteen seconds. The controls in this last case showed polyspermy, prob-
ably indicating that insemination was too heavy and that the eggs were
not in the best condition.

The results show that when a sperm enters one part of an egg. usu-
ally no other sperm enters any other part of the egg which is joined by
1 protoplasmic fusion, but another sperm does enter a part of the egg
which has only the same outer membrane in common, equally well as if
the parts of the egg were entirely separated and removed from one
another. Apparently the cortical change which normally prevents poly-
spermy is not conducted through the outermost membrane, the fertiliza-
tion membrane-to-be. It is conducted through the protoplasm within
this membrane.

The question naturally arises: To what extent must two fragments

CONDUCTION OF CORTICAL CHANGE

27

be associated in order that the fertilization of one shall prevent the fer-
tilization of the other? Is mere contact of the surfaces of two frag-
ments within the outer membrane sufficient or is fusion necessary? In
contrast with our extreme cases of complete separation and complete
fusion, which can be identified with certainty and in which the results
are clear cut, it is difficult to tell by observation in the border line cases
in which the two fragments are in contact precisely what degree of fu-
sion has taken place. Fragments which are in contact but not fused at
the time of insemination usually fuse soon after fertilization. Although
no statement can be made at present as to the minimum degree of contact
or fusion necessary to transmit the impulse, the following experiments
throw some light on the question.

Fifteen eggs were cut into two fragments which remained separate
for more than a minute and then came together so that at the time of
insemination they are classed as (1) touching, (2) in good contact, or
(3) fused at a narrow neck, the neck being from less than ten to thirty
microns through. The eggs are 200 microns in diameter. In five cases
the fragments were touching. In four of these cases one sperm entered
each fragment. The fragments subsequently fused, but retained an
ectoplasmic partition and separated at the time of cleavage, each com-

+ AN

FIG. 1. Diagrams before insemination, after insemination, and just before
cleavage, of eggs cut without severing the outermost membrane. (Sperm dis-
proportionately enlarged.)

a. b, c. An egg with halves separated at time of insemination.

a', b', c'. An egg with halves joined by a narrow neck at time of insemination.

D. M. WHITAKER

ponent dividing with a single amphiaster into two cells. The fifth case
admitted only one sperm and divided as a whole into two cells. In three
cases the fragments were in good contact, that is. about on the border
between touching and slightly fused. Two of these cases had one am-
phiaster and divided into two cells. The third admitted a sperm into
each half and divided directly into four cells. In seven cases the two
halves were truly fused by a narrow neck at one end of the cut. Shortly
after fertilization more complete fusion took place. In all of these
seven cases only one sperm entered the joined fragments. In four of
the cases normal division into two cells followed (Fig. 1, a', b' , f'). In
two cases the nucleus of the sperm, which had entered the non-nucleated
fragment of the egg, was not able to pass through the gelated partition
of the fused surfaces to fuse with the egg nucleus, and only the frag-
ment containing the sperm nucleus divided into a normal two-cell stage.
The other half separated off at the time of division but did not itself di-
vide. In the seventh of these cases the sperm entered the non-nucleated
fragment, which divided, and the nucleated fragment which apparently
received no sperm divided late, involving the egg nucleus only.

These results on fragments which are only in slight contact show
that a very narrow neck of truly fused protoplasm conducts the impulse.
Mere contact without fusion probably docs not.

SUMMARY

When eggs of the starfish are pinched into two cytoplasmic frag-
ments which lie within the same outermost membrane (the fertilization
membrane-to-be), and are inseminated while the two fragments are sep-
arate, each fragment receives one sperm. If the fragments completely
fuse before insemination, only one sperm enters the fused egg. If the
fragments are merely touching, each may receive a sperm. If they are
joined by a very narrow fused neck of protoplasm, only one sperm
enters.

CONCLUSION

The outermost membrane of the starfish egg (the fertilization mem-
brane-to-be) does not conduct the cortical change at fertilization which
prevents the cut ranee of a second sperm. This change is conducted by
the protoplasm within the membrane.

BIBLIOGRAPHY

CHAMBERS, Rom i i. l''Jl. Studies on the Organization of the Starfish Egg. Join:

Gen. Physiol.,4: 41.
CHAMBERS, ROIIKKT. 1021. M in •< dissection Studies, III. Some Problems in the

Maturation and Fertilization of the Echinodcrm Egg. Biol. Bull. 41: 318.

CONDUCTION OF CORTICAL CHANGE 29

HEILBRUNN, L. Y., 1928. The Colloid Chemistry of Protoplasm. Protoplasma-
Monographien.

JUST, E. E., 1919. The Fertilization Reaction in Echinarachnius parma. I. Cor-
tical Response of the Egg to Insemination. Biol. Bull., 36: 1.

JUST, E. E., 1923. The Fertilization- Reaction in Echinarachnius parma. VI.
The Necessity of the Egg Cortex for Fertilization. Biol. Bull., 44: 1.

JUST, E. E., 1929. Initiation of Development in Arbacia. IV. Some Cortical Re-
actions as Criteria for Optimum Fertilization Capacity and Their Signifi-
cance for the Physiology of Development. Protoplasma, 5: 97.

JUST, E. E., 1930. The Present Status of the Fertilizin Theory of Fertilization.
Protoplasma, 10: 300.

MOORE, A. R., 1929. The Function of the Fertilization Membrane in the Develop-
ment of the Larva of the Sea Urchin. Science, 70: 360.

WHITAKER, D. M., 1928. Localization in the Starfish Egg and Fusion of Blastulse
from Egg Fragments. Physio!. Zoo!.. 1: 55.

A STUDY OF THE EFFECTS OF MODIFICATIONS OF THE

CULTURE MEDIUM UPON LENGTH OF LIFE AND

FECUNDITY IN A ROTIFER, PROALES SORDIDA,

WITH SPECIAL RELATION TO THEIR

HERITABILITY

RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

(From the Zoological Laboratory of the Johns Hopkins University and the
Marine Biological Laboratory, Woods Hole, Mass.)

INTRODUCTION

Extensive investigations on the Infusoria have demonstrated that
these organisms, while undergoing continuous vegetative reproduction,
under certain conditions " run down,"-— become depressed. In many of
the earlier genetic studies this progressive depression was interpreted to
mean that the organisms have a definite life cycle with periods of youth,
maturity, and old age, representing through the course of many consecu-
tive generations of vegetative reproduction the usual life cycle of a single
metazoan organism. The results of later investigators, however, show
that in many cases the depression is partly or entirely due to unfavorable
environmental conditions. Enriques, Woodruff, and others found that
under adequate cultural conditions certain Infusoria can be cultivated
indefinitely without decline in vitality, although the same Infusoria un-
der poor cultural conditions exhibit a depression which may continue for
long periods, becoming more and more marked as generations pass and
eventually resulting in the death of most of the lines.

The inheritance of such a depression, after restoration to normal
conditions, was tested by Middleton (1918) and by Jollos (1921). Mid-
dleton found that a depressed condition induced in Stylonchia pKstulata
by long continued culture at a high temperature is inherited for at least
twenty days after restoration to normal temperature. Jollos produced
a depression of the fission rate- in Paramcciuin aiirclia by subjection to
calcium nitrate; he found that this lowered rate of reproduction is
heritable for four months. In most other cases it has not been possible
to test critically the inheritance of the depression after restoration to
normal conditions, but it is commonly assumed that inheritance occurs.
(For a detailed review of this type of work on the Protozoa, see Jen-
nings, 1929.)

The present investigation was undertaken to determine how far

30

MODIFICATIONS OF CULTURE MEDIUM

conditions in the lower Metazoa are comparable to those in the Infu-
soria, particularly as to the production of depression, its increase as
generations pass, and its inheritance after return to normal conditions.
The Rotifera are particularly suitable for such an investigation. They
resemble the Infusoria in size, habitat, and mode of life, but reproduce
by means of germ cells which undergo a process of development in the
production of the many-celled individual. They can be subjected to the
same conditions that induce depression in the Infusoria, and the de-
pression so produced can be studied in relation to its tendency to cumula-
tion and its persistence after return to normal conditions.

The particular species of rotifer chosen, Proales sordida Gosse, has
the additional advantage of reproducing exclusively by parthenogenesis,
that is, from single parents, as do the Protozoa in vegetative reproduc-
tion. The complications of biparental inheritance, common to most
types of germ-cell reproduction, are therefore avoided.

Insufficient feeding was the method used for producing depression :
certain lines were well-fed ; other lines were given insufficient food in
definite ratios to the amount given the well-fed group. This method is
the one used by Beers (1926) with the infusorian Didinmni. He found
that his well-fed group exhibited a high fission rate throughout his ex-
periment; and that his inadequately-fed group showed from the first a
lower fission rate, which after some time became still further depressed,
this depression steadily increasing until the line encysted.

Previous experimental work on the Rotifera has been concerned
chiefly with the problem of sex-determination; but there are a few pa-
pers which bear on the questions here considered. Among these are
four on the effects of food control during parthenogenetic reproduction.
Luntz (1926), in a critical study of the relation of food to the onset of
the sexual generation in Ptcrodiiia clliptica, describes briefly an experi-
ment on quantitative variation of food. Using Beneke's solution as the
culture medium, and Polytoma or Clilainydonwnas as the food, he varied
the amount of available food by varying the number of drops used. He
found that the number of offspring produced varied directly and pre-
cisely with the amount of food provided. He does not give data for
successive generations ; this work therefore furnishes no data on cumu-
lative effects.

Shull (1912) and Whitney (1912«) carried on extensive investiga-
tions bearing upon this matter, in successive generations of the rotifer
Hydatina senta. Both of these investigators, using various types of
culture media, found that races reproducing exclusively by partheno-
genesis decline in vitality, as indicated by their lowered fecundity.
WThitney suggests that this may be due to " the constant environment

3

*** \A V

-IRY)<

32 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

of the horse manure cultures" (1912a, p. 343). Lambert, Rice, and
Walker (1923), working under the direction of Whitney, show that
this decline in vitality is due to the unbalanced diet, not to the partheno-
genetic method of reproduction. None of these four investigations take
up the question of the heritability of these depressions.

Other investigations, in which alcohol is used as the modifying fac-
tor, do deal with the heritability of modifications produced. Whitney
(1912c) subjected Hydatina senta to alcohol and produced a decrease
in fecundity which continued for two generations after return to a nor-
mal culture medium, but not longer. Noyes (1922) with Proalcs dc-
cipiens, and Finesinger (1926) with Distyla incrmis, produced a cumu-
lative decline in fecundity by subjection to alcohol and found that on
return to normal conditions the depression continued to some degree for
two generations only. (Finesinger, however, found that in weak doses
alcohol acts as a stimulant and causes a slight increase in both fecundity
and longevity.)

In the present study, dealing with the effects of a reduced food sup-
ply, two main questions are considered: First, do adverse environmental
conditions produce the same effects in the Rotifera as in the Infusoria;
that is, do they bring about a decline which increases with succeeding
generations, so that the Rotifera run down as do the Infusoria? Sec-
ond, if they become depressed, is this depression inherited in succeeding
generations when the progeny are returned to a favorable environment?

Two characteristics, fecundity and length of life, were taken as in-
dices of vigor. The effect of the reduced food supply on these indices
was studied during the summer of 1928; the results of these experiments
are presented in Part I. During the summer of 1929 the experiments,
somewhat modified, were repeated ; their results are presented in Part
II. All of the experimental work was done in the Marine Biological
Laboratory at Woods Hole.

The investigation was undertaken at the suggestion of Professor
H. S. Jennings. It gives us great pleasure to express our appreciation
of his advice throughout the investigation and his help in the preparation
"f the manuscript.

THE ORGANISM

I'roalcs sariliiia is a rotifer well suited to this type of investigation.
It lives in pond> and sluggish streams, and can be cultivated with ease
in the laboratory in almost any infusion favorable to the growth of
Protozoa; it feeds on bacteria and other minute organisms. The usual
length of life is eight days, though occasionally individuals live as long
as twenty-five days. Each animal produces twenty-four to twenty-

MODIFICATIONS OF CULTURE MEDIUM

eight eggs, with an observed maximum of thirty-four eggs. The adults
average 300 microns in length. Reproduction is exclusively partheno-
genetic. The species consists entirely of females; no males have ever
been described and none have appeared during the history of our cul-
tures even under conditions known to be favorable to male production
in other rotifers. It was therefore possible, by starting with a single
ancestor, to have all the animals used in these experiments members of
a single clone, forming a group as genetically similar as it is possible
to obtain. (For a more detailed description of this organism see Jen-
nings and Lynch, 1928, I.)

METHODS

It is not only necessary, for this type of experimentation, to have
animals that are alike genetically ; it is also necessary to have them as
uniform as possible in other respects. Uniformity in age is particularly
important, since Jennings and Lynch (1928, I) have demonstrated that
the offspring of young parents have a lower fecundity than the offspring
of the same parents when old. Proalcs sordida has an average life of
only eight days ; it begins to lay eggs about forty-eight hours after hatch-
ing. Its eggs, laid in the one-cell stage, take from twenty to twenty-four
hours to hatch. A population uniform in regard to age may therefore
be obtained by including in it only such eggs as are laid within a five-hour
period by parents differing in age by less than twenty-four hours. To
meet these conditions the following procedure was employed : In starting
each experiment, several thousand eggs were removed from mass cul-
tures and placed in oatmeal infusion in culture dishes, three hundred to
a dish. Twenty-four hours later a small amount of infusion was added
to each dish ; at that time most of the eggs had hatched. After the lapse
of another twenty-four hours, when most of the animals had begun to
lay eggs, the animals were transferred to dishes of fresh culture medium.
The eggs laid within the next five hours were removed and placed on
two-concavity slides, one egg in each concavity in a single drop of cul-
ture medium. The animals hatching from these eggs were the first
generation of experimental animals. By this procedure, eggs of ap-
proximately the same age were obtained from parents of about the same
age, thus fulfilling the requirements outlined above. Throughout the
remainder of the investigation, except in rare cases to be mentioned at
the appropriate places, no eggs later than the third were used to give
rise to a new generation.

Without exception the animals were cultivated individually on
ground-glass, two-concavity slides in single drops of culture medium.
The slides were kept on glass supports in nine-inch crystallizing dishes,
twelve slides to a dish. Each dish was water-sealed by placing boiled

-U RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

spring water in the bottom of the larger portion and inverting over it
the smaller portion.

The animals were examined at approximately twelve-hour intervals.
In the morning each animal was transferred by means of a capillary
pipette to a fresh drop of culture medium on a clean slide. In the eve-
ning the eggs which had been laid during the day were removed from
the drops, but the animals were left undisturbed. The capillary pipettes
were sterilized in boiling water before each transfer. Pipettes of the
same size were used to place the culture fluid on the slides. The pipettes
were boiled daily in distilled water and the rubber bulbs were boiled at
frequent intervals.

A favorable culture medium for Proalcs is an oat infusion prepared
by boiling fifteen flattened flakes (% gram) of rolled oats ("Quaker
Oats ") for three minutes in 100 cc. of spring water, filtering this, and
allowing it to stand for twenty-four hours before using (cf. Jennings
and Lynch, 1928, I). The strength of this culture fluid may be easily
varied by changing the number of flakes of oatmeal used. Such varied
solutions were used in these experiments. In every case the media were
prepared simultaneously in the same way, the only difference between
them being the number of flakes of oatmeal used.

PART I. EXPERIMENTS OF 1928

In this series of experiments the following strengths of culture me-
dium were used :

Control Series C, full-strength fluid, 15 flakes oatmeal to 100 cc. spring water

Test //, half -strength 8 "

Q. quarter-strength 4 "

S, boiled spring water only

These media were prepared as described above and were left exposed to
the air for twelve hours before being corked.

In the preliminary experiments it was found that the animals would
not live beyond the second generation in boiled spring water without oat-
meal (S culture medium), nor in the four-flake solution (Q culture
medium).

In the main experiment, begun July 19, 1928, and continued for
three months, three groups of animals were studied: one group (C)
was cultivated in full-strength culture medium; a second (//) in half-
strength culture medium; and a third (Q) was put for its first genera-
tion into quarter-strength culture medium, into full-strength for its
second generation, and then again into quarter-strength culture medium,
where it remained until its death. Some individuals of the later genera-

MODIFICATIONS OF CULTURE MEDIUM

35

tions of both the H and the Q groups, and their descendants, were culti-
vated in full strength medium (HC and QC groups).

At the start, one hundred individuals in each generation of all three
groups were kept throughout their lives. But the rapid increase in co-
existing generations so increased the number of animals to be examined
daily that a change in procedure was necessary. In most cases the num-
ber of individuals to a generation was reduced, usually to about fifty (cf.
Tables I and II). In other cases, not only were the members of a given
generation reduced in number, but they were kept just long enough to
produce the next generation, and then discarded. Such generations, of
course, yielded no data on fecundity or longevity.

TABLE I

Experiment I, 1928 — Showing for the successive generations (Gen.) of the two
control series C and M the total numbers of eggs isolated (No.) ; the numbers of
these that did not hatch (N.H.), and the numbers that hatched (H.) ; and for the
latter the mean numbers of eggs laid and the mean numbers of days lived. Data
presented graphically in Figs. 1 and 2.

C Series (full-strength medium)

M Series (full-strength medium)

Gen.

No.

N. H.

H.

Mean
Eggs

Mean
Days

Gen.

No.

N. H.

H.

Mean

Eggs

Mr:m

Days

1

92

1

91

7.67

5.76

1

49

1

48

3.60

5.23

2

106

5

101

6.02

4.85

2

95

7

88

1.93

4.32

3

47

6

41

6.41

5.02

3

51

16

35

7.88

7.02

4

52

7

45

6.22

4.77

4

50

2

48

11.31

9.56

5

57

3

54

6.18

4.49

6

58

8

50

3.02

3.74

9

49

1

48

9.31

11.61

7

49

8

41

1.95

3.68

8

52

3

49

2.55

4.11

11

49

0

49

8.85

9.45

9

75

17

58

1.77

4.00

10

56

6

50

2.74

4.18

13

42

1

41

11.58

9.17

11

51

7

44

2.40

4.38

12

60

12

48

2.04

3.51

13

49

4

45

1.73

3.18

14

48

3

45

1.66

3.75

15

49

5

44

1.36

4.15

16

40

7

33

1.81

4.59

17

40

4

36

2.16

4.56

18

45

6

39

5.41

6.25

19

49

4

45

12.17

10.06

24

51

1

50

8.86

10.70

26

39

1

38

12.57

10.71

28

44

0

44

16.52

11.44

36 RUTH STOCKING LYNCH AND HELEN REREXICE SMITH

TABLE II

Experiment I, 1928. Showing for the successive generations (Gen.) of the H
series (cultivated in half -strength medium), the Q series (cultivated in quarter-
strength medium), and the HC and QC series (cultivated in full-strength medium),
the total numbers of eggs isolated (No.), the numbers of these that did not hatch
(N.H.), and the numbers that hatched (H.) ; and for the latter the mean numbers
of eggs laid and the mean numbers of days lived.

II Series (half-strength medium)

Q Series (quarter-strength medium)

Gen.

No.

X. II.

II.

Mean
Eggs

Mean
Days

< "n-ii.

No.

N. II.

H.

Mean
Eggs

Mean
Days

1

97

1

96

2.16

3.69

1

100

0

100

0.18

2.01

2

93

0

93

2.19

3.69

t(2)

17

0

17

7.76

5.50

3

87

6

81

2.54

3.16

3

55

8

47

0.95

1.96

4

52

10

42

3.97

3.78

4

17

0

17

2.00

3.52

5

46

2

44

3.43

3.95

5

34

2

32

1.68

3.09

6

62

0

62

2.41

3.64

6

49

1

48

1.02

2.70

7

45

3

42

2.11

3.42

7

49

0

49

0.4S

2.53

8

45

4

41

1.24

2.85

*8

13

1

12

0.00

1.75

9

43

6

37

0.94

2.75

10
*11

A4 1

18
7

2
0

16

7

0.62
0.14

2.50
3.00

2^ £

QC Series (full-strength medium)

*12
*13

10

4

0
1

10
3

0.50
0.00

.25
1.83

*1(8)

11

2

9

4.00

4.72

*2(9)

33

6

27

1.29

3.77

HC Series (full-strength medium)

*• \-f /

3(10)
*4(11)

34
35

4
13

30
22

1.30
0.90

3.30
2.95

*S(12)

20

3

17

0.64

3.26

*1(10)

14

2

12

2.50

3.95

*6(13)

9

0

9

0.77

3.88

*2(11)

27

4

23

1.04

3.19

*7(14)

3

0

3

2.00

4.33

•3(12)

23

7

16

1.62

2.90

•8(15)

2

0

2

5.50

7.75

*4(13)

24

3

21

2.90

4.57

*9(16)

8

0

8

8.12

9.18

5(14)

48

1

47

1.91

4.25

10(17)

44

4

40

5.50

7.80

6(15)

58

10

48

2.14

4.37

11(18)

46

3

43

6.06

8.38

7(16)

46

6

40

2.47

4.72

12(19)

47

0

47

5.48

8.70

8(17)

45

5

40

6.45

6.83

13(20)

54

0

54

6.98

10.24

9(18)

49

7

42

9.90

8.61

15(22)

47

0

47

9.68

9.95

14(23)

41

2

39

8.07

11.14

17(24)

50

0

50

14.34

11.01

16(25)

39

0

39

10.38

10.23

27)

49

0

49

15.14

11.15

' dn.un in full-strength medium.
* (iciin .itions comprising less than thirty animals.

MODIFICATIONS OF CULTURE MEDIUM

37

(icncral Description

I. The Control Series.

During the course of the experiment the control animals lived through
twenty-eight generations, showing great variability in length of life and
fecundity. This variation appeared to he correlated with the tempera-
ture fluctuations. The temperature was high when the experiment be-
gan and remained high for sixteen generations ; these generations showed
a progressive decline in both fecundity and longevity. The succeeding
generations (17th to 27th), subjected to a lower temperature, showed a
recovery which started with the first generation grown at the cooler tem-
perature. This recovery was so swift, both fecundity and longevity in-
creased so rapidly, that with the 20th generation it became necessary to
reduce the number of animals examined. This was done by testing only
occasional generations, as noted above. The number of animals in the
generations not tested (generations 20, 21, 22, 23, 25, and 27) was re-

n ;

2 3 •* 5 to 7 6 9 10 11 1? 13 14 15 16 1? 16 19 20 21 22 23 2* 25 26 27 28

FIG. 1. Experiment I, 1928. Comparison of the two control series M and C,
based on egg production ; horizontal axis represents successive generations ; vertical
axis, mean egg production.

Mot synchronous

1 2 } 4 5 « 7 » ? 10 11 1? JJ 14 15 1« IT 11 19 X !1 n 3) It n It, }T IS

FIG. 2. Experiment I, 1928. Comparison of the two control series M and C,
based on length of life ; horizontal axis represents successive generations ; vertical
axis, mean length of life in days.

38 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

duced to thirty, and these were maintained only long enough to provide
the next generation.

\Yhen the control (C) animals were in their most depressed condi-
tion (generation 15) a special effort was made to determine the cause of
their depression by comparing them with a second series of controls,
taken at this time from mass culture. This new control series (termed
M) was treated in every way like the original control series, so that the
only known difference between the two sets was the difference of date of
derivation from mass culture, resulting, of course, in a diversity in dura-
tion of cultivation on slides. If duration of slide cultivation is the ef-
fective cause of depression, the early generations of .!/ should not show
depression ; there should he a wide diversity between their records and
the records of the synchronous generations of C (16th to 28th genera-
tions undergoing slide cultivation). If. on the other hand, some current
environmental factor is the effective cause, synchronous generations of
the two series should yield like data. Figures 1 and 2 demonstrate the
latter condition. C and M show a similar and synchronous modality.
Figure 1 is based on egg production ; Fig. 2 on length of life. In both
figures the record of the second control series M is shown in two ways;
first, in relation to the records of the early generations of the C series,
and second, in relation to the synchronous generations of C. It is obvi-
ous that the record of the .17 series, comprising thirteen generations, fol-
lows closely the record of the generations of the C series that were living
at the same time ; not the record of the early generations of C. It is
therefore concluded that the cause of the depression in C was some fea-
ture of the current environment, probably the high temperature; and that
duration of slide cultivation had had no depressive effect.

II. The H Series.

The series cultivated in half-strength culture medium (H) was con-
sistently lower in fecundity and length of life than the controls. Con-
secutive generations showed some variation, but became progressively
depressed, and the series died out in the 13th gene-ration. In Table II
are given the records of series // for every generation, showing the mean
egg production and the mean length of life.

Fortunately the ultimate extinction of the // series was anticipated
and some eggs of the ninth generation were placed in the full-strength
culture medium, thus beginning a group of animals (series HC) in which
to test recovery from depression. At the end of the experiment the HC
group had been cultivated continuously for eighteen generations in the
full-strength culture medium. It showed complete recovery from the
depression (see Table 1 1. HC). Generations 19, 20, 21, 22, 24, and 26
were not tested.

MODIFICATIONS OF CULTURE MEDIUM

III. The Q Series.

In the preliminary experiments the animals cultivated in the quarter-
strength culture medium had not survived beyond the second generation.
The exceedingly low records of the first generation of the Q series of
the main experiment indicated that they were following a similar course.
Consequently, it was thought best to rear the animals of the second gen-
eration in full-strength culture medium. The records of this second
generation of Q, cultivated in the full-strength culture medium, greatly
exceeded those of the second generation of C animals. But the records
of the third generation, cultivated in the quarter-strength culture me-
dium, were again very low, although not as low as those of the first
generation. The subsequent generations were grown continuously in
the quarter-strength culture medium ; they declined consistently and died
out in the eighth generation.

A few eggs from the seventh generation were placed in the full-
strength culture medium to test the recovery of the Q series. This QC
series was continued in the full-strength solution until the end of the
experiment. They improved, but their records on the whole were lower
than those of the controls (C series) and lower than those of the HC
series. The generations did not overlap as much in this group as in
the other two groups, and at the end of the experiment, when the H-HC
series was in its 27th generation and the C series was in its 28th genera-
tion, the Q-QC series was only in its 24th generation. Generations 21
and 23 were the only ones not tested. The record for each generation
of the Q-QC series is given in Table II. Because of the low fecundity
of these depressed animals the number of individuals in many genera-
tions was small, although all the eggs laid were kept, including in some
cases eggs later than the third.

Temperature Variation

During the summer of 1927 the laboratory in which these experi-
ments were conducted had shown fluctuations of less than two degrees
(cf. Jennings and Lynch, I, p. 369), and it was expected that it would
maintain the same constancy during the summer of 1928. Unfor-
tunately this was not the case. The temperature of the Woods Hole
region was unusually variable from July until September of that year,
and the laboratory showed correspondingly greater fluctuations.

The temperature was read at twelve-hour intervals from a Maxi-
mum-minimum thermometer which was suspended above the rotifer
cultures. The data thus obtained are presented in Fig. 3, which is de-
signed to show the relation between the temperature fluctuations and the

40 RUTH STOCKING LYNCH AND HELEX BERENICE SMITH

records of the control animals (C series). On the vertical axis the tem-
perature is plotted in degrees centigrade ; on the horizontal axis the time
is plotted in twelve-hour periods, for the duration of the experiment.
The upper, shaded curve (a) shows the maximum and minimum records
for each twelve-hour period, the upper line showing the maximum, the
lower line the minimum. The second curve (b) is a simplified form of

Temperature Variation and Control bents

d Max nr\\iir\- Hummum Temperature
U It 10 M Jff September /ft ;* IS 21 J* October 9 « 76

0 Simplified Maximum-Temperature Curue,

uly IV J>

A

'<^»t" 9 II It JO JV ^9 Stytti.'w'/o" II 18 « i* (ictob"'^'

'1 /6

FH,. 3. Experiment I. 1928. Chart showing temperature variation and the
ords <>f control series ( . ( </ ) I'ppcr shaded curve: the maximum and minimum
temperatures at twelve-hour intervals from July 19th to October 16th; the upper
line shows the maximum temperatures; the lower line shows the minimum tempera-
tures. ( /• i Mi. Idle curve: a simplified form of the temperature curve giving only
the major points of maximum temperature, (f) Lower two curves: the record of
control scries (' i"i <rir production (solid line) and for length of life (dashed line).
The median dates of the twenty-eight generations produced from July 25th to Oc-
tober 5th are indicated on the horizontal axis ; the mean number of eggs produced
and the mean number of days lived by each generation are indicated on the vertical
axis.

MODIFICATIONS OF CULTURE MEDIUM 41

the upper line of the shaded chart and shows the highest records of the
summer plotted at intervals of approximately one week. The third
curve (c) shows the records of the control series (C) plotted on the
same basis as in the foregoing curves. On the vertical scale are plotted
the mean number of days of life (dotted-line curve) and the mean num-
ber of eggs (solid-line curve) for each generation. On the horizontal
scale is recorded the time, as in the upper curves. The records of the
controls are plotted by generations on their " Median Dates " ; that is,
on a date exactly midway between the date when the first animal of a
given generation was isolated as an egg and the date when the last animal
of that generation died.

A comparison of the three curves (a), (b), and (c), clearly demon-
strates a close relation between the records of the control animals and
the temperature. As the temperature became higher the animals became
more depressed, the lowest records of the controls for the entire ex-
periment being made immediately after the highest temperature of the
summer. The animals improved when the temperature became lower.
At the end of the summer, when the temperature stayed consistently
lower, they gradually recovered from their depression.

Analysis of the Data
I. Method.

Inasmuch as the temperature influenced the results in such a striking
way, valid comparisons of the different groups of animals could be made
only if the groups to be compared lived at the same time and were thus
subjected to the same temperature variations. An arbitrary method was
devised to meet this condition. This method was a day by day compari-
son made on a percentage basis with the control series C as the standard.-
The daily average record of the control series was taken as 100 per cent.
The averages for the other cultures on the same day were then expressed
as percentages of this. Thus if the average number of eggs produced
by H on a given day was one-half as great as the number produced by
the controls, H was given a rating of 50 per cent. In calculating the
averages for the controls, the generations as such were disregarded, since,
in an adequate medium, the number of generations cultivated on slides
had been shown to be unimportant (p. 38). In calculating the records
of H and Q, cultivated in the reduced media, each generation was con-
sidered separately.

The procedure for determining the percentage for each generation
of an experimental group was as follows, using H as an example : First,
the mean number of eggs for animals of a given generation of H was

1 This method was devised by Dr. Daniel Raffel of this laboratory.

42 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

computed by days on the basis of the day they were begun as individuals.
Then this mean for H was expressed as a percentage of the mean of C
for the same day. In this way percentages were calculated for each day

TABLE III

Experiment I, 1928 — Showing for the successive generations (Gen.) of the
four test series (H, Q, HC, QC) the numbers of animals (No.) and the percent-
ages of fecundity and longevity, computed as described in the text (pp. 41-43).
Data presented graphically in Figs. 4, 5, 6 and 7.

H Series (half-strength medium)

Q Series (quarter-strength medium)

Gen.

No.

Fecundity

Longevity

Gen.

No.

Fecundity

Longevity

per cent

per cent

per cent

per cent

1

96

28.16

64.06

1

100

2.34

34.89

2

93

38.22

75.78

t(2)

17

126.57

108.65

3

81

98.48

75.40

3

47

16.67

41.55

4

42

111.79

84.91

*4

17

40.98

82.81

5

44

85.69

94.54

5

32

70.47

82.68

6

62

80.67

92.01

6

48

55.61

71.99

7

42

115.65

95.60

7

49

35.85

62.92

8

41

69.01

70.67

*8

12

0.00

41.83

9

37

39.34

65.58

*10

16

24.72

59.22

*11

7

65.09

74.85

*12

10

46.88

67.60

*13

3

0.00

50.15

HC Series (full-strength medium)

QC Series (full-strength medium)

*1(10)

12

85.58

89.67

•1(8)

9

140.74

111.46

*2(11)

23

47.75

88.16

•2(9)

27

59.49

93.08

*3(12)

16

104.80

90.83

3(10)

30

83.79

100.47

*4(13)

21

190.96

118.52

•4(11)

22

51.21

78.81

5(14)

47

123.75

98.97

•5(12)

17

42.50

82.52

6(15)

48

115.89

94.68

*6(13)

9

46.99

87.35

7(16)

40

129.71

102.81

•7(14)

3

97.05

91.99

8(17)

40

114.86

110.12

•8(15)

2

261.62

155.04

9(18)

42

80.18

84.93

*9(16)

8

87.20

97.54

10(17)

14(23)

39

81.83

111.33

11(18)

No controls for comparison t

12(19)

16(25)

39

83.14

95.00

13(20)

54

79.06

93.03

18(27)

49

91.47

97.30

15(22)

47

70.49

92.62

17(24)

50

86.45

95.91

* Generations comprising less than 30 animals.
f(2) grown in full-strength medium.

t Gen. 17, 18, 19 of the QC scries were begun at a time when no control ani
mals were begun, so comparison on the percentage basis was impossible.

MODIFICATIONS OF CULTURE MEDIUM 43

included in that H generation. Finally, the percentages were weighted
according to the number of H animals included in each, and the true
mean of these percentages was found for each generation. Following
this method, the records of fecundity and longevity of each generation
of the experimental groups were calculated as percentages of the syn-
chronous records of the C group. The variation of these percentages
with the passing of generations is shown graphically in Fig. 4; here the
vertical axis represents the percentages, the horizontal axis represents
the generations, and the straight horizontal line at 100 per cent repre-
sents the standard of comparison, — the record of the C group.

A dependable interpretation of the data must take into consideration
the numbers of animals involved ; unless this number is fairly large, the
sample may not be representative. In this study, generations compris-
ing less than thirty animals have been considered of doubtful significance.
These generations are indicated by asterisks in Table III. The small
numbers in the later declining generations of H and Q were unavoidable,
since the fecundity of the depressed animals was very much reduced and
the rate of mortality was high. Fortunately, the records of these later
generations are relatively unimportant, as depression is clearly demon-
strated in the preceding, larger generations. But in the test of the
heritability of depression in the first HC and QC generations, the small
numbers available constitute a serious complication and must be taken
into account fully in the interpretation.

II. Depression: (a) Fecundity.

A comparison of the fecundity of the three groups of animals is pre-
sented in Fig. 4. The numbers of animals in the various generations of
the H and Q series and their fecundity percentages, computed as de-
scribed above, are given in Table III.

Both the H and Q groups showed some variation, but in general they
had a much reduced fecundity. The egg production of the first gen-
eration of the H series was very low ; approximately 25 per cent of the
C average. In the next six generations there was an improvement ; the
fecundity of the 4th and 7th generations of H was even higher than that
of the controls. But after that time there occurred a sharp decline, and
with two exceptions, each generation had a lower productivity than the
preceding one. The line died out in the 13th generation.

The Q series followed in general the same course as the H series, but
with an average fecundity consistently much lower throughout the ex-
periment. In the first generation grown in the weak culture medium,
the Q animals produced on the average less than three per cent as many
eggs as the control animals. In the next few generations they improved

44 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

and reached their maximum egg production in the 5th generation with
an average fecundity of about seventy per cent of the record of the con-
trols. From that time on there was a steady decline, each generation
having a lower fecundity than the previous one. The maximum depres-
sion was reached in the 8th generation with the complete dying out of
the line.

Subjecting Proalcs sordida to culture media of reduced strength for

no
lee

90
80
70
60
50
40
50
20
10

c A

/ \

A

/\ C

i \

/ \

( \

' ,

/

\

/

\

/
/
/

\ \ ,\

/

\ , \

/

\ / \

/H /

\ 1 V

V \
\

f

\

\

-Q.

. \

10 11 12 13

FK;. 4. Experiment I, 1928. Depression curve based on egg production.
Horizontal axis represents successive generations; vertical axis the percentages
with the record of the controls (C) as 100 per cent, and the records of H and Q
plotted with reference to that standard. In II the culture medium is half the
strength of that in the controls; in Q, it is one quarter of that strength.

many successive generations is thus shown to cause a decrease in fe-
cundity similar to that observed in other rotifers in adverse environ-
mental conditions, and similar in many respects to the lowering of fission
rate noted in poorly-fed Infusoria.

(b) Longevity.

There is very little in the literature on the effect of adverse conditions
on longevity. Most of the environmental studies on the Rotifera pre-
sent data dealing <mlv with the effect on fecundity, considered as an
index of vitality. For this reason the1 longevity data of these experi-
ments ;ire nt" special interest. Those for Experiment I are presented
graphically in Fig. 5 in the manner already described in the consideration
of fecundity. Table III gives a resume of the records for every gen-
eration.

The effect on the length of life of tin- i\vo underfed series was greater
than the effect on fecundity, discussed in the last section : this is especially
true of the H series. Neither the Q series nor the // series ever equalled
in average length of life the record of the C group; and the records of

MODIFICATIONS OF CULTURE MEDIUM 45

the three coexisting groups fall in precisely the same order, throughout
their history, as that of the strengths of their culture media.

In the H group, the first generation lived, on the average, about sixty-
five per cent as long as the contemporaneous members of the C group.
Generations 2 to 7 showed an increase in relative length of life, rising in
the 7th generation to about ninety-five per cent of the C record. Gen-
eration 8 showed a sharp decline which continued with two exceptions,
until the 13th generation, when the line died out.

In the Q group the average life of the first generation was about

90

80
70
60
50

40

30
20
10

c

c

^/T"\ \

/ ' '*• '

/vx.

'H / \ x,

/ \
•/ \

/

\

;'

\

. •** *

\

Q

\

\

\

\

\

\

9 10 U 12 13 14

FIG. 5. Experiment I, 1928. Depression curve based on length of life. Hori-
zontal axis represents successive generations ; vertical axis represents the percent-
ages with the record of the controls (C) as 100 per cent, and the records of H and
Q plotted with reference to that standard.

thirty-five per cent of that of the controls. The next few generations
showed an improvement, and in the 4th generation the average longevity
of the Q line practically equalled that of the 4th generation of the H
group. From then on, however, there was a rapid decline ; the line died
out in the 8th generation.

These results are diverse from those obtained on the effect of alcohol
by Noyes for Proalcs decipiens, and by Finesinger in most of his experi-
ments on Distyla incrmis. They did not find that the length of life was
appreciably decreased by subjection to alcohol, even though the depres-
sive effect on fecundity was very marked. Also, Finesinger found that
in very weak doses of alcohol, both fecundity and longevity were in-
creased. In the experiments here reported, on Proales sordida culti-
vated in inadequate media, both fecundity and longevity were strikingly
decreased.

In general, then, we may say that Proales sordida, when grown in
dilute culture media (quarter or half as strong as the optimum), and
under variable temperature conditions, after an initial immediate de-
pression, tends to adjust itself fur a few generations and pursue an up-

46 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

ward course, but eventually becomes progressively depressed and dies
out completely. Individuals grown in a quarter-strength solution have
a lower fecundity and longevity than those grown in a half -strength
medium, and in turn, individuals grown in a half-strength medium, for
the most part, have a lower fecundity and longevity than those grown in
a full-strength solution. In other words, the degree of depression and
the speed with which it occurs are, to a large extent, dependent on the
strength of the culture medium.

III. Recovery.

The question of the heritability of induced modification is of very
great interest; it forms the problem of chief concern in these experi-
ments. To test this, some of the progeny of each of the depressed
groups were placed in control medium (full-strength) and thc'ir sue-

rm

260
250

2«>

230

220
210
20O
190

ieo

170
160
150

uo

130
120
|llO

IK

90

•C
TO

60

30
?0
10

l\

* H 1 \ .

1 \

I \

l

A

I

\

u : .* :

\ c

'

: \

; ;

• . — '"

A ,

H\ ; '

0 : \ / v '

"•-'•''

V ; \ ' f.

6 9 10 11 12 1J 14 1$ 1ft 17 10

20 21 22 23 24 25 26 27

FIG. 6. Experiment I, 1928. Recovery curve based on egg production. Hori-
zontal axis represents successive generations ; vertical axis represents the percent-
ages with the record of the controls (C) as 100 per cent, and the records of H and
Q plotted with reference to that standard. Asterisks indicate the generations in
which return to normal medium occurred.

cessive generations were continuously cultivated in this medium. These
" recovery groups " ( HC and QC} were begun with eggs from the ninth
generation of // and from the seventh generation of Q. In the curves
showing the records of these groups (Figs. 6 and 7) the first genera-

MODIFICATIONS OF CULTURE MEDIUM 47

tions are marked with an asterisk. All computations were made on the
percentage basis described above, using the record of the controls as the
standard for comparison. ( )n the vertical scale are plotted the percent-
ages ; on the horizontal scale the number of generations. The record
of the controls is indicated by the solid straight line at 100 per cent.
The short solid lines preceding the dashed JIC line and the dotted QC
line represent the records of the generations of depressed animals (H
and Q) from which the recovery groups were begun. For the detailed
data on fecundity and longevity, and the numbers of animals comprising
each generation, consult Table III.

(a) Fecundity.

At the time when the return to normal medium was made, the H
series had an average fecundity about forty per cent of that of the con-
trol animals. As is shown in Fig. 6, the first generation in the normal
medium (generation 10, marked with an asterisk) showed a considerable
increase in productivity, rising to about eighty-five per cent. The next
HC generation (generation 11) fell to almost forty-seven per cent again.
The third HC generation (generation 12) improved markedly, with an
average egg production slightly higher than that of the control series;
and the next five generations, though showing a considerable fluctuation,
maintained an average fecundity consistently above that of the controls.
In the 9th generation in normal medium (18th generation of cultivation)
the HC group had an average fecundity of 80 per cent and stayed at
about that level for the remainder of the experiment.

The animals from the Q line which were placed in control medium
(full-strength) also showed great variation in fecundity. Ten of the
twelve QC generations had a fecundity lower than that of the controls.
The other two (first and eighth QC generations) showed a fecundity
much higher than that of the controls. The first QC generation showed
an increase from 36 per cent (the record of the preceding Q generation,
cultivated in quarter-strength culture medium) to about one hundred and
forty-one per cent, exceeding the control record by a wide margin. The
eighth QC generation (comprising only two animals) rose to 262 per
cent, more than doubling the record of the controls.

For a correct evaluation of the significance of these very irregular
records of the HC and QC series, the numbers of animals in the various
generations (given in Table III) must be taken into account. With at
most two exceptions the early generations of recovery are composed of
so few individuals as to render the significance of their data doubtful.
As far as they go, they indicate that recovery of egg production is imme-
diate and complete in the first generation grown in normal culture me-
dium.

4

48 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

Longevity.

On the whole, in both test lines, variation in length of life after the
return to normal culture medium was much less than variation in egg
production. The data are shown in Fig. 7, plotted on the percentage
basis as earlier described. The first generations in full-strength culture
medium are indicated by asterisks.

The average length of life of the first generation of the HC series

1U
150
120
110
'100
90

/ \ . X 'V

1 10 11 12 1J !» 15 It 17 18 19 » 21 „ „ „ „ !6 „

FIG. 7. Experiment I, 1928. Recovery curve based on length of life. Hori-
zontal axis represents successive generations; vertical axis represents the percent-
ages with the record of the controls (C) as 100 per cent, and the records of H and
Q plotted with reference to that standard. Asterisks indicate the generations in
which the return to normal medium occurred.

was considerably greater than that of the preceding H generation. This
increase in longevity was continued for several generations, the 4th gen-
eration of recovery ( 13th) exceeding the record of the controls. From
that point on, there was considerable variation, the average length of life
of the HC series fluctuating around the record of the controls. Here,
as in the case of fecundity, the fact that the early generations comprised
only small numbers of animals makes the significance of their data
doubtful.

In the QC series, the average length of life of the first generation
exceeded the record of the controls, indicating that recovery was imme-
diate and complete. In generations 2 and 3 (9 and 10) the longevity of
the (_>< series remained very close to that of the controls; but alter that,
except for one generation (15th) the average was slightly lower than
that for ( . Mere, also, the fact that only one (generation 10) of the
first nine generations of recover) in this group was significant from a
statistical point of view makes the results unreliable. However, the in-
dication is that the decreased length of life produced by insufficient food
is not inherited when the progeny are placed in an adequate environment.

MODIFICATIONS OF CULTURE MEDIUM 49

Discussion

The results of these experiments show that the Rotifera do respond
to an unfavorable environment as do the Infusoria; they become pro-
gressively depressed, show a gradual lowering of vitality, and eventually
die out. In the adverse condition used, an inadequate culture medium,
successive generations show this graded decline in both fecundity and
longevity.

The results bearing on the question of the heritability of induced
modifications, however, are not decisive, owing to their great irregularity.
It seems probable that this irregularity was largely due to the extreme
temperature variation and to the inconsiderable number of animals avail-
able in the critical generations. To control the temperature, and in the
light of the experience gained to test again the heritability of the depres-
sion, a second series of experiments was carried out. An account of
these follows.

PART II. EXPERIMENTS OF 1929

In 1929 the experimental work was conducted in a constant tempera-
ture room where fluctuations were very slight (20° to 22° C.). Varia-
tions from day to day were usually less than half a degree, and omitting
the last week of the experiment when considerable fluctuations occurred,
the total variation for the summer was less than two degrees.

Preliminary tests were made to find out whether cultivation under
uniform temperature necessitated any change in experimental procedure.
It was found that under constant temperature conditions rotifers reared
in the culture medium used for the controls in the former experiments
(15 flakes of oatmeal to 100 cc. spring water) had a very low fecundity
and longevity and appeared to be starved. Consequently, solutions of
16, 20, 24 and 32 flakes of oatmeal to 100 cc. of spring water were
tested. The 32 flake fluid was found to be the most satisfactory.

In Experiment II, therefore, the following culture media were used:

Control Series C, full-strength fluid, 32 flakes oatmeal to 100 cc. spring water
Test " H, half-strength " 16 "

Q, quarter-strength

In general these media were prepared as earlier described under
" Methods," except for the fact that the culture fluid was corked at the
end of three hours and placed in the constant temperature room, instead
of being corked at the end of twelve hours and subjected to variable
temperature. It will be observed that in Experiment II the half-strength
and quarter-strength culture media were twice as strong as the corre-
sponding media of Experiment I, and that the fluid for the control series
was more than twice as strong as the full-strength medium formerly

50 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

used. In all probability, tbe fact that the culture fluid was exposed to
the air for such a short time (three hours) explains the necessity for
increasing the number of flakes of oatmeal used.

General Description

This experiment was begun August 11. 1929, with a single series of
control animals which were cultivated on slides for three generations
before the test groups were started. These test groups, again called H
and Q, were begun with some of the eggs laid by the third generation of
this control series. In the 4th generation, therefore, there were three
series of animals: the controls (C) cultivated in thirty-two flake fluid;
the H series cultivated in sixteen-flake fluid ; and the Q series cultivated
in eight-flake fluid.

In general, Experiment II was conducted in the same way as Ex-
periment I, dealing again with the effects on fecundity and longevity
produced by continuous cultivation for many generations in weak culture
media. When the animals had shown a marked decrease in productivity
and length of life, the heritability of this environmental modification was
tested by placing progeny from both test groups in the control medium
of thirty-two flakes, and cultivating them in this full-strength medium
for several successive generations.

Analysis of the Data

Since the corresponding generations of all three groups of animals
occurred at the same time (except during the last part of the experi-
ment, when the control line was sufficiently in advance of the other two
groups to produce one more generation), and since the temperature con-
ditions were satisfactorily controlled, there was no reason for comparing
the groups of animals on any basis except that of corresponding genera-
tions. The results of the experiment are shown on this basis in Figs. 8
to 11. In these curves the horizontal axis represents the number of
generations, using 1, 2. 3 for the preliminary series, then re-numbering
1. 2, 3. etc., beginning with the first generation to comprise three groups
of animal>. The vertical axis represents either the mean number of
eggs or the mean length of life. The data for these curves will be
found in Table IV.

I. I )epres>ion.

In this experiment, conducted under constant temperature conditions,
and with stronger media than those used in Experiment I, the test lines,
// and Q. did not die out with the passage of generations. \Yhen the

MODIFICATIONS OF CULTURE MEDIUM

51

TABLE IV

Experiment II, 1929 — Showing for the successive generations (Gen.) of each
of the five series of the second experiment (Control, //, Q, HC, QC} the numbers
of eggs isolated (No.), the numbers of these that did not hatch (N.H.), and the
numbers that hatched (H.) ; and for the latter the mean numbers of eggs laid and
the mean numbers of days lived. Data presented graphically in Figs. 8, 9, 10, 11.

Control Series (full-strength medium)

Preliminary

Gen.

No.

N.H.

II.

Mean Eggs

Mean Days

1

24

0

24

20.75

10.00

2

21

0

21

16.95

8.66

3

39

1

38

17.28

8.89

Principal

1

48

1

47

13.02

7.67

2

52

0

52

12.34

8.17

3

41

3

38

14.00

9.13

4

40

1

39

15.38

9.01

5

40

0

40

13.87

8.17

6

54

3

51

14.31

7.93

7

30

3

27

10.48

7.62

8

34

0

34

9.23

7.26

9

30

2

28

9.53

6.94

10

35

4

31

10.19

6.30

11

29

0

29

8.13

5.79

12

28

0

28

6.46

5.32

13

30

7

23

6.26

6.19

14

39

12

27

9.03

7.35

15

32

10

22

8.59

9.00

H Series (half-strength medium)

1

50

1

49

9.97

7.34

2

52

1

51

8.25

7.58

3

43

2

41

5.63

6.76

4

39

0

39

6.74

7.55

5

40

4

36

6.97

7.09

6

43

2

41

6.04

6.71

7

51

5

46

5.21

6.71

8

42

4

38

5.63

6.82

9

41

1

40

6.22

6.13

10

39

3

36

4.36

5.97

11

42

3

39

3.82

5.33

12

42

9

33

3.54

6.12

13

36

6

30

5.53

6.71

14

27

4

23

8.52

8.50

HC Series (full-strength medium)

1(13)
2(14)

38
29

4

1

34
28

12.50
8.17

8.67
8.26

RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

TABLE IV — Continued

Gen.

No.

N.H.

H.

Mean Eggs

Mean Days

Q Series (quarter-strength medium)

1

41

0

41

5.19

5.74

2

38

2

36

2.75

5.13

3

43

7

36

1.33

3.68

4

31

0

31

4.29

6.51

5

51

1

50

5.46

6.59

6

48

4

44

3.00

5.30

7

40

1

39

2.66

5.66

8

40

4

36

1.83

5.00

9

42

3

39

1.64

4.24

10

62

11

51

1.39

4.00

11

63

9

54

2.33

4.93

12

54

5

49

3.51

5.72

13

28

0

28

2.10

4.60

14

34

4

30

1.53

5.05

QC Series (full-strength medium)

1(9)

24

1

23

9.43

6.04

2(10)

50

9

41

8.85

6.48

3(11)

41

16

25

6.72

6.18

4(12)

39

2

37

12.91

8.29

5(13)

27

4

23

10.39

7.50

6(14)

31

4

27

6.03

7.14

experiment was concluded they had lived fourteen generations and, while
showing some depression, they were reproducing well enough to make it
seem probable that they would live many more generations, if not in-
definitely.

(a) Fecundity.

Figure 8 presents graphically the data on egg production. It shows
that in general the records of the three groups of animals ran parallel to
each other, the half-strength (H) series regularly lower than the control
series, and the quarter-strength (Q) series regularly lower than the H
series. At first, the difference between the three groups was very great,
but at the end of the experiment this difference was much less, especially
between the // and C series. The control series showed considerable
variation, but its average fecundity gradually decreased as the experi-
ment progressed. The H line declined rapidly for the first three genera-
tions, and from then on continued to show a fairly constant though much
less extensive decline until the 13th and 14th generations. In these last

MODIFICATIONS OF CULTURE MEDIUM

two generations the H line showed a marked improvement, almost equal-
ling the average of the controls. The Q line likewise declined rapidly
when first subjected to the reduced culture medium, showing in the third
generation the maximum effect on fecundity, when the average egg pro-
duction fell to 1.33. The average fecundity stayed consistently low for

2 3 1 £ 5 4 5 6 7 8 9 10 11 12 13 14 It

FIG. 8. Experiment II, 1929. Depression curve based on egg production.
Horizontal axis represents successive generations ; vertical axis, mean egg pro-
duction.

the remainder of the experiment although showing considerable varia-
tion. It showed no improvement in its final generations such as the H
series showed; the 13th generation had an average fecundity of only 2.1
eggs and the 14th generation of only 1.5 eggs. The individuals were
small and thin.

Thus in this experiment, as in the former one, decreased food is
associated with a decrease in fecundity, the most marked decrease oc-
curring in the first three generations.

(b) Longevity.

The effects on longevity are presented in Fig. 9. The H series
showed a moderate and gradual decline for eleven generations, but then
began to improve; in the 12th, 13th, and 14th generations its average
longevity exceeded that of the controls of the corresponding generations.

In the Q line the effect of the adverse condition was more marked,

54 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

and for the first few generations the average length of life declined rap-
idly, reaching its lowest point in the third generation. From that time
on the Q line showed considerable variation hut remained consistently
below the other two groups except in the 12th generation, when the

J 2 ) 1, i 6 7 8 9 10 11 12 1J M 15

FIG. 9. Experiment II, 1929. Depression curve based on length of life.
Horizontal axis represents successive -nuTations; vertical axis, mean length of
life in days.

average length of life was slightly above that of C, but did not equal that
of//.

1 1 . Recovery.

The // recovery series was not begun until very late in the experi-
ment. \Ye wished to test progeny from the most depressed generation
and for that reason delayed beginning this series in anticipation of fur-
ther depression in //. which did not occur. The HC series was finally
started with eggs from the 12th H generation, a short time before the
experiment terminated, and so was carried through only two generations.
The Q recovery series was begun with eggs from the 8th generation in
the quarter-strength medium. This QC line was continued in normal
medium for six generations, when the experiment ended. The data on
these HC and QC lines is presented in Figs. 10 and 11. Asterisks indi-
cate the generations where the return to normal medium occurred.

( i/) Fecundity.

While the // line in the 12th generation had an average egg produc-
tion of 3.54, their progeny which were returned to the normal medium
had an average of 12.5, exceeding the average of the controls by 6.24
eggs. The .semi id (the last) HC generation fell to an average slightly
below that of the controls.

The Q—QC series increased from an average of 1.83 eggs in the 8th
generation in the quarter-strength medium to ().4.3 eggs in the first gen-
eration in control medium, practically equalling the controls of that gen-

MODIFICATIONS OF CULTURE MEDIUM

55

oration. They remained slightly he-low the controls for two generations
and then surpassed them hy an enormous margin in the 4th generation
(twelfth under cultivation). After this they again declined to a point
below the controls. The records of these two lines certainly provide

7 8 9 10 11 12 13 14 15

FIG. 10. Experiment II, 1929. Recovery curve based on egg production.
Horizontal axis represents successive generations; vertical axis, mean egg pro-
duction. Asterisks indicate generations in which return to 100 per cent medium
occurred.

conclusive evidence that the effect of insufficient food on fecundity is
not inherited.

(b) Longevity.

Figure 11 clearly shows that animals from both the test series, when
placed in full-strength culture medium, recovered a normal longevity at
once, their records even exceeding that of the control animals. The rec-
ord of the first HC generation exceeded by the same amount the record
of their parents (the last H generation) and the record of the corre-
sponding generation of the C series, and fell only slightly in their second
and last generation.

The longevity of the first generation of the QC line, grown in the
full-strength medium, almost equalled the record of the controls ; the
four generations following exceeded the controls, and the sixth (the
last) generation was only very slightly below the controls.

56 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

What has been said concerning the inheritance of decreased fecundity
is likewise true for longevity. Animals whose length of life has been
altered by long cultivation under adverse conditions produce progeny
that show a normal longevity when grown in a favorable environment.
Depression produced in this rotifer by the action of an inadequate cul-
ture medium lasts only as long as the stimulus which produces the de-
pression is present. This result differs a little from the results of \Yhit-

4
3-

2 •
1 -

9 10 11 12 13

15

FIG. 11. Experiment II. 1929. Recovery curve based on length of life.
Horizontal axis represents successive generations; vertical axis, mean length of
life in days. Asterisks indicate generations in which return to 100 per cent me-
dium occurred.

ney (1912) and Xoyes (1922). who found that the depression produced
by the action of alcohol persisted for two generations after removal to a
favorable medium.

Mechanism of tlic Effect

As to the precise way in which the environment operates to produce
depression, we have no evidence. 1 1 may be that in a medium deficient
in food, the rotifer grows at the expense of its own productivity; that
in the ordinary metabolic processes it uses up part or all of the yolk
^tore<l in the vitellarium, making the production of many eggs impossible.
thus bringing about a depression in fecundity. This process may be
correlated with the length of life of the organism to the extent that when
all the stored yolk has been used up in metabolism, the animal has no
more reserve on which to depend and consequently dies, thus bringing
about a lowering in longevity. Or, on the other hand, in such a weak
food medium the animal may use all the food it receives for its own
body metabolism and not have enough left to produce yolk for many
eggs, causing a reduction in fecundity but not in longevity, the animals

MODIFICATIONS OF CULTURE MEDIUM 57

having sufficient vitality to continue living for the usual length of time if
they do not use part of their energy in egg production.

It is possible that the amount of stored yolk is reduced in ill-fed ani-
mals, and that the eggs are smaller. If this were a cumulative process,
the eggs of consecutive starved generations becoming progressively
smaller, containing less and less yolk, the lines would eventually die out.
Measurements need to be made to determine whether there is any size
difference between the eggs of normal parents and the eggs of depressed
parents. As far as we could determine by superficial observation, the
eggs of all generations of all three groups were normal in every way.

No measurements were made on the animals in our experiments, but
some of the observations on comparative size and general appearance are
of interest. In the second experiment, particularly, the difference in
appearance between animals belonging to different lines was very strik-
ing. The controls were large and healthy in appearance from the first ;
the animals of the H series very soon became slightly smaller than the
controls, and quite transparent ; the Q individuals were extremely small
and thin. For several generations one could recognize at once to what
line an animal belonged merely by its size and type. As the experiment
proceeded, however, the H and Q animals improved in appearance to
such an extent that in the later generations they seemed as healthy and
normal as the control animals. In the last two generations some of the
control animals looked bloated and heavy, and died containing dark-
masses of material, and often one or two or even three well-developed
eggs which they had been unable to lay.

In both of the experiments, animals were observed with internal eggs
which they were unable to eject ; in numerous cases the egg developed
within the female and hatched within her body after her death. Some
of these young finally escaped from the parental corpses and lived there-
after as normal individuals, apparently not differing from animals that
had hatched normally. A single instance was observed where two such
young hatched out within the same female and later became liberated.
This behavior was not confined to one group or one generation, but in-
stances were observed in C, H, and Q, in the later generations of both
experiments.

SUMMARY AND CONCLUSIONS

This paper is a study of depression and recovery in the partheno-
genetic rotifer, Proalcs sordida Gosse, showing the response of the
organism to various dilutions of culture medium. All the animals used
were members of a single clone and hence were genetically alike. Oat-
meal infusions of half and quarter-strengths, as compared with the full-
strength medium used for the control animals, were employed. The in-

58 RUTH STOCKING LYNCH AND HELEN BERENICE SMITH

vestigation consisted of two series of experiments conducted separately
one year apart, together with some supplementary procedures. It was
found that with continuous parthenogenetic reproduction, rotifers in the
reduced strengths of culture media become depressed, decline in vitality
with the passage of generations, showing a much lowered fecundity and
longevity. Under conditions of variable temperature this decline is
cumulative and results in the ultimate death of the lines. Under con-
stant temperature conditions the animals show a cumulative decline in
the first few generations, reaching a low level of fecundity and longevity,
where they remain without further depression. The degree of depres-
sion and the speed with which it occurs depends on the strength of the
medium employed. This depression produced by means of an environ-
mental agent is similar in many ways to that produced bv the action of
adverse conditions in various Infusoria and in other rotifers. However,
in contrast to the results obtained by others in studies of the Rotifera.
longevity as well as fecundity is affected by the environmental agent.

The progeny of the depressed animals recover their normal fecundity
and longevity at once in the first generation placed in a favorable me-
dium. Apparently the food deficiency does not affect the germ cells so
as to produce a heritable modification.

The protozoan constitution seems to be extremely sensitive to en-
vironmental modifications and becomes changed to such an extent that
the alterations are passed on to the progeny of later generations. Some
of these changes are adaptive, so as to cause an increase in resistance to
certain agents, while others (such as those resulting from food insuffi-
ciency) are degenerative and result ultimately in the extinction of the
race. The situation in the Rotifera, as apparently in the other Metazoa,
is quite different. Though in Rotifera it is possible to produce and
maintain depression for a long period of time, and though the depression
becomes greater in later generations, it disappears at once, or within one
or two generations, upon restoration to normal conditions. The com-
plete agreement in this respect of the results of the present study with
those of \Yhitney, Noyes. and Finesinger on the injurious effects of
alcohol goes far to establish this as a definitive conclusion for the Roti-
fera. Injurious conditions acting on the individuals of many successive
generations and producing a progressive decline, do not alter the germinal
materials carried by these individuals.

It is possible that this difference between the Rotifera and the Proto-
zoa may be related to the differences in the details of their reproductive
pr< In tlie I'roto/oa the bodv divides into two equal parts; each

part becomes a new individual. In Proales, as in most Metazoa, a
smaller pieee separates from a larger. The larger piece is called the

MODIFICATIONS OF CULTURE MEDIUM 59

parent; the smaller the egg. This smaller piece goes through a twenty-
four-hour-long process of development, including numerous cell divi-
sions, into a new Proales individual. It does not seem obvious a priori
that this difference in detail must result in diverse genetic behavior.
These experiments on the heritability of an environmental effect indicate
that they do.

LITERATURE CITED

BEERS, C D., 1925. Encystment and Life Cycle in the Ciliate Didinium Nasutum.
Proc. Nat. Acad. Sci., 11: 523.

BEERS, C. D., 1926. The Life Cycle in the Ciliate Didinium Nasutum with Ref-
erence to Encystment. Jour. Morph., 42: 1.

FINESINGER, J. E., 1926. Effect of Certain Chemical and Physical Agents on Fe-
cundity and Length of Life, and on their Inheritance in a Rotifer, Lecane
(Distyla) inermis (Bryce). Jour. Expcr. Zool., 44: 63.

JENNINGS, H. S., AND LYNCH, R. S., 1928. Age, Mortality, Fertility and Indi-
vidual Diversities in the Rotifer Proales sordida Gosse. I. Effect of Age
of the Parent on Characteristics of the Offspring. Jour. Expcr. Zool.,
50: 345.

JENNINGS, H. S.. AND LYNCH, R. S., 1928. Age, Mortality, Fertility, and Indi-
vidual Diversities in the Rotifer Proales sordida Gosse. II. Life-history
in Relation to Mortality and Fecundity. Jour. Expcr. Zool., 51: 339.

JENNINGS. H. S., 1929. Genetics of the Protozoa. Bib!. Gcnctica, 5: 105.

JOLLOS, V., 1921. Experimented Protistenstudien. I. Untersuchungen iiber
Variabilitat und Vererbung bei Infusorien. Arch. f. Protistcnk., 43: 1.

LAMBERT, W. V., RICE, W. S., AND WALKER, H. C. A., 1923. Food and Partheno-
genetic Reproduction as related to the Constitutional Vigor of Hydatina
Senta. Biol. Bull., 44: 192.

LUNTZ, ALBERT, 1926. Untersuchungen iiber den Generationswechsel der Rota-
torien. I. Die Bedingungen des Generationswechsels. Biol. ZcntralbL,
46: 233.

MIDDLETON, A. R., 1918. Heritable Effects of Temperature Differences on the Fis-
sion Rate of Stylonychia Pustulata. Genetics, 3: 534.

NOYES, B., 1922. Experimental Studies on the Life-History of a Rotifer Repro-
ducing Parthenogenetically (Proales decipiens). Jour. Ex per. Zool., 35:
225.

SHULL, A. F., 1912. The Influence of Inbreeding on Vigor in Hydatina Senta.
Biol. Bull., 24: 1.

SHULL, A. F., 1912. Studies in the Life Cycle of Hydatina Senta. III. Internal
Factors Influencing the Proportion of Male-Producers. Jour. E.rpcr.
Zool., 12: 283.

WHITNEY, D. D., 1912a. Reinvigoration Produced by Cross Fertilization in Hy-
datina Senta. Jour. Ex per. Zool., 12: 337.

WHITNEY. D. D., 1912/?. Weak Parthenogenetic Races of Hydatina Scuta Sub-
jected to a Varied Environment. Biol. Bull., 23: 321.

WHITNEY, D. D., 1912r. The Effects of Alcohol not Inherited in Hydatina Senta.
Am. Nat., 46: 41.

WOODRUFF, L. L., 190P. Further Studies on the Life Cycle of Paramecium. Biol.
Bull, 17: 287.

THE ROLE OF THE BASAL PLATE OF THE TAIL IN

REGENERATION IN THE TAIL-FINS OF FISHES

(FUNDULUS AND CARASSIUS)

S. MILTON NABRIT
BIOLOGICAL LABORATORY, MOREHOUSE COLLEGE

Experiments on the tails of fishes from a standpoint of morpho-
genesis have been performed by Morgan (1900. 1902 and 1906), and
by the writer (1929). A review of the literature may be found in
Nabrit (1929).

The writer suggested in the previous paper, after comparing the
findings of Harrison (1918) and Detwiler (1918) in Amblystoma with
the results of his experiments in fishes, that a possible similarity existed
between the production of limbs in Amblystoma and tail-fins in fishes.
Each one seems to be an independently differentiating mesenchymal
system.

It was further concluded that the rate of regeneration from cut sur-
faces in the tail-fins of fishes is regulated by the cross-sectional area of
the fin rays exposed.

Broussonet (1786), Morgan (1906), and Morrill (1906) agreed
that rav stumps must be left for regeneration to take place. Morrill
suggested that regeneration does not take place when the remaining
stumps are too small. This point was left open for further investigation
by the writer and is considered in this paper.1

The experiments were carried out during the summer of 1929 at the
Marine Biological Laboratory at Woods Hole, Massachusetts, and dur-
ing the winter of W29-1930 at Morehouse College. At Woods Hole,
/•' mid ul us hctcroclitus was used and at Morehouse College, the goldfish,
Carassius auratus. In both cases, adult animals were used.

Rays were carefully picked from the tails by fine forceps. The tail
and the rays were examined under a binocular microscope to make sure
that no parts of the rays were left embedded in the tails.

After regeneration was complete, examinations were made upon the
living and then upon the fixed tails.

In three weeks after removing the fin rays, undilTerentiated mesen-

1 I wish to express my sincere thanks to Professor J. Walter Wilson of
Brown University for his very helpful suggestions and criticisms in this investi-
gation.

60

REGENERATION IN TAIL-FINS OF FISHES

61

chyme had filled in the gap made in the tails. In from four to six weeks,
rays appeared in the regenerated mesenchyme (a, Figs. 1 and 2). These
rays begin at the basal plate and differentiate distally. They seg-
ment and may branch as any other rays in the tail, though some rays do
not branch (b, c. Figs. 1 and 2). If they branch at all it is usually dur-

FIG. 1. Carasshis. Diagrammatic sketch of the base of the tail,
o. A newly regenerated ray (six weeks).

b. A ray six months after it started regeneration. It did not branch.

c. A regenerated ray after six months.

d. Basal plate.

e and e' ' . Proximal end of fin ray.

ff'. Region of a cut anterior to proximal end of fin rays.
y. Articulating region of basal plate.

ing the first two months. The anterior knobs or proximal ends of the
rays are completed in from five to six months (c, c' , Figs. 1 and 2).
The proximal ends are produced on both sides of the basal plate.

In the embryonic development of the tail in Fiindnlus, the rays streak
out from a common mesenchymal plate. They begin to appear between
the ninth and thirteenth days. From two to four appear at a time, dor-
sally and ventrally placed in respect to the previously formed rays.

When a cross-cut is made in the tail anterior to the proximal knobs
of the fin rays (ff' , Figs. 1 and 2), regeneration does not readily take
place. This led to the conclusion that the ray stumps must be left in
the tail for regeneration of the tail to occur.

When rays are picked out of the tail, they readily regenerate; the

62

S. MILTOX NABRIT

rate, however, is slower than it would be if ray stumps were present.
It would appear, therefore, that no parts of the rays are necessary for
regeneration.

Xew rays do not begin to appear at their original most proximal
point, but at the distal end of the basal plate. Physically this is neces-
sary unless two separate anlage are developed for each ray. as each ray
has a component on each side of the basal plate.

When a cross-cut is made anterior to the proximal knobs of the fin
rays, the cut also includes the anterior articulating portion of the basal

FIG. 2. 1-uudnlus. As al.ovc.

a. New ray, three and one-half weeks.
/'. A ray, six and one-half weeks.
c. A normal ray.
(/. Uasal plate.
c and c'. Proximal end of fin ray.

ff'. Region of a cut anterior to proximal end
g. Articulating region of the base of tail.

>f fin rays.

plate. In Mich a case the delay in regeneration would be so great that
the animal would be cast aside, even if, as is usually the case, the flesh
did not slough off and lead to death.

A cut ray will regenerate those parts distal to the level of the cut.
An anterior knob regenerates the entire ray. Axial heteromorphosis on
the knot) has not occurred. I'ut when an anterior knob is severed or
removed from the distal stump, axial heteromorphosis occurs in the dis-

REGENERATION IN TAIL-FINS OF FISHES 63

tal portion. When a ray is picked out, on the other hand, the new distal
portion appears at the end of the articulating portion of the basal plate.
Now instead of getting axial heteromorphosis with this distal portion,
there is added a normal anterior knob. The visible difference in these
two cases is that this new ray is in connection with the basal plate. It
thus seems evident that the basal plate gives rise to the parts distal to it,
or induces their development. This suggestion gains further support
from the fact that the articulating part of the basal plate is shaped like
the tail in Carassitis and in Funduhts, bilobated in the former and
rounded in the latter, and may still retain the original formative influ-
ences which cause the differentiation of the fin rays and hence the size
and form of the tail.

It seems that a further similarity is demonstrated between the mor-
phogenesis of tail-fins of fishes and the Amblystonia limbs. The articu-
lating plate, like the girdle, may give rise to parts distal to it. In regen-
eration as in embryonic development, the basal plate gives rise to or
induces the development of rays. The rays, like the limbs in Ambly-
stonia, and the basal plate, like the girdle, belong to a self -differentiating
mesenchymal system.

SUMMARY

1. By picking out the fin rays instead of cutting anterior to them, it
is shown that ray stumps are not necessary for regeneration in the tail-
fins of fishes.

2. When the stumps are removed it seems that the new rays appear
under the influence of the articulating portion of the basal plate of the
tail.

3. From the embryological development and mode of regeneration a
similarity is demonstrated between morphogenesis of limbs of Ambly-
stoma and the tails of fishes, each being a self-differentiating mesen-
chymal system.

LITERATURE CITED

BROUSSONET, M., 1786. Observations sur la regenerations de quelques parties
du corps des poissons. Historic de I'Acad. Roy. des Sciences.

DETWILER, S. R., 1918. Experiments on the Development of the Shoulder Girdle
and the Anterior Limb of Amblystoma punctatum. Jour. Exper. Zool.,
25: 499.

HARRISON, R. G., 1918. Experiments on the Development of the Fore Limb of
Amblystoma, a Self-differentiating Equipotential System. Jour. Exper.
Zool. ,25: 413.

MORGAN, T. H., 1900. Regeneration in Teleosts. Arch. Entiv.-mcch., 10: 120.

MORGAN, T. H., 1902. Further Experiments on the Regeneration of the Tail of
Fishes. Arch. Entzv.-mech., 14: 539.

MORGAN, T. H., 1906. The Physiology of Regeneration. Jour. Exper. Zool., 3:
457.

MORRILL, C. V., 1906. Regeneration of Certain Structures in Fundulus heteroclitus.
Biol. Bull., 12: 11.

NABRIT, S. M., 1929. The Role of the Fin Rays in the Regeneration in the Tail-
Fins of Fishes. Biol. Bull., 56: 235.
5

RELATIVE RESISTANCE OF FOURTEEN SPECIES OF

PROTOZOA TO THE ACTION OF CROTALUS

ATROX AND COBRA VENOMS

CHARLES H. PHILPOTT

(From the Department of Bacteriology and Immunology of the Washington
University Medical School and the Marine Biological Laboratory,

Woods Hole, Mass.)

Essex and Markowitz (1930) have offered data which lend sup-
port to the conception that Crotalus venom is a non-specific protoplasmic
poison that affects protoplasm wherever it comes in contact with it.
This is in contrast with the view that snake venom is a complex of
various constituents, each specific for certain kinds of cells. In a re-
cent publication (1930) I have presented data, on eleven species of
Protozoa, which indicate a wide difference in the degree of suscepti-
bility to the effect of Crotalus atrox venom. Some species were not
affected by the concentration of the venom used (0.00025 gram of dry
venom per cc. of medium), while some were killed quickly. No at-
tempt was made, however, to determine whether the races of Protozoa
found to be resistant would be affected by higher concentrations of the
venom. In the light of the conclusions of Essex and Markowitz con-
cerning the general nature of the poisonous effects of Crotalus venom
it has seemed advisable to continue the investigation of the effect of
venoms on Protozoa to determine if any species is wholly resistant to
their destructive action. The investigation is continued also as a phase
of a more extended study, which is in progress, of the resistance which
races of Protozoa acquire to venoms.

For each of the fourteen species of Protozoa used in this investiga-
tion a determination has been made of the least concentration of Cro-
talus atrox and of Cobra venom required to kill the animals within one
hour. This, for each species, is called the minimal lethal concentration.
The method used in determining this minimal lethal concentration is the
same as that described in a previous publication (1930). The minimal
lethal concentration of each venom for each of the species of Protozoa
is shown in the accompanying table.

An inspection of the data shown in the table reveals the fact that
the venoms at some concentration have a lethal effect on each of the
species of Protozoa used. It is evident, however, that the Protozoa

64

RESISTANCE OF PROTOZOA TO VENOMS

65

Minimum Lethal Concentration of Crotalus atrox and Cobra Venoms
for Fourteen Species of Protozoa

Minimal Letha

Concentration

Species of Protozoa

Cobra Venom,
gm. per cc.

Crotalus atrox Venom,
gm. per cc.

Amoeba dubia

000008

001500

Blepharisma undulans

000015

000100

Chilomonas paramecium . . . .

ooooio

000300

Coleps hirtus

oooo^o

000900

Colpidium campylum

.000006

.003000

Euglena gracilis

000040

000200

Euplotes patella

.000010

.000150

Oxytrichia fallax

.000040

.000150

Paramecium multinucleatum

.000010

.000050

Prorodon teres

000025

.000300

Stentor polymorphic

.000200

.000100

Trachelius ovum

.000060

.000900

Urocentrum turbo

.000004

.000100

Vorticella sp

.000100

.002000

exhibit a great variation in respect to the degree of the resistance of
each species to the venoms. All species are more resistant to Crotalus
atrox venom than to Cobra venom. Some species (Urocentrum, Para-
mecium, Euplotes, Blepharisma) are, in relation to other species, highly
susceptible to both venoms. One species (Vorticella') has relatively
great resistance to both. Some species (Colpidium, Amceba, Chilo-
monas, Coleps and Prorodon) are highly resistant to Crotalus venom
but quite susceptible to the effects of Cobra venom. On the other
hand, one species (Stentor) is, in relation to the other species, highly
resistant to Cobra venom but susceptible to the action of Crotalus
venom. The same is true to some extent with Trachelius.

Repeated titrations made in connection with previous studies have
shown that the minimal lethal concentration for Paramecium multi-
nucleatum of Cobra venom and of Crotalus atrox venom is respectively
0.0000016 gram and 0.00002 gram per cc., while in this study it is, re-
spectively, 0.00001 gram and 0.00005 gram per cc. This indicates that
the stock solution of Cobra venom used in this investigation was ap-
proximately one-sixth normal strength, while that for Crotalus atrox
venom was two-fifths of its normal strength. Repeated tests were
made with Paramecia to show that the stock solutions of the venoms
used throughout this study did not deteriorate in strength during the
time. The stock solution in each case consisted of 0.05 gram of dry
venom dissolved in 4.5 cc. of distilled water to which 0.5 cc. of glycerine
was added. The stock solutions were rendered neutral by the addition
of Na2HPO4.

66 CHARLES H. PHILPOTT

It may be concluded that, so far as this group of Protozoa is con-
cerned, the poisonous effect of Crotalus atrox venom and of Cobra
venom on Protozoa is general. There exists among the species, how-
ever, much variation in the resistance to the action of the two venoms,
some species having relatively high resistance to the one and low to the
other, while, in other species, the reverse relationship exists.

REFERENCES

ESSEX, HIRAM E., AND MARKOWITZ, J., 1930. The Physiologic Action of Rattle-
snake Venom (Crotalin), IV. The effect on lower forms of life. Am.
Jour. Physiol., 92: 342.

PHILPOTT, CHARLES H., 1930. Effect of Toxins and Venoms on Protozoa. Jour.
E.rper. Zool, 56: 167.

A DETERMINATION OF THE TENSION AT THE SURFACE
OF EGGS OF THE AXXELID, CH&TOPTERUS

E. NEWTON HARVEY

(From the Marine Biological Laboratory, Woods Hole, Mass., and the Physio-
logical Laboratory of Princeton University, Princeton, New Jersey)

By means of the microscope-centrifuge, recently described by Har-
vey and Loomis (1930), it is possible to observe, and to photograph,
using highest dry objectives, living cells while subjected to centrifugal
forces thousands of times gravity.1 The cells are mounted on a special
slide fixed over the objective, which rotates on the centrifuge at a dis-
tance of 11 cm. from the axis of rotation. By total reflecting prisms
the image is brought to the axis of the centrifuge and then vertically
upward where it is observed with a stationary ocular. The light, a
condenser discharge in mercury vapor, is arranged for stroboscopic illu-
mination, flashing on with each revolution of the centrifuge. The im-
age is perfectly clear and steady at 4000 R.P.M.

CJicetoptcrus eggs, as is well known from the work of F. R. Lillie
(1909), pull apart into fragments under the influence of high centrifugal
forces. The centrifuge-microscope allows us to observe clearly the
series of changes in this process and to make exact measurements from
which it is possible to calculate the tension at the surface of the egg.
This process is illustrated for unfertilized Chcetopterus eggs by the
photographs of Fig. 1.

The first three rows show ten stages in the formation of a fragment
or spherule containing oil, taken through the microscope-centrifuge while
revolving about 4000 R.P.M., 5 seconds' exposure. It will be noted
that No. 5 is a double exposure (by mistake) and shows the exact
change in position which takes place in 1 minute. At one stage the
protoplasm is drawn out into long filaments with the oil spherule at one
end. In the pulling off of this oil spherule the appearance is that of a
drop of molten glass slowly falling from a heated rod. When the oil
finally breaks away, it moves slowly (as long as it can be observed in

1 The microscope-centrifuge was devised in collaboration with Mr. Alfred L.
Loomis and constructed in his private laboratory at Tuxedo Park, New York.
The author expresses his deep appreciation to Mr. Loomis for the hospitality of
the laboratory. It is contemplated that the instrument will be placed on the mar-
ket by the Bausch and Lomb Company. For description see Harvey and Loomis
in Science, 1930, 72: 42.

67

T

FIG. 1. Photographs of unfertilized eggs taken with a Leica camera through
the microscope-centrifuge (radius 11 cm.), showing successive changes (examine
from right to left) in stratification of granules and in shape of eggs. Magnifica-
tion, 87 diameters. The direction of the centrifugal force is downward.

First three rows, Chcctopterus exposed for 5 seconds at 35, 55, 95, 165, 205
and 265 (double exposure), 345, 375, 450, and 480 seconds after starting centrifuge.
Speed of the microscope-centrifuge was 55 R.P.S. in the first row and 66 R.P.S.
in the second and third rows. At the highest speed about 330 flashes of light oc-
curred during tin exposure.

Fourth row, (.'nmingia exposed 10 seconds at 35, 95 and 155 seconds after
starting ccntrifuiM . Speed 38 to 55 R.P.S. The eggs are separated by jelly.

Fifth row, Arbacia exposed 10 seconds at 45, 150, and 285 seconds after start-
ing centrifuge. Speed 35 R.P.S. The eggs are partially separated by jelly.

TENSION AT SURFACE OF CPLETOPTERUS EGGS 69

the field of the microscope) to the surface of the sea water. Some-
times a clear fragment containing no yolk or oil globules will separate.
These clear fragments are denser than sea water and sink; in fact, a
clear fragment, one half containing oil globules, is heavier than sea
water.

The pulling away of an egg spherule containing oil occupies about
seven minutes at 1916 times gravity, and it will be noted that when an
oil spherule breaks away the stalk does not round up immediately but
remains for some time and only slowly rounds off. The stalk does not
behave as if it were elastic like rubber, which would immediately re-
cover when stretched, nor is it stiff, for one can observe these stalks
waving back and forth in currents in the sea water during centrifuging.
They appear like a very slow flowing plastic material such as tar.

The stalk is part of the original surface of the egg. Are we to
regard that as having purely a surface tension at its boundary or a
membrane of some sort with an elastic tension and a definite breaking
strength, or should we think of the breaking strength of the stalk per
cross sectional area? Perhaps the magnitude of this tension will allow
us to decide. If we determine the buoyant force of the oil tending to
draw apart the egg into fragments, we can equate this to a tension of
the egg considered as acting only around the surface (surface tension
or elastic tension of a membrane) or as acting over the whole cross
sectional area of the stalk (internal friction of a plastic fluid), or a
combination of the two.

Fortunately, it is easy to gain a rough idea of the buoyant force,
which tends to pull apart the egg, by the following considerations. A
maximum value for this force will be obtained if we consider oil rising
through egg fluid rather than egg spherule rising through sea water.
Let us assume the density of the oil is 0.915, the same as that of hen's
egg yolk oil. By measuring the diameter of the oil spherule (34 /A)
pulled away from the egg, we find that the volume of the oil is 75 per
cent2 of 2.05 X 10'8 cc. or 1.54 ; ; 10'8 cc., and its mass is 1.41 X 10'8
grams. The density of the medium can be taken as 1.044, the average
of Heilbrunn's (1926) values for Arbacia and Cumingia eggs, and
slightly greater than sea water (1.0238) at Woods Hole at 20° C.
Therefore, the buoyant force of the oil pulling inward toward the axis
of the centrifuge will be ( 1.044 --0.915)/0.915 ) ; 1.41 ; ; 1Q-8 = 0.2
X 10-8 grams X 980 X centrifugal force (1916 X g) = 375 X 10'3
dynes. Since the yolk rests against the bottom of the slide, this will
be the only force tending to pull the egg apart, and it will be counter-

2 Since the oil granules are spheres and spheres packed in a volume occupy
only 75 per cent of that volume.

70 E. NEWTON HARVEY

acted by the surface tension or elastic tension (T} of the egg, acting
around the circumference (irD) of the stalk of diameter (D) connecting
oil spherule and egg at the moment the stalk hreaks.3 Hence -n-DT =
3.75 X 10 3 dynes.

The diameter of the stalk is found to be 9 /x by measurement. Solv-
ing, T - - 1.32 dynes per cm. This is a maximum value, as it is likely
the oil might have been pulled away with somewhat lower centrifugal
forces, although it takes 7 minutes to pull apart at 1916 X g. One
dyne per cm. cannot be very far from the truth, perhaps 25 per cent
error at most. If we are to regard this as a surface tension, it is very
low, but might be compared to that of a water — isobutyl alcohol inter-
face whose T is 1.76 dynes per cm. at 20° C. On the other hand, the
value seems entirely too low for "an elastic membrane tension.

If we regard the stalk of Chcetoptenis eggs as made up of homogene-
ous material, its breaking strength, S, would be (-rr/4)D2S = 3.75 X
10~3 dynes or .5* = = 5900 dynes per cm.2, also an extremely small value.
The breaking strength of rubber is about 100 million dynes per cm.2

The long cylindrical stalk of centrifuged Chcetoptenis eggs would be
a very unstable figure according to classical surface tension interpreta-
tion. Plateau showed that a cylinder breaks into drops when its length
equals its circumference. This is a geometrical relation, independent
of the value of the surface tension, provided the fluids have a low vis-
cosity. In the derivation of surface tension formulas by the vibrating
drop and vibrating jet methods the viscosity of the fluid is neglected for
simplification. With very high viscosities the attainment of equilibrium
conditions might take so long a time as to make thin cylinders stable
figures for all practical purposes. With plastic fluids surface tension
might be too small to overcome the internal friction.

The value of 1.3 dynes per cm. does not allow us to decide definitely
whether the boundary of the egg exerts a surface tension or an elastic
tension, but it does show that there can be no very firm "pellicle"
around these eggs. The observed behavior of the stalk makes it quite
certain that we are dealing with a very plastic material. If the tension
were greater than fifty dynes per cm., a maximum value for liquid-water
interfaces, we might say quite definitely that it was elastic membrane
and not surface tension. Such may be the case in eggs like Arbacia
and Cumingia which are not torn apart by much higher centrifugal
forces, and show only a slight tendency to lengthen, as illustrated in
rows 4 and 5 of Fig. 1.

3 I am deeply indebted to Dr. Charles Zahn of the Physics Department, Prince-
ton University, for suggestions regarding tin- interpretation of some of these sur-
face tension phenomena.

TENSION AT SURFACE OF CH/ETOPTERUS EGGS

SUMMARY

71

A method is described, using the microscope-centrifuge, for calcu-
lating the tension at the surface of Chcctoptcrus eggs, which gives an
approximate maximum value of 1.3 dynes per cm.

LITERATURE

HARVEY, E. N., AND LOOMIS, A. L., 1930. Science, 72: 42.
HEILBRUNN, L. V., 1926. Jour. E.rper. ZooL, 44: 255.
LILLIE, F. R., 1909. BioL Bull., 16: 54.

THE EFFECT OF CERTAIN NARCOTICS (URETHANES) ON
PERMEABILITY OF LIVING CELLS TO WATER

BALDUIX LUCKfi

(From the Laboratory of Pathology, School of Medicine, University of Pennsylvania,
Philadelphia, Pa., and the Marine Biological Laboratory, Woods Hole, Mass.)

The experiments on narcotics reported in this paper were carried out
with the purpose of gaining further information on the factors which
regulate or influence permeability of the living cell to water. Former
studies on sea urchin eggs have shown that cell permeability to water
is not constant, but varies with a number of factors, an important one
being the chemical composition of the medium (McCutcheon and
Lucke, 1928). Thus permeability is regulated, at least in part, by the
sign and the number of charges on the ions of the medium, in the sense
that anions increase and cations decrease permeability to water; the
effects are the greater the higher the valence of the ions (Lucke and
McCutcheon, 1929). In the presence of cations of two or more
valences permeability is of very low magnitude. For example, in
hypotonic sea water cells, such as unfertilized eggs of the sea urchin,
have a numerical value of permeability of approximately 0.05 at 15° C.,
i.e., 0.05 cubic micron of water pass through each square micron of
cell surface per minute, under the driving force of one atmosphere of
osmotic pressure.1 This surprisingly low degree of permeability is
presumably due to the high concentration of calcium and magnesium
in sea water, for the same low value of permeability is obtained when
cells are placed in a non-electrolyte medium, of like osmotic pressure,
containing as little as 0.0001 molar CaCl2 or MgCl2. But all attempts
to further decrease permeability have so far been unsuccessful. How-
ever, several writers have reported that narcotics decrease permeability
to water (Winterstein, 1916; Lillie, 1918; and Anselmino, 1928).
The question therefore arose whether narcotics in the presence of sea
water, or narcotics in any solution containing cations of two or more
valences, would lower permeability beyond the value obtained in sea
water alone.

METHOD

A satisfactory method of studying quantitatively the effect of
various factors on permeability to water is to place a suitable cell in a

1 By permeability to water is understood the quantity of water passing through
unit area of cell surface in unit time under unit pressure.

72

EFFECT OF URETHANES ON CELL PERMEABILITY 73

hypotonic medium, thus causing water to enter the cell under the
driving force of osmotic pressure. The spherical unfertilized egg of the
sea urchin, Arbacia punctulata, is an excellent natural osmometer
(McCutcheon and Luck6, 1926; McCutcheon, Lucke and Hartline,
1931 a and &). When placed in a hypotonic solution it swells relatively
slowly, thus permitting accurate measurement of its diameter, from
which volume and surface area can be calculated ; it has a high degree of
semipermeability, retains its spherical shape during change of volume,
and allows ready determination of injury which may have taken place
during the experiment.

The narcotics selected for this study are certain urethanes and
carbamates.2 These compounds have the advantage of being non-
injurious over a considerable range of concentration; they are non-
volatile, penetrate almost instantaneously, and their narcotic effect is
easily demonstrable with the material used.3

In the first series of experiments the narcotics were dissolved in
ordinary (100 per cent) sea water. Unfertilized eggs of Arbacia were
exposed to these solutions for from 5 to 10 minutes. The cells were
then transferred to hypotonic sea water (usually 40 per cent sea water)
with the narcotic to be tested in the same concentration as in the iso-
tonic solution. The course of inflow of water at constant temperature
(15 ± 0.5° C.) was then observed by measuring the diameter of three
cells at minute intervals with a filar micrometer for six successive
minutes.4 The mean volume of the cells was plotted against time and
a smooth curve drawn through the points. The rate of passage of
water is given by the rate of change of volume, dV/dt, and is obtained
from the slope of the curve at a given time, /. The numerical value of
permeability to water is calculated from the equation :

Permeability = -~ /S(P - Pex),

where S is the surface area of the cell and (P — Pex) the difference in
pressure between the interior of the cell at time t and the medium.6

2 For the sake of brevity the compounds will be referred to in this paper as
urethanes. The propyl urethanes were not commercially obtainable, and were
kindly supplied by Dr. Ralph Major, then of Princeton University.

3 The narcotic effect of these compounds will be dealt with by Dr. E. B. Harvey
in a forthcoming paper.

4 The technic of measuring volume changes in Arbacia eggs by means of a filar
micrometer eye-piece has been previously described (McCutcheon and Lucke, 1926).

6 For details of calculation see McCutcheon and Lucke, 1928, 1929, and Lucke,
Hartline and McCutcheon, 1931&.

74

BALDUIN LUCRE

EFFECT OF URETHANES WHEN DISSOLVED i\ SEA WATER

Before determining the effect of these narcotics on permeability a
preliminary question needed to be answered: Do these compounds
affect the volume of cells in equilibrium with either an isotonic or
hypotonic medium? If they should not enter practically instantane-
ously the volume of the cells would decrease when transferred to iso-
tonic sea water containing a considerable concentration of the narcotic
(e.g., 0.2 m. ethyl urethane). On the other hand, if these compounds do
enter rapidly, but cause severe injury or death, the volume of the cell
in equilibrium with a hypotonic solution might be either too small
(from escape of dissolved substances) (Lucke and McCutcheon, 1930)
or too large (from splitting of substances in the interior of the cell)
(Lucke and McCutcheon, 1926). Now, in order to use the simple
equation for permeability given above it is necessary that the equilib-
rium volume of the cells in the solution to be tested should be the
same as the volume at equilibrium of control (unnarcotized) cells in
both isotonic and hypotonic concentrations of their natural medium,
sea water. This question was answered as follows:

Different lots of cells from the same animal were placed in sea water
in which had been dissolved a narcotic in the concentrations shown in
Table I. After 5 to 10 minutes' exposure a number of cells were meas-

TABLK I

Effect of Urethanes on Cell Volume at Equilibrium

The concentrations shown are the highest used in the experiments. V0 is the
mean volume of 25 cells measured after from 5 to 10 minutes' exposure to the narcotic
solution in isotonic (100 per cent) sea water, Fe is the mean volume of 25 cells in
equilibrium with a given hypotonic solution in which were dissolved the different
urethanes (measured 4 hours after transfer). It is seen that the volume of cells in
the narcotic solutions corresponds closely with the volume of control cells in the same
concentration of sea water. The figures must be multiplied by 100 to obtain volumes
in cubic micra.

Solution

Vo

re

First sea water control

1908

3?55

Second sea water control

1882

3230

Ethyl urethane 0.2 m

1933

3182

n-propyl urethane 0.1 in

1920

3230

i-propyl urethane 0.1 m ...

1876

3182

n-butyl carbamate 0.05 m

1884

3275

Phenyl urethane 0.002 m

1924

3273

tired and then transferred to a hypotonic solution containing the same
concentration of the urethanes. The volume at equilibrium was

EFFECT OF URETHANES ON CELL PERMEABILITY 75

determined after 4 hours' exposure at 22° C. Table I shows the result
of a representative experiment. It is seen that there is no change in
volume of the narcotized cells in isotonic sea water, and the final
equilibrium attained in hypotonic solutions corresponds to that of the
control (unnarcotized) cells.6 Other experiments gave similar results.

It was, therefore, possible to study the effect of these narcotics on
cellular permeability to water by the method outlined above. In
Table II are given three experiments, representative of a larger number.
The table shows that none of the urethanes over a considerable range
of definitely narcotic concentration caused significant change in per-
meability from the controls, excepting that increase in permeability
occurred when the cells became injured during the experiment.7

These and similar experiments lead to the conclusion that urethanes
in narcotic concentration when dissolved in sea water do not decrease
permeability of the living cell to water.

EFFECT OF URETHANES WHEN DISSOLVED IN A NON-ELECTROLYTE

MEDIUM

The result obtained with narcotics when dissolved in sea water, i.e.,
failure to decrease cell permeability to water, may possibly be explained
on the grounds that at a given temperature permeability can be re-
duced only to a certain value, and that this value is the one normally
obtained in sea water. Since, then, the presence of the bivalent cations
of sea water might, perhaps, mask the action of narcotics, the experi-
ments were repeated in hypotonic dextrose solution.

The experiments were carried out as follows: Unfertilized eggs of
Arbacia were washed in 0.95 molal solution of dextrose, isotonic with
sea water, to eliminate electrolytes from the medium. The cells were
then caused to swell in 0.38 molal dextrose solution, isotonic with 40
per cent sea water. Other eggs from the same animal were washed in
isotonic dextrose solutions containing in the one case urethane, in the
other calcium chloride; they were then measured during the course of
swelling in hypotonic dextrose solution to which had been added ure-

6 It should be pointed out that prolonged exposure to urethane solutions may
change the shape of cells. This is especially the case in the more concentrated
solutions of ethyl and propyl urethane, in which the normally spherical cells may
become transformed to bizarre amoeboid forms. No such changes occurred when
cells were exposed to less concentrated solutions, for a shorter length of time and at a
lower temperature. In the experiments here reported, cells retained their normal
shape.

7 The narcotics, in the experiments here reported, were used in three different
concentrations, of which the lowest was still definitely narcotizing and the highest
not toxic under the conditions of the experiment. Still higher concentrations proved
injurious and increased permeability.

The narcotizing effect was kindly determined by Dr. E. B. Harvey on the basis
of cleavage experiments.

76

BALDUIN LUCKE

TABLE II

Urethanes in Sea JVater

In these three experiments the narcotics in the molar concentrations shown are
dissolved in 40 per cent sea water (40 parts of sea water and 60 parts of distilled water).
In the last column of the table is given the value of permeability, which is the number
of cubic micra of water entering the cell per minute, per square micron of surface,
per atmosphere of pressure. The temperature was 15 ± 0.5° C.

It is seen that the various urethanes do not decrease permeability beyond the
value obtained in sea water alone. The increased values of permeability (indicated
by an asterisk) were obtained in injured cells, i.e., in cells which failed to cleave when
returned to ordinary sea water and inseminated at the conclusion of the experiment.8

Compound

Concentration

Permeability

Ethyl urethane 0.2

0.1
0.05

First sea water control . .

Second sea water control

n-propyl urethane 0.1

n-propyl urethane 0.05

n-propyl urethane 0.025

i-propyl urethane 0.1

i-propyl urethane 0.05

i-propyl urethane 0.025

First sea water control . .
Second sea water control

n-butyl carbamate 0.05

n-butyl carbamate 0.025

n-butyl carbamate 0.0125

i-amyl carbamate 0.01

i-amyl carbamate 0.005

phenyl urethane 0.005

phenyl urethane 0.0025

phenyl urethane 0.00125

First sea water control . .
Second sea water control

0.055
0.057
0.056
0.054
0.054

0.096*

0.052

0.056

0.057

0.051

0.055

0.058

0.053

0.089*

0.061

0.063

0.059

0.061

0.060

0.057

0.056

0.059

0.069

8 It has previously been shown (Lucke and McCulcheon, 1930) that injury
induced by high temperature and by anisotonic solution causes an increase in cellular
permeability to water. From the experiments given in the table as well as from
similar experiments, it is evident that injury induced by toxic concentrations of
narcotics also increases permeability to water. Higher concentration of narcotics
than those employed produced injury, and hence increased permeability.

thane or calcium, in the same concentration as was present in the iso-
tonic solution of dextrose. Permeability was therefore determined, in
each experiment, in pure dextrose solution, in dextrose solution con-
taining a narcotic, and in dextrose containing calcium.

EFFECT OF URETHANES ON CELL PERMEABILITY

77

The results of three such experiments are shown in Table III. It is
seen that the values of permeability in pure dextrose solution are about
twice as great as in calcium-dextrose solution, while permeability values

TABLE III

Urethanes in Dextrose

The narcotics in the molar concentrations shown are dissolved in 0.38 molal
solution of dextrose. In the top row is given the permeability of cells in pure dex-
trose solution and in the bottom row the permeability in dextrose solution containing
0.01 m. CaCU. Each permeability value in this experiment is based on measurements
of ten cells. The temperature was 12° ± 0.5° C.

It is seen that the narcotics cause a definite decrease in permeability which,
however, is not as great as the decrease effected by calcium.

Solution

Permeability

Dextrose 0.38 m

0.096

0.097

0.092

Dextrose 0.38 + n-butyl carbamate 0.025 m. .

0.062

Dextrose 0.38 + i-amyl carbamate 0.01 m. . . .

0.085

Dextrose 0.38 + phenyl urethane 0.0025 m. .

0.070

Dextrose 0.38 + CaCl2 0.01 m

0.041

0.047

0.041

of the narcotized cells lie about midway. A number of similar experi-
ments gave the same results. In every case exposure to the narcotic
caused a definite decrease in permeability, which was never of the
magnitude of the decrease effected by calcium. The conclusion may
be drawn that narcotics tend to decrease cell permeability to water but
that their effect may be masked by the presence of cations in the
medium.

DISCUSSION

Most studies of the effect of narcotics on cell permeability have
been concerned only with permeability to various substances in solu-
tion.9 The most important investigations on permeability to water,
as influenced by narcotics, are those of Winterstein (1916), Lillie
(1918), Heilbrunn (1925) and Anselmino (1928).

Winterstein in his first group of experiments used sartorius muscle
of frogs, employing the usual method of weighing. He found that in
hypotonic solution of sodium chloride containing alcohol in narcotic
concentration the weight increase of muscle was less than in the same

^ 9 The literature on this subject is reviewed by Gellhorn, E., Das Permeabilitats-

problem, Berlin, 1929; Winterstein, H., Die Narkose, Berlin 1926; Lillie, R. S.,
Protoplasmic Action and Nervous Action, Chicago, 1924; Bayliss, W. M., Principles
of General Physiology, 4th edition, London, 1924; Osterhout, W. J. V., Injury,
Recovery, and Death in Relation to Conductivity and Permeability, Philadelphia
'and London, 1922; Jacobs, M. H., in Cowdry, E. V., General Cytology, Chicago,
1924; Hober, R., Physikalische Chemie der Zelle und der Gewebe, Leipsic, 6th edition,
1926; von Tschermak, A., Allgemeine Physiologic, I, Berlin, 1924.

BALDUIN LUCKE

hypotonic solution containing no narcotic. In later experiments
\Yinterstein constructed artificial "cells"; glass cylinders were covered
with the thin abdominal muscle of female frogs. It was found that, in
agreement with the experiments on sartorius muscle, four different
narcotics effected a marked decrease of the water intake by the "cell."
The effects were reversible.

Lillie investigated the effect of narcotics on permeability to water of
fertilized eggs of Arbacia. He had previously shown that in the
Arbacia egg fertilization is followed by an approximately four-fold
increase in permeability to water. This increase of permeability was
found to be inhibited by various organic anesthetics. In all cases eggs
which were caused to shrink, two or three minutes after insemination,
in solutions of these compounds in sea water of the appropriate con-
concentrations, remained in the condition of low permeability char-
acteristic of the unfertilized egg. This effect of narcotics was readily
reversible.

Heilbrunn repeated Winterstein's experiments, studying the in-
crease in weight of frog's gastrocnemius muscle in distilled water, and
in distilled water to which two per cent by volume of ether had been
added. He found that water entered etherized muscle somewhat
less rapidly than normal muscle. In further experiments Heilbrunn
investigated the rapidity of swelling of unfertilized eggs of Arbacia
in hypotonic sea water containing one or two per cent of ether. His
curves show that the cells swelled even more readily in the presence of
ether than in its absence. From these experiments Heilbrunn con-
cluded that ether does not lower the permeability of sea urchin eggs to
water.10

Very recently, Anselmino investigated the effect of various nar-
cotics on permeability to water of dried collodion and of copper ferro-
cyanide membranes, and found, in agreement with the work of Winter-
stein on living membrane, a marked decrease of permeability to water.

From the experiments summarized above (excepting Heilbrunn's
experiment on Arbacia eggs) the conclusion was drawn that narcotics
decrease permeability to water. It would appear, however, that the
chemical composition of the medium is a factor of importance when
investigating the action of narcotics on cell permeability. The experi-
ments reported in the present paper indicate that narcotics, at least
in the case of the unfertilized egg of Arbacia, do not lower permeability
to water when they are caused to act in sea water, or in a medium
containing calcium.

"' It is not improbable that the more rapid swelling of the narcotized cells in
Ik-ilbrunn's experiments is due to injury.

EFFECT OF URETIIANKS ON CELL PERMEABILITY 79

SUMMARY

1. The effect of narcotics (urethanes and carbamates) on cell
permeability to water was studied by measuring the rate of swelling
of unfertilized eggs of the sea urchin, Arbacia punctulatu, in hypotonic
sea water and in hypotonic dextrose solution.

2. Narcotics in the presence of sea water do not decrease permea-
bility to water beyond the value normally found in sea water.

3. But narcotics have a tendency to reduce permeability to water,
being, however, less effective in this respect than are bivalent cations.
This tendency to decrease permeability is demonstrated when nar-
cotics are used in solutions free from bivalent cations, i.e., in hypotonic
solutions of dextrose.

4. The effect of narcotics on permeability to water depends on the
chemical composition of the medium in which the narcotizing compound
is dissolved.

BIBLIOGRAPHY

McCuTCHEON, M., AND LUCRE, B., 1928. The Effect of Certain Electrolytes and
Non-electrolytes on Permeability of Living Cells to Water. Jour. Gen.
Physiol., 12:429.

LUCRE, B., AND McCuTCHEON, M., 1929. The Effect of Valence of Ions on Cellular
Permeability to Water. Jour. Gen. Physiol., 12: 571.

WINTERSTEIN, H., 1916. Beitrage zur Kenntniss der Narkose. IV. Narkose und
Permeabilitat. Biochem. Zeitschr., 75: 71.

LILLIE, R. S., 1918. The Increase of Permeability to Water in Fertilized Sea-
Urchin Eggs and the Influence of Cyanide and Anaesthetics upon this
Change. Am. Jour. Physiol., 45: 406. Comparative Permeability of
Fertilized and Unfertilized Eggs to Water. Science, 1918, N.S., 47: 147.

ANSELMINO, K. J., 1928. Versuche iiber Permeabilitat und Narkose. Pfliiger's
Arch., 220: 524.

McCuTCHEON, M., AND LUCRE, B., 1926. The Kinetics of Osmotic Swelling in
Living Cells. Jour. Gen. Physiol., 9: 697.

MCCUTCHEON, M., LUCRE, B., AND HARTLINE, H. K., 1931a. The Osmotic Proper-
ties of Living Cells (Eggs of Arbacia punctulata). Jour. Gen. Physiol., 14:
393.

LUCRE, B., HARTLINE, H. K., AND MCCUTCHEON, M., 19316. Further Studies on
the Kinetics of Osmosis in Living Cells. Jour. Gen. Physiol., 14: 405.

LUCRE, B., AND McCuTCHEON, M., 1930. The Effect of Injury on Cellular Perme-
ability to Water. Arch, of Pathol., 10: 662.

LUCRE, B., AND McCuTCHEON, M., 1926. Reversible and Irreversible Swelling of
Living and of Dead Cells. Arch, of Pathol., 2: 846.

HEILBRUNN, L. V., 1925. The Action of Ether on Protoplasm. Biol. Bull., 49: 461.

THE CHANGES IH'RIXG DESICCATION AND REHYDRA-

TION IN THE BODY AND ORGANS OF THE

LEOPARD FROG (RAN A FTP TENS)

YERXON D. E. SMITH AND C. M. JACKSON
(From the Department of Anatomy, University of Minnesota, Minneapolis]

On account of the fundamental importance of water in the living
organism, it is of interest to know how much water the body can lose
from its storage depots and yet survive. The further question arises
as to whether there is, after recovery from desiccation, a normal re-
distribution of the water in the body as shown by the water content of
the various organs. Dehydration is difficult to produce in the mam-
malian organism, but is easily accomplished in the frog. This animal
was therefore used in the present experiments.

MATKRIAL AND METHODS

Forty common leopard frogs (Rana pipiens) were used in the experi-
ments. They were all males of similar weight and obtained from the
same location in southern Minnesota. They were caught in October,
and the experiments ran from October through February. The frogs
were kept in large concrete tanks through which ran fresh, cold tap
water at a temperature ranging from 8° to 1 1° C. The water was kept
at a depth of six inches. Bricks were put into the tanks in such a
manner that when the frogs rested on them, only their heads would be
exposed. The frogs had reached a constant body weight before being
used.

The forty animals were divided into four groups of ten frogs each.
These were designated as control group A (controls for normal tresh
<>iv,un weights); control group B (controls for normal total water
content); test group C (frogs desiccated for organ weights), and test
:^ioii|> I) (rehydrated for organ weights). The frogs in control group
A were anesthetized with ether and autopsied at once. Their fresh
and dry organ weights were determined. The control group B were
killed by the ether and their fresh and dry weights were determined for
the body as a whole. The test group C were desiccated so as to lose
forty per cent or more of their total body weight, then anesthetized
and autopsied to obtain the fresh and dry organ weights. The test
group D were desiccated to the same extent as group C, and then

80

DESICCATION AND REHVDRATION IN RANA PIPIENS XI

allowed to recover for ten days in water at room temperature of 25° C.
(range 20-28° C.). They were then anesthetized and autopsied to
secure the fresh and dry organ weights. The body weights of test
groups C and D were taken at frequent intervals (see Fig. 1) to de-
termine the rate of dehydration, and, in the case of group D, also the
rate of rehydration.

The frogs were desiccated by exposing them to evaporation, each in
a cylindrical glass jar 8 inches in height and width. These jars were
covered by a metal screen with three meshes to the inch. They were
placed in a room, the temperature of which was kept fairly constant
by means of a thermostat. The average temperature was 25° C.
(range 23-26° C.). The average relative humidity was 25 (range 20-
29), measured with a Tycos hygrometer.

Before weighing in all groups, the excess of moisture was removed
from the frogs by carefully blotting with paper towels. No allowance
was made for the contents of the urinary bladder, as the frogs usually
voided urine during the manipulation incident to drying them.

In all cases, ether was used to anesthetize the frogs. With the
exception of control group B, the (anesthetized) animals were killed
by cutting open the tip of the cardiac ventricle and suspending the
animals by the head and feet so that all the blood possible might drain
out into a bottle. The weight and the water content of the escaped
blood were determined.

The procedure at autopsy was as follows: The head was severed
from the body at the atlanto-occipital articulation. The limbs were
separated at the hip and shoulder joints, respectively. All parts were
placed in a moist chamber until dissected. The dissection was made
upon dampened blotting paper. The lungs were removed at the
junction with the trachea, and were opened to remove any parasitic
flukes found there. The heart was removed by cutting the large veins
at their termination, and severing the aorta at its origin from the bulbus
aortae. The chambers of the heart were opened and their contents (if
any) removed. The contents of the stomach and intestines were
removed before weighing these organs. The ano-cloacal junction was
taken as the lower border of the intestines. The suprarenal glands
were not removed from the kidneys. The liver was removed and
weighed together with the gall bladder. The bile in all cases was
allowed to drain out before the weighing. The tongue was separated
by cutting the muscles at their attachment to the hyoid bone. When
the integument was taken, the subjacent lymph was merely allowed to
drain off. The term "remainder" includes the various items such as
large blood vessels, nerves, connective tissue and other structures not

V. D. E. SMITH AND C. M. JACKSON

included under the other headings. The remainder does not include
the water in the subcutaneous lymph spaces.

All organs were weighed on a YVilkens- Anderson analytical balance
in previously weighed weighing bottles with ground glass stoppers.
The organs, and likewise the entire frogs of control group B, were dried
in a Thelco electric oven at about 95° C. until constant weights were
obtained.

RESULTS

General Observations

The frogs of all groups were healthy and normal in all respects.
The average and range of initial body weights in grams were as follows:
control group A, 43.59 (35.0 - 53.2); control group B, 40.64 (34.5
41.6); test group C, 43.57 (37.0 -- 53.1); test group D, 44.80
(42.7 49.0). It will be noted that in average body weight groups
A and C were nearly identical, and group D differed from these by
about only two per cent. In the calculations, it is assumed that the
initial organ weights were equal in these three groups.

It required 28.6 hours, on the average, to desiccate the test group
C to an average weight of 24.98 grams, with loss of 42.7 per cent of
their initial body weight, or 51.9 per cent of their total water content
(Fig. I).1 These frogs at first were quite active, but after a few hours
of desiccation they remained motionless, apparently almost lifeless.
The skin became very dry, and the animals appeared emaciated. The
frogs maintained the sitting position, and the joints of the limbs be-
came stiff. In addition to the ten surviving frogs in this group, four
others died during the test and were excluded from the experiment.

A period of 31.8 hours was required to desiccate the test group D to
an average weight of 25.40 grams, with loss of 43.3 per cent of their
average initial body weight or 52.7 per cent of their total water con-
tent.1 Up to this stage, the general appearances and reactions were
the same as in the test group C. Upon being put back into water at
room temperature (25° C.) for rehydration, the test group D at first
i< -in. lined inactive. After two or three hours, they gradually began to
move about, and soon became normal to all outward appearances.
Alter 2(>.2 hours these frogs had returned to a nearly constant body
\\ei-ht, averaging 36.74 grams (Fig. 1). This weight they main-
tained, \\ith slight fluctuations, for ten days before being autopsied.
Thi- lie. nly constant body weight averaged about 8.1 grams below the
original body weight. Figure 1 shows the general weight curve fol-

' I In--'- percentages fur the cstimaccd losses of water content are slightly too
high, siiu i i In -y iiu Imlr <'as shown later) a loss of about 0.51 gram in solids during the
experiment. Tin- im-n-iit-.i ]o-s ()f total water content is 50.5 percent for group C

and 5U per ( cut tor ;<roiip I ).

DESICCATION AND REHYDRATION IN RANA PIPIKNS

83

lowed by this group during desiccation and rehydration. In addition
to the ten surviving frogs in this group, five others died during the test
and were excluded from the experiment. These five frogs died during
the dehydration part of the experiment. No deaths occurred during
rehydration.

Teat 6roup C (Desiccated)

Tesh iroup D (Desiccated and rthudrafed)

30 35
Time in hours

60 4.5 6.5 &5 10.5 12.5
Time in

FIG. 1. Curves of average weight during desiccation and rehydration.
Test group C was desiccated 28.6 hours, decreasing in average body weight from 43.6
grams to 25.0 grams. This corresponds to a loss of about 43 per cent in body weight.
Test group D was similarly desiccated 31.8 hours, decreasing from 44.8 grams to 25.4
grams, with loss of about 43 per cent in body weight. This group was then replaced
in water and increased to 36.7 grams in 26.2 hours (end of 58th hour of experiment).
Thereafter the weight remained nearly stationary during a period of 10 days. Failure
to reach the original weight is ascribed to the increase in room temperature.

Observations on Individual Organs and Parts

Observations at Autopsy. — At autopsy the control group A appeared
normal in every respect with the exception of an occasional lung fluke.
The test groups C and D also occasionally contained one to three
flukes. In the desiccated test group C the blood was so thick that it
would barely run into the container. There seemed to be practically
no water in the tissue spaces and lymph sacs. The intestinal contents
were very slight. Most of the organs appeared very dry.

The rehydrated test group D at autopsy appeared quite normal
with the exception that the epidermis was shedding in large pieces
quite generally over the body. This process resembled the normal

84

V. D. E. SMITH AND C. M. JACKSON

ecdysis, although in this case it was caused by the desiccation. The
amount of water in the tissue spaces and lymph sacs appeared some-
what below that in the normal control group A. The blood seemed
normal in color and consistency.

Fresh Weights of Organs. — (Tables I and II.) — The average fresh
organ weights of control group A were taken as norms, and the average
fresh organ weights of test groups C and D were compared with these.

TAHLK I
Average Fresh Organ Weights in Grams

Control group A
(normal)

Test group C
(desiccated)

Test group D
(rehydrated)

Eyeballs

0.3980

03958

0 3987

Luniks

0.2146

0 1992

0 2134

Skeleton

5.2250

^ 0073

S 1 >46

Brain

0.1013

0 0934

0 1093

Kidneys

(I 1381

01151

0 1442

Heart

0.111 '

00922

0 1057

Spinal cord

0.0636

0.0498

0 0591

Testes

0.0424

0.0314

0 0456

Stomach-intestines

0.6182

0.4326

0 5812

Muscles

16.5533

11.4997

16 0331

Skin

5.7809

3.3976

4 7196

Tongue

0.3993

0.1892

0 3800

Spleen

00381

00170

0 0796

Liver

1.6586

0 7019

1 0137

Blood

1 4274

0 2739

1 3468

Fat bodies

0.0950

0 0251

0 046?

Remainder

0.71<H

04448

0 7808

Total

33.5844

22.9660

31 1317

As shown by Tables I and II, in the test group C, the eyeballs and
lungs remained nearly unchanged in weight, and the skeleton and
1'i.iin lost less than ten per cent through desiccation. The kidneys,
heart, spinal cord, testes, stomach-intestines and muscles form a group
losing moderately (16-30 per cent), but relatively less than the body as
a whole. Finally the skin, tongue, spleen, liver and (escaped) blood
lo-t l.-i si per cent, which exceeds the relative loss of the body as a
\\ hole.

In the rehydrated test group D, all the fresh organs, with the excep-
tion of the skin, spleen and liver, had regained practically normal
weight (within x percent). The skin remained about eighteen percent
below the noun, the spleen 22 per rent below, and the liver nearly .W
per cent belo\\ . The significance of these changes will be considered
later.

DESICCATION AND REHYDRATION IN KANA PIPIENS

TABLE 1 1

Average Percentage Difference in Fresh Organ Weights of Test Groups C and D,

Compared with Control Croup A

Test group C Test group D

(desiccated) (rehydratecl)

1. Organs with slight loss during desiccation

Eyeballs -0.6 ±0.0

Lungs -0.6 -7.2

Skeleton -4.2 - 1 .9

Brain -7.8 +7.9

2. Organs with moderate loss during desiccation

Kidneys - 16.7 +4.4

Heart -17.1 -5.0

Spinal cord -21.7 -7.1

Testes -26.0 +7.6

Stomach-intestines —30.0 —6.0

Muscles -30.5 -3.1

3. Organs with great loss during desiccation

Skin -42.8 -18.4

Tongue -52.6 -4.8

Spleen -55.4 -22.3

Liver -57.7 -38.9

Blood -80.8 -5.6

Water Content. — (Tables III and IV.) — The average percentage
water content found in the whole body of control group B was 82.14
per cent, ranging from 80.21 to 84.28 per cent. This average is taken
as the norm for the body as a whole.

The average percentages of water found in the organs of control
group A, and of test groups C and D, are given in Table III. For the
normal control group A, the organs are arranged in order of increasing
relative water content, from the skeleton (58.13) to the escaped blood
(88.97 per cent). The muscles in this group contain almost exactly
the same percentage of water as the average for the whole body. The
skeleton, liver, spinal cord, skin and spleen contain relatively less, and
the testes, "remainder," brain, heart, kidneys, stomach-intestines,
tongue, lungs, eyeballs and blood contain relatively more water than
the body as a whole.

In the desiccated test group C, the organs listed (Table III) have
all lost part of their water and hence appear lower in relative water
content than the corresponding organs of control group A. Assuming
that all the loss of weight in this desiccated group C (Table I) is due to
loss of water, we may estimate the theoretical water content of the
organs in group C. This theoretical water content should agree

86

V. D. E. SMITH AND C. M. JACKSON

fairly with that actually found in the desiccated test group C (Table
III). The calculations of the theoretical water content were made in
this way for every organ, but the figures in detail are omitted for
economy of space. The muscles on this basis show the closest agree-
ment with a discrepancy of only 0.2 per cent.2 Agreement (within 3.5

TABLE III

Average Percentage of Water in Organs

Control iirniip A
(normal)

Test group C
(desiccated)

Test group D
(rehydrated)

Skeleton

58. 13

57 <M

60 5>

Liver

76.08

72.10

76 41

Spinal cord

80.75

76.73

81 67

Skin

81.05

70.84

80 15

Spleen

82 13

77 84

81 39

Muscles

82.15

74 52

8? 17

Testes :

82.57

75.09

81.40

Remainder

S3 12

77 1(>

81 85

Brain

83.90

si. 28

85 09

Heart

M 07

82 >l

87 17

Kidneys

84.42

78 09

8406

Stomach-intesl ines

S5.00

75.21

83 36

Ton true

86.13

76. S3

87.87

Limes

87.11

82 52

86 15

Eyeballs

NT S3

85.38

88 16

Blood

88.97

77.33

88.79

per cent) is also found in the case of all the other organs except the
liver, spleen, tongue, lungs and skin. The percentage of water in the
liver of the desiccated group averages 28.7 per cent greater than that
to be expected from its loss of weight. The corresponding percentages
in the spleen, tongue, lungs and skin are, respectively, 17.9, 6.8, 4.5
and 4.0 per cent greater than those theoretically calculated from the
i lumge in organ weight. The significance of these discrepancies will
be discussed Liter.

The absolute amount of water given up by each of the various
organs or \<,ii\- in test group C is sho\\ n in Table IV. The amounts

I i ' method oi r.il< illation is as follows: As shown in Table III, every 100 grams

"I In -!i inn-,rlf in ilic normal control group A contains 82.15 grams of water and (ac-

' .mis of solids. Table II shows that in die desiccated test group C

imsol n HIM le had lost 30.5 grams in wreight, which is assumed to be due to

loss i«| \\.itrr. Sulit racl ing 30.5 from 82.15 gives 51.65 grams of water remaining in

tin- '''' nt nnisrk- in the desiccated frog. This gives a theoretical water

content nl 71.SJ p.-r cent. The actual water content, as found in i he oven-dried

mus< It- nl tin desiccated tf-t group C (Table III), was 74.52 percent, which differs

from the theoretical value !•> only 0.2 per cent.

DESICCATION AND REHYDRATION IN RANA PIPIENS 87

are estimated from the differences in relative water content between
groups A and C (Table III), assuming that the initial organ weights
were equal to those in control group A. The estimate for the lymph
spaces is explained later. The chief storage depots of water in the
frog's body are apparently the lymph spaces, yielding 8.51 grams (or

TAHLli IV

Organs and Parts of Test Group C, arranged according to the Absolute Amount of
Water in Grams Lost During Desiccation. Kst imat rd on basis of change in percentage
of water content shown in Table III and normal fresh organ weights in Table I.

Lymph spaces 8.5100

Muscles 5.0289

Skin 2.2794

Blood 1.0582

Liver 0.7559

Remainder 0.2547

Stomach-intestines 0.2001

Tongue 0.1985

Skeleton.. . 0.1363

Fat bodies 0.0294

Lungs 0.0225

Kidneys. . 0.0210

Spleen. . . ... 0.0181

Heart. . 0.0177

Eyeballs 0.0134

Spinal cord. . . 0.0132

Testes . 0.0112

Brain.. . 0.0091

Total. . 18.5776

45.8 per cent of the total 18.5776 grams of water lost by the body during
desiccation); and the muscles, yielding 5.03 grams (or 27.1 per cent).
Somewhat smaller contributions are made by the skin (2.28 grams or
12.3 per cent), and the blood (1.06 grams or 5.7 per cent). All the
other organs together give up only 1.70 grams, or 9.2 per cent.

In the frogs of the rehydrated test group D, the organs in general
have regained nearly normal water content (Table III). All the organs
average within 2 per cent of the normal water content found in control
group A, excepting the skeleton, which is about 2.4 per cent above
normal, and the heart, which averages 3.1 per cent above.

To determine the amount of water in the lymph spaces of control
group A, the total average fresh organ weight (33.58 grams, Table I)
was subtracted from the average original body weight (43.59 grams).
This gives a difference of 10.01 grams, due chiefly to the contents of the
lymph spaces. During desiccation, the test group C frogs lost 18.59
grams in average body weight. The organs of this group (excluding
the lymph spaces) apparently gave up 10.07 grams of water, as shown
by the estimates in Table IV. The remainder of the loss (8.52 grams)
was assumed to come from the lymph spaces, although a small part of
this loss is in solids, as will be shown later. It is assumed that the
lymph spaces of test groups C and D originally contained as much
water as the lymph spaces of control group A (10.01 grams). There-

V. D. E. SMITH AND C. M. JACKSON
*

fore, at the time of autopsy the lymph spaces of test group C should
theoretically contain 1.49 grams (10.01 minus 8.52 grams). If we
subtract the average fresh organ weight of this group (22.97 grams,
Table I) from their average final body weight (24.98 grams), we find
that the actual amount of water in the lymph spaces was 2.01 grams.
The apparent loss of solids in the various organs can be calculated by
comparing the theoretical losses of water (Table IV) with the decreases
in fresh weight (Table I). All the organs excepting the stomach-
intestines, lungs, eyeballs and testes show variable (usually small)
losses in weight beyond those due to apparent decrease in water con-
tent. The net decrease amounts to about 0.58 gram, or apparently
slightly more than sufficient to account for the difference between the
theoretical and the actual amount of water in the lymph spaces
(0.52 gram).

The total theoretical amount of water in the body, as a whole, in the
rehydrated test group D at time of autopsy was 30.18 grams. This is
obtained from the average final body weight (36.74 grams) multiplied
by the normal percentage of water in the whole body (82.14 per cent).
If we subtract the total water actually found in the organs (24.46
grams) from the total (theoretical) water content of the body (30.18
grams), we get the theoretical amount of water in the lymph spaces
(5.72 grams) of the rehydrated test group D. We can check this by
subtracting the average fresh organ weight of test group D (31.13
grams, Table I) from the average final body weight for this group (36.74
grams). This gives 5.61 grams as the amount of water found in the
lymph spaces, which is very close to the theoretical expectation.

In general, therefore, it appears that the lymph spaces of the desic-
cated test group C had given up approximately 80 per cent of their
initial water content, while the lymph spaces of the rehydrated test
group I) had not fully refilled, but were still about 44 per cent below
their estimated original water content. The explanation for this
difference will be discussed later.

1 )ISCUSSION

Iii .general, it must be remembered that the number of frogs (10)
in e.ich of the four groups is relatively small. The results may there-
loir be influenced somewhat by individual variations, so that in some

es the conclusions are to be regarded as tentative rather than final.

GENERAL ( )HSEKVATIONS

The average loss of weight (about 43 per cent) sustained by desic-
cation in our tc-t group- (' and D may be compared with the results of

DESICCATION AND REHYDRATION IN RANA PIPIENS 89

other investigators. Chossat (1843) noted a loss of 35 per cent in the
body weight of frogs subjected to evaporation. Similar observations
were made by Durig (1901). Hall (1922) subjected various species of
animals to exsiccation, and found that frogs survived a desiccation
period of 33 hours, with loss of 41 per cent in body weight, correspond-
ing to 50 per cent of the entire water content. De Almeida (1926)
reported that in frogs the loss in body weight before death from de-
hydration varies with the rapidity of evaporation. With rapid desic-
cation, death may occur with a loss of only 10 per cent, while with
slower experiments the loss may be 30 to 40 per cent. Further data on
dehydration in the frog are given by lizuka (1926) and Hug (1927).
Data for comparison as to storage and loss of water in other species
and in other forms of inanition are cited by Jackson (1925, 1929).
These include the effects of exsiccation on various invertebrates and
vertebrates.

The curves of body weight during dehydration in our test groups C
and D (Fig. 1) show a more rapid loss at the beginning, with a pro-
gressively slower rate of decrease later. The rate of rehydration fol-
lowed by test group D appears slower at first, but is more nearly uni-
form from the third to the 26th hour (58th hour of the experiment),
when the weight ceased to increase. This group, therefore, did not
return to their original body weights, but maintained a nearly constant
body weight averaging 36.74 grams, or 8.06 grams below their original
weight. This difference is apparently due chiefly to the partially un-
restored loss of water from the lymph spaces. It is probably caused
by the change in room temperature, which was 8° to 11° up to the be-
ginning of the experiment, but 25° C. during the test. Ott (1924)
found that frogs on transfer to cooler environment may absorb water
(chiefly in the lymph spaces), which is lost within a few days when
they are removed to warmer quarters.

INDIVIDUAL ORGANS AND PARTS

Fresh Weights. — The losses in fresh weights of the organs in the
desiccated group C agree in general with the findings of Durig (1901).
He reported that the musculature lost most heavily, but he did not
measure the loss from the lymph spaces, which we find to be even
greater. According to Durig, the brain and heart lost relatively less
than any other organ; but our data indicate that the eyeballs, lungs
and skeleton (organs not studied by Durig) lost relatively less fresh
weight than the brain or heart. Durig also found that the kidneys, in
spite of a marked decrease in water content, showed a relative increase
in weight, which he ascribed to retention of insoluble urinary constitu-

90 V. D. E. SMITH AXD C. M. JACKSON

ents. The kidneys in our desiccated group C lost only 16.7 per cent
of their fresh weight, so that they gained in relative weight, as did all
the other organs losing relatively less than the body as a whole.

Also in other species under conditions of dehydration, certain
organs such as the brain, eyeballs, skeleton and heart maintain their
organ weights remarkably well (in some species even increasing their
weight under these conditions) ; while other organs such as the skin,
muscles, stomach-intestines, liver and spleen lose a large proportion
of their fresh weight. (See review of the literature by Jackson, 1925,
1929.) A similar tendency was found by Ott (1924) in the frog during
inanition.

Ott found that the excess water absorbed when frogs were trans-
ferred to a cooler environment did not appreciably affect the weights
of the various organs, excepting the lungs and fat bodies. In the
present experiments, all the organs excepting the skin, spleen, liver and
fat bodies (as shown in Table I) in the rehydrated group D had returned
to nearly normal weight, and all excepting the heart and skeleton
(Table III) had returned to within two per cent of the original normal
water content. Data for the water content of the fat bodies are lacking
however.

Loss of Solids. — A comparison of the figures for fresh organ weights
in Tables I and II shows certain discrepancies between the desiccated
test group and the rehydrated test group D. All the organs of test
group D returned to within 8 per cent of the normal weight, excepting
the skin, liver and spleen. As previously mentioned, these discrepan-
cies are attributed chiefly to loss of solids during the experiment. The
liver during the brief period of inanition probably lost much of its
glycogen content, which could not be restored by the mere addition of
water. This would account for the nearly normal relative water
content of the liver in spite of its marked subnormality of weight. The
-kin in test group 1 ) started shedding (ecdysis) three to four days after
beginning rehydration, which probably accounts for its marked sub-
normality in weight, although the relative water content had returned
to nearly normal. Similarly the spleen may have lost solids (through
e-rape or destruction of cells), which would account for the subnor-
mality in \\eight in spite of nearly normal water content. The other
organs apparently lost but little or no solids, as was noted above.

\\'<ilcr Cmili-nt. — The average percentage of water content in the
organ- of our control group A (Table III) is in general agreement with
Otl's ligures lor his normal control frogs. The differences are within 3
per rent, excepting the spleen, which is higher in his series, and the
liver, skeleton and stomach-intestines, which are somewhat lower.

DESICCATION AND REHYDRATION IN RANA PIPIENS 91

In his fasting frogs with loss of 40 per cent in body weight, the per-
centage of water content is usually increased, in contrast with our
desiccated group C showing marked loss in water content.

Table IV indicates that the greatest absolute sources of water
during desiccation in the frog are from the lymph spaces, muscles,
skin, blood, liver, "remainder," and stomach-intestines, decreasing in
the order named. In relative loss, however, the corresponding order is
the lymph spaces, blood, skin, stomach-intestines and muscles.

The escaped blood of the test group C contained 77.33 per cent of
water, or 11.64 per cent below the water content of the escaped blood
in control group A (Table III). Yet the escaped blood of the rehydrated
test group D had regained practically the same relative amount of
water as the control group A. The markedly thickened blood of the
dehydrated frog is therefore rapidly restored to normal. Durig found
that the number of red corpuscles per cu. mm. was more than doubled
in the frog by desiccation .

After recovery from desiccation, nearly all the organs of the rehy-
drated group D return to within two per cent of their normal water
content (Table III), although insufficient water has been absorbed to
restore the original body weight. Ott (1924) similarly noted that the
excess water absorbed by frogs at lower temperature did not affect
the weight or water content of the organs, excepting the lungs and fat
bodies. The only exceptions in our rehydrated series at higher tem-
perature are the skeleton and heart, which gain an excess water content
of 2.39 and 3.10 per cent, respectively. The lymph spaces of the test
group D, however, are estimated (as previously mentioned) to have
remained 44 per cent subnormal in water content. It is evident that
the changes in water content of the frog's body with fluctuations of
temperature affect the amounts in the interstitial spaces rather than
the water content of the tissues and organs.

It is not the purpose of the present paper to compare the findings of
other investigators in various species, but it may be noted that, in
general, the main storage depots of water in the frog, aside from the
subcutaneous lymph spaces, appear to correspond in general to those
found in mammals. The amphibian is more tolerant of dehydration,
however, and can survive relatively greater losses. Moreover, the
water in the amphibian organism is stored in a more labile form, so that
it can be more readily removed and restored. Data for comparison
are cited by Jackson (1929).

SUMMARY

The principal results of the present investigation may be sum-
marized briefly as follows:

92 V. D. E. SMITH AND C. M. JACKSON

1. Male frogs (Rana pipiens) weighing about 44 grams were able
to recover after desiccation with average loss of about 43 per cent of
their body weight or about 51 per cent of their total water content.

2. The eyeballs and lungs of the desiccated frogs remained nearly
unchanged in average fresh organ weight, and the skeleton and brain
lost less than 8 per cent. The kidneys, heart, spinal cord, testes,
stomach-intestines and muscles formed a group losing moderately
(16-30 per cent), but relatively less than the body as a whole. The
skin, tongue, spleen, liver and (escaped) blood lost 43-81 per cent, or
relatively more than the body as a whole.

3. The average percentage of normal water content in the entire
body was 82.14 per cent (range 80.21-84.28) at temperature of 8 to
11°C.

4. In the normal organs the average percentage of water content
ranged from 58.13 in the skeleton to 88.97 in the escaped blood.

5. The decreases in percentage of water content in the desiccated
frogs averaged approximately as follows: skeleton 0.2, heart 1.8,
eyeballs 2.5, brain 2.6, liver 4.0, spinal cord 4.0, spleen 4.3, lungs 4.6,
"remainder' 5.9, kidneys 6.3, testes 7.5, muscles 7.6, tongue 9.3,
stomach-intestines 9.8, skin 10.2, and escaped blood 11.6.

6. The average absolute amount (grams) of water given up by the
various organs or parts of the desiccated frogs was approximately as
follows: lymph spaces 8.5, muscles 5.0, skin 2.3, blood 1.1, liver 0.8,
"remainder1 0.3, stomach-intestines 0.2, tongue 0.2, skeleton 0.1.
The remaining organs each lost less than 30 mgm. of water.

7. In the rehydrated frogs, all the fresh organs with the exception
of the skin, spleen and liver regained nearly normal wreight (within 8
per cent). In all the rehydrated organs, except the skeleton and heart,
the water content returned to within 2 per cent of the norm. How-
ever, the amount of water in the interstitial lymph spaces remained 44
per cent subnormal, probably due to the increased temperature of the
room.

8. The following peculiarities were noted in the rehydrated organs:
The fresh weight of the skin remained 18 per cent subnormal, ap-
parently due to ecdysis. The fresh weight of the spleen was 22 per
• (in subnormal, probably through escape or destruction of cells. The
liver ueivht remained 39 per cent subnormal, probably through loss of
glycogen. The relative water content in each of these three organs in
the rehydrated group was found to be within 0.9 per cent of the normal.

9. In general, the main storage depots of the frog, aside from the
lymph sp.n < appear to correspond in general to those found in mam-
mals. The amphibian is more tolerant of dehydration, however, and

DESICCATION AND REHYDRATION IN RANA PIPIENS

can survive relatively greater losses. Moreover, the water in the
amphibian organism is stored in more labile form, so that it can be
more readily removed and restored.

LITERATURE CITED

CHOSSAT, C., 1843. Recherchcs experimentales sur I'inanition. Memoires de

r academic royale des sciences. Tonic S des savants etrangers. Paris.
DE ALMEIDA, M. O., 1926. Sur Ics effets de la deshydration des batraciens produite

par la ventilation. Jour, de physiol. et de path. gen.. 24: 243.
DURIG, A., 1901. Wassergehalt und Organ funkt ion. Erste Mitth. Arch.f. d. ges.

Physiol., 85: 401.

HALL, F. G.', 1922. The Vital Limit of Exsirralion. Biol. Bull., 42: 31.
HUG, ENRIQUE, 1927. L'evaporazione dell'acqua attravcrso la cute della Rana in

varie condizioni d'ambiente. Arch, di scienze biol., 10: 322.

(Zoo/. Bericht, 1929, 18: 225.)
IIZUKA, N., 1926. Recherches sur la deshydration de la grenouille et son retentisse-

ment sur les echanges respiratoires. Ann. phvsiol. et physicochim. biol., 2:

310. (Biol. Abstr., 1929, 2: 1545.)
JACKSON, C. M., 1925. The Effects of Inanition and Malnutrition upon Growth and

Structure. Philadelphia. P. Blakiston's Son and Co.
JACKSON, C. M., 1929. Recent Work on the Effects of Inanition and Malnutrition

on Growth and Structure. Arch, of Path., 7: 1072.
OTT, M. D., 1924. Changes in the Weights of the Various Organs and Parts of the

Leopard Frog (Raua pipiens) at Different Stages of Inanition. Am. Jour.

Anat., 33: 17.

Vol. LX, No. 2 April, 1931

THE

BIOLOGICAL BULLETIN

PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY

OSMOTIC PROPERTIES OF THE ERYTHROCYTE

II. THE INFLUENCE OF pH, TEMPERATURE, AND OXYGEN TENSION
ON HEMOLYSIS BY HYPOTONIC SOLUTIONS

M. H. JACOBS AND ARTHUR K. PARPART

(From the Department of Physiology, University of Pennsylvania and the Marine
Biological Laboratory, Woods Hole, Massachusetts}

I

In the first paper of the present series (Jacobs, 1930), some of the
more important advantages of the erythrocyte as material for the study
of cell permeability were mentioned. At the same time it was pointed
out that because of its peculiar sensitiveness to certain environmental
factors whose effects upon ordinary cells are much less noticeable, the
erythrocyte has acquired an undeserved reputation for variability in
its osmotic behavior. In the present paper, the nature and magnitude
of the effects of three such factors, namely, pH, temperature and oxygen
tension, will be considered. It will be shown that at least two of them
have for the erythrocyte an importance entirely out of proportion to
that observed in the case of other cells and that their neglect in osmotic
studies on this type of material is certain to lead to serious difficulties.

A survey of the literature shows that the importance of the first of
the three factors was recognized by Hamburger, the pioneer worker
in the field of osmotic hemolysis, whose numerous earlier studies are
conveniently summarized in his book, " Osmotischer Druck und lonen-
lehre " (T902). Hamburger, to be sure, did not distinguish clearly be-
tween the titratable acidity or alkalinity of a solution and its true reac-
tion, now commonly expressed as pH ; but he did, nevertheless, show
conclusively that changes of the blood in the acid direction, whether

7 95

96 M. H. JACOBS AND A. K. PARPART

produced by carbon dioxide or by other acids, result in a lowered os-
motic resistance of the erythrocytes and that changes in the alkaline
direction have the opposite effect. He correctly associated these changes
in resistance with the volume changes, observed by von Limbeck (1805),
which occur under the same conditions. The volume changes he in-
terpreted in turn as being primarily of an osmotic nature, resulting from
changes in the amounts of base bound by the cell proteins and from a
certain type of ionic interchange between the cell and its surroundings
(1902, pages 307 and 335). In its details, Hamburger's theory was
rather indefinite and in several respects it has proved to be entirely er-
roneous, but it does nevertheless foreshadow to some extent the most
modern views on the subject.

.Among workers who followed Hamburger, there was for a time a
tendency to abandon the osmotic explanation of the effects of pH
changes and to attribute these effects to more or less vague colloidal
phenomena. Haffner (1920), for example, in a study in which actual
pH values are mentioned (for most rtf the values in question see Jodl-
bauer and IlalTner. 1920), and which in this respect represents an ad-
vance over the work of Hamburger, concludes that, "Da es sehr un-
wahrscheinlich erscheint, class der osmotische Druck der im Zellinnern
in wahrer Losung befindlichen Substan/en (lurch Yerschiebung der
I I Konzentration erhebliche, die obigen r.efnnde erkliirende Anderungen
crfahrt, so folgt, dass fiir das Quellungsverhalten der Zelle als Ganzes
der Quellungszustand — also der Ladungszustand — gewisser Zellkolloide
von ausschlaggebender P.edeutung ist." Similar views have been ex-
pressed by other workers.

More recently the osmotic theory of the effect of pH on the volume
of erythrocytes (from which its application to problems of hemolysis
is an easy step) ha- a-ain been brought forward, this time in a far
more definite and satisfactory form than previously, in the very impor-
tant papers of E. J. \Yarburg (1922) and of Van Slyke. \Yu, and
McLean (1923). I;or a somewhat briefer discussion of the theory
in question Van Slyke (1924, 1926) may be consulted. Not only did
this theory provide a plausible theoretical explanation of the known
facts but the observed and predicted magnitudes of the effects in ques-
tion, of which that on cell volume is the one which concerns us here,
were found to be in good quantitative agreement. So generally satis-
factory has this theory proved to be that it is most surprising that so
lit lie account of it lias as yet been taken by persons interested primarily
in problems of hemolysis. One of the purposes of the present paper
U to point out ii- general applicability to such problems.

'Mil- second of the three factors under consideration, temperature,

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 97

was also studied to some extent by Hamburger (1902, page 172), but
with results which, probably because of the crudeness of his methods,
were unsatisfactory. He found almost no measurable differences in
the degrees of hemolysis produced by hypotonic solutions at 0°, 14°
and 34° C. Such slight differences as he obtained seemed to indicate, if
anything, slightly more hemolysis at higher than at lower temperatures,
a result in the opposite direction from that observed when more accurate
methods are employed.

A much more satisfactory piece of work is that of Jarisch (1921),
in which a very clear and regular decrease in degree of hemolysis with
rising temperature was obtained for a number of species of mammals.
Neither the theory offered by Jarisch to account for this effect, how-
ever, nor the evidence upon which he based it are very convincing, and
it seems much more plausible to look for an explanation along the same
lines as those already shown to be useful in the case of pH effects.
While the indirect osmotic influence of temperature upon the volume
of the erythrocyte has apparently received only incidental attention, it
is in all but its details included in the general theory of such volume
changes.

Concerning the effect of the third factor, oxygen, on hemolysis,
little information is available. Ordinary ' fragility ' studies are of
very questionable value for reasons which will be discussed later.
Strangely enough, Hamburger, whose own work had done so much
to establish the importance of carbon dioxide as a factor capable of
influencing hypotonic hemolysis, at times appeared to attribute the in-
creased resistance of the ervthrocytes from defibrinated blood solely
to an increase of oxygen (for example, 1902, page 172), seemingly
forgetting the simultaneous loss of carbon dioxide. It is obvious that
conclusions of value can be drawn from such experiments only if the
effects of the different variables are properly separated, and this appears
not to have been the case in the work so far published in this field.
On the other hand, the effects of oxygenation of the hemoglobin upon
cell volume have been adequately dealt with by Van Slyke, Wu and
McLean (1923) and by Henderson, Bock, Field, and Stoddard (1924),
and the application of the results of these workers to problems of he-
molysis is obvious.

Summarizing our present knowledge of the subject, it may be said
that from the experimental point of view the general effects of pH and
temperature upon the degree of osmotic hemolysis are known, at least
qualitatively, even though they are only too frequently neglected in
practice. The effect of oxygen tension is known with less certainty.
On the other hand, the effects of all of these factors upon cell volume

o

fe/S -*-*^ ^

f r? R A R Y

98

M. H. JACOBS AND A. K. PARPART

can be predicted theoretically from the known properties of the erythro-
cyte, of which the base-binding power of its hemoglobin is one of the
most important; and the calculated and observed volumes have been
shown in the case of pH to be in good agreement (Warburg, 1922;
Van Slyke, 1924). It remains to determine how far the simple and
definite theory of osmotic volume changes is capable of accounting for
the effects upon hemolysis of the three factors in question and how
far less precise theories of " colloidal behavior " must be considered.
Only to the extent that the first type of explanation can be shown ex-
perimentally to be inadequate will it be necessary to invoke the second.

II

Before the presentation of the experimental results, it may be pointed
out that the method of hemolysis, when properly employed, is perhaps
the most delicate method at present known for investigating the osmotic
volume changes of cells. This highly desirable characteristic of the
method is, rather paradoxically, due to a property of the biological
material with which it is employed which is commonly thought of only
as a disadvantage, namely, — its variability. As is well known, the
erythrocytes in a given sample of blood form a highly heterogeneous
population with respect to their resistance to osmotic swelling. In Fig.
1 is represented the distribution of cells of different resistances in a

100-

80-

60-

40-

20-

.110

— i —
.100

.o&o

.070

.090
FIG. 1

Fie. 1. Variation in osmotic resistance of erythrocytes of the ox. Ordinates
represent percentages of hemolysis; ahscissrc. times in minutes. Conditions of
the experiment as mentioned in the text.

OSMOTIC PROPERTIES OF THK ERYTIIROCYTE 99

sample of ox blood. The data were obtained under the standard con-
ditions of 20° C., pH 7.4, the oxygen tension of room air, an exposure
to the solutions for an hour with stirring and the use of such small
quantities of blood that the composition of the solution does not undergo
any appreciable change during the course of hemolysis. The points
represented by circles, which were obtained by the method previously
described (Jacobs, 1930), and those represented by the two lower
crosses, which were obtained by the other author by a new method to be
described elsewhere, fall very satisfactorily upon the same smooth curve.
The uppermost cross is based upon a direct microscopic examination of
the suspension.

The cause of the osmotic variability of the erythrocytes is not defi-
nitely known. It may be associated with differences in the properties
of the membranes of different cells, either in resisting stretching or in
becoming permeable to hemoglobin when stretched. Another possi-
bility was suggested by Hamburger, namely, that different cells contain
different amounts of substances such as the materials of the stroma
and hemoglobin which occupy a part of the cell volume without them-
selves undergoing osmotic volume changes. The higher the percentage
of such substances the less the swelling of the cell as a whole in a given
hypotonic solution would be and the less, therefore, its tendency to
undergo hemolysis.

Whatever factor or factors may be responsible for the osmotic
variability of the erythrocytes, the important fact is that these cells are
so numerous (approximately 250,000,000 in an ordinary drop of blood)
that, on the one hand, successive samples from the same lot of blood are
almost identical and, on the other, the variation is practically continuous,
i.e., there are represented an almost infinite number of degrees of
osmotic resistance. Under these conditions, the variability of the cells
instead of being a source of error becomes an advantage by rendering
it possible to compare with a high degree of accuracy the osmotic prop-
erties of solutions whose concentrations differ only very slightly.

A fair idea of the degree of accuracv obtainable with the method of

o •*

osmotic hemolysis without the employment of any unusual precautions
is illustrated in Fig. 2. In this figure smooth curves have been drawn
through the points transferred from the original kymograph record (see
Jacobs, 1930). The slightly buffered NaCl solutions employed differed
in concentration by 0.001 M. With one or two slight discrepancies
the curves form a very regular series. It would obviously be possible
in the most sensitive part of the range of the instrument without further
precautions than those here taken to distinguish the osmotic effects of
solutions differing from one another by 0.0002 M or less, e.g., the effects

100

M. H. JACOBS AXD A. K. PARPART

of 0.0880 M and 0.0882 M NaCl. Such a difference in concentration
would produce in an osmotically perfect system a volume change of
less than 0.25 per cent and in an erythrocyte in which the volume of
the non-liquid part of the cell has been estimated at from 35 or 40
per cent (Ege, 1921) to 65 or 70 per cent (Gough, 1924; Krevisky,
1930) a visible change for the cell as a whole of only between 0.1 and
0.2 per cent. Differences of this order of magnitude, which are by

OSY

OSS

FIG. 2. Effect of concentration on the rate of hemolysis of ox erythrocytes
at 20° C. and pH 7.4. Ordinates represent scale readings of the instrument and
abscissae times in minutes.

no means the smallest that could with care be measured, could probably
not be dealt with successfully by any other method. Even the large
spherical Arl'ucia rgg. the most favorable cell in certain respects for
osmotic studies which has yet been employed, does not permit the ac-
curate measurement of volume changes ten times as great.

It follows that the hemolysis method under favorable conditions
is one of great delicacy. As with other delicate methods, however, the
effects of disturbing factors are correspondingly serious. Some of
these factors will now be considered.

Ill

The most important factor theoretically and the most difficult one
to control practically is the pH of the medium. It is safe to say that
more errors have been caused in studies of osmotic hemolysis by neglect

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 101

of this factor th;ui of any other. \Yhere a relatively large quantity
of blood is used with an unbuffered salt solution, the resulting pH will
be determined chiefly by that of the blood, which in turn will vary
greatly according to the amount of CO, which has been allowed to
escape, etc. If very small quantities of blood are employed the case
is no better, since now the final pi I will be more strongly influenced
by that of the solutions used, which because of absorption of carbon
dioxide from the laboratory air, etc. may be expected, in the absence
of buffers, to vary considerably. Under such conditions, even a breath
from the experimenter at the wrong time may completely ruin an ex-
periment. It is evident, therefore, that to obtain accurate results it is
important on the one hand to use very small quantities of blood and on
the other to work with buffered solutions of known pll. These precau-
tions have been used throughout this work.

In carrying out experiments upon the effects of pH changes in which
concentration differences of 0.0002 M are significant, the usual phos-
phate buffers are not entirely satisfactory. Solutions of Na2HPO4 have
considerably higher osmotic pressures than those of Nat^PC^ of equal
concentration, and mixtures of such solutions in different proportions
therefore involve an important variable in addition to pH. Theoret-
ically, it is possible to find a concentration of HC1 which when mixed
with a solution of Na2HPO4 will cause no appreciable change in os-
motic pressure, the resulting volume changes just balancing those in
dissociation ; practically, however, it is difficult to be certain that pH
standards made up in this way are exactly equivalent, since the ordinary
freezing point method for osmotic pressure determinations is not suf-
ficiently delicate for the necessary tests. It has proved better in prac-
tice, therefore, to use a solution containing a fixed amount of NaHCO3
and to vary the pH by adding different amounts of CCX, a substance
which not only causes no appreciable volume changes in the solution
itself but which distributes itself so rapidly between the solution and
the cells that its direct osmotic effect upon the latter can be considered
to be zero.

The disadvantages of the bicarbonate-CCX buffer system are, first,
that because of the volatility of the CO2 the experiments must be carried
out in tightly stoppered bottles in which stirring is somewhat difficult
to accomplish and from which samples cannot be removed at will ; and,
second, that the highest CO2 tensions involve a simultaneous decrease
in oxygen tension which in itself has indirect osmotic effects. As will
be shown later, however, these effects are small in magnitude and, on
the whole, the advantages of the NaHCO3-CO2 buffer system greatly
outweigh its disadvantages for experiments such as the one about to
be described.

102 M. H. JACOBS AND A. K. PARPART

•

The procedure employed in this experiment was as follows. A
mixture of MANaCl and MANaHCO;, (4: 1) was diluted to 0.12M.
0.11 M. etc., these concentrations referring to the two salts taken to-
gether. It is permissible to combine the concentrations of the salts in
question because of their great osmotic similarity. Taking first any
chosen concentration, a small part of the solution was almost saturated
with C(X. This saturated solution was then mixed with the unsaturated
solution in different proportions. For most of the mixtures the amount
of CO.,-containing«solution was so small relatively that the oxygen ten-
sion of the resulting combination was not greatly changed ; in the solu-
tions of lowest pH, however, it was considerably reduced. This re-
duction might have been largely prevented by the use of pure oxygen in
conjunction with the carbon dioxide, but in view of the comparatively
slight osmotic effects of even fairly large variations in oxygen tension
(see Section V). this refinement was not considered necessary.

As each solution was prepared, two carefully measured drops of
defibrinated blood were introduced into 50 cc. of it in a closed vessel.
As soon as thorough mixing had been accomplished a 30 cc. glass-
stoppered bottle was quickly filled with the suspension of crythrocytes
through a glass tube without unnecessary exposure to the air. The
bottle, completely filled and stoppered, was placed in a water-bath at
20° C. and rocked gently for one hour. A small glass rod in the bottle,
kept in continuous movement by the rocking of the latter, prevented
the erythrocytes from settling to the bottom. Immediately after the
bottle had been filled a pH determination was made upon a part of
the remaining solution with a quinhydrone electrode designed to prevent
escape of CO2. After a series of different pH values had been studied
for one concentration a second concentration was similarly employed and
so on until the entire range had been covered.

The results of this experiment are represented in Fig. 3 in which
percentage of hemolysis is plotted against pH, and concentrations are
represented by contour lines. It will be noted that starting at any given
point the percentage of hemolysis may be increased either by diminishing
the concentration or the ]>If, and decreased by changes in the reverse
direction. Furthermore, it is possible, by changing the two variables
in opposite directions, to secure an exact balancing of their effects. For
exam] ilc. by following the level of 50 per cent hemolysis from the
contour line' for concentration 0.110 M to that for 0.100 M it is seen
that a concentration difference of 0.010 M is here equivalent to a pH
difference of 0.45. Considering the figure as a whole, there is seen
to be some variation in the pH-concentratinn relation, though the order
of magnitude remains the same, being in the vicinity of 0.5 pH units

OSMOTIC PROPERTIES OF THE ERYTHROCYTE

103

for a concentration difference of 0.010 M. A comparison will later
be made between this observed order of magnitude and that predicted
by the osmotic theory of pH effects.

One further point is of interest, namely, that if, for example, a
concentration change from 0.110 to 0.100 is equivalent to a pH change
of approximately 0.5 and if, as is the case, a concentration difference
in this region of 0.0002 M or a concentration ratio of 0.9998 has a

100

20

FIG. 3. Effect of pH on osmotic hemolysis. Ordinates represent percentages
of hemolysis ; abscissae, pH values ; contour lines, concentrations in mols per liter.

visible effect upon the observed degree of hemolysis, then a pH change
of 0.01 should have approximately the same effect. In other words,
uncontrolled pH differences of more than this magnitude may be ex-
pected to be a source of error in experimental work.

IV

Experiments on the effects of temperature on the observed degree
of hemolysis are considerably easier to carry out than those of pH be-
cause a single phosphate buffer system may be used throughout, the
entire series. It is true that the necessary differences in the degree
of dilution of the buffer salts in such experiments will produce slight
pH changes, but the effects of these changes are almost negligible in
comparison with the large concentration differences involved.

In the experiment about to be described, which is a typical one
chosen from several of the same general type, the erythrocytes were

104

M. H. JACOBS AND A. K. PARPART

those of the cat. The stock salt solution employed consisted of M/j
NaCl and M/x Na_.IlPO4 in the proportion of 14 to 1, with the pH
reduced to approximately 7.0 by means of concentrated hydrochloric
acid so as to give upon dilution a pH in the vicinity of 7.4. A mix-
ture of this sort was, of course, not exactly equivalent osmotically
to M/t Nad. but the difference was not great and. in any case, it was
relative rather than absolute concentrations which were of importance
in this experiment. From the stock solution the necessary dilutions
were made, treating the original solution as M/^ The pH values for
the various dilutions, as determined with the quinhydrone electrode
both before and after the addition of the blood, had in previous experi-
ments been found in all cases to vary only very slightly from a value
in the vicinity of 7.4. In this connection, it may be mentioned that in
making ] (regressive dilutions of the stock solution the pH changes fairly
rapidly at first from its initial value of approximately 7.0. but by the
time concentrations such as those here employed have been reached,
at which the pH is in the neighborhood of 7.4, the effect of further
dilution is slight.

The blood in the proportion of two carefully measured drops to 50
cc. was introduced into the solutions, previously prepared and brought
to the desired temperatures, which in the case of those at 0° C. and
20° C. were kept constant within 0.1°, and in the case of the others

80

60

40

20

o-

.116 .110

10°

y

20°

J

30"

Z 4
/ A

40'

.106 .100

FIG. 4

.095 .090 .085

Fi<;. 4. KflVct of temperature on osmotic hcmolysis. Ordinates represent
<n In -moK -sis ; abscissae, concentrations in mols per liter and contour
1< mpi-raturrs in ° C.

within less than 0.5° in water-baths. The solutions were gently and
continuously stirred throughout the experiment. Determinations of the
percentages of hemolvM's attained were made at the end of one hour.
This time is not sufficient for the establishment of complete equilibrium

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 105

at 0° and 10° C., but the further changes at these temperatures are slight
and the disadvantages of experiments of longer duration probably out-
weigh their advantages.

The results of the experiment are presented in Fig. 4. The per-
centage of hemolysis is here plotted against concentration, and contour
lines are used to represent the different temperatures. It will be ob-
served that for the region of 50 to 75 per cent hemolysis, where the
erythrocytes are presumably fairly typical and where errors in the meas-
urements are small, a change in temperature of 10° C. is exactly bal-
anced by one in concentration from 0.0035 M to 0.0075 M. In other
words, for the region in question, a concentration difference of 0.001 M
is roughly equivalent to a temperature difference of the order of magni-
tude of two degrees. It follows, therefore, that if a visible influence
upon the observed degree of hemolysis is exerted in this region by a
concentration change of 0.0002 M, the same result should be obtained
by a temperature change of less than 0.5°. In the light of this relation,
it is not surprising that a lack of agreement is frequently found in ex-
periments carried out at " room temperature."

Reference may here be made to the results obtained by Jarisch
(1921). From his figures on page 256 it appears that in the case of the
ox, for example, a concentration of 0.473 per cent NaCl at 40° C. was
equivalent to one of 0.519 per cent at 15° C. The results for several
species of mammals are presented graphically on page 257 of the same
paper. It should be noted, however, that the method employed by
Jarisch was unsatisfactory in two respects. In the first place, only fifteen
to twenty minutes were allowed for the attainment of equilibrium and,
in the second place, no account was taken of anything short of complete
hemolysis (as judged by the eye). The use of more refined methods
would almost certainly have given a continuous fall in the curves repre-
sented on page 257 to 0° C. instead of only to 10° or 15° C. Making
allowance, however, for these and possibly other differences in technique
and remembering that the concentrations in the experiments of Jarisch
are expressed in percentages of NaCl rather than in mols per liter, it
appears that his results were of the general order of magnitude of those
here reported.

V

The effects of oxygenation upon the degree of hemolysis may next
be briefly described. On the whole, they prove to be much smaller in
magnitude than those produced by pH and temperature changes — in
fact, their practical importance as a source of error must be very small
since most laboratory solutions are in equilibrium with room air, and

&

106

M. H. JACOBS AND A. K. PARPART

in such solutions hemoglobin will 1>e almost completely oxygenated.
Only if a deliberate effort were made to reduce the hemoglobin would
effects of any important magnitude be expected.

Several experiments were tried at various times to determine the
effects of oxygenation at constant pH upon the degree of hemolysis.
The experiment here reported was a typical one. In it, ox blood was
placed at 20° C. in a series of different dilutions of the buffered stock
solution mentioned in Section IV. The degree of hemolysis observed

.090 .088 .086 .084 .082

FIG. 5

FIG. 5. Effect of oxygenation on osmotic hemolysis. Ordinates represent
percentages of hemolysis and abscissae concentrations in mols per liter. The
degree of oxygenation is represented as follows : open circles, complete reduction ;
solid circles, equilibrium with room air; squares, complete oxygenation.

at the various dilutions is represented by the solid circles in Fig. 5.
After this preliminary standardization, two lots of 50 cc. each of the
O.O.S; M, 0.0875 M and ( ).()')() M dilutions were placed in separate vessels
through which oxygen and hydrogen, respectively, were bubbled for
ten minutes. At the end of this time two accurately measured drops
of blood which had previously been introduced into the vessels but not
mixed with the liquid were submerged in the oxygen-rich and oxygen-
poor solutions. The gases were allowed to bubble slowly through the
resulting suspensions for an hour, and then determinations of the degree

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 107

of hemolysis were made in the usual way. The pi I of the various
solutions did not vary beyond the limits of error of the quinhydrone
electrode determinations. The hydrogen used contained approximately
0.05 per cent of oxygen, so reduction of the hemoglobin was probably
almost but not quite complete.

The degrees of hemolysis obtained under the different conditions are
represented in Fig. 5 by the open circles for the reduced and by squares
for the oxygenated blood. From the figure it is apparent that the max-
imum difference obtainable under these conditions is equivalent to one
in concentration of approximately 0.0016 M. Otherwise expressed, by
reduction of the hemoglobin the degree of hemolysis at the concentra-
tions in question is increased by about five to eight per cent. Neither
in this experiment nor in other similar ones were significant differences
observed between the effects of room air and those of pure oxygen,
which is what would be expected from the known character of the
oxygen dissociation curve of hemoglobin.

VI

In the three sections immediately preceding this one there have
been described and measured in terms of equivalent concentration
changes the effects upon osmotic hemolysis of variations in pH, temper-
ature, and oxygen tension. It remains to determine whether the mag-
nitude of these effects is such as can be accounted for wholly or chiefly
by osmotic differences within the cell resulting from changes in the
base-binding powers of the hemoglobin. Van Slyke, Wu and McLean
(1923) have treated at length from this point of view the effects of pH
and, to some extent, those of oxygen tension upon the related problem
of cell volume, but their equations apply to the more complicated situa-
tion where the cells are suspended in a relatively small volume of a
protein-containing solution (serum) whose composition varies with that
of the cells. We are concerned here with the simpler case, to which
these equations are not directly applicable, of cells suspended in a
protein-free salt solution whose volume is so great that its composition
may for practical purposes be considered to be constant. It will be
necessary, therefore, following in part the method of Van Slyke, to
derive an equation applicable to this type of system by which the theo-
retical volume changes associated with changes in the base-binding
powers of the hemoglobin may be estimated.

The substances of which account must be taken in the derivation
of the equation are the cations of the system (chiefly K° and Na°) to-
gether designated as B (base), and the anions exclusive of Hb' (chiefly

108 M. H. JACOBS AXD A. K. PARPART

Cl' and HCO3') together designated as A', hemoglobin ions Hb' and
hemoglobin uncombined with base, together designated as Hb and base
bound by hemoglobin, designated as BHb. The question of the degree
of oxygenation of the hemoglobin need not be raised until later. Fol-
lowing Van Slyke, the symbols (B}c, (A)c, (B)8, (A)s, etc. will be
used to refer to the amounts of B and A in the cells and solution, respec-
tively, in a given quantity of the suspension and [B]c, [A]r, [/?]«, [A]*,
etc. to the corresponding concentrations expressed per unit amounts of
water rather than per unit volume of solution, i.e., [Z?]c = = (Z?)c/(H.,O)c.
Van Slyke uses as his units the amounts of the substances other than
water in milli-equivalents and of water in kilograms for one kilogram
of blood. The resulting concentrations are therefore expressed in milli-
equivalents per kilogram of water. In the present treatment, where
relative results only are desired, the particular units employed are of
no consequence as long as they are used consistently throughout.

\Yith regard to the measurement of quantities and concentrations
of hemoglobin. Van Slyke, in the absence of definite knowledge at the
time his work was done of the molecular weight of this substance, used
as an equivalent of hemoglobin the amount that combines with one mol
of oxygen. This assumption leads to calculated osmotic pressures for
hemoglobin which, in the light of our present knowledge, are probably
too high. However, according to the equation given by Adair (1925,
page 533). it would appear that the discrepancy is certainly much less
than it would be if the true molecular weight of hemoglobin could be
employed in the usual way for calculating the osmotic pressure of con-
centrated Milutiim^. In view of this fact, and in the absence of any
very definite knowledge concerning the exact contribution, which in
any case is relatively small, of hemoglobin to the total osmotic pressure
within the erythrocyte, \vc may without serious error continue to employ
Van Slyke's convenient assumption, particularly in a case like the present
one where comparative results only are desired.

In order that osmotic equality inside and outside of the cell may
exist, which is known to be the case with the erythrocyte, whose delicate
wall is incapable of supporting in cither direction an excess of osmotic
pressure, the following relation must hold between the concentrations
of the various osmotically significant substances :

[B]e+\A]e+[Hb]r: \/'},+ [A]s (1)

or otherwise expressed,

(H20)c (H80).

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 109

Since the cell is impermeable to hemoglobin and to base but perme-
able to water, (Hb)c and (B)c will remain constant, while [Hb]c and
[B]c will tend to vary. Indeed, the values of the latter quantities may
be used as a measure of the volume of the " liquid " portion of the cell
and indirectly of that of the cell as a whole, provided that information
is available concerning the bulk of the " non-liquid " materials. The
erythrocyte is known to be permeable to the osmotically important
anions, so neither (A}c nor [A]c will be fixed but will in general vary
in such a way that the Donnan ratio, [/i]c/[^].s, will have the value
determined by the other properties of the system. (The use of con-
centrations in place of activities introduces no great error; for the justi-
fication for assuming complete dissociation of the various salts involved
-Van Slyke, Wu and McLean (1923) — may be consulted.)

If now in equation (2), (J5)c-- (BHb)c be substituted for (A)c,
2(A)S be substituted for (B}s-\- (A)8 and the terms be suitably re-
arraned we obtain :

2(A}

CH20)C~"2(5)C- -(BHb)c + (Hb)c'

Similarly, by substituting in equation (2) (A}c-\- (BHb)c for
2(A)S for (B)s-\- (A)s, and rearranging, we have:

_ — v •< i ; f i / j \

7TTO)7r : (H20), (H20),

Dividing both sides of equation (4) by 2(A)S/(H2O)S, substituting
the value of (H2O)g/(H2O)c from equation (3) and remembering the
definitions of [A]c and [A]s and of the Donnan ratio, r, we have finally :

cr

-0 + (Hb)/

Up to this point Van Slyke, Wu and McLean have been followed in
principle, and equation (5) is the same as their equation (14) except
that (BP}S, the base bound by the protein in the solution, is here zero,
there being no protein present in the solution. The base bound by
protein in the cell is also here designated as (BHb)c instead of as
(£P),.. Beyond this point a somewhat different treatment of the
problem has been found convenient.

If under any given conditions of pH, temperature and oxygenation,
the amount of base bound by unit amount of hemoglobin be repre-
sented by F, we may write F(Hb) instead of (BHb} and equation
(5) becomes:

2(B)c+'(Hb)e (1- ^"'

110 M. H. JACOBS AND A. K. PARPART

Substituting in equation (1) r[.-J]« for [A],- and F[Hb]c for [BHb]c.
we obtain :

«-'\ (7)

Introducing into (7) the value of r from (6) :

[M>}c=2(B)e2^,:r',';' ... ' (8)

As mentioned above, the concentration of hemoglobin [Hb] is de-
termined by the amount of water in some given quantity of cells, so
that l/[Hb] --kW, where W is the water in one cell and k is a con-
stant whose value need not be determined since it will subsequently be
eliminated. Representing the ratio (B)c/(Hb)c by R and substituting
C, the concentration of the salt in the external solution, for [A]*, we
obtain from equation (8) :

1 __ _ ,

-. (9)

[Hb\ 2C

It C is kept constant, it can be seen that the theoretical effect on the
amount of water in a single cell caused by a change of pH, temperature,
or oxygen tension will be governed by the extremely simple relation:

(10)

(^ 2K+1--F,
//'.. 2R4-1 — F'

where W \ and }]'., are the amounts of water in the cell and F} and F.,
are the amounts of base bound by a unit amount of hemoglobin under
any two chosen conditions, R as defined above representing the ratio
of base to hemoglobin within a single cell or any given quantity of
cells. If, on the other hand, F is kept constant and C is varied, we
have the usual osmotic equation,

(11)

/r, c}'

(I- is actually not entirely independent of C. but for the small concen-
tration differences here involved it may be considered to be.)

VI]

It is now possible to determine whether or not the observed effects
on osmotic hemolysis of the factors pH, temperature, and oxygen ten-
sion are of the order of magnitude of those predicted by the simple
theory developed in the preceding section. A convenient method for

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 111

making the comparison is to convert equation (9) into the following
form :

W ^R _|_ 1 . _ F

— - ' - ( \7\

W2~~2R+ 1--F, ' C,'

C and F have here both been allowed to vary simultaneously, it being
assumed that these variables are independent of one another, which
under the conditions in question is approximately true.

Two observed points on a line of equal hemolysis are now chosen.
By the osmotic theory equal percentages of hemolysis indicate the
entrance into the cells of equal quantities of water, I.e., under these
conditions W1 = H\, and the resulting value of the right-hand side of
equation (12) should be unity. If, therefore, on substituting in equa-
tion (12) the appropriate numerical values of C and F for the twq
chosen points together with the proper value of R for the blood in
question, the expression becomes equal to, or nearly equal to one, the
observed results may be said to be in agreement with the theory. Any
considerable departure from this value, on the other hand, will indicate
inadequacy of the theory.

Concerning the values of C there is no difficulty. The necessary
values of R and F, however, cannot at present be obtained with the
same degree of certainty, since the work here reported was done on
the blood of two species, the ox and the cat, for which exact data con-
cerning these quantities are apparently not yet available. Indeed, emi
if such data had been published, it is not likely that different animals
of the same species would fail to show some individual differences.
In view of the fact, however, that various simplifying assumptions
have been made in the derivation of the equations, which at best are
only approximate, and the further fact that, as far as is known, the
bloods of different mammals resemble one another fairly closely, it
would seem to be permissible to take from the literature such data as
are at present available, expecting no more than that the calculated and
observed results may perhaps prove to be of the same order of mag-
nitude.

With regard to R, in the absence of any more definite information
concerning the bloods actually used, we may take as a plausible value
that given for the horse by equation (15) of Van Slyke, Wu and
McLean (1923), namely, 6.0. This value will be used in all of our
calculations.

Before the magnitude of the theoretical pH effects can be estimated,
it must be noted that the pH values used in the calculations of the base-
binding power of hemoglobin are those of the interior of the cell, while

8

112 M. H. JACOBS AND A. K. PARPART

those actually observed in the hemolysis experiments are those of the
solution. The relation between the two by the Donnan principle is :

pi I cell - - pH solution = = log r,

where r can be obtained by means of equation (6), preferably first
converted to the simpler form :

2R + 1 — F'

To avoid the mathematical difficulties involved in attempting to de-
termine pH cell from pH solution, it is convenient to work in the op-
posite direction, assuming a number of values of pll cell and calculating
the corresponding values of pi I solution. These latter values may then
be plotted and a smooth curve drawn through them from which the
de-ired values of pH cell may readily be obtained for any observed
value of pH solution. In the calculations which follow this has been
done.

Fairly extensive data for the calculation of pi I effects are available
for the blood of the horse at 38° C. Temporarily disregarding the dif-
ference in temperature, we may use for the calculation of F the equation
given on page 152 of the paper by Hastings, Van Slyke. Neill, Ileidel-
berger and I larington (1924) with the substitution of the necessary
constants from Table XXIII for horse hemoglobin with a cation con-
centration of 145 m.M.

Referring now to Fig. 3, we may choose any two points such as
A and /,' which represent the same degree of hemolysis, for example.
75 per cent at pll 8.0 and pll 7.0 respectively. These particular points
have been selected in order to keep within the range actually studied
by Warburg (1922) and by Van Slyke. \Vu and McLean C1023) ami
at the >ame time to take advantage of the region where our method
of measuring hemolysis (Jacobs. 1''30) is most accurate. The concen-
trations represented are 0.107 M and 0.07'' M respectively. .Remem-
bering that the values of cell pi I on-responding to solution values of 7.0
and 8.0 are 6.92 and 7.75, and assuming complete oxygenation of the
hemoglobin at the tension of room air. we obtain by calculation F 7.0 =
0.97 and F 8.0 =--3.34. Substituting these various values, together with
that of R in equation (12) we have finally

W_A 12 + 1- - 3.34 _ 0.107 Q9

\\'B~ : 12+ 1--0.97 ' 0.07"

The agreement between observation and theory is therefore seen
to be fairly close. It may be noted that while the closeness of this

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 1U

agreement will vary somewhat with the positions of the selected points,
A and B, owing perhaps to the influence of unknown factors of sec-
ondary importance, the points actually chosen are hy no means the most
favorable that could have been selected for the theory. Furthermore,
over a considerable range of pH and concentration, the observed and
predicted results are at least of the same order of magnitude. This,
under the circumstances, is all that reasonably can be expected.

The calculation just made involves a temperature differing consid-
erably from the one at which the observations were made. It is not
likely, however, if direct observations on the value of F at 20° C.
were available, that their use would greatly change the ratio in question.
According to Stadie and Martin (1924), the effect of temperature on
the base-binding power of hemoglobin is exerted not upon the buffer
value of the hemoglobin but only upon its isoelectric point. In other
words, in the equation for reduced hemoglobin

[BHb]=(3R[Hb] (PH — pi);

a change of temperature causes a change in pi but not in PR. Condi-
tions for oxygenated hemoglobin are similar, though there is a very
slight departure within the range in question from a simple linear rela-
tionship. Under these circumstances, in correcting our calculations so
as to make them apply to 20° C., both numerator and denominator of
the fraction :

2R+1--F,

2R -f 1--F2

will be changed in such a way that the value of the fraction itself is
little altered. In other words, it appears to be approximately though
not exactly true that a given pH change produces the same relative
volume change at different temperatures.

The question last considered anticipates to some extent the dis-
cussion of the effects upon hemolysis of temperature changes at con-
stant pH. For data upon the base-binding powers of hemoglobin at
different temperatures we have used Fig. 1 of the paper by Stadie
and Martin to which reference has already been made. Since the tem-
peratures are here 38° C. and 15° C. we have selected as our own
points for comparison A and B in Fig. 4 representing these two tem-
peratures and a degree of hemolysis of 60 per cent. This percentage
is as near to the region of the greatest accuracy of our instrument as
the rather small amounts of blood used in this experiment permitted
us to go. The concentration values are 0.081 M and 0.092 M, re-
spectively. Since the relation between the pH of the cell and that of
the surrounding medium is almost the same at a solution pH of 7.4,

114 M. H. JACOBS AND A. K. PARPART

which is that which we employed, and a blood pH of the same value,
we have taken our figures directly from the graphs of Stadie and Martin
without further calculations.

Considering first the figures for reduced blood, which are based
upon actual observations, we find F 38° =2.6 and F 15° = 1.0. Sub-
stituting in equation (12), we have

WA 12+1-2.6 0.092 _

WB~ ' 12 + 1- -1.0 0.081"

If, instead of the observed figures for reduced blood, we take the cal-
culated ones for oxygenated blood, we have

//•^ 12+1-3.2 0.092

It's =12+1--1.6 0.081"

Finally, we may use the theoretical effect upon pi calculated by
Stadie and Martin from the heat of ionization of hemoglobin. They
estimate that a change from 38° to 20° causes a change of pi for re-
duced blood of approximately 0.3 in the alkaline direction. Taking,
therefore, values of pi of 6.8 and 7.1 for the two temperatures in
question, and one of 2.8 for PR we have for F, i.e., for [BHb]/[Hb],
F 38° l.uS and F 20° =0.84. The concentrations for 60 per cent
hemolysis for these two temperatures taken from our own data are
0.081 M and 0.089 M, respectively. Substituting these values as be-
fore in equation (12) we have

WA 12+1- -1.68 0.080

WB ~~ 12+ 1--0.84 ' 0.081 :

This calculation, strictly speaking, applies only to reduced hemoglobin,
but the results would not be very different tinder conditions of partial
or complete oxygenation.

Finally, the factor of oxygenation at constant temperature and pH
may be considered. Our own results upon hemolysis indicate a com-
paratively small osmotic effect of this factor. For the calculated effect
at 3X the equation of Hastings, Van Slyke, Neill, Ileidelberger and
I farington (1924) already referred to may be used. As mentioned
above, the difference in temperature in the two cases, while unfortunate,
may be expected to have only a minor effect on the- calculated result.
At the pll of our experiment, namely 7.36, corresponding to an intra-
llular pll of 7.23 we find Fn= l.'H) and FR -- =1.22. For points A
and />' in Fig. 5. which are typical, the concentrations are 0.0862 M and
0.0875 M. Substituting these various values in equation (12) we
obtain :

WA 12+1- l/'O 0.0875

12--1- - 1.22 (MM ,2

= 0.96.

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 115

Otherwise expressed, the value of the water ratio WO/WR cal-
culated from the theoretical base-binding power of the hemoglobin is
0.94; that calculated from the observed equivalent concentration differ-
ence is 0.98. These values are of the same order of magnitude and
both suggest the likelihood of much smaller errors in experimental work
from neglect of the factor of oxygenation than of either of the others
considered.

From the calculations which have been given concerning the theo-
retical osmotic effects of changes of pH, temperature and oxygenation,
it is apparent that the predicted results are at least of the order of
magnitude of those actually observed. The agreement is better in some
parts of the range of the variables considered than in others and is
nowhere perfect, as might be expected considering the various simpli-
fying assumptions made in the derivation of our equations, and the
necessity for using data obtained for other species of animals under
dissimilar conditions. The existence of factors other than the purely
osmotic ones is by no means excluded by our experiments ; indeed it
appears to be likely. We believe, however, that such factors are prob-
ably of secondary importance and that until the possibilities of the os-
motic principles here discussed have been exhausted by calculations in-
volving the use of more accurate data than are now available it will be
unprofitable to indulge in speculations concerning little-understood " col-
loidal properties " of the cell and its constituents.

VIII

It has been shown that the influence upon hemolysis of the factors
pH, temperature and oxygen can be at least largely accounted for by
the osmotic effects which they produce through changes in the base-
binding power of the hemoglobin. The question arises why the erythro-
cyte appears to stand alone in its sensitiveness to such factors. Hemo-
globin is not a unique substance in its ability to bind base, nor in the
effect upon its base-binding power of at least the factors pH and tem-
perature. The proteins of all cells share with it these properties and
yet within physiological limits other cells appear to be only slightly
affected by the factors in question. Indeed, in the case of pH effects,
Lucke and McCutcheon (1926) over a very wide range of reactions
were unable to obtain unmistakable evidence of volume changes in the
case of uninjured cells.

It is perhaps not possible in the present state of our knowledge of
cells other than the erythrocyte to give a complete explanation of these
differences, but attention may be called to a number of significant facts,

116 M. H. JACOBS AND A. K. PARPART

In the first place, even if a cell such as the Arbacia egg were freely
permeable to anions and not to cations as is the erythrocyte — which is
probably not the case — equation (10) would suggest a reason for a
lesser liability of the former type of cell to volume changes. Other
things being equal, the larger the value of R in this equation the less
will be the effect upon the ratio W\/W « of a given change in F. Now
the value of R for the Arbacia egg is probably several times as large
as that for the erythrocyte. We have little information about the con-
centration of its cell proteins as compared with that of hemoglobin in
the erythrocyte. but it is almost certainly lower; the concentration of
base, on the other hand, must be, roughly, three times as great, since
the isotonic XaCl solution for the Arbacia egg is approximately M/2
as compared with M/6 for the erythrocyte. A value of R five times
as great for the Arlnicia egg as for the erythrocyte would therefore
not be beyond the bounds of probability. Furthermore, the change in
F with change of pH depends upon the buffer value of the protein
in question as defined by Van Slyke (1922). Xow the buffer value
of hemoglobin is approximately three times as high as that of some of
the- commonest proteins such as albumins and globulins (Hastings, Van
Slyke, Neill, Heidelberger and Ilarington, 1924). The absolute values
of F, of course, are also of importance, but they cannot as yet be
compared very accurately for the two types of cells. In any case, other
things being equal, the high buffer value of hemoglobin tends to favor
the production of large osmotic effects by pH changes.

In the second place, it seems almost certain that external changes
in pll have far less effect upon the internal pi! of a cell such as the
Arbacia egg than of the erythrocyte. In the latter the simple Donnan
principle: pi I internal - pi Internal - log v. governs the relation between
internal and external reaction, however the latter may be produced. At
ordinary reactions the difference is not usually greater than 0.2 pTI units.
On the other hand, in the eggs of Astcrias and Echinarachnius, which
are probably very similar to that of Arbacia. the internal pll is not
onlv normally very different from that of the surrounding sea water.
. // .. '>.X as compared with 8.1 (Chambers. 1'L'S). but is relatively in-
dependent of the pH of the latter. Only freely penetrating acids, such
as carbonic, butyric, etc., or freely penetrating alkalies, such as am-
monia, are able, without injury to the cell, to cause visible changes in
the appearance of intracellular indicators or to affect the cleavage process
in a decided manner. According to Chambers (1928), even acids and
alkalies of the naturally penetrating type in moderate quantities do not
noticeably change the reaction of uninjured protoplasm but affect only
certain cell inclusions. ( )n the whole, therefore, the opportunities for

OSMOTIC PROPERTIES OF THE ERYTIIROCVTE 117

producing intracellular pll changes which alone could cause the type
of changes descrihed by equation (10) are very slight as compared with
those existing in the case of the erythrocyte, and it is not surprising
that the latter cell responds to such changes in a more or less unique
manner.

Less is known about the effects of temperature than of pH on cells
other than the erythrocyte. Obviously, however, the situation is very
much more complicated in cells with a high rate of metabolism and a
great variety of physiological activities than in the relatively simple
erythrocyte. It would be expected, therefore, that in ordinary cells
the effects predictable from equation (10) would be modified and ob-
scured by others of a different nature. As far as it is legitimate to
apply this equation, however, the remarks already made about the mag-
nitudes of R and of F would hold in this case also. As for oxygen,
the reasons for the unique behavior of the erythrocyte are so obvious
that they require no special discussion.

IX

It appears from the foregoing discussion that the erythrocyte is a
more or less unique type of cell. Its nature is such that apparently
insignificant changes in environmental factors such as pH and tem-
perature produce within it osmotic effects of a magnitude sufficiently
great to be absolutely fatal to the securing of reproducible results by a
method as sensitive as that of hemolysis.

It is now possible to appreciate the difficulties that have attended
all work upon osmotic and perhaps, to a lesser extent, other types of
hemolysis where the factors in question have been neither measured
nor controlled. If it be true, as appears to be the case, that pH changes
of 0.01 pH unit and temperature changes of 0.5° C. can produce meas-
urable osmotic effects, what is to be said of the numerous papers which
have been published upon the " fragility " of the erythrocytes under
almost all conceivable normal and pathological conditions, in which in
the absence of any information whatever concerning these factors an
uncertainty of as much as 2.0 pH units and 10° C. or more may exist,
and in which, in addition, there is frequently no assurance that the
experiments have been sufficiently long continued to secure approximate
equilibrium? It is disheartening to be forced to believe that a large
part of the work in this much cultivated field is of very doubtful value,
but there seems to be no escape from such a conclusion. The erythro-
cyte being what it is, attempts to use it for comparative osmotic studies
without controlling especially the factors pi I and temperature, and to

118 M. H. JACOBS AND A. K. PARPART

a lesser extent oxygen tension, are scientifically in the same category
with attempts to study volumes of gases without accurate regulation
of temperature and pressure.

In conclusion, a point mentioned in the first paper of this series
may again be emphasized. The erythrocyte is not in itself a highly
unreliable and capricious form of material, as it is frequently believed
to be. Such capriciousness as it may appear to possess is merely a
reflection of the carelessness of the experimenter or of his lack of
understanding of its true physiological nature. When used uncritically
merely as so much " material " for experiments in the field of General
Physiology, the erythrocyte is likely to be a source of much vexation
to the experimenter. When treated in a manner appropriate to a highly
specialized cell with unique functions and physiological properties, con-
cerning which more exact information is already available than is the
case with perhaps any other single type of cell, it is capable of yielding
results of an extremely satisfactory character.

SUMMARY

1. By the method of osmotic hemolysis an attempt has been made
to evaluate the indirect osmotic effects upon mammalian erythrocytes
of changes in the pH, temperature, and oxygen tension of the surround-
ing medium.

2. The observed effects of these three factors, within the range con-
sidered, are of the order of magnitude of those predicted by the equation :

" 2R+1--F, C^

II- , 2R+ \--F,' C,'

where W^ and W» are the amounts of water contained in an erythrocyte
under two given conditions, T7, and F., are the amounts of base bound
by one equivalent of hemoglobin under the same conditions. Cl and C.,
are the concentrations of the solutions in question and /\ is the ratio of
base to hemoglobin within the cell. It is probable that the effects of
the factors studied are primarily osmotic in nature, though smaller
effects of a different sort are by no means excluded.

3. Certain differences between the osmotic behavior of the erythro-
cyte and that of other cells are discussed.

4. It is shown that pll changes of as little as 0.01 pH unit and
temperature changes of as little as 0.5° C. may have a measurable effect
upon the ohMTvrd degree of hemolysis. It follows, therefore, that
" fragility " tests and other osmotic studies upon erythrocytes in which
these factors arc not properly controlled are of little value.

OSMOTIC PROPERTIES OF THE ERYTHROCYTE 119

BIBLIOGRAPHY

ADAIR, G. S., 1925. Jour. BioL Chcm., 63: 529.

CHAMBERS, R., 1928. Biol. Bull., 55: 369.

EGE, RICH., 1921. Biochcm. Zcitschr., 115: 109.

GOUGH, A., 1924. Biochcm. Jour., 18: 202.

HAFFNER, F., 1920. Pfliigcr's Arch., 179: 144.

HAMBURGER, II. J., 1902. Osmotischer Druck und loncnlehre. Wiesbaden. Vol.

I, pp. 161-400.
HASTINGS, A. B., VAN SLYKE, D. 1)., NEII.U J. M., HEIDELBERGER, M., AND HAR-

INGTON, C. R., 1924. Jour. Biol. ( linn., 60: 89.
HENDERSON, L. J., BOCK, A. V., FIELD, H., JR., AND STODDAKU, J. L., 1924. Jour.

Biol. Chcm., 59: 379.

JACOBS, M. H., 1930. Biol. Bull., 58: 104.
JARISCH, A., 1921. Pfliigcr's Arch., 192: 255.
JODLBAUER A,., AND HAFFNER, F., 1920. PflHycr's Arch., 179: 121.
KREVISKY, C., 1930. Biochcm. Jour., 24: 815.

VON LIMBECK, R., 1895. Arch. f. r.r/vr. Pathol. n. Phannakol., 35: 309.
LUCKE, B., AND McCuxcHEON, M., 1926. Jour. Gen. Ph\siol., 9: 709.
NASSE, H, 1878. Pfliiger's Arch., 16: 604.
REZNIKOFF, P., AND POLLACK, H., 1928. BioL Bull., 55: 377.
STADIE, W. C, AND MARTIN, K. A., 1924. Jour. Biol. Chcm., 60: 191.
VAN SLYKE, D. D., 1922. Jour. Biol. Chcm., 52: 525.
VAN SLYKE, D. D., 1924. The Colloidal Behavior of the Body Fluids. Chapter

IV in Bogue : The Theory and Application of Colloidal Behavior, Vol. I.

New York.
VAN SLYKE, D. D., 1926. Factors Affecting the Distribution of Electrolytes,

Water, and Gases in. the Animal Body. Philadelphia.

VAN SLYKE, D. D., Wu, H., AND MCLEAN, F. C, 1923. Jour. Biol. Chcm., 56: 765.
WARBURG, E J.., 1922. Biochcm. Jour., 16: 153.

Till-: I MOMENT OF VELELLA SPIRAXS AND
FIONA MARINA

BENJAMIN KROPP1
STAZIONE ZOOLOCICA, NAPLES

The siphonophore, V clclla s pirn us Iiscli. occurs in enormous numbers
during the months of March and April in the Mediterranean and ad-
jacent seas. Especially after two or three days of the sirocco, the en-
ervating tropical wind from the south and southeast, these beautiful
animals often appear in thousands. There are records of their being
cast up on the 1 teaches of Sicily in such large numbers as to form a
broad blue line hundreds of yards in length. The animal consists of
a flat ovoid plate of chitin surmounted by a triangular " sail " set per-
pendicularly to the plate and obliquely to its longest axis. The plate
is usually 6 to 10 cm. long and 3 to 6 cm. wide. The sail rises about
two cm. above the surface of the plate. Like all other members of
the order, I'clella is free-swimming and is carried by wind and current.

/ 'clclla is striking because of its color, — a very deep blue. The
pigment is present in varying amounts in all cells of the ectoderm, enclo-
derin and mesoglea on both oral and aboral surfaces of the disc. It is
especially dense in the tissues immediately surrounding the pneumato-
phore on the aboral side, but in the blastostyles, dactylozooids and sail
the pigment is more diffuse and imparts a pale blue color to those organs.

Associated with V clclla are relatively large numbers of two molluscs,
—the nudibranch Fiona marina Forsk., and the prosobranch Janthina
conimnms Lam. Each of these species may be- found by itself, but the
presence of Vclclla invariably means the presence, in greater or less
degree, of these two molluscs also. Both Fiona and Janthina were
found feeding on Vclclla, and the eggs of each were found on partly
devoured animals or on their cbitinous skeletons. Fiona creeps readily
at the surface inverted, but in a 1 'clclla swarm it is most often found
clinging to the oral or aboral surface of the siphonophore and occa-
sionally to the sail. Here it feeds on the soft tissues, beginning at the
periphery usually, and abandoning the skeleton when all else has been
eaten. When removed from /V/<7/(/, Fiona is deep blue and its color

1 Parker Fellow in Zoology, Harvard University.

120

PIGMENT OF VELELLA SPIRANS AND FIONA MAKIXA 121

is indistinguishable from that of Vclclhi. Only the dorsal papillae and
the head contain the hlue pigment, the ventral surfaces being in general
whitish or pale yellow. The attempt was made to determine the re-
lationship between the body color of Fiona and its feeding habits.

In the laboratory the nudibranchs were removed from Vclella and
when starved, or fed on fish scraps, they rapidly lost the blue color.
It was not possible to retain them alive until all the pigment had disap-
peared, but the loss of color was very pronounced after three days.
When such animals were again permitted to feed on Vclella, the deep
blue color reappeared in the dorsal papilla; two to four hours after
feeding began. The pigment was distributed through all the cells of
those structures. On removing the nudibranch and starving it again,
the color tended to disappear once more. Individuals, on being re-
peatedly starved and fed on / 'clclla, showed the characteristic variation
between light blue and the metallic blue of the food. In animals fed
on colorless fish scraps for four days or more the color became very
light blue or yellowish. Fiona- that were allowed to feed only on Velclla
retained their dense blue pigmentation.

Fragments of blue Vclella tissue were found in the digestive tracts
of all Fiona examined immediately after feeding on Vclella. In all
cases the walls of the tract were impregnated with the same color. In
these nudibranchs ramifications of the stomach and liver extend into
the dorsal papillae, and indeed food fragments of the same nature as
those found in other parts of the tract were to be recognized in the
lumina of the dorsal papillae. There is no doubt, therefore, that the
blue color of these Fiona: is derived directly from Vclella. The pig-
ment passes through the digestive tract and into the dorsal papillae un-
changed and thus imparts the characteristic blue color to the latter
organs. Fiona has also been found on floating seaweed and debris.
At low tide it may also be discovered under stones, where it feeds on
small jelly-fishes and crustaceans. Unfortunately, it was not possible
to obtain animals from these habitats for comparison. No function
can be ascribed to the pigment in Fiona.

There is present in Janthina a violet-blue pigment quite unlike that
in Vclella and Fiona in general appearance, distribution and chemical
nature. But since it fed on Vclella in much the same way that Fiona
did, Janthina was also investigated in the manner described above.
Feeding, with materials other than blue Vclella tissue, had no effect on
the abundance or intensity of the Janthina pigment. Nor did similar
periods of starvation effect any noticeable change in the pigment. In
brief, the pigment of Janthina is not related specifically to its food as
seems to be the case with Fiona.

122 BENJAMIN KROPP

II

The determination of the chemical nature of the Vclclla pigment
has been attempted several times. Lankester (1873) regarded it as
probably identical with that of other Hydrozoa, but this view is no
longer acceptable. A. and G. de Xegri (1877) distinguished the pig-
ment of F. limbosa spectroscopically from that of Murcx and Aplysia
by the absence of characteristic absorption bands. More recently
Haurowitz and Waelsch (1926), in the course of a chemical analysis
of the entire animal, examined the fresh animals spectroscopically and
found diffuse absorption in the red and violet, but no sharp absorption
bands.

The fresh tissues of V . spirans, when macerated in sea water and
allowed to stand for four to six hours, yield a solution that closely
approximates in intensity of color the blue of the living animal. The
pigment may also be extracted in distilled water, though in this case
only a portion of the contained pigment will appear in solution. These
solutions, especially that in distilled water, show a reddish opalescence
in reflected light. Near the neutral point it turns yellow, and suc-
cessively pink and reddish-brown as it becomes increasingly acid. The
pigment is not light-sensitive, — solutions kept in full or diffuse daylight
for several days respond precisely as those kept in the dark. On ex-
posure to air, the solution becomes acid with corresponding changes
in color. On drying in air or in a dessicator, a reddish-brown, flaky,
or amorphous crystalline mass results. Tins is soluble with difficulty
in distilled water and does not yield the original blue color on increasing
the pi I to its value in the intact animal. \Yarming, or treating with
alcohol, produces a flocculent white precipitate which is insoluble in
water but soluble in alcohol and ether. The filtrate is red. Alkalis
turn the solution blue-violet. Acids and boiling turn the solution red
with the production of a flaky white or yellow coagulum. the latter
being easily digested by pepsin. F.xcess of NaOl 1 gives a violet color.

It was possible to make only some preliminary spectroscopic ex-
aminations of this pigment. The number of animals available wras
small, and it is necessary to use freshly prepared extracts, since under
laboratory conditions they decompose rapidly by bacterial action. At
22° C. solutions retained their color for four days. Distilled water
solutions showed diffuse absorption in the red, — A.= 655/1/1-685/1/1, and
blue-violet A^ 425/x/x-475/*./x. The bands were not sharply demarked.
This agrees with the findings of Haurowitz and "Waelsch noted above.
Then- was reason to suspect, of course, that the blue color of Fiona;
which had been fed on ]'clclhi was chemically identical with the color
of the latter. The blue dorsal papilla- of / "elella- fed Fiona- were, there-

PIGMENT OF VELELLA SPIRANS AND FIONA MARINA 123

fore, extracted in the manner described above. These extracts, in both
sea water and distilled water, reacted in precisely the same way and
showed all the properties of Vclclla pigment solutions.

Several descriptions of animal pigments exist, none of which, how-
ever, corresponds in all details with the properties of the Vclella and
Fiona pigment. Apparently there is here present a protein combination
upon which the blue color depends, but there is as yet no data as to

its nature.

LITERATURE

HAUROWITZ, F., AND H. WAELSCH, 1926. Zcitsc/ir. f. physiol. Chew., 161: 300.
KRUKENBERG, C. FR. W., 1882. Vergl-Physiol. Studien, 2te Reihe, 3te Abt.
LANKESTER, E. R., 1873. Quart. Jour. Micr. Sci., 13: 139.
MAZZARELLI, G., 1928. Boll. Istituto Zoul. R. Univ. Messina., No. 2.
DE NEGRI, A., AND G.. 1877. Bcr. d. d. Chcm. Geselhch., Jahrg. X, S. 1099 u.
1100. (Cited from Krukenberg.)

Till; s|-;\ RATIO IX THE DOMESTIC FOWL IX RELATION

TO AXTECEDEXT EGG PRODUCTION

AND IX BREEDING

MORI.KY A. JCLL
(From the Bureau of Animal Industry, U. S. Department of Agriculture)

A study of factors effecting a variation in the normal sex ratio of
the domestic fowl should have an important bearing on the practical
aspects of the poultry industry and on current theoretical questions
concerning sex determination. During recent years considerable
progress has been made in ascertaining the normal sex ratio of the
domestic fowl, but this does not necessarily imply an increased ability
to control sex. The term "sex determination" has reference to the
causes which lead to the production of an individual of one or the other
sex, whereas the term "sex control" is understood to imply that the
causes which determine sex are more or less amenable to human

control.

FACTORS THAT DO NOT AFFKCT TIII-. SKX RATIO

A number of observations have been made concerning factors which
do not affect the sex ratio. That prenatal mortality does not affect
the sex ratio has been determined by Thomsen (1911), Pearl (191 7a),
Crew and Huxley (1923), Jull (1924), Lambert and Knox (1926),
Horn (1927), and Lambert and Curtis (1929).

That large eggs tend to produce male chicks and small eggs female
chicks was the observation of Lienhart (1919), but the observations of
Jull (1924) and Jull and (Juinn (1924, 1925) show quite definitely that
e-g weight bears no relation to sex ratio. Jull and Quinn (1925) found
no significant difference- between the weights of male and female chicks
at hatching time, and Jull and Heywang (1930) found that the per-
• riitage chick weight of initial egg weight is independent of the sex
of the chick. Jull and (Juinn (1924) also found that there is no
relationship between the absolute length of the egg and the sex of the
chirk h. itched from it nor between the relative length or shape of
the rgg ,iiid the sex of the chick hatched from it.

Jull (1924) showed that there is no relationship between yolk
weight (it egg and >ex nor between yolk water content of egg and sex,
and Jull and lleywang (1930) found that the percentage yolk weight
ot chick weight at hatching time is independent of the sex of the

chick.

124

CONTROL OF SEX RATIO IN DOMESTIC FOWL

125

Lambert and Curtis (1929) and Christie and Wreidt (1930) ob-
served no significant differences among sex ratios of families of hens
of different ages.

ANTECEDENT EGG PRODUCTION AND THE SEX RATIO

The results secured by Pearl (19176), Crew and Huxley (1923) and
Jull (1922 and 1924) show that the greater the egg production prior
to the hatching season the lower the sex ratio, although the results of
Mussehl (1924), Lambert and Curtis (1929), Callenbach (1929), and
Christie and Wreidt (1930) do not confirm these observations since
they found no relationship between antecedent egg production and
sex ratio.

The possibility of the sex ratio decreasing as antecedent egg pro-
duction is increased is important from the practical standpoint because
the continuous selection of high producing hens to be used as breeders
each year should tend to give a high proportion of pullets. This is
desirable from the economic standpoint since the major receipts in
poultry husbandry are obtained from egg production. In order to
secure further evidence on the relationship between antecedent egg
production and sex ratio, the sex ratios were determined of families
of ten or more chicks per family produced during the normal hatching
season in 1925, 1926, 1927, and 1928 in flocks of Barred Plymouth
Rock, Rhode Island Red, and White Leghorn yearling hens at the U. S.
Animal Husbandry Experiment Farm, Beltsville, Maryland. The
breeding females producing the families were divided into three classes
of antecedent egg production, 0-29 eggs, 30-59 eggs, and 60-89 eggs
laid prior to the hatching season from the beginning of the second
year record in the case of each hen. The results are shown in Table I.

TABLE I

The Sex Ratio in Relation to Antecedent Egg Production in the Domestic Fowl

Breed

Antecedent Egg Production Classes

0-29 eggs

30-59 eggs

60-89 eggs

Barred Plymouth Rock ....

50.66±1.50

51.76±0.67
47.41±0.79

5U.OOzb2.47
5 1.96 ±1.04
49.18±0.75

52.35±3.37
Sl.73zbl.65

Rhode Island Red ....

White Leghorns . . .

The data in Table I show that in the case of all three breeds there
is no relationship between antecedent egg production and sex ratio.
These results, therefore, are in conformity with those secured by

126 MORLEY A. JULL

Mussehl (1924), Lambert and Curtis (1929), Callenbach (1929), and
Christie and Wreidt (1930), and support the evidence tending to show
that there is no relationship between sex ratio and antecedent egg
production.

It is true that the observations of Jull (1924) were based on the
annual egg production records of Barred Plymouth Rock pullets
mated to Brown Leghorn cockerels, whereas the data given in Table I
are based on yearling females mated to cockerels of the same breed.
At the same time, it has already been observed that Lambert and
Curtis (1929), and Christie and Wreidt (1930) observed no significant
differences among sex ratios of the families of hens of different ages.
Furthermore, Christie and Wreidt (1930) used a sex-linked back cross,
Fi, males from a mating of a Black Minorca male with Barred
Plymouth Rock females mated to Black Minorca pullets, and found no
correlation between antecedent egg production and sex ratio, ante-
cedent egg production being based upon the full year's production as
in the case of Jull's (1924) work.

On the other hand, Christie and Wreidt (1930) observed an excess
of barred progeny produced during February, March, and April as
compared with the progeny produced during May, June, and July
and an excess of black progeny produced during August, September,
and October as compared with the progeny produced during May,
June, and July. In the back-cross mating, heterozygous barred males
on recessive black females, used by Christie and Wreidt, the progeny
would normally be produced in the proportion of one barred male, one
black male, one barred female and one black female. The excess of
barred progeny produced in the early part of the season and an excess
of black progeny produced during the latter part of the season is
similar to the results secured by Jull (1924), but in Christie and
Wreidt's case the sex ratio was not influenced. They offer the follow-
ing suggestion, based on their results: "It is shown in this experiment
that there is probably a seasonal effect of B (the sex-linked barring
factor) and Jull's results may have been affected by this phenomenon."
This matter is being investigated.

INBREEDING AND THE SEX RATIO

Callenbach (1929) secured results indicating that different strains
of domoiir fowl produce differing sex ratios. Mussehl (1924) secured
results that :-h<>w sex ratios from different males differing materially,
and Jull (1'JJI secured sex ratios of 91.67 and 66.67, respectively,
from two hen- Li-rd on their total annual egg production. Christie
and Wreidt (193()j ob-ri \cd that some of their females gave aberrant

CONTROL OF SEX RATIO IN DOMESTIC FOWL 127

sex ratios, six each giving a low sex ratio and three each giving a
high sex ratio.

Hays (1929) found that inbreeding tends to lower the sex ratio in
Rhode Island Reds.

Moreover, Pearl (1924) observed that among matings of humans in
Buenos Aires "there is some evidence of a significantly greater pro-
portionate production of males in the offspring from matings involving
different racial stocks than in the offspring from matings in which
both parents belong to the same racial stock." Pearl analyzed the
statistics of over 200,000 human births, extending over a period of
ten years, among the following matings:

Argentine cf X Argentine 9

Italian d" X Italian 9

Spanish cf X Spanish 9

Italian o" X Argentine 9

Spanish cf X Argentine 9

Conclusive evidence regarding the possible influence of inbreeding
and outbreeding on the sex ratio among the various classes of domestic
livestock is still very meagre but, as Pearl says, "Any data tending to
throw light on the significance of any supposed sex-influencing factors
can but be welcome." In the case of the domestic fowl it is a com-
paratively easy matter to secure relatively large families from single
matings. Moreover, it is possible to practice close inbreeding over a
period of years.

At the U. S. Animal Husbandry Experiment Farm, Beltsville,
Maryland, matings involving crosses between breeds and varieties
have been made and the sexes of the progeny recorded. Also, a
number of pens of Barred Plymouth Rocks and White Leghorns have
been inbred for a period of years and careful records kept of the sex
of the progeny. The data from these matings are given here as a
contribution to the general study of sex-influencing factors.

During 1927 a number of varieties of the Plymouth Rock breed
were intercrossed as follows :

1. Buff Plymouth Rock tf X White Plymouth Rock 9

2. White Plymouth Rock tf X Buff Plymouth Rock 9

3. Silver Penciled Plymouth Rock cf X Partridge Plymouth Rock 9

4. Partridge Plymouth Rock cf X Silver Penciled Plymouth Rock 9

5. White Plymouth Rock cf X Barred Plymouth Rock 9

In 1928 the progeny of each intercrossed pen were inbred, the
amount of inbreeding being mostly half-sister-and-brother and in a
few cases full-sister-and-brother matings.

9

128

MORLEY A. JULL

In Table 1 1 are given the sex ratios of the families of the original
crosses and of the families of the inbred matings made the following
year.

The sex ratio of the families of the inbred matings is lower in three
and higher in two cases than the sex ratio of the families of the original
crosses, but in no case is there a significant difference.

TABLE II

Sex Ratios of Families of Plymouth Rock Variety Crosses and
Subsequent Inbred Matings

Mating

Sex I

latio

Number

Original Cross, 1927

Inbred Mating. 1928

1

44.62±5.60

58.70±5.48

2

58.23±5.12

46.36 ±4.19

3

48.72±2.74

50.65db3.61

4

48.70±3.95

44.86±1.21

5

50.44±5.10

44.83 ±3.09

Total

49.77±2.02

47.32 ±1.62

In the fall of 1924 an inbreeding project was begun with each of
four pens of Barred Plymouth Rock and White Leghorn pullets, each
pullet being mated to an unrelated cockerel. Pedigree breeding was
practiced so that the progeny for each male and female mating was
definitely established. When the progeny of each pen reached ma-
turity in the fall of the year, five full sisters and at least two of their
full brothers were selected from the female breeder that had the
largest number of progeny at the time of selection. Three full sisters
were chosen from each of the two female breeders having the next
highest number of progeny at the time of selection. The five full
sisters and the two groups each of three half-sisters were mated to a
full brother of the five full sisters. This was the method of selecting
the breeders for each inbred pen in 1926, 1927, and 1928, the chicks
raised in 1925 being the progeny of unrelated males mated to the
original pens of females.

In Table III are given the sex-ratio data of the progeny of the
original mating- .nnl of the progeny of the half-brother-and-sister and
full-brother-and-sister matings the following year.

CONTROL OK SKX RATIO IN DOMESTIC FOWL 129

TAIJLK III
Sex Ratio of Families from Half-brother-and-sister and Full-brother-and-sister Matings

Mating

Sex Ratio

Original Barred Plymouth Rock mating

Half-brother-and-sister Barred Plymouth Rock mating.
Full-brother-and-sister Barred Plymouth Rock mating.

Original White Leghorn mating

Half-brother-and-sister White Leghorn mating

Full-brother-and-sister White Leghorn mating

53.54 ±1.89
50.21 ±2.99
55.79±4.46
44.15 ±1.33
50.65±1.82
48.37±2.52

The data in Tables II and III show that close inbreeding in the
domestic fowl, at least for one generation, apparently has no significant
influence on the sex ratio of the progeny.

In order to present further evidence bearing on the problem of the
sex ratio in the domestic fowl in relation to inbreeding, the coefficient
of inbreeding has been determined for the inbred families produced in
1926, 1927, and 1928 in the case of the inbred Barred Plymouth Rocks
and White Leghorns.

Wright (1923) has pointed out that the coefficient of inbreeding (F)
depends primarily on the number and closeness of the ancestral con-
nections between the sire and dam, and secondarily on the degree of
inbreeding of the common ancestors of the sire and the dam. " Every
chain of generations in the pedigree by which one may trace back
from the sire to a common ancestor and then forward to the dam,
passing through no animal more than once (within the given chain),
contributes to the inbreeding an amount equal to one-half used as a
factor one more time than there are generations in the chain, with
the qualification that this must be multiplied by a corrective term
(1 + FA), in case the common ancestor (A) is himself inbred."
Wright's formula for determining coefficients of inbreeding has been
used in the analysis of the data discussed in this paper. "Let Fx and
FA be coefficients for the individual and for a representative common
ancestor of his sire and dam, and letting n and n' be the number of
generations between the sire and dam, respectively, and their common
ancestor, we have as the general formula:

F, = S[(*)«+"'+1(l + /W

It should be clearly understood that the coefficients of inbreeding
given in this paper refer to the coefficients of inbreeding of the progeny
of a given mating. For instance, when A is mated to his half-sister B,
the coefficient of inbreeding of their progeny is 12.50 per cent. Also
when C is mated to his full-sister D, their progeny have a coefficient of
inbreeding of 25.00 per cent.

130

MORLEY A. JULL

In Table I\* the sex ratios of families representing different
coefficients of inbreeding are given for both the inbred Barred
Plymouth Rocks and \Yhite Leghorns.

TABLE IV

Sex Rntin <>f Families Representing Different Coefficients of Inbreeding

Coefficients of In-
breeding

Sex Ratios

Barred I'lymouth Rocks

White Leghorns

12.50

25.00
37.50

50.00

50.68±3.99
43.90±3.45
38.57±5.93

48.82±2.92
46.63 ±2.00
44.20±4.83
41.94±7.41

The data in Table IV are consistent to the extent that in both
breeds the sex ratio decreases as the coefficient of inbreeding increases.
The sex ratios, however, do not differ significantly, as may be observed
from the fact that in the Barred Plymouth Rocks the highest sex ratio
is 50.68 ± 3.99 and the lowest is 38.57 ± 5.93, but the difference,
12.11 ± 7.14, is not significant. It is true that the observations are
based on relatively small numbers of families and it is possible that
observations based on larger numbers of families might show a sig-
nificant relationship between inbreeding and the sex ratio.

I Burger numbers of families are represented in the class of coefficients
of inbreeding ranging from 21.51 to 29.50 and in the class ranging from
29.51 to 37.50 than in the coefficient of inbreeding classes given in
Table IV. The sex ratios of the two larger family numbers of
coefficient of inbreeding classes are given in Table V.

TAKI.K V
Sex Ratios of Families Representing TICK Runges of Coefficients of Inbreeding

Coefficients of In-
breeding

:atios

Barred Plymouth Rocks

White Leghorns

J1.51 !9.50
29.51 37.5(1

45.66 ±2. 41
43.62 ±5. 10

48.48 ±1.59
48.68 ±2. 20

In i he Kirred I'lymouth Rocks the class of higher coefficients of
inbreeding h.i> the lower sex ratio, but in the White Leghorns the
reverse is th< In neither case is the difference between the sex

ratios signilii .mi . Moreover, the sex ratio of 48.68 ± 2.20 given in
Table V for White Leijlioin- for the coefficients of inbreeding ranging

CONTROL OF SEX RATIO IN DOMESTIC FOWL 131

from 29.51 to 37.50 is higher than the sex ratio of 46.63 ± 2.00 given
in Table IV for White Leghorns for the coefficient of inbreeding
of 25.00.

It seems apparent, therefore, that inbreeding in the domestic fowl
can hardly be regarded as a sex-influencing factor, even when close
inbreeding is practiced.

LITERATURE REFERENCES

CALLENBACH, E. W., 1929. The Relation of Antecedent Egg Production to the

Sex Ratio. Paul. Sci., 8: 320.
CHRISTIE, W., AND C. WREIDT, 1930. Seasonal Effects on Mendelian Segregations

and Sex Ratios. Separat ur Hereditas XIV, pp. 181-188.
CREW, F. A. E., AND J. S. HUXLEY, 1923. The Relation of Internal Secretion to

Reproduction and Growth in the Domestic Fowl. I. Effect of Thyroid

Feeding on Growth Rate, Feathering, and Egg Production. Vet. Jour.,

79: 343.
DARWIN, CHARLES, 1874. The Descent of Man. Rev. Ed., Vol. 1, p. 262. D.

Appleton and Co., New York.
FIELD, G. W., 1901. Experiments on Modifying the Normal Proportion of the

Sexes in the Domestic Fowl. Biol. Bull., 2: 360.
HAYS, F. A., 1929. Inbreeding in Relation to Egg Production. Mass. Agr. Exper.

Sta. Bull., No. 258: 256.
HORN, E., 1927. Untersuchungen iiber die Moglichkeit einer Geschlechtsvoraus-

bestimmung beim Hiihnerei nebst einer Factorenanalyse der Befiederungs-

geschwindigkeit von Kucken. Zeitschr. f. Tierzilcht. u. Zuchtungsbiol.,

10: 179.
JULL, M. A., 1922. Can Sex in Farm Animals be Controlled? Proc. Am. Soc. An.

Prod., p. 92.
JUIL, M. A., 1924. The Relation of Antecedent Egg Production to the Sex Ratio

of the Domestic Fowl. Jour. Agr. Res., 28: 199.
JUIL, M. A., AND J. P. QuiNN, 1924. The Shape and Weight of Eggs in Relation to

the Sex of Chicks in the Domestic Fowl. Jour. Agr. Res., 29: 195.
JULL, M. A., AND J. P. Quinn, 1925. The Relationship between the Weight of Eggs

and the Weight of Chicks according to Sex. Jour. Agr. Res., 31: 223.
JULL, M. A., AND B. W. HEYWANG, 1930. Yolk Assimilation during the Embryonic

Development of the Chick. Poul. Sci., 9: 393.
LAMBERT, W. V., AND V. CURTIS, 1929. Further Studies on the Sex Ratio in the

Chicken. Biol. Bull., 56: 226.
LAMBERT, W. V., AND C. W. KNOX, 1926. Genetic Studies in Poultry. I. The Sex

Ratio in the Domestic Fowl. Biol. Bull., 51: 225.
LIENHART, M., 1919. De la possibilite pour les eleveurs d'obtenir a volonte des

males ou des femelles dans les races gallines. Compt. rend. Acad. Sci.,

169: 102.

MUSSEHL, F. E., 1924. Sex Ratios in Poultry. Poul. Sci., 3: 72.
PEARL, R., 1917a. The Sex Ratio in the Domestic Fowl. Proc. Am. Phil. Soc.,

56:416.
PEARL, R., 191 76. Factors Influencing the Sex Ratio in the Domestic Fowl. Science,

N.S., 46: 220.

PEARL, R., 1924. Studies in Human Biology. Williams and Wilkins Co., Balti-
more, Md.
THOMSEN, E., 1911. Die Differenzierung des Geschlechts und das Verhaltnis der

Geschlechter beim Hiinchen. Arch. Entwick. Organismen, 31: 512.
WRIGHT, SEWELL, 1923. Mendelian Analysis of the Pure Breeds of Livestock.

I. The Measurement of Inbreeding and Relationship. Jour. Hered., 14: 339.

Till-. EFFEl T OF FATTY ACID BUFFER SYSTEMS ON THE
APPARENT VISCOSITY OF THE ARBACIA EGG,
\\mi ESPECIAL REFERENCE TO THE QUES-
TION OF CELL PERMEABILITY TO IONS

EVELYN HOWARD 1

(Prom the Department of Physiology, University of Pennsylvania, Philadelphia,
and the Marine Biological Laboratory, Woods Hole, Massachusetts)

The subject of this study is the degree to which certain intracellular
effects of acids may be modified by the presence of the salts of the
acids. If a living cell is exposed to a penetrating acid and its salt,
conditions being such that the salt is unable to penetrate, the internal
reaction of the cell will be influenced by the absolute concentration of
the acid rather than by the pi I of the solution as such. Results which
were interpreted in this way, as evidence of the relative impermeability
of cells to salts, have been obtained for a variety of cells by Jacobs
(1920<7, I920b, 1922a), Beerman (1()24), Bodine (1025), and Lillie
(1926, 1927). Furthermore. Osterhout (1925) and Osterhout and
Dorcas (1926) have investigated the penetration of CO., and H2S in
a more quantitative manner on a particularly favorable material, Valonia,
with results which led them to state that ". . . it is only the undissoci-
ated molecules which penetrate. . . . Under ordinary conditions there
seems to be little or no exchange of ions."

On the other hand Smith and Clowes (19240, />). and Smith (1925.
1926) found that certain physiological effects of the free acid in buffer
systems of penetrating acids depended on the pi I of the external solu-
tion, a result which they were inclined to interpret (1924/?) as being
due to the penetration of the salt of the acid rather than to effects on the
external cell surface, since analogous external pi T variations produced
by non-penetrating acids were physiologically ineffective. Similar con-
clusions as to the penetration of certain salts have been reported by M.
M. Brooks (1923) and by ITaywood (1927).

The contradiction implied in the conclusions of these two groups
of workers led to the present study of the question, in which the effect
of acids on the apparent viscosity of the protoplasm of sea urchin eggs
has been used as a measure of the intracellular effectiveness of the pH
produced by various systems. The results obtained indicate that physi-

1 Moon- Fdli.w in Medical Sciences, University of Pennsylvania.

132

CELL PERMEABILITY TO IONS 133

ologically significant quantities of the salts of several penetrating acids
enter the cells studied within a few minutes, in contrast to the ions
of certain mineral acids, for which there was found no evidence of
penetration under similar conditions. In this respect my work confirms
the conclusions of Smith, but it has extended them to apply to consider-
ably shorter times, and has led to a different conception of the manner
in which the intracellular buffering action of the salt might be brought
about.

MATERIAL AND METHODS

The unfertilized eggs of the sea urchin, Arbacla pitnctulata, have
been used throughout these experiments, since they are adaptable to
semi-quantitative studies on the apparent viscosity of protoplasm. Be-
cause of the danger that the selective permeability of cells might be
altered by injury, care wras taken that the material should be as normal
as possible. Eggs which were in good condition, as indicated by the
prompt formation of fertilization membranes and by uniform cleavage
(95 to 100 per cent), stood centrifuging without appreciable cytolysis
and as a rule the controls maintained a uniform value of the apparent
viscosity throughout an experiment. During the beginning and the
end of the season, a sample of eggs of each lot studied was fertilized,
and those lots selected for use which showed prompt and uniform eleva-
tion of fertilization membranes, but in the middle of the season such
a selection was unnecessary. The occurrence of an apparent liquefaction
on treatment with acid is not incompatible with subsequent cleavage and
development after fertilization on return to sea water, although pro-
gressively less recovery is obtained as the degree of liquefaction is
increased.

An experiment consisted of exposing the cells to the solutions in
question, and determining the apparent viscosity of the protoplasm after
various intervals. The method of centrifugation employed for these
measurements was composed essentially of observations on the degree
of movement of the yolk granules through the cell under the influence
of a given centrifugal force acting for a constant time. Certain factors
concerned in the resultant apparent viscosity of the protoplasm of these
cells will be discussed elsewhere. For the present purposes such a
determination of the rate of granule displacement has been used as a
measure of the effects of acid, without attempting any analysis of the
actual changes occurring in the apparent viscosity, or of whether changes
occurred in the granules themselves. The term liquefaction is used
in this paper to describe the effects studied, because it is probable from
the results of various authors (Jacobs, 1922ft; Edwards, 1923; and

134 EVELYN HOWARD

Brinley, 1928) that at least a part of the physical effects of acid on
the cell consist in the production of an actual liquefaction.

The granule displacement was evaluated in the following manner.
After centrifiu;ing. in order to prevent return redistribution of the
granules in the cell, the eggs were immediately fixed in a solution of
0.04 per cent formaldehyde in sea water. The layers of material which
are separated in sea urchin ci^s l>y centrifugal force may he classified
as a cap of fatty material, a hyalin segment of clear protoplasm, a
segment of ynlk granules, and finally the red pigment granules. Counts
were made of the percentage of eggs showing a hyalin segment of an
altitude of one-fifth or over of the diameter of the egg, out of a total
number of 200 eggs which showed an axis of stratification perpendicular
to the axis of observation. The position of the axis of stratification
can be determined in a definite manner by locating the red granule
layer opposite to the clear area. Since the estimation of the position
of this axis is only possible in those eggs which show some granule
displacement, the percentage actually counted is not an absolute per-
centage representative of the entire lot of eggs, although it tends to
become so as its magnitude increases. The altitude of the hyalin seg-
ment was measured from the surface of the egg to the boundary be-
tween the hyalin and yolk segments, including within this segment the
small amount of fatty material. The measurement consisted simply
of a comparison with an ocular micrometer scale at a magnification
of I'M) diameters. The counts were reproducible with a probable error
of plus or minus one per cent. The percentage of eggs showing this
degree of stratification as determined in the manner described will here-
after be referred to as sigma.

Counts made on the same samples of eggs at intervals after the
fixation showed an initial increase of about eight sigma during the first
twenty minutes following fixation, after which the value remained
constant for about two hours. A slow decrease in sigma of the order
of M to 8 occurred after four to eight hours. Slightly lower concentra-
tions of fixative did not give the initial rise but gave a more rapid drop,
! higher concentrations tended to render the cells opaque. With the
eMiicciitratioii of fixative employed, counts were made between one-half
hour if. t\vo hours after fixation.

I 'or values between 5 and 40, sigma increases nearly linearly with
the time of centrifuging; but below 5 the curve of sigma against time
is co i" the time axis, and above 40 to 60 sigma the curve tends

to bee- MI ic concave. Treatment with fatty acids causes sigma to in-
crease until a point i- reached, with increasing concentrations of acid,
where coagulation begins to occur, which is associated with an abnormal

CELL PERMEABILITY TO IONS 135

general appearance of the eggs. Coagulating concentrations of acid
were not used in the present study of permeability.

Numerical values of sigma in different acid concentrations are of
relative significance only, since considerable variations in absolute mag-
nitude occurred among different lots of eggs. In the experiments de-
scribed below, however, the relative effectiveness of a series of solutions
was reproducible without exceptions. The effects described are based
on a series of observations on three or more lots of eggs, for each series
of solutions tested.

A centrifugal force of approximately 1000 times gravity was used,
but this was not entirely constant from day to day, since the line current
at Woods Hole fluctuates considerably. However, controls in sea water
or unbuffered saline were included in each test, and with material in
good condition sigma values were usually constant during each experi-
ment, within the limits of error of the counts. The centrifuge was ac-
celerated in a standard manner, twenty-five seconds being used to move
the rheostat to its final point, and the machine established speed in about
forty seconds after the current was turned on. Room temperatures
varied between 22 and 25° C. during the course of the work, but the
variations during any one experiment were not more than a few tenths
of a degree. The cells were centrifuged for 2.0 to 2.5 minutes, the
actual time suitable for each lot being determined in a preliminary test.

The technique of measuring the apparent viscosity differs from that
previously employed by Heilbrunn (for numerous references see 1927)
and by Earth (1929) in one fairly important respect. These workers,
by centrifuging the material for various times, determined the shortest
time required to produce granule displacement for comparison under
different conditions. For the present purposes, I found that in cells
centrifuged for any one time, with the technique used, the amount of
granule displacement varied in different cells to such an extent that it
was not possible to compare groups of cells centrifuged for different
times unless comparisons were made on the basis of those times at which
equivalent percentages of cells showed a given amount of granule dis-
placement. In practice it was found more satisfactory to compare the
percentages of cells showing a given amount of granule displacement
after equal times of centrifuging.

Solutions. — Stock solutions of acids were made up by titration
against standard NaOH, using phenolphthalein as indicator, dilution
to the desired concentration, and retitration. The solutions of the salts
of the fatty acids were made up by mixing 1.000 N NaOH and 1.000
N acid to a pH of 9.4 with thymol blue, this being the approximate
pH of salt hydrolysis, allowing for salt error of the indicator, at salt

136 EVELYN HOWARD

concentrations of 0.5 molar. The salt solutions were kept on ice and
made up fresh each week. Solutions for use on the eggs were made
up daily from these stock solutions by volumetric dilution, with an
accuracy of 0.5 per cent, in balanced isotonic saline media. For the
first part of the work the saline medium was prepared daily from sea
water, by treating it with 5 cc. of 0.5 X HC1 per liter and passing a
current of moist room air through it overnight. This procedure frees
the solution from carbonates and CO., except for traces in equilibrium
with room air. The solution was then filtered and neutralized with
0.25 N NaOH to an apparent pH of 7.2 with phenol red. which, cor-
rected for salt error, becomes pH 7.0. To this medium were then added
the acids to be tested on the eggs. However, for most of the work an
artificial saline was prepared daily by mixing 1.00 M NaCl 420 cc., 0.50
M KC1 18 cc., 0.50 M MgSO4 51 cc., 0.50 M MgCl, 46.7 cc., and 0.50
M CaCL 18.7 cc., and distilled water to one liter. These concentra-
tions of salts were obtained from data on the relative salt concentrations
of sea water, as given in the Tabulae Biologicse. The salinity was ad-
justed to the point where the medium caused no significant change in
sigma. when compared with \Yonds Hole sea water. No difference in
the results was observed between solutions made up in this saline and
in the carbonate-free sea water.

When the volume of hypotonic reagents added to the saline exceeded
one per cent, hypertonic NaCl was added in proper amount to restore
the osmotic balance. Two per cent of distilled water in sea water was
found not to affect sigma. while 10 per cent distinctly increased sigma.
When the volume of the solutions added to the saline exceeded 10 per
cent, the amounts of K, Ca, and Mg were adjusted to maintain the
normal cation ratio, unless otherwise stated. A ten per cent variation
in the cation ratio was not associated with appreciable variations in
sigma. The effect of larger variations is described below.

All pH values given for HC1 and phosphate solutions were deter-
mined with a quinhydrone electrode. The individual determinations
were subject to potential drifts of one to three millivolts. The values
iMven in Table I are averages of several determinations. During pre-
liminary experiments, the pi 1 of the solutions was found to remain
unchanged during an experiment. The pll of solutions of fatty acid
buffers were calculated from the known amounts of the free acid and
it- -alt added, n^ing constants of pK' as 4.48 for acetic acid and 4.56
for valeric. The constant for acetic was taken from data of Michaelis
and Kriiger i 1"_M ) as the value of the constant in 0.5 M NaCl with
1/50 M Cad... and the value for valeric was calculated from the ac-
cepted value for its pK, assuming that the effect of the salinity is the

CELL PERMEABILITY TO IONS 137

same as with acetic acid. For the second dissociation constant of phos-
phoric acid Michaelis and Kriiger's value of pK' in 0.5 M NaCl of 6.43
was used.

It will be noticed that the procedure followed in making up the ex-
perimental solutions differed from that used by Smith in his work on
marine ova in one fairly important respect. Smith's method was to add
standard sodium acetate to neutral carbonate-free sea water, and then
to divide the resulting solutions into separate portions, and adjust the
hydrogen ion concentrations to various values by the addition of HC1
of the appropriate strength. This technique obviously lowers the salt
concentration at the same time that it raises that of the free fatty acid,
thus changing simultaneously the three variables : salt, acid, and pH.
Under these circumstances the evaluation of the effect of each variable
separately is difficult, and can be attempted only by a combination of a
number of experiments. Aly procedure was so planned that each of
the variables, acid, salt, and pH was in turn held constant while the
other two were varied, thus rendering it possible to test Smith's con-
clusions directly in individual experiments.

THE EFFECT OF CERTAIN ACIDS AND BUFFER SYSTEMS ON THE

APPARENT VISCOSITY

1. The Effect of Mineral Acids on the Apparent Viscosity

One of the arguments that ions in general do not penetrate living
cells has been that cells are not injured or affected by the completely
dissociated strong acids or bases in pH ranges in which the solutions
of certain weak acids and bases, which are believed to contain a large
proportion of undissociated molecules, can exert marked physiological
effects apparently resulting from their ability to affect the hydrogen
ion equilibria within the cell. The work of Bethe (1909), Warburg
(1910), Harvey (1911), Jacobs (1920&), Chambers (1928) and others
has shown that weak acids and bases produce intracellular indicator
changes promptly under conditions where strong acids and bases are
ineffective. A variety of physiological effects in animal, plant, and
bacterial cells have been shown to follow the same rule in that, according
to Loeb (1909), Cohen and Clark (1919)$§eobs (1924), Smith and
Clowes (1924 a and b), Earth (1929') and others, they respond to weak
organic acids more readily than to strong mineral acids, although in
application the rule is not without its limiting conditions, especially when
high concentrations of acids are employed (Crozier, 1916).

In the present experiments this general principle has been confirmed,
in that HC1 and the phosphate buffer system have been found to be

1 vV

EVELYN HOWARD

ineffective in producing a liquefaction of protoplasm within the pH
range in which the organic acids were effective.

Eggs were exposed to acids for various times, and they were then
centrifuged and >igma values determined in the manner described above.
The re-tilt- «\ exposing eggs to 11C1 in unbuffered saline for fifteen
minutes aiv given in Table I.

TABLE I

The Effect of HCl on the Apparent Viscosity

Experiment

Molarity

pH observed

Sigma

1

1.1X10-5

—

15

2.2

5.42

12

4.3

5.08

16

control

—

12

2

4.3

5.08

17

control

—

17

3

5.4

4.41

12

6.5

4.34

18.5

8.7

4.28

15.5

control

—

14.5

Since- the variations of sigma in cells in HCl solutions are slight,
and since there is no regular variation with the changes in acid con-
centration, it is considered that no significant effects on the apparent
viscosity are produced by HCl at pll 4.3 or above. At about pi 1 4.3
the cgi;> tend to cvtolyze, so that in view of complications due to cell
injury, quantitative measurements of sigma below pH 4.3 were not
undertaken. At pi I 4.1 the protoplasm was not coagulated after fifteen
minute-.' exposure, and at pi I 3.0 coagulation occurred.

Similar experiments were performed with phosphate buffers. So-
lutions of ().(H lUJ and ().()()()(> molar NaH2PO4 in saline media were
adjusted to a series of pll values between 4.2 and 5.S with NaOH.
ilicant differences were observed between the apparent viscosity
of cell.- in the saline media alone and in these phosphate solutions, after
exposures of five to forty minutes. The ions of the phosphate system
therefore appear to resemble tin- ions of MCI in being unable to influence
the interior of the living Arbacia egg. The pTI ranges used with these
mineral acids are within those in which the fatty acids exert their ef-
fects; consequently the effects produced by the fatty acids cannot be
due to the hydrogen ion concentrations as such which are produced in
the external media.

CELL PERMEABILITY TO IONS 139

2. The Effect of Fatty AeiJs on the Apparent I 'iseosity

The results obtained with the saturated fatty acids are very different
from those just described for mineral acids. The effect of unbuffered
valeric acid on protoplasmic viscosity is shown in Fig. 1. In this and
all subsequent figures the ordinates represent the degree of liquefaction
in sigma, and the abscissae the time in minutes of exposure of the cells
to the various solutions up to the time of centrifuging. In Fig. 1 it
will be observed that with increasing time of exposure the degree of
liquefaction increases up to a certain time, beyond which it tends to
decrease. The magnitude of the maximum liquefaction and the rate
of change increase with the concentration of the acid, between 1 X 10~4
and 3 X 10~6 molar.

Since the action of valeric acid may be due in part to the liberation
of carbon dioxide from intracellular carbonates, a few observations on
the effect of carbon dioxide were made. The gas was passed through
sea water until a saturated solution was obtained. The latter \vas then
diluted with sea water to certain pH values, determined with the quin-
hydrone electrode. A marked liquefaction was obtained at pH 5.1,
and a coagulation at pH 4.9. These pH values correspond to carbon
dioxide tensions of approximately 400 and 700 mm. respectively, ac-
cording to the data of Henderson and Cohn (1916).

The magnitude of the liquefaction obtainable with unbuffered
valeric acid is not very great. Cytolytic effects usually begin to occur
above concentrations of 0.0001 molar, obscuring the possible effects
of larger concentrations of acid on protoplasmic viscosity. Out of a
total of seven experiments with unbuffered valeric acid, a liquefaction
was observed six times ; and in the lot of eggs in which no definite
liquefaction could be observed, appreciable cytolysis was produced by
0.0001 N acid.

Using acetic acid in unbuffered solutions, I failed to observe a lique-
faction in four experiments, using concentrations of 1 X 10~4 to 1 ", [ 10~6
normal, although considerable cytolysis was observed in the higher con-
centrations. Consequently it is obvious that acetic acid produces intra-
cellular effects less readily than valeric.

The effectiveness of different fatty acids was compared in buffered
solutions, since under these conditions it was possible to obtain a definite
liquefaction without cytolysis in all the acids used. The effect of the
length of the carbon chain was studied with the acids acetic, propionic,
butyric (a. mixture of normal and iso forms), and //-valeric. Because
of the almost equal strengths of the acids in question, such a comparison
has a greater significance than would otherwise be the case. The re-
sults of a typical experiment with short exposures are shown in Fig. 2.

140 EVELYN HOWARD

The concentration of the free acid was in each case 0.000517 N, in the
presence of 0.000993 mols per liter of its salt. In this experiment the
salt concentration was determined from the amount of standard NaOH
added to tin- respective standardized acid solutions, instead of the salt
having lu.-c.-n made up separately as described above for the other ex-
periments. The following pH values were calculated for these solu-
tions, using the dissociation constants obtained by Drucker (1905) and
assuming the correction of - —.249 applied to pK for acetic acid to hold
for each acid: acetate system 4.77, propionate 4.87, butyrate 4.81, and
vale-rate 4.84. It is evident from Fig. 2, that the rapidity with which
In inc.- faction is produced by such buffer mixtures increases with the
increasing length of the carbon chain, irrespective of the slight differ-
ences in pIT, which, except in the case of the propionic acid, would

PLATE I
Explanation of Figures

The figures show the effect of various acid solutions on the apparent viscosity
of the Arbacia egg. The ordinates are the values of the viscosity in sigma,
plotted in each case to the scale shown in Fig. 1, although only relative viscosities
rather than absolute values are comparable in separate figures. The abscissa are
the times of exposure of the eggs to the solutions, up to the times of starting
the centrifuging. The diagrams give the time in minutes, and it will be noticed
that the time scales are not Ihe same in all the figures. The curves are numbered
in the order of increasing hydrogen ion concentration. The solutions as given
were made up in balanced isotonic saline. The effects shown in Figs. 3 to 6 are
typical of both acetic and valeric buffer systems.

Curve 1 in each figure represents the controls in sea water or in the unbuffered
saline medium.

1. The effect of unbuffered solutions of valeric acid. Curve 2, valeric acid
2.9- 10-° normal; curve 3, 1.74-10-5 normal, curve 4, 1.16-1Q-4 normal.

2. The effect of the length of tin- fatty acid carbon chain on the rapidity
with which the liquefaction is produced. Curve A, acetic acid; curve P, propionic
acid ; curve B, butyric acid ; curve V , valeric acid. The acid was in each case
5.17-10-4 normal, in the presence of 9.93- 10-* mols of the respective salt.

3. The effect of varying the jill by varying the concentration of the free
1 in the presence of a constant amount of the salt of the acid. Curve 2, pH

free valeric acid 0.000341 normal; curve 3, pi I 4.8, acid 0.000852 molar;
nn valeratc in each case 0.00135 molar.

The effect of varying tin- pH by varying the concentration of the salt
add in the presence of a constant amount of free acid. Free valeric acid,
i normal. Curve 2, pH 5.6, sodium valcrate 0.00585 molar; Curve 3, pH
; ". '• 0.00176 molar; curve 4, pH 4.6. valerate 0.000585 molar.

The eftc-ct nf varying the salt of the acid in the presence of a constant

a higher pH range than that of Fig. 4, namely 5.4 to 6.7. Free

act • 'Minimi, n,,nnal, in the presence of sodium acetate of 0.008 to 0.16 molar.

'. I-' I ' ve 3, 6.4; curve 4, 6.0; curve 5, 5.7; and curve 6, 5.4.

6- 'I 1" increasing both the acid and the salt, while maintaining a

constant pH of 5.17. 2, free valeric acid 0.00029, valerate 0.00117; curve

3, acid 0.000870. . oiiftfSl molar.

CELL PERMEABILITY TO IONS

141

PLATE I

[80

o

co

H

—

Q

h— f

D

|20

140

j I

J 180

TIME IN MINUTES

142 EVELYN HOWARD

be expected to reduce rather than to enhance the observed effects of
the permeability differences.

.After three to live minutes' exposure to the fatty acid solutions, there
is shown a tendency of the viscosity values to reach the same equilibrium
point. However, an identical equilibrium is not realized, and even after
twenty minutes' exposure the acetic acid has usually produced somewhat
less liquefaction than the valeric, in spite of the fact that it is a slightly
stronger acid, this behavior recalling the comparative ineffectiveness of
acetic acid in unbuffered solutions.

The observation that the rate of production of the protoplasmic
liquefaction increases with the length of the fatty acid carbon chain
accords with the general principle of Overtoil, which seems to hold for
a great variety of material (Jacobs, 1924), that in homologous series,
the more highly the non-polar portion of a molecule is developed as
compared with the polar, the more readily are intracellular effects pro-
duced. The ineffectiveness of the mineral acids in this case cannot
lie due to their low concentrations, for IIC1 fails to produce a liquefac-
tion at a concentration of 8.7 ' ; 10~s, while valeric acid is effective at

; W~fi N. Furthermore, if we consider the buffering capacity of tin-
solutions used (AB/ApH, Van Slyke, 1922) over the pH interval of
4.77 to 6.80, the latter taken as the normal pi I of the cell contents
(as determined for other Echinoderm ova by Chambers, 1928), it
is seen that the effective acetate solution has a buffering capacity of
only 0.00025. while for the ineffective phosphate buffer it is 0.0023.

3. Penetration of the Cell by the Salts of Patty Acids

So far the results have not been in disagreement with the conceptions
that ions penetrate uninjured cells only with considerable difficulty, if
at all. If now we consider the salts of the fatty acids, which are be-
lieved to be completely ioni/ed. we obtain quite a different picture.
For the following experiments in this section both acetate and valerate
buffers have been used.

(a) The Effect of tJie Salt of flic Fatty Acid on tJie Apparent Viscosity

As a preliminary to the study of the effect of the acid in the presence
of the salt, observations were made on the effect of the salt alone.
Sodium acetate in the saline media without the presence of free acid,
other than the traces resulting from hydrolysis, was found to produce
no significant changes in sigma in concentrations of 0.001 to 0.004
molar. In solutions of 0.008 and 0.01 molar, a liquefaction was ob-
served after twenty minutes' exposure which became more marked after

CELL PERMEABILITY TO IONS 143

longer exposures. The effect was variable in degree, but the lique-
faction was at times as great as thirty sigma above the controls after
forty minutes' exposure to 0.01 molar sodium acetate. In the presence
of a potassium excess of 0.01 molar, the acetate effect was increased,
i.e., it was manifested sooner and in greater degree. Lithium acetate,
0.01 molar, produced an even greater effect than potassium acetate.
The lithium and extra potassium, when added in the form of chlorides
to the standard saline, caused no significant changes in sigma; therefore,
this liquefying effect of lithium or of excess potassium appears to be
associated with the presence of the acetate ion.

(b) 77/r Effect on the Apparent Viscosity of Changing the pH by

Varying the Free Acid in the Presence of Constant Salt

In a concentration range in which the salt alone produced no de-
tectable effects, it may be shown that the effect of the acid on the
apparent viscosity in the presence of the salt is similar to the effect
of the acid in the absence of the salt in the case of valeric acid described
above. The result of varying the free acid in the presence of a constant
amount of salt is shown in Fig. 3. It is seen that, in the presence of
the salt, as with the acid alone, increasing degrees of liquefaction are
associated with increasing concentrations of free acid. This result
would be expected whether or not the salt penetrated. The concentra-
tions of acid used are greater than those possible in unbuffered solutions
without gross injury; and the liquefaction produced is also greater and
considerably more prolonged.

(c) The Effect on the Apparent Viscosity of Changing the pH by

Varying the Salt while the Free Acid is Constant

If the salt penetrates readily, it would be expected that altering the
pH by varying the salt concentration in the presence of a constant
amount of free acid would result in immediate differences in the degree
of liquefaction produced by the acid. That such is the case is shown
in Fig. 4, in which the effect of valeric acid 0.000585 N is shown to be
modified by sodium valerate of 0.00585 to 0.000585 molar. It is seen
in Fig. 4 that the degree of liquefaction increases with the increasing-
hydrogen ion concentration and the decreasing amounts of salt. At
pH 4.6 to 5.6 this effect is marked in both the valeric and the acetic
solutions. The effect of the salt appears within four minutes of ex-
posure to the solutions, the shortest time tested. It is presumably
not a specific salt effect, since it occurs in salt concentrations of 0.002
to 0.0006 molar, which are below those where the salt effect was ob-

10

144 EVELYN HOWARD

served in solutions of sodium acetate without added free acid, and also
because the lie UK- faction increases rather than decreases with the de-
creasing salt concentration. It is probably not caused by an influence
of the salt on the rate of penetration of the acid, since the acid penetrates
very quickly and apparently establishes a distribution equilibrium within
three to eight minutes in solutions of pH 5.2 or below.

It seems most probable that the influence of the salt is due to an
actual buffering effect exerted intracellularly. If this is so, these ex-
periments indicate that the effect of the acid on the apparent viscosity
is exerted through some influence on the intracellular hydrogen ion
equilibria, rather than through a more specific molecular reaction, since,
when the total acid concentration is constant, variations in the hydrogen
ion concentration alter the effectiveness of the acid. It is believed,
therefore, that these experiments constitute evidence that the salts of
fatty acids are able to penetrate the living cell. In regard to the com-
parative rates of penetration of the acids and their salts, it is shown
in Fig. 2 that acetic acid does not exert its characteristic effect until
after about three minutes' exposure. Since the buffering action of
sodium acetate may be in evidence after four minutes' exposure (the
shortest time tested), it appears quite possible that the salt penetrates
the cell, in the presence of the acid, at a rate which is of the same
order of magnitude as that of the acid.

When a similar experiment is performed in which the free acid
is kept constant and the salt varied, within a higher pi I range, i.e.,
between 5.7 and n.7. somewhat different results are obtained. In order
to produce a measurable liquefaction at these pH values, both the acid
and the salt must be increased to within the range of salt concentrations
at which the salt produces specific effects on the apparent viscosity.
The results of such an experiment are -shown in Fig. 5. The free acid
in this experiment was 9.6 ) ; 10~4 N, and the salt concentrations were up
to 0.16 molar. The pi I values calculated were, for curve 2, 6.7; curve
-\ f>.4; curve 4, 6.0; for curve- 5. 5.7; and curve 6, 5.4. It is si-en that
the initial effect is a liquefaction which increases with the hydrogen
ion concentration, as was observed for the lower pi I ranges described
above. However, with increasing exposures, sigma values decrease in
the solutions of lower pi I and rise in the higher, with the result that,
after an hour or more of exposure, between pll 5.7 and 6.7 the order
of increasing fluidity lias become reversed and consequently the lique-
faction iii' with the salt concentration instead of with the external
hydrogen ion concentration.

Since the order of increasing fluidity may be reversed in this way,
it is possible that the liquefying action of the salt may be of a different

CELL PERMEABILITY TO IONS 145

character from that of the acid. In view of the many factors which
may affect the apparent viscosity, it is not unlikely that the seat of action
of the hydrogen ion may differ from that of the salt. In this connection,
the possibility was considered that the influence of the salt on the ap-
parent viscosity was the result of an osmotic effect. The solutions of
sodium acetate in saline media were made up to be isosmotic with normal
saline, but if the salt should penetrate and remain osmotically active,
these solutions would not be isotonic for the cell ; in fact, the solutions
used in the experiment shown in Fig. 5 would in effect be of various
degrees of hypotonicity down to 68 per cent of the isotonic strength,
and the expected changes in volume would result in an increased ap-
parent liquefaction such as was observed. That this is probably not
the case, however, is indicated by the fact that measurements of cell
volumes in such solutions, kindly made for me by Miss D. R. Stewart
by a technique which she will describe elsewhere,2 revealed no significant
changes. In this respect, then, the sodium salts of the fatty acids differ
from the ammonium salts, which Stewart (1929) has found to cause
osmotic swelling. These results may be due to the fact that the sodium
salts can enter only in osmotically negligible amounts, but, on the other
hand, it may be that their entrance is for some other reason not asso-
ciated with an increase in the total osmotic pressure of the cell.

(d) The Effect on the Apparent Viscosity of Increasing the Total Acid

at Constant pH

When the pH is kept constant and the free acid and the salt in-
creased simultaneously a greater liquefaction is produced as the free
acid concentration becomes larger, as shown in Fig. 6. Indeed, a coag-
ulation was obtained with free acid above 0.001 N at pH 5.2. This
result might be due to the fact that relatively less salt than acid pene-
trates the cell, so that the intracellular hydrogen ion concentration would
become greater with the larger amount of acid. On the other hand
there is an alternative possibility that the buffering capacity of the pene-
trating acid solution is one of the factors concerned in the magnitude
of the changes produced. That buffering capacity is of importance
in these experiments may be inferred from the work of Chambers
(1928), Reznikoff and Pollack (1928) and Pollack (1928), who find
that various cells have considerable buffering power, sufficient indeed
to prevent changes in the colorimetrically determined pH of the hyalin
protoplasm, within the observational limits of ±0.1 pH, when treated
with sublethal concentrations of CO2. Hence it would be expected that

2 See Stewart, BioL Bull, in press.

146 EVELYN HOWARD

the production of an acid liquefaction in the present experiments is not
associated with an equalization of the intracellular and environmental
pH. but only with a certain shift of the pH of the cell. The degree
of this shift would he a function not only of the pll itself, but also of
the buffering capacity of the penetrating components of the medium.
Consequently, on this basis, one would expect the results obtained in
this experiment, and such effects cannot be considered as evidence that
the salt penetrates less readily than the acid.

A POSSIBLE MECHANISM HY \\.IIK 11 THE SALTS OF PENETRATING ACIDS

MIGHT ENTER THE CELL

The apparent permeability of the cell to the salts of the penetrating
fatty acids presents a rather surprising contrast to its relative imperme-
ability to certain other ionixed compounds. This situation might be
due to specific differences in the behavior of the different ions toward
the cell membrane, but there is no independent evidence that the physical
properties of the salts of the various acids are related to the physical
properties of the acid molecules to an extent which would account for
these physiological peculiarities.

An alternative explanation may be developed from the following
considerations. In general, ionixed compounds do not penetrate cells
at all freely, but it is possible that the restrictions to the passage of
ions in many cases are imposed chiefly on ions of one sign. The ery-
throcyte, for example, uiukr ordinary physiological conditions, appears
to be a cell which is permeable to anions and impermeable to cations
Cfor references sec Warburg. l'L'2 and Van Slyke, \Yu. and McLean.
T'23). The fact that the erythrocyte is freely permeable to ammonium
chloride and similar ammonium salts lias been explained by Jacobs
i l''J7) as being due to the free penetration of the- cell by NH;, and
subsequent exchange of the internal ( )1 I ions for the external anions
of the salt. With the majority of cells, it is possible that the conditions
prevailing in the erythrocyte may be reversed, such cells under ordinary
conditions being more permeable to cations than to anions. The cvi-
dcnn- for other cells is less direct than in the case of the erythrocyte,
and is chiefly poteiitiometric. the sign or magnitude of the potentials
measured across membranes responding to changes in the concentration
or character of the salt present in the manner which would be expected
it the am >\\\<\ not pass through the membrane while the cations

tended to diffuse. The application of the principles involved has been
discussed by Michadis and others (1925, 1(>27, etc.). Potentiometric
evidence of this nature has been reported by Loeb and Beutner (1911)

CELL PERMEABILITY TO IONS 147

and by Fujita (1925) for the apple skin, by Moiid ( 1927) for the wall
of the stomach, by Amberson and Klein, (1928) for the frog skin, and
by Sunnvalt (1929) for the chorion of the Funditlns egg. Michaelis
and Fujita (1925) have furnished in addition chemical evidence of
the permeability to cations and the impermeability to anions of the
apple skin. It is generally known that a normal surface is positive
to an injured area in nerve, muscle, and apple, and this sign of the
demarcation potential is not incompatible with a greater permeability
of the normal surface to cations than to anions. Furthermore, Moiid
and Amson, (1928), from analyses of perfusing fluids, concluded that
frog muscle is permeable to certain cations, e.g., K- and Cs-, but is im-
permeable to chlorides.

If the Arbacia egg is permeable to certain cations, its apparent per-
meability to the salts of penetrating acids might result from the en-
trance of the cell by the organic acid in the undissociated form followed
by a transfer of cations, the external base being exchanged for the
hydrogen ion derived from the intracellular dissociation of the acid.
This process would result in the presence of acid anions and additional
base inside the eel! without an actual transfer of anions across the mem-
brane. There is evidently an analogy between this mechanism and
that of the penetration of the anion-permeable erythrocyte by a salt
whose cation enters in an indirect manner. However, conditions are
not completely analogous, since the penetration of the erythrocyte by
ammonium chloride causes marked osmotic swelling, whereas this does
not appear to be the case in the penetration of the Arbacia egg by po-
tassium ^acetate.

If the transfer of base takes place by an exchange of alkali cations
from without for hydrogen ions from within, the mobility of the cations
in passing through the membrane might be the limiting factor in the
rate at which the intracellular buffering action of the salt becomes mani-
fest. Michaelis and collaborators (1925-1927) have observed that in
the dried collodion membrane the mobility differences of the cations
are greatly exaggerated. Netter (1928), Brooks (1930), and others
have applied these considerations to explain the accumulation of K by
many cells as a result of a cation exchange of hydrogen ions, produced
metabolically, for K ions from the outside rather than for Na ions, since,
although the latter ions are present in much the greater concentration
in the environment, their diffusion may largely be restrained by a mem-
brane which permits the passage of K ions. The presence of a mem-
brane of this type in Arbacia eggs is suggested by the high ratio of
K to Na found in the cells in comparison with that in sea water.
Blanchard (unpublished results) reports a ratio of equivalents of K
to Na of 1.90 in Arbacia eggs, while in sea water the ratio is 0.0213.

148

EVELYN HOWARD

If this property of the cells results from characteristics of the mem-
brane, it would be expected that, in the exposure of the cells to an
acetate buffer in the balanced saline media, the base transferred would
be K rather than Xa, and that consequently changes in the K concen-
tration of the medium would affect the rate of entrance of the base,
and therc-li\- the degree to which a given amount of salt is able to exert
an intracellular buffering action after short exposures. This was ob-
served to be the case under the following conditions. Certain amounts
nf 0.05 X XaCl, KC1, or RbCl were added to solutions of acetate buffers,
with free acetic acid 0.00059 X at pH 5.0, in the usual saline media.
As already mentioned, variations in K concentration of this magnitude
in the absence of acetates produced no significant changes in sigma.
The effects on the degree of liquefaction produced during the first two
to nine minutes of exposure to such solutions are shown in Table II.
The results are given in sigma reduced to a common basis by taking
the values obtained in the high K concentration as twenty sigma.

TABLE II
The Effect of Various Cations on the Liquefaction Produced, by an Acetate Buffer

Concentration of

chlorides added

Kb

K

N'a

molslltter

0.002

—

20

35

0.002

7.5

20

27.5

0.01

14

20

28

0.004

13.5

20

—

0.003

—

20

32

In these experiments the absolute differences are small, but the rela-
tive effects of the cations always appeared in the order given at this
I'll. At higher pi I values, where the salt reversal of the pH effect

Hired, the cation series was in most case's reversed also. However,
if we consider only the pi I range where the salt appears to exert a
buffering effect which is not complicated by other factors, the degree
of initial liquefaction produced is in the order Rb<K<Na. After
longer exposures, there is a tendency for the same equilibrium point
to be reached, but this is not maintained for over twenty minute's in
some cases. It will be noted that the series obtained is compatible
with the view that the degree to which the salt is able to buffer the
acid intracellularly during the first few minutes of exposure, and there-
fore the rate of penetration of the salt, increases with the amount of
the more highly mobile cations present. Although physical properties

CELL PERMEABILITY TO IONS 149

other than mobility may he concerned, nevertheless the effect is sug-
gestive of an actual penetration of base rather than a removal of acid
as suggested by Smith.

When the amounts of acetate in the medium are as high as 0.1
molar, such a cation exchange would probably result in an abnormally
great accumulation of K in the cell relative to Ca. Chambers and Rezni-
koff (1926) have observed that the injection of NaCl or KG into cells
causes a liquefaction, whereas Ca causes a coagulation. This obser-
vation suggests that the liquefaction produced by the acetate salt may
be due, in part at least, to the accumulation of K.

DlSCUSSIO'N

Evidence has been presented that the salts of penetrating acids are
able to enter sea urchin eggs within four minutes or less, although
certain other ionized compounds are apparently not able to penetrate
these cells. The evidence is based on the observation that the salt of
a penetrating acid alters the degree to which the acid is able to affect
the apparent viscosity of the protoplasm, an effect presumably exerted
by virtue of the influence of the acid-salt mixture on the intracellular
hydrogen ion equilibria.

This evidence of the penetration of cells by the salts of penetrating
acids supports the conclusions arrived at by Smith with other* methods
on marine ova and cardiac muscle, and by Haywood on skeletal muscle,
so that this type of permeability seems to be found in a certain variety
of cells. These findings differ from those of Jacobs, Lillie, Osterhout,
and others on a number of other cells, and it seems reasonable to suppose
that the different results may be due to rather considerable differences
in the relative rates of penetration of certain acids and their salts into
the various types of cells in question. Since, however, it was noted
by Jacobs that the differential effects produced by CO2-bicarbonate
mixtures on Sviuphyfuin and on starfish eggs tended gradually to dis-
appear, the difference is probably only a quantitative one. This view
is substantiated further by the observation of M. M. Brooks (1923)
that bicarbonate penetrates into Valonia after eighty minutes' exposure.
Lillie's observations on starfish eggs were limited to exposures of five
to fifteen minutes, and under these conditions the pH of the medium
usually seemed to be a less important factor physiologically than the
changes in the concentration of acid available for diffusion into the cell ;
however, the experiments do not exclude the possibility of the penetra-
tion of salt.

The present observations of the effects of acids on protoplasmic

150 EVELYN HOWARD

viscosity have been discussed in terms of permeability. It has been
shown by other authors that many organic acids diffuse through certain
living tissues where mineral acids do not. Since organic acids applied
externally cause color changes in granules stained by indicators in marine
ova although mineral acids do not have this effect, it is probable that
the fatty acids actually penetrate. If fatty acid enters marine ova. it
is assumed, until proven otherwise, that the salt penetrates and buffers
the acid directly in the cell, rather than that it acts from a distance in
some unknown manner. Strictly, the term penetration has only the
justification of being a convenient description of the selective physio-
logical reactivity with which the present study is concerned.

SUMMARY

The effects of the pi I of the medium on the apparent viscosity
of Arbacia eggs have been studied by means of an adaptation of the
centrifuge method.

It was found that the degree to which fatty acids de-crease the
protoplasmic viscosity can be altered by the presence of the salt of the
acid, apparently by virtue of the influence of the salt on the intracellular
hydrogen ion equilibria. Similar pi 1 variations of the medium produced
by mineral acids do not affect viscosity.

These results offer confirmation, from a separate type of evidence,
of Smith's observation that the salts of peijetrating acids differ from
other ionized compounds in apparently being able to penetrate the living
cell much more easily. It is suggested that cell permeability to this
type of salt could be explained as resulting from a cation exchange of
external base for internal hydrogen ions from the penetrating acid, the
entrance of the salt therefore being accomplished without the direct
transfer of its anions.

It is a pleasure to acknowledge my indebtedness to Dr. M . II. Jacobs
for the suggestion of this problem, advice during its development, and
careful criticism of the manuscript.

BIBLIOGRAPHY

. YY. K.. AND KLEIN, H., 1928. Jour. Gen. Ph\s'wl. 11: 823.
i. 1. G., I1'-"'. I'rotnplasiiia. 7: 505.
BKERMV. 11, 1(>J4. Join: Il.v^cr. ZooL. 41: 33.
BKTIII. Y. ]'Ht>. I'jliificr's Arch. 127: IV).
BOI.INF.. .1. II.. l''_>5. Jour. Gen. Physwl. 7: 10.
RRIXLEY, P.. l"_'s. I'mtoplasmii. 4: 177.
BROOKS. M. M.. 1 ';_>.?. r. S. Public Health Ref>.. 38: 1470.
BROOKS. S. C., 1('3(). Protoplasma, 8: 389.

CHAMBERS. R., VND K'I/ IKOFF, P., 1926. Jour. Gen. Ph\sloL. 8: 369.
CHAMBKRS. R.. 1928. B'wl. Bull. 55: 369.
COHEN, B.. AND CLARK, W. M.. 1919. Jour. Ractcr'wl.. 4: 409.

CELL PERMEABILITY TO IONS

CROZIER. W. J., 1916. Jour. Bio!. Clicm.. 24: 255.

DRUCKER, K.. 1905. Zcitschr. Phys. Chen;.. 52: 641.

EDWARDS, J. G., 1923. Jour. /I.r/vr. Zool., 38: 1.

FUJITA. A.. 1925. Biochcm. Zcitschr.. 158: 11.

HARVEY, E. N., 1911. Jour. E.rfier. Zool.. 10: 507.

HAYWOOD, C, 1927. Am. Jour. Physiol, 82: 241.

HEILBRUNN, L. V., 1927. Quart. Rci\ Biol., 2: 230.

HENDERSON, L. J., AND COHN, E. J., 1916. Nut. Acad. Sci. Proc., 2: 618.

JACOBS, M. H., 1920<7. .-/;;/. Jour. Physiol, 51: 321.

JACOBS, M. H., 1920k. Am. Jour. Ph\siol. 53: 457.

JACOBS, M. H., 1922o, Jour. Gen. Physiol., 5: 181.

JACOBS, M. H., 19226. Biol. Bull.. 42: 14.

JACOBS, M. H., 1924. Chapter in Cowdry, General Cytology, Chicago.

JACOBS, M. H., 1927. The Harvey Lectures, Series 22: 14o.

LILLIE, R. S., 1926. Jour. Gen. Physiol, 8: 339.

LILLIE, R. S., 1927. Jour. Gen. Physiol.. 10: 703.

LOEB, J., 1909. Die Chemische Entwicklungerregung des lierischen Eis. Berlin.

LOEB, J., AND BEUTNER, R., 1911. Science. 34: 884.

MICHAELIS, L., AND KROGER, R., 1921. Biochcm. Zcitschr., 119: 307.

MICHAELIS, L, AND FUJITA, A., 1925. Bioclicni. Zcitschr., 158: 28, and later

papers of this series.

MICHAELIS, L., 1925. Jour. Gen. Physiol.. 8: 33.
MICHAELIS, L., AND PERLZWEIG, W. A., 1927. Jour. Gen. Physiol., 10: 575, and

later papers of this series.
MONO, R., 1927. Pfliiger's Arch.. 215: 468.
MONO, R., AND AMSON. K., 1928. Pflih/er's Arch., 220: 69.
NETTER, H., 1928. Pfliiger's Arch.. 220: 107.
OSTERHOUT, W. J. V., 1925. Jour. Gen. Physiol.. 8: 131.
OSTERHOUT, W. J. V., AND DORCAS, M. J., 1926. J our. Gen. Physiol., 9: 255.
POLLACK, H., 1928. Biol. Bull.. 55: 383.
REZNIKOFF, P., AND POLLACK, H., 1928. Biol. Bull.. 55: 377.
SMITH, H. W., AND CLOWES, G. H. A., 1924<r. Biol. Bull.. 47: 323.
SMITH, H. W., AND CLOWES, G. H. A., 1924/>. 'Am. Jour. Physiol., 68: 183.
SMITH, H., 1925. Am. Jour. Phvsio!.. 72: 347.
SMITH, H., 1926. Am. Jour. Phvsiol.. 76: 411.
STEWART, D. R., 1929. Anat. Rec., 41: 31.
SUMWALT, M., 1929. Biol. Bull.. 56: 193.
VAN SLYKE, D. D., 1922. Jour. Biol. Chem.. 52: 525.

VAN SLYKE, D. D., Wu, H., AND MCLEAN, F. C, 1923. Jour. Bio!. Chem., 56: 765.
WARBURG E. J., 1922. Biochcm. Jour.. 16: 153.
WARBURG, O., 1910. Zcitschr. f. physio! . Chem., 66: 305.

THE I '!•: KM I-.ABILITY OF THE ARBAC1A EGG TO NOX-

ELECTROLYTES

DOROTHY R. STEWART

( /><<w the Department of I'liysinlogy, University nf Pennsylvania and the Marine
Biological Laboratory, Woods Hole, Massachusetts)

I

The most extrusive systematic studies of the permeability of cells
t<> dissolvetl substances have been made upon plant material. Par-
ticularly noteworthy in this connection is the work of Overtoil (1895,
1907) and of P.arlund ( 1929) (see also Collander and Barlund. 1926).
Ann m- animal cells the mammalian erythrocyte has proved to be espe-
cially useful and has been most frequently employed (Gryns. 1896;
Hedin. 1897; Hamburger. 1<M)_>; M.md and Hoffman. 1'L'S; Fleisch-
man. 1('J8. and others). Hecause of the highly specialized nature of
this latter type of cell, however, it is desirable, before using the results
obtained with it as a basis for extensive generalizations, to supplement
them with similar observations upon a variety of other animal cells.
The present paper deals with experiments upon the relative rates of
penetration of nineteen different non-electrolytes into a cell of animal
origin whose permeability to dissolved substances has not hitherto been
systematically studied, namely, the egg of the sea urchin. Arbac'm
punctulata.

Among the peculiarities which make the Arbac'm egg a suitable form
of material for such investigations are its spherical shape, which permits
a study of permeability by direct measurements of volume changes, its
rnnvenient and almost uniform size, its even and not too rapid rate of
swelling, and the large numbers in which it may be obtained from a
single female. Its chief disadvantage lies in the amount of inert yolk-
material which it contains. This cell was first employed for perme-
ability studies by Lillie (1916), who compared the rates, before and
after fertilization, at which water entered the egg from anisotonic solu-
tion^. I'urther extensive work upon its permeability to water, under
a variety of conditions, has been done by McCutcheon and Lucke (1926
and subsequent papers in the same journal).

Though the present investigation is concerned primarily with the
penetration ot dissolved substances rather than of water, the actual
nature of the observations is the same in the two cases, consisting in

152

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 153

measurements of the progressive changes in the diameter of the eggs
during the course of the experiment. In the work of the previous in-
vestigators the movement of water, and the volume changes incident
to this movement, were determined solely hy the initial osmotic differ-
ences between the eggs and the surrounding solution. In these experi-
ments there are, in addition, such further osmotic inequalities as are
set up by the diffusion of the dissolved substance into the eggs.

According to the experimental procedure employed, the volume
changes accompanying penetration of a solute may be of two types.
The first involves swelling in a pure solution of a penetrating substance
and should theoretically continue indefinitely, but actually results in the
eventual cytolysis of the egg. The second occurs in solutions containing
both penetrating and non-penetrating substances; in it, following a
preliminary shrinkage, the swelling continues until the cells have at-
tained an equilibrium volume determined by the concentration of non-
penetrating substances in the solution.

In Fig. 6A are represented typical curves of the former type, ob-
tained by exposing eggs to solutions of ethyl and methyl alcohols ap-
proximately isosmotic with sea water. They illustrate the rapid and
continued swelling of the eggs in a pure solution of a readily penetrat-
ing substance. Figure 6R, for the same substances dissolved in sea
water, represents the other type of volume change which occurs in solu-
tions that are initially hypertonic, clue to the addition of some non-
penetrating substance. Both types of procedure have been used in the
present experiments, but the second has been found to be more generally
satisfactory.

The actual technique used in handling the eggs in these experiments
was essentially that described by McCutcheon and Lucke (1926). Eggs
from a single female were obtained with due care to avoid contamination
by either sperm or intestinal debris. They were rinsed several times
in filtered sea water and allowed to stand, without crowding, in dishes
set in running sea water. From these they were removed, as needed, by
means of a capillary pipette. Ordinarily about 100 eggs, in the smallest
possible quantity of sea water, were used for each experiment. All
measurements were made at room temperature, which varied from 21°
to 25° C. during the summer, but not over a degree on any given day.
The temperature of the running sea water in which the eggs were set
averaged slightly lower, varying from 20.5° to 23° C.

In each experiment a watch-glass containing 7 to 8 cc. of the solution
to be tested was placed on the mechanical stage of a microscope, the
eggs introduced by means of a capillary pipette, and the water immersion
lens quickly focussed. A group of five spherical eggs was selected and

154 DOROTHY R. STEWART

their diameters recorded by means of a filar micrometer. The total
magnification afforded by the system used was 188 diameters. The
eggs in each group were always measured in the same order and at as
nearly equal time intervals as possible. Eggs which were not changing
markedly in si/.e could be measured with an error of no more than
0.7 per cent of their total diameters, and even if readings were taken
at maximum speed on rapidly swelling eggs, the error was not greater
than twice this amount.

The necessary pi I determinations were made colorimetrically, using
phenol red and bromocresol purple as indicators. It was not always
possible to keep the reaction of a given solution constant during an
experiment, but readings were taken at the beginning and end of each
series and the results were discarded if the variation was found to be
too great. The eggs do not seem to be osmotically sensitive to small
changes in pll, as such, in regions not far removed from neutrality, so
that a variation of about 0.2 of a pll unit in half an hour was considered
allowable. Xo correction was made for the salt error of the indicators,
since it was merely desired that solutions which were to be compared
should lie made up to reproducible pll values.

The experimental results have been presented in the form of curves,
plotting diameter against time. Kach point on a curve represents the
average diameter of the five eggs under observation at that time. .All
the curves in any single figure were obtained on the same day and with
eggs from the same female. Since volume is a function of diameter,
the latter was used in plotting the curves in order to avoid laborious cal-
culations. Actual figures for the volumes were obtained, however, for
the preparation of Table I.

Since the eggs deteriorate on standing, only a limited number of
experiments can be done on those from any single female. On the other
hand, different lots of eggs vary enough in their size and permeability
so that results on material from different animals cannot fairly be
compared. As a result of these difficulties, the relative rates of pene-
tration of a long series of compounds can be arrived at only by piecing
t' Aether the results of a number of shorter, overlapping series.

I'.vrn in the same lot of eggs it is desirable, where accurate com-
parisons are to be made, that the particular group selected for measure-
ment in each solution shall agree in average initial si/e and final emii-
librium volume. Since the eggs in each set must be selected at random
in a very short time, and averages cannot be obtained until the end of
the experiment, it frequently happens that the observations must be
repeated several times before an unobjectionable comparison is possible.
The discarded results are not a total loss, however, since, while they

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 155

are not in themselves conclusive, they furnish a certain amount of
confirmatory evidence. A considerable number of experiments of this
type have been performed but are not reported in this paper, though
their results are in general agreement with those actually cited.

The best method of presenting the experimental data would oh-
viously be by the use of complete swelling curves. Since it is impossible,
however, to reproduce here more than a few of the curves that have
been obtained, it has seemed best to make comparisons of the times
required in different solutions for the eggs to reach some arbitrary
degree of swelling. A convenient volume increase for most substances
is that represented by a size halfway between the initial and equilibrium
volumes ; for more slowly penetrating substances 5 per cent or 20 per
cent increases have sometimes been employed. While this method of
presenting the results admittedly leaves something to be desired, it
should be remembered that in every case the entire curve has been
plotted and used as a basis for the conclusions drawn with regard to the
behavior of the various substances.

II

The results of the most conclusive experiments on the relative rates
of penetration of different substances are presented in tabular form
in Tables I, II, III and IV. Additional features of a few of the ex-
periments are shown graphically in Figs. 1 to 6, inclusive. In the fol-
lowing more detailed description of these results the various substances
studied are grouped according to their chemical relationships.

1. WATER

(50 experiments.) Experiments involving the entrance of water
alone were run frequently to serve as controls for those with other
substances. In Fig. 4 and in Tables I, II and III are given the results
of a number of such experiments. They showed comparatively slight
variation from day to day during two successive summers, the time
for swelling halfway to equilibrium in 60 per cent sea water averaging
very close to four minutes, with maximum variations of from 2.6 to
5.8 minutes. These figures agree well with those obtained by Mc-
Cutcheon and Lucke (1926) under similar experimental conditions.

2. MONOHYDRIC ALCOHOLS

The two monohydric alcohols that were tested both entered Arbacia
eggs with great ease. Cells placed in molar solutions of methyl and
ethyl alcohols in distilled water doubled their volumes in three or four
minutes (Fig. 6). If the alcohols were dissolved in sea water, swelling

156

DOROTHY R. STEWART

10 20

»Q -to

FIGS. 1-6: The swelling of Arbacia eggs in various solutions. The abscissae
arr times in minutes; the ordinatcs are diameters in arbitrary units, each of which
<iual to 1.0o4M. Each point represents the average diameter of five eggs except
in a lYw instances when ten eggs were measured. In all the experiments except
1 and o. I there was a ten-minute preliminary shrinkage of the eggs in a solution
of i ! made up in the same way as the solution to be tested. pH 8.2

1. Glycerol (•), crythritol (X) and dextrose (O) ; molar solutions made
tip in <>(i P<T (rut sea water. Fertilized eggs, ten eggs measured for each point.

J. l-'t!i\ !cnc ^lycol (•), urea (O) and glyccrol (X); molar solutions in
MI p. r cenl ;ea \\ au-r.

.v Moiiai-rtin (•). diacetin (X) and glycero! (O) ; molar solutions in sea
wat' Solutions brought to api>roximately pH 7.0.

4. l'.Mt\ raimdr (X)i propionamide (•) and acetamide (A), made up to half
molar concentration in od per cent sea water; (O) 60 per cent sea water alone.
The two bn .il. . in tin- Lutyramide curve correspond to times when abnormalities

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 157

was also rapid, being- faster than for glycerol, urea, ethylene glycol,
acetamide or formamicle, and at al)out the same rate as that for monacetin
or for water alone. Some of the more toxic compounds such as
ethylene glycol mon- and diacetates, cliacetin and monochlorhydrin
seemed to show a more rapid penetration, but this was perhaps due to
injury effects on the eggs.

Alcthyl Alcohol (16 experiments) regardless of solvent, invariably
entered the eggs more slowly than did ethyl alcohol, though the differ-
ences were slight (Fig. 6). In molar solutions in distilled water the
average of the experiments gave, for the times necessary for a 50 per
cent increase in volume, 2.9 minutes for ethyl and 3.9 minutes for
methyl alcohol (Table IV). Another experiment in which molar solu-
tions in distilled water were used gave values of 4.1 and 3.2 minutes,
respectively (Table I).

Ethyl Alcohol (10 experiments), as just mentioned, gave a rate of
swelling slightly greater than that for methyl alcohol. Sometimes this
rate appeared to be even faster than that for water alone ; in any case,
it was not significantly slower.

3. Dl- AND POLYHYDRIC ALCOHOLS

Three alcohols belonging to this group were studied and were found
to be slower in their rates of penetration than the monohydric alcohols ;
they covered the range from moderately rapid entrance to apparently
none at all, as the size of the molecule and number of polar OH groups
were increased.

Ethylene Glycol (70 experiments) is the most satisfactory to work
with of all the compounds tested, since it seems to be entirely non-toxic
to the eggs when dissolved in sea water and only very slightly so in other
solvents. It has been used in solution in sea water, 60 per cent sea water,
M/3 CaClo, M/2 KG, M/2 NaCl and distilled water. Its rate of pene-
tration is intermediate between that of substances such as glycerol or
urea, which enter very slowly, and those of the large group of rapidly
penetrating substances of which ethyl alcohol and diacetin are typical.
In solutions made up in 60 per cent sea water approximate equilibrium
is attained in about two hours, a period long enough so that the swelling
process can be studied in detail. The rate of penetration from sea water

occurring in the original eggs necessitated shifting the observations to different
individuals.

5. Fertilized (•) and unfertilized (O) eggs in ethylene glycol made up to
molar concentration in 60 per cent sea water; fertilized (X) and unfertilized
(A) eggs in 60 per cent sea water alone.

6. Ethyl (O) and methyl (X) alcohols. A, molar solutions pH 7.0; B, molar
solutions in sea water.

158 DOROTHY R. STEWART

solutions has been compared with that of each of the other substances
investigated. < M the- results obtained, two sets which permit com-
parisons with the very slowly penetrating urea and glycerol (Fig 2)
and with water > Fig. 5 ) have been reproduced in graphic form.

Because of their almost identical molecular weights, it is interesting
that the rates for urea and ethylene glycol should be very different, a
20 per cent increase in volume in one experiment (Table 1) requiring
4.S minutes for ethylene glycol and 100 minutes for urea. In another
experiment a 5 per cent increase in volume occurred in about two min-
utes in the one case and in thirty-six minutes in the other. A number
of additional experiments are included in Tables 1. II. Ill and IV. cov-
ering comparisons between ethylene glycol and fourteen other com-
pounds. It will be seen that the time required to swell halfway to equi-
librium in molar solutions of ethylene glycol in sea water varied con-
siderably, i.e. from 7.5 to 20.5 minutes, and averaged about twenty
minutes as compared with an estimated time of possibly as many hours
for glycerol and one to five minutes or less for most of the other
com] M iniids.

Glycerol (18 experiments) penetrates Arbacia cgg> slowly, but at
a measurable rate1, as is shown in Fig. 1. The eggs cea>e to swell in
hypotonic sea water in about an hour, but with glycerol present in the
external solution they do not live long enough to attain their equilibrium
volume; apparently more than twenty- four hours \\ould be needed for
the completion uf the process. Several of the experiments summarized
in Table I give actual figures for the times necessary to increase the
volumes of the eggs by a definite amount in glycerol solutions, as com-
pared with the- corresponding values for water, formamide, acetamide,
urea and ethylene glycol. It will be observed that of these substances
glycerol is by far the most slowly penetrating, an hour being required
in its solutions for a 5 per cent increase in cell volume, which occurs
in less than half of that time in urea solutions and in less than three
minutes in water or in solutions of formamide, acetamide or ethylene
glycol.

Erythritol (3 experiments), the highest member of the series of
polyhydric ali"ho1s studied, does not enter .-Irhucia eggs to an appreci-
able extent either from pure solutions or in the presence of salts, during
an exposure of five hours ("Fig. 1 ).

4. SUBSTITUTION PRODUCTS OF (ILYCEROI, AND ETIIYLFXF, di.YCoL

I he i nent of one' or more Oil groups in glycerol by either

a chlorine atom i,r by an acetic acid radical greatly increases the ease
with which the compound enters .•Irhncia eggs, but at the same time in-
creases its t. to the eggs.

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 159

Momicctin (9 experiments), in which one Oil group of glycerol
has been so replaced by an acetic acid radical, is the least toxic of the
substitution products that were tested. The effect of the substitution
in question is most striking, the time required for a 20 per cent increase
in volume being reduced from an average value of 250 minutes to one
of approximately two minutes. Monacetin enters eggs far more easily
than ethylene glycol but somewhat more slowly (i.e., one and one-half
to two times) than diacetin (Fig. 3). The experimental results ob-
tained varied so much that it was impossible to compare its behavior
quantitatively with that of the monohydric alcohols. It appeared on
the whole to penetrate more slowly than ethylene glycol mon- and
diacetates, or than monochlorhydrin, all of which are rather toxic
substances.

Diacetin (17 experiments). The addition of a second acetic acid
radical still further increases the ease of penetration of the Arbacia
egg. "\Yhen made up in sea water, the rate of penetration of diacetin
is very close to that found for ethyl alcohol, ethylene glycol mon- and
diacetates and faster than that for monochlorhydrin and monacetin.
It is also faster than that of any of the amides. The complete series :
glycerol < monacetin < diacetin is represented in Fig. 3.

Monochlorhydrin ('4 experiments). In this compound a chlorine
atom replaces one hydroxyl group of glycerol, thereby increasing the
penetrating power of the substance to an extent almost exactly the same'
as with monacetin. Monochlorhydrin is more toxic than monacetin,
however, and this may perhaps account for part of the increase (Table
III).

Eth\lci\e Glvcol Monacctatc (4 experiments) and Ethylene Glycol
Diacctatc (2 experiments) were both tested, but satisfactory results
could not l)e obtained because of their toxicity. It seems probable that
they both penetrate rapidly, though not so readily as the figures in Tables
III and IV would seem to indicate, since any toxic effect increases the
apparent rate of penetration.

5. SUGARS

Experiments with these compounds were difficult to perform in the
usual manner since the eggs tended to float in the solutions. By using
sufficiently dilute solutions of the sugars in sea water, however, it was
possible to show that no measurable penetration occurred.

Sucrose (3 experiments), as far as could be judged from such
measurements of diameter as are possible on floating eggs, failed to
enter the cells from a pure molar solution. It also gave no evidence
of penetration when made up to lesser concentrations in sea water.

Dextrose (7 experiments). Experiments with this substance were

11

160 DOROTHY R. STK\YART

continued for eight hours or more without any trace of its penetration
either in the presence or in the absence of salts in the external solution

(. Fig. 1 ) -

6. AMIDES

The four amides used form an instructive homologous series; but.
mi fortunately, they prove to be rather difficult to work with. \Yhen
molar Dilutions in distilled water are used, toxic effects invalidate the

lilts. If the solvent is 60 per cent sea water, then it is necessary,
in order to make the comparison a fair one. that all of the curves should
end at the same equilibrium volume. Ordinarily this can be accom-
plished by simply repeating any experiment which does not come out
at the proper level, but with the amides there are further difficulties
of a chemical nature which may now be mentioned.

Formamide (33 experiments), in particular, gives solutions which
arc strongly acid in reaction. Adjustment of the pi I necessitates the
addition of so much NaOH that the concentration of non-penetrating
substances in the external solution is definitely increased and the volume
of the eggs at equilibrium correspondingly decreased. In one experi-
ment in which the equilibrium value in a so-called molar solution of
formamide in N) per cent sea water was compared with those for a series
of various dilutions of sea water, it was found that the final volume
of eggs in the formamide solutions was the same as that in 68 per cent
sea water instead of 60 per cent. < )n the other hand, if the solution
is diluted so that its osmotic pressure is correct, then the concentration
of formamide is decreased. I'.ecause of these difficulties, it is impos-
sible to compare this substance satisfactorily with the other compounds.
The figures given in Tables I. II and IV show the considerable varia-
tion that occurs. At times formamide seems to enter the eggs more
rapidly than acetamide, but in other experiments more slowly.

Acetamide (17 experiments) gives solutions which are somewhat
acid, but it does not present the extreme difficulties that were encountered
with formamide. The penetration of acetamide was found to be more
rapid than that of glycerol. urea and ethylene glycol, but slower than
that of any of the other compounds studied, with the' possible exception

Formamide. The time required to reach the halfway point on the
swelling curve is about ten minutes for molar solutions of this substance
in (>H per rent sea water — a value only half of that for ethylene glycol
i I ables F. II and IV), but nearly twice that ordinarily obtained with
the least toxic of the more rapidly penetrating substances such as the
moiiohydric alcohols.

Propi on amide (11 experiments) enters Arbacia eggs more rapidly
than acetamide (Tables I and II; Ki^. I I but at a slower rate than

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 161

butyramide. It has not been compared directly with any compounds
except the other amides and water, but the average values for the time
in minutes for reaching a point halfway to equilibrium in such series
was found to be : water, 4.5 ; butyramide, 4.9 ; propionamide, 6.3 ;
acetamide, 10.3. Formamide is omitted from the series for the reasons
already mentioned.

Butyramide (7 experiments) enters the eggs at a rate which is con-
sistently faster than that for propionamide. Swelling in its solutions
is, however, noticeably slower than in water alone. Results such as
are given in Fig. 4 have been duplicated several times. Additional
experiments are given in Table II.

7. AMINO ACIDS

Glycocoll (4 experiments). Eggs left for twelve hours in solutions
of glycocoll made up in 60 per cent sea water failed to show any in-
dications of swelling.

8. OTHER COMPOUNDS

Urea (27 experiments) penetrates at a rate which is much the same
as that of glyceroi, though usually not quite so slow. The lower curves
in Fig. 1 illustrate the swelling of the eggs in solutions of these two
substances and Table I gives the actual times necessary for 5 per cent
and 20 per cent increases in volume in solutions of each substance. It
must be remembered in interpreting these figures that the slopes of
the curves are very slight and small differences in slope will mean
large differences in the times required to reach a given volume. Table
I also gives comparative values for urea and water, formamide, acetamide
and ethylene glycol. In Table III, additional comparisons are made,
but urea experiments could not be continued long enough to obtain
comparable figures.

The general behavior of urea is not that of an inert substance.
Instead, it acts as if it had a definite effect, either on the membrane
or on the protoplasmic consistency of the egg. Most of the eggs placed
in pure urea solutions (molar in distilled water) lose their spherical
shape after 20 to 50 minutes and become distinctly irregular. Later,
" pseudopodia " may form and the egg may move much as an amoeba
does. After a time it usually rounds up and becomes perfectly spherical
again — only perhaps to put out " pseudopodia " once more, after a little
while. This peculiar behavior of the eggs is considerably less marked
if the urea solution is made up in sea water.

Ethyl urcthane (3 experiments). This substance is so toxic to the
eggs that no satisfactory results could be obtained with it. It probably
penetrates rapidly. Reference to one experiment which was fairly
good is made in Table I.

DOROTHY R. STEWART

c
^ £•

"

i ~

p «

t- t)

~ .
'- o

N

~s '7.

•^ r.

•^™ Pj

'~ ZJ

— V.

" O

1°

I— I

W § O

.=: ~

- £

- V

- '7

— V

'

apiiucuuoj

tr- - I"M"-'I

v! U •£

2 t '-

= '"

•£. U

01 —

-t

00 01

H —
01

Ol IO

0?

8

*

o

0

5
1.

'-

1/3

01

IT)

>o

•ri

*

o
es

01

O
-t

T— t

o

u-j o

0
01

01

3

t+

<

-
4.

*

*— i

c

L

*

*

* — r
-t

q
i

M

10

o

00

10

00

_

r- oo

~ /

01

^

10

4

tfi

ro

10

10

-f

-t-

^^

10 .-<_

10

"

-t

i

r

01

. —

r

01

oo

"^«

01

01

o"
1—

Ol

01
01

01
01

ON

01
01

01

o"

OJ

01

^2.

c>

o"

d

o'

•
(

V*

t/!

O

o

U3

rt
u

_a
o

D

cr

._ o

^ £

•3 •_=

^ bjo
ri bfl

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 163

apiuiwioulojj

8

a

w
1-1

asojong

i 10

o"

CN

OO

q

06

(N

10

o

o

O CN

o\

00

0

^2.

00

o o
o"

CN

I

60

00

CN
CN

o

00

C3
r<3

o"

oo"

164

DOROTHY R. STEWART

TABLE II

Time in minutes required for eggs to well to a point halfiuay betzueen their
initial and equilibrium volumes. Substances made up to half molar strength in
sea water.

Water

Etliylene
Glycol

Formamide

Acetamide

Propionamide

Butyramide

7/12/20. . .

5.8

23.5

7/18/29...

3.5

6.2

7/19/29...

5.0

14.0

7/20/29 . . .

5.8

14.2

8/10/30. ..

6.6

4.0

8/11/30. ..

4.4

6.2 f

6.5

4.9

8/12/30...

4.7

9.7

10.5

11.0

5.7

5.7

8/13/30. . .

4.5

6.9

9.0
12.2

6.3

5.2 f

8/15/30. ..

2.8?
4.3

9.0?
9.7?

9.1

6.5

4.9

8/16/30. ..

3.7
3.6

6.1

t Eggs not swelling to the same equilibrium volume as in other solutions.

TABLE 111

Time in minutes required for eggs to sivcll to a point halfzvay between their
initial and equilibrium volumes. ( ) time required for a 5 per cviit inrrease in
Substances made up to molar strength in sea WUUT.

o
u

o
u

X

O

o
a

o
u

!

03

u

°S

01

'C

CJ ^"*

flj "

• PM

•o

2*

Diacetin

Jl §
|S
u

II

»•*• ^r

w

Formam

Acetami

§1

1

- 9 29

5.3

1.5

16.5

CO

3.1

1.7

L.O?

6.7t

?

2.8

•

» . .

4.0

i.O

2.4

6.7

CO

2.4

1.7

1.7

4.1t

?

2.7

(36)

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 165

TABLE IV

Time in minutes required to increase I'olitine of eggs 50 per cent. Molar solutions.

hol

G

e Glycol
cetate

ide

de

<L>
&

">.

J3

-t->

<u

s

<q
W

Ethylen

§

u
>>

O

Menace

Diacetit

C 2
^§

^S

w

H

B
g

tt
o
b

Acetami

^
§•5

§

1

7/8/29 . ,

2.2

1.5

8/5/29

3.8

2.8

slow

8.8

slow

8/6/29

4.0

3.0

8.6

6.3

*

slow

8/7/29

11.5 *

4.3

3.2 *

6.7

5.2

2.3 *

8/13/29..

7.0?

3.5

6.7

The eggs were first shrunk in a similar solution of glycerol in all experiments
except :

Table I Table II Table IV

7/12/29
7/10/30
7/13/30
7/20/30
7/21/30
7/24/30
8/ 9/30

7/12/29

7/8/29
8/5/29
8/6/29
8/7/29

III

The results that have been presented are, on the whole, in good
agreement with those of Overtoil (1895, 1907), Barlund (1929), and
others, on plant material. Certain substances, such as sucrose, dextrose,
erythritol and glycocoll, fail to enter either plant cells or Arbacia eggs
to an appreciable extent. Others penetrate at a measurable, but slow,
rate. Glycerol and urea belong to this group. A third group of sub-
stances, including ethylene glycol, acetamide, and perhaps formamide
and propionamide, penetrate both types of cells readily but at a rate
somewhat slower than that which is characteristic of the remaining
compounds. With the plant cells ethylene glycol enters even more
rapidly than propionamide, but with Arbacia eggs it is the slowest of
the four substances. A possible explanation of this difference will be
discussed later.

The group of compounds which penetrate the Arbacia egg most
rapidly includes butyramide, methyl and ethyl alcohols, monacetin and
diacetin. and a number of other substances whose ready penetration

166 DOROTHY R. STEWART

may be due in part to toxic effects upon the cells. Of the substances
in this group, monacetin appears to behave somewhat differently with
plant cells, its rate- of penetration into the cells of Rhoco, according to
Barlund, being of the same order of magnitude as that of glycerol,
though distinctly more rapid.

The results with Arbacia eggs are also in general agreement with
those of Mond and Hoffmann (1928) for ox corpuscles, the chief dif-
ference in the two cases being in the very ready permeability of the
erythrocytes to urea, which has also been noted by other workers, as
compared with the slight permeability of the eggs.

As to the factors governing the permeability of the Arbacia egg,
there have been found in this work no exceptions to Overtoil's principle
that high lipoid solubility of a compound is associated with rapid pene-
tration. As examples of the applicability of this principle the following
series of closely related compounds, in which lipoid solubility and rate
of penetration run parallel, may be mentioned:

glycerol < monacetin < diacetin.
acetamide < propionamide < butyramide.
methyl alcohol < ethyl alcohol.

It is probable, however, that factors other than lipoid solubility are
also significant. Of such factors, the size of the molecule, especially
of substances of low lipoid solubility, is thought by a number of workers
to be of much importance. Barlund (1929) and Mond and Hoffmann
(1928) have even gone so far as to estimate the molecular size below
which penetration is possible regardless of lipoid solubility. Expressed
in terms of molecular refraction, which they accept as an approximate
measure of molecular size, the critical values given are 15 for Rlioeo
cells and 25 for the ox erythrocyte, respectively.

The substances of low lipoid solubility whose behavior with the
Arbacia egg has been tested, arranged in the order of their molecular
weights (the order of their calculated molecular refractions, according
to Barlund, is almost the same), are: sucrose, dextrose, erythritol, gly-
cerol, glycocoll, ethylene glycol, urea, acetamide and formamide. Of
these, the Arbacia egg is impermeable, or practically so, to all down to
and including erythritol, slightly permeable to glycerol. impermeable to
glycocoll. fairly permeable to ethylene glycol, much less so to urea and
most permeable of all to acetamide and formamide.

While then- is in this series a general parallelism between molecular
weight (and molecular refraction) and penetrating power, two com-
pounds in particular, glycerol and ethylene glycol, stand out of their
proper position^. As a matter of fact, these compounds, though not

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES U>7

freely lipoid-soluble, do possess such solubility in a higher degree than
their immediate neighbors in the series and it is not improbable that
their behavior may be due to this fact. The more rapid penetration
frequently obtained with formamide, the substance of lowest molecular
weight and molecular refraction in the list, as compared with that of
the somewhat more lipoid-soluble acetamide, might be used as evidence
of the importance of molecular size, were it not for the toxicity of the
former substance and the difficulties mentioned above, connected with
its use.

It has already been mentioned that ethylene glycol enters Arbacia
eggs at a rate which is relatively slower than that at which it penetrates
cells of Rhoco discolor. This fact is of interest chiefly because the
molecular refraction of ethylene glycol, i.e. 14.4, is very close to the
critical refraction of 15 which Barlimcl believes marks the upper limit
in the size of molecules which can penetrate the plant cells by way of
pores. Since ethylene glycol enters cells of Rhoco more rapidly than
does propionamide, the smaller size of its molecule may perhaps more
than compensate for the greater lipoid solubility of the latter substance.
But with Arbacia eggs the much slower penetration of ethylene glycol
than of propionamide might perhaps suggest that its entrance by way
of pores is distinctly limited and that if such pores are present in this
type of cell their diameter averages somewhat less than that for the
cells of Rhoco. In general, it appears that for Arbacia eggs lipoid
solubility is definitely an important factor in determining the rate at
which any substance enters the cells. Molecular size also probably
plays a part, though the evidence which has been obtained on this point
is not considered to be conclusive.

IV

It was shown by Lillie (1916) that fertilized eggs of Arbacia swell
more rapidly in hypotonic and shrink more rapidly in hypertonic sea
water, than unfertilized eggs from the same female. These results
were interpreted as indicating an increase in permeability to water fol-
lowing fertilization. Since no attempt appears to have been made to
determine whether dissolved substances behave similarly, a number of
experiments of the sort previously described were carried out with
ethylene glycol.

In order that the results of such experiments may be strictly com-
parable, the various groups of cells must all approach the same average
equilibrium volume. As mentioned above, it is not easy to select
quickly, and without previous measurement, different groups which will
do this exactly. It is necessary, therefore, to repeat the experiments

168

DOROTHY R. STEWART

until a suitable comparison can be made. As a consequence of this
unavoidable difficulty only three sets of experiments, out of a much
larger number actually performed, gave results which were considered
to be entirely satisfactory. These experiments form the basis for Table
VI and Fig! 5.

In Fig. 5 are represented the swelling curves obtained in one such
experiment with fertilized and unfertilized eggs from the same indi-
vidual, in 60 per cent sea water and in a molar solution of ethylene
glycol in 60 per cent sea water. In each case there had been a prelim-
inary shrinking of the eggs for ten minutes in a glycerol solution made
up to the same concentration as the ethylene glycol, in order that no
shrinkage should occur during the experiment proper. Within such a
short period glycerol behaves practically as a non-penetrating substance.

It will be observed that in both solutions swelling is decidedly more
rapid with fertilized than with unfertilized eggs. In the case of the
hypotonic sea water, in accordance with Lillie's view, this may be
interpreted as indicating a greater permeability of the fertilized eggs
to water. But the fact that swelling in the ethylene glycol solution is
faster with fertilized than with unfertilized eggs is not in itself a proof
that the solute enters more rapidly in the former case, since the ob-
served rate of swelling depends upon the rate of penetration of water
as \vrll as upon that of the solute.

A mathematical attempt to separate the effects of the respective
rates of penetration of the solute and solvent in cases such as this has
recently been made by M. H. Jacobs and will be published elsewhere.
In the meantime, the data already obtained may be compared roughly
by referring to Table V, in which are shown the times required for

TABLE V

A comparison of the times, in minutes, required by unfertilized and fertilized
'"'/</.«• to swell to a volume halfway bctzveen their initial and equilibrium •volumes
~c!tcn placed in 60 per cent sea water and in ethylene glycol made »/> to molar con-
centration in 60 per cent sea ^vater.

Unfertilized

Ferti

Unfertilized

Eggs

KKC^

Fertilized

7 10 !0 .

Water

5.2

2.7

1 '»

Kt hyU-ne glycol

23.6

11.0

2.1

7/13/30

Water

3 7

i s

1 }

1 i li\ Icne glycol

27.3

17.7

1.5

7/16/30. . . .

Water

4 5

3 4

1 ?

1 .1 liylene glycol

13.3

9.0

1.5

PERMEABILITY OF ARBACIA EGG TO NON-ELECTROLYTES 1 <*>

fertilized and unfertilized eggs to reach a volume halfway between their
initial and equilibrium volumes, where swelling involved (a) the en-
trance of water alone, and (b) the entrance of both water and solute.
The ratios for the times in question for unfertilized and fertilized eggs
are given in column 5. It will be observed that the ratio for the ethylene
glycol solution is in each case greater than that for water alone. In
other words, there is, with ethylene glycol present, a proportionately
greater decrease in the time required for a given volume change after
fertilization than is found in its absence. This evidence, as far as
it goes, would seem to indicate an increased permeability to ethylene
glycol itself, though the differences in the observed rates are not very
large.

Attempts to carry out similar experiments with other substances
were not entirely successful. It was fairly definitely determined, how-
ever, that the increase in permeability which presumably occurs upon
fertilization was not sufficiently great to permit a measurable penetration
into fertilized eggs of substances such as glycocoll, erythritol and dex-
trose to which the unfertilized egg appears to be almost or entirely
impermeable.

It is a pleasure to acknowledge my indebtedness to Dr. M. H. Jacobs,
at whose suggestion this work was begun and under whose direction
it has been carried on; and to Dr. Balduin Lucke for his kindness in
demonstrating his methods of handling and measuring the eggs.

SUMMARY

1. The permeability of the Arbacia egg has been tested by following
the volume changes which occur in solutions of nineteen non-electrolytes.

2. The general behavior of this type of cell agrees fairly closely
with that of the plant cells studied by Overtoil and by Barlund and,
to a somewhat lesser extent, with that of the mammalian erythrocyte.

3. No exception has been found to the principle that compounds
which are freely lipoid-soluble readily penetrate the Arbacia egg. In
the case of substances which are only slightly lipoid-soluble, the size
of the molecule appears to be of importance, but unequivocal evidence
on this point is difficult to obtain because of the existence of complicating
factors of various sorts.

4. Some evidence has been obtained that following fertilization there
is an increase in permeability to ethylene glycol.

170 DOROTHY R. STEWART

BIBLIOGRAPHY

BARI.VXD. H.. 1929. Ada Dot. Fcnnica. 5: 1.

COLLAXDER, K.. AND !'>.\Ki.rxD, H., 1926. Soc. Scicnt. Fcnnica. Continental. BioL,
2: 1.

FLEISCHMAXX. W.. 1928. P finger's Arch., 220: 448.

GRYXS. G.. 1896. P finger's Arch.. 63: 86.

HAMnn«,i K. 11. I.. 1902. Osmotischer Druck und lonenlehre. \Yiesbaden. Vol.

I. pp. 161-400.
m, S. G., 1897. Pfliigcr's Arch.. 68: 229.

LILLIE, K. S.. 1916. Am. Jour. Pltysiol.. 40: 249.

Mi i'i i' HKON, M., AND LUCRE, B.. 1926. Jour. Gen. Physiol., 9: 697.

MONO, R., AXD HOFFMANN, F., 1928. Pfliigcr's Arch.. 219: 467.

I'MKTOX. I-!., 1895. Vicrtcljahrschr. d. Naturforsch. Gcs. in Zurich. 40: 159.

i IVI.KTOX, E., 1907. Nagcl's Handbuch dcr Physiologic des Menschcn. Braunsch-
weig. Vol. II, pp. 744-898.

THE PERMEABILITY OF THE ARBACIA EGG TO AM-
MONIUM SALTS

DOROTHY R. STEWART

(From I lie Department of Physiology, University of Pennsylvania, and the Marine
Biological Laboratory, Woods Hole, Massachusetts)

Salts, as a rule, appear to penetrate living cells with considerable
difficulty. A striking exception to this behavior is found in the case of
many ammonium salts which are known to enter mammalian erythro-
cytes with the greatest ease. (Gryns, 1896; Hedin, 1897). Jacobs
(1927) has offered an explanation of this unusual property of the am-
monium salts which is based on the behavior of the products of their
hydrolysis in aqueous solutions. In the case of NH4C1, for example,
the products of hydrolysis are largely undissociated ammonia, which
enters all cells with the greatest ease, and highly dissociated hydrochloric
acid, one of whose ions, Cl', is capable of being exchanged through the
anion-permeable wall of the erythrocyte for OH'. The final result is a
gradual accumulation within the cell of the salt in question, leading
eventually to hemolysis. Other cells, which lack the pronounced specific
permeability to anions that characterizes the erythrocyte, would not be ex-
pected to accumulate ammonium chloride, but only small quantities of
free ammonia. In accordance with this view Jacobs finds that Arbacia
eggs do not swell in isotonic solutions of ammonium chloride, the amount
of ammonia which must penetrate in order to establish equilibrium with
the external solution being too small to cause an appreciable volume
change.

With the ammonium salts of the lower fatty acids, however, the
situation is somewhat different. These salts form, upon hydrolysis,
ammonia and free fatty acid. The latter is largely undissociated and
is able to penetrate cells easily. Once inside them it unites with the
ammonia, which has entered independently, forming salt again. Jacobs
(1927) showed that the behavior of erythrocytes is in accordance with
this theory, and he predicted that, since apparently all cells are perme-
able to both ammonia and the lower fatty acids, the ammonium salts
of these acids should be able to penetrate any type of cell in this fashion.
The present paper furnishes evidence that Arbacia eggs are freely

171

172

DOROTHY R. STEWART

FK,S. 1-7: Swelling curves for Arbacia eggs in half molar solutions of various
ammonium salts. Abscissa?, times in minutes; orclinatcs, diameters in arbitrary
unit «|ii:il to 1 .064^. (See Footnote 1 concerning the pH values used.)

1. Ammonium formate (F), acetate (A), propionatc (P), butyrate (5)

and i /' > ; ]>II 7.0.

_'. Same; pH 7.8.

.v Sani' ; ],| 1 6.5.

Ammonium acetate at pH indicated beside each curve.

5. Ammonium propionate at pH indicated beside each curve.

6. Ammonium butyrate at pH indicated beside each curve.

7. Ammonium valerate .a pH indicated beside each curve.

PERMEABILITY OF ARBACIA EGG TO AMMONIUM SALTS 173

permeable to these ammonium salts, and that the mechanism of the
penetration is that predicted.

8

FIG. 8. The swelling of Arbacia eggs in M/2 solutions of ammonium acetate
of the pH values indicated beside each curve. Abscissa?, times in minutes ; orcli-
nates, diameters in units each equal to 1.064/". (See Footnote 1 concerning the pH
values used.)

The solutions of ammonium salts were made up by adding to half
normal NH4OH sufficient concentrated acid to give the pH desired.
The volume changes produced in this way were slight and the resulting
solutions could be considered, with little error, to have a salt concen-
tration of M/2, or approximately that of sea water. In some cases,
where hypertonic solutions were desired, the salts were made up in
bicarbonate-free sea water instead of distilled water. To obtain this,
2.4 cc. of normal HC1 were added to a liter of sea water, and the CO2
was driven off by aerating for twelve to fifteen hours. The pH was
brought up to neutrality by the addition of a little NaOH and then the
ammonium salt was prepared as before. The half molar solutions made
up in distilled water were isosmotic with the cells, and penetration of
salt from them could be detected by the swelling of the eggs, which
proceeded until they were destroyed. The solutions in bicarbonate-free
sea water, on the other hand, were initially hypertonic so that the eggs
shrank, regaining their original size gradually only if the salt penetrated.
The volume changes were studied by the method described in the pre-
ceding paper (Stewart, 1931).

II

The experiments were divided into two main series, one in which
the ammonium salts were made up in pure M/2 solutions and another
in which they were dissolved in sea water or half molar KC1. (NaCl
could not be used because of its toxicity in unbalanced solutions.) The
advantage of experiments of this second type lies largely in the fact
that in them the eggs must approach, and then maintain, an equilibrium
volume. Failure to do so gives definite evidence of injury to the cells.

174 DOROTHY R. STEWART

This method was used as a check on the results obtained by the first
one in which conditions were, on the whole, considerably more ab-
normal and for which no acceptable criterion of injury was available.
The results obtained by the two methods were in good agreement.

For must of the experimental work the ammonium salts of the
first five saturated fatty acids were used. Of the acids formed by the
hydrolysis of these salts, four, namely acetic, propionic. butyric and
valeric, are of very nearly the same strength. Furthermore, these four
acid> have dissociation constants that are almost the same as that of
ammonia. Consequently at pi I 7.0 there should be approximately equal
amounts of free acid and ammonia present in the solutions. Formic
acid is enough stronger so that the pH of the salt solution must be
lowered to about 6.5 before equal quantities of acid and ammonia are
obtained. For this reason the formate is not strictly comparable, at
the same pH, with the other salts.

A comparison of the swelling curves of Arbacia eggs placed in
M/2 solutions of the five salts at pi [ 7.0 (Fig. 1) shows the marked
differences in the rates at which the salts enter the eggs. Since am-
monia is known to penetrate these and other cells with extreme rapidity,
it is reasonable to believe that the fatty acids, which in pure solutions
enter cells somewhat more slowly, are to a certain extent the limiting
factor which determines the rate of swelling. Because of the fact that
at a given pi I the concentration of free acid is approximately the same
in the solutions of all of the salts except the formate, the swelling curves
must give an indication of the relative order of penetration of the acids
themselves. Numerous experiments indicate that the series is formic <
acetic<propionic<butyric<valeric. which is the order of their lipoid
solubilities, and the inverse of that for their molecular weights. Fur-
thermore, it is the order frequently found in experiments with pure solu-
tions of the acids, provided the results are not confused by injury effects,
i For references to the literature see Jacobs, 1927).

The same order of penetration is observed if the salts are made up
to half molar strength in either bicarbonate-free sea water or half molar
K< '!. Figure 9 gives a set of the curves obtained in experiments with

water solutions. It will be noted that the valeratc proves decidedly
toxic, tin continuing to swell in it until they are cytolyxed. All

the other curves approach an equilibrium, however, indicating the
lack "i any derided injury to the cells. The fact that the equilibrium
volume U not exactly the same as the initial volume of the eggs may
be due to slight osmotic differences in the solutions themselves or may
perhaps indicate iln- existence of secondary factors of some sort whose
effects are of relatively minor importance.

PERMEABILITY OF ARBACIA EGG TO AMMONIUM SALTS 175

B

N / >

FIG. 9. Swelling curves for Arbacia eggs in solutions of ammonium formate
(X), acetate (A), propionate (©), butyrate (•) and valerate (D) ; pH 7.0.
Solid lines, M/2 solutions ; broken lines, M/2 solutions in bicarbonate-free sea
water. (See Footnote 1 concerning pH values used.)

Ill

If the rate of entrance of free acid into the cell is the most impor-
tant factor in influencing the rate of penetration of the salt, then it
would be expected that the pH of the solution, which determines the
relative concentrations of free acid and ammonia, would have a very
important influence upon the swelling process. This proves to be the
case. In general the increase in volume of the cells tends to become
more rapid as the pH is lowered, down to about 6.0, and to become
progressively slower as the pH is increased above 7.0. The swelling
of the eggs in a solution of ammonium acetate at pH 7.8 is almost
imperceptible (Fig. 2), while at pH 6.5 (Fig. 3) — an acidity which is
not in itself markedly injurious — they increase in size so rapidly that
cytolysis occurs within five or six minutes. A striking illustration of
this effect of pH in changing the amount of free acid, and consequently
the rate of swelling of the Arbacia eggs, is given in experiments such
as that represented in Fig. 8. Here ammonium acetate solutions were
used and the pH varied, in small steps, between 7.7 and 6.4. There is
no exception to the orderly increase in rate with decrease in pH. The
experiment has been repeated several times, always with the same re-
sults. Similar, though less extensive, observations have been made with
each of the other salts (Figs. 5, 6, 7). The only essential difference
between them lies in the spread of the three curves, which becomes less
and less as the ease of penetration of the free acid involved increases,
until finally, with the valerate, the curves are not separated from each

176 DOROTHY R. STEWART

other by an amount that is significantly greater than the error of the
method. As a result the three valerate curves do not maintain any
definite relation to each other. Frequently penetration at pll 7.8 ap-
pears to be faster than at 7.0, or the curve for pH 6.5 is a trifle less
steep than that for 7.0. But in about half of the experiments the order
was found to be the same as that for the other salts, namely pH
6.5>pH 7.0>pH 7.8.

The effect of variations in pH is thus seen to parallel the changes
in the amount of undissociated acid that is present in the solution. As
has been mentioned previously, at pll 7.0 approximately equal quantities
of free acid and ammonia are present. At a more alkaline reaction
the amount of acid would be less than that of ammonia, and vice versa.
Since the acid, rather than the ammonia, tends to be the limiting factor
in determining the rate of swelling of the eggs, it is relatively easy to
diminish this rate by a change of the pH in the alkaline direction. It
is conceivable that at a sufficiently alkaline reaction almost no swelling-
would occur. This proves to be the case with an ammonium formate
solution at pH 7.8, in which eggs often failed to swell appreciably during
a 30-minute exposure. This failure was not due to any injury to the
eggs, for after much longer exposures than that they were capable of
giving typical swelling curves in an acetate solution. Attempts were
made to obtain similar results with the other salts by raising the pH
beyond 7.8, but toxic effects became so marked that the experiments
could not be carried out.

As the pH is lowered below 7.0, more and more acid is present in
the solution and swelling becomes increasingly rapid. If the pH is
lowered sufficiently, the amount of ammonia might theoretically become
so small that it would in turn assume the role of the limiting factor.
This actually seems to be the case with the ammonium acetate. The
maximum rate of swelling for this salt was obtained at about pH 6.0,
while below that point the rate decreased until at pll 5.5 the curve ob-
tained was almost identical with that at pi I 6.5 or 6.6. Similar experi-
ments with the other salts were impracticable because of the high tox-
icity of their more acid solutions.

As would be expected from the marked effect of pi I upon the amount
of free acid present in the solution, the divergence between the rates
of swelling for the five salts is most marked in the alkaline range. Here
the acids themselves arc the chief factor in determining the rate, and
differences in their ability to enter the eggs become most prominent.
At pi I 7.o (Fig. 2) the penetration of the formate is sometimes almost
impossible to detect, while the valerate enters so rapidly that cytolysis
occurs in three minutes. At lower pH values, however, where the acids

PERMEABILITY OK ARHAC1A KG(i TO AMMONIUM SALTS 177

are present in comparative abundance and where the ammonia becomes
a more important factor, the difference would he expected to he less.
Reference to Fig. 3 shows that as a matter of fact the spread of the
curves at pH 6.5 is much reduced, the three for the propionate, butyrate
and valerate being" almost superimposed.

IV

In addition to the experiments which have been performed with the
ammonium salts of fatty acids, a considerable number have been carried
out with other ammonium salts and with some sodium and potassium
salts of fatty acids. With ammonium nitrate and ammonium chloride
no change occurred, within half an hour, in the volume of eggs placed
in AI/2 solutions, though eggs from the same lot were found to swell
appreciably in ammonium formate at pH 7.7 during that time. Indeed,
eggs may be left in ammonium chloride made up in bicarbonate- free
sea water for 80 minutes without any change in size and at the end
of this time give normal swelling curves if transferred to a neutral
acetate solution. In other words, they appear not to have been injured
appreciably by the chloride and their failure to swell is probably due
to the fact that the acid cannot enter the cell and so no salt accumulates.

An attempt was made to work with solutions of ammonium benzoate
and salicylate, but they proved very toxic to the eggs and it was not
possible to obtain reproducible results with them. About all that can
be said concerning these salts is that they both seem to enter the cells
rapidly, though this apparent speed of penetration may be entirely due
to the effects of injury which are ordinarily apparent within ten min-
utes. Ammonium lactate, on the other hand, enters very slowly, there
being no appreciable penetration by the end of an hour from M/2 solu-
tions at pH 7.7 or 7.0, though at pH 6.5 the volumes have increased
to an extent that is just barely significant. At pH 7.0 a 10 per cent
increase in volume occurs in about three hours as compared with a
corresponding time of approximately five minutes for a similar solution
of ammonium propionate. Since lactic acid differs from propionic acid
only in the substitution of an OH group for a hydrogen atom, this
behavior illustrates the marked effect upon permeability of the intro-
duction of a polar group into a molecule.

Repeated experiments with sodium or potassium acetate and with
sodium valerate at varying pH levels failed to show any penetration
of these substances — or at least any penetration sufficient to cause a
change in cell volume.

I wish to thank Dr. M. II. Jacobs, under whose direction this study
has been made, for his many suggestions and his never-failing interest
in the progress of the work.

178 DOROTHY R. STEWART

SUMMARY

1. The rate of swelling of Arbacia eggs in solutions isosmotic with
sea water of the ammonium salts of the first five saturated fatty acids
has been found to be in the order: valerate>butyrate>propionate>
acetate > formate.

_'. In solutions of these salts made up in bicarbonate-free sea water
or in M 2 KG recovery of the original volume after a preliminary
shrinkage follows the same order.

3. Within the pH range from approximately 7.8 to 6.2 the rate
of -welling increases with increasing acidity. Under certain conditions
a further lowering of the pH may reverse the effect.

4. Xo significant changes in volume occur in solutions of ammonium
nitrate or chloride, though cells exposed to such solutions for over
an hour are still capable of swelling in ammonium acetate.

5. These results agree with the theory suggested by Jacobs, that
there is no appreciable penetration of either the salt or its ions as such
but that undissociated ammonia and fatty acid, formed by hydrolysis
of tin- salt, pei letrnte the cell separately, uniting again after their en-
trance.

6. The rate of entrance' of the acid, rather than that of the ammonia,
usually appears to be the limiting factor in determining the rate of
swelling of the cell.

BIBLIOGRAPHY

r.kvxs, G., 1896. rjli'Kicr's Arch.. 63: 86.
HEDIN, S. ])., 1897. Pfliiger's Arch.. 68: 229.
JACOBS, M. H., 1927. The Harvey Lectures, 22: 146.
STEWART, D. R.. 1'Ml. Hi«I. Hull. 60: 152.

1 Since the salt errors of the indicators for the solutions used are not exactly
known, it has been thought hi-st to use the uncorrected pH values throughout this
paper, particularly since relative rather than absolute figures are of importance.
The proper corrections may be made at any time when the necessary data become
available. Since these corrections would probably be about — 0.2. the values here
.yivcn are too high by approximately 0.2 pH units. The standards used for com-
I'.irison were the ordinary M/15 phosphate buffer solutions.

NOTES ON TREMATODES !• \<( >M A LONG ISLAND DUCK
WITH DESCRIPTION OF A NEW SPECIES

H. \V. STUNKARD AND F. W. DUNIHUE
BIOLOGICAL LABORATORY, UNIVERSITY ( "OU.KI.K, NEW YORK UNIVERSITY

In view of the fragmentary knowledge concerning the trematodes
of North American birds, and the extent to which these parasites are
distinct from those of other continents, it seems proper to record certain
observations which we have made.

Examination of the alimentary tract of a single duck, sold in the
New York market, yielded four species of trematodes. belonging to four
different genera. One of the species is new to science and two others
are recorded from America for the first time. The high degree of in-
festation is unusual, since previous examinations have shown that ducks
bought in the market are rarely parasitized by trematodes. Further-
more, the conditions under which most Long Island ducks are raised
and prepared for market tend to prevent the acquisition of such para-
sites. These facts indicate that the bird was cither a wifd duck or that
it had not been closely confined in breeding and feeding pens.

To substantiate the correctness of specific determination, brief de-
scriptions and figures of the parasites are included.

Echinostoma rcrolulinn (Froelich, 1802) Looss, 1899

Three specimens from the intestine are referred to this species.
The one shown in Fig. 1 measures 13 mm. in length and 1.9 mm. in
width. Measurements of internal structures are as follows : oral sucker,
0.245 mm.; pharynx, 0.18 mm.; acetabulum, 0.635 mm.; ovary, 0.43
mm. ; testes, 0.795 mm. long by 0.43 mm. wide. The collar of spines
is not complete in any of the specimens and consequently their number
can not be determined with certainty. In other features, however, the
worms agree so well with the descriptions of E. revolution as given by
Looss (1899). Liihe (1909), and Dietz (1910) that there appears to
be no doubt concerning their identity. E. revolutum has been previously
reported from North America and the life cycles of the species was
experimentally demonstrated by Johnson (1920).

Psilochasunts o.vvnnis (Creplin, 1825) Liihe, 1910

Since the description of Creplin (1825) this species has been studied
by von Linstow (1882), Braun (1902), Lube (1909), and Oclhner

179

180

H. \V. STL'NKARD AND F. \V. DUN1HUE
PLATE I

I--C8

0V

__-(«

IMC;. 1. ILchinoslottia rcrolnliini Fie. 2. Psilochasmus o.rynnis

AMtrcviations

Is lesti^
/)// pharynx

ac acclal)iiluin

cs cirrus sac

<//> "cnital pore

i» intestine

os oral sucker

sy shell gland

seminal vesicle

lit uterus

7'</ vitelline duct

77 vitellaria

TRMMATODES FROM LONG ISLAND DUCK

(1913). The material used in these investigations came from various
species of ducks taken in Northern Europe. A second species, Psiloch-
asinus loiigicirratns, was described by Skrjabin (1913) from Fulirjiila
nyroka in Russian Turkestan. A comparison of his description with
those of other authors, especially that of Odhner, and with the specimens
at hand shows no specific differences. The only feature in which there
is not actual overlapping is the length of the cirrus sac and reports do
not agree on this point. Braun stated that the cirrus sac is long and
slender with a seminal vesicle in its base. Odhner found the cirrus sac
ending a short distance before the ovary, while in the specimen shown
in Fig. 2 the sac extends to the ovarian level. Consequently, there is
no means of distinguishing between the species, and /•'. longicirratus may
be regarded as a synonym of P. o.vyunts.

Four specimens from the duck, one of which is shown in Fig. 2,
show essential agreement with the description of P. o.vyitnis and ac-
cordingly are assigned to that species. The worms measure from 4.6
to 5.5 mm. in length and from 1 to 1.25 mm. in width. The shape of
the body is characteristic and the terminal spike projects from a depres-
sion in the ventral surface at the caudal end. The spike is composed
of strong muscular fibers, continuous with those of the body wall. The
acetabulum is situated at the anterior fourth or fifth of the body and
measures from 0.54 to 0.6 mm. in diameter.

The oral sucker is 0.3 to 0.32 mm. in diameter and followed almost
immediately by a muscular pharynx about one-half its size. The esoph-
agus is lined with epithelium and bifurcates at the level of the genital
pore to form the intestinal ceca. The ceca terminate blindly just an-
terior to the caudal lobes of the vitellaria.

The testes are distinctly lobecl. They measure from 0.653 to 0.884
mm. in length and 0.461 to 0.530 mm. in width. They are situated one
behind the other in the anterior part of the caudal half of the body.
A vas deferens arises from the anterior, ventral surface of each and
passes forward, the duct from the anterior testis on the left and that
from the posterior testis on the right side of the body. The left duct
passes under the ovary and both open into the caudal end of the cirrus
sac. This portion of the sac is expanded and filled with spermatozoa.
From the seminal vesicle the ejaculatory duct makes a coil and then
passes in a straight course to the genital pore.

The ovary measures from 0.27 to 0.32 mm. in 'diameter. It is
slightly left of the median plane, about two-fifths of the distance from
the cephalic testis to the acetabulum. The oviduct arises at the caudal
end and passes backward to the ootype, which is enclosed in a large
" shell gland." From the ootype Laurer's canal leads to the dorsal

12

182 H. \V. STUNKARD AND F. W. DUNIHUE

surface. The vitellaria are lateral, dorsal and ventral, and extend from
the level of the acetabulum almost to the posterior end of the hody.
The lobes of the two sides are connected by a ventral commissure behind
the testes. and just in front of the cephalic testis ducts arise from the
ventral sides and pass dorsally and medially, meeting on the dorsal
side to form a common vitelline duct which passes forward to open
into the ootype. There is no seminal receptacle, but the initial portion
of the uterus is filled with masses of spermatozoa. The uterus passes
in a sinous course, backward to the level of the cephalic testis and then
forward to the genital pore, situated below the bifurcation of the ali-
mentary tract. Eggs in the uterus measure from 0.1 to 0.12 mm. in
length and from 0.07 to 0.077 mm. in width..

Xotofotylns </ihhns (Mehlis in Creplin, 1846) Kossack, 1911

This species, unknown from its discovery by Mehlis until the original
specimens were restudied by Kossack (1911), has not been reported
since that time so far as we can determine. The European specimens
were found in the intestine of Fiilica (tiro and (iallinnla chloropus.
Other species are known from Europe and two species, N. urbancnsis
and N. quinqueserialis have been described from North America. The
specimens from the Long Island duck agree most closely with the de-
scription of .Y. yibbits and apparently belong to that species.

Fifteen specimens (Fig. 3) were recovered from the washings of
the alimentary tract but their original location in it is uncertain. Since
members of the Notocotylidas ordinarily infest the ceca and terminal
part of the intestine, it is probable that these worms also were in that
region. They are much flattened and sexually mature specimens meas-
ure from 1.1 to 1.5 mm. in length by 0.5 to 0.6 mm. in width. The
ventral surface bears three rows of pits, 12-14 in each. Those near
the ends of the body are smaller than those near the middle.

The oral sucker is usually broader than long and measures from 0.07
to 0.11 mm. in diameter. The esophagus is short and bent, opening
almost immediately into the ceca. These structures end blindly between
the testes.

The testes are deeply lobed, 0.2 to 0.23 mm. in length and 0.12 to
0.14 mm. in width. Ducts from the testes unite in front of the ootype
to tonn a common duct which passes forward on the dorsal side of
tin- body. It becomes coiled as it approaches the middle of the body
and increases in sj/r to form a seminal vesicle. Immediately in front
oi" tin- uterine loops it enters the cirrus sac, which measures from 0.43
to 0.5 mm. in length.

The ovary li<-^ between and in front of the testes. It is deeply

TREMATODES FROM LONG ISLAND DUCK

183

lobcd and measures from 0.13 to 0.15 mm. in diameter. The ootype
complex lies anterior to the ovary and is of the usual notocotylid type.
Eggs measure from 0.02 to 0.024 mm. in length and 0.01 to 0.012 mm.
in width. They bear long filaments at both poles.

The points of difference between the description of Kossack and
the present specimens appear to be only variations that might be ex-
pected in a single species. His statement that the testes are some dis-
tance from the posterior end of the body and that the intestinal ceca
are relatively close together may be explained by difference in sexual
maturity. In specimens containing fewer eggs than the one shown in
Fig. 3, the testes are relatively farther forward and the ceca closer

PLATE II

3 4

Abbreviations as in Figs. 1 and 2

FIG. 3. Notocotylus fjibbus

FIG. 4. Parainonosloinitin parvum

together. In fact, in sexually immature, non-gravid worms, the testes
are proportionally farther forward than in the specimen shown in Kos-
sack's figure. It appears that when the uterus becomes filled with
eggs the other structures are somewhat displaced. Kossack reported
that the ventral glands are very weakly developed and that they number
6-8 in each row. Since their location is not shown in his figure, one
can not tell whether they were present in a section of the body or dis-
tributed uniformly throughout its length. The material available to
Kossack was probably in poor condition and the observations were ap-

184 H. W. STUNKARD AND F. W. DUNIHUE

patently not checked by study of serial sections. It seems very prob-
able, therefore, that some of the ventral glands were overlooked.

Paramonostomwn parvum n. sp.

The genus Paramonostomuin was erected by Luhe (1909) to contain
a single species, Monostomum alvcatum (Mehlis in Creplin, 1846), that
occurs iu several European birds. Earlier accounts of the species are

•••wed in the paper by Kossack (1911).

Barker (1916) questioned the validity of the genus on the ground
that it differs from Notocotylus only in the absence of ventral glands.
Careful examination shows, however, that the absence of ventral glands
is only one of a number of fundamental differences which distinguish
the genus. In shape of body, location and proportion of internal or-
gans, extent of uterus, and form of the copulatory organs, there are
marked differences between Notocotylus and Paramonostomum.

Harrah (1922) described a second species, Paramonostomum
echinus, from the intestine of the American muskrat, Fiber sibethicus.
Review of his description and figure shows that, in regard to the features
listed above, his specimens do not agree with P. alreatmn, type of the
genus. In these respects they are similar to Notocotylits. It appears,
therefore, that P. cchinum should be removed from the genus Para-
inonostomum. It may be regarded as a type of a new genus or the lim-
its of Notocotylus may be extended to contain it. The second of these
alternatives would probably invalidate the genus Catatropis.

Twenty specimens fmm the intestine of the Long Island duck agree
very closely, except for size, with /'. alvcatum. They may prove to be
only a variety of that species but the difference is too great, in the
absence of intermediate sizes, to a>-jon them to it, and they are regarded
a> a new species for which the name Paramonostouium pari'iim is
proposed.

Sexually mature specimens are broadly ovate, almost as broad as
long, concave ventrally, pointed anteriorly, and rounded posteriorly.
'Nicy nieaMire from 0.25 to 0.5 mm. in length and from 0.2 to 0.35
mm. in width. There are no papillae on the ventral surface, and the
pn if -pines on the cuticula is doubtful.

'I he oral sucker measures from 0.035 to 0.046 mm. in diameter.
1 here is no pharynx, the esophagus is short, and the ceca pass backward
in a wide arc. close to the lateral edges of the body. Near the posterior
end ot the body the ceca turn medially and then pass backward between
the testes and the ovary, terminating about the level of the caudal margin
of the ovary.

TREMATODES FROM LONG ISLAND DUCK 185

The testes arc oval, lobed organs, situated in the extracecal areas
at the posterior end of the body. They measure from 0.08 to 0.095
mm. in length and from 0.06 to 0.07 mm. in width. The vasa deferentia
unite in front of the ootype to form a common sperm duct that passes
anteriad. At the caudal end of the anterior third of the body the sperm
duct opens into a large, much coiled seminal vesicle. The cirrus sac is
curved, 0.035 to 0.06 mm. in length and 0.025 to 0.04 mm. in width.
The caudal part of the cirrus sac contains a small section of the seminal
vesicle.

The ovary is situated in the median plane, near the posterior end
of the body. It is a much lobed structure, 0.05 to 0.07 mm. in diameter.
The " shell gland " is anterior to the ovary and about one-half its size.
The female genital complex is similar to that of Notocotylus. The
uterus extends in transverse folds to the caudal end of the anterior
third of the body and other coils extend farther forward on either side.
The metraterm is short and opens at the genital pore, situated in the
median plane immediately behind the bifurcation of the alimentary
tract. The vitelline glands consist of small follicles which occupy the
extracecal areas from the level of the testes to that of the seminal
vesicle. Eggs measure from 0.021 to 0.024 mm. in length and from
0.011 to 0.013 mm. in width. They are provided with long polar
filaments.

DISCUSSION

The discovery of these trematodes in a North American duck affords
supplementary data concerning two little known species, previously
found only in Europe, and adds a second species to the genus Para-
monostoiiiuin, which supports the validity of that genus. It supple-
ments other data which tend to indicate that the parasitic fauna of
Europe are being introduced into North America and that exotic species
are being established here. Migratory animals, especially birds, may
acquire infections in one location and during their journeys distribute
infective stages of the parasites over wide areas. The presence of suit-
able intermediate hosts, and of conditions which permit their infestation,
result in the spread of parasites to new locations. In many instances
the parasites are able to utilize new intermediate and definitive hosts
and thus extend both their geographical range and list of host species.
It appears that there is more overlapping of the parasitic faunas of the
Eastern and Western hemispheres than has been realized.

186 II. \V. STUXKARD AND F. W. DUNIHUE

BIBLIOGRAPHY

BARKER. F. D., 1916. A New Monostome Trematode Parasitic in the Muskrat

with ;i Kt-v t«> the Parasites of the American Muskrat. Trans. Am. Micr.

Soc.. 35: "175.

BRAL-N, M.. 1902. Fascioliden der Vogel. Zoo/. Jalirb. Syst.. 16: 1.
CREPLIX. F. C. 1S.25. Observationes de entozois. Gryphiswaldise.
I i 1 . !"10. Die Echinostomiden der Vogel. Zool. Jahrb. Supfl, 12: 265.

RAH. K. C., 1922. North American Monostomes, primarily from Fresh-Water

Musts. ///. Biol. Monofir.. No. 7.
l"iiN?DN. T. C., 1920. The Life Cycle of Echinostoma revolutum (Froelich).

I'niv. of Calif. Pub. Zool., 19: 335.

Ki'SSAfK, W., 1911. liber Monostomidem. Zool. Jahrb. Syst., 31: 491.
LINSTOW, O. F. vox, 1882. Helminthologische Studien. Arch. f. Naturg., 48:

(Band I) 1.
Looss, A., 1899. Weitere Beitrage zur Kcnntniss der Trematoden-Fauna Aegyptens,

zugleich Versuch einer natiirlichen Gliederung des Genus Distomum Ret-

zius. Zool. Jahrb. Syst., 12: 521.
LriiE, M., 1909. Parasitische Plattwiirmcr. I. Trematodes. Die Siisswasser-

fauna Deutschlands, vol. 17.
ODHNER, T.. 1913. Zum natiirlichen System der digenen Trematoden VI. Zool.

Anc., Band 42: 289.
SKKJABIX, K. I... I'/l.i. V'ogeltrematoden aus Russisch Turkestan. Zool. Jahrb.

Syst., 35: 351.

THE PRODUCTION OF NORMAL EMBRYOS BY ARTIFICIAL
PARTHENOGENESIS IN THE ECHIUROID, URECIIIS

ALBERT TYLER

(From the William G. Kerckhoff Laboratories of the Biological Sciences, California

Institute of Technology, Pasadena)

INTRODUCTION

The eggs of annelids and mollusks have been the subject of
numerous parthenogenesis experiments, but only a few instances have
been reported in which normal development was said to have been
obtained. Lefevre (1907) described and figured parthenogenetically
produced embryos of Thalassema mellita which resembled the normal,
but he states (p. 116), "I have never observed a single instance where
the trochophores rose to the surface of the water." Just (1915)
reports the production of "normal-looking swimming forms — trocho-
phores scarcely to be distinguished from the normal" —from eggs of
Nereis exposed to high temperature. Morris (1917) obtained embryos
which she described as "fairly normal swimming larvae" from eggs of
Cumingia exposed to high temperature and hypertonic sea water.
In other cases the embryos obtained were described merely as swimmers
or even uncleaved swimmers, (Allyn, 1912; Loeb, 1908, 1912; Lillie,
1910; Kostanecki, 1908; Bullot, 1904; Scott, 1906; Treadwell, 1902;
Fischer, 1903). It is evident from the above citations that normal
embryos have rarely, if at all, been obtained. The question may then
be raised as to whether the sperm performs some special function in
the eggs of annelids and mollusks, such that in its absence normal
development is less likely to be obtained.

In the eggs of some annelids and mollusks there is evidence that
the entrance point of the sperm is instrumental in determining the
plane of bilateral symmetry of the embryo (Just, 1912; Morgan and
Tyler, 1930). Such an orienting point is lacking, of course, in eggs
treated with the usual parthenogenetic agents. It is important to
know, therefore, whether eggs, in which the entrance point of the
sperm is a factor in the determination of bilateral symmetry, are
capable of normal development after artificial activation.

For the eggs of Urechis it was found that the first cleavage plane
coincided with the entrance point of the sperm in the great majority
of cases. Since in spirally cleaving eggs the first cleavage plane bears

187

188 ALBERT TYLER

a definite relation to the plane of bilateral symmetry of the embryo,
it may be safe to conclude that the sperm entrance point determines
the oriental ion of the embryo in Urechis. The eggs of Urechis can be
activated by means of certain artificial agents; such as, hypotonic sea
water, hypertonic sea water and high temperature. Top-swimming
embryo- indistinguishable from those produced by normally fertilized
eggs were obtained from eggs treated with hypotonic sea water. How-
ever, the normal embryos \vere produced in very small numbers, the
\arious types of abnormal swimmers being far more abundant.

The eggs which are to cleave and develop can be distinguished and
i-olated from the others at a few minutes after treatment. The
manner in which these eggs respond to the treatment presents a
striking contradiction to the view (Just, 1922) that normal develop-
ment results only when the initial response of the egg to the artificial
agent is similar to its response to the sperm.

MATERIAL AND METHODS

Urechis caupo was found at Monterey Bay and described by Fisher
and MacGinitie (1928). Professor MacGinitie later found the same
species at Newport Bay and I am indebted to him for the information
on the location of the animals and the method of obtaining the sexual
products from them. This remarkable echiuroid is ripe practically
the year round, and one pair of animals may supply eggs and sperm
for a number of experiments. The ability to obtain eggs repeatedly
from the same individual would seem to cut down the variability of
different experiments. However, it was found that the eggs of a
segmental tube, from which samples had previously been removed,
decrease in size after a couple of weeks and their response to fertilization
i- not quite 100 per cent, as is the case when eggs are removed from an
untouched tube.

The agents used to activate the eggs were diluted sea water,
distilled water, concentrated sea water and high temperature, but
since tin- first was mostly used in these experiments, it alone will be
described. The dilutions of sea water were made up of distilled water
.UK! MM water taken at a definite height of tide. The eggs were
washed twice and allowed to settle in a centrifuge tube. The super-
natant sea \\ater was then removed and the eggs together with O.S to
0.5 cc. of sea \\aier transferred to 50 cc. of the diluted sea water.
Eggs were then removed after various intervals of time to 10 cc. of sea
water by means of a -mall bore pipette filled to 0.2 cc. Any dilution
of sea water up to so per cent was found to be capable of activating
the eggs. Controls of unfertilized eggs kept in covered dishes gave no

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 189

activation and controls of eggs inseminated at the time of the experi-
ment gave 95 to 100 per cent fertilization and development.1

NORMAL FERTILIZATION AND DEVELOPMENT

The development of the fertilized egg of Urechis is so similar to
that of Thalassema (Torrey, 1902, 1903) that only a brief description
need be given here for comparison with the parthenogenetic eggs.
The "fertilization reaction" differs somewhat from that described for
other marine ova, and since use was made of this reaction in the
entrance point observations, it will be described below.

The unfertilized egg of Urechis (Fig. 1) has a rather unique shape.
It has a large indentation which varies somewhat in size in different
eggs. This shape is similar to that which a spherical rubber ball would
assume if it were evacuated and filled with a volume of water about
one-fourth less than ^irRs. The ratio of surface to volume is thus
much greater than for a spherical egg. Eggs with two indentations
(Fig. 2) may also be found, but if such eggs are made to round out in
dilute sea water and returned to normal sea water only one indentation
reappears.

The unfertilized egg contains a large germinal vesicle and a single
nucleolus. The germinal vesicle is nearest the surface of the egg at
the innermost point of the indentation. This point, as will be shown
below, marks the pole of the egg, and the egg may be seen to be
radially symmetrical about this axis (Fig. 3).

The sperm enters the egg at any point on the surface with respect
to the indentation. Within three minutes after the attachment of
the sperm, a clear conical process appears on the egg immediately
below the point of attachment.2 The cone enlarges within the next
two minutes and the membrane immediately above it becomes thin
(Fig. 4). At this time the indentation begins to round out and the
fertilization membrane begins to separate from the surface of the egg.
The sperm head then enters the cone (Fig. 5) and in about 90 seconds
passes into the egg (Fig. 6), the tail remaining behind. During the
next five minutes the cone continues to enlarge and becomes cylindrical
rather than cone-shaped, with a rounded outer end (Fig. 7). It
enlarges more rapidly than the perivitelline space, apparently stretch-
ing the membrane above it. The "cone" then begins to narrow
considerably in width so that at about thirteen minutes after the
entrance of the sperm it has the appearance of a single filament

1 The insemination was usually done in another part of the room by Betty S.
Tyler.

2 This time schedule holds for 20° C., unless otherwise stated.

190 ALBERT TYLER

(Fig. 8). It finally disappears at about seventeen to twenty-two
minutes after the penetration of the sperm. The egg thus supplies
a marker of the entrance point which can be seen as late as twenty
minutes after insemination. Entrance point observations are therefore
quite easy to make.

The first polar body is extruded 34 minutes after fertilization and
the second at 44 minutes. Cleavage begins at 74 minutes after
fertilization.3 The first two divisions are equal. At the third cleavage
the first quartet of micromeres arises dexiotropically. They are very
-lightly smaller than the macromeres. In the formation of the embryo,
no essential differences were found from Torrey's description for
Thalassema.

THE INDENTATION AND THE POLE OF THE EGG

It is generally conceded that the polarity of an egg is established
in the ovary, the polar axis being indicated by the eccentric position
of the nucleus and radial arrangement of cytoplasmic materials about
it. In the egg of Urechis the germinal vesicle is nearest the surface
at the innermost point of the indentation (Fig. 1). If the indentation

PL ATI-: I

Photomicrographs of the living objects. The magnification is 330 diameters for
all the figures with the exception of Fig. 1 , for which the magnification is 300 diameters.

FIG. 1. Unfertilized egg in "side" view, showing indentation, large germinal
vesicle and nucleolus.

FIG. 2. Unfertilized egg in "side" view showing two indentations.

I n.. 3. Unfertilized egg, antipolar view.

FIG. 4. Formation of fertilization "cone" at point of attachment of sperm,
and rounding out of indentation.

FIG. 5. Showing sperm within cone, beginning of dissolution of germinal
vesicle and elevation of membrane.

IK;. 6. Showing sperm head half-way in egg, enlargement of "cone," and
further elevation of membrane.

I n.. 7. Showing persistence and enlargement of the "cone" five minutes after
entrance of sperm head into egg.

FIG. 8. Showing the shrinkage of the "cone" into a filament-like process,
thirteen minutes after entrance of sperm, and disappearance of germinal vesicle
except for nucleolus.

I i'.. '*. Showing eccentric position of germinal vesicle in an egg in which the
indentation was made to disappear by allowing it to swell in 50 per cent sea water.

I i'.. in. Shouing migration of remains of germinal vesicle to pole as marked
by indentation; entrance point of sperm shown by filament-like cone near antipole
of egt; iifteen minutes after entrance of sperm.

FIG. 11. Photographed in t lie saim- position as in Fig. 10; shows formation of
first polar body .it center "t formerly indented area 32 minutes after fertilization.

FIG. 12. Photographed in the same position as in Fig. 11; shows formation of
second polar body next to lirM ; I I minutes after fertilization.

8 At room temperature, which is generally about 23.5° C., the time for first
cleavage is 65 minutes.

NORMAL EMBRYOS I?V AKTI I-U'IAI. 1'ARTl 1 K.\< >< IKNHSIS 191

PLATE I.

10

192 ALBERT TYLER

is caused to round out by placing the egg in 50 per cent sea water,
the germinal vesicle is seen to occupy an eccentric position in the egg
(Fig. 9). The rounding out can be followed under the microscope,
and the results of such observations show invariably that the eccen-
tric-it y is the same as that of the original indented egg. This would
indicate that the indentation marks the pole of the egg.

In order to obtain a more definite check on this relation, the point
of extrusion of the polar bodies was determined on eggs in which the
location of the indentation had been previously noted. The method
used was identical with that described below for the entrance-point
observations. The eggs selected for observations were those in which
the indentation was on the side and thus distinctly visible. They
were followed almost continuously throughout the time of rounding
up and membrane elevation. Errors due to rotation of the egg were
also partly guarded against by the presence of extra sperm which
served as markers on the surface of the membrane. As the egg
rounds up, the remains of the germinal vesicle are seen to migrate
towards the surface of the egg in the region of the former indentation
(Fig. 10). The first polar body later appears at about the center of
that area on the surface (Fig. 11), and the second polar body is later
extruded next to it (Fig. 12). The results of the observations given
in Table I show four exceptions. These presumably represent cases
in which the egg rotated within the membrane. However, there does
not seem to be any general tendency for such rotation, since the polar
bodies may appear in any position on the surface of the egg and remain

in that position.

TAHLE I

Relation betict-cii Indentation and Point nf I-'.xtrnsion of Polar Bodies

Divergence No. of Eggs

0°- 10° 24

10°- Q0° 2

90°-180° 2

\<\ i \TION OF THE ENTRANCE POINT OF THE SPERM TO THE PLANE

OF BILATERAL SYMMETRY

'I he method used for making entrance point observations and the
precautions to IK- taken have been described by Morgan and Tyler
(1930). As has been mentioned above, the point of entrance of the
sperm is very easily located in Urechis even as late as twenty minutes
after fertilization. The first polar body appears ten minutes after
the disappearance of the "cone" and serves as a marker on the surface
of the egg. This diminishes the chances of unobserved rotation of
the egg occurring.

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 193

The eggs were placed within a square of vaseline on a slide and
the drop sealed over completely with a coverslip to prevent evapora-
tion. This transfer was generally done at about ten minutes after
fertilization. The first drawings were made five to ten minutes later,
at which time the fertilization membrane has lifted off to about its
full extent and the egg itself has become spherical. The position of
the polar bodies was then noted as they appeared. At the time of
cleavage the egg was watched almost continuously, since there is a
tendency for eggs that are elongating vertically to roll over.

A number of the observations were made by Mr. Jackson Gregory,
Jr., one of the students working at the Kerckhoff Marine Laboratory,
and are listed separately in Table II as Series 1. The results of the
observations given in this table show a coincidence of 71 per cent
between the entrance point of the sperm and the first cleavage plane.

TABLE II

Relation between Entrance Point and First Cleavage Plane

No. of Eggs

Series 1

Series 2

Total

0°-10°

61

66

127

10°-45°

13

24

37

45°_90°

8

6

14

This value is obviously very much higher than would be expected on
a pure chance basis and clearly indicates that the position of the first
cleavage plane is determined by the point of entrance of the sperm.
The exceptions may be due to the inability of completely removing
all sources of error or to the presence of abnormal eggs. The most
convincing cases are those in which the entrance point and the polar
axis lie in a horizontal plane on the slide. In order for coincidence to
be obtained in such cases the egg must elongate vertically and the
cleavage plane must cut through horizontally. Twenty-six eggs of
that type were followed and in every case the first cleavage plane
passed through the entrance point, although seven of them rotated
during the division.

In eggs with spiral cleavage the plane of bilateral symmetry is
given by the position of the 4d cell, which also marks the dorsal side
of the embryo. The position of this cell is determined by the plane of
first cleavage. In Urechis it can be located after its bilateral division

194

ALBERT TYLER

into -l/i and J/_>, and its position is such as to cause the plane of
bilateral symmetry to make an angle of about 60 degrees with the first

cleavage plane.

PARTHENOGENESIS EXPERIMENTS

Activation by Dilute Sea Water

The eggs of Urechis may be activated by dilutions of sea water
ranging from 80 per cent to distilled water. Table III gives the time
range for the different solutions over which activation may be obtained,
it is taken from a set of experiments run at temperatures between 20
and 22.5° C. The third column gives the time of exposure at which
activation begins, taken arbitrarily as 5 per cent activation. These
values are obtained graphically by connecting the two nearest points
above and below 5 per cent with a straight line. The fourth column
Liives the time of exposure necessary to obtain 100 per cent activation.
With longer exposures the percentage of activation drops again to
zero for certain concentrations of dilute sea water. Column five gives
the exposure at which the activation drops to 5 per cent, calculated
in the same manner as for column three. Solutions below 45 per cent
sea water cause cytolysis of the egg, the time for cytolysis varying
inversely with the amount of dilution.

TABLE III

Range of Exposure Time in which Activation May be Obtained

Dilution of

Sea Wat.-r

Temperature

i all 111 ite.l Tinu- for
5' , Aetivution

Time for
100'; Art.

Calculated Time
for 5% Activation

ffr cent

0

°C.
21.8

seconds
1

seconds
12

minutes

20

22.0

3

20

30

22.1

10

30

40

22.0

11

50

\-

20.1

21

60

12

50

22.5

15

60

10

55

20.1

43

100

10

60

21.5

35

105

15

65

22.3

is

ISO

16

70

21.2

(.1

180

19

75

20.8

87

2K)

9

80

22.5

71

210

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 195

It is readily seen from Table III that the time for activation de-
creases with increased dilution of the sea water. This is consonant
with an interpretation of activation based on the volume of water
taken in, since water enters more rapidly the greater the dilution of
sea water in which the egg is placed.

Two other points may be mentioned here which, together with
their interpretation, will be reported in detail in a subsequent paper.
The increase in activation occurs much more rapidly than the decrease,
giving a skew activation-time curve. The percentage of the activated
eggs which cleave varies with the time of exposure; but its variation
is such that as the percentage of activation increases the percentage
of cleavage decreases. Thus, when 100 per cent activation is obtained
no cleavage occurs, while to either side of that exposure time the per-
centage of cleavage increases.

PRODUCTION OF NORMAL EMBRYOS

Many of the cleaved eggs develop into swimming embryos, and a
few into top swimmers, some of which are indistinguishable from those
produced by normally developing fertilized eggs. Some of the experi-
ments in which normal embryos were obtained are listed in Table IV.

TABLE IV

Exposure

Top Swimming Embryos

Dilution of Sea
Water

Normal

Abnormal

per cent

minutes seconds

0

20

3

25

1

10

1

14

30

1

0

3

29

1

40

2

15

2

20

5

41

0

20

2

11

40

0

40

1

28

1

0

1

14

2

0

2

25

45

0

30

3

32

1

15

1

15

1

30

2

8

50

3

0

3

12

6

0

3

17

60

4

0

3

26

8

0

5

5

65

1

45

2

12

4

0

1

5

196

ALBERT TYLER

As is shown by the table, the normal embryos are not produced by
any particular dilution of sea water or any definite time of exposure.
The percentage of normal development obtained is very small, most
of the normal embryos listed occurring in dishes of five hundred to a
thousand cleaved eggs.

In Figs. 13-16 some normal parthenogenetically produced embryos
of various ages are presented. The embryos of fertilized eggs are not
I'muri-d, inasmuch as they are identical with the parthenogenetic ones.
The early gastrulation stages of the parthenogenetic embryos were
not obtained, since at that time the normal embryos are not readily
distinguishable from certain abnormal types to be described later.

The young trochophore of sixteen hours (Fig. 13) shows a large

13

ap

.

•:

lo

Fu;s. 13-16.

(if), apical plate; lil, anterior portion of original blastopore; en, enteron; in, in-
testine; nin, mouth: -•••. o sophagus; f>r, prototroch; si, stomach. Normal embryos
from artiln i.illy activated <^gs; drawn from total mounts of preserved specimens.

FK.. lv Sixteen-hour trochophore.

FIG. 14. Twenty-hour trochophore.

FIG. 15. T \u-in \ foin hour trochophore.

1 •!(.. id. I on • < hour trochophore.

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 197

enteric mass in which a small cavity has appeared. The prototroch
is well developed and the beginning of the cesophageal invagination is
present below the prototroch on the ventral side of the embryo.
In the twenty-hour trochophore (Fig. 14) the invagination has
progressed towards the anterior end of the embryo and the post-
trochal region has enlarged. The enteric cavity soon becomes divided
into stomach and intestine (Fig. 15), and the oesophagus reaches the
anterior end of the embryo. The whole embryo enlarges considerably
during the next 24 hours and becomes a relatively thin-walled structure
(Fig. 16). The anterior end of the oesophagus is turned to the right
side of the stomach where it acquires a ciliated opening into the latter.
Bilateral symmetry is evident in the sixteen-hour embryo. The
plane of bilateral symmetry is, as pointed out above, determined very
much earlier, and for fertilized eggs, at the time of entrance of the
sperm. The parthenogenetically activated eggs can evidently estab-
lish a plane of bilateral symmetry without the aid of a localized external
agent such as the sperm entrance point. However, this is done in
relatively very few instances, while in the great majority of cases,
even with apparently normal cleavage, abnormal embryos of various
types are obtained. Of particular interest in this connection is
the occurrence of embryos that are more or less radially symmetrical.

RADIALLY SYMMETRICAL EMBRYOS

Among the top swimming and bottom swimming embryos produced
from the artificially activated eggs, there occur a number of forms
which may be grouped in a radially symmetrical class. They retain
their radial form until their death, which occurs long after the attain-
ment of bilateral symmetry in the normal embryos.

Certain of these embryos have the form of a hollow blastula
(Fig. 17), resembling the blastula of the sea-urchin. The prototroch
is well developed and large entoblast cells are present below it, but
no gastrulation occurs at any time in its life history. Another type
(Fig. 18) shows the beginning of gastrulation, which in some instances
progresses to the formation of a large enteric mass (Fig. 19). In
certain cases the enteric cavity may be formed (Fig. 20). In none of
these types is there any outward evidence of bilateral symmetry.4
The blastopore does not undergo its antero-ventral migration, and
the oesophagus is not formed.

It is quite probable, although the cell-lineage has not been worked
out, that these forms arise from eggs in which the normal bilateral

4 Other forms have been observed which might be classed as transition types,
but their abnormal form makes them quite difficult to analyze.

13

1'JS

ALBERT TYLER

cleavages do not take place. Torrey (1903) has described eggs of
Thalassema that have more or less completely reverted to the radial
type. He states, "There is, also, strong reason for believing that this
type never develops into an adult; for I found a few gastrula stages in
which no A' group, no M cells, no larval mesenchyme cells from
the third quartet and no large anterior oesophagoblast (262.2+) could
be distinguished." It may be fairly safe to assume then that the cell
lineage of the radial embryos described above is similar to that de-
-nibed by Torrey, and that the failure to develop into a bilateral

17

18

Fn.s. 17 20. Radially symmetrical embryos from artificially activated eggs;
dr.i\vn from total mounts of preserved specimens.
FlG. 17. I lollow blastula of 24 hours.
I ii.. is. lU'ginning gastrula of 22 hours.
FlG. I1'. ' ..i-irula of 28 hours, showing large enteric mass.
I i',. 20. < ..isirula of 28 hours, showing enteric cavity.

trorhophore result > from the failure of certain bilateral cleavages to
occur, so that m> A ^nnip, M cells, etc. are formed.

Approximately one-quarter of the top swimming embryos and
about 5 per cent of the bottom swimmers were radially symmetrical

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 199

or approached the radial type. No radial embryos were observed
among the trochophores of the normally fertilized eggs. Their
occurrence as a result of artificial activation may be taken as an
expression of the inability of the parthenogenetically activated egg to
completely negotiate its problem of bilateral symmetry. The low
percentage of normal development obtained is then accountable by
the assumption that it is a matter of chance whether the response of
the egg is similar to that resulting from the entrance of the sperm or
whether it responds in an "abnormal" fashion to the artificial treat-
ment. Before this could be put in more concrete terms, we would
have to know much more about the response of the egg to the sperm
in cases where the entrance point determines the median plane of
the embryo.

MATURATION AND CLEAVAGE OF THE ARTIFICIALLY ACTIVATED EGGS

It has been previously noted that 100 per cent activation may be
obtained with certain times of exposure to different dilutions of sea
water, but that the percentage of cleavage decreased with increased
activation so that at the optimum point practically no cleavage is
obtained. The eggs at the optimum point were all found to extrude
two polar bodies,5 whereas on either side of that exposure time in-
creasing numbers of eggs with one and with no polar bodies were found.
Eggs were therefore isolated according to the number of polar bodies
formed in order to determine which type cleaved and developed.

TABLE V

Relation between Cleavage and Number of Polar Bodies Formed

Type of Egg Isolated

Number of Eggs
Cleaved

Number of Eggs
Uncleaved

No polar bodies .

876

156

One polar body

0

83

Two polar bodies

0

518

The results obtained by isolating the various types of eggs are
given in Table V. The eggs were isolated at 45 to 75 minutes after
treatment and cleavage recorded about three hours later. A binocular
microscope magnifying 125 diameters was used in making the isolations
and the eggs were rolled around to make sure that the presence of
polar bodies was not overlooked. None of the eggs with one or with

5 The first polar body may divide in about half of the eggs, as is the case in the
fertilized eggs, so that when eggs with two polar bodies are referred to in the text it
is meant to include also those with three polar bodies.

200 ALBERT TYLER

two polar bodies were found to divide, whereas 85 per cent of the eggs
with no polar bodies divided. This result is in accord with that
obtained by Morris (1917) on the eggs of Cumin gia. She found that
a high percentage of maturation resulted in a low percentage of cleavage
stages and larva?, and from 1215 eggs with no polar bodies obtained
five larva-, whereas from 313 eggs with one or two polar bodies one
embryo developed that was evidently uncleaved.

In Urechis the eggs with one or with two polar bodies can be
readily distinguished from those with none. The former are decidedly
more "normal" in appearance and in their reaction to the treatment
than are the latter. It is quite surprising then that only the abnormal-
looking eggs with no polar bodies develop, whereas the others do not
even cleave. These types of eggs will now be described in some
detail.

\Yhen the eggs are placed in the dilute sea water, the indentation
rounds out and the whole egg swells due to the intake of water
(Fig. 9).6 Upon return to normal sea water the egg shrinks and the
indentation reappears. The egg thus returns to its original shape.
This is true for all the eggs so treated regardless of their subsequent
history.

I'l.ATK I!

Photomicrographs of Ihing objects. The magnification is 330 diameters.

FIG. 21. Artificially activated egg five minutes after return to normal sea water,
showing beginning of membrane elevation and rounding of indentation.

FIG. 22. Artificially activated egg ten minutes after return to normal sea
water, showing beginning of dissolution of germinal vesicle.

FIG. 23. Artificially activated egg fifteen minutes after return to normal sea
water. Compare with Fig. 10 of fertilized egg.

FIG. 24. Artificially activated egg 34 minutes after return to normal sea water,
showing first polar body.

FIG. 25. Artificially activated egg 47 minutes after return to normal sea water,
showing second polar body.

FIG. 26. Artificially activated egg 50 minutes after return to normal sea water,
showing single giant polar body.

FIG. 27. Artificially activated egg 45 minutes after return to normal sea water,
showing presence of indentation, disappearance of germinal vesicle, and absence of
mi ml mine and of polar bodies.

I [G. _'S. Antipolar view of an egg of the type shown in Fig. 27, showing small
nucli

I IG. 29. Same egg as in Fig. 27, 40 minutes later, showing rounding up of egg,
eltnaiidii of membrane, and formation of cleavage spindle (indicated by clear area)
at right angle* t<> polar axis.

FIG. 30. Same r^g as '" l''g- 27, 5 minutes later, completely rounded up.

FIGS. 39 AND 40. Show reappearance of indentation in its original position
when egg is returned to normal sea water after treatment with dilute sea water.

'"' In dilutions of sea water above 65 per cent the indentation does not com-
pletely round out.

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 201

PLATE II.

•mm

202

ALBERT TYLER

It has been noted above that with certain lengths of exposures
to dilute sea water, the time of treatment depending upon the con-
centration employed, 100 per cent of the eggs become activated.
Such eggs upon return to normal sea water behave very much as
though they had been normally fertilized except that no entrance cone
is formed. The germinal vesicle breaks down, the indentation rounds
out, and the membrane is elevated very much in the same manner as
in the fertilized eggs (Figs. 21, 22, and 23). The first and second polar
bodies (Figs. 24, 25) appear at the same time as they do in the fertilized
eggs, if allowance is made for the time of return to normal shape in
sea water.

This last point is illustrated in Table VI. The figures given in

TABU; VI
Time of Polar Body Extrusion in the Artificially Activated Eggs

Time for First Polar Bd.lv

Time for Second Polar Body

Dilution of

Tempera-

Length of

Sea Water

ture

Exposure

Total

Calculated

Total

Calculated

per cent

0 C.

minutes

minutes

minutes

minutes

minutes

30

20.0

i

3

35

34!

46

45!

1

35

33i

47

45!

2i

36

32 j

50

46}

v.

41

.UJ

55

48i

Pert, eggs

34

44

45

23.5

1
3

30i

30

41

40!

1

31

29$

42

40$

2k

33

29 i

43

39|

4!

36^

29^

48

411

Pert, eggs

30

40

50

19.5

1

36

34*

48

4<>!

H

37$

35

•I')'.

47

2*

38

34J

51

4?i

4-J

393

32!

52!

45|

1 . rt. eggs

34 !

46

" The total time is taken from the beginning of the treatment.
t The calculated time is that obtained by allowing for the swelling and shrinking
of the egg.

column- l</ui ,ni<l >i\ of the table are for the time of appearance of the
polar bodice fiom the beginning of the treatment. If allowance is
nude IIM the time n| exposure, the time for appearance of the polar
bodie^ still increases with the length of exposure. \Ve must therefore
subtr.iet .in amount equal to the time for return to original shape.
A rough determination of the time of shrinkage shows it to be about

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 203

half the swelling time. Allowing for that, the time of appearance of
the polar bodies is comparable with the time of polar body extrusion
in the fertilized eggs. This holds for all eggs that form two polar
bodies, whether or not they have received the exposure resulting in
100 per cent activation. It is evident from these results that develop-
ment does not start in the dilute sea water but after its return to normal
sea water. This is in accord with the fact that the egg first assumes
its original form on return to normal sea water and then proceeds to
develop. It is also consistent with the fact (to be reported in a
subsequent paper) that eggs which have been "over exposed" to
dilute sea water do not become activated although they are still
capable of fertilization.

The eggs which extrude only one polar body occur in much fewer
numbers. The maximum amo.unt observed was about one per cent
of the activated eggs. When only one polar body is extruded it is

FIGS. 31-34. Early cleavages of artificially activated eggs, drawn from photo-
graphs of living objects.

FIG. 31. Two-cell stage of an artificially activated egg.

Three-cell stage of an artificially activated egg.

Four-cell stage of an artificially activated egg, showing polar cross-

FIG. 32.

FIG. 33.
furrow.

FIG. 34.

Five-cell stage produced by the egg shown in Fig. 32.

204

ALBERT TYLER

generally quite large, being approximately equal in size to both
polar bodies of the normal egg (Fig. 26). The average time of ex-
trusion of the polar body was 45 minutes, which is the time of second
maturation division in the normal egg.

The eggs that produce, no polar bodies respond to the treatment
in such a way that they were first classified as "poorly activated."
The germinal vesicle disappears much more slowly than in the others.
The rounding out of the indentation and elevation of the membrane
may not occur for one to two hours. Figure 27 shows such an egg
45 minutes after treatment. Although the germinal vesicle has dis-
appeared, no membrane is present and the indentation remains.

A small nucleus is later formed which persists until the time of
rounding up of the egg (Fig. 28). The time at which the egg rounds
up varies from 40 minutes to over two hours. As the indentation dis-
appears an irregular membrane is lifted oft from the surface of the egg
(Figs. 29 and 30) and a spindle develops at right angles to the egg axis.
\Yithin 20 minutes after the rounding up, the egg elongates and
divides (Fig. 31).

The first division may be equal or unequal. Of the 876 divided
eggs listed in Table V, 485 had formed an equal or nearly equal two-
cell stage. Four of the eggs had divided at once into three cells and
the rest had formed an unequal two-cell stage. The next division of
both the equally cleaved and unequally cleaved eggs may be into
three cells (Fig. 32) or into four cells (Fig. 33). The time of division
is quite irregular. Table VII gives the time schedule of division for

TABLE VII

Number of Eggs in Different Cleavage Stages at Various Times after Treatment —

ftom Isolation of Undeaved Eggs

Time after

Treatment

l-ccll

2-cell

3-ccll

4-cell

hours

minutes

1

10

300

0

0

0

1

20

298

2

0

0

1

35

270

30

0

0

1

45

267

31

2

0

2

0

202

96

2

0

2

30

112

160

24

4

3

0

30

201

41

28

300 eggs without polar bodies that were isolated before cleavage
began. The i.iMc -ho\\s that the behaviour of the eggs with respect
to time of <h\iHon i> tar Irom uniform; the second division in some
eggs may occur e\en before the first division in others.

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 205

The eggs which divide into three cells at the second division
generally divide into five at the next (Fig. 34), the original undivided
cell still remaining undivided. Those which divide into four cells at
the second division may divide into six, seven, or eight cells at the
third cleavage (Figs. 35, 36, 37, 38). The third cleavage is dexiotropic
as in the normal egg. The four-cell stage also shows evidence of
spiral cleavage in the presence of the cross furrow, which results from
the A and C blastomeres being higher than the B and D. The time
of third cleavage is also quite variable. In order to determine whether

FIGS. 35-38. Early cleavages of artificially activated eggs, drawn from photo-
graphs of living objects.

FIG. 35. Six-cell stage produced by the type of egg shown in Fig. 33.

FIG. 36. Seven-cell stage produced by the type of egg shown in Fig. 33.

FIG. 37. Eight-cell stage produced by the type of egg shown in Fig. 33; showing
dexiotropic cleavage.

FIG. 38. Later eight-cell stage showing dexiotropic cleavage.

a more uniform result might be obtained, equally cleaved eggs were
isolated according to whether they divided early or late. The results
of two such experiments are given in Tables VIII and IX. No
significant differences in types of cleavage obtained are evident, and
no greater uniformity in time of division results, when the divided

200

ALBERT TYLER

TAHI.K VIII

Number of Eggs in Different Cleavage Stages at Various Times after Treatment-
front Equally Cleaved Eggs Isolated Early

Time after Treatment

2-cell

3-cell

4-cell

S-cell

6-cell

7-cell

8-cell

1 hour 50 min. to 2
hrs 15 min.

190

o

o

o

o

o

o

' hours 30 min.

168

22

0

o

o

o

o

2 " 47 "

143

47

0

0

0

0

0

2 " 53 "

129

47

14

0

0

0

0

3 " 0 "

106

60

24

0

0

0

0

3 " 5 "

106

50

24

10

o

o

o

3 •• 10 "

92

59

26

10

0

3

0

3 " 30 "

56

48

39

77

6

11

8

4 " 0 "

21

36

31

38

22

17

25

eggs are isolated early or late. The normal-looking eggs in the eight -
cell stage were isolated from the early dividing and late dividing
series. The embryos produced were examined about 32 hours later.

TABLE IX

Number of Eggs in Different Cleavage Stages at Various Times after Treatment —

from Equally Cleaved Eggs Isolated Late

Time after Treatment

2-cell

3-ccll

4-cell

5-cell

6-cell

7-cell

8-cell

1 hours 20 min. to 2
hrs 45 min

J()7

0

0

0

o

()

()

2 hours 52 min.

184

0

23

o

o

o

o

3 " 0 "

159

25

23

0

0

0

0

3 " 15 "

72

SI,

49

0

0

0

0

3 " 35 "

M

86

57

(i

0

0

0

3 " 5d "

53

74

51

15

4

4

6

4 " 15 "

24

58

62

29

11

10

13

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 207

The results are listed in Table X. It may be seen that the retarded
eggs behave as well with respect to the type of embryos produced as
the early dividing ones. The table also shows that only very few of
the eggs that form a normal-looking eight-cell stage develop into

TABLE X

Embryos Produced by Normal-Looking Eggs Isolated in the Eight-Cell Stage

Type of Egg

Number
Isolated

Normal
Embryos

Top
Swimmers

Bottom
Swimmers

Early dividing

57

1

8

30

Late dividing

78

1

5

45

normal embryos. It might be claimed that this is due to irregularities

in the division of the chromosomes, but such irregularities would
probably manifest themselves in abnormal divisions before that time.
This point will, however, be checked when a study of the chromosomes
of the artificially activated eggs is made.

POLARITY OF THE ARTIFICIALLY ACTIVATED EGGS

The abnormal development might also be interpreted on the basis
of a disturbance in the polarity of the egg. However, assuming that
the indentation marked the pole of the unfertilized egg, it was found
that the polarity of the artificially activated eggs remained unchanged.
These observations were made by means of the vaseline-slide method.
When the eggs are allowed to swell in the dilute sea water, the in-
dentation rounds out, and the germinal vesicle is nearest the surface
at the formerly indented region (Fig. 9). Upon return to normal
sea water the indentation reappears. The reappearance of the in-
dentation can be followed quite easily (Figs. 39 and 40) with respect
to the germinal vesicle and proves to occupy the same position as in
the original egg. In the eggs which form polar bodies as a result of
the treatment, the indentation soon rounds out and the membrane
is elevated (Figs. 21, 22, and 23). The contents of the dissolved
germinal vesicle moves towards the pole (as determined by the in-
dentation) and the first polar body later appears at that point (Fig. 24).
Fourteen eggs were followed in the manner described and all gave
the same results. In the eggs which do not form polar bodies, the
indentation persists for some time. Eighteen eggs of that type were
followed to the first division and in every case the first cleavage plane
passed through the polar axis (as determined by the indentation).
However, if the original polarity is retained, the second cleavage plane

208 ALBERT TYLER

must also coincide with the polar axis. This was found to be true for
all of six eggs followed to the second division. It appears then that
the abnormal development cannot be attributed to altered polarity

of the egg.

DISCUSSION

In the eggs of Urechis the sperm appears to be important for the
determination of the plane of bilateral symmetry. The small per-
centage of normal embryos obtained from the artificially activated
eggs may then be interpreted as due to the failure of the processes
connected with the establishment of bilaterality to occur in all of the
eggs, although they may take place "by chance" in some of the eggs.
The production of radially symmetrical embryos from the treated eggs
is consistent with such an interpretation. In other eggs with spiral
cleavage (Nereis, Cumingia, and Chcetopterus) there is evidence for
the determination of the plane of bilateral symmetry by the entrance
point of the sperm. In such eggs development closely simulating
the normal has been shown to be quite difficult to obtain, although the
amount of normal-looking cleavage obtained is generally quite con-
siderable.

The view expressed here might presumably be tested by the use
of an artificial agent the action of which on the egg is more nearly
like that of the sperm. The needle of the puncture method may be
taken to be such an agent. Brachet (1911) has studied the relation
of the point of puncture to the plane of bilateral symmetry in the egg
of the frog. In the frog's egg the gray crescent has been shown to
form opposite the entrance point of the sperm, and the plane of
bilateral symmetry is determined by the position of the crescent.
Brachet found that in the punctured frog's egg the position of the
crescent has no definite relation to the point of puncture. In those
eggs that develop, the median plane forms in relation to the crescent.
This would appear then to be evidence against the idea expressed
.iliove, and Brachet interpreted the result to mean that the egg itself
lias a sort of labile bilaterality. However, he points out that the action
of the needle is not quite comparable to that of the sperm. He states
(p. .->57), "le stylet, si fin qu'il soit, est tin instrument grossier, dont
Faction CM plus brutale et surtout beaucoup plus rapide; si habile
que soit 1'optTateur, le fil de platine ou de verre a atteint le centre de
1'ceuf en une fraction de seconde. Aussi 1'oeuf reagit-il en masse, et
presque simultan£ment dans toutes ses parties, tout comme quand il se
sent penetre en iiu-me temps par 8 ou 10 spermatozoi'des." He
suggests using a very fine needle of one to five micra thickness, and
allowing it to take ten to fifteen minutes to reach the center of the
upper hemisphere of the

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 209

The failure of the eggs that extrude two polar bodies to develop
has been noted in one other case of artificial parthenogenesis in
annelids and mollusks. Morris (1917) found in Cumingia that very
few eggs with one polar body or with two polar bodies divide, whereas
those with no polar bodies may develop into swimming embryos.
This result was confirmed by Heilbrunn (1925). In other cases it
has been reported that development may proceed whether one, two
or no polar bodies are formed. This has been noted for Mactra
(Kostanecki, 1904, 1911), Thalassema (Lefevre, 1907), Asterias (De-
lage, 1901, 1904; Garbowski, 1903), Amphitrite (Scott, 1906), and
Cheetopter-ns (Allyn, 1912; Lillie, 1906).

In Urechis, one of the most interesting features of the eggs which
fail to extrude polar bodies is their slow and comparatively abnormal
response to the treatment. Those that extrude polar bodies respond
to treatment as well as though they had been fertilized. It is sur-
prising then that only the former should develop, since it has been
generally held that the degree of success attained by artificial agents
depends on the extent to which the initial response of the egg to the
treatment resembles its response to the sperm. Just (1922), for
example, states, "the highest per cent and normality of cleavage and
of plutei result when the membrane separation most closely simulates
the separation of the vitelline membrane as a cortical response to
insemination." In Urechis the eggs which extrude polar bodies have
undoubtedly received an optimum treatment, whereas the ones that
do not form polar bodies have apparently been incompletely activated.
But the former have retained only one fourth of the original chromatin
of the egg. The ability of the Urechis egg to divide would thus appear
to depend upon the amount of chromatin present relative to the
cytoplasm. However, it is conceivable that by means of other agents
or by additional treatment the eggs that extrude two polar bodies
may be made to develop, since in the sea-urchin egg fairly normal
plutei have been obtained from eggs which contain the haploid number

of chromosomes.

SUMMARY

1. The eggs of Urechis may be activated by hypotonic solutions
ranging from distilled water to 80 per cent sea water, and trochophores
indistinguishable from those produced by fertilized eggs may be
obtained.

2. The time of exposure necessary to bring about activation in-
creases with increased concentration of sea water.

3. The activated eggs may extrude two, one or no polar bodies.
When the exposure is such as to produce 100 per cent activation,

210 ALBERT TYLER

two polar bodies are extruded on practically all the eggs, but to either
side of this optimum exposure the proportion of activated eggs with
no polar bodies increases. Eggs with one polar body occur in small
numbers, and the single polar body is equal in size to the two polar
bodies of the normal egg.

4. Only the eggs with no polar bodies divide and form embryos.
Those with two polar bodies or with one polar body may sometimes
pi xluce uncleaved swimmers.

5. The eggs which extrude no polar bodies show a very poor re-
sponse in other respects to the treatment. The eggs which extrude
two polar bodies respond to the treatment in very much the same
manner as if they had been fertilized. The fact that only the former
cleave and develop is inconsistent with the view that for development
to be obtained the artificially activated egg must behave like the
fertilized egg in its initial response to the treatment.

6. The cleavage of the parthenogenetic egg is quite variable both
in form and in time of division.

7. The normal embryos are produced in very small numbers.
Even when normal-looking eggs in the eight-cell stage are isolated,
less than 2 per cent normal development is obtained.

8. Among the abnormal embryos produced by the parthenogenetic-
ally activated eggs, a number of larvae are found that may be classified
in a radially symmetrical group. Such embryos may have the form
of hollow blastula-, early gastrula?, or late gastruhc, and their develop-
ment ends without the appearance of bilateral symmetry.

9. The indentation of the unfertilized egg is found to mark the
pole, and with respect to that point the polarity of the artificially
activated egg is found to remain unchanged.

10. The first cleavage plane is found to pass through the entrance
point of the sperm in 71 per cent of the cases. It may be concluded
from this that the sperm is instrumental in determining the plane of
bilateral symmetry.

11. It is suggested that in the absence of a localized external agent
such as the sperm, it is a matter of chance whether the establishment
of bilateral symmetry and other processes associated with the orienta-
tion of the embryo may be effected. This would account for the low
prn-rnt.ige of normal development and the presence of radially

symmrt i ical embryos.

BIBLIOGRAPHY

ALLYN, II. M., 1('1_'. I In- Initiation of Development in Chu-topterus. Biol. Bull.,

24: 21.
BRACHI i. A., 1'Ml. I tudes sur les localisations germinales et leur potentiality

n rllr il.tns I'u uI parthenog6netique tie Kana fusca. Arch, de Biol., 26: 337.

NORMAL EMBRYOS BY ARTIFICIAL PARTHENOGENESIS 211

BULLOT, G., 1904. Artificial Parthenogenesis and Regular Segmentation in an

Annelid (Ophelia). Univ. of Cat. Pub. Physiol., 1: 165.
DELAGE, YVES, 1901. Etudes expc-rimentales chez les Echinoderms. Arch, de

Zool. Exper., 3me Ser., 9: 285.
DELAGE, YVES, 1904. Elevage des larves parthenogenetiques d'Asterias glacialis.

Arch, de Zool. Exper., 11: 27.
FISCHER, M. PI., 1903. Artificial Parthenogenesis in Nereis. Am. Jour. Physiol.,

9: 100.
FISHER, \V. K., AND G. E. MAcGiNiTiE, 1928. A New Echiuroid Worm. Ann. and

Mag. Nat. Hist., Ser. 10, 1: 199.
FISHER, W. K., AND G. E. MACGINITIE, 1928. The Natural History of an Echiuroid

\Yorm. Ann. and Mag. Nat. Hist., Ser. 10, 1: 204.
GARBOWSKI, M. T.. 1903. Uber parthenogenetische Entwickelung der Asteriden.

Bull. Internal. Acad. des Sci. de Cracovie (Jan. 1903), pp. 810-831.
HEILBRUNN, L. V., 1925. Studies in Artificial Parthenogenesis. IV. Heat Partheno-
genesis. Jour. Exper. Zoo!., 41: 243.
JUST, E. E., 1912. The Relation of the First Cleavage Plane to the Entrance Point

of the Sperm. Biol. Bull., 22: 239.

JUST, E. E., 1915. Initiation of Development in Nereis. Biol. Bull., 28: 1.
JUST, E. E., 1922. Initiation of Development in the Egg of Arbacia. I. Effect of

Hypertonic Sea-Water in Producing Membrane Separation, Cleavage, and

Top-Swimming Plutei. Biol. Bull., 43: 384.
KOSTANECKI, K., 1904. Uber die Veranderungen im Inneren des unter dem Einfluss

von KCl-Gemischen kiinstlich parthenogenetisch sich entwickelnden Eis

von Mactra. Bull. Internal. Acad. des Sci. de Cracovie (Jan. 1904), pp.

70-91.
KOSTANECKI, K., 1908. Zur Morphologic der kiinstlichen parthenogenetischen

Entwicklung bei Mactra. Zugleich ein Beitrag zur Kenntnis der vielpoligen

Mitose. Arch. Mikr. Anal., 72: 327.
KOSTANECKI, K., 1911. Uber parthenogenetische Entwicklung der Eier von Mactra

mit vorausgegangener oder unterbliebener Ausstossung der Richtungs-

korper. Arch. f. Mikr. Anat., Vol. 78, Abt. II, p. 1.
LEFEVRE, GEORGE, 1907. Artificial Parthenogenesis in Thalassema mellita. Jour.

Exper. Zool., 4: 91.

LILLIE, F. R., 1910. Studies of Fertilization in Nereis. Jour. Morph., 22: 361.
LOEB, J., 1908. Ueber die Entwicklungserregung unbefruchteter Annelideneier

(Polynoe) mittelst Saponin und Solanin. Pfltiger's Arch., 122: 448.
LOEB, J., 1912. The Comparative Efficiency of Weak and Strong Bases in Artificial

Parthenogenesis. Jour. Exper. Zool., 13: 577.

MORGAN, T. H., AND ALBERT TYLER, 1930. The Point of Entrance of the Spermato-
zoon in Relation to the Orientation of the Embryo in Eggs with Spiral

Cleavage. Biol. Bull., 58: 59.
MORRIS, MARGARET, 1917. A Cytological Study of Artificial Parthenogenesis in

Cumingia. Jour. Exper. Zool., 22: 1.
SCOTT, J. W., 1906. The Morphology of the Parthenogenetic Development of

Amphitrite. Jour. Exper. Zool., 3: 49.
TORREY, J. C., 1902. The Early Development of the Mesoblast in Thalassema.

Anat. Anzeig., 21: 247.
TORREY, J. C., 1903. The Early Embryology of Thalassema mellita (Conn).

Annals N. Y. Acad. Sci., 14: 165.
TREADWELL, A. L., 1902. Notes on the Nature of Artificial Parthenogenesis in the

Egg of Podarke obscura. Biol. Bull., 3: 235.

Vol. LX, No. 3

June, 1931

THE

BIOLOGICAL BULLETIN

PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY

OBSERVATIONS ON THE METABOLISM OF
SARCINA LUTEA. I

R. W. GERARD AND I. S. FALK

(From the Departments of Physiology and of Hygiene and Bacteriology,

University of Chicago)

The experiments reported in this communication were designed to
provide some information on the dynamics of oxygen diffusion into
living cells. It seemed desirable, a priori, to perform the observations
on very small, spherical cells. To this end, micrococci were the forms
chosen. After preliminary experimentation, it was found that a strain
of Sarcina lutea which had long been maintained on artificial culture
media fulfilled many of the requirements of the experiments: the
organisms are very small and discrete (this strain no longer forms
packets); the cells are spherical and highly uniform in diameter; and
in water or dilute aqueous solutions they remain sufficiently viable
for the requisite periods of time under the conditions used. As the
experiments progressed there were noted a number of interesting
characteristics of their metabolism, which are the main subjects of
this paper.

METHODS

1. Bacterial Cultures. — (a) Preparation of suspensions: The organ-
isms grow luxuriantly on plain nutrient agar prepared with peptone
and fresh meat infusion or beef extract, with or without the addition
of NaCl. After the completion of the preliminary experiments, the
medium used contained one per cent Difco peptone, 0.35 per cent
beef extract and 1.5 per cent agar in tap water. Agar slant cultures
in tubes or Kolle flasks were grown for 20-24 hours at 37° C. The
growth was washed off with distilled water, the suspension homo-
genized by gentle but prolonged shaking, the microorganisms precipi-
tated by centrifugation and then washed by two suspensions in and

15 213

-
tu i L ! B f

214 R. W. GERARD AND I. S. FALK

precipitation from distilled water. After the final washing, the secli-
mented cells were again suspended in distilled water and shaken to
give a concentrated homogeneous suspension.

(b) Enumeration of cells and pH measurements: The numbers
of bacteria per cc. were determined before and after exposure to the
specific conditions of the experiments by direct microscopical observa-
tion and by plate count. The former was performed by the hemo-
cytometer method, using a dilute solution of carbol fuchsin as a
diluting fluid; the latter as commonly used in the enumeration of
bacteria in water, milk and other fluids. The direct count includes
all cells visible under the microscope and the plate count only viable
cells which developed on the medium used into colonies visible to the
eye aided by a 5-10 X hand lens. Although all reasonable efforts
were made to insure accuracy, the usual presumptive errors of these
methods apply to our data.

The suspensions used in our experiments were, before dilution,
thick, yellow, pasty emulsions of the sedimented, packed cells. Micro-
scopical observations indicate a high degree of freedom from detritus.
The concordance of the direct and plate counts made before and after
performance of the experiments implies a high coefficient of vitality
and the essential absence of dead or senile cells in the suspensions
when first prepared.

Measurements of hydrogen ion concentration were performed with
indicators by a spot plate method. Inasmuch as the suspensions
consist of the bacterial cells in distilled water or dilute aqueous
solutions of low intrinsic buffering power, the stability of the pH of an
exposed suspension is controlled primarily by the buffering powers of
the cells. From preliminary titration experiments (after the method
of Shaughnessy and Falk, 1924) it developed that the strain of Sarcina
lutea used displayed rather slight buffering power. Hence, in per-
forming colorimetric pH measurements, we determined with which of
a series of solutions of very low buffer action, and varying but known
pH values, the test solution was isohydric. The standard solutions
were prepared by adding twice the usual concentration of indicator
solution to water and adjusting the pH of 10 cc. samples to various
levels (at 0.2 pH intervals) by adding appropriate quantities of HC1
or NaOH (cf. Fawcett and Acree, 1929). The colorimetric standards
were checked by electrometric measurements, using a buffer solution
standard and the quinhydrone electrode. The results of several
experiments appear in Table I.

It will be noted that the direct and plate counts generally agree
closely at the outset of an experiment and are not too widely apart at

METABOLISM OF SARCINA LUTEA

215

TABLE I

Hydrogen Ion Concentrations and Microbic Enumerations for Suspensions of

Sarcina Lutea

Date 1929

Suspension

pH

Number of cocci
per cc. ( X 10-8)

Direct

Plate

Apr. 27 Original 3.2 1.7

After 2 hours (room temperature):

Original 2.6 2.1

May 30 Original 7.4 143 140

After 6 hours:

In water shaken in air 6.8 107 430

June 6 Original 7.4 63 64

After 6 hours:

Original 7.2 150 30

In glucose shaken in air ? 150 30

In water shaken in N2 7.6 32 27

In glucose shaken in N2 7.5 41 27

In water shaken in air 8.0 40 7

July 5 Original 70 26

After 6 hours:

Original 102 29

In water shaken in air 23 9

In glucose shaken in air 27 11

In glucose shaken in 1% O2 13 14

In water shaken in 1% O2 58 13

Aug. 7 Original 85 13

After 11 hours:

Original 7.4 71 17

In water shaken in air 7.8 23 8

In glucose shaken in air 7.2 8 6

In water shaken in N2 7.2 6

In glucose shaken in N2 6.4 6

Oct. 16 Original 7.4 199 55

After 26 hours:

In water shaken in N2 7.8 182 29

In glucose shaken in N2 6.4 202 180

In glucose shaken in air 6.0 139 35

In water shaken in N2 7.8 170 35

In water shaken in air. 7.6 148 32

216 R. \y. GERARD AND I. S. FALK

the completion. When variations occur they are generally, as is to
be anticipated, in the direction of lower plate than direct counts.

(c] Size of the bacteria: The diameters of the bacterial cells were
determined by direct microscopical measurements with hanging drop
preparations and a filar micrometer. There were no significant
variations in the size of cells suspended in water and in dilute glucose
suspensions, with or without traces of dilute, neutral fuchsin. The
average diameter determined from several hundred measurements on
the cocci was 1.275 micra. In this series the maximum was 1.42,
the minimum 1.08. The distribution was approximately symmetrical
about the mean and the probable error of the mean was ± 0.07 M-

(d) Final handling: The thick bacterial paste obtained after
centrifugation was diluted 1 : 1 with distilled water or with 1.0 or
0.5 per cent glucose in water and well mixed by shaking (with a
glass bead). Of these suspensions 0.4 cc. (dry weight about 5 mgm.,
containing about 2 X 109 bacteria) or 0.8 cc. portions were placed
in the manometer vessels for determination of oxygen consumption,
etc., and a similar amount dried on a cover slip and weighed. In
certain experiments oxygen was bubbled through the suspension for
some time before beginning the observations. The pH was usually
determined before and after the manipulation by the method described
above.

2. Gases. — Tank nitrogen, of 99.7 per cent purity, was passed over
heated copper and through glass connections directly into the manom-
eters. Even this treatment did not give absolute freedom of oxygen,
but traces that remained did not affect the results (see below). Gas
mixtures containing 0.5 to 10 per cent oxygen were made by volumetric
mixing of the treated nitrogen and air in large bottles over water.
The nitrogen was always bubbled through first for a considerable
time to remove dissolved air, and the mixture was used at once after
preparation. To obtain accurate mixtures of 0.5 per cent or less,
a manometric method was used. This is described below.

3. Measurement of Oxygen Consumption, Carbon Dioxide Production,
-The Warburg methods were used throughout. Small chambers

of the cylindrical or conical type with an inset, used for studying
metabolism of nerve (Gerard, 1927), were found highly satisfactory.
With 0.4 cc. of bacterial suspension and 0.2 cc. of Af/10 NaOH in the
inset to absorb CO2, the constants were either about 0.25 or 0.35,
so that one cu. mm. change in gas volume gave a pressure difference
of 3 or 4 mm. (of Brodie's solution). Oxygen consumption, @Q,, is
expressed as cu. mm. consumed per hour per mgm. dry weight of the
bacteria.

MKTABOLISM OF SARCINA LUTKA 217

The respiratory quotient (R.Q.) was determined by comparison
of the manometer readings when CO2 was absorbed (NaOH in inset)
and when allowed to accumulate (water in inset) in parallel experi-
ments. Acid production was determined manometrically by the
liberation of COo from a bicarbonate-Ringer solution in the presence
of 5 per cent CO2 in nitrogen (Warburg, 1926).

All experiments were carried out at 20.1° C.

Gas mixtures with low concentrations of oxygen were obtained by
mixing in a manometer. The Brodie solution used in the manometer
capillary has a specific gravity so adjusted that 10,000 mm. equals
760 mm. of mercury. The pressure of the gas in the manometer
chamber is easily obtained from the barometer reading and the
difference of fluid level in the two limbs of the manometer. It is
usually sufficiently accurate, in determining the percentage of oxygen,
to assume 10,000 mm. Even the maximal correction for water vapor,
if dry air is admitted, is only 5 per cent and may ordinarily be omitted.
The oxygen admitted to the chamber filled with nitrogen is deter-
mined as follows: the fluid levels in the manometer are lowered (by
a screw at the bottom) until there is a "negative" pressure in the
chamber of circa 100 mm. (left side open to air = 50 mm., right
side to chamber = 150 mm.). The chamber is then opened to the
air for an instant by rapidly turning a stop-cock, the apparatus is
shaken for a moment, the fluid level on the chamber side is brought
to its initial setting and the level in the open limb read. This will
now be, say, 200 mm. At constant volume and temperature then
(the room air being nearly at the temperature of the thermostat and
coming to rapid equilibrium) the gas pressure has been increased by
150 mm. of air or 30 mm. of oxygen. With a little practice it is
possible to adjust the initial pressure difference and time of opening
the chamber so that any desired oxygen percentage up to 0.5 can be
obtained by one operation. This method permits the rapid and
accurate preparation of gas mixtures with very low oxygen concen-
tration and should be valuable in other studies where the influence
of oxygen pressure on oxygen consumption is to be determined.

RESULTS

1. Respiration in Air. — The rate of oxygen consumption for any
sample of the bacteria used is not constant with time, but tends to
be high at first and to fall rapidly at the outset and then slowly.
Sometimes it remains at a constant low level from the third or fourth
hour on; more often a slow fall continues through a 20-hour period.
A typical curve is shown in Fig. 1, and the results of several experi-
ments are given in Table II.

218 R. \v. GERARD AND I. S. FALK

The initial level of oxygen consumption and the amount of fall
have been fairly constant from one test suspension to another and
variations are not related to variations in the ages of the cultures used.
It might be anticipated that bacteria taken from a culture which has
passed the period of logarithmic rate of growth would show an approxi-

8.0

6.0

v^ «^^\.

\57o glucose

40

1.0

Hours

FIG. 1.

mately uniform level of oxygen consumption. Although viability
generally decreases in the non-nutrient menstrua, the rate of this
process is far too low to be invoked as the explanation of the decline
in oxygen consumption. The question arises: In what way does the
preparation of the suspension and its treatment in the manometer
chamber lead to the establishment of a "time zero"? In other words,
if the oxygen consumption of these cocci is assumed to have been
essentially constant in time, what causes it to begin a rapid fall
when its measurement is undertaken?

If a progressing mortality of the individual cells does not seem to
explain the findings, the early high respiration cannot, on the same
evidence (some variable fall in plate counts from start to end), be
interpreted as due to continued growth which was later suspended.
Continued agitation, gentle or vigorous, has been shown to have no
effect; and temperature and suspension menstruum can also be

METABOLISM OF SARCINA LUTEA

219

w

\o

OO vO ro
<^i ^-H CN

4J

OJ

E ,n

>O i— -t* r<i

r^ oo \O

I^ t^. t^ _i

1 £
.2 *

d/

^H ^H tN CN

(T5 rH CN

v— i ^— i ro fO

I'S

•0 4-*

X. i-

a cd

o o o o

000

^H CN rf I—

^ s

o fc
#£

CN CN IO ^

TjH tN ro

tN C^ -f ro

,-p rt
0- w

M

3

rr> <~O 00 -*

CN f*i r^

o\ I-H 10 o

K "

CN r^ fo <r>

IO CN fO

»-i r^l ** ^P

IO "O "O IO

CN O '-1 C^

00 O O

r^ r— rs ^

ro <~O 10 10

CN CN IO IO

vo CS lO

tN CN If) If)

O 10

\o \d

•* ro O; -^
•^ ^ ^O t~^

r*i o <~^ O

>O \O t^ '-i

*— i

o

' v. u

' u u,

•o

• • N?

. . to^

: :&5

1- LO

4J

•3

c

. ca nJ

• • c c

. CO OJ

• c c

. • *^

:d.s

O rt ^

Tod.s.s

o

o

Water suspension, in air.
W'ater suspension, in air .
0.2% glucose suspension,
0.2% glucose suspension,

Water suspension, in O2. .
Water suspension, in air.
0.5% glucose suspension,
0.2% glucose suspension,

Water suspension, in air.
Water suspension, in 1%
0.2% glucose suspension,

Water suspension, in 2.5^
Water suspension, in 1%
0.2% glucose suspension,
0.2% glucose suspension,

»

(N

0>

^H

I'

0

-*

rH

>.

Oj

s

o

ro

£*i

n!

2

o

<u

c

3
i — i

fi

j>,

"3
i — i

220

R. W. GERARD AXD 1. S. FALK

s
•-2

R

W

hJ
ffl

tn

•J5

O u; !/i sn »5

L*

•_ I

= _ _ _ _

ai O^

o c

s S-S^S,^

o

* A

^

f ;

: ||||

~ ~

*"" t *"

— r^ r^ \O VO

1

j . — — — »—

e

at

t

E
J I

1/5

tttt

.- 1- 3C 0

a v_

T

^- i— «— . rs

| 1

*

Y Y Y Y

"~J

A A

A I

k

0 *

1 I

\

3

-^ ^ 0

O v^

O

CO

-TfvOO — OO — —

1

T— i rvj f

1-

11

,. HI!

• — ' »O <"O IO

•y

Y

r^ CN r<^ -f

cs

> '

>

Y Y Y Y

\O "•!• oo '-i oo oo o-i <r;

CO ~ t-

, M ^ ^ ^ a,

•~

OC -f --•'-* ^-f — r^f^O

^H ON r-

„ (Vj -rf ^f O\ OO

-H ^-, <N r*5 ro

T— «

,

• t- ri '

rt X

XX K£

•

O 1J

•* o

. . o o c

£r • '

s (^

c = .=

.n . "

2 n

3

V« Cj j

3 en

<fl 'J3 _ -^

^ ' *s

0 u

'x 'X 33 3

•3 o \

- 3
3 PI

g i s o -

w

- C j^ j- w
rt cd /

e, a J

s —

CA CO

^^ l-l ^

>- >- u i_ >-

- 3 ..

,

33 CJ O> 3

3 C C

l) ^

0 O « *j C

J •*-<

J= ^= ^ "^ J=

J^ ^

2 *ts

C

_o

^* "^1* ^ ^ L-, ^^

-H — i

" — u

"O

^ Jj 'rt 'rt '« Jj

u u '•

4/ 4)

3 '5 '5

c

•^ ^ .S .S .S £

.-^ ,-p1 C

: c c

0

C3 r3 " "

- . — . —i

o

u u u cf c" C~ 1- ^

u u c

:" c~ i- c"

•^ *!^ '!2 O O O '^— '^
CQ Co Cu t . z. *M Co w

. — . — ,

2.2 '3 .2

cccScc cc

C C |

: c = c

"""""" a; aj <u "" ""

c- c- c- c. c. c. ._- _-

*****"• 10 CO (Q ^-fc-

^ (

i| c- |

.2 .2 :

3 3 .2 3

'3J '35 '35 us "J "> '35 '35

t/) in <•

I) U) ;/) I/l

ccc<ynja> cc

^ ^ ^ (A CiO (/I O O

O.O.Q.OOO a. c.

C C c

* £U

0 « 4) So
5 O Q. O

intninuou cnu)

c« « t

j o in o

333_3_3_3 33

3 3 :

33 3 3

U) U) —

- — U) —

uuu6060610 uu

u u '

i*0 bC u M

"cdrtnitf^^ ,«.rt

rt « ^

^ ^1^1

i.= :=:^ddd i^^

?- ?• c

5 O £- C/; O C/1

ex

IN

t- 0

-t

2.

•"*

»— <

u

ti ^J

>

M

Q

<s

JD

METABOLISM OF SARCINA LUTEA 221

excluded as factors leading to the fall in oxygen consumption. It
might then be assumed that the initial values are excessive, probably
largely as a result of partial asphyxia of the organisms induced in the
course of their preparation (and partly perhaps due to food reserves
stored from the nutrient medium), and that the lower values obtained
some time after an experiment is begun are to be regarded as the
"normal" ones rather than as "abnormal" levels resulting from some
depression. It may be noted that Callow (1924) obtained a Qo> of
7.0, constant for 10 hours, using Sarcina aurantiaca.

When the organisms are suspended in a 0.2-0.5 per cent glucose
solution instead of water, the same progressive fall of oxygen con-
sumption is observed. The Qo, values at each point in time are,
however, much greater in the former case (by 100 per cent to 300
per cent), so that the respiration curve for the bacteria in the glucose-
containing menstruum lies above but parallel to that for the bacteria
in water. (Fig. 1, Table II.) The influences of glucose and other
substances on the respiration of Sarcina lutea are reported upon in
more detail in the following paper.

The R.Q. in water and in glucose solution was determined in two
experiments. In water it remained constant at about 0.67 during
8 hours. (Gotschlich, 1912, cited quotients of 0.71-0.78 for bacteria
metabolizing in the absence of fermentable substances, and Stephenson
and Whetham, 1923, found R.Q.'s of about 0.7 after all added glucose
had been utilized.) In glucose solution, the R.Q. for the first hour
was 0.95 and then fell during seven hours to 0.71, the average for the
whole run being 0.81. These figures suggest that, even in glucose
solution, less than half the oxygen consumed is used to oxidize carbo-
hydrate, although the presence of glucose has more than doubled the
amount of oxygen taken up. (These values of R.Q. are lower than
those recorded for bacteria on a nutrient pabulum capable of sup-
porting their growth. Vide Soule, 1928.)

2. The Influence of Oxygen Pressure. — The oxygen consumption of
Sarcina lutea in equilibrium with oxygen at varying partial pressures
has been determined and it has been found that in air or pure oxygen
the respiration is essentially the same. When the bacteria were held
in water, with a lower rate of respiration, 2.5 per cent oxygen in the
gas or even one per cent oxygen, in two cases, sufficed as well as the
pure gas. \\ ith the glucose menstruum (and more rapid oxygen
consumption) one per cent oxygen was definitely too low a concen-
tration to support full respiration and 2.5 per cent of the gas was
sometimes inadequate. Concentrations below one per cent were
uniformly insufficient to support the rate of oxygen consumption
possible in air.

) ) )

R. \V. GERARD AXD I. S. FALK

When, after a period of complete anoxia, oxygen was admitted to
a concentration of 0.5 per cent or less, there was at once a moderate
oxygen consumption. Although the rate of respiration was already
far less than in air, it fell off considerably in a short time, much as
previously described for experiments in air (Fig. 1). The fall in this
case was not due to a diminishing supply of oxygen following utilization
of the small amount admitted, for it occurred when only 5 to 10 per
cent of the oxygen had disappeared. More than doubling the amount
of shaking did not affect the value of the critical oxygen pressure That
failure to attain equilibrium between gas and liquid phases did not
play a role is further attested by experiments made in another con-
nection. Hydrogen wras absorbed by an unsaturated fat in water
suspension under the influence of a catalyst at a rate over 100 times
that of the oxygen absorption here, yet the rate of shaking was ade-
quate.

3. Oxygen Debt. — It seemed possible that the regular occurrence of
a relatively high oxygen consumption directly after admitting this
gas, or at the start of an experiment as the bacterial suspensions
became well oxygenated by shaking, might represent the discharge
of an oxygen debt. That such a debt can be accumulated is easily
demonstrated: a suspension kept for some hours in nitrogen and

10

Hours

8 $ 10 //

FIG. 2.

then exposed to air takes up oxygen at a greatly increased rate (Fig. 2).
At the start, this rate may be considerably greater than observed
when the bacteria are kept in air throughout. It subsequently

METABOLISM OF SARCINA LUTEA

decreases just as under the conditions previously considered. This
burst of increased oxygen consumption is greater after long than after
short periods of anoxia, and for the same period is greater when the
bacteria are suspended in glucose solution than in water (Table II).
Glucose is capable, apparently, of stimulating the metabolism of
these organisms not only in the presence'f oxygen (leading to a more
rapid consumption of the gas), but also in its absence (leading to an
increased oxygen debt).

To determine whether previous anoxia was the cause of the high
respiration observed at the start of the simple experiment in air,
the following test was made. A particular bacterial suspension was
divided into two portions. Oxygen was bubbled through one and
nitrogen through the other for an hour and then each was rapidly
introduced into manometers in air. The usual high initial oxygen
consumption values were obtained with the sample treated with
nitrogen, while they were much lower for the oxygen-treated portion.
The slow fall, continuing for many hours after mounting, occurred
with both. This slow fall of oxygen consumption with time might
be related to the exhaustion of the substrate to be oxidized rather
than to the oxygen itself. In the case of water suspensions, at least,
the only material for oxidation that the cells have is their own sub-
stance and such food as may be stored during growth in a nutrient
medium. During 24 hours the carbon in the carbon dioxide given
off corresponds to 1 to 2 per cent of the dry weight of the organisms.

4. Acid Formation. — The oxygen debt these cells incur in an
atmosphere of nitrogen might be due to accumulation of oxidizable
metabolites or to exhaustion of an oxidizing reserve. The first case
is reminiscent of the metabolism of muscle and the role played by
lactic acid, the second suggests the metabolism of nerve. We therefore
attempted to ascertain whether acid is produced in the metabolism
of the microbic cells under the conditions of our experiments-
especially in the absence of oxygen.

The pH of the cell suspension was determined in many experiments
before and after manipulation in the manometers, using the indicator
method described above. There was no consistent change under any
of the conditions studied — water or glucose solution, oxygen present
or absent. The initial and final pH values were about 7.4, though
occasional final values were as high as 8.0 and as low as 6.0, especially
in glucose suspensions. Acid formation was also measured mano-
metrically by the liberation of CO2 from bicarbonate-Ringer solution
in equilibrium with 5 per cent CO2 in N2. This was studied in the
presence or absence of glucose. During eight hours' asphyxia, the

224 R. W. GERARD AND I. S. FALK

o liberated, if due to lactic acid formation, indicated a production
of 0.0023 mgm. of lactic acid per mgm. dry weight per hour in glucose
solution and 0.0001 mgm. lactic acid in water (one experiment each).
The quantities of lactic acid formed in eight hours would require,
respectively, 14.5 and 0.6 cu. mm. of O2 for complete oxidation.
For many cells, including bacteria, studied by Meyerhof (1927) (also:
Meyerhof and Finkle, 1925) and others, the oxidation quotient (the
lactic acid produced in nitrogen minus that formed in oxygen, divided
by the oxygen consumed) is 3 or more. That is, one equivalent of
oxygen consumed prevents the appearance of three times as many
equivalents of lactic acid as it could oxidize. In these experiments,
in eight hours' exposure in an oxygen atmosphere, the bacteria con-
sumed: in glucose solution, 24 cu. mm., and in water, 17.5 cu. mm.,
of oxygen. The oxidation quotients are, therefore, 0.6 and 0.03.
The lactic acid (?) produced during asphyxia was only one-fifth of what
might have been anticipated in the glucose solution and was practically
zero in the water.

These results may be expressed in another way. In other experi-
ments, in either water or glucose, following a six-hour asphyxia the
average extra oxygen taken up during 6 hours was 50 per cent of the
normal oxygen consumed by the same number of bacteria in the same
time. The total oxygen debt in these experiments may therefore be
assumed to have been 9 cu. mm. in glucose solution and 5 cu. mm.
in water. Since in the water suspension only 0.6 cu. mm. of oxygen
could have been used to oxidize acid metabolites (typified by lactic
acid), nine-tenths or more must have acted in some other manner.
In glucose solution, the extra oxygen might have been used entirely
to oxidize metabolites, though presumably here also some would be
used otherwise. The possibility must be recognized, of course, that
more acid may have been formed than is recorded by the COo libera-
tion, some being neutralized by other cell buffers. Also, intermediate
substances of too wreak acidity to result in the liberation of CO2
would escape detection by the methods used.

5. Oxidizing Reserve. — The existence of a large oxygen debt after
asphyxiation, with no demonstrable accumulation of incompletely
oxidized metabolites to account for it, suggests that the extra oxygen
is utilized to replenish an oxidizing reserve which had been drawn
upon to continue oxidations during asphyxia. If this were the case,
some CC>2 should appear during the asphyxial period and a corre-
spondingly smaller amount during the following oxygen period. The
R.(J. of the extra respiration, while the oxygen store is being refilled,
should be correspondingly low. This does not appear to be the case,

METABOLISM OF SARCINA LUTEA

for the R.Q. of suspensions in water is the same early or later after
commencing a series of observations, whether or not the extra oxygen
consumption is occurring. Also, the absence of CO2 liberation in
nitrogen from water-bicarbonate suspensions renders improbable the
production of CO2 as well as of non-volatile acids. We do not wish,
on the basis of the few experiments performed, to conclude finally
that the oxygen debt developed by Sarcina lutea is due to the accumu-
lation of non-acid metabolites and not to the depletion of oxidizing
reserves, though this seems most probable. The work unfortunately
could not be continued at the time and the present report must suffice
to record these preliminary observations. It is interesting to note
that Meyerhof (1912) found in some of his studies on Vibrio metchnikovi
that in the absence of oxygen all chemical activity and energy liberation
appeared to be simply suspended. The falling <2o2 with time observed
for Sarcina, in contrast to the constant respiration of bacteria observed
by Callow (1924), may be related to such differences.

SUMMARY

The respiration of a strain of Sarcina lutea, growing as individual
cocci of uniform size and of relatively high viability under our experi-
mental conditions, was studied by means of the Warburg technique.
Suspensions of thoroughly washed bacteria in water or in glucose
solution were enumerated by direct and plate counts, and pH determi-
nations were made before and after manipulation in the manometer.
A simple method is described for obtaining accurate gas mixtures
with a low concentration of oxygen.

The oxygen consumption of various suspensions of the washed
cocci was found to be fairly constant in water suspension at about
2.6 cu. mm. O2 per mgm. dry weight per hour, or approximately
7 fj? O2 per single cell per hour: that is, over three times its volume
of oxygen was consumed per hour by each micrococcus. This value
is for the nearly constant level attained some hours after the start
of the experiment. Oxygen consumption is considerably more rapid
at first and falls along a roughly arithlogarithmic curve towards an
asymptote. This early excess appears to represent an oxygen debt
due to partial asphyxia produced in the course of preparation of the
suspensions.

In 0.2 to 0.5 per cent glucose solution, the rate of oxygen con-
sumption falls along a similar curve, but the rates are 100 to 300
per cent greater than for suspensions of cocci in water.

The respiratory quotient in water is constant at about 0.67 ; in
glucose it falls from 0.95 to 0.71 in seven hours (one experiment each).

226 R. W. GERARD AND I. S. FALK

The oxygen consumed by a suspension in water is independent of
oxygen concentration when this is above one per cent. For glucose
suspensions 2.5 per cent O2 or more is required for maximal consump-
tion. Below the critical values of oxygen tension, oxygen consumption
becomes less with diminishing oxygen concentration. When air is
admitted to the suspension after a period of complete anoxia, there
is observed a high initial rate of consumption followed by progressively
declining rates during several hours until a relatively constant rate
is approached. When small amounts of oxygen (0.5 per cent or less)
are admitted after anoxia, consumption is greater at first and then
falls as when air is admitted; but the maximal rate is low, usually
less than the normal in air.

The oxygen debt developed by these organisms is not apparently
associated with the production of detectable acid metabolites (in
water suspensions, at least), nor with the formation of carbon dioxide
during asphyxia. It may result from the accumulation of non-acid
metabolites, or of amounts of acid metabolites that do not equal or
exceed the buffering capacities of the cells. The oxidation quotient,
even in glucose solution, in contrast to the data for most cells that
have been studied, is not over 0.6.

BIBLIOGRAPHY

CALLOW, A. B., 1924. Biochem. Jour., 18: 507.

FAWCETT, E. H., AND S. F. AGREE, 1929. Jour. Bacterial., 17: 163.

GERARD, R. W , 1927. Am. Jour. Physiol, 82: 381.

GOTSCHLICH, E., 1912. Kolle u. Wassermann, Handbuch d. Path. Mikroorg., 1: 99.

MEYERHOF, O., 1912. Pfliiger's Arch., 146: 159.

MEYERHOF, O., 1927. Jour. Gen. Physiol., 8: 531.

MEYERHOF, O., AND P. FINKLE, 1925. Chem. Zelle, 12: 15".

SHAUGHNESSY, H. J., AND I. S. FALK, 1924. Jour. Bacterial., 9: 559.

SOULE, M. H., 1928. Gas Metabolism of Bacteria. Chapter XVIII in The Newer

Knowledge of Bacteriology and Immunology. Edited by Jordan and Falk,

Chicago.

STEPHENSON, M., 1930. Bacterial Metabolism. Longmans Green and Co.
STEPHEN-SON, M., AND M. D. WHETHAM, 1923. Proc. Roy. Soc., Series B, 95: 200.
\YARBCRG, O., 1926. Uber den Stoffwechsel der Tumoren. J. Springer, Berlin.

OBSERVATIONS ON THE METABOLISM OF
SARCINA LUTE A. II

R. W. GERARD

(From the Department of Physiology, University of Chicago)

In the preceding paper the effects of anarobiosis on this organism
and the influence of glucose on its oxygen consumption have been
reported. It seemed desirable to extend these observations to the
influence of lactic acid, as one of the carbohydrate intermediates,
and also to study the influence of the respiratory catalyst, methylene
blue, and of the inhibitors, carbon monoxide and cyanide. The
results of these experiments have been surprising in several respects.
The methods used were exactly as previously described, and all
experiments were carried out at 22° C.

RESULTS

1. Water and Saline Suspensions. — This series of 34 experiments in
water shows an average Qo* of 2.5 during the relatively steady period
five hours after the start of readings (Table I). This is in good

TABLE I

(?o2 in Water Suspension

Later

value

Initial value Qoo

Hours

Qo2

8.1 (4) *

5

3.5

9.3 (4).

5

3.3

7.2 (4).

4.5

2.4

7.5 (4)

4 +

2.5

8.5 (2)

4.5

1.8

5.3 (4)

5

2.1

5.8 (2)

4.5

2.0

5.1 (3).

3.5

2.2

4.7 (2).

6

1.2

6.1 (1).

5.5

3.0

8.0 (4)

5

2.3

? 1.4 (2)

5

0.8

? 1.8 (3)

3

1.2

* Bracketed number represents number of separate runs.

agreement with the value of 2.6 for the first, smaller series. The
temperature was slightly higher in this series but the average was

227

R. W. GERARD

taken at a somewhat later time. The initial Qo, average is 7.1, a
value depending on the conditions of preparation, as previously
discussed. Two experiments at 37° C. compared with 22° C. gave a
Qv\ of approximately 1.7.

Barren and Harrop (1929) found a marked effect of crowding on
the Qo? of white blood cells, the respiration per unit of material
falling as more and more concentrated suspensions were used. The
possibility of a similar effect in these experiments was tested by
varying the density of the suspension from the start or by suddenly
diluting a water suspension during a run by tipping in more water.
No change was observed, so that this factor may be excluded. It
might well depend, in the case of the blood cells, on limitation of
motility of the undulating membrane in the heavier suspensions.

The effect of NaCl in concentrations up to 1.1 per cent was deter-
mined, as a control of osmotic and ion effects with other reagents.
The lower concentrations wrere without effect, and even the highest
used gave a maximum decrease of 20 per cent of the Qo< as compared
with water, and this only occasionally. Further addition of one part
to ten of M/15 phosphate buffer was likewise of no consequence for
the oxygen consumption; nor wras the pH, between 7 and 8. \Vhen
salt solutions were tipped into a water suspension of respiring bacteria,
a small brief acceleration was often observed, after which the original
curve was resumed (Fig. 3). \Yhether this is a true momentary
"stimulation" or some small experimental error has not been further
investigated; in either case the results serve as a control for the
observed changes when other substances are tipped in.

2. Methylene Blue. — Addition of methylene blue (0.2 per cent or
0.1 per cent methylene blue "for vital staining") to a water or buffered
saline suspension always raised the Qo, by about one hundred per cent;
from an average value of 1.9 to one of 3.9 in seven experiments, an
increase of two. The increased oxygen consumption tends to fall off
with time but remains above that of water for several hours. The
effect of methylene blue plus glucose, sodium lactate, or sodium
cyanide will be considered with those substances.

3. Glucose. — The effect of glucose addition has been followed
somewhat further than in the preceding paper, by tipping a glucose
solution into a water suspension of the cocci during a run. This
makes it possible to obtain the entire glucose effect, since no time is
lost after its addition, during which readings cannot be made. Table
II sums tip all the results with plain glucose. The average Qo2 of 2.2
before tipping in the sugar rose to 10.5 for the first period (one-halt
to one hour) after tipping and then fell. At three and one-half

METABOLISM OF SARCINA LUTRA

229

TABLE II

Respiration in water
before addition

Added

Maximum

after addition

Later value

Glucose

Mi'thylenc
blue

<2o2

Per cent

<3o2

Hours

<3o2

*/2.3

.5
.5

8.8
0.4

0.2
0.8
0.8

0.2
0.2
0.8
0.8

+

+

+
+

13.6
10.4

15.0
9.7

9.3
14.6
9.2

19.2
10.0
18.2
12.2

6
6

5
4.5

2
2
2

2 (18)
2 (18)
2 (18)
2 (18)

1.3
4.1

4.5
4.5

8.5
14.4
9.7

4.9 (0.5)
4.4 (1.2)
13.2 (0.3)
8.2 (0.5)

<
\2.5

2.1

2.0.

f 1-9

\ 2.5

1 2.3

(2.2.

2.3

/
] 2.3

I 2 3

* Parallel runs.

hours the average Qo« was six. The maximum reached was inde-
pendent of glucose concentration (though 8.8 per cent glucose gave
the greatest increase observed), but the subsequent fall was somewhat
slower with larger amounts than smaller.

The total excess oxygen consumed by the bacteria due to the
addition of glucose can be determined by subtracting the <2o2 curve
in water from that with glucose added. In at least two cases this
has been definitely greater than the amount required to completely
oxidize all the added glucose, which suggests a "specific dynamic
action" of this substance. Confirmatory evidence is obtained from
the R.Q. findings, discussed previously, and from the further observa-
tion that suspensions carried for 24 hours in glucose solutions have a
lower Qo> than similar ones in water, as if the higher rates at first
had more completely exhausted the cell reserves. Evidence of a
stimulating action of glucose in nitrogen has also been previously
presented. All these observations indicate a dynamic action of
glucose though they are hardly extensive enough to permit sweeping
conclusions.

Methylene blue plus glucose leads to a greater initial increase of
oxygen consumption than the sum of their separate effects, but this
excess becomes less with time (Fig. 1). Thus, in four parallel experi-

16

2.N)

R. \V. GERARD

ments, the average maximum in glucose was 11.2, in glucose plus
0.2 per cent methylene blue, 16.5. Seven hours later the values were
4.7 and 5.S, and the next morning, in all cases followed, the values
were higher in the suspensions in glucose without the dye. This
diminishing or reversing of the methylene blue increase with time is

similar to the results with water solutions of the dye or of glucose
alone. The absolute increase of @o» when methylene blue is added
to water suspensions is 2.0, but the percentage increase is about one
hundred. When the dye is added to glucose suspensions, the absolute
increase is over 5, the percentage increase fifty. The interpretation
comes to mind that these two substances act at successive steps in the
chain of oxidative reactions, thus leading to a product of their separate
effects, rather than concomitantly, which should give only the sum.
Much evidence exists that this is indeed the situation. The recent
findings of Warburg, Kubowitz and Christian (1930) and Wenclel
1(>30) indicate that reduction of methylene blue may lead to peroxide
tormation with cytochrome, which in turn oxidizes other substances
(as lactic acid to pyruvic acid), and leaves the iron in the Fe+" f stage.
The second reaction might well depend, if only indirectly, on the
amount of oxidizable material (glucose) present. The findings with
lartate additions, however, are not entirely in harmony with this view.
indium /.delate. — The greatest rates of oxygen consumption
obtained resulted from the addition of lactate to water or saline
suspensions. I he maximum Qo2 after tipping averaged over ten
times that ju>t before, and a more than twenty-fold increase has been
observed. Though the maximum Qo2 varied in different runs from

Ml' r.\!',< H. ISM i >K SAKi IXA [.I I I A

231

17 to over 50, the variation was low in any one run and not related
to the concentration of lactate added (Fig. 2). (In one experiment a
regular rise in maximum with concentration was observed.) The
subsequent course of the extra oxygen consumption, on the other

Added:
0.1% No. Lactate

hand, was intimately dependent on this concentration (Figs. 2 and 3,
Table III). Thus, the average of a series of twelve experiments
gave a QQ., of 2 in water, and a maximum of 23 after sodium lactate
was tipped. Five hours later the Qon values for various concentra-
tions were:

0.05% = 1.4; 0.1% == 2.5; 0.25% == 3.0; 0.55% ---- 6.0; 1.1% -- -- 7.0.

Such a falling off with time, of the lactate effect, is due in part to
the actual removal of lactate, but cannot be accounted for entirely
on this basis. A few hours after the addition of 1.0 per cent lactate
ions, the Qo2 will have fallen to, say, one-half of the maximum. It
can be inferred (see below) with considerable certainty that over
0.5 per cent lactate is still present. It has been found, however,
that when 0.5 per cent lactate is freshly added to a water suspension,
the Qo, reaches the same maximum as when 1.0 per cent is added;
so that the falling off in the first case must be complicated by other
factors. An occasional finding, which also indicates the complexity
of the system, is an immediate increase of respiration rate after lactate

232

R. W. GERARD

(or glucose) addition followed by a further rise for an hour or more
before the usual fall sets in. This appeared oftener with the more
concentrated additions, as if the strong lactate partly inhibited
respiration until some was metabolized away.

The amount of lactic acid actually burned may be estimated from
the excess oxygen consumed, on the assumption that just this extra
oxygen is all used to completely oxidize the acid. In most experi-
ments 0.2 cc. of a 0.2 per cent solution of lactate (as the sodium salt)

TABLE III

Added

Maxi-

before addition

Sodium
lactate

Other substance

after
addition

Later >

ralue

Qo,

Per cent

Go2

1 lours

Qo,

f35

.2

McthvK-ne blue

48

7

1.5

-1 4 3

.2

Methylene blue

52

7

0.5

U.5

.2

65

7

3.0

f2.6.

.2

Methylene blue

17

6

1.1

J2.5

.2

20

6

1.9

f 1 8

2.2

In water

17

4.5

8.3

J

114

1.1
1.4

In NaCl, I'()4 buffer, pi 17
Same, pi 18

18

4.5

4.5

7.4
8.2

J2 2

2.2

29

5

5.5

12 0

0.5

28

5

3.0

v -•

was tipped into 0.2 cc. of the bacterial suspension, giving a final
concentration of 0.1 per cent. This quantity of lactic acid, 0.0044
millimols, would require 298 cu. mm. of oxygen for kill oxidization.
The values actually obtained for the extra oxygen consumed (the area
under the Qo2 curve with lactate present less that with lactate absent
—the two curves usually joining some hours after the start) were:

210, 250, 200, 270, ,UO.

The average, 250, is somewhat less (15 per cent) than theoretical
tor the complete burning of all the lactate. This is probably a techni-
cal error due to failure to rinse all the fluid from the side bulb into the
main chamber, with the consequent exclusion of some ol the lactate
from the bacteria. I'nfortunately, the quantitative possibilities were
not in mind when the experiments were carried out, and the usual to
and fro pouring \\as ordinarily omitted because of the danger of

METABOLISM OF SARCINA LUTEA

233

spilling alkali — a narrow opening to the side-bulb made rather severe
tapping necessary in some cases. In general, the agreement is good
enough to indicate complete oxidation of small amounts of lactic acid.

With larger quantities, that is, more concentrated solutions,
oxidation was not complete even in twenty-four hours — when the
extra oxygen consumption had ended. Thus for a 1.0 per cent
solution 1500 cu. mm. of oxygen would have been required for full
oxidation. The excess oxygen totalled over 700 cu. mm., and the
total oxygen used (i.e., the lactate Qo, curve not corrected by the
water control) after the lactate addition only about 900 cubic milli-
meters.

The possibility at once suggests itself that the falling off of the
excess respiration, and the failure of complete oxidation of the larger
amounts of lactic acid are alike due to the accumulation of injurious
reaction products. A rise in pH, in particular, must result from the
conversion of lactate ion into CO2, two-thirds of which would leave
the bacterial suspension and be absorbed in strong alkali in the inset
of the vessel, while one-third would remain to combine with the free
Na+ ions, giving NaHCO3 in place of NaC3H5O3. Such a pH change
is not, however, a critical factor; for suspensions buffered at pH 7 or
8 show the same Qo* curve with added buffered sodium lactate as do

Kdded:

!•/% Na. Lactate pHdto fi/aCI
1-1% " " pH7 " "
1.1% •• " H7 "H20

Hours

water suspensions (Fig. 3, Table III). Also, in one series, a phosphate
buffer at pH 5.9 was added to the suspensions in 1/10 and 1/5 concen-
trations. After a long run with much added lactate, the 1/10 buffer
was exhausted and the suspension alkaline to litmus, while that with
the stronger buffer was still faintly acid. The Qo2 curves, however,
were almost perfect duplicates. Further, it is^doubtful if any kind
of inhibiting end-products can be involved. If such were the case,

234

R. \V. GERARD

dilution of a suspension which has had lactate added some time before

should cause the fallen Qo2 to again rise to some extent. Actually the

reverse is found; dilution promptly lowers the QQ, still further (Fig. 2).

Finally, it is interesting to note the effect of a double addition of

A suspension in 0.3 per cent lactate at an early

stage (1- ! hours) and with a markedly increased respiration is in no

way affected by the further addition of one-third its volume of 0.6

Added to:

&-o67o
l-0.2.7*NaL

at once
O2%A/aL-olonce

Hours

per cent lactate. There is no immediate increased oxygen consump-
tion nor any delay in the subsequent fall. If the second portion of
lactate is added at the end of nine hours, when the original lactate
action has ended and respiration is the same as in a water control,
there is some increased consumption but much less than in the control.
In one experiment, for example, the Qo, in both water and lactate
suspensions had fallen to 1.5. When new lactate was added to the
lactate suspension, this value rose only to 3, while the Qo, of the
\\ater suspension rose to eighteen. If the lactate addition is delayed
another 12 hours, however, the water and lactate suspensions being
shaken in the usual manner in manometers, the difference is much
marked. Thus, in the same set of runs, lactate added to the
water -ii-pen^ion 21 hours after the start again gave a Qo, of 18, and
added to i he lactate suspension, one of 12 instead of three. ( )1>\ iously,
a se< mid laet.iie addition has no effect when added soon after a first
one, some eilert \\hen added after the initial increase in respiration is
past, and close i<> a maximal effect when added the following day.

When the experiment is carried out with 0.1 per cent lactate and
0.2 per cent lactate added later, the time relations are altered. With

MKTABOL1SM OF SARCINA LUTEA 235

these dilutions, a second addition has some effect even two hours
after the first. Thus, in one case (Fig. 4), the fall in respiration was
delayed an hour or more and some actual increase may have been
present at once after the second addition. It is interesting to note
further that the actual Qo2 at the moment the addition was made
was the same as that of the suspension in 0.3 per cent lactate where a
further addition was ineffective. After nine hours, four hours after
the respiration of the lactate suspension had become identical with
that of the water control, the further addition of 0.2 per cent lactate
to each gave the following Qo2 values: water suspension, 19; lactate
suspension, fourteen.

These observations strongly suggest that the falling off in rate of
oxygen consumption with time is due less to accumulation of some
retarding factor than to a temporary exhaustion of some required one.
Certainly neither oxygen nor lactate are lacking, and no retarding
substance accumulates as the ultimate product. Either some inter-
mediate product is formed rapidly and but slowly further altered,
so as to act as a temporary obstacle to the first reaction; or the oxi-
dative mechanisms of the cell are somehow run down by an excessive
load and are only gradually restored to the equilibrium state. The
effects of sudden dilution are more in accord with the second possibility.

The effect of methylene blue combined with lactate is of interest
in this connection. When, in comparable experiments, buffered
solutions of lactate alone or with methylene blue are tipped into a
water suspension, the maximum Qo2 is consistently greater with the
dye absent (Table III). In eight experiments with 0.1 per cent
sodium lactate and 0.1 per cent methylene blue the maximum Qo2
averaged 34; in four without the methylene blue, forty. In every
run the depressant action of the dye was. seen, decreases ranging from
10 to 30 per cent. At two to three hours the curves approached or
even crossed for a short time, but at about six hours after the tipping
the values were: with methylene blue, Qo* : : 0.9, without, Qo, -- • 2.5.

Methylene blue has, to sum up, a marked early accelerating
action on the respiration of Sarcina suspensions in water or in glucose
solutions. After some hours the rate of oxygen consumption falls to
that of control suspensions and is ultimately depressed. With lactate
solutions, the depressant action of methylene blue is present at the
start, and, in percentage, becomes more marked with the passage
of time.

It may be pointed out that for equimolar amounts of glucose and
sodium lactate added, the time course of extra oxygen consumption
is very different. The glucose leads to a moderate prolonged rise,

236

R. \V. GERARD

the lactate to an intense brief one (Fig. 2). This would be in harmony
\\ith, though of course not necessitating, a slow split of the glucose
molecule to lactic acid or a closely related intermediate followed by a
rapid oxidation of this latter. Barren's view (1930) that methylene
blue catalyzes primarily the initial splitting of glucose and not the
later reactions would then fit the observed results fairly well. The
main difficulty with such an interpretation lies in the failure to obtain
an accumulation of lactic, pyruvic or other acid stronger than carbonic
under anaerobic conditions, as shown in the preceding paper.

Finally, simultaneous addition of glucose and sodium lactate leads
to the sharp maximum of lactate followed by the less intense enduring
glucose effect. There does not appear to be summation of the two
effects — the Qo, follows the higher single curve (lactate alone or glucose
alone) at any time. No separate experiments were performed with
optically active lactic acid, the racemate serving in these tests. Since,
however, all added lactate can be burned and the Qo, curve after
lactate addition gives no evidence of a discontinuity, it seems probable
that both the d and / forms are easily utilized by this organism.

5. Thioglycollic Acid. — A few experiments, made in another con-
nection, have shown that this substance in 0.2M concentration
doubles the oxygen consumption.

6. Sodium Cyanide. — Cyanide, in concentrations up to N/1QQ or a
little stronger does not inhibit the oxygen consumption of Sarcina
lutea. This is true for the usual respiration of suspensions in water,
saline, or phosphate buffer, as well as for the increases evoked by
methylene blue, lactate, or glucose (Fig. 5). The presence of cyanide

3
2

Hours

2. 3 4 J
FIG. 5.

fi

8

api in fact, i<> slightly increase the Qo, or at least to diminish

the I.IK- <it tall with time. The average Qot of five experiments in
water was 1.1 .it tlin-r hours; in A'/ 100 Xadi it was 1.4. The possi-
bility of IK'N distilling from the side bulb to the main suspension

METABOLISM OF SARCINA LUTEA 237

prematurely, and so masking a real effect, is excluded. In N/30
cyanide, respiration is depressed 1/2 to 2/3, and in N/1Q, 2/3 to 3/4.
Rubinstein (1931) has confirmed the absence of cyanide inhibition
with water suspensions.1

A much less complete absence of cyanide inhibition has been
described by Emerson (1927) for the alga Chlorella, for in the presence
of sugar 10~4 M cyanide gives 50 per cent inhibition. Lund (1918)
has claimed a cyanide insensitivity for Paramecium, though this has
not passed unchallenged (Hyman, 1919).2 Pitts (unpublished) has
found Colpidinm to be but slightly sensitive to cyanide. Burnet
(1927), studying growth and oxygen consumption of a number of
bacterial types under the influence of cyanide, found them divisible
into two groups. Most showed a usual degree of sensitivity, but
some, as streptococci, were insensitive to cyanide. These groups were
also different in their behavior with hydrogen peroxide. Even for
vertebrate tissues a complete inhibition of respiration by cyanide is
not the rule. -/V/100 does not inhibit the respiration of frog nerve
by more than 80 per cent (Gerard, 1930), and Dixon and Elliot
(1929) obtained similar results with a variety of mammalian tissues,
though Alt (1930) and Warburg (1930) sharply criticize their work.
The objections raised certainly do not apply to the present results.
Sarcina lutea apparently carried on a fairly typical aerobic metabolism,
using oxygen freely to burn the usual organic molecules. The almost
complete insensitivity of its respiratory catalytic system to cyanide
would seem to preclude a too broad generalization as to the ferroactive
nature of oxidative enzymes.3 The action of carbon monoxide is
likewise atypical.

7. Carbon Monoxide. — Carbon monoxide containing 5 per cent

1 Working with descendants of the original strain, over a year later, Barren ob-
tained definite inhibition of respiration by cyanide. I repeated my original experi-
ments, using the identical techniques previously employed, and also found inhibition
was now present. In water suspensions TV/100 cyanide gave depressions of 50 per
cent or more, in glucose even greater ones. The depressions were, however, of short
duration and respiration returned to normal in an hour and sometimes became ex-
cessive later. Presumably the organisms had altered in the interim due, perhaps, to
unknown differences in the culture conditions.

- This has now been confirmed. Gerard and Hyman, Am. Jour. Physiol., in
press.

3 It may be mentioned here that Dr. E. S. G. Barren and I have demonstrated by
means of the spectroscope that cytochrome is present in these cells. Further, it is
changed easily between the oxidized and reduced states by bubbling in oxygen or
nitrogen. In the presence of ./V/100 KCN, however, the pigment remains reduced
even in a stream of oxygen. This appears as evidence that cytochrome is not an
essential link in the chain of respiratory substances of this cell, since respiration
proceeds as normally after its oxidation is blocked. We are further investigating the
oxidizing mechanisms of Sarcina.

238

R. W. GERARD

oxygen was used in all experiments and the thermostat covered to
exclude daylight. A 200-watt tungsten bulb, immersed in the white-
walled thermostat, within 2 to 10 inches of the manometer chambers,
was turned on at intervals of 20 minutes to 1 hour.

\Vaier suspensions of Sarcina respired alike in air and in the
monoxide mixture. Light had no obvious effect on either. When
-lurose (1.0 per cent) or lactate (1.0 per cent) was added to such
\\ater suspensions, an inhibiting action of the CO appeared. In the
case of glucose this was slight, the increased oxygen consumption in
CO being over three-fourths that in air. \Yith lactate, the inhibition
was marked, the increase in monoxide being less than one-fifth that
in air, but inadequate oxygen may have contributed to the effect.
In both cases, glucose or lactate added, when the carbon monoxide

12.

Full line • dark

, NoL-CQ rep/need

Dotted line* light-

^^O

10

'

(Added NaL

8

Y or Glucose

•

^Oj.

CO +

NaL

6

(air + Glucose

Glucose -CO replaced

IH20-

in air

%,

^^^ bf air

0 orCO -4curv es

_

~~~~..^

^•s,

4

^/^

^vr--

.^^

^^^_/

2

^

^CO-t- Glucose

2 Curves

10 40

60 80 100
FIG. 6.

/60

mixture was later replaced by air, the full rate of oxygen consumption
promptly appeared. The presence of light, of the intensity used, had
no reversing action on the inhibited lactate oxidation. There was a
lion of such an effect in increasing the respiration of a suspension
in glucose in the presence of CO, but an entirely similar augmentation
appeared also in air (Fig. 6). Rubenstein (1931) has found that light
increases the respiration of Sarcina lutea at this temperature.

DISCUSSION

These results raise several interesting questions regarding the
chemical dynamics involved. Qu, is a measure of reaction velocity,
yet this is independent of the concentration of the primary reactants

MKTABOUSM OF SARCINA LUTEA 239

over a wide range. Thus 0.05 and 1.1 per cent lactate give the same
maximal rates, but both rates decline following the maximum. For
the lower concentrations, at least, the rate must fall with decreasing
amount of substrate. Probably at sufficiently low initial concentra-
tions, the maximum reached would vary with concentration, until an
asymptotic value had been reached, beyond which concentration has
no effect. The critical concentration of lactate cannot be over 0.05
per cent, but may be very close to this. The decline in rate after the
maximum, when considerably larger amounts of lactate are present,
cannot, as in the previous case, be due to diminution in concentration
of the lactate, for this must remain for long periods above the critical
value. Evidence has also been presented to show that accumulation
of end products is not the important factor leading to slowing, though
temporary piling up of intermediates might play a role. Since, in
the simplified reaction:

Substrate + O2(+ Respiratory Catalyst)

-> CO2 + H2O + End Products,

the velocity appears to be independent of substrate, oxygen, or end
product concentration, its fall might be attributed to interference
with the respiration-catalyzing system. This might mean destruction
or out-diffusion (suggested by the fall in Qo, on dilution) from the
cell of one of the key substances in the respiratory chain more rapidly
than it is replaced.

The situation with respect to glucose is analogous though quanti-
tatively different; and the ultimate depressant action of methylene
blue can also be laid at the door of an injured catalytic system.

Meyerhof (1912) has shown that the oxygen consumption of
acetone-yeast is doubled by the addition of methylene blue, whereas
that of the intact cells is depressed by it; and similar results were
obtained with staphylococci (1917). Barren (1930) likewise has
found little or no stimulating action of methylene blue on normal
tissues, though it does increase the respiration of cancer and other
aerobically glycolysing material. Gerard (1930) has found that
medullated nerve, on the contrary, though lacking aerobic glycolysis,
is stimulated by methylene blue to a marked respiratory increase,
and Chang (unpublished) has confirmed this for non-medullated nerve
and with cresyl blue. Sarcina in this respect resembles nerve. It
may be noted in passing that methylene blue largely or entirely
reverses the cyanide inhibition of respiration in the case of staphylo-
cocci (Meyerhof) and nerve (Gerard, Chang). There is no evidence
that the yellow pigment of Sarcina acts in a similar way, though such
activity might contribute to the cyanide insensitivity of this organism.

240 R. \V. GERARD

SUMMARY

The <2o: of washed Sarcina lutea in water suspension at 22° C. in
the relatively steady state averages 2.5. There is no untoward
effect of crowding in heavy suspensions. The presence of NaCl does
not modify the Qo, until M/5 concentrations are reached, when a
slight depression may result. The Qo, is also unaffected by pH,
at least between 7 and 8, or by the presence of phosphate buffer
mixtures.

( ilucose addition causes a marked increase in the Qo,, the maximum
being largely independent of glucose concentration. The extra oxygen
used under the influence of glucose may be more than that required
to fully oxidize it which, with other evidence, suggests a "specific
dynamic action."

Sodium lactate may increase the Qo, in water suspension over
twenty-fold. Its addition is always followed by a great rise of the
respiratory rate, the maximum reached being independent of lactate
concentration, at least between 0.05 and 2.0 per cent. The respiration
falls rapidly back to normal after addition of small amounts, and the
extra oxygen consumed accounts for full oxidation of the added lactate,
both d and / forms. With larger concentrations of lactate, the in-
creased respiration also falls after the initial maximum, but more
slowly. This fall is not primarily due to removal of lactate nor
accumulation of end products.

Methylene blue added to a suspension in water doubles respiration
at first, later depresses it. Added to one in glucose solution the same
sequence appears. The Qo, in glucose is increased only 50 per cent
by the dye, but this is an absolute increase over twice that in water;
so that glucose and methylene blue added together to a water sus-
pension cause a greater increase in respiration than the sum of their
separate effects. The dye added to lactate solutions seems to be
depressant from the start.

Thioglycollic acid doubles the respiration of a buffered water
suspension.

indium cyanide causes no inhibition of respiration up to concen-
tration-, of .1/100 or somewhat stronger. This is true for the low
respiration ot suspensions in water, saline, or phosphate buffer and
lor the increases evoked l>y methylene Mne, lactate and glucose.
I ctremelj strong < \.mide concentr.it ions do depress, but even .1/10
Ja< < doe- 111 >t .il.oli>h more than 2/3 to 3/4 of the total respiration.

( .ilium monoxide containing 5 per cent oxygen has no effect on
the respiration ot a \\ater suspension, but somewhat inhibits the

METABOLISM OF SARCINA LUTEA 241

increase in glucose and largely that in lactate. Light has little if
any effect on the inhibition.

The metabolism of Sarcina lutea is compared with that of other
cell types.

My thanks are due Miss Ruth Bilger for valuable technical

assistance.

BIBLIOGRAPHY

ALT, H. L., 1930. Biochem. Zeitschr., 221: 498.
BARRON, E. S. G., 1930. Jour. Exper. Med., 52: 447.
BARRON, E. S. G., AND G. A. HARROP, 1929. Jour. Biol. Chem., 84: 89.
BURNET, F. M., 1927. Jour. Path, and Bacterial., 30: 21.
DIXON, M., AND K. A. C. ELLIOT, 1929. Biochem. Jour., 23: 812.
EMERSON, R., 1927. Jour. Gen. Physiol., 10: 469.
GERARD, R. W.f 1927. Am. Jour. Physiol., 82: 381.
GERARD, R. W., 1930. Proc. Soc. Exper. Biol. and Med., 27: 1052.
HOLMES, E. G., AND R. W. GERARD, 1929. Biochem. Jour., 23: 738.
HYMAN, L. A., 1919. Am. Jour. Physiol., 48: 340.
LUND, E. J., 1918. Am. Jour. Physiol., 45: 365.
MEYERHOF, O., 1912. Pfliiger's Arch., 149: 250.
MEYERHOF, O., 1917. Pfliiger's Arch., 169: 87.

MEYERHOF, O., LOHMANN, K., AND R. MEIER, 1925. Biochem. Zeitschr., 157: 459.
RUBINSTEIN, B. B., 1931. Unpublished.
WARBURG, O., 1919. Biochem. Zeitschr., 100: 268.

WARBURG, O., KUBOWITZ, F., AND W. CHRISTIAN, 1930. Biochem. Zeitschr., 221:
494.

THE OXYGEN TENSION-OXYGEN CONSUMPTION CURVE
OF UNFERTILIZED ARBACTA EGGS

PEI-SUNG TANG1
(From the Marine Biological Laboratory, Woods Hole, Massachusetts)

Gerard (1931) has developed equations relating the external oxygen
pressure to the oxygen consumption, cell diameter and diffusion
constants for individual cells. This work was undertaken to obtain
experimental data under favorable conditions for comparison with the
theoretical derivation. For this the sea urchin egg is beautifully
adapted, as it is homogeneously spherical, regular, and its resting rate
of oxygen consumption can be suddenly increased by fertilization.
Its effective diameter can also be varied by allowing a greater or lesser
amount of the egg jelly to remain about each cell. It is our intention
to investigate the effect of varying these several factors, but for the
present the results with simple, clean, unfertilized eggs are to be

presented.

METHODS

Oxygen consumption was determined by the usual XVarburg
technique, using small conical vessels of about three cc. capacity, con-
taining an inset for alkali. The experiments were conducted at
24.7° C. and the shaking of the manometers was demonstrated to be
adequate for diffusion equilibrium even at the low oxygen pressures
used. Gas mixtures were prepared by mixing nitrogen and air over
water and, for the smaller concentrations, by direct mixing in the
manometer as described by Gerard and Falk (1931). \Yhen the
mixing was done over water, the composition was checked by analysis
in the Haldane apparatus. Passing the gas through the manometers
with gentle shaking for fifteen minutes sufficed to bring the fluid into
complete equilibrium with it.

The Arbacia eggs were obtained by picking out the gonads of
about twenty animals, allowing them to shed into sea water, and
straining through cheese cloth. The eggs were allowed to settle and
the Hiprrnatant water decanted, giving a heavy suspension. They
wen- not tint her washed, but controls showed no difference in respira-
tion betueen \\ashed and unwashed suspensions, and fertilization tests

1 It is .1 |ilr.!Mirr to acknowledge the guidance, help and suggestions of Professors
R. S. I.illi:- .ni(l \\. \\ . ( icrard, under whose supervision this work was done. To Ur.
Walter S. Root my ili,mL-> are due for many of the gas analyses.

242

OXYGKN TKNSION ( )\V( iKN CONSl'M I'TK )N CURVK 243

yielded over 90 per cent cleavage. The concentration of eggs in the
suspension was determined by means of a hemocytometer, and half
a cc. was used for each manometer. The egg diameters, as determined
by an ocular micrometer, averaged 77 micra, variation 72 to 80.

Significant readings were made for three and a half to four hours
and the viability of the eggs again checked by fertilization. In each
experiment parallel runs were made of the respiration in air or oxygen
and at the lower oxygen pressures, so that the latter values could be
expressed as a direct percentage of the former.

TABLE I

Oxygen Tension
mm. of Hg

Oxygen Consumption
per cent in air

No. of Experiments

20

43.0

13

6.1

54

7

9.9

69

2

20.5

92

3

42.0

92

5

86.0

101

3

160 (air)

100 (0o, = 33.6)

24

760

100

3

RESULTS

The average <2o2, expressed as cubic millimeters of oxygen consumed
per million eggs per hour, was 33.6 in air (extremes 17 to 51; 24 experi-
ments), the lower values being obtained toward the end of the breeding
season. At oxygen tensions of 70-80 mm. or higher, up to pure
oxygen, the consumption was the same as in air. Below 70-80 mm.
the consumption fell off with pressure, as indicated in Table I. Be-
ginning at the oxygen tension of 20 mm., the oxygen consumption
fell very rapidly with a decrease in oxygen tension. Surprisingly,
the same type of curve, with the turning point coming at the same
oxygen tension has been obtained by Amberson (1928), on fertilized
eggs by a different method.

The oxygen consumption was found to proceed at a constant rate
up to about four and one-half hours, at which time there was a sudden
increase in oxygen consumption, both in air and in lower concentrations
of oxygen. This seems to be related to a fairly abrupt cytolysis of
large numbers of the cells. A few experiments were carried out with
fertilized sea urchin eggs to determine the possible existence of a
rhythm of oxygen consumption analogous to the rhythm of production
of CO2 reported by Lyon (1904). The experiments showed the

244

PEI-SUNG TAXG

fertilized eggs to have a respiration five times that of the unfertilized
ones, confirming Warburg (1908) and Loeb and \Vasteneys (1910);
but, within the limits of accuracy of the method, the rate of respiration
was constant through the first, second and third cleavage periods.

o

Hm.Hg. 20 40 60 80 100

Oxygen tension

FIG. 1.

120

140

160

BIBLIOGRAPHY

AMBERSON, \V. R, 1928. Biol. Bull, 55: 79.

GERARD, R. W., 1931. Biol. Bid!., 60: 245.

GERARD, R. \V., AND FALK, I. S., 1931. Biol. Bull., 60: 213.

LOEB, J., AND WASTENEYS, H., 1910. Biochem. Zeilschr., 28: 340.

LYON, E. P., 1904. Science, 19: 350.

WARBURG, O., 1908. Zeilschr. physiol. Chem., 57: 1.

WARBURG, O., 1915. Pfl tiger's Arch., 160: 324.

OXYGEN DIFFUSION INTO CELLS

R. \Y. GERARD
(From the Department of Physiology, University of Chicago]

Krogh (1919) determined the coefficients of diffusion of oxygen
through surviving frog muscle and through fascia by using these as
tissue membranes and measuring the gas diffusing through under
given conditions. He applied the measured values in a formula
relating capillary size and number to the oxygen needs of muscle
in vivo.

These constants have been utilized by subsequent workers for
substitution in other equations applying to the equilibrium state.
Warburg (1923) developed an equation relating the necessary oxygen
pressure to the metabolism and thickness of a slice of isolated tissue.
Fenn (1927) and Gerard (1927) independently presented analogous
equations for a cylinder of tissue, and A. V. Hill (1928) has extensively
surveyed the general question of diffusion in tissues and developed
several formulae for the attaining of equilibrium as well as for the
equilibrium state. Applying these equations to the specific question
of diffusion of phosphate ion (Stella, 1928) and lactate ion (Eggleton,
Eggleton and Hill, 1928), the values for the diffusion constants
obtained suggested that these ions moved mainly in the intercellulai
fluid and with much greater difficulty across cell boundaries.

Should a similar situation obtain for oxygen, the calculations
based on Krogh 's constant might be considerably in error, and in
any case, since this value has been so widely used, any further checks
on its magnitude should be of value.

The simplest situation for studying oxygen penetration into living
cells is offered, of course, by the unicellular organisms, which present,
mainly, the problem of the sphere. It was the original intent of this
work to evaluate the diffusion constant for cells as the only unknown
in a simple equation, but it soon became apparent that a more searching
examination of the assumptions made in such a simplified derivation
was necessary. The elementary equation for diffusion into a sphere
is easily derived as follows:

At equilibrium, the amount of oxygen diffusing into a mass of
tissue must equal the amount consumed by it in the same time.
For unit time, the inward diffusion depends on the surface area
(47TT2 for the sphere); the diffusion coefficient, D; and the gradient

17 245

246 R. W. GERARD

of oxygen concentration along a radius, dC dr, C being the oxygen
concentration at radius r. The consumption of oxygen by a sphere
of indeterminate radius, r, is the product of the measured consumption

/ mass \

per unit volume - = mass X 0.9^ for most protoplasm , A,

\ sp. gr.

I
times the volume, - irrx. At equilibrium, then

<J

dr 3
or

dC .!/•

which gives on integration

A r~

C - - -\- K
6D

At r • - 0, C - K and X therefore represents the oxygen concen-
tration at the center of the sphere.

In the case of most tissues so far studied, the oxygen consumption,
A, has been found to be independent of the oxygen pressure over a
wide range. The assumption has been made in this derivation that
consumption in any one region is independent of oxygen concentration,
provided this is greater than zero. Then for the critical condition
where the oxygen concentration just reaches zero at the center of the
sphere, I\ - 0, and the oxygen concentration at any level in the sphere
is

C_A*

and at the surface, r •• - r0, the critical concentration of oxygen needed
to just insure a supply at the center is

Co = — (2)

6D

and

(3)

.1 may be fairly simply determined with one of the available
meilidd> <>l mea-nring oxygen consumption, and Gi is that concen-
n. n ion of oxygen in the medium which just gives full \alues of oxygen
consumption, \\ith lower concentrations the observed oxygen con-
sumption begins to fall oil. If the radius of the cell is known, a simple

OXYGEN DIFFUSION INTO CELLS 247

means is available for obtaining the oxygen diffusion coefficient, or
for bringing out the invalidity of assumptions as to the conditions
of respiration.

It may be stated at once that the results of applying this simple
equation to a variety of cells, as compared with results experimentally
obtained, show such great variation as to indicate that other factors
than the simple diffusion constants must be involved. It is necessary,
then, to examine in some detail the assumptions made in the above
derivation. Probably the most important is that just expressed,
namely that the oxygen consumption of any region is independent
of the concentration of oxygen at that region. This will be examined
later. Other assumptions of little probable importance in the case
of tissues may become extremely significant in the case of the individual
cell. These are, first, that the oxygen consumption per individual
region is a constant throughout the cell, that is, that the total con-
sumption divided by the total volume gives the true consumption of
each region. This is certainly not necessarily true, as the rate of
respiration of a nuclear region or a cortical region might clearly be
very much greater than that in other regions. It is also implicitly
assumed that the diffusion constant, whether equal to that of Krogh
or not, is at least constant throughout the cell. This also is not at
all certain; in fact, in view of the previous discussion, it is entirely
possible that the constant would vary widely across the membrane
and in the cell interior.1

The mathematical development which follows is designed first to
include all possible cases in which diffusion or consumption in two
concentric regions of the cell are not assumed to be equal; and, fol-
lowing that, the more difficult case of the dependence of consumption
on oxygen concentration will be considered. In this treatment I have
called freely upon the expert assistance, generously given, of Dr.
Walter Bartky of the Department of Astronomy at this University.
The final section of the paper will apply the equations derived to the
data available in the literature as well as to our own measurements,
in an attempt to evaluate the conditions actually existing as regards
respiration. Obviously, the derivations apply equally to other dif-
fusable substances formed in and leaving or entering and utilized by
the cell.

1 The further complication, in many cells, of protoplasmic streaming cannot be
handled rigorously. It is probably not significant in bacteria, yeast, or even Arbacia
eggs; and when it is present, the stirring action would act as a decrease in radius in
lowering the diffusion limit.

248 R. \Y. GERARD

1. Oxygen Consumption not Alike in All Portions of A Sphere
a. Limited to a central core (nuclear respiration). — Let Q -- the
total oxygen consumed by the cell; C • the oxygen concentration at
radius, r; ro --- radius at cell surface; /•' - radius of respiring region.
The oxygen diffusing into a sphere centered at the cell center and

of any radius, r, equals 4irrzD —r- . \Yhen r \ = r', all oxygen consumed
l>v the cell must pass through to the region of consumption, so:

4irr-D - - 0.
dr

The consumption per unit volume in the non-respiring portion of
the cell is zero; of the respiring portion:-

/3

When r < r', then, the oxygen passing the shell must equal the
volume within this shell times the consumption per unit volume, or

A ndC 4

4irr2D - - = -

Then

dC Q

dr 3 4 ,3
3"

dr 4irr2D '
dC rQ

dr r03 4vD '
Integrating:

when r ~= r', (4a)

when r ;= /. (4b)

For the critical condition, C - 0 when r • -- 0 and C > 0 for ,
r > 0, tlu-n KZ -'- 0; and since C is continuous at r --- r' ,

or

K
"

So that, from

c =

47T/ 2r' r

OXYGEN DIFFUSION INTO CELLS 249

At the cell surface r - r0 = " r' ,

(Cc — oxygen concentration needed at the surface, radius == r0, to
just insure full penetration, in this case when all the oxygen is con-
sumed within a sphere of radius r'.)

For the special case when r' - r0, or oxygen consumption is
uniform throughout the cell, this reduces to

r - Q

Co -

8irDr0
or, since A = consumption per unit volume,

A - Q

4 .'

3
and therefore

the original equation.
In general,

£ = ^-° - 2. (7)

Co r

It is readily seen from equation (7) that as r' decreases relative
to r0, the critical oxygen pressure needed, Cc, increases rapidly relative
to C0. When r' = ><r0, Cc ~- 46V when r' %r0, Cc 1060.

b. Oxygen consumption limited to an outer shell (cortical region}.
-By an entirely analogous derivation it can be shown for this case
that

v e v ' 0 ^ ' / v 0 ) / o \

~c = " ~ — ft- ' W

co r0- + r^r + r

From this equation it appears that as the respiring shell of thickness
(r0 -- r'} thins, -^decreases towards zero. When

r' = 0,=1; r' = &*, = 0.57; r' = 34>0f = 0.27;
Co Co Co

r' = 0.9ro. ^ = 0.10.

Co

c. Oxygen consumption present throughout cell, but more or less
intense in an outer shell. — -This will be considered in the general case
below. See Equations 17 and 18.

250 R. \V. GERARD

2. Diffusion Constant Different in Central and Cortical Parts of Cell

If the diffusion constant is smaller at one region than another—
as that of a relatively impermeable membrane is small compared to
that of a permeable cell interior — the following derivation becomes
necessary.

Let the diffusion coefficient be D' for the cell interior from r -- 0
to /• ~ r'\ and D for a superficial shell, r • - r' to r • - r0. Then, by a
familiar derivation,

,,.. , dC Ar

\\hen rSr, '

, dC 4 dC Ar

\\ hen r =± r', 4-rrr-D -j- - '

dr 3

Integrating between limits, and assuming C - 0 for r • - 0, and
C > 0 for r > 0,

67)"

'2

o - - r')
Co~

where C' is the concentration at r'. Combining,

_ .4 / ro2 - r'2 r'2\
Co • 6 V " ~D~ '&)'

which reduces to the simple ecjuation (2) when D ---- D' or r' -- r0.

It is obvious from (8) that if the cell membrane is thin compared
to the cell radius, that is, r' is nearly as large as r(), its diffusion constant,
D, must be very small in relation to that of the interior, /)', to affect
Co- For example, to double G> when r' -- 0.9;-n, D must be 0.167)';
and when /-' is 0.99r(), /) -- 0.027)'. To increase Co ten times: for
0.9r0) /> •••- 0.01 27>'; for r' ••-- ().09/-0, /) == 0.0027)'. Since it
appears from the work of the Hggletons and Hill (1928) that the
diffusion coefficient for ions across the cell surface may be one hun-
dredth as large as in the cell or intercellular fluid, this factor obviously
be of importance for oxygen diffusion. But unioni/ed molecules
to penetrate membranes much more easily than do ions, so
that tlic analogy may not hold.

In all the above derivations the critical oxygen pressure at the
cell sinlai e has been calculated. That is, assuming that the oxygen
consumption ot the \\lmlc cell remains at its maximum value, what is
the necess,n\ eMeinal prr-Mire as the various conditions within the

OXYGEN DIFFUSION INTO CFLLS 251

cell are changed? It is next desirable to investigate the oxygen
consumption-external oxygen pressure curve below this critical value
and to examine its shape for the simplest situation.

Let r\ represent the depth of oxygen penetration, i.e., C - 0 for
values of r such that r\ I = r =^ 0, and C > 0 for r > r\. And let C,
represent the oxygen concentration at the surface, r - - r<>.

Then, as the respiring volume - - ir(r3 -- r\x),

4
-- \

(IT O

dC A (r3 - -

c - A

l^s —

3D

r

Similarly, the oxygen consumed is:

rro

e = 4,J

Jr,

/-> Tc TT^T. /•?•!% / 1 < \

<2 = — T- (r° ~" ri/- ^ )

Take units such that r0 = • 1, and that Cs = C0 = 1 for ri = 0,
and also that Q = 1 = Q0 for n = 0. Note that with this selection

A 31

of units, -p: = 6, A = - - , D = —. The relation between Q and C

ix 4:1T

for the general case is then expressed in terms of r as a parameter:

From (10)

Ca = 1 -^(3 -- 2n),
and from (11)

(2 = 1- r,\

which give the following values (Table I). The corresponding curve
is given in Fig. 1.

Still more generally, assuming only that A is constant as regards C,
the following derivation includes all cases involving a single concentric
zoning of the cell as regards respiration and diffusion.

Let r' represent the surface separating two cell zones.

Let D' and A' represent the diffusion coefficient and unit oxygen
consumption for the nuclear region, r < r' ; and similarly D and A
represent these values for the cortical region, r > r'.

252

R. W. GERARD

TABLE I

c.

Co

ri
ro

0
Qo

1.00

0.0

1.00

0.90

0.2

0.99

0.65

0.4

0.94

0.50

0.5

0.87

0.35

0.6

0.78

0.22

0.7

0.66

0.10

0.8

0.49

0.04

0.9

0.27

0.00

1.0

0.00

Let r\ represent, as previously, the depth of oxygen penetration.
In the following it is assumed that r\ = = r' , for when r\ > r' the situa-
tion reduces to the simple cases previously considered, inasmuch as
no oxygen reaches the inner sphere where conditions are different.

1.0

0.8

0.6

Q

Qo

0.4

0 0.1 O.Z 0.3 0.4 0.3 0.0 07 0.8 0.9 16

Cs
Co

1 n,. 1.

Let C represent the oxygen concentration at r • -- r' \ Cs the oxygen
concentration at the cell surface, r -- r(}; and Cc the critical surface
concentration to just insure complete penetration. For the simple

, Cc --•- Co.
Then the derivation, along the previous lines, is:

(12o)

For ;• : /-', Irr, />' vA'(r* -- n3) ;

dr 3

r,

' + *A(r* - r). (126)

OXYGEN DIFFUSION INTO CELLS 253

Simplifying:

dC A' (r5 -- r:3)
For r <r', --

and integrating from ri to r':

A' / r'~ r,3 V2\

c = — i — 4- — - — l • r 1 n

3D' V 2 /•' 2 /

Similarly:

, dC A'(r'3 -

and integrating from r' to r0:

r r' A'(r'3 - rj3} / J. L | • A ( ^2 ^! ^'2

From (13) and (14)
A'(r'z - r

3D

/ 1 1\ A/"^2 ^ .? /A
\r'" rj ' 3D \ 2 ' r0 ' 2 Y )

Equation (15) may be explored as follows:
For TI = 0, that is, at the critical external oxygen pressure,

A'r'3 I 1 1 \ , A /;V r^ 3 /2\ , A'r

2 " rn ' 2 r

For ri = r' = 0

6D

For ri = 0, r' = r0

r ^4Vo2
6D'
For ri = r'

A I ?/2

Cs = . ±L ^o2 + — - 3r/:; ) . (See Equation 10.)

For r\ = r' - r0

Cs = 0.
For D = D'

C - (r* -

~

3D 2 ro

(A -A'y

1 + A /^! + ^ _
J 3D ( 2 r0 2

6D 6D

254 R. \V. GERARD

For /-, : = <), Cs -, /-o2 + — - 3/v . (See Equation 10.)

If ,1 = A', or /! - rf,

TQ

A - A'\ ,2/2r'

< '-

If A -A',

Co =
For^l -- A'

A |>02 -- r* ,/l 1 \] A /r'- r,s 3
r3Z)| ~2~ ril/"^/l "I^VT "P "2

2r!3

r<?

If D --•- D' or rl ••-- r',
I'or /-! = 0, Cs = - - ( r02 - — - Ir (See Equation 10.)

M) \ 6D

A / 2 '2

•" / ''O ^

+ ^-

1. 1 \'

(See Equation 9.)

if D = /r,

r

6

The e\i)rc'>sion for oxygen consumed per cell is:

Q -- 4wA' t r-dr + 4^4 f° rdr

t/r, «/r'

- /')]. (20)

l.iMc II and Fi.u. 2 i;ive the Q -- Cs curves obtained for various

r' D A

rrl.itioii^hi[» of , , ' , •

< onsider finally thr -ciu-ral case where AaF(C) and is not al\va\s
a constant, thoii;j,h approximately so for hiiji values of C\\. The

t y^*

amount of oxygen ci-o^sini; a spherical surface of radius r is 4irrD-j- ',

OXYCKN IMI-l-T'SION INTO CKLLS

255

therefore that consumed in a shell of internal radius == r, external
radius = r + dr, is

41 o ri ^ ^
TT { rL> -=-
dr

r+dr

ttf

This is equal to the consumption per unit volume times the
volume: A4irrzdr.

A is some function of C •- DF(C), where F(C) does not decrease
with increase in C, i.e., a monotonically increasing function of C.
Then

-r

dr

r+dr

-j-

dr

= DF(C)r2dr

or

dr

dr

= r*DF(C);

and, assuming D independent of C and r,

j i ->
d[ r2-r-
dr

~dr~

The total oxygen consumption of a cell is

X'o

(21)

(ro

F(Qr2dr.

(22)

Q as a function of Co is an experimentally observable relationship,
from which it is desired to obtain F(C). Consider the most general
case, that is, either C -- 0 for r •• -- ri, where r\ ^ 0, or C > 0 for all
values of r. For this latter, always take rt •- 0 in what follows.
Integrating (21) between r\ and r,

^= f
dr J TI

and integrating between r and r0,

r° i / r

C0--C- 4

Jr i \ J TI

or

rr° i / r \

C--C0- U r*F(C)dr)dr.

t/r * \tSrj /

(23)

256

R. W. GERARD

£Q

r r

^™* *^ ^^ ^^ _ ^ _ f^^ f^ f~^

x~s

_ _
II II
VQ

_ ^^ oo ^^ t*"* ^^* ^^ ^^* ^o if~~i ^^

^* ^« ^** ^^ <*""* ^^ ^^ ^^ 30 l^** '^^ ^i

-- fM_ ^ d d d c d d d

II II

VQ

" *-< 2^0' d o' o' d d d d o
•u

u

^^ r

>

r —

O^^"- •._;,_:_;_;,_;,_ -^_:

^^ *— ^^ f^ f^^ j^^ ?^p f^p f^ r^

— C^CCCCCCCC

1

z z
II II

^^ ~^ *" ^ 10 f*^ *^^ ^^j O" ^~~* ^^

~ ; C'^iCC'C^oC t^- W) ^f *^O

— ^— coccccc

s~*,

C5 ^0 r^l CO OO ^O ^O ^^
SB t ^^ O^ 00 ^G IO ^* ^"O ^^

*

o^\

^^^r^d ceded T*

0

IO

C *t" "". ^^ "0 oo t— rr; <*; — -" JJ

cTlc c '-•. ^ °°. '^ ^ ^ *~! c. ~. C

ii

do

II II

|

II II

VQ

^ — --1CCCOCCCC =

u

VQ

^~" ^^ (^ l^^> ^^ ^^ ^^ f^j

^M r'j ^-H ^^ ^^ ^^ f~^ ^^ f^: ^^ • — *

x —

~ r

^^ f^ \Q S^ ^^ ^\ ^^ ^^ ^-* ^^

^^ ^^ ^j ^^ r*^

r« ,«o <; oc -r —

^

— •

(d

•— •

c -
II II

VQ

- .ooooc c cc c

O

tfe

II II

»-< o_, o' o o

25

VQ

0 c x— > 10 o oo

— i

Q
II

Q

_; ^ d d c c d d d d 3

U

— c,c c c

U

£Q

C - — - oo oc t^ ^^ oo OC 'O ^i

•oq

r~? r^. O^ OO t^* ^O *O "^ f^ CS

ii

°

II II
VQ

— ooooooocc

do

^ W^H— .dodo'doo'd

-
'

x *""

II II
VQ

i^l°CNCi--'-0'-'~s- °

-4,-^dddddddd,3

0
cs

VQ

_. t^i r**! ^^ i"*» ^^* ^^> ^i' ^O *O ^^
^^ ^^ ^^ O^ ^^ ^^ OO i^* *"^ *O

_; r^ ^ o" c' o d d c d

r =
- z

C . — . 1~~ «— ' "2 <^5 CO O 'O IO C>l jj

i-Jovoooooooo.SpC

II

§O C -f — ' i^- f^l O
C C C^ C^ CC 00 1^

II II
VQ

do

"" ^-^&2,— .deedd

II II

- ^^ '^ ^^ ^\ OO t^* *O f**l
^.. ^, ^^ ^^ ^^ ^v ^\ ^\ ^\ ^\

c 2 5~ ~ x i"- o -t <^i ^^

— — d d d d d d c d

^•^ *^3 ^^ _ _ ^^ ^^ ^^

< )\Y(,I \ mi-l'T'SK IN IN I < » ( ELLS

257

Cs is the largest value of C and hence F(CS) is the largest value
F(C) can have. Replacing F(C) by this constant maximum value
and also, in the case r ^ 0. replacing r by 0 (this gives a larger range

1.0 r-

Curve

A
A7

D
D'

r'

ro

1

5.0

1.00

0.67

2

1.0

1.00

1.00

3

1.0

0.10

0.98

4

0.2

0.10

0.90

5

1.0

0.01

0.90

6

0.2

0.01

0.98

See text for details.

FIG. 2.

of integration, so that a larger value of the integral is obtained),
we have

c =- c - r -i( r

*-• = ^ s o I

i r \ i

J r \ i/o

And, completing the integration,

(24)

Hence, if F(C) is such that

^

<

C is greater than 0 for any r. That is, if F(Q is such a function that

258 R. W. GERARD

for a certain range of values of Cs, then in this range C is greater
than 0 for r '• I 0 for any Cs. In the further development this in-
equality will be assumed for the range of Ca employed. A first

s /"*

approximation may be obtained by assuming^C) <£i —2 . (See later.)

PO~

Then from (24) C is approximately equal to Cs. And, as a second
approximation,

rr° i / rr \

±( I r*F(C,)dr}dr
\ I/O

= C. - ^^ (r02 - r2).

Returning to equation (22), the first approximation gives

Q -- 4ir

A 7""^ *?

That is, for high Cs, Q == kF(Ct), where k = ~ : . Using the
second approximation,

Q = ^D f F T Co - ^^ (r02 - r2) 1 rW/".

F(CS)
Expanding, assuming Cs large compared with — -7-- (r02 - r2),

_ p[ ^(Q, , '//"CC,)1 M

Q = 4vr/) F(C.) - - (''o2 - r-} —- rdr

Jo \

< . .

" is ~ '

This shows that the effect of diffusion (in lowering C inside the cell
and therefore A and Q), represented by the term in brackets, is very

dF(C,} .

small unless ,r is very large.
(' C^a

Iv-.pfrimc'ntally, Q is ol)taincd as a function of Cs. Dropping
the Mib-rnpt, s,

. (25,,)

Let F(Q •- + ra~II(Q + higher powers of /•„ times functions of

OXYGEN DIFFUSION INTO CELLS 259

C. Substituting in (25a)

Q(Q = Q(Q +

plus higher powers of r0 times functions of C, or

\WdC
Hence

Q(Q i f°2Q(Q^(Q i

~~ ~^~~

0 dC
By Taylor's theorem, neglecting higher powers,

where

C" = C

and since

^(C) = DF(C),
therefore

(26)
and

C* = C + -- . (27)

207rZ)ro

It will be noted that the larger the value of D, the more nearly do
Q(C) and Q(C*) agree. Q decreases with decrease of C more rapidly
than does the true function A (C). This is due to the factor of diffusion,
making C at interior regions of the cell lower than Cs. Obviously
the closer C approaches CSJ the more nearly will the observed oxygen
consumption of the whole cell equal the theoretical consumption per
unit volume where this volume has an oxygen pressure of Cs through-
out. The higher the diffusion constant and the smaller the radius,
the closer will the approach be. The effect of D is apparent in equa-
tions (26) and (27), but an increase of radius, r0, might at first seem
to likewise reduce the diffusion factor. When it is recalled, however,
that Q(C), the total consumption of a cell, varies with r^, it is clear
that the diffusion effect increases as the square of the radius.

For unfertilized sea urchin eggs, Tang (1931) has obtained the
data necessary for the calculations. These indicate that the maximal

R. Y\r. GERARD

correction is 0.5 per cent, if the diffusion constant be as assumed.
For a constant 100 times smaller, the correction is important, and
this case is illustrated in Table III and the resultant A(C) curve

TABLE III

For unfertilized Arbacia eggs:

Qo = 6.7 X 10~10 cc. O2 per minute, per egg.

D =- 1.1 X 10~7 cc. Oa per minute, per cm.2 per diffusion gradient of one atmosphere

per cm. (1/100 Krogh's value assumed).
Cc = 0.1 atmosphere.

r0 = 3.8 X 10-3 cm.

1

2

3

4

5

6

7

8

Cs

Cs

Cc

Q(CS)
x to-">

Q(C8)

c*

Q(C*)

X 10-'o

CKC*)

A(O

X 10-'

Qo

Go

0.100

1.00

6.7

1.00

0.125

6.7

1.00

2.92

0.090

0.90

6.6 +

0.99

0.115

6.7

1.00

2.92

0.065

0.65

6.6 -

0.98

0.090

6.6 +

0.99

2.89

0.050

0.50

6.4

0.96

0.074

6.6

0.98

2.86

0.035

0.35

6.2

0.92

0.057

6.5

0.97

2.83

0.022

0.22

5.7

0.85

0.044

6.3

0.94

2.74

0.010

0.10

4.4

0.66

0.027

5.9

0.88

2.57

0.004

0.04

3.1

0.46

0.016

5.0

0.75

2.19

0.000

0.00

0.0

0.00

0.000

0.0

0.00

0.00

shown in Fig. 3. For luminous bacteria, because of the very small r0,
the correction turns out to be entirely negligible and the Q(C) curve
as observed is also the A(C] curve (Fig. 3).

Columns 1 and 3 are taken from the data on Arbacia eggs. These

OXYGEN DIFFUSION INTO CELLS 261

values are expressed as ratios in Columns 2 and 4. C* is calculated
by equation (27). Q(C*) is again obtained by interpolation on the
Q(C) curve for eggs, taking Q for each value of C* as if of C. Column 7
is obtained from 6 and Column 8 from 6 and equation (26). A, as a
function of C, is expressed as volume of oxygen per equal volume of
eggs per minute.

It may be observed that the assumption made in this last deriva-

tion, that A <C — ^- is fully justified for this case. Table IV shows

7*0

the actual values of each side of the inequality for several values of C.

TABLE IV

6CD

C

A

0.1

0.0029

0.480

0.05

0.0029

0.240

0.01

0.0026

0.048

0.004

0.0022

0.019

Examination of Available Data

There are a number of data in the literature for several types of
cells sufficiently complete to permit the calculation of the simple Co',
fewer giving as well the observed Co- Table V summarizes most of
this material. It will be noted that the observed Co is from 30 to
1000 times greater than the calculated one in all cases except that
reported by Warburg and Kubowitz (1929). These workers offer
evidence that inadequate shaking at the low oxygen pressures may
cause the liquid phase to lag behind the gas phase in oxygen pressure,
the dissolved oxygen being used by the cells faster than it can be
replaced from the gas phase. Where such a non-equilibrium state
obtains, the actual Co would be less than supposed, which might
account for a reported high C0. In the experiments of Gerard and
Falk (1931), as well as those of Tang (1931), at least, this factor was
considered and apparently excluded by controls of the rate of shaking.
As pointed out by these workers, wide variation in the conditions of
shaking did not affect the Co. In Amberson's experiments no gas
phase was present, only dissolved oxygen being available to the cells
from a large volume of liquid. Also, it may be noted that Stephenson
and Whetham (1924) found B. coli forming lactic acid in air but not
in oxygen; and Novy and Soule (1925) observed for the case of tubercle
bacilli that growth on the surface of an agar slant was retarded at

18

262

R. \V. GERARD

w

K
-

H

*

•n >a

*P eo

10

-f—f-

>O -r

w

—
CJ

r -

r -

o

ia

1 1
C 0

1
o

jj

XX

XX

X

O

X X

X

U

u

O) OO

— i -

o

"a

0

o

— f.

t^ oi

—

a

o

0

p

CN

^1

T

'O

A

<J

.

»o

o

o

V

C 0

cr o

c

c
o

0 0

o

01

d d

d d

d

d

d c

d

5

C CJ

=

0

~

O *J

E

t- ^ ^

c

•- 3

x— s. ' *

oj . *

rt

a.S

1 =

3
O

u

5 j

^c u

y

K o

o

O-^5

-f 01

bO

0 a

ID r—

C o1

'

g

C —

C 0

<"3

OJ

~-

as

d d

d d

d

d

d

X 3

C "

c

r r

C 0

o

^ — >

~

XX

0 re

X

ID

:-

ID J>"

^2 f^l

d
?'-

O

— o

04

o

C ID

01 -H

ID -f

H°

E

-:

re

ID

1S

i

/

O

d

00

X

d

^ x

00 o

X

• ID

M '

fj

•

04

(=3

c

tml

O

c>

3

-^

V

L-

P,

00

i—^

V

vO ^n

c

—

o

*M

NO

— ^3

u

V)

CO

u.

'H

OO
O '^i

S

01

c\ o

Os J

C

-z

•^5 C

o^ o

o —

tx

to ^

1 C.
1 1

•d

"* i-

- ^<

3 O*1

3 c"

U

u

S

=:!

i i

•es

— 'Z

~ 5

in'

o t^

c5

H<

&£

^

1> ^

[^J

S E

Te

'u

O

-2

*u
(U

0
Pi

to
to

0

u

JS

VJ

jj

^s

c

in

~

to g

^*

' ^n

3

u

n g

Sarcina

u

u

a

V

/

Lumino

Microco

o

« -8

in §

KS 5

f 1.

g 2

t*->

0

c

to

u

o

o.

8
-n

3
in

CJ

o

o.

aT

3

c
'£

O)

0.

to
c
'35

T3

6

6

in

a

T3
%

£

c.

X
m

O

X

T— i

o

4->

"rt

3

cr
<u

in

ca

3
O

to

3
|

C
rt

Arbacia eggs, Amberson (1928) finds C0 = 0.10 atmospheres. A cannot be obtained from his data, but he informs me
ed egg suspension consumed 0.25 cc. O2 per hour. If the eggs were just in contact, this would correspond to about
d A = 0.003 cu. mm. O2 per cu. mm. eggs per minute. This value is lower than that obtained by Tang by direct
ifertilized Arbacia eggs, and less than 1/7 the corresponding one for fertilized eggs. A similar value for A, 0.002, can be
irburg's data (1908) on Strongylocentrotus eggs, if the same cell radius as Arbacia and a specific gravity of 1.05 be

to
c

rt
H

"o

4V

cS

jr

<u
g

ID

en

u
U

j;
•a

1

o

Ui

3

to
vC

CJ

c
in

0)

C

o
to
c

in

tn

rt

'a

2

in
cd

in

'35

O

4J
U

U

'a

u

u

to

OJ

•a
u

N

4J

u
_<U

c

^

—

4V

=

^

j

-

C

<u

-.--
u

O)

X

4J

:
00

ca

1^-

•*• *

rt

3

HI

u

• - -

•i'

O

"

>-i in

4)

_

.j

U

ca

/
5

•M

0

-

C3

V

_ -

*O

V

-,

• —

u "

—

4t

h
0)

_:

a

'

U

te

o
O

CO

e

uremi

—
u

4V

-

0)

to to

c8 C

~~! 4_J

^- m

— . ^

a

-+--f

:

y

CO

P

c

-.-

+J

ca

10

-
•„

-
-

~

',

0)

CO

M

vH

B

ca

*J

OXYGEN DIFFUSION INTO CELLS

263

oxygen pressures below 0.06 atmospheres. In Meyerhof's experiments
(1916) also, vigorous shaking is mentioned, and Warburg (1926)
found the oxygen consumption of yeast to fall with oxygen pressures
below 0.2 atmospheres. On the other hand, despite the careful
experimental manipulation described by Warburg and Kubowitz,
extra oxygen may have entered their system by diffusion out from the
measuring column of Brodie's solution. It is possible that the dis-
crepancy between the results of these experimenters and those of all
others is to be attributed to the different cells used.

I am aware of only five sets of observations from which the entire
CS-Q curve can be obtained. The older observations of Amberson
(1928) on fertilized Arbacia eggs fit almost exactly with those of
Tang (1931) on unfertilized ones, even the absolute Co values agreeing.
This is hardly to be expected with a QQ., five to six times greater for
the fertilized than the resting eggs, and if this identity is confirmed
when using the same method on both materials, it will be of con-
siderable importance. Actually, the estimated Qo« for Amberson's
fertilized eggs is less than Tang's value for unfertilized ones (see
footnote to Table V), so little more can be said at this time. For
bacteria, the curve obtained by Shoup (1929) is available. The
complete Q-CS curve was not obtained in the case of Sarcina because
of the continually changing Qo, even in air. Meyerhof (1916, 1917)
has presented less complete data for the nitrifying bacteria. All

TABLE VI

Cs
To

Q

Arbacia
eggs

Luminous
bacteria

Nitrite
formers

Nitrate
formers

1.00

1.00

1.00

1.00

1.00

0.90

0.99

0.98

0.99

0.98

0.65

0.98

0.95

0.97

0.88

0.50

0.96

0.92

0.92

0.80

0.35

0.92

0.86

0.79

0.71

0.22

0.85

0.74

0.54

0.59

0.10

0.66

0.53

0.24

0.34

0.04

0.46

0.15

(.13)

0.00

0.00

0.00

0.00

0.00

these data are summed up in Table VI. The oxygen consumed by
each type of cell is expressed as a fraction of the maximum for that cell,
and all are aligned against C3 values expressed as fractions of C0,
this being likewise taken as unity for all.

264 R. \V. GERARD

It is especially interesting to compare these observed curves with
those calculated on the various types of assumption. For the luminous
bacteria, the observed curve agrees fairly well with the one calculated
on the assumption that there is no concentric zoning of the cell and
that oxygen consumption is independent of pressure, except as diffusion
becomes inadequate. Q falls with Cs, in other words, as would be
expected if the only factor involved were the depth of penetration.
Still the absolute value of Co as measured is far above that expected
on the grounds of the diffusion equation. Either, therefore, the
diffusion coefficient for oxygen through the bacterial cell is over
1,000 times smaller than that for fascia and muscle,2 or the fall of
Q with C, is quite independent of any diffusion factor.

In the case of sea urchin eggs, the situation is still more complicated.
The observed curve does not lit the simple calculated one. It does
fit, however, a curve calculated on the assumption that A is inde-
pendent of C but including the complication that A is not uniform in
all cell regions but is greater in a cortical zone than in the central
portion (Curve I, Fig. 2, and Fig. 3). A lower diffusion constant in
the cortical zone can not yield this shape of curve.

Again, the absolute value of Cc, 100 times smaller than that
calculated, makes it very difficult to interpret the agreement with the
curves calculated on a diffusion basis. If oxygen diffuses through
sea urchin eggs 100 times more slowly than through muscle, all the
results become quantitatively intelligible. But it is more probable
that failing diffusion is only a very minor factor, even in these larger
eggs, and that the direct fall of A with falling C is the primary one.

Assuming that A(C] is directly given by the Q(C) curve in the
case of luminous bacteria and is obtained by correcting for diffusion
in the case of Arbacia eggs, we are left with an empiric relation between
oxygen concentration and rate of oxidations. Shoup (1929) offers a
physical interpretation of his results in terms of adsorption of oxygen
on or evaporation from surfaces in the cell. He puts Langmuir's
(1916) equation for such gas adsorption in the form:

kip(l • - 6) •--- kz6

where k\ and k2 are adsorption and evaporation constants respectively
and 0 i-. the fraction of the total surface covered, and indicates that

this will lit hi> curve if 0 be assumed to represent the percentage of

A
maximal respiration, i.e., — (A0 being the rate of oxidation at oxygen

AQ

2 Each bacterial o 11, as any MM. ill part irk-, is surrounded by an envelope of fluid
which is essentially not stirred. This increases the effective radius of the cell and
hinders diffusion. This alone, ho\\r\rr, to increase C0 1,000 times, would require
a layer over 300 times the radius of the cell. (See Equation 7.)

OXYGEN DIFFUSION INTO CELLS 265

pressure Cc; A the rate at pressure C). Warburg and Kubowitz
(1929) also developed a relationship in terms of reaction velocities for
oxidation and reduction of an iron catalyst.

A somewhat similar development follows.

It may be safely assumed that oxygen combines with at least one
intermediate catalyst before acting upon the substrate. Then,
neglecting any details,

O2 + Int. -> Int.O2
Int.Oa + Sub. -> Sub.O2.

For an equilibrium state, the concentration of Int.O2 must remain
constant and the rates of its formation and reduction must be equal.
So

[O2][Int.]£0 ----A -- [Int.O2][Sub.>r, (28)

where A is the velocity of combination of oxygen, or rate of respiration,
and k0 and kr constants of oxidation and reduction.

In the presence of an excess of substrate, [Sub. 3 may be assumed
to be a constant, ks, and

A ••= [Int.O2]fcr*s.

The maximal rate of respiration, A0, will be obtained when all
Int., Int.t, is in the oxidized state; and the respiration as a fraction
of the maximum is then

A [Int.O2]

IT TI^T
From (28)

[Int.] krks K

[Int.02] " £o[(V] "" [02]
Adding 1 to each side and inverting,

[Int.Oo] [O2]

[Int.] + [Int.02]==# + [02]'

But

[Int.] + [Int.O2] - [Int.,].

So that

j. = K IM .3 (30)

3 If [Sub.] is not assumed to be constant, the final equation is

_£ [P.]

A0 A'[Sub.] + [02]

When [Sub.] is maximum, the effect of [O2] is greatest; as [Sub.] becomes less, it,
rather than [O-2], becomes the critical factor.

266

R. W. GERARD

But Lanpmuir's equation may be readily put in an identical form

by using the notation [O2] for p, - - for 6, and A" for -- • Obviously,

AO K\

an agreement between the curve of this equation and that obtained
experimentally fails to substantiate either of the assumed mechanisms
as the one underlying these oxidations. As a matter of fact, the
experimental data do not agree too well with the equation.
Shoup's data (as read from his curve) yield the following:

Per cent Oz

.4

-4o

K

3.00 and over

1.00

0.00

1.32

0.90

0.14

0.79

0.80

0.20

0.55

0.70

0.23

0.38

0.60

0.25

0.27

0.50

0.27

0.22

0.40

0.33

0.17

0.30

0.40

0.13

0.20

0.52

It is sufficiently obvious that K is even roughly constant only in
the middle range.

Tang's data, similarly handled, give:

Per cent O2

Q A

Qo Ao

K

10.0 and over

1.00

().()

9.0

0.99

0.1

6.5

0.94

0.3

5.0

0.87

0.8

3.5

0.78

1.0

2.2

0.66

1.1

1.0

0.4')

1.0

0.4

0.27

1.1

0.0

0.00

The agreement here is somewhat better. 7v is approximately
constant up to 80 per cent maximal respiration.

For the nitrifying bacteria, the curves of the nitrite and the
nitrate formers are uidely different, and both are far from agreeing
with the derived equation.

It seems to follow, clearly enough, that the Q — Cs curve is far
Irom identical for the various cells, and that the individual form is
probably but little dependent on the limits set by diffusion. No one

OXYGEN DIFFUSION INTO CELLS

267

NO-

NOs

Per cent O2

Q
Qo

K

0

Oo

K

20 and over

1.00

0.0

1.00

0.0

15

0.99 +

0.1

0.92

1.3

10

0.98

0.2

0.80

2.5

7

0.92

0.7

0.72

7.0

4

0.68

1.9

0.57

4.0

2

0.35

3.7

0.34

3.9

simple assumption as to oxidizing mechanism appears able to fit all
cases. Additional data for a much greater variety of cells are greatly
to be desired, and must precede any more elaborate attempt to
theoretically derive the important A(C) curve.

Further important questions concern the change in Qo attending
changed conditions. The fertilized sea urchin egg, for example, after
a tremendous initial burst of oxidation, continues to use oxygen at
five to six times the resting rate. Is this to be interpreted in terms
of a sudden increase of available Int.? None the less, so far as present
data show, there is no change in Co or in A(C). Sarcina, on the
other hand, shows an increase in Co approximately in proportion to
the increased Q when glucose is added. The complicated relations
between lactic acid and Qo have been previously discussed (Gerard,
1931). The suddenly increased respiration of muscle, nerve, and other
cells associated with activity also comes to mind in this connection.

Finally, it is to be noted that most cells exhibit a considerable
"factor of safety" in their rate of energy liberation. The respiration
may be very considerably depressed by low oxygen pressures before
the lessened energy interferes with the physiological activity of a cell.
Thus luminous bacteria do not begin to dim until the oxygen is
reduced to 0.25 per cent (Shoup, 1929), when respiration has been
reduced to half, and still luminesce at 0.0007 per cent O2 (Harvey,
1928). Arbacia eggs show no disturbance in cleavage at oxygen
concentrations over 1.5 per cent (Amberson, 1928) though 10 per cent
is required for maximal respiration.

Similar relations are difficult to demonstrate for tissues because of
the tremendously greater importance of diffusion in the larger masses.
For the case of nerve, it was shown (Gerard, 1927) that the Co calcu-
lated from the simple diffusion equation, assuming A independent of
C, was in good agreement with the observed value. It is probable,
however, that more careful examination of this point would reveal

268 R. \V. GERARD

some dependence of A on C. Less direct evidence indicates that here,
also, conduction is possible when the energy available is but a fraction
of the normal amount. (Gerard, 1930).

SUMMARY

The question of diffusion of oxygen into cells, in distinction to
tissues, requires examination of several factors not previously con-
sidered. The relation between external oxygen pressure and rate of
respiration may be complicated by the existence of different diffusion
constants and respiration rates in separate cell zones.

Equations are developed covering such cases for two concentric
zones in a sphere and relating rate of respiration to oxygen concentra-
tion, cell radius, and diffusion coefficient.

When the assumption, made in similar studies for tissues, that
consumption in any one region is independent of oxygen pressure if
above zero, is avoided, the observed consumption-pressure curve can
be corrected for diffusion effects to give the true consumption-pressure
relation. Equations for this correction are presented.

Available data on bacteria, marine eggs, and other material, are
examined quantitatively. It appears very probable that oxygen
consumption in each region is dependent on oxygen concentration up
to quite considerable values. The theoretical bases for such a de-
pendence are discussed.

BIBLIOGRAPIIY

AMBERSON, \V. R., 1928. Biol. Bull, 55: 79.

EGGLETON, ('.. P., EGGLETON, P., AND lln.i,, A. Y., 192S. Proc, Roy. Soc., Series B,

103: 620.

FENX, \Y. O., 1927. Jour. Gen. Physiol., 10: 767.
GERARD, R. \V., 1927. Am. Jour. Physiol., 82: 381.
i ,i HARD, K. \Y., 1930. Am. Jour. Physiol., 92: 498.
• rERARD, K. \Y, 1931. Biol. Bull, 60: 227.
i ,i RARD, R. \Y., AND E.\i.K, I. S., 1931. Biol. Bull., 60: 213.
HAKVI.V, E. N , 1928. Jour. Gen. Physiol., 11: 469.
HlLL, A. \ "., 1928. Proc. Roy. Soc., Series B, 104:39.
KROGH, A., 1919. Jour. Physiol., 52: 391.
LAV.MIIK, I., 1916. Jour. Am. Chem. Soc., 38: 2221.
Mi M IMI. ii, O., 1916. Pfliiger's Arch., 164: 353.
Mi M RHOF, <)., 1<)17. Pfliiger's Arch., 166: 2-10.
N«.\ \, I . < ,., AND SOULE, M. H., 1925. Jour. Infect. Dis., 36: 168.
PAN i IN, ( . I . A., 1930. Proc. Roy. Soc., Series B, 105: 538.
SIK.I P, ( . S., \'>2<>. Jour. Gen. Physiol., 13: 27.
STKI LA, G., 1<>28. Jour. Physiol., 66: 19.

STEI in NSON, M., AND \YHETHAM, M. 1) , 1924. Biochem. Jour., 18: 498.
TANG, I'. S., l«>31. Hi,,l. Bull., 60: 242.
WARHI RG, ' > , L908. /.citschr. physiol. Chem., 57: 1.
WARIH K(., ( )., 1(>23. Hinc hem. Zeitschr., 142: 317.
WARBURG, ()., 1(>_><.. /Ho, hem. Zeitschr., 177: 471.
WARULRG, ()., AND Ki BOWITZ, !•"., 1929. Biochem. Zeitschr., 214: 5.

THE EFFECT OF TEMPERATURE CHANGES UPON THE

PULSATIONS OF ISOLATED SCALE MELANOPHORES

OF FUNDULUS HETEROCLITUS

DIETRICH C. SMITH
ZOOLOGICAL LABORATORY, HARVARD UNIVERSITY l

When Spaeth (1916a) published a method for producing pulsations
in the isolated scale melanophores of Fimdulus heteroclitus he greatly
broadened the possible scope of investigations dealing with the direct
action of various physical and chemical factors upon such preparations.
Recognizing this fact, it was thought that perhaps an investigation of
the effects of temperature changes upon such pulsations would help in
extending our knowledge of how heat and cold act upon the pigment
cells. The melanophores of isolated Fimdulus scales will respond
directly to variations in temperature (Spaeth, 1913; Smith, 1928), heat
causing a contraction and cold an expansion of the pigment granules.
But beyond these descriptive facts our knowledge does not go. One's
interest is, therefore, directed towards ascertaining how these melano-
phores, when induced to pulsate through the use of Spaeth's method,
would react to temperature changes, with the hope of perhaps obtain-
ing information permitting a more exact analysis of this phenomenon.

Briefly, Spaeth's method of producing pulsations is as follows:
First the isolated Fimdulus scales are immersed in N/10 BaCl2 for 5
minutes, in which time their melanophores become punctate; then
they are transferred to N/10 NaCl, where within fifteen to twenty
minutes after their removal to this solution, the pulsations become
evident. These pulsations take the form of periodic migrations of
the pigment granules in and out of the processes of the melan-
ophores, the movements of the individual pigment granules being
readily seen with ordinary microscopic magnification. Hence, we have
a preparation which, in its rhythmical activity, is comparable to the
behavior seen in other effector organs such as cilia, heart muscle, and
smooth muscle. To this add the advantage that in the melanophore
we deal with an effector consisting of only a single cell. In studying
the effects of temperature changes upon pulsating pigment cells, our
interest is therefore centered on how this factor may alter either the
rate or the extent of these rhythmical migrations, distal and proximal,
of the pigment.

1 National Research Fellow in the Biological Sciences.

269

270 DIETRICH C. SMITH

Now it i- well known that the color changes exhibited by numerous
fishes in respon-e to variations in the tint of the background over which
they swim are determined in a large measure by the behavior of these
melanophores and allied pigment cells. However, in the intact fish the
melanophores do not pulsate; such alterations as take place in the
distribution of their pigment are relatively slow in occurrence and only
in response to a change in the color of the background or of the incident
light intensity. Therefore, in dealing with pulsating isolated melano-
phores, we are not concerned with a phenomenon which plays a part in
the natural economy of the animal.

As a cell the melanophore is a relatively large structure with num-
erous branching processes exceeding in length by three or four times the
diameter of the central body from which they ramify. These processes,
when empty of pigment, are transparent under ordinary conditions of
illumination. In fact, the visible manifestation of the shape of the
cell is due entirely to the pigment granules contained within itself, and
if this pigment is concentrated in the central body, the cell appears to
be spherical. The term punctate is applied to melanophores in this
state. But when under the proper circumstances these invisible pro-
cesses are invaded by a distal migration of the pigment granules, their
outlines are rendered visible. It is this migration of the pigment in
and out of the otherwise transparent processes that creates the illusion
of a change in the shape of the melanophore. The use of the terms
"contraction" and "expansion" of the melanophores must, therefore,
be considered as only convenient expressions describing a redistribu-
tion of the pigment granules within the melanophores.

An investigation of this nature requires an apparatus sensitive
enough to accurately measure these pulsatory movements within the
processes of the melanophores. Such an apparatus was modeled with
some modifications after one originally devised by Spaeth (19166) for
the same purpose, and employed by him in connection with his re-
searches upon the effects of electrical stimulation upon the isolated
melanophore. As my investigations were not concerned with the
effects of electrical stimulation, but were rather studies of the effects
<>t temperature variations, certain changes were necessarily introduced
in the original design, which is fully described and pictured elsewhere
sl>. ii-ili, 1 '>!<>/ ; the most important among these changes being the
addition ot a warm stage to the set-up. This was done by fitting to the
stage oi the microscope a water bath in which a small Stender dish
could be almost < "iii|>]etel\ submerged. This Stender dish was filled
with ? in \.i( 1, .UK! in this solution the scales containing the pulsating
melanopliou-- were placed. By altering the temperature of the water

PULSATIONS OF FUNDULUS MELANOPHORES 271

flowing through the bath, it was also possible to change the tempera-
ture of the NaCl. Any desired temperature between 0° and 45° C.
could be attained within two minutes in the Stender dish and main-
tained indefinitely with no variation greater than 0.2° C.

To measure the pulsations and to make a permanent record of their
variations, the microscope was equipped with an adjustable microm-
eter ocular in place of the usual eye piece, the adjusting screw of the
ocular being fitted to a pulley operating a lever capable of writing upon
a kymograph drum. In making a record, a scale was placed under the
objective, and one of its pulsating melanophores selected for observa-
tion, a convenient division of the ocular scale being brought level and
at right angles to the apparent end of one of the pulsating processes of
this selected melanophore. As the pigment granules in this process
migrated either distally or proximally, the ocular scale division kept
pace with those granules which marked the apparent distal boundary
of the process. Such action was accomplished by turning the adjusting
screw of the movable ocular scale. To insure accurate measurements,
it was absolutely necessary that the melanophore be kept motionless
throughout the experiment. This was accomplished first by bringing
a microscope clamp, attached to the warm stage, to bear lightly upon
the end of the scale, care being taken that the melanophores were not
damaged or stimulated by too great a pressure. Since such a contin-
gency was always imminent, this method for holding the scale was
eventually discarded in favor of another. The new method consisted
in laying a paraffine mould on the bottom of the Stender dish cut so
as to accommodate the scale. In this way the scale was snugly se-
cured without being subjected to any noticeable pressure. No effect
of the paraffine upon the pulsations could be demonstrated. As the
adjusting screw was turned to follow the migrations of the distal pig-
ment granules, the pulley attached to it moved the lever in contact
with the kymograph drum up and down. Consequently, with the lever
writing upon the smoked kymograph paper, a permanent and accurate
record of the pulsations could be obtained. The speed of the drum was
adjusted to one revolution in approximately forty-five minutes.

Records of the melanophore pulsations were not made until the
cells had been exposed to NaCl for at least one hour subsequent to their
immersion in BaCl2. While pulsations are first evident in fifteen to
twenty minutes after the removal of the scales from BaCl2, it usually
requires an additional thirty minutes before they become constant in
extent and frequency at any given temperature. The first pulsations
are extremely slight, but once started they gradually become more
pronounced until they finally reach the maximum characteristic for

272 DIETRICH C. SMITH

the particular temperature in question. An hour, then, gives ample
time to insure the attainment of this point. However, once this
maximum was reached, a new state of equilibrium incident upon a
change in temperature could be attained within 5 minutes. Therefore,
a record having been made at a certain temperature, 10 minutes being
usually -uiVicient for this purpose, another record at another tempera-
ture could be made 10 minutes after the change in temperature was
rflected. The extent of the temperature change between observations
was from 1° to 10° C., the usual variation being about 5° C. The ex-
periments were begun at either end of the temperature scale, and by a
series of four or five steps, the temperature was raised or lowered to the
opposite extreme, the result being the same regardless of whether the
initial record was made at a high or a low temperature. In this way
one melanophore was kept under continuous observation for two or
three hours. At the beginning, a single process of a single melano-
phore was selected, and the record made from the pulsations of this
process. The numerous melanophores of a single isolated scale vary
considerably in size. Though an attempt was made to confine the
experiments to melanophores approximately equal in size, this was
not always possible. However, such attempts were more or less nulli-
fied by the fact that it was impossible to predict the maximal length
of a process by its appearance when the cell was in an almost punctate
or even semi-expanded condition. As fully one-half the experiments
were begun under such circumstances, certain corrections in the data
were necessary before the results could be subjected to satisfactory
analysis.

I I Mi'KKATURE ClIANC.ES AND TIIKIK F.FKECT UPON THE EXTENT OF

TIII-. PULSATIONS

( >n analysing the results obtained from the records pertaining to
the relationship between the temperature and the amount of extension
of the melanophore processes during their pulsations, two general facts
emerge: (1) An increase in temperature causes a decrease in the extent
of the pulsations and (2) this relationship between temperature and
extension is a rectilinear one. Table I presents the data relative to
this relationship. The figures given in the temperature columns rep-
resent the average of the mean maximal and minimal lengths of a
single selected melanophore process during its pulsations at a given
temperature, obtained by measuring in centimeters the height of the
highest and lowest points of the curve traced on the kymograph records.
This averaging was necessary for two reasons: first because during
the pulsations the maximal extensions of a single process were not

PULSATIONS OF FUNDULUS MELANOPHORES 273

o

rt

fS

o
ir;

CN

1

t^
rs

«-H

O
<M

O
""}

C

r^

O
^-H

°0

<N

IO

0)

1^

ON

>o

06

^-

q

LO

»— 1

O
IT;

C-*

CO

1O

01

O

O\

O
C*3

IO

00

o
10

oc

o

CO

"0
01

o
»0

r^

IO

-h
rc

r^
c«)

o
!>.

~

LO

ro

i — i

— p
t*5

OO

cs

>o

01

o
m

\6

0

O

to

UO

01

*— 1

t^

"0

IO

06

1— 1

fO
04

o
10
10

^— i
01

o
10

LO

OO

r*3

tN

ro

10

LO

CN

UO

01

04

O
<n

o
10

•*

o

'^<

o

ITi

CO

"1

^H

o
1*5

•*

t*3

o

CN

oc
ro

o
10

o"

i

OO

c*3

cj

*— t

<M

ro

•*

IO

0

l^

00

ON

0

*-H

o;

c*5

-t^

IO

vO

t-—

>

1

^ 2

w
^
e

<

274

DIETRICH C. SMITH

D
M
tf)

C3

O
1/3

"?

«-H

P*!

10

1 — 1

0

If

r
r*)

-

>o

CO

^H

0

i/*.

r

M

•H

10

•x

t^-

'*"

O

r
rs

-

O

10

1O

es

o
i/l

X

-i

10

OC3

o
X

-

>O

d

^]

o
I/I

r*."

*i

to

O

00

.0
CXJ

t--
t--I

o
r**
N

o
i/~,

O
cs

CO

1O

crj

o

C

cs

0

^H

'^

'0

d

O

10

:

rvj

q

d

«— «

o
"-,

If
pg

-f

00

cs

f*5

i— «

o
I/I

10

O

10
1O

q

N

o

I/-

a

cs

c*5
rN

o
t

-1

i -

es

00

«H

re
es

r*?

^^

0

O

r-t

o

"":

«*5
r-i

o
**)

^•J

10

i -
»— <

ON

"?

q

-t
»— i

o
Ifl

^i

^1

"

0

1 -

»— '

••-.

Ch

"-.

c*5

CN

-t

10

d

•*

CC

»-H

u

cs

"V

-r

if.

C

i -

•

c>

O

-

2

2

S

10

0

i^.

>
<

^ C

34

PULSATIONS OF FUNDULUS MELANOPHORES 275

always constant at a given temperature, a fact which applies equally
to the minimal extensions, and secondly because at a point two or
three degrees above the lowest temperature at which pulsations oc-
curred (about 12° C.), the process would attain during its period of
greatest extension its maximal possible length. However, over these
same low temperatures the length of the process during the phase of
minimal extension was greater, the lower the temperature, the minimal
extension gradually approaching the maximal extension as the tem-
perature came nearer to the critical point. Similarity, as the tempera-
ture rose to within two or three degrees of the upper limit (about 32° C.)
at which pulsations stopped, the melanophores would appear punctate
during their phases of minimal contraction, though at these tempera-
tures the process would keep on periodically extending; the amount of
extension, however, decreasing as the temperature rose. On reaching
the upper limit all pulsations ceased and the cells remained perman-
ently punctate. Therefore, by using averages of the maximal and
minimal lengths, figures are obtained which at the extreme tempera-
tures vary in accordance with the temperature changes. This, of
course, necessitates treating the data obtained in the middle of the
temperature scale in the same manner.

As indicated in Table I, whenever observations were made upon
three or more different processes at the same temperatures, an average
of the figures expressing the relative amount of extension was made.
These averages are given in the bottom line of the table and are plotted
with the result shown in Fig. 1. From this curve it is clear that as the
temperature increases, the extension of the melanophore processes
during their pulsations decreases; a result to be expected, since it was
previously known that high temperatures, when they affect the melano-
phores directly, cause a withdrawal of the pigment into the central
body. Furthermore, it is quite clear from Fig. 1 that the relationship
between temperature and extension is a rectilinear one.

The observational data given in Table I represent the true relative
lengths of the melanophore processes as they were calculated from the
kymograph records. It is apparent, from an inspection of the table,
that there is considerable variation in the pulsatory activity among the
different processes in regard to the amount of extension shown at the
same temperature. Because of this variation, the figures cannot be
profitably compared unless they are all corrected on the assumption
that the amount of extension in any one process at 20° C. is equal to
twenty. When this is done and the results plotted as shown in Fig. 2,
it is obvious that the effect of temperature changes on all pulsating
melanophore processes is virtually the same so far as any relative alter-

276

DIETRICH C. SMITH

ation in the amount of extension is concerned. The one exception is
case 3 of Table I, where the corrected figures were so at variance
with the others that it was impossible to plot them upon the curve.
They have consequently been omitted from Fig. 2. But aside from

E
o

t-
o:
tt
o

u

o!5

10

16 20 24 0 Z&

TEMPERATURE C.

32

I 10. 1. Showing the rectilinear relationship between the temperature of the
medium (abscissae) and the average length of a melanophore process during a single
pulsation (ordinates). The points plotted are those given in the average column of
Table I.

tlii- mn-\|»l;iinr<l discrepancy, the curve substantiates the conclusions
dr.iuii Irom thr one vivrn in Fig. 1.

In tin- hi'jhci rc;i< lies of the temperature scale pulsations cease

PULSATIONS OF FUNDULUS MELANOPIIORKS

277

between 30° and 34° C., the average being approximately 32° C.
When this occurs, the melanophore assumes a permanent punctate
state. At low temperatures, however, the melanophore comes to rest
with all of its processes maximally extended, the pigment being evenly,
if thinly, distributed throughout the whole length of the process.
Such a condition is generally attained at about 12° C., slight move-

30

z
o

E
lu

tc

bJ

10

1Z

16 20 24- c 2&

TE MPER RTURE: C.

3Z

FIG. 2. Showing the result of plotting seventeen different experiments on the
relationship between the temperature of the medium (abscissae) and the average
length of the melanophore process during a single pulsation (ordinates) ; the average
length of the melanophore process being taken in all cases as equal to 20 at 20° C.,
the data in Table I being corrected accordingly.

19

278

DIKTRICH C. SMITH

o
0

(N

O
"1

<N

1*3

- —

oq
^H

0
(N

O

='

*t

(N

c>

r^i

°i

"*

10

00

O
If.

OC

rN

^^

O

a*

10
OC5

10

o

I/-.

00

o

00

"t

o

u~.
r~

01
10

"t;

o
r~

1/5

CO

00

OO

OO

°J

o
f.

C

o

<£>

i— i

OC

oq

**?

»—

q

*— 4

o
"-.

ir.

«*J

o

-r.

t«i

°i

30

-t

t-~

0

•/-,

•T

O

-r

o

w.

ri

c

o

f*3

fN

o
PS

o
m

»

i/-.

o

^^

CN

fO

-t

".

c

i -

'

0

O

~-

es

f*3

-f

'0

0

1^

^—

C2 *^

< |

t"^ •**»

•=

to

PULSATIONS OF FUNDULUS MELANOPHORES 279

o

fS

*o

O

o
u->

P*3

^

0

<^

OO

^^

o
10

O
**i

o

O
PO

**?

CN

0
od

0
u",

C>
N

q
ro

CN
C>4

""*!

CN

<^>
>o

CN

O

O^
N

tN

*— i

^— 1

r— 1

OO

q
«^

O

m

DO

es

r—

o

00
CN

*t
r^

q
10

O
u*>

r—

<N

GO
vH

CN

"t
rO

<r>
•*

CN

o

C-l

o
in

>d

CN

CO
1— (

"t

o
0
<N

q

CN

D

1T>

^— i

•*

-*

"h1
CN

OO
CN

\o
•*

CN

O
IT.

IO

(N

i^)

C~)

OO

fO

T— 1

r>-

00
»— (

O
IO

CN

^-H

rH

q

CN

t-~

^o

t*

CN
CN

O

U",

•*

es

^-H

CN

>O

o
-t

CM

(*3

»— i

*— i
CN

o

»-H

q

tq

T— 1

o
un

<^"
CN

O

e*3

fS

q

*— i

*— H

r^j

q
-t

PO
CN

o
i/3

PS
fN

-*

O

^— (

r^-

»— i

fN

CN

CN

^— i

CN

rH

T— 1

tN

rO

't1

10

O

t^

OO

ON

o

i-H

CN

^O

"*

'O

0

t^

>

<

-*> ^>
»s» •»**

•§ §

a

"a

5S

w 5
***

p-J S

C ft,

280 DIETRICH C. SMITH

merits persisting in some cells until 10° C.; while others come to rest
at temperatures as high as 14° C. In many cases Brownian move-
ment was observed among the pigment granules in all parts of the
process, even though the pulsations had entirely ceased.

TEMPERATURE CHANGES AND THEIR EFFECT UPON THE
FREQUENCY OF PULSATIONS

( >n considering the action of temperature changes upon the fre-
quency of pulsations, we find (1) that an increase in temperature from
12° C. to within the neighborhood of 27° C. causes an increase in the
rate of pulsations, wrhich is followed by a rapid decrease as the tem-
perature mounts higher, and (2) that the relationship between tem-
perature and the increase in rate during the rise from 12° to about 27° C.
is a non-rectilinear one. The data obtained from an analysis of the
kyomograph records are presented in Table II, where the figures given
in the columns under the temperature headings show the absolute
number of pulsations of a single melanophore process in one minute at
the temperature indicated. Every figure in the temperature columns
represents the pulsations per minute of a different process, no two
processes being attached to the same melanophore and every melano-
phore being selected from a different scale. Between different cells the
variation in the frequency of the pulsations at the same temperature
is considerable, as shown in Table II; at 22.5° C., for example, a proc-
ess of one cell shows 0.7 pulsation per minute, while a process of
.mother cell shows 1.9 pulsations per minute. In every case where
three or more observations were made at the same temperature, the
results were averaged, the averages being given in the bottom line of
the table.

In Fig. 3 these averages are plotted, such plotting bringing out the
previously-mentioned increase in rate of pulsation during the greater
part of the temperature rise, the relationship in this case, in contrast to
the relationship between temperature changes and the extent of the
pulsatim^, being clearly non-rectilinear in nature. The possible
significance of this will be discussed later. From this curve the
temperature coefficient of the melanophore pulsations can be cal-
culated, though such coefficients can have only an approximate value.
Between 15° C. and 25° C. the temperature coefficient is 3.57, while
between 20" C. and 30° C. it is 3.6. It is doubtful whether the differ-
ence between the two is significant, especially as the latter coefficient
can have only a >prculat i\ <• value, owing to the irregular appearance of
the depressant ellcct upon the rate of pulsations at the warmer end of
the temperature scale.

PULSATIONS OF FUNDULUS MELANOPHORES

281

As there is a wide variation in the behavior of different processes
at the same temperature, the data presented in Table II must be cor-
rected before they can be presented in intelligible graphic form. In
doing this it was arbitrarily assumed that the rate of pulsation of all
melanophore processes at 20° C. was equal to one, the necessary cor-
rections for all other temperatures being made on this basis. These

LJ

h

16

20

24

30

TEMPERATURE C.

FIG. 3. Showing the non-rectilinear relationship between the temperature of the
medium (abscissae) and the number of pulsations per minute (ordinates). The points
plotted are those given in the average column of Table II.

corrected data were then plotted as shown in Fig. 4, the curve so ob-
tained substantiating the relationship expressed in Fig. 3. Further-
more, the points of Fig. 4 group themselves closely about the curve
given in Fig. 3.

L ! B R A R Y

A

282

DIETRICH C. SMITH

As to the point where the frequency of pulsations decreases with
further increases in temperature, there is considerable variation among
the different melanophores, some showing no diminution in the in-

I 1C, 4. Showing the result of plotting seventeen different experiments on the
relationship between the temperature of the medium (abscissae) and tin- number of
pulsations per niinuii- (ordinatcs); the number of jiulsations per minute beiny taken
in .ill ' |u il in one at 20° ("., and the data in Table II corrected accordingly.

crease in rate ,it temperatures as high as 30° C., while in others this
depressant effect is initiated at temperatures as low as 25° C. In
general, this point occurs in the neighborhood of 27° C. Once a

PULSATIONS OF FUNDULUS MELANOPHORES 283

decrement in rate is established, a further rise in temperature produces
a fall, which is very rapid in comparison to the relatively slow incre-
ment which preceded it, the difference being graphically shown in
Fig. 4. The exact course of this reduction in rate is uncertain, as the
data at hand do not permit an accurate analysis.

DISCUSSION

A rise in temperature, then, affects the pulsating pigment granules
of an isolated Fundulus scale melanophore by increasing the frequency
of their pulsatory movements up to a certain point and at the same
time reducing the distance migrated by these granules during their
pulsations, which in effect gives the appearance of a shortening of the
melanophore process. Gray (1923) mentioned a somewhat similar
situation in his study of the action of temperature changes upon ciliary
activity when he observed a progressive reduction in the amplitude of
the ciliary beat between 34° and 38° C., though the frequency of the
beat increased in proportion to the rise in temperature. In Gray's
experiments, however, this reduction in amplitude was seen only above
34° C. Between 0° and 34° C. the amplitude of the beat remained
constant, though the frequency increased as the temperature became
higher. One must, therefore, make a distinction between cilia and
melanophores in this respect, for in the latter any temperature rise
capable of increasing the frequency of the pulsation also produces a
corresponding decrease in the extension of the granules during these
pulsations. Gray suggests that in his experiments the decrease in
the amplitude of the ciliary beat above 34° C. might be explained by
assuming "that at this temperature the rate at which the process of
contraction is induced in the cilium is more rapid than the process of
relaxation, so that relaxation begins before contraction is complete."

The decrease in apparent extension of the melanophore process
with an increase in temperature might also be explained in the same
way. Once a distal migration of the pigment granules has been estab-
lished, there is the possibility of its continuing until either factors are
brought into play, producing a movement in the opposite direction, or
until the granules reach the end of the process and are incapable of
further migration. But except with extreme low temperatures, prox-
imal pigment migration sets in before the limits of distal migration
are reached. Therefore, the factors producing this proximal migra-
tion are capable of overcoming those governing the distal migration,
or else the latter cease to operate before the pigment granules reach the
limits of the process. But whichever explanation is correct, the matter
is of secondary importance, as the factors determining the extent of

284 DIETRICH C. SMITH

the migration of the pigment granules must be independent of both.
This condu-ion i- deducible from the response of melanophore pulsa-
tions to variation?- in temperature, for such variations seem to influence
independently two separate series of reactions within the pigment cell,
one governing frequency, and the other the extent of the migration.
The fact that the distance migrated by the pigment granules during
their pulsations shows a rectilinear relationship while the frequency of
the pulsations varies in a non-rectilinear manner in relation to tem-
perature changes hardly warrants any other interpretation. There-
fore, if extension is determined by frequency, and it is by no means
clear that this is so, it is only because the factors leading to a proximal
migration dominate those which might favor further extension and
not because extension is a function of frequency. It is interesting to
note in this connection that Clark (1920) demonstrated in the isolated
rabbit's auricle a rectilinear decrease in the height of contraction with
an increase in temperature; and at the same time observed a non-recti-
linear increase in the frequency of the beat. In isolated frog's heart
this was not the case, both the diminution in the force of contraction
and the increase in frequency of the beat caused by a rise in tempera-
ture being of a non-rectilinear nature.

Since melanophores contract in the presence of an insufficient
amount of oxygen, there may exist a relationship bet\veen the decrease
in the oxygen tension of the water with an increasing temperature and
the decrease in extension of the melanophore processes. But Spaeth
(1913) has shown that in the isolated scale, melanophore contraction
of the pigment occurs at high temperatures both in media furnished
with an excess supply of oxygen and in media with the usual oxygen
i rnsion for the temperature in question. Therefore, the contraction
of the pigment granules with an increase in temperature cannot be due
to a lack of oxygen. But the contraction of pigment in a non-pulsating
niflaiiophore may be governed by factors other than those which con-
trol the amount of extension of the pigment granules in a pulsating
niflanophore. Unfortunately this question is left open as my obser-
\ation- did not include the effect of a decrease in oxygen supply
upon either the- frequency or the extent of pulsations and nothing is
known concerning this question from other sources. The sudden
decrease in the t'reqnenry of pulsations with temperatures above 27° C.,
following .1 < (insistent increase with a rise in temperature to this point,
may be concerned with a po»i|>le inadequacy in the oxygen supply to
the melanophoics at high temperatures. \Yhether or not this is true
for pulsating melanophores, it i> apparently untrue in the case of cilia,
for, according to Gra) • 1'JJ.i the oxygen consumption of the tissue

PULSATIONS OF FUNDULUS MELANOPHORES

keeps pace with the speed of the beat over the temperatures used in
his experiment. As a consequence of this, Gray doubts whether a rise
in temperature involves in well-ventilated tissues a lack of oxygen
except with very high temperatures, when a decrease in mechanical
activity might well be due to insufficient oxygen.

The effects of temperature changes upon pulsating melanophores
and the effects of temperature changes upon ameboid movement,
ciliary activity, and heart beat show a certain similarity. In all cases,
the effect is one of an increase in frequency, an increase which is more
rapid than a rectilinear function. But while the temperature coeffi-
cient of ameboid movement, ciliary activity, and heart beat are all in
fairly close agreement, the melanophore is peculiar in that it has a
temperature coefficient which in comparison to the others is remark-
ably high. As we have previously seen, the temperature coefficient of
a pulsating melanophore is 3.57 between 15° and 25° C., and 3.6 be-
tween 20° and 30° C., figures which can be compared with those given

in Table III.

TABLE III

Q,o 15°-25° C. 20°-30° C.

Ciliary activity 2.15 1.95 (Gray, 1923)

Ameboid movement 2.04 - (Pantin, 19246)

Heart beat 2.10 1.90 (Clark, 1920)

Melanophore pulsations 3.57 3.60

A determinative role on the part of viscosity changes has already
been assigned as a possible factor controlling the movements of pigment
granules within the melanophores and it is easy to imagine the same
factor assuming considerable importance in governing the frequency
and the extent of melanophore pulsations. But unfortunately our
knowledge of the exact effect of temperature changes upon the viscosity
of protoplasm is at present too meager and too contradictory to war-
rant any extended conclusions as to how much viscosity is involved in
responses of melanophore pulsations to variations in temperature.
Both Pantin (1924a) and Heilbrunn (1924) have interested themselves
in the relationship between temperature and its effect upon animal
protoplasm, but with no agreement as to results. Pantin, who worked
with the protoplasm of Nereis eggs, states that as the temperature
increases, the viscosity decreases, the relationship being a non-recti-
linear one with a temperature coefficient of 1.3 between 15° C. and
25° C. and 1.26 between 20° C. and 30° C. Heilbrunn, on the other
hand, working with Cumingia eggs found a sharp decrease in viscosity
between 0° C. and 2° C., followed, as the temperature rose further, by
a gradual increase which came to a maximum in the neighborhood of

286 DIETRICH C. SMITH

16° C. This in turn was replaced by a decrease until the temperature
reached about 30° C., when within the next degree or two there was a
very rapid increase. But on the basis of either Pantin's or Heilbrunn's
work it is clear that viscosity is not an important factor in determining
the responses of melanophore pulsations to temperature changes.
The difference between the temperature coefficients which Pantin ob-
tained and those found to apply to melanophore pulsations are so
great that the viscosity changes in the melanophores can hardly be of
importance in governing the variations in pulsations associated with
temperature changes. While, on the basis of Heilbrunn's results, if
viscosity changes within the melanophore were an important factor, the
curves obtained in these studies on the relationship between tempera-
ture changes and the amount of extension of the pigment granules
together with the frequency of their pulsations would hardly be as
regular as they are, but would show significant variations or breaks in
accordance with the curve obtained by Heilbrunn. For the present
at least, the conclusion seems inevitable that viscosity changes in the
protoplasm of the melanophores attendant upon variations in tempera-
ture are not responsible for the particular manner in which the pulsa-
tions of these cells respond to alterations in temperature. Whether
viscosity is of importance in governing the non-pulsatory changes which
occur in the distribution of melanophore pigment, as a fish adapts
itself to shifts in the color of its background, is quite possibly another
matter and one about which nothing can be stated here.

SUMMARY

1. Pulsations in isolated scale melanophores of Fundulus heteroclitus
occur in N/10 NaCl after treatment with N/10 BaClo at any tempera-
ture between 12° and 32° C.

2. Below 12° C. pulsations cease, the melanophore coming to rest
in the maximally expanded state.

v Above 32° C. pulsations also cease, the melanophore coming to
rest in the punctate state.

4. Between 12° and 32° (\ the extent of the pulsations decreases in
a rectilinear manner with an increase in temperature.

5. l.rtween 12° and 27° C. the rate of pulsations increases in a
non-rectilinear manner with an increase in temperature. With a
further rise in temperature above 27° C. there is a rapid decrease in the

rate of pul>at i<m>.

LITERATURE

CLARK, A. J., 1920. The Effort <>t .\lirr.itionsofTemperatureupontheFunctionsof
the Isolated Heart. Jour. I'liy^ol., 54: 275.

GRAY, J., 1923. The Mechanism of Ciliary Movement. III. The effect of tempera-
ture. Proc. Roy. Soc., London, 95#: 6.

PULSATIONS OF FUNDULUS MELANOPHORES

287

HEILBRUNN, L. V., 1924. The Colloid Chemistry of Protoplasm. III. The vis-
cosity of protoplasm at various temperatures. Am. Jour. Physiol., 68: 645.

PANTIN, C. F. A., 1924a. Temperature and the Viscosity of Protoplasm. Jour.
Mar. Biol. Ass., 13: 331.

PANTIN, C. F. A., 19246. On the Physiology of Ameboid Movement. II. The
effect of temperature. Brit. Jour. Exper. Biol., 1: 519.

SMITH, D. C., 1928. The Effect of Temperature upon the Melanophores of Fishes.
Jour. Exper. Zool., 52: 183.

SPAETH, R. A., 1913. The Physiology of the Chromatophores of Fishes. Jour.
Exper. Zool., 15: 527.

SPAETH, R. A., 1916a. Evidence Proving the Melanophore to be a Disguised Type
of Smooth Muscle Cell. Jour. Exper. Zool., 20: 193.

SPAETH, R. A., 1916&. The Responses of Single Melanophores to Electrical Stimula-
tion. Am. Jour. Physiol., 41: 577.

SPAETH, R. A., 1916c. A Device for Recording the Physiological Responses of Single
Melanophores. Am. Jour. Physiol., 41: 597.

IXFLl KXCK OF CYANIDE AND LACK OF OXYGEN ON THE

ACTIVATION OF STARFISH EGGS BY ACID, HEAT AND

HYPERTONIC SEA-WATER

RALPH S. LILLIE

(From the Marine Biological Laboratory, Woods Hole and the Laboratory of General

Physiology, University of Chicago)

In previous papers (Lillie, 1915, 1917, 1926) I have called attention
to the general resemblances between the phenomena of acid activation
and of heat activation in unfertilized starfish eggs, — a parallelism indi-
cating that both forms of activation are to be referred to the same kind
of change in the egg system.1 This primary or critical change in some
way transforms the egg from the quiescent into the automatically
developing or "activated" state. Since exposure to temperatures
above the physiological range is known to produce acid (especially
lactic acid) in many cells (as seen, e.g., in the phenomena of heat rigor),
the inference is natural that heat activation is in reality a secondary
effect resulting from the production of acid within the egg; and the
essential problem becomes that of determining the factors concerned
in acid activation. Experiments on the relative activating effective-
ness of a variety of weak penetrating acids having a wide range of dis-
sociation constants (Lillie, 1926 and 1927) indicate that the essential
factors in a given activating effect are (1) the production of a certain
definite degree of acidity in some region of the egg (probably cortical)
and (2) the maintenance of this acidity for a definite time. Some
progressive change, the "activation process" or "activation reaction,"
occur > within the egg during this time; and if the egg is to be rendered
capable of developing to an advanced stage, the process must advance
to a certain stage of completion. One of the most striking features of
the activation process in starfish eggs is that it is readily arrested at
any stage by -imply returning the eggs to normal sea-water; in such a
case it may be renewed and brought to completion by a second ex-
posure to heat or acid, applied after not too long an interval (Lillie,
1915, pp. 2X4 si •</. ). The rate of the activation process, in the case of a
particular arid, app»-ai> closely proportional to the concentration of

1 The chief p.millrls arc (1) tin- Minilnrity in the time-relations of partial and
complete activation by the two met hods: (2) the fact that both are completely effec-
tive only during the prematuration.il period of the egg; (3) the substitutability of one
for the other in experiments on partial activation, and (4) other parallels described in
the present paper.

288

ACTIVATION OF STARFISH EGGS

the latter in the external medium — more specifically to the concentra-
tion of its undissociated molecules, which alone appear to penetrate
freely to the site of the activation reaction (Lillie, 1926 and 1927).

Since apparently any non-toxic penetrating acid can activate star-
fish eggs, the special chemical nature of the acid would seem to be
immaterial; its presence merely enables the activation process to pro-
ceed in the egg system. In general our present data indicate that the
rate of activation is closely proportional to the acidity (cH above a cer-
tain critical level) attained at the site of activation. The most reason-
able assumption appears to be that activation is an effect of the chemi-
cal union of certain specific substances already present or available
in the egg,2 and that this interaction, the primary or key reaction in
the activating sequence, occurs only above a certain level of acidity
(i.e., within a certain pH range). The implication is that the product
of this reaction — which we may call the "activating substance" —re-
quires to accumulate to a certain definite level if complete activation is
to result. Eggs in which this critical quantity of activating substance
has been formed proceed normally with their development if placed
under appropriate external conditions. That the accumulation of
some material is a necessary condition of activation is indicated by the
progressive character of the activation reaction, and especially by the
possibility (already referred to) of arresting it at an incomplete stage
and afterwards renewing it.

The activation reaction cannot be characterized definitely in chem-
ical terms at present. The fact that responsiveness to activation,
either by parthenogenetic agents or sperm, begins normally to decline at
or about the time of separation of the first polar body and remains
incomplete during the post-maturational period indicates a loss or
destruction of some reactant or reactants at this time.3 Completely
mature eggs are capable of only partial activation: although such eggs
still form fertilization-membranes and cleave after treatment with
fatty acids or warm sea-water, only a small proportion (if any) develop
to swimming stages; similarly only a minority of eggs fertilized with
sperm during this period exhibit normal development. The capacity
for complete activation is lost, apparently permanently, during the
progress of the maturation divisions (Lillie, 1908).

2 This is also the assumption of the fertilizin hypothesis of F. R. Lillie, which is
quite consistent with the present view. Cf. Lillie, F. R.: " Problems of Fertilization,"
University of Chicago Press, 1919; also E. E. Just: "The Present Status of the Ferti-
lizin Theory of Fertilization." Protoplasma, 1930, Vol. 10, p. 300.

3 The term "prematurational period" is used in the present paper to designate
the period between the dissolution of the germinal vesicle and the separation of the
first polar body; the post-maturational period is that succeeding the separation of the
second polar body.

290 RALPH S. LILLIE

This change in the physiological properties of the egg, like the
change leading to its natural death if it is left unfertilized, (Loeb, 1902 a
and b] is intimately dependent on normal respiration, as is shown ex-
perimentally by the effects of cyanide or deprivation of oxygen. If
freshly removed starfish eggs are placed for two to four hours (at 20°) in
sea-water freed of oxygen or containing KCN (M/500, M '1000), and are
then returned to normal sea-water, they show an almost unimpaired
response to fertilization or artificial activation; while the response of
control eggs left for the same interval in normal sea-water is partial or
defective. Such facts indicate that some material essential to the
activation reaction (i.e., some precursor or precursors) is removed or
destroyed at a definite stage in the normal oxidative metabolism of the
egg; apparently this occurs most rapidly during the period at which
the polar bodies are being separated.

A further fact of significance is that the activation reaction, as such
(i.e., the reaction occurring during the actual exposure to fatty acid
or heat), is independent of immediate oxygen consumption; that is,
the eggs respond in a normal manner to heat or fatty acid while im-
mersed in oxygen-free or cyanide-containing sea-water — even if they
have previously been exposed for some hours to these media. Such
suppression of oxidative metabolism has, however, a characteristic
modifying influence on the rate of activation, as is shown by the
experiments to be described below.

1. Influence of Oxygen-lack or Cyanide on the Effect of a Second Activat-
ing Treatment Following Partial Activation by a First

Treatment

It was found in earlier experiments (Lillie, 1915, pp. 284 seq.} that
two successive brief exposures to either heat or fatty acid, separated
by a considerable interval (up to 30 minutes or more), may by a process
of summation result in complete activation, although either exposure
acting alone has only a partial effect.4 A single continuous exposure
lasting as long as the sum of the other two also causes complete activa-
tion. Such an experiment shows that the modification (whatever its
nature) produced by the first exposure persists for some time, and that
the second exposure induces further modification of the same kind.
II the first exposure is made early, e.g., within 5 to 10 minutes after the

1 In partial a< -tivation UK- eggs form typical fertilization-membranes, but cleave
slowly ami irregularly (if at all) and die before reaching the blastula stage. In
complete art i\. it ion cleavage approaches the normal in rate and regularity and the
great majority (in favorable cases all) of the eggs form active blastulae and gastrulse.
All gradation^ in dc-rcc of activation can be obtained by varying the length of ex-
posure.

ACTIVATION OF STARFISH FGGS

291

commencing dissolution of the germinal vesicle, an hour or more may
elapse without marked decline in the effect of the second exposure;
later this exposure becomes progressively less and less effective. This
decline in the response to a second activating treatment appears to
follow the same course as the normal decline of activability already
described. It also depends on the oxidative metabolism of the egg
and can be greatly retarded by cyanide or removal of oxygen; appar-
ently it is referable to a depletion of the reserve of activable material
(precursor or precursors of the activating substance) left unchanged by
the first treatment.

The following experiment (June 10, 1927) will illustrate. The
mixed eggs from three starfish were exposed during the prematura-
tional period for 6 minutes to .002 M acetic acid in balanced NaCl -
CaCU solution at 20° C. Part (A} was then transferred to normal sea-
water (at ca. 18° C.) and part (B] to sea-water containing M/1000
KCN. At the intervals indicated in Table I eggs from both lots were
again exposed for 6 minutes to the same acid solution and returned to
normal sea-water. The total exposure, 12 minutes, is approximately
the optimum for this solution at 20°. The intervals between the two
successive exposures varied from 11 minutes to two hours. The results

are shown in Table I.

TABLE I

Interval

Percentage of eggs forming blastulae as result of second exposure

between
exposures
(minutes)

A. Eggs kept in normal
sea-water

B. Eggs kept in M/1000 KCN
in sea-water

11

90

90

20

60-65

90

30

60-70

90

45

60-70

70-80

60

60-70

80-90

90

50

70-80

120

5

50

Eggs exposed for 6 minutes without any second treatment showed typical partial
activation, and a few (2-3 per cent) formed blastulse. Sperm-fertilized eggs devel-
oped normally.

In normal sea- water nearly all the eggs have lost responsiveness to
the second treatment after two hours, while of the eggs remaining in
cyanide-containing sea-water for the same interval a large proportion
developed. Table II describes a second similar experiment (July 11,
1927) with longer intervals between exposures.

292

RALPH S. LILLIE

TABLE II

July 11, 1027. The procedure was the same as in the previous experiment.
The acid solution used was .0015 M acetic acid in Xa-Ca solution at 20° C. A
control scries showed an optimum continuous exposure of 12-14 minutes, yielding
60-7') per rent blastulae. The first exposure was 6 minutes; eggs so treated without
a sen Mid exposure gave membranes and some cleavage but no blastulae. After each
interval the eggs kept in sea-water received second exposures of 6, 8, and 10 minutes;
those kept in KCN sea-water received 4, 6, 8, and 10 minutes.

Interval
between
res (hours)

Percentage of eggs forming blastulae and
optimum durations of second exposure

A. Eii'-i^ ~<-;>t in normal
sea-water

1',. Ku^ in M 1000 KCN in
sea-water

1

3

5

22

35-40 (8 and 10 min.)
few (1%) (6-8 min.)
0 (6-8 min.)
(all dead by 22 hrs.)

ca. 50 (6-8 min.)
50-65 (4-6 min.)
25-35 (4-8 min.)
0

Sperm-fertilized eggs developed normally.

Well-marked response to the second exposure after 5 hours in
M/1000 KCX was observed in four other similar experiments and also
in three experiments in which the eggs were left for 5 hours in oxygen-
free sea-water between exposures. Four experiments with periods of
18 to 22 hours in M/1000 KCN showed almost no response after this
interval, although in one experiment some slight effect was apparent
after 19 hours.

Sea-water deprived of oxygen (by prolonged boiling, restoration of
the original volume with boiled distilled water, and passage of a stream
of hydrogen or nitrogen overnight) prolongs similarly the possible
summation-interval. For example, on August 16, 1927, eggs kept for
2 12 and 5 hours in oxygen-free sea-water, after a first exposure of
5 minutes to .002 M acetic acid, yielded with both intervals 60-70 per
cent ol lil, i>i ul. r as the result ot a second exposure ol () to S minutes to
the same solution. The control eggs kept in normal sea-water showed
almost no response to the second exposure. The eggs which received
only tin ln>t exposure of 5 minutes showed typical partial activation
wiih no blast ul. !•, while those receiving continuous exposures of 12, LS
and 1 1 minutes at the same time (M) minutes after removal from the
animal- gave 75-90 per cent of blastuhe.

Seventeen experiments of this kind were performed during the
summer ul ]<)27 with results similar to those cited. In all cases the
interval between the fn>t inadequate and the second effective exposure
was prolonged to several hours by the suppression of oxidative metab-
olism in the manner indicated.

ACTIVATION OF STARFISH EGGS 293

These results arc in conformity with the hypothesis that some ma-
terial essential to the production of the activating substance disappears
normally from the egg in the course of the regular oxidative metabolism.
Hence the unused reserve of this material is retained within the egg for
a longer period, together with the activating substance formed by the
first exposure, if oxygen-consumption is prevented.

There is, however, some indication that the reserve of precursor
material (activable substance) remaining in the egg after the first
exposure is not retained entirely unaltered during the interval, but
that it undergoes gradual transformation into activating substance.
This is seen in the fact that partially activated eggs show an increase
in the degree of their activation (as compared with eggs of the same lot
left in normal sea-water), if they are kept undisturbed for some hours in
oxygen-free or cyanide-containing sea-water. For example, in the
experiment just cited (August 16, 1927) eggs left for two hours in
oxygen-free sea-water (after the initial exposure of 5 minutes to .002 M
acetic acid), and then transferred to normal sea-water, yielded without
any second treatment 10-15 per cent of blastulae, while after remaining
for 5 hours in oxygen-free sea-water more than 50 per cent formed bias-
tula^. In the latter case (5 hours without oxygen) a second exposure of
two minutes proved almost as effective as one of 6 minutes, both yield-
ing 65-70 per cent of blastulae. Such cases show that a partial activa-
tion may be completed in many eggs by simple exposure of one to seve-
ral hours to cyanide-containing sea-water. This effect is similar to
Loeb's completion of activation in sea-urchin eggs, after membrane-
formation by acid, by after-treatment with cyanide-containing sea-
water (Loeb, 1913). In starfish eggs a well-marked shortening of the
optimal duration of the second acid treatment was a constant effect of
one or more hours exposure to oxygen-free or cyanide-containing sea-
water. Apparently prolonged suppression of oxygen-consumption in
the partially activated eggs has a physiological effect similar to that of
treatment with acid or high temperature.

It seems probable that the production of acid within the egg under
the conditions of anaerobic metabolism is responsible for the increase
of activation shown in such cases. This supposition is consistent with
the fact that previous exposure of normal untreated starfish eggs to
oxygen-free or cyanide-containing sea-water also very definitely in-
creases the susceptibility to activation by either heat or acid, as will
be shown in the next section. I have, however, never succeeded in
activating such eggs, even to the extent of membrane-formation, by
this treatment alone.5 The acid formed under these conditions may

5 E. P. Lyon obtained parthenogenetic development in the Strongylocentrotus
egg at Naples by prolonged exposure to solutions of KCN (M/100 to M/1000) in sea-
water (cf., Am. Jour. Physiol, 1903, Vol. 9, p. 308).

20

294 RALPH S. LILLIE

not reach the critical level of concentration required for activation, or
it may l>e ineffective for some independent reason.

2. Activation after Previous Exposure of Normal Eggs to Cyanide and

Lack of Oxygen

In the experiments just described partially activated eggs after
remaining some hours in oxygen-free or cyanide-containing sea-water
showed a tendency to complete their activation spontaneously; i.e.,
the second effective exposure to acid was unexpectedly brief, and many
eggs formed blastulae without any second exposure. Similarly, normal
unactivated eggs kept under the same oxidation-suppressing conditions
show in course of time an increasing susceptibility to activation by
either heat or acid. Suppression of oxygen-consumption has thus a
double effect on the eggs: (1) prolongation of the time during which
they remained activable, and (2) increase of activability, as shown by a
decrease in the effective durations of exposure to the activating agent.

Cyanide-containing Sea-u'ater. — A typical series in which the eggs
were exposed to KCN-containing sea-water for 2 1/2 hours before the
activating treatment is the following:

July 3, 192(). Kggs from several starfish were placed, 35 minutes
after removal, in sea-water containing M/1000 KCN. Two hours, 35
minutes later, a part (B) was washed in several changes of normal sea-
water, left in sea-water for 26 minutes and then placed in M/260
butyric acid solution in sea-water. A second part (C) after 2 hours,
32 minutes in the KCN sea-water was transferred directly to KCN
sea-water containing also M/260 butyric acid.6 The essential differ-
ence between B and C is that in B the activating treatment occurred
after the eggs were washed free of cyanide, while in C the entire period
preceding and during activation was spent in the presence of cyanide.
At two-minute intervals eggs were transferred from the butyric acid

c In estimating the rate of activation in solutions of fatty acid containing KCN,
allowance must be made for the neutralizing action of this salt, which is very nearly
the same as that of NaHCOj. In general the rate of activation exhibited by starfish
eggs placed directly from sea-water in solutions of fatty acid containing also KCN
is closely similar to that of eggs exposed to solutions of the same non-neutralized acid
strength without KCN. This was shown experimentally as follows: Eggs exposed
to balanced NaCl-CaClj solution containing both acetic acid and KCN in equal
concentrations, .002 M, showed no membrane formation or other signs of activation
with exposures from 5 to 18 minutes. The control experiment with .002 M acetic
.11 id alone showed optimal activation at 10 minutes, with 75-80 per cent of eggs
ining blastulae. The same Nad-Cadi solution containing .002 M acetic acid
plus .001 M KCN gave a rate of activation similar to that of .001 M acetic acid with-
out K< \ . I he solution used above, sea-water containing M/260 butyric acid plus
M/1000 KCN, corresponds closely in its activating effect on normal eggs to a solution
of .003 M butyrir acid in sea-water (i.e., M/260, ca. .0038-.001 M).

ACTIVATION OF STARFISH KC.C.S

295

solutions to normal sea-water; later the proportions developing to
blastuhr were determined. The control (/I) consisted of eggs exposed
to M/260 butyric acid in sea-water, beginning 40 minutes after re-
moval from the animals.

Table III gives the approximate percentages of eggs forming bias-
tula? with the different exposures.

TAHLK III

Duration of
exposure
(20° C.)
(minutes)

Percentage of eggs forming blastulae

A. (Control without
KCN exposed to
normal sea-water con-
taining M/260
butyric acid)

B. (KCN 2H hrs.
exposed to
normal sea-water con-
taining M/260
butyric acid)

C. (KCN 2% hrs.
exposed to sea-water
containing M/260
butyric acid
plus
M/1000 KCN)

2

0

40-50

no observation

4

1

30-40

60-70

6

15-20

30-40

60-70

8

30-40

60-70

50

10

ca. 90

70-80

20-30

12

80-90

70-80

ca. 10

14

70-80

30-35

ca. 10

16

20-30

ca. 10

1

Nearly half the eggs in Part B formed blastulae with exposures as
brief as two minutes; also a large proportion continued to develop
favorably with longer exposures up to 12 minutes. The effect is even
more striking in Part C. Here the effective (non-neutralized) concen-
tration of acid in normal sea-water (ca. .003 M) typically requires 20
minutes or more for complete activation; yet after the preliminary
treatment with KCN, two-thirds of the eggs formed blastulae with
exposures of 4 to 6 minutes.

Two other series on July 15 and 16, with preliminary exposures of 3
and 5 hours (respectively) to KCN sea-water gave closely similar re-
sults. In the series of July 15 eggs exposed to M/260 butyric acid in
sea-water, after 3 hours in KCN sea-water, for periods of one to 6 min-
utes, gave 60-80 per cent blastulae. With the cyanide-containing
butyric solution one minute was ineffective, but exposures of 2 to 6
minutes yielded 50 per cent or more blastulae. The results of July 16,
with 5 hours previous treatment with KCN, were similar; butyric acid
in normal sea-water showed optimal activation at 2 minutes with 70-80
per cent blastuke; the same solution in KCN-containing sea-water was
ineffective at 2 minutes, but with exposures of 4 to 14 minutes gave 50
per cent or more blastulae. On July 17 and 18 similar series with 5

206

RALPH S. LILLIE

and 3 hours in KCN-sea-water showed the same effect with somewhat
fewer blast tike.

The shortening of the effective exposures under the influence of
\\( '\ i> appreciable although slight after 15 or 20 minutes and increases
with time. Table IV gives the results of a typical series in which the
preliminary exposure to KCN wras varied from 15 to 65 minutes. The
optimum exposure for the control eggs unexposed to KCX was 10
minutes (21° C.), with more than 90 per cent forming blastulre.

TABLE IV

Experiment of June 20, 1()3(). K^< \\ere exposed during the prematurational
period to M 260 butyric acid in normal sea-water at 21° C. after washing free of
i y.mide.

1'ii'vi. ms Exposun ti
M.ionii KCN (minutes]

Times of Exposure to Butyric Acid (minutes)
and Percentages of Eggs Forming Blastulae

2

3

4

5

6

7

8

9

10

None (control)
15

0
0

0
0

0

0

2-3
2-3

1 1

2-3

5-10
60-70

ca. 50
60-70

65-75

75 S5

90
80-90

26

0

ca. 1

ca. 1

ca. 1

25-35

50 (id

70-80

70-80

ca. 50

45

1

ca. 5

10 20

40-50

(>() 7(i

65-75

50-60

65

ca. 1

ca. 5

10-15

20-25

35-45

55-65

25-35

Three other similar series carried out between June 1 7 and June 23,
1930, with exposures to KCN ranging from 12 minutes to 5 hours, gave
the same general result. In general, the degree of shortening increases
with duration of exposure up to a maximum of 2 to 3 hours. After 5
hours there is usually a decline of responsiveness, and after 20 to 24
hours no blastulae were obtained, although there was some partial
;K tivation.

Oxygen-free sea-water. — Experiments with eggs remaining for 2 to
5 hours in oxygen-free sea-water before activation with fatty acid gave
.1 -imilar result.7 Kight successive series between July 23 and August 3,
]*>2'> showed uniformly a striking decrease in the activating exposures;
ol Midi eggs a large proportion and in some cases nearly all formed bias-
tula- after exposures of one to 2 minutes to M/260 butyric acid. Table
V gi\ es a Minmiary < >t t he tii>t four series. The eggs, after exposure to
the 'ii free sea-water (in HaM1^ through which a slow stream of

hydrogm was kept flowing) for the periods named, were treated for

7 In general ihe phvMolo-i.-.il effects of exposing starfish eggs to cyanide-
containing and to oxygi 11 /..HI i are strikingly similar. One constant effect

following exposure lor on< to several hours to either medium is to render the eggs
glutinous or sticky, so ili.it they cohere in masses, a result (apparently) of the secretion
of some water-swelling or glutinous rn.iteri.il.

ACTIVATION OF STARFISH EGGS

297

periods ranging from one to 16 minutes with M/260 butyric acid made
up in both normal and oxygen-free sea-water. The controls, as before,
consisted of eggs exposed during the prematurational period (usually
30 to 45 minutes after removal) to M/260 butyric acid,8 again both with
and without oxygen. These showed optima varying between 6 and 12
minutes (most frequently at 8 minutes) with 70 to 85 per cent of eggs
forming blastula?. The temperatures of the sea-water and butyric acid
solutions were 21° to 22.5° C.

TABLE V

Experiment
and date
(1929)

Control
(Effective
exposures
and percentages
of eggs
forming
blastulae)

Period
in 02-free
sea- water
before

activation

Exposures to M/260 butyric acid and
percentages of resulting blastulae

(a) 0-2-containing
acid solution

(b) O.-free
acid solution

1. July 23

mill. %
6 (35-40)
8 (70-80)
10 (50-60)

lirs. nt ni.

2 15

mhi. %

2 (ca. 50)
4 (35-45)
6, 8 (25-35)
10 (20-35)

min. %
2 (ca. 50)
4 (50-60)
6 (35-40)
8, 10 (25-35)
12 .(20-25)

2. July 24

4 (25-35)
6 (80-90)
8 (50)
10 ( 5)

5

1 (40-50)
2,3 (20-30)
4 ( 5-10)
6 (0)

1 (35-40)
2 (30-35)
3 (15-20)

4 (ca. 1)

3. July 25

6 (40-50)
8 (75-80)
10 (ra. 90)
12 (50-60)

4 25

1, 2 (80-90)

3,4 (75-85)
5 (70-80)
6, 7 (ca. 60)
8 (50)

1 (75-85)
2. 3 (80-90)
4, 5 (70-80)
6,7 (65-75)
8 (55-60)

4. July 26

8 (10-15)
10 (40-50)
12, 14 (70-80)

3 40

2, 4 (70-80)
6. 8 (60-70)
10 (ra. 50)
12, 14 (25-35)

2, 4 (70-80)
6 (60-65)
8 (ca. 50)
10 (30-40)
12 (ca. 5)

In three other series (August 1, 2 and 3, 1929) eggs kept previously
for equal times in both KCN-containing and oxygen-free sea-water
(respectively 2 1/4, 3 and 4 1/2 hours) were exposed to M/260 butyric
acid in sea-water. In all cases exposures of 1, 2 and 3 minutes gave
numerous blastulae (60 to 80 per cent in five of the six experiments);
exposures of 4, 5 and 6 minutes were less effective, and with 8 minutes

8 Two cc. M/10 butyric acid plus 50 cc. sea-water. Oxygen-free sea-water was
used in mixing solution (b) and hydrogen was bubbled through the mixture for some
hours.

298 RALPH S. LILLIE

almost no blast uhe were formed. The control eggs gave no blastulae
with the two-minute exposure, few with 4 minutes and a maximum
(60-80 per cent) with 6 to 8 minutes.

( >ne uniform feature of these experiments was that the eggs, after
j>n ili mged suppression of oxidative metabolism, not only gave complete
activation with brief exposures to acid, but also showed a less sharply
defined optimal period of exposure; e.g., in the experiments of July 25th
and 26th exposures ranging from one to 8 minutes were almost equally
effective in producing blastulae (65 to 85 per cent). This behavior
differs strikingly from that of eggs activated in normal sea-water soon
after removal ; in this case the curve relating percentages of blastulae to
durations of exposure rises rapidly to a maximum — usually at 8 to 10
minutes at this temperature (21°-22° C.) and concentration of acid—
and declines equally rapidly, appearing nearly symmetrical.9 l"nder
the anaerobic conditions just described, over-exposure seems less dele-
terious; the optimum is reached early, and with further prolongation of
the activating treatment there is a more gradual decline of effect; i.e.,
over-activation is caused less readily, an indication (possibly) of a
partial depletion of activable or precursor material during the interval
in oxygen-free sea-water. A similar prolongation of the optimum is
seen in eggs remaining for some hours in KCX-sea-water previous to
activation.

Controls of sperm-fertilized eggs were also kept during these experi-
ments: at the end of the period in oxygen-free or KCN-containing sea-
water eggs were transferred to normal sea-water, washed by changing
the latter, and immediately fertilized by sperm. In general, respon-
siveness to sperm fertilization was found to run closely parallel with
activability by artificial agents; i.e., normal fertilizability is also pro-
longed for several hours by exposure to anaerobic conditions. Typi-
cally, sperm-fertilization is complete and uniform only during the
prematurational period and falls off rapidly with the separation of the
polar bodies (Lillie, 1908). Both kinds of prolongation of responsive-
ness are to be regarded as expressions of the same kind of change in the

: in terms of the foreyoin^ hypothesis, the precursor material, which
is transformed into activating substance under the influence of both
acid and sperm, is retarded in its breakdown or removal by the sup-
pression ot oxidative metabolism.

->. Influence of Oz-lack and KCN on Activation by llcnt
The effective exposures to activating temperatures (e.g., 32°) are
similarly shortened by previous exposure to cyanide or lack of oxygen.

9 Cf. the curvi - in m\ .mi. h cm activation by acids in Jour. Gen. Physiol., 1926,
Vol. 8, p. 339.

ACTIVATION OF STARFISH EGGS

299

In general, however, I have found that in the case of heat activation the
decline in responsiveness during such exposure is more rapid than with
acid activation. This may indicate that the material which gives rise
to the activating acid (presumably carbohydrate) undergoes compara-
tively rapid depletion with time. In some cases, the eggs have pre-
served good responsiveness to heat for as long as 3 hours in KCN-con-
taining sea-water, although usually the response is greatly diminished
by this time. Apart from this difference the general phenomena are
closely similar.

The procedure was the same as in my earlier experiments on heat
activation (Lillie, 1908, 1915). The eggs, together with a small quan-
tity of sea-water, were placed in a beaker (or small flask) to which
100-200 cc. of sea-water at the required temperature (32°) was added.
The temperature was kept constant by a water bath. At intervals
eggs were transferred to a series of dishes containing normal sea-water.

The following experiment (August 6, 1929) is cited as an illustration.
One portion of starfish eggs (A} was placed, one half hour after re-
moval from the animals, in M/1000 KCN in sea-water; another portion
(B} in oxygen-free sea-water; in this case one to 2 cc. of a dense egg
suspension was added to 100 cc. of oxygen-free sea-water contained in a
flask through which hydrogen had flowed overnight, and a slow flow of
H2 was continued. After 3 hours eggs from both lots were treated with
normal sea-water at 32° for the times indicated. It had previously
been found that eggs left in normal sea-water for 2-3 hours, and then
warmed, show only partial response and never form blastulae (Lillie,
1908).

TABLE VI

Percentage of eggs forming blastulae after exposure to 32° C.

for the times indicated (minutes)

Treatment

1

2

3

4

5

6

7

8

1. Control: warm sea-

water, 30 min. after

removal . . .

0

1

2-4

25-35

75-85

75-85

60-70

25-35

2. (A) In KCN-sea-

water 3 hrs., 10 min.

before warming.

10-15

25-35

ca. 50

25-35

20-30

1-2

0

0

3. (B) In 02-free sea-

water 3 hrs., 10 min.

before warming

ca. 1

ca. 5

20-25

30-35

5-10

1-2

0

0

The preservation of responsiveness and the shortening of the effec-
tive exposures are both well shown in this series. In six other similar

RALPH S. LILLIE

experiments during tlie same season, with exposures of one hour or
more to l\( .\, the effective exposures were similarly decreased, but
few IT eggs formed blastulu?. \\'ith a briefer previous stay in KCN-con-
taining-sea-water (19 to 24 minutes in five experiments between June
25 ami July 1, 1930), the decrease was less marked but still definite.
The actual exposure to heat may be made either in normal or in KCN-
rontaining sea-water with the same result. Even without a previous
stay in KCN-containing-sea-water, eggs warmed in KCN-containing or
oxygen-free sea-\vater show a distinctly briefer optimal exposure than
eggs ot the same lot warmed in normal sea-water. Table YII gives a
Mimmary of experiments illustrating these effects.

TABLE VII

Procedure

Percentages of eggs forming blastulae after exposure to 32° for

the times iiuiicateil (minutes)

1

2

3

4

5

6

7

8

9

1. June 25, 1930.

(a) ( "out nil: nor-

0

0

0

0

10-15

10-15

20-25

30-35

ca. 10

mal sea-water at

(no

( few

(all

32°

mem-

ineiii-

form

branes)

branes)

mem-

branes)

(b) No previous

0

0

0

10-20

20-30

35-40

5-10

0

exposure to

(me m -

KCN; KCN-

lira nes

containing sea-

in

water at 32° at

nearly

same time as

all

control

eggs

2. June 2S, 1930.

(a) < diitrol:

0

0

0

I <!. \

25-35

60-65

ca. 60

10-20

imrmal >ea-

no

on-,

(all

water at 32°

mem-

mem-

mem-

bra "

branes) bram

(l>) I'revimi

0

0

5-10

15 25

60-70

20-30

1

0

poMirc t'i l\' N

(all

M-a-w.iier (Or 1 7

form

minute-; 1 ln-n di-

mem-

rect to KCN

bra I:

water at 32°

Lark ol oxygen ha- the -aine effect as KCN upon the rate of heat
activation. One uniform and readily observed effect is the shortening
ot the exposures n-<|iiired for nieml irane-formatioii. Typically eggs
exposed to normal sea \\ater at 32° for one minute do not form mem-

ACTIVATION OF STARFISH EGGS

301

branes or show any other signs of activation; exposures of 2 1/2 to 3
minutes are required to form membranes in a majority of eggs. But
if eggs are warmed in KCN-containing or Og-free sea-water (es-
pecially after previous exposure to either medium) for one minute,
all (or nearly all) form membranes (cf. Table VII). Restoration of
normal oxygen-consumption reverses this sensitivity, — i.e., eggs ex-
posed as above to oxygen-free or cyanide-containing sea-water and
then returned for a few minutes to normal sea-water are found to re-
quire 2 or 3 minutes at 32° for membrane-formation, like previously
untreated eggs.

Membrane-formation is an index of partial activation and is a
regular effect of brief exposure to either heat or acid, as well as to other
parthenogenetic agents such as solutions of cytolytic compounds and
pure isotonic alkali salt solutions. Complete activation, however, is
obtained only with heat and acid,10 but with these agents the exposures
required for the formation of blastula? are several times longer than
those required for membrane-formation alone. In confirmation of the
hypothesis that heat acts indirectly through the intra-cellular produc-
tion of acid, the following experiment is cited. Eggs were exposed to
neutral, acid and alkaline artificial unbuffered sea-water (van't Hoff's
solution) u at 34° as indicated in Table VIII.

TABLE VIII

Solution in which

Times of exposure (minutes} and percentages of resulting blastulae

(34°)

1

i'j

2

2 i .,

3

3l/2

4

41;

3

1. Neutral van't

0 (ca. 50%

2-3

20-30

65-75

50-60

ca. 50

15-20

10-15

1

Hoff

membranes)

2. van't Hoff plus

0 (40-50%

ca. 5

30-40

ca. 90

ca. 90

75-85

20-30

1-2

0

N/1000 NaOH

membranes)

3. van't Hoff plus

25-30 (ca.

55-65

75-85

20-30

1

0

0

0

0

N/1000 HC1

100%

membranes)

Neutral or moderately alkaline balanced isotonic salt solution at
34° has the same action as sea-water at this temperature, while acidula-
tion to this degree nearly doubles the rate of activation. This is

10 I.e., apart from hypertonic sea-water (cf. next section), which apparently acts
indirectly through its influence on metabolism. In numerous experiments with
cytolytic compounds (e.g., ether and chloroform), pure Na-salt solutions, ultra-violet
radiation and the electric current, I have never obtained more than partial activation,
whatever the duration of exposure.

11 Made by mixing M/2 solutions of the following salts in the proportions by
volume: 100 NaCl, 7.8 MgCl2, 3.8 MgSO4, 2.2 KC1, 2.0 CaCl2.

302 RALPH S. LILLIE

apparent ly a summation effect resulting from a partial penetration of
the external acid to the site of activation. The increased rate of ac-
tivation observed in KCN-sea-water may similarly be regarded as a
elimination effect, due to the increased intracellular formation of acid
under the conditions of asphyxiation. \Yith regard to the reversal of
hypersensitivity just described, it may be assumed that readmission of
oxygen restores the egg to its original condition through the oxidative
removal of the surplus of acid, as in the case of the lactic acid formed in
muscle cells under asphyxia.

Experiments in which the eggs were placed for 19-20 minutes in
sea-water saturated with pure oxygen from an oxygen-cylinder and
then warmed to 32° in the same medium showed no significant differ-
ence from the control. (Results of three series: July 15, 16 and 18,
1 ('30.) Apparently an increase of oxygen tension to 5 times the normal
does not affect the rate of heat activation.

4. Influence of 0*-lack and KCN on Activation by Ilypcrtonic Sea-Water

The case of hypertonic sea-water is peculiar. This agent, acting
alone, is not so effective with starfish eggs as heat or fatty acid, and its
rate of action is much slower. Nor is it so effective with these eggs
as with Arbacia eggs. It is unusual to obtain 50 per cent of free-
swimming blastula- with hypertonic sea-water alone, although in a few
experiments the proportion has reached 70 per cent or even higher.
The most favorable procedure is to expose the eggs during the pre-
maturational period to a mixture of 100 volumes sea-water plus 20
volumes 2.5 M XaCl for 1 1/2 to 2 hours (at 20-22°). Exposures dur-
ing the post-mat urational period are much less effective. In experi-
ments with a series of hypertonic mixtures, consisting of 100 volumes
sea-water plus (respectively) 10, 15, 20, 25 and 30 volumes 2.5 M NaCl,
membrane-formation and blastula- were obtained with all but the first
and last. The effective osmotic gradient thus has a well-defined range;
and within this range the optimal exposures show in general a decreas-
ing dur.it ion with increasing concentration, as in I.oeb's experiments
\\ith Strongylocentrotus.12 The best results have been obtained with
the 100 -f- 15 and 100 + 20 mixtures, corresponding to an increase in
oHnol.ir ronreiitration of 50 to 60 per cent. The second of these solu-
tion- was ii>ed in most experiments.

I'" des ,H ting as a single parthenogenetic agent, hypertonic sea-
\\ater 111. iv al>o be used to Mipplement and complete a partial activa-
tion 1,\ heat or fatty acid See l.illie, 1(M5, pp. 284 seq.).

\ characteristic peculiarity in the mode of action of this agent is
- J. Luc].: .\nilii-i.il Parthenogenesis and Fertilization, Chapter XI.

ACTIVATION OF STARFISH EGGS 303

that the physiological change which it produces in the egg during the
period of exposure is closely associated with the consumption of oxygen.
In this respect the action of hypertonic sea-water offers a contrast
to that of fatty acid and heat, both of which agents act independently
of immediate oxygen consumption, besides being more uniform, rapid
and complete in their activating effect.

Loeb showed many years ago that oxygen-free or KCN-containing
hypertonic sea-water was ineffective as an after-treatment in sea-urchin
eggs which had received the initial membrane-forming treatment.13 In
starfish eggs similar conditions are found, except that in order to de-
prive the hypertonic sea-water completely of its activating influence,
it is necessary to expose the eggs previously for some time to cyanide
or lack of oxygen. Eggs transferred from normal sea-water directly
to oxygen-free or KCN-containing hypertonic sea-water always show
some partial activation (i.e., membrane-formation and occasionally a
few blastulae) ; while if they are kept first in oxygen-free (or KCN-
containing) isotonic sea-water for 2 to 3 hours and then exposed to the
oxygen-free (or KCN) hypertonic sea-water for the usual time (1 1/2
to 2 hours) the great majority show no signs of activation. If such eggs
are then returned to sea-water and fertilized with sperm, a large pro-
portion develop to larval stages, showing that they have remained
essentially unchanged in their properties.

These conditions are illustrated by the experiments described in

Tables IX and X.

TABLE IX

June 26, 1929. The eggs were placed, 40 minutes after removal, in (.4) hyper-
tonic sea-water of the composition 100 volumes sea-water plus 20 volumes 2.5M
NaCl, and (B) in the same solution plus M/1000 KCN. Temperature 21° C. At
intervals of \]/2, \% and 2 hours portions were transferred to normal sea-water.

Duration of exposure
(hours')

Percentage of mature eggs forming blastulae

A. Hypertonic
sea-water

B. Hypertonic sea-water
plus M/1000 KCN

1H

70-80

0

\%

60-70

1-2

2

60-70

ca. a

The eggs in this experiment were exceptionally responsive to hyper-
tonic sea-water; it is in fact unusual to obtain any blastulae after treat-
ment with KCN-containing hypertonic sea-water. In a similar experi-
ment on June 27 the eggs exposed to cyanide-free hypertonic sea- water

13 Loc. cit., Chapter XI.

304

RALPH S. LILLIE

for 1 1 2 hours .it 2J?0 formed 25 per cent of blastukv, while with KCN
pre-ent, although all formed membranes and many cleaved irregularly
or fragmented, no blastuke were obtained. ()ther experiments gave
similar results.

The effect of placing in oxygen-free or KCN-containing sea-water
for some time previous to the hypertonic treatment is shown in the
following experiments (August 1, 1929).

T\BLE X

I': :

Rrsillt

A. Eggs exposed to hypertonic sea-vater

(100 + 20) alone, beginning 30
niin. after removal, for 1 ' •> and 1%
hours (23°)

B. Eggs in 02-free sea-water for 2 hours,

beginning 30 min. after removal;
thence to 02-free hypertonic sea-
water for \Yi and \% hours

C. Eggs in sea-water + M/1000 KCN

for 2 hours, lic^innin^ 35 minutes
after removal; thence to KCN-
containing (M/1000) hypertonic
sea -water for 1 ' _< and 1 :! [ hours

Activation in all; 25-35 per cent H.I--
tulae

C.reat majority show no activation (no
membranes); partial activation in a
few

Similar to Experiment H. No activation

except in .Miiall minority

A similar result was obtained in eight other experiments of the same
kind, live with KCN-containing and three with oxygen-free hyper-
tonic sea-water. It is evident that when the oxidative metabolism of
the egg is suppressed, the abstraction of water by hypertonic media
has little or no activating effect. The period of asphyxiation does not
in itself deprive the eggs of responsiveness to hypertonic treatment,
provided the oxygen is later restored. Eggs returned after 2 to 3
hours in KCN-containing sea-water to normal sea-water, washed
thoroughly in sea-water and then exposed to hypertonic treatment
show typical activation with production of blastuhe. As already
shown, such eggs also retain their responsiveness to activation by heat
and a< id and to sperm-fertilization.

Tin OKI 1 1. \i.

lining that the activation proce-ss consists in a regular sequence

"I interconnected physical and clH-iiiic.il processes, leading to some

delinite niodiiic.it ion of the egg-system, we have to inquire how such a

uence can be initialed by a temporary slight rise of acidity. In

general the experimental facts suggest that the lirst step in the chemical

ACTIVATION OF STARFISH EGGS 305

sequence of activation is a transformation (e.g., hydrolysis) of some com-
pound present in the surface layer or cortical zone of the egg, this
change leading directly or indirectly to the formation of some specific
metabolic product ("activating substance"), the accumulation of
which renders the egg capable of automatic development. Just how
such a substance may be regarded as producing its effect is uncertain;
it might contribute to special cell-structure, or alter already existing
structure (e.g., increase permeability), or it might be a catalyzer, or
form some necessary link in the chain of structure-forming reactions.
On the whole the last supposition seems the most likely one. Exam-
ples of the control of morphogenetic processes by special metabolic
products are well known (Y/., the case of hormones) ; and they show at
least that the hypothesis of an activating substance is consistent with
the general facts of developmental physiology.

It seems unlikely that the primary or releasing reaction (or reac-
tions) of activation can be widely different from those met with in
other types of reactive cell. In general the basic chemical processes
of protoplasm (as well as its basic chemical composition) show a well-
defined uniformity throughout the range of living organisms. Thus
we find that the carbohydrate metabolism of yeast cells resembles
closely that of muscle cells, with certain divergences in detail; the
utilization of oxygen by living cells depends on factors of a kind uni-
versally present in protoplasm (r/., the heme and sulphydryl com-
pounds) ; the course and end-products of protein metabolism are similar
in their general character. Such considerations lead us to expect that
the key reactions of activation may be simple and of a type already
known.

The prompt arrest of the activation reaction when eggs are re-
turned from acid to normal sea-water, and its equally prompt renewal
on re-exposure to acid, are facts suggesting that the rise of acidity
within the egg acts by releasing from combination some compound
which is necessary for the specific reaction of activation, and that this
releasing reaction is reversed on return to or toward neutrality, the
compound re-entering combination and becoming again unavailable.
The releasing reaction itself has apparently the character of a hydroly-
sis, and the dependent specific reaction which forms the activating
substance is to be regarded as keeping pace with this hydrolysis.

Certain readily hydrolyzable compounds widely distributed in
cells show a behavior of the general kind required by such a hypothesis.
Of chief interest are the phosphate compounds or esters, especially the
guanidin phosphates (phosphagens), which appear to be of special
importance in the metabolic reactions associated with excitation in

306 RALPH S. LILLIE

muscle. The enzymatic hydrolysis of these compounds in muscle
extract^ N Influenced by variations of acidity in a manner closely
paralleling the characteristic variation in the rate of activation of
eggs with concentration of acid. The creatin phosphate of vertebrate
muscle extracts splits into creatin and phosphoric acid at a slightly
acid minion (pH = 6.4), and is resynthesized at slight alkalinity
i pi 1 8.5 9) ; the analogous arginin phosphate of invertebrates shows
similar properties, but has a greater tendency to resynthesis at slightly
alkaline or even neutral reaction.14 The total phosphate content of
starfish eggs is very similar to that of muscle cells (Page, 1927), and it is
possible that in the activation and other reactions of these cells phos-
phate esters play a role of the same kind as in the analogous reactions
of muscle and presumably other cells. Meyerhof has suggested that
a special biological role of the guanidin compounds may consist in the
binding and splitting off of phosphate groups (Meyerhof and Lohmann,
1928), a view in accordance with the general importance of phosphates
in cell metabolism. The hypothesis that the splitting of phosphagen
or some analogous compound is the primary chemical event in activa-
tion requires, however, further experimental testing from both the
physiological and the biochemical sides.15

\Vith regard to the activation by hypertonic sea-water, it is to be
noted that abstraction of water from the egg means increase in concen-
tr.it ion of egg-constituents, and it is possible that locally (e.g., in the
cortical region) this increase may be considerable. Conditions favor-
able to the synthesis of complex compounds would thus arise. The
general theory of the dehydrolytic synthesis of complex molecules
through the combination of a number of smaller molecules has recently
been reviewed by \Yasteneys and Borsook (1930, 1931) in relation to
their work on the enzymatic synthesis of plasteins from the products
of peptic hydrolysis. The authors show both theoretically and experi-
mentally that increase in the concentration of such a digest may change
the prevailing direction of the reversible reaction from a hydrolysis to
5 nt hois; and it is reasonable to assume that conditions of the same
ieral kind exist in living cells. In the case of the starfish egg the
inference would be that hypertonic sea-water acts by promoting the
synt he-is of some compound (e.g., protein) which either is, or gives
rise to, the activating substance.

Mi- nd I.olmumi, Hiochem. Zeitschr., 196, 49, 1928. It might be ob-

ted thai onsare too slow to account for the rapid course of the activation

linn in si.irti>h rggs imnnT-i-d in arid sea-water, and its rapid arrest on return to
normal ter; Inn it i> well-known that many enzymatic reactions which are

>l"v. I r.ipidly in thr living cell.

, ii would In- import. mt to determine if the free phosphate is increased in
the egg during .n tivation, .1- in tin- stimulation of muscle and nerve.

ACTIVATION OF STARFISH EGGS 307

It is significant that the formation of the activating substance
appears to be promoted by two quite different external changes of
condition, hypertonicity and moderate acidity. As already described,
hypertonic activation proceeds relatively slowly as compared with
acid activation and is dependent on oxygen consumption. The essen-
tial problem is why dehydration combined with oxidation should lead
to the production of the same substance as slight local increase of
acidity. Experimentally we find that the two parthenogenetic proced-
ures may act additively or supplement each other; according to the
present hypothesis, however, they influence metabolism in different
ways, the one promoting synthesis of a complex compound, the other
releasing hydrolytically some relatively simple compound which enters
further combination. The fact that both procedures have the same
final physiological effect suggests that the activating substance is
formed by the combination of the two products. In such a case, ac-
cording to the mass action rule, increase in the concentration of the
one product would offset deficiency in the other; i.e., activation would
result from a sufficient increase in either compound, provided some of
the other were available. This would explain why either acid or hy-
pertonic sea-water acting alone may cause complete activation, since
both compounds may be assumed to be present in low concentration
in the resting egg. Whether the additive relations between the two
parthenogenetic procedures conform to such a hypothesis in detail
can be determined only by further experiment.

SUMMARY

1. Suppression of oxygen-consumption in freshly removed unfertil-
ized starfish eggs (by exposure to cyanide-containing or oxygen-free
sea-water) prolongs by several hours the period during which they re-
main responsive to artificial activation by heat, acid, or hypertonic
sea-water (as well as to sperm fertilization). The possible interval
between a first partial and a second completing activation may be
similarly prolonged.

2. During the exposure of the eggs to these asphyxiating conditions
their susceptibility to activation by heat or acid, as indicated by the
effective durations of exposure, undergoes a progressive increase.

3. Eggs kept for some hours in cyanide-containing or oxygen-free
sea-water and then exposed to acid or heat while immersed in these
media show normal activation. On the other hand, a similar suppres-
sion of oxygen consumption prevents activation by hypertonic sea-
water.

4. It is suggested that the activation by hypertonic sea-water de-

RALPH S. LILLIE

pends on the increased intracellular production, by dehydrolytic syn-
thesis, of some complex specific compound (e.g., protein); while in the
anarroliir activation by acid (or heat) the critical change is a hydrolysis
. of ,i phosphagen compound), yielding a product which combines
with the complex compound to form the specific activating substance.
The accumulation of this substance to a certain definite level deter-
mines activation. Two metabolic processes, respectively aerobic and
anaerobic, would thus cooperate in activation. The fact that either
hypertoiiic sea-water or acid (or heat), acting alone, can produce the
same physiological end-effect, complete activation, is shown to be con-
sistent with this hypothesis.

BIBLIOGRAPHY

JUST, E. E., 1930. The Present Status of the Fcrtilizin Theory of Fertilization.

Protoplasma, 10: 300.

LILLIK, V . k., 1919. Problems of Fertilization. University of Chicago Press
LILI.II-, K. S., 1908. Jour. Exper. Zool., 5: 375.
LILLIE, R. S., 1915. Biol. Bull., 28: pp. 260 and 284 seq.
LILLIE, R. S., 1917. Biol. Bull., 32: 131.
LILLIE, R. S., 1926. Jour. Gen. Physiol., 8: 339.
LILLIE, R. S., 1927. Jour. Gen. Physiol., 10: 703.
LOEH, J., 1902a. Biol. Bull., 3: 295.
LoEii, J., 19026. Arch. ges. Physiol., 93: 59.
LOEH, J., 1913. Artificial Parthenogenesis and Fertilization. University of Chicago

Press. Cf. Chapter IX, pp. 78 seq. Also Chapter XI.
LYON, E. P., 1903." Am. Jour. Physio!., 9: 308.
MEVEKHOI-, O., 1928. Arch, di Sc. biol., 12: 536.

Mi M KHOF, O., 1930. Die chemische Ynr^in-e ini Muskel J. Springer, Berlin.
Mi M kin IK, O., AND K. LOHMAXN, 1928. Biochem. Zeitachr., 196: pp. 22 and 49.
PAGE, I. II., 1927. Biol. Bull, 52: 168.
WASTENEYS, II., 1931. Biol. Bull., 60: 1.
\\ASII.\I.VS, II., AND II. BORSOOK, 1930. The Enzymatic Synthesis of Protein.

Physiol. Rev., 10: 110.

MODIFICATION OF THE RATE OF OXYGEN CONSUMP-
TION BY CHANGES IN OXYGEN CONCENTRA-
TION IN SOLUTIONS OF DIFFERENT
OSMOTIC PRESSURE

J. WILLIAM BUCHANAN
DEPARTMENT OF ZOOLOGY, NORTHWESTERN UNIVERSITY

INTRODUCTION

It has not been possible to derive a general rule describing the
relation between rate of oxidations in organisms and the oxygen
tension of the environment. Pfliiger (1872), reasoning from observa-
tions on the oxygen-absorbing power of the blood and the partial
pressure of oxygen in the alveoli of the lungs in the dog, concluded
that the rate of oxygen absorption of the tissues is independent of the
oxygen tension over a wide range. Thunberg (1905), testing the
oxygen absorption of Lumbricus, Limax, and Tenebrio larvae in
different oxygen tensions, found that the rate varies with the oxygen
tension. Thunberg assumed that the oxygen tension within the cells
is less than that of the milieu, but not zero, and concluded that his
results were in accord with the fact that the velocity of a reaction
depends on the number of molecules reacting; hence the velocity of
oxidations in the organism varies with the oxygen concentration of
the environment until a maximum is reached. Henze (1910), from
his experiments with pelagic forms, concluded that in organisms with
thin and delicate tissues the oxygen tension of the cells is that of the
surrounding water and always in excess of requirements; consequently
in such organisms the rate of oxygen utilization is independent of the
oxygen tension of the water.

Krogh (1916) amplified Henze's conclusion and states that whether
or not the rate of oxygen consumption is independent of the oxygen
concentration of the environment depends on two conditions: In
animals with well developed blood streams and respiratory systems,
and in animals with a low percentage of dry substance, the oxygen
tension does not affect the rate of oxygen utilization over a wide
range. The assumption is necessary in these cases that the oxygen
tension in the tissues is positive. On the other hand, if the organism
has a high proportion of dry substance and "imperfect" respiratory
and circulatory systems, the rate of oxygen utilization is highly

21 • 309

310 J. WILLIAM BUCHANAN

dependent on the oxygen tension of the environment until the oxygen
ten>ion is above the normal. It is assumed that the oxygen tension
of the tissues of such forms approaches zero.

Hyman (1929) suggests that, in certain cases at least, the body
surface acts as a regulatory mechanism; this appears to act according
to internal requirements and hence in the presence of excess oxygen
the oxidative rate remains constant. Shoup (1929) holds that the
size of the organism is important. In general, according to Shoup,
microscopic forms are small enough to allow complete diffusion of the
dissolved gases. In other words, they are independent of the oxygen
concentration, within limits, because oxygen is always present in
excess within the system. Shoup also adopts the view expressed by
Krogh (1916) that in large organisms the rate of respiratory exchange
depends on oxygen tension and the adequacy of the respiratory and
circulatory systems.

Thus far it is not possible to explain the experimental evidence
at hand by any or all of these theories. According to Krogh, the
water content of the organism is an important characteristic in
determining its response to changes in the oxygen tension of the
environment. Thus an organism with relatively low water content
and what Krogh calls "imperfect" systems of respiration and circu-
lation might be expected to consume oxygen in direct dependence on
the oxygen concentration of the environment. In the case of Planaria
dorotocephala certainly the mechanisms of respiration and circulation
are, so far as known, simple and rudimentary; yet Hyman (1929)
found that this form has a constant rate of oxygen utilization over a
range of environmental oxygen tension between 3 cc. and 9 cc. per
liter. If Krogh 's suggestion has a basis in fact, one would expect
that the water content of this animal is relatively high.

Plannrin afford suitable material for a further study of the relation
between water content and the degree of dependence of rate of oxygen
« oiiMimption on the oxygen tension of the milieu. More specifically,
the problem at hand is: What is the mechanism that renders the rate
ol oxygen utilization of this animal indifferent to changes in the
oxygen tension of the surrounding medium? Krogh 's suggestion
indicate^ the point of attack, namely, the relation between water
content and I he effect of changes in oxygen tension on the rate of
n utilization \\ith changes in free water content of the organism.

OXYGEN CONSUMPTION 311

METHODS l

Investigation of this problem resolved itself into (a) finding the
total free water content of Planaria dorotocephala; (6) increasing and
decreasing the water content by means of hypo- and hypertonic
solutions; (r) determining the rate of oxygen consumption of the
animals in different concentrations of oxygen in natural water, and in
hypo- and hypertonic solutions.

The free water content was determined by weighing the animals
after removal from water, then drying over sulphuric acid in a desic-
cator, and weighing repeatedly until no further decrease in weight
occurred. The first weighing was accomplished as follows: The
animals were poured in water into a properly folded hard-surfaced
filter paper, the water drained off, and the paper drawn across a
plate of glass until it adhered firmly. The animals were then picked
up with a thin spatula and placed in a weighing tube, the tube closed,
and the weighing completed as rapidly as possible. The proceeding
was timed so that the exposure of the animals to air was approximately
the same in different experiments. Weighing the animals while wet
introduces the error of the weight of water adherent to their surfaces.
Considerable success was realized in standardizing this error, for
during exposure to sulphuric acid the animals lost very nearly the
same percentage of total weight in all experiments; which indicates
that the original proportion of adherent water must have been much
the same in each first weighing. The total loss of weight as shown
by the data must be discounted by the weight of adherent water at
the first weighing. The actual water content of the animals is there-
fore somewhat less than the data indicate.

The degree of dryness attained by desiccation over sulphuric acid
depends on the relative affinities of the acid and of the animal tissues
for water. The method was used by Davenport (1899) in his analysis
of the changes in water content of the frog embryo during development.
For the present work the method is more satisfactory than heating
to dryness, for only free water is withdrawn from the organism.
Heating to dryness involves loss of bound water as well as other
substances. It is not assumed, however, that drying over sulphuric
acid actually removes all the free water from the system, for the
resistance of aquatic organisms to desiccation is common. However,
the proportion of water remaining after the treatment must be very
small; the residue is a brown mass that on pressure breaks up into a
coarse granular powder.

1 The experimental work on which this article is based was done in the Osborn
Zoological Laboratories, Yale University. Acknowledgment is here made for the
facilities and assistance provided.

312 J. WILLIAM BUCHANAN

The hypo- and hypertonic solutions employed throughout this
work were distilled water and Ringer's solution made up with buffer
and in concentration that is described as isotonic for amphibian tissue.

The effect of distilled water on the weight of Planaria dorotocephala
and the analysis of the distilled water used are given in a former paper
(Buchanan, 1930a). It was also shown that these animals in some
way condition distilled water; accordingly in this work the proportions
of animal tissues to distilled water were made approximately the same
in all experiments. Since the changes in weight in hypo- and hyper-
tonic solutions constitute an important phase of the problem at hand,
the method employed in obtaining the data given in the former paper
is restated.

Twenty-five animals of medium length were weighed and placed
in 250 cc. of distilled water, then weighed again after certain intervals.
The hydrogen ion concentration of the distilled water was reduced
by the addition of a minute amount of NaOH. The experiments
were conducted at room temperature, 20° to 23°. Control lots were
weighed after similar handling in tap water and the difference in
changes in weight between the distilled and tap water lots is regarded
as due to the effect of distilled water on weight. The uniformity of
the results was seriously interfered with by the fact that the distilled
water causes the animals to exude slime. It is difficult to remove
this slime during the weighing process. Furthermore, partial drying
on filter paper and the incidental exposure to air causes a withdrawal
of water from animals that have taken up water from a hypotonic
medium. Thus the data show wide variations in weight changes in
distilled water. It is clear, however, that after four hours in distilled
water, pH 6.8 to pH 7.6, the weight increases. Consequently for
respiration tests I employed animals that had been pre-treated with
distilled water for from four to six hours.

The effect of Ringer's solution on the wreight of Planaria doroto-
ifblmla is also given in a former paper (Buchanan, 19306). Ringer's
MI] ui ion was employed because in the writer's experience the animals
live longer in this hypertonic medium than in others that have been
tried. At 20° medium sized Planaria will live for several days in
Ringer's ^.Union. Data on the effect of Ringer's on the weight of
tin- .inimals were obtained by methods similar to those employed for
di-tilled \\ater. The results are fairly uniform, for the animals do
ii"t exude Clinic. It is shown that the weight decreases continuously
• lining tin- tn>i four hours of exposure, but after that time further
l"v- (.| \\ci-ht i> >li^U(. Hence, for tests of the effect of loss of water
lr»m tUe .mini, ils on their rate of oxygen consumption under different

OXYGEN CONSUMPTION 313

conditions of oxygen concentration, I employed animals that had
been exposed for four hours previously to Ringer's solution.

The oxygen content of the various liquids was determined by the
Winkler method. This method has recently been criticised by Shearer
(1930). His criticisms are based in kirge part on his statement that
when present in numbers a European planarian discharges substances
which should, according to Shearer, absorb iodine and thus interfere
with the titrations. Shearer assumes that this possibility has not
occurred to us. The fact is rather elementary that in iodometric
work the possible absorption of iodine by substances in the solution
must be taken into account. Years ago Hyman (1919) showed
conclusively that the substances which Planaria dorotocephala excretes
do not absorb iodine. Further evidence is to be found here. Planaria
exude slime in distilled water but none at all in Ringer's solution,
yet the titrations show that the oxygen consumption of the animals
is materially less under comparable conditions in distilled water than
in Ringer's.

Shearer also criticises work done with Planaria by this method on
the grounds that the motor activity of the animals is not controlled.
Shearer himself attempts to draw conclusions concerning the motor
activity of an American flatworm which he probably has never seen,
from quite unconvincing evidence obtained from a European form.
If the head of Planaria dorotocephala is removed the animal continues
to crawl about for some minutes. Then it comes to rest in a contracted
state and if undisturbed does not change position for many hours.
This is also true of Planaria maculata but not true of Phagocata gracilis.
A comparison of the rate of oxygen consumption of beheaded Planaria
dorotocephala with that of intact animals brought to rest by other
methods, namely, lowering the temperature slightly and excluding
light and other stimuli, shows no appreciable difference (Hyman,
1919a, 19196; Buchanan, 1926). It has thus far been impossible in
this, as in much of the work on standard metabolism, to exclude local
muscular twitching and the action of cilia, without recourse to profound
anesthesia. If the work on Planaria is invalidated by these internal
and localized motor activities, then the same standards must be
applied to a very considerable proportion of the work on respiratory
metabolism, including Shearer's. In the present work the factor of
motor activity of the animals is not seriously concerned, for the
respiratory differences are of such nature and so conspicuously different
that they cannot be explained on the basis of differences in motor
activity.

The possibility of the salts in Ringer's interfering with the titrations

314 J. WILLIAM BUCHANAN

required safeguarding. Since only the relative quantities of oxygen
in the medium before and after the animals had consumed some of it
were of major importance, the action of salts may be discounted so
loiu as it permits a sharp end point. To assure the accuracy of
tit i alii in, a casserole containing the same quantity of Ringer's as the
sample to be titrated was placed beside the casserole containing the
sample. To this Ringer's was added sufficient iodine to match the
color of the sample. It was then cleared with sodium thiosulphate.
This served as a check end point, and the sample was then cleared
to the same state of decoloration as the check. The end point in
Kin tier's was quite distinct and it was found unnecessary to provide
this check as a constant procedure.

A much more difficult problem was the control of motor activity
so that standard conditions obtained in the several environments.
In part this was solved by the fact that Planaria dorotocephala, when
changed from light to darkness, or when changed to a somewhat
lower temperature, remain quiet for some hours. In this work both
of these methods were employed. In both tap and distilled water
the factor of motor activity does not enter, for in both the animals
were quiet throughout the tests, except for the very brief time at the
start, when the animals were distributing themselves about the interior
of the flasks. In Ringer's solution the animals writhe for some time,
then become paralyzed, accumulating at the bottom of the flask in
abnormal positions. After two hours or less only occasional spastic
movements occur unless light is admitted or the flask agitated. Un-
questionably such tw-itching results in the consumption of oxygen.
If one may judge from observations on the extent of such localized
muscular activity, the amount of oxygen required for this movement
must be very small as compared with the total oxygen consumption.
Furthermore, there is no evidence in the observations that the extent
'it such movements increases with the increase in oxygen concentration
ot the Ringer's. After careful observation of the behavior of the
animal^ as compared with the oxygen consumption under the several
conditions, one must conclude that the dillerences in oxygen consump-
tion cannot be ascribed to differences in motor activity.

In carrying out a titration the quantity of clearing agent used in
reaching the end point is in practice slightly in excess or else slightly
under tin- exad quantity. Also in the necessary manipulations no
U\o lots «,| animals can IK- treated in exactly the same fashion. In
"ider t" compensate for such variations in procedure that result in a
spread in the results, duplicate lots of animals were carried simultane-
"ii-lv in all experiments and the results recorded are the average

OXYGEN CONSUMPTION 315

between them. The oxygen consumption of such duplicates varied
in amount, sometimes rather widely. But with one or two exceptions
both members were in the same direction from the control determina-
tions and fell within ranges that are consistent with the general data.

If the oxygen concentration of the medium is reduced materially
below saturation at room temperature, some oxygen will be absorbed
during a two-hour experiment, from the air or from the wall of the
flask, or by leakage through the connections. Likewise, if the oxygen
concentration is higher than saturation at room temperature, oxygen
will be lost during the experiment. Obviously, therefore, determina-
tions of the oxygen content of flasks in which animals had been kept
for two hours had to be corrected for changes in oxygen concentration
not due to the metabolism of the animals. A considerable series of
tests was made in an attempt to determine the magnitude of change
at various oxygen concentrations. The apparent oxygen consumption
of the animals was then corrected in each case in accordance with the
average oxygen gain or loss at approximately the same oxygen tension
in tap H2O in the absence of animals. Admittedly this does not
yield an absolutely accurate result, for the control tests just described
showed that the amount of oxygen gain or loss is subject to some
variation. However, this appears to be the most accurate and
convenient method at hand for establishing the necessary correction.
A tabulation of the average oxygen gain or loss at various oxygen
tensions which was used as a table of corrections is given as Table III.

A typical experiment to test the effect of differences in osmotic
pressure of the medium on the response of the organism to different
oxygen concentrations was carried out as follows: Twenty-four hours
before the test, if distilled water was to be used, a large quantity of
distilled water was treated with carbon dioxide-free air, the water
being treated for at least twelve hours. The H ion concentration
was thus decreased to pH 7.0 prior to use in the experiment. A
selected lot of animals was divided into two groups and their rates of
oxygen consumption in tap water were taken over a two-hour period.
In all cases the oxygen concentration of the tap water was at or near
saturation at 20°; the H ion concentration being always lower than
pH 7.0 but never as low as pH 7.8. The amount of oxygen consumed
per two hours under these conditions is regarded as the normal and in
comparison with the effects of other conditions is taken as 100 per cent.
The animals were then placed in 500 cc. of distilled water, or of
Ringer's, as the nature of the experiment required, the oxygen tension
of which was approximately that of the tap water in the control test.
After four or more hours a fresh supply with higher or lower oxygen

316 J. WILLIAM BUCHANAN

tension, but with the same H ion concentration, was siphoned into the
flasks, after appropriate washings, which were then sealed and im-
mersed in the constant temperature bath in darkness. Samples for
titration were taken of the stock solutions and of the contents of the
flasks at the end of two hours.

The oxygen concentration of the stock solutions was controlled
either by bubbling pure oxygen through the stock, or by driving out
most of the oxygen by heating and then cooling in a container so
sealed as to prevent the ingress of air directly to the supply to be
used. The time intervals were accurate to within one minute, the
temperature within 0.5°, and the H ion concentration ranged in the
\ arious experiments from pH 7.0 to pH 7.8. In the case of Ringer's the
pre-treatment with carbon dioxide-free air was not necessary, since
the solution wras buffered to pH 7.8. In all cases the test containers
were Erlenmeyer flasks, capacity 500 cc., in which were placed from
35 to 60 animals, depending on size. Over a two-hour period the
products of the animals increased the H ion concentration of distilled
water approximately pH 0.2, somewhat less in tap wrater, and not at
all in Ringer's. Therefore at no time was the H ion concentration
greater than pH 6.8, nor less than pH 7.8. According to Hyman
(1925), within this range the effect of H ion concentration on oxygen

consumption is slight.

DATA

Table I shows that the water content of Planaria dorotocephala is
approximately 78 per cent of the total weight, as determined by
drying over sulphuric acid. The accuracy of this value is of course
conditioned by the weight of water adherent on the surfaces of the
animals at the first weighing and by the distribution of water between
the drying agent and the organic material. The affinity of the latter
for water being undetermined, it is not justified to assume that all
the free wrater is removed by this method. With these reservations,
it is, however, certain that the water content of these animals is
relatively low. Comparison of these data with those given by Daven-
poi i (1896, pp. 58-59) for other invertebrates shows that the water
<on i cut of /'In mi ria is comparable with that of other soft forms,
although somewhat lower. It is much lower than that of hyaline
forms referred to by Krogh (1916, pp. 58-59). Thus the observed
facts do not support Krogh's hypothesis, that independence of rate
of < i utili/ation of environmental oxygen supply is conditioned

on a low content of dry substance.

The effect of distilled water and of Ringer's solution on the weight
• •I fl<i nurin is given in two former papers (Buchanan, 1930r/, 19306).

OXYGEN CONSUMPTION

31'

TABLE I

Loss of weight of Planaria dorotocephala over sulphuric acid. All animals starved
two weeks before weighing. In milligrams and percentage change.

Length

No. of Animals

Live Weight

Final Weight

Loss

mm.

mg.

mg.

per cent

15

5

46.2

9.8

78

((

8

61.9

14.1

77

II

11

74.6

20.0

73

it

14

109.1

23.7

78

1 1

18

131.2

28.1

78

t <

25

180.5

38.6

78

( i

20

142.4

34.0

75

u

20

134.4

31.7

76

1 1

20

143.7

28.2

80

1 1

20 +

160.5

34.4

78

16

20

196.7

36.7

81

( I

20

194.7

39.2

79

11

20 -

159.7

36.3

77

I *

20 +

223.3

44.7

79

Table II is abstracted from the data there given and consists of
weight changes in experiments in which the conditions were comparable
to those which were employed in testing the effect of these environ-
ments on respiration. As stated above (p. 312), it was found im-
possible to weigh animals from distilled water with accuracy, due to
the adherent slime and the accidental drying that occurs when the
animals are handled on filter paper. The average weight gain shown
by the data is approximately 15 per cent; the organism being 78
per cent water, this represents an increase in total free water of
11 per cent. No great confidence may be placed on these data,
however. At best they merely indicate that water is taken up by
the animal while in distilled water.

TABLE II

Changes in weight of Planaria dorotocephala after four- and five-hour treatment with
Ringer's solution and distilled water. Abstracted from Table III, 1930a, and Table
III, 19306. All animals starved at least one week. 14-18 mm. animals. Twenty-
five animals in 250 cc. of liquid in each case. Room temperature, Ringer's solution,
pH 7.8; distilled water, pH 7.0 to pH 7.6. In percentage change as compared with
controls in tap water.

Ringer's Solution Distilled Water

4 hours

5 hours

4 hours

5 hours

- 17.7

- 17.7

+ 1-9

+ 24.6

- 17.7

- 19.0

+ 3.7

+ 18.6

- 20.9

- 18.5

+ 4.5

+ 16.5

- 16.9

- 14.2

+ 1.9

+ 11.9

- 17.0

+ 0.6

- 2.8

- 10.0

+ 1.5

1.1

- 11.0

- 12.0

Average -- 16%

+ 15%

318 J. WILLIAM BUCHANAN

In the case of Ringer's, the data show clearly that the water loss
amounts to approximately 16 per cent of the total weight, or a reduc-
tion of approximately 20 per cent of the total water content.

Figure 1 shows in graphic form the effects of changes in the oxygen
ten-ion of the medium on the rate of oxygen consumption in distilled

•90
•60
•70
*€0
«50
•40
+30
«20
• 10
100
-10
-20
-30
-40

2 4 6 8 10 12 14 16 18 20 22 24 26 28

FIG. 1. The effect of increase in oxygen tension of the medium on the oxygen
consumption of Planar ia dorotocephala. Abscissae represent cc. oxygen per liter.
Ordinates represent changes in rate of oxygen consumption as compared with rate
in tap water with oxygen concentration at or near saturation at 20°, which is taken as
100 per cent. T, animals in tap water; D, animals in distilled water; R, animals in
Ringer's solution. Temperature, 20° ±0.5°. H ion concentration, pH 6.8 to pH
7.8.

water, tap water, and Ringer's solution. The points on the respective
graphs were determined by consolidating the data as follows: For
each oxygen concentration duplicate lots of animals were used and
the average between them is set down as the result at that particular
oxygen concentration. I'p to 8 cc. of oxygen per liter the determi-
nations within a range of 1 cc. of oxygen per liter were averaged and
plotted. Above 8 cc. per liter determinations of oxygen consumption
\\iiliin each concentration range of 2 cc. per liter were averaged and
plotted. The graphs, then, represent 52 separate determinations of
i consumption in distilled water, 54 determinations in tap water,
ami n<H determinations in Ringer's solution.

The graph of the controls in tap water, C in Fig. 1, shows that
\\ithin the range of oxygen tensions employed the rate of oxygen
• onMimption does not change materially. For a number of years,
in connection with other problems, data have accumulated on the rate

OXYGEN CONSUMPTION

319

TAKI.K III

Average change in oxygen concentration in sealed 500 cc. flasks of tap -water im-
mersed in a water bath at 20° for two hours. Corrections for the data plotted in Fig. 1
were taken from this table. Change in percentage.

Range of Oxygen
Concentration

No. of Experiments

Average Change

cc. Illter

20-30

12

per cent
-0.9

15-19

8

- 0.6

10-14

16

- 0.5

7-9

24

- 0.2

6

4

0.0

5

8

+ 0.1

4

4

+ 0.2

3

12

+ 0.6

of oxygen consumption of Planaria dorotocephala in oxygen tensions
varying around the saturation point of water at room temperature.
The data are in terms of the oxygen consumed per unit of wet weight
per unit of time. At hand are 282 determinations that were made
under comparable conditions; temperatures were between 18° and
24°; 208 experiments were conducted at 20°. The calculations
necessary to determine the relationship between rate of oxygen
utilization and the oxygen tension of the water have been carried out.2
The results are as follows:

Number of experiments, 282.

Highest temperature of an experiment, 24° (3 cases).

Lowest temperature of an experiment, 18° (9 cases).

Average temperature, 20°.

Oxygen concentration of the water:

Extremes, 3.90 and 7.19 cc. per liter.

Mean, 5.87 cc. per liter.

Standard deviation, 0.64 ± 0.018.
Oxygen consumption, per gram per two hours:

Extremes, 0.30 cc. and 0.64 cc.

Mean, 0.37 cc.

Standard deviation, 0.05 ± 0.0014.
Coefficient of correlation, oxygen consumption and oxygen tension,

0.09 ± 0.04.

While a statistical treatment of data of this sort is not free from
criticism, the fact that the coefficient of correlation is not significant,
when considered together with the graphic results shown in Fig. 1,

2 These calculations were made by Miss Rosalthea Sanders.

320

T. WILLIAM BUCHANAN

and the data of Hyman (1929), constitute convincing evidence that
the rau- of oxygen consumption of Planaria in natural water is inde-
pendent of the oxygen tension over a wide range.

1 i-mv 1 also shows that in distilled water after four to six hours
the rate of oxygen consumption is sub-normal and remains fairly
constant UM \\rrn oxygen concentrations of 2.65 and 5.35 cc. per liter.3

50

40

30

20

10

10

12

14

16

18

20

FIG. 2. Rate of disintegration of Planaria dorolocephula in distilled water with
different oxygen content. Ten animals in 500 cc. distilled water in each case. Room
temperature. Abscissa? represent hours; ordinates are derived from arbitral}
values assigned to portion of animals alive at each observation, e.g., 50 indicates that
all 10 within a flask are alive; 40 indicates that some have partially disintegrated.
.1 and B and a and b represent two different experiments. In A the oxygen concen-
tration of the distilled water was 11.94 cc. per liter; in II the oxygen concentration was
o.lo cc. per liter. In a the oxygen concentration was 17.75 cc. per liter; in b the
oxygen concentration was 6.51 cc. per liter.

It increases as the oxygen concentration is further increased, but is
In-low that in tap water until the oxygen tension is increased beyond

an apparent <li» n-p.nu y 1 >et wren these results for distilled water and
those of I less (192*') and those given in an earlier paper by the writer (Buchanan,'
\'>2'>in. The depressive action of distilled water shown here arises from the fact that
tin animals \\c-re prr-e.\ posed, after washing with distilled water, from four to six
h'uir-, ilien washed l.y siphoning distilled water through the flasks for several minutes,
then si-a|i-d in drilled water for two hours, thus bringing the total exposure to dis-
tilled v of ->i\ and eight hours, in large volumes. The earlier results

I1'-"' and tin •-,.• ..| I [ess (1929) agree that long exposure to distilled water lowers
the rate nt" oxygen ut ili/a t ion. The data here given are not in conflict with those
resul

OXYGEN CONSUMPTION 321

9 cc. per liter. In Ringer's solution the fact appears that the rate
of oxygen consumption is sharply dependent on the oxygen tension
of the solution until the tension reaches a high value, approximately
14 cc. per liter. The effect of increasing the oxygen concentration
above 14 cc. per liter becomes increasingly less; the data indicate
that in the highest concentrations employed the rate of oxygen
consumption is somewhat less than in lower concentrations.

Thus it appears that although the animals in natural waters are
independent over a wide range of oxygen concentration, this is not a
fixed condition, but is susceptible of being materially affected by
either the salt content of the water, or by the osmotic pressure, or
by both.

Figure 2 shows that animals in distilled water with a high oxygen
content disintegrate slightly less rapidly than do animals in distilled
water with oxygen content around the saturation point at room
temperature. The method of plotting these data has frequently been
described in detail and will not be re-described here. Reference to
Fig. 1 shows that the rate of oxygen consumption of animals in distilled
water containing corresponding concentrations of oxygen is higher in
the higher concentrations, the reverse of the order of disintegration.
In other words, in high oxygen concentrations the rate of oxygen
utilization is greater while the rate of disintegration is lower than in

normal oxygen tensions.

DISCUSSION

With regard to the original problem, what is the mechanism that
renders the rate of oxygen consumption of this animal indifferent to
oxygen concentration changes in the environment, these results
render certain hypotheses improbable.

In the first place, it is clearly shown that this characteristic, in
this particular animal at least, may be modified by the action of salts
and by the osmotic pressure of the environment. As a matter of
methodology, this fact is of considerable importance. For if one
were to measure the oxygen consumption of an animal in Ringer's
solution, for instance, with the oxygen tension slightly greater than
saturation at 20°, the data would appear to show that Ringer's
accelerates oxygen utilization. On the other hand, should the oxygen
concentration of the Ringer's be somewhat under saturation at 20°,
the results would indicate that Ringer's is a depressant. The effects
of various salts on respiratory metabolism have received considerable
attention, but there is little uniformity in the findings of various
investigators. A partial review of the literature is given by Hess
(1929).

322 J. WILLIAM BUCHANAN

•ondly, the results are important in showing that the absolute
size of the animals, i.e., the surface-volume relation, as suggested by
Shoup (1929), may be rejected as a general factor in determining the
relation between rate of oxygen utilization and oxygen tension.
Also the role played by the degree of perfection of the respiratory and
circulatory mechanisms does not appear to be of general importance;
it i- inconceivable that Ringer's or distilled water in any way converts
these mechanisms in Planaria.

With these hypotheses of limited application, attention must be
focused on the cellular oxidative mechanisms, not sufficiently empha-
sized in Krogh's (1916) discussion. The possibility must be considered
that the relation between rate of oxygen utilization and the oxygen
concentration of the environment is conditioned on the concentration
of oxidative enzymes in relation to the oxygen concentration within
the tissues. In the total absence of any information whatever as to
the oxygen tension in the tissues, the permeability of the tissues to
oxygen, and the effects of Ringer's solution and distilled water on
tissue permeability to oxygen, explanations of the causal factors in
these results must be in large part conjecture. However, sufficient
facts are shown by the data to justify a discussion of certain possi-
bilities.

In the normal animal in tap water, assuming that oxygen is present
and oxidizable substances are in excess in the tissues, the velocity of
oxidations is a product of a constant and the concentration of the
active enzymes. That is,

Velocity = kE.

So long as E remains constant, further increase in the oxygen tension
should have no effect on the rate of oxygen utilization. This consti-
tutes a possible explanation of the fact shown in Fig. 1, that increasing
the oxygen concentration of the tap water has no effect on the rate of
oxygen utilization within the concentration range employed.

But if the value of E is increased, thereby increasing the value
of I', the oxygen concentration in the tissues becomes a limiting factor
until it reaches a value equal to the oxygen-transferring power of the
en/yines as conditioned by the rapidity of adsorption of the substrates.
If the value of E is decreased, V is proportionately decreased and
remains constant so long as the oxygen tension is adequate to maintain
.1 • onstanl adsorption rate.

The d.it.t on both Ringer's and distilled water offer some support
for this interpretation, for it will be noted in Fig. 1 that when the
oxygen concentr.ition reaches a certain value the rate of oxygen

OXYGEN CONSUMPTION 323

consumption tends to become constant. In distilled water this value
is low, between 2.65 cc. and 5.35 cc. per liter, as would be expected
if the cell enzymes were diluted. It is unfortunate that exact data
on the water increase in distilled water have not been determined
with accuracy; a dilution of 11 per cent could hardly account for the
lowered rate of oxygen utilization as compared with the rate in tap
water. Still, there is no general law stating the relation between
enzyme concentration and the rapidity of an enzyme reaction except
that small amounts of an enzyme are relatively more effective than
are greater concentrations. The data here are in accord with this,
but their nature does not permit of further analysis in this direction.

The evidence is considerably more impressive in the data on the
effects of Ringer's solution. The rate of oxygen consumption tends
to become constant at approximately 14 cc. oxygen per liter, a result
to be expected if the cell contents are materially concentrated. Re-
duction of water content in Ringer's is approximately 20 per cent.
These suggestions regarding the effects of both distilled water and
Ringer's solution on the oxygen utilization necessarily include the
assumptions that the oxygen tension in the tissues is less than that
of the surrounding medium and increases with increase in environ-
mental oxygen concentration.

Under normal conditions, with environmental oxygen in excess of
requirements, undoubtedly one of the conditions which determine
the rapidity of oxygen utilization in the organism is the concentration
of the oxidative enzymes. It is not unreasonable to suppose, and the
results here support the view, that at least one of the conditions that
determine the degree of dependence of oxidative rate on oxygen
tension of the milieu is the relative concentration of the oxidative
enzymes.

Another interpretation of lines T and R in Fig. 1 is possible.
For this, two assumptions are necessary: (a) that in tap water with
oxygen concentration between the minimum and maximum employed,
the intra-cellular oxygen tension is greatly in excess of requirements,
and (b) that Ringer's materially reduces permeability to oxygen, thus
reducing intra-cellular oxygen tension. Under such conditions the
rate of enzyme-controlled oxidations within the cell in tap water
may be expected to remain constant with increase in environmental
oxygen supply, while in Ringer's increasing the environmental oxygen
concentration would be expected to effect corresponding increase in
intra-cellular oxygen tension, with consequent increase in rate of
oxygen utilization. The fact that in Ringer's the effects of increasing
oxygen concentration form a logarithmic curve rising well above the

324 J. WILLIAM BUCHANAN

rate in tap water must be referred to the effect of Ringer's in extracting
water and by concentration increasing the rate of oxygen transfer by
the enzymes.

However, the fact that the effects of increasing the oxygen tension
in distilled water above 5.35 cc. per liter are in general parallel to the
effects in Ringer's, in that the rate of oxygen consumption is no longer
independent of the supply, seems to indicate that some condition
common to both distilled water and Ringer's disturbs the normal
oxidative mechanism. Apparently, the ability of the oxidative
enzymes to transfer oxygen is not adversely affected, for in both the
rate of oxygen utilization increases above the controls with increase
in oxygen concentration. Since in time both distilled water and
Ringer's are lethal, it is conceivable that this common effect is an
indication of the onset of structural breakdown. The most probable
early injurious effect that would bring about this common effect is
interference with the normal cell permeability to oxygen. While
permeability effects are usually referred to the cell boundary and
there is substantial evidence that the cell boundary does regulate the
admission of oxygen (Warburg, 1911), permeability of the general
cytoplasm must also be considered.

An interesting fact appears in the data on the effect of increasing
oxygen tension in Ringer's solution. When the oxygen concentration
of the Ringer's is approximately at saturation for that concentration
of salts at 20°, the rate of oxygen consumption is approximately that
of normal animals in natural water. Oxygen being less soluble in
Ringer's than in water at the same temperature, this saturation point
therefore represents an environmental oxygen supply somewhat less
than at saturation in tap water. The significance of this fact is not
apparent; it may be merely incidental with this particular concentra-
tion of salts.

Figure 2 shows that oxygen in the concentrations employed is not
toxic for Planaria. The animals disintegrate slightly less rapidly in
abnormally high concentrations of oxygen in distilled water than in
distilled water with oxygen tension around the saturation point at
room temperature. Also, it may be pointed out, with reference to
I ig. 1, that in distilled water at oxygen concentrations corresponding
to those used in Fig. 2 the rate of oxygen consumption is higher in the
higher concentrations, while Fig. 2 shows that the rate of disintegration
i- -lightly loucr in the higher oxygen concentrations. In other words,
the higher the concentration of oxygen the less rapid the rate of dis-
integration and the higher the oxygen consumption. These facts
-ulMaittiate the conclusion expressed in a former paper (Buchanan,

OXYGEN CONSUMPTION 325

1930a), that disintegration in distilled water is not a direct result of
the action of the hypotonic medium on the oxidative metabolism of
the animal; that relative rate of disintegration in distilled water is
not an index of relative rate of oxidative metabolism.

SUMMARY

In tap water the rate of oxygen consumption of Planaria doroto-
cephala is independent of the oxygen concentration of the water
between 3.53 cc. and 13.84 cc. per liter. In so far as the methods
coincide, this confirms the results of Hyman (1929).

In Ringer's solution after four hours' exposure Planaria lose
approximately 16 per cent of their total weight, or approximately
20 per cent of their total water content. In Ringer's the rate of oxygen
consumption is sharply dependent on the oxygen concentration of the
solution between 3.79 cc. and approximately 14 cc. per liter. Above
this concentration the rate of oxygen consumption tends to become
constant.

In distilled water the animals imbibe water; irregular data indicate
that during the first four or five hours of exposure the weight increases
on the average 15 per cent, which represents an increase in water
content of approximately 11 per cent. The rate of oxygen consump-
tion in distilled water is lower than normal and constant in oxygen
concentrations between 2.65 cc. and 5.35 cc. per liter. Above this
concentration the rate of oxygen consumption increases as the oxygen
tension increases, rising above the normal at approximately 9 cc. per
liter, and continues to increase as the oxygen concentration increases,
up to 15 cc. per liter, the maximum employed.

Planaria dorotocephala contain approximately 78 per cent free
water, as shown by weight loss on desiccation over sulphuric acid.

The facts indicate that two conditions may be involved in the
regulation of the rate of oxygen utilization in relation to oxygen
tension in natural water: Concentration of water within the organism,
thus controlling the concentration of the oxidative enzymes; boundary
regulation of the admission of oxygen. In this animal, size (surface-
volume) and the degree of development of its circulatory and respira-
tory mechanisms do not appear to be important.

In abnormally high oxygen tensions in distilled water the rate of
oxygen consumption is higher and the rate of disintegration is lower
than in distilled water with oxygen tension at saturation at 20°.

22

326 J. WILLIAM BUCHANAN

BIBLIOGRAPHY

BUCHANAN, J. \V., \'>1<>. 1 impression of Oxidative Metabolism and Recovery from
]>ilute PiitasMiim ("yaniilc. Jour. Exper. Zool., 44: 285.

I '.i . IIANAV, J. \\"., I930a. The Nature of Disintegration Gradients. I. The
of a gradient in susceptibility to distilled water in Planaria.
Jour. I-'.xpcr. Zool., 57: 307.

BrciiANAN, J. \\ ., 1<>30/>. The Nature of Disintegration Gradients. II. The effect
of hypertonic solutions on the disintegration of Planaria by high tempera-
tures. Jour. l-'.xper. Zool., 57: 455.

I > AYKNi'ORT, C. B., 1896. Experimental Morphology. Part I. New York.

I >Avr M'lHiT, C. B., 1899. Experimental Morphology. Part II. New York.
lli\/i, M., 1910. Uber den Einfluss des Sauerstoffdrucks auf den Gaswechsel

rini^rr Meerestiere. Biochem. Zeitschr., 26: 255.

I 1 i.ss, O. T., 1929. The Effects of Pure Solutions of Sodium, Potassium, and Calcium

Chlorides on the Oxygen Consumption of Planaria dorotocephala. Physiol.
Zool., 3: 1.

I IYMAN, L. 1 1., 1919a. On the Action of Certain Substances on Oxygen Consumption.

II. Action of potassium cyanide on Planaria. Am. Jour. Physiol., 48: 340.

II V.MAN, L. 11., 19196. Physiological Studies on Planaria. I. Oxygen consumption

in relation to feeding and starvation. Am. Jour. Physiol., 49: 377.
1 1 \ MAN, L. H., 1925. On the Action of Certain Substances on Oxygen Consumption.

VI. The action of acids. Biol. Bull., 49: 288.
HVMAN, L. H., 1929. The Effect of Oxygen Tension on Oxygen Consumption in

Planaria and Some Echinoderms. Physiol. Zool., 2: 505.
KROGH, A., 1916. The Respiratory Exchange of Animals and Man. Monographs

on Biochemistry. London.
PFLUGER, E., 1872. Ueber die Diffusion des Sauerstoffs, den Ort und die Gesetze der

Oxydationsprocesse im thierischen Organismus. Pfluger's Arch., 6: 43.
SIIKARER, C., 1930. A Re-investigation of Metabolic Gradients. Jour. Exper. Biol.,

7: 260.
SHOUP, C. S., 1929. The Respiration of Luminous Bacteria and the Effect of

Oxygen Tension upon Oxygen Consumption. Jour. Gen. Physiol., 13: 27.
THUNBERG, T., 1905. Der Gasaustausch einiger niederer Thiere in seiner Abhiingi^-

keit vom Sauerstoffpartiardruck. M-i/»r/. Anh. I'hytiol., 17: 133.
\\AUHURG, O., 1911. I'ber Beeinflussung der Sauerstoffatmung. Zeitschr. f.

physiol. Chem., 70: 4H

OBSERVATIONS ON EUGLENA LEUCOPS, SP. NOV., A

PARASITE OE STENOSTOMUM, WITH SPECIAL

REFERENCE TO NUCLEAR DIVISION l

S. R. HALL
MILLER SCHOOL OF BIOLOGY, UNIVERSITY OF VIRGINIA

INTRODUCTION

A Stenostomum heavily infected with a colorless euglenoid parasite
was found in a collection made near the University of Virginia. The
morphology of this curious flagellate has led me to place it in the
genus Euglena Ehrbg.

My purpose is to describe this new form, its nuclear division and
its relation to its host; discuss its systematic position; and record
some further observations upon its life history.

HISTORICAL

Euglena-\ike parasites of rhabdocoels have been reported on
previous occasions. Haswell (1892) reported from an undetermined
species in Australia an intracellular, colorless flagellate that had
neither stigma nor flagellum. In 1907 the same author reported from
the same continent another colorless form, from a mesostomid. It
possessed a stigma and when liberated from the host, it no longer
progressed by metabolic or euglenoid movement but became somewhat
bottle-shaped and swam spirally by a flagellum. This parasite was
found not only inside the cells but also in the space between the gut
and body wall. He did not attempt to identify either of these forms.

In France, Beauchamp (1911) described Astasia captiva, a parasite
within the " pseudocode " of Catenula lemna. His description is
fairly complete. This form retained its flagellum in the host and
bore a colorless "rudimentary" stigma. It measured 30 to 40 microns
in length and even when liberated from the host moved only by a
rapid metabolic (euglenoid) movement. He speaks of, and his
figures show, the oblique surface striations and spoon-like depression
near the posterior end of the body. He mentions the "conduit
buccal" and nucleus as being typical of the "Eugleniens."

1 The writer is greatly indebted to Dr. B. D. Reynolds, under whose direction
this investigation was carried out at the Miller School of Biology and the Mountain
Lake Biological Station of the University of Virginia.

327

328 S. R. HALL

MATERIALS AXD METHODS

The infect i-<I Slenostomnm (a new species soon to be described)
was found in one of my aquaria in October 1929.2 Although there
were several other Stenostoma of the same, as well as of other species,
in the aquarium, none were found to be infected. Xo free-living
(Liucll.iU's were found in the water that resembled the parasite.
<Mher old aquaria in the laboratory as well as fresh collections were
then examined. Of the hundreds of rhabdocoels examined, only two
species were found infected and these were both species of Steno-
stomnm. The other Stenostomum found infected is 5. predatoriiim
Kepner and Carter, 1930). The infection is very rare in nature.

1 ortunately, one of these Stenostonia was cultivated very easily
in rather large glass dishes in which there were wheat cultures of small
flagellates and ciliates. A half dozen of these flatworms added to a
thriving protozoan culture will produce a hundred or more in two or
three weeks. The worms were examined frequently to see if any of
the free-living Protozoa had become established as parasites. Nega-
tive results are reported. I'pon the addition of two or three infected
Stenostonia, practically all the flatworms were infected in a week.
In this way it was possible to have on hand as many specimens of
Euglena leucops as were desired.

Various killing and fixing agents were employed. The best
fixation of the chromatin was obtained with Carney's aceto-alcohol
mixture. Absolute alcohol saturated with corrosive sublimate gave
good results, as did Bouin's picro-formol solution, Allen's modification
of the former, Schaudinn's and Zenker's fluids. Nuclear division was
best studied in sectioned hosts, although smear preparations were
also used. Heidenhain's iron alum haematoxylin was found to be the
best stain for the nucleus, flagellum and blepharoplasts. The alcohol-
xylol-paraffin method was employed.

I am greatly indebted to Dr. W. A. Kepner for help in the identification of the
rhabdocoels employed in this investigation.

PLATE I (X 8,000)

FlG. 1. I >IM \\ini; of ;i non lla^-llated living specimen of Englena leucops sp.
nov. as it appears imiiieili.it ely .itier liberation from the host.

Tin- two i.imi of the flagellum can be seen to end at the blepharoplasts (bl.)
at th<- lu-i- of tin- n-ser\oir i/.v.). The flagellar swelling (f.s.), a little anterior to
tin- Mi^ma (st.), is b.in-ly visible. The oblique surface striae are shown as well as
\xi\\\ t.i<e .mil side views of the puramylum grains. Four contractile vacuoles (c.v.)
.ne li-un-d, two extended ami two contracted. Details of the living nucleus other
n th( endosome can not be dearly made out. The stigma illustrated is composed
oi tour di-tiii. t parti- 1.

EUGLENA LEUCOPS

329

PLATE I

c.v..

AS.

c.v

' ''•

*1§m

— ' — ^i — — — St.

- c.v.

330 S. R. HALL

GENERAL MORPHOLOGY

The organisms vary as to size. In a contracted condition the
range in length is from 22 to 29.6 microns. Elongated, the range is
from S3 to 44.5 microns. In width, the range is from 9 to 13 microns
for resting individuals. In a resting condition the parasites are
approximately two and one-half times as long as wide and rather
1 >lunt on both ends. The anterior end bears a slight depression in its
contour where the gullet has its outlet. Most of a population are of
a\ erage size. The forms just before cell division present the maximum
H/e, but the daughter cells produced are often larger than others seen.
The wave formed during metabolic movement may be as great in
diameter as 18.5 microns. During metabolic movement the anterior
end decreases only slightly in diameter while the posterior end becomes
quite pointed.

Very fine surface striae run obliquely from left to right (Plate I,
Fig. 1).

There are no chromatophores in the cell.

A gullet and reservoir can be seen in the living animal. The two
together present a flask-shaped outline with a narrow passage, the
gullet leading anteriorly. Associated with the reservoir are a series
of contractile vacuoles that open into it. As many as four vacuoles
have been counted in living specimens (Plate I, Fig. 1).

Flagellum

\Yhen the parasite is in the host the flagellum does not extend
beyond the opening of the gullet (cytostome). Even in living indi-
viduals the flagellum can be seen to bifurcate as soon as it has passed
the narrow passage of the gullet, the two rami proceeding to the
blepharoplasts at the bottom of the reservoir. There is a "flagellar"
swelling at the point of bifurcation, which is approximately at the
level of the stigma (Plate I, Fig. 1).

When the parasite is liberated from the host and put into tap or
>pring water, a flagellum can be seen beginning to grow out of the

imp" in less than ten minutes. In twenty-live or thirty minutes
in this medium it has reached its maximum length, a little longer than
the body. I. \ni before it has attained a length one-half as great as
the body, the parasite ceases its metabolic movement and progresses
Irdily by the llagellum. In this condition the organism is dumb-bell-
shaped, moving forward irregularly in very small spirals. The cell is
>till blunt on the po>terior end at this stage. By the time the flagellum
ha- leached it-, full length, the shape of the cell has changed greatly.
It i- now bottle-shaped, moving vigorously and swiftly by means of
the new1} acquired organelle of propulsion (Text figure 1).

EUGLENA LEUCOPS

331

The animal displays a marked faculty to replace a lost flagellum,
since, as pointed out below, this structure when lost will be regenerated
if the conditions of osmotic pressure are made favorable.

TEXT FIG. 1.
sp. nov.

Outline drawing of a flagellated specimen of Euglena lencops

Paramylum Bodies

These bodies are distributed irregularly over the cell, being most
abundant in the posterior third. Only one to three small ones may
be seen anterior to the stigma (Plate I, Fig. 1). These granules
sometimes measure 4 microns at the greatest diameter. They are
found to be disc-shaped and to decrease regularly when the parasite
is subjected to inanition.

Stigma

The stigma, dark reddish-brown in color, stands out distinctly
in the anterior region, just lateral to the gullet. It is not large,
measuring less than two microns at the greatest diameter.

332 S. R. HALL

Much variation in size and form of the stigma has been observed
(Text figure The stigma is seen to be composed mainly of four to

seven di-tinct bodies which appear to be connected by fine strands
and the whole embedded in a colorless matrix. This morphologically
degenerate stigma may represent a stage in the complete loss of this
organelle.

A large number of observations were made upon the stigma of
dividing individuals. While the nucleus is in the early prophase, the
stigma enlarges slightly but usually the number of colored bodies of
which it is mainly composed does not increase. These granules push
apart somewhat and, keeping their identity, usually about one-half of
them migrate, probably attached to each other, to the opposite side of
the cell, never wandering far from the reservoir. They finally come
to lie in a position lateral to the new gullet which has formed in the
meantime.

This account of the origin of the daughter stigmata agrees some-
what with the observations of R. P. Hall and Jahn (19296) on Euglena
gracilis, except that they found that the stigma breaks up into a
large number of smaller granules which become more widely separated.

The stigma does not always divide equally. A number of times
the daughter cell has been seen to receive only two-fifths or even

&

Astf '.
•

5i

'* •» w *

TEXT FIG. 2. Variations in the stigma of Euglena leucops sp. nov. Drawings
from several specimens. X alxmt 16,000.

one-fourth of the dark granules. This unequal division aided by
n.iiural selection perhaps accounts for the decreased si/.e and degener-
al ion of the stigma.

Nucleus

The nucleus can be seen in the living organism as a relatively
large spherical body (8 to 9 microns in diameter) lying usually near
the center of the body. I'pon fixing and staining, its chief character-
istic U the presence of usually one comparatively large, deeply-staining,
<entr.il body, the endosome (karyosome of some authors). Immedi-
ately surrounding the endosome is a hyaline area, apparently lacking
in chroniatin. The t \\enty-two to twenty-five chromatin bodies which

EUGLENA LEUCOPS

have been called "chromosomes" by Keuten (1895) and which give
rise to the same number of chromosomes, remain distinct even in the
nucleus of the interphase or the resting nucleus. Their distal ends
lie close to and probably join the well-defined nuclear membrane.
Fine strands are distinctly seen connecting the proximal ends of these
bodies with the endosome (Plate II, Fig. 1).

These observations on the chromatin in the resting nucleus are
not in agreement with those of most other workers on the euglenoids.
Numerous small granules arranged at the nodes of a linin network
have been described by Hartmann and Chagas (1910) for Peranema,
by Tschenzoff (1916) for Euglena viridis, by Belaf (1916) for Astasia,
by R. P. Hall (1923) for Menoidum, by R. P. Hall and Powell (1928)
for Peranema, by Baker (1926) for Euglena agilis and by Ratcliffe
(1927) for Euglena spirogyra. It is not difficult to see how the
''numerous small granules" idea became prevalent, since this impres-
sion is frequently given by the nucleus after a casual observation,
particularly following fixation in Schaudinn's fluid, the usual method
employed. What I take to be the correct condition of the chromatin
is better followed after the use of Carney's fixative. I have been
able to observe the distinct "chromosomes" in the resting nucleus of
Euglena leucops following fixation with all the agents mentioned in
the section on materials and methods.

My observations are supported by Keuten (1895), who has figured
chromosomes in the resting nucleus and in the late telophase for
Euglena viridis. R. P. Hall (1925) has described the nucleus of the
dinoflagellate Oxyrrhis marina thus, "Around the endosome the
chromatin appears in the form of chromomeres, arranged in string-of-
beads fashion in rows, or chromosomes; such an organization seems to
be evident even in the resting nucleus, as characteristic of the dino-
flagellates 'where there appears in the nuclei ... to be a persistent
organization of beaded chromosomes with subparallel or even spiral
arrangements within the nucleus.'" 3

In preparations fixed with Schaudinn's fluid and stained with iron
haematoxylin, the endosome resists destaining very much longer than
the chromosomes. After Carney's fixative and iron haematoxylin,
the endosome decolorizes more readily than the chromosomes. The
ability to destain the endosome almost completely has made possible
a careful study of the resting nucleus and division stages.

A nuclear membrane is distinguishable in the resting condition
and can be followed through mitosis.

Frequently, the endosome is seen to be in two, three or more

3 Hall is quoting from Kofoid (1923).

334 S. R. HALL

fragments. 1\. 1'. Hall (Hall and Powell, 1928) noticed a multiplica-
tion of the endosome in Peranema trichophorum. \\ "enrich (1924), in
his clcj-rnption of Euglenamorpha Hegneri, says: "In the pellucid
variety there is a marked tendency for the nucleus to hypertrophy.
. . . This hypertrophy is accompanied by a multiplication of the
i ,ir\ "some, as many as four have been found in one nucleus. Hyper-
trophy apparently leads to amitotic division of the nucleus which is
probably followed by division of the body. Such amitotic stages have
not been found in the green variety." No explanation is offered here
for fragmentation of the endosome occurring in Euglena lencops. How-
ever, no evidence whatever for amitosis was found.

BEHAVIOR

Euglena leucops is stimulated to extremely vigorous movement by
light, heat, and a change either way in osmotic pressure.

Thigmotropism is exhibited in non-flagellated individuals. Flagel-
lated organisms were found to be negatively phototropic.

PLATE II (X 10,000)

All the figures were drawn from specimens fixed in the host with Carney's
aceto-alcohol mixture and stained with Heidenhain's iron alum haematoxylin.

FIG. 1. A typical nucleus of the interphase. Twenty-two chronuitin bodies
or "chromosomes " are figured. Note the attachment of these to the endosome.

FIG. 2. A very early prophase. The "chromosomes" and endosome are
beginning to enlarge.

FIG. 3. The nucleus is becoming noticeably larger and the "chromosomes"
and endosome are beginning to elongate.

FIG. 4. Note the blunt bifurcation of one end of the endosome.

FIG. 5. The nucleus, chromosomes and endosome are all distinctly elongated.
Most of the chromosomes have apparently lost their attachment to the endosome.

FIG. 6. This is the first stage in the metaphase. The chromosomes are quite
narrow and extend the length of the elongate nucleus. The blunt bifurcation of
one end of the endosome is distinctly pronounced by now and both ends have become
enlarged.

I i',. 7. Metaphase and the beginning of the anaphase. The chromosomes,
ha\in- spun out practically the length of the elongated nucleus during the early
metaphase, now divide i ransA ersely as the e\i n-mii ies of the nuclear membrane and
endoMime further pull apart. Note 1 he constriction in the membrane and endosome.
Fun- strands an- seen in the mitotic figure at this stage, most of which appear to be
a i i.iinei tii in between the two broken ends of the chromosomes.

I IG. S. The ( hromoMMiies arc beginning to round off somewhat and are
apparently ,cti r.u ted toward the two emU of the membrane. The nucleus and
eml'1-niin- are now decidedly elongated and constricted in the center.

1 i' ''. Tin- chronioMMiies are beginning to assume their interphase shape and
i heir at i.n hmeiit tn tlie ends of the dividing endosome again becomes evident.
N'Mii e the upward turn of the nuclear membrane and endosome.

I i' in. I In- md M|' tin- telophase. Most of the chromosomes have regained
ih' i iinii nt tn iln- eiido-ome. The daughter nucleus and endosome now

round nil as ill'- inierpha.-M- approaches.

I .mi indel.ied tu MUs M. L. Hill for help in the preparation of the drawings.

EUGLENA LEUCOPS

335

PLATE II

!

'.

7

336 S. R. HALL

The fact that Euglena leucops elaborates a flagellum soon after
being liberated from the host and is negatively phototropic, — the host
react> -imilarlv makes it particularly well adapted for swimming
.i\v,iy and finding a new host.

DIVISION OF THE BODY

When an infected host is macerated in spring water, at any hour
of the day, numerous dividing parasites are encountered after a few
minutes. Frequently as many as 10 per cent of the flagellates, after
ten to fifteen minutes exposure to this solution, will start dividing and
complete the process within seventeen to twenty minutes. Now if an
infected host is macerated in a solution approximately isotonic with
the host's plasma, no more dividing forms are seen in this medium
than normally appear in the host and twenty-five to thirty minutes
are required for the parasite to complete the division. Since the spring
water is of a lower molecular and ionic concentration than the host's
plasma, it would appear that the change in osmotic pressure both
initiates and accelerates cell division.

Division in Euglena leucops seems to be periodic, as is the case in
Hydramoeba hydroxena. Reynolds and Threlkeld (1929) observed
that division in this amoeba occurred most often about 3 A.M. .
( )bservations upon more than three hundred dividing forms of Euglena
I en co ps have shown that division, while sometimes occurring at other
times, most often takes place between 10 P.M. and 2 A.M., with the
peak about midnight.

A short time before the longitudinal split occurs, there is a dupli-
cation in the number of gullets, reservoirs, flagellar "stumps," stig-
mata, and nuclei. The split begins at the anterior extremity and
continues posteriorly along the median plane of the body until the
two daughters are completely separated. There is a continual
twisting and writhing of the binucleate body during the process of
division. Finally the posterior ends of the daughter cells are connected
only by a small strand, but this is apparently the toughest part of the
irll, for a veritable tug of war ensues, before the daughter cells are
separated.

The parasite has been under observation in the host for more than
a year and no evidence of encyst. ition has been observed.

NUCLEAR DIVISION

Prophase

The beginning of unclear division may be recognized in lour ways:
1 ini rease in size <>| ilu- nucleus, i2) migration of the nucleus anteri-

EUGLENA LEUCOPS 337

orly, (3) changes in the endosome and (4) changes in the chromatin
bodies or "chromosomes."

A considerable increase in the size of the nucleus is observed at
the onset of division.

By the end of the prophase the dividing nucleus has reached a
position in the anterior part of the organism, just posterior to the
reservoir.

The endosome increases in size and elongates. One end can be
distinguished from the other by a blunt bifurcation (Plate II, Figs. 4,
5, 6, 7, 8 and 9). Baker (1926), working with Euglena agilis, describes
the bud arising from one of the lateral points, as forming the blepharo-
plast of one ramus of the new flagellum and the connection of the
bud with the endosome as the rhizoplast. I find insufficient evidence
from Euglena leucops to support this.

By the end of the prophase the endosome is distinctly dumb-bell-
shaped (Plate II, Fig. 5). It would appear that at the end of the
prophase there is a brief pause in mitosis since this stage was en-
countered most often.

The distinct, rather large chromatin bodies or "chromosomes"
of the resting nucleus, keeping their identity, increase slightly in size
and stain more deeply early in the prophase (Plate II, Fig. 2). As
the nucleus and endosome elongate, these bodies, with one end appar-
ently remaining attached to the nuclear membrane, are spun out into
extremely long chromosomes which become arranged around the
endosome (Plate II, Fig. 5). The connection between the chromatin
bodies or "chromosomes" and nuclear membrane persists certainly
to middle prophase and perhaps until division is completed.

M eta phase

The endosome continues to elongate, its ends increasing in size at
the expense of the middle portion (Plate II, Fig. 6). One end still
can be distinguished from the other. The nuclear membrane follows
largely the contour of the endosome. The chromosomes, having
drawn out almost the length of the elongated nucleus, now break in
the middle.

There is apparently little doubt that the chromatin bodies of the
resting nucleus give rise directly to the same number of elongated
chromosomes that divide transversely. Twenty-two to twenty-five
bodies in the resting nucleus can be followed through the metaphase
when the long chromosomes produce a double number of daughter
chromosomes of approximately half the length of those in the late
prophase (Plate II, Fig. 7).

338 S. R. HALL

I have been unable to find the Y-shaped chromosomes of the
late pn "phase that have been described for euglenoids by R. P. Hall
(\<-)2.\ 1928), Baker (1926), and Ratcliffe (1927) which they take as
evidence fur a longitudinal split.

Afier a review of a large part of the literature on mitosis in the
Proto/oa. I have been able to find only one other case where a trans-
verse division in the chromosomes has been definitely demonstrated,
although others have mentioned it as a possibility [Belar (1916) in
Axtdsin and Borgert (1909, 1910) in Aul-acantha and Ceratium}.
Calkins (1929) describes a transverse division in the chromosomes of
the ciliate Uroleptus Halseyi on the third meiotic spindle.

A naphase

During this stage the ends of the endosome become quite far
apart but remain connected by their narrow, middle portion. The
bend is quite pronounced, the two ends of the endosome making a
sharp upward turn.

The forty-six to fifty daughter chromosomes now shorten but
remain radially arranged around the extreme ends of the daughter
endosomes. The membrane has become distinctly constricted in the
middle largely following the contour of the endosome (Plate II, Fig. 8).

Telophase

This stage begins when the ends of the daughter endosomes and
daughter nuclei have pulled apart. The original endosome breaks
in the center of its now narrow, lightly-stained, central portion.

The daughter endosomes and nuclei now round oft and the chromo-
somes arrange themselves at the periphery of the nucleus. The con-
nection between chromosomes and endosome is established early in
the telophase (Plate II, Figs. 9 and 10).

The accompanying diagram (Text figure 3) illustrates the history
of four chromosomes, in the resting nucleus and through mitosis.
Kater (1926, 1927, 1928) brings forth evidence that even in the higher
plant- and animals, reconstruction of the daughter nuclei is by vesicle
formation of contiguous chromosomes, that may be traced well into
the succeeding propha^e. He considers this as "good evidence for

1 tic chromosomal continuity." The evidence for this continuity
i- -till stronger in /-'.Helena lencops.

( .ilkin- IV < ha> the following to say concerning the chromosomes
ol the ciliate I 'rali-fttHs I/nLseyi: "In Uroleptus Halseyi the chromo-
somes are di\i<Ied transversely at each vegetative division. On the
theory ot the -cue these facts are possible only on the assumption of

EUGLENA LEUCOPS

339

a single type of gene for each chromosome and on this assumption
there would be not more than twelve types of genes in Uroleptus
Halseyi. Here there are twelve chromosomes in the third meiotic
spindle and twenty-four in the amphinucleus. There are forty-eight
in the first meiotic spindle and twenty-four in the second. We
conclude that each of the twenty-four chromosomes found in the
amphinucleus divides once to form the forty-eight of the first meiotic
spindle; that two of these are separated from two by this division,
and that one is separated from one at the second meiotic division,
thus leaving twelve in the third meiotic spindle in which each chromo-
some represents a single type of gene. It is immaterial, therefore,
whether division of each chromosome is transverse or longitudinal,

TEXT FIG. 3. Diagrammatic representation of the resting nucleus of Euglena
leucops sp. nov. and the history of four of its chromosomes during division.

Fig. 1 — the resting nucleus.

Fig. 2 — the middle prophase.

Fig. 3 — the beginning of the metaphase.

Fig. 4 — the end of the metaphase and the beginning of the anaphase.

Fig. 5 — the late telophase wherein begins the reconstruction of the nucleus of
the interphase.

See the text and Plate II for a more detailed description.

340 S. R. HALL

for it would be equational in either case." Carrying this theory a
step further, since the chromosomes are distinct even in the resting
nucleus in the case of Euglena leucops, a single type of gene can be
considered as recognizable at any stage in the life of the organism.

i'i KIMIMS ox RETARDATION IN GROWTH OF TIII-: FLAGKLLIM

A.S already stated, the flagellum of organisms liberated into tap or
-|)Hng water will make its appearance outside the gullet in five to ten
minutes. In 12 per cent Locke's solution, the organisms have been
carried fourteen days without ever elaborating a flagellum. In a
pure glucose solution with a molecular concentration approximately
equal to the Locke's solution they have been carried more than a
day or until death without showing a flagellum. The same thing is
true for creatine, a mixture of different amino acids and other soluble
protein derivatives. By varying the concentration of the molecular
or ionic substances it is possible to vary the time required for the
flagellum to be elaborated and for the appearance of the changes in
body form that accompany it. The hydrogen ion concentration
apparently has nothing to do with the retardation in the development
of the flagellum. Viscous substances such as thin starch paste, as
well as aqueous agar mixtures, which of course do not alter the molecu-
lar concentration of the media appreciably, have little or no effect on
retarding the growth of the flagellum.

By increasing the osmotic pressure of the medium in which Euglena
leucops elaborates and retains the flagellum, the organisms have been
made to settle down and again mo\e only by metabolic movement,
the flagellum apparently being broken off. But if the osmotic pressure
be once more decreased, the flagellum will reappear. This process
has been repeated four times with the same individuals and the
organelle grew out as readily the fourth time as it did the first.

It would appear therefore, that molecular or ionic concentration
is the main factor in retardation of flagellar growth and that this
organelle is capable of continued or repeated growth.

The osmotic pressure of the plasma of the host is apparently
great enough to prevent the elaboration of a flagellum.

l\l— I I.TS OF Aril-. Ml' IS TO CULTIVATE THE PARASITE

All attempts to cultivate the parasite in vitro failed. They were
carried fourteen days in Locke's solution diluted to 12 per cent at
pi I 7.6. At the end of this period practically all the paramylum was
consumed.

They may be kept alive three or four days in spring water.

EUGLENA LEUCOPS 341

METHODS OF INFECTION

The parasite may enter the host in three ways:

(1) By the ingestion of liberated parasites. The Stenostoma that
the writer has been able to infect, feed extensively upon Protozoa.
The parasites, either swimming actively or moving on the bottom of
the container by euglenoid movement, are ingested by the host and
enter the space between the gut and body wall, presumably, directly
through the wall of the enteron.

(2) The Stenostoma susceptible to infection are predatory and
cannabalistic. On a number of occasions I have observed the trans-
mission of the infection by an infected flatworm serving as food.

(3) Vegetative zooids are infected from the parent.

THE EFFECT OF THE INFECTION UPON THE HOST

Upon one occasion the writer observed an uninfected Stenostomum
ingest a portion of an infected one of the same species. About eight
parasites entered the enteron of the uninfected rhabdocoel. In less
than an hour approximately all of the parasites were seen in the
mesenchyme of the new host. The next day they had about doubled
in number. The host divided on the second day and the (approxi-
mately thirty) parasites were distributed about equally between the
two. At the end of a week the infection had become well established.

Stenostoma lightly infected show no ill effects whatever and appear
normal in all their reactions. No signs of "nervousness" are exhibited
even though the parasites can be seen passing over, under and in
contact with the cephalic ganglia. Kepner and Carroll (1923) found
this to be true in the case of Stenostomum leucops infected with the
ciliate, Holophyra Virginia.

Since the parasites evidently absorb their food from the plasma
in the mesenchyme, one would expect that rather heavily infected
animals would grow and reproduce slowly; but specimens infected
with as many as fifty or sixty parasites have been seen to divide
every second day under favorable conditions. This is as fast as
normal individuals reproduce.

Frequently the flatworms become infected with two or three
hundred parasites which causes them to appear bloated, due to the
ectoderm being lifted to make room for the increased contents of the
mesenchyme. Animals in this condition become sluggish, the ecto-
derm breaks in one or more places, the parasites are liberated and
death generally follows in a day or so.

Means for freeing the flatworms of the parasites have not been

found.
23

342 S. R. HALL

INFECTION EXPERIMENTS

Even though species of Stenostoma other than the two mentioned
above ingest the parasite, only one, 5. grande, has been infected
experimentally. All attempts to infect Catenida lemna, the host of
Asltisi<i cuptiva, failed, even though it ingested Euglena leucops in
large numbers.

SYSTEMATIC POSITION OF PARASITE

The apparent specificity of this parasite excludes the probability
of its being found in hosts other than rhabdocoels. The only euglenoid
ile-cribed as a parasite of this group is Astasia captiva, from Catenida
lemna, and this rhabdocoel apparently is not capable of becoming
infected with Euglena leucops. A. captiva differs further from E.
leucops in the following respects:

(1) A. captiva, according to Beauchamp (1911) is without a stigma,

save possibly a colorless rudiment.

(2) It has a flagellum while in the host.

(3) It does not change in shape or otherwise when freed into spring

water.

(4) A. captiva dies in a few hours after liberation.

(5) The paramylum grains of Beauchamp's parasite are elliptical in

outline.

The Euglena -like form that Haswell (1907) mentioned but did not
describe, resembles very closely Euglena leucops. Nevertheless it was
intracellular and a parasite of a mesostomid, the most highly specialized
of the Rhabdocoelida, and hence the farthermost removed from
Stenostomum.

\Venrich (1924) has shown that the colorless variety, pellucid a
of Euglenamorpha hegneri may arise from the green variety, simply
by loss of color. More conclusively Zumstein (1900) and Ternitz
(1912) have shown that Euglena gracilis will lose its chlorophyll and
become colorless when supplied with rich nourishment. They point
"in ill, it their results do away with the boundary line between the
genera /•'.iti'JcHii and Astasia which have been separated on the basis
"I color 'chlorophyll and stigma). Color, therefore, is not a sound
character on \\hich to base a classification of Euglend-\\ke forms.

K. I'. Hall and Jahn (1929a) say "A bifurcation of the flagellum
i- characteristic of the different species of uniflagellate Euglenidae

i mined and the llagellum in such species always shows a 'flagellar
-uelliiiL;' at the level of the stigma (which is always present) in
vegetative ges. . . . Such structural features, however, have not

EUGLENA LEUCOPS 343

been observed in the vegetative stages of any of the non-chlorophyll-
bearing species examined by us or by other workers."

Further, they believe that certain stigma-bearing but chlorophyll-
free flagellates, described in the literature as Astasia, should be put in
the genus Euglena. They made no observations upon the conditions
of the flagellum in these forms.

Belaf (1916) was of the opinion that Astasia captiva was an Euglena.
He possibly was influenced by the presence of a stigma in this form,
even though Beauchamp states that it was a colorless rudiment.

Because of the presence of a definite stigma and the bifurcation
of the flagellum, the writer is of the opinion that this new form should
be placed in the family Euglenidae Stein. If it is placed in this
family, obviously it belongs in the genus Euglena Ehrbg. because of
its morphological features, which would preclude its belonging to any
other genus of this family.

SUMMARY

A new species of chlorophyll-free Euglena, E. leucops, is described,
and though it lacks color, the presence of a stigma and the bifurcation
of the flagellum place it in the family Euglenidae.

A flagellum is not present while the parasite is in the host but is
elaborated quickly outside the host if favorable conditions of osmotic
pressure prevail. This organelle is capable of repeated growth.

The stigma has been followed through binary fission and found
to divide, but not always equally.

Nuclear division is described. The "chromosomes," which remain
distinct, even in the resting nucleus, divide transversely during the
metaphase.

Osmotic pressure apparently may initiate and accelerate cell
division.

Attempts to cultivate the parasite in vitro failed.

The host may be easily cultivated and the infection readily con-
tinued.

No evidence of encystation has been observed.

Observations upon the flagellates for more than a year have given
no evidence of a method of division other than binary fission involving

mitosis.

BIBLIOGRAPHY

BAKER, W. B., 1926. Studies in the Life Cycle of Euglena. I. Euglena agilis,

Carter. Biol. Bull, 51: 321.
BEAUCHAMP, P. DE, 1911. Astasia captiva, N. Sp. Euglenien parasite de Catenula

lemnae Ant. Dug. Arch, de Zool., Exper. et Gen., Ser. 5, T. 6.
BELAR, K., 1916. Protozoenstudien I. Amoeba diplogena, Astasia levis, Rhyno-

chomonas nasuta. Arch. f. Protist., 36: 14.

344 S. R. HALL

BORGERT, A., r">0'). t'ntersuchungen iiber die Fortpflanzung der tripyleen Radio-

l.irieii, spe/.iell von Aulacantha scolymantha II. II. Teil. Arch. f.

.>tist., 14: 134.
1 iRGERT, A., 1910. Kern-und Zcllteilung bci marincn Ceratium-Arten. Arch. f.

Protist., 20: 1.

CALKIN-, G. X , 1(>29. The Chromosomes of Uroleptus Halseyi. Science, 70: 562.
HAIL, U. I'., T'J.v Morphology and Binary Fission of Menoidium incurvum

(Fres.) Klebs. Univ. Calif. Publ. ZooL, 20: 447.
HAIL, R. P., 1925. Binary Fission in Oxyrrhis marina Dujardin. Univ. Calif.

I'n!>l. /.ool., 26: 281.
HALL, R. P., AND T. L. JAHN, 1929,;. On the Comparative Cytology of Certain

Euglenoid Flagellates and the Systematic Position of the Families Eu-

glenidae Stein and Astasiidae Biitschli. Trans. Am. Mir. Sor., 48: 388.
HALL, R. P., AND T. L. JAHN, 1929i. Dispersed Stages of the Stigma in Euglena.

Science, 69: 522.
HALL, R. P., AM> \Y. X. POWELL, 1928. Morphology and Binary Fission of Pera-

nema Trichophorum (Ehrbg.) Stein. Biol. Bull., 54: 36.
I IARTMANN, M., AND CHA<;AS, C., 1910. Flagellaten studien. Mem. lust. Osu. Cruz,

2: 64.
HA>\\I:LL, \\". A., 1892. Note on the Occurrence of a Flagellate Infusorian as an

Intracellular Parasite. Proc. Linn. Soc., N.S.W., 2d Ser., Vol. 7.
HASWELL, \\". A., 1907. Parasitic Euglenae. Zoo/. Anz., 31: 296.
KATER, J. McA., 1926. Chromosomal Vesicles and the Structure of the Resting

Nucleus in Phaseolus. Biol. Bull., 51: 209.
K \IF.K, J. McA., 1927. A Cytological Study of Dormancy in the Seed of Phaseolus

vulgaris. Ann. Rot., 41 : 629.

KATER, J. McA., 1928. Nuclear Structure and Chromosomal Individuality. So-
matic and Germ Nuclei of the Rat. Zeitschr. f. Zell. u. mik. Aunt., 6: 578.
Ki I'M.K, \Y. A., AND JENETTE S. CARTER, 1930. Ten New Species of Stenostomum.

Zoo/. Anz., in manuscript.
KEPNER, \\'. A., AND R. P. CARROLL, 1923. A Ciliate Endoparasitic in Stenostomum

leucops. Jour. Parasit., 10: 99.
KM UN, J., 1895. Die Kerntheilung von Euglena viridis Ehrbg. Zeitschr. f.

wiss. ZooL, 60: 215.

KOFOID, C. A., 1923. The Life Cycle of the Protozoa. Science, 57: 397.
KATCUFFE, H. L., 1927. Mitosis and Cell Division in Euglena spirogyra Ehrenberg.

Biol. Bull,, 53: 109.
REYNOLDS, B. D., AND W. L. THRELKELD, 1929. Nuclear Division in Hydra moeba

hydroxena. Arch.f. Protist., 68: 305.
TERNETZ, C., 1912. Beitrage zur Morphology und Physiology von Euglena gracilis

Klebs. Jahrb.f. wiss. Bot., 51.
TsCHENZOFF, I'.., 1916. Die Kernteilung bci Euglena viridis Ehrbg. Arch. f.

Protist., 36: 137.
\Vi-.\Kicii, D. H., 1924. Studies on Fuglenamorpha hegnrri n. g., X. Sp., a Eu-

;Jenoid l'"lagellatf found in Tadpoles. Biol. Bull., 47: 149.
ZUMSTEIN, HANS, 1900. /ur Morphology und Physiology der Euglena gracilis

Klebs. Jahrb.f. wiss. Bot., 34.

ALTERNATION OF GENERATIONS IN THE ROTIFER
LECANE INERMIS BRYCE

I. LIFE HISTORIES OF THE SEXUAL AND NON-SEXUAL GENERATIONS

HELEN MAR MILLER1

(From the Zoological Laboratory of the Johns Hopkins University)

CONTENTS

The Organism 345

Outline of the Change of Generations 346

Culture Methods 348

Life Histories of the Diverse Types of Individuals 349

The Amictic and the Unfertilized Mictic Females: Com-
parison 349

Length of life and distribution of mortality, p. 353.
Fecundity, p. 358. Duration of fecund period and rate
of egg-production, p. 361. Age at cessation of fecund
period and duration of the post-fecund period, p. 362.
Relation of length of life to fecundity, rate of egg-
production, duration of fecund period, p. 364. General
and comparative, p. 367.

The Fertilized Mictic Females 371

The Males 372

Summary 378

Literature List 380

THE ORGANISM

The small loricate rotifer, Lecane (Distyla) inermis Bryce, presents
many advantages for the study of the alternation of parthenogenetic
and bisexual generations, common among the rotifers, and of the life
histories of the diverse types of individuals that occur. It is hardy,
easily cultivated, multiplies rapidly and changes readily from the
non-sexual to the sexual condition and vice versa. The entire life
of the female lasts usually but nine or ten days, and the first offspring
appear on the second or third day. This organism is of interest as a
representative of a genus in which the occurrence of males has not

1 This paper was prepared for publication during the tenure of a National Re-
search Council Fellowship. The author desires to acknowledge her indebtedness to
Professor H. S- Jennings for advice and assistance freely given.

345

346 HELEN M. MILLER

hitherto been < It-scribed (\Vesenberg-Lund, 1930, p. 94). The male
of this specie- \vas first observed by Finesinger subsequent to his study
(1926) of the inheritance of certain environmental effects in a succes-
sion of gem-rations of parthenogenetic females.

Lecane inermis is one of the slow-moving rotifers, common in wet
-plia^num. It belongs to the family Euchlanidae, which is character-
ized by the possession of a lorica composed of two parts, a dorsal and
a \i-ntral plate, connected by lateral sulci. In Lecane inermis the
li'iii a is extremely flexible, so that dorsal and ventral plates are barely
distinguishable; the animal therefore resembles somewhat the more
primitive, soft-bodied notommatids, with which the euchlanids are
classified by \Vesenberg-Lund (1929). A full description of this and
other species of Lecane is given by Harring and Myers (1926). The
individuals (females) of the stock employed in the present investigation
are somewhat larger than the dimensions given for this species by
Harring and Myers. Adults are 175-190 n in length by 65-75 /j. in
breadth; newly-hatched individuals are about 130 p. in length by 33 /j.
in breadth.

As is well known, the rotifers of most common occurrence are the
parthenogenetic or amictic females, and specific descriptions are based,
in large measure, on these. In a single species there may occur,
however, three sorts of individuals: the common amictic or non-sexual
females, the mictic or sexual females, and the males. An outline of
the change of generations will facilitate the understanding of the
present study.

OUTLINE OF THE CHANGE OF GENERATIONS

(Compare Fig. 1)

The common non-sexual or amictic females (A) multiply exclusively
by parthenogenesis. They deposit eggs (/) that are ellipsoid in form,
with a thin, smooth, outer membrane. These eggs carry the diploid
number of chromosomes, are incapable of fertilization and produce
females, usually like the mother, so that multiplication by (diploid)
parthenogenesis may continue for many generations. But among the
females produced by such eggs are at times individuals (J/) that
bring forth small c.uu*- ( /// ) which develop into the small, male indi-
vidual-. These females are called mictic because their eggs are
capable of bring fertilized. They resemble outwardly the amictic
It-male-. Tin- small, male-producing eggs carry the haploid number
ol chromosomes. There is some evidence that the male embryo may
rithrr doublr its chromosomes during development (Asplanchna
intermedia, Tauson, \()2 1 i or develop entirely with the haploid number
(Asplanchna <i»i/>/!or<i, Whitney, 1929). When the mictic females are

I. IKK IllSTORIKS OF Till-: ROTIKKk LFA'ANK

347

fertilized they produce, instead of the small, haploid, male-producing
eggs, larger, dark eggs with a thick, horny shell. Such an egg has,
of course, the diploid number of chromosomes. It remains unhatched
for some time, and then produces a parthenogenetic or amictic female,

CD

FIG. 1. Diagram showing the alternation of sexual and non-sexual generations
in a bisexual rotifer, such as Lecane inermis. The most common type of individual,
the non-sexual "amictic" female (A), reproduces exclusively by diploid partheno-
genesis. Its eggs (/) develop invariably into females, which are usually amictic, so
that multiplication by diploid parthenogenesis may continue for many generations
(as indicated by the long, broken-lined arrows). But the daughters of the amictic
female may be "mictic" (M). Their eggs are haploid, and develop partheno-
genetically (m) into males, or, if fertilized, produce non-sexual amictic females, the
"stem mothers," with which the cycle begins anew.

with the diploid number of chromosomes. From such a female,
commonly spoken of as the stem mother, begins again the cycle of
diverse generations just described.

The purpose of the present investigation is to analyse this cycle
in the species Lecane inermis; to give an account of the different

348 HELEN M. MILLER

types of individuals composing it, and to discover the factors that
cau>e their production. This first paper presents a description of the
different types of individuals, with a comparative study of their life
hi.-tories. A later paper, now in preparation, will deal with the
factor> causing the change of generations.

CULTURE METHODS

Lf«inc inermis is readily cultivated in grain infusions of proper
strength, in malted milk, and in some standard inorganic culture
media to which are added suitable food organisms. Oat infusion,
used by Jennings and Lynch (1928 — I) for Proales sordida, is (as they
observe) a satisfactory medium also for Lecane inermis. An infusion
of suitable strength is prepared by boiling twenty flakes of rolled oats
for three minutes in 100 cc. of spring water, filtering while hot, and
allowing to stand 24 hours before using. This "standard" oat
infusion was employed in the comparative studies of the life histories
of the different types of individuals and in the early part of the experi-
mental work.

For precise experimental study of the influence of the environment
on the kinds of individuals that occur, a more exactly controllable
culture medium is necessary. Luntz (1926, 1929) has demonstrated
the usefulness of synthetic inorganic media, such as the Benecke and
Knop solutions, in the experimental investigation of the life cycle of
rotifers. The food is controlled by the use of pure strains of food
organisms, and the pH and initial salt concentration are likewise
under control.

Benecke's solution in 0.07 per cent concentration, with addition
of suitable food organisms, provides a most satisfactory culture
medium for Lecane inermis. To it were added pure strains of certain
bacteria, B. protens, B. sitbtilis and B. coli (obtained from the School
of Hygiene of the Johns Hopkins University). Such pure strains are
inadequate as food, the rotifers surviving in them for but three
generations. A mixture of two or three of these strains of bacteria
is mop- -.itisfactory. Addition of green alga- to the bacteria gives
still belter results. A number of different alga?, obtained from the
laboratory of I'ringsheim at Prague, were tested. Species of Chlaniy-
domonas and /-'.Helena were apparently too large to serve as food for
I.fuinr inermis. Chlorella vulgaris Beyerinck proved to be suitable,
pariicularly \\hcii combined with B. protens. The mixture of Chlorclln
ami B. proti-Hs lias been employed in much of the experimental work.
Chlorelln is -muii on |,eef agar, B. protcus on bacteriological agar.

LIFE HISTORIES OF THE ROTIFER LECANE 349

A fresh suspension of these two food organisms in the culture solution
is prepared daily immediately before the transfer of the rotifers.

In order to avoid contamination with other organisms, bacterio-
logical precautions are observed in the preparation of the culture
medium and food suspension, and in the handling of the rotifers.
The method is, with a few necessary alterations, that devised by
Raffel (1930) for the culture of Paramecium. The rotifers are culti-
vated individually on hollow-ground slides, placed in sterile Petri
dishes containing distilled water. They are transferred daily with
sterile, cotton-stoppered pipettes. A fuller account of the method of
culture will be given in a later paper on the factors causing the change
of generations.

LIFE HISTORIES OF THE DIVERSE TYPES OF INDIVIDUALS

Knowledge of the general biology of the bisexual generations of
rotifers is not extensive. This is doubtless due in part to the rare
occurrence of sexual individuals in many of the species that have
been employed for laboratory investigations. There have been few
detailed comparative studies of the lives and activities of the mictic
and amictic females, though such studies are much needed for an
understanding of the fundamental difference between them, and of the
different roles they play in the life history of the species. Knowledge
of the life history of the males is in little better case. Lecane inermis
is a favorable organism for supplying these needs, since it readily
yields an abundance of all types of individuals.

The Amictic and the Unfertilized Mictic Females: Comparison
The amictic and mictic females of Lecane inermis resemble each
other closely, but differ markedly in certain features of their life
histories, particularly in fecundity and length of life.

The mature females, whether amictic or mictic, present the
appearance shown in Fig. 2.

As in Proales sordida (Jennings and Lynch, 1928 — I), four periods
may be distinguished in the life of the individual: (1) The period
of embryonic development, from the deposition of the egg to hatching;
this requires, in Lecane inermis, about one and one-half days. (2) The
immature period, one of rapid growth and activity, concluding as a
rule on the third day with the deposition of the first egg. (3) The
period of fecundity, five or six days during which the eggs one at a
time are matured and deposited in regular succession. (4) The post-
fecund period, or period of old age, during which the activities of the
female gradually cease, structural degeneration sets in and death

350 HELEN M. MILLER

ensues, usually on the ninth day, though it may be deferred for
varying periods, depending on the conditions.

The amictic female. — The term amictic was first applied by Storch
](>2 I to the commonest type of rotifer female, since it cannot repro-
duce by amphimixis. The cytological investigations by Whitney,
Monli, Nachtwey and others, of several species of rotifers, have
shown that the eggs of the amictic female invariably develop by
diploid parthenogenesis and produce diploid females, which may be
either amictic or mictic.

The amictic females of Lecane inermis produce eggs that are
ellipsoid in form and measure about 64 n by 38 /z (Fig. 6). They are
covered with a thin membrane surrounded by a gelatinous layer
which swells, separates from the egg and disappears before the embryo
emerges, about 30 hours after deposition.

The mictic female is capable of reproduction by amphimixis. Its
eggs, as shown by the above-mentioned investigators, undergo reduc-
tion in the number of chromosomes. The mictic female of Lecane
inermis, although indistinguishable in appearance from the amictic
female, is immediately recognizable by the kinds of eggs it produces.

EXPLANATION OE PLATE

These drawings were reconstructed from many camera lucid. i sk-.'tchcs, for which
a Zeiss microscope was used with oc.lSx, obj.8.3, and with oc.lSx, (Leitz) obj.3. The
magnification of the seven figures is 518 diameters; this magnification was obtained
by doubling the average dimensions of sketches of eggs and of living individuals,
made with oc.lSx, (Leitz) obj.3.

Abbreviations

c, contractile vacuole />.#., prostate (?) glands
</, rudiment of anterior portion (if the o, ovary

digestive system oe, oesophagus

g, gastric glands r, retro-cerebral organ

fx, excretory (?) granules /, toes

;, intestine If, testis

»i, mastax v.d., vas deferens

1 K,. 2. Mature female (mictic or amictic); dorsal view.

I ii.. S. Young female, shortly after hatching; lateral view, showing prominent

, gastrii iJ.uuU, ee-Miph.igus, intestine and contractile vacuole.
IK.. 4. Male; dorsal view.

In.. 5. M.de; lateral view, for comparison especially with Fig. 3. The di-
• in i> re-presented only by an anterior rudiment; the contractile vacuole is
ni .

I •!«.. <>. 'I In- (female-producing) egg of the amictic female, several hours after

deposition; lateral view. The gelatinous layer has separated from the egg membrane.

I i'.. 7. 1 In- un. ile-producing) egg of the mictic female, before the initiation of

I ll'.IV

I i'-. x I In- tiTiili/e<l (female-producing) egg of the mictic female.

LIKE IIISTORIKS 01- Till-. KOTII'T.K U-.CANK

351

MW\
r \

1

v.d.

8

352 HELEN M. MILLER

If unfertilized, the eggs (Fig. 7) resemble the female parthenogenetic
eggs, except that they are smaller (44 n by 32 ju) and produce males.

The fertilized eggs (Fig. 8) are larger (64 /j. by 38 /*), covered with
a thick, horny shell, and produce amictic females. They are very
Mack when laid, but the color fades to light brown in the course of
several days. In this condition they remain for a period varying
from a few days to two months. They withstand drying, which does
not seem to alter their hatchability. Ordinarily, only 20 per cent to
30 per cent hatch in spring water or in oat infusion. Further investi-
gation of the hatchability of the fertilized eggs is much needed. The
course of development of these eggs has not been studied in detail.
Apparently, cleavage begins shortly after deposition, but the greater
part of embryonic development occurs during the 24 or 48 hours
prior to hatching. The embryo emerges through an irregular break
in the side of the shell.

The life histories of 108 amictic and 111 unfertilized mictic females
\\ere studied and compared in detail. These individuals were char-
acterized by a high degree of genetic uniformity, since they were
derived recently by parthenogenesis from a single common ancestor,
and since they were derived from young mothers, less than 4 days old.
This latter precaution eliminated the possibility of certain intrinsic
differences, in fecundity, embryonic mortality and in the duration of
certain periods in the life history, which result, in Proales sordida
(Jennings and Lynch 1928-1, II), from differences in the ages of the
parents.

The experimental females were cultivated individually under
similar conditions, in small groups, during a period of two months.
I iiiformity of cultural conditions was obtained as follows. Small
mass cultures, of 15 to 25 individuals in three drops of oat infusion,
were maintained, the individuals being transferred daily to new fluid.
I emale-producing parthenogenetic eggs were removed from these
cultures at ihe time of daily transfer, and isolated in depression slides,
each one in three drops of oat infusion. The eggs isolated at any one
time had thus been produced during the preceding 24 hours. The
ret oids of ages begin at the time of isolation, so that they are accurate
only to a period of within 24 hours.

I In eggs hatched as a rule during the first day after isolation.
The young females grew rapidly and began to deposit eggs on the
second day. \<>t until then could the individuals be recognized as
nnctic or amictic. They were transferred daily to clean slides and
twenty-four hour infusion, at which time the number of eggs produced
and other deiaiU wen- recorded for each mictic and amictic female.

LIFF HISTORIES OF TIIK ROTIFFR LECANF,

353

8 3

'£.rs

a

_3 *"
d) ^C2

"c3 '33

•£ «J

T3

•g

4) *

it: c

"" -2s

ll

bC U

O rt

"S.I3
U Ul

3 .

*O '^i

O .

U5

bo

tn **•"

b.a

11

JT N

»2i

- 3
O

O v3

5J C

S S 3

e <r> o

-T* 3 en

<a O K

<*.<!)?*>

^ C C3

vj cS "o

§g

+-> O

OJ

<4->

™" t/)

°'£
js-a

tin

c —
a>

t^

t—

t^

CO

o

1 — 1

v— (

t^-

O

^H

0

0

«— (

i*

V
f:
C in

E M

gg

Ov

^— *

C>

. — i

\O

-f

IO

o

vO

-f

o

3-s

o
H

O

"O

0

10

o

-*

o

rO

o

(S

o

-

0

T3

0

•a

o

o

0

O

•o

0

o

"O

o

&

c

o

o

o

T3

O

o

CO

-o
o

0

o

o

o

0

o

r^

-o
o

•a

*— *

-t-t
(/)

O

•a
o

v-H

0

0

-a
o

o

o

o

o

O

CN

ro

-*

P«J

rO

o

o

T— (

ro

o

o

i^

^f

•*

M3

fO

•*

•*

<T>

<r>

•>*

Tf

CN

OO

•*

\O

^^

>o

lO

10

•<*

^J^

10

Th1

•*

ro

•*

C«}

't1

LO

•*

^

IO

T^

•*

>*

Tj<

oo

rf

•*

<N

ro

CN

s-H

00

-*

t-H

fO

ro

ro

CN

•«*

^

«-•

O

O

J3

C

O

o

J=
C

f~^

O

O

O

^-H

*-H

Successive days of life

1 1

X

^

•s j

Irt I

<o J
bOw
bio '

4) •;

"3 '
1- t

•s I

c y

3

2

CNI

a

j

M

a
3

^

j
j
j

M

4

5

t>

>

H

ro

IO

^H
>» ,

•° •;

-D :

!si
i

Sbn

M *

<" C

«-*.

•o j

S3 \

^ J

" r

~ C
3 <•

^

CN
J

H

|J

: 5

ll

3 .U
+J

3

^
j

a

00

^

IO

\0

t--.

354 HELEN M. MILLER

Since the male embryos require 30 hours or more for development,
fertilization of the mothers by the sons was impossible, so that the
mictir females as well as the amictic females reproduced throughout
life only parthenogenetically.

During the two months of cultivation, the temperature varied
from 19° to 24° C. Daily variations did not exceed one and one-half
degrees. Although the mictic and amictic individuals were not
distributed throughout this period in equal numbers, certain striking
differences are observable in the life histories of the two kinds of
females that cannot be attributed to differences in temperature, since
they are exhibited by contemporaneous individuals subjected as nearly
as possible to the same conditions of culture. Large contemporaneous
cultures of the two types of females were at the time impracticable.

Table I gives the life histories with relation to the number of eggs
produced daily, and the length of life, for two small groups of con-
temporaneous females of the two types. From Table I it is at once
apparent that the mictic females (unfertilized) produce fewer eggs
than the amictic females, but usually live longer. Further, on the
whole, the egg-laying period of the mictic females terminates earlier.

Length of life and distribution of mortality

Table II gives for the individuals of the amictic and mictic popu-
lations the length of life in days (after their isolation as eggs); also
the biometric constants, and the proportions that lived for the modal
number of days, less than the mode and more than the mode.

As Table II shows, the largest number of individuals died, in both
populations, on the ninth day, so that this is the modal length of life
for both sets. But many of the mictic females (55.9 per cent) lived
beyond this age, while fewer of the amictic females (32.4 per cent)
did so. The maximum age reached by the amictic females was 15
d;iys. while 13 per cent of the mictic females lived beyond this age,
t\\o reaching the ages of 27 and 28 days respectively. The mean
length of life of the amictic females is 8.9 ± 0.11 days; of the mictic
females, 11.0 ± 0.28 days. There are thus very characteristic and
significant differences in the length of life reached by the amictic and
inn-tic k-mules respectively. Further, the duration of life is much
m<. re variable for the mictic than for the amictic females. For the
amictic females the standard deviation in length of life is 1.72 ± 0.08
days, and the cneiiicient of variation is 19.28 ± 0.92 per cent, while
for the mictic females the standard deviation is 4.37 ± 0.20 and the
coefii. ient ..I' variation 39.41 ± 2.04 per cent.

'1 hese dillerence-, in length of life are very noticeable in observation

LIFK HISTORIES OF THE ROTIFER LKCANK

355

w
_i

PQ

£-

~ C

5 "O

**" C
u •-

'J^ tn
u >>

•^ 03
C TD

. tn

— 41

*H 00

c o
ca ca

P -C

S.

g.

a

u

•4-> OJ

.

0) 3
J= 13

•°^

£ p

4)

c° rt

OJ 4)

J3
ca

s

a

VJ

OJ

C ^*~>

"«

4d

o
H

00
0

1— t

i— i
^— (

00

cs

*-H

t--

CN

*-H

>O

rN

o

"1

(S

o

a

i-H

W3
(N

^-H

(N
<N

T— 1

S

CN

O

CN

i— <

O*

cs

00

^— (

1 -

^— 1

c

CN

2

*-H

ro

s

O

T-H

M

•^

t— .

fS

IO

OO

-

cs

*— 1

IT)

o

>o

•r-*

>O

»

^•H
rO

O
CN

00

IO

*— 1

"*

r^.

0
CN

»— 1
^-H

O

r-

f*5

u->

^

-t

<»>

<N

-

tf)

'^

a
Q

Number Amictic Females

Number Mictic Females

^H

•0

>0

£3 V u

"^!

00

•^

o ** "^

rv)

10

rr^

, X 2

r*^

IO

CN

H rt

1

«rj

0

00

O -u73

q

0

u d 2

od

00

d

(S!

cs

-u

r-> ^

OO

m

10

o O "^

OO

^H

1--

>-— o

06

0

Cu""0

w

CN

""

01

•o

o

Ov

0

o

(S

-t1

T*

0

q

CN

-I-J >.

d

CN

CN

c S

•H

-H

•ti

ig'o'rt

00

_,

r^

a 1

f>

S

d

CN
1

.2

.3

OO

0

d

0

d

CN

d

Q

-H

-H

•n

_j

fVJ

j — .

IO

IH

f-~»

<r>

^o

ca

•d

^— (

^•^

c^i

c

a

i

53

GO

o

^— <

CN

CO

d

d

d

e

ca

•U

-H

-H

flj

o

o

o

i*S

CN

06

— ;

CN

1

tn

Jj

tn

4)

C^

r t

U

£

CU

u

CS
4>

to

C
4>

1

'-*_»

CJ

u

5

£

§

356

HELEN M. MILLER

CD u *•*- r*

en "•-• * J ^

3 - /j y «

CD -— Z .« i^-i

31

cs

—

|

0)

-
bo

^j

y

~~

A

c

g

1

(0

4^

4)

rt

(4

•o

U
U
0)

- _-

en

c

6.

4)

V

bo

«

.

c - b* c
"rt

111:

rj o eo

LIFE HISTORIES OF THE ROTIFER LECANE 357

of the living animals. The mictic females cease egg-laying early in
life, but most of the individuals continue to swim about actively for
several days after, a few of them for many days. In these long-lived
individuals physical decline is very gradual. The gastric glands and
mastax become blackened and locomotion wanes, typical signs of
structural degeneration. But the mictic females may linger on for
many days in this condition. The amictic females, on the other hand,
frequently show signs of age before the last egg is produced and all
are dead within two days thereafter.

The distribution of mortality is represented graphically in Figs. 9
and 10. In Fig. 9, curves Af and A show the proportions of the mictic
and amictic populations respectively that reached the different ages.
Figure 10 represents mictic and amictic individuals as though con-
temporaneous, and shows the proportions of each population present
at the beginning of each day. Embryonic mortality is negligible;
only one egg of the 220 eggs isolated failed to hatch. In this respect
early-born mictic and amictic females resemble early-born populations
of Proales sordida (Jennings and Lynch, 1928 — II, Fig. 15). None
die during immaturity; all live through the fourth day. All of the
amictics and 99 per cent of the mictics survive the fifth day after
isolation, that is, four days of fecundity. On every day thereafter,
fewer amictic than mictic females are still alive. In both populations
the mortality rate is low on the sixth day, higher on the next and
reaches a maximum on the ninth day. It thereafter decreases until
all the amictics are dead on the fifteenth day and the mictics linger on,
two or three dying on successive days until the twenty-eighth day.

To what extent is the difference in longevity of the mictic and
amictic females correlated with, or the result of, difference in fecundity?
Information regarding this question is obtained from examination of
the distribution of fecundity in the two populations, the rate of
egg-production, the duration of the post-fecund period and the age
at which the production of eggs ceases.

Fecundity

Inspection of Table I, giving the life history records for contempo-
raneous mictic and amictic females, has shown that the mictic female
deposits only 14 to 16 of the small, male-producing eggs; the amictic
female deposits 19 to 21 of the larger, female-producing eggs. Table
III presents a comparison of the distribution of fecundity in the two
populations, with the biometric constants; and Fig. 11 shows the
proportions of individuals that produced the different numbers of eggs.
The differences in the fecundity of the two types of female are striking.
The amictic females produce 13 to 24 eggs per individual; the modaLA^

.v-

HELEN M. MILLER

_^

"3

LIFE HISTORIES OF THE ROTIFER LECANE

359

m

bo

°n

3
T3

I

*0

8

a

en
bo
bJO

3
C

_o
rt

o

's

<u

N

c

3

O $

'5 |

03 •*-

<<-! CJ

O '^

•a

oo

•ti-S

0)

"3
o
H

oo

O

rH

~

"o ^

sO
-f

d

10

d

d

d

CN

sO

d

CN

so

'c —
.-!S

•n

•H

•n

41

41

CN

sO

fl

0^

es

o

0

ON
00

S

J^

0

ON

^ — i

ON

1

O

CN

t~-

O
cs

00

oo

sO

i-H

2

3

-afl

O

d

o

d

O

d

d

i— I

d

00

»o

rt'w

-H

41

41

41

41

s

CO

la

sO
°

ON

0

CO

CN

d

d

10

|

i — i

"""

-t

0

ON

cs

2

-

OO

O

5

sO

T-H

2

-

LO

O
4J

0

41

0

41

o
41

o
41

a

0

1 — 1

OJ

~

CM

J2

CN
CN

3

sO

o

sd

2

O

0)

"-1

•^

a

o

.

00

o

o"

-

o

s

so

o

in

OC

«

o

-f

o

cp

CJ
3

,0

o

-o

O

ts

<rH

Q,

a

-

0

3

•n

Number of eggs

Number amictic females

en

S
i*

OJ

z

(a) Amictic Females (108)

T-H

1

o

4-1

O

(c) Mictic Females, without indiv

Differences a-b

SJ

HELEN M. MILLER

fecundity is 21 and the mean 20.7 ± 0.13. The mictic females
produce 11 to 16 eggs (with the exception of one individual which
produced only two eggs and died on the fifth day); the modal fecundity
i< 15, and the mean 14.2 ± 0.11. Thus, the mictic females produce

tilurly only two-thirds as many parthenogenetic eggs as the amictic
females. The fact that the coefficients of variability for both popu-
lations are low, 9.9 per cent for the amictics and 12.2 per cent for the
mictics, emphasizes the difference in the fecundity of the two types
of females.

Excluding the single exceptional mictic female which produced
only two eggs, the mean fecundity for the mictic population (110
individuals) is 14.3 ± 0.08, the standard deviation 1.29 ± 0.06, which
is less than that of the amictic population by 0.77 ± 0.11, a significant
difference, and the coefficient of variability is reduced to 9.03 ± 0.41,
which is almost the same as that of the amictic females.

26
26

24
22
20
18
16
14
12
10

e

6

i

2

0

2 3 4 3 6 7 B 9 10 11 12 1J 14 15 16 17 18 19 20 21 22 25 24 25

I i'.. 11. Lecane inermis. Graphs showing the numbers of eggs produced by
individuals of the two populations, the amictic females (A) and the unfertilized
mictic females ( M). Horizontal scale, numbers of eggs produced; vertical scale,
- of the populations producing the different numbers of eggs.

The two populations differ further in the fact that a very much
greater proportion of the mictic females produced the maximum or
nearly the maximum number of eggs. This is clearly observable in
l-'iij. 11. \\hcrcas 4S per cent of the mictic females produced 15 or
L6 ej only 21 per cent of the amictics produced 23 or 24 eggs.
The mode and the mean fecundity more nearly coincide with the
maximum in the mictic population.

LIFE HISTORIES OF THE ROTIFER LECANE

361

Duration of the fecund period and rate of egg-production

It will be recalled that the mictic and amictic females mature and
deposit the first eggs at about the same time, during the second day
after isolation, also that the mictic female deposits but two-thirds
as many eggs as the amictic female. If the male-producing partheno-
genetic eggs are deposited at the same rate as the female-producing
eggs, the fecund period of the mictic female should be only two-thirds
as long as that of the amictic female.

TABLE IV

Lecane inermis. Duration of the period of fecundity of amictic and unfertilized
mictic females. The figures given in the body of the table indicate the number of
individuals in each population tha" required for the production of their eggs the
number of days indicated on the upper line.

Number of days

2

3

4

5

6

7

8

9

10

11

Total

Number of Amictic Females

4

SO

SO

19

14

Q

1

1

108

Number of Mictic Females

1

0

22

57

76

5

111

Mean

Mode

Standard
Deviation

Coefficient of
Variability

(a) Amictic Females (108) ....

6.42 ±0.09

5, 6

1.44 ±0.07

22.45 ± 1.08

(b) Mictic Females (111) .

5.10 ± 0.05

5

0.83 ± 0.04

16.22 ±0.75

(c) Mictic Females, without
individual that produced 2
eggs and died on 5th day
(110)

5.13 ±0.05

0.78 ± 0.04

15.13 ±0.70

Differences a— b

1.32 ± 0.11

0.61 ±0.08

6.23 ± 1.32

a— c

1.29 ±0.10

0.66 ±0.08

7.32 ± 1.29

In Table IV is given the number of days during which the females
of both classes continue to produce eggs. The minimum number of
days of the fecund period was the same for both (with the exception
of the one individual, which produced only two eggs and died on the
fifth day) ; but the total number of days ranged from 4 to 7 for the
mictic females, from 4 to 11 for the amictic females. The mode for
the mictics is five days; for the amictics 5 and 6 days are equally
common. For 23 per cent of the amictics the period of fecundity
exceeds the maximum period for the mictics. The mean for the
mictics is 5.10 ± 0.05 days, for the amictics 6.42 ± 0.09 days. The
standard deviation is, for the amictics 1.44 ± 0.07, for the mictics
0.83 ± 0.04. The coefficient of variability is, for the amictics 22.45

362

HELEN M. MILLER

± l.i is. for the mictics, 16.22 ± 0.75. The differences in the mean,
standard deviation and the coefficient of variability of the two popu-
1, n ions are statistically significant, regardless of whether the one
piional niictic female is included in the calculations (Table IV).
The mictic female requires on the average only one and three-
truths days less to produce its eggs, though these are fewer by one-third
than the number produced by the amictic female. This means that,
although both types of females produce 3 to 5 eggs per day, rarely 6,
the mictic female deposits its small (male-producing) eggs, on the
.1 \ erage, in less rapid succession than the amictic female deposits its
larger (female-producing) eggs. The mictic female deposits, on the
average, 3.8 eggs per day, or one every 8.6 hours; the amictic female
3.2 eggs per day, or one every 7.5 hours. These figures represent the
quotients of the mean fecundity divided by the mean duration of the
fecund period. Additional data regarding the rate of egg-deposition
are given later in the paper.

Age at cessation of fecund period and duration of post-fecund period

In Table V is given the duration of the post-fecund period for all
individuals of the two populations. Most of the amictic females
(59 per cent) died within 24 to 36 hours after deposition of the last egg;
33 per cent died within one day, all died within two days thereafter.

TABLE V

Lecane inermis. Duration of the post-fecund period of amictic and unfertilized
mictic females. Those individuals that produced eggs on the day on which they died,
or as many as four eggs on the day preceding death are considered as having a post-
fecund period of one day or less; if they produced no eggs on the last day, but one to
i liree on the preceding day, they are listed as having a post-fecund period of one to one
and one-half days; if they produced no eggs on the last two, three or four days, their
post-fecund period is correspondingly two, three or four days, etc.

Number of Days

( >nr day
(ir lrs>

I-

U

2

3

4

5

6

7

8

9

10

.\ mill » i •• .1 Amictic Females ..

36

64

8

N unilii-r til M i<i ii 1 rui.drs. .

2

is

Ml

IX

17

6

11

?

?

3

3

Nmiit.' r of ] >ays

11

12

13

14

IS

16

17

18

19

20

21

22

Total

Nimil.rr </l Ami. tic 1 emales. . .

108

.\uml.rr lit' Mi.-t i, 1 Vm.i Irs. . . .

1

1

3

1

2

0

0

2

0

0

1

1

111

Nol so with the mirtic females. Only 33 per cent of the entire popu-

ifiii died \\itliin two days after the cessation of egg-deposition.

LIFE HISTORIES OF THE ROTIFER LECANE

363

Of these, 20 individuals survived for two days, 15 for one and one-half
days, only two died within 24 hours after deposition of the last egg.
While two days after the cessation of egg-production is the commonest
period of death, three or four days are almost equally common; five
and six days less frequent. A few individuals survive 7 to 22 days
after the cessation of egg-production, there being, however, but one,
two or three individuals in each class. Yet 20 per cent of the mictic
females lived more than 6 days after the fecund period had ended.

As set forth in the preceding section, the mictic females cease to
deposit eggs earlier, on the average, than the amictic females. The
post-fecund period of the mictic females is thus extended in this way
by about 1.3 days. However, this accounts for only a small fraction
of the greater length of life of the mictic females after the cessation of
egg-deposition. The mictic females have, on the whole, a much
greater ability to survive the period of fecundity than the amictics.

This striking difference is shown for the entire populations in
Fig. 9, representing the distribution of ages at the cessation of egg-
production in comparison with the distribution of ages at death.

TABLE VI

Lecane inermis. Comparison of the ages at which amictic and unfertilized
mictic females cease the deposition of eggs.

Age in Days

4

5

6

7

8

9

10

11

12

13

14

Total

Number of Amictic Females

0

t

19

?s

?6

17

n

6

0

o

1

108

Number of Mictic Females

3

?4

4Q

16

is

4

111

Table VI gives for both populations the number of individuals
that cease egg-production at the different ages. The fecund period
of the amictic females terminates usually on the seventh and eighth
days; that of the mictic females usually on the sixth day. As stated
in a previous section, all of the amictic females and all but one mictic
female live through four days of egg-production. Mortality then
begins in both populations and rises to a maximum on the ninth day.
Thus, the maximum mortality in the amictic population occurs from
one to two days after the maximum number of individuals cease
egg-production; in the mictic population, three or four days thereafter.
After the ninth day, the distribution of mortality of the amictic females
follows the cessation of egg-production within one or two days. The
mictic females, however, have all ceased egg-production at the end of
the ninth day, and 56 per cent are still alive. These die at a slow
rate until the 14th day, the remaining 15 per cent at a fairly uniform
rate until the 28th day.

364

HELEN M. MILLER

Relation of length of life to fecundity and rate of egg-production

The foregoing data indicate that the difference in the length of
life of mictic and amictic females results, in part, from the difference
in their ability to survive the fecund period. This has been evidenced
by the relative duration, in the two populations, of the fecund and
post-fecund periods, and the relative ages at which egg-production
ceases.

\\'e may inquire further into the extent of the relationship betwreen
length of life and fecundity of mictic and amictic females, by comparing
the fecundity, rate of egg-production, and the duration of the post-
fecund period, in relatively short-lived, average-lived and long-lived
individuals of each population. Those that lived less than 9 days,
the mode, are classified arbitrarily as short-lived, those that lived 9
days as average-lived and those that lived more than 9 days as long-
lived. The. results of this comparison are presented in Tables VII

to IX.

TABLE VII

Lecane inennis. Comparison of the relation of life duration to fecundity
displayed by amictic and unfertilized mictic females. Individuals that lived less
than 9 days, the mode, are called short-lived; those that lived 9 days are called
average-lived, and those that lived more than 9 days, long-lived.

Short-
lived

Average-
lived

Long-lived

Total

Total

Those living
l>ryi>nd 15 days

Amictic Females

Number

42

31

35

0

108

Mean eggs

19.95

20.45

21.91

Mil t ii IViii, ill--,

Number

29

20

62

14

111

Ml MM t'ggS

14.0

14.05

14.2

14.64

Table VII gives the mean fecundity for these groups. The short-
li\c'l amirtir females produced, on the average, 19.95 eggs per indi-
\i'lual, the av< ii\ed 20.45, the long-lived 21.91. The difference

in tin- mean lei nudity of short-lived and long-lived amictic females is
The short-lived mictic females produced, on the average,
14.0 . ,er indi\ idual, the average-lived 14.05, the long-lived 14.2.

The diliereiHe in ti,c mean fecundity of short-lived and long-lived
mirti< It-male- i- \\hirh is only one tenth of the difference

•mid in tin- <iiiTe^|>oiiding groups of amictic females. The
mean feiundiiy of ihos,- miriir females that lived beyond 15 days,
and thu- -ur\i\<-<| all the ami. tic females, is 14.6 eggs; that is, it is

LIFE HISTORIES OF THE ROTIFER LECANE

365

greater than the mean fecundity of short-lived mictic females by only
0.6 eggs. The positive correlation between length of life and the
number of eggs produced is appreciably more marked for the amictic
females.

TABLE VIII

Lecane inermis. Comparison of the relation of life duration to rate of egg-
deposition (number of eggs per day) for amictic and unfertilized mictic females.
The "mean rate" of egg-deposition is the average of the rates at which all the
members of a particular group deposited their eggs. The rate of egg-deposition of
each individual is computed by dividing the total number of eggs produced, by the
number of days on which eggs were deposited by that individual.

Short-
lived

Average-
lived

Long-lived

Total

Total

Those living
beyond 15 days

Amictic Females

Number

42

31

35

0

108

Mean rate

3.88

3.33

2.73

Mictic Females

Number

29

20

62

14

111

Mean rate

3.0

2.78

2.78

2.87

The length of life of the amictic females bears a negative correlation
with the rate at which they produce their eggs (Table VIII). The
short-lived individuals are found to have deposited their eggs in more
rapid succession than individuals that lived 9 days or longer. They
deposited, on the average, 3.9 eggs per day, the average-lived 3.3 eggs
per day, the longer-lived 2.7 eggs per day. The short-lived mictic
females deposited 3 eggs per day, the average-lived 2.8, the long-lived
also 2.8 eggs per day. Those in the latter group that lived 16 days
or more deposited, on the average, 2.9 eggs per day. The length of
life of the mictic females apparently bears little relation to the rate
at which their eggs are deposited.

It should be noted also that each group of mictic females is char-
acterized by an average daily egg-production almost the same (within
0.3 eggs) as the group of amictic females that produced their eggs
most slowly, namely, the long-lived individuals. The amictic group
with the highest rate of egg-deposition (the short-lived individuals)
produce, on the average, 0.9 eggs per day more than the corresponding
group of mictic females. These facts confirm and supplement data
given in a previous section which indicate that the mictic female
produces its eggs more slowly than the amictic female.

We find, furthermore (Table IX), in the amictic population, that
whereas 25 (60 per cent) of the short-lived individuals died \vithin 24

366

HELEN M. MILLER

TABLE IX

Lecane inermis. Comparison of the relation of length of life to the duration of
the post-fecund period for amictic and unfertilized mictic females. The figures in the
body of the table indicate the number of individuals that survived after the cessation
of egg-production for the number of days indicated on the upper line.

Number of Days

1 or
less

1 to

li

2

3

4

5

6

7

8

9-22

Total

Short-lived Females

Amictic

25

16

1

42

Mictic

2

9

14

4

29

Average-lived Females

Ami< i i.

6

22

3

31

Mirtic

5

2

8

5

20

Long-lived Ft-nuiles

Amictic

5

26

4

35

Mictic

1

4

6

12

6

11

2

2

18

62

Total numr

Amictic

108

Mictic

111

hours after deposition of the last egg, only 6 (19 per cent) of the
average-lived and 5 (14 per cent) of the long-lived died within this
period. Twenty-two (71 per cent) of the average-lived and 26 (74
per cent) of the long-lived females survived for 24 to 36 hours after
deposition of the last egg. Only 8 individuals (7 per cent of the
entire population) survived as long as two days after the cessation of

production.

However, all of the mictic females, with the exception of two
short-lived individuals, lived more than 24 hours after the cessation
of egg production. Most of the short-lived mictic females survived
one .uid one-half or two days. The average-lived survived usually
three days, though one and one-half and three days were almost
equally common. The long-lived individuals, which constitute 56
P« -i" <<•'" »l ihe population, survived one and one-half to 22 days
'tier the on of egg-production, commonly 4 to 6 days. Only

1" I"-' cent survived les - than 4 days; 20 per cent survived longer
than <> da-

[n short, the amictic females that lived for less than 9 days produced

. on the average, in more rapid succession than those that

also pio.lm-ed fewer eggs and died usually within

hour- alier deposition of ihe last egg. These facts suggest that

rapid rate ol rodu< tion exhausts the short-lived amictic females,

rendering them incapable «| completing the transformation of their

LIFE HISTOKIES OK THI-: ROTIFER LECANE 367

germ cells into eggs, thus reducing their fecundity. So strenuous is
the fecund period that none is able to survive more than two days
after its cessation.

In the mictic population, as in the amictic population, the longest-
lived individuals are those that have, on the average, produced their
eggs more slowly, have produced a larger number of eggs and have
survived more than a day after the deposition of the last egg. How-
ever, the length of life of the mictic female is correlated in less degree
than that of the amictic female with the number of eggs produced and
the rate at which they are produced. The short-lived mictic females
usually survive as long after the cessation of egg-deposition as the
longest-lived amictic females; they very infrequently show signs of
physical exhaustion before the completion of egg-production. We
may conclude, then, that the greater longevity of the mictic females,
which involves a greater ability to survive the fecund period, results,
in part at least, from the fact that for them the process of egg-produc-
tion is less strenuous. This is to be expected in view of the fact that
they produce fewer, smaller eggs at a less rapid rate than the amictic
females.

General and comparative

Under similar conditions of cultivation amictic and unfertilized
mictic females of Lecane inermis differ markedly in length of life,
fecundity, rate of egg-production, duration of the fecund and post-
fecund periods and in the degree of correlation between fecundity and
length of life. How are these differences between the two types of
females produced?

In their investigation of the diversities that arise within a clone
of (amictic) females of the parthenogenetic rotifer, Proales sordida,
Jennings and Lynch (1928) concluded that differences in length of
life, among those individuals that hatch, are not an expression of
intrinsic diversity. They result, rather, from the accidents of life,
the interaction of the rhythmic processes of digestion and reproduction,
combined with disturbances in these processes imposed by the daily
transfer of the individuals to clean slides.

In Lecane inermis, the difference in the length of life of the mictic
and amictic female results largely from the difference in the severity
of the process of egg-production, hence in their ability to survive the
fecund period. Both mictic and amictic females lived most commonly
nine days, but many more mictic than amictic females lived beyond
this age, presumably because they produce fewer, smaller eggs than
the amictic female, at a slower rate, and cease egg-deposition at an

368 HELEN M. MILLER

earlier age. Thus, in Lecane inermis, differences in length of life are
probably to be considered of an intrinsic nature only in so far as they
result from intrinsic differences in fecundity.

The mictic females of Lecane inermis resemble the females of
Proales sordida in that some of them are able to live for many days
after the cessation of egg-production. The amictic females of Lecane
inermis, however, never survive more than 3 or 4 days thereafter.
Both mictic and amictic females differ from Proales sordida in the
degree of correlation between length of life and fecundity. For
individuals of Proales sordida that live entirely through the period of
fecundity (24 hours after deposition of the last egg) there is no corre-
lation between length of life and the number of eggs produced; longer-
lived individuals have, on the whole, produced neither fewer nor more
offspring than the average (Jennings and Lynch, 1928 — II, p. 371).
In Lecane inermis, on the other hand, among those individuals that
lived entirely through the fecund period, the longer-lived amictic
females and those mictic females that lived for more than 15 days
have a somewhat greater average fecundity than those that lived a
shorter time.

\\'hat is the origin of the difference in fecundity which is, in Lecane
inermis, largely responsible for the difference in longevity of the
mictic and amictic female? In Proales sordida, intrinsic diversities
in fecundity arise among the individuals of a clone in correlation
with the size of the eggs from which they have hatched. Smaller
(usually early-born) eggs produce less fecund individuals than larger
'later-born) eggs. In order to eliminate diversities that might arise
in this way, only early-born individuals were used in the present
investigation. Although measurements have not been made, there
U no observable difference in the size of the parthenogenetic eggs
that produce the two different kinds of females of Lecane inermis.
In the viviparous species Asplanchna intermedia, on the other hand,
tin- mictic female is apparently derived from a smaller embryo and
i- le-- fecund than the amictic female (Tauson, 1925, p. 144).

\- Jennings and Lynch point out (with respect to Proales sordida},
ditlen -IK <•- in fecundity must proceed either from differences in the
ability of the individuals to transform their germ cells into eggs
under the londition- of cultivation employed, or from differences in
the original number of their germ cells. There is (as they observe)
evident r tli.it the nunilici ol germ cdl>, as well as of the other cells
"I the bod\ ol rotifer-, i- fixed during the embryonic period. Cell
not observed later iii life. Nachtwey (1925,) has shown
'hat the rotili-r . 1 .\/)lnm linn priodonta has usually 32 germ cells,

LIFE HISTORIES OF THE ROTIFER LECANE 369

derived by five cell cleavages from the single primitive germ cell.
In view of this evidence, Jennings and Lynch (1928 — I) have suggested
that differences in the fecundity of the (amictic) females of Proales
sordida that are derived from smaller eggs and those derived from
larger eggs may result from differences in the original number of
germ cells. Late-born individuals may have usually 32 germ cells,
early-born individuals a smaller number, resulting from failure of a
portion of the fourth cell generation to undergo another division.

In Lecane inermis the observed difference in the fecundity of the
two types of female would result if the primitive germ cell of the mictic
female undergoes regularly only four successive divisions, producing
16 oocytes; and if, in the amictic female, one-half of the fourth cell
generation regularly proceeds to a fifth cleavage, thus producing 24
germ cells. Cytological investigation would, of course, be required
in order to determine whether there exists a difference in the original
number of germ cells.

It is possible that, under conditions other than those of this
investigation, the mictic female might be able to bring to maturity as
many germ cells as the amictic female. It should be noted, however,
that the differences in the fecundity and length of life of the two
types of females have been observed during cultivation under widely
different conditions; namely, in oat infusion and in Benecke solution
at temperatures ranging from 18° to 24° C.

The observations of Whitney (1907) regarding the fecundity of
mictic and amictic females of Ilydatina senta are of interest here.
At 20° to 22° C. both types of females produce about the same number
of eggs; at 24°-26° C. the mictic female produces twice as many eggs
as the amictic; at 26°-290 C. the mictic female produces four times
as many eggs as the amictic female. Thus, in Hydatina senta, differ-
ences in the fecundity of the two types of female result observably
from differences in response to the environment. The mictic female
lives at all these temperatures slightly longer than the amictic female
(Wesenberg-Lund, 1929). The relative longevity of the amictic and
mictic female of Ilydatina senta is not correlated with the relative
fecundity, as in Lecane inermis.

For other species, information regarding the relative length of life,
fecundity and rate of egg-production of the mictic and amictic female
is not extensive.

The life history records of three mictic females of Euchlanis
triquetra (Lehmensick, 1926) indicate that they produce the same
number of eggs as the amictic female (24) and live about the same
length of time, 21 days. But the mictic female produces its eggs

370 HELEN M. MILLER

more rapidly and thus lives longer after the cessation of egg-produc-
tion.

The mictic female of Asplanchna intermedia (Tauson, 1925)
produces fewer eggs, as mentioned above, and lives a shorter length
of time than the amictic female. The mictic female matures on the
first day, produces 10 to 12 eggs on the second day, which hatch on
the third day; the mother dies after the hatching of the last male.
The amictic tVmale lives 5 or 6 days, produces daily no more than
4 eggs (probably 16 to 20 in all). The mictic female produces its
eggs appreciably more rapidly.

In /'trrotlnni elliptica (Luntz, 1926), the mictic female produces
regularly six or seven eggs, the amictic female only five. There is no
available information regarding the length of life. Physiological
di\er>ity is manifested by the inability of the mictic female to with-
stand certain conditions of osmotic pressure and pH in which the
amictic female is perfectly normal.

YVesenberg-Lund states (1930, p. 31) that "if not fertilized,
investigations hitherto carried out seem to show that the number of
eggs laid by the two sorts of females is almost the same, but that
those of the mictic female are laid in a much shorter time (Lehmensick,
\(>2(>, I-.iK'lilnnis triguetra)"

\\'e find, however, that the mictic females of some species, e.g.,
Asylum hnti intermedia and Lecane inermis, are regularly less fecund
than the amictic females, whereas those of some other species, e.g.,
Pterntlimi rlli/)ti/<i and Ilydatina senta are, usually, more fecund.
With regard to the relative rates at which the eggs are deposited by
the two sorts of females, Lecane inermis is exceptional; the mictic
female produces its eggs less rapidly than the amictic female.

Wesenberg-Lund (1930) presents a comparison of the mictic and

amictic females on the basis of his extensive investigations carried

out largely in nature, but supplemented by laboratory observations.

ithor stresses the physiological and biological differences between

the inn t ic and amiciic female and gives some additional data regarding

then relative fecundity, rate of egg-production and length of life.

splanchna .wV/Wr// the mictic females mature and begin to deposit

earlier than the amictic females. In isolation cultures they

prodmed 11 to if, eggs and lived 1() to 12 days. Amictic females,

i'i mass • ultures, lived only S days and produced 8 to 10 eggs. The

author i- in. lined to U-lie\e (p. 15ui that in nature the amictic females

ve than the mictics, e-pe« ially at low temperatures. O>n-

i he "conjectures that at the same temperatures
live -ome days longer than the mictic ones;

LIFE HISTORIES OF THE ROTIFER LECANE 371

produce more young ones; and to a somewhat higher degree are able
to accommodate duration of life and time of production of young ones
to temperature" (p. 149). Regarding the rate of production of the
parthenogenetic eggs, Wesenberg-Lund states (p. 210) that, in several
species, "the mictic female is able to subdivide the yolk mass into
small amounts, and produce about 12 eggs simultaneously, whereas
the amictic females produce the eggs successively." Further evidence
of physiological diversity is presented by certain parasitic species in
which the amictic females are usually free-living, the mictic females
always parasitic. Peculiarities in behavior exhibited by mictic females
of some species after fertilization indicate that they are structurally
as well as physiologically diverse.

In summary, the evidence at hand is sufficient to demonstrate
that in different species the relative fecundity, length of life and rate
of egg-production of the mictic and amictic female vary greatly.
In every species which has been studied physiological differences
between the two types of female are demonstrable; but investigations
have not proceeded sufficiently far to admit of general conclusions
regarding the nature of the fundamental diversity that distinguishes the
mictic from the amictic female.

FERTILIZED MICTIC FEMALES

In the foregoing only unfertilized mictic females have been con-
sidered. The mictic female of Lecane inermis, like that of Pterodina
elliptica and Brachionus bakeri (Luntz, 1926, 1929), and Asplanchna
sieboldi (Wesenberg-Lund, 1930, p. 155), may be fertilized during its
immature period or after having deposited parthenogenetic eggs.
In some species the mictic female can be fertilized only when very
young (Eucldanis triquetra, Lehmensick, 1926).

The mictic females of Lecane inermis resemble those of most
species in that they may be fertilized by their own sons. The females
of Pterodina elliptica are exceptional in that they are incapable of
being fertilized by the male progeny of the same grandmother. In
Lecane inermis cloacal copulation occurs. The toes of both male and
female are bent ventrally, almost at right angles to the axis of the
bodies, which are extended in opposite directions. The female
continues to swim about, pulling the male along with her. Copulation
lasts for at least five to ten minutes. The cloacal method of copulation
is uncommon in rotifers, according to Wesenberg-Lund (1929). The
males of most species attach themselves to the body of the female and
in some manner pierce the body wall and deposit sperm directly in
the body cavity.

HELEN M. MILLER

As noted by all investigators of the bisexual rotifers, fertilization
produces a visible effect upon the vitellarium. This organ becomes
black during the growth of the fertilized egg. This, it is generally
believed, results from the deposition of fat and the production of
larger, darker yolk granules. In the course of 24 to 36 hours a fertilized
egg matures and is deposited. The vitellarium is then quite colorless.
The production of yolk and the deposition of a fertilized egg may be
repeated four or five times.

Information regarding the fecundity and length of life of fertilized
mictic females of this species is derived from the life history records
of 18 individuals fertilized during immaturity and 5 individuals
fertilized by their own sons. The individuals fertilized before ma-
turity produced only fertilized eggs, the total number not exceeding
five. They lived usually 4 or 5 days after the cessation of egg-
production; the total length of life ranged from 10 to 25 days.

The life history records of the five females fertilized by their
own sons at various times during maturity are given in Table X.
These individuals produced 5 to 9 parthenogenetic eggs before being
fertilized. They produced thereafter one to 4 fertilized eggs. They
are typical in their ability to survive the fecund period. The total
length of life ranged from 13 to 18 days.

Mictic females of this species have not been observed to deposit
parthenogenetic eggs after once having produced fertilized eggs.
This is not the case in Ilydatina senta (Shull, 1910) and in several

• ulier species. Lehmensick (1926) suggested that fertilized mictic
females return to the production of parthenogenetic eggs because all

• •I their germ cells were not fertilized at the first mating, and \Yesen-
berg-Lund has observed that after a second mating the production of
fertilized egg- may be resumed (1930 — II, p. 154). The failure of
fertilized mictic females of Lecane inermis to produce parthenogenetic

- probably indicates that more of their germ cells were fertilized

than they were capable of bringing to maturity; for, frequent ly, after

'he i ti< MI of egg-deposition the vitellaria contained dark yolk

and leriili/ed eggs were present in various stages of growth.

I i 1 1 ili/.iiion i niimioiily reduces the fecundity of the mictic female.
Lecane in,'r»ii\, which produces, when fertilized, only 5 to 10 eggs
rather than ](>. \~~ not exceptional in this respect. The length of life
i-. in 'hi- .jiparently not affected by fertilization.

Tine MALES

I'"' males »t thU species (Figs. 4 and 5) are strikingly different
in appearance I mm the females, and, like most male rotifers, are to

LIFE HISTORIES OF THE ROTIFER LECANE

373

X

W
C3

H

^n -c _c

rt £, >

•g «^

4-> fll 0)

*-> n
^ o
c -^

•>-g
o .£f

; 'in
0)

a
o

TD
OJ

hjO c ^*
3-H

P c

. be rt

^ in ^

'5 D -r
-a -Q .i

*r^ C "^

^ 5 £

3 = 0,

O *^ t p! en

IN HH b/3

cx c^ be

in . 4j

bjo ^^^ ™r^

be X '^ <U

O .2 O N

a, u -t-1

'•£ 3 >>4J

4J cj J3 ti

5-Srfl

rt^:~"c

£5s.i

C'C a-^
•§1^1

O „ 01 C

'_ in r-

p QJ *]~t Q

^i-£ c "

rt "*^ O 'o

a; in

a> "T3 3

^ 4^ n 3

p R) •"• c

S "rt <" •-

5 oj >< in

- N nj b£

O 'H

M in" S 3

•T1 U O , *~!

sl-a s

U C: oj n

" o rt

'5 &£

•llT

s 'o

"O

o .£

in in

i- i_
<u o

e £ s

333

•a c c

•— W)

P

oo

2

CO

i— i

'^

*— (

Fecundity

~a
H

-

-

O

0

a

Fertilized
eggs

«

-

•H

«

«

3 i

«,

-

0,

oo

0

o
J
'o

CTJ

Q

00

o

o

0

-

o

o

o

vO

o

0

o

2

o

0

o

2

o

0

o

-o
o

2

0

T3

0

o

o

0

2

o

O

—

o

o

-

0

o

o

o

o

2

0

-

o

o

0

B,

o

-

0

o

0

00

o

-

o

o

-

r^

o

-

-

1 — 1

-

*O

o

0

- *o

1 — 1

•«-H

-,

o

•
CM *b

-^

-b

- "b

*

-

ro'D

«-b

ro-b

-'b

«

-

N-b

^«0

«-b

esib

(N

10 'b

O

o

- "b

- 'b

-

o

O

c

c

o

c

-

^

-

'0

25

374 HELEN M. MILLER

some extent degenerate in structure. The females are somewhat
flask-shaped, flattened dorso-ventrally, and become pigmented as they
The males, on the other hand, are nearly cylindrical, transparent,
and are somewhat shorter and about one-half the width of the newly-
hatched female. The toes of the female are long and slim, those of
the male are short and thick. The lorica of the male, especially in
the posterior region, shows a greater degree of segmentation than that
of the Female, and the body is thus more flexible. One of the most
striking peculiarities of the male rotifer is a black, disk-shaped body.
It consists of a vesicle containing large, dark granules, which are
supposed by Leydig to be excretory in nature (\Yesenberg-Lund,
\{>2(), p. 327). This structure is located dorsally in the posterior
region of the body.

The iVm. iK's more than double their size through growth (compare
Figs. 2 and 3); the males do not grow. This lack of the power of
growth is characteristic of the males of most species of rotifers and is
associated with the absence of a functional digestive system. In
I.i'niue iuerniis, no traces of the mastax and digestive glands, so
prominent in the female, have been observed in the male. Apparently,
only the rudimentary anterior portion of the digestive tube persists,
.nid this serves, as in other species (Wesen berg-Lund, 1923, 1929),
,1- ,i siispensor ligament for the testis (compare Figs. 3 and 5).

The excretory system shows some structural reduction in the ab-
sence of the contractile vacuole.

The reproductive system is well-developed. The testis is large
and fills most of the body cavity. The vas deferens is also conspicuous.
Two large bodies lying ventral to the testis and vas deferens may
be the so-called prostate glands, two or four <>t which are usually
present in the male rotifers. The structure and method ot functioning
ol the copulatory organ has not been observed.

A prominent retro-cerebral organ and eye-spot are present, as in
the female. The activity and flexibility of the male indicate well-
de\e!..ped neiAoiis system and musculature.

The male- hatch from the small eggs of the mictic female after a

developmental period somewhat exceeding in duration that of the

Four m. ilc parthenogenetic eggs, allowed to develop in

-t.nid.ird oat infusion at 23° ('., required two hours longer to develop

than lour female part heiio^enel ic eggs laid at the same time and

subjected \<> identical conditions during development. The average

time required by the male eggs was 31 hours 1C) minutes, by the female

-'' li»m- 15 minute-, a dilference of one hour and 31 minutes.

In •' suspension <>l Chlnn-lln and B. proteus in Henecke solution of

LIFK HISTOklKS <)!• THK KOT1FKR LKCANK

375

0.07 per cent concentration, at 22°-24° C., the male embryos require
on the average four hours longer for hatching than the female embryos.
The duration of the embryonic periods of male and female partheno-
genetic eggs deposited by young parents and allowed to develop at
the same time are given in Table XI.

TABLE XI

Lecane inermis. Comparison of the duration of the embryonic periods of
contemporaneous male and female parthenogenetic eggs, deposited by young mothers,
and allowed to develop at 22° to 24° C. The eggs were isolated in pairs on hollow-
ground slides, usually a single male and female egg in each depression, in a suspension
of Chlorella vulgaris and B. proteus in Benecke solution of 0.07 per cent concentration.

Number of hours
after deposition

Number of contemporaneous eggs hatched

Female

Male

30

18

0

32

6

0

34

1

7

36

1

10

38

0

9

39 — or longer

0

6

Total 26

32

Male rotifers have been reported to be sexually mature at the
time of hatching or shortly thereafter. This is true also of the males
of Lecane inermis. During the first few days they are extremely
active. They copulate indiscriminately with fertilized or unfertilized
mictic females, or with the amictic females. This seems to be generally
characteristic of rotifer males. The duration of sexual activity has
not been investigated.

The males that hatch live usually 4 to 6 days. They are very
active. They swim dorsal side up, commonly in a less direct path,
and rotate less than do the females. The males of this species are
also typical in the lack of the power of growth. On the fourth or
fifth day, activity begins to wane until locomotion ceases entirely.
The organism lies on its side in a semi-contracted condition, the
anterior portion bent ventrally. This condition culminates in death
usually within 24 hours,.

For investigation of the length of life and rate of mortality, male

376

HELEN M. MILLER

eggs were isolated within 24 hours after deposition, from small mass
cultures during a period of three weeks. These individuals, as well
as the mictic and amictic populations previously described, belonged
to the same pure line (descendants by parthenogenesis of a single
stem mother). Of the 110 eggs isolated, 6 individuals that hatched
were lost in transferring, and 22 (20 per cent) failed to hatch. Nine
of those that failed to hatch were discarded before it was discovered
that in many cases the embryos developed completely and were very
active within the shell. Only one of the remaining 13 failed to
develop. The other 12 individuals lived and were active the usual
length df time within the shell; two lived longer than any of those

that hatched.

TABLE XII

Lecanc inermis. Length of life of 95 males, including 13 which developed but
failed to hatch. The length of life is reckoned from the time of isolation of the eggs
(within 24 hours after deposition). For comparison of the males with the females,
the differences in the biometric constants of the different populations are given at the
end of the table. The biometric constants for the female populations are given in
Table II.

Nutnlii-r nf days

i

2

3

4

5

6

7

8

Total

Those that hatched

0

1

1

4

?0

SO

ft

0

82

Those that developed but failed to hatch . .

1

0

0

0

1

7

2

2

13

Mean

Standard
Deviation

Coefficient of
Variability

Total studied (95)

5.69 ± 0.07

1.10 ±0.05

17.48 ± 0.88

: h.,nlu-d (82)

5.65 ± 0.06

0.83 ± 0.04

14.73 ±0.79

1 >i:t> • .•. 1 1 11

Mali-- and I-' i:

Ami. tii 108 M.d. 5 (95

3.21 ±0.13

0.72 ±0.09

1.80 ± 1.27

Midi. - 11 1 M.d. 95

541 ± 0 29

3 37 ± 0 M

M 93 ± ? ? 7

I .il.le XII -li.)\vs the life duration of those males, hatched and not
i'li'-'l '^ in<li\ Mual- ., I. .i \\hich records are complete. They lived
<""• to v <!,, Sixty per cent of the population died on the sixth
<l.iy, which was the modal length of life. Twenty-two per cent died
"" lll(1 I'ltli day .in.l only a very few on any other day. The mean
of lil<- ,,| tlu^r that hatched is 5.65 days; including those that

LIFE HISTORIES OF THE ROTIFER LECANE 377

failed to hatch the mean is 5.69 days. The males live three days
less than the amictic females, five and one-half days less than the
mictic females. The standard deviation is significantly lower than
that of both types of females; the coefficient of variability is only
slightly less than that of the amictic females, but less than that of
the mictic females by 22 per cent.

In Fig. 10, the rate of mortality of male and female populations
is compared. Those that developed but failed to hatch are included
in the graph. The rate of mortality increases very gradually until the
fourth day, at the end of which seven and one-half per cent of the
males, but no females are dead. During the fifth and the sixth days,
88 per cent of the male population die, while mortality is just beginning
in the female population. All of the males are dead one day before
mortality reaches its height in the female population.

The male and female populations differ in hatchability, in the
mode and mean length of life, and in the standard deviation; but
in the variability of life duration the males and the amictic females
do not differ appreciably.

Is the difference in hatchability of the male-producing and female-
producing eggs due to an inherent difference in fertility of the mictic
and amictic females, or to differences in the ages of the parents or to
the conditions of cultivation? The males were, for the most part
(77 per cent), derived from females under four days old; all the females
studied were deposited by individuals less than four days old. Most
of the male-producing eggs that failed to hatch were derived from
parents more than four days old, but six were early-born. Thus,
7 per cent of the early-born males and only 0.5 per cent (one in 220)
of the early-born females failed to hatch. The female embryos that
fail to hatch never live more than one or two days. Since the female
and male populations were not cultivated at the same time, however,
the failure of male embryos to hatch may have resulted from some
environmental difference, and cannot be regarded as conclusive
evidence that mictic females produce parthenogenetic eggs that are
regularly less fertile than those from amictic females. Information
regarding the relative fertility of the parthenogenetic eggs from mictic
and amictic females is not available in the literature, in so far as I
am aware.

The uniformly short length of life of the males in comparison with
that of the females has been observed in other lines investigated on a
smaller scale, in which, however, both sexes were cultivated at the
same time.

The males of Lecane inermis live longer than the males of most

HELEN M. MILLER

other species hitherto described (Wesenberg-Lund, 1930, p. 158).
The large males of the viviparous Asplanchna species live four or five
days. The moderate reduction in length of life, in comparison with
that of the females, is, in Lecane inermis, in accord with the moderate
degree of structural degeneracy. Differences in the degree of struc-
tural degeneracy of the males of different species may bear some
correlation with differences in the degree of reduction in the size of
the male-producing parthenogenetic eggs and in the rate at which
the eggs mature and are deposited, two factors which \Yesenberg-
Lund (1930) considers to be largely responsible for the structural
degeneracy of the male rotifers. Evidence in support of this view is
derived from a comparison of the males of Lecane inermis with those
of the plankton rotifers. The former, which are characterized by a
moderate degree of structural degeneracy and of reduction in length
of life, are derived from eggs which are appreciably smaller than the
female-producing eggs, but which are deposited less rapidly, and
require a longer time for development. The males of the plankton
rotifers, on the other hand, are derived from very small eggs, a large
number of which, according to \Yesenberg-Lund, are produced almost
simultaneously, and develop rapidly. These males are small, consist
of little more than reproductive system and live (in some cases) only
a few hours. It should be borne in mind, however, that the tendency
toward structural degeneracy and reduced vigor of the male rotifers
may result from the haploid condition of their chromosomes (Morgan,
l')26. Chapter X). The knowledge regarding these matters is at

present very meager.

SUMMARY

This paper is the first contribution of a series dealing with the
life cycle of the bisexual rotifer Lecane inermis Bryce. The three
t\pes of individuals, the mictic females, the amictic females and the
malt--, and the eggs from which they are derived, are described and
their life histories are compared statistically. The individuals studied
are intrinsically uniform members of the same genetic stock, cultivated
mulct uniform environmental conditions, although they were not all
contemporaneous. The comparison is based upon about 100 repre-
sentath es <>i each type of individual.

The ilnrr type:-, of individuals are characteristically different in
length o| lite. The mi-ail length of life is, for the amictic females,
- 9 0.11 days, for the mictic females, 11.1 ±0.28 days, for the
in. i n.iiT days. This difference is apparent in their life

Embryonic mortality is negligible in all three populations.
Mortality imie .(dually in the male population until the fourth

LIFK HISTORIES OF T1IK ROTIFKR LECANE 379

day, then increases suddenly, S3 per cent of the population dying on
the fourth to the sixth days. All are dead by the eighth day. All
the females survive through the fourth day; thereafter the rate of
mortality increases more rapidly in the amictic population. More
than half of the mictic females survive the modal life duration, which
is the same for both kinds; 12 per cent to 15 per cent survive the
longest-lived amictic female and die off very gradually until the
twenty-eighth day. The short life duration of the males is un-
doubtedly correlated with the structural degeneracy, especially of their
digestive and excretory systems.

The longer life of the mictic females, as compared with the amictic
females, results, in part at least, from the fact that egg-production is
for them a less strenuous process. This is evidenced by the following
facts. (1) Their male-producing eggs are smaller. (2) The mictic
female produces regularly only two-thirds as many eggs as the amictic
female. The maximum number is, for the mictic female, 16, excep-
tionally 17, for the amictic female 24. The mean number of eggs
per individual is, for the mictic female 14.2 ±0.11, for the amictic
female, 20.7 ± 0.13. A higher proportion of mictic females produce
their maximum number of eggs. The difference between the mode
and the maximum is, for the mictics, only one, for the amictics, three.
(3) The mean duration of the period of fecundity of the amictic
female is 6.4 ± 0.09 days, with a standard deviation of 1.44 ± 0.07
and a coefficient of variability of 22.5 per cent; the mean for the
mictic female is 5.1 ± 0.05 days, the standard deviation 0.83 ± 0.04,
the coefficient of -variability 16.2 ±0.75 per cent. (4) During this
time the mictic female deposits, on the average, one egg every 8.6
hours, the amictic female deposits one egg every 7.5 hours. (5) The
amictic females die usually within 24 to 36 hours after deposition of
the last egg; 19 per cent of the mictics lived more than six days there-
after. In summary, the mictic female produces in less rapid suc-
cession only two-thirds as many eggs as the amictic female, requires
for the production of her eggs 1.3 days less than the amictic female
and usually lives longer after the cessation of egg-production.

A comparison was made, for the mictic and amictic females, of the
degree of correlation between length of life and their fecundity, rate
of egg-production and ability to survive the fecund period. Short-
lived amictic females are found to have produced fewer eggs than
long-lived individuals, in more rapid succession, and to have died
within a few hours after the deposition of the last egg, apparently
exhausted by the severity of the process of egg-production. The
length of life of the mictic female is correlated only in slight degree

HELEN M. MILLER

with the number of eggs produced and the rate at which they are
produced.

The mictic female of this species may be fertilized during im-
maturity or after having produced parthenogenetic eggs. The fecun-
dity of the mictic female is reduced by fertilization, the total number
of eggs per individual, in 18 cases studied, not exceeding ten. The
maximum number of fertilized eggs produced by a single individual
was five. The length of life of the mictic female is not appreciably
altered by the production of fertilized eggs.

The differences cited above are evidence of fundamental physio-
logical diversity between mictic and amictic females and males of

line inermis.

LITERATURE CITED

1 INESINGER, J. E., 1926. Effect of Certain Chemical and Physical Agents on
Fecundity and Length of Life, and on their Inheritance in a Rotifer, Lecane
(Distyla) inermis (Bryce). Jour. Exper. Zool., 44: 63.

HARKING, H. K., AND MYKRS, F. J., 1926. The Rotifer Fauna of Wisconsin. III.
A revision of the genera Lecane and Monostyla. Trans. Wise. Acad., 22:
315.

JENNINGS, H. S., AND LYNCH, R. S., 1928. Age, Mortality, Fertility, and Individual
Diversities in the Rotifer Proales sordida Gosse. I. Effect of age of the
parent on characteristics of the offspring. Jour. Exper. Zool., 50: 345.
II. Life-history in relation to mortality and fecundity. Jour. Exper. Zool.,
51: 339.

Li II.VKNSK K, R., 1926. Zur Biologic, Anatomic und Eireifung der Radertiere.
Zeitschr.f. li'iss. Zool., 128: 37.

LINTZ, A. 1926. Untersuchungen iiber den Generationswechsel der Rotatorien.

I. Die Bedingungen des Generationswechsels. Biol. Zentralbl., 46: 233.
Li M/, A., 1929. Untersuchungen iiber den Generationswechsel der Radertiere.

II. Der zyklische Generationswechsel von Brachionus bakeri. Biol.
Zentmlhl., 49: 193.

M(IK<.AN, T. II., 1926. The Theory of the Gene. Vale University Press, New

Haven.
\\MII\M.Y, R., 1925. Untersuchungen iiber die Keimbahn, Organogenese und

Aii.iiuiiiic \(jn Asplanrhna priodonta Gosse. Zeiischr. f. u'iss. Zool., 126:

KAI i M , I >., 1930. The Effect of Conjugal ion within a Clone of Paramecium aurelia.
Hi»l. Hull., 58: 2<>3.

1 ., 1nll). Studies in the Life ( \.lc of llyl.nina senta. I. Artificial
"iniirol of the tr.niMtion from the parthenogenetic to the sexual method of
rrprndiirtion. Jour, I-'.xfer. Zonl., 8: 311.
", < '., 1924. I >ic Li/ellen der heu-rogonen Radertiere. Zool. Jahrb. Abt. Anal.

int., 45: MY>
I ' v>. <»., 1924. IM<- l\cihm--|iio/(--r der parthenogenetischen Eier von

Vsplanchna im<-r li.i lhi'l>. Zeitschr.f. Zell. u. Geweb., 1: 57.

< '., 1(>J5. Wirkimi; de> Mediums auf das Geschlccht des Rotators
!i« Im.-i intermedia lluds., Teil 1. Int. Rev. d. ges. Hydrobiol. n.
r., 13: ISii.

1923. ' • •iitribiilioiis to the Biology of the Rotifcra. I-
the Rot if era. Mem. Acad. Roy. Sci. Lett. Daneitiark, Copen.
, 4: 189.

1. 1 1-M-; HISTORIES OF THE ROTIFER LECANE 381

WESENBERG-LUND, C., 1929. Rotatoria. Handb. d. Zool., 2: 1. \\'alter de Gruyter

& Co., Berlin und Leipzig.
WESENBERG-LUND, C., 1930. Contributions to the Biology of the Rotifera. II.

The periodicity and sexual periods. Mem. Acad. Roy. Sci. Lett. Danemark,

Copenhague, Sec. Sci. Serie 9, 2: 3.
WHITNEY, D. D., 1907. Determination of Sex in Hydatina senta. Jour. Exper.

Zool., 5:1.
WHITNEY, D. D., 1929. The Chromosome Cycle in the Rotifer Asplanchna amphora.

Jour. Morph. and Phys., 47: 415.

PERMEABILITY AM) FATIGUE IX MUSCLE AM) ITS
BEARING ON THE PROBLEM OF ION ANTAGONISM

E. GELLHORX '
(From the Department of Animal Biology, University of Oregon, Eugene, Oregon)

It has been shown by numerous investigators that stimulation of
different tissues leads to an increase in permeability which is indepen-
dent of the particular structure and function.- In the case of muscle,
\Yeiss showed that even subminimal stimuli lead to an increase in
permeability which was measured by an increase in the output of
phosphoric acid by the muscle. When effective stimuli were applied,
a close relationship between the intensity of fatigue and the increase
in permeability was observed (Embden and Adler, 1922). Further-
more, there was a parallelism between the reversibility of fatigue and
that of the changes in permeability. When the muscle recovered from
fatigue and the original contractility was again obtained, the per-
meability was also normal, i.e., no phosphoric acid was given off to the
medium in which the muscle was suspended. But if no recovery took
place, the output of phosphoric acid remained the same or even in-
creased. These experiments prove that the physico-chemical changes
in permeability play an important part in the fatigue problem. It
seemed probable that fatigue might be delayed if it were- possible to
prevent or to diminish the increase in, permeability which ordinarily

"inpanies fatigue. The experiments described in this paper show
the correctness of the supposition. Fatigue can be delayed in a me-
dium by which the permeability of the muscle is decreased. The
eiticiency of different ions in that respect and the general importance
of thoc facts for the problem of ion antagonism will be discussed in this
paper.

METHOD

Sexer.il hundred experiments on the sartorius muscle of Rana cs-

'ciitn were performed from March 1929 to October 1930. The mus-

< le~. prep. in-d in the ii-u.il manner, were immersed in an isotonic salt

-i 'hit it Hi .UK! Mi^pended between two platinum electrodes connected

\\nli an i-oionic lexer \\lncli magnified the contractions seven times.

"I'-'l tin- V • \l«irri.s..n Prize in Experimental Biology for 1930 by the

Aiili-d li\ .1 ^i.ini Inun tin- Ki-M-.trrh l-'uml of the University of Oregon.
•• l!h«in, \<>1'>, pp. 106-197.

ION ANTAGONISM

In the first group of experiments the platinum electrodes were carefully
covered with lacquer (except on a point-like spot where the platinum
was in contact with the muscle). The muscles were stimulated with
rhythmical condenser discharges (apparatus of Scheminsky) 40-60
times per minute so that a decrease in the height of contraction oc-
curred. The twro sartorii of the same frog under the same conditions
showed an identical course of fatigue (Gellhorn, 1930). This made it
possible to study the influence of different ions. The solutions were
made up of NaCl, KC1 and CaCl2 solutions of A = 0.39°. In order
to hold the conductivity of the liquid constant and avoid any possible
effect on the strength of the stimulus, a part of the sodium chloride
solution was replaced by an equal volume of isotonic CaCl2 solution,
and thus the effect of different CaCl2 concentrations was studied.

In a second group of experiments another method was used. The
sartorii were not stimulated in a liquid but in a moist chamber. The
first series of stimuli was applied for five minutes and the muscle was
then allowed to recover in an aerated salt solution. Afterward a
second series of stimuli was applied until the muscle ceased to respond.
The first series of stimuli was designed to show whether both sartorii
had the same irritability and fatigability. Pairs of muscles which be-
haved alike in the first series of stimulation were the only ones con-
sidered as material for the determination of the influence of different
salt solutions. The effect of these solutions was apparent in the course
of fatigue during the second series of contractions. The advantage of
this second method consists in the independence of the strength of the
stimulus from the conductivity of the solution and furthermore in the
exact proof of the identical behavior of the muscles before they were
immersed in the salt solutions. In these experiments condenser dis-
charges obtained from the apparatus of Scheminsky (1930) in its im-
proved form were used. Maximal stimuli were given 90 times per
minute. Between the first and the second series of stimulations the
muscle remained unstimulated for nine minutes in the salt solutions.
The experiments of the second group were performed with curarized
and uncurarized muscles. The effect of Ca, Sr, Ba, and Mg was
studied in reference to the fatigue of muscle.3

EXPERIMENTS PERFORMED WITH THE FIRST METHOD
Figure 1 shows the course of the height of contractions of the
sartorii stimulated 40 times per minute for twelve minutes. One
muscle was in NaCl + KC1 + CaCl2 solution in which these salts had

3 After preparation the muscles remained in aerated Ringer's solution for 45-60
minutes before the experiment was begun.

384

E. GELLHORX

the same concentration as in Ringer's solution, while the other muscle
was bathed in XaCl + KC1 + CaClo which contained eight times as
much CaCl2 as the first one. It is apparent that, in spite of the identi-
cal height of the contractions at the beginning of the experiment, both

\

\

\

Ft*;. 1. The course of fatigue in a pair of sartorii stimulated in 0.111 M

NaCI + 10.2 M X 1(T4 KC1 + 8.1 M X 10^CaCl2( ) and in 0.103 M NaCl + 10.2

M in ' K< I - M.4 M X 10~4 CaCl. ^ -).

< ir.linatc: bright of contraction; Abscissa: time in minutes.

Maximal < <in<l<-nst'r discharges; frequency 40 per minute.

-how the "treppe" in the same amount: — the muscle immersed in a

M .hit ion with a high ('a(T> concentration showed a lesser fatigue than

the < "utrol muscle. In other words an excess of CaCI-2 delays fatigue.

Tlii- 1. n-t , which was the ki-is tor further experimentation, is illustrated

2, which reproduces a series of experiments performed with six

p.i irtorii. The control muscle was always stimulated in 0.1 1 1M

; o.iKMOJM KC1 + o.oooXIM CaCl2( the other muscle in

1V<>M or <).no«)72M CaCl2) the KC1 concentrations being identical

in l>otli cases .uid the \a('l slightly less in order to keep constant the
osmotic pressure and c onductivity of the solution. The contractions
were rej istered at inter\-al- of two minutes for a period of twelve min-
nte- ,ind ihc height of contraction of the muscle stimulated in the

ION ANTAGONISM

385

°/

7*

ito

ifo

zoo

180
160

/zo

too

o z + 6 e 10

FIG. 2. The influence of different CaCU concentrations on the course of fatigue
in sartorius muscle. The height of contraction of the control muscle (stimulated in
0.111 M NaCl + 10.2 M X 10~4 KC1 + 8.1 M X 10~4 CaCl2) is taken as 100 and the
percentage of the height of contraction of the muscle placed in a solution with higher
CaClo concentration is calculated.

0.106 M NaCl + 10.2 M X KT4 KC1 + 48.6 M X 10~4 CaCl2.

0.0996 M NaCl + 10.2 M X 10~4 KC1 + 97.2 M X 10~4 CaCl2.

386

E. GELLHORX

higher CaCl2 concentration was calculated as the percentage of the
height of contraction of the control muscle. The figure shows that
the height of contrartion in the beginning of the experiment exhibited
very slight variations if any, but in the further course of the experi-
ment the deviations became progressively stronger and reached in
most cases 100 per cent or more. This illustrates again the delay of
fatigue in solutions with higher CaClo concentrations and shows at
the same time that the increase of CaClo to twelve times its normal
concentration (i.e., the concentration in Ringer's solution) is still
favorable for the preservation of contractility and the delay of fatigue.
The fundamental importance of this fact is apparent when the

n

18

\

\

\

\

V-

I K.. .v The course of irritability (threshold) in unstimulated muscles.
1 MM a ni <• IM-I ttrrii primary and M-coiidary coil; abscissa: Time in hours.
—0.111 M NaCl + 10.2 M 1 i ' KC1 + 8.1 M X 10~4 CaCl2.
0.103 M NaCl + 10.2 M : ; in ' KC1 + (.4.4 M X 10"4 CaCl,.

Ordi-

d| the same -"lutions is studied on the irritability of the unstim-
ulated muscle. As was to be expected, the solution with the "normal "
• "in cut ration is much more favorable to the preservation of
irritability than a Dilution with higher CaCl2 content (Fig. 3). Therc-
tlif "ptmial i Mudiiidiis are dilTerc-nt for the stimulated and un-
-timnlated mu-( le. That is to say, ///c laics of ion antagonism which
Ifinl to tin- di\i »:', ry of the c</nil ihr<ilcil solutions do not hold quantitatively
Jor tin- stininliiti-il must I,-. These experiments make probable an even

ION ANTAGONISM

387

more general formulation that the efficiency of ions on cells depends
upon their degree of permeability.

EXPERIMENTS PERFORMED' WITH THE SECOND METHOD

In order to check these results more than 100 experiments were
performed with the second method. Figure 4 illustrates the course of
fatigue after the muscles have been immersed in solutions of different
CaCl2 concentrations. The curves I and II show the fatigue curve
of the two muscles when stimulated in the moist chamber. One recog-
nizes that they behave identically. After this they were immersed

FIG. 4. Stimulation of a pair of sartorii in a moist chamber with maximal
condenser discharges 90 times per minute. Then muscle I is immersed in 0.111
M NaCl + 10.2 M X 10~4 KC1 + 8.1 M X 10"4 CaCl2 (control fluid) and muscle II
in 0.111 M NaCl + 10.2 M X 10~4 KC1 + 81 M X 10~4 CaCl2 for nine minutes.
Thereafter stimulation in a moist chamber as before (curve la, Ha respectively).
The numbers at the curves indicate the time in minutes after the beginning of
stimulation.

for nine minutes in solutions of NaCl + KCL + CaCl2 which were
alike save for the CaCl2 concentration. This was 8.1 M X 10~4 in the
first case and 81 M X 10~4 in the latter one. Finally they were trans-

...

E. GELLHORN

ferred again into the moist chamber and stimulated in the same way
as in the first series. The experiments yielded the curves la and I la
and prove that the solution rich in CaClo preserved the contractility
of the muscle much better than that with the lower CaCl2 concentra-
tion. The height of contraction which was observed in the first muscle
(treated with 8.1 M ] [ 10~4 CaCl2) after five minutes was the same as

I 1C. 5. Procedure as in V\g. 4. Muscle I immersed in control fluid; muscle II
in (i.l 1 1 M Nad + 10.2 M X ICT* KC1 + 40.5 M X 10~4 CaCl>.

in il nd muscle (treated with 81 M X 10~' CaCl2) after fourteen

niiimii .

Am it her rv.miple is reproduced in Fig. 5. It is of particular inter-
est l>r< ause in the first series of contractions the second muscle shows
spontaneously ;i greater fatigability than the first one. Nevertheless,
t In- t rc.it incut • .1 tin- muscle with the solution containing a higher CaCl2
<••incrntr.it 11 -a < >\ ri < "inpensated the deficiency, since this muscle re-
main- contra* iilr IW •}() minutes and the control muscle for only five

ION ANTAGONISM

389

minutes. In a group of experiments the solution with the higher CaCl2
concentration contained the same NaCl concentration as the other one
so that the osmotic pressure of the first solution was slightly higher,
while in a second group the NaCl concentration was a little lower in
order to keep constant the osmotic pressure in spite of the differences
in CaCl2 concentration. In both groups the solution rich in CaCl2

FIG. 6. Procedure as in Fig. 4. Muscle I (Fig. 1 and la) immersed in control
fluid; muscle II in 0.111 M NaCl + 10.2 M X 10~4 KC1 +40.5 M X 10~4 CaCl2
(Fig. II and I la).

26

390

E. GELLHORN

delayed the fatigue, and this result was also obtained if by addition of
XaHCO3 the salt solution were completed to a typical Ringer's solu-
tion. Figure 6 shows this in a very striking way because, in spite of
the weaker contractions of the second muscle, the fatigue is delayed
after the muscle has been immersed in a Ringer's solution with five
times its normal CaCla concentration.

FK.. 7. Pr.M-r.lurc as in Fig. 4. Muscle I in 0.111 M NaCl + 10.2 M X 10~4
KCI + SI M J KIM ,,ri2 (1 and la); muscle II in control fluid (II and lla).

I li( LI', orable Hirrt of the solution with higher CaClo concentration
in delaying t.ni^iic m.iv be illustrated by Table I. in which the time of
the priori v.ttion (,|' ( mn ractility was determined for each pair of

ION ANTAGONISM

391

muscles after they had been immersed in salt solutions with different
CaClo concentrations.

TABLE I

The Dependence of the Duration of Contractility upon the CaCh Concentration

After a previous period of stimulation for five minutes, the muscles were im-
mersed in 0.111M NaCl + 0.00102M KC1 + different CaCl2 concentrations as
indicated in this table. After this, stimulation was continued in a moist chamber.

No.

Cads Concentration

Duration

CaClz Concentration

Duration

minutes

minutes

1

S.1M X 10-4

5

40.5M X 10-4

> 17

2

i t

5

40. 5 A I X 10-*

> 40

3

11

6

81 M X 1CT4

> 15

4

i (

4

81 M X 10-"

> 9

5

i i

4

SIM X 10~4

46

6

t (

6

81 M X ICT4

20

The experiments with the second method reveal still another fact
which seems to be of interest for the problem of the tonus of skeletal
muscle. The curves of Fig. 5 in particular show that after the inter-

FIG. 8. Procedure as in Fig. 4. Solutions as in Fig. 7.

392 E. GELLHORN

val during which the muscles have been immersed in salt solutions the
stimulation leads to a contracture which is slightly stronger in the
Ca-muscle. Furthermore, the diminution of this contracture is de-
layed in the Ca-muscle:1 In this experiment the contracture still in-
creased one minute after the beginning of stimulation while in the con-
trol muscle it diminished. This behavior is even more pronounced in
the experiment illustrated by Fig. 8. in which the Ca-muscle showed
an increase in tonus during almost the entire period of stimulation,
although the control muscle showed a rapid loss of tonus during this
time. The tonus effect of calcium becomes apparent when the CaCl2
concentration is 81 M X 10~4 or higher. In reference to the delay of
fatigue the muscles show an unequal behavior. In most cases the
fatigue was delayed in spite of the increase in tonus, but in some cases
< ompare Fig. 7) the tonus increased so rapidly that it led to an acceler-
ation of fatigue.

The behavior of the stimulated and unstimulated muscle in solu-
tions containing different CaCl2 concentrations was compared and it
\\as concluded that stimulation entirely changed the reactivity of the
muscle to this particular ion. This fact was explained by the increase
in permeability of the muscle. The observations on the effects of
calcium ions on the tonus of skeletal muscle seem to support this view
since Neuschlosz (1922) observed that the immersion of the resting
muscle in Ringer's solution with increased CaCl2 concentration lead
always to a decrease in tonus. It is probable that the muscle brought
into a state of high permeability by previous stimulation takes up CaCl2
in a greater amount than the resting muscle and this leads to entirely
different phenomena. It may be emphasized that about the same
concentrations were used in Neuschlosz' experiments and our own.
11 the interpretation is correct it is to be expected that far greater con-
centrations of CaCl-j would be able to increase the tonus in the resting
muscle. In fact it was found by Guenther (1905) that CaCl2 in one
per lent volution leads to an increase in tonus which occurs after a
latent period of several hours. Probably the irritability of the muscle
laded under these conditions. In contrast to that, the experi-
ments described above show that CaCl- may immediately raise the
tomi- of the irritable nui-cle provided that the permeability had been
increased by previous stimulation. Therefore not only by K as Neu-
-chlo-/ loimd, bui al-o by ('a the tonus of muscle can be increased
without decrease of irritability.

I i-.miiL.' a-ide the effects of CaCl2 on the tonus of muscle, the ex-

I "i 'I" i>ie\iiy tin expression Ca-musclc is used for the muscle which

>een m luii.iH loniaining the higher CaClj concentration.

ION ANTAGONISM

393

periments lead to the conclusion that increase in CaCl2 concentration
through a very wide range reduces the higher degree of permeability
produced by stimulation, and therefore delays fatigue. Further ex-
periments dealt with the question whether the Ca effect is specific or
whether it is replaceable by other cations. The effects of Sr, Mg and
Ba were examined.

The experiments showed that none of the alkali earths can replace
calcium. Rather small concentrations of Sr, Ba and Mg had only a
slight effect if any, and greater concentrations accelerated fatigue.
MgCl2 seemed to have the most harmful effect, since it reduced the
life duration of the muscle which had been previously immersed in a
solution of NaCl + KC1 + MgCl2 + NaHCO3, while the effect of Sr
and Ba was mostly limited to a decrease in the height of concentration.
The concentrations used are given in Table II.

TABLE II

Each solution ordinarily contains beside the salts with bivalent cations 0.111M
NaCl + 10.2M X 10~4 KC1 + 12.5M X 10~4 NaHCO3.

Concentration

Effect on Fatigue

Mg

15.4M X 10-4
23. 1M X ICT4
30. 8 M X ICT4

—

Ba

23.9M X ICr4
31.8M X 10-4

—

Sr

16.7M X 1CT4
32.5M X ID"4
40. 7 M X 10~4
81.4M X 10~4

—

In another series one muscle was immersed in a solution of NaCl -f-
KC1 -f- CaCl2 + NaHCOg, the composition of which was identical
with that used for the control muscle save for the addition of SrCl2,
BaClo or MgCl2. In these experiments it was found that all three
cations were able to delay the fatigue of the muscle but the effect was
much less than in the experiments described above in which an excess
of CaCl2 was studied. Furthermore, it was found that the favorable
effect was limited to a very small range of concentrations. When this
was surpassed the fatigue was accelerated. The quantitative data
are reproduced in Table III.

394

E. GELLHORX

TABLE III

The Effect of Sr, Ba and Mg on the course of fatigue if added to O.lllM NaCl
+ l0.2Kf X 10~* KCl + 8.1 M X 10-*CaCh + 12.5M X 10~* NaHCO3.

Concentration

Effect on Fatigue *

Sr

32.SM X Ur4
40.7M X 10-4

t

Ba

19.8M X 10-4
23.9M X 10-4
S1.SM X 10-4

i

Mg

15.4M X 10-4
19.2M X 10-4
30.8M X 10~4

±

* + indicates delay in fatigue; — , acceleration in fatigue; ±, no change.

DISCUSSION

The experiments described in this paper prove that the laws of ion-
antagonism differ quantitatively in the resting and the stimulated
muscle. The increase in permeability in the latter leads rapidly to
fatigue and this can be delayed when an excess of Ca is added which
reduces permeability. On the other hand, a Ringer's solution con-
taining an excess of CaClo is unfavorable for the preservation of ir-
ritability in resting muscle. The fundamental difference in the be-
havior of resting and stimulated muscle in reference to ions is further
illustrated by experiments in which the effect of other bivalent cations
was studied. It was found that Sr, Mg or Ba added to Ringer's solu-
tion with normal ('a('l-j concentration exerted a favorable influence on
the stimulated muscle, that is, the fatigue was delayed. But the same
solutions were unable to preserve the irritability of the unstimulated
nm-rle as well as Ringer's solution. The experiments permit one to
i i include that tin < nm posit ion of the optimal salt solution is not a definite
one tnr n ( erttiin type of tell hut is dependent upon its degree of permeability.

This implies the fact that the efficiency of ions differs quantitatively
i! applied to a < I'll in a state of low or high permeability. The experi-
ments mi the lontis effect of calcium described above show the correct-

- "I 'In- i OIK hiHon. The increase in tonus observed in the stim-
ulated miiM-lr begins immediately after stimulation, while, according
'" ( 'iM-nthei , it require:- a ( a concentration ten times higher and several
hours in rr-i ini; nni-clr.

I he experiments thn.\v an interesting light upon the specificity of
They indicate that, in contrast to the ion antagon-
iellhorn, l'»Jl .md 1926.

ION ANTAGONISM 305

ism between K and bivalent cations in muscle in which numerous
cations are of great antagonistic efficiency, calcium cannot be replaced
by any other bivalent cation in its effect of delaying fatigue. This is
astonishing because the suppression of the K contracture by bivalent
cations is based upon their power to reduce permeability (Gellhorn,
1928). The writer found recently in experiments on vital staining in
sea urchin eggs that the permeability of the cell to stains could be
prevented by Ca but not by Mg, Sr or Ba (1931). But the explanation
of the specificity seems to be quite different in sea urchin eggs and mus-
cle, since in the former Sr, Ba and Mg were unable to decrease per-
meability, while in muscle permeability is decreased. This is not
only indicated by observations on K-contracture but also by the fa-
tigue experiments of this paper, since it was shown that addition of Sr,
Ba or Mg to Ringer's solution containing CaClo was effective in reduc-
ing fatigue. The reason why Ca cannot be entirely replaced by Sr,
Ba and Mg seems to lie in the fact that Ca is probably the only ion
which diminishes the increase in permeability of muscle in a completely
reversible manner. However, when coarser changes in the surface of
the cell are concerned, such as are produced by K, several bivalent
cations act antagonistically. Fewer of these ions are effective if irrita-
bility is concerned, as in paralysis of muscle (Hober, 1917), than in the
case of K-contracture, which is rather independent of irritability. In
the former case Ca is the most antagonistic ion, in the latter case the
heavy metals (Gellhorn), because here the greater the power in reducing
permeability, the more efficiently is the contracture suppressed. The
impossibility of fine reversible changes in permeability, which are
necessary for the preservation of irritability, is not concerned here.
The intermediate case, the restoration of irritability after K-paralysis,
is influenced in a somewhat intermediate manner: calcium is most
effective, but other bivalent cations are effective to a lesser degree.
The importance of calcium for the preservation of irritability and con-
tractility is apparent. It is conceivable that the delay of fatigue re-
quires a still finer adjustment than the partial restoration of irritability
after K-paralysis. This may be the reason for the specificity of the
calcium effect in the fatigue experiments of this paper.

SUMMARY

1. The fatigue of the sartorius muscle of the frog can be delayed by
an excess of calcium chloride in Ringer's solution, although this
solution does not preserve irritability of resting muscle as well as Ring-
er's solution. Therefore, in contrast to the view generally accepted
today, the quantitative composition of the optimal solution depends
upon the permeability of the cell.

396 E. GELLHORN

2. This conclusion implies a quantitative difference in the effi-
ciency of ions in relation to permeability. This is illustrated by ob-
servations on the increase of muscle tonus in solutions with high CaCl2
content. The change in tonus takes place in irritable muscle.

S. The effect of calcium on fatigue is specific. Neither Sr, Ba nor
Mg can replace it. But these cations can delay the fatigue, although to
a lesser extent than calcium, when added to Ringer's solution contain-
ing 8.1 M '. \ 10~4 ('a('l-j. The cause of this specificity is discussed.

BIBLIOGRAPHY

EMBDEN, < '.., AND ALI.KR, E., 1922. Zeitschrift f. physiol. Chemie, 118: 1.

< .M.I. HORN, K., 1(>24. In Oppenheimer's llandbuch der Biochemie, 2: 98.

i ,11 i ii' 'K\, 1.., !''_'(). Nrutre Ergebnisse der Physiologic. Leipzig, Chapter 2.

GELLHOKN, I-".., l'J2S. 1' tinker's Arch., 219: 761.

GELLHORN, E., 1929. Das Permeabilitatsproblem, seine physiologische uml patholo-

gischc lit -ik-utung. Berlin.
^GEI.I.HORN, 1.., 1'MO. Contributions to Marine Biology, p. 115. Stanford University

Press.
Gi-:i.i.!i"k\, E., 1931. Protoplasma, 12: 66.

< .1 i.VMihk, A. I. , 1905. Am. Jour. Plivxiol., 14: 73.
H6BER, K., 1917. Pjlu-fr'.-i Arch., 166:531.

Nil » HLOSZ, S. M., \(>22. Printer's Arch., 196: 503.
N HEMINZKY, K., 1928. /.citxclir. /'. Hi»l., 87: 1S9.
SCHEMINZKY, F., 1 (>30. Pfl user's Arch., 225: 303.
WEISS, 1!., 1922. Ptlii^rs Arch., 196: 393.

THE RECOVERY CONTRACTURE IN MUSCLE; A NEW
GENERAL SALT EFFECT1

E. GELLHORN

(From the Department of Animal Biology, University of Oregon, Eugene, Oregon.)

In a study on permeability and fatigue in muscle and its bearing on
the problem of ion antagonism, the method used in a series of experi-
ments was as follows. The sartorius muscle was stimulated for five
minutes with condenser discharges about 60-90 times per minute.
Thereafter the muscle was allowed to rest for at least five minutes in
Ringer's solution with normal or increased Ca content and then a new
series of stimulations of the same frequency was begun. The muscle
was in a moist chamber when stimulated. Under these conditions a
peculiar phenomenon was observed. In the beginning of the second
stimulation series the muscle showed a contracture which became less
during the continuation of the stimulations and finally disappeared
more or less completely. The phenomenon is fundamentally different
from the fatigue contracture in muscle which increases with the con-
tinuation of stimulations. The experiments described below serve

to analyse the phenomenon.

METHOD

The experiments were carried out on m. sartorius and biceps of
Rana esculenta. The carefully prepared muscles were suspended
between platinum electrodes and stimulated by means of the apparatus
of Scheminsky, which allowed one to use condenser discharges over a
wide range of frequencies. The strength of the stimulus was varied by
a parallel resistance; there was also a resistance of 10,000 ohms in
series in order to make very slight changes in the conductivity of the
muscle ineffective. The stimuli were either just maximal or submaxi-
mal. The muscles were loaded with 2 g; the magnification of the
isotonic lever was six fold. The experiments were performed from
October 1930 to January 1931.

RESULTS

I. THE BASIC PHENOMENON

A typical example is reproduced in Fig. 1. In this experiment the
sartorius was stimulated 90 times per minute with a condenser of 0.5
mf. Between the periods of stimulation the muscle recovered in a

1 Aided by a grant from the research fund of the University of Oregon.

397

398

E. GELLHORX

well-aerated Ringer's solution. Each period of stimulation and
recovery lasted for five minutes. One recognizes from Fig. 1 that the
curves b-d show in the beginning contractures which decrease while
stimulation is being continued. There is a regular decrease in the

FIG. 1. Four stimulation periods on sartorius. Condenser discharges (0.5 mf.)
90 times per minute, in a moist chamber. Muscle is in Ringer's solution during
recovery period. The conditions in all succeeding experiments are the same unices
otherwise specified

hci-ht of contracture from b to d.~ This makes it plain that the phen-
omenon in question is basically different from the fatigue contracture
since (1) fatigue contracture increases with the continuation of stimu-
lation and (2) in spite of a slight decrease in the height of contractions
tiom /* to il, i.e., an increase in fatigue, the contracture becomes less.
Thi-- peculiar behavior of the contracture makes it rather probable
that it might In- due to a recovery and not to a fatigue process. As to
the height of contractions, it is characteristic that although the con-
traction-, decrease in the very beginning of the contracture until the
maximum of the contracture has been reached, hereafter a remarkable
increase in the height of contractions is accompanied by a decrease in
< ontrai i M

ihci< till other facts which support the hypothesis that the

phenomenon a recovery contracture. It is known that with in-
creasing fatigue the duration of the contraction increases. The de-
tailed -nid\ ha- -dio\vn that the change in duration chieHy concerns
relaxation. At firsl one may see that an elastic vibration ends the

I" ili' 1( •••' .mil iii tin- figures the different stimulation periods are designated
as a, 1 In- in -i ulni h i .HI -li<>\v recovery contracture is the b curve.

RKl'OVKKY CONTRACTl'RF IN Mt'SCI.K

399

period of relaxation. With increasing fatigue the elastic vibration
disappears and a slow contraction (designated as "Funke's Nose") is
apparent in the downstroke of the contraction. The more fatigue
progresses the earlier in the downstroke appears this slight contraction.
Since the relaxation period of the "Funke's Nose" is much longer than
in a normal contraction the duration of the contraction increases in
correspondence with the earlier appearance of the "nose." The regis-
tration of the recovery contracture on a faster revolving drum revealed
the interesting fact that the changes described as characteristic for
fatigue (Funke, Wachholder) also occur during the recovery contrac-
ture but in the reverse order as is shown in Fig. 2, in which tracings
obtained at intervals of one minute are recorded.

FIG. 2. 6-curve on a faster moving drum.

It was stated above that the recovery contracture decreases with
increasing number of alternating stimulation and recovery periods as is
apparent if the recovery contracture in Fig. 1, b is compared with 1, c or
1, d. Experiments on a fast revolving drum revealed that the changes
in shape and duration of the contraction correspond to simultaneous
changes in the size of contracture. In an experiment, for instance,
which was carried out in a similar manner as that reproduced in Fig. 1,
it was found that the disappearence of the "Funke's Nose" required
one minute in experiment d, four minutes in experiment c and more than
five minutes in experiment b. In other words, the recovery phenom-
enon judged in its intensity according to the height of contracture and
to the time during which the "Funke's Nose" appears decreases the

400 K. GKLLHORX

more the fatigue increases. That is, of course, to be expected, since
recovery is more marked in a slightly fatigued muscle than in one
which has been very much fatigued. Thus the phenomenon is un-
doubtedly a recovery contracture and the question arises as to what
factors are responsible for its occurrence.

II. TlIE INFLUENCE OF STRENGTH, FREQUENCY, AND DERATION OF STIMULI ON THE
OCCURRENCE OF RECOVERY CONTRACTURE

SyMematic studies showed that the recovery contracture is rather
independent of the strength of the stimuli since it occurs in experi-
ments with submaximal as well as maximal condenser discharges.
This makes it plain that the contracture is different from Tiegel's
contracture, which requires the application of supermaximal currents.
However, the frequency of stimulation had a marked influence.
Kxperiments were carried on in which the frequency of stimuli in the
first and second periods varied independently over a range of from 30
to 150 per minute. It was apparent that at least 45 stimulations
\\ere required in the first period to bring about the contracture
in the second, provided that during this period the frequency was
at least one hundred. Under these conditions, the higher the fre-
quency employed in the second period, the more marked the con-
tracture. But this is true only up to a frequency of 120 per minute,
-•ince beyond this value the contracture dropped slightly. It is easily
understandable that stimulations below 45 per minute (luring the
lust period did not produce a recovery contracture in the second even
if very high frequencies were used, since under these conditions in the
first period no fatigue was observed. Therefore, a recovery phenom-
enon must fail to appear.

In general it may be said that the higher the frequencies used in
both periods, the more marked the recovery. But even after a first
period of hiirh frequency the contract lire does not occur unless the nuis-

/' < urve in t\\'> -.irinrii of the same fro,j. l;rr<|iK-nry of stimulation in
•curvi minute; in the second, 120 per minute.

RECOVERY CONTRA CTURE IN MUSCLE

401

cle is stimulated at least 45 times per minute in the second "period.
The strength of the recovery contracture is not only apparent in its
height but also in the changes in the shape and duration of the contrac-
tion. This may be illustrated by Figs. 3 and 4. In Fig. 3 the recovery
contracture of two sartorii of the same frog are reproduced. The fre-
quency during the first period was the same in both cases (90 per min-

FIG. 4. i-curve. Frequency in a-curve was 60; in b, 90 per minute.

ute), while in the second period the first muscle was stimulated 60
times and the second muscle 120 times per minute. Fig. 4 shows that
the "Funke's Nose" disappears in the downstroke of the contraction
within one minute, frequency of stimulation being 60 in the first and

FIG. 5. 6-curve. The two sartorii were stimulated in the a period with a
condenser 0.1 mf.; in the b period of the first muscle a condenser of 2.0 mf. was used,
in that of the second a condenser of 0.1 mf.

90 in the second period, while if during both periods the muscle is
stimulated 90 times per minute, it takes much longer for the duration of
contraction to become normal. (Compare Fig. 2.)

The duration of the stimulus also seems to affect the recovery

402

E. GELLHORX

contracture. In Fig. 5 the recovery contracture of two sartorii of
the same frog is reproduced from an experiment in which strength and
frequency of the stimulus were identical, but the duration of the stim-
ulus was in the first muscle 70o- and in the second 20<r. ( )ne recognizes
in Fig. 5 that increasing duration of the stimulus increases recovery
contracture. But even with very short currents the phenomenon can
be observed, since recovery contracture is obtained in sartorius after
application of submaximal induction currents. It also shows under
these conditions the characteristic features mentioned above which
concern the occurrence of the " Funke's Nose" and the duration of the
single contraction.

III. Tut: INFLUENCE OF SALTS AND XON-ELECTROLYTES ON THE OCCURRENCE OF

RECOVERY CONTRACTURE

This paragraph deals with the problem of the physico-chemical
factors which determine the occurrence of the recovery contracture.
In the first group of experiments one muscle recovered in the moist
chamber while the other muscle of the same frog recovered in Ringer's
Mention as in the previously described experiments. In order to get

two ^.irtorii of the same frog. Upper tracing: recovery
in moist. ( lum!" r. I.UWIT tracing: recovery period in Ringer's solution.

RECOVERY CONTRACTURE IN MUSCLE

403

distinct results it is advisable to avoid soaking the muscle in Ringer's
solution for a long time when it is afterwards to be in the moist chamber
during the recovery period.

Duliere and Horton observed recently that a muscle which is
brought into a moist chamber without having been in contact with
Ringer's solution lost its irritabil ity. The observations of these authors
were confirmed and therefore the muscle was immersed in Ringer's
solution for two minutes before being placed in the moist chamber.
Under these conditions the irritability of the muscle did not change
and the very striking result was regularly obtained that the muscle
which recovered in a moist chamber did not show recovery contracture,
while the muscle immersed in Ringer's solution during recovery period
always showred the contracture. Figure 6 gives an illustration of the

FIG. 7. a-g curves of sartorius. Between a and d period, muscle in moist
chamber; between d and g, in Ringer's solution during recovery.

typical results for three periods (b-d). While the "Ringer" muscle
undergoes the typical recovery contracture, decreasing more and more
with each stimulation period, the "air" muscle does not show any con-
tracture.3 The change in the shape and duration of the contractions is
also restricted to the muscle which recovered in Ringer's solution and
showed recovery contracture.

Experiments of this type were performed in great number without
a failure. It is interesting to note that, as Fig. 7 indicates, alternating
conditions bring about alternate presence and absence of the contrac-

3 For the sake of brevity the muscle which remained in Ringer's during the
recovery period is referred to as the "Ringer muscle," that which remained in the
moist chamber, the "air musrle."

404 E. GELLHORN

ture of the muscle. In this experiment the muscle remained during the
first three recovery periods in a moist chamber and therefore did not
show a recovery contracture; after the fourth stimulation period it
always remained in Ringer's solution during the recovery period. The
ect is very significant. The long period in air influenced the be-
havior of the muscle so that one recovery period in Ringer's did not
immediately bring about the contracture. Therefore it is not notice-
able in experiment c, but it becomes apparent in / and still more in g.
Tin- time during which the muscle was kept in Ringer's solution
before the experiment was started is not immaterial to its reaction in
reference to the recovery contracture. Thus it was found that soaking
tin- muscle for 24 hours in Ringer's solution increases its tendency to
react with recovery contracture even if the muscle is kept in a moist
chamber during the recovery period. This may be illustrated by Fig.
8. The first curve showed a typical although very slight recovery

I [G. x. Both s;irtorii soaked for 24 hours in Ringer's solution. First muscle
immediately thereafter stimulated .i> in Fig. 1, the second muscle remained one hour
in moist chain! ><T lie-tore stimulation. Muscles in moist chamber during recovery
period, ^-curves reproduced.

<•< nit met ure. Hut the contracture was still smaller in the see. mil

sartorius of the same frog, which was held in a moist chamber for one

hour l)el<>re start ing the stimulation. It was, of course, also soaked for

24 hours in Ringer's solution. Besides, the second muscle shows but

i i lei ic, isc in the elastic vibrations during contracture while the

first one «lid not show them at all. This indicates, as was proven by

•• a fasl re\ olving kymograph, that in the first case the

< "Hi rat t ute was at ipanied by the occurrence of the " Funke's nose,"

which did not occur in the second case. Summing up these observa-
tions, it may be -aid i hat recovery contracture occurs if the muscle
kept in Kinger'- solution during the recovery period and it disap-
pear- it the muscle is in a moist chamber during that period. The
effects are reversible. The soaking of the muscle in Ringer's solution
tin llnT- the ret OVCry < "lit racture.

RECOVERY CONTRACTURE IN MUSCLE 405

Seeking for an explanation of this peculiar phenomenon it seemed
rather probable that the presence or absence of the contracture might
be due to differences in ion concentrations at the surface layer of muscle,
since the efficiency of ions to produce contractures and influence the
"tonus" of muscles is known (Neuschlosz, Gellhorn). But experi-
ments did not prove the correctness of this view. The recovery con-
tracture occurs not only if the muscle is kept in Ringer's solution but
also if the composition of that solution is much altered. An increase in
the K or Ca concentration to ten times its original value did not notice-
ably influence the strength of the recovery contracture. The phen-
omenon was also observed in isotonic NaCl, Na2SO4 and LiCl solu-
tions. If the K or Ca content of the muscle were responsible for the
phenomenon in question, changes in concentration of these ions should
be effective. Therefore it was thought that one might have to do here
with a general salt effect as was first observed by Loeb (for further
references compare Gellhorn, 1926 and 1929, pp. 153-163). In this
case the contracture should be diminished or should not occur at all if
the muscle were placed in a non-electrolyte solution during the re-
covery period. In fact, the experiments carried out with glucose,
sucrose and urea either alone, in isotonic solution or replacing different

FIG. 9. fr-curves. (m. sartorius.) Muscle 1 in Ringer's, muscle 2 in 3 cc.
Ringer's + 17 cc. isotonic urea during recovery period.

parts of Ringer's solution, yielded the expected results. Figures 9 and
10 show a typical experiment on sartorius and biceps. In Fig. 9 the
recovery contracture was much decreased if the muscle remained in a
solution consisting of Ringer's + urea during the recovery period. In
Fig. 10 it is shown that recovery contracture is suppressed completely
if the muscle is immersed in isotonic sucrose solution during the re-
covery period.4

4 In some of the experiments it was observed that the sartorious muscle showed
a small contracture when immersed in isotonic non-electrolyte solutions. Therefore
in most cases mixtures of Ringer's and non-electrolyte solutions were employed which
did not produce a contracture during the recovery period.

27

406 E. GELLHORN

As mentioned above, the experiments were carried out on the sar-
torius and biceps of Rana esculenta. The results were the same in
both muscles. This seems of interest in respect to observations of
Sommerkamp, who found that some groups of muscles react more to
reagents which produce contractures, especially to acetylcholin, than
others. This fact was explained by structural differences because the

FIG. 10. 6-curves. (m. biceps.) Muscle 1 in Ringer's, muscle 2 in isotonic
sucrose during recovery period.

content of "tonusfibers" is different in different muscles. The obser-
vations of this paper seem to indicate that in the recovery contracture
only the ordinary muscle fibers are involved since biceps and sartorius
behaved rather alike in spite of the differences observed by Sommer-
kamp in experiments with acetylcholin.

As to the explanation of the recovery contracture, several facts
support the assumption that the condition of the surface layer of the
muscle is the determining factor for its appearance. This is not meant
in the sense of the theory of Botazzi, who believes that sarcoplasm is
responsible for "tonic" reactions, but there are a number of observa-
tions which show more or less the dependence of contractions and con-
tractures upon the surface layer of the muscle cell and its permeability.
It will be recalled that fatigue in muscle can be delayed by an excess
of calcium in Ringer's solution (Gellhorn), and furthermore, that K
and SC contractures depend entirely on permeability (Gellhorn).

The importance of the behavior of the surface of muscle for the
recovery contracture is expressed by two groups of observations: 1 . It
-liown that the recovery contracture is increased by the applica-
tion of condenser discharges with relatively long duration of discharge.
As was recently proved by v. Gulacsy and Heller such discharges bring
about marked polarisation phenomena on the surface layer of muscle.
Experiments were described demonstrating that the recovery
diminished or suppressed in the presence of non-elec-
trolytes shows that a general salt effect is present as in the
experiments of Loci,. It still remains a question as to whether the
same process underlies the salt effect in Loeb's and in our own obser-

RECOVERY CONTRACTURE IN MUSCLE 407

vations. But the experiments seem to point out that the conditions of
the colloids in the surface layer of the cells are different when recovery
contracture occurs and when it is prevented.

SUMMARY

If the sartorius or biceps of Rana esculenta is stimulated with con-
denser discharges in periods of five minutes with a frequency varying
between 45 and 150 per minute and stimulation periods alternate
regularly with recovery periods of the same duration, a recovery con-
tracture is brought about in the beginning of the second and each
following stimulation periods provided the muscle is allowed to re-
cover in an aerated salt solution. Wide changes in the K and Ca
concentration of the Ringer's solution are without influence. The
contracture also occurs when isotonic solutions of Na or Li salts are
employed. The phenomenon is suppressed by replacing the salt with
a non-electrolyte. The following observations are in favor of the as-
sumption that the contracture is a recovery contracture:

1. The phenomenon does not occur if the frequency is too low to
produce even a slight fatigue.

2. It decreases from the second to the fourth stimulation period,
i.e., with increasing fatigue.

3. The changes in the shape of the contractions (Funke's Nose)
during the recovery contracture show a temporal sequence just the
opposite of that in fatigue.

It is assumed that changes in the surface layer of the muscle cell
bring about the recovery contracture. In favor of this hypothesis are
the facts:

1. The recovery contracture is increased with increasing duration
of the electrical discharges which were used to stimulate the muscle.

2. Non-electrolytes which suppress the phenomenon bring about
changes in the surface layer of cells.

BIBLIOGRAPHY

BOTAZZI, FIL., 1901. Ueber die Wirkung des Veratrins etc. auf die quergestreifte

Muskulatur. Arch. f. Physiol., p. 377.
DULIERE, W., AND HoRTOx, H. V., 1929. The Reversible Loss of Excitability in

Isolated Amphibian Voluntary Muscle. Jour. Physiol., 67: 152.
FUNKE, O., 1874. Ueber den Einfluss der Ermiidung auf den zeitlichen Verlauf der

Muskelthatigkeit. Pfliiger's Arch., 8: 213.
GELLHORN, E., 1927. Experimented Untersuchungen zur Permeabilitat der

Muskulatur. Pfliiger's Arch., 215: 248.
GELLHORN, E., 1928. Zur Kenntnis der Kaliumcontractur am quergestreiften und

glatten Muskel. Pfliiger's Arch., 219: 761.
GELLHORN, E., 1929. Das Permeabilitatsproblem, seine physiologische u. allgemein-

pathologische Bedeutung. Berlin. (Jul. Springer.)

408 E. GELLHORX

GELLHORX, E., 1931. Permeability and Fatigue in Muscle and its Bearing on the
Problem of Ion Antagonism. Biol. Bull., 60: 382.

GELLHORX, E., 1931. SCN Contracture in Skeletal Muscle: Further Studies in
Permeability of Muscle. Am. Jour. Physiol., 96: 203.

GELLHORX, E., 1931. Further Experiments on Direct and Indirect SC\ Contrac-
ture. Am. Jour. Physiol., 96: 477.

v. GULACSY, Z., 1930. Permeabilitat und Ermiidung. I Mitt. Pfliiger's Arch., 223:
407.

HELLER, R.. 1930. Permeabilitat und Ermiidung. Ill Mitt. Pfliiger's Arch., 225:
194.

LOEB, J., 1916. The Mechanism of the Diffusion of Electrolytes through the Mem-
branes of Living Cells. Jour. Biol. Chem., 27: 339, 353, 363.

XEUSCHLOSZ, S. M., 1922. Untersuchungen iiber die \\~irkung von Neutralsalzen auf
den tonischen Anteil der Muskelzuckung. Pfliiger's Arch., 196: 503.

SCHEMIXZKY, F., 1930. Uber Reizung mit rhythmischen Kondensatorentladungen.
Pfliiger's Arch., 225: 303.

SOMMERKAMP, H., 1928. Das Substrat der Dauerverkiirzung am Froschmuskel.
Arch.f. exper. Pathol. u. Pharmakol., 128: 99.

TIEGEL, E., 1876. Ueber Muskelcontractur im Gegensatz zu Contraction. Pfliiger's
Arch., 13: 71.

WACHHOLDER, K., 1930. Untersuchungen iiber tonische und nicht tonische Wirbel-
tiermuskeln. Pfliiger's Arch., 226: 274.

NOTE ON DR. JAN HIRSCHLER'S PAPER ON SPINDLE

BRIDGES, AND THE DE NOVO ORIGIN OF MOTH

SPERM ATOCYTE GOLGI BODIES

J. BRONTE GATENBY1
TRINITY COLLEGE, DUBLIN

INTRODUCTION

In a recent paper Jan Hirschler (1929) has described fully in
Macrothylacia rubi L., a remarkable relationship between certain
cytoplasmic elements, and the nucleus. He states that he has found
this in other genera of Lepidoptera, namely, Phalera, Dasychira, and
Dendrolimus.

Hirschler's technique consisted in Zenker fixation followed by iron
alum staining, and chrome-osmium acid fuchsin for Golgi bodies.
Apparently the Zenker method shows no cytoplasmic inclusions but
does show very remarkably a connection between hemispherical
nucleoli and spindle bridges, such as does not appear to have been
described before. In text figures A and B are reproductions of two

N

TEXT FIGS. A AND B
After Jan Hirschler (1929a)

A. Zenker fixed cells showing spindle bridges and attachments to nucleoli
("tres analogues aux heterochromosomes d'autres Insectes").

B. Osmic Golgi preparation, showing Golgi bodies (G) originating around a
spindle bridge space (C). Ordinary nucleolus, N. Both figures from Macrothylacia
rubi.

1 Theresa Seessel Fellow, Osborn Zoological Laboratory, Yale University.

409

410 J. BRONTE GATENBY

of Hirschler's figures illustrating in the first, Zenker hrcmatoxylin
prepared cells with two curious hemispherical nucleoli at the nuclear
membrane, each connected with an extension into the cytoplasm,
described by Hirschler as "residus fusoriaux." The nucleoli lie
claims to be analogous with the heterochromosomes of other insects.
Prepared by an osmic Golgi method, one gets the picture shown in
text figure B. The hemispherical nucleolus with its pale halo, has a
small object protruding from it into the cytoplasm; a normal or
ordinary nucleclus (TV) adheres to the hemispherical one. Now, just
in this region there is a pale zone (C) surrounded by Golgi bodies.
Hirschler says, "Nous considerons cette zone comme le residu fusorial
et nous pensons que, sous 1'influence de la substance nucleaire qu'elle
renferme et qui touche au cytoplasme, il se forme dans celui-ci, par
precipitation des elements de 1'appareil de Golgi qui peuvent ensuite
perdre contact avec le residu fusorial et se transporter en d'autres
regions de la cellule."

In a previous paper (Gatenby, 19310) the present writer has
referred to the same idea of Hirschler, by which a new formation of
secondary Golgi bodies is supposed to take place even in the spermatids
of a pentatomid. This hypothesis of Hirschler (19296) has been
dealt with in the previous paper, and is considered untenable so far
as it applied to insect germ cells studied by the present writer.

The Material and Methods. — The material used in the present
work consisted of a large series of normal, x-radiated and radium-
radiated testes of Abraxas. No material from other moths is at
present available, but a large number of lepidopterous testes from
other families has been studied in recent years.

The Spindle Bridge in Abraxas. — In Fig. 2 of the plate is a group

DESCRIPTION OF PLATE

Letters: B, bacteria. BR, branch of spindle bridge. D, dead cell. GB, Golgi
body. L, living cell. SB, spindle bridge. A', dying cell. Y, normal-looking cells.
All Abraxas and, except Fig. 3, by Champy haematoxylin.

FIG. 1. Group of spermatocytes, some normal-looking (Y), one dying (A'), as
the result of x-radiation.

FIG. 2. Rosette of spermatogonia showing live and dead cells as result of
gamma-radiation. Spindle bridge at SB. Such cells already contain Golgi bodies as
in next figure.

S'assanow method. Rosette spermatogonia with numerous Golgi
These cells are much smaller than those in text figures A and B, in which
Golgi bodies are supposed to be arising de novo.

Two spermatocytes, with metaphase spindles, showing dense bacterial
infection in one, and freedom from infection of the other.

Two spermatids, one badly infected, the other containing one bacterium.
!• IG. 6. Group of growing spermatocytes showing spindle bridges, mitochondria
and neutral red vacuoles (just above the bridges).

NOTE ON DR. JAN HIRSCHLER'S PAPKR

411

PLATE I

M

412 T. BRONTE GATENBY

of spermatogonia connected by blackly-staining spindle bridges.
When the cells hollow out to form the typical nest of growing spermato-
cytes, each unit is found to be connected to its neighbour by long
lighter-staining bridges which pass right across the cytoplasm, in
front of the nuclei, as shown in Fig. 6, SB. In no case has the spindle
bridge been found to lie in any other region.

In spermatogonia and young spermatocytes, the spindle bridges
stain sharply in chrome-osmium material which has been fixed over-
night and well washed out. As the cells become larger the spindle
bridges become more difficult to find under the Champy or F. \Y. A.
ha?matoxylin treatment. It is only in special regions that the bridges
are shown as clearly as in Fig. 4, at SB, in full grown spermatocytes.

Now in no case, in chrome-osmium haematoxylin preparations,
could the structure depicted in text figure A be found. The spindle
bridges were "solid" and never appeared to consist of two parts,
inner chromophile region, and an outer canal-like structure. It was
also in vain that in the chrome-osmium haematoxylin slides the
modified heterochromosomes (hemispherical nucleolus) were looked
for. In the case of the Kolatschew preparations nothing corresponding
to the hemispherical nucleolus was found. The writer is certain that
the oidinary nucleoli of Abraxas are not heterochromosomes (Gatenby,
19316), and has found as in Fig. 6, that they lie in all directions in the
nuclei, some towards, others away from the spindle bridge (SB).
Very careful examination failed to reveal any side branch coming
from the spindle bridge towards the ordinary nucleolus. Of course,
it is quite possible that Champy's fluid and iron alum haematoxylin
may have failed to show both the side branch and the hemispherical
nucleolus. It is not thought that Jan Hirschler's figures are not
correct, but it can only be stated that Abraxas must be very different
from the forms studied by him.

Then to refer to text figure B — nothing like this has ever been
found in Abraxas except in the beginning stages of the formation of
abortive acroblasts (Gatenby, Miikerji, and \Yigoder, 1929), which
was a purely experimental phenomenon.

The Origin of Golgi Bodies in Moths. — Hirschler makes the re-
statement that he believes that Golgi bodies appear in
\Iacrotliylacia, under the influence of nuclear substance, on the edge
of the spindle-bridge shown in text figure B at C. Here again the
present writer finds this somewhat difficult to credit. In Fig. 3 of
the plate are three very small spermatogonia from rosettes; these
already contain Golgi bodies. All the evidence which the writer
could find seems to indicate that Golgi bodies are to be seen in the

NOTE ON DR. JAN HIRSCHLER'S PAPER 413

smallest rosette cells in young testes, and do not originate as claimed
by Hirschler. Nor do the ordinary nucleoli in such cells bear any
relationship to the spindle bridges, the confines of which (Fig. 2) are
clearly demonstrable.

The Function of the Spindle Bridges. — It seems certain that the
function of the spindle bridges in such animals as the moths, is to
keep all the cells in one nest at approximately the same stage of
growth and physiological condition. For in the testis crowded with
all stages of spermatogenesis, some cells lying nearer the testicular
wall are better exposed to nutriment than others.

That there must be some sort of connection between the cells of a
nest seems well shown by the "experiment" in Fig. 4. Abraxas is
often infected with large bacteria which live in many of the cells of
the body, and quite often in the testis. Here are two spermatocytes
in first maturation division — both at the same stage, but one cell is
simply crammed with bacteria, so as almost to hide the spindle.
Now it appears unlikely that a cell so burdened could keep up with its
neighbours unless it received active help from them. Yet such
infected cells go successfully through maturation and produce sperma-
tids as in Fig. 5, which are perfectly normal except for the bacteria.
Finally, the latter come to lie on the sperm tails and are probably
instrumental in spreading the infection to non-infected eggs. In
testis cells, such bacteria are not symbiotic.

Radiation and Cell Nests. — Radiation by either x- or gamma-rays,
provides another aspect to this question, and seems to show that the
injury caused by this type of treatment cannot be made up by as-
sistance from neighbouring cells. In Fig. 2, is a rosette of cells,
which had been treated by gamma rays; the cells D, are dead, the
cells L, alive. In the same way the older cells in Fig. 1 had received
one dose of x-rays, and were left for some hours, yet they all differ
in the manner in which they have responded ; the cells Y are normal-
looking, whereas that at X has been badly hit.

In radiated testes of Abraxas left for hours or days, the badly
stricken spermatocytes drop out and degenerate, leaving gaps in the
nest. The survivors go on developing, often quite normally beside
the degenerate remains of their less fortunate neighbours.

DISCUSSION

Hegner (1914) shows in his figures of Leptinotarsa, clavate spindle
bridges, the broad ends of which impinge on the spermatogonial
nuclei, but have no connection with heterochromosomes or nucleoli.
In a previous paper (Gatenby, 1917) clavate spindle bridges in sperma-

414 J. BRONTE GATENBY

togonia are drawn, but these do not quite impinge on the nuclei
(1917, PI. 25, Fig. 46, Euchelia jacobaeae). In the case of three
young spermatocytes of Euchelia, there is certainly no connection
between the spindle bridges and the nucleoli (1917, PI. 25, Fig. 45).
In recent preparations made by the Nassanow method, it has been
found that Golgi bodies exist in the earliest spermatogonial rosettes
and in these the spindle bridges certainly do not touch the nuclei,
while no nucleoli other than the normal ones can be discovered.
The author's previous wrork on Euchelia and Smerinthus is also in
agreement. The spindle bridges were carefully worked out in a
number of species of moths without finding anything like Jan Hirsch-
ler's text figure B, and the only conclusion possible is that these
moths must differ very much from Jan Hirschler's genera.

On the comparison made by Hirschler between heterochromosomes
and the peculiar hemispherical nucleoli, the author is somewhat
doubtful but does not care to express a definite opinion.

Hirschler's text figure A seems too definite an arrangement to be
artifically produced, and the writer is willing to believe that a trial
with Zenker's fluid may show something similar in Abraxas. Against
this, however, is the fact that Hirschler figures the same thing in
text figure B, done by osmic methods, as used by the writer in his
collection.

On the question of the origin of Golgi bodies from nuclear sap,
the author finds himself in complete disagreement with Jan Hirschler.
It will be remembered that Hirschler (19296) has advanced a similar
idea in the case of the pentatomid spermatid worked out by him.
It is not suggested that mitochondria and Golgi bodies could not
originate from nuclear material, but that in the twro cases cited by
Hirschler, Golgi bodies were already present, before they were supposed
by him to arise by some form of condensation.

The response of cells in a nest to radiation, is peculiar and indicates
that some of the cells have been affected much more than others.
N'or do the unaffected units manage to save their stricken neighbours.
\\ hatever be the exact basis of communication via spindle bridges,
between the cells of a nest, it is evident that the radiation injury
cannot be repaired by ordinary nutrition as is presumed to be supplied
through the channels formed by spindle bridges.

SUMMARY

The spindle bridges of Abraxas spermatocytes do not appear to
have any relationship with the de noi'o formation of Golgi bodies.
No hemispherical nucleoli analogous with the heterochromosomes of

NOTE ON DR. JAN HIRSCHLER'S PAPER 415

other insects have yet been found in Abraxas. The Golgi bodies are
present in the smallest spermatogonial rosettes in which the newly-
formed spindle bridges are unrelated to the nuclei; in such cells the
nucleoli have no connection with the spindle bridges. The spindle
bridges, as has been generally concluded, are a means whereby sister
cells in a nest are kept at the same state of nutrition. Spermatocytes
vtry badly infected with bacteria, keep step with non-infected sister
cells, this, it is concluded, by virtue of the spindle bridges. Radiated
testes show nests with dying and apparently undamaged cells side by
side. This type of injury appears to be something different from
mere inadequacy of nutrition, such as the presence of large numbers
of bacteria could bring about.

BIBLIOGRAPHY

BUCHNER, P., 1928. Holznahrung und Symbiose. Julius Springer, Berlin.
GATENBY, J. BRONTE, 1917. The Cytoplasmic Inclusions of the Germ-cells. Lepi-

doptera. Quart. Jour. Micr. Sci., 62: 407.
GATENBY, J. BRONTE, 1931o. The Post-Nuclear Granule in Anasa tristis, with some

Remarks on Hemipteran Spermatogenesis. Am. Jour. Anat. In press.
GATENBY, J. BRONTE, 19316. Induced Multiplication and Growth of Golgi Bodies

and Alteration of Nebenkern Pattern and Nucleolus in Abraxas. In press.
GATENBY, J. B., MUKERJI, R. N. AND WIGODER, S. B., 1929. The Effect of X-

radiation on the Spermatogenesis of Abraxas grossulariata. Proc. Roy. Soc.,

Ser. B., 105: 446.
HIRSCHLER, JAN, 1929a. Sur la relation entre le noyau et les composants plas-

matiques (Appareil de Golgi, vacuome) dans les spermatocytes des Lepi-

dopteres. Compt. rend. Soc. de Biol., T. 101.
HIRSCHLER, JAN, 19296. Sur un Appareil de Golgi primaire et secondaire dans les

spermatides de Palomena. Compt. rend Soc. de Biol., T. 101.
HEGNER, R., 1914. The Germ Cell Cycle in Animals. Macmillan.

INDEX

ACTIVATION of starfish eggs, as

influenced by cyanide and oxygen

lack, 288.
Alternation of generations, in Lecane

inermis Bryce, 345.
Ammonium salts, permeability of

Arbacia egg to, 171.
Amphibian metamorphosis, 11.
Arbacia egg, effect of fatty acid buffer

systems on apparent viscosity, 132.
— , permeability to ammonium

salts, 171.

-, permeability to non-electro-

lytes, 152.

-, unfertilized, oxygen tension-

oxygen consumption curve, 242.

Artificial parthenogenesis, production of

normal embryos by, in Urechis, 187.

J^ASAL plate of tail, role of, in re-
generation in tail-fins of fishes, 60.

BUCHANAN, J. WILLIAM. Modification
of the rate of oxygen consumption
by changes in oxygen concentration
in solutions of different osmotic
pressure, 309.

(^ARASS I US, role of basal plate of tail

in regeneration in tail-fins, 60.
Cell permeability to ions, illustrated in

Arbacia egg, 132.

— , to water, as affected by

urethanes, 72.

Cells, oxygen diffusion into, 245.
Chaetopterus, determination of surface

tension of eggs, 67.
Cobra venom, resistance of fourteen

species of Protozoa to, 64.
Cortical change, conduction of, at fertil-

ization in starfish egg, 23.
kCrotalus atrox venom, resistance of

fourteen species of Protozoa to, 64.
Culture medium, modifications of, effect

on longevity and fecundity of

I 'males sordida, 30.
Cyanide, influence on activation of star-

fish eggs, J

[)''• IAL plicae formation in Rana

pipiens, 1 1.
Desiccation, changes during, in body and

organs of leopard frog, 80.

Domestic fowl, sex ratio, 124.
DUNIHUE, F. W. See Stunkard and
Dunihue, 179.

"pMBRYOS, normal, production in
Urechis by artificial parthenogenesis,
187.

Erythrocyte, osmotic properties, 95.

Euglena leucops, nuclear division, 327.

UALK, I. S. See Gerard and Falk,
213.

Fatigue in muscle and its bearing on inn
antagonism, 382.

Fatty acid buffer systems, effect on
apparent viscosity of Arbacia egg,
132.

Fecundity of Proales sordida, as affected
by modifications of culture medium,
30.

Fertilization in starfish egg, conduction
of cortical change, 23.

Fiona marina, pigment of, 120.

Fowl, sex ratio in relation to antecedent
egg production and inbreeding, 124.

Fundulus, role of basal plate of tail in
regeneration in tail-fins, 60.

Fundulus heteroclitus, effect of temper-
ature changes on pulsations of iso-
lated scale melanophores, 269.

QATENBY, BRONTE. Note on Dr.
Jan Hirschler's paper on spindle
bridges, and the de novo origin of
moth spermatocyte Golgi bodies,
409.

GELLHORN, E. Permeability and fatigue
in muscle and its bearing on the
problem of ion antagonism, 382.

— . The recovery contracture in
muscle; a new general salt effect,
397.

GERARD, R. W. Observations on the
metabolism of Sarcina lutca, II, 227.
— . Oxygen diffusion into cells,
245.

GERARD, R. W. and I. S. FALK. Obser-
vations on the metabolism of Sarcina
lutea, I, 213.

Golgi bodies, of moth spermatocytcs, de
novo origin, 40<>.

416

INDEX

417

T_I ALL, S. R. Observations on Euglena
leucops, sp. nov., a parasite of
Stenostomum, with special reference
to nuclear division, 327.

HARVEY, E. NEWTON. A determination
of the tension at the surface of eggs
of the annelid, Chaetopterus, 67.

HELFF, O. M. Studies on amphibian
metamorphosis, IX. Integumentary
specificity and dermal plicae for-
mation in the anuran, Rana pipiens,
11.

Hemolysis, as influenced by pH, temper-
ature and oxygen tension, 95.

HOWARD, EVELYN. The effect of fatty
acid buffer systems on the apparent
viscosity of the Arbacia egg, with
especial reference to the question of
cell permeability to ions, 132.

Hydrogen ion concentration, influence on
hemolysis, 95.

INBREEDING in domestic fowl, in
relation to sex ratio, 124.

Integumentary specificity in Rana pipi-
ens, 11.

Ion antagonism, and permeability and
fatigue in muscle, 382.

TACKSON, C. M. See Smith and

J Jackson, 80.

JACOBS, M. H. and ARTHUR K. PARPART.
Osmotic properties of the erythro-
cyte. II. The influence of pH,
temperature, and oxygen tension on
hemolysis by hypotonic solutions,
95.

JULL, MORLEY A. The sex ratio in the
domestic fowl in relation to ante-
cedent egg production and in-
breeding, 124.

J^ROPP, BENJAMIN. The pigment
of Velella spirans and Fiona marina,
120.

T ECANE inermis Bryce, life histories
of sexual and non-sexual generations,
345.

LILLIE, RALPH S. Influence of cyanide
and lack of oxygen on the activation
of starfish eggs by acid, heat and
hypertonic sea-water, 288.

Longevity, in Proales sordida, as in-
fluenced by modifications of culture
1 medium, 30.

LUCRE, BALDUIN. The effect of certain
narcotics (urethanes) on perme-
ability of living cells to water, 72.

LYNCH, RUTH S. and HELEN B. SMITH.
A study of the effects of modifi-
cations of the culture medium upon
length of life and fecundity in a
rotifer, Proales sordida, with special
relation to their heritability, 30.

X/jETABOLISM of Sarcina lutea,
1 213,227.

Metamorphosis, amphibian, 11.

MILLER, HELEN MAR. Alternation of
generations in the rotifer Lecane
inermis Bryce. I. Life histories of
the sexual and non-sexual gener-
ations, 345.

Moth spermatocyte Golgi bodies, de novo
origin, 409.

Muscle, permeability and fatigue in, and
problem of ion antagonism, 382.

, recovery contracture in, new

general salt effect, 397.

J^ABRIT, S. MILTON. The role of
the basal plate of the tail in re-
generation in the tail-fins of fishes
(Fundulus and Carassius), 60.

Non-electrolytes, permeability of Arbacia
egg to, 152.

Nuclear division of Euglena leucops, sp.
nov., 327.

("")SMOTIC properties of the erythro-
cyte, 95.

Oxygen consumption, modification of
rate by changes in oxygen concen-
tration in solutions of different
osmotic pressure, 309.

- diffusion into cells, 245.

lack, influence on activation of
starfish eggs, 288.

- tension, influence on hemolysis, 95.

— , oxygen consumption curve of
unfertilized Arbacia eggs, 242.

pARPART, ARTHUR K. See Jacobs

and Parpart, 95.
Parthenogenesis, artificial, production of

normal embryos by, in Urechis, 187.
Permeability and fatigue in muscle, 382.

- of Arbacia egg to ammonium salts,
171.

- of Arbacia egg to non-electrolytes,
152.

418

IXDKX

- of living cells to water, effect of
urethanes, 72.

PHILPOTT, C. II. Relative resistance of
fourteen species of Protozoa to the
action of Crotalus atrox and Cobra
venoms, 64.

Pigment of Yelella spirans and Fiona
marina, 120.

Proales sordida, effects of modifications
of culture medium on length of life
and fecundity, 30.

Protein synthesis, 1.

Protozoa, resistance to action of Crotalus
atrox and Cobra venoms, 64.

Pulsations of isolated scale melanophores
of Fundulus heteroclitus, as influ-
enced by temperature changes,
269.

13 ANA pipiens, changes during desicca-
tion and rehydration in body and
organs, 80.

, integumentary specificity and

dermal plicae formation, 1 1 .
Recovery contracture in muscle, 397.
Regeneration of tail-fins of fishes, role of

basal plate of tail, 60.
Rehydration, changes during, in body

and organs of leopard frog, 80.
Resistance of Protozoa to action of

Crotalus atrox and Cobra venoms,

64.

OARCINA lutea, metabolism,
227.

Sex ratio in domestic fowl in relation to
antecedent egg production and
inbreeding, 124.

SMITH, DIETRICH C. The effect of
temperature changes upon the pul-
sations of isolated scale melano-
phores of Fundulus heteroclitus,
269.

SMITH, H. B. See Lynch and Smith, 30.

SMITH, YERNON U. E. and C. M.
JACKSON. The changes during des-
iccation and rehydration in the body-
am 1 organs of the leopard frog (Rana
pipiens), 80.

Spindle bridges, note on Dr. Jan
Ilirschler's paper, 409.

Starfish egg, conduction of cortical
change at fertilization, 23.

— , influence of cyanide and
oxygen-lack on activation, 288.

Stenostomum, parasite of, 327.

STEWART, DOROTHY R. The perme-
ability of the Arbacia egg to am-
monium salts, 171.

— , - — . The permeability of

the Arbacia egg to non-electrolytes,
152.

STUNKARD, H. \V. and F. W. DUMHUE.
Notes on trematodes from a Long
Island cluck with description of a
new species, 179.

Surface tension of Chaetopterus eggs, 67.

'AIL-FINS of fishes, role of basal
plate of tail in regeneration, 60.

TANG, PEl-SuNG. The oxygen tension-
oxygen consumption curve of un-
fertilized Arbacia eggs, 242.

Temperature changes, effect on pul-
sations of isolated scale melano-
phores of Fundulus heteroclitus, 269.
— , influence on hemolysis, 95.

Tension at surface of Chaetopterus eggs,
67.

Trematodes from Long Island duck,
notes on, 179.

TYLER, ALBERT. The production of
normal embryos by artificial parthe-
nogenesis in the echiuroid, Urechis,
187.

TJRECHIS, production of normal
embryos by artificial parthenogene-
sis, 187.

Urethanes, effect on permeability of
living cells to water, 72.

WELELLA spirans, pigment of,

120.
Yiscosity of Arbacia egg, as affected by

fatty acid buffer systems, 132.

\yASTENEYS, HARDOLl'H. The
Second Reynold A. Spaeth Me-
morial Lecture. Protein synthesis,
1.

WHITAKKR, D. M. On the conduction
<>t the cortical change at fertilization
in the starfish egg, 23.

Volume LX Number 1

THE

BIOLOGICAL BULLETIN

PUBLISHED BY

THE MARINE BIOLOGICAL LABORATORY

Editorial Board

GARY N. CALKINS, Columbia University FRANK R. LlLLIE, University of Chicago

E. G. CONKLIN, Princeton University CARL R. MOORE, University of Chicago

E. N. HARVEY, Princeton University GEORGE T. MOORE, Missouri Botanical Garden

SELIG HECHT, Columbia University T. H. MORGAN, California Institute of Technology

M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University

H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University

E. E. JUST, Howard University EDMUND B. WILSON, Columbia University

ALFRED C. REDFIELD, Harvard University
Managing Editor

FEBRUARY, 1931

Printed and Issued by

LANCASTER PRESS, Inc.

PRINCE & LEMON STS.

LANCASTER, PA.

THE BIOLOGICAL BULLETIN is issued six times a year. Single
numbers, Si. 75. Subscription per volume (3 numbers), $4.50.

Subscriptions and other matter should be addressed to the
Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa.
Agent for Great Britain: \Yheldon & Wesley, Limited, 2, 3 and
4 Arthur Street, New Oxford Street, London, W.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, 240 Longwood Avenue, Boston, Mass.

-•red October iu, 1902, at Lancaster, Pa., as second-class matter under
Au i.| ( 'on:L;rr-,s of July 16, 1894.

CONTENTS

Page
WASTENEYS, HARDOLPH

The Second Reynold A. Spaeth Memorial Lecture. Protein
Synthesis 1

HELFF, O. M.

Studies on Amphibian Metamorphosis. IX. Integumentary
Specificity and Dermal Plicae Formation in the Anuran, Rana
pipiens 11

WHITAKER, D. M.

On the Conduction of the Cortical Change at Fertilization in
the Starfish Egg 23

LYNCH, RUTH STOCKING, and HELEN BERENICE SMITH

A Study of the Effects of Modifications of the Culture Medium
upon Length of Life and Fecundity in a Rotifer, Proales sor-
dida, with special relation to their Heritability 30

NABRIT, S. MILTON

The Role of the Basal Plate of the Tail in Regeneration in the
Tail-Fins of Fishes (Fundulus and Carassius) 60

PHILPOTT, CHARLES H.

Relative Resistance of Fourteen Species of Protozoa to the
Action of Crotalus atrox and Cobra Venoms 64

HARVEY, E. NEWTON

A Determination of the Tension at the Surface of Eggs of the
Annelid, Chaetopterus 67

LUCKE, BALDUIN

The Effect of Certain Narcotics (Urethanes) on Permeability
of Living Cells to Water 72

SMITH, 'VERNON D. E., and C. M. JACKSON

The Changes during Desiccation and Rehydration in the Body
and Organs of the Leopard Frog (Rana pipiens) 80

Volume LX

Number 2

THE

BIOLOGICAL BULLETIN

PUBLISHED BY

THE MARINE BIOLOGICAL LABORATORY

-

Editorial Board

GARY N. CALKINS, Columbia University

E. G. CONKLIN, Princeton University

E. N. HARVEY, Princeton University

SELIG HECHT, Columbia University

M. H. JACOBS, University of Pennsylvania

H. S. JENNINGS, Johns Hopkins University

E. E. JUST, Howard University

FRANK R. LILLEE, University of Chicago
CARL R. MOORE, University of Chicago
GEORGE T. MOORE, Missouri Botanical Garden
T. H. MORGAN, California Institute of Technology
G. H. PARKER, Harvard University
W. M. WHEELER, Harvard University
EDMUND B. WILSON, Columbia University

ALFRED C. REDFIELD, Harvard University
Managing Editor

APRIL, 1931

Printed and Issued by

LANCASTER PRESS, inc.

PRINCE & LEMON STS.
LANCASTER, PA.

THE BIOLOGICAL BULLETIN is issued six times a year. Single
numbers, $1.75. Subscription per volume (3 numbers), $4.50.

Subscriptions and other matter should be addressed to the
Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa.
Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and
4 Arthur Street, New Oxford Street, London, \V.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, 240 Long wood Avenue, Boston, Mass.

Entered October 10, 1902, at Lancaster, Pa., as second-class matter under
Act of Congress of July 16, 1894.

CONTENTS

Page
JACOBS, M. H., and ARTHUR K. PARPART

Osmotic Properties of the Erythrccyte. II. The Influence of
pH, Temperature, and Oxygen Tension on Hemolysis by

Hypotonic Solutions 95

«

KROPP, BENJAMIN

The Pigment of Vellela spirans and Fiona marina 120

JULL, MORLEY A.

The Sex Ratio in the Domestic Fowl in Relation to Antece-
dent Egg Production and Inbreeding 124

HOWARD, EVELYN

The Effect of Fatty Acid Buffer Systems on the Apparent
Viscosity of the Arbacia Egg, with Especial Reference to the
Question of Cell Permeability to Ions 132

STEWART, DOROTHY R.

The Permeability of the Arbacia Egg to Non-Electrolytes. . 152

STEWART, DOROTHY R.

The Permeability of the Arbacia Egg to Ammonium Salts. . 171

STUNKARD, H. W., and F. W. DUNIHUE

Notes on Trematodes from a Long Island Duck with De-
scription of a New Species 179

TYLKR, ALBERT

The Production of Normal Embryos by Artificial Partheno-
genesis in the Echiuroid, Urechis . 187

Volume LX Number 3

THE

BIOLOGICAL BULLETIN

PUBLISHED BY

THE MARINE BIOLOGICAL LABORATORY

Editorial Board

GARY N. CALKINS, Columbia University FRANK R. LlLLIE, University of Chicago

E. G. CONKLIN, Princeton University CARL R. MOORE, University of Chicago

E. N. HARVEY, Princeton University GEORGE T. MOORE, Missouri Botanical Garden

SELIG HECHT, Columbia University T. H. MORGAN, California Institute of Technology

M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University

H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University

E. E. JUST, Howard University EDMUND B. WILSON, Columbia University

ALFRED C. REDFIELD, Harvard University
Managing Editor

JUNE, 1931

Printed and Issued by

LANCASTER PRESS, Inc.

PRINCE & LEMON STS.

LANCASTER, PA.

THE BIOLOGICAL BULLETIN is issued six times a year. Single
numbers, $1.75. Subscription per volume (3 numbers), $4.50.

Subscriptions and other matter should be addressed to the
Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa.
Agent for Great Britain: YVheldon & Wesley, Limited, 2, 3 and
4 Arthur Street, New Oxford Street, London, W.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, 240 Longwood Avenue, Boston, Mass.

Fntcrcd October 10, 1902, at Lancaster, Pa., as second-class matter under
Act of Congress of July 16, 1894.

CONTENTS

Page
GERARD, R. W., and I. S. FALK

Observations on the Metabolism of Sarcina lutea. 1 213

GERARD, R. W.

Observations on the Metabolism of Sarcina lutea. II 227

TANG, PEI-SUNG

The Oxygen Tension-Oxygen Consumption Curve of Unfer-
tilized Arbacia Eggs 242

GERARD, R. W.

Oxygen Diffusion into Cells 245

SMITH, DIETRICH C.

The Effect of Temperature Changes upon the Pulsations of
Isolated Scale Melanophores of Fundulus heteroclitus 269

LILLIE, RALPH S.

Influence of Cyanide and Lack of Oxygen on the Activation

of Starfish Eggs by Acid, Heat and Hypertonic Sea- Water. . 288

BUCHANAN, J. WILLIAM

Modification of the Rate of Oxygen Consumption by Changes
in Oxygen Concentration in Solutions of Different Osmotic
Pressure 309

HALL, S. R.

Observations on Euglena leucops, fir nov., a Parasite of
Stenostomum, with special reference xo Nuclear Division. . 327

MILLER, HELEN MAR

Alternation of Generations in the Rotifer Lecane inermis
Bryce. I. Life histories of the sexual and non-sexual
generations 345

GELLHORN, E.

Permeability and Fatigue in Muscle and its Bearing on the
Problem of Ion Antagonism 382

GELLHORN, E.

The Recovery Contracture in Muscle; a New General Salt
Effect 397

GATENBY, J. BRONTE

Note on Dr. Jan Hirschler's Paper on Spindle Bridges,
and the De Novo Origin of Moth Spermatocyte Golgi Bodies 409

INDEX 416

MBI. WHO1 I.1KKAKY

UH 17IC 1