Full text of "The Biological bulletin"

See other formats

<i-
<*

THE

BIOLOGICAL BULLETIN

PUBLISHED BY

THE MARINE BIOLOGICAL LABORATORY

Editorial Board

GARY N. CALKINS, Columbia University E. E. JUST, Howard University

E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago

E. N. HARVEY, Princeton University CARL R MOORE, University of Chicago

SELIG HECHT, Columbia University „„ , „, ,,

LEIGH HOADLEY, Harvard University °EORGE T' M°ORE' Mlssoun Botamcal Garden

L. IRVING, Swarthmore College T- H- MORGAN, California Institute of Technology

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

H. S. JENNINGS, Johns Hopkins University F. SCHRADER, Columbia University

ALFRED C. REDFIELD, Harvard University
Managing Editor

VOLUME LXXX

FEBRUARY TO JUNE, 1941

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, W.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, Marine Biological Laboratory, Woods Hole,
Mass., between June 1 and October 1 and to the Biological Labo-
ratories, Divinity Avenue, Cambridge, Mass., during the remainder
of the year.

Entered 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. FEBRUARY, 1941

PAGE

IRVING, L., E. C. BLACK AND V. SAFFORD

The Influence of Temperature upon the Combination of Oxygen
with the Blood of Trout 1

MACGINITIE, G. E.

On the Method of Feeding of Four Pelecypods 18

HITCHCOCK, H. B.

The Coloration and Color Changes of the Gulf-weed Crab
Planes minutus 26

BURGER, J. W.

Some Experiments on the Effects of Hypophysectomy and
Pituitary Implantations on the Male Fundulus heteroclitus ... 31

HSIAO, S. C. T.

Melanosis in the Common Cod, Gadus callarias L., Associated
with Trematode Infection 37

GILMOUR, D.

Repayment of the Anaerobic Oxygen Debt in Grasshopper
Skeletal Muscle 45

KIDDER, G. W.

Growth Studies on Ciliates. VII. Comparative Growth Char-
acteristics of Four Species of Sterile Ciliates 50

HUNTER, F. R., S. B. BARBER AND A. P. CAPUTI

The Effect of Saponin on the Osmotic Hemolysis of Chicken
Erythrocytes 69

CULBRETH, S. E.

The Role of Tissues in the Anaerobic Metabolism of the Mus-
sel Anodonta hallenbeckii Lea 79

REDFIELD, A. C.

The Effect of the Circulation of Water on the Distribution of
the Calanoid Community in the Gulf of Maine 86

Fox, D. L.

Changes in the Tissue Chloride of the California Mussel in

Response to Heterosmotic Environments Ill

iii

53704

iv CONTENTS

PAGE

TARTAR. V., AND T. T. CHEN

Mating Reactions of Enucleate Fragments in Paramecium
bursaria 130

No. 2. APRIL, 1941

FERGUSON, J. K. W., AND E. C. BLACK

The Transport of CCX in the Blood of Certain Freshwater

Fishes 139

TUNG, T, S. Ku AND Y. TUNG

The Development of the Ascidian Egg Centrifuged Before

Fertilization 153

BERRILL, N. J.

The Development of the Bud in Botryllus 169

BERRILL, N. J.

Size and Morphogenesis in the Bud of Botryllus 185

JENNINGS, R. H., AND D. M. WHITAKER

The Effect of Salinity upon the Rate of Excystment of Artemia 194

CORNMAN, I.

Sperm Activation by Arbacia Egg Extracts, with Special Rela-
tion to Echinochrome 202

SCHNEIDER, B. A.

The Nutritional Requirements of Tribolium confusum Duval, I. 208

RAHN, H.

The Pituitary Regulation of Melanophores in the Rattlesnake 228

PORTER, K. R.

Diploid and Androgenetic Haploid Hybridization between Two
Forms of Rana pipiens, Schreber 238

FOWLER, C.

The Relation Between Hydrogen-Ion Concentration and Vol-
ume, Gel/sol Ratio and Action of the Contractile Vacuole in
Amoeba proteus 265

No. 3. JUNE, 1941

DAY, M. F.

Pigment Migration in the Eyes of the Moth, Ephestia kuehni-
ella Zeller 275

KLEITMAN, N.

The Effect of Temperature on the Righting of Echinoderms . . 292

CONTENTS v

PAGE
BOTSFORD, E. F.

The Effect of Physostigmine on the Responses of Earthworm

Body Wall Preparations to Successive Stimuli 299

ALBAUM, H. G., AND B. COMMONER

The Relation between the Four-Carbon Acids and the Growth

of Oat Seedlings 314

KlTCHING, J. A.

Studies in Sublittoral Ecology. III. Laminaria forest on the
west coast of Scotland ; a study of zonation in relation to wave
action and illumination 324

MORGAN, T. H.

Further Experiments in Cross- and Self-Fertilization of Ciona

at Woods Hole and Corona del Mar 338

HARVEY, E. B.

Relation of the Size of " Halves " of the Arbacia punctulata
Egg to Centrifugal Force 354

EVANS, T. C, H. W. BEAMS AND M. E. SMITH

Effects of Roentgen Radiation on the Jelly of the Arbacia Egg 363

TURNER, C. L.

Gonopodial Characteristics Produced in the Anal Fins of Fe-
males of Gambusia affinis affinis by Treatment with Ethinyl
Testosterone 371

GILMAN, L. C.

Mating Types in Diverse Races of Paramecium caudatum .... 384

DETHIER, V. G.

The Function of the Antennal Receptors in Lepidopterous
Larvae 403

HUNNINEN, A. V., AND R. M. CABLE

Studies on the Life History of Anisoporus manteri Hunninen

and Cable, 1940 (Trematoda: Allocreadiidae) 415

BENDITT, E., P. MORRISON AND L. IRVING

The Blood of the Atlantic Salmon during Migration 429

FOX, D. L., AND B. T. SCHEER

Comparative Studies of the Pigments of Some Pacific Coast
Echinoderms 441

Vol. LXXX, No. 1 February, 1941

THE

BIOLOGICAL BULLETIN

PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY

THE INFLUENCE OF TEMPERATURE UPON THE

COMBINATION OF OXYGEN WITH THE

BLOOD OF TROUT l- -

LAURENCE IRVING, EDGAR C. BLACK AND VIRGINIA SAFFORD

(From the Edward Martin Biological Laboratory, Sivarthmore College,

Swarthmore, Pennsylvania )

The air breathed by all species of mammals is the same in composi-
tion, and the affinity of their blood for oxygen is also much alike. But
the water in which fish live differs greatly in oxygen saturation in dif-
ferent places and seasons, and this variability is particularly conspicuous
in bodies of fresh water in temperate regions. Under these circum-
stances it is not surprising to find that the blood of various species of
fish varies greatly in affinity for oxygen. Krogh and Leitch (1919) first
remarked upon the differences among the eel, carp, plaice, cod, and trout,
and regarded these differences in affinity for oxygen as evidence for the
adaptation of the blood to the conditions in which each species lived.
But only a few species were examined and these were quite dissimilar
in form and habit as well as in respect to the habitat which they occupied.

The catfish, carp, bowfin, and sucker were found by Black (1940)
to have blood with different oxygen dissociation curves, which were
nevertheless related as if in one family. These freshwater fish are simi-
lar in form and are all of free swimming habit. Since their respiratory
requirements and their physical systems for providing oxygen are much
alike, it is reasonable to consider that large differences in the properties
of the blood of these species indicate the suitability of the blood for
respiration in different environments.

We have now examined the blood of eight species of freshwater fish
and find that the affinity for oxygen varies greatly among the species.
Furthermore, as Black (1940) pointed out for four of the species, the

1 We wish to acknowledge the kind assistance of Mr. C. O. Hayford, Super-
intendent of the New Jersey State Fish Hatchery at Hackettstown, in enabling us
conveniently to secure the blood from the trout used in these experiments.

2 The expenses of this investigation were in part provided by a grant from the
Rockefeller Foundation.

L. IRVING, E. C. BLACK AND V. SAFFORD

effect of CO, upon oxygen combination is also quite different in the
species. Table I shows the usual oxygen affinity and the effect of CO2
in the blood of these eight species. These characteristics would fit respi-
ration under quite different conditions.

These properties of the blood do not depict its natural suitability for
the transport of oxygen because there is no allowance for the variety

TABLE I

Oxygen affinity and effect of COz in blood of freshwater fish at 15°

(Jordan, 1929)

Species

P COz at J saturation

Limit of CO2 effect

P CO2 = l-2 mm.

P CO2 = 10 mm.

Percentage HbO2

Common catfish

1.4

(Black, 1940)

Ameiurus nebulosus

Bowfin

(Black, 1940)

Amia calva

Carp

85 (Black, 1940)

Carpiod.es cyprinits

Yellow perch

77 (Irving, unpub-

Perca flavescens

lished)

Common sucker

71 (Black, 1940)

Catostomus commersonnii

Chain pickerel

53 (Irving, unpub-

Esox niger

lished)

Rainbow trout

52 (Irving, Black,

Trutta iridea

and Safford—

this paper)

Brown trout

52 (Irving, Black,

Trutta trulta

and Safford —

this paper)

Brook trout

52 (Irving, Black,

Salvelinus fonlinalis

and Safford—

this paper)

Atlantic salmon (freshwater)

57 (Irving, 1939)

Salmo solar

Lake trout

40 (Irving, unpub-

Cristivomer namaycush

lished)

of temperatures in the natural habitats of the fish. In lakes in tem-
perate regions the temperature of a stratum of water may differ sharply
from the temperature above and below, and the seasonal changes are
rapid and large. The influence of temperature upon the oxygenation of
mammalian blood is such that at 20° it would be 95 per cent saturated
with oxygen by a pressure of 45 mm., while at 37° the pressure required
is about 100 mm. (Brown and Hill, 1923). According to the figures

OXYGENATION IN TROUT BLOOD

given for the blood of the skate (Dill et al, 1932), the hemoglobin would
be 95 per cent saturated by 40 mm. pressure of oxygen at 15°, but the
same saturation would require 170 mm. at 25°. At the higher tem-
perature part of the efficacy of the hemoglobin for oxygen transport
would be lost even in water which was saturated with air. The blood
of the eel would be 95 per cent saturated with oxygen by pressures of
12, 25, and 71 mm. at 5°, 17°, and 30° respectively (Kawamoto, 1929).
From general considerations concerning the nature of hemoglobin, as
well as from these two examples, it is to be expected that the function
of oxygen transport in fish blood is considerably influenced by the tem-
perature at which it occurs.

The combination of oxygen with the blood of fish is peculiarly sensi-
tive to carbon dioxide, as Krogh and Leitch (1919) first observed.
Even at 700 mm. pressure of oxygen the hemoglobin of the blood of
the tautog will not become saturated with oxygen in the presence of car-
bon dioxide (Root and Irving, 1940). In this respect the CO2 effect
in fish blood differs from the Bohr effect produced by CO2 in mam-
malian blood. In the practical absence of CO., the hemoglobin of fish
blood is saturated with oxygen at 150 mm. pressure. With increasing
pressures of CO2, oxygen saturation falls off until no further reduction
of oxygen saturation is secured beyond 60 mm. pressure of CO2 (Root,
Irving, and Black, 1939). At the limit of the CO2 effect on the blood
of trout only about 50 per cent of the hemoglobin can be oxygenated.

The limit of the CO2 effect is quite different in various species of
fish. A number of freshwater fish which we have examined clearly
show the variation in the limit of the CO2 effect in the blood of several
species (Table I, column 3). The differences in CO2 sensitivity of the
blood and the influence which the CO2 effect would exert upon conditions
essential for respiratory transport, however, have only been determined
at 15°.

The limit of the CO2 effect in trout blood changes with temperature,
as is shown by the observations recorded in Fig. 1. These determina-
tions were made by measuring the percentage of saturation of the hemo-
globin with oxygen at a pressure of 150 mm. and CO2 at 65 mm. or
more. The blood samples were secured from a number of brook, brown,
and rainbow trout. The limit of the CO2 effect falls at a lower level of
oxygen saturation with increasing temperature up to 25°. Above 25°
the blood cannot be saturated by oxygen at 150 mm. pressure even in
the absence of CO2.

The influence which temperature exerts upon the CXX effect shows
that the effect of CO2 upon affinity for O2 must be considered in making

L. IRVING, E. C. BLACK AND V. SAFFORD

a description of the blood of fish which will he useful in designating
its suitability for the performance of its natural function of oxygen
transport.

MATERIAL USED

For examining the effect of temperature upon oxygen affinity and
the CO., effect we have selected three common and related species of
freshwater fish; brook trout, Sah'clinus fontinalis; brown trout, Trutta

60-

(£>

o
in

50-

40-

0 Brown Trout
A Rainbow »
Q Brook »

Temperature
FIG. 1. Limit of the effect of CO2 upon oxygenation of hemoglobin in trout blood.

trutta; and rainbow trout, Trutta iridca. These fish live under similar
conditions and are alike in activity. They were raised in the same water
at the New Jersey State Hatchery at Hackettstown. All had received
the same sort of food, and each species had been raised from a homo-
geneous and selected stock. If they were as different in form and
habits as the toadfish and the tautog, all of the physical components of
the respiratory mechanism would obviously be different, and a com-
parison of the species on the basis of the properties of blood alone would

OXYGENATION IN TROUT BLOOD

not be justified. But since these three species of trout are so much alike,
against the background of general similarity of essential respiratory
devices either similarity or difference in the properties of the blood for
respiratory transport may acquire significance for the eventual picture
of specific respiratory adaptation.

METHODS

Male fish of from 1^ to 2l/2 pounds weight were bled by syringe
from the heart, yielding from 5 to 20 ml. per fish. Coagulation was
prevented by liquid heparin (Connaught Laboratory), the blood was
iced, and analysis was begun about five hours after drawing and com-
pleted within the next fourteen hours. About five of the eighty fish
used died but the others showed no ill effects from bleeding.

The blood was equilibrated for more than 20 minutes with suitable
gas mixtures. Half-milliliter samples were withdrawn and analyzed ac-
cording to the method of Van Slyke and Neill (1924). Extraction
of the blood in the apparatus was complete after six minutes. Samples
of the equilibrated gases were analyzed in the Haldane apparatus. He-
matocrit measurements were made with a high speed (6000 r.p.m.)
centrifuge.

Each sample of blood was analyzed after equilibration at from two to
five temperatures. If the procedure was prolonged, final results were
checked by repetition of an early equilibration. It was possible to keep
the blood without change at 30° for at least an hour, but after a short
time at 35° the oxygen capacity was not restored by equilibration at 15°.

Mixtures of the blood of several fish gave results comparable with
the blood of single fish. Blood samples from 23 brook trout, 23 brown
trout, and 34 rainbow trout were examined on five separate days for
each species. Each species was examined in March, October, Novem-
ber, and December. No fish were used soon after stripping the sperm
for breeding, and there was no apparent seasonal change in the blood.

The temperature of the natural spring water at Hackettstown was
uniform at about 12°, but in the pools it fell as low as about 6° without
noticeably altering the blood. With a good flow of water in the hatch-
ery and stock selected for a number of years conditions are especially
favorable for a degree of uniformity that has not been encountered in
our experience with several other species of wild fish.

OXYGEN CAPACITY

When equilibrated with air at temperatures below 25° and with
pressure of COL, less than 2 mm., the hemoglobin is saturated with oxy-

L. IRVING, E. C. BLACK AND V. SAFFORD
TABLE II

Oxygen capacity

Species

Number of fish

Hb02
ml. Oa per 100 ml.
blood

Cell volume
ml. O2 per 100 ml.
blood

Brook trout
Salvelinus fontinalis
Brown trout
Trutta trutta

23
23

11.7 (11.0-13.9)
12.2 (11.1-14.4)

Rainbow trout
Trutta iridea

13.8 (13.0-15.0)

gen. The oxygen combined with hemoglobin was determined by sub-
tracting from the total oxygen in the blood the dissolved oxygen, which
was estimated from the solubility coefficients for oxygen dissolved in

100

FIG. 2. Oxygen dissociation curves of trout blood at various temperatures.

mammalian blood (Peters and Van Slyke, 1935). The amount of
oxygen combined with the saturated hemoglobin is designated as oxygen
capacity. Its value is shown for the three species in Table II. Since
the samples were frequently mixed from several fish, the variations in
oxygen capacity are not the limits of those which may occur among in-
dividuals. The oxygen capacities were slightly larger than in the cat-

OXYGENATION IN TROUT BLOOD 7

fish, bowfin, carp, and sucker (Black, 1940), and distinctly larger than
in many of the marine fish examined by Root (1931).

OXYGEN DISSOCIATION CURVES

In our experience the oxygen affinity of fish hemoglobin is not per-
ceptibly diminished by CO., unless the tension exceeds 2 mm. The

90-

Dog 40
Man B8iH
40mm CO.

ODoq 10mm

Temperature

FIG. 3. The effect of temperature upon the pressure of oxygen needed to
half-saturate blood of brown trout at CO. tensions 1 and 10 mm. Human blood
(Brown and Hill, 1923) and some points for dogs' blood (Bohr, Hasselbalch, and
Krogh, 1904) are shown for comparison.

oxygen dissociation curves shown in Fig. 2 were made with blood having
tensions of CO2 less than 2 mm., drawing the best lines through points
from several samples of mixed blood of the brook trout. One sample

L. IRVING, E. C. BLACK AND V. SAFFORD

of mixed blood from two brown trout was examined over the range of
oxygen pressures at 30°. Considering that there are individual varia-
tions and that the construction of a number of complete curves for each
species at each temperature is not practical, it may be said that these are
representative curves, and that one family of similar curves depicts the
oxygen affinity of the blood of all three species at various temperatures.
In the gills of trout, loading naturally occurs with tensions of CO.,

O 50
.0

O 40

(T
O

O
Q.

BROOK TROUT

TEMPERATURE

FIG. 4. The effect of temperature upon the pressure of oxygen necessary to half-
saturate blood of brook trout at CO2 tensions 1 and 10 mm.

in the water which are less than 2 mm. (Ferguson and Black, 1940).
In the arterial blood of rainbow trout the tension of CO2 is about as
low, and so the curves represent the condition of the blood when loading
with oxygen occurs in the gills. At 30° the blood could not be satu-
rated with the oxygen pressure of the air, but at 25° the hemoglobin
could reach its full saturation.

At temperatures above 15° it is easy to see that the curves are
^-shaped. At lower temperatures the deviation is too small to be dem-

OXYGENATION IN TROUT BLOOD

onstrated with certainty. At any temperature the curve is sufficiently
close to a straight line so that a single determination of the oxygen
content and tension of blood between 40 per cent and 60 per cent satura-
tion establishes within 2 mm. the point where the curve cuts half satura-
tion. In this manner it was possible by determining the PO2 for half
saturation of the hemoglobin in one lot of blood at a number of tempera-

50-

CM
O

X 40'

CVJ

30-

20-

10-

Rainbow Trout

i
20

i
30

Temperature

FIG. 5. The effect of temperature upon the pressure of oxygen necessary to half-
saturate blood of rainbow trout at CO., tensions 1 and 10 nun.

tures, to estimate the effect of temperature upon oxygen affinity before
the blood deteriorated.

The effect of temperature upon half saturation of hemoglobin in
brown trout blood is showyn in Fig. 3, the upper curve at 10 mm. CCX
tension and the next lower at 1 mm. CCX. For comparison with human
blood the lower curve is drawn from the data of Brown and Hill (1923),
and points for dog's blood at 37° (Bohr, Hasselbalch, and Krogh, 1904)
show how CCX affects its affinity for oxygen. Data for the effect of

L. IRVING, E. C. BLACK AND V. SAFFORD

temperature upon blood of brook trout and rainbow trout are shown in
Figs. 4 and 5.

When the curves for the three species are superimposed in Fig. 6, it
is apparent that without CO2 there is no difference between the blood
of brook and brown trout, but the hemoglobin of rainbow trout requires
somewhat greater pressure of oxygen to secure half saturation, particu-
larly at lower temperatures. In the absence of CO2, the bloods of the

60 -

50-

O 4O

e
o

30 -

20-

10-

Rainbow Trout
Brown Trout
— — — — Brook Trout

— r
15

— T~
25

20
Te mperature

— i —
35

FIG. 6. Comparison of effects of temperature upon oxygen affinity in blood of

brown, brook, and rainbow trout.

three trout are quite similar, but with 10 mm. CCX tension significant
differences are apparent.

In the blood of several fish (Root, 1931; Willmer, 1934; Black,
1940), typical oxygen dissociation curves obtained in the absence of
CO2 are shifted to the right by CO2. The shape of the curve may be
altered as well as its position (Root, 1931), and one change in particular
appears in the failure of the hemoglobin to become saturated with oxygen
even at tensions of 150 mm. As a consequence of this situation the

OXYGENATION IN TROUT BLOOD

tension of O2 required for half saturation cannot serve as in the ab-
sence of CO2 to define the whole dissociation curve.

Up to about 65 per cent saturation, however, oxygen dissociation
curves in the presence of 10 mm. CO2 rise approximately straight from
the origin. If the oxygen tension for a single degree of saturation be-
tween 40 and 60 per cent is determined, that point may then be used
to designate the curve as far as the 65 per cent level without appreciable
error, and the tension at half saturation locates part of the curve as well
as the important physiological condition during unloading of oxygen.

The effect of 10 mm. tension of CO2 upon the oxygen tension needed
for half saturation is shown by the upper curves in Figs. 3, 4, and 5 at
each temperature. The distribution of points at any temperature
amounted to about 10 mm. in the blood of rainbow and brown trout,

30,-

BROOK

TEMPERATURE

FIG. 7. Increase in PO, necessary for half-saturation when PCO2 is increased from

1 to 10 mm. at different temperatures.

but was less in brook trout. The curves do not miss any point by more
than 5 mm., and describe the condition of the blood closely, considering
the number of samples of blood examined.

When the curves are superimposed in Fig. 6, it is apparent that at
15° the blood of brook and brown trout are alike with and without CO2.
The effect of CO2 upon rainbow trout blood is less by nearly 10 mm.,
which is a greater difference than is shown by individual variations.
Over the full range of temperature the curve representing brook trout
blood with CO2 differs from the curve for the other two species in posi-
tion and slope.

If the addition of CO2 to the blood occurred in the tissues, the dimin-
ished affinity of hemoglobin for oxygen would facilitate the diffusion of

oxygen from the blood into the tissues. The situation resembles that

12 L. IRVING, E. C. BLACK AND V. SAFFORD

produced by the Bohr effect in favoring unloading of oxygen from
human blood, but in the blood of many fish the influence of CO2 is much
greater. In Fig. 7 there is shown the influence which 10 mm. CCX would
have in raising the oxygen tension at half saturation and at various tem-
peratures. At 5° the addition of 10 mm. CO2 requires a rise in oxygen
tension for half saturation of 17 mm. in rainbow trout, 20 mm. in brook
trout, and 30 mm. in brown trout. The oxygen tension is, however,
three times increased by CCX in brook trout and only doubled in rainbow
trout. The practical bearing of specific differences of this size upon
the unloading tensions of oxygen would be important in respiratory
transport. It is indicated that the blood of the three species suits its
use under different conditions, particularly at temperatures away from
15°, but during unloading in the tissues rather than during loading in
the gills.

PROPERTIES OF THE ERYTHROCYTES RELATED TO THE CO2 EFFECT

Effect of Hemolysis

It is an easy custom to attribute differences among different specific
types of blood to specific properties of the hemoglobin involved. But
the pure hemoglobin of fish is not known, and there are some striking
illustrations that in fish blood the properties of hemoglobin depend upon
its condition within the erythrocyte. The effects of CO, upon CX affinity
largely disappear from the blood of the carp and sucker when the blood
is hemolyzed (Black and Irving, 1938). The CO2 effect likewise de-
pends upon the integrity of the erythrocytes in the blood of the tautog
and toadfish. On the other hand, the hemolyzed blood of the sea robin
still retains part of its sensitivity to CO2 (Root, Irving, and Black, 1939).
The blood of the Atlantic salmon (Irving, 1939) is still sensitive to CO2
after hemolysis, and so there are instances for the dependence of CO2
sensitivity upon the erythrocytes in some species and independence in
other species.

Samples of normal trout blood and of blood hemolyzed with saponin
were equilibrated with 150 mm. pressure of oxygen and several pres-
sures of CO2 up to about 80 mm., according to the procedure outlined
earlier for determining the limit of the CO2 effect. The CO2 effect in
hemolyzed blood proved to be only slightly less than in whole blood.
The limit of the CO2 effect was about 50 per cent reduction of the hemo-
globin in whole blood and about 40 per cent reduction in hemolyzed
blood. Evidently the hemoglobin of trout blood, like that of salmon and
sea robins, is sensitive to CO2 whether in the erythrocytes or in hemo-
lyzed blood.

OXYGENATION IN TROUT BLOOD

Changes in Cell Volume

We have observed that the erythrocytes of a number of fish swell to
a remarkable extent when the CO2 tension is increased. The erythro-
cytes of suckers (Black, 1940), tautog, sea robins, and toadfish (un-
published observations), and Atlantic salmon (Irving. 1939), and rain-
bow trout (Ferguson and Black, 1940) swell considerably, while the
cells of carp (Black, 1940) swell very little with CCX. In the rather
small number of species examined swelling is great in those in which the
CCX effect is large, and small where the CO2 effect is small.

Comparing the erythrocyte volume of the blood of the trout when
the blood was equilibrated with air, with the erythrocyte volume when
the same blood was equilibrated with 10 mm. CCX and about half satu-
rated with oxygen showed that CCX caused swelling in every case.
These volume changes are shown in Table III as the increase in per-

TABLE III
Increase in volume of erythrocytes produced by 10 mm.

Species

Number

Range of swelling

Average of swelling

per cent

Brown trout

1-21

Trutta trutta

Rainbow trout

4-27

Trutta iridea

Brook trout

1-24

Salvelinus fontinalis

centage over the original volume. Removal of the CCX reduced the
volume of the erythrocytes again. The swelling of fish erythrocytes
with CCX is more variable than would be expected if CCX tension is the
only variable factor which determines volume changes, and erythrocyte
volume is certainly labile toward factors or conditions other than CO2
tension alone.

CCX DISSOCIATION CURVES

The data from the determinations of CO, effects were plotted loga-
rithmically as content against tension of CCX. In all of the blood sam-
ples considered the hemoglobin was about half saturated. The points
were uniformly distributed, and the regularity justified the construction
of the average curves shown in Fig. 8. The points show the mean po-
sition of the results obtained with each species, and indicate that there
was no difference between the average blood of brook and brown trout.

L. IRVING, E. C. BLACK AND V. SAFFORD

Blood of rainbow trout has, however, consistently a somewhat smaller
CO2-combining power than the blood of the others. Judging from the
slope of the curves, the buffering of the blood of the three species is
similar at CO2 tensions greater than 10 mm. Up to 10 mm. the buffering
of brook and brown trout somewhat exceeds that of rainbow trout, al-
though the concentration of hemoglobin in the latter is usually larger
than in the other two species.

The amount of CCX combined with blood increases with diminishing

• Brook
O Brown
A Rainbow

P C02
FIG. 8. CO2 dissociation curves of trout blood at 3°, 15°, and 25°.

temperature more rapidly than does the solubility of CO, alone. If the
CO2 added to the blood increases the anion concentration, base must be
removed by CO2 from the protein buffers, which, as weak acids, must
then undergo a decline in strength in comparison with carbonic acid.
The amount of base lost by proteins and gained by CO2 is equivalent to
the change from 26.0 to 36.0 volumes per cent of combined CO2 between
25° and 3° or 4.5 millimols per liter. Along with this decline in base-
binding power of the protein, the affinity of the hemoglobin for oxygen
increases as the temperature declines.

OXYGEN ATION IN TROUT BLOOD 15

Increasing the temperature diminishes the affinity of the hemoglobin
for oxygen and increases the ability of the hemoglobin to bind base.
The affinity of hemoglobin for oxygen may also be diminished by acidi-
fication, but with an opposite effect upon ability to bind base. From
these contrasting relations between the effect of temperature and acidity
upon oxygen affinity it appears likely that change of temperature does
not alter the affinity of hemoglobin for oxygen by affecting its ionization.

DISCUSSION

The affinity for oxygen of the blood of the three species of trout is
scarcely distinguishable at 15° and in the absence of CO,. In the ar-
terial blood of trout the tension of CO2 is probably only one or two
millimeters (Ferguson and Black, 1940), and complete oxygenation in
efficient gills would be equally secured for all three species by atmospheric
tensions of oxygen in water cooler than 25°. At 20° half of the
atmospheric tension of oxygen would suffice to saturate the blood, and
it seems likely that under common natural conditions of temperature and
oxygen supply the blood of all three species would be sufficiently and
about equally oxygenated.

But when the effects of CO2 and changing temperature are consid-
ered, differences appear which distinguish each species. At 5° the nor-
mal tension of CO2, which is about 10 mm. in venous blood, would yield,
at half saturation, as unloading pressure, in brook trout 29 mm., in rain-
bow trout 32 mm., and in brown trout 39 mm. At 25°, the pressures
would be 54 mm. in brook trout, 47 mm. in brown trout, and 39 in rain-
bow trout. The condition of half saturation and with 10 mm. tension
of COo represents the situation in the blood when O2 is passing into the
tissues. At low temperature the tension of O2 available for unloading
would be greatest in brown trout, while the blood of brook trout would
offer greater unloading tension at higher temperature. The charac-
teristics of unloading are distinct for each species at all temperatures
except at 15°, and the change in unloading conditions with temperature
is likewise distinct in the blood of each species. The natural conse-
quence would afford to the brown .trout an unloading tension greater
than that of the other three at low temperatures. At high temperature
the brook trout would have the advantage of greater unloading tension.

The specific differences which have been shown might be attributed
to the possession of hemoglobin of a different type by each species, but
on this point there is no evidence. Another view could regard the dif-
ferences as based upon the influence of the milieu upon the oxygen af-
finity of a type of hemoglobin common to the three species. Hemoglobin
is extremely sensitive to the acidity and salts in which it exists, and in

16 L. IRVING, E. C. BLACK AND V. S AFFORD

natural conditions the slope of the CO., curve in trout blood is very
steep. The changes in acidity within the erythrocyte may be quite rapid,
and that the osmotic changes are apparently quite large is indicated by
the considerable swelling of erythrocytes produced by a 10 mm. increase
of CO2 tension. We have not seen any specific difference in the erythro-
cytes which would affect oxygen affinity, but the differences which we
have shown are small, and our observations upon the lability of the
erythrocytes are gross. The influence of temperature is undoubtedly
exerted directly upon the hemoglobin, but temperature probably influ-
ences the properties of the erythrocytes as well, and so indirectly affects
the hemoglobin by altering its milieu. The lability of the milieu and
the sensitivity of hemoglobin are such that the respiratory functions of
quite similar hemoglobins might be greatly modified by the milieu.

SUMMARY

Various species of fish possess blood with different characteristic
ability for combining with oxygen. These differences appear to fit the
blood of each species for the transport of oxygen under special condi-
tions.

In three closely related species of trout the characteristics of oxygen
combination are similar at the low CO2 tension characteristic of arterial
blood. The effect of rising temperature upon the combination of oxygen
with the blood in vitro of Salvelinus fontinalis, Trutta trutta, and Trutta
iridea is to diminish the oxygen affinity. At 15° their blood is half
saturated at 17, 17, and 18 mm. tension of oxygen respectively, and
changing temperature increases the oxygen tension required for half
saturation about 1 mm. per degree. This situation prevails when the
tension of CO, is about 1 mm., and only at lower temperature does the
blood of rainbow trout become distinguishable from the other two in
requiring slightly greater oxygen tension for half saturation.

CO2 greatly decreases the affinity of the hemoglobin for oxygen.
The limit of the effect of CO2 is reached at about 60 mm., and at that
tension at 15° the hemoglobin is only half saturated. Raising the tem-
perature diminishes the degree of oxygen saturation in the presence of
CO,.

When the CO2 tension is 10 mm., half saturation with oxygen re-
quires about twice the tension of oxygen needed in the absence of CO2.
The curves representing change in oxygen affinity with temperature when
the CO2 tension is 10 mm. are different in position or slope for each of
the three species. The differences are large enough to fit the blood of
each species for oxygen transport under different conditions.

OXYGENATION IN TROUT BLOOD 17

The erythrocytes of trout blood may swell 25 per cent when the CO2
tension is increased from one to 10 mm. The swelling is observed in
the blood of several species of fish having hemoglobin which is sensitive
to CO2. The CO2 dissociation curves of the three trout are essentially
alike and vary in the same manner with temperature.

The difference observed in the blood of these three species would
apparently provide different conditions for unloading oxygen in the
tissues, and the change of unloading conditions with temperature is
peculiar to each species of trout. Only at temperatures above 20° would
aeration at the gills normally be restricted.

LITERATURE CITED

BLACK, EDGAR C, 1940. The transport of oxygen by the blood of freshwater fish.
BioL Bull, 79 : 215-229.

BLACK, E. C, AND LAURENCE IRVING, 1938. The effect of hemolysis upon the af-
finity of fish blood for oxygen. Jour. Cell, and Cotnp. Physiol., 12 : 255-
262.

BOHR, C., K. HASSELBALCH, AND A. KROGH, 1904. Ueher einen in biologischer
Beziehung wichtigen Einfluss, den die Kohlensaurespannung des Blutes
auf dessen Sauerstoffbindung iibt. Scand. Arch. Physiol., 16: 402.

BROWN, W. E. L., AND A. V. HILL, 1923. The oxygen-dissociation curve of blood
and its thermodynamical basis. Proc. Roy. Soc., B , 94 : 297-334.

DILL, D. B., H. T. EDWARDS, AND M. FLORKIN, 1932. Properties of the blood of
the skate (Raia oscillata). BioL Bull., 62: 23-36.

FERGUSON, J. K. W., AND E. C. BLACK, 1940. Unpublished data.

IRVING, LAURENCE, 1939. Examination of the oxygen dissociation curves of blood
of the Atlantic salmon while living in salt and in fresh water. Am. Phil.
Soc. Report of Committee on Research, 243-244.

JORDAN, DAVID STARR, 1929. Manual of the Vertebrate Animals of the North-
eastern United States. World Book Co., Yonkers-on-Hudson, N. Y.

KAWAMOTO, N., 1929. Physiological studies on the eel. II. The influence of tem-
perature and of the relative volume of the red corpuscles and plasma upon
the haemoglobin dissociation curve. Sci. Rep. Tohoku Impcr. Univ., Series
4, 4 : 643-659.

KROGH, A., AND I. LEITCH, 1919. The respiratory function of the blood in fishes.
Jour. Physiol., 52 : 288-300.

PETERS, JOHN P., AND DONALD D. VAN SLYKE, 1935. Quantitative Clinical Chem-
istry. Vol. II. Methods. Williams and Wilkins Co., Baltimore.

ROOT, R. W., 1931. The respiratory function of the blood of marine fishes. BioL
Bull., 61 : 427^56.

ROOT, R. W., AND LAURENCE IRVING, 1940. The influence of oxygenation upon the
transport of CO2 by the blood of the marine fish, Tautoga onitis. Jour.
Cell, and Comp. Physiol., 16 : 85-96.

ROOT, R. W., LAURENCE IRVING, AND E. C. BLACK, 1939. The effect of hemolysis
upon the combination of oxygen with the blood of some marine fishes.
Jour. Cell, and Comp. Physiol., 13: 303-313.

VAN SLYKE, D. D., AND J. M. NEILL, 1924. The determination of gases in blood
and other solutions by vacuum extraction and manometric measurement.
II. Jour. Biol. Chcm., 61 : 523-573.

WILLMER, E. N., 1934. Some observations on the respiration of certain tropical
fresh-water fishes. Jour. E.vp. BioL, 11 : 283-306.

( >X THE METHOD OF FEEDING OF FOUR PELECYPODS

G. E. MACGINITIE

(From the William C. Kerckhoff Marine Laboratory of the California Institute of
Tcchnohniy, Corona del Mar, California)

INTRODUCTION

The feeding tracts of many pelecypocls have been described, but
while many of these accounts have, approximated the true conditions,
there is one important fact that has been omitted, and it is upon this
that an understanding of their feeding methods depends.

When the pelecypods used in this experiment were feeding, a sheet
of mucus entirely covered the gill structure, and it is this mucous sheet
which strains out food material from the water vascular current. Cilia
serve only to create the current and move the mucus. With the excep-
tion of highly specialized or quite primitive forms, it is probable that
this method of feeding is general in the class Pelecypoda.

Except in cases where the animal has been left sufficiently long (usu-
ally long enough to begin the regeneration of the shell and mantle), a
pelecypod which has been cut open is not feeding. When disturbed, the
pelecypods that I have investigated cease feeding at once, and, when
brought into the laboratory, several days may elapse before they will
feed naturally. Therefore, it is necessary to make certain that the ani-
mal which is being investigated is adjusted to its surroundings and is
carrying on its activities exactly as though it were in its natural habitat.
Many pelecypods will begin feeding shortly after being brought into the
laboratory, provided they are not mutilated in any way. In such cases
feeding may begin within a few hours, and, in time, these animals actu-
ally become somewhat adapted to oft repeated disturbances.

MATERIALS AND METHODS

Four pelecypods were used, namely, the gaper clam, ScJiizotJiacnis
nuttallii; the mud flat scallop, Pecten circnlaris; the native West Coast
oyster, Ostrea lurida; and the West Coast mussel, Mytllus calif or nianns.
These represent a burowing form, a surface form, an above surface
form, and an open coast form. The first three use detritus for food,
and the fourth uses plankton.

FEEDING METHOD OF FOUR PELECYPODS 19

A hole was bored in different positions in one side of the shell of
different individuals of each species, so that ultimately all regions con-
nected with the feeding activities of each of these pelecypods could be
observed. These windows were covered with a thin piece of glass, the
size of the window depending upon what region and how much area
of the region was to be investigated.

The openings were made by grinding a portion of the shell away,
removing the underlying mantle, and then cementing a piece of cover-
glass over the hole. The cover-glass was cut to shape by means of car-
borundum points and cemented in place with United Mender. Man}'
cements were used, but the United Mender made by the United Sales
Co. of New York, Dallas and Los Angeles proved to be by far the
best ; for windows cemented in place with it remained in place for as
long as six months, although they were continuously immersed in ocean
water. Before the cement was applied both the cover-glass and the
rim of the shell around the opening were wiped clean with a clean cloth
dampened with 95 per cent alcohol.

It is necessary to have the shell surrounding the opening perfectly
flat before the cover-glass is cemented in place. A sander grinds faster
and generates less heat than an emery wheel. While one is grinding the
opening the animal should be dipped often into ocean water. A small
hand rotor and dentist's drills can be used to advantage to cut the open-
ing, and then only the surface surrounding the opening need be sanded
flat.

A binocular and microscope lamp were suitably mounted next the
aquarium in such a manner that they could be adjusted to any position.

After a window was put in a shell the animal was left undisturbed
in running sea water for about two weeks, then watched carefully
through the windows with the binocular to determine if feeding was
being carried on in a natural manner. This was determined by intro-
ducing into the water some non-irritating material which constitutes the
natural food of the animal.

The food material used was either a diatom culture or detritus, the
latter being preferable. When the surface of the mud of an estuary or
of the ocean is disturbed, a grayish turbidity results. The material caus-
ing this turbidity consists of decaying organic matter which is rich in
bacteria, protozoa and other organisms such as rhabdocoeles, nematode
worms, larvae of many species of marine animals, and, in addition, on
mud flats there are usually surface diatoms and single-celled algae. This
surface sediment or detritus, which constitutes the main, or, in many
cases, the only source of food for burrowing or surface pelecypods, is

20 G. E. MACGINITIE

obtainable in any quantity and is readily eaten by most pelecypods.
When introduced as food it does not disturb the feeding activities of
the clams. (A diatom culture was found to be better for Mytilits
calif ornianus.}

FEEDING

In the four species listed above, when feeding begins either mucus is
secreted at the upper edges of the gills and is carried in a sheet by the
frontal cilia to the free edges of the gills, or it is secreted more or less
uniformly over the entire surface. It is then carried in strings along
the edges of the gills to the labial palps. The palps perform a selective
function, at least to the extent of partially removing undesirable parti-
cles, while allowing the rest of the material to pass intact with the strings
of mucus directly into the esophagus in the form of food-laden strings
of mucus. The main point to be stressed here is that the sheet of mucus
covers the entire gill in much the same manner as described for tunicates
(MacGinitie, 1939), and intercepts all particles from the current of
water which passes through the gills and out through the dorsal or ex-
current channels.

I consider a pelecypod to be feeding when a sheet of mucus entirely
covers the gills, at which times all particles in the water, however small
they may be, are strained out by the mucus. I am referring to undis-
turbed animals. Mucus may be made to flow copiously from any por-
tion of a gill by direct stimulation, but it is very difficult to determine
just where and when the secretion of mucus for feeding takes place.
For this reason it took more than two years of careful observation to
be sure of the main points set forth in this paper. Since mucus itself
is perfectly transparent, the presence of the sheet of mucus is shown
only by the included detritus. When feeding is actually going on, as
witnessed through a window in the shell, the sheet of mucus may not
follow the grooves, but, as shown by particles in it, may be deflected
somewhat in an anterior direction, the particles carried by it crossing-
over the grooves, for they are carried by the mucus and not by the cilia.
For example, when the mucus sheet is present in Mytilus calif ornianus,
the pull of the cilia of the free edge of the gills causes the sheet to be
deflected ahead, particularly near the lower edge of the gills. Thus it is
seen that the sheet of mucus, and not the cilia, carries the food particles.

When particles are moved by the cilia in the absence of the mucous
sheet, as in the case of an opened clam when one valve and mantle have
been removed, such particles follow the grooves. As it is the frontal
cilia which move the mucus, it is to be expected that when the mucous
sheet is absent, that is, when feeding is not taking place, the particles

FEEDING METHOD OF FOUR PELECYPODS 21

being moved to the free edge of the gills will move parallel to the
grooves. The current of water created by the lateral cilia bordering the
grooves tends to hold the particles in the grooves as they are being moved
by the frontal cilia towards the free edges of the gills. In some pelecy-
pods particles will move both down the ridges and up the grooves. The
ciliary mechanisms have been worked out in great detail by Atkins ( 1936.
1937, 1938) .

When a small amount of carmine powder is mixed with the detritus
and introduced with the incurrent water, it is usually ingested. In
Mvtiliis calif ornianns and ScJiizotJiacrus nuttaUii if carmine alone is
introduced in the same manner, some of it will find lodgment on the
sheet of mucus already formed ; but the sheet nearly always will lie cut
off at the upper edge and the carmine which thereafter collects on the
gills will be carried down the grooves, thence forward along the edge
of the gill, and then dropped by the labial palps into the anterior portion
of the mantle cavity. In Mytilus californiamts rejected material that is
dropped into the anterior portion of the mantle cavity is carried poste-
riorly by grooves just within the mantle edge and issues from the pos-
terior end in a continuous string termed pseudofeces. I have carefully
observed, through windows at the anterior portion of several Mytilus
calif ornianus, the rejection of the undesirable material by the labial palps.
The palps spread apart and assume a transverse position. The mucous
threads from the edges of the gills travel directly to the bases of the
ventral palps, down their anterior edges and thence into the ventral
grooves of the mantle edge, where the laden threads of mucus travel to
the posterior end and out as the pseudofeces mentioned above. When
Mytilus californianits is feeding, the palps are laid backward outside
of and close to the gills.

In Schizothaerus nuttaUii, rejected material will be forcibly ejected
from the mantle cavity by a sharp contraction of the adductor muscles,
which quickly brings the valves together and squirts the water and re-
jected material from the mantle cavity out through the incurrent siphon
or opening.

This activity undoubtedly accounts for some of the squirting by
clams on mud flats when the tide is going out. As the tide is leaving
the mud flats, clams that have long siphons and burrow deeply squirt
much more often than they do when the tide is in and they are covered
by water. When the mud flats are nearly bare a considerable amount of
sand and other undesirable material stirred up by the action of the waves
may cause material that will be rejected to accumulate rather rapidly.
Long-necked clams usually eject water much more forcibly than do those
clams which live much nearer the surface. ScliizotJtacnts nuttaUii

G. E. MACGINITIE

squirts water to a height of from 3 to 5 feet. Such removal of rejected
material, or squirting, is also much more frequent when the tide is first
coming in, as will be evident to anyone who will take the trouble to stand
knee deep in the incoming tide in a clam bed and observe.

The mantle cavities of the four pelecypods were never free of mucus,
and particles are at all times conveyed to and along the edges of the
gills by mucus, but it is only at feeding times that the gills are covered
by the sheets of mucus. While feeding is going on these sheets are be-
ing continuously secreted and move slowly toward the free edges of the
gills.

It should be noted here that some mucus with its included particles
may find its way into the mouth when the sheet of mucus is not present.
I think this is sometimes due to testing the mucus for suitable food, for
the secretion of the mucous plate often follows such testing. Abnormal
ingestion of material often follows disturbance or mutilation. This is
clearly shown by placing carmine or carborundum on the gills of a
pelecypod which has been opened. Under such conditions the introduced
material may be carried to the mouth and ingested. This never happens
in the case of a pelecypod which is feeding normally in an aquarium as
observed through a window, for the introduction of even small amounts
of carmine or carborundum causes the pelecypod to cease feeding imme-
diately. I consider carborundum particles the most undesirable of all
materials to use in feeding experiments.

In connection with the above, I am of the opinion that the function
of the osphradia as " water testing organs " is over-emphasized in the
pelecypods. The region which seems to me to have the highest tactile
and " olfactory " sense is the region where the incurrent opening is
located. In Mytilus, Pecten and Ostrea it is the edge of the mantle ;
in Schizothacrus it is the finger-like papillae which partially close the
entrance of the incurrent siphon and act as a coarse strainer. When a
valve has been removed, part of the reception area for stimuli has been
removed, and the nervous system of the pelecypod is rather badly upset,
to say the least. What an animal does with the mucous threads from
its gills at such a time had better be disregarded.

DISCUSSION

It is unwise to speak of feeding in a pelecypod unless it is actually
observed doing so. Pelecypods are very sensitive to stimulation, either
mechanical or chemical (Hopkins, 1932a, 1932&), and sometimes will
cease feeding at the least movement or change in food material. In
general, I think it may be said that small or juvenile members of any

FEEDING METHOD OF FOUR PELECYPODS

species are less sensitive than the older and larger ones, for they adjust
themselves more quickly to handling and begin feeding sooner after the
window is placed in them. For this reason it is best to use as young
specimens as one can conveniently. They are quite erratic for several
days after being moved into the laboratory, and also after any major dis-
turbance. However, when once fixed and left alone for a considerable
length of time, which varies in each individual, they become much more
uniform in their feeding activities, although apparently none of them
ever feed continually. The rate of intake of water varies considerably,
and this is usually or perhaps always accompanied by some contraction
of the gills. Complete contraction of the gills shuts off the current of
water entirely, just as a similar contraction does in the tunicate basket
(MacGinitie, 1939). It is impossible for large particles to pass through
the gills. Whether such small particles as bacteria pass through or not
depends on whether or not the pelecypod is feeding.

It is well known that cilia often have a selective function, fine ex-
amples being the cilia of the pouch and funnel of Stentor (Schaeffer,
1910), and the egg and sperm collectors of Urcchis (MacGinitie, 1935).
Nevertheless, the separation of solid material from water currents is
much more efficiently done by straining such water through mucus. It
is not surprising, therefore, that mucus plays a much more important
role in the feeding mechanisms of plankton and detritus feeders than
it has been given credit for doing. Certainly it never should be said
that a pelecypod is feeding just because it is pumping or maintaining a
current through the mantle cavity.

In the light of the information presented here, it is interesting to
read other papers concerned with the feeding of pelecypods and even
certain gastropods (Orton, 1912). When the use of a mucous sheet
for straining food material from the water is understood it will elimi-
nate many points of discussion that have arisen. Although many pa-
pers concerned with the feeding of pelecypods have been written, only
a few are listed in this paper, for the literature on this subject is rather
voluminous. Therefore, the reader is referred to the following papers
for complete bibliographies on the subject (Atkins, 1936-38; Galtsoff,
1928; Hopkins, 1936; Nelson, 1938; Orton, 1912; Yonge, 1936; and
ZoBell and Feltham, 1938).

When it is understood that the food material of pelecypods in general
is strained from the water as it passes through a sheet of mucus, and
that feeding is being carried on only when such a sheet is present, it
will clear up practically all points of uncertainty that one meets in read-
ing about feeding methods and feeding experiments in the pelecypod
mollusks.

24 G. E. MACGINITIE

SUMMARY

1. Opening's were made through the valve and mantle of four spe-
cies of pelecypods. These were made in various positions in the valves
of many individuals so that ultimately all outer portions of the feeding
mechanisms could he observed. These openings were covered with
pieces of cover-glass cemented in place so as to form windows through
which the feeding activities could be watched.

2. The feeding activities were observed through a binocular without
in any way disturbing the animals.

3. Evidence is given to show that when a pelecypod is feeding a
sheet of mucus covers the gills, and it is this mucus which strains the
food material from the water, the cilia affording mechanical means for
its transportation.

4. While the pelecypod is feeding this mucus is constantly being
secreted and is carried to the food grooves bordering the gills, along
which it is transported to the mouth as strings of food-laden mucus.

LITERATURE CITED

ATKINS, DAPHNE, 1936. On the ciliary mechanisms and interrelationships of

lamellibranchs. Part I. Some new observations on sorting mechanisms.

Quart. Jour. Micr. Sci., 79 : 181-308.
ATKINS, DAPHNE, 1937a. On the ciliary mechanisms and interrelationships of

lamellibranchs. Part II. Sorting devices on the gills. Quart. Jour. Micr.

Sci., 79: 339-373.
ATKINS, DAPHNE, 19376. On the ciliary mechanisms and interrelationships of

lamellibranchs. Part III. Types of lamellibranch gills and their food

currents. Quart. Jour. Micr. Sci, 79: 375-421.
ATKINS, DAPHNE, 1937c. On the ciliary mechanisms and interrelationships of

lamellibranchs. Part IV. Cuticular fusion, with special reference to the

fourth aperture in certain lamellibranchs. Quart. Jour. Micr. Sci., 79 :

423^45.
ATKINS, DAPHNE, 1938. On the ciliary mechanisms and interrelationships of

lamellibranchs. Part VII. Latero-frontal cilia of the gill filaments and

their phylogenetic value. Quart. Jour. Micr. Sci., 80 : 345-436.
GALTSOFF, PAUL S., 1928. Experimental study of the function of the oyster gills

and its bearing on the problems of oyster culture and sanitary control of

the oyster industry. Bull. Bur. Fish., 44 : 1-39.
HOPKINS, A. E., 1932a. Sensory stimulation of the oyster, Ostrea virginica, by

chemicals. Bull. Bur. Fish., 47 : 249-261.

HOPKINS, A. E., 1932b. Chemical stimulation by salts in the oyster, Ostrea vir-
ginica. Jour. Expcr. Zool., 61 : 14—28.
HOPKINS, A. E., 1936. Adaptation of the feeding mechanism of the oyster

(Ostrea gigas) to changes in salinity. Bull. Bur. Fish., 48: 345-364.
MACGINITIE, G. E., 1935. Normal functioning and experimental behavior of the

egg and sperm collectors of the echiuroid, Urechis caupo. Jour. E.rper.

Zool., 70: 341-354.
MAC&NITIE, G. E., 1939. The method of feeding of tunicates. Biol. Bull., 77 :

443-447.

FEEDING METHOD OF FOUR PELECYPODS

NELSON, THURLOW C., 1938. The feeding mechanism of the oyster. I. On the

pallium and the branchial chambers of Ostrea virginica, O. edulis and

O. angulata, with comparisons with other species of the genus. Jour.

Morph.. 63: 1-61.
ORTON, J. H., 1912. The mode of feeding of Crepidula, with an account of the

current-producing mechanism in the mantle cavity, and some remarks on

the mode of feeding in gastropods and lamellibranchs. Jour. Mar. Biol.

Assoc., N. S., 9 : 444-478.
SCHAEFFER, ASA ARTHUR, 1910. Selection of food in Stentor coeruleus. Jour.

Exper. Zool., 8 : 75-132.
YONGE, C. M., 1936. Mode of life, feeding, digestion and symbiosis with Zooxan-

thellae in the Tridacnidae. Scientific Reports Great Barrier Reef E.vp.

1928-29, 1 : 283-321.
ZoBELL, CLAUDE E., AND CATHARINE B. FELTHAM, 1937-1938. Bacteria as food

for certain marine invertebrates. Scars Found. Jour. Mar. Res., 1 : 312-

327.

THE COLORATION AND COLOR CHANGES OF THE
GULF-WEED CRAB, PLANES MINUTUS

HAROLD B. HITCHCOCK

(From flic Bermuda Biological Station for Research, Inc., and the Department of
Zoology, University of U7cstcrn Ontario1)

Many observers have noted the remarkable adaptations in color pat-
tern of the gulf -weed fauna. Perhaps the most common and colorful
crustacean associated with the gulf -weed is the little grapsoid crab,
Planes minutus (L*). Its predominant color is brown of many shades
from yellow to red, matching the weed to which it clings. The brown
of many individuals is broken by conspicuous white patches of various
shapes and sizes, some sufficiently large to cover the entire carapace.
These white areas appear to be in imitation of the conspicuous calcareous
tubes of the annelid worms attached to the gulf-weed. A color plate
showing a few of the variations is given in Murray and Hjort's book
(1912). Little experimental work has been done with this crab. Cro-
zier (1918), who found some mahogany-colored Planes on a drifting
tree of that shade cast ashore at Bermuda, was unable to detect any
change in coloration after they had been kept for six days on the much
lighter gulf -weed. The coloration and color changes of one of the
other members of the gulf-weed fauna, the shrimp, Latrcutcs fnconnn,
have recently been described by Brown (1939).

In an attempt to discover whether the coloration of Planes is a fixed
pattern or an active adaptation to background, crabs were kept on dif-
ferent backgrounds and observed microscopically as well as grossly. A
white background was furnished by white china bowls. Other back-
grounds were obtained by painting the outer surface of clear glass dishes.
Ordinary commercial paints were used for this, the green, for example,
being of a dark shade marketed as " window blind green." The crabs
were exposed to these backgrounds for one day in full sunlight from
dawn to dusk. At sundown the bowls were placed under a bright electric
light until the crabs could be examined. Microscopic observations were
made on the flattened surface of the fifth leg, a region easily viewed
under low magnification (X 62). The results of these observations
are summarized in Table I.

1 This study was made possible by a grant from the James F. Porter Fund of
Harvard University.

COLOR CHANGES IN GULF-WEED CRAB

There are three kinds of chromatophores in Planes inimitus: white,
black and yellow. Of these the most prominent are the white cells,
which usually appear to he larger and more numerous than the hlack
cells. The yellow cells are smaller than either the whites or blacks, and
are hard to distinguish except in the contracted state, both because of
their diminutive size and because of masking by the other chromato-
phores. In some individuals these cells appear to be almost orange in

TABLE I

Responses of chromatophores in Planes m-inufus to different backgrounds

O A B C Number of Crabs Examined

White Background

Black cells 0 1 4 27

White cells 8 18 4 2 32

Yellow cells 3 12 15

Black Background

Black cells 27 0 3 0 30

White cells 03918 30

Yellow cells 18 6 6 0 30

Red Background

Black cells 24 5 6 0 35

White cells 3 12 8 12 35

Yellow cells 10 4 15 6 35

Blue Background

Black cells 0 0 5 18

White cells 13 9 1 0

Yellow cells 2 0 4 17

Yellow Background

Black cells 0 1 4 21 26

White cells 12 13 1 0 26

Yellow cells 1 0 11 14 26

Green Background

Black cells 18 5 19

White cells 5554 19

Yellow cells 4 0 11 4 19

Symbols: O — pigment fully dispersed; processes indistinguishable.

A— pigment partially dispersed; arborizations visible.

B — pigment partially concentrated; stellate.

C — pigment fully concentrated; punctate.

Symbols after Kleinholz (1937). See his paper for illustrations of the four
phases of pigment distribution.

color, and when viewed in the contracted state through the darker areas
of the leg they appear to be ruby. It is probable that the yellow pigment
is astacin, a carotinoid commonly found in Crustacea. This pigment
appears red when concentrated. It may be that the ruby appearance is
caused in part by the fact that they are viewed through a region of the
exoskeleton which is dark-brown in color.

On most of the backgrounds tested each type of chromatophore re-

HAROLD B. HITCHCOCK

acted fairly consistently, its pigment tending to become either concen-
trated or dispersed. In general the black cells and the yellow cells re-
acted similarly, the only clear exception being when the animals were
exposed to a red background. White cells and black cells reacted oppo-
sitely, the pigment of one becoming concentrated when that of the other
became dispersed. The responses of the white cells to red and green
backgrounds were not consistent, nor were those of the black cells to a
green background. Perhaps examination of more animals or a longer
exposure would have eliminated these apparent inconsistencies. How-
ever, it has been observed by others that there is in crustaceans some

TABLE II

Responses of Planes mimitus compared with those of Portunus cmceps and

Portnnus ordwayi

(Data on Portunus from Abramozvits (1935)).

Red

White Pigment Cells

Yellow Pigment Cells

Black Pigment Cells

Pigment

Cells

Back-

ground

Planes

Portu-

Planes

Portu-

Planes

Portu-

minu-

nus

minu-

nus

minu-

nus

tus

anceps

ordwayi

tus

anceps

ordwayi

tus

anceps

ordwayi

White

Black

Blue

Red

Yellow

Green

—

Symbols: D — dispersed,

I — intermediate,
C — concentrated.

variability in chromatophoral response not only when different indi-
viduals are compared but also when different regions of the same animal
are studied. The condition of the yellow pigment was frequently very
difficult to determine, which introduced the likelihood of observational
error.

In spite of chromatophoral responses, Planes is unable to effect color
adaptation rapidly, for animals kept all day on white or yellow back-
grounds became but slightly lighter than those kept on black. Almost
every individual from one background can be matched in coloration in
a group from a contrasting background. It is interesting to note that
Abramowitz (1935) observed similar behaviour in two cancroid crabs.
One of these, Portunus anceps, has the same three pigments as Planes,

COLOR CHANGES IN GULF-WEED CRAB 29

while the other, Portunus ordwayi, has a fourth pigment, red. The red
and black pigments were found to react similarly except perhaps to a
green background, where the black became " dispersed " and the red
" intermediate." Table II gives a comparison of the responses of Planes
and the two species of Portunus. Since the responses of the cancroid
crabs did not agree in all cases (one type of chromatophore apparently
being concentrated in the one species and dispersed in the other for the
same background), complete agreement between them and Planes is not
found. However, if only those cases where the two species of Portunus
are in agreement in their response are compared with Planes, there is
lack of uniformity in only three instances : on a blue background for
both the white and the red cells, and on a red background in the case
of yellow cells. These differences are probably of no significance in
view of the lack of consistent response which Abramowitz reported.

Direct comparison between the chromatophoral responses of Planes
and Latreutes is difficult because the latter has red and blue pigments in
addition to the white and yellow of Planes, and lacks black. The two
species are alike in exhibiting a wide variety of colors and color pat-
terns, but whereas Planes shows almost no alteration in appearance,
Latreutes shows the effects of physiological color change almost immedi-
ately (Brown, personal communication).

The failure of Planes to effect a rapid color change in spite of its
active chromatophoral responses may possibly be explained, at least in
part, by a study of the moulted exoskeleton, which is a faintly colored
replica of the intact skeleton. Each dark area or white spot on the
intact animal is present in the exoskeleton, the intermediate areas being
pale yellowish brown. Extra-chromatophoral pigment is found also in
the hypodermal cells. Because of the distribution of this pigment, crabs
whose chromatophores have reacted to a certain background do not be-
come better color-adapted upon moulting. Until the diffuse pigment can
be elaborated or destroyed, according to the prevailing condition of the
chromatophore, it prevents the changed state of the chromatophores from
becoming evident in the general appearance of the animal. This process
apparently takes considerable time.

The pattern of the individual crab is probably genetic, as Brown has
suggested for Latreutes. This is borne out by the observation made by
Beebe (1928) that embryos of Planes have marked differences in pat-
tern before hatching (p. 194). Yet it is clear from Crozier's report
of the mahogany-colored individuals that Planes can in time become
adapted to new backgrounds. The experiments reported above show
that the chromatophores of Planes are responsive, but that extracellular

M) HAROLD B. HITCHCOCK

1 dement in the hypodermis and exoskeleton prevents the animal from
effecting an immediate change in appearance.

LITERATURE CITED

ABRAMOWITZ, A. A., 1935. Color changes in cancroid crabs of Bermuda. Pr»c.

Xat. Acad. Sci, 21 : 677-681.
BEEBE, W., 1928. Nonsuch : Land of Water. Brewer, Warren and Putnam, New

York.
BROWN, F. A., JR., 1939. The coloration and color changes of the gulf-weed

shrimp, Latreutes fucorum. Am. Nat., 73: 564-568.
CROZIER, W. T., 1918. Note on the coloration of Planes minutus. Am. Nat., 52:

262-263.
KLEINHOLZ, L. H., 1937. Studies in the pigmentary system of Crustacea. I.

Color changes and diurnal rhythm in Ligia baudiniana. Rial. Bull., 72 :

24-36.
MURRAY, J., AND J. HJORT, 1912. The Depths of the Ocean. Macmillan and Co.,

London.

SOME EXPERIMENTS ON THE EFFECTS OF HYPOPHY-

SECTOMY AND PITUITARY IMPLANTATIONS ON

THE MALE FUNDULUS HETEROCLITUS *• 2

J. WENDELL BURGER

(Front the Alt. Desert Island Marine Biological Laboratory, Salsbury Coi'c, Maine,
and Trinity College, Hartford, Connecticut}

While sufficient work has been accomplished to show that the hy-
pophysis of fish secretes a gonadotropic principle, the hormonal rela-
tionships involved in the piscine sexual cyle are not clearly understood
(compare Matthews, 1940 for a summary). For Fwidiilus Matthews

(1939) found that after hypophysectomy, the testes failed to continue
gametogenesis during the breeding period. The injection of mammalian
pituitary extracts had no decisive effect upon the gonads. Matthews

(1940) found, however, that the implantation of Funduhts pituitaries
into non-hypophysectomized immature Funduhts induced gametogenic
activity, especially in the male. The present investigations are com-
plementary to those of Matthews. Positive effects from pituitary
implantations were secured in hypophysectomized adult male Funduhts.

Effects of Hypophysectomy

On July 3-6, 73 freshly captured, mature male Fundulus were hypo-
physectomized. The opercular approach was used for this very simple
operation. Control fish were given blank operations. All fish were
maintained under identical conditions in running sea water, the tem-
perature of which varied between 11° and 19° C. The water for the
most part was near or below 15° C. The fish were fed almost daily on
chopped clams. This diet appeared adequate, since the operated fish
deposited fat as do fish in nature during the summer.

Mortality in operated fish was about 30 per cent. No significant
difference was found between the mortality of hypophysectomized fish
and those which received a blank operation. Over a two-month experi-
mental period, the loss of the hypophysis seems to have little to do with
the viability of Fundulus.

1 Aided in part by a grant from the Penrose Fund of the American Philosophi-
cal Society; this grant administered in 1939-40 by T. H. Bissonnette.

2 There is some doubt if this fish is strictly speaking, Fundulus hcteroclitus.
It may be a related heteroclitoid form.

J. WENDELL BURGER

At the time of hypophysectomy, the testes had passed their maximal
development for the annual sexual cycle (Fig. 1). In nature, the maxi-
mal development occurring during the spring is followed by testicular
regression. By late August, spermiogenetic transformations have almost
ceased and the testes are practically devoid of sperm. As pointed out
by Burger (1940), testicular regression does not occur as rapidly for
fish kept in cool water (11°-17° C.) in the laboratory, as it does in the
warmer water of the natural habitat. The degree of testicular regres-
sion attained by September 1, in fish which received blank operations,
is shown in Fig. 2.

The effects on the testis after hypophysectomy were similar to those
described by Matthews (1939) for other periods in the sexual cycle,
viz., after complete testicular regression (October-December), and at
the beginning of normal spermatogenesis (March- April).

The testes underwent a rapid reduction in size. The most obvious
effect was a cessation of sperm formation. This cessation was not
immediately a complete one. One month after hypophysectomy, how-
ever, only rare cysts of spermatids could be found. Two months after
the operation, spermatids were absent in the six testes examined. At
no time were spermatogonial multiplications suppressed. These divi-
sions formed a well-defined cortical zone of spermatogonia. Cross-
sections of testes from hypophysectomized fish are shown in Figs. 3
(August 2, one month after hypophysectomy), and 4 (September 1,
two mouihs after operation, cf. with control, Fig. 2). Thus it would
appear that the loss of the hypophysis results in inhibition of spermato-
genetic stages beyond those of spermatogonial division. Nevertheless,

FIG. 1. Cross-section of a testis at time hypophysectomies were performed
(July 3-6). This testis has passed its maximal sperm production. Black patches
are sperm.

FIG. 2. Cross-section of a testis from a control fish, killed September 1 ; this
fish has experienced a blank operation. Black patches are sperm. The cortical
zone where spermatogenetic stages are visible, has become narrow (cf. Fig. 1).

FIG. 3. Cross-section of a testis from a hypophysectomized fish, killed August

2. The cortical zone contains almost nothing but spermatogonia. Spermiogenesis
has practically ceased.

FIG. 4. Cross-section of a testis from a hypophysectomized fish, killed Sep-
tember 2. The cortical zone of spermatogonia has grown more narrow (cf. Fig.

3, one month earlier).

FIG. 5. Cross-section of a hypophysectomized fish implanted with twenty
pituitaries, beginning August 17 and killed September 2. The spermatogenetic zone
has deepened noticeably. Newly-formed sperm are visible. Compare with control,
Fig. 4.

FIG. 6. Cross-section of a testis from a fish which received a blank operation ;
this fish was implanted with twenty pituitaries beginning August 17 and killed
September 2. Compare with control, Fig. 2.

EFFECTS OF HYPOPHYSECTOMY ON FUNDULUS

it does seem true that once spermatogenesis has been initiated, spermio-
genesis can continue for some time, and to a limited degree, in the
absence of the pituitary. The fact that this spermatogenesis is not
maintained in any great volume even for as long as one month, indicates,

' ''U;

di « •"••-••, j-Wi

t if »*'„';>*. .

Stffi

:3^H

^^-mp^:-^^^

" ' ' -

PLATE I

34 J. WENDELL BURGER

as Matthews (1939) has suggested, that the later stages of spermato-
genesis are more sensitive to the absence of the pituitary than are the
stages of spermatogonial division.

Effects of Pituitary Implantations

After the testes of the hypophysectomized fish had involuted suffi-
ciently, both hypophysectomized fish and some operated controls were
implanted with pituitaries from freshly captured, mature male Fundulus.
The receptors selected were close to each other in size. The size of
the donor fish was also roughly standardized. The implantations were
made intraperitoneally. Ten hypophysectomized fish and three operated
controls each received five fresh glands on each of the following days :
August 17, 21, 26, and 29. These fish, which each received twenty
glands, were sacrificed on September 2. Five hypophysectomized fish
and two operated controls each received five fresh glands on each of the
following days: August 21, 26, 29. These fish were killed on Sep-
tember 4. During this phase of the work, the water was almost con-
stantly near 13° C.

The effects of the implantations were striking. Within five days after
the first injections, the fish began to show pronounced sexual display
antics. These peculiar swimming movements are common in fish during
the breeding season and occur occasionally throughout the year in fish
kept in aquaria. These movements were noticeably absent in the un-
treated hypophysectomized fish. By the tenth day after the first injec-
tion, the fish were in a frenzy of display.

Both series of implantations caused a recrudescence of the testes.
After two weeks, the average weight of the testes of the fish which re-
ceived twenty implants was slightly more than doubled, and the average
volume was quadrupled, when compared with the average weight and
volume of untreated hypophysectomized controls. The cortical zone of
spermatogonia had deepened, while new transformations into spermato-
zoa were well established (Fig. 5). The hypophysectomized fish which
each received fifteen glands likewise, formed new sperms. The weight
and volume increases were the same as those for the fish which received
twenty glands. The spermatogenetic stages were also the same. The
non-hypophysectomized fish which were receptors of implants reacted as
did the hypophysectomized receptors (Fig. 6). The control hypophy-
sectomized fish (Fig. 4), and the control fish which experienced blank
operations (Fig. 2) showed no testicular recrudescence during the ex-
perimental period.

These results clearly demonstrate that the pituitaries of adult male

EFFECTS OF HYPOPHYSECTOMY ON FUNDULUS

Fundulus contain, and the testes of hypophysectomized and normal adult
fish are responsive to, gonadotropic material. Matthews (1940) has
shown that the immature testis of non-hypophysectomized Fundulus
can be excited by Fundulus pituitaries implanted into the body cavity.

Discussion

The present study, together with that of Matthews (1940), permits
an interpretation of the pituitary rhythm involved in the normal sexual
cycle. During the period of sexual regression, pituitary secretion gradu-
ally declines. The pituitary does not abruptly cease to secrete, since
active spermiogenesis in a gradually decreasing volume occurs during
this regression. When the pituitary is removed, spermatogenesis is more
quickly checked. In the late summer, pituitary secretion seems to be
almost or entirely absent. Beginning in the fall and going through the
winter, spermatogonial multiplications take place. . The divisions can oc-
cur in the absence of the pituitary, but in the hypophysectomized fish
there is no progressive increase in the number of these spermatogonia.
Hence, the piling up in the testis of spermatogonia must be supported
by hypophyseal secretion. The amount of this secretion is low, since
very few spermatozoa are formed during the fall and winter. The
spring spermatogenesis is accompanied by the highest phase of secretion.
The gonadotropic material of the Fundulus pituitary is thus responsible
for two phases of spermatogenesis: (1) the great volume of spermato-
gonial proliferation, and (2) maturation phenomena. It has been shown
by Burger (1939) and Matthews (1939a) that the progressive phases at
least of the sexual cycle of the male are influenced by the temperature of
the water.

Summary

Adult male Fundulus, hypophysectomized shortly after maximal tes-
ticular development, show an inhibition of spermatogenesis for stages
beyond those of spermatogonial multiplication. Spermatogonial divi-
sions do not become numerous. The inhibition of the later stages is not
immediately affected, since a few cysts continue to form sperm for as
long as one month after hypophysectomy. These results confirm those
of Matthews.

Both hypophysectomized adult male Fundulus and fish which re-
ceived blank operations were maintained until sexual regression was well
established. Implantations of twenty or fifteen pituitaries from normal
male Fundulus caused within two weeks a recrudescence of the testes.
Non-implanted controls showed none of this activity. It is concluded

36 J. WENDELL BURGER

that the pituitary of the adult male Fundulus contains gonadotropic ma-
terial and that the testes of adult F undid us, hypophysectomized or not,
are responsive to this material.

The normal relation of the pituitary to the annual sexual cycle is
discussed.

LITERATURE CITED

BURGER, J. W., 1939. Some experiments on the relation of the external environ-
ment to the spermatogenetic cycle of Fundulus heteroclitus (L.). Biol.
Bull., 77 : 96-103.

— , 1940. Some further experiments on the relation of the external environment
to the spermatogenetic cycle of Fundulus heteroclitus. Bull. Mt. Desert
Island Biol. Lab., 1940: 20-21.

MATTHEWS, S. A., 1939. The relationship between the pituitary gland and the
gonads in Fundulus. Biol. Bull, 76 : 241-250.

— , 1939a. The effects of light and temperature on the male sexual cycle in
Fundulus. Biol. Bull., 77: 92-95.

— , 1940. The effects of implanting adult hypophyses into sexually immature
Fundulus. Biol. Bull., 79: 207-214.

MELANOSIS IN THE C( >M MON COD. GADUS CALLARIAS L.,
ASSOCIATED WITH TREMATODE IXFECTION1

SIDXKV C. T. HSIAO -

(I'roin the Museum of (.',nn punitive Zoology, Hari'ani l')iirersily, und the
Woods Hole Ocettn<>(inif>hic Institution. }\'oods Hole. Mass.)

A codfish displaying an unusual degree of melanosis is described in
this paper. The fish was caught one mile north of Race Point, Province-
town, in March, 1940, by Mr. f. \Y. Lowes, who sent it to the Museum
of Comparative Zoology. Mr. \Yilliam C. Schroeder asked the writer
to make a histological study of the tegumentary system of the specimen
in a search for a possible clue to the cause of its melanosis.

METHOD

Samples of skin of one-half to one centimeter square were taken
from different parts of the head, trunk and fins. The regions employed
are indicated by letters and numbers shown in Fig. 1. Each sample was
dehydrated, cleared and mounted in balsam. With a calibrated ocular
micrometer ruled into squares, 3 to 5 separate square millimeters from

FIG. 1. Diagram of left side of cod showing regions from which skin samples
arc removed for comparative study. A, anal fin; B, body or trunk; C, caudal fin;
D, dorsal fin; E, eye; H, head; P, pectoral fin; and I7, ventral fin.

each piece of skin were measured under a binocular microscope, using
reflected light, and the number of pigment cells per square millimeter
determined and recorded from each region.

Small pieces of skin from the first dorsal fin (region D 1), the trunk,
directly ventral to the first dorsal fin (region B 1), and the cornea (re-

1 Contribution No. 253 of Woods Hole Oceanographic Institution.

2 China Foundation Research Fellow.

SIDNEY C. T. HSIAO

gion E) were sectioned and stained with Heidenhain's " azan " stain
which gives a blue color to the connective tissue, reddish yellow to the
muscles, carmine to the cyst wall formed by the host tissue, blue to the
cyst wall secreted by parasites and blue and carmine to the parasites
themselves.

For comparison a normal cod was treated in the same way.

The trunk muscles, gills, oesophagus, heart and peritoneum were ex-
amined for parasites. As the fish had been eviscerated, only remnants
of the cardiac portion of the stomach and liver were examined. The
cysts were isolated, stained, and mounted and some of them sectioned
and stained.

OBSERVATIONS

Superficial examination of the whole fish and microscopic study of
sections of its skin show that general cutaneous melanosis in this fish
is associated with parasitic infection which attacks the whole tegumentary

FIG. 2. A. Normal cod. B. Dark cod described in this paper.

system and the gill filaments. No parasitic cysts were found in t he-
somatic muscles, the peritoneum, the heart, the oesophagus, the rem-
nants of the liver and the cardiac end of the stomach which happened
to be left in the fish after its evisceration. These uninfected regions
exhibit no melanosis when compared with the corresponding regions
in the normal cod.

MELANOSIS IX THE COMMON COD

In general appearance this fish is strikingly different from an ordi-
nary cod in the presence of so many melanophores in the corneae and in
the skin over the fins and the dorsal half of the body that these parts
are actually black. The contrast in color between this and normal cod is
shown in A and B of Fig. 2, which are printed from one photographic-
negative and hence are of identical exposure. Instead of being smooth
and shiny, the skin is warty and rough. The tiny excrescences which
produce the roughness are covered with more melanophores than the
surrounding tissues. Parasitic cysts appear as white specks among the
melanophores. Inside the cysts different stages of the metacercaria of a
heterophyid trematocle are seen.

The cysts and melanophores are so abundant in the cornea that the
fish is blind and the eye scarcely distinguishable from the rest of the
head. The melanophores on the body above the lateral line and on the
dorsal and caudal fins are so numerous that they form a continuous
sheet, making it impossible to ascertain their number per unit area of
skin. In the less densely pigmented regions the dark cod has, on the
whole, about six times as many melanophores per square millimeter on
the head and six to nine times as many on the paired and anal fins as
has a normal cod. Table I shows the number of pigment cells per square
mm. for each of the 23 corresponding samples of skin from the dark
and normal cod. The last two columns show that the melanophores of
the dark cod are smaller than those of the normal fish. In the dark fish
the melanophores are more uniform in size. In Fig. 3, A and B, two
equal pieces of skin from the pectoral fin of a normal cod and this dark
cod are compared. The normal cod has only one-sixth as many pigment
cells as the dark individual.

Parasitic cysts are present in the tegumentary system from the tip
of the snout to the surface of the caudal fin, including both corneae.
When examined under a dissection microscope, they appear as small
white dots among thick masses of melanophores — the tips of the cysts
being free from pigment cells. From Fig. 3, C, it will be seen that the
rugose appearance of the skin is produced by a mass of parasitic cysts
under the epidermis which is thrown into folds. These cysts occur both
above and below each scale, which, when pulled off from the body,
always has a mass of cysts firmly attached to its two surfaces. The
connective tissues are hypertrophied so that the skin is more than three
times as thick as the normal skin from a corresponding part of the body
(Fig. 3, D). The cornea is also infested. In Fig. 3, F, which is a
photomicrograph of a 4 p. thick section of a piece of cornea, 10 trematocle
cysts can be seen from a field 1.86 mm. long and 0.66 mm. wide. This

SIDNEY C. T. HSIAO

cornea is more than twice as thick as a normal one and has melanophores
throughout its whole thickness.

TABLE I

Comparison of melanophores between dark and normal cod

Body region

Number of melanophores

Ratio of
melanophores:
Dark

Size of melanophores in mm.

Normal

Dark

Normal

Dark

Normal

Head :

H 1

121

0.15-0.2

0.1

H 2

0.3

0.1-0.2

H 3

0.3

0.15

H 4

0.3

0.25

Trunk:

R 1

Too

0.2

0.05-0.1

B 2

numerous

0.2

0.05-0.1

B 3

0.2

0.05-0.1

B 4

0.1

0.1-0.15

B 5

0.3-0.4

0.1-0.15

B 6

0.3

0.15-0.2

B 7

0.?

0.2-0.25

B 8

0.4

0.15-0.2

B 9

Contracted

0.2-0.3

Dorsal tin:

I) 1

Too

0.15-0.2

0.1

D 2

numerous

0.2

0.1

D 3

0.15-0.2

0.1-0.2

Anal tin:

A t

Contracted

0.2-0.3

A 2

0.2-0.4

0.2

Caudal tin :

C 1

Too

0.1

C 2

numerous

0.15-0.2

0.1

Pectoral fin :

P 1

0.4

0.1-0.15

Ventral fin :

V 1

0.2-0.25

0.2

Eye:

K 1

0.2

The parasitic cysts are thick-walled, ovoid in outline and white to
the naked eye. The majority of the cysts measure 0.33-0.38 mm. along
one principal diameter and 0.24-0.28 mm. along the other. The cap-

MELANOSIS IN THE COMMON COD

V M •/, l^ea 7.
$•[. kL-£'=4jir^.«fe

* V ' .-/i , J^^^^jt.^ ^ i'-> i "

:auf!«^SI-i4»\ iv». •• f •

-«tKS^«- "-;-: ,-- - -: ^>- -*^5^>^Jv->.

^SS?W "J.'f £t-3L: ' -ii^x-^

m^^^Hv-/. ;

t' ** *^^->r •* *j - • " ^^*- • '

FIG. 3. Photomicrographs of preparations from cod skin.

A. Skin from the pectoral fin (same region as in B) of a normal cod.

B. Skin from pectoral region (PI) unstained, showing part of the outline of
two cysts and melanophores.

C. Section of a piece of melanotic cod skin from region B 1 (below the first
dorsal and above the lateral line), showing the wavy epidermis and clusters of cysts.
Section of one scale is shown in this figure. Three cysts are seen under the scale,
while above it there is a large number of cysts.

D. Skin of normal cod from the same region and under the same magnification
as in C above, showing one scale in section and part of two others and the smoother
epidermis and less connective tissue.

E. Section of a metacercaria.

/'. Section of the cornea showing cysts and pigmentation.

42 SIDNEY C. T. HSIAO

sules secreted by the parasites to enclose themselves measure 0.21 mm.
by 0.14 mm. in the two principal diameters. The cyst walls are very
resistant to mechanical injuries and are transparent in unstained skin
cleared in xylene and mounted /'// toto in balsam. The cyst wall secreted
by the host stains bright carmine with Heidenhain's " azan " stain and
is laid down in concentric layers. It measures 0.028-0.07 mm. in thick-
ness. The cyst wall secreted by the parasite, on the other hand, is only
about 0.007 mm. thick and stains blue. In the gill filaments many smaller
cysts are seen. This difference in size is due to the thinness of the host
wall, for the smaller cysts contain parasites which, when measured along
the parasite wall, are of the same size as those enclosed in the larger
cysts of the gill filaments or the skin of the trunk. The host wall about
the parasite increases in thickness, with more concentric layers, as the
external size of the cyst increases. There is comparatively very little
pigmentation about the cysts in the gill. All the melanophores present
in the gill are arranged about the cysts with thick host walls. As the
host walls are laid down about the parasites in concentric layers cen-
trifugally and as the melanophores are associated with the more periph-
eral layers, it is probable that these pigment cells appear some time after
the infection occurs.

The cyst contains the coiled body of the metacercaria of a trematode
whose suckers can be distinguished through the transparent cyst walls.
In stained sections (Fig. 3, E] the suckers and the spines on the pos-
terior part of the body wall of the parasite can be seen easily. Rut as
the metacercariae are still young, it is not possible to work out the struc-
tures of the reproductive system of our material.

From the absence of the parasitic cysts inside the body of the fish
except in the tegumentary system and the gills, it may be inferred that
the trematode larvae infected the cod by boring from the outside. The
presence of very thin-walled parasitic cysts in the gill filaments indicates
that these were the latest site of infection.

Compared with other infected fishes, this cod shows an extraor-
dinarily heavy infection. Not only is the whole tegumentary system
completely infested with parasites, but the parasitic cysts are gathered in
groups several layers thick under the epidermis. According to Dr.
Stunkard (verbal communication), a cunner kept in a laboratory aquar-
ium for six weeks with 50 infected snails giving off thousands of
trematode larvae does not get nearly so heavily infected. It is the more
surprising when we consider that although in the ocean the cod was able
to move about, it nevertheless contracted such an enormous number of
parasites.

MELANOSIS IN THE COMMON COD 43

DISCUSSION

Many cases of melanosis associated with parasitism have been ob-
served among freshwater fishes. There are also a few records of
melanosis in marine fishes parasitized by trematode larvae. In 1884
Ryder reported his observations on the darkening of the skin in para-
sitized dinners from Woods Hole and Cape Breton, N. S. He thought
that these cysts were formed by the cercariae of some trematode and
that the pigment cells about the site of infection were either formed
(/r novo or gathered there by migration. Linton (1900) observed para-
sitic cysts on the skin of dinners in 1889, and in 1901 lie reported
similar infections on tautog, winter flounder, torn cod and eel and " less
so on others." In 1915 he recognized the similarity between these en-
cysted forms and the trematode Tocotrema lingua (Creplin). The
presence of this species on the gills of sea raven and on the skin of
dinners from Passamaquoddy Bay was reported by Cooper (1915).
Stunkard worked out experimentally (1930) the life history of this
trematode and identified it as Cryptocotyle lingua (Creplin), belonging
to the family Heterophyidae — Tocotrema being long suppressed as a
synonym of Cryptocotyle. Smith (1935) described a hyperplastic epi-
dermal disease in two winter flounders associated with a trematode
infection which was probably due to Cryptocotyle lingua. All these
observations were probably concerned with the same species. A second
species was reported by Gamble and Drew (1911) from Plymouth. A
whiting infected by trematode larvae showed abnormal pigmentation in
the form of black specks scattered over its pigmented areas and over
the conjunctiva. They suggested that the trematode was a species of
Holostonntui, probably H. cuticola. No melanosis due to parasitism in
the cod has been reported. In the cod described here the heaviness of
infection and the intensive reaction of the host in the hypertrophy of the
skin and the development of excessive melanophores are very remarkable.

Smith's experiments (1931, 1932) on the evoking of melanophores
through mechanical injury and the eruption of corial melanophores and
general cutaneous melanosis strongly suggest that these reactions are
related to repair and defense of the tissues. The general eruption of
melanophores in this cod, whose tegumentary system was completely
infected by trematode metacercariae, is probably a defensive reaction
against the parasite. There are different views on the question whether
the melanophores in a parasitized fish migrate to the site of infection
or are formed dc nova. In this particular cod the number of melano-
phores is so much in excess of that found in an ordinary cod and they

44 SIDNEY C. T. HSIAO

so completely cover the whole body that migration of melanophores can-
not account for them. They must have developed anew.

The development of the metacercariae is not advanced enough to
allow an exact identification of the parasite. However, the mode of
reaction of the host and the structures of the parasite, as far as they
can be determined, appear very similar to what Linton described for
C>'\ptocot\'lc lintjna, which is found to infect a variety of fishes such as
cunner, tautog, torn cod, eel, sea raven, winter flounder, etc. They sug-
gest that this cod is parasitized by a species related to Cryptocotyle.

SUMMARY

A very melanotic codfish is described which proved to be heavily
infested with metacercariae of a heterophyid trematode. The number
of parasites and the intensive reaction of the host in the development
of melanophores and hypertrophy of the dermis are greater than any
recorded for parasitized fish. The parasite may be a form related to
Cryptocotyle.

LITERATURE CITED

COOPER, A. R., 1915. Trematodes from marine and fresh-water fishes, including

one species of ectoparasitic turbellarian. Trans. Rov. Soc. Canada, Ser.

Ill, Sect. IV, 9 : 181-205.
GAMBLE, F. W. AND G. H. DREW, 1911. Note on abnormal pigmentation of a

whiting infected by trematode larvae. Jour. Mar. Biol. Ass'n. 9: 243.
LINTON, E., 1900. Fish parasites collected at Woods Hole. Bull. U. S. Fish Coinin.

for ]$W. 19: 281 and 462.

— , 1915. Tocotrema lingua (Creplin). Jour. Parasitol., 1: 128-134.
RYDER, J. A., 1884. On a skin parasite of the cunner (Ctenolabrus adspersus).

'Bull. U. S. Fish Coinin. for 1884. 4: 37-42.
SMITH, G. M., 1931. The occurrence of melanophores in certain experimental

wounds of the goldfish (Carassius auratus). Biol. Bull.. 61: 73-84.
— , 1932. Eruption of corial melanophores and general cutaneous melanosis in

the goldfish (Carassius auratus) following exposure to X-ray. Am. .Four.

Cancer. 16: 863-870.
— , 1935. A hyperplastic epidermal disease in the winter flounder infected with

Cryptocotyle lingua (Creplin). Am. Jour. Cancer. 25: 108-112.
STUNKARD, H. W., 1930. The life history of Cryptocotyle lingua (Creplin), with

notes on the physiology of the metacercariae. Jour. Morph. and Ph\siol.,

50: 143-191.

REPAYMENT OF THE ANAEROBIC OXYGEN DEBT IN
GRASSHOPPER SKELETAL MUSCLE

DARCY GILMOUR *

(From the Department of Physiology, School of Medicine and Dentistry, the
University of Rochester, Rochester, Nezv York)

It has been shown (Gilmour, 1940) that in the roach Cryptocercus
pwictulatus Scudcler, the " oxygen debt " incurred during a period of
anaerobiosis is repaid threefold. Since such a large excess oxygen
consumption during recovery had not previously been demonstrated in
animal tissues, it seemed that further work on the phenomenon of oxygen
debt repayment in insects was warranted.

In order to simplify the problem, it was decided to limit this inves-
tigation, as far as was possible, to one tissue. To this end, the hind
femora of grasshoppers were employed as material. The femur of the
jumping leg of the grasshopper consists almost entirely of skeletal
muscle ; the amount of epidermal and other tissue present \vould account
for only a very small fraction of the total respiration.

MATERIAL AND METHODS

Two species of grasshopper were used: (1) Melanoplus femur-
rubruin (De Geer). Adults of this species were collected in the field
during the fall of 1939 and kept in the laboratory until used. (2)
Melanoplus difjcrentlalis (Thomas). A number of adults were raised
from eggs supplied by Dr. J. H. Bodine from the stocks at Iowa State
University.

The legs were removed from the bodies by cutting through the
trochanter. The tibiae were cut off close to the proximal end.

Oxygen uptake was studied in a differential volumeter designed
especially for following oxygen debt repayment, the " apparatus B "
previously employed by Rotta and Stannard (1939). The electrical
conductivity method of Fenn (1928) was used in following carbon
dioxide production. Oxygen uptake and carbon dioxide production of
the resting leg were first followed in air, after which the vessels were
filled- with nitrogen, by running through the pure gas for 20 minutes,

1 James King of Irrawang Scholar of the University of Sydney ; permanent
address : Department of Zoology, University of Sydney, Sydney, N. S. W.,
Australia.

DARCY GILMOUR

and the carbon dioxide production during anaerobiosis was determined ;
finally air or oxygen was re-admitted and the recovery respiration fol-
lowed until a steady rate of oxygen uptake had been reestablished.

It was not possible to use the respirometer designed for conductivity
determination for following both oxygen uptake and carbon dioxide
production during the recovery period, as the oxygen uptake readings
were disturbed by solubility effects. The anaerobic period was ended
by passing pure oxygen through the vessel for one minute, insufficient
time for equilibrium to be established between the gas mixture and the
relatively large volume of barium hydroxide required by the method.
Oxygen uptake was followed, however, at the beginning and end of

TABLE I

Repayment of oxygen debt in isolated hind femora of Melanoplus femur-rubrum.

The duration of anaerobiosis sometimes varied slightly from that shown in
column 2. Such variations have been considered in calculating the amount of
oxygen missed.

Experi-
ment

Time in

Initial
Oj uptake

Oj missed

Final
O2 uptake

Excess 62
consumed

Excess O2 consumed xx

O-2 missed

minutes

mm.3lgm.[hr.
218

mm.3lgm.

110

mm.^lgm.jhr.
180

mm.3lgm.
193

175

180

173

160

180

186

179

160

269

135

277

250

185

G12

191

180

262

275

158

162

173

110

140

166

196

140

G10

201

162

550

275

Average

193

185

190

each run with the conductivity apparatus, in order to establish the
respiratory quotient.

The temperature of the experiments was 23° C.

That an adequate supply of oxygen to the interior of the excised legs
was maintained by diffusion from air was demonstrated by the fact that
filling the vessels with oxygen caused no increase in oxygen consumption
above that measured in air. The legs survived without any apparent
disturbance in respiratory metabolism throughout the course of the
experiments (5 to 9 hours). The oxygen uptake usually remained
constant over long periods of time, although the final steady rate was
often slightly lower than the initial.

Oo-DEBT REPAYMENT IN GRASSHOPPER MUSCLE

EXPERIMENTAL RESULTS

M. fomur-rubntui

Either two or three legs were used in each experiment. The average
oxygen uptake was 180 cu.mm. per gram live weight per hour (26
determinations ranging between 140 and 277 cu.mm. per gram per hour).
The results of the oxygen debt experiments are shown in Table I.
" Oxygen missed," in this table, means the amount of oxygen the tissue
would have consumed in air, during the time it was in nitrogen. The
last column represents the percentage repayment of the oxygen debt.

TABLE II

Repayment of oxygen debt and retention of carbon dioxide in isolated hind
femora of Melanoplus differentialis.

Per-

cent-

Expected

Measured

Experi-
ment

Initial
O-> uptake

Final
Oi uptake

age
repay-
ment

C02
produced

in N2

excess
CO-
produc-

excess
CO.
produc-

CO:

retained

of O2

tion

debt

mm.zjgm.lhr.

mm.3/gm.lhr.

mm.3/gm.

min.3/gm.

»i»i.3lgm.

mm.slgm.

178

164

120

—

180

0.93

180

0.80

—

136

225

212

132

—

134

—

175

—

— .

—

142

0.79

132

0.85

—

126

178

174

198

—

186

—

115

—

212

0.88

216

0.79

— .

154

265

245

M10

205

0.94

183

0.85

—

116

256

168

M13

207

—

195

—

160

—

M14

243

0.67

234

0.67

—

100

304

166

138

All 5

253

—

212

—

M16

277

0.72

237

0.73

—

346

194

152

M17

202

—

201

—

110

—

M18

187

0.98

207

0.89

—

123

234

176

Average

201

0.84

191

0.80

125

120

111

There is no significant difference between the value for this obtained
from the half-hour experiments and that from the one-hour experiments.

M. differentialis

One leg was used in each experiment. The procedure was to use
the femur of one side of the grasshopper in the oxygen debt apparatus,
and that of the other side in the respirometer designed for conductivity
determination.

The average oxygen uptake was 197 cu.mm. per gram per hour (29
determinations ranging between 132 and 277 cu.mm. per gram per
hour). Table II shows the results for both oxygen uptake and carbon
dioxide production. The percentage repayment of the oxygen debt is
calculated in the same manner as in Table I. The duration of anaerobic-

48 DARCY GILMOUR

sis was one hour in all experiments. The " expected excess carbon
dioxide production " is the amount of carbon dioxide which would have
been given off, over and above that produced as the result of basal
metabolism, if the recovery process had had a respiratory quotient of
1.0, and there had been no retention. The carbon dioxide retained in
the tissues during recovery is the difference between this figure and the
measured excess carbon dioxide production. The figures obtained in
this way show an extremely wide range of variation, but have an average
which is approximately equal to the average amount of carbon dioxide
produced during anaerobiosis. The variation must be due largely to
the fact that oxygen uptake and carbon dioxide production during re-
covery were not determined on the same tissue. In determining the
expected excess carbon dioxide production it was supposed that the
repayment of oxygen debt had the average value (125 per cent) in each
case; that is, that the figure for excess oxygen consumed (and hence for
carbon dioxide produced) was 125 per cent of the original oxygen uptake
per hour. It might be supposed that a more accurate method would be
to use, in each of the carbon dioxide experiments, the figure for oxygen
debt repayment obtained from the opposite leg of the same grasshopper.
When this is done, however, the variation is as great, while the average
is practically unchanged (106 cu.mm. per gram). It appears, then, that
it is impossible to predict the actual percentage repayment of oxygen
debt of any leg, even from an experiment run on another leg of the same
grasshopper. The average is thus the only figure for carbon dioxide
retention that need be considered.

The respiratory quotient is somewhat low for muscle, but at the time
at which experiments M14 and M16 were run the insects were rather
inadequately fed, and the low respiratory quotients in these experiments
are probably the result of the utilization of reserve foodstuffs.

DISCUSSION

Since the chemical constituents of insect muscle are quite similar to
those of vertebrate muscle, it is not unreasonable to expect the anaerobic
processes of the two groups to be at least qualitatively similar. The fact
that in M. differentialis the amount of carbon dioxide produced during
anaerobiosis is equal to the amount retained during recovery suggests
that the amount produced in anaerobiosis is the result simply of the
buffering of acid by bicarbonate, and supports the conclusion that lactic
acid is the only important end-product of anaerobiosis. In frog muscle
70 per cent of the debt incurred during anaerobiosis is repaid (Rotta
and Stannard, 1939). In the insects used in this study more than 100

O2-DEBT REPAYMENT IN GRASSHOPPER MUSCLE 49

per cent is repaid. The problem of the removal of lactic acid thus seems
to be a more expensive one in grasshopper muscle than in frog muscle,
particularly in the case of M. femur-rubntiii, which uses more oxygen
in the recovery process than does M . differcntialis. It has already been
suggested that in Cryptocercus (loc. clt.) the threefold repayment of the
oxygen debt may have been due to the burning off of a large proportion
of the lactic acid supposed to have been produced by anaerobiosis. The
fact that in isolated muscle tissue, investigated at normal temperature
(as Cryptocercus was not), a repayment of the oxygen debt in excess of
100 per cent can be demonstrated lends support to this conclusion.

SUMMARY

The average oxygen consumption of isolated hind femora of Mclano-
plus fcniur-rubrum was 180 cu.mm. per gram per hour; that of femora
of M. differcntialis was 197 cu. mm. per gram per hour. The average
respiratory quotient of the latter was 0.82.

In M. fcimir-rubrum 190 per cent of the oxygen debt incurred during
anaerobiosis was repaid during recovery. In M. differential-is 125 per
cent of the debt was repaid, and the carbon dioxide retained in the tis-
sues during the recovery period was equal to the carbon dioxide produced
during anaerobiosis.

The end-products of anaerobiosis in grasshoppers are probably simi-
lar to those in vertebrates, but their removal seems to involve a greater
expenditure of energy.

My thanks are due to Dr. W. O. Fenn for making available the facili-
ties of his laboratory, and for his interest in the course of this work.

LITERATURE CITED

FENN, W. O., 1928. A new method for the simultaneous determination of minute
amounts of carbon dioxide and oxygen. Am. Jour. Pliysiol., 84: 110-118.

GILMOUR, D., 1940. The anaerobic gaseous metabolism of the roach, Cryptocercus
punctulatus Scudder. Biol Bull., 79 : 297-308.

ROTTA, A., AND J. N. STANNARD, 1939. Studies on the oxygen debt of frog tissues.
Am. Jour. Physiol., 127: 281-289.

GROWTH STUDIES ON CILIATES

VII. COMPARATIVE GROWTH CHARACTERISTICS OF FOUR SPECIES OF

STERILE CILIATES

GEORGE W. KIDDER
(Arnold Biological Laboratory, Broii'ii University)

During the past year experiments have been conducted on four spe-
cies of holotrichous ciliates in pure culture in order to establish their
nutritional requirements and some of their characteristics of growth.
Tt is now possible to report the results of these experiments and to at-
tempt an analysis of some of the factors of growth, both favorable and
unfavorable.

In the ever widening field of protozoan physiology the quest is going
on for more species which can be used for precise experiments. These
species should be able to grow and reproduce in the absence of other
microorganisms if complete control is to be obtained. Up to the present
time it seems likely that the only genus of ciliate which has remained
in successful pure culture is Tetrahymena (Furgason, 1940). Various
names have been applied to pure-culture ciliates by different authors but,
as Furgason has succeeded in showing, they were probably dealing with
strains of Tetrahymena geleii. In a previous paper of this series (Kid-
der, Lilly and Claff, 1940) a description was given of a saprozoic ciliate
which was referred to the genus Glaucoma. This organism (G. vorax)
was described before access was had to Furgason's excellent work. I
am now of the opinion that our ciliate belongs to the genus Tetrahymena
and therefore it will be referred to in the future as Tetrahymena t'ora.v.

Paramccinm bnrsaria was cultured bacteria-free by Loefer (1936)
but these cultures were subsequently lost. Because of the inclusions of
Chlorella in this species the status of "pure culture" is questionable.

The four species to be dealt with in the following report are Tetra-
hymena geleii (strain W), T. vora.v, Glaucoma scintillans and Colpidiuin
campylum. All of these organisms were sterilized and established in
pure culture in this laboratory and remain available to other investi-
gators who may be interested in them for experimental purposes.

GROWTH CHARACTERISTICS OF CILIATES

MATERIAL AND METHODS

Isolation and Sterilization

Tetrahymena gclcii (strain W) was isolated from Mill Pond in
Woods Hole, Massachusetts in July, 1939. It was sterilized in the
migration-dilution apparatus described by Gaff (1940) and established
in pure culture.

Tetrahymena vorax is the strain previously described from this labo-
ratory (Kidder, Lilly and Claff, 1940).

Glaucoma scintillans (strain A) was isolated from Mill Pond in
July, 1939. It was sterilized in the migration-dilution apparatus of
Claff and established in pure culture. Strain B was isolated from a
freshwater stream near Providence, Rhode Island in May, 1940. It

FIG. 1. Tetrahymena geleii (strain W). X 800. Total number of ciliary
meridians == 17 — 19.

FIG. 2. Glaucoma scintillans. X 800. Total number of ciliary meridians =
35 — 40.

FIG. 3. Colpidium campyhtm. X 800. Total number of ciliary meridians =
27 — 30.

was sterilized and established in the same manner as strain A. Strain B
will be discussed only in reference to adaptation to sterile conditions as
it was not used in any of the other comparative studies.

Colpidium campylum was isolated from a freshwater stream near
Providence, Rhode Island in September, 1939. It was sterilized by mi-
gration across a fluid-filled Petri dish (details of this method given else-
where, Kidder, 1940) and established in pure culture.

Description of Species

Because of the confusion which has resulted from lack of adequate
description of experimental material, three figures are presented which

52 GEORGE W. KIDDER

show the diagnostic characteristics of those strains which were used and
which have not been figured previously. These figures were prepared
from opal blue treated material and indicate the distribution of the
ciliary lines and the position of the mouth. Figure 1 is of Tetrahyuicna
geleii (strain W) and corresponds almost exactly to the figures given
by Furgason (1940) for this species. Figure 2 represents Glaucoma
scintillans and Fig. 3 Colpidium campylum. No figure is given of Tctra-
hyinena vorax as descriptions have been previously presented (Kidder,
Lilly and Claff, 1940).

Studies of Glaucoma scintillans and Colpidium campylum have been
made using the silver technique of Klein (1926) and the relief method
of Bresslau (1922). Comparisons with slides prepared from different
strains which were used in a previous study (Kidder and Diller, 1934)
show that the present organisms are the same species as those reported
at that time.

Conditions of Experiments

All qualitative studies were made from cultures grown in the spe-
cially designed Pyrex flasks described in detail elsewhere (Kidder,
1941). Organisms were counted by the direct method after appropriate
dilutions. Qualitative observations were carried out on material grown
in Pyrex test tubes.

Incubation of all experimental cultures was at 27° C. ± 0.2°.

Except in the experiments designed to test the effects of the age of
the inoculum, all cultures were started from logarithmic phase ciliates.
For uniformity the following ages of inocula were always used : Tctra-
hymena geleii (strain W) — 18 hours; T. vorax — 24 hours; Glaucoma
scintillans and Colpidium campylum — 48 hours.

Sterility tests on solid and in liquid media, incubated at room tem-
perature and at 37° C., were carried out according to the methods out-
lined in previous studies of this series (Kidder and Stuart, 1939 ; Kidder,
Lilly and Claff, 1940; Dewey and Kidder, 1940; Kidder, 1941) and,
unless otherwise stated, all cultures were bacteria- free.

Method of Evaluation of Data

Attention should be called to an important point regarding the
presentation and evaluation of data. The method often employed (Hall
and Elliott, 1935; Hall, 1939; Hall and Schoenborn, 1939; etc.) of
comparing the final concentration of cells (X) to the initial concentra-
tion (Xo) and expressing the result as the ratio X/Xo may lead to
erroneous conclusions. The time selected for the final concentration
count is arbitrary and may represent a point on the growth curve beyond

GROWTH CHARACTERISTICS OF CILIATES

the cessation of logarithmic growth. No information is obtained re-
garding the activity of the cultures during the earlier phases of growth.
The same criticisms apply to the method developed by Elliott (1939)
where total protoplasmic volumes are compared, unless estimations are
made in the early stages of the growth of the cultures. Therefore it
seems not only desirable but necessary to follow the growth of cultures
by taking numerous samples at regular intervals. The culture flasks
used in these experiments were designed for such a procedure (Kidder,
1941).

EXPERIMENTAL RESULTS

Physical Condition of Medium

Tctrahymena geleii (strain W) and T. rorax are both able to utilize
dissolved proteins. This fact was immediately apparent upon the initial
sterilization. They began rapid reproduction when placed in any of
the standard peptone media or in Difco yeast extract. The addition of
particles to such media did not increase the growth rate or the yield.
These ciliates. correspond to the other strains of Tctrahymena, therefore,
in their ability to grow and reproduce in dissolved materials. Evidence
is still lacking regarding LwofFs (1932) contention that saprozoic cili-
ates are able to take in polypeptides through the pellicle. We still do
not know whether extracellular enzymes are released which might hydro-
lyse proteins. If it can be shown that no such proteolytic enzymes are
released into the medium, then it seems fairly certain that nutritive mate-
rials, even in the dissolved state, enter food vacuoles by way of the cyto-
stome. This conclusion would be justified when it is noted that at least
five strains of Tetrahymcna (tested by V. C. Dewey in this laboratory)
have been found to exhibit perfectly normal growth characteristics in
dissolved casein. It seems highly improbable that whole protein mole-
cules could be absorbed through the pellicle.

Both Glaucoma scintillans and Colpidium cauipylnm are dependent
upon particles of nutritive materials. This fact was noted by E. and
M. Chatton (1923) for G. scintillans when they were able to obtain
growth on dead B. coli but not on dissolved proteins. Hetherington
(1933) reports the establishment of G. scintillans in yeast autolysate
but the ciliates failed to reproduce beyond a few divisions and the cul-
tures were presumably discarded.

When Glaucoma scintillans (strain A) was first sterilized single cili-
ates were placed in 2 per cent proteose peptone broth. After many days
only a few divisions had occurred and it was apparent that the medium
was inadequate. Difco yeast extract (1 per cent) and liquid yeast
autolysate (10 per cent) were no better. Those ciliates placed in par-

54 GEORGE W. KIDDER

ticulate Yeast-Harris (Kidder, Lilly and Claff, 1940; Kidder, 1940),
however, reproduced quite rapidly while those placed in a mixture of
Yeast-Harris and proteose peptone yielded thriving cultures (Fig. 4).
Yeast-Harris or the mixture with proteose peptone which had been ren-
dered particle-free by filtration gave no growth. It was later found that
strain B and Colpid'mm campyluui likewise require particles in the me-

» -v e -«• „ e

,-» e e

• YH- PP- D
0 YH-PP
e YH
o PP

) 24 48 72 96 120 144 168 192 216 240 264

HOURS

FIG. 4. Glaucoma scintillans. Growth curves constructed from the average
data of 5 separate experiments. YH == 1 per cent Yeast-Harris ; PP = 2 per cent
proteose peptone; YH — PP == 1 per cent Yeast-Harris plus 2 per cent proteose
peptone ; YH — PP — D == 1 per cent Yeast-Harris plus 2 per cent proteose pep-
tone plus 0.5 per cent dextrose.

dium and it was in this way that pure cultures of these strains were
established.

Glaucoma and Colpidium appear to possess feeding mechanisms
which are stimulated to ingestkm only by solid particles. The slight
amount of growth obtained in proteose peptone broth can be accounted
for by the few particles which are invariably present after sterilization.
When these particles are used up reproduction ceases.

After considerable experimentation the following basic medium was

GROWTH CHARACTERISTICS OF CILIATES

adopted as the best for general use in dealing with Glaucoma and
Colpidium—

Brewers Yeast-Harris 10 grams

Pyrex distilled water 1 liter

This is brought to a boil and filtered, first through cotton and then
through Schleicher and Schiill No. 595 filter paper. This does not re-
move the finer particles of the broken yeast cells and the resulting solu-
tion is slightly turbid. To this liter of 1 per cent Yeast-Harris is added
20 grams of Difco proteose peptone and the whole sterilized in the auto-
clave at 15 pounds pressure for 20 minutes. This forms the base for
the other experimental materials or may be used without additions. The
concentrations given appear to be near optimum for these two species
of ciliates as both higher and lower concentrations were inferior for
growth.

Chemical Condition of Medium

No experiments on the inorganic requirements are to be described
here as all of the media used contain sufficient inorganic constituents for
all species (Lwoff, 1932).

In all media used the peptones, proteoses and proteins offered an
adequate source of nitrogen. For the two species of Tetrahyuiena it
was possible to test the relative effectiveness of various types of protein
products on growth rate and maximum concentration. Various yeast
products were tested as a substitute for and in combination with the
standard proteose peptone. One per cent Difco yeast extract, one per
cent filtered Yeast-Harris and 10 per cent liquid yeast autolysate were
used. The rate of reproduction (as calculated by the formula

/ log 2'
" log b log a '

where g = the generation time and t = the time in hours during which
the population has been increasing, a = the number of cells per unit
volume at the beginning and b = the number of cells at the end of time,
t) in the case of T., geleii, was slightly lower in all three' types of yeast
media than in proteose peptone (Table I). The generation time for
T. vorax was more than doubled (as compared with 2 per cent proteose
peptone) in both yeast extract and yeast autolysate and was somewhat
greater in Yeast-Harris (Table I). Some product of yeast autolysis
(present in both the extract and the autolysate) seems to inhibit the
reproduction of this species.

In the above experiments the various yeast factors were presented
along with the yeast proteins. There remained the possibility that some

GEORGE W. KIDDER

TABLE I
Tetrahymena. Comparison of growth in protein media. Average of four experiments.

Medium

Generation time in hours
T. geleii
(strain W) T. vorax

1 per cent yeast extract

10 per cent liquid yeast autolysate

1 per cent Yeast-Harris (particular

2 per cent proteose peptone

3.51
3.34
3.65

2.78

7.07
7.90
4.61
3.54

of these factors might stimulate growth if more adequate protein prod-
ucts were present. Consequently a " yeast vitamin concentrate — Har-
ris " which is practically free of native protein was added to a basic
medium of proteose peptone. Various concentrations were tested on
Tetrahymena geleii. (strain W) and the results are given in Table II.

TABLE II

Tetrahymena gdeii (strain W). Comparison of growth after the addition of
various concentrations of Yeast Vitamin Concentrate (Harris) to a basic medium of
2 per cent proteose peptone plus 0.5 per cent dextrose. Average of two experiments.

Percentage Yeast
Vitamin Cone.

Generation
time

Population per
ml. at end of
log. phase

Maximum
yield

hours

cells /ml.

3.37

48000

310000

0.025

3.14

52000

330000

0.05

2.95

67000

380000

0.1

2.70

70000

400000

0.2

2.74

54000

365000

The reproductive rate increased with the concentration up to 0.1 per
cent but was slightly lowered at 0.2 per cent. The addition of yeast
concentrate consistently lowered the reproductive rate of T. vorax.

The addition of yeast vitamin concentrate to the particulate medium
used for Glaucoma and Colpidium had no significant effect up to a
concentration of 0.2 per cent although higher concentrations caused in-
hibition of growth. These observations are of little significance, how-
ever, as the basic medium must contain rather high concentrations of
the yeast factors.

The addition of a separate source of carbon to the basic proteose
peptone medium for Tetrahymena and the particulate medium for Glau-
coma and Colpidium increased the length of the logarithmic phase and
the maximum yield of all species except T. vorax. As was mentioned
in the initial report on T. vorax (Kidder, Lilly and Claff, 1940) dextrose

GROWTH CHARACTERISTICS OF CILIATES

TABLE III

Tetrahymena vorax. Comparison of growth after the addition of various con-
centrations of dextrose to a basic medium of 2 per cent proteose peptone. Average
of four experiments.

Percentage
dextrose

Generation
time

Population per
ml. at end of
log. phase

Maximum
yield

hours

cells /ml.

3.55

10000

70000

0.5

4.60

7000

42000

1.0

5.01

5500

32000

2.0

9.56

890

6000

decreases the division rate in direct proportion to its concentration (Ta-
ble III) and the maximum yield is lowered in the same manner.

Experiments were conducted to test the ability of the four species
of ciliates to ferment some of the more common carbohydrates. One
polysaccharide (Difco soluble starch), four disaccharides (Difco sac-
charose, Difco maltose, Difco lactose and Difco cellobiose), three mono-
saccharides (Difco dextrose, Difco levulose and Difco galactose) and
two pentose sugars (Special Chemicals arabinose and Difco xylose)
were used. To the two types of basic media 0.5 per cent of the above
carbohydrates and 0.02 per cent brom thymol blue were added. These
media were dispensed in Pyrex tubes and sterilization was accomplished
in the autoclave at 15 pounds pressure for 12 minutes. After cooling
each type of media was inoculated with the four species of ciliates and
the results were noted by the color change of the indicator after 96
hours in the case of Tetrahymena and 240 hours in the case of Glaucoma
and Colpidiuiu.

TABLE IV

Fermentation of carbohydrates. All carbohydrates added to basic protein
media in 0.5 per cent concentrations. Medium contained 0.02 per cent brom thymol
blue. Six experiments.

Carbohydrate

Organism
Tetrahymena
geleii Tetrahymena Glaucoma Colpidium
(strain W) vorax scintillans campylum

starch
sucrose
maltose
lactose
cellobiose
dextrose
levulose
galactose
arabinose
xylose

I I I I

58 GEORGE W. KIDDER

The results of these fermentation experiments are given in Table
IV. Colpidium alone failed to utilize starch and cellobiose. On the
other hand, Colpidium was able to utilize sucrose while the other three
species were not. None of the ciliates fermented galactose although
Colas-Belcour and Lwoff (1925) report fermentation of this monosac-
charide by their strain of Tetrahymena (Glaucoma piriformis). Galac-

2
tr4

O 3

0 24 48 72 96 120 144 168 192 216

HOURS

FIG. 5. Glaucoma scintillans. Effect on growth of the addition of carbohy-
drates to basic medium (1 per cent Yeast-Harris plus 2 per cent proteose peptone).
All carbohydrates added in 0.5 per cent concentrations. Average of 3 experiments.

tose, arabinose and xylose so inhibited the growth of all the species that
these carbohydrates were not used in the quantitative growth studies.
All of the other carbohydrates were re-investigated in growth flasks and
the cultures followed by frequent counts. The indicator was omitted
but otherwise the media were as above.

The division rate, length of logarithmic phase and maximum yield
were slightly increased by all of the carbohydrates fermented by Tetra-

GROWTH CHARACTERISTICS OF CILIATES

hymena gclcii (strain W). These increases were small but constant.
No significant differences could be detected between any of the media
containing fermentable carbohydrates. The acidity rose from an initial
pH 6.8 to a final pH 4.8 in those flasks containing starch, maltose, cello-
biose, dextrose or levulose while it fell in all others, including the control
flasks, to pH 7.2.

Tetrahymena vorax, although it was able to ferment starch, maltose,
cellobiose, dextrose and levulose (initial pH 6.8-final pH 5.4) was dis-

i 4

o
6 3

o
o

96 120

HOURS

144

168

192

216

FIG. 6. Colpidiuin campylum. Effect on growth of the addition of carbohy-
drates to basic medium ( 1 per cent Yeast-Harris plus 2 per cent proteose peptone ) .
All carbohydrates added in 0.5 per cent concentrations. Average of 5 experiments.

tinctly inhibited in its growth. Reproductive rate, length of logarithmic
phase and maximum yield were decreased wherever fermentation oc-
curred (see Table V for dextrose) and were unaffected (as compared
with controls) in media containing carbohydrates which were not at-
tacked (sucrose, lactose).

The most striking results of the addition of carbohydrates were
found in the cases of Glaucoma and Colpidium. These results are given
in Figs. 5 and 6.

The growth of Colpidium without carbohydrate was very slow and
the maximum yield was low (not greater than 800 per ml.). The addi-

GEORGE W. KIDDER

tion of 0.5 per cent dextrose, levulose, sucrose or maltose increased the
rate of growth during the logarithmic phase and the maximum yield was
increased to over 40,000 per ml. in some cases. For practical purposes,
therefore, an additional source of carbon is a necessity for this ciliate.
The situation is somewhat different with Glaucoma. The division
rate during the first 48 hours is not appreciably changed when a fer-
mentable carbohydrate is added. Without a separate source of carbon,
however, the end of the logarithmic phase is reached rather suddenly and

TABLE V
Summary of growth characteristics.

Organism

Medium

Genera-
Optimum tion
pH time

Popula-
tion at
end of
log. phase

Maximum
concen-
tration
per ml.

hours

Tetrahymena
grleu (W)

2 per cent pro-
teose peptone,
0.5 per cent
dextrose, 0.1

5.6-8.0 2.69

58000

395000

per cent Yeast
Vitamin Cone.

Tetrahynii-Hd

2 per cent

6.2-7.6 3.52

12000

110000

vorax

proteose

peptone

Glaucoma
scintillans

1 per cent
Yeast-Harris,

5.6-6.8 7.37

40000

270000

Colpidium
campylum

2 per cent pro-
teose peptone,
0.5 dextrose

5.4

11.56

7200

41000

the curve flattens, the concentration (about 42,000 per ml.) remaining
relatively constant thereafter for many days. The addition of sucrose
or lactose has no significant effect upon the cultures, but the addition of
dextrose, levulose, maltose, cellobiose or starch causes an increase in
the length of the logarithmic phase, a long phase of negative growth
acceleration and a final yield in excess of 200,000 per ml.

Optimum pH

A number of experiments were conducted to determine the optimum
pH limits for the four species of ciliates. For these experiments three

GROWTH CHARACTERISTICS OF CILIATES

types of media were used : 2 per cent pretense peptone for Tetrahymena
vorax; 2 per cent proteose peptone plus 0.5 per cent dextrose for T.
geleii (strain W) ; 1 per cent Yeast-Harris, 2 per cent proteose peptone,
0.5 per cent dextrose for Glaucoma and Colpidium. The pH was ad-
justed through a wide range of values (from pH 4.8 to pH 8.6) with
HC1 and NaOH.

The results of these experiments are contained in summary form in
Table V.

o -,

TETRAHYMENA G ELEI I CS TR Al N WJ
o TETRAHYMENA VORAX
• GLAUCOMA SCINTILLANS
o COLPIDIUM CAMPYLUM

72 96

HOURS

120

144

166

FlG.

7. Graphic comparison of the growth characteristics of four species, of
ciliates. Media as given in Table IX.

General Comparison of Groivth Characteristics

Several interesting facts are brought out when the four species of
ciliates are compared during the various phases of their growth. Figure
7 is a graph prepared from various experiments, each species growing
under optimum conditions. A summary of data is given in Table V.

The growth rate of Tetrahymena geleii (strain W) is higher than
any strain of ciliate in pure culture so far reported (# = 2.69 hrs.).

62 GEORGE W. KIDDER

Unlike strain P (Phelps, 1935; 1936) and strain H (Kidder, 1941) of
this species the negative acceleration period of strain W is quite long
and the stationary phase is short. The concentration declines rapidly
after the culture is approximately 60 hours old (initial inoculum of 100
cells per ml.) but the death rate decreases later (at about 120 hours)
so that a concentration of 30,000^1-0,000 cells per ml. is maintained for
many days.

Tetrahymena vorax grows at a regular rate (#==3.52 hrs.) only
during the first 24 hours. Thereafter a long negative growth accelera-
tion phase ensues and the stationary phase is not reached until the cul-
ture is approximately 96 hours old. There is no decline in concentra-
tion, however, for many days so, in this respect, T. vorax resembles
strains P and H of T. gclcii (Phelps, 1935, 1936; Kidder, 1941).

In general the shape of the growth curve of Glaucoma s c infill an s is
similar to that of T. vorax. The generation time during the logarithmic
phase (first 48 hours of growth) is 7.37 hours. This rate gradually
falls off and an extremely long negative growth acceleration phase takes
place during which time the concentration increases from approximately
40,000 cells per ml. to over 200,000 per ml. The shape of this curve is
reproducible under the conditions of these experiments.

The shape of the growth curve of Colpidium campylum is similar to
that of strain H of Tetrahymena gcleii (Kidder, 1941) although the
growth rate is very low (g- - 11.56 hrs.) as is the maximum yield (41,-
000 per ml.). This final concentration is maintained for many days.

Observations on Age and Size of Inoculum

As was stated earlier in this report, the ciliates used as inocula in
the foregoing experiments were all taken from their logarithmic growth
phases. Under these conditions no lag phase occurred. This corre-
sponds to previous findings on controlled cultures (Phelps, 1935 ; Dewey
and Kidder, 1940; Kidder, 1941). A lag phase invariably occurs if
the ciliates which form the inoculum are taken from cultures which have
passed the logarithmic growth phase. This statement holds for all
four species used in the present study and has been reported for other
strains and species (strains P and H of Tetrahymena geleii, Phelps,
1935 ; Kidder, 1941 ; Perispira ovum, Dewey and Kidder, 1940) . The
length of the lag phase increases, up to a certain point, in direct relation
to the age of the parent culture.

The course of the growth is not dependent upon the size of the
inoculum of Tetrahymena geleii, T. vorax or Glaucoma scintillans. Ex-
periments on this point were conducted using large growth flasks (1

GROWTH CHARACTERISTICS OF CILIATES 63

liter capacity) containing 500 ml. of media. Single ciliates were inocu-
lated into these flasks. In order to insure the inoculation of active,
single organisms they were first isolated into small containers made from
the lower portion of shell vials. These containers had been previously
placed in Petri dishes and sterilized. After the single ciliates had been
isolated and checked under the dissecting binocular for number and
activity, the whole container was lifted with sterile forceps and dropped
into the culture flask. The initial inoculum is, in this case, 0.002 cells
per ml. After growth has proceeded for sufficient time so that samples
include enough cells for determination of numbers, the generation time
is calculated and compared to the control flask which has received the
usual inoculum of 100 cells per ml. No significant difference between
the generation times in high and low inoculum cultures of the three spe-
cies was obtained (Table VI). In no case did these single ciliates fail
to establish perfectly normal cultures.

TABLE VI

Effect of size of inoculum. Figures represent generation time in hours.
Volume of medium = 500 ml.

No. of cells
per ml. inocu-
lated

Tetrahymena
geleii
(strain W)

Tetrahymena
vorax

Organism

Glaucoma
scinlillans

Colpidium
campylum

100

2.69

3.52

7.37

11.66

0.002

2.71

3.48

7.41

18.25

Colpidium campylum did not give the same results (Table VI). In
a number of the flasks inoculated with single ciliates no growth oc-
curred. In those cultures which became established the generation time
was significantly increased (18.25 hours as compared with 11.66 hours
in the controls) and, what is more striking, the maximum yield was al-
ways very low (8,000 per ml. as compared with 40,000 per ml. in the
controls). This species does not follow the same course as the other
three and would seem to correspond to the reports of Robertson (1921-
1927) on non-sterile organisms and of Mast and Pace (1938) on Chilo-
monas. Considering the general characteristics of Colpidium, however,
I believe that there may be an explanation of the apparent " allelocata-
lytic " effect. Some substance or condition of the medium may be
slightly detrimental to this ciliate. When large numbers of organisms
are introduced no single cell receives a lethal amount of the toxic mate-
rial. When a single ciliate is introduced into a large amount of medium
it accumulates enough of the toxic substance to cause its death in some

64 GEORGE W. KIDDER

cases or to injure it in others. When the injury is sub-lethal it never-
theless permanently affects the cells. The lowering of the maximum
yield in single-cell-inoculated cultures which become established seems
to favor this hypothesis.

DISCUSSION

No complete analysis of the protein requirements of any ciliate is
available at the present time due to a number of factors. Lwoff (1932)
and Lwoff and Lwoff (1937) have obtained some data on their strain
of Tetrahymena gelcii (Glaucoma piriformis} but until all of the sup-
plementary factors in relation to nutrition are more perfectly known this
knowledge must remain incomplete. Neither Glaucoma scintillans nor
Colpidiwm campylum appears to offer satisfactory experimental material
for studies along this line. They both require particles. The particles
obtained from powdered yeast cells are of nearly unknown chemical con-
stitution. About all we can say concerning these particles is that they
are very complex. To these particles appear to be adsorbed the mole-
cules of proteoses and peptones necessary for the optimum growth. An
attempt was made to substitute animal charcoal (Norit) for the pow-
dered yeast. Colpidium failed to ingest these particles and while Glau-
coma did ingest them at first (black food vacuoles), they later refused
to do so and very little growth resulted. Various other inert materials
which were tried proved no more successful. There may well be other
types of particles (such as precipitated proteins, etc.) which could be
substituted but no appreciable advantage would be gained. Casein, a
well-known protein, was used successfully to supply the particles but
the growth was never as good as in the Yeast-Harris medium, even
though a filtrate of the Yeast-Harris was added.

It appears strange that Tetrahymena vorax is inhibited by some
factor in yeast while the reproduction of T. gelcii (strain W) is acceler-
ated. This situation is also true of the fermentable carbohydrates. No
answer to the question of these specific differences is available at present.
It will be interesting to compare the supplementary requirements of these
two species with the yeast factor question in mind.

Living organisms as food have been found to be a necessity for a
number of species of ciliates (Phelps, 1934; Kidder and Stuart, 1939;
etc.). This is not the case with Glaucoma and Colpidium, however.
The most apparent difference between growth on a favorable bacterium
and in pure culture is rate of reproduction. The living bacteria accel-
erate growth. This is not true in the case of Tetrahymena, where no
species of food organism tested was as favorable for growth as the
dissolved protein materials.

GROWTH CHARACTERISTICS OF CILIATES 65

The observations on the carbohydrates are interesting in showing
specific differences in enzyme production. While all four species of cili-
ates used in this investigation produce an amylase and a maltase, none
of them produce lactase. Colpidiwn canipylum stands alone in produc-
ing invertase and failing to produce cellobiase. All species ferment dex-
trose and levulose and fail to ferment galactose, arabinose and xylose.

With the exception of galactose and the pentose sugars, the carbo-
hydrates which were not fermented did not influence the growth of any
of the ciliates, although Elliottt (1935) reports some cases where accel-
eration of growth resulted without acid fermentation. These cases,
however, must be questioned as he calculated acceleration by yield after
a given time (usually 72 hours). The reason for questioning the va-
lidity of this method has been given in a previous section of this report.

Galactose, arabinose and xylose were found to be inhibitory to all
four species of ciliates. Elliott (1935) reports inhibition of Tetra-
hymena gclcii (strains H and E) by galactose, while Colas-Belcour and
Lwoff ( 1925) record the fermentation of galactose by their strain of
T. gelcii but give no data regarding growth.

In the experiments designed to test the effect of the initial pH of the
medium upon the growth of the ciliates investigated there was no indi-
cation that two optima exist as was reported by Elliott (1933) for his
strain of Tetrahymena gelcii. In fact, there were no significant dif-
ferences in generation time, length of logarithmic phase or maximum
yield over a wide pH range in the case of T. gclcii (strain W), T. vorax
or Glaucoma scintillans. Colpidium canipylum reproduces faster, for
a greater length of time and to greater final concentrations when the pH
of the medium is low (pH 5.4).

No data are available from these experiments as to the factors which
limit the period of maximum reproductive rate or cause the death of the
organisms during the later stages of the cultures. It should be pointed
out that the growth characteristics given are valid only under the condi-
tions outlined and might well be changed somewhat by varying these
conditions. The accumulation of volatile products of metabolism, such
as CCX, or the reduction of O2 tension could be largely overcome by
aeration. Phelps ( 1936) found that aeration increased the length of the
logarithmic phase of Tetrahymena gelcii (strain P) but did not alter the
generation time in the early stages of growth.

A point of some interest which should be brought out is what Elliott
(1933) and Johnson (1935) called "acclimatization." These authors
reporf the necessity for gradually reducing the number of bacteria in the
process of sterilizing their ciliates (Tetrahymena). These observations

66 GEORGE W. KIDDER

were not confirmed on the ciliates used in this study. In every case
establishment after complete sterilization followed immediately upon the
presentation of an adequate medium. Another type of acclimatization
was noted, however, in the case of Glaucoma scintillans. The growth
rate (strain A) increased steadily during the first three months of sterile
culture. The first calculations were based upon cursory data so this
point was checked with strain B. One week after its initial isolation
(May 27, 1940) growth flasks were inoculated and the generation time
during the logarithmic phase was determined and found to be 12.21
hours. Cultures started June 10 grew more rapidly (generation time
10.64 hours) while those started on July 12 and September 20 were
increasingly rapid (9.81 hours and 8.98 hours, respectively). Strain B,
therefore, repeated what had been noted for strain A and although this
strain does not reproduce as rapidly as strain A, even after four months,
the same tendency of gradual adaptation to the sterile medium is shown.

SUMMARY

1. The growth characteristics of four species of holotrichous ciliates
(Tetrahymena geleii, T. vora.v, Glaucoma scintillans and Colpidium
campylum), grown in pure culture, are given.

2. The two species of Tetrahymena are able to utilize dissolved nutri-
tive materials while Glaucoma and Colpidium, are dependent upon par-
ticulate materials in the media.

3. The growth of T. geleii is slightly accelerated by some factor in
yeast and by the presence of fermentable carbohydrates (dextrose, levu-
lose, maltose, cellobiose and starch) while inhibition of the growth of
T. vora.v results when these materials are present.

4. The maximum yield of Glaucoma and Colpidium is greatly in-
creased by fermentable carbohydrates.

5. Colpidium fails to ferment cellobiose but, unlike the other three
species, does ferment sucrose.

6. Galactose, arabinose and xylose, while not fermented by any of
the four species of ciliates, inhibit the growth of all.

7. The optimum range of pH values for T. geleii (strain W) is
wide (pH 5.6 --pH 8.0) ; T. vora.v is slightly more limited (pH 6.2-
pH 7.6); Glaucoma is limited to the acid range (pH 5.6--pH 6.8),
while Colpidium grows best at pH 5.4.

8. In the cases of Tetrahymena geleii, T. vora.v, and Glaucoma scin-
tillans when single ciliates from the logarithmic growth phase are inocu-
lated into 500 ml. of media (initial inoculum = 0.002 cells per ml.)
there is no lag phase and the generation time is not reduced (as com-
pared with controls).

GROWTH CHARACTERISTICS OF CILIATES 67

9. Single Colpidium campylum inoculated into 500 nil. of media
often die. When a culture is established the generation time is longer
and the maximum yield is smaller than when many cells are inoculated.
It is suggested that these results are correlated with slight toxicity of
the medium.

LITERATURE CITED

BRESSLAU, E., 1922. Zur Systematik der Ciliatengattung Colpidium. Zoo/.
Anscig., 55 : 21-28.

CHATTON, E., AND M. CHATTON, 1923. La sexualite provoquee experimentalement
chez un Infusoire: Glaucoma scintillans. Predominance des conditions du
milieu dans son determinisme. Comfit. Rend. Acad. Sci., 176: 1091-1093.

CLAFF, C. L., 1940. A migration-dilution apparatus for the sterilization of pro-
tozoa. Physiol. Zoo/., 13 : 334-343.

COLAS-BELCOUR, J., AND A. LWOFF, 1925. L'utilization des glucides par quelques
Protozoaires. Comfit. Rend. Soc. Biol., 93: 1421-1422.

DEWEY, V. C., AND G. W. KIDDER, 1940. Growth studies on ciliates. VI. Diag-
nosis, sterilization and growth characteristics of Perispira ovum. Biol.
Bull, 79: 255-271.

ELLIOTT, A. M., 1933. Isolation of Colpidium striatum Stokes in bacteria-free cul-
tures and the relation of growth to pH of the medium. Biol. Bull., 65 :
45-56.

ELLIOTT, A. M., 1935. Effects of carbohydrates on growth of Colpidium. Arch.
f. Protist., 84: 156-174.

ELLIOTT, A. M., 1939. A volumetric method for estimating population densities of
protozoa. Trans. Amcr. Mic. Soc., 58 : 97-99.

FURGASON, W. H., 1940. The significant cytostomal pattern of the " Glaucoma-
Colpidium group " and a proposed new genus and species, Tetrahymena
geleii. Arch. f. Protist. (in press).

HALL, R. P., 1939. Pimelic acid as a growth stimulant for Colpidium campylum.
Arch. f. Protist., 92: 315.

HALL, R. P., AND A. M. ELLIOTT, 1935. Growth of Colpidium in relation to
certain incomplete proteins and amino acids. Arch. f. Protist., 85 : 443-
450.

HALL, R. P., AND H. W. SCHOENBORN, 1939. Fluctuations in growth-rate of
Euglena anabaena, E. gracilis, and E. viridis, and their apparent relation
to initial density of population. Physiol. Zoo/., 12 : 201-208.

HETHERINGTON, A., 1933. The culture of some holotrichous ciliates. Arch. f.
Protist., 80 : 255-280.

JOHNSON, D., 1935. Isolation of Glaucoma ficaria Kahl in bacteria-free cultures,
and growth in relation to pH of the medium. Arch. f. Protist., 86 : 263-
277.

KIDDER, G. W., 1940. The technique and significance of control in protozoan cul-
ture. In : Protozoa in Biological Research. Ed. G. N. Calkins and F. M.
Summers. Columbia Univ. Press, New York.

KIDDER, G. W., 1941. Growth studies on ciliates. V. The acceleration and inhi-
bition of ciliate growth in biologically conditioned medium. Physiol. Zoo/,
(in press).

KIDDER, G. W., AND W. F. DILLER, 1934. Observations on the binary fission of
four species of common free-living ciliates, with special reference to the
macronuclear chromatin. Biol. Bull., 67 : 201-219.

KIDDER, G. W., D. M. LILLY AND C. L. CLAFF, 1940. Growth studies on ciliates.
IV. The influence of food on the structure and growth of Glaucoma vorax
sp. nov. Biol. Bull., 78: 9-23.

68 GEORGE W. KIDDER

KIDDER, G. W., AND C. A. STUART, 1939. Growth studies on ciliates. I. The role

of bacteria in the growth and reproduction of Colpoda. Physiol. Zool.,

12 : 329-340.
KLEIN, B. M., 1926. Ergebnisse mit einer Silbermethode bei Ciliaten. Arch. f.

Protist., 56 : 243-279.
LOEFER, J. B., 1936. Bacter-free culture of Paramecium bursaria and concentration

of the medium as a factor in growth. Jour. Expcr. Zool., 72 : 387-407.
LWOFF, A., 1932. Recherches biochimiques stir la nutrition des protozoaires. Mono-
graphics de 1'institut Pasteur. Paris.
LWOFF, A., AND M. LWOFF, 1937. L'aneurine, facteur de croissance pour le cilie

Glaucoma piriformis. Comfit. Rend. Soc. BioL, 126 : 644-646.
MAST, S. O., AND D. M. PACE, 1938. The effect of substances produced by Chilo-

monas paramecium on rate of reproduction. Physiol. Zool., 11 : 360-382.
PHELPS, A., 1934. Studies on the nutrition of Paramecium. Arch. f. Protist., 82 :

134-163.
PHELPS, A., 1935. Growth of protozoa in pure culture. I. Effect upon the growth

curve of the age of the inoculum and of the amount of the inoculum.

Jour. E.rpcr. Zool, 70: 109-130.
PHELPS, A., 1936. Growth of protozoa in pure culture. II. Effect upon the growth

curve of different concentrations of nutrient materials. Jour. Expcr. Zool.,

72 : 479-496.
ROBERTSON, T. B., 1921. Experimental studies on cellular multiplication. I. The

multiplication of isolated infusoria. Biochcin. Jour., 15: 595-611.
ROBERTSON, T. B., 1924. The nature of the factors which determine the duration

of the period of lag in cultures of infusoria. Australian Jour. Exfi. Biol.

and Med. Sci., 1 : 105-120.
ROBERTSON, T. B., 1927. On some conditions affecting the viability of cultures of

infusoria and the occurrence of allelocatalysis therein. Australian Jour.

E.vp. Biol. and Med. Sci., 4: 1-23.

THE EFFECT OF SAPONIN ON THE OSMOTIC

HEMOLYSIS OF CHICKEN

ERYTHROCYTES x

F. R. HUNTER, S. B. BARBER AND A. P. CAPUTI

(From flic Zoology Department, Rhode Island State College,
Kingston, Rhode Island}

Although the hemolysis of erythrocytes by lytic agents such as sa-
ponin has received much attention, few investigators have studied the
effect of these agents prior to hemolysis. Ponder (1937) reported a
decrease in the fragility of rabbit erythrocytes which had been exposed
to sub-lytic concentrations of lysins. This effect, he believed, was
similar to the action of narcotics (cf. Jacobs and Parpart, 1932). Dav-
son and Danielli (1938) reported that saponin • caused no loss of K+
from erythrocytes in either sub-hemolytic or hemolytic concentrations.

The present experiments were performed to determine what effect
saponin would have on the penetration of small molecules prior to the
time that the membrane became permeable to the hemoglobin molecule.

Chicken blood was obtained by cardiac puncture and then defibri-
nated. One half a cc. of blood suspended in 5 cc. of Ringer-Locke re-
quired approximately 0.2 cc. of 1 per cent saponin in Ringer-Locke to
cause slight hemolysis in two hours. A control suspension was similarly
prepared, omitting the saponin. The slight difference in total volume
(0.2 cc. in 5.5 cc.) did not introduce any error, since the method used
to measure permeability is not influenced by variations in the number of
cells within a wide range.

In this way stock suspensions of control and experimental cells
were prepared. In most experiments 0.2 cc. of these suspensions were
added to 10 cc. of an isosmotic solution of the substance whose rate of
penetration was to be measured. Permeability, fragility, and swelling
measurements were made at a temperature of 37° ± 0.5° C, using the
photronic cell technique usually employed in this laboratory (cf. Hunter,
1936; Hunter and Pahigian, 1940). In every experiment sufficient

1 One of the authors (F. R. H.) is indebted to the American Association for
the Advancement of Science for a grant-in-aid.

70 HUNTER, BARBER AND CAPUTI

NaHCO, was added to insure complete clearing of the suspension
(Stringer et al., 1940; Hunter et al., 1940). Permeability measure-
ments were made until the experimental cells were found to hemolyze
more rapidly than the controls when placed in the isosmotic solutions of
penetrating substances. At this time the experimental solution was
diluted with Ringer-Locke and centrifuged. In some of the experi-
ments the supernatant fluid after centrifugalization was colorless and
all of the packed cells were red, which indicated that the saponin had
not caused any hemolysis. In others, the supernatant fluid was red
and some of the cells were white. This indicated that the saponin had
been allowed sufficient time to destroy the membranes of some of the
cells and liberate the hemoglobin. Experiments in which this had hap-
pened served to indicate that the cells whose permeability was being
studied had been exposed to hemolytic concentrations of the lysin, but
the process had been stopped by removing the lytic agent before many
of the cells had been hemolyzed. After the cells had been centrifuged,
the supernatant fluid containing the saponin was discarded and the
cells were resuspended in a volume of Ringer-Locke sufficient to give a
suspension containing approximately the same number of cells per unit
volume as the control. Additional washing with Ringer-Locke was
unnecessary, as these cells would remain unhemolyzed for several days.
These resuspended cells exhibited the same permeable properties that they
had had immediately preceding the centrifugalization. The control cells
were not centrifuged in every experiment, since this treatment had no
marked effect on their permeability.

Some of the molecules penetrated the cells so rapidly at 37° C. that
measurements could not be made using the photronic cell technique. In
these cases the rate of hemolysis was measured by eye at room tem-
perature (about 23° C.). The complete hemolysis curve could not be
obtained in this way, but it was possible to make comparisons by measur-
ing the times for a given percentage of the cells to hemolyze.

Permeability measurements were begun as soon as the control and
experimental suspensions were prepared. A comparison of the effect
of saponin on the permeability to glycerol and monoacetin was made
using the photronic cell apparatus, while comparisons using the other
method were made between various rapidly penetrating lipoid-soluble
and insoluble molecules. Since it usually required at least one hour
before the saponin produced any marked effect, there was sufficient time
to make several measurements before the cells were centrifuged.

After the suspensions had been centrifuged and the cells resuspended
in Ringer-Locke, hemolysis measurements were made using a number

SAPONIN AND OSMOTIC HEMOLYSIS 71

of different substances. Centrifuged cells were used to obtain the
swelling and fragility data. In some experiments, higher concentrations
of saponin were used which produced slight hemolysis in 10—15 minutes.
Cells obtained in this way gave the same results as those which had
been exposed to lower concentrations of the lysin.

Ill

The data have been analyzed in several ways. In Table I are listed
the average number of seconds required for 80 per cent hemolysis (one-
half the total deflection) of the experimental and the control cells when
placed in isosmotic solutions of the various substances studied. Since
the change in hemolysis time depends on a number of factors such as
the concentration of saponin, and the time allowed for its action, there

TABLE I

The effect of saponin on the osmotic hemolysis of chicken erythrocytes.

Average time in seconds
for 80% hemolysis

Per cent
change

Number of
observations

Substance

Control

Experimental

*Water

-61.1

* *Ethylene glycol

-19.2

*Diethylene glycol

-48.0

Triethylene glycol

130

-33.1

Urea

148

112

- 24.3

Thiourea

193

148

-23.3

Glycerol

193

104

-46.1

Malonamide

868

273

-68.5

Erythritol

1125

292

-74.0

*Diacetin

-32.3

Monoacetin

144

-56.3

*Acetamide

-42.1

*Propionamide

-42.1

* Measurements made by eye.

is great variability observed in this value. To reduce this variability,
only the figures obtained after the cells had been centrifuged are re-
corded. The differences in the percentage change which remain result
from the fact that the saponin acted for a longer time in some experi-
ments than in others. The table merely indicates that the permeability
to a number of different substances has been greatly increased. The
data for those substances which are starred were obtained from measure-
ments made by eye. Figure 1 presents a representative control and
experimental curve.

Having established the fact that treatment with saponin could in-
crease the permeability of the cell membrane without causing hemolysis,

HUNTER, BARBER AND CAPUTI

the data were analyzed in an attempt to demonstrate any differential
effects. The possibility that an increase in permeability to lipoid soluble-
molecules might appear first was considered. These data are presented
in Table II. Once again there is considerable variability, but the evi-
dence indicates that the experimental treatment increases the permeability

00-

(O-

cr
LJ

C\J-

ooo

0 I 23456

TIME IN MINUTES

FIG. 1. The effect of saponin on the osmotic hemolysis of chicken erythrocytes
in glycerol. O — control; • — experimental. (Scale reading represents degree of
hemolysis.)

to lipoid-soluble and lipoid-insoluble molecules similarly. The time at
which the increase is first noted and the degree of change are the same
for the two types of substances.

Although the increase in permeability to lipoid-soluble and lipoid-
insoluble molecules apparently occurs at the same time, it was thought

SAPONIN AND OSMOTIC HEMOLYSIS

TABLE II

Data comparing the effect of saponin on the osmotic hemolysis of chicken erythro-
cytes in lipoid-soluble and I ipo id-insoluble molecules after a short exposure to the saponin.

Percentage change

Date

Glycerol

Monoacetin

Ethylene glycol

Diacetin

3/6/40

-42.2

-62.2

3/8/40

-13.8

-22.7

3/27/40

- 7.2

-15.2

3/29/40

-18.3

-12.3

3/30/40

-11.1

4.5

+ 0.5

• 2.4

4/6/40

-16.2

-20.9

4/8/40

-18.0

- 9.5

TABLE III

Data comparing the. effect of saponin on the osmotic hemolysis of chicken erythro-
cytes in lipoid-soluble and lipoid-insoluble molecules after a long exposure to the saponin.

Glycerol

Time in seconds for
80% hemolysis

Percentage
change

Time in seconds for
80% hemolysis

Percentage
change

Control

Experimental

Control

Experimental

200

-75.0

205

110

-46.3

215

142

-33.9

205

150

-26.8

175

-86.3

175

-85.2

195

115

-41.0

195

-51.3

195

-51.3

195

-65.1

131

-86.3

175

-88.6

125

-60.0

-41.2

106

-24.5

140

-80.0

165

100

-39.4

110

-13.6

154

152

1.3

130

112

-13.8

255

-76.5

155

-75.5

210

-81.0

130

-30.8

185

140

-24.2

130

-69.2

185

- 70.3

130

110

-15.4

237

157

-33.8

175

-85.7

237

-66.2

205

135

-34.1

237

-89.5

205

-80.5

237

-85.2

205

-88.8

237

135

-43.0

146

-43.8

237

188

-20.7

146

130

-11.0

237

120

-49.4

146

-52.0

215

-83.7

127

-89.8

215

-87.9

127

-89.8

215

130

-39.5

127

-70.9

Average 192

-55.2

157

-57.3

Monoacetin

HUNTER, BARBER AND CAPUTI

possible that after the membrane had become considerably altered a
difference in rate of penetration might appear. The data in Table III
are presented as evidence on this point. These were obtained after the
saponin had completed its action on the cell and had been removed by
centrifugalization. The results indicate that under similar treatment
the permeability to a lipoid-soluble molecule such as monoacetin is in-

TABLE IV

Data comparing the effect of saponin on the rate of penetration of a small and a large

lipoid-insoluble molecule.

Glycerol

Malonamide

Time in seconds for

80% hemolysis

Percentage

80% hemolysis

Percentage

Control

Experimental

Control

Experimental

131

-86.3

670

204

-69.6

125

-60.0

578

126

-78.2

165

100

-39.4

805

644

-20.0

154

152

1.3

729

675

7.4

255

-76.5

1080

180

-83.3

237

157

-33.8

1080

-93.1

237

-66.2

920

540

-41.3

237

-89.4

920

210

-77.2

237

-85.2

920

-91.0

237

135

-43.0

920

110

-88.0

237

120

-49.4

920

150

-83.7

200

110

-45.0

730

390

-46.6

230

110

-52.2

810

330

-59.3

190

-81.6

570

-89.5

180

-63.9

480

-86.5

330

320

• 3.0

1190

1070

-10.1

340

315

7.4

1260

1080

-14.3

330

285

-13.6

630

270

-57.1

330

260

-21.2

690

210

-69.6

125

-64.0

400

120

-70.0

125

-64.0

640

150

-76.6

Average 221

120

-45.7

807

321

-60.2

creased the same amount as the permeability to a lipoid-insoluble mole-
cule such as glycerol.

Finally, a comparison was made of the effect of the experimental
treatment on the penetration of a large molecule such as malonamide
and a small molecule such as glycerol. In this case there appeared to
be a consistent difference, but in order to be certain, an additional series
of experiments was performed. All of the data are included in Table
IV. The last ten sets of readings were from the second series of

SAPONIN AND OSMOTIC HEMOLYSIS

experiments. These data indicate that the rate of entrance of malona-
mide is increased more by the experimental treatment than the rate of
entrance of glycerol.

1.0

0.8 0.6 0.4 0.2
PERCENT NACL

FIG. 2. The effect of saponin on the fragility of chicken erythrocytes. O — con-
trol; • — experimental (Scale reading represents degree of hemolysis.)

A decrease in the time for hemolysis does not necessarily indicate an
increase in permeability (cf. Hunter and Pahigian, 1940). In order to
test for a possible change in cell fragility, the following experiments
were performed. Twenty cu.mm. of blood were added to 10 cc. of
NaCl solutions of concentrations from 1.0 per cent to 0.2 per cent. The

HUNTER, BARBER AND CAPUTI

pH of these solutions was adjusted by the addition of two drops of
NaHCCX to each tube. A typical pair of curves is presented in Fig.
2. These data indicate that there is little, if any, change in the volume
at which the experimental cells hemolyze.

0123
TIME IN MINUTES

FIG. 3. The effect of saponin on the rate of swelling of chicken erythrocytes. O —
control; • — experimental. (Scale reading represents amount of swelling.)

As a final test, swelling experiments were performed. Figure 3 pre-
sents a typical pair of curves obtained when control and experimental
cells were added to an hyperosmotic solution of 0.3 M glycerol in
Ringer-Locke. It can be seen that the experimental cells swell more
rapidly than the controls. These data, then, definitely indicate an in-
crease in permeability resulting from the experimental treatment.

SAPONIN AND OSMOTIC HEMOLYSIS 77

Schulman and Rideal (1937) presented evidence which indicated that
the lytic action of saponin resulted from its reaction with cholesterol in
the cell membrane. This would suggest that the experimental cells in
the present investigation had altered membranes as a result of the action
of saponin on some portion of the membrane in which lipoids were
involved. The fact that the rate of penetration of lipoid-insoluble mole-
cules, as well as that of lipoid-soluble molecules, was increased, would
indicate that lipoid molecules in the membrane were in some way asso-
ciated with the channels through which both types of molecules pass. A
recent series of experiments by Ballentine and Parpart (1940), in which
the effect of lipase on the cell surface was investigated, gave similar
results. By making chemical analyses, these authors concluded that the
lipase split fatty acids from the phospholipids in the cell surface. They
suggested that these phospholipids were " an important structural unit in
determining the rate of penetration in the aqueous channels."

As a result of the experiments in which lipase was used and those
in which saponin was used, there is evidence to indicate that some of the
lipoids in the cell membrane influence the passage not only of lipoid-
soluble molecules but also of lipoid-insoluble molecules. It has been
demonstrated that phospholipids are one type of molecule involved and
the data contained in the present investigation suggest that cholesterol
may be another.

SUMMARY

1. Chicken erythrocytes exposed to low concentrations of saponin
have their membranes altered.

2. By removing the saponin and resuspending the cells in Ringer-
Locke, they will remain unhemolyzed for several days, even though the
membranes have been altered.

3. These cells are more permeable to both lipoid-soluble and lipoid-
insoluble molecules.

4. The penetration of both types of molecules appears to be affected
equally.

5. The rate of penetration of a large molecule such as malonamide is
increased more by this treatment than the rate of penetration of a smaller
molecule such as glycerol.

6. The fragility of these cells is not increased by this treatment.

78 HUNTER, BARBER AND CAPUTI

LITERATURE CITED

BALLENTINE, R., AND A. K. PARPART, 1940. The action of lipase on the red cell

surface. Jour. Cell, and Comp. Physiol., 16: 49-54.
DAVSON, H., AND J. F. DANIELLI, 1938. Studies on the permeability of erythro-

cytes. V. Factors in cation permeability. Biochem. Jour., 32 : 991-1001
HUNTER, F. R., 1936. The effect of lack of oxygen on cell permeability. Jour.

Cell, and Comp. Physiol, 9 : 15-27.
— , AND V. PAHIGIAN, 1940. The effect of temperature on cell permeability and

on cell respiration. Jour. Cell, and Comp. Physiol., 15 : 387-394.
— , L. D. STRINGER, AND H. D. WEISS, 1940. Partial retention of hemoglobin

by chicken erythrocytes. Jour. Cell, and Comp. Physiol., 16 : 123-129.
JACOBS, M. H., AND A. K. PARPART, 1932. Osmotic properties of the erythrocyte.

IV. Is the permeability of the erythrocyte to water decreased by narcotics?

Biol. Bull., 62 : 313-327.
PONDER, E., 1937. Effects of simple haemolysins in hypotonic systems. Proto-

plasma, 27: 523-529.
SCHULMAN, J. H., AND E. K. RiDEAL, 1937. Molecular interaction in monolayers.

II. The action of haemolytic and agglutinating agents on lipo-protein mono-
layers. Proc. Roy. Soc., B, 122 : 46-57.
STRINGER, L. D., H. D. WEISS, AND F. R. HUNTER, 1940. The effect of pH on the

hemolysis of chicken erythrocytes. The Biologist, 21 : 138-139.

THE ROLE OF TISSUES IN THE ANAEROBIC

METABOLISM OF THE MUSSEL ANO-

DONTA HALLENBECKII LEA

SARAH E. CULBRETH
(From the Zoological Laboratories, Duke University, Durham, N. C.)

INTRODUCTION

Lamellibranch mollusks possess the capacity for enduring anaerobio-
sis for a considerable time. Such a capacity is advantageous to tidal
zone forms which are exposed to air at low tide, and likewise to fresh-
water mollusks which may have to endure low oxygen content of pol-
luted water as well as exposure. Recognition of this peculiar ability
has led to investigation of anaerobic metabolic processes of mussels.
If the stream of water passing over the gills is cut off, the oxygen
supply fails while carbon dioxide accumulates. The manner in which
the mollusk deals with accumulating carbon dioxide has been the subject
of several investigations.

Collip (1921) showed that the marine form Alya arcnaria used cal-
cium to buffer carbon dioxide. Dotterweich and Ellsner (1935) showed
that in the freshwater mussel Anodonta cygnca most of the carbon di-
oxide formed during anaerobiosis entered into combination with cal-
cium to form calcium bicarbonate. A small amount was buffered by
calcium proteinate. They concluded that in general calcium in the
shell of mollusks may be utilized as an alkali reserve.

Recent investigation by Dugal and Irving (1937) indicated that
tissues as well as body fluids are involved in adjustment to oxygen lack.
Mantle tissue of Venus nicrccnaria was found to accumulate carbon
dioxide and calcium just as did mantle cavity fluid.

The work reported in the present paper was an investigation of the
adjustment of a freshwater mollusk to a disturbance of the acid-base
balance resulting from anaerobiosis. Particular reference was made to
the role of mantle and gill tissues in this adjustment. Determinations
of the carbon dioxide content gave results which indicated that mussel
tissues were able to buffer carbon dioxide. The relation of calcium to
the buffering process was studied. Observations were made on the
oxygen consumption of tissues taken from asphyxiated animals. Evi-
dence of an oxygen debt was found, showing that dissimilative processes
were continuing through the period of anaerobiosis.

80 SARAH E. CLJLBRETH

MATERIALS AND METHODS

Animals used were freshwater mussels taken in the vicinity of Dur-
ham, N. C. They were identified by Dr. Henry van der Schalie of the
University of Michigan Museum of Zoology as Aiwdonta hallcnbeckii
Lea.

Control animals were kept in tanks of running water. In this situa-
tion the valves remained open, allowing a constant stream of water to
pass over the gills. Experimental animals were removed from such
tanks and placed in a refrigerator with an air temperature of 6 to 8° C.
At this temperature clams survived about a month. When removed
from water Anodonta closed the shell valves. In this position ex-
change of gases between animal and environment was impossible. Any
opening of the shell was accompanied by leakage of fluids from the
mantle cavity. Leaking animals were not included in the experiments.

Tissues used were gill, mantle, and kidney tissue. Some observa-
tions were made on pallial muscle and foot muscle.

The rate of oxygen consumption of gill, mantle, and kidney tissue
was measured in a standard Warburg apparatus. Tissue samples weigh-
ing about 0.1 gram were suspended in a salt solution containing 0.153
per cent NaCl. Absorption of carbon dioxide was accomplished with
20 per cent KOH. The temperature was held at 25° C. Measure-
ments were made over a period of sixty minutes.

Carbon dioxide content of gill and mantle was determined by an
adaptation of the Van Slyke manometric method for the determination
of blood gases. The gas burette of the apparatus was modified from
that described by Ferguson and Irving (1929). A ground joint at
the lower end of the extraction chamber allowed the introduction of
tissue. A weighed sample of tissue was placed in the extraction cham-
ber, the burette put in place, and the joint made secure. Carbon dioxide
was liberated by 0.1 N HC1 introduced through the upper stopcock.
Usually complete extraction required 45 minutes of shaking. Carbon
dioxide was absorbed with air-free 1.5 N NaOH. The values P± and
P2 and the correction factor, c, were determined in the usual way.

Conversion of the observed pressure of carbon dioxide into cubic
centimeters of gas was made according to the formulae modified for use
with tissue samples by Ferguson and Irving (1929). Values for spe-
cific gravity were necessary for the conversion formulae. These values
as determined were: for mantle. 1.04; for gill, 1.12.

Care was taken to maintain constancy in the method of obtaining
and weighing tissue samples. It is felt that the values for carbon di-
oxide content are comparable, although they may not be absolute.

ADJUSTMENT TO ANAEROBTOSIS IN THE MUSSEL 81

Calcium content of clam tissues was determined from samples di-
gested in a mixture of concentrated nitric and perchloric acids. Cal-
cium was precipitated from the digest as oxalate, redissolved and ti-
trated with permanganate.

RESULTS
Oxygen Consumption

Respiration of tissues from aerohic and anaerobic animals was com-
pared. The results are given in Table I. The values for control ani-
mals are based on four determinations. They agreed closely. There
were wider differences in the determinations on tissues from 'anaerobic
animals. These values have been arranged by length of anaerobic pe-
riod, and also averaged into one value for asphyxiated animals.

TABLE I

Oxygen uptake of gill, mantle, and kidney tissue. Values are averages of two to
four determinations and represent cubic millimeters of oxygen consumed per hour per
milligram dry weight of tissue at 25° C.

Days out
of water

Kidney

Mantle

Gill

4.70

1.32

.486

3.66

1.34

.766

2.69

1.24

.652

2.82

1.45

.758

3.24

1.02

.660

2.82

.573

Average
Average of controls

3.32
2.11

1.27
1.02

.649
.421

Tissues removed from asphyxiated animals consumed more oxygen
per hour per unit weight than did tissues from control animals. This
was true for the first hour after removal. Determinations were not
carried beyond this point. It is therefore impossible to make any cal-
culation of the total extra oxygen required. However, the increase
noted suggests the paying off of an oxygen debt incurred during anaero-
biosis.

The respiration rates are referred to dry weight of tissue. It was
found that mussel tissues varied in water content from one individual
to another. There was no evidence of a correlation between dry weight
and anaerobic period. The observed percentages dry weight as aver-
aged from a large number of samples studied are given below :

mantle 3.9

kidney 8.6

gill 24.0

SARAH E. CULBRETH

It is interesting that the rate of oxygen consumption of kidney
tissue was much higher than that of other tissues studied. According
to Holmes (1937), the high rate of respiration of mammalian kidney
tissue is due to osmotic work done by excretory cells. Probably a
similar explanation fits the case of mussel kidney.

Carbon Dioxide Content

Results of the determination of the carbon dioxide content of gill
and mantle are given in Table II. The following points are to be noted :

TABLE II

Carbon dioxide content of mantle and gill. Values are expressed as cubic centi-
meters of gas at standard temperature and pressure and equivalents of carbon dioxide
in one hundred grams fresh tissue. Averages of several determinations are represented.

Days out
of water

Mantle

Gill

Mantle

Gill

cc./WO gr.

eqtiiv./lOO gr.

equiv./WO gr.

25.0

322

0.0022

0.0287

399

0.0356

32.2

372

0.0028

0.0332

35.0

369

0.0030

0.0328

32.0

405

0.0028

0.0376

34.4

430

0.0030

0.0392

43.6

455

0.0038

0.0406

44.2

487

0.0038

0.0432

499

0.0444

47.7

512

0.0042

0.0456

1. Gills contained approximately ten times as much carbon dioxide
as did mantles.

2. There was a steady increase in the amount of carbon dioxide
accumulating in gill tissue during anaerobiosis.

3. Carbon dioxide accumulated in mantle tissue in proportion to the
increase in gill tissue. The equivalents of carbon dioxide in mantle
doubled during asphyxiation.

For purposes of comparison with the amount of calcium present,
the values for carbon dioxide were converted into equivalents and are
also given in Table II.

Calcium Content

It was found that the calcium content of the tissues 'studied did not
vary significantly writh the period of anaerobiosis. Averages from a
large number of determinations are given below, expressed as milli-
grams of calcium per gram dry weight of tissue.

ADJUSTMENT TO ANAEROBIOSIS IN THE MUSSEL

foot muscle 8 mg./gram tissue

pallia! muscle 31

kidney 46

mantle 62

gill 175

Gill tissue contained a large amount of calcium as compared with
other tissues. This may be correlated with the relatively high dry
weight of gill tissue. The small amount of calcium found in foot
muscle is surprising when considered with the other values.

By using the percentage dry weight of mantle and gill tissue it was
possible to calculate equivalents of calcium per one hundred grams fresh
tissue. These were found to be: for mantle, 0.0045; for gill, 0.21.

DISCUSSION

Study of the functioning of animal tissue in buffering processes has
not been investigated in many species. Dottenveich (1933) showed
that the calciferous glands of earthworms were capable of giving up
calcium to buffer carbon dioxide accumulating in body fluids. Banus
and Katz (1927) found weak buffering by hind leg muscles of a cat.
A similar effect wras noted by Irving and Chute (1932) in muscle.

A buffer system in the tissues of Anodonta is indicated by a study
of the carbon dioxide and calcium content of certain tissues. Gill tissue
seems to be most active in this respect.

From the data given above, it is seen that one hundred grams fresh
gill tissue contain 0.21 equivalents of calcium, and 0.0287 equivalents of
carbon dioxide (see Table II). This proportion indicates that most,
possibly all, the calcium is present in some form other than carbonate.

During anaerobiosis the carbon dioxide level rises, increasing to
0.0456 equivalents at 14 days. This increase is not accompanied by an
increase in the hydrogen ion concentration. The hydrogen ion concen-
tration of the tissue was measured colorimetrically, and was found to
vary less than 0.05 from pH 6.8 for gill, 6.9 for mantle. Apparently the
accumulating carbon dioxide is bound in some w-ay so that an increase
in hydrogen ions does not occur.

It was suggested by Dotterweich and Ellsner (1935) that a calcium-
proteinate might act as an additional buffer in the fluid of Anodonta
cygnea. In that system calcium carbonate was the principal alkali re-
serve. In the tissues of Anodonta hallcnbcckii it would seem that cal-
cium-proteinate, or some other combination of a weak acid writh calcium,
is the chief buffer, with the carbonate playing at the most a minor role.

84 SARAH E. CULBRETH

fn the case of mantle tissue 0.0045 equivalents of calcium are present
in the normal mantle. Carbon dioxide increases from 0.0022 equivalents
in the normal tissue to 0.0042 equivalents in the asphyxiated tissue. The
calcium and carbon dioxide are then in a one-to-one ratio. This would
indicate a more limited calcium reserve in mantle than in gill.

Dugal (1939) has shown that in Venus the calcium reserve may be
augmented by calcium from the shell. Tissues of Anodonta maintain
a steady calcium level.

Calcium is not only the chief component of the hard parts of
mollusks but also forms a considerable portion of the alkali reserve.
The same factors which govern the precipitation of solid calcium in the
shell are responsible for the deposition of calcium in tissues. It is a
point of interest that freshwater clams possess large deposits of calcium
in their gills, and marine clams possess the larger deposits in mantle
tissues (McCance and Shipp, 1933). There may be some correlation
here with the fact that glochidia develop in the gill pouches of fresh-
water mussels and may derive calcium for their shells from the abundant
supply available.

Jatzenko (1928) showed that certain freshwater mussels build up an
oxygen debt during anaerobiosis. It is to be expected that individual
tissues would also show such a debt. All activity does not cease when
the clam is temporarily asphyxiated. Some of it continues. Ciliary
action such as accounts for a great deal of the oxygen consumption of
gill and mantle probably does decrease to some extent. Osmotic work
which is characteristic of kidney tissue continues and may even increase
during anaerobiosis. Data for individual tissues as presented in Table
I show that oxygen consumption of mussel tissues is higher immediately
after a period of asphyxiation than under normal conditions.

The source of energy for activities carried on during anaerobic
periods is generally laid to a glycolytic process. However, there has as
yet been no isolation of the tissue or tissues mainly responsible for the
glycogen reserve. The problem of the energy source and its localiza-
tion is a pertinent one to a complete explanation of the anaerobic
metabolism of mussels.

SUMMARY

Tissues of Anodonta hallcnbeckii are capable of buffering carbon
dioxide accumulating during anaerobiosis. Calcium compounds present
in gill and mantle serve as an alkali reserve. During anaerobiosis carbon
dioxide increased in the tissues studied while the hydrogen ion concen-
tration remained constant. It is concluded that accumulated carbon
dioxide was buffered by calcium present.

ADJUSTMENT TO ANAEROBIOSTS IN THE MUSSEL

Gills contain large amounts of calcium which is present in some
form other than carbonate.

Kidney tissue showed a very high rate of respiration. Mantle and
gill showed low rates. After anaerobic periods the rate of respiration
showed a tendency to increase. This may be taken as evidence that
these tissues continued to do work during anaerobiosis.

LITERATURE CITED

BANUS, M. G., AND L. N. KATZ, 1927. Observations of the role of tissues in
maintaining the acid-base equilibrium of the blood. Am. Jour. PhysioL,
81 : 628-649.

COLLIP, J. B., 1921. A further study of the respiratory processes in Mya arenaria
and other marine molluscs. Jour. Biol. Clicin., 49: 297-310.

DOTTERWEICH, H., 1933. Die Function tierisches Kalkablagerungen als Puffer-
reserve im Dienste der Reaktionsregulation. Arch. f. die ges. PhysioL,
232 : 263-286.

DOTTERWEICH, H., AND E. ELLSNER, 1935. Die Mobilisierung des Schalenkalkes
fiir die Reakstionsregulation der Muscheln (Anodonta cygnea). Biol. ZbL.
55: 138-163.

DUBUISSON, M., AND Y. VAN HEUVERSWYN, 1931. Recherclies histologiques et
chimiques sur les branchies d' Anodonta cygnea. Arch, dc Biol., 41 : 37-74.

DUGAL, L.-P., 1939. The use of calcareous shell to buffer the product of anaerobic
glycolysis in Venus mercenaria. Jour. Cell, and Comp. PhysioL, 13 : 235-
251.

DUGAL, L.-P. AND L. IRVING, 1937. Secretion de carbonate de calcium par les
Venus mercenaria f ermees hermetiquement. Conipt. Rend Soc. Biol., 124 :
526-528.

FERGUSON, J. K. W., AND L. IRVING, 1929. A method to determine the CO-, con-
tent of muscle. Jour. Biol. Chem.. 84: 143-153.

HOLMES, ERIC, 1937. The metabolism of living tissues. Cambridge University
Press, pp. x-235.

IRVING, L., AND A. L. CHUTE, 1932. The participation of the carbonates of bone
in the neutralization of ingested acid. Jour. Cell, and Comp. Ph\siol., 2:
157-176.

JATZENKO, A. T., 1928. Die Bedeutung der Mantelhohlenfliissigkeit in der Biologic
der Siisswasserlamellibranchier. Biol. ZbL, 48: 1-25.

McCANCE, R. A., AND H. L. SHIPP, 1933. The magnesium and other inorganic
constituents of some marine invertebrates. Jour. Mar. Biol. Ass. Ply-
mouth, 19: 293-296.

THE EFFECT OF THE CIRCULATION OF WATER ON THE

DISTRIBUTION OF THE CALANOID COMMUNITY

IN THE GULF OF MAINE 1

ALFRED C. REDFIELD

(From the Biological Laboratories, Harvard University, and the Woods Hole
Oceanographic Institution, Woods Hole, Mass.}

Damas (1905) has pointed out that the flow of water tends to dis-
sipate local populations of pelagic organisms, and that the permanence
of breeding stocks may 'be maintained by the existence of eddies. His
predictions have been strikingly confirmed by hydrographic observations
in the Norwegian Sea (S0mme. 1933). Along the margins of the Gulf
of Maine the permanence of the stock of Sagitta clcgans is correlated
with the stability of the hydrographic conditions which exist in different
regions (Redfield and Beale, 1940). Walforcl (1938) has indicated
the importance of fluctuations in the circulation on Georges Bank to the
fate of haddock eggs spawned in that region. These studies and that
of the author (1939) on the population of Limacina rctrovcrsa empha-
size the rapidity with which currents move pelagic organisms about
within the Gulf. It becomes a problem whether the community of the
basin of the Gulf is truly endemic, and by what mechanism a breeding
stock is maintained within the Gulf. S0mme (1934) has discussed this
question in regard to the copepod population of the Lofoten area.

Bigelow (1926), who has described the zooplankton of the Gulf in
great detail, considers that the species which form the bulk of the
pelagic population are endemic in origin, breeding with sufficient regu-
larity and abundance to maintain the local stock by local reproduction.
From its dominating member, Calanus finmarchicus, he has referred to
the population as the calanoid community.

We have measured the catches taken in the Gulf during a year-
round survey and will attempt to explain the distribution of numerical
abundance in terms of the pattern of currents obtained during the period
of observation.

DATA

The data employed in the present study were collected in the course
of cruises made by the research vessel " Atlantis " during the years of
1933 and 1934. The dates of these cruises and the numbers of the

1 Contribution No. 281 from the Woods Hole Oceanographic Institution.

CIRCULATION AND DISTRIBUTION

stations occupied are given in Table I. Thirteen cruises were made in
the course of fifteen months, with the result that 684 hydrographic
stations in the Gulf of Maine and its adjacent waters were occupied. At
no time did a period longer than two months elapse without observation.
The routine hydrographic and chemical data are published in the Bulletin
Hydrographique (1933, 1934).

A supplementary cruise was made in May, 1936 in order to confirm
certain observations made during the primary survey.

In the course of the cruises standard vertical hauls were made with
a 1.5 meter Heligoland net of No. 0 silk having 38 meshes to the inch
(Kiinne, 1933). The net was hauled from a point near the bottom to
the surface at all stations occupied, weather permitting. This type of
haul was selected in preference to the oblique haul in the belief that the
procedure could be carried out uniformly as a part of the routine duties

TABLE I

Cruise No.

Dates

Stations

Number

16 and 17

June 19-July 10, 1933

1643-1721

Sept. 2-Sept. 14, 1933

1741-1802

Oct. 17-Oct. 29, 1933

1803 1860

Dec. 2-Dec. 11, 1933

1861-1906

Jan. 8-Jan. 13, 1934

1907-1934

Mar. 21-Mar. 29, 1934

2019-2070

27 and 28

Apr. 17-Mav 13, 1934

2071-2164

May 21-June 3, 1934

2165-2215

June 25-July 1, 1934

2217-2236

Aug. 10-Aug. 11, 1934

2252-2259

Sept. 17-Sept. 27, 1934

2268-2303

May 14-May 19, 1936

2555-2583

of the ship's company. Unfortunately, it proved impossible to use the
net in rough weather, so that data are lacking from many stations, par-
ticularly those made in the winter months. The yields of the successful
hauls have been measured by collecting the plankton on filter paper in a
Buchner funnel. Suction was continued until the preserving fluid
ceased to flow, whereupon the " dry " plankton was introduced into a
measured volume of fluid and the resulting increase in volume noted.
The data so obtained were reduced to figures expressing the number of
cubic centimeters of " dry " plankton under each square meter of the
sea surface. Before filtering and measuring collections, any large gelat-
inous organisms were removed (Salpa, ctenophores, medusae) with the
result that the measurements reflect primarily the abundance of the
crustacean community.

Since most of the hauls with which we are concerned were made in

A. C. REDFIELD

FIG. 1. Volumes of zooplankton taken in vertical hauls between September,
1933 and June, 1934. Numbers represent the cubic centimeters taken per square
meter of sea surface. Contour interval 25, 50, 100, and 200 cc. per square meter.

CIRCULATION AND DISTRIBUTION

depths greater than 100 meters, above which level most of the population
may be expected to occur, these figures are thought to express the density
of population more precisely than numbers reduced to unit volume of
water strained. The general character of the results is not altered by
expressing the catch in terms of the yield per cubic meter.

The determination of the " dry " volume of the catch by the method
of filtration and displacement yields smaller values than are obtained by
the '" wet " method of allowing the animals to settle in a calibrated
container. In order that our results may be compared with those of
Bigelow and others who employed the settling method, a number of
samples have been measured by both methods. The wet method gave
values on the average 4.9 times higher than the dry method, the ratios

FIG. 2. Volumes of zooplankton taken in vertical hauls in September, 1934
and May, 1936. Numbers represent the cubic centimeters taken per square meter
of sea surface. Contour interval 50, 100, 200 cc. per square meter.

varying between 3.3 and 7. The ratio was smaller in the case of the
larger samples, due perhaps to the tighter packing of large samples in
the wet method and to the greater retention of water when large quan-
tities of organisms are filtered off in the " dry " method.

THE SEASONAL DISTRIBUTION OF THE ZOOPLANKTON POPULATION

The quantities of the catches obtained by vertical hauls, and their
positions during the most complete periods of survey, are entered on
the charts shown in Figs. 1 and 2. These charts show that the area
of maximum abundance shifts its position with the season. From late
summer until December the richest population is found in the northern
portion of the Gulf, centering off Mount Desert. During the winter
the center shifts to the west, coming to lie off the Massachusetts coast.

A. C. REDFIELD

In late spring and early summer the richest catches were obtained along
the southern margin of the Gulf, north of Georges Bank, extending
from the offing of Cape Cod, eastward and northward toward the Bay
of Fundy.

FIG. 3. Chart of the Gulf of Maine showing principal place names and the
sectors into which the area is divided for analysis of population distribution. Con-
tour encloses depths less than 100 meters.

In order to deal with the data statistically, the area of the Gulf has
been divided into seven sectors as shown in Fig. 3. Each sector includes
one of the principal lines of stations at which collections were regu-
larly made. The quantities of plankton taken at each cruise in each
sector have been averaged and the resulting number taken to represent
the density of population in that sector at the time. While the data are
frequently numerically inadequate, certain interesting regularities appear
from its analysis.

CIRCULATION AND DISTRIBUTION

Figure 4 shows the density of population in the various sectors at
each principal period of survey. It presents graphically the shift in the
center of population westward from the Mount Desert to the Massa-

FIG. 4. The average catch in each sector of the Gulf of Maine during the
period September, 1933 to September, 1934. Ordinates : cubic centimeters of
zooplankton per square meter of sea surface.

A. C. REDFIELD

chusetts sector in the course of the winter and its extension along the
southern sectors in May and June, followed by the reestablishment of a
maximal population in the northeastern sectors in September.

Figure 5 presents the same data in a form which brings out the sea-

1933

1934

VI I VII I VIII I IX I X I IX I XII I I II I III I IV | V I VI | VII I VIII I IX

BROWNS

SE6UIN

100
50

FIG. 5. Seasonal fluctuation of catch in each sector of the Gulf of Maine
during the period June, 1933 to September, 1934. Ordinates : cubic centimeters of
zooplankton per square meter of sea surface. The black bars indicate the actual
period occupied by each cruise.

MASSACHUSETTS

CULTIVATOR

CIRCULATION AND DISTRIBUTION

sonal fluctuation in each sector. The sectors along the east and north
sides of the Gulf are marked hy a pronounced seasonal fluctuation — most
extreme in the Mount Desert region. In contrast, the population is
much more uniformly distributed from month to month in the Massa-
chusetts and Cultivator sectors, which include the greater part of the
western basin.

The general features of the distribution appear to recur from year
to year, for our observations for September 1933 and 1934 show essen-
tially similar patterns, as do also those for May-June, 1934 when com-
pared with May, 1936. The seasonal distribution observed by Bigelow
over a number of years is also in agreement. He found the quantitative
fluctuations to be comparatively narrow from season to season in the
waters of the western basin and considered the plankton in that part
of the Gulf to be " rich " the year round. He reports the northern cor-
ner of the eastern basin, as well as the shallows off Cape Sable, to be
the site of a wide seasonal fluctuation (Bigelow, 1926, p. 89).

THE CIRCULATION OF THE GULF

The shift in the center of abundance of the zooplankton population
suggests that it is being borne about a great cyclonic eddy. We may
consequently examine the nature of the circulation of the Gulf to see
if it can account for the fluctuations in numbers in different places and
to learn to what extent the calanoid community may be carried into
and out of the Gulf by water movements.

The evidence marshalled by Bigelow (1927) — measurements with
current meters, drift-bottles, temperatures, salinities, distribution of
plankton, and dynamic calculations — can be harmonized with one type
of dominant circulation only, a general anti-clockwise eddy around the
basin of the Gulf. The demonstration of this, named by Huntsman
(1924) and by Bigelow the " Maine" or " Gulf of Maine" eddy, with
all it implies in its biological bearing, is perhaps the most interesting
result of their joint explorations of the Gulf. Observations made during
a series of years demonstrated that the center of the eddy shifted its
precise location from summer to summer, and that marked seasonal
variations in the circulatory scheme occurred. Observations of the
velocity of the non-tidal drift of the surface made in shoal water about
the margin of the Gulf indicated an average movement of seven miles
per day, at which rate some three months would be required to complete
the circuit of the eddy. No estimations were made of the velocity of
the deeper layers.

The hydrographic data collected in the course of the cruises in
1933-34 have been analyzed by Dr. E. E. Watson, who has kindly per-

A. C. REDFIELD

FIG. 6. Dynamic contour charts showing the theoretical circulation of the
Gulf of Maine at the surface between September, 1933 and September, 1934. The
heavy contours, taken from Figs. 1 and 2, indicate the relative density of population
at the time of each cruise.

CIRCULATION AND DISTRIBUTION

APRIL

LARGE SPECIMENS

FIG. 7. The distribution of Limacina rctrovcrsa in the Gulf of Maine between
December, 1933 and September, 1934. The numerals indicate the position and the
numbers caught per haul. Compare Fig. 6 for corresponding current diagrams
and the coincident distribution of the calanoid community.

96 A. C. REDFIELD

mitted me to use some of his current diagrams in advance of the publica-
tion of his full report. This study has not only confirmed the more
general conclusions of Bigelow, but gives the best available evidence of
the actual character of the circulation at the time of our collections.

The zooplankton population is not distributed uniformly in waters
of various depth but tends on the whole to congregate in the upper 100
meters (Bigelow, 1926, p. 93). In many species, particularly of the
numerically important copepods, there is a pronounced diurnal vertical
migration which has been studied in the Gulf of Maine by Clarke ( 1933,
1934). At a station in the deep part of the Gulf, he found that Calanns
and Metridia migrated to a depth of 120 meters or more during the
daytime and moved upward to levels of from 6 to 42 meters at night.
On Georges Bank Calanns was confined to the surface strata, under-
going very limited migration, but Metridia carried out an extensive ver-
tical migration. Consequently the population cannot be identified exclu-
sively with any particular layer and any attempt to correlate its distri-
bution with the drift of the water is complicated by the undoubted mi-
gration of the animals to and from layers of different depth moving
with different velocities and in some places without doubt in different
directions. As a first approximation, however, it is reasonable to assume
that considerable volumes of the more superficial water, unconfined in
its movements by shoals, will retain for appreciable periods a unity suf-
ficient to permit a definite population to be identified with it. The
horizontal movement of the water at a depth of 40 meters should be
fairly representative of the layers in which the zooplankton chiefly occur
within the Gulf. At this depth the water is unobstructed in its move-
ments by any considerabe shoals. We have reproduced the dynamic
contours at a depth of 40 meters in a recent paper (Redfield and Beale,
1940, Fig. 10). The surface circulation does not differ in important
detail from the charts representing conditions at 40 meters. It shows
a somewhat closer correlation with certain features of the plankton dis-
tribution. Charts showing the gradient currents at the surface have
consequently been employed in preparing Fig. 6.

Additional evidence of the character of the circulation, and particu-
larly of its influence in actually transporting a pelagic population, is
provided by the distribution of Limacina rctrorcrsa in the Gulf during
the period of this survey (Redfield, 1939). These organisms appeared
en masse in the Browns sector in December and their drift was fol-
lowed as they spread across the Gulf. In four months they had ar-
rived in numbers in the western basin, having spread along the northern
margin of the Gulf in the course taken by the receding center of the
zooplankton population (Fig. 7).

CIRCULATION AND DISTRIBUTION 97

THE RELATION OF POPULATION DISTRIBUTION TO THE HYDROGRAPHY

OF THE GULF

The following theory is proposed to account for the seasonal fluctua-
tion of the population of zooplankton. The superficial current, or non-
tidal drift, consists of a great cyclonic eddy. The eddy is augmented
by the inflow of water on the eastern side from over the Nova Scotian
Banks. The inflow is compensated for by the escape of water to the
south and east across the end of Georges Bank. The relative volumes
of inflow and outflow vary from season to season and year to year. In
the winter and early spring the inflow is sufficiently great to replace
a considerable part of the eddy with water new to the Gulf. This
" new '' water is relatively barren and does not develop a more con-
siderable population until conditions become favorable for growth and
reproduction in the spring, by which time it has extended over the entire
northern half of the eddy. Meanwhile an equal part of the older water,
which had been in the Gulf during the preceding summer, escapes from
the Gulf. The remainder occupies the southern half of the eddy. This
water supports a rich population grown up during the previous summer
and only moderately diminished by the conditions of the winter. In
spring and summer the inflow and outflow diminish and the southern
half of the eddy carries an increasing quantity of water northeasterly
toward the Bay of Fundy, with the result that this water enters a second
circuit of the Gulf, carrying with it a large population which enriches
the northern half of the eddy during the late summer and fall. This
region may also be enriched by an inflow of water from the Nova Scotia
banks which carries a considerable population at some seasons.

The adequacy of this theory is demonstrated in Fig. 6, in which con-
tours representing the areas of relative abundance recorded in Figs. 1
and 2 are transposed upon diagrams of the dynamic gradients in the
surface waters. The general distribution of these contours was estab-
lished without reference to the current diagrams.

In September, 1933, the center of population lay in relatively quies-
cent water along the northern margin of the Gulf and extended south-
westward to occupy a secondary eddy over the western basin. A more
scanty population occupied the Cultivator sector. Over the Nova Sco-
tian Banks a condition of slack water existed with no evidence of in-
draft except along the eastern margin of the Eastern Channel. This
water was scantily populated. A marked eddy occupies the Eastern
Channel with its offshore component lying on the western side. This
includes a tongue of richly populated water in which much plankton is
being carried out of the Gulf. During this period the population is
being impoverished by this outdraft.

98 A. C. REDFIELD

By December a strung indraft of water over the Xova Scotian Banks
has commenced carrying into the Gulf water containing a scanty popu-
lation. This water contains abundant Limacina which occupy the east-
ern area in which catches of less than 25 cc. of plankton occurred.
(Figure 7.) Compensatory movements must be expelling richly popu-
lated waters over Georges Bank.

By March scantily populated water has extended along the entire
northern margin of the Gulf, carrying with it the population of Liina-
cina. Considerable volumes of zooplankton were then taken only in
the southwestern quarter where relatively slack water is found. A small
eddy occupies the western basin and in it Limacina mingle with consid-
erable remnants of the copepod population. In the eastern half of the
Gulf the major eddy is well marked. Along its eastern side water still
enters the Gulf scantily populated with copepods and now containing
very few Limacina. Its western arm is carrying many Limacina out
to sea.

In April the character of the circulation changes abruptly from a
loop to a closed eddy. Invasion of water from offshore has come prac-
tically to an end and considerable numbers of copepods which occur at
the mouth of the eastern channel have no opportunity of entering the
Gulf. In the west a small concentration of copepods persists near the
South Channel. The Limacina population is now centered over the
western basin, but considerable numbers appear to have spread eastward
along the southern and eastern arcs of the great eddy.

Up to this time the movements of water and of the populations of
copepods and Limacina appear to be perfectly correlated. There can be
no doubt that the inflow of barren water from the east has displaced a
large part of the copepod population from the northern and eastern part
of the Gulf, forcing it out to sea over the eastern end of Georges Bank.

In May and June the loop-like character of the circulation reestab-
lishes itself, but a considerable eddy persists in the center of the loop.
Reproduction now increases the population everywhere. The copepod
population occurs in greatest numbers in the slack water of the western
basin. From there a rich band extends along an east-flowing current
out to sea over the Eastern Channel. A part of this eastward extension
has evidently been caught in the recurrent eddy and carried northward
toward Mount Desert. Richly populated water found in the North
Channel appears to be moving out of the Gulf. The only water entering
the Gulf at this time lies along the east side of the Eastern Channel and
appears to be scantily populated. It seems improbable that a consider-
able population is being recruited at this time. The increasing numbers

CIRCULATION AND DISTRIBUTION

observed in the eastern region apparently come from the southern and
western region.

The distribution of Liinacina in May-June agrees with this inter-
pretation of the water movements. The population of large specimens,
which had wintered in the Gulf, extends eastward along the southern
side of the Gulf, and northward along the eastern side of the eddy.
The small specimens, new to the Gulf, lie along the inflow and about the
center of the eddy. Others follow the eastern arm of the inflow which
recurves along the Nova Scotia shore.

By September, 1934 conditions have reestablished themselves much as
they were a year before. The southern half of the Gulf appears to be
occupied by the more scanty population, presumably derived from the
barren water which lay to the north in the spring. This is trapped in
a dead water. A large population has grown up in the eddy which
forms each summer in the northeast quarter and well-marked currents
exist to carry this population to the southeast. The current flow into
the Gulf is stronger than the year before and appears to bear an abun-
dant population from offshore into the eastern side of the Gulf. This is
the only indication that the copepod population is enriched by exchanges
with offshore waters in the course of the year.

The small numbers of Liinacina which occupied the Gulf in Sep-
tember, 1934 occurred in greatest numbers along the course of the inflow-
ing water.

It is unfortunate that a more complete survey was not made between
June and September. The events which are least clear are those leading
to the development of the exceptional populations in the Mount Desert
region in late summer. It is not certain that these may not have been
recruited from offshore during the summer. The circulation calculated
for May-June would appear to transport the richer water then found
in the south out to sea more effectively than toward the northeast. Pos-
sibly the rich population extending northward toward Mount Desert
arrived there before the loop-like eddy reestablished itself. There are.
however, several considerations which support the view that the popula-
tion of the northeastern sectors is recruited from the southern part of
the Gulf in early summer.

Fish (1936) records the invasion of the coastal waters of Maine by
Calanns finmarchicus larvae in June which he assigns to a " western
stock." These he believes to be absent from the eastern half of the
outer Gulf earlier in the season, and to have drifted in toward the Maine
coast from the southwest. The larvae of this stock greatly outnumber
those of an " eastern stock " which entered the Gulf from the Scotian
Banks in April.

100 A. C. REDFIELD

Drift-bottle observations indicate an actual movement of water from
the southern toward the northeastern quarter of the Gulf in summer
(Bigelow, 1927). There can be no doubt that the surface water does
move in this direction, dynamic calculations notwithstanding.'

It is possible that lateral mixing along surfaces of equal density may
permit rather extensive exchanges of water across gradient currents
(Iselin, 1939). In particular, according to a principle developed by Parr
(1936), stratification in turbulent waters leads to increased lateral mix-
ing. Thermal stratification in the Gulf of Maine was well developed
in May, 1934, and its onset may have facilitated the transfer of well-
populated waters across the eddy. It is noteworthy that during May a
considerable intermingling of Liinacina with water rich in copepods
occurred along the southern side of the Gulf. It is also noteworthy that
the distribution of Limacina became much more homogeneous in May
than it had been earlier. This was true also of the plankton population
as a whole, as Fig. 4 shows. Lateral mixing deserves more study by
biologists, as Iselin lias pointed out, for it may well be an important
factor in preventing local breeding stocks from being swept out of
embayments by directional currents.

In summary, the hydrographical evidence appears to support the
view that the scanty population of the eastern sectors in midwinter is
due to their occupancy by the barren water, which appears in the Browns
Bank sector in December and can be traced until it enters the Massa-
chusetts sector by May. The eastern and northern sectors thus receive
an influx of relatively barren water in midwinter when the climate is
unfavorable for further growth of the population. In the early summer,
on the other hand, water drifts from the southern part of the Gulf into
the northeastern sectors. This water 'is about to commence its second
circuit of the Gulf and carries with it a population which has already
grown to some magnitude in the sectors from Massachusetts to Georges
Bank by the end of May.

In contrast to this, the sectors of the western basin receive in winter
water which had acquired an abundant population in the Mount Desert
and Seguin sectors during the fall. Although there is some destruction
of the organisms at this season, it is not sufficient to reduce the numbers
greatly. With the coming of spring, this water moves on to be replaced
by the barren water found in the northern sectors during the winter.
But as this water warms, its population grows and rapidly conies to equal
that of the water it replaces. The uniformity of the population in the
western basin is due to the fact that a rich fauna arrives there coincident
with unfavorable conditions in winter, and a scanty fauna comes to
occupy the region during the period most favorable for growth.

CIRCULATION AND DISTRIBUTION

101

THE AVERAGE MONTHLY CATCH

The monthly catch obtained by averaging the mean values for all
sectors during each cruise is given in Fig. 8. The average monthly
catch remained constant at about 40 cc. per square meter from June to
December, 1933. The values fall markedly from January through April.
This is undoubtedly due to the destruction of the population by winter
cold. At this period, however, large quantities of water poor in popu-
lation are entering the Gulf and an equivalent quantity bearing a richer
population is leaving, thus accounting for much of the loss.

1933

1934

VI | VII | VIII I IX | X I XI

XII | I | II I III | IV I V | VI | VII I VIII | IX

100
80 h
60
40
20

FIG. 8. Average catch for entire Gulf during the period of survey. Ordinate:
cubic centimeters of zooplankton per square meter of sea surface. The black bars
indicate the actual period occupied by each cruise.

In May the average population begins to increase suddenly and its
growth continues at a diminishing rate until the last observations in
September, 1934. The population as a whole is then twice as great as
that encountered during the preceding year, — the average haul being 90
cc. per square meter.2

2 Bigelow found that the zooplankton was at its lowest ebb in late February
and the first half of March in 1920. At this time his catches varied from 75 to 25
cc. per square meter measured wet (equivalent to 15 to 5 cc. measured "dry").
Our catches in March averaged 13 cc. and for April 8 cc. per square meter meas-
ured "dry." The 1934 year appeared to be a late year and was initiated by an
unusually cold winter. The April observations are perhaps misleading since col-
lections were not made in the parts of the Georges and Cultivator sectors where
the highest population was expected. Bigelow considered 100 cc. wet (20 cc. or
more measured dry) to be representative of the Gulf in midsummer, a value much
smaller than our average of 40 cc. in 1933 and 80 cc. in 1934. His largest catch
of 425 cc. (85 cc. "dry"), made in September, 1915, does not exceed the average
value obtained in September, 1934 for all sectors. It must be remembered that he
employed a different method of measuring his catch and nets which doubtless dif-
fered from ours in efficiency.

102 A. C. REDFIELD

THE TOTAL ANNUAL PRODUCTION AND EXCHANGE

The average catch throughout the period of the survey in each sector,
—obtained by averaging the mean figures obtained at the time of each
cruise — is given in Fig. 9.3 It is noteworthy that the catch increases
progressively as one passes along the course of the water movement
from its inflow over the Browns Bank sector to its exit across the end
of Georges Bank. The longer the water has been in the Gulf, the
greater its population.

The average catch is greatest in the Mount Desert sector and is an
exception to the foregoing tendency. We suggest that this is due to
the movement of water, which has already completed the circuit of the
Gulf and is rich in plankton, from the southern sectors into the Mount
Desert sector during the summer. The Mount Desert sector thus virtu-
ally occupies a position at the end of the series during the months when
its population is greatest.

The average haul for all sectors of the Gulf and at all cruises is
about 40 cc. per square meter of "dry" plankton. If the area of the
Gulf be taken at 36,000 square miles, this would indicate a total popu-
lation of about 3.7 X 1012 cc. or some four million tons.4

The standing crop does not give a measure of the rate of production
of the population, since it reflects merely a balance between rate of
growth and death and the gains and losses in the population by move-
ment into or out of the region. It is clear from Fig. 8 that the crop
increased by nearly 80 cc. per square meter of surface between May
and September. This represents a net gain of some eight million tons.

The observations make it apparent that the Gulf loses at least one
half of its population through the escape of water over Georges Bank
and the Eastern Channel each winter. There is no evidence that water
enters the Gulf at any time carrying a richer population than that ob-
taining there at the time, and only in September, 1934 was water found
to enter the Gulf in which the copepod population was not distinctly

3 This method of averaging avoids overweighting the yield of these sectors in
which an unusual number of rich hauls were made at a time when the population
was particularly large, as would be the case if all catches were simply averaged.

4 In discussing the productivity of the Gulf of Maine, Bigelow concluded that
the population was greatest over a band extending from the Massachusetts coast
to Penobscot Bay and the Bay of Fundy. The areas occupied approximately by
our Seguin sector and the greater part of the Cultivator and Georges Bank sectors
were considered barren. His conclusion that the southern sectors are barren cer-
tainly rests on insufficient evidence, since only two hauls are recorded. Our richest
haul was obtained over the southeastern deep in June, 1934 and the Cultivator and
Georges sectors are the most populous sectors excepting Mount Desert throughout
the year as indicated. Our observations agree that the Seguin sector is relatively
unproductive. Its best season was the fall of the year, a time when Bigelow made
relatively few cruises. Several of our richest hauls were made off Seguin in
September, 1934.

CIRCULATION AND DISTRIBUTION

103

scanty. Unless further observations during the summer should prove
the contrary, it may be concluded that the Gulf is a region of production
for the calanoid community which supplies immigrants to the southern
banks in quantity, but receives relatively unimportant recruitments from
the regions to the eastward.

THE GROWTH OF THE POPULATION IN THE MOVING MASS OF WATER

If the interpretation which we put on the data is correct, it is certain
that observations made at a geographically fixed point, or standard
station, tell little about the fluctuations of any unit of the population.
We record simply a series of events distributed in space as they drift
past in the course of time. The conditions we observe today are not
determined by the events we observed yesterday. The curves of popula-
tion growth presented in Fig. 5 are grossly misleading if they are inter-
preted to represent the history of any biologically continuous unit.

60
50
40
30
20
10

FIG. 9. The average catch in each sector during the period June, 1933 to
September, 1934. Ordinate : cubic centimeters of zooplankton per square meter
sea surface.

By taking account of the rate of drift of the water, it is possible to
select appropriate stations to show the growth of population in a unit
volume of water as it is carried about the Gulf of Maine eddy at the
apparent rate of its non-tidal drift. While the result is both an ab-
straction and an approximation, it probably indicates the true history of
events better than the usual curves of population fluctuation. Figure
10 represents the apparent growth of the population of a unit mass of
water entering the Gulf in December and recognized by the presence of
Limacina. As the season advances it is carried across the northern sec-
tors, arriving by May in the offing of Cape Ann. During this period
the population decreases by about one-third, but in May as it crosses
the western basin rapid growth occurs with the result that in midsummer

104

A. C. REDFIELD

it has reached the value of 50 cc. per square meter, greater than the
yearly average for the Gulf as a whole. The unit may now drift out
of the Gulf across the end of Georges Bank in late summer or it may
drift northeasterly into the Yarmouth sector to commence a second cir-
cuit of the Gulf. Its numbers grow meanwhile to over 100 cc. per
square meter. After January 1, high mortality again reduces the num-
bers to about one-third or to 30 cc. per square meter, as it crosses the
Massachusetts sector in March. In May growth recommences and by

XII I I I II I III I IV I V I VI VII I VIII
FIRST CIRCUIT

IX X XI XII I II III IV
SECOND CIRCUIT

FIG. 10. The growth of population in a mass of water assumed to move along
course indicated in inset. Ordinate : volume of zooplankton in cubic centimeters
per square meter sea surface ; abscissa, time in months. The black bars indicate
the volumes caught at the selected stations. The positions of these stations and
the month of collection are indicated on the inserted chart.

midsummer it reaches the Georges sector, having attained the record
volume of 170 cc. per square meter. Its history may be terminated by
supposing it to be carried out of the Gulf at this time.

Figure 10 is presented because it illustrates the possibility of taking
account of the current system in an ecological analysis. It may be
pointed out that the life history of any individual species might be
treated in a similar way.

ANNUAL FLUCTUATIONS IN POPULATION

The average catch at all stations was more than twice as great during
the summer of 1934 as during the same period in 1933. Since the hauls
were numerous, widely spaced, and all made with the same technique,
there can be little doubt of the significance of this observation. The
difference is the more striking in that an unusually severe winter pre-
ceded the richer year.

CIRCULATION AND DISTRIBUTION 105

The suggested theory offers a tentative explanation of such fluctu-
ations. The poverty of the water in the northern half of the Gulf in
winter seems due to the introduction of a large volume of relatively
barren water from the Scotian Banks. It is suggested that more of this
water enters in some years than in others, and that the population of
the Gulf is impoverished in proportion to the magnitude of this inflow.
The longer water remains in the Gulf, before being replaced by new
water, the richer its population becomes.

A number of facts support this suggestion. It is well known that
the magnitude of the inflow varies from year to year. The inflow dur-
ing the winter of 1934 seems to have been of shorter duration than
usual, having been completely terminated by a strong movement of
water from the Gulf over the North Channel and adjoining banks in
May. The movement was apparently underway in the latter part of
April, as the current diagrams show (Fig. 6). Drift bottles set out by
Dr. Herrington off Cape Sable at that time were recovered to the east-
ward. In contrast to this, Bigelow observed the invasion of the Gulf
by Nova Scotian water to continue until May or June and to result in
a cooling of the eastern part of the Gulf long after the other parts had
commenced to warm. He speaks of the invasion as a phenomenon of
spring, whereas in 1934 it terminated before the end of winter.

It is also possible that differences in the circulation such as those
observed in September, 1933 and 1934 may cause the population of the
Gulf to be augmented from external sources to a different degree each
season.

Since the water flowing into the Gulf over the Nova Scotian Banks
is less saline than that occurring at like depth within the basins, a small
influx of this water should be followed by a summer of relatively high
salinity. The superficial water of the Gulf was exceptionally salt in
1934. It appears to have been as salt or salter than in 1915, the most
saline year recorded by Bigelow. In 1933, on the other hand, the water
was quite as fresh as in 1914 and 1916, the least saline of those he re-
corded. A correlation between salinity, productivity, and annual inflow
over the Scotian Banks is strongly suggested. Systematic annual ob-
servations would serve to test this relation, and should it prove general,
might lead to an understanding of the yearly fluctuations in commercial
fisheries.

THE DISTRIBUTION OF PETRELS AND MACKEREL

The zooplankton supplies food for various predators. Their distri-
bution may be expected to be influenced secondarily by the hydrographic
factors which determine the abundance of the calanoicl community.

106

A. C. REDFIELD

Petrels

These birds appear to feed upon zooplankton or their products.
They pick up whatever scraps of organic matter they can find, gathering
about fishing vessels, following ships, and feeding about the carcasses
of dead whales and seals. There is a general belief that they pick up
droplets of oil from the surface of the water, and their stomachs fre-
quently contain an oily fluid which they eject when captured. It seems

JUNE 25-
JULY I 1934

AUGUST 18-23
1936 —

FIG. 11. Numbers of petrels observed in different parts of the Gulf of Maine
during cruises at various times of year.

more probable that this is derived directly from their food. Wilson
(1907) states that the food of the Wilson's petrel, which he observed
in the antarctic, consists of minute crustaceans. The natural food of
the Leach's petrel, according to Bent (1922, p. 143) "includes shrimps
and other small crustaceans, floating mollusks, perhaps small fishes occa-

CIRCULATION AND DISTRIBUTION

107

sionally, and probably many other forms of minute marine animals which
are found swimming on the surface or in floating masses of seaweed."
Wilson's petrel is the common petrel of the Gulf of Maine. Leach's
petrel, though a breeder along the coast of Maine, is much scarcer.
Not more than one petrel in twenty or thirty observed at sea is of this
species. Wilson's petrel breeds in the south Atlantic during December,
January and February and does not reach the Gulf of Maine until May.
None were observed during the cruise of April 19-23, 1934. In cruises
in May, 1934 and 1936 petrels were observed in small numbers in the
southern and eastern regions, where at that season the largest zooplank-
ton hauls were taken, but were absent from the northern sectors which
at the time were occupied by relatively barren water (Fig. 11). During

MACKEREL FISHE
JUNE 1931-19

MACKEREL FISHERY
AUG. 15 -SEPT. 30
1931 - 1934.

FIG. 12. The areas occupied by the mackerel fishery during the early and late

summer 1931-1934.

a cruise from June 25 to July 1, 1934, petrels were present in larger
numbers in the southwestern quarter, but were not observed in the
Seguin sector. On August 18 to 23, 1936, however, petrels were ob-
served regularly at stations made in the northern side of the Gulf. In
September, 1933 petrels were observed everywhere in the Gulf except
at a few coastwise stations along the shore. At this time all sectors sup-
ported an abundant population of zooplankton.

It appears that when Wilson's petrel first arrives in early summer it
remains confined to those sectors of the Gulf which then support the
richest plankton.

Mackerel

In the Gulf of Maine these fish have long been known to feed on
calanoid copepods and are known to eat various other crustaceans which

A. C. REDFIELD

compose the bulk of the zooplankton (Bigelow, 1926, p. 102). It is not
surprising in consequence to find a correlation between the seasonal dis-
tribution of this fishery and of the local abundance of zooplankton.

Dr. Settee has kindly placed at my disposal the very complete records
made by the U. S. Bureau of Fisheries showing the places in which
mackerel have been taken during recent years. The mackerel fishery
begins in New England waters with the arrival of the fish south of
Cape Cod in May and June. The fish are first taken along the shores
of the Gulf of Maine in June, chiefly within the 100-meter contour from
Cape Ann to Cultivator shoal. At this time heavy catches are also made
along the east coast of Nova Scotia and northward to Gaspe (Settee
and Needier, 1934). The July fishery has much the same distribution,
though tending to spread farther east along Georges Bank and also in
some years along the western part of the coast of Maine. In August
and September a considerable fishery is conducted in the northern side
of the Gulf, from Mount Desert westward, and southward as far as
Cape Cod (Fig. 12). During these months the fishery in the Bay of
Fundy is at its height, 37 per cent of the total catch being made during
each of these two months, whereas less than 10 per cent is taken in any
one of the preceding or following months.

In interpreting these facts the peculiarities of the fishermen as well
as of the fish must be borne in mind. Mackerel are now marketed fresh
and are landed chiefly in Boston. The fishermen consequently do not
fare farther from this port than is necessary. The takings of mackerel
do not reflect accurately the total distribution of the fish, but only their
availability to the Boston market. It seems sufficiently clear that in
early summer mackerel are available chiefly along the southern shores
of the Gulf ; that by late summer their abundance has shifted to the
northern shores, including the Bay of Fundy. This is the distribution
of the maximum of zooplankton population as well. Whether the
mackerel follow the plankton as it drifts around the great eddy, or cut
across to meet its advance from the east in the late summer, as the fisher-
men undoubtedly do, cannot be told.

Thus there appears to be a general correlation between the distri-
bution of the zooplankton, the occurrence of petrels, and the capture
of mackerel.

SUMMARY

1. The seasonal and geographical fluctuations of the abundance of
the calanoid community of the Gulf of Maine are described.

2. The shift in the center of abundance is closely correlated with
the superficial circulation, deduced from hydrographic observations and
the drift of an invading population of Limacina.

CIRCULATION AND DISTRIBUTION 109

3. The principal factor influencing the distribution of population
density is the inflow of relatively barren water from the Nova Scotian
coast in winter.

4. The Gulf appears to be an area from which the calanoicl commu-
nity spreads to other waters, but which receives relatively small recruit-
ments from without its borders.

5. A breeding stock is maintained by the establishment of a recur-
rent eddy in the late spring.

6. Estimates of annual productivity and seasonal mortality are given.

7. The distribution of petrels and of the mackerel fishery appears
to be correlated with the distribution of zooplankton.

REFERENCES

BENT, A. C., 1922. Life history of North American petrels and pelicans and their

allies. U. S. Nat. Mits. Bull. 121, Washington, 1922.
BIGELOW, H. B., 1926. Plankton of the offshore waters of the Gulf of Maine.

Bull. Bur. Fish.. 40 (Part II) : Document 968.
— , 1927. Physical oceanography of the Gulf of Maine. Bull. Bur. Fish., 40

(Part II) : Document 969.
BULLETIN HYDROGRAPHIQUE POUR L'ANNKE 1933. Conscil perm. int. explor. de la

mer. Copenhagen, 1934.
BULLETIN HYDROGRAPHIQUE POUR L'ANNEE 1934. Conscil perm. int. explor. de la

mer. Copenhagen, 1935.
CLARKE, G. L., 1933. Diurnal migration of plankton in the Gulf of Maine and its

correlation with changes in submarine irradiation. Biol. Bull., 65 : 402-

436.
CLARKE, G. L., 1934. Further observations on the diurnal migration of copepods

in the Gulf of Maine. Biol. Bull., 67 : 432-455.
DA MAS, D., 1905. Notes biologiques sur les copepods de la Mer Norwegienne.

Publ. de Circonstance, No. 22, Copenhagen.
FISH, C. J., 1936. The biology of Calanus finmarchicus in the Gulf of Maine and

Bay of Fundy. Biol. Bull.. 70: 118-141.
HUNTSMAN, A. G., 1924. Oceanography. Handbook of tlic Brit. Ass. for the

Adv. Sci.. Toronto, pp. 274-290.
ISELIN, C. O'D., 1939. Some physical factors which may influence the productivity

of New England's coastal waters. Jour. Mar. Res., 2 : 74-85.
KUNNE, CL., 1933. Weitere Untersuchungen zum vergleich der Fangfahigheit

vershiedener modelle von vertikal Fischen den Plankton-netzen. Rapp. et

Proces-verb. des Reunions. Conscil perm. int. pour I'c.rplor. de la mer,

83 : 19 pp.
PARR, A. E., 1936. On the probable relationship between vertical stability and

lateral mixing processes. Jour, du Conscil, 11 (No. 3) : 299-313.
REDFIELD, A. C., 1939. The history of a population of Limacina retroversa during

its drift across the Gulf of Maine. Biol. Bull.. 76: 26-47.
REDFIELD, A. C., AND ALICE BEALE, 1940. Factors determining the distribution of

populations of chaetognaths in the Gulf of Maine. Biol. Bull., 79 : 459-

487.
SETTEE, O. E., AND A. W. H. NEEDLER, 1934. Statistics of the mackerel fishery off

the east coast of North America, 1804 to 1940. U. S. Bur. of Fisheries,

Invcstigational Report No. 19, pp. 1—48.
S0MME, I. D., 1933. A possible relation between the production of animal plankton

and the current-system of the sea. Am. Nat., 67 : 30-52.

110 A. C. REDFIELD

K, J. D., 1934. Animal plankton of the Norwegian coast waters and the open
sea. I. Production of Calanus finmarchicus (Gunner) and Calanus hyper-
boreus (Kroyer) in the Lofoten area. Report on Norwegian Fishery and
Marine Investigations, IV, No. 9, pp. 3-163.

WALFORD, L. A., 1938. Effect of currents on distribution and survival of the eggs
and larvae of the haddock (Melanogrammus aeglefinus) on Georges Bank.
Bull. U. S. Bur. Fish., 49 : 1-73.

WILSON, E. A., 1907. National Antarctic Expedition 1901-1904. Natural History,
Vol. 2, Zoology, Part 2, Aves.

CHANGES IN THE TISSUE CHLORIDE OF THE

CALIFORNIA MUSSEL IN RESPONSE TO

HETEROSMOTIC ENVIRONMENTS x

DENIS L. FOX
(From the Scripps Institution of Oceanography, La Jolla, California)

The adult California mussel, My tikis calif ornianus Conrad is a fairly
heterosmotic animal, typically marine in its environment, yet potentially
euryhaline in a striking degree.

It has been shown (Fox et al., 1936) that this mollusk can live in
the laboratory for long periods following sudden and continued immer-
sion in aerated solutions of natural sea salts, varying in Cl concentra-
tion - from about 0.94 per cent, or approximately half that of natural
sea water in the vicinity of La Jolla, to as high as 2.5 per cent, or about
34 per cent above normal values (1.86 per cent being the approximate Cl
concentration of sea water at La Jolla).

Sudden exposure to concentrations of sea salts below or above
these respective values proved fatal to mussels, but gradual alteration of
the water in which they were immersed, over periods of several weeks,
left the animals still alive in solutions diluted, on the one hand, to one-
third of the normal Cl concentration (0.62' per cent Cl) or concen-
trated, on the other hand, to nearly twice the normal value (3.50 per
cent).

The purpose of this investigation was to determine the chloride
concentration in the tissues of the mussel, under the widely differing
physiological conditions which must result from continued exposure to
concentrations of sea salts not encountered by these animals in nature.

Extensive reviews of both older and more recent work dealing with
homoiosmoticity and heterosmoticity of fishes and invertebrates have
been presented in papers by Dakin (1935) and by Schlieper (1935).

1 Contributions from the Scripps Institution of Oceanography New Series No.
117.

2 Ci concentration refers to the total halide ion concentration as determined in
sea water analyses. The same designation, used in the present discussion, refers
to the total halide concentration in parts per cent of both water and tissues, since
the amounts of both Br and I are relatively very small. Concentration of " chlo-
ride " refers hereinafter to that of halide ions.

Ill

112 DENIS L. FOX

Other useful discussions of the subject are given by Aclolph (1930)
and by Baldwin (1937).

Experimental Methods

Diluted sea water solutions were prepared by adding distilled water
to sea water, while relatively concentrated solutions were obtained by
the partial evaporation of ordinary sea water without precipitating any
salts.

Experimental solutions were contained in glass battery jars, each
of 2% gallon capacity. To avoid any injurious effects which might
result from overcrowding, the predetermined ratio of at least one liter
of water per animal was consistently adopted as a minimum. Mussels
of an average length of 10 cm. (varying between 9 and 11 cm.) were
employed in the great majority of experiments, although no biochemical
or physiological differences of direct bearing upon the work were recog-
nized in larger animals (12 cm.) or in somewhat smaller ones (7 to 8
cm.)

Experiments were carried out at room temperatures which varied
between the approximate limits of 19° and 22° C. All solutions con-
taining mussels were aerated continuously.

No attempts were made to adjust and maintain the pH of the various
solutions, since it was found that mussels lived in piped sea water
within the range of pH values encountered in the experiments. Figures
representing numerous pH measurements of environmental solutions
will serve as examples and are tabulated below. These measurements
were made with a Beckman glass-electrode pH meter through the
kindness of Mr. J. C. Hindman.

Solution pH

(1) Sea water of normal salinity from running supply in tank

containing mussels ; unaerated 8.63

(2) The same; aerated 8.43

(3) The same, drained from mantle-cavity of mussel; un-

aerated 7.51

(4) Sea water of normal salinity from stationary supply in

large jar containing mussels; unaerated 7.50

(5) Sea water, initially of normal salinity, diluted to maxi-

mum degree used in experiments, i.e. to one-third of
original concentration (ca 0.62 per cent Cl) ; aerated 8.52

(6) Sea water, initially of normal salinity, concentrated to

maximum degree used in experiments, i.e. to 54 per

cent of original volume (ca 3.50 per cent Cl) ; aerated 8.32

TISSUE CHLORIDE CHANGES IN THE MUSSEL 113

Since the mussels were therefore presumed not to have been exposed
to unfavorable conditions of acidity or alkalinity during the course of
the experiments, it was concluded that extensive changes in the concen-
tration of dissolved salts themselves constituted the limiting factors to
life in the environment. The possibility of disturbances in " salt bal-
ance " (in the physiological sense), of experimental solutions was con-
sidered. Because calcium was believed to be the chief element prone
to be precipitated as the carbonate from moderately concentrated sea
water, analyses were made for dissolved calcium in sea water samples
concentrated by boiling. Normally present in amounts close to 0.42
gram per liter, the dissolved calcium in a solution boiled down without
precipitation to 54 per cent of its original volume (number 6 in above
table) was present in the nearly theoretical amount of 0.80 gram per
liter. Since this solution was the most concentrated used in any of the
experiments, it was concluded that mussels were at no time exposed to
solutions " unbalanced " with reference to calcium.

Experimental animals were introduced into the various solutions of
altered salinity after first propping the valves apart by a few millimeters
with smooth glass plugs and draining the gill chamber of sea water.
Animals immersed under these conditions in sea-salt solutions varying
between the approximate limits of 0.95 per cent and 2.73 per cent Cl
usually relaxed their hold on the glass rod in a short time, parted their
valves in a normal manner, and resumed feeding activities.

The flesh of control and experimental mussels was prepared for chlo-
ride analysis in the following way. The mussels were removed from
their shells as rapidly as possible and with minimum cutting of tissues,
severing only the adductors and small muscles attached to the hinge
region of the shells. The flesh was blotted on absorbent paper to re-
move most of the adhering sea water, then rinsed briefly in 95 per cent
ethyl alcohol (and reblotted) in order to remove most of the remaining
sea water from body surfaces, constrict the gill capillaries, wash out sea
water expressed therefrom, and coagulate the cut surfaces to allay ex-
cessive bleeding and subsequent losses of chloride-containing body fluids.
The consistent adoption of this procedure resulted in a series of checks
which were quite close in normal control animals, in spite of the fact
that the work was done on the wet weight basis.

For chloride analysis, the method of Sunderman and Williams (1931,
1933) was followed, digesting the whole tissues in chloride-free KOH,
followed by further treatment of aliquot portions with concentrated
HNO, in the presence of an excess of dilute AgNCX solution. The
excess of Ag ion was finally titrated with standard NaCNS solution in

114

DENIS L. FOX

s
1

C/)

1)
j->
C

000°".

00°

II II II

oq 10

O O

II II

ON
O

[/)
rt

r^ ro TJ< NO
ON ON ON ON

O O O O
II II II II

^be o

^") O*~ £*"

-a -c

rt rt

C be

C C

o v, J£

O- O- ^
*u "u ^

O O

bo be

T 8 I

<D ^J Ul

^ CH-

"o "o O

OI <u

S
g

QJ bo

CL be— —

# ?

'u 'El

• —

c/; O rt rt

.S .S

13 15 "o "o

U U UU

rt rt

rt cd cd rt

LO LO LO

f \

LO LO

IO 1O

§NO NO *••< l^~ ON

O -^ ^ OO CN CN

t-^ 00 NO 00 O ON

^3 , — i \^ j^ r^ ^~H ^^ T^H ^) ro

ON O 00 t^ ON

ON NO ON >O OO IO

ON ON ON ON O ON

O ON O- O\ O^ O O^ O^ O\ O

OOOOOOO — O'-'OOO

O O O O O O

O O O O — < O

•^OOOO^OOO-^

10 00

O oq r~;
LO PD O

— < NO OO ON NO CO
^^ CM O** *~^ ^^ ^^

oo fS

NO NO

OO "0

" b/3

Ow. iu

- a,

u _ cue,

1" lie"

aj
h

_ >-i

11
1 s

CM-

3
+J
rt :

u> C en G <n C <"

b*,^ ii - bo qj be a; be- — -
ac-o o, be a. be a. be-o

c f u

cuauc

V
•Si

'b 'b

III

I I I

CO OO 00

OO
«*5

I
00

TISSUE CHLORIDE CHANGES IN THE MUSSEL 115

the presence of ferric alum according to the well-known method of Vol-
harcl. The Volhard method was also employed for the analysis of
chloride in sea-water solutions. Our experiments showed that the step
involving preliminary digestion of tissues by KOH gave consistently
higher values for chloride than did the ordinary " open Carius " deter-
minations (digestion with excess concentrated HNCX and AgNO3) em-
ployed by other workers. Sunderman and Williams (1933) report in-
complete recovery of chloride when the preliminary alkaline digestion
is omitted, and assign the low chloride values to interference by fatty
substances.

Preliminary Analytical Survey of Normal Animals

Because parts of this research were conducted in different seasons
of the year, i.e. especially in the summer and fall of 1937 and the winter
of 1937-1938, it seemed desirable to compare chloride analyses of normal
animals taken during August with those of animals comparable in size
and weight taken in December and January. Also, because it was im-
possible to differentiate the sexes without sacrificing the animals' lives,
and since it was conceivable that biochemical differences in sex might be
reflected in the chloride content of the tissues, attention was given to
the sex and relative degree of maturity throughout the same group of
animals.

The data of Table I reveal that differences between the chloride con-
tent of whole bodies of summer animals and those of winter animals
are of relatively small order, showing a departure of only -f- 0.06 per
cent in the grand average, in favor of the winter animals ; this difference
is the same in direction and extent whether one compares ripe winter
with ripe summer males, or ripe winter with ripe summer females.

Sexually immature animals exceeded in chloride content the grand
average (0.93 per cent) of the combined values of mature animals of
both seasons by the small departure of 0.05 per cent.

Because the demonstrated seasonal and sexual differences in chloride
content fell well within the departures recorded between individual
analyses, they were not regarded as significant in the experimental re-
sults. The sexual difference in chloride content appeared to be a real
one, although relatively small, and for the purposes of this work, in-
significant. It was doubtless due to the fact that the relatively heavy
ripe ovary contained only about two-thirds as much chloride as did the
male gonad.

116

DENIS L. FOX

Changes in Tissue Chloride Following Sudden Immersion in Heteros-

motic Solutions of Sea Salts within the Tolerated

Physiological Range of Concentrations

Reference to Figs. 1 and 2 brings out some rather consistent general
facts : While mussels in nature show close agreement among one another

CO
CO

1.7

1.5

|.3

0.7

0.3

o
o

Chloride in water = 2.52%

USdays

^3 days

Chloride in water = 1.86%

• i! .'
. .1

Chloride in water = 0.95%

•<>•

t ' • ; :

Summer Series: Aug. - Sept., 1937.

,[8 days

"0 4 8 12 16 20 24 30 40

Time in Hours

FIG. 1. Chloride analyses of whole mussel tissues following immersion of
living animals for increasing time-intervals in hypotonic and hypertonic sea water
within the limits of the tolerated range. Each point represents the analysis of a
single animal ; where two or more analytical values were identical, this is shown by
concentric circles.

in chloride content, wide individual variations are manifest in the rate
of change in tissue chloride concentrations in both kinds of new en-
vironment. This is particularly striking in the early hours following
immersion. Since mussels have been demonstrated to survive for long
periods at either the hypotonic or the hypertonic concentrations here

TISSUE CHLORIDE CHANGES IN THE MUSSEL

117

employed, and to eventually attain, in each group, respective chloride
levels in close individual agreement, the obvious differences in the early
hours are clearly due to individual variations in rate of water and salt
interchange.

2.0

1.8

1/1

_ 1.4

r 1.2

** I 0

o_' -u(
a;

_20.8
o

0.6
0.4
0.2

Chloride in water
= 2.52 to 2.73%

8 o
o o

Chlonde in water
= 1.86% (normalK

1 * .Chloride in water
= 0.94%

Winter Series:

Jan. -Feb., 1938.

0 4 8 12 16 20 24
Time in Hours

FIG. 2. Further chloride analyses, as in Fig. 1.
See Fig. 1 for meaning of points and concentric circles.

While four out of five animals which remained for the whole 24-hour
period in water containing 2.52 to 2.73 per cent Cl (Fig. 2) were filtering
water, feeding, and voiding feces and thus appeared normal, none of
the animals in the water containing 2.78 per cent Cl (Fig. 3) appeared

118 DENIS L. FOX

normal even after 30 hours ; they maintained a grip on the inserted glass
props, failed to filter water, and gave off much mucus, thus giving several
signs of physiological disturbance. A later lot, however, exposed to a
solution of 2.80 per cent Cl (Table IV) did not show any effects of
injury after 25 hours. These facts are taken as evidence that Cl values
between 2.70 per cent and 2.80 per cent are close to the threshold of
hypertonicity at which the mussels can withstand sudden immersion.

Figure 4, to which reference will be made below, reveals the close
fit to a straight line between water chlorinitics tolerated by mussels after
sudden immersion. Attention is directed, for the moment, only to the

2.4

2.2

s/l
9)

•zl.4

Chloride in waterr2.78%
- °

1.2

024 6 8 10 18 30

Time in Hours

FIG. 3. Further chloride analyses, as in Figs. 1 and 2.
See Fig. 1 for meaning of points and concentric circles.

portion of the graph lying between the ordinate values of 0.95 per cent
Cl and 2.8 per cent Cl, illustrating the data shown in Figs. 1 and 2,
and those of Table IV, when all average values of internal chloride
concentrations attained by mussels in 24 to 40 hours are plotted against
chloride concentrations of external water. The ratio between grams
Cl per 100 grams wet mussel tissue and grams Cl per 100 ml. sea salt
solution has an average value of 1 : 1.89.

The blackened point on Fig. 4 indicating the average chloride con-
centration in the tissues of 8 mussels maintained for 30 hours in water
containing 2.78 per cent Cl was obtained from Fig. 3. While this

TISSUE CHLORIDE CHANGES IN THE MUSSEL

average value lies in the vicinity of the curve, its departure is doubtless
due to the fact that the animals were in poor condition (from an un-
known cause), and so failed to maintain the ratio between internal and
external chloride exhibited by other lots of animals at this concentration.
Specimens placed, with valves propped apart, in distilled water for
15- or 16-hour periods underwent a drop in tissue chloride to values of
0.226 per cent or even 0.112 per cent. Under these conditions animals,
although not quite dead, were definitely moribund, and failed to recover
when placed in running sea water.

Tolerance by the Mussel of Gradually Altered Concentrations

of Sea Salts

While the immediately preceding experiments and earlier investi-
gations seemed to define fairly well the limits of hypo- and hypertonic
solutions withstood by mussels following sudden immersion, there were
grounds for believing that gradual changes in the salt concentration in
the animals' environment might not bring about signs of injury until
greater extremes in both directions were reached.

In order to investigate this question, and to determine if possible the
extremes of tissue chloride lost or gained by surviving animals, the fol-
lowing experiments were conducted. Two sets of mussels in separate
jars of normal sea water, continuously aerated, were subjected to a
gradual change in the chloride concentration of the water, one jar being
diluted, the other concentrated daily by slow steps over a period of about
six weeks. This gradual change was brought about in the following
manner. From one jar containing about eight liters of sea water (seven
mussels) some 250 ml. were withdrawn, discarded and replaced by an
equal quantity of distilled water daily; from the other jar (eight mus-
sels) the same volume of water was discarded and replaced by an equal
quantity of water containing sea salts concentrated by about four-fold
(i.e. Cl concentration of 7.725 per cent). These changes brought about
a gradual drop in Cl from 1.86 to 0.625 per cent in the first, and a
gradual rise from the same initial value to 3.48 per cent in the second
solution, over the experimental period. Under these special conditions
the mussels survived exposure to previously unexpected concentrations
at both extremes.

Table II shows the results of chloride analyses, and a restoration
of experimental animals to normal chloride levels after exposure to the
gradually altered solutions during five and six-week periods.

It is assumed from the foregoing experiments that it has been pos-
sible to ascertain the approximate thresholds of tolerance of the mussel
toward the osmotic effects of both hypotonic and hypertonic solutions of

120

DENIS L. FOX

S i- °!

e\ gx ea ,. <u o

KJfNt^OvtN Th ^ ctf

<»

OOO«^-t^.t-~ oo p_|j

^^^rt^ _ £Jjrt-oU

1 1
1 b

^£ .. ~ J5 « S "

S So J0-^^
.6 g _S £.5 £ ^

_c
'c
5

C Q; C 3 C o3 • — '

cfl > 03 O C -o ™

"0 *^ »-( 6Jc S vo a!

u
rs

is-

1L> fi

S >

C ON Ov

•" y

^ q o

1 1 *

— 'H

|ro

0) v^-
S U ^

C ^°

•s « S:

__ ._• ON

S Si

Co <U

•2 S
I-H -a 2
•— i s

w '*

rf S> c
w s
•< o

M 'vj

>vO >. O

^ ^ S c S "

- 0 ro 00 <" .£ rt ||

'*'*."? ^ S T> is

?°° ° 8 g rt-au

tn < 03 ^ C ..

g, "« 0- bo * "g
.H of .6 c£.C « N

"5 gj 'S 3 C o3 -^
03 > 03 0 CT, g

^ *^ "H t* u vo 03

^g"S ^ d

^5 ^l|js^l

™. § ^^^a^rt'r! 1

"in eSo'c^iu

C •- 1_ g >, N
^ 2 M § 03 >^

S-2 2"°^

^•s.s^ s

*»-4
t U

•**» Q-4->

1 !?

04 fN

w OO OO

^o \o

M '"'S

& 0 0

d d

S-g

tn tn

<U ~t

JS JS jS

<U <U (!)

<U QJ

<u a> S

* * *

<-s

O «-H CM

'~H

CN CM CM

o a

4J 5

Q Q

II 1

03 -fi
S'S

n
coming

Average

TISSUE CHLORIDE CHANGES IN THE MUSSEL

121

sea salts. It seemed hardly likely, however, that individuals kept in
solutions of concentrations below 0.94 per cent Cl or above 2.8 per cent
Cl would survive indefinitely, since the flesh of animals in solutions
beyond these respective concentrations appeared rather thin, sometimes
even emaciated. Furthermore, they failed, in these respective realms
of salt concentrations to maintain their constant ratio of tissue chloride
to sea water chloride as illustrated in Figure 4. This graph summarizes
much of the information reported above and indicates the relationship
which exists between the chloride concentration in the surrounding
media and that in the tissues of animals immersed therein. Each point

2.5

2.0

O
O

1.5

« I o

r-> I . \J
OJ

| 0.5
o

Po.o

((Threshold
I of tolerance)

(Indefinite survival)

lormal

>Fatal

-(Indefinite survival)
(Threshold of tolerance)

0 0.5

4.0

1.0 1.5 2.0 2.5 3.0 3.5

Grams Chloride per 100 cc Solution

FIG. 4. Relationship between internal and external chloride, in grams Cl per
100 grams tissue vs. grams Cl per 100 ml. water respectively. Points lying between
abscissal values of 0.95 and 2.8 represent results obtained after 24 hours or more
following sudden immersion. Portions of the graph lying below and above these
respective values represent the trend taken following gradual alteration of Cl con-
centration in the water over protracted periods, excepting wherein distilled water
was used (see text).

represents the average of the analyses of a number of animals immersed
in the respective solutions for 24 hours or more, save in the cases of
the two lowest points on the hypotonic side, which represent analyses
of two respective animals : a single survivor containing 0.26 per cent
Cl and the animal which showed the highest Cl value (0.226 per cent)
of the group that succumbed to immersion in distilled water (at zero
abscissal value). The maintenance of a ratio of 1:1.89 between in-
ternal and external chloride is noted in the straight line between the latter
concentrations of 0.94 per cent and 2.8 per cent Cl. The projection of
this straight line as a dotted line beyond these values shows, when com-

122 DENIS L. FOX

pared with the actual curve, the degree of departure in the animals'
maintenance of such a ratio. The final two points at each extremity of
the curve represent Cl values attained by mussels after gradual changes
in Cl concentration of the water to the corresponding values shown on
the abscissa. (See Table II.)

Content of Water and Chloride in Various Tissues: Exchanges of both
Water and Chloride Ions between Tissue Fluid and Environment

Since some question arose as to whether a considerable part of the
observed changes in tissue Cl might be assignable to the mere presence

TABLE III

Water and chloride contents of gills and bodies (minus gills) of normal mussels.

Gills: Wet weight 11.02 grams

Dry weight 1.09 "

H20 90.1%

Cl (wet wt.) 1.32%
Mols. Cl per liter

tissue water 0.413

Bodies

minus

Gills: Wet weight 59.25

Dry weight 7.30

H20 87.7%

;< '

Cl (wet weight) 1.075%

Mols. Cl per liter

tissue water 0.345

Combined

Total: Wet weight 70.27

Dry weight 8.39

H2O 88.06%

Cl 1.11%
Mols. Cl per liter

tissue water 0.355

of the medium itself in the capillary tubes of the gill structures, analyses
were made of water and chloride in (1) the gills and (2) the rest of
the tissues en masse, of six mussels taken directly from stock tanks of
running sea water. In consistency with the experiment, tissues could
not in this case be rinsed in alcohol prior to analysis (hence the slightly
elevated Cl values). Table III shows the results of these analyses, and
brings out the fact that the gills, as dissected for analysis, possess only
a slightly higher moisture content than do the other tissues ; further-
more while the chloride content of the gills was about 23 per cent higher
than that of the other tissues, their relative proportion of the total wet
body weight was only about 15.6 per cent, so that they were responsible

TISSUE CHLORIDE CHANGES IN THE MUSSEL 123

ir
W

O
<o

~se

3 Si

*• ^i r->

| §

a <o

"3 fe

o -ft.

I \\

s G

<-> c

1^3 "a

•^ ^

t I

nj >•

•- g yQ^ 3

d,g<«ffi^.S

fe«&Ss

£ =
US

t/J

"1 O FD ,O^O " OO
O — ' O O — < .,• O

^
£8

O-—

Os VO

"5 u->

oooooooooooortoo O\oooocooooOn3c»

LO 10

irj 10

O — i «M --

^OvOir>vl^t^'->O C>O\t~-O\oo\>Ov 1O u «*5 »O VO

OOOOOOOQOOoOctfOO OvoOOOOOoOOOcSOO OOOOOOCXJOO

ooioOOooo "«— i »ot--r-iot^f^3 "'f
r-~ t^ t-~ t— r- r^» rti~~ oooooooooooo rtoo

C-l OO O — H \O OO

IT)

OOOOOoiiO

OO rf

O O p p p p

OOOOOO

p p p p p p
o' o o o o o'

IO 1O

p p -; p ~

o o o o o

in if) \r>

OOO— '

t^* LO ^t^ ^^ OO C**]

C^J O^l C^3 C^J v~~ *~*

OOOOOO

— 03 —

i/; 10

OOOOOO

irj uo 10

O O O O O

8S
fc

QQQQQQ UUUUUU

124 DENIS L. FOX

for bringing the total tissue chloride up from 1.075 per cent to 1.11 per
cent, an increase of only about 3.5 per cent. Observed shifts in tissue
chloride were therefore not due to environmental solution mechanically
suspended in the gills.

The question of whether the change in tissue chloride concentration
might be due, at least in part, to the osmotic interchange of major quanti-
ties of water, with or without the migration of Cl ions as well, was in-
vestigated in an experiment involving eighteen animals of the usual
range of length (i.e. 92 to 110 mm.), the results of which appear in
Table IV. Six mussels were immersed in ordinary sea water, six
propped open in hypotonic water ("50 per cent sea water"; 0.97 per
cent Cl), and a third set of six propped open in hypertonic water wherein
the Cl concentration was 2.80 per cent. After being kept in the re-
spective, constantly aerated solutions for 25 hours, the animals were
analyzed. The foot of each was severed with a sharp razor blade and
analyzed separately from the other tissues, in order to determine whether

TABLE V

Weight of water and of chloride per unit weight of dry tissue; concentration of chloride

per 100 grams of tissue water.

Series

Grams FhO per
gm. dry tissues

Grams Cl per
gm. dry flesh

Grams Cl per
100 gm. tissue-water

foot

body

whole

foot

body

whole

foot

body

whole

3.18

6.46

6.35

0.030

0.081

0.079

0.94

1.25

1.24

5.025

8.34

8.22

0.021

0.046

0.045

0.42

0.55

2.98

5.69

5.55

0.051

0.101

0.099

1.72

1.77

1.78

such a relatively compact structure might show any considerable differ-
ence in chloride shift as compared with that of softer tissues. The dry
weights of all tissues were obtained by keeping them overnight in tared
containers placed in an electric oven at 105° C., then re-weighing; chlo-
ride was determined in the usual manner. In the table, the N series
represent the normal control animals kept in sea water, the D series those
placed in the diluted sea water, and the C series those immersed in the
more concentrated solutions. Some significant trends are observed.
The average water content of foot tissues and of whole bodies shows a
consistent increase in the order which would have been expected from
a consideration of the osmotic effects of the relative solutions, i.e..
D > N > C; the relative chloride values of foot tissues or of whole
bodies show the opposite order, i.e., C > N > D.

The exchanges of both water and chloride ions, under the conditions
of the preceding experiment, are brought out in a quantitative way in
Table V, showing for each of the three series, the ratios of average

TISSUE CHLORIDE CHANGES IN THE MUSSEL

125

water content and average chloride content, in grams per gram of dry
tissue.

Furthermore, by plotting the values for Cl concentration in grams
per 100 grams of tissue water, against Cl concentrations in grams per
100 ml. of sea water solutions, we arrive at the nearly linear relationships
shown in Fig. 5. The foot, which contains less water, less chloride,
and lower concentration of the latter, normally constitutes only about
2 per cent of the total body weight, and exerts no significant influences
upon the data collected from analyses of the whole body. Nevertheless,

2.0

•I

I 1.0

1.0

Cl % of sea water

2.0

3.0

FIG. 5. Relationship between internal chloride concentrations in grams per
cent of tissue water and in environmental water respectively. Sea water chlorini-
ties lie within the range of values tolerated by mussels on sudden immersion.

this compact muscular tissue is observed to exchange both water and
chloride with the aqueous environment in a manner similar to that of
the other tissues.

Under the conditions of the above experiments, the mussel alters,
throughout a physiologically wide range, the concentrations of its dis-
solved salts in such manner as to maintain a rather constant ratio be-
tween the Cl concentration of internal and that of external water, this
ratio appearing to have an average value of the order of 1 : 1.60. The
migration of both chloride and water occurs between the tissues and the
solution outside.

126 DENIS L. FOX

Discussion

The ready exchange of water and chloride between the mussels' flesh
and the environment may not involve any such profound changes in
tissue cells themselves as might seem to be the case. The tissues of
some organisms contain major quantities of the total chloride in the
interstitial fluid, the cells themselves possessing the ion in very small
amounts, and not readily undergoing alterations in water content or
chloride concentration. This seems to be especially true in certain mus-
cular tissues of mammals (Eggleton, Eggleton and Hamilton, 1937;
Bourdillon, 1937), but some investigators are of the opinion that in
other tissues (notably in frogs) considerable portion of the chloride is
intracellular (Amberson, Nash, Mulder and Binns, 1938).

Since the completion of the present work, some informative results
have been published by Steinbach (1940 a, b), who studied the content
and distribution of water and electrolytes in the excised retractor muscles
of certain marine invertebrates, i.e. the holothurian, Thione briareus and
a sipunculid worm, PJiascolosonta.

Steinbach immersed his material for a few hours in solutions con-
taining chloride in concentrations of from 52 milli-equivalents per cent
(i.e. full-strength sea-water) to nearly zero, employing isosmotic sucrose,
or alternatively NaNO3 solution as the sea water diluent. No hypertonic
solutions were used. He obtained linear relationships throughout the
experimental range of concentrations employed, the fresh Thyone tissue
being" 'richer in Cl (20.4 meq. Cl per cent) than that of Phascolosoma
(9.1 meq. Cl per cent when fresh, and 16 meq. Cl per cent when soaked
in sea water), and the former yielding steeper slopes of linear change.
The normal water-content was very similar in both species, i.e. 75.9
per cent in Thyone and 78 per cent in Phascolosoma muscle.

The present writer immersed whole mussels for 24 hours or longer
in normal, dilute, and concentrated solutions of natural sea salts within
the range successfully tolerated by the animals. Linear relationships
were apparent between the range of about 1.74 and 79.1 meq. Cl per
cent of environmental water, when dealing with whole bodies, while such
data as were collected upon the subsequently amputated foot alone re-
vealed a nearly linear function between environmental Cl concentrations
of 27.7 and 79.1 meq. Cl per cent.

It is also of interest to note that the initial concentration of Cl in
Steinbach's Thyone muscle is virtually identical with that of the muscular
foot of Mytilus, viz. ; 20.4 and 20.3 meq. Cl per cent respectively, and
that the water-contents of both are close, viz.: 75.9 per cent and 76.1
per cent. Furthermore, the Cl concentration in half-strength sea water

TISSUE CHLORIDE CHANGES IN THE MUSSEL 127

(Steinbach's fortified with isosmotic solutions of sucrose or NaNO3,
mine merely diluted) resulted in a decrease in Thyone muscle Cl to a
value close to that determined for Mytilus-ioot, viz.; 12 vs. 9.9 meq.
Cl per cent respectively.

Steinhach's experiments led him to conclude that muscular tissues of
both Thyone and Phascolosoma (classed as smooth muscle) contain
sodium chloride almost exclusively in the extracellular space, that it is
free to equilibrate by simple diffusion with external solutions, and that
little if any can penetrate the tissue-cells themselves. He discusses the
evidence in support of the conclusion that virtually all chloride is extra-
cellular in the material studied by him, and in striated muscles of frogs
and some other vertebrates as well, and that measurement of chloride
content of such tissues may be employed as a relative measure of the
extracellular space.

The muscular foot of Mytilus is observed to show osmotic behavior
closely similar to that of the retractor muscles of Thyone and Phascolo-
souia. and must be very like the latter tissues in biochemical constitution
and function.

The whole body of Mytilus shows a similar linear slope as does the
foot tissue within the range of environmental Cl concentrations com-
pared, but exhibits consistently higher water and Cl content. Deviations
from the linear relationship between whole tissue Cl and sea water Cl
which occur at extreme dilutions and extreme concentrations represent
the failure of the physiological mechanism to control any longer the
integrity of the chloride-free cellular space under relatively drastic
conditions.

Investigations of biochemical or physiological changes in tissues of
whole organisms, living successfully in controlled foreign environments,
allow, in the conclusions drawn, a degree of certainty less frequently
assured from data collected on isolated, surviving tissues. In the present
instance, it is of much interest to note some close parallelisms between
different species investigated by the respective experimental approaches.

The work reported in this paper was begun with the experimental
ecological viewpoint in mind. In summarizing briefly, it is recalled
that while the adult California mussel is able to adjust itself, with
accompanying changes in tissue constitution, to a considerable range of
salinities in the laboratory, this species, unlike M. edulis, is rarely if ever
found in bays and estuaries, even though such waters may be consid-
erably less diluted than were solutions which the animals have been
shown experimentally to tolerate for indefinite periods.

Our findings during the course of this work suggested some experi-
ments on the effect of dilute solutions of sea salts upon ripe sperm and

128 DENIS L. FOX

eggs of this species, the process of fertilization, and subsequent devel-
opment. Some preliminary experiments of this kind were accordingly
carried out at the Scripps Institution by Dr. Robert T. Young, whose
findings indicate that sea water diluted by more than 25 per cent may
exert injurious effects upon (1) the sex products themselves, (2) inci-
dence of fertilization, and (3) subsequent development. Doubtless the
dilution of water in bays and estuaries is not the only factor responsible
for the failure of Mytilus calif ornianns to colonize them, but the sensi-
tivity of eggs, sperm, and larvae to the dilute environments provides one
tangible clue which should prove helpful in further attack on this and
kindred problems.

Acknowledgments

I take pleasure in expressing grateful appreciation to Mr. Arthur
Raymond Holland (formerly Chemist in the Federal Works Progress
Administration Project No. 7039, California District No. 12), and to
Mr. Hiomi Nakamura, graduate student, and to Mr. John Cunningham
(Chemist, W. P. A. Oceanography Project No. 9964-D), each of whom
rendered technical assistance at various stages of this investigation, in-
cluding the carrying out of chloride analyses under my direction; appre-
ciative acknowledgment is also given to other members of the same local
W. P. A. project who provided valuable aid in collecting and care of
animals, typing, draughting, and library service.

Summary

1. Experiments indicate that the adult mussel Mytilus calif orniaus is
heterosmotic, yet potentially euryhaline to a considerable degree, although
sperm, eggs and larvae are highly stenohaline toward dilution of the
environment.

2. The tissue-chloride content is close to 1 per cent by wet weight,
varying only slightly with season. Mature males show slightly higher
chloride values than do the mature females, due to the higher chloride
content of testicular tissues than of ovarian tissues.

3. Mussels can survive for indefinite periods the sudden and con-
tinued exposure of their tissues to sea water diluted by 50 per cent (Cl
cone. 0.94 per cent) or water concentrated to half again its normal
value (Cl cone. 2.73 per cent to 2.8 per cent). Below or above these
two respective extremes, sudden immersion is fatal.

4. Within the limits of the physiologically tolerated range indicated,
the concentrations of tissue chloride are adjusted to concentrations of
chloride in the environment, with maintenance of an approximate value
of 1 : 1.60, calculated as grams per 100 ml. of internal and external water.

TISSUE CHLORIDE CHANGES IN THE MUSSEL 129

5. Considerable individual differences exist in the rate of establish-
ment of equilibrium between environmental and tissue chloride concen-
trations, when mussels are exposed to the indicated dilute and concen-
trated solutions.

6. While sudden immersion in solutions of sea salts below or above
the respective limits resulted fatally, it was possible for mussels to
survive in solutions considerably beyond such limits, i.e. in water diluted
to as low as 0.62 per cent Cl, or concentrated to 3.48 per cent Cl, if the
concentrations were altered by gradual steps.

7. Mussels surviving in sea water, gradually diluted to a chlorinity
of 0.625 per cent Cl, underwent a fall in their tissue chloride to values
of about 0.26 per cent to 0.38 per cent ; animals kept in sea water grad-
ually concentrated to a chlorinity of 3.48 per cent underwent a rise in
their tissue chloride to average values of 2.25 per cent ; at these respective
points animals were at their threshold of tolerance and showed incipient
sluggishness. Animals of such extreme chloride levels, however, recov-
ered if placed in running sea water, and readily underwent therein a
restoration of their tissue chlorides to normal values.

8. Exposure of mussels to the diluted or concentrated solutions
results in migrations of both water and chloride between internal and
external media.

LITERATURE CITED

ADOLPH, E. F., 1930. Living water. Quart. Rev. Biol., 5 : 51-67.

AMBERSON, W. R., T. P. NASH, A. G. MULDER AND D. BINNS, 1938. The relation-
ship between tissue chloride and plasma chloride. Am. Jour. Physiol., 122 :
224-235.

BALDWIN, E., 1937. An Introduction to Comparative Biochemistry. Macmillan
Co. New York.

BOURDILLON, J., 1937. Distribution in body fluids and excretion of ingested am-
monium chloride, potassium chloride, and sodium chloride. Am. Jour.
Physiol, 120: 411-419.

DAKIN, W. J., 1935. Presidential Address: The aquatic animal and its environ-
ment. Proc. Linnean Soc. Xcu' South Wales. 60 : vii-xxxii.

EGGLETON, M. G., P. EGGLETON AND A. M. HAMILTON, 1937. Distribution of chloride
in frogs' skeletal muscle immersed in saline solution. Jour. Physiol., 90:
167-182.

Fox, D. L., ET AL., 1936. The habitat and food of the California sea mussel. Bull.
Scripps Inst. of Occanogr., Tech. Scr., 4 : 1-64.

SCHLIEPER, CARL, 1935. Neuere Ergebnisse und Probleme aus dem Gebiet der
Osmoregulation wasserlebender Tiere. Biol. .Rev., 10: 334-360.

STEINBACH, H. B., 1940a. Electrolytes in Thyone muscles. Jour. Cell, and Co»if>.
Physiol., 15: 1-9.

STEINBACH, H. B., 1940fr. The distribution of electrolytes in Phascolosoma muscle.
Biol. Bull., 78: 444-453.

SUNDERMAN, F. W., AND P. WILLIAMS, 1931. Diminution in chloride measure-
ment after drying blood and tissues. Jour. Biol. Clicin., 92 : 99-107.

SUNDERMAN, F. W., AND P. WILLIAMS, 1933. The analysis of chloride in tissues.
Jour. Biol. Chan.. 102 : 279-285.

MATING REACTIONS OF ENUCLEATE FRAGMENTS IN

PARAMECIUM BURSARIA

VANCE TARTAR AND TZE-TUAN CHEN

(From the Department of Zoology, University of California a.t Los Angeles and
the Department of Zoology, University of Vermont}

INTRODUCTION

The presence of distinct mating types in Paramecium has been dem-
onstrated by several investigators (Sonneborn, 1937; Kimball, 1937;
Jennings, 1938; Sonneborn, 1938; and Jennings, 1939). Under appro-
priate conditions individuals belonging to different mating types in the
same " group " will, when they are placed together, immediately agglu-
tinate and later form pairs. Such agglutination has been called the
" mating reaction." The present study is designed to answer the ques-
tion: Do enucleate fragments of Paramecium manifest mating reactions?

In this investigation we have studied, for comparison, the following
phenomena: (1) mating reaction between whole animals, (2) mating re-
action between nucleate x fragments and whole animals, (3) mating reac-
tion between enucleate l fragments and whole animals, (4) mating reaction
between enucleate fragments. The results of these comparative studies
will be reported in the order given.

Preliminary work on this problem was originally begun at the Osborn
Zoological Laboratory, Yale University, in collaboration with Dr. R. F.
Kimball ; but circumstances unfortunately prevented its completion there.
The present study was carried on largely at the University of California
at Los Angeles and completed at the University of Vermont. A pre-
liminary report appeared in Science, 91 : 246 (1940).

MATERIAL AND METHODS

Paramecium bnrsaria — the green Paramecium — is especially favor-
able for this study for several reasons. (1) As far as we know,
bursaria is the only species of Paramecium in which enucleate fragments
are viable. (2) The tendency of animals of this species to creep slowly
over the bottom of the container facilitates cutting with a glass needle
to such an extent that large numbers of fragments can be obtained.

1 In this paper the word " nucleate " is used to indicate the presence of both the
macronucleus and the micronucleus ; the word " enucleate " meaning the absence of
both nuclei.

130

MATING REACTIONS— FRAGMENTS OF PARAMECIUM 131

(3) The single micronucleus of this species (especially in the races used
in the present investigation) is large and stains deeply with hematoxylin.

(4) Stability of mating type permits one to obtain uniform and constant
material for study.

For this study two races of P. bursaria — McD~ and GrlA — were
used.2 They belong to two different mating types of Group II (Jen-
nings, 1939). Under suitable conditions they give a marked mating
reaction when placed together, and permanent pairs are later formed.

Race McD?j was collected near Baltimore, Maryland ; race GV14 from
south of Greensboro, North Carolina. McDz is a large race, while Grl4
is somewhat smaller. In both races the micronucleus is large and stains
deeply with hematoxylin. After such staining the micronucleus or a
piece of macronucleus could thus easily be detected if present in any
fragment.

The animals were cultured in essentially the same manner as de-
scribed by Jennings (1939). A number of cultures of each race were
kept in the laboratory, and only those which gave the strongest mating
reaction were used.

The animals to be operated upon were placed in a depression slide
and cut with a fine glass needle under a dissecting microscope. When
the needle passed directly through the mid-region of the animal, the
nuclei were usually seen to be extruded from one of the fragments ;
while if the needle cut to one side of the mid-region one large nucleate
and one smaller enucleate fragment resulted. Only fragments one-half
the size of the original animal or smaller were isolated for testing since
these were most likely to be enucleate.

When a definite mating reaction had been observed, the fragments
were fixed in hot Schaudinn's fluid (95 cc. Schaudinn's fluid and 5 cc.
glacial acetic acid) or in Benin's fluid at room temperature. They were
stained in iron-hematoxylin, destained in aqueous picric acid, and
mounted in damar.

OBSERVATIONS
Mating Reaction between Whole Animals

Jennings (1939) has described in detail the mating reaction between
whole animals in P. bursaria. We have, however, noted an additional
feature. When the area of contact of one animal with another is small
there appears a distinct flattening if not an appreciable indentation of
that part of the cell (Fig. 1). At present an explanation of this phe-
nomenon cannot be given, but at least the response indicates that the

2 We are indebted to Professor H. S. Jennings for these two races of P.
bursaria.

132 VANCE TARTAR AND TZE-TUAN CHEN

union between the animals involves more than a possible adhesion of the

FIG. 1. Mating reaction of whole animals. Two individuals in this case have
become attached to a third animal. In each animal there appears a flattening of
the contour of the body at the region of contact. (This drawing and others to
follow are free-hand sketches from the living material.)

cilia. It is significant that if members of a reacting pair (two whole
animals, or a whole animal and a fragment) are gently separated with
the glass needle, these flattenings or indentations (Fig. 3, g-i) do not
disappear at once but only gradually round out to the normal contour.

Mating Reactions between Nucleate Fragments and Whole Animals

When P. bursaria is cut transversely with the needle there usually
results one larger and one smaller fragment. It is the larger fragment
which contains the nuclei, the nuclear complex being visible as a clear
sphere in the living fragments. Shortly after cutting, such nucleate
fragments of McDs were placed with whole animals of GV14. The
mating reaction took place at once, and on the following day intimate
fusion of fragments with whole animals, similar to conjugation between
whole animals, was observed. Conversely, nucleate fragments of GV14
were placed with whole animals of McDs, and such mixtures gave the
same mating reaction and intimate fusion (Fig. 2) as described above.

FIG. 2. Large nucleate fragments of GV14 fused with McD3 whole animals.

(Samples of such pairs were stained with aceto-carmine and a full
nuclear complement was invariably found in the fragment.) It has not

MATING REACTIONS— FRAGMENTS OF PARAMECIUM 133

yet been determined whether the intimate fusion between the nucleate
fragment and the whole animal initiates nuclear changes in either or
both partners.

It should also be noted that the nucleate fragments not only exhibited
mating reactions immediately but also on the second and third days
after cutting. After the mixture was made and mating reaction noted,
it was placed aside undisturbed in a moist chamber. This mixture was
again examined on the second and third days following. Such nucleate
fragments as had not already fused in conjugation were again found to
be exhibiting the mating reaction. Apparently the nucleate fragments
which did not fuse with the whole animals gave repeated mating reac-
tions.3 Thus the nucleate fragments can react not only on the clay when
they are prepared, but also on subsequent days ; and this is quite in con-
trast to enucleate fragments which give the mating reaction on the day
of cutting but never subsequently, though they may be active two or
more days later (see below).

Viability of Enucleate Fragments

Enucleate fragments of P. bursaria may remain alive for as long as
four days, and during that time exhibit a surprisingly normal behavior,
swimming actively in one direction, alternating the direction of move-
ment, spiralling, and coming to rest adjacent to masses of food (Tartar,
1938). Observations on 100 enucleate fragments each of McD3 and
GV14 showed that practically 100 per cent of such fragments were alive
and active after one day, and 50 per cent after two days. Such hardi-
ness of enucleate fragments made possible the investigation herein de-
scribed. In the present study the enucleate fragments were tested within
half an hour after cutting.

Mating Reaction between Enucleate Fragments and Whole Animals

Enucleate fragments of either race were found to give the mating
reaction with whole animals of the other race (Fig. 3, a-/). Enucleate
fragments never agglutinate with fragments or whole animals of the
same race. The mating reaction was observed in 131 small enucleate
fragments of McD3 mated with whole animals of GV14. A sample of
fifteen of these fragments which reacted was fixed in Schaudinn's fluid
and stained with iron-hematoxylin to test whether nuclei or parts thereof
might be present in the fragments. In no case was a micronucleus or
a piece of macronucleus found in a fragment.

3 The races studied show a diurnal reactivity, discontinuing mating reactions in
the late afternoon and not reacting again until the morning of the following day.

134

VANCE TARTAR AND TZE-TUAN CHEN

In the reciprocal cross enucleate fragments of Grl4 were placed with
whole animals of McD3. Such a mixture gave the same agglutination
as described above. Twelve such fragments which reacted were stained
and no trace of a nucleus or part of a nucleus was found in them.

FIG. 3. Mating reactions between enucleate fragments and whole animals.
Note the flattening or indentation of the body of the animal at the point of union
with a fragment (Fig. 3, a-/). Whole animals which were separated from their
attachment with enucleate fragments still retain typical indentation (X) at the
points of union (Fig. 3, g-i).

Apparently in no respect does the mating reaction between enucleate
fragments and whole animals differ from that between whole animals.
Such similarity is shown by the following observations : ( 1 ) The react-
ing fragment agglutinates with the first animal of the other race with
which it happens to collide. Subsequently, one, two, three or more ani-

MATING REACTIONS— FRAGMENTS OF PARAMECIUM 135

mals of the other race may become attached to the fragment, thus form-
ing the typical agglutinated clump (Fig. 3/). (2) The fragment re-
mains securely attached to the whole animal, i.e., the two do not rotate
upon one another. The direct medium of this union is not the surfaces
of the fragment and the whole animal since the partners remain sep-
arated by the distance of their cilia. Yet the union is more intimate
than might be suspected, for at whatever region the fragment attaches
to the whole animal, a flattening or slight indentation of the normal con-
tour of the whole animal is there produced (Fig. 3). (3) The pairs
or clumps formed by the agglutination of fragments and whole animals
break up at approximately the same time as that at which clumps of whole
animals mated simultaneously break up into conjugating pairs and single
animals. This latter point is the conclusion from a separate study of
the duration of agglutination of enucleate fragments with whole animals
in which two additional groups of McD3 enucleate fragments (total
number, 90) were mated with Grl4 whole animals. All enucleate frag-
ments became separated from the whole animals 123 to 140 minutes
after the beginning of agglutination and never again united with them.
This was approximately the time required for simultaneously mixed
whole animals, which formed large clumps, to break up into pairs and
single animals (123 to 153 minutes).

The mating reaction between enucleate fragments and whole animals
is thus in these respects altogether normal, but it was never follozved by
trite fusion as in conjugation. It remains to be determined whether
contact with the enucleate fragment is sufficient stimulus to initiate
nuclear changes in the whole animals.

Although mating reaction is thus shown by enucleate fragments
tested promptly after cutting, this is not the case if the enucleate frag-
ments are kept for 24 hours before they are mixed with whole animals
of the other race. Enucleate fragments from a McD3 culture were
prepared at the time when tests showed the animals of this culture to
be most reactive with GY14. These McD3 enucleate fragments were
placed with Grl4 whole animals 24 hours after they were cut. None of
the 115 fragments so tested in delayed mixing showed any mating reac-
tion although they were alive and active. At the time of mixing the
enucleate fragments with the whole animals, control experiments showed
that the McD3 whole animals from the culture in question were again
strongly reactive.

Mating Reaction between Enucleate Fragments

For a study of reaction between enucleate fragments, one fragment
of each race was placed in a very small drop of culture fluid and the

136 VANCE TARTAR AND TZE-TUAN CHEN

pair observed. In approximately half of the cases the enucleate frag-
ments agglutinated at once when, during their rapid movements, they
first collided with one another (Fig. 4). Mating reaction between
enucleate fragments was under these conditions apparently not so strong
as between enucleate fragments and whole animals, for the pairs were
easily separated by jarring, and even when not disturbed they remained
attached for not longer than ten minutes. That the response is not a
mere chance adherence, however, is shown by the fact that it never
occurred between fragments of the same race even though these were
observed to collide with one another. Although flattening at the point
of contact was not observed in these small fragments, they swam together
as one animal without rotating upon one another, a behavior typical of

FIG. 4. Mating reaction between enucleate fragments. (These fragments were
later fixed and stained and were found to be enucleate.)

the mating reaction. After the two reacting fragments separated, they
frequently again agglutinated and became attached.

Thirty enucleate fragments (15 of each race) were tested as de-
scribed; and of these 16 gave the mating reaction. Each pair of
reacting fragments was separately fixed and stained in iron-hematoxylin.
In no case was a nucleus or part of a nucleus found in any of the
fragments.

Another method of observing the mating reactions between enucleate
fragments consists of introducing under the dissecting microscope one
fragment of one race into a drop containing several fragments of the
other race and to watch the introduced fragment continuously. This
procedure greatly increases the probability of the fragment finding a
partner. Under these circumstances two more cases of mating reaction
were observed : in one case the introduced J\IcD3 fragment remained

MATING REACTIONS— FRAGMENTS OF PARAMECIUM 137

attached to two GV14 fragments for five minutes ; in the other case a
McD3 fragment paired with a GV14 fragment for twelve minutes.

Still another method was to place three to six enucleate fragments
of one race (GV14) into a drop containing many enucleate fragments of
the other race (McD3). Out of a total of 22 Grl4 fragments so tested,
15 reacted. Typical mating reaction occurred; pairs of enucleate frag-
ments and groups of three enucleate fragments were observed. When,
as here, the drop of water in which the reaction is followed is of sufficient
size that evaporation does not interfere and the reacting fragments need
not be disturbed by replenishment of the drop, the conditions may
be said to be most nearly normal ; and under these circumstances pairs
of enucleate fragments and groups of three enucleate fragments were
found to remain continuously united in the mating reaction for as long
as 34 minutes.

Thus a total of 25 cases of mating reaction between enucleate frag-
ments has been recorded. Of these, 12 sample pairs of fragments were
carefully stained and no nucleus found in either fragment. In every
experiment, it is to be emphasized, the unmixed enucleate fragments of
each race were conscientiously watched en masse, and in no instance did
fragments of the same race react with one another though collisions
between them were frequent.

Thus the cytoplasm alone (in the absence of the nuclei} exhibits
the reactivity and diversity of mating type. Of course, this reactivity
may be due to the retention of influence of the nuclei which have just
been removed. This possibility is strongly suggested by the fact that
enucleate fragments lose their reactivity within a day and do not regain
it thereafter.

SUMMARY

1. Comparative studies were made on the following phenomena in
two races of .Paramecium biirsaria belonging to two different mating
types: (a) the mating reaction between whole animals, (&) the mating
reaction between nucleate fragments and whole animals, (c) the mating
reaction between enucleate fragments and whole animals, and (rf) the
mating reaction between enucleate fragments.

2. In the mating reaction between whole animals a phenomenon
hitherto unreported was observed : When two animals become attached
there is a flattening or even an appreciable indentation at the point of
union which is most striking when the area of contact is small.

138 VANCE TARTAR AND TZE-TUAN CHEN

3. Nucleate fragments of either race show mating reaction with
whole animals belonging to the other race. The mating reaction may
be followed by an intimate fusion, as in conjugation between whole
animals, between the nucleate fragment and whole animal.

4. Enucleate fragments of either race give the mating reaction with
whole animals of the other race. Mating reactions never occur between
enucleate fragments and whole animals of the same race. Mating reac-
tion between enucleate fragments and whole animals appears to be iden-
tical with that between whole animals, but it was never followed by
intimate fusion as in conjugation between whole animals.

5. The mating reaction also occurs between enucleate fragments be-
longing to two different races. Controls showed that mating reactions
never occur between enucleate fragments of the same race.

BIBLIOGRAPHY

JENNINGS, H. S., 1938. Sex reaction types and their interrelations in Paramecium

bursaria. I and II. Proc. Nat. Acad. Sci., 24: 112-120.

— , 1939. Genetics of Paramecium bursaria. I. Mating types and groups, their
interrelations and distribution; mating behavior and self sterility. Genetics,
24 : 202-233.

KIMBALL, R. F., 1937. The inheritance of sex at endomixis in Paramecium aurelia.
Proc. Nat. Acad. Sci., 23 : 469-474.

SONNEBORN, T. M., 1937. Sex, sex inheritance and sex determination in Para-
mecium aurelia. Proc. Nat. Acad. Sci., 23: 378-385.

— , 1938. Mating types in Paramecium aurelia : diverse conditions for mating
in different stocks ; occurrence, number and interrelations of the types.
Proc. Am. Phil. Soc., 79: 411-434.

TARTAR, V., 1938. Regeneration in the genus Paramecium. (Abstract.) Anat.
Record (Supp.), 72: 52.

TARTAR, V., AND T. T. CHEN, 1940. Preliminary studies on mating reactions of
enucleate fragments of Paramecium bursaria. Science, 91 : 246-247.

Vol. LXXX, No. 2

THE

April, 1Q41

BIOLOGICAL BULLETIN

PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY

THE TRANSPORT OF CO2 IN THE BLOOD OF CERTAIN

FRESHWATER FISHES

J. K. W. FERGUSON AND E. C. BLACK

(From the Department of Pharmacology, University of Toronto, and the
Department of Zoology, Swarthmore College}

The effect of carbon dioxide (and other acids) on the affinity of the
hemoglobin in fish blood for oxygen varies greatly from species to
species (Redfield, 1933). In some species not only the affinity of the
blood for oxygen, but the oxygen capacity too, is greatly diminished by
relatively low pressures of CO, (Root, 1931). The physiological and
ecological implications of great differences in affinity for oxygen and
sensitivity to CO2 have become more apparent as a result of a series of
studies on freshwater fish, which are summarized in a recent paper by
Black (1940). A few of the findings may be recapitulated briefly as
follows. Carbon dioxide decreases the affinity of the blood for oxygen
to a greater degree in those fish inhabiting deeper and colder water.
The same bloods have also a lower affinity for oxygen at minimal pres-
sures of CO2. These two characteristics would act to offset the effect
of low temperatures, which is to lower the pressure at which oxygen is
available to the tissues. Another manner in which sensitivity to CO2
may be useful to fish inhabiting deep water, is in the regulation of
buoyancy at different depths, by facilitating the formation in the swim-
bladder of gases rich in oxygen, as suggested by Haldane (1922) and
Hall (1924).

The present study was designed with two objects in mind. The first
was to investigate the mechanisms by which the great differences in
sensitivity to CO2 are achieved. The second was to determine the
probable range of physiological tensions of CO2 in a series of fresh-
water fish comprising some of those used in the foregoing studies.
This information is difficult to obtain with any degree of accuracy in
fish, but even approximate determinations may be of use in evaluating
some of the relationships described above.

In pursuit of the first object, a detailed study of the characteristics

139

140 J. K. W. FERGUSON AND E. C. BLACK

of CO, transport was made on the blood of two species differing widely
in the sensitivity of their blood to CO2, namely rainbow trout, Salino
gairdnerii Richardson, and carp, Cyprinus carpio Linnaeus. Trout are
among the fish most sensitive to CO2 and carp, while not the most insen-
sitive, are relatively so. Availability in sufficient numbers to supply
enough blood was an important consideration in the choice of these
species. Data on physiological CO2 tensions were obtained on the same
species as well as on a few specimens of bullhead, Amemrus nebulosus
Le Sueur, and sucker, Catostomus commersonnii Lacepede, which hap-
pened to be available. The blood of the bullhead is even less sensitive
to CO2 than that of the carp, while those of the sucker and the trout are
of about the same order of sensitivity.

METHODS

Blood was drawn from the heart into a syringe containing heparin.
No fluoride or oxalate was used, since these have been found to cause
progressive swelling of the erythrocytes and eventual hemolysis in certain
fish bloods (Black and Irving, 1938; Hamdi and Ferguson, 1940). For
the construction of dissociation curves it was necessary to pool the blood
from several fish. Equilibration was done at 15° C., except in specified
cases. The blood was kept on ice throughout the day. Samples were
equilibrated one at a time for fifteen minutes in tonometers of the
original Barcroft type (1914). A centrifuge tube was attached to the
open end of the tonometer by rubber tubing. After equilibration the
blood was collected in the centrifuge tube and separated from the gas
phase by clamping the rubber tubing. The gas was analyzed for CO2
and O2 in a Haldane apparatus. The blood in the centrifuge tube was
covered with liquid paraffin and samples were removed without delay
for analysis of CO2 and O2 by the manometric method of Van Slyke;
of chloride by the open Carius method ; and of water content by drying
at 105° C. Another sample was centrifuged in a capillary tube for
twenty minutes at a centrifugal force of about 3000 times gravity for the
estimation of packed cell volume. The rest of the blood was also
centrifuged under oil for 20 minutes and samples of plasma removed
for the analyses listed above. A number of experiments were performed
to evaluate the error introduced by metabolism of the blood during the
foregoing procedures. Hourly analyses were done on blood kept in
sealed syringes at 15° C. and on ice. Trout blood was found to have
a somewhat higher metabolic rate than carp blood, and the highest rate
found in trout blood at 15° C. was 1 cc. of oxygen consumed and 1 cc.
of CO2 produced per 100 cc. blood per hour. The only estimation which

TRANSPORT OF CO2 IN BLOOD OF FRESHWATER FISHES 141

might have been appreciably affected by even this highest metabolic rate
was the estimation of cell volume in which centrifuging was done at
25° C. or thereabouts for 20 minutes. It was possible in the later
experiments on trout blood to minimize this source of error by reducing
the time of centrifuging to 5 minutes in a high speed hematocrit centri-
fuge. The results with both long and short times of centrifuging were
in substantial agreement.

In trout blood kept on ice, the highest metabolic rate was found to
be 0.5 cc. per 100 cc. per hour. This figure may be used to estimate
the maximum correction for the figures on O2 and CO2 content of the
venous blood of trout, because the trout blood was kept on ice for three
or four hours in transit from the hatchery to the laboratory.

Some experiments were done to determine the change in CO2 capacity
with time. No changes were found in blood kept on ice for as long as
12 hours. When the blood was kept at 10° C. in a refrigerator for 24
to 48 hours, small changes occurred. It is interesting that the CO2
capacity increased in blood (both carp and trout) kept fully oxygenated,
but decreased in blood which was partly or fully reduced, suggesting
something like a Pasteur reaction.

The gas content of mixed venous blood, drawn from the heart into
an oiled syringe containing heparin, was determined by analysis without
exposure to air. Three portions of the same sample were then equi-
librated with different gas mixtures and subsequently analyzed. One
portion was equilibrated with about 2 mm. CO2 in air. This gave the
oxygen capacity and one point on the CO2 dissociation curve. Another
portion was equilibrated with 8-12 mm. CO2 in air and the third portion
with 8-12 mm. CO2 in N2. These gave the positions of the CO2
dissociation curves of oxygenated and reduced blood respectively. The
equilibrations in these particular experiments were done at the tem-
perature of the water from which the fish were taken. Knowing the
oxygen content of the sample of venous blood and the oxygen capacity
of the same sample, it was possible to estimate the position of the CO2
dissociation curve of the venous blood between those of the fully oxy-
genated and fully reduced bloods. Then knowing the CO2 content of
the venous blood, its CO2 tension could be read off on the abscissa with
an error probably not more than 2 or 3 mm. Hg.

During the withdrawal of the venous blood, the gills were not aerated,
but only samples which flowed freely into the syringe at the first stab
were used. Since even the trout survived this procedure without ap-
parent harm, the estimated tensions may be regarded as well within the
limit of tolerance of the fish, though probably above the average for the
resting state.

142

J. K. W. FERGUSON AND E. C. BLACK

OBSERVATIONS
Physiological Gas Tensions

The results on the gas contents and tensions in circulating blood are
presented first, since they indicate the part of the CO2 dissociation curves
of greatest physiological importance. Table I gives the CO, content,
percentage saturation with oxygen and estimated CO2 tensions of sam-
ples of venous blood drawn from four species of fish. Only the trout
blood could have changed significantly between the time of drawing the
blood and the time of analysis. If maximum corrections for metabolism
are applied to the results on trout blood, the CO2 tension would be about

TABLE I

Venous blood gases

Trout

Temperature

CO2

P CO2

°C.

v.p.c.

per cent sat.

mm. Hg

21.0

24.8

19.4

22.8

Carp

36.4

31.2

28.6

Sucker

47.8

36.3

Catfish

21.4

1 mm. lower, the CO, contents about 2 vols. p.c. lower and the oxygen
contents about 2 vols. p.c. higher.

No measurements were made on the aerated blood from the gills,
but a rough estimate of the CO., tension in aerated blood can be made
in the following manner. Trout D in Table I is favorable for purposes
of illustration, because it had practically the same O, and CO, capacity
as the composite blood F in Fig. 1. If the blood of trout D were fully
oxygenated, it would gain about 11 vols. p.c. of 0. Assuming an R.Q.
of unity, the CO, lost would be 11 vols. p.c., leaving a CO, content of
about 8 vols. p.c. A CO, content of 8 vols. p.c. corresponds on the
curve for oxygenated trout blood in Fig. 1 to a CO, tension of about 3.5
mm. Similar calculations for all the bloods in Table I gave estimated

TRANSPORT OF CO, IN BLOOD OF FRESHWATER FISHES 143

CO2 tensions of 3-5 mm. for fully oxygenated " arterial " blood. These
figures indicate that the most important range of physiological CO2
pressure is between 3 and 10 mm. Hg in all four species. It is a matter
of considerable interest that the physiological range is so similar in fish
representing the extremes of variation in the effect of CO, on the
combination of oxygen in the blood. It may be concluded that the
consequences of differences in sensitivity to CO, are not evaded to any
extent, as might conceivably be the case if different species maintained
themselves at significantly different CO, pressures.

Another implication of the foregoing results is that CO, tensions
above 10 mm. could only be attained in any of these fish under condi-

VOL5
PC.

40-

+ - TRUE PLASMA OF REDUCED BLOOD
• - REDUCED WHOLE BLOOD
° = OXYGENATED WHOLE BLOOD

20 JO 40 50 60

P. C02 MM. He.

FIG. 1. CO2 dissociation curves of pooled blood of rainbow trout at 15° C.

tions of oxygen debt. Higher tensions might occur, however, in certain
tissues as a result of acid production locally. Consequently the rather
low CO., tensions found in mixed venous blood need not be regarded as
the maximal CO, tensions which may operate in the production of gases
in the swim-bladder in deep water.

CO, Dissociation Curves

Figure 1 shows composite curves of two batches of blood each from
twelve rainbow trout. They resemble the curves obtained by Root
(1931) and Root and Irving (1940) on various marine fish in the con-

144

J. K. W. FERGUSON AND E. C. BLACK

vergence of the curves for oxygenated and reduced blood at the higher
pressures of CO2. It can be seen in Table II that the oxygenated blood
is not fully oxygenated at the higher pressures of CO2 even though in
some cases it was exposed to pressures of oxygen as high as 670 mm. Hg.

The apparent disappearance of the effect of oxygenation on CO2
capacity (Haldane effect) at the higher pressures of CO2 can be due
only in part to incomplete oxygenation of the oxygenated blood because
even at the high pressures it is still half saturated with O2.

In Fig. 2 are shown the CO., dissociation curves of carp blood. Two
sets of curves are given representing the range of variation found in six
batches of carp blood. The greater variation shown by the carp may
be due in part to the use of fewer fish for each batch of blood (the carp

TABLE II

Complete data on blood of trout F at 15°C. To calculate combined 02 (Hb O2) the
Bunsen solubility coefficient of oxygen in the blood at 15° is assumed to be 0.036. The
meaning of r COz and r Cl is explained in the text.

COz content

Chloride

Water content

P CO2

PO2

HbOj

rCOs

Cell

rCl

Whole
blood

Plasma

Whole
blood

Plasma

Whole
blood

Plasma

mm. Hg

v.p.v.

v.p.c.

per cent

m. eq./l.

m. eq.jl.

g./WO g.

g./lOO g.

2.2

Air

6.1

8.0

9.9

0.33

32.8

—

140.0

0.60

86.0

94.5

2.7

11.9

13.8

0.0

0.71

33.0

116

139.5

0.61

86.0

94.6

10.7

124

16.3

18.0

7.3

0.90

35.3

118

140.0

0.65

86.5

95.5

12.5

23.8

26.4

0.0

0.95

47.0

117

136.0

0.81

85.9

94.3

32.7

168

28.6

31.5

5.0

0.94

43.0

116

138.0

0.78

86.0

94.9

38.5

31.8

34.2

0.0

1.02

44.0

117

139.5

0.77

85.8

94.5

61.0

675

37.3

40.6

5.5

0.94

37.4

116

139.0

0.68

85.3

94.6

70.5

40.4

44.0

0.0

0.98

43.8

116

138.0

0.77

84.8

95.0

being larger fish), and consequently less averaging out of individual
variations. The curves for carp differ from those of the trout in two
important respects. Firstly they are higher, indicating a higher pH at
a given pressure of CO2. Secondly the curves of oxygenated and re-
duced bloods are widely separated and show only a slight tendency to
converge at higher CO2 pressures.

The higher CO2 capacity of carp blood indicates a higher pH which
must be attributed to the regulation of the acid-base balance of the fish
as a whole at a more alkaline level. The CO2 dissociation curves of
true plasma in both species lie above the curves for whole blood, in this
respect resembling mammalian blood rather than dogfish blood (Fergu-
son, Horvath and Pappenheimer, 1938). They indicate a higher con-
centration of CO2 in the plasma than in cells at the same pressure of

TRANSPORT OF CO2 IN BLOOD OF FRESHWATER FISHES 145

CO2. The distribution of CO2 between cells and plasma is more pre-
cisely indicated in Table II in the column headed r CO2.

Distribution of Electrolytes

The distribution ratio r CO2 is the concentration of combined CO2
per gram of cell water divided by the combined CO2 per gram of plasma
water. Combined CO2 is calculated by subtraction of the physically dis-
solved CO2 from the total CO2, using the factors 0.125 and 0.105 for
plasma and cells respectively, which multiplied by the PCO2 in mm. Hg
give the concentrations of dissolved CO2 in volumes per cent. RC1 is

VOLS
PC

60-
50
40
30
20
10

CARP /

= TRUE PLASMA OF REDUCED BLOOD
• - REDUCED WHOLE BLOOD
0 - OXYGENATED WHOLE BLOOD

10 20 30 40
P. COa MM. He.

50 60

FIG. 2. CO, dissociation curves of carp blood at 15° C. Of six batches of pooled
carp blood, E had the highest CO2 capacity and I the lowest.

the analogous distribution ratio for chlorides. Little weight should be
given to individual values of the distribution ratios because of the large
number of possible errors that are introduced in their computation. On
the average, however, it is evident that in carp blood the values of r CO2
do not tend to exceed those of r Cl as they do in mammalian blood where
the higher value of r CO2 is probably due mostly to the presence of
carbamino compounds of CO2 with hemoglobin (Roughton, 1935). In
the trout blood the values of r CO2 do tend to exceed somewhat the
values of r Cl. Thus the distribution ratios provide no evidence for
the presence of carbamino compounds in the carp erythrocytes, but would
be consistent with the presence of a small amount in trout cells. They

146

J. K. W. FERGUSON AND E. C. BLACK

are consistent with the hypothesis that the combined CO2 in red cells is
largely in the form of bicarbonate and that the bicarbonate and the
chloride are partitioned between the red cells and plasma according to a
Donnan distribution. The figures on plasma chloride of carp blood in
Table III give further evidence of a Donnan equilibrium with chloride
and bicarbonate ions diffusible. The plasma chloride decreases regu-
larly (except for one figure) with each increase in plasma CO2 con-
sistent with a migration of chloride into the cells with increasing acidity.
Plasma chloride in trout blood also shows a decrease with increasing
CO2, but in a very irregular fashion which, however, acquires more
meaning when considered with the changes in packed cell volume.

TABLE III

Complete data on blood of carp I at 15°C.

COz content

rvii

Chloride

Water content

PCO2

PO2

Hb02

rCO2

rCl

Whole
blood

Plasma

Whole
blood

Plasma

Whole
blood

Plasma

mm. Hg

v.p.c.

per cent

m. eg. /I.

g-1100 g.

g./WO g.

3.4

158

12.9

15.2

12.3

0.77

39.6

124.5

146.0

0.94

82.5

94.8

3.7

2.3

19.8

23.8

1.8

0.77

41.2

125.0

144.0

0.88

83.9

94.8

10.5

148

17.0

21.2

12.7

0.61

38.8

125.0

146.5

0.90

84.1

95.1

12.6

29.3

33.7

0.4

0.88

40.6

125.5

142.5

0.91

83.5

94.6

32.2

145

27.9

31.6

10.8

0.90

39.5

125.5

144.5

0.85

83.2

95.0

36.6

41.1

45.9

0.95

40.2

126.0

141.5

0.94

83.4

94.5

66.2

224

38.1

42.2

10.8

0.98

40.9

125.0

142.0

0.92

83.6

94.4

63.2

45.0

51.3

0.88

41.2

124.5

140.5

0.95

82.7

95.2

Cell Volumes

In Fig. 3 the packed cell volumes of carp and trout blood are plotted
against pressure of CO2. In carp blood the cell volume decreases
slightly with the first increment of CO2 pressure, but with further in-
crease in P CO, the volume increases as in mammalian blood, and by
about the same amount for each increment of CO2 combined, namely
by about 5 p.c. of their volume for an increase of 10 m.eq. per liter of
combined CO2. The cells of TJ.F. (Henderson, 1928) increased by
6 per cent for an increase of 10 m.eq. of CO2 combined.

The volume changes for trout blood shown in Fig. 3 are the com-
posite data from three batches of blood. The trout cells show greater
changes in volume than do the carp cells. They reach their maximum
size between 10-20 mm. Hg of CO2 pressure. The reduced cells show
a much greater increase in volume, but at pressures above 20 mm. Hg
they decrease again with increasing P CO2. It can be seen in Table II

TRANSPORT OF CO2 IN BLOOD OF FRESHWATER FISHES 147

that the largest cell volume corresponds with the lowest concentration of
chloride in the plasma, suggesting that a migration of chloride from the
plasma, presumably into the cells, has occurred. The extent of this
migration is depicted in Fig. 4, where the chloride content in the plasma
of a liter of blood is plotted against the bicarbonate content of the
plasma of a liter of blood. The contents are calculated by multiplying
the concentrations per liter of plasma by the fraction of the whole blood
volume which is plasma. In carp blood it appears that for each increase
of 1 eq. of bicarbonate in the plasma about 0.75 eq. of chloride enters
the cells. This ratio is approximately that for mammalian blood (Van

£35

I-

• = OXYGENATED BLOOD
0 = REDUCED BLOOD

.---_ • CARP I •

o --. o

CARP F

TROUT A e,. f

-SK//

10 20 30 40 50
P. CC}, MM. HG.

FIG. 3. Packed cell volumes in carp and trout blood are plotted against the
pressure of CO2. The large changes in cell volume in trout blood are in striking
contrast to the small changes in carp blood.

Slyke, 1921). In trout blood, however, the loss of chloride from the
plasma at the maximum cell volume greatly exceeds the increase in CO,
in the plasma. This suggests that the excessive cell volumes and exces-
sive chloride shift may be due to the production in the blood of an acid
other than CO,. It is immaterial where this acid is produced, but it is
necessary to postulate that it is diffusible through the red cell membrane.
It is also necessary to postulate that it is produced by a reversible reac-
tion and that the equilibrium point is determined by the tension of
oxygen and of CO, in the blood. The optimum conditions for its
formation would presumably be at low oxygen tensions and a CO2
tension between 10 and 20 mm. To explain the cell volume changes by

148

J. K. W. FERGUSON AND E. C. BLACK

this production of acid alone would require the formation of about 30
m.eq. per liter of acid other than CO2. Such an hypothesis certainly de-
serves the utmost skepticism till the changes in cell volume can be checked
by a method other than centrifuging, but it does receive some support of
a qualitative nature from another and independent consideration, namely
the effect of oxygenation on CO2 capacity.

Oxygenation and CO2 Capacity

When the effect of oxygenation on CO2 capacity is expressed as
- ABHCO3/AO, at constant plasma pH the maximum values for this

o
LJ

Oxy. Carp Blood A
Red. Carp Blood A

Red. Trout Blood

Oxy. Trout Blood o

M.Eq. BHCQ,

FIG. 4. The chloride in the plasma of a liter of blood is plotted against the
bicarbonate in the same volume to illustrate the magnitude of the migration of
chloride into the cells. In carp blood ABHCO3/AC1 = — .75 as in mammalian
blood, but in trout blood the ratio is much greater than unity, suggesting that an
acid other than CO2 has been neutralized in the cells and diffused into the plasma
in exchange for chloride ions.

ratio are found at a pH of about 7.3 in both bloods. The maximum
value in each is about 1.2. This value is higher than any of those in
other species compiled by Redfield (1933), as it should be, in accordance
with the principle that the greater the effect of acid on the combination
of oxygen the greater should be the effect of oxygenation on the dis-
sociation of hemoglobin as an acid. But the fact that the value is the
same in trout and carp blood is apparently inconsistent with the prin-
ciple. If, however, the hypothesis of "extra acid " production in trout

TRANSPORT OF CO, IN BLOOD OF FRESHWATER FISHES 149

blood is correct, the apparent inconsistency would be only apparent, for
if an acid other than CO2 were produced in the blood at low pressures
of oxygen, ABHCO3 would not be a complete measure of the change in
base combined with hemoglobin, and hence would give too low an esti-
mate of the effect of oxygenation on the acid dissociation of the hemo-
globin in trout blood.

Buffer Power

The buffer power (ft) of whole blood is often represented by the
ratio — ABHCO3/ApHs, where ABHCO3 is the change in CO, com-
bined in whole blood and pHs the pH of the plasma or serum. This
procedure will be erroneous in trout blood if the hypothesis of the extra
acid production is correct. It is interesting, however, to make the cal-

TABLE IV

Buffer power of reduced bloods.

Range of
plasma pH

Range of
P COa

Buffer power
03)

Carp 15°

8.11-7.36
7.36-6.96

2-20
20-60

1.6
1.4

Trout 15°

7.90-7.23

2-20

2.3

7.23-6.85

20-60

1.7

"Human 38°

7.5-7.2

35-90

2.3

* Peters and Van Slyke, Quantitative Clinical Chemistry, Vol. I.

/3 = — ABHCOs/ApH/Hb. Concentrations are in milliequivalents per liter.

culation in the conventional manner and then to consider what change
in conclusion would be necessary if the hypothesis of extra acid produc-
tion were correct. The calculated buffer powers • of carp, trout and
human blood adjusted to equal concentrations of hemoglobin are given
in Table IV. Figures are given for reduced blood only, since fully
oxygenated trout blood cannot be obtained over a large enough range
of pH. Carp blood has about the same buffer power over the range
studied, but trout blood shows a greater buffer power over the more
physiological range of P CO2 of 2-20 mm., where it is equal to the
buffer power of human hemoglobin. Over the higher range of CO2
pressure the buffer power of trout hemoglobin is less, though still higher
than that of carp blood. The trout blood shows a concentration of its
buffer power in the physiological range, a characteristic which may be
attributed tentatively to closer grouping of the dissociation constants of

150 J. K. W. FERGUSON AND E. C. BLACK

the acid (or basic) groups in this range. If the anomalous increase and
decrease in cell volume in trout blood represents increase and decrease in
extra acid, the true buffer power over the range 2-20 mm. would be even
greater, and even less over the range 20-60 mm. In other words, the
tendency for the buffer power to be concentrated in a narrow range in
the trout blood would be even more striking.

Plasma pH

Values of plasma pH at CO2 pressures of 2, 20 and 60 mm. Hg
calculated from the smoothed data of the dissociation curves, assuming
a pK/ of 6.2 are shown in Table V. The changes in plasma pH may
be taken as paralleling fairly closely the changes in cellular pH. To
calculate these separately would merely introduce the uncertainties of
arbitrary values for pK/ in cells.

The main points of interest are: (1) an increase of P CO2 from
2 mm. to 20 mm. produces a greater change in pH in carp blood than

TABLE V

pH of true plasma of oxygenated blood calculated from the data of Tables II and III.

P CO2 Carp Trout

mm. Hg

2 7.91 7.66

20 7.23 7.15

60 6.84 6.84

in trout blood ; (2) at P CO2 of 60 mm. the carp blood is as acid as the
trout blood. In other words, the loss in oxygen capacity in trout blood
cannot be attributed either to a greater change in acidity for a given
increase in P CO., or even to a higher absolute acidity at the higher
pressures of CO2. However, the lower acidity in carp blood at lower
pressures of CO2 must be held partly responsible for the higher affinity
of the blood for oxygen in the absence' of CO2.

DISCUSSION

The results as a whole indicate that the great differences in the effect
of CO2 on the combination of oxygen in these two bloods and their
affinity for oxygen may be due to adaptations at three levels of physio-
logical organization. These seem to be : ( 1 ) specific differences in the
hemoglobin molecule; (2) differences in the environment provided for
the hemoglobin by the erythrocyte; (3) differences in the acid-base regu-
lation of the fish as a whole. It seems likely that differences in the

TRANSPORT OF CO, IN BLOOD OF FRESHWATER FISHES 151

hemoglobin molecules will prove to be the most important element in the
total adaptation, although no comparison of the two hemoglobins in
solution has yet been made. Until such a comparison is made, it cannot
be said that differences in the erythrocytes may not be equally important.
Certainly one of the most striking differences between the bloods has
been the behavior of the cell volumes. The effect of the intracellular
environment on the affinity for oxygen is marked, even among mammals,
and varies from species to species (Hill and Wolwekamp, 1936) in a
manner as yet inexplicable. In fish blood the effect of hemolysis on
affinity for oxygen is also large and cannot be explained by changes in
acidity alone (Root and Irving, 1940). The anomalous behavior in
trout blood of cell volumes, plasma chlorides and effect of oxygenation
on CO., capacity could all be explained, at least in part, by the production
(by a kind of Pasteur reaction) of acid at low tensions of oxygen and
an optimal tension of CO2. The production of extra acid cannot, how-
ever, explain the loss of oxygen capacity or the convergence of the CO2
dissociation curves of oxygenated and reduced blood. At most it could
only be a mechanism augmenting the effect of CO2 in maintaining a
high tension of oxygen at low contents of oxygen in the blood of trout.

SUMMARY

The venous blood from four species of freshwater fish, rainbow trout
(Salmo gairdnerii Richardson), carp (Cyprinus carpio Linnaeus), bull-
head (Ameiurus nebulosus Le Sueur) and sucker (Catostomus commer-
sonnii Lacepede) was analyzed and an estimate made of the probable
range of physiological CO2 tension.

A detailed study was made of CO2 transport in the blood of two of
these species, the rainbow trout and the carp, which differ greatly in the
effect of CO2 on the combination of oxygen in the blood. They differ
too, in their systems of CO2 transport. A curious feature of trout blood
is the great change in packed cell volume with changes in the pressure
of O, and CO2. An hypothesis is presented to explain in part these
anomalous changes in cell volume and other characteristics of the trout
blood. Carp blood shows less differentiation from general vertebrate
characteristics.

LITERATURE CITED

BARCROFT, J., 1914. The Respiratory Function of the Blood. Cambridge.
BLACK, E. C, 1940. The transport of oxygen by the blood of freshwater fish.

Biol. Bull, 79: 215-229.
BLACK, E. C., AND LAURENCE IRVING, 1938. The effect of hemolysis upon the

affinity of fish blood for oxygen. Jour. Cell, and Comp. Ph\sioL, 12:

255-262.

152 J. K. W. FERGUSON AND E. C. BLACK

BLACK, E. C., F. E. J. FRY AND W. J. SCOTT, 1939. Maximum rates of oxygen
transport for certain freshwater fish (abstract). Anat. Rec., 75: (Sup-
plement) 80.

FERGUSON, J. K. W., S. M. HORVATH AND J. R. PAPPENHEIMER, 1938. The trans-
port of carbon dioxide by erythrocytes and plasma in dogfish blood. Biol.
Bull, 75 : 381-388.

HALDANE, J. S., 1922. Respiration. New Haven.

HALL, F. G., 1924. The functions of the swimbladder of fishes. Biol. Bull, 47 :
79-126.

HAMDI, T. M., AND J. K. W. FERGUSON, 1940. Hemolytic action of fluorides on
certain nucleated erythrocytes. Proc. Soc. Exp. Biol. and Med., 44: 427-
428.

HENDERSON, L. J., 1928. Blood; a Study in General Physiology. New Haven.

HILL, R., AND H. P. WOLVEKAMP, 1936. The oxygen dissociation curve of haemo-
globin in dilute solution. Proc. Roy. Soc. London, B, 120 : 484-495.

PETERS, JOHN P., AND DONALD D. VAN SLYKE, 1932. Quantitative Clinical Chem-
istry. Volume I. Interpretations. Volume II. Methods. Baltimore.

REDFIELD, ALFRED C., 1933. The evolution of the respiratory function of the
blood. Quart. Rev. Biol, 8: 31-57.

ROOT, R. W., 1931. The respiratory function of the blood of marine fishes. Biol
Bull, 61 : 427-456.

ROOT, R. W., AND LAURENCE IRVING, with the assistance of Virginia Safford and
Henry Brown, 1940. The influence of oxygenation upon the transport of
CO2 by the blood of marine fish. Jour. Cell, and Comp. Physiol, 17 :
85-96.

ROUGHTON, F. J. W., 1935. Recent work on carbon dioxide transport by the blood.
Physiol Rev., 15: 241-296.

VAN SLYKE, DONALD D., 1921. The carbon dioxide carriers of the blood. Physiol
Rev., 1 : 141-176.

THE DEVELOPMENT OF THE ASCIDIAN EGG CENTRI-
FUGED BEFORE FERTILIZATION

TI-CHOW TUNG,1 SU-HWEI KU AND YU-FUNG-YEH TUNG

(From the Laboratory of Histology and Embryology, College of Medicine,
National Central University, China)

The development of centrifuged eggs of ascidians has been studied
by Duesberg (1926) and by Conklin (1931). In Styela and Ciona
Conklin found that when the eggs were centrifuged after fertilization,
three substances, namely the mitochondria, the hyaloplasm, and the yolk
may be displaced from their normal positions into three zones. If these
substances were so held until the beginning of cleavage, they were
distributed abnormally to the blastomeres. The tissues or organs of
the embryos derived from these eggs were also dislocated.

The unfertilized eggs of Styela and Ciona also were centrifuged by
Conklin. He noted that under strong centrifuging the spermatozoon
frequently did not enter the egg at all. No description of the develop-
ment of these eggs was given.

Recently Dalcq (1932, 1935, 1938) has fragmented unfertilized eggs
of Ascidiella scabra into two parts and then fertilized them. He found
that meridional halves of egg fragments may give rise either to appar-
ently normal and symmetrical larvae, or to lateral half larvae; while
larvae obtained from latitudinal halves may be deficient in one or more
kinds of tissues according to the level of the cut. These experiments
indicate that the unfertilized egg already possesses germinal localizations;
but at the same time it has a great capacity for regulation, as has been
demonstrated by Tung (1934) for the fertilized egg.

To gain further light on the organization of the unfertilized egg of
the ascidian and to learn whether, with sufficient force, the organ-forming
substances might be dislocated, we have made a study of the develop-
ment of Ciona eggs strongly centrifuged before fertilization. The ex-
periments on which this investigation is based were performed in the
summer and autumn of 1936.

1 This work was done while the senior author was on the tenure of a grant
from the Board of Trustees for the administration of the indemnity funds re-
mitted by the British Government for which he wishes to express his gratitude.

153

154 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

MATERIAL AND METHODS

The observations recorded here were all made on eggs of dona
intestinalis obtained in the vicinity of Tsingtao. During the months
from June to October the gonoducts of Clona are usually full of ripe
eggs and spermatozoa. The eggs were obtained free from sperm by
removing the integuments and carefully opening the oviduct which is
near the surface. The eggs were then removed from the oviduct with
a pipette and transferred to a glass dish containing fresh sea water. A
part of the eggs was used for centrifuging, leaving the remaining eggs
in the dish as a control.

The eggs to be centrifuged were placed in glass tubes with 10 cc.
sea water and rotated for ten minutes to one hour and fifty minutes at
a speed ranging from 2000 r.p.m. to 3800 r.p.m. These speeds repre-
sented centrifugal forces of about 716 X gravity and 2585 X gravity.
In order to prevent eggs from rotating during centrifugation, capillary
tubes were used as by Conklin. After centrifuging, the eggs were im-
mediately removed from the capillary tubes and fertilized by sperm of
another animal.

Individual eggs showing an abnormal distribution of ooplasmic sub-
stances to the first two or four blastomeres were picked out, sketched,
and placed in separate dishes of fresh sea water in order to study in
detail the location of various tissues in later development.

The larger part of the eggs and embryos was fixed in Bouin's fluid
and double-embedded in agar-parafmi (Chatton, 1927). Sections were
cut at 7 /A and stained with iron-haematoxylin, eosin, and light green.
Some of the material was fixed in Flemming's solution and mounted in
toto in order to determine the location of mitochondria.

RESULTS
Stratification of the Egg

The degree of the stratification of ooplasmic substances of the unfer-
tilized egg varies with the rate and duration of centrifugation. In
strongly centrifuged eggs three zones can be distinguished. These are :
(1) an alveolar or light zone at the centripetal pole; (2) a middle clear
zone containing the nuclear elements; and (3) a lower, heavy zone of
yolk. In eggs fixed in Flemming's solution and mounted in toto, the
light zone consists exclusively of large black, densely packed granules.
On the basis of the staining reaction, it corresponds obviously to the
" mitochondria " zone of Duesberg. Sections of the eggs show that
this zone is composed of an alveolar substance in which are embedded

DEVELOPMENT OF CENTRIFUGED ASCIDIAN EGG

155

a large number of blue-black granules ; these are taken to be mitochondria
(Fig. 1).

Under prolonged centrifuging, a fourth zone appears at the cen-
trifugal pole. Figure 2 shows an egg in which the four zones are

•^0^°®?^-°°^°

O.^:.O:P.P

^.o^o^o:?^? W#to P.b:'
'•a°5aSls?o£-<£

FIG. 1. Section of an unfertilized egg, centrifuged for one hour at 2193
times gravity, and fixed ten minutes later, showing a stratification of the egg
contents into three zones.

FIG. 2. Section of an unfertilized egg centrifuged for one hour at 2193 times
gravity and fixed ten minutes later, showing the fourth zone at the centrifugal pole.

FIG. 3. Section of an unfertilized egg, centrifuged for one hour at 2193 times
gravity and fixed fifteen minutes later. The stratification of the centrifugal zone
is not complete.

clearly separated. The centrifugal zone contains blue-black granules,
while in the centripetal zone the number of the same kind of granules
is greatly reduced. Figure 3 shows an egg with incomplete stratifica-
tion of the fourth zone. In this case a part of the granules remains

156 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

scattered throughout the clear and yolk zones and especially in the
latter. A comparison of the eggs illustrated in Figs. 1, 2, and 3 leaves
no doubt that the granules of the centrifugal zone are separated out
from the alveolar substance of the centripetal zone in which they were
formerly embedded.

Conklin described a fourth zone between the hyaline and yolk zones
and concluded that ". . . It (the substance of the zone), rather than
the mitochondria, is the myoplasm or formative substance for the future
muscles." Our conclusions are that it is the alveolar substance of the
centripetal zone that is the formative substance of the muscles ; the
alveolar substance is similar to the cytoplasm of muscle cells of the young
embryo both in structure and in staining reaction, whereas the granules

'•^•jim^F-;

"".fiSSSSSLSp

2oo

o°oo;

»^ •** <i
ooo

rt i*~i r^i /

FIG. 4. Section of a polyspermic egg centrifuged for one hour at 2193 times
gravity and fixed fourteen minutes after fertilization. Note sperm asters in clear
and yolk zones.

FIG. 5. Section of a polyspermic egg centrifuged for one hour at 2193 times
gravity and fixed fourteen minutes after fertilization. Note sperm asters in clear
and alveolar zones.

of the centrifugal zone are the mitochondria since in normal fertilized
eggs similar granules are found in the lower hemisphere at which the
sperm enters. In Physa heterostropha, Clement (1938) has recently
shown that the mitochondria which stratify between the clear and yellow
zones and are therefore heavier than the clear protoplasm are the last
to be segregated under centrifugal forces. In these respects our obser-
vations on Ciona agree with those of Clement on PJiysa.

Immediately after centrifuging, the eggs were cross-fertilized with
fresh sperm. As a check, in each experiment, the control eggs left
unfertilized in the original dish were examined. As has recently been
reported by Morgan (1938), cleavages were rare, showing that self-
fertilization occurs rarely in Ciona. Conklin found that in eggs which

DEVELOPMENT OF CENTRIFUGED ASCIDIAN EGG 157

had been strongly centrifuged before fertilization, the spermatozoon fre-
quently did not enter at all. He attributed the failure of fertilization
to the compactness of the yolk at the vegetative pole which blocked the
entrance of the spermatozoon. In our experiments, though we have
no detailed record, the percentage of fertilization in centrifuged eggs is
generally not very low in comparison with normally fertilized eggs.

In sections it appears that the spermatozoon may penetrate the egg
in the clear zone or between this zone and the yolk; for the most part
it enters the yolk zone (Fig. 8). This indicates that the compactness
of the yolk is not a factor in blocking the entrance of the spermatozoon.
In polyspermic eggs the sperm asters are found in almost any part of
the egg, even in the alveolar substance of the centripetal zone. Figures
4 and 5 show two such eggs that had been centrifuged for one hour at
2193 X gr. and were fixed fourteen minutes after fertilization. The
entrance points of the sperm are suggested by the positions of the asters.

The first maturation spindle of the strongly centrifuged eggs always
lies in the clear zone. After fertilization it moves to the periphery of
the zone where the polar bodies are given off. In some polyspermic
eggs, when a sperm-nucleus has already migrated to the center of the
egg, the first maturation spindle still rests in the middle of the clear
zone. In such cases probably no polar bodies will be given off; they
will be retained in the egg.

The entrance point of the spermatozoon may be found at any plane
with respect to the position of polar bodies. This is evident in Figs.
6 and 7, which represent two eggs centrifuged for ten minutes and fixed
one hour after fertilization.

In the normal egg a protoplasmic movement usually takes place
immediately after the entrance of the sperm. Such movements probably
also occur in the weakly centrifuged eggs with incomplete stratification
of ooplasmic substances, since after fertilization the different substances
of such eggs return to their normal positions. In strongly centrifuged
eggs the movement of the ooplasm is less marked. This is well demon-
strated in Fig. 8. This egg was centrifuged in a capillary tube for one
hour and twenty-four minutes at 2193 X gr-, and fixed forty minutes
after fertilization. The three principal zones are still quite clear. The
sperm aster is found in the yolk zone surrounded by a clear substance
and no mitochondria. It appears that the free movement of the sub-
stances is impeded by the stratification.

Cleavage

As a check, in every experiment some of the control eggs were
fertilized at the same time as the centrifuged eggs. The first cleavage

158 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

of the eggs of both sets occurred at the same time, regardless of the
rate and duration of centrifuging. It took place usually in about one
hour and a half after insemination, much later than the records of other
authors. Such delayed cleavage may be due to the low temperature of
the sea water at Tsingtao.

SO p

Q o

fe^^ty ;-W, .• ' \>V. y-t^rt :-rv " '"* -V ' Q-**-"~ • " -rSo ' ' '-. - " A

^Blfe^MPISfl^Sl^ft

^l§^^^^^^^

,^P-°:Oo:CXp^-o- , 0 ° no°bnP^-&x900.y

.,
°.°

G °'°.

QOO:,

- °.a°r
%$£&&&

.50

FIGS. 6 and 7. Sections of eggs centrifuged for ten minutes at 2193 times
gravity and fixed one hour after fertilization, showing the positions of the sperm
aster, the polar body (Fig. 6) or second maturation spindle and the mitochondria
crescent. There are no constant relations between them.

FIG. 8. Section of an egg centrifuged in capillary tube for one hour and
twenty-four minutes and fixed forty minutes after fertilization. The sperm aster
is in the yolk zone ; no mitochondria surround it.

The pattern of the cleavage planes of the centrifuged eggs differed
in no essential respect from those of normal eggs. The first two
cleavage planes were perpendicular to each other, while the third plane
was at right angles to both of the first two, resulting in the formation
of four micromeres and four macromeres. However, the first two
cleavage planes did not always pass through the point of attachment of
the polar bodies. After prolonged centrifuging, there was a significant

DEVELOPMENT OF CENTRIFUGED ASCIDIAN EGG

159

percentage of eggs whose polar bodies were not at the first cleavage
furrow. The deviation between them may be of any angle up to 90°
(Fig. 9). Since after centrifuging there was no definite landmark to
indicate the original position of the animal pole and since the orientation
of the eggs during the process of centrifuging could not be determined
accurately, it was impossible to ascertain whether the polar bodies were

IOO fl>,

FIG. 9. Section of an egg in 2-cell stage. The polar body is not located at
the cleavage furrow.

FIG. 10. Two entire eggs in 2- and 4-cell stages, showing mitochondria in one
blastomere.

produced at other than their normal position, or whether the first
cleavage plane passed through the original animal pole.

In most eggs the first cleavage plane appears to coincide with the
axis of centrifugation. As the stratified substances are rarely equally
distributed around the axis, it divided them into more or less unequal
halves. However, there is a significant percentage of eggs in which the

160 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

first cleavage appeared in any plane and at any angle with respect to
the stratification of substances. It may be oblique to the axis of cen-
trifugation or perpendicular to it. In the latter case one blastomere
contains only the yolk spherules and the other the clear and alveolar
substances. Figure 10 shows two such eggs in the 2- and 4-cell stages
respectively. They were fixed in Flemming's solution and mounted in
toto. The mitochondria or alveolar substance is confined to one blasto-
mere. Similar conditions are also shown in Fig. 9.

SO ft.

SOfL

775

FIG. 11. Section of an egg centrifuged for one hour at 2193 times gravity.
The three ooplasmic zones remain unchanged after fertilization.

FIG. 12. Section of two gastrulae, showing the abnormal positions of muscle
(a) and endoderm (b) cells. ««., muscle cell; end., endoderm; ns., neural cell.

FIG. 13. Section of a gastrula derived from egg centrifuged fifteen minutes
at 2193 times gravity. The polar body is situated posterior to the middle of the
ventral surface in its typical position.

It not infrequently happens that the first cleavage is quantitatively
unequal ; this leads to a formation of a larger macromere and a smaller
micromere. Sometimes the second cleavage plane does not intersect
the first at right angles, so that the four resulting blastomeres do not
lie in the same plane. The cleavage pattern in such eggs as well as in
those described above seems to be determined by the position of the

DEVELOPMENT OF CENTRIFUGED ASCIDIAN EGG 161

cleavage spindle. In cases where the three principal zones remain un-
changed after fertilization, the first cleavage spindle always lies hori-
zontally in the clear zone. In such eggs the cleavage plane coincides
with the axis of centrifuging. If a slight change of the stratification
takes place before the division, the position of the mitotic spindle is also
changed. The subsequent cleavage will then divide the egg in any plane
with respect to the axis of the centrifuging. Figure 11 shows a section
of an egg in which the ooplasmic substances were stratified into three
zones. The two pronuclei had come together in the center of the clear
zone. The first cleavage of this egg may be expected to approximate
the axis of stratification and will distribute the ooplasmic substances
equally between the two blastomeres.

In many eggs the division of the cell body is suppressed while the
division of centers and chromosomes continues. In such eggs are found
numerous nuclei and centers, confined £0 the alveolar and clear zones.
Such anomaly of cleavage is evidently not due to any direct effect upon
the mitotic figure, for the fertilization of the eggs took place after
centrifuging.

Later Development of Centrifuged Eggs

The later development of centrifuged eggs is particularly interesting
since the dislocated tissues or organs can be identified in stained sections.
Individual eggs showing abnormal distribution of ooplasmic substances
to the first two, four, or eight blastomeres were sketched and isolated.
The development of these eggs was studied. In general, it may be said
that the stronger and the longer the centrifuging, the more abnormal was
the subsequent development.

The gastrulation of strongly centrifuged eggs was rarely typical.
Figure 12 shows two abnormal cases in which a part of the mesoderm
and entoderm cells had not been invaginated. In normal development
the polar bodies are situated on the anterio- ventral part of the older
gastrula. In Fig. 13 a polar body is found posterio-ventrally on the
gastrula, indicating that gastrulation took place independent of the posi-
tion of the polar bodies.

Following prolonged centrifugation, a large number of eggs started
their development but most of them never reached a stage at which they
could be recognized as larvae. The embryos were so abnormal in form
that it was impossible to identify their parts and organs except by a
histological study. In these embryos, tissues and organs which will be
described separately in the following paragraphs were usually displaced.
Embryos are found composed of three types of cells, namely: (1)
muscle and mesenchyme cells; (2) ectoderm and neural cells; and (3)

162 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

notochordal and endoderm cells. These types correspond to the products
of the three zones (viz., alveolar substance, clear cytoplasm, and yolk)
which had been stratified by the centrifugal force.

Muscle Cells and Mitochondria. — In typical embryos and larvae, the
large muscle cells are arranged in the tail in three rows on each side of
the notochord. Owing to their large nuclei and specific staining reac-
tion, they are easily distinguished from other types of cells. In the
abnormal larvae derived from centrifuged eggs, isolated or aggregated
cells of this type may be found in the interoir or at the surface (Fig. 14).
They are rarely arranged regularly even when they are found alongside
of the notochord. In some cases these cells lie in the midst of the
endoderm or just under the neural tissue. Figure 15 shows a section
of a larva in which a part of the gut wall is formed of cells other than
typical endoderm. Though one cannot be certain as to whether these
are really mesoderm cells or not, the structure of the cytoplasm and the
size of the nucleus are similar to those of the latter.

The mitochondria which are normally embedded in the cytoplasm of
the muscle cells may have been driven into the regions subsequently
forming ectoderm (Figs. 16 and 19a) or neural tissue (Fig. 15). The
myofibrillae are often found in the muscle cells, but are never found in
those tissues into which the mitochondria have been driven. Our ob-
servations, therefore, confirm the view of Conklin, that in Ciona the
mitochondria do not give rise directly to myofibrillae.

Notochordal Cells. — Owing to the abundance of yolk spherules, the
notochordal cells of the young embryo are very similar to those of the
endoderm. In larvae, on the other hand, they are characterized by the
possession of large vacuoles and can be thus recognized wherever they
occur. In most abnormal larvae derived from strongly centrifuged eggs,
notochordal cells did not arrange themselves to form a rod, but instead
were displaced to various positions. They may be grouped together or
scattered. In some cases, they are found at the surface (Fig. 17) of
the larvae or just under the ectoderm (Fig. 16). In other cases they
are in the midst of endoderm or surrounded by muscle cells. In such
embryos no tail is formed.

The neural tissue which will be described in the next paragraph was
not always associated with the notochordal cells ; nor do the latter exert
any influence on the differentiation of the ectoderm cells with which they
are in contact. We agree, therefore, with Berrill and Conklin who have
concluded that the notochord of the ascidian does not act as an organizer
in the sense of Spemann.

Neural Cells and the Sensory Pigment. — The neural cells may occur
in any portion of these abnormal larvae. Sometimes they formed a

DEVELOPMENT OF CENTRIFUGED ASCIDIAN EGG 163

ect

enJ

ft ^ v»oo°iv ° 7c? o/ ":^\xr s?^

FIG. 14. Section of an abnormal larva derived from an egg centrifuged for
thirty minutes at 2193 times gravity and fixed twenty-eight hours after fertilization.
ins., muscle cell ; ch., notochord ; pg., pigment spot ; ns., neural cell ; ect., ectoderm ;
end., endoderm.

FIG. 15. Section of an abnormal larva derived from an egg centrifuged for
one hour and thirty minutes at 2193 times gravity and fixed twenty-four hours
after fertilization. A part of the gut wall contains alveolar substance, ms.,
muscle cell ; ns., neural cell ; end., endoderm ; ch., notochord.

FIG. 16. Section of a larva from an egg centrifuged for one hour and fifty
minutes, showing the mitochondria in the ectoderm and neural cells, m., mito-
chondria ; ms., muscle cell ; ch., notochord.

plate on the surface ; at other times they were grouped forming a mass
of neural tissue. Usually, however, they lined a large irregular cavity
extending into the interior of the larva. In larvae having well-developed
tails, such a neural cavity is always found at the junction of the trunk

164 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

and the tail. The latter often turned dorsally, giving the larvae a curva-
ture in an atypical direction. Figure 18 shows a larva of this kind.
The neural cavity with two spots of sensory pigment is surrounded by
the curved tail on one side. Conklin has suggested that the elongation
of the neural plate and tube depends upon the normal elongation of the
notochord. Our observations, however, do not lead to this conclusion.

So/i

FIG. 17. Section of a larva from an egg centrifuged in capillary tube for forty
minutes at 2193 times gravity, ch., notochord; ms., muscle cell; end., endoderm.

FIG. 18. Sagittal section of a larva derived from an egg centrifuged one hour
and thirty minutes at 2193 times gravity, showing the inverse body curvature.
ch., notochord ; ns., neural tissue ; pg., pigment spot ; ins., muscle cell.

FIG. 19. a. Section of a larva from an egg centrifuged in a capillary tube for
fifty minutes at 2585 times gravity. The left yolk-filled cells are endoderm and
the right mitochondria-filled cells are ectoderm.

b. Section of a larva from an egg centrifuged for fifteen minutes at 2193 times
gravity. The yolk-filled cells form a superficial epithelium, ch., notochord ; ect.,
ectoderm ; end., endoderm ; ms., muscle cell.

The sense organs are also structures of interest. In typical larvae
there are two sense organs, the eye and the otocyst, with their pigments
situated on the wall of the brain vesicle. Typical organs are rarely
formed in abnormal larvae ; instead, pigment spots are found in un-
expected places. The number of spots varies from none to four. Their

DEVELOPMENT OF CENTRIFUGED ASCIDIAN EGG 165

locations may be near together or widely separated (Fig. 14). In some
cases they project from the surface of the larva and in others they are
embedded in the neural cells. However, they are always associated with
neural tissue and notochordal cells.

Endoderm and Ectoderm Cells. — The histological characteristic of
endoderm is its large yolk-filled cells. In these abnormal embryos, how-
ever, it is not easily identified, for the yolk spherules are also found in
other cells of the embryo. In such a case as that shown in Fig. 19A
there is no doubt but that the superficial cells of the left side are endo-
dermal and those of the right side are ectodermal. In the case shown in
Fig. 195, on the other hand, one cannot be certain as to whether the
yolk-filled surface epithelium labeled " end " is true endoderm or is
ectoderm. Tung (1934), in experiments on the combination of blasto-
meres in Ascidiella scdbra, has shown that the development of ectoderm
and endoderm is not strictly mosaic. When the endodermal cells lie at
the surface of the embryo, they may form a regular superficial epithe-
lium. Similarly, if ectodermal cells come to lie in the interior, they may
transform into large irregular endoderm-like cells. Such a regulative
capacity of ectoderm and endoderm seems to exist also in the tissues of
dona embryos.

In these larvae having displaced organs, the endodermal cells may
be found in many abnormal locations. The fact that they are not always
associated with neural tissue and pigmented sensory spots indicates that
they do not serve as organizers.

Papillae. — The normal embryo possesses three papillae ; two are
paired and are situated dorso-laterally on each side ; one is median and
ventral. Papillae are also observed in some of the abnormal larvae.
Their number varies from none to three. It is of interest to note that
in spite of the dislocation of other organs, the development of papillae
is always in relation to the endoderm. Without exception in all 49 cases
observed, the papillae were formed in ectoderm underlain by endoderm.
Tung (1934), on the basis of experiments on Ascidiella scabra in which
the four animal blastomeres were rotated through 180° over the four
vegetal blastomeres, has suggested that papillae are evoked in the ecto-
derm by underlying endoderm. The present observation furnishes fur-
ther evidence in support of that hypothesis.

From the above descriptions it is obvious that the different tissues
and organs of the many larvae derived from eggs centrifuged before
fertilization, are out of their normal positions. Such placement of tis-
sues is undoubtedly the result of the placement of the ooplasmic sub-
stances from which the tissues are derived. In some cases the quantity
of certain tissues seems to be reduced. We have not, however, made a
count of the cellular components of these tissues.

166 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

DISCUSSION

Conklin (1905 a and &), in his classical studies on the development
of ascidians, described four different kinds of ooplasmic substances in
the fertilized egg, namely the ectoplasm, endoplasm, mesoplasm, and
chorda-neuroplasm. From these are derived the ectoderm, endoderm,
mesoderm, and the notochordal and neural tissue respectively. Accord-
ing to Conklin (1905c), the development of these tissues or organs is
strictly mosaic. When the eggs were centrifuged after fertilization,
tissues of the larvae which developed were displaced from their normal
positions (1931). Such displacements of tissues were attributed to
corresponding displacements of ooplasmic substance by means of cen-
trifugal force.

As has been shown in the present experiments, the stratification of
ooplasmic substances of eggs centrifuged prior to fertilization and the
displacement of the tissues or organs of the larvae derived from these
eggs are in general similar to those obtained from eggs centrifuged after
fertilization. It is reasonable, therefore, to conclude that the different
kinds of ooplasmic substance described by Conklin in fertilized eggs have
already existed in the egg before fertilization. This conclusion agrees
with that of Dalcq (1932, 1935, 1938), who likewise concluded that
organ-forming substances are already differentiated in the unfertilized
eggs. Conklin has suggested that the unfertilized egg of Cynthia pos-
sesses a bilateral symmetry. If so, it is undoubtedly due to the bilateral
arrangement of these ooplasmic substances.

In recent years the idea that the ascidian egg is strictly mosaic has
been called in question, though it is still insisted upon by Berrill (1932)
and by Cohen and Berrill (1936). Schmidt (1931) observed three
papillae in larvae derived from one-half blastomeres of dona and
Phylusia eggs. Reverberi (1931) has obtained normal larvae from
fragments of fertilized dona eggs. Tung (1934) has found in Asci-
diella scabra that the endoderm and ectoderm are relatively equipotential
and that the development of papillae and sensory pigments is always
associated with endoderm and notochord respectively. Recently, Von
Ubisch (1938) has described a normal embryo of Ascididla aspersa pro-
duced by fused eggs. All these investigations show that after fertiliza-
tion the ascidian egg has a considerable capacity for regulation.

Dalcq has reported a certain degree of regulation in the unfertilized
egg of Ascidiella scabra. A fragment of the egg may give rise to an
embryo very similar to the control. The number of cells of muscles or
notochord of a pair of embryos derived from two fragments of one egg
may be double the total number of a normal embryo. Moreover, in the
experiments reported here, we have demonstrated that the ectoderm and

DEVELOPMENT OF CENTRIFUGED ASCLDIAN EGG 167

endoderm constitute relatively equipotential systems and that the papillae
and sensory pigment do not appear to be self-differentiating organs.
These facts plainly show that the eggs of Ascidiella and dona are not
strictly mosaic prior to fertilization.

The recent work of Rose (1939) shows that the anterior vegetal
region of the Styela egg is the cerebral inductor. This region contains
materials essential to the differentiation of endoderm, notochord and a
part of the neural tissue. In the larvae with displaced tissues or organs
we have not found any typical relations between the cerebral vesicle and
either notochord or endoderm. In short, there is no evidence in the
present experiments to indicate that either notochord or endoderm acts
as a cerebral inductor. The development of the adhesive papillae has
been discussed by Cohen and Berrill (1936). They have found three
papillae in a lateral half larva of Ascidiella aspersa and interpret the
origin of such supernumerary papillae as a result of the mosaic pattern.
If this interpretation is correct, it might be expected that in larvae with
displaced organs, papillae would be found in a variety of abnormal
locations. On the contrary, papillae in abnormal larvae are always as-
sociated with the endoderm. This fact furnishes further evidence in
favor of the suggestion of Tung that papillae appear to be evoked in
the ectoderm by the underlying endoderm.

From the foregoing discussion we come to the conclusion that the
organization of the unfertilized egg of the ascidian is similar to that of
the fertilized egg. In both there exist different kinds of ooplasmic
substances from which different tissues develop. These substances, how-
ever, are not strictly mosaic ; they still possess a certain capacity for
regulation. Some organs, such as papillae and sensory cells, seem in-
capable of self-differentiation. Their development might be dependent
upon extrinsic factors.

SUMMARY AND CONCLUSIONS

1. The unfertilized eggs of Ciona intestinalis after centrifuging can
be cross-fertilized. The majority undergo cleavage but rarely develop
normally.

2. The first cleavage furrow may lie in any plane relative to the
position of polar bodies or the axis of centrifuging.

3. In larvae derived from strongly centrifuged eggs, tissues and
organs were often displaced from their normal positions.

4. Endoderm and ectoderm appear to be relatively equipotential.

5. There is no evidence that the differentiation of the neural tissue
is dependent upon other tissues with which it is in contact.

6. Mitochondria may be displaced from muscle cells and appear in

168 TI-CHOW TUNG, SU-HWEI KU AND YU-FUNG-YEH TUNG

the neural or ectodermal cells, where they are not transformed into
myofibrillae.

7. The number of papillae developed by the abnormal larvae varied
from none to 3. They are always adjacent to the endoderm which it
is suggested may evoke their formation in the ectoderm.

8. The number of sensory pigment spots in the abnormal larvae
varies from none to 4. They are always associated with notochordal
cells.

9. The organization of unfertilized eggs is found to be strikingly
similar to that of fertilized eggs in respect to ooplasmic substances and
the capacity of regulation.

10. The elongation of the notochord is not always accompanied by
the elongation of the neural plate or tube.

LITERATURE CITED

BERRILL, N. J., 1932. The mosaic development of the ascidian egg. Biol. Bull.,

63 : 381-386.
CHATTON, E., 1927. La methode de micro-inclusion double a 1'agar. Bull. d'Hist.

appl., 4 : 355-363.
CLEMENT, A. C, 1938. The structure and development of centrifuged eggs and egg

fragments of Physa heterostropha. Jour. Exper. Zool, 79 : 435-460.
COHEN, A., AND N. J. BERRILL, 1936. The development of isolated blastomeres

of the ascidian egg. Jour. Exper. Zool., 74: 91-118.
CON KLIN, E. G., 1905a. The organization and cell-lineage of the ascidian egg.

Jour. Acad. Nat. Sci., Philadelphia, 13: 1-119.
— , 1905&. Organ-forming substances in the eggs of ascidians. Biol. Bull., 8 :

205-230.

— , 1905c. Mosaic development in ascidian eggs. Jour. Exper. Zool., 2 : 145-223.
— , 1931. The development of centrifuged eggs of ascidians. Jour. Exper.

Zool, 60 : 1-120.
DALCQ, A., 1932. fitudes des localisations germinales dans 1'oeuf vierge d'ascidie

par des experiences de merogonie. Arch. d'Anat. Micros., 28: 223-333.
— , 1935. La regulation dans le germe et son interpretation. Comm. a la reunion

pleniere de la Societe de Biologic.
— , 1938. fitude micrographique et quantitative de la merogonie double chez

Ascidiella scabra. Arch, de Biol., 49: 397-568.
DUESBERG, J., 1926. fitude cytologique des oeufs centrifuges de Ciona intestinalis.

Arch, de Biol., 36 : 489-522.
MORGAN, T. H., 1938. The genetic and the physiological problems of self-sterility

in Ciona. I. Data on self- and cross-fertilization. Jour. Exper. Zool.,

78: 271-318.
REVERBERI, G., 1931. Studi sperimentali sull'uovo di Ascidie. Publ. Staz. Zool.

Napoli, 11: 168-193.

ROSE, S. M., 1939. Embryonic induction in the ascidia. Biol. Bull, 77: 216-232.
SCHMIDT, G. A., 1931. Die Entwicklung der Papen bei Ascidienhalbeilarven.

Arch. Zool Ital. (Torino), 16: 490^494.
TUNG, T. C., 1934. L'organisation de 1'oeuf feconde d'Ascidiella scabra au debut

de la segmentation. Compt. Rend. Soc. Biol, 115: 1375-1378.
— , 1934. Recherches sur les potentialites des blastomeres chez Ascidiella scabra.

Arch. d'Anat. Micr., 30: 381^10.
VON UBISCH, L., 1938. liber Keimverschmelzungen an Ascidiella aspersa. Roux'

Arch., 138: 18-36.

THE DEVELOPMENT OF THE BUD IN BOTRYLLUS

N. J. BERRILL
(From the Department of Zoology, McGill University, Montreal)

In the following account an attempt is made to describe the develop-
ment of the bud in Botryllus in a manner directly comparable with the
development of an egg, in an effort to bring out certain essential sim-
plicities in this direct type of development. The subject itself is not
new, but it is believed that the treatment is. Of the older papers those
of Hjort (1896) and Pizon (1893) are outstanding. Hjort's accounts
were concerned with the significance of the primary germ layers, or
rather their lack of significance, in the asexual development of Botryllus.
Pizon, on the other hand, described primarily the formation of young
colonies rather than bud development as such. Also his work was
marred by a completely erroneous account of the origin of the gonads,
gonads being considered to arise in the oozooid, in conformity with the
opinion of Weismann, and to be transported to the developing buds of
succeeding generations until sexual maturity was finally attained. This
was a false conception, and the description given here is very different.

Origin and Nature of Bud

The bud of Botryllus first appears as a small disc-like thickening of
the atrial epithelium on each side of the body, immediately anterior to
the gonads. A bud appears on each side, while just posterior to them
hermaphrodite gonads continue to develop. The atrial epithelium is of
ectodermal origin, and apart from its involvement in the process of
budding plays its part apparently only as a limiting membrane. There
is, in other words, no reason to suppose that the cells of the atrial
epithelium have become in any way specialized. Their formation as an
epithelium is a matter of tissue organization and implies nothing in
regard to cell specialization.

The disc of atrial epithelium concerned in budding is overlain ex-
ternally by a similar area of epidermis. This tissue is likewise an epi-
thelium and of ectodermal origin, but its association with test or tunicin
production suggests that it has special chemical activities in addition to
serving as a limiting membrane. That is, its constituent cells have
probably acquired a certain degree of individual specialization.

169

.170 N. J. BERRILL

The double disc of cells forming the initial bud is shown in Fig.
1, in optical section. In Fig. I, A it is shown in relation to its subse-
quent developmental cycle inasmuch as three stages are shown while yet
in organic continuity. In the largest and oldest of the three generations,
viewed from the ventro-posterior side, the vascular connection with the
circulatory system of the colony is clearly seen. The zooid is fully
active and the contained eggs have developed as far as the gastrula
stage. The zooid bears on its right side a bud about one-fifth its own
length in which the organization is virtually completely expressed but
is far from being functional. This bud has also formed vascular con-
nection with the colonial system. In turn it bears a bud in the first or
disc stage. Two features may be emphasized. A high degree of struc-
tural organization is attained at a relatively small size, and the size of the
bud in the disc stage is minute when compared both with its size at the
end of development and also with that of the developing egg. In Fig.
1, B the bud disc is shown, on a larger scale, in relation to the adjacent
structures of the parent bud.

Polarity

The question of origin of polarity in ascidian buds has been dis-
cussed before (Berrill, 1935, 1936). In every case where organic con-
tinuity is maintained between bud and parent, and Botryllus is no excep-
tion, the polarity of the bud is clearly a derivative of that of the parent.
Both the antero-posterior axis and the left-right axis coincide with those
of the parent zooid and must exist from the beginning.

Development of the Bud as a Whole

The simplest conception of the developing bud is that of a mass of
tissue expanding during a certain period. This is shown pictorially in
Fig. 2. The larger drawing in this figure represents accurately the
linear growth plotted against time. The various cross-sections within
the cone of growth represent certain developmental stages of special
interest. At 26° C. the time units are days, and at this temperature
development is completed on about the fifteenth day. If the value of
the time unit is coordinated with the temperature coefficient, the cone
becomes a constant expression of growth for all temperatures. The
growth curve is the usual sigmoid characteristic of developing organisms
in general.

At a given temperature, development to the complete functional stage
has a specific duration. Equally striking is the subsequent history of
the individual so formed. If the temperature is 26° C. and the devel-
opmental period fifteen days, the individual lives, feeds, and grows a

DEVELOPMENT OF BOTRYLLUS BUD

171

FIG. 1, A. Four generations representing complete cycle of zooid. fc1(
" ghost " of autolysed zooid of preceding bud generation ; b2, ventral view of active
zooid bearing right bud only and containing developing eggs in gastrula stage.
63, bud borne by active zooid, with rows of definitive stigmata about to become
perforate, and in turn bearing bud of next generation in its initial disc stage (&4).
Both the active zooid (&2) and its bud (63) are connected with the colonial circu-
latory system by their ventral ampullary vessel, v~.

B. Part of 63 at a higher magnification, showing general relationship of bud
disc, atrial epithelium, and gonads. This stage is drawn at the same magnification
as those of Fig. 4 and follows as a stage of bud development Fig. 4, F.

as, atrial sac ; ap, atrial epithelium ; bd, bud disc ; end, endostyle ; g, gastrula ;
ht, heart; ov, ovum; s, bud stalk; st, stomach; /, testis ; v, ampullary connecting
vessel.

172

N. J. BERRILL

A initial disc

maximal disc

closure
atrial horns
heart origin
Jjud rudiment
stigmata perforation

heart beat

stigmata and
siphons active

gonad maturation

dissolution

stage i
stage z

stage i \

period of gonad origin
stage 4 '.

stage 5

stage 6

stage 7

stage a

stage 9

stage 10

stage n

length

development
maintenance
dissolution

FIG. 2, A. Critical stages of bud development in relation to growth curve and
whole life cycle of zooid. The time scale is abbreviated for the maintenance or
active period.

B. Similar growth curves for five successive generations indicating over-
lapping of generations and the three phases of development, maintenance and
dissolution.

little, for about seven days. After this a period of dissolution and
autolysis occupies a further two days. In Fig. 2, A, these last two
phases are indicated against a condensed time scale. Their duration is

DEVELOPMENT OF BOTRYLLUS BUD 173

related to temperature in the same way as is the first or developmental
phase.

As shown in Fig. 1, A, the first or disc stage in bud development
appears when the parent zooid is itself but one-fifth grown and in turn
still attached to its parent. In other words, there is a great extent of
overlap in the life cycles of successive generations. This is shown in
Fig. 2, B. Five complete cycles are shown in the form of overlapping
growth cones related to one time scale. It can be seen that in any tem-
poral cross-section three generations in their respective existence phases
will occur. It may also be seen that the final phase of dissolution and
autolysis commences immediately after the bud of the succeeding gen-
eration reaches its full size and ends at about the time the bud attains
its maturity. For each generation the developmental phase is shown in
light line, the mature phase in heavy line, and the dissolution phase in
dotted line.

Early Development of the Bud

Throughout development the epidermis, arising in continuity with
the epidermis of the parent, plays almost no part other than to form
more epidermis conforming in area and shape to that of the organism
arising from the atrial tissue. For the most part therefore it will be
ignored.

The youngest stage so far detected consisted of eight atrial cells
forming a disc of three cells, more flat than cubical, in cross-section.
The whole disc was about thirty micra in diameter, and is shown in
surface view in Fig. 3, B, and in optical section in continuity with the
squamous atrial epithelium in Fig. 3, C. Figure 3, A represents an
arc of the egg of Botryllus drawn to the same scale to show the relative
size of the egg and the bud rudiment.

As the number of cells constituting the disc increases, the area of
the disc increases and the constituent cells change from a sub-cubical to
a columnar shape. This leads one to suspect that the columnar condition
is typical but in the earliest stage the transition to the surrounding
squamous atrial epithelium is so short that the columnar condition can
be only partially expressed.

In Fig. 3, D the disc stage is shown at its maximum size. With
further increase in area, or cell number, it curves into an arc, into a
hemisphere, and eventually into a closed sphere (Fig. 3, E, F, G, H).
The sphere becomes pinched off from the atrial epithelium from which
it originated, and the bud remains connected with the parent primarily
by an epidermal stalk. Two phases may accordingly be distinguished,
an expanding disc phase, and the phase of continuing expansion during

174 N. J. BERRILL

which the disc curves into a hollow sphere. A further feature of con-
siderable significance is associated with the second of these phases.
This is shown in Fig. 3, G and H, in which the gonads are already ap-
pearing. In the closed sphere stage shown in Fig. 3, H, the sphere
proper is shown in optical section. In addition certain cells stand out
clearly in surface view. These consist of three primary ova and a
number of small more ventrally placed cells destined to form the testis.
In the younger stage in Fig. 3, G, even before the sphere has closed, four
cells can be seen which, from their position and shape, are undoubtedly
four primary ova. These and later reproductive cells arise by extrusion
or delamination from the wall of the sphere. The gonads arise therefore
in a remarkably precocious manner.

In Fig. 3, H, the epidermis shows definite evidence of active morpho-
genesis, for the distal evagination is the rudiment of the epidermal stolon
that unites eventually with the colonial circulatory system. The disc,
hemisphere, and closed sphere stages are shown on one-half the scale of
Fig. 3 in Fig. 4. In this are also shown three subsequent stages. Apart
from the growth and elaboration of the gonads, which is described in
detail later in this paper, these six stages represent stages in a single
continuous process. The process of expansion and folding that changes
a disc into a hollow sphere continues so that the sphere becomes concave
along several facets or arcs. As the anterior arc continues to expand,
two vertical folds appear. At the same time an evagination appears in
the posterior arc. These are shown in Fig. 4, E. The anterior folds
gradually extend posteriorly until they divide the single vesicle into a
median and two lateral chambers. These three units are the central
pharyngeal chamber and the pair of lateral atrial chambers (Fig. 4, F).
At the same time the posterior evagination grows out to form the rudi-
mentary stomach and intestine. When the primary subdivision into
three chambers is complete, two small evaginations develop from the
central chamber. Median and anteriorly a small bulge becomes the
neural mass, while a somewhat larger evagination from the left posterior
wall represents the developing heart. Therefore, apart from the segre-
gation of lateral masses from the wall of the sphere to form the gonads,
all the principal divisions of the Botryllus zooid are produced, pharyngeal
and atrial chambers, intestine, heart and neural complex by a simple
process of progressive folding of an expanding sheet of tissue. These
are all clearly shown in Fig. 4, F.

Later development is primarily an elaboration of detail of each of
these divisions. Figure 1, B is of the same scale as Fig. 4 and demon-
strates both the extent of growth and elaboration that occur by the time
the developing atrial epithelium in turn has formed its bud disc. The

DEVELOPMENT OF BOTRYLLUS BUD

175

FIG. 3. Formation of bud.

A. Part of circumference of mature egg drawn to same scale for comparison
of size.

B, C. Surface view and optical section of initial disc (stage 1).

D. Optical section of maximal disc (stage 2).

E, F. Arching of disc to form sphere or vesicle.

G. Bud vesicle beginning to close proximally, and showing lateral segregation
of four cells destined to become mature ova, extruded into space between inner
and epidermal vesicles.

H. Later stage (stage 3) with vesicle closed, epidermal ampullary vessel
protruding distally, three presumptive mature ova and a number of male cells all
extruded from the lateral wall of the vesicle, to lie outside it. A similar condition
exists on the opposite side of the vesicle, not shown in the figure.

cp, epidermis ; cr/>, atrial epithelium.

N. J. BERRILL

0-1 mm

FIG. 4. Early development of bud.

A. Maximal disc (stage 2), from left side.

B. Hemisphere stage, from right side.

C. Closed vesicle (stage 3), from right side, with gonads extruded from wall.

D. Continued gonad segregation, stage viewed from ventral aspect.

E. Formation of atrial folds and intestinal outgrowth (stage 4), from ventral
side.

F. (Stage 5.) Origin of heart, intestine and neural vesicle, and completion
of subdivision of vesicle into central pharyngeal chamber and lateral atrial sacs,
from ventral side. The succeeding stage (stage 6) is shown on the same scale
as Fig. 1, B.

af, atrial fold; as, atrial sac; lit, heart; nv, neural vesicle; s, stalk; st, stomach;
v, epidermal ampullary vessel.

DEVELOPMENT OF BOTRYLLUS BUD 177

bud disc appears at a precisely definable stage in the whole development
and is to be regarded as an essential and definite constituent part of the
organization of a specific stage.

Later Development

Development of Stigmata. — Gill slits develop as perforations of the
combined pharyngo-atrial wall. This is the case both for the organism
developing directly from the egg and for the developing bud. In the
first case atrial sacs grow in on each side of the embryo and come into
contact with the pharynx wall. Gill slits appear only within the area
of contact. In the bud the equivalent double wall is formed, as already
described, by the downgrowth of the pair of anterior folds that divide
the primary vesicle into the central and lateral chambers, as shown in
Fig. 4, E and F. The two walls are shown in Fig. 5, A at a stage in-
termediate between the preceding two. Only when this double wall
expands to about ten times its linear size does stigmata formation become
evident. The first indication is the appearance of an alternating thick-
ening and thinning (spatially) of each of the two component epithelia
separately, as shown in Fig. 5, B. The thick ridges run dorso-ventrally
from the mid-dorsal line to the endostyle and each represents a row of
stigmata. Between the ridges the epithelia flatten out as the interstigmal
tissue. This condition is definitely associated with the stage bearing
the bud disc stage of the next generation.

As the two layers of ridges or thickenings increase somewhat in depth
they come into contact and fuse at a series of points along each pair
of ridges. Perforation occurs at these points to form the rows of
stigmata in their first definitive stage. The first perforate stage is shown
in Fig. 5, C and D. Subsequent development consists of an elaboration
of each of the units thus formed. No more will be added. Perfora-
tion of the fusing wall occurs at a definite and precisely definable stage
of development. At this same stage other features of the developing
pattern will be at a constant associated condition.

The stage of development of the bud of the next generation con-
forms to this relationship just as any other feature, and is near the
hemisphere stage of vesicle formation (Fig. 3, F and G}. In other
words, the bud itself is an integral part of the whole organization pattern
and the time and place of its inception are as sharply defined as that of
any other unit structure in the developing organism.

At perforation each stigma in surface view consists of a rosette of
about six cells (Fig. 5, D). Each constituent rosette of a row continues
its development as a unit. With multiplication of its seemingly un-
specialized cells the central aperture bordered by the cells expands and

178

N. J. BERRILL

elongates, so that the size and length of each stigma increase progres-
sively. When the cell-multiplication is almost terminated each cell de-
velops short cilia. Further growth of each stigma to approximately

0-05 mm

FIG. 5. Development of gill slits (stigmata).

A. Double layer formed of inner atrial wall and of pharyngeal wall.

B. Alternate thickening of atrial wall and corresponding thickenings of
pharyngeal wall, each such paired thickening representing cross-section through
ridge destined to become row of stigmata.

C. Equivalent section to B, but of perforate stage.

D. Surface view of C, showing three stigmata rosettes in initial perforate
condition.

E. F, and G. Three stages in subsequent growth and differentiation of a
single rosette to form a functional gill slit.

at, atrial wall ; pt, pharyngeal wall ; st, perforate stigmata.

double its size takes place as the result of change in cell size and shape,
and at the same time the cilia grow until they almost touch those from
the opposite side (Fig. 5, F and G).

DEVELOPMENT OF BOTRYLLUS BUD

179

Accordingly the following features are evident in the development
of the gill slits : There is a primary condition in which atrial and
pharyngeal epithelia are present and in virtual contact. The basic pat-
tern is expressed as a series of ridges, each with a series of swellings
in each tissue some time before stigmata formation. Perforate stigmata
appear at the points of fusion between the two tissues, while subsequent
growth consists first of a period of cell multiplication and then of a
terminal phase of individual cell expansion and cyto-differentiation.

Development of the Gonads. — The origin of the gonads has been
described already. Their subsequent development as a unit organ is,
however, of some interest, as is that of a single ovum. Figure 6 shows
sections through a number of stages. Figure 6, A represents a section
of a stage immediately following that seen as a whole mount in Fig. 3,
H. The originally thick wall of the internal vesicle is divided into the
thin atrial wall and the massive developing gonacl. The gonad here
consists of two primary ova and a mass of loose cells representing a
few rudimentary ova and many male cells. Cells are added to the col-
lection over a considerable period from some parts of the inner retaining
wall. In other words, as the lateral walls continue to grow, the splitting
into inner atrial and outer gonadial components continues in marginal
regions previously incapable of such splitting by virtue of insufficient
cells. This is shown in Fig. 6, B, a section passing transversely across
the anterior end of a bud at a considerably later stage, a stage inter-
mediate between those shown in Fig. 4, E and F.

The section shown in Fig. 6, C illustrates several points of interest.
The inner atrial epithelium is entirely distinct in kind and in space from
any part of the gonad. The form of the lobular testis becomes apparent
in spite of the small number of its constituent cells. And in the case
of both ovary and testis there is a residual mass of cells unincorporated
into those organs. In the case of the ovary, the small inner ova never
grow and mature. The residual cells of the testis may or may not
develop into testicular lobes, depending on the degree of belatedness of
their segregation. Virtually the complete form of the testis is to be
seen in the stage represented in Fig. 6, E, even though the testis here
is less than one-quarter its final size (in linear dimension). The form
is almost fully expressed, but its histo-differentiation is indiscernible.
In fact the final differentiation into condensed and tailled spermatozoa
occurs only after the full size of the developing bud is at last attained.

In the ovary those ova segregated from the vesicle wall in the first
phase of gonad formation (in number from one to four) grow and
mature. Those formed later remain close to the size at which they
were segregated. The primary ova, as far as can be determined, in-

FIG. 6. Development of gonad. All drawings from actual, not merely optical,
sections.

A. Left half of advanced vesicle (between stages 2 and 3) showing dif-
ferentiation of lateral wall into atrial epithelium, male cells, and presumptive
mature ova.

B. Frontal section of later stage (between stages 4 and 5) in anterior region,
showing continued segregation on left side of male cells from the lateral wall.

C. Left side of bud at stage 5 showing four presumptive mature ova, a few
undeveloping ova, and precocious tabulation of testis.

D. Two isolated ova with follicle and nurse cells.

E. Gonad of stage 6. showing lobular testis, and a single ovum with numerous
follicle and nurse cells.

F. Part of ovum at its maximum size, showing nurse cells, follicle cells, and
developing oviduct of same origin as follicle cells.

at, atrial epithelium; cp, epidermis; fc, follicle cell; Iv, left wall of primary
vesicle; m, male cells; ns, nurse cells; ov, ovum; ovd. oviduct; ovs, secondary ova;
/, testis.

DEVELOPMENT OF BOTRYLLUS BUD

181

elude other cells from the first. No stage, with the possible exception
of those shown in Fig. 3, has been seen in which the ova proper are
without accessory cells. These cells are of two kinds, a few flattened
follicle cells clinging to the surface, and an equally small number of
nurse cells completely within the cytoplasm of the ovum. Following
multiplication, the outer surface cells become columnar, as in Fig. 6, D,
and eventually flatten again as they give rise to the egg chorion. The
nurse cells also multiply and are clearly involved in the growth of the
ovum. They are eventually extruded into the perivitelline space as the

18 days

FIG. 7. Growth curves for linear growth of bud, and of single egg X 10. On
the same chart is also shown a cleavage curve indicating the geometrical increase
in cell number associated with growth.

inner follicle cells or "test" cells of the mature ovum (cp. Berrill,
1929).

Thus the development of the gonad as a whole is a comparatively
complex process. Yet the ova and spermatozoa attain actual functional
maturity at virtually the same time and almost at once after the bud as
a whole has become active. Either some factors external to the gonads
suddenly terminate growth and multiplication and enforce final differen-
tiation to coincide with that of the rest of the bud, which is unlikely, or
the development has from the first been approaching a condition of
equilibrium permitting final cyto-differentiation. That this last is the

182 N. J. BERRILL

case is shown by a comparison of the growth curves for the bud as a
whole and of a single ovum. The curves for the linear growth of the
bud and the ovum, when reduced to equivalent scales, are practically
identical, as may be seen in Fig. 7. Both are sigmoid curves indicative
of an approach to and attainment of a " steady state."

Cell Division and Differentiation. — In all tissues of which the cells
finally exhibit a marked degree of structural specialization or differentia-
tion, the structural details become visible only at or toward the close
of the phase of cell multiplication. The time at which this occurs varies
greatly. In the case of spermatozoa, condensation and elongation occur
only after the rest of the organism has as a whole become functionally
active. In the case of ova, as distinct from their associated follicle
cells, growth and differentiation without division occupy almost the whole
developmental period. Muscle cells of the heart and body wall stop
dividing and become greatly elongate when linear growth of the whole
organism is little more than half complete. Cessation of division and
subsequent formation of long cilia in stigma cells occurs very late, but
is complete before elongation of spermatozoa commences. In the case
of ova and muscle cells, at least, there is very considerable growth after
cell division has come to an end. Growth of the organism as a whole
accordingly conforms to a typical curve in spite of the fact that the
growth is in part clue to cell multiplication and in part to cell growth
without division. The growth curve for a single ovum is similar to
that of the whole organism. It seems clear that the developing bud
grows at a rate characteristic of an approach to a " steady state," and
that the growth of the parts, whether based on cell division or not, is
not a group of independent processes cooperating to form the whole,
but must be governed by the whole. Cell division as such becomes, in
this view, a condition and tool of the whole developmental process rather
than in any sense a basic cause. Otherwise the coordination of the
varying times at which different cell types cease division and differentiate
becomes virtually unaccountable. In a comparison of the growth of a
non-dividing ovum with the growth of a group of cells by multiplication,
it appears that the rate of volume-increase is quite independent of
division processes.

Summary and Conclusions

The bud arises as a disc-like thickening of the anterior atrial wall,
consisting of a small number of columnar cells transformed from the
atrial epithelium, overlain by an equivalent area of unmodified epidermis.
The polarity of the disc and subsequent organism is an extension of that

DEVELOPMENT OF BOTRYLLUS BUD

183

of the parental tissue, with regard both to the antero-posterior and lateral
axes. Development itself is fundamentally extremely simple and direct.
After the completion of development there is a phase of functional
activity and a phase of autolysis and dissolution. For any given time-
temperature scale the duration of these last two phases is as specific and
determined as that of the developmental phase. Of the two tissues
constituting the bud disc the epidermis forms only more epidermis,
though acquiring the form of a whole organism including the ventral
stolonic outgrowth. The atrial component of the disc forms everything
else. As the disc expands, by means of cell multiplication, it trans-
forms progressively into a hemisphere and eventually into a hollow
sphere attached by a narrow stalk to the parental tissue. Two folds
develop anteriorly and divide the vesicle into two lateral and one median
chamber. The lateral divisions represent the atrial chambers, the median
the pharyngeal sac and from it three evaginations are formed represent-
ing the heart, neural mass, and intestine respectively. Later development
is primarily an elaboration of these unit-regions. As an example, the
formation, growth and differentiation of gill slits in the pharyngeal wall
is described in detail. The essential pattern of the stigmata is apparent
even before they become perforate. Each stage in their development is
precisely correlated with specific stages in the development of the whole
organism.

The bud anlagen of the succeeding generation appear as discs in the
anterior wall of the left and right atrial chambers at a specific stage in
the development. This stage is that in which rows of stigmata, while
not yet perforate, are represented by ridges or folds of the pharyngeal
wall. At the time of perforation, the buds are approximately at the
closed vesicle stage. The buds, in fact, are to be regarded as essential
constituents of the organization pattern, appearing and developing in
time and place in a manner strictly analogous to that of any other unit
structure.

The gonads segregate as a mass from the lateral walls of the bud
at an extremely precocious period, even while the primary vesicle is in
process of formation. Once segregated, they in turn develop as a
seemingly independent unit structure. The testes show the final lobular
form virtually as soon as sufficient cells are present for its expression.
Ova, apart from the associated internal and external follicle cells, grow
and differentiate without dividing. They mature finally at the same time
as the spermatozoa which cease dividing and differentiate later than any
other tissue of the bud.

The development of each tissue is fundamentally the same. A period
of cell multiplication is followed by a phase of final differentiation. This

184 N. J. BERRILL

last phase may or may not include a period of cell enlargement, depend-
ing on the cell type to be formed. In the case of ova the multiplication
phase is barely present at all and the second phase occupies most of the
developmental period, involving enormous growth. In spermatozoa the
case is reversed and the final phase is extremely brief and actually in-
volves reduction in cell size. Muscle tissue lies between these two
extremes, while most other tissues approach more the condition of
spermatozoa.

The whole development of the bud and that of its component parts
is therefore as direct a process as can be conceived, without there being
any indication of the divergence to form tadpole larvae associated with
egg development. Cell multiplication continues to a greatly varying
extent in different parts and tissues, while the linear growth of the whole
or of a non-dividing ovum follows a regular sigmoid curve typical of an
approach to and attainment of a " steady state." In fact, the develop-
ment of the bud is essentially such a unitary process that " wholeness "
can be said to be the most outstanding feature of the organism not only
in its final functional phases but of every moment of its existence, and
especially of the beginning. It is virtually as though organization is
present from the first, though the extent of its visible expression is
closely correlated with and limited by quantity of available material at
every moment of development.

REFERENCES

BERRILL, N. J., 1929. Studies in tunicate development. Pt. I. General physiology

of development in simple ascidians. Phi!. Trans. Roy. Soc., B, 218 : 37-78.
— , 1935. Cell division and differentiation in asexual and sexual development.

Jour. Morph., 57 : 353-427.
— , 1935. Studies in tunicate development. Pt. IV. Asexual reproduction.

Phil. Trans. Roy. Soc., B, 225: 327-379.
HJORT, J., 1896. Germ layer studies based on the development of ascidians. Zool.

Res. Norwegian North Atlantic Exped. Christiania.
PIZON, A., 1893. Histoire de la blastogenese chez les Botryllides. Ann. Sci. Nat.

(7) Zool., 14: 1-386.

SIZE AND MORPHOGENESIS IN THE BUD OF BOTRYLLUS

N. J. BERRILL
(From the Department of Zoology, McGill University, Montreal)

The bud of Botryllus first appears as a disc arising in the anterior
wall of each of the atrial chambers. The disc grows a little and then
transforms into a sphere. The size to which the disc grows before it
starts to transform varies among different bud generations in a Botryllus
colony, increasing slightly with each successive generation of buds. The
present account is primarily a comparison of the development of the
buds arising from the smaller discs of early generations with the large
discs of late generations, within the colony formed from a single ferti-
lized egg. In addition to differences in size of bud primordia from early
and late generations in the colony, there are usually size differences
between the bud primordia of the right and left sides, that of the right
side being the larger.

The development of a bud has as its basis a continuous material
expansion from the small group of cells constituting the primary disc to
a functional bud of several thousand times its volume and cell number.
The significance of this expansion is paramount. As the disc expands
in area it becomes curved into a hollow sphere. As the sphere expands,
its surface folds inwards to divide it into three chambers, the major
territories of the body. With continued expansion, further surface
folding occurs to divide off smaller territories such as neural mass, heart,
and intestine. It can be said that for each successive size the material
(mass, area, cell number, etc., however it may be expressed) present at
that moment expresses virtually every character of the final organization
that is not inhibited by the limitation of size itself.

Each bud disc arises from the atrial epithelium as a group of cells
that gradually acquire a columnar form. The epidermis forms an
equivalent overlying component of the disc, but plays a relatively minor
part in the subsequent development. Since every disc of atrial cells has
to develop from the general atrial epithelium, there is almost certainly
no real minimum size that can be compared in different generations.
On the other hand, the disc in every case grows to a certain extent
before changes in form begin, and the size or cell number of the disc at
its maximum size, which is a precise stage, is a value readily compared.

185

186
A

minimal size
(right) (left)

maximal size

(right)

FIG. 1. Formation of bud vesicles, all in optical section. A, F, and / are
three maximal bud discs. A-E represents vesicle formation from small right
maximal disc from zooid of young colony. F-H, the smaller left maximal disc
of same series. I-M, vesicle formation from large right maximal disc from zooid
of a mature colony. M and N, right and left vesicles from same individual and
forming three and one mature ova respectively.

SIZE AND MORPHOGENESIS IN BOTRYLLUS 187

The smallest maximal disc commonly seen consists of about six cells
in optical section, the largest of about fourteen cells, or a difference of
about eight times in volume of tissue or number of cells present at this
stage. Figures 1, 2, and 3 represent a comparison of the development
of two sizes of maximal right bud discs, from early and late generations
respectively. In Fig. 1, G, H, and N, the smaller left buds are also
shown. In this figure several features of comparative interest are clear.
The relative difference in size of three maximal disc stages is main-
tained in the subsequent stages of hemisphere and sphere. In optical
section these maximal discs have 5, 8, and 14 cells respectively (Fig. 1,
F, A and 7), representing totals of about 21, 48, and 150 cells (ratios
1 : 2 • 1 : 7 • 1 ) . In the corresponding closed sphere or vesicle stages,
optical sections show 9, 15, and 25 cells respectively (Fig. 1, H, E and
M), representing cell totals of about 33, 75, and 210 (ratios 1:2-2:6-3).
From these values two facts emerge. The ratio of cell numbers repre-
senting the smallest and largest maximal disc illustrated is about 1 : 7.
The same ratio holds for the closed vesicle stage, and it is evident that
whatever the size of the maximal discs, the transformation is correlated
with an increase in cell number of about one and one-half times that of
the disc.

Morphogenesis is thus independent of absolute cell number, but/
closely dependent on relative cell number.

In Fig. 1 two other features are evident. The relation of morpho-
genesis to cell number is the same in the epidermis as in the atrial tissue.
The epidermis conforms in size and shape to the inner tissue, and as
may be seen in Fig. 1, E and K, the protrusion foreshadowing the ventral
ampullary vessel appears in both small and large vesicle stages, and in
spite of the very early stage in development as a whole.

It is of greater interest that gonad tissue is segregated from the right
and left lateral wall of the vesicle stage in the large forms but not in the
small. These two correlated variations in the vesicle stage, namely
degree of segregation of gonad tissue and the absolute size, produce an
increasingly marked effect on later stages of development. Figure 2
shows immediately succeeding stages drawn to the same scale as those
in Fig. 1 . Figure 3 shows still later stages at necessarily reduced scales.
In each case equivalent stages are shown for the development of both
large and small primordia.

The two primary differences, in size and gonad development, are
maintained in an increasingly obvious form. Thus in Fig. 2, the three
stages A, B, and C are morphologically equivalent to the stages D, E,
and F. In A and D the folding of the vesicle wall to delimit the primary
divisions of the body are just beginning. In B and E they are com-

N. J. BERRILL

fit

minimal size

maximal size

FIG. 2. Development, at same magnification, of series A-E and I-M of Fig.
1. A-C ', development of small vesicle, D—F of large vesicle, showing differences
in size of equivalent stages and in presence and absence of gonads.

af, atrial folds ; a.?, atrial sac ; hi, heart ; nv, neural vesicle ; ov, ova ; pc, pharyn-
geal cavity ; st, stomach ; /, testis ; v, ventral ampullary vessel.

SIZE AND MORPHOGENESIS IN BOTRYLLUS 189

pleted, and atrial chambers, pharynx, intestine, heart, and neural mass
already exist as unit regions. The difference in diameter of the closed
vesicle stage shown in Fig. 1, E and M, is fully maintained. In addition
there is the striking difference in gonad development. In the series of
stages associated with the small primordium none appear, in the larger
they continue to develop and become massive organs on each side between
the lateral wall of the atrial epithelium and the epidermis. These dif-
ferences become more and more pronounced, as may be seen in Fig. 3.

Size and General Organisation

While the structural consequences of the primary difference in size
become progressively more obvious, one feature needs to be emphasized
strongly. The great difference in cell number constituting maximal disc
stages is maintained at least in the later closed vesicle stage, the increase
being about one and one-half times. The difference in cell number is
expressed less markedly in disc and vesicle diameters. In the two main
series illustrated in Fig. 1, the diameter of the larger series is about one
and two-thirds that of equivalent stages of the smaller series. Excluding
gonads for the time being, this difference in linear dimension of the two
series is maintained closely in the later stages shown in Figs. 2 and 3
up to and including the active functional stage. Not only is the linear
size difference maintained throughout development, but it is equally
expressed in the number of such multiple structures as stigmata. In the
stages in which stigmata are just becoming perforate and in which they
are active organs, the number of rows of stigmata is six in the smaller
buds, ten or eleven in the larger, while the number of stigmata per row
in the smaller is 12 and in the larger 22. The number of stigmata does
not change during development. Thus in the two series the number
of stigmata formed is proportionate to size, since both linear dimension
of the whole, and the number of rows of stigmata and number of stig-
mata per row, vary as one to about one and two-thirds.

The difference in whole size of equivalent stages, which is expressed
numerically in multiple organs such as stigmata, applies equally to organ
size. This is the case for the heart, for in the three stages — primary
heart vesicle, initial beating, and final — the same relative size differences
are maintained.

Since the relative difference in size between the two series is main-
tained virtually at a constant level for all stages, it follows that each
stage represents a certain degree of expansion in terms of a preceding
stage, whatever may be the absolute size. A specific degree of expres-
sion of the complete organization is correlated with a certain size or

190

N. J. BERRILL

end-

tit

o-i o.g

minimal size

maximal size

FIG. 3. Continuation of same two developmental series, at two smaller mag-
nifications. A-B, E-F continued development of smaller bud disc and vesicle
shown in Figs. 1 and 2. C-D, and G-H, continued development of larger vesicle.

For purposes of comparison, A and C represent at the lower magnification the
stages shown in Fig. 2, C and G; in the same way E and H are reduced from B
and D. B and D are equivalent stages and have formed maximal bud discs. E
and H are also equivalent and have become active zooids. The difference with
regard to size, presence of gonads, and length and number of rows of gill slits
is obvious.

b, bud disc ; end, endostyle ; lit, heart ; ov, ova ; st, stomach ; t, testis.

SIZE AND MORPHOGENESIS IN BOTRYLLUS 191

material quantity. This size is not absolute and is not expressable in
actual measurement or cell number, but must be expressed in terms of
reference to the absolute size of the maximal disc stage. This is highly
significant and will be referred to again.

Development of Gonads and Initial Size

Confining the present account to the two extremes already illus-
trated, there is a spectacular difference in the condition of the gonads in
the two series, one producing mature gonads and the other none. This
difference goes back to the first stages of development. In the larger
series, the gonads are separated or extruded from the lateral walls of
the primary vesicle even before closure is complete. Once separated,
the gonads develop apparently as independent unit regions. The sep-
aration phase is comparatively brief, and there is no tendency to form
gonads except during this precise phase of the whole development. The
ova destined to become mature are the first tissue to be separated, testis
and prospective immature ova separating a little later.

In the smaller series, no separation of gonads occurs at all during
the equivalent phase, and no gonads appear at any later time. Conse-
quently the massive lateral bulges representing the growing gonads in
the developing buds of mature colonies are absent altogether in those
of very young colonies.

The correlation of presence or absence of gonad separation (and
therefore of subsequent development) with the size of the transforming
disc stages suggests at once that size itself may be the determiner.

Gonad tissue is separated during a very definite and specific period
of bud development, namely, during a period starting before closure of
the primary vesicle and lasting until the vesicle is more or less subdivided
into its three primary regions. At no other moment in development,
either earlier or later, is there any indication of gonad formation.
There is no absolute proof that at no other time under any conditions
can gonad tissue be separated, but it is reasonably certain that the ca-
pacity to produce gonad tissue is definitely limited to the period or phase
in which production always occurs. In other words, the gonad pri-
mordium is determined and formed at as precise a period in the whole
development as is the case for other organs, such as the heart.

Accordingly, if this assumption is made, it is easy to account for the
suppression of gonad formation in the development of small buds. The
situation is clearest if one compares the largest and smallest of the four
vesicle stages shown in Fig. 1. In the largest, three prospective mature
ova have separated from the vesicle wall. The extrusion of these par-

192 N. J. BERRILL

ticular cells commences immediately after the attainment of the open
hemisphere stage. The gonad material thus separated must represent a
certain minimum proportion of the lateral wall from which it arises, of
the order of one-quarter to one-half. It is separated when the vesicle
cells total about 160 and when there are about 22 to 26 cells in optical
section (Fig. 1, M). At this stage, in other words, the material from
which ova are separated is in the form of a sufficient number of cells for
individual cells representing individual ova to be pushed out.

In the equivalent stage of the smallest series, only 7 or 8 cells consti-
tute the optical section, and the region from which gonads and atrial
wall should be differentiated consists, on each side, of only 2 cells in
optical section. Accordingly the region constituting the prospective ova
is at this stage and size inadequately cellular and the separation of ova
at this moment becomes mechanically impossible.

Gonad formation, however, is not momentary but occupies a period
of time. In the smallest series even the late closed vesicle stage consists
of so few cells (Fig. 1, E) that at no time during the proper period can
gonad tissue be separated. In the largest series the prospective mature
ova are extruded before closure, male tissue and prospective immature
ova during and shortly after closure of the vesicle, and some additional
male cells after completion of the extrusion of female cells.

In series of developing buds of intermediate size all conceivable types
of immature gonads should be found. This is the case. In a series
slightly larger than the smallest, a sufficient size or cell number is at-
tained before the gonadial phase is completely passed, and some male
tissue is separated at the end of that phase. In a somewhat larger series
again, and a little sooner, more gonad tissue is separated consisting of
prospective immature ova and male tissue sufficient to develop into a
lobular testis of submaximal size. Similarly, in a larger but not maximal
series, there may be a separation at or before closure of one or two or
three prospective mature ova. In other words, a complete grading
from none to mature 4-ova gonads exists, correlated both with size of
series and with time of separation.

Conclusions and Summary

In successive bud generations within a colony there is a progressive
increase in the size of the bud rudiments and zooids subsequently de-
veloping from them. Dealing with material derived from a single ferti-
lized egg, it is possible to determine the relationship or importance of
rudiment size to morphogenesis. Development in every case is simple
and direct. It consists of the growth of the rudiment to a maximal

SIZE AND MORPHOGENESIS IN BOTRYLLUS 193

disc stage, the conversion of the disc into a sphere, the subdivision of
the sphere or vesicle into unit regions, the whole process being accom-
panied and conditioned by expansion or growth of tissue.

The maximal disc is a precise relative stage. The larger discs, typi-
cal of late and mature generations, may be seven times as large in area
as those of early generations.

Whatever the size of the maximal disc, succeeding stages bear to it
definite growth ratios. A certain percentage expansion or growth of
the disc tissue is correlated with a specific developmental stage, whatever
the absolute sizes may be. Absolute size must be determined during the
initial phase of development before the disc exhibits any tendency to
transform into a vesicle.

Ultimate size being thus initially determined, there is variability in
the following expressions of size. All organs and regions vary in abso-
lute size, maintaining their proportionate dimensions relative to the
whole. Multiple structures, such as stigmata, vary little in absolute size
for a given stage but vary in number in proportion to tissue, area. The
relatively massive gonads, fully formed only in the largest series, are
partially or completely inhibited in the smaller.

Gonad formation is essentially a serial separation of the various com-
ponents during a short phase of development, lasting from the open
hemisphere stage to the expanded closed vesicle stage. If the size of
the whole permits separation of each component as discrete cells at the
proper time for separation, maximal mature gonads will be formed and
develop. If size is so reduced that the various components cannot be
materially separated as cells, separation is inhibited and no gonads will
develop at the normal or any other time. With successive increases in
size from this last condition an adequate cellular state is reached, at
first including the later phases of the gonadial period and progressively
including the earlier, so that a series of immature gonads appear in the
inverse order of normal maximal development. Prospective mature
ova do not appear at a time normal for the appearance of prospective
immature ova or for male cells. Gonad components that do not separate
at their normal time do not appear at all.

THE EFFECT OF SALINITY UPON THE RATE OF
EXCYSTMENT OF ARTEMIA

R. H. JENNINGS AND D. M. WHITAKER

(From the Department of Biology, Stanford University)

•

INTRODUCTION

The phyllopod crustacean Artemia lives and reproduces in natural
and artificial brine pools and lakes in many parts of the world. It
tolerates an extreme range of salinity, pH, and other environmental
conditions, although it is relatively intolerant of certain substances such
as potassium1 (Martin and Wilbur, 1921; Boone and Becking, 1931).
Artemia does not require brine, or other environmental extremes, since
it completes the life cycle in ordinary sea water in the laboratory, but it
is defenseless in nature and is quickly destroyed by predators except in
environments which exclude them.

Artemia are abundant in evaporating ponds on the margin of San
Francisco Bay where salt is manufactured from sea water by solar
evaporation. This particular Artemia has been regarded as a variety of
A. salina, and also as a separate species, A. franciscana. Bond (1932)
suggests, after experiments on the effect of salinity on development,
that it should be regarded as a separate species.

The Artemia from the margin of San Francisco Bay reproduce by
two methods. In the presence of males the same females sometimes
produce viviparous nauplii, and at other times they release encysted
embryos which are encased in hard chitinous coverings or shells. These
encysted embryos will not ordinarily hatch in sea water until they have
first been desiccated. Air-dry cysts remain viable for many years and
when they are placed in sea water, the embryos hatch as swimming
nauplii.

Dry cysts are approximately % mm. in diameter. They are deeply
indented on one side, but when they are placed in sea water, or in sea
water of modified salinity in which they will hatch, they take up water
and round out to become spherical. After a time, which depends among
other things on salinity and temperature, the chitinous cyst wall or shell
splits, and the embryo emerges head first encased within a delicate trans-

1 This intolerance of potassium appears to be an important factor in the distri-
bution of Artemia in desert salt lakes (Boone and Becking, 1931).

194

SALINITY AND ARTEMIA EXCYSTMENT RATE 195

parent membrane or sac. This sac may remain attached at one end to
the shell or it may at once be free. During the emergence from the
shell and for some time thereafter the nauplius is quiescent within the
sac. The sac finally begins to soften and dissolve and the nauplius moves
its appendages. The nauplius completes its excystment by hatching or
escaping from the remains of the sac, after which it swims actively about.
At the time of hatching the nauplius contains an appreciable supply of
yolk and even in the absence of food it develops for several days and
undergoes the first moult to form a metanauplius. The external anatomy
of the developmental stages is completely described and figured by Heath
(1924).

The two stages or actions in the excystment, the initial emergence
from the shell, and the final hatching from the membranous sac, will
for convenience be referred to as " emergence " and " hatching " respec-
tively. Both emergence and hatching proceed rapidly compared with the
time lapse before emergence and between emergence and hatching. For
accuracy in determining rates, it is necessary to define these two stages
rather precisely even if somewhat arbitrarily. The emergence from the
shell is a discreet abortive process and a nauplius is considered emerged
if the eye can be seen. The first indication of hatching from the sac is
usually the projection of the first pair of antennae. Soon the large
second pair is also projected outside the sac and swimming or attempts
at swimming begin. The first movements of the appendages are often
intermittent and uncoordinated. A nauplius is considered hatched when
the first two pairs of appendages project outside the sac and are motile.
The third pair of appendages soon slides out and the remnant of the
sac is left behind.

The effects of specific ions and of ion antagonism on the excystment
of Artemia have been studied by Boone and Becking (1931), who also
have concluded that osmotic pressure has much less effect on excystment
tha'n chemical factors. Jacobi and Becking (1933) observed that excyst-
ment will not take place in natural sea water concentrates of three or
more molar equivalent. The present experiments were undertaken to
test the effect of total salinity on the rate of excystment in diluted and
concentrated sea water in which the proportion, or relative concentration,
of the ions contained in sea water remains essentially unaltered. In the
strongest concentration used (225 per cent sea water), there was no
visible precipitation of any kind of salt so that the ionic proportions
were unaltered except for second order differential effects on dissocia-
tion, and minor pH effects. The minimum salinity in which emergence,
hatching, and early development will take place has also been determined.

196 R. H. JENNINGS AND D. M. WHITAKER

METHOD
The Cysts

The cysts used in these experiments were generously provided through
the courtesy of Dr. Alvin Scale of the San Francisco Aquarium Society.
They were collected near Redwood City, California, June 10, 1937, and
the experiments were carried out in the winter of 1939.

About 20 per cent of the original sample of cysts excysted in sea
water. It was found by dissection that most of the remainder were
empty shells of previously excysted embryos, although some contained
embryos which were presumably dead. The empty shells are difficult to
distinguish by simple inspection, but it was found that they could be
partly separated out by rapid differential flotation since they contain air
until they have soaked. The cysts were shaken and suspended in dis-
tilled water in a test tube and those that floated were discarded. Some
good cysts were discarded by this method, and not all that sank were
good cysts, but a stock was obtained in which the percentage which
excysted had been increased from 20 to about 60. In this process the
cysts were exposed to distilled water for only three minutes and were
then dried on filter paper for five days at room temperature and 30 per
cent humidity. After this they were stored for some time in a stoppered
bottle before using. The brief washing in distilled water also served to
remove most of the salt on the cysts, which is important for the present
purpose. Otherwise, the differential flotation can be carried out as well
in sea water (Whitaker, 1940). The rate of excystment in normal sea
water varies with the duration of drying and storing after the washing,
and would no doubt differ in different samples of cysts for this and other
reasons. The present experiments were carried out on a single stock
of cysts during a period in which the rate of excystment in normal sea
water was practically constant.

The Media

Sea water (specific gravity, 1.025, pH 7.9-8.0) was collected at Moss
Beach, California, and was filtered before being concentrated or diluted.
The specific gravity 1.025 was taken as a base throughout and was con-
sidered to represent the salinity of what is called 100 per cent sea water.
Sea water was diluted by adding triple glass distilled water to prepare
the dilutions shown in Table I. It was concentrated by evaporating
under reduced pressure in a water bath at 45-50° C. The resulting brine
was diluted back with distilled water to prepare the salinities greater
than sea water which are also shown in Table I. Specific gravities were

SALINITY AND ARTEMIA EXCYSTMENT RATE 197

checked with pycnometers. The concentrating process removed gases
from the brine so the solutions more concentrated than sea water were
re-equilibrated by aerating for several hours with a sintered glass nozzle.
A glass electrode was used to measure pH. The rate of excystment is
practically unaffected by pH within the range 8.3-7.7, so that pH can
hardly be an important factor in the present instance except perhaps in
distilled water (see Table I).

Excystment

Small 1 cc. Syracuse dishes were used for excystment. Especially
in the solutions of high salinity, the cysts tend to float and to accumulate
in the meniscus where observation is difficult. Accumulation in the

TABLE I

Salinity and pH of media. For convenience in comparing, salinity is expressed
as a percentage of the salinity of normal sea water (specific gravity 1.025), and a solution
is described in terms of its relative salinity as the corresponding percentage of sea water.

Percentage Sea Water

Specific Gravity

225

1.0562

8.0*

200

1.0500

8.2*

175

1.0438

8.3*

150

1.0375

8.3

125

1.0312

8.2

100

1.0250

8.0

1.0187

8.0

1.0125

7.9

1.0062

7.9

12^

1.0031

7.7

1.0000

6.5

* Probably inaccurate due to effect of high salt concentration on glass electrode.
t Distilled water.

meniscus was prevented by dipping the dishes in hot, pure high melting
point paraffin. A thin coating of paraffin causes the water meniscus to
be inverted. In each experiment about twenty cysts were placed in 1
cc. of medium in each small dish, and ten small dishes were placed in
Petri dishes arranged as moist chambers to prevent evaporation. The
two most dilute solutions (0 and 12% per cent sea water) were changed
once in the course of the experiments so that the small amount of salt
on the cysts would not appreciably alter the salinity. No measurable
changes of salinity took place during the experiments. The moist
chambers and the solutions were kept throughout in a humid constant
temperature room at 25 ± %° C.

198

R. H. JENNINGS AND D. M. WHITAKER

After emergence began in a population, counts were made of the
numbers emerged and hatched at least every two hours until at least
60-70 per cent had hatched. The numbers that ultimately emerged and
hatched were also determined at about 96 hours, and the original empty
shells and non-viable cysts were excluded from consideration. The num-
ber emerged and the number hatched at the time of each observation
were treated as percentages of the number that ultimately emerged and

8 IOO

§
I

90
BO
70
6O
50
40
30
20
10
0

-o — o-

0 25 50 75 IOO 125 ISO 175 200 225

PERCENT SEA WATER

FIG. 1. The percentage of emerged embryos that hatched in diluted and con-
centrated sea water of various salinities. One hundred per cent sea water corre-
sponds to specific gravity 1.025, and 0 per cent sea water is distilled water (see
Table I).

hatched, respectively. The percentages obtained from successive obser-
vations were plotted against time to give signoid curves, and the times
at which 50 per cent had emerged, and at which 50 per cent had hatched,
were determined from these curves by interpolation.

RESULTS

Five to nine experiments (involving counts on a total of 500-1,000
viable cysts) were carried out at each salinity. Throughout the range
of salinities used, and in distilled water, approximately 60 per cent of

SALINITY AND ARTEMIA EXCYSTMENT RATE

199

the stock mixture of cysts and empty shells emerged (see method), i.e.
all of the embryos which are presumed to have been viable emerged in
all of the solutions. More than 96 per cent of the embryos that emerged
from the shell also hatched from the membranous sac in 50-225 per cent
sea water, inclusive, but in lower salinities this percentage decreased and
no true hatching at all occurred in distilled water (Fig. 1). In distilled
water the appendages of the emerged embryos did not move. The
membranous sac disintegrated after several hours in about one-fifth of

225

200

175

ISO
125

•*w

100

PERCEA

O EMERGED
0 HATCHED

10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

HOURS UNTIL 50 % EMERGED AND HATCHED

FIG. 2. The rates of emerging and hatching in diluted and concentrated sea
water at 25° C. One hundred per cent sea water corresponds to specific gravity
1.025, and 0 per cent sea water is distilled water (see Table I). The curves
show the time lapse until 50 per cent of the embryos in populations emerge, and until
50 per cent hatch (see text).

the cases, causing a sort of pseudo-hatching, but in these cases the
nauplii always swelled, and often burst near the first joints of the large
second antennae. Inactivity of the embryo probably interferes with
hatching. Occasional cysts burst within 20-30 minutes after being
placed in distilled water and aborted amorphous masses. This also oc-
curred rarely in 12% per cent sea water and in higher salinities as well.
The embryos that hatched throughout the range 12^4-225 per cent

200 R. H. JENNINGS AND D. M. WHITAKER

sea water were all normal, active, and viable in the salinity in which they
excysted. They moulted once before dying of starvation on about the
fourth day. No food was provided and no attempt was made to deter-
mine the salinity requirements of more advanced developmental stages.
The effect of salinity on food organisms is a complicating factor after
the yolk has been consumed.

The effect of salinity on the rate of emergence and hatching is shown
in Fig. 2. Each point represents the average of the recorded time lapses
until 50 per cent had emerged and until 50 per cent had hatched in the
several experiments at each salinity. The results of the individual ex-
periments were quite consistent. It may be seen in Fig. 2 that the rates
of emergence and hatching are little affected by osmotic pressure within
the salinity range 25-125 per cent sea water, but this is not true in higher
and lower salinity.

SUMMARY AND CONCLUSIONS

1. The excystment of Artemia takes place in two principal stages:
first, the quiescent nauplius emerges from the shell of the cyst within a
membranous sac, and then later the nauplius hatches from the sac and
swims actively about.

2. The excystment of Artemia obtained from the margin of San
Francisco Bay has been studied at 25° C. in diluted and concentrated sea
water over a salinity range from zero (distilled water) to 225 per cent
sea water (i.e., a solution in which the salt concentration is 225 per cent
of the salt concentration of sea water. No salts precipitated out).

3. The same percentage of embryos emerged from the shells in all
of these salinities, including zero salinity (distilled water).

4. In distilled water the emerged embryos are motionless and they
do not hatch from the sac. Some swell and burst.

5. In l2l/2 per cent sea water, 83 per cent of the emerged embryos
hatch; in 25 per cent sea water 93 per cent hatch. In 50-225 per cent
sea water 96-99 per cent hatch.

6. In 12%-225 Per cent sea water the nauplii that hatch are normal,
active, and viable. They moult to form metanauplii before dying of
starvation (in the absence of food) on about the fourth day.

7. In the salinity range 25-125 per cent sea water, the rates of
emergence and hatching are practically constant (and therefore inde-
pendent of change in osmotic pressure).

8. In the salinity range 150-225 per cent sea water the rates of
emerging and hatching decrease with increasing salinity, but the interval
between emergence and hatching is nearly constant throughout the range
25-225 per cent sea water.

SALINITY AND ARTEMIA EXCYSTMENT RATE 201

9. Emergence is accelerated in l2l/2 and 0 per cent sea water, but
hatching is retarded in 12% per cent sea water and is inhibited in dis-
tilled water.

BIBLIOGRAPHY

BOND, R. M., 1932. Observations on Artemia " franciscana " Kellogg, especially

on the relation of environment to morphology. Int. Rev. der. ges. Hydro-

biol. und Hydrogr., 28: 117-125.
BOONE, E., AND L. G. M. BAAS-BECKING, 1931. Salt effects on eggs and nauplii

of Artemia salina L. Jour. Gen. Physiol., 14 : 753-763.
HEATH, H., 1924. The external development of certain phyllopods. Jour. Morph.,

38: 453-483.
JACOBI, E. F., AND L. G. M. BAAS-BECKING, 1933. Salt antagonism and effect of

concentration in nauplii of Artemia salina L. Communications from the

Leiden Botanical Laboratory, No. 2.
MARTIN, E. G., AND B. C. WILBUR, 1921. Salt antagonism in Artemia. Am. Jour.

Physiol., 55 r 290-291.
WHITAKER, D. M., 1940. The tolerance of Artemia cysts for cold and high

vacuum. Jour. E.vper. Zoo!., 83 : 391-399.

SPERM ACTIVATION BY ARBACIA EGG EXTRACTS, WITH
SPECIAL REFERENCE TO ECHINOCHROME

IVOR CORNMAN

(From the Marine Biological Laboratory, Woods Hole, Mass.1)

It has long been known that the eggs of various marine invertebrates
secrete substances which markedly affect the behavior of sperm. One
has merely to rinse eggs in sea water and add this water to a sperm
suspension to produce striking changes. These effects are classified by
Lillie (1924) as activation, aggregation, and agglutination. Activation
is a stimulation of the sperm, bringing them instantaneously from an
inactive state (as in the testis) to a high pitch of activity. Aggregation
constitutes the gradual accumulation of the sperm within a region of high
concentration of various agents, and in the case of egg-secretions, ap-
pears to be a chemotaxis. Agglutination is the clumping of the sperm
exposed to egg-secretions. While some properties of the substances
active in egg-secretions are known (Tyler and Fox, 1940), the sub-
stances themselves have not been isolated from eggs in pure enough
form that we can attribute these properties to definite chemical entities.
In 1939, Hartmann, Schartau, Kuhn, and Wallenfels reported that
echinochrome, the pigment which gives Arbacia eggs their reddish color,
produces the same stimulation in Arbacia pustulosa sperm as does the
egg-secretion itself, and is effective in dilutions as great as 1 : 2,000,000,-
000. An attempt was made to duplicate these results, using Arbacia
punctulata sperm, and crystalline echinochrome kindly supplied by Dr.
E. G. Ball, which he had isolated from A. punctulata eggs (1934). No
stimulating effect could be detected. Subsequently there appeared a
fuller account by Hartmann and Schartau (1939), and a report by Tyler
(1939) of negative results with Strongylocentrotus purpuratus. Be-
cause of this, it seemed worthwhile to repeat and extend the experiments
with A. punctulata.

Echinochrome as the Activator

In preparing the solutions, both sea water and isotonic sodium chlo-
ride were used. In sodium chloride, the sperm do not agglutinate, which

1 Present address : Department of Zoology, University of Michigan. The
laboratory space at Woods Hole was obtained through the kindness of Dr. Robert
Chambers.

202

SPERM ACTIVATION BY EGG EXTRACTS 203

sometimes facilitates comparison of sperm activity. The egg-secretions
were obtained by suspending one volume of eggs in 100 volumes of sea
water (or NaCl) for a half -hour. The liquid, filtered free of eggs, is
generally referred to as egg-water, and is extremely effective in bringing
about activation and agglutination. The concentrations of echinochrome
ranged from 1 : 2,500,000,000 to 1 :25,000. At the latter concentration,
the echinochrome solution is pink in color and therefore is well above
the concentration of echinochrome in active egg-water, which was color-
less. The activating properties of the echinochrome solutions were
tested immediately after preparation, to avoid possible loss of the echino-
chrome through its precipitation as a calcium salt in sea water, or its
decomposition in alkaline solutions.

Sperm respond readily to differences in pH. Therefore, in testing
potency of echinochrome, particular attention was given to the control
of pH in all solutions used. Glycyl-glycine (.001 M) and piperazine
(.001 M) were used as buffers in preference to, phosphate, which tends
to precipitate calcium and magnesium from the sea-water solutions (Ty-
ler and Horowitz, 1937). In these experiments little strain is placed
upon the buffer systems, and measurements with the glass electrode
showed these low concentrations to be adequate. In any single series,
the egg-water and echinochrome solutions, and the sperm suspensions
were prepared with the same buffer and the pH was measured with a
glass electrode before and after the activation tests were made, as fur-
ther precaution against differences in hydrogen ion concentration. Most
of the tests were carried out at a pH level where the sperm were inactive,
but were readily activated when egg-water at the same pH was added.
This pH value was found to be in the neighborhood of 6.0 for the sea
water and 7.5 for the isotonic sodium chloride solutions. The actual
pH values varied with the individual sea-urchin, and increased if the dry
sperm was allowed to age. In these experiments the absolute value is
not important, since in every case a control test with egg-water was made
along with each test of echinochrome. Accordingly, if echinochrome is
the activating agent in egg-water, it should show activating properties
at the same pH as the egg-water.

Two methods of testing were employed. In one, the dry sperm, that
is, the sperm taken directly from the testis with a minimum of moisture,
was diluted to about 1 : 100 in buffered sea water or isotonic NaCl. A
drop of this was covered with a cover-glass, and the egg-water and
echinochrome pipetted into opposite sides of the drop. In this way the
slightest response of the sperm could be detected and a precise com-
parison made between the two solutions. Adding dry sperm directly
to the test solution sometimes gave more spectacular differences in re-

204 IVOR CORNMAN

sponse, but where small differences are involved, the first method is freer
from subjective interpretation.

In no case did echinochrome activate the sperm. Each test was ac-
companied by a test with egg-water at the same pH, in which activation
did occur.

Additional tests were carried out at higher pH values to supply more
nearly normal conditions for the sperm. The results are not as clear-cut
as with inactivated suspensions, since differences in speed of sperm are
hard to estimate. However, in no case could it be said that the echino-
chrome definitely produced an increase in motility greater than did mere
dilution with buffered sea water, whereas stimulation by egg-water could
usually be seen clearly.

Chemotaxis in Echinochrome

To check the reported chemotactic effect of echinochrome, a few tests
were made to compare the migration of sperm up glass capillary tubes.
Tubes of the same diameter were washed and filled with sea water, egg-
water, and echinochrome solution. The ends were then inserted into
buffered sperm suspension, and migration measured at various times.
The results were so variable that none of the solutions could be said to
be definitely chemotactic on the basis of these few trials. Variations in
the alkalinity of the glass probably played some part, since the volume
of solution was small in proportion to the surface of the tube, and the
buffer capacity of the solutions was low. Under such conditions a
shift toward alkalinity could occur and give an illusory chemotactic effect
by merely speeding the progress of the sperm.

Activity of the Echhwchrome-protcin Complex

The complex which echinochrome forms with proteins from the Ar-
bacia egg was reported by Kuhn and Wallenfels (1940) to have sperm-
stimulating properties in even greater dilutions (1:300,000,000,000)
than echinochrome alone, in uncombined form. A similar echinochrome
complex was extracted from Arbacia punctulata by their method. Eggs
were frozen, crushed, and extracted with sea water. After filtering, an
equal amount of saturated ammonium sulphate solution was added,
bringing down a rose-colored precipitate which redissolved in sea water
without leaving any residue. It was purified by repeated precipitation,
centrifugation, and decantation.

This complex, buffered and tested in the same way as echinochrome,
both activated and agglutinated punctulata sperm. In biological proper-
ties and solubility, therefore, it is the same as the tertiary complex ob-

SPERM ACTIVATION BY EGG EXTRACTS 205

tained by Kuhn and Wallenfels. In view of the tests with pure echino-
chrome, however, the activity of the complex would seem to be centered
in the protein moiety rather than in the echinochrome.

Separation of the Agglutinating and Activating Properties of Egg-water

To determine more of the nature of the activating agent, egg-water
was dialyzed against sea water for a half-hour, then both fractions buf-
fered and tested for activity. The dialysate stimulated without agglu-
tinating, while the residue both stimulated and agglutinated sperm.
Dialysis, then, can separate the agglutinating and activating agents.
Distillates obtained by gently boiling each fraction also showed activating
properties, but the activity disappeared shortly after the distillate was
buffered to pH 6. The original, unboiled dialysate retained its activity
24 hours, beyond which it was not tested. Similarly, a distillate from
a repeatedly precipitated and washed sample of the echinochrome-
protein complex could stimulate, when unbuffered (pH 9.0), and lost its
activity within an hour after it was buffered to pH 6.6. This disappear-
ance of stimulating activity from the distillates suggests that the activat-
ing agent had been altered during distillation. Improved methods of
separation will probably yield a stable stimulating fraction. At present,
it is important that distillates of egg-water, of egg-water dialysate, and of
the echinochrome complex are similar in that they contain an activating
substance.

DISCUSSION

The absence of visible response of A. punctulata sperm to echino-
chrome is in agreement with Tyler's investigations with Strongylocentro-
tus purpuratus. He found that echinochrome brought about no increase
in oxygen consumption of the sperm or eggs. On the other hand, these
results do not agree with the observations of Hartmann and Schartau on
A. pustulosa, which was found to be extremely sensitive to echinochrome
solutions. The difference in response of A. pustulosa and A. punctulata
could be attributed to species difference, although this would make the
similarity of A. punctulata to the more distantly related Strongylocen-
trotus appear somewhat anomalous. Another possibility is that Hart-
mann and Schartau did not control pH in their solutions, since it is not
mentioned in any of the papers on A. pustulosa. Their results, particu-
larly the activity of highly dilute solutions (1:2,500,000,000), suggest
that the activation is due to the normal alkalinity of the sea water used
as a solvent.

Whatever the final answer may be with regard to echinochrome, we

206 IVOR CORNMAN

must search farther for the answer to the general problem of sperm-
activation by egg-secretions. Echinochrome is limited in occurrence,
even within the class Echinoidea. Moreover, the egg-secretions from
pigmented eggs will stimulate sperm from unpigmented species, and vice
versa (cf. Woodward, 1918, Table I).

In the Arbacia egg there is some sperm-activating substance which
will dialyze through a collodion membrane. It can, then, be separated
from the agglutinating substance with which it is closely associated, but
which will not dialyze. However, the echinochrome-protein complex
carried with it through seven precipitations the power to activate as
well as agglutinate sperm. In view of the ease with which the activator
dialyzes, one might well expect it to be washed completely free from the
agglutinating substance. Tyler (1939) also reports that partial puri-
fication of agglutinin from the keyhole limpet does not free it from
activating properties. This leads one to suspect that there may be two
substances present which activate sperm : one closely attached to the
agglutinating substance, and one easily separated from it. On the other
hand, the activating properties of rough distillates from the egg-water
dialysate and the partially purified echinochrome-protein complex is some
evidence of similarity between the activating agents in both, but it does
not prove identity.

Earlier work on extracts from echinoderm eggs offers a possible
explanation of these observations. Woodward (1918) obtained, in ad-
dition to an ammonium sulphate precipitate of agglutinin, a barium
chloride precipitate which showed lipolytic activity. Glaser (1921)
pointed out that this lipolysin (and pancreatic lipase as well) could
activate sperm. Yet, if this lipase is the activating agent in egg-water,
it follows that the activator in the distillates must be some substance
other than the lipase, since a protein would not distil. In this respect
the work of Clowes and Bachman (1921) takes on added significance.
They were not only the first to obtain sperm-activating distillates from
egg-water, but also found that higher alcohols (propyl, allyl, and cin-
namyl) and related substances activate sperm. Bringing these findings
together, one might tentatively suggest that the immediately effective
agent in sperm activation is an alcohol freed by the lipase. Then the
presence of the alcohol or of the lipase would be adequate to produce
sperm activation. It remains to be demonstrated that the sperm-acti-
vator that follows the agglutinating fraction is the lipase, and the dialysa-
ble, distillable activator is the product of the activity of that lipase. The
hypothesis fits the framework of assembled facts, but substantiation will
require considerable further investigation of egg-water fractions and
egg extracts.

SPERM ACTIVATION BY EGG EXTRACTS 207

SUMMARY

Echinochrome is not the agent in A. punctnlata egg- water which
stimulates sperm. The echinochrome-protein complex precipitated from
the extract of crushed eggs by (NH4)2SO4 is an effective sperm stimu-
lator. From egg- water dialysate and from the echinochrome-protein
complex a distillate can be obtained which has sperm-activating proper-
ties. It is tentatively suggested that a higher alcohol freed by a hydro-
lytic agent in the egg-water is the stimulating substance acting directly
upon the sperm.

BIBLIOGRAPHY

BALL, E. G., 1934. Isolation of crystalline echinochrome (abstract). Biol. Bull.

67 : 327-328.

CLOWES, G. H. A., AND E. BACHMAN, 1921. A volatile, sperm-stimulating sub-
stance derived from marine eggs (abstract). Jour. Biol. Chem., 46:

xxxi— xxxii.

GLASER, O., 1921. Fertilization and egg-secretions. Biol. Bull., 41 : 63-72.
HARTMANN, M., AND O. SCHARTAU, 1939. Untersuchungen iiber die Befruchtungs-

stoffe der Seeigel. I. Biol. Zentralbl, 59: 571-587.
HARTMANN, M., O. SCHARTAU, R. KUHN, AND K. WALLENFELS, 1939. Uber die

Sexualstoffe der Seeigel. Natunviss., 27 : 433.
KUHN, R., AND K. WALLENFELS, 1940. Echinochrome als prosthetische Gruppen

hochmolekularer Symplexe in den Eiern von Arbacia pustulosa. Ber.

deutsch. chem. Ges., 73: 458-464.
LILLIE, F. R., AND E. E. JUST, 1924. Fertilization. Cowdry's General Cytology,

Sect. VIII, pp. 481-536. University of Chicago Press, Chicago.
TYLER, A., 1939. Crystalline echinochrome and spinochrome : Their failure to

stimulate the respiration of eggs and of sperm of Strongylocentrotus.

Proc.. Nat. Acad. Sci., 25 : 523-528.
TYLER, A., AND S. W. Fox, 1940. Evidence for the protein nature of the sperm

agglutinins of the keyhole limpet and the sea-urchin. Biol. Bull., 79 :

153-165.
TYLER, A., AND N. H. HOROWITZ, 1937. Glycylglycine as a sea water buffer.

Science, 86: 85-86.

WOODWARD, A. E., 1918. Studies on the physiological significance of certain pre-
cipitates from the egg-secretions of Arbacia and Asterias. Jour. Exper.

Zool, 26: 459-501.

THE NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

CONFUSUM DUVAL

I. THE SURVIVAL OF ADULT BEETLES ON PATENT FLOUR AND
COMPLETE STARVATION DIETS x

B. AUBREY SCHNEIDER 2

(From the Department of Biology, The Johns Hopkins University School of
Hygiene and Public Health, Baltimore, Md.)

INTRODUCTION

The attention of numerous workers has recently been drawn to the
problem of determining the nutritional requirements of insects. The
problem has been attacked along two general lines of procedure, namely :
(1) by observations on the ability of an organism to survive on a given
diet ; and (2) by analyses of the different regions of the digestive tract
of the organism to determine the presence or absence of the various
digestive enzymes. The present series of studies on Tribolium con-
cerns the former aspect of the problem. The importance of determining
the nutritional requirements of Tribolium was emphasized in an earlier
paper (Schneider, 1940) on thyroid feeding. There it was stated that
' When the fields of the nutritional requirements and the endocrinology
of insects have been thoroughly worked, we shall be in a much better
position to determine the effect of a given vertebrate hormone extract
on a given insect than we are at present."

Chapman (1924), Sweetman and Palmer (1928), Street and Palmer
(1935), Nelson and Palmer (1935), Bushnell (1938), and Chin and
McCay (1939), all working with Tribolium, studied the nutritional re-
quirements with respect to group biology or population growth and
maintenance. Since there are many factors in addition to the nutritional
quality of the food medium which affect population growth, i.e., popu-
lation density, cannibalism, and " conditioning " of the food medium
(Park, 1934a), it seemed expedient to study the problem from a more
fundamental viewpoint, namely that of determining the individual per-
formances of isolated beetles fed on various diets. From logical consid-
erations it appears that an approach to the problem from this point of

1 From the Department of Biology, School of Hygiene and Public Health,
Johns Hopkins University.

- The writer is greatly indebted to the late Dr. Raymond Pearl for his helpful
guidance and criticism during the progress of this investigation, and to Mr. Myron
L. Simpson for his willing assistance in the experimental routine.

208

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM 209

view is likely to lead to a sounder foundation of the nutritional require-
ments of insects per se, than can be attained from following any other
line of investigation.

The present investigation has indicated that the starvation control is
of considerable value in nutritional studies, and is particularly desirable
in this case in view of the long life (up to three years) of the organism
on the so-called " normal diet," flour. The term " normal diet " is here
used in the sense that the beetles have been found to grow and reproduce
in it, though it may not really be adequate for optimal physiological
processes. The value of a particular diet can be determined just as
accurately by comparison of the performances of the animals fed on it
with those of animals on complete starvation, as it can by comparison
with those of animals fed the " normal diet," and in this particular case
it is certainly less time-consuming. In other words, if we start with
the starvation diet as the zero point on a scale, all other diets, including
the present so-called " normal diet " can be referred to the zero point
in terms of either positive or negative values, i.e., they will be either
better or worse than no food at all.

STATEMENT OF THE PROBLEM

Pearl and Parker (1924) have stated that ". . . when we study
duration of life under normal conditions we are dealing with the com-
bined effects of two variable complexes, inborn organization, on the one
hand, and environment, including renewal of available energy and sub-
stance by food, on the other hand." The former complex, inborn
organization, designated by Ashby (1930) as the "capital" of the or-
ganism, has also been termed the inherent vitality of the organism.
Since the knowledge of the inherent vitality of Tribolium is primarily
essential to the knowledge of its general nutritional requirements, the
present investigation has been designed to deal principally with this
phase of duration of life. Hence, the problem may be stated as that of
determining the ability of the adult beetle, Tribolium confnsum Duval,
to survive under conditions of complete starvation. Several more spe-
cific problems dealing with the influence of various environmental factors
of pre-imaginal life on the ability of the adult beetle to survive conditions
of complete starvation are involved.

TECHNIQUE

Park (1934&) has described in detail the general technique for han-
dling Tribolium cultures in the laboratory, and there is no need for its
repetition here. It is sufficient to state that the beetle spends its entire
life history in flour, and by the use of sieves of various sized mesh, the

210 B. AUBREY SCHNEIDER

adults, eggs, and all immature forms can be separated from the flour for
observation. The sex is determined from the pupal characteristics.

Throughout the present investigation, when eggs were used to start
an experiment, a sufficient number was taken from a general stock
culture, and allowed to hatch in dishes containing a small quantity of
patent flour (Ceresota). The date of hatching was determined within
0.5 day, and the larvae were placed in bottles in a definite quantity of
flour (indicated in the experiments). The bottles were kept in a
darkened constant temperature incubator at a temperature of 29° C.,
and a relative humidity of approximately 40 per cent. The food was
not changed, and the bottles were not sifted until about the twentieth
day after hatching, when it became necessary to sift them once a day to
collect pupae. The date of pupation was recorded to the nearest 0.5
day (the color of the pupa indicating whether the pupal stage was reached
within the past 0.5 or 1.0 day). The pupae wrere sexed and placed in
sterile vials (1.5 cm. diam. X 2.5 cm. tall) and kept in the incubator
under the above-mentioned physical environmental conditions. The date
of emergence was recorded to the nearest 0.5 day, and the discarded pupa
cases were removed from the vials. The vials were closed with corks
covered with cellophane, the corks having a filed groove along the side,
and the cellophane being perforated so as to maintain a normal oxygen
supply within the vials. The experience of the writer in having beetles
escape by eating their way through the cork along the filed groove
prompted the use of the cellophane covering. This prevented the boring
activity in all but a very few vials where the cellophane was broken,
and of course these few cases were immediately discarded. The date of
death was recorded to the nearest day, and the period from emergence
to death (adult survivorship) was calculated.

When only adults of a known age and sex were needed for an
experiment, it was necessary to start by collecting pupae from a general
or specific stock culture (indicated in the experiments). These were
sexed and placed in vials, from which point the procedure was the same
as that described above.

In the one experiment where flour- fed controls were used, the vials
were supplied frequently with plenty of fresh flour, so that the organism
was always presented with an unlimited quantity of food. Data on the
weights of pupae, live and dead images, have been reserved for a later
publication.

The statistical data are based on frequency distributions of the varia-

bles studied. The ^- 1^ test for significance was used ; any value

r . iL, Dirt.

higher than 3 indicating that the difference was probably not due to
sampling errors.

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

211

OBSERVATIONS AND RESULTS

Survivorship of Adult Beetles, Taken as Pupae from a General Stock
Culture and Subjected to Flour and Starvation Diets upon Emergence

In the first experiment 432 pupae (216 of each sex) were isolated
from a general laboratory stock culture, and were observed daily for
emergence. Since these forms were taken from a general stock cul-
ture, they were reared under identical, though unknown factors of larval
density. Upon emergence, 100 of each sex were submitted to conditions
of complete starvation, and 116 of each sex were given a flour diet.
Daily observations were made for mortality, and the time of death was
recorded. The time intervening between emergence and death was
calculated as adult survivorship. Failure of the organism to respond
to agitation with a small camel's hair brush was taken as the criterion
for death.

Table I presents the data relative to the survivorship of these two
groups of organisms.

TABLE I

Survivorship of adult beetles on flour and starvation diets.

Group

Range

Mean
number days
of life at end
of 24 days

Median
number days
of life at end
of 24 days

Standard
deviation
(days)

Coefficient
of
variation
(per cent)

Min.
(days)

Max.

(days)

Starved adult males ....
Flour-fed adult males. . .
Starved adult females. . .
Flour-fed adult females .
Total starved adults. . . .
Total flour-fed adults. . .

2.5
2.5
3.0
2.5
2.5
2.5

21.0
24.0
23.5
24.0
23.5
24.0

14. 235 ±0.274
21.323 ±0.370
13. 535 ±0.314
22.964 ±0.240
13.885 ±0.213
22. 166 ±0.220

14.665 ±0.343
24.181 ±0.464
14.000 ±0.394
24. 221 ±0.301
14.593 ±0.267
24. 202 ±0.276

4.062 ±0.194
5.908 ±0.262
4.648 ±0.222
3.826±0.169
4.459 ±0.150
4.974±0.156

28.53 ±1.46
27.70±1.32
34.34±1.82
16.16±0.76
32. 11 ±1.19
22.43 ±0.74

100
116
100
116
200
232

The data are presented for the sexes separately and then combined.
Since all the beetles on starvation were dead at the end of 24 days, it
appeared justifiable to compare their survivorship performances with
those of the flour-fed group up to the end of 24 days. This procedure
obviated the necessity of observing the flour-fed group for their entire
life (which may have been up to three years). For a study of the
complete survivorship record of flour- fed Tribolium, the reader is
referred to Pearl, Park and Miner (1941). In the present case it was
not possible to calculate mean length of life of the flour-fed group;
instead there is calculated the mean number of days of life lived up to
the end of 24 days per beetle exposed to risk.

The difference between the mean length of life of the starved beetles
(13.885 ± 0.213 days) and the mean number of days of life at the end

212

B. AUBREY SCHNEIDER

-C5
•

- 1

Starved

Flour-fed

gether

0-2

..O

C o

• —

I S.|

,3
~a

V Ml" «-O

-° ^5 « 2

' "—'S

L ^ «-

"C* rt ^

p ° »

C3
P

Jf&l

5> MI'S OT3

•S ™ f3 Sf °

>> aj
°

OOO

OOOOi— '•

I-H ^^ i-l (M CO C

OtM*-iO
eD t^O I
-Tf lO O 1

ooooot— e

ffOCOCOC^<MC<lr>}C^

H^-tf-Hi— lOOOOOlOSOSOi

O O O O O5 C"j ~- O* '~j O^ O^ Oi O5 CT> O CT. O^ OS O) w) Oi Oi O^ OO OO

O O ^O GO O O OO Is*

COlCCOCCCC<Mi-l

1 1 1 ! 1 ! 1 I

lC^<Nc3<M

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

213

of 24 days of the flour-fed group (22.166 ± 0.220 days) is 8.281 ±
0.306 days (27.06 X P. E.). The sex difference, though not a signifi-
cant one in the case of the starved forms, indicates that the males on
the average survived slightly longer than the females. The opposite is
true for the flour-fed group, i.e., the females showed a higher mean
number of days of life at the end of 24 days than the males. The latter
difference amounts to 1.641 ± 0.441 days (3.72 X P. E.).

Survivorship Curves for Tribolium
Both $•>«• Tog«th«r

12 M

AGE IN DAYS

FIG. 1. Survivorship curves for the flour-fed and starved adults of Tribolium,
taken as pupae from a general stock culture. Flour-fed: jV = 232; observed 24
days. Starved : N = 200 ; observed entire life.

Table II is the ungraduated life table for these two groups of or-
ganisms. The four columns under each sex heading show: (1) the
absolute number of beetles living at the beginning of each day of age
and therefore exposed to the risk of dying during that day; (2) the
number of deaths actually occurring during each day of age; (3) death

214 B. AUBREY SCHNEIDER

rate for each day of age per 1000 living at the beginning of the day;
and (4) survivors per 1000 exposed to risk at the beginning of each
day of age.

It is true that a few females of the starved group lived longer than
any males of that group, but the fact that a greater number of males
survived through the 12- to 14-day period accounts for the higher aver-
age survivorship of the males. In the flour-fed group the females took
an early lead in survivorship and maintained that lead at an increasing

TOO

too

900

/Stoi

I
I
I

100

.Flour-ttd

I 2 3 Jt i 6 7 8 9 10 II 12 IS 14 15 16 17 18 19 20 21 22 23 24

AGE IN DAYS

FIG. 2. Observed (ungraduated) death rates (deaths per 1000 exposed to
risk) for flour-fed and starved adult Tribolwm, taken as pupae from a general
stock culture.

rate to the end of the 24-day period of observation. Figure 1 shows
the survivorship curves for the starved and flour-fed groups, sexes
combined.

The only other factor of importance in discussing this particular
experiment is the mortality rate. This may be studied from the life
table for these two groups of beetles (Table II) and from the mortality
curves presented in Fig. 2.

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM 215

From Fig. 2 it is evident that the mortality rates exhibited no great
differences between the flour-fed and starved groups of organisms up
to the twelfth day, though during this time the rate for the starved group
was slightly higher, on the whole, than that for the flour-fed group.
After the twelfth day the mortality rate of the starved beetles went up
very rapidly, while that of the flour-fed group continued to fluctuate
just a little above the zero line. The point of interest here is that
starving the adult beetle for the first 10 or 12 days after emergence does
not seriously affect the rate of mortality. If starvation is continued
beyond this point, however, the necessity for an exogenous supply of
matter and energy for the maintenance of normal physiological processes
manifests itself in a sudden rise in the rate of mortality. Whether or
not the adult beetle will recover from the effects of starvation if returned
to a diet of flour on the twelfth day is a problem for further investiga-
tion.

An observation worthy of note here is that the starved beetles pro-
duced fecal pellets throughout their entire lives.

The question may be raised as to whether the starved female adults
may not have laid eggs and eaten them, and hence may not represent a
completely starved group. To answer this point, it can first be stated
with certainty that no eggs were ever seen in the vials containing the
starved females. This means that the females either ate the eggs im-
mediately after laying them, or that they never laid any eggs at all.
Further experimental evidence points to the latter conclusion as the
correct one. Virgin females from the flour-fed group were observed
carefully for daily egg-laying. In no case were any eggs found under
12 days. At the end of 20 days the cumulative egg-laying record of
185 virgin females in flour was as follows:

49 females had laid 0 eggs at the end of 20 days.
29 females had laid 1 egg at the end of 20 days.
22 females had laid 2 eggs at the end of 20 days.

20 females had laid 3 eggs at the end of 20 days.

21 females had laid 4 eggs at the end of 20 days.
8 females had laid 5 eggs at the end of 20 days.
6 females had laid 6 eggs at the end of 20 days.
13 females had laid 7 eggs at the end of 20 days.
4 females had laid 8 eggs at the end of 20 days.

2 females had laid 9 eggs at the end of 20 days.

3 females had laid 10 eggs at the end of 20 days.

2 females had laid 11 eggs at the end of 20 days.

3 females had laid 12 eggs at the end of 20 days.

216

B. AUBREY SCHNEIDER

It is clear that over 76 per cent of these beetles had laid fewer than
5 eggs each at the end of 20 days.

From this sample of 185 virgin, flour-fed females, it is evident that
in the absence of the male, the beginning of egg-laying is decidedly
retarded (mated, flour-fed females begin laying 2 or 3 days after
emergence). This fact, as well as the low mean length of life of starved
females, in addition to the poor body nutrition of the starved group at
the time when they would have started laying, all point strongly to the
probability that the starved females never laid any eggs during their
lives. Furthermore, the fact that the mean length of life of the starved
females was slightly less than that of the starved males indicates that
even if the females did lay a few eggs and eat them, this action did not
contribute greatly to their supply of nutrition.

TABLE III

Length of larval period for various larval population densities.

Larval density

Range

Mean length
of larval
period
(days)

Median length
of larval
period
(days)

Standard
deviation
(days)

Coefficient
of
variation
(per cent)

Min.
(days)

Max.
(days)

(a) 100 larvae per 100
grams flour

25.5
25.5
29.5

40.0
52.0
64.0

34.095 ±0.106
35.922 ±0.149
37.460 ±0.229

33. 431 ±0.133
35.281 ±0.187
36.038 ±0.287

2. 719 ±0.075
3.726±0.105
5. 642 ±0.162

7.97 ±0.22
10.37 ±0.29
15.06 ±0.44

297
285
276

(6) 10 larvae per 10
grams flour

It is evident that when an experiment such as this is started with
pupae from a general stock culture, nothing is known of the population
factors under which the larvae developed. If pupae are taken from
various stock cultures where larval density factors are almost certain to
be unequal, and if larval density does have an influence on adult inherent
vitality, then it is clear that the experiment has not been well controlled.
In the present experiment, the pupae were all taken from the same stock
culture, and hence developed as larvae under identical conditions, though
these conditions were probably not optimal. In order to test the in-
fluence of larval density on adult inherent vitality, and to find the larval
density conducive to the greatest adult vitality, further experiments were
made.

The Influence of Larval Density on Inherent Vitality of Adult Tribolium

Larval Development. — For this investigation, eggs were collected
from a general laboratory stock culture of Tribolium, and allowed to

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM 217

hatch in a small quantity of flour. Every 12 hours the newly-hatched
larvae were collected, and placed in various sized bottles under three
different food-density relationships. Each group contained 300 larvae
and the density relationships were as follows :

(a) 100 larvae in 100 grams of flour.

(b) 10 larvae in 10 grams of flour.

The larvae in (b) were kept in 1-ounce bottles; those in (a) and (c)
in half -pint milk bottles. Not only were the survivorship performances
of the resulting adults compared for the three series, but also the
developmental periods for the larvae.

Table III presents data relative to the length of time spent in the
larval stage for the three larval population densities.

From Table III it is evident that the larvae in density (a) developed
in less time than those from either (&) or (c}. Comparing the mean
length of larval period for group (a) with that of group (&) and (c),
respectively, we find differences of 1.827 ± 0.183 days (9.98 X P. E.)
and 3.365 ± 0.252 days (13.35 X P. E.). Each difference is a statisti-
cally significant one, and each points to a more rapid larval development
of the (a) group. Comparing (6) and (c), the difference of 1.538 ±
0.273 days (5.63 X P- E.) is also a statistically significant one, and
indicates a more rapid rate of development for the (b) than for the (c}
group. From the standpoint of larval development, density (a) pro-
duced the most rapid rate; density (b) the next most rapid rate; and
density (c) the slowest rate. These observations are substantially in
accord with those of Park (1938). The standard deviations indicate
that the variation in length of larval period is positively associated with
the absolute length of larval period ; i.e., with an increase in mean length
of larval period comes an increase in the amount of variation. This
fact may be grasped more quickly by reference to Fig. 3.

The curve representing density (a) is high and narrow; that for
density (c) is low and broad, tailing off far to the right; while that for
density (b) falls between these two extremes. Not only is the absolute
variation greater for the greater length of larval period, but also the
variation relative to the mean (coefficient of variation) as indicated in
Table III.

Of the 300 larvae which were started in each series 297 reached
pupation in density (a), 285 in density (b), and 276 in density (c),
indicating that the larval mortality was influenced by the density factor
also. The percentage mortality for the three groups was as follows :
(a) 1.00 ± 0.57 per cent; (b) 5.00 ± 1.26 per cent; and (c) 8.00 ±

218

B. AUBREY SCHNEIDER

1.57 per cent. Comparing these percentages we find the following
differences: (a) vs. (b), 4.00 ± 1.38 per cent (2.90 X P. E.) ; (a) vs.
(c), 7.00 ± 1.67 per cent; (4.19 X P. E.) ; and (b) vs. (c), 3.00 ±
2.01 per cent (1.49 X P- E.). Only one of these differences is statisti-
cally significant; that existing between groups (a) and (c). Clearly
something in addition to chance fluctuation caused the heavier mortality
in group (c). In all probability, it was the larval density factor which
was responsible for this effect. The difference between (a) and (b)
(2.90 X P- E.) is a borderline case of significance, and hence cannot
be considered a real difference. The difference between (b) and (c) is

Denut, lot

ntltjiol

0«n.,t,(c)

24 26 W SO X 34 36 98 40 42 44 46 48 50 92 54 96 58 60 62 64 66

TIME IN DAYS AFTER HATCHING

FIG. 3. Frequency distribution of length of larval period for various larval

population densities.

clearly not significant. These observations on the density factor as
related to larval mortality are also in accord with those of Park (1938).
It is easy to understand the differences that exist between densities
(a) and (c), since there are three times as many larvae per unit of
environment in (c) as there are in (a). However, the differences be-
tween (a) and (b} are more difficult to fathom in view of the fact that
the two had the same population density per unit of environment. The
only difference lay in the size of the total populations ; one was made up
of 10 larvae in 10 grams of flour, the other of 100 larvae in 100 gram?

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

219

of flour. Just why the 100/100 condition was conducive to a lower
larval mortality, and more rapid larval development than the 10/10 is
not apparent at the moment, though the phenomenon is certainly worthy
of further investigation.

Adult Survivorship. — After reaching the adult stage, the organisms
from the three larval densities described above were isolated and sub-
jected to complete starvation. The survivorship (period from emer-
gence to death) data for the sexes separately, then combined for the
three experimental groups are presented in Table IV.

There was only a small and statistically insignificant difference in
survivorship between the sexes in densities (a) and (c). In density

TABLE IV

Length of life of starved adults, reared under various larval population densities.

Larval density

Range

Mean length
of life
(days)

Median
length
of life
(days)

Standard
deviation
(days)

Coefficient
of vari-
ation
(per cent)

Min.
(days)

Max.
(days)

Male
(a) 100 larvae per
100 grams flour
(6) 10 larvae per
10 grams flour
(c) 300 larvae per
100 grams flour

1.5
3.0
3.5

25.0
23.0

21.5

18. 739 ±0.118
16.086 ±0.159
17.007 ±0.154

19.353 ±0.148
16. 500 ±0.199
17. 568 ±0.193

2.086 ±0.083
2.887 ±0.1 12
2.632 ±0.109

11. 13 ±0.45
17.94±0.72
15.47 ±0.65

142
150
133

Female
(a) 100 larvae per
100 grams flour
(&) 10 larvae per
10 grams flour
(c) 300 larvae per
100 grams flour

3.0
8.0
4.0

24.0
23.5
22.0

18. 753 ±0.173
17.088 ±0.061
17. 547 ±0.169

19.375 ±0.217
17.202 ±0.076
18.333 ±0.212

3. 169 ±0.122
1.018±0.043
2.918±0.119

16. 89 ±0.67
5.96 ±0.26
16. 62 ±0.70

152
125
136

Total
(a) 100 larvae per
100 grams flour
(6) 10 larvae per
10 grams flour
(c) 300 larvae per
100 grams flour

1.5
3.0
3.5

25.0
23.5
22.0

18. 763 ±0.128
16.414±0.105
17. 181 ±0.115

19.393 ±0.160
17.046 ±0.131
18.032 ±0.144

3. 251 ±0.090
2.593 ±0.074
2.794±0.081

17.32 ±0.49
15. 79 ±0.46
16.26±0.48

294

275
269

(b) this difference is 1.002 ±0.170 days (5.88 X P. E.). In all three
cases, however, the differences indicate that the females survived longer
than the males. For the combined sexes, there are large differences in
survivorship between the adults reared as larvae, under different popula-
tion densities. Group (a) showed the best survivorship; group (c}
the next best; and (b) the worst. A comparison of the mean length of
adult life of group (a) with that of group (b) and (c) respectively,
reveals differences of 2.349 ± 0.166 days (14.15 X P. E.) and 1.582 ±
0.172 days (9.20 X P- E.). In each case group (a) exhibited a su-
periority in survivorship which is far outside the limits of chance fluc-
tuation. A comparison of mean survivorship of groups (&) and (c)

220

B. AUBREY SCHNEIDER

brings to light a difference of 0.767 ±0.156 days (4.91 X P. E.), which
difference is also a statistically significant one. In this case, group (c)
possessed the greater survivorship value. These differences are illus-
trated by the survivorship curves in Fig. 4.

The influence of larval density on adult survivorship is clear, but
the exact differential effect of the different densities is not so apparent.

12 14

AGE IN DAYS

FIG. 4. Survivorship curves of starved adult Tribolium reared in various larval

population densities.

It is true that group (a) which showed the most rapid larval develop-
ment also exhibited the longest record of survivorship. But group (b),
which showed the second most rapid larval development, survived the
shortest period of time on starvation; and, conversely, group (c), which
showed the slowest larval development, possessed the second best adult
viability. These results indicate the desirability of further work to
determine the differential effects on growth, development, and inherent

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

221

vitality of various sized total populations kept at a constant density per
unit of environment.

This experiment has demonstrated that larval density is a factor
which must be controlled in investigations on adult inherent vitality.
From this it follows that the " capital " of the newly emerged adult
Tribolium is not solely dependent upon the inborn organization of the
organism, but also upon the environmental influences which have been
effective throughout its immature stages. In a study such as that of
Greiff (1940), where the length of life of isolated adults of Drosophila
melanogaster and that of its mutant, Ebony, under conditions of com-
plete starvation were compared, and where it was admitted that "... the
ebony mutant fly was observed to do better under laboratory conditions
than the wild-type fly " (i.e., it produced higher larval densities in the
stock bottles from which the flies were isolated), it would obviously have
been desirable to control the larval density factor.

TABLE V

Length of larval period for descendants of parents of indicated ages.

Age of
parents

Range

Mean length
of larval period
(days)

Median length
of larval period
(days)

Standard
deviation
(days)

Coefficient of
variation
(per cent)

Min.
(days)

Max.
(days)

1 month
6 months

30.5

25.5

54.0
48.0

37.516±0.154
33.896±0.175

37.416±0.193
33.618±0.219

3.830±0.109
4.235±0.124

10.21 ±0.29
12.49±0.37

279
265

To separate the environmental factor of larval density from the
inborn organization variable, and to test the relationship between age of
parent and inherent vitality of offspring, further experiments were made.

Influence of Age of Parents on Vitality of Adult Offspring

Larval Development. — Eggs from two different stocks, one a month
old, the other six months old, were collected and kept separate for
hatching, whereupon 300 larvae from each group were placed in flour
in half-pint milk bottles at a density of 100 larvae to 100 grams of food.
Every factor in the experimental procedure was held constant except
that of the age of the parent generations.

The data on larval development of these two groups of organisms
descended from parents of different ages are set out in Table V.

The mean figures for length of larval period for the two groups
reveal a difference of 3.620 ± 0.233 days (15.53 X P. E.). The prog-
eny of the six-months-old parents developed at a rate significantly more

222

B. AUBREY SCHNEIDER

rapid than those of parents one month old. The frequency distribution
curves in Fig. 5 illustrate this point clearly.

o
tr

cr 10

Offspring of
6 Mo. Old Parents

\0ffsprlng of
\ I Mo. Old Parent!

I
I

I
\

24 26 28 30 32 34 36 38 40 42 44 46 48 50 ^ 52 54
TIME IN DAYS AFTER HATCHING

FIG. 5. Frequency distribution of length of larval period of beetles descended from

stocks of various ages.

It is apparent that the six-months-old parents, having reached a more
highly productive and more mature period of life, produced progeny
which developed more rapidly than those produced by the younger, less
productive, one-month-old parents. The variability in length of larval

TABLE VI

Length of life of starved adults descended from parents of indicated ages.

Age of parents

Range

Mean length
of life
(days)

Median
length of life
(days)

Standard
deviation
(days)

Coefficient
of variation
(per cent)

Min.
(days)

Max.
(days)

Male
1 month
6 months

11.0
6.0

21.5
22.5

16. 196 ±0.136
17.450 ±0.136

16. 160 ±0.170
17. 791 ±0.170

1.992 ±0.096
2. 129 ±0.096

12.30 ±0.60
12. 20 ±0.56

97
111

Female
1 month
6 months

4.0

2.5

26.0
24.5

16. 697 ±0.177
18.005 ±0.174

17.000 ±0.222
18. 555 ±0.218

2.684 ±0.125
2. 443 ±0.123

16.07 ±0.77
13.57 ±0.69

104
90

Total
1 month
6 months

4.0

2.5

26.0
24.5

16.455 ±0.124
17.701 ±0.112

16.716±0.155
18.019±0.140

2.618 ±0.088
2.345 ±0.079

15.91 ±0.55
13.25 ±0.45

201
201

period for the two groups of progeny exhibited no striking difference.
The larval mortality for the progeny of one-month- and six-months-old

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

223

parents respectively, was 3.67 ± 1.08 per cent and 11. 67 ±1.85 per
cent. The difference here is 8.00 ± 2.14 per cent (374 X P. E.) and
indicates a significantly greater amount of mortality in the progeny of
the six-months-old parents than for those of one-months-old parents.

Survivorship of Adults. — Upon emergence, 201 adults from each
group were subjected to complete starvation. The numerical data on

1000
900

eoo

TOO
600
500

400

100
90

AGE IN DAYS

FIG. 6. Survivorship of starved adult beetles descended from parents of indicated

ages. N = 201 in each group.

length of life of starved adults descended from parents one and six
months old respectively are gathered into Table VI.

The sex difference in survivorship of the two groups is not large,
though the females again showed a slight superiority. When the sur-
vivorship of the two groups is compared for either females, males, or
for the combined sexes, there appear large differences. For the females,
the difference is 1.408 ± 0.248 days (5.68 X P. E.) ; for the males it is

224

B. AUBREY SCHNEIDER

1.254 ±0.192 days (6.52 X P. E.) ; and for the sexes together it is
1.246 ± 0.167 days (7.46 X P. E.). All of these differences are highly
significant and in each case point to a longer survivorship for the progeny
of 6-months-old parents than for those of one-month-old parents. The
difference in survivorship of the two groups (sexes combined) is illus-
trated in Fig. 6.

The sex ratios of the two groups, though not significantly different
from each other, or from a 50-50 ratio, are of interest in view of the
fact that the parents which produced the most rapidly developing larvae
and the most viable adults, also produced the highest proportion of male
offspring. This observation is essentially in accord with that of Law-
rence (1940) on Drosophila melanog aster.

Evidently parental age is a factor which must be controlled carefully
in experiments on inherent vitality. It follows that parent age must be
considered in determining the basic foundation of our knowledge of
the nutritional requirements of Tribolium.

TABLE VII

Survivorship of adult Tribolium subjected, at the indicated ages, to a starvation diet.

Starvation begun

Range

Mean length
of life after
starvation
(days)

Median length
of life after
starvation
(days)

Standard
deviation
(days)

Coefficient
of vari-
ation
(per cent)

Min.
(days)

Max.
(days)

(a) At emergence. . .
(6) At age 20 days .
(c) At age 125 days
(d) At age 220 days

4.0
2.5
3.0
4.5

26.0
26.5
20.0
14.5

16.455 ±0.124
16. 841 ±0.107
11.676±0.174
9.625 ±0.221

16.716±0.155
17.000 ±0.134
11.687 ±0.218
9.250 ±0.277

2.618±0.088
3.390 ±0.075
2. 892 ±0.123
1.937 ±0.156

15.91 ±0.55
20. 13 ±0.46
24.76±1.39
20.12±1.68

201
456

125
35

In order to determine the influence of the age of the adult beetle on
its ability to survive conditions of complete starvation, a final experi-
ment was performed.

Influence of Age on Inherent Vitality of Adults

Isolated adults fed a flour diet from the time of emergence to the
time of starvation made up three of the four series of organisms in this
experiment. Starvation was begun at ages 20 days, 125 days, and 220
days respectively in these three groups. Another group representing the
controls were starved from the day of emergence. The beetles in all
four series were descended from parents of equal age (six months) and
were reared under identical larval densities (100 larvae in 100 grams of
flour). The three groups of organisms that were fed for a while and
then starved represent select groups, since only those remaining alive at

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

225

the desired time were used. The factor of selection does not impair the
value of the experimental data, however, since we are not interested in
the viability of those that died before a particular age, but only in the
ability of those living at a certain age to survive when transferred to a
starvation diet.

The data relative to the survivorship of these four groups of or-
ganisms are presented in Table VII.

1000
9OO
600
70O
EOC
500

300 -

200 •

> 100

5 90
in

60
70

60
SO

8 10 12 14 16 18

TIME IN OATS AFTER BEGINNING STARVATION

FIG. 7. Survivorship of adult Tribolium placed upon a starvation diet at various

ages.

From Table VII it is evident that there exist significant differences
between each two mean values except in the case of groups (a) and (&),
which is only 0.396 db 0.164 days (2.41 X P. E.). This difference,
though statistically not certainly significant, points to a better survivor-
ship of 20-day-old adults than newly emerged adults when both are
subjected to starvation. The beetles of the two higher age groups all

226 B. AUBREY SCHNEIDER

survived a shorter period of time on starvation than did those of the
two younger groups. The influence of age on the ability of adults to
survive under conditions of starvation is illustrated graphically in Fig. 7.
The results suggest that the age at which the beetle survives best
under starvation is something more than 0 and less than 100 days. The
optimal point is probably in the neighborhood of 20 days. Investigations
are now in progress to determine the exact age at which the longest
survivorship under starvation occurs.

SUMMARY

The experimental data set forth in this paper have served to show
that:

1. Isolated adults of Tribolium confusum Duval, subjected to condi-
tions of complete starvation upon emergence, or shortly thereafter, will
survive up to 26.5 days depending upon the conditions of the experiment.

2. Survivorship of starved adults is shortened and the period of
larval development is lengthened by an increase in larval population
density.

3. Survivorship of starved adults is significantly longer and the
period of larval development is significantly shorter for progeny of
6-months-old parents than for those of one-month-old parents.

4. Survivorship of starved adults is shortened with increasing age.

LITERATURE CITED

ASHBY, ERIC, 1930. Studies in the inheritance of physiological characters. I. A
physiological investigation of the hybrid vigor of maize. Annals of
Botany, 44: 457-467.

BUSHNELL, R. J., 1938. The relation of nutritional levels to the growth of popu-
lations of Tribolium confusum Duval. I. Growth of larval, pupal and
adult populations in corn-meal and in corn-meal supplemented with yeast.
Ann. Ent. Soc. Amer., 31 : 345-351.

CHAPMAN, R. N., 1924. Nutritional studies on the confused flour beetle, Tri-
bolium confusum Duval. Jour. Gen. Physiol., 6 : 565-585.

CHIU, S. F., AND C. M. McCAY, 1939. Nutritional studies of the confused flour
beetle (Tribolium confusum Duval) and the bean weevil (Acanthoscelides
obtectus Say). Ann. Ent. Soc. Amer., 32: 164-170.

GREIFF, D., 1940. Longevity in Drosophila melanogaster and its ebony mutant in
the absence of food. Amer. Nat., 74: 363-376.

LAWRENCE, P. S., 1940. Ancestral longevity and the sex ratio of the descendants.
Human Biology, 12 : 403-429.

NELSON, J. W., AND L. S. PALMER, 1935. The phosphorus content and require-
ments of the flour beetle, Tribolium confusum Duval and a study of its
needs for vitamin D. Jour. Agric. Res., 50: 849-852.

PARK, T., 1934a. Studies in population physiology. III. The effect of conditioned
flour upon the productivity and population decline of Tribolium confusum.
Jour. Exper. Zodl, 68 : 167-182.

NUTRITIONAL REQUIREMENTS OF TRIBOLIUM

227

PARK, T., 19346. Observations on the general biology of the flour beetle, Tri-
bolium confusum. Quart. Rev. Biol., 9: 36-54.

PARK, T., 1938. Studies in population physiology. VIII. The effect of larval
population density on the post-embryonic development of the flour beetle,
Tribolium confusum Duval. Jour. Exper. Zool., 79: 51-70.

PEARL, R., AND S. L. PARKER, 1924. Experimental studies in the duration of life.
X. The duration of life of Drosophila melanogaster in the complete ab-
sence of food. Am. Nat., 58: 193-218.

PEARL, R., T. PARK, AND J. R. MINER, 1941. Experimental studies in the duration
of life. XVI. Life tables for the flour beetle, Tribolium confusum Duval.
Amer. Nat., 75 : 5-19.

SCHNEIDER, B. A., 1940. Thyroid feeding and metamorphosis. I. The influence of
various concentrations of thyroid substance on metamorphosis and growth
of the flour beetle, Tribolium confusum Duval. Jour. Exper. Zool., 84 :
113-139.

STREET, H. R., AND L. S. PALMER, 1935. Requirement of the flour beetle (Tri-
bolium confusum Duval) for vitamins in the B. group. Proc. Soc.
Exper. Biol. and Med., 32 : 1500-1501.

SWEETMAN, M. D., AND L. S. PALMER, 1928. Insects as test animals in vitamin
research. I. Vitamin requirements of the flour beetle, Tribolium con-
fusum Duval. Jour. Biol. Chem., 77 : 33-52.

THE PITUITARY REGULATION OF MELANOPHORES
IN THE RATTLESNAKE

HERMANN RAHN

(From the Department of Zoology, University of Wyoming, Laramie, Wyoming)

INTRODUCTION

Only recently has the pituitary been linked with certain color changes
in reptiles (Noble and Bradley, 1933; Kleinholz, 1935, 1938a, b; Parker,
1938). All of these observations are concerned with the lizard meta-
chrosis and indicate that the pars intermedia of the pituitary gland plays
an important role in the regulation of the melanophores. The present
study reveals that a similar concept must be extended to the snake,
since in this animal the dispersed phase of the melanophores is likewise
dependent upon the pituitary secretion. Furthermore, certain aspects of
the melanophores' arrangement and activity of the snake offer an inter-
esting contrast to those described for the chromatophores of the lizard.

MATERIAL AND METHODS

This study deals primarily with the observations of the effect of
hypophysectomies and subsequent replacement injections on the activity
of the skin melanophores of the prairie rattlesnake, Crotalus v. viridis
Raf. These specimens were collected in the eastern part of Wyoming.
Similar experiments were later extended to several other species of
snakes.

The operative removal of the snake pituitary is relatively simple,
since the lower jaw can be retracted sufficiently to expose the entire roof
of the oral cavity. Intraperitoneal injections of 10 per cent Nembutal
served for anesthesia as recommended by Clark (1937). After the
skin and muscles have been removed over the basi-sphenoid region the
more exact site of the gland can usually be seen through the semi-
transparent bone. The latter is chiseled away and then the whole gland
may easily be removed. After this operation the bone capsule is replaced
and the skin stitched. A strictly aseptic technique is not necessary.

The pars intermedia is very pale and may easily be distinguished from
the highly vascular and pink pars anterior. Since the intermedia tissue
in the rattlesnake is approximately one-third to one-fourth the size of
the anterior lobe and rather loosely attached to the latter, it may be
removed independently and completely.

228

COLOR CHANGES IN THE RATTLESNAKE

The effect of intermedin (melanophore-dispersing hormone) on the
operated animals was studied by means of intraperitoneal injections of
an extract of the pars anterior of the chicken pituitary. This tissue
contains an unusually high concentration of melanophore-dispersing hor-
mone and was prepared and assayed according to the method of Klein-
holz and Rahn (1939, 1940).

The effect of these various operations upon the ophidian melanophore
could most easily be observed and studied in the living, anesthetized
animal with a dissecting microscope, since the branches of these large
chromatophores extend into the epidermal layer and are contrasted
sharply with the homogeneous yellow background of the dermis. For
permanent recordings of the various changes scales were clipped, fixed
in alcohol and prepared as whole mounts. To supplement this series
other scales were serially sectioned at 10 micra.

These operations yielded very striking results in the rattlesnake and
led to similar experiments in other species of snakes. Although their
gross color changes were not as striking, the melanophores responded in
a similar manner to both hypophysectomies and injections. Altogether
seven rattlesnakes were hypophysectomized and several of these animals
were observed for four months after the operation. The other operated
snakes included five garter snakes, Thamnophis ordinoides; one ribbon
snake, Thamnophis radix (B. and G.) ; one bull snake, Pituophis s. sayi
(Schlegel) ; and four Florida water snakes, Natrix sipcdon pictiventris
(Cope).

OBSERVATIONS

The background color of the prairie rattlesnake is a homogeneous
light yellow color covered by various melanophores and pigmented epi-
dermal cells whose groupings and concentrations form various light- and
dark-brown scales. The latter are responsible for the typical color
pattern of the rattlesnake (Fig. 2). Before describing the induced color
changes, it seems desirable to consider first the histology of the rattle-
snake skin, since the arrangement and activity of its melanophores do
not correspond in all respects with those described for the lizards (von
Geldern, 1921 ; Schmidt, 1917; et al.).

Epidermis

One of the outstanding features of this tissue is the occurrence of
typical, branched melanophores and their processes which in the lizard
seem to be confined almost entirely to the dermis. The epidermis is
divided conveniently into two layers, the stratum germinativum and the
stratum corneum.

230

HERMANN RAHN

The cells of the germinativum are cuboidal to squamous and ar-
ranged in two or three rows. Among these cells one finds the cell body
and branches of a small melanophore which will be referred to as the
epidermal melanophore (Figs. 1, 5), although the cell body may often
lie just below the basement membrane. Whenever it appears in the
epidermis, it is probably a migrant from below. One must recognize
two varieties of this rather evenly branched epidermal melanophore.
One is large (200 micra) and found chiefly on the dorsal side where it
contributes to the color of the dark-brown scales (Fig. 5). The other
epidermal melanophore is much smaller (130 micra) and confined chiefly
to the lateral scales. This cell is more delicate in structure and contains
relatively little melanin (Fig. 9).

MEL.

DERM._.<

-MU5C.

FIG. 1. Schematized section through a scale of the rattlesnake. The melanin
deposits, MEL., of the stratum corneum, STR.C., are laid down first in the stratum
germinativum, STR.G., near the branches of the epidermal melanophore, EP.MEL.,
and the dermal melanophore, D.MEL. The dermis, DERM., contains a thick layer
of white-reflecting guanophores, GUA.W., covered by a thin layer of yellow-
reflecting guanophores, GUA.Y. Muscle is denoted by MUSC.

The dermal melanophore (Figs. 1, 7) is discussed below. Yet, it
must be briefly considered here as part of the epidermis, since its
branches are found among the epidermal cells, while its cell body is
always restricted to the dermis.

In the upper layer of the stratum germinativum the epidermal cells
begin to keratinize and it is here that the first signs of intra and some
inter-cellular melanin granules appear. Above this region is found the
stratum corneum where the cells are completely keratinized and flattened.
This region is destined to be cast at the next shedding. It has long been
observed in older studies that this layer, the shedding skin, contains
melanin granules which correspond in their horizontal distribution to
the pigmented areas below. Whether or not these pigmented epidermal

COLOR CHANGES IN THE RATTLESNAKE 231

cells (Figs. 1, 6) represent independent melanin producers is still ques-
tionable. At least they contribute greatly to the color intensity of the
dark dorsal scales where they are especially prominent. However, they
are not limited to this region alone, since they appear wherever melano-
phores occur and thus suggest a dependence on these cells.

Dermis

In this region are found the largest melanophores, the dermal melano-
phores (Figs. 1, 7) whose branches terminate among the epidermal cells.
Their cell bodies lie imbedded among the dermal chromatophores re-
sponsible for the white-yellow background color of these animals. The
latter pigment cells never extend into the epidermis. Their ramifying
processes are filled with very small, alcohol-resistant crystals which re-
flect a white light and are doubly refractive under the polarizing micro-
scope (GUA.W., Fig. 1). Thus these cells may be regarded as guano-
phores according to the classification of Schmidt (1917). This white-
reflecting guanophore stratum is in most places covered by a very thin
layer of yellow-reflecting guanophores (GUA.Y., Fig. 1) responsible for
the almost homogeneous yellow background color exhibited by the scales.
In a few places, this last-mentioned layer is absent (Fig. 1), such as the
white-tipped dorsal scales and the regions directly above each dermal
melanophore. In these places the dermis reflects only white color.

Effect of Hypophysectomy

The total removal of the pituitary as well as the extirpation of the
pars intermedia alone causes a complete pigment concentration in all
melanophore types. This would indicate that the pars intermedia is the
only pituitary tissue responsible for the normally dispersed state of these
cells. Evidence of a successful removal of this gland may be observed
one hour post-operatively. Both types of melanophores begin to con-
centrate their pigment and the long branches of the dermal chromato-
phores give the impression of disappearing below the epidermis leaving
small white islands in the otherwise yellow-reflecting guanophore area.
After 4 to 24 hours the concentration is complete. The dermal melano-
phore pigment is clumped in an irregular fashion (Fig. 8). Granules
of the guanophores do not seem to be affected.

It must be pointed out, however, that the snake as a whole does not
necessarily appear lighter after the operation. This paling awaits the
shedding of the stratum corneum, since this layer contains an abundance
of already deposited melanin which may obscure any changes in the
melanophores themselves. As soon as shedding has occurred, usually

HERMANN RAHN

three to four weeks after the operation, a striking change is observed
(Fig. 2). (For similar observations on the induction of shedding in
the snake after removal of the pituitary, see Schaefer (1933).) This
paling is permanent, since little, if any, pigment is deposited hereafter.

Injection of Intermedin Hormone

The intraperitoneal injections of intermedin into an hypophysecto-
mized animal cause complete dispersion in all melanophores. At least
one hour is necessary before the first effects can be seen, yet 12 hours
or more are required for the completion of this process (Figs. 4, 9).
Large doses (several thousand Anolis units) will maintain complete
dispersion for several days. The dermal melanophore dispersion, how-
ever, seems to be maintained by a lower intermedin concentration, since
it may remain in this phase for several days after the epidermal melano-
phores have already completely concentrated their pigment.

It is interesting to observe that an hypophysectomized rattlesnake
which has attained maximal pigment dispersion resulting from an inter-
medin injection is notably lighter than a normal animal (Figs. 3, 4).
This emphasizes again that much of the color intensity of this animal is
due to the melanin deposits in the keratinized portion of the epidermis
and not entirely to the state of melanin dispersion in the chromatophores.

PLATE I
EXPLANATION OF FIGURES

2. Two normal (dark) and two hypophysectomized (light) prairie rattle-
snakes. The control animal (x) has just shed and appears lighter than the other
control animal which is about ready to cast its skin.

3. An hypophysectomized rattlesnake before the injection of intermedin hor-
mone (anesthetized animal).

4. Same animal as in Fig. 3 photographed 24 hours after the injection of inter-
medin hormone. Notice that in spite of maximal pigment dispersion it is still
lighter than the normal control animal in Fig. 2. (For explanation see text.)

FIGS. 5-9. Photomicrographs taken from whole mounts of various scales.
X 140.

5. The large variety of epidermal melanophore which occurs primarily in the
dark dorsal scales ; maximal pigment dispersion.

6. Melanin containing epidermal cells (m.) from a dark, dorsal scale sur-
rounded by epidermal melanophores (ep. m.) whose pigment is concentrated.

7. Dermal melanophores from a lateral scale, notice the bush-like appearance,
finer branches, and lack of definite pattern.

8. Lateral scale of an hypophysectomized animal showing pigment concentra-
tion in the small variety of epidermal melanophore (ep. m.) and in the dermal
melanophore (d. m.).

9. Lateral scale from same animal as Fig. 8; 6 hours after intermedin injection.
Notice partial dispersion of pigment in both types of melanophores.

COLOR CHANGES IN THE RATTLESNAKE

233

'.X

4
•- ' *

\ '

d.rn.

• 8

PLATE I

234 HERMANN RAHN

If intermedin injections were carried out for a considerable length of
time in an hypophysectomized animal, one would expect to obtain the
same dark color possessed by a normal animal due to the gradual depo-
sition of epidermal melanin.

Other Snakes

A procedure similar to that outlined above was carried out with the
other species of snakes mentioned. In all these forms suitable scales
can be found which have small epidermal melanophores superimposed
on a light dermis. These cells, however, account for very little, if any,
color changes, yet will respond in the same way as the rattlesnake to
hypophysectomy and intermedin injection. The color pattern of these
snakes is accounted for primarily by a great concentration of various
dermal chromatophores which have not been studied.

DISCUSSION

The primary concern of this study is to point out that the pituitary
and more specifically the pars intermedia of the rattlesnake is responsible
for the dispersed phase of both types of melanophores. Noble and
Bradley (1933) were first to associate the paling response of the lizard,
Hemldaciylus, with the removal of the pituitary. Similar observations
were later made in much greater detail on Anolis by Kleinholz (1935,
193Sfl, b), and on the horned toad, Phrynosoina, by Parker (1938).
Little seems to be known concerning the normal color changes in snakes.
The experimentally induced metachrosis described above is relatively
slow but definite, and this group of reptiles may now be included in the
ever increasing number of vertebrates which have a pituitary regulation
of the melanophores.

The occurrence of epidermal melanophores and dermal melanophores
whose processes reach into the epithelial layer has long been recognized,
but has received little attention (Kerbert, 1877; Krauss, 1906; Fuchs,
1914; Schmidt, 1917; Lange, 1931). They seem to be rare in the liz-
ards (Schmidt, 1917) and are probably obscured in most snakes by the
dark, dermal chromatophore layer. In the rattlesnake, however, these
melanophores are especially striking, since their grouping and distribu-
tion seem to be responsible for the whole color pattern superimposed
upon a uniform yellow, dermal guanophore layer. The melanophores
are not crowded and are easily seen, since all their branches extend into
the epidermal stratum. This species is consequently peculiarly adapted
for this study. The other snakes studied are on the whole rather dark-
colored forms and the activity of the epidermal melanophores can only

COLOR CHANGES IN THE RATTLESNAKE 235

be recognized in a few light scales and even here they contribute little
to the color pattern.

All authors who have studied the shed reptile skin, especially that
of the snakes (Leydig, 1873; Werner, 1892; Lange, 1931) have ob-
served the close coincidence between the melanin pattern of the shed
skin and the pattern of the underlying layers. This can be followed
especially well in the rattlesnake where the melanin deposits of the
stratum corneum appear to coincide with the spread of each individual
dermal and epidermal melanophore. Such a situation poses the still un-
answered question of how this melanin is deposited. Is the melanin
formed independently by the epidermal cells, as seems to be the case in
various regions of the amphibian skin, or is it actually " injected " into
the epidermal elements by the branches of the melanophores as Strong
(1902) has described for the birds? In the snake the close association
of all melanophore branches with pigmented epidermal cells suggests very
strongly a mechanism of melanin deposition as in the bird feather. If
this is actually the case, then one could expect complete cessation of all
pigment deposits in an hypophysectomized animal, since the pigment is
completely concentrated and would never reach the upper layers of the
stratum germinativum where it seems to be normally laid down. Al-
though all hypophysectomized animals remained permanently pale as long
as four months after the operation, closer examination revealed slight
melanin deposits in the epidermal cells. However, this does not neces-
sarily invalidate this theory, since all these snakes received periodic inter-
medin injections and consequently might have had an opportunity to
deposit pigment during these intervals of melanophore dispersion.

To what extent the color changes in the snakes can be compared with
the relatively sudden changes described for the lizard group is difficult
to state. In Anolis (von Geldern, 1921 ; Kleinholz, 1938a) the light
(green) and the dark (brown) phases are due to the concentration and
dispersion of pigment in the dermal melanophores. These melanophores
never extend into the epidermis and achieve their effect by exposing or
masking the green-reflecting chromatophores of the dermis. This meta-
chrosis may be accomplished in a matter of minutes. In the rattlesnake
the light and dark phases are accomplished experimentally in a similar
manner, but the processes of the melanophores lie primarily in the epi-
dermis, give rise themselves to definite patterns and are rather slow to
react. As pointed out before, another factor must be considered in the
light and dark phase of the snake. This is the heavy accumulation of
melanin granules in the epidermal cells, which plays a minor or ques-
tionable role in Anolis. The melanin accumulation is especially notice-
able in the darkening of snakes before they shed (Fig. 2), for at this

236 HERMANN RAHN

time the future stratum corneum has already formed under the old
layer and undoubtedly contributes to the darkened condition as the new
layer already carries considerable pigment. The amount of melanin
deposits, however, seems to be in some way correlated with the melano-
phore activity, since hypophysectomized animals lay down very slight
amounts of pigment, or no pigment at all, and remain permanently pale.
Exact studies concerning the effect of temperature and light on the
chromatophore activity in snakes have not been found in the literature.
The role of these two factors in the melanophore regulation of various
lizards has recently received much attention by Kleinholz (1938a, &),
Parker (1938), and Atsatt (1939), but whether or not snake melano-
phores will respond in a similar manner awaits further study.

SUMMARY

1. In the prairie rattlesnake, Crotalus v. viridis Raf., the background
of the skin is a homogeneous yellow-white color reflected from the evenly
distributed dermal guanophores. The dark pattern of this snake is
formed by various distributions of melanophores and pigmented epi-
dermal cells superimposed upon this background.

2. Two main types of melanophores are found in the skin. One is
relatively small and resides primarily in the epidermis ; the other is much
larger, structurally different, and retains its cell body in the dermis, but
sends its ramifying processes into the epidermal layer.

3. Both of these melanophores appear to be associated with the depo-
sition of melanin granules in the keratinizing portion of the epidermis,
since (a) the distribution of its melanin deposits coincides with the pat-
tern of the underlying melanophores, and (&) the rate of pigment depo-
sition is greatly reduced after hypophysectomy.

4. The removal of the pituitary or the pars intermedia alone causes
a permanent paling due to the complete concentration of the melanophore
pigment. This paling, however, is more evident after the shedding of
the old keratinized epidermal layer carrying previously deposited melanin.

5. Intraperitoneal injections of intermedin from the chicken pituitary
will produce complete melanin dispersion in the melanophores of an
hypophysectomized animal.

6. Preliminary observations on four other species of snakes indicate
a similar pituitary regulation of the epidermal melanophores.

LITERATURE CITED

ATSATT, SARAH R., 1939. Color changes as controlled by temperature and light
in the lizards of the desert region of Southern California. Publ. Univ.
Calif. Los Angeles, Biol, Set., 1 : 237-276.

COLOR CHANGES IN THE RATTLESNAKE 237

CLARK, H., 1937. Embryonic series in snakes. Science, 85 : 569-570.

FUCHS, R. F., 1914. Der Farbenwechsel und ihre chromatische Hautfunktion der

Tiere. Winterstein: Handb. vergl. Physiol., 3: 1189-1656.
VON GELDERN, C. E., 1921. Color changes and structure of the skin of Anolis

carolinensis. Proc. Calif. Acad. Sci., 10: 77-117.
KERBERT, C. 1877. Ueber die Haut der Reptilien und anderer Wirbelthiere.

Arch. mikr. Anat., 13: 205-262.
KLEINHOLZ, L. H., 1935. The melanophore-dispersing principle in the hypophysis

of Fundulus heteroclitus. Biol. Bull, 69: 379-390.
— , 1938a. Studies in reptilian colour changes. II. The pituitary and adrenal

glands in the regulation of the melanophores of Anolis carolinensis.

Jour. Exper. Biol., 15: 474-491.
— , 19386. Studies in reptilian colour changes. III. Control of the light phase

and behaviour of isolated skin. Jour. Exper. Biol., 15: 492-499.
KLEINHOLZ, L. H., AND H. RAHN, 1939. The distribution of intermedin in the

pars anterior of the chicken pituitary. Proc. Nat. Acad. Sci., 25 : 145-147.
, 1940. The distribution of intermedin: a new biological method of assay and

results of tests under normal and experimental conditions. Anat. Rec.,

76: 157-172.
KRAUSS, F., 1906. Der Zusammenhang zwischen Epidermis und Cutis bei Sauriern

und Krokodilen. Arch. mikr. Anat., 67: 319-363.
LANGE, B., 1931. Integument der Sauropsiden. Bolk, Goppert, Kallius und Lu-

bosch: Handb. vergl. Anat. d. Wirbeltiere, 1: 375-448.
LEYDIG, F., 1873. Ueber die ausseren Bedeckungen der Reptilien und Amphibien.

Arch. mikr. Anat., 9: 753-794.

NOBLE, G. K., AND H. T. BRADLEY, 1933. The relation of the thyroid and the hy-
pophysis to the molting process in the lizard, Hemidactylus brookii. Biol.

Bull, 64 : 289-298.
PARKER, G. H., 1938. The colour changes in lizards, particularly in Phrynosoma.

Jour. Exper. Biol., 15: 48-73.
SCHAEFER, W. H., 1933. Hypophysectomy and thyroidectomy of snakes. Proc.

Soc. Exper. Biol. Med., 30: 1363-1365.
SCHMIDT, W. J., 1917. Die Chromatophoren der Reptilienhaut. Arch. mikr. Anat.,

90 : 98-259.
STRONG, R. M., 1902. The development of color in the definitive feather. Bull.

Mus. Comp. Zool. Harvard, 40 : 147-184.
WERNER, FR., 1892. Untersuchungen iiber die Zeichnung der Wirbelthiere. Zool.

Jahrb., Abt. Syst., 6 : 155-229.

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDIZA-
TION BETWEEN TWO FORMS OF RANA
PIPIENS, SCHREBER1

K. R. PORTER

(From the Biology Department, Princeton University, and the Rockefeller

Institute for Medical Research)

INTRODUCTION

One of the better methods for examining nucleo-cytoplasmic rela-
tionships is to combine identical nuclei with different cytoplasms. This
can be achieved in a number of ways. For example, if the gametes of
two different species are brought together to form reciprocal diploid
hybrids, it is expected that at least in the early stages such hybrids will
have identical nuclei and different cytoplasms. This arises from the fact
that in most cases the maternal parent contributes practically all of the
cytoplasm. Differences which may appear in the development and
heredity of the reciprocals can therefore be related to differences in the
egg cytoplasms of the two parent forms. If such cytoplasmic differences
are observable and measurable, the possibilities are obvious.

The same end is achieved by combining the male nucleus of one spe-
cies, subspecies or race with the nucleus-free egg cytoplasm of the
same species and another species, subspecies or race. If the andro-
genetic or merogonic 2 homospermic haploid and heterospermic haploid
resulting from this procedure show dissimilarities, these must be related
to cytoplasmic differences.

It is scarcely necessary to point out that the results of such pro-
cedures seldom if ever satisfy the preconceived possibilities. The re-
sults from reciprocal diploid hybrids may be limited by an incompatibility
of combined nuclei or nuclei and cytoplasms ; or, where this is not the
case, by the absence of sufficient cytoplasmic difference to produce an
effect. And heterospermic haploids are usually less satisfactory. In
the best of circumstances haploid organisms develop poorly and can only
be produced in a limited group of materials. Apparently as a result of

1 Data obtained, in part, from experiments performed during tenure of Na-
tional Research Fellowship at Princeton University.

2 The term androgenetic refers to the development of the whole egg with only
the male nucleus functional ; merogonic refers to the similar development of an
egg-fragment (Wilson, 1925).

238

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS

a high degree of incompatibility between nucleus and cytoplasm, the
development of most androgenetic species hybrids is extremely abnormal
and ceases in the earliest stages.

Despite such results, the possibility remains that hybridizations with
heretofore untried material such as the North American Salientia may
reveal one or more compatible combinations with the desired qualities.
Experiments to test this possibility have been made and the following
pages report one such investigation. Two distinct but closely related
forms of the genus Rana have been combined reciprocally to form diploid
and androgenetic haploid hybrids. The results form an interesting addi-
tion to the existing data on nucleo-cytoplasmic relationships.

The author is very grateful to Prof. G. Fankhauser for helpful

suggestions and criticisms.

MATERIALS

The gametes for these hybridization experiments were derived from
two distinct forms of frog, one collected from the meadows of northern
Vermont, the other from the immediate vicinity of Philadelphia. Both
forms are commonly referred to as Rana pipiens and possibly represent
different races or subspecies of that species. More attention will be
given to their probable relationship in the discussion.

That the two forms are distinct is indicated by their general charac-
teristics (Figs. 1 and 2), and also by the results of these experiments.
These same features also indicate that the two are closely related.
Therefore, as a temporary assumption and to facilitate the description
of the experiments, the frogs are being considered as northern and
southern forms or races of the same species. As such they will be
referred to in the succeeding pages of this report.

A brief description, supplemented by Figs. 1 and 2, will indicate their
major differences and similarities.

The northern form (from northern Vermont) is generally larger
and, relative to its body size and weight, it has shorter jumping legs
than the southern form. The head is obtuse ; the vocal sacs on the male
are less apparent ; the dorso-lateral folds are broad ; the skin is thick ;
the palmation is full. Distinctive features of pigmentation include spots
that are larger and surrounded by a green or yellow border ; the cross-
bars on the tibia are generally complete ; the posterior border of the thigh
is marked by black spots on a continuous white background ; the tym-
panum does not show a central light spot with the same clarity as in the
southern form.

The soutlicrn form (from the vicinity of Philadelphia) is generally
smaller and, relative to its body size and weight, its legs are longer.
The head is more accumatc; the vocal sacs are thin-walled and usually

240

K. R. PORTKK

FIG. 1. Photographs of representatives of northern (left) and southern (right)
forms of R. pipicns used in these experiments.

., * -r

Fic. 2. Pliiitngraplis showing pigmentation <>f jumping legs, northern (left) and

southern (right).

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 241

apparent ; the dorso-lateral folds are narrow ; the skin is thin ; the
palmation is shallow. Distinctive features of pigmentation include spots
that are smaller and, under the same laboratory conditions as the north-
ern form, not surrounded by clear borders; the cross-bar markings on
the tibia are generally interrupted along the dorso-lateral surface ; the
posterior surface of the thigh is marked by white spots on a continuous
black background ; and the tympanum generally shows a central light
spot.

METHODS

The eggs were obtained in every case from frogs which had been
induced to ovulate by frog pituitary injections. As much care as pos-
sible was taken to avoid removing the eggs in the immature or over-ripe
condition. The sperm for insemination were obtained by macerating
the testes in 10 per cent Ringer's solution and every precaution was
taken against contamination of one suspension with sperm from another.

The eggs of each frog were inseminated in two batches, the first
with sperm of the same form, the second with sperm of the other form.
Thus, in all, four batches of eggs were inseminated. An interval of 15
to 20 minutes was allowed to elapse between each insemination to provide
time for removing the egg pronucleus from a number of eggs of each
batch. By this procedure 8 different types of embryos were produced.
These are listed below with the designation used for each in the balance
of this report.

Homospermic diploids of the northern form n

Homospermic haploids of the northern form n/2

Heterospermic (hybrid) diploids from eggs of northern form and sperm

of southern form ns

Heterospermic (hybrid) haploids from cytoplasm of northern form and

nucleus of southern form (n^s/2

Homospermic diploids of the southern form .y

Homospermic haploids of the southern form s/2

Heterospermic (hybrid) diploids from eggs of southern form and sperm

of northern form sn

Heterospermic (hybrid) haploids from cytoplasm of southern form and

nucleus of northern form ( s) n/2

The egg pronucleus was removed with a fine glass needle as described
in a previous report (Porter, 1939). Adequate numbers of pure and
hybrid haploids were thus easily prepared (Table I).

All embryos were kept under identical conditions of temperature
(19.4° C.) and space. In fixation of representative forms for a per-
manent record, a mercuric chloride, acetic acid, and formaldehyde mix-
ture was generally used. The same sequence and time intervals were
observed in fixation as had been observed in fertilization. Thus it was

242

K. R. PORTER

assured that all animals fixed at the end of a period of time were of
the same age.

RESULTS

The description which follows is based upon observations made in
the experiments listed in Table I. The possibility that the same results
could occur by coincidence in all four series of crosses is slight if not
negligible. The analysis is confined to such characteristics as were ap-

TABLE I3

Number of
homo-

Number of
hetero-

Exp.

Date

spermic

spermic
/u K*-;/-I\

Treatment

haploids
produced

(nyDriu)
haploids
produced

Jan. 9, 1939

23 n/2

41 (n)s/2

Preliminary comparison of living animals

21 si 2

35 (s)n/2

made throughout development. Repre-

sentative embryos fixed at end of 3, 5, 7,

9, 10 and 11 days.

Jan. 17, 1939

29 n/2

30 (n)s/2

Living animals compared throughout de-

37 si 2

44 (s)n!2

velopment. Representative forms fixed

at end of 2, 3,4, 5, 6, 7 and 8 days. Dip-

loid hybrids and controls carried through

metamorphosis for examination of in-

heritance.

3a.

Feb. 15, 1939

37 n/2

48 (n)s/2

Living animals compared throughout de-

24 s/2

43 (s)n!2

velopment. Special attention given to

gastrulation and neural tube formation.

Representative forms fixed at end of 36,

43, 48, 51, 53, 55, 57, 59, and 61 hours and

at 3 and 4 days.

3b.

Feb. 15, 1939

21 n/2

36 (n)s/2

Living animals compared. Special atten-

45 s/2

41 (s)nf2

tion given to study of older stages. Ma-

terial fixed at end of 32, 54, and 60 hours

and 3, 5, 6, 8, 9, and 10 days.

3 The same females were used as a source of eggs for experiments 3o and 3b.
Otherwise, different parents were used in each cross.

parent from external examination and only those characteristics which
were uniformly shown by the animals in all four groups are stressed in
the succeeding paragraphs.

For greater clarity the description of the 3-day-old embryos is pre-
sented first. With the differences of these in mind the descriptions of
the younger and older stages have greater meaning.

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 243

Three-day-old Embryos

The following account is illustrated by the outline drawings in Fig. 3
and to them reference is constantly made.

The homospermic (control) diploids of the two races develop at
approximately the same rate at 19.4° C. and, stage for stage, are com-
parable at the end of 72 hours. The differences, though real, are very
slight and were clearly recognized only after repeated examination of
material available. The northern diploids compared with the southern
diploids show larger gill plates, a larger sense plate and larger mucous
glands. Relative to body size the head of n is the larger. The neural
tube is broader and stands up more distinctly in n. The tail-bud in n is
smaller and directed more dorsally than in s, thus creating a deeper
depression in the back of n. In relation to head size, the abdomen of -s
is larger than that of n. To these differences it can be added that the
head flexure dorsal to the posterior margin of the gill plate is more
pronounced in .? than in «.

The homospermic (control} haploids of the two races, as is normal
for haploids, are retarded in their development. Compared with each
other they show in an exaggerated form the same differences that were
given for the diploid controls.

The heterospermic (hybrid} diploids show approximately the same
rate of development as the homospermic diploids and as each other.
They differ in body proportions and show in accentuation the differences
which are difficult to see between the pure diploids of the two forms. A
greater proportion of ns consists of head structures than in the reciprocal
hybrid. Conversely, a greater proportion of sn consists of abdomen and
tail-bud. The mucous glands and sense plate are larger in ns and,
posterior to the medulla, ns shows a smaller neural tube which terminates
in a smaller and more dorsally directed tail-bud.4

The heterospermic (hybrid) haploids show in most exaggerated form
the differences which have been referred to as existing between control
diploids and haploids and more distinctly between the hybrid diploids.
It is readily apparent that oral suckers, gill plates, and sense plate are
greatly enlarged in (n)s/2. Relative to head size, the abdomen and tail-
bud of (s)n/2 are much larger than the same structures of (n)s/2. It
can be further noted that the head of (s}n/2 is flexed more ventrally
than (n}s/2 and the back of the latter is convex while in the former it

4 If these experiments had been confined to the production and study of diploid
hybrids, it is doubtful if the differences would have been considered great enough to
warrant any conclusions. Supported by the evidence from androgenetic hybrids,
however, the significance of the differences is unquestionable.

244

K. R. PORTER

i (j y i i to

N slloS i

O. O/5 p '

Q^ C/} C/) QJ *-"

^N Jk.\ -4-> ' ~ . . *O "S *J !

C^N tl £ p'S ^ J2 co

**•* — £ >-• .2 o a"O ;

a) -

,£ £ ""^ rt §

fe ^! C-= «

C 2 ° o

^oIN !^

^"O t
Cv. 3 !

S)5

.S a,

i *Q qj c

rs. O* -*-• c O
• -T 3 C u

M-l *O O *-* *4H ]

oo«.P«

« o>

- «•> c

T3
rt

P^O « g '"^-O

r^--^ u- T3 , c!

i j3 .rr Q a; aj to

•o o. > P
o

P O <L>

P P C

PCs

u to b£"

~ «J

o u ,

>- o
o ^^-c u g

n°:§^{3 2

S £-2>p-?

S Irt 2*^^.0 *o

^ !

vr e
c

<u ^
O S

•-*s

S rt

rt
o

gci

p i

rt a

S 4

^ I-

S "

3 O

O^ c"o

P r- V- P (Nl

S S *j IS-

;— _ r-«

e P e

&it

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 245

is concave. These differences are the expression of the decidedly dis-
similar embryology of the two reciprocal heterospermic haploids and not
a difference in age or stage.

Summary. — In general the combinations which include cytoplasm of
the northern form are characterized by larger head primordia and smaller
posterior axial structures than are observed in those with southern cyto-
plasm. Such differences, only slightly apparent in the pure diploid
controls, become progressively more accentuated in the homospermic
haploids, the heterospermic diploids and in the heterospermic haploids.

It is of interest to observe now the earlier and later expressions of
these general differences as shown by an examination of the earlier and
later stages in the ontogeny of the various combinations.

Neural Tube Formation

The description under this heading is derived from a comparative
study of living material and of representative embryos of the eight
different types fixed at intervals of two hours from 51 to 61 hours after
insemination. Illustration is provided by outline figures 4, 5, and 6
which are respectively representative of developmental stages reached at
the end of 55, 59 and 61 hours.

The homospermic diploids, during this period, are very similar both
in character and rate of development. As the neural plate is outlined,
it becomes apparent that its anterior portion plus the sense plate are
larger in n than in s. These differences increase in clarity as the neural
folds are elevated and gill plates appear (Figs. 5 and 6). At this latter
stage, ^ flattens dorsally and shows a greater elongation of that portion
of the neural groove posterior to the gill plates. At the same time the
neural plate and folds are more distinctly elevated in s. Although the
neural plate and folds may be outlined in j slightly in advance of n, the
closure of the folds is more rapid in the latter. During neurulation the
blastopore of .$• is bounded laterally by distinctly thickened lips.

The homospermic haploids show in exaggerated form the slight dif-
ferences existing between the diploid controls. In equivalent stages (55
hours) the neural folds of n/2 are thicker, the sense plate and other
primordial head structures are larger than in s/2, whereas the latter
shows a greater elongation of the neural plate, especially that portion of
it determined to be spinal cord. In s/2 the neural folds show a greater
elevation, the dorsal surface straightens or flattens out, and pronounced
lateral lips bound the blastopore. Besides stage-for-stage structural dif-
ferences, the differences in rates of separate morphogenetic processes
are also accentuated. For example, it is noted that s/2 completes gas-

246

K. R. PORTER

"O I-

!-. D
O OJ

-o
a

- rt

E I*

53 3

: <u

C y-r

O Z

-a

cu r\]

OJ
N

G
O rt

O) vj_

> <z

§1

s js
•e §

*0 S
u,

I"
• rt

"g'S)

a; O

c/) i-«

^ bo

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 247

g
<D

2
"o

"
o

en
u
O

248 K. R. PORTER

trulation considerably ahead of n/2 and only slightly after s, whereas
in n/2 the neural folds appear to close slightly before they do in s/2.

The heterospermic diploids show clear-cut differences. Compared
with each other in the early stages of neurulation, it is apparent that
the ns forms have a shorter neural plate which is abnormally broad at
the anterior end. The reciprocal sn, on the other hand, has a long and
narrow neural plate. As the two differ from each other so do they differ
from their maternal control diploids though to a lesser degree. Other
features of dissimilarity include neural folds which are larger in ns than
in sn and which are more distinctly elevated in us than in the control
diploid, n. In this latter respect they approximate the condition noted
above as apparent in the paternal diploid control. As neural tube forma-
tion continues, the greater size of the head primordia and the shorter
neural plate and groove are maintained in ns. Though the neural plate
and folds are outlined almost simultaneously in these reciprocal hybrids,
the folds come together in ns slightly before they do in sn. The lateral
borders of the blastopore are swollen in sn to form lips as in ^ and s/2.

The heterospermic ha pi aids show very striking differences during the
development of the neural tube. At the end of 51 hours (not illustrated)
(s)n/2 has a dorsal flattened surface, abnormally straight from anterior
to posterior ends. The neural plate is clearly outlined and is very
narrow. Gastrulation has been completed, and there are extremely pro-
nounced lips on both sides of the blastopore. Contrasting with this,
the reciprocal (11) s/2 is considerably retarded. The yolk plug is still
apparent and the limits of the neural plate are not visible. The (n)s/2
embryos are flattened dorso-ventrally and present a large, swollen ap-
pearance. By the end of 55 hours the neural plate of (n)s/2 has been
outlined. It is as broad as it is long and that portion designated to be-
come neural tube is extremely short. The yolk plug persists. At this
same time in (s)n/2 the neural plate has lengthened and the neural folds
have approximated to some extent. At 59 hours (n~)s/2 continues to
show a short, broad neural plate, bounded by prominently elevated neural
folds. This latter feature is a characteristic of ^ embryos and its ap-
pearance in these (n}s/2 embryos represents the appearance of a specific
paternal character. In the reciprocal it is not shown. It is a feature
which will be easier of description and analysis when sectioned material
is available. By 61 hours the neural folds of (n)s/2 have started to
approach and subsequent observations have shown that once started this
process proceeds more rapidly here than in (s)n/2. At this time and
later there is little elongation of the neural plate in (n)s/2 and the yolk
plug still persists in some cases. The sense plate and gill plates are
abnormally large. The structure of the reciprocal hybrid (s)n/2 at 61

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 249

hours is characterized by neural folds about ready to close, an elongate
neural tube, and extremely small head primordia which foreshadow the
diminutive head size of later stages.

Summary. — In summarizing, a few generalizations can be made.
Those combinations which include cytoplasm of the northern race, in-
cluding the diploid controls, are characterized by : (a) neural plates
which when outlined tend to be shorter, and broader anteriorly, and (&)
head primordia which are larger. The reciprocal combinations with
cytoplasm of the southern race are, on the other hand, characterized by :
(a) longer and narrower neural plates, (b) smaller head primordia, and
(c) pronounced lateral lips on the blastopore. These differences become
increasingly apparent as one compares respectively the diploid controls,
the haploid controls, the reciprocal hybrid diploids, and the reciprocal
androgenetic hybrids.

At one stage in the development of the neural folds it is apparent
that they are more sharply delimited and distinctly elevated in the diploid
of the southern form. This characteristic is repeated in the hybrid
diploids and in the androgenetic hybrids containing the southern nucleus.
It seems to represent, therefore, an inheritable embryonic characteristic
capable of expressing itself in the foreign cytoplasm of the northern
race. More careful analysis of this phenomenon is needed.

There are also to be noted slight differences in the times of occur-
rence and rates of the same morphogenetic processes. Relative to blas-
topore closure the neural plate is outlined earlier in those combinations
with northern cytoplasm. Relative to time after fertilization, however,
this may be later. Once clearly outlined the neural folds of the com-
binations with the northern cytoplasm seem to close more rapidly.

Gastrulation

This phase of the embryology of these various combinations was
studied from living material and from representative forms fixed at the
end of 36, 43, and 48 hours. A few differences between the gastrulae
of those forms with northern cytoplasm and those with southern cyto-
plasm occur consistently (excepting the diploid controls where they are
not sufficiently pronounced to be clearly evident) and become progres-
sively more pronounced in haploid controls, heterospermic diploids, and
heterospermic haploids. Those combinations with the cytoplasm of the
southern race show a larger gastrular angle, a smaller completed blasto-
pore, epiboly largely from the dorsal and lateral borders of the blastopore,
and toward the end of gastrulation, an increasing thickening of the lateral
blastopore lips. Those combinations with northern cytoplasm show a
smaller gastrular angle, a larger blastopore, epiboly from all sides of

LIBRARY

250 K. R. PORTER

the blastopore, and thin blastopore lips. Gastrulation appears to begin
earlier in s, s/2, (n}s/2, and simultaneously in sn and ns. Observations
recorded on this feature and on the rate of gastrulation are not suffi-
ciently extensive to be conclusive.

It would seem that the greater gastrular angle and the greater epiboly
of the dorsal lip in those haploid and hybrid embryos with the southern
cytoplasm are the early abnormalities related to the longer neural plate
of later stages. It also appears that the thickened lateral blastopore
lips of these same forms are the early expression of the larger tail-bud
and somites of later stages (Bijtel, 1931). The opposites of these same
features in those forms with the northern cytoplasm are probably related
to the shorter neural plate and smaller tail-buds of their later stages.

Older Stages (4-10 Days)

The studies reported in this paper have been largely devoted to the
younger stages hence only the most general features of the older stages
will be described under this heading. Reference should be made to
Fig. 7.

The homospermic diploids of 4 and 5 days continue to show the
slight differences which existed between the 3-day-old embryos. In the
older stages, however, these differences become increasingly subtle.
Relative to body proportions, the head of n remains larger while the tail
of ^ is more elongate and larger in relation to the rest of the embryo.
The dorsal concavity of n persists in greater prominence than in s.

The homospermic haploids differ in the older stages, as they had in
the earlier stages, in relative size of body parts. The differences are
similar to but more distinct than those occurring between the diploid
controls.

The heterospermic diploids demonstrate more clearly the perpetua-
tion of early differences. The combination, us, persists in showing at
various ages a larger head with larger mucous glands and a smaller
dorsally directed tail. The converse of these features are shown by the
reciprocal. Such differences are retained into the later stages of devel-
opment, especially the relative head and tail size. Clear-cut appearance
of paternal characteristics is recognized first in stages showing chro-
matophore patterns.

The heterospermic haploids, as in the younger stages, show the most
striking differences. It is recognized, however, that these differences are
less pronounced in the older stages suggesting some regulation. It is
easily noted that the head of (n)s/2 and its component structures re-
mains larger and the tail remains smaller and directed dorsally. The
androgenetic hybrid, (s)n/2, on the other hand, is characterized by a

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS

251

•*_»
IH

.£3

-5
j>>

cij

•41

MH
O

tn
U

'> S

^ .2

rt *^j
_h eo

C1J

•*->
OS

252 K. R. PORTER

small head and a ventrally directed tail which is referred to as larger
because of the broad heavy somite mass at its base. These features
were foreshadowed in the early embryonic development.

Summary. — Those combinations with cytoplasm of the southern race
tend to have smaller head structures and, relative to body size, larger
tails than the reciprocals with northern cytoplasm. Such features are
doubtless the expression of earlier embryonic differences in the size of
head primordia and tail-buds.

Survival of Various Combinations

Since representative embryos were sacrificed for fixation at various
intervals, no definite data can be given to demonstrate survival value.
Nevertheless, the observations made permit the following statements.

The heterospermic hybrids, in the majority of cases, develop up to
and through metamorphosis. Beyond that stage no data are available.

The homospermic haploids of the two races demonstrate approxi-
mately the same viability. They continue their development, on the
average, for from 8 to 12 days up to approximately stage 24 (Shumway,
1940).

The heterospermic haploids develop through the early stages showing
only a small percentage of deaths. About one-fourth to one-third fail
to hatch and of those which hatch the majority live for from 7 to 11
days or up to stages 22 and 23 (Shumway, 1940). They are slightly
less viable than the haploid controls. In a small percentage of cases a
gill circulation is established and in scattered cases growth continues
sufficiently long to show the first guanophores. No differences in via-
bility were recorded as existing between the two reciprocal combinations.

Identification of Haploids

The haploids were identified as such solely on the basis of the type
of development. In an earlier study (Porter, 1939) it was shown that
embryos which arise from operated eggs can be expected to develop as
haploids in 90 per cent of the cases. Furthermore, such haploids were
found to show certain definite characteristics when compared with their
diploid controls. Hence, in these experiments, it has not been consid-
ered necessary to make a complete cytological examination of every
embryo which developed from an operated egg, especially since group
characteristics rather than individual characteristics have been considered.
The isolated cases of diploidy which did appear among the embryos from
operated eggs were readily identified by their development, cell size, etc.

In order to establish the chromosome count of the southern form,

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 253

the tail-tips of several haploids were examined. The examinations made
indicated 13 to be the haploid count. This is the same as for the
northern form determined in an earlier study.

Other Observations

In the case of two of the above-described experiments, surplus diploid
embryos both pure and hybrid were kept for examination as older tad-
poles and as metamorphosed frogs. The pure diploids showed differ-
ences characteristic of the northern and southern forms ; the hybrid
diploids showed blended inheritance with indications of stronger paternal
influence in certain features of pigmentation. Thus despite the blending
the reciprocal hybrids were distinguishable. These observations indi-
cate that at least the differences in pigmentation between the two parent
forms are related to differences in nuclear factors.

Two further observations, which, because of the small amount of
evidence supporting them, must be considered as very preliminary, are
briefly described. They are presented because of their interest as pos-
sible leads for experiments aimed at determining the nature of the
factors responsible for the peculiar development of the hybrid embryos
described above.

Since the frogs used were obtained from widely separated northern
and southern points, it was considered of interest to examine the effect
of high temperatures. Embryos representative of the 8 different com-
binations treated above we're placed in a warming oven at 28°-29° C.
Those combinations with cytoplasm of the southern form were not ap-
parently damaged and developed in the typical manner, whereas those
combinations with northern cytoplasm were markedly affected and only
a very small percentage of the original number developed through 6
days. Controls kept at 19.4° C. developed normally.

Cytological examination of a few pure diploid 3-day-old embryos
revealed some interesting differences in nuclear size and size of yolk
granules. Measurements of nuclei of identical tissues of the two forms
showed those of the northern to be the smaller. Measurements of the
yolk granules revealed those of the northern to be much the larger.

DISCUSSION

The discussion which follows will be confined to a consideration
of the probable relationship of the animals used and to the more general
aspects of the cytoplasmic and nuclear influences demonstrated. A de-
tailed and inclusive treatment must await the accumulation of data from
a more thorough study of these and similar hybrids. In a sense, then,
this constitutes a preliminary report.

254 K. R. PORTER

When the experiments were first undertaken the parents were thought
to represent two distinct species. This conclusion was based on differ-
ences which the animals showed and also on the authority of amphibian
taxonomists (Kauffeld, 1937; Stejneger and Barbour, 1939). An ex-
amination of the literature soon revealed, however, that considerable
confusion exists in the classification of the leopard frogs or frogs of
this type resident in the eastern states and possibly over a wider area.
Differences between those forms resident in northeastern and those in
the southeastern states have long been recognized, but it appears that
sufficient material from a variety of localities has never been examined
to make a conclusive analysis of the species. A brief reference to the
writings of a few authorities on Salientia classification will serve to
illustrate this confusion.

It should be recalled that the southern forms used in these experi-
ments were collected in the vicinity of Philadelphia and the northern
forms in northern Vermont. From its place of collection, the southern
form doubtless coincides in appearance with that type early described by
Schreber (1782) as Rana pipiens (Kauffeld 1936 and 1937). Later,
Cope (1889), from examination of forms collected in a variety of
localities, chose to describe the leopard frogs under three subspecies.
The southernmost type he called Rana virescens sphenoccphala; the
type from the Atlantic coast Rana virescens virescens (probably same
as Schreber's R. pipiens and the Philadelphia type of this study) ; and
the type of northern distribution he called Rana virescens brachycephala
(his description of which coincides perfectly with the northern form
used in these experiments). More recent authors (Wright, 1933;
Dickerson, 1906) have pictured and described the northern form as the
typical R. pipiens and both it and the southern form have been consid-
ered as such by teachers and investigators alike. Most recently Cope's
nomenclature has in part been revived, only instead of using a subspecies
classification, the three types have been placed in separate species. Thus
the most southern form is called R. sphenocephala, the Philadelphia form
falls within the range of R. pipiens and the northern becomes R. brachy-
cephala (Kauffeld, loc. cit., and Stejneger and Barbour, loc. cit.}.5 It
was on the basis of this latter classification that the frogs were originally
considered to represent two species, R. pipiens and R. brachycephala.
It is clear, however, that this classification is uncertain and consideration
of some further points increases this uncertainty.

In the first place, it would seem that the two forms hybridize too

5 In footnote, Stejneger and Barbour (1939) indicate that the whole spheno-
cephala-pipiens-brachycephala complex needs further examination and possible
revision.

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 255

successfully to be representatives of two distinct species. It is true
that a few distinct species of the Salientia have been successfully hy-
bridized (Born, 1883; Pfltiger and Smith, 1883; Heron-Royer, 1891;
Montalenti, 1933 ; Durken, 1938 and Moore, 1940) so that the successful
crossing of these two forms, even if they represent distinct species, is
not without parallel. What is unique is the result of androgenetic hybrid-
ization, for no case involving the Salientia has been reported in which
the development of an androgenetic or merogonic species hybrid con-
tinued to the advanced stages obtained in these experiments.6 In other
words, the compatibility of the two forms is greater than would be
expected of two distinct species.

In the second place, it can be said that the characteristics of the two
forms do not differ sufficiently to place them in separate species. Aside
from body proportions, which is dealt with below, the major difference
is one of pigmentation. This difference, it can be noted, does not
involve the pattern but chiefly the size of the markings and intensity of
the coloration. These are features which in other animals may vary
considerably among races.

Finally, recalling that the two forms were collected from different
northerly and southerly climates, and considering the points about to be
discussed, the differences in body proportions likewise do not support a
species relationship. Taxonomists have long been acquainted with cer-
tain generalizations known as the Bergmann and Allen rules pertaining
to differences in size and body proportions which can be recognized
between the northern and southern races of warm-blooded species. The
former of these states that northern races are larger; the latter, that
the southern races have relatively longer body projections. Within re-
cent years an increasing volume of research examining racial and sub-
species differences has shown that characteristics other than body size
and proportions may likewise vary in an orderly and predictable manner
with a variety of environmental gradients. Inclusive surveys of these
phenomena are to be found in the recent writings of Goldschmidt ( 1940) ,
Rensch (1936) and others. But among the species of animals examined
for chains of racial differences or " clines " (Huxley, 1938), it appears
that species of Amphibia have been regrettably absent. Schmidt (1938)
reviewed some measurements of species of Salientia and noted that

6 Baltzer (1920 and 1933) reports that from the combination of Triton taeni-
atus cytoplasm and Triton palinatns nucleus heterospermic haploids develop to stages
showing good eye formation, pigment, small branching gills, and pulsating heart.
Though this represents advanced development as compared with the usual result
with different Salientia species, the stage reached does not seem to be the equivalent
of that reached by the best of the heterospermic haploids obtained with these two
forms of Rana pipiens.

256 K. R. PORTER

relative to body size the leg length was greater for those representatives
of a species which were collected from the more southern localities.
The small number of animals examined and the preserved condition of
these did not, however, permit any definite conclusions. Measurements
of unselected groups of the two forms used in these crosses show the
same tendency of the northern form to have a heavier and larger body
structure relative to leg length.

It is possible that a thorough examination of the literature would
reveal additional references to racial differences between frogs. For
example, such differences are briefly mentioned in a paper by Pfliiger
and Smith (1883). Comparing the English race of R. fusca with the
Konigsburg race of the same species, they write :

' Der englische braune Grasfrosch ist etwas kleiner und schlanker als
der deutsche, weniger stumpfschnauzig und von zarterer Haut."

The similarity between these differences and those noted between the
Vermont and southeastern Pennsylvania forms of R. pipiens is obvious.
This similarity takes on added interest when it is noted that roughly
the same climatic differences (as indicated by mean annual temperatures)
exist between East Prussia (44° F.) and England (50° F.) as between
Vermont (43° F.) and southeastern Pennsylvania (52° F.).

In view of these observations and the fact that racial variations
accompanying climatic gradients have been found in a great many species
of both the animal and plant kingdoms, it seems probable that species
of frog when thoroughly examined will likewise show various clines with
regard to temperature and other environmental factors. In the mean-
time it can only be maintained that the two forms used in these crosses
probably represent two races of the same species.

If such is the case, the results of these experiments are of interest
in demonstrating that racial differences involving body proportions can
be recognized in early embryonic stages, and that at least some of the
factors responsible for these differences exist in the cytoplasmic organi-
zation of the egg (see below). This observation and others which will
probably be made from a more extensive examination of these and simi-
lar crosses may prove of interest to students who concern themselves
with factors involved in species formation.

Experiments examining the relative roles of the nucleus and cyto-
plasm in heredity have generally shown the nucleus to be the sole
bearer of factors controlling the appearance of specific adult and juvenile
characteristics. Some of these experiments have combined the nucleus
of one species with the cytoplasm of another to form merogonic hybrids,
attempting thus to demonstrate the presence of hereditary units in the

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS

cytoplasm. Among these, the studies of the Hertwigs, Boveri, Baltzer,
Haclorn, and Horstadius are well known and frequently reviewed. With
the possible exception of Hadorn's (1936) results, the demonstration of
cytoplasmic inheritance has not been conclusive. The development of
the merogonic hybrids generally ceases very early and even where it
continues to a stage showing distinct species characteristics, as in certain
sea-urchin merogons, the intermediate condition of the characteristic
can be considered as an abnormality resulting from a degree of incom-
patibility between the nucleus and cytoplasm (Horstadius, 1936). The
early cessation of development which characterizes amphibian merogonic
hybrids is probably also the result of a severe incompatibility.

It would appear that by using more closely related forms than those
belonging to different species this problem of incompatibility could be
overcome. To some extent this is probably true, but in using members
of different races or subspecies, it is necessary to sacrifice the clear-cut
distinctions which usually exist between the embryonic stages of different
species and which are not to be expected between different races. Hence,
the problem is fraught with difficulties and it is doubtful whether mate-
rial such as used in these experiments, though it should be thoroughly
examined, will supply any evidence in support of cytoplasmic-borne units
of heredity even if present.

As distinct from heredity, cytoplasmic influence on development has
been and can be demonstrated. This influence has been considered as
the effect of plasmatic organization and composition upon the expression
of nuclear factors. To this category of cytoplasmic activity the results
of these crosses probably belong. Experimental embryologists have long
recognized a high degree of cytoplasmic differentiation in a variety of
eggs and the maintenance of such differentiation undisturbed is known
to be essential in many cases for normal embryonic development. The
cytoplasm of the egg by its organization, therefore, exerts an influence
on the appearance of the adult in so far as this appearance is determined
by the characteristics of the early developmental stages.7 Needless to
say, the nature of this early cytoplasmic influence is not understood but
every new demonstration of its presence offers new possibilities for its
examination.

The consideration of the results of these experiments is facilitated
if the development of the two control diploids is visualized as paralleling
on opposite sides an average or mean type (Fig. 8). If the factors
responsible for this slight departure from the mean are nuclear and the

7 In respect to even this cytoplasmic influence, it is to he remembered that
considerable differentiation of the egg takes place in the presence of the maternal

nucleus.

258

K. R. PORTER

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS

cytoplasms are perfectly neutral to nuclear control, then the diploid
reciprocal hybrids would be expected to be identical and would in their
development occupy a position coinciding with the hypothetical mean.
Under the same conditions of nuclear control, the heterospermic haploid
with the southern cytoplasm and northern nucleus would be expected to
show the same development as the homospermic haploid of the northern
form. Neither of these results is obtained. Instead, it is noted in the
case of the reciprocal diploid hybrids that their development places them
on opposite sides of the mean and at points more distant from the mean
than their diploid controls. And in the case of the heterospermic
haploids, the hybrid with the northern nucleus is not only further from
the mean than the homospermic haploid of the northern form, but it is
on the opposite side. Since the diploid hybrids can be considered as
having identical nuclei and differing only in their cytoplasms, and since
the same difference holds between the homo- and heterospermic haploids
with nuclei of the same form, it follows that cytoplasmic influence is
responsible for the dissimilarities existing between them.8 Therefore,
the eggs of the northern and southern forms differ in some property or
properties of their cytoplasms.

Are the nuclei identical or do they also differ? If the nuclei are
considered as being identical and responding solely to cytoplasmic in-
fluence, then the development of the reciprocal diploid hybrids should
parallel the mean at the same distance as their respective diploid controls.
Or, under the same assumed conditions of identical nuclei, the hetero-
spermic haploid with the southern cytoplasm should be identical in ap-
pearance with the homospermic southern haploid. Again, the results
indicate that the assumed condition of identical nuclei cannot be valid.
On the other hand, the intermediate position of the diploid control
between the mean and the hybrid diploid with the same cytoplasm
indicates that the nucleus of each race has compensating factors for the
cytoplasm of that race. The same conclusion is also supported by the
intermediate position of the homospermic haploid relative to the mean
and the heterospermic haploid with the same cytoplasm. Evidently then,
the nuclei of the two forms also differ and do so in such a way as to
compensate in development for cytoplasmic differences.

Cytoplasmic and nuclear differences seemingly demonstrated, it is
of interest to determine which is responsible for the slight dissimilarities
between the control diploids, and the more distinct dissimilarities be-
tween the homospermic (control) haploids of the two races. It has
been shown that each diploid control in its morphogenesis is on the same

8 A heterozygous genome in the parent forms could not account for these
differences.

260 K. R. PORTER

side of the mean as the hybrid diploid with the same cytoplasm, though
not at the same distance. The homospermic haploids, in their develop-
ment, parallel the mean at a greater distance than the control diploids,
suggesting a lesser degree of compensation by the haploid nucleus. The
homospermic haploid in its morphogenesis shows the same tendencies,
though to a lesser degree, as the heterospermic haploid with the same
cytoplasm. These facts suggest that the cytoplasmic differences are
responsible for the slight dissimilarities between the diploid controls and
homospermic haploid controls of the two forms. Further study may
demonstrate whether or not these cytoplasmic differences are also related
to the dissimilarities of the two adult parent forms.

What is the nature of these nuclear and cytoplasmic differences?
There is not, of course, sufficient information available to answer this
question. The presence of some degree of cytoplasmic organization in
the amphibian egg has been shown to exist as early as 20 minutes after
insemination (Fankhauser, 1930) and before first cleavage (Brachet,
1906), but the nature of this organization has not been demonstrated.
Though the differences which are being examined cannot be described
in precise terms, one feature of their relative nature does become ap-
parent. It is clear from the results that some property or properties
of the cytoplasm of the northern form tend to make the embryos with
the cytoplasm of that form display certain features of development
which, relative to the mean type representing normal development, are
the exact opposite of those found in the embryos with the cytoplasm
of the southern form (Table II). This infers that the differences in

TABLE II

Northern cytoplasm Southern cytoplasm

1. Small gastrular angle 1. Large gastrular angle

2. Large completed blastopore 2. Small completed blastopore

3. Neural plate abnormally broad at an- 3. Neural plate abnormally narrow at

terior end anterior end

4. Neural plate abnormally short 4. Neural plate abnormally long

5. Small tail-bud 5. Large tail-bud

6. Large head primordia 6. Small head primordia

7. Small abdomen relative to head size 7. Large abdomen relative to head size

organization or composition, whether they be quantitative or qualitative,
are of opposite natures as measured in terms of what they tend to
produce in development. It has been noted further that the nuclei of
the two forms have properties which tend to compensate for the cyto-
plasmic differences. Therefore the nuclei may also be considered to have
properties of opposite nature. If this reasoning is correct, it seems that
the nucleus of one form should supplement or enhance the cytoplasmic

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 261

influence of the other form. This means that the development of the
reciprocal heterospermic haploids should be sufficiently different to sug-
gest the activity of something more than the cytoplasm. While there
is no unit of measurement by which the degree of difference can be
determined, it is clearly great (Fig. 8) and is probably contributed to
by a nuclear influence.

The differences in size of yolk granules and nuclei which preliminary
studies have shown to exist between the early embryonic stages of the
two forms constitute the only concrete dissimilarities between cytoplasm
and nuclei so far observed. What connection, if any, these may have
with the actual nuclear and cytoplasmic differences responsible for the
above results is not readily apparent.

It is of further interest to determine how these differences operate
to produce the results described above. This point is brought into this
discussion not because any definite answer can be provided but because
certain experimental treatments which could be expected to alter the
mode of operation of cytoplasmic and nuclear factors have produced
similar results. For example, if a temperature gradient is applied to
the developing frog egg, that portion at the warm end of the gradient
develops abnormally large structural units (Huxley, 1927; Dean, Shaw,
Tazelaar, 1928; Gilchrist, 1928. 1929, 1933). More specifically, if the
gradient is applied " adjuvantly " (Huxley, 1927) along the animal-
vegetal polar axis in blastula stages (i.e., with warm end of gradient at
animal pole) the tail-bud embryos from a blastula so treated have slightly
larger heads than the controls and those subjected to the reverse gradient
(Huxley and Dean, Shaw, Tazelaar, loc. cit.). It is further reported
by the same authors that an adjuvant gradient increases by several times
the normal difference in size existing between animal and vegetal cells
of the blastula stages. Gilchrist (1933) demonstrates that size differ-
ences of embryonic structures resulting from temperature gradient treat-
ments are not due solely to age differences but thinks rather that there is
an alteration in what he terms the " physiological pattern " of the egg.
In this same connection it can be noted that toxic agents applied to devel-
oping frog embryos can likewise produce a disproportion of parts most
noticeably influencing those regions having the highest metabolic activity
at the time of application (Bellamy, 1919).

With these results in mind, it is reasonable to suggest that the
differences between the cytoplasms of the eggs of these two geographic
forms or races are differences in factors which normally determine the
varying rates of metabolism and cell division in the various parts of the
developing blastula and possibly the induction processes in later stages.
Only one bit of experimental evidence bearing on the physiological

262 K. R. PORTER

properties of these eggs is available and this of a very preliminary sort—
the temperature tolerance is higher for the egg of the southern form.
This, it is logical to suppose, is related to the fact that the southern
embryos may be called upon to develop at higher temperatures than the
northern. From this, however, it is not possible to reason that other
physiological differences which may exist between the two eggs are like-
wise related to climatic influences.

It is realized that other subjects of interest could be discussed in
relation to the results of these experiments but it is felt that they may
be considered more successfully after more information has been accumu-
lated. For the present, it seems best to emphasize that the gametes of
two geographic forms probably of the same species differ slightly in their
cytoplasmic and nuclear properties and that by androgenetic haploid as
well as diploid hybridization the orderly and measurable effects of these
properties on early morphogenesis can be observed. The nature of these
differences, their mode of operation, the relation of the embryonic dif-
ferences they produce to the differences between the adults are among
the major problems which can be and should be examined later with the
same or similar materials and methods.

SUMMARY

1. Two distinct forms of frog, commonly referred to as Rana pipicns,
Schreber, are described, and evidence is presented to show that they
probably represent geographic races of that species, one from northern
Vermont, the other from southeastern Pennsylvania.

2. In the experiments described, the gametes of these two races have
been combined reciprocally to form diploid and androgenetic haploid
hybrids and the early development of these has been studied in detail.

3. The diploid hybrids developed through metamorphosis ; the andro-
genetic hybrids for 7 to 11 days, up to about stage 24 (Shumway, 1940).

4. A comparison of 3-day-old control and hybrid embryos reveals
that, in general, the combinations which include cytoplasm of the northern
form are characterized by larger head primordia and smaller posterior
axial structures than are observed in those with southern cytoplasm.
Such dissimilarities, only slightly apparent between the homospermic
diploid controls, become progressively more accentuated between the
homospermic haploids, the heterospermic (hybrid) diploids, and the
heterospermic (hybrid) haploids.

5. A study of gastrula, neurula, and older stages discloses the early
expressions and later fate of the dissimilarities shown by the 3-day-old
embryos.

DIPLOID AND ANDROGENETIC HAPLOID HYBRIDS 263

6. These results demonstrate :

(a) Cytoplasmic differences between the eggs of the two forms
which seem to have contrasting effects upon the same developmental
processes.

(b) Nuclear differences which, in homospermic diploid control de-
velopment, appear to compensate for the cytoplasmic differences.

(c ) An orderly cytoplasmic influence on early morphogenesis.

7. The possible nature and mode of action of these differences are
briefly discussed.

LITERATURE CITED

BALTZER, F.. 1920. Uber die experimentelle Erzeugung und die Entwicklung von

Triton-Bastarden ohne mutterliches Kernmaterial. Verhandl. Schweiz.

Naturforsch. Ges. Neiienburg, pp. 217-220.
— , 1933. Ueber die Entwicklung von Triton-Bastarden ohne Eikern. Verhandl.

dcutsch. Zool. Ges., 35: 119-126.
BELLAMY, A W., 1919. Differential susceptibility as a basis for modification and

control of early development in the frog. Biol. Bull., 37: 312-361.
BIJTEL, J. HUBERTHA, 1931. Uber die Entwicklung des Schwanzes bei Amphibien.

Roux' Arch., 125 : 448-486.
BORN, G., 1883. Beitrage zur Bastardirung zwischen den einheimischen Anure-

narten. P finger's Arch. ges. Physiol, 32: 453-518.
BRACKET, A., 1906. Recherches experimentales sur 1'oeuf non segmente de Rana

fusca. Roux' Arch., 22: 325-341.
COPE, E. D., 1889. The Batrachians of North America. Bull. U. S. Nat. Mus.,

No. 34.

DEAN, I. L., M. E. SHAW, AND M. A. TAZELAAR, 1928. The effect of a tempera-
ture gradient on the early development of the frog. Brit. Jour. Expcr.

Biol., 5 : 309-336.

DICKERSON, M. C., 1906. The Frog Book. Doubleday, Page and Co., N. Y.
DURKEN, BERNHARD, 1938. Uber die Keimdriisen und die Chromosomen der Art-

bastarde Rana arvalis Nils. uX. Rana fusca Ro's. Q. Zeitschr. f. indukt.

Abstammungs-und Vererbungsl., 74: 331-353.
EAST, E. M., 1934. The nucleus-plasma problem. Am. Nat., 68: 289-303; 402-

439.
FANKHAUSER, G., 1930. Die Entwicklungspotenzen diploidkerniger Halften des

ungefurchten Tritoneies. Roux' Arch., 122 : 671-735.
GILCHRIST, F. G., 1928. The effect of a horizontal temperature gradient on the

development of the egg of the urodele, Triturus torosus. Physiol. Zool.,

1 : 231-268.
— , 1929. The determination of the neural plate in Urodeles. Quart. Rev. Biol.,

4 : 544-561.
— , 1933. The time relations of determinations in early amphibian development.

Jour. Exper. Zool., 66 : 15-51.
GOLDSCHMIDT, RICHARD, 1940. The Material Basis of Evolution. Yale University

Press, New Haven.
HADORN, ERNST, 1936. Ubertragung von Artmerkmalen durch das entkernte

Eiplasma beim merogonischen Triton-Bastard, palmatus-Plasma X cris-

tatus-Kern. Verhandl. dcutsch. Zool. Ges., 38 : 97-104.
HERON-ROYER, 1891. Nouveaux faits d'hybridation observes chez les batraciens

anoures. Mem. Soc. Zool. France, 4 : 75-85.

264 K. R. PORTER

HORSTADIUS, SVEN, 1936. Studien iiber Meterosperme seeigelmerogone nebst be-
merkungen iiber einige keimblattchimaren. Mem. Musee Ro\. Hist. Natu-
relle Belgique, 2 Ser. Fasc. 3., pp. 801-880.

HUXLEY, JULIAN S., 1927. The modification of development by means of tem-
perature gradients. Roux' Arch., 112: 480-516.

— , 1938. Clines : an auxiliary taxonomic principle. Nature, 142 : 219-220.
— , 1938a. Clines: an auxiliary method in taxonomy. Bijdragcn tot de Dier-
kunde, Festnummer, Afl., 27 : 491-520.

KAUFFELD, CARL F., 1936. New York the type locality of Rana pipiens Schreber.

Hcrpctologica, 1 : 11.

— , 1937. The status of the leopard frogs, Rana brachycephala and Rana pipiens.
Herpctologica, 1 : 84-87.

MONTALENTI, GIUSEPPE, 1933. L'ontogenesi degli ibridi fra Bufo vulgaris e Bufo

viridis. Physiol. Zoo/., 6: 329-395.

— , 1938. L'ibridazione interspecifica degli Anfibi anuri. Archivio Zoologico
Italiano, 25, Suppl., Altualita Zoo/., 4: 157-213.

MOORE, JOHN A., 1940. In Press.

PFLUGER, E., WM. J. SMITH, 1883. Untersuchungen iiber Bastardirung der anuren
Batrachier und die Principien der Zeugung. I. Experimente iiber Bastardi-
rung der anuren Batrachier. P finger's Arch. ges. Phys., 32 : 519-541.

PORTER, K. R., 1939. Androgenetic development of the egg of Rana pipiens. Blol.
Bull, 77 : 233-257.

RENSCH, BERNHARD, 1936. Studien iiber klimatische Parallelitat der Merkmal-
sauspragung bei Vogeln und Saugern. Arch. f. Naturgeschichte, N.F.,
5 : 317-363.

SCHMIDT, K. P., 1938. A geographic variation gradient in frogs. Field Mus. of
Nat. Hist., Zool. Scr., 20 : 377-382.

SCHREBER, H., 1782. Beitrag zur Naturgeschichte der Frosche. Der. Natur-
forscher, Halle, 18: 182-193.

SHUMWAY, WALDO, 1940. Stages in the normal development of Rana pipiens.
Anat. Rec., 78: 139-147.

STEJNEGER, LEONHARD, AND THOMAS BARBOUR, 1939. Check List of North Ameri-
can Amphibians and Reptiles. Fourth edition, Harvard University Press,
Cambridge, Mass.

WILSON, E. B., 1925. The Cell in Development and Heredity. Third edition, The
Macmillan Co., New York.

WRIGHT, A. H., 1933. Handbook of Frogs and Toads. The Comstock Publishing
Co., Ithaca, N. Y.

THE RELATION BETWEEN HYDROGEN-ION CONCEN-
TRATION AND VOLUME, GEL/SOL RATIO AND
ACTION OF THE CONTRACTILE VACUOLE
IN AMOEBA PROTEUS1

COLEEN FOWLER

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

INTRODUCTION

Von Limbeck (1894) observed that erythrocytes swell if the concen-
tration of carbon dioxide in the blood is increased. Jacobs and Parpart
(1931) found that hemolysis in erythrocytes increases with increase in
acidity and that the effect of as small a change as 0.01 pH is measurable.
Lucke and McCutcheon (1926) maintain that the volume of eggs of
Arbacia in sea water does not change with changes in hydrogen-ion con-
centration between pH 4.2 and pH 9.8 unless the eggs are left so long
that they become injured.

Chalkley (1929), in observations on Amoeba protcus in balanced salt
solution, found that as the hydrogen-ion concentration decreases from
pH 6 the volume decreases to a minimum at pH 7 and then increases,
i.e. that there are two maxima, one in the acid range and another in the
alkaline; and Mast and Prosser (1932) found that as it decreases over
the range studied (pH 5.4—8) the gel/sol ratio decreases. Pitts and
Mast (1934) investigated the relation between gel/sol ratio and
hydrogen-ion concentration in single as well as balanced salt solutions.
They confirmed the results obtained by Mast and Prosser and conclude
that " in sodium or potassium salt solutions the gel/sol ratio decreases
as the hydrogen-ion concentration decreases, but that in calcium salt solu-
tions it increases in the more acid range (pH 5.0 to pH 5.9) then re-
mains constant or decreases slightly."

Thus it will be seen that there is considerable difference of opinion
concerning the effect of changes in hydrogen-ion concentration of the
medium on the volume of cells, that no observations have been made on
Amoeba concerning the relation between volume and hydrogen-ion con-

1 These investigations were carried out under the direction of Professor S. O.
Mast in the Zoological Laboratory of the Johns Hopkins University and the
Marine Biological Laboratory at Woods Hole, Mass. They were greatly facili-
tated by a grant from the Brooks Fund.

265

266 COLEEN FOWLER

centration in single salt solutions, and that the results obtained in ob-
servations on the gel/sol ratio in these solutions have not been confirmed.
It is the purpose of this paper to present detailed information concerning
the relation between hydrogen-ion concentration, volume and the gel/
sol ratio of Amoeba frotcus in salt solutions containing respectively
sodium, potassium and calcium as the only metallic cations.

MATERIAL AND METHODS

The amoebae used were selected, prepared, and measured for volume
and gel/sol ratio as described in the section on material and methods in
a previous paper (Mast and Fowler, 1935). The solutions used con-
sisted of a primary phosphate hydroxide buffer system in which the
concentration of the cation was identical in the phosphate and in the
hydroxide (Pitts and Mast, 1933). The stock solution of phosphate
was kept in a covered Pyrex flask and the stock solution of hydroxide
in a carefully sealed Pyrex flask open to the exterior through a soda
lime tube and through a 50 cc. Pyrex glass buret. These solutions were
standardized according to the method described by Pitts and Mast
( 1933). By mixing the phosphate and the hydroxide in various propor-
tions the desired hydrogen-ion concentration was easily obtained. The
hydrogen-ion concentration of each solution prepared was measured with
a quinhydrone electrode and a Leeds Northrup potentiometer.

VOLUME AND GEL/SOL RATIO
Sodium Salts

Ten amoebae were selected, put into modified Ringer solutions,2 and
left for approximately 24 hours. Then the volume and the gel/sol ratio
of each were measured as described above, after which they were trans-
ferred to a solution containing 0.002 M sodium as the only metallic
cation at pH 5.5, left 15 minutes and measured again, after which they
were measured at 15-minute intervals for 105 minutes. This was re-
peated for 60 other individuals, 10 in each of the following solutions :
0.002 M sodium phosphate buffer solutions at pH 6.0, pH 6.5, pH 7,
pH 7.5, pH 8.0 and pH 8.8, respectively. There was but little change
in either the volume or the gel/sol ratio of the amoebae after they had
been in these solutions 30 minutes. All the results obtained in the

2 3.3 cc. salt solution (0.35 gram NaCl, 0.14 gram KC1, 0.12 gram CaCl., 1000
cc. H.XD) + 5 cc. buffer solution (25 cc. 0.2 M KH2PO4, 12.5 cc. 0.2 M NaOH,
62.5 cc. H2O ; Clark, 1927) +91.7 cc. H2O. This solution is the same in composi-
tion as Chalkley's 1/60 Ringer solution (1929). The total concentration of salts
is 0.002 M and the hydrogen-ion concentration is pH 6.8.

pH AND VOLUME IN AMOEBA PROTEUS

267

measurements of volume and gel/sol ratio made at each hydrogen-ion
concentration were therefore respectively thrown together and the aver-
age calculated. These averages are presented in Fig. 1, A.

Figure 1, A shows that after the amoebae had been transferred from
modified Ringer solution 0.002 M, pH 6.8, to 0.002 M sodium phos-

.2
a

o
c/)

"33
O
•o

_3
O

+ io

+ r

l
j

+ 11U

i i nc

5
n

•

-i- 1 1 0 —

™

^-v-^-

^-v

C^^

,^Vv-

- — ^

+ A n ""n

- 1 5

•

)

4U °

4-a

;

OJ
CM

— L\)

_l_ c

3ff\

•*

)

20 0

c/5

+ 1 C

•

)

• t

10 ^j

0
4- 10 T3

•
\ \

451

)

| 1U g

4-52

..c

•••'-<•

/ c

S
0 "o

— 15

*•.

N
.<

5 •-

\'\

f'eo'

3 tu

- 10 M

•

U
- 15

FT*

k
\

5.0

6.0

7.0

8.0

5.0 5.5 6.0

FIG. 1. The relation between volume and gel/sol ratio in Amoeba proteus
and hydrogen-ion concentration in solutions which contain only one kind of metallic
cations. A, sodium 0.002 M; B, potassium 0.002 M; C, calcium 0.005 M; solid
curves, volume ; broken curves, gel/sol ratio ; 0, volume and gel/sol ratio in modi-
fied Ringer solution; +, increase in volume and gel/sol ratio; — , decrease in vol-
ume and gel/sol ratio; points on curves, averages of 3 to 50 measurements (see
text) ; 15', 15 min. after transfer from Ringer solution to Ca solution ; 30', 30
min. ; 45', 45 min. ; 60', 60 min. ; 75', 75 min.

phate buffer solutions at the various hydrogen-ion concentrations used,
the average volume calculated from measurements made at 15-minute
intervals for 105 minutes after transfer, decreased 5 per cent at pH 5.5,
0.3 per cent at pH 6.0, 7 per cent at pH 6.5, 10 per cent at pH 7.0,
7.3 per cent at pH 7.5 and 10 per cent at pH 8.0. (Measurements at

268 COLEEN FOWLER

pH 8.8 were impossible because the amoebae disintegrated within a few
minutes after they had been put into the solutions.)

This indicates that as the hydrogen-ion concentration in a sodium
solution decreases from pH 5.5 the volume increases to a maximum at
pH 6, then decreases to a minimum at pH 7, then increases to a sec-
ondary maximum at pH 7.5, and then decreases again.

Figure 1, A also shows that the gel/sol ratio decreased 5.5 per cent at
pH 5.5, 27 per cent at pH 7 and zero per cent at pH 8, and that it in-
creased 7.5 per cent at pH 6.5 and pH 7.5. This indicates that in a
sodium solution as the hydrogen-ion concentration decreases from pH
5.5 the gel/sol ratio increases to a maximum at pH 6.5, then decreases
rapidly and very greatly to a minimum at pH 7, then increases equally
rapidly and greatly to a second maximum at pH 7.5 and then decreases
again.

Potassium Salts

The experiments concerning the relation between volume and gel/sol
ratio and hydrogen-ion concentration in solutions containing potassium
as the only metallic cation were performed the same as those containing
only sodium, except that three specimens were used for each hydrogen-
ion concentration in place of ten. The results obtained show that there
was, as in the sodium solutions, but little change in volume and gel/sol
ratio after the amoebae had been in the solutions 15 minutes. The aver-
ages of all the results obtained concerning volume and gel/sol ratio were
therefore respectively calculated. These averages are presented in Fig.
1, B.

Figure 1, B shows that in amoebae transferred from modified Ringer
solution, pH 6.8, to potassium phosphate buffer solutions at various
hydrogen-ion concentrations, the volume decreased 10 per cent at pH
5.0, 16.1 at pH 6.0, and 3.7 at pH 6.5, and that the gel/sol ratio increased
4 per cent at pH 5 and about 1 at pH 6.5 and decreased 33.8 per cent
at pH 6. In neutral and alkaline solutions the amoebae disintegrated so
rapidly that it was impossible to measure them.

These results indicate that as the hydrogen-ion concentration in po-

.tassium solutions decreases from pH 5 the volume of Amoeba proteus

decreases slowly to a minimum at pH 6 and then increases rapidly ; and

that the gel/sol ratio decreases very rapidly to a minimum at pH 6 and

then increases equally rapidly.

Calcium Salts

The methods used in the observations on amoebae in solutions in
which calcium was the only metallic cation present are the same as those

pH AND VOLUME IN AMOEBA PROTEUS 269

used in the preceding experiments. The concentration of calcium was
0.005 M and ten amoebae were measured at each of four hydrogen-ion
concentrations : pH 5, 5.5, and 6. In lower hydrogen-ion concentrations,
it was impossible to maintain the concentrations long enough to make
the measurements.

The results obtained show that there was but little change in volume
after the amoebae had been in the solutions 15 minutes, but that the
gel/sol ratio changes radically with time. In reference to volume, the
average for all the measurements made at each hydrogen-ion concen-
tration was therefore calculated ; but in reference to gel/sol ratio the
average of the results obtained in the measurements made after each 15-
minute period at each hydrogen-ion concentration wras calculated. These
averages are presented in Fig. 1, C.

Figure 1, C shows that in the amoebae which were transferred from
modified Ringer solution pH 6.8 to calcium phosphate buffer solutions
the volume increased 0.8 per cent at pH 5 and 4.8 per cent at pH 6 and
that it decreased 1.4 per cent at pH 5.5. This indicates that as the
hydrogen-ion in calcium solutions decreases from pH 5 the volume de-
creases slightly to a minimum at pH 5.5 and then increases fairly rapidly.

This figure shows that during the first 15 minutes after the amoebae
had been transferred from modified Ringer solution the gel/sol ratio
increased 110 per cent at pH 5.5, 107 per cent at pH 6.5 and 12 per cent
at pH 6, and that it then decreased during the following 90 minutes to
16 per cent below the original ratio at pH 5, 15 per cent at pH 5.5 and
20 per cent at pH 6.

This indicates that after transfer from Ringer solution to calcium
solution the gel/sol ratio increases very rapidly and very extensively, if
the hydrogen-ion concentration is relatively high, and then gradually de-
creases and that the extent of the change in this ratio varies directly
with the hydrogen-ion concentration.

THE ACTION OF THE CONTRACTILE VACUOLE

The results presented by Chalkley (1929) and those presented in the
preceding pages show that the volume of Amoeba proteus is correlated
with the hydrogen-ion concentration of the surrounding medium. The
question now arises as to whether or not this correlation is dependent
upon the action of the contractile vacuole. This problem was investi-
gated as follows.

Fifty to one hundred amoebae were transferred successively through
three beakers each containing 50 cc. redistilled water and left in the last
for one hour. During this time many of the amoebae became radiate in

270 COLEEN FOWLER

form. About 25 of these were selected and put into 50 cc. 0.002 M
Ringer solution (pH 6.8) and left 12-15 hours, then an actively moving
specimen was selected and measured in the volumescope as previously
described (Mast and Fowler, 1935). It was then transferred to test
solution, in the depression on a Pyrex glass slide, and covered with a
cover-glass, after which the diameter of the vacuole, immediately pre-
ceding contraction, and the interval between successive contractions were
measured with a Filar micrometer ocular and with a stop watch re-
spectively. This was continued as long as desired, after which the
whole process was repeated with other individuals in this and in other
solutions. Then the average volume of fluid eliminated by the con-
tractile vacuole per amoeba per minute and the average volume elimi-
nated per minute in percentage of the volume of the amoebae were
calculated for each solution used.

There was considerable variation in given amoebae during the period
of observation in the different solutions used and in the intervals between
successive contractions, but these variations were not specifically corre-
lated with time in any of the solutions except the Ringer-lactose solution,
a solution in which the osmotic concentration was relatively very high.
In this solution the size of the vacuole decreased and the interval be-
tween contractions increased with time and there usually were not more
than five contractions before it ceased to function altogether. These
statements are substantiated by the following typical results.

In one of the amoebae transferred to Na buffer solution (pH 6.5),
the diameter of the first vacuole in the series of ten measured was 2.36 p,
that of the last 2.30 p,, that of the smallest 2.07 ^ and that of the largest
2.66 p.

In one of the amoebae transferred to the Ringer-lactose solution, the
diameter of the first vacuole was 2.42 ^ and those of the following three
were 2.79 /*, 2.69 /A and 2.38 /A respectively, and the intervals between the
successive contractions in the series were 2' 45", 3' 5", 5' 0" and 6' 30".
After this series of contractions was complete the vacuole was con-
tinuously observed for 45 minutes. It did not contract during this time
but it became smaller, the diameter at the end of three successive 15-
minute intervals having been 1.75 (i, 1.66ju, and 1.57 /A respectively. In
this amoeba the vacuole ceased contracting 17 minutes and 20 seconds
after transfer to the Ringer-lactose solution. The average time required
for cessation of contraction in this solution was 18.5 minutes.

The averages of the results obtained directly by observation of
amoebae in the different solutions used and those obtained by calcula-
tions are presented in Table I.

pH AND VOLUME IN AMOEBA PROTEUS

271

This table shows that the size of the contractile vacuoles in the
amoebae in all the different solutions used except the Ringer-lactose
solution and the Na buffer (pH 8) was essentially the same.

In the Ringer-lactose solution the vacuole decreased markedly with
time, as stated above, hence the low average diameter of 2.09 p. The

TABLE I

The volume of fluid eliminated by the contractile vacuole in Amoeba proteus in
various solutions. Temperature, 25° C. Ringer and Na buffer solutions are described
in the text. The volume of only three of the nine amoebae in pure water was
measured. The average volume for these is 4980 CM and the average elimination
0.30 per cent of this volume.

Amoebae studied

Contractile vacuoles

Solutions used

Number

Av. vol.
in 1000
CM

Number
meas-
ured

Av.
diam.
in M

Av. in-
terval
between
contrac-
tions

Av. vol. of
fluid elimi-
nated in CM
per min. per
amoeba

Av. per-
centage of
vol. of
amoebae
eliminated

•

per min.

Ringer

0.002 M

pH 6.8

1785

100

2.46

3' 49"

5670

0.31

Ringer

0.002 M

pH 6.8

-0.2 M

lactose

2.09

9' 47"

1500

0.008

Na buffer

0.002 M

pH 5

1753

2.29

4' 5"

4260

0.24

Na buffer

0.002 M

pH 6

1891

2.25

4' 10"

4620

0.24

Na buffer

0.002 M

pH 6.5

1619

2.22

4' 4"

3900

0.24

Na buffer

0.002 M

pH 7

2184

2.43

5' 31"

7380

0.17

Na buffer

0.002 M

pH 8

3141

100

2.76

4' 2"

7620

0.24

Pure water

1625

2.61

3' 36"

4980

0.30

high average diameter of 2.76 p. in the Na buffer (pH 8) appears to
have been directly correlated with the size of the amoebae in this
solution.

The table shows that the average interval between successive contrac-
tions was relatively small in pure water and Ringer solution and consid-
erably higher but essentially the same in all the Na buffer solutions

272 COLEEN FOWLER

except pH 7 in which it was relatively very high. It shows that the
rate of elimination per unit volume of protoplasm was relatively high in
pure water and Ringer solution and considerably lower but essentially
the same in all the Na buffer solutions except pH 7, in which it was
much lower.

The results obtained consequently indicate that the rate of elimina-
tion of fluid by the contractile vacuole in Amoeba proteus is practically
independent of the hydrogen-ion concentration except in the region of
neutrality where it decreases markedly and that it gradually decreases to
zero in hypertonic solutions.

DISCUSSION

The results presented in Table I show that in amoebae which have
been transferred from Ringer solution to sodium solution of various
hydrogen-ion concentrations, there was a change in the rate of elimina-
tion of fluid by the contractile vacuole of only 0.07 per cent of the
volume of the amoebae per minute. It is consequently obvious that the
action of the contractile vacuole was only slightly involved in the changes
in the volume of the amoebae observed, in relation to changes in
hydrogen-ion concentration (Fig. 1), and that these changes were con-
sequently largely due to the effect of the hydrogen-ion concentration of
the solutions used, on the permeability of the surface layer to water.

Table I and other evidence presented above show that in the amoebae
which had been transferred from 0.002 M Ringer solution to 0.002 M
Ringer solution plus 0.2 M lactose, the rate of elimination of fluid by
the contractile vacuole decreased from 0.3 per cent of the volume of the
amoebae per minute to zero in an average of 18.5 minutes. No obser-
vations were made on the action of the vacuole in amoebae which had
been transferred in the opposite direction, but it is highly probable that
after such a transfer the vacuole becomes active as rapidly and to the
same extent as it becomes inactive after the reverse transfer. If this
is true, decrease in volume of amoebae in hypertonic solution is consid-
erably augmented and increase in volume of amoebae in hypotonic solu-
tion is considerably retarded, owing to elimination of fluid by the
vacuole; that is, the amount of fluid which leaves the amoebae directly
through the surface in the hypertonic solution is equal to the increase in
volume minus the amount eliminated by the vacuoles and the amount
which enters the amoebae directly through the surface in the hypotonic
solutions is equal to the increase in the volume of the amoebae plus the
amount eliminated by the vacuoles.

pH AND VOLUME IN AMOEBA PROTEUS 273

Mast and Fowler (1935) calculated the permeability constant for
water from results obtained in observations on the increase in the volume
of amoebae in hypotonic solutions, but they did not consider the effect
of the action of the contractile vacuole on the volume. As stated above,
this probably amounted to 0.3 per cent of the volume of the amoebae
per minute soon after the transfer to these solutions. The calculated
value obtained by them (0.026) is therefore somewhat too small.

In the preceding paper it was demonstrated that after amoebae have
been transferred from Ringer solution to this solution plus 0.2 M
lactose, they decrease about 15 per cent in volume in 15 minutes and it
was demonstrated above that under these conditions they continue to
contract for about 18.5 minutes. It is therefore obvious that elimination
of fluid through the vacuole continues after considerable fluid has passed
out of the body by diffusion and that the action of the vacuole is not
immediately dependent upon entrance of fluid and turgidity of the cell.
This also obtains for other protozoa (Kitching, 1938, p. 148).

Pitts and Mast (1934), in observations on the gel/sol ratio in
Amoeba proteus, obtained results which in general support the conclusion
reached above, namely that the gel/sol ratio is relatively low in the region
of neutrality. They also found that the rate of locomotion is low in this
region and Table I above shows that the rate of elimination of fluid by
the contractile vacuole is also low in this region. This indicates that the
rate of locomotion and the action of the vacuole vary inversely with the
fluidity of the cytoplasm and that it is maximum at neutrality. This is
probably in some way correlated with the isoelectric point of a prominent
protein in the cytoplasm.

The extraordinary changes observed in the gel/sol ratio in calcium
solutions indicate remarkably rapid and extensive adjustment but con-
cerning the processes in this adjustment there is no evidence.

SUMMARY

1. As the hydrogen-ion concentration decreases from pH 5.5 the
volume of Amoeba proteus in solutions containing sodium as the only
metallic ion increases to a maximum at pH 6.0, then decreases to a
minimum at pH 7.0, and then increases to a second maximum at pH
7.5 ; and the gel/sol ratio increases to a maximum at pH 6.5, then de-
creases very extensively to a minimum at pH 7.0, and then increases
equally extensively to a second maximum at pH 7.5.

2. In solutions containing potassium as the only metallic ion the vol-
ume and the gel/sol ratio decrease to a minimum at pH 6.0 and then
increase.

274 COLEEN FOWLER

3. In solutions containing calcium as the only metallic ion the volume
remains nearly constant, but the gel/sol ratio increases very rapidly and
extensively and then gradually decreases ; but the extent of change in
this ratio varies directly with the hydrogen-ion concentration.

4. The rate of elimination of fluid by the contractile vacuole is prac-
tically independent of the hydrogen-ion concentration except in the re-
gion of neutrality where it decreases markedly. In hypertonic solutions
it gradually decreases to zero.

5. The change in the rate of elimination of fluid by the vacuole in
relation to hydrogen-ion concentration is so low in comparison with the
change in rate of passage of fluid directly through the surface that it is
negligible. The changes observed in the volume of the amoebae in
relation to the hydrogen-ion concentration were therefore almost entirely
due to changes in the rate of transfer of fluid directly through the sur-
face, i.e. to changes in the permeability of the surface to water.

REFERENCES CITED

CLARK, W. M., 1927. The Determination of Hydrogen Ions. Baltimore, 480 pp.
CHALKLEY, H. W., 1929. Changes in water content in Amoeba in relation to

changes in its protoplasmic structure. Physiol. Zodl., 2 : 535-574.
JACOBS, M. H., AND A. K. PARPART, 1931. Osmotic properties of the erythrocyte.

II. The influence of pH, temperature, and oxygen tension on hemolysis by

hypotonic solutions. Biol. Bull., 60: 95-119.
KITCHING, J. A., 1938. The physiology of contractile vacuoles. III. The water

balance of fresh-water Peritricha. Jour. Exper. Biol., 15: 143-151.
LUCRE, B., AND M. MCCUTCHEON, 1926. The effect of hydrogen-ion concentration

on the swelling of cells. Jour. Gen. Physiol., 9 : 709-714.
MAST, S. O., AND COLEEN FOWLER, 1935. Permeability of Amoeba proteus to

water. Jour. Cell, and Comp. Physiol., 6: 151-167.
— , 1938. The effect of sodium, potassium and calcium ions on changes in

volume of Amoeba proteus. Biol. Bull., 74 : 297-305.

MAST, S. O., AND C. L. PROSSER, 1932. Effect of temperature, salts, and hydrogen-
ion concentration on rupture of the plasmagel sheet, rate of locomotion,

and gel/sol ratio in Amoeba proteus. Jour. Cell, and Comp. Physiol., 1 :

333-354.
PITTS, R. F., AND S. O. MAST, 1933 and 1934. The relation between inorganic salt

concentration, hydrogen-ion concentration, and physiological processes in

Amoeba proteus.
— , 1933. Rate of locomotion, gel/sol ratio, and hydrogen-ion concentration

in balanced salt solutions. Jour. Cell, and Comp. Physiol., 3 : 449-462.
— , 1934. Rate of locomotion, gel/sol ratio, and hydrogen-ion concentration

in solutions of single salts. Jour. Cell, and Comp. Physiol., 4 : 237-256.
— , 1934a. The interaction between salts (antagonism) in relation to hydrogen-
ion concentration and salt concentration. Jour. Cell, and Comp. Physiol.,

4: 435-455.
VON LIMBECK, R. V., 1894. tiber den Einfluss des respiratorischen Gewechsel auf

die Rothen Blut Korperschen. Arch. f. cxp. Path., 35: 309-335.

Vol. LXXX, No. 3 June, 1941

THE

BIOLOGICAL BULLETIN

PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY

PIGMENT MIGRATION IN THE EYES OF THE MOTH,
EPHESTIA KUEHNIELLA ZELLER

M. F. DAY
(From the Biological Laboratories, Harvard University)

INTRODUCTION

The movement of pigment in the eyes of insects has been described
many times, but there is little information in the literature suggesting
what factors are involved in bringing about this movement. In certain
Crustacea it has been conclusively demonstrated (Kleinholz, 1936, 1938;
Welsh, 1939, 1941) that the movement of the pigment from the " dark "
position into the " light " position is dependent upon the action of a
hormone secreted by the eye stalk. In view of recent discoveries of
reciprocal effects between insect and crustacean hormones (Hanstrom,
1937; Brown and Meglitsch, 1940), the question arises as to whether
endocrines also regulate pigment migration in the moth eye. There are
at least three ways in which eye pigment migration in an insect might
be controlled, namely, by hormones, by nerves, and by the action of the
pigment cells as independent effectors. Possibly also a combination
of these three methods of control might be found in some insects.

Many of the early workers (see Demoll, 1911) believed that the con-
trol was nervous in nature. But the experiments of Exner and Demoll
(cf. review by Parker, 1932), and the subsequent work of Uchida
( 1934) on the long-horned grasshopper, indicated that the movement
was generally slow in comparison with most nervous responses. Friza
(1928) suggested, as a result of his study of Mantis religiosa, that the
movement was Immorally controlled. In addition, the work of Horst-
mann (1935), in which he corroborated the conclusions of the older
workers on the existence of a diurnal rhythm in pigment migration
under constant conditions of the external environment, indicates that the
response is not that of independent effectors, in which the pigment
cells respond directly to light. Collins (1934), however, concluded
that it was most likely that light acted directly on the pigment cells of
the moth, Carpocapsa.

In the following experiments, the movement of the eye pigment of
the moth, Ephcstia kneliniella Zeller, will be described, and an attempt

275

276 M. F. DAY

made to determine which of the three above-mentioned factors is re-
sponsible for the movement. The study is based on the examination of
serial sections of the heads of over 450 moths.

The moths were bred in large glass jars, each containing about two
inches of dry oatmeal. Under these conditions the life-cycle occupied
about six weeks, a plentiful supply of adult moths thus being available
at all times. Observations were made on sections of entire heads fixed
in alcoholic Bouin's fluid, which penetrated with sufficient rapidity that
no migration of pigment occurred. Results were comparable in this
regard with those obtained with fixation by hot water, but histological
preservation was better. Some sections were stained in Mallory's triple
stain or impregnated by Bodian's protargol method, and others were
depigmented in Grenadier's fluid, and subsequently stained or im-
pregnated.

It is a pleasure to thank Professor J. H. Welsh and Dr. L. H. Klein-
holz for suggestions during the course of this work.

ANATOMICAL CONSIDERATIONS

Umbach (1934) described the eye structure of Ephestia in detail.
The following description will therefore deal only with the pigment
cells and other points which directly concern our discussion. Features
of a single ommatidium are shown in Fig. 1.

The pigment in the eye is arranged in three distinct groups, but at-
tempts to homologize these with the pigment cells of Crustacea were
unsuccessful. One group of pigment granules surrounds the bases of
the rhabdoms and also extends below the basement membrane. Umbach
(1934) has shown that this pigment in Ephestia is not contained in
specific retinal pigment cells, as it is in many insects. Although in
Ephestia the pigment is never completely withdrawn beneath the base-
ment membrane, as in Vanessa (Demoll, 1917), it apparently exhibits
some movement. But the extent of this movement does not exceed ten
microns, and no attempt has been made to determine its cause.

A second group of pigment granules is contained in cells which have
a varied terminology. They are here called primary pigment cells, but
have been referred to as iris pigment cells by Snodgrass (1935), as
primary iris cells by Wigglesworth (1939), distal pigment cells by Col-
lins, and accessory pigment cells by Uchida. These cells closely sur-
round the crystalline cones. The nuclei are extremely flattened, and
the pigment granules are arranged in a thin layer around the cones.
Neither these cells nor their pigment granules have been observed to
migrate. Finally, the large accessory pigment cells (secondary iris cells,
Wigglesworth; principal pigment cells, Uchida) are very conspicuous

PIGMENT MIGRATION IN MOTH EYES

277

and contain most of the pigment granules. The entire cell moves, as
well as the pigment granules contained in it, for the nuclei have been
shown in depigmented preparations to have moved through a distance
of about 30 microns, with the mid-point at about the proximal end of

-CORNEAL LENS-

-PRIMARY PIGMENT.
CELL NUCLEI

-CRYSTALLINE CONE-

ACCESSORY PIGMENT.
CELL

•RETINULAR CELL-
NUCLEI

-FILAMENT FROM—
CONE TO RHABDOM

RETINULAR CELLS-

• RHABDOM-

-TRACHEOLES-

BASAL CELL

BASEMENT MEMBRANE'

NERVE FIBERS

FIG. 1. Longitudinal sections of two ommatidia from eyes of Eplicstia in (a)
dark-adapted, and (/>) light-adapted positions, respectively.

the crystalline cone. None of the pigment in the eyes of EpJiestia has a
reflecting function.

Another important aspect of the structure of the eye is the size of
the ommatidia. Umbach (1934) states that there are between 2,000
and 2,500 in each eye. Between 60 and 70 of these are seen in each
eye in a transverse section of the head (Fig. 2). The ommatidia vary

278

M. F. DAY

in size, the smallest being on the dorsal side, and the largest on the
ventral side of the eye, the difference between the two extremes being
about 20 per cent of their length. The accessory pigment cells migrate
a greater distance in the larger ommatidia than they do in the smaller
ones, as in Crustacea, where the greatest amount of migration occurs
in the largest ommatidia which are found on the dorsal side of the eye.

FIG. 2. Transverse section of part of head of Ephestia, showing the arrange-
ment and differences in size of the ommatidia. Pigment cells in light-adapted
position.

These differences make it necessary to select the central ommatidia for
comparisons between eyes of insects treated in different ways. Also,
the eyes of female moths are on the average slightly larger than those
of males. The average length of the central ommatidia of ten males,
measured from the basement membrane to the distal end of the crystal-
line cone, was 177.3 microns, while that from ten females was 186.8
microns, longer than the average of the males by 5 per cent. Therefore

PIGMENT MIGRATION IN MOTH EYES

insects of only one sex were used in experiments when they were 10 be
directly compared.

Two other details of anatomy should be mentioned. The first is
that no contractile fibers, such as were demonstrated by Welsh (1930)
in the eyes of Palaeinonctcs, have been seen in the eyes of Ephcstia.
Movement of pigment granules and of the pigment cells probably results,
as was suggested by Bennitt (1924), and Parker (1932), from proto-
plasmic streamings or surgings, and a similar mechanism will probably
explain the interesting movement (first demonstrated by Umbach) of
the retinular cells themselves. Finally, Bodian preparations of both
pigmented and depigmented eyes showed no sign of effector nerves
supplying the accessory pigment cells. The significance of such nega-
tive evidence is, however, questionable in view of the difficulty of
demonstrating nerves to sense hairs which occur in large numbers on
the eyes of some insects.

XORMAL MOVEMENT OF THE ACCESSORY PIGMENT CELLS

The normal movement of the accessory pigment cells, such as occurs
when a dark-adapted moth is exposed to light, can be followed by re-
ferring to Fig. 3. First, it is to be noticed that the extreme dark posi-
tion, as seen in A, is rarely found. Moths kept for several days in the
dark more frequently have their accessory pigment cells in a position
similar to that in B. In this stage pigment granules are evenly dis-
tributed and extend from the level of the distal ends of the cones to
about 15 microns beneath their proximal ends, the level of the grouped
retinular cell nuclei. The first movement of pigment in such an eye
usually becomes evident as an increasing aggregation of the granules in
the proximal part of the cells. This region of the cell then moves
proximally toward the basement membrane (Fig. 3, D), accompanied
by a simultaneous decrease in the size of the extensions of the pigment
cells between the cones. As the proximal movement continues these ex-
tensions become attenuated (Fig. 3, £), and the characteristic " frayed "
region at the distal end is eventually withdrawn (Fig. 3, F). Between
the two last-mentioned stages, movement of the pigment cell nuclei
occurs. The final stage of movement involves the almost complete with-
drawal of granules of the accessory pigment cells from between the
cones (Fig. 3, G). Light alone is incapable of causing greater proximal
movement than that seen in Fig. 3, G, but certain treatments described
below may induce more extensive movement.

The movement in Ephestia is not so simple a migration as was indi-
cated by Collins (1934, PI. 4) in Carpocapsa. Measurements of the
distance of the pigment cells from the cones do not indicate all the

280 M. F. DAY

changes that occur, since two types of movement, lateral as well as longi-
tudinal, are found. These are seen in stages E and F, and are equivalent
to those observed by Peabody (1939) in the isopod, Idothca. The fact,
well known in Crustacea, that the type of pigment migration varies in
different species, holds also for insects.

EXPERIMENTAL RESULTS
Diurnal RJiytJnn

The first experiments were made to determine whether a persistent
diurnal rhythm existed in the movement of the pigment cells, as previ-
ously recorded for many insects and Crustacea (for review, see Welsh,
1939). In Ephcstia, kept under constant illumination for four days,
the pigment cells were found in the light-adapted position at any hour
of the day or night. Likewise, insects kept in darkness, with all other
usually controlled factors maintained constant, had pigment cells in the
dark position during both day and night. This is believed to be the first
reported absence of a persistent diurnal rhythm in a moth. However,
in view7 of the marked interspecific differences recorded by Welsh
(1935), even within a single genus of crustaceans, and the absence of
diurnal rhythm of pigment migration in Palaemonetes (Kleinholz.
1936), the differences between various insects are perhaps not so sur-
prising.

The majority of previous workers on Lepidoptera have made their
observations on the changes of the pseudopupil, or on the presence of
glow in the eye when illuminated at night. It has rarely been possible to
demonstrate glow7 during the day. In Ephcstia in constant darkness the
eyes glow during the day as well as at night. The glow, however, dis-
appears rapidly upon illumination, — as shown for other insects by many
observers (see Merker, 1929). At a light intensity of 12 ft-candles,
1 to 2 minutes are required for the glow to disappear from dark-adapted
Ephcstia at 25° C. Sections of eyes which exhibited glow and those
from which it had just disappeared show7 that the changes in position of
the pigment cells are frequently almost imperceptible. Glow in EpJicstia
is not caused by reflecting pigment, as in some Crustacea, nor does it
seem likely that its appearance is due to the reflection of light from the
tracheal tapetum between the rhalxloms. The cause of glow in insects
deserves further study.

The Effect of Light

Intensity. — The effect of light of various intensities was determined.
A standard source (15-watt Mazda bulb, filtered through Wratten
Neutral Filter) was maintained at a constant distance from a group of
insects which were inclosed, as in all subsequent experiments, in thin.

PIGMENT MIGRATION IN MOTH EYES 281

2-clram glass vials, plugged with cotton wool, and standing, bottom
uppermost, in a container with white walls. All experiments were begun
with an excess of moths, and insects which were persistently active, or
which were settled facing away from the light, were not used. It was
soon found that several factors had to be controlled in order to obtain
consistent results. Since it seemed that the rate of response was in-
fluenced by the age of the moths, only newly emerged virgins were
used. A further complication was found in some moths due to the
presence of a parasite. In such cases many spores were found in sec-
tions of the head, usually in the hemocoele, but sometimes also in the
tissues. Such individuals were not considered in the experiments.

Light intensity was varied by accurately calibrated neutral filters.
The results are indicated in Fig. 4. In this method of presenting data,
the technique employed involves measuring the pigment migration of a
number of individuals (in this case, five), and then photographing an
eye whose measurement is nearest the average. After 30 minutes ex-
posure to an intensity of approximately 0.3 ft-candles, migration into
the light position is incomplete. Above 3 ft-candles a maximum re-
sponse is produced in the same period of time. While these results
are readily reproducible, single measurements of proximal migration are
insufficient to permit a determination of the exact type of relationship
between intensity of light and distance migrated. In all subsequent
experiments an intensity of 12 ft-candles was employed.

The Rate of Movement. — The next step was to determine the time
taken for the migration of the pigment cells in both directions, when
intensity, temperature, and other factors were kept constant. The only
method which is available in the case of Epliestia is to fix and section
eyes of moths which have been exposed to light for varying periods of
time. Since this method has been shown by Welsh (1930) to be un-
satisfactory in comparison with methods of direct observation which are
applicable to animals with stalked eyes, the times given below can only
be considered to indicate the approximate rates.

The results of one experiment are recorded in Fig. 3 (B—G). Under
the conditions of the experiment, movement could be demonstrated
within one and one-half minutes after the exposure of the moths to
light. Movement was most rapid between 4 and 7 minutes, and was
completed within approximately 12 minutes. If, after 30 minutes, the
light is removed, dark adaptation begins within 5 minutes, but takes
somewhat longer for completion than movement under the influence of
light, a conclusion in agreement with that of all previous workers on
Crustacea and insects. In these experiments movement back to the
dark position required about 20 minutes.

282

M. F. DAY

PIGMENT MIGRATION IN MOTH EYES

The Effect of Localized Light. — In the experiments of Bennitt
(1932) exposure of only a small region of one eye of a crayfish resulted
in pigment migration in all ommatidia of that eye and of the other eye
as well. Comparable experiments were performed with etherized
Ephcstia by covering both eyes except for a few ommatidia of one eye
with the opaque mixture recommended by Crozier, Wolf, and Zerrahn-
Wolf (1937). These moths were allowed to recover in the dark, and
after 24 hours were exposed to 12 ft-candles for 20 minutes. Under
these conditions the sectioned eyes frequently showed that only the
region which had been exposed to the light exhibited pigment migration
(Fig. 5). Bennitt's results would be expected on the basis of the
humoral theory of control of migration. The ease with which it is pos-
sible to obtain unequal movement of different pigment cells in the same
eye in Ephcstia suggests that the humoral theory of control may not be
applicable in this case. That other insects may react in a manner
similar to Eplicstia has been indicated by the work of Exner (1891)
and Demoll (1911).

The Effect of Temperature

Rate of Migration. — Experiments exactly simulating those reported
above on the rate of migration under constant light intensity were per-
formed at a temperature of 12.5° C. instead of at 25° C. It was found
that in the dark the pigment cells were extended over a greater distance
proximally at 12.5° C. than at 25° C. (Fig. 6). However, movement
of the pigment cells during light adaptation was in no way different at
the lower temperature, except that they did not have so far to travel
before reaching the light position. The indications are, then, that the
processes of movement have a temperature coefficient of 1.0 within the
range 12.5 to 25° C. (but compare the results of Bennitt (1924) on
Ganunarus).

Extremes of Temperature. — When the factor of temperature was
found to have an effect on the position of the pigment cells in darkness,
a series of experiments was performed to determine the effects of more
extreme temperatures. Moths were placed in light-tight containers at
temperatures of 3°, 5°, 10°, 12.5°, 25°, and 37° C. for two hours. The
moths at the lowest temperature became immobile and those at 37° were
extremely active. All were living, however, and were fixed in the
dark. Sections showed that the eyes of moths at temperatures from
10° to 25° were in the normal dark-adapted position. At 3°, however,
there was considerable movement toward the light position, while at 37°
the pigment cells were concentrated between the cones, thus showing a
more complete dark adaptation than is ever found under normal tern-

284 M. F. DAY

perutures (Fig. 7). This demonstrates for Eplicstia a conclusion
reached long ago for Crustacea by Congdon (1907), that low tempera-
ture produces the same effect as light, and that high temperature has an
opposite effect. In view of this latter finding, the result of two mutually
opposite stimuli was investigated by exposing moths to a temperature of
37° C. and a light intensity of 12 ft-candles at the same time. Under
these conditions the effect of light predominated over the effect of heat
in nine cases. However, in one moth, where the thresholds must have
been approximately equalized, some of the pigment cells migrated into
the extreme light position, while others migrated only slightly (Fig. 8).
Interestingly enough, corresponding ommatidia in both eyes exhibited
the same response, suggesting the possibility of a central control.

The amount of proximal migration produced in the dark by low
temperature is greater in the ommatidia on the ventral side of the eye
than in those on the dorsal side, as is the case in light adaptation at
normal temperatures. The suggestion that such differences in the light-
adapted eye are due to differences in the amount of light reaching each
ommatidium is therefore untenable.

The Effect of Mechanical Stimulation

The rather unexpected effect of shaking which Horstmann (1935)
reported on the phototactic responses of moths has been substantiated
with Eplicstia. If a vial containing a moth is held toward the light and
is tapped lightly, the moth exhibits a marked negative response. If the
tapping is continued for about 15 seconds the moth will suddenly turn,
progress towards the light, and is thereafter strongly positively photo-
tactic. Sections were made of both light and dark-adapted Eplicstia

PLATE II

EXPLANATION OF FIGURES

6. The effect of temperature, showing spreading of pigment granules after 1
min. in light at 12.5° C.

7. The effect of (A) darkness and 3° C., (B) light and 37° C.

8. Irregular movement of pigment cells induced by simultaneous action of high
temperature (37° C.) and light (12 ft-candles).

9. The effect of anaesthetization by ether. Note the clumping of pigment
granules.

10. The effect of the injection of chloretone, inducing greater movement into
the light position than is ever produced by light alone.

11. The irregular movement into the light position induced by high tensions of
carbon dioxide.

12. Transverse section of the head, illustrating the effect of severing the left
optic tract. Pigment in the left eye is in the light position, while that in the eye
on the uninjured side is in the dark position.

PIGMENT MIGRATION IN MOTH EYES

285

PLATE II

286 M. F. DAY

before and after shaking, and of those which were positively and nega-
tively phototactic. It was found that shaking had no demonstrable
effect on the position of the accessory pigment cells, nor was there any
effect of shaking on the rate of disappearance of glow. Conversely,
glow could not be made to reappear just after its disappearance by any
kind of mechanical stimulation. We can only conclude that in the case
of Ephcstia differences in phototactic response are not necessarily cor-
related with obvious changes in the position of the pigment cells.

The Effects of Certain Drugs, Anaesthetics, and Extracts

Methods. — Injections of approximately 0.001 ml. of fluid were made
into the thorax of moths by the mi croinj action method of Ephrussi and
Beadle (1936). Anaesthetization and injection, which occupied between
one and two minutes, were performed in dim light. The moths were
then kept 15 to 20 minutes in darkness before fixing the eyes. Since
it was possible that some of the injected substances might cause move-
ment into the dark position, some of the injected animals were exposed
to light.

EtJicr.- — In concentrations just sufficient to anaesthetize a moth, ether
has the effect of sending the pigment even further into the dark position
than dark alone. A secondary effect is noted when moths are exposed
to ether for longer periods of time. The pigment cells are then ap-
parently disorganized, and have a slight tendency even to move proxi-
mally in an irregular fashion. The pigment granules may assume a
clumped appearance (Fig. 9). The illumination of moths which are
under the influence of either of these effects results in apparently normal
proximal movement.

CJiIoretone. — Injection of substances causing proximal migration of
pigment cells will obscure the effects of ether. When a saturated
aqueous solution of chloretone is injected into a dark-adapted moth, the
pigment cells will migrate into the light position, even though the moths
are kept in the dark. Usually the movement will be comparable to that
produced by light, but sometimes the distance migrated is greater than
is ever produced by light alone (Fig. 10). Chloretone might inhibit
nerve impulses which would result in maintaining the pigment in the
dark position. This inhibition could be exerted either upon motor
nerves or on nerves innervating an incretory organ. A similar criticism
can be made of any experiments with anaesthetics or drugs.

The Effects of Adrenalin, Acctylcholinc, and Prostiginin. — If move-
ment of pigment cells is under the control of the nervous system, it
might be expected that the application of appropriate chemical mediators

PIGMENT MIGRATION IN MOTH EYES

would result in their response. Adrenalin and acetylcholine have both
been extracted from insects (von der Wense, 1938; Corteggiani and
Serfaty, 1939) but the existence of an adrenergic or cholinergic system
has not been proved. The injection, with appropriate controls, of
dilutions of 1:1,000 and 1:10,000 adrenalin, and 1:10,000 and 1:
100,000 acetylcholine produced no change in the position of the pigment
cells. Prostigmin, though having some of the effects on behavior of
EpJiestia comparable to those described for eserine in the mantis by
Roeder (1939), did not influence the pigment cells.

Head Extract. — Experiments comparable to certain of those of
Kleinholz (1936) were performed on Ephestia. The heads of ten light-
adapted moths were triturated in 1 ml. of insect Ringer. The extract
was boiled, cooled, and injected in the manner described above. Neither
this concentration of the extract, nor a dilution of one part of extract
to five of Ringer, produced any proximal migration of the pigment cells
when the insects were kept in the dark. Xor did such injections inhibit
movement when they were exposed to light.

Sinus Gland Extract. — The sinus glands of three specimens of I'ca
pngilator were ground in 0.5 ml. of insect Ringer. The extract was
boiled, cooled, and injected into dark-adapted moths. The results were
negative, as with extracts of Ephestia.

The Effect of Higli Tensions of Carbon Dioxide. — The striking
anaesthetic effect of carbon dioxide on insects has been known for a
long time. In Ephestia carbon dioxide produces a proximal migration
greater than is ever produced by light alone. But this movement is not
uniform, so that all cells do not react to exactly the same extent (Fig.
11). The effect on the position of the pigment is still marked after 20
minutes in the dark, by which time the insects have completely recovered
from the anaesthetic effects. Bennitt and Merrick (1932) have reported
a comparable result in Crustacea due to crowding, which they associate
with oxygen lack.

The Effects of Operative Techniques

With carefully sharpened No. 12 hard steel needles an incision was
made in the head capsule on the side of the f rons along the ocular suture
where the cuticle is easily punctured. From this approach the optic
tract can be severed. Twenty moths were successfully operated upon
in this way, and lived at least four days thereafter. It is of interest to
note that an operation of this kind apparently releases the female from
certain inhibitions to oviposition, for eggs were deposited in the vials
far more frequently by operated than by control females. Several

288 M. F. DAY

operations did not completely sever the optic tract, and the results of
such cases constitute controls. In every case in which the nerve fibers
were completely severed, the pigment in that eye was in the light-adapted
position, irrespective of the position of the pigment in the eye on the
uninjured side. Moreover, once the pigment was brought into the light
position by cutting the optic tract, it could never be caused to migrate
back to the dark position, although appropriate methods (heat, ether,
dark, etc.) were employed. Thus Fig. 12 shows a section of the head
of an insect which had been kept for four days in the dark after an
operation had been performed on the left eye (left side of the illustra-
tion). The pigment cells on the unoperated side occupy the extreme
dark position, while those on the operated side have taken up positions
characteristic of cells released from their usual control, i.e., they have
migrated further proximally than they would under normal light con-
ditions, and their movement is somewhat irregular. The difference
between two such eyes can be detected in the living moth.

DISCUSSION

Two of the above experiments provide strong evidence against a
hormonal control of movement. The effect of short exposures of light
on a small number of ommatidia results in the movement of only a few
pigment cells. A hormone in the blood stream would be more general
in its action. But this experiment does not differentiate between a
nervous control and the possibility that the cells behave as independent
effectors. This latter possibility is shown to be untenable, however, by
the results of cutting the optic tract, which also provides further evi-
dence against the theory of humoral control, since the blood supply to
the eye on the injured side is in no way impaired. There remains only
the theory of the nervous control of pigment migration.

The absence of diurnal rhythms of migration does not argue against
the nervous control, and neither do the effects of light. The rate of
movement is admittedly slow compared with speed of muscular move-
ments in insects, but the movement does not appear to be muscular in
nature. Reasons have been given above for discounting the evidence
concerning the apparent absence of a nerve supply to the pigment cells.
The remote possibility of the liberation of the appropriate chemical
mediator occurring as the result of antidromic impulses along the sensory
nerves should not be overlooked.

The action of chloretone, carbon dioxide, light, and low temperature
would be interpreted on the nervous theory as agents which cause cessa-
tion or decrease of impulses from the brain which normally maintain

PIGMENT MIGRATION IN MOTH EYES

the pigment cells in the dark position. There is actually no positive
evidence in favor of the nervous control since no method has been found
for any arthropod which induces movement into the dark position,
except darkness itself.

In addition, it has been shown in the above experiments that changes
in the phototactic behavior of Ephestia are not entirely dependent upon
the position of the accessory pigment cells. Nevertheless, the variation
in the position of the pigment should be considered in investigations on
the behavior such as that of Brandt (1934), and may explain some of
the irregularities which he reported. Likewise, the results reported by
Taylor and Nickerson (1940) on changes in retinal potentials during
light adaptation should also be considered in relation to the pigment
migration which doubtless occurs in Galleria as in Ephestia. The pos-
sibility of relating electrical response to migration of eye pigments has
already been suggested in the case of some beetles by Jahn and Cres-
citelli (1940).

SUMMARY

The accessory pigment cells in the eyes of Ephestia kuehniella mi-
grate from their distal position between the cones to a more proximal
position when exposed to light of sufficient intensity.

A comparable, or even more marked, effect is produced by low tem-
peratures, chloretone, high tensions of carbon dioxide, and by cutting
the optic tract.

Movement of individual cells can be induced by illuminating only a
few of the ommatidia. This suggests that a hormonal method of control
is unlikely. The fact that pigment cells can never be induced to migrate
into the dark position once the optic tract has been severed suggests that
the cells do not respond as independent effectors.

In view of these several lines of evidence, though most of it is ad-
mittedly negative, it is possible that the migration of the accessory pig-
ment cells in the eyes of Ephestia may be principally controlled by a
nervous mechanism. It should again be emphasized that any conclusions
based on the study of moths cannot necessarily be applied to other
insects, let alone to Crustacea.

LITERATURE CITED

BENNITT, R., 1924. The migration of the retinal pigment in crustaceans. Jour.
Exper. Zool, 40 : 381-435.

, 1932. Physiological interrelationship in the eyes of decapod Crustacea.

Physiol. Zool, 5 : 49-64.

BENNITT, R., AND A. D. MERRICK, 1932. Migration of the proximal retinal pig-
ment in the crayfish in relation to oxygen deficiency. Biol. Bull., 62:
168-177.

290 M. F. DAY

BRANDT, H., 1934. Die Lichtorientierung der Mehlmotte Ephestia kuehniella

Zeller. Zcitschr. vergl Physlol, 20: 646-673.
BROWN, F. A., AND A. MEGLITSCH, 1940. Comparison of the chromatophorotropic

activity of insect corpora cardiaca with that of crustacean sinus glands.

Biol Bull, 79: 409-418.
COLLINS, D. L., 1934. Iris-pigment migration and its relation to behavior in the

codling moth. Jour. E.vper. Zool, 69: 165-198.
CONGDON, E. D., 1907. The effect of temperature on the migration of the retinal

pigment in decapod crustaceans. Jour. Ex per. ZooL, 4 : 539-548.
CORTEGGIANI, E., AND A. SERFATY, 1939. Acetylcholine et cholinesterase chez les

Insectes et les Arachnides. Compt. Rend. Soc. Biol., 131: 1124-1126.
CROZIER, W. J., E. WOLF, AND G. ZERRAHN-WOLF, 1937. Critical illumination for

response to flickered light, with dragonfly larvae (Anax), in relation to

area of eye. Jour. Gen. Physiol., 21 : 223-247.
DEMOLL, R., 1911. Uber die Wanderung des Irispigments im Facettenauge. Zool

Jahrb. Abt. Allg. Zool. u. Physiol., 30: 169-180.
— , 1917. Die Sinnesorgane der Arthropoden, ihr Bau und ihre Funktion. 243

S. Braunschweig, 1917.
EPHRUSSI, B., AND G. W. BEADLE, 1936. A technique of transplantation for

Drosophila. Am. Nat., 70: 218-225.
EXNER, S., 1891. Die Physiologic der facettierten Augen von Krebsen und In-

sekten. 206 S. Leipzig u. Wien, 1891.
FRIZA, F., 1928. Zur Frage der Farbung und Zeichnung des facettierten Insek-

tenauges. Zeitschr. vergl. Physiol., 8 : 289-336.
HANSTROM, B., 1937. Vermischte Beobachtungen liber die chromatophoracktivier-

enden Substanzen der Augenstiele der Crustaceaen und des Kopfes der

Insekten. Kungl. Fysiogr. Sallsk. Handl. Lund, 47 : 1-10.

HORSTMANN, E., 1935. Die tagesperiodischen Pigmentwanderungen im Facet-
tenauge von Nachschmetterlingen. Biol. Zbl., 55: 93-97.
JAHN, T. J., AND F. CRESCITELLI, 1940. Diurnal changes in the electrical response

of the compound eye. Biol. Bull., 78 : 42-52.

KLEINHOLZ, L. H., 1936. Crustacean eye-stalk hormone and retinal pigment migra-
tion. Biol Bull, 70: 159-184.
— , 1938. Studies in the pigmentary system of Crustacea. IV. The unitary versus

the multiple hormone hypothesis of control. Biol. Bull, 75 : 510-532.
MERKER, E., 1929. Einfache Praktikumversuche zur Beobachtung der Pigment-

wanderung in den Augen von Tagfaltern und Dammerungsschmetter-

lingen. Biol. Zbl, 49: 186-191.
PARKER, G. H., 1932. The Movements of the Retinal Pigment. Ergcbn. Biol, 9:

239-291.
PEABODY, E. B., 1939. Development of the eye of the isopod, Idothea. Jour.

Morph., 64 : 519-554.

ROEDER, K. D., 1939. The action of certain drugs on the insect central nervous sys-
tem. Biol. Bull, 76: 183-189.

SNODGRASS, R. E., 1935. Principles of Insect Morphology. McGraw-Hill, N. Y.
TAYLOR, I. R., AND M. NICKERSON, 1940. Potential changes in the eye of the

moth, Galleria mellonella, during the course of dark adaptation. Anat. Rec.

Suppl. 78 : 92.
UCHIDA, H., 1934. Color changes in the eye of a long-horned grasshopper, Hotno-

rocoryphus lineosus, in relation to light. Jour. Fac. Sci. Univ. Tokyo, 3

(3) : 517-525.
UMBACH, W., 1934. Entwicklung und Bau des Komplexauges der Mehlmotte

Ephestia kuhniella Zeller nebst einigen Bemerkungen itber die Entstehung

der optischen Ganglein. Zeitschr. Morph. Okol Tiere, 28 : 561-594.

PIGMENT MIGRATION IN MOTH EYES 291

WELSH, J. H., 1930. The mechanics of migration of the distal pigment cells in the

eyes of Palaemonetes. Jour. Exper. Zoo/., 56 : 459-494.
— , 1935. Further evidence of a diurnal rhythm in the movement of pigment

cells in eyes of crustaceans. Biol. Bull., 68 : 247-252.
— 1938. Diurnal Rhythms. Quart. Rev. Biol., 13: 123-139.
— , 1939. The action of eye-stalk extracts on retinal pigment migration in the

crayfish, Cambarus bartoni. Biol. Bull, 77: 119-125.
— , 1941. The sinus glands and 24-hour cycles of retinal pigment migration in

the crayfish. Jour. Ex per. Zoo/., 86 : 35-50.
VON DER WENSE, T. F., 1938. Wirkungen und Vorkommen von Hormonen bei

wirbellosen Tieren. Leipzig, 1938.

WIGGLESWORTH, V. B., 1939. The Principles of Insect Physiology. Methuen,
London.

THE EFFECT OF TEMPERATURE ON THE RIGHTING

OF ECHINODERMS

XATHANIEL KLEITMAN
(From the Bermuda Biological Station, St. George's West, Bermuda}

The righting response of echinoderms has engaged the attention of
investigators since Romanes and Preyer first studied it in the eighties
of the last century, but the interest lay, in most cases, in analyzing the
activities of the nervous system, as they manifested themselves in the
reaction to a change in the position of the body. The representatives
of the phylum usually employed were the sea-urchin (Echinoidea), the
starfish or sea-star (Asteroidea), and the brittle-star (Ophiuroidea),
none of which possesses statocysts so important in the righting of higher
animals. The sea-urchin and the starfish turn over by the action of their
tube-feet, while the brittle-star, which has no tube-feet, depends entirely
on the muscular action of its arms.

Jennings ( 1907), in his extended report on the behavior of the starfish,
gave considerable space to the righting response, but did not touch upon
the time-element. Likewise, Hamilton (1922), in a paper devoted en-
tirely to the topic of righting in the starfish, dwelt only on the mechanism
of the process, but Fraenkel (1928) furnished some interesting figures
on the time it took certain starfish to right themselves. An earlier
paper containing time data is the one by Glaser (1907), who studied
movements of brittle-stars. Working on the sand-dollar, Parker (1927)
made some observations on the time relations of the various phases of
righting. His paper also contains numerous references to the literature
of body-righting in echinoderms. However, none of the authors men-
tioned attempted to relate righting-time to temperature, and in the chap-
ter on the echinoderms in Principles of Animal Behavior by Maier and
Schneirla (1935) there is no mention of a temperature factor in the dis-
cussion of righting. Barnes (1937), who lists a great number of bi-
ological processes for which temperature characteristics were obtained,
has nothing on the subject of body-righting, although he gives several
references on the effect of temperature on locomotion, of which, accord-
ing to Hamilton (1922), the righting of the starfish is one phase.

Methods and Results

In the present investigation recently collected echinoderms were kept
in aquaria, with the temperature of the water naturally varying from

292

TEMPERATURE AND RIGHTING IN ECHINODERMS 293

18° to 26° C., usually higher in the afternoon than in the morning.
The particular specimen to be observed was transferred to a very large
Petri dish, filled with sea water whose temperature was regulated by
the continual addition of chilled or warmed sea water and gentle
stirring, thus maintaining the selected temperature within one-half a
degree C. The animal was allowed to remain in the warmed or cooled
water for at least 15 minutes, and, judging by its behavior, it acquired
the temperature of the new medium during that period. The pro-
cedure followed was to lift the animal in the water and lay it in the
upside-down position, as symmetrically as possible, on the bottom of
the dish. • The righting was observed both from above and from the side.
In general, no righting response could be elicited below 10° C. and
above 30° C., and most of the observations were made at temperatures
varying from 14° to 26° C. As a rule, 10 to 20 trials were made at one
temperature level, then the water warmed or cooled, and another series
of tests made at the new temperature level. In some cases the righting-
times at half a dozen different temperatures were determined in succes-
sion, and the same animal observed again later in the day, or after an
interval of several days. The righting-time was measured by means of
a stop-watch, and in most cases the time required to turn through an
angle of 90°, as well as the total turning-over time, was noted.

Sea-urchins

These animals usually turned over in l%-2 minutes, with the greater
portion of the period (71-95 per cent) needed for the first 90°. Some-
times, after turning through an angle of 130-150°, the animal would
fall with a thud, apparently of its own weight. At low temperatures,
however, sea-urchins often failed to complete the righting, remaining at
an angle of 30-40° to the horizontal for a long time. On the other hand,
at high temperatures the animals were likely to remain in the dorsal
position, continuously executing translational or rotational movements
(about a vertical axis). At 15° C. it took, on the average, 3^2 minutes
for a sea-urchin to turn over, and the best performance was at 24—26° C.,
when the mean total righting time was about 80 seconds, with the first
half of the turn carried out in 66 seconds. The disparity between the
fractions of the total righting time taken to turn through the first, as
compared to the second 90° was greater at lower than at higher tem-
peratures.

There was no evidence of fatigue on subjecting the sea-urchins to
repeated righting. For example, mean figures in seconds for successive
series of ten trials in different specimens were 77 and 75, 83 and 93, 83
and 79. Figure la shows the relation between the reciprocal of the

294

NATHANIEL KLEITMAN

absolute temperature of the water and the logarithm of the speed of
righting (reciprocal of the righting time in seconds). Applying the
Van't Hoff-Arrhenius equation to the data upon which this figure is
based, a temperature characteristic of 19,000 calories was obtained.

Through the courtesy of Professor H. L. Clark, I was able to test
the effect of temperature on the righting of three spiny urchins. They

5T4R.F/5H

.358

.350

IOQ/TC48SS.)

FIG. 1. The relation of the speed of righting, expressed as the logarithm of
the reciprocal of the righting time in seconds, to the reciprocal of the absolute tem-
perature of the water: a, for the sea-urchin, Lytechiints varicgatus atlanticiis, with
one slope, and b, for the starfish, Stolasterias tcnuispina, showing a break in slope
at 22° C.

behaved in every way like the common sea-urchins, although they de-
pended for their turning on the movements of the large spines, rather
than those of tube-feet. Their performance was best at 26° C., when
they turned over, on the average, in less than 70 seconds.

Starfish

As already stated, the characteristic features of the righting response
of the starfish have been adequately described by Jennings, Hamilton,

TEMPERATURE AND RIGHTING IN ECHINODERMS

and others. Because several arms may participate in the turning process,
sometimes interfering with each other, it is much harder than in the
case of the sea-urchin to tell exactly when the starfish is half-turned.
Therefore a certain position of the upper pole of the animal was ac-
cepted as indicating the midpoint of the response. Furthermore, as the
arms often became entangled, the completion of the righting was also
judged by the position of the upper pole, rather than the complete spread
of the several arms.

At 6° C. the animals did not move at all, when in the upright position,
and one arm was usually characteristically twisted and curled. Between
8° and 10° C. they would begin to stir, but did not crawl. Above 11° C.
it was possible to obtain a righting response, first in 10-15 minutes, then
in less time as the temperature was raised. As in the case of the sea-
urchins, it took the starfish 66-93 per cent of the total righting time to
execute the first half of the turn, and also as with the sea-urchins, the
disparity between the fractions of time required to turn through the
first and second 90° was greater at lower than at higher temperatures.
The optimum temperature was 26° C., when the righting time was, on
the average, 27 seconds. This figure compares well with the figures of
Fraenkel (1928) for the righting time of " fast" starfish as 25 to 50
seconds, and of " slow " animals as one to three minutes.

Unlike the sea-urchins, the starfish showed evidence of some fatigue.
One animal, kept at 21° C., gave the following figures, in seconds, for
turning over, in two successive series of ten trials each : 41 to 60 and 54
to 122; another animal, at 23° C., showed a variation of from 55 to 89
seconds for the first ten trials, and 85 to 120 for the next. There was
also a greater day-to-day fluctuation in the righting time of a particular
starfish, at a certain temperature level, than there was in the sea-urchins.
This resulted in a greater scatter of temperature-righting-time data, as
plotted in Fig. Ib, but it is possible to discern two distinct trends in the
curve, with a break at 22° C., and a temperature characteristic of about
20,000 cal. at lower temperatures and one of 50,000 cal. at higher.

Brittle-stars

These animals made no attempts to turn over at temperatures below
10° C., although the arms would execute undulating movements, and
some could move short distances while in the dorsal position. At
14-15° C., they often required from 5 to 10 minutes to right themselves,
thus not differing greatly from sea-urchins and starfish. At higher
temperatures, however, they responded with great rapidity, their best
time being only 5-6 seconds. On the other hand, brittle stars tired very
quickly, first showing a marked lengthening of the righting time, then

296 NATHANIEL KLEITMAX

failing to respond altogether. For example, mean figures in seconds for
successive series of ten trials on one animal at 22° C. were: 13, 22, 29,
and in the fourth series only two responses of 27 and 53 were obtained,
before the animal stopped responding to being placed on its back.
Similar results were reported for brittle-stars by Glaser (1907), with
the best righting times of only 3-4 seconds and rapid fatigue in some,
though not in all animals tested.

The variability of the data as well as the fatigability of the animals
resulted in such a scatter of the temperature-righting-time figures as to
make it impossible to calculate a temperature characteristic of the
process.

Discussion

The application of the Arrhenius equation to the relationship between
the rate of biological processes and environmental temperatures has been
confined mainly to such activities as enzyme action, oxygen consumption,
carbon dioxide assimilation, embryological development, or to such
organ and tissue performances as breathing, heart rate, nerve and muscle
physiology. The only neuromuscular activities involving the organism
as a whole to which this equation has been applied were locomotion, as
cited by Barnes (1937), and reaction time of the human subject, as re-
ported by Kleitman, Titelbaum and Feiveson (1938). In this investi-
gation it was found that such a global process as body-righting is also
subject to the effect of temperature. The echinoderms studied all failed
to right themselves below 10-11° C., and at the lowest effective tempera-
tures all took from 5 to 15 minutes to turn over. Although the righting
time progressively decreased in all with a rise in temperature and the
optimum performance attained in all at the level of 24-26° C., the
shortest righting time was quite different for each of the three classes
of echinoderms. It was longest (80 seconds) for the sea-urchin in
which the process was 'least complicated, depending as it did on the ac-
tion of successive groups of tube-feet, brought into play as the animal
was turning over. It was shorter for the starfish (27 seconds), which,
although it essentially depended on tube-feet action, had to follow the
initiative of one or more of its arms, sometimes several arms working
against each other in attempting to turn the animal in opposite directions.
It was shortest (5—6 seconds) for the brittle-star that performed the
righting by muscular action entirely.

On repeated testing the slowest of the three, the sea-urchin, showed
practically no fatigue; the fastest, the brittle-star, tired very quickly;
while the starfish, in this respect, too, occupied a middle position.
Whether the ultimate failure to turn over was due to fatigue of the
receptor or central nervous mechanism has not been established.

TEMPERATURE AND RIGHTING IN ECHINODERMS

Concerning the temperature characteristics, it will be recalled that
the righting of the sea-urchin had one p. value of 19,000, while that of
the starfish had two : 20,000 below 22° C. and 50,000 above that tem-
perature. Barnes (1937) states that when there is a break in the slope
of the rectilinear relation between the logarithm of the rate of activity
and the reciprocal of the absolute temperature, the /x values in the
higher temperature range are usually smaller than those pertaining to
lower temperatures. In the righting of the starfish the reverse was true,
but there have been reports of many other biological processes with a
greater /*, value at higher temperatures, among them such a global
activity as locomotion of the ant, studied by Barnes and Kohn (1932).
Although it was impossible to obtain a definite ju value for the righting
of the brittle-star, the available data suggest a high /A value at the upper
temperature levels.

It may be added that, in taking 66-93 per cent of the total righting
time to turn through the first 90°, the sea-urchins and starfish behaved
like the sand-dollars studied by Parker (1927), who found that " the lift
from the horizontal to the vertical requires as much as 3 hrs. ; the drop
from the vertical to the horizontal about half an hour." The similarity
is particularly striking in that the sea-urchins and starfish turned in
water and their total righting times were expressed in minutes or sec-
onds, while the sand-dollars partly buried themselves in sand and took
several hours to turn over.

Summary

The speed of a global activity of three echinoderms, in the form of
body-righting, is related to their temperature, within the physiological
limits of 10° to 30° C. The speed increases as the temperature rises,
and optimum performance is obtained at 24—26° C.

ACKNOWLEDGMENTS

I wish to thank Professor H. L. Clark of the Harvard Museum of Comparative
Zoology for advice and assistance, and for furnishing me with the names of the
animals studied, which were as follows: common Bermudian sea-urchin — Lythe-
chinus variegatus atlanticus (A. Agassiz) ; stout-spined sea-urchin — Eucidaris tribu-
loldes (Lamarck) ; starfish or sea-star — Stolasterias tenuispina (Lamarck) ; and
brittle-star — Ophiocoma echinata (Lamarck).

It is a pleasure to express my gratitude to Dr. J. F. G. Wheeler, the Director
of the Station, who placed its facilities at my disposal.

BIBLIOGRAPHY

BARNES, T. C., 1937. Text-book of General Physiology, Philadelphia. Chapter
XIII and Appendix.

298 NATHANIEL KLEITMAX

BARNES, T. C, AND H. I. KOHN, 1932. The effect of temperature on the leg pos-
ture and speed of creeping in the ant Lasius. Biol. Bull., 62: 306-312.

FRAENKEL, G., 1928. Uber den Auslcisungsreiz des Umdrelireflexes bei Seesternen
und Schlangsternen. Zcitschr. vcrgl. Physiol., 7 : 365-378.

GLASER. O. C., 1907. Movement and problem solving in Ophiura. Jour. E.rpcr.
' Zoo}., 4 : 203-220.

HAMILTON, W. F., 1922. Coordination in the starfish. III. The righting reaction
as a phase of locomotion. Jour. Comfy. Psycho!., 2 : 81-94.

JENNINGS, H. S., 1907. Behavior of the starfish, Asterias forreri de Loriol (right-
ing and locomotion). Univ. of Calif. Publ. Zoo/., 4: 53-185.

KLEITMAN, N., S. TITELBAUM, AND P. FEIVESON, 1938. The effect of body tem-
perature on reaction time. Am. Jour. Physiol., 121 : 495-501.

MAIER, N. R. F., AND T. C. SCHNEIRLA, 1935. Principles of Animal Behavior, New
York. Chapter III.

PARKER, G. H.( 1927. Locomotion and righting movements in echinoderms, espe-
cially Echinarachnius. Am. Jour. Psychol.. 39: 167-180.

THE EFFECT OF PHYSOSTIGMINE ON THE RESPONSES
OF EARTHWORM BODY WALL PREPARA-
TIONS TO SUCCESSIVE STIMULI

E. FRANCES BOTSFORD

(From the Department of Zoology, Connecticut College, and the Alarinc Biological

Laboratory, Woods Hole, Mass.)

INTRODUCTION
The experiments reported in this paper were performed in an at-

tempt to analyze certain augmentation phenomena in the responses of
the body wall of the earthworm, Lumbricus terrestris, and to provide
some basis for an opinion as to their cause.

One type of augmentation concerned in these investigations is that
of summation of contraction. When a second stimulus follows the first
before the response to the first is completed, the second contraction is
superimposed upon the first to produce a greater muscular response.
With an increase in the frequency and number of stimuli a condition of
tetanus is produced.

Another type of augmentation is shown by the muscles of the body
wall when a second stimulation follows the first after the response to
the first is completed. In a series of stimulations each successive re-
sponse is greater, resulting in a " staircase " effect.

The various augmentation phenomena which are shown by striated,
smooth, and cardiac muscle of vertebrates have been studied extensively
by many workers. Among the invertebrates, the field is unexplored ex-
cept in coelenterates, echinoderms, mollusks, and crustaceans. Because
of the diverse neuromuscular mechanisms involved in these different
types of invertebrates, the augmentation phenomena themselves present
varying characteristics, some being comparable to those found in verte-
brates, while others are peculiar to a particular invertebrate group.

The first evidence for summation in the body wall muscles of the
earthworm was given by Budington (1902), whose records showed an
increase in response corresponding to an increase in the number of
shocks administered. After Budington there is no other mention of
augmentation phenomena in the earthworm until the studies of Bacq
and Coppee (1937), who included three experiments on the earthworm

299

300 E. FRANCES BOTSFORD

in their work on Sipunculus and the leech. They found that physostig-
mine increased the muscular response when the nerve cord of the earth-
worm was stimulated repetitively.

The purpose of this paper is to show, first, how the muscular re-
sponses of the earthworm body wall are affected by variations in the
frequency and number of single shocks and by the spacing and duration
of tetanic stimulations ; second, that the ability of the muscle to give
augmented responses depends upon these time factors ; and third, that
the effect of physostigmine upon the muscle responses suggests the par-
ticipation of acetylcholine as a facilitating factor.

An investigation of the muscular responses of the earthworm body
wall is of especial interest since this muscle seems to parallel vertebrate
striated muscle in some of its physiological characteristics. Certain of
these similarities have been referred to by Pantin (1935b) and by Wu
(1939).

MATERIALS AND METHODS

The specimens of Lmnbricus terrcsins used for these experiments
were kept in an ice-box in moist earth and only the large and healthy
specimens employed.

After partially anesthetizing the earthworm in 0.2 per cent chlore-
tone, a mid-ventral slit was made the length of the worm and the nerve
cord and digestive system were removed. This preparation of the body
wall will be referred to as the muscle strip. At one end it was pinned
to a paraffin block and at the other it was attached to a lever of spring
steel, which recorded the contractions of the longitudinal muscles on a
kymograph drum by a downward deflection of the lever. The approxi-
mate magnification of the lever was five times.

Another preparation used was the whole worm minus about the first
ten segments, arranged for recording muscle contractions in the same
way as described for the muscle strip. This will be referred to as the
whole worm preparation.

For stimulating both the muscle strip and the whole worm, a fine
silver wire electrode leading from a vacuum tube stimulator was in-
serted in each end of the preparation. The stimulator employed a gas
triode 885 arranged to deliver stimuli at frequencies from 1 to 100 per
second. The duration of the bursts of stimuli and the interval between
the bursts were controlled by a commutator in the circuit. Submaximal
stimuli were used in all the experiments described in this paper unless
indicated otherwise.

AUGMENTATION PHENOMENA IN THE EARTHWORM 301

RESULTS
The Response to Successive Single Shocks

When the body wall preparations were stimulated electrically by a
series of submaximal single shocks, the type of response was found to
be affected by the frequency of the shocks.

With a frequency of about 2 per second the second response showed
an increased contraction, but after that there was no further augmenta-
tion. (See Fig. 1, B.) With an increase in frequency there was a suc-
cessive increase in the magnitude of the first four contractions. A small
amount of tonus developed which persisted for a short time after the
stimulation stopped. (See Fig. 1, A.} Increasing the frequency to 6

FIG. 1. Effect of frequency of single shocks on augmentation of contractions.
Whole worm preparation. Frequency in shocks per second: A, 3 per second; B,
2 per second ; C ' , 6 per second.

per second caused an augmentation in the successive responses up to the
ninth, which was nine times greater than the first response. This was
followed by an increase in tonus and a diminution in the magnitude of
the individual responses. (See Fig. 1, C.) The record shows a defi-
nite " staircase " effect. Strictly speaking, the term " staircase " should
be reserved for the increased muscular responses brought about with
maximal stimuli, thus showing that the heightened responses are due to
increased contractions of the individual contractile units involved. Since
maximal stimuli were not used in these experiments, the descriptive term
of " augmentation of responses " is employed, and the determination
of the exact mechanism of the facilitating effect is left for future
investigation.

With a frequency of 10 per second a state of increased tonus is pro-
duced immediately because of the summation of the successive con-

302

E. FRANCES BOTSFORD

tractions, and with a frequency of about 14 per second, the response
shows a condition of completely sustained contraction or tetanus.

Summation of the Responses to a Series of Shocks

This muscle preparation gives a response to a single shock and so
can be called a single volley muscle. Records were made on a stationary
drum of the responses of the whole worm preparation to one, two, and
three single shocks delivered within a 0.3-second period. The response
to two shocks was about one and one-half times greater than it was to
one, and the response to three shocks was over twice as great.

0 1

8 9

FIG. 2. A. Frequency-response curves for muscle strip. Abscissae, number of
shocks. Ordinates, response in mm. Duration of bursts, 0.5 second. Interval be-
tween bursts, 1 minute. First series shown by dotted line. (Second series omitted.)
Third series, unbroken line. Fourth series, broken line. There is a % hr. interval
between series 3 and 4, a 10-minute interval between the other series.

B. Duration-response curve for whole worm preparation. Abscissae, duration
of burst. Ordinates, response in mm. Frequency, 14 shocks per second.

In order to show the effect of frequency upon the magnitude of
contraction, a series of responses to a single burst of stimuli was re-
corded. The duration of every burst was constant, but the frequency
of the stimuli within each burst was varied at random over a relatively
wide range. A long interval was allowed between the bursts in order
to prevent an effect of previous activity upon the response. The re-
sponses were recorded on a stationary drum and the length of each
measured in millimeters. From the results obtained, frequency-response
curves were made. Figure 2, A shows three of these curves for one
muscle strip. »

AUGMENTATION PHENOMENA IN THE EARTHWORM 303

The magnitude of the responses increases with each increase in the
number of shocks. The increase throughout the first series of trials
is nearly in direct proportion to the number of shocks. In the third
series the increase in responses is proportional to the number of shocks
except with the highest frequency where there is a marked decline in the
amount of augmentation. The fourth series was recorded three-quarters
of an hour after the third series and about two hours from the be-
ginning of the experiment. Although with the lower frequencies the
muscle gives greater contractions than before, with an increase in the
number of shocks the amount of augmentation declines, so that with the
higher frequencies the responses are lower than in the other two series.
This decline is probably due to the deterioration of the preparation.

These results show that the augmentation is proportionally less with
the higher frequencies. There is an indication of a slow cumulative
building-up process, since the responses of the second series were greater
than the first, and those of the third the largest of all.

In one experiment with the whole worm preparation in which the
duration of the bursts was gradually increased while the frequency was
kept constant, the magnitude of the response to a burst was increasingly
greater from the first response measuring 6 mm. up to the tenth measur-
ing 76 mm. A duration-response curve was made by plotting the mag-
nitude of the responses against the duration of each burst. (See Fig.
2, B}. Experiments in which the duration of the bursts was changed
at random showed the same effect of increased contractions due to
longer bursts.

These results show that the magnitude of a response to one isolated
burst of stimuli is affected by the frequency of the stimuli within the
burst and the duration of the burst.

The Response to Repeated Bursts of Stimuli

When the earthworm body wall is stimulated electrically with shocks
at a constant frequency delivered in repeated bursts at appropriate in-
tervals, the first contractions show a definite increase in each successive
response. When this augmentation ceases, it is not followed by a
plateau, but by an immediate but gradual decline in the magnitude of the
successive responses. The rate of decline varies under different con-
ditions of stimulation.

Figure 3 shows the first part of a normal curve for the whole worm
preparation. Since the frequency of stimuli within the bursts was 28
per second, these responses were tetanic in character. The first 22
records of contraction show an increase in the response to each suc-
cessive burst. Figure 4 shows the same phenomenon in a muscle strip

304

E. FRANCES BOTSFORD

when the same frequency is used. A longer interval was allowed be-
tween bursts in the case of the muscle strip preparation because the
muscle strip required a longer period to recover its original state of
tension after each response.

To investigate the characteristics of the " staircase " effect exhibited
by these muscles, many series of responses were recorded showing the

FIG. 3. Augmentation of contractions in whole worm preparation. Duration
of burst, 0.3 second. Interval between bursts, 7 seconds. Frequency, 28 shocks per
second.

FIG. 4. Augmentation of contractions in muscle strip. Duration of burst, 0.4
second. Interval between bursts, 27 seconds. Frequency, 28 shocks per second.
Drum stationary.

FIG. 5. Effect of frequency of stimuli within burst on augmentation of con-
tractions. Whole worm preparation. Duration of burst, 0.3 second. Interval be-
tween bursts, 7 seconds. Frequency of stimuli: A, E, G, 28 shocks per second
causes augmentation; B, 9 shocks per second; C, 14 shocks per second; D, F, 18
shocks per second.

effect upon the augmentation phenomenon of four easily variable condi-
tions : the intensity of the stimulating current, the duration of the bursts
of stimuli, the interval between the bursts, and the frequency of the
stimuli within the bursts. It should be noted that in some instances it
is impossible to make exact quantitative statements which apply to all

AUGMENTATION PHENOMENA IN THE EARTHWORM 305

the preparations, since the differing physiological states of individual
worms produced variation in response.

Difficulties were encountered in working with maximum intensities.
In the case of muscle strips, a single burst of stimuli of high intensity
induced a condition of tonus which was prolonged to such an extent
that successive responses could not be elicited. With the whole worm
preparations, high intensities often brought about strong spontaneous
contractions which make it impossible to continue with the experiment.
In the few successful experiments with high intensities there was no
augmentation of the successive responses. This is a crucial point which
should be investigated more thoroughly by further experimentation.
Because of the disadvantages presented by the use of high intensities,
submaximal stimuli of uniform intensity were used in the following
experiments.

The length of the interval between bursts has a very definite effect
upon the production of augmented responses. Using a frequency of 28
per second with the duration of burst of 0.3 second, there is a striking
increase in successive responses in the whole worm preparation when
the interval between bursts is 7 seconds. On doubling this interval,
there is still some augmentation, but as the interval is increased still more
this is less evident, until, with a 30-second interval, the successive re-
sponses show no increase.

The duration of the bursts also affects the production of augmented
responses. In one experiment in which a frequency of 18 per second
was used with bursts spaced at 14-second intervals, the responses were
not facilitated when the bursts lasted 0.4 second. Lengthening the
bursts to 1.5 seconds built up increasing contractions through six suc-
cessive responses, the sixth one being more than twice as great as the
first.

The frequency of the stimuli within the bursts is a third factor in
determining the production of augmented responses. A low frequency
does not bring about an increase in the successive responses. Figure 5
shows a long series of responses of the whole worm preparation to
bursts of stimuli. The evident arrangement in groups is due to the
different frequencies of the stimuli. A frequency of 18 per second or
less did not cause augmentation, but with a frequency of 28 per second,
three of the groups in the series show facilitated responses. In other
preparations, however, increasing contractility was sometimes brought
about with lower frequencies.

From the results of these experiments and many others, it is evident
that the phenomenon of augmentation is to a great extent dependent
on the number of stimuli applied in a unit of time and the grouping of

306

E. FRANCES BOTSFORD

these stimuli. As the interval between bursts is lengthened, either the
frequency of the stimuli within the bursts must be increased or the
duration of the bursts must be lengthened in order to produce a series
of increasing responses.

The Effect of Physostigmine upon tlie Production of
Augmented Responses

In order to determine the underlying cause of the facilitation phe-
nomenon which is manifested in the augmented contractions of the earth-

FIG. 6. Effect of physostigmine 1 gm./10,000 cc. on the response of whole
worm preparation to a series of single shocks. Ten successive series of shocks each
consisting of 7 single shocks in 5 seconds separated from one another by an interval
of 3 minutes. Series A, Ringer's drip. Series B, physostigmine drip.

worm body wall muscles, application of physostigmine was tried. Wu
(1939) has shown that the sensitivity of the body wall to acetylcholine
is greatly increased by physostigmine. From the results of his experi-
ments concerning the action of drugs on the earthworm body wall, he
postulated the presence of some factor which prevented the action of

AUGMENTATION PHENOMENA IN THE EARTHWORM 307

acetylcholine and which was antagonized by physostigmine. This he
thought was probably a high concentration of choline esterase.

If acetylcholine is produced by stimulation of the earthworm muscle
preparation, and if this is not completely hydrolyzed by choline esterase
before the next stimulation, the persisting acetylcholine could be the
cause of the augmentation of the responses. The application of phy-
sostigmine which prevents the action of choline esterase could, therefore,
show some effect upon the augmentation phenomenon. The following
experiments were devised to test this hypothesis.

FIG. 7. Effect of physostigmine, 1 gm./10,000 cc., on the responses of muscle
strip to a series of bursts of stimuli. Interval between bursts, 27 seconds. Fre-
quency, 40 shocks per second. Drum stationary. A, physostigmine applied by drip
method during experiment. B, muscle strip immersed in physostigmine for 15 min-
utes previous to stimulation.

The effect of physostigmine on the responses to a series of single
shocks was to produce a striking augmentation. The whole worm
preparation was stimulated for 5 seconds at a low frequency of 7 shocks
in 5 seconds. This was followed by a 3-minute period of rest during
which the preparation was given Ringer's solution by drip method.
Figure 6, A shows a series of five of these 5-second stimulations. Within
each of the five groups there is an increase in the successive responses,
but the facilitation decays during the 3-minute rest period, so that the
first response in each of the five groups is of the same magnitude.
Figure 6, B shows the result of repeating this procedure except that
physostigmine drip 1 gm./10,000 cc. was substituted for the Ringer's

ARY :

308 E. FRANCES BOTSFORD

during the 3-minute rest period. Here the first response in each group
is increasingly greater. The succeeding responses in one group are
smaller than the first response of that group but greater than the normal
responses shown in Fig. 6, A. A. state of maintained tension develops
during each 5-second burst of stimuli. The optimum action of the
physostigmine, judged by magnitude of response, was reached in 12%
minutes, as shown in the fourth series, in which the maximum response
is 8% times greater than the first response in the normal preparation.
The fifth series shows no further increase in magnitude of response.

In order to show the effect of physostigmine upon a series of suc-
cessive tetanic responses, physostigmine 1 gm./10,000 cc. was applied
to a muscle strip preparation continuously by the drip method during
a long series of bursts of stimuli. Figure 7, A shows the result. A
strong tonus was built up gradually in the first 18 responses, but even
with this increasing tonus there was an augmentation of the successive
individual contractions. On the other hand, when the preparation was

TABLE I

Effect of physostigmine on the magnitude of the responses. Two muscle strips
were used: A, with physostigmine drip 1 gm./10,000 cc.; B, with Ringer's drip.
Duration of bursts, 0.3 second. Interval between bursts, 3 minutes. Frequency,
40 shocks per second. The magnitude of response is recorded in the body of the
table in mm.

Successive responses

A. Physostigmine

B. Ringer's

put into a bath of physostigmine for 15 minutes before the stimulations
began, the first response was of normal magnitude, but following this
response an immediate tonus was manifested. The second response
was only three- fourths the magnitude of the initial one, and the suc-
ceeding responses showed a fatigue-like diminution. The long applica-
tion of physostigmine prevented any augmentation of contractions.
These characteristics can be seen in Fig. 7, B. Several control experi-
ments, in which the preparation was left in a bath of Ringer's solution
for 15 minutes previous to stimulation, showed a normal response.

Physostigmine was demonstrated to increase the length of the interval
between bursts of stimuli which is necessary for the production of aug-
mented responses. As has been stated above, there is a maximum in-
terval of about 30 seconds, which, if exceeded, does not allow augmen-
tation. When a normal muscle preparation is stimulated by bursts
which are separated by 3-minute intervals, there is no increase in the
successive responses. Table I shows the magnitude of the responses as

AUGMENTATION PHENOMENA IN THE EARTHWORM 309

measured by the length of the record in millimeters for two muscle
strips, one treated with physostigmine drip 1 gm./10,000 cc., the other
with Ringer's. When physostigmine drip was used on the preparation,
there was an increase in the magnitude of 8 successive responses even
though these were separated by 3-minute intervals.

Figure 8 demonstrates this effect in a single muscle strip stimulated
in the same way. The first five responses are separated by 3-minute
intervals of treatment with Ringer's drip. There is no increase in the
successive responses. After the fifth response, treatment with physo-
stigmine drip, 1 gm. /10,000 cc., is begun during the 3-minute interval
with the result that augmentation is produced as well as an increase in
tonus.

The experiments with physostigmine described above suggest that at
each burst of stimuli acetylcholine was formed and that the physostig-
mine acted upon the choline esterase to delay the breakdown of this
acetylcholine. This resulted in the persistence of a certain quantity of
acetylcholine which caused an increase in the next response of the
muscle.

It will be noted that a relatively high concentration of physostigmine
was used in these experiments. This was in order to favor the diffusion
of sufficient drug into the tissues to produce an immediate effect even
with the slow drip method employed.

Further indication of the formation of some facilitating substance at
the time of stimulation is shown by conditions of tonus succeeding the
muscular responses. Figure 9 shows a response in which the primary
contraction and partial relaxation due to the single shock is followed
by a smaller and slower contraction and relaxation. In Figure 10 is
recorded the response to a series of 7 single shocks after the preparation
had been in physostigmine 1 gm./ 100,000 cc. for 4 minutes. After the
seventh response, the stimulation having stopped, there is a long slow
contraction followed by relaxation.

The persistence of a facilitating effect is demonstrated in Fig. 11.
Here a series of augmented responses was produced using a frequency of
28 per second. This was followed, with no break in the intervals of
stimulation, by a series of bursts with a frequency of 9 per second, which
is not a facilitating frequency. The first four responses to the low fre-
quency bursts show a greater response than those which follow. The
preceding series of augmented responses had a facilitating effect which
lasted 29 seconds. This corresponds to the interval of 30 seconds deter-
mined to be the maximum interval within which facilitation can occur.
The response to a single shock is also increased when immediately pre-
ceded by a series of augmented responses.

310

E. FRANCES BOTSFORD

FIG. 8. Effect of physostigmine 1 gm./ 10,000 cc. in producing augmentation
of contraction in a muscle strip with an interval of 3 minutes between bursts of
stimuli. A' represents physostigmine drip begun. Duration of bursts, 0.3 second.
Frequency, 40 shocks per second.

FIG. 9. Response of muscle strip to a single shock followed by a change in
tonus.

FIG. 10. Responses of muscle strip to a series of 7 single shocks after 4
minutes in physostigmine 1 gm./ 100,000 cc. Following the responses there is a
change in tonus.

FIG. 11. Increased magnitude of the initial responses to bursts of low fre-
quency when preceded by a series of successively augmented responses. Whole
worm preparation. Frequency of stimuli: A, 28 shocks per second; B, 9 shocks
per second. Duration of burst, 0.3 second. Interval between bursts, 7 seconds.

AUGMENTATION PHENOMENA IN THE EARTHWORM 311

DISCUSSION

From the results of the experiments described it is evident that pre-
vious activity has a definite facilitating effect upon subsequent contrac-
tions of the longitudinal muscles in the body wall of the earthworm.

When preparations are stimulated with successive single shocks, three
types of responses are possible, depending upon the frequency of the
shocks : first, an increase in magnitude of successive separate contrac-
tions; second, a partial summation of the successive contractions; and
third, complete tetanus. Since, with increase in frequency of shocks,
the first type of response merges gradually into the second, and the
second into the third, it would appear that the same facilitating factor is
responsible for all. This is also indicated by the fact that with the ap-
plication of physostigmine it is possible to produce a summation of con-
tractions or tetanus with the lower frequencies.

In his work on Actinozoa, Pantin (1935o) found that in different
neuromuscular mechanisms he was able to demonstrate these three types
of responses depending upon the time relation between the duration of
a complete contraction and relaxation of the muscle, and the duration of
the persistence of the facilitating factor.

In the earthworm preparations the facilitating factor persists between
0.4 and 0.5 second after a single shock. With repeated shocks the
facilitation effect is cumulative, resulting in considerably heightened con-
tractions, as exhibited by the responses to repeated tetanic stimulations.
Under these conditions the facilitating factor may last nearly 30 seconds,
as shown by an increased response to a second tetanic stimulation within
that period of time.

This phenomenon of facilitation has been demonstrated in the mus-
cular reactions of other invertebrates. The number and frequency of
stimuli are the decisive factors in producing the augmented responses,
as is the case in the earthworm, but the cause of the facilitation seems
to be different in the various neuromuscular mechanisms.

Pantin (1935a), in his work on facilitation in Actinozoa, came to the
conclusion that in certain muscles the increasing magnitude of the re-
sponses in a series of stimuli is due to neuromuscular facilitation, by
which, with each succeeding stimulus more muscle fibers are affected.
He found no evidence at that time for the functioning of chemical
mediators in coelenterates. Ross and Pantin (1940), in their investiga-
tion of the effect of certain ions on facilitation in Actinozoa, found that
two factors were involved in facilitation. Although they did not de-
termine the nature of the facilitating process they concluded that it could
not be due solely to a transmitter.

312 E. FRANCES BOTSFORD

In crustacean striated muscle, according to Katz (1936), the fre-
quency to which individual muscle fibers respond varies, so that the
number of fibers which contract and consequently the magnitude of the
contraction, are controlled by the frequency of the nerve impulses.

In the present paper evidence has been offered to support the view
that in the earthworm body wall the facilitating factor is acetylcholine.
The effect of physostigmine in producing greatly augmented responses
points to this. The most convincing evidence is the ability of physostig-
mine to delay the decay of the facilitating factor, so that augmented re-
sponses are elicited even with long intervals between tetanic stimulations.

The location of this facilitating effect is a subject for further investi-
gation. Since the phenomenon showed the same characteristics in the
muscle strip as in the whole worm preparation, it cannot be dependent
on the nerve cord.

SUMMARY

1. The longitudinal muscles of the body wall of the earthworm (Lmn-
bricus terrestris) show augmented responses when stimulated by succes-
sive single shocks at low frequencies. The facilitating condition lasts
not more than 0.5 second after the response to a single shock.

2. A frequency of 14 per second results in a complete summation of
the contractions or a condition of tetanus.

3. The magnitude of a summated response elicited by a series of
shocks is proportional to the duration of the burst of shocks and to the
frequency of the shocks within the burst.

4. The tetanic responses to repeated bursts of stimuli show an in-
creasing augmentation of the initial contractions. The production of
this " staircase " effect is affected by the frequency of the shocks, the
duration of the bursts, and the length of the interval between bursts.

5. After a brief tetanus the facilitating condition persists for nearly
30 seconds.

6. The application of physostigmine increases the augmentation of
responses and tends to produce a condition of tonus.

7. Physostigmine delays the decay of the facilitating property so that
augmented responses are produced with intervals as long as 3 minutes
between bursts of stimuli, suggesting the role of acetylcholine in the
production of the augmented responses.

ACKNOWLEDGMENTS

The author wishes to express appreciation for the helpful advice of Drs. C. L.
Prosser and J. H. Welsh, and for the research grants from Connecticut College
during the summers of 1938, 1939 and 1940 which made the present experiments
possible.

AUGMENTATION PHENOMENA IN THE EARTHWORM 313

REFERENCES

BACQ, Z. M., AND G. COPPEE, 1937. Reaction des vers et des mollusques a 1'eserine.

Existence de nerfs cholinergiques chez les vers. Arch. Internal, dc Physiol.,

45 : 310-324.
BUDINGTON, R. A., 1902. Some physiological characteristics of annelid muscle.

Am. Jour. Physio]., 7: 155-179.
KATZ, B., 1936. Neuro-muscular transmission in crabs. Jour. Ph\siol., 87: 199-

221.
PANTIN, C. F. A., 1935a. The nerve net of the Actinozoa. IV. Facilitation and

the " staircase." Jour. Ex per. Biol, 12 : 389-396.
PANTIN, C. F. A., 19356. Response of the leech to acetylcholine. Nature, 135 :

875.

Ross, D. M., AND C. F. A. PANTIN, 1940. Factors influencing facilitation in Ac-
tinozoa. The action of certain ions. Jour. Expcr. Biol., 17 : 61-73.
Wu, K. S., 1939. The action of drugs, especially acetylcholine on the annelid body

wall (Lumbricus, Arenicola). Jour. Expcr. Biol., 16: 251-257.

THE RELATION BETWEEN THE FOUR-CARBON ACIDS
AND THE GROWTH OF OAT SEEDLINGS

HARRY G. ALBAUM AND BARRY COMMONER

(From the Department of Biology, Brooklyn College, the Department of Biology,
Queens College, and the Marine Biological Laboratory, Woods Hole, Mass.)

INTRODUCTION

Plant growth hormones are noted for the multiplicity of their effects.
This is particularly true of the auxins, which are known to influence
elongation of roots and shoots, determination of root number, produc-
tion of callus tissue, tropistic responses, correlative growth of plant
organs and protoplasmic streaming.

The most striking feature about the variety of these auxin effects is
the fact that any one concentration of the hormone will often stimulate
one process, inhibit another, and have no effect on a third. If the con-
centration is varied these relations may be completely altered. Such
effects, together with the fact that the auxins have an effect on proto-
plasmic streaming, have suggested that these hormones must play a
decisive role in some fundamental and common intracellular process.

For some time attempts had been made to demonstrate a link between
auxin and cell respiration. Experiments showed, however, that auxin
had no demonstrable effect on the respiratory rate of plant cells (Bonner,
1936; Van Hulssen, 1936) and the conclusion was drawn that the
hormone had no respiratory effect (DuBuy, 1940).

More recently, however, Commoner and Thimann (1941) have
shown that indole-3-acetic acid is directly related to the activity of a
specific respiratory process, the 4-carbon dicarboxy'lic acid system.
They showed that the salts of these acids (malic, fumaric, and succinic)
will enhance the effect of auxin on growth and that the inhibitory effect
of iodoacetic acid on growth is due to a specific inhibition of this system.
It was shown that the 4-carbon acid system normally accounted for only
5-10 per cent of the total respiration, but that all the growth is dependent
upon its activity. Thus it was demonstrated that auxin can stimulate
the respiratory activity of this system (and in this way increase the rate
of oxygen consumption) and that the effect of the 4-carbon acids on
respiration is enhanced by the presence of auxin.

This work dealt with a single effect of auxin, the elongation of ex-

314

FOUR-CARBON ACIDS AND GROWTH OF OAT SEEDLINGS 315

cised sections of the Arena coleoptile, but indicated that the effect of
the hormone is exerted through a common cellular oxidative process.

The work suggested that this effect may play an important role in
determining the nature and direction of the influence of auxin on the
other processes mentioned above. This indication was also supported by
the findings of Sweeney (unpublished data) that the 4-carbon acids
influence significantly the auxin effect on protoplasmic streaming in the
Avena coleoptile.

It was our purpose therefore to extend this evidence by examining
the effect of one of the 4-carbon acids on three different growth processes
known to be influenced by auxin in the intact Avcna seedling: shoot
elongation, root elongation, and determination of root number.

PROCEDURE

Seeds of Avena sativa L. var. Black Norway and Fulghum similar to
those used in earlier experiments were hulled and soaked in distilled
water for 20 hours at room temperature with continuous aeration. After
soaking, the oats were placed in beakers lined with moist filter paper, as
described in an earlier paper (Kaiser and Albaum, 1939), and to which
the test solutions had been added. The plants were then allowed to
continue their growth in the dark at room temperature. At various
times the coleoptiles were measured to the nearest millimeter under an
orange safelight. Growth of the first internode appeared to be com-
pletely inhibited ; no effort was made to measure this separately. At the
close of each experiment (i.e., when the coleoptile had ceased growing),
final measurements were taken on total root length, total root number
and final coleoptile length. All figures reported here are the averages
of at least twenty plants. The test solutions, indole-3-acetic acid
(Merck), iodoacetic acid (Eastman Kodak) and fumaric acid (Eastman
Kodak) were made up in distilled water and adjusted to pH 6.0 with
KOH at the beginning of each experiment. No change in pH occurred
during the course of the experiment. Measurements on the length of
epidermal cells wrere carried out at the close of the experiments by
stripping the epidermis from small pieces of coleoptile, mounting it in
water on a slide and measuring it with an ocular micrometer under the
low powers of a compound microscope.

RESULTS
The Effect of lodoacctatc on Coleoptile Groivth

These experiments indicate that the 4-carbon acid system influences
the effect of auxin on seedling growth in a manner similar to its effect
on the growth of isolated coleoptile sections.

6Q
50.

i 40J

- 30.
20.
10
0

OOOOI M I * OIMF

OOOOI M I t O05M F
CONTROL

OOOOI M I + OOI M F
OOOOI M I + OOOI M F

20 40

'60 80 100 120
TIME-IN-HOURS

140 160

80 100 120 140 160

TIME IN HOURS

FIG. 1. Growth curves of coleoptiles following poisoning with iodoacetate,
and recovery after addition of various concentrations of fumarate.

FIG. 2. Growth curves of coleoptiles after poisoning with various concentra-
tions of iodoacetate.

FOUR-CARBON ACIDS AND GROWTH OF OAT SEEDLINGS 317

Figure 1 shows that 10 r> M iodoacetate reduces the final size of the
coleoptile, although the growth curve characteristics are not affected
(the half-times of all curves in Fig. 1 are identical). If fumarate is
added the inhibitory effect of iodoacetate is negated ; in fact, use of a
concentration of .01 M fumarate results in a coleoptile even exceeding
in size that of the control.

Figure 2 demonstrates the effect of various concentrations of iodo-
acetate on coleoptile growth. Maximum inhibiton (50 per cent) is ob-
tained at a concentration of 5 X 10~5 M, identical with the maximum in-
hibiting concentration found by Commoner and Thimann (1941).

Thus it is indicated that iodoacetate influences the growth of the
coleoptile in vivo in the same way that it affects the growth of isolated
sections. Very low concentrations of this substance produce a sharp
decline in the rate of coleoptile growth and this inhibition may be com-
pletely negated by the presence of sufficient fumarate.1

TABLE I

The relationship between iodoacetate concentration, final coleoptile length

and epidermal cell size.

Iodoacetate
Concentration

Coleoptile
Length (mm.)

Epidermal Cell Size

(ocular units)

50.0

26.0

ID"6 M

48.1

22.6

1(TB M

41.2

17.4

5X10-5 M

27.3

12.6

10-" M

26.9

11.3

That this effect is exerted on the growth processes within each cell is
shown by the data of Table I. These data indicate a fairly close pro-
portionality between the effect of various concentrations of iodoacetate
on the length of the entire coleoptile and the average length of the epi-
dermal cells of these coleoptiles. Thus it is suggested that the effect of
iodoacetate and fumarate is on some general auxin-sensitive process in
the cell.

The Influence of Iodoacetate and Fumarate on
Auxin-sensitive Processes

A further investigation of this hypothesis was made possible by the
fact that auxin has several different effects on various parts of the

1 Commoner and Thimann (1941) were able to get complete inhibition with
5 X 10~5 M iodoacetate. In these experiments, we were able to get growth inhibi-
tions of only 50 per cent maximally. We believe that this difference is due to pene-
tration phenomena which would be different for the intact seedling as compared to
the sections.

318

H. G. ALBAUM AND B. COMMONER

growing plant. If the above effects are due to a direct influence on the
activity of the auxin processes, then it should be possible to demonstrate
an augmentation of all auxin effects by fumarate and an inhibition of
these effects by iodoacetate (in the proper concentration). The follow-
ing experiments bear this out.

Three auxin-sensitive growth processes were selected for compari-
son: coleoptile (shoot) length, total root length, root number. It is

COLEOPTILE
LENGTH

ro Co

ROOT
NUMBER

if en

ROOT
LENGTH

o
O

AUXIN

FUMARATE

AUXIN +
FUMARATE

IODOACETATE

IODOACETATE
4- AUXIN

FUMARATE+
IODOACETATE

FUMARATE+
IODOACETATE
+ AUXIN

FIG. 3. Effects of auxin (10 mg./l.), fumarate (.01 M), auxin plus fumarate,
iodoacetate (.00001 M), iodoacetate plus auxin, fumarate plus iodoacetate, fumarate
plus iodoacetate plus auxin on the coleoptile length, root length and root number
of Avcna seedlings var. Black Norway.

already well known that in certain concentrations auxin tends to in-
crease the coleoptile length, decrease the root length and increase the
number of roots. The influence of auxin, fumarate, iodoacetate, and
various mixtures of these substances was determined in the usual manner.
The data obtained are presented in Fig. 3.

This figure shows that, whatever the direction of the auxin effect
(i.e., whether inhibitory or stimulatory), the effect is augmented by the
addition of fumarate and inhibited by the addition of iodoacetate.

FOUR-CARBON ACIDS AND GROWTH OF OAT SEEDLINGS 319

Thus, auxin reduces total root length ; auxin and f umarate produce
an even greater inhibition, while auxin and iodoacetate show a smaller
inhibitory effect. Iodoacetate alone has no significant effect on root
length, but reduces the number of roots and the coleoptile length. Auxin
reduces this inhibitory influence of iodoacetate, in the case of the latter
effects, but increases it in the case of root length. Fumarate alone acts
like auxin in increasing coleoptile length and root number, but also has
a slight positive effect on root length. In every case fumarate an-
tagonizes the effect of iodoacetate.

Thus it seems clear that the effects of iodoacetate and fumarate are
exerted directly on these auxin-sensitive processes in the cells. The in-
fluence of these substances on the various growth relations in the oat
seedling is directly related to the effect of auxin on these relations. It
seems likely that the four carbon acids and auxin are together concerned
with the activation of these processes, while iodoacetate is a specific in-
hibitor (in the proper concentration) of these processes.

The Differential Effect of Growth Substances on the Various

Growth Processes

It has been suggested by Thimann (1937) that the effect of auxin on
the various growth processes is essentially identical, but that the sensitivi-
ties of these processes to varying concentrations of auxin are different.
He points out that the curve relating intensity of effect to auxin concen-
tration is essentially the same for coleoptile length, root length, and root
number, but that the zero points are different in each case. The sug-
gested relation between these curves is shown in Fig. 4, together with
the zero points indicated by the work of Kaiser and Albaum (1939) on
two varieties of Arena sativa, Black Norway and Fulghum. The latter
work showed that the variety Fulghum responded to auxin in a manner
that indicated a greater intrinsic content of growth hormone as compared
with Black Norway. The data on the effect of auxin in the table ap-
pended to Fig. 4 (taken from Kaiser and Albaum, 1939) indicate the
probable validity of Thimann's suggestion. However, it is clear that
the curves to the left of the zero lines were purely hypothetical since
there was no method of quantitatively removing auxin from the plant.

It is apparent, from the data presented above, that iodoacetate does
offer the possibility of investigation in the " negative " regions of these
curves. Consequently, the effect of various concentrations of iodoace-
tate on the growth processes of the two varieties was determined. The
data are shown in the table of Fig. 4.

It is clear that the response of these growth processes to iodoacetate
follows the course of the curves to the left of the zero lines.

320

H. G. ALBAUM AND B. COMMONER

AUXIN

IODOACETATE

IXIO"6

IXIO"5

3XIO"5

5X10-5

8XIQ-S

IXIO"4

BLACK
NORWAY

SHOOT LENGTH
ROOT LENGTH
ROOT NUMBER

31 5
324
3.9

300
312

251

320
39

1 99

244
40

195
220
40

201
2 18
40

198
224
39

FULGHUM

SHOOT LENGTH
ROOT LENGTH
ROOT NUMBER

428
256
43

405
278
33

304
285
31

258
282
33

23 1
265
30

208
264
32

205
214

AUXI N MG/L

,005

0 2

0 5

5 0

50 0

BLACK
NORWAY

SHOOT LENGTH
ROOT LENGTH
ROOT NUMBER

290
320
40

340

208

389

1 10
48

80
53

FULGHUM

SHOOT LENGTH
ROOT LENGTH
ROOT NUMBER

441
250
52

440

200

451

100
57

30
62

100

FIG. 4. Hypothetical relations between auxin concentration and its effect on
root length, shoot length and root number (after Thimann, 1937). The vertical lines
for Black Norway and Fulghum represent zero points or intrinsic hormone content
as postulated by Kaiser and Albaum (1939). The auxin data in the table below
are taken from the latter paper.

FOUR-CARBON ACIDS AND GROWTH OF OAT SEEDLINGS 321

Thus, examination of the curves to the left of the Black Norway zero
line indicates that progressive " removal " of auxin should show the
following effects : on shoot length : an immediate decrease ; on root
length : no effect, followed by a decrease ; on root number : no effect.

OO
OZ

CTLJ

325

275

225

Hid
oca

4 5

o
u

• WITH IODOACETATE

O WITHOUT IODOACETATE

0 0001

0005

001

FUMARATE MOLAR ITY

FIG. 5. Relationship between coleoptile length, root number and root length
and the concentration of fumarate with and without .00001 M iodoacetate.

The data obtained follow this course. The shoot length is reduced
by the lowest concentration of iodoacetate, a maximum inhibition being
reached at a concentration of 5 X 10~5 M. Low concentrations of iodo-

322 H. G. ALBAUM AND B. COMMON KK

acetate (up to 10~r> M) have no effect on root length, but larger amounts
are inhibitory. Finally, all concentrations of iodoacetate used produce
no significant change in the root number.

In the case of the variety Fulghum, the curves are also descriptive of
the effect of increasing iodacetate concentrations. Since this variety
contains a greater intrinsic concentration of auxin, the zero point is
farther to the right. Thus, in this case, low concentrations of iodoace-
tate increase the root length, while higher concentrations reduce it.
Similarly, low concentrations of the poison cause a slight reduction in
root number, higher concentrations having no additional effect. There
is an immediate inhibition of shoot length, again following the course of
the curve.

It seems apparent that iodoacetate affects the growth processes in
Avena by quantitatively inactivating the auxin originally present in the
plants.

Such an effect might be taken to indicate a stoichiometric reaction
between auxin and iodoacetate, resulting in inactivation of the hormone.
However, such an explanation is contraverted by the marked acceleration
of the growth of coleoptile sections by a concentration of 10~6 M iodo-
acetate (see Commoner and Thimann, 1941).

Furthermore, reversal of the iodoacetate inhibition by fumarate also
seems to rule out this suggestion.

This is clearly shown by the data presented in Fig. 5. These curves
illustrate the negation of iodoacetate inhibition by fumarate and show
that there is no clearly proportional relation between the effects of these
substances.

CONCLUSIONS

It is clear that the 4-carbon acid system is an important factor in the
activity of auxin in controlling plant growth. There seems to be a close
interaction between these factors, indicating that both participate in the
various auxin-sensitive processes that regulate plant growth.

It is not our purpose at the present time to offer a complete explana-
tion of this phenomenon, but the data presented demonstrate that the
4-carbon acids 'participate directly in the growth processes in the plant.
The conclusions reached by Commoner and Thimann can therefore be
extended to include many of the known effects of auxin.

LITERATURE CITED

BONNER, J., 1936. The growth and respiration of the Avena coleoptile. Jour. Gen.

Physiol, 20: 1-11.
COMMONER, B., AND K. V. THIMANN, 1941. On the relation between growth and

respiration in the Avena coleoptile. Jour. Gen. Physiol., 24: 279-296.

FOUR-CARBON ACIDS AND GROWTH OF OAT SEEDLINGS 323

DuBuv, H., AND R. A. OLSON, 1940. The relation between respiration, proto-
plasmic streaming and auxin transport in the Avena coleoptile, using a
polarigraphic microrespirometer. Am. Jour. Bot., 27: 401-413.

KAISER, S., AND H. G. ALBAUM, 1939. Early root and shoot growth in two varie-
ties of Avena sativa in relation to growth substances. Am. Jour. Bot., 26 :
749-754.

THIMANN, K. V., 1937. On the nature of inhibitions caused by auxin. Am. Jour.
Bot., 24 : 407-412.

VAN HULSSEN, C. J., 1936. Ademhaling, gisting en groei. Thesis, Utrecht.

STUDIES IN SUBLITTORAL ECOLOGY
III. LAMINARIA FOREST ON THE WEST COAST OF SCOTLAND; A STUDY

OF ZONATION IN RELATION TO WAVE ACTION AND ILLUMINATION

J. A. KITCHING

(From the Department of Zoology, University of Bristol, England)

INTRODUCTION

Rocky areas of the sea bottom in the shallow sublittoral region of
the coast of Britain are in general densely covered with a forest of brown
laminarian seaweeds. In this paper an account is given of the influence
of depth on the distribution of organisms within the Lmninaria forest,
and of the factors which mainly control this distribution. Observations
were made by means of a diving helmet (see Kitching, Macan, and Gil-
son, 1934), from low water of spring tides to a depth of about 12
meters below this level.

Carsaig Island, the place chosen for this work, is about 1 km. long
and Y5 km. wide, and lies in the Sound of Jura, on the west coast of
Scotland (see inset to Fig. 1). The waters of the Sound of Jura in
general reach a depth of 100-200 meters, and the tidal currents attain a
velocity of 7—8 km. per hour ; in windy weather there is a short choppy
sea. According to measurements made in open water during hot still
weather, the gradient of temperature within the shallow sublittoral region
is insignificant; and it is probable that gradients of salinity, oxygen
content, ca-rbon dioxide content, and hydrogen ion concentration must
also be slight. However, in places sheltered from wave action such
gradients may well be set up. The western shore of Carsaig Island is
fully exposed to wave action, but the eastern side is well protected. As
will be shown below, the main features of the distribution of organisms
within the shallow sublittoral region may reasonably be attributed to
gradients of wave action, with all its consequences, and to illumination.

DISTRIBUTION IN THE LAMINARIA FOREST
Introduction and Methods

The Laminaria forest was investigated by observation from a boat
with the help of a water telescope all around the coast of Carsaig Island,
and by diving at several selected stations. Finally a detailed study was
made, by diving, at one station (A on map, Fig. 1), down to a depth of

324

SUBLITTORAL ECOLOGY OF LAMINARIA FOREST

325

about twelve meters below low water of spring tides. At this last sta-
tion data were collected for the drawing of a diagrammatic section (Fig.
1) to show the vertical distribution of algae. Investigations were car-
ried on in September of 1932, and in August of 1933-1936 inclusive.
The general description which follows is based on all these observations.

Dominant Large Brown Algae

The upper margin of the sublittoral region, — that part which lies
just above low water of spring tides, — is very different in character from

METRES

FIG. 1. Section at position A (see inset) on coast of Carsaig Island. Inset :
map showing position of Carsaig Island in Sound of Jura, Argyll, Scotland.

the deeper layers. It is dominated by three large brown algae, Himan-
thalia lorea, Alaria esculenta, and Laininaria digitata (see Fig. 1).
Alaria is confined to the west side of the island, and favors vertical rock
surfaces as well as wave action. L. digitata and Himanthcdia range all
around the island. All three grow sufficiently densely to protect the
organisms amongst them to a large extent from desiccation, so that the
associated fauna is typically sublittoral.

Just below low water of spring tides lies the lower limit of Laniinaria
digitata, and, except in extreme shelter, the upper limit of Lauiinaria
Cloustoni. There is very little mixing of the two species. Laminaria
Cloustoni in its zone forms a dense forest, but does not reach its greatest
size at depths less than 4 meters. Here the innumerable vertical stipes

326

J. A. KITCHING

support an immense canopy of fronds, below which the light is dim even
when the sun is shining brightly above. In order to penetrate this forest
without risk of the air line becoming entangled, we had to cut a glade
with shears. The forest reaches a height of about 3 meters above the
bottom. Individual plants of L. Clonstoni reach a length of 2 meters
or more, the longest measured being 2.40 meters ; and scattered plants
of Saccorhisa bulbosa reach about the same length. The full height of
the forest is made up by epiphytic Laminaria plants growing on the
stipes of L. Cloustoni. These epiphytic plants are chiefly L. digitata,
which, however, is rarely if ever found growing directly on the sea bot-
tom below its normal zone. In general Laminaria spp. from deeper
down or from the sheltered side of the island were found to have un-

TABLE I

Population of Laminaria digitata from the sublittoral fringe

Individuals
per 5 square
meters of frond

Individuals
per 10
stipes

Individuals
per 10
holdfasts

On east (sheltered) side of Car-
saig Island:
Patina pellucida
Spirorbis spp.
Scrupocellaria reptans

31
6800
45

0
6
0

1
599
0

On west (wave-exposed) side of
Carsaig Island:
Patina pellucida
Spirorbis spp.
Scrupocellaria reptans

131
0
0

8
0
0

10
652
2

divided or incompletely divided fronds, in contrast to the plants growing
nearer the surface and on the wave-exposed side. Presumably wave
action is most vigorous near the surface of the water.

Between 6 and 12 meters below low water of spring tides the forest
opens out, so that it was possible to walk between the Laminaria plants
easily in spite of the rugged nature of the sea bottom ; we therefore
called this area the " park." In the park nearly every L. Cloustoni
was entwined with a large unattached plant of the brown alga Des-
inarestia aculeata. Although 12 meters was the greatest depth to which
we penetrated, it was possible through the misty water to see the park
extending downwards much further on the steeply sloping bottom.

In extreme shelter from wave action L. Cloustoni is replaced by L.
saccharina, or by a form, allied to this species, which is characteristic of
the sea lochs of the west coast of Scotland (Kitching, 1935). The

SUBLITTORAL ECOLOGY OF LAMINARIA FOREST 327

latter extends up to the lower margin of the zone of L. digitata in the
extreme shelter of the small bay on the eastern side of the island, but
southwards with increasing wave action it is progressively replaced from
above by L. Cloustoni.

Associated Flora and Fauna

The canopy of Laminaria fronds is relatively clean of epiphytic algae
in shallow water on the western side of Carsaig Island ; instead it carries
many colonies of the hydroid Obelia geniculata and the polyzoan Mem-
branipora inembranacea, and numerous gastropods, including Patina
pellucida. On the more sheltered eastern side of the island Obelia and
M. membranacea are much less abundant ; they are practically confined to
a narrow zone at the level of low water of spring tides, and they are
absent from the very sheltered Bay (see map in Fig. 1) ; whereas in
shelter filamentous epiphytes with their associated fauna, and the tube-
worm, Spirorbis (spp.), and the polyzoan, Scrupoccllaria rep tans, may
cover nearly all the space available on the Laminaria fronds. These
same differences, though in lesser degree, may be observed at greater
depths on the wave-exposed western side of the island. Here again it
may reasonably be claimed that depth affords shelter from wave action.

The undergrowth of the sublittoral fringe consists mainly of the
green alga Cladophora rupestris and the red algae Chondrus crispus,
Gigartina stellata, and Rhodymenia palmata. Of these Cladophora and
Rhodymenia also grow as epiphytes on the uppermost parts of the stipes
of Laminaria Cloustoni in shallow water only, and small plants of
Rhodymenia are sometimes found attached to the tips of Laminaria
fronds in shallow water. The distribution of these algae is such as to
suggest that they require a relatively high incidence of light.

The holdfasts and stipes of the Laminaria Cloustoni plants, except
the upper parts of those nearest to the surface of the water, and the rock
bottom between the holdfasts, are densely covered with an undergrowth
of red algae, which includes Membranoptcra alata, Phycodrys rubcns,
Odonthalia dentata, Delesseria sanguinca, Ptilota plumosa, and Crypto-
pleura ramosum. Callophyllis laciniata flourishes especially in the park,
where its bright red fronds may be seen from afar. Apart from this,
the composition of the undergrowth appears uniform from 2 to 12
meters. All these algae appear to favor a relatively weak illumination,
a conclusion which is supported by photoelectric measurements reported
later in this paper (Table II).

The tube-worms Spirorbis spp. and Salmacina incrustans, as well as
various polyzoans and colonial tunicates, are abundant on the inner parts
of the Laminaria holdfasts, even in places well exposed to wave action;

328 J. A. KITCHING

hut the outer surfaces of these holdfasts are usually clean. The hold-
fasts shelter an extensive motile fauna, which will not be described in
this paper. The composition of this fauna did not appear to be signifi-
cantly influenced by depth within the limits of our exploration.

Recolonisation of Denuded Areas of Laminaria Forest

In August, 1936 we were able to examine areas where the Laininaria
forest had been cut down with shears 12 months previously. The old
holdfasts had disappeared, and new Laminaria Cloustoni plants covered
the area very densely, and had grown to a height of about 1 meter. The
holdfasts were affixed very tightly to the rock. They were almost clean
of epibiotic organisms, but a few specimens of Spirorbis were found on

TABLE II

Depth

Illumination

Position of sea photometer

below
water surface

cut off
by seaweeds

(meters)

(per cent)

Under Saccorhiza, near Callophyllis

4.0

99.5

Among Chondrus and Cladophora, in zone of

Laminaria digitata

1.5

At bottom, in old forest

3.9

98.7-99.4

Under dense new growth of L. Cloustoni, one

year old

4.0

99.1-99.5

Under Laminaria Cloustoni in "park"

82-85

Under old L. Cloustoni forest

4.0

98.8-99.1

The first four observations listed above were made on August 13, 1936, between
12.25 and 1.00 P.M. G.M.T., under an overcast sky. The last two observations
were made on August 22, 1936, between 2:08 and 3:15 P.M., under a sunny sky with
light cirrhus cloud. On both days the wind was light, the water temperature about
12.5-12.6° C., and the air temperature 15-16° C. All these observations were made
at station A (Fig. 1).

them. Some Patina were found on the stipes and fronds, but little else.
The motile fauna associated with these young Laininaria plants was
much poorer in numbers than that of the older forest, but was in general
similar in constitution.

ILLUMINATION IN THE LAMINARIA FOREST
Apparatus and Methods

Measurements of the light intensity at a number of positions in the
Laminaria forest were made photoelectrically, according to the general
methods described by Atkins and Poole (1933). The light intensities
in air and at a chosen under-water position were measured simultane-
ously and compared. Two " photronic " rectifier cells, made by the

SUBLITTORAL ECOLOGY OF LAMINARIA FOREST 329

Weston Electric Company, were employed. Separate galvanometers
were used for the two cells, as this was found to be an advantage in
changeable light, and also reduced the time during which the diver had
to wait in the cold while readings were being taken. The cells were
mounted in water-tight containers, and could be screened either with a
flashed opal glass alone or with a selected color filter x underneath an
opal glass.

The two cells were kindly calibrated for me by Dr. W. R. G. Atkins
of the Marine Biological Laboratory, Plymouth. The calibrations were
carried out by daylight under a variety of light intensities, with the cells
mounted ready for use but dry, and under opal glass without color filters.
Submersion under a thin film of water made scarcely any difference.
For low light intensities, and with the low external resistance used, the
relation between current and light intensity was approximately linear.
For most purposes 10 ohm pattern L galvanometers made by the Cam-
bridge Instrument Company were used, but for very low light intensi-
ties, as recorded in the depths of the Laminaria forest, it was necessary
to use a 1000 ohm pattern LY galvanometer ; however, at these low in-
tensities of light the linear relation noted above was found still to hold
in spite of the increased resistance of the external circuit.

It is recognized that the lux is not an ideal unit for present purposes ;
and that the measurement of illumination over a wide spectral range,
with a photocell differentially sensitive with respect to wave-length, is
not an entirely satisfactory procedure. However, it is believed that
these objections do not invalidate the treatment of results which will be
followed below.

Results

As a result of observations made with the opal screen alone, the light
intensity in various situations in the Laminaria forest was found to be
only a very small fraction of that in air. A comparison was then made
of the light intensities within the forest and in open water at the same
depth by using curves obtained in open water (Fig. 2). In this way
the light intensity within the forest was expressed as a percentage of
what it would have been had there been no forest, and from this was
computed the percentage illumination cut off by the forest. A few typi-
cal results are given in Table II. In general the forest cut off about 99
per cent of the available light, and new growth one year old was about
as effective. However, in the park, only 82-85 per cent was cut off by
the seaweeds, and the illumination at the bottom was better than in the

1 Listed by the makers (Schott and Gen, of Jena) as BG 12 (blue, 2 mm.
thick), VG 2 (green, 4 mm.), RG 1 (light red, 2 mm.), and RG 8 (dark red, 2
mm.).

330

J. A. KITCHING

forest, even though the depth was greater. It is not possible to say to
what extent this condition is general on the British coast.

The illumination at any one position in the forest fluctuates continu-
ally owing to the movement of the fronds overhead. Therefore it was

100

ILLUMINATION

8O 6O 4O 2O

FIG. 2. Graph showing illumination at various depths in open water at position
A, on the west side of Carsaig Island (see inset to Fig. 1). A plain opal glass,
without color filters, was used. The illumination in air is rated as 100.

El August 12, 1936 (transmission 79-81 per cent per meter).
A August 13, 1936 (transmission 80-87 per cent per meter).
G August 19, 1936 (transmission 81-85 per cent per meter).

found impossible with our equipment to investigate in detail the quality
of the light in the forest, as compared with that in open water. How-
ever, since the Laminaria fronds are opaque and of a darkish brown
color, it seems probable that not much light of any wave-length is either
transmitted or reflected into the depths of the forest by the fronds ; pre-

SUBLITTORAL ECOLOGY OF LAMINARIA FOREST

331

suniably most of that which reaches the depths of the forest passes be-
tween the fronds. Therefore, results obtained with the opal alone prob-
ably represent approximately the fraction of light of any wave-length
which penetrates the forest, as compared with the illumination at that
wave-length in the same depth of open water. It is therefore possible to
estimate approximately the depths in open water at which the monochro-
matic illumination at various frequencies is equivalent to that in the
Laminaria forest just below low water mark. For this somewhat rough

100

ILLUMINATION

8O 6O 4O 2O

2
3

QlO

1 1

FIG. 3. Graph showing illumination at various depths in open water at posi-
tion A, on the west side of Carsaig Island (see inset to Fig. 1). Color filters were
used in addition to opal glass. The illumination in air is rated as 100. The obser-
vations were made on August 19, 1936.

Blue, 350-470 m/*, transmission 75-91 per cent per meter.
Green, 500-570 mju, transmission 87-93 per cent per meter.
Light red, 620-700 m/u. transmission 53-65 per cent per meter.
Dark red, 700-750 niju, transmission 27-28 per cent per meter.

and arbitrary procedure the illumination in the Laminaria forest has
been taken as 1 per cent of that in open water at the same depth, and
the transmission for light of various wave-lengths has been given the
values obtained from Fig. 3, although indeed transmission may actually
vary with depth. The results of this calculation are given in Table
III, and are discussed further on (p. 334). The preferential transmis-
sion of green light is in accord with the results of many other workers
in coastal waters (Knudsen, 1922; Atkins, 1926; Klugh, 1927; Clarke,
1936, 1939).

332 J. A. KITCHING

DISCUSSION OF ZONATION

The zonation of plants and animals is very strongly marked in the
littoral region of the British coast. The dominant algae are restricted
very obviously to narrow belts along the shore ; while the zonation of
animals, although usually less precise, is fully shown by studies of popu-
lation density and growth rate (Fischer-Piette, 1936; Moore, 1934).
Littoral zonation may reasonably be attributed in the main to desiccation
(Baker, 1910; Muenscher, 1915; Kanda, 1916; Johnson and Skutch,
1928; and Hatton, 1930, 1932), although in general decisive proof has
not yet been given, and to a smaller degree to the effects of heat and
light (Gail, 1919, 1922; Klugh and Martin, 1927). In response to the
severe gradation of the controlling environmental factors, the zonation
of organisms in the littoral region is sharply defined.

At the upper margin of the sublittoral region, along a strip of shore
only uncovered at low water of spring tides, there is a peculiar and char-
acteristic zone called appropriately by Stephenson ct al. (1937) the
" sublittoral fringe." The characteristic algae are probably restricted
at their upper limits by desiccation, and at their lower limits by deficient
illumination. Wave action is maximal here, and the larger brown sea-
weeds (Himanthalia, Alaria, Laminaria digitata), as is frequently
pointed out, are well adapted by their pliable stipes to withstand sudden
and violent stress. Perhaps of equal importance is the fact that at low
tides they lie prostrate and so in the main escape desiccation. After
low spring tides in very hot dry weather I have seen the fronds of Lami-
naria digitata around Carsaig Island scarred with dead patches where
they had projected from the water. However, at greater depths these
algae of the sublittoral fringe cannot compete with the erect and less
pliable Laminaria Cloustoni. It is noteworthy that L. digitata fails to
penetrate the true sublittoral region except as an epiphyte of L. Clou-
stoni; by this means it borrows the advantages of a tall erect stipe and
achieves " a place in the sun." On the other hand, L. Cloustoni, though
a true sublittoral form, may by its habit of growth expose its frond to
the dangers of emersion at low water of spring tides, and is, in fact, quite
probably limited by desiccation. The animals of the sublittoral fringe
are restricted at their upper limits by exposure to air, but in general
range downwards extensively into the shallow sublittoral, since they are
less dependent on light than are the algae. However, in certain cases
immersion may possibly be detrimental (Moore and Kitching, 1939).
In the sublittoral region proper the upper limits of distribution of
organisms are determined by excessive wave action, and perhaps by
excessive illumination in some cases ; whereas the lower limits are set by

SUBLITTORAL ECOLOGY OF LAMINARIA FOREST 333

deficient wave action and deficient illumination. The factors effective
in controlling zonation below the sublittoral fringe are not steeply
graded, and therefore zonation is not sharply marked.

The limitation of the vertical distribution of sublittoral organisms
by wave action can be demonstrated readily, because in such cases these
limits are raised where the coast-line provides greater shelter, depressed
wrhere it is more open to the waves. The upper limits of Laminaria
saccharina, in its sea-loch form, and the lower limits of Membrampora
rncnibranacea and Obelia gcniculata (see p. 327) are examples. The ef-
fect of wave action upon marine organisms is complex and obscure ; and
the amount of wave action necessary to support the existence of an or-
ganism may depend on other environmental conditions (Moore and
Kitching, 1939). Apart from its destructive mechanical effect, wave
action probably influences the settling of larval forms, promotes the
circulation of planktonic food, disturbs sediment, and obliterates extreme
fluctuations in the physical and chemical conditions of the water.

Limits of distribution determined by illumination, in contrast to
those set by wave action, are likely to be relatively independent of the
conformation of the coast-line over a small area, despite local variations
in the loss of light at the surface of the water. The transition from
the sublittoral fringe to the zone of Laminaria Cloustoni involves a de-
cline in illumination at the sea botttom so great and so steep that none
of the chief undergrowth-forming algae is common to both levels.
However, within the broad L. Cloustoni zone, the illumination is by
comparison almost uniformly dim, and over a range of 12 meters no
case was found in which limitation could be attributed to illumination.
Although the upper limits of the shade-loving undergrowth algae of
the L. Cloustoni zone coincide with a sudden gradient of illumination,
yet it is not clear whether these algae are restricted directly by excessive
illumination, or by some other factor, such as a brief exposure during
low equinoctial spring tides. It has been shown that in certain of these
algae, Delesseria sanguinca and Plocainiiun coccineum (Moore, Whitley
and Webster, 1923), and under certain conditions, the greatest photo-
synthetic activity takes place under moderate rather than very strong-
illumination ; but for present purposes much more experimental evidence
is required. It must also be remembered that this discussion is con-
cerned throughout with organisms in competition, and that competition
is likely to accentuate the sensitivity of these organisms to their environ-
ment (Beauchamp and Ullyott, 1932).

It is important to recognize that not only the quantity but also the
quality of the light varies with depth. Green and blue light, — in coastal
waters especially the former, — penetrate more readily than orange, red,

334

J. A. KITCHING

and ultra-violet ; and therefore with increasing depth they predominate
to an ever-increasing extent (see the review by Clarke, 1939). Atkins
(1926) has stressed the importance of this change in quality in relation
to the vertical distribution of algae. In spite of certain obvious excep-
tions, it is in general true that green and brown algae are restricted to
littoral and shallow sublittoral levels, whereas many red algae penetrate
to greater depths. For instance, red algae have been dredged in the
English Channel (Hamel, 1923) and off the Faeroes (Bjzfrgesen, 1908)
from depths of 45 and 50 meters respectively, whereas the brown algae
do not in general extend below 25 meters in Faeroese waters. At the
depths to which red algae penetrate the light is mainly blue and green.
It has been suggested (for references see Atkins, 1926) that by the na-
ture of their photosynthetic pigments they derive the greater part of
their energy from light of these colors ; and this hypothesis may be ap-
plied with reasonable safety to algae growing at depths where green and
blue light vastly predominate. However, various of these shade-loving
red algae, Phycodrys rub ens, Delesseria sanguined, Ptilota plumosa,
found at 40-50 meters off the Faeroes, are characteristic of the under-
growth of the Laminaria forest in the shallow sublittoral region, and are
also found in sea caves (B^rgesen, 1908; Rees, 1935). This implies a
wide tolerance of variations in the spectral composition of the light. In
spite of qualitative differences in illumination, there is clearly an eco-
logical similarity between the shades of the Laminaria forest, the half
darkness of caves, and the open sea bottom of the deeper sublittoral
region. The incidence of green and blue light is probably of the same
order in the deeper sublittoral and in the Laminaria forest (Table III).

TABLE III

Color of light

Blue

Green

Light red

Dark red

Range of filters (approxi-
mate)

350-470 ITIM

500-570 m/x

620-700 niM

700-750 niM

Transmission per meter
(see Fig. 4)

83%

90%

59%

27.5%

Depth at which monochro-
matic light intensity is
equivalent to 1% of that
in Laminaria forest near
low water mark

25 meters

40 meters

10 meters

3 meters

Although this suggests that within the Laminaria forest the green and
blue light alone are sufficient to support the shade-loving red algae, it

SUBLITTORAL ECOLOGY OF LAMINARIA FOREST 335

still remains uncertain to what extent use can be made of light of other
colors. It seems possible from the work of Klugh (1931) that some at
least of the deeper water red algae, though able to make an exceptional
use of green light, may nevertheless be able to utilize profitably a consid-
erable range of the spectrum. Much more work is needed on this
subject.

ACKNOWLEDGMENTS

I wish to express my thanks to the numerous helpers who assisted in the work
of diving and of handling equipment ; and especially to Mr. H. C. Gilson, Mr.
G. I. Crawford, and Dr. A. Haddow. In addition, Mr. Gilson has given me much
valuable advice. For assistance in transport and storage I am greatly indebted to
Mr. G. Brodie and Mr. D. J. Gillies. Numerous specimens of algae were kindly
examined by Miss C. I. Dickinson at Kew Herbarium. I am grateful to Pro-
fessor C. M. Yonge of Bristol University, and to Professor H. G. Jackson of
Birkbeck College, London, for advice and encouragement. I am indebted for a
grant towards expenses to the Challenger Society, and for a grant for apparatus
to the Dixon Fund Committee of the University of London.

SUMMARY

1. On the shores of Carsaig Island, Scotland, the sublittoral region is
densely forested with laminarian seaweeds down to a depth of at least 15
meters, and probably down to a much greater depth.

2. At the upper margin of the sublittoral region, exposed to air at
low water of spring tides, there is a characteristic " sublittoral fringe."

3. The dominant brown algae of the sublittoral fringe have pliable
stipes, so that they lie flat at low water and escape desiccation. The
dominant brown algae of the true sublittoral region have tall erect stipes,
which hold up the fronds to the light.

4. The Laminaria canopy at depths of 1-6 meters cuts off about 99
per cent of the available light. At 6-12 meters the forest is less dense,
and relatively more light penetrates. There is therefore a sharp decline
in illumination between the sublittoral fringe and the Laminaria forest
just below it; but within the forest the illumination changes much less
over a considerable range of depth.

5. Wave action is considered to be maximal in the sublittoral fringe,
and to decrease gradually with depth.

6. The undergrowth-forming algae fall into two clearly defined
groups — one group confined to situations of high illumination, such as
the sublittoral fringe and the upper parts of Laminaria stipes ; and the
other to dark places, such as the rock surface in the depths of the forest
and the lower parts of the Laminaria stipes.

7. The undergrowth of the Laminaria forest is practically uniform
in composition within a vertical range of about 12 meters, and probably
more.

336 J. A. KITCHING

8. Vertical distribution is determined chiefly by wave action in the
case of one laminarian alga, two polyzoa, one hydroid, and probably
other organisms.

9. It was found that on artificially denuded areas new Laminaria
Cloustoui plants grew to form a forest 1 meter high in 12 months.

10. It is concluded that whereas the steeply graded zonation of the
littoral region is to be ascribed in the main to desiccation, the more gentle
zonation of the sublittoral region depends on illumination and on wave
action.

LITERATURE CITED

ATKINS, W. R. G., 1926. A quantitative consideration of some factors concerned

in plant growth in water. Part I. Some physical factors. Jour. Cons.

Internal. Explor. Mer, 1 : 99-126.
ATKINS, W. R. G., AND H. H. POOLE, 1933. The photo-electric measurement of

the penetration of light of various wave-lengths into the sea, and the

physiological bearing of the results. Phil. Trans. Roy. Soc. London, B,

222 : 129-164.
ATKINS, W. R. G., AND H. H. POOLE, 1933. The use of cuprous oxide and other

rectifier photo cells in submarine photometry. Jour. Mar. Biol. Ass'n, 19 :

67-72.
BAKER, S., 1909. On the causes of the zoning of brown seaweeds on the seashore.

Neiv Phytol, 8 : 196-202.
BAKER, S., 1910. On the causes of the zoning of brown seaweeds on the seashore.

II. The effect of periodic exposure on the expulsion of gametes and on the

germination of the oospore. Nciv Phytol., 9 : 54-67.
BEAUCHAMP, R. S. A., AND P. ULLYOTT, 1932. Competitive relationships between

certain species of fresh-water triclads. Jour. EcoL, 20: 200-208.
B0RGESEN, F., 1908. The algae-vegetation of the Faeroese coasts, with remarks

on the phytogeography. Botany of the Faeroes, Part III, 683-834.
CLARKE, G. L., 1936. Light penetration in the western north Atlantic and its

application to biological problems. Ra[>. ct Proc. Cons. Internal. Explor.

Mer, 101 (part 2) : no. 3.
CLARKE, G. L., 1939. The utilization of solar energy by aquatic organisms. Publ.

Am. Ass'n Adv. Sci., No. 10, pp. 27-38.
FISCHER-PIETTE, E., 1936. Etudes sur la biogeographie intercotidale des deux rives

de la Manche. Jour. Linn. Soc. Zool., 40 : 181-272.
GAIL, F. W., 1919. Hydrogen ion concentration and other factors affecting the

distribution of Fucus. Publ. Puget Sound Afar. Biol. Sta., 2: 287-306.
GAIL, F. W., 1922. Photosynthesis in some of the red and brown algae as related

to depth and light. Publ. Puget Sound Mar. Biol. Sta., 3 : 177-193.
HAMEL, G., 1923. Sur la limite de la vegetation dans la Manche, d'apres les dra-

gages effectues par le " Pourquoi-Pas? " C. R. Acad. Sci. Paris, 176:

1568-1570.
HATTON, H., 1930. Sur la repartition des algues calcaires dans la zone des marees.

Bull. Lab. Mar. Saint-Servan, 5: 41-44.
HATTON, H., 1932. Quelques observations sur le peuplement en Fucus vesiculosus

des surfaces rocheuses denudees. Bull. Lab. Mar. Saint-Servan, 9 : 1-6.
JOHNSON, D. S., AND A. F. SKUTCH, 1928. Littoral vegetation on a headland of

Mt. Desert Island, Maine. II. Tide-pools and the environment and classi-
fication of submersible plant communities. Ecology 9 : 307-338.
KANDA, S., 1916. Studies on the geotropism of the marine snail, Littorina littorea.

Biol. Bull., 30 : 57-84.

SUBLITTORAL ECOLOGY OF LAMINARIA FOREST 337

KITCHING, J. A., 1935. An introduction to the ecology of intertidal rock surfaces

on the coast of Argyll. Trans. Roy. Soc. Edinb., 58 : 351-374.
KITCHING, J. A., T. T. MACAN, AND H. C. GILSOX, 1934. Studies in sublittoral

ecology. I. A submarine gully in Wembury Bay, South Devon. /. Mar.

Biol. Ass'n, 19 : 677-705.
KLTJGH, A. B., 1927. Light penetration into the Bay of Fundy and into Chamcook

Lake, New Brunswick. Ecology, 8 : 90-93.
KLUGH, A. B., 1931. Studies on the photosynthesis of marine algae. No. 1.

Photosynthetic rates of Enteromorpha linza, Porphyra umbilicalis, and

Delesseria sinuosa in red, green and blue light. Contr. Canad. Biol., N.S.,

6: 41-63.
KLUGH, A. B., AND J. R. MARTIN, 1927. The growth-rate of certain marine algae

in relation to depth of submergence. Ecology, 8: 221-231.
KNUDSEN, M., 1922. On measurement of the penetration of light into the sea.

Publ. Circ. Cons. Internat. Explor. Mcr, 76 : 1-16.
MOORE, B., E. WHITLEY, AND T. A. WEBSTER, 1923. Studies of photo-synthesis in

marine algae. Trans. Liverpool Biol. Soc., 37 : 38-51.
MOORE, H. B., 1934. The biology of Balanus balanoides. I. Growth rate and its

relation to size, season, and tidal level. Jour. Mar. Biol. Ass'n, 19 :

851-868.
MOORE, H. B., 1936. The biology of Balanus balanoides. V. Distribution in the

Plymouth area. Jour. Mar. Biol. Ass'n, 20: 701-716.
MOORE, H. B., AND J. A. KITCHING, 1939. The biology of Chthamalus stellatus

(Poli). Jour. Mar. Biol. Ass'n, 23: 521-541.
MUENSCHER, W. L. C., 1915. Ability of seaweeds to withstand desiccation. Publ.

Pugct Sound Mar. Biol. Sta., 1 : 19-23.

REES, T. K., 1935. The marine algae of Lough Ine. Jour. EcoL, 23 : 69-133.
STEPHENSON, T. A., A. STEPHENSON, AND C. A. DU TOIT, 1937. The South Af-
rican intertidal zone and its relation to ocean currents. I. A temperate

Indian Ocean shore. Trans. Roy. Soc. S. Africa, 24 : 341-382.

FURTHER EXPERIMENTS IN CROSS- AND SELF-
FERTILIZATION OF CIONA AT WOODS HOLE
AND CORONA DEL MAR

T. H. MORGAN

(Prom the William G. Kerckhoff Laboratories of the Biological Sciences,

California Institute of Technology, and the Marine Biological

Laboratory, Woods Hole, Mass.)

Of the many studies that have been carried out on the development
of the eggs of marine animals, it is surprising how little attention has
been paid to the development of 100 per cent normal embryos or larvae.
Even in the case of the sea-urchin, where hundreds of experiments have
been reported, the percentage of normal embryos is seldom, if ever,
recorded. Many workers are contented with batches of these eggs in
which the fertilization membrane is given off in practically all the eggs,
but even then the number of such eggs that produce normal plutei is
seldom stated. E. E. Just, almost alone, has emphasized the importance
of using normal sea-urchin's eggs, and also the need of paying the
strictest attention to the environment in which the eggs develop. Albert
Tyler has also taken great care to use only normally developing eggs in
his physiological experiments.

Sea-urchins brought back from the collecting grounds in crowded
jars or in buckets to the laboratory or those kept in floating cars, or kept
without food in aquaria are sometimes recognized as a source of ab-
normalities, but the condition of their eggs is, as a rule, ignored.

In the earlier work on artificial fertilization the occurrence of " swim-
mers " was often reported as indicating a successful result, but every
embryologist knows that " swimmers " are abnormal embryos. Only as
methods improved were normal plutei reported, but practically never was
the percentage of normals given.

Most of those who have used sea-urchins are familiar with the fact
that individuals are frequently met with whose eggs fail to give off a
normal fertilization membrane, although the eggs appear to be normal.
To what extent this is due to the eggs not being " ripe " or overripe is
seldom known, even though in sea-urchins the polar bodies have been
given off when the eggs leave the wall of the ovary and are presumably
mature. In the starfish, on the contrary, the eggs free in the ovary still
contain the large germinal vesicle which will disappear to form the polar
spindle when the eggs are removed to sea water. It is well known, how-

338

CROSS- AND SELF-FERTILIZATION OF CIONA 339

ever, that such eggs often develop abnormally when fertilized. When a
starfish spawns normally, the eggs, before extrusion, have matured, i.e.,
the germinal vesicle has disappeared, and it is recognized that such eggs
give, as a rule, normal development. Of course, it is known that poly-
spermy in the sea-urchin is one of the factors of abnormal development,
and this holds for other animals whose eggs are fertilized in the labora-
tory. Fortunately, in dona polyspermy is a relatively rare occurrence,
and this source of abnormal development is practically eliminated.
Nevertheless, sets of eggs that give 100 per cent two-cell stages some-
times give rise to some or to many abnormal embryos.

So far I have referred to what appear to be internal factors in the
eggs themselves which cause abnormal development. What these fac-
tors are, aside from immaturity or over-ripeness, is generally unknown
and has been little studied. On the other hand, it is well known that
developing marine embryos are extremely sensitive to external factors
such as temperature, salinity, impurities in the water, bacteria, etc. In
forms that develop slowly (outside the parent), the chances of unfavor-
able conditions appearing are much increased, but even in rapidly de-
veloping forms, such as dona, external factors may also play a signifi-
cant role. Fortunately, in this animal the completely formed tadpoles
develop at 22° C. in fifteen hours or less, and it is not difficult to control
the environmental factors during this time. But, as the following ex-
periments clearly show, unless the eggs are thoroughly washed and the
excess of sperm removed, abnormal development is apt to occur. Even
then, however, different individuals may give quite different results
when the environment is apparently the same for all. Perhaps this may
be expressed by saying that the eggs of different individuals respond
differently to the same environmental differences. If this is the correct
interpretation, as the evidence at hand seems to indicate, it is evident
that contradictions may appear when the eggs of different individuals
are treated in the same way. This possibility makes the problem diffi-
cult, but repetitions of the same kinds of experiments have helped to
clear away some of the apparent contradictions.

The most puzzling problem is the occurrence of both normal and
abnormal larvae in the same culture of Ciona. The numbers may vary
from one or two normals to 99.5 per cent normals. Sets of eggs of one
individual fertilized by sperm of one other individual tend to give the
same proportions, under external conditions that are the same, but even
here, exceptionally, the ratios may vary. Different samples of eggs may
account in part for these differences, even when an attempt is made to
make the samples the same. Differences in the position of the eggs in
the dishes may make for differences, but when few are present they space

340 T. H. MORGAN

themselves equally on the bottom, and even when several thousand are
present in a thin layer of water, practically 100 per cent may be normal.
;< Accidents " of development, such as the position of the cleavage
planes, may be another factor, but it is impossible to say what causes
these accidents, — whether they are internal or external.

The western Ciona, like the eastern one, sets free its eggs and sperm
in the early morning in response to a change from darkness to light. In
general, therefore, fewer eggs are found in the oviduct in the later morn-
ing than in the afternoon, and the former may be said to be younger
than the latter, although all have formed the polar spindle before leav-
ing the ovary. No difference has been observed between the younger
and older eggs in respect to the development of normal embryos. When
Cionas are brought into the laboratory and kept in running water, or
in aerated water, the eggs may accumulate, in some individuals at least,
until the oviduct is swollen with them. They may be two or three days
old, yet produce as many normals as eggs from freshly-caught indi-
viduals. If the water is changed daily and kept clean, and only a few
Cionas kept in the same jar, the eggs are good for at least three days.

My interest in the problem of abnormal development in Ciona is only
secondarily concerned with the problems mentioned above. It became
necessary to find out to what extent the abnormal development is due to
internal factors on account of its possible relation to the genetic prob-
lem of self-sterility and cross-fertilization in Ciona. If abnormal de-
velopment is due to inherited genetic factors, then its occurrence, if regu-
lar and internal, might be due to genetic lethals whose presence might
bear on the main problem.

Neiv Experiments ivith Ciona at Woods Hole, Mass.

The results of experiments with Ciona at Woods Hole which I car-
ried out during the years 1904 and 1910, are in some instances more
erratic or irregular than those obtained in recent years with what is said
to be the same species on the California coast. This is true, moreover,
of experiments that I carried out in 1904 at Coronado Beach, California,
on Ciona which is undoubtedly the same type that is found at Corona
del Mar. I was inclined to think that the methods of handling the eggs
in the earlier experiments at Woods Hole might account for the differ-
ences, and therefore when I had a chance to test out the Woods Hole
form during September, 1940 I carried out some experiments that I
hoped would show whether the differences are due to the earlier tech-
nique employed, or to differences in the material itself. Two kinds of
tests were carried out. There were five of the 5X5 cross- and self-
experiments, and five experiments of a different kind.

CROSS- AND SELF-FERTILIZATION OF CIONA 341

In the 5X5 tests a larger amount of water was used than in the
former experiments at Woods Hole ; also the eggs wrere washed in one or
two changes of water. The eggs were then concentrated in a small
amount of the water, and 7 to 10 drops of the eggs were then transferred
to Stender dishes containing 20 cc. sea water. The cleavages were noted,
and after 20 to 24 hours the kind of development recorded. In all cases
(100 in all) the cross-fertilized eggs gave 100 per cent cleavage. There
were no failures to cross-fertilize. In the 25 selfed lots most gave no
cleavage (except that in one case 95 per cent cleaved), but there were a
few cleavages in some lots. It is to be noted that only 5 drops of sperm
suspension were used, and the concentration of sperm was not large since
the individuals were small.

The first lots came from a float in the seal-pool of the Fish Commis-
sion. There were at least a thousand Cionas on the float when removed ;
most of them were only half grown, but a few had eggs and sperm.
Those reported in the first three sets came from this float ; those in the
two remaining sets came from the supply tank on the roof of the labora-
tory. These were also small individuals and only a few were mature
or had enough eggs for the tests.

I. (Sept. 16, 1940.) All of the cross-fertilized eggs gave 99 or 100
per cent cleavage (except one that gave 50 per cent). The five selfed
sets gave 0, 25, 25, 50, 40 per cent cleavage. The total number of eggs
in the dishes (20 cc.) was between 50 and 100. After 24 hours it was
found that three lots of eggs, viz.. c, d, e, gave very abnormal embryos;
one lot, b, gave all normal or late abnormal stages ; and another lot, a,
gave normal, late abnormal tadpoles, and early abnormal tadpoles. It
was very noticeable that the eggs rather than the sperm determined the
kind of devlopment that took place. Over and beyond this, however,
there are differences in the different lots of eggs of the same set that
seem to follow the sperm, or the combination of eggs and sperm.

II. (Sept. 16, 1940.) These Cionas came from the same float as the
last. They had been kept in the laboratory in running water for one
day. The 20 crossed lots gave 100 per cent normal 2-cell stages ;* rarely
one or two of the selfed lots divided. Five drops of sperm suspension
were used, and 15 cc. sea water. Taking the crossed eggs in the hori-
zontal lines, aB and aC gave abnormal embryos, aD normal tadpoles and
aE late embryos. The next lot, b eggs, gave all normal tadpoles as did
the c eggs. The c and the d eggs gave nearly all abnormal embryos.
The eggs appear responsible for the differences, and not the sperm, which
was the same in each of the vertical lots.

III. (Sept. 18, 1940.) These came from the same source as I, but
had been kept in running water in the laboratory for two days. Only

342 T. H. MORGAN

a few of the Cionas had enough eggs for the experiment. The sperm
duct was better. Five drops of sperm suspension were used, and the
eggs were in 16 cc. sea water when fertilized. Most of the supernatant
fluid was then drawn off the eggs, and 10 drops of eggs added to 20
cc. sea water in Stender dishes. There were 75 to 150 eggs in each dish
(fewer in d}. All 20 crosses gave 100 per cent regular cleavages, ex-
cept Ad and Bd, that had a few irregular cleavages. None of the eggs
selfed. The cross-fertilized eggs gave abnormal embryos that died
young (stage d and c). (See Morgan, 1938a, p. 305.)

IV. (Sept. 19, 1940.) These Cionas came from the tank on the
roof of the laboratory. The cleavage was not observed. The eggs were
washed twice in sea water and 12 drops transferred to each dish (20
cc.). Five drops of sperm suspension were added to each. Four of the
egg-lots gave normal and nearly normal tadpoles, one gave very abnormal
embryos. Two dishes, eA and eB, contained only 10 cc. sea water and
these gave very abnormal embryos, although the other two, eC and eD,
gave normal and abnormal tadpoles.

V. (Sept. 21, 1940.) This lot also came from the tank and had been
kept two days. The eggs were washed once ; then eight drops were
added to each Stender containing 20 cc. sea water, where they were ferti-
lized (5 drops). Four of the lots had some normal and bent tadpoles,
but even these dishes had, for the most part, abnormal embryos ; one lot
had only very abnormal tadpoles. It is not evident in this set whether
the great variability in each dish and between different dishes is due to
the condition of the eggs or to the presence of the sperm in the dishes.
One dish of selfed, eE, gave 113 short abnormal tadpoles, 12 late ab-
normal and 4 abnormal embryos ; the other selfed lots had nearly all
unfertilized eggs.

It is quite evident even from these few experiments that practically
100 per cent cleavage occurs if the eggs are washed and not too much
sperm added. Evidently these Cionas at Woods Hole behave in the
same way as those at Corona del Mar. The cleavage irregularities in
the 1904 and 1906 experiments must have been due to handling. It is
also clear here that despite the occurrence of the normal 2-cell stage, the
eggs of several of the individuals gave abnormal embryos and abnormal
tadpoles.

Some further tests were made on some of these Woods Hole Cionas
in order to study the effect of external conditions on development.

(Sept. 19, 1940.) The eggs were washed, then cross-fertilized (A
by b and B by a) by 5 drops of sperm suspension. All eggs in both
sets cleaved. In the two-cell stage the eggs (5 drops) were transferred
to 10, 20, 40 cc. of sea water. All sets gave abnormal embryos. These

CROSS- AND SELF-FERTILIZATION OF CIONA 343

results are clearly clue to the condition of the eggs and are nearly the
same as two of the lots in the 5 by 5 tests of the same date. The other
three lots of the same 5 by 5 test gave mostly normal tadpoles.

(Sept. 21, 1940.) Two small Cionas from the tank were used. The
eggs were washed once, and 10 drops were transferred to 10 and 20 cc.
sea water, where they were fertilized (5 drops). A by b, in 10 cc., gave
late abnormal embryos, and in 20 cc. 49 nearly normal tadpoles and 19
normals. B by a, in 10 cc., gave the same result as above ; and in 20
cc. there were 85 normals and 3 abnormal tadpoles. Here 20 cc. gave
distinctly better results than 10 cc. It will be noted that 5 drops of the
sperm suspension were left in each dish.

(Sept. 21, 1940.) The water was changed once on the eggs, then
they were cross-fertilized (5 drops). The supernatant water was then
largely removed and 8 drops of eggs added to each dish of 10 and 20 cc.
sea water. The A by b in 100 cc. sea water gave 95 per cent normals ;
and in 20 cc. gave 99.5. But B by a in 10 cc. gave 145 abnormal em-
bryos and in 20 cc. gave only 9 normal tadpoles, 6 bent and 109 abnormal
tadpoles. Here also the results were a little better when more water was
present, but there was a striking contrast between the reciprocal crosses.

In a repetition of the same experiment on the same date, A by b eggs
in 10 and in 20 cc. sea water gave abnormal tadpoles ; B by a eggs in 10
cc. gave abnormals as before, but in 20 cc. gave 393 normals, 109 twisted
tadpoles and 62 eggs or young embryos.

(Sept. 24, 1940.) The eggs were washed once. Ten drops of eggs
were then transferred to 20 cc. sea water in two dishes (a and a'). Here
they were fertilized with 5 drops of sperm suspension. After 15 min-
utes the water was removed from one dish (a') and replaced by new sea
water. All sets gave 100 per cent cleavage. The embryos in a died
young, but a' gave 6 nearly normal, 40 crooked tadpoles, and 76 very
abnormal tadpoles. Evidently the latter set went further than the
former (in which the sperm was left). The reciprocal cross, B by a,
gave somewhat better results. In b there were 11 normal, 12 crooked,
and 59 very abnormal tadpoles. In b' there were 14 normal, 14 crooked
and 49 very abnormal tadpoles. The results were the same in b and b',
and somewhat better than in a and a'.

It is very noticeable that with the better technique practically all the
cross-fertilized eggs in the set of experiments gave 100 per cent cleavage,
but nevertheless there was a high percentage of abnormal development.
The latter can only be ascribed to internal conditions in the eggs or
sperm, probably in the former. The Cionas were just reaching maturity,
i.e., about half normal size. The test was transparent and thin. It
seems safe then to ascribe the abnormal development to immaturity of

344 T. H. MORGAN

the animals although the eggs had every appearance of being normal.
There is here a rather sharp contrast between the normal behavior of the
eggs and sperm with respect to cross- and self-fertilization, and the
failure of many of the fertilized eggs to produce normal tadpoles.

Revieiv of Earlier Work

In the summer of 1903 (see Jour. Exper. Zool., 1904) I made four
5 by 5 experiments on Cionas at Woods Hole, Mass. In one of them
practically all the cross-fertilized eggs segmented. In another the sperm
completely failed in two cases and practically in a third (giving 30 per
cent in one case), although the eggs were good as shown in the other
two cases. In a third test the sperm was not very good (except in one
case). In a fourth test the sperm was not good in one case (except with
one set of eggs where it gave 75 per cent cleavage). It seems now evi-
dent that the failure to cross-fertilize was due to the sperm suspension,
and not to the conditions under which the tests were made.

Six years later (1909. See Roux's Archiv., 1910) abundant mate-
rial was available at Woods Hole and fifteen 4 by 4 tests were carried
out (180 crosses in all). Of these only two gave uniformly good re-
sults. In some, the four lots of eggs were obviously poor, in others the
sperm, but in all of them (except two) there were inconsistent results as
though certain combinations were incompatible, but in the light of the
results with the California type of Ciona it seems highly improbable that
these failures to cross-fertilize (i.e., to cleave) were due to the presence
of individuals with identical genetic composition, and there is no evi-
dence that the results were due to differences in the environment.

In 1932 (Sept. 19) a few large Cionas were brought to me at Woods
Hole. I tried out one 5 by 5 experiment for the cleavage. All cross-
fertilized eggs gave nearly 100 per cent (one 95 per cent). The selfed
lots gave 0, 0, 8, 0, 4 per cent. This is the same kind of result as with
the California type. Some other eggs were treated with crab juice (Cal-
lincctes) for four hours, and then selfed. They gave practically 100
per cent cleavage. When treated with acid sea water and selfed they
gave 100 per cent cleavage. In both respects these eggs agree with the
California Cionas. It is noticeable, in comparison with the later (1940)
experiments at Woods Hole on small immature Cionas, that these Cionas
were large, and the results were uniform and good.

In the summer of 1904 I spent a couple of weeks at Coronado Beach,
California, at the abandoned yacht club that had been used earlier as a
marine station of the University of California. There were quantities
of large Cionas on the float at the station. The laboratory room was

CROSS- AND SELF-FERTILIZATION OF CIONA 345

very warm during the daytime and the glassware insufficient. Only
small saltcellars were available for the eggs, etc. The results of eight
5 by 5 tests were published in 1905. On the whole, the results for
cleavage were poor and irregular. There were at least 17 cases where
the sperm gave no cross-fertilization or very little, and almost as many
cases where the eggs were poor, in the sense that they did not cleave
although they had every appearance of normal eggs. That the heat in
the laboratory was not the cause of the failure to cleave was evident since
in every set there were cases of normal (100 per cent) cleavage, but the
heat may account for the almost entire failure to give normal embryos.
The dishes were treated uniformly, although owing to the abundance
of sperm, too much may have been used, which, while not affecting the
cleavage, would affect the development. The only explanation I can
offer is that the temperature of the water in the basin around the float
had injured the reproductive cells before the Cionas were removed.

The Causes of Variability of Normal and Abnormal Development

From March to the end of June and again during October and
January, 1940-41, I repeated, with improvements in the technique,
some of the earlier experiments that had been made to find what con-
ditions determine whether normal or abnormal development takes place,
whether internal or external.

It had become evident from previous experiments that better, i.e.,
more uniform results, take place if the eggs are thoroughly washed,
then fertilized with a few drops of sperm suspension, most of the super-
natant fluid removed, fresh water added, and, after the eggs have settled,
most of this water also removed leaving only enough to supply a drop
of eggs to each of the dishes in which the eggs are to develop. An
experiment of this kind was made (Oct. 19) at Corona del Mar with
Cionas freshly brought in. The eggs were washed, and 10 cc. of fresh
sea water added. They were cross-fertilized by 5 drops of sperm sus-
pension, and after 10 minutes the supernatant fluid was drawn off
(except 20 drops). One drop of these fertilized eggs was added to
10 cc. sea water (in each of 10 dishes). The eggs were brought back
to Pasadena in closed Stender dishes, except the last four that were
brought back in closed vials. A by b gave the following figures :

94, 97, 99, 93, 99, 99, 99, 99.5, 99.5, 99.5 per cent tadpoles
The reciprocals, B by a, gave a much smaller percentage of normals :

49, 46, 28, 24, 3, 59, 38, 31, 35 per cent tadpoles
The latter figures are less than half the former and there is more vari-

346 T. H. MORGAN

ability, although the external conditions were made as like as possible.
The differences in the two cases seem to depend on the eggs or on the
sperm (or both). The amount of sperm carried over with one drop
of fertilized eggs must have been too small to affect the results, even
if there was some initial difference in the sperm suspensions.

The next day (Oct. 20) the same experiment was carried out with
fresh eggs from the adult Cionas that had been brought to Pasadena.
The results were very uniform, giving 99 per cent throughout (except
one, 95 per cent). Repeated four times (Oct. 20-21) all gave about
100 per cent normals.

In order to test further (Nov. 3) whether the transportation of seg-
mented eggs in closed Stender dishes gives the same results as do those
left at Corona del Mar, since in one case not reported here some of the
former were abnormal, the eggs of fresh Cionas were washed twice,
then fertilized, and after 15 minutes the supernatant fluid was taken off
(except about 20 drops). Then one drop of these fertilized eggs was
added to each of 20 Stender dishes (20 cc.). Ten of these dishes,
A by b, were taken (after 4 hours) to Pasadena. The percentages of
normal embryos were: 52, 74, 80, 95, 91, 52, 96, O,1 55, 80 per cent;
average 75.0 per cent. The other ten dishes left at the shore gave : 100,
99.5, 100, 99.5, 99.5, O,1 60, 100, 70, 100 per cent ; average 92.1 per cent.
In each set there was one dish that gave no normals ; both had dirt in
them and were disregarded. Those left at the shore had a higher level
of normals, which is probably not significant. Difference in temperature
was not involved here, since the car was cool. Six dishes of the re-
ciprocals, B X a, were also tested in the same way. Those taken to
Pasadena gave : 93, 87, 87, 93, 92 per cent. These average a little better
than the corresponding A by b. Those left at Corona del Mar gave :
100, 99, 99, 99.5, 93 per cent. The differences are probably not sig-
nificant since the percentages are estimates only.

The following experiment was made with Cionas (Nov. 11) that
had been kept for two days (with change of water). The eggs were
washed twice and all fertilized by five drops of sperm suspension.
After 15 minutes the supernatant water was changed, and one drop of
eggs put into each of ten Stender dishes containing 15 cc. sea water.
The reciprocal test was treated in the same way. The percentages in
the first dishes were: 68, 69, 69, 73, 79, 80, 70, 78, 76, 80 per cent.
The abnormals were of two kinds, viz., late abnormal tadpoles and early
abnormal embryos in about equal numbers. The reciprocal ten dishes
gave : 96, 97, 97, 96, 96, 98, 97, 99, 97, 99 per cent. The percentages

1 All were late abnormal tadpoles. The contrast with the others is not so
great as it appears to be.

CROSS- AND SELF-FERTILIZATION OF CIONA 347

are higher here than in the other ten dishes, and there is less variation.
Many of the eggs (4224, and 4624) were left over in the original
Syracuse dishes containing little water. Despite the small amount of
water and the large number of eggs, the percentages (A by fr==80 per
cent, B by a = =98 per cent) were not very different from those above.
The Cionas used in the next experiment had been brought to Pasa-
dena (Oct. 12). Two days later (Oct. 14) the eggs of one of them
were distributed in 10 dishes (10 cc. sea water), fertilized by sperm
of another individual (one drop). The sperm was left with the eggs
in this case, but only one drop to each dish. The actual counts were:

Normal Tadpoles 163 138 196 188 65 43 89 104 108 80

Abnormal tadpoles: 10 2 2 2 4 1 1 1

Abnormal embryos: 99 112 137 157 55 86 87 102 71 75

Counting both kinds of abnormals together, the percentages of normals
are 60, 55, 58, 54, 52, 33, 50, 50, 60, 52. The variability is not much,
but the percentages of normals are low.

The reverse (not reciprocal) test, with eggs that came from the same
individual that supplied the sperm above, fertilized by sperm from one
other dona, was much the same, but somewhat better :

Normal Tadpoles 45 78 50 65 102 70 100 124 59 165

Abnormal tadpoles: 29 12 12 2 8 29 11 13 30 16

Abnormal embryos: 15 13 88 21 22 18 22 16 30

The percentages are 50, 75, 71, 87, 78, 58, 78, 78, 56, 78. The per-
centages of normals are low in both these tests, and not obviously
environmental.

Since the experiments in which reciprocals are tested often show
differences in the proportions of normal embryos, some further tests
(Jan. 10 and 14) were made in which the eggs were first washed in
two changes of water, and, after fertilization, the supernatant fluid was
changed twice. At the two-cell stage one drop of eggs was added
either to 10 cc. in a Syracuse dish or to 20 cc. in a Stender. The dishes
had been washed with a weak hydrochloric acid solution, put into
running tap water for two hours and then rinsed in distilled water.
In one case ( 10 cc.) A by b gave 100 per cent ; B by a 80 per cent. In
another case (10 cc.) A by b gave 90 per cent and B by a 80 per cent.
There were two dishes of each which gave the same results. Also the
many left-over eggs in each case gave approximately the same ratios.
Furthermore, some of the unfertilized eggs kept in another dish and
fertilized by sperm of a third individual gave about the same kind of
results. The same statement may be made for two other reciprocal
pairs in 20 cc. sea water. These results are in line with previously re-

348 T. H. MORGAN

corded cases showing often different proportions of normal tadpoles in
reciprocals.

Finally another test (Jan. 13) of the same kind with the same pre-
liminary precautions was made in which ten similar dishes (20 cc.) were
made up of the cross and ten of its reciprocal. This experiment was
made as a further check on different dishes of the same sort which
should give more precise confirmation as to the reliability of the experi-
ments. The actual counts are given in Table I. In the first column
the number of normal tadpoles is given ; in the second the number of
abnormal tadpoles which were crooked or twisted ; and in the third the
abnormal embryos that had died at an early stage without evident
differentiation.2

TABLE I

A by b B by a

Abnormal Abnormal Per- Abnormal Abnormal Per-

Normal Tadpoles Embryos centage Normal Tadpoles Embryos centage

247

138

275

137

167

184

332

102

205

0.5

111

Several of the ratios in Table I call for comment. The ratios are
made up of the normals against the two kinds of abnormals taken to-
gether. Except for 8 and 10 in A by b, where the ratios are 37 and
30 per cent, the other ratios are nearly the same. In these two the low
ratio is due to an excess of " abnormal tadpoles." Again, in B by a
the ratios are nearly the same, except for 1, 3 and 8. Here also the low
ratios are due to excess of abnormal tadpoles and not to younger stages,
i.e. " abnormal embryos." Number 8 had only one normal tadpole.
The total average for A by b is 77.2 and for B by a 76.9. If number 8
is eliminated, the difference between the two crosses is small. I am in-

- The left-over eggs (about 2004) of A by b gave about 99.5 per cent normal
and 5 per cent early embryos, and B by a (about 700 eggs) gave about 80 per cent
normals, 15 per cent abnormal tadpoles and 5 per cent abnormal embryos. These
results are about the same as those in Table I, where fewer eggs and much more
water were present. In addition a few (150 and 100) eggs of a were fertilized
each by sperm from a third individual, and gave, respectively, C by a about 95
per cent normal and C by b 95 per cent normals, which are as good percentages as
the best in the table.

CROSS- AND SELF-FERTILIZATION OF CIONA 349

clined to think that the " abnormal embryos " are due to internal factors.
Polyspermy is rare and only a part of this kind of embryo can be as-
signed to it; the others are probably due to failure to cleave normally
in an early stage. On the other hand, most of the " abnormal tadpoles "
seem to be due to environmental factors, especially the extreme cases.
The individuals of this group have for the most part developed nearly
normally to a late stage and their defects are due to failure to straighten
out at the time of hatching. Nevertheless, since in the less extreme cases
of low ratios the large majority of the eggs have developed normally,
the environmental factors must have been nearly normal, which would
mean that some of the eggs were more sensitive to outside agents than
are others. This conclusion is borne out in some of the earlier experi-
ments where more normals were present when the eggs developed in a
larger volume of sea water, which would tend to dilute those external
agents that act injuriously. In the earlier experiments, where reciprocal
crosses were made, differences in the sperm suspension was probably
one of the factors that would account for the difference between the cross
and its reciprocal, but in these later experiments that factor is prac-
tically eliminated by washing the eggs and removing the supernatant
sperm suspension. The eggs were thoroughly mixed and one drop
only of the eggs was added to each 10 or 20 cc. of sea water. It is un-
likely then that the differences in the dishes of the same composition are
due to factors of this kind, and can only be ascribed to differences in
the dishes themselves. A number of further tests of reciprocals were
made which need not be recorded here, but which confirm the conclusion
that reciprocals may consistently give different ratios which can only be
due to internal factors in the eggs. Since no definite ratios appear in
the data, it is highly probable that there is no specific lethal genetic factor
involved.

The Development of Eggs from Different Levels of the Oviduct

In order to find out whether eggs from different regions of the ovi-
duct give different percentages of normals, the following experiments
were made. A very large Ciona, freshly collected (Nov. 23, 1940),
was opened and the long oviduct was tied off by two ligatures into three
sections, one near the outlet, one in the middle, and one next to the
ovary. The first section would be expected to contain the oldest eggs,
and the section next the ovary the eggs most recently set free from the
ovary. Each section was opened separately. The eggs were washed
twice, cross-fertilized with five drops of sperm suspension, which was
largely removed when the eggs had reached the two-cell stage. Fresh
water was added and then drawn off and replaced by fresh water ( 10

350 T. H. MORGAN

cc.). The first and second sections gave 100 per cent normal tadpoles.
There were about 2,000 eggs in each dish. The third section (nearest
the ovary) gave 85 per cent normals and 15 per cent bent and coiled
tadpoles ; 800 eggs in all. These differences are not significant as later
results showed.

The experiment was repeated the next day (Nov. 24. 1940) on
Cionas brought to Pasadena. The first and second (bracketed) are
reciprocals as are also the third and fourth.

Section 1 Section 2 Section 3

[80 90 99

199 70 70

99.5 100 100

99.9 100 100

There were from 800 to 1200 eggs in a dish (Syracuse) in only 10 cc.
water, yet the percentage of normals was very high. There is no ap-
parent difference in the percentages from the three different regions. It
follows that different percentages of normals and abnormals in other
experiments, when the eggs from the oviduct were used, are not due to
age differences in the oviducal eggs.

The Cleanliness of the Glassivare

The failure to get uniform results relating to normal development
when the conditions of the experiments have been made as uniform as
possible drew attention finally to the cleanliness of the glassware. As a
rule the dishes were used for no other purpose, and were washed in tap
water after each experiment, and drained, partly inverted, and not used
again until dry. It was noticed that whenever drops of water had re-
mained sticking to the glass and evaporated there a slight residue or stain
was left. There was not sufficient reason to suppose such a minute
amount of salts could affect the results, but, as a check, the dishes in one
test were put into cleaning fluid, washed in running tap water and rinsed
in distilled water, and used at once (Dec. 1). In four sets over 95 per
cent normal tadpoles developed, but in one set the tadpoles failed to
come out of the membrane. Clearly all the cleaning fluid had not been
removed in this set. As a check the experiment was repeated (Dec. 2),
omitting the cleaning fluid, and washing with distilled water and the
double distilled water used in the last experiment. Normal tadpoles de-
veloped, 95 to 100 per cent. Finally, 10 dishes were scrubbed with
trisodium phosphate, washed thoroughly in running water and drained.
Washed eggs in the two-cell stage were added to 10 cc. sea water in
Stender dishes. All gave late abnormal embryos. There were eight

CROSS- AND SELF-FERTILIZATION OF CIONA 351

control dishes washed only in tap water which gave normals (most of
them 100 per cent). When the same kind of experiment was repeated,
but, after scrubbing the dishes in trisodium phosphate, they were washed,
rinsed in 10 per cent HC1, and washed again, they gave 90 to 100 per
cent normals, as did the controls washed in tap water, as also did some
other dishes rinsed in HC1, and washed in tap water. As a further
check other dishes were scrubbed in the trisodium phosphate and very
thoroughly washed in running tap water. These also gave 99 to 100 per
cent normals. It is evident from these experiments that the developing
embryos are very sensitive to even traces of the chemicals used here in
cleaning the dishes. The ordinary procedure of washing in tap water
gives the best results, as a rule. Any salts that happen to remain as
stains do not interfere with normal development. In fact, unless the
utmost care is taken in removing the chemical material used to clean the
dishes, there is more risk of causing abnormal development than from
ordinary tap water.

Summary

A repetition of some of the earlier work on Cionas at Woods Hole
(in 1904, 1905, 1910), with improvements in the technique that the
later work at Corona del Mar had shown to be important, has made it
clear, so far as cross-fertility and self-sterility are involved, that the
Woods Hole type gives the same results as the California type. The
very eccentric results concerning normal and abnormal development
shown in the earlier experiments appeared again, and were found not to
be due entirely to differences in technique, but to differences in the eggs
of the Cionas themselves, connected, in part at least, with immaturity of
the animals, even when their eggs and sperm appeared to be normal,
and when the eggs cleaved normally, at least into two cells.

Some of the later work, carried out at Corona del Mar in 1940 when
greater care was taken in removing the egg-water (by washing the eggs)
and also in removing the excess of the sperm suspension, is reported.
In most cases the eggs were fertilized en masse, and, after washing again
and removing the excess of water, a drop or two only of the eggs was
added to the sea water (10 to 20 cc.) in Stender dishes. The reciprocal
cross was made in the same way. It was found that there is a marked
tendency for all the dishes of the same cross to give closely the same
percentages of normal tadpoles, but there were occasional dishes that
gave more extreme variations (usually more abnormals). These are due
to environmental factors. The cross, when compared with its reciprocal,
frequently gives different percentages of normal development when the
external conditions (water and dishes) are as nearly the same as possible.

352 T. H. MORGAN

Evidently, then, the eggs of different individuals in reciprocal crosses
may give different percentages, although from a genetic (chromosomal)
point of view the two are, on the average, identical after fertilization.
It seems to follow, then, that this difference must lie in the cytoplasm
of the eggs which in each case has been formed under the diploid condi-
tion of the eggs. There is also other evidence supporting such an inter-
pretation. It should be pointed out, however, that, as a rule, the per-
centages of the two reciprocals are the same or nearly so. When they
differ more markedly there are no definite ratios between them, so far
as the observations go.

A number of experiments were also made to test whether differences
in the dishes, used in the experiments after washing in tap water and
drying, are responsible for the variability sometimes found in the same
series. There is no evidence that this is the case if the dishes have been
carefully washed and drained. On the other hand, if they have been
cleaned by the ordinary chemical treatments there is evidence of effects
on the development unless great care is taken to remove every trace of
the cleaning fluid.

BIBLIOGRAPHY

EAST, E. M., 1934a. Norms of pollen-tube growth in incompatible matings of self-
sterile plants. Proc. Nat. Acad. Sci., 20 : 225-230.
— , 19346. The reaction of the stigmatic tissue against pollen-tube growth in

self-sterile plants. Proc. Nat. Acad. Sci.. 20 : 364-368.
— , 1935. Genetic reactions in Nicotiana I. Compatibility. Genetics. 20 : 403-

413.
FUCHS, H. M., 1914. On the conditions of self-fertilization in Ciona. Arch. Entiv.-

mech., 40: 157-204.
LAUG, EDWIN P., 1934. Retention of dichromate by glassware after exposure to

potassium dichromate cleaning solution. Jour. Ind. EIK/. Chcm. (Anal.

Ed.), 6: 111-112.
MORGAN, T. H., 1904. Self-fertilization induced by artificial means. Jour. Expcr.

Zoo/., 1 : 135-178.
— , 1905. Some further experiments on self-fertilization in Ciona. Biol. Bull., 8:

313-330.
— , 1910. Cross- and self-fertilization in Ciona intestinalis. Arch. Entw.-mcch.,

30 (2) : 206-235.
— , 1923. Removal of the block to self-fertilization in the ascidian Ciona. Proc.

Nat. Acad. Sci., 9: 170-171.
— , 1924. Self-fertility in Ciona in relation to cross-fertility. Jour. Expcr. Z067.,

40 : 301-305.
— , 1924. Dilution of sperm suspensions in relation to cross-fertilization in Ciona.

Jour. E.\-pcr. Zoo/., 40 : 307-310.
— , 1938a. The genetic and the physiological problems of self-sterility in Ciona.

I. Data on self- and cross-fertilization. Jour. Ex per. Zoo/., 78: 271-318.

— , 19386. The genetic and the physiological problems of self-sterility in Ciona.

II. The influence of substances in the egg water and sperm-suspensions in
self- and cross-fertilization in Ciona. Jour. E.rpcr. Zoo/., 78 : 319-334.

CROSS- AND SELF-FERTILIZATION OF CIONA 353

— , 1939. The genetic and the physiological problems of self-sterility in Ciona.

III. Induced self-fertilization. Jour. Expcr. Zool., 80: 19-54.

— , 1939. The genetic and the physiological problems of self-sterility in Ciona.

IV. Some biological aspects of fertilization. Jour. E.vpcr. Zool., 80: 55-80.
— , 1940. An interim report on cross- and self-fertilization in Ciona. Jour. Ex-

pcr Zool, 85 : 1-32.

PLOUGH, H. H., 1930. Complete elimination of self-sterility in the ascidian Styela

by fertilizing in alkaline solutions. Proc. Nat. Acad. Sci., 16 : 800-804.
— , 1932. Elimination of self-sterility in the Styela egg — a re-interpretation with
further experiments. Proc. Nat. Acad. Sci., 18: 131-135.

RICHARDS, OSCAR W., 1936. Killing organisms with chromium as from incom-
pletely washed bichromate-sulf uric-acid cleaned glassware. Physiol. Zool.,
9 : 246-253.

RELATION OF THE SIZE OF " HALVES " OF THE

ARBACIA PUNCTULATA EGG TO CEN-

TRIFUGAL FORCE

ETHEL BROWNE HARVEY

(From the Marine Biological Laboratory, Woods Hole, and the
Biological Laboratory, Princeton University)

The centrifugal force used in my previous work (1932-1940) to
obtain " halves " x of the egg of Arbacia pimctulata has been about
10,000 X g- With this force, the egg breaks into two halves : — a (light)
white half containing oil, nucleus, clear layer, mitochondrial layer and
some yolk; and a (heavy) red half containing most of the yolk and all
the red pigment granules. A small electric centrifuge was used, whose
highest speed is approximately 10,000 R.P.M., or 160 R.P.S. The
hematocrit tubes used to hold the material were 6.5 cm. long with an in-
side diameter of 0.35 cm. When placed in the arms of the centrifuge,
the radial distance from the axis of the centrifuge to the position of the
eggs while rotating was 10 cm. The centrifugal force is computed from
the equation

F = = .04XR X

where R = radius in cm. and the force is expressed in times gravity.

Approximately this same force ( 10,000 X g) has been used by others
to obtain the red and white halves for studies on permeability to water
(Lucke, 1932), respiration (Shapiro, 1935), indophenol oxidase (Navez
and Harvey, 1935), peptidase (Holter, 1936), and dehydrogenase (Bal-
lentine, 1940). The segregation of different materials into the two
halves has furnished a nice means of determining the location of various
cell activities.

In any one batch of eggs, the white halves (and the red halves) are
quite uniform in size. The relative size of the two halves throughout
many experiments has been remarkably uniform; the white half is
slightly greater in diameter than the red, and about one and one-third
times greater in volume, with the force of 10,000 X g at 23° C. This I
will call the standard force, since it has been so widely used.

Much greater centrifugal forces can be obtained with the air turbine.

1 The term " halves " is used incorrectly but purposely because there is no word
to express two, and only two, unequal fractions of a whole, constant in size. The
term " fragments " used by some writers implies variability in size, as well as in
number.

354

SIZE OF HALF-EGGS AND CENTRIFUGAL FORCE

355

The small Incite tubes used measure 1.4 cm. in length and have an inside
diameter of 0.3 cm. The radius from the center of the rotor to the
position of the eggs while rotating is 1.2 cm. The highest speed obtain-
able with the rotor used was 1,500 R.P.S., which gives a centrifugal
force of about 100,000 )< g. Somewhat lower forces are obtained by
using lower speeds of the air turbine ; the lowest speed of the air turbine
gives a slightly greater force than the highest speed of the electric centri-
fuge. A very low force can be obtained with a low speed of the electric
centrifuge. There is then, a range in the centrifugal force available of
from 4,000 X g, just sufficient to break the eggs, to 100,000 X g.

The relative size of the two half-eggs varies with the centrifugal
force used to break the egg in two. As mentioned above, with the stand-
ard force of 10,000 X g, the white half is one and one-third times the
volume of the red half. With a force of 60,000 X g, obtained by a

TABLE I

Arbacia punctulata. Size and force

Diameter

Volume

(M3)

Ratio

Whole egg

212,000

(approx.)

Minutes
to break

Force

(Xg)

W: R

4,000

180,000

36,000

5 : 1

10,000

125,000

92,000

4 : 3

60,000

107,000

1 : 1

3/4

80,000

92,000

125,000

3 :4

1/2

100,000

36,000

180,000

1 : 5

medium speed of the air turbine, the two halves are equal in size. With
a force of 80,000 X g, the size of the two halves is the reverse of that
obtained with the standard force, the red half being now one and one-
third times the volume of the white. With the greatest force obtainable.
100,000 X g, the red half is five times the volume of the white. The
white half is therefore very small and contains only the nucleus, some
oil and a little of the clear matrix. Correspondingly, with the lowest
force that will break the egg in two, 4,000 X g, obtained with a low
speed of the electric centrifuge, the white half is five times the volume
of the red. The greater the force, the larger the (heavy) red half and
the smaller the (light) white half.

The time required to break the eggs apart at the different forces is,
of course, different ; it requires 20 minutes with the lowest force, and
only one-half a minute with the highest force. In Table I will be found
the forces used, the sizes of the two halves, their approximate ratio, and

356 ETHEL BROWNE HARVEY

the time necessary to break approximately one-half the eggs into the two
parts at 23° C. With the very high speeds of the air turbine, some in-
accuracy and variability obtain, owing to the fact that the turbine must
be speeded up and slowed down somewhat gradually, and the interval
of full speed is so short (% min.).

Together with the difference in size of the two halves with different
forces, there is also a difference in the degree of stratification of the egg
just before breaking. With a low force applied for a long period — 20
minutes — the granules are entirely segregated into their respective layers
according to their density, and well packed. With a somewhat higher
force, the standard force, where 4-5 minutes are necessary to break the
eggs apart, the granules are segregated into well-defined layers, but they
are not so well packed, especially the pigment granules. With the higher
forces of the air turbine, requiring only % to one minute, the egg breaks
apart before the granules are entirely segregated so that there is no very
definite stratification. Photographs 1-10 show the whole eggs centri-
fuged with different forces, just before breaking in two, and the two
halves into which they break. With the greatest force, the materials are
so poorly segregated (Photograph 5) that the red half contains some
of all the materials present in the whole egg except the nucleus. By
re-centrifuging the red half at the standard force, a new stratification is
obtained just like that of the whole egg at this force (Photograph 12;
cf. Photograph 2). A similar red half obtained with the standard force,
has, when re-centrifuged with this same force, no oil or mitochondrial
layer, these materials having been completely segregated out into the
other half in the first centrifuging (Photograph 11).

It will also be noted that there is less elongation of the egg prior to
breaking, with the higher forces.

If centrifuged slowly at first with a low speed of the electric cen-
trifuge until well stratified, and then transferred to the air turbine and

PLATE I

EXPLANATION OF FIGURES
(Magnification, 275 X)

PHOTOGRAPHS 1-5. Stratification of Arbacia eggs at different centrifugal
forces, just prior to breaking in two.

PHOTOGRAPHS 6-10. The two halves into which the eggs break at different
forces.

PHOTOGRAPH 11. A red half obtained with 10,000 X g (Photograph 7) re-
centrifuged at 10,000 X g. Control to Photograph 12.

PHOTOGRAPH 12. A red half obtained with 100,000 X g (Photograph 10) re-
centrifuged at 10,000 X g. Note re-stratification similar to original egg at this
force (Photograph 2) ; nucleus, of course, absent. Mitochondrial layer stained
with methyl green.

SIZE OF HALF-EGGS AND CENTRIFUGAL FORCE

357

4,000 x g

10,000 x g

60,000 x g

80,000 x g

100,000 x g

PLATE I

ETHEL BROWNE HARVEY

centrifuged rapidly, the eggs break in just the same way as though they
were kept at the low speed; that is, the white half is much larger than
the red.

As has been noted previously (1936), several irregular batches of
eggs occur each summer in which the break is quite different from the
ordinary though the eggs themselves appear no different. In these eggs,
at the standard force, the white halves are very large, and the red halves
are very small (Photograph 16). The red halves (D == 28.5 /x; Vol.
= 12,000 /JLS) are only about one-seventeenth the volume of the white,
considerably smaller than those obtained from normal lots with the
lowest force capable of breaking the egg, and even smaller than the
pigment quarters obtained by breaking apart the usual red half with the
standard- force (Harvey, 1936, p. 103). These small red halves contain
little besides the pigment granules which have been determined by E. N.
Harvey (1932) to form 5.5 per cent of the materials in the egg. With
a medium speed of the air turbine (60,000 )< g) the red halves of the
irregular batches are much larger, but still smaller than the whites ; the
two halves are of about the same size as those from ordinary batches
at the standard force (Photograph 17, cf. 13). With the highest speed
of the air turbine (100,000 >( g), the two halves are about equal (Photo-
graph 18, cf . 14) . These irregular batches, then, show the same increase
in size of the red halves with increase in centrifugal force as the normal
batches, though starting from a different point.

It might be of interest to mention that these irregular batches can be
made to break into the normal sized halves with the standard force, by
keeping them in hypotonic sea water (80 per cent sea water, 20 per cent
distilled water) for an hour and then centrifuging them in a sugar-sea
water solution of the same tonicity.

The same general relationship between centrifugal force and the size
of the half-eggs was found previously (1933) for the Naples sea
urchins, Arbacia pustulosa, Sphaerechinus granularis and Paraccntrotns
livid us, within a much more limited range of forces then available. For
the egg of Parechinns microtuberculatns (from Naples) and Tripneustes

PLATE II

EXPLANATION OF FIGURES
(Magnification, 125 X)

PHOTOGRAPHS 13-15. Normal eggs broken at different forces. Control to
Photographs 16-18. The slightly larger sphere at the middle right in Photograph
15 is a whole egg, the others are halves.

PHOTOGRAPHS 16-18. Irregular batch broken at same forces as 13-15.

Note uniformity in size of halves at any one force. There is usually a slight
irregularity in normal batches at the highest force (Photograph 15).

SIZE OF HALF-EGGS AND CENTRIFUGAL FORCE 359

Normal

Irregular

w *
^

* m

10,000 x g

100,000 x g

PLATE II

360 ETHEL BROWXE HARVEY

csciilcnta- (from Bermuda), the reverse relation was found, i.e. the
greater the force, the larger the light half and the smaller the heavy half.
In these eggs, the yolk granules are lighter than the matrix and lie
under the oil, and the clear layer is formed more centrifugally ; this must
be the explanation of the apparent inconsistency in the relative size of
the half -eggs of these two species. Whenever the granules are well
segregated (low force) so that there is a large clear layer, the sphere
with this layer is larger. Conversely, when the egg pulls apart before
complete segregation of granules (high force), the clear layer is small,
and the sphere containing this layer is smaller.3

The fact that one can vary at will the relative sizes of the two half-
eggs and their composition by simply changing the centrifugal force
gives a new tool for investigating the properties and functions of the
different constituents of the egg. The principle may well be applied to
the study of other marine eggs, and to other types of cells. For any
accurate chemical work on, the two half-eggs, the centrifugal force by
which they are obtained must be given careful consideration. For com-
plete segregation of materials, a low force should be used for a long
period, even though one may actually break the egg in two subsequently
with a high force.

Development

The small white fragments obtained with very high forces will de-
velop when fertilized, if not too small. The smallest to develop were
32 n in diameter, having a volume of only 7 per cent of the whole egg.
Some of these formed skeletons though they did not become perfect
plutei. The early cleavages of the small white fragments are quite
regular.

The large red halves obtained with the high centrifugal forces cleave

2 Data in original notes, 1932 (not published) for Tripneustes cscitlcnta:
At 3,000 X g for ft hr. Light half, D. = 76 M Vol. = 230,000 /a3

Heavy half, D. = 51

At 7,000 X g for M hr. Light half, D. = 79

Heavy half, D. = 43

Vol. = 69,500 M
Vol. = 258,000 /*•
Vol. = 41,600M:

The whole egg, D. = 83 /* ; Vol. = 300,000 M3.

3 Recent studies at Pacific Grove show that Strongyloccntrotus franclscanus
is stratified like Par echinus and Tripncustcs, with the yolk granules in the light
half ; this half increases in size with increased force. In Strongyloccntrotus pur-
purattts, two types of eggs occur, sometimes in the same batch, one type with the
clear layer in the light half (like Arbacia), and the other type with the clear layer
in the heavy half (like 5". franciscainis). The heavy half becomes larger with in-
creased force when it is granular, and the light half becomes larger with increased
force when it is granular. There seems no doubt, therefore, that the distribution
of the granules determines the size of the halves in relation to the centrifugal
force; the more granular half becomes larger with a greater force.

SIZE OF HALF-EGGS AND CENTRIFUGAL FORCE 361

more regularly and develop very much better, both fertilized and par-
thenogenetic, than those previously obtained with lower forces. This is
surprising in a way, because one might suppose that these tremendous
forces would completely disorganize the cytoplasm. It has, however,
been found by Beams and King (1936) that Ascaris suum eggs could
be centrifuged at 400,000 X g for an hour in the 2- and 4-cell stage and
90 per cent of these would still continue to develop when removed ; and
they would develop also after being centrifuged at 150,000 >( g for 4%
days. They would even cleave while being centrifuged at 100,000 >< g.
These eggs are, of course, protected from disruption by a very heavy
shell. The red halves of Arbacia are also apparently not disorganized
by these high forces, and they are almost as large as the whole egg
(Photograph 15) and contain some of each kind of material present in
the whole egg. Only the nucleus, a little of the oil and a little of the
clear matrix are lacking. This is, then, a nice technique for practically
separating the nucleus from the rest of the egg. So far, these non-
nucleate halves, when activated artificially — the parthenogenetic mero-
gones — have not developed further than the blastula. If such proves
finally to be the case, we may conclude that eggs can cleave and pass
through the early stages of development involving cell multiplication,
without nuclei, but that for the later stages involving differentiation,
nuclear material is necessary.

SUMMARY

1. Arbacia punctulata eggs have been broken into '" halves " with
centrifugal forces ranging from 4,000 X g to 100,000 X g, the higher
forces being obtained with the air turbine.

2. The relative size of the two halves varies with the centrifugal
force used; the higher the force the larger is the (heavy) red half, and
the smaller the (light) white half. With low forces, the white half is
larger than the red ; with high forces, the reverse holds.

3. The degree of stratification of the eggs just prior to breaking also
varies with the force used. With low forces, applied for a long period
(20 min.), the eggs are very well stratified. With high forces, the egg
breaks apart (%— 1 min.) before the materials are completely segregated
into layers.

4. The red half obtained with the highest force available (100,000
X g), is only slightly smaller than the whole egg and contains some of
all the materials in the original egg except the nucleus. It develops
much better, both fertilized (fertilized merogone) and parthenogenetic
(parthenogenetic 'merogone), than the red half obtained with lower
forces previously used.

362 ETHEL BROWNE HARVEY

LITERATURE CITED

BALLENTINE, R., 1940. Quantitative cytochemistry. The distribution of the re-
ducing systems in the egg of Arbacia punctulata. Jour. Cell, and Comp.
Physiol, 16: 39-47.

BEAMS, H. W., AND R. L. KING, 1936. Survival of Ascaris eggs after centrifug-
ing. Science, 84 : 138.

HARVEY, E. B., 1932. The development of half and quarter eggs of Arbacia punc-
tulata and of strongly centrifuged whole eggs. Biol. Bull., 62: 155-167.

HARVEY, E. B., 1933. Development of the parts of sea urchin eggs separated by
centrifugal force. Biol. Bull, 64: 125-148.

HARVEY, E. B., 1936. Parthenogenetic merogony or cleavage without nuclei in
Arbacia punctulata. Biol. Bull, 71 : 101-121.

HARVEY, E. B., 1940a. Development of half-eggs of Arbacia punctulata obtained
by centrifuging after fertilization, with special reference to parthenogenetic
merogony. Biol. Bull., 78 : 412-427.

HARVEY, E. B., 1940&. A comparison of the development of nucleate and non-
nucleate eggs of Arbacia punctulata. Biol. Bull., 79 : 166-187.

HARVEY, E. N., 1932. Physical and chemical constants of the egg of the sea
urchin, Arbacia punctulata. Biol. Bull., 62: 141-154.

HOLTER, H., 1936. Studies on enzymatic histochemistry. XVIII. Localization of
peptidase in marine ova. Jour. Cell, and Comp. Physiol., 8 : 179-200.

LUCRE, B., 1932. On osmotic behavior of living cell fragments. Jour. Cell, and
Comp. Physiol., 2 : 193-199.

NAVEZ, A. E., AND E. B. HARVEY, 1935. Indophenol oxidase activity in intact
and fragmented Arbacia eggs (abstract). Biol. Bull., 69: 342.

SHAPIRO, H., 1935. The respiration of fragments obtained by centrifuging the
egg of the sea urchin, Arbacia punctulata. Jour. Cell, and Comp. Phvsiol.,
6: 101-116.

EFFECTS OF ROENTGEN RADIATION ON THE JELLY OF

THE ARBACIA EGG a- -

T. C. EVANS, H. W. BEAMS AND MARSHALL E. SMITH

(From the Departments of Radiology and Zoology, State University of loiva,
and the Marine Biological Laboratory, Woods Hole)

Freshly shed Arbacia eggs are enclosed by a layer of transparent
jelly which can be easily demonstrated by adding a few drops of Janus
green B (Harvey, 1939) to the water containing the eggs. When the
Janus green is added, the jelly is first outlined, then is stained com-
pletely, eventually becomes brittle, and finally separates into long shreds
which settle to the bottom of the container. The jelly is slowly lost,
but when the eggs are kept in cool fresh sea water it can still be demon-
strated after several hours. The jelly affords some mechanical protec-
tion to the egg, contains sperm agglutinin (Lillie, 1915; Just, 1930; and
Tyler and Fox, 1939), and probably facilitates the preliminary stages of
fertilization. Lillie (1915), Woodward (1918) and Tyler and Fox
( 1940) have shown that the agglutinin is non-dialyzable. Glaser (1914)
and Woodward (1918) applied a number of common protein tests and
obtained no reaction except that of weak xanthroproteic test. Tyler
and Fox (1939, 1940) have strong evidence for the protein nature of
the sperm agglutinin of the sea-urchin Strongylocentrotus purpuratiis.
These investigators also found that the jelly was dissolved by the pro-
teinase, chymotrypsin. Tyler (1940) concluded that the agglutinin is
either the jelly itself or a component of it. Although the above evidence
is indirect as regards the jelly, it appears likely that it is protein in
nature. The effect of Roentgen radiation on the jelly is so striking
that we believe it is worthy of note even though little is known of the
chemical nature of the jelly or of the mechanism involved in this
phenomenon. We shall describe briefly the experiments from which the
following results have been obtained. The radiation, under conditions
described, removed the jelly from the eggs. Heavier dosages produced
a change in the staining reaction of the Arbacia jelly which had been
removed from the eggs by mechanical means.

1 Aided by a grant from the Rockefeller Foundation for Research on Physi-
ology of the Cell.

2 We wish to thank Drs. E. B. Harvey, J. H. Bodine, Gordon Marsh and
Jay A. Smith for suggestions concerning the experiments and the manuscript.
We also wish to thank Dr. E. P. Little for his kind cooperation in doing the
irradiations.

363

364 EVANS, BEAMS AND SMITH

MATERIALS AND METHODS

Unless otherwise specified, the eggs employed were those of Arbacia
punctulata, collected during June and July at Woods Hole. About 0.5
ml. of eggs in 10 ml. of solution was used for each sample and were
irradiated in covered dishes made of Turtox plastic. Before and after
the irradiation the eggs were kept in ringer bowls at room temperature.
For experiments involving jelly alone the eggs were allowed to stand for
a time in sea water, were then shaken, centrifuged gently, and the super-
natant fluid (containing the jelly) was poured off. The solution was
irradiated immediately in the plastic dishes and then poured into Pyrex
centrifuge tubes. Determinations of pH were done as soon as possible
after the irradiation and were usually completed within an hour after
the treatment. The pH measurements were made at room temperature
(22-27° C.) with a Leeds and Northrup potentiometer-electrometer No.
7660, equipped with a small modified Maclnnes glass electrode (Mac-
Innes and Belcher, 1933). The electrode was calibrated with " standard
acetate " for which the value pH 4.64 was taken (Maclnnes et al, 1938).

The radiation characteristics were: 192,000 volts; 20 ma.; 6,100
roentgens per minute ; target distance 9.5 cm. ; filters consisting of glass
walls of tubes and 5 mm. of bakelite (equivalent inherent filter equal to
0.2 mm. copper). Two tubes were used simultaneously. They were
opposed, parallel and self -rectifying.

The buffers used were : acetate at pH 4.0, 4.4, 4.9 and 5.3 ; phosphate
at pH 5.7 and 6.0 ; glycylglycine at pH 7.4, 9.0 and 9.4 ; and piperazine
at pH 10.2.

EXPERIMENTAL RESULTS

Removal of the Jelly from the Eggs by the Radiation

Relation betiveen Dosage and Loss of Jelly. — The first experiments
served to confirm the results reported in preliminary notes (Evans and
Beams, 1939; and Evans, 1940). As the dosage of radiation was in-
creased, the halo of jelly around the egg became progressively thinner
until it was entirely gone. The relation between the dosage and number
of eggs completely denuded of jelly is shown in Fig. 1, curve 1. The
individual eggs varied as to amount of jelly present at the beginning of
the experiments and this probably explains why some eggs were denuded
by low dosages whereas others required heavier irradiation. Curve 2 of
Fig. 1 shows the results of a similar experiment with the eggs of Asterias
forbesii, and it can be seen that this jelly is more resistant than that of
the Arbacia egg. The dosages required to remove the jelly from all of
the Arbacia and Asterias eggs are high enough to produce abnormal de-
velopment subsequent to fertilization. Roentgen radiation, therefore,

X-RAYS ON JELLY OF ARBACIA EGG

365

is not suggested as a means of removing the jelly from eggs in which
normal development is desired.

Is the Jelly Affected Indirectly by the Radiation? — A number of ex-
periments have been performed in an effort to determine whether the
effect on the jelly was produced directly by the radiation or indirectly
through the action on some other substance in the medium. The fol-
lowing results seem to indicate that the effect was due directly to the
radiation. (1) There was no apparent latent period. (2) The removal
action was not prevented by low temperature (0° C.). (3) The jelly
was removed by the radiation in solutions buffered at pH 7.4, 9.1, and
9.6. (4) Eggs were placed in heavily irradiated sea water and in water
from eggs previously irradiated, and in both cases the jelly remained

100

0 5 10 15 20 35
DOSAGE IN 1000 ROENTGENS

FIG. 1. A graph showing the relation between the dosage of radiation and the
percentage of eggs devoid of jelly. Curve one is drawn through the values found
in nine experiments on eggs of Arbacia pnnctulata. Curve two is drawn through
the values found in one experiment on eggs of Astcrias forbcsii.

intact. (5) In two experiments female Arbacia were irradiated, and it
was found that the eggs shed later were devoid of jelly.

Action of the Radiation on the Jelly after its Removal from the Egg

At the suggestion of Dr. Robert Chambers, we attempted to study
the effect of the radiation on the jelly after its removal from the egg.
It was found if the jelly was removed from the eggs, that upon addition
of a few drops of 2 per cent Janus green the jelly would react vigorously
with the stain and would accumulate into one large clump at the surface
of the water. As is shown in Table I, the dosages required to produce
a negative test are greater than that necessary to remove the jelly from
the eggs.

366

EVANS, BEAMS AND SMITH

Effect of pH on the Jelly. — Inasmuch as acid sea water has been used
to remove the jelly from the eggs (Harvey, 1939; Just, 1939; Tyler,
1940; Tyler and Fox, 1940), it seemed of interest to determine whether
the radiation might increase the acidity of the egg medium enough to
account for the removal of the jelly. One of the first steps in this in-
vestigation was to determine what pH range was effective in removing
the jelly. It was found that at a pH of 4.9, or below, nearly all of the
eggs lost their jelly within five minutes. Jelly alone was placed in solu-
tions buffered at different pH values and it was found that at pH 4.0
the Janus green test was negative, and at pH 4.9 it was very faint, but

8 0

7 5

7 0

X
Q_

6 5

6 0

0 20 40 60 80 100
DOSAGE IN 1000 ROENTGENS

I 20

FIG. 2. Effects of Roentgen radiation on the pH of certain media. Curve
one — jelly in medium buffered at pH 7.58. Curve two — jelly in sea water. Curve
three— jelly in NaCl-KCl mixture. Curve four — eggs alone in the NaCl-KCl
mixture. Curve five — supernatant fluid from cytolyzed eggs.

at pH 6.0 to pH 10.2 the test was strongly positive. These results indi-
cate that a marked change in the properties of the jelly is brought about
by solutions whose pH values are as low as 4.0. This finding is in
accord with that of Tyler (1940), who used solutions of pH 3.5 to
dissolve the jelly.

Effect of the Radiation on tJie pH of the Solution. — Sea water was
irradiated and in two experiments where the dosage was 121, 000 r the
maximum lowering was only 0.06 of a pH unit. Fresh eggs placed in
such irradiated water retained the jelly layer intact.

Irradiation of freshly-shed eggs in sea water apparently produced
some acidity in the medium. In a representative experiment the pH

X-RAYS ON JELLY OF ARBACIA EGG 367

was changed from 7.9 to 7.3 by an irradiation of 157,500 r. Conditions
which were found to alter the pH change were : ( 1) amount of radiation,
(2) length of time eggs had been in sea water before irradiation, (3)
length of time eggs had been in sea water after the irradiation and
before the determination of the pH, (4) mechanical injury to the eggs
during the preparation of the supernatant fluid for pH determination,
and (5) original concentration of eggs (and jelly).

Irradiation of eggs (with jelly) in unbuffered solutions of 19 parts
0.52 M NaCl to 1 part 0.53 M KC1 produced a marked lowering of the
pH. In a typical experiment the pH was lowered from 6.6 to 6.2 by an
irradiation of 122,000 r.

It appears that irradiation of either eggs or jelly produces some
acidity. As indicated in curve 3 of Fig. 2, when jelly alone was irradi-
ated (61,000 r) in the NaCl-KCl mixture the pH was lowered about 0.4
units. The same dosage on jelly in sea water (curve 2) lowered the pH
about 0.3 units. When eggs were separated from the jelly and irradi-
ated in the NaCl-KCl solution (curve 4) again some acidity was pro-
duced in the supernatant fluid.

It is interesting to note, in connection with the question of the
mechanism of the radiation action on the jelly, that a dosage sufficient to
cause the negative Janus green reaction did not affect the pH of the
buffered solution (curve 1). This indicates that a pH change is not
necessary for the radiation action on the jelly.

Chambers and Pollak (1927), Pandit and Chambers (1932), and
Krahl and Clowes (1938) agree that injury to the starfish and Arbacia
egg lowers the pH which is usually changed from about 6.8 to 5.3. The
present writers cytolyzed eggs by shaking them vigorously in various
dilutions of sea water with distilled water and the lowest observed pH
was 5.5. When the initial pH was this low, as shown in curve 5 of
Fig. 2, the radiation produced no further shift toward the acid side.

The greatest change in pH produced by irradiation was obtained by
adding the supernatant fluid from irradiated eggs (with jelly), in the
NaCl— KC1 solution, to a fresh lot of eggs and repeating the radiation
treatment. In one experiment where the addition of fresh eggs and
jelly was done twice, the pH was lowered by the radiation (total of
274,500 r) from 7.9 to 7.2. It is interesting to note here that the lower-
ing of the pH by the radiation was not enough to remove the jelly from
freshly-shed eggs. These findings are in agreement with those pub-
lished in a preliminary report (Smith and Evans, 1940).

Effect of the Radiation on the Sperm-agglutination Property of the
Jelly. — Richards and Woodward (1915) destroyed the sperm agglutinin
of Arbacia egg-water by means of Roentgen radiation. We performed

368 EVANS, BEAMS AND SMITH

some experiments to determine whether the radiation would affect the
sperm-agglutination property of the jelly in the same manner as the
destruction of the ability to give the positive Janus green reaction. The
results are shown in Table I, and it can be seen that the dosages of
radiation required to destroy these two properties are somewhat similar.
In a few instances it was possible to obtain a weak agglutination test
where the Janus green reaction was negative. The response of these
two properties of the jelly to the radiation are somewhat similar, yet
their reaction to acidity is quite different. We have observed that an
acid medium destroys the ability of the jelly to stain with Janus green,
whereas Tyler (1940) has observed that the acidified egg-water is rich
in agglutinin titer.

TABLE I

Exp.
no.

Lot

Medium

Amt. of
agglutination

Jelly

control

sea water

heavy

present

1 22,000 r

*i ( i

none

control

If U

heavy

present

61,000 r

II II

light

absent

122,000 r

II tl

none ?

absent

control

50% NaCl-KCl

heavy

present

30,500 r

it n

heavy

some present ?

61,000 r

U It

light

absent

91 ,500 r

li U

absent

122,000 r

it 11

none ?

absent

DISCUSSION

Proteins in suspension are denatured and coagulated after long expo-
sures to Roentgen radiation (Clark, 1936). The literature of radiation
action on agglutinins has been reviewed by Brooks (1936) and he sum-
marizes the suggested hypotheses as : ( 1 ) oxidation following primary
activation of oxygen, (2) electrical discharge of colloid particles by
alpha and beta rays, and (3) reduction processes. It is impossible at
the present to give definite evidence for any mechanism of the action of
the radiation on the Arbacia jelly, but the following observations may
be suggestive. (1) The action is rapid. (2) It is not prevented by
low temperature. (3) The increase in acidity is small. (4) The re-
moval of the jelly from the egg, the loss of the Janus green reaction in
solution, and the loss of the agglutination property appear to be three
stages in the degree of separation of the jelly particles. In view of the
observations noted above it may be that the radiation action involves a
change in the charge on the jelly particles causing them to be repelled
from each other. The action might also be expressed as an increased

X-RAYS ON JELLY OF ARBACIA EGG 369

affinity for water. The mechanism might he the same as the action of
Roentgen and radium radiation on the jelly of the Nereis egg. Packard
(1915) irradiated this egg with heta rays of radium and observed an
increase in the perivitelline space subsequent to fertilization. Redfield
and Bright (1918, 1919, 1921) and Redfield, Bright, and Wertheimer
(1924) exposed Nereis eggs to radium and Roentgen radiation and also
observed an increase in the perivitelline space after fertilization. These
authors were able to demonstrate a quantitative relationship between the
radiation and amount of swelling. The intensity of the radiation and
temperature were also found to be factors. The swelling of the jelly
was attributed to the absorption of an abnormal amount of water. Cos-
tello and Young (1939) suggest that the radiation initiates the outflow
of cortical jelly precursor and alters either the vitelline membrane or
the jelly in such a way that the passage through the membrane is com-
pletely or partially prevented.

SUMMARY AND CONCLUSIONS

Roentgen radiation (6,100 r) removes the jelly from the Arbacia egg.
When jelly is mechanically removed from the eggs it may be demon-
strated in the supernatant fluid by its reaction with Janus green. An ir-
radiation of 61,000 r will alter the jelly so that the Janus green test be-
comes negative. The jelly is not affected by solutions previously irradi-
ated. The radiation produces a slight shift in the pH of unbuffered
egg- water toward the acid side. The ability of the jelly-water to ag-
glutinate sperm is greatly decreased by an irradiation of 61,000 r.

LITERATURE CITED

BROOKS, S. C, 1936. The Effects of Irradiation on Venoms, Toxins, Antibodies,
and Related Substances. Chapter X. Biological Effects of Radiation.
Edited by Duggar. McGraw-Hill Co., New York.

CHAMBERS, R., AND H. POLLAK, 1927. Micrurgical studies in cell physiology.
IV. Colorimetric determination of the nuclear and cytoplasmic pH in the
starfish egg. Jour. Gen. Physiol, 10: 739-755.

CLARK, J. H., 1936. The Effect of Radiation on Proteins. Chapter VIII. Bio-
logical Effects of Radiation. Edited by Duggar. McGraw-Hill Co., New
York.

COSTELLO, D. P., AND R. A. YOUNG, 1939. The mechanism of membrane eleva-
tion in the egg of Nereis (abstract). Biol. Bull., 79: 311.

EVANS, T. C., AND H. W. BEAMS, 1939. Effects of Roentgen radiation on certain
phenomena related to cleavage in Arbacia eggs (Arbacia punctulata) (ab-
stract). Biol. Bull, 77: 331.

EVANS, T. C., 1940. Effects of Roentgen radiation on the jelly of the Arbacia egg.
I. Disintegration of the jelly (abstract). Biol. Bull., 79: 362.

GLASER, O., 1914. A qualitative analysis of the egg-secretions and extracts of
Arbacia and Asterias. Biol. Bull., 26: 367-386.

HARVEY, E. B., 1939. Arbacia. Collecting Net, 14: 180-181.

370 EVANS, BEAMS AND SMITH

JUST, E. E., 1930. The present status of the fertilizer! theory of fertilization.
Protoplasma, 10 : 300-342.

JUST, E. E., 1939. Basic Methods for Experiments on Eggs of Marine Animals.
P. Blakiston's Son & Co., Philadelphia.

KRAHL, M. E., AND G. H. A. CLOWES, 1938. Physiological effects of nitro- and
halo-substituted phenols in relation to extracellular and intracellular hy-
drogen ion concentration. II. Experiments with Arbacia eggs. Jour.
Cell, and Comp. Physiol, 11: 21-39.

LILLIE, F. R., 1915. Sperm agglutination and fertilization. Biol. Bull, 28: 18-33.

MAC!NNES, D. A., AND D. BELCHER, 1933. A durable glass electrode. I ml Eny.
Chem., Anal Ed., 5 : 199-200.

MAC!NNES, D. A., D. BELCHER, AND T. SHEDLOVSKY, 1938. The meaning and
standardization of the pH scale. Jour. Am. Clicm. Soc., 60: 1094-1099.

PACKARD, C., 1915. The effects of the beta and gamma rays of radium on proto-
plasm. Jour. Expcr. Zool., 19 : 323-347.

PANDIT, C. G., AND ROBERT CHAMBERS, 1932. Intracellular hydrion-concentration
studies. IX. The pH of the egg of the sea urchin Arbacia punctulata.
Jour. Cell, and Comp. Physiol, 2 : 243-249.

REDFIELD, A. C., AND E. M. BRIGHT, 1918. A quantitative study of the effect of
radium radiations upon the fertilization membrane of Nereis. Am. Jour.
Physiol, 45 : 374-387.

REDFIELD, A. C., AND E. M. BRIGHT, 1919. Temperature coefficient of the action
of beta rays upon the egg of Nereis. Jour. Gen. Physiol, 1 : 255-259.

REDFIELD, A. C., AND E. M. BRIGHT, 1921. The physiological changes produced
by radium rays and ultra-violet light in the egg of Nereis. Jour. Physiol,
55 : 61-85.

REDFIELD, A. C., E. M. BRIGHT, AND J. WERTHEIMER, 1924. The physiological-
action of ionizing radiations. IV. Comparison of beta and X-rays. Am.
Jour. Physiol, 68 : 368-378.

RICHARDS, A., AND A. E. WOODWARD, 1915. Note on the effect of X-radiation on
fertilizer Biol Bull, 28 : 140-147.

SMITH, M. E., AND T. C. EVANS, 1940. Effects of Roentgen radiation on the
jelly of the Arbacia egg. II. Changes in the pH of egg media (abstract).
Biol Bull, 79 : 362.

TYLER, A., AND S. W. Fox, 1939. Sperm agglutination in the keyhole limpit and
the sea-urchin. Science, 90: 516-517.

TYLER, A., AND S. W. Fox, 1940. Evidence for the protein nature of the sperm
agglutinins of the keyhole limpet and the sea-urchin. Biol. Bull., 79:
153-165.

TYLER, A., 1940. Sperm agglutination in the keyhole limpet, Megathura crenulata.
Biol. Bull, 78: 159-178.

WOODWARD, A. E., 1918. Studies on the physiological significance of certain pre-
cipitates from the egg secretions of Arbacia and Asterias. Jour. E.rpcr.
Zool., 26: 459-502.

GONOPODIAL CHARACTERISTICS PRODUCED IN THE

ANAL FINS OF FEMALES OF GAMBUSIA AFFINIS

AFFINIS BY TREATMENT WITH ETHINYL

TESTOSTERONE l

C. L. TURNER
(From the Department of Zoology, Northwestern University)

INTRODUCTION

The gonopoclium of the male of poeciliid fishes is a highly modified
anal fin used in the intromission of sperm. During juvenile stages the
anal fins of males and of females are practically identical in structure
but the fins of the two sexes become structurally divergent with the ap-
proach of maturity. Rays 3, 4 and 5 of the male fin undergo elonga-
tion and six differentiation areas arise within which characteristic hooks,
spines, serrae and plates develop. In the female fin there is no marked
elongation of the 3-4-5 ray complex and the differentiation areas do
not arise to elaborate the morphological features peculiar to them.

Regnier (1938) was able to induce the development of some of the
gonopodial characteristics of the male in the anal fins of the females of
Xiplwphorus hclleri and Lcbistcs reticulatus by intramuscular injections
of testosterone propionate, thus proving that the genetic factors for a
gonopoclium are borne by the female but that, in the absence of an
androgenic hormone, they are not expressed. Grobstein (1940) ob-
tained the same result in Platypoecilus maculatus but was able to demon-
strate in addition that an anal fin of an adult female produced the fea-
tures of the male gonopodium more exactly if that portion of the fin
which gives rise to the male characteristics was cut away and allowed
to regenerate during the treatment with testosterone propionate. Ever-
sole (1939) was able to secure the formation of atypical gonopodia in
females of Lcbistcs reticulatus at various stages by injecting testo-
sterone propionate peritoneally and by adding testosterone propionate
to the ration. The writer is indebted to Eversole for the information
that pregneninolone (ethinyl testosterone) and other androgenic hor-
mones will produce a similar effect.

The writer has studied in detail the normal development of the anal
fins in males and females of Gauibusia affinis affinis (Turner, 1941a)
with particular attention directed to the origin of the marked differences

1 This research has been supported by a grant-in-aid from the Graduate School
of Northwestern University.

371

C. L. TURNER

between the mature fin of the female and the gonopodium of the male.
In a second study (Turner, 1941&) fins or parts of fins were excised
during development and allowed to regenerate. Some conclusions were
drawn concerning the factors which control the normal development of
the characteristic structures of the gonopodium. The discovery by
Regnier, Grobstein and Eversole that the female anal fin could be made
to develop into an atypical gonopodium has made it seem worth while to
the writer to carry out similar experiments in Gambusla affinis with the
objective of comparing in detail the development of the male features in
the fins of treated females with those developed by males normally and
under hormone treatment. From the results obtained in these experi-
ments some conclusions can now be drawn concerning the effects of
an androgenic hormone upon the development of the various structures
of the gonopodium in females of Gambusia affinis which apparently pos-
sess genetic factors for the development of gonopodia but do not develop
them in the absence of a hormone.

In the experiments ethinyl testosterone 2 was added to the water of
the aquaria in which the fishes were kept in the dosage indicated by
Eversole (1.25 mg. of the hormone to 2000 cc. of water). All speci-
mens were maintained at a temperature of 18° to 21° C. during the
experiments.

NORMAL DEVELOPMENT OF FEMALE ANAL FIN AND
GONOPODIUM OF MALE

The anal fins of males and females in early juvenile stages are prac-
tically identical in structure. As the female fin develops, all of the ten
rays increase in length and segment regularly. A primary bifurcation
of rays 4 to 9 occurs at the 10 mm. stage, a secondary bifurcation at ap-
proximately the 22 mm. stage and a tertiary bifurcation in rays 8 and 9
when a specimen has reached a length of 40 to 45 mm. Growth and
segmentation of rays are continuous throughout the life of the specimen.
There is no fusion of the branches of the rays. Some joints between
the segments are obliterated by anchylosis beginning with a single joint
in each segment at about the 8 mm. stage. The anchylosis proceeds with
age until the basal parts of rays and about one-half of the primary
branches are involved in old specimens 48 mm. in length. Rays 3, 4,
5 and 6 are longest and rays 5, 6 and 7 contain the largest number of
segments.

The anal fin of the male differs from that of the female in late
juvenile stages in the differential growth of the rays, in bifurcation of rays

2 A supply of the hormone was obtained through the kindness of Dr. Erwin
Schwenk of the Schering Corporation, Bloomfield, New Jersey.

MODIFIED ANAL FINS OF FEMALE GAMBUSIA

373

and in the degree of anchylosis of basal joints. Rays 3, 4 and 5 become
greatly elongated during a period of rapid growth in which the growth
of the other rays is subordinated. A single bifurcation takes place in
all rays except 1, 2, 3 and 10 and the anchylosis of basal joints is very
limited. During the latter part of the period of accelerated growth the
differentiation areas appear (Fig. I, A). Tissue lying along the ventral

FIG. 1. Arabic numerals indicate fin rays numbered from the ventral side of
the fin. Roman numerals indicate location of differentiation areas.

A. Rays 3 to 8 and differentiation areas in the anal fin of a normal mature
male of Gambusia affinis affinis. The basal portions of the rays are not shown.

B. Rays 3 to 8 and differentiation areas in the anal fin of a 15 mm. female of
Gambusia affinis affinis after 30 days of treatment with ethinyl testosterone.

C. Rays 2 to 5 in the anal fin of a normal 22 mm. female specimen of Gam-
busia affinis affinis.

side of ray 3 thickens the ray (Area I). Area VI encompasses part of
the branched portions of ray 5 and of rays 6, 7 and 8 at the same level.

374

C. L. TURNER

Fusion of the branches and thickening of the segments of rays takes
place in this area. Areas II, IV and V arise near the end of the dif-
ferentiation period and Area III arises last. The hooks, spines, plates
and serrae which have their origin in these areas are indicated in A of
Fig. 1. Growth of the rays of the gonopod is concluded and terminalized
when the specialized structures within the differentiation areas have been
formed.

GONOPODIA OF TREATED MALES

Females exposed to treatment with ethinyl testosterone are obviously
placed in a situation markedly different from that of males which are
untreated, especially since the dosage of androgenic hormone used is
much in excess of that necessary to produce the male characters. It is
better to compare the gonopodia induced in females with those of males

TABLE I

Ray

Number of segments

Normal male

Treated male at

Normal untreated

specimen in

conclusion of

male at conclusion

16-segment

gonopodial

of gonopodial

stage

development

Ray

38-43

Ray

Rav

Ray-

Ray

which have been subjected to the same treatment rather than with males
which have been allowed to develop normally. For a first point of de-
parture, therefore, a comparison is made between the gonopodia of
normal untreated males and those of males treated during development.

Males in the 16-segment stage ( 16 segments in the undivided third
ray) will complete their gonopodial development and differentiation
within a period of 35 to 60 days. The specific morphological features at
the conclusion of development are shown in Fig. 1, A. Males in the
same stage of development will complete their development in 30 to 40
days when treated with ethinyl testosterone.

Table I shows the segment number in each ray of the anal fin of
normal specimens at the beginning of the experiment, of the treated

MODIFIED ANAL FINS OF FEMALE GAMBUSIA

specimens at the conclusion of the experiment and of fully mature un-
treated inales. The outstanding difference between the fully developed
gonopodia of normal and of treated males, as far as segment number is
concerned, is in the 3-4—5 ray complex. The explanation of the dif-
ference lies in the relation of growth and segmentation in these rays to
the formation of the differentiation areas. In normal gonopodia de-
velopment of a low concentration of androgenic hormone from the testis
induces accelerated growth and segmentation in the 3^4—5 ray complex.
As the testis develops further and more hormone is liberated from the
testis the developing fin responds by«giving rise to differentiation areas
and with the development of these, growth is terminalized. In the treated
specimens the concentration of the hormone is so high that the develop-
ment of the differentiation areas is quickly evoked and, although growth
in the 3-4-5 ray complex is stimulated, it is soon curtailed. Rays 1 and
2 on the ventral side of the fin and rays 9 and 10 on the dorsal side are
not involved in terminalizing differentiation areas and the growth-
stimulating effect of the hormone becomes evident in slight increases in
length and segment number in these rays as compared to the same rays in
the normal gonopodium.

Some of the differentiation areas develop normally in the treated males
while others do not. Area I in untreated males lies along the ventral
side of ray 3 and adds new tissue to the segments as they are formed
with the result that the rays are thickened. When Area II appears the
effect upon ray 3 is to add a series of about eighteen new segments
within 25 days. The new segments become progressively shorter and
thinner and on the ventral side develop characteristic spines and plates.
The transition in the character and size of segments from Area I to
Area II is gradual. In the treated males there is an abrupt change in
the length of segments in ray 3 between Area I and Area II. There
is an indication that Area II has an immediate terminalizing effect upon
further growth and regular segmentation of the ray, so that there are
fewer segments in the ray when terminalized and there is no gradual
transition in length of segments where Areas I and II overlap. The
spines and plates which are developed on the ventral sides of the seg-
ments in Area II are fairly normal. Area VI in the treated male is
almost identical with that of a normal untreated male. The appearance
of Areas IV and V on rays 4 and 5 have the same effect of terminalizing
growth and segmentation. From nine to eleven fewer segments are
formed in rays 4 and 5 than in a normal gonopodium. Area IV, within
which in the normal gonopodium a terminal hook develops on the dorsal
branch of ray 4 and a second smaller, closely associated terminal hook

376 C. L. TURNER

on the ventral branch of ray 5, arises in the treated specimen in the usual
position with reference to the ends of rays 4 and 5. However, the
relative lengths of rays 4 and 5 are not quite normal, ray 5 being shorter.
When the terminal hooks form on the two rays they are well separated
in the treated specimen. Area V in normal untreated specimens is at
first a condensed mass of tissue on the dorsal branch of ray 4 beginning
about six segments from the end and extending basally for approxi-
mately eight segments. Within it arise vertical curved serrae which are
laterally paired and placed one to each segment. Occasionally in the
anterior segments of the series two low elevations may arise instead of
one. In the treated specimen the development and final appearance of
the structures of the area are normal except for a tendency to include a
few additional segments at the anterior end. The relation of Area V
to Area IV is constant with four segments intervening. Area III is
the last to appear in both normal and treated specimens. It is likely to
be lacking altogether in treated specimens or to be poorly developed.

When ethinyl testosterone is administered to males in juvenile stages
before the 3^1—5 ray complex has begun to predominate in growth, the
same result is obtained in the early evoking of the differentiation areas.
However, growth in the 3-4-5 ray complex is stimulated by the same
hormone and the earlier the stage treated, the greater the degree of
growth and segmentation of the rays before terminalization. There
will actually be a smaller number of segments in the 3-4-5 ray complex
in a completely differentiated gonopod of a younger treated specimen,
but the relative number of segments added to the rays after the be-
ginning of the treatment will be greater. For example, a young normal
male in the 8-segment stage will have 8 segments in ray 3, 11 segments
in ray 4 and 12 segments in ray 5. Treatment with ethinyl testosterone
induces the early formation of a gonopodium which will have, when
complete, 25 segments in ray 3, 25 segments in ray 4 and 26 segments
in ray 5. Ray 3 will have added 17 segments and rays 4 and 5 will
have added 14 segments each. The treated specimen already described
(at the 16-segment stage) will have added only 15 segments to ray 3,
11 segments to ray 4 and 8 segments to ray 5.

The effect of treating males during the development of the gono-
podium with large dosages of ethinyl testosterone are: (1) to evoke pre-
maturely the formation of the differentiation areas. (2) To terminalize
prematurely growth in the 3-4—5 ray complex. Growth in these rays is
normally terminalized when differentiation areas appear but the early in-
duction of the differentiation prematurely terminalizes growth and differ-
entiation. (3) To produce secondarily some abnormal conditions in the

MODIFIED ANAL FINS OF FEMALE GAMBUSIA 377

differentiation areas themselves. The normal formation of an area which
involves parts of two rays (Area IV) depends upon the attainment of
a specific growth stage in each ray by the time differentiation occurs. If
treatment with hormone is started at a stage before one of the rays has
attained the usual growth stage the differentiated structure may be ab-
normal. (4) To stimulate growth and segmentation in the 3^4—5 ray
complex if the hormone is administered before these rays have entered
their accelerated growth period.

GONOPODIAL STRUCTURES INDUCED IN THE ANAL FINS OF FEMALES

The extent to which an anal fin of a female can be modified during
its development depends upon the extent to which the fin has become
structurally fixed in the female pattern. Modification by treatment with
ethinyl testosterone does not produce a regression of any structure al-
ready formed. Addition of new tissue to old structures or renewed
or accelerated growth of structures already present are the mechanisms
of modification. A closer approximation to the pattern of the mature
gonopodium of the male may be induced more readily in a young female
specimen, where a mature pattern of neither male nor female has been
laid down, than in an old female where growth, segmentation, additional
bifurcation and anchylosis of basal joints of rays have fixed the struc-
ture of the fin in the adult female pattern.

Female Fifteen mm. in Length

In the anal fin of a normal female at this stage rays 4 and 5 have
bifurcated once and there is some anchylosis of the basal segments of the
rays. Rays 5 and 6 are the longest and have the largest number of
segments.

After 23 days of treatment with ethinyl testosterone rays 3, 4 and 5
have elongated and segmented so as to form a definite lobe and all the
differentiation areas except Area III have made their appearance. At
the end of 31 days differentiation is complete. An untreated male of this
size and stage of development would have required about 45 days for
the complete development of a gonopodium. At 23 days ray 3 added
11 segments, ray 4 added 13 segments and ray 5 added 6 segments.
Differentiation of areas occurred prematurely and the growth of the
3^—5 ray complex was terminalized sooner than would have been the
case in a normal untreated male. In the complete gonopod of the treated
female rays 3, 4 and 5 had 24, 29 and 24 segments respectively while
an untreated male would have had 41, 43 and 42 segments in rays 3, 4
and 5 respectively.

378 C. L. TURNER

All differentiation areas develop in normal positions with respect to
the ends of rays 3 to 8 (Fig. 1, B). As a result of the early development
of the fin in a female animal and the development of the terminal phases
under the influence of the androgenic hormone, the base, which is un-
changed, is that of a typical female while the terminal portion is, with
some modifications, that of a male. At this stage there is little difference
between the anal fins of males and females both of which have been
treated with the hormone. In the female specimen the elbow of the ven-
tral branch of ray 4 (Area III) is not so well developed and in Area V
the vertical serrae are lower and the anterior segments of the gonopodium
are inclined to have two pairs of serrae instead of the one occurring in
males. An interesting variation of Area IV suggests that the synchroni-
zation of the end of the growth period with the onset of the differentiation
areas which operates smoothly in the normal specimen is poorly inte-
grated here. Some days in advance of the final stage of differentiation
there appeared in the ventral branch of ray 5 what seemed to be the
terminal hook. However, after the hook was formed the ray continued
to grow and to segment (Fig. 1, B) until the terminal segment came into
contact with the recurved terminal hook at the end of the dorsal branch of
ray 4. In this position it formed a second smaller hook. Both branches
of ray 4 also continued to grdw a little beyond the position of the usual
terminal hook and, after adding a long segment, formed small but definite
second terminal hooks. Apparently, here the axial growth of the rays,
which is terminalized in normal gonopodial development with the dif-
ferentiation of the terminal hooks, was not terminalized completely and
continued on so as to add new segments beyond the differentiated areas.

Females Twenty to Twenty-eight mm. in Length

Anal fins of females at this stage of development are rounded in
outline with the central rays the longest. The unbi furcated bases of
rays 4 to 9 are partially or wholly solidified by the anchylosis of the joints.
Secondary bifurcations may occur in rays 7 and 8.

In specimens treated with ethinyl testosterone at this stage the 3-4—5
ray complex is visibly elongated within 18 days and within 31 days
differentiation of the fin is complete. Ray 3 adds 13 short segments
during this time and rays 4 and 5 add 7 and 8 segments respectively to
the ends of the branches.

Figure 1, C represents the terminal parts of rays 2, 3, 4 and 5 of
a normal female fin at this stage prior to treatment. Figure 2, D
indicates the changes that take place within 28 days and Fig. 2, E is an
enlarged camera drawing of the terminal parts of rays 3, 4 and 5 of
another specimen in the same stage of development. Area II con-

MODIFIED ANAL FINS OF FEMALE GAMBUSIA

379

tains a smaller number of segments than in the normal male but the
ventral spines, while atypical, are clearly like those in Area II in normal
gonopodia. The elbow of Area III is hardly developed. Area IV,
containing the terminal hooks of rays 4 and 5, is well represented but

FIG. 2. D. Rays 3 to 5 and the differentiation areas in the anal fin of a 20
mm. female of Gambiisia affinis affinis after 28 days of treatment with ethinyl
testosterone.

E. Enlarged drawing of terminal parts of rays 3 to 5 of the anal fin of a 22
mm. female of Gambusia affinis affinis after 28 days of treatment with ethinyl tes-
tosterone. Differentiation of the fin is not quite complete.

F. Rays 3 to 5 and differentiation areas in the anal fin of a 38 mm. female of
Gambusia affinis affinis after treatment with ethinyl testosterone for 36 days.

the' terminal hook of ray 5 is poorly formed and is well separated from
that on ray 4. Area V with its vertical serrae is developed in the same
position with reference to the end of ray 4 and to about the same extent
as in a treated 15 mm. female. Area VI is not as well developed as in
the 15 mm. specimen. The basal portions of the rays with their

380 C. L. TURNER

anchylosis were laid down while the specimen was developing as a normal
female and no change in this portion of the fin is produced by the
introduction of the hormone.

In general, the transformation of the fin at this stage is specifically
male-like in the elongation of rays and the addition of segments to the
3^4—5 ray complex and in the development of specific differentiation
areas characteristic of the male fin. In none of the differentiation areas
at this stage is the transformation as completely male-like as in the
specimen treated with ethinyl testosterone at an earlier stage.

Females Thirty-five to Thirty-nine mm. in LcngtJi

At this stage the anal fin of the female is rather definitely fixed in
its structural pattern. It is still growing slowly and is adding new
segments to the rays. The anchylosis of segments has proceeded apically
from the base of each ray so that the unbifurcated basal portion and a
considerable part of the secondary rami of the divided rays are solidified.
Secondary bifurcation has taken place in the dorsal branch of ray 4
and in both branches of ray 5 with the result that there are now three
termini in ray 4 and four in ray 5.

Treatment of specimens with ethinyl testosterone results in a spe-
cific but incomplete development of the differentiation areas of the gono-
pod but the addition of new segments is extremely limited. From one
to four new terminal segments may be added to ray 3, a very small
number as compared to the number that is added to ray 3 in the younger
stage, but it is important that the specific response occurs even slightly.
The terminal branches of rays 4 and 5 may add one or two short seg-
ments. It would appear that the capacity to respond to hormone stimu-
lation by growth had nearly reached its limit at this stage.

The formation of the differentiation areas is also specific as in the
anal fins of the treated females at younger stages but the development of
the areas is less complete. Fig. 2, F represents the terminal parts of
rays 3, 4 and 5 in the anal fin of a 38 mm. female which has been treated
for 36 days. Area II is represented by a slight roughening on the
ventral side of the last seven segments of ray 3. Area III in this
specimen does not appear but in some specimens there is a slight thicken-
ing of two or three segments in the correct position on the ventral
member of the dorsal branch of ray 4. Area IV is represented by a
strong terminal hook on the dorsal member of the dorsal branch of
ray 4. The terminal hook on the branch of ray 5 nearest to ray 4 is
sometimes present but is always well separated from that of ray 4.
The vertical serrae which form Area V are present in some specimens

MODIFIED ANAL FINS OF FEMALE GAMBUSIA 381

but are usually low. They are developed upon the usual number of
segments (5 to 7) and the area is in the usual position beginning about
four segments from the end. Area VI is poorly represented. In some
instances there is a slight thickening of the dorsal members of ray 5 or
there may be some fusion of the terminal branches but the modification
does not extend to rays 6, 7 and 8.

Modification of the anal fin in females at this stage is much more
limited than in younger specimens and consists of the addition of a few
segments to rays 3, 4 and 5 and of a specific but incomplete development
of the differentiation areas.

DISCUSSION AND SUMMARY

The structural modification of the anal fin in a female treated with
ethinyl testosterone follows the same course of development as that
of the development of the normal gonopodium of the male, as far as
possible, considering that the anal fin is already fixed to some extent
in the female pattern and that the ethinyl testosterone is administered in
a high dosage. In the development of the normal gonopodium a small
amount of androgenic hormone, released from the testis, produces the
initial effect of accelerated growth in the 3-4-5 ray complex. At the
same time two of the six differentiation areas are induced to arise.
Later, a larger amount of hormone is released from the testis and four
other differentiation areas make their appearance. With the origin of
these four differentiation areas growth and segmentation in the 3^1—5
ray complex is terminalized. When ethinyl testosterone in the dosage
used here is introduced, the period of accelerated growth is induced, but
before the 3-4—5 ray complex has become elongated to the normal ex-
tent, the second phase of development, terminalizing differentiation, is
induced and growth is curtailed.

A specific pattern in the normal gonopodium is dependent upon a
normal growth in the various rays at the time differentiation occurs. In
the treated specimens, lacking the full extent of growth in each of the
rays concerned, there are likely to be abnormalities in the differentiation
areas themselves, particularly in Area IV where the two terminal hooks
found in the area become spatially separated.

If the hormone is administered to a juvenile female the amount of
growth in the 3-4-5 ray complex is greater and the resemblance of the
fin in general and detail is nearer to that of the normal mature male
than is the case if the hormone is administered to a specimen in which
the fin is in a later stage of development. In old females there is prac-
tically no growth and the extent to which differentiation areas are formed

382 C. L. TURNER

is extremely limited. The differences in transformation at the different
stages are due to two conditions: (1) The rate of growth of the fin be-
comes diminished in older specimens, presumably because of the onset
of a slower metabolic rate in the tissue. It has been shown by the writer
in regeneration experiments (Turner, 1941/?) that capacity to regenerate
diminishes in old specimens and that after the formation of the dif-
ferentiation areas this capacity is lost entirely in those rays which are
involved in the differentiation areas. In the same study it is shown that
capacity to form new differentiation areas after they have been excised
diminishes and disappears entirely once the areas have developed fully.
Grobstein's experiments, in which it is demonstrated that, in older
females, the formation of the differentiation pattern is more complete in
regenerated than in normal senescent tissue of the anal fins of older
females, indicate that tissue with a higher metabolic rate is more sus-
ceptible to transformation. (2) Since there is no dedifferentiation nor
any absorption of tissue already formed in the transformation of the
female fin, the development of the normal anal fin in the direction of the
female pattern becomes a definite deterrant to transformation. Changes
occur only by growth and segmentation of rays already formed and by
the addition of new structures within the differentiation areas to the old
or new segments. In old specimens, since there is little capacity to
respond to hormone stimulation by growth, the modification of the fin
is limited to such changes as can be accomplished by the addition of new
structures to the framework of segments already elaborated. Further-
more, the reduced capacity in old specimens of the differentiation areas
themselves to form the characteristic structures reduces the degree of
change.

In general, it may be stated that the capacity for a female anal fin
to become modified in the direction of the typical gonopod of a male is
high in juvenile specimens, less in specimens somewhat older, and very
limited in older specimens because of a fixation of the fin in a female
pattern and because of a diminished capacity for growth and differen-
tiation in older specimens. In older specimens the fin will be female
in pattern but there will be additional structures formed within the six
differentiation areas to the extent to which it is possible in a fin already
fixed in the female pattern.

BIBLIOGRAPHY

DULZETTO, F., 1931. Sviluppo e struttura del gonopodio di Gambusia holbrookii

(Grd.). Pnbbl. Staz. Zoo/. Nafoli, 11: 62-85.
EVERSOLE, W. J., 1939. The effects of androgens upon the fish (Lebistes reticu-

latus). Endocrinology, 25: 328-330.

MODIFIED ANAL FINS OF FEMALE GAMBUSIA

GROBSTEIN, CLIFFORD, 1940. Endocrine and developmental studies of gonopod dif-
ferentiation in certain poeciliid fishes. I. The structure and development
of the gonopod in Platypoecilus maculatus. Unir. California Publ. Zoo!.,
47: 1-22.

GROBSTEIN, CLIFFORD, 1940. Effect of testosterone propionate on regenerating anal
fin of adult Platypoecilus maculatus females. Proc. Soc. Exp. Biol. and
Mcd., 45 : 484-486.

REGNIER, M. T., 1938. Contribution a 1'etude de la sexualite des Cyprinodont vivi-
pares (Xiphophorus helleri, Lebistes reticulatus). Bull. Biol. de France
et Belg., 72 : 385-493.

TURNER, C. L., 1941a. The morphogenesis of the gonopodium of Gambusia affinis
affinis. (In press.)

TURNER, C. L., 1941 b. Regeneration during morphogenesis of the gonopodium in
Gambusia. (In press.)

MATING TYPES IN DIVERSE RACES OF
PARAMECIUM CAUDATUM

LAUREN C. OILMAN
(From the Zoological Laboratory, Johns Hopkins University)

INTRODUCTION

Investigations on Paramecium aurelia (Sonneborn, 1937, 1938 a
and b), and on P. bursaria (Jennings, 1938 a and b, 1939 a and b) have
recently shown that these species consist of a number of mating types.
As a rule, and possibly always, individuals of the same mating type will
not conjugate with each other ; but when cultures of certain diverse mat-
ing types are mixed together there follows under appropriate conditions
an immediate agglutinative reaction leading to conjugation between ani-
mals of diverse types. These phenomena are of interest in themselves,
and in relation to sexuality and self-sterility ; and they provide a means
by which the genetics of these organisms may be rapidly developed. It
therefore appears desirable to investigate from this point of view a large
number of diverse species so that there may be available a broad com-
parative body of knowledge of these phenomena.

For this purpose, Paramecium caudatuin, a species not hitherto
studied from this point of view, was selected for intensive investigation.
This species was chosen because it is one of the commonest and most
intensively investigated species of Paramecium and because its nucleus
and chromosomes are moderately favorable for the cytological work that
must eventually become correlated with the genetic analysis.

Four major problems have been attacked experimentally. The first
and basic problem is the occurrence, interrelation and geographical dis-
tribution of the mating types. The second problem, for which no final
answer is available, is the inheritance of mating type during vegetative
reproduction. The third is the influence of various environmental fac-
tors (nutrition, time of day, and temperature) on conjugation following
mixture of different mating types. The final problem was to discover,
if possible, morphological or physiological differences between the diverse
groups that could be distinguished by their breeding behavior.

MATERIALS

The material used in the present work was derived from collections
of Paramecium can-datum obtained from twenty-six natural sources in

384

MATING TYPES IN PARAMECIUM CAUDATUM

Canada and in the states of California, Connecticut, Georgia, Kansas,
Maryland, Massachusetts, and Pennsylvania. Soon after each collec-
tion reached the laboratory, one or more individuals were isolated and
from each individual a large stock culture was developed. These ninety-
three stock cultures we;r e the ones employed in all the following experi-
mental work. All the clones used were identified as P. caudatum by
examination of temporary aceto-carmine preparations or, in a few cases,
of permanent Feulgen preparations to determine the number and type
of micronuclei.

I am indebted to the following people for supplying me with collections
of Paramecium caudatum: Dr. T. T. Chen, Dr. Harold Finley, Father
J. A. Frisch, S. J., Dr. A. C. Giese, Mrs. R. W. Oilman, Mr. C. B. Metz,
Dr. T. M. Sonneborn, Mr. Samuel Steinberg, Dr. Vance Tartar, and
Prof. D. H. Wenrich.

METHODS

The basic culture fluid, a lettuce infusion medium, was prepared as
described by Sonneborn (1936), save that .75 grams of dried lettuce per
liter was used instead of 1.5 grams. This fluid was lightly inoculated
before use with a single unidentified species of bacteria grown on agar
slants. This bacterium was isolated in the early stages of the work from
a thriving culture of the paramecia.

The paramecia were cultured either as isolation lines on depression
slides with daily transfer of single animals or as mass cultures in glass
caster dishes with periodical transfer of a number of the animals to a
fresh dish. In some cases, the mass cultures were fed by adding a grain
of pearl barley, or a small piece of coagulated egg yolk to induce bac-
terial growth.

No effort was made to maintain absolutely sterile conditions but pre-
cautions were taken to insure the predominance of the desired bacterium
in the culture. The glassware was sterilized by boiling or autoclaving,
and the cultures were exposed to the air only long enough to allow the
removal of animals for transfer or for experimental purposes.

At times, heavy bacterial growths (presumably of a contaminating
bacterium) caused the appearance of heavy clouds of bacteria in the bot-
tom of the slides or caster dishes. At other times, some of the cultures
became contaminated by a small flagellate. The cultures were effectively
purified of the contaminating organisms by running single animals in
isolation lines for four days in succession. In making the transfers, the
mirror of the microscope was tilted so that no light fell on the objec-
tives and the contaminating organisms appeared as luminous dots. In

386 LAUREN C. OILMAN

this way it was possible to draw back up into the pipette most of the con-
taminating organisms transferred with the paramecium. This method
reduces considerably the risk of injury to the animals which the repeated
transfers used in washing by the method of Parpart (1928) involve.
Although the method described merely insures a predominance of the
desired bacterium, it was found to be entirely satisfactory.

TESTING CULTURES FOR MATING TYPES

The fundamental observation on which the concept of mating types
is based is simply this : certain cultures in which conjugation does not
occur when separate, conjugate when mixed together. Two such cul-
tures that do not conjugate alone but do conjugate when mixed are said
to be of different mating types. In order to ascertain whether there
occur in P. caudatum mating types such as those found in P. aurelia,
P. bursaria and other species of Paramecium, it was necessary to obtain
cultures within which conjugation did not occur, to mix representatives
of these in all possible combinations of two, and to observe whether
conjugation occurred in the mixtures or not.

The cultures to be tested for mating types were the 93 from the
various collections mentioned in the section " Materials." Previous
work by Sonneborn (1938a) and Jennings (1938a) on other species of
Paramecium has shown that it is unnecessary to make all possible com-
binations of two among the cultures examined, for they found that all
cultures of the same mating type behave alike when mixed with any
other culture. Therefore, in the present work, after mating types had
been discovered, only one representative culture of each mating type was
used for mixture with new cultures of unknown mating type. The rule,
therefore, was to mix every unknown culture with every other unknown
culture and with representative cultures of each known mating type.

OCCURRENCE, NUMBER, AND INTERACTION OF THE MATING TYPES

As a result of mixing the various clones it was found that certain
mixtures regularly gave conjugation while others regularly gave no con-
jugation. It was concluded, therefore, that mating types were present
in P. caudatum. This agrees with the findings of Giese and Arkoosh
(1939) who reported the presence of two mating types in P. caudatum.
When the results were collected, it was found that the clones studied
could be divided into at least four and probably five groups of two
mating types each. The groups were numbered one to five in order of
their discovery. The mating types of Group 1 were designated I and
II, those of Group 2, III and IV, those of Group 3, V and VI, those of
Group 4, VII and VIII, and those of Group 5, IX and X. Although

MATING TYPES IN PARAMECIUM CAUDATUM 387

this is the same nomenclature used by Sonneborn for P. aurclia, it im-
plies no connection between the corresponding groups and types in the
two species.

The interaction of groups and types in P. caudatutn is shown in
Table I. Each mating type in a group conjugates only with the other
mating type in the group. Thus type I conjugates only with type II and
not with other clones of type I or with clones of types III, IV, V, and
VI, and so on for the other groups.

TABLE I

The relations among types and groups in P. caudatum. Conjugation is repre-
sented by a plus, absence of conjugation by a minus. A blank indicates that no
mixture was made when both of the groups involved were known to be in reactive
condition.

Group

1 2
Type I II III IV

345
V VI VII VIII IX X

I +

— —

II +

— —

HI +

_ _ _ _ _ _

IV +

_ _ _ _ _ _

VII

- - + - -

VIII

-. - +

X -

It is not certain as yet that Group 1 is separate from both Group 4
and Group 5, since so far Group 1 has not been in condition to con-
jugate at the same time as Groups 4 and 5. However, from other con-
siderations, it appears highly probable that there really are five groups
of mating types. There remain, however, two collections from which
no clones have so far conjugated when mixed with each other or with
any of the groups of mating types. It is possible that these two collec-
tions may represent one mating type of a sixth group.

If samples of the two mating types in a group are mixed when in
the proper physiological condition, there follows immediately the pro-
nounced agglutinative mating reaction described by Sonneborn (1937)
for P. aurclia and by Jennings (1939a) for P. bursaria. When the
animals are put together they stick to those of the opposite mating type
with which they chance to come in contact and form clumps which later
break down into pairs which complete conjugation.

In regard to the geographical distribution of the various groups, it
is to be noticed that Group 1 was found in two collections from Balti-
more, Maryland, but not in collections from other localities. Group 2

•^(LIBRARY

LAUREN C. OILMAN

was found in six collections from Baltimore ; one collection from Wood-
stock, Md. ; one collection from New Haven, Conn. ; one collection from
Stanford, California; and two collections from Falmouth, Mass. Group
3 was found in three collections from Baltimore, three from Baldwin
City, Kansas, one collection from Atlanta, Georgia, and one collection
from an unknown locality in Connecticut. Group 4 was found in one
collection from Baltimore, one collection from Waterville, Conn., and
one collection from New Haven, Conn. Group 5 was found in a col-
lection from Hamden, Conn. The two collections whose group is still
undetermined were from Philadelphia, Pa. and Canada near Buffalo,
N. Y.

In general, it can be said that there is no definite evidence of geo-
graphically isolated non-interconjugating groups of mating types. It is
true that Groups 1 and 5 have been found in only one general locality,
but since there is little material available for these groups this cannot be
considered significant. Animals of the other groups have been found in
widely separate localities.

CONJUGATION WITHIN A CLONE

There are some apparent exceptions to the rule that any clone in a
group conjugates with only one of the two mating types in the group.
However, in all these apparent exceptions, conjugation has also occurred
in one of the control cultures, so that it is not a question of a clone
conjugating with clones of the two mating types of a group but rather
a question of conjugation between members of the same clone. As an
example, some results of mixtures of five clones of type IV (A2 to A6)
with a type III clone (SI) and a type IV clone (Al) will be given.
In the mixtures with the type III clone, from 61 to 72 pairs were formed
with from 7 to 11 pairs of conjugants in the mixtures with the type IV
clone. However, no conjugants were found in the type III (SI) con-
trol, while ten pairs were present in the type IV (Al) control. Some
conjugants were found in the A2 to A6 controls, ranging from none in
A6 to eleven in A4. Thus it is seen that the conjugation in the mixtures
with Al was the result of conjugation within the various clones and not
the result of mixture.

In the clones in which conjugation has been observed to occur in the
controls set up when making mixtures, conjugation has also been ob-
served in the source cultures. Many of the cultures have conjugated at
times without mixture. There is a great deal of variation with respect
to this phenomenon among different clones. In some clones there have
been large percentages of conjugants at intervals of two weeks to a

MATING TYPES IN PARAMECIUM CAUDATUM 389

month. Each time the conjugation occurred in a culture which had
been started from a single non-conjugant animal at the time of the
previous occurrence of conjugation. The proportion of animals con-
jugating varied in these cultures from about 20 per cent to nearly 100
per cent. Other clones have conjugated at only one time during a period
of nine to twelve months, and in these the proportion of animals con-
jugating varied from 5 per cent to 10 per cent. In still other clones, no
conjugation has ever been observed except upon mixture with another
clone of the proper mating type. Examples of the kind of clone in
which large numbers of conjugants occur at short intervals have been
found in both Groups 1 and 2. In Group 1, all three of the type II
clones so far discovered are of this kind while in Group 2 two of the 7
type III clones and five of the 12 type IV clones are of this kind.
Clones of the second kind in which small proportions of conjugants occur
at long intervals have been found in two groups. In Group 1, there
are four examples all belonging to type I, while in Group 3, two type V
clones and two type VI cones are of this kind. In each of these, only
a small proportion of the animals in one culture have conjugated at one
time during the entire period of eight months to a year and a half that
they have been under observation. Examples of clones which have never,
while under observation, conjugated without mixture are found in all
three groups: in Group 1, one type I clone; in Group 2, two type IV
clones and one type III clone; and in Group 3, one type V and one type
VI clone. The length of time that this condition has held true differs
with different clones since some have been kept in the laboratory for
longer periods than others. Four of the clones have been under ob-
servation for a year to a year and a half, and two for eight months.

Since conjugation does occur in this manner within a clone, the prob-
lem presents itself of whether or not two mating types have been pro-
duced within the clones as Sonneborn (1937) found in P. aurclia at
autogamy. In order to test this matter, it was necessary to do two
things: first to see if any autogamy or endomixis occurred (endomixis
in P. caudatinn has been reported by Erdmann and Woodruff, 1916, and
by Chejfec, 1930), and if so, to see if conjugation in a culture without
mixture depended upon its prior occurrence in the culture; second, to
find out whether or not both of the mating types in a group were pro-
duced in a culture originally of one mating type and whether or not the
differentiation into two types occurred at autogamy or endomixis.

In order to test the possibilities just mentioned, it was necessary to
attack the problem in two ways. In the first place, to find if autogamy
or endomixis were occurring, it was necessary to stain cultures daily
with aceto-carmine. In pursuance of this plan, twenty-four clones were

390 LAUREN C. OILMAN

run in daily isolation lines with daily staining for a period of thirty days.
During this time one-fourth to one-half of the animals in each of the
depressions were removed each day, stained with aceto-carmine and ex-
amined with the compound microscope for nuclear changes indicative of
autogamy. Although examples of clones which had, in caster cultures,
been producing conjugants at intervals of from ten days to two weeks
were included, no evidence of any nuclear changes was observed during
the month the cultures were under observation.

Although the results indicated that conjugation was occurring in
these cultures without previous autogamy, it was felt, since the tests for
autogamy were made under isolation line conditions in depression slides
and the conjugation which had occurred had been under mass culture
conditions, that the environmental conditions in the two situations were
sufficiently different so that autogamy might have occurred in the parent
caster dish cultures even though none occurred in the isolation lines de-
rived from them. It was therefore decided to stain representative
samples from a caster dish culture that was originally derived from one
animal of clone D4, a clone in which conjugation occurred frequently.
This animal was allowed to multiply in a depression slide. As soon as
a sufficient number of animals was present, they were transferred to a
caster dish and sixty drops of culture fluid added. As soon as several
hundred animals were present in the dish, a grain of pearl barley was
added to give a constant supply of food, since it was under these condi-
tions that conjugation was previously observed to occur when the para-
mecia in the isolation lines showed no indication of autogamy.

The procedure used in testing for autogamy was to stain a sample
from the culture every day and examine for nuclear changes. When
only a relatively small number of individuals was present in the culture
one-fourth were stained with aceto-carmine for the daily examination
until at least ten were being stained. From that time onward from ten
to one hundred animals were examined daily. No evidence of nuclear
reorganization was ever observed in this culture but practically 100 per
cent of the animals were conjugating twelve days after the start of the
experiment.

These results indicate (for clone D4 at any rate) that the conjugation
observed is not the result of a previously occurring autogamy. There re-
mained to be answered, however, the question of whether or not both mat-
ing types were present in a clone when conjugation occurred. In order
to answer this question, pairs of conjugants which were not yet firmly
united were separated by squirting them violently from a small bore
pipette. The cultures derived from the animals separated in this way
are known as split-pair cultures. Because of the difficulty of finding

MATING TYPES IN PARAMECIUM CAUDATUM 391

pairs which are not yet firmly united in cultures in which only a small
proportion of the animals conjugated at one time during the course of
the experiment, all the work with split pairs was carried on with those
cultures in which large proportions of the animals conjugated at rela-
tively short intervals. In Group 1, type II, clones D4 and C2 were used
mainly with some additional work on clone D2. In clones D4 and C2,
a number of pairs (17 in D4 and 18 in C2) were split and cultures
grown from each member of the pair. Besides the split-pair cultures, a
number of cultures were started from single non-con jugant animals from
cultures in which conjugation was occurring. In Group 2, clone P (type
III), twelve pairs of conjugants were split and cultures were grown
from them. The clones were tested for mating type as soon as the popu-
lation in the cultures had reached a sufficient density to make tests for
type possible. However, conjugation again occurred in many of the
cultures before it was possible to make the tests. The cultures which
were tested were always of the same type as the parent culture. Among
the split pairs from type II cultures, cultures from both members were
tested and reacted in the case of two split pairs from D4 and one split
pair from C2. Both members of each pair were type II, the same as
the original culture. In the split-pair cultures from clone P both mem-
bers of four pairs reacted as type III the same as the original culture.
In addition, those split pairs in which only one of the resulting cultures
gave a reaction and those cultures derived from isolated non-conjugating
animals behaved as the same type as the parent cultures when tested
for mating type.

In summary, it can be said that the results of the work on two clones
in Group 1 (C2 and D4, type II) and one clone in Group 2 (P, type III)
indicate that there is no permanent change of type when conjugation
occurs within a culture since, whenever cultures from both members of
a pair have given a test for mating type, both have been of the same type
as the parent culture. Furthermore, the parent cultures still react as
the same mating type they were when first isolated, even though they
have been subcultured many times during the year to a year and a half
that they have been in the laboratory.

These results do not allow any definite conclusion about possible
temporary changes of mating type during vegetative reproduction such
as Kimball (1939) found in P. aurelia. If such changes occur, they are
temporary and the animals quickly revert to the mating type character-
istic of the clone. The possibility that conjugation in these cases is be-
tween animals of the same mating type cannot be ruled out but seems
unlikely in view of the previous work on Paramecium.

Although the results presented indicate that conjugation within a

392 LAUREN C. GILAfAN

clone (selfing) is not the result of a change of mating type following
autogamy, there are several points of interest in connection with such
selfings. They occur only in cultures which have been in the same cul-
ture dish for a period of time — two weeks approximately if a grain of
pearl barley is added to the culture, longer if the cultures are fed by the
addition of lettuce infusion. If animals from such selfing cultures are
removed during the period before selfing occurs and mixed with the
appropriate mating type conjugation will take place, however, and con-
tinues to occur even when selfing has commenced in the culture. It can
thus be seen that being in the proper condition to conjugate when mixed
is not enough to induce selfing and that some additional factor is in-
volved. Successive selfings can be readily obtained at approximately
two-week intervals if cultures are started from single non-con jugant or
split-pair animals and maintained under the conditions described.

There is one case of conjugation within a clone (Gilman, 1939) in
which the situation seems to have been different from that so far re-
ported. In this clone (M, Group 2, type III) when conjugation was
observed the first time in an unmixed culture, four con jugant pairs were
split before they had gone through conjugation and the clones derived
from the two members were tested for mating type. In all four, one
member gave rise to a type III and one to a type IV clone. Since this
phenomenon did not recur, it is impossible to tell whether it was due to
a change of mating type as a result of autogamy or some other nuclear
change or whether it resulted from an accidental contamination of the
original type III culture by type IV animals from another source.

The subsequent histories of the two types isolated from this culture
were very different. The four type IV clones continued to react as type
IV and no conjugation occurred in them without mixture during the year
they were kept under observation. The four type III clones, however,
contained conjugants again approximately six weeks after the pairs were
split. New split pairs were obtained from these type III clones. The
clones derived from the two members of these pairs were all type III.
It appears, then, that conjugation within a clone may be the result of
the production of animals permanently of both types as in clone M at
the time of its first spontaneous conjugation or it may occur without the
production of clones permanently of two types as in the subsequent con-
jugation in clone M and also in clones C2, D2, D4 (type II) and P (type
III).

CONDITIONS NECESSARY FOR CONJUGATION

Nutrition

Some observations of considerable interest have been made on the
influence of the nutritive state on the clumping or mating reaction and

MATING TYPES IN PARAMECIUM CAUDATUM

on the subsequent conjugation. No detailed observations have been
made with respect to this problem in Groups 3, 4 and 5. A little has
been done with Group 1, but since the results obtained were essentially
similar to those obtained with Group 2 and since more detailed observa-
tions were made with Group 2, only this group will be considered. In
Group 2 five stages in nutritive decline associated with characteristic
changes in the mating reaction have been observed. When cultures of
types III and IV in these diverse conditions are mixed the following
immediate behavior is observed :

( 1 ) Animals very well fed and plump : No immediate mating reac-
tion and no conjugants present at the end of 24 hours.

(2) Animals well fed but not markedly plump: A weak immediate
mating reaction; a few animals cling together in pairs but break apart
in a short time ; no conjugants present at the end of 24 hours.

(3) Animals of moderate size, not well fed : Strong mating reaction ;
many clumps form ; these later disintegrate into pairs which remain to-
gether and complete conjugation.

(4) Animals small and thin : Strong mating reaction ; many clumps
form; these later disintegrate, but few or none of the animals proceed
to conjugate.

(5) Animals very small and starved: No immediate or later mating
reaction and no conjugation.

The mixtures used in the above observations were kept for only 24
hours so that it is not known whether conjugation would have occurred
in the mixtures in the first two stages if they had been kept for a longer
time, but it seems probable that conjugants would form as the paramecia
reached stage 3. The various stages of nutritive decline can be seen
successively in a culture to which a considerable amount of food is added
and the culture then allowed to decline without the addition of more food.

These observations appear significant in that they indicate that the
conditions under which the mating reaction occurs are not necessarily
favorable for conjugation; i.e. the mating reaction is much less sensitive
to nutritive conditions than conjugation so that the paramecia give the
mating reaction before they have reached the proper condition for
conjugation and also after they have passed this condition.

Diurnal Periodicity

Since a diurnal periodicity in the mating reaction has been found in
certain groups of P. bursaria (Jennings, 1939a) and P. aurcUa (Sonne-
born, 1938a), and furthermore, since it has been stated by Maupas

394 LAUREN C. GILMAN

(1889) that P. caudatum conjugates at about 4:00 A.M., this question
was investigated with all five groups of mating types.

In the investigation of periodicity the mixtures were examined im-
mediately for mating reactions and twelve hours later for conjugants.
In Group 1 hourly mixtures with immediate mating reactions and later
conjugation were made only between 11 P.M. and 6 A.M., but since
pairs just beginning to form have been observed at various hours of the
morning and afternoon, it seems fair to conclude that there is no diurnal
periodicity in this group. In Groups 2 and 3, mixtures followed by im-
mediate clumping and later conjugation were made at all hours of the
clay and night, so that it is obvious that there is no periodicity in these
two groups. In Groups 4 and 5, immediate clumping and later pairing
was obtained in mixtures made between 8 A.M. and 2 A.M. Although
mixtures were not tried between 2 A.M. and 8 A.M., strong mating reac-
tions were obtained at both ends of this period and it seems probable that
clumping and conjugation would have occurred if mixture had been
made at these times. It therefore appears that there is no diurnal pe-
riodicity in any of the five groups of mating types so far discovered in
P. caudatum.

Temperature

It was desired to investigate the effect of temperature upon conjuga-
tion in order to ascertain what temperature was most favorable for con-
jugation, whether the temperature to which the paramecia had been ex-
posed before mixture had any effect on the number of conjugants
formed, and whether any differences existed among the groups in their
response to temperature. In investigating this problem all factors ex-
cept the temperatures used were kept as constant as possible. One repre-
sentative clone of each mating type of Group 1, Group 2, and Group 3
was used in this work. Four cultures of each mating type were left at
room temperature (24° to 28° C.) for two clays. At the end of this time,
additional culture fluid was added and they were placed at the various
temperatures — in Groups 1 and 3: 9°, 18°, 24°, and 31° C., in Group 2:
9°, 20°, 24°, and 28° C. The cultures were left at the various tempera-
tures for forty-eight hours. For Groups 1 and 3, at the end of this time
four mixtures of the two types from each temperature were placed at
all the temperatures used. Thus, four mixtures from 18° were placed at
9°, four at 18°, four at 24°, and four at 31° C., and in the same manner
for each of the other temperatures. In the case of Group 3 the experi-
ment was repeated, giving a total of eight mixtures in all. In no case
were any mixtures made between cultures which had previously been
kept at two different temperatures ; all were between two cultures kept at

MATING TYPES IN PARAMECIUM CAUDATUM

the same temperature. In Group 2, six mixtures were made instead of
four and the experiment was performed twice, giving twelve mixtures
in all. The mixtures were examined at twelve-hour intervals and all
conjugants present were removed with a pipette and the number present
recorded. When no conjugants had been found in the mixtures for the
three previous twelve-hour periods, the remaining animals were removed
and counted. This number was used in calculating the percentage of
conjugation.

The results of the experiments are given in Table II. In Group 2,
it will be seen that the greatest percentage of conjugation occurred when

TABLE II

The effect of various temperatures on the number of pairs of conjugants and
the percentage of conjugation in mixtures of Group 2 and Group 3 animals. The
means of eight or twelve mixtures are given in the table.

Group 2
Temperature after mixture of the two types

28° 26° 20° 9°

28°

2.2%

18.3%

15.0

.4%

26°

3.5%

18.3%

16.4

.8%

20°

6.7%

33.0%

33.5

.8%

9°

0.0%

27.5%

33.7

-5%

Temperature
for two days
before mixture Group 3

31° 24° 18°

31°

4.5

9.0%

73.8

86.7%

63.2

88.2%

24°

0.0

0.0%

68.2

88.5%

54.5

83.8%

18°

0.0

0.0%

61.2

78.5%

49.5

71.2%

9°

0.0

0.0%

53.2

63.7%

42.2

49.7%

animals kept at 20° C. were put after mixture at 20° C. In general, the
results indicate that the temperature at which the animals were kept both
before and after mixture affect the amount of conjugation. Either very
high or very low temperatures after mixture either decrease the per-
centage of conjugation markedly or prevent it entirely. Low tempera-
tures (9° or 20°) before mixture appear to be more favorable for con-
jugation than high temperatures (26° or 28°).

In Group 3, the highest percentage of conjugation occurred when
mixtures of animals from 24° C. were kept at 24° C. In general, it can
be said that, as in Group 2, very high or low temperatures after mixture
are unfavorable for conjugation. Unlike Group 2, high temperatures
before mixture appear to be more favorable than low.

When the numbers of conjugants formed in each of the mixtures

396 LAUREN C. OILMAN

during each of the twelve-hour periods after mixture until no more con-
jugants were formed were considered, it was found that in Group 2 most
of the conjugants were formed in the first twelve hours after mixture.
In only one set of mixtures, — those put at 20° after two days unmixed
at 9°, — is the time of greatest conjugation shifted to the period between
twelve and twenty- four hours after mixture. At 24° in Group 3 most
of the conjugants were found between twelve and twenty-four hours
after mixture while at 18° most of the conjugants were found 36 to 48
hours after mixture.

In Group 1, the results obtained are not complete enough to justify
inclusion but they indicate that the lower the temperature before mixture
the greater the conjugation, and that most conjugants are produced when
the paramecia previously kept at 9° C. are put at 20° C.

GROUP DIFFERENCES

An attempt was made to ascertain whether there were any differ-
ences between the groups besides the primary group difference in the
mating type. With respect to size, it was found that the animals in
Group 3 are characteristically smaller than the animals in Groups 1, 2,
4 and 5. Under isolation line conditions, these differences are hardly
noticeable but when grown in mass cultures become very striking. In
Group 1, it was found that on the average the type II animals were
smaller than the type I and of a slightly different shape, being rather
shorter and broader. This difference becomes very obvious in con-
jugating pairs where the type I members may frequently be twice as
long as the type II member. So far only two collections of eight clones
in all have been found for Group 1 so that it is possible that the con-
dition is not general for the group. However, all the clones so far ex-
amined (three of type II and five of type I) show the difference. In
Group 2 characteristic size differences between clones have been observed
but without any correlation with mating type.

There are also characteristic differences in Group 1 as to selfing.
Type II animals self readily while the type I animals self very rarely.
In Group 2, both selfing and non-selfing clones are found but they are
not correlated with mating type. Group 3 seems to have rather less
selfing than either of the other two groups since only a few instances of
conjugation without mixture have been found in this group.

Under good conditions for conjugation, there are, as mentioned
previously, differences in the proportions of the animals in mixtures
which conjugate. Thus the greatest proportion of conjugation in mix-
tures occurred in Group 3 and the lowest proportion of conjugation in

MATING TYPES IN PARAMECIUM CAUDATUM 397

Group 1. Another already-mentioned difference with regard to conjuga-
tion is that low temperature before mixture causes more conjugation in
Groups 1 and 2 but less conjugation in Group 3.

A comparative study was made of the adverse effects on P. caudatum
of mixture with three clones of P. aurclia (H, G, and 47). These
clones are known (Sonneborn, 1938&, 1939, and unpublished) to cause
certain other clones of P. aurclia to die in characteristic fashion when
they are mixed with them. These three clones were mixed with 73
clones of P. caudatum and the mixtures were observed daily until dead
or until all the other mixtures which showed an effect were dead. In
all cases, control groups of P. caudatum were kept without mixture with
the clones of P. aurclia.

The effect on P. caudatum which is produced by G is first indicated
by the avoiding reaction and spinning by the affected individuals. The
spinning takes the form of rapid rotation on the longitudinal axis with
little or no forward movement. In some instances, the animals revolve
(without moving forward) around an axis parallel to the longitudinal
axis of the paramecium so as to describe a cylinder or a segment of a
cone. These manifestations do not occur continuously but alternate with
periods of quiescence or normal swimming. In addition to the altera-
tion in behavior described above, morphological changes also occur; the
paramecia become thin, flattened and gradually become transparent. As
they become transparent, crystals become visible in the cytoplasm. There
is no regularity about the position of the crystals, which are sometimes
in the anterior end, sometimes in the posterior end, and sometimes dis-
tributed about the periphery. No characteristic differences in the re-
action to G were noted between the groups of mating types.

Stock H has no marked effect on behavior of P. caudatum but the
morphological changes are much more striking than those produced by
Stock G. The affected animals stop feeding and lose all their food
vacuoles. They then become filled with clear vacuoles. Generally there
were two large vacuoles in each animal in the region of the contractile
vacuoles and several smaller vacuoles. In some, only one large vacuole
was formed. The vacuoles gradually enlarge until the ectoplasm be-
comes widely separated from the rest of the cytoplasm and the paramecia
appear blistered. Frequently, just before death, the macronucleus be-
comes visible as a round body. In many dead animals both the blistering
and the visible macronucleus are evident while in others only one is to
be seen. In other cases, only the vacuolization was apparent. No char-
acteristic differences were noted among the first three groups in their
response. However, both Groups 4 and 5 were resistant to the effect of
H and remained normal.

398 LAUREN C. GILMAN

The effect of Stock 47 was as striking as that of H in a somewhat
different way. The first effect wras that the animals became shorter and
thicker than normal. This was followed by enlargement of the pos-
terior end. This enlargement appeared to be caused by the massing of
the cytoplasm and macronucleus at the posterior end toward one side of
the animal. The animals frequently became almost spherical before
death. In some instances a single vacuole appeared at the posterior
end but this was not a constant effect. Much more common were small
vacuoles under the ectoplasm giving the animals a rough appearance.
None of the clones tested were resistant to Stock 47, and it is of con-
siderable interest that the clones which were resistant to H appeared to
be more quickly and strikingly affected than the other clones.

In summary, it can be said that there appear to be characteristic dif-
ferences between the groups in size and in various physiological char-
acteristics. In Group 1, there also appear to be characteristic differences
between the two mating types in size. In the other groups, there were
no such characteristic differences between the mating types.

DISCUSSION

The condition in P. caudatuin with respect to the number of types
within a group is like that in P. aurclia, since only two mating types have
been found in each group; not four or eight as in P. bursaria. P.
caudatuin differs from both P. aurclia and P. bursaria in that it has
certainly four and in all probability five groups of mating types while
they have only three groups.

The wide occurrence, under certain conditions, of conjugation within
a clone is of considerable interest in connection with the question of
whether or not the presence of both mating types is necessary for con-
jugation to occur. In one clone (M) when pairs were split, it was
found that both mating types of the group were indeed present. This
may have been due to a change of mating type at autogamy, such as was
found to occur in P. aurclia by Sonneborn (1937) and Kimball (1937).
The possibility of accidental contamination cannot be excluded, however.
In all the other clones of P. caudatuin in which it was possible to separate
conjugating pairs, the cultures derived from both members of a split
pair were of the same type as the original culture. It has been impos-
sible as yet to ascertain whether two mating types were present at the
time conjugation occurred, one of which changed type again to become
the same as the original clone or whether conjugation was occurring
between two animals of the same mating type. Kimball (1939) has
shown that in P. aurclia there may be a temporary change of mating type

MATING TYPES IN PARAMECIUM CAUDATUM 399

during vegetative reproduction but that animals which have thus changed
their mating type give rise to clones of the original mating type. It
appears possible that a similar process occurs in P. caudatinn. That the
repeated occurrence of conjugation within a clone was the result of a
preceding autogamy appears unlikely in view of the fact that the two
members of split pairs gave rise to clones of the same mating type and
in view of the failure to find evidence of nuclear change in these clones.
In some clones selfing occurred frequently while in others it occurred
only once or not at all. This difference between clones is probably the
basis of the conflicting results on the effect of environmental changes
on conjugation in P. caudatum. Zweibaum (1912) found that he could
induce conjugation in the clone of P. caudatinn with which he was work-
ing by adding various salts in certain concentrations. He concluded,
since he used only this one clone, that conjugation was dependent solely
on environmental factors. Hopkins (1921) and Ball (1925) tried
Zweibaum's methods on a number of clones of P. caudatum and found
that they gave conjugation with some clones but not with others. These
conflicting results can be explained by assuming that Zweibaum had a
clone of animals in which selfing occurred rather readily, while Hopkins
and Ball, since they used several clones, had some wrhich selfed readily
and some which would not self.

The nutritive requirements observed for conjugation appear to agree
very well with those reported by Maupas (1889), Calkins and Cull
(1907), Jennings (1910), Zweibaum (1912), Calkins and Gregory
(1913), Ball (1925), Chatton and Chatton (1931), and Giese (1935).
The paramecia conjugate when not too well fed, yet not starved. This
is, of course, the condition produced when the food supply is suddenly
decreased or when animals which have exhausted the food in the medium
are given a limited supply of food. The most interesting result of the
work on the effect of nutrition on conjugation is the fact that the mating
reaction, clinging together and clumping, occur under conditions which
are not satisfactory for conjugation. This phenomenon of strong
clumping with the production of few or no pairs of conjugants was ob-
served in several mixtures of clones in Groups 1 and 2. It is especially
marked in those animals which have passed the point of optimum nutri-
tional condition for conjugation and somewhat less so when the animals
have not yet reached this condition.

SUMMARY

1. An investigation of the numbers and interrelations of the mating
types in Paramecium caudatum in cultures derived from single animals

400 LAUREN C. OILMAN

isolated from wild cultures was carried out. Of 93 clones from 26
natural sources the mating types are still to be identified in 3 clones
from 2 natural sources.

2. The clone cultures could he divided into mating types in several
non-interbreeding groups. Animals from cultures of different groups
did not conjugate when mixed with one another. Within each group
two mating types were found. Animals from cultures of different mat-
ing types belonging to the same group conjugated when mixed together.
Four non-interbreeding groups of mating types have been definitely
established and the occurrence of five groups is highly probable.

3. In regard to the geographical distribution of the mating types, no
evidence was found for the formation of local groups of mating types
which would not conjugate with animals from other localities.

4. Ordinarily, conjugation occurred only when animals from two
different mating types were mixed but under certain conditions some
clones conjugated without mixture.

5. It was found that ordinarily such conjugation was not the result
of the production of two mating types at autogamy as in P. aurclia.

6. In one case (clone M, type III) both mating types of a group
were produced in a clone but it was not possible to correlate this fact
with a preceding autogamy.

7. It was found that the mating reaction itself (clumping of the
animals) would occur under nutritive conditions which would not per-
mit the completion of conjugation.

8. The temperature both before and after mixing the mating types
has a definite effect on the proportions of the animals conjugating.
Very little or no conjugation occurred when the animals were kept at
the extremes of temperature (9° and 28° and 31°) after mixture. A
low temperature prior to mixture caused more conjugation in Group 2,
less in Group 3.

9. None of the five groups of mating types gave any indications of
a diurnal periodicity.

10. Group 3 animals are obviously smaller than Group 1 or Group 2
animals.

11. In Group 1, there is a difference in size between the mating types ;
type I is larger than type IT.

12. Type II animals " self ' (conjugate without mixture) much
more frequently than type I animals.

13. In mixtures between animals of different types the largest per-
centage conjugate in Group 3 and the smallest in Group 1.

14. No differences were found between Group 1, 2, and 3 in their
response to the toxic effects produced by races G, H, and 47 of P.

MATING TYPES IN PARAMECIUM CAUDATUM 401

aurelia. The clones which form Groups 4 and 5 are resistant to the
lethal effect of H.

15. A possible explanation of the conflicting results on the effect of
environmental factors on conjugation obtained by earlier workers on
P. caudatum was presented.

LITERATURE CITED

BALL, G. H., 1925. Studies on Paramecium. II. The behavior of a conjugating
race of Paramecium caudatum. Univ. Calif. Pnbl. Zoo]., 26 : 387-433.

CALKINS, G. N., AND S. W. CULL, 1907. The conjugation of Paramecium aurelia
(caudatum). Arch. f. Protist., 10: 375-415.

- AND L. H. GREGORY, 1913. Variations in the progeny of a single ex-conjugant

of Paramecium caudatum. Jour. Expcr. Zool., 15 : 467-525.

CHATTON, E., ET M. CHATTON, 1931. La conjugaison du Paramecium caudatum
determinee experimentalement par modification de la flore bacterienne
associee. Races dites conjugantes et non conjugantes. Compt. Rend.
Acad. Sci., 193 : 206-208.

CHEJFEC, M., 1930. Zur Kenntnis der Kernreorganisations-prozesse bei Para-
mecium caudatum. Arch. f. Protist., 70: 87-118.

ERDMANN, R., AND L. L. WOODRUFF, 1916. The periodic reorganization process in
Paramecium caudatum. Jour. Exper. Zool., 20 : 59-83.

GIESE, A. C, 1935. The role of starvation in conjugation of Paramecium. Ph\siol.
Zool., 8: 116-125.

- AND M. A. ARKOOSH, 1939. Tests for sexual differentiation in Paramecium

multimicronucleatum and Paramecium caudatum. Ph\siol. Zool., 12 :

70-75.
GILMAN, L. C., 1939. Mating types in Paramecium caudatum. Am. Nat., 73 :

445-450.

HOPKINS, H. S., 1921. The conditions for conjugation in diverse races of Para-
mecium. Jour. Exper. Zool., 34 : 339-384.
JENNINGS, H. S., 1910. What conditions induce conjugation in Paramecium?

Joitr. E.vper. Zool., 9: 279-300.
— , 1938a. Sex reaction types and their interrelations in Paramecium bursaria.

I. Proc. Nat. Acad.'Sci., 24: 112-117.
— , \938b. II. Clones collected from natural habitats. Proc. Nat. Acad. Sci,

24: 117-120.
— , 1939a. Genetics of Paramecium bursaria. I. Mating types and groups, their

interrelations and distribution ; mating behavior and self sterility. Genetics,

24: 202-233.
— , 1939b. Paramecium bursaria : mating types and groups, mating behavior,

self-sterility; their development and inheritance. Am. Nat., 73: 414-431.
KIMBALL, R. F., 1937. The inheritance of sex at endomixis in Paramecium aurelia.

Proc. Nat. Acad. Sci., 23 : 469-474.
— , 1939. Change of mating type during vegetative reproduction in Paramecium

aurelia. Jour. Expcr. Zool., 81 : 165-179.
MAUPAS, E., 1889. Le rajeunissement karyogamique chez les cilies. Arch, dc

Zool. Exper. et Gen., 2e Sen, 7 : 149-517.
PARPART, A. K., 1928. The bacteriological sterilization of Paramecium. Biol.

Bull, 55: 113-120.
SONNEBORN, T. M., 1936. Factors determining conjugation in Paramecium aurelia.

I. The cyclic factor; the recency of nuclear reorganization. Genetics, 21:

503-514.

402 LAUREN C. OILMAN

— , 1937. Sex, sex inheritance and sex determination in Paramecium aurelia.

Proc. Nat. Acad. Sci, 23 : 378-385.
— , 1938a. Mating types in Paramecium aurelia : diverse conditions for mating

in different stocks ; occurrence, number, and interrelations of the types.

Proc. Amcr. Phil. Soc., 79: 411-434.
— , 1938ft. Mating types, toxic interactions, and heredity in Paramecium aurelia.

Science, 88 : 503.
— , 1939. Paramecium aurelia: mating types and groups; lethal interactions;

determination and inheritance. Am. Nat., 73: 390-413.
ZWEIBAUM, J., 1912. La conjugaison et la differenciation sexuelle chez les In-

fusoires (Enriques et Zweibaum). V. Les conditions necessaires et suf-

fisantes pour la conjugaison du Paramecium caudatum. Arch. f. Protisi..

26 : 275-393.

THE FUNCTION OF THE ANTENNAL RECEPTORS IN
LEPIDOPTEROUS LARVAE1

V. G. DETHIER

(From John Carroll University, Cleveland, Ohio)

The problem of correlating the different structural types of sensilla
with the various sensory perceptions recognized in insects has, with the
exception of vision and the mechanical senses, met with only fair suc-
cess. Olfactory sensilla are still so called primarily because of their
structure. Two of the greatest difficulties encountered in determining
experimentally the function of certain sensilla have been their wide
distribution over the body and the proximity of many different types on
a single appendage or given area. In this respect the antennae of lepi-
dopterous larvae are ideal organs for study since the sensilla thereon
are ten in number and comprise only five structural types. Advantage
has been taken of this situation to study not only the functions of the
antenna but also to attempt to assign some function to each of the
various sensilla located there.

Two sensory faculties have been ascribed to the larval antenna.
Blanc (1889) considered it simply a tactile organ. Nagel (1897) main-
tained that it was primarily an olfactory organ. Both of these authors
based their assumptions on structure alone. The purpose of this paper
is to point out the probable functions of the antenna as indicated by
experiments on several species of caterpillars and to attempt to correlate
this sensory activity with the five morphological types of sensilla borne
there.

MATERIAL AND METHODS

The species of caterpillars employed were determined largely by
their seasonal and numerical availability. Since the antenna is funda-
mentally the same in all larvae of this order (cf. Dethier, 1941) and the
reactions predictable irrespective of the species used, it was deemed un-
necessary to restrict experiments to a single species. Similarly inter-
specific comparisons seemed allowable. Third, fourth, and fifth instar
larvae of the following species were employed : Isia isabella A. & S.
(Arctiidae), Liparis dispar L. (Liparidae), Malacosoma disstria Hbn.
(Lasiocampidae), Cingilia catenaria Dru. (Geometridae), Pieris rapae
L. (Pieridae), and Nymphalis antiopa L. (Nymphalidae). It soon be-

1 The writer wishes to express his appreciation to Professor C. Ladd Prosser
of the Department of Biology, University of Illinois, for his generous assistance
in the electrical phases of this problem.

403

404 V. G. DETHIER

came apparent that age had no effect on the results obtained, hence larvae
that had attained size convenient for dissection became fit subjects for
experimentation. Approximately five hundred specimens were used.

It was proposed : ( 1 ) to record and measure action potentials on the
antennal nerve when the long hairs were stimulated by bending; (2) to
study the character of the olfactory sense and ascertain the effect of
antennal extirpation on the general threshold for response to odors ;
and (3) to record and measure action potentials on the antennal nerve
when the end organs were stimulated by odorous substances.

Action potentials on the antennal nerve were recorded photographi-
cally by means of a Matthews oscillograph and a resistance-capacity
coupled amplifier. Conduction of the experiment was rendered difficult
by reason of the toughness of the head capsule and general inaccessibility
of the antennal nerve as well as the extreme shortness of the nerve. A
larva to be tested was fastened ventral side down to a block of paraffin
by means of fine insect pins. The anterior surface of the head capsule
comprising the vertex and adfrontal areas was excised with a pair of
iridectomy scissors. Removal of this piece of cuticle left a window in
the front of the head capsule exposing the brain and its nerves. Ex-
perience showed that neither Ringer's solution nor a moist chamber prep-
aration were necessary because there was practically no drying out of
the nerves during the first hour of experimentation. After one hour
had elapsed, a fresh specimen was prepared. Due to the shortness of
the nerve (1.75 mm.), all recordings were mono-polar. A fine silver
wire hooked under the nerve raised it into the air. A thicker silver
wire led down into the tissues and body juices of the head. This lead
was used alternately as ground and grid. Unfortunately, recordings of
this nature rendered impossible any interpretation of wave form, and
spike heights were purely relative.

The general threshold for response to olfactory stimuli was measured
with a specially designed olfactometer. Caterpillars give no recognizable
response to attractant odors nor do they respond to repellent odors
unless the concentration is high. For these reasons it was necessary to
stimulate with repellent odors and to deliver measured concentrations
directly to the larvae.

A stream of nitrogen gas of constant velocity was saturated with the
test odor by being bubbled through a C. P. grade of the odorous liquid.
It was then diluted to the desired concentrations by the addition of
oxygen gas and the resultant mixture delivered to the chamber contain-
ing the larva. Complete saturation of the nitrogen was insured by the
use of three all-glass saturators in series. These were adaptations of the
type designed by v. Bichowsky and Storch (1915). The flow of gas in

FUNCTION OF LARVAL ANTENNA OF LEPIDOPTERA 405

each case was regulated by needle valves and measured by glass flow
meters (Benton, 1919). The gases were mixed in a chamber which
also acted as a valve preventing the more rapidly flowing gas from
blocking the flow of the less rapidly flowing one. A larva was confined
in a glass tube 8 mm. in diameter and 45 mm. long. A cap of fine-mesh
copper screening at either end prevented the escape of the animal yet
allowed the free passage through the tube of all gases. This tube was
then placed within the glass test chamber to which the mixed gases were
delivered and from which they escaped into the room. Saturators, mix-
ing chamber, and test chamber were immersed in a constant temperature
bath maintained at 20° C. This temperature was found to be optimum
for larvae and convenient for calculations of concentrations. A light
was suspended over the test chamber to illuminate the animal and cause
it to remain in a central position where it was observed with the aid of
a large magnifying glass clamped over the chamber. All connections
were glass to glass. Rubber tubing and stoppers, even when treated,
emitted odors. Cork stoppers were used wherever frequent breaking
of connections was necessary. They were replaced several times daily.
Inability to compress or secure compressed air necessitated the use
of a cheap, easily obtained gas. Nitrogen filled these requirements.
The addition of oxygen was necessary for the maintenance of life.
Two variables, the oxygen/nitrogen ratio and the gas velocity, consti-
tuted possible sources of error. Carefully controlled experiments indi-
cated that changes in the oxygen concentration of the gas mixture rang-
ing from 20 per cent to 100 per cent oxygen had no effect on the response
to odor. The average thresholds for larvae maintained at atmospheric
concentrations of oxygen did not differ significantly from those for
larvae maintained in a gas mixture containing about 99 per cent oxygen.
Therefore, the small changes in oxygen concentration introduced in
the course of the experiments (80 per cent to 99 per cent) required no
further control. Responses to variations in gas velocity accompanied
large and abrupt changes only. Thresholds could be determined at dif-
ferent gas mixture velocities since concentrations depended upon the
ratio of the rate of flow of oxygen to the rate of flow of nitrogen and
not on the total velocity. Thus it was found that the threshold of re-
sponse to odors remained unaffected by small velocity changes occurring
under usual experimental conditions. Concentrations were calculated
in terms of grams of solution (in this case benzaldehyde) per liter of
gas mixture from the following equation :

W= pa'Ma

Pb - Pa

406 V. G. DETHIER

where W • the number of grams of benzaldehyde per liter of gas
mixture, />a= = the vapor pressure of benzaldehyde in millimeters of
mercury at the temperature of the solution, 71/fa = = the molecular weight
of the benzaldehyde, PI, = the barometric reading in millimeters of
mercury, 7? = = the gas constant (0.08207 liter atmospheres), T- =the
absolute temperature, Fo2= = the ratio of the rate of flow of oxygen to
the rate of flow of nitrogen in liters per minute, and FN* = the rate of
flow of nitrogen in liters per minute. Humidity wras not controlled and
may have had some slight effect on the threshold values.

THE ANTENNAE

The antennae are located on the ventral lateral surface of the head
arising from the region of the postgenae near the bases of the mandibles.2
They are inserted into a membranous area in the head capsule known
as the antacoria. Each antenna is three-segmented (Fig. 1). The first
or basal segment sometimes contains four sensilla campaniformia. Upon
the second segment are located most of the antennal sensilla. At the
proximal end approximately in line with the larger hair is located a
single sensillum campaniformium. Next in order are two long thick-
walled hairs (sensilla trichodea). They are true hairs arising from-
articulation sockets. Distally there are always three sensilla basiconica,
two large and one minute. Segment three usually contains four sensilla
apically, a sensillum styloconicum, a large sensillum basiconicum, and
two small sensilla basiconica. The three large, hollow, thin-walled sen-
silla basiconica possess elaborately sculptured surfaces.

Three discrete sets of muscles inserted on the anterior mesal edge
of the base of the proximal segment effect the withdrawal of the an-
tenna. These muscles originate in the head capsule in the parietal re-
gion laterad of the adfrontal area. Extension of the antenna is regulated
by blood pressure.

A single nerve from the deutocerebrum innervates the antenna. A
basal branch terminates in the head capsule adjacent to the insertion of
the antenna; the antennal branch innervates the antenna proper. Each
of the two hairs is innervated by a single bipolar sense cell. Also in-
nervated by single bipolar sense cells are the sensillum campaniformium
and the small sensilla basiconica. Four bundles of primary bipolar sense
cells fill the greater part of the antenna distally. Each of the three large
sensilla basiconica is innervated by one bundle (Fig. 2). The fourth

2 A complete description of the antennae has been given in a previous commu-
nication (Dethier, 1941). Figures 1 and 2 are reproduced through the courtesy
of the Bulletin of the Museum of Comparative Zoology, Harvard College.

FUNCTION OF LARVAL ANTENNA OF LEPIDOPTERA 407

FIG. 1. Semidiagrammatic longitudinal section of the antenna (after Dethier,
1941). 1, 2 and 3, first, second and third segment respectively; Ml, M2 and M3,
first, second and third muscle bundle respectively ; T, trachea ; N, nerve ; SB, sen-
sillum basiconicum ; SS, sensillum styloconicum ; SC, primary bipolar sense cells.
(This figure is reproduced by courtesy of the Bulletin of the Museum of Com-
parative Zoology.}

bundle apparently innervates sensilla on the third segment. As many
as twenty-five cells have been counted in each bundle.

THE TACTILE SENSE

Scattered over the bodies of caterpillars are numerous thick-walled
hairs (sensilla trichoclea) differing markedly in length and diameter. It

408

V. G. DETHIER

SB_

ST_ _.

FIG. 2. Longitudinal section through the second antennal segment showing
the innervation of the hair and sensillum basiconicum (after Dethier, 1941). SB.
sensillum basiconicum; ST, sensillum trichodea ; 77, trichogen ; TR, tormogen ;
V, vacuole; SC, primary bipolar sense cells; N, nerve; HY, hypodermis ; TG,
tracheolar glomerulus. (This figure is reproduced by courtesy of the Bulletin of
the Museum of Comparative Zoology.)

has been demonstrated repeatedly that many of these subserve a tactile
function. The resemblance in structure of the two hairs characteris-
tically found on the antenna to tactile hairs of the body very early led
to the belief that they also were tactile hairs. The function of the
remaining sensilla was largely ignored.

By observing directly the response to tactile stimulation of the dif-

FUNCTION OF LARVAL ANTENNA OF LEPIDOPTERA 409

ferent sensilla and by recording action potentials on the antennal nerve,
it has been possible not only to verify the function of the large hairs but
also to determine the functional nature of some of the other sensilla.
Action currents were recorded by means of the apparatus and prepara-
tions already described. Hairs were stimulated first by touching gently
with another hair from the body of the animal, second by bending with
a fine glass needle. In the former case no responses were observed on
the intact animal nor any action potentials recorded on the antennal
nerve of the dissected animal. When the long hair was bent sufficiently
to cause movement of the hair within its socket, the intact animal re-
sponded by quickly withdrawing the antenna into the antennal socket.
Impulses were recorded when stimulation of this sort was applied to the
dissected animal. Since the nerve was severed centrally, there were no
spontaneous discharges nor motor activity from the brain. Spike
heights averaged 8 microvolts. After the long hair was cut away at its
base, stimulation of the short hair was possible. The tactile threshold
of this hair, judging from the amount of bending necessary to produce
a response, is much lower than that of the long hair. Following re-
moval of the short hair the remaining sensilla were stimulated tactually
without interference. Attempts were made to touch each sensillum
individually. This proved exceptionally difficult. A micro-manipulator
and fine glass needles were tried without success. Finally stimulation
was accomplished by means of a fine hair from the animal's body. When
the sensilla on the third antennal segment were touched, immediate with-
drawal of the antenna resulted. It was impossible, however, to touch
the large sensillum basiconicum without also touching the sensillum
styloconicum. Attempts to touch either of the large sensilla basiconica
on segment two without also stimulating the minute sensillum basiconi-
cum were not entirely successful. It was possible with an animal pos-
sessing an abnormal antenna which lacked large sensilla basiconica to
stimulate the minute sensilla basiconica. This much is certain. Light
tactile stimulation of the minute sensilla basiconica resulted in a re-
sponse; bending of the large sensilla basiconica sufficiently to cause
strains in the surrounding surface cuticle also caused a response. No
other part of the antenna appeared sensitive to tactile stimulation.

When the cuticle in the region of the sensillum campaniformium was
deeply depressed by pressure with a glass needle, withdrawal of the an-
tenna resulted. This region, however, was not sensitive to light touches.
Movement of the antenna as a whole on the dissected animal with the
nerve cut centrally resulted in a burst of impulses. These probably
resulted from stimulation of the sensillum campaniformium since no
muscle receptors have been found associated with the antennal muscles

410

V. G. DETHIER

THE OLFACTORY SENSE

The existence of an olfactory sense in caterpillars was first demon-
strated by Mclndoo (1919). Larvae of several species responded to
the odors of essential oils. End organs on the distal segments of the
antennae and maxillae were suggested in a previous communication
(Dethier, 1937) as the probable olfactory receptors. The ability of the
larvae to respond to essential oils is impaired or destroyed by removal of
these areas. Gotz (1936) maintained that the olfactory sense was not
localized in the antennae or maxillae insomuch as larvae continued to
feed following extirpation of these appendages. This conclusion does
not seem justifiable because the initiation of feeding is not always de-
pendent upon chemotaxis. Caterpillars frequently attempt to eat odor-

TABLE I

Olfactory response and longevity of Cingilia catenaria following operations to the

antennae and maxillae

Specimen
no.

Date of
operation

Response

Date of
operation

Response

Date of
pupation

Date of
emergence

182

Maxillae

Excellent

Antennae

None

7/25/37

8/8/37

removed

6/30/37

7/3/37

197

Antennae

Excellent

Maxillae

None

7/15/37

7/28/37

removed

6/30/37

7/9/37

198

Antennae

None

—

7/12/37

7/26/37

and

maxillae

removed

7/1/37

less cellulose materials. They are known to eat their way to freedom
when confined in cardboard containers.

Repeated experiments have shown that extirpation of the antennae
does not abolish responses to essential oils or odorous liquids held close
to the head. Nor does removal of the maxillae completely destroy the
olfactory sense. Only when both pairs of appendages are removed is
the sense of smell destroyed. As long as the operation was carefully
executed a larva suffered no permanent ill effects. Parts were removed
by cutting with a microscalpel. Recovery from surgical shock was com-
plete thirty minutes after cutting. As a precautionary measure, how-
ever, experiments were not conducted until twenty- four hours had
elapsed. Best results were obtained without the use of anesthetics.
Two sets of experiments demonstrated that failure to respond to odors
following removal of both antennae and maxillae was not due to more

FUNCTION OF LARVAL ANTENNA OF LEPIDOPTERA 411

severe surgical shock than the removal of either pair alone. (1) Re-
generation of antennae or maxillae in subsequent instars was accom-
panied by return of the olfactory sense. (2) Operations involving in-
jury to more and larger nerves, as removal of all thoracic legs or de-
capitation, did not result in shock of sufficient severity to impair the ol-
factory sense. Mortality rates were less than 1 per cent. It may be
seen from Table I that these operations were not sufficiently shocking to
interfere with the normal life processes of the insect.

Two facts were noted in the course of these experiments which sug-
gested that a determination of the threshold for response to olfactory

-40

-30

-20

-I 0

385

425

465

505

545

585

625

665

GRAMS OF BENZALDEHYDE PER LITER OF GAS (x 1 0~7 )

FIG. 3. Change in the threshold sensitivity of larvae of P. rapac L. to the
odor of benzaldehyde following amputation of the antennae. The open histogram
represents the threshold of response of normal larvae ; the cross-hatched histogram,
that of larvae lacking antennae.

stimuli might shed further light on the nature and locus of the olfactory
sense. (1) Larvae lacking maxillae apparently responded more rapidly
than larvae lacking antennae. (2) Occasional specimens lacking both
antennae and maxillae gave questionable, feeble responses when strong
odorous liquids, as turpentine or ammonia, were held not more than one
millimeter away from the head for a period of sixty seconds or longer.
Accordingly the threshold values for response to benzaldehyde were
determined for normal larvae of Pieris rapae as described above. It
was found that the average threshold for response was 580 X 10~7
± 16.49 X 10~7 grams of benzaldehyde per liter of gas mixture. For
larvae from which the antennae had been removed the threshold was

412

V. G. DETHIER

770 > ' 10~7 ± 3.7 ] ; 10~7 grams per liter. This difference in the aver-
age thresholds is eleven times the standard error of the difference
(Fig. 3). No significant rise in threshold could be demonstrated for
larvae from which the maxillae had been removed. Unilateral extir-
pation of either appendage likewise caused no rise in the threshold value.
This would seem to indicate, first that the antennae did actually possess
olfactory end organs, and second that the olfactory threshold of the an-
tennal end organs was lower than that of the maxillary end organs.

In 0.5 per cent of all of the tests on animals lacking both antennae
and maxillae a response to benzaldehyde could be obtained. The lower
limit of the threshold for response in these cases was 1 X 10~3 grams
per liter or approximately 17 times the same value for a normal indi-
vidual. Although this could be interpreted as meaning that olfactory

lii iliii ili

FIG. 4. a. Record of responses on antennal nerve to stimulation of tactile hairs
on antenna. Time signal at top of record shows intervals of 0.01 second.

b. Amplifier baseline with calibrating signal of 8 microvolts. Time signal,
0.01 second.

end organs are also borne on parts of the body other than the antennae
and maxillae, it seems more probable that odors in such high concen-
trations acted as irritants and were stimulating organs similar in nature
to the common chemical sense of man.

A further effort was made to confirm these findings by recording
action potentials on the antennal nerve when the receptors on the tip of
the antenna were stimulated by strong odors applied from a short dis-
tance. Turpentine, oil of wintergreen, oil of cloves, and benzaldehyde,
which cause the animal to react under normal conditions, were used as
test substances. Recording methods and preparations were as described
above.

With this preparation no action potentials were recorded upon stimu-
lation of the antennal sensilla. Spikes due to accidental tactile stimula-

FUNCTION OF LARVAL ANTENNA OF LEPIDOPTERA 413

tion of the hairs on the tip of the antenna did appear on the record.
They were of the order of magnitude of 8 microvolts. Failure to detect
potentials from chemoreceptors was thought to be due to the fact that
the potentials from these receptors were of too low a magnitude to be
distinguished from the ordinary fluctuations of the base line. That
this was actually the case is proven below by application of Erlanger
and Gasser's equation which states that the recorded potential of action
varies as the square of the outside diameter of the axon.

A study of longitudinal as well as cross-sections of the antenna has
shown that the largest sensory nerve fibers in the antenna innervate the
tactile hairs. The average outside diameter of these fibers is 3 micra.
The average diameter of the remaining fibers is 0.8 micra. The am-
plitude of spikes resulting from stimulation of the tactile hairs was of
the order of magnitude of 8 microvolts. Measurements of spike am-
plitude were taken on a calibrated record of potentials from this nerve.
Application of these values to the equation shows that the value of spikes
from any of the other fibers would be of the order of 0.57 microvolts.

D? ZV 0.64 9

Pi ' P2 x ~ 8

With the electrodes far apart on fine nerves raised into the air resistance
was exceedingly high. As a result there was a high noise level of the
order of magnitude of 2 microvolts. Thus it was impossible to dis-
tinguish potentials of the order of magnitude of 0.5 or even 1.0 micro-
volt when the fluctuations of the base line approximated 2.0 microvolts.

DISCUSSION

The present experiments confirm the assumptions of earlier workers
that the antennae of lepidopterous larvae are tactile and olfactory organs.
All of the evidence favoring the tactile nature of the large sensilla
trichodea is direct. It is hardly necessary to add that these hairs by
their structure are ideally adapted for the reception of tactile stimuli.
Any object or vibration impinging upon either hair of sufficient mag-
nitude to cause movement of the shaft within its socket may be an ade-
quate stimulus. Of such a nature are air currents, vibrations of the
substratum, and preeminently, shocks imparted to the hairs by explora-
tory movements of the antenna. Stimuli of this last sort serve to in-
form the animal of obstacles in its path and of the contours of the sub-
stratum. In this respect other hairs on the feet and mouthparts are of
nearly equal importance. Finally the antennal hairs protect the more
delicate antennal sensilla in that by stimulation thereof the animal is
made aware of potentially injurious objects. The antennal hairs are

414 V. G. DETHIER

not alone in this service. Stimulation of numerous structurally similar
tactile hairs on the parietal region of the head capsule results in with-
drawal of the antenna.

It is to be expected that the longer hair would have the higher
threshold since it is continuously stimulated. Stimulation of the short
hair signals proximate danger to the antenna.

Direct evidence likewise indicates that the sensillum campaniformium
responds to bending of the adjacent cuticle. It is likely that this sen-
sillum is a proprioceptive organ. Movement of the antenna within its
socket results in a burst of impulses on the antennal nerve (severed
centrally). These undoubtedly originate with the sensillum campani-
formium or with the minute nerves ending freely in the region of the
hypodermal bulb. These may also be proprioceptive in nature.

The olfactory nature of the large sensilla basiconica still rests on
indirect evidence. All experiments point to the olfactory function of the
antenna. By a process of elimination the large sensilla basiconica must
be olfactory end organs. Responses to tactile stimuli are observed,
however, when these end organs are bent with a hair. Some doubt re-
mains as to whether the response is due to direct stimulation of the
sensilla basiconica or to transmission of the stimulus through the sur-
rounding cuticle to the small tactile sensilla basiconica. It is also pos-
sible that in the aggregate of sense cells innervating the large sensilla
basiconica there are tactile receptors as well as chemoreceptors. In
other words, a large sensillum basiconicum may be a composite end
organ. Histological examination, however, reveals no differences among

the cells.

LITERATURE CITED

BENTON, A. F., 1919. Gas flow meters for small rates of flow. Jour. Ind. and

Eng. Chcm.. 11 (7) : 623-629.
BICHOWSKY, F. R. VON, AND H. SxoRCH, 1915. An improved form of gas-washing

bottle. Jour. Am. Chem. Soc., 37 (12) : 2695-2696.
BLANC, L., 1889. La tete du Bombyx mori a 1'etat larvaire. Travaux du Labora-

toirc d'Etudcs de la Sole, Lyon, pp. 163-340.
DETHIER, V. G., 1937. Gustation and olf action in lepidopterous larvae. Biol. Bull..

72 (1) : 7-23.
DETHIER, V. G., 1941. The antennae of lepidopterous larvae. Bull. Museum

Comp. Zool, 87 (6) : 455-507.
ERLANGER, J., AND H. S. GASSER, 1937. Electrical Signs of Nervous Activity.

Philadelphia.
GOTZ, B., 1936. Beitrage zur Analyse des Verhaltens von Schmetterlingsraupen

beim Aufsuchen des Putters und des Verpuppungsplatzes. Zcitschr. f.

vergl. PhysioL, 23 (3) : 429-503.
MclNDOO, N. E., 1919. The olfactory sense of lepidopterous larvae. Ann. Ent.

Soc. Amer., 12 (2) : 65-84.
NAGEL, W. A., 1897. Vergleichend physiologische und anatomische Untersuch-

ungen iiber den Geruchs- und Geschmacksinn und ihre Organe. Bibl.

Zool, 18 : 1-207.

STUDIES ON THE LIFE HISTORY OF ANISOPORUS

MANTERI HUNNINEN AND CABLE, 1940

(TREMATODA : ALLOCREADIIDAE)

A. V. HUNNINEN 1 AND R. M. CABLE

(From Oklahoma City University, Purdue University, and the
Marine Biological Laboratory)

INTRODUCTION

In August, 1939, the authors discovered a new cotylomicrocercous
cercaria emerging from the marine snail, Mitrclla lunata (Say) collected
at Waquoit Bay, Cape Cod, Massachusetts. During the summer of
1940, it was found that this cercaria penetrated marine amphipods and
developed into progenetic metacercariae of a new species of Anisoporus,
for which the name Anisoporus manteri was proposed in a preliminary
abstract (Hunninen and Cable, 1940).

The genus Anisoporus was erected by Ozaki (1928) to contain A.
cobraeformis from the intestine of Diacocus pctersoni. Two other
species of Anisoporus from marine fishes have been described by Manter
(1940), — A. eucinostomi and A. tliyrinopsi which is described tenta-
tively, being possibly a developmental stage of A. eucinostomi. Aniso-
porus possesses a small accessory sucker anterior to the acetabulum,
thereby differing from the closely related genera Opecoclus and Opegaster.

There has been considerable difference of opinion concerning the
taxonomic significance of anal openings such as those found in Aniso-
porus, Opecoclus, Opegaster and several other genera of digenetic
trematodes. Ozaki (1925) regarded anal openings as fundamental
characters and accordingly proposed the family Opecoelidae to include
the genera Opecoclus and Coitocaccum. Later (1928), he placed the
new genera Anisoporus and Opegaster in the Opecoelidae but removed
Coitocaccum, making it the type of a new family, Coitocaecidae. In
the same paper, he also erected the family Diploproctodaeidae, with
Diploproctodacum La Rue as type genus and included two new species
of the new genus Diplopoms. Odhner (1928) erected the genus
Opccoeloides and expressed the opinion that Opecoelus and Opecocloides,
because of their resemblance to Podocotyle, were aberrant allocreadiids.
Ozaki (1929) reaffirmed his belief that anal openings were fundamental

1 This work was assisted by a grant-in-aid to the senior author from the So-
ciety of the Sigma Xi.

415

416 A. V. HUNNINEN AND R. M. CABLE

characters. Stunkard (1931) has reviewed the literature concerning
trematodes with anal openings and expressed the opinion that these
structures have arisen independently in several families and are not of
great taxonomic importance. He stated that the members of Ozaki's
families Opecoelidae and Coitocaecidae might well be regarded as a
subfamily, Opecoelinae, of the family Allocreadiidae. La Rue (1938)
concludes that anal openings have little more than specific or generic
value at the most in the taxonomy of digenetic trematodes.

The described species of freshwater and marine cotylomicrocercous
cercariae and the life histories of trematodes known to have this type
of larva are listed by Cable (1938, 1939) and Dobrovolny (1939a).

MATERIALS AND METHODS

Material was collected in abundance from Waquoit Bay, near Woods
Hole, Massachusetts, and studied mostly while living. Stained whole
mounts and serial sections of certain stages were prepared, using con-
ventional technics. Cercariae were studied with the aid of neutral red
and Nile blue sulphate supravital stains. The morphology of the meta-
cercaria and adult was observed in specimens from both naturally and
experimentally infected amphipods and fishes. Amphipods were in-
fected experimentally by placing them with infected snails in finger
bowls which were covered with cheese-cloth and placed in slowly running
sea water. All measurements given below are in millimeters. Eleven
adults were measured as stained whole mounts ; all other measurements
were made on living material under light cover-glass pressure.

OBSERVATIONS
Experimental Proof of the Life History

Technical difficulties and brevity of the season made it impractical
to rear parasite-free amphipods and fishes in the laboratory. For this
reason, proof of the life history is based on the fact that penetration
of the cercariae into amphipods could be induced at will and subsequent
development followed as a continuous process from very young to
mature, progenetic metacercariae with eggs in the uterus and discharged
from the body into the surrounding cystic fluid. Amphipods exposed
to cercariae in the laboratory usually contained a larger number of
metacercariae and always showed a much higher incidence of infection
than did amphipods collected in the field. Experimental infections
superimposed on natural infections could be detected by differences in
cyst size.

f
LIFE HISTORY OF ANISOPORUS 417

Fishes were infected experimentally by feeding them large numbers
of amphipods. Worms recovered from experimentally as well as
naturally infected fishes were little further developed than were the
progenetic metacercariae.

Description of Stages in the Life Cycle
Adult (Figs. 7-10)

Specific Diagnosis. — Small, elongate worms with characters of the
genus Anisoporus, Total length 0.74—2.32 (average 1.54) ; width 0.28-
.46 (0.35). Oral sucker width 0.07-.14 (0.114) ; acetabulum in anterior
third of body, 0.12-.17 (0.144) wide, provided with three anterior and
two posterior papillae; sucker ratio approximately 2 : 2.5. Prepharynx
very short, pharynx spherical, 0.07-.11 (0.09) in diameter; esophagus
length 0.07-.18 (0.124) ; ceca extend almost to posterior end of body,
uniting with excretory bladder, the excretory pore functioning as an anal
opening. Testes tandem, in posterior half of body; anterior testis
0.13-.2 (0.16) wide and 0.07-.16 (0.12) long; posterior testis 0.14-.2
(0.18) by 0.07-.17 (0.13) ; cirrus sac lacking. Ovary ovoid, median,
anterior to testes, 0.09-.14 (0.11) wide and 0.05-.12 (0.09) long;
seminal receptacle lacking ; Laurer's canal present. Uterus anterior to
ovary, with few coils ; vitelline follicles large, beginning just behind
acetabulum and extending to posterior end of body, the lateral fields co-
alescing posteriorly ; two accessory vitelline ducts extending trans-
versely, one anterior to ovary, the other posterior to testes. Eggs 0.062-
.068 (0.065) by 0.035-.04 (0.038). Excretory vesicle sac-shaped; ex-
cretory formula 2[(2 + 2) + (2 + 2)].

Hosts. — Northern pipefish, Syngnathus fuscus Storer ; flounder,
Paralichthys dentatits (Linnaeus) ; sand dab, Hippoglossoides plates-
soides (Fabricius) ; four-spined stickleback, Apeltes quadracus (Mitch-
ill) and the killifishes, Fundulus heteroclitus (Linnaeus) and F. majalis
(Walbaum).

Locality. — Waquoit Bay, Cape Cod, Massachusetts, U. S. A.

Type Specimens. — Holotype No. 36781 and Paratype 36782, Hel-
minthological Collection, U. S. National Museum.

The body is elongate, tapering slightly at both ends. The cuticula
is aspinose and is modified near the anterior end of the body to form
small papillae (Fig. 9), each set with a very delicate "hair." The
papillae are visible only in living material.

The ventral sucker is embedded in a large, stalk-like protrusion of
the body, making it very difficult to mount worms so that the sucker is

418 A. V. HUNNINEN AND R. M. CABLE

not displaced to one side. This displacement always causes distortion
which alters the relationships of various structures. The characteristic
lobes on the margin of the ventral sucker are shown in Fig. 6. The ac-
cessory sucker (Fig. 7) is seen more distinctly in living than in fixed
and stained specimens and hence may have been overlooked in species
at present assigned to genera other than Anisoporus. It is ventral in
position, about midway between the acetabular stalk and the pharyngeal
level, and appears to lie slightly to the left of the midventral line. The
sucker has no connection with the genital pore which lies at the posterior
end of the pharynx.

The shape of the oral sucker and pharynx depends on their state of
contraction, being either subspherical or slightly wider than long. The
short prepharynx is evident only in extended specimens. The esopha-
gus bifurcates at the level of the acetabulum. In the living worm, a
patch of tiny papillae is seen where the ceca join the bladder. These
papillae are in the bladder proper and in sections resemble the inner
processes of the muscle cells in Ascaris. They are especially noticeable
during rhythmic contractions of the posterior end of the body.

The testes are intercecal ; in moderately contracted worms, they lie
close together, one behind the other, are definitely wider than long, and
without notches or lobes. In extended specimens, the testes lie some
distance apart and are spherical in shape. The vasa efferentia extend
anteriorly and unite to form a very short vas deferens. The seminal
vesicle is long, beginning well behind the acetabular level, almost as far
back as the ovary in contracted specimens. The vesicle is continuous
with a narrow, delicate ejaculatory duct which is difficult to trace as it
approaches the genital pore.

The ovary lies in front of and in contact with the anterior testis.
From the anterior surface of the ovary, the ciliated oviduct (Fig. 10)
bends abruptly to the right, extends a short distance, then turns an-
teriorly and is joined immediately by the Laurer's canal. The canal
crosses to the left of the median line and opens dorsally. From the
junction of the oviduct and Laurer's canal, the ootype and uterus extend

EXPLANATION OF PLATE I
(All figures concern Anisoporus manteri)

FIG. 1. Cercaria, ventral view.

FIG. 2. Stylet of cercaria, dorsal view.

FIG. 3. Metacercaria, 2-day infection.

FIG. 4. Amphipod with a moderately heavy infection with metacercariae.

FIG. 5. Metacercaria, showing excretory system and other details of structure.

FIG. 6. Ventral view of acetabulum, showing characteristic papillae.

PLATE I

420 A. V. HUNNINEN AND R. M. CABLE

anteriorly as a moderately sinuous tube. The uterus usually contains
only a few eggs but as many as 80 have been counted. On each side
of the body, an anterior and posterior vitelline duct join to form the
transverse common vitelline duct which is expanded medially to form
the vitelline reservoir. This reservoir joins the ootype just posterior
to Mehlis' gland. The longitudinal vitelline ducts are connected by a
pair of transverse accessory ducts (Fig. 7), one anterior to the ovary,
the other posterior to the testes. These transverse ducts are clearly
visible only when rilled with vitelline material and may be overlooked
when empty.

The excretory vesicle is a long, simple tube with small cells scat-
tered over its inner surface. It extends to the anterior border of the
anterior testis. The main excretory tubules are ciliated for about
three-fourths their length and reach from the anterior end of the
vesicle almost to the acetabular level where each divides to form an
anterior and a posterior collecting tubule. Each collecting tubule re-
ceives two secondary tubules, each of which is joined by the capillaries
of two flame cells. The excretory formula remains unchanged during
post-cercarial development. The flame cells are large, averaging 0.012
mm. in length.

Anisoporus mantcri is compared with described species of Anisoporus
in Table I. A. nianteri differs significantly from A. cobraeformis in

EXPLANATION OF PLATE II
(All figures concern Anisoporus manteri)

FIG. 7. Ventral view of an extended adult specimen (vitelline follicles
omitted).

FIG. 8. Sagittal section of adult, showing junction of intestinal ceca and ex-
cretory vesicle.

FIG. 9. Dorsal view of adult specimen.

FIG. 10. Details of reproductive system, drawn freehand from living speci-
men.

FIG. 11. Amphipod appendage containing five metacercariae.

ABBREVIATIONS

A, anus. OO, ootype.

AC, anterior vitelline commissure. OV , ovary.

AP, acetabular papillae. PC, posterior vitelline commissure.

AS, accessory sucker. PH, pharynx.

E, esophagus. PP, prepharynx.

EG, egg. SV , seminal vesicle.

EV ' , excretory vesicle. U, uterus.

GP, genital pore. V, vitelline follicle.

/, junction of ceca and excretory ves- VD, vitelline duct.

icle. VE, vas efferens.

LC, Laurer's canal. VR, vitelline reservoir.
MG, Mehlis' gland.

PLATE II

422

A. V. HUNNINEN AND R. M. CABLE

size of body, suckers, testes, ovary and eggs, and the position of the
genital pore. A. eucinostonii and A. inanteri are similar in all respects
except egg size and shape of the ovary. All three species differ widely
in respect to host and locality.

Metacercaria (Figs. 3 and 5)

Metacercariae occur in the haemocoele throughout the body of the
marine amphipods, Carinogammarus mucronatus (Say) (Fig. 4) and
Amphithoe longimana Smith. Of 239 amphipods (143 C. mucronatus
and 96 A. longimana) examined for natural infections, 10 per cent were

TABLE I

Comparison of Species of Anisoponis

Species

A. cobraejormis
Ozaki,
1928

A. eucinostomi
Manter,
1940

A. manteri
Hunninen and
Cable, 1940

Length (mm.)

4.3-7.2

1.222-2.497

0.74-2.32

Width (mm.)

0.33-.5

0.345-.465

0.28-.46

Oral sucker
width (mm.) ....

0.16-.21

0.109-.144

0.07-.14

Ventral sucker
width (mm )

0 2 28

0 12 17

Ovary

width (mm )

0 13 23

0 09 14

shape

. . . . globular

stibtriangular

ovoid

position

. . . . separated from

close to an-

Testes
width (mm )

anterior testis
0 16- 33

terior testis

terior testis
0 14 2

Eggs
length ....

0.044-.046

0.041-.048

0.062-.068

width

0.03-.033

0.025-.029

0.035-.04

Genital pore
position ....

. . . closer to ventral

very close to

Locality

than to oral sucker
. . . . Japan

pharynx
Galapagos Is.

pharynx
Massachusetts,

U. S. A.

positive. The number of cysts per amphipod varied from one to seven,
averaging between two and three.

The encysted worm is folded on itself (Fig. 3) and closely sur-
rounded by an elastic cyst membrane 0.004 mm. thick. The membrane
is easily ruptured with a needle. The size and shape of the metacer-
caria depend on the age of the infection, young cysts being spherical and
less than 0.15 mm. in diameter; older ones are ovoid and measure as
much as 0.785 by 0.675 mm. In large cysts the worms become sexually
mature and eggs are laid within the cyst. One metacercaria contained
24 eggs, most of which were free in the cyst fluid. These eggs measured
0.061-.66 (av. 0.064) mm. in length by 0.035-.39 (0.038) mm. in width

LIFE HISTORY OF ANISOPORUS 423

and appeared to be as normal as those in worms removed from the
definitive host.

Amphipods were easy to infect experimentally. In one group of 1 1
amphipods exposed for four days to cercariae, one was negative on ex-
amination while ten contained 4, 4, 7, 7, 8, 8, 13, 13, 15 and 17 small
metacercariae, all of about the same size. In another experiment, 11
amphipods were exposed to cercariae for five days and found upon
examination to contain 1, 1, 6, 7, 7, 10, 10, 12, 14, 28, and 334 cysts.
Two amphipods in this group were naturally infected, one with two
cysts and the other with one, all three being considerably larger than
the metacercariae from experimental infections of the same amphipods.
Exposure of amphipods to large numbers of cercariae results in heavy
infections in which the cysts are found throughout the body, even in
the antennae and legs (Figs. 4, 11).

In 2-day-old metacercariae (Fig. 3), the stylet is still present but is
absorbed between the third and fourth days. Soon after encystment.
the relatively long prepharynx of the cercaria begins to shorten so that
in older metacercariae it is visible only when the body is extended.
Dobrovolny (1939fr) has described a similar shortening of the pre-
pharynx in Plagioporus lepomis during post-cercarial development.

Cercaria (Figs. 1—2)

Specific Diagnosis. — Modified cotylomicrocercous type. Body con-
tracted 0.12 long, extended over 0.325, average 0.225; cuticula aspinose ;
oral and ventral suckers lined with fine spines ; a circlet of about 10
papillae with sensory " hairs " surrounding mouth. Tail extended
0.048-.056 long and 0.018 wide near base, moderately contracted 0.028-
.039 long and 0.031 wide, filled with glands the ducts of which form a pro-
trusible papilla. Oral sucker 0.035 long and 0.031 wide; stylet double
pointed, 0.01-.011 long and 0.005 wide; prepharynx long and slender,
pharynx 0.011 in diameter. Ventral sucker 0.032 in diameter. Three
pairs of cephalic glands with a single lateral and two median ducts on
each side. Oval excretory vesicle filled with granular masses. Main
excretory tubules ciliated, dividing at level of acetabulum to form an-
terior and posterior collecting tubules. Excretory formula 2 [(2 + 2)
-}- (2 -(- 2)]. Develop in sausage-shaped sporocysts with terminal birth
pore.

Host. — Mitrella lunata (Say).

Locality. — Waquoit Bay, Cape Cod, Massachusetts, U. S. A.

The cercaria of A. manteri differs from described cotylomicrocer-
cous cercariae in that the tail, instead of forming a hollow sucking cup,
is filled with large gland cells which secrete a sticky substance. The tail
becomes so firmly attached to objects that the cercaria is not dislodged

424 A. V. HUNNINEN AND R. M. CABLE

by water currents of considerable force. In addition to extension and
contraction of the body during the waving, exploratory, and inch-worm
movements characteristic of most cotylomicrocercous cercariae, the
larvae of A. mantcri have a peculiar type of behavior, commonly ob-
served in attached specimens. The extended cercaria loops on itself so
that its shape is approximately that of the letter " e " written vertically.
The body is then straightened with a sudden spiral movement. This
coiling and uncoiling movement sometimes gives the illusion that the
cercaria reverses ends.

Sporocysts of A. ina uteri occur in the branchial region and digestive
gland of the snail. They are simple, sausage-shaped forms with a pro-
trusible anterior end bearing the birth pore. The largest sporocyst
measured was 0.65 mm. long and contained 35 apparently mature

cercariae.

DISCUSSION

The present study demonstrates that Anisoporus and probably the
related genera Opecoelus, Opecoeloides, and Opegaster are co-familial
with other trematodes having cotylomicrocercous cercariae and do not
constitute a distinct family as maintained by Ozaki (1925, 1929).
Hence, Odhner (1928) and Stunkard (1931) are supported in their
opinion that anal openings in these trematodes are of significance only
in the separation of genera and species.

If, in the phylogeny of the trematodes, convergent evolution has
occurred, one would expect to find evidences of it in the parasites of
fishes, the oldest class of vertebrates. Convergent evolution would ac-
count for similarity of adult stages of species having fundamentally dif-
ferent larvae as now well demonstrated in the Allocreadiidae. On the
other hand, divergent evolution ^apparently has led to the separation of
adults which appear to be distantly related but actually have similar larval
stages. An excellent illustration is afforded by the separation of the
Acanthocolpidae from the Allocreadiidae on the basis of spines in the
cirrus and metraterm. According to Martin (1939), the acanthocolpid,
Stephanostornwn tenue, has an ophthalmoxiphidiocercaria while the au-
thors (Cable and Hunninen, 1940) have found that another acantho-
colpid, Deropristis inflata, has a trichocercous larva. Not only are these
larvae dissimilar, but each displays characteristics common to certain
members of the family Allocreadiidae. Hence a classification based on
actual relationships will require the combination and reclassification of
the Acanthocolpidae, Allocreadiidae, and possibly the Monorchiidae.

The task of reclassifying the group would be a formidable one and
must evaluate both larval and adult characters with great care. Larval

LIFE HISTORY OF ANISOPORUS 425

and particularly cercarial structures may actually be misleading in some
cases, particularly in groups in which only a few life histories are known.
The elimination of a free-swimming period during cercarial life is often
accompanied by extreme reduction and even complete loss of the tail ;
such modification may occur even in cercariae which emerge from
aquatic hosts. A good illustration is afforded by the brachylaemids
whose cercariae were known until very recently only from pulmonate
gastropods, many of which are terrestrial. In these hosts the tails of
the cercariae are extremely rudimentary or lacking altogether. Re-
cently, however, Allison (1940) has found in the prosobranch snail,
Campeloma, a furcocercous brachylaemid cercaria which, as he states,
strongly suggests a relationship between the Brachylaemidae and other
trematodes having furcocercous cercariae.

When considered only as formulae, excretory systems also may be
misleading, as exemplified by Cercaria coronanda Rothschild, 1938, the
cercaria of Exorchis ovifonnis as described by Komiya and Tajimi
(1940), and an undescribed species discovered by the authors, all of
which are true pleurolophocercous cercariae but with an excretory
formula of 2[(2 -f 2) _|_ (2 + 2)]. This formula is the same as that
of the microphallids but a consideration of other larval characteristics
gives no reason to consider the cercariae mentioned as intermediate types
between the pleurolophocercous larvae of the Opisthorchioidea and the
xiphidiocercous larvae of the Microphallidae. It is clear that the ex-
cretory pattern must be correlated with other larval characteristics, all
of which must be considered as a whole.

There are pleurolophocercous cercariae without eye-spots, cotylomi-
crocercous forms without stylets, microphallid larvae without tails, and
strigeid fork-tails without pharynges. Yet other larval characteristics
give definite clues to relationships.

Much has been accomplished by the study of adult material as indi-
cated by the manner in which life-history studies have confirmed rela-
tionships postulated on the basis of adult morphology. Structure of
adults may be the deciding factor in the classification of species having
cercariae so modified that they might be considered either as aberrant
members of some well-defined group or as a distinct larval type. For
example, the cercaria of Monorcheidcs ciiiuingiac' (Martin, 1938) differs
from typical trichocercous species in the nature of the tail and excretory
pattern and may be a separate larval type. Determining the relationship
of this species to other trematodes, and hence the validity of the family
Monorchiidae, may depend as much on adult as larval characters, particu-
larly until more than one life cycle in the family is known,

426 A. V. HUNNINEN AND R. M. CABLE

It is concluded that a revision of the Allocreadiidae is needed and
at present could be proposed to the extent of defining families and plac-
ing certain genera in them. Even so, our knowledge of the morphology
and life histories of many genera is so incomplete that their allocation
to families would be a matter of conjecture.

We believe that the trematodes at present assigned to the family
Allocreadiidae represent at least three distinct families, possibly belong-
ing to more than one super family. This belief is justified if the hetero-
phyids and opisthorchiids, with practically identical cercariae, are cor-
rectly regarded as separate families of the same super family.

Very recently, Hopkins (1941) has accepted the validity of the
family Opecoelidae and included in it all allocreadoid genera having an
excretory formula of 2 [(2 + 2) + (2 + 2)], and cercariae of the coty-
lomicrocercous type. Leaving all other genera in the family Allocre-
adiidae, he includes in the Opecoelidae the genera Cymbephallus, Podo-
cotyloidcs, Enenterum, Dactylosomum, Coitocaccum, Genitocotylc, Ni-
colla, Ozakia, as well as the more typical opecoelid genera Opecoelus,
Opegastcr, Opecoeloldcs, Anisoporus, and Opecoelina. Hopkins also
implies but does not definitely state that Helicomctra, Plagioporus,
Hamacreadium, Sphaerostoma, and Podocotyle also should be included.
Since Mathias (1937) has found that Allocreadlum angusticollc has a
cotylomicrocercous cercaria, the genus Allocreadlum also should be in-
cluded in the above group. Obviously, the name Allocreadiidae is avail-
able only for the family including the type genus Allocreadlum; this
genus must be regarded as co- familial with the genera Hopkins allocates
to the family Opecoelidae. Either the name Allocreadiidae or Ope-
coelidae must be suppressed. Since Allocreadiidae is an older and
more familiar name than Opecoelidae and Ozaki proposed the family
Opecoelidae without knowledge of either excretory systems or life his-
tories, but separated it from the Allocreadiidae on the basis of characters
which have not been generally accepted as valid, it is proposed that the
name Allocreadiidae take precedence over Opecoelidae. This proposal
simply means that the family Allocreadiidae is restricted to include only
those forms having cotylomicrocercous cercariae and a simplified excre-
tory pattern. In any event, it will be necessary to propose new families
or redefine existing ones to include the genera excluded from the re-
stricted family. Such a revision is beyond the scope of the present
paper.

SUMMARY

The life history of Anisoporus manterl Hunninen and Cable, 1940,
has been traced experimentally. The cercaria, a cotylomicrocercous type,

LIFE HISTORY OF ANISOPORUS 427

develops in sporocysts in the marine snail, Mitrella lunata (Say) and en-
cysts in the marine amphipods, Carinogammarus mucronatus (Say) and
Amphithoe longimana Smith. Old metacercariae contain eggs in the
uterus and cystic fluid. Adult worms occur in the intestine of the
marine fishes, Syngnathus fuscus Storer, Paralichthys dentatus (Lin-
naeus), Hippoglossoides platessoides (Fabricius), Apeltes quadracus
(Mitchill), Fundulus heteroclitus (Linnaeus) and F. majalis (Wal-
baum).

It is proposed that the family Allocreadiidae be restricted to include
only trematodes having cotylomicrocercous cercariae and simplified ex-
cretory patterns, since the type genus, Allocrcadium, would be included
in the restricted family. Consequently, the family name, Opecoelidae,
would be suppressed as a synonym of Allocreadiidae sensu stricto.

ADDENDUM AND CORRECTION

In respect to the uroproct, the present species is more like Opecoc-
loidcs than Anisoporus. Odhner (1928) erected the genus Opecoeloides
to contain a single species, Distomum furcatum Bremser, and described in
this form the union of the ceca with the excretory vesicle as has been
observed in the present study. In Anisoporus, as defined by Ozaki
(1928), instead of joining the excretory vesicle, the ceca unite posteriorly
to form a median tube with an anal opening independent of the excre-
tory pore. In other respects, the genera Opecoeloides and Anisoporus
are identical. While the present paper was in preparation, the writers
were inclined to share with Manter ( 1940) doubt as to the generic sig-
nificance of the uroproct as well as sucker papillae, characters used ex-
tensively in the separation of genera. The paper in which Odhner pro-
posed the genus Opecoeloides appeared Nov. 13, 1928, while Ozaki's
definition of Anisoporus was published Dec. 31 of the same year.
Whether or not these genera are regarded as synonymous, Opecoeloides
is the older name and the present species must be placed in that genus ;
accordingly, the correct designation is Opecoeloides manteri (Hunninen
and Cable, 1940). The species for which Odhner erected the genus has
never been mentioned in the literature except as Distomum furcatum; its
proper designation is Opecoeloides furcatum (Bremser in Rudolphi,
1819). The writers have been unable to find a satisfactory description
of this species, but from Odhner's paper it is possible to differentiate
0. furcatum and O. manteri, the only species in the genus, on the basis
of sucker papillae; there are six in O. furcatum and five in O. manteri,
three anterior and two posterior. Since the present study supports the
observations of Odhner and adds a second species of Opecoeloides, the
validity of the genus is strengthened considerably. It therefore seems

428 A. V. HUNNINEN AND R. M. CABLE

advisable to maintain Anisoporus and Opecoeloides as distinct genera,
particularly since proposing synonymy for them would cause a certain
amount of confusion.

LITERATURE CITED

ALLISON, L. N., 1940. Life history of Cercariacum constantiae Mueller (Trema-
toda: Brachylaemidae) from the snail, Campeloma. Jour. ParasitoL, 26
Supplement) : 38.
CABLE, R« M., 1938. Studies on larval trematodes from Kentucky with a summary

of known related species. Am. Midi. Nat., 19 : 440-464.

CABLE, R. M., 1939. Two new species of cotylomicrocercous cercariae from In-
diana. Trans. Am. Micros. Soc., 58: 62-66.

CABLE, R. M., AND A. V. HUNNINEN, 1940. Studies on the life history of Dero-
pristis inflata (Molin) (Trematoda: Acanthocolpidae). Jour. ParasitoL,
26 (Supplement) : 37.

DOBROVOLNY, C. G., 1939a. Life history of Plagioporus sinitsini Mueller and em-
bryology of new cotylocercous cercariae (Trematoda). Trans. Am.
Micros. Soc., 58 : 121-155.
DOBROVOLNY, C. G., 1939&. The life history of Plagioporus lepomis, a new trema-

tode from fishes. Jour. ParasitoL, 25 : 461-470.

HOPKINS, S. H., 1941. The excretory systems of Helicometra and Cymbephallus
(Trematoda) with remarks on their relationships. Trans. Am. Micros.
Soc., 60: 41-44.

HUNNINEN, A. V., AND R. M. CABLE, 1940. Studies-on the life history of Aniso-
porus manteri sp. nov. (Trematoda: Allocreadiidae). (Abstract.) Biol.
Bull, 79 : 373-374.

KOMIYA, Y., AND T. TAJIMI, 1940. Study on Clonorchis sinensis in the district
of Shanghai. 6. The life cycle of Exorchis oviformis with special refer-
ence of the similarity of its larval forms to that of Clonorchis sinensis.
Jour. Shanghai Sci. Inst., Sec. 4., 5 : 109-123.

LA RUE, G. R., 1938. Life history studies and their relation to problems in taxon-
omy of digenetic trematodes. Jour. ParasitoL, 24: 1-11.

MANTER, H. W., 1940. Digenetic trematodes of fishes from the Galapagos Islands

and the neighboring Pacific. Allan Hancock Pacific Expeditions, vol. 2.

MARTIN, W. E., 1938. Studies on trematodes of Woods Hole: the life cycle of

Lepocreadium setiferoides (Miller and Northup), Allocreadiidae, and the

description of Cercaria cumingiae n. sp. Biol. Bull., 75 : 463-474

MARTIN, W. E., 1939. Studies on the trematodes of Woods Hole. II. The life

cycle of Stephanostomum tenue (Linton). Biol. Bull., 77: 65-73.
MATHIAS, P., 1937. Cycle evolutif d'un trematode de la famille des Allocreadiidae
Stossich [Allocreadium angusticolle (Hausmann)]. Compt. Rend. Acad.
Sci. Paris, 205 : 626-628.
ODHNER, T., 1928. Weitere Trematoden mit Anus. Ark. Zool. Stockholm, 20:

6 pp.
OZAKI, Y., 1925. Preliminary notes on a trematode with anus. Jour. ParasitoL,

12 : 51-53.

OZAKI, Y., 1928. On some trematodes with anus. Ja[>. Jour. Zool., 2 : 5-33.
OZAKI, Y., 1929. Note on Coitocaecidae, a new trematode family. Annot. Zool.

Jap., 12 : 75-90.

ROTHSCHILD, M., 1938. The excretory system of Cercaria coronanda n. sp. to-
gether with notes on its life-history and the classification of cercariae of
the superfamily Opisthorchioidea Vogel 1934 (Trematoda). Novit. Zool.,
41 : 148-163.

STUNKARD, H. W., 1931. Further observations on the occurrence of anal open-
ings in digenetic trematodes. Zcitschr. Parasitcnk., 3 : 713-725.

THE BLOOD OF THE ATLANTIC SALMON
DURING MIGRATION

EARL BENDITT, PETER MORRISON AND LAURENCE IRVING

{From the Edivard Martin Biological Laboratory, Swarthmore College,

Swarthmore, Pennsylvania )

There are many alterations in the habits and metabolism of salmon
during their migration from the sea into the rivers. Among the meta-
bolic changes, one can be well defined in relation to an equally distinct
alteration of the environment of the fish. In fresh water the freezing
point depression of the blood of salmon is less than it is while they are
in salt water (Greene, 1904; Smith, 1932). As a consequence of the
reduced salinity of the blood, which is indicated by the reduction of the
freezing point depression, it might be expected that the condition of the
blood for the transport of oxygen would be affected ; for it is known
that the affinity of hemoglobin in solution for oxygen is diminished by
increasing concentrations of salt (Barcroft and Camis, 1909).

The affinity of the hemoglobin in the blood of several freshwater fish
for oxygen is greater than in the blood of some saltwater fish, as is
shown by the pressure of oxygen required for half saturation of the
blood of a few marine and freshwater species in Table I. The freezing
point depression of the blood of freshwater fish is usually less than in
saltwater fish, and the blood of fish migrating from salt into fresh water
undergoes dilution, as is shown by the examples in Table II. The ex-
amples quoted are too few to warrant more than the suggestion of the
effect of salinity upon the blood, and there are many influences beside
the salinity of the environment which will operate to differentiate the
blood of various species.

There is an advantage in examining the blood of individuals of the
same species in two environments separated by so short an interval as
that which comes between the salmon in the brackish estuary and in the
lower fresh water reaches of a stream. We have found it possible to
distinguish the conditions for oxygen combination with the blood of
Atlantic salmon, Salnw salar, caught in salt water from those of the
fish which were caught in the rivers. The changes in oxygen affinity
occurred as the freezing point was changed by the passage of the fish
into fresh water. The changes observed are large enough to suit the
blood for respiratory transport under somewhat different conditions,
and may be critical in determining respiratory ability in certain natural
situations.

429

430

BENDITT, MORRISON AND IRVING

MATERIALS AND METHODS

The blood of the Atlantic salmon, Salino salar, was chosen because
of its accessibility. These fish spawn and are hatched in the rivers which
drain into the coastal waters of the Province of Quebec, Newfoundland,
Nova Scotia, and New Brunswick. They spend the first two to five
years of their lives in the rivers and then migrate to the ocean, where
they spend one, two, or three years before returning to the rivers to
spawn. The major spawning migrations occur in the spring and fall

TABLE I

A comparison of the tensions of oxygen required for half saturation of the blood of

some salt and freshwater fish

Fish

Tension at

half saturation

mm. Os

pC02
mm.

Tempera-
ture

Fresh water
Bowfin *

Amia calva
Common sucker *

Catostomus commersonnii
Carp*

Carpiodes cyprinus

Pikej

Esox lucius

4
12
5
3.5

0-1
0-1
1-2

7.5

15
15
15
18

Salt water

Codt

7.5

Gadus callarias

Sea robin J

Prionotus carollnus

Mackerel %

16-17

Scomber scombrus

Toad fish %

Opsanus tau

* Black.

t Krogh and Leitch, 1919.

JRoot, 1931.

of the year, but fish in greater number and of larger size come in the
spring (at least in the St. Lawrence region (Belding and Prefontaine,
1938)). Our samples of fish were obtained in the late spring and early
summer from the gill nets of commercial fishermen situated around the
mouth of the York River, which empties into Gaspe Bay, Province of
Quebec. These fish were in brackish water and, unfortunately, no true
saltwater fish were obtained. Freshwater fish were had from the lines
of the sport fishermen on the St. Jean River, which also empties into
the Bay. Two of the freshwater fish were kindly supplied to us from
the salmon of the Gaspe Hatchery.

BLOOD OF MIGRATING SALMON 431

The fish were all two- and three-year " sea life " salmon averaging
roughly 10 and 20 pounds in weight respectively. They were bled by
heart puncture, 20 to 80 cc. being obtained from a single fish. Heparin
was used as the anticoagulant throughout, and the blood was stored on
ice from the time of drawing until it was used.

The blood was equilibrated with gas mixtures in the special tonom-
eters designed by Irving and Black (1937). Gas analyses on blood
were done by Van Slyke's manometric method. The temperature of
equilibration was 15° C. ± 1.0°.

Freezing point determinations were made with a micro-Beckmann
thermometer on 2 cc. samples of plasma. Despite the small samples,
most duplicate determinations agreed to within .005° C.

Hematocrit determinations of relative erythrocyte volume were made
on all specimens in the usual capillary tubes in a centrifuge operating at
5000 r.p.m.

TABLE II

A comparison of the freezing point depressions of the blood of fresh, salt water and
migrator}- fish. Modified from Smith (1930)

Fish

Medium

Freezing Point
°C.
Fish plasma

Amia calva

.03

.54

Lepidosteus osteus

.03

.57

Angnilla rostra ta

.08

.63

H it

1.85

.82

Conger vidgaris

2.15

1.03

Salmo salar

.64

n «

.87

.77

OXYGEN CAPACITY

The oxygen capacities (Table III) of the bloods of 15 brackish water
fish varied between 10.5 and 14.9 volumes per cent, with an average
capacity of 12.3 volumes per cent. The proportion of erythrocytes in
the blood of these fish varied from a minimum of 24.4 per cent cells to
a maximum of 47.5 per cent cells, averaging 39.4 per cent. In 6 fresh-
water fish the oxygen capacities varied between 6.7 and 10.0 volumes
per cent with an average of 8.8; the proportion of erythrocytes ranged
from 19.3 to 28.4 per cent, averaging 24.8. If we calculate the oxygen
capacity of 100 cc. of cells for the two kinds of fish, we find that for the
brackish water fish the average value is 31.6 cc., whereas for the fresh-
water group it is 34.9 cc. It seems from these figures that the propor-
tion of cells in the blood of freshwater fish was decreased without dimi-
nution of the oxygen capacity of the cells.

432 BENDITT, MORRISON AND IRVING

OXYGEN DISSOCIATION CURVES

Points on the oxygen dissociation curve of the brackish water fish
show a considerable scatter. The points are taken from 14 different fish.
It can be seen from Fig. 1 that at half saturation the spread is from
about 17.5 to 28.8 mm. of O2 tension. The mean curve has a half
saturation at 23 mm. O2 tension. The curve is similar to those found
for other marine fish by Root (1931). The curve shows an interesting
tendency to have an " S " shape similar to that of mammalian bloods,
and is in most other respects similar to these well-known curves.

In contrast to the brackish water fish, it is to be noted that with the
exception of one point the points obtained from 4 freshwater fish fall
nicely on a continuous curve. This curve is rather steeper than the

TABLE III

A comparison of the data on brackish and fresh water salmon

Brackish water fish Fresh watei fish

Cell volume 39.4% 24.8%

Variation 24.4-47.5 (15 fish) 19.3-28.4 (6 fish)

Oxygen capacity 12.3 vols. % 8.8 vols. %

Variation 10.5-14.9 (15 fish) 6.7-10.1 (6 fish)

Maximum CO2 effect 62.8% Sat. 57.2% Sat.

Variation 56-67 (5 fish) 56-58 (4 fish)

Cell volume increase 9.0% 8.2%

Variation 8.1-13.9 (5 fish) 6.2-11.1 (4 fish)

Depression of freezing point 0.77° C. 0.64° C.

Variation 0.72-0.80 (5 fish) 0.60-0.68 (7 fish)

others and lies to the extreme left of them with a half saturation at about
19 mm. of O2 tension (Fig. 1).

THE EFFECT OF CO2 UPON OXYGENATION

Carbon dioxide prevents the saturation of salmon blood with oxygen
at 150 mm. oxygenation (Fig. 2). This has been observed in the blood
of other marine and freshwater fish (Root, 1931 ; Black and Irving,
1938; Irving, Black and Safford, 1941).

There is a considerable spread in the points obtained from the brack-
ish water fish, the maximum effect of CO2 restricting the oxygenation
of hemoglobin to between 56 and 67 per cent saturation. In the fresh-
water fish CO2 restricted oxygenation to about 58 per cent saturation
for the maximum effect. Neither curve for salmon blood seems to
flatten out quite as quickly as the curve of carp blood (Black and Irving,
1938), and even beyond pressures of 80 mm. of CO2 there appears to

BLOOD OF MIGRATING SALMON

433

be some further depression of oxygenation. Hemolysis does not abolish
the effect of CO., either in the brackish or freshwater fish, and causes a
decrease of not over 15 per cent in its magnitude. Hemolysis of trout
blood likewise does not much reduce the effect of CO., upon oxygenation

1OO r

0 BRACKISH WATER FISH
• FRESH WATER FISH

PO2 mm

FIG. 1. The O2 dissociation curve of the blood of salmon taken from brackish
and fresh water. P CO;, lessv than 1 mm.

(Irving, Black and Safford, 1941). There are evidently two categories
of fish blood, in one of which hemolysis abolishes the CO2 effect (Black
and Irving, 1938; Root and Irving, 1940), while in the other hemolysis
has little influence. Hemolysis by either freezing and thawing or by
the use of saponin produces the same result.

434

BENDITT, MORRISON AND IRVING

The effect of CO2 on the oxygen dissociation curve of brackish water
fish is shown in Fig. 3. The pressure at half saturation of the hemo-
globin is 40 mm. of oxygen when the CO., tension is 13-14 mm. com-
pared with the average of 23 mm. O2 tension with the CO2 tension of
1 mm. or less. The effect of CO2 upon the oxygen dissociation curve
of freshwater salmon was not determined, but from the maximum effect
of CO2 on the oxygen capacity in fresh and saltwater fish it seems likely

100 p

:8O

c
o

l£

FRESH WATER FISH

o Whole blood

• Hemolysed blood

3O 4O 5O 6O 7O SO 9O

.0,

100 n;

BRACKISH WATER FISH

'O 2O 5O <JO SO 6O 7O BO 90

pCO,
FIG. 2. The effect of CO; upon oxygen dissociation curves of salmon blood.

that the effect of CCX on oxygen dissociation curves would be similar
in salmon from salt and fresh water.

EFFECT OF CO2 ON CELL VOLUME

Carbon dioxide has a marked effect in causing the cells of salmon
blood to swell (Fig. 4), as was shown for the blood of some freshwater
fish by Black and Irving (1938), and for the blood of trout (Irving,

BLOOD OF MIGRATING SALMON

435

Black and Safforcl, 1941). The magnitude of the effect on the cell
volume in brackish water fish averaged 9.0 volumes per cent of cells, with
a variation of 5.1 to 13.9 volumes per cent in 5 fish. In the freshwater
fish the swelling averaged 8.2 with a variation of 6.2 to 11.1 in 4 fish.
The reversal of the effect on complete oxygenation of the blood without
CO2 is demonstrable.

Most of the swelling is produced at low tensions of CCX, and it indi-
cates some 25 per cent enlargement of the erythrocytes. The consid-
erable swelling of the cells is of a much greater magnitude than the
osmotic changes which are described in mammalian blood, and there
seems to be no satisfactory explanation for this phenomenon.

100 r

SO -

— BRACKISH WATER SALMON COZ pressure 1 mm or less
FRESH WATER SALMON COj pressure I mm or less

-°- BRACKSH WATER SALMON CO, pressure 13to|4 mm

FIG. 3. The maximum effect of CO2 upon the degree of oxygen saturation at
P O2 == 150 mm.

THE COMBINATION OF CO., WITH BLOOD

In Figure 5 is shown the curve describing the combination of CO.,
with the blood. This curve indicates that in contrast to the reduction
in oxygen capacity which occurs in fresh water, there is no significant
difference in CO2 combination.

FREEZING POINT OF THE PLASMA

The effect of the change of external environment upon the blood
plasma is demonstrated in the change in the freezing point depression.
In the plasma of brackish wrater fish freezing points varied from -- 0.717

436

BENDITT, MORRISON AND IRVING

t0 --0.800° C, with an average value of - -0.765° C. for the 5 bloods
examined. In the freshwater fish 7 determinations fell between -- 0.597
and — 0.675° C., with an average of -0.638° C. One other fish

ioo r

I.- Ot CAPACITY

*HEMOLYSED

WHOLE BLOOD

2- CO2 CARRYING POWEP

q 1rt

- /

i i i i i i i i i

E
0)
X 4O

3. -CELL VOLUME

FISH *2.7

_l L__

2.0

mm CQ Pressure

FIG. 4. The effect of CO, upon: (1) oxygen saturation, (2) CO2-carrying
power, and (3) cell volume in the blood of one salmon.

showed a value of - -0.775° C. This fish was caught much lower down
in the river than any of the others, although still in the fresh water. It
was caught at the very end of the season when it is known that the

BLOOD OF MIGRATING SALMON

437

late stragglers spend very little time in becoming acclimated to the
brackish water, and this observation may explain the individual dis-
crepancy.

DISCUSSION

We would like primarily to answer the question in this paper of
whether or not there is a difference in the properties of the blood of a
migratory fish when it is in equilibrium with a marine and freshwater
environment, and the changes which we are particularly interested in are
those of the system for respiratory transport.

There is undoubtedly a drop in the concentration of cells of the
salmon in fresh water, and concomitantly there is a drop in the oxygen
capacity of the blood (Table III). There appears to be no difference
in the hemoglobin concentration of the cells of the two kinds of fish.

SO r

• FRESH WATER FISH
o BRACKISH WATER FISH

pressure mm Ha
FIG. 5. The combination of COo with salmon blood at P O2 = 150 mm.

Dilution of the serum also takes place when the fish moves into fresh
water, as shown by the decrease in the freezing point depression. If
we examine the oxygen dissociation curves of the fish in its different
habitats, there appears evidence that here too there is a change with mi-
gration. The curve for the freshwater salmon lies on the outermost
border of the scatter of points belonging to the brackish water inhab-
itants (Fig. 1), and the points hold well to a smooth curve.

Considering the brackish water fish in relation to their environment,
we find : ( 1 ) that the environment varied somewhat in its osmotic con-
centration with respect to the tide and the location in the bay in which
they were caught; (2) that the fish had been in this environment of
lower salinity than the sea x for various lengths of time and were conse-
quently in various stages of acclimitization. In view of this we might

1 The freezing point depression of a sample of water from the mouth of the
York River, at which point many of the fish were taken, was 0.87.

438 BENDITT, MORRISON AND IRVING

expect to find the variation which was actually observed, and also to
predict that the blood of true marine fish should give points lying at the
extreme right of the group. We are inclined then to believe that in this
important characteristic, the affinity for oxygen, the blood of salmon
living in fresh water differs from those taken from salt water.

The provision for transporting oxygen in fresh water is less since the
oxygen capacity of blood is only about two-thirds of what it is in the
brackish water, and probably less than that in relation to the oxygen
capacity of true marine fish. One is inclined to wonder if the fresh-
water environment imposes a more sedentary mode of life because of
the changes which are brought about in the internal conditions of the
salmon.

In affinity for oxygen the blood of fresh water salmon equals that
of three freshwater salmonoid fish, — the brook trout, brown trout, and
rainbow trout (Irving, Black, and Safford. 1941). In all of these fish
atmospheric tensions of oxygen saturate the hemoglobin at 15°. But
as the temperature is raised, the affinity of trout blood was found to
diminish, until at about 25° atmospheric pressures of oxygen could no
longer secure saturation. It is quite likely that a similar temperature
effect prevails in salmon blood, which should therefore be suitable for
oxygenation at 25° in thoroughly aerated water. The blood of salt-
water salmon would probably fail to saturate at about 20°, but it is
likely that the well-circulated tidal waters are always well enough aerated
for adequate oxygenation. In the warm water of the rivers in summer
even the blood of the freshwater salmon encounters temperatures which
are near the limit permitting saturation. If the oxygen tension in the
warm water is depleted below atmospheric pressure, then the blood can-
not become saturated with oxygen in the gills. In the rivers it is likely
that stretches of warm water, partcularly if they are not well aerated,
act as barriers by hindering the transport of oxygen. Under such condi-
tions the effect of temperature upon oxygen affinity may have a critical
influence in determining where the fish can exist.

We see that the plasma of freshwater fish has a lower freezing point
than the brackish water fish, and that this is due to a reduction in the
electrolyte concentration has been shown by Homer Smith (1930). In
our freshwater salmon, the serum has been diluted and we may surmise
that in order to reestablish osmotic equilibrium, water has diffused into
and perhaps salts out of the cells. That the latter is so would seem to be
borne out by the values of oxygen carried by 100 cc. of cells. Long ago
Barcroft and Camis (1909) showed that when hemoglobin solutions are
dialyzed, the oxygen dissociation curve for the solution, when compared
with the curve for the undialyzed solution, is shifted some to the left, and

BLOOD OF A1IGRATING SALMON 439

hence saturates at a lower pressure. It should be pointed out that they
were working with rather dilute solutions of hemoglobin and also were
dialyzing off the last portion of electrolyte and therefore the conditions
in the salmon blood are hardly comparable.

It is rather remarkable that the CCX dissociation curves for the
freshwater and saltwater fish (Fig. 5) should be the same in spite of
the difference in oxygen capacity. There are large differences in CO2
capacity of the blood of different species of fish which appear to be
quite unrelated to oxygen capacity.

The changes which have been shown in the blood of migrating salmon
are large enough to be important to the economy of respiratory metabo-
lism. The relation of these changes in the blood to the change in en-
vironment suggests how the detailed physiology of the salmon changes
with the varying environment.

SUMMARY

The blood of Atlantic salmon caught in the brackish water of Gaspe
Bay has been compared with the blood of salmon caught in the fresh
water of the rivers draining into the Bay. In brackish and fresh water
the average properties of the blood are respectively : oxygen capacity,
12.3 and 8.8 volumes per cent ; cell volume, 39.4 and 24.8 per cent ; oxy-
gen tension for half saturation at T CCX = 1 mm., 23 and 19 mm.;
freezing point of the serum, - -0.79 and - -0.64. The oxygen com-
bination at P O.2= 150 mm. in the presence of large tensions of CCX
is reduced to about 60 per cent of saturation. Hemolysis does not much
reduce the CO2 effect. The cells swell greatly as the CCX tension is
increased. There appears to be a dilution of the blood as the fish goes
from salt to fresh water. This is seen in the decrease in cell volume,
oxygen capacity, and freezing point depression of the blood. It seems
also that in fresh water the affinity of the hemoglobin for oxygen is
greater than in salt water. The changes observed in the blood may be
related to the change in salinity of the environment. In the warm water
of rivers in summer small changes in temperature and oxygen saturation
may be critical in determining whether or not the blood can be saturated
with oxygen.

ACKNOWLEDGMENT

We gratefully acknowledge the assistance of the Bureau of Mines and Fish-
eries, Province of Quebec, extended in particular through Mr. Charles Lindsay,
and the hospitable provision of laboratory space by the authorities of Le Seminaire,
Gaspe, Quebec.

Part of the expenses were defrayed by a grant from the American Philo-
sophical Society.

440 BENDITT, MORRISON AND IRVING

BIBLIOGRAPHY

BARCROFT, J., AND M. CAMIS, 1909. The dissociation curve of blood. Jour.

Physiol., 39: 118-142.
BELDING, D. L., AND G. PREFONTAINE, 1938. Studies on the Atlantic salmon. II.

Contrib. dc I'lnst. dc Zool. de I'Univ. de Montreal, No. 3.
BLACK, E. C. Unpublished observations.
BLACK, E. C., AND L. IRVING, 1938. The effect of hemolysis upon the affinity of

fish blood for oxygen. Jour. Cell, and Coinp. Physiol., 12 : 255-262.
GREENE, C. W., 1904. Physiological studies of the Chinook salmon. Bull. U. S.

Bur. Fish., 24 : 429-456.

IRVING, L., AND E. C. BLACK, 1937. A convenient type of tonometer for the equili-
bration of blood. Jour. Biol. Chem., 118: 337-340.
IRVING, B., E. C. BLACK, AND V. SAFFORD, 1941. The influence of temperature

upon the combination of oxygen with the blood of trout. Biol. Bull.,

80: 1.
KROGH, A., AND I. LEITCH, 1919. The respiratory function of the blood in fishes.

Jour. Physiol., 52 : 288-300.
ROOT, R. W., 1931. The respiratory function of the blood of marine fishes. Biol.

Bull., 61 : 427-456.

ROOT, R. W., AND L. IRVING, 1940. The influence of oxygenation upon the trans-
port of CO2 by the blood of the marine fish, Tautoga onitis. Jour. Cell.

and Comp. Physiol, 16 : 85-96.
SMITH, H. W., 1930. The absorption and excretion of water and salts by marine

teleosts. Am. Jour. Physiol., 93: 480-505.
SMITH, H. W., 1932. Water regulation and its evolution in the fishes. Quart.

Rev. Biol, 7 : 1-26.

COMPARATIVE STUDIES OF THE PIGMENTS OF SOME
PACIFIC COAST ECHINODERMS *

DENIS L. FOX AND BRADLEY T. SCHEER
(From the Scripps Institution of Oceanography, La Jolla, California)

From the standpoint of comparative biochemistry, the echinoderms
represent an interesting and little-explored phylum. Prominent among
the biochemical features of this group is the conspicuous manifestation
of body-pigments, striking in their intensity and color-varieties, and ri-
valling the beauty of those displayed by the sessile coelenterates.

The integumentary colors of sea-stars and brittle stars are due pre-
ponderantly to carotenoids, while certain of the echinoids manifest in-
stead considerable quantities of pigments of the echinochrome class, first
reported by MacMunn (1883a) and recently shown by Kuhn and Wal-
lenfels (1939) to be naphthoquinones. Kuroda and Ohshima (1940)
have crystallized three distinct spinochromes, each from a different spe-
cies of Japanese sea-urchin, and have found the natural pigments to be
very similar to synthetic hydroxynaphthoquinones. Some echinoids
yield echinenone, a unique carotenoid which is a provitamin A (Lederer
and Moore, 1936). Long ago, MacMunn (1883&, 1886) reported the
presence of " enterochlorophyll " in the alimentary organs of a number
of carnivorous echinoderms.

Studies of Abeloos (1926) and Lonnberg (1931, 1932, 1933) give
qualitative indications of the nature and distribution of carotenoids in
echinoderms. Euler and Hellstrom (1934) and Euler, Hellstrom and
Klussman (1934) made chemical studies of the carotenoid proteins of
asteroids, and isolated a new pigment, asteric acid. Karrer and Benz
(1934) and Karrer and Solmssen (1935) isolated astacene from both
an ophiuran and an asteroid. Lederer (1938) has studied the pigments
of the echinoid Strongylocentrotus livid us.

Numerous writers emphasize the importance of the question as to
whether some of the lower animals may be able to synthesize specific
carotenoids from simpler molecules. Among the various ecological fac-
tors and physiological activities which may influence the pigmentation
of animals, it is probable that food exerts the closest and most direct
effect, although habitat and various physiological adaptations inseparably

1 Contribution from the Scripps Institution of Oceanography, New Series No.
132.

441

442 D. L. FOX AND B. T. SCHEER

associated with nutrition are necessarily important in a survey such as
we have undertaken.

Of the rather scanty information available regarding the physiology
of adult echinoderms, facts concerning their feeding mechanisms and
digestive enzymes are probably the chief entries. Yonge (1928) dis-
cusses the diversity of feeding mechanisms employed by the Echinoder-
mata, and divides the phylum into two main groups as regards feeding
and digestion, namely : ( 1 ) the Asteroidea and Ophiuroidea which are
exclusively carnivorous, and (2) the Echinoidea and Holothuroidea,
chiefly herbivorous and to some extent omnivorous. He suggests
(1931) that the ciliary-feeding Crinoidea may be primarily herbivorous.

In the present study, we have attempted to make a preliminary classi-
fication of the various echinoderms into biochemical (pigment) types,
with correlative differences in intraphylar class, sex (where practicable),
habitat and nutritional habits occupying collateral positions of importance.

The survey, which included four species of three genera of echinoids,
one species of holothuroid, four species of three genera of asteroids, and
three species of two genera of ophiuroids, revealed numerous familiar
and a few new carotenoids. Purple echinochrome pigments were ob-
served only in echinoids. Small amounts of green " enterochlorophylls "
were found among the asteroids and ophiurans. We have considered
the echinoderm classes in ascending evolutionary series within the two
chief nutritional groups named by Yonge (1931) i.e. (1) herbivore-
omnivores: (a) echinoids, (b) holothuroids ; and (2) carnivores: (a)
asteroids, (b) ophiurans. Consistencies in this classification will be ap-
parent in Table I.

The animals were collected at several localities in this vicinity of the
Southern California coast, and in some cases obtained by dredge hauls.
We are indebted to Mr. P. S. Barnhart, curator, Dr. C. E. Moritz, visitor
from Dartmouth College, Professor G. E. MacGinitie of the California
Institute of Technology, Mr. Granville Ashcraft of the Hancock Foun-
dation, University of Southern California, Professor H. J. van Cleave,
visitor from the University of Illinois, and Mr. L. D. Pratt of the Kelco
Company, National City, for much of the material used. Mr. Sheldon
C. Crane rendered technical assistance during the latter part of our
carotenoid investigations.

METHODS

With but few exceptions noted below, the animals were placed
temporarily in running sea water in laboratory aquaria, to permit foreign
material to be evacuated from the intestine. The carotenoid pigments
were extracted from the ground tissues with acetone, passed into petro-

PIGMENTS OF PACIFIC COAST ECHINODERMS 443

leum ether, subjected to partition between the latter and 90 per cent
methanol, and separated into individual components by chromatographic
adsorption, in accordance with regular procedures described in more
detail by Scheer (1940) and by Fox and Pantin (1941).

Individual pigments were provisionally identified by comparison of
certain of their properties with those of known pigments ; this identifi-
cation must wait upon further studies for full confirmation. The fol-
lowing properties were used: (1) Behavior in partition between im-
miscible solvents, i.e., petroleum ether (which constitutes the epiphase
and dissolves carotenes, xanthophyll esters and a few monohydroxy-
xanthophylls) and 90 per cent methanol (the hypophase which selectively
removes xanthophylls). (2) Adsorption behavior on Tswett chromato-
graphic columns, xanthrophylls and their common esters being adsorbed
from benzene solution on calcium carbonate, while carotenes pass
through this but are selectively adsorbed on calcium hydroxide (Zech-
meister and Cholnoky, 1937; Strain, 1938). (3) Positions of spectral
absorption maxima. (4) Behavior toward the partition test after treat-
ment with hot alcoholic potassium hydroxide in an inert atmosphere ;
carotenes remain unchanged; xanthophylls, epiphasic when esterified,
are rendered hypophasic when hydrolyzed ; astaxanthin, free or esterified,
neutral before treatment, is transformed into astacene, with definite
acidic properties.

Positions of absorption maxima were determined with a Hartridge
reversion spectroscope and with a Bausch and Lomb spectrophotometer.
The two instruments show good agreement excepting in the case of acidic
carotenoids like astacene, whose single broad maximum is more accu-
rately determined with the latter instrument. Carbon disulphide was
employed as the solvent unless otherwise specified.

Relative concentration of mixed carotenoid pigments (i.e., whole epi-
phasic or whole hypophasic fractions) were estimated in terms of " (3-
carotene equivalents," the extinction coefficient of /3-carotene at 485 mju,
being determined by other writers (Smith, 1936) and ourselves.

Echinochrome pigments were readily extractable from the echinoids
by treatment of the whole tests with dilute hydrochloric acid under a
layer of diethyl ether which readily dissolved the pigments (see Tyler,
1939). These were subsequently examined spectroscopically in ether
or chloroform solutions.

Green pigments (" enterochlorophylls ") were recovered in small
amounts from the digestive diverticula of two of the asteroids, Pisaster
ochraceous and P. gigantens, and from the whole-body extracts of all
three ophiurans. The material recovered from the ophiurans differed
in certain solubility properties from that yielded by the asteroids, but all

444 D. L. FOX AND B. T. SCHEER

showed closely agreeing absorption spectra. Actual quantities were so
small that no identifying tests were practicable excepting those em-
ployed ; hence the descriptive term coined by MacMunn has been used
provisionally in the discussion of these green pigments.

RESULTS

A summary of the distribution of carotenoids, echinochromes and
enterochlorophylls is given in Table I, followed by more detailed infor-
mation regarding the separate pigments.

Carotenoids

The distribution, kinds, and some relative quantities of individual
carotenoids encountered in the various species are summarized in Table
II.

Echinochromes

These pigments, present exclusively in the echinoids, showed a few
interesting variations in tissue distribution as well as in chemical and
spectroscopic properties, as shown in the following outline:

Dendraster excentricus: Purple aggregates of the pigment were pres-
ent in ectodermal and endodermal tissues lining the shell of this purple
sand-dollar, while similar bodies in mature male and female gonad
tissues and in anterior portions of the gut were red. The posterior part
of the gut lacked echinochrome. The gelatinous egg-cases contained red
echinochrome bodies.

The pigment was purple in neutral or alkaline media and red in acid
(see also Lederer, 1940). It was readily soluble in aqueous acetone,
giving clear filterable solutions, and was also extractable from dilute
acid digests of the shell with diethyl ether (see Tyler, 1939).

The absorption maxima were as follows :

In acetone : 524, 490 m/x.

In chloroform: 533, 496, 465 m/x (cf. Kuhn and Wallenfels, 1939).

Strongyloccntrotus franciscanus: Much purple pigment was yielded
to acetone by treatment of this large purple-red urchin. The pigment
decomposed with bleaching, however, before it was given any study.

Strongylocentrotus purpuratus: This purple urchin, like the sand-
dollar Dendraster, showed many aggregates of echinochrome in the ecto-
derm (purple) as well as in endoderm, coelomic fluid and gut wall (red
in all). Neither male nor female gonad tissues, however, contained any
of the pigment.

It showed the following absorption maxima :

PIGMENTS OF PACIFIC COAST ECHINODERMS 445

In water (neutral, colloidal) : 526.5, 591 m/x.

In water (acidic) : 497 HI/A (single, diffuse band).

In chloroform: 525, 490.5m/* (cf. Lederer and Glaser, 1938).

Lyt echinus pictus: This pale urchin yielded far less of the purple
pigment than did Strongylocentrotus or even Dendrastcr. Pigment ag-
gregates were observable, however, in parts of the skin lining the shell,
in the gut and in gonad tissues of both sexes.

The pigment, insoluble in acetone or water, readily bleached by
alkali, and slightly soluble in dilute acid, gave the following spectral ab-
sorption bands :

In water (colloidal suspension) : 526, 491 m/x.

In diethyl ether : 533, 497, 467 m/*.

The echinoids contained not only different quantities of echinochrome
pigments, in the order Strongylocentrotus > D end-raster > Lyt echinus,
but the pigments were somewhat different chemically. Lederer (1940)
lists the distribution of a number of pigments of the echinochrome class
in several echinoid species, and in their separate body parts.

Green Pigments

The " enterochlorophylls," found only in the digestive diverticula of
two of the five species of asteroids and in whole extracts of the three
ophiuroids, showed certain chemical differences.

Asteroidea

Pisaster ochraceous: A small quantity of the pigment was recovered
in acetone from the digestive diverticulum, being absent from stomach,
skin or other parts. Completely insoluble in petroleum ether, it was
readily soluble in alcohol, acetone and diethyl ether. Its non-acidic char-
acter was demonstrated by its extractability from diluted alkaline alcohol
with ether. Dissolved in absolute ethanol, it manifested a single sharp
absorption band in the red at 661 m//,.

Pisaster gigantcus: The green pigment in this species, recovered in
traces from the same tissues as in P. ochraceous, was similar in its chemi-
cal properties and also yielded a single absorption band in pure ethanol
at 661 m/x.

Astropectcn calif ornicus and Patiria miniata failed to yield green
pigments.

Ophiuroidea

Ophiopteris papillosa: Extracted with acetone and soluble also in
diethyl ether, the pigment was dissolved readily in petroleum ether (un-

w
J

CQ
<

•
a

•*•*

o , -

*- A "

QJ u Vi

1 1 1

op,

o <y

£ O

-)-

-j-

+ + 1

fi!

1 * ^

i i jr

1—

o c -*-1

'522
111

OJ
0)

4->

cT c S "S • ^ ^

-C jj tjj_ W

e j- -o g <u S S

« (A rt S g £ gj

N 8^ w a

.0 , .g.i: o

"il?l

•

I . : :

" ° a

'So

e">> £

R r* -*J

rt-5 co

A A

r -\ ( >

O M

J33

+ 1 flj fli
1 ' 1 4-4--I-04--4-U -1-

+ 1 i 1 | \ \ 1+t \ \ —t
\
U, iZ
•4-1 -4-)

tf>
<u
C
IV

•.

t/1 C/3
1 1 1 1 1 1 1 ^j [ 1 ^ #

++ ++ + ++2++S +

i ' j i

•SU

r—*~^

^-A-^ r- *— t

v- nj <»

.2-c rt
a a -c

CTf O O.

CN ID

P^ ID -t< OO O 00 Os
* ^^1 * ^^ ^^ ^^ ^O ^1 '**5 10 ^^ OQ ^O \^^
* t^i * ^H t~^ ^^ ^\ r^ r^ T^H T— H ^-^ t*** ^^

^'I "

o'd

r-iO tNO OOOO-nOO'" O— '

•l-S.-So

— i 00

„, u, -HO-DCNOOr^

_^ rt o •-* o

*-H r/^

Jj** o"^ IOIDOOOO — csoi-^ -^ CN

^"n. & c^l

o o

t^O 03^ ID"*-~HCV]QLOIOO OO

£a°-t

do"

isooiioo oo-*oo-oo oo

•- bb

C* o i *T§Q

fN to

^ $O<D-H01Os (N IDU->

^ '55 O •— o

to o

_2 j^ rt C ---.

o —

*-+

rtOsO^iDsO -POO-FOOO'DTfO OO

oa £bj

d d

^cNfsHtNO ddirt'dd'D'*d do'

^ «N -O

^O t^1* O\ ^O OJ

_£i

lO^EJ ^* ®1 ^H ^vo '"'ro t*» ^ rt

*"^ ^5

• - ^". • - »~i ro t*^« • »^» /-*j • - • f~ C^l OO *""

PO <^1

t-s)

• (U OO (U »~H ^— i *O QJ ^^ ^* Q) ^^ ^* ^^ ^O rj

. C .. ! C .. C _r^ c ..t^- 03

QJ o

_aT

• | i ^£2 " !._' 'U ^ ^J -^J ^3 ^^ 4 ' ""^ ^^ 1J QJ ^

"o "o

J3 -C

.SSc.S^c x^^S'S'Jc'S 11!!^

t/) .i^ bJO c/3 . 13 O/) ^ I> . Z- b/3 -4—* -b- OJO •*—• !? JS

mal (number analyzed)

VORES;OMNIVO
INOIDEA

endraster excentricu
(3?)
(2ef)

'rongylocentrotusfra
canus
(Icf)

CO f*5 '«

s U ^

S -_g---^ — i-, IS"

•S -"~* "^"o t ""^ "V> O S3

O 0 "0 ^vO(D *-"" -H 2 ^-^-,

"?•» O vs

^ ffi Q

CO ^ i-J C^

CQ (J

i w

O i -

«3 P"w

S.2^

1 1

1 + 1 + 1 +

+ + +

-w-s -a

IE
s P

1 1

1 1 1

o p

H-g

0 £ £

S 4) . 0)

I -

15 c d

•S-5 .S g

O. • C J£ s '/>

rt"o S

l"5 .£•£

C-1-

* r- C • r- ^

c "5 -5 c — T;

.5 -c .S •-' .S ^'

'§|,«

•tJ 03

^ c x 15 c'£

j= ^ -n a j= a

e « c ° = °

c x ex

CJ 03 <y 03

a. N

zeaxan
astax

c x.-= "c x £

03 03 -^ 03 03 C

g 03 S g « 0

N N

.3 X 3 J« S —

03 /-, 03 4-> 03 4->

X X c X c

2 £ 23 23

03 *• rt ^ aj ^

^ £•§!! ^

r- -, -

^ o

" ° Q,

c/) (/)

"5 *~* u

0) <U

*v* ^* ^
rS

1 1

1 1 1 1 - -

1 1 1 1 rt rt

u —

4-1 4-t

+ + +

c ji1

t !

HH tJ

t t i

c
o

++++ ++

1 1 1

_o
'<« .«

A A A

v- cd w

.2-c KJ
aS o a

r<o to

00 —
CN rrj GO OO -f -*

ON P* Q, P* D. P* O.

a; Q '5.

-H ,-H

•"•

»— i »— l *— i »-(

*-O ^j~| ^ ^ r^ r*^

•HUsg

oo ^

-+i -t <^1
O 00 — 0 t-

a. d. d,
o S S E

"^ ,C J-" O vM

l^1* ^^

f*^l ^^v f^J ^^ ^^v ^O

Os O O O

o o ° ^ ti
£_ a g

d d

d CN d CN d -H'

• u o u

*^ QJ <D <D

-a -a -a

.— bit

a.*.^o
— m £'3 °

«* OO
•* CN
0 0

rq t^ -f vo t-~ -H

fN SO CN "O SO O

O sO O -t< O -*1

oo 10 o o

o"a"2"bi

d d

0 O 0 O 0 O

do — ' — '

H E

^H
^T* _^_ _j> *

-^ "^ OO "^ so —i

P o

CTO ^H Cu ^ frt
_ u .5 u .So

' ^ ^ ^ i' t> Q^

-H ^

.£P

0 r-

^ tJ W U ^ ^

°°. C 00 °°.

!>• i — i

. — i

+ u+ 0 +0

Os «-i -H OO

aT aT

en ..H tfj « •— " c/j««

O CJ O QJ

0 0

So o

"3 "3 "o "o

J3 J=
S !S

-C

u "T^ t- "^ u "^
Csl O.fN O. CN O.

jsjz -c j:
^ ^ is £

t
^

4->

3 C/)

^^ * — '

•o

*2 o *" — ^

^ — ' 4-1

txj Q

rt
1

«> S-s^

^ .Si rt SU

'«** H 4-* U 1- Ui

"3 -- a;^: ^3 r

< G 1'io0-^

W §r^ |>~*S

39^^ So>s^^

§ § ! |u

^ u^ ^ ^

4-" * — s

jy -H

rt GJ

« g>

Sfe

s s ^ -1 *
I'll § 1

IH 5 -L i

* U Q 2° °*— >— i

>^ O - — ' - — - s^ • — - f~s

^> ^— ^} *^
"^ , , t/3 H—

a sJ 13
."^ .">

" S? "*** ^i
*^< VJ

-S, •§.

"*^ •^^PO .^ !^5 .^ OO

J 13
^ -S -S
•<x, <x >x,

O O O

fe; en

(^ ^

o
c

•o

448

D. L. FOX AND B. T. SCHEER

w -

j <o

« ^

•< o

•&
f

H
s
co

't?

o<^3

o c

c p

ic carotenoids]

111

Q. 0 g

zeaxanthin

lutein
lutein (isomer

petaloxanthin
antheraxanthi
zeaxanthin

petaloxanthin
antheraxanthi
lutein

pectenoxanthi

antheraxanthi
astacene or
metridene
pectenoxanthi

astaxanthin^
astacene
zeaxanthin

•a

">>

.2 o>

Wj'S

00 CU

'p'
^o

r^ t^ t;

00 00

rs O r^.
OOl^t^

Sf^'S

in 1 in rj<

£"S c^

"^ i-

.-— O

Tt -^ -T

**" *-* 'o ^

offi o o

•5

8 a

W w

6~— "£

— 0000

(N — * O

—

ox ?6"

^O^""

& c

in O

in in o

m .. .-

in in in

in~^ m

O^ — oo

"*•

•|

^i2

^-*

in in

4-1

•

•a— -

•a a> "a

Sit!

E S'H^'-'S

'^S'

„

,_,

•5 g

"rjfi

5 M ° °.s §

1-1 s- 5

"""* H-t

1-1 '-i S

S~5
Hi

i— t

1-1 £H

Q.S i.*

o v

Pi
B

o — £ M

»P -»-» *p •*-• e

U o E o 3

1*3

•*-»

•<*

H 3 N

C*3

Tt*

^ -2

P g C

^_^

O > CU

ao g

•

fc 'S,

Si CU CU

ogg

to ^ CU CU

V p ^_( *->

o c c

p
o
p

cu
p

p22

^ P 2 1-1

cu o o

a >- (-

tfl
cu

"ri cC rt

•g v v

CUCQ.CQ.

*CU "^ CO CO

i; js o o

_"^ ^( CQ_ (jj

•S CO CO
J3 u O

11 03. S

^_^

p
o

ir,

^-^

Q nj

Tf '5

00-*-*

QQ r/- ^'

^J1 \O t~~

Tj<

1"- *

55o

00 CO 00

r^ 00 1-

00 001^

o •* •* •*

.- Ov .- ..

O TO

mco

^O 't oom

^ O O t^

C^J

.2 E

"•> "3

CN C4 CS

g ^o

in in >n

CO
u

1/5

"55

i .

"T3 ' — *

*o - — *

•U CU O
CT3 _«, g

S M M °-S °

«S>

M^s>>

HH1-IS>

.5 cd

•2-
E x

.S g

»-i Jr 2? v ^

0 <"

QO LJ *-< ^\

•"•; <U tn

I—

-C 4-' r* L*~l ci

vfi

H|°S

<*)

_to

endraster
excentrici

'rongylo-
centrotus
francisca

iSl
"?, g s

K" S ^

to
!S to

ichopus
parvimen

o-S
•i °

PIGMENTS OF PACIFIC COAST ECHINODERMS

449

— ^

a c

>, >,

ic carotenoids)

« CO

c c c

O, O C

GO i2 c
g5.»

mytiloxanthin

astaxanthin— »
astacene
zeaxanthin

03 cu *;
t- "O C

cu •- a

•S £ X

C r- ftj

rt c N

taraxanthin
a new xantho
pectenoxanthi

a a

O 0

.ex;

n c
o! rt
X X

cu cu
c c

t decomposed

•o

•°

^^"o -*-• .^

.11

X!
a
o

O oj

'*j £
a.5 a.

s^slss

m cj in 1 m •*

o|3 oK o o

Sill^f

c oo "O1^ o

•*in

o"rt
rt C
£ o!

o a ^

.SE

O-'C^ — 00

o "") ^ , £

i— o — 1_^ rt ^ "^

— f> t; ><M

CC gjO^

in 'n

o
-c

.y o

CO "*-

.S o js10^

'^^_CJ

nja

*~^

•^ . —

c !S

'a'a

in o

cu rt

C ^

•o cu a

E £'2_-^'<G

t_< 1— 1

,- "-•

"^ >

*co

C^ O O o ii 5

*"H I-H |~]

^ HH

""* H~l ^H •— '

cu bO *** S

•a

r? O C .fe

2 """

<rj

E u

O ^^ 4> 0]

^ rt O ^j

cu
O

f2f°l

CO CO

'co 'co ^j

'co CO

"3*3 2

O O co

•O'O —

•T3T3 CU

*O *j

•^

>. >i x

>• >>»

c g c

O S CU

a o s
gig

o
U

CU
•*-»

C (u

CO |

C C o
0 Ox

'o'u rt
rt ol X

c c —

0 O >>

'S'S'a

CJ O O

rt rt x

03
CJ

rt u

D i

S rt
-*-» i-
co rt
cu ~>

CU OJ CU

c c c

£ £ c

CU CU OJ
C C X

-t

\§ ^

—

Tj" _**?

O cfl

ff) *&

•<*

•• £

C
_O

CJ
OJ

?'§ 1

|IS

00 CC

ooT

..!J

r- in .-

o c cu

*o in o
•*•*•*

co-*^

17 CU

**-

'co

•d^

"c-a •§. M

1 § °° * °

.M cu

c ^i

£*P,

^ JP 5? K ^

0 ^

o Q- *•

tr *^

o •— ' q^ 05

5 gpl

"c3

*s
o

'S ^

•2 ~

.«

V «

*- ^

^"5

I*,

~_^

'S §,

§'s.

.= •-

.g-S

•§ °

.1 5>

"§,*•

•«_ S

450 D. L. FOX AND B. T. SCHEER

like those encountered in the asteroids) hut migrated quantitatively from
this solvent to 90 per cent methanol in the partition test. It was ad-
sorbed from petroleum ether solution at the top of a calcium carbonate
column and was eluted therefrom with difficulty by methanol and acetic
acid. Its solution in absolute ethanol gave a single band in the red at
663 m/A.

Ophiothrix spiculata: A yellow-green pigment remained in the diluted
and salt-containing acetone extract from which all carotenoids had been
transferred to petroleum ether. Insoluble in the latter solvent, even in
the presence of dilute acids, this pigment was readily extracted with
diethyl ether and transferred to absolute ethanol, in which it showed a
single absorption band at 663 m/x.

Ophiothrix rudis yielded small quantities of a green pigment similar
in solubility properties to that recovered from 0. spiculata.

MacMunn's green pigments extracted from the radial caeca of So-
laster and Urastcr displayed several spectral bands in alcohol solution,
one in the vicinity of a single band found in the red by ourselves, and
the others in additional regions similar to those shown for chlorophyll.
He considered that such green pigments were breakdown products of
chlorophyll (MacMunn, 1886).

We have at present no new suggestions to add to MacMunn's con-
clusions. While the spectra of our green pigments were far simpler
than those of MacMunn or than those of chlorophyll, it is likely that
these pigments in the digestive organs of carnivores represent porphyrins
from an original source of chlorophyll.

DISCUSSION

The major results obtained in this study are summarized in Table I.
The animals are grouped according to their food habits. The echinoids
feed on kelp, on detritus that is predominantly of vegetable orgin, or
upon fixed algae, while the holothurians are bottom dwellers that sub-
sist on the latter two classes of food. The asteroids feed exclusively
on animal matter, especially mollusks, while the ophiuroids feed mainly
on very small participate animal matter.

A few calculations based upon information given in Table I reveal
some interesting qualitative and quantitative differences between the
carotenoids of herbivores and those of carnivores. For example, a total
weight of 847 grams of tissue from herbivorous forms yielded some 4.27
mg. of carotenoid pigment (/3-carotene equivalents), or an average con-
centration of about 0.50 mg. per 100 grams of fresh tissue. In the
carnivores, on the other hand, 916 grams of tissue yielded a measurable
quantity of some 14.61 mg. of carotenoids (/^-carotene equivalents),

PIGMENTS OF PACIFIC COAST ECHINODERMS 451

giving an average concentration of about 1.60 mg. per 100 grams of
fresh tissue. The average ratios of hypophasic to epiphasic pigments
were, in the herbivores, 0.86, and in the carnivores, at least 13. Fur-
thermore, account is to be taken of the fact that the hypophasic pigments
in the ophiuroids were, as in the other carnivores, far in excess of the
epiphasic ones, but that, in three out of four analyses, the color of the
hypophasic fraction bleached before quantitative estimations could be
made. If we assume that the hypophasic pigments in the second catch
of Ophiopteris papillosa exceeded the epiphasic fraction by the same
ratio as was found in the first catch (i.e. about 6) and that this ratio
could be applied also to the other two ophiuroid species, Ophiothrix
spiculata and 0. rudis, we then arrive at a figure of 2 mg. of carotenoids
per 100 grams of fresh carnivore tissue.

Among the carnivores, the ophiuroids contained oxygenated caro-
tenoids exclusively, the epiphasic fractions yielding no carotenes, but
only esterified xanthophylls and acidic carotenoids. The asteroids like-
wise showed a great preponderance of oxygenated over hydrocarbon
carotenoids. Pisastcr gigantcits contained an epiphasic esterified xantho-
phyll which behaved like zeaxanthin. In the herbivores, the echinoids
too contained some xanthophyllic pigments without exception, and most
of them yielded an oxygenated carotenoid, echinenone, in the epiphasic
fraction. One of three catches of Strongylocentrotus piirpuratus yielded
an esterified xanthophyll resembling lutein.

In summary, carnivorous species contained three to four-fold the
quantity of carotenoid pigments found in herbivores. The oxygenated
type of carotenoid, including xanthophylls and acidic compounds, pre-
ponderated vastly over carotenes - in carnivorous species. However, the
herbivores, with the exception of the urchin Strongylocentrotus pur-
puratus which showed a slight excess of xanthophylls over carotenes,
possessed predominating quantities of the latter type of pigment. Fi-
nally, the non-hydrocarbon carotenoids stored in carnivorous animals
were more heavily oxygenated than those in the herbivores, the former
yielding numerous carotenoid acids, as well as taraxanthin and similar
xanthophylls, none of which were encountered in any of the herbivores.

The apparently greater capacity of the carnivorous, as compared
with the herbivorous echinoderms, to assimilate or store carotenoids,
with perhaps a certain degree of oxidative modification, may be associ-

- It should be understood that the term "carotenes" as used here means those
epiphasic pigments resistant to alkaline hydrolysis, and hence may include such
mono-hydroxy xanthophylls as cryptoxanthin and the ketone echinenone. The
possible presence of cryptoxanthin was usually eliminated, however, by employing
in the partition test 95 per cent methanol, in which this pigment is preferentially
soluble.

452 D. L. FOX AND B. T. SCHEER

ated with less metabolic utilization of the oxidized pigments than of the
carotenes. At present nothing seems to be known either of the re-
quirements of marine invertebrates for vitamin A or its precursors,
or of specific roles played by carotenoids in lower phyla.

The carotenoid acids are of especial interest and may be classified
into three types. In vivo some are neutral xanthophylls but treatment
with alcoholic alkali produces molecular rearrangements and oxidations
which lead to the formation of acids (cf. astaxanthin). This type was
found only in the asteroids but was present in all such specimens exam-
ined. The second type is acidic as it is extracted from the tissues. It
was found only in Pisaster, and since it resembles strongly the mytilo-
xanthin described by Scheer (1940) from the California mussel, it is
likely that Pisaster derives this pigment from the mollusk, upon which
it feeds extensively. Finally, acids may occur in vivo as esters, in which
case they are epiphasic before hydrolysis, their acidic properties becom-
ing evident after treatment with alcoholic potash. These esters were
found only in ophiurans and gave on hydrolysis a new type of pigment,
which is acidic, and has a single absorption maximum at 460, 466 or
475 m/x. In addition, an acid-yielding ester of the more conventional
type with a single absorption maximum at 505 m^u appeared.

As in the lobster, carotenoid acids were found to be combined with
protein in one case. Pisaster giganteus contains purple, blue and yellow
chromoproteins in the skin, the purple and blue patches of which yielded
an unesterified carotenoid, which formed an acid on treatment with
alkali. This pigment resembled the anemone pigment metridene (Fox
and Pantin, 1941).

Table II shows some similarities between certain of the echinoderm
carotenoids and known pigments. A commonly occurring one showed
a spectrum like that of /^-carotene or zeaxanthin ; pigments in this group
were found in all classes except the Ophiuroidea, which yielded instead
pigments with spectra like taraxanthin, not found in the other classes.
Spectra like that of a-carotene or lutein were not frequently encountered,
except among the echinoids, wherein both carotenes and xanthophylls
with this type of spectrum were found. Pigments resembling echine-
none have been mentioned above, notably in the epiphase of echinoid
extracts (Lederer, 1938). The unchromatographed extracts, both epi-
phasic and hypophasic, of Stichopus also showed the echinenone type
of spectrum ; again the alkali-resistant epiphase and the hypophase of
one of the carnivores, i.e., Astropecten, as well as certain hypophasic
pigments (rare xanthophylls such as pectenoxanthin) in other groups,
manifested the same absorption maxima.

PIGMENTS OF PACIFIC COAST ECHINODERMS 453

Where individual tissues were studied, the highest values were found
in the digestive tract, excepting in the testes of Lytechinus; certain as-
teroid skins would also show relatively high values were they readily
separable from the colorless skeleton.

The few correlations which could be observed between sex and carot-
enoid content in tissues were limited to three of the echinoid species.
In Dendrastcr, males contained about twice the concentration of epi-
phasic and thrice the hypophasic pigments found in females. Estima-
tions of pigments in individual tissues were not made. In Strongylo-
centrotus purpuratus, the skin of each sex yielded only traces of carot-
enoids. In the intestine, epiphasic pigments were plentiful and of the
same order of concentration in both sexes ; the same was true of the
hypophasic compounds. However, carotenoids in ovaries exceeded
those in testes by three-fold, and were entirely epiphasic, xanthophylls
being absent. In Lytechinus, whole males yielded somewhat more of
each class of carotenoid than did whole females. This was due to a
five-fold excess of epiphasic and nearly a two-fold excess of hypophasic
carotenoids in testicular over ovarian tissues. Intestines, considerably
lighter in mass than gonad material, were also fairly rich in pigment,
females yielding nearly a third again as much epiphasic and almost three
times as much hypophasic pigment as males. The skin of Lytechinus,
like that of Strongylocentrotus, contained only traces of carotenoids.

Females of Strongylocentrotus and Lytechinus contained some 25
and 27 per cent, respectively, of their carotenoids in the ovaries, whereas
corresponding males mobilized approximately 17 and 77 per cent, re-
spectively, to their testes, although the sperm itself was not colored.

The carotenoid distribution ratio between intestine and ovary is
doubtless subject to variation with discharge of ripe eggs. There still
remains the striking mobilization of carotenoids to testicular tissues,
especially in Lytechinus. Also remarkable was the complete absence of
xanthophylls from the gonads of both sexes of Strongylocentrotus.

Three types of non-carotenoid pigment were encountered in this study.
Insoluble pigments, presumably of the melanin type, appeared in Sticho-
pus and Pisaster ochraceus, but were not studied. Green pigments, the
" enterochlorophylls " of MacMunn, were obtained from certain of the
carnivores, but were never found in starved herbivores. These pigments
invariably had absorption maxima at 661 and 663 m/i in absolute etha-
nol. From all the echinoids, but from no others, much red to purple
pigment of the echinchrome type was extracted. It was found as cal-
cium salt in the testes and spines, and as purple (basic) or red (acidic)
irregular bodies dispersed in the tissues of ectoderm, endoderm, intes-
tine, and in some cases, in the gonads. Although quantitative studies

454 D. L. FOX AND B. T. SCHEER

were not made, echinochrome was clearly most abundant in Strongylo-
centrotus, quite abundant in Dendraster, and present in relatively small
amounts in Lyt echinus.

In conclusion it may be reiterated that the echinoderms represent a
structurally and ecologically diverse group of much interest for further
investigations of a comparative biochemical nature, especially regarding
the metabolism of colored compounds.

SUMMARY

1. The pigments of echinoderms, belonging to twelve species, nine
genera and four classes, have been studied qualitatively and in part
quantitatively, with the aid of certain standard methods.

2. Carnivorous species contained more carotenoids in the aggregate
than did herbivores, by some three or four- fold.

3. Among the echinoids, Dendraster and Lytcchinus males yielded
more total carotenoids than did females, while in Strongylocentrotus
purpuratus, the concentrations in each sex were similar. Females of
Strongylocentrotus and Lytechinus contained about a fourth of their
carotenoids in the ovaries, whereas corresponding males mobilized about
one-sixth and three-quarters, respectively, to their testes.

4. Oxygenated carotenoids, including xanthophylls and acidic com-
pounds, preponderated vastly over the hydrocarbon type (carotenes) in
carnivores, while in most herbivores epiphasic pigments, including caro-
tenes and echinenone, showed some degree of predominance.

5. The presence of the ketonic carotenoid echinenone was indicated
in most of the echinoids ; its presence was also regarded as likely in an-
other member of the herbivore-omnivore group, i.e., the cucumber
Stichopus, and in one carnivore, Astropecten.

6. Carotenoid acids, or compounds which yield carotenoid acids on
treatment with alkali, were found only in carnivorous species. A new
type of epiphasic pigment, with a single absorption maximum in the
violet at values of from 460 to 475 m//,, yielding an acid on hydrolysis,
was found in the ophiuroids.

7. Xanthophyll esters were found consistently in the Ophiuroidea
and in one of the Asteroidea, Pisaster gigantcus. In the ophiuroids,
they were of the heavily oxygenated type, and replaced carotenes, which
were completely lacking. One of three catches of Strongylocentrotus
purpuratus yielded some esterified lutein-like xanthophyll.

8. The occurrence of echinochromes, found only in the echinoids,
and of green pigments, " enterochlorophylls," found in the intestines
of carnivores, is discussed.

PIGMENTS OF PACIFIC COAST ECHINODERMS 455

LITERATURE

ABELOOS, M., 1926. Sur les pigments tegumentaires des Asteries. Compt. Rend.

Soc. Biol, 94: 19-21.
EULER, H. v., AND H. HELLSTROM, 1934. Uber Asterinsaure, eine Carotinoidsaure

aus Seesternen. Zcitschr. Physiol. Chcm.. 223 : 89-97.
EULER, H. v., H. HELLSTROM, AND E. KLUSSMAN, 1934. Uber den Carotinoid-

gehalt einiger Evertebraten. Ibid., 228: 77-89.
Fox, D. L., AND C. F. A. PANTIN, 1941. The colours of the plumose anemone,

Metridium senile. Phil. Trans. Roy. Soc. London, Ser. B, 230: 415-450.
KARRER, P., AND F. BENZ, 1934. Uber ein neues Vorkommen des Astacins. Ein

Betrag zu dessen Konstitution. Hclv. Cliim. Acta., 17 : 412-416.
KARRER, P., AND U. SOLMSSEN, 1935. Uber das Vorkommen von Carotinoiden bei

einigen Meerestieren. Ibid., 18: 915-921.
KUHN, R., AND K. WALLENFELS, 1939. Uber die chemische Natur des Stoffes,

den die Eier des Seeigels (Arbacia pustulosa) absondern, um die Spermato-

zoen anzulocken. Bcr. Deutsch. Chcm. Gcs., 72 : 1407-1413.
KURODA, C., AND H. OnsHiMA, 1940. The pigments from the sea urchins and the

synthesis of the related compounds. Proc. Imp. Acad. Japan, 16 : 214-217.
LEDERER, E., 1938. Recherches sur les carotenoides des invertebres. Bull. Soc.

Chim. Biol., 20 : 567-610.
LEDERER, E., 1940. Les pigments des invertebres (a 1'exception des pigments

respiratoires). Biol. Rev., 15: 273-306.
LEDERER, E., AND R. GLASER, 1938. Sur 1'echinochrome et la spinochrome. Compt.

Rend. Acad. Sci., 207 : 454-456.
LEDERER, E., AND T. MOORE, 1936. Echinenone as a Provitamin A. Nature, 137 :

996.
LONNBERG, E., 1931. Untersuchungen uber das Vorkommen carotenoider Stoffe

bei marinen Evertebraten. Ark. Zool., 22A (No. 14) : 1-49.
LONNBERG, E., AND H. HELLSTROM, 1931. Zur Kenntnis der Carotenoide bei

marinen Evertebraten. Ibid., 23 A (No. 15) : 1-74.
LONNBERG, E., 1932. Zur Kenntnis der Carotinoide bei marinen Evertebraten.

II. Ibid., 25A (No. 1) : 1-17.
LONNBERG, E., 1933. Weitere Beitrage zur Kenntnis der Carotinoide der marinen

Evertebraten. Ibid., 26 A (No. 7) : 1-36.
MACMUNN, C. A., 18830. VI. Studies in animal chromatology. Proc. Birm. Phil.

Soc., 3 : 351-407.
MACMUNN, C. A., 18S3/>. Observations on the colouring-matters of the so-called

bile of invertebrates, etc. Proc. Roy. Soc., 35 : 370-403.
MAcMuNN, C. A., 1886. V. Further observations on enterochlorophyll, and allied

pigments. Phil. Trans. Roy. Soc., 177 (Part 1) : 235-266.
k SCHEER, B. T., 1940. Some features of the metabolism of the carotenoid pigments

in the California sea mussel (Mytilus californianus). Jour. Biol. Chcm.,

136 : 275-299.

SMITH, J. H. C., 1936. Carotene. X. A comparison of absorption spectra meas-
urements on alpha-carotene, beta-carotene and lycopene. Jour. Am. Chcm.

Soc., 58 : 247-255.
STRAIN, H. H., 1938. Leaf xanthophylls. Carnegie Institution of Washington :

Publication No. 490, 147 pp.
TYLER, A., 1939. Crystalline echinochrome and spinochrome : their failure to

stimulate the respiration of eggs and of sperm of Strongylocentrotus.

Proc. Nat. Acad. Sci, 25 : 523-528.
YONGE, C. M., 1928. Feeding mechanisms in the invertebrates. Biol. Rev., 3 :

21-76.
YONGE, C. M., 1931. Digestive processes in marine invertebrates and fishes. Jour.

Cons. Intern. E.rpl. Mcr., 6 : 175-212.
ZECHMEISTER, L., AND L. v. CHOLNOKY, 1937. Die Chromatographische Adsorp-

tionsmethode. Vienna: J. Springer, 231 pp.

INDEX

A CIDS, and growth of oat seedlings,

"• 314.

ALBAUM, H. G., AND B. COMMONER.
The relation between the four-
carbon acids and the growth of oat
seedlings, 314.

Amoeba proteus, pH and volume,
gel/sol ratio and action of con-
tractile vacuole, 265.

Anisoporus manteri Hunninen and Cable,
1940, life history, 415.

Anodonta hallenbeckii, role of tissues in
anaerobic metabolism, 79.

Antennal receptors, function in lepi-
dopterous larvae, 403.

Arbacia egg extracts, sperm activation
by, with relation to echinochrome,
202.

jelly, effects of Roentgen
radiation, 363.

- punctulata egg, size of "halves"
and centrifugal force, 354.

Artemia, effect of salinity on rate of
excystment, 194.

Ascidian egg, development of, centri-
fuged before fertilization, 153.

T3ARBER, S. B. See Hunter, Barber

and Caputi, 69.
BEAMS, H. W. See Evans, Beams and

Smith, 363.
BENDITT, E., P. Morrison and L. Irving.

The blood of the Atlantic salmon

during migration, 429.
BERRILL, N. J. Size and morphogenesis

in the bud of Botryllus, 185.
— , -. -. The development of the bud

in Botryllus, 169.
BLACK, E. C. See Ferguson and Black,

139.
— , -. -. See Irving, Black and

Safford, 1.
Blood, combination of oxygen with,

influence of temperature, in trout, 1.

- of Atlantic salmon during migra-
tion, 429.

- of freshwater fishes, carbon dioxide
transport in, 139.

BOTRYLLUS bud, development of, 169.
— , size and morphogenesis in bud, 185.

BOTSFORD, E. F. The effect of physo-
stigmine on the responses of earth-
worm body wall preparations to
successive stimuli, 299.

BURGER, J. W. Some experiments on
the effects of hypophysectomy and
pituitary implantations on the male
Fundulus heteroclitus, 31.

ABLE, R. M. See Hunninen and
Cable, 415.

Calanoid community, distribution of,
effect of water circulation, 86.

CAPUTI, A. P. See Hunter, Barber and
Caputi, 69.

Carbon dioxide, transport in blood of
freshwater fishes, 139.

Centrifugal force, and size of "halves"
of Arbacia punctulata egg, 354.

— , effect of, before fertilization,
on development of ascidian egg, 153.

CHEN, T. See Tartar and Chen, 130.

Ciliates, comparative growth character-
istics of four species, 50.

Ciona, cross- and self-fertilization of, 338.

Circulation of water, effect on distribu-
tion of calanoid community, 86.

Coloration and color changes in gulf-
weed crab, 26.

COMMONER, B. See Albaum and Com-
moner, 314.

CORNMAN, I. Sperm activation by
Arbacia egg extracts, with special
relation to echinochrome, 202.

CULBRETH, S. E. The role of tissues
in the anaerobic metabolism of the
mussel, Anodonta hallenbeckii Lea,
79.

"T\AY, M. F. Pigment migration in

the eyes of the moth, Ephestia

kuehniella Zeller, 275.
DETHIER, V. G. The function of the

antennal receptors in lepidopterous

larvae, 403.

457

458

INDEX

Distribution of calanoid community,
effect of water circulation, 86.

T7ARTHWORM, responses of body

wall to stimuli, effect of physo-

stigmine, 299.
Echinochrome, and sperm activation by

Arbacia egg extracts, 202.
Echinoderms, Pacific coast, pigments of,

441.

— , temperature and righting, 292.
Ephestia kuehniella, pigment migration

in eyes of, 275.
Erythrocytes, chicken, effect of saponin

on osmotic hemolysis of, 69.
Ethinyl testosterone, effect on gono-

podial characteristics produced in

anol fins of Gambusia affinis affinis,

371.
EVANS, T. C., H. W. BEAMS AND M. E.

SMITH. Effects of Roentgen radia-

tion on the jelly of the Arbacia egg,

363.
Excystment, rate of, effect of salinity on,

in Artemia, 194.

pEEDING, method of, in four pele-
cypods, 18.

FERGUSON, J. K. W., AND E. C. BLACK.
The transport of carbon dioxide in
the blood of certain freshwater
fishes, 139.

FERTILIZATION, cross- and self-, of
Ciona, 338.

FOWLER, C. The relation between
hydrogen-ion concentration and vol-
ume, gel/sol ratio and action of the
contractile vacuole in Amoeba pro-
teus, 265.

Fox, D. L. Changes in the tissue
chloride of the California mussel in
response to heterosmotic environ-
ments, 111.

— , -. -., AND B. T. SCHEER. Com-
parative studies of the pigments of
some Pacific coast echinoderms, 441.

Fundulus heteroclitus, effects of hypo-
physectomy and pituitary implan-
tations on male, 31.

callarias, melanosis in, asso-
ciated with trematode infection, 37.
Gambusia affinis affinis, effect of ethinyl
testosterone on gonopodial char-
acteristics produced in anal fins, 371.

GILMAN, L. C. Mating types in diverse
races of Paramecium caudatum, 384.

GILMOUR, D. Repayment of the an-
aerobic oxygen debt in grasshopper
skeletal muscle, 45.

Growth characteristics, comparative, of
four species of sterile ciliates, 50.
- of oat seedlings, four-carbon acids
and, 314.

Gulf of Maine, effect of water circulation
on distribution of calanoid com-
munity, 86.

•LJARVEY, E. B. Relation of the
size of "halves" of the Arbacia
punctulata egg to centrifugal force,
354.

Hemolysis, osmotic, effect of saponin on,
of chicken erythrocytes, 69.

HITCHCOCK, H. B. The coloration and
color changes of the gulf-weed crab,
Planes minutus, 26.

HSIAO, S. C. T. Melanosis in the
common cod, Gadus callarias L.,
associated with trematode infection,
37.

HUNNINEN, A. V., AND R. M. CABLE.

Studies on the life history of
Anisoporus manteri Hunninen and
Cable, 1940 (Trematoda: Allo-
creadiidae), 415.

HUNTER, F. R., S. B. BARBER AND
A. P. CAPUTI. The effect of
saponin on the osmotic hemolysis
of chicken erythrocytes, 69.

Hybridization, diploid and haploid,
between two forms of Rana pipiens,
238.

Hydrogen-ion concentration and volume,
gel/sol ratio and action of contractile
vacuole in Amoeba proteus, 265.

Hypophysectomy, effects of, and pitui-
tary implantations on male Fundu-
lus, 31.

IRVING, L. See Benditt, Morrison

and Irving, 429.

— , -., E. C. BLACK AND V. SAFFORD.
The influence of temperature upon
the combination of oxygen with the
blood of trout, 1.

JENNINGS, R. H., AND D. M.
WHITAKER. The effect of salinity
upon the rate of excystment of
Artemia, 194.

INDEX

459

BIDDER, G. W. Growth studies on
ciliates. VII. Comparative growth
characteristics of four species of
sterile ciliates, 50.

KITCHING, J. A. Studies in sublittoral
ecology. III. Laminaria forest on
the west coast of Scotland; a study
of zonation in relation to wave
action and illumination, 324.

KLEITMAN, N. The effect of tempera-
ture on the righting of echinoderms,
292.

Ku, S. See Tung, Ku and Tung, 153.

T AMINARIA forest, zonation in rela-
tion to wave action and light, 324.

Lepidopterous larvae, function of an-
tennal receptors, 403.

Light, effect on zonation, in Laminaria
forest, 324.

]y/[ACGINITIE, G. E. On the method

of feeding of four pelecypods, 18.
Mating reactions of enucleate fragments

in Paramecium, 130.
types in diverse races of Para-

mecium caudatum, 384.
Melanophores, pituitary regulation of,

in rattlesnake, 228.
Melanosis in common cod, associated

with trematode infection, 37.
Metabolism, anaerobic, role of tissues in,

of Anodonta hallenbeckii, 79.
MORGAN, T. H. Further experiments

in cross- and self-fertilization of

Ciona at Woods Hole and Corona

del Mar, 338.
Morphogenesis and size in Botryllus

bud, 185.
MORRISON, P. See Benditt, Morrison

and Irving, 429.
Muscle, skeletal, repayment of anaerobic

oxygen debt, grasshopper, 45.
Mussel, California, tissue chloride

changes in response to heterosmotic

environments, 111.

^•UTRITIONAL requirements of Tri-
bolium confusum Duval, 208.

seedlings, growth of, four-carbon
acids and, 314.
Oxygen, combination of, with trout
blood, effect of temperature, 1.
- debt, anaerobic, repayment of, in
grasshopper skeletal muscle, 45.

pARAMECIUM bursaria, mating re-
actions of enucleate fragments, 130.
- caudatum, mating types in diverse
races of, 384.

Pelecypods, method of feeding, 18.

Physostigmine, effect on responses of
earthworm body wall preparations
to stimuli, 299.

Pigment migration, eyes of Ephestia,
275.

Pigments of Pacific coast echinoderms,
441.

Pituitary implantations, effect on male
Fundulus, 31.

regulation of melanophores in
rattlesnake, 228.

Planes minutus, coloration and color
changes, 26.

PORTER, K. R. Diploid and andro-
genetic haploid hybridization be-
tween two forms of Rana pipiens,
Schreber, 238.

DAHN, H. The pituitary regulation
of melanophores in the rattlesnake,
228.

Rana pipiens, diploid and androgenetic
haploid hybridization between two
forms, 238.

Rattlesnake, pituitary regulation of
melanophores, 228.

REDFIELD, A. C. The effect of the
circulation of water on the distribu-
tion of the calanoid community in
the Gulf of Maine, 86.

Righting, temperature and, in echino-
derms, 292.

Roentgen radiation, effects on Arbacia
egg jelly, 363.

gAFFORD, V, See Irving, Black and

Safford, 1.
Salinity, effect on rate of excystment of

Artemia, 194.

Salmon, Atlantic, blood during migra-
tion, 429.
Saponin, effect on osmotic hemolysis of

chicken erythrocytes, 69.
SCHEER, B. T. See Fox and Scheer, 441.
SCHNEIDER, B. A. The nutritional

requirements of Tribolium confusum

Duval, I, 208.
Size and morphogenesis in Botryllus

bud, 185.
SMITH, M. E. See Evans, Beams and

Smith, 363.

460

INDEX

Sperm activation by Arbacia egg ex-
tracts, 202.

'"TARTAR, V., AND T. T. CHEN.
Mating reactions of enucleate frag-
ments in Paramecium bursaria, 130.

Temperature, effect on combination of

oxygen with trout blood, 1.
— , effect on righting of echinoderms,
292.

Tissue chloride, changes in California
mussel in response to heterosmotic
environments, 111.

Tissues, in anaerobic metabolism of
Anodonta, 79.

Tribolium confusum Duval, nutritional
requirements, 208.

Trout blood, effect of temperature on
combination of oxygen with, 1.

TUNG, T., S. Ku AND Y. TUNG. The
development of the ascidian egg
centrifuged before fertilization, 153.

TURNER, C. L. Gonopodial character-
istics produced in the anal fins of
females of Gambusia affinis affinis
by treatment with ethinyl testos-
terone, 371.

\/yAVE action, effect on zonation in

Laminaria forest, 424.
WHITAKER, D. M. See Jennings and

Whitaker, 194.

VONATION, with respect to wave
action and light, in Laminaria
forest, 324.

Volume LXXX Number 1

THE

BIOLOGICAL BULLETIN

PUBLISHED BY

THE MARINE BIOLOGICAL LABORATORY

Editorial Board

GARY N. CALKINS, Columbia University E. E. JUST, Howard University

E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago

E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago

SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden

^EITG^°A?LEYA Harva'd.University T. H. MORGAN, California Institute of Technology

L. IRVING, Swarthmore College

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

H. S. JENNINGS, Johns Hopkins University F. SCHRADER, Columbia University

ALFRED C. REDFIELD, Harvard University
Managing Editor

FEBRUARY, 1941

Printed and Issued by

LANCASTER PRESS, Inc.

PRINCE & LEMON STS.

LANCASTER, PA.

Biology Materials

LATEX INJECTED SPECIMENS

All specimens previously injected with
starch, including starfish, lobsters, dogfish,
squid, frogs, turtles, necturus, cats, etc., are
now available with latex injection.

MARINE AQUARIA SETS

In the past 7 years we have sent out nearly
2000 marine aquaria sets with not more than
3% loss. During the period between Novem-
ber 1st and March 1st we guarantee delivery
on these specimens to points indicated in the
living material section of our current cata-
logue. The sets listed in our catalogue have
proved most successful, but should any cus-
tomer wish substitutions on the sets or mate-
rials not listed, we would be pleased to quote
prices on same.

Catalogues sent on request.

Supply Department

MARINE
BIOLOGICAL LABORATORY

Woods Hole, Mass.

WHY ABSTRACTS?

A good abstract:

"(1) enables the reader to determine whether or not a paper is of sufficient
interest to him to warrant his reading it;

(2) it enables the teacher and general reader to keep themselves informed
on the trend of current biological investigations with a minimum of effort; and

(3) it enables any reader to peruse the article more intelligently than he

could otherwise.'

(Smith, R. C., Jour. EC. Ent.
31 (5): 564. N 11, 1938.)

USE BIOLOGICAL ABSTRACTS

A non-profit service by biologists for biologists, a modern necessity for every biology
department library. Are you primarily a zoologist, botanist, geneticist, bacteriologist ? Sub-
scribe personally for your special section and keep up with progress. Complete annual index
goes to every section subscriber.

Send your order now!

BIOLOGICAL ABSTRACTS, University of Pennsylvania, Philadelphia, Pa.

Enter my subscription to: fj BIOLOGICAL ABSTRACTS (Vol. 14, 1940) at $25, postpaid.

Abstracts of Q General Biology at $4.* D Experimental Animal Biology at $8.*

Q Microbiology and Parasitology at $5.* Q Plant Sciences at $6.* Q Animal Sciences at $5.*

Name Address

* Add 50c per section postage to foreign subscriptions.

LANCASTER PRESS, Inc.

LANCASTER, PA.

THE EXPERIENCE we have
gained from printing some
sixty educational publica-
tions has fitted us to meet
the standards of customers
who demand the best.

We shall be happy to have workers at

the MARINE BIOLOGICAL LABORATORY

write for estimates on journals or
monographs. Our prices are moderate.

A Perfect Illustration

Or the lack of it, may make
or mar a scientific paper.

For 65 years we have specialized in
making reproductions by the Helio-
type process of the most delicate
pencil and wash drawings and photo-
graphs;and by the Heliochrome proc-
ess, of paintings and drawings in
color.

Ask the editor to whom you submit
your next paper to secure our esti-
mates for the reproduction of your
illustrations.

The Heliotype Corporation

Est. 1872

172 Green St., Jamaica Plain,
Boston, Mass.

THE BIOLOGICAL BULLETIN

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, W.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, Marine Biological Laboratory, Woods Hole,
Mass., between June 1 and October 1 and to the Biological Lab-
oratories, Divinity Avenue, Cambridge, Mass., during the re-
mainder of the year.

INSTRUCTIONS TO AUTHORS

Preparation of Manuscript. In addition to the text matter, manuscripts
should include a running page head of not more than thirty-five letters.
Footnotes, tables, and legends for figures should be typed on separate sheets.

Preparation of Figures. The dimensions of the printed page (4*4x7
inches) should be borne in mind in preparing figures for publication. Draw-
ings and photographs, as well as any lettering upon them, should be large
enough to remain clear and legible upon reduction to page size. Illustrations
should be planned for sufficient reduction to permit legends to be set below
them. In so far as possible, explanatory matter should be included in the
legends, not lettered on the figures. Statements of magnification should take
into account the amount of reduction necessary. Figures will be reproduced
as line cuts or halftones. Figures intended for reproduction as line cuts
should be drawn in India ink on white paper or blue-lined coordinate paper.
Blue ink will not show in reproduction, so that all guide lines, letters, etc.
must be in India ink. Figures intended for reproduction as halftone plates
should be grouped with as little waste space as possible. Drawings and
lettering for halftone plates should be made directly on heavy Bristol board,
not pasted on, as the outlines of pasted letters or drawings appear in the
reproduction unless removed by an expensive process. Methods of repro-
duction not regularly employed by the Biological Bulletin will be used only
at the author's expense. The originals of illustrations will not be returned
except by special request.

Directions for Mailing. Manuscripts and illustrations should be packed
flat between stiff cardboards. Large charts and graphs may be rolled and
sent in a mailing tube.

Reprints. Authors will be furnished, free of charge, one hundred re-
prints without covers. Additional copies may be obtained at cost.

Proof. Page proof will be furnished only upon special request. When
cross-references are made in the text, the material referred to should be
marked clearly on the galley proof in order that the proper page numbers
may be supplied. Manuscripts should be returned with galley proof.

Entered October 10, 1902, at Lancaster, Pa., as second-class matter under
Act of Congress of July 16, 1894.

CONTENTS

Page
IRVING, L., E. C. BLACK AND V. SAFFORD

The Influence of Temperature upon the Combination of
Oxygen with the Blood of Trout 1

MACGINITIE, G. E.

On the Method of Feeding of Four Pelecypods 18

HITCHCOCK, HAROLD B.

The Coloration and Color Changes of the Gulf-weed Crab
Planes minutus 26

BURGER, J. WENDELL

Some Experiments on the Effects of Hypophysectomy and
Pituitary Implantations on the Male Fundulus heteroclitus . . 31

HSIAO, SIDNEY C. T.

Melanosis in the Common Cod, Gadus callarias L., Asso-
ciated with Trematode Infection 37

GILMOUR, DARCY

Repayment of the Anaerobic Oxygen Debt in Grasshopper
Skeletal Muscle 45

KIDDER, GEORGE W.

Growth Studies on Ciliates. VII. Comparative Growth Char-
acteristics of Four Species of Sterile CUiates 50

HUNTER, F. R., S. B. BARBER AND A. P. CAPUTI

The Effect of Saponin on the Osmotic Hemolysis of Chicken
Erythrocytes 69

CULBRETH, SARAH E.

The Role of Tissues in the Anaerobic Metabolism of the
Mussel Anodonta hallenbeckii Lea 79

REDFIELD, ALFRED C.

The Effect of the Circulation of Water on the Distribution of
the Calanoid Community in the Gulf of Maine 86

Fox, DENIS L.

Changes in the Tissue Chloride of the California Mussel in
Response to Heterosmotic Environments Ill

TARTAR, VANCE AND TZE-TUAN CHEN

Mating Reactions of Enucleate Fragments in Paramecium
bursaria . 130

Volume LXXX

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
LEIGH HOADLEY, Harvard University
L. IRVING, Swarthmore College
M. H. JACOBS, University of Pennsylvania
H. S. JENNINGS, Johns Hopkins University

E. E. JUST, Howard University
FRANK R. LILLIE, 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

F. SCHRADER, Columbia University

ALFRED C. REDFIELD, Harvard University
Managing Editor

APRIL, 1941

Printed and Issued by

LANCASTER PRESS, Inc.

PRINCE & LEMON STS.

LANCASTER, PA.

ANNOUNCEMENT

//N INDEX of the Biological Bulletin, Volumes 61 to 80,
^— ^ -* will be published in October, 1941. This Index will
contain an alphabetical list of authors, showing the titles of their
papers, and a classified index of subjects. The published abstracts
of papers presented at the Marine Biological Laboratory are also
indexed in this volume.

The edition of the Index is limited. Orders should be sent
to the Marine Biological Laboratory, Woods Hole, Mass.

The price of the Index is $2.50 postpaid. The Index to Vols.
1-60 and the Index to Vols. 61-80 are now offered at a combi-
nation price of $5.00.

ORDER FORM

MARINE BIOLOGICAL LABORATORY,
WOODS HOLE, MASSACHUSETTS.

Please enter my order for one copy of the Index of the

Biological Bulletin, Volumes 61 to 80, at the price of $2.50.

(check here)

Please enter my order for the Index to Vols. 1-60 and

the Index to Vols. 61-80 at a total price of $5.00.

(check here)

NAME
ADDRESS

WHY ABSTRACTS?

A good abstract:

"(1) enables the reader to determine whether or not a paper is of sufficient
interest to him to warrant his reading it;

(2) it enables the teacher and general reader to keep themselves informed
on the trend of current biological investigations with a minimum of effort; and

(3) it enables any reader to peruse the article more intelligently than he

could otherwise." /c ... „ n T *? r? *

(Smith, R. C., Jour. EC. Ent.

31 (5): 564. N 11, 1938.)

USE BIOLOGICAL ABS.TRACTS

Send your order now!

BIOLOGICAL ABSTRACTS, Umversi ty of Pennsylvania, Philadelphia, Pa.

Enter my subscription to: D BIOLOGICAL ABSTRACTS (Vol. 14, 1940) at $25, postpaid.
Abstracts of Q General Biology at $4.* Q Experimental Animal Biology at $8.*

n Microbiology and Parasitology at $5.* Q Plant Sciences at $6.* Q Animal Sciences at $5.*

Name Address

* Add 50c per section postage to foreign subscriptions.

LANCASTER PRESS, Inc.

LANCASTER, PA.

THE EXPERIENCE we have
gained from printing some
sixty educational publica-
tions has fitted us to meet
the standards of customers
who demand the best.

We shall be happy to have workers at

the MARINE BIOLOGICAL LABORATORY

write for estimates on journals or
monographs. Our prices are moderate.

A Perfect Illustration

Or the lack of it, may make
or mar a scientific paper.

For 65 years we have specialized in
making reproductions by the Helio-
type process of the most delicate
pencil and wash drawings and photo-
graphs; and by the Heliochrome proc-
ess, of paintings and drawings in
color.

Ask the editor to whom you submit
your next paper to secure our esti-
mates for the reproduction of your
illustrations.

The Heliotype Corporation

Est. 1872

172 Green St., Jamaica Plain,
Boston, Mass.

THE BIOLOGICAL BULLETIN

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, W.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, Marine Biological Laboratory, Woods Hole,
Mass., between June 1 and October 1 and to the Biological Lab-
oratories, Divinity Avenue, Cambridge, Mass., during the re-
mainder of the year.

INSTRUCTIONS TO AUTHORS

Directions for Mailing. Manuscripts and illustrations should be packed
flat between stiff cardboards. Large charts and graphs may be rolled and
sent in a mailing tube.

Reprints. Authors will be furnished, free of charge, one hundred re-
prints without covers. Additional copies may be obtained at cost.

Entered October 10, 1902, at Lancaster, Pa., as second-class matter under
Act of Congress of July 16, 1894.

Biology Materials

PRESERVED SPECIMENS

for

Zoology, Botany, Embryology,
and Comparative Anatomy

MICROSCOPE SLIDES

for

Zoology, Botany, Embryology,
Histology, Bacteriology, and
Parasitology.

LIVING SPECIMENS

for
Zoology and Botany

including Protozoan and
Drosophila Cultures, and
Animals for Experimental and
Laboratory use.

Catalogues promptly sent on request.

Supply Department

MARINE
BIOLOGICAL LABORATORY

Woods Hole, Mass.

CONTENTS

Page
FERGUSON, J. K. W., AND E. C. BLACK

The Transport of CO2 in the Blood of Certain Freshwater
Fishes 139

TUNG, T, S. Ku AND Y. TUNG

The Development of the Ascidian Egg Centrifuged Before
Fertilization 153

BERRILL, N. J.

The Development of the Bud in Botryllus 169

BERRILL, N. J.

Size and Morphogenesis in the Bud of Botryllus 185

JENNINGS, R. H., AND D. M. WHITAKER

The Effect of Salinity upon the Rate of Excystment of Ar-
temia 194

CORNMAN, IVOR

Sperm Activation by Arbacia Egg Extracts, with Special Rela-
tion to Echinochrome 202

SCHNEIDER, B. AUBREY

The Nutritional Requirements of Triboliumconfusum Duval, I. 208

RAHN, HERMANN

The Pituitary Regulation of Melanophores in the Rattlesnake 228

PORTER, K. R.

Diploid and Androgenetic Haploid Hybridization between
Two Forms of Rana pipiens, Schreber 238

FOWLER, COLEEN

The Relation Between Hydrogen-Ion Concentration and
Volume, Gel /sol Ratio and Action of the Contractile Vacuole
in Amoeba proteus 265

Volume LXXX

Number 3

THE

BIOLOGICAL BULLETIN

PUBLISHED BY

THE MARINE BIOLOGICAL LABORATORY

Editorial Board

E. E. JUST, Howard University
FRANK R. LULIE, 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

F. SCHRADER, Columbia University

ALFRED C. REDFIELD, Harvard University
Managing Editor

JUNE, 1941

Printed and Issued by

LANCASTER PRESS, Inc.

PRINCE & LEMON STS.

LANCASTER, PA.

ANNOUNCEMENT

INDEX of the Biological Bulletin, Volumes 61 to 80,
will be published in October, 1941. This Index will
contain an alphabetical list of authors, showing the titles of their
papers, and a classified index of subjects. The published abstracts
of papers presented at the Marine Biological Laboratory are also
indexed in this volume.

The edition of the Index is limited. Orders should be sent
to the Marine Biological Laboratory, Woods Hole, Mass.

The price of the Index is $2.50 postpaid. The Index to Vols.
1-60 and the Index to Vols. 61-80 are now offered at a combi-
nation price of $5.00.

ORDER FORM

MARINE BIOLOGICAL LABORATORY,
WOODS HOLE, MASSACHUSETTS.

Please enter my order for one copy of the Index of the

Biological Bulletin, Volumes 61 to 80, at the price of $2.50.

(check here)

Please enter my order for the Index to Vols. 1-60 and

the Index to Vols. 61-80 at a total price of $5.00.

(check here)

NAME

ADDRESS

WHY ABSTRACTS?

A good abstract:

"(1) enables the reader to determine whether or not a paper is of sufficient
interest to him to warrant his reading it;

(2) it enables the teacher and general reader to keep themselves informed
on the trend of current biological investigations with a minimum of effort; and

(3) it enables any reader to peruse the article more intelligently than he

could otherwise."

(bnnth, K. L., Jour. be. Ent.

31 (5): 564. N 11, 1938.)

USE BIOLOGICAL ABSTRACTS

Send your order now!

BIOLOGICAL ABSTRACTS, University of Pennsylvania, Philadelphia, Pa.

Enter my subscription to: Q BIOLOGICAL ABSTRACTS (Vol. 14, 1940) at $25, postpaid.
Abstracts of Q General Biology at $4.* Q Experimental Animal Biology at $8.*

G Microbiology and Parasitology at $5.* Q Plant Sciences at $6.* Q Animal Sciences at $3.*

Name Address

* Add 50c per section postage to foreign subscriptions.

LANCASTER PRESS, Inc.

LANCASTER, PA.

THE EXPERIENCE we have
gained from printing some
sixty educational publica-
tions has fitted us to meet
the standards of customers
who demand the best.

We shall be happy to have workers at

the MARINE BIOLOGICAL LABORATORY

write for estimates on journals or
monographs. Our prices are moderate.

A Perfect Illustration

Or the lack of it, may make
or mar a scientific paper.

For 65 years we have specialized in
making reproductions by the Helio-
type process of the most delicate
pencil and wash drawings and photo-
graphs; and by the Heliochrome proc-
ess, of paintings and drawings in
color.

Ask the editor to whom you submit
your next paper to secure our esti-
mates for the reproduction of your
illustrations.

The Heliotype Corporation

Est. 1872

172 Green St., Jamaica Plain,
Boston, Mass.

THE BIOLOGICAL BULLETIN

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, W.C. 2.

Communications relative to manuscripts should be sent to the
Managing Editor, Marine Biological Laboratory, Woods Hole,
Mass., between June 1 and October 1 and to the Biological Lab-
oratories, Divinity Avenue, Cambridge, Mass., during the re-
mainder of the year.

INSTRUCTIONS TO AUTHORS

Preparation of Figures. The dimensions of the printed page (4}4x7
inches) should be borne in mind in preparing figures for publication. Draw-
ings and photographs, as well as any lettering upon them, should be large
enough to remain clear and legible upon reduction to page size. Illustrations
should be planned for sufficient reduction to permit legends to be set below
them. In so far as possible, explanatory matter should be included in the
legends, not lettered on the figures. Statements of magnification should take
into account the amount of reduction necessary. Figures will be reproduced
as line cuts or halftones. Figures intended for reproduction as line cuts
should be drawn in India ink on white paper or blue-lined coordinate paper.
Blue ink will not show in reproduction, so that all guide lines, letters, etc.
must be in India ink. Figures intended for reproduction as halftone plates
should be grouped with as little waste space as possible. Drawings and
lettering for halftone plates should be made directly on heavy Bristol board,
not pasted on, as the outlines of pasted letters or drawings appear in the
reproduction unless removed by an expensive process. Methods of repro-
duction not regularly employed by the Biological Bulletin will be used only
at the author's expense. The originals of illustrations will not be returned
except by special request.

Directions for Mailing. Manuscripts and illustrations should be packed
flat between stiff cardboards. Large charts and graphs may be rolled and
sent in a mailing tube.

Reprints. Authors will be furnished, free of charge, one hundred re-
prints without covers. Additional copies may be obtained at cost.

Entered October 10, 1902, at Lancaster, Pa., as second-class matter under
Act of Congress of July 16, 1894.

BIOLOGY SUPPLIES

The Supply Department of the Marine Biological Labora-
tory has a complete stock of excellent plain preserved and
latex injected materials, and would be pleased to quote prices
on your summer school needs.

PRESERVED SPECIMENS

for

Zoology, Botany, Embryology,
and Comparative Anatomy

LIVING SPECIMENS

for
Zoology and Botany

including Protozoan and
Drosophila Cultures, and
Animals for Experimental and
Laboratory Use.

MICROSCOPE SLIDES

for

Zoology, Botany, Embryology,
Histology, Bacteriology, and
Parasitology.

Catalogues promptly sent on request.
Supply Department

MARINE
BIOLOGICAL LABORATORY

Woods Hole, Mass.

CONTENTS

Page
DAY, M. F.

Pigment Migration in the Eyes of the Moth, Ephestia kuehni-
ella Zeller 275

KLEITMAN, NATHANIEL

The Effect of Temperature on the Righting of Echinoderms . . 292

BOTSFORD, E. FRANCES

The Effect of Physostigmine on the Responses of Earthworm
Body Wall Preparations to Successive Stimuli 299

ALBAUM, HARRY G., AND BARRY COMMONER

The Relation between the Four-Carbon Acids and the Growth
of Oat Seedlings 314

KITCHING, J. A.

Studies in Sublittoral Ecology. III. Laminaria forest on the
west coast of Scotland; a study of zonation in relation to wave
action and illumination 324

MORGAN, T. H.

Further Experiments in Cross- and Self-Fertilization of Ciona

at Woods Hole and Corona del Mar 338

HARVEY, ETHEL BROWNE

Relation of the Size of " Halves " of the Arbacia punctulata
Egg to Centrifugal Force 354

EVANS, T. C., H. W. BEAMS AND MARSHALL E. SMITH

Effects of Roentgen Radiation on the Jelly of the Arbacia Egg 363

TURNER, C. L.

Gonopodial Characteristics Produced in the Anal Fins of
Females of Gambusia affinis affinis by Treatment with
Ethinyl Testosterone 371

GILMAN, LAUREN C.

Mating Types in Diverse Races of Paramecium caudatum . . 384

DETHIER, V. G.

The Function of the Antennal Receptors hi Lepidopterous
Larvae 403

HUNNINEN, A. V., AND R. M. CABLE

Studies on the Life History of Anisoporus manteri Hunninen
and Cable, 1940 (Trematoda: Allocreadiidae) 415

BENDITT, E., P. MORRISON AND L. IRVING

The Blood of the Atlantic Salmon during Migration 429

Fox, DENIS L., AND BRADLEY T. SCHEER

Comparative Studies of the Pigments of Some Pacific Coast
Echinoderms. 441

MBI. WHO! LIBRARY

UH 17IX M

'•:);)!!:

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "The Biological bulletin"

See other formats