Viriitge
“Gag 1A pes
ASE Piet Fs y,
ee
Gee Une bbe’
rs
eee)
Tete
“Tense?
ae ;
nied
Paid
er er et ee
A,
ia 7
CUT es
SEDO 3s
3
Hei
Pes BRD
ite
sew
th
‘
wih seee
agntN
alte
esa 4)
care aiany
A jrtices|
Heit
res
¢
sity
ts reed Saleh:
i
dats ate
brig ind a dn boty Shak Day
il bret i yen A
Ne
?
rh is]
iets)
ah
ita
rf
ts lbg >
+
vt Shei $25)
yt SS
aie ay:
STts
foaye
fi
Dart
saa ieeed&
iaien 438
Shes inoe bs tees
*
aS se
Rechte
f Rin
ie J.
ited
Porpeht ey
ve
telat
apie
ry posit
et
gies
‘= na
3
oe SATIRE Tate
a ieveriegeate
Waray aes
(5
Shewes
ef
a a neo ie
SBR a Reh Teochateh
Beane als
(haved 1% ph)
iy Fe FAGoaste! sponte (it th bee
: yh ate
send
Chie
re
aaths bye
sy eaeay
vs
ae MET reve)
yg Rms vi
Tighe te
me Natalee
na earl Ha VER
DT aah sesteket
eh 39nd AM ot
che,
ae
¥
fe. ES
Hates
fe As
wei
red)
©
betayel
Lyi
ARSC Nae
TR seg At be Free
Rise hatek pita ra ee
Wis 1e
ei rdghe
i ts
i
Ky
$y
bai at
ne nt Se aby;
iyiei
ae
.
As
be
+ art
Hig di 4 aya
staleboagetts
ads an
anouaty
bd
age:
stint 8 Ge
dvinneys iia
at
oH
boa
a
51
He}
<t
‘e
ty
thts
ae
mich}
chase a on
sptieeet
nied tag pode
carts Teer
renin
Digitized by the Internet Archive
in 2009 with funding from
University of Toronto
http://www.archive.org/details/journalofexperim14broo
tron SER a
ni
4
THE JOURNAL
OF
EXPERIMENTAL ZOOLOGY
EDITED BY
WILLiaAM E. CastTLe FrankK R. LILLIE
Harvard University University of Chicago
Epwin G. ConKLIN Jacques LoEB
Princeton University Rockefeller Institute
CuHarLEs B. DAVENPORT THomas H. MorGan
Carnegie Institution Columbia University
Horack JAYNE GeorGcEe H. PARKER
The Wistar Institute Harvard University
HERBERT S. JENNINGS Epmunp B. WILSON,
Johns Hopkins University Columbia University
and
Ross G. HARRISON, Yale University
Managing Editor
VOLUME 14 ,
O
1913 ay
a /\s
\ Sy
THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY
PHILADELPHIA, PA.
uA
oa Ue ye
|
y 4
2 Co b-2
j a
fe oe Ma 0
Ts Fe:
. . THE WAVERLY . ,
vF : BALTIMORE, U.S.A,
; ‘ :
“= ik
CONTENTS
19138
No.1 JANUARY 20
Lron J. Cote. Direction of locomotion of the starfish (Asterias forbesi)... 1
T. H. Morean anv Evets Catrety. Additional data for the study of sex-
Hmkecunheritance im Orosophilas. ty «2. ecb ce. mew oaks ee oe ee ce 33
A. H. Srurtevant. The linear arrangement of six sex-linked factors in
Drosophila, as shown by their mode of association........................ 43
Eruet NicHotson Browne. A study of the male germ cells in Notonecta.... 61
L. B. Nice. Studies on the effects of alcohol, nicotine and caffeine on white
ee ee eH CCUS OH AGCULVILY. &. 005 ck. silos Wed dacs atecs cle ascenetucee: 123
No.2 FEBRUARY 20
C. M. Cuitp. Studies on the dynamics of morphogenesis and inheritance
in experimental reproduction. V. The relation between resistance to
depressing agents and rate of metabolism in Planaria dorotocephala and
iimvalue/as a method of investigation...2...:..-...e.4eceee.s.see ces ss aloe
V. &. SHELForD anp W. C. Atuter. The reactions of fishes to gradients
DacSsolved: AbMOSPheLiC) SASS... o.28s on os <2 Gl Seuws ss eae owe meses oe 207
JacevEs Lors anp F. W. Bancrort. The sex of a parthenogenetic tadpole
SIMRAN Se Me ai INL LS oy SOD 1 AY Oe ee «5 atu ee 267
Frank D. Lutz. Experiments concerning the sexual difference in the
wing length of Drosophila ampelophial...............6......00s 2. ene e ees 271
No.3 APRIL 5
H. S. Jennines. The effect of conjugationin paramecium. Two figures..... 279
H. S. Jennincs anp K. S. Lasuiey. Biparental inheritance and the ques-
tion of sexuality in paramecium. Two figures...................--.. ... 393
No. 4 MAY 20
AuFrEeD O. Gross. The reactions of arthropods to monochromatic lights
of equal intensities... Mervy-five figuress.cci4. nw. 22 wee. een ee eee 467
Frank R. Liviie. Studies of fertilization. V. The behavior of the sperma-
tozoa of Nereis and Arbacia with special reference to egg-extractives.
vedi Cues... 5)..-. eee Pap Jake, stn bot eee 515
LoranvE Loss Wooprurr. The effect of excretion products of Infusoria
on the same and on different species, with special reference to the proto-
BanresequenCe if INTUSIONS? . . 3 <a asehednge Re sere <> <2 = 2s SP ees a 575
ru
a
Al eer inf yeie
3
AUTHOR AND SUBJECT INDEX
CTIVITY. Studies on the effects of alcohol,
nicotine and caffeine on white mice. II.
Effects on 123
Alcohol, nicotine and caffeine on white mice. Stu-
dies on the effectsof II. Effects on activity. 123
ALLEE, W. C., SHELFORD, V. E. and The reactions
of fishes to gradients of dissolved atmospheric
gases. 207
Arbacia with special reference to egg-extractives.
Studies of fertilization. V. The behavior of the
spermatozoa of Nereis and Sole
Arthropods to monochromatic lights of equal inten-
sities. The reactions of 467
ANCROFT, F. W., LOEB, JACQUES and The
sex of a parthenogenetic tadpole and frog 267
Biparental inheritance and the question of sexuality
in Paramecium. 393
Browne, Eruet NicuHortson. A study of the male
germ cells in Notonecta. 61
AFFEINE on white mice. Studies on the effects
of alcohol, nicotine and II. Effects on a
ity.
Car?rety, EvetH, Morean, T. H. and Additional
data for the study of sex-linked inheritance in
Drosophila. 33
Cells in Notonecta. A study of the male germ 61
Cuitp, C. M. Studies on the dynamics of morpho-
genesis and inheritance in experimental repro-
duction. V. The relation between resistance to
depressing agents and rate of metabolism in Pla-
naria dorotocephala and its value as a method
of investigation. 153
Cour, Leon J. Direction of locomotion of the star-
fish (Asterias forbesi).
Conjugation in Paramecium. The effect of 279
EPRESSING agents and rate of metabolism in
Planaria dorotocephala and its value as a
method of investigation. Studies on the dy-
namics of morphogenesis and inheritance in
experimental reproduction. V. The relation
between resistance to 153
Dissolved atmospheric gases. The reactions of fishes
to gradients of 207
Dorotocephala and its value as a method of investi-
gation. Studies on the dynamics of morpho-
genesis and inheritance in experimental repro-
duetion. V. The relation between resistance to
depressing agents and rate of metabolism in
Planaria 153
Drosophila. Additional data for the study of sex-
linked inheritance in 33
Drosophila ampelophila. Experiments concerning
the sexual difference in the winglengthof 271
Dynamics of morphogenesis and inheritance in ex-
perimental reproduction. Studies on the V.
The relation between resistance to depressing
agents andrate of metabolism in Planaria doro-
tocephala and its value as a method of investi-
gation. 153
TE eee ere Studies of fertilization.
V. The behavior of the spermatozoa of Nereis
and Arbacia with special reference to 515
Excretion products of Infusoria on the same and on
different species, with special reference to the
protozoan sequences in infusions. The ses
Hiss in Drosophila, as shown by their mode
of association. The linear arrangement of six
sex-linked 43
Fertilization. Studies of V. The behavior of the
spermatozoa of Nereis and Arbacia with special
reference to egg-extractives. 515
Fishes to gradients of dissolved atmospheric gases.
The reactions of ; 7
Frog. The sex of a parthenogenetic tadpole and 267
ERM cells in Notonecta. A study of the male
61
Gross, AtrreD O. The reactions of arthropods to
monochromatic lights of equal intensities. 467
NFUSIONS. The effect of excretion products of
Infusoria on the same and on different species,
with special reference to the protozoan pence
in
Infusoria on the same and on different species, with
special reference to the protozoan sequence in
infusions. The effect of excretion products a
Inheritance and the question of sexuality in Para-
mecium. Biparental Bit)
Inheritance in Drosophila. Additiona] data for the
study of sex-linked 33
Inheritance in experimental reproduction. Studies
on the dynamics of morphogenesis and YV. The
relation between resistance to depressing agents
and rate of metabolism in Planaria dorotoceph-~
ala and its value as a method of paves cation:
ENNINGS, H. S. and LASHLEY, K. 8. Bi-
parental inheritance and the question of sex-
uality in Paramecium. 393
JENNINGS, H.S. . The effect of conjugation in Para-
mecium. 279
ASHLEY, K. S., JENNINGS, H.S. and Bi-
parental inheritance and the question of “ie
uality in Paramecium. 93
Lights of equal intensities. The reactions of arthro-
pods to monochromatic 467
Litiie, Frank R. Studies of fertilization. V. The
behavior of the spermatozoa of Nereis.and Ar-
bacia with special reference to 2 egal
o
Locomotion of the starfish (Asterias forbesi). pa
tion of
Logs, Jacques and Bancrort, F. W.
a parthenogenetic tadpole and frog.
Lutz, FRANK D. Experiments concerning the sex-
ual difference in the wing length of Drosophila
ampelophila. 271
The sex of
267
ETABOLISM in Planaria dorotocephala and its
value as a method of investigation. Studies
on the dynamics of morphogenesis and in-
heritance in experimental reproduction. V.
The relation between resistance to depressing
agents and rate of 153
Monochromatic lights of equal intensities. The
reactions of arthropods to 467
Morean, T. H. and Carrert, Evers. Additional
data for the study of sex-linked inheritance in
Drosophila. 33
vil AUTHOR AND SUBJECT INDEX
Morphogenesis and inheritance in experimental re-
production. Studies on the dynamics of V.
The relation between resistance to depressing
agents and rate of metabolism in Planaria doro-
tocephala and its value as a method of investi-
gation. 153
EREIS and Arbacia with special reference to
egg-extractives. Studies of fertilization. Ve
The behavior of the spermatozoa of 515
Nice, L. B. Studies on the effects of alcohol, nico-
tine and caffeine on white mice. II. Effects on
activity. 123
Nicotine and caffeine on white mice. Studies on
the effects of alcohol. II. Effects on pene
Notonecta. A study of the malegermecellsin 61
ARAMECIUM. Biparental inheritance and the
question of sexuality in 393
Paramecium. The effect of conjugation in 279
Parthenogenetic tadpole and frog. The sex of a 267
Planaria dorotocephala and its value as a method
of investigation. Studies on the dynamics of
morphogenesis and inheritance in experimental
reproduction. V. The relation between resist-
ance to depressing agents and rate of metabo-
lism in 153
Protozoan sequence in infusions. The effect of ex-
cretion product of Infusoria on the same and on
different species, with special reference to the 575
EACTIONS of arthropods to monochromatic
lights of equal intensities. The 467
Reactions of fishes to gradients of dissolved atmos-
pheric gases. The 207
Reproduction. Studies on the dynamics of mor-
phogenesis and inheritance in experimental. V.
The relation between resistance to depressing
agents and rate of metabolism in Planaria doro-
tocephala and its value as a method of investi-
gation. 153
Starfish (Asterias forbesi).
Resistance to depressing agents and rate of metzbo-
lism in Planaria dorotocephala and its valueas
a method of investigation. Studies on tte
dynamics of morphogenesis and inheritance i,
experimental reproduction. V. The relation
between 153
EX-LINKED factors in Drosophila, as shown
by their mode cf association. The linear ar-
rangement of six 43
Sex-linked inheritance in Drosophila.
data for the study of
Sex of a parthenogenetic tadpole and frog. The 267
Sexual difference in the wing length of Drosophila
ampelophila. Experiments concerning the 271
Sexuality in Paramecium. Biparental inheritance
and the question of 393
SHELFORD, V. E. and ALLEE, W. C. The reactions
of fishes to gradients of dissolved atmospheric
gases. 207
Spermatozoa of Nereis and Arbacia with special
reference to egg-extractives. Studies of fertili-
zation. V. The behavior of the 515
Direction of locomo-
Additional
33
tion of the 1
Sturtevant, A. H. The linear arrangement of six
sex-linked factors in Drosophila, as shown by
their mode of association. 43
f eae and frog. The sex of a partheno
genetic 267
HITE mice. Studies on the effects of alcohol,
nicotine and caffeine on II. Effects on ac-
tivity. 123
Wing length of Drosophila ampelophila. Experi-
ments concerning the sexual difference in >
Wooprtvrr, LoRANDE Loss. The effect of excretion
products of Infusoria on the same and on differ-
ent species, with special reference to the proto-
zoan sequence in infusions. 575
DIRECTION OF LOCOMOTION OF THE STARFISH
(ASTERIAS FORBESI):
LEON J. COLE
College of Agriculture, University of Wisconsin, Madison, Wis.
NINE FIGURES
CONTENTS
ih. Theta he Ri. 3h oe eee ee eee 1
EL. DY tein ering [reir ove 0s Re a co 3
PPE RMe HaMOeGia OCOMOUMON Gs a5: 4. ks. SQs ko iS Petes et Aes oa nc sos ees 6
0) DUPE EPD ET PLL SSRN LG SOS a oe 8
Pe DITE CTO OF (GEOMOLIONN <2 ar see Acie Aoi ae sbeystsernie 2 Vi he 8
2. Relation of direction of locomotion to length of arm............. 12
SEAL EHCC IGT GE ATAPUISE, ony.) hh fom Eien’ oe Se eles yee swe 5 2 oes 16
Neu PbTOTINGE TRC TODUISE: 6a one hoc sa as ow Sige cls ens fens ee eee 17
5. Relation of arms used in righting to direction of locomotion..... 19
V. Comparison with other echinoderms................5.5.---22++eee eee: 23
1_ Preyer’s experiments on starfishes...........-.....5.022+-24--+5- 24
PaSRChinolds and Other echinoderms, 2 oi. 2. 22s -Joascnclate = See 26
M@are latins GISCUSBION.. .2.0. 5 25-25-52. Dele cs ew oe He Ree we 28
EIT TE eps 8 hes So eis: hia eens he Se peas: oo 31
SS RRCRIEL LC Cte ae ye ei ee eee Re Fe ee oie ieee Cerca eer 31
I. INTRODUCTION
Jennings (07), in his comprehensive paper on the behavior
of the starfish Asterias forreri de Loriol, devotes considerable
attention to the righting reactions of this animal. One of the
results stated in this connection is as follows (loc. cit., p. 144):
There is for some reason a general tendency, seen in all the speci-
mens, to use certain definite rays for the pulling over. A strong tend-
ency is evident toward using the rays lying close to the madreporic
1 For a preliminary statement of some of the results detailed in the present
paper, see Cole 710.
1
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
JANUARY, 1913
2 | LEON J. COLE
plate. The ray e is used 89 times out of the 95, and the next greatest
numbers are shown by the two rays lying on either side of e, namely
a (56) and d (43). The combination of the two rays lying at the sides
of the madreporic plate (a + e), was used 37 times. On the other
hand, the rays lying opposite (b and c) were used but rarely, and not
once in the whole 95 experiments was the pair b + ¢ used in combination.
To make this statement intelligible, it will be necessary to
explain the notation used by Jennings in designating the various
rays. This is shown in figure 1, which is reproduced from the
diagram given by Jennings. Beginning with the arm at the
Fig. 1 Diagram of starfish, showing notation of rays with respect to position
of madreporic plate (from Jennings ’07). Preyer designated arms a, b, c, etc., as
1, 2, 3, ete. respectively (cf. p. 24). For comparison with the system adopted
by Lovén, and in common use by. specialists, see figure 9, p. 28.
right of the madreporite (the right hand member of the ‘bivium’)
and going clockwise, the arms were designated arbitrarily by |
the letters a to e respectively. Thus it will be seen that b, c, d,
compose the ‘trivium’, while e is the left hand member of the
‘bivium.’ R .
The question at once suggests itself: What is the reason
that the rays of a certain region are used most often in the right-
ing process? Does it mean possibly that there is something
akin to an antero-posterior? differentiation? It would be inter-
* It must be borne in mind that ‘anterior’ and ‘posterior’ are here used in rela-
tion to the assumed position of the starfish in its natural locomotion, and not
in their morphological sense.
DIRECTION OF LOCOMOTION OF THE STARFISH 3
esting to know in what direction, that is, with what ray forward,
these starfish would have crawled if left undisturbed after right-
ing themselves. The question then naturally occurs: Do star-
fishes naturally crawl with any particular region of the body
(radius or interradius) forward? It is a well known fact that
they may crawl with any part in advance. As Jennings (’07,
pee 155) put it:
The starfish is not hampered by any considerations of anterior and
posterior; it may move with any one of its rays in the lead, or with
any interradius in advance, or indeed in any intermediate direction,
so that its possibilities as to variations of direction of locomotion are
really unlimited.
But, other things being equal, is this a matter of indifference?
In certain echinoderms, e. g., holothurians and spatangoids,
there is a well defined antero-posterior differentiation with respect
to locomotion, accompanied, in most cases, by decided secondary
morphological bilaterality. It was in the attempt to throw
some light on these questions that the experiments reported in
this paper were undertaken.
II. MATERIAL AND METHODS
While it was hoped to test the matter not only with asteroids,
but with representatives of the other classes of Echinodermata
as well, the only animal with which more than tentative experi-
ments were accomplished was one of the common starfishes of
the New England coast, Asterias forbesi (Desor). The work
was begun at the United States Fisheries Laboratory’ at Woods
Hole in the summer of 1909, and was.continued later in the same
season at the Zoological Laboratory of the Sheffield Scientific
School of Yale University, at New Haven.
As it was desired to determine the direction of locomotion
in the absence of external stimuli of a directive nature, it became
- necessary to make the conditions of the experiments as uniform
3] wish to express my indebtedness to Dr. Francis B. Summer, Director of
the Laboratory, and to the authorities of the United States Bureau of Fisheries
for the facilities of the laboratory.
4 LEON J. COLE
as possible around an axis passing through the center of the
starfish and perpendicular to the surface upon which it was
crawling. The presence of all external stimuli (mechanical,
chemical, thermal, and light) can no more be eliminated than
can those of internal origin, but so long as they are uniform
over the entire animal, or bear to it the radial relation mentioned
above, they need not be considered. :
The earlier experiments were performed in a large circular
glass dish, 36 em. in diameter, 13 cm. deep, and having a flat
bottom. At the beginning of each period of experimentation
this dish was filled with fresh sea water to a depth of about 10
tol5em. It will be seen that at the central point on the bottom
of this dish the requirements stated above regarding mechani-
cal, chemical and thermal conditions must have been very well
met. Light could, of course, have been excluded entirely, ex-
cept for its desirability in enabling the experimenter to observe
the behavior of the subject. Accordingly the dish was illumi-
nated by the light from an incandescent electric bulb at the
ceiling some 6 or 7 feet above, and directly over the center of
the dish. Extraneous light and unsymmetrical reflection were
prevented by means of a cylinder of black cloth, supported on
hoops, and pendant from the ceiling, with the lamp at its center
above, and enclosing the glass dish below. As a further pre-
caution, the experiments were all performed at night, when out-
side the cloth cylinder was darker than within. A small peep
hole permitted observation of the interior, while, by passing
the hand beneath the cloth, the specimen could be manipulated.
At New Haven the experimental conditions were essentially
the same, except that a glass dish of sufficient size not being
available, a small pressed paper wash tub was used in its place,
and the experiments being conducted in a dark-room, the cloth
cylinder was dispensed with.
In the experiments made for determining simply the direc-
tion of locomotion, ten starfishes (designated Nos. 1 to 10) were °
used, with each of which fifty trials were recorded. Owing to
the limitations in size of the receptacles employed, compara-
DIRECTION OF LOCOMOTION OF THE STARFISH i)
tively small specimens (from 5 cm. to 14.5 em. in diameter)
were chosen; but except for this, no selection was made, further
than that care was taken to see that all individuals used were
normally active and that all their arms were normally devel-
oped, specimens with one or more arms decidedly deficient in
size,* or In any way noticeably abnormal, being rejected. The
method of procedure in the trials was as follows. When all
was ready the starfish was picked up and held inverted with
the aboral disc resting on the operator’s finger tips. It was
held in this position usually from a half-minute to two minutes,
or until the arms of the starfish drooped down aborally. The
specimen was then placed oral side down in the center of the
dish of sea water, care being exercised to handle it, in so far as
possible, by the disc alone, in order to avoid the possibility of
an unequal stimulus which might result from handling the ani-
mal by the rays. Furthermore, that any possible unilateral
stimulating effect of the environment might be eliminated, the
specimen was rotated one-fifth of its circumference in each suc-
ceeding trial, the arms a, b, c, et cetera, being turned toward the
observer successively. Between each successive trial it was
removed from the water, being handled in the same way, and
held inverted on the fingers for a period as described above.
It was felt that if the starfish were simply moved back to the
center of the dish each time, this action might not be enough
to break up the action of the impulse under which it was crawl-
ing in a given direction in the previous trial. In other words,
any trial might then be considered merely a continuation of
the previous one. Jennings (’07, p. 141) for example, found
in studying the righting reactions of Asterias forreri that the
animals tended to right themselves in the same way as in the
preceding case, that is, the impulse was retained from a previous
reaction. To what extent the method employed to break up
their impulse was successful in the present experiments will be
discussed later (p. 16).
4Tn the case of specimen No. 3 it was noted that arm c was slightly shorter,
while in No. 5, arm d was noticeably shorter.
6 LEON J. COLE
Ill. METHOD OF LOCOMOTION
Before considering in detail the direction in which the star-
fish crawls it may be well to consider briefly the method of loco-
motion. In the paper to which reference has already frequently
been made Jennings (’07) does much to correct the popular
conception of this process. The important point is that except
on vertical or overhanging surfaces, the tube foot of the star-
fish acts essentially as does the leg of a higher animal, that is,
‘as a lever for swinging or shoving the body forward, not as a
rope for hauling it forward.’’ The sucker functions chiefly in
giving the foot a firm hold and preventing its slipping. As a
result of this use of the tube feet, the locomotion is not a per-
fectly even glide, but may often be seen to consist rather of a
series of very short lunges. Furthermore, these lunges are not
always exactly in the same direction but now a little to this side,
now a trifle to that, resulting in a slight zig-zag movement, so
slight, however, that it is apt to be overlooked without close
observation.
Different starfishes show considerable individuality in’ their
methods of progression. Certain specimens show a tendency
as they crawl, even over a short course, to change the direction
of locomotion, not by a turning of the body, but by a change of
front, so to speak. In an indisturbed individual this change
is usually very small; for example, if it starts out with ray e
in advance, it may change more or less to the interradial area
ea, but the arm e points in the same direction as before. In
the experimental trials, this change of direction was usually so
slight that it was inconsequential; when it was greater, the pre-
ponderating direction was recorded. In only four instances in
the course of these experiments was an undisturbed crawling
starfish, so long as it remained on the bottom of the dish, observed
to change its course radically, or to stop and then go off in a
new direction. In these four cases (specimen No. 38, trials 38
and 47; specimen No. 6, trials 12 and 16) there was no obvious
’ In some cases there may be a slight turning of the animal as well as this
change of direction.
DIRECTION OF LOCOMOTION OF THE STARFISH rf
explanation of the departure from the normal behavior. Cer-
tain specimens showed a tendency always to change their course
in the same direction, while others swayed a little first to one
side and then to the other.
My observations still further confirm those of Jennings (’07,
pp. 97, 115) with regard to the ‘unified impulse’ in locomotion.
When a starfish once establishes progression in a definite direc-
tion, the tube feet of all the rays are extended with reference
to the movement in that general direction, and not according
to their relation to the particular rays.
As to the position of the rays during locomotion, there is
again much variation; two rays may be pushed in advance,
é
b
Fig. 2 Diagram of starfish crawling with two rays in advance and three fol-
lowing. Arrow indicates direction of locomotion.
with three following (fig. 2) or three may be in advance and two
behind (fig. 3). Sometimes the arms are all flattened to the
surface on which the starfish is crawling; at other times, -or in
other individuals, they are curled upward at the tips. In cer-
tain instances it was noted that flexible specimens were more
active than the more rigid ones, and such specimens would go
through fifty successive experimental trials often with no appre-
ciable slowing in the rate of locomotion. The more sluggish
individuals sometimes slowed down towards the end of the series,
and in one case (No. 5) it was necessary to fill out the last five
trials of the fifty on the succeeding day.
8 LEON J. COLE
d
pes
]
C
a
b
Fig. 3 Diagram of starfish crawling with three rays in advance
IV. EXPERIMENTAL RESULTS,
1. Direction of locomotion
In recording the individual trials in the experiments note
was made as to which arm was directed toward the observer
at the beginning of the trial (i.e., the orientation of the speci-
men) and the part of its own body which was in advance as it
crawled. Since the orientation was changed in the successive
trials, the actual direction of movement was of course different
if the animal crawled with the same ray in advance. The orien-
tation at the beginning of the trial was recorded that it might
serve as a check on the uniformity of the environment; but since
the position of original orientation showed no effect in the results
of the experiments, it will not be further considered, and we
need concern ourselves only with the direction of advance with
respect to morphological relations. This was recorded only for
radii and interradii. There were bound of course to be some
cases of doubt; but these were comparatively few, and an error
here would be of little consequence, since it would merely throw
the reading to the next radius or interradius to right or left, and
these would tend to balance each other. That such was actu-
ally the case is.indicated by the evenness of the totals for the
different adjacent positions as shown in table 1, which presents"
DIRECTION OF LOCOMOTION OF THE STARFISH 9
a summary of the results of the whole set of experiments. In
the horizontal rows are recorded the results of trials with each
of the ten individuals. The vertical columns show the number
of trials in which the respective radii and interradii were in
advance. Inspection of the totals shows that on the whole there
was a considerably larger number of trials in which arm e was
in advance than any other arm or interradius. It will be noticed,
furthermore, that there is in general a falling off as one goes in
either direction from this position. This is shown more clearly
TABLE 1
Distribution of crawling trials
|
|
- RADIUS OR INTERRADIUS IN ADVANCE
INDIVIDUAL 7 = Sp SSS TOTAL
| c cd.e |. sd WW ade e ea a ab b be
ia A 2 eal eae MT hits © SR pes 3 1 2 2 | 50
2 5 5 8 2. | eho 4 4 1 "i 4 | 50
3 > Bet she 1 2 6 5 5 7 4 | 50
4 Fed PAD "8 7) is 1 0 3 2 TEN Sp)
ee ee i Weta iy Ae spe fees: 5 1 3 5 | 50
Bee A Bale 4 Sate alll 9 4 5 5 3 | 50
meio. 8 | ee 1 eel Se at | eo | 3 | 8 0
8 | 0 1 8 9 7 7 9 | 1 | 50
Ch ae ame 5 38 6 Abia) ied 6 4 | 6 6 | 50
1 | 8 ii 9 2 Sei G 8 DA ask 3 | 493
Wotal....| 38 | 43>] 64 |. 48 |-85 | 61 | 55 | 35 | 40 | 40 | 499
1 Through oversight only 49 trials were made with specimen No. 10.
in figure 4, where the totals of table 1 are shown plotted in the
dotted line. In this figure the abscissal divisions represent
the radii and interradii from left to right, while each of the
ordinal spaces represents ten trials.
If we consider all the movement in relation to the radii alone,
the preponderating direction of movement stands out some-
what more clearly. This may be done legitimately by appor-
tioning to each of the radii, in addition to its own records, one-
half of those for each of the adjacent interradii to right and left.
The result of such a redistribution is shown in table 2. It is
here easily seen that the preponderating direction in which star-
fish No. 1 crawled was with arm e in advance. The same is true
10 LEON J. COLE
TABLE 2
‘ Crawling trials apportioned to radii
pormus | ——7— a | PREPONDERATING
1 6.0 | 14.0| 20.5) 6.0) 3.0 50 e
2 9.5| 11.5 | 13.0| 6.5| .9.5 50 e
3 6.51 16.0| 5.5| 10.5 | 11.5 50 | d
4 | 14.5] 11.0| 15.5| 2.0] 7.0 50 | e
5 7.5| 1.5 | 23.0|12.0| 6.0 50 | e
6 | 6.5| 6.5|17.0/ 11.0] 9.0 50 e
7. 3.5] 6.5 | 11.5 | 18.5 | 10.0 50 | a
8 1.0 | 13.0| 15.0}14.5| 6.5 50 | e
9 6.5 |13.5| 9.0| 10.0| 11.0] 50 d
10 13.01 13.5| 7.0| 12.0| 3.5| 49 d
77.5 499.0
Total..........| 74.5 |107.0 137.0 103.0
for Specimen 2; for No. 3 it is arm d; and so forth, as indicated
in the last column to the right in the table. This column shows
that in six of the ten individuals used, the preponderating direc-
tion of crawling was toward arm e; in three it was toward arm
d; while in one arm a was most often in advance. In no case
was the preponderating movement toward either b or c. This
tendency to move in the direction of arm e is plainly indicated
when the records for the ten individuals are added, as shown in
the bottom row of the table. The same fact is represented
graphically by the solid line in figure 4, where these totals are
plotted; the ordinal spaces in this case being given a valuation
of twenty trials in order to make the two curves comparable.
The ‘mode,’ as before, is clearly at e; but in this case. the ‘curve’
drops off smoothly and with remarkable symmetry on the two
sides.
In figure 5 these valuations are referred to a diagram of a
starfish, where the symmetry of the ‘curve’ in figure 4 shows
Fig.4 Plotting of number of times each radius and interradius was in advance
in the crawling experiments. Dotted line from data given in table 1; solid line
from data of table 2.
Fig. 5 Diagram of starfish, showing the number of trials accredited to each
arm as ‘director’ (ef. fig. 4 and table 2). AA, line of bilateral symmetry with
respect to the records; an approximately equal number of trials fall to each side.
The line BB has 347 records ‘ahead’ of it and only 152 ‘behind.’
DIRECTION OF LOCOMOTION
5 181% 137
OF THE STARFISH
180%
12 LEON J. COLE
an interesting relationship to the animal, for a ine AA drawn
through arm e and the interradial area be divides the crawling
records almost equally into halves; 137 of the 499 records le
on the line, while of the remaining 362 trials, 181.5 are on the
left and 180.5 on the right. Or if a line BB, at right angles
to AA be drawn through the center of the disc, 347 records
are in a direction ‘ahead’ of this line, while only 152 are
‘behind’ it.®
From the foregoing it may be concluded that, although the
starfish Asterias forbesi may move with any ray in advance, in
the absence of directive stimuli, as shown by a large number of
trials, it was most often the one lying next to the left of the madre-
poric plate which went ahead. Using direction of movement as
a criterion, this may then perhaps be considered the ‘physiologi-
cal anterior’ of the animal.
If we turn again now to the experiments of Jennings, it will
be recalled that in his study of the righting reactions he found
the ray e used most often, namely 89 times out of 95, ‘‘and the
next greatest numbers [were| shown by the two rays lying on
either side of e, namely a (56) and d (43).’’ His results are thus
directly comparable to those on the direction of movement,
and in each case a greater activity of the ray e and the rays
on either side of it is probably indicated. A somewhat similar
determination of direction of movement dependent upon a dif-
ferential activity of the organs of locomotion has been demon-
strated by the writer (Cole ’01) in the pyenogonid, Anoplodactylus
-lentus, in which the legs assume an essentially radial position.
2. Relation of direction of locomotion to length of arm
A possible explanation of the greater activity of the ‘anterior’
rays (if we may so call them) is suggested by the observation
mentioned in the footnote on page 5 that in some of the speci-
mens used the arms b and ¢ (the ‘posterior’ arms) were notice-
ably shorter. Unfortunately the actual specimens used in the
* If the rays had been used indifferently, approximately 100 records would be
expected for each ray.
DIRECTION OF LOCOMOTION OF THE STARFISH 13
experiments were not accurately measured nor preserved, but
in order to ascertain whether perhaps this condition of shorter
‘posterior’ arms held generally, though in a less striking degree
(since otherwise it would have been noted) careful measurements
of the length of arm have been made on a series of 116 specimens
of Asterias forbesi from the Woods Hole region, and of about
the size of those used in the experiments. A brief statement
of the results of these measurements will suffice for the present
discussion. The specimens measured were a selected sample
only in that individuals about the size of those used in the exper-
iments were included and that those having arms of obviously
disproportionate lengths (probably regenerating arms) were ex-
cluded. This was done also in selecting individuals for the
experiments. Nevertheless, in measuring this somewhat re-
stricted lot, one was most impressed by the considerable varia-
tion in arm length and its apparent irregularity as to position.
It was quickly discernible that no one arm, nor pair of arms,
was regularly the longest or the shortest. The number of times
each arm occurred as the longest or as one of the two longest
(in cases where the longest two measured the same) and as the
shortest or one of the two shortest, was as follows:
NUMBER OF TIMES
ARM
a b c d €
Longest or one of two longest. . 53 41 25 22 36
Shortest or one of two shortest. . 30 28 40 41 34
These figures correspond in a general way to the results on
locomotion, and in so far would appear to show a correlation
between length of arm and the frequency with which it moves
in advance. Thus it will be observed that arms c and d are
longest the least number of times, while arm a is most often
longest (or one of the two longest). Conversely arms ¢ and d
are the shortest more often than the others, and arms e and a
are shortest the least often. This may be seen most readily
where the figures are plotted as in figure 6, which should be
compared with figure 4. It will be noted that except for the
14 LEON J. COLE
Fig. 6 Plotting of number of times in 116 specimens in which a particular
arm occurred as the longest or one of the two longest (solid line) or as the short-
est or one of the two shortest (dotted line).
fact that the mode falls at a instead of e, the curve for the ‘long-
est’ arm, represented by the unbroken line in figure 6, corres-
ponds closely to the curve for direction of locomotion in figure
4, while in a general way the curve for the ‘shortest’ arm (dotted
line) is the converse of these.
If now we consider the mean lengths of the respective arms
of the 116 specimens, although the differences are very small,
and perhaps insignificant, nevertheless we find that their dis-
tribution corresponds to that just given, namely, e and a have
the greatest mean length, cand d the shortest. The exact fig-
ures are:
;
|
ARM c | d e a b
Mean length in millimeters?....) 46.30 | 46 .23 | 46.91 46.93 46.74
7 Later computation of the probable error of the differences between the mean
lengths of the respective arms as here determined makes the value of these differ-
ences extremely doubtful. The striking fact is that, considering the often very
DIRECTION OF LOCOMOTION OF THE STARFISH 15
As noted above, however, the axis which marks the approach
to bilaterality in these determinations (AA, fig. 7) does not accord
~ exactly with that with respect to direction of locomotion (AA,
fig. 5); for whereas in that case it passed through arm e and
between arms b and c¢, here it obviously passes through arm a
and the interradius cd.
(46.91)
36
34
(46.23)
(46.30) (46.74)
25 . 41
GO 28
Fig. 7 Diagram of starfish, showing number of individuals in 116 specimens
in which a particular arm was longest or one of the two longest (black face fig-
ures) and in which it was shortest or one of the two shortest (italic figures). Com-
pare with figure 6. Figures in parentheses are the mean lengths in millimeters
of the respective arms of the 116 specimens as determined. AA, line of sym-
metry with respect to these figures.
Considering the considerable amount but irregular distribu-
tion of the variation in arm length and the very small difference
in mean length of the different arms, as well as the difference
marked difference in arm length of a given individual, the mean arm lengths of
the 116 specimens should be practically equal. The greater individual varia-
tion giveS somewhat greater value to the tabulation on p. 13 of the number of
times a particular arm occurred as the longest or shortest than to the mean lengths.
But even if these figures could be relied upon the fact would still need to be con-
sidered that they do not refer to the actual specimens used in the experiments.
16 LEON J. COLE
in the axes of bilaterality, it seems very doubtful whether the
tendency to bilaterality in this respect can have significance in
accounting for the noticeable bilaterality with respect to loco-
motion, in spite of the fact that the evidence seems to show that
on the whole there is a slight tendency for arms e and a to be
longer than the others. One is inclined to believe rather that
both these results may be a more or less complete expression of
factors or tendencies of development or organization.
3. Persistence of the impulse
Reference has already been made (p. 5) to the ‘impulse’ of
a starfish to move in a certain direction, once it has estab-
lished movement in that direction—a sort of momentum of
physiological reaction which causes a certain behavior to persist
for a time even against an adverse stimulus (Jennings ’07, p.
115). This fact has more recently been confirmed by Cowles
(11, p. 103), who remarks:
A characteristic of the starfish is that when once the impulse to
move in a certain direction is formed, the starfish is quite persistent
in its behavior and continues to move in that direction; so when the
creature reaches the wall it ascends owing to the persistence of the
impulse
This inherent stubbornness of Asterias forbesi may be seen
in table 3, which displays in detail the results of the whole series
of trials made in the present experiments, in the order of their
sequence. Here it is evident that the precaution taken to break
up the locomotive impulse (ef. p. 5) was inefficient, for although
there is some variation in the part advanced in successive trials,
in general the same one remains the temporary ‘anterior’ for
a considerable number of trials. Specimen No. 4 may be taken
as an example. In this case the subject first crawled with the
interradial area ea as ‘anterior,’ then followed seven trials with
e in advance, next one with de, one with d, another with de,
three with e, again de, three more e, and so on. The degree of
this persistence is shown by the records better than it can be
described.
Deiail of ci
vidual No. 1 No. 2 No. 3 No. 4
Inter-radius ¢ cd d dee caaabbbe ccdd dee caaabh be ccd d dee caa abb bejc cd d dee ea a ab
wn ON TFL
10
gO
1 Period of 3 to 5 minutes interval between trials.
2 Turned on back for period of about 2 minutes between trials.
3 Changed direction abruptly during trial; went ea first for some distance, then d, :
4Started a and changed to b.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
Wah eh id . ble naweted barrie.
one pe oni abntetiod oles? hand ree,
> eek aby aaior 01 Beall os tiww jp fared gi
: ge hal
wr, . L cor 0d gti tel gal aaa
DIRECTION OF LOCOMOTION OF THE STARFISH 17
4. Rotation of the impulse
Inspection of table 3 reveals another interesting relation exist-
ing in the trial records. This is what may be called a tendency
to ‘rotation of the impulse.’ Not only is there a slight vacilla-
tion in the direction of the impulse, but there appears to be a
well-marked tendency for the impulse as a whole to shift gradu-
ally around one way or the other. This was more pronounced
in some individuals than in others. It is especially marked
in the records of specimens No. 4, No. 5, No. 7, No. 8 and No.
9. With the records spread out flat as they are here the records
of successive trials tend to trail out diagonally across the space
allotted to the individual, as is indicated by the diagonal dotted
lines enclosing them. But to express the proper relationships
they should be plotted on a cylinder (for on the animal ¢ and
be are adjoining) and the tendency of the impulse to rotate in
one direction or the other would then be expressed by the spiral
path of the records around the cylinder, to right or left as the
case might be. This would bring the figures lying on each side
of the flat diagrams together and make the result appear more
striking.
In the case of five of the starfishes (Nos. 1, 3, 6, 9 and 10)
the direction of rotation of the impulse around the animal is
plainly to the right, or clockwise; in four (Nos. 4, 5, 7 and 8)
it is as evidently to the left, or counter clockwise. In only one
specimen (No. 2) is there definitely shown a change in the direc-
tion of this rotation during the recording of the fifty trials. This
individual was apparently right-handed in this respect at first,
and changed to left-handed after about the fifteenth trial. There
is nothing recorded in the notes which would appear to furnish
an explanation of this change.
It will be observed that the width of the ‘paths’ between the
dotted lines in the table is in a degree a measure of the ‘inten-
sity,’ if we may so call it, of the impulse; in some cases the im-
pulse remained so nearly fixed in a particular region that trial
after trial the starfish crawled with the same ray forward. Take
for example, specimens Nos. 3, 4, 5, 7 and 8 (after about trial
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
18 LEON J. COLE
19). On the other hand, specimen No. 6 has a broad ‘path,’
indicating that although the impulse was directed toward the
same general region it vacillated considerably from side to side.
It was noted that specimens Nos. 4 and 5 were relatively stiff
or rigid and comparatively inactive, while No. 6 was flexible
and moved at a good rate throughout the fifty trials. This
fact may or may not be of significance.
The slant of the ‘path’ is similarly a measure of the ‘rotation’
of the impulse as a whole. This was not correlated in any defi-
nite way with the condition of the starfish, so far as could be
observed, but during the course of the experiments the impres-
sion was gained that an individual was often comparatively
inactive during the first few trials—that it had to be put over
the course a few times before it became ‘waked up to the work,’
so to speak—and in a number of cases at least it slowed down
toward the end of the experiments as if becoming fatigued® (for
example, see No. 5, which refused to crawl after the forty-fourth
trial). In this connection it may be pointed out that a consider-
able number of the cases (Nos. 2, 5, 6, 7, 8, 9 and 10) seem to
show a tendency to start at the sides of the diagrams (in the
regions or arms b and c), to cross the space diagonally, and to
end again at the opposite side below. If the impression stated
above is true, then this means that in general when the experi-
ments were started with the various specimens they were com-
paratively inactive and tended to crawl ‘backward’ (if we admit
that a physiological ‘anterior’ has been established), 1.e., with
the general region be in advance; gradually the impulse swings
around until they are crawling ‘forward; and finally as they
become fatigued, it moves on around to the original position.
This matter should have been tested by continuing the trials
much further, but it was not realized at the time, and there
appeared no obvious reason why fifty trials with each individual
were not sufficient.
Whether or not the above relations prove to be general, never-
theless the records do seem to establish beyond a doubt the
* Glaser (’07, p. 206) mentions changes from active to sluggish behavior and
the reverse, in Ophiura, for which there was no obvious explanation.
DIRECTION OF LOCOMOTION OF THE STARFISH 19
gradual rotation of the impulse. In seeking an explanation foi
this two possible ones suggest themselves, of which one is physi-
ological in the strict sense, the other psychological, or based on
behavior. According to the first of these, the change in direc-
tion may result from fatigue (whether of muscles or controlling
centers) which results from activities being continuously directed
one way, and as a result they shift to a new position, somewhat
as the control of pulsation in a medusa passes from one center
_ to another around the periphery. On the other view, this would
probably be a type of modifiability of behavior analagous to
that of Stentor and other lower organisms, which, as Jennings
(06, chapter 10) and others have pointed out, change their mode
of reaction when one mode has been tried for a time and proves
ineffectual. In this case the impulse in any one direction is
ineffectual for the starfish because it fails to ‘get anywhere,’
and the impulse accordingly shifts to a new position, just as
Stentor may bend first one way and then another to avoid an
unfavorable stimulus. This ‘regular’ rotation of the impulse
would probably not take place under any except uniform con-
ditions; if the starfish were to come to an uneven surface or to
food, these new stimuli would undoubtedly influence the further
direction of the impulse. Like most cases of modifiability of
behavior, this behavior of the starfish may be adaptive in nat-
ure—at least it would prevent its travelling continually in a
straight line.’ Indeed, if we could imagine a starfish on a stretch
of level, smooth sea-bottom where directive stimuli were absent,
it would, according to these results, tend to swing around in a
large circle and come to rest near where it started out.
5. Relation of arms used in righting to direction of locomotion
In table 4 are shown the results of a number of tests to deter-
mine what relation exists between the arms used in righting
when the starfish is placed on its aboral surface and the direction
of locomotion previous to and subsequent to the righting reac-
tion. The data may be summarized as follows:
° As has been mentioned (p. 6) the straight course is also frequently swerved
from somewhat, even though the same ray remains in the lead.
20
LEON J. COLE
ARMS
cd a|de| e ea| a ab| b | be
Crawling previous to test
Arm or arms used in righting.................
Crawled subsequent to righting......,........
5} 1
Dae
2 16
| 3 2
2
|
=
| !
es | 2
| |3
This shows that whereas the four specimens used in these
tests righted themselves on arms ea sixteen out of twenty-four
times, previous to the trials they had been in nearly all cases
crawling in a direction nearly opposed to these arms, and more-
over, they continued locomotion in the same general direction
An examination of the individual
records reveals the same relations in the great majority of cases.
after righting themselves.
TABLE 4
Relation of arms used in righting to direction of previous and subsequent crawling
pore reser uei jit obo sal Ea aaa
No. 9—after trial 50............ d ea
No. 10—before trial.1........... ea Cc
No. 10—after trial 10........... a e(b) cd
No. 10—after trial 16 ........... c e(ab) be
No. 10—after trial 28 ........... cd ea cd
No. 10—following day........... cd be
No. 10—following day........... be be a
No. 10—following day...........| a cd d
OL 2— Crisler Se ee ee be a
Bont trial Sch). Cee a ea de
INGOs oti ai hh raodeioe eee de ea cd
Jairo 7 hg C1: i ede cd ab Cc
INOS 12 —OITAL'D. res se cl en c ea cd
INO IA — tral Uses bc ee ea d
Gr 1 SS hats Se d ea d
Unt aj] a: 4S inca | d ea cd
INO. WA ——“OFlE @. occ cols const ace cd ea cd
NO-t4——prigled. 8, a Hee cd ea cd
IN Oa TETAS Ooi he dns nieve bocce cd ea d
i Kopp eo Aa hy ie d ea d
MWe AOL oo iy cae ca xv ee aie d ea cd
iy Co ea EE) ot ee cd cd —!
Wow 14—trial 10. .2.7.4.<.!. ea be
No. 14—+trial 11............ be ea cd
1 Remained quiet 12 minutes after righting itself.
DIRECTION OF LOCOMOTION OF THE STARFISH 21
The subsequent locomotion in a direction opposite to the
arms which have been used in righting is easier to understand
in many cases than why the starfish should use those rays in
turning over. But even here there is difficulty in reconciling
the direction of the co-ordination in righting with that of the
subsequent locomotion. Let us-suppose for example a starfish
placed on its ‘back,’ that is, oral side up. Let us suppose, fur-
thermore, that it has already succeeded in twisting rays e and
*. a and attaching them to the substratum, and that a co-ordinated
impulse toward righting on these arms has become established.
The tube feet of all the rays would accordingly be extended in
that direction, as indicated in the accompanying diagram (fig.
8 a). In some cases these two rays will swing all the others
over freely, in which case it would be expected that the impulse
would be still towards ae, and that the animal would crawl in
that direction. It much more frequently happens, however,
that one or more of the other arms become attached before the
animal is completely righted. Thus in figure 8 }, e and a are
represented as still attached; c has swung over and attached at
its tip between e and a; d has twisted and attached before cross-
ing over, and 6 is free. Ray c now becomes the dominating
factor in pulling the other arms over. To do this its tube feet
must action by pulling toward the tip of the ray, whereas previous
to its attachment they were directed towards its base (cf. fig.
8a). Arm ), being free, is now pulled over into position between
a and c; but as d is attached it. remains crossed over e, even after
the starfish is completely righted. It is now, however, swung
over, with the help of the movement of the whole animal c-ward,
and locomotion now continues in that direction. It will thus
be seen that between the time the creature is in the position
shown in figure 8 a, and when it is completely over and moving
in a general direction opposite to that of the rays on which it
turned, there must be a reversal in the direction of the co-or-
dinated impulse of the tube feet. The details of this were not
studied as carefully as they should have been owing to the fact
that the data on the righting and locomotion were not tabulated
until there was no opportunity to continue the experiments.
22 LEON J. COLE
No simple explanation appears to offer itself as to why the
converse of the above should be true, namely that a specimen
which had been crawling c-ward should, when placed on its
8b
Fig. 8 Diagrams of starfish righting itself. Arrow indicates direction of
turning; note that all tube feet are extended in that direction. Arms attached
to substratum are indicated by X.
aboral surface, use more often the rays opposite (e and a) in
righting itself. This is a point deserving further study.
At best the results in this connection seem rather paradoxical.
Jennings found that his starfishes used ray e more often than
DIRECTION OF LOCOMOTION OF THE STARFISH 23
any other in righting themselves; in the present experiments it
was found that Asterias forbesi moves more often with ray e
in advance than with any other. Furthermore it was found in
the comparatively few experiments which were made with respect
to righting that the East Coast starfish also used rays e and a
most frequently for this purpose, but these individuals, in the
great majority of cases, then crawled with the opposite rays in
advance. Which way would Jennings’ starfish have crawled?
V. COMPARISON WITH OTHER ECHINODERMS
It would be interesting to know whether the tendency for a
predominance of movement to be with a particular region of the
body forward obtains in other’starfishes and in the other groups
of (practically) radial echinoderms. The statement is not un-
commonly made categorically by authors, based only on general
observations, that no preference is shown for a particular ray or
region. Thus Cowles states in a preliminary report (’09, p.
128): ‘‘Echinaster in locomotion does not show any tendency
to use a special ray, or pair of rays, as directors.’’ He admits
later, however, that although this conclusion was “‘based on a
considerable number of observations with directive light ex-
cluded,” it was ‘“‘not a careful statistical study’? (Cowles ’11,
p. 98). Grave (’00, p. 86) remarks similarly: “‘No preference
as to which arm should precede could be found in an adult ophi-
uran, each arm being equally capable of going before, making
the stroke, or following behind.’”’ But while each arm might
be ‘equally capable’ of performing these actions, it does not fol-
low that they would be used equally for that purpose. Bohn
(08, p. 29) says regarding the relative value of the arms in
Asterias rubens that commonly they are the same functionally,
but if any are shorter or are mutilated they have a relatively
smaller value. Further on in the same paper (p. 43) he asserts
that the small specimens of this species (rays 1 cm. in length),
which are found on the Zostrea, and which are very active and
in general show a positive phototropism, use all their arms with
practical indifference. Larger specimens, however, sometimes
24 LEON J. COLE
appear to evince ‘‘a sort of preference for certain arms;”’ never-
theless the general conclusion appears to be that no ray has
especial value as a director. This, like many of Bohn’s con-
clusions, one wishes might be based on a more extended series
of observations.
E. C. and A. Agassiz (’65) state that ‘‘Cribrella [Cribrella
oculata = Henricia sanguinolenta] moves usually with two of
the arms turned backward, and the three others advanced to-
gether, the two posterior ones being sometimes brought so close
to each other as to touch for their whole length.” They make
no statement, however, as to whether it is usually the same
arms which take these positions.
1. Preyer’s experiments on starfishes
Preyer (’86-7) appears to be the only person who has pre-
viously investigated this question in anything like a statistical
way, and although he concluded negatively as to the preferen-
tial use of any particular arm or arms as directors, his results
are worthy of examination in some detail. His method of experi-
mentation was careful and ingenious, and one feels inclined to
place confidence in what he speaks of as these ‘sehr zeitraubenden
Versuchsreihen.’ He laid the animals on the top of a glass
support, which was hemispherical above, and so arranged that
while the greater parts of the rays of the specimen were immersed
in the water, their basal part and the central disc were exposed
to the air. The support was filled with water except for an air
bubble at its top, by means of which it could be accurately
leveled. Finally the specimens were oriented with a given ray
to the north, east, south and west in successive trials, in order
to neutralize the possible interference of any inequalities in the
surroundings. The starfishes and brittle stars used naturally
tended to crawl down one side of the glass support so as to im-
merse themselves completely in the water, and Preyer recorded
what ray was in advance in this movement. He numbered the
rays 1, 2, 3, 4 and 5, beginning with the one which has themadre-
porite at the left of its base and counting around clockwise,
these numbers therefore corresponding respectively to rays a,
DIRECTION OF LOCOMOTION OF THE STARFISH pA
b, c, d, and e, as denominated by Jennings. In case two adja-
cent rays took an equal part in the lead, ‘one-half time’ was ac-
credited to each.t° Except for Luidia, which is a seven-rayed form,
Preyer’s (’86—7, p. 218) results are brought together in table 5.
Preyer states that the brittle-stars Ophiomyxa and Ophio-
derma showed just as little preference for any particular ray;
but a careful inspection of table 5 will show that this conclu-
sion regarding the starfishes was scarcely justified by the results
presented. In fact these results show a striking similarity to
TABLE 5
Summary of Preyer’s results on the direction of locomotion
5 NUMBER OF TIMES EACH ARM WAS | NUMBER OF
SPECIES fe eres ope SRS SEE TOTAL | cane IF
TESTED 3 USE WERE
a(1) | 6 (2) c(3) | d (4) | e (5) INDIFFERENT
; pe | sae
Astropecten
bispinosus 6 16.5} 14.5] 4.0; 9.5] 9.5| 54.0] 10to 11
Astropecten |
pentacanthus + 7.5} 3.5) 7.0) 14.0] 18.0/| 50.0} 10
Astropecten |
aurantiacus.. 3 18.5; 8.0] 9.0} 12.5) 8.0] 56.0; 11 to 12
Asterias
glacialis..... 1 5.0| 4.51) 5.0] 4.0] 5.5] 24.0 | 4to 5
3] j | a fi
a 14 47.5 | 30.5 | 25.0 | 40.0 41.0 184.0 | 36 to 37
those presented in the earlier part of this paper. Not only is
this apparent in the individual records for each species (note
the small number of times c was used as compared with the
ray opposite it), but is especially noticeable when comparison
is made of the totals (cf. table 5 with table 2, p. 10). In the
case of Preyer’s results, however, the plane of bilaterality would
not pass through the radius e; but the plane which would have
most nearly an equal number of records on each side, would
cut through the interradius ea and ray c. Such a plane has
eighty-one records to the left, seventy-eight to the right. So
if we thus lump together then the four species of starfishes with
10 This accords with the treatment of the data in the present paper (cf. table
1, and fig. 4).
26 LEON J. COLE
which Preyer worked (which seems permissible considering the
essential similarity of the records) the ‘physiological anterior,’
as determined by the direction of locomotion, is the interradius
ed.
Finally, if we combine Preyer’s results with those obtained
on Asterias forbesi, making a total of 683 trials with five differ-
ent species of starfish, the result is as follows:
ES d e a b
Preyer s fesulés..-,-.-2s-.--2ean.| . 254) 40.0 | 41.0 | 47.5 30.5
if
Aaterias forbesh sca: hess n.d hese 4.6 107.0 | 187:0 | 103.0 dgse
Rotaliekst pee ee ee) “9g el sree ete aee 150.5 | 108.0
Here it will be seen that the plane of bilaterality passes through
e, which is by this token the ‘physiological anterior.’
2. Echinoids and other echinoderms
Aside from the few experiments of Preyer on ophiurans, and
which he mentions only in a general way as giving negative
results (see p. 25), no effort appears to have been made to
determine definitely whether preference is given to a particular
ray or radius in the other groups of echinoderms outside the
asteroids. Grave has already been quoted (p. 23) as saying
that in Ophiura ‘each arm is equally capable’ of acting as direc-
tor, and this appears to be the general impression.
What has just been said of the ophiurans may apparently be
applied equally well to the radial sea-urchins—that is, that
they may crawl with any part of their circumference in advance.
There are, however, certain echinoids which have assumed a
secondary bilateral symmetry, and are accordingly especially
adapted to locomotion in one particular direction. These are
the spatangoids, in some of which the spines all point back-
wards, imparting to the creatures somewhat the appearance of
a hedgehog. The anus, furthermore, is commonly shifted back-
DIRECTION OF LOCOMOTION OF THE STARFISH 27
ward from the dorsal position until it comes to lie on the pos-
terior border of the periphery, while conversely, the mouth may
be located considerably forward of the mid-ventral position.
Now it is of considerable interest to note that the region which
is anterior is an ambulacrum, corresponding to a ray of the
starfish, and that the madreporite always lies in the inter-ambu-
lacral area next to the right of the anterior ambulacrum. If, there-
fore, we designate the different areas with relation to the madre-
porite as we have done in the starfish, it will be observed that
the anterior ambulacrum ts e (III)! and this is the ray which was
found to function as the “physiological anterior’ in Asterias. These
radii in the two cases are accordingly clearly analogous physio-
logically, but whether they are morphologically homologous, it
would be hazardous to state, since the complications of embryo-
logical development make this point practically impossible
to ‘determine. The physiological relationship may be readily
understood by comparing figure 9, which is a diagrammatic
representation of a spatangoid, with figure 5 (p. 11) which
similarly represents the starfish.
In holothurians the morphological and axial relations are so
different from those of the forms we have been discussing that
there would be little of value in a comparison. It is of interest
to note, however, that Pearse (’08, p. 269) found in Thyone
a preferential use of certain tentacles. As to locomotion
however, he states (p. 264) that ‘the animal may move in any
direction,’ and (p. 266) that
Individuals move with the posterior end in advance as often as with
the anterior end, and although the long axis of the body is as a rule
approximately parallel with the direction of locomotion, animals often
move a long distance (as much as 12 cm.) with the body at right angles
to the direction of movement, that is, they move straight toward the
right or left.
11 The anus consequently lies in the inter-ambulacral area (5) corresponding
to the inter-radius be of the starfish. In certain of the radial sea-urchins (e. g.
Strongylocentrotus) the anus occupies an eccentric position in the periproct,
lying nearer to the border opposite interambulacrum 5 (bc). A line drawn through
ambulacrum 111 (e) and interambulacrum 5 (bc) therefore marks the beginning
of a bilateral symmetry and presages the condition found in the spatangoids.
28 LEON J. COLE
He apparently made no test, however, of the proportion of
locomotion forward and backward in the absence of directive
stimuli.
III (¢)
A
Fig. 9 Diagram of aspatangoid. 1—yv, Lovén’s designation of the ambulacral
areas; (a)—(e), corresponding arms of the starfish with respect to the position of
the madreporic plate, as designated by Jennings (cf. fig. 1); i—4, interambulacral
areas, according to Lovén. an., anus; mo., mouth; m.p., madreporic plate;
AA, line marking bilateral symmetry of form. This diagram should be com-
pared with fig. 5.
VI. CONCLUDING DISCUSSION
Granting that the observations which have been presented
establish satisfactorily the fact of a ‘physiological anterior’ in
the starfish—that a particular part tends to precede most often
in locomotion—three possible explanations to account for this
phenomenon suggest themselves. First, there may. be a defi-
nite morphological relation between the bilateral larva and the
adult which establishes what shall be ‘anterior’ in the latter;
second, it may depend upon a proportional relationship in the
length of the rays; and third, it may be related to the condition
of some other set of organs, such as the nervous system or the
water vascular system.
DIRECTION OF LOCOMOTION OF THE STARFISH 29
In considering the first of these theories we are at once con-
fronted by the great complications which take place in the meta-
morphosis of the starfish, and which render extremely difficult
the correlation of planes and the orientation of parts in the larva
and in the adult. It is true that a number of years ago Goto
(98, p. 241) believed he had proven “‘a direct connection between
the two principal planes in question, sagittal of brachiolaria
and of adult,’ and “‘that this connection is that of exact coinci-
dence.”’ He believed his studies made it clear (p. 242) that
The sagittal plane of the larva euts the disc of the star at right angles
and passes through the water-pore and the centre of the disc, that is
to say, the sagittal plane of the larva and the plane of bilateral symme-
try of the star are coincident. The arms of the starfish may therefore
be justly spoken of as the median ventral, the right and left dorsal,
and the right and left ventral, arms. It need hardly be added that
the oralside is anterior, and the aboral side posterior
as they are in the holothurian.
It is clearly evident that if this simple relationship were true
it would serve nicely to explain the results on locomotion, since
the plane of bilaterality with respect to direction of crawling
practically coincides with that which he believes separates the
symmetrical halves of the brachiolarian larva and the adult,
at least if we consider the physiological ‘anterior’ to be in the
general direction of arms e and a rather than exactly through
a. Later researches have, however, apparently failed to corro-
borate Goto’s conclusions, the complications of metamorphosis
being much more. difficult to unravel than would appear from
his description, so that we are probably not justified in accepting
this as an explanation for the preponderance of locomotion in
one direction.
Similarly, the measurements which have been presented, al-
though they appear to show a very slight correlation between
mean length of arm and direction of locomotion, would seem to
indicate that this explanation too must be rejected. Further-
more, if such a correlation actually exists, it will not serve as
a satisfactory explanation, for we should still have to account
for the proportional relationship of the arms.”
12 Goto’s contention, if true, would of course satisfactorily account for this.
30 LEON J. COLE
The third suggestion is that the observed facts regarding
locomotion may be explained by some peculiarity of the nervous
or water vascular systems. As to the former of these, the ner-
vous system appears to offer no peculiarity which could account
for a differential in the locomotion. On the other hand, in the
water vascular system we do find a special structure which
breaks up the radial symmetry and which bears a positional
relationship to the physiological differentiation observed in the
locomotion. This structure is the madreporite, which connects
the water vascular system with the exterior.
It will be recalled that whereas the experimental results esta-
lished for Asterias forbesi a plane of physiological bilaterality
through arm e and between 6 and c¢ (ef. fig. 5), arm a was found
to be most frequently the longest or one of the two longest (fig.
6 and table, p. 13), and also had the greatest mean length.
These facts were shown in figure 7, which shows the axis passing
through arm a and between arms c and d. Preyer’s experi-
ments on other starfishes, on the other hand, indicated a plane
passing between e and a (thus intersecting the madreporite)
and through arm c¢ (p. 26). The significant fact appears to
be that all these data seem to indicate a plane passing through
or near the madreporic plate. Proximity to the madreporite
seems then to be associated in some way with a greater activity
or other functional superiority of the tube feet on the arms so
situated, for such is probably the reason for the preponderance
of locomotion in this direction. Just why this should be so is
not so clear, though it may be purely mechanical; this portion
of the ambulacral system may be able to adjust itself more read-
ily to changes in volume of contained water, consequent upon
the activity of the ampullae and tube feet, because of its near-
ness to the outside reserve, that is, the outside water surrounding
the starfish, just as a better supply is had near the mains of any
water system than in the more remote branches."
'® It would be interesting to determine whether there is possibly a difference
in caliber of the radial tubes in different arms, and whether this is related
to their nearness in origin to the stone canal.
DIRECTION OF LOCOMOTION OF THE STARFISH 31
VII. SUMMARY
The principal points brought out in this paper may be sum-
marized as follows:
1. Experiments indicate that the starfish (Asterias forbesi,
and probably other species), in the absence of directive stimuli,
does crawl more frequently with a particular part of the body
in advance, namely the part in proximity to the madreporite.
This demonstrates a ‘physiological anterior,’ and a plane passed
through the madreporite or one of the adjacent rays divides
the animal into symmetrical physiological halves.
2. The correlation between direction of locomotion and mean
arm length is doubtful and the results obtained are probably
not significant.
3. A definite ‘impulse’ is established on account of which the
starfish tends to crawl in the same general direction in succes-
sive trials.
4. There is a tendency for this impulse to shift or ‘rotate’
gradually around the body in one direction or the other.
5. After righting itself, the starfish more often crawls in the
general direction of the rays opposite to those which it has used
primarily in righting.
6. Preyer’s results from his experiments fall substantially in
line with those on Asterias forbesi.
7. The ‘physiological anterior’ of the starfish corresponds to
anterior in the spatangoids, with respect to the position of the
madreporite.
8. The position of the madreporite may perhaps be what
determines ‘anterior,’ and it is possible that this may be from
purely mechanical causes.
LITERATURE CITED
Agassiz, ELizABETH C., AND ALEXANDER 1865 Seaside studies in natural his-
tory. 157 pp.
Boun, G. 1908 Introduction A la psychologie des animaux 4 symétrie rayonnée.
Deuxieme Mémoire. Les essais et erreurs chex les Etoiles de mer et
les Ophiures. Bul. Inst. génér. psychol., 8e année, pp. 21-102.
Cote, L. J. 1901 Notes on the habits of pyenogonids. Biol. Bull., vol. 2,
no. 5, pp. 195-207.
32 LEON J. COLE
1910 Direction of locomotion of the starfish (Asterias Bios Sci-
ence. N. S., vol. 31, no. 795, p. 474.
Cowtes, R. P. 1909 Preliminary report on the behavior of echinoderms. In
annual report of the Director of the Department of Marine Biology,
year book of the Carnegie Institution of Washington, no. 8, pp. 128,
129.
1911 Reaction to light and other points in the behavior of the star-
fish. Papers from Tortugas Lab. Carnegie Inst. Wash., vol. 3, pp.
95-110.
Guaser, O. C. 1907 Movement and problem solving in Ophiura brevispina.
Jour. Exp. Zodél., vol. 4, no. 2, pp. 201-220.
Goto, Serraro 1898 The metamorphosis of Asterias pallida, with special
reference to the fate of the body cavities. Journ. Coll. Sci., Tokyo,
vol. 10, pt. 3, pp. 239-278, pls. 19-24.
GRAVE, CaswELL 1900 Ophiura brevispina. Mem. Nat. Acad. Sci., 1900, pp.
77-100, pls. 1-8. Mem. Biol. Lab. Johns Hopkins Univ., vol. 4, no. 5.
Jennincs, H.S. 1906 Behavior of the lower organisms. The Macmillan Com-
pany, xvi + 366 pp.
1907 Behavior of the starfish, Asterias forreri de Loriol. Univ. Calif.
Publ., Zoél., vol. 4, no. 2, pp. 53-185.
Pearse, A. 8S. 1908 Observations on the behavior of the holothurian, Thyone
briareus (Leseur). Biol. Bull., vol. 15, no. 6, pp. 259-288.
Preyer, W. 1886-7 Uber die Bewegungen der Seesterne. Mitth. a. d. zool.
Stat. zu. Neapel, Bd. 7, 1. Heft, pp. 27-127, 2. Heft, pp. 191-233,
Tal. 7
-ODTTIONAL DATA FOR PHN STUDY OF SEX-
LINKED INHERITANCE IN DROSOPHILA
T. H. MORGAN anp ELETH CATTELL
From the Department of Zoélogy, Columbia University
In a paper dealing with Data for the Study of Sex-linked
Inheritance in Drosophila (Jour. Exp. Zo6l., vol. 13, no. 1, 1912)
we described seven crosses in which three pairs of the sex-linked
factors were involved. Three crosses that belonged to the same
series were withdrawn because, as stated, the results were anom-
alous in certain points. It seemed almost certain that an error
had crept in somewhere. The new results show, in fact, that
these crosses are consistent with the other results concerning eye
color, body color and wing characters. The new data, added
to those of our former paper, to those of Morgan’s paper for
1911, and to those of Dexter’s paper that has just appeared,
give numbers large enough to show the ‘coupling strength’ of
some of the factors involved.
THE HEREDITY OF THREE CONTRASTED SEX-LINKED
CHARACTERS
In the former paper (page 89) the second, third, and fourth
combinations were the ones omitted. They are given here in
sequence. The same symbols are used and the same method
employed in writing out the analyses that were used before.
Short, red, black by long, white, yellow
This is the reciprocal of the cross already published (1912,
page 90). When the female, LWY is mated to the male SRB
all the female offspring are long, red, gray and all the males
long, white, yellow. The numerical results for this and the F;
generation are as follows:
33
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
T. H. MORGAN AND ELETH CATTELL
SRB o& by LWY 9
LRN @ = 2838
BR, LWY @ = 157
LRN ~ o> = 391
LRN icte—s 9+
Ig 2 = iby
IA op = Gal
SEIN oto
SRB oot
F, LWY 9 = 248
Wis ot 215
LWBr 2 = 86
LWBr oh = 45
SMC. toi SS
SWBr ot — 10
TAR OP see
SRY “oi=
SR BrGlE— nS
WIN Oe — ane
LWN oci= 4
The analyses follow:
LWbYBrxX — LWbYBrx
SRByBrX — SWbyBr
LWbYBrX — SRByBrX Long, red, normal. _
LWbYBrX — SWbyBr
Long, white, yellow.
LWbyBrX — LWbYBrX — SRByBrX — SRBYBrX
Gametes | SWbyBrX — SWbYBrX — LRByBrX — Peet Eges
LWbyBrX — LWbYBrX — SWbyBr — SWbyBr Sperm
LRN 6
Males
LWBEr 1 SWBr 1
LWY 3 SWY 3
IRS) 1 SRB 1
LRN 3 SRN 3
There were 20 cases of crossing-over counting both males and
females in a total of 662, or 1 to 33. It is legitimate to count
here because both
both sexes
males and females lack the
SEX-LINKED INHERITANCE IN DROSOPHILA 35
dominant characters (C and R). The number of cross-overs is
relatively high. The converse cross gave 1 to 97 (see Jour.
Exp. Zodl., p. 91, vol. 13, 1912).
Of the cross-overs there were 14 in one direction to 6 in the
opposite. Moreover, it is the short, red, browns and yellows,
often assumed to be the less viable combination, that give the
larger number.
Short, red, yellow by long, white, black.
The two crosses recorded under this heading, both absent from
the preceding paper, were Nos. 3, and 4 of the series (page 89).
When the female LWB is crossed to the male SRY all the
female offspring are long, red, gray and all the males long, white,
gray. The numerical results follow:
SRY @ by LWB 9
F, LRN ? = 80
LWN ¢@ = 7
LRN 2? = 572
LRB 9 = 194
LWN @ = 586
LWN ¢ = 314
LWB 9 = 151
- LWB ¢ = 99
a. SWN ¢ = 112
SWB oc = 33
LRY o& = 1o4
LRBr o¢ = 66
SRY o = 295
SRBr o@ = 105
SRN o = til
LWY o@ =
The analysis follows:
LWByBrX — LWByBrX
SRbYBrX — SWbYBr
FP LWByBrX — SRbYBrX
1
LWByBrX — SWbYBr
36
Ga
T. H. MORGAN AND ELETH CATTELL
LWBYBrX — LWByBrX — SRbYBrX — SRbyBrX \
metes SWBYBrX — SWBYBrX — LRbYBr — LBbyBrX f Eggs
of F,
LWBYBrX — LWByBrX — SWbYBr — SWbyBr Sperm
F,
90
Females Males
LWN 6 LWN 3 SWN 3
LRN 6 SRYel 3 BEY. 3
LWBI 2 LWBI 1 SWB 1
LRBI 2 SRBr 1 * LWBr 1
There were 13 cases of crossing over in the 1178 males or 1 to
. Of these crossings 11 were in one direction, and 1 in the
other.
The reciprocal cross gave the following results:
SRY 2 by LWB ¢@
r LRN Or — A989
, SRY ao = 18!
LRN @ = 445
LRB @ = 52
LWN o = 150
LWB oi = 4
SRN @ = 45
SRB. ¢. = 21
SWN o = 58
r SWB oc = 16
; LRY 92? = 72
LRY ot = 1.
LRBr 9 = 21
ERBrsoe—alo
SRY o = 107
SRY 9? = 88
SRBr op = 41
SRBr 9 = 64
SRN oc = ll
SRB c= 5
LWY c= 3
LRN c= 1
1 One LWY found in F,
There were 20 cases of cross-overs amongst 500 males, or 1 in
25.
SEX-LINKED INHERITANCE IN DROSOPHILA 37
The analysis follows:
SRbYBrX — SRbYBrX
LWByBrX — SWbyBr
SRbYBrX — LWByBrX
SRbYBrX — SwhyBr
SRbyBrX — SRbYBrX — LWByBrX — LWBYBrX | ,,,
Gametes LRbyBrX — LRbYBrX — SWByBrX — SWBYBrx f{ ~®®
of F,
Fy
SRbyBrX — SRbYBrX — SWbyBr — SWbYBr Sperm
Females Males
SRBr 1 LRBr 1 SRBr 1 LRBr 1
Sane st RS SRYel 3 LRYel 3
SRBI 1 LRBr 1 LWBI 1 SWBI 1
SRN 3 LRN 3 LWN 3 SWN 3
This last cross gave, the first time it was tried, the results
recorded below. These results were not published because of
the very high number of the cross-over class, SRN = 17. In
the repetition of the experiment the class is again large, but
not so excessive as in the former instance. As there were no
grounds for suspecting these results in other respects we feel
that the only fair course is to add them here. In the sum totals,
however, they have been omitted.
SRY 2 by LWB ¢@
F LRN @ = 76
: SRY o = 86
ERAN Oe — 3
ERB “9 =" 44
LWN o = 101
LWB-o'.= 33
SRN 2 = 25
SR Bu Ga 29)
SWN oc = 30
F SWB o@ = 13
; LRY ¢° = 48
TRAYS xo) =) 44
LRBr @ = 9
ERBr-o =" 73
RAYS aOR 47)
SEOt ie = ai!
SRBr 9 = 17
SRB ou 2
“ SRN gf = 17
SRB a= 1
38 T. H. MORGAN AND ELETH CATTELL
The linkage of RY and WB
In order to obtain a larger amount of data for linkage of R
and Y versus W and B, when the result is not complicated by
other factors, we repeated one of our former crosses that was
too small for the matter in hand. Long, red, yellow females
were crossed with long, white, black males and gave LRN ¢ and
LRY ¢. The numerical results for F,; and F, are given below.
LEY 9. Wee
Fr LRN @ = 622
: LRY #@ = 521
LRN and BI @ 2034
' LRY and Br 2 1740
2
LRY and Br o& 1682
LWN and BI ¢@ 1561
LRN and Bl Ss 48
LWY and Bro’ 10
In the total of 3243 males there were 53 cases of cross-over
or 1:61. In the former experiment there were 260 males and
no cross-overs.
In order to obtain still more data some of the F, LRN eof
the last cross were mated with the double recessive LWY ¢.
Since the latter contains both of the recessive factors, w and b,
all the offspring of the cross, both males and females, may be
counted. The results were as follows:
LWN o and 9 = 2731
LRY co and 2 = 2923
LRN cand @ = 438
LWY cand? = 34
In a total of 5644 there were 77 cross-overs or 1 to 73.3.
SEX-LINKED INHERITANCE IN DROSOPHILA 39
GENERAL CONCLUSIONS
It is not our intention to discuss here the many questions that
arise in connection with these results, but rather to put the
data on record for future reference. There are three topics,
however, that require summing up:
1. The total record of crossing-over for the combination YW
and BR and the reciprocal.
2. The total record of crossing-over for LW and SR and the
reciprocal.
3. The record of crossing-over for LY and SR and the recip-
rocal. .
In all three are six records of crossing over of YW and BR.
The simplest cases are those in which both parents have long
wings. We have Dexter’s large count, counts in two pre-
ceding papers (Morgan ’11, Morgan and Cattell ’12) and the
counts of the present paper.
F2 | CROSS-OVERS | RATIO
Titi ieee 16002 | «189 1: 84:7
DOERR «glee RRC eee ny en 3253 | 5 1 : 650.0
Cul DIDS Gee TS) Ea oa 1806 15 1: 120.4
Mon mam © a celle we cle bse ke ees 8897 | 130 > 68:4
WOtell a, chek Tees ee ace ee eS eae 29958 339 12 88h3
In these counts, the second in the table gave anomalous results,
the cross-overs falling below the other cases. The last three
cases contain brown which has a high mortality; yet the total
result is not far fregm Dexter’s results. In the next table the
results are complicated by the presence of short wings, which
run behind long wings in some combinations, particularly where
short, yellow, white is expected.
F2 CROSS-OVERS | RATIO
ier beeVLor gan, 1912.00.52 taken eat 4890 55 - 1: 89
Morgan-Cattell, 1912. . 2.) wee es: | 10748 ~ $0 il : 134
UGS. bee 5 eee eee ENE Pech ero ee 15538 135 ioe AUR nee a ss
40 T. H. MORGAN AND ELETH CATTELL
In this case the number of cross-overs is considerably smaller
than in the last, but since many classes are included and some
with high mortality the former estimate is probably more cor-
rect.
If we add the two results together we get a total of 35496
counts in addition to the 474 cross-overs, which, on the basis
of calculations here employed, gives a ratio of 1 to 95. Whether
the difference in these ratios obtained at different times stand
for variations in the process itself, or whether they represent
chance results in the sense that for such a relatively rare event
the probable error will account for the differences can be more
profitably discussed at another time when the question of the
gametic ratios has been more fully studied.
The gametic ratio of long and miniature wings versus red and >
white eyes (LW and SR) gives more even results, because the
crossing-over is much more frequent and smaller numbers give
significant results. In our former paper (’12) there were 6829
eases of no crossing-over to 3573 cross-overs; a ratio of 1:1.8.
In the present paper there were 1835 cases of no crossing to 816
cross-overs; a ratio of 1:2.2. In the paper of 1911 there were
3177 cases of no crossing to 1713 cross-overs; a ratio of 1:1.8.
NO CROSSING CROSS-OVERS RATIO
6829 3573 ee
1835 816 1 os
3177 1713 | 1: 8
11841 6102 | 1: 1.94
Adding these cases together, the linkage ratio is 1: 1.94. This
means that the chance is about twice as great that the grand-
parental combination will hold as that it will break. These
results will need to be corrected for viability, and for other
disturbances as well, but there can be little doubt that the results
give approximately the gametic ratio for this combination. No
attempt has been made here to separate those cases where LW
SEX-LINKED INHERITANCE IN DROSOPHILA 41
and SR formed one pair and SW and LR another pair, because
the results, as far as they go at present, do not seem to give very
significant difference in relation to which way the original couple
was made, but here again a more critical examination may be
called for. In these latter figures we have omitted the ‘small
classes’ containing the cross-overs of YW and RG. Their omis-
sion does not affect materially the sum total although they
should be included, with certain corrections that can not now
be made.
Finally, these same data give the linkage ratio of S and L to
B (black and normal) and b (yellow and brown). This is the
linkage between the factor in question for wings and that for
body color. We should expect that this ratio would’ closely
approximate to the last since the first calculation showed that
crossing-over between eye color and body color occurred only
once in 88.3 cases.
In the paper of 1911, there were 1713 cross-overs to 3175
coupled cases, a ratio of 1:2.4 In the 1912 paper (Morgan-
Cattell) there were 2921 cross-overs to 5383 coupled cases, a
ratio of 1:1.8. In the present paper there are 527 cross-overs
to 1155 coupled cases, a ratio of 1:2.2. The sum of all these
cross-overs is 5161 and of the coupled cases, 9713, a ratio of
about 1: 1.98.
CROSS-OVERS COUPLES RATIO
spy atid § a .
Iilereeaaing IG NSE see a eee Ce ee eae ome | 1713 3175 ied Be
NMorean—Cattell, 19125... 0... 2.002. nse. es | 2921 base | ee 8
IMorean-Cattell, 1912). 2.0.62... 06.02 ee | 527 1155 | fl, S242
ripe ee okra Gene, | 5161 o713, -| 121.88
|
In these counts the ‘small classes’ are not included. The ratio
of the total count is 1:1.88, which is almost identical with the
ratio 1: 1.94 for wings and eye color.
The difference to be expected between the gametic ratios of
the last two cases (wings and eye color and wings and body color)
would, on the basis of the first case considered (eye color and
42 T. H. MORGAN AND ELETH CATTELL
body-color) be only 1 to 88. This difference is so small that it
affects the data given above only slightly. It could be accu-
rately estimated only by adding in to the two last calculations the
cross-over and coupled cases given in the ‘small classes,’ but this
addition involves the consideration of another matter, namely,
double-crossing-over. Without going further into the details of
this question the addition of the small classes would be mis-
leading. As the matter is especially considered by Mr. A. H.
Sturtevant in a paper appearing at the same time as this one,
the whole discussion may be left to his handling.
THE LINEAR ARRANGEMENT OF SIX SEX-LINKED
FACTORS IN DROSOPHILA, AS SHOWN BY
THEIR MODE OF ASSOCIATION
A. H. STURTEVANT
From the Zoélogical Laboratory, Columbia University
HISTORICAL
The parallel between the behavior of the chromosomes in
reduction and that of Mendelian factors in segregation was
first pointed out by Sutton (02) though earlier in the same year
Boveri (’02) had referred to a possible connection (loc. cit., foot-
note 1, p. 81). In this paper and others Boveri brought forward
considerable evidence from the field of experimental embryology
indicating that the chromosomes play an important réle in devel- —
opment and inheritance. The first attempt at connecting any
given somatic character with a definite chromosome came with
McClung’s (’02) suggestion that the accessory chromosome is a
sex-determiner. Stevens (’05) and Wilson (’05) verified this
by showing that in numerous forms there is a sex chromosome,
present in all the eggs and in the female-producing sperm, rae
absent, or represented by a smaller homologue, in thr ~miale-
producing sperm. A further step was made when Vaorgan (10)
showed that the factor for color in the eyes of thee fly Drosophila
ampelophila follows the distribution of the seéx-chromosome al-
ready found in the same species by $%tevens (08). Later, 08
the appearance of a sex-linked wine £ mutation in Drosophila,
Morgan (710 a, ’11) was able toJ make clear a new point. By
crossing white eyed, long winggp-q flies to those with red eyes and
rudimentary wings (the ne 7, ‘sex-linked character) he obtained,
in F., white eyed rudiment... winged flies. This could happen
43
44 A. H. STURTEVANT
only if ‘crossing over’ is possible; which means, on the assumption
that both of these factors are in the sex-chromosomes, that an
interchange of materials between homologous chromosomes occurs
(in the female only, since the male has only one sex-chromosome).
A point not noticed at this time came out later in connection
with other sex-linked factors in Drosophila (Morgan ’11 d). It
became evident that some of the sex-linked factors are associated,
i.e., that crossing over does not occur freely between some fac-
tors, as shown by the fact that the combinations present in the
F, flies are much more frequent in F;, than are new combinations
of the same characters. This means, on the chromosome view,
that the chromosomes, or at least certain segments of them, are
more likely to remain intact during reduction than they are to
interchange materials... On the basis of these facts Morgan
(11 ec, 711 d) has made a suggestion as to the physical basis of
coupling. He uses Janssens’ (’09) chiasmatype hypothesis as a
mechanism. As he expresses it (Morgan ’11 ¢):
If the materials that represent these factors are contained in the
chromosomes, and if those that “‘couple’ be near together in a linear
series, then when the parental pairs (in the heterozygote) conjugate
like regions will stand opposed. There is good evidence to support
the view that during the strepsinema stage homologous chromosomes
twist around each other, but when the chromosomes separate (split)
the split is in a single plane, as maintained by Janssens. In consequence,
*he original materials will, for short distances, be more likely to fall
othe same side of the split, while remoter regions will be as likely to
fali Strv_same side as the last, as on the opposite side. In consequence,
we find cutling in certain characters, and little or no evidence at all
of coupling inther characters, the difference depending on the linear
distance apart oie chromosomal materials that represent the factors.
Such an explanatioa®/i2ccount for all the many phenomena that I
have observed and will expin equally, I think, the other cases so far
described. The results are ample mechanical result of the location
of the materials in the chromeé™es, and of the method of union of
homologous chromosomes, and théroportions that result are not so
much the expression of a numerical Stem as of the relative location
of the factors in the chromosomes.
1 It is interesting to read, in this connectionpck’s (’06, p. 248-253) discussion
of the matter.
SEX-LINKED FACTORS IN DROSOPHILA 45
SCOPE OF THIS INVESTIGATION
It would seem, if this hypothesis be correct, that the proportion
of ‘cross-overs’ could be used as an index of the distance between
any two factors. Then by determining the distances (in the
above sense) between A and B and between B and C, one should
be able to predict AC. For, if proportion of cross-overs really
represents distance, AC must be approximately, either AB plus
BC, or AB minus BC, and not any intermediate value. From
purely mathematical considerations, however, the sum and the
difference of the proportion of cross-overs between A and B and
those between B and C are only limiting values for the proportion
of cross-overs between A and C. By using several pairs of
factors one should be able to apply this test in several cases.
Furthermore, experiments involving three or more sex-linked
allelomorphic pairs together should furnish another and perhaps
more crucial test of the view. The present paper is a prelim-
inary report of the investigation of these matters.
I wish to thank Dr. Morgan for his kindness in furnishing
me with material for this investigation, and for his encouragement
and the suggestions he has offered during the progress of the
work. I have also been greatly helped by numerous discussions
of the theoretical side of the matter with Messrs. H. J. Muller,
E. Altenburg, C. B. Bridges, and others. Mr. Muller’s sugges-
tions have been especially helpful during the actual preparation
of the paper.
THE SIX FACTORS CONCERNED
In this paper I shall treat of six sex-linked factors and their
inter-relationships. These factors I shall discuss in the order in
which they seem to be arranged.
—< B stands for the black factor. Flies recessive with respect
to it (b) have yellow body color. The factor was first described
and its inheritance given by Morgan (’11 a).
C is a factor which allows color to appear in the eyes. The
white eyed fly (first described by Morgan ’10) is now known to
be always recessive with respect both to C and to the next factor.
46 A. H. STURTEVANT
O. Flies recessive with respect to O(o) have eosin eyes. The
relation between C and O has been explained by Morgan in a
paper now in print and about to appear in the Proceedings of the
Academy of Natural Sciences in Philadelphia.
P. Flies with p have vermilion eyes instead of the ordinary
red (Morgan ’11 d). :
R. This and the next factor both affect the wings. The nor-
mal wing is RM. The rM wing is known as miniature, the Rm
as rudimentary, and the rm as rudimentary-miniature. This
factor R is the one designated: L by Morgan (’11 d) and Morgan
and Cattell (’12). The L of Morgan’s earlier paper (11) was
the next factor.
M. This has been discussed above, under R. The miniature
and rudimentary wings are described by Morgan (’11 a).
The relative position of these factors is B, G P, Ry. Magee and
O are placed at the same point because they are completely linked.
Thousands of flies had been raised from the cross CO (red) by
co (white) before it was known that there were two factors —
concerned. The discovery was finally made because of a mutation
and not through any crossing over. It is obvious, then, that
unless coupling strength be variable, the same gametic ratio must
be obtained whether, in connection with other allelomorphic
pairs, one uses CO (red) as against co (white), Co (eosin) against
co (white), or CO (red) against Co (eosin) (the eO combination
is not known).
METHOD OF CALCULATING STRENGTH OF ASSOCIATION .
In order to illustrate the method used for calculating the
gametic ratio I shall use the factors P and M. The cross used
in this case was, long winged, vermilion-eyed female by rudiment-
ary winged, red-eyed male. The analysis and results are seen
in table 1.
It is of course obvious from the figures that there is something
peculiar about. the rudimentary winged flies, since they appear
in far too small numbers. This point need not detain us here,
as it always comes up in connection with rudimentary crosses,
\
SEX-LINKED FACTORS IN DROSOPHILA 47
TABLE 1
Long vermilion 9—MpX MpxX
Rudimentary red o—mPX
Fi MpX mPX—long red Q
MpX —long vermilion
Eggs —MPX mPX MpX mpX
Gametes F; SpermMpX
MPX Mpx \
mPX Mpx f{
MpX MpxX \
mpX MpX /{
MPX —long red «—105
mPX —rudimentary red o—33
MpX —long vermilion “~—316
mpX —rudimentary vermilion «—4
—long red 9—451
—long vermilion @—417
F,
—*
and is being investigated by Morgan. The point of interest at
present is the linkage. In the F, generation the original com-
binations, red rudimentary and vermilion long, are much more
frequent in the males (allowing for the low viability of rudiment-
ary) than are the two new or cross-over combinations, red long
and vermilion rudimentary. It is obvious from the analysis
that no evidence of association can be found in the females,
since the M present in all female-producing sperm masks
m when it occurs. But the ratio of cross-overs in the gametes is
given without complication by the F, males, since the male-
producing sperm of the F,; male bore no sex-linked genes. There
are in this case 349 males in the non-cross-over classes and 109
in the cross-overs. The method which has seemed most satis-
factory for expressing the relative position of factors, on the theory
proposed in the beginning of this paper, is as follows. The unit
of ‘distance’ is taken as a portion of the chromosome of such
length that, on the average, one cross-over will occur in it out
of every 100 gametes formed. ‘That is, percent of cross-overs
is used as an index of distance. In the case of P and M there
occurred 109 cross-overs in 40% gametes, a ratio of 26.9 in 100;
26.9, the per cent of cross-overs, is considered as the ‘distance’
between P and M. }
/
j
1
i catt A. H. STURTEVANT
TABLE 2
PROPORTION OF
FACTORS CONCERNED ee ee ae
PER CENT OF CROSS-OVERS
1.2
0.5
26.9
SEX-LINKED FACTORS IN DROSOPHILA 49
THE LINEAR ARRANGEMENT OF THE FACTORS
Table 2 shows the proportion of cross-overs in those cases
which have been worked out. The detailed results of the crosses
involved are given at the end of this paper. The 16287 cases
for B and CO are from Dexter (’12)._ Inasmuch as C and O are
completely linked I have added the numbers for C, for O, and
for C and O taken together, giving the total results in the lines
beginning (C, O) P, B (C, O), ete., and have used these figures,
instead of the individual C, O, or CO results, in my calculations.
The fractions in the column marked ‘proportion of cross-overs’
represent the number of cross-overs (numerator) to total avail-
able gametes (denominator).
As will be explained later, one is more likely to obtain accurate
figures for distances if those distances are short, i.e., if the asso-
Be PR M
00 10 807 337 5786
Diagram 1
ciation is strong. For this reason I shall, in so far as possible,
use the percent of cross-overs between adjacent points in mapping
out the distances between the various factors. Thus, B (C, O),
(C, O) P, PR, and PM form the basis of diagram 1. The figures
on the diagram represent calculated distances from B.
Of course there is no knowing whether'or not these distances
as drawn represent the actual relative spacial distances apart of
the factors. Thus the distance CP may in reality be shorter
than the distance BC, but what we do know is that a break is
far more likely to come between C and P than between B and C.
Hence, either CP is a long space, or else it is for some reason a
weak one. The point I wish to make here is that we have no
means of knowing that the chromosomes are of uniform strength,
and if there are strong or weak places, then that will prevent
our diagram from representing actuai relative distances—but,
I think, will not detract from its value as a diagram.
Just how far our theory stands the test is shown by table 3,
giving observed per cent of cross-overs, and distances as calcu-
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
~
50 A. H. STURTEVANT
lated from the figures given in the diagram of the chromosome.
Table 3 includes all pairs of factors given in table 2 but not used .
in the preparation of the diagram.
It will be noticed at once that the long distances, BM, and
(C, O)M, give smaller percent of cross-overs than the calculation
ealls for. This is a point which was to be expected, and will be
discussed later. For the present we may dismiss it with the
statement that it is probably due to the occurrence of two breaks
in the same chromosome, or ‘double crossing over.’ But in the
case of the shorter distances the correspondence with expectation
is perhaps as close as was to be expected with the small numbers
that are available. Thus, BP is 3.2 less than BR, the difference
TABLE 3
FACTORS | CALCULATED DISTANCE | Ee Ceo ae el
(So) a, Aaa on ee ee 30.7 32-2
15122) 6 TN Re hc A ee Oe aa 7 35.5
(3!) ROOMS, atl ne a 57.6 37.6
(OSO)R. eras See ea aw By 33.7
(SGA CMR Spa SN? nee. en etneainy Pee 56.6 45.2
expected being 3.0. (C, O)R is less than BR by 1.8 instead of by
1.0. It has actually been found possible to predict the strength
of association between two factors by this method, fair approx-
imations having been given for BR and for certain combinations
involving factors not treated in this paper, before the crosses
were made.
DOUBLE CROSSING OVER
On the chiasmatype hypothesis it will sometimes happen, as
shown by Dexter (’12) and intimated by Morgan (’11 d) that
a section of, say, maternal chromosome will come to have paternal
elements at both ends, and perhaps more maternal segments
beyond these. Now if this can happen it introduces a com-
plication into the results. Thus, if a break occurs between B
and P, and another between P and M, then, unless we can follow
P also, there will be no evidence of crossing over between B and
SEX-LINKED FACTORS IN DROSOPHILA 51
M, and the fly hatched from the resulting gamete will be placed
in the non-cross-over class, though in reality he represents two
cross-overs. In order to see if double crossing over really does
occur it is necessary to use three or more sex-linked allelomorphic
pairs in the same experiment. Such cases have been reported
by Morgan (711 d) and Morgan and Cattell (’12) for the factors
B, CO, and R. They made such crosses as long gray red by
miniature yellow white, and long yellow red by miniature gray
white, etc. The details and analyses are given in the original
papers, and for our present purpose it is only the flies that are
available for observations on double crossing over that are of
interest. Table 4 gives a graphical representation of what hap-
pened in the 10495 cases.
Double crossing over*does then occur, but it is to be noted that
the occurrence of the break between B and CO tends to prevent
that between CO and R (or vice versa). Thus where B and CO
did not separate, the gametic ratio for CO and R was about
1 to 2, but in the cases where B and CO did separate it was
about 1 to 6.5.
Three similar cases from my own results, though done on a
smaller scale, are given in the table at the end of this paper.
The results are represented in tables 5, 6 and 7.
TABLE 4
NO CROSSING OVER SINGLE CROSSING OVER DOUBLE CROSSING OVER
B iB |B B
co co | co CO
R R | | R R
6972 3454 60 9
TABLE 5
NO CROSSING SINGLE CROSSING OVER DOUBLE CROSSING OVER
0 lo 0 0
P P P P
R R R [R
194 102 rn 1
ae A. H. STURTEVANT
TABLE 6
NO CROSSING SINGLE CROSSING OVER | DOUBLE CROSSING OVER
B B B | B
O O O | O
M M M M
278 160 1 0
TABLE 7
B B Bp | Be a
om O O O O oO] | "-o/ 4 ae
P i 1 12 1 iE 1e 12
R R Rihas ve R R R Jie
393 203 19 6 2 1 l, ne
It will be noted that here also the evidence, so far as it goes,
indicated that the occurrence of one cross-over makes another
one less likely to occur in the same gamete. In the caseof
BOPR there was an opportunity for triple crossing over, but it
did not occur. Of course, on the view here presented there is
no reason why it should not occur, if enough flies were raised.
An examination of the figures will show that it was not to be
expected in such small numbers as are here given. So far as I
know there is, at present, no evidence that triple crossing over
takes place, but it seems highly probable that it will be shown
to occur.?
Unfortunately, in none of the four cases given above are two
comparatively long distances involved, and in only one are there
enough figures to form a fair basis for calculation, so that it seems
as yet hardly possible to determine how much effect double cross-
ing over has in pulling down the observed percent of cross-overs
in the case of BM and (C, O)M. Whether or not this effect
is partly counter-balanced by triple crossing over must also
remain unsettled as yet. Work now under way should furnish
answers to both these questions.
2A case of triple crossing over within the distance CR was observed after
this paper went to press. i
SEX-LINKED FACTORS IN DROSOPHILA 53
TABLE 8
(The meaning of the phrase ‘proportion of cross-overs’ is given on p. 45)
BO.
F.:
BP.
F3:
F::
F3:
Fs:
BR.
Fs:
Pi: gray eosin 2 X yellow red &@
F,: gray red Q X gray eosin @
99, g.r. 241, g.e. 196
og, er. Oi ge: 176; yar. 195, y:e. 2
1 ti f ae
roportion of cross-overs, 373
P,: gray red 9 X yellow vermilion 7 —
F,: gray red 9 X gray red & :
2 9, g.r. 98;
o'o', gr. 59, g.v. 16, y.r. 24, y.v. 33
Back cross, F; gray red? 2 from above X yellow vermilion 77
OOP etoile. Save UL yer: LZ, y-v. Al
Bp, er. 20, £V.i3, yr. &, y.y. 21
P;: gray vermilion 9 X yellow red 7
Fi: gray red 2 X gray vermilion 7
2: 29, g.r. 199, gv. 182
io’, gr. 54, g.v. 149, y.r. 119, y.v. 41
P,: yellow vermilion 9 X gray red &
Fi: gray red 2 X yellow vermilion ~7
© 9, g.r. 472, g.v. 240, y.r. 213, y.v. 414
Oo’, g.r. 385, g.v. 186, y.r. 189, y.v. 324
\Fy: gray vermilion X yellow red (sexes not recorded) ‘
Fi: grayred 92. These were mated’to yellow vermiliond ¢ of other stock
2 9, g.r. 50, g.v. 96, y.r. 68, y.v. 41
oo’, g.r. 44, g.v. 105, y.r. 86, y.v. 47
-
Proportion of cross-overs, adding 2 2 from BOPR (below), =
P; miniature yellow @ X long gray @
F,: long gray 2 X miniature yellowc
OPO e da ley ee. 7, Sys 0;
‘ff lg. 10, ly. 1, m.g. 6, m.y. 8.
P,: long yellow 9 X miniature gray 7
F,: long gray .? X long yellow 7
»: 9 9, lg. 148, ly. 130
oc’, Lg. 51, ly. 82, m.g. 89, m.y. 48
115
Proportion of cross-overs, 304
54 A. H. STURTEVANT
TABLE 8 (continued)
BM. P,: long yellow 2 X rudimentary gray
Fi: long gray 2 X long yellow 7
Fi: 2° 9, lg. 591, ley. 549
iat, ke: 228; Liy..3i!, re 20) rsya3
Pi: long gray 9 X rudimentary yellow @
F,: long gray @ X long gray &
F.: 9 9, Lg. 152
IH, lg. 42, Ly. 29, rg. 0, ry. 0
é 260
Proportion of cross-overs, 693
COP. Pi: vermilion 2 X white @
Fi: red Q X vermilion @
F2: 2 9, r. 320, v. 294
oo’, r. 86, v. 206, w. 211
(7 of the vermilion 2 2 known from tests to be CC, 2 known to be Ce. 7 white
oo Pp, 2 pp.)
Back cross, F; red 99 from above X white oo’, gave
Fe: 2 9, 7.195, w. 227,
oo), F106; vi 164, w. 164
Out cross, F; 2 2 as above X white o’o recessive in P, gave
Fs: 9 9, r. 35, v. 65, w..98
Gig, Te 30, V. 10, We-9S
Proportion of cross-overs, 748
COR. Pi: miniature white 2 xX longred &
F,: long red 2 X miniature white ¢7
Fe: 9 9, Lr. 193, l.w. 109, m.r. 124, m.w. 208
oo, ler, 2025 low. 114s merioli2sssmeawe li:
P,: long white 2 X miniature red @
F,: long red 2 X long white 7
F.: 9 9 lr. 194, 1. w. 160
oo lr. 52, 1. w. 124, m.r. 97, m.w. 41
563 ;
Proportion of cross-overs, 7-775 OT, adding such available figures from
Morgan (711 d) and Morgan and Cattell (712) as are not complicated
16438
by the presence of yellow or brown flies, 4749
COM. P;: long white 2 X rudimentary red @
F,: long red 9 X long white 7
Ra? 2°9:, Lr. 157; law. ee
o'a', Lr. 74, low. Se2,shUere on Cueweee
76
Proportion of cross-overs, 161
)
OR.
F.:
Fi:
OM.
iis
CR.
Fy:
SM.
SEX-LINKED FACTORS IN DROSOPHILA
TABLE 8 (continued)
P,: black red @ & black eosin-vermilion oi
Fi: black red @ X black red &
: (all black), 9 @, r. 885
Glo’, fe o2l, v.11125, et 122,1e:-v. 268
247
Proportion of cross-overs, 836
P,: long red 2 X miniature eosin
F,: long red @ X long red &
QQ, Lr. 408
oo", l.r. 145, l.e. 67, m.r. 70, m.e. 100
P,: long eosin @ X miniature red
F,: long red @ X long eosin &@
Mole Were MOTO), We 9)
oot, kr. 27, le. 54; mr. 56; me. 19
. 183
Proportion of cross-overs, 255
” 538
P;: long eosin 2 X rudimentary red
F,: long red @ X long eosin @
9 9, Lr. 368, le. 266
oo", Lr. 194, l.e. 146, ru.r. 40, ru.e. 24
218
Proportion of cross-overs, 404
P,: long white @ X miniature eosin <7
F,: long eosin 2 X long white o
: 9 9, Le. 185, l.w. 205
oo’, le. 54, lw. 147, m.e. 149, m.w. 42
P,: long eosin 9 X miniature white
F,: long eosin 2 X long eosin <7
@ Q, Le. 527
oo’, l.e. 169, l.w. 85, m.e. 55, m.w. 128
; 236
Proportion of cross-overs, 399
P,: long white 2 X rudimentary eosin
F,: long eosin 2 X long white @
One heazs, Lawaom
oo’, le. 112, l.w. 217, ru.e. 4, ru.w. 0
112
Proportion of cross-overs, 333
0
56 A. H. STURTEVANT
TABLE 8 (continued)
PR. P,: long vermilion (yellow) @ X miniature red (yellow) &
_ F,: long red yellow 2 X long vermilion yellow o
Fe: (all y.) 2 9, l.r. 138, lv. 110
ao, lr 8) éve 117, mar 7 meet
P,: long vermilion (gray) @ X miniature red 7
F,: long red 2 X long vermilion @
B,:.9 9, Lr. 116, Lv. 110
aig’; Lr.~ 2, l-vs, 81, mir... 96, anv!
P;: miniature red 2 X long vermilion ~
F,: long red 2 X miniature red 7
Bis 2 9] brs45) mor. 49
Sic, lara we ives27 mer 26, mv O
F, long red 2 2 from above X miniature red oc of other stock, gave
m2 OO Lr 74 omer 52 Z
Citra, 1-VeL00,n. POM vaek
. 17
Proportion of cross-overs, 573
PM. P:: long vermilion 9 X rudimentary red 0
Fi: long red @ X long vermilion @
Fo: 9-9, lr. 451, Lv. 417
oo’, L.r. 105, l.v. 316, ru.r. 33, ru.v. 4
109
Proportion of cross-overs, 405
405
OPR. P:: long vermilion 9 X miniature eosin
F;: long red 2 X long vermilion @
He QO ter 2050 Ixvelse
oo’, Lr. 1, l.v. 109, l.e. 8, l.e.-v. 53, m.r. 49, m.v. 3, m.e. 85, m.e.-v. 0
BOM. P;: long red yellow 2 X rudimentary eosin gray
; F,: long red gray 2 X long red yellow &
Bs3. 9 9, lrg: 530; ley. 453
od’, Lr.g. 1, lr.y. 274, le.g. 156, Lesy. 0, ru.r.g. 0, ru-r.y.-4, tore
Tu.e. y. 0
BOPR. P;: long vermilion brown @ X miniature eosin black @
F,: long red black 2 X long vermilion brown
F2: 2 9, Lr.bl. 305, L.r.br. 113, l.v.bl. 162, lv.br. 256
J, Lr.bl. 0, l.r.br. 2, L.v.bl. 3, l.v.br. 185, l.e.bl. 9, l.e.br. 0, l.e.-v.bl.
127, l.e.-v.br. 0, m.r.bl. 1, m.r.br. 76, m.v.bl. 1, m.v.br. 10, m.e.bl.
208, m.e.br. 3, m.e.-v.bl. 0, m.e.-v.br. 0
SEX-LINKED FACTORS IN DROSOPHILA 57
POSSIBLE OBJECTIONS TO THESE RESULTS
It will be noted that there appears to be some variation in
coupling strength. Thus, I found (CO)R to be 36.7; Morgan and
Cattell obtained the result 33.9; for OR I got 34.0, and for CR, 28.5.
The standard error for the difference between (CO)R (all figures)
and CR is 1.84 per cent, which means that a difference of 5.5
per cent is probably significant (Yule 711, p. 264). The observed
difference is 6.1 per cent, showing that there is some complication
present. Similarly, BM gave 37.6, while OM gave 54.0—and
BOM gave 36.7 for BM, and 36.5 for OM. There is obviously
some complication in these cases, but I am inclined to think that
‘the disturbing factor discussed below (viability) will explain
this. However, experiments are now under way to test the effect
of certain external conditions on coupling strength. It will be
seen that on the whole when large numbers are obtained in differ-
ent experiments and are averaged, a fairly consistent scheme
results. Final judgment on this matver must, however, he with-
held until the subject can be followed up by further exp*tiicits.
Another point which should be considered in this connection
is the effect of differences in viability. In the cas of P and M,
used above as an illustration, the rudimentary winged flies are
much less likely to develop than are the longs. Nowif the via-
bility of red and vermilion is different, then the longs do not give
a fair measure of the linkage, and the rudimentaries, being present
in such small numbers, do not even up the matter. It is probable
that there is no serious error due to this cause except in the case
of rudimentary crosses, since the two sides will tend to even up,
unless one is very much less viable than the other, and this is
true only in the case of rudimentary. It is worth noting that
the only serious disagreements between observation and calcu-
lation occur in the case of rudimentary crosses (BM, and (CO)M).
Certain data of Morgan’s now in print, and further work already
planned, will probably throw considerable light on the question
of the position and behavior of this factor M.
58 A. H. STURTEVANT
SUMMARY
“
It has been found possible to arrange six sex-liked factors in
Drosophila in a linear series, using the number of cross-overs
per 100 cases as an index of the distance between any two factors.
This scheme gives consistent results, in the main.
A source of error in predicting the strength of association be-
tween untried factors is found in double crossing over. The
occurrence of this phenomenon is demonstrated, and it is shown
not to occur as often as would be expected from a purely mathe-
matical point of view, but the conditions governing its frequency
are as yet not worked out.
These results are explained on the basis of Morgan’s
application of Janssens’ chiasmatype hypothesis to associative
inheritance. They form a new argument in favor of the
chromosome view of inheritance, since they strongly indicateethat
the factors ie are arranged in a linear series, at least
mathematically.
te en + \
. ‘
November, 912.
LITERATURE CITED
Boveri, T. 1902 Ueber mehrpolige Mitosen als Mittel zur Analyse des Zellkerns.
Verh, Phys.-Med. Ges. Wiirzburg., N.F., Bd. 35, p. 67.
Dexter, J. S. 1912 On coupling of certain sex-linked charaters in Drosophila.
Biol. Bull., vol. 23, p. 183.
JANSSENS, F. A. 1909 La théorie de la chiasmatypie. La Cellule, tom. 25, p.
389.
Lock, R. H. 1906 Recent progress in the study of variation, heredity, and evo-
lution. London and New York.
McCune, C. E. 1902 The accessory chromosome—sex determinant? Biol.
Bull., vol. 3, p. 43.
Morean, T. H. 1910 Sex-limited inheritance in Drosophila. Science, n.s., vol.
32, p. 120.
1910 a The method of inheritance of two sex-limited characters in
the same animal. Proc. Soc. Exp. Biol. Med., vol. 8, p. 17.
1911 The application of the conception of pure lines to sex-limited
inheritance and to sexual dimorphism. Amer. Nat., vol. 45, p. 65.
SEX-LINKED FACTORS IN DROSOPHILA 59
1911 a The origin of nine wing mutations in Drosophila. Science,
n.s., vol. 38, p. 496.
1911 b The origin of five mutations in eye color in Drosophila and
their modes of inheritance. Science, n.s., vol. 33, p. 534.
1911 c Random segregation versus coupling in Mendelian inheritance.
Science, n.s., vol. 34, p. 384.
1911 d An attempt to analyze the constitution of the chromosomes
on the basis of sex-limited inheritance in Drosophila. Jour. Exp.
Zool., vol. 11, p. 365.
Moraan, T. H. anp Catretyt, E. 1912 Data for the study of sex-linked inher-
itance in Drosophila. Jour. Exp. Zodl., vol. 13, p. 79.
Stevens, N. M. 1905 Studies in spermatogenesis with special reference to the
‘accessory chromosome.’ Carnegie Inst. Washington, publ. 36.
1908 A study of the germ-cells of certain Diptera. Jour. Exp. Zodl.,
vol. 5, p. 359.
Sutton, W. 8S. 1902 On the morphology of the chromosome group in Brachy-
stola magna. Biol. Bull., vol. 4, p. 39.
Witson, E. B. 1905 The behavior of the idigchromosomes in Hemiptera. Jour.
Exp. Zoél., vol. 2, p. 371.
1906 The sexual differences of the chromosome-groups in HLetniptera,
with some considerations on the determination and inheritance of
sex. Jour. Exp. Zodl., vol. 3, p. 1.
Yue, G. U. 1911 An introduction to the theory of statistics. London.
= — 3 a
+? a =
= 7 "~«
, :
‘o ‘
+
.
= HO fe =
- .
a .
= *
‘i
.
7 CG
im
+
e =
=
4
.
'
J
- ‘
.
c *
a
=
=
*
’
‘
~< '
7
° <
e
aa
, —- s_¢
A STUDY OF THE MALE GERM CELLS IN NOTONECTA
ETHEL NICHOLSON BROWNE
From the Zoélogical Laboratory, Columbia University
TEN PLATES
CONTENTS
eet rOdUCtlON sens ones ert ee ee nek oe Ea ER es, eee 62
Mien Wesierich Arig heChnigue.. sy cme. | seneee eis sree aK Biel Pookie ab, es tem 63
melibice Chromosomes. 5. se date ecticis oe Oot es ee POS RS SO Ae EIS Ae 64
PAPE SCLV Ail OMS ethos NPY a. < oe SE EN EE EM tae 2 ore eee a oe 64
eiNotonecta undulatas isons eee tere ea ee ee 65
DEP INSIELORAUA IS, (oeeT avs pater aber TREE Feo Dees aes CEES 66
Ease TN Tiras 0d IY Hae ade en eg Gin, ko cn Oey NE ote ae eR are 67
Ae Ne eiaican (eanteloanGdusitiety) aaceeein. c9a ie eects <eceie er 69
Be IDISCUISSLOTIORe Ee aero ety ee eth oo itt treet ee os ee ee 70
1. Relation between chromosome number and species......... 70
2. Temporary fusion and separation of chromosomes......... 75
Stel Bl aVs) DG) 5 OYA cate Baha NE oP SS ee reer en A i in eS ee een 77
LW, LEE INY OS 0) AD POU & G< s RL GEE bi aie Oe epee ae ee cee SC ae 78
ANS AISNE SIS TO YST EEN TEN 2 Se aed Nam ot En NEL = et cee ROR eR a ee aa ES ge 7
1. Early growth stages and formation of karyosphere......... i
PRs DescE pion OMkaryvOspnere: 7.08 ker a nee ee 81
Ses incoluiiony Of Kany OSp Mere wy....0 x, c%- <0. 00 ees eee eee 82
18); IN) the ULE ADE ete nines & biti ccc te Se ERE anit tons Anant aire ae g ocsis 82
1. Early growth stages and formation of karyosphere........ 82
2. Description and dissolution of karyosphere.......:........ 88
CRIN PREETOT AUS aa See ee eee eer tose. Shey te eae eee eee 4 84
ee conelusious ad, COMIPALISOUS:...< ~. 2.02255 e ee ie 6 eaten 84
\j'n) LEALOFSL TENSES EG) 5 eee saat LS eC a A ge eR eee See 86
UME COPTAMGESCLID LOU t 245-326 ne ake cee whe Se eis Sera ee aa an ees 86
: SURI Nir SS Cl Gt is hn Sy A cee) Sh tegen SN re ee ED tanec rae SO 86
is AN] STE OTH: Se Dai teal Rs Ar Loe, SO a eee Oe el Se 86
Soe NEU Ulta byte bo eee eet ene Loti SETS wah 7
Bey Wetsrledrdeserip tions: 2%. 4.44 asset ee oe eee eae Sees Ge 87
ft GINS ts Ul aa nacheoncky. coo, REE Te Na eects oe dS 87
Be VERIO ee, sy PEERY ete ies ORE ES Beate cen. eftcve. ots 87
DF Grosses. ie ee ie ois ler en eo habe 89
CA Doublerod--eeareee eee tee ase ees J) soos 90
5 PAE oe BNE ore St i c Gein ee eee 90
22 NEITTORA ase ee tee ee Ei cn ax Fs le thoe c's os = 91
SEIN SATLING Ul] uate eae ney Re ae ais chee Sis. 2 eis" 91
Cre DISCUSSTON ee ire HN Ae ee a rie eine pee stereiieae S Sini 91
62 ETHEL NICHOLSON BROWNE
Viv Mitochondria... 2.0.65: : acc os soe 2 oo. eo oe ee eee 94
A Observations :...:d..<0%.ss3eSon Sete: fo eee eee eee 94
1. Late’ growth and division stages... 2.2... = 2 =s.22 eee 94
2. Karly growth stages: nuclear plate.:..2.:. 50... 255 eee
Bi: DISCUSSION: «oo, Avec e ee eA oe ee eee 97
VER Summary... .. 605... 5 05.0 pe oie. Scam cee ole ee een eee nee 98
Keiterature cited. 2h. c's oh cee Hee os oe De ee eee eee 99
I. INTRODUCTION
Gilson, in his study of the spermatogenesis of the arthropods in
1885, passed over Notonecta with the remark, “‘ Les phénoménes
de la spermatogénése y sont fort simples et presentent peu de parti-
cularités dignes d’étre mentionées” (op. cit., p. 123), adding that
N. glauca possesses the longest and largest spermatozoa known.
More recently, Pantel and Sinéty (’06) have published a copious
memoir on ‘‘Les cellules de la lignée male chez le Notonecta
glauca L.,’’ and they, unlike Gilson, have found themselves “en
presence d’un assez grand nombre d’images d’un aspect nouveau,
parfois trés inattendu ou méme déconcertant” (op. cit., p. 90).
Further work on this genus seemed to be warranted by the very
peculiar appearances described by these authors, as well as by
the acknowledged slight treatment of the maturation divisions in
favor of the stages concerning the transformation into the sper-
matozobn. The problem was suggested to me by Prof. E. B.
Wilson, to whom I wish to express my most sincere thanks for
his valuable advice and criticism during the course of the investi-
gation. This study is based on the three common American
species, kindly identified by Mr. E. P. Van Duzee as Notonecta
undulata (Say), N. insulata (Kirby) and N. irrorata (Uhler);
the form used by the French authors was the European species,
N. glauca. While the American species agree with N. glauca
in presenting many very puzzling appearances, they differ from
it in several important respects and also differ considerably
among themselves. The two facts of main interest are, first,
the presence of a karyosphere or body in which the chromatin is
aggregated during the growth stages in all three species, as was
noted also in N. glauca by Pantel and Sinéty; and secondly,
the relation of the chromosome number to the species, a brief
MALE GERM CELLS IN NOTONECTA 63
summary of which has already been published (Browne ’10).
The present study deals only with the growth stages and matura-
tion divisions, no attempt having been made to treat the later
stages which have been elaborately worked out by Pantel and
Sinéty.
II. MATERIAL AND TECHNIQUE
The material, consisting of the three species already mentioned,
was collected during four summers at Woods Hole, Massachusetts.
These species differ considerably from one another in size, in
wing coloration and markings, and in other characters. N.
insulata is the largest, with brown wings usually marked with
two black bands. N. irrorata is slightly smaller, and its wings
are black, mottled more or less with brown. N. undulata, the
most common species, is considerably smaller than the other two
two, or three black bands. In respect to germ cell production,
there are two types. In N. undulata, all the stages of the sper-
matogenesis occur in the adult and even in the very young larva
throughout the summer. In N. irrorata and N. insulata, during
the greater part of the summer, the testis of the adult and late
larva is filled with cells in the late growth stages, the younger
cysts being empty except those at the very tip of the testis where
a few spermatogonia occur. For only about a week during the
summer are division stages found in these two species; after this
the testis is filled with spermatids and spermatozoa. Pantel and
Sinéty have noted the same slow evolution of the germ cells in
N. glauca. Probably owing to this long period of growth, the
cells of N. irrorata and N. insulata are larger than those of N.
undulata. The great size of the cells coupled with the diagram-
matic clearness of the spindle fibers and asters make the material
exceptionally fine for the study of the maturation divisions.
The French authors find it otherwise for N. glauca, stating that
“La figure chromatique est d’un type malingre, dans les cinéses
maturatives du Notonecta, et peu favorable 4 une analyse
detaillée des phénoménes morphologiques” (op. cit., p. 136).
64 ETHEL NICHOLSON BROWNE
The only difficulty with my material has been the scarcity of
spermatogonial divisions and early growth stages.
The testes are bifurcated coiled tubes lying on either side of
the alimentary canal. They were dissected out in Ringer’s
solution and transferred at once to the fixing fluid. Flemming’s
strong fluid, Bouin, Carnoy, Gilson and corrosive sublimate were
used with results that are favorable in the order named. Heiden-
hain’s haematoxylin was used almost exclusively as a stain, though
some saffranin preparations were made. In order to demon-
strate mitochondria, some of the testes were fixed with Benda’s
modification of Flemming’s fluid, and these were subsequently
treated with his mitochondrial stain of sulphuralizarinate of
sodium and crystal violet, used according to his original method.
’The results of the fixed preparations have been controlled by
observations of the living cells both with and without intra-vitam
stains. By dragging the testis over a slide and mounting in a
drop of ‘Ringer’s solution, very good results were obtained.
The mitochondria and karyosphere may at once very clearly be
seen; and after half an hour or so, the chromosomes in division
stages come out very clearly. This is probably due to the fact
that some change has taken place in the chromosomes, and it
may be that they are not visible in the living state. Such would
seem to be the case from the fact that constant observation of
anaphase spindles failed to reveal any progression of the chromo-
somes toward the poles. In some cases it was possible to count
the chromosomes in these preparations.
III. CHROMOSOMES
A. Observations
As pointed out in my preliminary paper (’10), the study of
the chromosomes in Notonecta has proved of much interest from
the fact that the change in number from species to species can
here be attributed to the relations of a particular chromosome.
Briefly the results are as follows. In all three species there is
present an unequal X Y-pair of chromosomes which divide sepa-
rately in the first spermatocyte division but are united in the
MALE GERM CELLS IN NOTONECTA 65
second, thus making the total number of separate chromatin
elements one greater in the first than in the second division. In
N. undulata, there are 14 chromosomes in the first division, 13 in
the second, including two small chromosomes. In N. irrorata,
there are 13 in the first and 12 in the second, including only one
smallone. In N. insulata there are either 14 or 13 in the first, and
12 in the second; when there are 14 in the first, there are two small
ones, when 13 there is only one free small one, but the other small
one can often be detected attached to the largest chromosome.
This species thus appears to be intermediate in respect to the
chromosomes between N. undulata with a larger number, and N.
irrorata with a smaller number.
1. Notonecta undulata. In N. undulata, the typical, and I
am inclined to believe, the invariable, arrangement in the first
spermatocyte division is a ring of 12 surrounding two very small
chromosomes. ‘This is shown in polar view in figures 1 and 2
and in side view of a spindle from two adjoining sections in figure
3 A,B. Very frequently side views present the appearance
shown in figure 4 B, the two pairs of small chromosomes lying in
a straight line, as though on the same spindle fiber (A, B, C are
serial sections of the same spindle). This is probably due to
the fact that they lie very close together and the smaller of the
two pairs usually precedes the other in division. In the periph-
eral ring can be distinguished one chromosome larger than the
rest, one very small one slightly larger than the central ones,
and ten of intermediate and intergrading sizes.
In the second spermatocyte division, side views clearly show
the presence of an unequal X Y-pair (fig. 5 B). Since these chro-
mosomes have divided separately in the first division, as is the
case in many other Heteroptera, there should be one chromosome
less in the equatorial plate of the second division. That there
are 13 chromosomes in this division is shown in side view in
figure 5 A, B, C (from the same spindle), and in polar view in
figures 6 and 7. (In the latter figure the X-chromosome is seen
at a lower focus). In this division, X and Y always take up their
position in the center of the spindle, as they do in other Hemip-
tera. A rather interesting phenomenon occurs in Notonecta
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
66 ETHEL NICHOLSON BROWNE
with regard to the XY-pair. The two components frequently
fail to conjugate, and lie in the second metaphase side by side, on
separate spindle fibers (figs. 8, 9). A little later stage is shown in
figure 10 A, where the small component is evidently going to one
pole, the large one to the other; these are frequently connected
at this time by an oblique fiber (fig. 11). The size relations of the
chromosomes are evident from an inspection of figures 5, 6, 7, 9,
10. The two smallest chromosomes which were in the center in
the first division are now in the peripheral ring. The third small
one of the first division was apparently the Y-chromosome. ‘The
largest chromosome is again evident in the peripheral ring. The
X-chromosome is one of the larger chromosomes, probably the
third largest. The size relations come out very clearly in figure
12 A, B, which are sister anaphase groups from the same spindle.
It is apparent from these groups that X is present in one of the
resulting cells, Y in the other, and since these cells develop directly
into the spermatids and thence into spermatozoa, the latter must
be of two types in respect to the chromosome content.
In the spermatogonial groups, there are 26 chromosomes
(figs. 13, 14), among which a largest and a smallest pair can easily
be recognized. There are two pairs of very small ones, evidently
corresponding to the two small bivalents in the center of the spin-
dle in the first spermatocyte division. There are two very large
chromosomes corresponding to the one large one in the haploid
groups. Then there is a pair slightly smaller than these and an-
other odd large one. This is evidently the X-chromosome, and
Y is distinguishable as the fifth small chromosome which clearly
has no mate of its own size.
2. Notonecta irrorata. In N. irrorata, the typical arrangement
is a ring of 12 chromosomes surrounding one small one (figs.
15, 16). The second small chromosome which occurred inside the
ring in N. undulata is here lacking. Serial sections of a spindle
in side view showing the total number and the typical arrange-
ment, are represented in figure 17 A, B,C. A few cases have been
observed where the components of the central pair apparently
fail to conjugate and lie on separate fibers in the metaphase;
these are distinctly univalent in contrast to the other bivalents
MALE GERM CELLS IN NOTONECTA 67
(fig. 18). Their behavior resembles that of the components of
the X Y-pair in the second division, which may or may notcon-
jugate before going to the poles; and it is also analogous to that
of the m-chromosomes of the Coreidae which conjugate very late
and do not fuse. As in N. undulata, one chromosome in the
peripheral ring is larger than the others, and it is here in some
cases longitudinally split (fig. 16). There are in this species two
small chromosomes in the peripheral ring.
In the second division, the presence of an unequal X Y-pair
in the center of the spindle is evident from side views (fig. 19 B).
Serial sections of a spindle in side view (fig. 19 A, B, C), and
polar views (fig. 20), show that there are 12 chromosomes, in
contrast to the 13 of N. undulata. Here too, the components
of the X Y-pair may fail to conjugate before the second division,
and lie on separate spindle fibers in the center of the spindle
(figs. 21 B, 22). It is evident that X and Y are less unequal in
size than in N. undulata, X being comparatively smaller and Y
larger. The largest chromosome is distinguishable among the
others, and also the three small ones of the first division (the one
in the center and the two peripheral ones). The fact that the
result of this division will be two kinds of cells (ultimately sper-
matozoa) differing in chromatin content in respect to one chro-
mosome is apparent from sister anaphase groups (fig. 23 A, B).
Only one clear spermatogonial group has been found (fig. 24);
the number here is 24, including three pairs of small chromosomes,
corresponding to the three small ones of the spermatocyte
divisions; the largest pair corresponding to the large one of the
haploid groups; two other large pairs, and one odd large one.
This is doubtless the X-chromosome; the Y-chromosome is indis-
tinguishable, but must be one of the smaller intermediate ones.
3. Notonecta insulata. N. insulata has proved an extremely
interesting species from the fact that two distinct types of chro-
mosome groups occur in the first divisicn, in approximately equal
numbers and side by side in the same cyst. One type has 14
chromosomes including two small ones in the center, like N. undu-
lata (figs. 25, 26); the other type has 13 chromosomes, including
only one small one in the center, like N. irrorata (figs. 27, 28).
68 ETHEL NICHOLSON BROWNE
The discrepancy in number was very perplexing until consecu-
tive sections of complete spindles were examined as they appeared
in side view. It was then discovered that in many cases the dis-
crepancy is accounted for by the fact that the second small chromo-
some which appears in the center in the 14-type is frequently found
attached to the largest chromosome in the 13-type. In figures 29-
31 A, B, C are shown serial sections of three spindles which have
only one small chromosome in the center, the other small one being
attached to the large chromosome forming the compound chro-
mosome Ma (macrochromosome + small autosome). Polar views
of the compound chromosome are rather difficult to obtain owing
to the small size of the smaller component. Such a view, from a
spindle cut somewhat obliquely, is given in figure 32 where both.
components show very clearly. In figures 33, 34 A, B, C, are
shown spindles of the other type, where both small chromosome
pairs are in the center and the large chromosome is not compound.
In over thirty cases where the chromosomes have been counted
in consecutive sections in side view, the apparent 13-type has
been found to be due to the attachment of the second small chro-—
mosome to the large chromosome. Jt is always this particular
chromosome, the largest one, with which the little one is associated.
In many cases, however, when only one small chromosome appears
in the center, the compound character of the large one cannot be
detected, the two components having probably fused beyond rec-
ognition. When there are two small ones in the center, there are
14 chromosomes, and the large chromosome is never compound.
Besides the two small chromosomes at the center of the spindle
(or one in the center and the other attached to M ) it is clear from
an inspection of the figures that there is always another small
chromosome in the peripheral ring. Attention may also be called
to the fact that the largest chromosome is usually longitudinally
split, as it is occasionally in N. irrorata (figs. 25-28, 32).
In the second division, the number of chromosomes is always
12, so far as I have observed (figs. 35-37). As in the other two
species, an unequal X Y-pair is here present in the center of the
spindle (fig. 37 A), and the components are frequently found
side by side, having apparently failed to conjugate (figs. 38, 39).
MALE GERM CELLS IN NOTONECTA 69
They are more nearly equal in size than in N. undulata, in this
respect resembling those of N. irrorata. The invariable number
12 is accounted for on the assumption that the second small
chromosome which in the first division is sometimes separate and
sometimes associated with the large one, has fused with it in all
cases before the second division. Additional evidence is given
by the fact that in most cases two of the chromosomes are con-
siderably smaller than the others, one of these corresponding to
the small central one, and the other to the small peripheral one
of the first division (figs. 35-39). The large chromosome, how-
ever, presents an unexpected appearance. It gives no evidence
whatever of its real composition of two very unequal parts, but
it appears in the metaphase as.a large quadripartite chromosome,
as though each part into which it divides were composed of
two equal parts (p. 89). The longitudinal split which was very
noticeable in the first division in polar view marks the division
plane of the second division. In figure 40 A, B are shown two
sister plates of an anaphase; the groups are identical except for
the middle chromosome, and it is evident that on this account
two kinds of cells are produced which give two kinds of sper-
matozoa.
Unfortunately no spermatogonial groups have been found of
which a satisfactory count could be made. The expectation
would be either 26 single chromosomes, or 24 including two
compound ones.
4. Notonecta glauca (Pantel and Sinéty). According to the
account of Pantel and Sinéty (’06), there are in N. glauca some-
times 12, sometimes 13 chromosomes in the first division. They
state that they are unable to account for this difference, but they
also say, ‘‘elle (la couronne équatoriale) comprend un anneau
périphérique, plus une ou deux unités situées au centre” (op. cit.,
p.139): No figures of polar views are given but it seems probable
from the statement that the discrepancy is here due to the pres-
ence or absence of a second small chromosome in the center, as
in the case of N. insulata. They do not state in the text the
number present in the second division, but figure 12 chromosomes.
The writers mention no unequal X Y-pair in the second division.
70 ETHEL NICHOLSON BROWNE
If this pair is absent, N. glauca differs radically from the other
three species, and it may be that this discrepancy is analogous to
that found in Metapodius (Wilson ’09 a) where in different indi-
viduals of the same species a Y-chromosome may be present or
absent; or in the mosquitoes, where there is a.typical unequal
XY-pair in Anopheles punctipennis, while in two other genera,
Culex and Theobaldia, the differential chromosomes are absent
(Stevens ’11). It is possible that the X and Y chromosomes are
present in N. glauca, but are of practically equal size as in Onco-
peltus and Nezara hilaris (Wilson ’11), and have been overlooked.
Pantel and Sinéty call attention to the presence of an extra large
chromosome, which they call the ‘chromosome exceptionelle,’
suggesting that it may be an accessory chromosome, but they are
convinced that it participates in both divisions. This body seems
quite similar in appearance and behavior to the large chromosome
described in N. insulata. It is unfortunate that N. glauca cannot
be brought into line with the three American species in regard to
chromosome number, but the account of Pantel and Sinéty is
inadequate to permit the attempt.
B. Discussion
1. The relation of chromosome number to species. It is doubtful
whether every chromosome in the three species can be homolo-
gized individually, for the size relations are different in some
respects. By comparing the spermatogonial groups of N. undu-
lata and N. irrorata (figs. 13, 14, 24) it is evident that there are
5 large chromosomes in the former, and 7 large ones in the latter;
and the X Y-chromosomes are of different relative size in the three
species. On the other hand, one largest chromosome can be
traced throughout the history of all three species; likewise the
smallest chromosome, not only by its size but especially by its
position in the first spermatocyte division. We may also homolo-
gize the second small chromosome which is present in the first
division of N. undulata, in the center of the group, with the one
of similar size which is sometimes present in N. insulata in the
same position. And since the steps in the process of fusion
MALE GERM CELLS IN NOTONECTA rg
can actually be observed in N. insulata, it seems reasonable to
attribute its absence in N. irrorata to its permanent association
with the largest chromosome. Representing the large chromo-
some, or macrochromosome by M, the two small autosomes by
a, the unequal chromosomes by X, Y, and the larger autosomes by
A, we may schematize the results as follows:
PRODUCTS OF THE FIRST SPERMATOCYTE PRODUCTS OF THE SECOND SPERMATOCYTE
DIVISION DIVISION
M+9A+Y+acta_ (l3)
either
(IM+9A+X+Y-+a+a (14)| (Ma+9A+X+a (12)
. insulata ..... 4] or | and
IMa+9A+X4+Y+a (13)
\|
Ma+9A+Y+a (12)
\
Ma + 9A + X + 4 (12)
.irrorata........|Ma+9A+X+Y+a_ (13)|4 and
Ma+9A+ Y+a (12)
The scheme shows the intermediate condition of N. insulata
between N. undulata with a larger number of chromosomes and
N. irrorata with a smaller number. The large and small
chromosomes in N. undulata are always separate, in N. insulata
sometimes separate and sometimes associated, and in N. iror-
ata are presumably always associated. This may represent a
progressive (or regressive) series, or the three forms may represent
different modifications of a single original type.
The somatic characters do not afford decisive evidence concern-
ing these three possibilities, although the wing color fits in with
the view that N. insulata is an intermediate species. By sub-
stituting brown pigment for the white of N. undulata, the wing
coloring and pattern of N. insulata is obtained; further, by sub-
stituting for this brown pigment, black, but leaving some of the
brown as mottling, the wing pattern of N. irrorata is obtained.
On the other hand, N. irrorata is intermediate in size between the
othertwo, and N. undulata is intermediate in respect to the distance
1 This scheme is identical with that published in my preliminary paper (710)
except that X, Y have been substituted for J, 7.
| {M+9A+X-+a+a (13) 9
. undulata....... M+9A+X+Y-+a+t+a (14) and
of
2
i2 ETHEL NICHOLSON BROWNE
between the eyes. It may be that the wing color is directly cor-
related with the fusion and separation of the two chromosomes,
and that the other somatic differences are connected with other
chromosomes, which, as I have stated, differ in size in the differ-
ent species.
In the Acrididae, McClung (’05) has found that a particular
genus, Hesperotettix, is distinguished from others of the family
by a special arrangement of the chromosomes, by which the acces-
sory is always associated with another chromosome forming a
multiple element. He concludes that this arrangement “‘is
genetically connected with the subsequently appearing char-
acters” (op. cit. p. 326). This correlation of a multiple chromo-
some element and a generic difference in the Acrididae is directly
comparable to the correlation of a multiple chromosome element
and a specific difference in Notonecta.
The correlation of a definite number of chromosomes with a
particular species is a well established fact throughout the animal
and plant kingdoms, and is admitted by practically all cytologists,
with only a few exceptions. In several cases, however, it has
been shown that the number is constant for the individual, but
differs for different individuals. This is sometimes due to the
presence of ‘supernumerary’ chromosomes, as in Metapodius,
Banasa ealva, Diabrotica and Ceuthophilus ( Wilson ’09 a, Stevens
12a, b), and sometimes to the fact that two types of chromo-
some groups occur within the species, one with twice the number
of the other, as in the well known cases of Ascaris megalocephala,
Echinus microtuberculatus and Helix pomatia, and as in Artemia
salina, as recently pointed out by Artom (11). Cyclops viridis
is apparently a species in which different numbers occur in differ-
ent varieties (Chambers ’12). With these and possibly a few
other exceptions, the number of chromosomes is a specific char-
acteristic, although occasional fluctuations may occur.
Further, there are many cases where closely related species
have the same number of chromosomes. For example, five spe-
cies of Euschistus have the same number (Montgomery, Wilson) ;
four species of Sagitta (Stevens 710), and three species of Ceresa
(Boring ’07). In other cases, related species differ only slightly
_
MALE GERM CELLS IN NOTONECTA 73
in number. For example, three species of Podisus have the
diploid number of 16, two species 14 (Montgomery, Wilson).
In this category belong the three species of Notonecta. But
much wider differences in related species may occur, as for exam-
ple, in two closely similar species of Banasa, of which B. dimidiata
has 16 chromosomes, B. calva has 26 (Wilson ’09b). So in
Thyanta, in which a distinction between two species has just
recently been rediscovered by Barber, two types of chromosome
groups occur, 27—28 in T. calceata, and 16 in T. custator. Among.
the phylloxerans, there is considerable variation; four species
have 6 chromosomes in the diploid groups, one species has 8, one
12, and one 22 (Morgan ’09). Similarly in the aphids, the hap-
loid number ranges from 3 in the willow aphid to 16 in the maple
aphid (Stevens 06). Likewise Braun (’09) found in fifteen spe-
cies of Cyclops a wide range of number, from 6 to 22, although
several species have the same number. In the Oenotheras,
mutants have been found with 14, 15, 21 and 28 chromosomes
(Lutz 712).
When we come to groups less closely related than species,
marked differences in the chromosome number frequently occur.
In the family Jassidae, the diploid number varies from 15 to 23
in different genera; in the Cercopidae, from 15 to 27; in the Mem-
bracidae, from 17 to 21 (Boring ’07). In ten genera of the Chryso-
melidae, there is a range from 16 to 36 (Stevens); in the Coreidae,
from 13 to 27 (Montgomery, Wilson). These are a few of the
many cases of divergence within a family. There are on the other
hand, a few cases where a constancy obtains throughout as wide
a group as a family, as for example, in ten genera of the orthop-
teran family, Acrididae (McClung ’05, et al.), and in four genera
of the opisthobranch mollusks. The constancy in number goes
still further in some of the Amphibia, where all the urodeles, so
far as examined, apparently have 24 chromosomes in the diploid
groups.
It is therefore evident that while in some cases the chromosome
number is the same for members of rather a large group, it is not
necessarily the same for even very closely related forms. It is
true in general, however, that closely related forms have the same
74 ETHEL NICHOLSON BROWNE
or very nearly the same number of chromosomes. ‘This fact has |
led Montgomery and McClung to the view that the number and
arrangement of chromosomes should be considered as an important
character in taxonomy. More recently, McClung (’08) has
expressed this view forcibly in his paper on “Cytology and Tax-
onomy.” It is of interest in this connection to find that in the
Oenotheras, according to Lutz (712), all individuals having a given
type of vegetative character have the same number of somatic
chromosomes, irrespective of the origin of these individuals,
whether hybrid or mutant.
That one method by which a change in the chromosome num-
ber has taken place is by the fusion or separation of particular
chromosomes seems highly probable from the evidence given by
Notonecta, where we have all the stages in the process in the
three species. Such a process may also be indicated by the d-
chromosome in Nezara (Wilson 711). A somewhat similar idea
was put forth by Montgomery (’01) before the relation of the
X-chromosome with sex had been established, to explain the
occurrence of an odd number of chromosomes in the spermato-
gonial groups of some of the Hemiptera; the odd number represent-
ing, he believed, a transition stage between two even numbers.
A change in number by a process of fusion has been advocated
by McClung (’05) in regard to the multiple chromosomes in the
Orthoptera. A change in number by a process of splitting has
been advocated by Payne (’09) in the case of the multiple X-ele-
ment in the reduvioids, and this may likewise apply to that
of many other forms, such as Phylloxera, Syromastes or Ascaris
lumbricoides, as has been indicated by Wilson (711). A second
probable method of change is by a process of progressive reduc-
tion and final disappearance of particular chromosomes, as was
originally suggested by Paulmier (’99) in the case of the small
m-chromosomes of the Coreidae, and later by Wilson in the case
of the Y-chromosome.
These two methods will account for gradual and slight changes
in the chromosome number. Such wide variations as occur in
closely related species, e.g., in Banasa, Thyanta, and the phyl-
loxerans, must be accounted for in some other way. Wilson (’11)
MALE GERM CELLS IN NOTONECTA 75
suggests that some sudden mutation has taken place, involving a
new segregation of the nuclear material, and causing a change
in number and size relations of the chromosomes, but not in their
essential quality.
A fourth method by which either a slight or a radical change in
the chromosome number might take place is by an abnormality
occurring in mitosis, as has been suggested by several authors,
either by an unequal distribution of the chromosomes to the
daughter cells, or by an arrest of cell division after a division of
the chromosomes. The former abnormality has actually been
observed in the case of Metapodius (Wilson ’09 a), and the Oeno-
theras (Gates ’08, et al.). The possibility of the occurrence of
the second abnormality is shown by the experiments of Gerassi-
mow (’01) on Spirogyra, of Némec (’04) on Pisum and of Boveri
(05) on sea-urchin eggs, in which a monaster was produced
instead of an amphiaster, the chromosomes dividing but not the
nucleus, and the double number of chromosomes remaining in
subsequent divisions. To this cause has been attributed the
occurrence of triploid and tetraploid mutants in Oenothera(Gates
09, Lutz ’12), and it seems probable that many of the cases of a
double number of chromosomes occurring in closely related forms
of some animals (e.g., Ascaris megalocephala), and many plants
have been brought about in this way.
2. Temporary association and separation of chromosomes. ‘The
condition of temporary association and separation of particular
chromosomes which occurs in N. insulata is of especial interest
in comparison with other forms. In the first place, there are
cases where the union and separation concerns the sex-chromo-
somes only. In these cases the X-element may consist of two or
more components—in Acholla (Payne ’09) and Ascaris lumbri-
coides (Edwards 710) as many as five—that appear as separate
chromosomes in the diploid nuclei but become associated in the
spermatocyte divisions and behave as a single accessory.
In a second category may be placed those forms where there is
a temporary or permanent association of the sex chromosomes
with other chromosomes. Sinéty (’01) was the first to describe
a case of this sort in the phasmids Menexenus and Leptynia.
76 ETHEL NICHOLSON BROWNE
Here the accessory becomes attached to another chromosome in
th first division, and goes over with it to one pole. McClung
(05) describes a similar relation for the acridian Hesperotettix
and the locustid Anabrus. In Hesperotettix he found that it
is always the largest chromosome with which the accessory is
associated. In Mermeria, another acridian, a similar multiple
element becomes further associated with another tetrad, and in
division this complex acts as a single bivalent, with the anoma-
lous result that entire tetrads pass to one pole. More recently,
Boring (09), Boveri (09) and Edwards (10) have found in the
case of Ascaris megalocephala that the accessory may be free or
may be indistinguishably united with another chromosome.
Stevens (711) similarly finds in one of the mosquitoes a close
union of X and Y with a pair of autosomes in the spermatocyte
divisions, while in the spermatogonia they may or may not be |
closely united with them.
In a third category we may place the form N. insulata where
there is a temporary association and separation of two ordinary
chromosomes (autosomes). That this association has some sig-
nificance can scarcely be doubted when we consider that it is
always two particular chromosomes that are united. If the
union of the two chromosomes is the primitive condition, then the
secondary separation might mean that certain characters are
being segregated from other characters. If the separation
is primary, the fusion of the chromosomes might, mean that a
certain series of characters which were entirely independent of
another series have become linked with them. It is possible that
just as the association of a sex-chromosome and an autosome may
serve as a morphological basis for sex-linked inheritance as pointed
out by Wilson (11), the association of two autosomes may give
the morphological basis for the cases of coupling that have
recently been made known in experimental work. For example,
Bateson and Punnett (’11) find that in the sweet pea, bluecolor
and long pollen are usually combined, red color and round pollen,
etc.; and in Primula, according to Gregory (’11), magenta color
is coupled with short style. In Drosophila also, a linkage of
a color and a wing factor has been found by Morgan and Lynch
a
MALE GERM CELLS IN NOTONECTA 77
(12). In order that the linkage may take place, however, in
N. insulata, we must assume that a particular chromosome of
one pair always associates with a particular one of the other pair
and never with its mate. When these two chromosomes are
permanently associated, as is probably the case in N. irrorata, one
chromosome might serve as the basis of the linked characters.
3. The XY-pair. The observations on Notonecta add nothing
new to the main facts in regard to the X Y-chromosomes. The
difference in size between the two components is much more
marked in N. undulata than in the other two species; similar differ-
ences between related species have been found in Nezara, Euchis-
tus and other Hemiptera: The only departure from the usual
behavior of the X Y-pair is in the failure of the two components to
conjugate. Usually in the Hemiptera the components come to-
gether in the prophase of the second division, in contrast to all the
other chromosomes which have paired before the first division.
In Notonecta, however, in all three species this pairing frequently
does not take place, and the two components of the X Y-pair lie
side by side in the metaphase of the second division and pass to
opposite poles. A similar condition has been seen by Mont-
gomery (710) in Euchistus, but here it is apparently very excep-
tional for he found it only in one case out of 672. As to the time
of conjugation of chromosome.pairs, there is a graded series. The
autosomes conjugate in the general synaptic period; the m-chro-
mosomes undergo a late synapsis in the prophases of the first
division; the X Y-chromosomes in most Hemiptera do not finally
conjugate till the end of the first division; the X Y-chromosomes
of Notonecta frequently do not ever form a definitive dyad. It
is of interest, from the point of view of the mechanics of division,
to find that a linear arrangement of the components of a chromo-
some pair is not necessary for their distribution to the opposite
poles of the spindle.
78 ETHEL NICHOLSON BROWNE
IV. KARYOSPHERE?
A. Notonecta insulata
1. Formation of karyosphere. The growth stages of the pri-
mary spermatocyte are probably separated by a considerable
interval from the last spermatogonial telophase. In the earliest
spermatocytes observed (fig. 41), the chromatin is massed in a
single large body, the karyosphere, and thin strands of linin are
scattered through the nucleus. The first change to take place is
the accumulation of chromatin on the linin threads (fig. 42).
Although the source of this chromatin cannot be definitely ascer-
tained, it seems most likely, from a study of many of these nuclei,
that it comes from the karyosphere following the course of the
linin strands and tending to aggregate at particular points. The
threads from the karyosphere are more or less twisted, and show
a distinct radial arrangement. The tendency of the chromatin
to aggregate in clumps becomes more marked until the nucleus
is filled with small chromatin masses connected with each other
and with the karyosphere by thin deeply staining strands (fig.
43). During this process the chromatin masses are frequently
approximated in pairs (figs. 42, 48). This fact suggests that the
masses represent chromosomes which are conjugating. Although
the evidence is not conclusive that synapsis takes place at this
time, the whole process of the formation of these chromatin masses
seems unintelligible otherwise. The number of the chromatin
masses varies considerably in nuclei of the same cyst; the maxi-
mum number is however greater than the reduced number of
chromosomes, although probably not as large as the somatic
number. This is easily accounted for on the assumption that the
pairing of different bodies takes place at different times, as seems
* The term ‘karyosphere’ is used in this paper in the sense in which Blackman
(03) first used it to denote a structure consisting of chromatin and other sub-
stances, such as linin and karyolymph. It is thus a broader term than karyosome
or net-knot or chromosome-nucleolus which is usually applied to a mass of pure
chromatin. ‘Karyosphere’ is practically identical with Carnoy’s ‘nucléole-noyau,’
or miniature nucleus; it is however difficult to determine the presence of a mem-
brane as is required by his definition. Although it is impossible to tell at all
stages whether accessory material is present with the chromatin, the term karyo-
sphere will be used throughout the discussion.
MALE GERM CELLS IN NOTONECTA 79
to be the ease. This process in N. insulata is somewhat similar
to that described by Arnold (’08) in Hydrophilus, and by Davis
(08) in some of the Orthoptera. The masses are perhaps
comparable with those described in many plant cells as
‘prochromosomes.®
After the stage of the scattered chromatin masses in N. insulata,
a process of absorption sets in. The masses gradually decrease
in size and the connecting strands become thicker, especially
in the region of the karyosphere (fig. 44). By an absorption of
all the chromatin masses, a spireme of approximately uniform
thickness is formed which is irregularly coiled about in the nuclear
cavity (fig. 45). The spireme is not formed here by an unravel-
ing process as described by Janssens, Davis and others, but
by a uniform distribution of the chromatin material along the
connecting strands. A somewhat similar formation of the lepto-
tene spireme has been described by Gérard (’09) in Stenobothrus.
The spireme now becomes arranged in loops which are more or
less oriented toward the karyosphere and may connect with it
at their apices (fig. 46). The karyosphere seems to act as a center
of activity like the chromoplast of Eisen and Janssens. Later
the loops take up a position on the nuclear wall, receding from the
karyosphere which remains in the interior (fig. 47). The loops
then become somewhat irregular, coiled and thicker; their stain-
ing capacity gradually diminishes until in faintly stained prepara-
tions only the karyosphere and a few scattered remnants of the
loops take the chromatin stain (figs. 48-50). This change can
be readily appreciated by comparing these three figures, which
are from the same slide. The nuclear cavity is, however, filled
with a flocculent reticulum, which is quite faint in lightly stained
preparations, but is very noticeable and takes a deep chromatin
stain in preparations that are less extracted.
The foregoing facts in Notonecta are extremely perplexing.
Since the spireme is formed from the scattered masses, it must
apparently at one time contain the essential elements of the chro-
mosomes. A transfer of these elements into the karyosphere
3 These are evidently similar to the ‘massive bodies,’ recently described in
other insects by Wilson (’12), as occurring in Stage b.
80 ETHEL NICHOLSON BROWNE ,
may be afforded in two ways; by a flow along the whole length of
the spireme and (for a brief period) by the connection of the curves
of the loops with the karyosphere. After the loops become dis-
connected and oriented, they give somewhat the appearance
of the ‘bouquet,’ though this is never so regular or clearly marked
as in Batrachoseps, Tomopteris, etc. The polarized loops of
the bouquet stage are in other forms the forerunners of the chro-
mosomes. This may also be the case in Notonecta, but if this
be so, the conclusion seems unavoidable that the fundamental
material must subsequently return to the karyosphere, for, as will
be shown, the chromosomes later arise directly from the latter.
It is possible that this material flows back into the karyosphere
along the faintly staining threads that can usually be traced from
the loops; but it seems more probable that after the loops are
disconnected, their substance does not enter the karyosphere.
This conclusion is based on the fact that the loops withdraw from
the karyosphere, and that some of the more remote threads keep
the chromatin stain after the ones near the karyosphere have lost
it. If this be so, considerable ground is given for the view that
there are here two kinds of chromatin, corresponding with those
designated by Lubosch (’02) as trophochromatin and idiochro-
matin. The chromatin, which is later to form the chromosomes
is transferred to the karyosphere in one or both of the two ways
suggested, while the rest of the chromatin becomes disconnected
from the karyosphere and is represented by the flaky reticulum
in the nuclear cavity.t In the case of N. glauca, Pantel and
Sinéty have likewise concluded that the material in their ‘monili-
form cords’ becomes achromatic and is spread through the nuclear
cavity, taking no part in the chromosome formation. This ma-
terial has no doubt a metabolic function during the enormous
growth of the spermatocyte. From a comparison of figures 51 and
53 drawn to the same scale, it is evident that the surface of an
equatorial plane of the full grown spermatocyte nucleus is approx-
imately five times that of the youngest one, which means that its
‘ The paper of Vejdovsky (’12), containing very interesting observations, bear-
ing on this and other subjects here treated, unfortunately came to my notice too
late for his results to be incorporated in this article.
MALE GERM CELLS IN NOTONECTA 81
volume is nearly twelve times as great. About half of the increase
in size takes place before the disappearance of the spireme, the
other half after the karyosphere is fully formed (figs. 51, 52, 53).
The growth of the spermatocyte in Notonecta is so great as to
be comparable with that of an oocyte (as was noted also by Pantel
and Sinéty), and it must involve a similar metabolic activity.
It is well known that in the eggs of many forms, some of the chro-
matin is eliminated during the growth or maturation divisions;
this is probably correlated with the metabolism of the cell. The
diminution of the chromatin by a casting off of the ends of the
chromosomes in Ascaris megalocephala and A. lumbricoides is
probably of a similar nature. The ring in Dytiscus (Giardina
01) which during four divisions passes to only one of the resulting
cells, i.e., the oocyte, has likewise been interpreted by Boveri
(04) and Goldschmidt (04) as representing chromatin that is
concerned in the nutrition of the cell. For a comprehensive
account maintaining the existence of two kinds of chromatin,
basichromatin and oxychromatin, see Stauffacher (710).
2. Description of karyosphere. The appearance of the karyo-
sphere varies considerably with the stage of growth, with differ-
ent fixing fluids and stains and with the amount of extraction. In
many preparations, especially those stained deeply with haema-
toxylin, no structure is evident; it is merely a round or approxi-
mately round mass of vesicular appearance (fig. 54 A). In other
preparations an irregular contour and difference in staining capac-
ity in different regions gives it a spongy appearance (figs. 54 B,
55). In haematoxylin preparations well extracted and in saf-
franin preparations the structure is quite definite. The karyo-
sphere consists apparently of dense, compact bodies of varying
size embedded in a less dense matrix (figs. 53 A, 54 C, D), the for-
mer being probably the chromatin proper. When crowded these
bodies give somewhat the appearance of a continuous spireme
closely convoluted (fig. 53.4). In the younger stages, the karyo-
sphere tends to have a vesicular appearance, ahd later the spongy
or granular structure is more evident. In the living material,
the differentiation of two sorts of material in the karyosphere is
perfectly evident, the denser substance taking the form of com-
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
82 ETHEL NICHOLSON BROWNE
pact masses, either isolated or continuous, embedded in a less
dense substance (fig. 56). In this condition the karyosphere
persists during the whole growth period.
3. Dissolution of karyosphere. Just prior to the formation of »
the aster, the karyosphere tends to assume a more definite out-
line appearing in lightly stained preparations as a more or less
spherical body in which darkly staining bodies are embedded.
Figure 57 is from the same slide as figure 58, the former being an
earlier stage. After the formation of the aster, the chromatin
masses leave the karyosphere as compact bodies, either irregular
in shape or threadlike (figs. 59-61). It is perfectly evident that
the karyosphere is breaking up into its two constituents; the
darkly staining chromatin bodies are passing out of the karyo-
sphere, leaving the less dense, paler material which becomes
rounded and now appears as a typical plasmasome. As the
masses leave the plasmasome, they quickly proceed to the
periphery of the nucleus where they take on the form of double
threads, as will be described later. In figure 60 the plasmasome
is entirely free of chromatin, staining a pale gray; some of the
chromosomes are seen along the nuclear margin. The mass of
chromatin that has just left the plasmasome, is the large chromo-
some, M, which has been mentioned previously as one element of
the compound chromosome typical for this species. The plas-
masome has usually one or several small vacuoles in the interior.
The body gradually decreases in size and disappears in the late
prophase.
B. Notonecta undulata
1. Formation of karyosphere. In the earliest spermatocyte,
a small karyosphere is present, and the rest of the nuclear cavity
is filled with a reticulum of linin (fig. 62). The reticulum in-
creases slightly in staining capacity and takes on the appearance
of a very thin spireme, often twisted in spirals; a leader is usually
to be seen running from the karyosphere (fig. 63). This is
undoubtedly the leptotene stage. The spireme becomes more
heavily staining and tends to contract to one side of the
nucleus; the threads are still very thin, and in some cases appear
MALE GERM CELLS IN NOTONECTA 83
to run parallel in pairs (fig. 64). This is probably a synizesis
stage and it is likely that a conjugation of the parallel threads
takes place, although this cannot be conclusively demonstrated.
’ In the next stage, the spireme is somewhat thicker and becomes
arranged in loops, oriented toward the karyosphere (fig. 65). The
loops next become attached to the karyosphere at their apices;
this gives the appearance of arms radiating out from the karyo-
sphere (fig. 66), and affords an opportunity for the transfer of
material from the loops to the karyosphere. Very frequently
the threads give the appearance of being longitudinally split.
The loops now recede from the karyosphere and become arranged
on the nuclear wall while the karyosphere remains in the in-
terior (fig. 67). The loops gradually disappear, and the nucleus
is filled with an irregular reticulum which is appreciable when
the staining is dark, but pale when the stain is more extracted,
in contrast to the dark karyosphere.
The early history of N. undulata thus differs markedly from
that of N. insulata. There is no formation of scattered chro-
matin masses, but instead a leptotene stage which is followed by a
synizesis or contraction, after which a spireme is present which
is quite similar to that of N. insulata. The fate of the material
in the spireme offers the same difficulties as in N. insulata, but it
seems probable that here too some chromatin passes into the kary-
osphere, and other chromatin is segregated out and furnishes the
material in the nuclear cavity.
2. Description and dissolution of karyosphere. The karyosphere
of N. undulata appears very much as it does in N. insulata; it
tends to be vesicular in the early stages, especially in darkly
stained preparations, but later appears granular or rope-like
(figs. 68, 69). At the approach of the prophase the karyosphere
becomes broken up into a variable number of small round bodies,
giving the appearance of a mass of marbles (fig. 70). The mass
loosens and from it project one or two longitudinally split threads
(figs. 71, 72). That the thread may be formed directly from
the balls which become arranged in pairs is evident from figure
72. As the double threads form, they go to the nuclear wall where
they become the diffuse prophase chromosomes which will be
84 ETHEL NICHOLSON BROWNE
described later. In figure 73, some of the chromosomes already
lie at the periphery while others are still being transformed from
balls into threads in the interior. From this process it seems
clear that the material of one particular chromosome may be —
broken up into more or less isolated bodies, which later arrange
themselves and fuse together to give rise to a continuous struc-
ture. As the last chromosomes form, some of the substance of
the original karyosphere is left behind as a plasmasome. This
body must apparently be formed from material that was in the
balls which segregates out in the process of thread formation.
The plasmasome is not colorless as it is in N. insulata, but remains
dark even in well extracted preparations, probably owing to the
fact that some of the chromatin is left behind. The plasmasome
has a vesicular appearance and is frequently vacuolated. It
gradually decreases in size and disappears in the late prophase.
C. Notonecta irrorata
The earliest stage occurring in my material is represented in
figure 74. This shows the presence of a looped spireme on the
nuclear wall, more or less oriented toward the karyosphere. The
spireme gradually disappears as in the other two species, probably
giving rise to the flaky material in the nuclear cavity. The karyo-
sphere at this stage is distinctly vesicular. Just before the aster
forms, however, it breaks up into a mass of balls of an inconstant
number which form threads, very much as described for N.
undulata (figs. 75, 76, A, B). A plasmasome is formed during
this process which sometimes takes a heavy chromatin stain
(fig. 77), and sometimes appears grey and vacuolated; it gradually
decreases in size and disappears in the late prophase.
D. Conclusions and comparisons
A karyosphere is apparently present in the three species of
Notonecta which I have examined, throughout the entire history
of the spermatocytes. In the very early stages and in the later
stages, this is the only body in the nucleus that takes a deep
chromatin stain, but there is an intervening stage when a chro-
matic spireme is present. It would appear that the chromatic
MALE GERM CELLS IN NOTONECTA 85
material comes from the karyosphere, and that later, at least
that part of it which contains the essential elements, returns to
the karyosphere. This flow of material back and forth from the
karyosphere seems highly remarkable, and is probably concerned
with a rearrangement of the chromatin particles, for which a
particular structure (i.e., the spireme) is necessary. It is unfor-
tunate that I have not been able to determine more definitely
how the definitive karyosphere is formed. In N. glauca, accord-
ing to Pantel and Sinéty, a pale nucleolus is present in the early
stages and the chromatic material from the ‘moniliform cords’
condenses around it, a process similar to that described for some
of the myriapods (Blackman ’05 b, ’07) and the dragon fly (McGill
06). In other myriapods, the accessory chtomosome is the cen-
ter around which the other chromosomes are deposited to form
the karyosphere (Blackman ’05 a, ’07, Medes ’05).
In Notonecta, the chromatin remains massed together in the
karyosphere, in an apparently inactive state during a long growth
period. It is usually possible at this time to distinguish the chro-
matin material as distinct bodies, not necessarily the individual
chromosomes, embedded in a less dense (plasmasome) material.
Such an intimate association of plasmasome and chromatin mate-
rial, where the latter is recognizable as distinct bodies, has been
described in some of the myriapods (Blackman, Medes), and in
the case of the XY-chromosomes in some of the reduvioids
(Payne ’09) and in certain Coleoptera and Diptera (Stevens).
In some cases, e.g., in Scolopendra heros (Blackman ’05 a) and
in Hydrophilus (Arnold ’08), there is apparently no plasmasome
material associated with the karyosphere. The distinction of
two sorts of material is extremely apparent in N. insulata in the
early prophase, when the chromatin leaves the karyosphere as
compact masses, and the remaining material becomes a typical
pale plasmasome. In N. irrorata and N. undulata, and appar-
ently also in N. glauca, the dissolution of the karyosphere takes
place a little differently, by breaking up at once into a number of
separate elements. In either case, there can be no doubt that
the material which forms the chromosomes comes from the karyo-
sphere. The events described for Notonecta do not seem to me
at variance with the hypothesis of the genetic continuity of the
86 ETHEL NICHOLSON BROWNE
chromosomes; it seems, on the contrary, altogether reasonable
to suppose that the essential chromosome elements retain their
identity throughout the entire process.
V. PROPHASES
A. General description
1. Notonecta insulata. It is of interest to trace the history of
the chromatin from the dissolution of the karyosphere until the
formation of the definitive chromosomes. After the irregular
masses of chromatin have left the plasmasome, they pass from the
interior of the cell to the nuclear membrane; and here the
chromosomes pass through a diffuse stage before assuming their
finalform. At first they appear on the nuclear wall as longitudin-
ally split rods, long, thin and somewhat curved (fig. 78); the
rods are apparently made up of a linear ‘series of granules
(chromomeres) which give them an irregular contour. The usual
prophase figures, rings, crosses, etc., are formed from the longitu-
dinally split threads (fig. 79); they will be described in detail later.
By a process of condensation are formed the definitive chromo-
somes which are typically dyad-like in appearance; their tetrad
nature cannot be made out unless they lie in a favorable position
and are very critically observed (fig. 80). During these stages,
the chromosomes have remained close against the nuclear mem-
brane, and it is from this position that they are drawn on to the
spindle in the late prophase. In figure 81 they are seen irregularly
arranged on the spindle, prior to their final grouping around the
equator. Attention may be ealled to the fact that frequently in
the late prophase, the small chromosome is found attached to the
large one, forming the compound chromosome, to which reference
has been made in an earlier part of the paper (figs. 80, 81).
2. Notonecta irrorata. The history for this species is practically
the same. The thin longitudinally split rods (fig. 82) on the nu-
clear wall give rise to rings, crosses, ete. (fig. 83). While still on
the nuclear wall, they condense into the definitive chromosomes,
which later become irregularly arranged on the spindle (fig. 84).
MALE GERM CELLS IN NOTONECTA 87
3. Notonecta undulata. In this species also, the first indication
of the final chromosomes is the presence on the nuclear wall of
thin, more or less coiled threads (fig. 85). The prophase figures
which these form are quite different, however, from those of the
other two species. At first these have a very vague, spongy
appearance and are coarsely granular (fig. 86). By a process of
condensation, they become more compact and more definite in
outline (fig. 87). The figures are quite irregular in shape, but
in general consist of two bars, diverging or united at one or both
ends. While in this stage, the nuclear membrane breaks down
and the spindle fibers begin to form. The chromosomes are still
quite irregular in shape after the spindle is fully formed (fig. 88),
and do not assume their definitive form until the full metaphase.
B. Detailed description
1. Notonecta insulata. a. The ring. The M-chromosome is
usually the last one out of the plasmasome, and is therefore in the
interior of the nucleus at the time that all the other chromosomes
are in a diffuse condition on the nuclear wall (fig. 60). Owing
to'this fact and also to its greater size, its history can be traced
throughout the prophase and also during the first and second
maturation divisions. Whereas the other chromosomes come out
of the plasmasome in more or less irregular masses, the M-chromo-
some has the form of two rods, somewhat coiled about each other,
but in general taking the same direction (fig. 89 A—D). The two
rods untwist, and open out in the middle, usually becoming or
remaining united at the ends (fig. 90 A—D). By opening out
still more, a small ring is formed (fig. 91 A—D); frequently at this
stage and occasionally earlier, a longitudinal split is present, in
one or both half rings. If we term the original line of separation
between the two rods a longitudinal split, this is the second longi-
tudinal split. By this time, the M@-chromosome has reached the
nuclear wall, and at once a process of expansion sets in. The
ring opens out until the enclosed space becomes relatively very
large and the ring itself correspondingly thin (fig. 92 A, B). It
88 ETHEL NICHOLSON BROWNE
is usually broken at this stage into two half rings, each one corre-
sponding to one of the original rods; and each half ring is clearly
longitudinally split. The two bars of each half ring frequently
intertwine, thus making a quite remarkable figure (fig. 92 A).
There seems to be no fusion at the points of crossing and no con-
nection of the elements of the two half rings inter se, so that the
figure lends no support to Janssens’ (’09) chiasmatype theory.
After the stage of maximal expansion, a process of condensation
sets in during which the enclosed space becomes smaller and the
bars thicker; the first stage of the process is shown in figure 93 A-F.
The second longitudinal split has become so pronounced that it
entirely separates each half ring into two distinct elements. The
quadripartite nature of the ring is especially noticeable at the
juncture of the two half rings, for here the longitudinal bars
diverge considerably. In the very late prophase, the /-chromo-
some appears as shown in figure 94 A—D; the space enclosed in
the ring has become very much reduced, and the second longitudi-
nal split is still in evidence. The chromosome becomes arranged
on the spindle with its first longitudinal split in the plane of the
equator and its second longitudinal split in the plane of the spindle
axis. Ina side view of the chromosome on a metaphase spindle,
therefore, the second longitudinal split is not visible, since it
lies in the plane of the paper (fig. 95 A, also figs. 29 C, 30 C, 31 C,
33 C,34 A). If however one obtains an end view of the chromo-
some as it lies on the periphery of the spindle, i.e., so that the
place of union of the four elements is in the line of vision, the
second longitudinal split is clearly seen at right angles to the
first (fig. 95 B). Also, in polar view of a metaphase plate, the
second longitudinal split is so clearly marked, that the M-chro-
mosome seems to consist of two distinct parts (fig. 95 C; also
figs. 25-28, 32). The first division plane passes through the
first longitudinal split. The second longitudinal split remains
during the anaphase; figure 96 A is a view of the chromosome
cut obliquely so that one of the components is at a higher level
than the other; figure 96 B shows the compound chromosome
Ma in end view. Figure 97 is a late anaphase showing the bi-
partite nature of the M-chromosome. This is also evident in
MALE GERM CELLS IN NOTONECTA 89
figure 98 where the chromosomes are being pulled on to the second
spindle immediately after the completion of the first division. On
the spindle, the M-chromosome lies so that the longtitudinal
split is in the plane of the equator; the split therefore marks the
line of division (fig. 99). The daughter groups in the late ana-
phase of the second division are shown in figure 100. The M-
chromosome in the second division has rather a peculiar form for
this stage. In the metaphase it looks like a tetrad, and after
division like a dyad, but this bipartite appearance of the single
element has probably no significance. The four chromatids have
been distinct since long before the first division and each has
retained much the same form throughout its history; this form
happens to be a dyad-like structure.
To sum up: the M-chromosome starts as a double rod which
opens out to form a ring; a second longitudinal split appears.
In the first division, it divides along the first longitudinal split
into what were two half rings. In the second division each part
divides along the second longitudinal split which has remained
since it was formed.
It is not only the M-chromosome that forms a ring, but the
next largest chromosome goes through a similar history, as far as
it can be traced. Starting with the open ring which is longitudi-
nally split (fig. 101 A), it passes through stages in condensation,
exactly parallel with those of M (fig. 101 B—D). In the meta-
phase of the first division, in side view, the longitudinal split
is not visible since it lies in the plane of the paper (fig. 101 £),
but in polar view it divides the chromosome in two halves (fig.
101 fF). The line of division coincides with the plane between the
two half rings. In figure 102 is shown a late anaphase group in
which one may distinguish the largest and the next largest chro-
mosomes, both longitudinally split. The split in both cases
marks the line of separation for the second division.
In addition to the two large rings, there is a small ring in the
prophase, which also has a longitudinal split (fig. 103 A, B).
Its history has not been traced.
b. The cross. At the time that the M@-chromosome is leaving
the plasmasome, the other chromosomes are on the nuclear wall
90 ETHEL NICHOLSON BROWNE
in the form of thin double rods (fig. 60). The genesis of the cross
from these bodies is as follows. The two segments open out into
-a V, each arm of which becomes longitudinally split (fig. 104 A).
The arms of the V open out still further so as to form a double
straight rod, the original space between the arms (i.e., the first
longitudinal split) being represented only by the small opening
in the middle of the two bars (fig. 104 B). There are evidently
two methods by which a tetrad may be formed from this figure.
The double bars may condense, while the connection around the
central opening becomes very thin (fig. 104 C, D), or the connec-
tion around the central opening may become pulled out trans-
versely, so as to form the cross-bars of a typical cross (fig. 104
E, F). In this case, half of each long arm and half of each short
arm condense to form one element of the tetrad (fig. 104 G).
The end result is the same in either case, a tetrad is formed in
which the original longitudinal split is represented by the division
line through the short axis and the second longitudinal split by
the line through the long axis. In the metaphase, the tetrad .
lies with its long axis parallel_with the spindle and its short axis
in the plane of the equator. The first division therefore separates
the two components of the original double rod. The vertical
split is usually rather difficult to make out with certainty in the
metaphase but in some eases is quite clear (fig. 104 H). This
split becomes very distinct in the anaphase, and marks the line of
separation of the second division (fig. 104 J).
ce. The double rod. By a process of condensation, the original
double filament forms a thick double rod, the two components
of which lie parallel (fig. 105 A, B). These become united at
one end, and straighten out to form a dyad (fig. 105 C, D). There
is no clear evidence of the presence of a second longitudinal split.
In the metaphase, the chromosome lies with its original longitudi-
nal split in the plane of the equator, so that the first division
separates the two components of the original double rod.
d. XY-pair. In figure 106 A is shown a diffuse cross which
differs from the ordinary cross dascribed above only in the fact
that its longitudinal bars are unequal. It is possible, of course,
that this is an ordinary cross of which part of one bar has been
MALE GERM CELLS IN NOTONECTA Q]
cut off in the section. The same may be said of figure 106 B
where the cross is more condensed. But the probability that
these represent stages in the history of the X Y-pair is suggested
by the occurrence of an unequal tetrad in the late prophase
(fig. 106 C, D). The two small components have, in this event,
arisen from the longitudinally split short vertical bar of the cross,
and the two large components from the split long vertical bar.
The small components represent the Y-chromosome and the
large ones the X-chromosome. Edward (’11) figures in Ascaris
felis the X Y-pair in the prophase quite similar to my figure 106 B.
As stated previously, the X- and Y-chromosomes are separate
in the first division and in figure 106 D from a late prophase they
are already somewhat separated. A preliminary separation of the
members of the X Y-pair therefore takes place first in the prophase
in advance of the other chromosome pairs. This may be corre-
lated with the fact stated previously that the X- and Y-chromo-
somes are frequently found in the second metaphase side by side
instead of joined together to form the usual unequal dyad.
2. Notonecta irrorata. ‘The history of the ring in this species
is the same as that described for the M-chromosome of N. insu-
lata (fig. 107 A—D). ‘The second longitudinal split remains here
also during condensation, and although not seen in lateral view
(fig. 107 C) is frequently visible in polar view of the metaphase
(fig. 107 D;,see also fig. 16). The crosses are likewise similar
to those of the other species (fig. 108 A—H).
3. Notonecta undulata. <A detailed study of the prophase fig-
ures in this species has not been attempted. They are evi-
dently very different from those of N. irrorata and N. insulata
and their irregular shape renders them difficult to trace. Some
of these in the diffuse stage are shown in figure 109 A-—F, and
after ay have condensed in figure 110 A—F.
C. Discussion
The prolonged discussion that has followed Flemming’s original
discovery of the open ring type of bivalent chromosome is even
now not terminated, and the same is true of the cross described
92 ETHEL NICHOLSON BROWNE
by Paulmier and other early observers of the insects. In some
respects Notonecta is not well adapted for the elucidation of this
problem, owing to the difficulties attending the study of the chro-
mosomes during most of the growth period. On the other hand,
this form offers certain advantages in the fact that the formation
of the rings and crosses may be clearly followed during the pro-
phases. The facts here seen seem to leave no doubt that the
rings are formed in essentially the same way as in the Amphibia
and Tomopteris, though their relation to the original spireme can
not be traced.
Those observers (Grégoire, the Schreiners, and many others)
who accept a side-by-side conjugation, or parasynapsis, regard the
ring as originating by the opening out of the longitudinally split
spireme. Those observers (Paulmier, McClung, et al.) who accept
an end-to-end conjugation, or telosynapsis, regard the ring as
originating by the bending together of the split spireme at the two
extremities. In either case, the final result is the same as far as
the real significance of the ring is concerned. The plane between
the two half rings passes through the synaptic point and there-
fore, according to most observers, a division in this plane means a
reduction division, the division in the plane of the ring dividing
it into two whole rings is longitudinal and equational. According
to some observers, e.g., Paulmier, Montgomery, Farmer and
Moore, and also most of the adherents of parasynapsis, the first
division is reductional. McClung and his students, however, be-
lieve that in most Orthoptera it is the second division that is re-
ductional. Bonnevie holds that the ring divides in its own plane
in both divisions and that therefore there is no reduction; the
rings of Enteroxenus, however, have been differently interpreted
by the Schreiners (07).
In Notonecta it is impossible to trace the chromosomes through
the greater part of the growth period when they are aggregated
in a karyosphere, but the evidence seems in favor of the hetero-
homeotypic scheme of Grégoire. The ring is formed from two
parallel rods which probably represent univalent chromosomes.
' The first division separates the ring into two half rings, and is
therefore probably a reduction division. The second division is in
MALE GERM CELLS IN NOTONECTA 93
the plane of the ring and is therefore probably an equation
division.
In many cases the ring goes on to the metaphase spindle with-
out further modification, e.g., in the vertebrates and higher
plants, annelids, ete., but in some cases, e.g., the insects and
copepods, it condenses into a tetrad, as first explained in detail
by Paulmier (’98). In other forms where the ring condenses, the
split is either entirely lost before the second division or is only
faintly indicated by an indentation. In most forms in which no
condensation takes place, the identity of the chromosomes is lost
in the interkinesis, although in Tomopteris, the Schreiners (’06a)
have traced the second longitudinal split with some degree of cer-
tainty to the second division. In Notonecta, the ring condenses
to form a tetrad, but the longitudinal split remains most distinct.
In N. insulata in the case of two chromosomes and in N. irrorata
in the case of one chromosome, in the first metaphase, the chromo-
some is completely divided into two parts and it remains thus
until drawn on the second spindle. The second division follows
directly on the first, the telophase of the first being the prophase
of the second. Since this split can be traced from the early
prophase of the first division to the second metaphase when it
lies in the equatorial plane, there can be absolutely no question
as to its identity with the division line of the second division.
The cross is in principle the same as the ring, as first pointed
out by Paulmier (’98) and as more recently discussed from the
point of view of parasynapsis by the Schreiners (’06 a,b) and
Montgomery (’11), the difference in form being due to the diver-
gence of the two parallel rods from the ends (cross) instead of
in the middle (ring). In the process of condensation the cross
becomes a typical tetrad in contrast to the ring-tetrad, whose
quadripartite nature is not detectable in lateral view since the
second longitudinal split lies in the plane of the spindle. The
similarity between the two is easily seen however, if we compare
a lateral view of the cross-tetrad with an end view of the ring-
tetrad (cf. fig. 104 H with 95 B). The first division plane passes
across the short axis of the cross-tetrad, and if we consider each
original parallel rod as a univalent chromosome this is a reduc-
94 ETHEL NICHOLSON BROWNE
tion division for this figure as well as for the ring-tetrad. The
second division plane coincides with the second longitudinal
split and probably means an equation division. In the case of
the parallel rods, similarly, the first division separates the two
original components. The evidence therefore, is in favor of the
first division acting reductionally for all the autosomes. This
is not true, however, for the X Y-pair, the two components of which
are finally separated in the second division. The fact that the
reduction and equation divisions are reversed in the case of the
X Y-pair and of the X-chromosome has been noted in many other
cases. With this exception and with the exception of the multiple
element of Mermeria (McClung ’05), and possibly a few others
(Blackman 710), all the chromosomes are believed to undergo a
qualitative division at the same time.
There is some evidence from N. insulata for Baumgartner’s
(04) view that the form of individual chromosomes in the pro-
phase is constant. The two largest chromosomes assume a ring
shape, several of the large ones become crosses, one of the large
ones a double rod, and a small one a ring; the smallest ones could
not be traced. The Schreiners (’06) have also concluded that to
a certain point the form of a particular chromosome is constant.
Davis (’08) in the Orthoptera and Blackman (710) in the myri-
apods, hold the same view. On the other hand, Bonnevie (’07)
in Nereis and Foot and Strobell (’05) in Allolobophora, believe
that the form of the chromosome is merely a matter of chance.
Bonnevie states that rings are limited to chromosomes of a certain
size, and Robertson (’08) has attempted to show in Syrbula that
shape is dependent on size. From the fact that both very large
and small rings occur in N. insulata, it seems that in this case,
form is not dependent on size.
VI. MITOCHONDRIA
A. Observations
1. Late growth and division stages. An exhaustive study of this
subject has not been undertaken, but a brief treatment is given
because of a few observations that I have to offer. Owing to the
MALE GERM CELLS IN NOTONECTA 95
precarious nature of the mitochondrial stain of Benda, only a
few clearly differentiated slides were obtained. The general tone
of both cytoplasm and karyoplasm is a pale lavender or rusty
red, the chromatin is a brick red and the mitochondria a deep
purple. In the drawings, the lavender is represented as a pale
grey, the brick red as a darker grey, and the purple mitochondria
as black and dark grey. The earliest stage of N. insulata ob-
tained is shown in figure 111; the chromatin is in compact masses
in the karyosphere, and the mitochondria are scattered through
the cytoplasm. A mass of mitochondria from which project
fibers, is attached to the nuclear wall. A slightly later stage is
shown in figure 112, where the karyosphere is breaking up, and
the nuclear plate of mitochondria has disappeared. From these
figures, it is evident that the mitochondria are of two distinct
kinds, fibers and spheres. The spheres occur chiefly around the
nuclear periphery, and frequently form a complete circle about it.
The fibers usually occur further out in the cytoplasm and tend
to aggregate in several dense clusters. The relation between the
fibers and the spheres is shown in figure 113; the spheres have a
curved rod at the periphery extending about half way around the
circumference, the rest of the sphere is less deeply staining. By
a gradual disappearance of this less dense substance, the sphere
is converted into a fiber, or rather, the fiber which was already in
the sphere becomes free. Whether the fibers always originate
in this way, it is impossible to say. In figure 114 is represented
a metaphase of the first division in side view. As the asters form,
the mitochondria become pushed away from their vicinity
although a few of the fibers take up a position along the astral
rays. In the division stages, the mitochondria are quite evenly
distributed through the cytoplasm between the mitotic figure
and the cell wall, though there is usually a clear area at the periph-
ery of the cell. The spheres and fibers are more intermingled
than during the growth stages. When the cell divides, the mito-
chondria are divided en masse, so that each daughter cell receives
approximately the same amount (fig. 115). There is no evi-
dence that individual fibers or spheres divide, except possibly in
the region of constriction. In the interkinesis the mitochondria
96 ETHEL NICHOLSON BROWNE
are distributed through the cytoplasm, so that in the second divi-
sion they are arranged as in the first. They are divided again
en masse when the cell divides.
2. Early growth stages: nuclear plate. In all three species of Noto-
necta, there is present during the greater part of the growth period,
a characteristic deeply staining mass applied to the nuclear wall.
This takes the chromatin stains of haematoxylin and saffranin,
but is purple when stained according to Benda’s method, and is
evidently of mitochondrial nature. In the earliest growth stages,
the body is more or less spherical, and may be closely applied to the
nuclear wall, or may lie free in the cytoplasm (figs. 43, 44, 63-66).
The mitochondrial body flattens down so as to form a plate on
the outside of the nuclear membrane; it is in this form during the
spireme stage (figs. 45-53, 67, 74). In N. undulata, at the time
when the spireme is disappearing, there is a peculiar bulging of the
nuclear membrane at the place where the nuclear plate is attached
(fig. 68). On the nuclear side, chromatic substance is present in
the swelling, and in the cytoplasm there is a mass of mitochondria
in this region; this differentiation is clear with the Benda stain
(fig. 116). Up to this time, there are practically no mitochondrial
bodies present except the nuclear plate. The mitochondrial
mass which appears outside the nuclear plate is composed of small
spheres and fibers. The bulging very soon disappears, the plate
flattens down again with the membrane (fig. 69), and the mito-
chondria become distributed through the cytoplasm (fig. 117).
The nuclear plate gradually disappears; in N. insulata it becomes
conical or spherical in the later stages and apparently may sepa-
rate from the nuclear membrane (fig. 55).
In haematoxylin preparations, the plate when viewed from
above appears as a spongy mass (fig. 53 B). When viewed from
the side one or two small granules in many cases are seen pro-
jecting from the surface; these may be centrosomes. This is
suggested further by the peculiar modification of the protoplasm
in their vicinity, giving the appearance of an idiozome or attrac-
tion-sphere which lies as a cap over the nuclear plate. The origin
of these granules cannot be conclusively shown, although in the
early stages a small granule may be often detected in the modified
MALE GERM CELLS IN NOTONECTA 97
protoplasm near the mitochondrial body. It seems altogether
probable that the centrosome becomes embedded in the mito-
chondrial mass at an early stage and remains in connection with
it during the growth period. It is of interest to note that the
orientation of the spireme is not toward the nuclear plate but
toward the karyosphere; these two bodies may lie in any position
relative to each other, the karyosphere being usually eccentri-
cally placed.
B. Discussion
Mitochondria have been found in many invertebrates and
vertebrates in both germ and tissue cells by many observers;
Fauré-Frémiet (10), Prenant (10) and Montgomery (’11) have
recently given comprehensive reviews of the subject so that only
a few points will be touched on here. Most commonly, mito-
chondria appear as fine granules which have a tendency to
arrange themselves in rods or chondromites (Meves 00). In some
forms, the mitochondria form long fibers called by Meves (’08)
‘chondriokonts.’ In a few cases the mitochondria have been
described as vesicles with a dense shell, e.g., by Meves in Pygaera
(00) and other forms, by Meves and Duesberg (’08) in the hornet,
and by Gérard (’09) in Stenobothrus. In the latter case the mito-
chondria occur both in the form of vesicles and fibers and bear a
striking resemblance during the growth period to those of Noto-
necta. In Notonecta it is perfectly clear that the fibers are
formed not by chains of granules but directly from the vesicles
by a disappearance of the surrounding substance; the dense shell
described by the above named observers is probably the mito-
chondrial fiber in the sphere. Loyez (’09) has found in the egg
of tunicates that the mitochondrial fiber develops into a yolk
sphere; this is practically the reverse of what occurs in Notonecta.
In division the mitochondria usually appear to be divided en
masse, as they are in Notonecta; but in some of the Protozoa,
according to Fauré-Frémiet (710), the individual mitochondria
divide at the time of the division of the micronucleus. In some
of the Metazoa they form a mantle of long fibers at the side of
the spindle and are divided individually and equally. According’
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 1
98 ETHEL NICHOLSON BROWNE
to the observations of Benda (’02) and others, no mitochondria
occur in the spindle itself. In Notonecta this is true to a certain
extent, but a few mitochondrial fibers lie along the inside spindle
fibers.
In regard to the source of the mitochondria, the evidence
from Notonecta leads to the conclusion that the first mitochon-
drial body is of cytoplasmic origin, that this becomes applied to
the nuclear wall, and that by an interaction of this material and
some of the chromatic material of the nucleus, the numerous
mitochondrial bodies of the later growth period are formed. The
mitochondria are not of nuclear origin in the sense of Hertwig
for the chromatin in the nucleus and the mitochondria outside
are of very different appearance. But it seems quite probable
that their chief elaboration takes place under the influence of the
chromatin since this accumulates in the region where they are
formed and at the time of their formation. It seems to me proba-
ble that the mitochondria are merely early formed cytoplasmic
structures which function in the mature sperm.
The observations on the nuclear plate in the American species
confirm in the main the observations of Pantel and Sinéty on
N. glauca. The ‘archoplasmic vesicles’ which they find scat-
tered through the cytoplasm are no doubt the mitochondrial
spheres described above; these form part of the acrosome of the
sperm. The granular masses, ‘matérial nebenkernien simple’
which forms the principal foundation of the nebenkern, are prob-
ably identical with the masses of mitochondrial filaments which I
have described; the fibrous nature of the masses is brought out
by the Benda stain. ;
VII. SUMMARY
1. The most suggestive result of the foregoing observations
is to show that in the case of Notonecta the change in the number
of the chromosomes from species to species can be explained by
the relations of two particular chromosomes. In N. undulata
these two chromosomes are always separate, in N. irrorata always
united to form a single body, while in N. insulata they may be
separate in the first spermatocyte division, but are united in the
second.
MALE GERM CELLS IN NOTONECTA 99
2. In all three species all the chromosomes are aggregated
during the growth period to form a massive karyosphere, which
consists of chromatic bodies embedded in plasmasome material.
The precise origin of this body is somewhat difficult to ascertain,
but the evidence indicates that it contains at least part of the
early spireme.
3. In the prophases the chromosomes are formed from the kary-
osphere, which gives rise to dense chromatic bodies, which form
diffuse double threads; these condense to form ring- and cross-
tetrads, etc., whose entire history can in some cases be traced.
_ 4. Mitochondria are present in the form of a flat plate in the
early stages, and of spheres and fibers later; the fibers may arise
directly from the spheres. The mitochondria are divided en
masse with cell division.
September, 1912
LITERATURE CITED
ARNOLD, G. 1908 The nucleolus and microchromosomes in the spermatogenesis
of Hydrophilus piceus. Arch. f. Zellforsch., Bd. 2.
Artom, O. 1911. La sistematica del genere Artemia in relazione col numero dei
cromosomi, etc. Biol. Centrb., Bd. 31.
Bateson, W., AND PunNeETT, R. C. 1911 On the interrelations of genetic fac-
tors. Proc. Roy. Soe., vol. 84.
BAUMGARTNER, W. J. 1904 Some new evidences for the individuality of the
chromosomes. Biol. Bull., vol. 8.
Benpa, C. 1902 Die Mitochondria. Ergebn. d. Anat. u. Entw., Bd. 12.
Buackman, M. W. 1903 The spermatogenesis of the myriapods. II. Biol.
Bull., vol. 5.
1905a Ibid. III. The spermatogenesis of Scolopendra heros. Bull.
Mus. Comp. Zool. Har. Coll., vol. 48.
1905 b Ibid. IV. On the karyosphere and nucleolus in the sperma-
tocytes of Scolopendra subspinipes. Proc. Am. Acad. Arts and Sci., vol.
4]. §
1907 Ibid. V. On the spermatogenesis of Lithobius. Ibid., vol. 42.
1910 Ibid. VI. An analysis of the chromosome-group of Scolopen-
dra heros. Biol. Bull., vol. 19.
BonneEvig, K. 1907 ‘Heterotypical’ mitosis in Nereis limbata. Biol. Bull.,
vol. 13.
Bortne, A. M. 1907, A study of the spermatogenesis of twenty-two species of
Membracidae, ete. Jour. Exp. Zodl., vol. 4.
1909 A small chromosome in Ascaris megalocephala. Arch. f. Zell-
forsch., Bd. 4.
5 In order to abbreviate the literature list, many papers have been referred to in
general, and are not included in the list.
100 ETHEL NICHOLSON BROWNE
Boveri, TH. 1904 Ergebnisse iiber die Konstitution der chromatischen Sub-
stanz des Zellkerns. Jena.
1905 Zellenstudien V. Uber die Abhiingigheit der Kerngrésse und
Zellenzahl der Seeigellarven von der Chromosomenzahl der Ausgangs-
zellen. Jen. Zeitschr., Bd. 32.
1909 Uber ‘“‘Geschlechtchromosomen’’ bei Nematoden. Arch. f. Zell-
forsch., Bd. 4.
Braun, H. 1909 Die specifische Chromosomenzahlen der einheimischen Arten
der Gattung Cyclops. Arch. f. Zellforsch., Bd. 3.
Browne, E. N. 1910 The relation between chromosome-number and species
in Notonecta. Biol. Bull., vol. 20.
CHAMBERS, R. 1912 A discussion of Cyclops viridis Jurine. Biol. Bull., vol. 22.
Davis, H. 8. 1908 Spermatogenesis in Acrididae and Locustidae. Bull. Mus.
Comp. Zo6él. Har. Coll., vol. 53.
Epwarps, C. L. 1910 The idiochromosomes in Ascaris meg. and Ascaris lumb.
Arch. f. Zellforsch., Bd. 5.
1911 The sex-chromosomes in Ascaris felis. Ibid., Bd. 7.
Faurft-Fréimiet, M. E. 1910 Etude sur les mitochondries. Arch. d’Anat.
erie. treed lil
Foot, K. anp StropEett, E. C. 1905 Prophases and metaphases of the first
maturation spindle of Allolobophora foetida. Am. Jour. Anat., vol. 4.
Gatss, R. R. 1908 A study of reduction in Oenothera rubinervis. Bot. Gaz.,
vol. 46.
1909 The stature and chromosomes of Oenothera gigas. Arch. f.
Zellforsch., Bd. 3.
GfraRD, P. 1909 Récherches sur la spermatogénése chez Stenobothrus bigut-
tulus. Arch. Biol., t. 24.
Gerassmmow, J. 1901 Uber den Einfluss des Kerns auf das Wachstum der
Zelle. Soc. Imp. Nat., Moscow. ,
GriarpiINnA, A. 1901 Origine dell’oocite e delle cellule nutrici nel Dytiscus.
Inter. Monat. Anat. and Phys., vol. 18.
Gitson, G. 1885 Etude comparée de la spermatogénése chez les arthropodes.
La Cellule, t. 1.
GoupscumipT, R. 1904 Der Chromidial-apparat lebhaft funktionierender Ge-
webzellen. Zool. Jahrb., Bd. 21.
Grecory, R. P. 1911 On gametic coupling and repulsion in Primula sinensis.
Proc. Roy. Soc., vol. 84.
JANSSENS, F. A. 1909 La theorie de la chiasmatypie. La Cellule, t. 25.
Loyrz, M. 1909 Les premiers stades de la vitellogénése chez quelques Tuniciers.
Assoc. Anat. Nancy.
Lusoscu, W. 1902 Uber die Eireifung der Metazoen insbesondere iiber die
Rolle der Nukleolearsubstanz und die Erscheinungen der Dotterbil-
dung. Anat. Hefte, Bd. 11.
Lutz, A. M. 1912 Triploid mutants in Oenothera. Biol. Centralbl., Bd. 32.
McCuiuna, C. E. 1905 The chromosome complex of orthopteran spermatocytes.
Biol. Bull., vol. 9.
1908 Cytology and taxonomy. Kan. Univ. Bull.
MALE GERM CELLS IN NOTONECTA 101
McGi11, C. 1906 The behavior of the nucleoli during oogenesis of the dragon-fly
with special reference to synapsis. Zool. Jahrb., Bd. 23.
Mepes, G. 1905 The spermatogenesis of Scutigera forceps. Biol. Bull., vol. 9.
Meves,F. 1900 Uberdenvonla Valette St. George entdeckten Nebenkern (Mito-
chondrienkérper) der Samenzellen. Arch. Micr. Anat., Bd. 56.
1908 Die Chondriosomen als Triger erblicher Anlagen, ete. Ibid.,
Bd. 72.
Meves, F. unp DursserG, J. 1908 Die Spermatocytenteilung bei der Hornisse,
ete. Ibid., Bd. 71.
Montecomery, T.H. 1901 Astudy of the chromosomes of the germ cells of Meta-
zoa. Trans. Am. Philos. Soc., vol. 20.
1910 On the dimegalous sperm and chromosomal variation of Euschis-
tus with reference to chromosomal continuity. Arch. f. Zellforsch.,
Bd. 5.
1911 Thespermatogenesis of an hemipteron, Euschistus. Jour. Morph.,
vol. 22.
Morean, T. H. 1909 A biological and cytological study of sex-determination
in phylloxerans and aphids. Jour. Exp. Zoél., vol. 7.
Moraan, T. H., AND Lyncu, C. J. 1912 The linkage of two factors in Drosophila
that are not sex-linked.. Biol. Bull., vol. 23.
Nimec, B. 1904 Uber die Einwirkung des Choralhydrats auf die Kern- und
Zellteilung. Jahrb. wiss. Bot., Bd. 38.
PaNTEL, J., ET Sinfity, R. pe. 1906 Les cellules de la lignée male chez le Noto-
necta glauca L. La Cellule, t. 23.
Pautmier, F. C. 1898 Chromatin reduction in Hemiptera. Anat. Anz., Bd.
14.
1899 The spermatogenesis of Anasa tristis. Jour. Morph., vol. 15, sup.
Payne, F. 1909 Some new types of chromosome distribution and their relation
to sex. Biol. Bull., vol. 16.
Prenant, A. 1910 Les mitochondries et l’ergastoplasme. Journ. Anat. et
Physiol., t. 46.
Rosertson, W. R. B. 1908 The chromosome complex of Syrbula admirabilis.
Kan. Univ. Bull.
Scurerner, A. unp K. E. 1906a Neue Studien tiber die Chromatinreifung der
Geschlechtszellen. I. Die Reifung des minnlichen Geschlechts-
zellen von Tomopteris onisciformis. Arch. Biol. t. 22.
1906 b Neue Studien.” II. Die Reifung der miinnlichen Geschlechts-
zellen von Salamandra, ete. Ibid.
1907 NeueStudien. IV. Die Reifung der Geschlechtszellen von Eute-
oxenus Oestergreni. Videns. Selsk. Skr. II.
Sinéty, R. pe 1901 Recherches sur la biologie et l’anatomie des phasmes. La
Cellule, t. 19.
STauFFACHER, H. 1910 Beitriige zur Kenntnis der Kernstrukturen. Zeit. wiss,
Zool., Bd. 95.
Srevens, N. M. 1906 Studies on the germ-cells of aphids. Carneg. Inst. Pub.
no. dl.
1910 Further studies on reproduction in Sagitta. Jour. Morph., vol.
21.
102 ETHEL NICHOLSON BROWNE
Stevens, N. M. 1911 Further studies on heterochromosomes in mosquitoes.
Biol. Bull., vol. 20.
1912 Supernumerary chromosomes and synapsis in Ceuthophilus.
Ibid., vol. 22.
1912b Further observations on the supernumerary chromosomes and
sex ratios in Diabrotica soror. Ibid.
Vespovsky, F. 1911-12 ZurProblemderVererbungstrager. K6n. Béhm. Gesel.
d. wiss. Prag.
Witson, E. B. 1909a Studiesonchromosomes. V. Thechromosomes of Meta-
podius, ete. Jour. Exp. Zoél., vol. 6.
1909 b_ Differences in the chromosome-groups of closely related species
and varieties, etc. Proc. Seventh Internat. Zodl. Congr.
1911 Studies on chromosomes. VII. A review of the chromosomes of
Nezara, with some more general considerations. Jour. Morph., vol. 22.
1912 Studiesonchromosomes. VIII. Observations on the maturation-
phenomena in certain Hemiptera, etc. Jour. Exp. Zoél., vol. 13.
PLATE 16
EXPLANATION OF FIGURES
Notonecta undulata
X 2250
1-2 Metaphase of first division, polar view, showing 14 chromosomes, two small
ones in center.
3 A,B Serial sections of spindle in side view, early anaphase, first division.
4 A, B,C Same, two central pairs arranged linearly.
5 A, B,C Serial sections of spindle in side view, initial anaphase, second divi-
sion, showing 13 chromosomes, XY in center.
6-7 Metaphase of second division, polar view.
8-9 X and Y on separate fibers.
10 A,B Same in complete spindle.
11 X and Y connected by oblique fiber.
12 A,B Sister anaphase groups of second division, from same spindle.
13-14 Spermatogonial groups showing 26 chromosomes.
6 All the figures (except fig. 113) were drawn with the camera lucida. In some
cases, for the sake of clearness, overlying chromosomes have been displaced.
MALE GERM CELLS IN NOTONECTA
ETHEL NICHOLSON BROWNE
|
NH
8 NY VWF
é Be
. Wh
sf @ é : AW,
AAW _ AWS
f | h i\ \\ Ail | \\ \
co] 8 Ka | !
x WW ‘ NW ‘
AN
« » | ‘¢ U vi
NW : NW) fare
~* 9 AN
es ih N
eae AGH?
9 %6 . | a NW
5,05
82> of
“uo?
ii\\ saree Pi e.
WY, i2 oe B
103
PLATE 2
EXPLANATION OF FIGURES
Notonecta irrorata
X 2250
15-16 Metaphase of first division, polar view, showing 13 chromosomes, one
small one in center.
17 A, B,C Serial sections of spindle in side view, early anaphase, first divi-
sion.
18 Components of central pair on different fibers.
19 A, B,C Serial sections of spindle in side view, initial anaphase, second
division, showing 12 chromosomes, XY in center.
20 Metaphase of second division, polar view.
21 A, B,C Complete spindle, X and Y on different fibers.
22 Same, polar view.
23 A,B Sister anaphase groups of second division, from same spindle.
24 Spermatogonial group, showing 24 chromosomes.
104
a ® -
a °« ~?
fs | a O.e?
cae Pp
FZ
EE Fe SS —
— oS
—— a BEE JE Pen
=
X a",
SS a=
PLATE 3
EXPLANATION OF FIGURES.
Notonecta insulata
X 2250
_ 25-26 Metaphase of first division, polar view, showing 14 chromosomes, two
small ones in center.
27-28 Same with 13 chromosomes, one small one in center.
29-31 A, B,C Serial sections of entire spindles in side view, showing one small
pair in center, and compound chromosome Ma, consisting of largest chromosome
and second small one.
32 Polar view, showing same.
106
PLATE 4
EXPLANATION OF FIGURES
Notonecta insulata
xX 2250
33 A, B,C Serial sections of entire spindle in side view, showing two small
pairs in center, components of Ma separate.
34 A, B,C Same, two small pairs arranged linearly.
35-36 Metaphase of second division, polar view, showing 12 chromosomes,
XY in center.
37 A, B Complete spindle, initial anaphase, side view.
38-39 X and Y on separate fibers.
40 A, B Sister anaphase groups of second division, from same spindle.
108
PLATE 5
EXPLANATION OF FIGURES
’ Growth stages. N. insulata
41 Very young spermatocyte. X 2250.
42-43 Formation of scattered masses of chromatin; mitochondrial body lies
outside nuclear membrane in fig. 48. XX 2250.
4445 Formation of spireme; mitochondrial plate in fig. 45. X 2250.
46 Oriented spireme. X 2250.
47 Withdrawal of spireme loops from karyosphere. X 2250.
48-50 Disappearance of spireme loops. X 2250.
51 Same as fig. 41, drawn to scale of figs. 52-53, to show increase in size of
nucleus during growth. X 1350.
52-56 Structure of karyosphere during late eee period. X 1350.
53 B Nuclear plate viewed from above. X 2250.
56 From living material. X 1350.
57 Irregular karyosphere before formation of aster; mitochondrial body has
disappeared. X 1350.
58 Karyosphere rounded, after formation of aster. X 1350.
59 Chromatin leaving plasmasome. X 1350.
110
MALE GERM CELLS IN NOTONECTA PLATE 5
ETHEL NICHOLSON BROWNE
54
oO
«
Oo
ey
oO
56
& &
?
a7 58 59
Browne, del.
111
PLATE 6
EXPLANATION OF FIGURES
Dissolution of karyosphere. N. insulata.
60 Last chromatin mass (M-chromosome) leaving plasmasome, diffuse chro-
mosomes at periphery. X 1350.
61 A—J Chromatin leaving plasmasome. X 1350.
Growth stages. N. undulata
62 Very young spermatocyte. X 2250.
63 Leptotene spireme; mitochondrial body outside nucleus. 2259.
64 Synizesis stage. >< 2250.
65-66 Oriented spireme. XX 2250.
67 Withdrawal of spireme loops from karyosphere; mitochondrial plate. X
2250.
68 Karyosphere fully formed; protrusion from nuclear wall in region of mito-
chondrial plate. X 1350.
69-70 Karyosphere breaking up into balls. 1350.
71-73 Thread formation. 1350.
Growth stages. N. irrorata
74 Looped spireme. XX 2250.
75 Karyosphere broken up into balls. 1350.
76 A, B,77 Thread formation; deeply staining plasmasome in fig. 77. X 1350.
PLATE 6
MALE GERM CELLS IN NOTONECTA
ETHEL NICHOLSON BROWNE
60 .
Sa » _
J, on
62 63 : 64
To F. Ge
; 76
75 By or
113
THE JOURNAL OF EXPERIMENTAL ZOOLOGY. VOL. 14. xo. 1
80
85
PLATE 7
EXPLANATION OF FIGURES
X 2250
Prophases. N. insulata
Early prophase, chromosomes diffuse on nuclear wall, M condensed.
Later prophase, rings and crosses.
Late prophase, chromosomes fully condensed.
Just before metaphase.
N. irrorata
Early prophase, diffuse stage.
Later prophase, rings and crosses.
Very late prophase, spindle forming.
N. undulata
Early prophase, diffuse stage.
86-87 Condensation, irregular figures.
88
Very late prophase.
114
MALE GERM CELLS IN NOTONECTA PLATE 7
ETHEL NICHOLSON BROWNE
O17 ‘ Se
SWAT AWS Z44)\\WXS
Bi WY a ZINN _ Sift INS
PLATE 8
EXPLANATION OF FIGURES
X 2250
Ring tetrad. N. insulata
89 A-D M-chromosome as it leaves plasmasome.
90 A-D Opening out of two bars from middle.
91 A-D Small ring formed; in B, second split has come in.
92 A, B Open double ring.
93 A-F Ring condensing.
94 A-D Condensed ring of late prophase.
95 Ring tetrad in metaphase; A, side view; B, end view; C, polar view.
96 Ring tetrad in initial anaphase; A, slightly oblique view; B, end view, show-
ing compound nature of chromosome.
97 - Late anaphase, each part of M split.
98 M asit is drawn on second spindle.
99 Initial anaphase of second division, showing M dividing along split.
100 Late anaphase.
101 A-F Second largest chromosome forming ring tetrad; D, late prophase;
E, metaphase, side view; F, polar view.
102 Telophase of first division, showing two largest chromosomes longitudin-
ally split.
116
PLATE 8
MALE GERM CELLS IN NOTONECTA
ETHEL NICHOLSON BROWNE
92
~ -@
= ei. 8S =|
— Pe < @ 8 Roo ee Jn
= \ i /Q
Sy - Gwe _ Le
ae L——_——
“oO” 8 S| 0
Reet a-
—— (a)
Cae
NSS
I\@e : 7.)
Ny an
\ I; Vif,’ Fre)
0)
A)
~
‘
11
PLATE 9
EXPLANATION OF FIGURES
X 2250
Tetrad formation. N. insulata (continued)
103 A, B Small ring.
104 A-IJ Formation of cross tetrad; H, in metaphase; J, in anaphase.
105 A-D Condensation of double rod.
106 A-—D Condensation of XY; D, late prophase, elements separating.
Tetrad formation. N. irrorata
107 A-~D Condensation of ring; C, in metaphase, side view; D, polar view.
108 A-H Formation of cross tetrad; D, in metaphase.
Prophase figures. N. undulata
109 A-F Diffuse stage.
110 A-F Condensed stage.
, MALE GERM CELLS IN NOTONECTA
ETHEL NICHOLSON BROWNE
Ee ( \ i i
9 104 A B C 0.
B
' care
E = Wo H 0,
d U >, > ~ Raa
L70%. Ff 4
meet t2¢
y
PLATE 10
EXPLANATION OF FIGURES
x 1350
Mitochondria. Chromatin represented by dark grey, mitochondria by black and
grey. N. insulata
111 Late growth, mitochondrial mass attached to nuclear wall.
112 Later stage.
113. Transition stages between spheres and fibers. Free hand drawing.
114 Metaphase, first division.
115 Division of mitochondria en masse with cell division.
N. undulata
116 Early growth, formation of mitochondria near nuclear bulge.
117. Mitochondria becoming evenly distributed.
120
MALE GERM CELLS IN NOTONECTA PLATE 10
ETHEL NICHOLSON BROWNE
114 He 17
Browne, del.
STUDIES ON THE EFFECTS OF ALCOHOL, NICOTINE
AND CAFFEINE ON WHITE MICE
II. EFFECTS ON ACTIVITY
L. B. NICE
From the Laboratory of Physiology in the Harvard Medical School
THREE FIGURES
CONTENTS
EGG Caller eee cued ie, we ese a Src Nat hac Wop ae Stat vid fe SE ane 123
INAGILVGYO Saeko Sin ee ope a Ci NE Seon ort ee te Se nes 125
Lid: GENES LTACTES Bye area aE a an ee oe ae A a 127
Rr resslaga rs Cala aO Pe NE UMMICES csi Aes 2S ken 2 MA cg Lete es = os Sh oe ane Secs weant 129
Sipe erie Bernlixmees oes pas Peet ee SSA eek: ANCA SEs 3. dl WETS 139
Comparison of the total activity of all the mice............................ 143
Discussion of the effects of alcohol, nicotine and caffeine on the activity of
VTLS THE SS eee UE A oe rte er 146
So DLEP Lag Ac Set URE ead oe a ee eee eee ee 149
Re RRTENAN te NAPS ca hi tt crs oegs eras that Chi Ee SS ig eR we 150
HISTORICAL
During the past twenty years the effects of drugs on the activity
of men and animals have been studied by many investigators.
With but few exceptions however, these studies have extended
over short periods of time and the drugs have been given inter-
mittently.
Alcohol
Hodge (’03) compared the spontaneous exercise of a pair of
alcoholic and a pair of control dogs and found that the alcoholic
female exercised 57 per cent as much as the control female and the
alcoholic male 71 per cent as much as the control male. Their
activity was measured by means of pedometers in their collars.
123
124 L. B. NICE
Stewart (98) gave two grey rats 20 per cent alcohol to drink.
The activity of the alcoholic rats, as measured by revolving cages,
surpassed that of the controls. Thirty per cent alcohol decreased
the activity of white rats.
In investigations on men Lombard (’92), Frey (96), Kraepelin
(99), Rossi (’94) and Schumburg (’99) showed that small doses of
alcohol increased the amount of work done with the ergograph.
Schnyder (’03) observed that alcohol when taken in a fasting
condition, increased the amount of work done with the ergograph
but when taken after or during a meal decreased it. With Hell-
sten (’04) 80 grams of absolute alcohol diminished the amount of
work he could perform. These doses were so large, however, that
they produced disturbances of digestion. Aschaffenburg (’96)
found that wine decreased the efficiency of typesetters. Rivers
(08) considers that the increase of work noted by the above inves-
tigators under the influence of alcohol was due to faulty methods.
The interest of taking the alcohol stimulated the subjects to
extra exertions. His own experiments were carried on with con-
trol mixtures so the subjects did not know when they were taking
alcohol. He states that ‘‘small doses, varying from 5 to 20 ce.
of absolute alcohol have no effect on the amount or nature of the
work performed with the ergograph, either immediately or within
several hours of their administration.”
Nicotine
So far as I can find, no experiments have been made on testing
the effects of nicotine on muscular activity. Tobacco was found
by Lombard (’92), Harley (’94), Féré (04) and Rivers (708) to
decrease the amount of muscular work as recorded by the ergo-
graph, although the pleasurable sensations connected with smok-
ing would be expected to stimulate the subject and thus increase
the amount of work done.
The fact that tobacco is forbidden to athletes when in training
for tests that require great muscular strength shows that it is
generally considered to have a depressing effect on muscular
activity.
EFFECTS OF DRUGS ON WHITE MICE 5
Caffeine
Caffeine was shown by Mosso (’93), Koch (’94), Hoch and
Kraepelin (’96), Schumburg (’99) and Rivers (’08) to increase the
capacity for work with the ergograph. From the results of these
experiments Rivers (’08) says:
This stimulating action persists for a considerable time after the sub-
stance has been taken without there being any evidence, with moderate
doses, of reaction leading to a diminished capacity for work, the sub-
stance thus really diminishing and not merely obscuring the effects of
fatigue. When taken in excess the stimulating action may be so transi-
tory, and followed by so great a decrease that it may legitimately be
spoken of as an accelerator of fatigue.
- METHODS
This study was undertaken to find the effects of alcohol, nico-
tine and caffeine on the spontaneous activity of white mice when
kept under the influence of these drugs in moderate quantities
all the time.
Sixteen male mice eight weeks old were used in the experiment.
They were all descendents of one pair of mice whose offspring had
been inbred for four generations. These mice belonged to the
fourth generation and came from four different lots, each lot being
the young of one male and several females. One mouse from
each lot was placed in each of the experimental lines. Thus
mice of the same sex, the same age, and very closely related were
the subjects of this investigation.
Four lines were carried: one was given alcohol, a second nico-
tine, a third caffeine and a fourth was carried for controls. The
alcohol and nicotine were given in the same proportions as in
preceding experiments, as on these strengths the mice seemed to
keep in good health. The caffeine was given in a 1:300 solution.
Each mouse in the alcohol line was given 35 per cent alcohol
to drink instead of water, and every other day, 3 cc. of 35 per
cent alcohol was added to its crackers and milk.
Each mouse in the nicotine line received 1:1000 nicotine sul-
phate solution to drink instead of water, and had 3 ce. of 1:1000
nicotine sulphate added to its crackers and milk every other day.
126 L. B. NICE
In the caffeine line each mouse drank 1:300 caffeine citrate
solution instead of water, and every other day 3 ce. of 1:300 eaf-
feine solution was added to the crackers and milk.
All of the sixteen mice were given the same food which consisted
of buckwheat and oats, every other day crackers and milk, and
once or twice a week meat.
The experiment continued from November 18 to J une 8. Dur-
ing the winter months the room was heated with hot water and
remained at about 65° F.
To study the spontaneous activity of these mice revolving
cages were devised. ‘These cages are similar to those used by
Stewart (98) and later by Slonaker (’07, 12) in their studies on
rats. The cages are 6 inches wide by 10 inches in diameter, and
made of 8-mesh galvanized wire. Each cage is fastened to an
axle which revolves with the cage. The axle is } inch in diam-
eter and 18 inches long. The ends of the axle are pointed and
set into the end of a bored out set screw forming a pinion which
by reducing friction permits the cages to revolve very easily. By
means of turning the set screw the pinions can be adjusted in
case of wear. The cages are mounted as shown in the accom-
panying photograph (fig. 1).
The revolutions of each cage are recorded by means of an alarm
clock whose balance wheel had been removed. A wire about 6
inches long is attached to the escapement lever of the clock and
to one end of a wooden lever which rests on the axle near one
end of a cage. In one brass hub of each cage two pins are set
on opposite sides of the axle, and 13 inches from it. These pins
are parallel with the axle. As a cage revolves the end of the
wooden lever is raised by each pin in turn causing the clock to
register. Each revolution of a cage corresponds to one second
on the clock.
Each cage is supplied with a nest box made of galvanized tin
2{ inches wide, 2$ inches long and 2 inches deep. These are
swung to the axle by two wire hooks attached to the top of the
nest box near its ends. A tunnel 1} inches square having a wire
mesh floor leads to the opening in one end of the nest box where
the mouse enters. The floor of the nest box is a hinged door.
EFFECTS OF DRUGS ON WHITE MICE OA
On the top of the nest box is a feed box 3 inches long and 1 inch
wide with two compartments, one for grain, the other for crackers
and milk. The feed box is held in place by a spring clip. An-
other clip holds a small wide mouthed bottle which is inverted.
This bottle contains the water or drug which the mouse drinks.
In the mouth of this bottle is a rubber stopper with an opening
through it 3 inch in diameter. Through this opening a glass tube
Fig. 1 Revolving cages and recording clocks
2 inch in diameter inside is inserted. The lower end of the glass
tube is drawn towards a point making an opening + inch in diam-
eter. This device compelled all the mice to drink directly from
the bottles.
THE GROWTH OF THE MICE
The mice were weighed at the beginning of the experiment
when they were eight weeks old. They were weighed once a
week during the next two months and once each month the next
five months.
“YOoM PUuOddS-A}.ITYY
24} Ul PL ‘ON puv YoomM YPANoj-APATYY OY} JO PUD OY} 4B 9 ‘ON ‘Yoo YJUoAoS-AQUIMy VY4 JO pus oY} 4B porp G “ON osnoyy ,
008
0 96
Q 26
G 1G
0&3
0 06
0 &%
0-06
0°61
NICE
B.
L.
O VG
OG
0-96
OGG
SUD LO
SMAUM OF
LV LHDIGM
128
9G
GGG
0°96 GGG
OGG | OG
Or | oF
OSG i) Sere
C96 | O26
10 G6 | ¢ GZ
10 61
G VG G VG
OnGe wale nOkas
O GG 0 GG
O GG G VG
SUDLb SUDLB
- SOLO.
LY GHDIOM | LV GHDTaM
SMUUM ZF
G&G
Gc:
lode
GG
SUuLDeb
SMAAM PZ
LY LHOTAM
O16 GIG | 002 | ¢°6r 0 61 0 ST GST OT! |
0&6 g2ccr.- | SOLGL | O81 0 21 0 91 0 91 CT | - guTayRd
G06 250G |) 0) ol OSI G'st 0 LT Oz Pl ; i
0 F% C26 Orr 0°02 0°02 0 LI Gor | et
0 06 AOE | CO) Koi iP (0) Se G LT (0) 4 0 0G GI |}
02% Oe IOsOs. >| Sst 0'21 0'LT 0°61 TINY suigoong
I ORO} ei sit OSI 0 ZT 0 ST OSI Ol he a
IG 0 GG i 0 16 ¢ 6l 0 0G a)! 6
GGG Ole | OTZ 0 61 0 61 0 61 ST 8
g° 8% 02% | G02 0°61 G61 0°61 aU ener tc
G83 G 1G | 0 61 OSI 0°06 OST “SI 9
G 1G CeOGe ORO 0 61 0 61 0 0G GOT ¢
0 GG G 1G | OLOGe |S) Sil 0 61 0 LT 0 61 t
0°93 Onde | xO0Re 03% 02 0°61 BOG) C1) noses
G &¢ SeGG) | ONUGe se OROG 0 06 OSI C61
O FG G 1G 0 0G | 0 0G 0 61 0 0G 0 61 L |
SULDAD SULDLD SUDLB | SuMDnib SUDO SULDLB SUuDLo
SUUTIM (IZ SMOAAM OT ITEM Z] | SHAAM IT. SMX
LY LHSIGM LV CLHDSIaM) LY LHOIGM | LY LHDIGAM LY LHDOIGAM LV LHOIGM LY LHOIGM
sMnIb UL aovU ay) Jo JY Br9 44
T WITaVL
SM
MOL sx
GMS
5 = ANI
GAsSiOWw
EFFECTS OF DRUGS ON. WHITE MICE 129
At eight weeks the control mice averaged 19.5 grams in weight,
the alcohol mice 18 grams, the nicotine mice 18.5 grams and the
caffeine mice 16 grams. At twenty-four weeks the control mice
averaged 24 grams, the alcohol mice 24 grams, the nicotine mice
22 grams and the caffeine mice 23 grams. At thirty-six weeks
the control mice averaged 27.3 grams, the aleohol mice 19.5
grams, the nicotine mice 22 grams and the caffeine mice 27.6
grams.
The control mice gained 4.5 grams each in the first sixteen
weeks of the experiment and 3.3 grams in the last twelve weeks.
The alcohol mice had gained 6 grams each when twenty-four
weeks old, but had lost 4.5 grams on an average at the end of
tne experiment. The nicotine mice gained 3.5 grams each by the
twenty-fourth week, and had gained no more at the end of the
experiment. The caffeine mice gained the most of all, 7 grams in
the first sixteen weeks and 4.6 grams in the last twelve weeks.
The average gain for the alcohol mice was 1.5 grams, for the
nicotine mice 3.5 grams, for the control mice 7.8 grams and for
the caffeine mice 11.6 grams.
WEEKLY ACTIVITY OF THE MICE
The number of revolutions registered by each clock was recorded
at 12.30 p.m. every day. In the following tables the number of
revolutions that each mouse ran are recorded by weeks.
Mouse No. 1 exhibited its greatest activity from the tenth to
the eighteenth week, then became somewhat less active, increased
again at the twenty-fifth week and then decreased rather gradually
till the end of the experiment. Its maximum run was 139 078
_ revolutions at the thirteenth week and its minimum run was
48 380 revolutions at the thirty-fourth week. Its average run
per week was 92 004 revolutions.
No. 2 exhibited its greatest activity from the eleventh to the
seventeenth week and then had a second period of high activity
from the twenty-fifth to the twenty-ninth week. Its maximum
run was 147 576 revolutions at the fourteenth week and its mini-
mum run was 59 320 revolutions at the thirty-fifth week. Its
average run per week was 95 622 revolutions.
130 Li. BY NICE
TABLE 2
Number of revolutions per week of the mice in the control line
AGE OF MICE IN | MOUSE | MOUSE MOUSE MOUSE
WEEKS no. 1 | NO. 2 NO. 3 no. 4
9 | 80243 «| 78 785 55 718 63 878
10 120 935 82 528 71 041 123 370
11 122 119 118 306 99 070 139 978
12 132593 | 125001 | 100003 120 495
13 139078 | 140830 81004 | 101626
14 126355 | 147576 98966 80 772
15 | Mes 057? |) iso 81 620 84 200
16 94870 | 129 195 80028 94 145
17 | 92 598 119 668 85 370s 109 275
18 — 108 405 91 304 69584 = —s- 129 304
19 | 86990 | 84 900 73100 | 144913
20 | 66117 | 83 987 78 744 105 902
21 | (A287 | 85 647 58 647 135,978
22 | 72890 | 64 281 58119 | 130395
23 82120 | 85955 76740 | 125572
24 | 88454 | 93540 | 54 545 68 147
25 107148 | 109515 | 54 532 126 018
26 91722. | 101 215 60 060 116 095
27 93892 | 74 380 54104 | 105780
28 88755 | 110640 55180 | 100500
29 | 67 071 119 481 41058 | 78 013
30 77 439 63 365 40 881 84 547
31 | 88 028 81 600 47 324 94 140
32 | 85 345 85 102 52164 | 105615
33 | 72 652 74825 | 36155 ° | 91 057
34 48 380 72 453 39: S2ar *| 77 830
35 74 460 59 320s 40746 | 62 100
36 | 69195 — | 63 554 41247 | 57 633
Average.......| 92004 | 95 622 63770 | 102045
No. 3 was the least active of all the controls. It reached its
maximum of 100 003 at the twelfth week and decreased steadily
to the end of the experiment. Its minimum run was 36 155 revo-
lutions at the thirty-third week. Its average weekly run was
63 770 revolutions.
No. 4 was the most active of the control mice. It showed high
activity from the tenth to the thirteenth week, from the seven-
teenth to the twenty-third week, from the twenty-fifth to the
twenty-eighth week and finally at the thirty-second week. Its
EFFECTS OF DRUGS ON WHITE MICE LST
maximum run was 144 913 revolutions at the nineteenth week and
its minimum run was 57 633 revolutions at the thirty-sixth week.
Its average run per week was 102 045 revolutions.
Mouse No. 5 ran its maximum. amount of 125 301 revolutions
during the eleventh week and decreased rather rapidly till the
twenty-seventh week when it died. Its minimum run was 2464
revolutions the twenty-sixth week. Its average run per week
was 57 615 revolutions.
TABLE 3
Number of revolutions per week of the mice in the alcohol line
AGE OF MICE IN MOUSE MOUSE MOUSE MOUSE
WEEKS NO. 5 No. 6 No.7 no. 8
9 57 634 41 370 | 28 430 45 395
10 T4915 | 81694 | 76 230 122 193
11 125 301 58444 | 132113 131.473
12 56 542 15 404 96 680 93 965
13 | 104 745 31 693 142 762 61 901
14 34380 | 32 522 128 495 83 629
15 80 092 63 128 129 393 86 196
16 69 722 79258 | 113,560 83 327
17 74074 86.895 . |. 207-005 83 658
18 56969 69 616 99 138 72 748
19 52 830 71 204 96 925 68 684
20 57087 | 57 318 112 247 81 660
21 51785 54 427 97 093 61 738
22 47297 | 63.423 | 89 195 69 655
23 45628 | 66 502 96 835 77 245
24 43680 | 71 510 16-185} 67 364
25 5683t 75 558 97 210 67 607
26 2464 | 54 630 85 420 66 063
27 27003, | GsailGe si 67 990 29 535
28 69 428 67 600 27 432
29 54997 | 51 083 12 728
30 46795 | 49 736 177
31 70095 | 71 074 25 638
32 | 69675 | 74 660 50 189
33 | 33 820. | 54 982 29 610
34 72 th Ey a) 60 107 3 610
35 | | 58 600 5 012
36 | 62 650 13 001
Avetage.......| 87 615° | 62 435 | 86 934 57 908
1 Nos. 5 and 6 died at the end of the twenty-seventh and thirty-fourth week,
respectively.
132 ; L. B.° NICE
No. 6 ran uniformly low; its maximum run was 86 895 revolu-
tions at the seventeenth week and its minimum run for an entire
week was 15 404 revolutions the twelfth week. It died the thirty-
fourth week. Its average weekly run was 62 435 revolutions.
No. 7 was the most active of the alcohol line. It showed
great activity from the eleventh to the twentieth week. Its
maximum was 142 762 revolutions at the thirteenth week and its
minimum, not counting the first week, was 51 083 revolutions
during the twenty-ninth week. Its average run per week was
86 934 revolutions.
No. 8 ran its maximum of 131 473 revolutions at the eleventh
week. It decreased steadily and near the end of the experiment
made some very small runs, of which 177 revolutions at the thir-
tieth week was the least. Its average weekly run was 57 908
revolutions.
No. 9 showed great activity from the tenth to the twentieth
week, from the twenty-fourth to the twenty-seventh week and
again the thirty-second week. Its maximum run was 134 027
revolutions at the eleventh week and its minimum run, not count-
ing the first week was 67 089 revolutions at the thirty-fifth week.
Its average weekly run was 99 693 revolutions. °
No. 10 was the least active of this line; it rose gradually to its
maximum run of 106 697 revolutions at the fourteenth week and
then decreased to its minimum run of 10 253 revolutions at the
twenty-ninth week, but increased somewhat later. Its average
run per week was 58 843 revolutions.
No. 11 shows a reversed record, for it ran low until its twenty-
seventh week. Its maximum run was 111 132 revolutions at the
thirty-fifth week and its minimum run 14 947 at the thirteenth
week when many of the mice were running their highest. Its
average weekly run was 65 006 revolutions.
No. 12 exercised the most of all the sixteen mice. It
reached its maximum of 192395 revolutions at the thirteenth
week and kept up its great activity throughout the experiment,
not going below 100 000 revolutions until the thirty-fourth ‘week.
Its minimum run was 54 392 revolutions at the twelfth week. Its
average run per week was 124 886 revolutions.
EFFECTS OF DRUGS ON WHITE MICE
TABLE 4
Number of revolutions. per week of the mice in the nicotine line
133
AGE OF MICE IN
MOUSE
MOUSE
MOUSE
MOUSE
WEEKS no. 9 no. 10 no. 11 no. 12
9 25 227 36 370 44 240 75 805
10 112 286 68748 | 31 218 69 467
11 134 027 53430 si 37 226 122 325
12 127 474 75451 | 19 167 154 392
13 107 162 98 414 14 947 192 395
14 119 168 106 697 40 222 175 052
15 122 093 94 483 33 564 145 488
16 133 615 93 275 29 488 160 474
17 122 570 80 278 61 232 135 952
18 123 667 77 205 69 841 150 792
19 110 990 82355 | 72 800 169 845
20 110 770 68310 64 275 134 322
| 94 691 74252 | 51 091 163 060
22 91 428 19 104 71 540 157 709
23 92 467 46 020 49 708 152 201
24 * 102 065 41 973 62 118 139 931
25 108 910 66 335 80 280 148 290
26 106 312 55715 | 66 680 109 654
27 104 340 30 960 99 900 134 390
28 80 052 54 510 100 124 125 528
29 79 221 10 253 72 718 100 308
30 83 505 22 382 68 433 103 102
31 85 974 50 024 53 315 100 280
32 102 207 58 168 110°585 112 674
33 81 638 51 922 108 188 101 015
34 78 748 34 335 95 980 97 185
35 67 089 38 428 111 132 90 943
36 83 434 58 214 100 286 74 228
99 693 58 843 65 006 124 886
Average.......
No. 13 ran low during the entire experiment.
run was 79 613 revolutions at the sixteenth week.
run was 14704 at the thirty-second week.
run was 40 052 revolutions.
No. 14 showed its greatest activity from the eleventh to the
seventeenth week, its maximum run being 135 316 revolutions
at the fifteenth week. After that it decreased rapidly till its
death in the thirty-second week.
Its maximum
Its minimum
Its average weekly
Its minimum run for an entire
134 L. B. NICE
TABLE 5
Number of revolutions per week of the mice in the caffeine line
AGE OF MICE IN MOUSE MOUSE MOUSE MOUSE
WEEKS no. 13 no. 14 no. 15 no. 16
9 27 395 25 160 51 796 51 991
10 43 205 79 935 22 685 79 617
11 41 960 102 615 33 712 143 242
12 55 474 121 164 44 478 83 853
13 34 623 87 722 44 476 141 494
14 53 403 97 O81 66 508 149 307
15 47 585 135 316 €4 128 128 433
16 79 613 110 157 79 980 105 980
17 48 182 104 390 46 347 112 592
18 56 228 55 084 60 028 88 598
19 51 672 82 158 58 585 103 555
20 66 583 65 340 72 885 123 895
21 53 286 66 561 38 847 87 510
22 49 292 83 704 45 256 92 468
23 58 332 61 774 48 048 72 504
24 45 728 61 374 38 114 « 90 872
25 35 560 26 366 47 217 111 157
26 33 015 12 512 40 410 74 635
27 33 246 4 230 43 822 70 652
28 23 602 12 657 38 868 77 981
29 24 812 451 36 690 66 dif
30 18 800 1 182 29 057 44 623
31 21 940 1 007 44 248 54 748
32 14 704 re 51 550 72 710
33 22 393 59 412 66 271
34 26 256 42 240 71 520
35 26 797 41 616 77 685
36 27 790 45 855 54 098
Average..... 40 052 47 036 48 815 89 218
1 No. 14 died in the thirty-second week of the experiment.
week was 451 revolutions at the twenty-ninth week. Its average
run each week was 47 036 revolutions.
No. 15 ran much the same as No. 13, reaching a maximum of
79 980 revolutions at the sixteenth week. Its minimum run was
22 685 at the tenth week. Its average weekly run was 48 815
revolutions.
No. 16 was the most active of the caffeine line. It showed
great activity from the eleventh to the twentieth week and rose
EFFECTS OF DRUGS ON WHITE MICE 135
again at the twenty-fifth week. Its maximum run was 149 307
revolutions at the fourteenth week, and its minimum run, 44 623
revolutions at the thirtieth week. Its average weekly run was
89 218 revolutions.
Comparison of the weekly activity of all the mice
Although all these mice were of the same sex, the same age and
closely related, yet even in the same lines they showed great indi-
vidual variations in their activity.
Some time was necessary for the mice to get accustomed to the
eages. The first week’s runs are low, although the animals had
been in the cages almost a week before the experiment was begun.
All but one of the mice showed their greatest activity in the early
part of the experiment. The decline in the latter part was prob-
ably due to increase in age. Slonaker (’07, ’12) found that white
rats are most active in early life.
The mice may be divided into two types, with one exception.
Type | had a period of high activity falling within the tenth to the
twenty-third week, and a second period of great activity occur-
ring between the twenty-fourth and the thirty-second week. This
would seem to be normal. This group includes six mice, controls
Nos. 1, 2 and 4, nicotine Nos. 9 and 12 and caffeine No. 16.
These six mice were more active than any of the others. None of
them died.
Type 2 exhibited only one period of high activity which fell
within the tenth to the twentieth week; the runs then decreased
steadily to the end of the experiment. This includes nine mice,
control No. 3, all of the alcohol mice, nicotine No. 10 and caffeine
Nos. 18, 14 and 15. Only one of these, aleohol No. 7, is above
the average in activity; all the others fall below the average.
Three of these mice died. It is evident that these mice had less
vitality than those of the first type.
No. 11 of the nicotine line is an exception to these two types.
Its period of high activity did not begin until the twenty-eighth
week, and its maximum run came at the thirty-fifth week.
136 L. B. NICE
Weekly activity of the average of each line
In order to compare the activity of the different lines the aver-
age of each line is recorded in table 6 and figure 2. Some of the
mice in the alcohol and caffeine lines died during the course of
the experiment. In such cases the last entire week of activity
was counted in making the averages.
From table 6 and figure 2 it will be seen that the controls
reached their maximum run of 119 523 revolutions at the twelfth
TABLE 6
Number of revolutions per week of the average of each line
AGE OF MICE IN CONTROL | ALCOHOL NICOTINE CAFFEINE
WEEKS MICE MICE | MICE MICE
9 69656 | . 43195 45410 | 39 063
10 | 99 468 88 758 45429 | 56 110
tt 118-648" 5)" eas ae 752 80 382
12 119523 65648 69121 | 76 242
13 | 115634 | 85 325 103254 | 79 579
14 113 417 69 756 110 289 91 575
15 105263 89 702 98907 | 93 865
16 | 99559 | 86 467 108213 | 93 932
17 101728. | <x 90408 | 100008 | 77 878
18 99 649 74630 | 105 376 64 984
19 | 97276 | 72 411 108997 73 992
20 83 687 77 078 94 419 82 176
21 88639 | 66 261 95773 | 66 551
22 81421 | 67 392 84945 67 680
23 | 92.592 71 552 85099 | 60 164
24 90 639 64822 86522 59 022
25 | 99 303 = FSI 100954 55 075
26 67 273 52 144 84 590 | 40 143
27 82 039 53 547 92307 | 37 987
28 88 769 54 820 90055 | 38 277
29 76 405 39 603 65625 | 32 017
30 66 558 32 236 69355 23 415
31 77 773 55 602 72398 | 30 485
32 79 556 64 841 95908 46 321
33 | 68 672 39 471 85691 | 49 359
34 59 622 25 O11 76562 | 46 672
35 | 59 156 31 806 76898 | 48 699
36 | 57 907 37 825 79040 | 42 581
Average.,.:.,.-| 87 493 63 767 86 321 59 079
137
EFFECTS OF DRUGS ON WHITE MICE
‘SUOTJNJOADL JO TOGUINU OY} SE}YVUIPIO oY} ‘SyooM UT 9oTUT oY} Jo oSe oY} Juosoidad
SUSSIOSG’ OYJ, ‘OUI, YOVS Jo oFvAOAV OY} JO AYTATJOV A[YOOM OY} Burmoys oAIND Z| ‘Shy
9 ve (A>) os 8% 9% ve 14 0z% 81 9| tl ra 01 8
-0
“N (fe aul) auIeyeQ—..—-- =, fj 000008
SEN a i 9U}} @UNOSIN — - —— :
on ‘\ 7 he sae: s SONOS A aterm Uf
r rt No cee eur) }3u0F ——___ “
ORS ed) \ Ss jj 00000%
nat Sa / \ x es ;
me J NS er Gok ifs / a)
\ S f
We \ SS me I 000009
Sit Ay “~ atk MIN \ I /
= . Wid Sore Le \G ys mh { t
me St : . Sires rN eet
ae aN a x CA Sage Se ee 000008
; : es
Bare =i \ Nae) eat se nd ; e
™ <r Ronan 0000001
Meee << ee “a vs
00000 31
138 L.°B. NICE
week and then showed a rather gradual decline in running to the
end of the experiment. Their last run, 57 907 revolutions, was
their least. Their average weekly run was 87 493 revolutions.
The alcohol line rose almost immediately to their maximum of
111 833 revolutions at the eleventh week, but after that declined
rapidly and ran low throughout the experiment. Their last run
was 37 825 revolutions and their minimum 25 011 revolutions
at the thirty-fourth week. Their average weekly run was 63 767
revolutions.
The nicotine line reached their maximum of 110 289 revolutions
at the fourteenth week. They retained a fairly constant level
with a slight decline at the end of the experiment. Their last
run was 79 040 revolutions and their least, if the first two weeks
. are left out of account, 65 625 revolutions at the twenty-ninth
week. Their average weekly run was 86 321 revolutions.
The caffeine line rose slowly to their maximum of 93 932 at the
end of the sixteenth week, and showed a rather steady decline
to the end of the experiment. Their last run was 42 581 revolu-
tions and their minimum run, 23 415 revolutions at the thirtieth
week. Their average weekly run was 59 079 revolutions.
The control and nicotine lines ran close together for a large
part of the experiment. The controls however reached a higher
maximum and reached it sooner than the nicotine mice, but after
the fifteenth week the nicotine line was generally slightly ahead
of the controls. These two lines correspond to type 1.
The alcohol and caffeine lines ran close together, except at the
maximum of the former. After the thirteenth week they ran
entirely below the control and nicotine mice. Most of the time
the alcohol mice ran somewhat higher than the caffeine mice.
These two lines correspond to type 2.
The alcohol mice varied more than any of the other lines.
The average weekly decrease from their maximum run to their
final run was 3360 revolutions, that of the caffeine mice 2566
revolutions, the controls 2566 revolutions and the nicotine mice
1420. The nicotine line kept a more constant level than the
others because both No. 11 and No. 12 were very active in the ~
latter part of the experiment.
EFFECTS OF DRUGS ON WHITE
TOTAL ACTIVIFY
MICE
139
The total activity of each mouse is shown in tables 7, 8, 9
and 10.
AGE OF MICE IN
WEEKS
9
10
11
12
13
14
15
16
36
TABLE 7
Total work in revolutions done by the control mice
MOUSE
no. 1
80 243
201 178
323 297
455 890
594 968
721 323
845 280
940 150
1 032 748
1 141 155
1 228 145
1 294 262
1 368 549
1 441 439
1 523 559
1 612 013
1719 162
1 810 884
1 904 776
1 993 531
2 060 602
2 138 041
2226069 _
2 311 414
2 384 066
2 432 446
2 506 906
2 576 101.
MOUSE
MOUSE MOUSE
NO. 2 NO. 3 No. 4
78 785 95 718 63 878
161 313 126 759 187 248
279 619 225 829 327 226
404 620 325 829 447 721
545 450 406 836 549 347
693 026 505 802 630 119
$24 300 - 587 422 714 319
953 495 667 450 808 464
1 073 163 752 820 917 739 e
1 164 467 $22 404 1 047 048
1 248 567 $95 504 1 191 956
1 232 554 974 248 1 297 858
1 418 201 1 032 895 1 433 836
1 482 482 1 091 014 1 564 231
1 568 437 1 167 754 1 689 803
1 661 977 1 222 299 1 757 950
1 771 492 1 276 831 * 1883 968
1 872 707 1 336 891 2 000 063
1 947 187 1 390 995 2 105 843
2 057 827 1 446 175 2 206 343
2 177 208 1 487 229 2 284 356
2 240 573 1 528 110 2 368 903
2.322 173 1 575 434 2 463 043
2 407 275 1 627 598 2 568 658
2 482 100 1 663 753 2 659 715
2 554 553 1 703 578 2 737 545
2 613 873 1 744 324 2 799 645
2 677 427 1 785 571 2 857 278
The total runs are 1 785 571, 2 576 101, 2 677 427 and 2857 278 revolutions. Three
mice in this line showed great activity while one was much less active. No. 3
ran 62 per cent as much as No. 4, No. 1 ran 90 per cent and No. 2 94 per cent as
much as No. 4.
140 Te Be NIee
TABLE 8
‘Total work in revolutions done by the alcohol mice
AGE OF MICE IN MOUSE | MOUSE | MOUSE | MOUSE
WEEKS no. 5 No. 6 NO. 7 | no. 8
9 57 634 41370 | 28 430 | 25 345
10 132 549 123 064 104 660 167 538
11 257 850 181 508 226773 | 299 011
12 314 392 196 912 333 453 392 976
13 419137 | 228 605 476 415 454 877
14 453 517 | 261 127 604 910 538 506
15 533 609 324 255 734 303 624 702
16 603 331 403 513 847 863 708 029
17 677 405 490 308 | 964 868 791 687
18 734374 | 559 924 1 064 006 864 485
19 | 787 204 | 631128 | 1160931 | 933 169
20 | 844 291 688446 | 1273178 | 1014831
21 896 076 | 742 873 1370271 | 1076569
22 943 373 806296 | 1459466 | 1146 224
23 989001 872 798 1556301 *| 1 223 469
24 1032681 | 944308 | 1633034 | 1290833
25 1 089 512 1019876 | 1730244 | 12358488
26 _ 1091976 | 1073906 | 1815664 | 1424501
27 | 1 094 6761 1137022 | 1883 654 1 454 036
28 1 206 520 1 951 254 1 481 468
29 1261517. | 2002387 | 1494196
30 1 308 312 2052073 | 1494373
31 1 378 407 2 123 147 1520011
32 | 1448082 | 2197807 1570 200
33 1481902 | 2252789 | 1599810
34 1503217! ‘| 2312896 1 603 420
35 2371496 ' 1608 432
36 | 2484146 | 1621 433
1 No. 5 and No. 6 died at the end of the twenty-seventh and thirty-fourth week
respectively.
No. 5 had run 1 094 676 revolutions at its twenty-seventh week, No. 61 503 217 at
its thirty-fourth week, and Nos. 8 and 7, 1 621 483 and 2 434 146 respectively at the
end of the experiment. The record of No. 5 was 58 per cent of No. 7 at the twenty-
seventh week, No. 6 65 per cent of No. 7 at the thirty-fourth week and No. 8 66
per cent at the end of the experiment. Three of this line showed little activity,
while one was much more active.
EFFECTS OF DRUGS ON WHITE MICE 141
TABLE 9
Total work in revolutions done by the nicotine mice
AGE OF MICE IN ASS MOUSE MOUSE MOUSE
WEEKS no. 9 no. 10 no. ll no. 12
9 25 297. ~'| 36 370 44 240 75 805
10 Piolo 105 118 75 458 145 272
11 271 540s 158 548 112 684 267 597
12 399014 | 233 999 131 751 321 989
13 506176 | 332 413 146 698 514 384
14 625 344 439 110 186 910 689 436
15 747 437 533 595 220 474 834 924
16 881052 | 626 870 249 962 995 398
17 1 003 622 707 148 311 194 1 131 350
18 1 127 289 784 353 381 035 1 282 142
19 1 238 279 866 708 453 835 1 451 987
20 1 349 049 935 018 518 110 1 587 309
21 1 443 740 1 009 270 569 201 1 749 369
22 1535168 | 1 028 374 640 741 1 907 O78
23 1 627 635 1 074 394 690 449 2 059 210
24 1729700 | 1 116 367 752 567 2 199 210
25 1 838 610 1 182 702 832 847 2 347 500
26 1944922 | -1258 417 899 527 2 457 154
27 2 049 262 | 1 269 375 999 427 2 591 544
28 2129320 | 1 323 887 1 099 551 2 717 072
29 2208 541 | 1 334 140 1 172 269 2 817 380
ou 2292046 | 1 356 522 1 240 702 2 920 482
31 2378020 | 1 406 546 1 294 017 3 020 762
32 2480227 | 1 464 714 1 404 602 3 133 436
33 2561 865 | 1 516 636 1 512 790 3 234 451
» 34 ; 2640613 | 1 550 971 1 608 770 3 331 636
35 2 707 702 1 589 399 1 719 902 3 422 579
36 2791 136 | 1 647 613 1 820 188 3 496 807
The mice in the nicotine line ran 1 647 613, 1 820188, 2 791 136 and 3 496 807
revolutions during the experiment. One showed very great activity, another
great activity while two were rather inactive. No. 10 ran 47 per cent as much as
No. 12, No. 11 ran 52 per cent and No. 9 80 per cent as much as No. 12.
142
L. B. NICE
TABLE 10
Total work in revolutions done by the caffeine mice
AGE OF MICE IN
MOUSE
MOUSE
MOUSE
MOUSE
WEEKS no. 13 No. 14 no. 15 no. 16
9 Zt ue eos 25 160 51 796 51 991
10 70600 | 104 095 74 481 131 608
11 112560 | 206 710 108 193 274 850
1 168 034 327 874 152671 | 358 703
13 202 657 415 596 207 147 | 500 197
14 256 060 512 677 273 655 | 649 504
15 303 645 547 993 337 783 | 771 937
16 383 258 658 150 417 763 | 883 917
iW 431 440 762 540 464110 | 996 509
18 487 668 837 624 524138 | 1 085 107
19 539 340 919 782 589,722= | 1 188 662
20 605 923 985 122 655 608 1312557
21 659 209 1 051 683 714 455 1 400 067
22 708 501 1 135 387 759 711 1 492 535
23 766 833 1 197 161 807 759 1 565 039
24 $12 561 1 258 535 845 873 1 655 911
25 848 121 1 284 901 893 O80 1 766 068
26 , 881 936 1 297 413 933 490 1 841 703
27 914 382 1 301 643 977 312 1 912 355
28 937 984 1 314 300 1 016 170 1 990 336
29 962 796 1 314 751 1 052 860 2 056 453
30 981 596 1 315 933 1 081 917 2 101 076
31 1 003 536 1 316 940 1 126 165 2 155 824
32 1 018 240 1317 O12 1177 715 2 228 534
33 1 040 633 1237 125 2 294 805
34 1 066 889 1 279 367 2 366 325
35 1 093 686 1 320 983 2 444,010
36 1 121 476 2 498 108
1 No. 14 died at the end of the thirty-second week.
One mouse in the caffeine line showed great activity and three were much less
active. Nos. 13, 15 and 16 ran 1 121 476, 1 366 838 and 2 498 108 revolutions during
No. 14 ran 1317 612 revolutions to its thirty-
No. 13’s record was 49 per cent of No. 16. No. 15 was
the course of the experiment.
second week when it died.
1 366 838
55 per cent and No. 14 was 60 per cent of No. 16 at the thirty-second week.
EFFECTS OF DRUGS ON WHITE MICE 143
COMPARISON OF THE TOTAL ACTIVITY OF ALL THE MICE
TABLE 11
In the following table the mice are arranged in the order of their total activity.
Total activity of all the mice in revolutions and in miles
MOUSE NO. NUMBER OF REVOLUTIONS MILES RUN
SOrmneINE ©... 2... 13 1 121 476 556.06
(CANT (re 15 1 366 838 677.72
ENO 0) Lr 5 1 094 676 (at 27 weeks) 542.77 (at 27 weeks)
Bemaboimne. 2a... 14 1 317 012 (at 32 weeks) 653.01 (at 32 weeks)
Mleonol oo 2s... ek. 6 1 503 217 (at 34 weeks) 745.35 (at 34 weeks)
ZNICGO1C0) Lae 8 1 621 433 803 .96
IMC OEING. 4. ss.c4 25>: 10 1 647 613 $16.94
COTA) Se re 3 1 785 571 885 .34
Nicotine... f)......- iil 1 820 188 898.72
PMG OHO, fice. fe | 7 2 434 146 1206.92
Waiteines.-.. 2)... : 16 2 498 108 1238.65
Clorn ito) | a ee 1 2 576 101 1277 .31
‘Chitin aa 2 | 2677 427 1327.56
Nicobime:.:.....:. >. 9 2791 136 1383 .94
Gomicolewy e. o..:. - AS IF 2.857 278 1416.73
INTEOMNE 2222... .2.. 1 3 496 807 1733.83
PROTA GES oc... 28 af: B15...) (2 O76 622 1010.30
1 Nos. 5, 14 and 6 died at the twenty-seventh, thirty-second and thirty-fourth
weeks respectively.
From table 11 it will be seen that below the average in activity are 3 caffeine
mice, 3 alcohol mice, 2 nicotine mice and one control. Above the average are 1
caffeine mouse, 1 alcohol mouse, 2 nicotine mice and 3 controls. Caffeine No. 13
was the least active of all the mice. Nicotine No. 12 was the most active. No.
13 exercised only 32 per cent as much as No. 12. No. 13 is 54 per cent and No. 12
is 169 per cent of the average.
When the revolutions are reduced to miles the average daily run of all the mice
is 4.93 miles. The average daily run of the mouse that ran the least is 2.84 miles,
and of the mouse that ran the most 8.99 miles. The greatest run in one day of
any mouse is 16.8 miles.
144 L. B. NICE
Total activity of the average of each line
TABLE 12
Total work in revolutions of the average of each line
Sree a IN CONTROL ALCOHOL NICOTINE CAFFEINE
9 69 656 43 195 45 410 39 063.
10 169 124 131 953 90 839 95 173
it 287 742 243 786 177 591 175 555
12 407 265 309 434 246 712 251 797
13 522.899 | 394 759 349 966 331 376
14 636316 464 515 460 255 422 951
15 | 741579 | 554 217 559 162 516 816
16 841138 | 640 684 667 375 610 748
17 942 866 731 092 767 383 688 626
18 1 042 515 805 722 872 759 753 610
19 1 139 791 878 133 981 756 827 602
20 1223 478 | 955 211 1 076 175 909 778
21 1312117 1021472 | 1171948 976 329
22 13935388 | 1088864 | 1 256893 1 044 009
23 1486130 | 1160416 | 1341992 | 110¢0e
24 1576769 | 1225238 1427514 | 1168195
25 1 676 072 1 299 539 1528468 | 1 218270
26 1 743 345 1351683 | 1613058 1 258 413
27 1 825 384 1 405 210 1 705 455 1 296 400
28 1914153 | 1460030 1 795 510 1 334 677
"29 1 990 558 1 499 633 1 861 135 1 366 694
30 2057116 | 1531 869 1930490 | 12390109
31 2 134 889 1 587 471 2 002 888 1 420 594
32 2214445 | 1652362 | 2098796 1 466 915
33 DPS 117 1691833 | 2184487 1 516 274
34 2332739 | 1716844 2261049 | 1562946
35 2391 895 1 747 650 2 337 947 1 611 645
36 2 449 802 1785475 | 2416987 | 1654 226
In table 12 and figure 3 the total activity of each line is averaged.
The average total activity of the caffeine line was 1 654 226
revolutions, of the alcohol line 1 785 475 revolutions, of the nico-
tine line 2 416 987 revolutions and of the control line 2 449 802
revolutions. The caffeine average is 68 per cent of the control
average, the alcohol average is 73 per cent and the nicotine aver-
age is 99 per cent of the control average.
14
MICE
WHITE
DRUGS ON
EFFECTS OF
‘SUOTJNJOADL JO LOQUINU [8404 OY} SoyvUIpAO oy ‘SyooM UT doTUT OY} JO OSV oY yuoserdo.
SUSSLUSGR oT, “oul, YORo Jo osRAoAT oY} JO APLATJOR [VJO} OYJ Burmoys oaang | By
ee ve Ze o€ 82 92 bz te 0z 8| 91 +1 rl 0! 8
od = =r 1 Be 7 —T a t ad ; —0O
A,
7
ae 000002
BUI) BUIAJCD — ~~ — ee g % fe
aul] QUOD} ——- —— oy
~—<——— = ot A
eul ae ——— , Mp f O00000V
ae
Pe
Dp 000009
Gar
Ze
“es 000008
Eee
i ae
ee
Ane 000000!
he aa
Pita 7 .
nite eM ao <A 0000021
ia} Ps is
ime ‘ Pica Uy
Se Be a 00000FI
Se. ce’ a
= se
ihe ce ve 0000091
abies Ve
Eire L
= uk 00008!
oa
Tee 0000002
a
a
VE 00000zz
A,
.
te 00000%z
: 0000092
146 | L. B. NICE
From figure 3 it will be seen that the control line leads through-
out the entire time. The nicotine line excels the caffeine line
at the sixth week and the alcohol line after the ninth week, and
then gradually increases, keeping about an even distance from the
control line, until the last two months when it more nearly ap-
proaches the control line. The alcohol average excels the nico-
tine line for eight weeks, but after that rises much more slowly.
The caffeine average equals the nicotine average for five weeks,
but after that is the lowest of all.
DISCUSSION OF THE EFFECTS OF ALCOHOL, NICOTINE AND
CAFFEINE ON THE ACTIVITY OF WHITE MICE
Control line
The control mice have the highest total activity of all the
lines. They also have the highest average maximum. ‘Three
of them, Nos. 1, 2 and 4, belong to type 1, having a second period
of high activity. These three are well above the average in total
activity. No. 3 is less active, coming below the average. It
belongs to type 2. It weighed 35 grams while the other control
mice weighed 24, 24 and 25 grams respectively. Only one other
mouse, caffeine No. 16, weighed as much as 30 grams. ‘There
may be some correlation between the inactivity of control No. 3
and its weight. None of the controls died. They gained on an
average 7.5 grams throughout the experiment.
Alcohol line
_ Of the mice subjected to alcohol three are decidedly below the
average in total activity, while one is slightly above it. They all
belong to type 2, none having a second period of high activity,
as the six most vigorous mice did. Three of them, Nos. 5, 7
and 8, start out well, but all except No. 7 rapidly decrease in activ-
ity. No. 7 seems to be an average mouse of type 2; however
it loses 20 per cent of its weight during the last month. No. 5
has a maximum of 125 301 revolutions but decreases very rapidly
until its death when twenty-seven weeks old. No. 6 runs low
all its life, its maximum being only 86 895 revolutions. It died
EFFECTS OF DRUGS ON WHITE MICE 147
when thirty-four weeks old. No 8 drops from its maximum of
131 473 regularly until the twenty-seventh week; after that its
runs are irregular and some are very small. At the same time,
from the twenty-eighth week to the end of the experiment it
loses 30 per cent in weight. This great loss of weight and its
small irregular run indicate that it would have died soon.
This line of mice started out well. They reached a high average
maximum of 111 833 revolutions during the eleventh week. This
is much higher than the nicotine or caffeine line at the same time
but a little lower than the maximum of the control line. In fact
the nicotine line never reaches quite as high an average maximum,
and the caffeine maximum is only 93 932 revolutions. But after
that the alcohol line drops abruptly and only once exceeds 90 000
revolutions. Their total activity is 73 per cent of that of the
controls.
The alcohol mice gained 6 grams each up to the twenty-fourth
week, which is more than the control mice had gained during the
same time. After this, however, they lost 4.5 grams on an
average. None of the other lines lost any in weight except the
nicotine mice.
The activity of all the mice in the alcohol line seems to have
been checked and lessened. The viability of all the mice was
weakened, for two died, one was evidently going to die soon and
the fourth lost 20 per cent in weight. This loss in weight oc-
curred during the last month of the experiment. The decreased
activity began to show after the mice were twelve weeks old, but
the loss of weight and lessened viability did not manifest them-
selves until after the mice were twenty-five weeks old. Thus it
appears that alcohol had a markedly injurious effect on the via-
bility an activity of these mice and that these effects were cu-
mulative.
Nicotine line
The mice in the nicotine line show more variations than the
mice in any of the other lines. No. 10 is the least active and
belongs to type 2. No. 11 is below the average in total activity.
It has a very unusual record, running its least at the twelfth week
148 L. B. NICE
when other mice are running their highest and reaching its maxi-
mum at the thirty-fifth week when the activity of all the other
mice is decreasing. No. 9 and No. 12 belong to type 1 and both
exhibit great activity. No. 12 has a remarkably high record,
much higher than any of the other mice in the different lines of
the experiment. It exercised 169 per cent as much as the average
and 122 per cent as much as control No. 4, the second most active
of all the mice.
The average total activity of these mice is almost equal to that
of the controls.
None of these mice died. They gained 3.5 grams during the
experiment, which was less than half what the control mice gained.
The last month they showed an average loss of 1.5 grams.
Nicotine did not seem to affect the health of these mice but
may have slightly checked their growth.
Whether the wide variations in the activity of these mice
were caused by nicotine cannot be known without further experi-
ments. Nicotine may have had a stimulating effect on activity,
shown particularly in No. 12 with its remarkable record and in
No. 9. No. 10 would be an exception to such a theory. No. 11
might be explained as being naturally a very inactive mouse but
that the cumulative effects of nicotine stimulated him to activity.
Or it is possible that all these variations were due to chance.
Caffeine line
Three of the mice in this line have low records of total activity.
Nos. 13 and 15 run very low throughout the experiment. They
have the lowest records of total activity of all the mice used
in the experiment. No. 13 ran 54 per cent as much as the average
and 39 per cent as much as No. 4, the most active mouse in the
control line. No. 14 starts out well but soon decreases in activity
and dies when thirty-two weeks old. These three mice belong
to type 2. No. 16 belongs to type 1 and is above the average in
total activity.
The average activity of these mice is the lowest of all the lines.
Their average maximum weekly run is only 93 932 revolutions
EFFECTS OF DRUGS ON WHITE MICE 149
and their minimum weekly run is 24415. Their average total
activity is 68 per cent of the controls.
The caffeine mice, with the exception of No. 14, gained 11.6
grams each, which was more than any other line gained. But
as they started out the smallest of all the mice and at the end of
the experiments equalled the controls in weight, they apparently
grew normally. No. 14 gained 3.5 grams and lost it again before
its death.
The growth of three of these mice does not seem to have been
affected by caffeine. One, however, seems to have been injured,
for it died when thirty-two weeks old.
Caffeine appears to have decidedly lessened the activity of the
mice.
SUMMARY
1. The control mice gained 7 grams on an average during the
experiment. None of them died. Three were above the average
in activity. Their total activity was greater than any other line.
2. The alcohol mice gained 6 grams on an average up to the
twenty-fourth week, but lost 4.5 grams later. Two died and one
probably would have died soon. Three were below the average
in activity. Their total activity was 73 per cent of that of the
controls. Alcohol appears to have had a markedly injurious
effect on the viability and activity of these mice.
3. The mice subjected to nicotine gained 2 grams each on an
average. None died. Two were below the average in activity
and two above, one being far more active than any other mouse in
any of the lines. Their total activity was 99 per cent of the
controls.
Nicotine apparently did not injure the health of the mice, but
seems to have checked their growth.
Nicotine may have had a stimulating effect on the activity of
three of the mice. Or it is possible that the variations shown in
this line were due to chance.
4. Three of the mice subjected to caffeine gained 11.6 grams
each on an average, but since they started out the smallest of all
150 L. B. NICE
the mice and at the end of the experiment equalled the controls
in weight it appears that they grew normally. One mouse died.
These mice were the least active of all the lines, their total
activity being 68 per cent of that of the controls. Three were
below the average in activity and one was above.
Caffeine seems to have had no influence on the growth of three
of the mice, but apparently had an injurious effect on one mouse,
resulting in its death.
Caffeine seems to have greatly lessened the activity of these
mice.
BIBLIOGRAPHY
ASCHAFFENBERG, GusTAy 1896 Praktische Arbeit unter Alkoholwirkung.
Kraepelin’s Psychologische Arbeiten, Bd. 1. S. 608-626. Leipzig.
Fért, Cu. 1904 Travail et plaisir. p. 316. Paris.
FREY. 1896 Ueber den Einfluss des Alkohols auf die Muskelermiidung.
Mitth. aus Klinken’u. med. Instituten der Schweiz. Reihe 4, Heft 1.
Basel u. Leipzig.
Harvey, V. 1894 The value of sugar and the effect of smoking on muscular work.
Journ. Physiol., vol. 16. pp. 97-122.
Heuusten, A. F. 1904 Ueber den Einfluss von Aklohol, Zucker und Thee auf
die Leistungsfahigkeit des Muskels. Skand. Arch. f. Physiol., Bd. 16.,
S. 139-221.
Hocu, A., AND KRAEPELIN, E. 1896 Ueber die Wirkung der Theebestandtheile
auf kérperliche und geistige Arbeit. Kraepelin’s Psychologische Arbei-
ten, Bd. 1, S. 378-488
Hopae, C. F. 1903 The influence of alcohol on growth and development. In
Physiological aspects of the liquor problem. 1 Ed. by J. 8. Billings.
Houghton-Mifflin, New York, pp. 359-375. 396 pp.
KRAEPELIN, E. 1899 Neure Untersuchungen iiber die psychischen Wirkungen
des Alkohols. Miinch., med. Wochenschrift. Jg. 46., S. 1365-1369.
Kocu, W. 1894 Inaug. Diss. Marburg.
Lomparp, W. P. 1892 Some of the influences which affect the power of volun-
tary muscular contractions. Journ. Physiol., vol. 13. pp. 1-58.
Mosso, U. 1893 Action des principes actifs de la noix de kola sur la contraction
musculaire. Arch. ital. de Biol. t. 19., pp. 241-256.
Nice, L.B. 1912 Comparative studies on the effects of alcohol, nicotine, tobacco
smoke and caffeine on white mice. I. Effects on reproduction and
growth. Jour. Exp. Zoél., vol. 12, no. 1, pp. 133-152.
EFFECTS OF DRUGS ON WHITE MICE 151
Rivers, W.H. 1908 The influence of alcohol and other drugs on fatigue. Arnold,
London, 186 pp.
Rossi, C. 1894 Ricerche sperimentale sulla fatica dei muscoli umani sotto
Vazione dei veleni nervosi. Rivista sper. di Freniatria, vol. 20, pp.
442-480.
ScoumMBuRG. 1899 Ueber die Bedeutung von Kola, Kaffee, Thee, Maté und
Alkohol fiir die Leistung der Muskeln. Arch. f. Anat. u. Physiol.
Abth., Suppl. Bd. S. 289-313.
Stewart, C. C. 1898 Variations in daily activity produced by alcohol and by
changes in barometric pressure and diet, with a description of record-
ing methods. Am. Jour. Phys., vol. 1, pp. 40-56.
Stonaker, J. R. 1907 The normal activity of the white rat at different ages.
Jour. Comp. Neur., vol. 17, pp. 342-359.
1908 Description of an apparatus for recording the activity of small
mammals. Anat. Rec., vol. 2, no. 3, pp. 116-123.
1912a The normal activity of the white rat from birth to natural death
its rate of growth and the duration of life. Journ. Animal Behavy., vol.
2, no. 1, pp. 20-42.
1912b The effect of a strictly vegetable diet on the spontaneous
activity, the rate of growth, and the longevity of the albino rat. Le-
land Stanford Junior Univ. Pub., Stanford Univ., Cal., 36 p.
Scunyprer, L. 1903 Alkohol und Muskelkraft. Arch. f. d. ges. Physiologie,
Bd. 98, 8. 451-484.
©,
é
PA es
z
PO |
Bue
r
q
a
=
=
*
.
*
°
i
j
ol,
2 ¥
é-
~
4
>
; -
2 + paws 4
‘ isk
oo aA eee
L Vis ‘ee
ied: 76 emul ath
+e
+
~ :
> nh — 7
ne P ¥
re
ae
»~
Sis
E
a
?
Tide
= ’
€ =F
=m = ng
rae, 2m
=
a
*—<
1 y
=. Ww
>
Pad
*®
a
” "/*
eA => . ~
é 7 : a ~ S,
© et
> *
- 7 we
5 ean nag
- a ©, =
: ? T
=a |
A i ; Oh
Pe a ha
Perl
STUDIES ON THE DYNAMICS OF MORPHOGENESIS
AND INHERITANCE IN EXPERIMENTAL
REPRODUCTION
V. THE RELATION BETWEEN RESISTANCE TO DEPRESSING AGENTS
AND RATE OF METABOLISM IN PLANARIA DOROTOCEPHALA AND
ITS VALUE AS A METHOD OF INVESTIGATION
C. M. CHILD
From the Hull Zoélogical Laboratory, University of Chicago
TWO FIGURES
CONTENTS
TL, LnGAHRGYOKDYOUNG TRS ct eae Rieke c acinar nts ier RCI re Pichi Uae 2 aCe 154
Il. The physiological resistance method of comparing rates of metabolic
TROVONIOIIL, 0 ho Ss Ane Ee OSES Sd SES loro eee OE tte TeeTeR OCS acre tere coc 155
fem GeneralmoutlimerotathesmethOG ee sem cee sis sues ei ere cutee crea mene 155
2. The relation between death and disintegration in depressing agents 156
3. The relation between length of life in depressing agents and rate of
REQOUD, oS eo eee RO ee Ee oe one AU RETR? teers ac OAC 5 157
Ashe technique of the-direct method. ......<. oc. css beeen at hen 161
Do whe technique of the indirect method '../.-!.- 2. 22224. 6, ane eee es 166
III. The evidence for the relation between resistance to depressing agents
SINC TAL CLO Lee LION MPa ters ci cpe ste-cte, cet saiandisnettos cs im oe as ete eee 167
IPANeSIStanceran Gusti aglOleet e. crt sa che cen “hells einen = erences aera as 167
2. Temperature experiments with alcohol and cyanide............... 171
3. Temperature experiments with benzamid......................-. 182
4. The evidence from animals of different age........................ 190
Heahurcther miscellaneous) evidence.) wee eee eerie era 193
6. The value and the limitations of the resistance method............ 196
IV. The action of depressing agents in general. ................2.20200005- 200
1. The nature of the action of depressing agents on Plaparia.......... 200
2. Certain differences in the action of different reagents.............. 203
SS EPTaT TT 1p 7 RN es Fr TTS = SO AALS eee ea eae ee 205
PRIA NVR TEADO IN Vict tye eh cos os RP ee 0 GR RES SE ER AYE Peeyr sigia "el ae 206
153
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, NO. 2
FEBRUARY, 1913
154 Cc. M. CHILD
I. INTRODUCTION
In the preceding paper of this series (Child ’12) it was shown
that the dynamic processes concerned with the regulatory mor-
phogenesis of Planaria dorotocephala, or at least some of them,
differ in certain respects at different levels along the main axis.
The evidence obtained by subjecting the regulating pieces to
depressing agents and conditions suggests that the existing dif-
ferences or certain of their essential factors are quantitative in
nature. In fact all the results of experiments along these lines
indicate the existence of a gradient in the metabolic processes or
of certain fundamental processes along the main axis, the rate
being highest in the anterior region and decreasing more or less
regularly in the posterior direction.
But more definite evidence is necessary to determine first
whether such a gradient actually exists and second, whether it is
the essential or only an incidental feature of the axial factor in
morphogenesis and function.
Besides the axial gradient my experiments have demonstrated
the existence of various other dynamic differences depending on
the age and nutritive condition and on various external factors,
and in pieces on the size of the piece, the region of the body from
which it came and the degree of regulation or reorganization
which has occurred in it. On the other hand certain morpholog-
ical characters are closely associated with certain quantitative
dynamie factors.
If we can determine positively whether a quantitative factor
exists in any given case and what part it plays we shall have made
a real step in advance in our knowledge of development and
inheritance.
The character of certain results obtained in my work on Pla-
naria with depressing agents such as alcohol, KCN, etc., led me to
spend much time in comparing and analyzing these results with
the aid of various external and internal factors. The result of
this work is the development of a method which enables us to
compare in a general way the rates of metabolic reaction in differ-
ent animals, pieces or regions of the body and so makes it possible
to answer certain questions concerning the dynamics of organ-—
DYNAMICS OF MORPHOGENESIS 155
isms, which have not heretofore been open to investigation. With
this method we are able to determine in a general way where
differences of rate exist and it is often possible also, with properly
devised experiments, to determine what part these differences in
rate play in phenomena of development (Child 713).
Since this method is essential for the most important results to
be considered in later papers, it is necessary, before going farther
in the analysis of .experimental reproduction in Planaria, to de-
scribe the method and its application in detail. The first part
of this paper is therefore devoted to a description of the method,
then follows a consideration of the evidence which constitutes the
foundation of the method, and finally the question as to the
nature of the action of the agents used is briefly discussed and
some interesting lines of investigation are pointed out.
II. THE PHYSIOLOGICAL RESISTANCE METHOD OF COMPARING
; RATES OF METABOLIC REACTION
1. General outline of the method
The method is concerned with the length of life, i.e., the physio-:
logical resistance of the animals or pieces in certain reagents
which decrease metabolism or in sufficient concentration kill.
As will appear below, a relation exists between the length of life
of an animal or piece in a solution of such a reagent of given con-
centration under standard external conditions and the rate of
metabolic reaction in the animal or piece. This being the case,
it becomes possible by standardizing the concentration of the
reagent used and the external conditions, to compare the rates of
reaction in different individuals, regions or pieces. The present
and following papers will show that this method is capable of
wide application and that it gives us a new means of attack on
various problems and opens up certain fields which have hereto-
fore been inaccessible.
Assuming for the moment the correctness of the method, it is
evident that we compare by means of it, not the rates of single
simple chemical reactions but rather the total amounts of the
reactions or processes concerned which occur in a given length of
156 Cc. M. CHILD
time. This total may be made up of many individual reactions
of different or of the same rate, or it may consist of a continuous
reaction with variable or uniform rate. But from the amount
of reaction occurring in a given time we may determine the aver-
age rate for that time. If the method under consideration is
correct, it enables us to determine whether the average rate dur-
ing a given time is greater or less in one case than in another.
‘The rate of reaction’ as the term is used here, is then analogous
to the term ‘rate of flow’ as applied to a current of fluid. It is
simply epee And finally, the method in its present form is
time
only comparative: it serves merely for the comparison of differ-
ent rates without giving any information as to what the rate is in
any case.
2. The relation between death and disintegration in depressing
agents
The reagents most used thus far in work along these lines are
ethyl alcohol and KCN. Enough work has been done with
ether and chloretone to demonstrate that they give results essen-
tially similar to those obtained with alcohol.
The first point of importance is that when an individual or
piece of Planaria dies in a not too highly concentrated solution
of any of these reagents it undergoes disintegration within a
short time after death. The process of disintegration consists
first in the breaking open and disappearance of the body epithe-
hum and second of a gradual swelling and separation of the tissues
until finally nothing remains but minute particles suspended in
the fluid or lying on the bottom of the vessel. The swelling and
increase in translucency of the tissues apparently follows almost
at once after the death of the part concerned and probably results,
at least in part, from the increase in permeability which occurs
at the time of death. After this stage the process does not con-
cern us so closely for it consists merely in the gradual separation
of the dead cells and supporting tissues.
The length of time between apparent death and disintegration
and the rapidity of disintegration vary according to temperature,
DYNAMICS OF MORPHOGENESIS 157
concentration of the reagent, etc., and also with various internal
factors. In worms or pieces in similar physiological condition
and under given external conditions, the time when disintegra-
tion begins and its-rapidity are uniform to a high degree.
The close relation between death and disintegration is shown in
various ways. For example, in KCN 0.001 m. at a temperature of
20°C. distinct movements of a given part can often be induced
within fifteen minutes of the time when disintegration of that part
begins. When concentrations of the anesthetics are used which
are sufficiently high to kill the animals almost at once, disinte-
gration occurs within a few minutes after the animals are placed
in the reagent, sometimes beginning within five or ten minutes.
In cases where different regions of the animal or piece die at
different times, we usually find certain parts of the piece still
showing active movement, while others are already disintegrating
or completely disintegrated.
In general then death is quickly followed by disintegration.
This fact affords an easy means for determining approximately
the time of death of an animal, a region of the body or a piece in
a given concentration of alcohol, KCN, ete.
3. The relation between length of life in depressing agents and rate
of reaction
Early in the course of my experiments it was found that the
length of life in a given concentration of the agent used was dif-
ferent according to the physiological condition of the animals and
in pieces, according to the size of the piece and the region of the
body from which it was taken. Moreover, in different concen-
trations of a given reagent the relation between certain animals
or pieces was not the same. At first the results appeared hope-
lessly complex, but I was convinced that there must be some way
of discovering the factors upon which they depended and finally,
after some eight months of work it became evident, first that a
relation existed between the physiological resistance of the ani-
- mals or pieces and their rates of reaction, and second, that the
character of this relation was dependent upon the concentration
of the reagent used. These relations between length of life or
158 Cc. M. CHILD
resistance, rate of reaction and concentration of reagent are
briefly as follows:
1. In relatively high concentrations in which the maximum length
of life is only a few houra, the length of life or resistance varies
inversely as the rate of reaction: the higher the rate, the earlier death
and disintegration occur and vice versa. This form of the method
which requires relatively high concentrations, I have called the
‘direct resistance’ method.
2. In relatively low concentrations in which the animals remain
alive for days or weeks and in which some degree of acclimatization
occurs the length of life or resistance varies directly with the rate of
reaction except in certain cases where incidental factors modify the
result: the higher the rate, the more complete the acclimatization
and the greater the length of life. This form of the method which
requires low concentrations and long times and which determines
the resistance indirectly through the degree of acclimatization,
I have called the ‘indirect resistance’ method.
3. Between these two extremes of concentration of the reagent
the results vary in character with the concentration and the rate
of reaction. For any two different rates of reaction it is possible
to find a concentration of the reagent in which the resistance will
be approximately the same: above this concentration the relation
is that of the direct method, below, it is that of the indirect ©
method.
These conclusions are drawn from thousands of experiments
with alcohol and KCN by both the direct and the indirect methods.
Ether and chloretone have been used to a sufficient extent to
show that with them the relations are essentially the same.
Undoubtedly there are many. other reagents particularly the
anesthetics, which would give similar results, but since I have
been primarily concerned with certain other problems, I have not
as yet taken the time to test any large number of anesthetics
or other substances with respect to this point. Of the different
substances used, KCN has proved to be the most satisfactory.
The differences in rate of reaction appear more clearly in most
cases in KCN than in alcohol or other anesthetics and the con-
centrations used are so low that various incidental factors are
DYNAMICS OF MORPHOGENESIS 159
practically eliminated. From the three general rules stated
above concerning the relation between resistance, rate of reac-
tion and concentration, it is evident that care must be exercised
to use concentrations sufficiently low or sufficiently high, other-
wise wholly misleading results may be obtained. For example,
if the concentration is too low in a test by the direct method, the
animals or pieces with the higher rate of reaction may become
~ acclimated to some extent and so may live longer than those with
the lower rate which do not become acclimated to any appreci-
able extent. In this case the observed relation between the resist-
ances would be the reverse of what it should be.
On the other hand, if the concentration is too high in a test by
the indirect method, the animals with the higher rate of reaction
may be killed by the direct action of the reagent and so die earlier
than those with the lower rate of reaction, which become accli-
mated to some extent. Here again the results will be the reverse
of what they should be.
These complications connected with the concentration on the
one hand, and on the other the fact that in my earlier experi-
ments only the indirect method, where further complications due
both to internal and external factors may arise, are responsible
for the long time and the large amount of work necessary for the
attainment of definite results.
It is, however, a simple matter to determine the proper limits
of concentration for either the direct or the indirect method.
When the relation between the resistances of the animals or pieces
compared does not undergo inversion with further increase of
concentration, then the concentration is sufficiently high for use
by the direct method and the factor of acclimatization is not
involved. By decreasing the concentration from this point until
the factor of acclimatization does appear clearly we can deter-
mine a concentration for use by the indirect method. In prac-
tice of course concentrations sufficiently far above or below the
critical concentration are used so that there is no danger of con-
fusing the direct and indirect effects of the reagents.
Of the two the direct method is the simpler and requires only
a few hours, where the indirect method may require days or weeks
160 Cc. M. CHILD
or even months. Moreover, the direct method affords no oppor-
tunity for the complication of the results by various factors which
may play a part in the indirect method, e.g., starvation.
In my earlier experiments the indirect method with alcohol was
used because I desired first of all to determine the effect of this
and other substances on morphogenesis: the existence of the rela-
tion between length of life and rate of reaction was discovered
by this method (Child ’11 a, pp. 568 to 571). Later, as I became
more clearly aware of the general significance of this relation, the
effects of different concentrations were compared and the inver-
sion of the relation was discovered. Later still it was found that
more exact results could be obtained with KCN than with alcohol
and this reagent has since been used to a large extent.
For Planaria dorotocephala the following concentrations have
been found to be most satisfactory. For the direct method KCN
0.001 m. serves, although concentrations considerably lower than
this may be used without altering anything but the time factor.
For the indirect method very low concentrations of KCN must be
used, 0.00004 m. or lower, i.e., acclimatization to KCN occurs
only in very low concentrations.
In the case of alcohol a 4 per cent solution of absolute alcohol
is commonly used for the direct method though higher concen-
trations may of course be used. For the indirect method a 1
per cent solution is sufficiently high when loss is prevented. In
my earlier experiments, where there was some loss 1.5 per cent
was used: in tightly closed flasks with very small air space this
concentration is too high.
For ethyl ether 2 per cent or higher serves for the direct method
and a 0.3 per cent or lower for the indirect. Chloretone has been
used only for the indirect method thus far with concentrations
of 0.0014 m. or lower.
This resistance method in general is applicable not only to
Planaria but to any forms in which the skeleton or the connec-
tive tissue are not sufficiently developed or too closely coherent
to permit the occurrence of disintegration very soon after death.
I have obtained results of great interest by this method with
various planarians, with coelenterates and with a number of
DYNAMICS OF MORPHOGENESIS 161
embryos, including those of the amphibia, and it can undoubtedly
be used for many other forms. With the higher organisms the
chief difficulty connected with its use lies in the determination of
the time of death. If the time of death in such forms or in their
parts can be determined by any other simple method than that
of disintegration, there is no apparent reason why it should not be
possible to compare rates of reaction by determining the physio-
logical resistance to certain reagents of such forms or their parts.
In the following sections the practical technique of the two
methods is described.
4. The technique of the direct method
In my own experiments it has proved most convenient to use
lots of ten animals or pieces for each test. When larger numbers
than ten are used the examination of a lot often requires too
much time so that it is difficult to avoid falling behind in keeping
the records. In many cases the comparison of single individuals
or pieces gives perfectly definite and constant results, but the
use of the larger number obviates the necessity of frequent repeti-
tion and also permits slight differences to appear which might not
be discovered in the comparison of single individuals.
In the case of whole animals all ten of one lot are taken from
the same stock, i.e., they have been kept in the same vessel, have
received the same food and have been subjected to the same exter-
nal conditions. Moreover, worms of as nearly as possible the
same size are selected. In the case of pieces the ten of a lot are
from animals of the same stock and the same size and the pieces
are as nearly as possible of the same length and from the same
region of the body. Every one of these factors is important for
the result and in order to obtain definite results it is absolutely
necessary that the material be standardized in this way.
When the animals are selected or the pieces cut they are usually
placed as nearly as possible simultanecusly or at like intervals in
the required concentration of the reagent used: in certain experi-
ments the animals must be placed in the reagent and the pieces
eut in it. I have found Erlenmeyer flasks convenient except
where the animals or pieces are so small that the use of a compound
162 Cc. M. CHILD
microscope is necessary. In most of my experiments by the
direct method 100 cc. Erlenmeyer flasks have been used. After
the worms or pieces are introduced the water is poured off and
they are filled with the solution to be used and corked, leaving
only a small bubble of air beneath the cork to prevent bursting
with slight changes of temperature. In this way loss of the
substance is reduced almost to zero. Moreover, the flasks pos-
sess another great advantage: objects inside the fluid-filled flask
except those on the inner surface of the glass on the side toward
the observer, are magnified to a considerable extent. With a
little practice the flask serves as well as a dissecting microscope
and the condition of small animals or pieces can be seen very
clearly.
In using the direct method, where death and disintegration
occur within a few hours, we may either record only the time of
disintegration, i.e., either of the beginning of disintegration or of
complete disintegration, or we may follow the course of disinte-
gration and compare different stages. Since death and disin-
tegration occur at different times in different regions of the body
the second method gives more satisfactory results: instead of
recording only the beginning or the final stage, it gives a series of
observations on the same material and so not only permits closer
comparison of the different lots but increases the value of the
results obtained.
In the course of my work with this method I have gradually
come to distinguish five stages. The limits of each stage are of
course arbitrary and some of them may, if desired, be further sub-
divided. ‘These stages are as follows:
Stage I. Intact, not showing any appreciable disintegration.
Slage II. This stage is intended to record the first appearance,
of disintegration in any part of the animal or piece. In whole
animals the first traces of disintegration usually appear in the
head reagion, sometimes in the most posterior zodid. At this
stage the disintegration is usually sharply localized and other
parts of the body are intact and often show motor activity.
Slage III. This stage is not very sharply marked off from
Stages IL and IV. It is intended to include that interval between
DYNAMICS OF MORPHOGENESIS 163
Stage IIT and the time when disintegration of the marginal regions
of the body is completed. In Stage III the disintegration has
spread from where it first attacked the animal or piece and new
areas of disintegration may have appeared; the lateral margins
begin to disintegrate but the original form is still maintained.
Parts of the body may still show motor activity at this stage.
Stage IV. The characteristic feature of this stage is the com-
plete disintegration of the marginal regions and the loss of the
original form ‘which follows. The whole animal or the longer
piece usually becomes more or less cylindrical, the shorter piece
a rounded mass. During this stage the epithelium and pigment
disappear, the dorsal surface preceding. This stage passes into
the following.
Stage V. This is the last stage on which observations are made.
It is reached when the epithelium and pigment are completely
gone and when all parts have undergone the swelling and change
in appearance. This stage I believe marks the completion of
the process of dying which began in Stage II. It is followed
within a short time, ranging from a few minutes to several hours,
according to temperature, age of worm, ete., by separation of the
tissues, disintegration of cells and gradual disappearance of the
mass until all that remains are microscopic particles suspended
in the fluid or on the bottom.
In whole worms, where certain regions of the body die much
earlier than others, the different regions pass through the various
stages at different times. The head, for example, may undergo
complete disintegration before the middle regions of the body,
i.e., the posterior regions of the first zodid, are dead. In record-
ing such cases Stage II represents the first appearance of disin-
tegration in any region, Stage III the beginning of marginal dis-
integration behind the head or,in the posterior zodids, and Stage
IV the completion of the marginal disintegration and the change
in shape. Here then Stage III may be disproportionately long
since different regions of the body possess different resistance.
It makes little difference, however, just what each stage includes,
provided it includes the same things in all cases. All that is
desired is to determine as accurately as possible the time of death.
164 : Cc. M. CHILD
In many series with whole animals I have found it desirable to”
divide Stage II into two stages, II a, and II b, IT a including only
the earliest appearance of disintegration at any point and II b
the period when the eyes and cephalic ganglia have become
involved in disintegration, but other parts of the first zodid
have not yet been attacked. In such series two important periods
instead of one are recorded, viz., the death of the head and the
death of the last part to remain alive.
In my experiments by the direct method no attempt has been
made to determine the exact time of the entrance into a given .«
stage. That would of course be very difficult, but we may avoid
the difficulty simply by examining each lot at regular intervals.
At 20°C. half-hour intervals serve for KCN 0.001 m. and in most
cases for 4 per cent alcohol. With higher concentrations or higher
temperatures and sometimes with extremely small pieces shorter
time intervals are often desirablé and with lower concentrations
and lower temperatures the time interval may be increased.
But even with this method of procedure it is of course sometimes
doubtful whether a certain case should be recorded under one stage
or another. My general rule in cases of this sort is to record the
case under the earlier of the two stages in question: before the
next observation it has passed the critical point.
In this manner then we can determine approximately the time
when disintegration begins in each individual or piece and in each
lot and we can also follow its course. The condition of each piece
in each lot is recorded at every period of observation, i.e., com-
monly every half-hour and a comparison of these records brings
out with much greater sharpness than a single record could the
essential differences of different lots.
The record of a comparison between old and young worms is
given in table 1 by way of illustration. Lot 1 consists of ten phys-
iologically young worms 5 to 6 mm. in length, Lot 2 of old worms
18 to 20 mm. in length.
In this table the column ‘Length of time’ gives the length of
time in the reagent in hours and minutes at each observation, the
column headed ‘Lots’ gives the numbers of the different lots
composing the series and the headings I to V under the general
DYNAMICS OF MORPHOGENESIS 165
heading ‘Stages’ indicate the five stages of disintegration. The
first time given in the table shows the length of time in the reagent
when disintegration was first observed in any case. The numbers
in each horizontal column are the numbers of worms of each lot
in each stage at each observation. The conclusion of the obser-
vations on any lot is marked by a broken line as at 4.15 for lot
1 and at 5.45 for Lot 2.
It is evident from table 1 that the young worms begin to die and
disintegrate earlier and that they disintegrate more rapidly than
the old worms. The results as they appear in the table are
perfectly definite and clear and must have some very definite
meaning.
Since the development of the direct method beyond its early
stages, all the records obtained by means of it have been kept in
this form. From these tables the results on any series can be
seen almost at a glance.
TABLE 1
Series 557 I (Nos. 1 and4a). In KCN 0.001 m. 10.15 a.m., October 17, 1912
STAGES
LENGTH OF TIME LOTS ] ]
no
—
oO
De
or
(2) or)
oo
_
2 9 1
ce 1--}-----|----}----|---- L_ — 10
4.15 2 3 -
4.45 2 10
166 Cc. M. CHILD
5. The technique of the indirect method
This method is chiefly useful where it is desired to follow the
morphological features as well as to compare the rates of reaction.
As stated on page 158 above, the results with this method are the
inverse of those obtained by the direct method. There the resist-
ance varies inversely, here it varies directly as the rate of reac-
tion. The results by this method really represent the degrees of
acclimatization to the reagent used.
In my earlier experiments, where the morphological changes
were followed in animals and pieces, this method, usually with
1 to 1.5 per cent absolute alcohol as the reagent, was used exclu-
sively. It was only after the relation between the rate of reaction
and the resistance was discovered that the direct method was
developed.
The data in my ‘Study of senescence and rejuvenescence’
(Child *11 a) were all obtained by this method. The procedure
used at that time is described on pages 538 to 540 of that paper.
Since then I have found it more convenient to use 1-liter Erlen-
meyer flasks instead of the Stender dishes in sealed jars, as there
described. The lots of worms or pieces, ten each in most cases,
are placed in the flasks, which are filled with the reagent and
corked, leaving an air space beneath the cork of some 10 mm. in
depth. The fluid is renewed every four days or oftener, prelimi-
nary experiments having shown that in such a flask containing
well aérated water twenty-five large worms would live for a week
or ten days at a temperature of 20°C. without showing any bad
effects. In the experiments with depressing agents the liquid is
always well aérated at the time the worms are added because of
the thorough shaking necessary for a uniform mixture of the
water and the reagent, moreover the rate of reaction in the ani-
mals is much lower in the depressing medium than in water, so
that it is impossible that lack of oxygen or harmful accumula-
tion of metabolic products should occur at ordinary temperatures
within four days.
In these experiments the lots are examined each day, or in
many cases every forty-eight hours and the number of. worms or
pieces remaining intact is recorded. Attempts to follow stages
DYNAMICS OF MORPHOGENESIS 167
of disintegration are unsatisfactory here because in many cases
disintegration involves certain regions, days or weeks before it
does others and sometimes it involves only those regions of the
body having the lowest rate of reaction, i.e., the posterior region
of the first zodid, and the body may separate into two pieces,
which then undergo some degree of regulation and attain a some-
what higher rate of reaction. In other words, disintegration is
often’ only partial and does not necessarily lead at once to the
death of the whole. For these reasons it has been found best
to record at each observation merely the number of individuals
which remain intact.
_ These records can be most readily presented in graphic form
as in my earlier paper (Child ’11 a). Figure 1 is a reproduction
of figure 2 of that paper. The starting point a of the curves of
the axis of ordinates represents 100 per cent of the number of
worms used, each small space of the cross section paper along the
axis of ordinate representing 2 per cent of the total.
Along the axis of abscissae each small space of the paper repre-
sents one day. The ordinates of the various points of the curves
show the percentage of worms intact at any time during the exper-
iment, the curves being plotted from observations forty-eight
hours apart.
In figure 1, the curve ab represents the resistance to 1.5 per
cent alcohol of fifty physiologically old worms 20 to. 25 mm. in
length, the curve ac, the resistance of fifty younger worms 12 to 15
mm. inlength. These results can also of course be tabulated in
numerical form.
III. THE EVIDENCE FOR THE RELATION BETWEEN RESISTANCE
TO DEPRESSING AGENTS AND RATE OF REACTION
1. Resistance and stimulation
One of the simplest ways of demonstrating the relation between
the physiological resistance to a given reagent and the rate of
reaction is the comparison of stimulated and unstimulated animals.
Various possibilities are open here, the increase in rate of reaction
following a cut or a sudden change of temperature, mechanical
stimulation, ete., may with proper care be demonstrated. But
CHILD
M.
C.
168
‘kep ouo syuoseidor
jovds Yous ovsstosqe jo sIxv oy} SuOTe :40] Yous UT g[eurTUT Jo Laqunu [e}0} BY} JO FU99
god z syuosoidas soded worye8-ssor9 OY} JO aouds [[BVUIs YORE So}VUTIPIO JO SIXB OY} SuOlLy
‘sumz0m AUJY TO paseq STOAIND You “TOOo]B FU09 jad g'] ul vpeydoooyou0p Biivue|d JO
‘gp oAIND T “BLA
gjenpratpur Suno¥ jo ooUV4ySISot OY} ‘op dAINO ‘plo Jo vdUBYSISOL OY}
crane
:
aa
Sl
ea
an
fe
oe aa
EE SEEEEEECEEENEE
aaa
aeaee
sage
ate
coe
as
Ho
ise
seeeece
EEC
aa
Co
ent ta!
reer ase Et tasenetd ese Htc
DYNAMICS OF MORPHOGENESIS 169
these results will be presented in full at another time: at pres-
ent only the results of a comparison between animals stimulated
mechanically to motor activity and those left undisturbed will be
considered. For experiments of this sort KCN is far more satis-
factory than alcohol or other anesthetics since it does not inhibit
motor activity so rapidly and so completely as do they. Thus
far only the direct method has been used in experiments of this
kind, for the indirect method would require frequent stimulation
of the worms during weeks, while with the direct method this is
necessary only for an hour or two.
Series 515. Two lots of ten worms each from the same stock
and of the same size (18 to 20 mm. in length) were placed in 500
ec. Erlenmeyer flasks in KCN 0.001 m. Lot 1 was left undis-
turbed in diffuse daylight and the worms soon came to rest and
remained almost wholly quiet until death. Lot 2 was shaken
every five to ten minutes during two hours and the worms were
dislodged by currents of water from a large pipette. In order to
do this it was of course necessary to uncork the flask, so that
some loss of KCN may have occurred in this case, but as table 2
shows any such loss certainly did not interfere with the result.
The worms were examined every half-hour, but in the table only
the figures for hour intervals are given.
It is evident from the table that Lot 2, the stimulated worms,
begin to disintegrate before Lot 1 and continue in advance of it
during the whole course of the experiment. The difference is not
extreme but is sufficiently large to leave no doubt of its existence.
Two of the worms of Lot 2 live as long as any of the worms in
Lot 1, but the average length of life in Lot 2 is distinctly less than
in Lot 1. The alternate readings omitted from the table show the
same relation in every case.
That stimulation and motor activity increase the average rate
of reaction, there can be no doubt and we see that the worms with
the higher rate die first in the KCN. The only question which
can be raised concerning these results is as to the possibility of
fatigue in the stimulated lot and consequently a lower rate of
reaction. This possibility can undoubtedly be excluded, for I
have at various times attempted to produce fatigue by repeated
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 2
170 Cc. M. CHILD
stimulation of this same sort, i.e., at intervals of five to ten min-
utes, and have not succeeded. Much shorter intervals or prac-
tically continuous stimulation are necessary to accomplish this
result. Moreover, the actual movement in the KCN is very
much less in amount than would occur with the same stimula-
tion in water, because the animals become less and less sensitive
‘to stimulation.
Other similar experiments have without exception given similar
results.
TABLE 2
Worms placed in KCN 0.001 m. 11.00 a.m., March 8, 1912
STAGES
LENGTH OF TIME LOTS | : | é | = — =
ait { kdl : 2
3.30 { ; | “ eerie
4.30 { : | : | . 5 |
sm | cE alretelt ai
}
wee stro core se
7.30 : faye ere
| |
8.30 ‘| : ? | : 6
9.39 : | 2 | s
es (ae aetee Saas
DYNAMICS OF MORPHOGENESIS ve!
2. Temperature experiments with alcohol and cyanide
The rate of metabolic reaction increases with rising tempera-
ture: if, therefore, my conclusions concerning the relation between
resistance to certain depressing agents and the rate of reaction
are correct, we should expect to find that the resistance of worms
as measured by the direct method decreases with rising tempera-
ture and increases with falling temperature and this is actually the
case. ;
But certain possible complicating factors must be considered.
First there is the possibility of increased chemical activity of the
reagent as a factor in the result, in consequence of the increase in
dissociation at the higher temperature. In the case of KCN,
however, and other substances may act in the same way, there
is every reason to believe that the action is primarily upon the
metabolic process at some point or points, rather than upon the
relatively inactive structural substances of the organism. If this
is the case the change in the rate of reaction in the organism, with
a change of 10°C. in temperature, for example, must be of much
greater importance in determining the observed changes in phys-
iological resistance than the change in a 0.001 m. KCN solution
under the same conditions. However, the evidence upon this
point from the temperature experiments by the direct method
alone is not demonstrative because the changes in the solution
and in the organism are in the same direction. But, as will
appear below, the evidence given by the indirect method in
temperature experiments, as well as the evidence from experi-
ments in various other lines leave no doubt that the rate of reac-
tion in the organism is the chief factor in determining the results
at different temperatures obtained by the direct method.
The change in the coefficient of distribution of the substances
used, with change in temperature is another factor which may
play a part in determining the results. According to the theory
of narcosis developed by Overton and Meyer the coefficient of dis-
tribution is the most important factor in determining anesthetic
action. In his third contribution Meyer (’01) shows that the
coefficient of distribution of ethyl alcohol between water and
olive oil increases from 0.026 and 3°C. to 0.047 at 30°C., Le., it
rt Cc. M. CHILD
nearly doubles. In the case of chloral hydrate a much greater
increase occurs. Certain other anesthetics, salicylamid, benza-
mid, monacetin, on the other hand, show a decrease in the coeffi-
cient of distribution with rise in temperature. Meyer finds
that the anesthetic action of all these substances varies at differ-
ent temperatures with the coefficients of distribution. Thus the
anesthetic effect of alcohol and chloral hydrate increases with
rising temperature, while that of salicylamid, benzamid, and
monacetin decreases.
In Meyer’s experiments tadpoles were used and the concen-
tration of the anesthetic which would just produce complete
anesthesia was determined. In other words, these results were
obtained with vertebrates in which the nervous system contains
large quantities of lipoid. Moreover, only the minimal concen-
tration capable of producing narcosis was determined. It is
very probable that in such an experiment with such material the
coefficient of distribution is an important, perhaps the most
important factor in determining the narcotic action, but it does
not necessarily follow that this is the case in all other organisms.
In Planaria, for example, there is no such accumulation of lipoids
in the central nervous system as in vertebrates, and the narcotics
which I have used have little if any greater effect on the nervous
system than on other cells of the body, so far as can be determined.
In fact it was shown in the preceding paper (Child *12 a) that
development of the nervous system might go on in a concentration
of narcotic which practically inhibited completely all other devel-
opmental processes.
Moreover, planarians which have been fed to repletion or dur-
ing a considerable period on mammalian brain tissue do not show
any decrease in physiological resistance to the action of alcohol
in the concentrations used with the direct method: on the con-
trary, their resistance may be much greater than that of other
animals which have been less heavily fed or fed on other kinds of
food. The accumulation of lipoid substance in the bodies of the
‘animals apparently does not decrease their resistance below that
of animals fed on lean beef: the beef-fed animals usually show
the lower rate of resistance because their rate of reaction is higher.
DYNAMICS OF MORPHOGENESIS 173
In the tadpole the concentration of the narcotic in the central
nervous system may be greater than in other parts of the body,
but there is no evidence that this is the case in Planaria. The
assumption of the universal importance of the coefficient of dis-
tribution in the action of narcotics is not justified from experi-
mentation of higher animals alone.
Meyer found that the coefficient of distribution of alcohol
between water and olive oil almost doubled with a rise of temper-
ature from 3° to 30°C., i.e., a change of 27°. For a change of 10°
in temperature the change in the coefficient would then be com-
paratively slight. This brings us at once to the question whether
the change in the coefficient of distribution with change in tem-
- perature is sufficient to account for the observed differences
in physiological resistance. If we find that the differences in
physiological resistance of two similar lots of planarians at tem-
peratures 5° or 10° apart are considerable, we must at least admit
the probability that the coefficient of distribution is not the only
and perhaps not the most important factor involved. The two
following series show the character of results obtained with alcohol
by the direct method.
Series 561. Worms 18 to 20 mm. in length from same stock.
Lot 1, ten worms in alcohol 5 per cent at temperature of 20° to 21°C.
Lot 2, ten worms in alcohol 5 per cent at temperature of 10° to
(Ee ted Oe Z
Worms in alcohol 11.00 a.m., October 29, 1912. Observations
were made every half-hour but only alternate observations are
given in the table.
Table 3 shows that Lot 1 at the higher temperature begins to
disintegrate first and disintegrates much more rapidly than Lot
2 at the lower. In Lot 1 disintegration begins after three hours,
in Lot 2 it begins in one worm after five hours, but not until
after nine hours in the others. In the single case in this lot where
disintegration began after five hours it remained localized in a
very small area on the preocular region of the head and did not
begin to advance further until after nine hours. Undoubtedly
this region was a region of high rate of reaction resulting from
some slight injury in handling the worms. It is possible to induce
174 C. M. CHILD
localized disintegration by means of such slight injuries. Leay-
ing this case out of account, the length of time to the beginning
of disintegration is three times as great in Lot 2 as in Lot 1.
The worms of Lot 1 all reach Stage V within eight hours, those
of Lot 2 require twenty-three hours, i.e., disintegration requires
three times as long in Lot 2 as in Lot 1.
TABLE 3
STAGES
LENGTH OF TIME | LOTS
I «| II III IV Vv
f 1 6 4
3.00 1 ; i |
10
4.00 3 a |
|
/ gy) TERE hee
5.00 9 9 1 |
5 4 1
6.00 ; ‘ i |
|
% 1 | 3 7
ce { 2 9 1
a2 te Pee he Le Sn reeete Ete a) SA eee i
se { 2 9 1
9.00 ; 2 6 4
10.00 ONG 2 8
11.00 2 8 2
12.00 2 5 5
19.00 2 1 2 1 6
20.00 2 2 1 7
21.00 2 2 8
22.00 2 2 8
23.00 eee ae Rae 3 Ts A Seer ES ee 10
DYNAMICS OF MORPHOGENESIS 475
These differences in resistance occurring at temperatures only
10° apart are certainly far greater than we should expect as the
result of a change in the coefficient of distribution. With an
increase of 10° the coefficient of distribution would increase less
than one-third, if its increase is uniform. If this factor alone
were concerned, the difference in resistance of the worms at 10°
and at 20° would be slight.
Series 562 below gives results obtained with a temperature
interval of only 5°C. Here young worms with a higher rate of
reaction were used and also a higher concentration of alcohol, so
that all times are shorter than in the preceding series.
Series 562. Worms 7 mm. in length from the same stock. Lot
1, ten worms in alcohol 6 per cent at 20°C. Lot 2, ten worms in
aleohol 6 per cent at 15°C. Worms in alcohol 2.30 p.m., October
29,1912. Observations every half-hour, table 4.
TABLE 4
LENGTH OF TIME LOTS ae
I baat ur IV v
3 7
asi 2 10
i | 1 5 |
z
hi 2 10
3 | 1 10
2.30 3 : : |
penis 1 10 |
a0 \| 2 10 | |
| |
| ‘poe ee eee ee aa L - 10
3.30 { : ; i
4.00 | 2 1 Bo Herre 4
|
4.30 2 tah = 9c 7
|
5.00 2 | 2 8
5.30 | bE See ee aa erly i 10
BEM enh Duif 3) 2S I eee ee oe ee pa se
176 Cc. M. CHILD
In spite of the earlier beginning and the more rapid progress of
disintegration the temperature difference is perfectly clear. Dis-
integration begins in Lot 1 after one and one-half hours: in Lot 2
at a temperature 5° lower it begins after two and one-half hours.
All the worms of Lot 1 reach Stage V within three and one-half
hours, while those of Lot 2 require five and one-half hours. In
short, the length of time to the beginning of disintegration and
to complete death (Stage V) at the lower temperature is almost
twice that at the higher temperature. Differences in resistance
resulting from differences in the coefficient of distribution would
be scarcely appreciable with a temperature interval of only 5°,
but the observed differences in resistance are almost 100 per
cent. Manifestly they must be due to some other factor.
It is of interest to note that the differences in resistance in
both of the above series are of the same order of magnitude as the
usual temperature coefficient of chemical reaction for the tempera-
ture intervals used.
The following KCN series with a temperature interval of 13° to
15° gives essentially the same results.
Series §21 II. The worms used had been kept for three months
at low temperature: during the first month it fell from 10° to
5°C. and ranged between 4° and 5° during the following two
months. From this stock worms 16 to 18 mm. in length were
taken for the test. Lot 1, ten worms, was placed at a tempera-
ture of 20° for twenty-four hours and then brought into KCN
0.001 m. Lot 2 was kept at 5° in KCN 0.001 m. The tempera-
ture interval is then 15°.
Worms in KCN 0.001 m. 10.30 a.m., March 20, 1912. Table
5 gives hourly observations.
The difference in resistance is striking. After two and one-half
hours four worms of Lot 1 have begun to disintegrate, but Lot
2 does not reach this condition until after ten and one-half hours,
i.e., a little over four times as long.! Lot 1 reaches Stage V
! Owing to the fact that observations were made only once an hour in this series,
the earliest traces of disintegration in Lot 1 were not observed. They probably
occurred about two hours after the worms were placed in KCN, but no record is
made until two and one-half hours. On the other hand, in Lot 2, where disinte-
gration proceeds much more slowly, the earliest traces are recorded at seven and
DYNAMICS OF MORPHOGENESIS 77
within six and one-half hours, Lot 2 within twenty-eight and
one-half hours. In other words, Lot 2 requires a little more than
four times as long as Lot 1 to reach a given stage of disintegra-
tion. As in the cases of the two alcohol series the differences in
physiological resistance in the two lots are of the same order of
magnitude as the temperature coefficient of chemical reaction for
15°. Moreover, they are certainly far greater than any possible
differences in the coefficient‘of distribution for the temperature
interval used.
And finally, the most interesting fact of all is that when the
differences in resistance .as expressed in the times required to
reach a certain stage in the three series, 561, 562, and 521 are
reduced to the same terms, e.g., to a temperature interval of
10°, they are practically identical. This fact is seen in table 6.
The table shows the relation between the times required to reach
given stages in the two lots of each series. For Lot 1 in each case
the time is taken as unity.
The table shows that, at least for the stages considered, the
rapidity of disintegration increases about three times for a rise
in temperature of about 10°C. The close correspondence of the
figures is all the more striking when it is remembered that in
Series 561 large old worms were used, the temperature interval
was 10° and alcohol 5 per cent was the reagent, while in Series 562
small young worms were used, the temperature interval was 5°
and alcohol 6 per cent was the reagent, and finally, in Series 521
II large old worms were used, the temperature interval was 15°
and KCN 0.001 m. was the reagent. This constancy of results
indicates that the essential factor is in all these cases the same.
The temperature coefficient of physiological resistance of Planaria
dorotocephala to alcohol and KCN is then about 0.33 for a rise
one-half hours, but Lot 2 does not reach a stage corresponding to the first recorded
stage of Lot 1, with six worms intact and four beginning to disintegrate until ten
and one-half hours. Evidently then, the full difference between the two lots is
given only by comparison of these corresponding stages. This gives us a relation
of 1 : 4.2 while if we take the first recorded times of disintegration for both lots
we obtain a relation of 1 :3. The discrepancy between these figures is due simply
to the fact that the time of the earliest stages of disintegration in Lot 1 was not
recorded. There can be no doubt that the proportion 1 : 4.2 is more nearly
correct than the other, 1:3.
178
Jae a
2.30 ;
3.30 :
4.30 { :
5.30 3
6.30 a =
7.30 | 2
8.30 | 2
9.30 9
10.30 2
11.30 2
12.30 2
13.30 2
sear ts + te
20.30 2
21.30 2
22.30 2
23.30 | 2
24.30 2
25.30 2
26.30 | 2
27.30 2
28.30 ne
Cc. M. CHILD
TABLE 5
STAGES
i It it IV ai
6 4
10
7 2
10
|
er 8 ren]
10; |
6 4
10 |
biped Sia oe S— = SS See
| 10 ;
9 1
9 1
lege eam.
Gye 4
5 5 |
2 6 2 |
3 9 |
* * * * * * * * *
4 6
3 7 |
|
| 2 6 2
Zs 6 3
|
7 3
6 4
5 5
S 7
Ls — — aE ee eee ee ee 10
DYNAMICS OF MORPHOGENESIS 179
TABLE 6
The relation between the resistances at different temperatures in Series 561, 562 and
521 II, reduced to a temperature interval of 10 C.
BEGINNING OF DISINTEGRATION STAGE V
rue Lotd, 25 Lot 2 Lot 1 Lot 2
SCHIES GOLA scae2 . 2tkee <a: led. 1 | 3.0 1 2.8
Bericd ooo! oni.) 2.. 22). Ae 1 3.33 1 3.33
DeLresi 21ers. Ass. setae 1 2.8 1 2.9
in temperature of 10°C. The fact that this coefficient is the
reciprocal of the temperature coefficient of chemical reaction,
together with the fact that other lines of experiment show that the
resistance depends upon the rate of reaction, justify the conclu-
- sion that this temperature coefficient of resistance is essentially
dependent upon the rate of reaction in the planarian body.
Nevertheless, when substances whose coefficient of distribution
increases with rising temperature are used with the direct method,
the possibility remains in temperature experiments that the
coefficient of distribution may be an important factor in certain
species or certain organs rich in lipoids. It should be possible,
however, to show whether it or the rate of reaction in the organ-
ism is the important factor in a given case by using some of the
depressing agents which show a decrease in the coefficient of
distribution between water and fat with rising temperature. If
we should find that in spite of the lower coefficient of distribution
the resistance of the animals was less at higher than at lower
temperatures, or even the same at both temperatures, we
should be forced to conclude that the coefficient of distribution
was not the essential factor in determining the result. This
question will be considered in the following section.
The results of temperature experiments by the indirect method
also afford further evidence bearing upon the point in question.
By the indirect method with alcohol as reagent the resistance is
less at lower and greater at higher temperatures. It is evident
that only the differences in the rate of reaction in the organism can
be responsible for this result. The animals become acclimated
to the reagent less readily and less completely at the lower than
at the higher temperature.
180 Cc. M. CHILD
On pages 564 to 568 of the paper on senescence (Child 711 a)
the records of such a temperature series were given. Figure 2 is
a reproduction of figure 15 of that paper. The series (Series
140) was prepared as follows (Child ’11 a, pp. 565-6):
The worms used were 15-18 mm. in length and were well nourished
when collected (Nov. 25). They were kept in the laboratory 24 days at
a temperature of 18°-22°C. without food, during which time they used
up most or all of their reserves. At the end of this period a stock of
about 100 worms was placed in dishes surrounded by running water at a
temperature of 8°-10°C. From this stock two sets of 10 worms each
were taken after 12 days, 22 days, 37 days and 65 days at the low tem-
perature. One of these sets in each case was placed in alcohol 1.5 per
cent at the same temperature, 8°—-10°C., at which the worms had been
kept, the other in alcohol 1.5 per cent at room temperature, 18°-22°C.
The two curves in figure 2 are plotted from these two lots of
forty worms each. Since each lot of forty worms is made up of
worms taken at different periods during sixty-five days, the
curves do not show the changes in resistance which occurred in
the stock during this period, but since these are the same at any
given time for all the worms of the stock, it is better for the pres-
ent purpose to eliminate them. The two curves then show simply
the effect of placing worms which had been kept in water at a
certain temperature, in alcohol at different temperatures. The
lower curve ac shows the resistance of the worms at the lower
temperature, the upper curve ab the resistance at the higher tem-
perature. As in figure 1, each small space of the cross section
paper along the axis of ordinates represents 2 per cent of the total
number of worms used, and each space along the axis of abscissae
represents one day. The curves show that the worms at the
higher temperature possessed greater resistance.
In another similar series the stock was kept at room tempera-
ture, 18° to 22°C. and from this parallel lots were taken at inter-
vals, one lot in each case being placed in alcohol at 18° to 22°C.,
the other in alcohol at 8° to 10°C. In this series, as in Series 140,
the worms in alcohol at the lower temperature always showed
less resistance than those at the higher.
There can be no doubt, I think, that the difference in resistance
under these conditions is due to the difference in the rate of reac-
181
MORPHOGENESIS
DYNAMICS OF
‘Aep ouo syuoseidea coeds
Yow OVSSIOSG’ JO SIX” OY} Buoys :4O] Yow ur s[vurtuR jo coquinu ][v}0} oYy Jo yuo aod Z syuosoido.t
todvd u01}008-ssor9 04} Jo oovds [[RUIS YOR SoJVUIPIO Jo STIX” OY} SUOTY ‘oANyesodut9} LOMO] OY} YR
99UBISISOL OY} ID OAINO ‘LOYSIY OY} 7B oOUVSISOL OY} ‘QD OANA) “SULIOM A410} WO pose ST AAAND YORiT
‘sounyetodure} JudIeyIp 4B joYyoo]e yuoo aod Gy ul vpeydoooyodop viIvUB[_ JO ooUBISISod OUT, FJ “SL
S2Ge0 00 R00 SO G00 0050000 0e02Ge Ge 0000 Re500 5s ne0ssse05e0e 050550 se0eneee se esscueeccesauee
te SN
an 4 4 AN meses 7 CN Fer ee
Srtretiest TUE ie area NEE
oe ae [ ae HEEL Ho : ian Bae ze me
se re eee ee eae eee HEE EEE
SUBSE ECorae sn se) to os in| esl sl estel| mel :
Ed lfued] ff etd lee feed eased adi ia enme SOU add nnreeeDfageziee
Sesssdabvaedeautstanttt onuttfaeet faces feces fece# facut foasinicouit anit weit
iE ee eee ee ee eee eee
Efze ceed ease et laet eat ane eed fane ad tae eet facet as tet eee
Hae eid faee eH tnt eae td eee tae EEN
Hace ai nae ee? el fae Hee aE Haein aase
TH att aRrastirsszsecee SSEctaE
sivtastteasttosttocst
ile alia E
os
DHoenES
age
a
aap
aaa
Be ao
Cc]
ries
im
t
ro
pola
conga
Sneee
Henn
a C]
aBp=aul
aaaee
HH
YZ
7
a0
He
Hoe
patie
noes
25a
1 \f T |
ERSRSEaS
a5:
san
cy
tt
EHH
182 Cc. M. CHILD
tion in the organism at the different temperatures. If the coeffi-
cient of distribution were the most important factor, the resist-
ance would be lower at the higher temperature. As a matter of
fact, the animals with the higher rate of reaction determined by
the higher temperature become more readily and more completely
acclimated and therefore live longer.’
Taken as a whole, the evidence from temperature experiments
with alcohol and KCN for the existence of a relation between
physiological resistance and rate of reaction is highly conclusive.
Within certain limits, the higher the temperature, the higher the
rate of reaction, the less the resistance by the direct method, and
the greater the resistance by the indirect method. —
How far these relations will hold for the higher animals, and
particularly those in which the nervous system contains a large
amount of Hpoid material, can be determined only by experiment.
At present I can state that they hold for all flatworms that have
been tested, some five species, for Corymorpha palma among the
coelenterates and for early embryonic stages of various species,
so far as tested, including the amphibia.
3. Temperature experiments with benzamid
On page 179 above, attention was called to the bearing upon the
question of the relation between resistance and rate of reaction
of experiments with a substance whose coefficient of distribu-
tion between water and fat increases as the temperature falls.
In temperature experiments with such a substance the animals
should die earlier at the lower temperature if the coefficient of
distribution or some other factor similarly affecting its concen-
tration in the cell is the chief factor in its effect, but if the rate of
reaction in the organism is the chief factor, then the animals
should die earlier at the higher temperature in spite of the lower
coefficient of distribution at this temperature.
The coefficient of distribution of benzamid as determined by
Meyer is 0.672 at 3°C. and 0.487 at 36°C. (Meyer ’01, pp. 341-
344). In accordance with this difference in the coefficient of
distribution, Meyer found that the minimal concentration which
would produce complete narcosis in tadpoles was si0 m. at
DYNAMICS OF MORPHOGENESIS 183
3°C. and soa m. at 30°C., ie., the narcotic effect was greater at
the lower temperature. In his experiments Meyer determined
merely the minimal concentration that would produce complete
narcosis and apparently did not attempt to determine what
would occur with higher concentrations.
My own results with different concentrations are as follows:
with concentrations near or somewhat above the narcotic mini-
mum (viz., 0.02 m.) the narcotic effect is greater and disinte-
gration occurs earlier at the lower (10—11°C.) than at the higher
temperature (20—-21°C.) With concentrations considerably higher
than this, on the other hand, the temperature relations are just
the reverse; the narcotic effect is greater and disintegration occurs
earlier at the higher than at the lower temperature. With these
higher concentrations of benzamid the temperature relations are
the same as with the higher concentrations of aleohol and KCN
and with the lower concentrations they are the same as with
the lower concentrations of aleohol and KCN. In spite of the
higher coefficient of distribution at the lower temperature, the lower
rate of reaction in the organism at this temperature determines with
higher concentrations of benzamid a higher resistance than at the
higher temperature where the rate of reaction is higher.
The question at once arises as to how far the coefficient of dis-
tribution and how far other factors are concerned in these rela-
tions. When low concentrations of alcohol and KCN are used
the animals with the higher rate of reaction become more readily
and more completely acclimated and so die later than those with
the lower rate. But such concentrations of alcohol and KCN are
below the minimum which produces complete narcosis. In the
case of benzamid the narcotic effect with minimal concentrations
is greater at low and less at high temperatures. It is certain
that the acclimatization factor is not involved in this primary
narcotic effect. There are, however, indications that acclimati-
zation to benzamid .occurs very rapidly; in sufficiently low con-
centrations the animals show signs of recovery from the partial
or complete narcosis in less than twenty-four hours at 20°C.
In disintegration experiments, therefore the acclimatization fac-
tor may play a part as it does with alcohol and KCN, but the
184 Cc. M. CHILD
temperature relations of the primary narcotic effect must be due
to some other factor.
Since the temperature coefficient of distribution of benzamid
changes with change of temperature, in the opposite direction
from that of alcohol the next step is to determine how far this
factor is responsible for the observed results.
When the concentration of the solution is below a certain limit
the benzamid does not enter the cell with sufficient rapidity or
in sufficient quantity to produce any appreciable physiological
effect; it may be oxidized or disposed of in some other way as
rapidly as it enters. Assuming that the substance enters the
cell through the lipoids, it is evident that with certain low concen-
trations the presence or absence or the degree of the narcotic
effect may depend on the coefficient of distribution. With such
concentrations, the higher the coefficient of distribution, the higher
the concentration of benzamid in the cells and the greater the
effect. On this basis we might interpret the greater narcotic effect
of the benzamid and the lower resistance of the worms at the
lower temperature when concentrations near the narcotic mini-
mum are used, as due to the higher concentration of the substance
in the cell at the lower temperature in consequence of the higher
coefficient of distribution.
But when concentrations considerably above the narcotic
minimum are used it is evident that in spite of differences in the
coefficient of distribution, within a wide range of external condi-
tions the concentration of the substance in the cell is sufficient
in all cases to produce complete narcosis and sooner or later death.
Tn such concentrations the coefficient of distribution must become
a factor of minor importance and the rate of reaction in the organ-
ism is the chief factor. If these conclusions are correct, it is
evident that the coefficient of distribution is not the essential
factor in narcotic action in any case. The lipoids may be largely
responsible for the entrance of the narcotics, at least in certain
cases, but their effect after they have entered is apparently pri-
marily chemical. This conclusion is in essential agreement with
the most widely accepted theory of narcosis.
With a closer scrutiny of the temperature experiments, how-
ever, certain obstacles to this simple interpretation appear. If
DYNAMICS OF MORPHOGENESIS 185
the temperature coefficient of the rate of chemical reaction in the
organism is of the same order of magnitude as that found for many
other chemical reactions then the rate of reaction in the organism
must increase two to three times with a rise of 10°C. On the
other hand, according to Meyer, the coefficient: of distribution .
decreases only about one-third with a rise of about 30°C. In
short the temperature changes in the rate of chemical reaction
are so much greater than those of the coefficient of distribution
that it is difficult to understand why the first factor does not
overbalance and mask the second. For example, in 0.02 m.
benzamid the animals are narcotized both at 10° and at 20°,
but more rapidly at 10° and they also die earlier at 10°. Evi-
dently at 10° enough benzamid enters the cell to produce the
full physiological effect. At 20° the coefficient of distribution
is only very slightly lower than at 10°, but the rate of reaction in
the organism is supposedly two to three times as great as at 10°.
The temperature coefficient of distribution of benzamid for 10°C.
is so small that it cannot possibly account for the observed
results which are constant and distinct.
There can be little doubt, I think, that with sufficiently low
concentrations of benzamid the increase in rate of reaction in the
organism is the chief factor in determining that the narcotic effect
is less at the higher than at the lower temperature. At 10° the
rate of reaction is so low that the concentration of benzamid in
the cell is sufficient to produce the full physiological effect: at
20° the concentration in the cell as determined by the coefficient
of distribution is only very slightly lower, but the rate of reaction
in the animal is now two to three times as great as before, i.e.,
it is now so great as compared with the concentration of benzamid
in the cell that only a fraction of the total reaction volume can
be affected by the benzamid, consequently the physiological
effect is less at the higher than at the lower temperature.
And here the factor of acclimatization enters. We have seen
that for KCN and alcohol, the higher the rate of reaction, the
greater the degree of acclimatization, provided the concentration
of the reagent is sufficiently low. The same rule holds good for
benzamid. In this case it is impossible to determine where the
THE JOURNAL OF EXPERIMENTAL ZOGLOGY, VOL. 14, No. 2
186 Cc. M. CHILD
primary effect ends and acclimatization begins, for both show the
same relation to temperature changes and one merges into the
other.
But if we accept the above interpretation of the results obtained
- with Planaria, how are we to interpret Meyer’s results with tad-
poles, viz., that in temperature experiments the narcotic effect
varies with the coefficient of distribution. It is evident that if
only minimal concentrations of the reagents are used, if the vol-
ume of lipoids in an organism or tissue is very great and if the
total volume of chemical reaction is relatively small, then the
coefficient of distribution may become the chief factor in deter-
mining the physiological effect of a narcotic. In the vertebrate
nervous system the volume of lipoids is very great and, except in
the earlier stages of development, the actual volume of chemical
reaction is small. Moreover, Meyer’s results are for minimal
concentrations only. Undoubtedly the concentration of the nar-
cotic in the nervous system of the tadpoles is the chief factor in
determining the narcotic effect. It is at once apparent that this
is a case where the coefficient of distribution may be the determin-
ing factor, but it is also apparent that generalization on the basis
of this case alone can lead only to wrong conclusions. If Meyer
had worked with some of the lower invertebrates as well as with
the tadpoles and if he had used higher as well as minimal concen-
trations, he would have reached very different conclusions.
My experiments permit only the conclusion that the action of
benzamid on Planaria is of essentially the same character as that
of aleohol, KCN, ete. In higher concentrations the resistance
of the animals varies inversely as the rate of reaction; in lower
concentrations it varies directly as the rate of. reaction, except in
certain cases where incidental factors such as the coefficient of
distribution, nutritive condition, etc., play a part. Under the
usual conditions and in organisms where the differentiation of
tissues, and especially the accumulation of lipoids is not very great
the rate of reaction in the organism is by far the most important
factor in determining the resistance and the other factors become
practically negligible. In general then the experiments with
DYNAMICS OF MORPHOGENESIS 187
benzamid afford a complete confirmation of the conclusions based
on the work with alcohol and KCN.
The records of the two following series show the results with a
concentration considerably above the minimum (Series 566) and
with a concentration near the minimum (Series 564).
Series 566, A 1,B1. Worms 8 to 9 mm. in length, all from the
same stock. Lot A 1, ten worms at 21°C.; Lot B 1, ten worms at
11°C. Condition recorded every fifteen minutes, table 7.
The worms used here were rather small and young, with a
relatively high rate of reaction. Lots A 2 and B 2 of the same
series consisted of very large old worms: in those disintegration
proceeded much more slowly, but the worms at the higher tem-
- perature died first, as in table 7. In other series similar results
were obtained.
The table shows clearly that Lot A 1 disintegrates more rapidly
than Lot B 1, although the coefficient of distribution is lower in
the case of Lot A 1. On the other hand, it is evident that the dif-
ference in rate of disintegration between the two lotsis very much
less than in alcohol and KCN series with the same tempera-
ture interval. At first glance, the obvious conclusion seems to be
TABLE 7
Worms in benzamid, 0.04 m. at 12.00 noon, November 15, 1912
STAGES
LENGtH OF TIME | LOT : = ae = y
1.15 { aa i :
1.30 | ms 10 :
fe ff fae aera:
Tv leaaes 0 |
on | A he Pps ree |
188 Cc. M. CHILD
that the two factors, coefficient of distribution and rate of reac-
tion balance each other to some extent and that since the rate of
reaction has the greater temperature coefficient it determines the
result. Undoubtedly this is to a certain extent correct, but it is
difficult to understand how the relatively small temperature
coefficient of distribution can so nearly balance the much greater
temperature coefficient of rate of reaction. Iam strongly inclined
to believe that another factor is involved here. The macerating
effect of benzamid is very great; the tissues seem almost to dissolve
in it. High concentrations of alcohol produce the same effect to
some extent and it appears to a greater extent in ether. It is
probable that disintegration in high concentrations of these and
many other substances which are highly fat-soluble is not solely
the result of the narcotic action, but in part of a change in physi-
eal condition in consequence of the solution of the substance in
the lipoids. The cells and tissues are undoubtedly dissolved to
some extent. This physical effect apparently hastens disintegra-
tion and often decreases the differences due to different rates of
reaction. With KCN this factor is eliminated for all practical
purposes, if it exists at all in that case. The concentrations of
KCN used are so very low that everything except the chemical
factor disappears from the result. For this reason results ob-
tained with substances which must be used in high concentration
or which are very highly fat-soluble should always be checked
by KCN.
The concentration of benzamid used in Series 566, viz., 0.04
m., is more than double the minimal narcotic concentration for
either of the temperatures used. This concentration is near
saturation at 10°C., so that-higher concentrations can be used
in temperature experiments only when higher temperatures
are used. But with higher concentrations the physical factor
undoubtedly becomes still more important, so that we should
expect the differences due to rate of reaction in the organism to
become less and less marked with increasing concentration.
For comparison with Series 566 another series is given in which
the concentration used was lower, 0.02 m.; this is only slightly
above the minimum.
DYNAMICS OF MORPHOGENESIS 189
Series 564, A 1, B.1. Worms 8 to 9 mm. in length, all from the
same stock. Lot A 1, ten worms at 20° to 21°C. Lot B1, ten
worms at 10° to 11°C. Only hourly observations recorded.
Table 8.
Observations were not carried further on this series. Table
8 shows two gaps in the observations, one during the night, when
no observations were made, the other after the death of B 1
where the records are omitted from the table as not essential.
These gaps do not interfere in any way with the definiteness of
the results. *
Here the worms show earlier and more rapid disintegration
at the lower temperature: apparently some other factor than the
rate of reaction is the chief factor here. Hereagain the tempera-
ture coefficient of distribution is certainly not sufficient to account
for the marked difference between the two lots. As a matter of
TABLE 8
Worms in benzamid, 0.02 m. 11.45 a.m., November 11, 1912.
STAGES
LENGTH OF TIME | LOTS =
I II III IV 7 Vv
Al 10
#18 { Bl 6 4
es fy At 10
ia \| Bi 4 6
: f| Al PRA
bela i Bt 10
Seis f| Al 10
ite h) “Ba | 5 bid
Al E70 |
Sule thee 8 rei A
* * * * * * cd * * * * * ! * * * * * *
Al 3 3 2 2
P Al 3 2 2 2 1
22.15 { Bl 10
* * * * * * * oS * * * * * *
190 C. M. CHILD :
fact, these results correspond to those obtained by the indirect
method with alcohol and KCN: there is no doubt that a certain
amount of acclimatization occurs in Lot A 1 of this series. At
22.15 hours the three worms of Lot A 1 which were still intact
showed some slight recovery from the complete narcosis of the
preceding day. In other words, this concentration of benzamid
gives us indirect results in which the acclimatization factor is
involved, consequently here the resistance of the worms varies
directly as the rate of reaction.
These two series are sufficient to show that the Coefficient of
distribution of benzamid is of little importance in determining
the narcotic effect on Planaria and the resistance of the animals
to it. “Here as in the case of alcohol, the rate of reaction is the
most important factor. It must not be forgotten, however,
that in the vertebrates with the great volume of lipoids in the
nervous system the coefficient of distribution of a narcotic may
be a much more important factor in determining its physiological
effect. But there as elsewhere, this factor remains a condition,
not a cause.
4. The evidence from animals of different age
When young and old animals are compared by the direct method
the resistance of the young animals is always much lower than
that of the old. This method has been used with both KCN
and alcohol on Planaria dorotocephala and with KCN on P.
maculata, P. velata, Phagocata gracilis, Mesostomum sp. and
embryonic and larval stages of Amblystoma. In all cases the
same result was obtained, the younger animals died and disinte-
grated earlier than the older. In the cases of Planaria doroto-
cephala and P. velata both the younger and older animals were
undoubtedly the products of asexual reproduction, as sexual
reproduction has not been observed in P. velata and only in a
single individual in P. dorotocephala during the years that I have
had these forms under observation. In Planaria maculata, on
the other hand, the young worms used in experiment were raised
directly from eggs laid in the laboratory and in the other forms
mentioned above asexual reproduction does not occur. As a
DYNAMICS OF MORPHOGENESIS 191
matter of fact, there is no fundamental difference, though there
may be a difference in degree between planarians asexually pro-
duced and those arising from eggs. The following series gives
characteristic results for Planaria maculata.
Series 51 II and VI. The younger worms (Lot 1) were 3 to
4 mm. in length and had emerged from egg capsules in the labora-
tory during the last few days preceding the experiment; they had
been fed with earthworm once, two days before the test was made.
The older worms (Lot 2) had hatched several weeks earlier and
had been fed with earthworm until they had attained a length of
8 to 9 mm.
Each lot consisted of ten worms. Table 9 gives the data
in the same form as the preceding tables. In this series observa-
tions were made every fifteen minutes, but since the results at all
stages are perfectly uniform and definite, only the alternate
readings are given in the table.
In Lot 1 disintegration begins earlier than in Lot 2 and all the
worms are completely disintegrated within two hours after being
TABLE 9
Animals in KCN, 0.001 m. 1.35 p.m., September 1, 1912
STAGES
LENGTH OF TIME LOTS |
I I } i EV: Vv
1 6 4
ae { 2 9, | ol
5 2 3
E80 { 2 7 in 2
| |
ee ee ee ee 10
|
20 { 2 gia 6 1 | 1
|
2.30 2 ey Syl 1 1
3.00 2 | ee 3 3
3.30 2 | 1 3 6
4.00 2 1 9
192 . ' ¢. M. CHILD
placed in the KCN, while the worms of Lot 2 remain intact in
most cases longer than those of Lot 1 and require four and one-
half hours to reach Stage V.
That the rate of reaction is higher in the more er hatched
worms cannot be doubted. They are more active and grow more
rapidly than the older worms and there is not the slightest reason
to doubt that if we could measure their metabolism directly, as
has been done for higher animals, we should find that they, like
the young of higher animals, show a higher rate of reaction per
unit of body weight than the older animals.
The following simple experiment also indicates that the rate
of reaction in the young worms is higher than that in the old. If
a miscellaneous stock of several hundred worms, including both
young and old is placed in water in an Erlenmeyer flask, which is
then tightly corked, the worms begin to die within a few hours
and it is always the young worms which die first. That death
in this case is due to lack of oxygen rather than to the presence of
CO, or other products of metabolism is indicated by the fact
that the water from such a flask in which worms are dying rapidly
will not kill other worms, provided they have access to a small
bubble of air. The effect of KCN 0.001 m. is almost exactly simi-
lar to that of lack of oxygen: in both cases the worms with the
higher rate of reaction (the young worms) die first.
The possible objection that the smaller size of the younger
worms may in some way determine the result is met by the
following facts: the younger worms do not simply disintegrate
faster than the older, they begin to disintegrate earlier; it is diffi-
cult to see how difference in size alone can account for this differ-
ence. Secondly, in certain experiments on nutrition to be de-
scribed elsewhere, the larger worms disintegrate earlier than the
smaller, because of a higher rate of reaction resulting from a differ-
ent nutritive condition. In fact a large number and variety of
experiments to be described will demonstrate beyond a doubt
that size alone is a factor of comparatively little importance. It
is, in fact, one great advantage of the method that it is at least
very largely independent of size.
DYNAMICS OF MORPHOGENESIS 193
All results obtained thus far from experiments by the direct
method on animals of different ages, whatever the species, are
essentially similar to those of table 9. On page 165 the records
of a series with asexually produced young and old individuals of
Planaria dorotocephala are given to illustrate the method of
recording data. There also the young animals begin to disin-
tegrate earlier and disintegrate more rapidly than the older.
When we compare these results with the other lines of evidence
it is clear that all are in essential agreement. The only possible
factor that can be responsible for the observed differences in
physiological resistance is the rate of reaction in the organism.
In the paper on senescence two series are presented showing
_ the differences between young and old animals from nature by the
indirect method (Child 711 a, pp. 544-547, figures 1 and 2; figure
2 is reproduced as figure 1 of the present paper). Other figures
show similar differences produced in a variety of ways: for exam-
ple, figures 3 to 7 of that paper show differences in resistance, L.e.,
in rate of reaction produced by differences in nutrition and figures
9 to 14 show how pieces may become physiologically young as
the result of regulation. .
By the indirect method the animals with the higher rate of
reaction show the higher resistance, viz., they become more readily
and more completely acclimated.
5. Further miscellaneous evidence
Further evidence for the existence of a relation between physio-
logical resistance and rate of reaction is obtained from various
other lines of experimentation which will be considered fully in
other connections.
By means of the direct method it is possible to distinguish the
change in rate due to various forms of stimulation. For example,
a piece isolated by cutting has, during the first few hours after
the section, a much lower resistance, i.e., a much higher rate
of reaction, than the same region of the uninjured body. This
method shows further that the resistance of such pieces gradually
194 Cc. M. CHILD
increases, namely, their rate of reaction decreases during the
first twenty-four hours after section, until, except in relatively
large pieces, it is greater (i.e., the rate of reaction is lower) than
that of the corresponding region in the uninjured animal. In such
a sequence of events we see first the sudden rise in rate due to the
cutting; after this the rate gradually falls as the effect of the cut-
ting gradually decreases, until finally the rate in small pieces is
lower than when they formed part of the uninjured whole. This
low rate of small pieces as compared with the uninjured whole, is,
as I shall show later, a result of isolation, viz., of the absence or
decrease in the action of physiological correlative factors which
before isolation played an important part in maintaining the
average rate of reaction at a certain level. Manifestly, such
changes in physiological resistance occurring in an isolated piece
within twenty-four hours or less, cannot be readily or consistently
interpreted on any other basis than that of the rate of reaction,
especially when they are compared with results obtained in other
ways.
Moreover, it can be shown by the same method that the decrease
in resistance (i.e., the increase in the rate of reaction in pieces
following cutting) can be largely prevented by partial anesthesia
at the time of cutting.
And finally, as regulation proceeds in the piece, the resistance
gradually decreases, namely, the rate of reaction gradually rises,
until in cases where the piece forms a new whole, the resistance
becomes much lower, i.e., the rate becomes much higher than
it was originally when the piece was a part of the uninjured
animal. In short, the new whole is physiologically, as well as
morphologically, younger than the part of the animal which it
originally represented.
Again, the change in resistance in a piece as compared with the
corresponding region of the uninjured animal, varies with the
degree of mutilation, viz., the rate of reaction increases tempora-
rily as the degree of mutilation increases. And the greater the
amount of regulatory reorganization in a piece, the lower the
resistance as measured by the direct method, becomes; in other
words, the higher its rate of reaction.
DYNAMICS OF MORPHOGENESIS 195
Turning to another line of experiment, we find that in pieces of
a given size from a given stock of worms, the different types of
head show different resistances by the direct method. In general
the normal head shows the lowest resistance, that of the terato-
phthalmic head is somewhat higher and the increase continues
through the teratomorphic and anophthalmic to the headless
type (for a description of these different types of head, see Child
’11 b’11 ce). I have already shown (Child ’11 ¢ 712 a) that we
can induce the appearance of the more abnormal types of head in
place of the less abnormal by decreasing the rate of reaction in
the head-forming regions and can bring about changes in the
opposite direction by increasing the rate of reaction in this region.
According to the results of the direct method of determining the
resistance, the normal head shows the highest rate and from this
the rate decreases through the various forms to the headless type.
Here then the two lines of experiment—production of the abnormal
types of head with the aid of external factors and determination
of the physiological resistance of the different types—are in com-
plete agreement and we cannot doubt that the difference in resist-
ance in the different types is determined by the differences in
rates of reaction.
It remains now to mention a widely different line of evidence
which serves to confirm the results of the resistance method.
Dr. Tashiro, an assistant in Dr. A. P. Mathews’ laboratory at
the University of Chicago, has recently devised an exceedingly
delicate apparatus which makes it possible to determine and com-
pare the amounts of CO.-production in small or nearly quiescent
organisms or pieces of tissue. Dr. Tashiro has been kind enough
to make a number of comparative tests of CO.-production in indi-
viduals and pieces of Planaria under different conditions; in every
case, the results obtained are parallel to my own, obtained with
KCN and alcohol. Animals or pieces which by the direct method
show a lower resistance show in this apparatus a higher rate of
CO.-production. Direct comparisons by means of this apparatus
are limited to animals and pieces of approximately the same size,
so that as far as my own results concern animals and pieces of
different size they cannot be directly confirmed in this way.
196 C. M. CHILD
Thus far comparison of the rate of CO,-production has been
made between relatively long anterior and posterior pieces,
between short anterior and posterior pieces, between animals in
different nutritive condition, between early and later stages of
regulation and between moving and quiescent animals and in
every case the result obtained paralleled that obtained by the
direct method. Moreover, in almost every case the result was
obtained without knowledge of the result which I had obtained
with the resistance method. I am under great obligation to Dr.
Tashiro, both for his kindness in making the tests and for permis-
sion to make this statement.
These data obtained by a totally different method from my
own afford a most valuable confirmation of the results of the resist-
ance method. The only possible conclusion is then, I believe,
that the resistance of Planaria—as well as of various other forms
—to certain reagents is in general a measure of the rate of meta-
bolic reaction and can be used as a basis for comparing the rates
of reaction of different animals and pieces under different internal
and external conditions.
6. The value and the limitations of the resistance method
It is evident that so long as disintegration is the criterion of
death, this method can be used only in cases where death is fol-
lowed within a short time by disintegration. So far as my experi-
ence goes, this occurs only in those forms where a highly differ-
entiated connective tissue or a well developed skeleton is absent.
For example, the method gives very definite results in the earlier
stages of amphibian ontogeny, but by the time the animals
hatch, the skeleton and connective tissue have attained a con-
sistency such that disintegration does not occur for days after
death. If we can find some other satisfactory criterion of death,
we can of course apply the method much more widely. Failing
this, it may be possible to use the indirect method to some extent
in such eases, for with that method determination within a day or
two of the time of death is in most cases sufficient. As yet I have
not attempted to develop the method along this line.
DYNAMICS OF MORPHOGENESIS 197
For those forms where disintegration follows soon after death,
the direct method is of much greater value.than the indirect.
In the first place, it is more accurate and permits the determina-
tion of smaller differences in rate of reaction. With the direct
method the rate of reaction during the first few moments after
the worms are placed in the reagent, or at most an hour or two, is,
the chief factor im determining the time of death. With the indi-
rect method the rate of reaction during days or weeks is a factor
in the results, but during this time the rate may change in conse-
quence of gradual starvation, or in pieces in consequence of regula-
tory processes, moreover, external conditions, e.g., temperature)
may also alter the rate and so influence the result. To take a
- ease in point, suppose we compare large relatively old worms with
very small young worms. By the direct method the young worms
show a much higher rate of reaction than the old, but by the
indirect method the factor of starvation may give the results a
wholly misleading character; in other words, the very small worms
with a much higher rate of reaction may, in spite of this rate, die
of starvation before the large old animals, with much lower rate,
die from inability to become acclimated. Nevertheless, it is of
interest to note that except in certain extreme cases of this kind
the smaller, younger animals live longer, even though they have
less material available for nutrition. It is perhaps possible that
in alcohol, with which most of my work by the indirect method was
done, these younger worms with their higher rate of reaction are
able to make some use of the alcohol as a nutritive substance.
This complicating factor is of course absent when KCN is used.
If a concentration of alcohol near the limit at which acclimati-
zation becomes impossible, is used, the temperature factor may
appear very clearly: after the animals have been in the solution
for some time and are less vigorous than normal, a rise in temper-
ature of a few degrees may change the action of the reagent from
what I have called the indirect to the direct and the animal dies.
All of these and doubtless other factors also, interfere with the
accuracy of the indirect method, but in the direct method they
ean be eliminated with little difficulty. It is always desirable
to check by the direct method results obtained by the indirect.
198 C. M. CHILD
There are also in certain cases regional factors which make the
application of the indirect method difficult. When, for example,
the rate of reaction in different parts of the body is very different
it is sometimes difficult to find a single concentration which will
give results by the indirect method for both.
In animals like many of the coelentera and turbellaria the con-
tinued existence of the characteristic structural features after
they have once developed is more or less directly dependent on
the maintenance of a certain relatively high rate of reaction.
Decrease in the rate below a certain limit is followed either by
dedifferentiation or more usually by death. It is difficult to com-
pare rates of reaction by the indirect method in such cases for
the direct depressing effect of the reagent often kills or hastens
the death of certain relatively highly differentiated parts, even
when the concentration is so low that it has little or no effect on
other parts. The hydranth region of the hydroid Corymorpha
is a case in point. In all my tests by the indirect method the
hydranth region dies before the stem, although young hydranths
live longer than old ones in the same solution. These cases need
further investigation, but apparently we have here simply a
structure in which the range of acclimatization is narrowly lim-
ited by the high degree of differentiation which has resulted from
the relatively high rate of reaction. It will probably be found
true in general that structural differentiation, especially where
it occurs in consequence of a high rate of reaction, limits the
range of adaptation to depressing media. Moreover, the older
such a structure becomes the more narrowly is its range lim-
ited. In Tubularia, for example, the mere change from open
water to the laboratory usually brings about the death of the
hydranths, apparently in consequence of the decrease in rate of
reaction accompanying the change in conditions: use of a depress-
ing agent hastens the death of the hydranth still further because
it brings about a still further decrease in the rate of reaction.
And finally there are the factors of time and labor to be consid-
ered in connection with the indirect method. New solutions must
be made up every day or two and the renewal of the solutions in
the flasks requires much time. The direct method requires only
DYNAMICS OF MORPHOGENESIS 199
one supply of solution and observations extend over only a few
hours.
But although the indirect method is much less valuable than
the direct for the comparison of rates of reaction alone it is of
great value in the analysis of morphogenesis for it enables us to
determine with some degree of certainty the relative rates of
reaction connected with different morphogenetic processes and to
inhibit the processes with lower rate (Child *12). Moreover,
I believe it will prove of value as a method for the experimental
study of acclimatization. Certain points of considerable interest
have already appeared in the course of my work with this method,
although I was chiefly concerned with other problems. For
example, the fact that old animals always die at a larger size than
young ones in certain low concentrations of alcohol must have
a very definite physiological significance. Under these conditions
the old animal is not able to use as large a proportion of its own
substance for nutrition as is the young animal. Possible inter-
pretations of this fact will be considered elsewhere.
As regards the relative value of different reagents, there can
be, I believe, but one conclusion, viz., that the cyanides are far
more valuable than any of the others. The very low concentra-
tions which are used, as well as the constitution of the cyanides
practically eliminate various factors which may complicate the
results obtained with the narcotics in the stricter sense and leave
only the chemical factor. The results obtained with the cyanides
must, I think, be taken as the basis of the method and other
results must be checked by them.
With pure reagents, water of constant constitution, constant
temperature, care in making solutions and the proper care of
stocks and selection of animals the direct resistance method is a
method of great delicacy and the complicating factors which
influence the results by the indirect method can be practically
eliminated in the direct method. The method is undoubtedly
capable of much further development as an exact method than I
have yet attempted. By standardizing the conditions of the
experiment and adopting a certain unit as a basis of measurement
we may obtain definite dynamic expressions for different ages,
200 C. M. CHILD,
different conditions of nutrition, different degrees of regulation,
ete.
It is of course possible that further use of the method may
bring to light further complicating or limiting factors but it is
certain that the method is capable of very wide application. But
in addition to the immediate results the method gives us a means
of attack on various problems which have not heretofore been
open to investigation.
During the last few years I have used both the direct and the
indirect methods in a very large number of experiments and with
a variety of forms. The results obtained afford a new insight
into the dynamics of living organisms, they throw light on the
problem of physiological polarity and symmetry, they afford a
dynamic basis for the law of antero-posterior development and
they have demonstrated that changes in the rate of metabolic
reaction may bring about changes in the number, localization,
degree of differentiation, etc., of definite morphological characters.
Besides this they have made it possible to demonstrate the essen-
tial similarity between the process of regulation by which a piece
gives rise to a whole and the other forms of reproduction and
development in nature. They have given us a new viewpoint
from which to consider the questions of senescence and rejuvenes-
cence. And finally, certain of the results obtained have a very
direct bearing upon the problem of inheritance. Some of the
more important results have already been briefly presented
(Child ’12 b) and these and others will be more fully considered
in following papers..
IV. THE ACTION OF DEPRESSING AGENTS IN GENERAL
1. The nature of the action of depressing agents on Planaria
It is not my intention to discuss at length the problem of nar-
cosis, but merely to call attention to certain points.
The poisonous effect of the cyanides is very generally regarded
as due, at least in large part, to a retardation or inhibition of
some sort of the oxidation processes. According to Geppert
(89) they render the tissues incapable of uniting with oxygen.
DYNAMICS OF MORPHOGENESIS 201
Loeb and various others have used the eyanides extensively dur-
ing later years as.a means of retarding or inhibiting the oxidation
processes.
It is also certain that the planarian in KCN shows a decrease
in CO.-production as compared with a normal animal.
At any rate, it is evident that the CN radical affects some of
the most fundamental metabolic reactions. Moreover, the fact
that the effect of the cyanides varies according to the rate of
reaction in the organism suggests that the action is primarily
chemical. Apparently the effect of a KCN-solution depends
upon the number of chemical bonds in the organism which are
opened up in a given length of time; in other words, the higher
the rate of reaction, the greater the opportunity for the KCN to
produce its effect.
Apparently the KCN acts by entering the metabolic complex
at some point or points and altering certain essential features of
it so that it cannot continue. That the point or points of en-
trance lie somewhere along the course of the oxidation processes,
is the most generally accepted view.
But how are we to conceive the process of acclimatization to the
eyanides? It occurs only in very low concentrations, but that
it does occur there can be no doubt. Why does the relation. be-
tween the capacity for acclimatization and the rate of reaction
exist, 1.e., why does the individual or piece with the higher rate
of reaction become in general more readily and more completely
acclimated? It is evident that a metabolic factor is involved in
the process of acclimatization, but we are at present far from any
real knowledge as to the nature of the process. The relation
between the rate of reaction and the degree and rapidity of
acclimatization must have a very definite meaning and it may
perhaps serve as a basis for further work along this line.
We have seen that with alcohol, ether, chloretone and benzamid
essentially the same relation between physiological resistance and
rate of reaction in the animals exists. Since my investigations
have had thus far another object, I have not as yet attempted to
determine the relations for any very large number of substances,
but the occurrence of the same relation between resistance and
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 2
202 Cc. M. CHILD
rate of reaction with all the substances used indicates clearly a
certain similarity in the nature of their action.
Moreover, the effects of alcohol, ether and chloretone upon
morphogenesis are in their essential features similar to those pro-
duced by KCN and on the other hand to those produced by low
temperature, products of metabolism in the water, ete. (Child
11 ¢’11d). Evidently all these substances act in some way to
decrease the rate of the metabolic processes. Similar conclu-
sions have been reached by various authors and on the bases of
various lines of investigation. The similarity in the effects of
KCN and alcohol, ether, etc., suggests that all act in some way
on the oxidation processes.
But that the action of the alcohol is not the same as that of the
cyanides is clearly indicated by the difference in the capacity
for acclimatization to the two substances. Alcohol 4 per cent
kills the worms within a few hours, but in alcohol 1 per cent most
worms become acclimated. KCN 0.001 m. kills the worms in
about the same time as alcohol 4 per cent, but no appreciable
degree of acclimatization occurs in concentrations higher than
0.00004 m. ‘To what is this difference due?
The planarian is exceptionally good material in many ways
for determining the physiological effect of chemical substances,
andparticularly for purposes of comparison of different substances.
In many cases also two different aspects, the physiological and
the morphological, of the effect may be compared with each other:
this possibility in turn gives us a method of attack on certain
morphological problems which have scarcely been accessible
heretofore.
The results of the experiments with depressing agents on Pla-
naria have an important bearing on the general theory of narcosis
for they indicate very clearly that the coefficient of distribution
of the narcotic between water and fat has no necessary relation to
its narcotic action. In Planaria, where there is no great accumu-
lation of lipoids in any organ, we have seen that the coefficient
‘of distribution is a factor of very little importance as compared
with the rate of reaction in the organism. But attention has
already been called to the point that in the vertebrates, where the
DYNAMICS OF MORPHOGENESIS 203
accumulation of lipoids in the nervous system is very great, the
coefficient of distribution may be a very important factor in
determining the concentration of the narcotic in the nervous
system and so in determining its physiological effect. But it is
evident that in such cases the fundamental relation between the
physiological effect of the narcotic and the rate of reaction in the
organism or organ is simply masked by the incidental factor of
coefficient of distribution.
2. Certain differences in the action of different reagents
Notwithstanding the general similarity in the physiological
and morphological effect of the different substances used thus far,
certain more or less characteristic differences exist.
In the first place, there is a marked difference in the relation
between mortality and morphological effect. In alcohol, ether,
chloretone a considerable proportion of the pieces die in any
concentration high enough to produce a marked morphological
effect. In KCN, on the other hand, it is possible to obtain ex-
treme morphological effects without losing a single piece in large
series of several hundred pieces. Such a difference as this must
have some very definite meaning.
But certain other minor differences appear to be more or less
characteristic of the different substances. For example, KCN
inhibits or retards the formation of the optic pigment to a greater
extent than any other reagent used. In many cases eye-spots
consisting only of the unpigmented areas are formed (Child ’12,
pp. 124-125). The outgrowth of the auricles is also greatly re-
tarded or completely inhibited in KCN, although the characteristic
unpigmented sensory area develops in the same manner as when
the auricle grows into its normal form. In general the consist-
ency and color of the new tissue formed in KCN differs to some
extent from that of tissue which develops in other reagents.
Ether apparently inhibits the development of new tissue at
the cut surface to a very great extent: in some cases a new head,
or at least a ganglionic mass may develop in ether with scarcely
a trace of outgrowth from the cut surface (Child ’12, pp. 120-
204 Cc. M. CHILD
121). Thus far, this éondition has been observed only in ether.
In alcohol, on the other hand, some development of new tissue
occurs in any concentration which does not kill.
The morphological effect of chloretone is very similar to that
of aleohol but appears with much lower concentrations (alcohol
1-1.5 per cent; chloretone 0.02—0.025 per cent).
As already mentioned above (p. 188) a much greater degree of
maceration and apparent solution of the tissues occurs in alcohol,
ether, chloretone and benzamid than in KCN; this may be due
to the different effects of these substances on the lipoids.
The effects of low temperature are in general similar to those
produced by the substances mentioned, but in low temperature
the growth of new tissue is merely retarded, not inhibited. The
relation between the formation of new tissue at the cut surface
and the redifferentiation back of the cut is much the same as under
the usual conditions. In this respect low temperature differs in
its action from all the substances used.
Thus far I have not been able to discover any features strictly
_ characteristic of the animals which develop in water containing
an excess of their own metabolic products. Conditions of this
sort which produce marked morphological effects usually produce
also a high mortality as is the case with alcohol and the morpho-
logical effects resemble those of alcohol.
More attention directed to this point will doubtless bring to
light other morphological or physiological effects characteristic
of the different substances. At present, however, there seems to
be no ground for believing that any of these features which are
more or less characteristic of one or the other substance are specific
chemical effects. Some of them may conceivably be such, but
it is much more probable that they are the result of incidental
physical factors, e.g., osmotic conditions, coefficient of distribu-
tion, ete. In general the morphological effects of these substances
and conditions are due primarily to differences in the rate of
reaction in different developing organs and regions.
In more highly differentiated animals than Planaria, where the
difference in constitution in different organs is much greater than
here it should be possible to produce morphological effects with
DYNAMICS OF MORPHOGENESIS 205
features more distinctly characteristic of the different reagents.
The slight differences of this sort in Planaria are, I believe, good
indications of the low physiological specification of the tissues.
V. SUMMARY
1. In Planaria and in many other forms in which there is no
highly specialized skeletal or connective tissue, death is followed
within a short time, varying from a few minutes to a few hours,
by disintegration. This fact makes it possible to determine with |
some degree of accuracy the time of death of animals, regions of
the body or pieces.
2. Experiments with stimulated and unstimulated animals,
with animals at different temperatures, with old and young
animals, with animals in different nutritive condition and with
pieces of different size, from different regions of the body and
under different conditions demonstrate the existence of a relation
between the length of life (physiological resistance) of the animals
or pieces in KCN, alcohol and various other anesthetics and the
rate of the metabolic reactions or certain of them, probably the
oxidations.
3. In relatively high concentrations, in which the maximum
length of life is only a few hours, the length of life (resistance)
varies inversely as the rate of reaction in the animals orpieces.
The higher the rate of reaction, the earlier .does disintegration
begin and the more rapidly it proceeds and vice versa. This is
the direct method of comparing rates of reaction.
4. In relatively low concentrations, in which the animals
remain alive for days or weeks and in which a greater or less
degree of acclimatization occurs, the length of life (resistance)
varies directly as the rate of reaction, except in certain cases
where incidental factors contribute to the result. The higher the
rate of reaction in the animal the more complete the acclimatiza-
zation, at least in most cases, and the greater the length of life.
This is the indirect method.
5. With concentrations between these two extremes the results
differ according to the concentration of the reagent used and the
206 C. M. CHILD
rate of reaction in the organism. For any two rates of reaction it
is possible to find a concentration of the reagent in which the
resistances will be approximately the same. In order to avoid
misleading results it is necessary to be certain that with the direct
method the concentration is sufficiently high and with the indirect
method, sufficiently low.
6. These two methods, and especially the direct method, afford
a means of attacking various problems and they serve particu-
larly to give us some insight into the dynamics of morphogenesis.
November, 1912.
BIBLIOGRAPHY
Cuitp, C. M. 1910 Analysis of form regulation with the aid of anesthetics
Biol. Bull., vol. 18, no. 4.
1911 a A study of senescence and rejuvenescence, based on experi-
ments with planarians. Arch. f. Entwickelungsmech., Bd. 31, H. 4.
1911 b Studies on the dynamics of morphogenesis and inheritance in
experimental reproduction. I. The axial gradient in Planaria doro-
tocephala as a limiting factor in regulation. Jour. Exp. Zodl., vol.
10, no. 3.
1911 c Experimental control of morphogenesis in the regulation of
Planaria. Biol. Bull., vol. 20, no. 6.
‘1911 d Studies on the dynamics of morphogenesis, ete. IJ. Physio-
logical dominance of anterior over posterior regions in the regulation
of Planaria dorotocephala. Jour. Exp. Zodl., vol. 11, no. 3.
1912 d Studies on the dynamics of morphogenesis, etc. IV. Cer-
tain dynamic factors in the regulatory morphogenesis of Planaria
dorotocephala in relation to the axial gradient. Jour. Exp. Zoél., vol.
13, no. 1.
1913 Certain dynamic factors in experimental reproduction and their
significance for the problems of reproduction and development. Arch.
f. Entwickelungsmech. Bd. 35, H. 4.
GepPERT, J. 1889 Uber das Wesen der Blausiurevergiftung. Zeitschr. f. klin.
Med., Bd. 15.
Meyer, H. 1901 Zur Theorie der Aikohobieniose: 3. Mitteilung: Der Einfluss
wechselnder Temperatur auf Wirkungsstairke und Theilungscoefficient
der Narkotica. Arch. f. exp. Pathol. u. Pharm., Bd. 46.
THE REACTIONS OF FISHES TO GRADIENTS OF
_ DISSOLVED ATMOSPHERIC GASES
VICTOR E. SHELFORD AND W. C. ALLEE
From the Hull Zoélogical Laboratory, The University of Chicago
SEVEN FIGURES
ils, TERRE CG REPEC VOT Re ree St Gee NR Ded, ee = Pe eee eee eee 208
SPANOS PD heniGsrASea: sere ea ss ee ipo tes ao Rese aties a eee see ee 208
IPO CCUTREN COR Rete eee aetiecic bo ato SOS wind «tease ae 208
2. Experimental control of gases in solution......................05. 209
TENT TANG ISPS TIPE ace SR Cee eae RI eee on a ae eS 219
fem UOC Keen een a Sera A ne Sicko hes ore bc eCoMSTe te eas eiele Hiss 3 BVM AOE NS SoG 219
2 pena vier sad physiology Of HENES.: 2 2: fb .. pd se alein ee a ke 220
IV. The physiological effect of gases upon fishes...................000000008 221
1. The effect of a great excess of nitrogen or oxygen................. 221
Pediheetiech oi a GeHnClency Of OXY LCD: «<i 5.<.2 snes F< ect nse eee ee 223
Sheen eet OmCalvOn GiOMGer. 4020. . ise. Benes 4 ose ciel eee 224
Veslvenctionion tisaato) easesam SOlutlOnee...- co. 4442-2 ea alae 225
ieViet hod: on Experimenta thon j).:543'¢ saicicie eocils se 2 F< seem ony s cee eee 225
PD BREACHION LOIS IN GlesTaASeS satame re A ate seta ant hel <, 5 cua Sse Paso RPI 229
ae OAT OES ORI ES otro eh ee ons ccc rsyl ie aa, ois cas Ee aie eR ee 230
[is WOsait5 15d eRe aA aL Nee Rae eRe tren Sree 236
Ga Nitrogen’... 022.02) 2% BRC ees, Sk Te ee oe 238
Saveserionwto combinations Of factors: os: ..)ks 0. 8 he ee ne ean ee 241
SPA OME GR WATCHES sate ys re tes So Cai) 2] 2 be Nees x Seek ered 241
be Boiled water, with oxygen addedi.........2....-5:4954202e-- 249
e~ Boiled water andresebon dioxide. <= vcs once wee cic «2 A oe 250
dFeAcidvandaimmontanes ease oe eee eee Oe ree ene 252
4. Comparison of reactions..... hota A An Apia ee tiene eo ACRE SE *, . 254
a. Degree of reaction to, different factors. .u sos 0c 0.550 6s eae 254
b- Adult and-jurventle fishess Sarees ace ee ate Re iam ic 5 sees 258
Wile General diSCUSSLON=<o--.2nc- eee ee RIS cee ee oe 259
Ree SMInMAr a2! 7 5: . eh Scene eee a eeeee eae gate ee Pe eRe eyed GS ee oa 262
LEY ol bKoyea) of dh eee seein eieetto.c a bic ts parca cic.e ok biti o Coc Ee CRO DERE EEE 263
207
208 VICTOR E. SHELFORD AND W. C. ALLEE
I. INTRODUCTION
The importance of atmospheric gases to aquatic animals has’
been coming more and more to the attention of biologists through
the work of Marsh and Gorham (’05), Marsh (10), and Reuss
(10) on the effect of gases upon fishes; through the studies of
reversal of behavior reactions by varying amounts in solution
(Loeb ’04, Mast ’10, Wodsedalek ’11, and Allee 12); and through
the survey of the distribution of gases, which has been carried
on in the Wisconsin lakes by Birge (’04, ’07a, ’07b, 710) and by
Birge and Juday (’11). The effect of the different gases, or of
varying amounts of a particular gas upon animals has been but
little studied. There are many facts concerning the distribution
of aquatic animals which do not seem to be explainable on the
basis of amount of oxygen considered as a life and death matter
(Shelford ’11b, and 711d); the junior author experienced diffi-
culty in the control of gases in solution in connection with his
study of isopod behavior. We accordingly decided to design gas
control apparatus and to study the reactions of some group of
animals to gas gradients, making it a joint investigation with
important bearings upon our separate interests. Fishes were
selected because their physiology, habits, and distribution are
well known.
II. ATMOSPHERIC GASES
1. OCCURRENCE
The chief facts concerning the occurrence of gases in nature
and their solubility under experimental conditions are shown in
table 1. The standard method of expressing quantity of gas
in solution is in cubic centimeters per liter at 0°C. and 760 mm.
of mercury. All values are therefore given in these terms.
The amount of each gas that will go into solution from the
atmosphere or under experimental conditions, is determined by
its solubility and partial pressure and by temperature. The rela-
tive solubility of the atmospheric gases is indicated in table 1,
by relative amounts going into solution at 20°C. and 760 mm. of
mercury. If we desire to increase the total amount of gas in
REACTIONS OF FISHES TO ATMOSPHERIC GASES 209
solution we must either increase the total pressure, decrease the
temperature, or substitute a more soluble for a less soluble gas.
TABLE 1
Showing the distribution and solubility of atmospheric gases
} GAS VALUES IN CC. PER LITER AT
| 0°C. AND 760 MM. MERCURY
COMPOSITION At temperature 20°C. Maximum 4 E
aa OF AIR IN 760 mm. mercury amounts Tat ie absent ee
PER CENT OF found in S22 DS ER
SASS | taza fish PRECEDING COLUMN
Water Water Tea
absorbs from | absorbs pure cha a
at aS springs ex-
cepted
Nitrogen,ete...... 79.02 12.32 15.00 19 Lakes; Birge and Juday
(11, p. 52)
Oxypen sn. ss 25< 20.95 6.28 28.38 24 Streams, lakes in winter,
and with green algae
Carbon dioxide... 0.03 0.27 901.00 30 Ponds
Ammonia......... small traces very large 14 Sewage contaminated;
locally quantities Nichols (’94, p. 62)
Methane........-- small traces 34.00 10 Bottom of lake in Sept.;
locally Birge and Juday ’1l, p.
101
2. EXPERIMENTAL CONTROL OF GASES IN SOLUTION
The two methods commonly employed in the control of gases
in solution in water are (a) The reduction of gas content by
boiling or by vacuum pumps, and (b) The increase of some one
gas by bubbling it through water.
a. Gases used
The analyses of the commercial gases were made with the
Hempel (’02, chap. 3) apparatus. The carbon dioxide was
absorbed with 334 per cent potassium hydroxide; the oxygen
with yellow phosphorus except when it constituted more than
50 per cent of the total, in which case alkaline pyrogallol was
used. Oxygen and carbon dioxide in solution were determined
by the methods given by the Committee on Standard Methods
of Water Analysis of the American Health Association (’05, pp.
72-77), and by boiling. Boilings were made in a boiler like that
described by Birge and Juday (’11, p. 7), which holds two liters.
The gas obtained was analyzed by absorption and was corrected
210 VICTOR E. SHELFORD AND W. C. ALLEE
for temperature, pressure and tension of aqueous vapor (Hempel
02, p. 64). The results of the analyses are shown in table 1 A.
TABLE 1A
Analyses of commercial gases used in experiments
ANALYSES, IN PER CENTS OF TOTAL
GAS | Cc SOURCE DEALER
2 rT arbon
| Oxygen Nitrogen disids
Oxygen.......| 99.0 0.95 trace | Liquid air Chicago Calcium Light
| Co.
Carbon diox- |
116 (es eM | 0.6 | 99.4 | Coke ovens Liquid Carbon Dioxide
Co.
Nitrogen..... 7.5 92.57 51 | Liquid air Linde Air Products Co.
b. Hydrogen, ammonia and methane
No experiments were performed with hydrogen. Shull (712)
has found that hydrogen manufactured electrolytically and on
sale by the Lind Air Products Company commonly contains from
2 to 4 per cent of oxygen. Hydrogen which is generated in the
usual way and used for the purpose of displacing oxygen, prob-
ably always contains a quantity of oxygen. The ammonia used
in these experiments was the chemically pure solution diluted
and allowed to flow into the apparatus at a uniform rate. The
effect of methane upon animals has probably not been studied.
Crocker and Knight (’12) manufacture it and have found that
it slightly affects plant growth.
c. Bubbling gas through water
Gas was bubbled through water in order to learn something of
the cause of gas bubble disease (Marsh and Gorham ’05) which
developed in the stock of fish while the experiments were being
conducted. Under these conditions, not one but several factors
were varied. The gas which is bubbled through is added to the
water and each bubble being a partial vacuum for the other
gases, takes them up until they are nearly exhausted from the
water. The results of some of the bubbling experiments are
REACTIONS OF FISHES TO ATMOSPHERIC GASES mrt
TABLE 2
Showing the effect of bubbling gas through a square vessel 30 cm. in depth, holding
four and one-half liters of water, and with 150 sq. cm. exposed to the air. The
gas was introduced at the bottom. Where not given, average flow was about 50 cc.
per min.
GENERAL CARBON
: “ | OXYGEN
___ CONDITIONS | DIOXIDE | x “ 0, | TOTAL
WATER USED| GASUSED Ey Ex sOllieor Es Sunline gil a SAT on
’ ‘ ae SO Ss | cas eeu ss Rg | GEN aa || age | BOER
a ep feats et ia | ee ee P=
Aqua dist...../exposed to | |
atmosphere| 24 16.5 3.0} 1.1 6.8 rea | isicae
Tap water... .| Ne | 20 192001 020) eLatr 3:53 |) e525 | Re 6.40 | 21
Aqua dist.....) | Ne | 24 30| 18.5] 0.5, 1.27] 4.6 } 19.6 | 6.47 =) TRS
Aqua dist.....| Oz | 24 168 | 18.5 1.5) 0.3 | 25.8 | 12.6 6.47 25
Tap water... .| CO2 | 27 180 | 18.5 662.7) | aaa | 12.6 6.47 | 600
0.4 6.3 12.5 6.40 | 500
|
Aqua dist.....| CO2
54 19.0 | 495.9)493.1
shown in table 2. The flows of gas as given are averages as the
tanks were not supplied with valves giving constant flows and
it was necessary to adjust the flows every few hours.
The second test given in the table shows that exposure to the
atmosphere, necessitated by filling the boiler with sufficient water
to allow it to run through only once, made accurate determi-
nation impracticable, with the apparatus at hand. We note that
when oxygen is bubbled through water, the oxygen supply is
not only increased but carbon dioxide and nitrogen are decreased.
When distilled water is used the carbon dioxide is lowered, but
when tap water is used, if the tap water contains bicarbonates,
the bicarbonates are largely changed to carbonates (McCoy and
Smith 711, McCoy and Test 711), the water becomes alkaline
and a slight cloud sometimes appears. ‘This indicates that some
carbonates are precipitated and free and half-bound carbon
dioxide are both removed just as in boiling. This changes the
number of alkaline metal ions in solution. The general results
of bubbling nitrogen through tap and distilled water are respec-
tively similar except that nitrogen is increased instead of oxygen.
When carbon dioxide is bubbled through, it may reach 662 ce.
or more if the flow is rapid. All carbonates are kept in solution;
oxygen is greatly reduced. Loeb (04, p. 7), Mast (11, p. 179)
and Wodsedalek (’11, p. 270) bubbled carbon dioxide through
212 VICTOR E. SHELFORD AND W. C. ALLEE
water and ascribed the results to the carbon dioxide though all
the gases were affected. Loeb in his experiments on Amphipods,
says that the water had an acidity equal to M/500, which is
about 44.8 ce. per liter. None of the others give carbon dioxide
determinations. ‘The presence of fishes modifies the results of
_ bubbling.
TABLE 3
Showing the results of bubbling gas through jars of water containing fishes, the jars
being 50 cm. high, with an exposed surface area of about 43 sq. cm. and holding
about 2.2 liters. The gas was introduced at the bottom. The nitrogen was not
determined because of the small quantity of water. Average flow 50 cc. per min.
|
GAS FISH | }
EE snd used) mee | Due, No, | Length ae Cex biome oN nomi
| min. | hours
Eni@e 43 | 43 5 5-15 |20 | 11.9
ee | 28 2 20 |16.5| 4.6 27.5
EEA, Neue 54 2 7-10 | 20 ey By ale = 22.7
Iv| N: 54 | 2A BON Gy 4905) 44 | aD 19.5
Ul NG 23 4 5-10 | 19 0.5 |
78) Es) leas al | eee Ge iD i a
Vit COs. |] 2") “OF erg ioe “Wak gy mera
VIII standing | 4 5-10 | 20 | 1.1]
| water
d. The reduction of gas content and the addition of particular gases
A laboratory water supply is usually supersaturated with gases
when it comes from a body of water exposed to the atmosphere.
It is often unsuitable for animals because of its low oxygen and
high nitrogen content when it comes from wells (Marsh 710). In
the first case the animals suffer ill effects from the excessive
amount of gas, while in the second they suffer from the deficiency
of oxygen. Either of these difficulties may be remedied by pass-
ing the water through a series of perforated vessels or allowing
it to flow in thin sheets over large rough surfaces (Marsh ’10).
To reduce the gas content to a point below saturation, requires
either the use of a vacuum pump or boiling. In either case all
the gases are affected and it is often necessary to add certain
gases in order to obtain the desired conditions. The exhausting
Per Ay
REACTIONS OF FISHES TO ATMOSPHERIC GASES Pel Ws
of gases by pumps or boiling, offers no difficulties if only small
quantities of water are desired but if a continuous flow is needed
the difficulties are greatly increased. The exhausting of gases
with pumps is to be preferred, as it interferes less with the plank-
ton, the bicarbonates; it would, however, no doubt remove some
of the half-bound carbon dioxide and thus affect the bicarbonates
present in solution to some extent.
A piece of apparatus to give a constant flow of water from
which the gas has been exhausted by a vacuum has been devised
but a good air exhaust and compression pump and considerable
experimentation would be necessary to perfect the first machine
and as yet this has been found impracticable. Apparatus which
boils water and gives a constant flow is less expensive, particu-
larly because less skilled labor is required to. build it. A com-
bined atmospheric deaérating and boiling device installed for the
control of gases is shown in figure 1. The principle is as follows:
The hot water of the laboratory system is passed through a
float valve of high grade, which regulates the flow into a storage
or delivery tank. Above the storage tank and between it and
the float valve, are two open boilers, which bring the water to
the boiling point, and a series of sieves which lowers the gases
to saturation when the boilers are not used. The water is with-
drawn from the storage tank through two tap water coolers.
When the boilers are used, the water comes off with a very low
gas content.!
With the aid of figure 1 we may follow the course of the hot
water through the apparatus when a flow of water at cold tap
temperature is desired. First the water passes through valve A
to the float valve Fv which is opened by the lowering of the water
in the receiving tank. The water then passes through the ‘return
bend’ to valve F which is set so as to give about the flow desired.
After passing this valvé the water rises to the top of the appa-
1The cost of such a piece of apparatus varies greatly with the conditions and
methods of installation. The total cost of the present piece, including labor,
sink, drain and small water table, was about $300. The plans were drawn and all
the parts ordered by the authors, who also supervised the installation. $500 to
$600 would be a fair contract price in Chicago if detailed plans were furnished.
214 VICTOR E. SHELFORD AND W. C. ALLEE
ratus and is distributed on the topmost sieve. From there it
flows through the series of sieves to the collecting pan P. From
here it is conducted to boiler H by means of several small alumi-
num tubes which reach to the bottom of the boiler and thus
introduce the water as near the heat as possible. The water is
conducted from boiler H through a large pipe attached to its
Fig. 1 Apparatus for the control of gas content of water. For description of
the course of the water, see text, p. 218. The water supply pipes are 2? inch
reduced to 2 inch above the branch going to the cooler from the cold water pipe
and to the three way valve from the hot. The valves A, B, and M, are 3 inch
straight way valves. Above valves A and B are swing check valves C, D to pre-
vent the water’s backing into the coolers when a mixture is used. The float valve
is one manufactured by L. Wolf Manufacturing Company, with ground brass
union*soldered to the delivery tube. The shape of the valve makes necessary a
Pi eGamae a, alles Bis
: oe
oa eae |
Fig. 1
| = A eee
a ae. To
| PR A
| hee
|| Zable
i Nd 2)
fata 2 DR >:
~
REACTIONS OF FISHES TO ATMOSPHERIC GASES po I
return bend of ? inch galvanized iron, with a third swing check valve (#) inserted
at the level of the float valve, to hold the weight of the water in the pipe, as the
float valves are not constructed so as to withstand back pressure. The valve F
is a tee handled valve with a brass bar 7 inches long bolted to the tee, a piece of
metal plate clamped to the pipes, lies directly behind the brass bar and the
position of the bar for 1200 cc. and 3600 cc. per minute under average pressure is
marked on the plate. Four } inch iron pipes are used to distribute the water
over the top sieve. The sieves are 18-inch garbage-can lids, with about 1,200
1 mm. holes punched in, 1 cm. apart. The collecting pan is of galvanized iron,
being also a garbage-can lid, a little larger than those mentioned above, with
4 inch holes. Aluminum tubing } inch inside diameter is threaded into iron
nuts, which are soldered to the lid directly beneath the holes.
The boilers, H and J are aluminum saute pans without handles, purchased from
the Aluminum Cooking Utensils Company. They are drilled and the drain pipes
are lock nutted into position. The pipe leading from the upper to the lower’boiler
is of galvanized iron. The small perforated pipe is of brass. The delivery tank
is of galvanized iron but for durability an aluminum vessel should have been used.
The gas burner a is a water heater burner; b is made up of six small unit burners,
so that the amount of heat may be fully controlled; ¢ is a gas range burner which
keeps the water in the delivery tank at the boiling point. The delivery tank is
supplied with an emergency overflow (Ho). The withdrawal tubes K and K of the
delivery tank are ? inch galvanized iron pipe, which fits the smallest size three
way valve (3 Wv); (for withdrawals 2 and 3, see text). The first cooler contains
15 feet of block tin pipe 2 inch inside; the second cooler contains 60 feet of block
tin pipe, ~; inch inside. The tin pipes are soldered to brass fittings. The valve
(R), at the final outlet, is a } inch straight way valve. The coolers are supplied
with tap water through } inch iron pipe attached with unions. Valves O, N,and
L, are } inch brass gas service cocks with tee handles. The coolers are made by
setting one galvanized iron tank into another, the pipe being fastened to the wall
of the inner tank. The overflows of the tank lead to a lead covered trough which
empties into the sink (S). The inner space in the lower tank, is intended to be
used as an ice pack in summer when a temperature lower than the tap water is
desired.
The gas introducer A, is a brass chamber (cm.) with half unions on the ends, and
a plate glass front pressed against rubber packing by means of screws and a brass
frame. On its lower side are two small chambers (above Ch), which are connected
with the main chamber by very small holes drilled in the apices of small cones
hammered into the metal. Each small chamber communicates with a cock which
may be attached to a tank of gas. The pipes in the two coolers are connected with
the gas introducer by one-half unions. Many brass ground joint unions, where-
ever practicable, make possible the detachment of any part. The entire appara-
tus occupies a space above the table 97 by 130 em., the highest part reaching to the
ceiling. It is supported on a frame of iron pipe made by screwing together cut
pipe and fittings described on pages 188 A-189 A, of Crane Company’s catalog No.
40. The sieves are supported on wire pins placed in holes drilled through both
walls of the four pipes which make the tower. Because of the small space avail-
able and because the completed apparatus is the result of several experiments,
there are unnecessary complications.
216 VICTOR E. SHELFORD AND W. C. ALLEE
side to the'bottom of boiler J. In this boiler, the water usually
reaches the boiling point before flowing into the receiving tank.
It flows out through double downward curved tubes K, to the
cooling coils under the pressure of a column of water of the height
of the delivery tank. The curves in the pipe in the coolers cut
down the pressure particularly in the lower cooler where the
pipe is zs inch, so that the flow from the outlet valve cannot
exceed 1200 ce. per minute. The treated water may be at the
exact temperature of the tap water and may contain as little as
0.3 ec. of oxygen per liter. If the water in the delivery tank falls
below the lower ends of the withdrawal tubes K and the coils fill
with air, the water does not flow. To clear the pipes of bubbles,
the hot or tap water may be forced through them by turning the
three-way valve beneath the delivery tube and opening valve O
or the corresponding tap water valve.
At all times of the year, but especially in the winter and spring,
the tap water from Lake Michigan is supersaturated with gases,
and in an open vessel the gases come off in bubbles which soon
cover the bodies of animals in experiments, and thus render their
activities abnormal. Fish kept in aquaria, which are supplied
with this water, die within a few days or even hours, from gas
disease. To remedy this and to supply water which is free from
excess gas, cold water is allowed to run through the apparatus
and is withdrawn from either the second or the third withdrawal.
The second withdrawal is supplied with a union by means of
which iron pipe may be attached to the aquarium in which fish
are kept, while the third is supplied with a hose-end near the
surface of the water table. With this apparatus we have been
able to secure water of any desired temperature and any desired
gas content, nitrogen excepted, within the limits and needs of
our problem.
(1) Effect of the apparatus upon the water. Here we are con-
cerned with three things: (1) The normal tap water, (2) the hot
water, and (3) the water that has passed through the apparatus.
(1) The cold water supply is from the Chicago water system
which uses water taken from Lake Michigan at a distance of
23 miles from shore and from a depth of 6 feet where the water
REACTIONS OF FISHES TO ATMOSPHERIC GASES yA Wy
is 90 feet deep. (2) The hot water is from the same general
supply but passes through a heater into which live steam is
introduced. It comes from the tap in the laboratory, where the
apparatus is installed, at a temperature of from 40 to 60°C. and
is highly supersaturated with gases; this water was used for
boiling.
A large number of metals were used in the apparatus, princi-
pally because the materials were at hand, but it was also thought
that the use of such metals followed by chemical analysis would
show what effect they have upon the water. It may be noted
that the water supply of most laboratories is pumped with pumps
which are lined with brass and copper and is then carried for
long distances through iron pipe, galvanized and ungalvanized,
so that there is little use of making the last 60 feet of pipe of some
special metal except where heat is applied.
TABLE 4
Showing the effect of the apparatus upon water. Analysis: solidsin parts per million
by Mariner and Hoskins and gases in cc. per liter
HOT TAP COLD TAP | APPARATUS
(OSS VLIETT 5 3515 Sema ae coe ge 10.46 0.74
LUD: EET ce a oe ae oe Cc 18.45 3.33
(Cait cera ChOp-a6 (Re ee te eae ee ee ares nee 2.50 On%3
Half-bound carbon dioxide............... 31.85 35 31.34
Free ammonia......... ees RN hs 0.00 0.00 0.00
PUBUNNNOIG aMMMONIA.. 00... [22 ble | 0.04 0.04 0.05
LC TEIAE 2 Jie 5 Rote ee a | 1270," | OL7Z0 5% 0.60
= TIANSU 29). aS ee 0.00 | 0.00 0.00
Chin ieee eee oe 12.00 |, -* 10/00 |" 12.00
Zine, copper, aluminum and tin.......... 0.00 0.00 | 0.00
TROD 3% ie! th RRS ee Cle Se eee 05. | 0.06 0.11
Tai COA 0) a ea nee en nee rer 43.60 | 48 .40 43 .60
Mieronestay (NGG ON sche. eee et in shies. soe L6xot) | 18.82 20.011
LGC, ¢ aan ee See ee ane pele eer ree oi eae 0.02 0.01 0.01
Seiupnunic acid’ (SOs). 4). 24, gene hoo 0.03 0.04 0.071
1 Increases shown are due to resolution of precipitated salts as shown by repeti-
tion of analysis of the water and analysis of scale from the boilers by M. M.
Wells. The magnesium in the boiled water probably varied from time to time
and the statement that it was reduced (Shelford and Allee ’12) is probably incor-
rect for most of the experiments.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 2
218 VICTOR E. SHELFORD AND W. C. ALLEE
The analysis shows that the differences between the mineral
content of tap water and the water that passed through the
apparatus are very slight. No traces of the various metals used
in the construction of the apparatus appear in the water. There
is a loss of 4.8 parts per million of calcium. Iron, magnesium
and sulphuric acid were slightly increased. The absence of any
traces of the metals contained in the apparatus is largely due to
the fact that they quickly became covered with sediment and
scale and the water did not actually come in contact with them.
The boilers are of aluminum which also becomes covered over
with aluminum oxide which is insoluble in water and only slightly
soluble in acids. Aluminum is probably the best metal from
which to make apparatus for biological purposes, principally for
this reason.
The Chicago tap water usually contains quantities of algae,
some rotifers, Protozoa and Entomostraca, the last usually dead.
The boiling process kills and cooks any plankton in the water
so that plankton-feeding fish might be able to secure food from
the tap water but not from the boiled. This, however, is prob-
ably of no importance in the cases of the fish used in these experi-
ments (see p. 220).
(2) Introduction of gases. Gases are introduced from tanks
between the upper and lower coolers and by means of the gas
introducer (fig. 1 A). The tank of gas is attached to the intro-
ducer by means of a rubber hose. The gas enters the chamber
(ch) and passes through a number of the small openings. Large
quantities of oxygen, carbon dioxide, or other soluble gases can
thus be added. If it is desired to add a soluble gas directly to
the tap water, valve G is opened and valve F is closed, thus
allowing the water to flow directly into the delivery tank. It
is necessary, when untreated tap water is used, to allow the gas
to enter only in such quantities as will go into solution in the
cooler coil, for if it comes off in bubbles other gases are also
‘removed. When gases are to be added to water that has been
boiled, but little gas is removed by any bubbles which may pass
through, and here it is sometimes desirable to allow an excess to
escape as bubbles by inserting a tee with the stem projecting
upward between the rubber hose and the withdrawal cock.
i
REACTIONS OF FISHES TO ATMOSPHERIC GASES 219
Gas escapes in bubbles through the tee and bubbles are not
introduced into the apparatus, where they interfere with the
experiment.
Ill. MATERIAL
The material used in these experiments consisted of the fol-
lowing: Young of the common sucker (Catostomus? commersonii
Lac., 8-11 em. long, adult length 45 cm.); adults and young of
the golden shiner (Abramis crysoleucas Mit., 6-15 em. long.,
adult length 15-20 em.); a few adults and many young of the
common shiner (Notropis cornutus Mit., 5-9 em. long, adult
length 12-20 cm.); young of the river chub (Hybopsis kentuck-
iensis Raf., 7-10 cm. long, adult length 15-20 em.); young of
the black bull-head (Ameiurus melas Raf., 13-15 em. long, adult
length 30 ecm. of more); adults of the mud minnow (Umbra limi
Kirt., adult length 10-13 em.); young of the rock bass (Amblo-
plites rupestris Raf., 4-10 cm. long, adult length 20-35 cm.);
young and adults of the blue spotted sunfish (Lepomis cyanellus
Raf., 4-6 cm. long, adult length 10-18 em.); young of the small-
mouthed black bass (Micropterus dolomieu Lac., 8-9 em. long,
adult length 30-38 cm.); and adults of the rainbow darter (Ethe-
ostoma coeruleum Stor., adult length 5 cm.). The stock also
included one or two adults of Ameiurus nebulosis LeSur.,
Schilbeodes exilis Nel., Boleosoma nigrum Raf., Hadropterus
aspro C. and J., and one or two young of Campostoma anomalum
Raf. and Notropis atherinoides Raf. On some occasions noted
in the text, one or two of these were included in experiments
with other closely related species. Others were used for pre-
liminary experiments which are barely mentioned.
1. STOCK OF FISH
The Ambloplites, Hybopsis, Micropterus, Notropis, Ethe-
ostoma and Schilbeodes used in these experiments, were collected
from Hickory Creek, New Lennox, Illinois, October 30, 1911.
The Boleosoma, Lepomis, and Catostomus were taken at Floos-
moor, Illinois, from Butterfield Creek, November 13. One or
2 In the body of the paper the fishes are sometimes referred to by the generic
names only, when it is to be understood that the species listed above are meant.
220 VICTOR E. SHELFORD AND W. C. ALLEE
two Lepomis taken from the Fox River at Cary, Illinois, October
21, were included. The Ameiurus, Abramis, and Umbra came
from a pond at Pine, Indiana, November 11. This stock was
divided and one part was put in a standing-water aquarium and
supplied with boiled water from time to time. This part of the
stock will be referred to as the ‘low oxygen stock.’ All other
fish were kept in aquaria supplied with running water from the
tap. During the period of experimentation, the fish kept in
good condition, with low mortality.
Most of the deaths were due to fungus attacking slight injuries
or lesions, due to gas bubbles or handling. The fish fed largely
upon small minnows which were present in the aquaria in large
numbers; also upon fish foods and pieces of fresh water mussel.
After December 1 the stock of fish was put into a large basement
tank, supplied by the overflow from a large artificial pond on the
campus. This water contained plankton and occasional inverte-
brates. The fish were kept all winter and noticeable mortality
did not begin until April.
2. BEHAVIOR AND PHYSIOLOGY OF FISHES
Fishes usually remain active all winter (Abbott ’75) appar-
ently carrying on their regular activities as in warmer weather.
The temperature of streams and of larger bodies of water prob-
ably does not fall below 4 to 6°C. before the end of December.
Our experiments were performed in November and December,
only a few being conducted in January. The behavior of fishes
in autumn is not modified by the breeding activities and fishes
may be brought to the laboratory in very large numbers without
mortality and kept alive.
In mode of locomotion the fishes studied fall into two main
classes. The first class comprises those that rest on the bottom
much of the time, swimming by darts; this type includes Boleo-
soma, Etheostoma, and young Catostomus, the latter being some-
what more like the other fishes. The second class is made up
of fishes that swim at a uniform rate, starting slowly. Umbra
Hypobsis, and Ameiurus often rest upon the bottom; the other
species do so rarely.
REACTIONS OF FISHES TO ATMOSPHERIC GASES 22t
The fishes studied represent all degrees of gregariousness. The
Abramis, Hybopsis, and Notropis are strongly gregarious. While
they tend to follow any small fish, the most compact schools
are made up of fishes of about the same size. Fishes larger or
smaller than the majority are most likely to stray. Very com-
pact schools of different species may be maintained if the fishes
are about the same size. Lepomis, Ambloplites and Microp-
terus are only slightly gregarious in captivity, Lepomis least of
all. The Umbra and Ameiurus are least gregarious of the swim-
ming fishes, two fishes rarely moving together. Of the darting
and resting fishes, none are more than slightly gregarious.
IV. THE PHYSIOLOGICAL EFFECT OF GASES UPON FISHES
The physiological effects of gases upon fishes have been but
little studied experimentally. Nothing has been done upon the
species of fish which we used in the gas gradient experiments.
While not a part of our main problem, we considered a knowl-
edge of the effects of gases upon the species studied of impor-
tance and accordingly conducted some preliminary experiments.
Some typical results of these are included here.
1. EFFECT OF A GREAT EXCESS OF NITROGEN AND OXYGEN
As has already been stated, the laboratory water supply con-
tains an excess of gas at all times, this being especially true in
the winter and spring. Fish kept in water which contains a
large quantity of gas, usually develop gas bubble disease. Bub-
bles of gas, consisting largely of nitrogen (Marsh and Gorman ’05)
collect in the fins, beneath the skin of the head, behind the eyes,
thus producing ‘pop eye,’ and in the circulatory system, espe-
cially in the heart, where they interfere with the circulation so
as to cause death. Hitherto, the disease has been noticed espe-
cially in marine fishes.
Gas bubble disease developed in the stock of fishes during the
progress of the ‘experiments. The excess gas in the aquarium
water ranged from 1 to 2 ce. per liter of both nitrogen and oxygen.
Gas bubbles developed in the fins of Ambloplites, Hybopsis,
y VICTOR E. SHELFORD AND W. C. ALLEE
Notropis, Lepomis, Umbra and Ameiurus. No bubbles devel-
oped in Abramis but they were experimentally produced on
several occasions. A typical experiment consisted of raising the
temperature of the water from 8 to 17°C. without loss of gas,
and allowing it to flow into an aquarium. This gave a large
excess of gas. Nine out of seventeen fish developed the disease
in nine hours; four of these did not recover. The fishes of a
standing water control at the same temperature, showed no
signs of the disease. The cure of gas bubble disease was accom-
plished by bubbling gases through water in which diseased fish
had been placed. Two Hybopsis and one each of Notropis,
Catostomus and Ambloplites were cured of the disease in eighteen
hours, by the bubbling of oxygen through a tall jar which con-
tained the fishes. Two Umbra with large bubbles in their fins
were placed in water through which nitrogen had been bubbled
for thirty hours and the bubbling continued. Both fish were
entirely cured in twenty-two hours. Table 3, experiment 3,
shows that, making a very liberal allowance for exposure to the
atmosphere, the nitrogen was probably increased 2 cc. per liter;
(compare table 2). On another occasion, two Hybopsis which
were badly affected, were cured in twenty-four hours in the same
manner (table 3, experiment 4). Two Ambloplites and four
Abramis were kept in water through which nitrogen was bubbled,
for twenty-three hours, and neither showed signs of gas bubbles
upon dissection, though one died of asphyxia due to the low
oxygen content, 0.5 ec. per liter (table 3, experiment 5).
We find in these experiments no suggestion that fish develop the
disease as a result of a simple increase of gas when one gas is dis-
placed by another, under one atmosphere of pressure, but rather
that the disease appears only where the gases are so much in
excess that bubbles collect on any rough or warm object in the
water. This excess may be due to a rise in the temperature of
the water or a decrease in pressure, or both. It is probably
essentially a laboratory disease (Birge and Juday ’11, p. 134).
REACTIONS OF FISHES TO ATMOSPHERIC GASES 223
2. EFFECT OF A DEFICIENCY OF OXYGEN
Duncan and Hoppe. Seyler (’95, p. 165) found that the Euro-
pean cyprinid (Tinea vulgaris) is not affected by a prolonged
exposure to oxygen reduced to 3 to 4 ce. per liter, but that when
the oxygen supply is reduced below 1 cc. per liter they come to
the surface and breathe heavily, violence of respiration increasing
as the oxygen decreases. They were able to keep these fish
alive, in a total absence of oxygen, for twenty-four hours. Trout
were strongly affected by 1.7 to 0.8 ec. per liter in two or three
hours. Reuss (10) found that an increase of oxygen decreased
breathing frequency. For further experiments and observations
concerning the relation of fishes to gases, see Knauthe (’98, p.
785, ’07, p. 148); Konig (’99, p. 32) and Marsh (’07, p. 346).
Several experiments were performed to determine the oxygen
minimum for the different species. Table 5 represents our
experience in this matter, together with current preferences
which may roughly represent oxygen content of the natural
environment.
The different rates of flow represent the following conditions:
Swift water is high oxygen content; the category ‘variable’ may
be construed as representing a condition in which stagnation
oceurs at times:and accordingly represents conditions which
TABLE 5
Showing the relative time of succumbing to low orygen content; and the current
preferences of the same species after Forbes and Richardson (’08)
TURNING CURRENT PREFERENCES
| LENGTH
IN/CM- earn. Biase oe Variable
WI CTODLEDUS ts sone cos set oe | 20 | 55 18 | 27
[CS OTe: ee 12 320 | |
Ambloplites, average............ 6 340 | 55 | 15 30
WISER Sos tah ciel pee 10 :| 355 | Bene eo hel! VIS
MMbioplnbed. ©... 602.2 2.4.58. 00 4 | -380.-|
ENN 4 34 53). Beyo4, tins SAR 10 | 376 | 45 | 36 19
ou, PLES Neer oe 105i} 400 | 32 oie 11
_ OUTDO; Be See 1080 | 37 53 10
1The ‘turning time’ is the time before the fishes turned ventral side up
224 VICTOR E. SHELFORD AND W. C. ALLEE
would be as detrimental to fish in the long run as a more con-
stant low oxygen content. In our experience with low oxygen
content, the smaller Amploblites are more hardy and the small
Abramis are more sensitive than the adults. The data at hand
suggest that there is some relation between habitat preference
and the amount of oxygen necessary to maintain life.
3. THE EFFECT OF CARBON DIOXIDE AND AMMONIA
Many physiologists hold that carbon dioxide is more important
than oxygen as a stimulant for respiratory action. In general
its action is that of a narcotic, stimulating in small quantities,
intoxicating in larger quantities, and producing death when taken
in very large quantities. Similar behavior results have often been
obtained with carbon dioxide and with acids. Both acids (Marsh
10, p. 896) and carbon dioxide are fatal to fishes when present
in quantity. Reuss (’10, p. 555) worked with the effect of vary-
ing amounts of carbon dioxide upon the rainbow trout and found
its general effect to be entirely similar to that with higher verte-
brates. Up to about 15 ce. per liter, carbon dicxide acted as a
stimulant to respiratory movements. Beyond this breathing did
not become stronger. Staggering occurred with a concentration
of from 25 to 41 ce. per liter and a total loss of equilibrium at
from 44 to 53.5 ec. per liter. Weigelt (’85, p. 82) working with
carp and trout, found that 35 to 37.5 cc. of carbon dioxide per
liter had no effect but that 50 cc. per liter was sometimes harm-
ful, while 100 ce. per liter was toxic at all times. He later (’03)
reported 5 cc. toxic to Tinea. See also Knauthe (’07, p. 125).
A series of experiments including most of the species at hand
confirmed the results of these workers with carbon dioxide. An
Ameiurus was narcotized in 163 ec. of carbon dioxide per liter
with oxygen at 1 cc. per liter. In this case the fish was placed
in tap water and the amount of carbon dioxide present was
gradually increased until anaesthesia was produced. Individuals
of nearly all the species at hand were dropped into the water
with the gas content as just given. They were all greatly stimu-
lated at first, but lost correlation of movements in a few moments
and died in ten to fifteen minutes.
fei
/
REACTIONS OF FISHES TO ATMOSPHERIC GASES 225
Weigelt worked with ammonia (NH;) and reports that 10 to 17
mg. of ammonia per liter had no effect on small fish and even
30 mg. per liter did not affect large ones, but smaller ones were
affected by less than this amount. Fourteen milligrams are not
uncommon in sewage, while 43 mg. have been reported.
V. REACTION OF FISH TO GASES IN SOLUTION
1. METHOD OF EXPERIMENTATION
It is impossible to study the reaction of fishes to gases with-
out first establishing a gas gradient. Long boxes 120 em. by
14 em. by 20.5 em. with screen partitions 5 em. from the ends
were constructed (figs. 2 and 3). The drain D placed at the
center near the top consisted of a tube with screen bottom,
opening outside. Water was allowed to flow in at both ends at
the same rate (usually 600 cc. per minute) through tees perfor-
ated so as to distribute the flow across the tank (fig. 3). The
two currents, too slight to interfere with the behavior of the
fishes, met at the center of the box. The temperature and flow
of the water into the ends must be the same if the gradient is
to be perfect.? :
Water which has been treated in the gas control apparatus,
was introduced at the end A (fig. 3) so that a gradient was estab-
lished in tank AB, while the control tank BB was alike at both
ends. The oxygen content of the water in a typical boiled water
gradient is shown in figure 2. The water coming from the tee
3 Tap water was introduced from }-inch cocks into the tee (7) introducers at
the end of the tanks marked (B, fig. 3) and at the same rate into end A from the
gas control apparatus. The flow of water was regulated by shoving the rubber
tubing connecting the tee introducer with the hose-end, onto the hose-ends as
far as possible and ligaturing them securely with copper wire. Each rubber tube
was wrapped with bicycle tape for a distance of about 3 cm. on each side of the
termination of the metal hose-end. Another ligature was usually applied over the
tape. A screw pinchcock was then placed over the wrapped portion of each tube.
With the metal valve wide open each pinchcock was screwed down until the desired
flow was secured; by measuring twenty seconds flow in a graduate the rate of flow
was determined. The flow does not have to be adjusted more than once or twice
a day and is almost constant for several hours. A metal valve is set for a given
flow with much difficulty and the flow soon falls off because of the accumulation
of sediment in the valve.
226 VICTOR E. SHELFORD AND W. C. ALLEE
contained less than 1 ce. per liter but was increased by exposure
to the atmosphere in the box. Samples taken from the boiled
water end of the experimental tank contained more oxygen than
was normally present at this end, due to the unavoidable dis-
turbance of the gradient while the sample was collected. The
experimental box was divided into three parts. Hach end third
was filled with water flowing in at that end, while the middle
was a gradient between the two.
The complete apparatus used in the experiment is shown in
figure 3. Two tanks as described above and shown in the figure
were placed side by side in the bottom of an aquarium (in the
absence of a suitable water table). The aquarium was enclosed
beneath a black hood with side curtains as shown in figure 3.
These side curtains hung loosely so that the observer could see
into the hood and into the experimental tanks from either side.
Fishes were placed in each of two dishes which were set above
the experimental and control tanks respectively. Sheets of trial
balance paper were prepared by recording kind, size, number and
previous history of the fishes used, together with temperature,
gas content of the water, et cetera. Certain vertical rulings on the
trial balance paper were taken to represent the ends, center and
thirds of the tanks. Usually the right-hand side of the paper
was used for the main record. Vertical distance was used
roughly to represent time.
Fig. 2 Shows the arrangement of the experimental tank in optical section, with
distances and dimensions. WS, tap water supply; C, hose end cock; Rt, rubber
tubing for tap water; Re, rubber tubing leading to the apparatus; Spc, screw pinch
cock over the tapped end of the hose; HE, level of the observer’s eye, 45 cm. above
the surface of the water; L, the 4-candle-power lights, 30 cm. above the surface of
the water and above the center of the respective halves of the tank (the two lights
being the only source of illumination) and between the observer and the fish; Ta,
top of aquarium wall; 7’, is the tee introducer; Sc, the screen partition; D, drain;
Rg, region of gradient; d, division points marked on the tank and corresponding
to the red rulings of the trial balance paper. The series of figures, 3.1 to 9.4 ec.,
show the amount of oxygen present in collections taken from the gradient immedi-
ately above the location of the figures.
Fig. 3 Sketch of the experimental tanks and hood in position, with the arrows
indicating the direction of the flow of water. The position of the lamps (L) and
of the observer (£) is indicated. The tank lettered (A) at the right hand is the
one usually used as the experiment in which case conditions are identical in the
ends marked (B). Other lettering as in figure 2.
lor)
~
fe
$
~
es
Sy
~~
SS
Up Rae eee oS
|
Pat)
C@---------
E
//p
{Rt
-L- 30 cm------>+
KY feigta
4.15
9.3
94
2258 VICTOR E. SHELFORD AND W. C. ALLEE
© One observer took charge of the control, the other of the
‘experiment. At a time agreed upon, the fishes were emptied into
the tank near the center and were watched continuously from
twenty to ninety minutes. At first they nearly always moved
back and forth exploring the tank, and their movements were
recorded on the trial balance paper, in the form of a tracing
similar to the graphs on pp. 233, 237, 239 and 240, but with time
written in, instead of represented by a scale. Full notes on the
reflexes, risings to the surface, et cetera, were recorded on the
left half of the sheet. From these records, particularly from
the graphs, the total and fractional time spent in each end, the
number of turnings in the gradient and in the corresponding
position in the control were determined. These are shown in
the tables to follow. Some of the graphs were transcribed and
are shown (charts 1 to 4) with the actual vertical seale correctly
represented.
At the end of the first period of experimentation, the fishes
were often removed to the small dishes while the observers
changed places. The control fishes were then placed in the
gradient tank and vice versa. Thus the same fishes were ob-
served by the same person in both the experimental and control
conditions for the same length of time. At the outset we under-
took the study of the reactions of fishes to boiled water, trying
various experiments of varying lengths and employing all the
species which were available. In this way we became familiar
with specific peculiarities and acquired skill in observing and
recording results. Some long experiments were performed with
larger numbers of fishes and their positions were read and re-
corded at five or ten minute intervals. The results are expressed
in percentages for the three major divisions of the tanks.
In general our index of the effect of treated water upon the
fishes was its effect upon the rate and vigor of respiratory move-
ments, movements of the mouth, rising to the surface, et cetera.
These aspects of reactions were observed in detail in three series
of experiments in which fishes were put into glass boxes (Reuss
10) 18 em. by 13 em. by 3 em. Two individuals were used in
REACTIONS OF FISHES TO ATMOSPHERIC GASES 229
each experiment; one was placed in a box into which normal
water was allowed to flow and another in one into which treated
water was allowed to flow. The fishes were each observed in
detail and a full record of all their movements made for a period
of ten minutes for each fish, when they were interchanged and
the experiments repeated. A detailed series of such experiments
were performed with boiled water, a partial series with carbon
dioxide, and carbon dioxide in boiled water. The respiratory
movements were increased in vigor or in number, usually in both
when the fishes were in carbon dioxide or low oxygen water.
They rose to surface often and gasped and gulped. Although
the details of these experiments formed an important part of
- the method of work, they are not presented here because their
inclusion would burden the paper with much detail not of prime
importance, from our point of view.
The chief sources of error in this gradient method lie in the
attempts at studying the fishes at temperatures higher than that
of tap water. It was found that a difference of 1° C. between
the ends of the experimental tank very seriously interfered with
the gradient. 'The movement of the fishes back and forth in the
tanks tended slightly to mix the water of the two ends but we
found no evidence that this interfered with the main gradient.
2. REACTIONS TO SINGLE GASES
It is usually difficult to vary a single factor in an experiment,
In the experiments in which we intended to vary a single gas.
the same kind of water was used in both ends of the experimental
boxes, gas being added to the water at one or both ends, as
necessity demanded. In adding a gas to the water, a very small
quantity of other gases were also added (table 1A) and any
effect of increased gas pressure upon salts in solution (McCoy
and Test ’11) of necessity took place. However these sources
of error probably in no way interfered with the essential char-
acter of the reaction of the fishes to the gas used.
330 VICTOR E. SHELFORD AND W. C. ALLEE
a. Reaction to carbon dioxide
The reactions of fishes to carbon dioxide in water are shown
in table 6, with illustrative graphs in chart 1. The experiments
fall into the following three classes: (a) Those with a difference
of 5 to 10 ec. between the two ends; in these experiments the
principal gradient was in the two center divisions as shown for
oxygen in figure 2, page 227. (b) Those with a difference of
about 19 cc.; in taking the readings of these, four central divi-
sions were regarded as constituting the principal gradient. (c)
TABLE 6
Showing the reactions of fishes to a gradient of carbon dioxide in tap water. The
experiments bearing numbers only, were performed with fishes taken directly from
the stock aquaria, while the experiments bearing numbers and the letter A were
performed with the same fishes, control and experimental individuals being inter-
changed. The control fishes of the first period of observation, after having become
accustomed to going back and forth in the uniform tank were exposed to the gradient,
while fishes accustomed to meeting the gradient were put into a tank where no
gradient existed. For description of the apparatus, et cetera, see pp. 225-229; cor-
responding ends of the controls are designated by the same letters. H indicates
high carbon dioxide and L low carbon dioxide.
| co | cRoss- | PER CENT OF | oy ae
| PE |ED CEN- | TIME IN HALVES FROM HIGH (H)
EXPT eee ee ee p=! oS ORLOw(L) | TEMP.
mot SPECIES ess pie as Pert DEG. C.
= le | Expt. |Control| Expt. | Control
| L|H} x = j ay
| |-| 8 |H|L|H|L/H|L|H|L
85) | Abramiss. Ac 2t sin ole 7 | 58 | 93 | 15 (34 | 66 | 54 46 s| 6] 0| 0 4
87 |Ambloplites..........| 2.|2 | 7|30/11] 15 | 40] 60/80/20] 8] 3] 0| 2 4
87 X | Micropterous......... [rate 7)13) 6| 15 | 46/ 54] 79} 21) ra 0 | 1 4
86 X |Hybopsis.............| 1] 2 |7-13) 12 | 9.5| 15) 85 eee a | 4
86 | Notropis......... Br fee FW oa Hy | 9.5 | 26| 74|58| 42} 7] 0} O| 0 4
46 |Hybopsis.............] 4 2.5 21] 48) 22| 30 | 24|76/45|/55| 8| O| 2| 3} 6
46A |Hybopsis.............| 4| 2.5)21| 37/16] 30 | 18|g2|35|65| 6| o| 1| 0 6
44 |Ambloplites......... 3|7 | 72/23/13} 30 |18| 82/57/43] 1| o| o| 1] 6
45 |Abramis..............) 3|7 |72|71| 65] 30 |13| 87| 84/16/19] Q] 0| 1| 6
45A |Abramis........:.....] 3] 7 | 72] 28] 52] 30 8/92) 53) 47) 6) 1) 11) 5 6
47 - |Lepomis..............| 4 |12 | 80] 69|101| 30 | 42/58| 50/50) 16} 2) 4| 6] 6
7A |Lepomis..... | 4l12 |80|46|58| 30 | 8/92|70|30|19| 1] 1] 3] 6
48 Amelurys.......... 2 |12 | 80 | 34) 46 30 | 12] 88} 45 | 55/13) 0 6 4 6
48 A |Ameturus........... | 2}12 | 80/15] 24] 30 | 10] 90 32 69 9| 0] 12| 5 5
49 |Umbra....... | 2}12 | 80/14) 23] 30 | 5|95|40| 60/10] 2} 4] 2] &
S/R Anita. ooh ooo: 2/12 | 80|22| 37] 30 | 16| g4| 38] 62] 13] 0| 16| 12 5
52 |Catostomus......... | 3/15 | 80|100| 4] 30 | 25|75| 2 | 98} 11} o| o} 4) 5
52 A |Catostomus...... “a 3 15 | 80) 39/85] 30 | 23] 77|40| 60/17] 4] 2) 5] 5
53 Notropis........ | 3/6 | 71) 46/133} 20 | 13| 97| 48) 52/41) 7] 5| 4] 5&
53. | Notropis......... ...| 3/6 | 71/64) 44} 90 | 20) go| 40 60} 41) 2] 0} 2| 5
~ 2 Lda ss OS! OS ! 2h Yaa eS es ee ed Ce | ed a ad et ——
(SS le ae . | 54 | 1501081] 788/760 | 494 |416 1584871 (929 |260 | 29 | 64 | 60 |
BVMMNB sce nichizeoes ness | [7-5 | 54] 39] 38) 24 | Qt] 79) 48) 52) 13) 2)3 | 3)
REACTIONS OF FISHES TO ATMOSPHERIC GASES 2a)
Those with a difference of about 65 ec. between the two ends,
the whole tank being regarded as gradient because the carbon
dioxide drifted across to the tap water end so as to raise the
content from 5 to 13 ec. The amount of carbon dioxide was
markedly less at the surface, due to loss into the atmosphere.
Occasionally the fishes, when put into the center, moved into
the tap water end and came to rest without encountering the
high carbon dioxide. In such cases they were driven to the
center with the hand or were roused by jarring the tank and a
corresponding number of drivings or tappings was effected in
the control. The reactions of the fishes to the carbon dioxide
gradient fall into the following three main classes: (a) They
~ entered the carbon dioxide water with little perceptible hesitation
and gave some definite reaction while there. (b) They gave an
avoiding reaction upon reaching the increased dioxide. (c) They
reacted by rising to the surface.
The reactions of the first type were given by most of the
swimming species usually at the beginning of the experiment in
which carbon dioxide exceeded 20 ce. per liter. With the excep-
tion of the Abramis, the fishes gave a coughing reaction in which
the mouth was thrown wide open and the jaws slightly protruded
with a sudden jerk. The same reaction sometimes took place
more slowly and may be characterized as a yawn. In all cases
observed, the gill movements were increased. The second type
of reaction usually occurred after the fishes had tried the highest
carbon dioxide and had given one of the reactions just mentioned.
The commonest of these reactions may be characterized as a
testing or backing-starting reaction. The fish moved forward and
acted as though it encountered a sheet-rubber wall which it
carried forward for a short distance but which in turn through
its elasticity, caused the fish to rebound for an equal distance,
probably 2 to 10mm. _ In other words, the fish suddenly stopped,
backed a very short distance and immediately moved forward
the same distance again, usually repeating several times. ‘The
fishes sometimes turned back after giving this reaction, and
sometimes turned back without giving it, and without any other
characteristic movement. The third type of reaction was some-
times given following the other two but was especially charac-
232 VICTOR E. SHELFORD AND W. C. ALLEE
teristic of Umbra and Notropis. For example, Umbra in experi-
ment 49 (table 6, p. 230) tried the high carbon dioxide during
the first three minutes, began rising to the surface after about
two-and-one-half minutes and turned back from the center dur-
ing the next three minutes, then tried the high concentration
during the two minutes following. After this the fishes remained
for eleven minutes in the low end, spending much of the time at
the surface and turning back often only a short distance away
from the low end. In chart 1, experiment 48 A, the bullheads
tried the highest concentration during the first two minutes, went
only to the center during the next five minutes and then rested
in the low end for fourteen minutes. They then tried the higher
concentrations during the four minutes following and came to
rest in the low end remaining there until the end of the
experiment.
The invasion of the high carbon dioxide was in nearly all
cases followed by several turnings back at the center or resting
in the low end. The sensitiveness of the fishes appeared to be
in some way affected by exposure to the high concentration and
the invasion of the high concentration was not repeated for some
time. This is true in a general way of all the fishes and for all
the concentrations. This tendency appears in the reactions of
Abramis to the low concentrations (experiment 85, chart 1).
An inspection of the graph here shows that in the second half of
the experiment considerable time was spent in the high end,
Chart 1 Showing the reactions of fishes to a carbon dioxide gradient in tap and
in boiled water. For concentrations, see tables 6 and 7. The space from right to
left between the scales represents the length of the tanks in which the fishes went
back and forth. The broken vertical line represents the centers of the tanks;
the character of the water flowing into the ends of the tanks is indicated at the top
to the right and left sides of thespace. The vertical distance represents numbered
minutes, which are subdivided into ten second periods. The lines as drawn, where
horizontal, represent the movements of the fishes; and vertical, the resting of the
fishes. The method of representing the facts is shown in the experimental portion
of experiment 55. The fishes rested most of the time in the tap water end but
made occasional excursions to or beyond the center of the tank; the duration of
these is indicated by the scale at the left. The number of fishes used is indicated
in front of the scientific name of the species. The dotted lines represent the move-
ments of a single fish; the solid lines of two or three fishes as the case may be.
REACTIONS OF FISHES TO ATMOSPHERIC GASES 233
Experiment 85 Control! 85 Experiment 48 A Contro! 48A Experiment 55 Control 55
3 Abramis 3 Abramis 2 Ameiurus 2 Ameiurus 3 Amblioplites 3 Ambloplites
Tap+ Boiled +
Tap Tap Tap CO: Tap Tap
SS =a
—- 1 -
: : =)
jontg
(PARE
pit
|
Jim
peyert
potter)
Hibbliepegpass
ee
H
jinent
|
tfireit
|
UO NCOCOT COLL OCL L
bikin
2,
yi Vt
o
Piyleeieienin
pobeboeelini
TO ee
|
Te
yiyer
{
(I
iat
im
Oy
Control 86
trite
Calculated 95 min.
2 Notropis 3 Notropis
Tap Teap+CO. Tap Tap
na
it Oc)
|
|
rreptrsesetaeiendeat
reeenty
Heetedieeeetiree tia
Seg P= nN on oye eee erg he a er So a
THE JOURNAL OF EXPERIMENTAL ZOGLOGY, VOL. 14, No. 2
234 VICTOR E. SHELFORD AND W. C. ALLEE
which was followed by resting in the low for two minutes.
In experiment 53 Notropis showed similar alternations of. inva-
sions and avoidances which took place more rapidly because
the fishes moved back and forth more rapidly. Notropis
(experiment 86, chart 1) invaded the high concentration only
once after the first minute. In the.controls for all the experi-
ments, the fishes moved back and forth with more or less regu-
larity and on the average with little apparent preference for
either end.
An inspection of table 6 and the graphs (pp. 230-233), shows
that activity is much greater in the experiments where low con-
centrations were used, as indicated by the number of crossings
of the center. The activity where high concentrations were used
was variable. The percentage of time in the low half of the ex-
periment is in all cases much greater than in the high. Time per-
centage and number of turnings back in the gradient portion are
the best indications of reaction. The number of turnings back
from the higher concentration was greatest in all cases. In the
controls, the number of turnings was usually nearly equal.
The reactions of all the species of fish experimented upon are
quite similar although they belong to several taxonomic groups.
The data in table 6, if averaged, show an almost equal number
of crossings of the center in the experiment and control, but
there is a marked time preference for the low carbon dioxide
half of the experimental tank and a nearly equal division of
time between the two ends of the control. The average number
of turnings back from the higher concentration is thirteen as
opposed to one from the lower. The average turnings in the
controls are three from each half.
Preliminary experiments were tried with Schilbeodes, Ethe-
ostoma, Boleosoma, and Hadropterus. While the experiments
were not carried far enough to give results of definite value they
suggest that these swift water fishes which probably encounter
very little carbon dioxide, may react less definitely to it than
the fishes which live more often in the presence of carbon dioxide.
Table 7 shows the reactions of Abramis from the low oxygen
stock and Hybopsis to a carbon dioxide gradient in boiled water
REACTIONS OF FISHES TO ATMOSPHERIC GASES 235
with oxygen 2.4 ec. per liter. The reactions were more marked,
as shown by the graph (chart 1, experiment 55, p. 233), but
when compared with other experiments, the depressing effects
of the low oxygen and carbon dioxide are suggested. The fishes
showed greater disturbance in these experiments than in the
presence of more oxygen.
TABLE 7
Showing the reactions of fishes to a carbon dioxide gradient in boiled water. Oxygen
content of the water was from 1 to 2.4.cc. per liter. Other data as in table 6.
CO2 | TURNED BACK |
| CROSS- PER CENT OF
IN CC. | ep CEN-| TIME IN HALVES| LN GRADIENT |
PPR TER | TOTAL OF TANK Oe ate (2) |
oe SPECIES NOs |e eee TIME or Low (L) esate
5 | =| IN MIN. =
eseall ey Expt. Control} Expt. | Control)
L|H| #| 3 |
As Tae | 1G 4) Ise |) bral ASEM ast) ib:
| | | ee oa
70 Ey bopsise....--aee|) 3 2.5)/50.0) S110" 10) |) 2 98 | 54) 46) 9) O| 3 6 | That)
70 A Ey bopsist. ss ..sssce: 3 | 2.5|50.0) 15 | 2 10 4/96/80] 20] 9] 0} 0} O 7.8
84 Abbramistecc.-seceole ole so|e4, 0136: |) 87 27 «=| 12) 88 | 29| 71| 75) 4/41) 4) 8.5
BIvAy | -Abramis. 2). ... 2... : Si leeeo! 24.0) 21) 76] 15 | 9] 91} 63 | 37 | 34] 0| 4] 24 8.5
1 | | | |
TRoG | aera Ste Bape re beak Colona eee 80 |275 | 27 |373 |226 i174 127 | 4] 48} 29 |
Peeeen ee fee E ONO. as Se Ae es | 20 | 69 7| 93 | 56| 44) 32] 1] 12].7 |
TABLE 8
Showing the reactions of fishes to a carbon dioxide gradient in experiments lasting
forty minutes or more. The numbers from 1 to 6 refer to six equal longitudinal divi-
sions of the tanks (indicated by the d’s in fig. 2) when counted from the low (L) car-
bon dioxide end of the experimental tank and the corresponding end of the control
tank. The numbers beneath them represent the percentage of total individuals
(number of individuals times number of readings) recorded in each division at the
time of the readings. Readings were taken every five minutes.
COzIN cc. ie
5 4 | i Sana L—EXPERIMENT—H | CONTROL
SPECIES A al | l 7 Vener]
| = i444) Expt.
Bele (hte (Se legates | alpen | ee lds | 2) | aban
e | 2 | 2 age L|a | |
Ambloplites...... 63 | 15 | 8 | 3 | 6/50/69) 7|21| 3) 0) 0/42) 5/10) 12) 5) 26
Lepomis..........| 64 12 SSP het 6 | 50 | 63 | 22; 15) 0] O| OJ 23 | 3 | 11] 11 | 36
Alpranmisess..- 2...) * 65 8 12 elb | 6 |50|88| 12) 0} O| 0| 0} 34 | QAO on elon) 22
PANVODAPES PEL CONG: << .lsc'te ais s.6e.ck eile ee erals Sn a TB YAy ate a ad PAM at | 0| 0} 33 | 14) 8) 9) 11) 25
\
The results of several long experiments with readings every
five minutes are shown in table 8. Here the fishes showed a
236 VICTOR E. SHELFORD AND W. C. ALLEE
marked preference for the low carbon dioxide, which was main-
tained for more than an hour. They showed no tendency toward
becoming acclimated to the carbon dioxide during this period.
b. Reactions to oxygen
In the experiments in which oxygen alone was varied, boiled
water was used at both ends and oxygen added at one end.
Titrations of collections from the ends of the tank showed a
gradient of 4 to 10 ce. per liter, but as has already been stated,
these collections tended to disturb the gradient so that exact
differences could not be determined. The water as it left the
deaérating machine contained about 1 ce. of oxygen per liter and
this probably gave a minimum gradient of 5.cc. per liter at the
bottom of the experimental tank. The reactions of the fishes
TABLE 9
Showing the reactions of fishes to an oxygen gradient in boiled water. Numbers and
abbreviations as in table 6, p. 230. Experiment 72x was run with different fishes
from Experiment 72 because the fishes in 72 were not accustomed to seeing the experi-
menters and the results were interfered with by fright. Abbreviations as in Table 6.
TURNED BACK
CROSS- PER CENT OF
IN CC. | aD CEN- TIME IN HALVES | IN GRADIENT
ee TER | TOTAL OF TANK FROM HIGH (H)
EXPT. =SnSE no.| LITER TE AND Low (L) | TEMP.
No. 5 Eckay : pE@. C.
eine Expt. |Control) Expt. | Control
L/H|@\¢
ica} 8 FES is SE | | |
Exp.9
75 hepomiss..0.. scan 3 | 1.2/11.7| 59 | 22) 20 58 | 42 | 44] 56} 11) 4] 2] 1 Con.7
(¢2 me. || INOUGDISS....2 ne ee 3 | 3.7 7.9) 67 | 33) 20 75 | 25 | 66.) 34.) 12) 43) 2) 2 8
72 Wotropisicctc.ce eee 3 | 3.7] 7.9) 68 9| 20 46 | 541) 48 | 52) 6}12|) 1] 1 8
73 Hybopsis............| 3 3.4] 7.9} 38 | 49} 20 78927 |'-42)}\58))— 34) Ba) 4a 9
73.A | Hybopsis............| 81] 38.7] 7.9) 80 | 125) 20 56 | 44| 54/46) 4/13] 6] 0 9
74 Catostomus.......... 3 | 3.7 7.9) 29 | 163} 20 56 | 44/53/47) 1] 7) 7] 4 9
74.A | Catostomus.......... 3 | 3.7] 7.91185 | 3| 20 | 48| 522] 41/59] 5] 30] 0] 0 9
76 Abrams... 6...) Oil Legilioa| Sago) aera 27 | 733| 46 | 54) 2] 0} 0} 0 9
69 Ambloplites......... 3} 1.2/12.0) 6 | 15) 20 68: 47 | 215) 78) as) ae ee 8
69 A | Ambloplites......... 3 | 1.2/12.0) 9] 19) 20 71 | 29 | 46; 54) 0} 2] O| 1 8
VAVOPALO 5.5 o5.< oof sie tle s vv. Oe 2.7| 9.5/52.6) 47) 20 56 | 44 | 46 | 54] 4.5) 12) 2] 1
1 One fish lay in the low oxygen over four minutes; cause unknown; control
often resting.
2 No evidence of a reaction.
* Fishes swam at the surface of the low oxygen end for seven minutes and thus
were giving a reaction to the vertical, but not to the horizontal gradient.
REACTIONS OF FISHES TO ATMOSPHERIC GASES 237
Experiment 72X Control 72 Experiment 75 Control 75 Experiment 80 Contro! 80
3 Notropis 3 Notropis 3 Lepomis 3 Lepomis 3 Lepomis 3 Lepomis
Eoved Boiled Tap Tap Sta Boiled Tap Tap See Tap Boiled Tap
2
ees
Oy
mney
any
ON
ate
hint
in)
|
|
boo} tg
|
\
n=
TT wa
== 1 == == 1 == ES
== == a5 z= | =
= i a == = = = ==
z i =z 2 == = S
22 ee as a =
== i == = = ==
rat \ === aac ==. = =
SS ! abe = I == = = ==
= = Fm == == Wase=e == ==
aS i —— =e i == = =
—— aS tO! f= l =- -—
== - = it 1 = == ==
=10— = a 4 i sl= | S0— re
== == == ! == =< ==
= == = ' == == ==
= ——s ==> == “} a es
== = -- - - = ' = =
== == == =) ese } ==
== == a == Y== ==
== = == = =) ==
= =H 5 _ Soe ee
a= = S=- == st =
== == i 22 == = aes" ==
= Sates" 2s = SS 00 Se
a == ie == 25. pees 2 ==
— == ~~ = == === == ==
== == as Se =A (pe == = = == ==
== == SS fi == == == ; ==
== == i j = S coal == == z= , ==
== =p == I == == = 3 ==
—10 SS ee a SVB SS sso 5555 = 10 age = = - - ~~ - - 7
Chart 2 Showing the reactions of fishes to an oxygen gradient in boiled water
and to the effect of boiling. For oxygen concentration, see table 9; for experiment
80, table 15. For a description of the method of charting, et cetera, see chart 1,
p. 233.
(table 9) were in general very indefinite, but the activity was
greater in the experiment, due to the stimulating effect of the
change in character of the water in passing from one end of the
tank to the other. Time preference was not strongly in favor
of the high oxygen end except in experiment 72x (chart 2). In
72 fright appeared to enter into the behavior and there
was a time preference for the low oxygen end. In all cases,
turnings in the gradient were more numerous from the low than
the high, though the differences from the control were not so
great as in the case of carbon dioxide. Chart 2, experiment 72x,
C. ALLEE
238 VICTOR E. SHELFORD AND W.
shows the graph of the reaction of Notropis, which was the most
decided response given. It will be noted that Notropis turned
back from the low concentration several times during the first
fifteen minutes of the experiment but became very indefinite
during the last five minutes. A typical reaction of Lepomis is
shown in chart 2, experiment 75, of Hybopsis in chart 4, experi-
ment 73. The former was plainly indefinite in its reactions
throughout, but turned back more often from the higher concen-
tration. Ambloplites and Abramis did not turn back in the
gradient and gave no definite reaction. For the other fishes,
the number of turnings in the gradient indicates some reaction
to the absence of oxygen.
TABLE 10
Showing the reactions of fishes to a nitrogen gradient. The gradient was estabished
by running boiled water into both ends of the experimental tank and adding the high
nitrogen atmosphere at one end and enough oxygen to balance the oxygen added with
the nitrogen, at the other. Controlintap water. The difference in nitrogen secured
was only 3 cc. per liter.
O2_
AND Nz CROSS-
IN CC. |ED CEN-
PERCENT OF | TURNED BACK
TIME IN HALVES FROM HIGH (H)
PER TER TOTAL OF TANKS OR Low (L)
ge niags SPECIES NO.) LITER | | Tie | TEMP.
NO. - — IN MIN.
Os), Ose -aGo | Expt. Control, Expt. Control
2.2812.44, 2] 3 Nel Ne
pe aiedlise HiL) =|) |e 4 2 ee
| Ne| N2| |
66 | Ey bopsis 225.22 35 Se o a 7 | 33} 13 | 20 | 17 | 83 | 52) 48 1 0}) .<40) 28 10
66'A. | Hybopsis: 22.2 4..-4]/ St} (45 7) 12 69 20 | 31 69 32 | 68 0 1 0 | 3 10
67 | Notropis.............| 3| 4] 7/60) 90) 20 | 26| 74) 59) 49| 9) 3) 8] 12) 10
| i {
c. Reactions to nitrogen
We were unable to secure pure nitrogen and did not succeed
in putting enough of this inert gas into solution under one atmos-
phere pressure to duplicate the high nitrogen content which is
sometimes found in the deeper waters of lakes. Two preliminary
Chart 3 Showing the reactions of low oxygen Abramis to boiled water and of
Ameiurus to boiled water with acetic acid or ammonia added. Ammonia was
used in experiment 39; acetic acid in experiment 42. For detailed discussion,
see pp. 247, 252 and tables 13, 17 and 18. For a description of the method of
charting, see chart 1, p. 233.
Experiment 25X Experiment 42 Controls 42+39 Experiment 39
Calculated 30 min.
Control 24
Experiment 24
3 Abramis 2 Ameiurus 2 Ameiurus 2 Ameiurus 2 Ameiurus
Ta
3 Abramis
Ta
Boiled +
Ammonia
SS Sin Ses Bs
Tap
Tap
eS ee
—
Boiled
Tep
Tap
Boiled Ls}
p
f ' |
iW OC PE RY at Fe en nN Pet
|
bas utes saa
Wirrrefrrerefisereperees
neni pee eit aaa Perce ee eea eee tanes
crprsrt iid Jiirie lott fo
OGRE ICED CRT WTS WCNC ACL
PO COPEL Ore Cee creer eed
| ve | |
Lye] tit
He a
CO
iy y
ea Eee aT ea rE Ae Peter oer Are alia GARE SitPnGe
prec aniten| in
ry i =
Peer beere yee DOT COC CITI CACC. ICICI CCL RI PUPPET
mitt none
BTU WL PCL A
‘ins i funni
=
TN a mre ene Veen ey Hrbattts
Ne Ge pecge peerreriye
Ste oe:
1 TY a
thetarti
PO Qreerp icra ty et
Ty ANN
9 ee ee
Be CCRT TTL COCCULNCY RULICICATS CIAO SEA
=
= We UT} j
Savane
sz &
A) wee
5 -}-—- —-|-]|—— — -fJ- —- —- —-— - — — —-]-
2a 4|4
| fo}
ay OY
itech | vottfiirer
tril binntpeniti ftir ze S TO
i ' +t o ll ,
H a ° {
: H ' GO —€ ao '
Sa) om '
1 SW ee bet al I [= fe (
- } ; o 3 =\-W-|- == — == a at =\l- = He — — I ee
j H Bi gk !
+3 f =
4 i 5 o 2 :
j id , 2) ©
/ ) ' x Lead
peeberenep te WwW
\b
CO em
} bd | | = Tere
ENN MAN CRU AMON CN MM OC OC CC
Hotter} ieee pedir TO OO) OO OC vit
| Pree ep Pade siaduiorclent
sel ovine Ps mentee fin Peper peepee
Experiment 46 A Experiment 56 A
Experiment 71 Experiment 71 Experiment 73
Composite
(Control)
control
rotjitiirl
TM
POL b | reir
Tap
Tap
a ee tan
TN NN ee
frist
Rae aC
een A
|
ttt
Hitt
recirtti
Cer fitter etree} cei
\e
HPere yay
| loti
UL LL Ou
a
With ii ficiirfinyee]
i bt CT PI
' ES Te
1
i me tift
15
1 [iinet fits
<
es
as
oh
HPUETEEDGEEPE eater
4 — — ay
REACTIONS OF FISHES TO ATMOSPHERIC GASES 241
experiments were run with the two most sensitive species
Hybopsis and Notropis). The oxygen content of the water was
raised by the addition of the atmosphere from the tank (nitrogen
92 parts; oxygen 8 parts) and a small quantity of oxygen was
added at the other end. This made the difference in the oxygen
content of the two ends only 0.16 cc. per liter with the most
oxygen at the high nitrogen end. In the three experiments,
there were no turnings but a time preference for the low nitrogen
end. The graph of the reaction of Hybopsis gives no good evi-
dence that the fish reacted to nitrogen but only a suggestion
that they may react to a nitrogen gradient, and that they may
select the low concentration. However the fishes may have
been avoiding a slight oily odor which was detectable in the
high nitrogen atmosphere. Because of this and the small differ-
ence in nitrogen that could be obtained and the difficulty of
manipulation, the experiments were not carried further.
3. REACTIONS TO COMBINATION OF FACTORS
a. Boiled water
The effect of boiling the water in the apparatus is shown in
table 4 (p. 217). The water lost most of its oxygen and much
of its nitrogen, the nitrogen content being reduced from 18.45
to 3.33 ec. per liter. The free carbon dioxide was reduced from
2.5 to 0.7-cc. per liter; 1.2 cc. per liter of half-bound carbon
dioxide was lost in addition to the changes in salt content already
discussed. The fishes were then reacting to a difference in salts,
carbon dioxide and nitrogen and oxygen. When the higher
oxygen was chosen, the fishes of necessity selected the higher
nitrogen and the higher carbon dioxide. We have clear evidence
that the fishes selected the lower concentrations of carbon diox-
ide when the minimum was that of tap water but we have no
evidence concerning the optimum amount of carbon dioxide for
fish. It is therefore difficult. to interpret the results of such
Chart 4 Showing the relative intensity of reaction to the various factors
employed in experiments with Hybopsis. The control given is a graphic repre-
sentation of the average of all the controls.
242 VICTOR E. SHELFORD AND W. C. ALLEE
experiments as we are about to describe. The stock used was
divided into two parts; one in tap water, and the other in water
with less than 1 cc. of oxygen per Liter.
1. High oxygen stock. Nearly all the fishes kept in tap water
reacted with some distinctness to the boiled water, but con-
sidering only time preferences, the Etheostoma and the Amblo-
plites did not react to the gradient, since they showed a time
preference for the boiled water in some experiments and for tap
water in others. All the others gave a preference for the tap
water, though in many cases it did not exceed that shown for
one end of the control where the two ends were identical. For
example, the control individuals of Hybopsis (experiment 10)
spent 90 per cent of the time in one end of the tank, the experi-
mental fishes only 76 per cent in the tap water of the experiment.
Such cases are explainable when we consider the amount of
activity as indicated by crossing the center in experiment and
control. The control fishes in this case crossed the center only
one-sixth as many times as the experimental fishes. This was
characteristic and was probably due to the stimulating effect of
encountering a change of water.
The fishes which reacted with greatest precision to the boiled
water gradient were Hybopsis (chart 4, experiment 71), Microp-
terus, and Notropis (table 11). The reaction of Lepomis in experi-
ment 13 serves as a typical case of the reactions to the boiled
water. The fishes showed a time preference for the tap water
and ten turnings back from the boiled against six from the tap.
The turnings of Lepomis and Hybopsis, which react similarly,
were not characterized by any striking movements of the mouth
or opercles though in the experimental tank the fishes showed a
disturbance due probably to lack of oxygen. That is, they gave
characteristic risings to the surface and gulpings with emission
of air bubbles, in the boiled water end, and some hesitation in
crossing the center, not shown in the control. Micropterus is
apparently one of the most sensitive of fishes. Our original
stock consisted of only four specimens, three of which died in
water containing less than 1 cc. of oxygen, while confined there
in the second experiment attempted (see p. 245). In the one
REACTIONS OF FISHES TO ATMOSPHERIC GASES 243
experiment performed the experimental fishes tried the boiled
water three times during the first eight minutes. During the last
half of this period they showed some disturbance in excess of that
shown in the control, by rising to the surface and opening the
mouth. At the end of eight minutes the fishes showed their
first tendency to turn back at the center. After trying the
boiled water three times more, they began turning quite regu-
larly. This was continued twenty-six minutes with trials of the
boiled end being made every five minutes. The trials were
accompanied by some of the typical avoiding reactions, especially
by rising to the surface until the fins protruded.
The graph of the reaction of Notropis (experiment 2, table 11)
showed less definite turnings than Hybopsis or Lepomis and
TABLE 11
Showing the reaction of fishes from the high oxygen stock to a boiled water gradient.
The meaning of abbreviations, et cetera, is essentially as in table 6, p. 230. Controls
in tap water. Gradient between tap and boiled.
}
PER CENT OF TURNED BACK
IN CC. eee TIME IN HALVES FROM BOILED (B)
| PER aa (TAP, aeons OR TAP (T) IN
EXPr. | _. |LENGTH| “TER ) TIME ) SDE
NO. | Eee als NO-| IN cM. IN MIN. =
a! TBE?) Bay | Bee
| | | ° 5.4 hs He = j
| an ee BlBl¢ ¢ BlBi g/g
laid Ala |O;|O|s/a/O|o
Iaeee aI Re OTIN Ss. <6 <- -o 3| 45 |9.4|3.1|112177| 60 | 67|33|49|51| 6|10| 7! 6
3 Ambloplites......... See eSe tke (A139 |) 60" W730 271 3) [e2ze| eo ee
9 Micropterus......... 2} 8-9 9.4 3.1) 43) 29 40 64 | 36|85/15/] 4] 8] O| 0
21 Etheostoma......... 3] adult | 9.4.3.1) 77/59) 40 | 554) 45/38) 62) 1) 2) 7| 4
21A | Etheostoma......... 3| adult | 9.4/3.1) 33/57| 40 | 411| 59/64) 36| 6] 3] 4| 2
12 Catostomus.......... 3 9.4) 3.1 106) 3 30 | 64 | 36 | 28 7237 Mette Ola
4 Ameiurus............ 3113-15 |3 |1 | 55|52) 30 | 65| 35 | 49 BE |) ion 34
1 Albramis:..50..025<<0: 3| 6-10 | 8.1/0.5) 2737) 60 | 80| 20) 41) 59/ 0} 0| 3| 3
IVAN \pAlbramis--222526--|' 3.) 6=10) | 8:1])0.5) 21 60 | 62] 38 0| 3
11 Ambloplites.........| 3] 5-6 | 9.4) 3.1) 40) 69| 60 | 272| 73 | 69 Str OMmete le || 8
10 Hybopsis............| 3] 7-9 | 9.4) 3.1) 270) 46| 40 [ie 18 10} 90) 3|78| 2] 3
2 | Notropis?....:....... | 3 8.0 1.0 248) 36 {ite} | 69 31] 14|86| 0| 73] 0| 0
] | | | } |
3 = r | ia |
LESS a SRE Mer ROLLE tie bce | 95 | 51| 43 62| 38) 47) 53| 3 | 14/3 | 3
1 No reaction to the gradient.
2 The fishes first encountered the boiled water which depresses the general activ-
ity of Ambloplites, p. 245.
3 Two N. atherinoides were included in the experiment here by mistake, which
may account for a more definite reaction than was shown in table 15, experiments
77, 77 A and 79, where no atherinoides occurred.
244 VICTOR E. SHELFORD AND W. C. ALLEE
avoidance did not begin until the fishes had tried the boiled
water about twenty-five times during the first twenty-five min-
utes. In the Hybopsis experiments the fishes began to turn
back after about fifteen trials of the boiled water during the
first ten minutes. The testing or backing-starting reaction was
given when the boiled water was encountered and while the
fishes were turning back. ‘The experimental fishes began rising
to the surface and taking the surface film for respiration at about
the time the turnings at the center became frequent. In spite
of the marked difference in reflexes, et cetera Ameiurus reacts
similarly, as shown in experiment 4, table 11. The control fishes
went back and forth quite symmetrically, turning at the center
when going in either direction and showing no marked preference
for either end. The experimental fishes reacted to boiled..water
by showing a time preference for the tap water, by vigorous
opercular movements, and by gaping and rising to the surface.
There were no turnings back from the boiled water and the two
from the tap water are due to the fact that the fishes thus turn-
ing were swimming at the surface of the water and encountered
the center drain.
The behavior of Abramis (experiments 1 and 1B, table 11)
was in some respects similar to that of Notropis. The control
behavior was very similar, being simple symmetrical back-and-
forth movements in both cases. The behavior of Abramis in the
experiment was however different, in that the fishes rarely
turned back more often from one-half than the other but estab-
lished an apparent preference for one end and did not move
out of it.
Ambloplites is the most peculiar of the swimming fishes which
we have studied. In experiment 2 (table 11) the control graph
would make a good experimental graph for most of the fishes.
In this case the fishes acquired an apparent preference for the
end of the control corresponding to the tap water of the experi- .
ment and turned back eight times from the half corresponding
to the boiled water. In the case of the experimental fishes, on
the other hand, two of the three moved into the boiled water
when they were put into the center at the beginning of the experi-
REACTIONS OF FISHES TO ATMOSPHERIC GASES 245
ment. The third, which first moved into the tap water, soon
joined the other two fishes and all remained in the boiled water
for thirty-four minutes. Then all began moving back and forth
showing a time preference for the tap water but without turn-
ings or other characteristic activity. In experiment 3 (table 11),
on the other hand, all the fishes moved to the tap water end of
the tank at the beginning and remained four minutes except for
a single excursion into the gradient. After this time they began
to go into the boiled water but showed a marked time preference
for the tap water end. In the box experiments (p. 230) and else-
where, the boiled water decreased the activity of Ambloplites
so that if for any reason the fishes came to rest in the boiled
water for a time, their tendency to leave was decreased rather
than increased, as is usually the case.
The darting fishes, Catostomus and Etheostoma, reacted to the
boiled water and in fact to all differences in water in a somewhat
different way from the swimming fishes. In the experiment the
reactions of the fishes were most erratic. All of their movements
were dartings and restings, accompanied by risings to the surface.
Catostomus gave off bubbles of air and in two or three cases
leaped out of the water. However, they finally worked out an
apparent preference for the tap water.
Long experiments with five or ten minute readings were con-
ducted with most of the species. Since some of the fishes, such
as Abramis, often established an apparent preference for one
end or the other without testing both, it was thought advisable
to confine the fishes in the boiled water for a time to permit
them to become affected by the boiled water before the readings
began. While this later proved to be of little advantage and
often undesirable, it was continued to make the series uniform.
The results are shown in table 12, page 246.
Notropis which reacted definitely in the closely observed ex-
periments (tables 11 and 15) here shows a decided preference
for the boiled water. The fishes began rising to the surface—
one of their very definite reactions to unsuitable water—and
kept this up throughout the experiment. They tend to move
about in circles near the same spot when giving this reaction and
246 VICTOR E. SHELFORD AND W. C. ALLEE
while it is a definite reaction to the boiled water, the fact is not
evident from the table alone. Lepomis were indefinite in their
reactions, showing a slight preference for the boiled water in one
case and for the tap in another. Abramis showed a clear pref-
erence for the tap water or tap water and gradient; Ameiurus
showed a preference for the boiled water in the second two
experiments. The Ambloplites showed the same reversal of
preference as was shown in the earlier experiments. Catostomus
showed a slight preference for the tap water while Etheostoma
showed a preference for the boiled water, as in one of the experi-
ments described above. When we note the decided preference
for one end or the other in the controls, we are justified in con-
cluding that on the whole, the reactions of the fishes to the
boiled water during the longer periods was indefinite rather_than
clear-cut.
TABLE 12
Showing the reaction of the fishes from the high oxygen stock to a boiled water gradient
in experiments lasting an hour or more. Since some of the fishes tended to rest
in the tap water end of the experiment and not to come into contact with the boiled
water, they were confined in the sixth of the tank nearest the boiled water end for
thirty minutes or more. At the beginnings of the readings they were released and
usually went back and forth in the tanks. For the distribution of the gradient in
these experiments, see figure 2, p. 227. The boiled water at the point of inflow,
contained less than 1 cc. of oxygen per liter. Nearer the surface the amount was
probably larger. Controls in tap water; confined as experiments.
| EXPERIMENT CONTROL
EXPT | | commrtin Felgen aS = | no. TIME) no. READ-
mer. | SPECIES BOILED a | 3 2 | 8 | ue Fish BIW: INGS
| | IN MIN. = 3 | a B |Os Os | i
ee)” a ee o S oF eee eee =
18 | Abramis..........| 35 | (Seb )| yo |eae | solute a) ea | ame ie 4
18A | Abramis........... 35 55 1 | 44 | 25 | 41 | 94 Ale & 13
18A | Abramis........... 35 41 | 52 7 | 88 | 37 |. 25 Tulane 13
19 | Lepomis......... 5 | 51 | 9 | 40 | 21 | 25 | 54 5 5 12
19.A | Lepomis..... ae 30 | dde if) 242) 4 4a I) 82) (29 | 39 pees 19
16 | Ambloplites....... 50 4341 10) 4701) 55 | 20 | 25 Ble 12
16A | Ambloplites....... 30% | age | Aaa Ph a0 | oad tdl Ogee Mage 13
| |
14 | Catostomus. ..... 4 52 faa fi30 0h ea) oe eee Ce} 28
15 | Catostomus........) 30 34 | 40 | 26 | 25 | 21 | 54 | 6] 2 44
17 Ameiurus...... He 35 39 33 28 31 22 | 47 | 3 5 | 12
17 A | Ameturus....... 36 7 0 | 93 | 28 | 61 11 3 5 | 12
17B | Ameiurus........ ; 50 7 0 93 22 50 28 | 3 6 12
20 | Notropis......... 35 2 87 11 38 15 Gt Weal 5 18
20 A | Notropis........... 30 0 0 |} 100 | 97 0 orl, =o D 12
30 Etheostoma....... 32 21 1 | 785) Ado od SRB Be) ab 13
30.4 | Etheostoma....... 13 | 27 | 6 | 13 | 90 | 67 | 3 | 5 16
REACTIONS OF FISHES TO ATMOSPHERIC GASES 247
In the closely observed experiments, Abramis, Hybopsis and
Notropis may be said to have reacted with considerable pre-
cision, either by rising to the surface or by spending more of the
time in the high end or by turning back definitely from the
center. In the long experiments, the amount of rising to the
surface was not definitely noted in all cases although this reac-
tion may dominate over all others. Ameiurus and Catostomus
sometimes give this reaction so the preference of these fishes
for the boiled water as shown in table 12 is open to some
question.
2. Low oxygen stock. The low oxygen stock consisted of a
number of Abramis, Ameiurus and Umbra. They were put in
a small glass-sided aquarium and supplied with water from the
boiling apparatus, which water was changed every few days and
which always showed an oxygen content of less than 1 cc. per
liter. The carbon dioxide at times was as high as 8 cc. per Liter.
Tables 13 and 14 give the results of experiments upon this stock
TABLE 13
Showing the reactions of fishes from the low oxygen stock to a boiled water gradient.
The data are arranged as in the preceding tables; compare with table 11. Controls in
tap; gradient between tap (T) and boiled (B); corresponding ends of the control
designated by the same letters. Tap water contained 8 cc. per liter. Boiled less
than 1 cc. per liter.
| cRoss- 4 TURNED BACK
ED CEN- Plait ee FROM TAP (T)
| TER x oy Z OR BOILED (B)
EXPT. | TIME TEMP. IN
SPECIES No.| ———_| __ — - -
he 3 EEN Expt. |Control) Expt. Control DEG. C.
n 16 ga fad alee Sl id es Re ct |
A|o | ¥
24 | PADTAMVISl. sas ca aren | 3 | 330) 438 40 49.5/50.5) 55 | 45 | 4 9 1 17
24 Al SOURIS oie sciecats atone. |p 1O 106 87 20 44 156 | 79] 21) 1 6/ 0; 0 17
252 Wa@imibmae f2 och cin cece Mest 45 | 40 jot 46 | 1 3 } l 17
25 A fetimbras-ereenee eee 3 68) 106 40 59/41 52} 48] 6 (ol Ce) al | 17
22 Abrainvise: 3=.csce oe: 3 | 128) 63) 10 (67 33 | 35|65| 6] 10] 6 | 6 19
23 Abramise 5s. aoce esc 3 | 425 369) 30 41 |59 | 50/50/19! 12) 4 6 | 12
7 UWimibrass cet ee 3 23 29) 40 61 (39 | 34| 66) 3 5 LG 7
25 X | Ameiurus...... 93} 231 3, 40 [64 |36 | 40/60) 3] 1| 0] o| 18
| |- =
Pere ee oe cot or nen 143 155) 12 65 | 45/49/51) 5] 2| 2] 3
1 Fish same as 24, left in tank one hour and forty minutes, and read again.
2 Control reading incomplete.
3 Three fishes in control.
248 VICTOR E. SHELFORD AND W. C. ALLEE
in low oxygen. Abramis was apparently affected by the con-
tinued exposure to low oxygen. In the high oxygen stock, there
was a definite apparent time preference for the tap water while
here in three of the four trials the fish gave an apparent time
preference for the boiled water. In one experiment the turnings
were more numerous from the tap than from the boiled water.
When all the turnings are considered as of the same value as the
time preference, positive reaction is markedly less than that
shown by the high oxygen stock. The same thing is shown in
the long experiments (table 14). This difference in reaction may
TABLE 14
Showing the reactions of fishes from the low oxygen stock, to a boiled water gradient.
Data arranged as in the preceding tables. Compare with table 12, p. 246. These
fishes were confined in the tap water and for 30 minutes or more. Controls in tap
water.
! j
| CONTROL
ieee DHEMENE T CORRESPONDS TAP
EXPT. er wo! _T2MP Tap, T; BOILED, B |B CORRESPONDS BOILED scumaraaee
NO ‘| pDEG.C. | | READINGS
[op Gra- B No. T Cor. B | No.
dient read Grad. |read.
|
OS (cAbramige...cosctl ott 21 26 27 47 14 35 26 39 14 5
28 A | ALDTAMIS: 2) cecal at 19-21 19 2 79 14 43 | 14 43 14 5
Ihe | Umbra: ss. 6. Fe 3 19 28 46 26 13 59 18 23 13 5
26.A| Umbra.......... fe 19 41 | 19°)-40) 13: °) Yea) Oe |) ap aae 5
27. +| Ameiurus........| 3 16 | 20 49 31 13 4 36 60 13 5
27A Ameiurus........| 3 16 13 49 38 13 8 69 23 13 5
5 Ameiurus........| 5 6 39 4 57 | 20 50 | 14 36 | 20 5
a i | ]
be due to an acclimatization to low oxygen. It is fully as prob-
able, however, that the change was due to the fact that in the
low oxygen aquarium the fishes formed the habit of swimming
much of the time at the surface and thus in the experiment the
gradient would not be noticed. This habit of swimming at the
surface was necessary because the fishes died if confined below
the surface of low oxygen water.
We had only a low oxygen stock of Umbras for boiled water
experiments but their reactions are similar to those given by
other fishes from high oxygen. Low oxygen Ameiurus showed
about the same preference for the tap water in one watched
experiment as did the high oxygen stock. A limited number of
observations in the glass boxes showed the high oxygen stock
OO ——————
REACTIONS OF FISHES TO ATMOSPHERIC GASES 249
to be more active and probably more stimulated by the experi-
mental conditions. In the long experiments Ameiurus showed
a preference for the boiled water end but the high oxygen stock
showed the same in two out of three trials.
b. Boiled water with oxygen added
These experiments were essentially like the preceding ones but
oxygen was added to the boiled water end so that the amount
in solution was equal to that in the tap water. Thus the fish
encountered changes in water as described for the boiled water
gradient experiment with the exception of oxygen. The con-
trols were different from the controls of other experiments, in
that boiled water was introduced into one end of the control
tanks and tap water into the other. The control was then an
experiment and the amount of reactions to factors other than
oxygen was determined by comparing the experiment and con-
trol. The general results are shown in table 15. Lepomis (ex-
periment 80) Hybopsis (experiment 78) Catostomus (experiment
82) and Notropis (experiment 77x) showed a time preference for
the boiled water with oxygen added, but the average time prefer-
ence was for the tap water, these being exceptions. The turn-
ings back in the gradient were not accompanied by any of the
characteristic reactions described for carbon dioxide. As a rule
the fishes turned back from the boiled water with oxygen added,
oftener than from the tap water, but Notropis, experiment 77 A,
furnishes an exception.
In the controls (really experiments with boiled against tap
water) the fish gave practically the same reactions that have
already been discussed in treating of the reactions to the boiled
water gradient (p. 241). In general, the time preference was
much greater for the tap water while the number of turnings
was about the same for each end. The sharper reaction as
shown in the matter of time, shows the effect of oxygen upon
the reactions of fishes.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, NO. 2
250
VICTOR E. SHELFORD AND W. C. ALLEE
TABLE 15
Showing the reactions of fishes to boiled water with oxygen added to balance the oxygen
of the tap. The gradient consisted of a few parts per million of calcium, et cetera,
15 cc. of nitrogen, 2 cc. of free and 1 cc. of half-bound carbon dioxide.
Controls were gradients between tap and boiled
arranged as
in the preceding tables.
Data
water with no oxygen added, so that they are like the experiments tabulated in
table 11, p. 243.
| PER CENT OF TURNED BACK
| O2 IN CC. PER LITER CROSS- | TIME IN HALVES | FROM WATER AS
| ED CEN-| , TAP; INDICATED
= TER B, BOILED Se ES
gen ‘| SPECIES NO. SIZE Expt | Control | TIME WATER Expt. |Control) pra.
| |
ZK | We ee hese teal | B B
fissile l|aleia| |tTltltT Bl tlt|tl|B
led Bs id ta s BSa eos | Oe
i | } | | |
83 X |Micropterous.| 1 9) 9.16 9.91 | 9.16 | 4.16 | 13 41| 20] 85] 15] 65|35| 4] 11] 3] 10] 88.5
80 |Lepomis...... 3| 5| 9.16] 9.91|9.16| 4.16 | 48| 13 | 20 | 461] 54| 64] 36/19/19] 3] O
77 A |Notropis...... 3 | 9.16 | 9.91 | 9.16 | 2.83 | 86 | 52| 20 | 54] 46 | 64/ 36] 9| 4] 14] 30] 9
Flac eee el | 9.16 | 9.91 | 9.16 | 2.83 | 31] 73| 20 | 462] 54 | 48] 52] 1] 10| 22] 30] 9
79 |Notropis......| 3 (9.16 | 9.91 | 9.16 | 2.83 106 | 56) 20 | 55 | 45 | 56 | 44 | 10 | 13 | 12| 18| 9
78 | Hybopsis..... 3 9.16 | 9.91| 9.16 | 2.83|19| 76| 26| 318] 69/84] 16] 2] 2] 2] 7/9
83 |Ambloplites..| 3 4-8 | 9.16 | 9.91 | 9.16 | 2.83 | 40 |109| 20) 60| 40| 63) 37} 1| 9| 1| 9/ 88.5
83 A |Ambloplites..| 3 4-8 | 9.16 | 9.91] 9.16 | 2.83 | 88) 52| 20) 65| 35| 73 | 27} 4]10| 1] 13] 8-8.5
81 |Abramis...... 3 | 9.16 | 9.91 | 9.16 | 2.83 | 33 | 62| 20/59) 41|84/16| 1| 4] O| 6/9
82 ._|Catostomus..) 3 | 9.16 | 9.91 | 9.16| 2.88] 14] 48] 20] 484) 52 | 55) 45] 6] 3] 3]17/8
e {E45
; | } 8
71. ‘| Hybopsis..... | 3 9 9 9 2.4 310 | 44 eal 67 | 33 | 57| 43 | 5| 30] 4} 11
ERE ik iygci os wt Biase a ee ..| 72 | 35 | 56 | 44 | 63 | 37| 5.5 10| 6 | 13 |
| |
1 Fish stayed in boiled water plus oxygen. After finding the tap water they
spent most of the time in that end.
2 Fish did not encounter tap water until the end of the first five minutes. After
the tap water was encountered the fish spent 62 per cent of the time in that end.
’ Fishes driven asymmetrically.
* No evidence of a reaction.
c. Boiled water and carbon dioxide
The results of these experiments are shown in table 16. The
oxygen content of the boiled water was less than 1 ce. per liter
as it flowed into the tank. The carbon dioxide content was kept
as nearly as possible at 50 cc. per liter at the boiled water and
3 cc. per liter at the tap water end. As shown by the table and
by the graphs, chart 4, the reactions of all the fishes were very
decided. In every case they showed a strong time preference
for the tap water and turned back from the treated water much
oftener than from the tap. All the testing (backing-starting)
—— =
REACTIONS OF FISHES TO ATMOSPHERIC GASES 251
TABLE 16
Showing the reactions of fishes to a gradient of boiled water accompanied by high
carbon dioxide. Data arranged as in preceding tables. Controls in tap water.
Gradient between tap and boiled plus carbon dioxide.
| j |
Lees CROSS- rere sf | TURNED BACK FROM
| pen |2D CEN
LITER a Ses Expt. Control Expt. Control
EXPT. =
ear SPECIES No. ; — Cares to Ree
zs | "| L.O2| H.O2| sponding | sponding
re) High) Low
ai 2/3 High| Low |__| CO: | CO: |
Fi m)/ 818 CO: | COz High Low "| High| Low
eq /la/o | COz CO:z
57 Lepomis 3] 3 | 50] 28 | 64 20 6 | 94 66 34 24 al ao 2
57 A |Lepomis........ 3| 3/50/29] 5/ 20 | 17} 83 | 66 | 34 | 2] 9] O| 0
Notropis........ 37 3: DOM eG) poe Dea at 99 50 50 | 80 0} 6 5
61 (Ameiurus 2| 3) 50] 6| 33 20°} ASST 895 33 67 | 10 0) 22 4
62 |Umbra.......... 2) 3) 50/13/15! 20 | 19 4088 | 67 | 337. 9 1 1 6
59 Catostomus 3‘ 3) 50 | 8| 3 20 4 96 48 52 | 25 aT 0
55 Ambloplites 3] 3|50)20|) 5] 20 11 | 89 By; 63 | 16 0 0 0
55 A | Ambloplites 3) -3o.50 [02a | 173) 220 5 Obi, [e238 77 9 0 u 1
58 |Abramis peers 3) | (3° ESO ST) 38 15 920 8 92 24 76 1l 0 0 0
58 A | Abramis........ 31-3] 50)19|37] 20 8 92 41 59 5 0 1 4
56 | Hybopsis EEE 3} 3) 60:}33) 30) 20 3 97 63 37 7 0 5 0
56 A | Hybopsis....... 3 | 3} 50} 20) 32) 20 8 92 68 32 3 0 0 0
Meese. 2. s-e koa | 19 26 7A cee page | Seo). Jeet bes. |. 2
coughing, gasping and gulping reactions were given with greater
frequency and greater intensity than in any of the other experl-
ments.
When observed in the glass boxes, Hybopsis, Lepomis Ameiu-
rus, Umbra and Ambloplites showed great disturbance in the
boiled water with carbon dioxide added. There was uniformly
greatly increased activity, increased opercular movements, and
special reactions, such as gulping, rising to the surface, et cetera,
as described for boiled water alone. These were accompanied
by some lack of coérdination and in one case (Ambloplites) by.-
falling on the side. peal“
We have in the experiments good evidence that fishes turn
back from waters high in carbon dioxide and low in oxygen with
precision and vigor. Also that if they enter such localities, they
cannot behave normally and may soon dié. When we compare
these results with those on boiled water or oxygen and with the
results on carbon dioxide alone we see that carbon dioxide is the
most potent factor yet studied in this series of experiments.
252 VICTOR E. SHELFORD AND W. C. ALLEE
d. Acid and ammonia
The only two species of fishes used reacted to a combination
of acetic acid and boiled water (table 17, chart 3) in a manner
comparable to the response given to boiled water and increased
carbon dioxide. Since the reactions were so clearly parallel, no
further experiments were tried. This acidity was far greater
than when carbon dioxide was used (table 16) and since the
reactions were not markedly different, it is probable that the
fish react to some factor other than acidity in the case of the
carbon dioxide. Thus it would seem that the narcotic action
of carbon dioxide is more important in fish reactions than its
action as an acid.
TABLE 17
Showing the reactions of fishes to boiled water accompanied by acetic acid. Data
arranged as in preceding table. Controlsintap water. Experiments with gradient
between tap and treated water.
| TURNED BACK
4 | PERCENTOF | FROM
| | é as pa CROSSED |'TIME IN HALVES! BOILED (B)
Sees R) CENTER | | acrp (TAP)
EXPT. | | eae een TIME
NO. Senet age IN MIN.| Expt. |Control) Expt. | Control
pare = hea) ee) Ls
| Sule | FE) Ses) & S| Bese
z | es 3 8 | magia O84 mga mls
= — Sl Gael ee ieee
41 Abramis..........:..| 2| 3.36] 12 | 28 | 22 | 30 15 | 85 | 56 | 44) 29) 0 1 0
42 | Amefuriis:, 25: ..sceses Z| 4,00 | AZ 67 | 142 | 40 19 | 81 | 56 | 44 | 26| 5) 21 | 13
} |
eS WES? AS SeeRI NS ae — | == pas || 22
Average, ii5ct SPs aes cee a eee | 48 82 17 | 83 | 56 | 44 28 | 3 1 7
The reactions to ammonia and boiled water are shown in
table 18. Although their reactions to boiled water were marked
Abramis did not react to boiled water and ammonia in a con-
centration which caused them to turn on their sides after an
hour or more. With the higher concentration this was true also
but more markedly since the fishes showed a time preference for
the high ammonia, though free to go back and forth. The three
Abramis, died after from nineteen to twenty-three minutes. The
only definite response given was the testing reaction which
occurred very often in both ends of the tank. <A single Notropis
put in with the Abramis died after thirty minutes.
REACTIONS OF FISHES TO ATMOSPHERIC GASES 2a
TABLE 18
Showing the reactions of fishes to a gradient of boiled water with ammonia added.
Data arranged as in preceding tables. Controls in tap water. Gradient between
tap and treated water.
PER CENT OF TURNED BACK
3 TIME IN HALVES FROM
A co. | CROSSED ————, (NH,OH+
PER LITER| CENTER Expt. Control BOILED) oR TAP
as SPECIES NO —— i
No. on IN MIN. - | Corre- ne | Control
= = spond a fem S71
a = E © EY - -
Fel Be lok) a fH) Sie) Sle) Sie] Se
4 3 5 5 Z\/e8\/Z4/\/8/|4/e |\42)e&
36 ADT aMnIS 2 2 ere ae 3 1 |} 8 29 27 30 42) 58158| 42] 0| 0| 1] 0
39 PMEMIUS <2. -<.2<20s6- 2 1 | 102 44 | 108 30 60 | 40 | 62| 38; 2| 3] 18} 10
EVER AL Ce ot as 52k al hoin (OE win ne 3 37 68 51 49) 60 40 1 2:10) 5
Ameiurus showed some signs of stimulation, such as gulping
but gave no movements which tended to bring the fish into
better conditions. As in Abramis, no avoiding or regulatory
reactions were given. After going back and forth for twenty-
nine minutes, the two fishes came to rest, one in the high ammonia
and low oxygen, the other in the low ammonia and high oxygen.
The former was apparently dead at the end of forty minutes
though it recovered after a week or more in clear water. The
one in the lower ammonia lived and showed no sign of having
been affected.
‘While these two species probably rarely encounter acid media
except carbonic acid, they react to the acid in much the same
manner as to carbon dioxide. This result is in accord with
much experimental work in animal behavior.
Fishes must encounter ammonia in very weak concentrations
quite often in primeval nature, but the species studied in these
preliminary experiments, appear to be unable to react to the
concentration used, at least when it is accompanied by low
oxygen content. Low oxygen accompanies ammonia in sewage
and if the results obtained with these fishes are the rule, the
relation of fishes to ammonia and low oxygen is a life or death
matter.
254 VICTOR E. SHELFORD AND W. C. ALLEE
4. COMPARISON OF REACTIONS
We have discussed the reaction of several species of fish to
various factors and the combination of these factors, without
especially considering the differences of reactions given by fishes
of different species or unlike age, to any one set of factors. In
this section we will discuss comparative aspects.
a. Degree of reaction to the different factors
Table 19 shows the average reactions of the controls. It
indicates that some of the fishes, as Ambloplites and Catostomus,
sometimes spent a greater part of their time in one end of the
control tank when the two ends were alike as far as we could
know. The majority of the controls show almost a balanced
time average, and in four of the ten species studied this is exactly
balanced. The number of turnings is more variable and in some
TABLE 19
Showing the average control responses for each species. The turnings are given in
* percentage of the total number and a rating is given of the degree of asymmetry of
response or the apparent preference for one end. The ratings are obtained by
subtracting the percentages given for the two ends and dividing their sum by two.
For example, in the case of Ameiurus, 49 from 51 gives 2 and 42 from 58 gives 16.
Since the turnings do not agree with the time preferences they mut be considered of
the opposite sign. Adding +16 and +2 gives +14 which, divided by 2 gives the
rating as +7. This is a numerical expression of the time spent and the turnings
from each end when they are considered of equal value.
| | AVE. PERCENT AVE. PER CENT
eel on | AVES a OF TIME OF TURNINGS
: CR |
SPECIES ete HO. “ise 2 Bs E E gs | gs RATING
ANOTUTNIS fs. sy.02 ote be Zin Bie 60 31 51 49 58 | 42 | =7
PDEAINIBS | os tae cas | Leal, gehen alee at oes 51 | 49° | 62 |; Saale
Hybopsis............ 11 | 3 |47 {22 | 50 | 50 | 50 | 50 | +0
Ambloplites........... 8 | 3 |30 |31 | 42 | 58 | 22 | 73 | +90
Catostomus..........| 6 | 3 | 43 | 25 37 | 63 | 43 | 57 | £6
MDE ei.cin.cy isl 1 Be 1a tee 46 | 54 | 50 | 50 | +4
Wotropisc............|' 00'S") aa ae 47 | 53 | 45 | 55 | +8
BIEMOOUG aes. i. os a Ve 3 43 29 42 58 50 | 50 | +6
Etheostoma........... 2 | 3 |58 | 40 51 | 49 | 64 | 36 | #13
Micropterus.......... 2 Lb. 17 SQ S Sila ae 0 0 | =18
REACTIONS OF FISHES TO ATMOSPHERIC GASES 259
cases counterbalances the apparent time preference for one end.
To be consistent, there must be more turnings from the end
in which the least time is spent. When this does not occur, the
apparent time preference is neutralized to a greater or less
extent, as the case may be. The figures in the column marked
‘rating’ in table 19 are attempts to express numerically the inten-
sity of reaction when both turnings and time preferences are
considered. The ratings were obtained by adding the difference
in the percentage of time spent in the two ends, to the difference
in the percentage of total number of turnings. This sum divided
by two gives the rating average. In the experiments where the
fish avoided the tap water or the water nearest like that in which
they had been kept, they were rated as negative. Thus the rating
for Ameiurus from the low oxygen stock in relation to the boiled
water is in favor of the boiled water end of the experiment, so
they are rated as +18.
In the experiments the degree of negative reaction to the vari-
ous factors and combihations of factors is shown in table 20.
The ratings of the boiled water, oxygen, et cetera, group are very
conservative, as the averages upon which they are based include
cases where fishes reacted by coming to the surface, et cetera, and
which neutralized the reaction in experiments where our methods
are effective (p. 246). In some cases few experiments were run
with a given set of conditions because the reactions were very
decided, so that further experiments were unnecessary for the
purpose of this paper. Considering the averages as they stand,
it will be noted that the greatest vigor of reaction is shown to
earbon dioxide in boiled water and to the combination of carbon
dioxide and boiled water with tap water at the other end of the
gradient. The reactions to carbon dioxide in tap water and that
to acetic acid in boiled water are about equal and stand next
in rank. Of the experiments where carbon dioxide and acetic
acid are not concerned, the reaction is most definite to an oxygen
gradient in boiled water. Fishes vary greatly in the vigor of
their reaction to both boiled water and to oxygen alone.
Hybopsis proved most sensitive of all the fishes tried. The
graphs in chart 4 show the manner of reaction and the movements
aaliod aNv 79
aaiog +
700 HO SONILVH AO “GAY
dVL NI %QO
£6
299 40
“09 C JO SISVE VY NO DNILVE
06
8+
dVL‘SA Gai110od + VINOWNY
.
‘SidxXad ON
—
—_
|
=
dVi “SA HSLV A
aaiiod OL MOOLIS*7Q MOT
“SLldXa "ON
68
far)
oO
| UALVYAM dVL “SA
| agONVIVatO BLIA GaATIOg
NANAK AN KH eS
"SLdXad “ON
£46
6b
ve
20 + ULV A
| GUTIOg "SA UTLYA GaATIO"€ |
‘
“SIdXa ‘ON
‘ggz ‘d jo w10440q pur ‘g ayou AT[VIDedso ‘eFz ‘d ooS f {TT O]qBI YAIA p1O99B UT YON +
Mis
IIe
dVi ‘SA HaLVAM GATIORg
8¢ OL
FIL 4 Lag
69
if 68
€ 09
4
6 19
Z ; 2
7 OF
4 68 I
€ 6h I TZ
€ 96 (4 18
sof] 8
by a)
: p+) OF y)
TORE
By 4
oa Ss
wo 5
=| w
be |
8
Ba :
Nn
iF
“SIdX@ “ON
197
dVL ‘SA
UaLYA Gag1M0d + IILAIV
"SIdXa “ON
88
‘SA UALYM GAO + 7D
aw
“SLdX@ ‘ON |
16
est
S8
~
=
daqai0@
‘SA HALVM GTTIOG+ 7QD
“SIdXa ‘ON |
TOuL
-NOO JO GNA GANO 4Od
8+
GONGUSADTHd IWLINGACIOOV |
“* OSRIDAY
fee ee,
Nonreoen mo &
-_
_- Oo
_— =
S1OHULINOO “ON
So eruBqy
** BUI} S094 FAT
“*** BNUIOFSOPBO,
ain Se symodary
“*"saqy[dolqury
“ysnseydosoy Py
“+++ g1dou0N
“oes stsdoqAyy
Salogds
“Buinojjof wunjoo ou 07 soyddy uaarb syuamriadxa fo LaqUunny
“LOIDN PIJVIt] AY) 07 WOYIDAL AY) AWDOLPUL SUBIS ALaYN DIMOWUP 07 SUOL1]IDAL BY) PUD 4ayDM Ppa7L0g 07 Y90}8 Uabhxo mo} ay) fo suoN IDAs
ay) fo saspo ay) Ur jdaoxe suf pauoyuowm sayom fo pury oy} fo avUuvproan aaynjas Afrubis sainbyf yuasaf{ip ay] spLom Layjo UL faavjobau
ap sburyos ay) payfroads asunsayjo alayn ydaoxy “6 21qQD) UL sp paurwjiqo buyns sojanf pun sarvads fiq uoyonas fo Lobia aaynjas Buunoyy
0¢ WTAVL
256
REACTIONS OF FISHES TO ATMOSPHERIC GASES 257
of this fish to the various factors. As shown by the chart,
Hybopsis clearly avoids water which has lost some of its salts,
carbon dioxide and nitrogen (p. 240), and reacts a great deal
more vigorously to water which has also lost most of its oxygen.
However the reaction to an oxygen gradient in boiled water is
less definite than to tap against boiled water. The reactions to
earbon dioxide are all definite, being most so to carbon dioxide
in boiled water.
Table 20 not only gives relative vigor of reaction when read
from right to left but enables one to compare species by reading
from top to bottom, in so far as the various factors have been
worked. Notropis and Abramis stand second after Hybopsis.
The data are far too incomplete and the experiments too few in
number to justify general comparison. ‘The column at the ex-
treme right suggests the reactions of the different species to
combinations of oxygen and carbon dioxide. The amounts of
carbon dioxide used were much higher than the animals com-
monly encounter in nature, so that these figures could not be
used for ecological ratings and comparisons even if there were
enough experiments to justify the attempt. The ratings on the
basis of a gradient of 5 ec. of carbon dioxide per liter given in
the column to the left of the last would come nearer to rating
the fishes according to their distribution in clear, well-aérated
water and stagnant and foul waters. The reactions to oxygen
are clearly greater than the rating of accidental preference for
one end or other of the control, while the reactions to the effect
of boiling, with oxygen added to balance, are variable and the
rating is less than the errors of the controls for Lepomis, Notropis
and Catostomus.
When compared with the habitat preference data of Forbes
and Richardson (’08, pp. 79-85), we are unable to see definite
correlations between the occurrence of fishes with reference to
size of stream or pond; or with reference to current or kind of
bottom. However, it appears significant that Hybopsis, which
proved to be the most sensitive fish in these experiments, is
shown to have the most limited habitat preference of any of the
fishes studied. Since the environment of fishes is a complex of
many factors we cannot expect correlation to be possible until
experimental study has been carried much farther.
258 VICTOR E. SHELFORD AND W. C. ALLEE
b. Adult and juvenile fishes
We have had only a little opportunity to note the differences
in behavior of adult and juvenile fishes. Wiegelt (85) found
that young fishes were more sensitive to ammonia than adults.
As a rule, in our experiments the younger and smaller fishes
were more easily affected by the various stimuli employed than
the adults of the same species. In a general way, the adults and
where used, the young also of the hardier fishes (Abramis, Umbra,
and Ameiurus) react clearly in a negative manner to carbon
dioxide, acid, and with less vigor, to lack of oxygen. In all
probability the adults of the food and game fishes react in a
manner comparable to the young, but with the equipment at
hand, we were unable to obtain recordable results.
Among other observations of this kind, a large Leporhis was
placed in the tank with the smaller individuals in an experiment
with a carbon dioxide gradient in tap water. During the entire
period of experimentation, the young fishes went back and forth,
both turning back from the center and spending a longer time
in the tap water. The large fish came to rest in the carbon
dioxide end, and although clearly affected, as shown by gulping
and rising to the surface, remained in the carbon dioxide for
twenty-six-and-one-half minutes before encountering the tap
water end. During the next seven minutes, it remained most
of the time in the tap water but after that went back and forth
for seven minutes at a very rapid rate and without stopping in
either end. In a thirty-minute control, the same fish did not
cross the center at all. . The greater speed of this large fish prob-
ably carried it to the end of the tank before it could be expected
to turn, after having been affected by entering the carbon dioxide
water. In such a case it could not have been possible to obtain
recordable results. The smaller size of the tanks, in proportion
to the size of the fish probably makes the surroundings much
more unnatural. In .connection with the study of adult fishes,
we conclude that for the adults of the food and game fishes, the
tanks should be about three to five times as long as ours and
probably twice as wide and deep. We must, however, leave the
matter as a special subject for investigation.
REACTIONS OF FISHES TO ATMOSPHERIC GASES 259
VI. GENERAL DISCUSSION
Previous to this series of almost a hundred experiments, we
spent much time in devising means of applying the gradient
method of experimentation to our particular problems. By the
methods used we have ascertained that certain fishes react clearly
to varying amounts of oxygen, carbon dioxide, acids, and to the
general effect of boiling. The possibility of more accurate, more
detailed, and more comprehensive work is evident, and is now
being undertaken in this laboratory. These experiments were
planned only as an introduction to the subject.
We have noted that each species has differences in details of
movement and of resting but we have made no attempt to pre-
sent here a detailed statement of these differences. While recog-
nizing these differences, we have found a clear correspondence in
the general avoiding reactions given under certain conditions
Such reactions are gasping, rising to the surface, increased re-
spiratory. activity, and turning back from the disturbing condition
of the water. These general characters of reaction clearly domi-
nate over the more specific details in the matter of successful
avoidance of otherwise stimulating conditions. This makes spe-
cific peculiarities of minor significance in the success of fishes
and points clearly to physiological characters comparable to_
generic, family, and ordinal characters but which bring together
fish in nowise taxonomically related. These groupings on the
basis of physiological or ecological characters have no relation
to groupings made on the usual bases of taxonomy. The major-
ity of investigators are commonly impressed with the detailed
structural characters of the organism and such peculiarities of
behavior as go with them, quite forgetting the physiological
processes and groupings which are clearly general in the sense
that they belong to whole groups of organisms (Shelford ’12).
While the possibility of groupings is clearly suggested by dif-
fering vigor of reaction on the part of different species, it is not
possible to outline such groups definitely on the basis of our
data, for reasons outlined on page 257. Furthermore such group-
ings may have to be based primarily on the reactions during
the breeding season. Reactions to solutes must be considered in
260 VICTOR E. SHELFORD AND W. C. ALLEE
connection with the breeding period and breeding condition, if we
are to explain the distribution of fishes in nature and make our
groupings ecological. ‘The number of ecological factors is clearly
great. The entire life history of the species must be studied
both by experiment and by observation and first, as we believe,
with particular reference to the physiological characters of the
greater magnitude, leaving the specific aspects until later. For
these reasons we have studied several species, hoping to get a
hint concerning these physiological characters of the higher order.
That is, we were hunting for the generalities of behavior that
would apply to ten species of fishes widely distributed taxo-
nomically, rather than the specific details of the behavior. of
any one species. Had we chosen to work entirely on Abramis,
the abundant and easy species, we could have completed a
detailed bit of work of a type highly approved by investigators
but which would clearly have led us into error, because of the
specific peculiarities of the species. These peculiarities would
have led to incorrect interpretations of the behavior of the indi-
vidual fish and to a much greater error if the data had been used
as a basis for generalizations. Thus we would have accumulated
a mass of details of doubtful application to current problems,
however interesting they might have been of themselves. Had
we studied only darters we should have erred in a still more
dangerous direction.
Turning to the practical application of our conclusions to
current biological problems, we find that they fall under three
main heads. First, the economic and distribution problems; sec-
ond, the problems of fish physiology; and third, the problems of
behavior and psychology. From the economic point of view, it
appears to us from these experiments, that emphasis has been
wrongly placed upon environmental factors as matters of life
and death to the fishes concerned. Clearly, fishes are often
absent from accessible situations which upon inspection appear
favorable, and where an examination of the water shows condi-
tions entirely compatible with life. It appears also that the
importance of oxygen in determining the distribution of fish, has
been too much emphasized. The oxygen optimum of all the
REACTIONS OF FISHES TO ATMOSPHERIC GASES 261
fish studied, as was that of the fish studied by Duncan and
Hoppe-Seyler (’93), is clearly low. Fishes react to oxygen gradi-
ents, though usually indefinitely.
On the other hand, the importance of carbon dioxide in fish
distribution has been largely overlooked. It is significant that
even in tap water, all the fish tried reacted very definitely to
an amount of carbon dioxide that is scarcely greater than that
often found in ponds. Increased carbon dioxide is usually accom-
panied in nature by low oxygen and it is to the combination of
lack of oxygen (boiling) and increased carbon dioxide, that the
fish react most definitely. We accordingly feel justified in stat-
ing that the carbon dioxide content of the water (not excessively
alkaline) is the best single index (Shelford and Allee 712) of the
suitability,.of water for fishes. Half bound carbon dioxide may
be of some importance in alkaline waters but our evidence tends
to show that in neutral or acid water it has little effect. Cer-
tainly in survey work designed to determine the suitability of
water for fishes, the determination of the carbon dioxide content
and the study of the conditions necessary for breeding (Shelford
11) should not be omitted.
From the standpoint of the physiology of fishes we have con-
tributed little but have added some confirming data to the obser-
vation that adult fishes are less sensitive than juvenile ones of
the same species, and that carbon dioxide acts as a narcotic and
in small quantities stimulates the respiratory center (Reuss 710).
The experiments indicate also that the fishes detect differences
in the character of water but the localization of the reception
of such stimuli has not been studied.
From the standpoint of the behavior and psychology of fishes,
we note that fishes are able to react to stimuli by simple turnings
back, and that as a rule, they remain longer in water which does
not clearly influence the details of their activities. That the
formation of associations may enter into the latter type of reac-
tion and perhaps also into the former is suggested by the more
decided avoidance of treated water which comes with repeated
entrances into it. It seems, however, that such results may
possibly be otherwise interpreted. It is possible that time is
262 VICTOR E. SHELFORD AND W. C. ALLEE
required for the fishes to become affected by the lack of oxygen,
carbon dioxide, et cetera, and that when the system has once been
affected, for example, by carbon dioxide, a slight increase may
have a more pronounced effect than at first when the blood
supply is relatively free from this substance.
VII. SUMMARY
1. In the experimental control of gases, several and not one
factor are commonly varied; the varying of single factors is
unusually . difficult, being essentially impossible, when gas is
bubbled through water (p. 210).
2. Fishes are clearly affected by lack of oxygen; species usually
die in the order of their relation to these factors in nature (p. 223).
3. Carbon dioxide, in concentrations probably used to produce
reversals of reaction in some of the invertebrates, is poisonous
to fishes, producing death very quickly (p. 224).
4. Fishes are not seriously affected by high nitrogen except
when gas is in excess under one atmosphere pressure and comes
off in bubbles on rough and warm objects; under these conditions
gas bubble disease occurs (p. 222).
5. Fishes react negatively to a gradient of decreasing salts,
nitrogen, and 1.5 ec. per liter of carbon dioxide, in combination;
to a decrease in oxygen and other effects of boiling in combina-
tion; to carbon dioxide, to carbon dioxide in combination with
the effects of boiling, and to carbon dioxide in boiled water at
both ends of the experimental tank. The precision and definite-
ness of the reaction is indicated by the order in which the factor
and combinations are given the most definite reaction being to
carbon dioxide (p. 229).
6. The negative reaction of the fishes is evident through longer
stays in the tap water end of the gradient tanks, by turnings
back from the center, by risings to the surface, or by any combi-
nation of the three (p. 231).
7. Such reactions are accompanied by backing-starting reac-
tions, coughing, gasping and gulping, directly proportional to the
degree of avoidance of the treated water (p. 231).
REACTIONS OF FISHES TO ATMOSPHERIC GASES 263
8. The two species tried reacted negatively to a gradient of -
acetic acid in boiled water but some of the species did not react
to ammonia in concentrations which produce death (p. 252).
9. Some fish establish ‘preferences’ or apparent preferences for
one end of the experimental tanks for reasons obviously due to
the experimental conditions (pp. 246, 261).
10. There is a large similarity of reaction among fishes that
differ widely taxonomically, which indicates the possibility of
groupings which are of generic, ordinal, or family value, but
which bear no relation to existing taxonomy (p. 259).
11. The carbon dioxide content of the water is probably the
best single index of the Ease of the water for supporting
fishes (p. 261).
The Writers are indebted to Dr. H. N. McCoy and Mr. C. H.
Viol, of the Department of Chemistry, for advice in connection
with the control of gases, and to Dr. W. Crocker for the loan of
apparatus used in some of the work. Also to Dr. 8. E. Meek,
of the Field Museum, for the identification of the fishes, and to
Mariner and Hoskins for chemical analyses without charge.
July 29, 1912.
BIBLIOGRAPHY
Aspott, C. C. 1870 Notes on fresh water fishes of New Jersey. Am. Nat.,
vol. 4, pp. 99-117.
1875 Notes on some fishes of the Delaware River. B. Notes on the
winter habits. Report U.S. Fish Commission; 1875-76, pp. 841-845.
ALLEB, W. C. 1912. An experimental analysis of the relation between physio-
logical states and rheotaxis in Isopoda. Jour. Exp. Zodl., vol. 18, pp.
269-344.
AMERICAN HeauttTH AssocIATION 1905 Report of the committee on standard
methods of water analysis. Jour. Inf. Dis., May,’Supl., pp. 1-141.
Birce, E. A. 1904 The thermocline and its biological significance. Trans’
Am. Micro. Soc., vol. 25, September, pp. 5-33.
1907 a Oxygen dissolved in the waters of Wisconsin lakes. Report
Wis. Com. of Fisheries, pp. 118-139.
1907 b The respiration of an inland lake. Trans. Am. Fisheries,
Soc., pp. 223-241.
264 VICTOR E. SHELFORD AND W. C. ALLEE
_Brrce, E. A. 1910 Gases dissolved in the waters of Wisconsin lakes. Bull. Bur.
of Fisheries, vol. 28, pp. 1278-1294.
Brree, E. A., AnD Jupay, C. 1911 The inland lakes of Wisconsin. The dis-
solved gases of the water and their biological significance. Wis. Geol.
and Nat. Hist. Survey. Bull. no. 22., Sc. Series, vol. 7.
Crocker, W., AND Knigut, L. I. 1912 Effect of gaseous impurities on etiolated
seedlings of sweet pea. Bot. Gaz., vol. 53.
Duncan, C., AnD Hoppe-Seyter, F. 1893 Beitrage zur Kenntnis der Respira-
tion der Fische. Zeitschr. Physiol. Chem., vol. 17, pp. 165-181.
Epincer, L. 1901 Have fishes memory? Smithsonian Rept. for 1899, pp. 375-
394.
Forses, 8. A., AND RicHarpson, R. E. 1908 The fishes of Illinois. Nat. Hist.
Surv. of Ill., Ichthyology, vol. 3.
Gitt, T. 1904 A remarkable genus of fishes, the Umbras. Smithsonian Misc.
Coll., April, pp. 295-305.
Hankinson, T. L. 1907 Walnut Lake, Michigan. Biological Survey of Michi-
gan. Lansing, State Board of Geological Survey, Rep. for 1907, pp.
157-288.
1910 a Ecological notes on the fishes of Walnut Lake, Michigan,
Trans. Am. Fish Soc., 1910, pp. 195-206.
1910b An ecological study of the fish of a small stream. Tr. Ill. Ac.
of Science, vol. 3, pp. 1-8.
Hempezt, W. 1911 Methods of gas analysis. (Translation by L. M. Dennis)
New York.
Homes, S. J. 1911 The evolution of animal intelligence. Especially pp. 219-
230 and citations. New York.
Jupay, C. 1908 Some aquatic invertebrates that live under anaérobic condi-
tions. Wis. Ac. of Arts, Sci. and Lett., vol. 16, pp. 10-16.
Jupay, C., AND WaGNER, GEO. 1908 Dissolved oxygen as a factor in the dis-
tribution of fishes. Wis. Ac. Se. Arts and Lett., vol. 16., pt. 1. pp. 17-22.
JOHNSTONE, JAMES 1908 Conditions of life in the sea. Cambridge Univ. Press.
Kenpatu, W. C. 1902 The habits of the commercial cat-fishes. Bull. U. S.
Fish Com., pp. 399-415.
KnautHe, K. 1907 Das Siiszwasser, Neudamm.
1908 Der Kreislauf der Gase in unseren Gewissern. Biol. Centr.,
Bd. 18, pp. 785-805.
Koro, C. A. 1903 The plankton of the Illinois River. Part 1., Quantitative
investigations and general results. Bull. Ill. St. Lab. of N. H., vol. 2,
Article 2.
K6niG, J. 1899 Die Verunreinigung der Gewisser. Berlin. Band 2.
REACTIONS OF FISHES TO ATMOSPHERIC GASES 265
Lake Micu1GAn WATER Commission 1911 Second Report. Urbana, IIl.
LavauDEN, L. 1905 Recherches sur la physiologie du poisson chat. (Ameiurus
nebulosis) C. R. Soe. Biol., Paris, pp. 256-258.
Lors, J. 1904 The control of heliotropic reactions in fresh water crustaceans
by chemicals, especially COz. Univ. of Calif. Publ. Physiol., vol. 2,
pp..1-3.
1905 Physiological effects of lack of oxygen. General Physiology,
Chicago Decennial Publications, pp. 370-425.
Lyon, E. P. 1904 On rheotropism. I. Rheotropism in fishes. Am. Jour. of
Phys., vol. 12, pp. 149-161.
Mast, S.O. 1910 Light and the behavior of organisms. New York.
Marsu, M. C. 1907 The effects of some industrial wastes upon fishes. U. S.
Geol. Surv. Water Supply and Irrigation Paper no. 192 (The Potomac
River Basin), pp. 337-349.
1910 Notes on the dissolved content of waters in its effects on fishes.
Bull. Bur. Fish., 1908 (Internat. Fisheries Congress), pp. 893-906.
Marsgu, M. C., anp Gornam, E. P. 1905 The gas disease in fishes. Rept. of
Bureau of Fisheries, 1904. Washington, pp. 345-376.
McCoy, H. M., ann Test, C. D. 1911 Equilibrium between sodium carbonate,
sodium bicarbonate and water. Jour. Am. Chem. Soc., vol. 33, pp.
473-476.
McCoy, H. M., anp Smitu, H. J. 1911 Equilibrium between alkali-earth car-
bonates, carbon dioxide and water. Jour. Am. Chem. Soc., vol. 33,
pp. 468-473.
MEEK, S. E., AND HitpEBRAND, S. F. 1910 Synoptic list of the fishes known to
occur within fifty miles of Chicago. Field Museum of Natural History,
Chicago, Publ. 142. Zodl. Ser., vol. 7, no. 9.
Nicuots, W. R. 1894 Water supply. New York.
Parker, G. H. 1903 Hearing and allied senses in fishes. Bull. U. S. Fish
Com. for 1902, pp. 45-64.
1910 Effect of explosive sounds such as those produced by motor boats
and guns, upon fishes. U.S. Bureau of Fisheries for 1910. Document
no. 752.
Pizron, Henri 1908 Del’ influence de l’oxygen dissus sur le comportment des
invertebrés marins. C. R. Soc. Biol., Paris, (pp. 886-888; 955-957;
1020-1022; 1061-1063; 1161-1162).
Reuss, H. 1910 Die Wirkung der Kohlensdure auf Atmung der neideren Wirbel-
tiere im besonderen der Fische. Zeit. fiir Biol., vol 53, pp. 555-587.
Scuremenz, P. 1901 Weit. Stud. i. d. Abw. d. Zuckerfabr. Ztschr. f. Fisch.,
vol. 10. pp. 147-185.
THE JOURNAL OF EXPERIMENTAL ZOGLOGY, VOL. 14, No. 2
266 VICTOR E. SHELFORD AND W. C. ALLEE
ScuiEMEeNz P. (witH CronHEIM, W.) 1901 Die Schadigung der Fischerei i. d.
Obra (Zuckerfabrik; Stirkefabrik). Deutsch f. Fischerei. H. 1 and 2.
ScHttmann. 1902 Die Verunreinig. d. 6ff. Gewasser z. Berlin. Deutsch Viertel-
jahrsschrf. off. Gesundheitspfl, H. 2, pp. 226-250.
SHELForD, V.E. 1911 a Ecological succession. I. Stream fishes and the method
of physiographic analysis. Biol. Bull., vol. 21, pp. 9-35.
1911 b Ecological succession. IJ. Pond fishes, L. ¢., pp. 121-149.
1911 c Physiological animal geography. Jour. Morph., vol. 22, pp.
552-618.
1911 d Ecological succession. III. A reconnaissance of its causes
in ponds with particular reference to fish. Biol. Bull., vol. 21, pp.
127-149.
1912 Ecological Succession. V, Aspects of physiological classifica-
tion, L. c., vol. 23, pp. 331-70.
SHELFORD, V. E., AND ALLEE, W.C. 1912 An index of fish environment. Science,
N.S., vol. 36, no. 916, p. 76.
Suutz, C. A. 1911 The oxygen minimum and the germination of Xanthium
seeds. Bot. Gaz., vol. 52, pp. 453-477.
Sumner, F. B. 1906 Physiological effects upon fishes of changes in the density
and salinity of water. Bull of Bureau Fish., vol. 25, pp. 53-108.
WeiceE.t, C. 1885 Schidigung der Fischerei. Arch. f. Hygiene, vol. 3, p. 82.
1903 L’assainessement et le replenement des rivieres Brux.
WopsEDALEK, J. E. 1911 Phototactic reactions and their reversal in the may-fly
nymphs. Biol. Bull., vol. 21, pp. 265-271.
Zortuout, W. D. 1899 On some analogies between the physiological effects of
high temperature, lack of oxygen and certain poisons. Am. Jour.
Physiol., vol. 2, pp. 220-242. ;
EXPERIMENTS CONCERNING THE SEXUAL DIFFER-
ENCE IN THE WING LENGTH OF DROSOPHILA
AMPELOPHILA
FRANK E. LUTZ
The American Museum of Natural History, New York
TWO FIGURES
Two of the forms of the fruit fly, Drosophila ampelophila,
which have been isolated recently by Prof. T. H. Morgan are
distinguished from the normal by lesser wing length. One,
called ‘wingless’ in Professor Morgan’s papers, really possesses
vestiges of wings which appear to consist largely of modified basal
portions of normal wings. The other, called ‘miniature,’ pos-
sesses all the veins of the normal wing in approximately normal
condition but the wing is only about two-thirds the normal length.
‘Winglessness’ is recessive to normality according to the simple
Mendelian formulae. In F, all individuals, both male and female,
are hybrid no matter which parent bears the abnormal character.
The ‘miniature’ wings are also recessive but are sex limited in
their inheritance. If the mother have miniature wings and the
father be normal, only the females of F,; will be hybrids while
all the males will be pure recessives. In the reciprocal cross the
females will again be hybrids but the males will be pure domi-
nants. ‘The reasonable explanation which Professor Morgan has
advanced of these phenomena is that the factor for miniature
Wings (using such an expression in lieu of a better) is contained
in, or in some way connected with, that chromosome of which
the female possesses two and the male but one, while the factor for
winglessness is connected with something which is shared equally
by both sexes. This idea is shown diagrammatically in figure 1
in which the composition of pure stock of the three forms and
that of two of the cross are shown, the X-chromosomes being
represented by squares.
267
268 FRANK E. LUTZ
Pure Females
Pure Males
. Normal Wingless Miniature
Normale x
Wingless é
Normal e x
Miniature #
Fig. 1 Theoretical composition of zygotes. The squares represent the X-
chromosome. The small circles represent the rest of chromosomes. Where they
are plain their composition is supposed to be normal.
SEXUAL DIFFERENCE IN WING LENGTH 269
TABLE 1
Average dimensions. See text.
| NORMAL 2 NORMAL 9 NORMAL 2
| NORMAL WINGLESS MINIATURE
Win Me 71.404+0.180 | 65.644+0.210 | 70.932+0.139
ae es 64.025+0.137 | 58.568+0.151 | 64.737+0.149
Perna Jie 73.059+0.162 | 69.695+0.236 | 71.045+0.157
See |) # | 69.194+0.170 | 68.053+0.160 | 69.651+0.174
Wine : Femur. d| 2% | 29-199=0.128 | 983.851+0.156 | 98.400+0.145
aa wey 91.783+0.116 | 85.990+0.175 | 92.767+0.133
| 9-c) 7.416+0.173 7.861+0. 234 5.642+0.197
It will be seen from table 1 that the mean wing length of homo-
zygous normal females is considerably greater than that of similar
males—their brothers. The average length of the middle femur
is also greater in the females than in the males but the sexual
dimorphism with respect to wing length is relative as well as
absolute, as is shown by the fact that the ratio of wing to femur
among the females greatly exceeds that among the males, the
difference between the averages being nearly fifty times the error
of the difference. Even so, this would not prove the existence of
a fundamental sexual dimorphism if there were a tendency for
generally large flies to have the ratio large. Without entering
the maze of spurious correlation caused by using indices, we can
see from the regression lines (fig. 2) that in both sexes, but espe-
cially in the male, there is a tendency for the wing to get propor-
tionately smaller as the general size of the insect, as measured by
the size of the middle femora, increases. The dimorphism is
therefore real. The sexes are built on different plans.
It is in all ways probable that there is a large complex of fac-
tors concerned in the development of a normal wing. It is pos-
sible that the abnormal forms considered here are caused by the
dropping out of certain of the factors from this complex. It
would seem that the normal females get a double dose and the
males but a single dose of those factors of the normal complex
which are connected with the X-chromosomes. It is well known
that many factors do not cause as great a somatic development
when in a simplex condition (for example, heterozygous) as these
270 FRANK E. LUTZ
Win g
| [st « 2 & Bh eR Ee ee
Se lp Sy ee eee
-—---™“N-----r---so- --- s - - As ee roe re ore ee ee
Femur
1
! -- - -=— --
|
|
|
|
+
|
i)
|
\
\
!
1
|
(
!
|
1
sof = Mis
1
'
\
i
I!
\
1
\
i]
|
‘
'
4
'
i)
'
1
\
4
!
Middle
ey ee ee
See een
Fig. 2. Regression lines for homozygous normal flies.
same factors cause when in a duplex or homozygous condition.
May it not be, then, that the greater wing development in normal
females is caused by their getting a double dose of certain germinal
elements, one dose for each X-chromosome, while the males get
but a single dose? The following experiments were tried in the
hope of getting some answer to this question.
All the flies in the three sets of experiments were reared at the
same time, given an abundance of food from the same jar of fer-
menting banana, kept in the same sort of bottles which were placed
side by side on the table. Furthermore, the flies for each experi-
ment were the offspring of several score of freely interbreeding
parents and were reared in a number of different bottles so
that the chance variations of ancestral bias among the parents and
SEXUAL DIFFERENCE IN WING LENGTH Tal
of environmental effects from bottle to bottle would tend to be
equalized in the three sets of offspring. This care was taken
because of the marked influence environmental conditions, at
least, have upon general body size.
_ Delcourt and Guyenot' have suggested that in my note on
the failure of disuse to decrease wing length, I should have given
the ratio of wing length to body length. It did not, and still does
not, seem to me to be necessary since all the factors involved,
except disuse of wings, tended to decrease the general body size
as much as or more than that of the wing. In this paper, however,
it does seem desirable to have some other character with which to
compare the wing. Among all that are feasible the body length
is the worst because it may change from hour to hour in the living
insect, can be measured only with great difficulty and changes
greatly after death. I have used the length of the middle femora
because it has none of these disadvantages. ‘The wings and legs
used here were removed from freshly etherized flies and immedi-
ately mounted in balsam. Measurements were made and are
recorded in units of ss mm. for the wings and too mm. for the
femora.
Taking up first the offspring of normal females x wingless
males, we find that, while the wing is approximately of normal
length in both sexes, both it and the femora are significantly
smaller than in the homozygous normal material. This may
possibly be an environmental effect which was not entirely avoided
by the cultural methods used. However, the ratio of the wing
to femur is also significantly smaller in both sexes. It is smaller
even though, as was pointed out above, there is a tendency for
small flies to have the ratio larger than in the case of large flies.
1 Bull. Scient. France et Belgique, 7th Serie, tom. 45, no. 4. Ina general denun-
ciation of all the work hitherto done with this insect, they deplore the fact that
the results have been obtained without the extreme refinements of bacteriological
and physiological methods which they recommend. Their criticisms, insofar
as they have any value, can be applied only to the study of fluctuating variants
such as the characters considered here. All attempts to get heritable abnormal
venation or such forms as wingless and miniature by purposely using extreme
‘environmental conditions have failed. It is, therefore, absurd to lay stress in
such cases upon the slight variations of environment from bottle to bottle.
PA We FRANK E. LUTZ
The heterozygous wings are to all appearances normal but they
are really not relatively as long as normal.
As is indicated in figure 1, both sexes of these flies have only
one dose of that part of the normal wing complex which presum-
ably dropped out to give a wingless fly, whereas both sexes of
normal flies have two doses. In wingless flies, and hence in half —
of each zygote from which these hybrids came, something has
been changed, at least, so that the tendency to develop the wings
to normal proportions has been lost. It is probable that the
result of this experiment, a phenomenon usually referred to as
‘incomplete dominance,’ is due to this cause.
_ However that may be, the interesting point for the present
discussion is that the sexual dimorphism has not been changed.
Environmental effects are practically ruled out here since the
males and females of a given experiment grew up together. Nor-
mal females have a wing-to-femur ratio 7.42 greater than the
males and in these flies it is 7.86. The difference of 0.44 is less
than twice the probable error and certainly is not significant.
On the hypothesis stated above, this is what is to be expected
since the germinal changes are alike in the two sexes.
Conditions are theoretically quite different in the cross between
normal females and miniature-winged males. It will be seen
from figure 1 that normal females get two doses of that part of
the wing complex which is connected with the sex chromosome
whereas the males get but one. In order that miniature wings
may appear, this part of the complex must be changed, either by
the dropping out of a factor or in some other way. In the cross
just mentioned the male offspring are, according to theory,
perfectly normal in their germinal make up. Their single X-
chromosome has it full share, and no more, of wing factors. The
female, however, has only one normal X-chromosome. The other
either lacks a factor or, less probably, has a new modifying fac-
tor. At any rate the second X-chromosome is not equipped for
full wing development. Hence the sexes are more nearly alike
in their germinal make up than are the normal. The results
show that they are also more nearly alike in their somatic condi-
tion.
SEXUAL DIFFERENCE IN WING LENGTH PAS)
The difference between these males and the normal males with
respect to the wing-to-femur ratio must be considered as an
environmental effect since they are supposed to be germinally
the same. It is 1.01 + 0.15 and is doubtless significant in spite
of the precautions taken. However, as was pointed out above, we
escape even this difficulty when comparing brothers and sisters
since there is no evidence and it is not believable that a given
environmental condition will operate to increase the relative size
of the brothers’ wings and decrease that of the sisters’ as is the
case in this cross. The sexual difference is only 5.64, that is,
1.78 = 0.25 less than normal. The sexual difference in these
flies is still nearly thirty times its probable error but in the normal
lot it is about fifty.
An explanation of this remaining and still considerable sexual
difference is not difficult of framing on the hypothesis here fol-
lowed. In fact, a wiping out of all sexual difference would have
proved too much. It is only recently that any sex-limited char-
acters have been known. It is more than likely that there are
many factors concerned in wing development. Certainly all
the factors have not been isolated since wings have not been en-
tirely done away with even in the so-called wingless strain. We
are free, then, to postulate that it is these remaining factors which
cause the remaining sexual dimorphism of wing length.
Therefore it seems that, while proof is lacking, indications
have been found that the greater wing development in the normal
females of this fly than in the males is due to the females getting
two sets of those wing factors which are connected with the X-
chromosome while the males get but one, the double set causing
a greater somatic effect than the single.
aA
ss
“3
;
.
>
‘
noe
at,
—
*
t
patent 2a %
7
slaw Fe +s
2 \ 7aS
= 72"
; rh ye ba! st fe i
Ae iy ¢ 7 is, ‘+e Lee,
‘J
i 7s
‘
‘ ei
a
Te
7 4
ia
es 8 ive
7 las
+ -yue® Toby Eg POR “y
7 }
— , :
q aad
i . ;
£ «
- * ‘ Be
*
; ei
’
} ld
er
<a
® %,
’
S
wy
»
~
=
"
.
THE SEX OF A PARTHENOGENETIC TADPOLE AND
FROG
JACQUES LOEB AND F. W. BANCROFT
The Rockefeller Institute for Medical Research, New York
THREE FIGURES
Bataillon has shown that the unfertilized egg of the frog can
‘be caused to develop by puncturing it. Last spring we tried the
experiment in a large number of eggs of various species of anura.
The females were separated from the males, carefully washed
with water and with alcohol, and then opened. The eggs were
taken out of the uterus with sterilized instruments without com-
ing in contact with the surface of the frog. About 20 per cent of
the unfertilized eggs were kept as controls and 80 per cent were
punctured. A few eggs were fertilized with sperm. Not a
single unfertilized control egg segmented or developed. The
number of unfertilized eggs which began to segment after punc-
ture was greater in the wood frog than in the leopard frog, and
amounted in the most favorable cases to about 40 per cent in the
former. Only 2 of about 10,000 punctured eggs of the wood
frog reached the tadpole stage, but these died before they were
able to swim. The percentage of eggs of the leopard frog which
reached the tadpole stage was greater. From 700 punctured
eggs of the southern leopard frog, 13 good morulae were isolated
the next day. On the third day, when the fertilized controls
were in the gastrula stage, 13 unfertilized punctured eggs were
also in the gastrula stage and 4 more eggs were developing abnor-
mally. On the fourth day, 8 of the parthenogenetic eggs had good
medullary folds and 4 had irregular folds. On the sixth day,
most of the fertilized eggs hatched and 8 of the parthenogenetic
eggs hatched also. Of these latter, 4 were developing regularly
and 4 irregularly. Those that had not hatched were abnormal.
275
276 JACQUES LOEB AND F. W. BANCROFT
On the eighth day, the larvae arising from the fertilized eggs
were swimming. Among the larvae arising from the unfertilized
punctured eggs only 3 were normal, and their development was
slightly retarded, perhaps one day. In addition, 6 partheno-
genetic larvae were abnormal but still alive.
On the thirteenth day, 2 of the parthenogenetic larvae were
feeding and these were the only ones which survived definitely.
The other parthenogenetic larvae all died during the next few
days. Of the 2 surviving larvae, one went through metamor-
phosis after five months. When it died, the tail was almost com-
pletely absorbed (fig. 1). Its death was probably accidental.
The other lived a month longer and formed small hind legs, but
died in the tadpole stage (fig. 2).
The sex glands of the frog were taken out, hardened in Tellyes-
nicki’s fluid and sectioned; those of the tadpole were removed
after it had been preserved in formalin for several months.
It was found that both parthenogenetic tadpole ang frog were
females (fig. 3).
This result should be expected if the frog gE = to that
group of animals in which the female is heterozygous for sex.
Part of these experiments were made in the laboratory of the
University of North Carolina, and we take pleasure in thanking
Prof. H. V. Wilson for the many courtesies shown to us.
SEX OF A PARTHENOGENETIC TADPOLE 2th
Figs. 1 and 2. Parthenogenetic frog and parthenogenetic tadpole (natural size)
Fig. 3 Section of the ovary of the parthenogenetic frog (magnification 253)
: ’
«
és f
a
¢
’ ~
oy 7
» ia
3 e. 4
ah
- .
7
~
“*
ar
“é
es
. a
a > i
oo
*
'.
ae
Ty
THE EFFECT OF CONJUGATION IN PARAMECIUM}!
H. S. JENNINGS
Zoélogical Laboratory of the Johns Hopkins University
TWO FIGURES
CONTENTS
11, JERE a oie eee eee sate ook tect oe RMI ahah 1 rate amare Se 280
Il. Methods: ‘Split pairs’; ‘pairs’; designation; culture; records; tabu-
Intion and tabless.0 53.502: ast 2 A. Fy CeO ae a ee 282
Ill. Effect of conjugation on multiplication, survival and variation........ 285
Experiments with wild cultures: Paramecium caudatum............ 286
Experiment 1: rate of fission; variation; mortality; abnormalities.. 286
Experiment 2: rate of fission; variation, mortality; abnormalities;
SIMMATN UU OR MES UGS ear act ye bsg: iis eatin ois) Noes oc ema eee 293
Experiment 3: effect of high temperatures.............../........ 300
Experiments with: pure-strame. 4.0056... 8b eek stv ales. thd oo 302
ENC Oa It) Sear ee tk ay tT Pea eR cia Soke. Renae tree eek oh 302
RixpenimentAssLaramecium aurelia-... eo. 28. 4. okie ee na eee ee 304
Experiments 5 to 14; comparative effects of repeated conjugation
and of long abstention from conjugation...................... 305
Diagram of history of these experiments......................5. 307
IEE CTUMETMIRO RR ees re Nein aie ad cc eae ISIS eo SMI see Oem 308
EIRPICTAMEN biOee. ose toe ew SFr a iskoytortrce, ts Ae RAE Reap eee 309
SSeS MINION Lian Se eae oe Oe os Teas ak ce Boe ae PE Bas eee i Ge 314
LBeS TCO IET GT: Sp) ei hf a ie a nee ts OME Treen, LOSE e: 314
Experiment 9: fission rate; dimensions; determination of conjuga-
UTC 2 cs PR ae el eid ome Maier he SLT tN LS oS as 315
fieamectimeiml Ue 4 ornate se PAT Th ae eee tee Mere) eee 317
Paperiment tL les Mibs te. 1 ds thes tt fe ihe 7 ee Shy S18
Experiment 12: Pominpanion in a doped stock.. Bre Bote cos BEC AL
Experiment 13: Production of inherited iinet eman i conju-
RUIOMS oye orc AA EWE Se I, MO Re ae aT, 329
Hight selffertilimagizous? (- 40s to ae ES ie 330
Experiment 13 a: rate of fission; variation; mortality........... 332
Experiment 13 6: inherited differentiations in the pure strain.... 333
Experiment 13 c: conjugant lines; non-conjugant lines........... 340
Sumintiry’ of ‘experiment a oe ee 342
1 Fourth of a series of papers on Heredity, Variation and Evolution in Protozoa.
279
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
APRIL, 1913
280 H. Ss. JENNINGS
Hixpertment J4....... 5.225... 3. Steed 1. a 343
Experiment 15: Paramecium caudatum; inherited differentiation
produced by conjugation... :...... °..325. 2%. gaeess. 5 eee 343
Culture methods necessary to secure uniformity of conditions.... 345
Records; explanation of tables 34 and 35; results................ 346
IV. Résumé of results: discussion and conclusions......................... 355
Effect of conjugation on rate of reproduction....................... 356
Effect. of conjugation on mortality. ..22.22--. =) - 3c. .+...:% 28 29a 359
Effect of conjugation on abnormalities.......................--...-- 361
Effect of conjugation on Variation. 22.1) 02o5. 5. ee eee 361
Conjugation and biparental inheritance. ...<..2......:....02.5..2) eason
Conjugation and the theory of rejuvenescence...................... 367
General conclusion.<).<:), a6 nse oe aoe Oe ee 376
Literature cited)..c.2<0255... 2202 aS eee ee Sn ee ee 378
Appendix: fundamental tables (tables 29 to 35).................9:....:..----- Og
I. INTRODUCTION
What is the effect of conjugation on the individual or stock
that undergoes it? This question reduces experimentally to the
following: In what respect does a stock that has conjugated
differ from one that is in other respects similar, but has not con-
jugated? What difference is produced if half a given stock is
allowed to conjugate while the other half is not?
When one examines the evidence for the conclusions commonly
drawn as to the effects of conjugation on the stock in the in-
fusoria, it is curious that almost no direct experimental evidence
is found. The conclusions as to the rejuvenating or other physio-
logical effects of conjugation are based almost exclusively on
reasoning of the following character: ‘‘Since without conjuga-
tion such and such processes of degeneration (or other phenomena)
occur, it must be that conjugation has the effect of preventing
or curing this degeneration (or these other phenomena). The
experimentation is almost all devoted to testing the premise,
while direct experimental demonstration that the conclusion is
correct, that conjugation actually does rejuvenate (or the like),
is almost unattempted. Such exceptions to this generalization
as exist we shall later take up in detail.
There is then need of an investigation in which conjugation
itself—rather than what happens without conjugation—shall le
at the center of experimentation. Such a study this paper pre-
EFFECT OF CONJUGATION 281
sents. The fundamental experiment is to divide a given stock
into two parts, kept under identical conditions, permitting one
part to conjugate and preventing the other; to keep these further
under identical conditions, and to determine in what respects
they differ. In the few attempts that have heretofore been made
to observe directly the results of conjugation, the control series
(of the same stock without conjugation) has almost invariably
been omitted, so that it is uncertain how far the phenomena
observed would have occurred equally if there had been no
conjugation.”
In the investigation here set forth this fundamental experiment
has been many times repeated, with careful study of the various
characteristics of the set that have conjugated, as compared with
those of the set that have not conjugated. Besides an account
of the results of this fundamental experiment, the paper deals
with certain other problems connected with the physiology of
conjugation. The effect of conjugation on the size of the indi-
viduals of the stock has been set forth in a former paper (Jennings
eEL).
Thus the matters dealt with in the present paper are mainly
the following: the effects of conjugation on the rate of multi-
plication; on survival and mortality; on the general vigor; its
relation to ‘rejuvenescence;’ the effects of conjugation among
close relatives; the effects of continued inbreeding; the results of
allowing a stock to conjugate many times in a given period, as
compared with causing it to multiply without conjugation for
the same period; the relation of conjugation to inheritance, and
the effect of conjugation on variation.
Each experiment gives evidence on most of the matters just
mentioned, so that it is not possible to separate fully the dif-
ferent subjects. The experimental results will first be presented
systematically, with more particular reference to the effects of
conjugation on vigor, multiplication, survival and variation, then
each of the topics will be taken up, and an analysis given of the
experimental evidence bearing upon it.
2 The only exception to this that I have found is in the experiment of R. Hert-
wig, briefly set forth in his paper of 1889; this will be taken up later.
282 H. S. JENNINGS
Il. METHODS
Most of the experiments followed a general plan, the chief
features of which may be here set forth. The object was, to
compare, under similar conditions, a set of the animals that had
conjugated with another set that was ready to conjugate, but
was prevented from doing so. Both Paramecium caudatum and
Paramecium aurelia were employed in the work. Abundant con-
jugation was obtained in the way described by Maupas (’89),
Calkins and Cull (’07), and others. In the evening large numbers
of the animals were taken. from the large cultures and placed
in watch glasses; early the following morning they were usually
beginning conjugation. Paramecium caudatum is_ especially
favorable for obtaining with certainty the first stages of the
process, since as Maupas (’89, pp. 171, 182) has noted, this
animal conjugates in the early morning, commencing at about
five o’clock. If therefore there were no conjugations when the
watch glasses were set, one can be certain that any pairs found
early the following morning have just united.
Split pairs. At the beginning the pairing animals fit loosely
together; they at first, as a rule, adhere together only by their
anterior ends. At such a time it is easy to separate them, by
drawing them repeatedly into a fine pipette. The separated
individuals are then isolated and cultivated separately.
Pairs. Other pairs are allowed to complete conjugation.
They separate spontaneously after about twelve hours; the two
members are then isolated and cultivated separately, under the
same conditions as the members of the ‘split pairs.’
In this way two sets are obtained, taken from the same cul-
ture, both ready to conjugate and beginning the process at the
same time; the only difference between them lies in the fact that
one is allowed to complete the process, while the other is not.
By cultivating the two sets under identical conditions it becomes
possible to determine what difference is made by conjugation.
Designation. The terms ‘pairs’ and ‘split pairs’ will be used
in referring to the members of the two sets, and to their progeny.
The two members of any pair, or of any split pair, will be desig-
EFFECT OF CONJUGATION 283
nated a and b. It is important to understand that these desig-
nations do not imply any characteristic differences, the two
letters being assigned arbitrarily and at random to the two mem-
bers, in order to make it possible to speak of them separately.
But the individual and its progeny to which a given letter is
assigned of course retain this designation throughout. The pairs
(and split pairs) of any experiment are designated by serial
numbers, so that any individual is indicated by a number and
a letter; thus 8b signifies the individual b of pair 8. The lines
of progeny from a given individual receive the same designation
as the parent individual, so that in later stages of the experiment
8b signifies the line of progeny derived from the individual }
of the pair 8.
Culture. The isolated individuals were transferred to the
concavities of hollow ground glass slides, each concavity con-
taining two or three drops of culture fluid. Thick sldes with
two concavities were found most convenient. At the beginning
the animals were usually left for one or two days in water from
the dish in which they were found, in order not to disturb the
processes of conjugation by the shock of removal to a different
fluid. For the later cultivation an infusion of pure Timothy hay
was usually employed. This was made by boiling one gram of
Timothy hay for ten minutes in 100 cc. of tap water, then adding
to this infusion, after it was cool, 100 cc. of filtered but unboiled
water. Sometimes this filtered water was taken from the parent
culture; a procedure that, in some cases, though not in all, works
well. The infusion was tried in varying strengths at different
times, but all the animals of a given experiment were treated
throughout in exactly the same way. The infusion was invariably
made up fresh just before it was used.
In the last experiments tried, it was found that ;; per cent
Horlick’s malted milk, as recommended by Peebles (’12), was
preferable in some respects to the hay infusion, particularly in
summer. It was found necessary, for the best results, to make
up this culture fluid fresh each day.
The animals were transferred to two drops of fresh infusion,
on a clean slide; in some experiments every day, in others every
284 H. S. JENNINGS
other day. In some experiments two individuals of each line
were transferred to the new slide, in others only one; the remainder
being destroyed.
The slides were kept in moist chambers, on strips of glass
which were supported above water covering the bottom of the
vessel.
Records. At each transfer the number of fissions undergone
since the last transfer was recorded; so that these records were
made either every day or every other day. Since either one or
two individuals had been left on the slide at the previous trans-
fer, there was no difficulty in determining how many fissions had
occurred. In some cases of course the same number of fissions
had not occurred in the two individuals left the day before, but
this made absolutely no difficulty in practice. If, for example,
ten individuals were found on the slide, invariably four of these
were distinctly smaller than the other six. This showed that
each of the two original individuals had divided twice, producing
8, and that two of these 8 individuals had divided again, giving
the four small specimens out of the ten. If therefore two of
these small individuals were transferred to the new slide (as would
usually be done in such a case), the number of fissions was re-
corded as three.
For keeping records of large numbers of cultures (the numbers
ran up to 480 in some cases), the following procedure is con-
venient. Procure hollow ground slides of which the upper sur-
face has been ground, so that one can write on them with a lead
pencil. Then at transfer write upon the new slide (besides the
designation of the line) the number of specimens found in the
old slide, the number left in the new one, and an indication of
the generation to which they belong. Thus in the case just
cited, where 10 individuals were present and two of the smaller
ones were transferred, the legend on the new slide would be simply
‘10 (16)-2’; which indicates that 10 were present; that 2 were
retained, and that if the fission from which these two resulted
were complete, there would have been 16 on the slide. After
all the slides have been thus transferred, labeled, and placedin
EFFECT OF CONJUGATION 285
the moist chambers, the latter are examined and the records on
the slides copied to permanent records on large sheets of paper.
From these records the exact number of fissions can be obtained
at any time; thus, since in the above case the generation contain-
ing 16 had been obtained from 2, it is clear that three fissions
had occurred.
Tabulation and tables. For analysis, the records of fissions
have in most cases been tabulated for definite periods, as of one
week or of ten days. In analyzing the records it has been nec-
essary to make from the original tables of records a very large
number of secondary tables, particularly correlation tables. In
place of publishing these secondary tables, the original tables of
record will be published in the Appendix of the present paper.
These contain all that would be found in the correlation and
other secondary tables, and anyone who desires can reconstruct
the latter from them, so that they furnish every possibility for
testing the results here given. Furthermore, the original record
tables show much that is lost when they are transformed into
correlation tables; particularly do they show much that is of
interest from the point of view of ‘pure line’ studies.
Besides these tables giving the original records, the present
paper will contain as a rule only tables giving the data—the
constants, et cetera—resulting from the analysis, by biometrical
methods, of these records; these are found in the body of the paper.
Ill. EFFECT OF CONJUGATION ON MULTIPLICATION, SURVIVAL AND
VARIATION
As previously noted, the experiments described below furnish
evidence on other matters besides those set forth in the heading
above, but we shall examine them first from this point of view.
We shall divide the experiments into two sets; the first including
those dealing with ‘wild’ cultures; the second those dealing with
pure strains.
286 H. S. JENNINGS
EXPERIMENTS WITH WILD CULTURES: PARAMECIUM CAUDATUM
Experiment 1: May 4 to June 7, 1909
As giving typical results, an experiment which was in progress
from May 4 to June 7, 1909, will first be presented.
The animals were taken from a ‘wild” culture of Paramecium
caudatum, which was brought from a pool on May 3, 1909. It
was found to swarm with the infusoria, and on the evening of
May 3 numbers of them were placed in watch glasses; at that
time none were conjugating. Early the next morning conjuga-
tion was beginning. Thirty-five pairs in the first stages of union
were separated in the way already described, but in 11 of these
one of the members was killed or lost, so that there remained 59
individuals that had gone through the first stages of pairing;
among these 59, both members were present in 24 of the ‘split
pairs.’
Thirty-one pairs were allowed to complete conjugation, then
the two members isolated. One member of one pair was lost,
so that from the pairs there were derived 61 lines of propagation,
as against 59 from the split pairs.
In this experiment the 120 lines of propagation were changed,
and the records taken, every other day. The numbers of fissions
were grouped in weekly periods for each line.
Thus we have before us 120 individuals undergoing propaga-
tion; one set of 61 have just conjugated, while another set of
59 were ready to conjugate, but were prevented from doing so.
They differ in no other way but in regard to conjugation. What
later difference does this make in the two sets?
1. The first thing that we discover is that the individuals which
were ready to conjugate but were prevented, are by no means
in a depressed, degenerate condition, unable to propagate farther.
On the contrary, they continue to propagate in an active, healthy
manner. They continued to do this till the experiment was
discontinued five weeks later.
2. Secondly, we notice that those which have conjugated multi-
ply less rapidly than those which have not. This difference is
EFFECT OF CONJUGATION 287
very great, and will be well brought out by examining side by
side certain weekly records for the first fifteen individuals of
each set (table 1).
TABLE 1
Experiment 1. Paramecium caudatum. Number of fissions per week for the first 15
lines of each set (d = dead).
First week. [eis |
LI Ag oa ea }O/1/5)5/4/2)1)5}0/5/6)/3|2/4])0
Srpspeins a8 iG GH FSH AS sae ieee eh ae eT Mee eS TUES: WA reeds
Second week: le A Te ae
oir. a eee 1/2)6/5|/6/6/0)5/2)4/5|/5)1/2]6
SCIOTO Us eee a 21°63) G7 16.) See | G6) 6 4)5/6/5/ 6/6
pilin weeks: 2.08) eon ae | | Pe en
OES eee ey ae oe ey tr ee GAGs RS. |aOel soa OR GOR ON RGHeGMleo lon lnaalis
STG Crise eae 7 6/9|7/6/6|8]6/5)9| 10 9)8)|7
3. All those which have not conjugated multiply, while among
those that have conjugated are a considerable number that either
never divide again, although they may live for a long time; or
divide but few times. This will be evident from examination of
the general record table (table 29, Appendix).
4. A considerable number of the lines derived from those that
have conjugated die out, while none of the others die out (table
29).
5. It is evident on a cursory examination of the records that
among the lines derived from the conjugants there is much
greater variation in the rate of fission than among those derived
from the individuals that have not conjugated.
Each of these points will now be taken up in detail and the
facts precisely brought out.
The weekly records for the entire experiment are given in
table 29 (Appendix), which serves as a basis for the following
discussion.
Fifth week not typical. One point should be brought out at
the beginning of the analysis. During the fifth week the lines
of propagation were in an unhealthy condition owing to extrane-
ous reasons. On May 31, at the end of the fourth week, the
experiment was tried of mixing a little starch from boiled bread
288 H. S. JENNINGS
with the culture fluid, in the hope that this would improve the
latter. It had the reverse effect, making the animals unhealthy,
and almost or quite stopping multiplication. As a result, the
figures for the fifth week are very low and irregular for both
sets. It would beyond doubt give a more correct idea of the
real relations if we should exclude the fifth week entirely, and
consider the experiment as ending with the fourth week. But
I have not felt justified in suppressing any part of the record,
and the main results are clear in spite of the irregularity due to
the exceptional conditions of the fifth week. But it will be well
to keep in mind the fact that the results obtained from the fifth
week have little or no significance on our main problems.
Rate of fission. Table 2 gives the mean number of fissions for
each of the two sets, for each week and for various combinations
of weeks; also the ratio of the means for those that have not
conjugated (split pairs) to the means for those that have con-
jugated.
TABLE 2
Experiment 1. Mean numbers of fissions per week and for certain other periods,
in those that have conjugated (pairs) as compared with the same for those that
have not conjugated (split pairs); also ratio of the means for the two sets.
MEMBERS OF PAIRS MEMBERS OF SPLIT RATIO OF MEAN
PAIRS FOR SPLIT PAIRS
= = ee 7 TO MEAN FOR
No. Mean No. Mean PAIRS
BATSb WEEK oe eer eae ae 61 3.279 = 0.148) 59° 6.729 += 0.099 2.052
Second week.................| 56, 4.661 = 0.205 59, 5.932 = 0.093 1.273
Third week.................. 45) 5.000 + 0.264 59, 6.678 = 0.199! 1.336
Pourth weele:.2-4.%. 0. #,. dae 42} 3.976 = 0.210) 59, 5.102 = 0.130 1.288
Fifth week...................| 38) 2.787 + 0.156] 59| 2.593 = 0.113) 0.947
First two weeks......... _..., 56, 7.589 + 0.363) 59 12.661 + 0.144. 1.668
Second two weeks............| 42; 9.306 = 0.415, 59 11.780 + 0.287 1.266
Four weeks:
(a) Those that lived
Grough ee | 42, 18.857 + 0.526] 59 24.441 + 0.352; 1.296
(b) All, including those '
that did not live to |
ey ee eS .....| 61 13.918 = 0.775| 59 24.441 = 0.352) 1.756
Five weeks: |
(a) Those that lived | fost
Mershupl.s'ss hc Shan | 38) 21.842 + 0.602) 59, 27.034 = 0.384 1
Pe ed Pre shi ea, 61 15.902 + 0.833) 59) 27.034 + 0.3841.
EFFECT OF CONJUGATION 289
Table 2 shows that in every week (save the fifth), and in every
combination of weeks, the average number of fissions was greater
for those that had not conjugated than for those that had conju-
gated. The fifth week, as we have already seen, gives, for
extrinsic reasons, atypical results. In that particular case the
difference in the means is not significant, as is shown by the
probable errors in the two cases. For the entire four (or five)
weeks, the average number of fissions was about 25 per cent
greater in those that have not conjugated. If we take the total
number of fissions for each line that was alive at the beginning
of the experiment, we find that the average number of fissions
was 70 to 75 per cent greater for those that had not conjugated.
This is of course partly due to the fact that none of the latter
died before the end of the experiment, while a considerable
number of the conjugant lines died out early.
It is of interest to compare the number of progeny produced
by the two sets. This is of course obtainable from the number
of fissions. The potential progeny produced by the two sets in
each of the five weeks is given in table 3.
As the table shows, each line of those that have not conjugated
produced weekly on the average almost exactly two-and-a-half
times as many progeny as a line of the conjugants.
The 61 lines derived from the conjugants had a potential pro-
duction all together during four weeks of the experiment of
TABLE 3
Experiment 1. Potential number of progeny from those that have conjugated, as
compared with those that have not, based on the number of fissions in table 2.
PAIRS SPLIT PAIRS
WEEK ] Te = =
| Number of Number of Average per Number of | Number of Average per
| ines progeny aes | Eines a progeny | 4 up
£ %) 61 991 16.246 | 59 | Rogge) Y |) MIB5L729
2 56 3038 54.250 59 4296 72.813
3 45 4712 | 104.711 | 59 17066 | 289.254
4 | 42 1489 | 35.452 | 59 2962 | 50.203
5 38 479 12.605 | 59 537 | 9.102
Average per week..........| 44.653 | 111.420
290 H. S. JENNINGS
1 billion, 256 million progeny, while the 59 lines derived from those
that had not conjugated had a production of 48 billion, 467 mil-
lion, so that the non-conjugants produced somewhat more than
38 times as many progeny as the conjugants. ‘The very great dif-
ference between the two in this respect arises from the fact that
many of the conjugant lines died out before the end of the experi-
ment and the further fact that the number of progeny increases in
geometrical ratio as the number of fissions increases in arithmeti-
eal ratio. To this latter fact is due also the seemingly excessive
differences in the number of progeny produced in the different
weeks, as shown in table 3. Unfavorable temperature or culture
medium, decreasing the number of fissions by a small number,
decreases the progeny enormously. |
To sum up on this point, the experiment shows clearly that
those that have not conjugated multiply more rapidly than those
that have conjugated, and the difference persists for at least four
weeks after conjugation.
Variation. A careful examination of the data given in table
29 will show that there is more variation (among the different
lines) for the number of fissions in any given period, for those
that have conjugated than for those that have not. To deter-
mine accurately the differences in this respect, it is necessary to
determine the standard deviations and coefficients of variation for
each period. These are given in table 4 together with a compari-
son showing what the ratio of the variation among the non-con-
jugants is to that among the conjugants.
Table 4 shows that in every week, and in every combination
of weeks, without exception, the variation is greater in those
that have conjugated than in those that have not. It is greater
in the conjugants, whether measured absolutely, by the standard
deviation; or relatively to the mean, by the coefficient of varia-
tion. In many of the periods the coefficient of variation is for
the non-conjugants but one-third to one-fourth of that for the
conjugants.
It is then a simple statement of fact to say that in this case
conjugation increased greatly the variability in the fission rate.
Examination of table 29 shows that this great increase of varia-
EFFECT OF CONJUGATION 291
tion in the progeny of the conjugants is due mainly to the fact
that many of the lines descended from them multiply but slowly,
while others multiply at nearly the same rate as do the progeny
of non-conjugants. Among the 59 lines of non-conjugants, there
are but two that gave fewer than 20 fissions in the five weeks,
while among the 38 lines of conjugants that lived through the
entire five weeks there are 12 that fall below 20. On the other
TABLE 4
Experiment 1. Relative variability in number of fissions for given periods, in those
that have conjugated (pairs), and those that have not (split pairs).
|
| RATIO OF
PAIRS SPLIT PAIRS | SPLIT PAIR
TO PAIR
ee = E g
No,| Standard | Coefficient of [No Standard Coefficient of | E @|ss
deviation variation ‘| deviation variation | 5 z ee a
|} pt jo os
fea Stee ee
First week...... 61 | 1.709 = 0.104 | 52.131 + 3.955) 59 | 1.132 = 0.070 | 16.829 = 1.074 | 0. 662| 0.323
Second week...| 56 | 2.270 + 0.145 | 48.704 + 3.769 | 59 | 1.055 + 0.066 | 17.792 + 1.139 | 0.465 0.365
Third week....| 45 | 2.625 + 0.189 | 52.494 + 4.648 | 59 | 2.266 + 0.141 | 33.929 + 2.337 | 0.863) 0.646
Fourth week...| 42 | 2.018 + 0.149 | 50.743 + 4.506 | 59 | 1.481 + 0.092 | 29.027 = 1.948 | 0.734) 0.572
Fifth week..... 38 | 1.427 + 0.110 | 52.135 = 5.011| 59 | 1.290 = 0.080 | 49.759 + 3.778 | 0.904) 0.954
First two weeks| 56 | 4.030 + 0.257 | 53.103 + 4.232| 59 | 1.643 + 0.102 | 12.975 + 0.819 0.408) 0.244
Secondtwo |
Weeksiics S.455.5 42 | 3.991 = 0.294 | 42.870 + 3.689| 59 | 3.268 + 0.203 | 27.743 + 1.850 | 0.819) 0.647
Four weeks:
(a) Those that |
lived |
through 42 | 5.055 = 0.372 | 26.806 + 2.110| 59 | 4.010 + 0.249 | 16.405 = 1.046 | 0.793) 0.612
(oy PAN ae 61 | 8.901 + 0.544 | 63.951 = 5.430/ 59 | 4.010 + 0.249 | 16.405 + 1.046 | 0.451) 0.257
Five weeks:
(a) Lived |
through 38 | 5.499 = 0.425 | 25.174 = 2.068) 59 | 4.376 = 0.272 | 16.188 + 1.031 | 0.796| 0.643
(199) VNU ee 61 | 9.571 = 0.585 | 60.186 + 4.829 | 59 | 4.376 = 0.272 | 16.188 = 1.031 | 0.457] 0.269
Experiment 1.
Paramecium caudatum.
TABLE 5
Number of lines that died out during
different periods, among those descended from the pairs (conjugation consummated).
Per cent of those alive at be-
ginning of week............|
Per cent of all...
|
ees
[ah Yo
| 0 | 5
lol 8.2]
| O SE2e
ee | FOUR FIVE
Leer | a. ———| WEEKS WEEKS
3 | AE a ei
) 11 | a a yi 23
| 19.6 |. 6.7] 9.5]
1800.\0 469 | G6.) Bl is, | 37-7
{
292 H. S. JENNINGS
hand, the upper extreme for the non-conjugants (38) is higher
than that for the conjugants (32), but most of the non-conjugants
are so grouped near the high figure that the variation is relatively
small.
Mortality. None of the 59 lines of non-conjugants died out
during the five weeks of the experiment. Of the 61 lines of
conjugants, on the other hand, 23, or 37.7 per cent, died out
during the experiment. The number of lines descended from
conjugants that died out during each week is given in table 5.
Thus in this case conjugation greatly increased the mortality.
Although the ‘split pairs’ were ready to conjugate, and had
actually taken the first steps in the process, they are not in the
least injured by being prevented from consummating the process;
while those that finished mating showed a high mortality.
Abnormalities. Besides the actual deaths, the descendants of
those that had conjugated showed many abnormalities, while
among the descendants of the non-conjpugants there were none.
For example, on May 14, I noted that there were among the
descendants of the conjugants 24 abnormal individuals, belonging
to 12 different lines, while in the other set there were none.
The abnormalities take the most diverse forms: bodies of
irregular shape, crooked, truncate, or with projections; double or
multiple monsters: some are abnormally large, others extremely
thin. The structural abnormalities are in many cases connected
with abnormalities in fission. In some cases the ex-conjugants
do not divide for many days after separation. During this time
they grow larger till they reach an immense size, many times
greater than that ever reached at other times. Some of these
immensely large individuals never divide again, and after living
a week or two die. Others after a time divide irregularly, pro-
ducing progeny of diverse sizes and forms. Thus the individual
10a, of the pairs, in this experiment did not divide until eight
days after the separation from its mate. It then divided during
the night into seven, of four diverse sizes. The individual 176
divided immediately into two specimens, which became im-
mensely large; these did not divide again for six days, then each
produced two large abnormal individuals which soon died.
EFFECT OF CONJUGATION 293
Abnormal individuals appear again and again in certain of the
lines derived from conjugants, while in others they do not appear
at all. The conditions which induce them are thus evidently
inherited from generation to generation in the fissions. As a
rule, a given abnormality is not inherited in its special form, but
only the tendency to produce abnormalities of various sorts.
Lines which show abnormalities in structure commonly have a
slow rate of fission, are thin, succumb easily to unfavorable con-
ditions, and in general, appear to lack vitality. Often they die
out after a number of generations.
There are likewise found lines which show the thinness, slow
fission rate, and general lack of vitality, without structural
abnormalities.
It appears probable that these abnormalities have a cytologi-
cal basis, and are due to irregularities in the nuclear processes
accompanying conjugation. A precise study is greatly needed,
as to the minute characteristics of these abnormalities, their
heritability, their experimental cause, and their cytological basis.
Such a study I hope will soon be made.
The data obtained from this experiment, and presented in
table 29, will be analyzed in later papers with reference to the
problems of sexuality, and of uniparental and biparental inheri-
tance.
We may summarize the results of this experiment, so far as
they bear on the problems now under consideration, as follows:
Conjugation decreases the rate of fission, causes a great increase
in variation in the fission rate, brings about many abnormalities,
and greatly increases the death rate.
Experiment 2: April 7 to June 7, 1909
This extensive and long continued experiment was the first
one undertaken for comparing the fission-rate and vitality of
animals that had conjugated and animals that had not. Owing
to lack of experience the method of culture was not good, so
that the mortality was very high; this makes the results less
sharp and clear than in the experiment just described. The chief
mistakes in culture were: (1) the culture fluid was not made up
294 H. S. JENNINGS
by measure, so that it varied in strength from day to day. All
the specimens were treated alike at each change, so that no dif-
ference between the sets resulted from this; but the changes in
concentration of the infusion caused many deaths. (2) The hay
of which the infusion was made was not sorted over, to exclude
all but Timothy; thus at times injurious plants were included,
increasing the mortality.. (3) As a rule, only one individual of
each line was retained at each change, so that if this individual
died, the line became extinct. As a consequence of all these
things, the number of lines decreased rapidly in all the different
sets.
However, such an experiment, lasting eight weeks, is not likely
to be often repeated, and the results are of much value in certain
relations.
Three sets of the animals were employed in this experiment.
One set (‘pairs’) consisted of individuals that had just conjugated;
a second (‘split pairs’) included individuals that had begun to
unite for conjugation, but were separated, in the manner pre-
viously described; the third set (‘free’) consisted of ordinary
individuals that had not begun union, taken from the same cul-
ture as the others. This third set is known not to have conju-
gated recently, since they were taken (like the others) in the
early morning, from watch glasses which contained no conju-
gants when set the evening before. Comparison of the ‘split
pairs’ and the ‘free’ will show whether entrance upon the condi-
tion preparatory to conjugation alters the animals in any way,
such as to affect their multiplication and vigor.
The ‘free’ specimens were given paired designations, each two
being called a and b, as in the ‘pairs’ and ‘split pairs’. In this
case of course a and b were related in no way; the paired desig-
nations were given at random, in order to test the question
whether one individual of -a pair of conjugants dies or is weak
more often than occurs as a result of mere chance causes, in
specimens paired at random and merely by designation. This
is a matter that will be dealt with fully in a later paper.
The three sets were treated in exactly the same way, the slides
of pairs, split pairs and free alternating in the same moist cham-
EFFECT OF CONJUGATION 295
bers. The culture fluid was changed as a rule every other
day.
Since usually members of pairs of conjugants do not divide till
the second day after conjugation, the comparison of the rate
of fission for the three sets was not begun till this second day.
Thus, the animals were isolated on the morning of April 8, but
the tabulation of the fissions begins, for all sets, on April 10.
For purposes of comparison, the fissions were tabulated by
weeks for each of the three sets. The experiment may best be
divided into two periods, the first comprising the first two weeks;
the second the last six weeks. At the end of the second week a
considerable number of each set were lost by accident, so that
the number to be dealt with is much smaller in the second period.
The experiment included at the beginning 57 lines (284 pairs)
of those that had finished conjugation; 39 lines (193 pairs) of
split pairs, and 58 lines of ‘free’ individuals.
The actual number of fissions per week is given for the first
two weeks in table 30; for the last six weeks in table 31 (Appendix).
It should be noted that the data given are, so far as numbers
of fissions go, of little value after the sixth week, and particularly
is this the case for the seventh week. Pressure of other duties
forced me to neglect these experiments at that time, so that
during the seventh week the slides were changed but once; as a
result they hardly multiplied at all. The figures for the seventh
and eighth weeks are given only in order not to suppress any
part of the record.
It is evident, as in the previous experiment, that the animals
which were ready for conjugation were by no means in a de-
pressed or degenerate condition. The split pairs continue to
multiply, somewhat more rapidly than those that have conju-
gated. We shall examine in detail the rate of fission, the varia-
tion, and the mortality, in the three sets.
Rate of fission. Table 6 gives the mean numbers of fissions
in each set, for each week, and for certain other periods.
As this table shows, in practically all of the 15 means given,
the rate of fission is less for those that have conjugated than for
those that have not. The only exception is in the seventh
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
296 H. Ss. JENNINGS
week, where the rate is nearly the same, with a slight excess in
favor of the conjugants. But as we have already noted, the
animals were changed and (records made) but once that week,
and in consequence there was almost no multiplication; the
figures for that week are of no significance. In all the other
cases (14 out of 15) the non-conjugants show a greater rate of
fission, the excess varying from 6 to 85 per cent, with an average
of 23 to 31 per cent.
The split pairs and the free individuals show no significant
difference so far as rate of fission is concerned; so that the speci-
mens that have taken the first steps in conjugation do not differ
in this respect from those that have not.
TABLE 6
Experiment 2. Mean number of fissions for each week, and for certain other periods,
in the three sets, together with a comparison of all that have conjugated with all
that have not.
COMPARISON OF
ALL NON-
| PAIRS | SPLITPAIRS FREE | CONJUGANTS pee Siete:
| Sent -|-FESE CONJUGANTS
A 1 2 3 eet 4 5 6
WEEK | 2 | | 3 S | 2 |8e
| a [| a i Le | S8a
| = |= = |= | Besa) Ras
| ‘§ | Mean number | 6% | Mean numper | 4 | Mean number | ‘§ | Mean number § $28) 9 FI
| = | Offissions | x of fissions |§ x of fissions % | offissions ‘$5 Zs = = ps
| 3 | 1s 3 5 | as~ 5) ea
Ea 'E E E | g25¢ 82s
|B | 3 5 : 5858 585
| | 1 | Z A | 1S} | Px
1 50 | 4.080 = 0.150 | 37 | 7.378 = 0.162 | 54 | 7.685 + 0.1381 | 91] 7.560 + 0.102 | 3.480} 85.3
2 34 | 2.000 + 0.126 | 21 |. 2.810 + 0.174 | 30} 2.233 = 0.126 | 51 | 2.471 + 0.106 | 0.471 | 23.6
3 | 22) 3.591 = 0.236 | 10 | 4.800 = 0.230 | 19 | 4.947 = 0.197 | 29 | 4.897 = 0.160 | 1.306) 36.4
4 19 | 3.211 + 0.335 9| 4.666 + 0.237 | 17 | 4.471 = 0.178 | 26 | 4.538 = 0.241 | 1.327] 41.3
5 14 | 7.143 = 0.609 | 8 | 9.125 = 0.186 | 15 | 9.333 = 0.164 | 23 | 9.261+0.126 2.118 | -29.7
6 11 | 5.182 + 0.272 | 8) 5.125 + 0.512 | 14 / 5.714 = 0.159 | 22 | 5.500 = 0.215} 0.318 6.1
7 11} 1.455 = 0.133 | 8) 1.125 = 0.186 | 14] 1.429 = 0.089 | 22 | 1.318 = 0.091 —0.137 | —9.4
8 9| 4.111 + 0.430 | 6) 5.333 = 0.130 | 11] 4.636 + 0.264 | 17 | 4.882 = 0.185 0.771; 18.8
land 2 | 34| 6.176 + 0.229 | 21 | 10.381 = 0.296 | 30 | 10.100 + 0.281 | 51 | 10.216 + 0.206 | 4.040| 65.4
38and4!19| 7.158 = 0.520 | 26 | 11.692 + 0.253 | 4.534 | 63.3
5 and 6 | 11 | 13.727 + 0.451 22 14.773 = 0.267 | 1.046 7.6
6 weeks | 11 | 29.364 = 0.645 | 8 | 35.000 = 0.860) 14 | 35.143 + 0.486| 22 | 35.091 + 0.440] 5.727 19.5
Bweeks | 9 | 34.111 + 0.962 6 | 42.500 + 0.910| 11 | 41.545 + 0.645 | 17 | 41.882 = 0.532) 7.771 | 22.8
4 2
ots | }
fz d | | ae oe
b OD | |
aos f 9| 4.264 6] 5.313 111] 5.193 117) 5.235 0.971 | 22.8
Soe ||
g55|| | | |
ae | |
So >on | | } /
aye |} 3.846 5.045 5.056 | | 5.053 1.207 | 31.4
$3e fen
aE | /
EFFECT OF CONJUGATION 297
TABLE 7
Experiment 2. Relative variability in fission rate for those that have conjugated
and those that have not.
NON-CONJUGANTS (‘SPLIT PAIRS’)
AND (‘FREE’)
WEEK == == === =) = aaa
CONJUGANTS (‘PAIRS’)
‘No. Standard devi- Coefficient of x, | Standard “dev | Coefficient of
ation variation ation variation
= foe eae ut RNs EE || LE Aa
1 50) 1.573 + 0. 106 38.549 + 2.962| 91) 1.447 += 0.072) 19.134 + 0.991
2 34) 1.085 + 0. 088 54.235 + 5.590) 51, 1.126 + 0.075) 45.592 + 3.623
SG ies 34 1.977 + 0.162| 32.011 + 2.874) 51] 2.181 + 0.146] 21.350 + 1.489
3+4 | 19 3.360 + 0.368) 46.944 + 6.166) 26] 2.671 = 0.250| 22.847 + 2.246
5+ 6 11] 2.219 + 0.319 16.168 + 2.385) 22) 1.857 + 0.189) 12.570 + 1.298
6 weeks..../ 11) 3.170 + 0.456 10.796 + 1.571) 22, 3.059 + 0.311) 8.716 + 0.893
8 weeks....| 9) 4.280 + 0.680 12.548 + 2.026| 17) 3.252 + 0.376 7.764 + 0.904
The split pairs and the free may therefore properly be con-
sidered together, as non-conjugants, as in the fourth column of
‘table 6. With these total results for all the individuals that
have not conjugated may then be compared the results for those
that have conjugated, as in columns 5 and 6 table 6. As there
shown, for the first week the excess of the non-conjugants was
85 per cent, so that their rate was nearly double that of the con-
jugants. After this the excess for the non-conjugants decreased,
although even in the eighth week it is 18.8 per cent. Thus the
conjugants had not regained a rate equal to that of the non-conju-
gants even after so long a period.
Variation. The variation in the rate of fission is shown com-
paratively for conjugants and non-conjugants in table 7. In this
table the two classes of non-conjugants—the ‘split pairs’ and the
‘free’—have been put together, since we have already seen from
table 6 that there is no characteristic difference between them.
This fact is shown equally if we compute the variation separately
for the two classes. Thus, for the first week the split pairs give
a standard deviation of 1.458 and a coefficient of 19.767, while
the corresponding figures for the free are 1.425 and 18.542. For
the second week the figures are: split pairs, 1.180 and 41.998;
free, 1.023 and 45.785; first two weeks, split pairs, 2.011 and
19.375; free, 2.285 and 22.628. Throwing the two together, as
in table 7, gives the great advantage of larger numbers.
298 H. S. JENNINGS
As the table shows, although the means for the conjugants are
throughout less (table 6), their standard deviations are as a rule
greater than the standard deviations for those that have not
conjugated. As a result, the coefficient of variation (standard
deviation divided by the mean) is in every case much greater for
those that have conjugated. For the first week the variation, as
measured by this coefficient, is twice as great in the conjugants.
For the entire eight weeks it is nearly twice as great.
Thus in this experiment, as in the former one, conjugation has
the effect of greatly increasing the variability of the fission rate.
Mortality. Owing to the high general mortality, due to imper-
fect culture methods, the distribution of deaths in this experiment
is of much less significance than in Experiment 1. It is sum-
marized in table 8.
As the table shows, the death rate was greater in those that
had conjugated than in those that had not, in every week save
the second. In the second week I tried the experiment of adding
to the cultures water from a pool that was extremely foul, but
contained many Paramecia. It proved disastrous; many of my
lines were killed, and among these were a larger proportion of the
split pairs and free than of the pairs. I doubt if the distribution
in such a catastrophe is 6f any significance; though possibly it
indicates that those that have conjugated are more resistant
to such decidedly injurious conditions.
Throughout the remainder of the experiment (as throughout
the entire time in Experiment 1), the mortality was highest among
those that had not conjugated. For the entire eight weeks
together the mortality is nearly the same for all three classes,
but is a little greater for those that have conjugated.
There appears to be no significant difference, as to mortality,
between the split pairs and the free. There is thus no indication
that prevention of a conjugation that had been initiated is in any
way injurious.
Abnormalities. In this experiment, as in Experiment 1, I
noted frequent abnormalities among the progeny of the conju-
gants; none among the other sets. No detailed study was made
of these.
EFFECT OF CONJUGATION 299
TABLE 81
Experiment 2. Death rate in pairs, split pairs and free individuals.
aie S) |
Be Be a g Ba
Zz | (=) | & Z a | &
First week: | First two weeks
LS aa yeep meg eee Na Ey) (yl ed peer ely bial sees
Split pairs 39 2 Syea ly | tSyol ae oy Wiss\- oe eae af) Sh) 18 | 46.2
Rineer ts oS. 2: BS) 4 GO ebrees Se. Faso.) SB ah 28) aes
Second week | Last six weeks.....
ais oo. 50 | 14 | 28.0] Pairs 28 49" (16759
Split pairs... ...| 37 16 | 48.2 | Spltippairssee eee? 13 7 NV ease
inet ees ee ed aa dale eerste er tle eon | 12 ateoaee
Third week:...... Last five weeks
LE ee 23 | 6 121.4] Pairs 22 13a*) 59st
Spit pairs:...¢.| 43° |. cS. | 2627 |, Split pairs:........ | 10 4 | 40.0
recto s. 2. saz] 23 3 | 13.0 || rece neeegs es onl LO S421
Fourth week | Total eight weeks
IESE Sa. 24 Sods 22 oe) lowe I MANES oo. <2 Re JS 49 40 | 81.6
Split pairs...... 10 1 nO) Split pairs... 31) 25 «(1280.6
nee atsee 2.22) ) 19 Pet OROM eC Che cae cn tee. se eel AO 39) [le 6
iftheweele# =... =...
IBAUESHES hy was a St 19 5) 2623
Splitpairss.----| -9 Ihe apikilenl
Wreer emetic. ce eh Le De altel |
Sixth week:
[PANDAS SOS eee 14 i ibe t3e [e214 |
Splitupairsts.../ 8 0 0
reese 8... lo 1 Git
Seventh week..... |
[ESSy Hera Cee ee iN ey em)
Splet pairs. 7.6) 8 | 0
reese cn: ec || 44 0 |
Fighth week:.....
eh 11 3. (273
Spit pairs. 2...) ~ 8 \geeee | 25.0 |
Tig s nde Aeeeeiem iy a | at |
|
1 Eight of the pairs, 8 of the split pairs, and 9 of the free, were accidentally
lost during the experiment, so that their disappearance is not accounted for
in this table.
300 H. S. JENNINGS
Summary of results. Experiment 2 gives the following general
results:
In three sets of individuals taken from the same culture and
treated in the same way, one set that was allowed to complete
conjugation, another separated before union was complete, and
a third that had not yet begun conjugation:
1. Those that had completed multiplied throughout less
rapidly than those that did not complete conjugation, and less
rapidly than those that had not begun conjugation. This dif-
ference persisted throughout the experiment; that is, for eight
weeks after conjugation had occurred.
2. The descendants of those that had completed conjugation
were much more variable in their rate of fission than those that
did not conjugate.
3. The mortality was slightly greater among those that had
completed conjugation than among the others.
4. There was no marked difference in these respects between
the set that were separated after beginning conjugation, and the
set that had not yet begun.
Experiment 3: June 20 to June 24, 1909: Effect of high
temperatures
This experiment lasted but four days, and was designed
primarily to test the relative variability in the dimensions ot
the progeny of conjugants and of non-conjugants. The results
on this point have been given in my paper of 1911, on Assortative
mating, et cetera (table 32, p. 99). Here will be given the results
of the experiment so far as they bear upon the comparative vitality
and the rate of reproduction in conjugants and non-conjugants.
In the experiments which we have thus far described, the
conjugants reproduced more slowly than the non-conjugants,
while at the same time the mortality of the conjugants was
higher. It appears possible that under some conditions the
greater rapidity of fission of the non-conjugants might be dis-
advantageous, causing greater mortality in them than in the
conjugants. This possibility is realized in the present experi-
EFFECT OF CONJUGATION 301
TABLE 9
Experiment 3. Paramecium caudatum. Relative number of fissions for the con-
jugants and non-cenjugants: during the four days, June 20 to June 24, 1909. (In-
cluding only those that lived throughout the four days.)
NUMBER OF FISSIONS
TOTAL | MEAN
2} 3] 4 SEES eB a
|
8
Number of conjugant lines. alae Aare atali! | | | 36 | 6.20
|
| |
Number of SA a |
A 1)2)3/5|3)2) 16 10.813
|
ment. The temperature during the four days that the experi-
ment lasted was excessively high, the thermometer standing
much of the time above 90° F. (above 32°C.). The non-conju-
gants multiplied with furious rapidity, at the rate of two to four
fissions a day (one fission in 6 to 12 hours), so that the average
for all that lived through was a little over two and a half per day
(one fission in 9% hours). The conjugants, on the other hand,
multiplied much less rapidly, the rate being but one-and-a-half
per day, or one fission in eighteen hours.
Correlative with this excessively rapid rate of reproduction,
the non-conjugants showed a very high mortality. At the end
of four days, 35 of the original 51 lines were dead, so that the
mortality was 68.6 per cent. In the conjugants, on the other
hand, of the original 47 lines, only 11 died, or but 23.4 per cent.
The data for the rate of fission of the conjugants and non-con-
jugants in this experiment are given in table 9.
The results of this experiment agree with those of all the others
in showing that conjugation decreases the rate of fission. They
differ from those of all others in the fact that the mortality is
much greater in the non-conjugants. The result, due, as it
evidently is, to the excessive rate of fission induced in the non-
conjugants by the very high temperature, shows that under
certain conditions conjugation may have a directly protective
effect, owing to its decreasing the rate of multiplication. Inter-
esting results would be obtained by comparing conjugants and
non-conjugants of the same stock under diverse conditions;
high and low temperatures, different chemical conditions, et
302 H. Ss. JENNINGS
cetera. Possibly it would be found that under all conditions
tending to cause excessive rapidity of fission, conjugation is
protective by decreasing this rate. |
The usual relations are found as to the relative variability of
the conjugants and non-conjugants. In my paper on Assorta-
tive mating (711, p. 99), I have shown that the progeny of the
conjugants are in this*experiment much more variable in size
than the progeny of the non-conjugants, for at least seven genera-
tions. Here we need to consider only the variability in fission
rate.
Of the non-conjugants, as we have seen, but sixteen lived
through the four days. Their mean rate of fission is 10.813 =
0.233, the standard deviation is 1.379 = 0.164, and the coeffi-
cient of variation is 12.756 + 1.546. Of the conjugant lines,
thirty-six lived through; their mean number of fissions was
6.222 + 0.205, the standard deviation 1.827 + 0.145, and the
coefficient of variation 29.369 + 2.528. Thus the variation is
both absolutely and relatively much greater in the conjugants;
if we measure it by the coefficient of variation, the variability in
fission rate was more than twice as great in the progeny of the
conjugants as in that of the non-conjugants.
EXPERIMENTS ON PURE STRAINS: CONJUGANTS ALL DESCENDED
FROM A SINGLE INDIVIDUAL
A large number of experiments, some of them extensive and
long continued, were undertaken with cultures descended from
' a single individual. The conditions in such pure strains (or
‘pure lines’, as I have called them in previous papers), are of
special interest in some respects, while the results bear likewise
upon the same general problems as does the work with wild
cultures.
Race k. These experiments were mostly carried on with the
race k, some account of which has been given in my previous
papers of 1910 and 1911. This race, belonging to the species
Paramecium aurelia, is distinguished by a tendency to conjugate
frequently, making it most favorable material for a study of
EFFECT OF CONJUGATION 303
matters connected with conjugation. It is easy to induce
epidemics of pairing at intervals of about a month, and they some-
times occur at much shorter intervals. In order to make clear
the conditions with which we are dealing, it is necessary to give
a brief account of the history of this race.
As a pure strain, the race k is derived from a single ex-conju-
gant isolated November 9, 1908. This individual itself came
from a culture derived from eight pairs of conjugants taken
February 4, 1908, these eight pairs being themselves derived
from 10 single individuals of similar size, taken from a wild
culture January 29, 1908. Thus even before the destruction
of all but this single individual of November 9, the race k was
derived from few individuals, which very possibly all came from
one. Our first experiment given below (Experiment 4) made use
of this race k& before its absolutely certain derivation from a
single individual; all the rest employed k as known to be a pure
strain.
Epidemics of conjugation were observed in this race eight
times between January 29 and its absolute purification on Novem-
ber 9. Since November 9 a great number of conjugations have
been observed; records have been kept of at least twenty.
That portion of the race still in existence (July, 1912), and the
part on which some of the chief experiments were performed, has
descended from many successive conjugations, in which all the
surviving progeny were derived from a single member of a pair.
Eight such conjugations have been observed, so that all the
animals now existing (and used in Experiments 13 and 14, below)
are derived from eight generations of the strictest inbreeding—
all the members of each of these generations being derived from
the fission of a single individual. This inbred race seems healthy
and vigorous, so long as cultivated in mass culture. But it
appears to have lost the ability, which it had at the beginning, to
propagate for any considerable period on slides. On this account
it has of late become unavailable for comparative work on such
questions as the rate of reproduction and the way this is affected
by conjugation or by other conditions; a fact which has caused
much trouble and loss of time. Many extensive experiments
304 . H. S. JENNINGS
have been undertaken, but after two or three weeks of intense
labor, all representatives of this race k cultivated on the slides
have become unhealthy and died. Earlier in its history it was
kept on slides for months in succession, multiplying vigorously
throughout. Whether its present peculiarity in this respect has
any connection with the long continued inbreeding, or whether
it may be due only to weakening from previous long cultivation
on slides, it is difficult to say; there is some indication, as we shall
see, that the latter is the case.
Experiment 4: October 19 to November 8, 1908: Paramecium
aurelia
The first experiment on conjugation in race k was designed
primarily to permit a comparison of the dimensions of the progeny
of conjugants and non-conjugants in the same race. The results
so far as dimensions are concerned are given in my paper of 1911
on Assortative mating (pp. 96-97). Incidentally, the records
kept give data as to the relative rate of fission, and as to mortality.
As we have noted above, in this experiment (alone of all those
with &), the race is not yet known to be absolutely pure, in the
sense of derived from a single individual, without admixture from
others; it is, however, extremely homogeneous, and probably
quite pure, even at this time.
The experiment was begun with 46 paired individuals (23
pairs), and 46 that were non-conjugants, derived from the same
culture. During the course of the experiments 4 of the conju-
gant and 8 of the non-conjugant lines were accidentally lost,
leaving 42 of the former and 38 of the latter.
This was one of the earliest experiments of the sort that I
tried, and the mortality was very high, doubtless owing to inex-
perience in handling. Of the 42 lines derived from the conju-
gants, but 17 lived throughout the twenty days of the experiment,
while of the non-conjugants 18 lines lived through. The total
number of fissions for each of these 35 surviving lines is given in
table 10.
For the 17 conjugant lines the mean number of fissions is
13.294 + 0.670, with a standard deviation of 4.098 + 0.474 and
EFFECT OF CONJUGATION 305
a coefficient of variation of 30.828 + 3.890. For the 18 non-
conjugant lines the mean is 13.500 + 0.425; the standard devi-
ation 2.672 + 0.300, the coefficient of variation 19.792 + 2.310.
Thus here, as in all other cases, the rate of fission is a little
greater in the non-conjugants, while the variability is much
greater in the progeny of the conjugants than in that of the non-
conjugants.
TABLE 10
Experiment 4. Paramecium aurelia. Number of fissions for the conjugants and
non-conjugants during the twenty days of experiment 4 (October 19 to November
8, 1908).
NUMBER OF FISSIONS
fo) 2108) [4s eee Cle 7 18.) 9) tO ead 2 Sea 15 | 16) 17
Number of con-
jugant lines... 1 1 1 2{/ 4| 3] 2| 3/17) 13.294
Number of non-
conjugant
it Bee sapere 1 1 1 SAS 2) 2)18) 13.500
Where the mortality is so high as in this case, it is doubtless
due mainly to extrinsic causes, so that its distribution is of little
significance. The facts are these: of 42 lines descended from
conjugants, 25 died out during the twenty days of the experiment,
a mortality of 59.52 per cent. Of the 38 lines descended from
non-conjugants, 20 died out, a mortality of 52.63 per cent. Thus
the mortality is, as usual, greatest among the descendents of
the conjugants.
Experiments 5 to 14: Comparative effects of repeated conjugation,
and of long abstention from conjugation
During the year 1910 a very extensive series of experiments
was carried on with the pure race k, for testing the relative effects
of conjugation and of abstention from conjugation.- The whole
series was so bound together that it might well be considered one
prolonged experiment; it will be convenient, however, in giving
an account of it, to designate as separate experiments the various
phases of it.
306 H. S. JENNINGS
Diagram of history -of these experiments. We have already
(page 302) given some account of the race k. ‘To make clear the
conditions in the present series of experiments, I give a diagram
(fig. 1) showing the history of the various divisions of this race
with which we are dealing; reference to this diagram should
frequently be made in reading the text. The race k as dealt with
in these experiments was derived from a single ex-conjugant of
November 9, 1908; before these experiments were undertaken,
it had passed through three self-fertilizations, or conjugations
with inbreeding; that is, all the surviving members of the race
were descendants by fission of a single individual of the preceding
conjugation; so that the two individuals that make up a pair were
thus descended from one. (In a fourth conjugation in the
series, on March 9, 1909, 25 pairs were saved, so that all came
from these; see diagram, fig. 1.)
After the fourth conjugation, of May 24, 1909, the culture was
allowed to rest till January 29, 1910; during the interval there
may have been many conjugations, in which of course the indi-
viduals would mate at random. On January 29, 1910, a pair
was isolated, from a single member of which came the line of cul-
tures which we shall call B; it forms the branch designated B
in our diagram (fig. 1).
In the remainder of this culture a new conjugation occurred
March 4, 1910. At this time there were isolated certain ex-con-
jugants, one of which gives the series forming the branch C, one
of the three main branches in our diagram; there were also
Fig. 1. Diagram showing the nature and history of Experiments 5 to 14, and
the history of the pure strain k, employed in this work. The rectangles (bounded
by broken lines) indicate each an experiment, and show what sorts of individuals
were compared, with the history of each. A united pair indicates the progeny of
conjugants; a single individual, the progeny of non-conjugants. Thus, in Experi-
ment 9, there were compared the progeny of non-conjugants (of branch A), and of
conjugants (branch #8) that had gone through four conjugations since the others
had conjugated. The dates beside each rectangle show the length of time that
this experiment lasted. The numbers in or (near) the rectangles give the number
under which the experiment is described in the text. The pairs at the left show
(with dates) the known self-fertilizations that the race had undergone before these
experiments began, the survivors being in each case derived from a single ex-con-
jugant, save after the conjugation of March 9, 1909.
307
‘
EFFECT OF CONJUGATION
[ dINST yy
é i161
-zzue
e ee" We uvady 22 '994
018z YE 0, 9 '99G 91-9 °99G ol6l
Pe eg get ote Lounr OL say p‘qe4
02-01 AeW 04 62 4dy OF-6 vdy 0) 62 weW 01 62 ‘uer
6061 6061
vez Aew Kew
0) 6061
6 sew
6 “AON
1, deg L =e 4
0) OL ‘6ny 82 INP
oy € “Unt
Ros 21 dy era!
308 H. Ss. JENNINGS
isolated certain split pairs, in which conjugation had not been
consummated; one of these gives the long line forming the branch
A of our diagram.
This branch A was then propagated without conjugation,
from some period before March 4, 1910, while the other two lines
B and C conjugated repeatedly; the experiments consisted mainly
in comparison of the progeny from these conjugations with the
progeny of the non-conjugating branch A. Finally, on June 4,
a part of A was allowed to conjugate, the rest not, and these
two parts compared. Again, on August 10, a part was allowed
to conjugate, and compared with the part that had propagated
from the beginning of the experiment without conjugation.
In the following an account is given of the results of these various
comparisons.
Experiment 5: January 29 to February 4, 1910: Paramecium
aurelia
Experiment 5 was a brief one, dealing with 10 pairs (20 indi-
viduals) and 10 split pairs (20 individuals), belonging to the
branch B of race k (fig. 1). These 40 lines were cultivated side
by side under the same conditions for six days. Of the descend-
ants of the pairs, three were accidentally lost during the course
of the experiment, so that we have finally but 17 lines descended
from conjugants; 20 descended from non-conjugants.
Of the 17 lines of conjugants, 8 died out during the experi-
ments, a mortality of 47.06 per cent. Of the 20 lines of non-
conjugants, 9 died out, a mortality of 45 per cent.
The number of fissions for the six days of the experiment is
given for the surviving conjugant and non-conjugant lines in
table 11, while the results are summarized in table 12.
Thus, as the tables show, this experiment, so far as it goes,
illustrates the usual conditions:
1. Those that have conjugated multiply less rapidly than those
that have not.
2. The rate of fission is much more variable among those that
have conjugated—both the standard deviation and the coeffi-
EFFECT OF CONJUGATION 309
TABLE 11
Experiment 5. Paramecium aurelia. Number of fissions for the descendants of
pairs and split pairs, during the six days from January 29 to February 4, 1910.
| TOTAL | MEAN
of iat (taelhs eal eat ine"
Conjugant lines (pairs)..............| 1 PA al Se | | 9 | 3.889
Non-conjugants (split pairs) ........ 2S, ori a BL | 5.364
| |
TABLE 12
Experiment 5. Summary of results as to mortality, rate of fission and variability,
in conjugant and non-conjugant lines, for the six days of the experiment.
Ica ee 17| 8 | 47.06} 3.889 = 0.272) 1.663 = 0.192) 42.762 = 5.781
Splitipairs..........| 20; 9 | 45.00) 5.364 = 0.116) 0.771 = 0.082) 14.382 = 1.565
cient of variation being more than twice as great as for the non-
conjugants.
3. The mortality is higher among those that have conjugated.
Experiment 6: March 5 to June 25, 1910: Paramecium aurelia
From March 5 to June 25, 1910, a period of sixteen weeks, a
further experiment was carried on with this pure line k, giving
results as to the difference between those that have conjugated
and those that have not. The experiment was primarily a study
of the fission rate and its inheritance in different races and under
different conditions, so that it included a large number of lines
of propagation, of diverse character. Among these were twelve
lines from the same culture, six beginning with members of pairs
that had just conjugated, and six others derived from individuals
that were beginning conjugation, but were separated before the
process had been accomplished (split pairs). It is only with the
results from these twelve sets, bearing on the effects of conjuga-
tion, that we shall deal here, reserving the remainder of the
experiment for a paper dealing with the inheritance of the rate
of fission.
310 H. S. JENNINGS
A watch glass culture of the race k showed on March 4 the
beginnings of conjugation, many of the individuals being observed
in the act of uniting; (for the history of the race before and after
this time, see the diagram, fig. 1). Three pairs were separated
as they were beginning conjugation; the component individuals
of these three split pairs were called 1, a and 6; 2, a and 3b; 3,
aandb. Three other pairs, designated 6, 7 and 8, were allowed
to finish conjugation, then their members (a and b) were isolated
and cultivated on slides side by side with the others. Thus we
have six lines that have just conjugated; six others that were
attempting to conjugate, but were prevented.
The experiment may be divided into three stages. The first
stage is from March 6 until April 11 (87 days). (Those that had
conjugated did not divide until March 6, so that the fissions of
those that had not conjugated are counted, for comparison only,
from that day also.)
All the six lines of those not allowed to conjugate lived and
multiplied vigorously throughout this period of five weeks and
two days. Of the six lines derived from those that had conju-
gated, on the other hand, four died out completely within two
weeks, while the others multiplied more slowly than did those
that had not conjugated. The detailed records for the 12 lines
are given in table 13. Here the classification by weeks begins
March 8, so as to omit in the weekly record the irregularities
due to the first two or three days after conjugation; this gives
us just five weeks.
As table 13 shows, even the two lines of conjugants that lived
multiplied less rapidly than did those that had not conjugated,
the weekly average for all the former being 8.8, while for the
latter it is 10.2. The next stage of the experiment was of a
character to determine whether this difference in rate of fission
continues beyond five weeks, as well as to decide whether it
might be due to accidental causes, or was a result of inherent
differences. Of the non-conjugants, the pairs 1 and 3 were con-
tinued till April 12 (31 days additional), while the conjugants were
represented by 7a only. Several lines of each set were kept in
progress. The results are given in table 14 for two periods of
EFFECT OF CONJUGATION 311
two weeks each, also for the total, thirty-one days; and for the
entire sixty-eight days from the beginning.
As table 14 shows, the progeny of the conjugant line 7 a still
reproduced somewhat more slowly than did the non-conjugant
lines, during these last four weeks of the nine weeks during which
the experiment had lasted. The average in the conjugants is
less in each of the partial periods, as well as in the period as a
whole. The average rate of fission is somewhat less in all lines
than during the first five weeks; this is a common result of con-
tinued cultivation on slides. :
In the third portion of the experiment two of the lines of non-
conjugants (1 a and 1b) and one of the lines of conjugants (7 a)
TABLE 13
Experiment 6. Paramecium aurelia. Comparative number of fissions in six lines
derived from conjugants and in six derived from non-conjugants (split pairs),
from a watch glass culture of the race k; for five weeks and two days. The fisstons
are given by days for the first nine days; for the rest only by the week. The num-
bers give the number of divisions that occurred in the time specified. The first
week is counted from March 8 to March 14. (d = died out.)
DAILY RECORD; DAYS OF MARCH RECORD BY WEEKS 3 2
6 | 7 | 8] 9 | 10| 11 | 12] 13 | 14 | 1st} 2d] 3d [4th [5th O53
A. Those that have conju-
gated:
Ge cease ecioeeiae Te 2a Te OP @) |d|d
De. 1 TE PE Oat, Oa ae se Iie?
U -2is.5 tro Cae Eee AE eB Se ee IE SE OL 29) hed |, Fan RO ae aes
Cin 8 6 ae eee ae eee d d
SMM ee SSP ichs,5 Se 3: 25 ONO a d
Dold clomtoere es ee ee TN A a 7 2{|11|1 | 10} 7 | 10} 10) 10) 48
Mean per week, 8.80; mean for 87 days, 45.00
B. Those that did not conju- |
gate
PEt ries ee eta 1)/2);2)1);2)1}2) 2) 0 | 10) 11) 10) 13) 11) 58
Dee eae Re 1} 27211) 212) 4) 2) 1 | 1) 6) 11) 12) 10) 52
EI eI ors tenens iter choke fe ee lene ae tees eer elas ate eS hOl en LS th o9
| Dinvercatarr eters nomen, Sacemn La ete al eo eteete al Oi Ol Ont st “O50
MOI revit sacl ex hoa eee 2 le eee ete Ob LO ia) (91.54
De Sra tnd Sats ok 2 Sale le WO) eOmmOW coe Si LOKAT
_Mean per week, 10.20; mean for 37 days, 53.33.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
312 H. S. JENNINGS
were cultivated six weeks longer, making a total of sixteen weeks
or 112 days. The comparative rates of reproduction for this
period, as well as for the entire time, are shown in table 15. One
point in this table requires explanation. During this period of
forty-four days there were left of each set for various purposes
separate lines of propagation which lasted less than the total
period; for example, one line was continued seven days, another
twelve, et cetera. ‘These diverse periods have been summed for
TABLE 14
Experiment 6. Comparutive rates of fission for certain lines that had conjugated
March 4, and for others that were prevented from conjugating at that time. The
number of fissions is given by periods of two weeks, for weeks 6 to 9; also for the
last 31 days of a period of 68 days; and for the entire period.
WEEKS WEEKS LAST | TOTAL
6 AND 7 8 AND 9 31 Days 68 DAYS
A. Conjugant
2 oy CHINO D6 8 te tan 13 14 30 ie
(ine 2) sec) ke 13 15 30 |
(limes hx heen ce 15 il |
Mean for 7a...... 13.7 13.3 30 72
B. Non-conjugant |
ar (hy Se Sees ose 15 16 35 88
CS ee, ee ee 15 17 36 94
(3) ohare ene 13 14 31
(4)EE chee eee 15 leis :
Meant ff 28 haa 14.5 | 15.7 34 | 91
DD: (1) es ein eee 18 15 37 88
(2) 3.0 ee eee 15 15 35 |
(S) i suki eee Mie Ro 20 |
(4).. 18 ? |
Mean ars 15 36 88
Se RR GB IN a gle: | 17 13 36 | 90
3b (1) 12 15 31 | 78
(2). 13 16 32 |
(8).. Ras, Sees, | 4 14 : han
14 STs h a ame tk pe | 12.5 15 31.5 78
Mean of means for all non-!|
conjugants...............| 15.94 14.68 34.38 | 86.75
EFFECT OF CONJUGATION 313
each set, and are presented, with the total number of fissions
during the periods, in the entry numbered 2.
Examination of table 15 shows that the line 7 a, derived from
a conjugant, no longer differs in any very marked or constant
way in its rate of fission from -the two derived from the non-
conjugants; it is certainly not slower in its rate than the others.
So far as the experiment goes, it indicates that after about two
months the rate of fission of the conjugants, which had been made
slower by conjugation, has regained about the usual rate. Owing
to the small numbers of diverse lines involved, such a conclusion
is of course not very secure.
TABLE 15
Experiment 6. Paramecium aurelia. Relative numbers of fissions in certain per-
iods, and rates of fission, in certain conjugant and non-conjugant lines, for the
last 44 days of an experimental slide culture that lasted 112 days from the time
of conjugation; also totals for the entire 112 days of the experiment.
CONJUGANT NON-CONJUGANT | NON-CONJUGANT
LINE 7a LINE la LINE 1b
Days /|Fissions Daly Days /|Fissions Dally Days Fiasiong ow!
A. Changed every 48) | | | |
hours during |
last 44 days |
1. Single consecu-) |
tive line}... 44 48|1.091|} 44] 38] .864| 44 45 | 1.023
2. Sum of diverse |
periods for
parts of line..| 157| 155 | .987| 141| 123] .872| 135] 189 | 1.030
3. Single consecu-| |
tive line from | |
beginning.....| 112 120 1.071) 112 | 133 | 1.188) 112 | 1833 | 1.188
B. Changed every 24) |
hours during
last 44 days: |
1. Single consecu-
tive line....... | 44) 62|1.409} 32; 40/1.250) 44; 60)| 1.364
2. Sum of diverse) |
periods, for |
parts of line..| 140 176 | 1.257, 88
3. Consecutive line |
from beginning 112 | 134) 1.196 100| 128 | 1.280 112 | 148 | 1.321
314 H. S. JENNINGS
The experiment as a whole shows the fact that after conjuga-
tion the organisms are in a condition such that many may die,
while those that have not conjugated live; and the further fact
that the rate of reproduction is made slower by conjugation,
remaining in this condition for about two months. This is true
even when all the lines concerned belong to the same race (derived
originally from the same single individual).
Experiment 7: March 29 to April 10, 1910: Paramecium aurelia
On March 29 there was a conjugation in the progeny of a
single ex-conjugant of January 29; the relation of these to the
remainder of the experiments will be seen from the diagram,
figure 1. These animals belong to the branch B of the diagram.
They have conjugated once, and probably twice, since those of
the branch A, which are known not to have conjugated since
some period before March 4, and to have been ready for conju-
gation March 4. A comparison as to rate of fission was made
between these non-conjugants of branch A and the conjugants
of March 29, branch B, lasting for nine days (April 1 to April 10).
Of the conjugants of branch B, 19 lines were in progress; their
average rate of fission per line for the nine days was 1.409 per
day. Of the non-conjugants (branch A), 21 lines were in progress;
their mean rate for the nine days was 1.455 per day.
Thus the non-conjugants of branch A give no indication as
yet of injury through having omitted conjugation. The differ-
ence in rate between conjugants and non-conjugants was slight,
but in favor of the non-conjugants.
Experiment 8: April 9 to April 30, 1910: Paramecium aurelia
On April 9 there was conjugation among the progeny of one
of the ex-conjugants of Experiment 7 (conjugation March 31)
shown in branch B of figure 1. There have now been seven
generations of inbreeding in this branch; and it has conjugated
twice (probably three times) since those of branch A have con-
jugated at all. What difference will this make between the rate
of fission in the members of the two branches?
EFFECT OF CONJUGATION 315
Eight lines descended from these conjugants of April 9 (branch
B) were kept under observation till April 30. As the first fission
did not occur till April 12, this gives nineteen days during which
the rate of fission was determined for these. Of the non-conju-
gants of March 4 (branch A, fig. 1), fifteen lines were in progress
at this time.
The average number of fissions for the nineteen days was, in
the conjugants of branch B, 21.375, while the average rate of
fission was 1.125 per day. In the non-conjugants of branch A,
the average number of fissions for nineteen days was 21.533, the
daily rate 1.133.
There was thus no appreciable difference in rate of fission.
The branch A shows no sign of injury as a result of having
omitted several conjugations which the branch B has undergone.
Experiment 9: April 29 to June 7, 1910: Paramecium aurelia
On April 29 there was another conjugation in branch B (9, fig. 1)
in the same direct line as the conjugations of Experiments 7 and 8.
That is, the conjugants of our present experiment are all descended
from a single ex-conjugant of Experiment 8, these from a single
ex-conjugant of Experiment 7, and soon. Thus there have now
been in this branch B three conjugations (probably four), since
there has been a conjugation in branch A. The members of
branch A have been cultivated on slides since March 4, while
those of branch B have been cultivated part of the time on slides,
part of the time in watch glasses.
Precise comparison of branch A (fig. 1) with the progeny of one
of the ex-conjugants of April 29 in branch B was not made till
about two weeks after the conjugation of the latter. During this
time observation with the eye seemed to indicate that the mem-
bers of B were a little larger than those of A. In view of this
apparent differentiation between the two, experiments were set
on foot for comparing the fission rate and the dimensions.
Fission rate. Four separate series or lines of propagation were
carried on from May 22 to June 7, both for the conjugants (B) and
the non-conjugants (A).
316 H. S. JENNINGS
We may divide this time into two periods of eight days each.
The results for the two sets are given in table 16, the four parallel
lines of each set being numbered (1) to (A).
The table shows that for the total period, each of the four
lines of A multiplied more rapidly than any of the four lines of B.
For any of the eight-day periods, the lowest record for A is at
least equal to the highest for B, save in one single case. The
lines of A average for the entire period 20.5 per cent more
fissions than those of B.
When we recall that B has conjugated recently, and three
times since A has conjugated at all, we see that the dropping
out of the conjugations has not unfavorably affected the rate of
reproduction in A.
Dimensions. To the eye it appeared that B was a little larger,
under the same conditions, than A. As these belong to the same
pure strain, this is a matter of* interest, as it would show that.
hereditary differences in size may arise within the pure strain
possibly as a result of conjugation. A careful comparison of the
dimensions was therefore made. Keeping all under the same
cultural conditions, I first measured a number of individuals of
each set at the same age, choosing the age of thirty minutes after
fission. Three other measurements were taken; the results of
all are given in table 17.
The fact that, as table 17 shows, B was larger at each of the
four measurements, seems to indicate that there has indeed
arisen a slight hereditary differentiation in size within the pure
TABLE 16
Experiment 9. Paramecium aurelia. Comparative number of fissions May 22 to
June 7, for two sets, one of which (B) has conjugated three times in series since
the other (A).
reas | aoe ea goer, | geo | ee
B. Line (1); 18 8 21 As Line. (iq) abe al ae 26
(2)} 14 5 (2)} 15 9 24
(3), 14 7 21 (3)| 14 13 27
(4)| 14 8 29 (4)) 14 12: | Sa
Meam 13.75 7.0 20.75 | Mean| 13.75] 11.25| 25.00
EFFECT OF CONJUGATION ole
TABLE 17
Experiment 9. Paramecium aurelia. Comparative lengths in microns of A and B.
These belong to the same pure strain, but A has been cultivated for a long time
without conjugation, while B has conjugated at least three times in succession
since A. :
A B
No. Mean length No. Mean length
May 19: age, 30 minutes..................| 10)116.600 = 0.551
11/125.454 + 1.063
May 22: adults, ill fed....................| 46, 99.000 = 1.195] 50}123.680 + 0.913
May 28: adults, well fed..................| 64/135.469 = 1.042) 70141.429 + 0.785
41|132.237 + 0.747
PATTEM ER LATEGGE, oc, <5, pose eines <-obhaserstehau sy ec 65131.877 = 0.616
line k. However, the fact that the difference was so extremely
small at the last measurement taken admonishes us not to lay
too much stress upon this case; the matter must be tested further.
Determination of conjugation. With A and B (fig. 1) a study
was made as to the relative influence of external and internal con-
ditions in inducing conjugation. In B, as we have seen, there
had been at least four successive conjugations since there has
been one in A. Will A be therefore readier to conjugate than B?
If conjugation depends mainly upon an internal condition of need,
then certainly we should expect this.
To test this, watch glasses of A and B were set side by side
May 13, two weeks after the last conjugation of B, and the con-
ditions for inducing conjugation supplied, so far as possible.
On June 3 conjugation occurred in both A and B.
Thus under the proper conditions both sets conjugate at the
same time, in spite of the fact that one has conjugated at least
four times since the other. The experiment indicates that the
recent external conditions are of more importance in determining
conjugation than a progresive internal need arising through the
fact that conjugation has not lately occurred.
Experiment 10: May 10 to May 20, 1910: Paramecium aurelia
On May 10 there was a conjugation in branch C (fig. 1, page
307), making the third in series in this division since any conju-
gation occurred in A. Four pairs were isolated from this new
conjugation in branch C, and experiments were set on foot for
318 H. S. JENNINGS
comparing these as to vitality, reproductive power, et cetera,
with A (which had not conjugated for some months), also with
those of branch B (which had conjugated two weeks before).
The four ex-conjugants of two pairs of C were placed on slides
and treated like the non-conjugants of branch A. Two of these
ex-conjugants divided once, two did not divide at all, and all
died after three to ten days. Meanwhile, the members of branch
A multiplied actively, at about the rate of once per day.
Two other pairs from branch C were allowed to multiply in a
watch glass. This they did very slowly, so that on May 13 but
10 individuals had been produced from the four. These were
then carefully brought into identical conditions with an equal
number of specimems of A, and of B, the three being placed side
by side in watch glasses.
On May 16, all the specimens of C were dead, while the mem-
bers of A and B were flourishing.
Thus in this case, the recent conjugants (C) multiplied very
slowly or not at all, and soon died; while others that had not
conjugated so recently nor so often (A and B) flourished.
Experiment 11: June 3 to July 28, 1910: Paramecium aurelia
Comparison of conjugants and non-conjugants of the branch
A (fig. 1, page 307).
On June 3 there was conjugation in a watch glass (taken from
the slides May 15) of members of the branch A (fig. 1), derived
from a split pair of March 4. Other divisions of this same stock
(branches B and C, fig. 1) had conjugated four times in succes-
sion since any conjugation in A. Thus we have in A a set that
has gone long past the normal conjugation period. Part of it
now (June 3) conjugates, while the remainder (part in the watch
glass, part on slides) does not. Thus we have an opportunity to
test the effects of conjugation on the vitality and reproductive
power of a stock that has long gone without it. Experiments
for this purpose were conducted on slide cultures, and also in
watch glass cultures.
To make clear the conditions in these experiments (which
when described in words alone are a little confusing), I give in
EFFECT OF CONJUGATION 319
figure 2, a diagram, which shows the relations of the various
parts of the experiment. If the reader will make frequent refer-
ences to this, he will have no difficulty in following the account
of the experiments, and appreciating their bearings.
We have, derived from the branch A of figure 1, at first two
divisions, designated D and FE in figure 2; D has been cultivated
continuously on slides, since March 4, while # was transferred
on May 15 from the slides to a watch glass.
+ Exp.12
E Paes ee |
s Jun. 22-25 Jul.6-28 Jul.16-19 1 Do2
Slides Watch: giassesi:g 5 eae nn tnify ann ish !
|
!
,Di
|
Mar. 21 K
------ aes
Aug. 11 -
!
!
!
1
| C=!
7 Jun. 22-25 Jul.16-28
I
'
1
|
1
Fig. 2. Diagram showing the history and the nature of the comparisons made
in the various divisions of Experiment 11, and in Experiment 12. Each rectangle
represents one of the parts of Experiment 11, save the one to the right, which
represents Experiment 12. The lower case letters a to g, within the rectangles,
show the letters under which the various parts of Experiment 11 are described in
the text; thus the one marked e shows Experiment 11 e. The capital letters A to K
show the diverse branches of the race k, described in the text. The duration of the
experiments is given by the dates below or above each rectangle. The legends at
the branches D, EH, etc., show the manner of cultivation; the branches D and #
being separated on May 15, the former placed on slides, the latter in watch-glasses,
ete.
Now, on June 3, this branch £ is divided into four parts, which
we may call fF, G, H and K. The first two are kept on slides;
the second two in watch glasses. The lots F and H are non-
conjugants, while G and K are conjugating pairs kept under the
same conditions. Comparison of F and G we may call Experi-
ment 11 a; comparison of H and K, Experiment 11 6b.
320 H. 8. JENNINGS
Experiment 11 a: cultivation on slides. From the watch glass
(fig. 2, FE), thirty-two ex-conjugants, from 16 pairs (G, fig. 2) and
thirty-two non-conjugants (F, fig. 2) were isolated. These 64
lines were. cultivated side by side, on slides, under identical
conditions.
In both sets the mortality was high (as I have invariably
found to be the case in attempting to cultivate Paramecium
aurelia in hot summer weather). By June 13, 25 of the 32 lines
of ex-conjugants (@) had died. By June 16 all the ex-conjugant
lines were dead, while nine of the non-conjugant lines (Ff) were
still alive.
In the ex-conjugants the average number of fissions, for those
that lived to June 13, was 5.445, while in the non-conjugants, for
the same period, it was 7.857. The average rate of fission for
all the ex-conjugants (reckoning the rate for each one as long as
it lived), was 0.5485 per day, while for the non-conjugants,
reckoned the same way, the average rate was 0.7365 per day; so
that the rate for the non-conjugants was 34.27 per cent greater
than for the ex-conjugants.
Experiment 11 b: cultivation in watch glasses. Parallel to the
slide cultures of 11 a, two watch glass cultures, one of conjugants
(fig. 2, K), one of non-conjugants (H), were propagated. One of
these contained at the beginning 30 ex-conjugants, the other 30
non-conjugants. On each of the following days the animals were
removed one by one to a new watch glass, counted, and the
number reduced so as to be the same for each. The ratio of the
number present on each day to the number present the day
before was thus obtained, this may be called the multiplication
ratio. It was as follows for eight successive days, beginning
June 5:
DAY 1
i)
ot)
~
eo
a
x
io)
=
13]
>
Z
Progeny of conju- |
gents (KK)... 2. 1.67) 1.84 | 2.17 | 110.) Yd) 1:26) tt See
Progeny of non-con- | |
jugants (H)........ 2.80 | 2.08 | 1.92 | 1.52 |.1.80 | 1.80 | 1.74 | 1.56 | 1.90
EFFECT OF CONJUGATION 321
Thus the rate for the progeny of the non-conjugants was
greater every day except one; and the mean rate for the non-
conjugants was almost exactly 20 per cent greater than for the
progeny of the conjugants.
It is clear therefore that the progeny of the non-conjugants
have a great advantage, both as to rate of reproduction and as
to mortality.
Experiment 11 c. The two sets in the watch glasses (fig. 2,
H and K) were allowed to multiply till June 22, then 10 individ-
uals were removed from each, isolated on slides, and their rates
of fission followed individually and compared. All multiplied
vigorously, in two days 8 of the conjugant progeny had divided
five times; 1 three times, 1 six times. Of the non-conjugant
progeny, 5 had divided six times, 2 five times, 2 three times,
while 1 died. Thus the two are now nearly equally vigorous,
the rate of fission being still a trifle higher for the non-conjugants.
Experiment 11d. The non-conjugants dealt with thus far
(Experiments 11, a to c) had come originally from the same
watch glass as did the conjugants (that is, from branch Z£, fig. 2).
But there was under propagation at the same time, another set
of non-conjugants (branch D, fig. 2), cultivated on slides since
March 4, while those just described (F to K, fig. 2) had been cul-
tivated in watch glasses (hence with more fluid) since May 15.
On June 7 a watch glass culture of this slide series (D, fig. 2) was
made, and left uniform with a watch glass culture (branch K,
fig. 2) derived from the conjugants of June 3. On June 22, 10
individuals were taken from each of these two watch glasses
(D and K, fig. 2), and cultivated on slides, in order to compare
their fission rates and general vigor. The results were strikingly
different from those thus far obtained; on account of their great
interest I give them in detail.
Table 18 shows that, contrary to all our previous results, the
progeny of the conjugants K are much more vigorous than those
of the non-conjugants D.
Now, in the experiment which just preceded this (11 ¢), we saw
that in a test made on the same date as the present one, and with
conditions identical, the non-conjugants (H) with the same history
322 H. S. JENNINGS
TABLE 18
Experiment 11 d. Paramecium aurelia. Number of fissions for three successive
days, in ten non-conjugants from the slide series D (of fig. 2), and in ten progeny of
the cenjugants (K, fig. 2), that had lived in watch-glasses since May 15 (d = dead).
NON-CONJUGANTS (D) CONJUGANTS (kK)
LINE i Tae ara A ae June
____—___—_— Total | Total
23 24 25 23 | 24 25
ieee 1 1 ee ily A) 3 4 2 4
2. 1 d= |e at ad (0)
eee 1 d | (Deal = Sing Jee 1 6
4. d Poy fc Sage Gate 1 6
Bese) 0 d Gt See 1 6
6. 2 0 0 2 So) he 1»). iheaa
re eee: 2 0 0 2 3 3 L- Ate
Spee d (0) 3 | 2 2 | 7
9. 0 d | (0) 2 eet ide 2) ae
Wie <4! 2 1 Dry Hs ay Pere 3 | 8
as the conjugants (A) (cultivation in watch glasses since May 15)
were not less vigorous than the progeny of the conjugants, but
at least equally vigorous with them. The only difference between
the non-conjugants of this present experiment (D) and those of
the previous one (H) is that those in the present experiment were
cultivated about a month longer on slides (May 15 to June 7).
Apparently this is the cause of the weakness of these animals.
But it is clear that the progeny of the non-conjugants of this
slide series D are now in a weakened, depressed condition, while
both conjugants and non-conjugants from watch glass cultures
are vigorous. This gives us an opportunity to determine the
effects of conjugation in such a depressed culture. Before under-
taking this, two additional tests were made to see if the depressed
condition of the progeny from the slide series D was beyond
doubt.
Experiment 11 e. The first of these consisted again of ten lines
of the non-conjugant slide series D, ten from the conjugants
(watch glass series, K). The experiment continued from July 6
to July 28.
In this experiment, as soon as any line of either set died out,
it was replaced from some other line of the same set. The number
EFFECT OF CONJUGATION ane
of such necessary replacements will give a comparative meas-
ure of the mortality in the two sets. In the non-conjugants D
(shde series) there were necessary 34 such replacements; in the
conjugant progeny K (watch glass series), there were 22. Of the
conjugant line (K) six lived to the end (21 days), the average
number of fissions for these being 21. Of the non-conjugant
lines (D), four were alive at the end, their average number of
fissions being 16.25. On the sixteenth of July ten lines were
alive in each set; the average number of fissions at that time was
for the conjugants (K) 15.6; for the non-conjugants (D), 13.00.
Experiment 11 f. On July 16, an additional comparison was
made, taking fourteen each of the conjugants (A) and non-
conjugants (D). In the nature of the results this experiment
lasted but a short time. Of the non-conjugants D (slide series),
only two divided, and all were dead by the third day. Among
the conjugants (K), all divided; three died out during the three
days; the remaining 10 averaged six fissions in three days. The
non-conjugants (D) are clearly depressed and weak.
Experiment 11g. A crucial question is whether the two sets
of non-conjugants of unlike history—the slide series D and the
watch glass series H—are still unlike in their vigor, as comparison
with the conjugants (Experiments 11 a-—f) indicates. Therefore,
on July 16 comparative tests were made of these two, and also
of the progeny of the conjugants of June 3 (K).
The three lines compared are those designated D, K, and H,
figure 2. Of line D (non-conjugants cultivated on slides up to
June 7) two watch glasses, containing five specimens each, were
taken. Of the other non-conjugant line H (cultivated in watch
glasses since May 15), the same number was propagated, in the
same way. Of the ex-conjugants of June 3—the line K, with
the same cultural history as the non-conjugants of line H—three
‘ watch glasses of five specimens each were propagated. In all
three series the conditions were made exactly alike, all the animals
being washed in the same water before they were introduced into
the culture fluid. At intervals of some days the animals were
removed one by one to a new watch glass, counted, and the
number reduced, so as to be the same in dll. The ratio of those
324 H. S. JENNINGS
present to those that had been introduced was taken at each;
this gives the ratio of multiplication for the two days. These
ratios are given in table 19.
From the results given in table 19 the following are clear:
(1) the non-conjugants, D, cultivated on slides till June 7, are
still much depressed, and much less vigorous than either the
non-conjugants (H) which came from the watch glass with the
conjugants, or than the conjugants themselves (K); (2) the
non-conjugants (H) which have the same cultural history as the
conjugants (A) are still somewhat superior in vigor to the con-
jugants. In every one of the four periods their ratio of multi-
plication is greater.
The results of this experiment are then throughout consonant
with those of Experiments 11a, 116 and llc. From all, the
TABLE 19
Experiment 11 g. Paramecium aurelia. Ratio of multiplication for a number of
periods of time, in watch glass cultures of the three sets D, H and K, described in
the tect.
| }
JULY 16-18 | suLy 18-20 | JULY 20-24 | JULY 24-28
| |
(D) Non-conjugants, cultivated on| | |
slides till June 7. |
Culture ih ins% yc phe ee |) DS) A ee ee
GCuiltiite 2222. Ss¢ at pis ho aves } ma 1. a
Mesa. crsiorn) oes 14 | 1.989 | 3.65 | 1:25
(H) Non-conjugants from same watch- |
glass as the conjugants of K, set | |
May 15. |
Culture dai: ts ote poe 2.8 | 2.957 | 6.5 6.2
Culture 2.) oo ok cc RLS ane 4.02 2.476 | 5.5 | 10.0
Weéan<) 1.u Ae ciiash sees eee 3.41 | 2.667 |. 6.00 | 8.1
|
(K) Conjugants of June 3, from same |
watch-glass as the non-conjugants
of H.
Culture $2. ser.'ac do eee 3.00 1.533 3.9 6.6
Culbare 2.25 .. a5t. street 2.00 | 2.110... 62 3.7
Culture 3. 3.00 1.867 5.9 6.4
EFFECT OF CONJUGATION ao
conclusion is evident that the reason why the conjugants (A) are
more vigorous than the non-conjugants (D) of the slide series,
is because the latter were cultivated for a month longer on slides.
For the non-conjugants (H), cultivated throughout in the same
way as the conjugants, are still more vigorous than the latter.
To this the only objections that could be raised in favor of
the view that conjugation has caused rejuvenescence would be
as follows: (1) It might be held that it is not possible to be
certain, in taking free specimens from a watch glass containing
conjugants, that one is not taking specimens that have already
conjugated. I believe that there is no ground for this objection
- in the present case, since the cultures in question were watched
with the utmost care in order to detect conjugation at its begin-
ning. Further, Experiments 11 a@ and 11 b show that there was an
actual difference between conjugants and non-conjugants, of the same
sort as that found in all our other experiments, the non-conjugants
being the more vigorous. It is clear therefore that in this case
we were actually dealing with non-conjugants. (2) It might be
said that it would be difficult to be certain that in the period that
had elapsed since June 3 there had not been conjugation in the
vessels of ‘non-conjugants.’ It may be admitted that this diff-
culty exists, though in this case the organisms were inspected
daily, and I believe that no conjugations occurred. But this
objection is in fact negated by the experimental results themselves.
The greater vigor of the non-conjugants seen in Experiment 11 g,
of July 16, is of just the same character as that seen immediately
after conjugation (Experiments 11 a and 11 6) and as that charac-
teristic of the non-conjugants in all our other experiments, to
which the present objection is not applicable.
But is the explanation given above—the fact that the non-
conjugants of one set (D) were cultivated a month longer than
-the other (H) on slides—a credible one? I believe that everyone
who has had extensive experience with experiments of this sort
will agree with me that long continued cultivation on slides does
produce a depressed condition. There are some stocks that will
not stand it at all, though they live perfectly in mass culture.
And in fact, my experimental records show that in this partic-
326 H. S. JENNINGS
ular stock of A (fig. 1) those cultivated on slides became very
unhealthy in June, so that all died out June 18; the branch was
preserved only because some of them had thus been removed to
watch glasses on June 7, forming our present stock D (fig. 2).
Thus all the evidence, from many distinct sources, points to
the explanation we have set forth above, that the non-conjugants
of the slide series D are less vigorous than either the non-conju-
gants (H) or the conjugants (K) of the watch glass series, because
they were cultivated longer on slidés.
Experiment 12: Conjugation in a depressed stock: August 10 to
September 7: Paramecium aurelia .
Whatever the cause, we now have on hand, after the Experi-
ments 11 a to 11g, a much depressed stock, which has omitted
at least four normal conjugations. Now, it might be maintained
that our uniformly negative results thus far as to rejuvenescence
by conjugation are due to the fact that we were not dealing with
depressed stocks. It is of interest, therefore, to determine the
effects of conjugation within a stock thus known to be depressed.
There is, however, great difficulty in carrying out such an experi-
ment, for such depressed stocks cannot easily be induced to
conjugate. The main condition for conjugation appears to be,
that there shall be a period of rapid multiplication, followed by a
decline in the conditions inducing it. But in such depressed
stocks it is almost impossible to induce rapid multiplication.
After many efforts, I finally succeeded on August 10 in getting
a scanty conjugation in a watch glass culture of this depressed
set of the slide series of D (fig. 2) which has descended without
previous conjugation from the split pair of March 4. I was able
to obtain for study but three pairs, the ex-conjugants of course
forming when isolated six lines of propagation. I call these
sets D 1.
From the same watch glass culture I isolated at the same time
ten of the individuals that were not conjugating, and from these,
ten lines of propagation were derived, which were kept under the
same conditions as those from the conjugants. These may be
called D 2.
EFFECT OF CONJUGATION ooh
For further comparison, I placed beside these, ten lines derived
from a culture of this same series that had conjugated June 3,
and had lived in small mass cultures since. These belong to
the line K, figure 2.
As this experiment gives the only results obtained that could
be interpreted as showing a favorable influence of conjugation
on survival and reproduction, it appears best to give in some
detail the records for the lines of propagation. This is done in
table 32, in the Appendix.
To understand table 32, the following must be considered:
Each line of any of the three sets was started with a single indi-
- vidual August 11. The animals were changed to new fluid
every day or every other day (all being alike), and all those that
had been produced were retained, save on certain days, when the
number was reduced, by removal of certain of the animals. To
give the essential facts in the history of the cultures, it is there-
fore necessary only to give the number of individuals on these
dates, before and after reduction. This is what is done in table 32.
Thus, in line 1 of set K, the single individual of August 10 had on
August 14 produced 6, of which 4 were removed, leaving 2. On
August 16, these 4 had produced 8, of which all but 2 were
removed, et cetera.
The last column of table 32 gives the total number of fissions
undergone by the line in question, up to September 7, or to its
death. As the lines of set K were all obviously vigorous, only
five were kept under observation till September 7, when the last
line of set D 2 died out.
To grasp the results, it will be best to examine first the facts
for set K, which had lived in mass cultures since May 15, and
had conjugated June 3. In this case, as will be observed, all the
ten lines flourished well. Number 4 was lost by accident August
28, and numbers 3, 5, 7 and 9 were discontinued September 3,
because the results were clear.
Now, compare with these the results given for the conjugants
and non-conjugants of August 10 (the depressed race) in sets D 1
and D2. It is clear from the data of set D 1 that the conjugation
of August 10 has by no means restored this depressed series to the
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
328 H. S. JENNINGS
level shown by the other set K of table 32. In spite of the utmost
care, and the division of the lines as soon as possible, so as to have
more than one from each of the ex-conjugants, the progeny of
three of them had died out within eight days, and a fourth died
out later. Only two of them survived till September 7, when this
experiment was discontinued; and the nine lines derived from
these two were then multiplying much less rapidly than those
derived from set AK (table 32). It should be stated further that
two tests made respectively three and four months later (one
December 1, 1910, the other January 7, 1911), showed that the
members of set AK (conjugants of June 3) were still far more
vigorous than conjugants of August 10 (set D1). In the test
comparison of December 1, all the twelve lines of set D1 (con-
jugants of August 10) died out after a week of cultivation on
slides, while those of set K flourished. .
When however we compare the records for the conjugants of
August 12 (set D1) with those for the non-conjugants of the
same culture (set D2) in table 32, we find that the conjugants
have a decided advantage. The non-conjugants ceased multi-
plication almost entirely, after the first week, and gradually died
out, the last one dying on September 7. At this time the descend-
ants of two of the ex-conjugants were multiplying well, so that
an indefinitely large number of progeny were later produced
from them.
Thus in this case two of the six conjugants were more vigorous
than any of the non-conjugants of the same stock and cultural
history. Most of the conjugants died out, but in the natural
course of events the entire set would have been replaced by the
progeny of the few more vigorous lines.
This is the only case, out of a very large number of experiments,
that gives any indication of a beneficial effect of conjugation on
vigor and survival. Just what has happend here? First, atten-
tion should be called to the fact, already set forth, that conjuga-
tion in this depressed stock was very scanty; in connection with
the further fact that the condition for producing conjugation is a
period of rapid multiplication, followed by a check. Now, from
this and from the data of tables 18 and 19, it is evident that there
EFFECT OF CONJUGATION 329
were few in this depressed race that could be induced to multiply
sufficiently to furnish the conditions required for conjugation.
Those that did conjugate evidently represent then those members
of the stock that are most vigorous and active in multiplication.
Their later vigor and survival, as compared with the non-conju-
gants, may therefore have been due to this, and not to the con-
jugation; in other words, conjugation may have been the effect,
not the cause, of their greater vigor. If the same individuals
that conjugated could have been cultivated without conjugation,
it is probable that they would have multiplied equally well or
better.
However this may be, it is clear that conjugation did not cause
rejuvenescence in any simple direct way, since the majority of
the conjugants died out, and those that survived were weak.
But in one respect this experiment gives the same results as all
others. Conjugation resulted in an increase of variability, as
regards vigor and rate of reproduction. Among the extreme
variates were some whose vigor was sufficient to keep them alive,
while among the more uniform non-conjugants all died. The
advantage of the conjugants, so far as it did not exist before con- °
jugation, is then in this case due to the effect of conjugation in
increasing variation.
Experiment 13: Production of inherited differentiation by conjuga-
gation: December 6, 1910, to May 15, 1911:
Paramecium aurelia
A very extensive and long-continued series of experiments was
carried on in the winter and spring of 1910-1911, with the same
pure strain k, of Paramecium aurelia, that was used in the experi-
ments just described (Experiments 6 to 12). The main purposes
of this new set were, to determine whether as a result of conju-
gation differentiations may arise within a pure strain, and to
bring out the rules of inheritance within the pure strain. Most
unfortunately, in the later and most critical part of the experi-
ment the conditions became such that multiplication almost
ceased, and this made futile a large part of the work, particularly
that designed to discover the rules of inheritance. Whether this
330 H. S. JENNINGS
cessation of reproduction was due to something in the cultural
conditions, or to weakening of the stock as a result of long con-
tinued culture on slides is perhaps not absolutely clear, though
the evidence is strong that the latter alternative is the correct
one. But in spite of this, the experiment gave definite results
on some important questions. I shall give the experimental
data only so far as they throw light on definite problems.
From many other experiments the general impression had been
obtained that conjugation produces inherited differentiation even
within the pure strain. By ‘pure strain’ is meant here simply a
series of animals all derived from one single individual. Experi-
ments set forth in previous papers indicate that no inherited dif-
ferentiation within such a pure strain arises, as a rule, during
multiplication by fission; and this agrees essentially with most
other work on inheritance in vegetative reproduction. The
evidence, so far as Paramecium goes, was based mainly on studies
of the inheritance of size. If, as these indicate, heritable dif-
ferentiations do not arise in fission, then the question comes up
as to how the existing differentiations into diverse races do arise.
The indications just mentioned, that conjugation produces such
differentiation, then of course call for investigation; this was
attempted in the present series of experiments.
If inherited differentiation does result from conjugation, this
might be held to be due to Mendelian inheritance, or something
similar. If the individual with which the pure strain began was
a heterozygote, and its progeny through fission were identical
heterozygotes, then of course when these interconjugated, new
combinations of various sorts might be produced, exactly as
differentiations may arise by self-fertilization of heterozygotes
in plants.
Hight self-fertilizations. To avoid, so far as possible, the
heterozygotic condition, I used the race k, of Paramecium aurelia,
already described in connection with Experiments 5 to 11 (see
diagram of its history, fig. 1, page 000). At the time when the
present series of experiments (13 and 14) begins, self-fertiliza-
tion had occurred in this race eight times in series. That is, the
progenitor of the race was a single individual; its progeny con-
EFFECT OF CONJUGATION 331
jugated among themselves; from these conjugants a single
ex-conjugant was taken and allowed to multiply till there was
conjugation among these. A single member of a pair was again
allowed to propagate till there was conjugation; and thus the
process was repeated eight times, all the members of each of the
eight non-sexual series being the progeny of a single ex-conjugant
of the previous series. The known history of this race is illus-
trated in the diagram of figure 1. This diagram shows also the
relation of the organisms employed in the present experiment to
those used in previous experiment. They belong to branch B of
figure 1, and are derived from a single ex-conjugant of the conju-
gation of April 29; they are thus the same stock as the conjugants
employed in Experiment 9.
Self-fertilization for eight generations in succession, of course
goes far in getting rid of heterozygotism in most characters. East
and Hayes (712) have given the general formula for determining
what proportion of the organisms would be homozygotic with
respect to any given number of characters after a given number
of self-fertilizations; this being based on the formula originally
given by Mendel (’66). In a recent note, written before the
paper of East and Hayes had appeared, I went into some details
on the matter (Jennings 712). If we call x the proportion of the
organisms that will be homozygotic, letting n be the number of
successive self-fertilizations and m the number of pairs of char-
acters, then the formula for use is
Gras
Dn
From this formula we find that after eight successive self-
fertilizations the proportion of the organisms that would be
homozygotic for any one, two, or more characters, up to ten, is
as follows:
NUMBER OF PROPORTION NUMBER OF PROPORTION
CHARACTERS HOMOZYGOTIC CHARACTERS HOMOZYGOTIC
Ih ad Rai Sees ees 0.99609 Poe eR oe 0.97679
7c et ee ee : 0.99220 Vag eel PE : 0.97297
Shoat Ou Aa a ae 0.98833 pis 3 ae ES aie ae 0.96917
Ae See 2S 0.98447 OR ae ee 0.96539
SE Oe es, 22m > ees 0.98062 LORS Fe ee 0.96162
5594 H. S. JENNINGS
Thus, after eight self-fertilizations, more than 96 per cent of
the organisms would be homozygotic with respect to all ten
characters.
Of course we do not know on how many independently heritable
characters depends the rate of fission (which was the character-
istic chiefly examined). If it depends on not more than 10 such
characters, the chances are thus at least 26 to one that we are
dealing with a pure homozygotic organism, when we select a
single individual after the eighth successive self-fertilization of
the line.
The above analysis is based on the view that there is no sepa-
ration of the zygotic constituents in the reproduction by fission,
this being indicated by the evidence thus far brought forward.
If it were not true, then we would expect the organisms consti-
tuting a pure strain (descended by fission from a single indi-
vidual) to become more and more diverse as fission was repeated,
for as any individual became homozygotic with respect to any
character it could produce forever after only progeny that were
homozygotic in that respect. The result would be in the course
of 20 or 30 generations to produce a set of individuals, each
of which was homozygotic with respect to all the characters it
bore, though the different ones would have diverse homozygotic
characters. Selection among such individuals would then give
rise readily to diverse races; this is opposed to the evidence
hitherto obtained.
The eighth of the conjugations in succession took place April 29,
1910. A single ex-conjugant gave rise to a tulture, which propa-
gated without admixture, till this experiment was begun, Decem-
ber 6, 1910. On the evening of December 5 a watch glass of the
animals was taken from the large culture; on the following
morning those in this watch glass were conjugating, while those
that remained in the large culture dish were not.
Experiment 13a. Fifty-two pairs were taken from the watch
glass, 100 non-conjugants from the culture dish; all these were
isolated on slides, in the way already described. Thirteen of the
pairs were later lost by an accident. This left 78 lines derived
from animals that had conjugated, 100 from animals that had
EFFECT OF CONJUGATION 333
not conjugated. Of the conjugants 20 died during the first week;
of the non-conjugants 18, leaving 58 and 82 respectively in the
two groups.
RaTE oF Fission. The ex-conjugants, as usual, began divid-
ing the second day after conjugation. Beginning for both sets
at this time, daily records were kept of the number of fissions in
each line. Table 20 gives the fissions for the first week, in each
of the two sets. As in all our other experiments, the rate of
fission was somewhat greater in those that have not conjugated.
TABLE 20
Experiment 18 a. Paramecium aurelia. Comparative number of fissions in con-
jugants and non-conjugants of the same culture, for a period of one week, begin-
ning two days after the separation of the pairs.
IE Oy Se | 5 MEAN STANDARD COEFFICIENT
7 B DEVIATION | OF VARIATION
pelea ay | 4s ab. | 6 5) Fe |"8 9 [hasty | |
Conju-
gant
lines | 4 1} 2] 8| 4] 6)12)14! 5] 2) 58) 6.155 = 0.220 | 2.484 = 0.106 | 40.850 + 2.910
Non- | | | |
conju-
gant : |
lines 2) 5/|16/33)15| 8| 3) 82] 7.098 + 0.093 | 1.246 += 0.066 | 17.550 = 0.953
| |
VaRIATION. As in other cases, the variation in the rate of
fission is much greater among the descendants of the conjugants
than among those of the non-conjugants. The standard devia-
tion is twice as great, and the coefficient of variation two-and-a-
half times as great, in the descendants of the conjugants (table 20).
Mortauity. Of the 78 conjugant lines, 20 died out during the
first week, or 25.6 per cent. Of the 100 non-conjugant lines,
18 died out, a mortality of 18 per cent. _
Experiment 13 b: inherited differentiations in the pure strain.
At the end of this first week, those lines of each set that showed
indications of differentiation in rate of fission were selected for
farther propagation. That is, from both the conjugant set and
the non-conjugant set the extreme lines were taken; also certain
of the intermediate ones. Thus, from the conjugant lines there
were selected the four that had not divided at all (table 20), the
334 H. S. JENNINGS
one that had divided but twice; two whose record had stood at 3
fissions, six at 4, two at 5; then two at 8, two at 9, and the two
at 10. In the non-conjugants, the two with a record of 4 were
taken, three at 5, three at 6, four at 7, two at 9, and two at 10.
In all, 21 of the conjugant lines and 16 of the non-conjugant lines
were thus continued. Two of the former (15 and 16) were, how-
ever, derived originally from one ex-conjugant.
The purpose of continuing these 37 lines was to determine
whether the varying rates of fission are inherited; this would show
that inherited differentiation had arisen within the pure line, in
this respect at least.
All but one of those conjugant lines which had not divided
during the first week died out during the second week. Sixteen
of the conjugant lines and fourteen of the ‘non-conjugant ones
were cultivated under identical conditions from December 6 to
February 27, a period of ten weeks; a few of these lines died out,
however, before the end of the period. It will be well to divide
the period into five homogeneous divisions of about two weeks
each, giving the fission rates for each line in each of these divisions.
During the first four periods the organisms were changed daily;
during the last one, every other day, a regimen under which they
did not.thrive. On this account, the first four periods are more
characteristic and significant than the last. The fissions for
these five periods are given in table 33 (Appendix). I have
arranged them in the order of their relative rates of fission, as
determined by comparing the total numbers of fissions in the
first three periods (given in the last column).
Examination of table 33 shows clearly that in some cases at
least the different rates of fission are inherited. Compare for
example among the conjugants, line 4 and line 15 (or 16). In
every one of the five periods, line 4 shows a higher rate of fission
than does line 15. The same thing appears in other lines, of
which details will be taken up later.
These constant differences appear in spite of the fact that all
of the lines were treated in exactly the same way throughout the
ten weeks’ experiment. All were kept together, in the same
moist chambers, and in the same culture fluid. In order that
the drop belonging to one line should not have a continuously
EFFECT OF CONJUGATION 335
different bacterial content from that of another, the animals of
different lines were frequently interchanged; line 1 being trans-
ferred into a drop in which line 2 has been living, and vice versa.
The drops were for the same purpose frequently intermixed.
If two such lines as No. 4 and No. 16 (of the conjugants) showed
in the long run about the same rate of fission, but with accidental
fluctuations from period to period, then of course in some periods
No. 4 would show a greater number than No. 16, while sometimes
the reverse would occur. When we find however that such a
line as No. 4 has uniformly a greater number of fissions than
another, and this continues for so long a time as ten weeks, with
no external differences to cause these results, we must conclude
that the lines themselves are differentiated.
We may make then as the test of inherited differentiation the
condition that one line shall show in every one of the five periods
of table 33 a distinctly higher fission rate than another. This is
an extremely severe test, and one that is beyond question more
than sufficient to show actual inherited differentiation. In the
slow process of experimentation these repeated differences are
most striking and surprising. Our first period covers eighteen
days; during this time one finds that conjugant No. 1 divides
more rapidly than No. 8 or No. 16. To test the matter, the three
are kept under identical conditions for twelve days longer (second
period). Again No. 1 shows the highest rate, No. 8 a lower one,
No. 16 a still lower one. To make assurance doubly sure, we
keep them fourteen days more (third period); again they show
the same relative rates. We keep them a fourth period of twelve
days; a fifth one of fourteen days; these confirm the differences
shown in the first three periods. There can be no question but
that the cause of the diversities is in the lines themselves; in
' other words there is differentiation inherited from generation to
generation.
On this basis it is clear that among the conjugants, Nos. 1,
8 and 16 represent three lines with inherited differentiation in
rate of fission. It is hardly doubtful but that other differentiated
lines exist, accidental fluctuations bringing these equal to one
336 H. S. JENNINGS
of the above three at one of the five periods. But we may hold
rigidly to our test and still demonstrate the existence among the
conjugants of the three diverse lines 1, 8 and 16. As will be
observed, the fission rate is on the average more than twice as
great in No. 1 asin No. 16.
Among the non-conjugants also there are inherited differentia-
tions. In every period but one, non-conjugant line 1 has twice
as high a fission rate as line 14. On the basis of our severest test,
it is clear that lines 1, 12 and 14 are diverse in their inherited
rate of fission.
It is clear therefore that heritable differentiations do arise
within the pure line, so far as the rate of fission is concerned.
How are these differentiations brought about?
At this point a weak spot in the plan of the present experiment
appears. All our experiments show that conjugation increases
the variability in the rate of fission; this is true both in wild cul-
tures and in pure lines, and holds for the present experiment, as
table 20 shows. It would appear probable therefore that some
of these variations are inherited, and that this is precisely what
the results given in table 33 demonstrate. But we find inherited
differentiations also, as we have seen, among what we have
called the non-conjugant lines of the present experiment. The
weak point mentioned relates to the applicability of the term
‘non-conjugant’ to these lines. As already set forth, the last
previous recorded conjugation took place for this line k on
April 29. The present experiment began December 6. Now,
it is almost certain that in the intervening time the animals had
conjugated one or more times, since this race k conjugates once in
one or two months, when conditions are favorable. Therefore,
if conjugation produces differentiations, my ‘non-conjugants’ of
the present experiment have had much opportunity to become
differentiated in that manner; they are not properly ‘non-conju-
gants’ for present purposes. That they have become in some way
differentiated is clearly shown by comparison of No. 1 with No. 14
in the non-conjugants of table 33.
This, of course, does not vitiate our main result, that inherited
differentiation does arise within a pure line, and it leaves it
EFFECT OF CONJUGATION S30
probable, or perhaps certain, that such differentiation arises in
consequence of conjugation. But it leaves unsettled the ques-
tion whether such inherited differentiations may not arise also
in other ways.
To give clear results on this point, the experiment should have
been performed as follows: A single individual of & should have
been isolated, allowed to multiply by fission; watched continu-
ously till the first conjugation occurred, then the experiment
should have been performed with these conjugants and non-
conjugants. If inherited differentiation appeared among _ the-
non-conjugants in such a case it could not be held to be due to
conjugation.
These conditions are fulfilled in Experiment 15 on another
race, to be described. They were likewise fulfilled in the latter
part of the present experiment, and repeatedly in experiments on
race k in 1912. But unfortunately race k has lost its power to
flourish in slide cultures; in every case with the later experiments
on this race all the lines have died out after a few weeks of cul-
ture. It would-be of interest to carry out the experiments with
race k, in view of its history of eight repeated self-fertilizations,
and efforts will be made to find a successful method of slide cul-
ture for it. In the meantime the results of Experiment 15, with
race HE, give clear results on the main questions at issue.
_ The results of the present experiment therefore leave open the
possibility that heritable differentiations may arise in other ways
than by conjugation. Do they furnish positive evidence that
heritable differentiation actually does arise as a result of conju-
gation? As we have seen, all our many experiments show that
conjugation increases the variation in rate of fission between the
lines. This is true (as already set forth) for the first week of the
present experiment. Furthermore, if we compare the variability
of the conjugant and non-conjugant lines of table 33, we find again
that the conjugants are much more variable. We are of course
not here dealing with random samples, but since both sets were
selected to give as much variation as possible, a comparison of the
variations may be of significance. The means, standard devia-
tions and coefficients of variation for various periods are given
338 H. S. JENNINGS
for the conjugant and non-conjugant lines, in table 21. The
constants are given, not only for each of the five periods of table
33, but also for certain of these periods taken together; likewise
for the first week (column 1).
Table 21 shows that: (1) in every case the mean rate is higher
in the non-conjugants; (2) in every case the standard deviation
(measure of the absolute amount of variation) is greater in the
conjugants; (3) in every case the coefficient of variation (measure
of the variation relative to the mean) is much greater in the
conjugants.
Since these measures are based on the number of fitssions for
long periods under identical conditions, they can hardly be held
to represent meaningless accidental fluctuations, but rather actual
differentiations. They show further that these differentiations
are much greater in those that conjugated during the last epidemic
than in those that did not. This conjugation therefore caused
inherited differentiations within the pure line. Whether the
fewer inherited differentiations among those that did not con-
jugate during the last epidemic are due to previous conjugations
we cannot tell in this case, but must refer the reader to the account
of Experiment 15.
The inheritance of the rate of fission in these cases may be
demonstrated, for those that prefer this method, by working out
the coefficients of correlation. The numbers we are dealing with
are of eourse small, but significant, owing to the great number of
generations dealt with. We may take the fissions during the
first two periods of table 33 and by determining their correlations
with the fissions in the same lines for the second two periods, get
a numerical expression of the inheritance. For the conjugant
lines we find that the coefficient of correlation thus taken is
0.5031 + 0.13846. For the non-conjugant lines it is 0.5627 =
0.1331.
(A full treatment of the inheritance of the fission rate, by
biometric methods, with adequate numbers, will be given in
another connection; together with an analysis of the relation of
this method of measuring inheritance to other ways of dealing
with the matter.)
339
EFFECT OF CONJUGATION
OST GFS8Zh OT
GIL OFGOL Te
G18 OF6LE'9
618 OF ELE 2
OPI T98L 88
£hS TOL8 Sé
€ OL T
sao1udd
918° 3960 0%
990 PLP 66
$69 OF TPO'S
962 OF LEG 9
286 OF ERO G6
NBT TPIS 16
f INV ¢
saoludd
OG8 Toss FI
$66 GFS86L £6
89 OF 899° £
9b9 OF FIPS
199 OF 98L 92
£16 O02 8%
Z ONV [
saouimd
68h F668 8
GbE OL Ohh £9
68 OF 90L'°S
689 OFZI6'E
099 OF $98 6
092 OF LOT 9
Lo-ET
AUVOUHHA
aomtd
Hoda
9
128 GEL OT
8Lb S618 LE
OLE OF E96 °%
&LP OF LOLS
68h OF LIP IT
899 OF 98L'6
Zl AUVOURGA
-0¢ AUVONVS
aormmd
HOMO
g
998 GL 8
OOP POLE EE
Ob OF 6G9'E
Shh OF EELS
£99 0298 GT
669 OF 881 IT
60-81
AUVONVE
aqoad
aun
b
699 LES ST
690° 3828 9T
991 OF 108 T
981 OF 199 T
GES OF EPL OT
£92 0092 '6
PI-Z
AUVONVE
aod
anooas
£
GOb GES ST
916 GF E68 Eb
LOE OF E88
oh OF 199
069 OF 98L ST
806 OF LP9 GI
96-8
ugqanaoad
aod
Isuta
z
£96 0099 ZT
O16 G0SE OF
990 OF OFS 'T
OST OF P8b'S
£600 860°2
006 OF GST 9
SOOM Lula
if
fuop-toN
ee eeeeenee tmOQ BA “Je0O
Jive mt 88s fEOD-UON
nae ames (uog—aeq ‘pung
Cede nneeneeeneaen fuop-u0N
Pichraeor acl We aphre htc ca ‘fuop—uBoW
n
‘(wypuaddy ay) ur ‘eg argv) fo ppp
AY} UO LapuYoUads AY) L0G 2790] fo DIDP AY) WO pasDg au UWNIOI I8LYf BY) UL SJUD)SUOD AY.)
-uou fo puv sjunbnluoa fo syuppuaosap sof ayos uorssif ur fpyrqnrawa aaynundwmoy
1 UTUVL
‘sporwad snoripa sof ‘syupbnluoo
“‘DYIIND WNVIUWDAD J
Q ST quauiisaday
340 ‘ H. S. JENNINGS
Experiment 13 c. In order to test more fully the inheritance
of the differences in fission rate shown in table 33, certain lines
were next selected for propagation on a more extensive scale.
Beginning January 28, 1911, the attempt was made to propagate
16 parallel lines each of conjugant numbers 1, 2, 3, 11, 14 and 16
(of table 33), the purpose being (1) to determine whether the
results with 16 lines of a given number confirm those obtained
with but one line; (2) to discover whether there arise differentia-
tions within any of the series derived from a single individual.
This second point, as we have before seen, is fundamental for a
full understanding of the results thus far reached.
These objects were not fully attained, owing to the cessation
of active propagation on slides in the race k, but certain results
of importance were reached.
CoNJUGANT LINES. The sets derived from the different ex-
conjugants of table 33 showed great differences in vitality as
well as in rate of fission. Lines 1, 2, 3 and 11 began strongly,
16 parallel sets being derived from the original single set in one
to three days. With conjugant lines 14 and 16, on the other
hand, there was great difficulty in getting 16 sets established;
multiplication was extremely slow, and many of the sets died out
almost as soon as they were isolated. It was a week from the
beginning of the experiment before 16 sets were in operation in
conjugant lines 14 and 16.
The relative rates of fission that had characterized the various
lines from the beginning continued to show themselves in the
sets of 16 from each line. The slower lines showed much greater
mortality than the faster ones. As fast as any set of a given line
died out it was replaced from another set of that line. The
number of deaths for each line was thus recorded. It will be
instructive to give for each of these lines of ex-conjugants the
number of fissions and the number of deaths, up to ites 15.
This is done in table 22.
The mortality in the slow lines increased from February 15 on,
so that on February 16 there were but three sets left (out of 16)
in line 14. By February 26 all the 16 sets of lines 14 and 16 were
dead, so that these two lines became extinct. In the meantime,
EFFECT OF CONJUGATION 341
TABLE 22
Experiment 13 c. Paramecium aurelia. Number of fissions, and number of deaths,
in each of the sets of 16 parallel cultures belonging to six of the conjugant lines
of table 33, between January 28 and February 15. For each line, the minimum
and maximum number of fissions in the cultures that lived through the period are
given. Thus, in line 1, one of the 16 sets gave 15 fissions, another 20.
LINE
| 1 ie AP: 11 14 16
Fissions, January 28 to February 15.... | 15-20) 9-15 | 15-16} 10-13) 5-6 | 3-11
|
INumperomdeaths:--& 4... sts sesdensiecie | 1 4 0 8 | 19 20
the 16 sets of line 1 were flourishing under precisely the same
treatment.
Later, lines 2 and 3 began also to die out. On March 7, line 2
was extinct, while of line 3 one set still existed. Line | continued
to flourish; 16 sets still existed March 7. For other purposes 64
sets of line 1 were kept in propagation till April 17.
‘NON-CONJUGANT’ LINES. Later, the rapid line 1 and the slow
line 14 of the ‘non-conjugants’ (of table 33) were compared
similarly, beginning February 26. Sixteen sets of each were put
in progress. Line 1 continued to multiply rapidly, line 14 slowly;
by March 20 the maximum number of fissions in the former was
17, in the latter 6. At this date an attempt was made to increase
the numbers to 64 parallel sets for each line. But it was found
impossible to get 64 sets of line 14 into existence, owing to the
great number of deaths. On March 31 the last set died out, and
the non-conjugant line 14*became extinct. At this time there
were 64 sets of line 1, which were continued till April 17, when the
experiment was abandoned. During the last two weeks there
were few fissions even in line 1.
Thus all of the lines having a slow rate of fission died out, even
though the attempt was made to keep up 16 parallel sets; and
this under conditions in which the lines with rapid fission con-
tinued to flourish. Although all were treated alike, only the two
most rapid lines, No. 1 of the conjugants and No. 1 of the non-
conjugants, continued to live till the close of the experiment,
April 17.
342 H. S. JENNINGS
This appears to indicate that the lines with slower fission are
defective in some way. Of course it is possible, perhaps probable,
that under more natural conditions they would have continued
to exist, in spite of their slow multiplication. The extremely
slow line 16 (conjugant) had lived from December 6 to Feb-
ruary 26, a period of two months and twenty days, comprising
forty successive generations. But slow multiplication and high
mortality are decidedly correlated.
It had been planned to employ the 64 sets that were kept for a
number of different lines in biometrical studies of the inheritance
of the fission rate; and in an attempt to determine whether
heritable differentiations in fission rate arise in the progeny of a
single individual multiplying by fission. But the death of all
the slow lines, and the extremely slow multiplication of the others
for the last weeks of the experiment rendered the extensive data
obtained valueless.
Summary of Experiment 13. We may summarize the results
of this entire experiment as follows:
In a pure strain, all the individuals derived originally from a
single one; and all derived from eight successive conjugations with
self-fertilization of the strain:
1. Conjugation decreased the rate of fission.
2. Conjugation increased greatly the variability in rate of
fission.
3. The differences in rate of fission were found to be inherited,
so that in this respect heritable differentiations arise within the
pure strain. ;
4. These heritable differentiations are due partly, if not
entirely, to conjugation, since the latter increases greatly the
variability. But whether such heritable differentiation may
arise within the pure strain by other means is not determined in
this experiment.
5. A low fission rate is correlated with a high mortality.
Conjugation produces many lines with low fission rate; these
lines die out in the course of time, if the conditions become
severe, although the lines with rapid fission continue to live.
But the slow lines may live for many generations (forty in this
experiment).
EFFECT OF CONJUGATION 343
Experiment 14: Paramecium aurelva
This was a direct continuation of the foregoing, dealing with
the rapidly multiplying line 1 of the non-conjugants. On March
20, 1911, a considerable number of these were placed in a watch
glass; on March 22, conjugating pairs were found among these.
Of these 48 pairs were isolated, making after separation 96 lines,
which were cultivated on slides as usual. At the same time there
were in progress 64 lines of those that had not conjugated, and
to these were now added 48 more. Thus we have now propagat-
ing, under identical conditions, 96 lines of ex-conjugants and 144
lines of non-conjugants, all derived from individual No. 1 of the
non-conjugants of the previous experiment.
Conditions were unfavorable for multiplication, the tempera-
ture being low and the university buildings not heated. Of the
96 lines of conjugants, all but four died without dividing; that
is, 95.8 per cent. Of the 48 non-conjugants set at the same time,
27, or 56.25 per cent died without dividing.
By March 31, nine days after conjugation, all but 4 of the 96
conjugant lines were dead, while 27 of the 48 non-conjugant lines
were dead. Of the entire 144 non-conjugant lines, kept under the
same conditions as the conjugant ones, 37 had died during the
same period. Thus the proportion of deaths was, for the con-
jugants, 95.83 per cent; for the non-conjugants 25.69 per cent.
This experiment shows that under such unfavorable conditions
the animals that have recently conjugated are much less resistant
than those of the same descent that have not recently conjugated.
Experiment 15: inherited differentiation produced by conjugation:
Paramecium caudatum
In the summer of 1912, after several months spent in vain
attempts to repeat with the race k the essential features of
Experiment 13, under such conditions as would show beyond
question whether all the inherited differentiations were due to
conjugation or not, a successful experiment for this purpose was
carried through with a race of Paramecium caudatum which I
called E.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
344 H. S. JENNINGS
The race E was derived from a single individual taken July 31
- from a wild culture of Paramecium caudatum. This individual
and its progeny were allowed to multiply on slides till a large
number were obtained. On August 19 many of these were
transferred to a mass culture, and on August 22 watch glass
cultures containing many individuals were removed from this
mass culture. Early the following morning conjugation was
beginning in these watch glasses. In the way set forth in our
general account of methods (page 282), I picked out 67 pairs and
68 split pairs (pairs which had begun to unite, but which were
separated before conjugation was consummated). The two
members of each pair (and of each split pair) were designated
a and 6. The products of the first division of each of these
were retained, becoming the progenitors of two lines which I
called x and y. Thus from each pair (and each split pair), four
lines were propagated, two from a and two from b. This of
course gave 268 lines derived from conjugants and 272 derived
from the non-conjugants of the split pairs.
During the heat of summer the cultivation of many lines of
Paramecium is very difficult, owing to excessively rapid develop-
ment of bacteria in the drop cultures. This has the effect of
inducing a high mortality, and also of making it very difficult to
keep the environmental conditions uniform throughout a large
number of lines. This latter condition is essential in the present
experiment, since if it is not fulfilled, differentiations in fission
rate due to environmental conditions simulate those due to
heritable or intrinsic differences in the diverse lines.
Owing to these difficulties the mortality among the conjugants
was high, and the measures required for making the conditions
uniform were so time-consuming that I was compelled to abandon
a large number of the lines of propagation of the non-conjugants,
so that I succeeded in keeping to the end of the experiment but
88 lines of conjugants, derived from 44 original ex-conjugants,
and 174 lines of the non-conjugants, derived from 87 original
members of split pairs. These however were sufficient for the
solution of the problem that gave rise to the experiment. These
262 lines were propagated from August 24 to September 16
EFFECT OF CONJUGATION 345
(non-conjugants) or September 18 (conjugants), a period of 24
(or 26) days.
Culture methods necessary to secure uniformity of conditions.
In order that the conditions should be uniform throughout the
large number of lines, the following method of culture was found
necessary :
As a culture medium, Horlick’s malted milk was employed,
following the example of Miss Peebles (12), one-sixteenth of 1
per cent being found the most favorable proportion. This was
made fresh each day, with boiling water. The animals were
changed every other day. The chief difficulty in making the
conditions uniform throughout all the lines is as follows: A
number of diverse bacteria are found in the cultures, falling into
them at the time of changing, or reaching them in other ways.
Some of these multiply strongly in certain of the slide cultures
while others get a better foothold in others. The effect of the
diverse bacteria on the rate of reproduction differs greatly; as a
result therefore some of our lines of Paramecium multiply rapidly,
others slowly, even though there is no intrinsic differentiation
among them. Now, in transferring with a capillary pipette a
single individual to a new drop, as is done at the time of changing
the animals, inevitably a certain amount of the bacterial culture
is transferred with them, serving to infect the new drop. Thus
one line will be accompanied always by the bacterium z, causing
rapid multiplication; another by the bacterium y, causing slow
multiplication. The results simulate those of inherited differen-
tiation in the fission rate.
Experience showed that this difficulty is obviated by the fol-
lowing method of procedure, which was adopted for the present
experiment on August 28:
The new fluid (1/16 per cent malted milk) was made sterile
by boiling. It was then infected with bacteria from a mass
culture of the race #, in which the animals were flourishing
strongly. This was done by filtering (through two thicknesses
of filter paper) a quantity of the fluid from this culture (in order
to remove the Paramecia). I added four pipettes full of this
filtered fluid to 100 cc. of the fresh culture fluid.
346 H. Ss. JENNINGS
For changing the animals to this, two fresh slides are prepared,
each containing three drops of this fresh fluid. A vessel of boil-
ing water is at hand; also a supplementary vessel of the fresh
culture fluid. The capillary pipette is first dipped in boiling
water, then into the fresh culture fluid, then a single individual
is removed with it from the old slide to the first new slide. The
pipette is then again disinfected in boiling water and washed in
the supplementary dish of culture fluid. Meanwhile, the removed
Paramecium has been swimming about violently in the three
drops of fresh culture fluid, thus washing itself largely free from
the bacteria introduced with it. Now, with the cleaned pipette,
it is retransferred from this wash water to the second slide of
fresh fluid. (In much of my work I gave each animal a second
washing in the same way.)
A‘new ‘wash slide’ is then prepared, the pipette is disinfected
and washed as before, and we proceed to transfer in the same
way an individual from the second slide to the wash water and
then to its definitive slide. After every transfer the pipette must
be disinfected and washed, and new wash water must be used for
every individual transferred.
Experience shows that all the details of this painful process
are quite necessary if the conditions are to be kept uniform in a
large number of lines. Carrying this out for some 250 lines for
nearly a month I found so exhausting as to make it practically
impossible to continue the experiment for a longer period.
Records. The records of the conjugant and non-conjugant
lines for this experiment, conducted in the manner just described,
are given in tables 34 and 35 (Appendix). The results of this
experiment are of so fundamental an importance for the subject
with which the present series of papers deals that I feel it neces-
sary to give the records in detail, showing the number of fissions
that occurred in each period of two days. These records will be
used farther in studies on inheritance, to follow the present
paper.
Explanation of tables 34 and 35. These tables give, for the conju-
gants and non-conjugants, respectively, of the pure strain HZ, the records
of fissions for each line for the entire period (twenty-four days for the
EFFECT OF CONJUGATION 347
conjugants; twenty-one days for the non-conjugants). The records
given are the numbers of fissions that have occurred during the two days
ending on the date at the head of the column. (In only one case, for
the non-conjugants, in the column headed November 6, is the elapsed
period three days instead of two.)
Each pair or split pair consisted of the two mated individuals a and b.
From each a and 6 the two sister lines x and y were kept in progress.
Thus from each pair there were derived four lines, az, ay, bx and by.
But the lines from both a and 6 were kept throughout the experiment in
but few cases (16 in the pairs, 22 in the split pairs).
The lines from a and b were kept in separate moist chambers and
changed at different times, so that there is no opportunity for resem-
blance between them to arise through special similarity of treatment.
The two lines x and y, from a single individual, were however kept in the
two concavities of the same slide, in the same moist chamber, and
changed in succession. (This was for convenience in replacing one
from the other, but in repeating such an experiment, x and y should be
kept in separate moist chambers and handled separately; otherwise the
significance of any correlation between x and y is not entirely clear.)
In working out constants of variation, the period August 27 to Sep-
tember 6 (twelve days) was considered the ‘first half’ for the conjugants;
August 28 to September 6 (eleven days) for the non-conjugants. The
second half for the conjugants included twelve days (September 8 to 18);
for the non-conjugants, ten days (September 8 to 16).
The blanks left in the column under certain dates indicate that the
line in question died out on that date, and its place was supplied by
taking an individual from the sister line x or y, derived from the same
parent, that is, from the same a (or b, as the case may be). Buta blank
in the final column of totals indicates only that the line in question did
not live independently throughout the experiment, but was supplied
from its sister line at some date, indicated as just set forth.
In determining mean, standard deviation or coefficient of correlation
for any period or periods, only lines that lived independently throughout
that period are included. However, for the entire period, the few totals
included in parenthesis, in the last column, are employed also, since the
lines for which they stand coincided with another for only two or three
fissions at the beginning.
In working out coefficients of correlation, for successive periods, it is
of course necessary to correlate any line with its real ancestral line, and
to do this it is necessary to pay careful attention to the blanks left in
certain columns.and the replacement of certain lines which they indicate.
Thus, if in table 34 we wish to correlate the fissions in the first half of
the entire period with those in the second half, then when, for example,
we enter the fissions for the second half of the time (September 8 to 18)
in line 8 by, we see that this second half descended partly from 8 bz;
there is no difficulty, however, in determining exactly how many fissions
occurred in the first half. We take in this case for the first half of the
period the sum of the fissions for 8 bx to September 2, plus those for 8 by
348 H. S. JENNINGS
for September 4 and 6 (that is,0 + 2+2+1+3+3 = 11); for the
second half, 8 by, September 8 to 16 (=17); and similarly for all analo-
gous cases. Thus in the correlation tables a given preceding period
may sometimes be counted twice, since it gives rise to two lines of prog-
eny, and is therefore correlated with both. This of course introduces
no error into the coefficient of correlation. The other constants (given
in table 23) were computed with each period counted but once.
Results. Comparative examination of tables 34 and 35 shows
a very great difference between the progeny of the pairs, and
those of the split pairs, in respect to variation and differentiation
in the rate of fission. ‘This is well shown by observing the range
of variation in the two cases. In the split pairs the slowest lines
show in twenty-one days 18 fissions, the fastest, 28 fissions.
In the pairs the range is (for twenty-four days), from 10 to 35.
If we reduce these latter numbers by one-eighth, in order that
thay may compare directly with those for the split pairs, the
range becomes for the pairs 8.75 to 30.6, as compared with 18
to 28 for the split pairs. From September 8 to 16 the range for
those derived from the pairs is 1 to 17; for those from the split
pairs, 8 to 15.
Working out the mean number of fissions, with the standard
deviation and the coefficient of variation for the pairs and split
pairs during a number of different periods, we obtain the results
shown in table 23. Here the data are given for the first and
second halves of the experiment; also for the entire period. In
order to have certain periods which are absolutely identical in
every respect for the two sets, I give also for the pairs the data
for the ten days extending from September 6 to September 16,
this period coinciding with the ‘second half’ for split pairs.
As the last columns of table 23 show, the variability in fission
of the lines descended from conjugants was, for the entire period,
four times as great as that for those descended from non-conju-
gants. In the first half of the time it was about twice as great;
in the second half five times as great.
Examining tables 34 and 35 to discover the cause of this great
difference in variation, we find that the descendants of those that
have conjugated are differentiated into a number of distinct
lines, with different rates of fission. This will at once be evident
EFFECT OF CONJUGATION ; 349
if one compares, among the pairs (table 34) line 1 with line 6, and
the latter again with line 4. Line 1 az shows a total during the
twenty-four days of 33 fissions, line 6 ax of 16 fissions; line 4 ax
of 13 fissions. Corresponding differences are shown in the other
divisions of lines 1, 6 and 4, the differences extending even to
lines descended from the two mates of a pair. Thus line 1 az has 33
fissions; line 1 bx, descended from its mate, 31 fissions; line 6 az
has 16 fissions; line 6 by, descended from its mate, 17 fissions;
line 4 ax has 13 fissions; line 4 bx, descended from its mate, has 12.
If we compare similarly the two lines x and y, derived from a single
member of a pair, we find that their fission rates are close together,
while lines derived from different pairs differ greatly.
Certain peculiarities of the fission rate are evident. During
the first five or six days after the beginning of fission the different
lines descended from the conjugants are more nearly uniform in
their rate. Then a number of the lines, such as those belonging
to pairs 4, 5, 15 and 26, show a marked decrease in the fission
TABLE 23
Experiment 15; Pure strain E. Constants of variation in fission, for the lines
descended from conjugants (pairs), and for those descended from non-conjugants
(split pairs), for certain periods of time. The total time is, for the pairs, 24
days, for the split pairs, 21 days. The first half includes for the pairs the twelve
days, August 27 to September 6; for the split pairs eleven days, August 28 to Sep- °
tember 6; the second half, September 8 to 18 (pairs), September 8 to 16 Es ie
piper ve MEAN NUMBER OF FISSIONS DAILY RATE
| Pairs ere Pairs Split pairs Pairs puns
Motaletime-...%)..2: 5: 5.2 69 145 | 26.333 + 0.613) 24.034 = 0.096) 1.097] 1.144
Wrstehalics cases 04 she 78 158 |} 12.154 + 0.081) 11.424 + 0.072) 1.013} 1.039
Second half......2.%....-||-8o 171 | 14.241 + 0.444) 12.614 + 0.056) 1.187] 1.261
September 8-16........| 83 171 | 12.060 + 0.367| 12.614 + 0.056) 1.206) 1.261
STANDARD DEVIATIONS | COEFFICIENT OF VARIATION .
Pairs Split pairs | Pairs | = pairs
otal time. ..+:-. 7.544 + 0.433 | 1.712 = 0.063 | 28.650 = 1.734) 7.122 = 0.284
Binstehalf <<. sss: 2.370 = 0.128 | 1.347 + 0.051 | 19.501 + 1.065) i 789 = 0.453
Second half...... 6.003 = 0.314} 1.083 = 0.039 42.154 + 2.885) 8.585 = 0.315
September 8-16..| 4.954 + 0.259:) 1.083 = 0.039 | 41.074 + 2.485) 8.585 = 0.315
350 H. S. JENNINGS
rate, which persists throughout the remainder of the experiment.
In other lines, such as those derived from pairs 1, 2, 3, 7, 12, the
rate remains high throughout the entire period.
The inherited differences between the lines will perhaps be
best brought out if we divide the twenty-four days of the experi-
ment into four periods of six days each, and give the number of
fissions for each of these periods, for a number of diverse lines.
We shall re-group these lines in such a way as to bring out strongly
the diversities. The results are shown in table 24. It is observ-
able, for example, that in every one of these four periods the line
3 ax has a greater rate of fission than 6 az; similarly 3 ay and
3 bx show in every period a greater rate than 6 ay and 6 bz.
Comparison of other lines shows the same relations.
TABLE 24
Experiment 15. Pure strain E. Fissions in certain of the conjugant lines, for four
successive periods of six days each, so arranged as to exhibit the differences between
the lines.
LINE AuGust 27-31 SEPTEMBER 2-6 SEPTEMBER 8-12 SEPTEMBER 14-18
A Se Se 6 | 10 9 9
Gute ees 3 3 5 5
IK AL ete 7 4 1 1
2S, ae ee | 6 | 9 9 10
Giday 7 ee Se 3 3 5 5
Te ye (7) 4 1 0
5 | 0): ia en ee 6 9 9 8
Gila. crake few 3 5 (5)
71) >, Ce ar Le 6 5 1 0
11) te Ae 6 7 10 9
WIDER aes 4 5 1 0
11 oop Se Raa | 4 9 10 8
BEAD Beaks « ocntne x | 7 5 | 2 1
|
ok | 6 8 11 9
7.24) 5 5 5 6
SOE, eS... | 6 5 | 3 1
EFFECT OF CONJUGATION 351
In some eases a line begins with a high rate of fission, then runs
down to a very low one, dividing but once or not at all during
the last six-day period. Such is the case, for example, in the
lines derived from pair 4. In such eases careful and extended
tests were made to determine whether the slow fission rate was
characteristic of all the members of the given line. Thus, of
line 4 ax, seven sets; of 4 ay, nine, and of 4 bz, eight sets; were
kept in progress during the last twelve days of the experi-
ment; all showed the same extremely low rate of fission charac-
teristic for the lines derived from pair 4 in tables 34 and 24.
In the same way, eleven sets of No. 5a, eight of No. 56,
twenty sets of No. 656, six sets of No. 26a, and eight sets of
No. 15 6 were kept in progress during the last twelve days of the
experiment; all of them showed slow rates of fission corresponding
closely to those given for the lines in question in table 34.
Of the rapid lines, No. 9a was tested by keeping eighteen
parallel lines in progress during the last twelve days of the experi-
ment. All divided rapidly, giving 18 to 20 fissions during the
twelve days.
It is thus clear that the lines descended from the ex-conjugants
are differentiated in their inherited characteristics, some having a
rapid rate of fission, some a slow rate, and some an intermediate
one (although all were kept under absolutely identical conditions).
One result of this inherited differentiation is the production of the
very high coefficients of variation shown in table 23.
Are there likewise inherited differentiations among the lines
derived from the non-conjugants—the members of the split pairs?
Examination of the coefficients of variation in table 23, as
well as a general inspection of table 35, shows at once that if
there is any such differentiation, it is very slight compared with
that among the descendants of the conjugants. If we compare
very carefully the records of the different lines in table 35, we
find a few cases in which it is doubtful whether there may not be
inherited differentiation. Line 3 az, for example, shows a rate
of fission somewhat below that of most others, while 6 by, 22 ax
40 ay and 44 ax show rates rather above the average. But the
differences between even the extreme cases are very small com-
352 H. S. JENNINGS
pared with those between the diverse lines derived from the
conjugants. Furthermore, taking the most extreme case of
tine 3 az, with but 18 fissions, we find that the sister line, 3 ay,
derived from the same parent, does not show a low rate of fission;
so that the slow rate is not characteristic of this entire line. In
the conjugant lines, on the other hand, the rates of the two or
more sets derived from a single individual we found to correspond
closely, showing that the characteristic is an inherited one. It
would then appear on the whole probable that all the differences
seen among the lines derived from the non-conjugants are simply
the slight fluctuations unavoidable where a large number of lines
are cultivated.
The question may be tested for both sets in another way. If
the differences between different lines are matters of inherited
differentiation, then of course lines having a fast or a slow rate
in one part of the period of the experiment should have a cor-
responding rate in the other parts. That is, the rates of fission
for earlier and later periods should be correlated. We may there-
fore determine the coefficients of correlation for successive
periods, in both the conjugants and non-conjugants; this will tell
us whether the rates of fission are, as a rule, inherited in the
different lines.
I have worked out for both sets the correlation between the
numbers of fissions in each line (1) in the first half of the experi-
ment compared with the second half; (2) in the second third
(September 3 to 8 or 4 to 10) compared with the last third (Sep-
tember 10 to 16 or 12 to 18). This latter comparison was made
owing to the fact that the direct physiological effect of conjuga-
tion appears to obscure the characteristic differentiations, for
some days after conjugation.
Furthermore, I have worked out, for the entire period of the
experiment, the correlation between the sister lines, x and y,
derived originally from a single member (of a pair or split pair).
If the differences in rate of fission are inherited, these two sister
lines should of course be similar, giving a positive coefficient of
correlation. The correlation, for both conjugants and non-
conjugants, is given in table 25.
EFFECT OF CONJUGATION 30a
As table 25 shows, the lines derived from conjugants give an
extremely high correlation. In other words the fact that we are
here dealing with lines differentiated in inherited characters is~°
demonstrated by this method as well as by the other evidence.
After passing the disturbance due to the direct physiological
effect of conjugation, the correlation between successive periods
rises to 0.8957 (practically to 0.9), an extraordinarily high coef-
ficient. The correlation between the sister lines x and y is like-
wise 0.9, showing an almost perfect correspondence.
In the lines descended from the non-conjugants, on the other
hand, there is no correlation between the numbers of fissions in
successive periods, the coefficients being practically 0. That is,
so far as this method can show, the diversities in fisston rate are
not inherited, among the members of a pure race which have not
conjugated.
On the other hand, the numbers of fissions for the sister
lines x and y, do give a small coefficient of correlation (0.2119).
TABLE 25!
Experiment 15. Pure strain E. Coefficients of correlation in number of fissions,
for successive periods, and for the sister lines x and y, in the descendants of
conjugants (pairs) and of non-conjugants (split pairs).
PAIRS SPLIT PAIRS
Num- Num- | ’
ber of | Coefficient ber of Coefficient
lines | lines
ey =
First half (10-12 days) with second
(ICAI CEN) aeenoueneasoo nena ct 82 | 0.5743 = .050 | 174 .0120 = .051
Second third (6-8 days) with last
HOIRGl: 6 Le aees eRe ee tenes 81 | 0.8957 + .015 | 172 |—.0020 = .051
MROtAlmuIMe sco Wilt lay sees a ene 50 | 0.9017 = .018 | 58 .2119 = .085
1For the pairs the ‘first half? comprises the 12 days August 27-September 6
of table 34; the ‘second half’ the remaining 12 days. For the split pairs the
‘first half’ is 11 days, August 28-September 6; the ‘second half’ is the remain-
ing 10 days. The ‘second third’ comprises September 4-10 (pairs) or September
3-8 (split pairs); the ‘last-third,’ September 12-18 (pairs), or 10-16 (split pairs).
The total time is, for the pairs, 24 days; for the split pairs, 21 days. The cor-
relation tables for successive periods are formed by taking the number of
fissions of a given line in an early period and entering this on the table in
connection with the number of fissions for the same line in the later period.
354 H. S. JENNINGS
Whether any significance is to be attached to this is doubtful,
since the value of the coefficient is but two-and-a-half times its
probable error; and a coefficient of this amount would occur once
in ten times as a result of chance distribution. Further, the two
sister lines x and y were kept in the two concavities of the same
slide, and one was changed immediately after the other. The
result of this may have been to keep the two under slightly more
uniform conditions than prevails for two individuals in different
moist chambers, giving rise to the slight correlation. The matter
will be investigated farther, but in any case it is clear that any
differentiation that may exist between the non-conjugant lines
is extremely slight; so that correlating the fissions of successive
periods gives no trace of it.
The present experiment therefore clears up the difficulty left
by the results of Experiment 13. In that experiment, as shown
in table 33, the ‘non-conjugants’ exhibited inherited differentia-
tions, as did the conjugants. It seemed practically certain how-
ever that these ‘non-conjugants’ had gone through previous
conjugations, so that the observed heritable differentiations were
probably due to these previous conjugations. On page 337,
I pointed out the necessity for an experiment in which this matter
should be controlled. Our present experiment supplies this need;
we know that our non-conjugants here have not conjugated since
they came from a single parent individual. And our results show
that the inherited differentiations in Experiment 13 were indeed
due to conjugation; they do not appear when we deal with actual
non-conjugants (lines which have not conjugated since they were
all derived from a single individual).
Even if it should turn out that the slight correlation shown by
x and y in the non-conjugants of the present experiment is due to
real differentiations between the lines, this result would not modify
our present conclusion in any essential way, since the differentia-
tion so indicated would be so slight as to be of quite a different
order from that produced by conjugation, the latter giving rise,
as we have seen, to coefficients as high as 0.9. Even a slight dif-
ferentiation arising during vegetative reproduction would be of
the highest interest, but it would not alter the positive fact of the
EFFECT OF CONJUGATION 355
immediate production of strongly marked heritable differences
by conjugation.
The data of our present experiment, given in tables 34 and 35,
bring to light many other important relations, which will be
dealt with in subsequent papers. For our present purposes it is
sufficient that the experiment demonstrates that conjugation
produces within a pure race heritable differentiations; so that as
a result races diverse in their heritable characters arise from a
single race with uniform heritable characters.
Our previous experiments had shown that conjugation increases
variation; and that the variations observed to follow conjugation
are heritable. The present experiment puts the finishing touch
on this demonstration by showing that these heritable variations
do not arise without conjugation.? Thus we find that one method
of producing new strains is by conjugation.
We have now in hand the essential facts for drawing conclusions
as to the actual effects of conjugation on the stock.
IV. RESUME OF RESULTS: DISCUSSION, AND CONCLUSIONS
In the foregoing sections are detailed the results of a large
number of experiments in which conjugants were compared with
non-conjugants of the same stock and the same cultural history.
What effects do we find conjugation to produce?
The prevailing view as to the effects of conjugation is that it
produces rejuvenescence in the stock. This view is excellently
stated in Calkins’ recent Protozoology (’09), particularly in chap-
ter II]. The essentials are somewhat as follows:
If we could take such an entire succession of cells thus formed from
the repeated divisions of a fertilized protozoén, and if at any given
period could combine them in one mass of cells, we should have the
analogue of a metazoén and would find that the protoplasm represented
by the agregate of cells would manifest the same successive periods of
vitality as those of youth, adolescence, and old age in Metazoa. We
would find that the young cells divided more rapidly than they do later
in the cycle; we should find that after a certain time they become
sexually mature and are able to conjugate and so to perpetuate the
3 There remains the possibilty that heritable variations of a totally different
(lesser) order of magnitude may arise during vegetative reproduction.
356 H. S. JENNINGS
race; and we would find that, ultimately, evidences of weakened vitality
and degeneration appear in the aggregate of cells, and that they finally
die of old age (p. 103).
It is conjugation that reinvigorates the stock; for succinct,
explicit statements of this we may quote from other papers of
Calkins and his associates:
Conjugation between two cells results in the complete reinvigoration
of all activities, both physiological and germinal (Calkins and Cull ’07,
page 376). As with the fertilized egg of a metazodn, the copula or
fertilized egg of a protozoén is endowed with a great power of cell
reproduction and with a high potential of vitality, and this is the main
characteristic of the first period of the life cycle (Calkins ’06, page 233).
As with the metazo6n so with the aggregate of protozoa cells, we note
a period of youth characterized by active cell proliferation; this in both
groups of organisms is followed by the gradual loss of the division
energy accompanied by morphological changes in type of the cells pre-
liminary to conjugation and fertilization and to the renewal of vitality
by this means (Calkins ’06, p. 232).
The experiments described in the present paper constitute an
examination as to how far conjugation actually exhibits these
effects in Paramecium; as well as how far it shows other results.
We shall here summarize and discuss the evidence as to the
effects of conjugation on the rate of reproduction; on the vigor
or vitality; as evidenced by the comparative mortality; on
abnormalities; on its production of variation; on inheritance;
and the relation of the results as a whole to the theory of reju-
venescence.
EFFECT OF CONJUGATION ON RATE OF REPRODUCTION
Practically all the experiments show that the average rate of
reproduction is less after conjugation than before. That is, if
we take two sets of animals of the same stock and history, both
ready to conjugate; permit one set to conjugate, and prevent the
other, we find that those which have conjugated divide there-
after on the average less rapidly than the others.
In most cases the rate of fission was very considerably greater
in those that had not conjugated, the excess usually varying from
25 per cent up to 80 per cent or more. In some cases, however,
EFFECT OF CONJUGATION 357
the difference is very slight. In no case did the conjugants
have a higher rate of fission, although in Experiments 4, 7, 8,
12 and 14 the difference between conjugants and non-conjugants
was so small as to be without significance. But in the majority
of the experiments, and particularly those which included many
cases and were little disturbed by extrinsic factors, those that had
not conjugated showed a fission rate higher in a marked degree.
And this higher fission rate of the non-conjugants persisted for
weeks and months (see the results of Experiments 1, 2 and 6).
So much has been said of the greater reproductive power, the
‘“‘active cell proliferation,” et cetera, of the period following con-
jugation, that this result appears surprising. Yet those inves-
tigators who have examined the matter with the greatest care,
came long ago to the same result. Maupas insists again and
again, at great length, in opposition to the prevailing views, that
conjugation does not increase the rate of reproduction. Since
the matter is an important one, and one on which incorrect ideas
are prevalent, and since Maupas had evidently done much careful
work on the question, it may be worth while to give a résumé of
the points he makes. The following passage might well be
designed as a statement of the present condition of affairs:
On a affirmé que la faculté fissipare des Ciliés était modifiée par la
conjugaison, et que cet acte sexuel avait, pour principal effet, de la
renforcer et de l’accélérer. Les Ciliés, au sortir de la‘ conjugaison, se
multiplieraient beaucoup plus rapidement qu’ils ne le font plus tard.
Cette opinion est devenue courante, et on la trouve reproduite dans
les Mémoires et les Traités Généraux, comme une vérité definitivement
acquise. Elle a été émise pour la premiére fois, par Biitschli en 1876,
et reprise ensuite par Balbiani, en 1882, qui s’en est emparé, et a
méme cru en avoir fourni la démonstration expérimentale (’88, pages
254-255).
Maupas then examines the supposed evidence of Biitschli and
Balbiani, showing that it amounts to nothing. He sets forth that
in his own records of fissions, beginning in a namber of cases with
ex-conjugants, there is no indication of a greater rate of fission
in the early part of the cycle. He says of the fissions:
Elles se succédent avec une marche uniforme, modifiée uniquement
par les variations de température. Je ne me suis pas contenté de cette
358 H. S. JENNINGS
unique expérience. J’ai isolé d’autres ex-conjugués de la Stylonychia
pustulata, puis de l’Onychodromus grandis, de |’Euplotes patella, du
Paramecium aurelia et de la Leucophyrys patula. J’ai suivi, jour par
jour, les générations successives de leur descendants, pendant des durées
de temps qui ont varié depuis quinze jours Jusqu’a un ou deux mois.
Chez aucune de ces espéces je n’ai constaté la moindre différence dans
la succession de bipartitions. Anciennement ou nouvellement con-
jugués, tous les individus se sont comportés de la meme fagon. (88,
pages 255-256).*
In the paper of 1889 Maupas details experiments with Para-
mecium aurelia (p. 227), Colpidium colpoda (p. 247), Leucophrys
patula (p. 261), Onychodromus grandis (p. 321), Stylonychia
pustulata (p. 329), and Euplotes patella (p. 353), all showing that
after conjugation these animals do not reproduce more rapidly
than later in the history of the strain. In Onychodromus and
Stylonychia, indeed, Maupas found that those which had recently
conjugated multiplied more slowly, but he believed this to be due
merely to individual variations, and to have no connection with
conjugation. He sums up, in opposition to Biitschli, as follows:
J’ai affrmé, en outre, que cette puissance de multiplication se main-
tient reguliére et égale pendant le cycle entier, sans qu’il se produise
un affaiblement graduel depuis la premiére géneration post-syzygienne,
jusqu’au retour d’une nouvelle période de maturité karyogamique.
Autrement dit, je nie que les Infusoires, au sortir de la conjugaison,
jouissent d’un faculté de reproduction plus énergique que plus tard
(89, p. 504).
Richard Hertwig (’89) came, through an experimental study,
to similar conclusions, save that he discovered the fact that
animals which have conjugated actually reproduce more slowly
than those which have not. He was apparently the first to per-
form the experiment employed on a large scale in the present
paper, of separating pairs before conjugation was completed, and
comparing these members of split pairs with specimens that had
finished conjugating. He gives only a general account of his
‘ Maupas notes that of course in the last stages of morpholog#al degeneration
just before death, there is a cessation of fission; but when this condition is reached
rejuvenescence is no longer possible. ‘‘Je suis convaincu que si, dans les généra-
tions d’un cycle, il se produit un ralentissement, celui-ci se fait sentir seulement
dans la période effectée de dégénerescence sénile; e’est a-dire, lorsque les Infu-
soires sont devenus incapable de rajeunissement karyogamique”’ (’89, p 504).
EFFECT OF CONJUGATION 359
experiment, not even mentioning the number of cases examined;
but some cultures obtained from the split pairs were kept as long
as three months:
Als erstes Resultat ergab sich mir eine auffallende Fruchtbarkeit der
an der Conjugation verhinderten Thiere; obwohl ich meine Versuche
noch nicht abgeschlossen habe, so méchte ich jetzt schon hervorheben,
dass die kiinstlich getrennten Thiere lange Zeit tiber sich energischer
theilten als Paramaecien, welche die Conjugation durchgemacht hatten
(’89, p. 223).
These observations led Hertwig to endeavor to save the theory
of rejuvenescence through conjugation, by holding that lack of
conjugation results in a rate of fission so great as to be harmful;
conjugation. would then rejuvenate by slowing and regulating
this immoderate rate of reproduction (’89, p. 226).
But the facts appear to be clear, so far as the infusoria go.
In view of the large number of experiments made by Maupas on
this point, the absolute agreement of his results with those of
Richard Hertwig; the fact that these men are perhaps the most
thorough investigators that have ever worked along these lines;
the further fact that there exist no careful experimental results
opposed to these; and finally, the very large body of evidence
presented in the present paper, all giving the same results—is it
not time that the statements or implications that in the infusoria
conjugation results in increased reproduction should disappear
from the literature of science?
EFFECT OF CONJUGATION ON MORTALITY
The experiments show that as a rule mortality is much higher,
under the same conditions, among those that have conjugated
than among those that have been prevented from conjugation.
This is true both for conjugation among unrelated individuals, and
for that among individuals belonging to the same pure strain.
Accidental influences increasing the death rate quite without
relation to conjugation are so numerous, especially in experiments
carried on under unfavorable conditions, that here the principle
is particulary important that one extensive experiment carried
through under ideal conditions, without extrinsic disturbing
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
360 H. S. JENNINGS
factors, gives a truer insight than many imperfect experiments.
Such a model experiment is, for present purposes, Experiment 1.
In this experiment none of the lines descended from non-conju-
gants (split pairs) died out during the five weeks of the experiment.
Of the lines descended from conjugants, though kept under exactly
the same conditions, 38 per cent died out during the same period.
In the other experiments some of both sets died, though as a rule
with more deaths among the lines derived from conjugants.
In two out of the ten or twelve experiments in which this
matter was tested the usual relation was reversed; in both these
experiments we are dealing with exceptional conditions. In
Experiment 3 the temperature was abnormally high, standing
much of the time above 32°C. I have found by long experience
in Baltimore that it is not possible to carry on slide cultures of
Paramecium at such a temperature; from whatever source, the
animals rapidly die out. Thus, in this experiment the conditions
were so bad that a large proportion of both conjugants and non-
conjugants died within the four days of the experiment. But
under these conditions the lines descended from non-conjugants
died out still more rapidly than those descended from conjugants.
Of the former 68.6 per cent died in the four days; of the latter,
but 23.4 per cent. The difference seemed clearly due to the
furious rapidity at which the non-conjugants multiplied, while
the conjugants (as is the rule after conjugation) divided but
slowly. There is little doubt but that under usual temperature
conditions the advantage would have been, in this case also; with
the non-conjugants. The fact, however, that conjugation may
be physiologically advantageous under very exceptional condi-
tions is an important one.
The other case in which the advantage was with the conjugants
is Experiment 12. Here we are dealing with a much depressed
stock, in which reproduction is slow and mortality high before
conjugation. Such a stock can hardly be induced to conjugate;
so that but three pairs could be obtained from it. With the six
lines derived from these were compared ten lines derived from
non-conjugants of the same culture. It is important to note that
the latter were not split pairs; in other words, they were not
EFFECT OF CONJUGATION 361
ready to conjugate (as were the non-conjugants in most of our
experiments). Such split pairs could not be had in the present
case. Of the six conjugant lines, four (or 66 per cent) died out;
but of the non-conjugant lines all died.
As set forth in the account of this experiment, the ground for
this difference seems to lie in the fact that a certain vigor and
power of multiplication are a prerequisite for conjugation; so
that in the only three pairs that conjugated are included the only
vigorous members of the stock; the others died for the same reason
that they did not conjugate.
What are the grounds for the greater mortality of the conju-
gants, found in the great majority of cases? Two possible
grounds occur to one:
1. Conjugation involves extremely complex and delicate cyto-
logical processes. It seems possible that these processes are
easily diverted into abnormal courses, resulting in abnormalities
and death.
2. Conjugation, like fertilization, is a process of uniting diverse
germ plasms; of producing new combinations of germ plasm
(evidence bearing directly on this will be given in a paper to
follow the present one). Possibly some of these combinations
are incompatible; or produce results not fitted for continued
existence under the conditions.
EFFECT OF CONJUGATION ON ABNORMALITIES
Throughout the experiments it was observed that frequent
abnormalities of all sorts occur among the descendants of the
conjugants, while among the descendants of non-conjugants
such are relatively rare. The grounds just set forth as possibly
accounting for the greater death rate of the conjugants, perhaps
play a part also in the production of abnormalities.
EFFECT OF CONJUGATION ON VARIATION
The most striking effect of conjugation that appears in com-
paring the conjugants and non-conjugants, is the great increase
in variability in the rate of reproduction. In all of the experi-
ments the conjugants are much more variable in this respect than
362 : H. Ss. JENNINGS
are the non-conjugants. It will be well to summarize here the
coefficients of variation for conjugants and non-conjugants in
certain periods of each experiment. This is done in table 26.
As the coefficient of variation was computed for several dif-
ferent periods in most of the experiments, it hardly appears
practicable to bring together in table 26 all the coefficients given
in the tables of the body of the paper. I have therefore selected
the longer periods, with some typical partial periods.
As table 26 shows, the difference in variability between con-
jugants and non-conjugants.is not a slight one, but is very great.
The coefficient of variation averages at least twice as great for
the conjugants, and in some of the cases given in table 26 it is
three or four times as great. There can be no question but that
conjugation increases greatly the variation in rate of reproduction,
both in wild cultures and in pure races.
If in place of studying the variation relative to the mean rate,
as shown by the coefficient of variation, we examine the absolute
amount of the variation, as shown by the standard deviation, we
TABLE 26
Comparative variability, as measured by the coefficient of variation, for the lines
descended from conjugants and for those descended from non-conjugants, in num-
bers of fissions to a given period; for various experiments of the present paper.
| CONJUGANTS NON-CONJUGANTS
(PATRS) (SPLIT PAIRS)
EXPERIMENT TIME Axum Bremen Fy a
ber of Coefficient ber of | Coefficient
| lines | lines |
(Wild Cultures) | |
1 Hirst, 2 weeks: «|. ss. reece oes } 56 | 53.103 = 4.232 59 | 12.975 = 0.819
1 | Second 2 weeks...............-| 42 | 42.870 + 3.689 59 27.743 = 1.850
2 | First 2 weeks. ... 2; ..0..--+s0.. | 34 | 32.011+2.874 | 51 | 21.350 + 1.489
2 | Second 2 weeks...... Gb Accll 19 46.944 = 6.166 26 22.847 = 2.246
3 Foumdays:. 3. dome ae eae 29.369 + 2.528 | 16 12.756 = 1.546
(Pure strains)
Twenty days..... Pe 17 30.828 = 3.890 | 18 19.792 = 2.310
5 [Six dayB....0......ccge02c022.) 200 | 48,762 5,781 | BU) 9) oka een
13a Seven days... .<5.cmc0.2 58 40.350 + 2.910 | 82 17.550 + 0.953
13b 18 Days, December 8-26...... 14 43.893 = 5.976 | 16 | 18.263 + 2.405
13b | January 2-14........ : RE 14 16.878 + 2.069 16 | 12.832 + 1.662
13b January 30-February 12...... 12 37.899 + 5.478 14 | 19.735 + 2.821
13b | December 8-January 29....... 16 21.675 + 2.715 | 14 16.428 = 2.150
15 First 12 (11) days...... Ade aele. 168 19.501 + 1.065 | 158 11.789 = 0.453
15 Later September 8-16......... |. 83 41.074 + 2.485 | 171 8.585 = 0.315
15 24 '(21) days. -.s...55.- | 69 28.650 + 1.734 | 145 7.122 = 0.284
{
EFFECT OF CONJUGATION 363
shall come to the same result, finding that in every case the varia-
tions are not only relatively, but absolutely, greater among the
conjugants. The standard deviations corresponding to the coef-
ficients of variation given in table 26 will be found in the tables
included in the body of the paper, under the different experiments.
In just what way is the variation increased in the conjugants?
That is, do we find that after conjugation there are more speci-
mens with a lower rate of fission, or with a higher rate of
fission, or with both? What is the nature of the distribution
of the fission rates in each case?
The fact that the mean rate is lower for the conjugants would
cause us to suspect that the increase in variation is at least partly
due to a decrease in the rate of fission of some of the lines, while
others remain high. Examinations of the data shows that this
is largely true. To bring out this point, it will be well to note the
comparative range of variation in number of fissions, for the
conjugants and non-conjugants, in the various experiments.
This is exhibited in table 27. In this table are included the
number of fissions for only the lines that lived through the period
specified.
As table 27 shows, at the lower extremity the conjugant lines
range much farther than the non-conjugant lines; in every case
_ the lower extreme for the conjugants is below that for the non-
conjugants, and in many cases the difference is. very considerable.
At the other extremity of the range no such great difference is
found. The maximum is, as a rule, higher for the non-conjugants,
but this is not invariable; in some cases the maximum for the con-
jugants is equal to that for the non-conjugants; or even a little
greater.
It appears therefore that conjugation increases the variation
mainly toward the lower extremity of the range; it produces many
lines whose rate of fission is lower than that for the non-conju-
gants, while others remain high. But even in the middle regions
of the range, the conjugant lines are less heaped up about the mean
than the non-conjugants. These peculiarities may be illustrated
by examination of the distribution of the variations in the experi-
ments with larger numbers, as given in table 28, for Experiments
364 H. Ss. JENNINGS
1 and 15. With this table may also be examined tables 9, 10,
11 and 20, which give the same data for various other experiments.
The spreading out toward the lower end of the range in the
lines descended from pairs is very striking in table 28. In some
cases it appears that the lines descended from conjugants tend
to differentiate into two groups, one with a low fission rate, the
other with a higher one. This is particularly notable in the data
for Experiment 15, in table 28, but is observable also in Experi-
ment 1.
TABLE 27
Comparative range of variation in lines descended from conjugants, and in those
descended from non-conjugants, for fissions in a given period.
CONJUGANT NON-CONJUGANTS
Minimum Maximum
number of | number of Minimum | Maximum
fissions fissions
EXPERIMENT TIME
(Wild cultures) | |
1 | First 2 weeks...........| 0 12 ai 15
1 | Second 2 weeks: -..-.-2- 0 | 36 6 20
1 | Total 4 weeks.......... Ob)" Be ioe aes ie
2 | First}? weeks -;: 2.2. Oh Noha eg 15
2 | Second 2 weeks......... 0 7-8 6 11
2 | Eight weeks............| 25 | 38 37 47
3 | Pour days’. Glire fo. 2 10 8 13
(Pure strains)
4 | Twenty days...........| 1 17 6 17
5 Sis Ways yatta ey a | 6 4 6
il Mine GSy8i. 0. cakes OO 14 li 16
9 Sixteen days........... 19 | 22 23 27
13a Sever days.) acne wae 0 | 10 4 10
13b | December 8-16......... 4 21 12 21
13b | January 2d a i 8 12
13b January 30—February 12) 3 je il 9 14
13b December 8-January 29) 21 | 46 24 50
15 | de) dase ne ee 6 | 16 8 15
15 | September 8-16......... 1 | i 8 15
S 10 35 18 28
15 | 24 (21) dayst........... { (8.075 30.06)
‘In the last entry, for experiment 15, the time for the conjugant lines is 24
days, that for the non-conjugants but 21. If we reduce by one-eighth the fis-
sions for the conjugants, they will then be comparable with those for the non-
conjugants; this gives the figures shown for this case in parenthesis.
EFFECT OF CONJUGATION 365
It will be observed from table 28 that even in that part of the
range where the non-conjugant lines are found, the conjugant
figures are much less heaped up near the mean than are those for
the non-conjugants. This shows clearly that the greater variabil-
ity of the conjugants is not due alone to an extension of the range
of variation toward the lower end; but also to a scattering of those
lying near the mean. If, for example, we omit in Experiment 15
all the conjugant lines lying lower (in table 28) than any of the
non-conjugant lines, we still find the variation for the conjugants
to be much greater than that for the non-conjugants. In Experi-
ment 15, making the omission mentioned, the coefficient of varia-
tion for the conjugants would be 16.776, as compared with but
8.585 for the non-conjugants.
Conjugation, then, increases variability in reproductive power.
The next question is: Are these differences inherited, so that
in this way differentiated races are produced? To this question
were mainly dedicated Experiments 13 and 15, and, as the account
given in the text shows, the differences thus produced are inherited.
In wild cultures, such as that of Experiment 1, this question
cannot be answered so clearly, since the differences in fission rate
existing before conjugation are likewise inherited and the effect
of conjugation is only to increase the number and extent of these
TABLE 28
Distribution of the number of fissions for the lines of descendants of conjugants, as
compared with those from non-conjugants, for certain periods in experiments 1
and 15. (The table shows, for example, that in the first two weeks of experiment
1, three of the conjugant lines did not divide: four divided once, etc.)
NUMBER OF FISSIONS |
oj}1]2/3|4|s5{e|7/8 9 }10] 11 12 | 13 | 14 | 15 | 16 | 17
a = iat | | ea | | | inert
Experiment 1— bh) | |
First two weeks: re et enue “I
Agr) Ser -e4- pr Bili ads et 1 em Wc fod ek UE ee ae 56
Split pairs...... | 2) eee deen sol dead s | 13 5 | 59
Second two weeks) | | | |
Pairs cece 2 1 [22 2 PO S| wes ie 1D aero Blog 42
Split pairs....... Big ira | Seg it pan oles 3 | 59
Experiment 15— | | }
September 8-16 |
Pairs........-..| 6) 5| 2) 1 | 2] 2 2| 5| 9/13/20] 12) 4] 83
Split pairs. .... | | | 4] | 6]13] 52] 68] 27} 4] 171
366 H. S. JENNINGS
inherited differences. But working with a pure race, as in
Experiments 13 and 15, it is found (1) that differences in rate of
fission among those that have not conjugated since they were
derived,from a single parent are not inherited (unless possibly
certain differences of a minimal character are to be excepted;
differences of an order of magnitude far below those with which
we are dealing); (2) that conjugation among the members of such
a pure race does result in differentiations that are inherited, so —
that from a race homogeneous with respect to fission rate, we
get many races, differing in their rates. The hereditary dif-
ferences thus produced are not small and inconstant but so decided
as to give coefficients of correlation up to 0.9 between earlier and
later generations, in spite of fluctuations due to environmental
differences.
To what is due the production of inherited differentiations by
conjugation? Here for the present we can only speculate. It would
seem probable that we have before us something of the process that
we see in Mendelian inheritance. If the members of a culture
differ in their germinal make-up, conjugation among them would
produce many new combinations of germinal characteristics.
The fact that we find such heritable differentiations produced by
conjugation among the members of the same pure race would be
accounted for if the members of the race are heterozygotic,
although all alike in germinal composition. Interconjugation
among such similar heterozygotes would, on Mendelian principles,
produce many new combinations of germinal constituents, just
as happens in the self-fertilization of higher organisms. °
In connection with such a view of the matter, it needs to be
recalled, however, that in our Experiment 13 we were dealing
with a stock that had gone through eight successive self-fertili-
zations, the stock being derived, after each of these, from a single
ex-conjugant. Such a series of eight self-fertilizations would, as
set forth in the account of Experiment 13, go far in getting rid of
heterozygotism, unless the character we are studying depends on
a very large number of independent factors.. In that stock we
nevertheless found that inherited differentiations as to fission rate
were produced by conjugation. This may indicate that Men-
EFFECT OF CONJUGATION 367
delian recombination is not the whole secret of the matter; it
does not, however, demonstrate this. :
In a previous paper (711), I have shown that conjugation like-
wise increases variability in size. In the account of Experiment 9,
of the present paper, some data are given indicating that the size
diiferentiations so produced are likewise inherited. But the
results there given are by no means conclusive, the matter
requires further study.
It seems best to reserve for a later paper on inheritance a com-
parative review of what is known as to the production of variation
by conjugation in other organisms, with the various theories that
have been held.
CONJUGATION AND BIPARENTAL INHERITANCE
In a paper to be published at once, by the present author and
K. 8. Lashley, it will be shown that conjugation results in inherit-
ance from both the parents that enter into the pair. All details
are reserved for the paper referred to; the matter is mentioned
here merely to complete the outline as to effects of conjugation.
CONJUGATION AND THE THEORY OF REJUVENESCENCE
The chief positive results from the present investigation are:
(1) that conjugation increases variation, giving rise to heritable
differentiations; (2) that it results in biparental inheritance (to
be taken up in a separate paper); (8) that the fission rate is lower
after conjugation; (4) that the mortality is as a rule higher, and
abnormalities are more common, among the descendants of con-
jugants than among those of non-conjugants.
What is the relation of these results to the theory that conju-
gation produces rejuvenescence?
A number of diverse things have been included under reju-
venescence, the theory meaning for some authors one combina-
‘tion of these, for others another combination. The main points
included appear to be the following:
1. The structural changes—the replacement of the old macro-
nucleus by a new structure derived from the micronuclei—has
368 H. S. JENNINGS
sometimes been held to constitute a visible rejuvenescence, a
rejuvenescence of the macronucleus. This in some forms is
accompanied by a renewal of other structures, as for example,
of the bodily appendages in the Hypotricha. Engelmann (’76)
emphasized these changes as constituting in themselves “ein
wahren Verjiingung”’ (p. 629).
This actual replacement of old structures by new no one will of
course deny, and it seems not inappropriate to call it a reju-
venescence, if we mean by this word nothing more than these
observed facts.
2. But the theory has as a rule gone far beyond these observed
facts. Thus, Maupas says, after a statement of these structural
changes:
Ce nouvelle appareil nucléaire agit sur tout l’organisme, auquel il
appartient, comme une sorte de ferment régénérateur, lui restituant,
sous leur forme parfaite et intégrale, toutes les energies vitales carac-
téristiques de l’espéce. Cet étre se trouve done rajeuni dans le sens
littéral et absolu du mot. II peut dés lors redevenir le progéniteur
d’un nouveau cycle de multiplications agames, dont toutes les géné-
rations successives seront douées des mémes facultés rajeunies, Jusqu’a
ce que celles-ci s’usent et s’affaiblissent peu & peu, par leur exercice
méme, et en arrivant ainsi 4 ressentir le besoin réparateur d’une nouvelle
période d’activité fécondatrice (’89, p. 434).
Now, in the passage, we have quoted, Maupas evidently
affirms certain things that by no means follow from the structural
changes observed, but can only be demonstrated by the results
of experimentation. We shall have to inquire how far these have
been thus demonstrated. But before doing this, we must pro-
ceed to one farther development of the theory.
3. As we have seen (page 357), Maupas did not hold that the
vigor and rate of reproduction are increased by conjugation,
although such a general statement as the one above quoted would
seem to imply that this is true. This idea has, however, been
held by many as a fundamental part of the theory of rejuve-
nescence. The rate of reproduction has been held to become
less and less as the number of vegetative generations increases,
until by a new conjugation it is brought back again to its original
level (see the quotations on page 356).
EFFECT OF CONJUGATION 369
This part of the theory of rejuvenescence which holds that the
vigor of reproduction is increased by conjugation appears to be
definitely a mistake, for the infusoria, as we have already shown
(page 359). We shall therefore consider it no farther.
The experimental results of the present paper of course do not
alter the facts as to the ‘structural rejuvenescence’ if one desires
so to call it. Certain points are worthy of notice in this connec-
tion. !
1. So far as the ‘rejuvenescence’ or renewal of structures other
than the macronucleus is concerned (locomotor organs, et cetera),
this takes place equally in vegetative reproduction. It furnishes
- therefore no foundation for a theory that conjugation is in any
special way a rejuvenating process.
2. The replacement of the macronucleus by parts of the
micronuclei of the two individuals of the pair is of course thor-
oughly in consonance with the results of the present study,
furnishing not the slightest difficulty for interpretation. The
micronuclei are to be conceived as corresponding to the nuclear
apparatus of the germ cells of higher organisms, each one con-
sisting of a certain combination of ‘determinants’ or ‘genes.’
When the macronucleus is replaced by parts of two micronuclei,
a new combination of ‘determinants’ is thus produced; the progeny
may therefore differ from the parents. In other words, ‘varia-
tion’ is induced in conjugation—through the production of many
new combinations, in different cases. Again, since the new ma-
cronucleus is produced by the union of parts from two diverse
individuals, the progeny may inherit from these two; in other
words, conjugation results in biparental inheritance, as we have
actually found to be the case.
Now, it is a priori not impossible that the effects of the renewal
of the macronucleus are completed in those two results; it is not
a priori certain that the new macronucleus must otherwise func-
tion any better than the old one.
We must therefore inquire as to the experimental ground for
the assertion made in the quotation given above from Maupas, to
the effect that this new apparatus acts on the entire organism as a
370 H. S. JENNINGS
sort of regenerating ferment, restoring all its vital energies, et
cetera.
The grounds for this view have consisted, almost exclusively,
not in actual observation of any such rejuvenizing action by con-
jugation, but in the observation that during vegetative repro-
duction under experimental conditions the organisms become
depressed, degenerate, and finally die. From this it was con-
cluded that conjugation must be what remedies this.
This line of argument has, however, quite lost its force, in view
of the modern work of Calkins, Enriques, Woodruff, and others.
These authors’ results demonstrate that the very limited periods
within which Maupas observed degeneration has no significance
for the question as to whether degeneration is an inevitable
consequence of continued reproduction without conjugation, for
they kept vegetative reproduction in progress for periods many
times as long as those which Maupas found to result in degenera-
tion. The work of Woodruff, in particular, seems to show that
Paramecium may be kept multiplying vegetatively for an indefi-
nite period. Furthermore, the work of Enriques and of Woodruff
has shown to what the degeneration observed by Maupas was
due. Under proper nutritive and chemical conditions no such
degeneration appears.
It is not necessary to review in detail this vast subject, but
there will hardly be any dissent from the statement that the
modern work has largely, if not entirely, deprived of its force this
argument for the necessity of conjugation.
All the more therefore we are driven to examine the direct
evidence as to the rejuvenating effect of conjugation. And in
doing so, we must reflect that if the argument above mentioned
were valid, there should be no difficulty in observing experimen-
tally the rejuvenating effect; so that a fortiori we must demand
what this direct evidence is.
In reading Maupas’ great works (’88, ’89) in search of this
direct evidence for a rejuvenating effect of conjugation, one is
astonished at the way it eludes one at every step. Most of the
actual observations that bear on the matter at all, seem indeed
EFFECT OF CONJUGATION atl
opposed to the rejuvenating action of conjugation. Maupas
demonstrated, as we have seen, by extended experimentation,
that conjugation is not followed by an increase in the vigor of
multiplication. He found, in repeated observations, that con-
jugation within his degenerating stocks: did not help them, but
attributed this to their being closely related. But he observed
further that when the depressed stocks that interconjugated were not
related, they still died after conjugation, so that conjugation did not
remedy degeneration in the one case or the other (89, p. 409).
He found that conjugation is often sterile (followed by death)
in wild cultures of Stylonychia (’89, p. 331). He found that
ex-conjugants of Spirostomum, Climacostomum and Didinium
did not reproduce farther (’89, pp. 277, 295, 297). In Leucophrys
a large proportion of the conjugants die (’89, p. 254-255). He
found that in some cases a second conjugation follows a first one
after but a few generations (Leucophrys, ’89, p. 409). He found
that animals which are ready to conjugate may be prevented, and
they will then continue to multiply with uninterrupted vigor
(89, p. 306). All these observations speak against rather than
for the idea of a regular cycle of vegetative reproduction, result-
ing in degeneration, and requiring conjugation at a certain stage,
this remedying the degeneration.
Has Maupas absolutely no evidence that conjugation reju-
venates? He seems possibly to have held that the following fact
is evidence of this effect. In his long continued cultures, he
found that when the animals derived from a single parent inter-
conjugated, they later died. It is notable that this result has
not been confirmed by later investigation, and Maupas himself
noted certain exceptions. But Maupas found that when he
mixed individuals from different cultures, the pairs were fertile
(provided both did not belong to degenerated cultures). It
would appear that Maupas supposed that rejuvenescence had
taken place in these cases. But of course there is absolutely no
evidence that such has occurred, unless it is shown experimentally
that the ex-conjugants are more vigorous and propagate longer
than similar parents who did not conjugate. In view of the
372 H. S. JENNINGS
results given in the present paper, where the reverse is shown to
be the rule, it is clear that these observations of Maupas do not
touch the matter at all.
One single case only Maupas has which makes even an approach
to the form of this necessary demonstration, and this, as we shall
see, really gives no evidence at all. This is the case of one of his
cultures of Stylonychia pustulata (’88, pp. 196-201). A line of
propagation was begun with a single individual, November 1,
1885. This line died out on March 26, 1886, after 215 genera-
tions. On February 22 a single specimen of the 156th generation
was taken from this line and allowed to conjugate with an indi-
vidual from outside. Maupas tells us on page 323 of his paper
of ’89 that these individuals from outside, which he mixed with
those from the long-continued cultures ‘‘were taken at hazard
in my small aquaria.’””’ Thus such an individual had not been
living under the peculiar conditions of these experiments. De-
rived from this pair a new line of propagation was continued
for 316 generations (till July 10, 1886), while the old line from
which one of these ex-conjugants came, died out after but 59
generations more.
Now, the work of Enriques, Woodruff, Baitsell (712), et cetera,
has shown that the conditions with which Maupas worked result
after a time in depression of the vital functions, but that animals
kept under more favorable conditions do not show such depres-
sion, even though they have lived as long without conjugation
as the depressed race. The depression is due to the conditions,
not to lack of conjugation. What Maupas did was to take from
outside a fresh, vigorous specimen, and mate it with one of these
depressed ones. He then found that the progeny were vigorous.
He does not note whether the line of progeny he used came from the
depressed member of the pair, or from the vigorous one, although
this is an absolutely essential point for determining whether the
depressed stock was rejuvenated even by conjugation with a
vigorous one. The probability is strong that the new line of
propagation came from the new, vigorous individual. But such
an individual would have given an equally long series of vegeta-
tive propagations if it had not been mated at the beginning. Its
EFFECT OF CONJUGATION 3163
vigor was due to the fact that it had been living under favorable
conditions, not to conjugation.
There is absolutely nothing in this experiment to demonstrate
that a partially exhausted race is rejuvenated by conjugation.
A real test would be the following: Two unrelated lines should be
allowed to multiply till both become depressed. Then they should
be allowed to conjugate, to determine whether the conjugation
remedies the depression. It will manifestly not do, in testing
the question whether conjugation remedies depression, to take a
vigorous, undepressed specimen as one member of the pair.
According to the cyclical theories, all lines of propagation become
depressed after a series of vegetative reproductions, so that if
conjugation is to maintain the race, it must be effective when it
occurs between two lines, both of which are depressed.
Now, as we have briefly mentioned above, Maupas performed
this crucial experiment. He kept lines of propagation of Stylony-
chia of diverse origin till they became depressed, then allowed
them to conjugate one with another. This fact is briefly set
forth on page 409 of his paper of 1889. Such cross-conjugation of
two diverse lines did not result in rejuvenescence; the animals died
just as happened when the two members of a pair came from the
same parents. Speaking af sterile conjugations, Maupas says,
“Blles s’effectuent, en effet, aussi bien entre individus apparte-
nant 4 un méme cycle ou proches parents, qu’entre individus
étranges l’un A l’autre et provenant de cycles différents” (89,
p. 409).
Anyone who goes critically through the 480 pages of Maupas’
two great papers for the purpose of finding out what evidence
there is that conjugation rejuvenates, will, I believe, be forced, as
I have been, to realize that they contain no evidence for this
whatever, although they do contain evidence against it. Maupas’
conclusion was evidently due to the supposed theoretical neces-
sity for something to remedy the degeneration induced by long
vegetative reproduction under the conditions of his experiments.
All that his experiments show is that long continued propagation under
the given conditions results in injury to the stock—and this equally
whether there is or is not conjugation within the stocks, or between
374 H. S. JENNINGS
two such stocks of diverse origin. There is thus not even any indi-
rect evidence that conjugation rejuvenates, since the stocks that
conjugated underwent the same fate as those that did not.
So far as I have been able to discover, there is no experimental
evidence from any other source that conjugation rejuvenates.
In Miss Cull’s paper entitled ‘‘Rejuvenescence as the result of
conjugation” (’07), the evidence consists merely in showing that
a considerable fraction of those that had conjugated continued
thereafter to multiply. But control experiments show, as set
forth in the body of the present paper, that they would have con-
tinued equally if they had not conjugated; in fact a larger propor-
tion would have continued to multiply if they had not been
allowed to conjugate. There is thus in these results no evidence
of rejuvenescence through conjugation; and this must be said of
all observations which merely show that some of the ex-conju-
gants continue to multiply. Control experiments with animals
prevented from conjugating are necessary for a correct under-
standing of the results.
In the long series of studies set forth in the present paper as
Experiments 5 to 14, the effects of conjugation were studied when
one division of a race is allowed to conjugate frequently, while
another is kept from conjugating; also the effects of conjugation
in a race that is actually depressed. As to the first point, the
animals that did not conjugate were found throughout to be more
vigorous than those that conjugated frequently.
With regard to the effects of conjugation in a depressed race,
it is to be recalled that Maupas had repeatedly tried this experi-
ment, finding always that conjugation has no beneficial effect
under such conditions. The question might then be regarded as
settled, since there is no expectation of beneficial effect even
accepting the views of the great upholder of the theory of reju-
venescence; positive results would be directly opposed to the
experimental results of Maupas.
Yet it was in one of these experiments alone that any result was
‘reached that could possibly lend themselves to an attempt to
maintain that conjugation has a beneficial effect on vigor and
vitality. In Experiment 12 the stock was so depressed that it
EFFECT OF CONJUGATION 375
multiplied scarcely at all, and the mortality was high. It was
almost impossible to get conjugation among its members, since
a prerequisite to conjugation is a period of rapid multiplication.
The necessary conditions were fulfilled only for three pairs.
From these, six ex-conjugants were obtained. The six lines of
propagation derived from these were compared with ten lines from
individuals that did not conjugate (and did not attempt to do so).
There was no general rejuvenescence due to conjugation.
Three of the six conjugant lines died out within a week, and a
fourth a little later; so that two-thirds of the conjugant lines were
dead. But two continued to multiply. But in the mean time
all of the ten non-conjugant lines died out.
What has happened here? We can hardly speak of rejuve-
nescence where two-thirds of the ex-conjugants die out. The
survival of some of the conjugants may have been due to the
greater vigor that was a prerequisite to their conjugation, the
lack of which caused the others not to conjugate. Aside from
this we can only say that the results of conjugation were here the
same as usual; it induced variation in the reproductive power.
As always, some lines derived from the conjugants had a low
reproductive power and died at once. Two out of the six had
greater reproductive power; they therefore continued to multiply.
In the meantime, the uniform non-conjugants, retaining the
original depressed condition, all died out after a short time.
This experiment therefore gives, in fact, the same result as all
the others, an increase of variation as a result of conjugation.
It differs from the others merely in the fact that in two of the six
cases the extremes of variation reached a higher level than that
which characterized the animals before conjugation. The same
result is reached in Experiment 15, where the conjugants at the
upper extreme of the range exceeded in their rate of reproduction
the uniform non-conjugants. But in this latter case there is no
temptation to speak of rejuvenescence, since the non-conjugants
still continue to multiply vigorously.
Thus, under exceptional conditions the production of variation
by conjugation results in preserving some representatives of a
stock which would otherwise die out completely.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, NO. 3
376 H. S. JENNINGS
The results of the present investigation on the effect of con-
jugation need to be considered in connection with the results of
the investigations of Calkins, Enriques, Woodruff, and others,
on the results of long continued vegetative reproduction. The
two lines of work complement each other and lead to harmonious
and definite conclusions. In a recent brief paper (’12 a) I have
reviewed the two in their relation to each other. Here I shall
not attempt to review the work on vegetative reproduction but
merely to summarize the common result of both lines of work.
GENERAL CONCLUSION
Comparing conjugation with the fertilization of higher animals,
we find the following to be the state of the case:
In higher animals fertilization has two diverse effects, which
recent investigation, particularly that of Loeb and his associates,
has clearly disentangled. (1) On the one hand, it initiates develop-
ment; it prevents the egg from dying, as it would do if not ferti-
lized. This function of fertilization is the one that is replaced by
the processes which induce artificial parthenogenesis. (2) But,
secondly, fertilization brings about in some way inheritance from
two parents. When there is inheritance from but one parent,
the inheritance is as it were complete; the child as a rule resembles
its parent in all hereditary characteristics; this is the result of
the so-called ‘pure line’ work. But when we have biparental
inheritance, a great number of different combinations of the
characteristics of the two parents are produced, so that the proc-
ess of fertilization is one that in this respect completely alters
the face of organic nature, producing infinite variety in place of
relative uniformity.
These two functions of fertilization, the initiation of develop-
ment, on the one hand, and the production of inheritance from
two parents, on the other, are logically independent; they might
conceivably be performed at different times and by different
mechanisms. The fact that in many organisms the same mech-
anism that brings about biparental inheritance is likewise the one
that initiates development might from certain points of view be
EFFECT OF CONJUGATION 377
called an adaptation. Its result is to insure that in all organisms
that develop there shall be inheritance from two parents, not
from one. In the work on artificial parthenogenesis these two
functions have been separated experimentally; the initiation of
development takes place alone.
Now, in endeavoring to understand conjugation, attention has
been given hitherto almost exclusively to the first of these two
functions. It was held that the function of conjugation must
be to make possible life and development where it was otherwise
impossible, just as fertilization arouses the egg to further life
and development. But it turns out that in the infusoria con-
jugation, instead of having this one of the two functions of fer-
tilization, has the other. The two functions are in the infusorian
separated, just as they are in artificial parthenogenesis, but it is
the second, not the first, that we have before us. Conjugation
is not necessary in order that life and reproduction shall continue;
they continue without it. There is no evidence that conjugation
in the infusoria increases the reproductive power, or rejuvenates
the organism physiologically in any way.
But the life which thus continues is uniform and unchanging.
To give biparental inheritance, with varying mixtures of the
characteristics of the two parents; to produce these new combina-
tions in great variety, conjugation is necessary. And when this
happens under such conditions that the original combinations
were not adapted to survival, then some of the new combinations
produced often are adapted to the conditions; conjugation then
results in a survival of an organism that would have been com-
pletely destroyed without it. It is most interesting in this
connection to observe that conjugation is usually induced by an
unfavorable change of conditions, a change of such a nature that
the organisms begin to decline. Thereupon conjugation occurs,
so that new combinations are produced, adapted to varied con-
ditions, some of which may survive.
Thus the whole series of investigations on vegetative reproduc-
tion and on conjugation leads to a unified result, and one that is
in consonance with what we observe in higher animals.
378 H. S. JENNINGS
Our main results may then be summed up as follows: So far as
physiological effects are concerned, conjugation does not produce
rejuvenescence, for after conjugation most of the animals are less
vigorous than before. What conjugation does is to bring about
new combinations of germ plasm, just as is done in the sexual repro-
duction of higher animals. One result of this is to produce biparental
inheritance; another is to give origin to many variations, in the sense
of inherited differentiations between different strains. Some of the
new combinations are better adapted to the existing conditions than
others; these survive while the others die out.
LITERATURE CITED
BaITsELL, G. A. 1912. Experiments on the reproduction of hypotrichous infu-
soria. I. Conjugation between closely related individuals in Stylony-
chia pustulata. Jour. Exp. Zoél., vol. 13, pp. 47-76.
Catxins, G. N. 1902. Studies on the life-history of Protozoa. I. The life-
cycle of Paramecium caudatum. Arch. f. Entw.-Mech., Bd. 15, pp.
139 -186.
1906. The protozoan life-cycle. Biol. Bul., vol. 11, pp. 229-244.
1909. Protozoology. 349 pages. New York and Philadelphia.
Cauxkins, G. N., anp Cutt, S. W. 1907. The conjugation of Paramecium aurelia
(caudatum). Arch. f. Protistenkunde, Bd. 10, pp. 375-415.
Cutt, 8. W. 1907. Rejuvenescence as the result of conjugation. Jour. Exp.
Zool., vol., 4, pp. 85-89.
Kast, E. M., anp Hayes, H. K. 1912. Heterozygosis in evolution and in plant
breeding. Bul. 243, Bureau of Plant Industry, U. S. Dept. Agr., 58
pages.
ENGELMANN, T. W. 1876. Ueber die Entwicklung und Fortpflanzung von
Infusorien. Morph. Jahrb., Bd. 1, pp. 573-634. F
EnrIQUES, P. 1903. Sulla cosi detta ‘degenerazione senile’ dei protozoi. Moni-
tore Zool. Ital., vol. 14, pp. 349-351.
1907. La conjugazione e il differenziamento sessuale negli Infusori.
Arch. f. Protistenkunde, Bd. 9, pp. 195-296.
Hertwic, R. 1889. Ueber die Conjugation der Infusorien. Abhdlg. d. II. Cl.
d. kénigl. bayr. Akad. d. Wiss., Bd. 17, (Abth. 1), pp. 151-233.
JenninGs, H. S. 1910. What conditions induce conjugation in Paramecium?
Jour. Exp. Zoél., vol. 9, pp. 279-300.
1911. Assortative mating, variability and inheritance of size, in the
conjugation of Paramecium. Jour. Exp. Zodél., vol. 11, pp. 1-134.
1912. Production of pure homozygotic organisms from heterozygotes
by self-fertilization. Amer. Nat., vol. 46, pp. 487-491.
1912a. Age, death and conjugation in the light of work on lower
organisms. Pop. Sci. Monthly, June, pp. 563-576. (Also in the Harvey
Society Lectures for 1911-12, pp. 256-276).
EFFECT OF CONJUGATION 379
MeEnNpDEL, G. 1866 (Experiments in plant hybridisation; transl. in Bateson’s
Mendel’s principles of heredity, 1909.)
Maupas, E. 1888 Recherches expérimentales sur la multiplication des infu-
soires ciliés. Arch. d. Zool. Exp. et Gén. (2), T. 6, pp. 165-277.
1889 La rajeunissement karyogamique chez les ciliés. Arch. d. Zool.
Exp. et Gén., (2) T. 7, pp. 149-517.
PEERLES, F. 1912 Regeneration and regulation in Paramecium caudatum.
Biol. Bul., vol. 23, pp. 154-170.
Wooprurr, L. L. 1911 Two thousand generations of Paramecium. Arch.
f. Protistenkunde, Bd. 21, pp. 263-266.
APPENDIX: FUNDAMENTAL TABLES
TABLE 29
Experiment 1. Paramecium caudatum. Number of fissions per week in the 61
lines derived from conjugants, and in the 59 lines derived from those that have not
conjugated. May 4 to June 7, 1909. Numbers in. parenthesis indicate that the
line in question had died out before the end of the experiment. (d = died out,
during the week indicated.)
Pairs (conjugation consummated)
10
11
=| | a} i] FS]
< | < | <
DO WEEK | ae WEEK } p WEEK
a | LP ese’ 8 | Peal 2
2 2 fel &| Zi ealla2 3
Qa ie =] al j BH | ela &
Fe Si Sel Oe aia | a2 sat a lee Tse | ge) | see es. | 9
a | oO} 1\@ fiz} a} 6] 5| 3] 4] 2] 20 jas}a] 4] 5] 5] 2] @| a6)
pm ee a (3) | 5 fe ee ed Hl be) 5iieG ly oleae) 8
29) a oe Ne 2 Psa iS) F| we by 4) -28" | Oe eae eee aie | 20
b| 5] 5| 6| 4] 1d] (21) beeen ao | ose Ps, 174 b}| 1] 0] a (1)
ae 68 6 | 2) 26.) a bat 7) 6) Gi si 26. 25) a | 4.) 2) ae (9)
BP 2). 6) 2} 1) @ jdo):| biet Gil 7 |) sap es foes b| 4] 6] 4] 5] 0} 19
acl i) 0 0} a sacl th oa] a2 (2)//26)/ a] 3] 6] 5] 6] 2] 22
Bap ols) fh 2) 20-1 ei a a a (7) Boe eae Rabel. 30
a| 0| 2] 0| @ (2) | 16] a | 3] 2d) (5) }27/ a | 0] @ (0)
by} 5) 4/6] 4) 2) 21 | Bulk Sp Sosa) S pale | b | 4| 4d (8)
sep 5) 6) 4[°7|-28 farted Sta o7] 2) Bi ae | os] a | wh aioe (8)
Bales, 5 |S 0%) 6) 7 Bice ey (1) | b| 4| 6] 3d (13)
al 2) 1) 3] sf 4) 13 fisfa! SS) eho! 7] 2] 29 f2ola || 7m! 6} sd 26
b| 4| 2| 4 | (6) | bh Si Fie) 4] El 48 bj 4] 8] 9] 5] 4] 30
a| 0] 6| 8| 4/ 3] 21 | 19/a | 5] 6] 0} 0} 2] 13 | 30] a] 3) 5] 7] 6] 3] 24
by 4ie ees) 51 e3| 23 Be yoo Or. OP 2h 12 by ee cele 3 | 2D
au} 0 |) 0)) a CO) 20h 4s baa eah, alt Oy ails tern nee (0)
B00 |) 2 | (0) Bi 23) Ee Screg ee aa |
a| 4| | | 5 OGLE fea th 9591) wf 28] ua
Bale Geer Bl 8 | 22 b}| 5] 8} 8] 7] 4] 32 |
a| 4| 6) 7| 3] @|(@0)/ 22} a] 5] 6] 4] 4/ 1} 20 |
Bo a S| 61 5 | 3h 23 |b] 3] 7/9} 7] 4] 30 |
JENNINGS
s.
H.
380
TABLE 29 (ContinvED)
Split pairs (conjugation not consummated)
TyLOL| PVLAASKEH OHNO AH LCRANLA
AANA MANAINANMANDARAA
1p | HO OD ND OD OD OD I Hon HOD I OD OD
ke Oe 6 ee ee
we | CO Hm XH HID | oD CO IN 1 ON A IN 1
Md ——-
a op | 2 OM AO 1B © 1B/ OS | & C19 1 O&O 4 XH OO
is I —
a| Oorroeonnrnrolnmonoonrntto
q|ROOROnKRO|IN HEH AMODONH HM
IVOQIAIGNI SQ AN GO a O
- | a oO ~H ID OMmMADARMOHAM HD
ulvyd IWS | GQ AN NAN AN AAA AA © OD oO oO OD
nvgon| ®BSMtxASRResSsnrneonwxoeooesoneon
DAAADAAANAAAIUAAMRMAAMDMHAA
ip | HHANHAAMDMAAMONAAA HOD OO
<H mM 1 Hd ™ 1D Hd 1D OH OD OD HO CO Hl DH H tH
td : — SS —_—— = _—
te ey See cp Ore 9) e009) Nes Hes Gp Ep) GS ieee He
a wononor~rnr~rnronrnrnrtovoonmnm © © 19
~ | corner dtrn nnn nmonnnononnr ©
IVOAIAIAGNI SOganaontnaeotogoegos®Og oO
4 a on 19 ce} r Co) a o
aviot| PA RAR SAMSKFPOHSBH OH oe Oo
AMDAMNANANMANANNMMNDANNAADA
in MD MAMANVAMNM AM MMOTAAAAWMA MH
| OR OOHOSCHH OW OM 1D OH 1 1H © 1H &
td es = ss ESS eS
Ha mo | rH POArOSOHDONMAOANMHNMAOrA
an| Hooronwoowrnonnonnoonr
= COOn~ ne ~~ OM OM & & ™ ™ DO O'R
IVOAIAIGNI arnanmanteaonsnanoagno SOadgos o
Riva tiad pr Ok Nie SS ee
TABLE 30
Number of
April 10 to June 4, 1909.
fissions, first two weeks, for those that have conjugated (pairs) and those that
Experiment 2. Paramecium caudatum.
i)
2
td _ et al
J e-sonannene
ress — SS Oe
iva)
‘'~
= a ee ae
s
= | a| ae aun es
a and ae =
§ E = | ioiy ee
& a
8 ee
a nvadiaani| acaodoaoeo
2 de ee S ee a =
é ©
S urva | © 5 @ 2
a = a
me
S38 Ts ee
Ps) qvton| MEME See ae ZSnGvarnm
Ee
3 8 ip". Dae s : re
ee | =a Boos Gin Sie ew FE Sey
2 5 na | oe a ae
< 3 | —_-—_______-——
ato. E ~ | MOKRHBMOMHMMODHO OM HH
oo = "
= 5 QNVOQIAIGNI| sanonoaongaentsoadanoacoal
So ait. — z = Pp er, eS
Lewaer mvaj/5 2 288 8 R B
A = —
SIS pes EEE
J& nvton | MMOon Bron aSwaoSEr Sr
ie
nw ~
o v
eed ; Saeed ay em) i ee aan
aml Sn on an _ oo
me Bom ue cb tol" Beare
ss E a | ANMNMHN KH HN AMOBNH HUH
Sw =e A var ME Mab =
§ et a a
oe NIVOGIAIGNI| CO ae@oanoaoasnotgonoao a4
~~ SS
Xs = = es es ae. =
s8 ent oe
as —— —_————_ — - —- — = ——————_———
Wes AVLOL coanSroonownenSor
~
ae
m 2 i = = = ~ a ee
Ls rv) ACMHM HMOABAABA )
iS) nN a
~ 'S | “ = =.
= 8 E al *TOKTOSOVMDMOWH WHO OD TO
sie _ é a i 3 ree
~
Sas JVOGIAIGNI|] @eQo anoadnanendgnoenoaso
Rie = = _
3 eee tS
ae) urva mt a oo bal io) o ~ o
381
TABLE 30 (ContINvED)
Split pairs
EFFECT OF CONJUGATION
(9)
(8)
ontLanllantans
IVLOL Sie Col eld
a CEES Sey ei BBONMAMSD oss sd 38
tn |
eS
RrRWWRDOAHOOWEAMAV]!8H
ee
seaaenaN dg AMaN ts QeO eo
a | CON M tH HO DO
avaararanr | SOSOSOBAA
a>
Ulvd Litas | & = ce) =) S| ee Se
3
4
25
26
27
=
TVGOL || £2 0 2 oo 2
i CQ |S eat SES CNRS CN NIC RSA aN. oS SIS, tS eS
a = nce 3 : aa =
Ee =! onr~ooroonrme © Ee Bee), 00) 8) CO aS) FS! 00) "SO
IVOGIAIGNI SOesgo eno ega 80 SANBGANAABOSOBO BO DO
N ~
Sivauidds |e Siok se eS ie toe Sy RS) anes ae ees
os =: - = ———— — S = ———— - == ==
S
os a EN R, Atnmnowmarnmonoas
viol) JY Sen eees Pee cat aa Ree
4 a wn oo 8 ~VSuz GOV GS) SE) Cet SS) tel GG ae} OY tea) OU aS
a “ei
a | XxOwnVORONMAON OAMHOKNAFrENOHBAAM SO
eS
TV OGIATGNI 8 2 SNsGCOaAaN GSO SO AGO SO
8
9
0
Ss
COACeIGC CT ko Eb EI Co cl ress
1
11
Sues aanara ©
at Oo Hin Nd o
IVLOL Sntonoanons it Wana St ot
east geal hes ceed Sst mS ae:
ANANNDBDMM SD
i4 nN po Balad ost ino Gris! v=J.0) Beh Sat eh ats] Soh Sal as} 3
a) =
E —y CHOOMDHOnr-AwOD oromonmnonornroennr
on] oO aes
(a and b have no relation, but were arbitrarily designated at the beginning)
IV OCIAIGNI SOs Q80O009O 8 2
Ulvd LI1d8 = co oF — ar) - N or) aa 12> © ~
JENNINGS
Ss.
H.
382
TABLE 31
Experiment 2. Number of fissions, last 6 weeks of the period of eight weeks, for
those that have conjugated (‘pairs’) and for those that have not (‘split pairs’ and
(d = dead.) Numbers in parentheses give total fissions for those that
died before the end.
‘free’).
~ wo tC _ on
sxaaM g-1¥LO5 | cs sd 3 Se emery ea |
collie Lt} © in O19 19 1 SS
| om Cal = N NAN TN
| 2 Ue ibe tela Pq eel Pe) MAAS MOS EN OSS Se aC
ap | Pam eRin ae et tot oe aR ID OI 1D DO ID O O eH HO eH 1D
IVOQIAIGNI| OD© oon gcosgasoh Oacas S®HOeSQOsgosds
Uulvd aoaoornrn OH 1d rh tH 10 © ™~ oO —N OD Hd Oo
Son I oe | SS Ss et sl = = et = =
nN o ™~
SHAAM § IVLOL oo 0 % 2aao a 3S 8 oo Oo SS x
oo as) N19 30 RB 1D fo oO iD in VS Ow He OO tt Oo om ons
a Oe ee
~ i>] or ANN mr NN NOonAN OC Sees Net mNN et
a a a a ee
I © o IO Had st mE oO <H rh ~-eNON =H wm Oo Oo © oo © 19
Q
B 19 oO ~moooowonrsv O17 0 =m OO COD Sa OO oO on Om
son Sen A oe oe Nn Sal Se I oe A oe oe ~~
aa ~~ SH Hid 19 19 oO HID OO o> oD © iO Hid Hid HC WH iD © OO
|
co CoO HH HO IO +H OD H ID stomp oO N19 019 SH Hm & 1 CO
QIVOGIAIQNI|; go BQNOSQO8O0G SB SBQNNAS NHN SGSHOS FO
utya | = N oO =H a> orm _ N oO aM mm NN OO tH AO o~ on
|
|
Split pairs...
RRO s: sidatists +s
Rairss Sx. 2.55
Experiment 12.
moved to leave but 4 for future multiplication.
TABLE 32
Paramecium aurelia.
days of the experiment, with the reductions made.
EFFECT OF CONJUGATION
383
Number of individuals present on certain
Such entries as ‘14-4’ signify
that 14 were present on the day in question, and that a sufficient number were re-
No individuals were removed
except in the numbers and on the dates shown. Each line started with a single
individual, August 11.
Set K. Progeny of the conjugants of June 3. Cultivated in watch glasses since
May 16.
AUGUST SEPTEMBER
LINE
20 25 27 29 1 3 5 6 7
sof) 16-4 | 8-4 | 8-4 | 12-4 | 16-4 13-4 | 14-4 7 13
2 16-4 | 20-4 | 8-4 | 16-4 | 21-4 8-4 | 30-4 8 19
3 16-4 | 13-4 | 8-4 | 16-4 | 32-4 disco|ntinuled
4 12-2 | 21-4 | 12-4 | 8-4 | lost
5 8-2 | 12-4 | 10-4 | 14-4 | 13-4 | 19-4 disc ojn tinuled
_ 6 8-2 | 12-4 | 11-4 | 7-4 | 18-4 | 29-4 8-4 | 19-4 8 14
7 4-2 | 8-4] 15-4] 6-4 | 10-4 | 12-4 discojntinujed
8 4-2 | 5-4 | 14-4 | 11-4 | 12-4 | 13-4 4 8-4 6 8
. 9 7-2 | 32-4 | 12-4 | 8-4 | 16-4 | 14-4 discojntinuled
10 2 | 13-4 | 15-4 | 16-4 | 16-4 | 20-4 8-4 | 16-4 7 12
Set D 1.
TOTAL
FISSIONS
Conjugants of August 12, from the same culture as the non-conjugants .
of set D 2. Cultivated on slides from March 4 till June 7; from that time in watch
glasses. (1 a and 1 b constitute the two members of pair 1, etc.)
3-2 |23-4]| 6 8 8 5 | 5 DA 8
22 | 46-10) 9 12 |discojntinuled (8)
a 8 zl Mg | 7
10 d 8
1b d | | 2
2a } Ml 0
7 | 77-10) 19 | 67 |disco|ntinuled | (13)
file er 84a? d 6
100) P1440 5 1g} ||, 42 Cap Si a8 8 11
a 6 | 10 | 10 | 10 12
a 10 12 12
5 | 10 | 11 12
3a | 0
3b 3 d | 5
d 3
6 | 9-4] 6 Te t26 A ise} iB} i! Te 12
Ns a ati 4) a 12
A || 36 1) 3@ |) 18 12
4 Go] ahh || rh 12
AS NAT. |p 5 12
\
TABLE 32 (ContinvEp)
Set D 2. Non-conjugants of August 12, from same culture as the conjugants of
set D1. Cultivated on slides from March 4, till June 7; from that time in watch
glasses.
AUGUST SEPTEMBER TOTAL
LINE FISSIONS
14 | 16] -20,] 25 -|-a7 |20) 4 fee a aaah a
1 g-2| 6-2| 4 1 1 d 6
2 8-9) S23 4a gl ae 4 4 3 1 d 6
mY | 0
4 gs-2| 2| 5 |134] 4 d 6
2} 42 | 4 | 94] 6 1 1 d 1
2 f4-2| 42) 5] ad | 6
|\42] 8-2] @ 6
6 d | 0
7 Ga) Oe ees 5
8 ARE ay Ber | 5
9 ee ae aR AN NE 5
HOS 201) e2an Ue S=2Gl Aa 28 a2 4 8 7 Cri ila4: Ba) ba 7
| NGS | een te Alea 7
| | « ‘
TABLE 33
Experiment 13 b. Paramecium aurelia. Comparative number of fissions in the
selected conjugant and non-conjugant lines, for five periods, between December
8 and February 27. (d= died out.)
CONJUGANTS NON-CONJUGANTB .
DESCENDANTS OF | DESCENDANTS OF
ela/S]a/3/3 Fla (8) a8] 3
LINE Al Tn i a eT LINE Se tick % b>) S =
ABE ae zi |5 pel a|2
£l1a|a@iea|\5|° Hlé@| elise) S|
e) 2 |: (gz 5| 2 £|2| 2 (22/2 | 3
a A ad a ALS|S 55 |
1 21} 11} 14) 9) 13) 46 1 20) 12} 18) 14) 12) 50
2 19} 12) 13) 9 44 2 21) 11) 14) 12) 9} 46
3 20; 9) 14] 12) 7 43 3 15) 12) 15} 12} 12) 42
4 18) 10 13) 14; 9) 41 4 17| 11] 14] 12) dj 42
5 14! 11) 15! 13} 10) 40 5 17| 8 16) 12! 111 41
6 TS SiS ed 37 6 15} 11) 15} 13} 11) 41
7 12} 10} 13) 15) 11) 35 7 19} 10} 11) 9} 9] 40
8 14) 8 13) 6] 5) 35 8 16) 11] 13) 11) 6] 40
9 14| 8 11) @ 33 9 15) 9] 15) 13) 10} 39
10 11] 9} 11; 10) 8 31 10 16} 11) 12) d 39
11 14) 7.9) 7 2| 30 11 12} 10) 14 12 8} 36
12 11} 10; 8&| d 29 12 13] 9) 13} 12) 12) 35
ils} 5] 10] 13) 12) 3) 28 I: 15] 9} 4) d 28
14 16} 7} 4| 5| d| 27 14 10) & 6 5| 3) 24
15 4, 8 10) 3] 3) 22 :
16 CO AP ral teaeid fe (0) PA
17 0| d
18 d
19 d |
20 d
7A d | |
384
385
EFFECT OF CONJUGATION
TABLE 34
Record of number of fissions by two-
Experiment 16. Pure strain E; conjugants.
day periods, for each of the 88 lines descended from 28 conjugating pairs, through-
For full explanation, see ‘Explanation
out the twenty-four days of the experiment.
of tables 34 and 35,’ page 346.
TOTAL 24 DAYS
SEPTEMBER
AUGUST
LINE
10) (2s, cha eee
—~ vam
Bawa Bea RS Oe a SSeS ses ee citeece ect
cs) MNNANTMMMNMANMHWMHOCOAOCSTOGOAOONANA AN © MD OD OD % O39 09 0 1 HOO OO He N
< OD OD OD OD OD CYD NMAHMMODCOC CO OCF Fw AN OM OD OD OD OD OD OD OD OD OD OD OD OD CY
sit Beslesicencal ca earcerea ice eaice He OS StS Sicko aA NI OD OO 0D OD OD ON NI OD OD OD OD OD OD CD
ry MMOD MDMOMMMAMMNNNDTODOTOTOONNATWMMANMNMDAN DAA © MM ON
S Hat MM MM OHO HAH RAAT HHT NNN NOD OD HO OD HOO HH OO HOD HH OD HH
boo (Sai Coie a TA aot oO So ao Sm nDNA OO OD OO SH OD OD OO NOD SH OH on oD
= HHAMAMNMMNMNNMNMNDNRA ON HH ONH OA TAA TH HHO OOD OD HH OO HN OD
~ ODA MOMAMMA HH HH NNNOCOCONATAAAAN O HH HO ON OD SH OD OD OD <H
Nn NNNOHMTANANNANNANANA TAN See NNN ANN Ne MONMON
oo NATTA NN TAT TAHNONNANANNATANNTANTOTANANANANANNOCA
a ANMMANANAMNANNANNMANANANA TS N NANNN NANNANAN
BL OCOCCONUNMHMANANMHANNNMNMOHOOCOCOSCOSCH rc Tr CN rt
H. S. JENNINGS
386
TABLE 34 (ContINUED)
a
a
— |
no eal oD Vo m@r) co Lael ~~ ™~ (ve) co
a 088 BABSBBARAR A BABKRASHSS BAARRARZA A B AS a
a
°
&
| 69 60 CL CL AN OD 69 0D OD LN OD HEY 09 9 OO O_O 6 09 OD HAI OD OD OD OD OD OD OD OD OD OD I OD HOD OD I OD
Zl MDM MANN MMMM MANNDANSOOOM 1) DS OD OD OD OD CD OD OD OD OD OD OD OD OD OD OD OD XH OD
% ae aay ae |
Z| AM MHAMANMDMOANAN MAMMHOOTNOMMMMMANNANANMMM ONO AQ aN |
ra S| 6D 09 CX 4 CD OD OD ED OD OD ED OD OD EI OD ED OD OD OO OO OD OD AI OD OD OD OD CI OD OD OD OD OD OD rH M4 OD OD OD I OD OD
‘| ae
Q Zl op 69 09 09 60 OF 6D OD EI OD OD OD OD OD OD HOD CD Rm tH HH 0D 00 OD SH MH MH OD 09 OD OD SH OD OD HOD OD OD oH oH
Ly Ss
Bolo} HO HOHHMMMHAMDMANMDMMMMOOCHHAWMMAMMANNDMANH OOOO ON +
J ON MDMOMON DAN MANNA TMNMNN MON CO sw Me HOON H OD OO OD RH Or A oO OD Nest OOon,
+ | oD OD H CD OD SH OD OD AI OD OD OD OY OD 0D OD OD HO HN OD XH ID HH OD OD XH OD OD OY OD OD HH SH OH cH OH OH OH
AL AANMANNANNNOAN Sal NNANNTONANTAAMMATATAONNTANANANTAANM
te NONCHHANANT HHH HONTANTANANNTDANTAANTAANNTATAAAANANANNTAAAAN |
Dm — ~ -
8 AL NMNMANANTAMANNANANAAN MMMMANNDOONANANANMANMMAAANAAN !
eB |———____— ibaa eee ria aed Pa
TRI NANANNAANNMDAMNNANNANANANNOANNANANDOCHONNTONANNANANAANAN
Q SP OWAC ap tee cates ea age ee ee Re ae ae Nr FL ie Minne ie er) am WSS ee it cee ge eer
7 SD ae REON os oy EAC NRE yOu ty Os Ok EA Na RT COMM ne |i Meo Ttie er eer i me ee a ee We a my, eee any iP a SS
| Te ee. Seam APL ERR A ee Sa RN AE. hed Prat Ae = tC he Dk Slee Mee ES Cee ee ee TEE here Eee |
MPR DK RR RK KKK KR RR KKM KKK BKK Om
Q 3 Q 3 fe) 3 Q 3 Q r fe) 3 ate 3 a 3 3 3 3 3 a 3
oO ; ' =, ae | oD > 12
=! 2 2 a Si pte == SS ee: So 2 aie eee
387
16
TOTAL 24 DaYs
TOTAL 21 DaY¥s
0/0 0
Record of number of fissions by
SEPTEMBER
SEPTEMBER
TABLE 35
For full explanation, see ‘Explanation of tables
OD OD OD OD OI GI OD Of OI OO OD ON OD OD ON On OANA HOM AN
ANANMAANNANAANMONMANNNAN OA A
MOMMA NMMOMMOAAANA NAM MANN DMA OM AA
NAANMANAAN MAHAN MM NAAN ON A OO ON OO OD ON
sH <H CD OI CD OD HOD Ht OI OD OD OI OI OD OD OO OD OI oH OD OD OO eH OD NY
tw | Hos HoH HoH 60 HOOD oH cH cH HOD 09 mt OD mt oH mH OO CN oH HOD 08
| AND ANDAAANDAAA AAA Ao oO A
EFFECT OF CONJUGATION
TABLE 34 (Continvep)
AUGUST
AUGUST
LINE
24 (40s eee Soe ee
two-day periods, for each of the 174 lines descended from 65 split pairs, throughout
the 21 days of the experiment.
34 and 35° page 346.
Experiment 15. Pure strain E; non-conjuganis.
eee te |
7: TRS oe ele |e eT i
m= OI NN NN BANNAN SN eS ON me OA NSN me OOO ON OT
HOR TH HH HTH ON TR TAANAMNANAN NANDA HO
MDMA AMMANNMAMARADMDANMDMANRN AAA
a a ee ee ee ee ee
me ein eg Rainy SA - ray mg (ime em Ble
ol a“ on =H wD © ~
H. S. JENNINGS
388
TABLE 35 (ContTINUED)
m
b
A
co 4 1 oH MIDNA OOMNAHAMMMHANMOD HO sH xt xH OD Te) oD 1 OD <H Wo)
: AQ “A aN AAANRARARAAARANANRAATAN ANNAN aa ANNAN AQ
&
°
&
SB] cx op 0D oD OD 09 CL EI OD EL ELEN OD EY OD OD OD OD OD OY OLED OD HI OY OD 09 09 OD OD OD OD OD OD EL OLED VOD NN OD OD
Sl ONNANANMANMMDNMDNMANANANANANANTAANANANTANANANANANANAMANANAAANAN
Be NL AMANMMANMNNMANANNAANAMDMDMDMDIAANMDMAMAINMANADADMANANMDAAN
Hs i
g SIL MMMMANMDMANAMDMANMAMMAMNAMNAMDMAMANMDMAANANANA MONA OO AO OO A OD OD OD AN OD OD OD OD
=>)
a Ol] HHMOANMDMDMAMANAMAMADMDMMDMAMMADNM MMOD MDN AN OO AN HO OD OD OD OD OD OD OD OD OD
n
wo] 0 OD OD OD SH SH OD SH OD SH OD OD 19 H OD OD OD OD SH OOD HH OD OD OD H HH OD OD OD OD OD HO HH HOH HH SH SH OH HH OOD oH
oo | oD AMNANAMANAANANMMANNAMMANANANMANANNTANDMANDANANAAMD OH
al; ano NANNNNTAAMNOCHAANTANANNANNOCANA MANO ANNTAATDANAS
= Raid E
3 SBlLaANNNTO NANTANANANANRNTAONAANS NANMNONN oo oD 4 OD NN oD
oO
RB lA] DMONNAMMANANANAMMDNANNOAANNANNAMNDNANDONMASDOAMOOWN
es CMR tans Ste, ae diy Sen ate 5 MAR a ae eae Oe a Sa eee le i I ee a RAE SS a hw. Fs
8 Shy Tie ie. ai it eee ee aie a Se eS ee ee ae Eig a er aor. Ss oe ane meee — cee eee Cee ae
MK KM AK KR KR RMR KKK RK KKK Ow OM
Q 3 Q 8 Q a Q 3 eo 3 Q 3 2 3 fe) 3 Q 3 2 a = 3
™~ (oa) for) o 4 “N on H Ve) we) ~ 00
nm m= re re re So | — a _
389
TOTAL 21 pa¥s
NI OD OD OD OD OD OD OD OD OD
oD OD OD
1
SoM aR is aaCigoC so. FOLIO: boi aliocet sn bao soe Acie bere kOe
AAAAAAACAAADAAAAAAAAAAA
NNUNNNNMAANNAN NN OD
DODO ADA DDO AAMANAAN OANA
oD
ine)
NON MH M MO MON N
N
12)3)2
13;2/3|/2)3
OM ON HON NAN CL LY OD ODTED ON ON OD CY OD 09 OO EON OD NN OD NN 08
PROTA O CRCRO TGC ROC EDS Bock ube Naridor kein occ
EFFECT OF CONJUGATION
TABLE 35 (ContINuED)
sH sH CD OD <H
09 60 09 HO 09 HOD OD OD <H aH cH cH HH HOD HHO
N
“4
OD OD OD OO MI OD OD OD OD OD SH OD rN OH
sH
N
Op CIES. BIS
N
AATAADANAAAAAAAAA
N
baal
rt
SCONN NR ANMOMN
RE exes chee ys ae Oe ae
1D) SC ae a ie a A
oD
Pos a as TS i MSF GMa os sik io ion ie Digs os
NAN AAHANSHANNSANS
09 69 09 60 69 NOD CD HN CY ON OD OO ON OD OO
TOTAL 21 DAYs
23
25
26
24
23
24
26
27
25
22
24
24
25
27
19
23
23
23
26
25
23
24
24
25
22
27
25
24
25
26
26
23
21
23
26
H. S. JENNINGS
390
TABLE 35 (ContTINvED)
SEPTEMBER
| AUGUST
LINE
eS NM MM MN ON oD oD o> | ANNAN ror an a rere co 8 on 09 0 ep c0 co om oD eoUeD ON rrr
a NAAMDANNANARNANANHAANANAAANANANAANAMAMANAANAAAAAN
im | cag em 65 Ganeb PBA ee. ta Gb ESTO CACO GU NUE CS, om GOON BLL Gace CULES GCN GUGacaN ow Gd culed GuLea- gy ea one
= GU_CY CY 6D CY COD GI ED OO CY GD EOD OD 9 ED.D TLE OD LGN OD GO CY CY CY OD 09 69 09 OD 69 cD OD OD AOD OD C2 OD OD OO
« | Note Memeo oo os oo 9 6) ON ON 6 O0 Cl 60 MOM OADM MMA MDM MDH AMON To
Se ee ee ee
9 Bi eo cp oes POL AN eS OLN GULGN GN GGA cm co Gn co LQU-CNCUDON GG OP AN oD SalGN cu cn GLLGWEOLQILGR GL Ga Op eq eS
4 a ALAS a ae An MAAN AGA NO Ol OYPGT AL. oo Pee GOLAN GN ON-ea coca aN
3 a NANHHMANNANAANAN Er eae rr eon oe ee
B | 6 60 CU GON HALAL ENT HH NED ALL LAN EN GN ED AN AN 6 6 AW GN ED ED OD GL ED OD LCL CL LAL AN &
bao By ii MPa BA es cd a ad Be cbs a be) bs Bes bs Pook Ps a acd Pcie: Bs Be fe es
3 CsI 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 a
2 5-8 8 Se eee Se ee
391
EFFECT OF CONJUGATION
TABLE 35 (ContTINUED)
@ |
os] |
4
= 1 HOH O1In MID N © O10
ay a RARARANANAAAAA |
a |
&
fo}
is
am = = aif
©) ow ANMAMMMMMMNA MONO
IMI AINANANANAMAMAMAN
we || ANNANMANNAAANARAAA |
es
B 1S | cm 00 op 00 I MN OD OD OD OD OD OD I OD OD OD
= pe: seed
Sy Ae AI 0D 0D 00 00 OD HOH NI OD NH HH OD OD OD
tak ap Fokus > |
wo | HH oD OD H OD H OD OD OD OD OD OD CH OD OD
DIL OMMNANNMMANANMMAMANAN
IH | ANAM ANANATANANANAN
a
B Bi MANANATAANNANNANABDHAAN
S :
PIRI AMONANNANANTANAAA
g a Te ED AN a Men ARO BRO Bs eC
4 Ope ie Cote gO ie oh eee ie hehe Se a. EAS Go
-
4 ee ere Peay ek he eke Ay Me «SMa cmt
MPM DM eM mM Mm Om
3 3 3 3 3 3 3 ey
oe) > =) = a oD be 10
Ye) Yes Yo) © © © To)
© ou. 14, No. 3
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, V
BIPARENTAL INHERITANCE AND THE QUESTION OF
SEXUALITY IN PARAMECIUM}
H. 8. JENNINGS AND K. S. LASHLEY
The Zoélogical Laboratory of The John Hopkins University
TWO FIGURES
CONTENTS
PSM e see Denys EAS Se, EIN NETO GGhs: a7 ennai A I ne SO tty oo eS ae 394
PEO GAT CIO Meyer erat or Pa AE CM ITT an or ah tees on ces irk ONE A eet 394
Metnodof analysis thormulae 4.) My. ts se add see Bes alee ened 394
Anakyais: of therobserved results. 0s. 462 oe Vedic hs co ks dawclios As dca. 410
Distribution of survivals and deaths among the members of pairs......... 410
MS ERC IC SHER PELUBIONES oe ar 3.0 SER a SA hase sda ata oie APS ayssae ae et Pe 410
SO UIT REX POLITE MLS tert cs hee ere < ee esau oe OMS wojace ene he aa 5 apes 415
PP XPerimMents: Of Ge AULMOTor «tea. Yes. Pte sarod, s hed oe bo 4 Ov ole w aleteees he 415
Summary on the relation of the survivals and deaths to pairing.......... 420
Comparative rate of reproduction in the two members of pairs............ 422
ESUNOO ROR ana ENR d are” Ore a klar, fee teclt> ty, Ue Bree 7 ae CE ae 422
NMitsse@ ulleskexp erinteniGeerede a 1s ihe anus se ore See ee a
REpernimMentsy by Line AMOR 27 ei. eae Pace Srinlnchle de oa be ee ee 424
Pap abeny a sIMherebaneenacei: cx ars bes bes ckaes ects y Sceteadho sg soe nal ae ce Ae 428
Analysis OL thevexperimental results...5. ..h< 00 ooh cee eva Hs awe Soe Denes 428
Patio by HS. JONNINGS AtipeK. SO. ASHLEY... .¢. 00.2005: -ah eee Oe
Experiment 16: 241 pairs of conjugants propagated separately for forty days 433
Distribution of survivals and deaths among the members of pairs........ 436
Comparative rate of reproduction in the members of pairs............... 439
Experiment 17: 239 split pairs propagated separately for twenty days...... 443
SULVvvalsrandkdeaghsmiumet Mens polite p aise mere ees pares ere) eee 445
EME COUT ON GALT yes toes Neves cee ocd kat cae Ite h tneytenees Fey CLR Ee a egal tl 445
Distribution-ot survivals and. deaths. 7. ..00 ct sa encivies ee eae oe 7
SS UTA TILT Ae Mee oe NA eT eae ee ess RS coe aR 449
Comparative rate of reproduction in the members of split pairs........... 450
1'The results in the present paper are based to a certain extent on an analysis
of data presented in the tables of the preceding paper, on The Effect of Conjuga-
tion in Paramecium (this Journal, vol. 14, 1913, page 279). To facilitate reference
to the tables in that paper and to avoid confusion, the tables in the present contri-
bution are numbered consecutively with those of the former paper, beginning thus
with table 36. All references to tables numbered below 36 are therefore to those
in the preceding paper.
393
394 H. S. JENNINGS AND K. S. LASHLEY
PART Tih” By H> So JENNINGS {5 2200e ie os oa eee oe eee 451
Conjugation within a pure strain: Experiment 15................9........ 461
SS HREATTIRTY 6 ocala Wis oss «wine SN die wate Reels se ae 457
General SUMMATrY .. 2.06 6.6< Heys oes SR Gon ea. oP eee eee ee ee 458
Witerature Cited. «occ. ccc ok oeew a ue Pe os le Se 460
Appendix:
Rormulae used inithejanalysis:je.2¢ ss. eee a ree ee eee eee
Mable Si... 2 ncdcwinccbs tlic 0G. Lee woe ale 2 ee cone eee 463
PART I
H. S. JENNINGS
INTRODUCTION
When the individuals of Paramecium that have paired are
kept separate and allowed to reproduce under favorable conditions,
as a rule a considerable number of the lines of progeny produced
by them either die out, or reproduce very slowly; while the others
live and reproduce freely. This has been studied by Calkins
(02) and by Cull (’07) and is held by them to indicate that there
is at least an incipient sexuality in Paramecium. Of the two
members, a and ), of a pair, one is held to reproduce freely, thus
corresponding to the female, while the other, reproducing little
or not at all after conjugation, represents the male. Since the
life of individuals that do not reproduce is short, the ‘males’ will
in many cases die in a brief period after conjugation.
METHOD OF ANALYSIS: FORMULAE
Now, of course the mere fact, in itself, that a considerable
number of the lines of progeny are weak or die out after conjuga-
tion does not show that there is a tendency to sexual differentia-
tion in the members of pairs. There might be causes of weakness
or death that are quite independent of such sexual differentiation.
How can we determine whether the observed cases of weakness
- and death do indicate a sexual differentiation?
For answering this question the following considerations apply:
If the causes of death are quite independent of sexual differentia-
tion between the two members of the pairs, then sometimes both
members of a given pair will die or be weak, and it should be pos-
BIPARENTAL INHERITANCE IN PARAMECIUM 395
sible to determine how frequently, on the average, this will occur.
This would enable us to determine further in how large a propor-
tion of the cases both members of pairs should be represented
among the survivors, in this case where the distribution of deaths
has no relation to the pairing.
On the other hand, if the weakness and death are due to the
fact that one member (the ‘male’) of a pair does not reproduce
well, then it would be unusual for both members of a pair to show
weakness or death; from this cause acting alone it would never
occur. There would probably, of course, be acting also other
causes of weakness or death, that were independent of the pairing,
so that we should sometimes find that both members of a pair
die or are weak, owing to chance causes, but the number of cases
where both members die would be smaller, in proportion to the total
number of deaths, than if the distribution of deaths were purely
random so far as the pairing goes. For one of the chief causes for
the deaths would affect only one member of each pair. Thus
among the survivors there would be a greater proportion of cases
where only one member of the pair exists than would be the case
if sexual differentiation played no part in the matter.
This gives us a method of testing the matter. If, owing to
sexual differentiation, one member of the pair is more liable to
weakness or death than the other, then the number of cases where
both members of the pair survive will be found less than would be
probable if the distribution of deaths were due to causes that were
independent of the pairing.
To take a concrete case, Miss Cull (07) found that out of the
progeny of 93 pairs (186 lines), there remained at the end of a
month 103 lines, and among these 103 lines were representatives
of both members in 38 pairs. Is this number of entire pairs
still living less than would be expected if the distribution of
deaths had no relation to the pairing?
The same question may be dealt with by taking up the cases
that die, in place of those that survive. Thus, in the case cited
above, out of the progeny of the two members of 93 pairs (186
lines), 83 lines had died out, including 28 cases where both mem-
bers of the pairs had died. Is this number of pairs dead less than
396 H. S. JENNINGS AND K. S. LASHLEY
would be expected if the distribution of the 83 deaths had no
relation to the method of pairing?
It is important to realize that these two ways of putting the
question are identical. The probability that 38 complete pairs
should survive is in this case the same as the probability that 28
complete pairs should die, and this is true in general. Therefore
we need deal explicitly with only one of these questions; the answer
we obtain will hold for both. This will be demonstrated later.
Miss Cull (07) appears to leave out of consideration all pairs
in which both members die, dealing only with those which survive.
At the end of a month, since 28 pairs out of the 93 had died out, it
follows that there were left representatives of but 65 pairs, and it
is only these 65 that she considers in summing up the evidence in
favor of sexual differentiation:
It may be broadly stated that of the sixty-five pairs which I have
observed one conjugant either died or left a weak strain in which the
descendants were half as numerous and much less vigorous’ than those
of the stronger ex-conjugant. . . Here we have indications
that one gamete gives up its vitality to and loses its identity in the egg
where its presence forms a stimulus to development analogous to the
rajeunissement and greater activity in cell division which follows con-
jugation. There is little reason to doubt that a physiological and per-
haps a physical difference exists between the two unicellular organisms
which unite in conjugation and a difference of the same nature as that
expressed morphologically in the case of Adelea ovata, where the male
gamete does not fuse with the female, but dies after delivering one of
its four pronuclei (pp. 88, 89).
Now, even if we deal only with the survivors, as in our first
way of putting the question, we come to the same result as when
we deal only with those that die (as set forth above, and as will
be demonstrated later). But in any case, the results given in
my paper on the effects of conjugation (’13) show that there is
no justification for omitting the cases in which both members of
the pair die. Miss Cull did this apparently on the ground that
these were instances where conjugation was unsuccessful in pro-
ducing rejuvenescence, so that the animals died as they would
have done if there had been no conjugation. She says of the death
of these 28 pairs: ‘‘These facts confirm Calkins’ observation
BIPARENTAL INHERITANCE IN PARAMECIUM 397
that conjugation is by no means always successful in producing
rejuvenescence” (p. 87). But I have shown in the paper (713)
which precedes this one that there is no ground for supposing
that these would have died if conjugation had not occurred; in
our Experiment 1, for example, none of those that were prevented
from conjugating died, the deaths being limited to those that did
conjugate; and the other experiments give evidence in the same
direction. This takes away all ground (if there ever was such
ground) for trying to exclude the cases in which both members of.
the pairs died. What we must inquire, so far as deaths go, is,
whether the number of complete pairs that survive (or, if we pre-
fer, the number of complete pairs that die) is less than would be
expected if the deaths were due to causes that had no relation to
the pairing.
To answer this question, we must know, first, how many com-
plete pairs would probably have occurred among the survivors
(or the non-survivors), if the deaths had taken place at random.
In Miss Cull’s case, cited above, we must ask: How many com-
plete pairs would have survived among the 103 lines (out of 186),
if the deaths had occurred at random? Would the number have
been greater or less than 38 (the actual number)? Or: How
many complete pairs would have died among the 83 deaths (out
of 186), if the deaths had no relation to the pairing?
To deal with this and similar cases, we are compelled to take
up this general problem: Suppose that we have a given number of
pairs, from which a given number of individuals are drawn at
random; what is the most probable number of cases where both
members of pairs will be drawn? We may realize such a case
concretely by throwing a lot of serially numbered tickets into a
hat, there being two tickets bearing each number (these constitu-
ting a pair), then drawing out a certain number of the tickets.
What we wish is a formula for determining how many entire pairs
(both members) we shall probably get for any given number of
tickets taken from the hat. This will at the same time determine
how many pairs will be left. The question may be putalgebraic-
ally thus:
398 H. S. JENNINGS AND K. S. LASHLEY
Given m numbers, forming pairs; from these n numbers are
drawn. What is the most probable number of entire pairs that
will be drawn? And what is, consequently, the most probable
number of entire pairs that will be left?
This problem I did not find explicitly taken up in any of the
books on probabilities which I consulted. It may be attacked
directly in the following way:
Suppose that the total number m is 20 (forming thus 10 pairs),
so that we have in the hat the series 1 to 10, twice repeated. Now,
if we draw out one number, obviously no pair will be obtained.
There are then 19 numbers left, and if we draw out one more,
there is one among the 19 that will, with the first one drawn, make
one pair. Thus the chance for getting one pair when two mem-
bers are drawn is in this special case 1/19, and in general, it is
1
m-—l.
That is, if we repeat the process of drawing 2 from 20
av 1
a great number of times, we shall get a pair in 1/19 (or Sm i)
of all cases, while we shall get no pair at all in the remainder, or
—2
18/19 (= aa of all the cases.
m—1
Now, consider the case where one more number is drawn, mak-
ing 3. Before the third one is drawn, there remain 18 (or m—2).
Now, as we have already seen, a pair will have been obtained with
rai of all cases; in these cases
no additional pair can be obtained when the third is drawn. But
in the cases where the first two drawn did not form a pair (that is,
nD)
in 18/19, pe of all cases), there are two numbers out of the
18 (or m — 2) remaining that, with the two already drawn, will
form pairs. Thus there are now two chances out of 18 for getting
a pair when the third number is drawn. But this is true only for
78)
18/19 (or = a of all cases. So the total chance from this
the first two drawn in 1/19 (=
2 mM
a ai) hicks
m—l1
ae 18
source for drawing a pair is 2/18 of io (or
9 m—2
BIPARENTAL INHERITANCE IN PARAMECIUM 399
equals 2/19, ( or ) We had found from the previous source
m—1
a chance of 1/19 for a single pair, so the total chance for one pair
when 3 are drawn is 1/19 + 2/19, or 3/19 (in general, z: :
—]
2 3 ;
——- = . The remainder of the chances ( 16/19, or
m—1l1 m—1
paid
Ae) are for 0 pairs when 3 are drawn. Thus, when we draw
3 from a given number m, and repeat the drawing many times,
e
the average number of pairs obtained will be
By a continuation of this line of reasoning, which becomes
somewhat complicated as the number drawn becomes larger,
we discover that the average number of pairs that will be obtained
when n units are drawn is :
1424344... . to(m—1)_ }n(u-1)
eT | ~ m—l
Thus, when 9 are drawn from 20, the average number of pairs to
be obtained, if the drawing is repeated an indefinitely great num-
ber of times, is
H(O5<8) 86" 17
i 19 19
This average, like any other average, would be obtained in a
concrete case, by multiplying each number of pairs by the num-
ber of times it is drawn, and dividing by the number of drawings.
Thus, in the case given above, if there are a great number of
drawings of 9 from 20, the various number of pairs will be ob-
tained in the following proportions:
0 pairs in 384 drawings (= 0 pairs)
1 pairs in 3,456 drawings (= 3,456 pairs)
2 pairs in 6,048 drawings (= 12,096 pairs)
3 pairs in 2,520 drawings (= 7,560 pairs)
4 pairs in 189 drawings (= 756 pairs)
12,597 drawings 23,868 pairs
400 H. S. JENNINGS AND K. S. LASHLEY
Thus a total of 23,868 pairs will have been obtained in 12,597
V7 :
drawings, giving an average of 1 19 (or 1.8947) pairs for one
drawing. In an actual case of 100 drawings of 9 from 20, the
average was 1.91 pairs.
This gives us our first formula. If we let k be the average
number of pairs, our formula for determining it is
in(n—1)
m—1
k= (1)
As a rule this average number of pairs gives also the most
probable number of pairs that will be drawn—this being the
integral number nearest the average. In the above case, for
example, the most probable number of pairs when 9 are drawn
from 20 is 2. In Miss Cull’s case, already cited, where 83 died
out of 186, the average number of pairs included would be
4 (83 X 82)
185
(in place of 28, the number actually found). (Five actual draw-
ings of 83 tickets from 186 gave respectively 16, 18, 18, 18, 20 pairs,
the most frequent being 18, as our formula demands.) For most
practical purposes formula (1) is quite adequate for determining
the most probable number of pairs.
But it happens in rare cases that the integer nearest the aver-
age is not the most probable number of pairs. It is apparently
always within 1 of the probable number, and this is usually a
sufficiently close approximation, since the probability will be
nearly the same for the two numbers that differ only by unity.
To take an example, when we draw 5 from 20, the average num-
(5 X 4)
ber of pairs to be obtained from repeated drawings is ae
10
- ; fo :
Bet Now, the nearest integer to 198 1, yet complete analysis
shows that 0 is slightly more probable than 1 (the relative proba-
bilities of 0, 1 and 2 pairs are in this case as 168, 140 and 15).
. or 18.39, so that the most probable number is 18
BIPARENTAL INHERITANCE IN PARAMECIUM 401
For a formula which for even numbers drawn gives correctly
the most probable number of pairs I am indebted to Dr. A. B.
Coble, of the Mathematical Department of The Johns Hopkins
University. This formula is (if & represents the most probable
number of pairs):
(n + 1) (Ge 1)
k= m+3 (2)
In this case the nearest integer below the result gives the most
probable number of pairs (if the result is itself integral, then this
and the integer below it are equally probable). Thus, if we draw
10 from 20, the most probable number of pairs is given by
so that the most probable number of pairs is 2. But this formula
is not available when odd numbers are drawn. We shall see later
an indirect method of determining with absolute certainty the
most probable number of pairs when an odd number of units is
drawn.
But it is often needful to know what is the relative probability
of a given number of pairs being drawn, even though this may
not be the most probable number. For example, in Miss Cull’s
case, cited above, we found that the most probable number of
entire pairs that will be included when 83 out of 186 die is 18,
while the actual number of pairs included is 28. What is the
probability that we should get 28, if the distribution of deaths
has no relation to the pairing?
For a formula to determine the probability of any given num-
ber of pairs when a given number of units is drawn, I am again
indebted to Dr. A. B. Coble, to whom I wish to express my
thanks. If x represents the probability of any given number of
pairs k, then Coble’s formula is as follows:
n! m—n! 1! 222k
== 3
mikin—2k!l—-n+k! (3)
x
402 H. S. JENNINGS AND K. S. LASHLEY
where
m = the total number of individuals
1 = the total number of pairs (so that 1 = } m)
nm = the number of individuals drawn
k = the number of pairs whose probability we desire
n! = the product of all integers up to n (thus 186! is the product of all integers
from 1 to 186)
This formula forms the basis for further formulae which I
have developed from it (notably number (4), given below),
and it is indeed the basic formula for the greater part of our work.
It may be employed to determine directly either the probability
of any given number of pairs among those drawn, or of any given
number of pairs among those not drawn.
In a concrete case the formula works out as follows: Suppose
that out of 20 individuals (10 pairs), 9 individuals are drawn.
What is the probability that there will be just 3 pairs included
among those drawn?
Here m = 20; n = 9; k = 3. The formula (3) therefore be-
comes
9111! 10!23
~ 90131314!
which gives x = 0.2005, so that there is almost exactly one chance
in five of getting 3 pairs when 9 are drawn from 20. We may put
the question in the reverse way by asking: What is the probabil-
ity of there being left 4 pairs when 11 are left out of 20? (If
9 are drawn from 20, and 6 of the 9 from 8 pairs, then the other
3 belong to other pairs, so that 6 pairs are represented among
those drawn, leaving 4 complete pairs among those not drawn).
In this case m is 20, n is 11, while & is 4, and formula (8) there-
fore becomes
11!9110! 28
~ 90141313!
which is identical with the formula (given above) for 3 pairs when
9 are drawn from 20. Thus, the probability of the number of
pairs actually drawn is of course bound to be the same as that for
the number of pairs actually left, as was previously mentioned.
BIPARENTAL INHERITANCE IN PARAMECIUM 403
Determining by formula (3) the probability that there should
be 38 pairs among the 103 left from the 186 lines in Miss Cull’s
case, above cited, we find this to be but 0.00004556; or the odds
against this number are 21,948 to 1.
The result at this stage however does not give us what we need
to know. If 21,949 tickets were placed in a box and one of these
drawn out, the odds would be 21,948 to 1 against any particular
number, yet some particular number would be drawn, and there
would be no ground for surprise that it should be one particular
number rather than another. Our present case is not entirely
similar to this, since a certain number (18) of pairs 7s more proba-
ble than any other. Yet formula (3) shows that the probability
for this most probable number is but 0.16685, so that this precise
number of pairs would be found in but one case out of every 6,
and for any given case the odds against it are 5 to 1.
What we require to know is, not only what is the most probable
number of pairs, and the deviation from this most probable num-
ber, but (and this is the essential point): How probable is it
that there should be so great a deviation from this most probable
number as that which we find?
Thus, in Miss Cull’s case already cited, the most probable
number of surviving pairs we know to be 28, while the actual
number surviving is 38, so that the observed number deviates
by 10 from the most probable number. We require to know how
probable it is that there should be so high a deviation as 10.
The probability of a deviation so great as that observed may
be determined directly as follows: Determine the probability
of all cases showing a deviation less than the given deviation
and compare the sum of these probabilities with the sum of all
cases showing a deviation as great as the given deviation. In
Miss Cull’s case, where the most probable number of surviving
pairs is 28 and the deviation is 10, we should have to find the
sum of the probabilities for all numbers deviating less than 10
(that is, of all numbers from 19 to 37), and compare this with the
sum for all numbers deviating more than 10 (all numbers below
19 and above 37). 5
404 H. S. JENNINGS AND K. S. LASHLEY
Since the sum of all the probabilities is 1, it is practically only
necessary to find the probabilities for all numbers showing less
than the given deviation (in this case, for all numbers from 19 to
37) and subtract their sum from 1; the remainder will be the sum
of the probabilities for all deviations as great as (or greater than)
the given deviation. This procedure is that which in the Appen-
dix is formulated as rule (5).
For finding the probabilities for a series of successive numbers
(as 19 to 37, in the case above), the following formula (4) will be
found convenient. First find the probability by formula (8)
for the lowest number of pairs (in this case 19). Call the proba-
bility for this number x; Then the probability for the next
higher number of pairs (which we may call k»; in this case i» is 20)
is given by multiplying the probability 2, by
(n—2 rot il) (n — 2 ke+ 2)
4ke(1—n-+ ko) @)
The probability for the next higher number of pairs (in this
case 21) may then be found from this result by using the same
formula anew, and so on for the entire series of numbers. (Or
if we desire we may begin with the higher numbers, and find the
probability for succeeding lower numbers by using the expres-
sion (4) inverted.)
An example may be taken. In the case we have cited, we
find that the probability for 19 pairsis, by formula (3), 0.000059733.
To find the probability for 20 pairs we must multiply this value,
64 X 65
according t ich gi : 1061
according to formula (4), by ee el which gives 0.0003106
as the probability for 20 pairs. To find the probability for 21
: : 625<¢: 63 se
pairs we must now multiply this value by ote = x 11’ giving
0.001313 as the probability for 21 pairs.
Proceeding in this way, and adding together the value of the
probabilities obtained, we find that the total probability for all
numbers from 19 to 37 is 0.99995242; this then is the probability
that the deviation shall be less than the ebserved deviation 10,
if the distribution of deaths has no relation to the pairing. The
BIPARENTAL INHERITANCE IN PARAMECIUM 405
difference between this value and 1 gives 0.00004758 as the proba-
bility that we should have any deviation as great as 9. Dividing
the former value by the latter, we find that the odds against so
great a deviation as that actually observed are 21,016 to 1, unless
there is something in the pairing that causes the two members of
the pairs to be more alike in their fate (not less alike, as the theory
of sexual differentiation held).
It would be of interest and value if it were possible to find for
the present case some standard of value corresponding to the prob-
able error or standard deviation of the common ‘normal’ curve
of probabilities. It may be of interest to examine a curve for
such cases as those we are here dealing with. To get such a
curve, we must determine the probabilities for each of the possible
numbers ef pairs (in the way just set forth); then these may be
plotted on some convenient scale. In doing this, one or two sim-
ple considerations will aid. If the number drawn is lessthan
half of the total number (that is, if m is less than 4 m), then evi-
dently it is possible that all the numbers drawn should be differ-
ent, so that we must begin with 0 pairs. Further, the number of
1, if n is odd), but cannot be greater
than this. So in such a case we must find the probability for
n
ips
I have plotted such a curve for one of the cases described by
Miss Cull. She found that after twenty days, 51 lines had died
out, from the entire 186; and among these were 13 pairs. In
this case the possible numbers of pairs range, in accordance with
the considerations just adduced, from 0 to 25.
Determining by formulas (3) and (4) the probability for all
these numbers, we obtain the results given in table 36, page 412.
We employ these probabilities as ordinates. while the numbers of
pairs are laid off on the abscissa. This gives a curve or polygon
such as is shown in figure 1. The sum of all the ordinates is here
equal to 1 (or if we prefer we may take the total area of the poly-
gonasl). Itis evident that this polygon bears some resemblance
to that obtained from distributions following the normal law, but
pairs might bes (or
each number of pairs from 0 to
406 H. S. JENNINGS AND K. S. LASHLEY
differs from it; particularly in the fact that its upper limit is
much farther from the mode than is the lower limit. This isnot
characteristic of all polygons obtained in this way, but apparently
so only of cases where m and n are rather large. In man'y cases
we have, in place of such a polygon, merely a point, a straight
line, or a broken line, which may be of various forms. Some of
these are illustrated in figure 2.
In view of the form of the polygons obtained, it would appear
to be difficult to obtain any simple formula to express the proba-
bility of a given deviation: one must use such a method as that
set forth above (see rule (5), Appendix).
The procedure which we have just described of course makes
it possible to determine with absolute accuracy what is the most
probable number of pairs when an odd number of units is drawn;
a point which we left undecided in our previous account. When
this is the only question to be answered, it is done as follows:
Find by formula (1) the average number of pairs to be obtained
when the given odd number is drawn. Then find by formulae
(3) and (4) the probabilities for the two numbers nearest this
average; this will of course show which of the two is the most
probable.
Formula (3) gives unmanageable numbers when the operations
indicated are directly performed; thus 186! gives a number con-
sisting of 343 integers. In practical work therefore logarithms
must be employed. The logarithm for any factorial number
n! is of course the sum of the logarithms of all integral numbers
from | up to and including n (since n! is the product of all numbers
from 1 ton). With a table of such sums of logarithms the com-
Fig. 1 Polygon showing the relative probabilities for obtaining the different
possible numbers of pairs, when 51 specimens are drawn from 186 (93 pairs). The
ordinates give the probabilities in per cent, for each of the numbers of pairs
indicated on the base line. (These probabilities are the percentage of all draw-
ings in which the given number of pairs would be obtained, if the drawing of 51
from 186 were repeated a great number of times.) They are plotted from the data
given in table 36, page 412.
The polygon extends at the left to 0, at the right to 25 pairs, but the probabilities
for 0 and for all numbers from 15 to25 pairs areso minute that the ordinates donot
appear at all in a polygon drawn to this scale, so that in these regions the outline
of the polygon as drawn coincides with the base line.
BIPARENTAL INHERITANCE IN PARAMECIUM 407
22 23 24 25
19 20 21
OMe OMnti lien oan an 4S) ehOemnl yen 8
8
Number of pairs drawn
ee) PONE a COs Re UT Set nn CON Om os, Oe Sa an) Ser on
eS = — eS =
Probability, in per cent
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
Se!wiigeqos 4
Number of pairs drawn
Fig. 2 Examples of the form of graphs for the distribution of the probabilities
of the various possible numbers of pairs, in certain cases. The figure is designed
to illustrate the variety of forms of such graphs, and to show how little they resem-
ble, in many cases, the ‘normal curve.’ The ordinates are probabilities; the ab-
scissas are numbers of pairs, from 1 to 10. The graphs are drawn, like the usual
polygons of variation, by connecting the tops of the ordinates representing the
probabilities. :
The line A-B-C-D-E shows the probabilities for the different possible numbers of
pairs obtainable when 20 specimens are drawn from 1000 (500 pairs). From 3 up
to 10 pairs (D to E) the probabilities are so small that they do not show in a figure
drawn to this scale, so that the line representing them coincides practically with
the 0 line.
The remaining graphs (numbered 10 to 20) show all those obtainable when any
possible number of specimens is drawn from 20 (10 pairs). The numbers adjacent
to the graphs show the number of specimens drawn; thus the graph numbered 10
shows the probabilities for the various possible numbers of pairs (0 to 5) when 10
specimens are drawn from 20. As the figure shows, the graphs for 17 and 18 are
straight lines; those for 19 and 20 are points.
These graphs (10 to 20) are identical with those to be obtained when the numbers
0 to 10 are drawn from 20, save that in the latter case: (a) they stand in the re-
verse order, the graph for 1 being the same as that for 19, the graph for 2 the same
as that for 18, etc.; (b) all would begin at the left on the 0 line, so that they would
be crowded together. Otherwise they have the same form, slope, and dimensions
as those shown, for drawing 10 to 20.
408
BIPARENTAL INHERITANCE IN PARAMECIUM 409
putations become fairly simple. Such tables have been pub-
lished by De Morgan (’45) (six-place logarithms), and by Degen
(23) (twelve place logarithms of n! for all numbers from 1 to
1200), but I have not been able to obtain these. Pearl and Mc-
Pheters (711) have recently published a useful table of such sums
of logarithms for the numbers from 1 to 100. This table will be
found convenient when the numbers dealt with fall within it.
For use in the present investigation I have prepared such a table
up to 248!.
But the value of n! for any number may be found with sufficient
accuracy from Stirling’s formula. This, in form for practical
use, is as follows:
nD
7 V 20 n
en
Here:
e = 2.7182818 (log. 0.4342944819)
3 = 3.1415927 (log. 0.4971498726)
V2x = 2.506628 (log. 0.3990899342)
All the operations should be performed by the aid of logarithms
(even where a computing machine is available), the best course of
procedure being as follows:
(1) Find the logarithm of n; multiply this by n.
(2) Multiply log. 0.4842944819 by n, and subtract this pro-
duct from the result of (1).
(3) To the result of (2) add log. 0.3990899342.
(4) To the result of (3) add 4 the logarithm of n.
This gives the logarithm of the given number n!, which may be
used in formula (3). For example
186186
€ 186
186! = V 2x V 186 = log. 342.8844688
In using formula (3) it is of course necessary merely to add
together the logarithms of the factors in the numerator, and sub-
tract from the result those of the denominator; the result is the
logarithm of the probability.
410 H. S. JENNINGS AND K. S. LASHLEY
ANALYSIS OF OBSERVED RESULTS
The view that there is a sexual differentiation between the
two members of pairs has been based upon two sets of facts:
(1) upon the distribution of survivals and deaths among the
ex-conjugants; (2) upon differences in the rate of fission between
the progeny of the two members of pairs. We shall deal with
these separately, taking up first the deaths.
DISTRIBUTIONS OF SURVIVALS AND DEATHS AMONG THE MEMBERS
OF PAIRS
Miss Cull’s experiments
We may analyze first the data given by Miss Cull (’07).
Ninety-three pairs of conjugants, giving 186 separate lines, were
isolated; each line was kept in a vial. At the end of seven days
32 of the lines had died out, including both members of pairs in
6 cases (that is, ‘6 entire pairs’). After twenty days 51 of the
lines had died out, including 13 pairs. After 30 days 83 lines
had died out, including 28 pairs. We have then three cases to
deal with. Is the number of complete pairs dead in any or all
of these cases less than would be expected if the distribution of
deaths had no relations to the pairing, as would be required if
this is to be considered evidence for sexual differentiation? Or,
what is the same thing, is the number of pairs surviving less than
would be expected if the distribution of deaths had no relation
to the pairing? We need to deal explicitly with but one of these
two questions, since the answer is absolutely identical in the two
cases. As a rule therefore we shall take up the question only in
the form first stated.
First case. Here, after seven days
m = 186
n = 32
number of pairs dead = 6
Now, applying our formula (1) or (2), we find that the most
probable number of pairs, if the distribution of deaths is random,
is 2, although 3 is nearly as probable. The average number
BIPARENTAL INHERITANCE IN PARAMECIUM 411
of pairs that will be obtained if successive drawings are made is
by formula (1), 2.68. Five actual drawings of 32 tickets out of
186 gave respectively 4, 2, 1, 2, 4 pairs; the average being thus
2.6; very close to what the theory calls for.
The observed number 6 is therefore greater than would be
expected, not less, as the theory of sexual differentiation would
expect. By formula 3, the probability for 2 pairs is found to be
0.27233; that for 3 is 0.26808; for 6 is 0.02013. The observed
number (6) deviates from the most probable number (2) by 4;
we require to know what is the probability that there should be
so great a deviation as this. By our rule (5) (Appendix), we
find that the total probability for a deviation as great as 4 is
0.0248, while for deviations less than 4 it is 0.9752, so that the
chance is 39.3 to 1 against there being a deviation so great as that
observed, if the distribution of deaths is independent of the pair-
ing.
Thus, so far as this case goes the probability is very strong
_that the distribution of deaths is not independent of the pairing.
_ But its dependence is of a character the reverse of what the
theory of sexual differentiation requires; the number of pairs is
greater, not less, than we should expect from a random distribu-
tion of deaths. Thus it appears that the members of the pairs
are more alike, not less alike, in this respect, than would be antici-
pated. The drawing of further conclusions may be deferred until
other cases are examined.
Second case. After twenty days, Miss Cull found that 51
lines had died, out of 186, and among these were 13 pairs.
By formula (1) the most probable number of pairs is found to
be 7, the average number if a great number of drawings were made
being 6.892. Five drawings of 51 tickets out of 186 gave respec-
tively 6, 9, 8, 8, 8 pairs, the average being 7.8, which is very close
to theory; the excess of nearly 1 showing what may happen when
but few drawings are made.
The observed number of 13 pairs exceeds the most probable
number by 6 pairs.: What is the probability that there should
be so great a deviation as this from the most probable number of
pairs?
412 H. S. JENNINGS AND K. S. LASHLEY
In this case I have worked out by formula (3) the value of the
probability for all possible numbers of pairs; the results are given
in table 36.
From this table is plotted the curve of figure 1 (page 407).
The curve shows directly to the eye how great is the divergence
of the observed result from what should be expected if the dis-
tribution of deaths has no relation to the mating; the ordinate of
13, the actual number of pairs, is far from that of the most prob-
able number of pairs. By our rule (5) we find that the probability
of so great a deviation as 6 from the most probable number is
i
TABLE 36
Probabilities for all possible numbers of pairs, in the case where 51 individuals die
out of a total of 186.
NUM-
BER OF PROBABILITY
PAIRS
0 .0000676
1 -.0009644
2 0064441
3 0258003
1 069407
5 133352
6 189841
a .205061
8 170717
9 110648
10 .056170
11 022402
12 .007018
13 .001722
14 .000329
15 0000487.
16 -.0000055
17 .0000005
18 .00000003
19 -000000001
20 .00000000004
21 .0000000000009
22 -00000000000001
23 -00000000000000008
24 .0000000000000000002
25 .0000000000000000000002
BIPARENTAL INHERITANCE IN PARAMECIUM 413
0.00224, while the probability that the deviation will be less than
this is 0.99776. Hence the chances are 445 to 1 against a devia-
tion so great as is actually observed, if the distribution of deaths is
independent of the pairing.
Thus it appears that as further deaths take place, the tendency
for both members of a pair to die, if one dies, becomes greater;
after seven days the chances were but 39 to one against so great
a deviation as occurs, while after twenty days the chances against
the actual result are more than ten times as great as they were.
It is therefore clear that among those who died in the interval
between the seventh and the twentieth days there were a greater
number of mates of those that had previously died, than could
be expected as a matter of chance. This is clear when the exact
figures are considered. At the end of seven days there were left
154 individuals, including 134 that were still paired, while 20
were odd (their mates dead). Now, in the next thirteen days
there died 19 more, and of these 19, no less than 14 were mates of
those that had already died (or that died during the entire twenty
days), while but 5 fall to the 134 paired individuals! It is clear
that whatever is causing the death of one individual of a pair
has a strong tendency to cause likewise the death of its mates;
the mates are much more alike in this respect than are two lines
taken at random. And in accordance with what has previously
been set forth, we may make the same statement for the sur-
vivals. Whatever tends to cause one individual of a pair to
survive has likewise a strong tendency to cause the survival of its
mate—the numerical relations as to survival being identical with
those above set forth for the deaths.
Third case. After thirty days Miss Cull found that out of
the 186 lines, 83 had died out, including 28 pairs. We have al-
ready dealt with this case, but may here recapitulate the facts.
By our formula (1), the average number of pairs obtained when
83 are drawn from 186 is 18.39, from whence it may be concluded
that the most probable number of pairs is 18. The excess over
the most probable number has therefore now grown to 10 pairs,
though at the end of 20 days the excess was but 6 pairs. This
414 H. S. JENNINGS AND K. S. LASHLEY
shows that the tendency for mates to die continues into this
third ten-day period. At the end of the second period (twenty
days) there had been left 110 paired lines and 25 odd ones. There
die in the next ten-day period 32 additional lines, and of these,
no less than 30 are mates of those already dead, or of those that
die during the present period, while but 2 are mates of those that
still live at the end of this period!
As one would expect, the chances for such a result, if the dis-
tribution of deaths has no relation to the pairing, is infinitesi-
mally small. By formula (3), we find that the chance for getting
28 pairs is but 0.0000455623. The chance for getting any num-
ber whatsoever that deviates from the most probable number
(18) as much as does 28 is almost equally small. By our rule (5)
we find that the probability that the deviation from 18 will be
less than 10 (the observed deviation) is 0.999952421, while the
chance that it will be as much as 10 is but 0.00004758. Thus the
chances are 21,016 to 1 against there being so great a deviation
as was observed, unless the distribution of deaths depends upon
the pairing.
Thus a more complete demonstration that the distribution
of deaths depends on the pairing could not be demanded; but
the relation is throughout the reverse of that called for by the
theory of sexual differentiation. There is not a tendency for
one member of the pair to live and for the other to die; on the
contrary, there is a most marked tendency for both members to
have the same fate. The effects of this tendency show themselves
more and more strongly as the lines of progeny successively die
out after the pairing. Supposing that such a tendency does not
exist, the chances, after seven days, are 39 to 1 against getting so
great a deviation as exists; after twenty days the odds are 445 to
1; after thirty days they are 21,016 to 1.
It appears from the figures given by Miss Cull that there is
something in the process of pairing that causes the members of
pairs to become alike, so far as survival and death are concerned.
This will now be tested by other experiments and in other rela-
tions.
BIPARENTAL INHERITANCE IN PARAMECIUM 415
Calkins’ experiments
Calkins in his paper of 1902 gives some statistics bearing on
this matter. In his table 3 he gives records for the progeny of
40 pairs of conjugants, 24 being ‘exogamous,’ 16 ‘endogamous.’
These were not all taken nor cultivated at the same time, so that
it is doubtful how far we should attempt to use them for our pres-
ent purposes. However, some of the facts may be given, in order
to show any relation they may have to our problem. Calkins
notes that a large proportion of the lines die out before the tenth
day. We may then select this period for examination. It is
found from Calkins’ table that by the end of the tenth day 40 of
the 80 lines derived from the 40 pairs have died out. Here we
have m = 80; n = 40. The most probable number of pairs is
found, by our formula (1) or (2), tobe 10. The actual number of
pairs dead is 14. Thus in this case, as in those described by Miss
Cull, the number of pairs dead, and the number of pairs living,
is greater than would be expected; not less, as the theory of sexual
differentiation demands.
At the end of one month but five lines were living, out of the
80. Among those living there were no pairs; this was tobe
expected on any theory. :
The data given by Calkins perhaps hardly warrant farther
analysis for our present purposes.
Experiments of the author
In my paper on the Effect of Conjugation in Paramecium (’13),
which immediately precedes the present one, I have given an
account of a number of experiments which yield data for attack-
ing our present problem. I will here analyze these data, refer-
ring to the paper just mentioned for all details not bearing directly
upon the question now in hand. I shall refer to the experiments
by the numbers given to them in the paper cited.
416 H. S. JENNINGS AND K. S. LASHLEY
Experiment 1: Paramecium caudatum
This Journal, volume 14, 1913, page 286
In Experiment 1 there were 31 pairs under propagation; one
member of one of the pairs was accidentally lost, so that for our
present purposes this pair must be excluded from consideration.
We have thus 30 pairs to deal with. None of them died during
the first week, showing that no deaths were due to handling or
the like while the pairs were united. The facts are given in table
29 of the paper just referred to (this Journal, page 379).
It will be well to examine the condition of affairs at certain
intervals.
At the end of the second week, five lines had died out, including
one pair. By formulae (1), (2) and (8), it is found that the most
probable number of pairs is 0, and that the chances are 5.06 to 1
against any pairs whatsoever being included. That is, the chance
for 0 is 5.06 times as great as for all other possibilities together.
At the end of the third week 15 lines out of the 60 were dead,
and among these fifteen were 4 pairs. Formula (1) shows that
the most probable number of pairs is 2; and formula (3) shows
that 2 pairs will occur in somewhat more than one third of all
cases, while 4 pairs would occur but once in 25 cases, if the dis-
tribution of deaths had no relation to the pairing. The deviation
of the actual number from the most probable number is 2: by
our rule (5), we find that the chances are 5.96 to 1 against so
great a deviation as this, unless the distribution of deaths is actu-
ally connected with the pairing.
At the end of the fourth week there are 18 lines dead, including
5 pairs. The most probable number is 3 pairs, which would occur
in 31 per cent of all cases, while 5 would occur in but 4 per cent,
if the distribution is random. The deviation from the expected
number has not increased during the fourth week.
At the end of the experiment, five weeks from the beginning,
22 lines were dead, including 5 pairs, so that the number of pairs
has not increased. As set forth in the account of this experiment
in the paper referred to (page 287), the deaths during the fifth
week were due to mistaken cultural treatment; their distribution
BIPARENTAL INHERITANCE IN PARAMECIUM 417
had no relation to the pairing. The most probable number of
pairs when 22 lines die is 4; the actual number 5. The probability
for four is 0.306; for 5 it is 0.214, or about two-thirds as great as
for 4. Five pairs might therefore well occur purely as a matter
of chance distribution.
Thus in this experiment the number of pairs included among
those that died is a little greater than would be expected if the
deaths had no relation to the pairing. The relation is again the
reverse of that which the theory of sexuality reauires.
Experiment 2: Paramecium caudatum
This Journal, vol. 14, 1913, page 293
In this experiment, lasting for eight weeks, 56 lines derived from
members of pairs, 38 from split pairs, and 58 from unpaired
individuals, were cultivated. The designations of the unpaired
lines were arbitrarily grouped in pairs at the beginning, in order
to discover whether such arbitrary grouping would give the same
relations between its members as appear in the actual pairs.
The death rate was very high in this experiment, for reasons
set forth in the preceding paper (p. 293); on this account it does
not furnish the best of data for determining whether there is any
relation between liability to death and the pairing. It will be
worth while however to set forth the facts in the case; this is done
in table 37. I have included the facts, not only as to the pairs,
but as to the split pairs, and the free specimens. In the split
pairs, the two individuals that had begun pairing are called a
pair. In the free specimens, the pairing is purely artificial; these
are for control purposes.
As the table shows, in no case does the actual number of pairs
dead differ from the most probable number by more than a single
one. This is precisely such a result as might be expected in
chance single cases. The deviations from the number that is
absolutely the most probable one occur, as will be seen, in the
case of the free specimens, in which the pairing is purely arbi-
trary, as well as in the other classes. I have given in the table the
418 H. S. JENNINGS AND K. S. LASHLEY
TABLE 37
Death rate in relation to patring, in experiment 2.
MOST | PROBA-
WHOLE | PROBA-
ean | pIED | PAIRS DIED na seach et
LINES apes NUMBER peat
a = ee a.| fe = Le a4 : =
First week: |
PES ee La 56 di 0) 0 0.654 0.654
Split, pairs... 22 s!e"- 38. | 2, 0 0 | 0.973 0.973
Pres.gias 28 ee 58 | 2M 0 - 0) | 0.856 | 01856
First and second weeks:
Dani oie eee 5G 0 20 ae 3 | 0.168 | 0.534
Split pairs........... SG. UCaLy | 4 | 4 | 0.345 | 0.345
Bree. 235.0. hee ee 58a 28 6 TON 10:268, 0.281
Eight weeks: | |
RIT ert sae oes 48 39 | 17 | 16 | 0.189 | 0-449
Splipairs....2.>.<: 26 20™ | Saya 7 | 0.447 0.477
Free: ....:-ee.cd-icns{ ae 46°. | 0850. 43 Lo Mia | oer
probability for the occurrence of the most probable number, as
well as of the actual number, in order that these may be compared.
Thus the data of this experiment give no evidence for sexual
differentiation, nor for any other special relation of the two
members of pairs; they are what might be expected if the distri-
bution of deaths were purely random. This is not surprising, in
view of the high mortality; the deaths were evidently due mainly
to extrinsic causes.
Experiment 3: Paramecium caudatum
This Journal, volume 14, 1913, page 300
In this experiment there were 46 lines, derived ftom 23 pairs.
Out of these, 11 lines died out during the four days of the experi-
ment. The number of pairs among these was 1, and this is the
most probable number that would be found, if the incidence of
death had no relation to the pairing.
There were also in this experiment 50 lines derived from 25
split pairs; of these 35 lines died out in the four days. Among
these were 13 split pairs. Applying our formulae, we find that
the most probable number of pairs is 12, although 13 is nearly as
probable. The probability for 12 is 0.369; for 13 it is 0.260, so
that 13 would occur little less often than 12.
BIPARENTAL INHERITANCE IN PARAMECIUM 419
The mortality in this experiment gives then no indication of
sexual differentiation, nor of any other special relation between
the members of pairs. The very high mortality, taken in con-
nection with the conditions of the experiment, shows that the
causes of death were mainly extrinsic.
Experiment 4: Paramecium aurelia
This Journal, volume 14, 1913, page 304
In this experiment with the race k, there were included 40 lines
derived from 20 complete pairs (2 other lines having lost their
mates through accident). Of these, 23 lines died out during the
twenty days of the experiment, and among these were both mem-
bers of 8 pairs. The most probable number of pairs when 23
die out of 40, is 6, if the deaths have no relation to the pairing,
although 7 is nearly as probable. Thus the number of pairs that
died is greater than would be expected if the deaths had no rela-
tion to the pairing; not less, as would be required in order to give
evidence of sexual differentiation. ‘The excess is however small,
the probability of getting eight pairs being 0.144, so that this
would happen in one case out of seven, purely as a matter of
chance. No positive conclusions could therefore be drawn from
this result, taken by itself.
Experiment 5: Paramecium aurelia
This Journal, volume 14, 1913, page 308
In this experiment on the pure strain k, lasting six days, 8
lines died, out of 16 (derived from 8 pairs). Among these were
both members of 3 pairs. The most probable number of pairs,
if the deaths are distributed at random, is 2. The chance of 2
is 0.522; that of 3 is 0.174.
Among the 9 lines that died from among 20 derived from 10
split pairs, there were both members of 2 split pairs. The most
probable number, on a chance distribution of deaths, is likewise 2.
420 H. S. JENNINGS AND K. S. LASHLEY
Thus the results in this experiment are, so far as they go,
against the hypothesis of sexual differentiation, but do not give
definite positive evidence of any other relation between the
members of pairs.
Experiment 13: Paramecium aurelia
This Journal, volume 14, 1913, page 329
In this experiment, carried on with the pure strain k, there were
78 lines derived from 39 pairs of ex-conjugants; these were culti-
vated for only a week. During this time there died out 20 lines,
including 4 pairs. The most probable number of pairs, with a
distribution of deaths having no relation to the pairing, is 2. The
probability of 4 pairs is 0.142, so that they would occur, as a
result of chance, once in seven Cases.
Thus the number of pairs that died is here, as usual, greater
than would be expected if the distribution of deaths had norela-
tion to the pairing.
Summary: on the relation of the survivals and deaths to pairing
In order to give evidence of sexual differentiation between
the members of pairs, it is necessary that the number of cases
in which both members of pairs die or survive should be less than
would be expected if the distribution of deaths had no relation to
the pairing. In no case, among those of which statistics are
available, as set forth in the preceding, is this the state of affairs.
The nearest approach to it is in my experiment 2 (page 417), where
after two weeks the number of pairs dead was one more than the
most probable number, though at the end of the experiment the
number dead was one less than the most’ probable number. In
all other cases the number of pairs that died and the number that
survived, was equal to, or greater than, the number that would be
expected if the distribution of deaths had no relation to the pair-
ing. In most cases it is greater than would be expected.
Thus the distribution of survivals and deaths gives no evidence
of sexual differentiation between the members of pairs.
BIPARENTAL INHERITANCE IN PARAMECIUM 42]
On the contrary, the fact that the number of pairs that sur-
vive (or of those that die) is as a rule greater than would be
expected as a result of a random distribution of deaths indicates
that the two members of pairs are more alike than individuals
taken at random; that there is thus in the pairing something that
tends to make the two members have a similar fate.
This second conclusion could not be drawn with any great
degree of confidence from my. own experiments, summarized in
the preceding section. In most of these the number of pairs
was indeed greater than would be expected, but the excess was
usually not so great but that it would occur once out of seven or
eight cases purely asa matter of chance. Itis Miss Cull’s experi-
ment that thus far gives the only incontrovertible evidence for
something causing a like fate in the members of pairs.
Miss Cull’s experiment is of course much better adapted for
testing this particular matter than are my own, or those of Cal-
kins. In my experiments the animals were cultivated on slides,
for determining the rate of reproduction, so that it was possible
to keep only a small number of representatives of each line. If
these few died, the line was extinct. In Miss Cull’s experiment,
however, all the progeny of a given line were retained, in small
mass cultures. They were thus practically assured against acci-
dental extinction; if any died out, it indicated some intrinsic
weakness. In Miss Cull’s experiment, as we have seen, the num-
ber of complete pairs that died, and the number that survived,
was enormously greater than would be expected; so that the result
cannot be attributed to chance. Particularly in the latter part
of the experiment does this appaer, so that the pairs tend to be-
come grouped into two classes, in one of which both members
die out, while in the other both members live. The chances
are 21,000 to 1 against the result actually reached, if there is not
something in the pairing tending to give the two members a similar
fate; in other words the chances are 21,000 to 1 that the latter is
the case.
In Part II of this paper further evidence will be given on this
point. °
422 H. S. JENNINGS AND K. S. LASHLEY
COMPARATIVE RATE OF REPRODUCTION IN THE TWO MEMBERS
OF PAIRS
Method of analysis
The second ground for holding the two members of pairs to be
sexually differentiated is the fact that often one of them is vigor-
ous and reproduces freely, while the other is weak and reproduces
slowly.
Here arises a question parallel to that which we have dealt
with in the preceding section. Differences in vigor and rate of
fission are common among the exconjugants, and are of course
found likewise between individuals that are not members of pairs.
In order to be considered evidence for sexual differentiation,
the differences between the members of pairs must be more marked
than the average differences between individuals taken at random.
Otherwise they can only be considered instances of the variability
in this respect among the members of a population.
It is true that organisms might be sexually differentiated in
other respects, and still the members of pairs might show in this
particular respect differences not greater than the average differ-
ences between members of the population. But in the present
case it is this difference alone that is the ground for the supposed
sexual differentiation (particularly since we have shown above
that the relative mortality gives no such ground). Unless there-
fore the difference is more marked than the average differences
between individuals taken at random, no ground exists for holding
that the two members of the pairs are sexually differentiatied.
The mere fact that there is variation in rate of fission, distributed
without relation to pairing, would certainly be no indication of
sexual differentiation.
We may therefore proceed to determine whether the differences
in this respect between the two members of pairs are greater
than the average difference between individuals.
One special point will be of interest. If the two members of
pairs are sexually differentiated, then the one that reproduces
most rapidly will have to be taken as representing the female,
and the other as representing the male. It will then be possible
BIPARENTAL INHERITANCE IN PARAMECIUM 423
for us to determine the average differences between the indivi-
duals representing the same sex; between the diverse ‘females,’
and between the diverse ‘males.’ If the differences between the
two members of pairs are not distinctly greater than those between
members of admittedly the same sex, then a fortiori these would
constitute no basis for holding the two members to be sexually
different.
With these considerations in mind we may examine the data.
Miss Cull’s experiment
In Miss Cull’s experiment it was found that there is in most
cases a difference between the two ex-conjugants with respect to
vigor and reproductive power.
At the end of the month, of the 65 pairs then represented by living
cells, in twenty-seven pairs, or 41 par cent, one of the exconjugants only,
or the off-spring from it were alive; in fifteen pairs, or 23 per cent, the
progeny of one exconjugant was three times as large as that of the other;
in six pairs, the descendants of the one were twice as numerous as those
of the other organism; and in only five cases had both conjugants given
rise to the same number of offspring. The twelve remaining pairs
showed a wide diversity in the number of paramecia produced by any
two conjugants. (Cull ’07, p. 88).
The differences were determined by examining the number of
progeny present in the vial culture at the end of the month.
It needs to be remembered that a single additional fission
doubles the number of progeny in one conjugant as compared with
the other; so that a doubling of the number of progeny by no
means signifies a large difference in the rate of fission, in the course
of a month, or even in the course of a few days. Such a difference
in the number of progeny might be produced even though the
rate of fission in the two were only slightly different. Now, we
have no further statistics on this case, for determining how far
there are variations in the rate of fission even when we take the
organisms quite at random, so that we cannot tell how the differ-
ences between the members of pairs compare with the differences
we should find if we took two individuals at random.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No.3
424 H. S. JENNINGS AND K. S. LASHLEY
Experiments by the author
In my Experiment 1 (this Journal, vol. 14, 1913, page 286)
we have the data for determining this point. In this experiment
there were 120 lines derived from animals that began conjugation;
61 from those that were allowed to complete the process, 59 from
those that were separated before the union was consummated
(split pairs). In the five weeks during which the experiment
was continued, none of the strains from the split pairs died out,
showing that any deaths among the progeny of the pairs were not
due to lack of a revivifying effect of conjugation. We have al-
ready dealt with the deaths among the conjugants (page 416).
Here we will deal with the differences between members of pairs
in the rate of reproduction.
The number of fissions for five successive weeks are given for
the progeny of each member of the pairs in table 29 of the paper
on the effects of conjugation (this Journal, vol. 14, 1913, page
379). From this table we may find the average difference be-
tween members of pairs in number of fissions for any given period;
and also the average difference that would exist if the organisms
were paired at random (that is, the average difference between
any two individuals of the lot).
The method by which this is done may be illustrated as follows,
taking the fissions for the first week as an example. Here the
fissions are distributed among the 60 members of pairs as shown
in the following table 38 (data taken from table 29 of the paper
on the effects of conjugation ’13).
The differences in number of fissions between the two members
a and b of the 30 pairs are distributed as shown in the following
table 39 (data from table 29 of the preceding paper).
TABLE 38
‘ , f FISSIONS -
0. [ 7 Boi 3 | + | E> ve
Most rapid member of pairs (a)..... 1 | 1 | Ee | eos: 11 4 11 2
Slowest member of pairs (b).... 5 5 | 2 7 | 8 | 3 |
Rpts sieos ea i A aie: 6 | 6-| 3 | 10 | 19 4 | 2
BIPARENTAL INHERITANCE IN PARAMECIUM 425
TABLE 39
DIFFERENCES IN NUMBER OF FISSIONS
Pies he ee 4 | 5
Number of pairs....
|
Piste es 3
mv | 12 5 | roi ae Bey. bliehy La)
From table 39 it is found that the average difference between
the two members of the actual pairs is for the first week 1.533.
To find the average difference between any two members (or
the difference that would exist if the animals were paired at
random), we must pair each member with every other. It will
- be found that for any given number n of organisms, the number of
diverse pairings that can be made is 1 +2+3+... up to
one less than the given number; this is equivalent to the expres-
sion 3 n (n — 1). Thus, for the 60 organisms of the first week,
the number of possible pairs is $ (60 * 59) = 1770.
To make all possible pairings we proceed as follows, taking
the ‘totals’ of table 38 as an example: Arrange a table for the
possible differences, as in the first row of table 40. Now, in table
TABLE 40
DIFFERENCES
0 1 2 3 4 5 6 Ls
() 15. 36") 18-7) 60>) 114° |. 8 12
15 | e486) Gothia e421 49
STURN SN 4e 6
Number of pairings..... 45 190) 1 140 9) 20. |
14745 |" 2661.) 38
Di, 28
1
Total pairings.......... 341 | 568 | 313 | 236 | 204 | 96 12=1770
38 there are 6 members showing 0 fissions. If these 6 are paired
among themselves, the differences between them will likewise
be 0. The number of diverse pairings among the 6 is given by
the formula 4 n (n — 1); for n = 6 this gives 15 pairs. We enter
this 15 beneath the 0 of table 40.
426 H. S. JENNINGS AND K. S. LASHLEY
The 6 members at 0 fissions must likewise be paired with the
6 at 1 fission, giving 36 in which the difference is 1; we thus enter
this 36 beneath the difference 1, in table 40. In the same way
we successively pair the 6 (at 0 fissions) with all the other totals
of table 38, entering the products under the proper differences in
table 40; this gives the first row of ‘number of pairings’ in this
table.
We proceed in the same way with the remainder, each ‘total’
of table 38 giving under the difference 0 the entry found by the
formula 4 n (n — 1), and successively under the other differences
the product of this total with each of those following it. This
gives, for the entire first week, table 40. We next sum the num-
ber of pairings for each difference, as in the last row, and get their
total sum (1770). <A control for the accuracy of the work up to
this point is given by the fact that this total sum must be equal
to 4 n (n — 1), where n is the total number of diverse lines (in
this case 60). ;
Next we multiply each difference of table 40 by the sum of the
‘total pairings’ under it, and add these products, giving the total
sum of differences for all the pairings. In the present case this
total sum is 3270. Dividing this by the total number of pairings
(1770), we find that the average difference between members,
when all possible matings are made, is 1.8474.
In the same way we may if we desire find the average differ-
ence between the most rapidly dividing members. (‘females’)
of pairs; and between the slowest members (‘males’).
Carrying out these operations for the five weeks of Experiment
1 (table 29 of the paper on the effect of conjugation, page 379),
and for certain combinations of these periods, we obtain the
results given in table 41.
The results given in table 41 are surprising. Jn every case the
difference between the members of actual pairs is not greater, but
less, than the average difference between two individuals taken at
random.
Thus we find, for the rate of reproduction, as for the distribu-
tion of deaths, and of survivals, that the two individuals do not
resemble each other less, but more, than the average individuals.
BIPARENTAL INHERITANCE IN PARAMECIUM 427
There is something in the pairing that causes the two individuals
to resemble each other more than usual.
The condition is again the reverse of that called for by the
theory that the two members of pairs are sexually differentiated.
I pointed out on a preceding page the interest of the question
as to whether the two members of the pairs differ from each other
more than do members of what would have to be considered
members of the same sex. I have worked out this matter for a
number of periods, for the same organisms dealt with in table
41. The results are given in table 42.
Table 42 shows that the difference between the two members of
pairs (‘male and female’) is on the average distinctly less than
the difference between those that would have to be considered
members of the same sex in different pairs. It is less than the
difference between the ‘females’ of table 42 in two of the four
weeks of table 42, greater in two, and averages practically the
same. It is decidedly less than the difference between the ‘males’
of different pairs, in three of the four weeks; greater in one; it
averages distinctly less than for the ‘males.’
Thus, so far as the conditions in experiment 1 are typical,
we must conclude that the differences in rate of reproduction
TABLE 41
Experiment 1. (This Journal, vol. 14, 1913, p.286). Mean differences in number of
fissions between the twomembers of pairs, as compared with the mean differences of
members paired at random. The table includes only those pairs, both members of
which lived through the period in question, save the last entry, which gives the
data for the actual number of fissions for all the pairs that began the experiment
whether they lived to the end or not.
DIFFERENCE
NUMBER OF
PAIRS
Members of Random
actual pairs pairing
|
First week..... Bi ee ee OL eee | 30 Iebasee (ee L847
SeECONG. Weekeuse a: Ae Sock eee gate 26 16155 2.364
Mbbirdcweskeitr 84.) (Ike. omy lee) wane tte 2 19 2.895 | 3.060
Hourtheweekwsssaie as ee Ae eee eee 17 1.882 2.474
Birst GwOoweeks-e.--6.. 5 ons4. RE a ae 26 Ztol: 3.728
Baim weeks). tie) sv) See net ee | 17 5.000 5.966
Hive weeksintiays (xu eats eee ee 13 6.154 7.246
Rivesweeks all pairs’. cn 2 93a eto ty oe 30 8.333 10.998
428 H. S. JENNINGS AND K. S. LASHLEY
TABLE 42
Experiment 1. Average difference in fission rate between the members of pairs, as
compared with the average difference between the ‘females’ of different pairs; and
also between the ‘males’ of different pairs. The most rapidly reproducing member
of a pair is designated ‘female’; the slower one ‘male’. Based on data of table 29
of the paper on the effect of conjugation, (page 379).
AVERAGE DIFFERENCE BETWEEN
Members of ‘Females’ of ‘Males’ of
actual pairs diverse pairs diverse pairs
Birstiweek..: -:: So ek ee eee 1.533 1.349 1.910
Second week: Vacs. ns ee he ee or eee 1.615 1.902 7-5) ae
"Rhirdcwee kexts ae eR ee ds noe 2.895 2.363 2:666
Bourthcweeke. 4 aoe eect ee a ee 1.882 POA eh 2.059
irs 2 Wee KS ore eal ee eo OE rt a PASE DATES 4.043
Mean-for the 4: weeks). ...2..¢.. ss... + 2558 1.981 1.970 2-285"
between the members of pairs cannot be considered evidence of
sexual differentiation. For this difference is actually less than
that between individuals taken at random; and not more than that
between individuals that would, on the sexual theory, have to be
considered members of the same sex. |
Biparental inheritance?
Thus in studying the question of sexuality, we have come upon
a matter that is of much interest quite independently of this
question. The fact that the progeny of two individuals which
have conjugated resemble each other more than do those of two
individuals taken at random would seem to indicate that they
must inherit from both the conjugants.. If this be the case, we
have here the first evidence that has been presented of biparental
inheritance in Protozoa, and particularly in connection with
that form of conjugation in which both the conjugants continue to
reproduce.
Analysis of the experimental results
From this point of view we may examine the facts further.
Since in Experiment 1, as we have seen, the two strains derived
from a pair resemble each other in rate of reproduction more than
BIPARENTAL INHERITANCE IN PARAMECIUM 429
others, we may expect to find that there is a positive coefficient
of correlation between them, in this respect. To determine the
coefficient of correlation is a more accurate and elegant way of
discovering whether the two members of a pair are more alike
than strains taken at random, than that which we have thus far
employed. I have therefore computed the coefficients of corre-
lation for Experiment 1, and also for other experiments, to be
set forth. In these cases we shall as a rule not employ the method
thus far used, but consider that the determination of the coeffi-
cient of correlation suffices for answering the questions in which
we are interested.
In studying the coefficient of correlation in rate of reproduction,
we are dealing with integral variates; namely, the number of fissions in a
given period; these can be counted exactly. Furthermore, the two mem-
bers of the pairs are alike, so that we have the condition for which sym-
metrical correlation tables have been used. For determining the co-
efficient, I have employed mainly the ‘difference method’ (Harris ’09),
which is peculiarly applicable to cases of the present character. The
formula for the coefficient of correlation is for such cases by this method
where x = the number of fissions, and v = the difference in number of
fissions between the two members of a pair.
In this formula it is important to remember that c,?is given by the sum
of the squares of the positive differences between the members of pairs,
divided by the number of pairs. The formula can be written
where 7 is the number of pairs, and v is the positive difference.
In using this formula, it is not necessary to arrange the data in a
correlation table, since all that one requires is, the standard deviation
of the number of fissions (for all members together), and the standard
deviation of the differences between the two members of pairs. These
differences can be readily taken directly from such a table as table 29
of the preceding paper on the effects of conjugation (page 379), and ar-
ranged at the same time as in table 39 of the present paper.
430 H. S. JENNINGS AND K. S:. LASHLEY
Our chief data from the work set forth in the preceding paper
on the effects of conjugation are from Experiment 1. In the
first column of table 43 are given the coefficients of correlation
between members of pairs for the various periods of this experi-
ment, as well as certain coefficients from Experiments 2 and 3.
The coefficients from Experiment 1 are based on the data
given in table 29 of the preceding paper (page 379); they can
be recomputed from that table by anyone desiring to test the work.
Besides the coefficients for weekly periods, I have given also those
obtained by taking for each member the number of fissions for
periods of two weeks, for four weeks, and for five weeks. It is
important to understand that these latter are not averages,
derived in any way from the coefficients for the separate weeks,
but are computed directly from the data as to the number of
TABLE 43
Coefficients of correlation in rate of fission between the two strains derived from the
two members, a and b, of pairs, and of split pairs, in experiments 1, 2 and 8 of the
preceding paper (Paramecium caudatum). The coefficients for experiment 1 are
computed from the data given in table 29 of the preceding paper, page 379.
PAIRS: CONJUGATION SPLIT PAIRS: CONJUGATION
CONSUMMATED NOT CONSUMMATED
| Number of Correlation of a with b Number of Correlation of a with b
pairs split cars
Experiment 1:
Pirsu weeks-es es 30 0.1155 = 0.0859 24 0.2980 = 0.0887
Second week’........ 26 0.4925 = 0.0708 24 0.3082 = 0.0881
Third week i. '..... 19 | 0.13834 + 0.1074 24 0.2322 = 0.0921
Fourth week.........| 17 0.4355 = 0.0937 24 0.4966 = 0.0733
First two weeks...... 26 =| 0.4450 + 0.0708 24 —0.0686 = 0.0969
Second two weeks... .| 17. | «~0.3377 = 0.1025 24 0.4362 + 0.0788
Four weeks..........| 17 | 0.2734 +0.1070 24 0.2758 + 0.0899
Five weeks aaa be 13 0.3074 + 0.1198 24 0.1296 = 0.0957
Baar depekk: “alk peel 30 0.4788 = 0. 0671
Five weeks, all pairs. 30 =| 0.40388 + 0.0729
Experiment 2:
Two weeks......... 10 | 0.4825 = 0.1157 | 6 —0.1800 + 0.1884
Two weeks, all pairs. 28 | 0.1027 = 0.0892
Experiment 3:
Four days.. 13. | 0.4628 + 0.1039
BIPARENTAL INHERITANCE IN PARAMECIUM 431
fissions in each strain for the period set forth; it is entirely possible
that two weeks should when considered separately each give a
positive coefficient, while taken as one period they should give a
negative one, or vice versa, so that the coefficients for the longer
periods furnish independent data. In all cases except the three
entries where the words ‘all pairs’ are added, there is included in
each case only the strains in which both members of the pair
lived through the period in question. In the three where the
words ‘all pairs’ are added, I have given the coefficient which
appears when we include the number of fissions for all the strains
that entered the experiment, whether they lived through the
period in question or not; taking thus, in the case of Experiment
1, all the ‘totals’ of table 29 of the preceding paper.
In every case, as table 43 shows, there is a positive coefficient
of correlation between the progeny of members of pairs. In
one or two cases this is small and would hardly be significant
in comparison with its probable error, if it stood alone; but in
most cases the coefficient is of very considerable value. The
coefficient persists and is even in Experiment 1 increased, when
we include the number of fissions up to the time of death in the
strains that died before the end of the period in question, as in
the last two entries for Experiment 1. This is due to the fact
that there is a similarity in the two strains as regards length of
life; a point more fully brought out in the study of the distribu-
tion of mortality, already made.
The data from Experiments 2 and 3 of the paper on the effects
of conjugation are less full than those for Experiment 1, the
members of pairs being still smaller. The coefficients are given
in table 43 for what they are worth; they confirm so far as they
go the results from Experiment 1.
The question which next arises is as to the cause of the fact that,
as shown by these positive coefficients, the progeny of the two
members of a pair resemble each other in rate of reproduction
more than do those of unpaired individuals. Here there are two
possibilities:
1. The resemblance may be due to inheritance from both par-
ents. This would fully account for the facts, and appears a highly
4392 H. S. JENNINGS AND K. S. LASHLEY
probable explanation. The resemblance would in this case be a
result of conjugation.
2. The resemblance might be due to assortative mating. We
know that assortative mating does occur in Paramecium, so far
as size is concerned, individuals of nearly equal size tending to
mate together (Pearl ’07, Jennings ’11). It appears possible that
the similarity in size might be accompanied by a similarity in
physiological characteristics, resulting in similar rates of fission.
This would be especially notable if there were a tendency for mem-
bers of the same or related strains to mate together; and this
tendency does exist, as I showed in my paper of 1911. Thus it
appears not improbable that the correlation is partly due to assor-
tative mating. So far as this is the correct explanation, the
resemblance would exist before the mating took place; would not
be a consequence of conjugation.
It seems on the whole probable, a priori, that both inheritance
and assortative mating play a part in bringing about the greater
resemblance in the progeny of the two members of pairs.
How can we test the validity of these explanations? If assor-
tative mating plays a part in the matter, then the resemblance in
the two strains derived from a pair ought to exist, to a certain
degree at least, if the two members of the pairs are separated
before conjugation is consummated. Now, this operation was
performed, for other purposes, in Experiments 1 and 2 of the
preceding paper. In these cases we had the progeny both of
pairs and of ‘split pairs.’ By determining whether there is corre-
lation between the two individuals, a and b, of the split pairs (indivi-
duals that were uniting for conjugation), we can tell whether assor-
tative mating plays any part in the matter. The coefficients are
given in the second column of table 43.
As there appears, there is in fact a positive coefficient of correla-
tion, in most cases, between the progeny of animals that were
beginning to pair, but did not complete the process. The corre-
lation is on the whole of similar value to that found between the
two members of pairs that had completed conjugation. If we
average the eight coefficients for Experiment 1 in each case, we
find that the mean for the conjugants is 0.3176; for the non-con-
BIPARENTAL INHERITANCE IN PARAMECIUM 433
jugants 0.2635; so that that for the conjugants is about 20 per
cent greater.
It would seem therefore that assortative mating accounts for
at least a part of the similarity between the progeny of the pairs. ~
Whether any part of it is due to inheritance can hardly be deter-
mined from our data up to this point, since the probable errors
are in all cases large, owing to the small numbers of pairs in the
experiments.
What we need clearly is, data based on larger numbers. Such
will be supplied in Part II.
PAs Ll
H. S. JENNINGS AND K. 8. LASHLEY
EXPERIMENT 16: 241 PAIRS OF CONJUGANTS PROPAGATED SEPAR-
RATELY FOR FORTY-SEVEN DAYS
On the work presented in Part I the criticism may be made that
the numbers dealt with are hardly sufficient to place beyond all
doubt the conclusions to which they lead. This applies particu-
larly to the results shown in the measurements of the similarity
in rate of reproduction between the descendants of the two mem-
bers of pairs; the coefficients of correlation are variable and their
probable errors are large. It hardly applies to the study of the
distribution of the deaths among the exconjugants, since here we
had the relatively large number of Miss Cull’s valuable experi-
ment (93 pairs) to work with, but these results should be tested
by others. It is true that the difficulty in dealing with large
numbers is very great when the rate of reproduction for each
strain must be recorded for long periods. But the fact that it is
difficult to work with large numbers does not lessen the insecurity
of results drawn from small ones.
Furthermore, all the results thus far have been drawn from
experiments designed for other purposes; the question arises as
to whether all conditions have been fulfilled for getting accurate
results on the present problems. One point in particular suggests
itself. In the experiments thus far, the two members a and 6
434 H. S. JENNINGS AND K. S. LASHLEY
were cultivated side by side in the two concavities of a single slide.
In changing the animals to new fluid, it appears possible that the
pipette, after transferring a, would retain some of the bacteria
from the a drop, and mingle it with the 6 drop, when 6b was
changed. ‘Thus possibly the bacterial content of the cultures of
the two members of pairs might be on the whole a little more
uniform than that of two cultures taken at random; this might
cause the rate of reproduction to be a little more alike in the two
members of pairs or of split pairs, producing a correlation.
For these and other reasons it was determined to undertake
a very extensive experiment, giving numbers sufficiently large
to make the results reliable, and at the same time fulfilling all
conditions which the experiment demands. For this purpose
the two authors of Part II joined forces, since it is physically
impossible for one person to care for so many cultures as are
required.
We shall call this Experiment 16, in order that there may be
no confusion with Experiments 1 to 15 described in the senior
author’s paper on the Effects of Conjugation (this Journal, vol.
14, 1913, pages 279-391), since in the present paper it is neces-
sary to refer frequently to these fifteen experiments. In Experi-
ment 16 as carried out, 482 strains, derived from the two members
(a and b) of 241 pairs of conjugants, were propagated forty-
seven days, an exact record being kept throughout of the num-
ber of fissions for each strain; also of the dying out of strains.
The important facts as to the culture methods used are as
follows:
On March 23, 1912, 250 pairs were isolated from watch glasses
that had been taken the night before from a wild culture of Para-
mecium caudatum. Accidents later reduced this number to
241 pairs. As soon as the two members separated, they were
transferred to separate slides, giving us thus 482 distinct strains.
To each pair a number was given, while the two individuals form-
ing the pair were called a and b; the progeny of each of these
received the same designation. Each strain was therefore desig-
nated by a number and a letter, giving precisely its origin and
relationship; the designations running from 1 to 241, a and b.
BIPARENTAL INHERITANCE IN PARAMECIUM 435
Until they separated the pairs were kept in the same fluid in
which they were found. After separation the individuals were
removed to a fluid consisting half of filtered culture water from
the original culture from which they came, half of fresh hay
infusion. The latter was made by boiling 1 gram of pure timothy
hay for ten minutes in 200 ec. of water. The animals were cul-
tivated throughout the experiment in this mixture, which has the
great advantage of supplying a uniform bacterial content to all,
thus preventing certain strains from developing a peculiar bac-
terial flora, which would differentiate them from the others.
The two members, a and ), of a given pair, were placed on separ-
ate slides and kept in separate moist chambers. All the a’s were
kept together in one series, all the b’s in another. This effectu-
ally prevented the induction of a resemblance between the two
by special similarity of conditions, or by transfer of bacteria
from one to the other. Of course all the moist chambers were kept
near together, under as nearly identical conditions as possible.
The culture fluid was changed every other day, by transferring
the animals to two drops of new fluid on a fresh slide. As a rule,
one of the authors transferred the a series, the other the 6 series,
so that there was no opportunity for a correlation to arise by spec-
ial similarity of treatment of a and b in certain pairs.
In addition to the slide cultures, it was necessary to keep small
stock or reserve cultures, in the mass, of each strain; this gave us
482 such mass cultures to handle, in addition to the 482 slides.
These mass cultures were necessary in order to give us accurate
knowledge regarding the mortality of the different strains, since
the slide culture of a given strain might die out owing to accidental
causes, while the mass culture still persisted; this would show that
the death was not due to intrinsic causes. No strain is recorded
as dead save when both slide culture and mass culture have died
out. When the slide culture alone died, it was replaced from the
stock culture.
Owing to the irregularities in fission following immediately upon
conjugation, it was thought best to make no use of the record of
fissions before March 30, one week after conjugation had occurred.
The record was of course taken every other day, the exact number
436 H. S. JENNINGS AND K. S. LASHLEY
of fissions being recorded. ‘The records were then grouped by ten-
day periods, and the number of fissions in these periods were made
the units in analyzing the results.
We have thus for each of the 482 strains (so far as they lived
through), four records, of the number of fissions for four succes-
sive ten-day periods, extending from March 30 to May 8, 1912.
These records are given in table 51, which will be found in the
Appendix. This table contains everything that might be dis-
tributed in a great number of correlation tables, and at the same
time shows much that would not appear in such tables. From it,
anyone who so desires can construct tables or repeat the compu-
tations on which the results hereafter are based, in order to test
their accuracy. Further, we hope that table 51 may serve as a
storehouse of material, from which independent study can be
made by others on the problems of reproduction and inheritance
in these organisms. We shall ourselves make much farther use of
it, in papers to appear on inheritance in Protozoa.
Distribution of survivals and deaths among the members of pairs
We may first examine the question as to whether the two mem-
bers of pairs tend to have the same, or a different, fate, so far as
mortality is concerned. Is there a tendency for one to die and the
other to live, or are their fortunes alike?
The pertinent data for the four periods of table 51 and for a
subsequent period of twenty-one days, are given in table 44.
Since the periods of table 51 do not begin till seven days after
conjugation, we have in table 44 the data as to the deaths for
five periods, of respectively seventeen, twenty-seven, thirty-seven,
forty-seven, and sixty-eight days after conjugation. For the
first four periods the data are extracted from table 51; for the fifth
period of sixty-eight days, the data of table 44 are independent
of table 51.
The understanding of table 44, which condenses the results of a vast
amount of work, will be facilitated by the following explanation of the
various entries. .
The total number of strains at the beginning was 482; or for the lower
half of the table, where 8 pairs are omitted, 466. Columns (2) and (6)
BIPARENTAL INHERITANCE IN PARAMECIUM 437
TABLE 44
Experiment 16. Paramecium caudatum. Distribution of deaths among the mem-
bers of the 241 pairs of table 61, for five periods following conjugation, with the
probabilities of the observed results, if the distribution of deaths has no relation
to the pairing. The table is divided into two parts: in the upper half are given
the data when we include all of the 482 strains that were cultivated (m = 482). In
the lower half are given the data in case we omit entirely the members of the eight
complete pairs that died during the first period, in order to exclude any possibility
that their deaths may have been due to the handling of the pairs while still united
(see text). This leaves a total of 466 to be considered, in place of 482. See the
detailed explanation following the table.
SURVIVING
oO
ie]
b
i}
PROBA-
Zz
8 a. 12 7 a , 2 BILITY OF A
az 2 Les ES 2 1¢4a/] 8s DEVIATION ODDS AGAINST SO
eel 8 $s8/ 2a 5S |ose8!] 3a SO GREAT |GREAT A DEVIATION
27 cc fale Mh Alagoa eS = Be ipa eas > | AS THAT |
ao > eae | Sa = > Se) Se = OBSERVED
Bel. | Sea) es |) e | 3S hees| bat 8
a ea |S | < S pe = < =
(1)| (2) (3) (4) (5) (6) (7) (8) (9) (10) | (11)
Wy) 55) 8 |. 8 5 | 427] 189| 194} 5 | 0.00580} 191 tol
21.| 73 5 ate 409} 173| 180| 7 | 0.00242 | 411.9 tol
3779 6 : 9 403 | 168 177 9 | 0.000236) 4232 tol
47 | 92 9 18 9 390 | 158) 167 9 | 0.000319) 3130.9 to 1
68 161 27 40 13 321 | 107| 120 13 | 0.000343) 2915.7 to 1
(Omitting the 8 pairs of the first period, m = 466)
|
SsT aso) 14. er 1s) 409 | 179 |) 180). 1 | 075815 | | 1 to 84
Bag Awlhe 7 dee 3tr) 408. | 174 | 177\|,..3|-0.15350!| 5.5 tad
47| 76 | 6 | 10 | 4 | 390| 163] 167] 4 | 0.08824] ,10.3 tol
68|145 | 22 | 32 | 10 | 321| 110] 120] 10 7.6 to 1
0.00402 | 247.
give the number of individuals dead, and alive, respectively, at the end
of the periods given in column (1). Columns (3) and (7) give the most
probable numbers of complete pairs that would be included among these
if there were no relation between pairing and the distribution of deaths;
this is determined by formula (1), page 461.
Columns (4) and (8) give, on the other hand, the actual number of
eomplete pairs included; while columns (5) and (9) show how much the
actual number exceeds the most probable number. (It is to be observed
that this excess is the same for those dead as for the survivors.) Column
(10) gives the probability that any deviation so great as this (either by
excess or by deficiency) should occur. This is determined in accord-
ance with rule (5), in the Appendix. (If we should compute the proba-
bility for so great an excess, it would be much smaller than the proba-
bility given in column 10.) This probability shows directly in how
many case out of a thousand or a million, et cetera, so great a deviation
438 H. S. JENNINGS AND K. S. LASHLEY
would be found as a matter of chance; thus, for the first period, so great
an excess would occur but 58 times in 10,000 cases. Column 11, giving
the odds against so great a deviation as that actually observed, is ob-
tained in accordance with rule (5).
Attention should be directed first to the upper half of the table
only; the lower half is designed to meet a possible difficulty, to be
taken up later. Table 44 shows that in this experiment, as in
that of Miss Cull, analyzed in Part I of this paper, the number of
complete pairs among those dead, as well as among those living,
is throughout much greater than would be expected if the dis-
tribution of deaths had no relation to the pairing. That is, there
is a strong tendency for the fate of the two members of a pair to
be alike. The condition found is the reverse of that demanded
by the theory of sexual differentiation.
Table 44 shows also, in columns (5) and (9), that the deviation
from the most probable number increases as time passes; in other
words, there is a tendency in the later periods for deaths to occur
among the mates of those that have already died out. This
tendency we saw likewise in analyzing Miss Cull’s experiment.
The probability of getting the results observed, as a matter of
chance distribution, is so excessively small, as shown by columns
(10) and (11), that it must be left quite out of consideration.
This probability is in fact even much less than that computed
for columns (10) and (11), since what we give there is the prob-
ability for any deviation of this amount, whether plus or minus.
But all the deviations are plus; this decreases the probability
greatly.
It is then absolutely clear that there is something in the pairing
which tends to cause the two members of a pair to have the same
fate.
There is one possible source of error that deserves considerations
In removing the pairs to a slide while still united, it is conceivable
that both might be injured, in such a way that both would later
die. Thus the deaths of a certain number of the complete pairs
might be accounted for.
While there appeared to be no ground for supposing this to be
the case in the present experiment, it will be worth while to analyze
BIPARENTAL INHERITANCE IN PARAMECIUM 439
the matter in such a way as to exclude even this remote possi-
bility. To do this, we may exclude from consideration all the
pairs that died before the end of the first period. If the deaths
do not occur within seventeen days of the time the common hand-
ling occurred, there is evidently no danger that they were due to
this handling; particularly in animals that produce a new genera-
tion every twenty-four hours. There were 8 complete pairs that
died during these first seventeen days. Excluding these com-
pletely, we have 466 lines (233 pairs) instead of 482 lines (241
pairs)to consider. If now we examine the mortality in the later
periods among these 466 lines, omitting the 8 pairs that died during
the first period, we have the results given in the second half of
table 44.
As the table shows, essentially the same relations are shown
when we proceed in this manner, as when we include all cases.
In every period the number of pairs whose members have the same
fate is greater than would be probable if the distribution of
deaths had no relation to the pairing. The excess becomes greater
and greater in the later periods, showing that there is a tendency
for the mates of those dead to die, and of those alive to live. The
only difference made by omitting the eight pairs of the first period
is to make the numerical expression of the improbability less.
Yet even thus it rises to 247.6 against 1, in the last period.
Thus there is absolutely no escape from the conclusion that
there is something in the pairing which tends to make the two
members of the pairs alike in their fate.
Comparative rate of reproduction in the members of pairs
We may next examine the rate of reproduction in the members
of pairs, as shown in table 51, in order to determine whether the
members of pairs are more alike or less alike in this matter than
would be expected.
The best way to do this is to determine the coefficients of
correlation between the members of the pairs, for the diverse
periods of the experiment. It may possibly be of interest, how-
ever, to first examine directly for some typical period the question
,
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
440 H. S. JENNINGS AND K. S. LASHLEY
whether the differences between the two members of the pairs
is less or greater than that between two strains taken at random.
For this purpose we will select the first twenty days of the experi-
ment (sum of the first two periods in table 51). There were 179
pairs (358 lines) that lived through this period. The average
difference in the number of fissions between the two members of
these 179 pairs, as determined by an examination of table 51, is
2.514 + 0.122 fissions.
The average difference between strains taken at random is
determined in the way set forth on. page 424. Mating each of
the 358 strains with every other gives us 63,903 pairings; and the
average difference turns out to be 3.493 fissions. This is greater
than the average difference between the members of pairs by
1.429 fissions, or 56.8 per cent.
Thus again we find that the average difference in fission rate
between members of pairs is not greater than that between strains
taken at random, as would be required in order to give evidence
of sexual differentiation; on the contrary it is very much less.
The progeny of members of pairs are more alike than are the pro-
geny of individuals that have not paired.
We may now turn to an examination of this matter by the aid
of the coefficient of correlation.
In table 45 are given the coefficients of correlation in number of
fissions between the two members, a and b, of the pairs, for the
various periods of table 51. It is important to recall again that
the coefficients for the longer periods (20 days, 40 days, etc.) give
evidence independent of that for the shorter periods (or at least
not bearing any simple and evident relation to the latter).
As table 45 shows, the correlation in rate of fission between
the members of pairs does indeed show itself equally when we
experiment with very large numbers, and in such a manner as to
exclude the possibility that the similarity of the two members is
due to environmental similarity. In this experiment on 482
lines of propagation, as set forth in our description of methods,
the two members a and b, of a pair, are kept on separate slides,
in separate moist chambers, and handled separately—as a rule,
by different persons. All the experimental conditions are such
BIPARENTAL INHERITANCE IN PARAMECIUM 44]
that any irregularities in the experimental treatment would tend
to decrease any correlation that might intrinsically exist. Yet in
every period there is a marked coefficient of correlation, which as
a rule is ten to twenty times as great as its probable error.
The effect of possible similarity in conditions in the case of
strains lying adjacent to each other in the culture chambers was
tested in the following way: In the present experiment, where
the a’s formed one series, the b’s another, the two lines la and 2a
occupy the two concavities of a single slide; the lines 3a and 4a
the next slide, and so on. Hence the a of a given odd numbered
pair, and the a of the succeeding even numbered pair are through-
out immediately adjacent, and therefore under similar conditions.
The same relations hold for the b’s. If therefore, this similarity
of conditions due to adjacence in the experiment has any effect
in producing correlation, as suggested on page 434, then this
should give a positive coefficient of correlation when we compare
the odd numbered line with its following even numbered line.
TABLE 45
Experiment 16. Paramecium caudatum. Coefficients of correlation between the
number of fissions for a and that for b (a and b being the two members of a pair),
for the various periods embraced in table 51. In each case, save where the word
‘all’ is appended, the coefficient is based on the fissions of pairs of which both
members lived through the period in question. Where the word ‘all’ is appended,
the actual number of fissions for the entire 241 pairs is dealt with, including those
that died out before the end of the period in question.
CC ee
Ebinsta ON Sante este a ae ee i 193 0.3068 = 0.0311
First 10 days, all.. a 241 0.3770 = 0.0264
First 10 days, niche fone that did abe
Givers eAe roe Ms hes Pee) ae ee 183 0.3319 + 0.0314
Second glOMd ays. asec a tucks see eee oe see: 179 0.3862 += 0.0305
Ming MOK aya te sees me Es cee oe 175 0.2690 + 0.0334
Third 10 days, omitting those that did
MOUILVAC Clee os ks ee eee eae cee 174 0.30388 = 0.0328
Mourth TOM ays oie Mies cp cae ae 167 | 0.1021 + 0.0365
HITSe COAA YS. Scat Le fa eee ne al 179 0.4793 += 0.0275
SECON 20 GAYS scone: cee ree eee 166 0.2094 = 0.0354
Wer alAO days..5ss58C ot hie Mametnies wae ana 166 0.3450 = 0.0326
Rie pal sO rdayastall ya sor in eoae twee | 241 0.3842 + 0.0263
4492 H. S. JENNINGS AND K. S. LASHLEY
We carried out this operation for the second ten-day period
(April 8 to 18). There were 346 cases where pairs could be made
up in this way, giving a total of 173 pairs. The numbering in
the original experiment was not the same as in table 51, since nine
of the original 250 pairs were lost by accident; the numbering had
therefore to be altered to close up the gaps. Hence the computa-
tations were not based on comparing the odd and succeeding even
lines in table 51, though the latter would presumably give a simi-
lar result.
The coefficient of correlation between the odd numbered line
a (or 6), and the even numbered line a (or 6) lying adjacent to it
in the moist chambers was, for these 346 lines but 0.0117 + 0.0363
That is, there was no correlation whatever. The evidence is
therefore strong that this adjacence in position has nothing to
do with the production of correlation in the previous experiments,
where a and b were adjacent.
All together, the result of this experiment, with its 482 lines,
excludes the possibility that the correlation observed in former
experiments was without significance owing to the small numbers
employed; or owing to any possible similarity in the environmental
conditions of the two members of pairs. The correlation is
certainly real, and due to some intrinsic special similarity between
the animals that have conjugated.
The data of table 51 show many other relations of extreme
interest, bearing on inheritance; on the possible origin of heritable
variations; on the fate of the diverse lines of exconjugants; on
changes in reproductive power with the lapse of time, and many
other points, some of them of great importance for a full interpre-
tion of the significance of conjugation. But in order not to com-
plicate the present paper, which deals mainly with the general
question of whether conjugation produces biparental inheritance,
it appears best to reserve the analysis of these matters for a later
paper on variation and inheritance of these animals. We shall
therefore take up here only matters which bear upon the question
whether conjugation actually does result in biparental inheri-
tance.
BIPARENTAL INHERITANCE IN PARAMECIUM 443
EXPERIMENT 17: 239 SPLIT PAIRS PROPAGATED SEPARATELY FOR
TWENTY DAYS
The only possible explanation for the positive correlation
between the members of pairs other than that it is the result of
conjugation, is that based on assortative mating, as already set
forth on page 432. The two conjugants which united in a single
pair are known to be similar in size before union; they might also
be similar in other respects, as in fission rate. The coefficient
of correlation between the two members might then exist before
conjugation, not being due to the latter.
To properly test this matter, it became necessary to therefore
carry on an extensive experiment with split pairs, comparable to
Experiment 16 with pairs. It would be desirable if the pairs and
split pairs could come from the same culture and be propagated
side by side. With the large numbers we wished to employ,
this was however impraticable.
We therefore collected material from the same ditch from which
came the pairs of Experiment 16; put this under the conditions
for producing conjugation, and in this way obtained on May
9, 1912, 239 split pairs (pairs of which the individuals were separ-
ated before conjugation was consummated). These were propa-
gated in the same manner as were the conjugants of Experiment
16, so that we may refer to page 434 for an account of the methods
employed.
Unfortunately the mistake was made in this case of collecting
the material from near the mouth of a sewer, where the animals
were living in a dense mass, in company with bacteria of putre-
faction. Paramecia which come from such conditions do not
live well under the conditions necessary for slide cultures, so that
the mortality was high. Both members of 190 split pairs however
lived through the first ten-day period, and both members of 155
split pairs through a period of twenty days. This gives ample
material for making the test we desire.
The data obtained, so far as they bear upon correlation in
fission rate between the two members of split pairs, are given in
tables 46, 47 and 48. It may be desirable in a later paper on
444 H. S. JENNINGS AND K. S. LASHLEY.
TABLE 46
Experiment 17. Paramecium caudatum, split pairs. Correlation between the two
members (a and b) of 190 split pairs, in number of fissions, for the first 10 days
(May 9-19) of the experiment. (The more rapid member is designated a, the less
rapid b.)
Fissions of a
7 | 8 9 10 11 | 12
} }
4 1 | 1
; ZvS 1 1
676 a 1 1 4
ee 10 7 9 4 33
23 ee Me! 18 9 63
2 9 Lee Al ey 13 55
40 12 13 5 eee
11 | 4 1 5
6 | 28 | 47 | 66 |°43 | 5 | 190
TABLE 47
Experiment 17. Paramecium caudatum, split pairs. Correlation in number of
fissions between the two members (a and b) of 155 split pairs for the second 10
days (May 20-30.)
| ante
| ; Fissions of a
9 | 10 | 1 | 12 13 | 14 | 15 | 16
4 i 1
5
6
ae, 1 1
ie ee ae | 1 ee
2 9 Ag is fa ae ob tk 8
2 10 Bia Sma he | 5 Oe lee:
2 11 226 | dd. tne | elena alae
yale oe We are fe! 1 | 59
13 Bee ie wince.
14 | 1 Shee Ae hee
15 | 3 3
1 | 2 | 84 ogo) "ae | ae) ae Sar aoe
BIPARENTAL INHERITANCE IN PARAMECIUM 445
TABLE 48
Experiment 17. Paramecium caudatum, split pairs. Correlation in number of
fissions between the two members (a and b) of 155 split pairs, for the entire 20
days (May 9-30) of the experiment.
Fissions of b
Fissions of a
lz 18 19) 20 21 Wee 23 24 20 26 PET
15 1 1 1 3
16 1 1 2 1 5
Zs 1 1 2 1 5
18 2 2 4 3 1 1 13
19 | 2, 5 6 6 2 1 3 25
20 hes 4 4 3 6 2 21
ail 1 10 2 5 6 24
22 | Cc a Mee eae a 28
23 7 8 4 1 20
24 4 3 2 1 10
25 1 ]
i 2 5 12 17 31 22 | 30 24 4 2 155
inheritance to publish the complete record, as is done for Experi-
ment 16 in table 51, but the three correlation tables, supplemented
by further facts to be given in the text, are sufficient for present
purposes.
We shall take (1) the distribution of deaths, and (2) the rate
of reproduction, in the split pairs.
Survivals and deaths in the split pairs
Of the 478 lines with which the experiment began, 57 died out
during the first ten days, including both members of 8 split
pairs. During the second ten days 42 additional lines died out,
with both members of 7 additional pairs. Thus during the
entire twenty days, 99 lines died out, including both members of
15 split pairs.
Rate of mortality. The data just given show that the mortality
in this experiment was considerably higher than in Experiment
16, where we were dealing with pairs that had completed conjuga-
tion. In the present experiment, 99 out of 478 members of split
pairs, or 20.7 per cent, are dead at the end of twenty days, while
446 H. S. JENNINGS AND K. S. LASHLEY
in Experiment 16 but 73 out of 482 members of pairs, orl5.1 per
cent, are dead at the end of twenty-seven days (see table 44).
The question will naturally be asked whether this difference
has anything to do with the completion of conjugation in the one
case, and its lack in the other; whether comparison of the two
experiments indicates that the ex-conjugants are more vigorous.
To this question the answer no must be given; no conclusion what-
ever can be drawn on this point from comparing the two experi-
ments. The grounds for this are as follows:
1. The matter was directly tested by a later experiment in
which both pairs and split pairs came from the same source as
the split pairs of Experiment 17, the two sets being subjected
throughout to identical treatment. This later experiment,
beginning June 11, 1912, was planned for the purpose of comparing
the fission rate and mortality of members of pairs and of split
pairs for a long period, under identical conditions; as it turned out,
it gave data only on the relative mortality of the two. The experi-
ment included 130 pairs (260 lines), and 122 split pairs (244 lines),
making in all 504 lines of propagation.
The mortality was much higher than in experiment 17, and
affected pairs and split pairs to nearly the same extent, though
with a slight advantage in favor of the split pairs. On June
17 there remained alive but 34 out of 260 lines descended from
pairs, or 13.1 per cent; of the split pairs 48 out of 244, or 19.7
per cent still survived. The experiment was then abandoned,
since the survivors were too few to give valuable data on the
fission rate.
The difference between Experiments 16, 17, and this later
one in respect to mortality was clearly due to differences in the
conditions. In Experiment 16 the animals came from relatively
pure water, and the experiment was carried on in cool weather;
the mortality was low. In experiment 17 and the later one the
animals came from extremely foul water (mouth of a sewer),
and the experiments were carried on in hot weather; the tempera-
ture being still higher in the later experiment (June 11 to 17)
than in Experiment 17 (May 9 to 30). The mortality became
greater as summer came on; a result in accordance with much
BIPARENTAL INHERITANCE IN PARAMECIUM 447
other experience. Hence no conclusions are warranted as to the
comparative mortality rate of the descendants of pairs, and of
split pairs, from comparing Experiments 16 and 17.
2. This is merely one example of a general principle which is
impressed on one throughout all the many experiments on which
the present series of papers is based. For comparison as to mor-
tality, vigor, fission rate and the like, one must always compare
two lots that are cultivated together, so that all external conditions
are the same for each set. Comparisons of such matters in sets
cultivated at different times, and therefore necessarily under
different conditions, are bound to give fallacious results.
- Distribution of survivals and deaths. The experiments which
precede the present one show that after conjugation the two mem-
bers of a given pair tend to have the same fate (either survival
or death for both). In the present experiment with split pairs
we have the same conditions before us, save that the two paired
members have not conjugated. Do such split pairs show the
tendency to have the same fate?
In determining this, certain facts are to be noted. It is neces-
sary to use a certain amount of violence in separating some of
the split pairs, and this may cause injury. If so, this injury
will probably affect both members of the split pair, so that in
consequence both may die. Thus one might expect to find among
any that die shortly after separation, and without having divided,
a greater number of split pairs than would occur if the distribu-
tion of deaths were not influenced by this common injury to the
two members of the split pairs. It is therefore to the period
following the first three or four days of the experiment that we
must look for evidence as to any intrinsic relation tending to
induce the same fate in the two members. In the case of pairs
in which conjugation was completed, we found that the tendency
for the two members to have a similar fate increased greatly in
the later periods of the experiment.
In the first four days of the present experiment with split pairs
there died 26 lines, all save one without fission. Among these
26 lines were both members of 5 split pairs, all dying with-
out fission. Applying our formula (1), we find that when 26
448 H. S. JENNINGS AND K. S. LASHLEY
lines die out of 478 the most probable number of pairs is but
1. The actual number is five times as great as the most probable
number, so that it is clear that among those dying during these
four days something tends to cause both members to die if one
dies. Common injury to the two in separating them would have
this effect.
Compare this with the relations in the next six days of the first
ten-day period. There are left at the beginning of this six days
218 complete pairs (486 lines); of these 29 died during the six
days and these included 2 pairs. Thus the number of deaths is
greater than in the first four days, but the number of pairs in-
cluded is less than half as great. By formula (1) we find that when
there die 29 out of 486, the most probable number of pairs is 1, so
that here we have an excess of 1 pair. The probability for 1 pair
is 0.8942; for 2 pairs, 0.1811, so that 2 pairs would occur, as a
matter of chance, about half as often as 1. The deviation from
the most probable number is 1, and by rule (5) we find that prob-
ability of so great a deviation is 0.60582. Hence it is more prob-
able that we should find so great a deviation as this than that
we should not, even if the distribution of deaths is quite independ-
ent of the mating. There is thus no positive evidence here that
the two members of the split pairs resemble each other more than
any two individuals taken at random.
In the last sixteen days of the experiment (following the first
four-day period), 70 of the lines died out, including 8 pairs. By
formula (1) or (2), the most probable number of pairs when 70
lines die, out of 486, is 6, so that we have here a deviation of 2
pairs. By rule (5) we find that the probability of so great a
deviation is 0.3038, so that we should find such a deviation in
about one case out of three, though the distribution of deaths be
quite independent of the mating. Such a result cannot be con-
sidered to give any positive evidence that the two members of
the split pairs tend to have the same fate.
We may also examine the results for the entire twenty days,
comparing it with the results for the first four days. Of the en-
tire 478 lines, 99 died out during the twenty days, including 15
pairs. The probable number of pairs is 10, so that we have an
BIPARENTAL INHERITANCE IN PARAMECIUM 449
excess of five pairs over the most probable number. At the end of
four days the excess was four pairs, so that in the succeeding six-
teen days the excess has increased by but | pair; a result that
might readily be produced by chance.
Thus in the split pairs it is clear that there is an excess in the
number of pairs included among those that die immediately after
separation, without fission; and this is what might be expected
from the violence sometimes necessary in separating them. But
in the.,remainder of the experiment there is little evidence of a
tendency for the two members of the pairs to have a common fate.
There is a very slight excess in the number of cases where both
members of the split pairs died. If this is not due to chance, it
may be the result of the assortative mating which we know to
occur.
But when we compare the split pairs with the pairs that had
completed conjugation, as in Miss Cull’s experiment, and in our
experiment 16, we find a very great difference in this respect.
_ In the conjugants, the tendency for the fate of the members of
pairs to be alike becomes greater and greater as time passes, until
finally we get such extreme results as are found in the third period
of Miss Cull’s experiment (page 413), or in the later periods of
our experiment 16 (table 44, page 4387). If in experiment 16
we examine for the conjugants the deaths in the first four days
of the experiment, we find but five, including no pairs whatever.
It is clear, therefore, that deaths due to common injury of the two
members played no part in the case of the conjugants, yet in the
experiment as a whole the tendency for the two members to have
the same fate is much greater than in the split pairs, where such
injury certainly plays a large part.
Summary. The difference between the split pairs and the pairs
that have completed conjugation is then in this respect very
great, showing that in the conjugants something has occurred
to make the two members of the pairs more alike than they were
before conjugating. It is clear therefore that by conjugation the
progeny of the two members of pairs are made alike in vitality, so
that they tend to have a similar fate, both surviving or both dying
out.
450 H. S. JENNINGS AND K. S. LASHLEY
Comparative rate of reproduction in members of the split pairs
The number of fissions for the progeny of the members of the
split pairs was recorded for twenty days. We are here interested
in the question whether the two members of a split pair tend to be
alike in their rate of reproduction, as is the case with the two
members of pairs that were not separated before conjugating.
For determining this, we have formed tables of correlation for the
number of fissions of each member a, as compared with its pros-
pective mate b. The results for the 190 split pairs during the first
ten days are given in table 46; during the second ten days (155 split
pairs) in table 47; during the entire twenty days (155 split pairs),
in table 48. These tables include only the lines in which both
members of the split pairs lived throughout the period in question.
The coefficient of correlation for such tables is found in accordance
with the method set forth on page 429.
For the first ten days the coefficient of correlation for the 190
split pairs (table 46) is 0.1802 = 0.0335. The correlation is
thus slight, but it may have some significance, since it is about ~
six time its probable error.
For the second ten days (table 47) the coefficient is 0.0438
+ ().0382, a result which, based as it is on 155 pairs (310 lines),
indicates that there is no correlation between the two members of
the split pairs.
Finally, for the entire twenty days (table 48) the coefficient
of correlation for the 155 pairs is 0.2620 = 0.0356. Here again
we have a‘small positive coefficient which comparison with its
probable error indicates to be significant. 7
There appears thus to be a-slight tendency for the prospective
members of a given pair to be a little more alike than usual, in
rate of fission, even before conjugation has occurred. That is,
there appears to be a slight degree of assortative mating with
respect to fission rate. This, if it actually exists, is doubtless
a secondary result of the assortative mating in size that is known
to occur (see Jennings *11). Possibly individuals of similar size
tend also to have a similar fission rate.
But this very slight pre-existing correlation, whose existence is
even doubtful, since it did not show itself in the second ten days
BIPARENTAL INHERITANCE IN PARAMECIUM 451
of the present experiment, is evidently too small to account for
the very marked coefficients found when conjugation has actually
occurred. While assortative mating may account for a very
small degree of similarity between the members of pairs, the pro-
cess of conjugation itself results in a marked increase of that
similarity.
That is, the two members of a pair are more alike in their herit-
able characters after conjugation than before. As a result, their
progeny resemble each other in fission rate more than they would
have done if conjugation had not occurred. Conjugation results
in biparental inheritance.
We shall see the demonstration of this conclasion completed
by -the results of Experiment 15, to be described in Part III.
PAR TU
H. 8S. JENNINGS
CONJUGATION WITHIN A PURE STRAIN: EXPERIMENT 15
The results thus far given require to complete them a study
of biparental inheritance in the case of conjugation within a
pure strain, where all the conjugants are derived by fission from
a single individual. Such an experiment, including both pairs
and ‘split pairs, was carried through in August and September,
1912. This experiment has been described in my paper on
the Effect of Conjugation (this Journal, 1913, vol. 14, page 343),
and the records of the experiment are given in full in tables 34 and
35, in the Appendix of that paper. ‘The results bearing on bipa-
rental inheritance, to be given here, are drawn from analysis of
the data given in the tables just mentioned. We shall refer to the
experiment as Experiment 15, in accordance with the designation
used in the former paper, but it was carried through subsequent
to Experiments 16 and 17 of the present paper.
Referring the reader then to the paper just cited for a descrip-
tion of the experiment and the detailed record, I shall here men-
tion only essential points. In this experiment 88 lines derived
from ex-conjugant members of pairs, and 174 lines derived from
452 H. S. JENNINGS AND K. S. LASHLEY
non-conjugant members of split pairs, were cultivated for twenty-
four (or twenty-one) days, with extraordinary precautions for
keeping the conditions uniform in all the lines. The two members
of given pairs or split pairs were kept in separate moist chambers
_and handled separately.
Unfortunately, for reasons set forth in my former paper (713),
in but a comparatively small number of cases were descendants of
both members of pairs or split pairs propagated. Where this
was done, however, two lines of propagation were retained from
each member. The results are so extremely marked that the
number of lines thus obtained turns out to be amply sufficient
for solving the problem in which we are here interested.
The question for answer is: Do the progeny of the two members
of a pair give evidence of biparental inheritance, in case all the
conjugants are originally derived from a single parent? In other
words, are the progeny of the two members of the pairs more
alike than they would have been if their parents had not conju-
gated together, but had conjugated with other individuals? |
To answer this question for the rate of reproduction, which
was the character studied, we must determine (1) whether the
progeny of the two members of a pair that have conjugated show
any unusual likeness in their rate of reproduction; and (2) whether
this degree of likeness occurs also in the members of the split
pairs. If we find such an unusual likeness in the members of the
pairs, and not in those of the split pairs, this will show it to be the
result of conjugation.
An inspection of the ‘totals’ for the numbers of fissions of the
diverse lines during the entire experiment, as given in tables 34
and 35 of my paper on the Effects of Conjugation(this Journal,
vol. 14, page 385) will show at once that there 7s a most striking
resemblance in rate of fission between the progeny of the two
members of the pair; and that this resemblance is nearly or quite
lacking in the case of the members of the split pairs. To bring
this out clearly, it will be well to give here a table showing the
total numbers of fissions in both cases. The period of time was
for the pairs twenty-four days; for the split pairs twenty-one days;
from each member (a and b) of a pair or split pair there was kept
BIPARENTAL INHERITANCE IN PARAMECIUM 453
so far as possible two lines of propagation, x and y; thus we have
in the complete case four lines of propagation from each pair. It
is in comparing the number of fissions in the lines descended from
a of a pair, with those descended from the other member 6} of
the same pair that the striking mY is seen. The data are
given in table 49.
It will be agreed, I think, that the resemblance shown between
the two members of pairs is astonishing. Compare for example
pair 3, where the number of fissions for a were 34 and 34, for b,
32 and 33, with pair 4, where the numbers are 13 and 12 for a,
12 and 13 for 6; with pair 6, where the numbers are 16 and 16
- TABLE 49
‘ Experiment 15. Paramecium caudatum. Numbers of fissions during a period of
24 (21) days, for the progeny of the two members, a and b, of the pairs (conjugation
completed), and for the split pairs (conjugation not completed). Each member is
represented, when possible, by two independent lines of propagation, x and y.
PAIRS } SPLIT PAIRS
PAIR Pr “ b | SPLIT PAIR a | b
le y x y x y ead ake >
rm >| 33°|. 37) at lees | 1 25 | 25 | 23
2 ag° |) si 30 | 2 Bae | oh (COR Meas
3 34 | 34 | 32 | 33 | 3 ig | 24 | 24 | 22
4 NST ie ie oa es 4 DAMN FIA oot
5. 11 | | 10 | 5 23. | 23y,| 2am oe
6 dan P16. || 17 6 fees cr dial 7
7 30 | 32-| 33 | 32 | 7 22 | 26 |
8 | 29 | 27 8 21 25 |. 24
9 35.) 381" 32 9 ||. Se oar a) a5
10 | 29 | 31 | 30 10 22°) FA) Gr. 225
ll 32 | 31 | 28 1 23 | 92 | 4 | 25
12 33/0 Wes2e fh 20: Bs 12 250 | ORs Nee 22
13 29 | 26 13 D2! MAW seme |. 4
14 By ie tee a ess Wag ee 14 26 | 2A
15 1 Wee I ee a a 15 94 | 2 | 23
16 33 | 33 | 27 16 | 26 | 24 | 23
| | 17 | 25 | 23 | 24
18 | 26 | 24
19 7d fe ae
| 20 | 25 | oe
21 [P26 oleeeme esi’) 21
22 Hse uae 2amiy 26
454 H. S. JENNINGS AND K. S. LASHLEY
for a, 17 for b; et cetera. When it is recalled that all the lines
were propagated with the most extreme precautions for keeping
them uniform; that the progeny of a and 6 were cultivated in
separate dishes, handled separately, and their records separately
kept, to be compared only at the end, it will be evident that there
is a most striking degree of resemblance, due to intrinsic causes,
between the progeny of the two members of pairs.
This comes out strongly when we ask the question whether the
differences between the two members of the pairs are greater or
less than the differences between two individuals compared at
random. This is readily worked out from table 49 by the method
described on page 425. The average difference in number of fis-
sions for 24 days between the two members (a and 6) of the pairs .
(comparing each a with each 6 of its own pair) is found to be 2.349
fissions. The average difference when each individual is com-
pared with every other is 8.590 fissions. Thus the average differ-
ence between the members of actual pairs is less than one-third
as great as that between individuals taken at random. The
pairing has increased the resemblance between the progeny of the
two members of the pairs in a high degree.
When, on the other hand, we turn to the split pairs we find a
complete contrast in these respects. To begin with, we are
confronted by the fact, brought out in my previous paper (’13,
page 351), that there is no differentiation in rate of fission among
the different lines of propagation (or at least extremely little).
Thus there is little or no opportunity for any special degree of
resemblance between the two members a and 8, of the split pair.
But proceeding to examine the matter directly in the second half
of table 49, we find indeed that no special similarity between the
fissions for the a and 6 of split pairs is seen. The contrast in
this respect with the conditions found in the pairs is most striking.
If we determine the average difference between the two members
of the split pairs we find it to be not less that that between individ-
uals taken up at random. Indeed, for the particular case of
table 49, it turns out that the difference between the members of
split pairs is slightly greater than for random comparisons, the
BIPARENTAL INHERITANCE IN PARAMECIUM 455
former being 2.125 fissions, the latter 1.895. The difference is
without significance.
To obtain a precise expression of the similarity between a and
b in each case, and to compare accurately the pairs and split
pairs in this respect, we must determine for each the coefficient
of correlation between the number of fissions in a and that in b.
‘The data for this are given in table 49. To determine the corre-
lation between a and b, we must enter, in the correlation table
both representatives (x and y) of a mated with both representa-
tives (x and y) of b, giving us (in the complete case) four entries
for each pair (or split pair). In the incomplete cases there will
be either one, or two, such entries for each pair. That is, for pair
1 (table 49) we should enter in the correlation table the matings;
3a X31; 36: 25,37 xX 3ljand 37 <° 23. Proceeding in this
way, we obtain, for the entire period, 43 entries for determining
the correlation between a and b of the conjugants (pairs); 56
such entries for the correlations of a and 6 of the split pairs.
(In practice it is simpler to compute the correlation by the differ-
ence method, without constructing a correlation table, as set
forth on page 429, but the above gives the guiding principles,
whatever the method used).
The correlation between a and 6 was also determined, in this
manner, separately for the fissions during the first half of the
experiment; during the second half of the experiment, and for
the entire period. This gives us three independent determina-
tions, for both pairs and split pairs. The data for these correla-
tions in the partial periods are given in tables 34 and 35 of my pre-
ceding paper (13), so that they need not be repeated here.
The coefficients of correlation between a and 6 for the pairs,
as compared with the same for the split pairs, are given in table 50.
Table 50 expresses in figures the surprising difference between
pairs and split pairs, that is evident on inspecting table 49.
Between the two members of the pairs we have for the entire
period the extraordinarily high correlation of 0.9238, while be-
tween the two members of split pairs there is no correlation what-
ever (the coefficient being as near to 0 as could be expected).
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 3
456 H. S. JENNINGS AND K. S. LASHLEY
TABLE 50
Experiment 15. Paramecium caudatum. Coefficients of correlation in numbers
of fissions for given periods, between the descendants of the members (a and b)
of pairs, and of split pairs,—in the case of conjugation between individuals of
the same pure race.
PAIRS SPLIT PAIRS
Jol nelle ae iss Nene g
Se | BE | = | Be
ro 2 Correlation between | O§ 25 Correlation between
Bes Ee aandb | Eo ge aand
Alas | rate
Total umMeEersss. | 24 86 | 0.9238 = 0.0107 21 112 |—0.0690 = 0.0634
First half........| 12 | 106 | 0.6325 + 0.0393 11 120 |—0.3175 + 0.0554
Second half...... 14 | 104 | 0.9517 = 0.0062 10 168 |—0.17385 = 0.0505
The shorter periods are less significant, since fluctuations in
the environmental conditions produce in so short a time disturb-
ances affecting the coefficients. Yet they show essentially the
same relations as does the entire time; the positive coefficients for
the pairs are very high; while for the split pairs there is no positive
correlation.
The fact that we find for the first half of the experiment in the case of
the split pairs so great a negative coefficient as — 0.3175 naturally calls
for remark. The other coefficients for the split pairs are not greater
than might be expected in the absence of all correlation. But to what
is due this negative coefficient of — 0.3175?
It can, I believe, be affirmed positively that there was nothing in the
conditions of the experiment nor in the manner of keeping the records
that could give rise to any correlation, negative or positive, between the
two members a and b. They were kept in separate sets, handled separ-
ately, and their records were kept separately. With the great number
of experiments in progress, it was impossible for the experimenter to
have any idea as to whether a correlation was appearing or not; or con-
sciously or unconsciously to manipulate the records in such a way as to
tend toward either negative or positive correlation. Moreover the
facts to be entered in the record are so entirely clear as to leave no sphere
of action for unconscious personal bias. There remain then but two
conceivable explanations; the first is that there might be assortative
mating of such a character that a slowly reproducing specimen tends to
mate with a rapidly reproducing one, giving negative correlation. But
aside from the fact that there is no other evidence of this, and that it is
almost inconceivable how it could be brought about, (1) it is known that
the members of the pure race are not differentiated before conjugation
BIPARENTAL INHERITANCE IN PARAMECIUM 457
into slow and rapid lines (see the preceding paper (713), pages 351-354),
so that such assortative mating is impossible; (2) it is not consistent with
the result for the entire period in the present experiment, where the corre-
lation for the split pairs is practically 0. We are therefore driven, as the
only possible alternative, to the explanation that this negative correla-
tion is merely a chance result, such as would occur now and then if the
experiment were repeated many times. Comparison of the coefficient
(~ 0.3175) with the probable error (0.0554) given in the table would
seem to raise difficulties for this, but just what the probable error should
be is extremely doubtful. The probable errors given in the table are
based on the number of entries or cases compared, these being given in
the first and fourth columns of table 50, for the coefficient under discus-
sion this number is 120. But whether we should use in such cases the
number of pairs of entries (60 in place of 120), or the number of actwal
‘lines compared in the experiment (44), or the number of actual pairs
(22), for computing the probable error, appears not to be established.
If the last named figure is the correct one, the probable error for the co-
efficient — 0.3175 would be + 0.1293, which, being more than one third
the coefficient, would readily reconcile the latter with the explanation
given. Such a change in the method of computing the probable error
would not, however, cast any doubt on the validity of the high correla-
tions found for the members of the pairs. For the pairs the probable
errors would have to be based on the number 16; this would give for
the entire period the coefficient 0.9238 = 0.0247; for the first half;
0.6325 + 0.1012; for the second half, 0.9517 =0.0161—so that the security
of the main results is not altered.
Summary. The results of this experiment with a pure strain
therefore complete the demonstration, given in former experi-
ments, that conjugation results in bringing about a resemblance
between the progeny of the two members of a pair. If any doubt
was possible in the case of conjugants derived from wild cul-
tures, there remains none what ever with the results from a pure
strain. Here all the prospective conjugants are alike before con-
jugation; and there is no positive correlation whatever between
the progeny of the prospective members of pairs. But after con-
jugation such correlation comes into existence and rises to an
extraordinarily high figure, the progeny of the two members show-
ing a most surprising correspondence in rate of fission.
Question may be raised as to the applicability of the term
‘biparental inheritance’ to this result. If employed, one of course
must not understand by it the production of a resemblance to
both parents, for in the present experiment all the parents were
458 H. S. JENNINGS AND K. S. LASHLEY
alike in respect to the character studied; yet from some pairs the
progeny had a high rate of fission, others a low one. But all
recent work has emphasized the fact that the limiting of the con-
cept of inheritance to resemblance is purely artificial; the pro-
geny often inherit from a parent something which makes them
quite unlike the parent. In the present case it appears clear that
both parents do affect both sets of progeny, otherwise the latter
would not be alike; it appears proper therefore to speak of this as
biparental inheritance.
To fit the results in this case to any scheme of inheritance known
for other organisms appears difficult. This is the first demonstra-
tion of biparental inheritance in cases where the two cells that
conjugate both continue to exist and reproduce. If the two par-
ents were ordinary heterozygotes, alike in their germinal charac-
ters, it does not appear clear why the two sets of progeny should
resemble one another so closely. One might expect that they
would often receive different combinations of germ plasm. But
it is not worth while to speculate on this aspect of the matter till
the facts are better known. An experiment with larger numbers
of pairs and split pairs, derived from a single individual, would
help greatly in the interpretation. But the fact of biparental
inheritance in the conjugation of infusoria is clearly established
by the present results.
GENERAL SUMMARY
After conjugation in Paramecium, usually a considerable
number of the lines of progeny descended from the conjugants
die out or are weak.
Since in many of these cases the lines descended from the two
members of a pair differ in their fate, one dying or reproducing
slowly, while the other lives and reproduces vigorously, it has
been held that this indicates an incipient sexuality—the ‘male’
reproducing little or not at all, the ‘female’ reproducing vigor-
ously. The two members of the pair were thus held to be less alike
in their vitality and reproductive power than would be the case
if the deaths and the variations in reproductive power were dis-
tributed without reference to the pairing.
BIPARENTAL INHERITANCE IN PARAMECIUM 459
A method of analysis for determining whether this is correct
was worked out, and is described in Part I. Analyzing in this
way the data from a large number of experiments performed
partly by the present authors, partly by Miss Cull (’07), the
following is found:
1. As to survival and death the fate of the two members of a
pair is more alike, not less alike, than would be expected if the
distribution of deaths has no relation to the pairing. If one mem-
ber of a pair survives, the other member tends to survive also;
if one dies out, the other tends to die out also. Thus conjugation
has the effect of making the progeny of the two members resemble
each other in vitality.
This effect of conjugation is very decided, so that the number
of pairs in which both members survive is much greater than would
be the case if the distribution of deaths and survivals were inde-
pendent of the pairing.
The fact that this likeness in vitality is a consequence of con-
jugation, and does not exist before it, was shown by an extensive
experiment (IExperiment 17) with 239 split pairs. The two pros-
pective members, separated before conjugation occurs, show no
such strong teridency to likeness in fate as is shown by the two
that have conjugated.
2. As to the rate of reproduction, the two members of a pair
are more alike, not less alike, in their rate of reproduction, than
would be the case if the variations in reproductive vigor are dis-
tributed independently of the pairing. Thus conjugation has the
effect of making the progeny of the two members alike in their repro-
ductive power.
Thus these relations give no evidence for sexuality considered
as a tendency for the two members of a pair to be diverse in vital-
ity and reproductive power; the condition actually existing is the
reverse one.
What they show is that biparental inheritance occurs as a result
of conjugation, the vitality and rate of reproduction being affected
by both parents, so that the progeny of the two resemble each
other in these respects.
460 H. §. JENNINGS AND K. S. LASHLEY
Biparental inheritance in rate of reproduction was tested by
an extensive study of the coefficient of correlation between the
numbers of fissions in the descendants of the two members, for
given periods. A decided positive correlation was found, ranging
usually at about 0.3 to 0.4, but rising as high as 0.9, in the case of
conjugation among the progeny of a pure race.
In the case of mixed cultures it was found that there is a slight
and varying degree of correlation due to assortative mating, so
that it is shown by the progeny of members of split pairs. Thus
assortative mating occurs with reference to reproductive vigor; this
is probably a secondary consequence of the assortative mating
based on size, which was previously known to exist.
But this correlation is increased by conjugation; showing that
conjugation produces biparental inheritance.
Demonstration of this is most striking in the case of conjuga-
tion among the descendants of a single individual (a ‘pure strain’).
Here there is no assortative mating with respect to fission rate;
all the individuals having before conjugation the same rate.
There is thus no correlation in fission rate for the members of
split pairs.
But after conjugation there is a differentiation as to fission
rate among the pairs, and the two members of a given pair show
a most striking correspondence in rate. The coefficient of corre-
lation rises in such cases to 0.9.
Thus biparental inheritance is (for the first time) demonstrated
to exist as a result of conjugation in infusoria.
LITERATURE CITED
Caukins, G. N. 1902 Studies on the life-history of Protozoa. I. The life-cycle
of Paramcecium caudatum. Arch. of Entw.-mech., Bd. 15, pp. 139-186.
Cutt, 8. W. 1907 Rejuvenescence as the result of conjugation. Jour. Exp.
Zool., vol. 4, pp. 85-89.
Decen, C. F. 1824 Tabularum ad faciliorem probabilitatis computationem
utilem Enneas. Copenhagen.
De Moraan, A. 1845 Article on ‘Probabilities,’ Eneyclopedia Metropolitana,
vol. 2.
Harris, J. A. 1909 A short method of calculating the coefficient of correlation
in the case of integral variates. Biometrika, vol. 7, pp. 215-218.
BIPARENTAL INHERITANCE IN PARAMECIUM 461
Jennines, H. 8. 1911 Assortative mating, variability and inheritance of size,
in the conjugation of Paramecium. Jour. Exp. Zoél., vol. 11, pp. 1-134.
1913. The effect of conjugation in Paramecium. Jour. Exp. Zodl.
vol. 14, pp (279-391)
Peart, R. 1907 A biometrical study of conjugation in Paramecium. Bio-
metrika, vol. 5, pp. 213-297.
PEARL, R., anD McPuHetTers, L. 1911 A note on certain biometrical computa-
tions. Amer. Nat., vol. 45, pp. 756-760.
APPENDIX
FORMULAE USED IN THE ANALYSIS
For convenience of reference there are brought together here
the formulae and rules employed in the foregoing paper, for analy-
zing the question as to the number of pairs included, when a cer-
tain number survive (or die), out of a number of paired individ-
uals. In these formulae:
m = the total number of individuals
l = the total number of pairs (so that l= a)
n = the number of individuals drawn
k = the number of pairs obtained
x = the probability for any given number & of pairs
1. To determine the average number (xk) of pairs (and for all
practical purposes the most probable number of pairs), to be
obtained when 7 individuals are drawn from m individuals form-
ing l pairs:
= in (n = 1)
m—-1
(1)
The integral number nearest this average is (with rare excep-
tions) the most probable number of pairs; where this is not the
case, it differs by but 1 from the most probable number.
2. To determine with certainty, for even numbers of n, the most
probable number (k) of pairs:
(x + 1) (5+ 1)
m+3
k= (2)
where the integral portion of the result is the most probable
number of pairs.
462 H. S. JENNINGS AND K. S. LASHLEY
3. To determine the probability « of any given number &
of pairs, when n individuals are drawn:
n! m—n! I! 22-2k
= 3
3 mik!in—-2k!l—n+k! (3)
4, Having the probability x, for any given number of pairs
k,, to find the probability x. for the next higher number kz:
(n—2 ke +1) (n—2 ke +2)
4ko-(1— n+ ko)
I2= 1% (4)
5. To determine how probable it is that there should occur a
deviation from the most probable number of pairs as great as
that observed:
Determine by (8) and (4) the probabilities for all numbers of
pairs deviating less than the given deviation. ‘The sum of these
probabilities is the total probability for deviations less than that
observed; the difference between their sum and 1 is the total
probability of a deviation as high as that observed or given.
Dividing the probability for deviations less than that observed,
by the probability for deviations as high as that observed gives
the odds against a deviation so high as that observed.
6. Stirling’s formula for finding n!:
' n>
n! = — V2r n
en
where
BIPARENTAL INHERITANCE IN PARAMECIUM 463
TABLE 51
Experiment 16. Paramecium caudatum. Number of fissions for the 482 lines
descended from the two members, a and b, of 241 pairs of conjugants, for forty
days (March 30 to May 8, 1912). For each line the number of fissions ts given for
each of four successive periods of ten days each; also the total for the entire forty
days. Numbers in parenthesis, as (2), show that the line in question died out
during the period indicated, after the number of fissions shown in parenthesis.
Totals in parenthesis are for lines that did not live through the entire forty days.
e | 2] 2-2 2 m [el] 2) 2] 2 2 | 2 e | 2] a] 2] 2 mn
eseees Be |) oe | | Si Seka Be a ye eee See | oe
BE) leclesiis| BS | so | 8) 2 lesiesies| So [28 |e] 2 lesiesias| BS | 28
Siena |e | & |e lale ee pes fetal e lal ie le let [lee
i/a}~s} io} 9 8 | 35 | 19a] 10) 9 7 «9 | 35 | 37a} 10} 9 8 10 | 37
|b} 9 11] 9 12 | 41 | |b} 8 7 8 in | 34 1b} 8 910} 10 | 37
Ha} 711} 8 8 | 34 | 20a] 8 1011] 12 | 41 | ag}a! 9 10, 10) 1Z | 41
Pera, ai) 2h) 2) as | b.| 10} 9} 9} 10 | 38 /b| 9 9 12) 13 | 48
3) | 9| 12) 5) (7) | (88) |} 21] a] 10) 10) 8} 9 | 37 | aoa} 7 9 9 9 | 34
b| 11] 12} 9} 14 | 46 | |b{10/10' 8} 10 | 38 }b| 7. 8 9| 10 | 34
Ala} 911} 8 12 | 40 | 22} a] 11) 19) 8] 11 | .42-) 40] | 12) 911] 10 | 42
ib] 9 8| 9} 10 | 36°) |b] 41} 10 7} 9- | 37 |b | 10} 11) 11) 14 | 46
5] a] 10) 11) 9} 12 | 42 | 23) a} 10; 6 2 (S) | (23) | 41] a] 10] 11) 10] 12 | 43
b| 910 8 9 | 36 | |b!| 9/10 9 12 | 40 b | 10] 10} 12] 15 | 47
6) a | 8| 12) 8 11 | 39 | 24) a} 12/10) 10) 12 | 44 || 42} a| 10/10 9} 12 | 41
E | 10 11} 10} 13 | 44 | |b | 11} 10| s| 13 | 42 b| 81011) 16 | 45
Jai) | | | (1) | 25a) 9 9 9 8 | 35 | 43| a| 10) 10) 6| 5 | 31
b |(0) | ©) | |b} 710) 8 8 | 33 b| 10) 9 9} 13 | 41
8} a] 11) 12) 9 10 | 42 | 26a) 8 9 7] 11 | 35°} 44} a} 10) 9 9} 13 | 41
b |(0) } ©) |} |b{o} | | (0) b| 11} 11] 11] 13 | 46
Sa} 9 7 8| 8 | 32 | 27} a} 11) 192/19] 13 | 48 | 451 a] 9110-9] 15.| 43
lb| 8 6 6| 9 | 29 | |b/ 10} 9 9 9 | 37 b| 12) 9} 2} 3 | 26
10} a| 11} 11} 6 4 | 32 | 28a] 11) 12) 9 9 | 41 | 4e at oo)! | (0)
b | 11] 10} 9} 12 | 42 | |b] 10) 11) 11) 13 | 45 lb] 5 a] a} @) | (@)
Hija} 9 11/10) 9 | 39 | 29a] 9] 8! 7] 4 | 28 | 47 a] 5] 5] 5} 4°) 19
bj} i] 9 9 12 | 41 | |b} 11/41) 12} 11 | 45 b:| Spies aie sf 127
3 | 4 10 (11) | (41) |) 30) 2 | 11) 10 10 13 | 44 | 48) a] 11) 11) of 14 | 45
jb} 9 9 8 11 | 37 |b} 11} 9} 8 12 | 40 b | 10) 11) 11) 14 | 46
13) a-|(0)| |} (0) | 31) a] 8 8 7 10 | 33 | 49a} 71) | | (10)
b gs} 9 7 11 | 35 | |b] 9 8-8 10 | 35 | |b|-6 ala) - (10)
I4)a} 911) 6 3 | 29 | 32) a) 11/11) 9 9 | 40 | 50) a} 10) 9 8 11 | 38
|b | 10, 11) 8} 10 | 39 | |b] 910 8 9 | 36 | |b] 10 9 s| 8 | 35
15} a) 11] 6 2} 4 | 23 | 33) a|(0)| | (0) | 51] a} 11) 10) 8} 12: | 41
|b | 1} 10} 9 11 | 41 | |b \() | (0) b}10 9 10} 10 39
16 | 12) 12] 6 4 | 34 | 34a] 7 9113] 13 | 40 | 52a] gio 7] 9 | 35
|b} 9 10/10, 12 | 41 | |b! 7 2 of co) | (9) | |b] 10/11) 12) 12 | 45
172|(0) | | | ©) | 3a] 19 10} of 13 | 42 | 59] 0 10 10 9) 14 | 43
|b] 10 1, 9) 11 | 40 | |b} 911) 10-10 | 40 | | bj 12) 41] 9 13 | 45
oa (CO) | (0) | 36 a | 10) 11) 8 12 | 41 | 54) a] 10) 10) 10) 13 | 43
bj, 11] 6 5 8 | 30 | |b]. 910/10) 4 | 33 | |b} 11/101 10) 9 | 40
464 H. S. JENNINGS AND K. S. LASHLEY
TABLE 51 (ContTINUED)
TOTAL
40 DAYS
Peale ef ek |. a | Slok| £) we z le] Sloe] Sl ok
la lediedias] 8 | 32 | |) a ledlzaiadl fa | da | a | & lealgalaal aa
Ble i2o\Solko| Bo | Ho S| = BO|So|Ho| Bo Ho = a BO OS|HS| Bo
AISI ETIET| 27 | 8" a Be eS) Bo ee Nieto ere
55} a| 10) 8 10) 10 | 38 | 77) a] 2 5) 3} 5 | 15 ‘V99,a] 9 8 9) IF
b| 11/11} 9} 10 | 41 b |(6))_ | (6) b] 9} 9] 12) 14
56, a| 12] 11/11) 15 | 49 | 78} a] 10112) 8| 5 | 35 {x00} a| 10) 11) 11) 14
b| 10} 9/12 13 | 44 | |b] 9} 10) 11) 12 | 42 b| 7 910} 10
57| a| 11] 10| 10) 14 | 45 | 79} a |() (0) |101! a |(0)
b| 10] 11} 11} 12 | 44 | b| 9]. 9 8 (2) )G3)M Gkb | 8) 10) 40) as
58} 2| 10 9 9! 10 | 38 | 80} a| 11) 10 9 15 | 45 |102) a |(0)|
b| 9 10-6) 122 | 37. | |b’) 11) 10) a aa |b | 10} 12| 12} 14
591a| 9] 10 11] 13 | 43 | 8i}a] 9 9} 8| 11 | 37 |no3la] 9} 5] 4) 5
bid) 12) i} 16. 10.) 4} eS -ziel ey 28 b |(0)
60} a} 8} 9} 10| 13 | 40 | 82) a} 11) 11) 9} 12 | 43 |104) a] 9) 11) 10) 11
bt 1} Ta at) 1s |e 45 b] 7 5).5) 3)-] 20 | <1) bid oS) sestoneame
61) a | 13} 12) 9} 14 | 48 || 83] a/(3) (3) |105| a | 10) 10) 9} 12
b| 9 10) 12} 13 | 44 | 10). 8) 12.12") v42 | b | 11] 10) 13] 14
62} a} 11) 11) 11] 18 | 46 | 84a] 10) 9 9 18 | 41 106) a/ 12) 9 10) 8
b | 10) 10} 11) 11 | 42 bal 10), 3) 6) eS at 27 ib} O10)
63} a | 12| 11} 9| 11 | 43 | 85a} 6| 1) 3} 4 | 14 107) a] 10) 11) 10) 11
al tal 69) Ti) 1S 43 i Oey Taal Al OO) | b |(0)
64/ a | 10} 11) 10} 14 | 45 | 86] a] 11] 10/11) 14 | 46 |108 a| 11) 11] 13) 13
b| 911) 10} 12 | 42 || |b] 11) 10) 7 10 | 38 b | 11) 13] 12) 13
65| a | 11] 13, 11) 14 | 49 | 87}/a| 9/11) 8| 13 | 41 |109| a |(0)
bi} 10} 975) 8.) SB ew aera) -9).20)| 13 4) 240 b| 7 9 6 07%
66} a] 11) 10) 8| 13 | 42 | 88 a] 10) 11) 9} 12 | 42 |110/ a] 6(0)
b| 10, 9 8 9 | 36 B} 10) di) <8]. 140"), a9) | BAO)
67)a] 6 6 5 4 | 21 | 89} a] 12) 11) 10) 14 | 46 111} a] 11) 11] 11] 18
b| 10} 9| 8| 12 | 39 b | 10} 10} 11] 11 | 42 | |b] 10] 10) 13) 11
68} a | 11} 12} 12] 15 | 50 | 90} a] 9] 11) 10} 13 | 43 |112ia| 7 6 3 38
b| 11) 11) 10) 12 | 44 b| 9/11/12 15 | 47 b.| | 7) 1208s
69 a] 5) 6 5) 4 | 20 | 91] a| 10] 10) 10) 13 | 43 |413}a] 9} 8} S| ft
b| 11) 10) 9| 10 | 40 b | 10} 10} 13; 17 | 50 b| S| 7) Aaa
70| a | 10} 10; 9| 14 | 43 | 92} a] 10} 10} 9| 13 | 42 |1114) a /(1)
be 10) 10; 9] 12-40") bse hot Tal ia ae a b| 4\(0)
71} a] 9 8 2} 10 | 29 | 93} a] 11) 11) 9} 13 | 44 |115) a] 9} 11] 10) 14
b}] 4, 1) 4 5 | 14°] |b] 10 10) 11) 12 | 48 |b} 8} 9} 11} 12
72} a| 10} 11/12) 12 | 45 | 94a] 8} 6 5) 9 | 28 |li6)a| 8] 9} 9 11
b| 810) 10, 11 | 39 b| 0(0) (0) b | 10} 10] 11) 12
73) a | 14} 10) 11) 14 | 49 | 95} a| 11) 11) 10) 13 5 |/117| a | 10) 10) 11) 14
|b | 12) 11) 10) 10 | 43 |b} 11) 8) 12) 14 | 45 | |b} 10) 11) 13 9
74) | 13) 10) 8 9 | 40 | 96) a) 11) 11) 11) 13 | 46/118) a| 10) 10) 11) 12
b| 11) 11) 10, 12 | 44 | b | 10) 10} 11] 12 | 43 | |b] 9} 10 13) 18
75| a. | 12} 12) 11} 14 } 49 || 97) a/ 10) 7 9 9 | 35 119} a} 8} 9 9 12
b{ 11) 9/11) 13 | 45 /b| 9} 9) 9) 108) 387 fh } bh B) SS Seo
76 a | 10, 13; 9) 12 | 44 | 98) a/ 10 8 6 9 | 33 {120 a| 11) 11) 11) 15
b | 10) 11) 13 15 | 49 b | 0\(0) | (0) |b} 10] 10} 9} 12
iS
BIPARENTAL INHERITANCE IN PARAMECIUM
|
le
| >
| &
ES
13)
or
1 (CONTINUED)
|
|
ae ele sey cen) So Pee lesiecleeiee Pp eaahe leeSiecia sine | 38
Ree eee ye ullicact = (acleo ton ae Meceeoclca ele late |e bet
121] a | 10} 9} 9| 13 | 41 |143/ a| 12/10] 9}. 9 | 40 |165}a| 8] 10 12) 16 | 46
b | 10} 10/11} 10 | 41 bi Ostia 13s 40) vie hboe Sie St Ol sg S36
122) a |(1) @) 144) a | 10/0; Toad 4 itee|"a Ih Si 8} 8) 413 87
Bb} 10).10) 12) 11 | 43 b} 910} 8| 10 | 37 | |b| oO) (0)
123] a | 10} 10} 9} (11) | (40) 1145} a | 10] 12/11) 14 | 47 |167)/a/ 8 9) 9! 13 | 39
eee Wo) 11 |) 86) beh Of SY Ofca3) W839 b (0) (0)
124} a| 10) 5} 4) (2) | (21) |146) a} 8} 8} 3) (3) 7 (22) l63} a] 9) 11) 11] 15 | 46
Beles 7 5) 5) |) 25 b |(0) (0) Cm i Ne) ata
125| a |(0) (0) 147} a} 7) 51/10; 9 | 31 [led a} 9 9] 9] 15 | 42
pr) “9|-9[ S| S| BA b| 3} 1)(0) (4) b| 6 2/(0) (8)
126| a | 10] 10; 10} 10 | 40 |148)/ a] 9/11/10) 18 | 43 |170/a| 8| 9] 9] 12 |: 38
Bei) 9), 8-7) <6 130 be PopdOneshamie Ns aor) sb 7s Shel <3" le eng
27a | 8) 5) 3) 2 | 18 [149] a| 10) 10| 8| 10°) 38 |171/a| 1) *| 14
Baieeriet ly TO} (110, bec30 yO Sie Fi le 235 D/) Sheot Sige |. 20
128} a| 9} 10) 11) 12 | 42 150) a |(0) (0) 172) a] 8 8} 9} 10 | 35
b | 10} 10 10) 9 | 39 b |(1) (1) b] 8 10) 9 8 | 35
fee ab) 4h) 19 M5tha (10h S|. 9 10 | 88) -I173}-a.| OFS) 8 13° | 38
Peon 21a? 6° 16 BOP stOne Ol. 14s |) Aa |e OV shi sit eae |" 3s
130} a} 7 9 11/9 | 36 [152}a} 9| 7 8 11 | 35 174) a} 5/0) (5)
b| 5! 2\(0) (7) Di OS seit |) 32 ho he (0) 1s TO) sey | 44
131} a| 5) 4) 4) (1) | (44) 153} a] 8} 5! 7] 8 | 28 (175) a] 10) 9] 11] 15 | 45
ee Sie i sh th | Se = Eb Or- Si Sih 10 | 37 be | 10) LOCOMIA 42
132) a| 8| 9| 8| 12 | 37 |154) a] 9} 10/10] 13 | 42 {176 | 11) 11) 11) 15 | 48
b| 810 7 10 | 35 | |b 910 8 14 | 41 | |b CO) (0)
133} a| 6} 5) 6] (11) | (28) |1155| a (0) (0) 177] a | 11} 10) 10] 15 | 46
Ot 7) TAT | eb hes 92'S)" 12. | 37 bi | 105), 2a 23
134, 2/10) 6 7}. 7 | 80 |156) a} 9} 10) 12) 13 | 44 [178 al! 5] 9f 7] 13 | 34
b| 0(0) (0) b| 7 7 6 4 | 24 | |b) (0)
i5\a) 9} 7 10; 12 | 38 lz a| 8/11/10) 13 | 42 |179| a |(0) (0)
Bay spe oie sh 13% 1° 88) I 1@) (0) | |b/() (0)
136) a | 4) 10, 11) 14 | 39 {1158} a] 8] 10; 10) 11 | 39 |180/a/ 192) 6 5| 4 | 27
b| 0\(0) (0) b | 10} 11) 10) 14 | 45 b | 10} 10, 9| 10 | 39
137) a |(0) (0) /159) a |(0) (0) |181)a/ 6 7] 6 9 | 28
Ee oh Oe '7) IS. 87 b |(2) (2) | +| bj) (1)
138] a} 11) 9 9} 13 | 42 160) a| 9| 10, 12) 13 | 44 |182)a| 6] 9) 11) 14 | 40
b| 8/10) 9| 13 | 40 B,) 1OaIRE Sie iat | 244 Bel Sia alee: | 93
139} a | 10; 10) 9} 10 | 39 {1161| a |(O) (0) |183; a| 9) 10) 11) 15 | 45
b |(0)) (0) b}| 6/(0) (6) |b| 9/11) 11) 12 | 43
140, a| 7) 9 5| 6 | 27 1162.2) 7 9 9 8 | 33 [184 a/ 10/11) 9} 13 | 43
b |(0)). (0) 2S Sobek | Soa by Seo te) 12 |” 32
141, a/ 9 10) 11} 11 | 41 {163} a] 9/10) 9} 11 | 39 [185) a} 11] 11/11] 16 | 49
by li) 9) 8) 14 | 42 | Sb] Spdordile1h | 40° |< |b I1@) (3)
142} a| 10, 10| 9) 13 | 42 164) a| 10) 11) 12) 15 | 48 |186 a} 9 8 12- We "37
|b} 811) 8 14 | 41 | |b] 10} 10) 9} 11 | 40 b| 8} 9/11] 12 | 40
466
TABLE 51 (ContTINUED)
. JENNINGS AND K. S. LASHLEY
| 10 pays
THIRD:
10 pays
MEMBER |
10 Days
SECOND |
10 DAYS
10 DAYS
MEMBDR
SECOND —
10 DAYS
FOURTH
~ | PIRST
|
MEMBER
SECOND |
10 DAYS
o
191
_b
192) a
Fi
oo
a
Sorc
a
sIST Oo © © = © ©
a |(0)
_
Oo © 0
Serer oe Of op
STrmrmprerrer oer Of
TeTrTeTrerTerTerrTerTe Te oper rerTe rere rere oe of
— ae ers a i -
S
~—
6) 5] 4
4| 1] 2
(1)
co) |
8 ‘ 2
\(1)
| 9) 10] 10
0\(0)
10) 10) 11
9} 11) 11
li] 8 8
9} 9 10
FOURTH ©
10 DAYS
CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT
HARVARD COLLEGE, NO. 235.
THE REACTIONS OF ARTHROPODS TO MONOCHRO-
MATIC LIGHTS OF EQUAL INTENSITIES
ALFRED O. GROSS
FORTY-FIVE FIGURES
CONTENTS
PIE ROUUC ILO MMC mets. eT Re eM Se Wo. tials e arebe se steroersrota ore Sys 468
2 TERETOA Chl 6 A285 Recs \y ol eee en Ae aS Fe A pe Ao Be RE 469
ARVN HENRI Se Pea et Bye Ae aPee RT) SAL, 28 Learn Sue IRENE ioc nove, cs Sart Se TNR oie be 5 475
AMUSE LULONS. 15. es era es PS no scion Mag pm sets eietbadee cag 476
A. Calliphora erythrocephala Meigen (larva).................-...-e006: 76:
Heme V Tea eet Senet eae pepe ts es Sosy BI 6 Fc aus, Segre ae tote oad ete 476
Zee (NOUS hes ee rere eh a Sof . 3. SR Rae ore ieee 477
VOR UGS el te: << RAIA oo oso Bi blogs SO Sere erie coe i Om eee ATT
Be Calliphorajerythrocephala’ Meicen (adult). . 2... 26... J... s- 25. +e 486
ie Wiahentales. 2. peer amen. fo, Rs. <=. a1: Se eaberee hs ele ee Ree 486
2: Methods. 254.55 stn. 5b lh eh RE 5 PI EN Sane Ae 2 487
SS ALVCS Ul GS serpent Rees cee ce ca, oi tis sons je OR sam «eeu bee see 487
©. Draspphila ampelophila Loew: (adult)... ...... 04025. bs ee se cone ene 490
PICU LES CIG es cn. MMC E NT TNS Gekko «0 SY igs Ra ele « os REN RD
De NTE TNO Sprit sey o's MES Ara ns tes 6 tect eee ee OO 5 SEAS Siar een 491
DURLVCSULISEs SIE). 3). Hes e acls @e tots 2h. Slee 28 <3 ee Se a 491
> Reucera pyrina Lanneé (Wary) 2... ee oo ced Ss aco se a ene 495
APU LEST GAL ome! 22 bate RMIT oc>, way ERIS es ena « dk = ete he SE 495
De IVIETINOGS ../\Serscaets 2s Ome eee ORME ao ee ee eae 496
oe: Resullti#y . 2. POPE, «cal ONO S PEE eae 2 ea PERL Gta ie ace ne ee 496
Boweltia subpethica taworth. (adit). issn sas Pn et < oA ee: 500
I VE UDETU A eee he oo <<. ee get ey ec cn ulate aSe 500
2. Methods:,.2 7... ee Coes ees ari SS MR RDS ce 500
BUCS Gas: 2 Uae. ASME MD IDL a Be ES SP ee 501
F. Periplaneta americana Tanné-(adult) .... 2.05.00... 620 osc ol. 501
LeQN) BYR S Oe ee, See ee Ren aes, Te Ce. Ue 502
Rat US 7] 10 LR pe ni 2 OU =P a 502
pe 215101 Rie i a4 Ce SLY oA UN aed a a | os ee 503
see DIRCUASION fans is ae tie Set LVAD CAL 1 Cot Erne HRS Bre Go? 5 Sh eee 506
Reco TTI 4 25 5 hae a A ie A RES AEE RET tins ah) oe eee 511
FRI READY.) « <x 2 acorns SOR es Oe oa. FOR ee §12
467
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 4
MAY, 1913
468 ALFRED O. GROSS
1. INTRODUCTION
Because of the abundance of arthropods, especially insects,
the ease with which they may be secured and their adaptability
to experimental work they have been used in many investiga-
tions of light reactions. A considerable number of these investi-
gations, made both in the field and in the laboratory, have
included tests upon the reactions of these animals to lights of
different colors. The light used in the field consisted of light
reflected from colored objects, such as flowers, colored glass,
paper, or cloth. For the experiments performed in the laboratory
the different colors were secured by filtering white light through
colored solutions or glass, or else from a prism spectrum.
It is very apparent that light reflected from objects in the
field, though having the advantage of a nearly natural condition
for insects, is very unsatisfactory since the different colors thus
produced are not optically pure or of the same intensity. One
could not be certain under these conditions whether the response
of the organism was due to the intensity of the light or to its
specific quality. Furthermore there are so many other uncon-
trollable factors which may be involved in experiments performed
in the field that the results of these investigations are of little
value in solving the problems of the reactions of animals to
lights of different colors.
Colored lights produced by means of screens of glass or solu-
tions are also unsatisfactory since these lights with the possible
exception of the red are never pure but contain more or less of
a mixture of rays. Such screens have been found to transmit
the invisible heat rays also which when present undoubtedly
have their effect in stimulating organisms.
For these reasons the use of colored light reflected from dif-
ferent surfaces or that produced by means of screens is to be
avoided in experiments involving a careful analysis of the reac-
tions of animals to this form of stimulus. At present the spectrum
is the best means of securing monochromatic lights because of
the purity of its colors, but it is little better than light produced
REACTIONS OF ARTHROPODS TO LIGHTS 469
by means of screens, unless the different colors are made equal
in intensity.
Many investigators have ignored the important factor of
intensity and have ascribed the effect of colored lights on the
organism as due solely to the quality of the light. Others have
recognized the importance of intensity, but as far as I know the
only experiments on animals in which this difficulty has been
successfully overcome are those described in two very recent
papers published by Day (711) and Laurens (711). These two
investigators measured accurately the intensity of monochro-
matic spectral lights by means of an extremely delicate instru-
ment, the radio-micrometer of Boys. This apparatus has opened
the way to the correct solution of many interesting problems
involving the reactions of animals to colored lights, investigations
which heretofore have yielded so many conflicting and perhaps
questionable results.
This present investigation was taken up at the suggestion of
Prof. G. H. Parker and whatever success has been attained is
due to his untiring interest and helpful criticism throughout the
whole course of the work. Iam also indebted to Dr. H. Laurens
for codperation in the construction of the light generators used
in the experiments.
: 2. HISTORICAL
Much of the earlier work on the reactions of arthropods to
colored lights was taken up from a purely psychological point
of view. The investigators seem to have worked with the sole
purpose of answering the question of whether the lower animals
are able to perceive colors as such and, if so, do they perceive
the same colors of the spectrum as seen by the normal human
eye.
The first recorded experiments, to my knowledge, upon the
reactions of the arthropods to colored lights are those of Paul
Bert who published an account of his work on Daphnia in 1868.
He discovered that the Daphnia responded to each of the visible
colors of the spectrum. When the entire spectrum was thrown
470 ALFRED O. GROSS
on a trough containing these animals the majority of them col-
lected in the green and yellow, the most luminous region of the
spectrum as judged by our eyes. From the results of these
experiments Bert concluded that the vision of the lower organ-
isms is the same as it is of the eye of the normal human being.
Merejkowsky (’81) experimented with spectral colors upon
Balanus larvae and the marine copepod Dias longiremus. He
recognized the importance of the intensity of the light and.
attempted to eliminate this factor by equalizing the luminosity
of the respective colors as judged by his own eye. Under these
conditions Merejkowsky found the animals distributed equally
in the different colors. He opposes the view of Bert concerning
the vision of the lower animals and concluded that the lower
crustaceans cannot see the different colors and are conscious of
only one color in variations of intensity. ‘‘ Nous percevons les
couleurs comme couleurs, ils ne les pergoivent que comme lumiére”’
(81, p. 1161).
Lubbock (’79, ’81 a) tested the reaction of ants with lights
produced by means of colored glass placed over different parts
of artificial nests. From his numerous and ingenious experi-
ments he concludes that ants have the power of distinguishing’
color, that they are very sensitive to the violet and that their
perception of color is very different from ours. In two papers
published in 1881 and 1883 he describes experiments made with
spectral light upon Daphnia pulex. His results are in agreement
with those obtained by Bert except that he found the Daphnia
to be responsive to the ultra violet—a fact denied by Bert. In
a more recent publication (’04) are described an extensive series
of experiments upon bees made in the field. Lubbock placed
drops of honey on pieces of colored paper and glass and observed
the bees which visited the different colors. The results of his
investigations, he thinks, prove that bees can distinguish colors
and that they exhibit a decided ‘preference’ for the blue. Lub-
bock’s view, that insects can perceive the different colors, is in
general agreement with the results of the work by Forel (’88, ’01,
04) on ants and bees; the Peckhams (’87, ’94) on wasps and
REACTIONS OF ARTHROPODS TO LIGHTS A471
spiders; Perez (’94) and Bethe (98) on bees; and Fielde (’02)
on ants.
Graber (’83, ’84) made an exhaustive series of experiments
upon fifty-three different species of animals. Among the arthro-
pods used for these tests twenty-seven were insects and two were
spiders. Graber also attacked the problem from a psychological
point of view in an effort to determine whether the lower animals
are able to distinguish colors and intensity differences... He studied
the distribution of the animals in an apparatus of two compart-
ments illuminated by lights of different colors. Graber used
_ sereens to obtain his colored lights. From the results of his
large number of experiments he reached the following general
conclusion—‘‘dass die leukophilen oder weissholden Tiere mit
geringen Ausnahmen alle blauliebend, die leukophoben oder
dunkelholden hingegen rotliebend sind” (’84, p. 245).
Graber also experimented on blinded cockroaches in an effort
to ascertain whether these insects are able to perceive «olors
and intensities through their chitinous integument. He blinded
the cockroaches by covering the surface of their heads with a
layer of warm black wax about 3 mm. thick and found when
red and blue screens were used the greater numbers still congre-
gated on the side of the apparatus illuminated with red. Graber
concludes from this experiment—‘‘dass die geblendeten Kiichen-
schaben auch farbenempfindlich resp. blauscheu sind” (’84, p. 307).
Gratacap (’83) in a short paper discusses some experiments
made, in the open at night, upon the responses of nocturnal
Lepidoptera. He placed colored tissue-paper cylinders over kero-
sene lamps and found the moths exhibited no marked ‘prefer-
ence’ for one color over any other. More moths were attracted
to the white light than to the colored lights, probably because
the light transmitted by the white paper was of greater intensity.
Loeb (’90, ’93, ’05) studied the reactions of animals to colored
lights from the standpoint of the effectiveness of the different
rays in orienting the organisms. He objected to the application
of the expressions used by the so-called ‘anthropomorphists’ that
animals ‘love’ or ‘prefer’ certain colors and ‘hate’ or ‘dislike’
472 ALFRED O. GROSS
others. Loeb placed caterpillars of Porthesia chrysorrhoea under
a blue glass and noted that they oriented to the blue light in the
same way as when placed in white light. . If, instead of the blue
glass, he used a red one the larvae remained indifferent. He
performed a number of varied experiments on this caterpillar
and repeated similar ones on moths of Sphinx euphorbiae and
Geometrica piniaria, plant lice, Musca larvae, larvae of Tenebrio
molitor and of the June-bug, Melolontha vulgaris, Limulus poly-
phemus, Polygordius and several species of copepods. In all
these experiments he used only two colors, red and blue, obtained
by means of screens. Loeb (’05, p. 182) concluded from the
results of his investigations that ‘‘The more refrangible rays of
the visible spectrum are exclusively or more effective, than the
less refrangible rays, in causing the orientation of the animals
as is also the case in plants.”
Plateau (’97, 99) following the work of Lubbock and Forel,
studied the behavior of the bees in an effort to determine whether
the insects, in their visits to the flowers, were guided by the
different colors. Plateau, instead of using artificial colors, experi-
mented on natural flowers the colors of which, he thought, were
of equal brightness. He believed that bees are directed not by
color but by the sense of smell, a view in opposition to that
held by Lubbock and Forel. Plateau (’97, p. 41) concludes that
‘“Tls (the insects) ne manifestent aucune préference ou antipathie
pour les couleurs diverses que peuvent présenter des fleurs des
différentes variétiés d’une méme espéce ou d’espéces voisines.”’
The view that bees are not directed to the flowers by the
different colors is also held by Bonnier (’79) and Bulman (99).
Yerkes (99) made a series of experiments upon the small
crustacean Simocephalus vetulus with spectral light. When the
gas spectrum was thrown on a trough containing these animals,
the majority collected in the regions of the yellow and red light.
He then placed a prismatic glass box containing an India ink
solution between the trough and the spectrum. The greatest
depth of the solution was placed over the red and yellow so
that the intensity of this end of the spectrum was equal or less
than that of the blue and green region. Under these conditions
REACTIONS OF ARTHROPODS TO LIGHTS 473
Simocephalus showed a movement toward the violet. Yerkes’
(99, p. 182) conclusions are “‘Simocephalus prefers the orange
and yellow of a gas spectrum. This response to the spectrum
is a photopathic reaction, and is not, so far as is known, chromo-
pathic.”
Hess (710) has made some recent investigations on the reactions
of invertebrates to spectral light. A large part of this work
was devoted to experiments upon insects and small crustaceans.
Hess placed larvae of Porthesia chrysorrhoea at the bottom of a
parallel-walled glass vessel. In the dark the larvae remained
on the bottom but when the vessel was illuminated from the
side by the spectrum, the animals immediately started to crawl
upwards,in the yellow and green, the most luminous region of
the spectrum. In the blue and red the less illuminated areas the
larvae remained below. When only two colors, the red and blue
of either spectral or screened light, were used, the larvae were
most responsive to the blue. This agrees with Loeb’s results
secured by tests on various animals with these two colors. Hess,
however, does not agree with Loeb in believing that the more
refrangible rays are the most effective of all the spectrum, since,
he has shown that the larvae are more responsive to the green
and yellow regions of the spectrum than to the blue end. Hess
found that Hyponomeuta variabilis, Dasychira fascelina, Lasio-
campa potatoria and Phragmatobia fuliginosa responded to the
spectrum and to the colors in a way similar to that of the Porthesia
larvae. Among other arthropods, he tested the effect of colored
light on the movements of the eye in Daphnia and the aggrega-
tion in the spectrum of the crustacean Cypridopsis, the larvae
of Culex pipiens, Musca and Chironomus plumosus, the adult
insects Coccinella septempunctata, Culex pipiens, bees, house
flies and ichneumon flies and the marine crustaceans, Podopsis
slabberi and Atylus swammerdamii. In all of these forms Hess
found the yellow-green region of the spectrum most effective in
stimulating the animals. The animals which were positive in
white light aggregated in the yellow and green; those which are
negative, in the violet and red. As to the vision in the inverte-
brates Hess maintains “dass bei allen bisher untersuchten Wirbel-
474 ALFRED O. GROSS
losen die Kurven der relativen Reizwerte der verschiedenen
homogenen Lichter anniiherend oder ganz ubereinstimmen mit
der Helligkeitskurve fiir den total farbenblinden Menschen bei
jeder Lichtstirke und fir den dunkel adaptierten normalen
Menschen bei entsprechend lichtschwachem Reizlichte.”’
In this brief review of the literature pertaining to the reactions
of the lower crustaceans and insects to colored lights, we find
a great diversity of results and opinions. Merejkowsky (’81)
maintains that the difference in response exhibited by the ani-
mals to the different regions of the spectrum is due not to the
quality of the light but to the relative intensity of the colors.
Hess (’10) believes that the relative attractive power for the
different colors approach or correspond with the brightness curve
of the totally color-blind persons. Gratacap (’83), Plateau (’97,
99), Bulman (’99) and Bonnier (’79) believe that insects are
not guided in their movements by color, and Yerkes (’99) states
that the factor of intensity has the more important role in the
aggregation of Simocephalus in the yellow and red regions of
the spectrum. On the other hand Bert (’68), Lubbock (’79, ’81,
’81 a, ’83, ’04), Graber (’83, 784), Loeb (’90, ’93) and others
believe that the lower animals perceive or are stimulated by the
different colors as such. With the exception of the work on
Simocephalus, about which there seems to be much difference
of opinion, the results of these investigators in general support
the conclusion that the blue or the more refrangible rays of the
spectrum are most effective in producing a response in the ani-
mals which they tested. It is apparent from the work thus far
done that it is extremely doubtful just what part the intensity
and what part quality of the light has in stimulating the organ-
ism. The great variations in results and the conflicting opinions
concerning the relative efficiency of the colors is doubtless due
to the fact that in none of these experiments was the intensity
factor eliminated. Furthermore the lights used in many of the
experiments were produced by means of screens which have been
shown to be unreliable.
It is the purpose of this paper to present the results of inves-
tigations on the reactions of some of the lower animals to spec-
REACTIONS OF ARTHROPODS TO LIGHTS 475
tral lights of equal intensity, measured, not by the human eye,
but by accurate physical means. In this work are used five
species of insects belonging to three different orders as follows:
the larva and the adult of Calliphora erythrocephala Meigen and
the adult Drosophila ampelophila Loew of the Diptera, the
Zeuzera pyrina Linné larvae and the adult of Feltia subgothica
Haworth of the Lepidoptera, and the adult Periplaneta americana
Linné of the Orthoptera.
3. METHODS
The apparatus used in these investigations is the same as
‘that described by Laurens (’11, p. 258) in his paper on the reac-
tions of toads to monochromatic light. An account of the con-
struction of the generators is given in greater detail in a paper
published by Day (11, pp. 310-315). This paper also includes
a short description of the radiomicrometer used in measuring
and equalizing the intensities of the colored lights.
The essential features of the light apparatus are two light
generators placed at opposite ends of a dark chamber 80 cm. deep,
130 em. long and 70 em. high. In addition to the dark chamber
suitable screens and reflectors were employed to exclude, from
the animals during the experiments, any light not proceeding
directly from the prism. In order to reduce the amount of
diffuse light from the outside to a minimum the entire apparatus
was constructed in a large dark room arranged for the purpose.
The special accessory apparatus for exposing the animals to the
light, which was adapted according to the nature of each organ-
ism to be tested, will be briefly described in the account of the
experiments given for each species.
The colored lights used in this work were four in number as
follows: blue, 420 to 480 wu; green, 490 to 550 uu; orange-yellow,
570 to 620 uy; and red, 630 to 655 uu. These colors were
obtained by cutting down the spectrum by means of diaphragms of
blackened cardboard with narrow vertical slits of appropriate size.
The sources of, the lights were Nernst glowers on a 220-volt
circuit. In order to equalize the intensity of the four colored
lights it was found desirable to used one glower for the red, two
476 ALFRED O. GROSS
glowers for the yellow, and three glowers for the green and for
the blue lights. The finer adjustments of the intensity were
accomplished by regulating the size of the diaphragms until each
light gave the same reading on the radiomicrometer, that is, the
lights were made to contain the same amount of radiant energy.
When the lights from each generator were thus equalized any
consistent difference in the responses of the animals to the dif-
ferent colors was believed to be due not to the quantity of the
light but to its quality.
4. OBSERVATIONS
A. Calliphora erythrocephala Meigen (larva)
Because of the reversal of their phototropism during develop-
ment, certain species of flies such as the blow-flies are very inter-
esting from the standpoint of their reactions to light. When
the blow-fly larva first emerges from the egg, it is either indiffer-
ent, only slightly negative, or, as Herms (11, p. 177) has shown
for aggregate larvae, even positively phototrophic to light. As
it grows, it becomes more and more responsive to directive light
and by the time the feeding period is ended, it is very strongly
negative in its response to light. When the adult fly emerges,
it is no longer negative but strongly positive. This complete
reversal of its phototrophic behavior is closely correlated with
the habit of the animal.
1. Material. The blow-fly, Calliphora erythrocephala Meigen
is easily reared in the laboratory throughout the year and hence
it is excellent material for experimental purposes. For the inves-
tigations about to be described the larvae were kept in culture
jars which were constructed by placing large lamp chimneys on
earthenware plates of suitable size, and filling them with about
two or three inches of moist sand. This sand served to absorb
the excess of liquid from the food. Codfish heads were found
to be a convenient and desirable food for the larvae. The heads
were renewed every few days in order to keep the cultures in
the best condition. After the adult flies had deposited their
eggs on a piece of fish, the latter was placed in a culture jar
REACTIONS OF ARTHROPODS TO LIGHTS A477
which was left in a warm dark room of comparatively uniform
temperature. The larvae selected for experimental purposes
were about five or six days old, an age at which they seemed
most active and most responsive to light. The animals were
always dark-adapted and were carefully guarded from light for
several hours before the experiments.
2. Methods. The accessory apparatus for the experiments on
the blow-fly larvae was comparatively simple, being merely a
small table 18 cm. in height, supporting a thin piece of slate
30 em. long and 25 em. wide. This piece of apparatus was
placed inside the dark chamber midway between the two gener-
ators. The apparatus was arranged so that an animal placed
at the center of the slate was illuminated by lights from two
opposite sources, each of equal intensity. The surface of the
slate was frequently moistened with warm water to facilitate
the movements of the larvae and to prevent them from following
the old courses of other individuals. In addition to these tests,
the larvae were made to trace their own courses on paper with
dilute solutions of methylene blue as described by Herms (’11,
p. 189). The rate at which the larva moved was indicated by
marking on the paper the position of the animals at the end of
each ten seconds.
Different individuals, even when of the same age and reared
in the same culture under identical conditions, varied more or
less in their responsiveness to light. The cause of this difference
was not ascertained but probably depended on some physiologi-
cal condition at present unknown.
3. Results. The animals were first tested with each of the
four individual colors to determine whether they were responsive
to the various wave lengths of light of an intensity used in these
experiments. In each test the larva was placed next the side
of the slate nearest the source of light. Its subsequent course
after orienting was traced on the slate or else on the paper with
the methylene blue solution. Each individual was given two
_tests, the first with the source of light on one side, the second
with it on the other. The larva was then put aside and a new
478 ALFRED O. GROSS
one taken for the next two tests, et cetera. This use of both
lights eliminated from the results such errors as might arise
from defects in the apparatus, odors, diffuse light, et cetera.
The paths plotted by the larvae under the stimulus of blue,
green, yellow and red respectively -are shown in figures 1 to 4
inclusive. Under such conditions the larvae oriented and moved
away from the light on a course approximately parallel to the
direction of the rays. It is apparent from these experiments
that the blow-fly larva is responsive to each of the colored lights
when of an intensity used in these investigations. As far as I
was able to determine from these simple tests, however, the larvae
showed no appreciable difference in their response to the several
lights.
Opposed lights of equal intensity and of the same spectral
quality were then used in testing the larvae. When two lights
were used, the larva was allowed first to orient definitely under
the influence of one light’ and after this had been accomplished
the other light was thrown on. The position of the larva at the
time the second light was turned on was indicated on the record
in order to know over what part of the course it was under the
influence of the two lights. As the larva changed its position
the slate or the record paper, on which it was crawling, was
shifted a corresponding amount to keep the animal in the center
of the illuminated area and equi-distant from the two sources
of light. In general the larvae turned at right angles to the
direction of the rays when they were exposed to both’ lights.
Figures 5 to 8 inclusive are records of larvae tested with pairs of
blue, green, yellow and red lights respectively. Here again I was
unable to perceive any marked and consistent difference in the
responses of the larvae to the first three of the above lights. In
these cases the angle marking the change in the course was
generally sharp and more or less abrupt. When red lights were
used the angle in the path made by the larvae was seldom so
abrupt but there was more usually a uniform and prolonged
curve as shown in figure 8. However, even in red light, if the
larva is permitted enough time in which to crawl, it eventually
orients so that its median plane comes to lie at right angles to
REACTIONS OF ARTHROPODS TO LIGHTS 479
Figs. 1 to4 Paths traced by Calliphora larvae, in dilute solutions of methyl-
ene blue, in response to the single monochromatic lights, blue, green, yellow and
red respectively.
Figs. 5 to 8 Paths traced by Calliphora larvae in response to balanced pairs
of monochr omatic lights of equal intensity and of the same spectral quality.
Pairs ot blue, green, yellow and red lights respectively.
480 ALFRED O. GROSS
the direction of the rays. This peculiarity in the responses of
the animals to red light as contrasted with their reaction to
other balanced pairs of colors indicates that red is less effective
in orienting the organism than are the other lights, a conclusion
borne out by other lines of experimentation.
The foregoing experiments, though showing that the larvae
are responsive to all the lights and perhaps least effected by the
red, do not show the relative potency of the colored lights. To
determine this relation, the four monochromatic lights in all
their possible combinations were used in testing the responses
of the larvae. Typical examples of the records selected to repre-
sent the average of the course taken by the larvae under the
respective pairs of lights, are shown in figures 9 to 20. Where
red was balanced against blue, green or yellow it is evident
from an examination of the records (figs. 9 to 11), that red
is least effective in orienting the larvae. A larva started in the
red is completely reversed in its direction of crawling when
opposed by any one of the other lights. On the other hand a
path of the same larva started in the blue, green or yellow is
only slightly if at all altered in direction by the red light. Similar
results are shown in the records of the reactions under the same
pairs but with the lights reversed in direction respectively (figs.
20, 17, and 14).
In pairs where yellow is opposed to the other lights, the red
is less effective than the yellow in orienting the animal, but the
blue and green are more so (figs. 12 to 14). When the larva is
started in the yellow it turns sharply and reverses its course at
the point where the blue or green is turned on (figs. 12, 13).
In these tests where the larvae were started in blue or green
several trial movements were frequently made by a larva when
it received the stimulus from the yellow light. The subsequent
direction of its course in this case is not reversed but it may be
at an angle to the direction of the rays. This experiment shows
that yellow when compared with blue and green has an appre-
ciable effect on the larvae although it is much less potent than
either of these lights. A larva oriented in the red is reversed
in its course by the yellow and one started in the yellow is to
REACTIONS OF ARTHROPODS TO LIGHTS 481
all outward appearances unaffected by the red (fig. 14). These
reactions of the larvae to pairs including yellow light demon-
strate that yellow is more stimulating to the larvae than red
but less than either blue or green.
A study of the records made by the larvae when under the
influence of green in opposition to blue, yellow or red (figs. 15 to
Figs.9to14 Paths traced by Calliphora larvae, in dilute solutions of methyl-
ene blue, in response to balanced pairs of monochromatic lights of the following
pairs:
Fig. 9 Red-blue Fig. 12 Yellow-blue
Fig. 10 Red-green Fig. 13 Yellow-green
Fig. 11 Red-yellow Fig. 14 Yellow-red
17) show that green not excepting blue is the most powerful as
an agent for orientation. As might be predicted from the pre-
vious experiments the course of a larva started in the yellow or
red is reversed when the green is turned on. The course of the
same larva oriented in the green is not effected by the red and
482 ALFRED O. GROSS
only slightly by the yellow (figs. 16 and 17). When green and
blue are opposed, the course of a larva oriented by two lights is
at an angle to the direction of the rays as shown in figure 15.
If the course of the larva under the stimulus of these two lights
had been at right angles to the rays the potencies of green and
blue, so far as this experiment is concerned, would have been
Figs 15 to 20 Paths traced by Calliphora larvae, in dilute solutions of methyl-
ene blue, in response to balanced pairs of monochromatic lights of the following
pairs:
Fig. 15 Green-blue Fig. 18 Blue-green
Fig. 16 Green-yellow Fig. 19 Blue-yellow
Fig. 17 Green-red Fig. 20 Blue-red
shown to be equal. The larva, however, regardless of the light
in which it was first started was oriented toward the blue in a
direction at an angle with the direction of the rays.
In a similar way figures 18 to 20 show that blue is more effec-
tive than yellow or red, but less than green. A larva started
in the yellow or red is completely reversed in its course by the
REACTIONS OF ARTHROPODS TO LIGHTS 483
blue. In the blue and green pair of lights, the results are similar
to those of the green and blue already described. A larva ori-
ented in the blue is turned back at an angle by the green and
one started in the green is driven into the blue. The course
D E F (fig. 18) plotted by a larva in a blue-green pair is inter-
esting since the larva was oriented four different times and each
time it took practically the same angle with respect to the direc-
tion of the rays. It would be impracticable to reproduce repeti-
tions of the examples used in the foregoing account of the reactions
of blow-fly larvae. But since it is desirable to have more than
a single average record on which to base an opinion of a result,
the records of the respective pairs have been simplified and
condensed into diagrams. The construction and use of these
diagrams can be explained best by taking an actual example of
a reaction record, as for instance, red opposed to blue. In the
example (fig. 21, A) a larva was placed in a drop of methylene
blue at a and allowed to creep away from the red light to b.
When at b the blue light was suddenly turned on and as soon
as the larva received the effect of the greater stimulus it took
the new direction be. In order to classify the records the approxi-
mate angle the larva took with reference to the direction of the
rays was determined in the following manner. From b as a
center, where the second light was thrown on, an are with a
radius of an arbitrary length of 8 cm. was drawn intersecting
the path, taken by the larva, at x A line was drawn from 6
parallel to the direction of the rays intersecting the are at l.
The line b m was drawn perpendicular to the line 6 1 thus cutting
off an are of 90°. This are which is now definitely oriented with
respect to the direction of the rays was divided into four parts
for convenience in combining the records. <A circle of convenient
size (fig. 21 C) was then divided into four quadrants, each of
which, as the quadrant in figure 21 A, was further divided into
four parts and oriented with respect to direction of the rays.
The paths traced by the larvae may now be classified on this
simple diagram by indicating the position of x on the arc of the
corresponding sector of the circle. Thus the record of the path
a 6 e of the larva in figure 21 A is indicated on figure 21 C by 2’.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 4
484 ALFRED O. GROSS
In the converse of the above experiment, blue opposed to red,
the larva was oriented at c by the blue light. When at d the
red light was turned on but the blue being a much greater stimu-
lus than the red, the larva continued apparently on an unin-
terrupted course to f intersecting the are of the circle at y. In
a similar way this course may be approximately recorded on the
m
Fig. 21 A typical record of a larva in response to the red-blue pair
of lights. A and B are the paths followed by the larva started from the left
side and from the right side respectively. The arrows indicate the position of
the larva when the second light was turned on. The dots show the position of
the larva at ten-second intervals. The quadrants in A and B and figure 21 C
are used to illustrate the method of constructing the diagrams of the reaction
records of Calliphora larvae shown in figures 22 to 33 inclusive.
diagram by y’ as any number of additional such records may
be. A complete diagram of the reaction records of the larvae
to the balanced red and blue lights is shown in figure 22. Here
out of 14 trials all of the animals were driven by the blue into
the red as in the typical record, figure 21 A. The records of
a second series of experiments in which the colors were inter-
3| 32 33
Figs. 22 to 33 Diagrams illustrating the approximate courses taken by Calil-
phora larvae in response to the various pairs of monochromatic lights:
Fig. 22 Red-blue Fig. 26 Red-yellow Fig. 30 Yellow-green
Fig. 23 Blue-red Fig. 27 Yellow-red Fig. 31 Green-yellow
Fig. 24 Red-green Fig. 28 Yellow-blue Fig. 32 Green-blue
Fig. 25 Green-red Fig. 29. Blue-yellow Fig. 33 Blue-green
486 ALFRED O. GROSS
changed are shown in figure 23 and are similar to those shown
in figure 21, with the exception that three larvae which were
started in the red were not turned in the reverse direction by the
blue. They were turned at an angle when exposed to the blue
but did not turn completely before creeping off the paper. When
such larvae are allowed to continue on another sheet, they usu-
ally orient in the course of time in a direction away from the
blue light. Such slight irregularities, which may occur in any
series, are probably due to corresponding differences in the
responsiveness of the organisms. For this reason experiments
in which the larva is started first in one light and then in the
other are highly important in serving as an effective check
throughout these records. In each diagram (figs. 22-33), though
the records are plotted all on one ‘side, they represent records
one-half of which are from larvae started on the right side and
the other half on the left side in the respective colors of each
pair. A greater number of observations were made for the green
and blue pairs than for any other, since the relative differences
of the stimuli produced by these lights, is apparently less than
that in the lights of the other pairs.
These combined records of the blow-fly larvae reactions sub-
stantiate the results previously shown in the reproduction of
the paths plotted by the larvae, namely, green is the most effec-
tive and the red the least so, of the four monochromatic lights,
in orienting the larvae. The blue and the yellow are intermediate
between these extremes but the blue light is much stronger in
its effect than the yellow. The effectiveness of the four colors,
therefore, is in the order, beginning with the strongest; green,
blue, yellow, red.
B. Calliphora erythrocephala Meigen (adult)
1. Material. The adult blow-flies used in these experiments
were reared in the laboratory from the cultures of larvae used
in the foregoing experiments. The flies were easily kept alive
and in good condition by feeding them on sugared water. They
were placed in a large screened cage provided with a device for
REACTIONS OF ARTHROPODS TO LIGHTS 487
removing them to small glass jars as required. In order to
dark-adapt the flies, the jars containing two to five individuals
each were placed in the dark at least an hour before using them
in the experiments.
2. Methods. The essential part of the apparatus used in these
tests was an elongated glass cylinder or tube 6 cm. in diameter
and 40 em. long supported on a base 14 cm. high. The base
was a black box, 16 cm. wide and 20 em. long, open on one side
to allow the experimenter to place or remove the glass jars con-
taining the flies. This apparatus was placed midway between
the generators in such a way that the center of the cylinder
was directly in the center of the field of light and its axis parallel
to the direction of the rays. A funnel opened into the cylinder
from the dark chamber of the box below. Since the flies are
strongly negatively geotropic, they readily crawl upwards when
freed in the funnel. As they leave the narrow opening leading
into the illuminated region they are oriented to one side or the
other depending on which light exerts the greater stimulus.
Mechanical counters were used for recording the number of flies
as they left either end of the apparatus. The flies were not
used a second time but a fresh lot of dark-adapted individuals
was taken for each new set of records.
3. Results. The flies were tested first with single lights to
determine whether they were responsive to each of the four
colors of an intensity used in these experiments. In all of these
tests the direction of light rays was interchanged for each set
of individuals, but for a matter of convenience all the records
of any one color are combined. When the single lights are thus
compared with darkness the flies were found to be distributed
as shown in table 1. The records of the experiments shown
in table 1 demonstrate the strong positive phototropism of the
blow-fly to each of the four monochromatic lights. A greater
number of flies were negative or indifferent to the red or yellow
than there were to the blue or green, a result which indicates
that the more refrangible rays are more effective than those at
the opposite end of the spectrum. ‘This relative efficiency of
488 ALFRED O. GROSS
TABLE 1
Reactions of Calliphora adult to single monochromatic lights
NUMBER OF REACTIONS
PERCENTAGE OF
coLoxs aa ae > wo POSITIVE REACTIONS
Positive Negative
Dives tere 2 4- Peee 113 1 99
green..... A RN te cs 3 : 110 5 95
yellows so" s. Sscce : 150 9 94
FOG se see! ees eee : 106 26 80
the colors is corroborated by the results of further experiments,
to be described later, on the adult blow-flies with balanced lights
of different colors.
Flies were next tested in the cylinder illuminated by paired
lights of the same color and same intensity. About 100 flies
used in each of four such pairs were found to be distributed to
the two sides as shown in table 2. Theoretically the distribution
ratio of the responses of the flies to balanced lights of the same
colors should be 50%: 50%, but when comparatively small
numbers are used slight deviations, as in the above results, may
be expected. The percentages, however, even with these limited
number of trials, vary only from one to three units from the
expected ratio and show the degree of reliability one may place
on the records of the reactions of the adult flies to balanced lights.
With each new set of trials with any pair of colors the direction
of the rays was reversed to eliminate from the results any errors
which might arise from defects in the apparatus.
TABLE 2
Reactions of Calliphora adult to pairs of monochromatic lights of the same quality
and of equal intensity
COLORS NUMBER OF REACTIONS
= Sa ayes gus PERCENTAGE OF
REACTIONS TO THE LEFT
Left Right Left Right
blue.... blue 47 53 47
green........ green 55 49 51
yellow......| yellow 65 61 61
atts ae red 62
69 48
REACTIONS OF ARTHROPODS TO LIGHTS 489
The records of the responses when the flies were subjected to
the respective combinations of different colors are given in table
3. For convenience in comparing the relative potencies of the
colors, as expressed in per cents of reactions, table 4 is intro-
duced, in which each color is readily compared with every other
color and with darkness.
In table 4 the colors heading the columns are compared with
those facing the lines. For example, if we wish to know what
percentage of the flies went to the green in the green-yellow pair
of lights we look for the number in the column headed green
and in the line opposite the yellow. The number 79 indicates
that in the experiment with this pair of lights 79 per cent of
the individuals tested went to the green away from the yellow.
TABLE 3
Reactions of Calliphora adult to pairs of monochromatic lights of equal intensity
PAIRS OF MONOCHROMATIC NUMBERS OF FLIES TOTAL NUMBERS OF =
LIGHTS GOING TO EACH ee FLIES TO EACH LIGHT Pia ae
3 More Less REFRACTIVE
mene _ | Right | Left refractive | refractive LIGHT
blue reen 92 50
ee sabe | Sa 171 100 63
green blue 50 79
blue | yellow | 55 22
hae : : 106 40 72
yellow blue 18 51
blue | red 33 6
red blue 8 SO — 89
reen yellow ri) 35 &
3 } 3 pe eee 167 43 79
yellow green 8 | 97
reen red 93 | 18 vs
. 52 ee 167 23 87
red | green 5 74
ellow | red 66.) 4 22 LRT
_ | | 129 35 79
red yellow 13a 63
TABLE 4
Combtned records of Calliphora adult reactions
COLORS BLUE GREEN | YELLOW RED DARKNESS
Ble eae Az 37 petal ear 1
RREETTG Sh ost 63 51 | 21 13 4
EMG Woeliied 26 We Ad cc lgeery eeaes | 21 6
27) Ce ey 89 87 79 AS oe 20
@arkness....:.<.< 99 96 94 | 80
490 ALFRED O. GROSS
In the reverse of this pair, yellow-green, the number 21 in the
column headed by yellow and in the line faced by green indicates
that 21 per cent of the flies used went to the yellow. In a like
manner the results obtained with any pair of lights may be
readily ascertained.
From the above series of records it is at once apparent that
the adult blow-flies are responsive to all the colors, and, of the
lights used in these experiments, are affected most by those
containing the more refrangible rays. The effectiveness of the
colors in stimulating Calliphora, therefore, are in order, begin-
ning with the strongest: blue, green, yellow, red.
C. Drosophila ampelophila Loew (adult)
The pomace or little fruit fly, Drosophila ampelophila Loew,
is a very common insect for experimental work, not on light
alone, but in a great diversity of lines of investigation, because
of the ease with which vast numbers can be reared and handled
in the laboratory. Carpenter (’05) has demonstrated the strong
positive reaction of Drosophila in its response to white light,
but no one, as far as I am able to discover, has made careful
investigations of the reactions of these insects to monochromatic
light. The tests made with Drosophila in the following investi-
gations were carried out in practically the same way as those
made with the adult blow-fly. Since the pomace fly is so closely
related to the adult blow-fly and has the same type of visual
mechanism as that insect, one would naturally expect in experi-
ments with it to obtain results similar to those from the blow-
fly.
1. Material. A continuous culture of Drosophila derived from
an original stock secured during the summer was maintained
in the laboratory throughout the winter. They were reared in
large glass jars which contained a supply of decaying bananas
on which the insects fed and deposited their eggs. By inverting
a large glass funnel over such a culture and directing it upwards
and towards a strong light the flies can, because of their strong
reactions, be easily conducted through the funnel into small jars.
REACTIONS OF ARTHROPODS TO LIGHTS 491
About 15 to 20 individuals were thus transferred to each of
about 25 small 4-ounce wide-mouth jars for each of the series
of tests. The flies were not reared in the dark room but were
always dark-adapted before they were used in any of the light
experiments and, as in the case of the adult blow-flies, they were
never used a second time immediately after having been exposed
to the light of the apparatus.
2. Methods. The apparatus used for the Drosophila was essen-
tially the same in principle as that used for the blow-flies and
differed only in some minor details. The tube leading from the
funnel in this apparatus was smaller and its opening into the
illuminated cylinder was partially obstructed by two pieces of
cork which prevented the flies from going upwards too rapidly.
There was also a device to enable the experimenter to close
completely the aperture when so desired.
TABLE 5
Reactions of Drosophila to single monochromatic lights
NUMBER OF REACTIONS |
| PERCENTAGE OF
porone | | POSITIVE REACTIONS
Positive Negative
|
|
[DISTR SS Selb pee Corea a 191 | 10 95
Pe Rae Ee ORE 183 42 81
Walllon7iis SAG ee ee 160 55 74
TRELG |, oS ck Spee A et ele ree 152 60 71
3. Results. To determine whether all the lights used were
effective in stimulating the animals the flies were permitted to
enter the cylinder illuminated by a single monochromatic light
from one end only. In each of these tests about 200 individuals
were used, which were found to react as represented in table 5.
This experiment demonstrates the strong directive effect that
each-of the colored lights used in these tests have on Drosophila,
and gives further evidence of the strong positive phototropism
of these insects. A greater percentage of the individuals were
negative when tested with the yellow or red lights than when
the blue and green were used. This result suggests the fact
subsequently established that the lights having the more refran-
gible rays exert the greatest directive stimulus on the pomace flies.
492 ALFRED O. GROSS
The flies were then tested with balanced lights of the same
colors to determine to what extent the percentages of the responses
can be depended on to express the relative potencies pf the lights
used. The results of this test are shown in table 6.
The extreme amount of deviation in the percentages in table
6, which is only 4 units less than the expected theoretical ratio
of 50 % : 50 % is found in the red.
TABLE 6
Reactions of Drosophila to pairs of monochromatic lights of the same quality and
of equal intensity
COLORS NUMBER OF REACTIONS PERCENTAGE OF
~ REACTIONS TO THE
Left Right Left Right eo
plaer 0 <) plue 82 86 49
PIEPN ees. green 41 40 50
yellow...... yellow 46 43 52
Reis ess red 53 62 46
In the following experiments on the responses of Drosophila
to balanced lights of different colors not less than 300 and some-
times as many as 500 individuals were used in each combination
of colors. The large numbers used tend to reduce the size of
the error of the results and to give a more correct representation
of the relative efficiencies of the four colored lights. Table 7
is a summary of the reactions of the pomace flies to balanced
pairs of monochromatic lights of different spectral qualities.
A number of tests were made in which the flies were caught
in the retaining cylinder and run through a second series of
trials. This was done to determine whether or not flies exhibit-
ing a negative reaction to the individual colors or to the more
refrangible color of any pair of lights are permanently negative
or indifferent. The factor of mechanical stimulation of the flies
was eliminated in these experiments by allowing the flies to rest
and to become dark-adapted before using them in a second test.
The results of five such tests are shown in table 8.
The results shown in tables 5, 6 and 7 are combined and
condensed into the simple table 9.
REACTIONS OF ARTHROPODS TO LIGHTS 493
TABLE 7
Reactions of Drosophila to pairs of monochromatic lights of equal intensity
PAIRS OF MONOCHROMATIC
NUMBERS OF FLIES | TOTAL NUMBERS OF |
LIGHTS GOING TO EACH LIGHT | FLIES TO EACH LIGHT | TEER,
‘ if 3 REFRACTIVE
Left Right Left ESTE | con oe ee LIGHT
sats es ee) |
blue green 2597 79 |
Ze 2 7
green blue AQT |) 92438 a | = w
blue yellow 162 | 35 a |
yellow blue 4 106 aia Mia | =a
blue red 206 17) |
9
red blue 14 217 a | a ce
green | yellow 124 33 | ot =
yellow _—_ green 69 199 oe | eg ae
| |
green red 5 UG 42 | = Fs
391 ite
red _ green 73 239 | 5 if
yellow | red 174 og) | 5 S
: 3 174
red | yellow 105 192 . | eB ; ee
TABLE 8
Records of tests of the re-distribution of Drosophila adults
= > Gi Sa Sar | q |
NUMBERS OF FLIES Nee OF FLIES | DISTRIBUTION OF FLIES
COLORS sie Soa CAUGHT ee dees | IN SECOND TEST
| | |
left | Right Teft | Right Left | Right | Left | Right
| i 9 ; 4.
blue | red Gy | 9 | | 6 5 | 1
green red 42 | 13 | 8 ieee) | 3
red yellow 2A ile 34 he | Mec: Sa me
red | green 20 a | | 7 1 | 6
TABLE 9
Combined records of Drosophila reactions
COLORS BLUE GREEN YELLOW RED DARKNESS
blue 49 21 13 7 5
green 79 50 24 23 19
yellow 87 76 52 32 26
REIN ee Rene kona 3 93 ie 68 46 2
darkness-o--a-.0:. 95 81 74 71
494 ALFRED O. GROSS
The colors indicated on the lines are compared with those of
the columns as was done in the table of percentages for the adult
blow-fly. For example in the blue-red pair the number 93 in
the blue column opposite the red denotes that in this test 93
per cent of the flies went to the blue and away from the red. In
a similar manner the result of the experiment with any pair of
lights may be readily determined.
100
90
80
70
60
50
40
30
20
10
B G Dg R
Fig. 34 Curves representing the percentages of the responses of the pomace
flies to the four monochromatic lights. Percentages of responses of animals to
the respective colors as ordinates and wave lengths as abscissae. The wave
lengths of the middle band of each light are indicated by points marked on the
axis of the abscissa.
The results presented on the foregoing pages may be expressed
graphically in a series of curves in which each color is compared
with each of the other four colors (fig. 34).
The results of these experiments demonstrate that of the four
colors the ones containing the more refrangible rays exert the
REACTIONS OF ARTHROPODS TO LIGHTS 495
greatest directive stimulus on Drosophila. The effectiveness of
the colors beginning with the strongest is, therefore, in the fol-
lowing order: blue, green, yellow, red.
D. Zeuzera pyrina Linné (larva)
The lepidopterous insects are, as far as I know, unlike the
dipterons in that there is no reversal of their phototropism during
development. Both the caterpillars and the adults of the species
of moths and butterflies!‘ known to be responsive to directive
light are distinctly positively phototrophice. .
1. Material. The larva of the European wood leopard-moth,
Zeuzera pyrina Linné, used in these experiments, was accidently
introduced into this country from Europe, and has since proven
to be a serious pest to our shade trees and shubbbery. This
species of moth has not, to my knowledge, ever been the sub-
ject of light experiments, but Chapman (711, p. 15) states:
“The larvae when they first hatch exhibit a marked positive
phototaxis for they make their way to the tips of small branches
to the axis of leaves or to nodes and buds near the tips, and at
once bore into the woody tissue.”’
This suggests an interesting correlation of the positive photo-
tropism of these insects with their habits during the early stage
of their life history.
Unlike the blow-fly larvae the leopard-moth larvae are most
responsive during the early part of their free life. They become
more and more sluggish in their reactions to light as they approach
the pupal stage which, according to Chapman, is not attained
until the second year.
The larvae were abundant in the twigs and branches of many
of the trees and shrubs, especially the lilacs in the vicinity of
the Museum of Comparative Zoédlogy, Cambridge. The lilacs
provided an abundant and convenient source of material. A
considerable number of the larvae used in these experiments
were secured by Mr. Chapman from the trees and shrubbery
of Boston Common. Since the larvae are somewhat troublesome
1The reactions of Venessa antiopa Linn. is an apparent exception. These
butterflies when alighted turn away from bright sunlight. Parker (’03).
496 ALFRED O. GROSS
to keep in the laboratory in a good vigorous condition new
specimens were secured each time it was desired to make a
series of tests. Parts of branches were taken into the dark room
from which the contained larvae were not removed until they
were required in the experiments. These investigations were
made during October and November, a time too late for speci-
mens of newly hatched larvae but all the individuals of the first
winter exhibited a very strong positive phototropism.
2. Methods. The apparatus used for these experiments on the
leopard-moth larvae was essentially the same as that employed
for the larvae of the blow-fly. The paths followed by the larvae
were, however, traced with a pencil, since methylene blue could
not be used to an advantage because of the waxy nature of the
larval integument. The pencil tracings were made by following
the course of the animal in such a way as not to interfere with
its movements or with the rays of light impinging on them.
The larvae when cold did not respond very readily to the light,
hence it was frequently found desirable to warm the slate, by
means of an electric heater placed beneath the apparatus, to a
uniform optimum temperature.
3. Results. The accurate responsiveness of the larvae to the
different colors is demonstrated best in experiments with single
colors in which the direction of the light is suddenly reversed.
A typical record of the result of such an experiment, in which
a larva was tested with blue light, is illustrated by figure 35.
On this test the larva was placed at A, in the position indicated
by the arrow, in the midst of a beam of light coming from the
left. The larva immediately oriented and was permitted to
crawl towards the light undisturbed, to the position at B. At
this point the direction of the light was suddenly changed, where-
upon the larva immediately reversed its course in a direction
towards the new source of light at the right. The similar results
of many such tests made with each of the colors demonstrate
the positive phototropism and the marked responsiveness of the
leopard-moth larvae to lights of an intensity and quality used
in these experiments. The relative efficiency of the four lights
was obviously not determinable from the results of these simple
REACTIONS OF ARTHROPODS TO LIGHTS 497
preliminary tests with single colored lights. . However, when the
larvae were tested with red, they seemed to orient less accurately
than when stimulated by any one of the other colors, a result
indicating that the less refrangible rays are least effective.
In the experiments with balanced pairs of monochromatic
lights the larvae were started in the middle of the slate or sheet
of paper and in the center of the field of light. They were placed
with their axes at right angles to the rays of light. This posi-
tion gave the anterior end of the animal, which, presumably,
is the only part sensitive to light, equal opportunity to stimula-
tion by the lights of equal intensities coming from opposite
sources.
When the larvae were exposed to paired lights of the same
colors, they oriented, not with their axes perpendicular to the
light rays, as was the case with the blow-fly larvae, but toward
one light or the other approximately parallel to the direction of
the rays (figs. 36 and 37). When paired lights of different colors
were used the larvae turned toward the light containing the
more refrangible rays (figs. 38-43).
In addition to the method used in the experiments just de-
scribed, the larvae were started at one edge of the slate and
allowed to orient definitely to the rays of one light before turn-
ing on the opposing light. Figures 44 and 45 are records of the
paths taken by a larva when thus tested with the green-yellow
pair of lights. In the records reproduced in figure 44, the larva
was started at A and was allowed to crawl towards the yellow
light to the position at B, when the green light from the oppo-
site was turned on. When thus exposed to the influence of
two lights the larva did not, in this case, reverse its course imme-
diately but continued to O before orienting to the green. The
distance crawled by the larvae after the green light was turned
on varied considerably in the several tests with the same and
different larvae. In some cases the response was immediate and
in a very few tests the larvae crawled off the sheet of paper
towards the yellow light without exhibiting the least evidence
of the effect of the green. The visual organs, the larval eyes
or ommata, on the anterior end of the larvae are not in a favor-
35
36
37
a a
ss
39
Figs. 35 to 45 Paths followed by Zeuzera larvae in response to monochro-
matic lights of equal intensity, as described on pages 496 to 500 inclusive. The
approximate length of the larvae as well as their position and direction of travel
is indicated by the arrows. Action of lights successive in figures 35, 44 A and 45
A;in all other cases simultaneous.
Fig. 35 A Blue alone at left
Fig. 35 B Blue alone at right
Fig. 36 Blue-blue
Fig. 37 Red-red
Fig. 38 Blue-green
Fig. 39 Green-blue
498
Fig
Fig.
Fig
Fig.
. 40 Green-yellow
41
. 42
43
REACTIONS OF ARTHROPODS TO LIGHTS 499
ee
ats
_ ae \
—S
ae
3 ae
B
ee a eg ee
45
Yellow-green
Yellow-red
Red-yellow
Fig. 44A Yellow alone at right
Fig. 44B Green-yellow
Fig. 45 A Green alone at left
Fig. 45 B Green-yellow
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 4
500 ALFRED O. GROSS
able place to be stimulated by any color when the body is oriented
away from its source, hence in the above experiment, the varia-
tion in time of response probably depends, to a certain extent,
on the accidental exposure of the ommata to the green light.
In the record shown in figure 45 the larva was started from
the right at A in the green light and was allowed to crawl to
B before the yellow light from the opposite side was turned on.
In no ease of the many tests made was the larva reversed in
its course by the yellow when the latter was opposed with green.
In similar experiments made with each of the remaining pairs
of lights the larvae were shown to be most responsive to the
colors having the more refrangible rays. The effectiveness of
the colors in stimulating the leopard-moth larvae corresponds
to their sequence in the spectrum, namely, beginning with the
strongest: blue, green, yellow, red. |
E. Feltia subgothica Haworth (adult)
1. Material. Adult specimens of the leopard-moth could not
be obtained at the time of these experiments, but I was fortunate
in securing numbers of another species of Lepidoptera (Feltia
subgothica Haworth). These moths were very common around
the are-lights at night during the month of September. Through
the kindness of the Bussey Institution at Forest Hills an excel-
lent are-light insect trap was placed at my disposal by means
of which the necessary material for the following experiments
was secured. The moths thus collected were placed in ventilated
pasteboard boxes, which were shielded from the light until the
specimens were required in the experiments. In addition to this
precaution, since it was desirable to have each individual dark-
adapted, none of the moths were used in two succeeding experi-
ments.
2. Methods. The chief feature of the apparatus was an elon-
gated glass chamber, 10 x 10 x 45 em., which had an opening at
the middle of the side nearest the experimenter through which
the moths could be liberated from the paper boxes. This appa-
ratus, as that of the previous experiments, was placed in the
middle of the field of light midway between the two generators.
REACTIONS OF ARTHROPODS TO LIGHTS 501
A moth allowed to enter the chamber was free to go to one
side or the other, but since it is positively phototrophic it would
be expected, other factors being equal, to go to the light exerting
the greater stimulus.
3. Results. The moths reacted to the various pairs of lights
as shown: in table 10. The numbers represented in table 10,
though too small to warrant a comparison of the percentages
of the responses to the various colors, nevertheless show that
of the four colors used those of the more refrangible rays are
more effective in stimulating and orienting the adult Feltia sub-
gothica. The effectiveness of the four colors, therefore, is in
the order beginning with the strongest: blue, green, yellow, red.
F. Periplaneta americana Linné (adult)
The adult cockroach, Periplaneta americana, has, so far as
I know, never been used in experiments with spectral lights.
Graber (’84) has, however, made a series of investigations with
a closely allied form, Blatta germanica Linné, with colored lights
produced by means of screens.
TABLE 10
Reactions of Feltia to pairs of monochromatic lights of equal intensity __
PAIRS OF MONOCHROMATIC LIGHTS Cro tee ce aaa Raines Sta oo
| | : Moo 1 Ss
a Right | Left Right refractive raftictive
blue blue 20 18
green green a )
yellow yellow | 6 9
red red 10 13
blue green 12 6 23
green blue 1 11 sy :
blue yellow reals 3
yellow Blues) i eceoad 18 eS ey
blue red 11 3
red blue 6 23 at
green yellow 12 | 0 29 5
yellow green 5 10
green red 14 3 33 =
red green 4 19 :
yellow red | 16 | 9 :
red yellow | , 17 oe My
502 ALFRED O. GROSS
1. Material. The material used in these experiments was
obtained from the dark basement store rooms of a sugar refinery.
No attempt was made to rear the cockroaches in the laboratory,
but a fresh lot was secured each time it was desired to make
a series of tests. They were always dark-adapted before they
were used in the experiments.
2. Methods. <A plain piece of slate 28 em. square was used
in exposing the cockroaches and thus the strong thigmotrophic
response of these insects called forth by sharp corners or con-
cave surfaces was avoided. ‘The slate supported by a small table
16 cm. in height was placed inside a glass box 40 cm. square by
15 em. in height. The purpose of the box, which in no way
interfered with the rays of light, was to entrap the cockroaches
when they ran off the edge of the slate. When a large number
of specimens had thus accumulated they were re-collected but
were not used in a second series of tests until they had become
dark-adapted. The entire apparatus including the outside of the
glass box was painted dead black to reduce to a minimum the
reflection of any light.
A cockroach to be tested was placed at the middle of the
surface of the slate, with its axis perpendicular to the direction
of the rays. In this position’ there was an equal chance of its
going to one side or the other unless influenced by the light.
The direction of movement of the cockroach was not always
directly towards or away from the source of light, but was fre-
quently to one side of the illuminated field to the darkness.
Some of the erratic movements of these highly excitable and
nervous insects may be explained in part by what seems to be
their intense fear while in the light. In such cases, where the
orientation was not direct, the response was considered positive
or negative according as the cockroach left the slate on the side
toward or away from the source of light: If it left the slate on
the side of the median line away from the light, the reaction
was considered negative no matter if the course taken by the
insect was in the direction of the rays or at an angle with them.
If the response was in any direction toward the light it was
counted as positive.
REACTIONS OF ARTHROPODS TO LIGHTS 503
3. Results. Single monochromatic lights were used first to
determine the relative effectiveness of each light in stimulating
the cockroach. The direction of the lights was reversed with
each ten trials, but for a matter of convenience the records are
combined.
From the records of the experiments shown in table 11 it may
be inferred that the cockroach is positive to blue, negative to
green and yellow and indifferent to red, results which are borne
out by experiments with balanced lights of different colors to
be described later.
When paired lights of the same colors were used, there was
practically an equal distribution of responses to the two sides.
The records in table 12 while of no value in determining the
relative potency of the colors, do show the degree of reliability
that can be placed on tests of cockroaches with balanced pairs
of lights.
TABLE 11
Reactions of Periplaneta to single monochromatic lights
NUMBER OF REACTIONS |
a _ PERCENTAGE OF
OORT: NEGATIVE REACTIONS
Negative Positive
bitieh eee et eee 30 ie 29
iREO Ma ogea Borel OEE ee 89 AT 65
wre lllowienrtcc eevee: tlascn: 108 65 63
FOG ce Wy Bete eae 55 60 48
TABLE 12
Reactions of Periplaneta to pairs of monochromatic lights of the same quality and
of equal intensity
COLORS NUMBER OF REACTIONS | anne
= OF REACTIONS TO
Left Right Left Right TEE TEET
plue. 2 aos. blue 62 54 53
green....... green 68 61 53
yellow...... yellow 57 59 49
TECH 0c a eG 57 58 50
504 ALFRED O. GROSS
The results of the experiments with balanced monochromatic
lights of different colors are represented in table 13.
In each record of table 13 where blue is compared with any
one of the other colors or with darkness the majority of the
cockroaches were oriented in a direction towards the blue. There
is, however, no consistent gradation in the percentages of re-
sponses that corresponds to the spectral position of the color
opposed to the blue. —
‘ In the contrasts of green or yellow with other colors the major-
ity of the responses of the cockroach are negative to the green
and to the yellow. The percentages of the responses in these
cases are shown for greater convenience in comparison in table
14.
TABLE 13
Reactions of Periplaneta to pairs of monochromatic lights of equal
intensity
parma or uononmmoxinn |, Symmmmnarotee,” | saumceance || ae
EACH LIGHT EACH LIGHT | COCKROACHES TO
; | More | Less Saree EERE
Left Right Left Right refractive | refractive |
A (came } |
blue green 86 AS
| |
green blue 44 | 79 166 ee | a
blue yellow 88 18 a
yellow | blue | 61 | 103 me e -
blue red i 72 a as 2
red blue Aa 95 iz fe
green yellow 52 | 56
yellow | green | 56 | | LOO ieee ba
green red leis: Saab | 76 |
red green fr | 4 of ag so
yellow red eee ert |
red yellow | 84 © | 44 ig | ye | has
TABLE 14
Percentages of reactions of Periplaneta to each of the four colors and to darkness
when compared with green and with dase
2 a -
COLORS BLUE | GREEN YELLOW | | DARKNESS
Lae - = Le) |
|
(green............ 66 53 | 50 64 65
WOU isos 5. % » 71 63
REACTIONS OF ARTHROPODS TO LIGHTS 505
In the sets of pairs in which green and yellow are compared
with. the other colors the percentages of negative responses in
each pair are practically equal. Thus the theoretical result when
yellow is balanced with yellow or green with green is 50 % : 50 %,
while the actual result is 53 % : 47 % and 49 % : 51 % for the
green and yellow combinations respectively.
The percentages of responses when green-yellow and yellow-
green are opposed are 49% : 51% and 50% : 50% respec-
tively. These results indicate that yellow and green are practi-
cally equal in potency as far as their effectiveness in orienting
the cockroach is concerned.
TABLE 15
Percentages of reactions of Periplaneta to each of the four colors when compared
with red and with darkness
COLORS | BLUE | GREEN | YELLOW RED
=2) re | 69 36 | 34 50
Garkness::...:.... 71 35 37 52
The reactions of the cockroach to the pairs of colors in which
blue is balanced with green and yellow, the percentage’ of the
responses to the blue is not greater than when blue is used alone.
This result is somewhat inconsistent in view of the fact that the
cockroach is negative to the green and yellow and positive to
the blue. In using such pairs of lights one would naturally
expect a larger percentage of the cockroaches to go to the blue
than in the case when blue is used alone.
In table 15 are shown the percentages of responses for the
pairs red or darkness with each of the colors. The results exhib-
ited in this table demonstrate that red of an intensity and quality
used in these experiments, has no more effect than darkness in
stimulating the cockroach.
The reactions of the cockroach to light are unique in that
they are positive to the blue, negative to the green and yellow
and indifferent to red.
506 ALFRED O. GROSS
5. DISCUSSION
The results of these experiments with lights of measured inten-
sity demonstrate conclusively that the effectiveness of the dif-
ferent colors of the spectrum does not correspond to the relative
intensity of luminosity of the lights, as Merejkowsky (’81) has
stated, but to their specific quality. Merejkowsky’s method of
equalizing the intensity of the lights, by judging the relative
luminosity with his own eye, is very inaccurate, because the
different colors of the spectrum have an effect on the eye which
is not proportional to their energy content. Furthermore, the
maximum brightness of the spectrum differs with different degrees
of illumination as is well known from the Purkinje phenomenon.
In the bright spectrum the region of greatest luminosity lies in
the yellow, in a spectrum from a weaker source, it is in the green.
Merejkowsky’s results, therefore, cannot be considered seriously
as opposing the view, that the efficiency of the different colors
of the spectrum to stimulate the lower animals, is independent
of intensity.
The view of Hess (10) that the relative attractive power of
the different homogenous lights approaches or corresponds with
the brightness curve of the color-blind person, is not funda-
mentally different from the view held by Merejkowsky. Hess
has shown, contrary to Loeb’s hypothesis, that the reactions of
animals are not the same as those of plants in their response to
the spectrum, but he has not proven that the yellow and green
have the greatest stimulating efficiency when the factor of inten-
sity is eliminated. His experiments show that the yellow and
green are more efficient than the orange and red of the spectrum,
since the latter contain a much greater amount of radiant energy.
As the yellow-green rays contain much more energy than the
blue and violet, the seemingly greater effectiveness of the yellow-
green rays is probably due to the greater energy of this region
of the spectrum. ‘The results of the experiments by Bert (’68)
and by Lubbock (’81, ’83) upon the reactions of Daphnia to the
spectrum are similar to those obtained by Hess, but unfortu-
nately these investigators also ignored the factor of intensity.
REACTIONS OF ARTHROPODS TO LIGHTS 507
I do not mean to say that either Daphnia or many of the other
arthropods or the other animals investigated by Hess are not
more responsive to yellow-green rays than to those of shorter
wave length, but I do maintain that in none of these experiments
have the investigators proven that the yellow-green rays are
more potent than the blue rays, when the factor of intensity is
eliminated.
The results of the present investigations have shown con-
clusively that when lights of equal intensity are used, the adults
of Calliphora, Drosophila and Feltia and the larvae of Zeuzera
are more responsive to blue than to green or yellow.
The reactions of the animals named in the preceding paragraph
agree with the statements of Loeb (’05) and of Davenport (’97)
that the more refrangible rays of the spectrum are more effective
stimuli than the less refrangible rays. But the hypotheses of
Loeb and of Davenport are not in accord with the results of the
experiments upon the Calliphora larvae, which are most respon-
sive to the green rays of the spectrum. Loeb’s hypothesis was
based on the results of experiments which involved the use of
only two colors, namely, blue and red. In his experiments upon
larvae and other animals he did not compare the efficiency of
the blue end of the spectrum with the green and yellow rays.
Loeb therefore did not have sufficient evidence, as has been
pointed out by other authors, on which to base such a general
conclusion.
In order to have an unobjectionable basis for comparing the
results of the experiments upon Calliphora larvae with those on
adult insects under the blue-green combination of lights, the
adult Calliphora and also adult Drosophila were each tested
with the same lights and adjustments that were used in taking a
series of records similar to those shown in figures 15 and 18 and
immediately after the observations on the larvae. Out of 37
adult Calliphora used in five tests, 24 went to the blue and only
13 to the green; likewise in a second check experiment out of
42 Drosophila 33 went to the blue and only 9 to the green. These
tests demonstrate that the adult flies are more responsive to the
blue under conditions of illumination identical with those in
508 ALFRED O. GROSS
_which the Calliphora larvae are more affected by the green rays
and answer any possible objection that the results shown in the
records were caused by slight differences in the apparatus, inten-
sities, et cetera, employed for the larvae and for the adults.
This experiment proves that the relative stimulating efficiency
of the rays of different colors is not the same for all animals
nor for different ages of the same animal, and that the more
refractive rays are not always the most effective.
Mast (711) in developing the same idea, cites as evidence the
following investigations among others: Wilson (’91), who found
hydra to be more responsive to blue than to violet; his own work
(00) on amoeba, an animal which is also affected most by the
blue; the work of * Bert (’68), Lubbock (’81 a, ’83) and Yerkes
(99, ’00) on Daphnia, a crustacean, which aggregated in the
yellow-green region of the spectrum; and the investigations of
Engelmann (’82), who discovered that Bacterium photometricum
collected in the infra red of the spectrum. These investigations
do not prove that hydra and amoeba are affected most by the
blue rays, the Daphnia most by the yellow-green and Bacterium
photometricum most by the infra red independent of the inten-
sity factor. They do show, however, that the animals tested
respond differently in degree to the various rays of the spectrum,
since presumably the spectrum ‘used was similar in each case.
If the results of these investigations are reliable, then, in the
same spectrum amoeba and hydra are affected most by the blue,
Daphnia most by the green-yellow and Bacterium photometricum
most by the infra red. This work, therefore, in spite of the fact
that intensity was ignored, supports the view that the stimulating
efficiency of the rays of different wave lengths is not the same
for all animals.
The reactions of Calliphora larvae to colored lights are similar
to those exhibited by these larvae to white light, as demonstrated
by Herms (’11). The larvae, when opposing balanced colors of
the same wave length are used, crawl, after orientation, in a
direction approximately perpendicular to the direction of the
rays, as I have shown in the observations presented in this paper.
If lights of the same intensity but of different wave lengths
REACTIONS OF ARTHROPODS TO LIGHTS 509
are used the larvae proceed at an angle to the direction of the
rays, as they do when they are stimulated by opposed white
lights of unequal intensity; but if the efficiency of one of the
lights is much greater than that of the opposing light, as, for
example, blue against red, the larva is oriented directly towards
the latter by the dominating stimulus of the more potent color.
The larvae of Zeuzera pyrina, which are positively phototrophic,
never crawl at an angle to the direction of the rays, but, after
once having been oriented, proceed directly toward the source
of light. When paired lights of the same wave length are used
the larvae instead of crawling at right angles to the direction
of the rays, as the Calliphora larvae do, go directly to one or
the other of the two sources of light. When the lights are of
different wave lengths, the larvae are oriented by the more potent
color and proceed in a direction approximating that of the rays.
The reactions of Zeuzera agree with the general statement made
by Loeb (’05, p. 82) concerning the reactions of animals to white
light from two sources: “If there are two sources of light of
different intensities, the animal is oriented by the stronger of
the two lights.’”’ Loeb’s statement applies to the reactions of
Zeuzera larvae but it certainly does not hold for the reactions
of Calliphora larvae nor for that of many other animals, a criti-
cism of Loeb’s work which has also been expressed by others.
Adults of Drosophila and Calliphora, insects with image-forming
eyes, creep or fly toward the lights and, like Zeuzera larvae,
when balanced lights of different wave lengths are used, are
oriented by the more potent color. In all of these tests with
adult flies, however, a certain percentage of the individuals go
away from the source of the light which to the majority of the
flies is the most effective stimulus. It is difficult to explain this
difference in organisms of the same species, reared under identi-
cal conditions and tested with the same lights. To say it is
due to a difference in physiological condition brings us no nearer
to the solution. It is evident from the results shown in table 8
that the flies which go away from the more effective light at the
first trial do so not because of a permanent difference in their
phototropism, but because of mere chance or some unknown
510 ALFRED O. GROSS
factor operating in their initial orientation. If the fly on enter-
ing the chamber illuminated by opposing colors of nearly the
same wave length (as, for example red and yellow) is headed
directly away from the more effective light, the chances are it
will remain under the influence of the less potent light, since
more of the elements of the compound eyes are effected by this
light than by the other. In a case where there is a great differ-
ence in the efficiency of the two opposing lights (as, for example,
when red and blue, or blue alone is used) the more refrangible
rays, though impinging upon relatively few elements of the eye,
are so effective that the insect will be oriented towards the blue
light even when it enters the cylinder facing away from the more
potent light, or in a direction oblique to the direction of the rays.
The reactions of Periplaneta americana in response to colored
lights of equal intensity are remarkable in being positive to blue
and negative to green and yellow. It is not an uncommon
occurrence to produce a reversal of phototropism among the
lower organisms by changing the nature of their environment or
by using extreme differences in the intensity of illumination, but
I know of no recorded observation in which an animal has been
shown to be positive to one monochromatic light and negative
to another of the same intensity, but of different wave length.
Graber (’83), working with a closely allied species of cockroach,
Blatta germanica, found that these insects collected on the red
side of a two-chambered compartment illuminated respectively
with blue and red sheets of glass. From these results one would
infer that Blatta is negative to blue, not as Graber (’84, p. 152)
interprets a “blauscheues resp. rot-holdes Insect.’’ No one has .
ever verified with spectral light Graber’s results on the cock-
roach so the discrepancy in our results is in all probability due
to difference in method. The blue glass used by Graber trans-
mitted not only the blue rays but red and green, some yellow,
and the invisible heat rays. Furthermore Graber did not state
in just what manner the aggregation of Blatta took place.
At the present state of our knowledge, it is difficult to offer
an explanation of the reversal of the phototropism of the cock-
roach to the different colors.
REACTIONS OF ARTHROPODS TO LIGHTS 511
If it is assumed that the reactions of Periplaneta are brought
about by chemical changes produced by light on the photore-
ceptors of the organism, it is conceivable that there exist within
the organs photo-chemical substances, not unlike triphenylfulgid,
which have reversible reactions. In blue or violet triphenylfulgid
changes from orange-yellow to a dark brown compound; in the
red and yellow rays this reaction is reversed. Though no such
substances are known to exist in the body of the cockroach,
it is nevertheless an interesting suggestion and when our knowl-
edge of the chemistry of the reactions of animals to light is better
known, it is possible that explanations of such complicated reac-
tions as those of Periplaneta, as well as those of other organisms,
may be easily explained.
e
6. SUMMARY
1. The larvae of Calliphora and Zeuzera and the adults of
Calliphora, Drosophila, and Feltia are responsive to each of the
four colors of the spectrum used in these investigations.
2. The colors in the order of their efficiency in stimulating
the larvae of Calliphora are beginning with the most effective:
green, blue, yellow, red.
3. The order of the effectiveness of the colors in stimulating
the larvae of Zeuzera and the adults of Calliphora, Drosophila,
and Feltia is the order of the natural sequence of the colors in
the spectrum: blue, green, yellow, red.
4. The relative stimulating efficiency of the rays of any part
of the spectrum is independent of intensity, and is not the same
for all animals nor for different ages of the same animals.
5. The more refractive rays of the spectrum are not always
the most effective in stimulating an organism.
6. Periplaneta americana is responsive to blue, green and
yellow, but is indifferent to red. It is positive to blue and nega-
tive to both green and yellow. Green and yellow are practically
equal in their stimulating efficiency, as far as the reactions of
Periplaneta are concerned.
512 ALFRED O. GROSS
7. There is a reversal of phototropism in the reactions of
Periplaneta to the different colors of the spectrum which, as far
as the tests in these investigations indicate, is independent of
the intensity of the light and of any chemical condition of the
environment.
BIBLIOGRAPHY
Bert, P. 1868 Les animaux voient-ils les mémes rayons lumineux que nous?
Mem. Soc. Sei. phys. et Nat. Bordeaux, tom. 6, pp. 375-383.
Betue, A. 1898 Diirfen wir den Ameisen und Beinen psychische Qualitaten
zuschreiben? Arch. f. ges. Physiol., Bd. 70, pp. 15-100.
Bonnier, G. 1879 Les Nectaires, étude critique anatomique et physiologique.
Ann. sci. nat. Bot., sér. 6, vol. 8, pp. 46-48.
Buitman, G. W. 1899 Bees and the origin of flowers. Nat. Sci., vol. 14, pp.
128-130.
CaRPENTER, F. W. 1905 The reactions of the pomace fly (Drosophila ampelo-
phila Loew) to light, gravity, and mechanical stimulation. Amer.
Nat., vol. 39, pp. 157-171.
CuapMan, J. W. 1911 The leopard moth and other insects injurious to shade
trees in the vicinity of Boston. Cont. Ent. Lab. Bussey Inst., Har-
vard Uni., no. 48. Boston Svo, 51 pp., pl. 1-7.
Davenport, C. B. 1897 Experimental morphology. New York, 8vo, pt. 1,
XIV + 280 pp.
Day, E. C. 1911 The effect of colored light on pigment migration in the eye
of the crayfish. Bull. Mus. Comp. Zodl., vol. 53, pp. 305-348, pl. 1-5.
ENGELMANN, T. W. 1882 Ueber Licht- und Farbenperception niederster Organ-
ismen. Arch. f. ges. Physiol., Bd. 29, pp. 387-400.
FIELDE, ADELE M. 1902 Notes on an ant. Proceed. Acad. Nat. Sci., Phila-
delphia, vol. 54, pp. 599-625.
Foret, A. 1888 Expériences et remarques critiques sur les sensations des in-
sectes. Rec. Zool. Suisse, tom. 4, pp. 1-50; 145-240; 515-523.
1901 Sensations des insectes. Pt. III. Revista biol. gen., tom. 3,
pp. 7-62.
1904 Ants and some other insects. Transl. by W. M. Wheeler. Chi-
cago, 8vo, 49 pp.
GratacaP, L. P. 1883 Color preferences in nocturnal Lepidoptera. Amer.
Nat., vol. 17, pp. 791-792.
REACTIONS OF ARTHROPODS TO LIGHTS 513
GraBER, V. 1883 Fundamentalversuche ueber die Helligkeits-und Farbenemp-
Herms, W
Hess, C.
LAURENS,
Lors, J.
LUBBOCK,
Mast, S.
findlichkeit augenloser und geblendeter Thiere. Sitzb. Akad. Wiss.,
Wien, math.—naturw. Cl., Bd. 87; Abt. 1, pp. 201-236.
1884 Grundlinien zur Erforschung des Helligkeits- und Farbensinn
der Tiere. Prag. und Liepsiz, 8vo, vi11 + 322 pp.
.B. 1911 The photic reactions of sarcophagid flies, especially Lucilia
caesar Linn. and Calliphora vomitoria Linn. Jour. Exp. Zodl., vol.
10, pp. 167-226, 25 fig.
1909 Untersuchungen ueber den Lichtsinn bei wirbellosen Tieren.
Erste Mitteilung. Arch. f. Augenheilk., Bd. 64, Ergingzungsheft, pp.
39-61.
1910 Neue Untersuchungen iiber den Lichtsinn bei wirbellosen Tieren.
Arch. f. ges. Physiol., Bd. 136, pp. 282-367, Taf. 5-10.
H. 1911 The reactions of amphibians to monochromatic lights of
equal intensity. Bull. Mus. Comp. Zodl., vol. 53, pp. 253-302.
1890 Der Heliotropismus der Thiere und seine Uebereinstimmung
mit dem Heliotropismus der Pflanzen. Wirzburg, 8vo, 118 pp.
1893 Ueber kiinstliche Umwandlung positiv heliotropischer Thiere in
negativ heliotropische und umgekehrt. Arch. f. ges. Physiol., Bd.
54, pp. 81-107.
1905 Studies in general physiology. Pt. I, Chicago, 8vo, x1 +
423 pp.
J. 1879 Observations on ants, beesand wasps. Pt. V. Ants. Jour.
Linn. Soc., London, vol. 14, pp. 265-290.
1881 Observations on ants, bees and wasps. Pt. VIII. Jour. Linn.
Soc. London, vol. 15, pp. 362-387.
1881 a On the sense of color among some of the lower animals. Jour.
Linn. Soc., London, vol. 16, pp. 121-127.
1883 On the sense of color among some of the lower animals. Pt.
II, vol. 17, pp. 205-214.
1904 Ants, bees and wasps. New York, 8vo, x11 + 435 pp.
O. 1911 Light and the behavior of organisms. New York, 8vo,
xi + 410 + 17 pp.
MEREJKOWSKY, C. DE. 1881 Les crustacés inférieurs distinguent-ils les couleurs?
Compt. Rend. Acad. Sci., Paris, tom. 93, pp. 1160-1161.
ParRKER, G.H. 1903 The phototropism of the mourning-cloak butterfly, Vanessa
antiopa Linn. Mark Anniversary Volume, New York, 4to, pp. 453-
469.
514
PECKHAM,
PEREZ, J.
PLATEAU,
RAp1, E.
ALFRED O. GROSS
G. W., anp E. G. 1887 Some observations on the special senses
in wasps. Proceed. Nat. Hist. Soc., Wisconsin, pp. 105-112.
1894 The sense of sight in spiders with some observations on the
color sense. Trans. Wisconsin Acad. Sci., Arts and Letters, vol. 10 pp.
231-261.
1894 Notes zoologiques. De l’attraction exercée par les odeurs et
les couleurs sur les insects. Acta soc. Linn., Bordeaux, tom. 7, pp.
245-253.
T. 1897 Comment les fleurs attirent les insects. Recherches expéri-
mentales. Troisiéme partie. Bull. Acad. roy. Belgique, ser. 3, tom.
33, pp. 17-41.
1899 La choix des couleurs par les insectes. Mémoires Soc. Zool,
France, tom. 12, pp. 336-370.
1901 Untersuchungen iiber die Lichtreaction der Arthropoden. Arch.
f. ges. Physiol., Bd. 87, pp. 418-466.
Witson, E. B. 1891 The heliotropism of Hydra. Amer. Nat., vol. 25, pp.
413-433.
YERKES, R.M. 1899 Reactions of Entomostraca to stimulation by light. Amer.
Jour. Physiol., vol. 3, pp. 157-182.
1900 Reaction of Entomostraca to stimulation by light. II. Reac-
tions of Daphnia and Cypris. Amer. Jour. Physiol., vol. 4, pp. 405-
422.
STUDIES OF FERTILIZATION!
V. THE BEHAVIOR OF THE SPERMATOZOA OF NEREIS AND ARBACIA
WITH SPECIAL REFERENCE TO EGG-EXTRACTIVES
FRANK R. LILLIE
The Hull Zoélogical Laboratory, University of Chicago
‘
FIVE FIGURES
CONTENTS
POISE OCUICUION A: © 7. SEIIEES os oo oees 1 MEE on os MRR EG cleo girs ols sp osaees 516
PePaGHE CALTON PHCNOMEH AME Mates. . 252.0 Pe.e... sso cidoeeacues eknlnds wen 519
JN INSIST ERE Seen 2 ho 6 Sls 5 Se ods eee, SORT EA NPR Nees anats aimee ie eae 519
Pe anne Ae Per AMER CAUINON,.. OF 5. s-c.< o<jag> ea es opt oheea = wet 2's: 519
2. Individual movements of the sperm.....................--+-+- 522
Swe PESTN CSS Ae eee EE es. ket Scene ete reo estos 523
4 SREMIPCEAGCUEC eR ERr et Ras < ode ot Re ee ese sale ad ee 524
5. Chemical constitution of the medium........................ 525
RCACTO Sees poe ETM ooo. sw cers Shee eS Tn Se Te EAT® Sn TSE 526
PepAtoglitnen 7. : see net). 3. ssc wd to tideres sean oem 528
GrAlconolband Gimerssaseaes . .s oe te Oe e e 529
Geb po— nad, Wyper=bGniGIbY 2 as6). 5... 5 «is waidiac 2» Asean ees 530
Le JT DON 6 Oe I. ee Se) et aT ee > 531
Pia PerecAtrOn PHCHOMEDA. Fp. sens kes coe so ee been de euteemeeoee 532
Jie, INGEST Reams Cole Pe Ro) SE | Voie CB se | a I Pea 533
i pAreregationswith rererener to CO; «sc 5.850 ¢ 0/4 > sina eee 533
2. Interpretation of the aggregation reaction................... 538
ae enetignaie OLNER- BEIGs ne. . o oc. Sas Does s casino a 541
4. Behavior with reference to alkalis.........................5. 541
9.-eactions. to other sabshanees. .....0.%,. 2..-0% «+n cee: 542
Gs CERErMOLRXIN: 44.00). Pee Re ee a Bere hckc Sees 543
Tet eee OU RIN ong eee yes, 1.4 ccs dca oe ois Coe io ee 543
Sc WAPICIONS- Of PEACION ot 40 sos se wi os ee ee 544
oj Chemotaxis tolere-necretionse... <.icl-< eons c.g) eek: 546
B. Arbacia. .
1 The earlier studies of this series bore the general title ‘‘Studies of fertilization
in Nereis.”’
515
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 4
516 FRANK R. LILLIE
IV. Agglutination phenomena and reactions of spermatozoa to egg-secre-
4I0DS. oc. eee oe ck cc. « Re RRE aes ee ee Ae ee eee 548
1: introduction. .,. ..33<5c2 2 celeste ase eee: Bee eee 548
2. Initial experiment... 22. 605:..-2. oe oes eee Soe ee 550
3. Ova alone produce the agglutinating substance.............. 552
4, Fixation of the agglutinin by spermatozoa.................. 554
5. Nature of the effect on the:sperm:.. 9. .<>- .-- eee 555
6. Thermo-resistance of the agglutinin........................- 557
7. Fertilizing power of agglutinated sperm..................... 558
8. Conditions of formation of the agglutinin by the eggs....... 559
9. Hetero-agglutination and the question of specificity......... 561
10. Interpretation of some phenomena of normal fertilization of
Nereis: 2.02.06 Gee 35 56.55 Belge ot 3 eee eee 567
ii. Summary: Part TV... 3.2.0.2 22..2-58- eee ee 568
Vi DISCUBSIONS=.55.- hese eo es ee. eee 569
hiterature Cited yas: Se aoe ce eee Fs eo en: eee ee 574
I. INTRODUCTION
The earlier studies of this series dealt exclusively with Nereis
and concerned the cortical changes of the egg, partial fertiliza-
tion, the morphology of the normal fertilization and the fertilizing
power of portions of the spermatozoon. They yielded certain
positive results which I need not review, but they had convinced
me that other methods than the ones usually in vogue, includ-
ing the methods of artificial parthenogenesis, are needed for a
closer approach to some fundamental problems of fertilization.
Some incidental observations drew my attention to the study of
the behavior of the spermatozoa, and investigation of the subject
soon showed that the reactions of these minute active reproduc-
tive elements might furnish evidence of considerable significance.
This study was begun in the summer of 1911 and continued
throughout the summer of 1912 at the Marine Biological Labora-
tory.
With the publication of Loeb’s first study on artificial partheno-
genesis the study of fertilization entered upon a new phase which
has not yet run its entire course. The tendency during this
phase of investigation has been to regard initiation of develop-
ment as the fundamental problem of fertilization; and the aim
has been to discover the way in which the spermatozoon induces
development of the egg. Hence the term ‘chemical fertilization’
STUDIES OF FERTILIZATION aie
has come to be used loosely as practically synonymous with artifi-
cial parthenogenesis, as though a salt solution could take the
place, and play the réle, of the spermatozoon. This it can do
obviously only with reference to the initiation of development,
which, so far from being the only function of fertilization, is more
properly to be regarded as a secondary function, or better a separ-
ate phenomenon which is sometimes associated with fertilization,
sometimes not. On the one hand we may have initiation of
development without fertilization, as in parthenogenesis and all
asexual modes of reproduction, and on the other hand the phe-
nomenon of fertilization without initiation of development is
extremely common, as in the so-called winter eggs of Cladocera,
Aphids and Rotifera, where fertilization is followed by a long rest-
ing period; the Protozoa and unicellular algae also offer many in-
stances in which fertilization is the immediate prelude to a long
resting stage.
The study of initiation of development by chemical means has
yielded results of prime importance, and the consequent absorp-
tion in these problems has been an inhibiting factor in the analysis
of other problems of fertilization. ‘Thus, as spermatozoa are not
necessary for “‘chemical fertilization,’ the study of their behavior
has been largely neglected. The problem of specificity has as a
consequence been left almost entirely out of account, for there
is no specificity in salts, or even in the blood sera of animals of
other phyla; nevertheless specificity in reaction of sexual products
is a much more nearly universal phenomenon of fertilization than
initiation of development, and it is quite possible that the solu-
tion of this problem may furnish a valuable clue in the study of
the latter problem. In any event, the time seems ripe for the
development of new methods of attack on the fundamental prob-
lems of fertilization. The present contribution is a step in this
direction. I have taken up the study of the behavior of the sper-
matozoa, because it represents, after all, considered in a broad
sense, one-half of the problems of fertilization, and it seems prob-
able that these small motile cells may prove better indicators of
some of the reactions involved in fertilization than the slowly
reacting egg.
518 FRANK R. LILLIE
There are three categories of behavior exhibited by spermatozoa
that seem to me of importance for the problem of fertilization,
because all are exhibited in reponse to egg secretions. These
are (1) activation, (2) aggregation, (3) agglutination. The phe-
nomena of activation are involved in those conditions that affect
the activity of spermatozoa. The phenomena of aggregation are
positive taxic responses, for the most past chemotactic. The
phenomena of agglutination are exhibited in the presence of sub-
stances that cause the spermatozoa to adhere in masses. In a
preliminary paper I have described some aspects of these phenom-
ena (Science, N. 8., vol. 36, October 18, 1912). We may con-
sider the subject matter under these three heads.
The spermatozoa of marine animals in which fertilization takes
place in the sea-water offer advantages for study probably greater
than those of any other forms, because the conditions of normal
activity are given in the sea-water itself, no secretions of accessory
glands either of the male or of the female being requisite. More-
over, the spermatozoa may be obtained in large quantities. They
offer, thus, material directly accessible to experimental work with
the simplest possible facilities. Among the forms available for
work, Nereis and Arbacia were soon found to be best adapted
because the breeding season extends through most of the summer
and they furnish material in large quantities. The present paper
is therefore confined almost exclusively to these forms.
Suspensions of the spermatozoa in sea-water formed the mate-
rial for all of the experiments. The reactions vary somewhat
according to the density of the suspensions, and it may be impor-
tant in future experiments to find some quantitative method of
expressing the variations in density. But for the purposes of
this paper it will be sufficient to indicate the extremes as opales-
cent, milky and creamy, with intermediate qualifications. An
approximation to uniformity was attained in many of the experi-
ments by adding a certain number of drops of the dry sperm?
to measured quantities of sea-water.
* By ‘dry sperm’ is meant the sperm as it comes from the testes without the ad-
mixture of fluid.
STUDIES .OF FERTILIZATION 519
II. ACTIVATION PHENOMENA
. A. NEREIS
1. The aggregation reaction
We may begin the discussion of the behavior by describing
a phenomenon which was used throughout the experiments with
Nereis as a test of the activity of the spermatozoa. A drop of
dry sperm from a mature Nereis is mixed in about 6 cc. of sea-
water in a Syracuse watch crystal, making a uniformly milky
suspension; in a few seconds clouds begin to appear and in fifteen
to forty-five seconds these usually draw together in dead-white
solid-looking masses uniformly spaced throughout the fluid. The
intervening fluid becomes quite clear and the masses quickly
settle on the bottom. The rate of formation of these masses,
their number and size, depend on temperature, ‘freshness’ of
the sperm and other conditions discussed beyond. Any sperm
suspension that exhibits the aggregation phenomenon will be
called ‘aggregative’ in the experiments that follow.
The appearance is of course due to the aggregation of the sperm
in closely packed masses. Under a low power of the microscope
each mass appears like a swarm of bees, owing to the intense activ-
ity of the peripheral spermatozoa. But those in the interior of
the dense mass must be quiescent or the masses would break
apart. After the aggregations have settled to the bottom of the
crystal they tend to flatten out and may run together in time to a
greater or less extent.
If, immediately after settling of the aggregations, the sperm
is mixed up with a pipette, a perfectly uniform milky suspension
is again produced, which may aggregate a second time, but more
slowly than the first time, and in fewer aggregates; and the inter-
vening fluid remains quite opalescent, showing that all the sper-
matozoa have not joined the aggregates. A third mixing up is
not usually followed by aggregation until after the spermatozoa
have settled to the bottom, and then only a very partial aggrega-
tion results.
A considerable number of variations of this theme can be pro-
duced by using sperm suspensions of varying density contained
520 FRANK R. LILLIE
in vessels of varying form, et cetera; under certain conditions the
aggregations may arise in conformity with the water currents set
up by the last emptying of the pipette, et cetera. But a descrip-
tion of these variations would be useless without the analysis of
the causes of the phenomenon, which is taken up later.
All the experiments on Nereis to be described beyond were made
with aggregative sperm, so that there was always a test, which had
the advantage of being macroscopie and quick, of the activity of
the sperm used in the experiments, and this has much to do with
uniformity of results.
To give a more concrete idea I reproduce three photographs,
natural size, of the phenomenon of aggregation. The first (fig. 1)
was taken ninety seconds after mixing a drop of dry sperm in
about 8 cc. of sea-water. The aggregations are quite uniformly
distributed except in the upper right quarter where their arrange-
ment marks out original currents produced by mixing with the
pipette. Figure 2 was likewise taken ninety seconds after
mixing; the effect of water currents on the arrangement of the
aggregations is shown here quite well on the left. Figure 3 was
taken three minutes after mixing, and the separate aggregations
are beginning to fuse together on the bottom.
I propose to discuss in this section simply the conditions which
modify the activity of the spermatozoa. In the case of Nereis
such conditions may be inferred from two kinds of observations,
namely: (1) The appearance of activity presented to the eye under
the microscope and (2) the rate and degree of the aggregation reac-
tion which is macroscopic. Nereis is the only form with which I
am familiar that exhibits the latter reaction in any marked way.
Its sperm is therefore better adapted than that of any other species
for study of conditions of activity. The observations of different
samples of sperm under the microscope are very difficult to com-
pare as to degree of activity, as one is never sure of the successive
subjective impressions, but in the case of Nereis these can be
checked by the aggregation reaction.
The principle conditions that affect activity are ‘freshness,’
temperature, and the chemical constitution of the medium.
These conditions will be considered not exhaustively at all, but
STUDIES OF FERTILIZATION 521
Fig. 1 Photograph of aggregation of a Nereis sperm suspension, taken 90 see-
onds after mixing the suspension; natural size; description in text.
Fig. 2 Another suspension photographed at 90 seconds; description in text.
i rhe FRANK R. LILLIE
Fig.3 Another suspension photographed 3 minutes after mixing; description in
text.
only to the extent that they appear to be significant for the
phenomena of aggregation and agglutination, which are the main
problems for our consideration.
2. Individual movements of the sperm
To explain the various reactions of the sperm it is necessary to
consider first some of the more obvious features of locomotion of
individual spermatozoa. In their free movements through the
water they describe, as is well known, spiral paths. In Nereis the
successive turns of the spiral are rather close set. As soon as a
spermatozoon comes in contact with a surface, the movements of
translation cease, and circus movements begin. ‘The sperm moves
round and round in a circle of varying diameter in contact with
the surface. In the case of a preparation beneath a cover slip
on a slide, those in contact with the slide rotate anticlockwise,
those in contact with the cover-slip clockwise. The direction of
rotation is always the same. It is associated no doubt with struc-
tural asymmetry which I described briefly in Study III (Lillie
STUDIES OF FERTILIZATION 523
12). The tail of the spermatozoon is attached not to the center
of the middle piece, but on one side.
The movement is of course due to successive beats of the tail
and it is a very interesting fact that under certain conditions of
aggregation the successive beats of the spermatozoa forming an
aggregation may become synchronous, and under such circum-
stances the number of beats approximates 120 a minute at
temperature of 21°C., if the sperm be fresh. This phenomenon,
which I have not yet attempted to analyze in detail, follows after
the aggregations of sperm from a fresh suspension have settled on
the bottom of the container and begin to spread out of their own
weight. Its appearance may be accelerated by gentle agitation
of the dish, which tends to spread out the aggregations. Syn-
chronous movements appear when the sperm spread out in a kind
of membrane from an aggregated condition. In such a case the
synchronous beats spread over the membrane thus formed, like
waves of contraction over a ciliated epithelium. In fact, a kind of
synthetic ciliated epithelium is then established. The interest of
the phenomenon in the present connection is that it furnishes a
clear demonstration of the successive beats of the tail of the sper-
matozoa, which are not readily distinguishable, and certainly
cannot be counted, in the case of a single spermatozoon.
The movements of the spermatozoa are, then, due to succes-
sive beats of the tail, which is so placed as to cause rotation in a
definite direction. The movement when freely suspended in
water is in a spiral path, but when in contact with a surface
the translatory component of the locomotion is almost entirely
eliminated.
The following account of the behavior does not deal directly
with individual movements, but always concerns mass-reactions,
from which the behavior of the individual spermatozoa may be
inferred.
3. ‘Freshness’
The spermatozoa are absolutely immotile while they are in
the body of the male, but become intensely active when suspended
in sea-water. This expresses itself in the formation of aggrega-
tions; but, as already noted, aggregations form more slowly after
524 FRANK R. LILLIE
a second mixing up, and only to a slight extent or not at all after
a third mixing. This condition of relative inactivity, or staleness,
is reached in a few minutes, but varies more or less according to the
density of the suspension, a very dense suspension exhibiting it
more quickly than one less dense. The activity of the sperm
may be restored, partly at least, by the addition of fresh sea-
water, which shows that the staleness is not due to exhaustion,
but to the accumulation of by-products of activity in the sea-
water. Of these the chief is probably COs., as will be shown by
experiments described beyond. The formation of CO, by the
activity of the spermatozoa themselves is indeed one of the chief
causes that limits their activity when sufficiently concentrated to
form milky suspensions. To obtain the best results with the
experiments described it is necessary to work with fresh sperm;
otherwise, the accumulation of CO, may obscure the reactions.
4. Temperature
In 1911 a series of observations were made on the effect of tem-
perature on the aggregation reaction of fresh sperm. In general
the results as tabulated are:
13°C. No aggregations form
15°C. Slight signs of aggregation in 4 minutes
18°-19°C. Aggregation in from 2 to 4 minutes; much fewer in number than at
higher temperatures
20.5°C. Aggregations, numerous, in 1 minute
23.5°C. Aggregations, yet more numerous, in 30 seconds :
26.5°C. No aggregations form at this temperature, but they form as the tem-
perature falls to 23°C,
In general temperatures from 20° to 23.5°C. are optimum for
the aggregation phenomenon. At 15°C. the movements of the
spermatozoa are too slow, and at 26.5°C. the movements are
extremely active, but apparently uncoordinated, so that the aggre-
gation reaction is not given. These figures possess no absolute
value, but they indicate approximately the limits of temperature
within which the reaction may be expected. The normal tem-
perature of the sea-water varies from about 18° to 22°C. at Woods
Hole during the breeding season.
On
nN
Or
STUDIES OF FERTILIZATION
5. Chemical composition of the medium
_ The effect of the chemical composition of the medium upon the
activity of the spermatozoa is a very complicated subject, and no
attempt has been made to analyze it farther than was necessary
for comprehension of the forms of behavior studied. Even the
simplest experiments furnish convincing proof of the dependence
of activity of the spermatozoa upon a constant chemical com-
position of the medium; and this extreme susceptibility is
certainly a prime factor in the behavior of the spermatozoa. To
determine something of its limits becomes therefore necessary.
One of the first questions that presents itself is obviously the
relation of the activity to the various salts of the sea-water. This
is, however, in itself a problem of so much complexity that I have
hesitated to undertake it; especially as it is unnecessary for our
present purpose, seeing that the behavior to be studied takes
place in sea-water as its medium. The few observations made
demonstrate that spermatozoa of Nereis are paralyzed in pure
M/2 solutions of NaCl, KCl, CaCl, and MgCl:. As these are
the principal salts of sea-water, it is obvious that the activation of
the spermatozoa in the sea-water is a question of balance of
salts. I therefore tried Van’t Hoff’s solution, namely: 100 cc.
M/2 NaCl + 2.2 ec. M/2 KCI + 2 ce. M/2 CaCl, + 12 ec. M/2
MgCl, but the spermatozoa did not activate in this solution either.
Some other experiments were made, which did not materially
help the problem, which was not followed farther. The later
experiments all assume the sea-water as the given medium.
Some early observations in the course of this work had shown
that the female excretes certain substances in the sea-water that
have a strong inhibiting effect upon the activity of the spermato-
zoa. This is so marked that sperm suspensions made up in sea-
water sufficiently charged with secretions of the females never
exhibit the aggregation phenomenon, and their fertilizing power
is markedly reduced. This fact repeatedly observed suggested
tests on the susceptibility of the spermatozoa to CO, dissolved in
the sea-water, and this formed the beginning of a series of tests
that involved acids and alkalis and some other substances.
526 FRANK R. LILLIE
a. Susceptibility to acids, including CO.. The susceptibility of
spermatozoa of Nereis to acids was tested by opening a male
Nereis in a dry watch crystal, and mixing a drop of the thick
sperm, which flows out, in 6 to 8 cc. of the solution to be tested.
The effect on the movements of the spermatozoa was then ob-
served as rapidly as possible, first with the naked eye to control
the aggregation reaction which is given only by very active sper-
matozoa, and second with the microscope to note the degree of
activity if their movements were sufficiently slowed down to
prevent the aggregation reaction. For each experiment a control
suspension of spermatozoa in normal sea-water was run, and only
those experiments are taken into account in which the control
aggregated in ninety seconds or less. The acid solutions were
made by adding a sufficient quantity of N/10 dilutions in dis-
tilled water to a measured quantity of sea-water to reach the dilu-
tions tested. These were always so weak as not to involve the
question of osmotic changes in the results. The results may be
tabulated as follows:
H2SOx HC! HNO; | CH:.COOH
ANOS cc ee 0 0 0 0
12000. oe Se | (hate i= 1 0
Mr /S000 2G ae: be: Jae | 2 | vee 1 | 0
ING S000) ace 6 colo oe | 3 4 3— 1
INV TOO00 So. Beecher ae | 4 4 4 | 4
In this table (0) stands for complete paralysis
(1) represents minimum amount of movement; usually only a few
spermatozoa moving
(2) Fairly active, but no aggregations form
(3) Active; aggregations form, but are few in number and require
over two minutes to appear
(4) represents maximum activity, aggregations forming at least as
rapidly as in the control, i.e., in less than ninety seconds
While the four grades of activity noted are readily to be dis-
tinguished in Nereis, their relation to the grades of acidity in ques-
tion is not to be taken as fixed and invariable. As a matter of
fact the various observations show considerable variation with
reference to the intermediate dilutions, N /2000 to N/5000, in the
case of the mineral acids. But in the case of the extremes N /1000
STUDIES OF FERTILIZATION 527
always produces complete paralysis very quickly, and N/10000
always permits maximum activity. The range of activity with
reference to these acids is thus marked out fairly well. It will
be noted that acetic acid has a greater inhibiting effect than the
mineral acids.
It is an interesting question in this connection whether there is
a certain optimum amount of acidity which increases rather than
decreases the activity of the spermatozoa. In the experiments
now under consideration this could not be determined certainly.
In some cases spermatozoa aggregated more rapidly in weakly
acid solutions (N/5000 and under) than in the control; in others
at the same rate or at a slightly less rate. In the experiments on
chemotaxis, however, which involve an acid gradient, there is
possible evidence of stimulation of weak solutions.
CO.. The sensitiveness of the spermatozoa to COs, is con-
sidered separately because of its probable biological significance
and also because it was impracticable to state the strength of the
solution in molecular terms. ‘The solutions were prepared empiri-
cally as follows: A certain quantity of sea-water was supersatura-
ted with CO, in ‘Sparklet’ siphons. The charged sea-water was
drawn as desired, and after the effervescence ceased it was diluted
with measured quantities of sea-water, and the dilutions were
expressed as percentages of the charged sea-water. These solu-
tions were always prepared fresh for each experiment, and kept in
stoppered bottles or otherwise covered as far as possible. The
uniformity of the reactions obtained is adequate proof that the
solutions used in the different experiments were equivalent.
A very large number of experiments was made with CO, during
the course of the summer, so that the relations of the spermatozoa
to CO. were more adequately ascertained than for any other
substance. Here the question is only of the relation of CO»
tension to the activity of the spermatozoa, and the results may be
stated as follows:
One per cent of the charged CO, sea-water paralyzed the sper-
matozoa immediately; or rather a suspension of a drop of dry
sperm in 6 to 8 cc. of this strength of CO, does not exhibit any
activity. This is, however, very near the minimum paralyzing
528 FRANK R. LILLIE
dilution, and some samples of sperm will exhibit slight movements
in it.
In 0.75 per cent CO, sea-water no aggregations take place,
but the spermatozoa move feebly.
In 0.5 per cent CO» sea-water aggregations usually form slowly,
but the activity is usually less than the control.
In 0.33 per cent CO, sea-water there is apparently no inhibition
of activity as compared with the control.
Whether lower dilutions stimulate more than normal sea+water
is difficult to say by the method used here. But the chemotaxis
experiments possibly indicate stimulation at a certain optimum
(see p. 535).
The sensitiveness of Nereis spermatozoa to CO: is thus surpris-
ingly great, and it operates within very narrow limits. This is
the more surprising when comparison is made with spermatozoa
of other species. Thus I ascertained that the sperm of Loligo will
move, though feebly, in 50 per cent CO.-charged sea-water, and
that it is very active in 20 per cent, though less so than in normal
sea-water. In the case of Chaetopterus it requires about 33}
per cent to 40 per cent of the CO, sea-water to completely para-
lyze all the spermatozoa, though 10 per cent inhibits considerably.
Arbacia sperm on the other hand is much more sensitive to COsz,
being completely paralyzed in 3 per cent. But Nereis is very
much more sensitive than any of these, and this involves some
very interesting forms of behavior described later on.
b. Sensitiveness of spermatozoa of Nereis to alkalis. Alkalis
above a certain concentration agglutinate the spermatozoa of
Nereis, and cause them to stick together in masses. This is
never seen in acids, however strong. I can best state the sen-
sitiveness of the spermatozoa to KOH by giving the protocol of
a single experiment (June 23, 1912) which followed some prelimi-
nary determinations. N/10 KOH in distilled water was used as
the standard solution. Added to sea-water this solution produces
a precipitate which redissolves up to about N/2500 KOH.
In the experiment a drop of dry sperm was stirred in about 8 ce.
of each of the following dilutions in sea-water, and observations
made as noted.
STUDIES OF FERTILIZATION 929
1. N/2500 KOH. Produces very rapid agglutination of the
spermatozoa; free sperm between show some movement.
2. N/5000 KOH. No agglutination; no aggregation; sperms
fairly motile.
3. N/7500 KOH. No agglutination; no aggregation; sperms
more active.
4. N/12500 KOH. No agglutination; no aggregation; sperms
very active. ;
5. N/25000 KOH. No agglutination; aggregations form slow-
ly; but sperms are extremely active.
6. Normal sea-water. Control. Aggregations form in half
minute. Very active sperm (maximum).
This experiment was carried out with the sperm of one male at
one time, the solutions being prepared in advance. The limits
of the agglutination effect are given. But it is improbable that
the inhibiting effect extended to the lower limit, although aggre-
gations were formed so slowly in N/25000. The reason for this,
as will be shown later, is that the aggregation effect is due to posi-
tive chemotaxis to a weak acid, probably CO:, produced by the
spermatozoa themselves. This is neutralized by the KOH so
that in spite of the great activity noted in N/25000 KOH aggre-
gation cannot take place until after neutralization of the alkali.
If the behavior of the spermatozoa be observed under the micro-
scope at the moment they are put into N/2500 KOH, there is
seen momentary great activity of the spermatozoa followed
quickly by agglutination as described.
The relations to NaOH were essentially the same. There was
slight agglutination in N/5000 NaOH, and the slightest appear-
ance of aggregations in N/25000 NaOH. .
c. To alcohol and ether the sensitiveness is as follows:
Alcohol:
(1) 5 per cent, sperm are paralyzed
(2) 2 per cent, some activity; no aggregations
(3) 1 per cent, more active, some aggregations may form in five minutes
(4) 0.5 per cent, few aggregations in about three minutes
(5) 0.2 per cent aggregations in forty-five seconds
(6) 0.1 per cent, aggregations in thirty-two seconds
(7) Control, sea-water; aggregations in thirty seconds
530 FRANK R. LILLIE
(4), (5), (6) and (7) came from an experiment with the same sam-
ple of sperm.
The sensitiveness to ether is essentially the same, though the
sperm did not aggregate even at 0.33 per cent. The chemotaxis
experiments with ether indicate a possible stimulation of the
sperms at an optimum concentration (see beyond).
As stated before, no attempt was made to carry the analysis of
the relation of activity. of the spermatozoa to known chemical
substances very far. Experiments on chemotactic and other
behavior phenomena of the spermatozoa were in progress at the
same time, and the determinations already given seemed fairly
adequate for the purposes of analysis.
6. Sensitiveness of spermatozoa to hypo- and hyper-tonicity
of the medium
As regards the sensitiveness of spermatozoa to hypo- and
hypertonicity of the medium, the following determinations may
suffice:
August 18, 1911. Sperm of one male; one drop mixed in each
of the following solutions, with results noted:
(1) 5 ce. sea-water + 2.5 ce. distilled water. The sperm are fairly active, but
no aggregations form.
(2) 5 ee. sea-water + 1 ce. distilled water
(3) 5 ee. sea-water + 0.5 ee. distilled water
(4) 5 ee. sea-water
(5) 5ec. sea-water + 1 ec. 5/2 N NaCl; sperm paralyzed
Aggregations form in one minute;
a little better in 4
The spermatozoa will thus stand considerable decrease in
osmotic pressure without much modification of activity. But
increase in osmotic pressure induced in the experiment by addi-
tion of NaCl and in others by KCl, CaCL, or MgCl, rapidly
paralyzes. The addition of this amount of KCl paralyzed
every sample of sperm used and its effect is undoubtedly toxic;
but some samples of sperm exhibited considerable, though
decreased, activity, when the other salts were used.
STUDIES OF FERTILIZATION Son
B. ARBACIA
The tests concerning the relation of the activity of spermatozoa
of Arbacia to the chemical composition of the sea-water were not
so extensive as in the case of Nereis. Moreover we do not have
any definite aggregation reaction here to serve as a measure of
activity. However, a sufficient number of tests were made to
show that Arbacia is much less sensitive to variations in inorganic
constituents than Nereis.
CO,. To afford comparisons we may use three forms here,
Chaetopterus, Arbacia and Nereis:
100% COz 40% CO2 20% COsz 10% COz 5% COs
Chaetopterus.... paralyzed few move manymove active | active
| slightly
AED ACR «Sete ss: paralyzed | paralyzed paralysed paralyzed | paralyzed
z 7
Nereis...........| paralyzed | paralyzed paralyzed | paralyzed | paralyzed
— = = — = — la = = —— — — —— = = = ———————————————
2.5% COz 1% COz 0.5% COz NORMAL SEA-WATER
Chaetopterus.... active active active active
AYPOSCIA. 2.2... ..'- traces of — fairly active maximum activity
movement active
INGTEIS ote ot ae paralyzed paralyzed fairly maximum activity (ag-
active gregation)
Thus, while compared to Chaetopterus, Arbacia, is extremely
sensitive to the presence of COs, compared to Nereis it is relatively
insensitive.
To other acids, H.SO;, HCl, HNO;, and CH; COOH, Arbacia
is also less sensitive than Nereis, exhibiting a fair degree of activ-
ity in N/1000 solutions in sea-water (compare table for Nereis,
p:526).
The sensitiveness to alkalis does not differ materially from that
of Nereis. Agglutination of the spermatozoa is caused by N /2500
KOH, giving a very pretty precipitation picture in a vial. Such
agglutinations are irreversible. Spermatozoa between the agglu-
tinated masses may be in motion.
The relatively slight sensitiveness of Arbacia sperm to CO:2 is
correlated with absence of any such striking aggregation effects
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 4
532 FRANK R. LILLIE
as are exhibited by suspensions of Nereis spermatozoa. But indi-
‘ations of the same kind of reaction may be seen under certain
circumstances. Thus, a fresh-suspension mounted beneath a
raised cover will soon exhibit cloud effects due to differences in
the density of aggregation, and this corresponds to the first stage
of aggregation in Nereis. In the course of an hour or so all the
spermatozoa retract from the edges into a central dense aggrega-
tion, and this is due, I believe, to the rising CO, tension towards
the center. Reasons for this opinion are given under the head of
the aggregation phenomena.
Fig.4 Reaction of spermatozoa of Nereis to a drop of 1 per cent CO, sea-water;
from an experiment of June 18, 1912. The original sperm suspension was made at
3.09: it aggregated clearly at 3.10 (fig. 1). It was then mixed up with a pipette and
some drops mounted on a slide beneath a raised cover-slip as in 1. The drop of
1 per cent CO, sea-water was introduced at 3.12 (on left) and a drop of pure sea-
water as control (drop to right). Figure 4, 7, shows the reaction at 3.13; 2, at
3.14; 3, at 3.144, and 4, at 3.16. In 4 the general suspension has aggregated. The
final position of the reactive sperm is in the center of the introduced drop. No
reaction takes place with reference to the drop of sea-water, which gradually be-
comes obliterated by inwandering of sperm. The figure shows also that the sper-
matozoa retract from the margin of the suspension.
III. AGGREGATION PHENOMENA
The spermatozoa of Nereis and Arbacia show very definite
positive chemotaxis toward acids and egg-extracts of the same
species, which may be demonstrated with striking clearness by
the method first introduced by Jennings in studying the behavior
of Paramecium. ‘The method as applied to behavior of spermato-
STUDIES OF FERTILIZATION 533
zoa consists in mounting some drops of a sperm suspension be-
neath a long cover slip supported by glass rods, and injecting a
drop of the fluid to be tested into the suspension. It then forms
a clear drop within the milky suspension, and reaction at once
begins at its borders. This method gives incomparably more
delicate results than Pfeffer’s method of using capillary glass tubes.
The drop is confined above by the cover and below by the slide
and diffusion takes place only at its margins; in this way a gradi-
ent is established. In the case of the capillary glass tubes diffu-
sion is so slight from the open ends that no delicate reaction can
be expected. So after a number of trials the capillary tube
method was abandoned and the injected drop method was used
exclusively.
A. NEREIS
1. Aggregation with reference to CO,
As introduction, the reaction of a fresh suspension of the sper-
matozoa of Nereis to a 1 per cent dilution of sea-water saturated
with CO, will first be described from a specific experiment:
June 26, 1912. A ripe male Nereis was placed in a dry watch
erystal and snipped with scissors; two drops of the dry sperm were
mixed in 10 cc. sea-water at 9.115 a.m. and made a milky suspen-
sion which aggregated freely in thirty seconds. Some drops of
this were then mounted beneath a cover slip supported by glass
rods about 1 mm. in diameter. A drop of the 1 per cent CO,
sea-water was injected at 9.13. In withdrawing the pipette a
trail of the CO, sea-water is left extending to the margin. Ina
few seconds the following configuration developed (fig. 4-7). It
consists essentially of a dense aggregation of very active sperma-
tozoa in the form of a ring within the margin of the original drop,
and a line extending from the drop to the edge where the pipette
was introduced and withdrawn. In this case the ring is open
below and a linear aggregation extends from the opening towards
the margin of the suspension. The ring and the linear aggrega-
tion are separated from the general sperm suspension by a clear
area devoid of spermatozoa 1.5 to 2 mm. in width. This area
534 FRANK R. LILLIE
belonged mainly to the original territory of the sperm suspension,
and the ring owes its origin to migration of spermatozoa towards
the drop. They do not, however, penetrate at first to the center
of the drop, but their movements are arrested, hence the formation
of thering. The ‘tail’ of the ring is due to migration of spermato-
zoa to the trail of CO. sea-water left behind in withdrawing the
pipette, leaving a clear zone marking the range of the effective
stimulus. Control: No reaction is given to a drop of pure sea-
water similarly introduced.
The migration of spermatozoa to the first formed ring continues
for a short time; the ring thus grows broader and tends to close
in the center (fig. 4-2 and 3). Shortly after the ring and tail
aggregations have formed with reference to the introduced drop
of CO, sea-water, the usual aggregations of the sperm, 1 to 2 mm.
in diameter, form in the remainder of the suspension outside the
drop evenly spaced throughout, if the sperm suspension is per-
fectly fresh (fig. 4-4). But if it is a little stale the general sus-
pension remains homogeneous.
The detail of form of the ring and tail aggregations vary accord-
ing to whether the introduced drop simply displaces a certain
amount of the suspension, or is more or less mixed in the introduc-
tion; and this depends obviously on the size of the opening of the
capillary pipette and the rate at which the drop is introduced.
But the general form of the reaction is always the same.
The spermatozoa in the CO, aggregations are never in the least
agglutinated and their behavior is in all essential respects the same
as in the aggregations formed in any fresh suspension. I there-
fore early formed the hypothesis that the aggregation phenomenon
is a chemotactic reaction to CO: produced by the spermatozoa
themselves, and this hypothesis has been abundantly confirmed,
as the series of experiments to be described will show.
The formation of the described configuration in a suspension of
active spermatozoa with reference to an introduced drop is due
to positive chemotaxis to the drop. If the clear margin be ob-
served during the formation of the ring, the spermatozoa may be
seen swimming across it to the ring head first. Under the low
power of the microscope they appear to drift across it with a
STUDIES OF FERTILIZATION 535
dancing motion like motes in a sunbeam owing to their spiral
path. Ifthe external edge of the clear zone be carefully observed,
the spermatozoa can be seen to detach themselves one by one
from the general suspension and pass straight over to the ring.
But only those freely suspended make the direct path; those in
contact with the slide or cover continue their circus movements;
the chemotactic stimulus seems unable to overcome the thigmo-
tactic reaction.
The reaction is given most clearly and rapidly by a fresh sperm
suspension, although one which has passed the aggregation stage
still gives it; however, as the sperm suspension becomes stale
the reaction becomes slower, and eventually ceases. Spermato-
zoa killed by gentle heat give no such reaction, thus excluding any
purely physical diffusion effect as cause of the phenomenon.
In the case of this reaction in a somewhat stale non-aggregative
suspension the movements of the spermatozoa on the outer mar-
gin of the ring are decidedly more vigorous than in the general
suspension. ‘This would appear to indicate that, at this place in
the CO, gradient marked by the clear zone, the concentration of the
CO, is stimulating rather than depressing; but when we consider
that the CO. gradient must rise from the suspension across the
clear zone to the ring, and that the relative inactivity of the sperm
in the suspension is due, partially at least, to CO. the conclusion
is not so clear. In any event, if we attribute a stimulating action
to a given CO, concentration on such evidence, we must regard
the depression of activity in the general suspension as due partly
to other excreta.
The conditions established by the experiment may be repre-
sented diagrammatically as follows (fig. 5). The injected drop
is represented by the continuous line circle and continuation, the
general suspension by the shaded area. By diffusion from the
injected drop a CO: gradient is established outwards, and this must
extend into the drop a certain distance because the gradient is
established by loss of CO: from the drop. The concentric broken
lines represent the gradient, or at least that part of the gradient
which is affective in the reaction. The thick open circle and the
similar linear extension represent the aggregations of the sperma-
536 FRANK R. LILLIE
tozoa. At the same time there is of course diffusion of substances
peculiar to the general suspension towards the introduced drop;
but that conditions thus arising are ineffective is shown by the
fact that no reaction is given to the introduced drop of pure sea-
water. We may, therefore, leave this centripetal diffusion out of
account. It should be remembered that 1 per cent CO, sea-water
is the minimum paralyzing strength for Nereis sperm.
Fig.5 Diagram of the reaction of asperm-suspension of Nereis to an introduced
drop of 1 per cent CO, sea-water; explanation in the text.
The diagram therefore shows, that migration of the spermato-
zoa proceeds up the gradient to, or near to, the point of paralysis
of the spermatozoa; for in the case of the drop of 1 per cent CO,
sea-water the ring forms well within the original margin of the
drop. With higher and lower dilutions of CO, the width of the
clear margin is practically the same.
The effects of greater and less CO, concentration than the
1 per cent used in the initial experiment are interesting. In general
the use of a greater concentration involves a larger aggregation,
and of a less concentration a smaller aggregation. Thus if a
drop of sea-water saturated with CO, be introduced into a sus-
pension of fresh sperm beneath a raised cover slip a border of
dead or paralyzed sperm forms at its margin, and shortly a clear
zone forms external to it; the spermatozoa migrate in large
STUDIES OF FERTILIZATION 537
numbers across the clear margin to the ring and are speedily
paralyzed by the diffusing CO2; this process continues very rapidly
and as a consequence the central aggregation expands, and may in
time absorb all the spermatozoa of the suspension.
The same phenomenon in less pronounced form is exhibited by
the reaction to a 1/10 dilution of the saturated CO. sea-water.
Here we may give a definite experiment with measurements of
growth of the aggregation: July 1, 1912. Into a fresh sperm sus-
pension beneath a raised cover a drop each, (a) of 1 per cent, and
(b) of 10 per cent CO, sea-water was injected some distance apart.
_ Drop (a) measured 5 mm. in diameter and drop (b) 3 mm. immedi-
ately after injection at 2.28 p.m. The aggregations caused by
these drops measured at 2.32: (a) 3 mm. (b) 5mm. That is to
say, the aggregations formed inside the drop in case of the weaker
solution, and outside in the case of the stronger. At 2.35 (a)
still measured 3 mm. and (b) now 6mm. In a repetition of this
experiment drops (a) and (b) each measured 3 mm. at 2.37. The
aggregations caused by them measured at 2.40, (a) 2 mm., (b) 5
mm.; at 2.47, (a) 2 mm., (b) 10 mm.
Th is clear from these ese ns that the spermatozoa are
positively chemotactic in a CO, gradient where the tension is
above a certain point, and that the aggregation caused by the
more concentrated drop grows because the diffusion of CO, from
the center forms a widening ring of the necessary concentration.
To furnish a gradient the concentration must exceed the CO;
tension in the general suspension, which is a function of the age
of the sperm suspension and the activity of the spermatozoa in it,
and on the other hand a limit is set to the differential which fur-
nishes the reaction by the fact that a concentration of about
1 per cent of the saturated CO, sea-water paralyzes the spermato-
zoa. The gradient that furnishes the reaction must, therefore,
operate within very narrow limits.
Greater dilutions of the CO. sea-water than 1/100 will act
positively in the case of fresh sperm suspensions. The ring forms
within the margin of the drop in the case of 1/200 dilution, but
it remains narrow, and tends to break into bead-like aggregations,
proving that the spermatozoa by their own activity have produced
538 FRANK R. LILLIE
a greater CO. concentration within the ring, which furnishes
centers of aggregation positive to the 1/200 concentration of the
drop. Drops below this concentration or drops of pure sea-water
furnish no reaction. °
Summarizing; it would appear that the spermatozoa of Nereis
follow a CO, gradient to the point of paralysis (about 1/100 sat-
uration). The clear zone outside an aggregation represents the
effective CO. gradient in every case. The various forms of
reaction to drops of different concentrations follow from this
simple principle.
2. Interpretation of the aggregation reaction
We are now prepared for the interpretation of the aggregation
phenomena exhibited by fresh sperm suspensions described on
page 519. The spermatozoa as they come from the body cavity
are absolutely quiescent; as soon as they are suspended in sea-
water they become intensively active, and consequently produce
CO, very rapidly. Any area of greater concentration of sperma-
tozoa, by producing more CO, than other areas, becomes a cen-
ter of attraction, and aggregations of the spermatozoa once begun
are bound to proceed to the limit, because the closer the aggre-
gation the greater the CO: production and consequently the
greater the chemotactic stimulation. If aggregations once formed
are broken up and the spermatozoa evenly suspended once more,
the CO, tension in the suspension is greater than at first and is
evenly distributed. Hence, in the first place the activity of the
spermatozoa is reduced, and, in the second place, the differential
of the gradient between that of the general suspension and the
point of paralysis is greatly lessened. Therefore aggregation
takes place more slowly and less completely than before; and,
after a second and a third stirring up, the CO, tension in the entire
suspension has become too great to permit of sufficient activity
to react to the slight possible differential gradient.
It is obvious that such a reaction can take place only in the
case of spermatozoa that exhibit extreme sensitiveness to COs.
The spermatozoa of Nereis possess by far the greatest sensitive-
ness to CO, of any studied, as we have already seen. No other
STUDIES OF FERTILIZATION 539
spermatozoa exhibit the aggregation reaction in any marked
form so far as I know; Arbacia which comes next to Nereis in
point of sensitiveness to CO: among the forms studied, shows,
under certain conditions, a cloud-like formation similar to the
initial stage of aggregation in Nereis. These will be referred to
beyond.
It may perhaps be objected that the aggregation reaction in
Nereis is not necessarily caused by CO, excretion, but possibly by
some other substance produced by the spermatozoa. And it
would be difficult to meet this objection in any absolutely con-
clusive way. But the following considerations render the con-
clusion extremely probable. In the first place it can be proved
that the spermatozoa exhibit positive chemotaxis towards some
substance that they themselves produce. Thus July 31, 1912,
the following experiment was made:
With the dry sperm of one individual two suspensions were made
at the same time (8.57 A.M.) namely: a one drop of sperm in 9 cc. sea-
water, b two drops of sperm in 6 cc. sea-water; the activity of the sperm
in both suspensions being the same b should produce any attractive sub-
stance in much greater amount than a. This was tested by making
preparations of a and b on separate slides beneath raised covers. A drop
of b was then injected into slide a, and a drop of a into slide b at 8.59.
On slide a there was a very quick beautiful positive reaction to the intro-
duced drop, that is a clear border formed about the drop owing to positive
chemotaxis of the spermatozoa a to the drop of denser suspension b.
On slide 6 not only was such positive reaction absent, but the drop of
introduced sperm actually lost its spermatozoa and became clearer,
owing to the positive chemotaxis now being away from the drop. The
same results were obtained also by using two suspensions of equal den-
sity, one of which was older than the other. The fresher suspension
reacted positively to the older suspension.
In the second place, it is of course certain that the spermatozoa
becoming suddenly active in the sea-water must produce CO,;
and as we have seen that the spermatozoa of Nereis react even to
a 1/200 dilution of a saturated solution of CO, in sea-water, if it
can be proved that a standard suspension of spermatozoa produces
an equivalent amount, the probability that CO. is the agent
involved in the aggregation effect become very great. Experi-
ments directed to this end showed that a 1/100 dilution of CO,
540 FRANK R. LILLIE
lies near the limit of demonstrability both by color reaction tests
and also by gas-burette estimation. A number of tests of sperm
suspensions were made with acid color indicators. In the case of
neutral red a dilute solution in sea-water has a decided orange
tinge due to slight normal alkalinity of the sea-water. The same
dilution made from a standard concentrated solution by the addi-
tion of a sperm suspension, shows a decided rose color without any
trace of orange. The spermatozoa then aggregated in the vial
used and the aggregations sank to the bottom, forming a bright
red precipitate, and the supernatant fluid, now merely opalescent
on account of the few sperms remaining in it, was faint rose.
There is thus a decided acid reaction of the sperm suspension.
Tests with azolitmin and tropaeolin 000 No. 1 also gave clear indi-
cations of acid. The sperm suspensions were tested within two
minutes or less after their preparation; the liberation of the acid
takes place therefore very suddenly. It is liberated only when
the sperm become active, and the change of color is not given if
the sperm remain inactive. It is therefore very probable that CO,
is the acid revealed. :
Finally a large number of tests for CO. were made of the air in
closed flasks containing considerable quantities of active sperm
suspensions of Arbacia. The details of these tests made with a
gas burette need not be given. They extended over a week, using
the sperm of Arbacia which could be obtained in larger quantities
than Nereis. Although the determinations came very near the
limits of experimental error, there could be no question as to the
presence of CO, in quantities above that contained in the air
or in normal sea-water.
In consideration of the facts (1) that sperm suspensions of
Nereis produce a substance to which spermatozoa of Nereis
react positively (2) that an acid is present in the suspensions (3)
that the production of CO. by the suspensions can be demon-
strated and (4) that spermatozoa of Nereis react positively to
dilutions of CO, in sea-water which are barely detectable by
color indicator, or gas burette, it can hardly be questioned that
the aggregation reaction in Nereis is due to positive chemotaxis
to COs..
STUDIES OF FERTILIZATION 541
3. Reaction to other acids
The sperm of Nereis exhibits the same positive chemotaxis
to other acids as to CO».. It is hardly necessary therefore to
enter into details. Sulphuric, hydrochloric, nitric, and acetic
acids were tested. They agree very closely with respect to the
effective dilutions; N/1000 dilution of any of these acids is an
effective chemotacticum agreeing quite closely in the degree of
the effect produced with 1 per cent CO.; N/2000 may cause slight
ring formation in a drop introduced into a fresh sperm suspension,
but, if the suspension has reached the non-aggregative stage, no
reaction ensues, owing to the fact that the acid concentration in
question furnishes no gradient.
Drops of stronger concentration cause a ring-shaped aggrega-
tion which continues to grow until diffusion eliminates the acid
gradient. None of these acids cause the least sign of agglutina-
tion of the spermatozoa whatever their strength.
A drop of N/10 acid introduced within a sperm suspension
beneath a raised cover kills all the spermatozoa in its immediate
neighborhood, as the acid diffuses the zone of dead sperm increases
but as the margin of the diffusing acid reaches a dilution that is no
longer fatal it becomes marked by a clear border which is due to
the migration of sperm to it, even though they are carried into a
fatal concentration; and so the drop continues to grow so long as
an acid gradient remains.
There is never the least sign of negative chemotaxis with respect
to any concentration of any acid, nor indeed of any other agent
tested. This being the case it is obvious that the aggregation of
the spermatozoa can not be by any trial and error method of
behavior, but must take place through orientation.
4. Behavior with reference to alkalis
The spermatozoa of Nereis do not exhibit any chemotactic
reaction, positive or negative, to drops of KOH or NaOH injected
into a suspension beneath a raised cover-slip. The drop remains
empty at first, and spermatozoa that enter it by chance are
agglutinated, so that in a short time the drop becomes filled
542 FRANK R. LILLIE
with agglutinated sperm masses. The alkalis were used in
concentrations of N/2500 and N /5000.
Though no chemotatic reaction takes place, yet an interesting
reaction may be procured by injecting a drop of N/5000 KOH into
a sperm suspension made in 1/100 dilution of saturated CO,
sea-water. The sperm in this are nearly or quite paralyzed, and
will not of course react to an injected drop of 1/100 CO:. But if
a drop of N/5000 KOH be injected into such a suspension beneath
a raised cover glass, motility of the spermatozoa returns at a short
distance from the margin of the introduced drop, due evidently
to neutralization of the acid, and aggregations of the active sperm
may form outlining the KOH drop a certain distance from the
margin. ‘This experiment furnishes at once an interesting demon-
stration of recovery from CO, narcosis, and of the nature of the
contrast between the acid and alkali reaction of the spermatozoa.
5. Reactions to other substances
It lay entirely outside of the scope of this work to attempt an
exhaustive analysis of the behavior of the spermatozoa with
reference to chemical substances. The investigation was under-
taken to analyze the relations of the behavior of spermatozoa
to fertilization; and when the principle of chemotaxis was once
demonstrated, and the relation of this chemotactic reaction to the
very striking aggregation reaction, it seemed better to turn at
once to the subject of the behavior of the spermatozoa towards
eggs and egg-secretions of the same species. It may be remarked
incidentally that tests with alcohol and ether gave negative
results, that is, no evidence of positive or negative chemotaxis
was found. In some cases there was slight indication of ring
formation near the margin of an introduced drop of 5 per cent
alcohol or ether in sea-water, which may indicate a stimulating
effect of the spermatozoa at an optimum concentration, but which,
in the absence of a clear margin external to the ring cannot indi-
cate chemotaxis.
STUDIES OF FERTILIZATION 543
6. Thermotaxis
The spermatozoa of Nereis do not exhibit any positive response
to drops of sea-water at higher temperatures. Into suspensions
of spermatozoa in sea-water at 21°C. under raised cover-slips
drops of sea-water at 44°C., 52°C., and 84°C. were injected succes-
sively. No ring formation occurred with reference to any of these
drops. They did not fill up with sperm as rapidly as drops of
sea-water at room temperature, but this is no doubt due to the
paralysis that sets in, as previously noted above about 28°C.
The only observable effect of the heated drops was that aggrega-
tions formed a little earlier in the general suspension near the
margins of the introduced drop; and this is attributable to in-
creased activity of the sperm owing to rise of temperature, hence
increased CO, production in the zone of increased activity; and
more rapid aggregation as a consequence.
7. Thigmotaxis
In contact with any solid object Nereis spermatozoa tend to
carry out circus movements in an anti-clockwise direction, when,
fresh, but may soon come to rest. In any field of the microscope
in a suspension beneath a cover glass one sees many of the sper-
matozoa in contact with the slide at rest, and many others carry-
ing out the circus movements, while those that are freely sus-
pended swim in spiral paths. The thigmotatic reaction then
appears first to be the exaggeration of the rotation component of
the ordinary locomotor movements, and second rest.
This reaction may of course come in conflict with the chemo-
tactic reaction, as for instance in the clear margin external to the
ring of spermatozoa produced in response to a drop of 1/100 CO,
sea-water. Within this area all the freely suspended spermatozoa
swim directly towards the ring, but those in contact with slide or
cover-slip may continue their circus movements without any
apparent directive effect from the CO, gradient. The thigmo-
tactic stimulus appears thus to be more effective than the CO,
gradient.
544 FRANK R. LILLIE
This thigmotactic reaction may be the starting point apparently
of some of the aggregations formed in suspensions. Thus aggrega-
tions tend to form in the angle between the glass rod and the slide,
which appear to owe their origin to the thigmotactic reaction;
but when a considerable number of spermatozoa have accumulated
in the angle their CO, excretion acts as a positive chemotactic
stimulus on the sperm of the suspension, and a dense swarm soon
forms along the rod, filling the angle and extending beyond it.
Such a continuous swarm then tends to break into evenly spaced
masses still in contact with the rod, owing to variations in CO,
production. Thus thigmotaxis in this case is the initial cause of
aggregations, which owe their subsequent growth to chemotaxis.
It may be that the thigmotactic reaction is a frequent cause of
aggregations, particularly in suspensions that have produced
considerable CO, when aggregations form only slowly and always
in contact with the substratum. But in fresh suspensions this
cannot be the case, for the aggregations first formed are freely
suspended. In many cases aggregations may be seen to form with
reference to firm strands or fibers of mucus in a suspension, and
in such cases it appears probable that thigmotaxis and chemo-
taxis are combined.
8. Variations of reactions
So far as observed the behavior of sperm suspensions in sea-
water may be quite fully explained by the forms of reaction
described, and this brings the present section to a natural conclu-
sion. But we may finally note certain variations of the reactions.
The sperm suspensions were usually made, as stated, by mixing
a drop or two of dry sperm with about 8 to 10 ce. of sea-water in a
Syracuse watch crystal. This was done with a pipette, drawing in
the suspension and squirting it out again until the sperm was
evenly mixed. If this is done from one side, as is usually the case,
a current is made across the dish to the opposite side and back
along both sides, creating miniature whirlpools. Such currents
of course come to rest in a few seconds, but when the aggregations
become visible, ten to forty or more seconds later, they define
very accurately the original currents. At first I thought natur-
STUDIES OF FERTILIZATION 545
ally of some rheotactie reaction. But on more careful examina-
tion and consideration the following explanation appears much
more probable. Microscopic aggregations must begin to form
while the water currents are still moving; they are then elongated
by the friction in the direction of the current, and as they grow
to macroscopic size the aggregations tend to preserve this form.
The definition of the currents is due to the form of the aggrega-
tions rather than to their arrangement, and as they contract to
spherical form the current-figures become less pronounced and
very largely disappear.
Very interesting configurations may be produced in a sperm
suspension of Nereis by dropping in dilute acids. Ina few seconds
quite complex wreath-like or festooned aggregations of sperma-
tozoa appear at the site of the entering drop marking out accur-
ately the distribution of the acid in the suspension. These of
course vary with the strength of the acid, and the distance from
which it is dropped.
If a few drops of a suspension of active Nereis sperm be mounted
beneath a raised cover slip, it will be observed that the outer
margin of the suspension for a width of 1 to 2 mm. soon becomes
free from spermatozoa, thus tending to concentrate the suspen-
sion the same distance from the margin (fig. 4). This concen-
trated ring of the suspension then tends to form aggregations more
rapidly than the more central parts. In the case of a suspension
that is not perfectly fresh, aggregations may form only in this
ring. The withdrawal of spermatozoa from the margin of the
drop might at first thought be attributed to a negative chemotaxis
towards oxygen. However, it is almost certainly not this, but
a positive reaction towards the higher CO, tension of the interior
of the drop. If a drop of sea-water saturated with oxygen be
injected into a suspension beneath a raised cover, the spermatozoa
avoid it in the same way that they do the free margin of the
suspension.
The spermatozoa of Nereis make an acid indicator more deli-
cate than any of the chemical dye indicators. In the course of
some experiments I discovered quite accidentally, thus avoiding
an awkward mistake, that the first few drops of water through any
546 FRANK R. LILLIE
available filter paper give a distinct acid test, using the spermato-
zoa of Nereis as indicator: A drop injected into a fresh sperm sus-
pension invariably gave the ring formation with outer clear border,
which is the characteristic and unmistakable acid reaction. This
suggests the possible use of such cells as indicators in certain
classes of experiments; some preliminary observations which I
have made concerning CO, production of dividing eggs by this
method are distinctly promising, though the results are compli-
cated by the usual presence of other substances.
9. Chemotaxis to egg-secretions
The spermatozoa of Nereis exhibit positive chemotaxis to egg-
secretions, which may be demonstrated in the same way as the
positive chemotaxis to acids, but this subject, which is of course
the most important part with reference to the fertilization
problem, may be postponed to the next section dealing with
agglutination phenomena, because it is always associated with
agglutination.
B. ARBACIA
The reactions of Arbacia spermatozoa are essentially the same
in principle as those of Nereis, but on account of the lesser sensi-
tiveness of the spermatozoa, as noted in the section on activation
phenomena, the reactions are much slower and less clearly defined.
This may be illustrated by the notes on a single experiment:
July 20, 1912. <A fresh suspension of Arbacia sperm was made by
mixing two drops of the dry sperm with 9 ee. of sea-water. The
suspension appears milky, and the spermatozoa are decidedly
active under the microscope. <A portion was immediately moun-
ted under a raised cover-slip and two drops of 5 per cent CO,
sea-water were injected some distance apart. At first there
appeared to be no reaction, as contrasted with Nereis in which
ring formation is almost instantaneous under such circumstances.
In two minutes the sites of the drops beeame more cloudy than
the rest of the slide, and a faintly defined clear margin began to
appear surrounding the drop. The picture gradually gained in
definiteness until it became very clear. The central aggrega-
e
STUDIES OF FERTILIZATION 547
tions at first showed radiating strands of sperm visible to the naked
eye, but then closed to form a solid drop, and this grew at its
margins, preserving the clear external zone, until the clear margins
of the two drops, at first several millimeters apart, ran together.
Outside of the influence of the drops cloud formations appeared,
corresponding to an early stage of the aggregations of Nereis.
The clear margin of the drops is not by any means so well
defined as in Nereis. Moreover, the spermatozoa are so small
that it is difficult to observe their behavior in the clear margin.
However, there can be no doubt that the phenomenon is essen-
tially the same as in Nereis, and that the aggregation in the drop,
the appearance of the clear margin, and the growth of the aggrega-
tion are due to positive chemotaxis to COs.
This reaction is given clearly only by a fresh sperm suspension.
One ten minutes old does not give it, owing presumably to for-
mation of CO, in the suspension.
A considerable number of tests were made. In some the reac-
tion was much more rapid than in the experiment described.
In one of these tests I injected drops of 20 per cent, 4 per cent and
1 per cent of the CO, sea-water near together. In the case of the
20 per cent a ring with external clear margin was formed in a few
seconds. The ring did not close. The 4 per cent formed a ring
which closed in its center. There was no reaction to the 1 per
cent. The same suspension gave no reaction twenty-five minutes
after it was mixed. In another case I got a faint reaction to
1 per cent CO, sea-water.
The fact of positive chemotaxis of Arbacia sperm to CO,
dissolved in the sea-water was repeatedly demonstrated. In
the case of a drop mounted beneath a raised eover it expresses
itself by a gradual aggregation of the sperm towards the center,
leaving the margins clear.
As is to be expected from the slower and less delicate reaction
to CO;, as compared with Nereis, spermatozoa of Arbacia react
also to other acids, but more slowly and not to so great dilutions.
Thus in tests of N/10, N/50, N/250 and N/1000 H.SO,, strong
positive reactions were obtained for the first three, whereas only
a faint shadowy reaction is given to N/1000. In the case of Nereis
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, NO. 4
548 FRANK R. LILLIE
it will be recalled that N/2000 gives a distinct reaction. Posi-
tive reactions were not secured with N/1000 HCl or HNO,,
but were with N/250. The ring formed in response to N/250
H.SO, grows somewhat owing to diffusion of the acid; in the case
of N/50 and N/10 H.SO, the growth is much greater, in general
in proportion to concentrations.
As in Nereis alkalis agglutinate, but cause no aggregation.
KOH was tested in two ways: (1) addition of a small quantity of
N/2000 KOH in sea-water to a vial of fresh sperm suspension
caused macroscopic agglutinations which are irreversible. (2) a
drop of N/2000 KOH injected into a suspension beneath a raised
cover fills in a short time with sperms that agglutinate; but there.
is no chemotaxis, and in a short while the drop is left to one side
by aggregation of the sperm away from it, owing to rise of CO.
tension elsewhere. |
Reaction to egg-secretions
As contrasted with the slowness of reaction of Arbaciasperma-
tozoa to acids, the reaction to egg-secretions is instantaneous
and clear cut. There is a most pronounced positive chemo-
taxis, as tested with drops even of very weak egg-extract injected
into a sperm suspension mounted beneath a raised cover-slip;
but this is always associated with agglutination, and is, therefore,
best considered under that head.
IV. AGGLUTINATION PHENOMENA AND REACTIONS OF
SPERMATOZOA TO EGG-SECRETIONS
1. INTRODUCTION
Following the demonstration of definite chemotactic behavior
of spermatozoa of Nereis and Arbacia the question naturally
arises what relation, if any, has this form of behavior to the union
of egg and spermatozoon in fertilization. As is well known, the
mere observation of the interaction of the sexual elements in
fertilization led long ago to the theory that the egg attracts the
spermatozoon to itself by chemotaxis, and fusion results from
active penetration of the ovum by the spermatozoon. But in
recent years there has been a tendency to deny both of these .
STUDIES OF FERTILIZATION 549
principles as factors in the union. Chemotaxis has fallen into
disrepute; and the theory that the spermatozoon bores into the
egg has been rejected by several observers.
If chemotaxis is concerned in the union of ovum and spermato-
zoon the medium in which fertilization operates must contain the
substance concerned. In the case of the eggs of Nereis and Arba-
cia, therefore, the hypothetical substance which attracts the
spermatozoa must exist in sea-water which has been in contact
with fertilizable eggs; and it must be possible to obtain a sufficient
concentration of the substance in question in sea-water to demon-
strate its presence by reaction of the spermatozoa, because, ex
-hyp., the substance exists in effective amounts in the sea-water
surrounding the eggs. If it were impossible to demonstrate the
presence of an agent to which spermatozoa of the same species are
positively chemotactic by such means the theory of chemotaxis
would have to be abandoned. However, the presence of such a
substance is readily demonstrated both in Nereis and Arbacia.
In the second place, if the union of the ovum and spermatozoon
after they have come in contact operates not mechanically, but
through some bio-chemical reaction between spermatozoon and
ovum, the sea-water in which eggs have been standing should
contain a substance also capable of reaction with the sperm, which
should be an efficient indicator for it.
I was guided by some such ideas as these in the series of experi-
ments which follow, and which showed at the very first trials that
sea-water which has stood with fertilizable eggs of Nereis or
Arbacia contains a substance to which the spermatozoa of the
same species are positively chemotactic, and also a substance
which agglutinates the spermatozoa of its own species. It may
be that one substance is concerned in both reactions, but it is
more probable that two are present. It is perhaps worth empha-
sizing here, for this is the fact that struck me at the start, that the
sea-water which has stood over eggs* combines both the effects of
3 To avoid the frequent repetition of such a circumlocution we may call sea-
water, which has contained eggs and is charged with their emanation, egg sea-
water; and the concentration of the substance in the sea-water may be expressed
by writing the relative bulks of eggs and sea-water as a fraction. Thus ‘egg sea-
water 1/3’ would indicate that the bulk of eggs was one-third the volume of the
sea-water. Time is also a factor in the concentration, of course.
550 FRANK R. LILLIE
an acid (aggregation) and also an alkali (agglutination) on the
spermatozoa. The comparison may, of course, be superficial,
but it serves at least to emphasize the double action of the egg-
secretion.
As contrasted with the difference in rapidity and delicacy of
reaction between the spermatozoa of Nereis and those of Arbacia
to inorganic substances, we may note in advance that the reactions
to the egg extractives are as rapid and clear in the one case as in
the other, and are entirely similar in principle, though there are
certain secondary differences that will be noted in the proper
place.
2. INITIAL EXPERIMENT
We may begin by describing the reactions to be observed in
the case of an Arbacia sperm-suspension freshly made and
mounted beneath a raised cover-slip, into which a drop of Arbacia
egg sea-water 1/10 to 1/20 about half-an-hour-old is injected.
The naked eye observation shows almost instaneous formation
of a ring at the margin of the drop, with simultaneous formation
of a clear external zone about 1.5 to 2 mm. wide; the ring then
breaks up into small agglutinated masses and so becomes beaded.
The trail of substance left in withdrawing the pipette extends to
the margin of the cover-slip. It also is a center of attraction and
the ring is therefore prolonged by a chain of agglutinated masses
to the margin.
One can observe the details of the reaction best under the micro-
scope, using a low power, by bringing the point of the pipette into
the field of the microscope and blowing in the drop with the aid of
a flexible rubber tube held in the mouth, while looking through the
microscope. The reaction takes place so rapidly that it requires
repeated observations to observe all the details. In the first
second the spermatozoa are aroused to intense activity and form
small agglutinated masses within the drop; these then appear.
actually to ‘rush’ together (to use the language of my note book)
to form larger agglutinations for a period of three to five seconds,
after which no more fusion of masses takes place. The aggluti-
nated masses thus range from relatively large to relatively small.
While this has been going on in the interior of the drop, a ring
STUDIES OF FERTILIZATION 5b1
has formed at the margin, and a clear zone arises external to it.
The ring is at first continuous, but it ruptures in numerous places
in two or three seconds and each segment contracts quickly to an
agglutination mass.
The agglutinated masses in the interior of the drop are smaller
than in the ring, owing to the relatively low concentration of the
sperm suspension within the drop, and they break up very quickly
while the sperm is still extremely active. The movements of
the spermatozoa then gradually slacken; in a few minutes the
larger and more firmly agglutinated masses of the ring also begin
to break up and in ten or fifteen minutes all are resolved.
This preliminary observation demonstrates a three-fold action
of the egg-extractive: (1) it activates the spermatozoa; (2) it
aggregates them through positive chemotaxis; (3) itagglutinates
them. The phase of increase of activity lasts only a short time,
a minute or two at the most, after which movements of thesperm
slacken and become less than the control, or cease entirely. Posi-
tive chemotaxis (aggregation) is shown by formation of a clear
zone external to the marginal ring. This is always a sign of
chemotaxis, as we have seen in the preceding section.
The agglutination phenomenon is fundamentally different
from the aggregation; in the latter the spermatozoa are merely
loosely associated, and slight agitation is sufficient to scatter them.
In the agglutinated masses the spermatozoa are stuck together
and are not separated by shaking. In the case of Nereis where
the agglutination is firmer than in Arbacia the masses may be
broken up into smaller masses by needles, or preserved en masse
-in killing fluids. The breaking up of the ring into separate masses
is a characteristic agglutination effect ; the rings formed in response
to an acid do not break up unless the acid is very weak (see p. 537).
Finally an agglutinative substance produces its effect when
shaken up and evenly distributed in a vial of sperm suspension,
but an aggregative substance cannot of course exert a chemotactic
effect in the absence of a gradient.
The same experiment succeeds well with Nereis. The eggs of
this form give off a substance (or substances) into the sea-water,
which causes aggregation and agglutination of the spermatozoa
552 FRANK R. LILLIE
when a drop of sea-water so charged is injected into a fresh sperm
suspension beneath a raised cover slip. The activation is not
so pronounced in this case as in Arbacia. The aggregation phe-
nomenon is the same. The agglutinations are substantially
permanent in Nereis; the spermatozoa stick together much more
firmly.
In what follows we may leave the activation and aggregation
phenomena out of account for the most part and confine ourselves
to the problems of agglutination. The substance which causes
agglutination of the spermatozoa we shall call the sperm agglu-
tinin. The agglutination may be shown very strikingly in a
vial of fresh sperm suspension. In the case of Arbacia the
addition of two or three drops of egg sea-water 1 /4, which has stood
half-an-hour, to about 2 ec. of a fresh milky sperm suspension
causes formation of agglutinations 1 to 2 mm. in diameter in a
few seconds. The agglutination may be so strong that the
fluid between the white agglutinated masses appears perfectly
clear. The masses gradually fade from view in a few minutes,
but microscopic agglutinations may remain half-an-hour or
more.
The degree of agglutination is of course dependent on the den-
sity of the suspension. This is shown by the following experi-
ment: Ovaries and eggs of Arbacia were cut up in about four times
their own bulk of sea-water and allowed to settle. Two cubic
centimeters of the supernatant fluid was put in each of three vials.
To one was added 3 drops of a fresh milky sperm suspension, to
the next 12 drops of the same, to the third 36 drops. No visible
agglutinations formed in the first; in the second agglutinations
became visible to the naked eye almost immediately, in the third
agglutinations were larger, more numerous, and apparently more
solid. ’
3. OVA ALONE PRODUCE THE AGGLUTINATING SUBSTANCE
The eggs of both forms thus produce an agglutinin in the sea-
water. The next question is whether the agglutinin is specifically
an egg-product. A considerable number of experiments prove
that this is the case. The large body cavity of Arbaciais filled
with abundant coelomic fluid and this may be supposed to
STUDIES OF FERTILIZATION 5S
contain substances from various tissues. But it invariably
proved perfectly neutral to spermatozoa of Arbacia, even when
taken from females with large ovaries, showing that the substance
concerned in agglutination does not escape from the ovaries of
the intact animal, or, if its does, that it is promptly destroyed.
Nor was it possible to extract a sperm agglutinating substance
for Arbacia by extracting the intestine in sea- or fresh-water.
These experiments were repeated a sufficient number of times to
be conclusive.
As illustrations: (1) Intestine extractives. August 31, 1912. The
intestines of several Arbacia were cut up in about twice their bulk of
‘distilled water, and were allowed to stand in it about an hour. The
strongly amber-colored fluid was filtered off and rendered isotonic with
the sea-water by addition of concentrated sea-water (four parts of the
latter to six of the intestine extract). This fluid causes no agglutination
in sperm suspensions. A similar extract of the ovaries caused immediate
large dense agglutination masses. (2) Coelomic fluid: As is well known,
the coelomic fluid of Arbacia contains large numbers of densely pigmen-
ted corpuscles. Outside the body the fluid quickly forms a loose clot
which includes many of the corpuscles. The others can be separated
from the remaining serum by centrifuging. The corpuscle-free serum
was sometimes used, sometimes simply the clot-free serum. Repeated
tests were made both by injecting drops into fresh sperm suspensions
beneath raised cover glasses, and also by mixing with fresh sperm
suspension in vials. Whether the coelomic fluid came from males or
females it proved invariably negative, except for the faintest sort of
agglutination reaction in one or two cases only, which may have indi-
cated some individual differences. It is interesting in view of these
facts that outside of the body the coelomic fluid becomes heavily
charged with the sperm agglutinin if eggs are placed in it. It must be
supposed therefore that the ovarian membrane is impermeable to the
agglutinating substance in the intact animal.
4 The experiments this year simply opened up the problems, and it was impossi-
ble to make any quantitative tests or adequate chemical examination. For the
purpose of the biological problem of the behavior of the spermatozoa with refer-
ence to the eggs the question of immediate importance was the behavior with
reference to egg-extractives or secretions in the sea-water tested. Stronger agglu-
tinating solutions were made by increasing the quantity of eggs with reference to
the sea-water, or by crushing the eggs in sea-water. Distilled water was shown to
extract more agglutinin from a given bulk of eggs than sea-water; and the coe-
lomic fluid of Arbacia also proved to be a better medium for extracting agglutinin
from the eggs than sea-water. It would be possible of course to establish quanti-
tative values for all of these relations, and this problem should receive attention.
The problems of solubility of the agglutinin in various media, and other chemical
questions, also present themselves.
554 FRANK R. LILLIE
In the case of Nereis it was not possible to reach such conelu-
sive results, because the sexually mature female is practically a
bag of eggs, and one cannot obtain other organs for testing. I
cut up five females that had shed their eggs in 10 cc. of sea-water.
In spite of efforts to get rid of all eggs, a considerable number
were in the water. For control I used the eggs of two females in
100 cc. sea-water. On test in half-an-hour the fluid from above
the eggs was found to be about ten times as agglutinative as the
fluid from the bodies of the spent females. So that it is certain
that other tissues do not produce much sperm agglutinin and it is
probable that they do not produce any. The small amount
present could be accounted for by the few eggs included, and per-
haps by egg secretions absorbed by the tissues.
4. FIXATION OF THE AGGLUTININ BY SPERMATOZOA
The next question was whether the agglutination reaction as
described has the usual characters of a chemical reaction? The
general result is (1) that an agglutinated sperm suspension in
which reversal has occurred is not capable of re-agglutination by
addition of more of the agglutinating substance and (2) that
the agglutinating substance disappears from an agglutinated
suspension if not present originally in excess.
As regards the first point, the earlier experiments were con-
cerned entirely with the form and conditions of the reaction, and
the agglutinating substance was always used, as later results
showed, in excess. It was not possible to get a repetition of the
agglutination reaction under these circumstances. But one can
get a repetition of the reaction in a sperm suspension by addition
of successive small amounts of the agglutinating substance, until
the reaction is complete, as the following experiment shows:
September 4, 1912. Arbacia. 2 cc. of a creamy active sperm suspension
was agglutinated with 5 drops of an egg-extract prepared as follows: The
ovaries of three females were cut up in about three times their volume of
distilled water and allowed to stand about thirty minutes. Then the
water was filtered off and made isotonic with sea-water by the addition
of concentrated sea-water (proportions of 58 to 42 parts). This made a
very strong agglutinating extract. After reversal of the agglutination
described above the agglutination was repeated by addition of a drop of
STUDIES OF FERTILIZATION 555
another egg-extract. This time the agglutination was complete for it
could not be repeated a third time. Other sperm suspensions gave
similar results.
The result might be interpreted as a purely biological reaction,
that is to say in terms of stimulation, were it not for the fact that
the agglutinating substance disappears from an agglutinated
sperm suspension, as shown by the following experiment:
September 12, 1912. Arbacia. Nine parts of a thick activesperm
suspension was agglutinated by one part egg-extract. The agglutina-
tion produced was so strong that the fluid between the white masses
appeared clear tothe eye. In three or four minutes reversal of the agglu-
' tinations had begun. The agglutinated sperm suspension was then
centrifuged until practically all the sperm was precipitated. The super-
natant fluid was tested and agglutinin was shown to be absent. As
control, a dilution of one part of the same egg-extract with 9 parts of
sea-water was tested with the same sperm and proved to be strongly
agglutinative. Three tests were made with each with uniform results.
There can be no doubt, as the result of this and other observa-
tions also, that the spermatozoa fix in some way the agglutinating
substance, and it will be simplest to assume as a working hypothesis
that the fixation is due to chemical union. I have not yet had the
opportunity to ascertain if the agglutinin could be regained from
the sperm precipitated in the centrifuge.
5. NATURE OF THE EFFECT ON THE SPERM
We have noted four effects of the egg-extracts on sperm of the
same species, namely: (1) Stimulation of intense activity, which is
of brief duration. This is more marked in the ease of the sperm of
Arbacia, than in the case of the naturally extremely active sperm
of Nereis; (2) An orienting effect expressed in positive chemotaxis;
(3) An agglutinating action; (4) Following these effects more or
less complete paralysis of the sperm.
To what kind of change in the individual spermatozoa is the
agglutination reaction due? We may note in the first place that
the agglutination is between the heads of the spermatozoa, and
that the tails are apparently unaffected, at least at first; it is only
in later stages of the action of the agglutinating substance that
the locomotor function is injured. The adhesion of the heads
556 FRANK R. LILLIE
demonstrates some change in the membrane that renders them
sticky. The cells are so minute that it is difficult to observe
any microscopical change in the case of Arbacia; in Nereis the
spermatozoa are larger, and it can be seen that in agglutinated
masses the heads of many of the spermatozoa are swollen into
spherical form and have lost the normal strong refringibility.
The change is in this case a very characteristic one, indicating a
great increase in permeability. The spermatozoa which have
undergone this change are usually motionless, and, when not
fused with one another, appear to be glued to the slide or cover
slip, never freely suspended.
Agelutination in itself is in no sense a specific reaction, butone
that may be expected to accompany certain superficial changes
of the spermatozoa, however caused, under conditions that bring
the spermatozoa into contact. It occurs, to a limited extent,
spontaneously in sperm suspensions that have stood for some time.
It is particularly noticeable in Nereis under the following con-
ditions: A fresh sperm suspension is allowed to aggregate on a
slide beneath a raised cover slip and the aggregations remain
undisturbed. In the course of ten or fifteen minutes small agglu-
tinated masses may form around the margins of theaggregations
or beneath the aggregations in contact with the slide. There
may be twenty to fifty or more such masses associated with a
single aggregation, and they are quite similar in their general
appearance, to those produced suddenly by the aeelutini of the
egg, though much smaller, on the average.
It is not probable that the agglutination is in any real sense toxic
or cytolytic. It is true that the agglutinin inhibits movement
after a few minutes, and it certainly lessens the fertilizing power of
the sperm. But if an agglutinated mass of spermatozoa of Nereis
be crushed under the microscope many of those liberated are
active. Moreover, if a small quantity of an agglutinated suspen-
sion of spermatozoa be added to a relatively large quantity of
sea-water fertilizing power is partially regained. This might be
either because some of the spermatozoa of the agglutinated sus-
pension had escaped combination with the agglutinin and were
alone concerned in the actual fertilization, or because of recovery
STUDIES OF FERTILIZATION 557
from the agglutination effect. It is very difficult to form a
definite opinion as to the real nature of the agglutination effect.
6. THERMO-RESISTANCE OF THE AGGLUTININ
The agglutinating agent is slowly destroyed at 95°C.:
August 28. Arbacia: (1) Ovaries and eggs of Arbacia were cut up in
three times their bulk of sea-water, and let stand about an hour. The
supernatant fluid is strongly agglutinative on Arbacia sperm suspensions.
(a) Part of it was now taken and boiled about thirty seconds, and cooled.
On test it proved as agglutinative as before. (b) Some more was then
boiled five minutes and cooled. Its agglutinative power was apparently
undiminished. (c) Another egg-extract similar to the first was then
boiled and put in a beaker of boiling water for thirty minutes; the tem-
perature stood about 95° during this process. On test its agglutinative
power was shown to be greatly diminished. (d) In a fourth test some
egg-extract was kept at 95° for sixty-six minutes. It still exhibited some
agglutinative power, which, however, was very slight as compared with
the control. The heated egg-extract exhibited a considerable change
of color from the yellowish red of the control to a much brighter red.
Simultaneously with the loss in agglutinating power, it appeared
also to gain in aggregating power: Drops of the heated and the
unheated egg-extract were injected into the same sperm extract
beneath a raised cover. The drop of the heated expanded its
sphere of influence shown by immigration of the sperm about
twice as rapidly as the unheated; this was tested several times;
so that it would appear that the aggregative and agglutinative
agents are probably distinct, and that the agglutinin inhibits
aggregation to a considerable extent.
Nereis: The eggs of three females were inseminated in 9 cc. sea-water;
the supernatant fluid has a slight amberish-green color, and is strongly
agglutinative on Nereis sperm. (a) After keeping at 95° C. for ten
minutés its agglutinative power was much reduced. (b) After twenty-
two minutes at 95°C. the agglutinin was entirely destroyed. The color
was entirely destroyed also.
Thus the agglutinating substances, whatever they may be,
are either volatile being gradually driven off by heating, or they
are slowly coagulated or disintegrated chemically by a temper-
ature of 95°C. The agglutinin of Nereis is either more volatile
or more labile than that of Arbacia. It is impossible to say defi-
nitely to what class of chemical substances these agglutinating
5d8 FRANK R. LILLIE
substances belong. It is, however, extremely improbable that
they possess a degree of chemical simplicity sufficient to allow of
volatilizing; it is more probable that they undergo slow chemical
disintegration at the temperature employed. The thermostabil-
ity of these sperm isoagglutinins is relatively very high, and this
perhaps makes it doubtful whether they can belong to the same
class of substances as the haem-agglutinins of vertebrateblood
sera. This matter must therefore remain undecided, and it
should be understood that the term agglutinin is used in the pres-
ent paper in a purely descriptive sense.°®
7. FERTILIZING POWER OF AGGLUTINATED SPERM
The powerful effect of the egg-extract on spermatozoa of the
same species may be shown by a complete loss of motility as
we have already seen, and also by a corresponding loss or diminu-
tion of the fertilizing power. The following experiments illustrate
this:
1. Arbacia. The egg-extract used was made by cutting up the ripe
ovaries in about three times their bulk of distilled water; in half-an-hour
the water was filtered off and was then made isotonic with sea-water by
the addition of 42 parts of condensed sea-water to 58 of the egg-extract.
Five small watch crystals in a series contained (1) 8 drops of the egg-ex-
tract (2) 4 drops egg-extract, + 4 drops sea-water (3) 2 drops egg-extract,
+ 6 drops sea-water, (4) 1 drop egg-extract, + 7 drops sea-water (5) 8
drops sea-water. ‘To each of these 3 drops of opalescent sperm suspen-
sion was added; and after twelve minutes a drop of a suspension of
fresh eggs was added to each. We thus had the same quantity of eggs
in sperm suspensions of the same density, but in graded amounts of
egg-extract. The sperm suspensions were so dense that in the control
(no. 5) the jelly became packed with sperm, forming dense halos around
the eggs. In (4) a very few (about 1 per cent) had slight halos of sperm
in the outer layer of the jelly, but in (1), (2) and (3) the paralysis of the
sperm was so complete that they did not enter the jelly of the eggs at
all. Ninety-seven minutes later none of the eggs in (1), (2), (8) or (4)
had segmented; whereas at least 5 per cent of the control were now in the
two-celled stage. The lot of eggs was rather poor in this case, but fertili-
zation was confined entirely to the control.
If the experiment be made in another way some recovery of the
spermatozoa from their state of paralysis may be observed. Thus: An
§ It should be borne in mind that but little is known concerning lysins or agglu-
tinins of invertebrates. It is perhaps not to be expected that they should exhibit
the same degree of thermolability as those of vertebrates.
STUDIES OF FERTILIZATION 559
active sperm suspension was divided in two parts, and one part was
agglutinated by the addition of about 40 per cent of its own volume of
the egg-extract described above, to the other an equal amount of sea-
water was added. The first was strongly agglutinated; after reversal
both suspensions were stirred up, and beginning thirteen minutes after
agglutination a series of fertilizations were carried out by adding one
drop of the agglutinated sperm suspension to a measured quantity of
eggs in about 9 cc. of sea-water at 10 minute intervals. Each fertiliza-
tion had a control of the same quantity of eggs fertilized with one drop
of the control sperm. The consistent result was that about 16 per cent
of the eggs fertilized with the agglutinated sperm segmented and at
least 33 per cent of the control. The non-agglutinated spermatozoa are
about twice as effective as the agglutinated. But a considerable degree
x recovery of some spermatozoa of the agglutinated suspension is
shown.
2. Nereis. Experiments with Nereis did not give such a marked
reduction of the fertilizing power of agglutinated sperm as in Arbacia.
There was, however, a.marked delay in the formation of jelly when agglu-
tinated sperm was used as compared with normal sperm.
It is somewhat difficult to make a satisfactory interpretation
of the effect of agglutination on fertilizing power. On the one
hand we may suppose that a certain proportion of spermatozoa
resist the agglutination effect, and are alone concerned in any
fertilizing power of an agglutinated suspension; on the other hand
it might be supposed that the agglutinating effect does not modify
fertilizing power except as it decreases the motility of the sperma-
tozoa or that the effect is reversible under the condition of the
experiments. Either assumption would be consistent with the
facts.
8. CONDITIONS OF FORMATION OF THE AGGLUTININ BY THE EGGS
The conditions of excretion of the agglutinating substance by
the eggs into the sea-water is quite different in the two forms.
In Arbacia the agglutinin is excreted continuously by unfertilized
eggs in such amounts that I have not succeeded even by repeated
washings in removing it all. Fertilization does not appear to
increase or decrease the quantity. In Nereis, on the other hand,
unfertilized eggs secrete but little of it, and one or two washings in
‘sea-water will completely remove it so that the eggs secrete no
more in detectable quantities; but at the moment of fertilization,
on the other hand, it is poured forth in advance of the jelly in
560 FRANK R. LILLIE
large quantities and the eggs then appear to have disposed of their
entire store, for washed fertilized eggs no longer produce it. The
conditions are important in their bearing upon the question of
permeability of the egg-membrane with reference to fertilization.
The spermatozoa are very efficient indicators of substances leay-
ing the egg. In the case of Nereis it can be shown that there is a
sudden increase of permeability at the moment of fertilization,
but in Arbacia such evidence is lacking.
The conditions in Arbacia may be shown by the following experi-
ment: Eggs were washed free from all fragments of ovary and placed in
about 20 times their own bulk of sea-water and divided in three lots,
a, b and c. The sea-water over them agglutinates Arbacia sperm in-
stantaneously. This test was made at 9.30 a.m. The supernatant fluid
was then removed and the eggs washed in 20 times their bulk of sea-
water. At 9.53 the supernatant water was tested and found agglutina-
tive. 9.58, a and 6 were again washed. 10.04, supernatant fluid again
agglutinative. 10.05, lot a was fertilized with 2 drops of sperm. 10.10,
a and b tested again; both agglutinate. 10.16, a and b washed again.
10.23, supernatant fluid of both agglutinates. 10.30, a and b washed
again. 10.36, both agglutinate sperm; b is rather more effective. 10.40,
washed a andbagain. 10.45, both agglutinate; b more effective. 10.51,
washed a andb again. 11.00, both agglutinative; a more than b. 11.08
to 11.20, other tests of a and b show a somewhat more effective.
The experiment shows both fertilized and unfertilized eggs of
Arbacia to be constantly secreting a substancé into the sea-water
which agglutinates the spermatozoa. The substance must be
effective in very minute quantities for the amount of water used
in each of the eight washings was at least ten times the bulk of the
eggs, yet as soon as the eggs had settled to the bottom of the vials
used the supernatant fluid contained the agglutinin in appre-
ciable amounts. How long this process keeps up in Arbacia I
cannot say; and too much reliance cannot be placed on the result
for the fertilized eggs because a large proportion of the eggs failed
to fertilize or at least to segment.
The conditions in Nereis are quite different; the experiments
showed that shortly after the eggs are taken they charge at least
ten times their bulk of sea-water with an easily detectable amount
of sperm agglutinin. But if the eggs are washed once usually no
more agglutinin can be demonstrated. If now the eggs are stirred
STUDIES OF FERTILIZATION 561
up in the vial, fertilized and allowed to settle, the supernatant
fluid is very powerfully agglutinative. The experiment was
repeated a sufficient number of times to make certain of this
result.
A number of tests were also made with the object of determining
if the eggs of Nereis continued to produce agglutinin after fertili-
zation. These showed that the eggs cease very quickly their
production of agglutinating substance. None could be detected
during the maturation period, but apparently there is a second
production about the time of the first cleavage. On this point I
wish to be understood to speak with reserye. The swelling of
the jelly secreted at fertilization makes the eggs very bulky, and
the jelly itself takes up any egg secretion, so that there are consid-
erable technical difficulties in making satisfactory tests.
The fact that stands out perfectly plainly in the case of Nereis
is the sudden increase in secretion of agglutinin into the sea-
water just after insemination, followed by cessation of its pro-
duction. The spermatozoa give absolutely positive tests. As
will be shown later, the observations on the normal fertilization
are in complete harmony with this.
The difference between Nereis and Arbacia in these respects
is thus sharply marked. However, it should be remembered that
the unfertilized eggs of Arbacia have formed both polar bodies,
whereas those of Nereis are in the stage of the germinal vesicle.
It may be that eggs of Arbacia in the germinal vesicle stage are
relatively impermeable in the same sense as those of Nereis.
9. HETERO-AGGLUTINATION AND THE QUESTION OF SPECIFICITY:
REACTIONS BETWEEN NEREIS AND ARBACIA
The demonstration of intraspecific sperm-agglutinating sub-
stances derived only from the ova having been made, the question
arose whether these substances were essentially the same in both
species, or different. If the same, the egg-extract of each should
agglutinate the sperm of the other. A number of tests were
therefore made which demonstrated conclusively that the sub-
stances are decidedly different with reference to their cross-agglu-
562 FRANK R. LILLIE
tinating effects, and this result has raised a number of questions,
the most important of which relate to the question of specificity,
but which could only be defined in the time at my disposal and
not definitely answered. ,
The first suggestion that the sperm agglutinating substances of
Arbacia and Nereis are different came from the following experi-
ment: A raised coverslip preparation of Nereis sperm was made
in the usual way and into it were injected a drop each of (1)
Arbacia egg-extract in sea-water, (2) drop a of coelomic fluid from
a female Arbacia, (3) a drop of coelomic fluid from a male Arbacia.
All three caused very extensive firm agglutinations of the Nereis
sperm.
Thus the egg-extract of Arbacia contains an agglutinating sub-
stance for the Nereis spermatozoa as well as for its own; but the
coelomic fluid of Arbacia also causes agglutination of the Nereis
spermatozoa, whereas it is perfectly neutral with respect to its
own. This demonstrated, therefore, the existence of at least two
sperm-agglutinating substances in Arbacia, namely: One in the
egg-extract agglutinative for its own sperm and that of Nereis,
and one in the coelomic fluid not agglutinative for its own sperm,
but agglutinative for the foreign sperm of Nereis. The prob-
ability is, therefore, that the egg-extract contains both substances
seeing that it is agglutinative for both kinds of spermatozoa.
This was afterwards demonstrated.
The reciprocal experiment proved the difference of the Nereis
and Arbacia agglutinating substances conclusively, for it was
shown that Nereis egg-extracts strongly agglutinative for sperm
of Nereis had no agglutinating effect on Arbacia sperm within the
limits of attainable concentrations (several experiments). The
Arbacia fluids are extremely toxic apparently to the Nereis sperm
for the agglutinations were more solid than those caused by the
iso-agglutinating substance. The absence of a reciprocal effect,
that is, of Nereis extracts on Arbacia sperm, is therefore all the
more striking.
The same difference may be shown by the reactions of sperm
suspensions of both species to one another. The experiment was
made in the following manner:
STUDIES OF FERTILIZATION 563
July 28, 1912. a. A suspension of active Arbacia sperm was made at
11.23 a.m.
b. A suspension of active Nereis sperm was made at 11.24 1/2.
The two suspensions were made of equal density as far as possible.
Part of each suspension was then mounted on a slide beneath a raised
cover-slip.
A drop of the Nereis sperm was then injected into the Arbacia slide
(slide 1) and a drop of Arbacia sperm into the Nereis slide (slide 2).
Slide 1 gave a very faint reaction; only slight evidence of a ring for-
mation at the margin of the Nereis drop.
Slide 2, on the other hand, gave a very pronounced reaction due to
inwandering of the Nereis sperm into the Arbacia drop, followed by
agglutination of the inwandering sperm.
The difference in reaction is due to two circumstances: (1) The
Nereis sperm exhibit a more pronounced chemotaxis than the
Arbacia sperm, hence they tend to enter the drop of Arbacia
sperm, whereas on the other slide the Nereis sperm tend to diffuse
from the drop into the Arbacia suspension. (2) The Nereis
sperm that wander into the drop of Arbacia sperm are agglutin-
ated, but there is no reciprocal reaction of the Nereis sperm on any
inwandering Arbacia spermatozoa.
It is demonstrated, therefore, first that Arbacia fluids in general
are toxic for Nereis spermatozoa to the extent, at least, that they
cause agglutination; but on the other hand, that no secretion of
Nereis appears to cause agglutination of Arbacia spermatozoa;
and second that the eggs of each species produce an agglutinin for
the sperm of its own species.
Agglutination is not in itself a specific process; it may take place
spontaneously to a certain extent under some conditions; it is
caused by increase of alkalinity of the sea-water in the case of
’ Nereis and Arbacia, or by the action of certain foreign sera as in
the case of the action of Arbacia fluids on Nereis sperm. On the
other hand, the class of specific immune agglutinins, character-
ized by limitation of their action to the specific form of blood or
sperm used as antigen is well known. The question naturally
arises, therefore, to which class the iso-agglutinating substances
produced by ova of Arbacia and Nereis belong.§®
6 In the latter case, fertilization itself would have to be regarded as an immuniz-
ing process, the sperm acting as antigen after entrance into the egg. It seems, in
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 4
564 FRANK R. LILLIE
The fact that the egg secretions of Arbacia cause agglutination
of Nereis sperm as well as Arbacia sperm seems at first sight
to indicate the iso-agglutinating substance of Arbacia is not
specific. But the fact that other Arbacia fluids likewise aggluti-
nate the sperm of Nereis, but not those of Arbacia, raises the
question whether the egg-secretion does not contain in addition
to the iso-agglutinating substance, also another which is aggluti-
native for Nereis sperm like the substance in the coelomic fluid,
but not for its own sperm.
If there are two substances present in the egg secretion it ought
to be possible to separate them by various means. They might
exhibit different heat-lability, so that one might be destroyed at
a temperature that would leave the other still active. Of if they
have different affinities it might be possible to fix the Nereis agglu-
tinating substance by Nereis sperm, leaving the iso-agglutinating
substance intact. Neither of these experiments could be per-
formed this year, owing to the disappearance of the necessary
material.
However, it was possible to show in another way that the Nereis-
agglutinating substance of Arbacia egg-extract is distinct from the
iso-agglutinating substance: An egg-extract of Arbacia seventeen
days old was found to have entirely lost its power of agglutinating
Nereis spermatozoa, while it retained undiminished its power of
agglutinating Arbacia spermatozoa. Originally it agglutinated
both kinds of spermatozoa. Now the change that took place in
the egg-extract on standing is not a mere weakening of action as
might be supposed, because the iso-agglutinating action was
noted as undiminished, whereas the hetero-agglutinating action ,
was entirely lost. The only possible conclusion, therefore, is
that the egg-extract contained two agglutinating substances at
least, namely: An iso-agglutinin and a hetero-agglutinin, and
that the latter is relatively labile, the former relatively stable.
Unfortunatelly the experiment could not be repeated onaccount
of the total disappearance of the material.
fact, an almost necessary conception on the general principles of immunity phe-
nomena that the sperm should so act. The question would be, of course, whether
there is a connection between any antibodies so formed and the sperm iso-agglu-
tinins produced by the next generation of ova.
STUDIES OF FERTILIZATION 565
The experiment in detail was as follows:
September 15, 1912. One small male Nereis available. Its sperm
is aggregative and very active. The Arbacia egg-extract was of August
29 and was made by cutting up ovaries and eggs of Arbacia in four times
their volume of sea-water. After settling of the eggs the supernatant
fluid was poured off; and had been kept in a stoppered vial since.
1. A drop of the Nereis sperm and a drop of the egg-extract were
placed side by side on a slide and connected; the sperm diffused into the
egg-extract and swam around in it; no agglutination.
2. A raised cover mount was made of the Nereis sperm and a drop of
the egg-extract injected. The sperm entering the drop swam around,
and were not agglutinated.
Controls: 1. The sperm of Nereis was agglutinated immediately by
extract of Nereis eggs kept since September 8.
2. The egg-extract of Arbacia agglutinated Arbacia spermatozoa with
no apparent diminution in the strength of the reaction.
Conclusion: The nonspecific agglutinating substance has been de-
stroyed by the chemical changes in the extract in the course of seventeen
days; but the iso-agglutinating sperm substance still remains.
This experiment does not demonstrate that the sperm iso-agglu-
tinin of Arbacia egg-extract is specific, but merely that it is with-
out effect on the Nereis sperm, just as the iso-agglutinin of Nereis
eggs is without effect on Arbacia sperm. It is of course still
possible that the iso-agglutinating substances might have an
agglutinating effect on some other varieties of sperm, and a crucial
test of specificity must await the securing of new material.
However, it seems to me that the probabilities in the case lie
strongly on the side of specificity of these sperm iso-agglutinating
substances. Quite apart from the value of the evidence already
adduced, we must consider the general fact that ova and sperma-
tozoa of the same species do behave in a specific way with refer-
ence to one another in the process of fertilization. This must have
some chemical basis and on the chemical side the only reactions
that exhibit a corresponding degree of specificity are those between
antigens and anti-bodies in the field of immunity. We have two
parallel instances, therefore, and the slight evidence which I have
so far been able to bring forward in favor of the specificity of the
sperm iso-agglutinins of ova gains immensely in weight by its
association with the universal principle of specificity in fertili-
zation, and the known class of specificities in agglutination re-
actions.
566 FRANK R. LILLIE
The agglutination of spermatozoa is, of course, in itself of no
significance for the problem of fertilization; the spermatozoa
unite in fertilization with the egg, not with one another. The
agglutination reaction is, however, an indicator of an important
change in the spermatozoon in the presence of egg secretions, and
therefore evidence of a change that any spermatozoon must
undergo when it comes in contact with the egg. The adhesive
property that the sperm develops under these circumstances may
be an important factor in binding the sperm to the egg until it
can be incorporated. But, if the reaction be specific, it is much
more than this; it is evidence of an intimate chemical combination
of sperm and egg constituents which begins at the very moment
of union.
Von Dungern’s experiments (’02) are the only ones, so far as I
know, in which the production of sperm agglutinins by ova was
investigated, and he discovered only hetero-agglutinins, no iso-
agglutinins. He did, indeed, describe the loss of motility of
spermatozoa in egg-extracts of the same species, but he entirely.
missed the phenomenon of agglutination and its reversal. He
reveals the reason for this failure by his remark that he always
examined for the effect of the ‘egg-poison’ about half-an-hour
after its addition to the sperm; but the phenomenon of agglutina-
tion and its reversal are completed in about five minutes.
Von Dungern also made experiments on the production of
immune sera by injection of ova and spermatozoa separately
into rabbits, and found that both caused the production of a
sperm agglutinin in the rabbit’s serum. From this he concludes
that both kinds of reproductive elements possess chemically
identical complexes of molecules in the protoplasm. While this
may be admitted as at least a very probable conclusion, his farther
conclusion that fertilization does not depend upon any specific
antagonism between ovum and sperm, but is conditioned by the
similarity of their protoplasms, is not well founded, for the egg is a
very complicated chemical system, and it certainly contains mole-
cules antagonistic to sperm, even if, as Von Dungern’s experi-
ments indicate, it also contains some that are not.
~J
STUDIES OF FERTILIZATION 56
10. INTERPRETATION OF SOME PHENOMENA OF NORMAL
FERTILIZATION IN NEREIS
To observe all the details of normal fertilization it is desirable
to inseminate in a suspension of India ink which will define the
transparent substances exuding from the egg on insemination.
the following observations can then readily be made on mixing
a drop of the eggs in the ink suspension with a drop of opalescent
sperm suspension: Hundreds of spermatozoa become attached to
each egg almost immediately; those in contact with the egg do
not show much activity, but are usually definitely oriented radi-
_ally; the spermatozoa external to these are in active movement.
In about a minute a clear fluid begins to exude from the egg and
surrounds all attached spermatozoa and involves the immediate
neighbors. The first exudate is quite fluid for it flows around the
spermatozoa and does not sweep them away, but the movements
of all the spermatozoa within the exudate cease suddenly. The
flow continues, and then most of the spermatozoa are swept away
from contact with the egg, for the later exudate is gelatinous in
consistency. However, a good many spermatozoa remain in
contact with the egg for some time, but these are detached one by
one as the flow of the jelly continues, until only one remains.
Some of the supernumerary spermatozoa are not carried away
until five or more minutes after insemination.
The immediate prevention of polyspermy in Nereis appears
to be due to the paralyzing effect of the egg-exudate poured out
in response to the stimulus of the first effective spermatozoon.
Polyspermy could take place in Nereis only under two conditions,
namely, (1) if two or more spermatozoa simultaneously give the
stimulus to the egg that causes excretion of the agglutinin; for
the condition of stimulus appears to be that the spermatozoon be
securely anchored to the egg; and all spermatozoa not securely
attached at the moment the egg begins to secrete are prevented
from securing attachment by the resulting paralysis; (2) if the
reaction of the egg be slow and therefore localized at first to the
region of an effective spermatozoon, opportunity will be afforded
for attachment of other spermatozoa.
568 FRANK R. LILLIE
In a short time the egg develops a physiological condition in
which union of spermatozoa is no longer possible. The immediate
protection against supernumerary spermatozoa is, however,
afforded by the paralyzing action of the egg-secretion.
Union of ovum and sperm, prevention of polyspermy, and the
attainment of a condition of insusceptibility to other spermatozoa
are phenomena so closely related in time sequence that a casual
connection must be postulated. These can be brought under
one head, in the case of Nereis at least, if we assume that the sub-
stance that paralyzes all the sperm in the vicinity of the egg is
necessary for the actual fusion of the spermatozoon and egg and
is completely used up in the cortical changes that follow immedi-
ately on insemination: the condition of insusceptibility would be
due to loss of a necessary substance, the immediate prevention of
polyspermy to paralysis of all ineffective spermatozoa, and the
penetration of the sperm to a chemical change of the effective
sperm and the neighboring egg-cytoplasm involving physical alter-
ations in surface tension, viscosity, et cetera.
In the case of Arbacia, I have been unable to demonstrate an
increase of secretion from the egg into the sea-water at the mo-
ment of insemination nor yet cessation ofsuch secretionsoon after
insemination; however, as indicated before (p. 560), this failure
may be due to the presence of unfertilized eggs in the experiments,
which require to be repeated.
11. SUMMARY: PART IV
1. The ova of Nereis and Arbacia give off into the sea-water
a substance (or substances) which agglutinates the sperm of their
ownspecies. The sea-water, which has the agglutinating substance
in it, has also a substance to which the spermatozoa of the same
species are positively chemotactic.
2. The eggs alone produce the sperm agglutinating substance;
It cannot be extracted from other tissues.
3. The agglutinin disappears from a mixture of sperm suspen-
sion and agglutinin if not present in excess; the disappearance is
attributed to chemical combination.
STUDIES OF FERTILIZATION * 569
4. The agglutinating substances are highly thermostable, but
are slowly destroyed by temperatures above 95°C.
5. In the presence of excess of the agglutinating substance
spermatozoa of Arbacia lose their fertilizing power.
6. Eggs of Arbacia give off the agglutinating substance in
the sea-water in large quantities prior to insemination; but eggs
of Nereis give off only smal] quantities until inseminated, or until
the cortical change analogous to membrane formation in the sea-
urchin egg is somehow produced.
7. The egg-extract of Nereis does not agglutinate Arbacia
spermatozoa.
8. The substance in the egg-extract of Arbacia that aggluti-
nates Nereis spermatozoa is distinct from the iso-agglutinating
substance.
9. The coelomic fluid of Arbacia contains a substancewhich
agglutinates the spermatozoa of Nereis but not of Arbacia. Pre-
sumably this substance is the same as the hetero-agglutinating
substance of Arbacia egg-extract.
10. Two arguments in favor of the specificity of the iso-agglu-
tinative reaction were brought forward, namely, (a) The fact
that the iso-agglutinin of Arbacia is distinct from the hetero-
’ agglutinin in the case of Arbacia and Nereis; (b) that fertilization
is fundamentally a specific reaction, and that the phenomena of
agglutination belong in a class of phenomena in which specificity
exists, and between elements which react specifically in fertiliza-
tion. While admittedly not demonstrative these srernent
appear to me to be cogent.
V. DISCUSSION
Since Pfeffer’s fundamental investigations concerning chemo-
taxis of spermatozoa of ferns and mosses with reference to the
secretion of the archegonia a similar explanation of the behavior
of animal spermatozoa with reference to the eggs of the same spe-
cies has been anticipated, and indeed has been postulated by
many writers without any other experimental basis. However
such actual experiments as have been performed have not been
favorable to such an interpretation. Thus Buller (00) experi-
570 FRANK R. LILLIE
mented with the spermatozoa of sea-urchins by the method of
Pfeffer and came to the conclusion that ‘“‘the spermatozoa of the
Echinoidea are not attracted to the egg by means of any special
substance excreted by the latter. The vast number of spermato-
zoa and the large size of the eggs are sufficient to ensure the neces-
sary contact taking place.” Von Dungern (’02) also rejects
the idea of an egg-secretion attracting the spermatozoa in the
ease of sea-urchins and starfish. Morgan, Payne and Brown
(10) also accept Buller’s interpretation, and there has recently
been a tendency among biologists to reject chemotaxis of the
spermatozoon as a factor in the fertilization of the egg.
Previous observers have worked either with Pfeffer’s capillary
tube method, or with the eggs themselves. - I made a sufficient
number of experiments with capillary tubes to convince myself
that this method of experimentation is many times less effective
than the method I employed. As illustration: July 4, 1912:
I filled three pieces of capillary tubing with a concentrated solu-
tion of CO, in sea-water, with 10 per cent and 1 per cent of this
solution and broke off short pieces—neither end of which had
been in the solution—to be tested. These pieces were then intro-
duced into an active sperm suspension of Nereis beneath a raised
cover-slip. In the course of a few minutes a decided positive
reaction was obtained with the-first and second tubes, the sperm
appeared to stream into the open mouth of the capillary tubes and
soon formed white plugs at the mouths of the tubes. The tube
containing | per cent CO, sea-water and a control tube with sea-
water alone showed no reaction. In a repetition of this experi-
ment the 1 per cent CO, sea-water and control were negative, and
the 10 per cent showed only slight reaction. The diameter of
the lumen of the tube was 0.48 mm. A much finer tube of 10
per cent CO, sea-water showed no reaction at all. With tubes of
the size first used 10 per cent CO, sea-water is near the minimum
for a positive reaction. But a drop of 0.5 per cent CO: sea-water
injected into a similar suspension causes chemotactic response.
Tubes of the size used are therefore about twenty times less effec-
tive as indicators of the chemotactic reaction, stated in terms of
percentage of CO, required, than the injected drop method. It
STUDIES OF FERTILIZATION Bee
is obvious that the size of the tube is a fundamental condition of
the experiment, both because the diffusion is a factor of size
and also because of the interference of thigmotactic reactions of
the spermatozoa at the mouth of the tube with the purely chemo-
tactic response. Buller does not state what was the size of the
tubes that he used in his experiments. But, if the delicacy of the
reaction was reduced to one-twentieth by the tube, his failure to
get a positive reaction in tubes containing sea-water taken from
over eggs is not surprizing.
Von Dungern drew his conclusions from observing the behavior
of spermatozoa mixed with eggs. Many embryologists, like
~myself, have made hundreds of observations of this kind; but it
is obvious that the conditions thus created render an analysis of
the behavior of the spermatozoa impossible. In some experi-
ments I introduced a drop of eggs in sea-water into a sperm
suspension beneath a raised cover, and obtained the typical ring
formation of spermatozoa with reference to the group of eggs
considered as a whole. But within the group any evidence
of chemotactic reaction is clearly impossible.
As to the réle that chemotaxis as a principle may play in the
fertilization of the ova in nature it is difficult to form a clear con-
ception. It may be little and it may be considerable. In the
first place it may be noted that, although the echinids have been
favorite subjects for research, but little appears to be actually
known concerning their breeding behavior. In the second place
we do not know the distance to which the secretion from an iso-
lated egg will diffuse. But even if we assume that it extends
effectively only a short distance in terms of the egg-diameter the
result would be essentially to immensely increase the chances of
scattered spermatozoa to become entangled in the jelly of the
egg. Measurements of the effective radius of diffusion of the
egg-secretion could, I believe, readily be made by the method
employed in my work and the results of this might enable us to
form some clearer idea of the possible significance of chemo-
taxis taken by itself in the meeting of the germ-cells.
The present results merely show that it may be a factor of some
significance. The quickness and readiness of the reaction of
572 FRANK R. LILLIE
spermatozoa of Arbacia and Nereis to the secretion of their own
kind of eggs is certainly surprising.
Another effect of their secretions that should be taken into
account in this connection is the stimulating effect on the sperma-
tozoa. This is more marked in some animals than in others.
Thus the spermatozoa of Nereis are so active in the sea-water
alone that but little effect of the egg-secretions can be noted; in
the case of Arbacia, although the sperm are quite active in pure
sea-water yet the egg-secretions greatly increase their activity
for a brief time. In the case of the star-fish, according to Von
Dungern’s account, the spermatozoa tend to be very inactive in.
pure sea-water, but are aroused to intense activity by the secre-
tions of the ova.
In different animals, therefore, we may expect to find some
difference in the effect of egg-secretions on the activity ofthe
spermatozoa. But the fact that in such widely separated forms
as Arbacia and Nereis secretions of the egg cause strong positive
chemotaxis of the spermatozoa inclines one to the view that such
a reaction may be very wide spread. In a form in whichegg-
secretions are both activating and directing in their action, the
importance of such secretions in favoring the preliminary steps in
fertilization can hardly be doubted.
The experiments, like Pfeffer’s earlier ones, indicate that the
factor of specificity is probably subordinate in the purely chemo-
tactic response. CO, and acids are in no sense specific, but they ©
are very effective chemotactic agents with Nereis spermatozoa.
But the case of Arbacia serves to indicate that substances of the
egg, whether specific or not, are more generally effective than
simple chemical substances, for its requires such substances,
apparently, in the case of Arbacia to produce a reaction of the
spermatozoa comparable in quickness and precision to the reac-
tion of Nereis spermatozoa to acid, COs, or the secretions of its
own eggs.
We have seen that in some respects the chemotactic behavior
of spermatozoa of Nereis and Arbacia is different depending on
their relative sensitiveness to CO, and other agents. The much
greater power of resistance of spermatozoa of Chaetopterus and
STUDIES OF FERTILIZATION 573
Loligo to CO, (see p. 528) indicates still greater differences in
characteristic behavior. And it may be that the differences of
chemotactic behavior of the spermatozoa of various animal
phyla will turn out on investigation to be extensive. I am very
far, therefore, from wishing to generalize any of the principles
that we have found to hold true for Nereis and Arbacia. Sound
generalizations must be based on much more extensive work. I
have made some observations on the sperm of Platynereis mega-
lops which demonstrate great differences as compared with Nereis
in spite of the close relationship, correlated, no doubt, with dif-
ferences inbreedingbehavior. While the common form of organ-
ization of flagellated spermatozoa points to fundamental principles
of behavior in common, yet it must not be forgotten that each
kind of spermatozoa has the chemical composition of the species,
and may therefore have entirely specific forms of behavior.
The agglutination of the spermatozoa by normally formed
egg-exudates of the same species indicates the possibility of study-
ing the chemistry of fertilization directly through use of the
spermatozoa as indicators. The very few and incomplete results
which I was able to obtain in the time at my disposal seem to me
to indicate a fruitful line of work. It would be interesting, for
instance, to investigate whether or not the ova of hermaphrodite
animals produce a sperm auto-agglutinin, that is, an agglutinat-
ing substance for spermatozoa of the same individual. Morgan’s
work on Ciona has shown that the failure to self-fertilize is in this
ease due to failure of penetration of the spermatozoon. It is
difficult, as he points out, to explain this on any mechanical
grounds. But if a specific agglutination is a necessary step in
union of ovum and spermatozoon, the failure to produce an auto-
agglutinin would explain the failure of self-fertilization. We
would have in this event a precise parallel to the usual failure to
produce blood auto-agglutinating substances in experiments on
immunity, though iso-agglutinins are readily produced.
Godlewsky (’11) has shown that there is an antagonism between
the sperm of certain animals (Chaetopterus and Echinids) which
destroys the fertilizing power of each when mixed together for a
certain length of time. He compares this to the antagonistic
574 FRANK R. LILLIE
action of heterogenous haemolytic sera on one another; and con-
cludes that his results strongly confirm Loeb’s theory that the
spermatozoon initiates development by means of a lysin.
Without discussing the interpretation, and considering only
Godlewsky’s most interesting results, another parallel is furnished
to immunity phenomena. We may confidently expect, therefore,
that study of the reactions of spermatozoa will break a new path
into the field of fertilization.
LITERATURE CITED
Buuuer, A. H. Reainatp 1900 The fertilization process in Echinoidea. Re-
port Br. Ass. Adv. Sci., p. 387.
von DuNnGERN, Emit 1902 Neue Versuche zur Physiologie der Befruchtung.
Zeitschr. f. allgem. Physiologie, Bd. 1.
_Gop.Lewski Jun., Emit 1911 Studien iiber die Entwicklungserregung. II.
Antagonismus der Einwirkung des Spermas von verschiedenen Tier-
klassen. Arch. f. Entw. Mech. der Organismen, Bd. 33.
Linuig, Frank R. 1911-1912 Studies of. fertilization in Nereis. I. and II,
Jour. Morph., vol. 22; III and IV, Jour. Exp. Zool., vol. 12.
1912 The production of sperm iso-agglutinins by ova. Science, N.S.,
vol. 36, pp. 527-530. October.
Morean T. H. 1910 Cross and self-fertilization in Cionaintestinalis. Arch.
f. Entw. Mech. der Organismen., Bd. 30, Teil 2.
MoreaNn, T.H., Paynn, E., AND Browne, E.N. 1910 A method to test the hypo-
thesis of selective fertilization. Biol. Bull., vol. 17, No. 2.
THE: EFFECT OF EXCRETION PRODUCTS OF INFUSO-
RIA ON THE SAME AND ON DIFFERENT SPECIES,
WITH SPECIAL REFERENCE TO THE PROTOZOAN
SEQUENCE IN INFUSIONS
LORANDE LOSS WOODRUFF
Sheffield Biological Laboratory, Yale University
It is well known that a hay infusion presents a kaleidoscopic
series of phenomena from its inception until it finally reaches a
stage of sterility, or, in the presence of sunlight, of practically
stable equilibrium in which animals and green plants become so
adjusted that a veritable microcosm exists. In an attempt to
elucidate some of the complex factors involved in the faunal and
floral changes of typical infusions a series of observations of a
considerable number of infusions was made,! and the following
conclusions, among others, were reached:
1. In hay infusions, seeded with representative forms of the chief
groups of Protozoa, there is a definite sequence of appearance of the
dominant types at the surface of the infusion, that is, Monad, Colpoda,
Hypotrichida, Paramaecium, Vorticella and Amoeba.
2. The sequence of maximum numbers and of disappearance is identi-
cal with that of appearance, except that apparently the position of
Amoeba advances successively from the last (sixth) place to the fifth
place and then to the fourth place.
3. Emphasis is put upon the strictly biological interrelations (for exam-
ple, those involving food and specific excretion products) of the various
forms as the most important determining factors in the observed
sequence.
The interdependence of the organisms of a hay infusion is so
complex that, taken as a whole, it is quite beyond the possibility
of analysis, and accordingly the logical method of approach to
the subject was to study the effects of isolated organisms on them-
selves and on each other. ‘The first essay in this direction con-
17. L. Woodruff, Observations on the origin and sequence of the protozoan
fauna of hay infusions. Jour. Exp. Zodl., vol. 12, no. 2.
-——-~
ov19
576 LORANDE LOSS WOODRUFF
sisted of a study of the effects of media contaminated with the
excretion products ofa single form (Paramaecium) on its rate of
reproduction.2?. The results showed that paramaecia excrete sub-
stances which are toxic to themselves when present in their
environment and reduce the division rate. This fact was inter-
preted as indicating that excretion products play an appreciable
part in determining the period of maximum numbers, rate of
decline, et cetera, of this animal in hay infusions.
In view of these results it was of interest to attempt to answer
the following questions:
1. Are the excretion products of paramaecia specifically toxic
to.this el eaaed or do they also influence the reproductive rate of
other species?
2. Do other species in the observed sequence of forms of the
infusion fauna produce excretion products which are specific, or
otherwise, in their action?
3. Do species which succeed each other in the observed sequence
produce substances which influence mutually their development
and tenure of life in the infusion microcosm?
The present paper presents results which, it is believed, con-
tribute to the solution of these questions with respect to Para-
maecium and two hypotrichous forms, Stylonychia pustulata and
Pleurotricha lanceolata. The series of experiments described
were made in an endeavor to determine the influence of media
contaminated with the excretion products of paramaecia on the
rate of reproduction of Paramaecium and of the hypotrichous
forms; and also to determine the influence of media contami-
nated with the excretion products of the hypotrichs on the rate
of reproduction of these forms and of Paramaecium.
It was decided to study the problem with respect to paramaecia
and the hypotrichs because the latter have been shown, in the
earlier paper, to occur in maximum numbers just before the
dominance of the paramaecia, and because they lend themselves
readily to such experimental methods as the present study
involves.
* L. L. Woodruff, The effect of excretion products of Paramaecium on its rate of
reproduction. Jour. Exp. Zodl., vol. 10, no. 4.
a |
EFFECT OF EXCRETION PRODUCTS OF INFUSORIA au
The paramaecia used in this work were from my pedigreed
race of Paramaecium aurelia (I). This was at the 2800th genera-
tion in January, 1912, when the study was started and had
attained the 3450th generation by the conclusion of the experi-
ments in December of the same year. For the hypotrichous
forms which were employed I am indebted to Mr. George A.
Baitsell who supplied them from his pedigreed races. Emphasis
is placed on the fact that the animals which formed the subjects
of the experiments had been under observation for considerable
periods of time (in the case of the paramaecia, for five years) and
consequently their rate of reproduction and the exact conditions
-to which they had been subjected before the experiments were
known. Further, since the pedigreed races were each originally
started with a single individual, all the specimens of the respective
species were ‘sister’ cells and therefore all the experiments were
made on the ‘same protoplasm.’
A weak extract of hay was employed as a culture medium,
which was made by boiling about 10 grams of hay in a liter of
water for five minutes. The hay was then immediately filtered
off and the liquid distributed equally in three flasks and allowed
to cool. About a dozen paramaecia and a dozen hypotrichs were
then isolated from their respective pedigreed cultures and allowed
to mingle together in a watch glass-of hay infusion. Three clean
cover glasses were taken and on one was placed a drop of the
infusion from the watch glass with a few of the paramaecia, on
another was placed a drop with a few of the hypotrichs, and on
the third was placed simply a drop of the infusion. One of these
cover glass preparations was dropped into each of the three flasks
of culture medium already prepared, which were accordingly
designated P, H and O respectively, signifying paramaecia
seeded, hypotrich seeded, and minus protozoa. The flasks were
then allowed to stand at room temperature for from five to ten
days, or until an enormous growth of the seeded forms had
appeared, and then the media were ready to be studied. These
three flasks thus contained the same culture medium and bacte-
rial flora, and one differed from the other only in the presence
of the protozoon with which it was seeded at the start.
od
578 ; LORANDE LOSS WOODRUFF
The length of the individual experiments varied with the
development of the organisms in the culture medium flasks.
Obviously, it was necessary that there should be very heavy and
essentially similar growths of both the paramaecia and the hypo-
trichs before the test of the media began, and frequently the
medium had to be discarded and a new lot started, owing to the
failure of heavy growths to develop. The more rapid rate of
division of the hypotrichous forms rendered it expedient to seed
the cultures with a few less individuals of these forms than of
paramaecia in order to have the development of the two flasks
synchronous. Again, the actual experiments could be continued
only while heavy growths of the respective species persisted and
consequently the length of the experiments varied, though a great
majority comprised five days. In many cases it was found impos-
sible to conduct series of paramaecia and hypotrichs simulta-
neously—but all data which are directly compared to determine
the results, were run at the same time and consequently tempera-
ture changes are without influence.
The general conduct of the experiments was similar to that
employed in my former work on the subject and therefore it is
unnecessary to describe it in detail again. In each individual
experiment four animals were isolated with the aid of a Zeiss
binocular microscope and placed on separate depression slides and
bred in the medium to be tested, which was supplied fresh daily
during the period of the experiment. The average rate of division
of these four lines, again averaged for the number of days of the
experiment, afforded the data on which this paper is based.
The details of a couple of typical experiments will best illus-
trate the modus operandi. On January 18, 1912, sixteen para-
maecia were isolated from the main pedigreed culture and each
one was placed on a clean depression slide. Four of these were
supplied with medium from flask P from which the paramaecia
had been entirely removed,* four were supplied with culture
medium from flask H from which the hypotrichs had been care-
3 The infusoria were removed by filtering the medium through filter paper and
then picking out with a pipet any of the animals which happened to pass through
the paper. Medium O was also filtered in order to make the treatment the same for
all the types of culture media.
EFFECT OF EXCRETION PRODUCTS OF INFUSORIA 579
fully eliminated, and eight (comprising two sets) were isolated in
media from flask O which was uncontaminated with paramaecia
or hypotrichs. At the same time sixteen Stylonychia were
similarly isolated on separate slides and treated identically the
same as the paramaecia just described. There were then in this
experiment thirty-two lines of cells forming eight sets designated
as follows:
Pol = paramaecia on uncontaminated medium
Po2 = paramaecia on uncontaminated medium
Pp = paramaecia on paramaecia contaminated medium
Ph = paramaecia on hypotrich contaminated medium
Hol = hypotrichs on uncontaminated medium
Ho2 = hypotrichs on uncontaminated medium
Hh = hypotrichs on hypotrich contaminated medium
Hp = hypotrichs on paramaecia contaminated medium
This first experiment gave the following results which repre-
sent the average rate of division of the four lines of cells of each
set, again averaged for the ten days during which they were
continued:
Pol = 2.25 divisions per day
Po2 = 2.31 divisions per day
Pp = 1.95 divisions per day
Ph = 2.46 divisions per day
Hol = 4.50 divisions per day
Ho2 = 4.45 divisions per day
Hh = 4.15 divisions per day
Hp = 3.95 divisions per day
This result is interpreted as showing that the P medium
decreases the division rate while the H medium increases the
division rate of Paramaecium. Further the H medium and the
P medium decreases the division rate of the hypotrichs.
As the second example, Experiment 4 (February) may be cited.
This was continued for four days with the following result:
Pol = 1.80 divisions per day
Po2 = 2.05 divisions per day
Pp = 1.385 divisions per day
Ph = 1.95 divisions per day
Hol = 3.73 divisions per day
Ho2 = 3.75 divisions per day
Hh = 3.40 divisions per day
Hp = 3.50 divisions per day
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, No. 4
580 LORANDE LOSS WOODRUFF
These data indicate that the P medium reduces the division
rate of both Paramaecium and the hypotrichs, while the H
medium reduces the rate of the hypotrichs and is without effect
on that of the paramaecia.
At the end of the four days, Pol and Po2 were continued for
four days more on the same medium while Pp and Ph were trans-
ferred to the O medium. Thus during this second phase of the
experiment all lines were on the uncontaminated medium. The
results were:
Pol = 2.10 divisions per day
Po2 = 2.20 divisions per day
Ppo = 2.10 divisions per day
Pho = 2.00 divisions per day
From these figures it is apparent that placing the Pp on the
O medium brought about in the Ppo series a division rate (within
the limit of error) the same as that of the lines (Pol and Po2)
on the O medium from the start, thus proving that the reduction
of the rate in the Pp series was a result of the P medium.
A comparison of the data from these two experiments shows
that the results are not in entire agreement, that is, in Experi-
ment 1 the H medium favored the division of Paramaecium, while
in Experiment 4 it was without influence. Such discrepancies
are not surprising when the large number of factors involved in
procuring heavy growth of the forms are taken into considera-
tion, but they make it apparent that results of value cannot be
secured without many repetitions of the experiment so that inci-
dental disturbing factors can be eliminated. Accordingly, a
series of experiments were made which involved the observation
of more than 8000 individuals and the isolation of over 2000 ani-
mals. The results of the entire series may be best interpreted
from the following brief table. In the first column (designated
‘minus’) the average division rate of the series was below that of
the controls and beyond the limits of error as indicated by the
differences between the two controls. In thesecond column
(designated ‘neutral’) the average division rate was the same as
the controls or within the limits of error as indicated by the differ-
ences between the two controls. In the third column (designated
EFFECT OF EXCRETION PRODUCTS OF INFUSORIA 5&1
‘plus’) the average division rate was above that of the controls
and beyond the limits of error as indicated by the difference
between the two controls.
MINUS NEUTRAL PLUS
ics. 2 4 0 0
TB [tts iene - o crcaSieap SER err ran if il 0
JETS ss nice Ree te ee eee 3 4 5
= Je eee 3 4 1
It is apparent from this table that the four experiments which
repeat the work already published on Paramaecium substantiate
completely the conclusions there reached that paramaecia excrete
substances which are toxic to themselves when present in their
environment and inhibit their rate of reproduction. It is also
evident that the hypotrichs excrete substances which inhibit their
rate of reproduction since the division rate was reduced in seven
out of the eight experiments—the exception indicating neutrality.
The results from the reciprocal action of the P and the H media
are not so conclusive. Taken at their face value the figures would
indicate that paramaecia excretion products are, on the whole,
neutral or slightly inimical to the hypotrichous forms, while the
latter’s excretion products are, on the whole, slightly favorable
to paramaecia.
Accepting for the moment this interpretation of the BecciPe the
data are, a priori, decidedly interesting. It having been shown
in an earlier paper that, on the average, the hypotrich maximum
at the surface of an infusion is passed before the corresponding
period in the development of the paramaecia is attained, one
would expect, if the hypotrichous species and the paramaecia
mutually influence each others life in the infusion microcosm that
the hypotrichs’ products would produce a favorable medium for
the approaching paramaecium maximum, while the products of
the increasing hordes of paramaecia would render the medium
unfavorable for the hypotrichs and so contribute to their decline.
Obviously, however, the data secured are not sufficiently concor-
dant to render this hypothesis secure though they trend in that
582 LORANDE LOSS WOODRUFF
direction. It is perhaps too much to expect that experiments of
this nature would show conclusively such a delicate adjustment
as must exist between mutually associated forms in the inconceiv-
ably involved infusion complex.
These experiments, however, show conclusively, it is believed,
that both paramaecia and hypotrichs produce substances which
are specifically toxic to themselves. This result is made doubly
secure by the very lack of concordance of the results secured from
the study of their reciprocal action. The data show, for example,
that media which have supported very heavy growths of hypo-
trichs are clearly inimical to the reproduction of the hypotrichida,
while the same media are favorable or neutral to the development
of Paramaecium. ‘This rules out completely the possibility, dis-
cussed in the earlier paper,‘ that the depressing effects observed
in this and the previously published experiments are the results
of changes in the quantity or quality of the bacterial flora (on
which the animals are dependent for food) in the protozoa seeded
and protozoa free media.
SUMMARY
1. Paramaecia excrete substances which are toxic to themselves
and these tend to inhibit the rate of reproduction.
2. Hypotrichs excrete substances which are toxic to them-
selves and these tend to inhibit the rate of reproduction. .
3. These excretion products are essentially specific in their
action since their presence does not uniformly influence the rate
of reproduction of other species.
4. The data secured, therefore, emphasize the importance of
specific excretion products as a factor in determining the limits
of development of individual forms in the infusion microcosm
but do not indicate that these specific products are of great im-
portance in relation to the rate of development of associated
species.
*Page 578.
BINDING —~_ >. JUN 2 2 1966.
QL The Journel of experimental
pi zoology
J68
V.14
cop.2 i 1%
Biological
& Medieal
Serials
PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET
UNIVERSITY OF TORONTO LIBRARY
Pata hes
getsiata
any
taste
*
age Be)
arenes
aes EI
i de
et
foonincene
petites
witrare as
Pee MiN TRS, 8 ceo
r
2
=
i>
ne
a
4
veayce'
iz?
Pe edited
7
S ;
peratpoeatias
Paes
*
Sets
Si
3
oefa nts
ee inte
ap erence tee:
nee
PLR
<
roy
ae i
veieseye
Ties
er poray.
Pie
seta
v
0
‘a
Be oF
ifal ve
ident & NAb Laie
aria
D
¢
tibet
Mere thin.
At as
Tapes
Aids mitt
pe
bi
steaes
oe Se
mS
ate.
Theh
Gphek?
week
Mts}
eats
\ ee
mart
ava?
bint