peat 


or 
a 


Yn asia 
unin 


Hs 


bids 


ee 


, Se. ML the) hh fen) ha aa pare 


Digitized by the Internet Archive 
in 2009 with funding from 
University of Toronto 


http://www.archive.org/details/journalofexperim11broo 


THE JOURNAL 


OF 


EXPERIMENTAL ZOOLOGY 


EDITED BY 


WILLIAM E. Caste FRANK R, LILurr 

Harvard University University of Chicago 
Epwin G. CoNKLIN Jacques LorB 

Princeton University Rockefeller Institute 
Cuar_es B. DAvENPORT Tuomas H. Moraan 

Carnegie Institution Columbia University 
Horacr JAyYNe& GrorGcEe H. Parker 

The Wistar Institute Harvard University 
HERBERT 8S. JENNINGS Epmunp B. WItson, 

Johns Hopkins University Columbia University 

and 


Ross G. HARRISON, Yale University 
Managing Editor 


VOLUME 11 
1911 


‘ 


THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY 
PHILADELPHIA, PA. 


call Gp 


ye 


Rhone a ee y , 7 
A RY 4 \ \ Mi, Pad rs 


4 : an 


me Zs yd it. 
18% CoN AE & 


We NSS Ae iene ify 

shee Lif ae | 

x ey Osi 
NEN 


CONTENTS 
1911 


No.1. JULY 5 


H.S. Jennrnes. Assortative mating, variability and inheritance of size, in 
the conjugation of Paramecium. Sixteen figures.......... 


LoranpE L. Wooprurr AND GuorGe A. Baritsetit. The reproduction of 
Paraemecium aurelia ina ‘constant’ culture medium of beef extract. 
TINS liVGUROS <0 claheo bp GEER TOR RIO OceOo Goon ae ae bce ares 


No. 2. AUGUST 20 


Ruty B. Howranp. Migration of retinal pigment in the eyes of Branchipus 
eli Us eee OUI PUITES He yore ere feuroyereieis chat erie eras acre te Tere ERE crete ols test « 


Davenport Hooker. The development and function of voluntary and 
cardiac muscle in embryos without nerves. Fifteen figures 


No. 3. OCTOBER 5 


C. M. Cuiip. Studies on the dynamics of morphogenesis and inheritance in 
experimental reproduction. II. Physiological dominance of anterior 


over posterior regions in the regulation of Planaria dorotocephala. 
ae tksyy 


NEMO TTS. oe od oo ORE nA Motos ord Goes Neca Tear meee 


C. M. Cutty. Studies on the dynamics of morphogenesis and inheritance in 
experimental reproduction. III. The formation of new zodids in Plana- 


Rid eaNGOUNerstOnmMs smn y=SiKd PULES= joe (efota cele a's of-:s «= \cie\« alalelaaleeemies ‘ 


H. V. Witson. On the behavior of the dissociated cells in Hydroids, Aleyo- 


ATI AC AN CMASLOLIAA MMC DIT byt PUTER cele se). sisicicis <),ec0'~ oiels phnietnelalelateiebnrets 2 


ili 


143 


159 


lV CONTENTS 
No. 4. NOVEMBER 20 


LoranpvE Loss Wooprurr AND GEORGE ALFRED BarrsELL. Rhythms in the 
reproductive activity of Infusoria. Thirteen charts.................... 339 


G. H. Parker anp H. M. Parsuiry. The reactions of earthworms to dry and 
to. moist: surfaces .aigieaerenPe renee mies. siaecieciclcdeehee eee eae 361 


T. H.Morcan. An attempt to analyze the constitution of the chromosomes 
on the basis of sex-limited inheritance in Drosophila. One colored plate 
Gour: figures) ia 7sepeepreen cee oer eaoe ocr eer Ae ee Oe 365 


E. J. Lunn. On the structure, physiology and use of photogenic organs, 
with special reference to the Lampyridae. Nine figures................. 415 


Stewart Paton. Experiments on developing chicken’s eggs..:.............. 469 


INDEX 


Arenas and Asterias, On the behavior of 
the dissociated cells in hydroids, 281. 

Asterias, On the behavior of the dissociated cells in 
hydroids, Aleyonaria and, 281. 


Base: G. A., L. L. Wooprurr and. The 
reproduction of Paramaecium aurelia in a 
‘constan:’ culture medium of beef extract, 135. 

Baitset, G. A., L.L.Wooprurrand. Rhythmsin 
the reproductive activity of Infusoria, 339. 

Beef extract, The reproduction of Paramaecium aure- 
lia in a ‘constant’ culture medium of, 135. 

Branchipus gelidus, Migration of retinal pigment in 
the eyes of, 143. 


Csr muscle in embryos without nerves, The 
development and function of voluntary and, 
159. 

Cells in hydroids, Aleyonaria and Asterias, On the 
behavior of the dissociated, 281. 

Chickens’ eggs, Experiments on developing, 469. 

Cuixp, C. M. Studies on the d¥namics of morpho- 
genesis and inheritance in experimental repro- 
duction. IJ. Physiological dominance of an- 
terior over posterior regions in the regulation 
of Planaria dorotocephala, 187. 

III. The formation of new zodids in Planaria 
and other forms, 221. 

Chromosomes on the basis of sex-limited inheritance 
in Drosophila, An attempt to analyse the con- 
stitution of the, 365. 

Conjugation of Paramecium, Assortative mating, 
variability and inheritance of size, in the, 1. 

Culture medium of beef extract, The reproduc ion 
of Paramaecium aurelia in a ‘constant,’ 135. 


OMINANCE of anterior over posterior regions in 
the regulation of Planaria dorotocephala. II. 
Physiological, 187. 

Drosophila, An attempt to analyse the constitution 
of the chromosomes on the basis of sex-limited 
inheritance in, 365. 

Dynamics of morphogenesis and inheritance in 

experimental reproduction, Studies on the, 
187, 221. 


Vv 


Be worss to dry and to moist surfaces, The 
reactions of, 361. 

Eggs, Experiments on developing chickens, 469. 

Eyes of Branchipus gelidus, Migration of retinal 
pigment in the, 143. 


Hoes: Davenport. The development and 
function of voluntary and cardiac muscle in 
embryos without nerves, 159. 

Howranp, Ruta B. Migration of retinal pigment 
in the eyes of Branchipus gelidus, 143. 

Hydroids, Aleyonaria and Asterias, On the behavior 
of the dissociated cells in, 281. 


nFusoriA, Rhythms in the reproductive activity 

of, 339. 

Inheritance in Drosophila, An attempt to analyse 
the constitution of the chromosomes on the 
basis of sex-limited, 365. 

Inheritance in experimental reproduction, Studies 
on the dynamies of morphogenesis and, 187, 
221. 

Inheritance of size in the conjugation of Parame- 
cium, Assortative mating, variability and, 1. 


Enns. H. S. Assortative mating, variability 
and inheritance of size, in the conjugation of 
Paramecium, 1. 


AMPYRIDAE, On the structure, physiology and 
use of photogenic organs, with special reference 
to the, 415. 
Lunp, E. J. On the structure, physiology and use 
of photogenic organs, with special reference to 
the Lampyridae, 415. 


Miceatios of retinal pigment in the eyes of 
Branchipus gelidus, 143. 

Morean, T. H. An attempt to analyse the consti- 
tution of the chromosomes on the basis of 
sex-limited inheritance in Drosophila, 365. 


vl INDEX 


Morphogenesis and inheritance in experimental 
reproduction, Studies on the dynamics of, 
187, 221. 

Muscle in embryos without nerves, The development 
and function of voluntary and cardiac, 159. 


Peeectem, Assortative mating, variability and 
inheritance of size in the conjugation of, 1. 

Paramaecium aurelia in a ‘constant’ culture medium 
of beef extract, The reproduction of, 135. 

Parker, G. H. and Parsutey, H. M. The reac- 
tions of earthworms to dry and to moist sur- 
faces, 361. 

ParsHuey, H. M., Parker, G. H. and. The reac- 
tions of earthworms to dry and to moist sur- 
faces, 361. 

Parton, Stewart. Experiments on developing 
chickens’ eggs, 469. 

Photogenic organs, with special reference to the 
Lampyridae, On the structure, physiology 
and use of, 415. 

Planaria and other forms, III. The formation of 
new zodids in, 221. 

Planaria dorotocephala, IT. Physiological domin- 
ance of anterior over posterior regions in the 
regulation of, 187. 


EGULATION of Planaria dorotocephala, IT. Physi- 
ological dominance of anterior over posterior 
regions in the, 187. 


Reproduction of Paramecium aurelia in a ‘con- 
stant’ culture medium of beef extract, 135. 

Reproduction, Studies on the dynamics of morpho- 
genesis and inheritance in, experimental, 187, 
221. 

Reproductive activity of Infusoria, rhythms in the, 
339. 

Retinal pigment in the eyes of Branchipus gelidus, 
migration of, 143. 

Rhythms in the reproductive activity of Infusoria, 
339. 


Seeuncren inheritance in Drosophila, An at- 
tempt to analyse the constitution of the chro- 
mosomes on the basis of, 365. 


VV orentarr and cardiac muscle in embryos without 
nerves, The development and function of, 
159. 


Wo; H. V., On the behavior of the dissoct- 
ated cells in hydroids, Aleyonaria and 
Asterias, 281. 
Wooprurr, L.L. and G. A. Barrsett. Rhythms 
in the reproductive activity of Tnfusoria, 339. 
Wooprtrr, L. L. and G. A. Bartsety. The repro- 
duction of Paraemecium aurelia in a ‘constant’ 
culture medium of beef extract, 135. 


Zoers in Planaria and other forms. III. The 


formation of new, 221. 


ASSORTATIVE MATING, VARIABILITY AND INHERIT- 
ANCE OF SIZE, IN THE CONJUGATION OF 
PARAMECIUM' 


H. 8. JENNINGS 
From the Zoélogical Laboratory, Johns Hopkins University 


SIXTEEN FIGURES 


CONTENTS 

Problemsioutlimed ey. cresctecers rare erste rete eter eretetolelole Cieverolaistate re aheleteielsle\s @ eie' ele 3 

I. The facts as to relative size, variability, and assortative mating in con- 
UGH) Vise Soc chaeesoqueueEDpouseLocchboadicodeubr-acsoo Coo UpoOOOOIaE 6 
MIGHNOOE: <3 .aocdaabdGapSCsoagooen OE snOGo oduUSEUneOScd dd Cguect.oc. OsnnaeMEE 7 
Hundamentallmeasurementss)). 5 0... sic1c cc a tnseu s+ siecle cee cieleinie sie ssie we 9 
Relative size of conjugants and non-conjugants...........-.-6+++52seseeee 11 
: Relative variability of conjugants and non-conjugants.........-.....5--+5 16 
Causes of the lessened variability of conjugants ............+..-20.500005 17 
iGametredifferentiations assets eect toot teet retire Astle = «= 17 
PA JOG EIIt cholo lated iach GoDoriO DoGHnoOd nodes Saee aU cod do LoD aE eE gor 18 
Bi (Condo cocceoncoictobo paodo ond cee rEn en toon.caac scp soso.00 Cnoeoeee ols 
Increase in size and ceuability of the conjugants before fission. .... 20 
Am IVACtAINCITErENCeS he nes cee cease ech eias sia ase soi cietncrns as * os + 22 
ANOUK DIC WR op COA SeS GaGa RG sonnet Come on oon aodee eoneac Onn iene 23 
ReaniisvexplanahiOMerre tem citrc pei cis erect ststevate stleeietereia stsia. =o) oe =! 23 
Observation of the process of conjugation............-.-.-.eeeeeeee rere: 24 
Correlation in size in the members of pairs of conjugants.............--- 2 228: 
Meaning of the coefficients of correlation.....:........++eeeee eee eeeeeetee 28 
Correlation in different classes of cultures...............00eeee eee ee eee SH 
Wild cultures, of unknown racial composition..............5...05++ 5555 37 
Cultures containing pairs belonging to two different species..........-- 39 
Conjugation within pure races... 2. ..2c.22sscccc eet eres snes eens 41 
Mixturestof tworknoOwn-Tacess.-... 2-2-2200 s0s+s RAR Reno pCoC 44 
To what is due the incompleteness of correlation?...........---++++s55055 46 
1 Slight differences of less effect than great ones..........-----+-+55050 46 
2 Different categories of pairs following different rules........-..+.+-- 51 
Unevenness at the anterior ends.............. MRP teased o 8 nth Cac 52 


1 Third of a series of papers on Heredity, Variation and Evolution in Protozoa. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 1 
Jury, 1911 
il 


2 H. S. JENNINGS 


Causesvof the; correlation: {04-4 Meee seen: 3.) «oe oe «ae 60 
Lg pAssortative mating... cians an eos, eee. « = eee s (a0) 
a Correlation between the length anterior to the mouth and the 
total length 5.4 ae eet ects conte Ce ee ee Oe 61 
b High variability of the posterior region........................ 63 
ce Low correlation of parts before and behind mouth.............. 64 
d Correlation of pairs greater for parts anterior to mouth ........ 64 
e Low correlation of parts posterior to mouth.................... 64 
f Region in front of the point where the two conjugants separate... 65 
2 Equalization during mating............. taGinged eon ee Bs as 
Correlation of pairs after eepamion POEM He eM oe os Aap Stic. c 67 
Change in variability and correlation of parts after separation ..... 70 
Correlation of pairs not due to equalization....................... 73 
3 \Changelofisizeduning union’... sete: coe ceo eee eee 7: 
4 Differential contraction due to killing fluid......................... 77 
5 Local or temporal differentiations in the culture.................... 80 
6) Othersuggested causes a2 cc. -..c.2e Ree en eee oe noe Renee 82 
@ (@orrelationsin /breadthe.......s.cs.ceseeeee PR ODA SrA c 82 
a Flattening at the time of conjugation........................02- 83 
b Correlation in breadth in pairs after separation................ 84 
ec Correlation in breadth not due to equalization.................. 85 
Siebostoricalyandtcomparatlv.er sane eee eee ee Eee 85 
Se Conclusionsiontassortativie:matim ge...) veeee aera eric eee ea reielo sia 89 
II Consequences of the differentiation of the conjugants and of their assort- 
ative-mabin pieces ore en eet Eee eae oer 39 
Consequences of the decreased size and lessened Tariab ity SARS aD aac turne 90 
1 Are the extreme specimens excluded from the new generation?........ 90 
2 Relative size of progeny of conjugants and non-conjugants; changes in 
size due to conjugation.......... FS ANS | RPE eas 92 
a Comparison of progeny of conjugants and non-conjugants in pure 
0: "=< RE fo an ey a te cs ae iaeio LG eo toe shoneleHs Buccs eee merged 93 
b Comparison of progeny of conjugants and non-conjugants in a 
wildvcultures. 28.0 tc. Mae nes eenas ecnetet ate os vensters onteses:cesbareereecrs 97 
NED GLK ald: Grenier. nao ORE Tan Ratton Goma cer acta nator 100 
Cabixceptionalicase ac. nmecnia nin erie tie ase cee er 101 
CG ASUMIMAL Ys. -Spem tee ic rice Sree Te ere eel il Poe eae ena eaters Cleese: 103 
3 Increase in variability as a result of conjugation.................. .. 104 
4 Do pairs of different size give progeny of different size? .............. 104 
5, Unkeritancesiromiunequal pairs. see ateee meets eit: = 106 
Summary, eaysereee a Naa ROe Teor Or aac Un GUb aS dpobodjen coda auemeans cc 106 
Literature ‘cited sete rise vcce cet ee se i337 ee 109 
110 


Appendix: tables of eanrerienia Ee So fC Ads COUCUC ORE IEC Oo o 


CONJUGATION IN PARAMECIUM 3 


INDEX OF TABLES 


LE hee hot ete arene sactsuats re atare 9 1 he ere Pei oko ot ie en ree 62 
7 c.g SED PIDICES CRO DCE rae 10 19S EPR peers soe ate ss 69 
SMe ne nre ere secrefeia erator 11 PAD es 55 ac gr DRA Onn 72 
Caley Rear iG BRR DE EKO oS Corer: 12 he cv SST Bakes 73 
Seen gee Ontos avin sche Ceeeetd 12 PRE Ee rho S's RROEe 75 
Gs ete cticeistctny Stonsrarny Seapets: eos sire 17 D8 sow asg ae ee ete 76 
Ua PEERS Tech oeee erate ome alain 2 toate tes 21 yA. EE Grito aD SOCK 76 
IE Para aod an Dosh Gon naom nee 32 25)... i. 9G anes eee 7 
(2) An ene, ltr ho et ete eee ee ary eT 36 Ps hams doCo dae 84 
(Oitint o dace ombo toanor om com een 40 QE sss svatavgatee ES Ceres 84 
1b a teed Coa benhocAconnaonone 47 28). os cil. See ee 94 
ID eee sme srs ties crcnccerete ke tepercters 50 7 RREPAGAT nS hic 505 00S 95 
NSP Sey men rae peel. Reacanaveyovarete oie 54 BO wc saenick-s Scares ee ee 96 
VAN eee is) pepsi) sya iskrie Swear ein Ae 55 0) Pe ee co oct 97 
hk tar Aen AaB an Aeterna 56 Sores os ces aistotente-os aoe 99 
1G Lays Recodo bckciae ream 57 DO ele aaa ahuadta ares ats, coronas 102 
Le, RPI E eee it oe ns 59 34-68, Appendix.......... 110 


PROBLEMS OUTLINED 


The second paper of this series dealt with variation and inherit- 
ance of size in Paramecium during reproduction by fission. 
The present paper is an experimental and observational study 
of the size relations in conjugation. A later contribution will 
deal with the relation of conjugation to vitality and reproduction. 

Results presented with biometrical analysis have unfortunately 
come to incur in many quarters the suspicion that mathemati- 
cal treatment has been substituted for accurate experimental 
and observational investigation. The numerical analysis of 
results should of course be an addition, not a substitution; but 
the experimenter should realize that without this addition ex- 
perimental results may sometimes be as misleading as statistics 
without experimentation (“which is putting it strong’). Adequate 
experimentation with adequate numerical analysis is the ideal; 
toward this ideal my efforts, however short they may fall, have 
been directed in the present investigation. 

In taking up the relation of conjugation to genetic problems, 
I have thought it best to become acquainted at first hand, so 


4 H. S. JENNINGS 


far as possible, not only with matters that have not been hitherto 
treated, but even with those that have been dealt with by previous 
investigators, in the latter case confirming or criticizing their 
results. This does not imply a precedent suspicion as to the 
accuracy of the work thus gone over; it is done only in pursuance 
of a general policy, for one often finds matters of great import 
where they are least expected. Furthermore, the recent dis- 
covery of the existence of many diverse races in Paramecium 
makes it needful to reéxamine many phenomena in relation to the 
part played in them by these different races. In any case, in 
this difficult field independent confirmation of another’s results is 
decidedly worth while. 

It will be well to set forth here an outline of the questions 
with which a thorough investigation of the size relations in con- 
jugation would have to deal. 

To Pearl (07) we owe the discovery of certain most interest- 
ing relations between the conjugating individuals of Paramecium. 
By an elaborate statistical investigation he showed (1) that the 
conjugants of a culture of Paramecium are much less variable 
than the non-conjugating population, and have (as had before 
been noticed) a smaller mean size; (2) that there is a marked 
degree of correlation in size between the members of pairs in Para- 
mecium; smaller individuals being found mated with smaller, 
larger with larger. With these important matters, particularly 
in their relation to the existence of diverse races, we shall have 
to deal thoroughly. The precise questions here are as follows: 

1. What are the facts as to the relative variability and size 
of conjugants and non-conjugants, and what is their relation to 
the existence of races of diverse size? 

2. What are the facts as to assortative mating; its determin- 
ing conditions, its peculiarities and limitations; its relation to 
the existence of diverse races? 

3. What are the results, in inheritance, variability and evo- 
lution, of the smaller size and decreased variability of the conju- 
gants, as compared with the non-conjugants? If we breed from 
a number of the conjugants, do they give progeny that are (a) 
smaller, or (b) less variable, than the progeny of the larger, more 


CONJUGATION IN PARAMECIUM oO 


variable non-conjugants? Does conjugation thus act as a pro- 
cess of excluding from the line of evolution individuals that vary 
from the usual size? 

4. What are the results, in inheritance, variability and evolu- 
tion, of the assortative mating of Paramecium? If we isolate 
(a) large, and (b) small, pairs of conjugants, keeping them under 
identical conditions, do they produce, respectively, large and 
small races? Is there any difference between the progeny of 
(a) pairs in which the two members are equal, and (b) pairs in 
which the two members are unequal? 

5. What relation has conjugation as a physiological process 
to the size of the individuals of the stock undergoing it? Does 
the size differ characteristically in different parts of the life cycle, 
as is sometimes set forth? Are the individuals at the ead of the 
life cycle (just before conjugation) larger or smaller than those 
at the beginning of the cycle (just after conjugation)? 

6. What are the facts of inheritance in conjugation? If the 
two conjugants of a pair differ, are the progeny of these two 
conjugants alike and intermediate between the two? Or will 
for example the larger member continue to produce large individ- 
uals like itself, the small one small individuals like itself? Or is 
there some third possibility? The laws of inheritance have never 
been worked out for this peculiar reciprocal fertilization, where 
both parents may continue reproduction. 

These questions we shall take up in detail. On most of them 
I hope to present data of importance, though on the extremely 
important problem last raised I have as yet been unable to get 
clear results on some of the points of greatest interest. 

In dealing with most of these questions, the existence of diverse 
races of Paramecia, as set forth in former papers,? will be found 
of extreme importance. In Paramecia multiplying by fission 
there are many races or lines, differing in size and in other respects. 
A considerable number of these races were isolated; the mean 
length of the largest being more than double that of the smallest, 
with many intermediate races. As will be recalled, each race 


2 See Jennings ’08, and Jennings and Hargitt ’10. 


6 H. S. JENNINGS 


shows within itself many variations, due to differences in growth 
and environmental action, but these variations within the race 
are notas arule inherited, and under the same conditions of growth 
and environment the race is uniform and constant. In all work 
with conjugation, the question whether we are dealing with a 
pure race or with a mixture of races is of the greatest importance; 
the significance of the phenomena observed is quite different 
in the two cases. 


I. THE FACTS AS TO RELATIVE SIZE, VARIABILITY AND ASSORTA- 
TIVE MATING IN CONJUGATION 


The data given by Pearl (’07) would seem amply sufficient to 
show that in a conjugating culture the conjugants are smaller 
and less variable than the non-conjugant population, and that 
there is a high degree of assortative mating in Paramecium. A 
further study of the facts is needed, however. On the one hand 
it is desirable that Pearl’s interesting results should be confirmed 
by independent observation, or refuted. The correctness of 
some of his results has been called in question (Lister ’06, Pear- 
son ’06, Pearl ’07). Further, there are a number of conditions 
not dealt with by Pearl that might produce a correlation in size 
between the members of pairs; these need to be subjected to 
experimental test. Beyond this, many important relations in 
this matter remain as yet unknown; we need a knowledge of the 
variations and limitations of assortative mating, of the conditions 
on which it depends, of its relation to the existence of diverse 
races, and particularly of its consequences in the later history of 
the stock. 

I have therefore examined and made measurements of a num- 
ber of conjugating cultures with reference to these matters. After 
a number of cultures had thus been studied, it became necessary 
to test by comparative examination of cultures under controlled 
conditions, one after another, many factors that suggested them- 
selves as possibly producing the observed relations (particularly 
the correlation between members of pairs). Asa result, the quan- 
tity of material available for study of these matters becomes 


CONJUGATION IN PARAMECIUM Ui 


very great, consisting of more than thirty lots, averaging more 
than one hundred pairseach. Some of these lots were ‘wild’ 
cultures, containing a number of diverse lines or races, belonging 
in some cases all to caudatum or all to aurelia; in other cases 
belonging partly to caudatum, partly to aurelia. Other lots 
consisted entirely of members of a single race or ‘pure line,’ 
having descended from a single individual; other lots consisted 
of mixtures of two known races. The relations observed are 
naturally somewhat different in these diverse cases. I have 
dealt mainly with the measurements of length, since it is here 
that the phenomena of primary importance appear; certain stud- 
ies of the breadth relations will however be found on later pages 
(table 20). 


METHODS 


The special methods for the diverse experiments will be men- 
tioned in the course of the paper. Here it will be well to mention 
mainly the methods of killing and of measurement. 

The animals to be measured were brought into a drop of water 
at the bottom of a watch-glass, then overwhelmed with the kill- 
ing fluid. For killing I used mainly Worcester’s fluid (10 per cent 
formalin saturated with corrosive sublimate). I have later found 
that chrom-osmic acid (1 per cent osmic in | per cent chromic) 
has advantages in some respects. Both these fluids kill without 
distortion if properly used. The animals were measured either 
in the killing fluid, or after transference to water or to 25 per 
cent glycerine. Careful comparative measurements before and 
after transference showed that no change is made by placing in 
water or weak glycerine. The most satisfactory method is to 
remove with pipette a portion of the killing fluid from the watch 
glass, then to add 25 per cent glycerine; in this the specimens 
are kept till measured. With the Worcester’s fluid there is some-_ 
times an objectionable deposit of fine crystals, in the course of 
time; this does not happen with the chrom-osmic. 

In the early part of the work the animals were measured on the 
slide, with the ocular micrometer. This becomes very wearing 


8 H. S. JENNINGS 


on the eyes; later the measurements were made by the aid of the 
Edinger drawing and projection apparatus, which cannot be too 
highly recommended for the purpose. The animals are projected 
much enlarged, on the drawing board, where they are measured 
with a millimeter ruler. I used a magnification of 500 diameters, 
so that each millimeter of the ruler corresponded to 2 microns 
(0.002 mm.). The best method I found to be as follows: the ani- 
mals were placed on a thin slide in a flat drop of the 25 per cent 
glycerine, with no cover (so that there was no danger of flatten- 
ing), projected, and measured. Without the glycerine in the 
fluid this method cannot be used, owing to the convection cur- 
rents and the rapid evaporation produced. 

In the measurements of conjugants the unit of grouping was 
4 microns, so that each group in the tables contains individuals 
varying from 2 microns below to 2 microns above the dimen- 
sion at the head of the column or row. In the original measure- 
ments, in many cases, the unit employed was but 2 microns. 

The constants of variation were computed according to the 
methods and formulae set forth in my paper of 1908 (page 397). 
In the present paper however we are dealing with cases where 
the two things to be compared (the-two members of a pair) are 
alike, so that either one might be entered in the rows or in the col- 
umns of the correlation table. In such cases double or symmet- 
rical tables have commonly been employed. In a recent note 
(11) I have shown that this is unnecessary, and that the compu- 
tations are performed with much less labor without the use of 
symmetrical tables. The method of computation set forth in 
this note was used in the present paper. In accord with this, | 
have formed the tables of correlation by entering in every case the 
larger member of the pair in the vertical columns, the smaller in 
the horizontal rows. 


CONJUGATION 


IN PARAMECIUM 


FUNDAMENTAL MEASUREMENTS 


Table 1 gives the important constants for the length of con- 
jugants as compared with non-conjugants in a number of cul- 
tures developed in material brought into the laboratory from 
ponds or pools, so that the racial composition is unknown. Table 
2 gives the same constants for cultures consisting entirely of a 
single ‘pure line’ or race,—all the individuals being derived from 
the fission of a single one; also those for certain mixtures of known 
Table 3 gives the constants for a number of lots of con- 


races. 


TABLE 1 


Constants of variation in length for conjugants and non-conjugants of Paramectum, from 


wild cultures, of unknown racial composition. 


(The original measurements of length 


for all these will be found in table 34 of the appendix; the tables of correlation for the 
conjugants, in the appendiz, are indicated in the column headed ‘table’ ) 


tw 


~ 


ou 


wo 


LOT 


DATE 


Dec. 13, 


Mar. 21, 


Mar. 26, 


Feb. 17, 


\Jan. 29, . 


Apri) ost 


Nov. 4, ’07 | 
June 20,’09 
07 


"08 


"08 


08 


| 1 D 1 fu & i 
| 1535 BA zis ill uae eS | fEze 
Bl pe ras AS Bra Bg | faae 
RBZ q| aoe a ags Sipe Beege 
leeles|a|eee| 32 Z88 2 | Seas 
all aaa teed cei ae Bee 8 oa 
A. ‘‘Wild” cultures | | 
a. Conjugants........ | 360 180) 11 148-260199.02+0.54 15.28+0.38 7.68+0.190.39840.030 
b. Non-conjugants ..| 360) 34 132-320 222.86+0.87 24.57+0.62 11.03+0. 28 
a. Conjugants........ | 284 142) 36 130-206 165.04+0.53) 13.20+0.37, 8.00+0.23/0.268+0 057 
b. Non-conjugants.... 87] 34) 134-198 164.12+1.00 13.84+0.71 8.43+0.43) 
a. Conjugants........ 164, 82) 37 164-236196.100.71) 13.54+0.50) 6,91+0.260.507+0.049 
b. Non-conjugants.., 176 | 34 164-288 228.36+1.23) 24.20+0.87| 10.60+0.39 
a. Conjugants ....| 84) 42) 38 116-160/139.29+0.70) 9.56+0.50) 6.86+0.36(0.499+0.055 
b. Non-conjugants ...) 152 34 100-244155.40+1.39 25.42+0.98 16.36+0.65 
a. Conjugants, de- 
scended from 4-a.., 16, 8 34 116-144130.50+1.33) 7.90+0.94 6.05+0.72 
b. Non-conjugants 
descended from 4-a., 100 34 104-200 143.80+1.29 19.08+0.91 13.27+0. 64 
a, Conjugants........ | 272) 136] 39 128-216181.49+0.54) 13.32+0.39 7.34+0 21/0.428=0 033 
b. Non-conjugants ...| 318} 34) 132-248)/186.10+0.79| 20.97+0.56) 11.27+0.31) 
fa. Conjugants........ | 158| 79| 40) 128-204 168.71-0.63| 11.66+0.44 6.91-+0.26(0.333+0.048 
| (b. Non-conjugants ..) 131) | 34) 148-224'182.20+1.03| 17.43+0.73) 9.57+0.40) 
| B. Descended from se-| | 
lected parts of wild) | 
cultures } 
2a.Conjugants de- 
scended from 10 
BM oa nocoods 54) 27) 42 120-152134.89+0.64 7.02+0.46 5.20+0.340.612+0.057 
b. Non-conjugants, | 
( froin same......... 56 35) 124-168148.93+0.98, 10.89+0.69) 7.31+0.47 


10 


H. 8S. JENNINGS 


TABLE 2 


Constants of variation in length for conjugants and non-conjugants of Paramecium, from 
cultures of pure races, descended from a single individual or a single pair. or from 


mixed cultures of known racial composition. 


(Lhe measurements of length for these will 


be found in table 25 of the appendix; the tables of correlation for the conjugants, in the 


appendix, are indicated in the column headed ‘table’ ) 


| 


Lil 4 ee le ees ; 
Jaen Bei | cau eecemulncees 
S28 | |Fza] 29 ae |) ae Z2h2 
Paar) ra AS Hz wp Be Bean 
Helge ,|a5¢| 73 | 382 | ES | SBEee 
a 3 SelB) Cees) Ae zee | 28 ESfa 
5 < Z papa 2 o<2 ae oi) ay Boman 
a a fof a 2 |B| & a nD o is) 
A. Pure lines from 
| one individual } 
| | (all aurelia) 
alsapes 28,071 - Conjugants...... 250, 125 43) 120-180 150.50+0.48, 11,130.34, 7.39+0.22) 0.132+0.042 
tay ib. Non-conjugants 200 35] 124-200 188.80+0.88 18.38+0 62 11.58+0. 40 
10 Feb. 20, 08 z Conjugants...... 52, 2644) 112-144 127.23+0.67, 7.14+0.47  5.61+0.37\—0. 193+0.090 
es b. Non-conjugants | 43 —35| 120-160 140.19+0.99 9.65+0.70, 6.89+0.50 
| 
11 Feb. 26, '08) i k. Conjugants...... | 98) 1445 06-136 116.71+1.13 8.87+0.80 7.60+0.69|—0.137+0.125 
Forenoon b. Non-conjugants | 100 35 96-152 133.68+0.79 11.64+0.56 ».70+0.42 
12\Sept. 12,'08) & tle Conjugants...... 156 78 46 104-140 124-08+0.41 7.51+0.29 6.05+0.23) 0.367+0.047 
Se eae eel b. Non-conjugants | 100, 35) 104-180 143.72+0.96| 14.25+0.68| 9.91+0.48} 
| | | 
Afternoon k a. Conjugants...... 336 168,47 92-156 129.58+0.40 10.96+0.29) 8.46+0.22) 0.184+0.036 
13 Sept. 12, '08) b. Non-conjugants 100 35] 85-168. 140. 20+0.97 14.35+0.68, 10.23+0.49 
1alMar. 31, 08 Nfs (/" Conjugants...... 42, 21,35] 124-148 136.95+0.58 5.53+0.41) 4.03+0.30| 0.295+0.095 
i piak iw | b. Non-conjugants 34 BE 132-168 144.59+0.92 7.92065 5.48+0.45 
islApr. 4 “08 WE a. Conjugants...... | 50 25.25) 116-148 132.88+0.64 6.66+0.45, 5.01+0.34) 0.257+0.089 
Paget * \'b. Non-conjugants | 31, —(/35| 120-188 147.61+2.20| 18,181.56) 12.31+1.07 
1eSept.14, "08 Cs a. Conjugants...... | 300 150,48} 100-160 128.67+0.47  11.97+0.33 9.30+0.26] 0.507+0.029 
ert * |b. Non-conjugants 100 35 96-168 134. 2041.04 15.37+0.73, 11.45+0.55 
17Sept. 16°08 Cp {2 Coniugants...... | 188 6949 88-148 121.91+0.66 11.46+0.47 9.40+0.39) 0.318+0.052 
AOE eres Me lb. Non-conjugants 110 [35] 104-164.132. 180.87) 13.53+0.62) 10.2340. 47 
| | | 
Forenoon | If Conjugants...... | 168, 8450) 104-152 123.57+0.41) 7.80+0.29 6.31+0.23) 0.251+0.049 
1gSept. 25, '08 7 th. Non-conjugants | 100 [35] 88-180 131.32+1.21) 17.97+0.86 13.680. 67 
| | | | 
Afternoon | a. Conjugants. ..... | 174 87)51| 96-152 118.28+0.£2) 10.11+0.37, 8.55+0.31! 0.323+0.046 
19 Sept. 25, 08 te. Non-conjugants | 118 35) 88-180 135.35+1.02 16.49+0.72) 12.18+0.54 
| B. Mixtures of two | 
| known races | | 
aiieaens 108 C2 4. / 8 Conjugants are 124 6253) 100-156 129.58+0.62 10.31+0.44, 7.95+0.34) 0.115+0.060 
la |/b. Non-conjugants | 149/35) 84-176 122. 44+1.37, 24.71+0.97 20.18+0.82 
SNe ie ns ta+k [ Conjugants...... 98 4954) 108-136 120.25+0.46 6.68+0.32 5.56+.27| 0.4080. 064 
ia jeans Sean b. Non-conjugants. 156 ia 104-264173.10+2.25 41,671.59) 24.07+0.97 
jugants where the corresponding non-conjugants were not 


examined. ‘The original measurements on which these constants 
are based will be found in the tables of the appendix; the more 
important ones in tables 34 and 35. 


CONJUGATION IN PARAMECIUM J] 


TABLE 3 


Constants of variation in length for a number of lots of conjugating Paramecia in which 
the non-conjugants were not measured. The column headed ‘table’ gives the number of a 
table to be found in the appendix, in which the distribution of the measurements is 
shown. The measurements are here given in microns 


, 4 he & fa fe 
z Fe E 3 OF EO 
pb = 5 z z = 
Satemlonlt gj ag go Rake 
jog B| ee é 4 me Bie Baga 
| ja alag| . ag << Ss | 88pea 
Q asians) & On z a Be Beads 
g| & Balpa|&| 3 F 23 Bs | aSaza 
3 | A 2 \% | a| # 2 a 8 8 
| 
| | | | 
| 1910 | | 
22 Aug. 31 Wild culture, caudatum) 204) 102 55 | 152-208176.14+0.48 10.08+0.34 5.72+0.190.359+0.041 
| | 22 
| | | ing 
23} Sept. 21 ‘Wild, caudatum ...... 296) 148.and) 152-208179.80+0.41 10.42+0.29 5 80+0.160.245+0.037 
| | 23 | 
24 Sept. 11 |/Racek (aurelia)........ | 244 122 59 100-144.118.92+0.33, 7.54+0.23 6.34+0.150.210+0.041 
26 Sept. 11 Mixed caudatum ca | 
erauitvel tals rereiyeerareretereserete | 62) 31 84-200145.81+2.44 28 491.73 19.54+1.230.939+0.010 
27 Sept. 13 Mixed caudatum and) | 
BUTELIAUS ascs scheint 340, 170 10 108-236161.19+1.27 34.70+0.84 21.53+0.540.940+0.004 


With certain exceptions, to be considered later, the results given 
in these tables confirm Pearl’s results (1) that the conjugants are 
smaller than the non-conjugant population of a culture; (2) that 
they are less variable than the non-conjugants, and (3) that there 
is a marked positive correlation in size between the members of 
the pairs, so that on the whole larger individuals are found mated 
with larger, smaller individuals with smaller. We shall take up 
these matters separately. 


RELATIVE SIZE OF CONJUGANTS AND NON-CONJUGANTS 


If we examine in the seven ‘wild’ cultures of table 1 the relative 
mean lengths of the conjugants and non-conjugants of a culture, 
we find, as shown in table 4, that in every case save one the con- 
jugants are smaller than the non-conjugants. In lots 1, 4 and 5, 
the difference between the means for the conjugants and non- 
conjugants is about 10 per cent of the mean length of the non- 
conjugant population; in lot 3 it is 14 per cent. But in lot 6 
the difference is shght, being but 2.5 per cent of the mean for the 


12 H. S. JENNINGS 


non-conjugants, and in lot 2 the mean length of the conjugants 
is actually the greater, by a very small amount, though here the 
difference is not significant in comparison with the probable error. 
In this culture then the conjugants are not perceptibly differ- 
entiated in size from the non-conjugants. 

In the four lots studied by Pearl (07) the conjugants were in 
all cases smaller than the non-conjugants, by amounts varying 
from 11.5 per cent to 16.4 per cent of the mean of the latter, and 


TABLE +4 


Differences in length between conjugants and non-conjugants of wild cultures. (The 
‘relative difference’ in the fourth column shows what percentage the difference is of 
the non-conjugant mean) 


NON-CONJUGANT 


LOT MEAN CONJUGANT MEAN ABSOLUTE DIFFERENCE RELATIVE DIFFERENCE 
Per cent 

1 | 222.86+0.87 199.02+0.54 23.83+1.02 10.7 

2 164.12+1.00 165.04+0.53 | —0.93+1.13 —0.54 

3) 228.36+1.23 196.10+0.71 | 32.27+1.42 | 14.1 

4 155.40+1.39 139.29+0.70 | 16.11+1.56 | 10.4 

5 | 143.80+1.29 130.50+1.33 | 13.30+1.85 9.2 

6 | 186.10+0.79 181.49+0.55 | 4.61+0.96 2.5 

7 | 7.4 


182.20+1.03 168.71+0.63 


13.49+1.20 


TABLE 5 


Differences in length between conjugants and non-conjugants of pure races and 
mixtures of pure races 


Per cent 
9 c 158.80+0.88  150.50+0.48  8.30+1.00 5.2 
10 k 140.19+0.99 | 127.23+0.67 12.95+1.20 9.2 
11 k 133.68+0.79 | 116.71+1.13 | 15.97+1.38 11.9 
12 k 143.72+0.96  124.08+0.41  19.64+104 | 13347 
13 k 140.20+0.97 | 129.58+0.40 | 10.62+1.05 7.6 
14 Nfo 144.59+0.92 136.95+0.58| 7.64+1.08 5.3 
15 Nf 147.61+2.20  132.88+0.64 | 14.73+2.29 10.0 
16 C2 134.20+1.04  128.67+0.47) 5.5341.14 Ae 
17 C2 132.18+0.87  121.91+0.66 10.27+1.09 WES 
18 g 131.32+1.21 123.57+0.41 7.75+1.28 5.9 
19 g 135.35+1.02  118.28+0.52) 17.08+1.15 12.6 
20 Co4t 122.44+1.37  129.58+0.62 —7.14+1.50 —5r8 

30.6 


21 Lak 173.10+2.25  120.25+0.46 52.86+2.30 


CONJUGATION IN PARAMECIUM 13 


it has been practically the universal testimony of observers that 
the conjugants are smaller than those not conjugating. Our 
own results, as we have seen, confirm this for most cases, but not 
for all. How is the fact to be accounted for that in some cultures 
the conjugants are not smaller? 

Light on this question will best be obtained by examining the 
relative sizes of conjugants and non-conjugants in cultures com- 
posed of pure races, and in mixtures of known racial composition. 
The data are given in table 2 and in table 5. In all the eleven 
cases of table 2 in which we can compare the conjugants and non- 
conjugants of a pure race, we find the conjugants smaller, by 
amounts varying from about 4 per cent up to more than 12 per 
cent, of the mean for the non-conjugant population. All of 
these are races of aurelia, as I had no opportunity to make a 
careful study of the conjugants of a pure race of caudatum. But 
the fact that in wild cultures consisting mainly if not entirely of 
caudatum, as was the case with all of Pearl’s material, and of 
our lots 1,3, 4and 5, the conjugants are as a rule markedly smaller 
than the non-conjugants, indicates strongly that this would 
hold generally for caudatum also. We may then take it as 
established for aurelia, and practically so for caudatum, that 
within any given race the conjugants average smaller than the non- 
conjugants. Why in some wild cultures the conjugants may not 
be found smaller is shown by examination of the data for our 
mixed cultures (lots 20 and 21, table 2). Lot 20 consisted of a 
mixture of two races of aurelia,i and (>. The race 7 was smaller, 
averaging usually about 100 microns in length, while the usual 
mean for C, was about 125 microns. When conjugation took 
place in this mixture, the conjugants were all of the size charac- 
teristic for the conjugants of C, (as shown by comparing lots 16, 
a and 20, a of table 2), measuring 129.58 microns. Conjugants 
in a pure race of 7 had been found to be much smaller, varying 
from 92 to 98 microns in length. Thus it was clear that in the 
mixture only the race C, was conjugating, and the measurements 


1 For measurements of these races uader various conditions, see Jennings ’08, 
and Jennings and Hargitt 710. 


14 H. S. JENNINGS 


for the conjugants are of that race alone. But the random sample 
of non-conjugant population contains representatives of both 7 
and C2, and its mean size (122.44 microns), therefore lies between 
that of the two races. It is therefore less than that of the con- 
jJugants. The conditions in this case are illustrated in fig. 1. 
They are well brought also by acomparison of the measurements of 
the conjugants and non-conjugants of lot 20, as given in table 35. 


Al 
JIN 


/ i 


Fig. 1 Typical group of specimens from a culture consisting of a mixture of 
the large race C2 and the small race 7 (both aurelia). Conjugating pairs all (>. 
X 333. 


The reverse condition is given by the mixture of LZ. and k 
(lot 21, tables 2,5 and 34). Here the conjugation was only in the 
smaller race k, while the non-conjugant sample includes also many 
of the large race Z2, As a result the conjugants are very much 
smaller than the mean for the mixture as a whole, the difference 
being 30.6 per cent this mean. 


CONJUGATION IN PARAMECIUM 15 


These cases show that when more than one race is present in 
a mixture, the members of one race alone may conjugate. If this 
is a large race, the mean size of the conjugants may be equal to 
that of the population as a whole, or even larger than this. This 
is doubtless the explanation for lots 2 and 6, tables 1 and 4; here 
we are dealing with cultures of unknown racial composition. As 
a matter of fact I did, by selection and propagation, isolate a 
number of races of diverse size from lot 6; from it came the races 
k and Ls, as well as a number of others. 

Thus when we are dealing with a single race the conjugants 
are always smaller than the non-conjugant population, by amounts 
varying from 4 per cent to 13 per cent (or more) of the mean 
for the latter. The variation in the proportional difference be- 
tween the two in different cases is readily accounted for the fact 
that sometimes multiplication is occurring during an epidemic 
of conjugation, at other times not. In the former case many 
young will be present, reducing the mean length for the non- 
conjugants, but not affecting that for the conjugants. It is to 
be noted however that the mean size of the conjugants may differ 
a certain amount under different conditions in the same race. In 
the race k the mean size at the epidemic of February 26, 1908, 
was 116.71 microns (table 2, lot 11) while at the epidemic of Sep- 
tember 12, 1908, it was, in the afternoon, 129.58 microns (lot 13), 
a difference of 11.02 per cent of the smaller mean size. This is 
the greatest difference in mean size observed between the con- 
jugants of a given race. It is to be observed that this difference 
is not one arising progressively over long periods, for but a week 
before the minimum, the conjugants of this same race showed 
practically the maximum size (compare lots 10 and 11, table 2). 
The difference is undoubtedly due to the varying nutritive condi- 
tions at the time of conjugation. 

But in cultures of unknown racial composition, the conjugants 
may be very much smaller than the average for all (as in lot 21), 
or they may be equal to or larger than the average for all, depend- 
ing on the relative size of the races present, and upon which of 
these races the conjugants belong to. 


16 H. S. JENNINGS 


The consequences in heredity of the decreased size of the con- 
jugants will be taken up later. 


RELATIVE VARIABILITY OF CONJUGANTS AND NON-CONJUGANTS 


Examination of tables 1 and 2 confirms further Pearl’s dis- 
covery that the conjugants are not only smaller, but also less 
variable than the non-conjugant population. A comparison of 
the original measurements for the conjugants and non-conju- 
gants, as given in tables 34 and 35 of the appendix, renders this 
difference in variability at once evident to the eye. In every one 
of the cases given by these tables, both the absolute variation 
(as shown by the standard deviation) and the relative variation 
(as shown by the coefficient of variation) are less in the conju- 
gants. In a few cases the difference is but slight and would 
perhaps be hardly significant, taking each of these cases sepa- 
rately, in comparison with the probable errors. But the fact 
that it is always the conjugants which show the lesser variability 
is very significant, especially when we consider the much more 
numerous cases in which the variability of the conjugants is much 
less than that of the non-conjugants. 

The difference between the variability of conjugants and non- 
conjugants itself varies greatly in different cases; in other words, 
sometimes the conjugants are but little less variable than the 
non-conjugants, while in other cases they are very much less vari- 
able. This is exhibited in table 6. In some of the wild cultures 
the coefficient of variation of the conjugants is but 5.1 per cent 
less than that of the non-conjugants (lot 2, where the difference is 
indeed of no significance in comparison with the probable error) ; 
from this minimum it varies uv to a difference in lot 4 of 58.1 
per cent,—the coefficient of variation for the conjugants being 
less than half that for the non-conjugants. In the pure races 
the least difference between the coefficients for the conjugants 
aod non-conjugants is 8.1 per cent of that for the non-conjugants 
(lot 17), rising to 59.3 per cent in lot 15. If we make an average 
of these numbers showing the difference in variability for the eleven 
lots of these races in table 6, we find it to be 33.01 per cent. 


CONJUGATION IN PARAMECIUM 17 


CAUSES OF THE LESSENED VARIABILITY OF CONJUGANTS 


What is the cause of this lessened variation among the conju- 
gants? A uumber of different possible factors may be considered. 


1. Gametic differentiation 


Lister (06) suggested that the cause was as follows: The 
conjugants are differentiated gametes. In measuring them we 
are dealing only with this particular class, while in the non-con- 
jugant population we include many gametes, many in the process 
of differentiation into gametes and many that are not gametes; 
hence the non-conjugants are a heterogeneous lot and must give 


TABLE 6 


Difference in variability between conjugants and non-conjugants 


STANDARD DEVIATION COEFFICIENT OF VARIATION 

LOT | RACE ; a 7 ‘ ian 

Non- ; Abs. Rel. dif- Non- - Abs. Rel. dif- 

conjugants| CoMugants difference | ference | conjugants | Conjueants difference ference 

A. Wild cultures 
| | | % I | | % 
1 | 24.57+0. 62) 15.28+0.38 9.29+0.73) 37.8 11.03+0 28) 7.630. 19) 3.35+0.34 30.4 
2 13.84£0.71| 13.20+0.37, 0.64+0.80) 4.6 8.43+0.43) 8.00+0.23) 0.434049) 5.1 
3 | 24.20+0.87| 13.54+0.50 10.66+1.00 44.0 10.60+0.39) 6.91+0.26) 3.69+0.47 24.8 
4 | 25.42+0.98 9.56+0.50 15.86+1.10 62.4 16.36+0.65) 6.86+0.36 9.5040 74 58.1 
5 | 19.08+0.91) 7.90+0.94 11.18+1.3i; 59.0 i 13.27+0.64 6.05+0.72) 7.22+0.96 54.4 
6 | 20 970.56 13.32+0.39 7.65+0.68 36.5 | 11.270 31) 7.34+0.21) 3.93+0.37) 34.9 
7 17.43+0.73| 11.66+0.44 5.77+0 85, 33.1 9.57+0.40) 6.91+0.26) 2.65+0 48) 27.7 
8 10.89+0.69) 7.02+0.46 $.87=0:83) 35.5 7.31+0.47, 5.20#0.34) 2.1140.58 28.9 
B. Pure races 
9 } ¢ 18.38+0.62| 11.13+0.34, 7.45+0.71) 40.6 11.58+0.40| 7.39+0.22) 4.19+0.46 36.2 
10 | k 9.65+0 70) 7. 140.47) 2.51+0.84 25.9 6.89+0.50) 5.61+0.37) 2.28+0.62, 33.1 
11 k 11.64+0.56) 8.87+0.80) 2.77+0.98) 23.8 || 8.70+0.42) 7.60+0.69) 2.10+0.81) 24.1 
12 | k 14.25+0.68 7.51+0.29) 6.74+0.74 47.3 | 9.91+0.48) 6.05+0.23) 3.86+0.53, 39.0 
13 k 14.35+0.65) 10.96+0.29) 3.39+0.74 23.7 | 10.23+0.49 8.46+0.22) 1.77+0.54) 17.3 
14 | Nfz | 7.92+0.65) 5.53+0.41) 2.39+0.77 30.2 5.48+0.45) 4,030.30) 1.45+0.54! 26.5 
15 | Nfe | 18.18+1.56) 6.66+0.45 11.52+1.62, 63.4 || 12.31+1.07) 5.01+0.34) 7.3041.12 59.3 
16 Ca | 15.27+0.73) 11.97+0.33) 3.40+0.80) 22.1 11.45+0.55) 9.30+0.26 2.15+0.61 18.8 
17 C2 | 13.53+0.62) 11.46+£0.47| 2.07+0.78) 15.3 10.23+0.47| 9.40+0.39) 0.83+0.88 8.1 
18 | g 17.97+0.86 7.80+0.29 10.17+0.91) 56.6 13.68+0.67) 6.31+0.23) 7.37+0.71, 53.9 
19 | 9 | 16.49+0.72) 10.11+0.37, 6.38+0.81) 38.7 12.18+0.54) §.55+0.31) 3.53+0.62) 29.0 
C. mixed 

20 | Cr4s) 24.71£0.97) 10.31+0.44| 14.4041.07) 58.3 | 20.1840 82, 7.95-+0.34 12.23+0.99| 60.6 


21 La+k| 41.67+1.59| 6.68+0.32) 34.99+1.62) 84.0 24.07+0.97 5.56+0.27 18.511.01) 76.9 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 1 


18 H. S. JENNINGS 


a greater coefficient of variation. This suggestion contained 
perhaps the germ of a correct explanation, but erred in empha- 
sizing a supposed differentiation of the gametes from ordinary 
adults. Pearl showed that in cultures which are not conjugating, 
nor near a period of conjugation, the coefficients of variation are 
as great as for the general population of those containing conju- 
gants (’07, p. 231), and my own extensive data (’08) confirm this 
fully. In such cultures there are no gametes and no specimens 
in the process of differentiation into gametes, so that heterogene- 
ity on this score cannot account for their greater variation as 
compared with the conjugants. 


2. Equalization 


Pearl (’07, p. 262) discusses the possibility that a process of 
equalization has occurred during conjugation; ‘“‘that the pro- 
conjugants were simply a random sample from the general pop- 
ulation having equal variability with it,” and that the decrease 
in variability is due to ‘‘a pronounced tendency toward equali- 
zation in size of the two members.’”’ He adduced evidence to in- 
dicate that such a process of equalization does not occur to any 
considerable degree. A tendency to equalization does exist, as 
we shall see later, but (as will later appear) it is not of such a 
character or degree as to account for the greatly decreased varia- 
bility and smaller size of the conjugants. Moreover, as we shall 
immediately see, such an explanation is gratuitous, since there is 
a fully adequate explanation on other grounds. 


3. Growth 


In my paper of 1908, I showed that a large proportion of the 
variation in an ordinary culture of Paramecium is due to growth. 
A random sample of the population includes young individuals, 
that are very small; individuals in all stages of growth up to the 
largest sizes and individuals that have again decreased in length 
preparatory to fission. Now, the conjugants include neither the 
young, small individuals, nor the largest ones. Hence they show 


CONJUGATION IN PARAMECIUM 19 


much less variation than the population as a whole. The same 
thing would be found true (in possibly a less degree) for man or 
any higher animal; mated couples would be found on the whole 
less variable than a general sample of the population that included 
children. In one respect the condition in the infusorian is 
peculiar; the conjugants do not grow so large as the individuals 
that are to undergo fission without conjugation. Thus the conju- 
gants represent a rather definite, limited stage of growth, exclud- 
ing the extremes at both ends. That this is fully sufficient to 
account for the lesser variability of the conjugants is demonstrated 
by the fact that non-conjugants at a definite growth stage show as 
little variability as do conjugants. The coefficients of variation 
in length for conjugants range as a rule from 5 to 8 per cent, as 
shown in tables 1 and 2, and in Pearl’s tables. In my paper of 
1908, I showed that individuals beginning fission (and therefore 
at a fairly definite growth stage) show coefficients of variation as 
small as those for conjugants, and differing as much from those of 
the general population (p. 454). Coefficients of variation as 
low as 4.5 were found for samples of a definite age. The low vari- 
ability of conjugants is then fully accounted for by the fact that 
conjugation does not occur till a certain stage of growth has been 
reached; and that conjugation occurs before the animals have 
reached the largest size, that precedes fission. 

One of the most striking things to be observed in a conjugat- 
ing culture is the existence along with the conjugants of many 
individuals of much greater size than the conjugants. This will 
best be seen by comparing the measurements of conjugants and 
non-conjugants of given lots, as exhibited in tables 34 and 35 of 
the appendix; it is indicated in tables 1 and 2 by the fact that 
the range of variation invariably extends to much higher limits 
in the non-conjugants than in the conjugants, as well as by the 
fact that the mean is higher for the former. Fig. 2 shows a col- 
lection of conjugants and non-conjugants from a culture of the 
race k; the specimen marked e shows one of the very large non- 
conjugants. It is easy to isolate from a conjugating culture many 
non-conjugants that are larger than any of the conjugants. 


20 H. S. JENNINGS 


Increase in size and variability of the conjugants before fission. 
Further evidence as to the significance of the decreased size of 
the conjugants will be reached by asking the following question: 
Do the conjugating individuals remain smaller than the non-con- 
jugants; or do they increase in size before they divide, till they 
finally reach the size of the large non-conjugants? A parallel 
question may be asked regarding the variability of the conju- 
gants; does this increase to the normal amount before the ex- 
conjugants divide? 


A 7 Li 
a b c d e 


Fig. 2. Conjugants (b and d) and non-conjugants (a, c and e) from the race k, 
showing the relative sizes. > 333. 


On these points observations were made for a number of cul- 
tures; those for lot 6 (of table 1) are the most complete. From 
this lot I selected fifty of the largest non-conjugants, all larger 
than any of the conjugants, and compared their dimensions with 
the dimensions of conjugants about 86 hours after separation; 
also with random samples of conjugants and non-conjugants. 
The results are given in table 7. 

As this table shows, the conjugants before dividing increased 
in size till they were fully equal to the selected largest non-con- 


CONJUGATION IN PARAMECIUM 21 


jugant individuals. The mean dimensions of the conjugants 
increased before fission from 181.49 by 48.11, to 214.48 by 64.24. 
These later dimensions of course exceed greatly the mean for the 
non-conjugant population. 

More extensive data on the increase in size of the conjugants 
before fission, though without comparison with the non-conju- 
gants, is given in table 19, on page 69. Here the conjugants of 
the two wild lots (22 and 23) had increased in length 20.62 per 
cent and 11.76 per cent, respectively, after separation; those of 


TABLE 7 


Changes in size and variability of the conjugants before they divide, in comparison with 
non-conjugants, from lot 6 (wild culture) 


LENGTH BREADTH 


DATE : 
Standard 


Standard. oem 
deviation 


Coefficients 
Re; Mean deviation of 


° 
variation variation 


Jan. 30, '08  Non-conju- 
gants, ran- 
dom sample. .) 318186.10+0.79 20.97+0.56 11.27+0.31) 200) 49.98+0.49 

Feb. 1, '08 Largest non- 
conjugants . £0211.52+1.99 20. 84+1.41) 9.85+0.67) 50 62.48+1.17 

Jan. 30, '08Conjugants, | | 
random sam- 
io) ee Gnaeritad | 272151.49+0.54 13.32+0.39, 7.34+0.21) 38, 48.11+0,57 

Feb. 1, ‘08Conjugants, 36 
hours after 
separation....| 50.214.48+2.09 21.92+1.48 10.22+0.70| 50 64.24+1.06 


So 


21+0.34, 20.43+0.72 


wo 


290.82, 19.66+1.37 


94+0 61 16.51+1.31 


= 


09+0.75) 17.27+1.20 


race k (lot 24) had increased 21.82 per cent. Comparing the 
size of the separated conjugants of race k as given in table 19 with 
the size of the non-conjugants as given in table 2 (lots 10-13), 
we find that the corjugants have become before fission fully 
as large as the non-conjugant population. 

Tables 7 and 19 show also that the variability of the conjugants 
likewise increases considerably after they separate. In lot 6 
(table 7) the variability in length of the separated conjugaats 
approaches that of the non-conjugants. The coefficients of 
variability for the separated conjugants of table 19, while con- 
siderably larger than those for the united conjugants, are some- 
what below those usual for the non-conjugating population, as 


22, H. S. JENNINGS 


shown in tables 1 and 2. There is of course an intelligible reason 
for their variability remaining less than that of the non-conjugant 
population, for the latter includes young individuals as well as 
old ones, while the ex-conjugants consist entirely of old individ- 
uals. There is no indication of a real differentiation of the con- 
jugants from the non-conjugant population in respect to varia- 
bility save as a result of this exclusion of young specimens from 
the conjugants. 

The general conclusion would therefore seem to be Justified 
that the conjugants are not differentiated from the non-conjugants 
of the same race, save temporarily, the differences disappearing 
practically before the first fission of the conjugants. 


4. Racial differences 


We have seen the cause of the lessened variability of the con- 
jugants within a pure race. Another ground for less variability 
in the conjugants, when the culture is not limited to a single 
race, is seen on examination of the data for the mixed cultures 
(lots 20 and 21, table 2). Here the coefficients of variation for 
the conjugants are respectively 60.6 per cent and 76.9 per cent 
less than that for the non-conjugants; or to put it another way, in 
lot 20 the variability of the non-conjugants is 2.54 times as great 
as that for the conjugants; in lot 21 it is 4.33 times as great. The 
remarkable differences become evident to the eye on comparing 
the measurements of the conjugants and non-conjugants of these 
lots, as shown in tables 34 and 35 of the appendix. These great 
differences in variability are due to the fact that but one of the 
two races present conjugated; so that the non-conjugant sample 
included members of two diverse races, the conjugant sample 
but one. As shown elsewhere (Jennings ’10), it frequently hap- 
pens that in a mixture of races but one race conjugates. This must 
often greatly affect the relative coefficients of variation in wild 
cultures of unknown racial composition. 

We conclude therefore that the less variability of conjugants 
as compared with non-conjugants is due (1) to the fact that the 
conjugants include only a limited number of growth stages, inter- 


< CONJUGATION IN PARAMECIUM 2B 


mediate between the largest and the smallest; (2) to the fact that 
in mixed cultures not all the races conjugate at the same time. 

The consequences of the lessened variability of the conjugants 
will be considered later. 


ASSORTATIVE MATING 


Pearl’s explanation 


Pearl (07) concluded from his study of conjugating Paramecia 
that there is a marked degree of assortative mating in these 
animals, z.e., that large individuals tend to mate with large 
ones, small individuals with small ones. This is an extremely 
important matter (as Pearl well recognized) for the understanding 
of heredity, variation and evolution in these organisms, and we 
must examine into the matter with care. Is there actually assort- 
ative mating? What are its degrees and limitations; what are 
its causes; what its effects in heredity and variation? 

In the study of these matters the method used by Pearl, and 
the one we shall to a large degree follow him in using, is to study 
the correlation between the members of pairs. We shall however 
endeavor to put to the test of experiment such questions as are 
open to it, and to combine the statistical and experimental study 
with the results of direct observation. We shall take up first 
Pearl’s explanation of how assortative mating occurs; then give 
an account of direct observations of the process of conjugating, 
in their bearing on this explanation. ‘ We shall then enter upon 
a statistical and experimental analysis of the facts. 

Pearl showed that when we measure the individuals making 
up the pairs in a Jot of conjugants, there is a rather high correla- 
tion between the two members; that is, large individuals are found 
mated with large, small with small. The details and degrees 
of this we shall take up later. Pearl’s explanation of this cor- 
relation is that there is a real assortative mating—larger indi- 
viduals mating with larger, smaller with smaller. The way in 
which this assortative mating is brought about, according to 
Pearl, is essentially the following: In a typical conjugation the 


24 H. S. JENNINGS 


two individuals first place their anterior ends together; these 
adhere. Then the two bodies are brought in line, and if the two 
mouths come in contact, they adhere, and conjugation becomes 
complete. Now it is evident that if the two animals are not of 
approximately the same length, mouth and anterior tip will not 
both come into contact, and conjugation will therefore not be 
completed; in this case ‘the individuals separate again or die, and 
no conjugation results.” Hence it is only individuals of approx- 
imately the same size that will conjugate. 

This explanation of Pearl’s is based to a certain extent on 
observation of the process of conjugation, andits essential correct- 
ness seems highly probable. But Pearl himself was able to make 
but few observations on the behavior in the process of conjugat- 
ing (as he notes on p. 266 of his paper), and it will be well there- 
fore to add what we can along this lme. Some of the details 
are of much importance for understanding the limitations as well 
as the potentialities of the assortative mating. We shall examine 
first the typical cases, then some of the variations. 


OBSERVATION OF THE PROCESS OF CONJUGATING 


The first contact between the individuals about to conjugate 
is very commonly at the anterior tips, and I am able to give figures 
of a number of pairs in which the process had gone no farther than 
this (fig. 3, a, b,c). The two tips where they meet form projec- 
tions and depressions, which interdigitate and hold the two 
together. This I have often noticed in the living specimens, and it 
is indicated in fig. 3, at b. Then the bodies themselves are brought 
together. There is at first no union except at the anterior tip, 
until the mouths are reached. These then unite, so that the ani- 
mals are held together at two points only. This stage in the 
process is represented by e, f, g (fig. 3), drawn from preserved 
material. 

Then the bodies become closely applied to each other through- 
out the stretch from anterior tips to mouth, and for a certain dis- 
tance behind the mouths. Apparently however the union is 
not so firm elsewhere as at the anterior tips and the mouth; 


CONJUGATION IN PARAMECIUM 25 


A 
DD 
Dy 


Fig. 3 Characteristic attitudes in conjugation. All are drawn with the pro- 
jection apparatus from fixed specimens, save h, which is a free-hand sketch from 
a living pair. a, e, f belong to the race k; c, d, g, i, 7 to the race C2; b, k and lL 
to the race c (all aurelia); a, b, and c, stages with only the anterior tips in contact; 
e, f, g, anterior tips and mouths in contact; h, reversed pair, at beginning of con- 
jugation; 7, 7, most complete union; k, view of pair in profile, showing the crossing 
(also seen in 7); 1, last stage, just before complete separation. Magnification (for 
all but h), 333 diameters. 


26 H. S. JENNINGS 


specimens not in contact elsewhere than at these points are fre- 
quently seen. These facts are noted by Pear] (’07). 

But this state of full contact is by no means reached without 
variation and active movements. On the contrary, there is 
frequently a period of twisting, turning, contracting, and shift- 
ing about, before the final result is reached. Certain points are 
important. 

1. At first apparently only the anterior tip (and possibly the 
mouth) is adhesive. The anterior tip of one specimen may ad- 
here, at least temporarily, to almost any part of the body of an- 
other; certainly to any part of the oral surface. Thus one often 
sees most irregular adhesions, the first specimen perhaps trans- 
verse or oblique to the second, adhering by its tip to the middle 
of the oral groove of the second; or one specimen is attached to 
the posterior half of another, trailing behind it; or three or four 
may be irregularly attached. Sometimes the animals become 
attached in reversed position—the anterior tip of one to the pos- 
terior part of the other, as in fig.3,h. Such irregular attachments 
have frequently been described and figured. 

2. If irregularly attached, the animals begin to pull, twist, 
contract, and shift, till the relative positions are many times 
changed. Not rarely one sees a pair completely separate, to 
reunite in a new position. More often they remain in contact, 
but the relative position and the parts in contact are changed 
by gliding and pulling. I have frequently seen pairs in the re- 
versed position of h, fig. 3, that finally came into the normal posi- 
tions, and underwent a typical conjugation. On the whole the 
period of ‘fitting’ in conjugation is one of great and varied activity. 

3. Among the movements at this period of ‘fitting’ are con- 
tractions and bending. One gets the impression that the animals 
are making an active ‘effort’ to bring the mouths into contact. 
A certain amount of curving of the anterior tips is common even 
before the mouths have come in contact. If when the bodies 
are brought in contact the mouths do not match, the curving and 
bending becomes very marked as in e, f, g, h, fig. 3. This is of 
course most necessary when the two individuals are not of the same 
length; the longer then may become curved (as in f and g, fig. 3, 


a 


CONJUGATION IN PARAMECIUM 27 


and in e and fh, fig. 15); so that a considerable degree of equaliza- 
tion in length may occur. In measuring the avimals it is difficult 
to detect or allow for this curving, as the animals in conjugation 
are slightly oblique to each other in any case (owing to the oblique- 
ness of the oral grooves). In actual practice, most such equalized 
pairs are measured by simply taking the distance from anterior 
to posterior tips, in the two specimens. This tends of course 
to produce a correlation that did not exist before the union. 

It is however important not to exaggerate the generality or 
amount of this equalization, and especially to remember that it 
is only one phase of a process of fitting, the remainder of which 
would lead to real assortative mating. 

4. All of this shifting and contraction may be insufficient for 
producing a proper fit: in such cases the animals separate. Such 
separation is often observed. This is of course an essential point 
for producing real assortative matiog. It appears clear that 
individuals of nearly the same size musi fit readily, and that the 
more unequal they are, the less likely they are finally to fit and 
remain united. 

5. From the description thus far, it is clear that the first attach- 
ment may not be at the anterior tips of both individuals. It is 
more likely to be here, because both anterior tips are adhesive, 
while most of the rest of the body is not. But the anterior tip 
of one may come in contact with the oral surface of the other 
some distance behind the anterior tip of the latter. If the two 
animals are equal, of course the two mouths will now not come 
together, until the positions have been shifted; but if the two ani- 
mals are unequal,—if the one lying more to the rear is shorter,— 
the mouths may then come in contact, and complete conjugation 
will take place between unequal specimens. Thus the assorta- 
tive mating, or the correlation in size between two members of a 
pair, cannot be expected to be complete, since many unequal 
pairs are found. A number of such are shown in fig. 15, page 53. 

Thus on the whole direct observation of the process of conju- 
gation and of the conjugated pairs is favorable to Pearl’s view 
that real assortative mating occurs, and to his explanation of the 
way it occurs. One important point, not brought out by Pearl, 


28 H. S. JENNINGS 


results from our description,—namely, that there is a real tend- 
ency toward equalization in length of the two members of the 
pair. This of course tends to produce correlation where it would 
otherwise not exist; in other words it gives, so far as it goes, an- 
other explanation of at least a portion of the correlation, and one 
which would deprive the process of its significance for variation, 
heredity and evolution. So far as direct observation goes, how- 
ever, this factor may be of very slight value, accounting for but 
little of the observed correlation, but it must be kept in mind in 
the statistical and experimental work, and evidence as to its real 
value obtained if possible. 


CORRELATION IN SIZE IN THE MEMBERS OF PAIRS OF CONJUGANTS 


We will now examine the facts as to the correlation in size of 
the members of a number of lots of conjugants. The data are 
given in tables 1, 2 and 3, pages 9 to 11, table 3 giving the facts for 
a number of lots in which only conjugants were measured, while 
tables 1 and 2 relate to lots in which a comparison was made be- 
tween conjugants and non-conjugants. In all these tables the 
data are for conjugating pairs measured while still united; later 
will be found the facts regarding correlation in pairs measured 
after separation (tables 19 and 20). 


MEANING OF THE COEFFICIENTS OF CORRELATION 


In all these tables the degree of correlation is stated in terms of 
the coefficient of correlation; coefficients will be found varying 
from —0.193 to + 0.940. In order to grasp clearly what is 
meant by correlation, and by the different coefficients of correla- 
tion, it will be helpful to examine the pairs shown in fig. 4, and 
the diagram of fig. 5. 

Fig. 4 shows a number of pairs selected in such a way as to 
exhibit the condition of affairs in the case of marked positive cor- 
relation. Where one member of a pair is large, the other is like- 
wise large, and at the other end of the series both members are 


PARAMECIUM 


IN 


ATION 


x 
I 


CONJU 


‘sdajauIVIp GZ ‘UOTYVOYTUsSe IY 
‘yoyered ose ,g-g pue Q-V seat oy “seyeUT ][VUIS sO [[VUIS ay} ‘S97 aZIv] SUIAVY S[ENPIAIPUI oFAe] OY} “OZ1S UL WOTPBTOLIO9 
aAT}ISOd ay} 4IGIYxXe 0} SB posed os “go6T ‘8 Youve ‘(eljaine) %) 9ov1 oyz Jo eINy[Nd B WLOTT sired jo dnois poyopes YF “B14 


LO 


va 


30 H. S. JENNINGS 


small. Thus if we select the pairs according to the size of one 
member only, we shall nevertheless find the other member to 
have nearly the same size. 

If in all the pairs of varying size the two members of each pair 
were precisely equal in length, then the correlation would be com- 
plete; the coefficient of correlation would be unity (1.000).4 But 
what is meant by coefficients of correlation less than 1.000; by 
such coefficients as 0.398 (lot 1, table 1), or 0.507 (lot 3, table 1)? 
And what is the difference between the two latter cases? Why 
should lot 3 be given precisely the coefficient 0.507, in place say 
of 0.275 or 0.860? It is this question that the diagram of fig. 5 
is intended to assist in answering. 

Suppose we take all the 360 individuals forming the 180 pairs 
of lot 1 (table 11, page 47) and group them according to their 
lengths into groups at intervals of four microns. We thus ob- 
tain 29 groups, the smallest 148 microns long, the largest 260 
microns long. We then place these 29 groups side by side, be- 
ginning with the largest and proceeding to the smallest, as in fig. 
4. But instead of giving the actual outlines as we did in fig. 4, 
we use merely lines giving the length of each group. This gives 
us the vertical lines of fig. 5,—the length of the largest individuals 
being represented by the line A-B, that of the smallest by the line 
C-D, the others by the intermediate lines. Then the ends of these 
lengths form the oblique line B-D. The average length of 
all is shown by the line 0-0’. 

Now suppose we examine the mates of the individuals of these 
groups—getting the average length of the mates for each group. 
If correlation were complete (coefficient 1.000) the mates would 
be of the same lengths as the first individuals (which we may call 
the principals) ; their upper ends would lie on the same line B-D. 


4It is worthy of special notice that in the particular case of correlation with 
which we are dealing, where the two members are of the same sort, this absolute 
equality of the two members of the varying pairs is the necessary condition for 
producing the coefficient 1.00. In other cases of correlation this is usually not the 
case. Thus it would be possible to study the correlation between the bodily stat- 
ure and the length of the little finger in man. In this case complete correlation 
(coefficient 1.00) would mean that the same proportion of one to the other was 
present in all the varying cases. 


260 


240 


220. 


180 


160 


140 


120 


100 


80 


60 


40 


20 


CONJUGATION IN PARAMECIUM 31 


A Cc 


Fig.5 Diagram to illustrate correlation in size between the members of pairs in 
lot 1 (see explanation in the text and compare fig. 4). The vertical lines represent 
the lengths of individuals of the different classes arranged in order from longest 
(260 microns) to shortest (148 microns), their terminations tracing the oblique 
line B-D. The distance from A-C to O-O’ shows the average length for all. The 
distance from A-C to the broken line H-F shows the lengths of the mates of the 
classes of individuals represented by the vertical lines. The general trend of this 
line E-F is shown by the line X-Y (‘regression line’). The value of the coefficient 
of correlation is given by the proportion which the line O-X is of the line O-B. 


But in fact we find that the mates are not of exactly the same 
lengths as the principals. Thus in lot 1 (table 11, page 47), 
the two individuals having the length 160 microns (40 units) 


32 H. S. JENNINGS 


have mates respectively 164 microns and 196 microns in length, 
so that the average of the mates is 180 microns. We thus work 
out the average length of the mates for each length of the prin- 
cipals; this gives us the results shown in table 8. As this table 
shows, the length of the mates increases on the whole as the 
length of the principals increases (though not at the same rate), 
so that larger individuals have somewhat larger mates. 

Now we mark on the diagram of fig. 5 the lengths of the mates 
corresponding to the lengths of each group of principals repre- 
sented by the vertical lines; we find that these lengths of the mates 


TABLE 8 


Mean lengths in microns of the mates for the individuals of diverse given lengths, in 
the 180 pairs of lot 1 


NUMBER OF PAIRS LENGTH OF PRINCIPAL MEAN LENGTH OF MATES 
1 148 188.0 
1 152 192.0 
2 160 180.0 
4 164 181.0 
2 168 182.0 
4 172 192.0 

11 176 188.4 
23 180 192.3 
19 184 190.0 
29 188 194.9 
36 192 194.2 
35 196 198.9 
41 200 200.6 
38 204 202.4 
34 208 204.2 
31 212 205.0 
15 216 207.5 
15 220 201.6 
6 224 205.3 
4 228 206.0 
4 232 202.0 
1 236 248.0 
1 240 228.0 
2 248 222.0 
1 260 212.0 


A CONJUGATION IN PARAMECIUM 33 
are distributed on the irregular (heavy) line H-F. The course 
of this line shows that the smaller individuals, near the side 
C-D, have mates larger than themselves; that the larger individ- 
uals (near A-B) have mates smaller than themselves, while the 
intermediate individuals have mates of nearly their own size. 
But it is on the whole clear that the line H-/' does slope a little 
in the same direction as B-D, only less; large individuals in the 
left half of the diagram do have larger mates than the smaller 
individuals, in the right half. That is, there is a certain degree 
of positive correlation between the size of individuals and the size 
of their mates. To show how marked this is, we may draw a 
straight line X-Y, showing the general trend of the slope of the 
broken line E-F. (The method by which this line is drawn will 
be taken up later.) The line X-Y shows approximately what 
would be the course of the line H-F if we had an infinite number 
of cases; the irregularities in H-F’ are due to the limited number 
of pairs with which we must deal. We may therefore look upon 
X-Y as showing us the real mean lengths of the mates of the 
individuals having the lengths shown by the vertical lines. 

The position of this line X-Y with relation to the position of the 
line B-D is now the important point for determining the degree 
of correlation. We see that X-Y rises above the mean (O-O0') 
where B-D rises above it, and falls below the mean where b-D 
falls below it, thus sloping in the same general direction as 5-D. 
If X-Y did not slope with B-D, but were instead quite horizon- 
tal (coinciding with O-O’), then there would be no correlation 
(coeflicient 0), since this would show that small individuals and 
large individuals had mates of the same average size. On the 
other hand, if X-Y not only sloped in the same direction as B-D, 
but actually coincided with it (so that all specimens had mates 
equaling them in size), then the correlation would be complete 
and the coefficient would be 1.00. But as a matter of fact X-Y 
falls neither at O-O’, nor at B-D, but between the two,—and 
its precise position is what determines the numerical value of the 
coefficient of correlation. The line X-Y cuts off at X just 0.398 
of the entire distance from O to B (that is, it cuts off 0.398 of the 
angle between the lines O-O’ and B-D); therefore the coefficient 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. Il, NO. 1 


34 H. S. JENNINGS 


of correlation is 0.398. If X fell half way between O and B, the 
coefficient of correlation would be 0.500; if it fell nine-tenths of the 
distance from O to B, the correlation would be 0.900, ete.° 

In place of measuring the proportion of the distance O-B cut 
off by X, we could of course measure on any of the vertical lines 
of the diagram the portion of the distance from the line O-O' 
to the line B-D that is cut off by X-Y; the result would be the 
same. 

Fig. 5 may be used further to illustrate negative correlation. 
If the line X-Y sloped in the opposite direction from B-D, falling 
below O-O’ where B-D rises above it, this would of course show 
that the larger the individual the smaller its mate; 7.e., we should 
have negative correlation. To produce complete negative cor- 
relation (coefficient, —1.00) the line X-Y would make the same 
angle below O-O’ that B-D makes above it, and vice versa. The 
degrees of negative correlation would then be determined in the 
same way as those of positive correlation. 

All this may be clearly illustrated if we make a diagram rep- 
resenting only the upper part of fig. 5, above D (that is, including 
only. the varying portions of the lines); such a diagram is given 
in fig. 6. On this diagram are shown the various positions of 
the line X-Y corresponding to different degrees of positive and 
negative correlation when the heavy line B-D shows the dimen- 
sions for the principals (also complete positive correlation in the 
mates).* (The diagram in fig. 6 is made on a somewhat dilfer- 
ent scale from that of fig. 5, the vertical distances being greater 


5 This exposition would not hold, in its present form, for eases where we seek 
the correlation between unlike things (as between the stature and the length of 
the finger in man). In such cases the coefficient of correlation depends partly on 
the relative variability of the two sets of things compared. In the case of corre- 
lation between likes, with which we are dealing (where in fact the two classes 
compared are composed of the same individuals), this complication does not come 
in; the means, standard deviations, and coefficients of variation are the same for 
the two classes, so that the coefficient of correlation is identical with the coefficient 
of regression. For such cases our exposition holds without modification. 

® Some authors report coefficients of correlation greater than 1.00. This is of 
course due to arithmetical errors, since when the measurements all fall on the same 
straight diagonal line passing through the mean, the coefficient is but 1.00, and it 
is decreased when any of the measurements fall elsewhere than on this line. 


CONJUGATION IN PARAMECIUM 39 


i. 
.9 
8 
7 
NG 
Es 
a \ 
3 NN 
SSA 
a2 SSN 
RB BSS S 
2 === 
sesS5 
te Liz 
| Gi 
ay 
as 7) 
~.6 
237 
-.9 
=a 


Fig. 6 Diagram illustrating the significance of different coefficients of correla- 
tion (should be compared with fig.5,tothe upper portion of whichit corresponds). 
The vertical lines are the terminal portions of the lines representing the lengths of 
various classes of individuals, arranged according to their size, the largest (A-B) 
at the left. The terminations of these diverse lengths trace the oblique line B-D. 
O-O marks the position of the mean length of all. The lines B-D and O-O and the 
other oblique lines show the positions of the theoretical average dimensions for 
the mates of individuals of the given sizes, in case of different coefficients of cor- 
relation. If the dimensions of the mates fall in the same line B-D as those of the 
principals, the correlation is 1.000; if they fall at the mean O-O the correlation is 
0; if they fall at right angles to B-D, the correlation is —1.00. If they fall in a 
line dividing O-B into equal parts, the correlation is 0.500. The coefficients, from 
—1. to +1., by tenths, are illustrated in the diagram, the numbers at the 
extremity of a given line showing the coefficient to which it corresponds. 


in proportion to the horizontal distances.) We shall use similar 
diagrams for showing the correlation in the various lots studied. 
From figs. 5 and 6, it will readily be conceived what is meant 
when such coefficients of correlation are mentioned as appear in 
the last columns of tables 1, 2 and 3. 

The position of the line X-Y is determined in practice simply 
by finding the coefficient of correlation, then marking off the 
equivalent proportion of O-B above O, and of O’-D below 0’, 


36 H. S. JENNINGS 


and connecting these two points by a line. (In the case of cor- 
relation between unlike things, modification of this procedure 
would be necessary; a coefficient of regression is derived from the 
coefficient of correlation, and this gives the position of the line 
X-Y. But in such eases as we are dealing with the value of the 
coefficients of correlation and of regression are the same.) 

The method of finding the coefficient of correlation is of course 
described in text-books of statistical methods. For computing 
correlation under the particular conditions with which we are 
here dealing, an improvement over the usual methods is described 
in a recent paper by the present author (711). 


TABLE 9 


Coefficients of correlation in length between the members (A and B) of pairs in the 
different classes of cultures of Paramecium (compare tables 1-3) 


WILD CULTURES, OF MIXED RACIAL E 
JRE RACE: 
COMPOSITION PU CES 


ae 2 ge. 
eee 2 2 2s e23 
1 11 | 180 0.398+0.030 9 c 43 125 0.132+0.042 
2 36-142 0.268+0.057 | 10 k Ad 26 —0.193+0.090 
3 37 82 0.507+0.049 11 k | 45 14. —0.137+0.125 
4 38 42 0.499+0.055 | 12 k 46 78 0.367+0.047 
6 39 | 136 0.428+0.033 | 18 k 47 =: 168 0.184+0.036 
7 40 7 0.333+0.048 || 24 k 122 0.210+0.041 
22 | 55 | 102 | 0.359=0.041 | 14 Nfe 21 0.295+0.095 
23 | 22 | 148 | 0.245+0.037] 16 Nps 25 0.257+0.089 
and 16 G 48 150 0.507+0.029 
23 17 Cs 49 69 | 0.318+0.052 
18 g 50 84 0.251+0.049 
Mixtures of two species, both con- 19 g 51 87 0.323+0.046 
jugating 
26 31 | 0.939+0.010 | | 


27 10-170 0.940+0.004 | 


Mixtures of two species, only one 
conjugating 


21 54 49 0.408+0.064 


lod 


CONJUGATION IN PARAMECIUM 37 


CORRELATION IN DIFFERENT CLASSES OF CULTURES 


Turning now to an examination of the coefficients of correla- 
tion of the various lots, as shown in tables 1, 2 and 3, we find that 
we may distinguish four different classes: (1) wild cultures, 
of unknown racial composition; (2) wild cultures known to con- 
tain pairs of the two species, aurelia and caudatum; (3) cultures 
consisting of a single pure race; (4) mixtures of two known races. 
It will be well to consider these separately. The coefficients of 
correlation for these classes are summarized in table 9. 


Wild cultures of unknown racial composition 


Tables 1, 3 and 9 give us the correlation for eight lots of this 
class. We find positive correlation in every case, the coefficients 
ranging from 0.245 to 0.507, with an average for the entire eight 


B 


D 


Fig.7 Diagram illustrating the extent of correlation in the eight ‘wild’ cultures 
of tables 1 and 3 (at a to 6), and in the mixture of two species in lot 27 (atc). Be- 
ginning at a the lines show in order the coefficients of correlation for lots 23, 2, 7, 
22, 1, 5, 4, 3, (at b); then that for the mixture of two species at c. 

O-O, mean and line of no correlation; B-D, line of complete positive correlation 


(compare figs. 5 and 6). 


38 H. S. JENNINGS 


of 0.380. In the five series studied by Pearl the coefficients were 
higher, ranging from 0.430 to 0.794, with an average of 0.614. 
Fig. 7, a to b, illustrates the correlation in our eight wild cultures. 
The actual relations of the individuals of various lengths to their 
mates are shown for lot 1 in table 11 and fig. 5; for lot 3 in table 
37 and fig. 8. 


B 
1 240 


230 


220 


ail 160 


Fig. 8 Diagram illustrating correlation in lot 3 (‘wild’ culture). B-D, lengths 
of the classes of individuals in order of size. H-F, lengths of the mates for these 
classes. X-Y, regression line showing the general trend of the broken line E-F, 
and indicating by the proportion of the line O-B which it cuts off, the value of the 
coefficient of correlation (0.507). O-O, mean length for the entire lot. The num- 
bers at the right show lengths in microns. 


Since we now know that wild cultures often contain a number 
of races of diverse sizes, the question arises whether the corre- 
lation may not be a result of this fact. If members of large races 
conjugated only with other members of large races, and members 
of small races likewise mated only together, there would result 
a positive correlation, provided that more than one race were 
undergoing conjugation at the same time. If such interracial 
selectiveness were the only basis for the correlation, we should 


CONJUGATION IN PARAMECIUM 39 


of course find no correlation on studying conjugating pairs that 
all belonged to a single pure race. To get light on this matter, 
we may examine the correlation (a) in cases where there were 
known to be two greatly differing races (two so-called species) ; 
(b) in cases where members of but one race are present. 


Cultures containing pairs belonging to two different species 


In two cases I was able to obtain conjugants from cultures 
containing both Paramecium caudatum and P. aurelia. These 
two species, as is well known, differ considerably in size, but very 
little in other external features.’ What happens when members 
of the two species, mingled together, conjugate at the same time? 

Simultaneous conjugation of the two was obtained as follows: 
Material known to contain Paramecium caudatum in large num- 
bers was brought into the laboratory, and mixed with cultures 
of the aurelia race k. About a week after the mixture was made, 
the conditions became favorable for conjugation, and both species 
mated. The resulting matings are shown, for the culture from 
which the largest numbers were measured, in table 10. 

From this table it is clear that the large individuals of caudatum 
mated exclusively with other caudatum; the small individuals 
of aurelia only with other aurelia. The two cultures of this sort 
that were examined gave coefficients of correlation in length, 
of 0.939 +0.010 and 0.940 = 0.004, respectively ; so that the cor- 
relation was almost perfect. Careful examination of all the 
pairs measured gave no single case in which it appeared, to the 
practiced eye, that caudatum had mated with aurelia. 

Thus when caudatum and aurelia are present together in a cul- 
ture, they do not intermix in conjugation,—certainly not to any 
marked extent, and apparently not at all. (Simpson, ’01, saw 
two cases of what he believed to be conjugation between aurelia 
and caudatum. The ex-conjugants died after one fission. It is 
of course possible that crosses might be induced by proper isola- 


7 For a detailed account of the differences between the two, see Jennings and 
Hargitt, 710. 


40 H. S. JENNINGS 


tion, even though they do not occur in nature, or occur there but 
rarely). 

Our results with two species are then favorable to the idea that 
correlation is produced by members of related races conjugating 
together. We now turn to the records for conjugation within a 
single race. 

TABLE 10 
Correlation table for the lengths of 170 pairs from a culture contain- 
ing pairs of both aurelia and caudatum (correlation, 0.940 +0.004). 
The unit of measurement is four microns (0.004 mm.,) (so that 


the first pairs to the left have members measuring 112 and 108 
microns in length) 


eile | heal l Ree ae Sse 
128.2930 31/32/33 34 3530137 3830-40 4142.43 44 45,46 47 48 49)50)51 52 53,54 55 56 57 58.59) 
j i St Pe | | J | | 


— tt SS Ses Se tes i — 


| Beeea at 4 
| || | Peers Pad lid 


a de bo 
el 
a 

DAR OWN NL & Ome 


ae ee 
ie) 


ee) 


= 


— ee 


_ 
i 


ee 
es Ce os 


bots 


Nww ree 
Ney ke ee 


ym bo to 
~ 
- 
Oe 


_ 


| | | 
| | | | | | | 

55 | Ved yea | [1 
| 


35 510 710124M\10, 56 5 2.21 1 3l 1) 1) 9| ss! gal e| 2| 3| 5 1 1) 170 
| ' 1 


CONJUGATION IN PARAMECIUM 41 


Conjugation within pure races 


The coefficients of correlation in length for twelve conjugations 
within pure races (all the conjugating individuals being derived 
originally, in each case, from a single specimen) are given in tables 
2, 3 and 9; they are illustrated in the diagram of fig. 9. 

Examination of the tables and the diagram shows that 
the correlation is indeed considerably less in the pure 


1.8 


-1. 
A D 


Fig.9 Diagram illustrating the correlation in the eleven lots composed of pure 
races, of table 2. O-O, mean and line of no correlation; B-D, line of complete 
positive correlation. a, lot 9; b, lot 10; c, lot 11; d, lot 12; e, lot 13° g, lots 15 and 
18;h, lot 16;7, lots 17 and 19. Just beneath 7 is the line for lot 14. 


caces than in the wild cultures. This is seen in. comparing 
fig. 9 with fig. 7. From the first three coefficients of table 2 
(0.182, —0.193, and —0.137) we might conclude indeed that sig- 
nificant correlation is quite absent within a pure race. But on 
examining the entire twelve lots from pure races, we find that this 
conclusion will not hold. In all save the two lots just 
mentioned the correlation is positive; and these two in which it 


42 H. S. JENNINGS 


was found negative are small, containing respectively but 26 and 
14 pairs; they have been included in our account merely to illus- 
trate the results reached if insufficient numbers are employed. 
Other, much larger lots of this same race k (lots 12, 13 and 24 of 
tables 2 and 3) gave significant positive correlation. Further, 
we have in lot 16, from the pure race C;, a large lot (150 pairs) 
giving the high correlation of 0.507 + 0.029. 

We must conclude then that there is really positive correlation 
in length in the members of pairs even when all belong to the 
same race. The fact that it is much less in amount than in mixed 
cultures is however of much significance. Fig. 10 illustrates 
the very slight correlation in the large sample of lot 9 (race c) 
while fig. 11 shows a similar diagram for lot 17 (race C,). The 
average coefficient for the twelve lots from pure races is but 0.251, 
as against 0.380 for our eight wild cultures, and 0.614 for Pearl’s 
wild cultures. The difference in these averages is illustrated 
in fig. 12. 

The smaller correlation found in pure races, as compared with 
mixed cultures, casts much light on the causes of the correlation. 
It indicates most directly, of course, that in wild cultures indi- 
viduals belonging to the same race, or to races of similar size, 
tend to mate together. The bearing of this most important con- 
clusion on other problems we shall bring out later. Here we shall 
take up certain other evidence indicating this tendency of mem- 
bers of the same or like races to conjugate together. 

In wild cultures it is frequently found that members of diverse 
races are conjugating at the same time. This is demonstrated 
by isolating pairs of different sizes, and finding that they produce 
progeny of permanently different characteristic sizes. We may 
cite here a single case, described in an earlier paper (Jennings, 
‘08, p. 494). Six races of diverse size (including the large race L, 
with average length of about 200 microns and the small race C:, 
with average length of about 125 microns) were derived from six 
differing pairs of conjugants taken from a single culture January 
29, 1908. 


CONJUGATION IN PARAMECIUM 43 


ae 200 
160 

F 

x 

a O 

Né 
140 
-! 120 

D 
10 


1 


Fig. 10 Diagram for correlation (0.132) in the pure racec of lot 9. For expla- 
nation, compare fig. 8, page 38. 

Fig. 11 Diagram for correlation (0.318) in the pure race Cy of lot 17. Lettering 
as in fig. 8, page 38. 


44 H. S. JENNINGS 


A D 


Fig 12 Diagram showing the average correlation in length between the mem- 
bers of pairs in wild cultures of one and of two species, and in pure races. O-O, 
mean; also line of no correlation; B-D, line of complete positive correlation. E, 
average correlation (0.250) for the twelve pure races of tables 2 and 3. F, aver- 
age correlation (0.380) for the eight wild cultures of the present paper (tables 1 and 
3). G, average correlation (0.614) in the five wild cultures studied by Pearl (’07). 
H, average correlation (0.940) for the two lots of table 3 where two species were 
present, both conjugating. 


Mixtures of two known races 


When cultures are formed by mixing two or more diverse 
races of known characteristics, it is extremely difficult to induce 
the members of both races to conjugate at the same time. I 
kept many such mixtures for months but only once did I succeed 
in getting both races to conjugate at once. This was in a culture 
containing the races 7 and / (both aurelia). The race 7 is very 
small, averaging but 95 to 100 microns in length, while x is larger, 
averaging about 125 microns in length.’ The two races had been 
living together in the same culture about five months (from 


5 For measurements of these races, see Jennings ’08. 


CONJUGATION IN PARAMECIUM 45 


November 8, 1908) when conjugation was observed in the cul- 
ture March 1, 1909. Unfortunately the conjugation was scanty, 
so that only five pairs were found. Three of these measured 
respectively 78 x 76, 78 x 76, 82 x 76 microns; the other two 138 
x 138 and 138 x 140 microns (compare fig. 13). It is clear that 
the first three pairs belonged to 7, the last two to k. Had we a 
large number of pairs of these two races, it is clear that we could 
form a correlation table from the culture as a whole that would 


i 
ip re Ay 


k k k i i 


Fig. 13 Conjugants and non-conjugants from a culture composed of a mixture 
of the two aurelia races k and 7, showing that each pair is composed of individuals 
of but one race,—either of k or 7. X 333. 


show a high degree of correlation. The 7’s would all fall in the 
upper left-hand corner of the correlation table, the k’s in the lower 
right-hand corner. 

Thus in this case the individuals of each race conjugated only 
with members of their own race. In most cases where conju- 
gation occurs in a culture containing two known races, it is lim- 
ited to the members of one race, as I have already set forth 
(p. 22). This fact that two races conjugate at different times has 
of course the same tendency as assortative mating, in preventing 
admixture of races. 


46 H. S. JENNINGS 


Before proceeding to further discussion of the causes and effects 
of the correlation of the conjugating individuals, it is important 
to examine certain other points. 


TO WHAT IS DUE THE INCOMPLETENESS OF CORRELATION? 


As we have seen (tables 1, 2, 3, ete.), correlation between the 
members of pairs is never complete, but in most cases is a cer- 
tain positive amount, less than 1.000. It is important to realize 
as nearly as possible to what biological relations this corresponds. 
Why, if there is correlation at all, should it not be complete cor- 
relation? 

Incompleteness of correlation may be due to a number of differ- 
ent conditions, the more important of which we must consider. 


1. Slight differences of less effect than great ones 


Perhaps the most probable cause for incomplete correlation, 
at least in such material as we are dealing with, would be this: 
that slight differences in size between the members of pairs do 
not prove a marked obstacle to their union, while greater dif- 
ferences prevent their uniting. If this is the state of affairs, 
we should find but slight correlation in collections or parts of col- 
lections in which there are but small differences among the 
individuals; higher correlations where the differences between 
individuals are great. The highest degree of correlation would 
be found in the case of a collection composed of two sets, those 
of each set not differmg much among themselves, but the two 
sets differing considerably. 

That the condition we have sketched is the real condition in 
the conjugation of Paramecium is shown in two different ways: 

a. We have already seen that when the conjugating culture 
contains two sets differing greatly (as when aurelia and caudatum 
are both present), the correlation is very high; when numbers 
of less differing races are present, as in ordinary wild cultures, 
the correlation is less, but still marked; when only the closely 


CONJUGATION IN PARAMECIUM 47 


similar members of the same race are present, the correlation is 
still less. 

b. It is shown further when we compute the correlation for 
portions of the various lots. Take for example lot 1 (tables 8 
and 11.) Here we have 360 individuals varying from 148 to 260 
microns in length; in the tables these are arranged in 29 classes 
in order of size. The mean lies at almost exactly 200 microns. 


TABLE 11 


Correlation table for the lengths of the pairs in lot 1, with lines to show the ‘medium’ 
and ‘extreme’ pairs described in the text and illustrated in fig. 14. 

(1) The pairs containing exclusively medium specimens are those in the small square 
enclosed by the lines a-b-c-d (correlation 0.104). (2). Pairs at least one member 
of which is of medium size (correlation 0.229) are between the lines c-g and c-e on 
the one side; b-h and b-f on the other. (3) Pairs at least one member of which is 
extreme (correlation 0.439) include all those outside the square a-b-c-d. (4) Pairs 
consisting exclusively of extreme specimens (correlation 0.678) lie outside all the 
lines drawn within the table. 

The unit of measurement is four microns—so that the length 37 for example signifies 
148 microns. 


10014 6 9161319 21222312146 3400 1200 1/180 


48 H. S. JENNINGS 


Now, suppose that in this culture there existed no specimens 
near this mean size, but only the larger and smaller specimens. 
What would be the nature of the correlation? To determine 
this we must omit all pairs containing medium specimens, and 
compute the correlation only for those containing extreme in- 
dividuals. If there were no correlation, no tendency for like to 
mate with like, we should among these extreme individuals find 
large specimens matched as often with small as with each other; 
the correlation would be 0. 

Let us consider as ‘medium’ specimens all those included in the 
three groups above and the three below the group containing the 
mean; that is the seven groups nearest the mean, in table 11; 
these are marked off by the lines a, b, c, d. Then the extreme 
specimens are those lying entirely outside these lines. There 
are twenty-three pairs containing only such extreme individuals; 
computing the correlation for these, we find it to have the high 
value of 0.678 + 0.054. For the lot as a whole the correlation 
is but 0.398 + 0.030. If in the same way we compute the cor- 
relation for all pairs in which no extreme individuals are present 
(the 87 pairs within the small square enclosed by the lines a, b, 
c, din table 11) we find a still smaller coefficient, of but 0.104 
==) 0/041. 

To complete the picture, we may ask what the correlation is 
when we select all medium individuals as principals, and compute 
their correlation with their mates, whether the latter are ‘medium’ 
or ‘extreme,’ and do the same for all extreme individuals. We 
find that there are 157 pairs belonging to the former group, 93 
to the latter. The 157 ‘medium’ individuals are correlated with 
their mates to the extent of 0.229 + 0.036; the 93 ‘extreme’ 
individuals are correlated with their mates to the extent of 0.439 
+= (0.056. 

Thus we find that there is a steady increase in the positive cor- 
relation as we include more and more ‘extreme’ individuals, the 
coefficient beginning at 0.104, and becoming successively 0.229, 
0.398, 0.439 and 0.678. ‘This is exhibited in the diagram of fig. 14. 

These relations are general. I have worked out the correla- 
tion for the ‘extreme’ and ‘medium’ specimens for a number of 


CONJUGATION IN PARAMECIUM 49 


a 


Or Nh Wh 
| OO 


A D 


Fig. 14 Diagram showing the different degrees of correlation between members 
of pairs in lot 1 (table 11), according as we examine the mating of individuals of 
medium sizes or of those of extreme sizes. O-O, mean and line of no correlation; 
B-D, line of complete positive correlation. 1, Correlation (0.104) for pairs con- 
sisting exclusively of individuals of ‘medium’ size. 2, Correlation (0.229) when 
we take all specimens of medium size as principals, and compare them with their 
mates. 3, Correlation (0.398) whea all specimens, of all sizes, are included. 
4, Correlation (0.439) when we take all specimens of ‘extreme’ size as principals 
and compare them with their mates, of whatever size. 5, Correlation (0.678) for 
pairs consisting exclusively of individuals of extreme sizes. (For the sizes con- 
sidered ‘medium’ and ‘extreme’ see table 11). 


the lots mentioned in tables 1 and 2. The results are shown in 
table 12. Column 1 answers the question: What would be the 
correlation if no extreme specimens existed? Column 2 gives 
the correlation for the culture as a whole, while column 3 answers 
the question: What would be the correlation if no specimens 
of medium size existed? ; 

As appears from the table, in every case the extreme individ- 
uals show higher correlation than the ‘medium’ ones, as well as 
higher correlation than the culture as a whole. In all cases save 


THE JOURNAL Of EXPERIMENTAL ZOOLOGY, VOL. Il, NO. 1 


50 H. S. JENNINGS 


TABLE 12 


Correlation between the members oj pairs (a) for extreme individuals only, as com- 
pared with the correlation for (b) medium individuals only, and (c) for the entire 
collections, including both medium and extreme specimens. (Column IV shows 
what are arbitrarily considered ‘medium.’ specimens for the lot in question; the 
extreme spectmens are those lying outside the limits designated medium) 


I | II II IV 


| 
| - = | a 
| 
non MEDIUM PAIRS | ALL PAIRS EXTREME PAIRS 
re j = —S= saa | SES CONSIDERED 
um- | *“MEDIUM’ 
lars Gocticlent: of Nom, Correlation pia Correlation ‘ ; 
(Wild culture) 
l 87 0.104+0.051 | 180 0.398+0.030 | 20 0.678+0.054 198-212 
2 45 0.320+0.064 | 142 0.268+0.057 | 29 0.518+0.092 15/-173 
3 39 0.015+0.076 82 0.507+0.049 | 16 0.830+0.052 154-204 
6 59 0.421+0.072 136 0.428+0.033 19 -0,6140.096 172-192 
7 27 0.021+0.130 79 | 0.333+0.048 | 22 0.477+0.111 160-172 
(Pure races) | 
9 36 —0.'27+=0.111 | 125 | 0.132+0.042 | 34 | 0.176+0.112 144-156 
16 49 0, 256+0.090) 150 0.507+0.029 | 41 | 0.843+0.022 120-136 
19 WeexA 0.052+0.129 87 | 0.323+0.046 | 16 | 0.681+0.090 112-124 
(Random mating)| | | 

Au 45 0.051+0.100 | 105- | —0.108-+0.046 | 16 | —0.192+0.162 | 163-177 
Cu 29° 0.187+0.121 | 101 | 0.045+0.047 | 27 | 0.0000.092 167-182 


one, the correlation for the ‘medium’ individuals is less than that 
for the culture as a whole. Further, on the whole, the greater 
the correlation of the lot as a whole, the greater the difference 
between the correlation of the ‘medium’ and the ‘extreme’ speci- 
mens. 

It may be well to note that this greater correlation of the ex- 
tremes is by no means a necessary result of mere arrangement 
in a correlation table. If the individuals are merely paired at 
random, there is no significant correlation either in the culture 
as a whole, or in the extreme pairs. This may be illustrated from 
the random pairings made by Pearl (’07). Pearl wrote on sep- 
arate slips of paper the lengths of all the individuals concerned 
in the pairs; mixed these together, and drew out two at a time, 
forming thus ‘random pairings.’ He did this for two lots, A 
and C, containing respectively 105 and 101 pairs of conjugants. 
I have worked out the correlation for the medium and extreme 
specimens for the tables so formed (Pearl’s tables A 11 and C 


CONJUGATION IN PARAMECIUM oil 


11). The results are given in the last two rows of our table 
12. As there appears, the extreme pairs do not show positive 
correlation, any more than do the rest of the collection. 

It is further to be noted that this higher correlation between 
the extreme specimens than between the specimens of medium 
size is not a necessary consequence of the existence of a consider- 
able degree of positive correlation in the table as a whole,—though 
it is doubtless a very common accompaniment of such positive 
correlation. But it is easy to form tables showing a marked 
degree of positive correlation, in which the correlation of the 
extreme parts is not greater than that of the medium parts. 

Why small differences between the individuals should not act 
so precisely in determining correlation as do large differences 
will be clear to anyone who considers carefully the process of 
mating, as described on previous pages. The difficulty in pair- 
ing caused by slight differences between the two individuals 
concerned is readily overcome by slight curving, shifting, etc., 
while great differences are not so easily remedied. Hence speci- 
mens differing much do not often unite, while those differing but 
little unite readily. Thus the correlation tables may be expected 
to exhibit many pairs in which the two members differ slightly ,— 
and this of course prevents the correlation from being complete. 


2. Different categories of pairs following different rules 


A second condition that would result in incompleteness of cor- 
relation would be the existence in the lot of different categories 
of pairs, following different rules. A certain set might, taken by 
themselves, give complete or nearly complete positive correla- 
tion, while another set, following different rules of union, might 
show little or no positive correlation, or even negative correla- 
tion. The lot might then give as a whole bit a moderate degree 
of positive correlation. 

Is there any ground for suspecting the existence of such diverse 
categories of pairs in our material? 

Careful examination of the pairs shows that there is such 
ground. The assumption on which is based the explanation 


52 H. S. JENNINGS 


of the existence of positive correlation is that the pairs in conju- 
gation place their anterior tips in contact, so that in the pairs as 
we find them the anterior ends of the two members should be 
even, as in fig. 3, c, e, z, ete. Pearl (07, p. 267) notes that this 
is on the whole approximately true in most pairs; he does not 
give measurements on this point. But examination of a large 
number of cases shows that (as Pearl further noted) the anterior 
tips are not always even. 

Now, if placing the anterior tips evenly together results (as it 
should) in high positive correlation, then if in any cases the an- 
terior tips are not placed evenly together; if the anterior tip of 
one individual is placed some distance from the tip of the other 
individual (as in fig. 15, b, c, e), then this would naturally result, 
for such pairs, in less positive correlation, or in no correlation, 
or even perhaps in negative correlation. 

We might perhaps then expect to find at least two categories 
of pairs, giving different results so far as correlation is concerned: 
(1) those with anterior tips even; (2) those in which the anterior 
tip of one individual projects beyond that of the other. 

I have made an analysis of certain cultures with relation to 
this matter, with the following results: 

First, as we have before seen, observation shows that the two 
members of a pair are by no means always equal, but that num- 
bers of unequal pairs are found. A number of such are shown in 
fig. 15. Cases of extreme inequality sometimes occur, but such 
are rare. In one of the pairs of lot 2 (table 1), the anterior tip 
of one individual extends forward thirty microns beyond that 
of the other —that is, about one-fifth of the length of the latter. 
Fig. 15, 6, shows a pair in which the smaller is less than three- 
fourths the length of the larger. 

Unevenness at the anterior ends. In four lots of conjugants I 
undertook to measure the differences between the anterior tips 
of the pairs. The measurements taken are necessarily somewhat 
gross in comparison with the minute absolute amounts that 
one individual projects beyond the other, but by using large 
numbers we may get results that will be accurate enough to 
indicate the real conditions. We may call the individual that 


CONJUGATION IN PARAMECIUM 53 


Fig. 15 Unequal pairs. a tod, pairs of the aurelia race c; e and /, pairs of cau- 
datum; g and kh, pairs of the aurelia race C2. a, b, and c give each two views of 
certain unequal pairs. X 333. 


54 H. 8. JENNINGS 


projects farther anteriorly A, the other 6; the measurements 
then show, beside the total length of A and B, the amount that 
A projects in front of B. One of the four lots thus analyzed 
was a ‘wild’ culture, the others were pure races. 

The wild culture examined was lot 2 (table 36); the unit of 
measurement in this case was 2 microns. Pairs in which the 
difference at the anterior ends was less than 1 micron were con- 
sidered even. The results for lot 2 are given in table 13. 


TABLE 13 


Distance that one member (A) of a pair projects in front of the other (B), in a ‘wild’ 
culture (lot 2, table 36) 


5 
TOTAL NUMBER | MEAN LENGTH 


OF ALL, IN 
Ces MICRONS 
Projection of A in mi-/ 0| 2, 4/ 6{| 8 | 10/ 12 | 14| 16 | 18| 30 
CIOUSssasepislewiclecis ea | | 
Number of cases......|61| 9 23/15/15|10| 4) 2/1] 1] 1 We Hie 
Per cent of total 
number of pairs.....|43.0) 6.3)16.2)10.6:10.6) 7.0) 2.8) 1.4) 0.7) 0.7) 0.7 


Thus of the entire 142 pairs, 61 (or 483 per cent) were even at 
the anterior tips, while 81 (57 per cent) were more or less un- 
even. The amount of projection of A was rather slight; in 123 
pairs, or 86.6 of all, it was less than 5 per cent of the mean length 
of the individuals. 

From pure races there were examined from this point of view 
one lot from race C, and two lots from race g. These are both 
races of aurelia, and of about the same size, the mean length of 
the conjugants falling, in all three lots, between 118 and 124 
microns. In these cases the unit of measurement was 4 microns 
instead of 2. The facts are given in table 14. 

In these cases we find respectively 33.3, 45.2 and 46 per cent 
in which the anterior ends are even. The amount of difference 
at the anterior end is again rather slight. A difference of 8 
microns is about 6.5 per cent of the means length in these cases; 
this difference is exceeded, in the three lots, respectively by 11.5, 
1.2 and 8 per cent, of all. 


CONJUGATION IN PARAMECIUM 55 


TABLE 14 
Amount that one member (A) of a pair projects beyond the other (B), in three lots 
from pure races 


| 
NUM 
TOTAL UMBER| MEAN LENGTH 


OF PAIRS | 
| | 
Projection of A, in microns..... 0 4 8 12 16 20 | 
Lot 17: Race C2 , | | | 
Number of cases...........- 23 24 14 E 3 69 121.91 
Per cent of all.....:........5. | 33.3 | 34.8/ 20.3) 7.2 4.3 
Lot 18: Race g | | | 
Number of cases 4 .+..| 38 37 8 1 84 123.37 
Per cent of all... sine of 45.2.)/) 4450/1) -925)| 12 
Lot 19: Race g | | | | 
Number of cases.... ..-| 40 30 104} 56 1 87 118.28 
Percent otall...............-| 46.0 | 34.5 | 11.5] 6.9 | 14 


Now, from the theory as to the cause of the correlation, we 
should expect to find that the individual A, projecting in front 
of B, is as a rule larger than B. (We shall later show that the 
length from anterior tip to mouth is so correlated with the entire 
length that this must be true.) This being so, we should per- 
haps expect that the individual A, which projects farthest for- 
ward, should likewise project farthest backward, thus overlap- 
ping B at both ends. (If however the two specimens merely 
came together at random, and any parts of the oral surfaces 
united, then there is no reason why the specimen projecting ante- 
riorly should be larger, and as a rule the specimen that extended 
farthest forward would not extend so far backward,—one speci- 
men being merely displaced forward as a whole, with reference 
to the other). 

Examination of a number of cultures from this point of view 
shows that as a rule it is true that the specimen extending far- 
thest forward is the larger and likewise extends farthest back- 
ward. The facts for five cultures are given in table 15. 

As the table shows, the specimen A, projecting anteriorly, is 
larger than B in from 83 to 91 per cent of all unequal pairs, while 
it is smaller than B in but 2 to 11 per cent. Further, A projects 
beyond B backward as well as forward in 51 to 67 per cent of 
all, while B extends beyond A in the rear in but 10 to 28 per cent 
of all. 


56 H. S. JENNINGS 


TABLE 15 


Proportional number of cases in which the individual A, which projects in front, is 
larger than B, and projects behind it as well as in front 


A LARGER A SMALLER AL ae pete fea) 
TOTAL | NUMBER | ; i 
Lor NUMBER | OF PAIRS ar 3 —— = 
| OF PAIRS | UNEQUAL | Abso- Abso- 5 Abso- Abso- 

lute | cont | 'ae | cent | Inte | cont | Ite | dene 
2 142 81 68 | 84.0 9 11.1 54 | 66.7 23 «| 28.4 
7 79 «| 36 30) 83.3 5 13.9 | 26 | 72.2 6 16.7 
17 69 | 46 40 87.0 2 4.3 | 28 | 60.8 8 17.4 
18 84 46 41 89.1 1 2.2| 31 67.4 5 10.9 
19 87 47 43 | 91.5 1 2.2)| 24 | 61d 5 10.6 


Thus, where one member of a pair extends further forward 
than the other, that member is usually larger. We should there- 
fore expect that in pairs where one member extends further for- 
ward than the other the difference in length between the two 
members would be greater than in the case where the two are 
even at the anterior end. We should further expect that the 
difference in size between the two members would be greater, 
the greater the amount that A projects anteriorly beyond B, 
The facts with regard to these points are given in table 16. 

The difference between the two members was taken in units 
2 or 4 microns, in different lots; in the table the grouping is by 
intervals of 4 microns. The number of pairs showing the larger 
differences is of course small; on this account I have thought it 
well to give the probable errors, as well as the number of pairs 
in each case. 

The table shows clearly that the difference between the mem- 
bers of the pairs is greater in pairs in which one individual pro- 
jects anteriorly in front of the other, and the greater the projec- 
tion, the greater the difference. These things are by no means 
matters of course; if it were merely necessary for the animals 
to coalesce by any part of the oral surfaces that came in con- 
tact, they would not be true. 


Ou 
“I 


CONJUGATION IN PARAMECIUM 


TABLE 16 


Average difference in length, in microns, between the members of pairs, in relation to 
the distance that one member (A) extends in front of the other (B) 


Ze x | 
Mand 
= = ro) < 0 | 2-6 | 6-10 10-14 14 18 18-22 
ogks 
ZRz 5 (ANTERIOR 
<0°0 Paine 
B2&yz| ENDS EVEN) 
2 ame 
a 
- on | oD 
8 2 2 2 2 5 | 5 
& = 2 | 3 z PS 
la res a MEAN & MEAN a MEAN Pa MEAN | ~ | EXCESS 
2 Leanne is eee oF) & (es OF | 5 eieaes OF iS Recess Bi cee 5 
% p INTENGTH) 3 | LENGTH | § | LENGTH 8 | LENGTH &% | pencta | § Lenora & 
45 a Aa ms a | a 
alas 2 2 3 2 a | = 
3 |o% 2 5 4 D 5 b 
<a Z| Zz z Zz Zz Zz 
| 
| | é 
2 142 | 8.62£0.70 | 61 | 10.20+1.37 40 | 12°30+1.57 27 18.80+3.35 10 9.33+3.50) 3) 34.00 1 
7| 79 | 8.56+0.66 | 43 | 10.46+1.26 26 | 12.00+2.02 9 12.00 1 
17 69 6.96+0.94 | 23) 6.00+0.79) 24 | 17.16+1.20 14 | 19.20+1.13) 5 | 16.00+5.52) 3 
18 | 84 3.79+0.35 | 38 |) 8.86+0.76 37 | 15.00+1.00 8 16.00 1 
19 87 5 0 20.00+1.39 6 9.00 1 


6440.52 | 39 | 7.87+0.84 31 | 15.20+1.45 1 
| | 


We may now pass to an examination of the correlations in the 
uneven sets, as compared with those for the even sets, and for 
the cultures as a whole. As before remarked, we should expect 
the correlation to be greatest in pairs that are even at the anterior 
end, least in those uneven at the anterior end. This is because 
pairs of the former sort will, from what we have already shown, 
have members nearly equal, which is the condition that produces 
high correlation; while the latter set will be unequal, giving low 
correlation. The facts for four cultures are given in table 17. 

To understand table 17, it is necessary to recall the method 
of computing correlation in the case of pairs composed of two 
similar members. In all such cases the correlation is referred 
to the mean of all the individuals; it shows whether, when one 
individual diverges from the common mean of all, the other in- 
dividual tends likewise to diverge from this mean in the same 
direction. The correlations computed in this way are shown in 
the first three columns of the table; only the values given in 
these three columns are strictly comparable. As shown in 
these three columns, the correlation in the case of the pairs in 


58 H. S. JENNINGS 


which one individual extends anteriorly beyond the other (as 
in fig. 15, b, c, e) is much less than in the pairs that are even at the 
anterior end, when the correlation is measured in the same way 
in the two cases. It is also much less than in the lots taken as 
wholes (column 1). In three of the four lots no correlation is 
detectable in the uneven pairs (column 3), the coefficients com- 
puted being less than the probable errors. 

In the case of the uneven pairs (such as shown in fig. 15), the 
two members have of course a distinguishing mark, since the 
individual A projects in front of the individual B. We may 
then properly ask whether the assortative mating does not show 
its effect even in these cases, if we compare the A’s with the B’s. 
As the projecting individual A becomes larger, does not the other 
individual B likewise become larger? This is not shown by the 
correlation computed with reference to the common mean of all, 
as in the first three columns. To discover whether, when A 
increases in size, B likewise tends to increase, we must compute 
the correlation using the independent means and standard de- 
viations of A and B, as in ordinary correlation work with unlike 
units in thetwoclasses. Theresults areshown in the fourth column 
of table 17. Here we see that the correlation is rather high; as 
A becomes larger, B likewise becomes larger.? The effects of 
the assortative mating are therefore seen in the pairs that are not 
even at the anterior ends, as well as in those that are. 

Another point is worthy of particular notice. Although the 
correlation in total length, in the case of pairs that are even at 
the anterior end, is greater than for the culture as a whole, it - 
is by no means complete, or even approximately complete. Now, 
in these pairs, the distance from anterior end to mouth is the 
same in the two individuals. Since the correlation in total 
length is not complete, it is evident that the remainder of the 
length, behind the mouth, is not the same in the two individuals. 
This is further shown in table 16, which gives in the first column 


9 In the case of the infusoria in which it is known that a large individual regu- 
larly mates with a smaller one, probably the correlation would have to be com- 
puted as in the fourth column of the table. This would certainly be the case if 
the two members had other distinguishing marks. 


CONJUGATION IN PARAMECIUM 59 


TABLE 17 


Correlation in length of members of pairs (1) in which the two anterior ends are even; 
(2) in which the anterior end of one individual (A) extends in front of that of the 
other (B) 


CORRELATION , -=4 ANTERIOR END OF A PROJECTING IN FRONT 
& |FOR ALL PATRS ANTERIOR ENDS EVEN CREATOR 
g |— — : 
a | EB a 3 4 
2 | ° ° 
z % = Correlation re- | Correlation referred to 
= 2 1 2 ag ferred tomean ‘the separate means 
i. a5 Be of all (A + B), of A and B 
Pa [led Mises == as in columns 1 
3 a a a and 2 
| bs = —— 
2 36 | 142 | 0.268+0.057 61 0.547+0.043 34 —0.033+0.082 0.315+0. 104 (see note) 
7 40 79 | 0.333+0.048 | 43 0.449+0.058 36 0.232+0.075 0.552+0.078 
17| 49 69 | 0.318+0.052 47 0.591+0.045 22 |—0.091+0.101 0.318+0.052 (see note) 
18 50 84 | 


0.251+0.049 38 0.762+0.032 46 | 0.032+0.070 0.503+0.074 


Note: In lots 2 and 17 the method of subdivision was varied, in order to deter- 
mine the effect on the correlation. In lot 2 the set with ‘anterior end of A pro- 
jecting’ includes only those in which A extended beyond B eight microns or more. 
In lot 17 the set ‘anterior ends even’ includes those in which the anteriorend ofA — 
extended not more than six microns beyond that of B, the other set the rest. In 
the other lots the division was as precisely as possible into even and uneven sets. 


the average differences in total length for the individuals whose 
anterior ends are even. Thus conjugant individuals that are 
equal in length from the anterior end to the mouth are not neces- 
sarily equal in entire length. This might be due either to (1) an 
equalizing of the parts in contact, in individuals where they were 
not originally equal, or (2) to variation in the proportion of the 
parts anterior to and posterior to the mouth, in different individ- 
uals, before conjugation began. The point will come up again 
later. 

Thus, we have shown that the incompleteness of the correla- 
tion in length between the members of pairs is due (1) partly to 
the fact that small differences between the individuals do not 
much affect the mating, while large differences do; (2) partly to 
the fact that while in certain pairs the anterior ends are placed 
evenly in juxtaposition, giving high correlation, in others they 
are not, giving little or no correlation. But even in the latter 
cases we found that there is marked correlation if we ask whether 


60 H. S. JENNINGS 


when A increases in size, B does not increase also, thus referring 
the correlation to separate means. 

We shall find other factors affecting the degree of correlation, 
in the next matter to be treated. 


CAUSES OF THE CORRELATION 


From our account thus far, there can be no doubt of the exist- 
ence of some degree of correlation in size between t1e members 
of pairs, in the collections of conjugants which we measure. We 
have seen that Pearl explains this as due to assortative mating. 
But this is not the only condition which might produce such cor- 
relation, so that we must now enter upon a critical examination, 
with experimental tests, of the various possible explanations. 
The following have been suggested, or are possible causes: 

1. Assortative mating. 

2. Equalization during mating. 

3. Change of size during union. 

4. Differential contraction due to the killing fluid. 

5. Local or temporal differentiations in the culture from which 
the pairs are taken. 

These we shall take up in series, concluding with a discussion 
of the hght thrown on this matter by the correlation in breadth. 


1. Assortative mating 


We have already seen much of the evidence for holding that 
the correlation is due to actual assortative mating. Here we 
enter upon a more precise analysis of the relations involved. 

If assortative mating occurs and is the cause of the correlation, 
as set forth by Pearl, then, as Pearl remarks, it depends primarily 
on the relative lengths of the animals from the anterior end to the 
mouth, since these are the regions that must fit together in mat- 
ing. The correlation between the total lengths of the conjugants 
would be a secondary result of this; it would be due to the fact 
that ‘‘individuals in which the distances from the anterior end 
to the mouth are equal would not be greatly different in total 


CONJUGATION IN PARAMECIUM 61 


length, and hence their lengths will be correlated”’ (Pearl’07, 
p. 267). That is, the total length should be on the whole pro- 
portional to the distance from anterior end to the mouth; in other 
words, there should be a marked correlation between these two 
dimensions. 

Examination of our figures will show that this proportionality 
is by no means invariable or absolute. In fig. 15, d and f, for 
example, where the distance from anterior end to the mouth is 
the same for the two members of each pair, the total lengths are 
very different. 

a. Correlation between the length anterior to the mouth and the 
total length. In order to determine what degree of correlation 
actually exists between the length anterior to the mouth and the 


| 

} 

u . 

d fo b a 
Fig. 16 Diagram to illustrate the measurements taken. See text. 


total length, I made measurements of the two dimensions in ques- 
tion, and of certain other dimensions, in a number of lots of con- 
jugants and non-conjugants. The other measurements made 
were, for certain cases, the distances from the anterior end to the 
point where the bodies of the two conjugants separate behind. 
The measurements made and the constants deduced from them 
throw an important light on conjugation, as well as on the general 
variation of Paramecium. They are therefore presented in 
table 18. 

The measurements given in table 18 will be understood from 
an examination of fig. 16, where the measurements taken are 
shown for one individual of the pair. The measurements are; 
(1) the total length; (2) the distance from the anterior end 


62 H. S. JENNINGS 


TABLE 18 


Dimensions, proportions and correlations of certain lots of conjugants and non- 
conjugants, with relation to the problems of assortative mating. The two individuals of 
a pair are denominated A and B 


j 


WILD CULTURE, LOT] | RACE G, Lots 18 anp 19 
= —_—_ — = —j— = ——— 
| 168 174 118 non- 
158 94 non- conjugants, conjugants, | conjugants, 


conjugants, | conjugants, | 


(table 40) (table 41) lot 18 lot 19 lot 19 


(table 50) | (table 51) (table 52) 


| 
———— —= 
| 
| 


Mean! .-<a.octeare 168.71+0.63 | 183,791.24 | 123.57+0.41 118.28+0.52) 135.35+1.02 
= Standard deviation) 11.66+0.44 | 17.79+0.88 | 7.80+0.29 10.11+0.37) 16.49+0.72 
(A) Total length Coeffi The | 
Soefficient of varia. | 
(tions. te ct aoasme | 6.91+0.26 9.68+0.48 | 6.31+0.23 §.55+0.31) 12.18+0.54 
Means tascmyts.< ea | 108.81+0.34 | 115.49+0.70 | 74.21+0.30| 82.61+0.62 
(B) Length in Standard deviation| 6.30+0.24| 10.11+0.50_ 5.77+0.21| 9.98+0.44 
front of mouth | Coefficient of varia- | 
HON erections 5.79+0.22 8.760.483 7,780.28) 12.08+0.54 
Meanie sn Ge sececncs | 59.90+0.44 | 68.30+0.69 | 44.07+0.32 52.75+0.50 
(C) Length  be- |Standard deviation! 8.24+0.31 | 9.92+0.49 | 6.31+0.23) 7,980.35 
hind mouth Coefficient of varia- | | 
[eRetlonexe-S5..¢ ct 13.76+0.52 14.53+0.73 14.31+0.52) 15.14+0.68 
(D) Length to {Mean...... saiteseece | | 99.77+:0.39|  92.78£0.47| 
poin: where the } Standard deviation | | | 7.62+0.28 8.68+0.32) 
pairs separate | Coefficient of varia- | 
behind lon scnecisciseees= 7.54+0.28 9.360.385 
(E) Length Bee | Mea tiiaiassarciastesaceiaeete 23.81+0.32) 25,580.37) 
rior to point | Standard deviation | 6.22+0.23) 7.18+0.26) 
where psirssep-| Coefficient of varia-_ | | | 
arate behind | tion.............. 26.10+1.02 28.09+1.11) 
(F) Proportion of entire length lying in| | | | 
front of mouth (mean index)........ | 65.00+0. 1676) 62.95+0.19% | 62.85+0.15%| 61.10+0.19% 


| 
(G) Coefficients of Correlation | 


1. Total length of individual A—| 


WIG SAME OF Bs 5.1... 0jc:0sie0 cielz sins e 0.333+0.48 | | 0,251+0.49) 0.3230. 046 
2. Part before mouth in A—with: | | 
BALES lara a crcicvoinnnieleiiee ciaioate 0.840+0.016 | | 0.626+0.031) 


3. Part behind mouth in 
SATOCUSAINE IT, Bi. s vsctsieiesse.e vate 
4. Part before mouth, with total | 
length of same individuals.....) 0.743+0.024 | 0.894+0.014 | 0.832+0.016 0.919+0 010 
5. Part before mouth, with part 
behind mouth of same indi-) | 
VAGUAIS ceria ccreisciecere anne | 0.261+0.050 | 0.570+0.047 | 0.383-+0.046) 0.671+0.034 
6. Part anterior to point of separa- 
tion with total length of same) | 
MAL VIG WA se scieejeie1aeyniocejeie.t a!ioie'6 | | 0.676+0.028 0.723+0.025 
. Part anterior to point of sepa-| | | 
ration with part posterior to | 
SAMOS. ene asecestet seer eee jet aoa Oates 


0.0360.053 | 0.1570.050) 


_ 


CONJUGATION IN PARAMECIUM 63 


a to the mouth b (the distance was taken to the posterior 
angle of the mouth); (3) the distance from the mouth (posterior 
side), to the posterior tip (b-d, fig. 16); (4) the distance from the 
anterior tip a to the point c, where the bodies of the two conju- 
gants separate behind; (5) the distance from this point ¢ to the 
posterior tip d. These two dimensions (4) and (5) were not taken 
in all cases.t° The measurements illustrated in fig. 16 were of 
course made for both members of the pairs, as well as for non- 
conjugants, so far as applicable to these. Such measurements 
were made for one wild culture, of unknown racia] composition 
(lot 7 of table 1), and for two lots of the pure race g (lots 18 and 
19 of table 2). The results are given in table 18. 

The first point in which we are interested is the correlation 
between the entire length, and the length from anterior tip to the 
mouth, as given at G, 4, in the table. This correlation is high, 
the coefficients being 0.743 and 0.832, respectively, for the two 
lots of conjugants, 0.894 and 0.919 for the non-conjugants. It 
would therefore seem amply sufficient to account for the degree 
of correlation found between the total lengths of the two members 
of the pairs. Indeed, possibly we should expect, on this basis 
alone, that the latter correlation should be higher than it is, since 
it is but 0.333 and 0.323 in the two lots of conjugants Just men- 
tioned. We have however seen certain reasons why the corre- 
lation in total length is relatively low, and a further study of the 
data of table 18 gives other grounds for this. 

b. High variability of the posterior region. We find in the con- 
jugating pairs that while the length of that part of the body lying 
in front of the mouth is relatively uniform (coefficients of varia- 
tion 8.76 and 7.78 in lots 7a and 19a), that part lying behind the 
mouth is much more variable (coefficients 13.76 and 14.31 in lots 
7a and 19a); the variation is almost twice as great in the latter. 


10 Tt is often very difficult to determine the eyact position of the mouth, so that 
it was at first thought that the point where the two members of the pair separate 
behind (c, fig. 16) would serve the same purpose. This turned out not to be the 
case, since the two do not follow the same rules. It was later found that by plac- 
ing the animals in weak glycerine the position of the mouth could usually be 
determined. The measurements to the point of separation have some interest in 
themselves, and are therefore given in the table. 


64 H. S. JENNINGS 


c. Low correlation of parts before and behind mouth. Further- 
more, the correlation between the length in front of the mouth, 
and that behind the mouth is low, amounting to but 0.261 and 
0.385 for the conjugants in question (table 18, G, 5). Thus two 
individuals that are equal in the distance from anterior end to the 
mouth will often be unequal in the remainder of the body. As a 
result, although they conjugate accurately, they will be unequal 
in total length, thus tending to make the correlation of length 
to length less than 1.00 in the collection to which they belong. 
This point came up earlier, in connection with tables 16 and 17 
(p. 58). 

d. Correlation of pairs greater for parts anterior to mouth. The 
point just made is further illustrated by another fact. If we com- 
pute for the pairs the correlation of the lengths anterior to the 
mouth in the two individuals (table 18, G, 2) we find this correla- 
tion much higher than that for the total lengths (G, 1). The 
correlation of the two individuals for lengths anterior to the mouth 
is 0.840 and 0.626, respectively, as against 0.333 and 0.323 for 
total lengths. The great decrease in the latter as compared to 
the former is evidently due to the variable proportions of the 
posterior part of the body to the anterior part,—this variable 
proportionality being either original, or produced by equaliza- 
tion of the parts anterior to the mouth during conjugation (a 
matter for later consideration). 

e. Low correlation of parts posterior to mouth. Conversely to 
the point made in the preceding paragraph, we find the correla- 
tion of the parts posterior to the mouth for the two members of 
a pair (table 18, G, 3) to be much less than that for the total lengths. 
In lot 7a correlation between the posterior parts is lacking or too 
slight to be detected (coefficient, 0.036 +0.053); in lot 19a it is 
very small (coefficient 0.157 +0.050). 

The correlation of total length in the two members is then due 
almost entirely to the correlation between the distances from 
anterior tip to the mouth in the two. This correlation might be 
due either to equalization in the process of uniting, or to assorta- 
tive mating; a point which we shall take up in a moment. 


CONJUGATION IN PARAMECIUM 65 


f. Region in front of the point where the two conjugants separate. 
With regard to the distance from the anterior end to the point 
where the two individuals separate behind (a-c. fig. 16), table 18 
shows (1) that this distance is more variable than that from the 
anterior tip to the mouth (and more variable than the total length) 
(lot 19); (2) that its correlation with the total length (table 18, 
G, 6) is less than in the case of the distance from anterior tip to 
mouth; (3) that the length of the part behind the point of separa- 
tion is extremely variable, the coefficient of variation rising to 
26.10 and 28.09 in the two lots examined; (4) that the correlation 
between the length anterior to the point of separation, and that 
behind the point of separation is negative (table 18, G, 7),—so 
that on the whole the greater the length of the region where the 
two animals are in contact, the shorter the region behind this, 
where they are free. All this shows, of course, that the point to 
which the union of the two individuals extends is not a constant 
and predetermined one, but that on the contrary, the relative 
length of the united regions differs greatly in different cases. 
If this were not the case, the correlation between the length 
anterior to this point and that posterior to it would be positive, 
as is the case for the length anterior to the mouth and that pos- 
terior to it. Evidently, while in conjugation there is always 
union as far back as the mouth, the further distance to which the 
union extends is extremely variable. 


2. Equalization during mating 


Certain important points as to what occurs in conjugation and 
how the positive correlation is brought about, appear on compar- 
ing in table 18 the conjugants with the non-conjugants. 

We find that about the same proportion of the body lies in front 
of the mouth in both conjugants and non-conjugants, the mean 
proportion varying from 61 to 65 per cent in the different cases 
(F, table 18). In the conjugants, as we have seen, this region 
anterior to the mouth is much less variable than that posterior 
to the mouth,—the coefficients of variation from the former being 
5.79 and 7.78, as opposed to 13.76, 14.53 and 14.31 for the latter. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 1 


66 H. S. JENNINGS 


But in the non-conjugants the variability of the part anterior 
to the mouth is much less reduced; the coefficients are 8.76 and 
12.08, as opposed to 5.79 and 7.78 for the corresponding conju- 
gants. The variability of the part posterior to the mouth is 
about the same in the non-conjugants as in the conjugants (co- 
efficients for the former 14.53 and 15.14; for the latter, 138.76 
and 14.31). Thus, in the conjugants the variability of the region 
antertor to the mouth has been reduced, while that posterior to the 
mouth has not. Furthermore, the correlation between the length 
anterior to the mouth, and that posterior to the mouth, (table 18, 
G, 5) is much diminished in conjugants (coefficients of correlation 
for the conjugants, 0.261 and 0.383, as compared with 0.570 and 
0.671 for the corresponding non-conjugants). Remembering 
that it is the parts anterior to the mouth that become fitted 
together in conjugation, while those behind the mouth do not, it 
is clear that these two facts indicate strongly that there has been 
an equalization of the two regions in contact, by curving, contrac- 
tion, stretching, ete., for this would reduce both the variation, 
and the correlation with the unchanged posterior part, exactly 
as we find to be the case. It would likewise result in making 
the correlation between the posterior parts of the two conjugants 
much less than that of the anterior parts, as table 18 shows to be 
the case. Such equalization we have before seen to occur, by 
direct observation of the conjugants (see above, p. 27); here we 
see that it affects the measurements and constants to a marked 
degree.!!. This agrees with a number of facts previously set forth 
in this paper; we have repeatedly had occasion to point out that 
an equalization in the region anterior to the mouth would wholly 
or partly account for some of the relations described. To it are 
partly, if not mainly, due the fact that individuals which con- 
jugate with anterior ends even are not completely correlated 
in length (p. 59); the related fact that conjugants that are equal 
in the region from anterior end to the mouth are not equal in the 


1 It is perhaps worth while to note the fact that the account of the equalization 
based on direct observation (pp. 26, 27) was written exactly as it stands, before 
the computations that indicate the same thing were made, and before I had even 
suspected that the measurements would show it. 


CONJUGATION IN PARAMECIUM 67 


remainder of the body; the much greater variability of the part 
posterior to the mouth, as compared with that anterior to the 
mouth, in conjugants; the low correlation between the part 
anterior to the mouth and that posterior to the mouth, in conju- 
gants; the very much greater correlation between the lengths of 
the two conjugants from anterior end to mouth, than between 
total lengths, or between parts posterior to the mouth. 

We have here a fact of capital importance; there occurs in con- 
jugation a process of equalization, of such a character as would 
produce correlation in length between the two individuals, even 
if such correlation did not exist before the two united. And this 
equalization affects strongly our statistical results. 

The question of course at once arises: Is not this the entire 
explanation of the correlation between the members of pairs? 
Is there any reason for assuming assortative mating at all? 


Correlation of members of pairs after separation. To answer 
this question requires further experimentation. If the correla- 
tion is due to the contraction, extension and curving of the body 
undergone in the process of union, it will be found only in the 
animals that are actually united. As soon as the contraction, 
curving, etc., ceases, the correlation will disappear. Now, care- 
ful observation shows that the contraction, curving, etc., do cease 
after the pairs separate; the animals regain completely their 
normal form, and increase much in size, as we have seen. If 
therefore, we measure the members of the pairs some consider- 
able time after they have separated, we shall be able to determine 
whether the correlation is due merely to the actual conditions 
during the period of union. 

I have carried out this operation in a number of cultures, some 
of them pure races, others ‘wild’ mixtures. For this purpose 
the pairs are isolated in hollow ground slides and kept, all under 
identical conditions, till the two animals separate. After sep- 
aration they are. further kept until about time for the first fission 
to occur. Then the pairs are killed separately and measured, 
usually in comparison with a considerable number that had been 
killed before separation. In this way, further, one can study 


68 H. S. JENNINGS 


the correlation of the parts anterior and posterior to the mouth, 
in the separated conjugants, determining whether the peculiar- 
ities in these dimensions found in the conjugating individuals 
are due to the fact of union or to actual differentiation of the con- 
jugants. 

The constants for variation and correlation for the two members 
of pairs that have separated some time before the measurements 
were made are shown in table 19, together with comparative data 
for the unseparated conjugants. 

The pairs were isolated in the morning. A certain number of 
them were killed at once. The rest were kept alive. At 6 p.m. 
it was found that the pairs were separating, many being already 
apart. The animals were then kept till noon the next day. 
Thus they were killed at least twelve, and probably eighteen, 
hours after separation. 

Examination of table 19 shows clearly that the correlation is 
not due to the contraction, extension and curving involved in 
the process of conjugation, nor does it depend in any way upon 
the state of actual union, for it persists undiminished for twelve 
hours or more after separation has occurred. Careful examina- 
tion of the animals shows that the contraction, curving, etc., 
has quite disappeared in the separated individuals (and this is 
further demonstrated by the measurements to be set forth later). 

Indeed, not only is the correlation undiminished in the sepa- 
rated pairs; it is actually greater in every case than in the pairs 
still united. This greater correlation after separation might be 
due to a number of different causes: 

a. Possibly it indicates that the alterations in form during con- 
jugation actually decrease the correlation instead of increasing 
it. If the correlation of the undistorted members were consider- 
able, then irregular contraction and curving such as occurs in 
the process of union, would decrease it. 

b. A more probable ground for the greater correlation of the 
separated pairs lies in the greater accuracy with which they can 
be measured. The two members of a united pair are not 
accurately parallel, but lie oblique to one another, owing to the 
obliqueness of the oral grooves by which they are united. This 


69 


PARAMECIUM 


IN 


CONJUGATION 


| 


Te 0*L0 ZI 


¥£0 OF TEZ 082 0F80 6 SF 0F96 ZETS9I-OOT SI+L9SLT ge | UoIyerEdas 1oqye 
| | sanoy ZI ynoqe 
| parry syuvsntuoD (61 “4ydag 
| (wyjaane) y 990Y cz 
£80 OF ZEF'OGT OF79'9 82 0F6S°6 OF OFLE HTS9T-88 I19+09FET 's9z wa ieee eae mOT}, 
| | | | beredas 1ayye sun0y 
| | | | | ZI qnoqe aurg'q (ZT ‘3dag 
| | | riences payun 
TF OF0IZ'0 G1 0FFE'9 83° OFFS'L LE OFZ 'SITFFI-OOT | 6G [eer [FFs {ITS Syuvsnfu0H “IT “qdag 
| | | | (wyoune) Yy aoe | 14 
9£0'0+82 082 OFFS'6 9S°OF8T OI GL°OFFG OOZCSC-ZEI | Lg |FET 992 | eo aE 
| | H | -eibdas 19}jesInoy 
| | | Zl ynoqe amug'q |7z 4dag 
180 OFSFZ'09T'0F08'S 62°0FZF OT IF O08 6LTSOZ-CST €Z+C78FT 962 | peyun | 
| | | | TMs szuvsnfuopH *e | 1z ‘3dag 
| | UINZBpPNBs: PITA — | &% 
FOOT TIF O66 006 4 6F 0OFL6 91 69 OFOF ZIZ09C-9ST | 99 SET 1922) | Gaiman ae “mol; | 
| | | | | peindasaayyesinoq | 
| | | ZI qynoqe oureg-q | 7 4ydag 
TF0'0+66E'061 OFZL°S FE OFSO'OT SF OFFI GAT8OZ-ZST gS ZOT FOZ | srs paqun | 
| | [Ws siuvsnftuop -e | Tg “any 
| | | WUNYBPNV | PITAL re 
| | 
3 8 2 x vw | a | vlig | 9 2 
a0 of 8 =] bee lll “ee Ietetil tes aS ° 
28 aie | ee B 6| 2 | & lreige ea |" 
Bel 43 <0 Zz 2g 8 [5 o)45 —) 
Eo >9 a c i Felice! 
Biel fs 53 5 BO eeeahone 
> 5 Bg Zz ag 
ee aE 5 s | 8 ges 
oe ies Zz 8 | 2 z | 
2° g A | 17) | 
5 Zz { | | | 


61 AIAVL 


Pa}VUN [US apy pa7]r4y 
sainjjno aps ay? wouf sind 1of asoy, yyim paundwoo sp ‘uoynundas sajfo s.inoy awos 
paqy squnbnluoo fo survd fo ssaquawu Lof yjbua) ur W01D)]aLL09 pup woUDLLDA fo sjuDj}sUo) 


70 H. S. JENNINGS 


obliqueness is very evident when the pairs are examined with a 
binocular microscope; it is indicated in fig. 3, 7 and k. Asa 
result of it, when the entire length of one member of a pair is 
shown, the other will be a little foreshortened. Thus the measure- 
ments taken will indicate that the two are less nearly equal than 
they really are, and this of course will reduce the correlation. 
In the case of the separated pairs, this difficulty is not met, since 
the two individuals are measured separately. 

c. Another cause, which might actually increase the measured 
correlation after separation, would lie in the increase of size which 
takes place after the separation of the pairs (see page 21). If 
the interval of time since separation varies in the different pairs 
(as is of course the case), then some will have increased in size 
more than will others. But the interval since separation will 
of course always be the same for the two members of a given pair. 
The result will then be a series of pairs differing in size, giving 
thus marked positive correlation between the members. It 
seems probable that this is the chief ground for the actual increase 
in correlation after the pairs separate. 

Observation shows however that the relative sizes during union 
are retained after separation (so far as this can be shown with- 
out actual measurements of the living united pairs). In many 
cases I noted that certain pairs, when first seen, were unusually 
large or unusually small, or unequal. When they were meas- 
ured, twelve to eighteen hours after separation, this was still 
true. 

In any case it is clear that the correlation in the united pairs 
is not due to the temporary change of form during union, since 
it persists after union ceases. The correlation is not due to the 
equalization in mating. 

Change in variability and in correlation of parts after separation. 
In a preceding section we saw that the individuals of the united 
pairs differ very markedly in certain respects from non-conjugants. 
We have now an opportunity to determine whether these pecul- 
iarities of the united conjugants are due to a differentiation of 
these conjugant individuals, existing quite independently of the 
act of union; or whether they are merely the result of the exist- 


CONJUGATION IN PARAMECIUM ial 


ence of the union, with its accompanying change of form. These 
peculiarities of the conjugants were as follows: 

a. The variability of the part anterior to the mouth is greatly 
reduced, as compared with the case of the non-conjugants. 

b. The correlation between the length anterior to the mouth, 
and that posterior to the mouth, is greatly reduced in the united 
conjugants. 

ec. The correlation of the two members of a pair is much greater 
for that part of the body anterior to the mouth than it is for the 
total length of the body. 

If these peculiarities are, as our previous examination had 
seemed to indicate, merely the result of the changes during union, 
then they should disappear after the pairs have separated and the 
individuals have reassumed their usual form. We may therefore 
get light on this matter by comparing these relations in the con- 
jugants before separation and in those after separation. The 
data for this comparison are given in tables 18 and 20. These 
show the following: 

1. The variability of the part anterior to the mouth. is not 
ereatly changed after the conjugants separate (table 20, B). 
In the wild culture of lot 22 the variability did indeed increase 
after separation, from 5.60 to 7.28, but in the pure race (lot 24) 
it remained substantially the same. Of course the separated 
conjugants include no young specimens, while the non-conju- 
gants (table 18) include both young andold. It is therefore not 
to be expected that the variability will reach the same degree in 
the former as in the latter. The data do not indicate that there 
is any great change in the variability of this part, owing to the 
mere act of conjugation. 

2. The reduction of the correlation between length anterior 
to the mouth and that posterior to the mouth quite disappears 
after the cessation of union. In the united conjugants we find 
in tables 18 and 20 coefficients of correlation between these parts 
of but 0.261, 0.383 and 0.277, while in the separated conjugants 
the correlations are 0.649 and 0.488 (table 20, G, 5). The same 
lot that shows a coefficient of 0.277 while conjugating, has a ec- 
efficient of 0.488 after conjugation has ceased. This is additional 


2 H. S. JENNINGS 


TABLE 20 


Dimensions, proportions and correlations of conjugants after separation, as compared 
with those for pairs still united. The two individuals in a pair are denominated A and 
B. Compare table 19 


(WILDCAUDATUMOFAUG.| WILD CAUDATUM OF | RACK k (AURELIA) OF | tees 
3l-sepT. 1, 1910 SEPT. 21-22, 1910 SEPT. 11-12, 1910 Oman 
Lor 22 Lor 23 Lor 24 19, 1910 


| Conjugants 134 Pairs 134 Pairs | 178 pairs 

102 pairs | after sep- ai eae, after separ-. ee ates after sep- | after sep- 
/still united |aration, 276) > Tables ation | Tables aration | aration, 
Tables Tables 60+ Tables 


Table 55 | specimens FE 
Table 56 | 22+23 | 57 and 58 | 5% 62, 64 |61 ¢3, 65, 66| 67-+68 
| | | 
(Mean......... 176. 14£0. 48/212. 46+0. 6¢|179. 80+0. 41200. 940.79 118. 92+0. 33/144. 87+0. 40/132. 96-+0. 43 


pie { Stand. dev...| 10.08+0.34) 16.97+0.4¢| 10.42+0.29| 19.18+0.56 7.54£0.23} 9.590. 28) 12.07+0.31 
aie (Coef. of var..| 5.72+0.19] 7.99+0.23) 5.80+0.16 9.540.283) 6.34+0.19] 6.62+0.19| 9.08+0.23 


B. Length {Mean.........|111.02+0.44/128.64+0.38 71.33+0.20| 86.76+0.23 
in front |S. dey...) 6.21+0.31] 9.37+0.27 | 4.58+0.14) 5.60+0.16 
of mouth |Coef. of var.., 5.60+0.28) 7.28+0.21 | 6.42+0.20] 6.45+0.19 
C. Length ee ee 66.13+0.51) 83.83+0.38 47.59+0.21| 58.10+0. 22) 
behind 4Stand. dev...) 7.22+0.36| 9.34+0.2i 4.75+0.15| 5.43+0.16) 
mouth | Coef. of var..) 10.91+0.56) 11.14+0.32 9.98+0.31) 9.35+0.28) 


D. Proportion of entire 
length lying in front of | 


the mouth........ z 63.0% 60.6% 60.0% 59.9% 
E. Mean.........| 29.21+0.18} 43.08+0.19 47.96+0.27) 29.020. 16) 45.970. 16) 
Breadth 4 Stand. dev...) 2.66+0.12) 4.58+0.12 6.59+0.19 2.614£0.11) 3.89+0.11 
Coef. of var..) 9.11+0.43) 10.63+0.31 13.75+0.41) 8.98+0.39 8.47+0.25, 
F. Proportion of mean | | 
breadth to mean length. 16.6% 20.4% 23.9% | 24.4% 31.7% 
G. Coefficients of correla-) | 
tion 
1. Total length of 
individual A with! | 
same of B....... 10.359+0.0410.411+0.0340.245+0.0370.358+0.0360.210+0.041) 0.432+.033/0.231+0.034 
2. Part before mouth) | 
in A with same | 
Thy ee recdanrenoe 0.400+0.034 (0.435+0.0350.370+0.036 
3. Breadthof A with) 
sameofB......... 0.356+0.034 0.295+0.038, 0.325+0.037 


4. Partbeforemouth | 
with total length | 
(of same indi- | 
vidual) 22csrtetecra 0,9120.007 '0.820+0.0140.886+0.009 

5. Part before mouth 
with part behind 
mouth (of same 
individual)....... | 0.6490 024 

6. Total length with) 

total breadth (of | 

same individual) | 0.622+0.025) 0.648+0.024 (0.504+0.031 


| 
| 0.277+0.040/0.488+0.031 
| 
| 


CONJUGATION IN PARAMECIUM to 


proof of the fitting and change of form in the anterior region 
during the union. 

3. This same thing is still more clearly shown by the change 
in our third point. In the two members, A and B, of the united 
pairs, the correlation of the parts anterior to the mouth (that 
of A with that of B) is double the correlation between the total 
lengths of A and B. But after separation the correlation is prac- 
tically the same for the two dimensions. ‘This is shown in table 
21. 


TABLE 21 


Correlation of parts anterior to the mouth, as compared with correlation of total length, 
in conjugants (1) during union, and (2) after separation 


CORRELATION BETWEEN PARTS CORRELATION BETWEEN TOTAL 


1. CONJUGANTS DURING UNION ANTERIOR TO MOUTH LENGTHS 
Lot 7, table 18 0.840+0.016 0.333+0.048 
Lot 18, table 18 0.626+0.031 0.323+0.046 
Lot 24, table 20 0.435+0.035 0.210+0.036 


2. consUGAN'S 12 HOURS AFTER 
SEPARATION 


Lot 22, table 20 0.400+0.034 0.411+0.034 
Lot 24, table 20 0.432+0.033 0.370+0.036 


From all these facts it is clear that the parts anterior to the 
mouth are fitted to each other during union, giving high correla- 
tion. After separation the change of form that brought about the 
fitting disappears, so that the correlation for this part becomes 
no greater than that for the body as a whole. The correlation 
for the entire body however persists undiminished. 

Correlation of pairs not due to equalization. These facts of 
course demonstrate completely that the correlation in total length 
is not due to the contraction, curving, stretching, etc., that takes 
place in fitting one conjugant to the other, for this correlation 
is fully as great after the change of form due to this fitting has 
quite disappeared. There is clearly an assortative mating that 
is quite independent of this fitting process, larger individuals 
mating with larger, smaller with smaller. 


74 H. S. JENNINGS 


Another conceivable method of equalization is discussed by 
Pearl (’07). He suggests that this might occur by a passage of 
fluid from the larger to the smaller individual of the pair. He 
points out that this would be a process requiring some time, and 
that therefore, if this were the explanation of the correlation, the 
latter should be higher in pairs that have been in conjugation 
for some time, than in those in the early stages of the process. 
By sorting out those in early stages of the process (as shown by 
the nuclear conditions), and comparing them with those in later 
stages, Pearl showed that this is not the case. Thesame thing 
will be shown experimentally later in this paper; pairs taken in 
the earliest stages of conjugation give as high a coefficient of cor- 
relation as those which have been united several hours (see table 
24 and the adjoining discussion). 

Of course there is absolutely no indication from any source 
whatever that any appreciable quantity of cytoplasm passes from 
one member of the pair to the other. Thus the suggestion of 
equalization by this means is one so entirely without foundation 
as to hardly require for its disproof the evidence above given. 


3. Change of size during union 


A eause that might conceivably produce correlation between 
the members of pairs is the following: The conjugants might 
change in size in some typical way, either decreasing or increasing, 
during the period in which union continues. Thus, if they de- 
crease in size, the pairs at the beginning of the period of conju- 
gation would be large; those near the end of the period would be 
smaller, and intermediate ones would show intermediate sizes. 
If we took then a collection containing paws in various stages of 
this process, we should find a marked degree of correlation in 
s1ze. 

It is therefore important to determine whether such a typical 
change of form occurs during union. For this purpose I tried 
the following experiment. A wild culture of Paramecium cauda- 
tum was placed, on the evening of September 20, 1910, under 
conditions favorable for conjugation (many specimens in a shal- 


CONJUGATION IN PARAMECIUM 75 


low watch glass). At this time there were no conjugants in the 
lot. Now, as Maupas (’89, p. 171) has noted, under such condi- 
tions Paramecium eaudatum begins conjugation in the early 
morning. At five o’clock the next morning rare scattering pairs 
were found. Others were beginning to unite, so that as I watched 
them the number of pairs increased rapidly. Beginning at six 
a.m. I picked out about 200 pairs as quickly as possible; all these 
were isolated, with no admixture of non-conjugants, before seven 


TABLE 22 


Correlation table for lengths of conjugants of lot 23 in 
the earliest stages of conjugation. (Unit of measure- 
ment, 4 microns) 


40 41 42 43 44 45 46 47 48 49 50 51 52 


38 1 1 
39 | 2 1 1 4 
40/1 1 1 3 
41 1);2)2/1 2 8 
' 42 3/4/2/2/2)1 1 15 
43 GN 2) 2 7 V3) 4 16 
44 2|)2/5 1 1 11 
45 3) 4/213) 1 1 14 
46 1 3/3 2 9 
47 2)6)42) 1) 71 )1 2 
48 1 1 
49 1 1 


3 1 | 7 |15 |11 14 111 117 | 8) 5,2) 1 95 


a.m. At this hour I killed about one-half the lot; their measure- 
ments are given in table 22. This includes pairs in only the early 
period of union. The remainder were kept for five hours, till 
noon; then these were killed; their dimensions are given in table 
23. If there is any typical change of size during conjugation, 
comparison of these two tables will show it, since in the latter 
table the animals have been in conjugation five hours longer than 
in the former. Furthermore, if such change of size is the cause 
of the correlation, then a table comprising both setstogethershould 
give a higher coefficient of correlation as well as a higher coefficient 
of variation than either of the tables alone; and notably higher 


76 H. S. JENNINGS 


than the coefficients for the pairs in the early stages of union 
(table 22). The data for the two lots separately and for both 
together are given in table 24. 

As table 24 shows, there was no change in size during the five 
hours of union between the taking of the first and second sets. 


TABLE 23 


Correlation table for lengths of conjugants 
of lot 23 in later stages of conjugation. 
(Unit of measurement, 4 microns) 


42 43 44 45 46 47 48 49 50 51 
40 | | |i 
41/2/11] 2 
42|2/1 1 
43 3 
44 } 1/2 
5 
1 


— 
a 
ro 
me 


45 
47 | | | 
48 | 
49 oh a 


}4}2/6/4\14/9/6/4]3]1 53 


to 


i 
an 
ry 
WNOaIntwoar 


TABLE 24 


Constants of variation for the pairs of lot 23, after different periods of union 


BER MEAN STANDARD COEFFICIENT OF erases 
s a 2 , q - = 
OF DEVIATION VARIATION IN LENGTH 


Pairs in the earliest 

stage of union: 

F AM Sane scree site 95 179.79+0.53 10.79+0.37 6.00+0.210.303+0.044 
Pairs that have been 

united more than | | 

five hours: 12 m.... 53 179.81+0.64 9.72+0.45 5.41+0.250.118+0.065 
Pairs in early and late 

stages of union taken 

together (sum of the 

two foregoing)..... 148 179.80+0.4) 10.42+0.29 5.80+0.160.245+0.037 


CONJUGATION IN PARAMECIUM Ul 


The mean lengths for the two sets are perhaps nearer together 
than one would usually hope to get them for two samples of the 
same material. Further, there is no increase of variation nor of 
correlation when we unite the pairs that have been but a short 
time in conjugation with those that have been united a long time. 
In fact the coefficients for the two sets taken together are slightly 
less than for the lot in the first stages of union, though the differ- 
ences are merely those that might be expected when two samples 
of the same material are examined. It is clear therefore that the 
correlation is not due to any characteristic change in size as union 
continues. 


4. Differential contraction due to killing fluid 


Another cause which might conceivably produce correlation 
between the members is the following: A contraction or other 
change of form due to the action of the killing fluid might vary 
among the different pairs, owing to the impossibility of all coming 
in contact with the same concentration of the fluid at the same 
relative instant. If such differences occurred, they would always 
be between different pairs, and not between two members of the 
same pair, because both members of a pair keep together, and 
would therefore be subjected to identical influences. Thus 
we should have some pairs that were contracted and therefore 
short; other longer; the result, when all were taken together, 
would be to give a correlation in length between the members 
of pairs. 

Any change of form due to the killing fluid is evidently very 
slight, if it oceurs at all; nevertheless, Paramecium does possess 
a certain contractility and it seemed best to test carefully this 
possibility. For a time indeed I was inclined to attach consider- 
able importance to it. An opportunity for a test is given by the 
experiments already detailed, in which the conjugants were meas- 
ured after they had been separated at least twelve hours. After 
the pairs have separated of course the two will no longer be sub- 
jected to identical action of the killing fluid; all correlation should 
therefore disappear, if it depends on this action. As a matter 


78 H. S. JENNINGS 


of fact, as we have seen, the correlation does not disappear, and 
is not even lessened under these conditions. 

However, in such experiments, the two members of a given 
pair are usually killed together in a single minute drop—each pair 
separately. It might still be maintained that there are accidental] 
differences in the action of the killing fluids on the different 
drops, causing different degrees of contraction in the different 
pairs; this would therefore still give us correlation. 

In order to test this, in a number of lots, I killed the two in- 
dividuals of each pair separately, in separate drops. If the fac- 
tor we are discussing plays any part, it would now act strongly 
against correlation, since the two members of a pair would con- 
tract diversely. There should be no correlation in such pairs, 
if correlation is due to diversity in action of the killing fluid. 

In certain Jots I had three sets of pairs: (1) those killed while 
united; (2) the two members separated, but killed together: (3) 
the two members separated and killed separately. Comparison 
of these will enable us to set at rest absolutely the question of 
the part played by any differential contraction due to the killing 
fluid. The pertinent data are given in table 25. 

As this table shows, the correlation persists even when the 
separated members of pairs are killed separately. Thus the cor- 
relation is certainly not due to any differential action of the kill- 
ing fluid on the different pairs. 

One peculiar case in the table requires mention. In lot 24, 
we find that for the separated pairs in which the two members 
were killed in the same drop, the coefficient of correlation (0.506) 
is much greater than for the case where the two members were 
killed in separate drops (0.230). This taken by itself would seem 
to indicate that the killing of the two members together does have 
a decided influence in increasing the correlation. But further 
examination shows that there is no such indication. First, in 
this same lot the correlation for the separated pairs killed sepa- 
rately is greater than for the pairs still united. It is clear therefore 
that the positive correlation of the latter is not due to the fact 
that the two members of the pair were killed together; and this 
is precisely the question we are putting to the test. Second, in 


79 


PARAMECIUM 


CONJUGATION IN 


'ye0'0- 182 0.8z 0+80°6 
‘8400 20Z 0£E 0-+6T 6 
‘gh0'0-F 82 028 006'8 


| 
| 


1e°0-L0°ZI SFO 96 CET) SLI 
Fr'OFS0E ICI OFEH SEI] 39 [16 | AoyuIUdes ports !poywaudo: 
Zh OFFL IT 09086 'iSI| L9 (28 


-*- 1994080} pal] ‘payeredag *e 


Shes qpuevjowng ‘o} 


>| &% 


(4) eIfeny 


: | | 99 
180'0SZE 0880 0 ZEF OGL 0FZO°9 [820696 OF OFL8 FT T9+09/FET “-°(9-+9) payeandas [TV *P] 
960'0F 082 OIG OFL0'S [IE OFTF'L [FF OFL6 SFI) 19 oo. Aleyeredes pertry tpayuredag °O| | z 
ZE0'0F 90S OTE OFOL'L |SF OFOT IL FO OFES EFT) 09 69. |" "10430709 party! paresedag -q{ | "° 
TF0'0*0IZ O\6T OFFE'9 [8Z'OFFG'L [S206 BIT] GF \CoT payun ins *8) | 
| | (4) BIEN 
88° OFG6Z 0 980 0F8GE08Z OFF G6 [99 OFT GT 62 OF FG 000 Lg pet | Ajoyeredas poly tpaqyuandas “q) | ez 
LEO OF SEZ O9T OF08'S [60 0FZF OT Th O08 GLT/E2-+esFI | payan ag “ef | 
| | | UIN}EPNB! PITAL 
$20 0-998 OFEO'OFIIF OLS 0F66'L [6F'OFLGE°OT 69 OF 9F GIG 9 [SET | 1099709 PATIL *poyeredag q| | __ 
TF0' 0698 OT OFZL'S |FE°0F80 OI SF OFFI OLT) SF (COT poyrunypyssyaemntuog “ef | °° 
| | al UINgepNed! PTT | 
| | 
ee lee oie we ied E 
. | | a q | 
Q4eSe/Sesse|) BSS | wae | Bis Ve 
waebe |egeea| See | 2o8 pee 
B8a|~scea|] Ses S25 g Fs 
wo | Bed 4 z i) | 
meee | AS&e | Ree Aad x om 
aaa Z| ie all o4 ° a | 3 | 
ron] 88e8 | 44 ag fc ” 
gage) gaze/ ssc | 4 Fs I | 
He ally see nae g emis | 


pouyuasagip ybnowy} wornjasoa fo woyonpoud ajqussod D 389} 


“pungf Surry 247 fo woxon 


so WIAVL 


07 syupbnluos fo uosr.wwdwo)y 


8O H. S. JENNINGS 


the case of the separated pairs killed together, not only is the 
coefficient of correlation greater, but the variation is likewise 
much greater than in the case of those killed separately. In the 
former the coefficient of variation is 7.76, and the range of varia- 
tion was from 88 to 168 microns, while in the latter the coefficient 
of variation is but 5.07, and the range was from 128 to 164 microns. 
Now this great difference in variation cannot possibly be due to 
the fact that the two members of the pairs were killed together 
in one case, separately in the other; there is absolutely nothing 
in this procedure to change the variation. Clearly, the differ- 
ence between the two lots is purely accidental. Among those 
taken for killing the two members together, it happened in this 
particular case that all the extreme specimens fell, leaving the 
medium individuals for the other lot. Both samples are rather 
small, so that this is not particularly surprising. Now, we have 
seen on previous pages that collections containing only specimens 
of medium size give a lower degree of correlation than do col- 
lections containing extreme specimens. This is clearly the ex- 
planation of this peculiarity. Such fluctuations in the coefficient 
of correlation due to the accidents of random sampling are not 
particularly rare. In one case, random matings of the compo- 
nent dimensions of a series of 125 pairs gave me a negative coefh- 
cient of correlation of somewhat beyond —0.500, though there 
was absolutely nothing in the process of drawing the numbers from 
a hat that would tend to induce correlation of any sort. 


6. Local or temporal differentiations in the culture 


An important possible cause of the production of corre’ation 
in the members of pairs is set forth by Pearl (’07) as follows: 


It might be maintained that since at different pomts in the culture 
and at different times the environment no doubt differs slightly, there 
would be a corresponding local differentiation of the Paramecia in each 
local culture unit. Then, even though the pairing were quite at random 
in each locality, yet if the records for several such localities were mixed, 
a spurious homogamie correlation would arise (p. 256). 


CONJUGATION IN PARAMECIUM 81 


Larger pairs would come from one part of the culture, smaller 
from another, giving of course a positive coefficient of correlation. 
This result would be much accentuated if the pairs were taken 
from the cultures at different times, as was actually the case in 
the study made by Pearl. 

Pearl attempts to meet this by taking the two non- conjugant 
individuals nearest to each pair, and hence from the same environ- 
ment, pairing these, and making a table of the results. Such 
pairs give no positive correlation. 

I am not convinced that Pearl’s procedure fully meets this 
difficulty. In the non-conjugant population we have to deal 
with the great variability due to growth. Of the two single 
individuals lying nearest a pair, one might be old and large, the 
other young and small. The difference might be much greater 
than any differential effect of the environment. In the case 
of the conjugants, on the other hand, this difference in growth 
plays little or no part. That is, in the non-conjugants we have 
an important source of variation that is independent of the en- 
vironment, while in the conjugants we have not. It would appeas 
likely therefore that local environmental differentiations would 
have much more effect on the correlation of conjugants than on 
that of non-conjugants. 

In the case of the conjugants studied in the present paper this 
source of error was completely excluded for most of the lots ex- 
amined, through the method by which they were taken. On the 
evening before the day the conjugants were desired a great num- 
ber of individuals, none of whom were conjugating, were taken 
from the large culture and placed together in a small watch glass. 
They were thoroughly stirred and mixed in the process. The 
watch glass contained only water, besides the infusoria. On the 
following day the animals were conjugating in multitudes. There 
had been absolutely no opportunity for local differentiations such 
as might give origin to correlation, yet collections so made showed 
the same degree of correlation as did others taken from larger 
vessels. In the case of collections of conjugants taken from large 
cultures, I removed individuals from only one small region and 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 1 


82 H. S. JENNINGS 


all at the same time, so that the results of local differentiation 
were avoided in these cases also. 

The fact that the coefficients of correlation for my samples 
are notably less than they were for Pearl’s (see page 38) perhaps 
indicates that these local differentiations may have played some 
part in Fearl’s.results. As Pearl himself remarks ‘‘the samples 
used in this (his) work were taken in just such a way as would 
make most pronounced any spurious correlation due to local 
differentiation resulting from time or place factors. Small 
samples, a drop or two of culture fluid, were taken from different 
parts of the culture at intervals of time”’ (l.c., p. 256). But the 
results of the present paper show that the effects of this were 
limited to increasing to a certain extent the degree of correlation 
observed. They do not account for the existence of the correla- 
tion. 


6. Other suggested causes 


Pearl raises the question ‘“whether these correlations represent 
any true assortative pairing or merely arise because conjugation 
goes on within a limited, differentiated portion of the population, 
which portion, as has been shown above, is much less variable 
than the non-conjugant population”’ (07, p. 254). 

It is not apparent how this last mentioned condition could 
produce correlation, since the latter is not affected in any way by 
the characteristics of that part of the culture which is not conju- 
gating, and low variation has no tendency to cause correlation. 
However, Pearl demonstrates fully that this is not the cause of 
the correlation. He makes random pairings between the measure- 
ments of the conjugating individuals, and shows that these ex- 
hibit no correlation, as they must do if the explanation just 
mentioned is correct. 


7. Correlation in breadth 


In the present paper I have dealt mainly with correlation of 
the two conjugants in length. It is here that the most important 
relations show themselves. Furthermore, as Pear] pointed out, 


CONJUGATION IN PARAMECIUM 83 


the measurement of breadth in conjugating pairs is very inaccur- 
ate, owing to the flattening that takes place as conjugation occurs. 
Also, the ridge forming the right side of the boundary of the oral 
groove of one specimen fits into the oral groove of the other, pro- 
ducing an interlocking which makes it almost impossible to deter- 
mine correctly the breadths of the two individuals. 

On these accounts I have not thought it worth while to deal 
with the breadth in any such full way as I have dealt with the 
length. Anyone who is interested in a careful analysis of the 
breadth relations of the two conjugants will find it in the paper 
of Pearl (07). I shall take up only a few points not brought out 
by Pearl. 

a. Flattening at the time of conjugation. Pearl states that at 
the time of conjugation there is frequently a dorso-ventral flat- 
tening of the two individuals, such as would result from pressing 
the two together; he gives however no measurements on this 
point. I have endeavored to obtain some precise data on the 
matter in certain cases. 

In the small race ¢ (aurelia) (lot 9 of table 2), I measured the 
amount of flattening in fifty pairs. This was done as follows: 
The breadth was first measured as the two members of the pair 
lie side by side, in the usual position. This gives the dorso- 
ventral breadth. Then the animals were turned till one lies 
directly above the other; they were kept in this position by plac- 
ing them in a jelly made from quince seeds in water. In this 
position the lateral breadth (at right angles to the dorso-ventral 
breadth) was measured. In almost all cases the lateral breadth 
was notably greater than the dorso-ventral breadth, showing that 
the animals were distinctly flattened. The two breadths are 
shown for these one-hundred conjugants in table 26. 

I made the same measurements also for fifty non-conjugants 
of this same culture. These were likewise flattened in the same 
way, indicating that the flattening was not due to the pressing 
of the two conjugants together, but that it existed in this culture 
independently of conjugation. The dorso-ventral and lateral 
breadths are given for the non-conjugants in table 27. The con- 
jugants are flattened a little more than the non-conjugants; 


84. H. S. JENNINGS 


TABLE 26 TABLE 27 
Lateral and dorso-ventral breadths (or Lateral and dorso-ventral breadths (or 
thicknesses) of 100 conjugants of lot thicknesses) of 50 non-conjugants of 
9 (race c). (Unit, 4 microns) lot 9 (race c) 
Lateral Lateral 
7|8| 9 (10 /11 (12 113 7 | 819 (10 |11 12 113 
3 6| |3]1 4 4 1 1 
Se oa eGR aati 15 5/1 1 9 
S 8/1) 2 12 11) 3)1 30 3S 6/1 9 3 
2 9 15 20/4] | 39 ss. 7a) Oulena eam liak 8 
aS 10 Op) ely als) dit Sg 21613 ll 
1 1 1 2 9 2/5 1 8 
i | Pai S 10 4 4 8 
2 | 8 119 30 34 | 6 | 1 {100 11 poe ete ar bal (C33 
3 12 2 2 
13 i) 3 
2 50 


4/3 116|9{|6 10 


in the former the mean dorso-ventral breadth is 83.4 per cent of 
the mean lateral breadth; in the latter it is 86.5 per cent. 

Thus in certain conjugating cultures there is a notable differ- 
ence between the dorso-ventral and lateral breadths, even in 
non-conjugants. It is important however to bring out clearly 
the fact that in most cultures of Paramecium, whether conjugating 
or not, there is no such flattening. I have made measurements in 
the same way of several other cultures, and have examined many 
more, and in no other case have I found any such marked flatten- 
ing. In this culture of the race c, the animals were thin and 
showed every appearance of being in process of starvation. I 
believe that this is the explanation of their exceptional dorso- 
ventral thinness. In most cultures the animals are nearly or 
quite circular in cross-section. 

b. Correlation in breadth in pairs after separation. Pearl gives 
the coefficients of correlation in breadth for three lots of pairs 
that are still united. It seems worth while to add here the breadth 
correlations for certain lots measured after the pairs have sep- 
arated. In such eases there is no such difficulty in making accur- 
ate measurements of breadth as we find in the case of pairs still 


CONJUGATION IN PARAMECIUM : 85 


united. The correlation in breadth for three such separated lots 
is given in table 20 (G, 3). The values for the coefficients (0.356, 
0.295 and 0.325) are not far from those found by Pearl] for unsep- 
arated pairs (0.218, 0.342, and 0.349). 

ce. Correlation in breadth not due to equalization. It is worth 
pointing out that the correlation in breadth of the two members 
of a pair could not be produced by that equalization of the pairs 
through stretching, contraction, curving, ete., which we have 
discussed so fully in the foregoing pages. For as the leneth of 
the two was made more nearly equal by this process, the breadths 
would inevitably become more unequal. The shorter member, 
stretching to match the longer, would become more slender; 
the longer specimen, contracting to match the shorter, would 
become broader, thus increasing the differences already existing 
(existing as a result of the fact that the longer specimens are al- 
ready broader than the shorter specimens, a fact demonstrated 
by Pearl (’07) and the present author (08). Thus the existence 
of marked correlation in breadth between the members of pairs 
is in itself a demonstration that the correlation is not due to the 
equalizing during union; a demonstration made on other grounds 
elsewhere in this paper (pp. 65-74). 


&. Historical and comparative 


Pearl (07) was apparently the first to notice the assortative 
mating in any infusoria, and his study of the matter was more 
thorough than any other made previously to the present paper. 
We have dealt so fully with his work in the body of this paper that 
we need not dwell further on it here. 

Collin (09) observed that in the conjugation of Anoplophrya 
branchiarum, a parasitic ciliate living in the blood of Gammarus, 
the two members of a pair areusually of nearly equal size, although 
the different pairs differ much in size. He thinks that there is 
real assortative mating, larger individuals mating with larger, 
smaller with smaller, as in Paramecium. However, a part of 
the greater similarity of the two members of a pair is in Anoplo- 
phrya due to the fact that as conjugation progresses the individ- 


S6 H. S. JENNINGS 


uals of the culture become smaller. Hence the first individuals 
that conjugate form large pairs; later ones form smaller pairs, 
still later ones still smaller pairs,—(although the size of a given 
pair does not change during conjugation). This does not, Collin 
thinks, account fully for the similarity in size of the two members 
of a pair, for even among the individuals that conjugate at any 
given period one finds much greater resemblance between the 
two members of a given pair, than between members of diverse 
pairs. Collin does not give measurements. 

Enriques (08) studied conjugation in Chilodon uncinatus, with 
relation to the problem of assortative mating. He found that a 
change in form takes place during conjugation, by which the 
left member of the pair becomes shorter than the right; further, 
the evidence indicated that the left-hand member was even before 
conjugation somewhat smaller than the right-hand one. This 
latter relation would result from the fact that in the process of 
conjugation, owing to certain peculiarities of form in Chilodon, 
‘the larger individual always becomes the right-hand member, 
the smaller one the left-hand member. 

Assortative mating was studied with relation to the question 
whether the two members of a pair tend to be of the same size or 
not. That is, if the member A diverges from the mean of all, 
does the other individual B Jikewise diverge in the same direction 
from the same mean? Enriques studied this question by deter- 
mining the mean difference between the two individuals of the 
pairs, and comparing this with the mean difference in length when 
matings are made at random. In ease of assortative mating in 
the sense above defined, the former value should be less than the 
latter. Enriques found that in most of his (rather small) samples 
it was not less; so that there was no indication of assortative 
mating in the sense defined. But in samples taken toward the 
end of an epidemic of conjugation, the members of pairs were 
more alike than were individuals taken at random, so that a 
degree of correlation is indicated. For two of his lots Enriques 
worked out the coefficient of correlation; for the early sample 
the coefficient was zero; for the later one it was 0.4. Thus in 
the later periods of an epidemic there is an actual correlation in 


CONJUGATION IN PARAMECIUM 87 


size between members of pairs. Enriques explains this difference 
between the early and the late stages of the epidemic in the fol- 
lowing way: In the late stages of the epidemic many of the pairs 
are formed (wholly or partly) of individuals that have already 
conjugated once in the same epidemic. This fact is determined 
by studying the nuclear conditions of the pairs; some individuals 
are found to be undergoing the changes consequent on a previous 
conjugation. Now, Enriques found in Chilodon (as we have 
found in Paramecium) that the ex-conjugants increase in size. 
Hence those that are conjugating for the second time in any epi- 
demic are larger than those that are conjugating for the first 
time. Further, Enriques found (by studying the nuclear con- 
ditions) that for some reason these ex-conjugants are likely to 
conjugate together. Thus we get certain paws, consisting of 
ex-conjugants, in which both members are large; other pairs, 
consisting of individuals that have not before conjugated, in 
which both members are small. This of course results in the pro- 
duction of some considerable degree of positive correlation when 
the entire lot is examined. 

Evidently, the question of main interest is: Why do the large 
ex-conjugants tend to conjugate together on the one hand; the 
small individuals on the other? There is clearly assortative mat- 
ing of some kind. If (as appears probable) it takes place on the 
basis of relative size, as in Paramecium, then the interpretation 
of the facts in Chilodon would be the following: In the early 
stages of the epidemic the differences in size among the individ- 
uals are not great enough to affect the mating, which therefore 
takes place at random. But in the later period of the epidemic, 
owing to the appearance of the large ex-conjugants, the differences 
in size become so great as to prevent the union of the largest and 
smallest individuals. Hence assortative mating occurs, giving 
rise to a certain degree of correlation between members of pairs. 

Enriques raises the question whether re-conjugation among 
larger ex-conjugants may not be the cause of the correlation in 
Paramecium. It is quite clear that this is not the case. (1) 
Pearl (’07) shows that the correlation exists when one includes 
only individuals showing early stages in the nuclear processcs 


88 H. S. JENNINGS 


attendant on conjugation. (2) We have shown above (p. 75) 
that when pas are taken in the very first stages of anepidemic, 
before it has been in progress more than two or three hours, there 
is still correlation in size in the members of pairs. It may be 
noted that there appears to be no evidence that in Paramecium 
re-conjugation ever takes place among ex-conjugants. 

Attention should be called to the fact that the method used 
by Enriques does not furnish a test for another possib'e kind of 
assortative mating. In Chilodon, the right and left members 
of the paw are morphologically differentiated and the right one 
is usually larger. Now, it is possible that assortative mating 
may so oceur that when the right member is larger than usual, 
the left member must also be larger than usual, though it need 
not be so large as the right one. That is, when the right hand 
member is above the average size for right-hand members, the 
left member may be above the average size for left-hand members. 
This might still be true, even though the average size for right- 
hand members were considerably greater than that for left-hand 
members. Ii this were the case, then the average difference 
between the two members of a pair would be greater than the 
average difference between members of random matings, and yet 
this would not be evidence against the kind of assortative mating 
we have mentioned. This kind of assortative mating could be 
tested by determining the coefficient of correlation between all] 
the right-hand members (entered as X) on the one hand, and all 
the left-hand members (as Y) on the other, each set being referred 
to its own mean (as in the usual computation of correlation), 
instead of to the mean of all. A parallel case for this is set forth 
for Paramecium on p. 58 of the present paper. Unfortu- 
nately, Enriques has not given us the correlation tables for the 
actual measurements of his pairs, so that the existence of this 
kind of correlation can not now be tested for Chilodon. 


ies) 
ve) 


CONJUGATION IN PARAMECIUM 


9. Conclusions on assortative mating 


The results of our examination of the causes of the correlation 
between the members of pairs (an examination that can perhaps 
fairly be designated exhaustive) 1s to confirm Pearl’s conclusion 
that there is actual assortative mating in Paramecium—larger 
individuals mating with larger, smaller with smaller. This shows 
its effect in three main categories: 

1. When the two species caudatum and aurelia are conjugating 
at the same time in the same culture, they do not inter-cross; 
caudatum conjugates only with caudatum, aurelia with aurelia. 
Since the two species differ in size, this gives very high coefficients 
of correlation (0.940, in length). The only reason that the co- 
efficient is not actually 1.000 is the fact of incomplete correlation 
within each of the two component species. 

2. When races of different size are present, as is usually the 
case in ‘wild’ cultures even of a single species, the members of 
the larger races tend to conjugate together, on the one hand; of 
the smaller races on the other. The correlation (in length) thus 
arising is less than when two species are present; it averages about 
0.380, in the wild cultures of the present paper. 

3. When members of but a single race are present (all descended 
from a single individual), there is still a notable correlation, larger 
individuals mating with larger; smaller individuals with smaller. 
But this process of assortment is considerably less accurate than 
when diverse races are present; for pure races the coefficient of 
correlation averages but about 0.250, as against 0.380 for mixt- 
ures of races, and 0.940 for mixtures of species. 


II. CONSEQUENCES OF THE DIFFERENTIATION OF THE CONJU- 
GANTS AND OF THEIR ASSORTATIVE MATING 


“ What effects in the further history of the organisms result 
from the occurrence of conjugation? What are the consequences 
in inheritance, of the decreased size and variability of the conju- 
gants, as compared with the non-conjugants? What con- 
sequences follow from the assortative mating of the conjugants? 


90 H. S. JENNINGS 


CONSEQUENCES OF THE DECREASED SIZE AND LESSENED 
VARIABILITY 


1. Are the extreme specimens excluded from the new generation? 


Does the decreased size and lessened variability of the conju- 
gants indicate that extreme individuals of the race are excluded 
from taking part in the new generation, conjugation tending thus 
to maintain the average racial type unchanged? 

As we have seen, the non-conjugant population includes many 
individuals that are larger, and some that are smaller than any 
of the conjugants. Are these smaller and larger individuals 
excluded from participation in the further development of the 
race? 

a. As to the smaller individuals, examination of the compar- 
ative tables (tables 34 and 35 of the Appendix) will show that 
the non-conjugant population usually contains but few individ- 
uals smaller than the conjugants, its main differentiation lying 
in the opposite direction. Now, we have already shown that these 
small non-conjugants are merely young specimens, that have not 
yet reached the size for conjugation. A few hours will of course 
remedy this. The small individuals are then not excluded from 
conjugation. 

b. The main difference between conjugants and the general 
population is that the latter contains many specimens much 
larger than the conjugants. Are these larger individuals excluded 
from conjugation, and so from the further progress of the race? 
Cr are these larger individuals merely temporarily differentiated, 
their progeny becoming conjugants later? 

This matter was tested in a number of cases by removing from 
a conjugating culture a considerable number of the large non- 
conjugants, each much larger than any of the conjugants, and 
keeping them under conditions favorable for conjugation. 

Thus, January 30, 1908, I removed from the conjugating cul- 
ture of lot 6 (table 34) one hundred of the large non-conjugants. 
On the next day fifty of these were killed and measured; they 
had a mean length of 211.52 = 1.99 microns; a mean breadth of 


CONJUGATION IN PARAMECIUM 91 


62.48 +1.17. They had multiplied a little during the night, so 
that this is really less than the original size. The conjugants 
at the same time had a mean length of but 181.49 += 0.54; mean 
breadth 48.11 + 0.87. The remaining large non-conjugants were 
allowed to multiply, and on February 24 their progeny were found 
to be conjugating. The conjugants which they gave were of 
sensibly the same size as those of the culture as a whole. The 
mean length, from 21 pairs, was 179.91 = 0.88 microns; mean 
breadth 48.19 + 0.51 (as compared with 181.49 by 48.11 for the 
culture as a whole). 

Thus in this case the large non-conjugants were by no means 
excluded from conjugation and the farther development of the 
race. They represented merely a temporary stage, not as yet 
prepared for conjugation, but ready to enter upon it after a few 
fissions. 

Similarly, from a wild culture NV there were isolated on March 
21, 1908, one hundred of the largest non-conjugants, and a con- 
siderable number of pairs of conjugants. Half of each were killed 
at once; the fifty large non-conjugants showed mean dimensions 
of 224.16 by 53.36 microns, while the sixty-six conjugants meas- 
urec. 139.94 by 38.55 in mean length and breadth. On the fol- 
lowing day numerous conjugants were found among the progeny 
of the remaining fifty non-conjugants. The large individuals 
were thus but temporarily excluded from conjugation. 

A similar experiment was tried with the small pure race c 
(aurelia). On September 27, 1907, I isolated from a conjugating 
culture ten of the larger non-conjugants (larger than any conju- 
gants); also five pairs of conjugants. These were cultivated side 
by side, under the same conditions. On September 30 there 
were conjugants among the progeny of the ten large non-conju- 
gants. (On the following day there were likewise conjugants 
among the progeny of the pairs.) 

Again, from the pure race Nf, I isolated, March 31, ten of the 
largest non-conjugants; on the following day there were conju- 
gants among these. 

Thus, it is clear that the large non-conjugants may later divide, 
take on the size characteristic of conjugants, and themselves 
conjugate. 


92 H. S. JENNINGS 


2. Relative size of progeny of conjugants and non-conjugants; 
changes in size due to conjugation 


Do the progeny of the large non-conjugants differ in size and 
variability from the progeny of the (much smaller) conjugants? 

The answer to this question is bound up with that to another, 
so that the two will be considerea together. This other question 
Is: 

Are there characteristic changes in size resulting from con- 
jugation, so that arace Just before conjugation is smaller or larger, 
or otherwise characteristically different from the same race just 
after conjngation? 

The answer to this question will tell us whether there are char- 
acteristic changes in form and size in the different periods of the 
reproductive cycle (from conjugation to conjugation), for just 
before conjugation the animals are at the end of the cycle; just 
after they are at the beginning of it. 

To answer these questions experiments were carried out on a 
number of different cultures. The plan of experimentation 
was in most cases as follows: From a conjugating culture a 
sample of non-conjugants was removed; also a sample of conju- 
gating pairs. Part of each of these was killed and measured, 
thus giving the initial size for each. Then each set was allowed 
to multiply farther, under rigidly identical conditions for the 
two sets. Samples of the progeny of each were killed at intervals, 
and compared as to size. The method of work was variedin 
different cases, in ways that will be set forth. 

This plan of work, simple as it sounds, in reality presents great 
difficulties in its execution. Each individual must be kept sepa- 
rate, in order that we may know that there has been no conju- 
gation save in the case of the original pairs; and also in order that 
we may know to what generation of the progeny a given individ- 
ual belongs. Owing to these and other difficulties, it is not 
possible to obtain under these absolutely controlled conditions 
large numbers for comparison of the progeny of conjugants and 
non-conjugants, though the numbers given in the following are 
sufficient for answering the main questions. 


6 


CONJUGATION IN PARAMECIUM: 93 


It will be well to examine first the results in the case of pure 
lines or races, then to take up ‘wild’ cultures. 

a. Comparison of progeny of conjugants and non-conjugants in 
pure races. The first comparison was made between the progeny 
of conjugants and non-conjugants all descended from the single 
individual Nfs. In the progeny of this individual conjugations 
occurred March 31 and April 4, 1908. In each case a random 
sample of the conjugants and non-conjugants was killed and 
measured; the relative sizes and variabilities are given in the first 
four lines of table 28. The non-conjugants were as usual con- 
siderably longer, and more variable than the conjugants. On 
March 31 there were picked out five pairs of conjugants and ten 
of the largest non-conjugants. In this case the individuals were 
not kept isolated, but the ten conjugants were kept together in 
one watch-glass, the ten large non-conjugants in another, under 
identical conditions. Ten days later (April 10) samples of each 
of these were measured, giving the results shown in the fifth and 
sixth rows of table 28. The progeny of the pairs were now a 
little larger than the progeny of the non-conjugants. 

The progeny of the very smallest pair observed on March 31 
was kept separate from the remainder. On April 20 its progeny 
were compared with those from the ten large non-conjugants 
taken on the same day. The results are given in the last two rows 
of table 28. The size of the two sets was now very nearly the 
same (those derived from the small pair were a trifle longer and 
thinner). The variability of the two sets was now practically 
identical. 

All together this experiment shows the following facts: 

1. The progeny both of the small conjugants and of the large 
non-conjugants increased somewhat in size beyond the original 
mean for the non-conjugants. This was probably due to the 
nutritive conditions. 

2. The progeny of conjugants became fully equal in size both 
to the original non-conjugants and to the progeny of these non- 
conjugants. They were in fact a little larger than the progeny 
of the non-conjugants, and the difference appears to be significant 
in comparison with the probable error (at least on April 10). 


JENNINGS 


Ss. 


He 


94 


{ | | j 
Ze OFC8 EF) 98-98 GE°OFOL'L [So°0FS0 ZI SL: OFSF 9ST\Z6I-ZET | OT | [dy 
| | | jo syursnfuod-u0ou 
| | | | JIB] OT Jo Auaso1g go, ‘0% [dy 8 
| | 
GE OFFO' GH 9S-9E [$6 0F89'L [LE OF9E ZI [0S OFG8 O9T|SSI-OZT | 80T | # [dy jo ayed 
| | | Tews | Jo Auatorg 80, ‘0g WAdy 2 
| | | } 
82°000'FF | ¥9-8z |00°0F L001 [28°006 FT OTT 46 LETIGLI-FOT | 99 | Te 2eIy | 
| | | jo  sjyuvsnfuoo-uou | 
| | jesivl OT Jo Auasorg 90, ‘OT ey 9 
£9008 GF | 0298 |12°0FS9°CI 60 TOs SFI|ZLI-OcE | 19 | Te “aeN 30 sured | | 
| | | | yews ¢ jo Auas014 80, ‘OT ney | c 
£9 OF 01 GE FF-8Z |LO TFS CI 9S TST SI [0 GFIO LFISSI-O8I | 18 ““syuBsnfu0d-u0N|g0, ‘F [4dy) & 
0€ 080 TE | OF-8Z FE OFIN'S CFOF99D 9 OFS ZETSFI-GII Og ““squvsn{u0D|g0, ‘F Tudy ¢ 
SOF 00 FE | FSS GP OFFS GNF L [6 0F6S FHI SII-CEL | FE | “syUBNfuoD-aoyl\r9, ‘Te “~AVIN| & 
09°0*2S'SE | FE-FZ 6 OFL0F [TF OFESE 9C 06 OETSFI-FEI | ZF squvsntuoD gd, ‘Te “Ve 1 
| i} 
z | = Q a | z a) 2 | | © 
| 2 i} Pa | sel ed | > 
eo eee ee BE B |SER2| Ee 2 
= BOG Reta! fio 4 g459| 25 
c 154 zs 2 2 agz | #8 
= egieng ob ee || g | Hg} 
a HZ OZ | x “ig il 10, 
> < 24 0 fe Rit i) 
o > ies] = Z >| nD 
te} mt iS} < ) bh 
a fn - 7 Boe 5 
| | | 7 a 


(prjainv) *f nN Jonprarpur ajburs ay? wos fjjoULb1L0 paarsap 
ny ‘fuabosd way) fo pun syunbnCuos-uou puv spunbnluos fo fiqyiqnisa pun azis aaynjay 


86 ATAVL 


CONJUGATION IN PARAMECIUM 95 


3. The variability of the progeny of the conjugants increased 
and became sensibly the same as that of the progeny of the non- 
conjugants. 


From September 16 to September 26 a similar experiment was 
carried out with the race of Paramecium aurelia that I have called 
C,. The measurements for conjugants and non-conjugants of 
September 16 are shown in the first two rows of table 29; the non- 
conjugants are considerably larger. At this time four pairs of 
conjugants were placed in each of twelve watch-glasses; eight of 
the large non-conjugants in each of twelve others; the two sets were 
kept under identical conditions and treated in the same way. 
The table gives the sizes of the progeny at certain later dates. 

As the table shows, the progeny of the conjugants were larger 
than those of the non-conjugants on September 18; less so on 
September 23, and smaller on September 26. Possibly the only 
conclusion clearly justified is that the progeny of the conjugants 
are not on the whole smaller than the progeny of the (larger) 
non-conjugants. 

From September 25 to 29, 1908, a similar experiment was under- 
taken with the aurelia race g, which is closely similar to C,. In 
this case the second and third generations of the progeny were 


TABLE 29 


Relative size of conjugants and non-conjugants and of their progeny from the pure 
line Cy (aurelia) 


C 
Sa ies a " 
a a a 
a |s # < 5 
& > 5 > S| 
o | 84m z fs 2 
x one a = = 
82) eaoo Ms i) : 
2 m2) ou Zz on a Zz 
& 2a | 3498 < Zoek < 
< pa} sa a <<A a 
a Z h 2 m = 
1 Sept. 16, 0&8 Conjugants 138 88-148121.91+0. 66 
2 Sept. 16, '08 Non-conjugants... ... 110 | 104-164)132.18+0.8 34.91+0.34 
3 Sept. 18, '08)/Progeny of conjugants 15 108-140128 00+1.6 31.20+1.19 
4 Sept. 18, '08 Progeny of non-conjugants F 70 96-136/116. 170.7% 31.20+0.45 
5 Sept. 23, '08 Progeny of conjugants 56 | 120-156132.93+0.84 28-48 | 35.84+0.42 
6 Sept. 23, '08 Progeny of non-conjugants , 37 | 108-152)129.62+1.13 24-48 | 35.52+0.67 
7 Sept. 26, '08|/Progeny of conjugants...... 11 92-136/112.36+1.18 20-40 | 29.50+1.34 
8 Sept. 26,08 Progeny of non-conjugants........ 52 100-148 122.15+1.11 28-44 | 34.81+0.61 


96 H. S. JENNINGS 


compared for the two sets. The results are shown in table30. 
In this case the progeny of the conjugants were larger in the 
second generation; smaller in the third. The numbers obtained 
were small in this experiment, and under such conditions the size 
depends much on how recently fission has taken place. The con- 
jugants divided more slowly than the non-conjugants, and the six 
individuals of the third generation were all apparently young. 
The results of this experiment then merely show that there is no 
very marked preponderance of size in the progeny of either set. 


TABLE 30 


Relative size of conjugants and non-conjugants and of their vrogeny in the second and 
third generations, in the race g (aurelia) 


a LENGTH BREADTH 
4 
oe - — - 
DATE 2 5 - ‘3 
Q , 
Z a = faa} = 
1 Sept. 27, ’08) Conjugants (table 34, lot 19).. Z 174 96-152)118.28+0.52 
2 Sept. 27,’08 Non-conjugants (table 34, lot 19) .| 118 88-180)135.35=1.02 
3 Sept. 29,’08 Generation 2f from conjugants 15 140-180 154.93+2.11 40-60 45.33+0.97 
4 Sept. 29,’08 Generation 2f from non-conjugants 9 132-168)148.89=2.52) 40-56 50.22+1.05 
5 Sept. 30,'08 Generation 3f from conjugants.... 6 128-148|/135.33+2.24 32-40 36.00+0. 64 


6 Sept. 29, 08 Generation 3f from non-conjugants 10 144-168 156. 401.64 40-56 49.60+0.95 


A more extensive and precise experiment was carried out with 
the race & (aurelia) from October 19 to 30, 1908. In this experi- 
ment the progeny of each of the original individuals was kept 
isolated. The lengths of conjugants and non-conjugants of this 
race on September 12 are shown in the first two rows of table 31, 
while in the remainder of the table the first seven generations of the 
progeny are compared, generation by generation, for the two sets. 

In this case the progeny of the conjugants were larger in the 
first and second generations. Thereafter the results varied, 
neither set having a uniform preponderance. On October 30, 
the last day of the experiment, the progeny of the conjugants 
averaged a little larger than did those of the non-conjugants. 

Thus in these experiments with pure races, we find that the 
progeny of the (small) conjugants are not smaller than those of 


CONJUGATION IN PARAMECIUM 97 


TABLE 31 


Relative size of conjugants and non-conjugants and of their progeny for seven genera- 
tions, in the race k (aurelia) 


LENGTH BREADTH 
| Do 
| i 2 
oe pare 2 
DATE rs Z | n es 

aa =) q op r=] 

A a 3s a 3 

pa a o 3 o 

Zz oa] = ia = 
1 Sept. J2 08) Conjugants (table 34, lot 13) 336 92-156 129.58+0_ 40 
2 Sept. 12, '08) Non-conjugants (table 34, lot 13).) 100 | 88-168140.20+0.97 
3 Oct. 21 08) Generation lf—from conjugants.... 37 | 132-172152.00+0.87, 36-60 | 45.84+0.75 
4 Oct. 21, '08) Generation If from non-conjugants) 17 | 136-156144.94+0.91, 36-52  44.47+0.71 
5 Oct. 22-25'08) Generation 2f from conjugants... -| 32 | 120-164 146. 50+1.18 36-56 | 44.00+0.60 
6 Oct. 22, '08| Generation 2f from non-conjugants) 31 | 120-156138.19+1.1£ | 38.58+0.70 
7 \Oct. 25, 08) Generation 3f from conjugants. .| 27 108-160)130.52+1.48 5. 44+0 80 
8 \Oct. 24, '08) Generation 3f from non-conjugants) 17 | 124-160143.76+1.60 88+0.81 
9 Oct. 25, '08) Generation 4f from conjugants.../ 5 | 104-144 120. 00+5.00 00+1.53 
10 Oct. 24, 08) Generation $f from non-conjugants) 27 | 116-156133.19+1.38 32-56 | 42. 96+0.91 
11 Oct. 26-28,) Generation 5f from conjugants {| 32 | 108-156 180. 2 35.38+0.52 
12 Oct. 28, ’08) Generation 6f from conjugants ‘| 39 | 112-156)131. 5.490. 52 
13 (Oct. 28, '08| Generation 6ffrom non-ecnjugants| 10 | 128-148 136 601.09 
14 Oct. 28, ’08) Generation 7f from conjugants. ...| 95 96-156 129 5.75+0.41 


15 Oct. 28, 08) Generation 7f from non-conjugants 140.00 | | 40.00 

16 Oct. 30, 08) All existing progeny of conjugants) 25 100-152/128 16+1.59 24-52 | 36.320 96 
17 Oct. 30, '08) All existing progeny of non-conju-, | 

PPATLES aiatefelsteterelsisisjerefelefarstotalo/olaletatte rc! 28 104-144123.71+1.26 28-44 | 34. 14+0 60 


the (larger) non-conjugants; that in fact the first measurements 
taken after fission show in every case the progeny of the conju- 
gants to be a little larger; and that the animals of the culture 
as a whole are a little larger after conjugation than before. With 
these should be compared the following results of an experiment 
on a wild culture, in which the conjugants and non-conjugants 
compared were fundamentally the same. 

b. Comparison of the progeny of conjugants and non-conjugants 
in a wild culture. On June 20, 1909, a wild culture that appar- 
ently consisted entirely of caudatum was found to be in conjuga- 
tion. Samples of the conjugants and of the non-conjugants were 
measured, with the results given in the first two rows of table 32. 
At the very beginning of the conjugation, in the early morning, 
a large number of pairs were picked out. Many of these had 
as yet become united only by the anterior parts of the body. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 1 


98 H. S. JENNINGS 


These were separated by manipulation with a pipette, and the 
two individuals isolated. Other pairs were allowed to com- 
plete conjugation, the individuals being isolated only after sep- 
aration in the natural course of events. 

Thus we have in this experiment two sets for comparison, both 
of which were ready for conjugation; both indeed had taken the 
first steps in actual union. Both sets of course came from the 
same culture, and had thus been living under identical conditions. 
Both sets were kept throughout the experiment under identical 
conditions, the progeny of each individual in a separate concavity 
of a hollow ground slide, the culture fluid changed every day. The 
only difference between them is that one set was allowed to com- 
plete conjugation, while the other was not. We have therefore 
the conditions perfectly fulfilled for determining Just what dif- 
ference conjugation makes. 

The pairs which had been separated before conjugation was 
complete may be designated ‘unpairs’; of these there were sixty- 
four of the original specimens. Of the ‘pairs’ (conjugation com- 
pleted) there were forty-eight individuals. Thus 112 lines of 
propagation were followed. I compared generation by genera- 
tion, so far as possible, the first seven generations produced by 
fission. Furthermore, the individuals present in each set on cer- 
tain days (June 22, June 23) were compared as to size, without 
regard to the generation to which they belonged. It may be 
noted that at this time the weather was extremely hot, so that 
fission took place in both sets with very great rapidity. The 
results of the experiment, so far as size is concerned, are summar- 
ized in table 32. 

Examination of table 32 shows that in all of the nine sets com- 
pared, except one, the progeny of the individuals that had been 
allowed to conjugate are larger than the progeny of those that 
had not been allowed to conjugate. In all these cases the differ- 
ence is very marked, even in comparison with the somewhat 
large probable errors (due to the relatively small numbers). 
In the single case where the progeny of those that had conju- 
gated are smaller (second generation), the difference is small in 
proportion to the probable error, and therefore without much 


CONJUGATION IN PARAMECIUM 99 


TABLE 32 


Relative size of the progeny of pairs that had completed conjugation (‘pairs’), and of 
those that were separated at the beginning of mating (‘unpairs’). The firsttwo lines 
give the dimensions for the conjugants and non-conjugants of the culture from which 
all came. The measurements given are the lengths, in microns. 


| | | | DIFFERENCE IN FAVOR | 5 

q | I eter oF “PAIRS” E 

e| 3 eas eal Taree dé 

las a) e/g lee g 23 gs 

a re) g | 4 | G la 2 % 2 aac | £& 

a g p/2Z/siee] a E mer Ae 

o a Z)a| 3 | = & pens | a) 
June 20 Conjugants....... 284 130 206 76165.04+0.53 

Non-conjugants.. 87 134 198 64164.12+1.00 

fi June 22 Palrsee cena ...., 380) 156) 240 84201.20+2.48 33.35+3.14 20.17 9.990. 88 
June 21 Unpatrs.. .. 27) 148) 196) 48167.85+=1.93 §.85+0.82 

fz | June 22-23) Pairs....... DoS 30, 128 224, 96155.87+3.14 —5. 5743.31 —3.45 16.37+1.46 
June 22 Unpairs............, 64; 120) 184 64161.44+1.06 7.75+0.47 

faaltdunes22—26| SPaAlrs. neree<B <n 2,20 37| 140) 228) 88189.73+2.62) 25.73+4.01 15.69 12.47+0.99 
June 22 Wanpairs:= 3... .| 22) 128) 200 72)164.00+3_03 12.85+1.33 

fs | Sune 22-23) Pairs..............| 61) 124} 216) 92)178.89-1.79) 26.4042. 15 17231 11.56+0.72 
June 22 | Unpairs........... 74 108) 188 80152 49-+1.20 | 10.05+0.56 

fs | June 22-23) Pairs.............- 58) 128) 212) 84174.83+1.60) 20.92+2.39/ 13.60 | 10.35+0.66 
June 22 Unpairs 23) 132) 180) 48153.91+1.78 | 8.22+0.82 

fs June 23 Para eraetetatsresrs 95| 136, 212) 76)175.62+0.94) 12.12+2. 43) 7.41 | 7.71+0.38 
June 23 Unpairss...).- | 24 125 196 68163.50+2.29 | 9.97+0.98 

fz | June 23 let et oorigopdGo 36) 164 228 64/187.44+1.68) 24.34+1.89 14.92 7.98+0. 64 
June 23 Wmpaitsssn.- 1-1: 62 144 196 52163.10+0. 86 | 6.13+0.37 
June 22 Allipairs: seemed 72) 124) 240) 116172.17+2.45 17.28+2.66| 11.16 17.86+1.04 
June 22 Allunpairs........) 119) 108) 200) 92154.89-+1.03} 10.77+0.48 
June 23 Ali pairs: ey cryscie 275 128 228 100181.02+0.73) 17.81+1.14 10.91 9.92+0.29 
June 23 All unpairs....... 86) 128 200, 72163.21+0.88 7.41+0.38 


significance. Owing to the small numbers measured in some of 
the cases, differences in the absolute age of the individuals must 
cause considerable fluctuations in the mean measurements, so 
that on the whole it is rather surprising that the results should 
be as uniform and unambiguous as they are. A considerable 
proportion of the thirty specimens in the second generation from 
the pairs were evidently young animals. 

Not only is the mean size practically uniformly greater in the 
progeny of those that were allowed to conjugate, but in every 
case without any exception the extreme size reached is greater 
in the descendants of the conjugants. 

Furthermore, the range of variation is in every case greater 
in the descendants of those allowed to conjugate. If we measure 
the variation by the coefficient of variability, we find that in all 


100 H. S. JENNINGS 


cases except two the descendants of the conjugants are the more 
variable. In these two cases (generations 3 and 6) the differ- 
ence between the coefficients of variability for conjugant and non- 
conjugant progeny is small, while in a number of cases the excess 
for the conjugant descendants is very great. The greater varia- 
bility of the descendants of the conjugants is on the whole marked 
and unmistakable. 

I may add that a considerable number of other experiments sim- 
ilar to that summarized in table 32 were carried out for other 
purposes, and in all cases the greater size and greater variability 
of the descendants of those that had conjugated was very notice- 
able, although measurements were not made. 

Historical. There has been considerable divergence of state- 
ment regarding the relative sizes of infusoria at different periods 
of the life cycle, so that precise measurements were much needed. 
It is of course generally agreed that the conjugating individuals 
are smaller than the average, but this is a point somewhat diverse 
from the one we are now considering. Woodruff (’07) emphasizes 
the fact that there are in Oxytricha variations in size in different 
periods of the cycle, but does not state at what periods the size 
is Jarge; at what period small. Calkins and Cull (’07, p. 380) 
state plumply that the smallest forms in the life history are ‘‘ char- 
acteristic of the first two generations after separation and before 
the young individuals have had an opportunity to take food.” 
This is quite in opposition to our measurements and if the im- 
plication is that the supposed small size is due to the fact that 
no food is taken between the time of the separation of the conju- 
gants and the first two fissions, this also is a mistake; in individ- 
uals of race k food was taken within two hours after separation, 
as shown by the ingestion of India ink. (Maupas, (’89, p. 199,) 
notes that the ex-conjugants take food ‘some hours’ after separa- 
tion.) Richard Hertwig (89, p. 189) remarked, with his usual 
correctness, though without giving measurements, that after 
conjugation the animals grow rapidly, so that they may reach a 
size not otherwise seen. 

The fact that the size is greater immediately after conjugation 
appears opposed to the common experience of students of the life 


CONJUGATION IN PARAMECIUM 101 


cycle; the remark is often made that the animals are smaller after 
an epidemic of conjugation has occurred. This, as a matter of 
fact, under the conditions usually prevailing, is true. If we 
measure a sample of a culture just before (or during) conjugation, 
and another sample after, we shall usually find that the size has 
decreased. Thus, in a culture of the aurelia race c the mean 
size (of the non-conjugants) during an epidemic of conjugation 
was 158.80 = 0.88 by 32.78 + 0.17 microns. Five days after 
the epidemic, the mean size was 129.64 = 0.87 by 35.44 + 0.40. 
But this great decrease in length is due to the cultural conditions, 
not to the fact that conjugation has occurred. I have shown in 
a former paper (710, p. 29) that conjugation commonly occurs 
under conditions in which the animals are decreasing in size; 
conditions of decrease of nutrition, as has often been noted. 
These conditions produce their effect on the size irrespective of the 
occurrence of conjugation. But if the nutritive conditions are 
kept good, the descendants of the conjugants actually increase 
in size, as compared with the descendants of the non-conjugants. 
The only way that we can determine the effect of conjugation 
itself on size is of course to compare two sets which differ in no 
other way, save in the fact that one has been allowed to conju- 
gate, the other not. In such experiments, as we have seen, the 
progeny of the conjugants are, for a time at least, the larger. 

c. Exceptional case. We are now prepared to understand an 
apparent (though not a real) exception to the results set forth. 
We have seen that in the case of a pure race, the large non-conju- 
gants do not give larger progeny than the small conjugants. We 
have however previously seen that in mixed cultures sometimes 
one race conjugates while another does not. Numerous ex- 
amples of this are given in my paper of 1910, p. 283. Now, if it 
so happens that a small race conjugates, while a large one does 
not,—then of course we shall find that the progeny of the small 
conjugants are smaller than the progeny of the large non-conju- 
gants,—contrary to what we find ina pure race. ‘The difference 
is however due, not to the conjugation, but to the difference 
in race. This case is realized in certain experiments on the wild 


102 H. S. JENNINGS 


culture N, carried out March 21 to April 6, 1908; these are sum- 
marized in table 33. 

In this case the mean length of the conjugating pairs was 139.29. 
A hundred of the largest non-conjugants were selected, all larger 
than any of the conjugants; their mean length (from a sample of 
fifty) was 224.16. The remainder of these large non-conjugants 
were allowed to multiply, side by side and under the same condi- 
tions, with an equal number of the conjugants, from March 21 
to March 26. On the latter date part of each were killed and 


TABLE 33 


Relative sizes of the conjugants and large non-conjugants, and of their progeny, in 
the wild culture N (lot 4, table 1) 


LENGTH BREADTH 
DATE 5) 5 
1908 2 & = 2 & e 
2 a 3 3 a S 
Z fac = Z fo = 
i Mar. 21 | Conjugants, sample S4 116-160139.29+0.70 86 24-48 | 37.91+0.37 
2 Mar. 21 Non-conjugants, sample . 152 | 100-244155.40+1.39 148 25-64 | 43.11+0.38 
3 Mar. 21 Largest non-conjugants 50 | 200-252224.16+1.06 50 32-72 | 55.36+0.62 
4 Mar. 26 Progeny ofconjugants 100 =: 104-200 143, 80=1.29 100 32-56 39.28+0.40 
5 Mar. 26 | Progeny of largest non-conju- 
gants..... -...| 100 | 148-240 198.52=1.40 100 36-72 | 48.68+0.52 
6 April 6 Progeny ofconjugants.... 100 100-172 137.16+0.84) 100 24-56 | 34.24+0.38 
7 April 6 | Progeny of largest non-conju- 


gants te : ase 100 =136-208 172.48+0.28) 100 32-72 47.02+0.52 


measured, as shown in the table, while the rest were kept till 
April 6. The progeny of the non-conjugants were much larger 
than the progeny of the conjugants—the former averaging 
172 to 198 microns, the latter 137 to 148. It is clear that 
there existed in the culture diverse races, and that only certain 
smaller races were conjugating. Thisis confirmed by the fact that 
I actually isolated from this culture a number of lines that were 
permanently diverse. It is further shown by the fact that when 
conjugation occurred in the progeny of the large non-conjugants 
(which happened March 22), the pairs were large, averaging 182.33 
microns in length, as opposed to 139.29 for the pairs first taken. 
Thus what we have really done in this case is to compare large and 


CONJUGATION IN PARAMECIUM 103 


small races; these (as is the rule) retained their relative sizes 
irrespective of conjugation. Where the racial composition of 
conjugants and non-conjugants is the same, the progeny of the 
conjugants are the larger, as we have seen. 

d. Summary. We may summarize our results from the exper- 
iments with pure races, and that with a wild culture, as follows: 

1. There is no tendency for the progeny of the (small) conju- 
gants to be smaller than the progeny of the (larger) non-conju- 
gants. 

2. On the contrary, in every case the first measurements taken 
after fission show the progeny of the conjugants to be larger than 
the progeny of the non-conjugants. In the wild culture of table 
32, this difference was still very marked in the seventh genera- 
tion after conjugation; in other cases the two sets had apparently 
become equalized at this time. 

This greater size of the progeny of conjugants is apparently 
due, partly, if not entirely, to the slower fission after conjugation. 
The conjugants usually do not divide for 24 to 48 hours after 
separation, and in the meantime they grow rapidly large, as we 
have before seen (table 19, ete.). Conjugants of k were observed, 
by the use of India ink, to begin feeding within two hours after 
separation. Meanwhile, those that have not conjugated are 
dividing regularly, so that they have no opportunity to become 
so large as the ex-conjugants. When the latter divide, their 
progeny for a number of generations are therefore larger than the 
progeny of the non-conjugants. The slower fission of the ex- 
conjugants may continue for a long time, thus giving them an 
opportunity to remain larger for many generations, as in table 32. 

3. There is a characteristic difference in size between indi- 
viduals at the end of the cycle (before conjugation), and those at 
the beginning of the cycle (after conjugation). The size is some- 
what greater at the beginning of the cycle. The difference, in 
the case of a pure race, is not very great, and is not such as_ to 
conceal the more marked differences between different races; 
the latter persist throughout the cycle. 


104 H. S. JENNINGS 


3. Increase in variability as a result of conjugation 


In the wild culture of table 32, the progeny of the conjugants 
are on the whole decidedly more variable than the progeny of 
the non-conjugants. It appears therefore that conjugation in- 
creases the variability. This result is of extreme importance; 
for this reason I have made an extensive study of the relations 
involved. This matter is inextricably bound up with the prob- 
lem of the significance of conjugation and its relation to vitality 
and rejuvenation. It appears best therefore not to enter upon 
a detailed account of these things in the present paper, which 
has grown to a considerable length, but to reserve them for a 
separate paper. The questions to be answered from our present 
standpoint are: 

a. Is it the rule that conjugation increases variation? Is it 
true both for wild cultures and for conjugation within pure 
races? 

b. In what characteristics is the increased variation shown? 
Does it appear in other matters besides size? 

c. Are the variations thus produced heritable? Are there in 
this way new races produced, differing in heritable characters 
from the race or races that entered conjugation? 


Reserving then these important questions for another paper, 
we take up here certain other points that are naturally treated 
in the present connection. 


4. Do pairs of different size give progeny of different size? 


On the answer to this question of course depends the decision 
whether assortative mating has any effect in inheritance; in the 
production or perpetuation of permanent differentiations. The 
question has already been answered in my paper of 1908. It is 
there shown that six pairs, of different sizes, isolated from culture 
M, gave six races differing permanently in size—four belonging 
to caudatum, two to aurelia (pp. 494-496). This fact, taken in 


CONJUGATION IN PARAMECIUM 105 


connection with the demonstration that larger specimens mate 
with larger, smaller with smaller, shows that assortative mating 
has clearly the effect of keeping differentiated races from mixing; 
of perpetuating such racial differences as already exist. When 
a culture containing two species conjugates, the two as a result 
of the assortative mating remain quite distinct, as we have be- 
fore seen. When races of diverse size are present, these like- 
wise tend to remain distinct in spite of the occurrence of con- 
jugation. 

But what is the result when all of the conjugating individuals 
belong to the same race? Do large pairs then give large races, 
small pairs small races—so that as a result of conjugation we get 
hereditary diverse strains from a single race? 

The answer to this question turns out to be bound up with the 
general problem as to the physiological effects of conjugation on 
the stock. I therefore reserve its full treatment for a paper 
which is to deal with that matter, merely stating here in a sum- 
mary way certain points. 

a. When we select large and small pairs from a culture of a pure 
race (all derived from a single individual), we do not find that as 
a rule the larger pair gives a larger race, the smaller a smaller 
one. Usually the two give a race of the same size. 

b. Yet hereditary differentiations do arise as a result of conju- 
gation within a pure race; certain pairs or individuals give strains 
that differ permanently from the strains produced by others. 
These differences are most noticeable in the matters of general 
vitality and rate of multiplication, but differences in sizes are 
likewise to be observed,—possibly due to the differences in gen- 
eral vigor. The extreme importance of these matters requires 
a full treatment on a paper devoted to the subject, where the 
experimental data may be properly set forth, in their relation to 
other phenomena. The indications are that these differentiations 
within a single race are not connected with assortative mating, 
and hence that the latter has no marked consequences when 
occurring within the limits of one race derived from a single 
individual. 


106 H. S. JENNINGS 
5. Inheritance from unequal pairs 


The last question raised in our introductory outline was in 
regard to the rules of inheritance when the two members of a 
pair are of different size. This matter again turns out to be com- 
plicated by its relation to the problem of the physiological effect 
of conjugation on the stock; furthermore, on some of the most 
important points I have as yet been unable to get experimental 
data. Adequate treatment must therefore be deferred; and I 
give here merely categorical statements on certain points. 

a. The assortative mating of course tends to prevent the mating 
of individuals differing in size. Yet this does occur at times. 

b. The only way to get the rules of inheritance in a satisfac- 
tory way would be to get matings between individuals belonging 
4otwo diverse stocks, the inherited characteristics of each of which 
were known beforehand. Such matings between known races 
I have not yet succeeded in getting. This is owing to the fol- 
lowing facts: (1) diverse races usually conjugate under differ- 
ent conditions, so that it is very difficult to induce two such to 
conjugate at the same time. (2) In the rare cases where two 
races have conjugated in the same culture at the same time, 
they have not inter-crossed, owing to assortative mating. Hence, 
in spite of much work, I have not been able to determine this 
point. (3) The progeny of the two members of a pair often differ 
hereditarily, particularly in vigor and rate of reproduction; some- 
times also (as a secondary result?) in size. Such differences in 
vitality have been previously noted by Calkins (’02, p. 175) and 
by Miss Cull (07). Their significance is bound up with the prob- 
lem of the effects of conjugation on the stock; they will therefore 
be taken up in a later paper on that matter. 


SUMMARY 


1. The members of conjugating pairs are, as Pear! had set forth, 
smaller and less variable than the non-conjugants of the same 
culture. This is true in cultures consisting entirely of the pro- 
geny of a single individual. It is likewise true, as a rule, in cul- 


CONJUGATION IN PARAMECIUM 107 


tures consisting of a mixture of races. But exceptions may occur 
in the latter case, owing to the fact that only one of the races 
present may conjugate. If this happens to be a large race, the 
conjugants may be larger in size than the non-conjugants. 

2. The small conjugants increase in size after conjugation, un- 
til they are as large as the large non-conjugants of the same race. 
Also, the large non-conjugants later divide and themselves con- 
jugate. Hence there is no permanent or fundamental distinction 
between the two sets, but only a temporary physiological differ- 
entiation, which is of no consequence in the later history of the 
stock. 

3. The lessened variability of the conjugants is due (in the case 
of a single race) to the facts that: (1) they include no young in- 
dividuals; (2) they do not grow so large as the non-conjugants. 
The latter, including young (and therefore small) individuals as 
well as large ones, together with all intermediate stages, must 
show more variation than the conjugants. But this greater vari- 
ation is due to temporary physiological differentiations, of no 
consequence in the later history. The progeny of the conjugants 
are fully as variable as the progeny of the non-conjugants (see 
paragraph 7). 

In cultures containing several races, the conjugation is often 
limited to but one of these. The less variability of the conjugants 
is then due chiefly to this fact. The progeny of the conjugants 
(all belonging to a single race) will naturally then be less variable 
than the progeny of the non-conjugants (belonging to several 
races). In this case therefore the less variability of the conju- 
gants shows its effect in the later history. 

4. As Pearl sets forth, larger individuals are usually found to 
be mated with larger, smaller individuals with smaller, so that 
there is a considerable degree of correlation in size between the 
members of pairs. This correlation is greatest in cultures con- 
taining pairs belonging to the two species, caudatum and aurelia; 
considerably less in cultures consisting of many races of a given 
species; still Jess in cultures consisting of a single race, though 
even here the correlation is considerable. 


108 H. S. JENNINGS 


An exhaustive analysis of the various factors which might pro- 
duce such correlation (the details of which are given in the text) 
shows that it is due to real assortative mating (as Pearl main- 
tained) ; larger individuals mate with larger, smaller with smaller. 
There is a certain amount of equalization occurring during the 
process of conjugation, but this is not sufficient to account for 
the correlation; the latter persists after the equalizing process 
has ceased to be effective; and correlation is shown in certain 
dimensions that would not be affected by the equalization. 

The assortative mating shows itself in the three following 
effects: 

a. When the two species caudatum and aurelia are present in 
a conjugating culture, caudatum conjugates only with caudatum, 
aurelia only with aurelia. Hence crossing is prevented. 

b. When races of different size are present (as is usually the 
case in ‘wild’ cultures), the members of large races tend to con- 
jugate together, on the one hand; of the smaller races on the 
other. Thus the assortative mating tends to prevent the admix- 
ture of races and their reduction to a single type. 

ce. Even where members of but a single race are present, larger 
individuals tend to mate with larger, smaller with smaller. It 
is not clear that this has any consequences for the later history, 
though this is not yet disproved. 

5. Isolation from a wild culture of large pairs gives large races; 
of small pairs, small races. Thus it is clear that assortative mat- 
ing has most important consequences in the later history of the 
organisms, tending to preserve the existing differentiations of 
species, races, ete. 

6. Conjugationresultsin certain slight but characteristic changes 
in size. For a few generations after conjugation the progeny of 
the conjugants are a little larger than the progeny of those of the 
same race that have not conjugated. This is apparently due to 
the slower fission of the progeny of the conjugants. This dif- 
ference in size is much less than the difference in size between 
diverse races, and after a few generations the progeny of con- 
jugants and non-conjugants as a rule show no characteristic 
differences in this respect. 


CONJUGATION IN PARAMECIUM 109 


7. The progeny of conjugants are more variable, in size and 
in certain other respects, than the progeny of the equivalent non- 
conjugants. Thus conjugation increases variation (a matter to 
be dealt with in full in another paper). 

8. Hereditary differentiations arise as a result of conjugation 
within members of the same race (all derived from the fissions of 
a single individual) (to be dealt with in full later). 

9 The progeny of the two members of a pair sometimes show 
hereditary differences (to be dealt with in full later). 


LITERATURE CITED 


Cauxins, G. N. 1902 Studies on the life-history of Protozoa. I. The life- 
cycle of Paramcecium caudatum. Arch. f. Entw.-mech., Bd. 15, pp. 139- 
186. 


Cauxins, G. N. anp Cutt, 8. W. 1907 The conjugation of Paramecium aurelia 
(caudatum). Arch. f. Protistenkunde, Bd. 10, pp. 375-415, pl. 12-16. 


Couuin, B. 1909 La conjugaison d’Anoplophrya branchiarum (Stein) (A. cir- 
culans Balbiani). Arch. d. Zool. Expér. et Gén. (5), tome 1, pp. 345- 
388. 


Cuz, Sara W. 1907 Rejuvenescence as a result of conjugation. Jour. Nxp. 
Zool., vol. 4, pp. 85-89. 


Enriqups, P. 1908 Die Conjugation und sexuelle Differenzierung der Infusor- 
ien. Zweite Abhandlung. Wiederconjugante und Hemisexe bei Chilo- 
don. Arch. f. Protistenkunde, Bd. 12, pp. 213-276, Taf. 17-18. 


Hertwia, R. 1889 Ueber die Conjugation der Infusorien. Abhdlg. d. II. Cl. 
d. koénigl. bayr. Akad. d. Wiss., Bd. 17, (Abth. 1), pp. 151-233, 4 Taf. 


Jennines, H.8. 1908 Heredity, variation and evolution in Protozoa. II. Hered- 
ity and variation of size and form in Paramecium, with studies of 
growtb, environmental action and selection. Proc. Amer. Philos. 
Soc., vol. 47, pp. 393-546. 


1910 What conditions induce conjugation in Paramecium? Jour. 
Exp. Zool., vol. 9, pp. 279-300. 


1911 Computing correlation in cases where symmetrical tables are com- 
monly used. Amer. Nat., vol. 45, pp. 123-128. 


Jennines, H. S. anp Harairr, G. T. 1910 Characteristics of the diverse races 
of Paramecium. Jour. Morph., vol. 21, pp. 497-561. 


Lister, J.J. 1906 Biometry and biology; a reply to Professor Pearson. Nature, 
vol. /4, pp. 584-585. 


110 H. S. JENNINGS 


Maupas, E. 1889 La rajeunissement karyogamique chez les ciliés. Arch. d. 
Zool. Expér. et Gén. (2), tome 7, pp. 149-517, pl. 9-23. 


Peart, R. 1907. A biometrical study of conjugation in Paramecium. Biome- 
trika, vol. 5, pp. 213-297. 


Pearson, K. 1906 The latest critic of biometry (etc.) Nature, vol. 74, pp. 
465-466; 608-610; 636. 


Srmpson, J. Y. 1901 Observations on binary fissioa in the life-history of Ciliata. 
Proc. Roy. Soc. Edinburgh, vol. 23, pp. 401-421. 


Wooprurr, L.L. 1907 Variation during the life-cycle of infusoria in its bearings 
on the determination of species. Science, vol. 25, pp. 734-735. 


APPENDIX 
TABLES OF MEASUREMENTS 


In the following tables the measurements are given, except in certain cases 
specifically mentioned, ia units of 4 microns each, so that to get the measurements 
in microns or thousandths of a millimeter the measures given are to be multiplied 
by 4. Thus, in table 34, lot 1, the fir:t individual is 33 units or 132 microns (0.132 
mm.) long. But in certain measurements of breadth (tables 57 and 65), and in 
lot 2 of tables 34 and 36 the unit of measurement is 2 microns; this is specified in 
the proper places. 

In the tables of correlation for the pairs, the larger individual is entered always 
in the vertical columns, the smaller in the horizontal rows, each pair being entered 
but once. This is done in accordance with my note of 1911, showing that double 
or symmetrical tables are unnecessary for such cases. 

On some of the lots a number of different measurements were taken, and these 
were divided up and classified variously, so that a large number of the ordinary 
correlation tables would be required for bringing out the diverse relations devel- 
oped. In such cases I have given, in order to save space, a classification of the 
original measurements, from which correlation tables can readily be formed for 
all the dimensions dealt with. Such is the case in tables 36, 40, 49, 50, 51, 55 and 
56. 

The two individuals of a pair are denominated A and B. In cases where one 
of the pair projects anteriorly in front of the other, this projecting individual is 
called A, the other B. 

Tables 1 to 33 are found in the text. 


CONJUGATION IN PARAMECIUM 111 


TABLE 34 


Comparative measurements of conjugants and non-conjugants from wild cultures 
(lots 1-7) and from a mixture of two species (lot 21). C,conjugants; N, non-conju- 


LOT iY ais} 2 4 5 6 if 21 
: wid \pai eps. areca SS = 
Length |C | N|C|N|\Length| C | N|\Lengthhc |N/|C|N|C|N|C|N|C|N 
me | ta eS = tee BS | Le 
33 1 | 64 1 25 1| 
37 1 | 65 26 } | 3 hae 
38 1 | 66 2 27 | 6) 1 
39 Pol} | Weer 1} 28 1 3 | 18 
40 | 2] 1] | 68 2 29 1} 4] 1] 3 | |19] t 
41 4/1] 2] 1]} 69 1} 1} 30 ih A \/.-2 Rh eo) 
42 Pe ae 70 n 31 4] 8) 1] 4 i |16] 3 
43 4/2] 5 ral 32 Oma aay Qed 1 Al) 2 
44 {at pt | alles | 3) 33 | 10) 1) | 3} 10 1 | 9} 9 
45 | 23] 3] 7 73 2] 34 MATA 2) A) | 8t) | 1] 10 
46 } 19] 3] 17) 2]) 74 | 41] 6] 35 | 19] 12} 1] 10 1} (14 
47 29/ 9/15) 2] 75 3/ 2] 36 13} 9/ 2} 5] 1) 2) 4 9 
48 | 36/14/13] 1]/ 76 |25| 4]) 37 Halle) Buletalwedy | 5)| 4 7 
49 35/14) 21] 7] 77 6/ 2] 38 Tel ATi |G!) £3) Gs) 2 |) Lk 8 
50 | 41 | 14 22| 7|| 78 | 26] 4] 39 2) 10 | 9} 6] 9} 9] 6] | 9 
51 38 | 14] 17 | 12 79 18] 5 || 40 3 | 13 | 5| 6}11} 11] 9 | 8 
52 34/33/18] 8] 80 8] 21) 41 8 | 4] 12] 19/22) 8 7 
53 VES elbonl alas ee fectall| viet) 4 5 | 20 | 20| 31 | 14 7 
64 15/29) 4/13] 82 | 11] 7] 43 12 | 32 | 21 | 20 | 10 4 
55 115) 25| 6|i4] 33 22] 5 || 44 2 U| 27 | 26 | 22 | 11 1 
56 §{23) 3/12] 84 |18| 3]| 45 5| | 2/30}18] 5] 8 2 
57 | 4]/18] 2/13] 85 ;12] 3] 46 | 6 | 25] 8| 10/11] 2 
58 4) 20 | 11] 86 21| 6 47 3 | 3|31})21| 3] 8| 
59 | 1/16] 1] 9] 87 | 9| 6] 48 2| 31/25] 6| 12| 
60 | 1] 18 6|| 88 | 15] 4) 49 L7|27)|, 1) 8) | 3 
61 | 16 9] 89 | 7] 4] 50 3 1| 18 | 25 53] 1 
62 2| 16 6 || 90 Ta eh eo ETI] SA fae Oi fe 
63 Wee exGy fee a (ie 2H (eon Siloti) 162 1 4 | 16 4 | 1 
64 | 10 41 92) |.7 | 53 12 | 2 3 
65 } 1] 8 Z|) ©9384 54 2] 7} 6 3 
66 | 4 | 7] 94 4 | 55 1} By hale {125 
67 3 | 4 || 95 3/1] 56 2M) Waa | 8 
68 Het | 4 96 | 4] || 57 | 5 3 
69 3 | 4] 97 2 | 58 1 | 3 
72 | | 2} 98 | 2] 1] 59 2 | 1 | 4 
73 | 1 99 | 1] 1] 60 2] | | 
77 1 100 | 1 | 61 1 : 1 ih 
78 | WeLOIe OH at) ge ry ih Mee 
79 | 1 io2 | 1) | 62 | eo al 2 
80 | i | 64 ; | 4 
| | i | 65 ee 
66 |} 2 


| | | | | | | | 
Total number. 360 360 164 176 | 284 | 87) 84 /152 | 16 100 272 /318 |158 |131 | 98 |156 


= 
GFT) FZT STI) FLT) O01 SOT) OTT, SET 


00T 


OO TE0¢ 


PE 


GF OOT 0OT) 9ST OOTSzeFZs) 00Z, 


988 “goquinu [B40], 


ASCWDMDMOM OD — 


AAMTANH HM DONABDAT OHH 


Arnns 


aAamN 
aed 
= = 
DY 00. 9 CRN6D, A S69) 4 ec 


eo) 
a 


JENNINGS 


H.. 8. 
mMaAMmMnDmDoDMDn ON 
1 


= 
mARM THA MOANKR SO 
i = 


a 


saa 


ao 
a 
nmnnontaa 


AAHODOAAA 
=a 


| o¢ 
| 6F 
SF 
Lt 
OF 
cr 
FF 
et 
ch 
If 
| OF 
68 
| 88 
Le 
9g 
ce 
| FE 
£& 
| aa 
| 1 
0g 
| 63 
8¢ 
| L 
| | 9% 
| cS 
id 
&% 
aa 
1% 


ertON HE MO 
— 


om 
= 
= 
i) 


S 
* 
eourne 


mot NnKMET SOMA 


o 
= 


a4 


AMM ANTDONNAMW 
ot 
aq 


Sr) 


ee 
AtOFOonrt oe 


Oa InN! 


> i So Be I) 


= 


a 


mmoreanonawtwon 


amo aoKr nana 


| 
| 
| 
| 
| 


yyouaTy 


112 


fo Uy 


‘squpbnluos-wou ‘xy fspupbnluoo 


SsUoLoU F “Wuauiaunspaut 
‘(saunqjna yons fo saunjxiu Lo ‘DyainD 


‘ 


0) 


fo saonu aind fo sainqjna Ayurvw) oe¢-g $10) fo sjunbnlu0s-uow puv syunbnluoa fo syybuaT 


st ATAV.L 


CONJUGATION 


TABLE 36 


IN PARAMECIUM 


113 


Measurements (in units of two microns) of the 142 pairs of lot 2 (wild culture), classi- 
fied according to the distance that A projects in front of B. The first column gives 
a classification of the various lengths of the projecting individual A; the other col- 
umns give the lengths of their mates B for each pair, under the different anterior pro- 
jections of A. Thus, there are eight pairs with A 76 units long; their mates B 
measure respectively 75, 73, 74, 76, 76, 72, 74, 74 units; in the first four the anterior 
tips of A and B are even; in the others A projecis in front of B 2,4and7 units, 


respectively 


LENGTH OF MATES, B, WHEN THE DISTANCE A PROJECTS BEYOND B, 1s 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 1 


| 8 


| 
0 1 O93 
| 
70 
72, 73 
(2)23, 74, (2)76 | 72 74 
| 
(2)68, 72, (2)74, (3)76 73, 74, 76 | 78 72 
78, 79 | 
1(2)76 | 72 
78, (2)79, 80 |76.79 | 80 
76, 81 78 74 
78, 79, 82 79 | 80 | 75, 76, 78 | 78, 81 
74, 76, 78, 79 77, 83 77 69, 74, 84 
75, 85 81 /79, 81 79 
| 75, 79, (2)83, 84, 85, (3)86 | 77 85 76,83 | 85 
77, 86 83 79 
76, (2)83, (2)84 88 | 82 87 83 
85, 86 86 
87, 88 |80,88 | 79 
84 79 82 
84, 90 78, 87, 92 | 
85 82 
82, (2)93 76 
92 
90 78 /79 
83 179 
| 88 | 
78 
om 


xan 
oe 


6 


76 


| 7 


a 


72 


114 


Lot 8 (wild culture). 


43 44 45 46 47 


EH: 5: 


TABLE 37 


48, 49 50 5 


bo oo be 


] 
at 
1 
2\1 
1 2 
S28 [2 
V4 3 
1/2/1 
2\3 
5|5|5 |11 |11 22 |12 


TABLE 38 


JENNINGS 


bo 


Length of A by length of B 


52| 53 54 55 56 57 58 59 


1 
| 
| 1 
1 
1 
1 
3 | 2 1 


Lot 4 (wild culture). Length of A by length 
of B 


31 32 33 34 35 36 37 38 39 40 


29 1 1 
30 1 1 2 
Sie 22 3 
32 eal 3 5 
33 1 3/2 1OHi7, 
34 D211 5 
35 D3) 3 1/9 
36 1p a 5 
37 ihe 14 
38 1 

1 | 4|3]2)10|8)3)6 3 |42 


_ 


45 


49 


CONJUGATION IN PARAMECIUM 


TABLE 39 


35 36 37 38 39 40 41 42 43 44 45 46 47 48 


1 
Pali age 
1 
aly Pal i} 
ial Later tel 1 
1H a ef Lt if 1 
2) 4 2 
Dee Ds Sats) S25 hel 
SHSy wor Ieee) lay LN pe a a 
463. \eo.2") a 
1)/4)4/4)3 
BE | Pet) a) a Es} 
CT Cs ha Da fe | 
1); 4/2 
4 


| 1| 3) 2 18 12 |14 16 16 20 13 |13 


Lot 6 (wild culture). Length of A by length of B 


49 50 51 52 53) 54 


115 


116 H. S. JENNINGS 


TABLE 40 


Lot 7 (wild culture). Measurements, length from anterior end to mouth, and total 
length; classified with relation to the distance that A projects in front of B. (The 
lengths ‘‘To mouth’’ are not repeated when successive pairs are the same in this 
respect.) (1 unit = 4 microns) 


ANTERIOR ENDS EVEN ANTERIOR ENDS EVEN PROJECTION 1 PROJECTION 1 


To Mouth Total To Mouth Total To Mouth Total To Mouth Total 


B A B A B A B 


B FGM bats) A B|A|B A | 
25 | 25 | 37 | 36 46 | 42|| 25] 24 | 39 | 37 51 45 
40 | 37 | 28] 28] 41] 41 | 41 38 Sa ee 
PROJECTION 2 
42 40 42 40 42 36 i z 
26 | 26 | 39 | 38 | 42/41] 26] 25] 40/39] 25/ 23] 38] 36 
| 40 | 39 | | 43 | 40 40 | 41 | 27| 25 | 39 | 36 
| 41 40 | 43 | 42 27 26 38 | 39 41 | 41 
42 41 43 | 42 40 | 38 | 28| 26] 44 | 38 
42 | 42 43 | 42 43 | 37 44 38 
42 | 42 43 | 42 44} 41) 29 27 | 42| 43 
43 39 43 | 43 | 44 | 42 44 40 
43. 41 44 | 42 44 | 42 48 44 
44 | 37 48 | 44 | 28) 27] 41 | 39 48 | 45 
45 | 42] 29] 29) 43 | 42 41 | 42 ms 
PROJECTION 4 
46 41 46 | 42 | 42 | 41 = 
46 | 41 46 43 42| 43 26| 22] 44 | 32 
27 | 27 | 41 | 41 46 44 44 | 41 
42 41 47 | 42 44 | 42 
42| 41) 30| 30) 46) 45 44 | 42 
43 40 47 | 42 44 42 
43 41 48 | 45 48 | 40 
44 | 41 49 16) 29} 28 | 43 | 39 
44| 43) 31] 31) 47 | 44 | 46 44 


CONJUGATION IN PARAMECIUM nlalz¢ 


TABLE 41 


Lot 7 (wild culture). Non-conjugants; length before 
mouth with length behind mouth 


Length before mouth 


24 25 26 27 28 29 30 31 32 33 34 35 

SSS 
12h il Via el ae ht 2 
See] Eh as ene Ba 17 
S14) 2h pa | 4 
2 15)| si 2h) ti 11 9 
m3 16/1/1/2/3)5)1/2/3)1 19 
S17 3) |1)2/4/2/4)1 17 
'$ 18 0) ed li Une |e 2a 13 
2 19 | (2piideel ed rales | |6 
220 Ay ze 3 elk 
a) 21 | ome i | pa eA ale FOG 
22 he 3 | | |3 
23 | 1 | 1 | 2 

mist aiediete pee laid 
2 {10 | 8/11 [12 [10 | 9 [15 14) 1] 1 | 1 94 

| { cere | Keel 


TABLE 42 
Lot 8. Length of A by length of B 


31 32 33 34 35 36 37 38 
Se ae 
30 2 | | 2 
Sine Belin | ee 3 
32 let 1 
33 | 1 2 Teo 7 1 | 9 
34 (BT ete seh a 5 
35 | Pye eal 5 
36 | ipl he ga faale 9 

| 

: | aa) 
1\/7/2)4)6)5|1)] 1 (27 


118 


H. 8S. JENNINGS 


TABLE 43 
Lot 9 (race c). Length of A by length of B 
33 34 35 36 37 38 (39 140 41 42 |43 lea las 


1 1 2 

}1|t 2 | 4 

1 | | 4 

2 eal Ot) 13 10 

Lael ae | Swale alartun Marg 

2/3/4/5| |3 17 

217|3|5|6 23 

oH Aes: 3 20 

2/2/5)5|6/1 | 1 | 22 

i be es 7 

Pa 3s I 8 

1 | 1 

L 1 

2/3 6/7 (18 18 12 23 13 16/3/31 (125 
TABLE 44 


Lot 10 (race k). Length of A by 
length of B 


30 31 32 33 34 35 36 
28 Ne a 1 
29 iy p34) al Al 
30°), 1: C2 yeoTl ae) 5 
31 Esl) Aa | Lae 7 
32 Ppl aba at (ete 5 
33 2 | 2 
34 | 1 1 


CONJUGATION IN PARAMECIUM 


TABLE 45 


Lot 11 (race k). Length of A by 
length of B 


LS) 
oo 


29 30 31 32 33 34 


24 1 1 
25 | 1 1 
26 | 1 1 
27 | ty} 2 2 
PAST ul Pena (ea Ee PG an 5 
29 al 
30 1 Le) (2 
31 1 1 


bo 


2 243 1 | 14 


TABLE 46 
Lot 12 (race k). Length of A by 
length of B 
28 29 30 31 32 33 34 35 


26 1 1 2 
DNe2 ste) (len a 5 
28 Sesame sys 10 
29 TOV t RAs OE VE, er 9 
30 31316 12 
31 3771915 24 
32 Sulu a2 8 
33 4} 2) ).2))\ 8 

3 | 3] 8 |14 22 116 110 | 2 | 78 


119 


120 


H. S. JENNINGS 


TABLE 47 
Lot 18 (race k). Length of A by length of B 


126 [27 28 [29 /30 |31 32 33 34 35 36 37 38 39 
Ate aa Ay Teta a eal 
23 | 1 | 1 
24 | 
25 | J | 1 
26 | Pe ela 1 | | 5 
7 [at | il fe ea | 5 
28 | 1132) 1S Sa" st 11 
29 | AOA 10) Gey Ze lea | 1¢ 
30 1/12|6/3/6 1| ¥| 30 
310 3.) 4 cba Syl eels alee 
32 | 5/5|4/3|5/3/2|] | 27 
33 | 4\/4/6/2/2/2) | 20 
34] | }5/3)4/2] | 1] 35 
Bt) 04 1j1lalal 4 
36 W2elcoul oI 5 
37 | | ja 1 
110 i i 5 17 17 25 36 19 20 14 6 | 2 168 
TABLE 48 
Lot 16 (race C2). Length of A by length of B 
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 
| | ees (oma eat | 
= | [ | heals lal rte 
25 | 1 | | | } 1 
26 | |_| | | Wess ie 1 
27) ns ey | 1 | 7 
28 1 Ve WS atrouiea fete | 15 
29 | | lslzisl2lsialaiiia | | 28 
30 | 3/5) 4/4/2) | | | 18 
31 | SR Oe RS Sa Oa a 
32 | Ss anes aad | 20 
33 | Wpeliai felts 534 3 6 
34 }5 73/2/11 11 
35 | Hee eeu salen 10 
36 | | Hee lari Nay 7 8 
37 | | (sya) | 4 
| | | | | | t 
11] 1/4 [13 a5 {14 (19 |17 |17 |12 16 [13 | 7 3/1 180 


CONJUGATION IN PARAMECIUM " 121 


TABLE 49 


Measurements of length for the 69 pairs of lot 17 (race C2), classified according to the 
distance that A projects in front of B. Unit, 4 microns. (Compare the explanation 
of table 36) 


LENGTH OF MATES, B, WHEN THE DISTANCE A PROJECTS BEYOND B 1s 


A es ae = 
0 1 | 2 3 4 
26 | 24 22 
2 27, 28 | 
28 28 27 | 25 
29 29 24, 25, 28, 29, | 27 
30 
30 26, 28, 30, 32 | 28 25, 27, 28 
ol 26, 30 (2)31 27, 28 
32 (2)31, 30 (3)29, 20, 31 31 
33 24, 30, 31, (2)32) 30, 31, (2)32 27, (2)28, 30 27, (3)28, 30 
33 
34 (2)32, 34 33 
35 (2)33, 35 33, 34 29 26 
36 34 
37 | 28 
Total num- 


ber......| 28 24 /14 5 3 


TABLE 50 


Measurements of length for the 84 pairs of lot 18 (race g), classified according to the 
distance that A projects in front of B. Unit of measurement 4 microns. (Com- 
pare the explanation of table 36) 


LENGTHS OF MATES, B, WHEN THE DISTANCE A PROJECTS BEYOND Bis 


A — - oo 
oO 1 4 2, 3 
28 27 
29 (2)28, 29 27, 28 
30 28, (4)29, (4)30 | 27, 28, 29, (2)30, 32 27 
31 (2)29, (6)30, (3)31 | 28, 29, (2)30, 31 
32 (2)31, 32 27, 28, (5)29, (3)30, 31, 32 | 27, 28, 29, 30 
33 30, (2)31, (2)32 26, 29, 31, (3)32 30 29 
34 (2)33, (2)34 30, (2)31, 33 
35 32, 33 (2)30 
36. 31 
37 
38 33 


Total num- 
ber 202. 38 37 8 1 


TABLE 51 


Lot 9 (race g). Measurements of conjugants in length from anterior end to mouth, 
and total, classified according to the distance that A projects in front of B. (The 
lengths ‘‘to mouth” are not repeated when successive pairs are the same in this re- 
spect.) Unit of measurement, 4 microns. (Thus, in the first pair, the distance 
from anterior end to mouth is 64 microns in both individuals, the total length of A 
is 100 microns, that of B, 96 microns) 


ANTERIOR ENDS EVEN PROJECTION 1 PROJECTION 2 
To Mouth Total To Mouth Total To Mouth | Total 
A B A B A B A B A B | A B 
16. |) 16) | -25° | "24 | a7 | 16.270 26 | 17 | 15 | 290 | 26 
28 | 26 | 27 | 26 18 | 16 | 208 eu 2i7 
ge || alge eee Ih Beye 2 7auli 26 19 | 17 | 28° | 2 
| OT Oy 7 | 27 | 29 | 95 
| 28 27 e2Sianlee 30 | 29 | 98 
| 98 | 28 | 18 |.17 | 27 | 26 | "84." ) 729 
18 18 27 27 | 2433 27. | | 34 30 
27 | 27 | 28; 27 || 20 | 18 | 32 , 27 
2 8ia1n026 29-28 34 29 
PA) || yy | 208 | o8m |) 2 | 219" 836 | 29 
28) | 30) | 30 | 30 ~ 
29 | 28 | 31 29 | PROJECTION 3 
29 |. 30 | 19 | 18 | 28 | 27 |— 
29") #30 99 || 25 || 17° | 14 | 29 | 24 
BY Il) ey 20), |) 227 19ealGmleesl | 28 
30 | 28 | ZOU 28a POM Taniese: | 28 
10} | eel 290, 298) SOTA ola Seeca, | 28 
29 | 29 31 | 29 35 29 
29 | 30 30a 228 23 ele OOM NSS! | sol 
30 | 29 SO es28 a, lee ee ee 
30 30 32 29 PROJECTION 5 
31 | 29 32 | 30 |————— 
Sie COMME 2ONN |e 1O OOUN Su ooeueeti: N35 || 26 
| 31 | 30 29) || 129 | 
31 | 30 S2anle 28 
Bile |) Sel Hero Qala 
20a 20 lecstn || 38 33 | 28 | 
32 29 33 31 || | 
3230 34 | 29 | | 
32 | 30 35 29 | 
SOMO Meme o1e | 200 1933) 32 il 
3380 
oe | S10 | 
33 | 33 
340] 939 
Oe 21533) 130 
35 30 


124 H. S. JENNINGS 


TABLE 52 


Lot 19 (race g). Non-conjugants; length behind 
mouth with length before mouth 
Length behind mouth 


9 |10 /11 /12 |13 |14 |15 16 |17 |18 19 [20 


yaya abjal | | 2 
aw 
Se | 
2 15/ | | 
2 16/1) 1 2 
= 17 ji 3/3/1] | 8 
FS WD BD ile Nat 10 
S19) |1] ft)3}3 18 
= 20; | |4/2 7 3 15 
91 | | 1) ae 4) (l/s Pata | | 19 
22 1} 2])1/6)2)2 1} wea ils 
PEI al Naud [ee 1 12 
24 | Dil ala 55 2 Tt eeal| tel 
25 Tae lid) 4 
26 1 | 1 | 2 


| 2] 8 [12 (24 j23 22 14 | 7 | 2/2/11] 1 1118 


CONJUGATION IN PARAMECIUM 


TABLE 32 


Lot 20; conjugants yrom mixture of races Cz 
andi. Length of A with length of B 


] | 
29 30 31 32 33 34 35 36 37 38 39 


RPnNN Ne 
i 
— 
— 


rFPoWn ere 
em 
NOONAN AWWH FE 


bo 
eS woe be 

e 

— 


1]4\/3i9lon7i|6|7\|4\1/1]| 6 


TABLE 54 


Lot 21; conjugants from mixture 
of races Ly (caudatum) and k 
(aurelia). Length of A with 
length of B 


28 29 30 31 |32 33 34 


27:15 1 | 6 
28)1 3) 1 eal se id 
29 3 loa 4 ite 2 13 
30 5|5)3)1 14 
31 reeled |i 5 
32 Hat | a 2 
33 2 2 


126 H. S. JENNINGS 


TABLE 55 


Lot 22; 102 pairs of unseparated conjugants; measurements of length to mouth, 
total length, and breadth, for A and B of each pair. (In some of the pairs 
only the total length was measured, as the table shows.) Unit, 4 microns 


TO MOUTH TOTAL LENGTH BREADTH TO MOUTH TOTAL LENGTH BREADTH 
A B A B A B A B A B A B 
26 25 44 45 7.5| 7.51) 30 28 45 42 8 9 
26 26 42 42 i is 47 44 foal 
43 43 8 8 30 29 47 43 Said 
27 25 43 43 ub 6.5 49 48 7.5 8 
44 42 7 6.5 31 28 49 44 7.5: \7 
27 26 39 40 6:5 | 7 31 29 47 45 tC 
43 40 cf 7 32 29 48 44 
46 43 8 7 32 30 47 43 fhe AALS 
27 27 43 43 ef vi 33 26 48 42 8 6 
43 42 7.5| 6.5 39 38 fe lif 
44 42 8 6.5 43 42 6.5 6 
44 43 7.5] 8 43 43 8 7 
44 44 7 6.5 45 42 Tan 
45 44 8.5| 8 45 43 87 
46 46 7 tf 47 45 8 7.5 
46 46 7 8 49 44 a | 
28 25 47 44 a 6 51 45 te ANT, 
28 26 41 2 ih ia = 
42 44 7 8 Only total length measured 
43 42 theaye | wie) iol. = ae 
28 Q7 45 44 7 7 eet A LENGTHS OF MATES, B. 
46 41 8 6 ——$<$— 
46 44 7 6.5 t 41 38, (3)40 
47 43 8 iq 3 42 39, 40, 41 
28 28 46 43 8.5 | 6 7 43 39, 40, (3)42, (2)43 
29 26 A 44 8 7.5 6 4 41, (2)42, 48, (2)44 
45 41 7 7 10 45 40, (2)41, 42, (2)48, 
29 27 46 41 if 7 (2)44, (2)45 
46 43 6.5 | 7 8 46 38, (3)42, (2)43, 
29 28 43 39 7 7 (3)45 
45 43 8 7 6 47 42, 44, (2)45, (2)46 
46 46 7 7.5 2 48 44, 45 
46 47 7 8.5 2 50 44, 48 
47 47 8 10 1 52 49 
50 43 9 7.5 
29 29 49 45 8.5 | 8 


CONJUGATION IN PARAMECIUM 127 


TABLE 56 


Lot 22; 188 pairs of conjugants about 12 hours after separation. Measurements of A and 
B of each pair, for length to mouth, total length, and breadth. Unit, 4 microns 


| | 
es | ela |e pleeie | fle} eye) 
B Sr eee 5 elt ee 8 | § Biel ett eas 
cS) atts < g Biel ai) lilt eR lees Pie eels = 
a Pac a 5 ga = 5 Ba | @ B 54 che 
2 | a ) \ a a } es | & | «@ } BR & a 
alp\a B/A|B| A B|A|B\A| Bi Ale lA B\AB A) B\A|B, 4\B 
| : if fi ia || | | Viner ie | iri {|(Gaeal | | ‘ i e= 
25 | 24) 41| 39] 9.510 | 32| 30| 54| 57 | | 8.5) 33 | 33 52 | 54 [12 9.5 | 35 | 32) 58) 50 12 {10.5 
27/27/45) 46 10.511 | | 53 48 11.5 10.5, | 53 | 58 10.512.5) 58 | 53 11 11.5 
28 | 26 | 46 | 43 | 9.5) 8.5] 54 | 48 |11.5)10 54 53 10.511 | 5856 11 1 
29 | 27| 46 | 40 8.5) 9 54/49 |10 | 9.21) 34/29 | 54/46 11 | 9 |} 60 | 52 14 12 
29 | 29) 45| 49 10 |10 | 32) 31 | 49 | 49 |10.5) 10) | 57 | 48 |12.9| 9.5) 60.) 53 42.511 
30 | 28 | 48 | 46 11 | 9 | 51 | 50| 9 salle 54 | 52 |11.5]11.5) | 60 54 11 11 
50 | 45 11 | 8.5) 52] 49 |10 | 9.5] 56 | 46/11 | 9 || 35 | 33 | 57 | 55 12 (11 
| | 50] 48 ]11 [10 | 52 | 50 /10.5)10.5) 34 | 31 | 54 | 52 /10.5)10 60 | 55 10.5115 
30 | 30) 52 53 9.511 | 52| 51 11.511 | 54] 52/11 10 | | 61/58 11 | 9.5 
31/26 48/4319 | 9 55) 51 11,510 | 155 48 11.5) 9 62 | 58 12.510 5 
| | 52] 42 |10 | 9 57 | 52 |10 | 9.5 | 95 | 57 12.512 | 35 | 34 52 59 11 A 
BL 27 | 82 49 [12 /12 | 32} 32 | 51 | 52 /10 |10 | 57 | 49 /13.5|10 || | 5955 11 10.5 
31 | 28 | 54) 5010 | 9.5) | 52 | 52/12 11.5) | 57 | 52 |11.5/10.5]/ — | 59 | 57/11/11 
|__| 49 | 48 |10.5) 9 ] 53] 52 |t0 jlo | 58 | 52 10.5 10.5 | 60} 55 12 1 
31 | 29 | 51 | 48 110 | 9 53) 52 |l1 |9 | 58 | 52/12 |10.5|| 35 | 35 | 55 | 60 \10 13 
j | 51 | 47 {11.5} 9.5) 53) 55 11 13 | | 58} 5312 j11 | | 57 | 56 10.5)10.5 
50 | 47 10.510 | 53 | 56 |11 [12 | 159] 52 (13. [a1 | 60 58 11 11 
31/30 50 | 48 11.511.5 "| 55 | 53 11 |10 | 34 | 32 | 54 | 52 11.5) 9 |) 36 | 31 | 61 | 52 11 10.5 
|53!50/9 |9 | 33! 28) 56/45 !12 [10.5 54 | 52 12 [10 |) 35 | 32 | 59! 54 |11.5111 
| | 51 | 47 {10.5} 9 1 33] 29| 49] 49 jst lat | 56 50 11.510 6l 54 13 11.5 
| |54/53\10 Ho |s3}31|s4|srfio fio | | | 56} 54/12 [0.5] 36 33 | 57 | 55 {13 |t1 
} | 52!) 52:).9° H11 54/51 10 (10 | 57 | 52 10 10 | 58 | 53 12/14 
| 51 46 10.511 | 54) 51 (11 10 | B8)|/63 10.5, 9 | 36 | 34) 57 55 12 10.5 
| | 53} 49 [11 [11.5 55| 52 |12 [12 | 34| 33 | 52 | 56 j10.5|11 58 | 56 12 {11.5 
| | 5250} 9.5)11 | 57 | 54 |10.5)11 54 | 52 [11 | 9.5) 59 | 57 |15 |12 
31 | 31 | 50) 51 | 9.5/10 | 33 | 32 | 50/50 |10 | 9 (55 55 11 10 60 | 58 113.512 
52 | 50 10.511 | | 52| 55 9.5 | 9.5) | 56/53 9.510 | 36 35 58 36 12 2 
52| 54/9 [11 | 53 | 52 |11.5/11 | [56] 57/10 [12 || | 59 | 59 |12 [11.5 
| | 54] 52 |11 [10 | (54/54 11 10.5 | 57 | 52 /10.510 || 36 | 36 57 | 59 112 [13 
32 | 28 | 57 | 48 [11.5) 9 || | 54] 54/11 [11 | 34 | 34 | 56 | 56 [12. 5|12. 5} 59 | 62 |12.5|13 
32 | 29 | 51 44 /10.5) 9 54/55 | 9 |10.5 58 | 58 12 |13 || 38 | 30| 60 | 50/13 {11 
| 52| 47 10 |10 | | 55 | 53 12° |14 | 59 | 58 |12.5/11.5| 38 | 34 | 62 | 55 [10.5] 9.5 
52 49 10.511 |55/53/9 jlo | 35/27/57] 46 \12 |10 | 40/ 31] 65 52 |13.5\11 
53 | 45 10 | 9 |} | 55 | 54 |13 |11 | 35 | 30) 56) 50 11.511 | 
| 53 | 47/| 9.5) 9 58 | 51/11 [10 


vetlt| lelelelolelsleleielmenrels|zielolr1 1/2 I 


nN 


N 


S. JENNINGS 
ao 
Lon ma OM es 
4 a 
aA 
Onn 


HH. 


AMMNAMHOONMANAAAHARMRANDKRM AS 
a al al Sol 
cl 
ol 
nN 
al 


£9 79. 19 09 6S 8g) 2g “od oo v6 ee ZS 1G OS 6F SF Lp fie + oF iF €F cP TF OF 


| 
suolaul ¥ iuyQ “s4ind fos Suaquiaue on} ay} fo yj6uay ur uoynjasio; Uu0y 
-pundas sajfo sinoy gy jnoqn (sdoup aywundas ur) pappry syupbnluog  *(ainz]na ppm) EZ 107 


io) 
ol 46 @IAVL 


CONJUGATION IN PARAMECIUM 129 


TABLE 58 


Lot 23 (wild culture). Correlation in breadth of A with B in pairs 
about 12 hours after separation, the individuals killed in separate 
drops. Unit, 2 microns 


= —————— = * 


16 17 \18 [19 20 21 22 23 |24 25 26 27 28 29 '30 31 32 


16 1 1 2 
17 | 150 1 3 
18 1) 1 She2 7 
19 | 3 3 | 1 1 8:5 
20 IS ele (RGA el,| <4 ye es 115 
21 ne lea 5) 2/1 1 {11 
22 AM ANITON E21 5.2 | 2 2 1 | 28 
23 HOBO By 2.) 0 | 1 1| 14 
24 Aantal 3 | dy 1 18 
25 ABN Tp) 181k | 10 
26 al) eit ez! | % 
27 2) 11] | 3 
28 Sale | 1] 5 
29 | 1) 2 
30 | | | Foie Sih Pen aan Ae 

e = ; | | * eS ee a isl 

1; 1111/3! 210) 8 29 14 22 6 16 TAT 6 134 


rd 
| 


TABLE 59 


Lot 24 (race k). Unseparated pairs, length 
of A by length of B. Unit, 4 microns 


26 27 28 29 30 31 32 33 34 35 36 | 


25 1a 2 
26 | 1 1 | 1 3 
27 2/3/4/3)2 het 15 
28 3/12;5)4)1/38 1 29 
29 8 /11;}6)4/2)1 32 
30 Reda Pea fn ol sees 2|3 | 29 
31 Gree 2 9 
32 2 2 
33 1 1 


1| 3 | 8 [24 \28 |25 113-/12 | 21 3 | 3 [122 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 1 


130. H. S. JENNINGS 


TABLE 60 


Lot 24 (racek). Separated conjugants, both members of 
a pair killed in the same drop of fluid. Length of A 
by length of B 


9 30 31 32 33 34 35 36 387 38 39 40 41 \42 


9 | 
P aks 
| 
| 


rh 
© 


Now oe 


Nr owe Hw 
bo 
eee Se 
RPWONANWWNW hd tv 
. 


CONJUGATION IN PARAMECIUM 131 


TABLE 61 


Lot 24 (race k). Separated pairs, 
the two members of each pair 
killed in separate drops. Length 
of A by length of B 


34 35 36 37 38 39 40 41 


| 


awe 
wow w : 
RPNwWe 
— 
NAN RS oon 


— * = 
2/64/14 14/ 9/3) 3 [6s 
| 


TABLE 62 


Lot 24 (race k). Unseparated 
pairs; length to mouth in A by 
same in B 


16 |17 |18 |19 20 '21 |22 


1o}-3 | 2] 1 | 6 
16 | 2 |10 | 6 1 19 
17 10 17 | 4/3 34 
18 27 15 16) 1} 1) 50 
19 3) 2/1 11 
20 ibe id 2 


Length Before Mouth. 


H.’S. JENNINGS 


TABLE 63 


Lot 24 (race k). Conjugants about 12 
hours after separation; length before 


| | | | 7 { 
jt? 18 119 20 21 22 23 24 25 


ai aha 1 
15 | 

16 1 1 
17 lima Pe 1 
18 1|2 3 
19. 3 1 ale (ped fens 
20 9/4/5)/5/1/)11)18 
21 | 11 33) 9}1 54 
ap) 17 (13'|-8.} 11.39 
23 4|5)1110 
24 i eI 


1 1,6 18 56 31 17. 4 184 


TABLE 64 


Lot 24 (race k). Unseparated conju- 
gants; length behind mouth by length 
in front of mouth, in the same in- 
dividual 


Length Behind Mouth. 


8 | 9 10 [11 |12 |13 /14 |15 |16 


Eg a i ees 6 
16, }) , |S, L039) Nae 225. 
17 |} A |S 25s 2 56 
18 2/12 25 40 /16/4/2! (101 
19 4/6 |15|8|4/1 38 
20 | 1 EST 2} | 15 
21 a) (te 3 
22 1 1 


| 1| 4 [23 56 95 |45 |13 | 6 1 (244 


Length Before Mouth 


TABLE 65 


Lot 24 (race k). Conjugants about 12 hours after 
separation; length behind mouth with length before 
mouth in the same individual 

Length Behind Mouth 


6|7|8 9 10 11 12 13 14 15 16 17 18 


TU SS ality | 1 1 
15 
16} 1 | i 
17 |1 1 2 
18 }2) 1) 3 
19 2 SS aN ta 6 
20 | | 4 Bye G. [Gs fa | 24 
21 | 2 /18 j21 22 | 9 | 72 
22 | 1| {15 30 129 |15 |} 4) 1) 95 
23 | 2 | 6 13 |16 4 41 
24 | 1| 1/2\10/4)1 19 
25 Helles 4 
1 | (1:13 43 67 83 47 8 5 268 
TABLE 66 


Lot 24 (racek). Conjugants about 12 hours 
after separation; breadth of A by breadth 
of B. Unit, 2 microns 


25 26 27 28 | 


19 20 21 22 23 24 


1441] | 1 
15 | | | 
16 

Ta\ete| he i 
TSA Mn et ah LGM ley ha 1 
1OM S| | Pag tae Po 1 
20) |2)1)7)/2)/2)1)83 18 
21 11/6 4) 11 
22 20 17|2|6 54 
23 1/15|1)4/1/ 2) 24 
24 4|3|6 ml ales 
25 Tele 1) 4 
26 ie 1 
27 
| Telimel 


134 


H. 8. JENNINGS 


TABLE 67 


Lot 25 (race k; aurelia). Correlation table for lengths of 
conjugants about 12 hours after separation, the two mem- 
bers of any pair killed together, in the same drop 


| | { 


| ) | 
27 \28 29 30 31 32 33 34 35 36 37 38 39 40 41 


159 bate fsa | | 3 
26 | 1 1 
27 1 1 2 
28 1 2 3 
29 ie) a peeiee | al | 5 
30 | Bele | 3) | 2 l 12 
31 1) age Jee fe es Ws gs 13 
32 | | P| 2) 4)-24).34 as) aide a 17 
33 | | A} 2) 2°) 1 earl ga aie 14 
34 | Bt | 412m | | 10 
35 | 29. 1/5 
36 | | 1h Len aL | 2 
1}1/3]1) 3] 8 |15 14] 8 13 to) 5 | 2/2) 1 | 87 


TABLE 68 


Lot 25 (race k; aurelia). Correlation table for lengths of con- 
jugants about 12 hours after separation, the two members of 
any pair killed in separate drops 


27 28 29 30 31 32/33 34 35 [36 37 38 39 40 41 42 

25 | 1 1 |) Aiea 3 
26 | | | 1 1 
ia Wat tr ay | 2 
28 nen | 2 
29 1] |2 | 3 
30 | ee ta a Pe 3 2 | 1 10 
31 | 123} 0 ayes | 1 1 12 
32 [2)7)riajair)eya 21 
33 | Ala) 32) | 1 14 
34 | 1 512,144 10 
35 (tal eal 1 7 
36 ee yab es 1 4 
37 | 1 1 2 
—— ta <= — —'l = < = 

1 | S0een | 4 1|9|8 (13 12 1211 |9 5 | 3 1 | 91 


THE REPRODUCTION OF PARAMAECIUM AURELIA 
IN A ‘CONSTANT’ CULTURE MEDIUM OF 
BEEF EXTRACT 


LORANDE LOSS WOODRUFF anp GEORGE ALFRED BAITSELL 


Sheffield Biological Laboratory, Yale University 
TWO FIGURES 


Previous work! with pedigree cultures of Paramaecium aurelia 
and Paramaecium caudatum has apparently shown that the life 
history of these forms, when bred continuously on infusions of 
hay made up exactly the same from day to day, tends to run in 
a cycle which terminates with the death of the culture. Previous 
work has also shown that Paramaecium aurelia’? may be bred 
indefinitely on a culture medium which is frequently varied. 

In view of these results the following question suggests itself: 
Is the longevity of Paramaecium on a ‘varied environment’ 
dependent upon intrinsic stimuli from the frequent changes of 
the medium, or is a ‘constant’ medium of hay infusion unfavor- 
able because it lacks some elements which are essential for the 
continued existence of this protozoén? 

To test this point it is necessary to find, if possible, a suitable 
‘constant’ culture medium which contains all the elements which 
the organism demands, and to determine its effect on the vitality 
of Paramaecium when subjected to it for a considerable length 
of time. If such a suitable medium is secured on which para- 
maecia will live indefinitely, it is apparent that the possible con- 
tinual daily stimulation afforded by ‘varied’ culture media is 


1 Calkins: Jour. Exp. Zodl., vol. 1, no. 3, 1904. Woodruff: Biol. Bull., vol. 
17, no. 4, 1909. 

2L. L. Woodruff: Two thousand generations of Paramaecium. Archiv fir 
Protistenkunde, Bd. 21, 3, 1911. 


135 


136 LORANDE LOSS WOODRUFF AND GEORGE ALFRED BAITSELL 


not the crucial factor in the determination of the longevity of 
paramaecia cultures. Further, and aside from this interesting 
theoretical consideration, such a favorable ‘constant’ culture 
medium would be valuable for breeding paramaecia in various 
lines of experimental work, since it is clear from many investi- 
gations that the reactions of paramaecia to various reagents, etc., 
are greatly modified by their past and present environment.® 

In the present paper are briefly outlined the results which have 
been secured, thus far, in an effort to answer the question sug- 
gested above, and to provide a suitable culture medium which 
investigators can employ in breeding cultures of this organism. 

The animals used in this study were from the pedigree culture 
of Paramaecium aurelia which one of us* has had under daily 
observation for more than four years, and which has attained, 
up to the present time (May 1, 1911), 2370 generations under 
the conditions of a ‘varied environment,’ without conjugation 
or artificial stimulation. 

The favorable results secured first by Calkins® with strong solu- 
tions of beef extract as a temporary stimulant for his degenerating 
cultures of Paramaecium caudatum in infusions of hay, and later 
by Woodruff* with Oxytricha fallax under similar conditions, 
suggested the use of a weak extract of beef as a ‘constant’ cul- 
ture medium. Further, beef extract should afford all the elements 
required for the continued life of protoplasm. The results of 
chemical analyses have shown that Liebig’s extract of beef is 


3 For example, Greeley (Biol. Bull., vol. 6, 1904, p. 1), in a study of the effect 
of various chemicals on the protoplasm of Paramaecium, wrote: ‘‘Maximal dilu- 
tions can only be approximate, as the action of identical solutions is not the same 
on paramaecia from different cultures, because no two are exactly alike in respect 
to chemical composition and osmotic pressure;’’? and Miss Towle in a similar study 
(Amer. Jour. Physiol., vol. 11, no. 2, 1904, p. 235) said: ‘‘The first step toward 
a clearing of the haze that envelops the subject will be found, I believe, when an 
effort is made to unify the conditions under which different investigators are 
working.” 

4 Woodruff: Loc. cit. 

§ Calkins: (Archiv fir Entwick.-Mechan., Bd. 15, 1,1902). ‘‘The lean beef was 
boiled in tap-water for fifteen minutes and allowed to stand until cool. The clear 
fluid was then used without dilution.” 

6 Woodruff: Jour. Exp. Zo6l., vol. 2, no. 4, 1905. 


REPRODUCTION OF PARAMAECIUM IN BEEF EXTRACT 137 


remarkably constant in composition, and therefore this standard 
preparation, which is available for all investigators, was used 
as the basis of our culture medium. 

Having decided on Liebig’s extract of beef, it was necessary 
to make a series of experiments to determine the strength of 
solution which was most favorable for Paramaecium. The 
solutions were made by weighing out one gram of the extract 
and diluting this with varying amounts of distilled water. The 
different concentrations of beef extract showed that a solution 
of approximately 0.025 percent gave the best results. Accordingly 
a quantity of this solution was made up which was sufficient to 
provide culture medium for the organisms for a period of seven 
months. This length of time was decided upon for this work 
as the final results of Calkins’ experiments led him to conclude 
that the cycle of Paramaecium caudatum, in a constant environ- 
ment of hay infusion, was not of more than six months duration.? 
The medium when made was put into over one hundred test 
tubes, and these were plugged with cotton and sterilized. The 
solution in the various tubes remained sterile until it was used, 
and the inoculation of the medium with bacteria which were 
transferred with the paramaccia afforded an ample supply of 
food for the animals. 

The regular experiment was begun on October 1, 1910, by the 
isolation of a specimen from each of the four lines of the pedigree 
culture I of Paramaecium aurelia at the 2012th generation. 
Each of the organisms was placed on a depression slide in five 
drops of the beef extract solution, and in this manner was started 
a culture designated Paramaecium IB. This culture was con- 
tinued by isolating an organism every day from each of the four 
lines of the culture, and placing it in fresh culture medium on a 
sterile depression slide. The number of divisions during the 
previous twenty-four hours was recorded at the time of isolation. 
From this record the average daily rate of division of the four lines 


7It may be noted that animals from this same pedigree culture of P. aurelia 
were bred on a ‘constant’ medium of hay infusion from February to June, 1909, 
and died out after 107 days subjection to this condition. (Cf. Biol. Bull., vol. 
17, no. 4, 1909, fig. 4). 


138 LORANDE LOSS WOODRUFF AND GEORGE ALFRED BAITSELL 


Ott Hov. Det. Jan. Feb. Mar Apr. 


Fig. 1 Graph showing the rate of division of Paramaecium aurelia, Culture I 
and Culture IB, from October 1, 1910, through April 29, 1911. The ordinates 
represent the average daily rate of division of the four lines of the respective 
cultures, again averaged for ten day periods. Culture I = ; Culture IB 


of the culture, again averaged for five- and ten-day periods, was 
computed, and the result is graphically shown in figs. 1 and 2. 

The original pedigree culture on a varied environment was, of 
course, continued, and served as a control for the culture on beef 
extract, since the method of carrying on the two cultures was 
identical, except that the medium used was not the same in each 
case. The original pedigree culture was bred on infusions of 
grass, hay, pond weeds, etc., made up with water from various 
sources. The infusions were boiled before being used to prevent 
the introduction of ‘wild’ paramaecia into the pure lines.® 

The various preparations were kept in moist chambers to 
prevent evaporation, and the temperature of the air in these 
chambers was recorded by a maximum and minimum registering 
thermometer. Obviously this method of recording the tempera- 
ture gives only the extremes to which the cultures were subjected, 


> For further details of the method employed, cf., Woodruff, loc. cit. 


REPRODUCTION OF PARAMAECIUM IN BEEF EXTRACT 139 


and does not take into account the length of time during which 
any particular temperature was maintained. However, the 
method is sufficiently exact for the problem at hand. Studies 
on the relation of temperature to the ‘rhythms’ in the rate of 
reproduction of paramaecia are now in progress. During the 
first three months of the work, culture I (‘varied’ culture medium) 
and culture IB (beef) were kept in different rooms, and therefore 
during this time the cultures were subjected to different temper- 
atures. From January 1 to the present time, both cultures were 
kept in the same place and consequently each was subjected to 
the same temperature. 

There were, then, two pedigree cultures of Paramaecium, each 
comprising four lines, being conducted simultaneously. One 
of these had been bred on a ‘varied’ culture medium for forty- 
one months, and was continued under the same conditions dur- 
ing the following seven months, 7.e., to the present time. ‘The 
other culture, isolated line by line from the first culture, was car- 
ried on for seven months (to date) on a ‘constant environment’ 
of beef extract. The chemical composition of this medium was 
identical from day to day as it was all made up and sterilized 
at the same time. The only variation, therefore, in the medium 
used for these organisms on beef extract was the fluctuations in 
the bacterial flora due to infections from the air, and slight vari- 
ations in the multiplication of the bacteria due to temperature 
changes. This, however, was so small that it is negligible from 
the standpoint of these experiments. 

A study of figs. 1 and 2 gives a clear idea of the comparative 
rate of division of the two cultures, and shows that, at the end of 
the seven months work, the rate of division, and therefore presum- 
ably the vitality, of the two series of animals is practically the 
same. Neither of the cultures shows any indications of loss of 
vigor, and the rate of division of each at the end of the experi- 
ment is practically the same as at the beginning—such fluctua- 
tions as have occurred being merely ‘rhythms’. 

During the first three months of the work, the rate of division 
of the beef series was considerably lower than that of the other 
series, but this is obviously explained by the considerably higher 


LORANDE LOSS WOODRUFF AND GEORGE ALFRED BAITSELL 


140 


aunyyng ! =] onyng « 
IO} podvlaae urlese ‘saangpno aaryo 
‘poqoofqns a1aM Samnjpnd at 


Jdy Ie 


{OJ9Q WOA] ‘Gy aIngyny pue JT ainypny 


ad sutpuodsas109 


YoIyM OF vinjetodur 


424 


ayy 10 


t 9G} jo soul] mmoj oq} jo UOISTAIpP JO 94¥1 


9} 99R. 


‘uer 


J (q,.) 2npRtedure, ssvi9oAv ay} OSTR puv 
ASLIOAB OY} JUK 


INAV VY} (@BAOGV) oOs[e pue ‘TIGL ‘6% [Ady | 


‘OI6T 
GSI 


a! 


REPRODUCTION OF PARAMAECIUM IN BEEF EXTRACT 141 


temperature to which the latter was subjected during this period 
(cf. temperature curves in fig. 2). During this period of three 
months, culture I advanced from the 2012th to the 2188th gen- 
eration, or 176 generations, whereas culture IB advanced from 
the 2012th to the 2120th generation, or 108 generations. A 
comparison of the rate of division from January 1 to April 29, 
1911, when both cultures were subjected to identical temperature 
conditions, shows that culture I advanced from the 2188th to the 
2365th generation, or 177 generations, while culture IB advanced 
from the 2120th to the 2287th generation, or 167 generations. 
Therefore the net vatiation in the number of generations attained 
by the two cultures during the last four months, when under the 
same conditions of temperature, was only ten. Further, the ap- 
pearance and behavior of the paramaecia in the two cultures 
were identical. 

It is evident, then, that the ‘constant’ medium of beef extract 
employed has proved (during the seven months of this experi- 
ment) to be practically as favorable a medium for the reproduc- 
tion of this pedigree culture of Paramaecium aurelia as the ‘varied 
environment’ medium,* and therefore, the conclusion seems jus- 
tified that this culture of Paramaecium can, in all probability, 
be continued indefinitely on this ‘constant’ medium. It there- 
fore appears fair to conclude that it is the ‘composition’ of the 
medium rather than the ‘changes’ in the medium which is con- 
ducive to the unlimited development of this culture without con- 
jugation or artificial stimulation. 

It is not suggested that every culture of Paramaecium would 
have the potential to attain more than two thousand three hun- 
dred generations under the conditions of a ‘varied environment,’ 
nor is it suggested that every culture of Paramaecium would 
thrive for over seven months on a ‘constant’ environment of beef 
extract. ‘‘For undoubtedly there are strong and weak strains 


’ Experiments are now in progress to determine if this culture of Paramaecium 
will develop indefinitely on infusions of hay which are not made up the same from 
day to day, 7.e., if hay infusion, in which there is a slight daily variation, may 
not be substituted for the decidedly varied culture medium which has been used 


during the past four years. . 


142 LORANDE LOSS WOODRUFF AND GEORGE ALFRED BAITSELL 


among Infusoria as among other classes of animals. Again, it 
is possible that the different races of paramaecia which Jennings 
has been able to isolate may have a physiological as well as a 
morphological basis of distinction.’'° But it is believed that these 
experiments clearly show that beef extract (in the concentration 
used) is a suitable environment for the continued reproduction 
of this pedigree culture of Paramaecium, and that beef extract, 
in this or closely similar solutions, will prove to be a favorable 
medium for use in many investigations on the physiology of 
Paramaecium.!! 


10 Woodruff: Biol. Bull., vol. 17, no. 4, September, 1909, p. 303. 

1 After the completion of the seven months experiment (April 29, 1911), a new 
lot of the beef extract medium was made up exactly the same as before and the 
culture has been continued on this without noticeable change in its reproduction to 
date of correcting proof, June 20, 1911. 


MIGRATION OF RETINAL PIGMENT IN THE EYES 
OF BRANCHIPUS GELIDUS 


RUTH B. HOWLAND 


FOUR FIGURES 
INTRODUCTION 


The general subject of pigment migration has received much 
attention during recent years. While it is not the purpose of 
this paper to deal with the movements of melanophore pigments, 
we can not disregard the effect which the study of this problem 
has had upon the study of the migration of retinal pigment. 
There can be no doubt that an intimate relation exists between 
the movements of the two forms of pigment, as there is no doubt 
concerning the morphological similarity existing between pigment 
spots and the eye spots of many of the lower animals. 

The striking changes in coloration exhibited in chameleons 
early led to investigations as to the physical factors involved. 
Conclusive evidence showed that the movements of melano- 
phore pigment was brought about largely by variations in light 
intensity and temperature. The work of Carlton (’03), Parker 
and Staratt (04), and Parker (’06) is sufficiently well known to 
need only brief mention. Carlton found, in the case of Anolis, 
the outward migration of pigment in the pigment spots to be 
directly dependent on the action of the sympathetic nervous 
system, and the inward migration simply a return to the resting 
state. 

In chameleons the migration was found to be under the control 
of the spinal nerves. 

As these observations have been made only upon the marine 
crustacea, it was suggested to me by Professor C. W. Hargitt 
that a series of comparative studies upon a fresh water crustacean 
might be of some value. The occurrence of the Phyllopod, 
Branchipus, in a pond near Syracuse, and its especially promi- 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. LI, No. 2 
AuGust, 1911 


143 


144 RUTH B. HOWLAND 


nent eyes made it a very favorable object for the investigation. 
In the spring of 1908 these crustacea could not be found at Syra- 
cuse, but material for experiments was obtained from Jordan, 
a small town seventeen miles west, and this supply was supple- 
mented by material from Potsdam and the West, and fresh speci- 
mens from Syracuse in the spring of 1909. Experiments were 
made and subsequent study conducted with a view to determining 
the effect of varied heat and light on the migration of retinal 
pigment in these eyes. 


HISTOLOGICAL METHODS 


The methods of preparation which gave the best results are, 
in brief, as follows: 

1. For killing and fixing, hot picro-acetie acid was used. The 
animals were dipped in and held for a few seconds, then trans- 
ferred to 80 per cent alcohol. Formalin proved unsatisfactory 
for fixation, and also rendered the material hard to stain in some 
eases. Corrosive sublimate fixation was also inferior to picro- 
acetic. 

2. a. Ehrlich’s haematoxylin in toto or section gave the best 
results of any of the stains tried. If used for staining in toto, a 
a period of at least sixteen hours was necessary. When the sec- 
tions were stained upon the slide, a minimum staining of ten 
minutes was allowed. The one stain was sufficient to show all 
features clearly, but eosin was occasionally used as a counter 
stain. . 

b. Iron-haematoxylin with counter stain of Bordeaux red also 
gave satisfactory results. The slides were treated as follows: left 
in iron-haematoxylin, a minimum of fifteen minutes; plunged 
into a FeCl, solution until sufficiently differentiated; washed in 
water; stained in aqueous solution of Bordeaux red for three min- 
utes; dehydrated; cleared and mounted. 

c. Borax carmine failed to give the desired differentiation. 

d. Haidenhain’s triple stain was temporarily satisfactory, but 
was not permanent. 

3. Xylol proved to be the best clearing agent. Cedar oil 
was not so good. Material 7m loto required an hour at the mini- 


EYES OF BRANCHIPUS GELIDUS 145 


mum for clearing. Infiltration of at least one and one-half 
hours was necessary. Paraffin melting at 62° C. permitted sec- 
tioning at from 3 to 5 microns. 

4. To see clearly the nuclei of the retinular cells which in the 
normal eye are obscured by large masses of pigment, the follow- 
ing method of depigmentation was used: 

Two or three drops of hydrochloric acid were poured over a few 
crystals of potassium chlorate. When the green color of the 
evolving chlorine appeared, a few cubic centimeters of 70 per 
cent aleohol were added. The eyes were allowed to stand in this 
bleaching fluid over night and subsequent staining revealed the 
retinular nuclei clearly. 


NORMAL STRUCTURE OF THE EYE 


Before entering upon a discussion of the experimental work, 
a brief description of the normal eye is essential. 

The eye of this form is one of the simplest of compound crus- 
tacean eyes. It is stalked, and of such large size as to be a very 
prominent feature. The structural elements are fundamentally 
the same as those of other crustacean eyes, the unit being the 
ommatidium. 

The histology of the eye of Branchipus vernalis is described 
in brief by Parker (’91) and others, and a more detailed account 
of the eye of Branchipus stagnalis is given by Nowikoff (’05). 

The outer surface of the eye is covered with a structureless 
cuticula, presumably a secretion of the underlying hypodermal 
cells. In Branchipus vernalis it is obviously facetted, having 
the form of concavo-convex elevations. 

According to Patten, the hypodermal cells in B. grubei are 
of indefinite arrangement, but Claus and Nowikoff find in B. 
stagnalis a regular arrangement which the latter shows diagram- 
matically in the following way: the hypodermis covers the distal 
end of each ommatidium as a cuticular cap made up of six equal 
cells extending radially from the center to the circumference of 
the circle. The nuclei are oblate, and lie upon one side. The 
intervening space is filled by two cells with round nuclei. These 
nuclei are extremely large and extend at an acute angle into that 


146 RUTH B. HOWLAND 


part of the cell lying between the cone cells. This arrangement 
as well as the facetted condition of the cuticula (both disputed 
points) have been observed in my preparations of the species 
found here. 

The cone cells are four in number, of equal size and regular 
form. The crystalline cones are ellipsoidal and their mass in 
comparison with that of the cone cells is large. The nuclei of 
the cone cells were not found by Claus in the adult, and only 
indistinctly seen in the larvae. Nowikoff, however, states that 
they are well defined in his preparations, lie close under the erys- 
talline cones at the promixal ends and are sphere shaped, while 
Patten finds these nuclei lying over the distal end of the cells 
like a cap. 

In regard to the exact location of these nuclei, it seems to me 
that there has been a great deal of confusion arising from incom- 
plete knowledge of the histology of the hypodermal cells. Patten, 
as I have previously stated, regarded the hypodermis as consist- 
ing of a “layer of indefinitely arranged cells’ and the four cone 
cell nuclei as forming a cap over the crystalline cones. These 
nuclei, described as forming a cap over the crystalline cones, 
might be those of the hypodermal cells in their characteristic 
radial arrangement as shown by Nowikoff. If this were the case, 
Patten’s description would agree with B. stagnalis as to the loca- 
tion, but not as to the number of the hypodermal nuclei. Now- 
ikoff, however, goes a step further in identifying the cone cell 
nuclei. 

I regret to have to say that in my own preparations I have been 
unable to confirm Nowikoff’s or Patten’s statement as to the 
location of these nuclei, though I have directed special attention 
to this point. It is a most peculiar circumstance that cells of 
this size and importance should fail to show clearly a definite 
nuclear structure. 

A peculiar condition in one series may be worth considering 
in connection with this point. This series, cut at 5 microns, 
shows, under the one-sixth inch objective, structures which even 
upon critical examination appear as nuclei of the cone cells at 
the level of the distal ends of the retinular cells. A slight enlarge- 


EYES OF BRANCHIPUS GELIDUS 147 


ment of the rhabdom at this place furthers the impression that 
we are looking at nuclear structures. Similar structures, though 
less evident, can be seen scattered elsewhere at various levels. 
They are of comparatively large size and appear round and very 
similar to the nuclei pictured by Nowikoff in B. stagnalis. Nu- 
clear structures, if present in this level, would easily escape even 
careful observation; for unless the section be extraordinarily 
thin and exactly in the plane parallel to the rhabdom, the ends of 
the pigmented retinular cells would obscure them. 

However, a closer study under a one-twelfth inch oil immersion 
lens shows these structures rather as reticulated fibrils of the 
cytoplasm of the cone cells with scattered pigment granules 
forming pseudo-nucleoli. Had these structures been constant 
in all of my slides, I would have been assured of the fact that they 
were nuclei, but their searcity, the reticulated structure of the 
cytoplasm of these cells, and their appearance under the oil im- 
mersion lens forces the conclusion that they are artifacts. 

The rhabdom is slender and in continuation with the cone cells 
(Nowikoff) ; it extends approximately three-fourths of the entire 
length of the ommatidium terminated proximally by the base- 
ment membrane, and is surrounded its entire length by the five 
retinular cells. The retinular cells contain a large quantity of 
pigment granules. Their cytoplasm is fibrillar, extending below 
the basement membrane as fibrils. The nuclei are oval and lie 
in the larger distal ends of the cells. 

The nucleated basement membrane separates the retinal and 
nerve-fiber regions. It is perforated where the retinal cells pene- 
trate it (Nowikoff, ’05). 

Blood corpuscles are found in varying numbers between the 
ommatidia, and especially upon the ventral side of the eyes. 
This was particularly noticeable in a large number of prepara- 
tions, and is doubtless due to the close proximity of a large ventral 
sinus. 

A comparison of sections of the eyes of B. vernalis and B. 
gelidus with the plates and descriptions of B. stagnalis given by 
Nowikoff show no differences of structure. The nuclei of the 
cone cells I was, of course, unable to compare. 


148 RUTH B. HOWLAND 


An article by Dietrich (09) came to my notice after the com- 
pletion of the above discussion. It comprises an extensive 
description of the facetted eyes of diptera, and contains points of 
general interest in their bearing upon the problem under dis- 
cussion. Mention is made of the fact that deep-sea crustacea 
have a greater number of elements in each ommatidium—pos- 
sessing as a rule seven—than do the crustacea of a more pelagic 
habit, whose ommatidial elements have been reduced to five. 
An explanation of this is offered by the fact that the difference 
in intensity of light which these two extremes receive has possibly 
led to the reduction as an adaptive feature. 

Another feature which calls forth comparison with the structure 
of the phyllopod eye is the location of the nuclei of the cone cells. 
In some species of diptera, these are found directly beneath the 
pseudocones, while in others they le further proximally. Their 
appearance in the latter case closely resembles the structures in 
my preparations already described, which closer study revealed 
as artifacts. 


EXPERIMENTS 


An account of the experiments tried may perhaps best be 
given in tabulated form: 


SPECIES HOW KILLED LIGHT OR DARK TIME TEMP. 
Deg. C 
Bip weliduse av3a2 4: Corrosive sublimate Dark 9 hrs. 525 
B. gelidus.. . : Corrosive sublimate Dark 8hrs. | 20.0 
3. gelidus. ...... Hot picro-acetic Dark 43> hrs. 24.0 
B. gelidus ... Corrosive sublimate — Diffuse light S8hrs. 17.0 
B. gelidus .. Corrosive sublimate Diffuse light 9hrs. | 55.0 
B. gelidus : Hot picro-acetic Sunlight Killed at 19.0 
six o'clock 
B. gelidus ae Hot picro-acetic Sunlight 12hrs. } 21.0 
B. gelidus. . . Hot picro-acetic Dark 16hrs. | 21.0 
B. vernalis Hot formalin Dark 5hrs. i 
B. vernalis Picro-acetic Dark 5hrs. 2? 
B. gelidus. . Hot picro-acetic Dark 30 min. 19.0 
B. gelidus..... Hot picro-acetic Dark 15-20 min. | 19.0 
B. gelidus........... Hot ipero-acetic Dark 2 hrs. 5.0 


B. gelidus...........| Hot picro-acetic Dark 2hrs. | 21.0 


EYES OF BRANCHIPUS GELIDUS 149 


The animals were taken from the pond and experiments made 
upon them without delay, for the character of all reactions was 
noticeably modified by keeping them for twenty-four hours in 
an aquarium. A loss of vitality, fading out of color, etc., became 
evident within a day unless the aquarium was kept at very low 
temperature, with a supply of soft mud in the bottom. 

In all the experiments, the largest, most active specimens of 
both sexes were chosen. The experiments in light and dark were 
conducted as follows: 

a. A few animals were placed in the direct sunlight, after allow- 
ing them to stand in a warm room until the water had been grad- 
ually raised to room temperature. In the early part of the 
season this method was found to be impossible. The maximum 
temperature at which they will live changes with the advancing 
season, and while early in the season the room temperature of 
21° C. will kill them almost immediately, later the maximum is 
raised to from 26° C. to 29° C. The range of temperature be- 
tween the optimum and the maximum for these animals, however, 
does not seem to vary. With a higher optimum, a higher maxi- 
mum temperature ensues, but the range between these two ap- 
pears to have a certain constancy of 5-8° C. 

b. On a cloudy day, an aquarium with a few specimens was set 
out of doors and allowed to stand for nine hours. The average 
temperature for that time was 5.5° C. and the few degrees of 
variation above or below were not considered of sufficient value 
to appreciably modify the result. The same experiment was 
twice repeated with aquaria upon a table in the center of the room, 
out of the direct rays of the sun, at temperatures of 17° C. and 
20° C. for periods of eight hours. 

c. In conducting the experiments in the dark, the greatest care 
was taken to exclude every means of access of light rays. The 
animals were placed in a tall glass dish, which was wrapped in 
light-proof paper. This was then either placed in a tin box and 
covered tightly, or left in the photographic dark room for the 
required length of time. The paper was then removed, and a 
flash of light thrown into the jar to assure the fact that all were 


150 RUTH B. HOWLAND 


alive. The water was at once drained off and the specimens 
thrown into the killing fluid, heated to 60° C. 

In preparing the eyes for sectioning it was found that the lia- 
bility to injury from handling was greatly reduced by dehydrat- 
ing and clearing the entire animal before separating the eyes from 
the body. After infiltration was complete, the eyes were removed 
from the body and imbedded with special attention as to orienta- 
tion for transverse or longitudinal sections as desired. 


CONDITIONS FOUND 


A discussion of the conditions found in the material sectioned 
falls naturally under three heads: first, the effect of ight; second, 
the effect of heat variation; and third, the relation of pigment 
migration to phototropism. 

1. In these examinations care was taken to compare those 
sections cut at equal thickness, and to select those parts of the 
sections (long) where the plane of cutting was parallel to the long 
axis of the rhabdom. 

a. In eyes which have been killed after exposure to the sun for 
a period of one and three-fourths hours the pigment is in the fol- 
lowing condition: In the distal parts of the retinular cells the 
granules are accumulated in great quantities, so that it is impos- 
sible to distinguish the nuclei of these cells, which are located here. 
The entire length of the rhabdom is closely protected with a 
heavy layer of pigment, and here the lateral portions of the cells 
are much clearer as compared with the distal part. In the prox- 
imal third of the cells there are proportionately smaller quanti- 
ties. Pigment is also heavily deposited below the basement 
membrane, and for a short distance along the nerve fibers. In 
eyes exposed to sun for longer periods, the conditions are identical. 
(Figs. 1 and 3.) 

b. In diffuse light for eight hours, the pigment granules are 
still heavily deposited in the distal ends of the cells, and along 
the rhabdom. ‘The lateral portions of the retinular cells, as in 
the case of sunlight, are much clearer than the densely pigmented 
region close to the rhabdom. The direct sunlight thus seems to 
produce no appreciably greater response of pigment than the 
diffuse light of a partly cloudy day. 


EYES OF BRANCHIPUS GELIDUS 151 


Fig. 1 Pigment conditions after exposure to sun until 6 p.m. on a sunny day 
Rhabdoms covered with pigment. 

Fig.2 Pigment conditions after remaining in dark four and one-half hour 
Readjustment of pigment through a lateral migration leaves rhabdoms exposed 

Fig.3 Transverse section of ommatidia after exposure to sun for five hours 
Conditions as 1n fig. 1 (Oil immersion, camera lucida 

Fie. 4 Transverse sec.ion of ommatidia after remaining in dark four and one- 


half hours. Conditions as in fig. 2 One-twelfth oil immersion, camera lucida 


152 RUTH B. HOWLAND 


c. Eyes sectioned after remaining in the dark fifteen minutes, 
twenty minutes, and a half hour, show no appreciable movement 
or migration of pigment granules. The conditions do not vary 
from those in diffuse light to any appreciable degree. Pigment 
is heavily deposited for a short distance down the nerve fibers 
below the basement membrane. 

d. In the eyes which have been in the dark for two hours, the 
pigment has begun its lateral migration and forms an intermediate 
stage between the two extremes a and e. 

e. The eyes which have been in the dark four and a half hours, 
show the pigment closely packed in the distal ends. The granules 
are not packed close to the rhabdom as in sun and diffuse light, 
but are scattered laterally in the cytoplasm by a lateral migration 
or a readjustment. The outline of the rhabdoms is thus not so 
clearly defined by a heavy boundary of pigment granules, but 
their surfaces are more exposed. ‘The entrance of light rays would 
result in a more intense stimulus than under the previously 
described conditions. A dense accumulation of pigment occurs 
below the basement membrane and proximally down the nerve 
fibers (fig. 2 and fig. 4). 

f. In eyes left in the dark from eight to sixteen hours, these 
variations in pigment location become less and less evident, and 
in general the eye assumes the appearance of those killed in direct 
sunlight. The distal pigment is closely packed around the rhab- 
dom and the retinular nuclei, the pigment in the proximal part 
lying close to the rhabdom. Eyes which have been in the dark 
for eight hours or more were thus not considered in drawing con- 
clusions as to the effect of light and dark. 

The adaptive movement of pigment granules in these eyes would 
thus appear rather as a rearrangement or readjustment of the 
pigment granules in the retinular cells than as a pronounced 
migration proximally in the dark, and distally in the light. A 
proximal migration in the dark would result necessarily in an 
accumulation of pigment below the basement membrane, which 
would increase gradually from the light conditions to the com- 
pletion of the migration. The pigment masses beneath the 
membrane would thus vary in a graded series, a condition which 


EYES OF BRANCHIPUS GELIDUS 153 


does not oecur in my preparations. The pigment deposited be- 
neath the membrane is quite as dense in sunlight as in dark. 
The distal pigment in all my preparations is densely packed around 
the rhabdom and the retinular nuclei, and shows no sign of move- 
ment due to light or temperature changes. 

The protective function of the pigment is accomplished in a 
different way. A lateral movement of the granules exposes the 
rhabdom in the dark, while an opposite movement in the light 
brings them close against the rhabdom. Since these eyes have 
no accessory pigment cells, which in other crustacean eyes serve 
as a reflecting apparatus in the dark, the cytoplasm of the retin- 
ular cells must perform this function. 

So, unlike the eyes of Cambarus and Gammarus, the proximal 
and distal migrations of pigment granules are not found in B. 
gelidus, but are replaced by a lateral migration, outward in the 
dark, centrally toward the rhabdom in the light, while the pro- 
tective function of the pigment remains the same. 

2. The temperature changes, so far as I have been able to note, 
produced no appreciable effect on the migration of the pigment. 
High temperatures were of course impracticable in experimental 
work, causing almost immediate death. In eyes exposed to dif- 
fuse light at 5.5° C. and 17° C. no difference could be noted in the 
distribution of the pigment. 

3. Experiments were made with a view to determine the de- 
gree of phototactic response, and also its possible bearing on the 
movement of the pigment. 

a. A dozen animals, which had just been brought in from the 
pond, were put into an aquarium in a large amount of pond water 
at 21°C. Of this number, five were young specimens, the others 
adults. Light from a 16-candle power lamp was thrown into the 
aquarium from above with the following results: 

11 were strongly positive. 1 female was indifferent. 

The two youngest specimens made frantic efforts to approach 
nearer the light beating their heads against the sides of the aqua- 
rium. The light was then placed below the jar. Nine responded 
at once by following the light. One male, one female, and one 
young specimen were indifferent for a period of three minutes, 


154 RUTH B. HOWLAND 


when they too went to the bottom. If the light were held below 
the jar, the tendency to orient themselves with their long axes 
perpendicular to the source of light was stronger than the habit 
of swimming with the appendages upward, and they stood upon 
their heads making futile efforts to reach the light. 

b. A larger number of animals was then put into the jar and 
the same experiments repeated. The younger ones showed a 
much quicker and more positive response than the adults. The 
depth of the water effected no change in the character of the re- 
sponse. Special attention was given to the occurrence of the 
animals in their natural habitat, to determine whether the re- 
sponses given in the laboratory were merely artificial or whether, 
even in the natural environment, the direction of the light rays 
had a noticeable effect on the location of these crustacea. On a 
bright, sunny afternoon, when the water was 11° C. a much 
larger number of specimens was found swimming on the west 
side of the pond than on the east side. They were not, however, 
very active, on account of the low temperature, but for the most 
part were swimming slowly along the bottom about an inch above 
the mud. For some time, the direction of the wind was thought 
to have a marked influence on the distribution in the pond, but 
the wind on this day was blowing from the northwest, and the 
animals must have gone almost directly against the wind in order 
to appear in such large numbers on the western side of the pond. 

c. The temperature of the water in one jar was gradually raised 
by the addition of small amounts of hot water. Pond water 
was heated for this purpose to guard against the addition of any 
chemical which might influence the response. When the water 
reached 24.9° C. to 25.5° C., the light was placed above the jar. 
The animals responded positively as in b. They became very 
active, and the branchipeds greatly increased their rate of motion. 
As the temperature reached 27° C., convulsive jerkings became 
frequent, jerking the animals almost out of the water when they 
came to the top in response to light. Between 27° C. and 29° 
C. they became less and less active and sank to the bottom of 
the jar. At 31° C. they were all dead. The increase in temper- 
ature does not, therefore, cause a negative response, but merely 


EYES OF BRANCHIPUS GELIDUS 155 


a greater activity which finally leads to spasmodic muscular con- 
traction and death. 

Numerous reports of photomechanical changes in retinal pig- 
ment have appeared referring to amphibia, cephalopoda, and 
arthropoda. Exner (’91) and Parker (’97 and 799) record the 
effect of changes in light intensity on retinal pigment in Palaemon 
and Gammarus. 

These reports have led without exception to the conclusion 
that the same laws of migration which hold true in the case of 
melanophore pigment apply as well to the migration of retinal 
pigment under various degrees of illumination. The uniformity 
of the effect of heat, however, has only recently been acknowl- 
edged. Kiihne (’79) observed that the retinal pigment in frogs’ 
eyes in darkness was withdrawn further proximally in a high than 
inalow temperature. Herzog (’05) confirmed and amplified these 
results, and further stated that while increased heat produced a 
proximal migration, decreased temperature a distal migration 
between 0° and 18° C., above this temperature exactly the reverse 
took place. The question of temperature influence was resumed 
several years later by Congdon (’07) in the case of Arthropod 
eyes. Experiments on Palaemonetes and Cambarus confirmed 
the previously published results of Ktihne and Herzog. Response 
to raised or lowered temperature in these crustacea is much weaker 
than photomechanical response, and probably of no functional 
importance, as the migration due to temperature change occurred 
much above normal conditions. 

As to the physiological importance of retinal migrations due 
to varied light conditions, there is no doubt that the distribution 
of the pigment along the sensitive parts of the eye protects it 
from too intense illumination by the absorption of light rays; 
while on the other hand, the withdrawal of pigment from them 
gives easy access to the non-injurious rays of diffuse light. In 
eyes which possess, in addition to the dark pigment cells, whitish 
accessory pigment cells (Palaemon, according to Congdon, ’07) 
these cells serve to reflect light rays into the rhabdom on with- 
drawal of the pigment in diminished light. 


156 RUTH B. HOWLAND 


In addition to this protective result the migration of retinal 
pigment has a bearing upon the phototropic response of animals 
in which this occurs. Gammarus annulatus shows (Smith, ’05) 
pronounced migration of pigment, and accompanying this a 
marked change in behavior toward the light stimuli. The close 
correlation of these two facts in matter of time, points towards 
the fact that the change from negative phototropism to a marked 
positive response is due to the exposure or protection of the rhab- 
dom by movement of the pigment. 

d. The effect of a moving light was then tried. A light was 
held above the jar until the usual response was given, and the 
number of negative or indifferent specimens counted. Then it 
was moved slowly around the jar. In every case the positive 
response was greater, and more definite with the moving than with 
the stationary light. 

e. A larger number of animals was placed in an aquarium and 
allowed to stand in the dark for an hour at room temperature. 
They were then exposed to light. They remained indifferent for 
a short period before any definite response was noted, but then 
gave a positive response. 

Another aquarium was tightly wrapped in black paper and 
made thoroughly light-proof. This was placed out of doors 
(about 19° C. and allowed to remain for five hours. It was then 
taken to a photographie dark room and opened under a light of 
16-candle power intensity. The majority of the specimens were 
negative for a period of from two to three minutes, but after 
remaining in the light, gradually became positive as under the 
usual conditions. 

The same specimens were at once placed in the dark and left 
for two hours longer. When exposed to the light, the negative 
reaction was even more definite than before.! 


‘In a recent paper on phototropic responses of Branchipus (McGinnis, ’11) 
the statement is made that Branchipus is never negative to light, even after ex- 
posure to darkness, but no record is made of the time necessary to bring about 
the normal response after this exposure. The peculiar ‘‘reversal of geotropic 
response’’ recorded in the same paper (p. 237) may have been influenced by nega- 
tive response to light, the other positive responses to gravity being, as suggested, 
simply the natural falling of bodies. 


EYES OF BRANCHIPUS GELIDUS 157 


The reversal of the response in my experiments is obviously 
related to the movement of the pigment in the eyes. As the 
previous results have shown the condition of the eye at the end 
of five hours to be one in which the sensitive rods are exposed to 
the fullest extent, the sudden entrance of a strong light would 
naturally cause a negative reaction. The fact that in the first 
experiment an exposure to 16-candle power light did not cause 
this negative response is doubtless due to the incomplete migra- 
tion of the pigment in the shorter period of time. 


SUMMARY OF RESULTS 


1. The effect of light and dark on the movement of pigment 
granules in the eye of Branchipus gelidus is in the nature of a 
readjustment rather than a proximal and distal migration. 

2. The distal pigment is not influenced by variation in light 
intensity. 

3. In light, the pigment granules collect closely around the rhab- 
doms, protecting them from too intense stimulation. 

4. In the dark, the granules move laterally and are readjusted so 
that they become more evenly distributed through the eytoplasm 
of the retinular ¢ells. 

5. The time occupied in complete readjustment is between four 
and one-half and five hours. 

6. The cytoplasm of the retinular cells serves as a reflecting 
apparatus in a weak light in the absence of accessory cells. 

7. Changing temperatures have no appreciable effect upon pig- 
ment migrations, higher temperatures causing almost instant 
death. 

8. Branchipus gelidus is positively phototropic. Animals ex- 
posed to light after remaining in the dark five hours were nega- 
tively phototropiec. 


158 RUTH B. HOWLAND 


BIBLIOGRAPHY 


Cariton, F. C. 1903 The color changes in the skin of the so-called Florida 
chameleon, Anolis Carolinensis Cuv. Proce. Amer. Acad. Arts and Sci., 
vol. 39, no. 10, pp. 237-276, 1 pl. 


DAHLGREN AND Kepner 1908 Principles of animal histology. 


Dierricu, WitaeELM 1909 Die Facettenaugen der Dipteren. Zeit. fiir wiss. 
Zool., Bd. 92, Heft 3. 

Hay, O. P. ann W. P. 1889 A contribution to the knowledge of the genus Bran- 
chipus. Am. Nat. February, vol. 23, no. 266, pp. 91-95. 


Herzoc,H. 1905 Experimentelle untersuchungen zur Physiologie der Bewegung- 
vorgiinge in der Netzhaut. Arch. f. Anat. u. Physiol., Physiol. Abt. 
Jahrg. Heft 5-6, pp. 413-464, Taf. 5. (Original not available.) 


Ktune, W. 1879 Chemis*he Vorgiinge in der Netzhaut. M. L. Hermann, 
Handbuch der Physiol., Bd. 3, Theil 7, pp. 235-3842. (Original not 
available.) 


McGinnis, Mary O. 1911 Reactions of Branchipus serratus to light, heat and 
gravity. Jour. Exper. Zoél., vol. 10, no. 2, pp. 227-239. 


Nowrkorr, M. 1905 Uber die Augen und die Frontalorgane der Branchiopoden. 
Zeit. fiir wiss. Zodl., Bd. 79, Heft. 3, pp. 432-464, Taf. 2. 

Packarp, A.S. 1878 A monograph of the phyllopod crustacea of North America, 
with remarks on the order phyllocarida. U.S. Geol. and Geog. Survey 
of the territories of Wyoming and Idaho, 1879, pp. 295-514. Part 1, 
39 pl. 


Parker, G. H. 1891 Compound eyes in crustacea. Bull. Mus. Comp. Zodl., 
Harvard Coll., vol. 21, no. 2, pp. 45-140, 10 pl. 
1897 Photomechanical changes in the retinal pigment cells of Palae- 
monetes, and their relation to the central nervous system. Bull. 
Mus. Comp. Zo6él., Harvard Coll., vol. 30, no. 6, pp. 273-300, 1 pl. 
1899 The Photomechanical changes in the retinal pigment of Gam- 
marus. Bull. Mus. Comp. Zo6l., Harvard Coll., vol. 35, no. 6, pp. 141- 
148, 1 pl. 
1906 The influence of light and heat on the movement of the melano- 
phore pigment, especially in lizards. Jour. Exper. Zoél. Vol. 3, no. 
3, pp. 401414. 


Parker, G. H. and SvarattT,S. A. 1904 The effect of heat on the color changes 
in the skin of Anolis Carolinensis Cuy. Proc. Amer. Acad. Arts and 
Sci., vol. 40, no. 10, pp. 455-466. 


Rogers, C. G. 1906 A chameleon-like change in Diemyctylus. Biol. Bull., 
vol. 10, no. 4, March. 


SmitH, Grant 1905 Effects of pigment migrations on phototropism of Gam- 
marus annulatus. Am. Jour. Physiol., vol. 13. 


THE DEVELOPMENT AND FUNCTION OF 
VOLUNTARY AND CARDIAC MUSCLE IN 
EMBRYOS WITHOUT NERVES! 


DAVENPORT HOOKER 
From the Sheffield Biological Laboratory of Yale University 


FIFTEEN FIGURES—ONE PLATE 
INTRODUCTION 


The influence of the nervous system upon development has 
been much studied and discussed, and it is natural, owing to the 
complicated nature of the subject and the close relation of the 
nervous system to the rest of the body, that the views that have 
been held have been diverse and conflicting. 

The literature on this subject up to 1904 has already been fully 
reviewed by Harrison (’04) and Goldstein (’04), and the reader 
is referred to their papers. 

Harrison (’03, ’04), using much earlier stages than Schaper 
(98) had used, removed the entire spinal cord from embryos of 
Rana sylvatica, virescens and palustris before visible differen- 
tiation of the axial musculature had begun and found that the 
muscle tissue subsequently differentiated normally. The irri- 
tiability of the nerveless muscle was not fully tested, however, 
though one of the operated embryos did give response to uni- 
polar electric stimulation, showing its independent. irritability. 
Harrison also found that muscle tissue would develop normally 
when embryos of the frog were kept anaesthetized in acetone- 


1The results detailed in this paper were reported in brief at the May meeting of 
the Society for Experimental Biology and Medicine, 1910, and an abstract appears 
in its proceedings. A more detailed account was reported before the American 
Association of Anatomists, at the Ithaca meeting in 1910. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 2 
159 


160 DAVENPORT HOOKER 


chloroform during their development. In this way, the motor 
nerve impulses normally transmitted to the developing muscle 
were eliminated. 

Goldstein (04) published a comprehensive paper after repeat- 
ing Shaper’s experiments with improved technique. The embryos 
which he used were of the same age as Shaper’s and, even though 
the whole spinal cord and hind-brain were removed, the experi- 
ments are therefore open to the same objection as Shaper’s, viz.: 
that the first nervous connections had been established before 
the operation. Goldstein found that the operated embryos de- 
veloped normally and that their muscle tissue would respond 
to direct stimulation. 

Wintrebert (’03—’05) removed the cord from frog and Ambly- 
stoma embryos at various stages and found that they developed 
normally. He described the presence of contractility in the muscle 
tissue before the nerves reach it. He states that he observed 
not only contraction of the myotomes, but also movement on 
direct stimulation of the fingers of the hind extremity of the frog 
larva, from which extremity all nerve tissue had been removed. 

The most recent work directly bearing on this problem is that 
of Paton in 1907. He concluded that the first movements of 
the embryo take place before there is any true nervous connection 
between the muscle tissue and the central nervous system, and 
used this as an argument in favor of the Hensen theory of nerve 
development. 

The purpose of the experiments detailed in the present paper 
was to determine the following: (1) Does differentiation of muscle 
fibrillae and the establishment of the nervous connection with 
the central nervous system precede or follow the acquisition 
of contractility in the myotomes in the normal embryo? (2) 
Will voluntary muscle which has developed without the influence 
of nerves respond to stimuli and, if so, must the stimuli be applied 
directly or may they be transmitted by non-nervous paths? 
(3) Will cardiac muscle differentiate and function independ- 
ently of nervous control? (4) Will the gross form of the heart, 
developing under such conditions, be normal? 


MUSCLE IN EMBRYOS WITHOUT NERVES 161 


METHODS 


The operations were performed on frog embryos at the stage 
of development immediately following the closure of the neural 
folds (fig. 1). At this time they are from 2.25 to 3.75 mm. 
in length, according to species, and there are absolutely no 
nerve fibers in the central nervous system nor any traces of 
peripheral nerves. The embryos operated upon by Shaper and 
Goldstein were much older and the chances of nerve contamina- 
tion were correspondingly greater. 

The instruments used were: a pair of Noyes’ iridectomy scis- 
sors, an iridectomy knife, forceps and needles, the points of all 
being ground down to a more than hair-like fineness. All oper- 
ations were performed under a Zeiss binocular microscope. 


Fig. 1 Embryo of R. sylvatica, to show the stage of development used in 
the beginning of the experiments. 93. (After Harrison.) 


For the study of the voluntary muscle, it was necessary to 
remove the source of nerve supply to the myotomes. The cord, 
hind-brain and skin of the dorsal region were removed by Harri- 
son’s method. The wound surface produced by this operation 
is so small and heals so readily, that no skin graft was made. 

For the experiments to study cardiac muscle, a much more 
complete operation is necessary. Not only must the cord, 
hind-brain and skin of the dorsal region be removed, but the fore- 
and mid-brain, the primordia of the cranial ganglia (fig. 2) and 
the skin of the entire head region must be taken out as well. 

‘The cardiac plexuses receive such a large contribution from 
the tenth nerve, according to the work of Kuntz (’09), that it 
is especially necessary to extirpate its ganglia. In operating, 
the first step is to remove the dorsal structures as in the previous 


162 DAVENPORT HOOKER 


experiment. The embryo is then laid on its side and the skin 
cut from the dorsal wound surface to the ventral aspect of the body 
behind the branchial region. When this has been done on both 
sides, the skin is dissected away from the entire head down to 
the suckers and removed by a circular cut just above them. The 
remainder of the nervous system of the embryo which has not 
been previously removed is thus completely exposed and may be 
taken off with the knife or needles. 

As may be well imagined, the cut surface in such an operation 
is considerable, and, to prevent disintegration, a skin flap from 
the abdomen of another embryo is grafted over {the wound. 


Fig. 2 Diagrammatic sketch to show the cranial ganglia present at the stage 
of operating—They are, from before backwards, the ganglia of the 5th, 7th—8th 
complex, 9th and 10th. 


The flap is held to the embryo by piling silver wire about it to 
prevent movement, after the method of Born. 

The embryos were operated in 0.5 per cent saline solution and 
were kept in it for one or two days after. They were then trans- 
ferred to water. 

Histological examination showed that the nervous system was 
entirely removed (fig. 3), except in four cases in which a small 
portion of the infundibulum was found to have been left in. These 
cases were carefully examined for nerves, but no trace of any 
were found. In view of the fact that no differences were 
found between the action of the heart in these and in the embryos 
absolutely without any nerve tissue, they were not rejected. 


MUSCLE IN EMBRYOS WITHOUT NERVES 163 


Mechanical and electrical stimulation were used to determine 
the irritability of the voluntary muscle. Theembryos were stimu- 
lated with a human hair, according to Coghill’s method, and with 
finely pointed steel needles. All stimulating was done under 
the binocular microscope. 

For observations upon the heart, the embryos were placed on 
the stage of an ordinary microscope and the region just behind 


Fig. 3 Diagrammatic cross-section of the body of an embryo of R. palustris 
from which the entire nervous system was removed. H, Heart; N.C., Notochord; 
notice the absence of the spinal cord above it. 


the mouth watched for the beat. This method is a very accurate 
one, as the slightest action of the heart is noticeable by the move- 
ment of the skin. At later stages in palustris embryos, the 
skin becomes transparent and the heart itself may be seen. 

The embryos were killed in a sublimate-acetic mixture, sec- 
tioned and stained either with iron haematoxylin or with Teld’s 
molybdic acid haematoxylin. Those stained by the former 
method were usually counterstained with Congo red. 


164 DAVENPORT HOOKER 


The experiments detailed in this paper were performed on the 
embryos of Rana sylvatica, palustris and pipiens. The embryos 
of Rana sylvatica are the least desirable for operating, owing to 
the strong dorsal curvature of the body. Rana palustris and 
Rana pipiens have nearly straight backs, but in the former there 
is a greater contrast between the nervous tissue and the surround- 
ings, so that they may be easily distinguished from one another 
even macroscopically. For this reason palustris is preferable 
to the more diffusely colored pipiens. 


MOVEMENTS OF NORMAL EMBRYOS 


The movements of frog embryos in response to stimulation 
vary greatly according to their age. If embryos are carefully 
watched in their development and stimulated at short intervals, 
a stage will be found where some will respond to stimulation and 
some will not. The reaction is characteristic. As pointed out 
by Harrison (’04) “it consists of a sharp tonic contraction of the 
myotomes on the same side of the body and immediately at the 
point of the application of the needle prick.”” It is necessary that 
the stimulus be given with a needle sharp enough to penetrate 
to the myotomes. The stimulus, then, is a direct one, but it is 
likely that the mechanical effect of the contraction of one myo- 
tome may sometimes stimulate neighboring myotomes to contract. 

This non-nervous type of response is very different from the 
nervous type which is found in slightly later stages. The char- 
acteristics of this later type were pointed out by Harrison in 1904 
and have since been more thoroughly studied by Coghill (709). 
At this stage, the response is asymmetrical, the contraction first 
being away from the side stimulated, then toward it, the result 
being a swimming motion. 

A number of embryos taken at the first appearance of the early 
or non-nervous type of response, some of which moved and some 
did not, were fixed separately and cut into serial sections. On 
examination, it was found that the development both of those 
which responded and of those which did not was at essentially the 
same stage. In both, the fibrillation of the voluntary muscle 


MUSCLE IN EMBRYOS WITHOUT NERVES 165 


had begun and motor nerve connections between the spinal cord 
and myotomes were present. Thus it was evident that contrac- 
tion on stimulation is a function normally acquired in the axial 
musculature of the frog just after the beginning of the process 
of fibrillation and after establishment of the connection with 
the central nervous system. In the muscles of the limbs the 
case is different, for Braus (05) found that the muscles in Bom- 
binator do not respond to stimulation until a much later phase 
of development is reached and not until spontaneous contraction 
takes place. 

Harrison (’04) showed that fibrillation was not dependent on 
nervous connection. For proof that contraction is not depend- 
ent on it, we must turn to the experiments. 


MOVEMENTS OF CORDLESS EMBRYOS 
Effect of mechanical stimulation 


For the experiments on mechanical stimulation of voluntary 
-muscle, 77 embryos were operated, of which 34 died, were abnor- 
mal, or were imperfectly operated. Of the remaining 43 embryos 
which showed no abnormalities, 27 moved on stimulation and 16 
did not. It is interesting to note that three-quarters of the latter 
died later from some organic deficiency, while less than half of those 
that moved failed to live, indicating, perhaps, that the relatively 
large number which did not react as compared with the number 
which did was due to vital disorganization of some sort. 

Each of these embryos was placed on the stage of the binocular 
every twelve to twenty-four hours and stimulated with a human 
hair (Coghill, 09) or with a very finely pointed needle. The 
former method was finally rejected, since the operated embryos 
did not respond to skin stimulation as do normal ones with an 
intact reflex are. The latter method was very satisfactory, as it 
permitted the direct stimulation of very minute areas of muscle 
tissue without any profound injury to the embryo. 

Of those embryos which responded to stimulation, the irri- 
tability began on the first day after the operation in 9 individuals, 
on the second in 16 and on the third in 2. In no ease did any 


166 DAVENPORT HOOKER 


embryo begin to show reactions later than the third day after 
the operation. The reaction was a single quick contraction 
toward the side stimulated with the point of stimulation as the 
center of contraction. In no case was there a response unless 
the needle point penetrated the skin. Three specimens gave a 
tremor-like response, but microscopic examinations showed these 
to have nerve tissue present in the trunk region. Spontaneous 
movement was never observed in operated individuals. This is 
contrary to the results of Shaper and Goldstein, both of these 
authors having claimed that spontaneous movement is to be 
observed in the ventral half of embryos which have been cut 
longitudinally. I have repeated their experiments, operating 
at an earlier stage, however, but have not been able to confirm 
their results in this respect. 

One of the most striking features of this experiment is that the 
embryos do not continue to be irritable for an indefinitely long 
period but cease to respond after a day or two. The irritability 
to mechanical stimulation lasted but one day in 20 individuals, 
two days in 5, and but one embryo responded during a three- 
day period and one during a period of four days. 

Embryos killed at the end of the first day and examined for 
their histological structure, show that muscle fibrillation is well 
under way and that the entire development of the body is at the 
same stage as those of normal embryos in which contraction in 
response to stimulation is beginning to occur. The nervous 
system is, of course, lacking, but in general the stage of develop- 
ment is the same. 

Microscopical examination of the embryos killed and preserved 
from six to eight days after the operation, shows, as brought out 
by Harrison, that the muscle differentiation is practically com- 
pleted, though not quite so fully as in the normal embryo, that 
there is no nervous tissue posterior to the mid-brain and that, 
with the exception of the absence of the hind-brain and cord, the 
embryos are essentially normal. The principal abnormal feature 
of these embryos is their oedematous condition, plainly visible 
in the embryos as a whole and shown microscopically by the 
presence of many vacuoles in the muscle tissue. 


MUSCLE IN EMBRYOS WITHOUT NERVES 167 


Effect of electrical stimulation 


For the experiments on electrical stimulation of voluntary 
muscle, 15 embryos were operated, all of which remained in the 
best of condition until killed. They were each stimulated, at least 
once daily, by a weak faradic current. The embryos were care- 
fully watched through a binocular microscope while the stimu- 
lation was applied. Each embryo was placed on a small plat- 
inum plate in water, a fine platinum wire being used to form the 
other electrode. On each stimulation, the embryo responded 
by a single tonic contraction toward the side stimulated, the 
point of stimulation being the center of contraction. The type 
of contraction exhibited was exactly the same as that shown 
when the operated embryos were stimulated mechanically. 

The embryos were killed after four or five days, during which 
time response to mechanical stimulation had ceased. Up to the 
time they were killed, all the specimens responded in the manner 
described above whenever stimulated electrically. 

Microscopical examination of sections of these embryos shows 
that the cord was entirely removed. The embryos were in every 
respect similar to those used in the experiments on mechanical 
stimulation. 


EXPERIMENTS TO DETERMINE THE MODE OF TRANSMISSION 
OF THE STIMULI 


Wintrebert (04, ’05) has described a ‘sensibilité primitive’ 
existing in embryos of Rana esculenta during a period of four 
days beginning at the stage following the closure of the neural 
folds. He says that if an embryo during this stage is cut through 
the back in the posterior part of the body, so that the spinal 
cord, notocord and part of the yolk are severed, leaving the two 
halves connected by the skin and yolk of the ventral region only, 
the front part will respond on stimulation of the posterior part. 
After the heart begins to beat, a cut near the anal region severs 
the animal into two independent parts, of which the anterior 
will not respond on stimulation of the posterior. Up to this time 


168 DAVENPORT HOOKER 


there are no peripheral nerves, and the yolk will not transmit 
stimuli, so Wintrebert concludes that it must be the skin which 
acts as the transmitter. 

To test these results, three sets of experiments on frog larvae 
and additional ones on Amblystoma were made. They were (1) 
a repetition of Wintrebert’s experiments, (2) exposure and stimu- 
lation of the yolk, and (3) girdling of the skin around the body 
and stimulation of the posterior half with and without cutting 
the cord. 

Embryos with the cord cut. In repeating Wintrebert’s experi- 
ments (table 1 and fig. 4), it was found that no reaction was ob- 
tained in normal embryos of the first stage named by him. A 
4.5 mm. Rana palustris embryo in which stage the tail bud is 
quite pronounced was the smallest from which a response was ob- 
tained. The discrepancy may, however, be due to the difference 
between European and American forms. Furthermore, in all 


TABLE 1 


Repetition of Wintrebert’s experiments 


EMBRYO SIZE RESPONSES 

| mm | | | | | 
ike 4.25 |0/0/0|0/0/0/0/0)0/0)0* 0 
2a 4.50 |0/0/0/04 0/+70/0/0/0/0/0 
Se ORM acs: 4.75 |0/0/0/0/0/0/0j;0/0/0]0]0 
AE Dees. | 4.75 |0/0)/0/0/0/0/0/0/0/0)0}0 
is a en arom ete | 5.00 }0/0]*/|*/0/+970/+970/0/0/0 
Honbuan nearest: EU CO 0) Herel AO eaters) (2) | | CON er 
sar eecn ee gat ee see] 6200) [ESE 94-14-59] 0) | 01/00 0/010 
Se eee al) 16200 0/07 0/0/0)0/0)0/0)0/0/0 


| 
| 


Intervals of stimulations 15 seconds. 


*Spontaneous contraction without stimulation. Where this is found above a 
negative response, it indicates that a contraction took ,place in the interval of 
stimulation following. 

+ Contraction slow in following stimulation. 

°Contraction obviously due to shaking. 


Total number of stimulations............ oe ees Ob 
Total number of positive responses. 7 aS eer SLO 
Total number of spontaneous Contractions, seta 5 5 


Total number of contractions due to shaking.................. epee 


MUSCLE IN EMBRYOS WITHOUT NERVES 169 


but one case, the number of times when responses were obtained 
from operated embryos was much less than the number of times 
when no reaction occurred. The older embryos frequently move 
spontaneously, however, and the probabilities are that some will 
move during the time that they are under observation. Many 
times the embryo moved just before it was to have been stimu- 
lated. That the movement may occur apparently as a result of 
stimulation, when it would have taken place without it, is abso- 
lutely certain. Several times, while an embryo was being stim- 
ulated at regular intervals and a large number of reactions was 
being obtained, the omission of two or three stimulations showed 
that the embryo was contracting more orlessrhythmically, whether 
stimulated or not, the periods of the rhythm being the same as 
the intervals of stimulation. 

This, however, is not the only explanation which can be offered 
for the contractions which take place. On careful analysis of 
the conditions preceding the contraction, three factors can be 
seen at work: (1) tension of the skin over the whole body pro- 
duced by pressure with a dull needle, (2) shaking the embryo 
during the stimulation, and (3) hindering the locomotion of the 
embryo due to the action of its cilia. Whenever a needle sharp 
enough to penetrate the skin without increasing the tension 
elsewhere was used, no contraction resulted unless the myotomic 
area was stimulated. The shaking of the operator’s hand will 
clearly cause contractions, so that great care must be taken to 
avoid this. Frequently the needle sticks after penetration and 
the embryo is shaken as it is withdrawn. This also produces 
contractions. Where the embryo’s progress, caused by its own 
cilia, is suddenly stopped, a contraction almost always ensues. 
This is avoidable by slowly checking the motion. 


Stimulation of the yolk 


In the second series of experiments, an opening was made in 
the skin at the side of the animal (fig. 5) and the yolk stimulated 
through it. Out of a total number of 100 stimulations, 29 con- 
tractions were obtained. Of this number, 14 were obviously 


170 DAVENPORT HOOKER 


due to shaking the embryo by running the needle through the 
body to the skin of the opposite side. This impaled the embryo, 
withdrawal of the needle was diffieult and in every case the embryo 
was shaken. One embryo was contracting freely between the 
stimulations and the large number of responses given by it may 
be accounted for to some extent in this manner. Table 2 demon- 
strates the details. 

The results show that the yolk, per se, as Wintrebert stated, 
will not transmit impulses, but contractions may result from shak- 
ing the embryo during stimulation. 


6 


Fig. 4 Diagram of a frog embryo to show position of cut (by dotted line) used 
in the repetition of Wintrebert’s experiments. 

Fig. 5 Diagram of a frog embryo to show opening through which the yolk was 
stimulated in the second series of experiments to determine the mode of trans- 


mission of stimuli. 
Fig. 6 Diagram of a frog embryo to show “‘girdling”’ of the skin around the 
body and the area in which stimuli were effective when the cord was uncut. 


Girdled embryos 


In the third series of experiments, the skin only was cut around 
the entire body, thus girdling it (fig. 6). 

Stimulation in the region of the cord (within the dotted area 
indicated in the figure) when the latter is uncut, will nearly always 
produce a response in the anterior end of the body (table3). If the 
cord is cut, there is no transmission of stimuli from the posterior 
part to the anterior. When the skin is girdled, the cord alone may 
serve for the transmission of mechanical shocks to the anterior 


MUSCLE IN EMBRYOS WITHOUT NERVES 


half, such as those produced by shaking the embryo. 


171 


If the 


cord is also cut, the yolk has not consistency enough to admit 
of shaking the anterior half by disturbances in the posterior, con- 


sequently no contractions result. 


TABLE 2 
Stimulation of the yolk only 


EMBRYO! SIZE RESPONSES 

gee 5.00 |+/+| 0/0 Pe Ae a] 04 0/0 +%+/0/ 0 
gions 5.50/0 +40 /0|/+*0/0/0/0/0/0/0)/0\0 0/+*0/0/0/ 0 
hone 5.75|0/0)0)0/0/0/0/0/0/0\0l+4+4 0 00 44/0/ 0/0 
ae. 6.000 0 00/00/00 0 +1440 +40 /F} 0 +10 0/0 
Ges 6.00/0/0/0/O}+*+4+ 544444444 0 00 +*/0|+*10] 0/0 


Intervals between stimulations 15 seconds. 


{Spontaneous contraction without stimulation. 
*Needle pierced animal and embryo was shaken. 
Embryo 1 contracted frequently between stimulations. 


otalénumberkoras tins tlOMG esereceyteas i cetsrele ce ets tetedepete avatcteiey tose ie elet ciel 100 
Total numberof positive TEsPONseS soci <r. ccc oer wpeiei eye oo woiee eo 8 
Total number of spontaneous contractions....................2.0.. 7 
Total number of contractions due to shaking..................... a 314 


TABLE 3 


Stimulation of the posterior portion of the embryo after ‘girdling.’ 


Cord intact. 


EMBRYO SIZE RESPONSES 
| | 
ae ee ae 
Oar see even NATE LoS ics | 4.50 |0/0/0/0|0/0/0/}+*4+5+4+%+* 
Det arse yeie st oeepaisioc, debiaht esatsia ty 5.00 Bevb i ilues cea a eae hates 
SUE oem pean eics i 8 | 5.50 0 0/0 0/0 ON rai viata ck 
ATS ee Te eT Tee ec |} 5.50 |+* 0 PO aut 0 
Buen. Seats wear et esed tae! 6.00 o/o/0/0|0|+/0|0/0H-*+*X0 
Intervals of stimulation 15 seconds. 
*Stimulations within dotted area shown in fig. 4. 
pLOtall MUM bers Oms tM Ula GONG ere,se-c eteiee sie letene ele sists tein x ote c ve nla sveleia- ove 60 
Total number of positive responses..........-----0+escceseeeeeeens 1 
Total number of spontaneous contractions...............0.00e00 ee 0 
Total number of contractions due to stimulation of the cord........ 24 


172 DAVENPORT HOOKER 


Judging from the results obtained, it would appear that, as 
Wintrebert claimed, the yolk will not transmit stimuli, but on 
the other hand, contrary to Wintrebert, there is no evidence that 
the skin will act as a transmitter of impulses other than those 
of a mechanical nature. 


EXPERIMENTS ON THE HEART 


For the experiments on the heart 45 embryos were operated. 
Of these, 16 died; in 2 it was impossible to see whether the heart 
functioned or not; and in 2 others the operation was imperfect. 
Of the remaining 25, the heart-beat was observed in 21, while it 
was found not to beat in 4 individuals. 


The pulse rate 


The beat usually becomes visible on the second or third day 
after the operation, though in a very few cases it begins earlier 
or later than this. The beating of the heart begins in normal 
individuals about twenty-four hours before it is visible in the 
operated embryos. The pulse-rate in the normal ones begins at 
25 to 40 per minute and gradually increases to from 60 to 70. 
This latter rate is held approximately throughout the entire 
developmental life of the embryos. It is, however, difficult to 
give any definite rate as being the normal, owing to the fact that 
certain factors tend to vary it greatly. Temperature changes 
cause great fluctuations, and activity on the part of normal 
individuals tends to accelerate the heart. The comparisons be- 
tween normal and operated embryos are not affected by tem- 
perature since controls and operated embryos are kept under the 
same conditions, but the muscular activity of the control, which 
often cannot be prevented, may seriously alter the validity of a 
comparison with the operated individual in its enforced perma- 
nent inactivity, the higher rate of the active normal being more 
nearly equivalent to the pulse of the operated embryo than is 
the rate in an animal which has been resting quietly. The rate 
in the operated individuals begins at 28 to 42 and rises slowly to 


MUSCLE IN EMBRYOS WITHOUT NERVES 173 


from 60 to 65 during the first few days, then the rate increases 
gradually to as high as 80, dropping soon after to about 50, 
shortly after which the embryo dies. 

Two reasons may be given for this rise of pulse-rate. 

Up to the time when the external gills would normally form, 
the rate is normal; but since the epithelium of the branchial region 
is removed by the operation, no gills ever develop, and a slow 
asphyxiation of the embryo results. This would ordinarily bring 
about acceleration of the pulse. Evidence of asphyxiation is 
found in the fact that all the embryos were oedematous, a con- 
dition, according to Fischer (710), produced by a lack of oxygen 
to the tissues. To aid in aeration of the blood, some embryos 
were kept in water under an atmosphere of oxygen. These were 
not as oedematous as embryos not under oxygen, and they lived 
longer. The rise of pulse-rate finally took place, however. 

Another explanation which can be given for the increased 
pulse-rate is that the time when the rate begins to increase beyond 
the normal is nearly coincident with the establishment of the 
connection of the vagi with the heart in normal embryos. It 
would be interesting if it could be proved conclusively that the 
absence of the inhibitory action of the vagi is responsible for the 
increased pulse. 

A number of experiments were performed to attempt an elu- 
cidation of this point. In one series the skin of the branchial 
region was removed, leaving the embryos with an intact nervous 
system, but preventing the formation of external gills. The pulse- 
rate was carefully counted in both operated and control individ- 
uals. While the beat was distinctly higher in the operated 
embryos during the first few days, the difference decreased as time 
went on, probably in consequence of the development of inter- 
nal gills, which would tend to bring the operated embryos back 
to normal condition. The results of these experiments indicate, 
then, that it is the disturbance of the respiratory function that 
causes the increase in the pulse-rate, though they are not conclu- 
sive with reference to the possible further effect of the withdrawal 
of the action of the vagi. 


174 DAVENPORT HOOKER 


In another set of experiments, certain embryos were grafted 
by the dorsal wound surface to the lateral abdominal region of 
normal individuals, in the hope that the nurse would aerate the 
blood for the operated specimen. This was in a measure suc- 
cessful, the life of the parasite being prolonged, though not for 
an indefinite period. The position of the parasite upon the host 
was not favorable, however, for the observation of the heart, and 
the nature of the observations desired precluded the use of 
anaesthetics. 


Differentiation of the heart muscle 


Histological examination of these embryos demonstrated that 
the heart was in every case well developed and that the differen- 
tiation of the cardiac muscle was progressing normally. In the 
heart of normal individuals, the differentiation of the muscle 
tissue had progressed much farther and many more striated 
fibrillae were present, but one would naturally expect, as was the 
case with the voluntary muscle described by Harrison, that the 
differentiation of the heart would be retarded by the abnormal 
condition of the embryo. Much yolk is present and a few vacu- 
oles are to be found here and there, but many sharply defined 
striated fibrillae may be seen in all parts of the heart (fig. 7). 
These differences between the normal heart tissue and that of 
the operated embryos also correspond to those found by Har- 
rison (’04), in the case of the voluntary muscle tissue. 


PLATE I 


EXPLANATION OF FIGURES 


8, 9,12 and 18, from a model of the heart of anormal embryo of Rana palustris 
19 mm. in length. Magnified 50 diameters. 

10, 11, 14 and 15 from a model of the heart of an embryo of Rana palustris from 
which the entire nervous system had been removed at the stage following the clos- 
ure of the neural folds. Magnified 50 diameters. 

Ao, Aortic arch. S. V., Sinus venosus. 


Au, Auricle. 


T. A., Truncus arteriosus. 


Ao', Aortic arch which ends blindly. V, Ventricle. 


V.O., Opening of sinus venosus into auricle. 


8 View of the right side of the model of a normal heart. 
9 View of the front of thesame model. The liver (fig. 7, L) has been removed 
the model. 


from 
10 
113 
12 
13 
14 
15 


View 
View 
View 
View 
View 
View 


of the 
of the 
of the 
of the 
of the 
of the 


right side of the model of the heart of an operated embryo. 
front of the same model. 

interior of the model shown in fig. 7, looking anteriorly. 
interior of the model shown in fig. 7, looking posteriorly. 
interior of the model shown in fig. 9, looking anteriorly. 
interior of the model shown in fig. 9, looking posteriorly. 


MUSCLE IN EMBRYOS WITHOUT NE 


RT HOOKER 


XPERIMENTAL ZOOLOGY, VOL. 11, 


net. wf nO 
ibpathied from yay ; 
nNeC rom e g ; 


we” The model 


2 
Whe) 


MUSCLE IN EMBRYOS WITHOUT NERVES 177 


The gross form of the heart 


Three methods were used in studying the gross form of the 
heart: examination of sections, wax-plate reconstructions and 
dissections. In all of these careful comparison was made between 
hearts of normal and of operated individuals, and the results 
obtained from the.different sources were the same. 

The models from which the figures were taken were made from 
a normal individual of Rana palustris, 10 mm. in length and from 


Fig. 7 Portion of the heart wall of the same embryo as Figure 3, showing the 
striated muscle fibrillae. Zeiss homo. immersion 2 mm., 2 oc. 


an operated embryo of the same species, killed seven days after 
the removal of its nervous system. The two embryos were in 
very nearly the same stage of development. The model of the 
normal heart was made at a magnification of 133} diameters, 
that of the nerveless organ 100 diameters; both have been reduced 
to 4 magnification of 50 in the figures. 

The heart of the normal frog larva shows clearly the relations 
of the various chambers in the embryonic organ (figs. 8, 9, 12 
and 13). The venous sinuses are more wholly within the tissue 
of the liver than is the case in the adult organ, a fact due to 
the close relation of the liver to the heart in this stage. The 


178 DAVENPORT HOOKER 


sinus venousus is a large cavity emptying into the right side of 
the auricle, which occupies a dorsal position on the heart. The 
auricle opens into the ventricle which occupies a ventral position. 

The hearts of the operated individuals are much like the nor- 
mal, except that in nearly every case some of the aortic arches 
end blindly. This is due to injury received in the operation. 
When one considers the very close relation of the heart in its 
pharyngeal position to the tissues of the brain and cranial gan- 
glia, one can readily see the liability of the heart to injury during 
the removal of the nervous tissue. The necessity of removing 
the branchial epithelium with the consequent injury to the bran- 
chial arches, causes a disturbance of the normal formation of 
the aortic arches and leaves the systemic trunk alone functional. 
As a result of this, the truncus arteriosus does not usually divide, 
but remains single and runs directly into the dorsal aorta. The 
venous supply is normal. 

The most striking thing about the gross form of the heart in 
the operated embryos is its relatively great size (figs. 10, 11, 14 
and 15). Both the auricle and ventricle are dilated and hyper- 
trophied. Asis usual, the dilatation of the auricle is greater than 
that of the ventricle but the hypertrophy of its walls is less. 

Dilatation of the heart isin general produced by two causes, 
working separately or together, viz: increased internal pressure 
and cardiac insufficiency. There seems to be ground for the 
belief that there is an hydraemia in these embryos, a condition 
which would create an increased internal pressure on the heart. 
As already noted, these embryos are oedematous. Fisher, (’10) 
says ‘“‘every condition that makes for a state of lack of oxy- 
gen... .. be this through disturbances in the affected 
parts themselves, or in some very distant organ, as the heart, 
makes for an increase in the severity of the oedema.’’ The entire 
animal must suffer through lack of oxygen due to the disturbances 
in the arterial supply of the body and the non-development of 
gills in the operated embryos, which causes difficulty in properly 
oxygenating the blood. Krehl (’07) states that “‘it is possible that 
an hydraemia should be caused, not only by a primary reduction ~ 
of the proteid constituents of the blood, but by a primary increase 


MUSCLE IN EMBRYOS WITHOUT NERVES 179 


in the amount of water. The hydraemias associated with 
ee vc cardiac insufficiencies are probably partly of this 
nature.” 

In general, wherever dilatation is present, hypertrophy occurs, 
but the auricle is much behind the ventricle in the compensatory 
thickening of its walls. Another factor though probably not 
such an efficient one, in the hypertrophy of the heart is its rapid 
beat. Tachycardia, or any other condition which increases the 
work of the heart, tends to produce hypertrophy. As pointed 
out above, it is impossible to prove definitely whether the tachy- 
cardia here present is due to dyspnoea or to the lack of the 
controlling power which the vagi normally exercise, though the 
evidence at hand points toward the former. 


DISCUSSION OF RESULTS 


It is evident from the experiments that in normal frog embryos 
the differentiation of the fibrillae in axial muscles precedes the 
acquisition of contractility and that contractility in response to 
stimulation appears before spontaneous contractility. Further, 
just before the embryos become irritable to stimuli, the myotomes 
become connected with the central nervous system by motor 
roots. The fact that contractility begins almost synchronously 
with the establishment of nervous connections in the muscles of 
the body wall might suggest that there is a causal connection 
between the two and, as a matter of fact, this mode of reasoning 
has already been employed by a number of authors. The inse- 
curity of such arguments is clearly demonstrated, however, by 
the experiments on cordless embryos whose muscles react to 
stimuli without ever being connected with the nervous system. 

It seems quite clear that the contractions of voluntary muscle 
tissue seen after mechanical stimulation of cordless frog larvae 
are due to direct stimulation, since the central nervous system was 
removed before there were nerves of any kind growing from 
it, and since contraction takes place only when the muscle is 
actually punctured. This phenomenon is, however, explained 
by Wintrebert, Shaper and others on the basis of a mode of non- 
nervous transmission of impulses. It is, of course, impossible 


180 DAVENPORT HOOKER 


to prove that a non-nervous conduction path does not exist but, 
on the other hand, its presence has never been actually demon- 
strated by any of the experiments which have been adduced in 
its favor by previous investigators and all the phenomena which 
have been supposed to demonstrate the existence of such conduc- 
tion paths are quite as readily explained on the basis of more 
familiar factors. Mechanical tension of the skin and shaking 
or jarring of the embryo will account for Wintrebert’s results. 
The burden of proof in this question rests entirely upon those 
who claim the presence of a conducting mechanism other than 
those whose existence is already clearly established. 

There is such a close parallelism between the normal and cord- 
less embryos that we are fully justified in attributing the first 
movements of normal individuals to the same cause as that which 
produces responses in the operated embryos, i.e., the direct 
irritability of the muscle tissue. When stimulated mechani- 
eally or electrically, the operated embryos respond by a single 
quick contraction toward the side stimulated, the point of stim- 
ulation being the center of contraction. The parts of the ani- 
mal anterior and posterior to the parts stimulated do not enter 
into the response. In stimulating normal embryos at a very 
early age to separate those which responded from those which 
did not, it was quite noticeable that the former always contracted 
on the side stimulated. A sharp needle was used and the stimu- 
lation was given directly to the myotomes, as it was in the cord- 
less embryos which were mechanically stimulated. In this stage 
normal embryos have in the Rohon-Beard nerves a fairly well 
developed sensory system, but it is doubtful whether the central 
connections with the motor apparatus are established. At a 
somewhat later stage a considerable number of fiber tracts are 
developed in the cord and then the embryos respond to the ligh* 
est touch in the manner described by Harrison and Coghill. Cog- 
hill used a human hair with which to stimulate, but this would 
have no effect unless the sensory system were developed, and 
there must be a complete sensory-motor are to produce a contrac- 
tion on the opposite side of the body from the one stimulated. 
A simple contraction toward the side stimulated, involving an 


MUSCLE IN EMBRYOS WITHOUT NERVES 181 


extremely small area, indicates that the stimulation was limited 
to one or two myotomes and was given directly. That the irri- 
tability of the operated embryos to direct mechanical stimula- 
tion should cease after a period of active response is extremely 
perplexing, because its irritability to electrical stimuli persists 
considerably longer and, as is well known, adult muscle responds 
to direct mechanical stimulation. 

The results of the study of the heart action in embryos deprived 
of all nervous tissue give very clear cut evidence in favor of the 
myogenic theory of the heart beat. The two principal theories 
of the nature of the heart beat, from myogenic or neurogenic 
causes, have each many strong arguments in their favor. To 
prove conclusively that the myogenic theory is the correct one, 
three things are necessary. (1) It must be shown that undiffer- 
entiated cardiac muscle possesses the ability to pulsate rhythmi- 
cally before the nervous system establishes connections with it. 
It must be proved (2) that cardiac muscle when differentiated 
will continue to contract rhythmically without ever having been 
connected with the nervous system and (3) that a heart which 
has so developed will act physiologically in the same manner 
as a normally developed heart of equal age. 

The first requirement for a conclusive proof of the myogenic 
theory has been clearly demonstrated already. It has long been 
known that the embryonic heart will beat before it is connected 
with the nervous system. Fano (’85) found the beating heart 
in the chick on the second day of incubation. Chandler (’08)* 
found the heart beating in chick embryos of 11 somites and Lillie 
(08) mentions the occurrence of the heart beat in chicks of 10 
somites. Paton (’07) has described the heart beat in Pristiurus 
embryos of 4 mm. length. In all these instances, the occurrence 
of the heart beat has preceded connection of the heart with the 
central nervous system, though Paton believes that there is a pro- 


*Perley B. Chandler, 1908, ‘‘A study of the neurofibrillae of the central ner- 
vous system.’’ Thesis presented for the degree of M. D. in the Department of 
Medicine of Yale University. Dr. Chandler was killed in a railway accident shortly 
after graduation and his thesis was never prepared for publication. I am 
indebted to the Medical Department for the use of his paper. 


182 DAVENPORT HOOKER 


toplasmic conducting path for impulses to it. Further, as Shaper 
(98) says: “We know that the anlage of the heart has a period 
of rhythmic pulsation before there can be any specific differentia- 
tion of nervous substance.” It seems evident, then, that undif- 
ferentiated cardiac muscle possesses the ability to pulsate rhyth- 
mically before the nervous system establishes connections with it. 

The second requirement for a proof of the myogenic theory is, 
I think, adequately met by the experiments detailed in this paper. 
My results are fully in accord with those of many other investi- 
gators, but the method used here gives a much more rigorous 
proof than has heretofore been offered. The destruction of the 
nervous system by chemical or physical means or by trusting to 
the atrophy of the remainder after the removal of a part is not 
as sure of producing unquestionable results as careful total 
removal of the nervous system at an early stage by surgical means. 
Wintrebert (’05) notes the fact that he found the heart beating 
in embryos of Rana viridis after the total removal of the nervous 
system. Vernoni (’10) finds that the heart of the chick will 
continue to beat for several days after the nervous system has 
been destroyed by exposure to radium at a very early age. Fi- 
nally, Burrows (’11) has succeeded in totally isolating the living 
heart from chick embryos into plasma clots and finds that they 
will beat rhythmically for three days. It is evident from the 
experiments that not only will the embryonic heart of the frog 
begin to beat without the influence of the nervous system, but 
will continue to beat till conditions entirely extraneous to the 
heart, such as difficulty in aérating the blood and inability to 
feed, kill the organism. During this period the tissue of the heart 
is normally differentiating, and the heart as a whole is passing 
through its normal development. It is, then, evident that car- 
diac muscle when differentiated will continue to beat rhythmically 
without ever having been connected with the nervous system. 

The third point necessary for complete demonstration, viz. : 
That a heart which has so developed will act physiologically in 
exactly the same manner as a normally developed heart of equal 
age, has not as yet been proved. Indeed, it is so difficult of proof 
in the case of animals like the frog, owing to the extremely small 


MUSCLE IN EMBRYOS WITHOUT NERVES 183 


size of the embryonic heart, that it is a question whether it can 
ever be adequately demonstrated. Until such time as this point 
is proved, we must content ourselves with evidence derived from 
the adult heart under conditions such that nervous control is 
eliminated. The first two points may be considered proved, the 
third can only be said to be probably true. 

The results of the experiments detailed in this paper, together 
with the work of Harrison, Braus and Paton, present certain dif- 
ferences in the various kinds of muscle tissues. All muscles, 
cardiac, axial and appendicular, differentiate independently of 
nervous control, but in their ability to function there is a graded 
difference. Cardiac muscle, which we have reason to believe is 
the most primitive, will function spontaneously and rhythmically 
without nervous control. The axial muscles of the frog, on the 
other hand, will not function spontaneously in the absence of 
the nervous system, though they will respond to direct stimula- 
tion. In appendicular muscle, however, the muscle fibers are 
still more dependent on the nervous system and do not respond 
to stimulation till late in their development. The work of Braus 
(05) demonstrates conclusively that normal and transplanted 
limbs of Bombinator have a complete nervous supply to the mus- 
cles before they move spontaneously. He found that he could 
obtain no contraction on stimulation with a faradic current until 
spontaneous contraction had begun to appear, and unipolar stim- 
ulation of the limbs in very early stages gave no results. It is 
interesting to note that Paton has shown that the axial muscles 
of Pristiurus function spontaneously, before the development 
of nerves, in a rhythmic manner similar to that of heart muscle. 
It may very well be that the limb muscles of this form present a 
condition similar to that of the axial muscles of the frog. The dis- 
covery of this point and the determination of the period of the 
transition in both axial and appendicular muscles in the phylo- 
genetic scale would greatly broaden our knowledge of muscular 
development and functioning power. 

In conclusion, it is a pleasure to acknowledge my very great 
indebtedness to Dr. R. G. Harrison for suggesting this problem 
and for his invaluable aid as the work progressed. 


184 DAVENPORT HOOKER 


SUMMARY 


1. In normal frog embryos, the differentiation of muscle fi- 
brillae and the establishment of nervous connection with the cen- 
tral nervous system precede, though but slightly, the acquisition 
of contractility in the myotomes. 

2. Voluntary muscle, which has developed without nervous 
influence, will not contract spontaneously but will respond to 
mechanical stimuli directly applied and to electrical stimuli. 
Irritability to electrical stimuli persists considerably longer than 
irritability to mechanical. 

3. There is no evidence that stimuli may be transmitted along 
non-nervous structures, such as the skin and yolk, though ten- 
sion of the skin and shaking the embryo may produce responses 
by bringing about direct mechanical stimulation of the myotomes. 
The contraction of one myotome may, however, mechanically 
stimulate others near it within very close limits. 

4. The heart will function normally in the total absence of 
the nervous system and its tissue will differentiate normally. 
The differentiation of cardiac muscle, like that of voluntary mus- 
cle, is slower than in normal individuals, corresponding to the 
retarded development of the body as a whole, but the fibrilla- 
tion is absolutely normal. The differences between normal and 
operated individuals are those of degree, not of kind. 

5. From the morphogenetic standpoint, the complete removal 
of the nervous system has little effect on the heart. The abnor- 
malities which do appear after the operations described are directly 
traceable to functional disturbances of which the general oedem- 
atous condition of the body is an index. This condition is 
caused by the difficulty in properly oxygenating the tissues owing 
to a disturbed arterial supply and the absence of external gills, 
for which the operative technic is entirely responsible. 


MUSCLE IN EMBRYOS WITHOUT NERVES 185 


BIBLIOGRAPHY 


Born, G. 1896 Uber Verwachsungsversuche mit Amphibienlarven. Arch. f. 
Ent.-mech. Bd. 14. 

Braus, H. 1905 Experimentelle Beitrige zur Frage nach der Entwicklung per- 
ipherer Nerven. Anat. Anz. Bd. 26. 

Burrows, M. T. 1911 The growth of tissues of the chick embryo outside the 
animal body, with special reference to the nervous system. Jour. 
Exp. Zo5l., vol. 10. 

Cocuiti, C. E. 1909 The reaction to tactile stimuli and the development of 
the swimming movement in embryos of Diemyctylus torsus, Escholtz. 
Jour. Comp. Neur., vol. 19. 


Fano, Giutio 1885 Sullo Sviluppo della funzione cardiaca nell’embrione. Lo 
Sperimentale. 


FiscHer, Martin’ 1910 Oedema. Wiley and Sons, New York. 


GoupsTEIN, Kurr 1904 Kritische und experimentelle Beitrage zur Frage nach 
dem Einfluss des Centralnervensystems auf die embryonale Entwick- 
lung und die Regeneration. Arch. f. Ent-mech., Bd. 1S. 


Harrison, RossG. 1903 On the differentiation of muscular tissue when removed 
from the influence of the nervous system. Am. Jour. Anat., vol. 2. 
1904 An experimental study of the relation of the nervous system to 
the developing musculature in the embryo of the frog. Am. Jour. 
Anat., vol. 3. 


Hetp, Hans 1909 Die Entwicklung des Nervengewebes bei den Wirbeltieren. 
Leipzig. 


Hooker, Davenrorr 1910 The development and function of cardiac muscle 
in embryos without nerves. Proc. Soc. Exp. Biol. and Med., vol. 7. 


Kreni, Lupotpx 1907 The principles of clinical pathology. Trans. by A. W. 
Hewlett. 


Kuntz, ALpert 1909 The role of the vagiin the development of the sympathetic 
system. Anat. Anz. Bd. 35. 


Linu, F. R. 1908 Development of the chick. New York. 


Paton, Stewart 1907 The reaction of the vertebrate embryo to stimulation 
and the associated changes in the nervous system. Naples Mitth., 
Bd. 18. 


SHaper, A 1898a Experimental studies on the influence of the central nervous 
system upon the development of the embryo. J. Bost. Soc. Med. Sciences. 
1898b Experimentelle Studien an Amphibienlarven. Arch. f. Ent- 
mech., Bd. 6. 


Vernoni, Guipo 1910 Studi di Embryologia sperimentale.  [.’azione del radio 
sull’ uovo di pollo, Arch. f. Entmech. Bd. 31. 


186 DAVENPORT HOOKER 


WINTREBERT, M. P. 1903 Influence du systéme nerveux sur !’ontogenése des 
membres. C. R. Acad. Sci., Paris. Tome 137. 
1904 Sur |’éxistence d’une irritabilité excitomotrice primitive, indé- 
pendante des voies nerveuses chez les embryons ciliés des Batraciens. 
Comptes Rendus de la Soc. de Biol., vol. 57. 


1905a Sur le développement des larves d’Anoures aprés ablation ner- 
veuse totale. Ibid., vol. 58. 


1905b Sur le développement de la contractilité musculaire dans les 
myotomes encore dépourvus de liason nerveuse réflexe. Ibid., vol. 59. 


1905e Nouvelles recherches sur la sensibilité primitive des Batraciens. 
Ibid., vol. 59. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 
AND INHERITANCE IN EXPERIMENTAL 
REPRODUCTION 


II. PHYSIOLOGICAL DOMINANCE ‘OF ANTERIOR OVER POSTERIOR 
REGIONS IN THE REGULATION OF PLANARIA DOROTOCEPHALA 
C. M. CHILD 
From the Hull Zoélogical Laboratory, University of Chicago 


TWENTY-ONE FIGURES 


CONTENTS 

Introduction...... : : Pence anne Jaen Sg, 
Experimental data....... sor od kshY) 
1. Regional differences in head formation atone ihe axis et LS 
2. Regional differences in the position of the pharynx . F ... 192 
3. The later course of regulation . : : . 192 

4. The dominance of anterior regions in the peas itory development of pharnyx 
and intestine. . Ba MESSE ETc a Nena siaele So eines Miata = coer maesretsiacls 194 
5. The position of the pharynx mien experimental conditions. PEE O33 LOS, 
A. In pieces from the old postpharyngeal region. .... beatae os ee LOS 
B. In pieces from the old prepharyngeal region. Sane OE .. 204 
Conclusion and summary....... si . 210 
Bibliography . . : 220 

INTRODUC] ION 


It has long been recognized by the botanists that the apical 
region is physiologically dominant over other regions to a greater 
or less extent in higher plants. In Tubularia (Child, ’07) and 
various—perhaps all other—hydroids the hydranth region is 
dominant over more proximal regions so far as they are not beyond 
the limit of effectiveness for the given conditions. In the present 
and following papers data will be presented which establish, I 
think, beyond a doubt that in Planaria dorotocephala a relation 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 3 
OCTOBER, 1911 


187 


188 Cc. M. CHILD 


of the same sort exists between anterior and posterior regions. 
The head region dominates and controls all regions posterior to 
it, though the degree of this dominance differs at different levels 
and under different conditions. 

It is primarily the processes in the anterior region of the body 
which determine the localization and differentiation of more pos- 
terior parts and the formation of a second zoéid and often of addi- 
tional zodids at the posterior end of the body is due to the fact 
that the correlative factors originating in the anterior region are 
effective only within a certain distance, the effective distance, 
i.e., they lose in energy or effectiveness with transmission from 
the point of origin, so that if increase in length beyond a certain 
point occurs certain regions become in greater or less degree phys- 
iologically isolated from the dominant region and begin to react 
in much the same way as when physically isolated. 

In short, we shall find that if a head is formed in regulation, 
the development of other organs follows necessarily, so far as 
nutritive material is available. On the other hand, the regula- 
tory formation of a head in Planaria is not due to any ‘inherent 
tendency’ of any sort whatever of the piece to form a new whole. 
So far is this from being the case that we find that new heads form 
more frequently in pieces from old than in pieces from young 
animals and under certain conditions we can even increase the 
capacity of a piece to produce a new head by decreasing the rate 
of dynamic processes in it by means of alcohol and various other 
anesthetics, by low temperature and by various other means. The 
data to be presented in this and following papers will show that 
the formation of a new head in an isolated piece of Planaria is to 
a certain extent opposed to rather than correlated with other 
activities of the piece. 

When once the process of head formation is well under way 
then the remodelling and redifferentiation of the other parts into 
a new whole begins. In general no piece of Planaria is abie to 
give rise to structures characteristic of regions of the body anterior 
to that which it originally occupied, wnless a new head region forms 
or begins to form first. The new whole arises from the piece in 
such cases not by a process of ‘restitution’ of the missing parts, 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 189 


but in a manner rather closely comparable to the formation of a 
new bud and a new axis from a region of differentiated tissue in a 
plant. We may say that in general the conditions which favor 
the regulatory development of a dominant part are opposed in 
character to those which favor the development of subordinate 
parts. Under certain conditions weakness or depression of the 
old system constitutes one of the most favorable factors for the 
development of a new dominant part, while vigor, i.e., a high 
rate or intensity of reaction favors the formation of subordinate 
parts. In other words, physiological isolation (Child, ‘1la) is 
the essential factor in the formation of a new dominant part, 
while the formation of subordinate parts is determined by corre- 
lation. That correlative factors are not entirely without influ- 
ence in the formation of dominant parts is shown by the fact that 
the frequency of head formation in Planaria varies with the length 
of the piece (Child, ’11b), but it is also a fact that short pieces of 
Planaria from any region will give rise to heads alone, provided 
only that the rate of reaction is sufficiently high. In the same 
way isolated short pieces of Tubularia and Corymorpha stems 
give rise to hydranths alone or even to only the distal portions of 
hydranths. Manifestly then the dominant part is capable of 
forming without any correlation with other parts, though correla- 
tive factors may increase or decrease the rate of its formation. 

The data obtained in the course of my experiments which bear 
upon this problem are in part morphological in part physiological: 
the presentation of the morphological data precedes that of the 
physiological because the former concern visible features of the 
processes of regulation while the latter afford a physiological basis 
for their interpretation. 

EXPERIMENTAL DATA 


’ 
1. Regional differences in head formation along the axis 


These regional differences were discussed at some length in the 
preceding paper and need be only briefly referred to here. If we 
compare pieces of the same length and within certain limits of 
length in sequence from the anterior end posteriorly, we find first 


190 C. M. CHILD 


| 


y 


o) 


Fig. 1 Outline of Planaria dorotocephala, showing 
various levels of section. 


Figs. 2-5 Pieces of different length from different 
regions, showing the different methods and results of 
regulation. Tig. 2, regulation in pieces above a certain 
length from the anterior region of the body: head large, 
pharynx near posterior end. Fig. 3, regulation in pieces 
above a certain length from the posterior region of the 
first zodid: head small, pharynx anterior to middle. Fig. 
4, one type of regulation in very short pieces from the 
anterior region; a ‘tailless head.’ Fig. 5, usual type of regulation in short pieces 
from posterior region of first zodid; ‘headless.’ 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 191 


that the rate of head formation and the size of the head formed 
decrease with increasing distance of the level concerned from the 
anterior end. Fig. 2, an anterior piece including the region ac, 
fig. 1, and fig. 3, a piece from the posterior region of the first zo6did 
(df, fig. 1) illustrate the extreme terms of these differences. But 
these regional differences, which in longer pieces are merely quanti- 
tative become in the shorter pieces qualitative, i.e., in the morpho- 
logical sense, and we often find very short pieces from the extreme 
anterior end giving rise to tailless heads (fig. 4), while short pieces 
from the extreme posterior end of the first zodid, developing under 
the same conditions as the others, may give rise to completely 
headless forms (fig. 5). 

In addition to these differences which concern the head as a 
whole,we find that teratophthalmic! (Child,’11b) and teratomorphic 
heads (Child, ’11d) show a similar relation to the main axis: not 
merely the capacity to form a head but the capacity to form a 
‘normal’ head decreases posteriorly when pieces are compared 
under uniform conditions. These and other data cited in the pre- 


‘Tn my earlier experiments 1 distinguished only three types of anterior regula- 
tion in the reconstituted pieces: ‘normal,’ with two equal eyes symmetrically 
placed; ‘teratophthalmic,’ with unequal, unsysmmetrically placed, partially fused 
or single eyes; ‘headless,’ including all cases without eyes and auricles whether 
a distinct outgrowth was present or not. In the preceding paper (Child, ’11b), 
which was almost wholly concerned with my earlier work, only these three types ot 
anterior regulation were distinguished. As my work went on, however, it became 
evident that different kinds ot teratophthalmic heads occurred and that many of the 
pieces which failed to form eyes and auricles were nevertheless not strictly speaking 
headless. 1 found it particularly desirable to distinguish those cases in which only 
the eyes were abnormal from those in which the form of the head was also aimormal, 
The term ‘teratophthalmic’ was therefore used for the former type and the term 
*teratomorphic’ for those cases in which the anterior median region of the head was 
reduced in size or failed to develop so that the auricles were brought close together or 
fused on the front of the head. Similarly the cases where eyes did not appear were 
grouped under two heads, the ‘anophthalmic’ pieces in which a distinct outgrowth 
of new tissue developed at the anterior end, and ‘headless’ in which the growth of 
new tissue was limited to closure of the wound. For most purposes this grouping 
of the results is sufficient, but in certain cases where a more extended anatysis of 
the conditions determining the formation of eyes is the object in view the terat- 
ophthalmic type must be still further subdivided. 

In a recent summary of certain parts of my work (Child, ’l1ld) the five types 
mentioned above are described and figured. 


192 Cc. M. CHILD 


ceding paper constitute the basis for the conclusion that a dynamic 
gradient of some sort exists along the chief axis of the planarian 
body. 


2. Regional differences in the position of the pharynx 


The pharynx serves as a visible landmark for a region of the 
body which possesses certain functional characteristics. In the 
contraction and extension of the planarian body the prepharyngeal 
and postpharyngeal regions behave differently. In the former 
contraction occurs in such a manner that the intestinal contents 
are in general forced posteriorly, while in the latter the intestinal 
contents move anteriorly as contraction occurs. In extension the 
contents move in the reverse directions in the two regions. In 
other words, the pharynx appears at the posterior end of a charac- 
teristic region, the prepharyngeal region. The longer this region, 
the farther from the head does the pharynx appear and vice versa. 

When we compare pieces of the same length from different 
regions of the body, we find that the prepharyngeal region is 
longest and the new pharynx farthest from the head in the most 
anterior pieces (fig. 2) and that the length of the prepharyngeal 
region decreases and the pharynx becomes more anterior in succes- 
sive pieces from the anterior regions of the body backward (fig. 3). 
In general then, there is a very definite relation between the rate 
of head formation, the length of the prepharyngeal region and the 
position of the pharynx; the greater the rate of head formation 
the longer the prepharyngeal region and the more posterior the 
pharynx and vice versa. 


3. The later course of regulation 


It has commonly been stated that pieces of Planaria gradually 
acquire in the course of regulation the proportions of the original 
animal. Such a statement can be accepted as true only in a very 
general sense. For example, the proportions of young small 
worms are different in various ways from those of old large ani- 
mals. In general, the shorter the piece the more closely do its 
proportions after regulation approach those of the young animal 
and vice versa. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 193 


But the point of chief importance in the present connection is 
that the change in proportions which occurs during the later stages 
of regulation is a very different process in pieces from the old 
prepharyngeal region from that in pieces from the old post- 
pharyngeal region. Under anything approaching natural con- 
ditions the change in proportions in the postpharyngeal piece 
is rapid (figs. 10 and 11). The head increases in size rapidly after 
its first appearance and the pharynx seems to migrate posteriorly 
because of the elongation of the new prepharyngeal region at 
the expense of the postpharyngeal region. Sooner or later the 
animal cannot be distinguished from a normal worm of medium 
size. Changes of this character occur in pieces from the post- 
pharyngeal region whether food is given or not, but in the presence 
of agents which decrease the rate of metabolism and consequently 
the rate of formation of the head the changes in proportions are 
also retarded and in extreme cases are limited to the newly formed 
prepharyngeal region of the piece (figs. 13, 18). 

In pieces from the old prepharyngeal region on the other hand, 
whether they retain the old head or develop a new one, we find that 
the short new postpharyngeal region (fig. 2) grows very slowly 
when no food is given. If such pieces are not fed, we usually 
find, that even after months the head is disproportionately large 
and the prepharyngeal region disproportionately long. 

Briefly stated then the facts are these: a new prepharyngeal 
region formed in pieces arising from the old postpharyngeal 
region attains full size even in starving animals, but a new post- 
pharyngeal region formed in pieces from the old prepharyngeal 
region remains indefinitely of relatively small size unless an 
excess of nutritive material is present. 

These facts seem to me to indicate that in growth as well as in 
localization and differentiation the prepharyngeal region is domi- 
nant over the postpharyngeal. If we carry the experiments of 
this character farther and compare different parts of the pre- 
pharyngeal region with respect to the degree of their domir ance 
over more posterior regions, we shall find that the more anterior 
prepharyngeal regions are more completely dominant than others 
over newly formed postpharyngeal regions. In other words, the 


194 Cc. M. CHILD 


growth of the new postpharyngeal region at the expense of the 
old prepharyngeal region proceeds more slowly and ceases at an 
earlier stage in pieces from the most anterior region than in pieces 
farther from the old head. 

My observations upon these points include hundreds of pieces 
and a large number of measurements. I have not attempted to 
give the data in full because I believe that all who are familiar 
with the course of regulation in Planaria and related forms will be 
able to confirm my statements. Morgan observed this relation 
between prepharyngeal and postpharyngeal regions in his earlier 
work on Planaria (Morgan, ’00, pp. 62-3) but merely recorded the 
fact without attempting to interpret it or to connect it with other 
facts. 


4. The dominance of anterior regions in the regulatory development 
of pharynx and intestine 


The facts to be considered under this head are briefly these: 
pieces from the postpharyngeal region of the first zoéid of Planaria 
never develop a new pharynx or prepharyngeal region except in 
cases where some approach to the formation of a head occurs 
at the anterior end. On the other hand, pieces containing any 
part of the original pharyngeal or prepharyngeal region always 
develop a new pharynx—or retain the old—whether a head 
develops or not. 

All who have worked with Planaria are familiar with the fact 
that under the usual conditions of experiment in all pieces where 
regeneration of a head occurs a pharynx also develops except in 
cases where the piece retains the old pharynx, or when it is so 
small that it produces a ‘tailless head.’ The differences in such 
pieces as regards rate of development of the head and the length 
of the prepharyngeal region have been noted above. 

But the pieces which fail to produce a head differ widely in 
their ability to produce a pharynx. During 1905 I made an 
extensive study of short pieces in which large individuals were cut 
into from ten to twenty-four pieces, a record being kept of the 
approximate region of the body from which each piece came and 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 195 


of the history of each individual piece. These experiments 
established the fact considered briefly in the preceding paper 
(Child, ’11b), viz., that the capacity for forming a new head 
depends to some extent upon the length of the piece and that this 
factor in head formation becomes increasingly important with 
increasing distance of the pieces from the anterior end of the 
original worm back to the region of fission. In addition to this 
evidence for the existence of an axial gradient of some sort, the 
experiments established the important fact the pieces which fail 
to produce a head do not produce any regions of the body anterior 
to their original level. My records show that in every case where 
pieces from the prepharyngeal or pharyngeal region remain head- 
less they nevertheless develop a new pharynx, or the old pharynx 
remains: even pieces which include at the anterior end only the 
old mouth and a very small part of the pharyngeal region such, for 
example, as ac, fig. 6, are capable of developing a new pharynx 
whether a head develops or not. But pieces from a region slightly 
farther posterior than this which includes no part of the pharyn- 
geal region (bd, fig. 6) never produce a new pharynx when they 
remain headless. If, however, some slight approach to head 
formation occurs a new pharynx always appears in such pieces. 
Further experiments in later years confirmed these results. My 
earlier records concerning this point include some fifty pieces. 

During the present year I prepared a series of pieces for a fur- 
ther test concerning this point. Well-fed worms 15 to 18 mm. in 
length were used and from these pieces like bd, fig. 6, were cut, 
50 such pieces being prepared. The results are given inpercen- 
tages in the following table: 


NORMAL | TERATOPHTHALMIC T naroMonrne ANOPHTHALMIC | HEADLESS 


wis Sol eas ee al 
| 
With pharynx..... 4 8 | 0 
Without pharynx. 0 | 0 46 
4 | 8 | 10 32 «| 46 


The table shows that all of the normal, Renate and 
teratomorphic pieces developed new pharynges. Of the 50 pieces 


196 C. M. CHILD 


Qos 


9 


Fig.6 Outline of pharyngeal and postpharyngeal region 
showing various levels of section. 


Figs. 7-9 ‘Anophthalmic’ pieces from posterior region 

of first zodid. Fig. 7, anterior outgrowth small, no 

x pharynx formed and no anterior regulation of the alimen- 

6 tary tract. Figs. 8 and 9, anterior outgrowth larger, new 
pharynx present, short prepharyngeal intestine formed. 


16, 32 per cent, are anophthalmic, i.e., they show some outgrowth 
at the anterior end but do not develop eyes. The amount of the 
anterior outgrowth differs widely in different pieces. Of these 
anophthalmie pieces 13, 26 per cent of the whole series and more 
than three-fourths of this type, developed new pharynges, while 
only 38 pieces, 6 per cent, remained without pharynges. More- 
over, my records show that in every case where a pharynx failed 
to develop in these anophthalmic pieces the anterior outgrowth 
was small like that in fig. 7 or even smaller, while in all but two 
cases where a pharynx was formed the outgrowth was visibly 
greater and of the types shown in figs. 8 and 9. In short it was 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 197 


evident from examination of the pieces that in general those 
which developed pharynges were less completely headless than 
those which did not. 

The regulatory changes in the intestine run parallel with the 
other changes: in those cases where a new pharynx is formed a 
short median prepharyngeal intestinal axis is formed, while in 
those pieces which do not give rise to a pharynx noappreciable 
regulatory changes in the intestinal tract occur. Figs. 8 and 9 
show the intestinal structure in pieces which develop a pharynx 
and fig. 7 that in pieces without a pharynx. It is sufficiently 
clear that the development of a pharynx is associated with the 
presence or new formation of a prepharyngeal region. 

The strictly headless pieces of the series, twenty-three in num- 
ber, 46 per cent of the whole series, did not in a single case give 
rise to a new pharynx. The alimentary tract in these pieces is 
essentially similar to that in fig. 7. It should perhaps be added 
that in the examination to determine the presence or absence of 
a pharynx each piece, after it had been kept long enough for regu- 
lation to proceed as far as it it would go—twenty days at 20° C.— 
was subjected to compression under the microscope. By this 
means it is possible to detect the pharynx even when it is very 
minute and otherwise quite invisible. My records also include 
hundreds of headless pieces from the pharyngeal and prepharyn- 
geal regions and in no single case among them where the old 
pharynx was not retained has a pharynx failed to develop. 

These facts seem to me to be of great importance in connection 
with the question of dominance of anterior over posterior regions. 
They show that where a part of the pharyngeal or prepharyngeal 
region is present a new pharynx develops whether a head forms or 
not, while in the postpharyngeal region of the first zodid no new 
pharynx appears unless at least a start toward head formation 
occurs. Evidently the process of head formation determines at 
a very early stage the localization of a new prepharyngeal and so 
of a pharyngeal region. In short a new pharynx appears in an 
isolated piece only when some part of the original pharyngeal or 
prepharyngeal region is included in the piece or is formed anew in 
the course of regulation. And furthermore, the formation of a 


198 C. M. CHILD 


new prepharyngeal and pharyngeal region is dependent on the 
formation of a head or something which approaches a head in 
physiological character. Discussion of the question as to the 
nature of this capacity of the head region to determine the devel- 
opment of other regions and of the capacity of regions at any level 
to determine regions posterior to them, in short, the question as 
to the nature of antero-posterior physiological dominance, 
requires the consideration of other data of different character 
from those presented here and must therefore be postponed. 


5. The position of the pharynx under experimental conditions 


A. In pieces from the old postpharyngeal region. The relation 
described above between the region of the body from which the 
piece is taken, the size and rate of development of the head and 
the position of the pharynx is by no means fixed in character, and 
can be altered very readily by a great variety of experimental con- 
ditions. 

In describing a few experiments of this character we may con- 
sider first pieces from the original postpharyngeal region: in such 
pieces the process of regulation involves the formation of a pre- 
pharyngeal and pharyngeal region from the anterior part of the 
old postpharyngeal region. In such pieces we can alter the 
size and rate of development of the head, the length of the pre- 
pharyngeal region and the position of the pharynx almost at will. 

For experimental purposes pieces including the whole post- 
pharyngeal region (bz, fig. 6) afford the most striking results, 
though shorter pieces from the posterior region of the first zodid 
may be used. As a basis for comparison I have used in my 
experiments the postpharyngeal regions isolated from worms 
15 to 18 mm. in length which have been sufficiently well fed to 
keep them from either decreasing or increasing in size to any 
marked extent. Regulation of these control pieces occurred at a 
temperature of about 20° C. in water containing an excess of 
oxygen for the requirements of the animals. The results from 
such pieces are uniform in high degree. The pieces produce a 
large head with normal eyes and the pharynx appears at about 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 199 


ve 
| 


LO 


Figs. 10 and 11 Postpharyngeal pieces showing regulation in water at tempera- 
ture of 20° C. Fig. 10, early, fig. 11, later stage. 


one-third of the length of the piece from the anterior end. Fig. 10 
shows an earlier, fig. 11 a later stage in the regulation of such a 
piece. The only appreciable variation from this result which I 
have observed, when material and conditions are as above stated, 
is the very rare occurrence of partly fused eye spots. In ten 
series of 50 pieces each, a total of 500, 497 pieces were essentially 
similar to fig. 11, while 3 pieces (0.6 per cent) differed only in 
possessing partly fused eye spots. It is evident from these data 
that these pieces under constant conditions give a very definite 
characteristic result. 

Similar pieces from worms in similar condition which undergo 
regulation in alcohol or ether or other anesthetics at the same 


200 Cc. M. CHILD 


72 IZ Ig 


Figs. 12-14 Postpharyngeal pieces showing relation between regulation and 
external conditions. Figs.12 and 13, regulation in alcohol 1.5 atter twetve days. 
Fig. 14 shows the changes in these pieces after return to water. 


temperature as the controls give strikingly different results. My 
records include more than a hundred such pieces and show uni- 
formly that the head is always smaller, develops more slowly than 
in the controls, and is very commonly teratophthalmic and often 
teratomorphic: in most cases also redifferentiation plays a larger 
part in the formation of the ,head than in the controls. The 
pharynx is also small and is always much nearer the anterior end 
than in the controls. Figs. 12 and 13 show two characteristic 
pieces after twelve days in 1.5 per cent alcohol. In fig. 12 the 
head is more nearly normal and a small pharynx is present: in 
fig. 13 the head is almost entirely the result of redifferentiation and 
possesses only a single eye in the median line and the pharynx 
is even smaller and nearer the anterior end than-in fig. 12. It 
should also be noted that the changes in proportion in these 
pieces extend posteriorly only a short distance from the head. If 
these pieces are kept in aleohol they remain in approximately the 
condition figured until death. If, however, they are returned to 
water the development of the head proceeds, it becomes larger, the 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 201 


prepharyngeal region becomes longer and the elongation and 
decrease in width gradually extend posteriorly (fig. 14). 

The intestinal regulation in these pieces occurs only in the region 
anterior to the pharynx, whether this region is long or short. In 
eases like figs. 12 and 13, for example, we find a prepharyngeal 
median axial intestine in the short prepharyngeal region and with 
the elongation of this region after the return to water (fig. 14) the 
prepharyngeal axial intestine also elongates. The pharynx is 
then an excellent morphological and physiological landmark. 

Pieces which regulate in ether give very similar results. <A 
series,of postpharyngeal pieces after eighteen days in five-tenths 
per cent ether showed forms ranging from the type of fig. 12 to 
that of fig. 15. In the extreme type of fig. 15 anterior regenera- 
tion is limited to the extreme tip of the head region, other parts of 
the head being the result of redifferentiation. No auricles are visi- 
ble, only a single eye is present, there is no elongation and change 
in proportions and no pharynx is present. In such pieces no intes- 
tinal regulation occurs and no prepharyngeal axial intestine is 
formed from the union of the lateral postpharyngeal branches in 
the anterior region of the piece. When a pharynx appears the 
prepharyngeal axial intestine is always present, though it may be 
short. In eases like fig. 15 it is evident that practically all regu- 
lation except the redifferentiation of the head region has been 
inhibited by the ether. The fact that this can go on under con- 
ditions which stop allother regulatory processes is itself suggestive. 
Its bearing will be discussed later. 

If such pieces are returned to water extensive changes occur. 
During the first four or five days the head region enlarges and a 
pharynx appears near the anterior end of the piece (fig. 16), but 
the prepharyngeal region soon begins to elongate and becomes 
more slender as the activity of the head increases. After eight 
days in water the pieces resemble fig. 17. The postpharyngeal 
region is not yet under complete control and drags along behind 
the rest of the worm like a dead mass, except when the animal is 
very strongly stimulated. The prepharyngeal region, on the 
other hand, resembles that of the pieces in water. Incidentally 
it may be noted that when the piece develops only one median 


202 Cc. M. CHILD 


fe) v40) 


IS t9 
Figs. 15-19 Postpharyngeal pieces showing the effect of ether and metabolic 
products on regulation. Fig. 15 regulation in ether 0.4%. Fig 16, four days in water 
after ether. Fig. 17, eight days in water after ether. Figs 18 and 19, regulation 
in water from planarian cultures. 


eye in alcohol or ether it usually develops two more in the normal 
position (fig. 17) after its return to water. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 203 


These cases of regulation in alcohol and ether show very clearly 
that when the formation of a head at the anterior end of a piece 
is retarded by external conditions the length of the prepharyngeal 
region is decreased and the pharynx arises nearer the head: in 
extreme cases regulation is confined to the head region alone and 
neither pharynx nor prepharyngeal region appear. On returning 
these pieces to water we can see that the regulatory changes pro- 
ceed in the posterior direction. 

It is possible to alter the rate of formation and the size of the 
head and the position of the pharynx by various other means. 
For example CO, and other products of metabolism in the water 
in which the pieces are kept are very efficient factors in altering 
the course of regulation. The following series will serve as an 
example: 50 postpharyngeal pieces (bz, fig. 6) were placed in old 
culture water in which a stock of several hundred worms had been 
kept for nine days. Since this experiment was merely prelimi- 
nary, no attempt was made to determine whether the CO, or 
other substances were the more important factors in determining 
the results: more exact data will appear in a later paper. As 
a control 50 similar pieces were placed in fresh, well aérated 
water at the same temperature. The results in percentages, so 
far as they concern our present purpose, are as follows: 


HEADS | NORMAL EYES PHARYNGES 
Iniold:ewltureswatert.- cass se sec eens 54 6 48 
Jn fresh water........... 2 Ak a ee ee 100 100 100 


The differences are evident at once. Certain features of the 
results however are not apparent from the table. In no case did 
a pharynx appear in a piece which did not form a head and some 
of the pieces which formed the smallest and most abnormal heads 
did not produce pharynges at all. In all cases where a head 
formed it was relatively small and was commonly teratophthalmic 
or teratomorphic and the pharynx when present was only a short 
distance posterior to it. Fig. 18 shows a characteristic piece from 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 3 


204 C. M. CHILD 


this series and fig. 19 a more extreme case with single median eye, 
teratomorphic head and without a pharynx. 

Differences which are similar in character though less extreme 
may be obtained with different temperatures. At high tempera- 
tures postpharyngeal pieces produce large heads with normal eyes 
and relatively long prepharyngeal regions. At lower tempera- 
tures the heads are smaller and more frequently teratophthalmic 
and the prepharyngeal regions shorter, i.e., the pharynges are 
further anterior. 

As regards still other methods by which similar results may be 
obtained, further data will be presented at another time. 

The important point for present purposes is that in a piece 
from the original postpharyngeal region the rate of development 
and the size of the head, the length of the new prepharyngeal region 
and consequently the position of the new pharynx all vary with the 
rate of the dynamic processes concerned with the formation of the 
new head and can be controlled experimentally. 

B. In pieces from the old prepharyngeal region. Internal con- 
ditions are very different in the pieces from the original pre- 
pharyngeal region from those existing in the postpharyngeal 
pieces. In prepharyngeal pieces the greater portion remains 
prepharyngeal in structure and function and only a short pos- 
terior portion becomes postpharyngeal. At the posterior end of 
the newly determined prepharyngeal region the new pharynx 
appears (fig. 20). Here then a prepharyngeal region exists from 
the beginning and it is merely a question as to how much of it 
shall become postpharyngeal. 

A superficial consideration of this case might perhaps lead us 
to expect from analogy with postpharyngeal pieces that a decrease 
in the rate or intensity of the dynamic processes in the head region 
would result here also in a decrease in length of the new prepharyn- 
geal region or more strictly speaking in the determination of a 
shorter prepharyngeal region and the consequent localization of 
the pharynx more anteriorly. This result would occur if the head 
region alone were dominant over all other regions and these latter 
co6rdinate in character. But as a matter of fact each level of the 
body is dominant in some degree over the levels posterior to it 
so far as they are within the effective distance. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 205 


It was shown in the preceding section that in the pieces from 
the original prepharyngeal region the formation or presence of a 
head is not necessary as it is in postpharyngeal pieces for the 
development of a new pharynx. The presence of any part of the 
original prepharyngeal or pharyngeal region is sufficient to deter- 
mine the formation of a new pharynx whether a head forms or 
not. In such pieces the localization of the new pharynx will 
depend on the length of the region which undergoes redifferen- 
tiation to form the missing posterior regions, for the new pharynx 
appears on the boundary between the region which retains its 
original prepharyngeal character and that which becomes post- 
pharyngeal in character. Since the prepharyngeal. region is 
dominant over regions posterior to it, pieces from the prepharyn- 
geal region develop the new pharynx near their posterior ends 
(fig. 2) and the new postpharyngeal region is formed almost wholly 
by regeneration and remains short for a long time unless the 
animal is fed. But when we compare prepharyngeal pieces from 
different levels we find that as the level of the pieces approaches 
the level of the old pharynx, i.e., the posterior end of the pre- 
pharyngeal region, the degree of dominance of the anterior 
regions of the piece over its more posterior regions decreases, a 
fact that will appear more clearly later, and in the course of regu- 
lation the new: postpharyngeal region becomes longer and the 
pharynx appears more anteriorly in the piece. 

In P. dorotocephala the degree of dominance of the prepharyn- 
geal over the postpharyngeal region is such that the formation of 
a new postpharyngeal region in a prepharyngeal piece occurs 
slowly, is limited to the extreme posterior region of the piece and 
is almost wholly a process of regeneration in the stricter sense. 
Consequently in all pieces from the prepharyngeal region the new 
pharynx arises at or very near the posterior end of the old tissue. 
As a matter of fact its level does differ somewhat according to the 
level of the piece. In pieces from the extreme anterior region of 
the body (ab, fig. 1) the new pharynx may appear in the posterior 
new tissue (fig. 20), i.e., in such cases not only the postpharyngeal 
region but the pharyngeal region also is the product of regenera- 
tion. As the level of the piece becomes more posterior the new 


206 Cc. M. CHILD 


o, 


V 


20 27 


Figs. 20 and 21 pieces from different parts of the prepharyngeal region to show 
the position of the pharynx. Fig. 20, from the region immediately behind the old 
head. Fig. 21, from the region just anterior to the old pharyn. 


pharynx appears more anteriorly until at a level just anterior 
to the old pharynx the new pharynx appears in the old tissue 
somewhat anterior to its posterior end (fig. 21). In P. maculata 
the corresponding differences in the level of the new pharynx 
are somewhat greater since the axial gradient in this species 
differs in certain respects, as will appear later, but in both cases 
the principle involved is the same. 

In general these differences in the level at which the new 
pharynx appears in the prepharyngeal pieces are merely expres- 
sions of the different degrees of stability of different prepharyn- 
geal levels. The most anterior levels of the prepharyngeal region 
are more fixedly prepharyngeal than other levels and a new post- 
pharyngeal and pharyngeal region arise only from their most 
posterior parts. In fact, if we isolate very short pieces from just 
behind the old head they give rise in most cases under standard 
conditions, i.e., when taken from large well fed worms and 
allowed to regulate in well aérated water at a temperature of 20 
C., to ‘tailless heads’ (fig. 4) in which neither pharyngeal nor 
postpharyngeal regions are formed, the new tissue at the posterior 
end ofthe piece being limited to closure of the wound, while ante- 
riorly such pieces produce large heads. At more posterior levels 
the pharyngeal and postpharyngeal regions are formed and as 
the level of the piece in the original body becomes more and more 
posterior they become larger and the new head becomes smaller. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 207 


The question as to why a new postpharyngeal region should 
arise at all in such a piece was discussed in connection with the 
analysis of regulation in another form (Child, 05). In that form, 
Cestoplana, almost no regeneration occurs posteriorly and the 
new postpharyngeal region arises from the old prepharyngeal re- 
gion by redifferentiation and its length differs at different levels in 
the same way asin Planaria. In my account of regulation in that 
form I called attention to the fact that the first visible change in 
the process of formation of a postpharyngeal region from a part 
of the original prepharyngeal region is not a change in structure 
but a change in function, in the motor reactions of the part. The 
visible structural changes follow this functional change and they 
occur, as I believe, in consequence of the altered dynamic condi- 
tions in this region of the piece. 

In Planaria, however, where the new postpharyngeal region is 
formed largely or wholly by regeneration conditions are somewhat 
different and we must consider the question as to why a new post- 
pharyngeal region should arise by regeneration at the posterior 
end of a prepharyngeal piece. 

In the first place a cut surface, a new terminal region, has been 
formed by the operation and that means that certain character- 
istic metabolic conditions and functional conditions in the stricter 
sense must exist or arise in the posterior region of the piece. 
These result first of all in dedifferentiation, i.e., loss of the old 
structural characteristics and renewed or accelerated growth of 
the cells near the wound. In this way a small region of more or 
less ‘embryonic’ tissue is formed. ‘The later fate of this tissue 
depends mainly upon its correlation with parts anterior to it, 
for if the law of the dominance of anterior over posterior levels 
holds good in this case, and I shall show later that it does, this 
region is subordinate to the regions anterior to it and the course 
of its development is determined chiefly by those more anterior 
regions. The correlative factors to which this region is subjected 
are essentially similar to those to which the old postpharyngeal 
region was subjected and it develops into a postpharyngeal region 
in consequence of the influence upon it of these factors. The rate 
of metabolism of the dedifferentiated cells which form the starting 


208 C. M. CHILD 


point of regeneration is accelerated by the process of dedifferentia- 
tion which is essentially a process of rejuvenescence (Child, ’11c) 
and in consequence of this acceleration of metabolism these cells 
are able to grow and divide for a time at the expense of the regions 
anterior to them. This growth continues until in the course of 
redifferentiation the cells again grow physiologically older and 
their rate of metabolism decreases to such an extent that they can 
no longer grow at the expense of more anterior regions and there 
regeneration ceases unless the piece contains an excess of nutri- 
tive material. In short the conditions which determine the regen- 
eration of a postpharyngeal region in a prepharyngeal piece are: 
first, the dedifferentiation of the cells in reaction to the wound 
and the resulting increase in capacity for metabolism and growth; 
second the correlative factors to which they are subjected in con- 
sequence of their physiological continuity with more anterior 
regions and which determine that growth and redifferentiation 
shall actually occur. In consequence of all these conditions a 
new postpharyngeal region arises at the posterior end of the pre- 
pharyngeal piece, while other parts of the piece remain almost 
unchanged structurally and functionally. The growth of the 
postpharyngeal region after its determination as such is then 
essentially a process of ‘functional’ growth and the final cessation 
of growth is the result of an equilibration in the rate of metabo- 
lism between the new part and the old. 

The occurrence of ‘tailless heads’ in very short pieces from 
extreme anterior regions, in other words, the failure of such pieces 
to form pharyngeal and postpharyngeal regions depends upon the 
fact which will be demonstrated later that a higher rate of me- 
tabolism is concerned in the formation of a head than in that of a 
posterior part: in these short pieces the posterior regions do not 
form because the process of head formation with its higher rate 
of reaction uses up the available material so rapidly and to such 
an extent that the formation of a posterior end is inhibited or 
rather prevented. With the aid of proper experimental condi- 
tions it is possible to produce tailless heads in longer pieces as well 
as in pieces from other regions of the body. In Planaria the 
dominant morphogenic reaction, i.e., the morphogenic process 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 209 


with the highest rate of metabolism is that of head formation, 
just as in Tubularia the dominant morphogenic reaction is that of 
hydranth formation—moreover, in Tubularia the distal hydranth 
regions are dominant over the proximal—and whenever this domi- 
nant reaction begins in a piece the other reactions are controlled 
by it and if the amount of material available as a source of energy 
is insufficient the subordinate reactions are decreased or completely 
inhibited. In short pieces of the Tubularia stem, particularly in 
those from the more distal regions, the product of regulation is 
more and more exclusively distal in character, the shorter the 
piece. Similarly in the anterior regions of Planari& the product 
of regulation becomes more and more exclusively anterior as the 
length of the piece decreases. Certain apparent exceptions to 
this law, e.g., the formation of headless pieces and biaxial ‘heter- 
omorphic’ tails in Planaria and other forms will be considered 
later and will be shown to be only apparent exceptions. 

The origin of a new dominant part like the head from any sub- 
ordinate part, i.e., any region posterior to the head is a problem 
of somewhat different nature and one which has not been consid- 
ered in any of the attempts at analysis of the process of form regu- 
lation which have been made, because the dominance of this 
region has not heretofore been recognized. My experiments 
have led me to the conclusion that the formation of the head in 
Planaria, the hydranth in Tubularia, etc., are in no sense restora- 
tions of missing parts, restitutions or anything of that kind, but 
rather that the new head or the new distal region as the starting 
point of a new individual arises from the mass of old tissue in a 
manner closely comparable to the formation of a new bud from 
differentiated tissue in a plant The new individual, which is 
at first merely a head, lives at the expense of the old parts and at 
the same time makes them over into parts of a new worm or uses 
the energy which it obtains through their destruction in the 
development of new parts. The new individual simply forms 
and grows head first out of the old mass. In Planaria the position 
of the new head commonly shows a definite relation to the old 
axis, though this is not always the case, but in various coelenter- 
ates, e.g., Corymorpha (Child, ’1la, pp. 112-119) and Harenactis 


210 C. M. CHILD 


(Child, ’09, ’11a) the new axes may arise ‘adventitiously.’ In 
other words, if the original axial gradient is sufficiently obliterated 
or if external conditions are sufficiently powerful to overcome its 
influence new gradients, i.e., new axes may arise without definite 
relation to it. 


CONCLUSION AND SUMMARY 


Although the consideration of the question as to the nature of 
the dominance of anterior over posterior regions in Planaria must 
be deferred until further facts have been presented, we may at 
this time consider briefly the significance of such dominance. 

The dominance of one region over another is of course relative 
rather than absolute. To say that a given region is dominant over 
others means merely that it influences and determines the proc- 
esses and conditions in them to a greater extent than it is influ- 
enced by them. ‘This, as will appear in later papers of this series, 
is the case in Planaria. 

In the present paper certain of the facts of regulation in Planaria 
have been presented which indicate that the head region controls 
and determines the development of regions posterior to it to a 
very large extent and that each level of the body is to a certain 
extent dominant over more posterior levels. This, however, is 
only the first step in the presentation of evidence. 

If the head region of Planaria and the distal region of Tubularia 
and various other forms are dominant over other regions as I 
have maintained, it follows that these regions develop more 
independently than any other part of the body. The formation of 
the head in Planaria, of the distal region in Tubularia, is the most 
fundamental, the most characteristic morphogenic reaction of 
the protoplasm of those species. Other reactions depend upon 
this to a much greater extent than it depends upon them. In 
general the first morphogenic regulatory change in an isolated 
mass of planarian or tubularian protoplasm is the formation of a 
head or a distal region or the initiation of this process. This 
fact will become more and more apparent as further experimental 
data are presented. As we know from numerous experiments, 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 211 


many isolated pieces of the planarian body appear to be incapable 
of producing heads. In such pieces I have been able to induce 
head formation experimentally by simply subjecting them to 
conditions which increase the rate of metabolism (Child, ’11d) and 
can demonstrate that in these pieces the absence of the head 
under the usual conditions is due, not to absence of the necessary 
‘organization’ or to lack of certain ‘formative substances’ or 
anything of that character, but merely to an insufficiently high 
rate of metabolism in the region concerned. By increasing the 
rate sufficiently in any way heads appear on pieces which would 
not otherwise produce them. By altering the conditions in the 
opposite direction it is also possible to induce the formation of 
teratophthalmic in place of normal heads or to inhibit head for- 
mation completely. 

But little attention has been paid to the matter of rate of reac- 
tion as a factor in ontogenetic and regulatory morphogenesis. 
When pieces fail toregenerate certain parts it has usually been taken 
for granted that the necessary ‘organization’ is lacking. This, 
however, is by no means always the case. The formation of a 
new whole from a piece or the failure to form such a whole, as 
well as the character of the whole formed, e.g., normal, teratoph- 
thalmic and teratomorphic wholes in Planaria may be the result 
of differences which are fundamentally purely quantitative rather 
than qualitative in nature. Undoubtedly differences in organiza- 
tion do exist and do play a part in many cases, but they are cer- 
tainly not the only nor the chief factors in many other cases. 

In the absence of the head region the most anterior regions 
present are dominant over more posterior regions within a cer- 
tain distance and to a certain degree. In general we find that 
while any region is a very important factor in determining what 
shall go on in regions posterior to it, it has but little influence, 
though it can be shown to possess a certain amount, in determin- 
ing what goes on in regions anterior to it. The morphological 
characteristics of Planaria are determined chiefly by the correla- 
tive influence of more anterior upon more posterior regions. 

It is evident that this point of view gives a very different con- 
ception from that generally held of the process of formation of a 


212, Cc. M.. CHILD 


new whole from a piece of a planarian body. Instead of being 
a process which shows almost infinite possibilities of adjustment 
to the conditions existing in a particular case, it is essentially 
one and the same reaction in all cases. In pieces where the head 
is present the posterior parts arise in consequence of correlation 
with more anterior parts. Where the head is absent a new head 
arises from the region involved in reaction to the wound, provided 
merely that the rate of reaction in this region becomes sufficiently 
rapid. When the process of head formation attains a certain 
stage the new head region begins to dominate the rest of the piece 
and makes it over according to certain definite and unchanging 
laws. All that is necessary for the formation of a planarian is 
first a cell or a group of cells capable of initiating a characteristic 
series of reactions which result in what we calla head and second 
an excess of nutritive material as a source of energy for growth. 

To put the matter in still another form, it is not too much to 
say that the capacity for head formation is all that exists in the 
planarian egg, all that is inherited. Given this, together with an 
excess of nutritive material, which in the case of Planaria exists 
in the yolk cells and the characteristic form of Planaria must 
result. 

And this brings us to the question as to how far similar relations 
obtain in regulatory and the ontogenetic development of other 
forms. At present I desire only to call attention to certain 
points in this connection. Tubularia and Corymorpha are essen- 
tially similar to Planaria as regards the axial gradient. In the 
first place it has been shown by various investigators that if a 
* tubularian stem is cut into a series of pieces of equal length the 
oral hydranth develops most rapidly and is largest in the most 
distal -piece and its size and rate of development decrease with 
each successive piece from the distal end proximally. These 
differences are similar in character to the differences in head forma- 
tion at different levels of the first zodid in Planaria. 

Secondly, as regards the dominance of distal over proximal 
regions Tubularia likewise resembles Planaria. In asexual repro- 
duction in nature the factor of distance is apparent, 1e., the tip 
of the stolon gives rise to a new hydranth when it has reached a 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 213 


certain distance, varying with conditions, from the original distal 
region and has become to some extent physiologically isolated 
from the latter (Child, ’1la, pp. 95-96, 101-112). When pieces 
are cut from the stem only vigorous pieces give rise to stolons at 
their proximal ends (Child, ’07) and as the piece becomes less and 
less vigorous the terminal region of the stolon, or the proximal 
end of the stem if astolon was not formed, tkecomes physiologically 
isolated from the dominant distal region and gives rise to a hy- 
dranth. Axial heteromorphosis in long pieces of the stem is 
merely the result of physiological isolation of the proximal region 
of the piece in consequence of increasing weakness, i.e., decreasing 
rate of metabolism and therefore decrease in length of the stem over 
which the distal region is dominant (Child, ’1la, pp. 101-112). 

Furthermore, as the length of the piece cut from the stem 
of Tubularia decreases the parts to which the piece gives 
rise become more and more exclusively distal: the longer pieces 
produce hydranths and stems, somewhat shorter produce hy- 
dranths alone, still shorter only manubrium and distal tentacles 
and so on, until finally the shortest pieces give rise only to the 
distal region of the manubrium with the distal tentacles (Child, 
07). These facts demonstrate that the distal region is able to 
form without correlation with other parts, but the more proximal 
regions have never been seen to arise except in connection with 
considerable regions distal to them. 

Corymorpha is essentially similar to Tubularia in all these 
respects except that it does not reproduce asexually by transfor- 
mation of the end of the stolon into a hydranth. This difference 
is probably due to the fact that the proximal region of the stem of 
Corymorpha is a much more highly specialized region than in 
Tubularia. So far as data are available, the relations seem to be 
similar in many if not in all other hydroids. 

Rand (11) has recently stated that in Hydra the peristome 
region controls morphogenesis and I have found in Cerianthus and 
Harenactis an axial gradient of the same character as in Tubularia 
and Planaria and in these forms also the peristome region appears 
to be dominant in regulation. The distance factor is more diffi- 
cult to demonstrate in these actinians since the rate of metabolism 


214 Cc. M. CHILD 


at the proximal end is so low that reproduction does not occur 
under ordinary conditions even when these regions are physically 
isolated. 

As regards plants, it is a well-known fact that in at least most 
forms the apical region of the axis is to a very large extent domi- 
nant over more proximal regions and I have endeavored to show 
(Child, ’1la) that the factor of distance is in many cases a factor 
of great importance in reproduction in plants as well as in animals. 

These facts as well as many others which might be cited show 
very clearly that an axial gradient exists in a large number of 
organisms and that in many cases at least the apical or anterior 
region is dominant in regulation. 

Turning now to normal ontogeny, we find that in most if not 
in all animals the visible phenomena of development begin in the 
region of the so-called animal pole of the egg and that this region 
in most if not in all cases becomes the anterior, distal or apical 
region of the resulting organism. I believe that these facts in 
themselves are highly suggestive when taken in connection with 
the facts concerning the dominance of the distal and anterior 
regions in the regulation of various forms. In fact it seems prob- 
able that dominance of anterior or distal over posterior or proxi- 
mal regions is a very general law of organic development, not only 
in animals but in plants as well. A more extended consideration 
of this question will be undertaken elsewhere. 

Critics of such a conclusion will at once cite those cases in which 
different parts of the developing egg appear to be almost wholly 
independent of each other, e.g., the annelid and mollusk and the 
amphibian. It seems at least probable that in such cases the 
characteristics of the different parts once determined through 
correlation at a very early stage are stable to a high degree and do 
not change when the parts are isolated. Even in such cases the 
‘animal’ region of the egg may be dominant over other regions 
at some stage in the history of the egg. The experiments of 
recent years on the nemertean egg afford strong evidence in favor 
of the view that independence of parts in later stages is preceded 
by a condition in which at least certain parts are determined correl- 
atively. According to this view the eggs which show the extreme 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 215 


mosaic type of development are merely cases of early differentia- 
tion or of absence of regulatory capacity due to one cause or 
another, rather than a type of constitution fundamentally dif- 
ferent from the extreme correlative type. 

In regulatory reproduction we find all stages between the two 
extremes. For example, in Planaria simplicissima the first forma- 
tion of the tail region depends as in P. dorotocephala upon 
connection with more anterior regions, but when it is once formed 
this region is more definitely determined as a tail region than in 
P. dorotocephala and when isolated often gives rise, as Morgan 
first observed (Morgan, ’04), to a tail at its anterior end. Some- 
what similar conditions exist in the earthworm, except that there 
the ‘tail region’ includes the greater part of the body length. 

In Stenostomum, on the other hand, posterior regions remain 
posterior in structure only so long as they are under the complete 
control of the dominant head region. When this control decreases 
below a certain limit, either in consequence of increase in length 
of the animal or other conditions (Child, ’1la) a new head begins 
to develop in the posterior region of the body, even though organic 
continuity with the original head region is not interrupted. 

It seems probable then that we shall find in ontogenetic as well 
as in regulatory development that anterior or distal regions are 
very generally dominant over posterior or proximal regions, at 
least at some stage, and that the egg as well as the piece capable 
of regulation represents primarily the anterior or distal region, 
together with an excess of nutritive material or some means of 
obtaining such an excess as a source of energy for growth. 

The most important facts of the paper and the conclusions to 
which they point may be summarized as follows: 

1. As was shown in the first paper of this series, the most 
important regional differences in the course of regulation in pieces 
of Planaria taken in sequence along the axis from the anterior 
end posteriorly consist first, of decrease in the size of the head, the 
rate of its formation and the frequency of normal eyes; second, of 
decrease in the length of the prepharyngeal region and therefore 
the formation of the pharynx at a more anterior level of the piece. 
These differences indicate the existence of an axial gradient of 


216 Cc. M. CHILD 


some kind and their character suggests that this gradient concerns, 
at least in part, the rate or intensity of certain processes along the 
axis. 

2. A head or prepharyngeal region is capable of growing or of 
maintaining itself at the expense of more posterior regions in the 
absence of other food to a much greater extent than posterior 
regions can grow or maintain themselves at the expense of more 
anterior regions. These facts also suggest differences in the rate 
or intensity of certain processes at different levels. 

3. In pieces of Planaria dorotocephala which contain a part of 
the original prepharyngeal or pharyngeal region and in whichthe 
old pharynx is not present or does not persist, a new pharynx 
develops whether a new or old head is present or not. In pieces 
from the postpharyngeal region, on the other hand, a new pharynx 
or a new prepharyngeal region never forms unless a head region 
begins to form first. 

In general no piece of the planarian body is capable of giving 
rise to structures belonging to levels anterior to that which the 
piece originally occupied in the body, unless the formation of a 
head or the first stages of this process occur first. But a piece 
can give rise to parts characteristic of levels posterior to that which 
it originally occupied, whether a head region is present or not. 

4. The formation of a head at the anterior end of a piece may 
be retarded or inhibited by various agents and conditions, e.g., 
alcohol, ether and other anesthetics, CO. and other products of 
metabolism, KCN and even by low temperature and insufficient 
nutrition. In all pieces from the postpharyngeal region of the 
body in which the formation of a head is thus retarded the length 
of the new prepharyngeal region is less than in pieces under the 
usual conditions and the pharynx appears nearer the head. When 
the depressing agents or conditions are used in higher concentra- 
tion or intensity so as to produce the more extreme effects, or 
when the animals are in such physiological condition as to be more 
sensitive to their effects, the regulatory changes may be limited to 
the early stages of head formation, all regulatory changes being 
completely inhibited in regions posterior to the head. The fact 
that the regulatory processes concerned in head formation are 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 217 


able to go under conditions which completely inhibit all other 
morphogenic regulatory processes is of great importance. It 
constitutes very strong evidence in support of the conclusion that 
regulation at the anterior end of a piece is initiated by the process 
of head formation or by the beginning of this process. 

In general we find that in pieces from the original postpharyngeal 
region the size of the head the length of the new prepharyngeal region 
and in extreme cases its presence or absence, and consequently the 
position or presence or absence of the pharynx are all very closely 
correlated with the rate of the dynamic processes or certain of them 
which are concerned in head formation. 

5. In pieces from the original prepharyngeal region a new post- 
pharyngeal region develops much more slowly, is much shorter 
and is to a much larger extent a product of regeneration in the 
stricter sense than is a new prepharyngeal region developing in 
a postpharyngeal piece. The new pharynx in pieces from the pre- 
pharyngeal region appears near the posterior end of the old tissue. 
Only when excess of nutrition is present does the new postpharyn- 
geal region grow to the proportions characteristic of animals in 
nature. 

6. The facts point to the conclusion that the regulatory for- 
mation and development of the head in Planaria, as well as the 
hydranth in Tubularia and the anterior or distal regions of various 
other forms are not in any sense a restoration of missing parts, a 
‘restitution’ a return to a condition of wholeness or anything of 
that sort. Such a process consists rather in the formation 
of a new individual, beginning with the head or distal region: this 
new individual simply grows out of the mass of old tissue head first 
and as it grows either uses the old tissue as a source of energy or 
brings about its redifferentiation until the dynamic equilibrium 
characteristic of the specific system and the existing conditions is 
attained. The regulatory changes at the anterior end in such 
‘ases begin in the region of the developing head and proceed pos- 
teriorly. By means of external factors we can determine the 
distance from the new head to which they shall proceed and where 
they shall produce certain results. 


218 C. M. CHILD 


7. Attention is called to the fact that this dominance of anterior 
or distal over posterior or proximal parts is similar in character 
to the relations between parts which exist in most if not in all 
plants. The regulatory formation of a new head in Planaria or 
of a new hydranth in Tubularia or Corymorpha, together with 
the effects of such a process on other parts is not essentially 
different in character from the formation of a new apical cell or 
vegetative tip in a plant and the resulting development of the 
new axis. In the plant, however, the new dominant region is in 
most cases unable to bring about to so great an extent as occurs in 
many animals the redifferentiation of masses of old tissue adjoining 
it, consequently the new plant axis remains short and small 
except when excess of nutrition is present. A considerable degree 
of redifferentiation of the old tissue under the influence of the 
new vegetative tip does occur, however, in many cases. 

If my conclusions are correct the processes of form regulation 
in the animal and in the plant follow the same law and this law 
finds a general expression in the statement that any given region 
along the axis in which dynamic processes are occurring dominates 
more or less completely regions proximal or posterior to it and is 
dominated by regions anterior to it. 

8. The very general if not universal formation of the distal or 
anterior region of the organism from the animal pole of the egg or 
from some region near it and the fact that many larvae consist 
essentially of only the most anterior regions of the body, together 
with the fact that it is the animal pole which initiates ontogenetic 
development in those cases where a difference in time along the 
axis can be observed in the earliest stages, and finally the very 
general progression of morphogenesis in the posterior direction, 
all these facts as well as many others suggest that the dominance 
of anterior or distal over posterior or proximal regions is a very 
general law of organic life. When we review the facts now at 
hand it appears probable that all organisms, except perhaps the 
simplest, are fundamentally systems of this character. More- 
over, such a conception affords a most satisfactory and logical 
basis for a physico-chemical theory of development. Undoubt- 
edly in many eases secondary isolations of parts occur, new correl- 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 219 


ative factors arise and many other conditions combine to alter the 
original condition. But even in the adult organism the funda- 
mental fact appears in the dominance of the apical region of the 
plant and in the functional dominance of the head region in 
aninals. 

9. It follows further, if the above conclusions are correct that 
in all cases where development is of this type the process of inher- 
itance concerns primarily the anterior or distal regions. An iso- 
lated mass of protoplasm of a given species which is capable of 
continued existence and synthesis, no matter whether it is a mass 
of cells from the soma or an egg, represents primarily the domi- 
nant distal or anterior region, i.e., its specific type of reaction 
results always in the formation of this region and then if excess of 
nutritive material is present or obtainable so that growth may 
occur other parts arise in consequence of growth and of the correl- 
ative influence of the dominant region. In such cases the repro- 
ductive element, whether germ cell, somatic cell or mass of cells 
may represent merely a single specific reaction complex from which 
others arise as continued metabolism brings about the establish- 
ment of certain characteristic internal and external conditions. 

So far as the formation of new distal or anterior regions 1s con- 
cerned, the process of form regulation in such cases consists essen- 
tially first in the return or approach of somatic cells or cell masses 
to the specific type of reaction in consequence of isolation from a 
dominant part which had previously determined some modifica- 
tion of this type of reaction in these cells, second of the formation 
of a new dominant region in consequence of this change, and finally 
of the development of subordinate parts under the control of the 
new dominant region so far as material or energy is available for 
such development. 

In cases where only the formation of proximal or posterior parts 
is concerned form regulation consists first in the local reaction of 
cells to the absence of other parts, i.e., to altered correlative con- 
ditions, and second in the renewed growth and differentiation of 
such cells under the influence of more distal or anterior regions 
which dominate them. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 11, NO, 3 


220 


C. M. CHILD 


Experimental data which throw light upon the problem of the 
nature of the dominance of certain regions over others will be 
presented in a later paper. 


CHILD, C. 


\ 


BIBLIOGRAPHY 


M. 1905 Studies on regulation. rx. The positions and proportions of 
parts during regulation in Cestoplana in the presence of the cephalic gan- 
glia. Arch. f. Entwickelungsmech. Bd. 20. 

1907 An analysis of form regulation in Tubularia. 1. Stolon formation 
and polarity. Arch. f. Entwickelungsmech. Bd. 23. tv. Regional 
and polar differences in the time of hydranth formation as a special 
case of regulation in a complex system. Ibid. Bd. 24. v. Regulation 
in short pieces. Ibid. vi. The significance of certain modifications 
of regulation: polarity and form regulation in general. Ibid. 

1909 Factors of form regulation in Harenactis attenuata. ur. Regu- 
lation in ‘rings.’ Jour. Exp. Zodl., vol. 7. 

1910 Physiological isolation of parts and fission in Pianaria. Arch. 
f. Entwickelungsmech. Bd. 30 (Festband fiir Roux) Teil 2. 

1911a Die physiologische Isolation von Teilen des Organismus. Vortr. 
u. Aufs. u. Entwickelungsmech. H. 11. 

1911b Studies on the dynamics of morphogenesis and inheritance in 
experimental reproduction. 1. The axial gradient in Planaria doroto- 
cephala as a limiting factor in regulation. Jour. Exp. Zodl., vol. 10. 
191le A study of senescence and rejuvenescence based on experiments 
with Planaria dorotocephala. Arch. f. Entwickelungsmech. Bd. 31. 


-191ld_ Experimental control of morphogenesis in the regulation of 


planaria. Biol. Bull., vol. 20. 


Morean, T. H. 1900 Regeneration in planarians. Arch. f. Entwickelungsmech. 


Ranp, H. 


Bd. 16. 
1904 Regéneration of heteromorphic tails in posterior pieces of planaria 
simplicissima. Jour. Exp. Zodl., vol. 1. 


W. 1911. The problem of form in Hydra. Science, vol. 33, no. 845. 
In report of meeting of December, 1910, of Am. Soe. Zool., Hastern 
Branch. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 
AND INHERITANCE IN EXPERIMENTAL 
REPRODUCTION 


III THE FORMATION OF NEW ZOOIDS IN PLANARIA AND OTHER 
FORMS 


©. M. CHILD 
From the Hull Zoological Laboratory, University of Chicago 


THIRTY-SIX FIGURES 


CONTENTS 

1 The origin of new zodids in nature.. RUD SS Uda MaomOe ore duces Oe 222 
2 The formation of a second zodid meden Eperimental conditions. . 223 
3 Theresults of feeding and growth in teratomorphic anophthalmic and hee nae 

NESS)POIECES| cere ysis seers tveks ute eisesreresoreie s/2)cbsystscstsestererccorave she ences hs Giee oie Sena cpl 
4 The fate of nosterioe zooids in ‘Siae AEC) ha OT EC ES iA Bi cicearre nent are 235 
Hee Lhertormationof chains) of: 200108: ccc y. teiee Ae a -eseiaeate cata sind 242 
6 The factor of distance in the formation of new nonids! and its relation 

to the stage of development of the dominant region..................--. 256 
7 The disappearance of zodids in consequence of decreased distance from a 

headenerlonay eee ine rice les leis : F Firth rhein Ate DOORN CLL 
8 Conclusion and summary................. aa Sen eI ce rege ek: 3 
BU Tio eresayy tay severe ester ecaycs hey sct teh sc esta) os ee ay p= Peoe lake atsbeeuoterabs emt svarels sp oceis ots wre) sraxoxsr st ota 6 280 


In an earlier paper (Child, ’10) the relation between the act of 
fission and physiological isolation in Planaria was considered and 
it was shown that the process of fission may be controlled experi- 
mentally in various ways. The present paper is devoted to the 
question of the origin of new zoéids, their fate under various 
conditions and the rédle which the factor of distance from the 
dominant region plays in determining their formation. <A brief 
account of the formation of new zodids in Stenostomum is added 
to show that the process in this form follows the same general 
laws as in Planaria. 


221 


222, Cc. M. CHILD 


The purpose of the present paper is to show that the formation 
of new zodids at the posterior end of the body in Planaria and 
related forms is essentially a process of regulation, resulting from 
the physiological isolation of the region concerned from the 
dominant region in much the same manner as development of a 
new whole results from the physical isolation of a piece of the 
planarian body. 


1. THE ORIGIN OF NEW ZOOIDS IN NATURE 


The facts which indicate the presence of one or more zoéids in 
the posterior region of Planaria dorotocephala have been discussed 
in the first paper of this series (Child, ’11b) and elsewhere (Child, 
10). Up to the present time I have not had opportunity to 
examine newly hatched individuals of P. dorotocephala, but in 
P. maculata immediately after hatching, or in animals removed 
from the egg capsule by artificial means just before hatching I 
have been unable to discover any evidence that the second zodid 
is present. The usual method employed to determine the pres- 
ence of one or more zo6ids in the posterior region is the separation 
of this region into a series of short pieces: the pieces which repre- 
sent the anterior region of a zoéid or levels just anterior to it 
show a greater capacity to form heads than those pieces which 


Y) 


I 2, 


Figs. land2  Fig.1,newly hatched Planaria maculata. Fig.2, Posterior half of 
newly hatched worm, which remains headless when isolated. 


represent other regions. In the newly hatched worms the post- 
pharyngeal region is much shorter than the prepharyngeal (fig. 
1) and even the whole posterior half of the body, including the 
pharyngeal region usually remains headless when isolated (fig. 2), 
i.e., it resembles the posterior region of the first zoéid in larger 
animals. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 223 


At the stage just after hatching the animals are about 2 mm. 
in length but growth is rapid at ordinary temperatures and they 
attain a length of 4 or 5 mm. in two or three days. By the time 
they have reached this stage the second zodid is clearly present 
as is shown by the fact that the capacity of pieces to form heads 
increases suddenly a short distance posterior to the pharynx. 
Moreover, it is possible to induce fission experimentally in animals 
not much over 5 mm. in length (Child, ’10). 

It is evident from these facts that the second zoéid arises in the 
posterior region of the body during the course of development. 
After its formation it continues to grow more rapidly than the 
first zodid and its own posterior region may give rise to other 
zooids. The second zodid consists essentially in a more or less 
definitely localized region of increased capacity of head formation 
and posterior to this a gradientresembling the axial gradient in the 
first zodid (Child, ’?11b). In other words, it represents the first 
stage in the formation of a new dominant region like the head of 
the whole. 

If my conclusions concerning the réle of physiological isolation 
in reproduction (Child, 11a) are correct, the origin of the second 
zooid in Planaria may readily be conceived as the result of physio- 
logical isolation of the region concerned from the dominant head 
region. In consequence of such isolation the piece begins to 
develop into a new individual with its own dominant anterior 
region. Development is not complete so long as organic con- 
tinuity with more anterior parts persists because the isolation is 
not complete. If the formation of the second zodid actually 
occurs in this manner then it should be possible to induce or in- 
hibit the formation of a second zoéid experimentally by altering 
the degree of physiological isolation of the posterior region of 
pieces. That this may be accomplished is shown in the following 
section. 


2, THE FORMATION OF A SECOND ZOOID UNDER EXPERIMENTAL 
CONDITIONS 


If we isolate the anterior end of a worm including the old head, 
e.g., the region anterior to the level a in fig. 3, regulation occurs 


224 Cc. M. CHILD 


in the manner indicated in figs. 4and 6. A short postpharyngeal 
region is formed but its growth is slow unless the animal is fed. 
If we now isolate the posterior half or two-thirds of this new post- 
pharyngeal region we find it incapable of producing a new head. 
When the whole postpharyngeal region of such pieces is isolated 


Sa 


re « Hh 


ag 


ena 


ee 
6 


©) 


Figs. 3-6 Fig. 3, diagram showing levels of section. Figs. 4 and 6, different 
stages of anterior pieces with large heads. When their posterior regions are 1so- 
lated they remain headless as in fig. 5. 


it usually gives rise to a head, but the region where we should 
expect to find the anterior end of the second zodid if it is present 
shows no indication of increased capacity for head formation. 
By way of illustration the records of a series of this sort are given. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 225 


The anterior regions including the old heads of sixty well fed 
worms 16 to 18 mm. in length were cut off at the level a in fig. 3 
and allowed to regulate for ten days at a temperature of 20°. At 
the end of this time they had reached the stage of fig. 4 and at 
this stage a second operation was performed to obtain the new 
posterior ends. The attempt was made to cut the posterior ends 
from twenty of the pieces at the level indicated by the line b in 
fig. 4. This operation is not easy since these pieces move with 
considerable rapidity and the new posterior end is still short. 
Examination of each of the twenty pieces after cutting showed that 
fifteen of them had been cut approximately at the level a in fig. 
4 or even further anteriorly. This could be determined without 
difficulty because such pieces showed a region of the old pigmented 
tissue at their anterior ends. The remaining five pieces contained 
nothing but the new unpigmented tissue. The results are as fol- 
lows: 


HEADS HEADLESS DEAD 


Twenty pieces......... 14 (longer) 5 (shorter) 1 


The five headless pieces (fig. 5) are really the only part of the 
series which represents successful operations for they are the 
pieces which correspond to the region where the second zodid 
would be if it were present. . 

The remaining forty of the original pieces were left for seven 
days more, seventeen days in all after the operation. At the end 
of that period they had reached the stage of fig. 6. From these 
also the posterior ends were removed for comparison with the 
first, in order to determine whether a second zoéid had arisen 
during the later stages of regulation. The results of this second 
operation were as follows: 


HEADS HEADLESS DEAD 


Nineteen pieces......... 6 (longer) 7 (shorter) 6 


226 Cc. M. CHILD 


The nineteen pieces include all in which the cut was made any- 
where the desired level. Among these thirteen were cut at or 
near the level 6 in fig. 6 while the others were cut at the level a 
or anterior to it. Of the thirteen pieces cut at 6 six died andsix 
remained headless. The pieces that died lived for several days 
after the operation and showed no signs before death of head for- 
mation. These pieces then behave like pieces that are physiologi- 
cally as well as morphologically posterior: they resemble pieces 
from the posterior region of the first zodid in larger worms; such 
pieces always remain headless unless they are of considerable 
length. 

Pieces of the same length as those in this series, or even those 
of half that length from the posterior ends of worms in which 
the formation of new zodids has occurred will give 100 per cent or 
nearly of normal heads. 

This series seems then to indicate that in short pieces from the 
anterior end and possessing the old head the second zoédid does 
not form, even though such pieces are longer, in the present case 
7 to 8 mm., than the shortest normal worms in which the second 
zooid can be distinguished (5 mm.). In these anterior pieces the 
posterior region is physiologically as well as morphologically a 
posterior end and not a second zoéid. 

But if we feed such pieces so that growth occurs the second zoéid 
appears sooner or later, as might be expected. 

In these anterior pieces the regulatory development of the new 
postpharyngeal region occurs under conditions different from those 
which exist in normal development, 1.e., it develops at only a short 
distance from a head which represents a much more advanced 
stage of development. The obvious inference from the facts is 
that the new postpharyngeal region does not give rise to a new 
zooid under these conditions, simply because it is too completely 
subordinated to the head region with which it is connected: 
physiological isolation of this region does not attain the stage 
where formation of a new zo6id is possible until or unless the pieces 
attain a considerably greater length. 

Taken by themselves, these and similar experiments are open 
to various objections. The possibility that the posterior pieces 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 227 


are ‘too small’ or that they are not sufficiently advanced in devel- 
opment to be able to undergo regulation when isolated cannot 
be ignored. It is necessary therefore to compare the posterior 
ends of such anterior pieces possessing large and active heads with 
the posterior ends of headless pieces. 


V 


o w 


LO Teh, EZ, 


Figs. 7-12 Fig. 7, a headless piece. Its posterior end when isolated at a, b, 
or ¢ produces a normal animal. Fig. 8, headless piece with posterior zodids indi- 
cated. Wigs. 9and 10, anophthalmic pieces: their posterior ends when isolated at 
a,b, or ¢, fig. 9, produce normal wholes. Fig. 11 shows the small size of posterior 
outgrowth in pieces which form a head. Fig. 12, teratomorphic piece: posterior 
outgrowth intermediate in size between that of headless pieces and pieces with 
heads. 


One noticeable characteristic of headless pieces in general under 
natural conditions is the large amount of new tissue which develops 
at the posterior end (fig. 7). Frequently this posterior new tissue 


228 Cc. M. CHILD 


becomes equal in length to the old tissue of the piece (fig. 8) and 
in consequence of the inactivity of these pieces the posterior end 
is always broad and blunt instead of slender and tapering as in 
pieces with heads. 

If such pieces are kept at a temperature of 20° C. or lower they 
show comparatively little ‘spontaneous’ motor activity and it is 
often rather difficult to stimulate them to locomotion. But 
even under such conditions headless pieces undergo fission more 
or less frequently. At a temperature of 25°, however, they are 
much more active and very often divide. Division occurs approxi- 
mately at one of the two levels indicated by the dotted lines 
a and 6 in fig. 8. In ease fission oecurs at b a second fission often 
oecurs a few days later, indicating that the posterior region of 
these pieces consists of at least two zoédids. In case the fission 
oecurs at aa second fission does not occur unless the pieces are 
fed and grow (see Section III below). 

Fission of these pieces can also be induced in many cases by 
mechanical stimulation. In consequence of such stimulation 
locomotion takes place and an independent motor reaction of the 
posterior region is very likely to oecur. 

In all cases of fission of headless pieces the posterior product of 
fission develops rapidly into a normal whole. 

It is evident from these facts that the posterior regions of head- 
less pieces are not physiologically posterior ends but rather new 
zooids. Since, as I showed in an earlier paper (Child, ’10), the 
act of fission is the result of an independent motor reaction of the 
two zodids concerned, the second attaching itself to the substra- 
tum while the first advances, fission can occur in these headless 
pieces only when their motor activity is sufficient to accomplish 
the rupture of the tissues. 

The anophthalmie pieces (Child, ’11d) resemble the headless 
pieces as regards their posterior ends (figs.9 and 10). Fission often 
occurs in them as well as in the headless pieces. 

All of these facts show very clearly that in the anophthalmic 
and headless pieces the slight development or the complete absence 
of the head region permits the very early physiological isolation of 
the posterior end and one or more new zodids develop early in the 
course of regulation. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 229 


On the other hand, pieces of the same length and from the 
same region of the body as the headless and anophthalmic pieces, 
but which develop normal or teratophthalmic heads, show much 
less new tissue at the posterior end (fig. 11) and very rarely 
undergo fission unless they are fed and increase in length occurs. 
In these pieces a very short second zodid may sometimes arise 
at the posterior end without feeding, but the degree of physiological 
isolation is insufficient to permit the occurrence of fission except, 
as above stated, when such pieces are fed and grow. 

In most of the series of experiments which bear upon this 
point the pieces were taken from the middle region of the body. 
A large number of my experimental series consist of fifty pieces 
each of the regions 1, 2: 2, 3: 3, 4 in fig. 3. These three sets of 
pieces are designated A, Band C. When well fed.-worms are used 
and when regulation occurs at a temperature of 20° or above in 
sufficiently aérated water all of the A pieces produce heads and 
show no fission, but in the B and C pieces the anophthalmic 
types are of frequent occurrence and it is among these that the 
fissions occur. The data for a few series will serve for purposes 
of illustration. 

In these series the pieces were taken from worms 15 to 18 mm. 
in length which had received all the food they would take for at 
least a week before section. Each set of B or C pieces was fifty 
in number. The table shows the results of regulation and the 
frequency of fissions for these pieces in percentages. 


< | n= =| z 
z ao a a = 
Ps | ° & “ < a 
3 4 = 5 z 
; (B 2OMMSG E10 2401 Si.) 22 |e 
Series 32 4 alles, 
eriesd21 4 ¢ Aeileide coe tes || ene ae | 6 
| 
(Be 2 36 s 34 Is 2 2 
tea ) 

Series 387, I, 1 1G 0 { 9 8 78 8 16 
2 4 IB Sere. 2 48 8 30 12 0 6 

Series 387, II,2 < 
Series 387, 1,2) \q_ Operas | 4 io || 74 0 26 


230 C. M. CHILD 


In all of these series every case of fission occurred in either the 
anophthalmic or the headless pieces. In the three series together 
there were sixty-three anophthalmic and headless pieces among 
the B pieces and of these five underwent fission, 1.e., about 8 
per cent. The anophthalmic and headless pieces among the C 
pieces were one hundred and fourteen in number and of these 
twenty-nine, about 25 per cent, underwent fission. This is a 
fair sample of my results at temperatures of about 20°. My 
records include more than two thousand of these B and C pieces 
and among these more than 95 per cent of the fissions which 
oecurred took place in anophthalmic and headless pieces. 

It is of some interest to note that fissions occur more frequently 
in the C than in the B pieces. The C pieces represent a more pos- 
terior region ofthe original body than the B pieces and in general 
show a much lower frequency of head formation. If the forma- 
tion and separation of the posterior zodid is the result of physio- 
logical isolation we may expect to find it occurring more frequently 
in the C than in the B pieces because the latter possess in general 
less capacity to form heads at their anterior ends. 

But instead of waiting for fission to occur in the normal manner 
we may examine the capacity of the posterior ends of headless and 
anophthalmie pieces for head formation by observing the regula- 
tion of pieces cut from them at various levels. This experiment, 
which I have repeated many times, shows that such pieces produce 
normal wholes in practically every case. The result is the same 
whether the cuts are made at the levels a, b, or ¢ as indicated in 
figs. 7 and 9. In fact it is difficult to cut from the posterior end 
of such pieces a piece so small that it will not produce a normal 
whole. Moreover, if the region posterior to the level a in fig. 7 
and 9, i.e., about the anterior end of the second zodéid is cut into a 
seriesof very short pieces a sudden increase in the capacity to form 
heads appears in its posterior region, indicating that, as in most 
cases the posterior region of the second zodid has undergone fur- 
ther physiological isolation and represents one or more additional 
zooids. 

When we contrast the behavior of the posterior regions of these 
headless and anophthalmic pieces with that of the posterior regions 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS Dell 


of the anterior pieces with large heads described above, the differ- 
ence is sufficiently evident to require no comment. I believe that 
these experiments justify the conclusion that the more advanced or 
the more rapid the development of the head in any piece of given length 
the less frequent is the formation of a new zodid at its posterior end 
and vice versa. 

Moreover, this difference is not closely connected with the 
amount of nutritive material available. The posterior zodid 
arises in starved headless pieces almost as readily as it does in 
those in which excess of food is present. On the other hand, when 
the head is present and well developed, excess of nutritive 
material does not determine the formation of a second zodéid, 
except in so far as it induces growth. 

In the teratomorphic pieces (fig. 12) the head is of small size, 
but it is much more highly developed than in the anophthalmic 
and headless pieces. Histological examination shows a gan- 
glionic mass which is often almost as large as that of the normal 
head. In accordance with this condition of the head region we 
find that teratomorphic B and C pieces undergo fission but 
rarely. The posterior outgrowth in these pieces is often some- 
what larger than in pieces of the same length with normal heads 
(compare figs. 11 and 12) but it is always less than in anophthal- 
mic (figs. 9 and 10) and headless pieces (figs. 7 and 8) of the same 
length. In some cases, however, the posterior regions of terato- 
morphic pieces give rise to new zodids but the development of the 
head is in almost all cases sufficient to prevent such zodids from 
attaining the stage of development at which separation is possible. 


3. THE RESULTS OF FEEDING AND GROWTH IN TERATOMORPHIC, 
ANOPHTHALMIC AND HEADLESS PIECES 


It is a well known fact that fission can be induced in normal 
animals by feeding: the worms increase in size, the region of the 
posterior zodid or zodids growing more rapidly than others, until 
sooner or later fission occurs (Child, ’10). 

The teratomorphic, anophthalmic and headless pieces are unable 
to make their way to food which is placed near them. As the 
extractives diffuse from the meat these forms often extrude their 


202 Cc. M. CHILD 


pharynges and these make searching movements in the water and 
over the substratum, but the pieces do not move toward the meat. 
If, however, they happen to come into contact with it by chance 
they almost invariably feed. 

In the teratomorphic pieces this inability to direct the move- 
ments toward the food is of particular interest. These pieces are 
very evidently excited by the diffusing extractives from the meat: 
they lift their heads and move them about in much the same 
manner as normal animals, but this reaction is not followed by 
movement toward the food. The movement which occurs may 
be in any direction and discovery of the meat seems to be purely 
a matter of chance. In all the teratomorphic pieces which I 
have used thus far in such experiments the auricles were partly 
or wholly fused in the median line at the anterior end of the head 
and it seems probable that the inability of these pieces to direct 
their movements toward the food is due to the fact that they 
possess only one median instead of two lateral organs of chemical 
sense. 

These sense organs are usually absent in the anophthalmic and 
always absent in the headless forms. In these latter the only 
marked reaction to food which I have observed is that of the 
pharynx. 

Since even the headless, as well as the anophthalmic and terato- 
morphic pieces will feed when brought into direct contact with 
the food, it is possible to keep them indefinitely and to bring 
about growth. In feeding such forms pieces of meat are placed 
in the dish containing them and the pieces are then picked up with 
a pipette or a needle, to which they usually adhere, and deposited 
on the meat. There they feed and sometimes creep away after 
they have finished. After two or three hours they are removed 
from the meat if still on it, the meat is taken out and the water 
changed. By these means such pieces can be kept with little 
more difficulty than normal worms. The possibility of keeping 
and ‘breeding’ such pieces, either through natural fission or by 
section opens up an interesting field of experiment. At present, 
however, we are concerned only with the occurrence of fission in 
connection with growth: other results will be considered at another 
time. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 233 


In some cases teratomorphic heads redifferentiate in later 
stages into heads of normal form and two eyes symmetrically 
placed develop in addition to the single median eye. In some 
cases also pieces which are at first anophthalmic later become 
teratomorphie or teratophthalmic. Occasionally these changes 
occur even when the pieces are not fed, particularly if they are 
kept at a temperature of 25° C. or higher; below this temperature 
I have not observed them. When the pieces are fed such changes 
occur more frequently. 

In general, however, these redifferentiations of the head occur 
only within the first two or three weeks after section. If the 
piece remains teratomorphic or anophthalmic during that time 
it remains so indefinitely, no matter how much food it receives. 
I have kept some of these pieces for more than three months with- 
out the occurrence of any trace of change. 

In pieces which were strictly headless ten or twelve days after 
section further development of the anterior end has never occurred 
within the limits of my observations. The important fact in all 
of this for present purposes is that many of the teratomorphic, 
anophthalmie and headless pieces do not develop further when fed 
but remain what they were. Evidently such pieces have attained 
a new dynamic equilibrium different from that of the normal 
animal. 

Since it is only during the last few months that I have attempted 
to feed pieces of this sort my experiments have not as yet pro- 
ceeded very far, but one result of interest in the present connec- 
tion has already been obtained. When fed abundantly so that 
growth occurs, the headless pieces usually attain a length of 5 to 
6 mm. before fission occurs. The anophthalmic pieces attain 
on the average a somewhat greater length, 6 to 7 mm. before fis- 
sion and the teratomorphie pieces often attain a length of 9 to 
10 mm. before dividing. 

After the first period of growth and fission no further increase 
in size of the anterior regions of these pieces has been observed. 
All further growth is in the posterior region where after the first 
fission another new zodid forms and grows until it attains about 
the same size as before, when another fission occurs. Up to the 


234 Cc. M CHILD 

present I have seen pieces of this kind produce new zoéids and 
divide four times and there is every reason to believe that they 
will continue to do the same thing indefinitely, or at least for a 
much longer time. 

There is of course considerable variation in different individual 
pieces, but in general there is no doubt that the length which the 
piece attains before fission varies with the degree of development 
of the head region: the higher the development of the head region 
the greater the length which the animal attains before fission. 
Evidently the more highly developed head region is capable of 
dominating a longer body than the less highly developed. If the 
pieces were left entirely undisturbed the posterior zodids in the 
anophthalmie and headless pieces might attain a considerably 
greater length before fission took place, but of course pieces which 
are fed cannot be left undisturbed. When the pieces are fed every 
alternate day or at regular intervals their removal from the meat 
and the change of water stimulate them to movement and so aid 
in bringing about the act of fission. The teratomorphic pieces, 
however, show a much greater degree of ‘spontaneous’ activity 
than the others, therefore the fact that they attain a greater 
length before fission is certainly not due to less motor activity. 

Such pieces do not attain anything like the length of animals 
with normal heads before fission. The greatest length observed 
thus far in the teratomorphic pieces is 9 to 10 mm., somewhat 
more than half the length attained by normal animals under simi- 
lar conditions before fission. 

These various facts seem to me to indicate very clearly the 
importance of the factor of distance in the dominance of the 
head region and the réle of physiological isolation in the formation 
of new zodéids. 

One further fact may be mentioned incidentally. After fission 
the posterior piece, whether it arose from a teratomorphic, an 
anophthalmic or a headless piece, develops rapidly into an animal 
with normal head and eyes. In short it does not inherit the peculi- 
arities of its parent. In later papers I shall show from other 
experimental data that the various types of the head are due 
primarily to quantitative rather than qualitative differences in 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS Dire 


the dynamic processes concerned. Teratomorphic, anophthalmic 
and headless pieces appear where the rate of reaction in the 
developing region falls below a certain minimum necessary for 
the development of a normal head. In the posterior zoéid after 
separation conditions are quantitatively different and the rate 
of reaction is sufficient for the formation of normal heads. 


4. THE FATE OF POSTERIOR ZOOIDS IN STARVATION 


It has long been known that Planaria and various other species 
of turbellaria may be reduced by starvation to a small fraction 
of their original size before death occurs. In general the post- 
pharyngeal region decreases in length more rapidly than the pre- 
pharyngeal during starvation. In the very large animals the 
postpharyngeal region is much longer than the prepharyngeal: 
Fig. 13 shows the proportions of a worm 22 to 24 mm. in length. 
As reduction through starvation occurs in such worms the pro- 
portions remain at first about the same, but when the animal is 
reduced to about half its length (10 to 12 mm., fig. 14) it is usually 
found that the pharynx is nearer the middle of the body than 
in longer worms, although the postpharyngeal region is usually 
still the longer. When the worm is reduced to 4 to 5 mm. in 
length the pharynx is usually approximately in the middle 
(fig. 15) and as further reduction occurs the postpharyngeal 
region may become slightly shorter than the prepharyngeal 
(fig. 16). These changes in proportions are the reverse in direc- 
tion of those which occur during the normal increase in size of the 
animals, but in general the reduced animal of a certain length pos- 
sesses a longer postpharyngeal region than the well fed growing animal 
of the same length. This difference is particularly noticeable in 
the smaller worms. In a young growing worm of 5 mm. in length 
the pharynx is usually posterior to the middle of the body (fig. 
17), while in the starved worm of the same length it is in the 
middle or anterior to the middle (fig. 15). In very young worms 
2 mm. in length, i.e., shortly after hatching, the pharynx is near 
the posterior end (fig. 18) while in starved worms of the same 
length the pharynx is only slightly, if at all posterior to the middle 
(fig. 16). 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 3 


236 


13 


Cc. M. CHILD 


I4 16 


Is 
Lyf, 


Figs. 13-18 Fig. 13, animal 24 mm. in length: postpha- 
ryngeal region about twice as long as prepharyngeal. Fig. 
14, a worm reduced by starvation from 24 to 12 to 14 mm: 
postpharyngeal region somewhat longer than prepharyngeal. 
Fig. 15, starved worm of 5mm., postpharyngeal and prepharyn- 
geal regions about equal. Fig. 16, starved worm of 2 mm. 
Fig. 17, growing worm of 5 mm., postpharyngeal region shorter 
than prepharyngeal. Fig. 18, newly hatched Planaria macu- 
lata: postpharyngeal region very short. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 237 


Comparison of the different stages of starvation shows that in 
most cases the anterior zodid decreases Sightly more rapidly than 
the posterior or at about the same rate during the earlier stages, 
but that later the decrease in length is more rapid in the posterior 
zooid or zodids. Apparently in the earlier stages neither part 
is capable of maintaining itself at the expense of the other to any 
ereat extent, or if there is any difference in this respect it is the 
posterior younger part which has the advantage. Later, however, 
as the worm becomes shorter the degree of dominance of the head 
region over the posterior region increases as the distance between 
the two regions decreases and the independent development of 
the posterior zodid is retarded or ceases. From this time on the 
anterior zodid has the upper hand, so to speak, and lives to some 
extent at the expense of the other, therefore in the later stages of 
starvation the posterior zodid decreases more rapidly than the 
anterior. 

A comparison of the different regions of the first zodid shows 
what has been noted by other authors, viz., that the head region 
decreases in size somewhat less rapidly than other parts. Conse- 
quently the head appears disproportionately large in the small 
reduced worms. On the other hand, repeated comparison of 
starved and young worms of the same size leaves no doubt that 
the head of the young worm is slightly larger than that of the 
starved animal. I have not attempted measurements in con- 
nection with this point, but have observed the difference many 
time. 

The facts indicate then that in the course of reduction through 
starvation the posterior zodid or zodids become less and less 
completely physiologically isolated from the dominant head region 
as the distance between them decreases. The zodids which arose 
during growth become less and less true zodids as they come more 
completely under the control of the dominant region: they gradu- 
ally return to or approach the condition of physiologically posterior 
regions and as this change occurs the more anterior regions 
become more and more able to maintain themselves on the mate- 
rial of the posterior parts. 

But if the longer postpharyngeal region and the smaller size 
of the head in the extreme stages of starvation as compared with 


238 Cc. M. CHILD 


young animals of the same length means anything it means that 
after the second zodid has once arisen it possesses a certain capa- 
city to maintain itself even in conflict with other parts. In other 
words, after a new zoéid has once arisen, whatever the conditions 
of its origin may have been, the control of the dominant head 
region of the whole over such a zoéid is not as complete as its 
control over a region situated at the same distance from it but in 
which no new zoéid has arisen. The relation between the ante- 
rior and posterior zodids in starvation is somewhat similar to 
that between a weaker and a more powerful or a less and a more 
active individual in a struggle for food when the supply is in- 
sufficient for both. At first the weaker individual succeeds in 
obtaining a certain amount but in the end it is eliminated before 
the other. 

At any given stage of starvation then the dominance of the 
anterior region of the animal over the body as a whole is some- 
what less complete than in growing animals of the same length. 

There seems to be no reason for a teleological interpretation of 
these relations such as Schultz (’04, ’06, ’07, ’OSa, ’OSb) and cer- 
tain other authors have suggested. If the relation between mor- 
phological structure and metabolism is of the character which our 
present knowledge seems to indicate (Child, ’1le, pp. 571-578; 
‘lle, pp. 173-178) it is evident that in starvation those parts of 
the body whose rate of reaction decreases least will maintain 
their structure and size most completely, for they will obtain their 
energy in part from other organs whose rate of reaction has 
decreased to a greater extent. In the part with the higher 
rate of reaction the loss of structural substance, which takes place 
more or less at all times is more completely compensated by the 
new structural substance which is formed in the course of further 
metabolic reactions than in the part with the lower rate. In the 
organ with the lower rate, on the other hand, the chemical con- 
ditions must favor an increased breaking down of structural 
material which thus becomes available as a source of energy and 
incidentally for the formation of new structure wherever the 
demand is greatest, i.e., wherever the rate of reaction has under- 
gone least decrease. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 239 


In Planaria the head region, where the chief sense organs which 
are affected by external stimuli are situated, and the central 
nervous system, the chief organ of conduction and correlation, 
must maintain more nearly their characteristic rate of reaction 
during starvation, not because they are more essential than other 
parts, but simply because they are more often or more intensely 
stimulated than other parts. So long as the animal remains 
alive in a changing external environment these organs, primarily 
because of their relation to the external world and to other parts 
must maintain more nearly their characteristic rate of reaction 
than other parts. Both their importance for the life of the organ- 
ism and their persistence in starvation are due to this relation. 
In general it is the most essential organs which are most constantly 
or most frequently the seat of dynamic processes, but the occur- 
rence of dynamic processes in any part depends not simply upon 
the part itself but upon its correlation with other parts and its 
relations to the external world, in short upon its environment. 
The most ‘essential’ part in any organism is then merely that 
part in which under the conditions of existence of the organism 
characteristic dynamie processes are most constantly or most 
frequently induced by the internal or external environment. In 
the simpler organisms like Planaria, which in the absence of food, 
are able to use their own structural colloids as a source of energy 
to a very large extent the most essential parts according to this 
definition are the head region and the central nervous system. 
When the dynamic processes in these fall below a certain level the 
organism ceases to exist as such. To maintain that the organism 
is able to control and guide the destruction or maintenance of 
parts in starvation in a manner involving the idea of finality is 
to ignore the facts which are already at hand. 

In the absence of the head region the relation between parts 
in starvation is very different from that which exists when the 
head region is present. We are able to compare the course of 
starvation in normal animals and in headless pieces of Planaria 
and we find that in headless pieces it is the posterior instead of the 
anterior zodid which maintains itself at the expense of other 
parts. Fig. 19 represents a headless piece corresponding to the 


240 C. M. CHILD 


IQ 20 2 22 

Figs. 19-22 The changes in a headless piece during starvation. Fig. 19, 
twelve days. Fig. 20, twenty-six days. Fig. 21, forty-eight days. Fig. 22, one 
hundred and twelve days. 
region 2, 3 in fig. 3 as it appeared twelve days after section. The 
boundaries between the new and the old tissue are indicated at 
the two ends of the piece. Fig. 20 shows the same piece twenty- 
six days after section: the posterior new tissue, i.e., the region of 
the posterior zodid has increased in size relatively to the anterior 
region. Fig. 21 shows the same piece after forty-eight days: 
here still farther increase in size of the posterior region at the 
expense of other parts has occurred. And finally fig. 22 shows 
the piece after one hundred and twelve days of starvation. 
Throughout the course of starvation the posterior zodid has 
continued to increase in size as compared with the anterior region. 
At the stage of fig. 22 the pharynx, which decreases in size less 
rapidly than the parts about it, has been forced posteriorly 
through the tissues until its posterior end lies a considerable 
distance posterior to the mouth as indicated in the figure. Shortly 
after the stage shown in fig. 22 the piece died. In general headless 
pieces die of starvation at a much larger size than wholes. This 
difference is undoubtedly connected with a difference in the rate 
or intensity of the dynamic processes and will be more fully con- 
sidered elsewhere. At present we are concerned with the fact 
that during starvation the posterior zodid decreases in size less 
rapidly than that part of the anterior zoéid which is present: inthe 
absence of a more highly developed head region anterior to it the 
anterior region of the posterior zodid has become to some extent 
the dominant region of the whole piece. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 241 


A similar relation exists under certain conditions in Stenos- 
tomum, as I showed in an earlier paper (Child, 03). There the 
cephalic region of any zodid, even if the more posterior parts of 
the zodid are not present, will bring about the resorption and 
destruction of portions of zodids deprived of the cephalic gangha 
which are attached to its anterior end. Moreover, when pieces 
are cut so that they consist of one or more younger zoéids with an 
older zo6id posterior to them, the older, most posterior zodid is domi- 
nant and the younger zoéids anterior to it, instead of continuing to 
develop, undergo resorption and the posterior zodid grows at their 
expense until finally its cephalic region becomes the anterior end of 
the whole piece. Only occasionally and when the younger anterior 
zooids are themselves well advanced in development do they 
succeed in maintaining their individuality and continuing their 
development to the stage of separation and even under these 
conditions they usually undergo considerable reduction in size. 
In Stenostomum this process of destruction is not a matter of 
long continued starvation, but is completed in two or three days 
and is much more striking than in Planaria. The enteric cavity 
and pseudocoel of the posterior zodid in Stenostomum often 
become greatly distended with the cellular débris from the dis- 
integrated anterior zodids or parts of zodids. In both cases, 
however, the posterior zodid becomes dominant and if the pieces of 
Planaria did not die before the process was completed we should 
undoubtedly find the head region of the posterior zodid becoming 
the anterior end of the piece in Planaria as well as in Stenostomum. 

Both of these cases demonstrate that the head region of a pos- 
terior zodid exercises some sort of correlative influence upon 
parts anterior to it, at least in the absence of more highly devel- 
oped dominant parts in those regions, as well as upon parts pos- 
terior to it. In the case of Stenostomum, however, starvation 
is not necessary to bring about the decrease in size and resorp- 
tion of the anterior zodids; this occurs so rapidly that the posterior, 
dominant zoéid is often packed almost to bursting with the prod- 
ucts of disintegration which it only gradually makes use of as 
nutrition. It is perhaps searcely worth while at present to 
speculate concerning the exact nature of the influence of the pos- 


242 C. M. CHILD 


terior zodid upon the parts anterior to it in Stenostomum, but it 
seems not improbable that the rapid resorption and disintegra- 
tion of the anterior parts is due to a condition of more or less 
complete functional paralysis. Theyare reduced by the operation 
to the condition of mere excrescences upon the anterior end of 
an organism. If they receive impulses from the dominant 
region posterior to them these must conflict with and contribute 
to the destruction of their own codrdinations since the region 
from which they come is posterior instead of anterior. More- 
over, the existing conditions do not result in the incorporation of 
these parts into a new functional whole as in many cases where 
parts are grafted together and both are undergoing regulation, 
for the functional whole is already present in the posterior zodid 
and merely interferes with the function of these parts anterior to 
it. My suggestion is then that in this conflict of conditions corre- 
lation in these parts is destroyed or interfered with to such an 
extent that they cease to exist as individuals and their cells, 
which do not necessarily die at once pass into the cavities of the 
body of the dominant individual as so much foreign maiter. 


5. THE FORMATION OF CHAINS OF ZOOIDS 


In my paper on physiological isolation and fission in Planaria 
(Child, 710) I mentioned the probability that long worms often 
consisted of more than two zodids. Since that time it has been 
possible to obtain more definite evidence upon this point and it 
is now certain that animals of more than sixteen millimeters in 
length usually consist of at least three and probably in most cases 
of more than three zodids, for the extreme posterior end appar- 
ently represents a sort of ‘growing tip’ and probably consists 
of a number of very short zodids. This region is perhaps compar- 
able to the ‘growing region’ at the posterior ends of various anne- 
lids where a considerable number of minute segments are often 
visible. 

It is possible to demonstrate the existence of several zodids at 
the posterior end of long worms by inducing fission at different 
levels and this can readily be done, as I shall show, by cutting 
pieces of different lengths so that the new head will arise at a 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 243 


certain distance from the region where it is desired to induce 
fission. It is also possible to induce the disappearance of zodids 
already formed by cutting the piece so that a head will form a 
short distance anterior to the head region of the zodid. 

But it is possible to determine approximately the positions of 
the anterior ends of zodids in the body without waiting for fission 
to occur. If we cut the postpharyngeal regions of long worms into 
series of short pieces of as nearly as possible equal length we 
find that the pieces which represent the level of the head region of 
a zpoid show a greater capacity for head formation than the others. 
In general the results obtained by this method correspond so 
closely with those of the experiments on fission that I believe we 
are justified in concluding that such regions of increased capacity 
for head formation, wherever they occur in the postpharyngeal 
region represent the anterior ends of existing zodids. In Planaria 
these zodids do not develop far enough to become visible mor- 
phologically, but I think there can be no doubt that they do exist 
as more or less definite regions of codrdinated dynamic activity 
which represent the earliest stages of the redifferentiation of this 
part of the body into new individuals. Under natural conditions 
an independent motor reaction sooner or later brings about the 
complete isolation of these regions and then the redifferentiation 
proceeds rapidly as in regulation. 

My experiments include a large number of series of pieces of 
this kind, but most of them show the existence of two or at most 
three zodids. Below are given the records of three series, the 
first of which shows the worms used to consist of at least three 
zooids, while the other two series indicate the presence of four 
or more. The method is of course not an accurate one since it is 
impossible to cut pieces of exactly equal length and length is a 
factor in the eapacity for head formation (Child, ’11b): moreover, 
as the length of the piece decreases its capacity for regulation 
decreases and the differences between different levels often become 
less marked. However, I have used the method merely as a means 
of showing that the planarian body may consist like Stenosto- 
mum and various other forms of a chain of zodids. 

For the sake of convenience the notation which was used in 
my earlier work on Stenostomum is adopted here. In a chain 


244 C. M. CHILD 


consisting of two zodids resulting from the division of a single 
individual the two zodids are 1. and 2., 1. being the anterior. 
When zoéid 1. divides it gives rise to zodids 1.1. and 1.2. and 
zooid 2. divides into 2.1. and 2.2. A third cycle of divisions gives 
ZOOIGS Wala ol ees De DED ee a eleeweues 

Series 333. During two weeks before the experiment the stock 
received maximum feeding: food was placed in the jar every day 
and after a few days of such feeding the stock as a whole showed 
but little hunger, i.e., on any one day only a few animals would 
feed. The postpharyngeal region in all worms used for the 
experiment was much longer than the prepharyngeal, an indica- 
tion that the posterior zodid or zodids had reached a considerable 
size. 

On February 8, 1911, ten worms as nearly alike as possible and 
all 18 to 20 mm. in length were selected from the stock and the 
postpharyngeéal region of each of these, beginning at the mouth, 
was cut into six pieces as nearly as possible equal in length. 
Each ten pieces representing approximately the same region of 
the body from each of the ten worms were placed together. The 
results of the regulation of these pieces appear in the following 
table in which the pieces are numbered 1 to 6 from the anterior 
end of the postpharyngeal region backward. Theresults are given 
in actual numbers, but since each set consists of ten pieces multi- 
plication of these numbers by ten will give percentages. 


| 
| 
PIECES NORMAL TERATOPHTHALMIC | TERATOMORPHIC ANOPHTHALMIC | HEADLESS 


1 1 0 0 0 9 
2 0 1 2 2 5 
3 3 1 0 1 5 
4 0 4 1 1 4 
5 8 2 0 0 0 
6 ¢f 3 0 0 0 


These results are presented graphically in fig. 23. In this 
figure the larger spaces of the cross section paper along the ordinate 
represent number of pieces, ten being the total. The abscissa 
represents the axis of the postpharyngeal region of the worms and 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 245 


i IE t 22+ 


Fig. 23 Curves of regulation in postpharyngeal regions of worms of 18 to 20 
mim. to show the existence of postpharyngeal zodids. Postpharyngeal region cut 
into six pieces. Ordinate represents number of pieces, abscissa the different levels 
of the postpharyngeal region at which cuts were made. The curve in unbroken 
line is the curve of reconstitution and includes all cases in which any approach to 
head formation occurred. The curve in long dashes is the curve of eye formation. 
The curve in short dashes is a ‘summation’ curve to show the different degrees of 
reconstitution at different levels. The line mz below the curves represents the 
axis of the postpharyngeal region: on it the approximate boundaries of the zodids 
are indicated and the lineage of the zodids given. 


246 Cc. M. CHILD 


the numbers 1 to 6 indicate the position of the anterior ends of 
each of the six sets of pieces. The starting point of the curves at 
the left corresponds to the anteriorend of the prepharyngeal region 
and their end at the right to the anterior end of the most posterior 
set of pieces. These curves then, like figs. 40 and 41 of the first 
paper of this series (Child, ’11b) show the differences in certain 
results of regulation at different levels along the axis. 

Of the three curves in the figure the one drawn in continuous 
line is what we may designate as the curve of reconstitution: 
its ordinates represent the number of pieces at each level or the 
body where a cut was made, that show any approach to head 
formation. It includes all except the headless pieces in the above 
table and serves merely as a general expression of the capacity 
of the pieces to initiate the process of reconstitution, or more 
specifically the process of head formation, which is, as I showed in 
the preceding paper (Child, ’J1f) the first step in reconstitution 
in such pieces. 

The curve drawn in long dashes is the curve of eye formation: 
its ordinates represent the number of pieces at each level which 
form eyes, whether normal or abnormal. 

The third curve drawn in short dashes is somewhat different in 
character and may for convenience be called the summation curve. 
As a basis for this curve each of the regulatory types in thetable 
is assigned an arbitrary numerical value which is selected first 
with reference to the different degrees of approach to the forma- 
tion of a normal head and second with reference to convenience 
in plotting the curve on the same scale as the others. The numer- 
ical values assigned to the different types are as follows: 


Normal heads 5 Meramorphicyes-bic.a- + ses 
Teratophthalmic 4 Amoph thalmie esti ccicuttieriee aera 


In determining the length of the ordinate of the curve at each 
level each of these values is multiplied by the number of pieces 
at that level which show that type of regulation and the sum of all 
these products for the different types of pieces at any one level 
gives the length of the ordinate. In order to draw this curve on 
the same scale as the others each unit of the arbitrary values is 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 247 


made to correspond to one of the small spaces of the cross section 
paper. Thus if all ten pieces at any level develop normal heads 
the ordinate of the curve at that level will equal fifty of the small 
spaces and will coincide with the ordinate of the curve of recon- 
stitution and that of eye formation at the same level, both of 
which will equal ten of the large spaces. Referring to the table 
we see that the first ordinate of the summation curve is 5, the 
second 12, the third 20, ete. 

This curve is merely an attempt to show to some extent the 
relative capacity of different regions to form heads. At one level, 
for instance, three pieces may form normal heads and the others 
remain headless while at another level perhaps three pieces pro- 
duce anophthalmic forms and the others remain headless. If we 
count each of the three cases of normal head formation andeach 
of the three anophthalmic forms merely as one, as in the curve of 
reconstitution in fig. 23 we take no account of the difference in 
degree of reconstitution. Three anophthalmic pieces do not 
represent the same capacity for reconstitution or for head forma- 
tion as three normal heads. The summation curve then attempts 
to take account of the different degrees of reconstitution. It is 
of course purely arbitrary in character. 

Examination of the curves in fig. 23 shows that the course of 
all three is in general similar. Each shows first a rise and then a 
decrease in steepness, then a second rise. In other words this 
series shows two levels in the postpharyngeal region where a 
marked increase in the capacity of pieces of a given length to 
form heads appears. I believe that these two levels indicate 
approximately the levels of the anterior ends of zodids. They 
correspond closely with the levels at which fission occurs in ani- 
mals from the same stock and of the same length. 

The absence of a fall after the second rise suggests that the 
extreme posterior region of these worms is in reality not a single 
zooid but rather a series of zodids too short to be distinguishable 
from each other in pieces of the length used in this series. If the 
posterior end consisted of but a single zoéid we should expect the 
axial gradient (Child, ’11b) in this zodid to appear in the most 
posterior set of pieces. The failure of this gradient to appear 


248 C. M. CHILD 


suggests that there is at least one other head region posterior to 
the level of the second rise in the curves. 

The line mx below the curves in fig. 23 represents the axis of 
the postpharyngeal region of the worms of this series drawn to 
the same seale as the curves, the mouth being at m. The two 
dotted lines drawn across this axis indicate the levels at which 
the steepness in the rise of at least two of the curves begins to 
decrease, i.e., they indicate approximately the two levels of 
ereatest capacity of head formation in the postpharyngeal region. 
As regards the more posterior level the position indicated on the 
line ma is more or less arbitrary for at ordinate 5 all the curves 
rise almost to 100 per cent and none of them shows any ap- 
preciable fall posterior to that level. 

The more anterior of these two levels coincides approximately 
with the level at which fission usually occurs in these worms and 
the more posterior level with that at which fission occasionally 
occurs in whole worms and very frequently in the posterior prod- 
ucts of a preceding fission (Child, 710). All the facts taken 
together seem to me to indicate very clearly that these levels 
represent more or less exactly the anterior regions of two zodids 
and that consequently the worms used in this series consist of 
at least three zodids, the anterior, bearing the fully developed 
head, and two others posterior to it. The extreme posterior 
region may represent still another zodid or more than one. 

The impossibility of observing the origin of the zodids directly 
prevents us from determining the lineage of the different zodids 
with certainty, but the facts already at hand concerning the domi- 
nance of the head region and the distance factor and their rela- 
tion to the degree of development of the head region make it 
appear at least probable that the two posterior zodids have arisen 
by the physiological isolation of the posterior region of a single 
posterior zodid rather than through the division of the anterior 
zooid. If this inference is correct then the three zodids present 
in these worms are 1., 2.1. and 2.2., as indicated in fig. 23. The 
probability that zoéid 2.2. has undergone still further division is 
indicated by the plus sign following the designation 2.2. in fig. 25. 

Series 298. This series of ten worms was taken from a stock 
collected on November 21, 1910. At this season the temperature 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 249 


of the water in which the worms live is so low that they are rather 
sluggish and their reactions are in general slow. In consequence 
of this and of the slow growth at such temperatures fission does 
not occur frequently, but the animals are still sufficiently active 
to feed. Because of these and other internal conditions to be 
discussed below the animals attain a much greater length under 
these conditions than at higher temperatures. From this stock 
as collected a hundred or more of the longest worms, 20 to 22 
mm. in length were selected and placed in a dish, the bottom and 
sides of which were covered with a layer of vaseline. The vase- 
line serves to some extent to prevent fission for the animals cannot 
adhere to it as to a glass or metal surface and consequently the 
rupture of the tissues is impossible. The worms live perfectly 
well under these conditions and stocks have been kept for several 
months in such dishes. 

These worms on vaseline were provided with an excess of food 
during twenty-six days and by that time some of them had at- 
tained a length of 25 to 30 mm. without fission. The postpharyn- 
geal region was twice as long, or in a few cases more than twice as 
long as the prepharyngeal region. 

If the increase in length is a factor in determining the formation 
of new zoéids then the conditions in these worms are favorable 
for the formation of a larger number of zodids than in shorter 
animals. 

On January 5, 1911, ten worms about 25 mm. in length were 
selected from this stock and the postpharyngeal region of each of 
these was cut into eight pieces as nearly as possible equal in length. 
The ten pieces corresponding to each level were placed together 
and the results of regulation recorded. These results are given 
in the following table, the sets of pieces being numbered 1 to 8 
from the anterior end of the postpharyngeal region posteriorly. 


| 
PIECES NORMAL | TERATOPHTHALMIC TERATOMORPHIC ANOPHTHALMIC HEADLESS 


1 0 0 0 0 10 

2 0 1 0 ta) 4 

3 0 1 1 2 6 

4 8 2 0 0 0 

5 0 0 0 3 7 

6 0 0 0 i 7 
2 


250 Cc. M. CHILD 


The data given in this table are presented graphically in fig. 
24. Here, as in fig. 23, three curves are plotted; the curve of 
reconstitution, drawn in continuous line and including all cases 
which show any approach to head formation; the curve of eye 
formation, drawn in long dashes and including all cases in which 
eyes of any sort appear: and the summation curve drawn in short 
dashes and plotted in the same manner as in the preceding series. 

All three of these curves show certain resemblances. The curve 
of reconstitution shows three very strongly marked rises, of which 
the first is not as high as the other two. The summit of the first 
rise in this curve corresponds to a decrease in steepness at ordinate 
2 in the summation curve, but the curve of eye formation does not 
show any summit at this level of the body. In all other respects 
the three curves correspond closely. 

Fig. 23 shows then the presence in the postpharyngeal regions of 
these worms three distinct regions of increase in head formation: 
the first of these is less distinct than the others, the second is very 
strongly marked and is followed by an equally well marked fall, 
while the third is even more strongly marked than the second 
but is not followed by a fall. If these regions of increase in head 
formation represent the anterior regions of zodids then these 
worms consist of at least four zodids. As in fig. 23, there is no 
fall in that region of the curves which corresponds to the extreme 
posterior region of the body and this again suggests that this 
region consists of more than one zo6id. 

The approximate levels of the different zodids are indicated on 
the line mx below the curves in fig. 24 and their probable lineage 
is given. The designation 2.2.+ for the most posterior region 
and the arbitrary division of this region by the dotted lines merely 
serves to indicate the probable further division of this part. 

The anterior region of the zoéid 2.1. corresponds to the most 
clearly marked maximum in the curves and it also coincides 
closely with thelevel at which fission usually occurs in these worms. 
It probably represents therefore the anterior region of the zodid 
which is most advanced in development and that must be a de- 
scendant of zodid 2., in this case 2.1. Posterior to zoéid 2.1. is at 
least one more zoéid which probably arose from the earlier physio- 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 251 


+ im [al 
ine 
: arate 
CEH ERE a 
- | IGS oe 
ital 
ime TAR On Df 
DEE Eoeeeeaeaeeeee 
B (2 aa | | i Z| | 
EEE PEEPS eee eer 
fala iv ime | 
J a= To Tahal 
H if 
T 
EE IC 
| ' tt 
W jah ttt 
meaeran 
oe a aH 
He Ge Swipe 
Ete Ina! AE 
ry eee 
ialeia 5 
Paw ic =f 
Coot t 
4 
aT ae 
ap Ne aE 
! [eh zl 
ielalcislale ial 
LT 2 3 
Mt LI. i 1.2, a: Dy jet : 


Fig. 24 Curves of regulation for postpharyngeal regions of worms of 25 mm.: 
eight pieces: all details as in fig. 23. 
logical isolation of the posterior region of zodid 2. And in this 
case we find still another zoéid indicated anterior to zodid 2.1. 
This I regard as resulting from a new division of zoéid 1. and it is 
therefore designated 1.2. in fig. 24. The reasons for so regarding 
it are these: it is short and the increase in head formation which 
corresponds to its anterior region is less strongly marked than 
that corresponding to other zoéids. Both of these facts indicate 
that it is less advanced in development than the others and there- 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 3 


252 Cc. M. CHILD 


fore has probably arisen later. Moreover, if the process of fission 
in Planaria follows the same general laws as in Stenostomum, and 
the probability is that it does, as I shall show below, then this 
lineage of the zodids is what might be expected. The animal 
divides first into two zodids: then, since in Planaria the posterior 
of these two does not undergo rapid development and differentia- 
tion as it does in Stenostomum, its posterior region soon becomes 
physiologically isolated as a third zoéid and this process continues 
as elongation goes on. Meanwhile the anterior of the first two 
zooids has also elongated to some extent and since it possesses the 
fully developed head which is able to dominate a much greater 
length of body than the head regions of the posterior zodids it 
attains a much greater length than these zodids before its posterior 
region becomes physiologically isolated. Finally, however, it does 
become isolated and the development of zoéid 1.2. begins. This 
is the condition which I believe is represented in this series. 

Series 297. The worms of this series were taken from the 
same stock as those of the preceding and at the same time. They 
were the ten longest worms among those which had been kept on 
vaseline. Their length was somewhat greater than that of the 
worms of Series 298, being 27 to 28 mm. In the present series 
the postpharyngeal region was cut into twelve pieces of equal 
length. The following table gives the results: 


THERATOPH- THERATO- 


PIECES NORMAL Bea aie Moneiie | ANOPHTHALMIC) HEADLESS DEAD 
1 0 0 0 0 10 0 
2 0 0 0 0 10 0 
3 0 1 0 1 8 0 
4 0 0 0 0 10 0 
5 0 0 0 3 6 1 
6 0 0 0 1 8 1 
7 1 0 0 1 8 0 
8 0 4 0 1 5 0 
9 2 1 0 2 5 0 

10 3 5 0 2 0 0 
11 6 3 0 1 0 0 
12 8 2 0 0 0 0 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 253 


Bi [eid edd fee ced ce 


! 2, i 211. ' fal bed WW area 


Fig. 25 Curves of regulation for postpharyngeal regions of worms of 28 mm.: 
twelve pieces: all details as in fig. 23. 


In fig. 25 the three curves are plotted from these data in the 
same manner as in figs. 23 and 24. All of these curves show four 
distinct regions of increase in head formation. The first two 
increases correspond rather closely in position with the first two 
in fig. 24 and may doubtless be regarded as representing the ante- 
rior ends of the same zodids. The most posterior rise also corre- 
sponds closely in position to the last rise in fig. 24 and represents 
the region of the posterior zoéid or zodids. But between these ©: 


‘ 


254 Cc. M. CHILD 


ordinate 8 in fig. 25 the summit of another rise in the curves 
appears which is not present at all in fig. 24. This summit is 
only slightly marked and, if my conclusions are correct must 
represent the anterior end of a rather young zodid, i.e., one 
recently formed. When we examine the line maz below fig. 25, 
on which these different summits are indicated, and compare it 
with mx in fig. 25 it becomes evident at once that the region 
corresponding to zodid 2.1. in fig. 24 has divided in fig. 25 into two 
zooids which must be 2.1.1. and 2.1.2. This is what we might 
expect since the worms of the present series are somewhat longer 
than those of the preceding, from which fig. 24 was plotted. More- 
over, in fig. 24 zodid 2.1. is longer than the other postpharyngeal 
zooids and since its head region remains inan early stage of develop- 
ment its further division might be expected if it increased further 
in length. 

We see then that the curves in fig. 25 indicate the presence of 
at least five zodids in these very long worms. The extreme 
posterior region shows the same characteristics as before and 
suggests a division into a number of short zodids. 

Series 298 and 297 seem to me to confirm each other in a strik- 
ing manner. The only difference between the worms of the two 
series is that those of Series 297 are slightly longer than the others 
and this series shows one more zodid than the other. The levels 
of the anterior ends of the other zodids correspond very closely 
in the two series and afford additional proof if such is needed 
that we are not dealing with mere chance differences in regulatory 
capacity or with differences due to unequal length of the pieces. 
Moreover the data from which the curves are plotted are not 
from single worms but from ten individuals in each case and the 
uniformity in the results constitutes a practical demonstration 
that we are concerned here with real physiological differences at 
different levels of the body. And finally, when we compare these 
data with those on the occurrence of fission at the different levels 
under natural and experimental conditions, there can I think be 
no doubt that these regions of increased capacity for head for- 
mation really represent the anterior ends of zodids which have 
arisen in this region but which have not developed far enough to 
become visible. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 255 


In any comparison of fig. 23 with figs. 24 and 25 it must be 
remembered that fig. 23 is plotted from much shorter worms than 
the other two figures and the line mz in fig. 23 represents there- 
fore a much larger scale of magnification than in figs. 24 and 
25. The difference in scale between figs. 24 and 25 is slight. 

If we accept the results of these series of experiments we must 
conclude that divisions occur in general more rapidly in the post- 
pharyngeal zoids than in the anterior zodid with fully developed 
head. This difference is probably connected with the fact that 
the postpharyngeal zoéids in Planaria remain at an early stage 
of development as long as their continuity with more anterior 
parts persists. Doubtless individual differences in the sequence 
of physiological isolations exist as they do in other formswhich 
undergo fission in a similar way. Such differences are doubtless 
connected with differences in the dynamic activity of the various 
head regions, in the length of body over which these regions are 
dominant and in the rate of growth. Indications of these differ- 
ences appear in the sequence of fissions in different cases. When 
very long worms like those of Series 298 and 297 undergo fission 
the first separation may occur at the anterior end of zodid 2.2. + 
and a second fission a day or two later at the anterior end of 
zooid 2.1. In ease this zodid has already divided, as in fig. 25, 
into 2.1.1. and 2.1.2. these two zodids may separate later. On the 
other hand the first fission of the worm may occur at the anterior 
end of zodid 2.1. (2.1.1. in fig. 25) and the posterior product of 
this fission may divide later at the anterior end of 2.2. + or at 
the anterior end of 2.1.2. Separation between zodids 1.1. and 
1.2. rarely or never occurs at the first fission of these worms, 
but after the other posterior zodids have separated zodid 1.2. 
sometimes separates without further feeding and growth, though 
in most cases this fission does not occur except after growth. 

Incidentally the three series give us further data on the relation 
between length of piece and regulatory capacity (Child, ’11b). 
This relation is particularly well shown in the difference in height 
of the summits of the curves in figs. 24. and 25. In the two series 
worms of almost the same total length were used and in the one 
(fig. 24) the postpharyngeal regions were cut into eight pieces, in 


256 Cc. M. CHILD 


the other (fig. 25) into twelve pieces. Both the tables and the 
curves show that in the latter case the frequency of head forma- 
tion is much less than in the former. 


6. THE FACTOR OF DISTANCE IN THE FORMATION OF NEW ZOOIDS 
AND ITS RELATION TO THE STAGE OF DEVELOPMENT OF THE 
DOMINANT REGION 


In Planaria the new posterior zodid remains at an early stage 
of development until separation occurs, consequently the boun- 
daries of the different zodids do not become visible as they do in 
the Microstomidae and in various annelids which undergo fission. 
It is therefore impossible to determine the exact position of the 
anterior ends of different zodids or the exact sequence of their 
formation. 

But if we combine the facts obtained by inducing fission in 
various ways (Child, ’10) with the results of experiments like 
those described in the preceding section it becomes evident that 
the formation of new zodids in Planaria takes place according to 
a definite law. 

If the formation of a new zodid at the posterior end of an indi- 
vidual or another zoéid in Planaria isthe result of the physiological 
isolation of the region concerned from the dominant head region 
then two factors must determine when and where new zodids 
shall form. These two factors are, first: the length of body which 
the head region is capable of dominating and second the rate of 
growth. 

As regards the first of these factors, we find that the length of 
body which the head region is capable of dominating is by no 
means a constant quantity, but differs at different stages of devel- 
opment and under different physiological conditions. In general 
this length inereases as development proceeds, at least up to a 
certain point. Figs. 26, 27 and 28 in which the boundaries of 
the posterior zodids are indicated in animals of 5, 12 to 14 and 
28 mm. show this very clearly. In fig. 26 the young animal of 
5 mm. has already divided: as a matter of fact, it is possible to 
induce fission in animals not much larger than this (Child, *10) 
and below the limit where fission occurs the behavior of pieces 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 257 


from the postpharyngeal region in regulation shows very clearly 
that a second zoéid is already present there. In the animal of 
12 to 14 mm. (fig. 27) the anterior zodid has about doubled its 
length but has not divided again, while the posterior zodid, 
although much shorter than the anterior, has already divided. 
In some eases, and especially if growth is slow, this division of 
the posterior zo6id does not occur until the worms are somewhat 
longer than this, but in all cases, so far as my observations go 
it occurs before the division of the anterior zodid. In worms of 
28 mm. (fig. 28) the anterior zodid has divided again, but the 
anterior product of this division, zodid 1.1. is three times as long 
as the whole worm in fig. 26. Evidently the length of body over 
which the head region is dominant increases greatly during devel- 
opment. In the posterior zodids, on the other hand, where the 
head region remains at an early stage of development the zodids 
divide at lengths which correspond much more closely to the 
length of the young animal at the time of its first division. 

The facts indicate clearly then that an extension of the domin- 
ance of the head region over parts posterior to it occurs during 
development in Planaria. Physiological isolation of the poste- 
rior end is therefore possible in much shorter animals in early 
stages of development than in later stages. 

Turning to the second factor in the formation of new zodids, 
the rate of growth, it is evident that the rate at which growth in 
length occurs will determine the rate at which divisions occur. 
But the question at once arises as to whether the rate of extension 
of dominance always coincides with the rate of growth in length. 
The facts indicate that it does not. My own observations on 
normal worms show very clearly that the slowly growing animal 
attains a much greater length without division than the one which 
is growing rapidly. By inducing very rapid growth fission can 
be made to occur in the natural way and without special stimula- 
tion in animals of 12 to 14 mm. or even shorter, while animals 
kept under conditions where growth is slow often attain a length 
of 25 mm. without fission. In general, the higher the rate of 
growth the shorter the animal when it divides. At the time my 
paper on physiological isolation and fission in Planaria (Child, 710) 


~ 


LIN) 


20 


27 
Worms of different length, showing approximate position and num- 
Fig. 27, 14 mm., at least three zodids- 


Figs. 26-28 
ber of zodids. Fig. 26, 5 mm., two zooids. 
Fig. 28, 28 mm., at least five zodids. 

258 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 259 


was written I had not recognized this fact clearly, although some 
of the data at hand at that time point to such a conclusion. 

This fact is of great importance for the understanding of the 
process of formation of new zodids, not only in Planaria but in 
other forms as well. It means that an individual or a zodid does 
not necessarily attain either a certain constant length or a cer- 
tain stage of development before it divides. It may divide at 
any length above a certain minimum or at any stage of develop- 
ment and the lengths and stage of development at which it does 
divide depend upon the relation between the rate of extension of 
dominance and the rate of growth in length and this is determined 
by the conditions under which the individual or zodéid is living. 
With a high rate of growth in length division occurs in shorter 
animals and at an earlier stage of development; with a lower 
rate of growth the animals attain a greater length and a more 
advanced stage of development before dividing, and when the rate 
of growth is very low they may attain a maxmum size without 
dividing at all, for under such conditions the rate of extension 
of dominance may keep pace with the rate of growth in length 
until the limit of growth is attained. 

To sum up: the facts as regards division in Planaria are these: 
first the length which the individual or zodid attains before it 
divides varies with the stage of development of its anterior domi- 
nant region and second, the higher the rate of growth in length the 
less the length attained before division and vice versa. These 
facts are readily accounted for by the hypothesis that the rate 
of growth in length and the rate of extension of dominance are 
not necessarily identical. According to this hypothesis division 
may occur at any length above a certain minimum and at any 
stage of development, according to the relation between the rate 
of growth and the rate of extension of dominance. In Planaria 
division does occur in individuals and zodids of very. different 
length and in very different stage of development but this hy- 
pothesis accounts for all the facts and I believe we cannot account 
for them in any other way. The question as to how and why the 
extension of dominance occurs during development is one of 
considerable interest and we shall return to it below after a brief 
review of the process of fission in Stenostomum. 


260 Cc. M. CHILD 


In Stenostomum and Microstomum and in various annelids 
chains of zodids arise in much the same manner as in Planaria and 
in those forms the morphological differentiation of the zoéid 
as a new individual takes place while it is still in organic continu- 
ity with more anterior regions and each zoédid becomes visible 
at a very early stage of its development. Some years ago, in con- 
nection with experiments on regulation I devoted considerable 
time to a study of the process of division in Stenostomum most 
of the results of which have not as yet been published. Since the 


L LLL. 
Ll. 
I2. LL2 
P23 
T2 
2 


29 


JO 


N) 


Figs. 29-31 Stenostomum chains showing sequence 
of divisions and lengths of zodids under natural con- 
ditions. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 261 


process of division in this form is in many respects similar to 
that in Planaria and since the visibility of the zodids makes its 
analysis less difficult a brief comparison of the two forms will 
serve to bring out some points of interest. In general the same 
law holds for both forms, although certain differences exist. 

In Stenostomum the first division of a single individual occurs 
in the posterior region of the body as indicated in fig.29. There 
is more or less variation in the level of the fission plane in different 
individuals at the time of its first appearance, but in well fed 
animals at a temperature of about 20° C. and with sufficient 
oxygen the first fission plane appears posterior to the middle. 
After its formation both zodids inerease in length and in the sec- 
ond division the anterior zodid, zodid 1., with fully developed 
head divides into a long anterior zoéid, 1.1. and a short posterior 
zooid, 1.2. (fig. 30). Somewhat later zodid 2. divides into two 
zooids nearly equal in length (2.1. and 2.2., fig. 31). Comparison 
of figs. 30 and 31 shows that the anterior products of these two 
divisions are unequal in length, the zodid 1.1. (fig. 30) with fully 
developed head being much longer than zoéid 2.1. (fig. 31) in 
which the head region is not fully developed, while zodid 2.2. 
(fig. 31), even without including the slender tail, is longer than 
zooid 1.2. (fig. 30). In the third series of divisions the anterior 
zooid with developed head again attains a greater length than 
others before division although its division precedes the division 
of other zodids in time and in this third division it once more 
divides into a long anterior and a short posterior zoéid (fig. 31, 
1.1.1. and 1.1.2.). Zodid 1.2. (figs. 30 and 31) divides next into 
nearly equal parts (fig. 32, 1.2.1. and 1.2.2.). Moreover, this 
zooid in which the head region is still at a rather early stage of 
development divides when its total length is much less than that of 
the anterior zodid with fully developed head. The figures show 
this point clearly: the combined length of zodids 1.2.1. and 1.2.2. 
in fig. 32 is much less than that of the anterior zoéid of the chain 
at any stage, and the difference between the anterior product of 
this division (1.2.1., fig. 832) and the zodid 1.1.1. in the same figure 
or zoéid 1.1. in fig. 80 or even zoéid 1. in fig. 29 is striking. 


262 Cc. M. CHILD 


The next zodid to divide in the third series of divisions is zodid 
2.1. (fig. 31). Its total length when it divides is greater than 
that of the zodid anterior to it (zodid 1.2.) but less than that of 
the most anterior zo6id of the chain. Moreover, zoéid 2.1. divides 
into a longer anterior and a shorter posterior part (fig. 32, 2.1.1. 
and 2.1.2). The development of the most posterior zodid of a 
chain is always slow and the third cycle of divisions does not 
usually occur in this zodid until after separation at the fission 
plane between zodids 1.2.2. and 2.1.1. (fig. 32) has occurred. 
Occasionally, however, and under certain conditions its division 
occurs somewhat earlier. In fig. 33 a case of this kind is shown. 
This zodid, zodid 2.2. of figs. 31 and 32, has divided in fig. 33 
into zodids 2.2.1. and 2.2.2. and of these the anterior is shorter 
than the posterior, or if we exclude the tail they are of about the 
same length. 

After the separation of the chain into two parts further divi- 
sions continue to occur in the same way. The figures of Steno- 
stomum chains are all drawn from measurements made with a 
micrometer. Thelength of theStenostomum chain when extended 
is fairly constant and with a little practice measurements can be 
made with a considerable degree of accuracy. In any case there 
is no room for doubt that the differences in length of the zodids 
in the figures are not in any way connected with errors of measure- 
ment. These differences can be seen without the slightest diffi- 
culty in any chain undergoing division under normal conditions. 
My measurements were made merely for the purpose of avoiding 
possible exaggerations and other errors which might result from 
the freehand drawing of figures. In the zodids of Stenostomum 
the ciliated pits, which are shown in the figures serve very well 
as an index of the stage of development of the head region, for 
their development is closely associated with that of the cephalic 
ganglia, which are the most essential part of the head. In figs. 
29 to 33, instead of drawing in the cell masses of the ganglia 
and the developing pharynx which are clearly visible in the living 
animals and which were shown in the drawings made in my notes, 
I have used the ciliated pits alone as an index of the stage of 
development of the head. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 263 


LEE Ss 
LEI GE 


LLZ 


LL2 


T21 L221, 


L22. 


2LL 211, 


212. 


22. 


Sih 


JS4 


Figs. 32-34 Stenostomum chains showing sequence of divisions and lengths of 
zooids. Fig. 32, natural conditions. Fig. 33, heavy feeding. Fig. 34, low tem- 
perature. 


264 Cc. M. CHILD 


This brief review of the sequence and the positions of the 
different divisions brings out certain facts of importance. In the 
first place we see that the more advanced the development of the 
head region of any zoéid the greater the distance between this 
head region and the new head which arises by the division of the 
zooid. An examination of the figures will show that this is true 
for all zodids and for all divisions. Stating this fact in other 
words, we may say that in any pair of zodids resulting from divi- 
sion of a single zodid, the length of the anterior member of the 
pair is proportional to the degree of development of its head region. 

Furthermore, a comparison of the posterior zoéids of the differ- 
ent pairs shows that the posterior products of division differ 
in length at the time of their appearance very much less than the 
anterior products (fig. 29, 2.; fig. 30, 1.2.; fig. 31, 1.1.2., 2.2.; 
fio. 32.012 1:22. Dal. fies 33. eae? at ole see 
As a matter of fact we do find that the posterior member of the 
most posterior pair in a chain (fig. 29, 2.; fig. 31, 2.2.; fig. 33, 
2.2.2.) is always somewhat longer at its first appearance than the 
posterior member of the most anterior pair (fig. 30, 1.2.; fig. 31, 
1.1.2.; fig. 33, 1.1.1.2.).. This difference is due in part to the fact 
that the posterior zodid tapers to a long slender tail at its posterior 
end, but even if the tail is left out of consideration the most pos- 
terior zooid is longer than the posterior member of the most ante- 
rior pair in the earliest stages. Examination of a very larger num- 
ber of chains has convinced me that the differences in length 
between these two zoids are merely the extreme terms of a graded 
series along the axis of the chain. In other words, the length of 
the posterior member of a pair of zodids at its earliest visible 
stage increases as its distance from the anterior end of the chain 
increases. It is somewhat difficult to attain certainty concerning 
this point, for the zodids begin to increase in length as soon as 
they are formed and it is necessary to exercise great care in select- 
ing only the very earliest stages or like stages of development for 
comparison. My notes include a large number of measurements 
of chains which were made before my attention was called to 
this point and in these the difference in length of newly formed 
posterior members of pairs at different levels appears so frequently 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 265 


that I cannot doubt that it really exists. Fig. 33, in which the 
posterior members 1.1.1.2., 2.1.2. and 2.2.2. appeared almost 
simultaneously shows this difference clearly: of these three zodids 
1.1.1.2. is the shortest, 2.1.2. is next in length and 2.2.2. is the 
longest. 

If this fact means anything it means that the minimal length 
of a region which is capable of forming a new individual when 
physiologically isolated from the dominant region increases with 
increasing distance from the anterior end of a chain. In the 
course of my experiments on the regulation of pieces of Steno- 
stomum I found that the minimal length of pieces capable of giving 
rise to new individuals when physically isolated by section was 
almost exactly the same as the length of these posterior members 
of pairs at their first appearance. Moreover, the comparison of 
pieces from different regions of chains indicated that under con- 
stant external conditions the minimal length of such pieces was 
somewhat greater in posterior than in anterior regions of the 
chain, that is, at more posterior levels a somewhat longer piece 
is necessary for the formation of a whole than at more anterior 
levels. This statement I make with some reserve for the differ- 
ence is not sufficiently great to make the conclusion certain. If it 
is correct then the results obtained from the study of natural 
division are in complete agreement with the results of experiment 
with isolated pieces and both indicate the existence of an axial 
gradient similar to that in Planaria, though much less strongly 
marked. 

Another fact of interest in this connection is that in Stenosto- 
mum growth in length occurs more rapidly in anterior than in 
posterior zodids. In general the rate of increase in length in the 
zooids of a chain decreases from the anterior end of the chain pos- 
teriorly. This is shown by the fact that any particular cycle 
of divisions begins in the most anterior zodid and the division of 
the other zodids follows in order. This sequence of divisions is 
shown in figs. 29 to 32. Occasional deviations from this order 
occur in nature or may be produced experimentally. In fig. 
33, for example, which represents a chain that was heavily fed, 
the sequence is less strongly marked than usual. In this chain 


266 Cc. M. CHILD 


the anterior zodid is in advance of all others as is usual and the 
anterior half of the chain is in advance of the posterior half, 
though less so than usual, but in the posterior half the last two 
divisions have occurred almost at the same time, as indicated by 
the stage of development of the posterior members of the two 
pairs. Usually the more anterior of these two divisions occurs 
considerably in advance of the other. The usual sequence of 
divisions, which is connected with the rate of increase in length also 
indicates the existence of an axial gradient of some sort, apparently 
quantitative. 

These facts concerning division in Stenostomum show that 
in this form as, in Planaria, division of an individual or a zoéid 
may occur at very different stages of development and at very 
different lengths above a certain minimum. The Stenostomum 
zooid does not necessarily attain either a certain stage of develop- 
ment or a certain length beforedividing. Moreover Stenostomum 
affords an ocular demonstration of the extension of dominance 
with advancing development. In every pair of zoéids resulting 
from division of a single zoéid, the length of the anterior member 
of the pair is proportional to the stage of development of its ante- 
rior end. This can only mean that the more advanced the 
development of the anterior region of a zodid the greater the 
length of body which it is capable of dominating. In Steno- 
stomum, however, the head region of the original individual com- 
pletes its development and growth before fission begins, conse- 
quently the length of the most anterior zoéid of a chain remains 
approximately constant and no extension of dominance occurs. 
Figs. 29 to 32 illustrate this fact. Fig. 33 is not strictly compara- 
ble with the others because the chain in this case developed under 
conditions which produced a somewhat different zodid-length. 
In Planaria, the head continues to increase in size long after divi- 
sion begins and in correspondence with this fact we find that the 
length of the anterior zodid in Planaria increases greatly during 
erowth. We are certainly justified in concluding that in Stono- 
stomum as in Planaria the occurrence of division depends upon 
the relation between the rate of increase in length and the rate of 
extension of dominance. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 267 


But the processes of division differ in certain respects in the 
two forms: in Planaria the posterior zodids do not develop mor- 
phologically to the point of becoming visible so long as they remain 
in continuity with more anterior parts, while in Stenostomum 
morphological development is very rapid and is practically com- 
plete when separation occurs. And secondly, we have seen that 
in Planaria the posterior zodids apparently precede the anterior, 
at least in most cases, in division, while in Stenostomum the 
sequence is the reverse. I believe that both of these points of 
difference between the two forms are connected with a difference 
in the relation between the rate of increase in length and the rate 
of extension of dominance. 

In the first place, it was pointed out above (p. 257) that the 
oceurrence of actual fission in Planaria is favored by rapid growth. 
Rapidly growing animals undergo fission when much shorter than 
the length at which fission occurs in those that are growing slowly 
and successive fissions follow with greater rapidity in the former 
than in the latter. When growth is very slow in Planaria the 
animals may attain more than twice the usual length and may be 
maintained at that length indefinitely without the occurrence of 
fission, although the isolation of pieces in series according to the 
method described in the preceding section shows that a number of 
zoids may be present in such animals. These facts indicate that 
when the rate of growth is slow the rate of extension of dominance 
may more nearly keep pace with it and so prevent the physiologi- 
cal isolation of posterior regions from reaching the stage at which 
the independent motor reaction and consequent separation of 
the posterior zoéid. Very frequently one sees in these very long 
worms of slow growth what appear to be abortive attempts at 
fission. ‘The region of the posterior zodid attaches itself to the 
substratum, i.e., a motor reaction of a certain degree of independ- 
ence does occur, and the anterior zodid pulls against the resist- 
ance to advance which is thus produced. But as the attempts of 
the anterior zodid to pull itself away become more violent the 
posterior zodid usually relaxes its hold before separation occurs 
and begins to behave as if under the control of the anterior zoéid, 
i.e., its further movements are coérdinated with those of the 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO.3 


268 Cc. M. CHILD 


latter. My interpretation of such cases is that the region of the 
posterior zo6id is physiologically isolated from the anterior under 
ordinary conditions of activity, but that the impulses connected 
with the violent attempts of the anterior zodid to pull itself 
away increase in intensity and effective distance until they finally 
bring the posterior zoéid into coérdination once more. Actual 
fission occurs only when even these impulses fail to overcome the 
independence of the posterior zodid. Apparently then, not only 
the formation of zodids in Planaria but their development to the 
stage where separation is possible depends upon the relation be- 
tween the rate of growth and that of extension of dominance. 

In Stenostomum posterior zodids often show independent 
motor reactions at comparatively early stages but the consistency 
of the tissues is apparently such that rupture does not occur until 
morphogenesis at the fission plane is well advanced. Moreover, 
in Stenostomum the rate of growth is very rapid as compared 
with that of Planaria under the most favorable conditions and I 
believe that the chief differences between the two forms are 
connected with this fact. In Stenostomum physiological isola- 
tion of the posterior region of a zodid or individual and the redit- 
ferentiation of this region into a new zodid occur so rapidly that 
the process is completed before less rapid extension of dominance 
brings the region to a greater or less degree under the control of 
the old head region. In Planaria, on the other hand, the degree 
of physiological isolation is sufficient to initiate the process of 
zooid formation but before this has advanced very far it is at 
least retarded and perhaps inhibited by the extension of dominance. 
The posterior zodids in Planaria then are not as completely physio- 
logically isolated as they are in Stenostomum, consequently their 
development doesnot proceed beyondavery early stage untilactual 
separation occurs and if growth is slow they may even be prevented 
from reaching the stage where separation is possible. 

If these suggestions are correct it should be possible to inhibit 
or accelerate division in Stenostomum as it is in Planaria by alter- 
ing the rate of growth. I have found that it is possible to do this 
in various ways. I hope to describe my experiments along this 
line more fully at another time: for the present a few facts will 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 269 


suffice. The simplest method of inhibiting the formation of 
new zodids in Stenostomum is by keeping the animals at low 
temperature. Fig. 34 shows a chain of two zodids which devel- 
oped from a stage like fig. 29 at a temperature of 10 to 15° C. 
The length of the zodids is very much greater than the usual 
length, as a comparison with figs. 29 to 32 shows. When this 
animal was removed to a higher temperature the new zodids 
produced were of the usual length, and fission planes appeared in 
the zodids that had developed at the low temperature. Similar 
differences can be produced by regulating the food supply, though 
this method is less satisfactory than the temperature method, for 
the animals are voracious feeders and it is difficult to adjust the 
food supply to the proper amount so that growth shall be retarded 
but not entirely inhibited. On the other hand, animals which are 
very heavily fed show accelerated division, i. e., division is not 
only more frequent but the zodids do not attain so great a length 
before dividing. Fig. 33 shows a case of this sort. The total 
length of the chain in fig. 33 is not greater than that in fig. 32, 
but more divisions have occurred in fig. 33. The anterior zoéid 
has begun the fourth cycle of divisions and the third eycle has 
occurred in the posterior zodid and all the zodids between are 
more advanced in development than in fig. 32. Moreover the 
difference is strikingly shown in the difference in length of the 
anterior zo6id of the chain. This is much shorter in fig. 33 than 
in fig. 32. Under like conditions the length of this zodid remains 
fairly constant as is shown by figs 29 to 32. These cases are 
sufficient to show that it is possible to inhibit or accelerate division 
in Stenostomum as in Planaria by altering the rate of growth. 
Theoretically it should be possible, if my conclusions are correct 
to induce the morphological development of zodids in Planaria 
by accelerating the rate of growth, provided it is possible to 
accelerate it sufficiently. I have shown above that we can control 
the degree of development of the posterior zodids in Planaria 
to some extent by altering the rate of growth, but as yet I have 
not attempted to push the acceleration of growth to its limit. 
The second point of difference in the process of division between 
Stenostomum and Planaria, viz., the different sequence of divi- 


270 C. M. CHILD 


sion, is, I think connected with the first. Since the posterior 
zooids in Planaria remain at an early stage of development after 
their formation, in consequence of the extension of dominance 
from the anterior zodid, the length of body which their own head 
regions can dominate is not great and does not increase to any 
great extent. Therefore, as growth occurs these zodids will divide 
while still short and even if the rate of growth in them is the same 
as in the anterior zoéid they will precede it in division. Thus the 
sequence of division in Planaria is due primarily to the retarda- 
tion or inhibition of the development of the posterior zodids after 
their formation. 

In many of the annelids the process of formation of new zodids 
is very similar to that in the flatworms and, so far as my oberva- 
tions go, follows the same laws. In these forms also the relation 
between the rate of growth in length and the extension of domi- 
nance determines the occurrence of division. In some cases the 
extension of dominance reaches a limit early in others late and 
these differences determine differences in the localization or se- 
quence in different forms. Doubtless various other conditions, 
such as advancing senescence which may lead sooner or later to a 
decrease in the length of body over which the head region is 
dominant (Child, 11a, ’11e) also play a part in some cases. In 
all cases, however, the new zoGid forms in consequence of the physi- 
ological isolation of some part, however that may be brought 
about. 

In the present section the extension of dominance has served 
to account for various facts. In conclusion some suggestions as 
to how and why it occurs are perhaps not out of place. As 
regards the fact that the length which a zoéid attains before divi- 
sion is under constant conditions proportional to the degree of 
development there can, I think, be no doubt. It is the question 
as to why this should be so that we have now to consider. Accord- 
ing to my conception the extension of dominance is similar to the 
phenomenon of ‘Bahnung” in the nervous system of higher 
forms. As the zodid develops the paths of correlation undergo a 
‘functional adaptation’ to the passage of impulses. This means 
simply that the passage of impulses over them alters them in some 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 271 


way so that later impulses pass more readily or to a greater dis- 
tanee. If the impulse consists of a transmitted chemical change 
the functional adaptation of the path may mean simply an 
increasing uniformity of constitution which determines that less 
of the energy of the chemical change is lost in other reactions 
which have nothing to do with transmission. Moreover, the 
dominance of the head region in the turbellaria is undoubtedly 
largely dependent in postembryonic stages upon the nervous 
system. The extension of dominance is probably therefore 
largely a matter of the development or “wearing” of paths 
along the nerve cords. In the higher forms the rate of such pro- 
cesses falls far behind that of growth and there is every reason to 
believe that it does so here. 


7. THE DISAPPEARANCE OF ZOOIDS IN CONSEQUENCE OF DE- 
CREASED DISTANCE FROM A HEAD REGION 


In Planaria it is possible to bring about the disappearance of a 
zooid already formed by inducing the development of a head 
near its anterior end. If for example, the postpharyngeal regions 
of long worms like fig. 28 are isolated by section at the level indi- 
cated by a in fig. 35 the pieces become new wholes like fig. 36 with 
normal heads and eyes and with pharynges lying at first anterior 
to the middle. In this process of regulation the first two zodids 
posterior to the cut disappear for they come under the control of 
the new head region and undergo redifferentiation into the pre- 
pharyngeal and pharyngeal regions of the new individual. If the 
body is cut into a series of pieces after regulation is completed we 
find that these regions are similar to the prepharyngeal and 
pharyngeal regions of other individuals and that no trace of the 
previously existing regions of increased head formation which 
marked the anterior ends of the zodids remains. 

But the more posterior zodids, those posterior to 0 and p in 
fig. 35 may continue to exist since they are farther away from the 
new head and fission may even occur before the new head develops 
far enough to control them. In pieces of this kind fission at the 
levels indicated by o and p in fig. 36 is of frequent occurrence 
while the new head is still young, but if fission does not occur at 


272 Cc. M. CHILD 


se 


Figs.35 and 36 Fig. 35, diagram of postpharyngeal region of long worm showing 
zooids and different levels of section. Fig. 36, regulation of region posterior to 
level a in fig. 35: the more anterior zodids are obliterated but the more posterior 


remain. 
that time it never occurs unless the animals are fed and growth 
occurs. In the same way any other zodid can be made to dis- 
appear or its separation prevented. 

In Stenostomum the results of such an experiment are different. 
A new head does not develop near the anterior end of a zodéid 
which is already visible, but the posterior zodid brings about the 
resorption of the headless region anterior to it (Child, ’03). This 
difference is due to the fact that the development of the zodid 
is more advanced in Stenostomum than in Planaria and it becomes 
dominant over the whole piece before a new head can develop 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 273 


at the anterior end. If we shoulduse pieces of Stenostomum con- 
taining only zodids at a stage where they were not yet visible we 
could probably bring about their elimination as in Planaria. 

The results of a series of experiments on inhibition of fission 
in Planaria also serve to show the influence of the new head. 

Series 380. I. Fifty pieces including the region posterior to 
the level 6 in fig. 35; these were cut from well fed worms about 
20 mm. in length possessing two or three well marked zoéids in 
the postpharyngeal region. 

II. Fifty pieces from similar worms but including the region 
posterior to the level ¢ in fig. 35. 

Sets I and II differ only in that the new head forms in Set II 
at a slightly less distance than in Set I from any zodids which 
may be present in the posterior region of the pieces. Fissions 
occurred in the two sets as follows: 


NUMBER OF PIECES! ONE FISSION SECOND FISSION | UNDIVIDED 
bi ‘ a ate 7 | 
Deets sce = nets 50 46 2 | 4 
1} Edo gocrc ota oe 50 25 0 | 25 


Most of these fissions were at the level o in fig. 35, but in at 
least two cases in Set I fission occurred first at the more posterior 
level p and then again at 0. The smaller number of fissions in 
Set II can be due only to the fact that the new head developed at 
a shorter distance from the zodid concerned and so prevented its 
separation. ‘These experiments demonstrate clearly enough the 
importance of the factor of distance in the dominance of a head 
region over more posterior parts. By altering the distance be- 
tween the anterior end of a zodid and a head region we can deter- 
mine either the continued existence of the zoéid and its final 
separation or its redifferentiation as a part of the new animal. 


8. CONCLUSION AND SUMMARY 


In the present paper I have attempted to show that the forma- 
tion of new zodids in Planaria is the result of physiological isola- 
tion of posterior regions of the body from the dominant head 


274 Cc. M. CHILD 


region. If my conception of the process is correct the formation 
of new zodids is not essentially different from the process of regu- 
lation of pieces into wholes after physical isolation. In Planaria 
the development of the zodid beyond a certain stage is prevented 
so long as continuity with more anterior parts persists, but this 
is merely a matter of the degree of physiological isolation. In 
many other forms, e.g., Stenostomum, the development of new 
zooids proceeds almost as if they were physically isolated. 

I believe the idea of physiological isolation will prove very useful 
in accounting for the phenomena of reproduction in both animals 
and plants. It must be remembered, however, that the distance 
limit of dominance and the degree of physiological isolation are 
dependent upon dynamic processes and undoubtedly vary from 
moment to moment. There is no fixed limit of dominance: the 
limit is undoubtedly greater for powerful than for weak impulses. 
In other words the very long planarian, consisting of several 
zooids is undoubtedly more nearly a single individual when it 
reacts very strongly to some stimulus than when it reacts only 
slightly. We must regard the limit of dominance as continually 
changing in the living organism, nevertheless it changes within 
certain limits. As the length of the individual or zoédid approaches 
the limit a critical »eriod must occur during which the posterior 
region is sometimes physiologically isolated to a certain degree and 
at others more completely a part of the original whole. Dur- 
ing the times of greater isolation the processes which would lead 
to reproduction if continued may begin, only to be inhibited and 
their effects removed by a change in the limit of dominance. As 
the periods of greater isolation become longer or more frequent 
with increasing length or with any change in conditions which 
leads to a more permanent decrease in the limit of dominance, the 
processes which lead to the formation of a new zodid continue 
during longer periods or occur more frequently and the new zoéid 
begins to acquire a more permanent character. Sooner or later 
it becomes a continuously existing system and there is no doubt 
that after a certain stage it opposes a certain degree of resistance 
to the impulses which reach it from the dominant region, i.e., 
its receptivity is changed (Child, ’1la). It might almost be said 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 275 


to have acquired a certain momentum. What has probably 
occurred is that new correlations have been established and the 
minuter structure has by this time been so far altered by the 
changed processes that it cannot return at once to its original 
condition. In this way the new system gradually emerges from 
a part of the old. The process is undoubtedly not a continuous 
one in its early stages but a series of oscillations to and fro between 
greater isolation and more complete subordination. Even after 
the new system has become what may be called continuously 
existing it is not necessarily completely isolated. Extreme con- 
ditions may bring it more or less completely under the control of 
the old dominant region either temporarily and only occasionally 
or perhaps permanently. But under ordinary conditions its 
independence gradually increases, still with many oscillations to 
and fro until finally the changes in the structure become great 
enough to be visible. Here the visible morphological develop- 
ment begins, but even this may be retarded or inhibited, or under 
extreme conditions the whole process may be reversed by the 
influence of the old dominant region. Reproduction in con- 
sequence of physiological isolation is not then a simple continuous 
process but rather a series of oscillations. The new zodid may be 
clearly defined dynamically before any evidunce of morphological 
development can be discovered: as I have shown, this is the case 
in Planaria. 

The continued oscillation between subordination and physio- 
logical isolation may be a very important factor in determining 
the development of a zone of structural weakness between the 
new zooid and other parts. The frequent changes in the correl- 
ative factors to which this region is subjected may lead to de- 
generation of its structure so that rupture occurs more readily 
there than elsewhere. 

The relation between the rate of growth and the rate of exten- 
sion of dominance which was discussed in Section VIT seems to me 
to be a factor of great importance in determining the sequence of 
the formation of zodids and the stage of development attained 
before division occurs. The formation of new zodids is most 
likely to occur in animals where the rate of growth is high as 


276 C. M. CHILD 


compared with that of extension of dominance. In a given 
species under normal conditions it is more likely to occur in young 
animals where growth is rapid than in old animals with a slower 
rate of growth, although, as I have shown elsewhere (Child, 710, 
‘11la) it may be induced experimentally by decreasing the domi- 
nance of the anterior region. 

The formation of new zodids in Planaria, as well as in Steno- 
stomum affords further evidence in support of the conclusion 
reached in the preceding paper (Child, ’11f), viz., that in each of 
these relatively simple organisms there is one characteristic or 
fundamental morphogenic reaction which occurs in every isolated 
mass of the specific protoplasm that is capable of continued exist- 
ence and synthesis, provided its rate of reaction is sufficiently 
high. The morphological result of this reaction is the formation 
of an anterior or distal end. 

Experimental data to be described later will show that the rate 
of reaction is a very important factor in morphogenesis. In 
order actually to form a distal or anterior region the isolated mass 
must not only remain alive but the dynamic processes going on 
within it must maintain or exceed a certain minimal rate. More- 
over, it is possible to determine great differences in morphological 
structure, e.g., the teratophthalmiec, teratomorphie and anophthal- 
mic heads, by changes in conditions which have primarily a purely 
quantitative effect. 

If we accept the conclusions reached in this and earlier papers 
of the series; it follows that the formation of new zodids in Planaria 
is essentially a process of regulation resulting from physiological 
isolation as the regulation of pieces results from physical isolation. 
I believe that most if not all forms of asexual reproduction are 
fundamentally similar in character (Child, ’11a). 

Moreover, it seems entirely unnecessary to assume the contin- 
uous existence of any sort of ‘germ plasm’ in connection with 
such processes. The germ plasm of Planaria is the specific 
reaction complex which leads to the formation of a head region and 
in Tubularia the germ plasm is the reaction complex which leads 
to hydranth formation, or more strictly to the formation of the 
distal region of a hydranth. There is not the slightest reason to 


fe leded 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS PAS 


suppose that this reaction complex is continuously existent, for 
the assumption that a mass of protoplasm may react in one 
way at one time and in another at another time certainly presents 
no greater difficulties than the assumption that a specific chemical 
substance may go through one series of reactions under certain 
conditions and through a different series under other conditions, 
and this we know to be a fact. In such a ease we do not regard it 
as necessary to assume that the specifie substance is continuously 
present in all these reactions. In the case of the organism it 
seems much more logical and certainly more in accord with the 
facts to believe that germ plasm arises de novo from somatic plasm 
every time dedifferentiation occurs in consequence of altered 
correlation or other conditions. It is not necessary that the dedif- 
ferentiation should be complete in every case, the new develop- 
ment need not start at the very beginning. A dedifferentiation 
of any degree is merely a more or less close approach to the fun- 
damental dynamic system of the species. In fact, when we think 
of life in dynamic instead of in static terms the conception of 
germ plasm as a continuously existing entity becomes really 
impossible. And finally, that conception does not help us in any 
way to understand or interpret the phenomena or the processes 
of development, it rather makes them more difficult to under- 
stand and all so-called interpretations based on this conception 
are simply paraphrases of the facts observed. 

The group of cells which gives rise to a head in a piece of Pla- 
naria seems tome to constitute as truly a germ as does the egg 
cell. In fact the chief difference between the egg and these cells 
is that the egg is so highly differentiated and physiologically so 
old (Child, ’11e) that a special stimulus from without is neces- 
sary before it can dedifferentiate into germ plasm, while in the 
pieces of Planaria the dedifferentiation follows automatically 
upon physiological or physical isolation. In short the piece of 
Planaria is in reality much nearer the condition of a germ plasm 
than is the egg before fertilization. 

According to my conception, the process of reproduction in- 
volved in the regulatory formation of a head and so of a new whole 
in Planaria, a new hydranth in Tubularia or a new bud in a plant 


278 Cc. M. CHILD 


is a much simpler and more primitive form of reproduction than 
the sexual form. In such simple forms of reproduction the prob- 
lem of heredity appears in its simplest terms and the process of 
inheritance itself becomes accessible to investigation. 

The chief points of the paper are summarized as follows: 

1. In pieces of Planaria possessing a large, fully developed 
head and below a certain length a second zoédid does not appear 
at the posterior end except after feeding and growth, but in head- 
less pieces of the same length and under the same conditions the 
second zodid appears at once and often another zodid forms and 
fission may occur. 

2. If teratophthalmic, anophthalmic or headless pieces are 
fed so that growth occurs they attain a certain length far below 
that of the normal animal under the same conditions and then 
undergo fission and this process is repeated indefinitely. In 
general the teratomorphic pieces attain a greater length before 
fission than the anophthalmic and these a greater length than the 
headless pieces. The second zodids arising from such pieces 
produce normal animals and do not inherit the characteristics of 
their parents. 

3. In starvation in normal animals the second zoéid or the 
posterior zodids gradually come again under the control of the 
dominant head region as the length of the animal decreases and in 
the later stages of starvation they may almost entirely disappear. 
In starving headless pieces, on the other hand, the posterior zodid 
or zodids maintain themselves at the expense of the more anterior 
headless region. 

4. Planaria dorotocephala often consists of four or five or more 
zooids which arise in definite relations to each other. 

5 In Planaria and Stenostomum, as well as in various other 
forms the distance from the head region in any zodid to the 
head region of the new zodid arising at the posterior end of the old 
increases as the development of the anterior head region advances, 
at least up to a certain stage. 

6. The formation of new zodids and fission are favored by rapid 
increase in length. In slowly growing animals division not only 
occurs less frequently than in rapidly growing, but the zodids 
attain a greater length before dividing. 


STUDIES ON THE DYNAMICS OF MORPHOGENESIS 279 


7. It is probable that the paths of correlation in the dominant 
region and between this region and other parts undergo a ‘func- 
tional adaptation’ during development. In Planaria and related 
forms nerve paths are undoubtedly largely concerned in this proc- 
ess, which is similar to the ‘Bahnung’ observed in the nervous 
systems of higher forms. In consequence of this change in the 
paths of correlation an extension of the dominance of the anterior 
region occurs during development and the length of body over 
which the anterior region is dominant increases with advancing 
development up to a certain stage. 

8. The relation between the rate of growth and the rate of exten- 
sion of dominance determines whether and where new zodids 
shall be formed and whether they shall undergo morphological 
development after their formation. According to this law a 
zooid or an individual may divide at any stage of development or 
at any length within certain limits. 

9. Zodids already formed can be made to disappear in Planaria 
or their further development and separation can be prevented by 
decreasing the distance between them and a dominant head region. 

10. The formation of new zoéids in Planaria, Stenostomum and 
various other forms is essentially a process of regulation resulting 
from physiological isolation of parts, just as the formation of 
new wholes from pieces results from physical isolation. 

11. The ‘germ plasm’ of Planaria is essentially the specific 
dynamic system that gives rise to a head region, whether this is 
contained in a single cell or in a mass of cells. Such a system may 
arise de novo wherever sufficient dedifferentiation occurs and there 
is absolutely no reason for assuming that this system is continu- 
ously existent in all parts of the planarian body which are capable 
of forming a head. 

12. The form of reproduction in which a new whole, beginning 
with the dominant region is formed from a part in consequence of 
physiological or physical isolation is a relatively simple and primi- 
tive mode of reproduction as compared with the sexual type, in 
which the egg is so highly differentiated and so old, except in 
cases of parthenogenesis, that a special stimulus from without is 
necessary to initiate the process of dedifferentiation. In such 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 3 


280 


Cc. M. CHILD 


simple forms of reproduction the process of inheritance itself is 
accessible to investigation. 


BIBLIOGRAPHY 


Cuitp, C. M. 1903 Studies on regulation. III. Regulative destruction of 


ScHULTz, 


zodids and parts of zodids in Stenostomum. Arch f. Entwickelungs- 
mech., Bd. 17. 

1910 Physiological isolation of parts and fission in Planaria. Arch. f. 
Entwickelungsmech., Bd. 30 (Festband fiir Roux), Teil IT. 

191la_ Die physiologische Isolation von Teilen des Organismus. Vortr. 
u. Aufs. ii. Entwickelungsmech., H. 11. 

1911b Studies on the dynamics of morphogenesis and inheritance in 
experimental reproduction. I. The axial gradient in Planaria doroto- 
cephala as a limiting factor in regulation. Journ. Exp. Zoél., vol. 10. 
191le A study of senescence and rejuvenescenct based on experiments 
with Planaria dorotocephala. Arch. f. Entwickelungsmech., Bd. 31. 
1911d Experimental control of morphogenesis in the regulation of 
Planaria Biol. Bull., vol. 20. 

191le The regulatory processes in organisms. Journ. Morphol., vol. 
22. 

1911f Studies on the dynamics of morphogenesis and inheritance in 
experimental reproduction. II. Physiological dominance of anterior 
over posterior regions in the regulation of Planaria dorotocephala. 
Journ. Exp. Zoél., vol. 11, No. 3. 


E. 1904 Uber Reduktionen. I. Uber Hungererscheinungen bei 
Planaria lactea. Arch. f. Entwickelungsmech. Bd. 18. 

1906 Uber Reduktionen. II. Uber Hungererscheinungen bei Hydra 
fusea L. Arch. f. Entwickelungsmech., Bd. 21. 

1907 Uber Reduktionen. III. Die Reduktion und Regeneration des 
abgeschnittenen Kiemenkorbes von Clavellina lepadiformis. Arch. 
f. Entwickelungsmech., Bd. 24. 

1908a Uber Reduktionen. IV. Uber Hunger bei Asterias rubens und 
Mytilus bald nach der Metamorphose. Arch. f. Entwickelungsmech., 
Bd. 25. 

1908b Uber umkehrbare Entwicklungsprozesse und ihre Bedeutung 
fiir eine Theorie der Vererbung. Vortr. u. Aufs. ti. Entwickelungs- 
mech., H. 4. 


ON THE BEHAVIOR OF THE DISSOCIATED CELLS 
IN HYDROIDS, ALCYONARIA, AND ASTERIAS! 


H. V. WILSON 
From the Zoological Laboratory of the University of North Carolina and the Beau- 
fort Laboratory of the U. S. Bureau of Fisheries 


THIRTY FIGURES 


INTRODUCTION 
1 


It is known that monaxonid sponges may be broken into their 
constituent cells and these will recombine to form restitution 
masses which have the power to transform into perfect sponges 
(Wilson, ’07b, ’11b; Miiller, ’11). Several kinds of cells enter 
into the composition of these masses, not only the indifferent or 
totipotent amoebocytes, but also differentiated elements. What 
is the fate of the latter elements? Do they undergo a process of 
de-specialization, passing into an indifferent or totipotent state, 
in which they persist as part of the restitution mass when it begins 
to transform? Or are they incorporated and digested by the 
amoebocytes? In general terms, in the development of such 
masses does regressive differentiation of cells into an indifferent 
condition play an important part? The large number of amoebo- 
cytes, at any rate in the monaxonida, makes it difficult to decide 
this question in the case of sponges. 

It seemed that a study of the coelenterates might throw hght 
on the matter. In the hydroids we have two tissue layers, ecto- 
derm and entoderm, and a comparatively small amount of mate- 
rial corresponding to the totipotent amoebocytes of sponges. 
The question was formulated: if all the anatomical connections 

‘Published with the permission of Hon. Geo. M. Bowers, U. 8S. Commissioner 
of Fisheries. The experimental work was carried on at the Beaufort Laboratory 
of the U. S. Bureau of Fisheries during the summer of 1910. Brief mention of 
the results has already been made (Wilson, ‘lla; ‘11b). 

281 


282 H. V. WILSON 


and physiological interrelationships, due to position, between the 
cells composing the hydroid layers, were broken up, would the 
cells recombine and form masses of totipotent regenerative tissue? 

Experiments were conducted on two hydroids, Pennaria tiarella 
and Eudendrium carneum. The phenomena were essentially the 
same in the two forms. The hydroid or, in one experiment, the 
coenosare only was cut into small fragments and these were 
pressed through gauze (fine meshed silk bolting cloth). The flesh 
is thus broken up into cells and small cell aggregates. Fusion 
between these begins quickly and large masses are produced 
which exhibit a slow amoeboid change of shape. Continued 
fusion goes on with the formation of massive lumps of tissue or in 
other cases of sheets. The size and shape of such masses are under 
control. These bodies acquire a smooth surface and secrete a 
perisare within about one day. They are solid and of a syncytial 
structure, although cell bodies are here and there marked out. 
During the two or three days following their formation, the 
bodies are subject to great mortality. Many isolated masses or 
nodules of larger masses, however, remain alive and differentiate 
ectoderm and entoderm layers, together with a central yolk mass, 
much as in the case of a coelenterate planula. These now send 
out cylindrical coenosarcal outgrowths which under favorable 
conditions continue to grow and produce well formed hydranths. 
Such hydranths appear to be normal. In the case of Pennaria 
they developed both the characteristic sets of tentacles and were 
equal in size to the smaller subapical hydranths of the adult, and 
were of about the same size as hydranths obtained from metamor- 
phosing planulas. In Eudendrium also hydranths of adult 
size with the characteristic hypostome and number of tentacles 
were obtained in this way. 

We apparently have here in the hydroids a plain case of the de- 
specialization of tissue elements and their union to form masses of 
totipotent regenerative tissue. The indifferent elements which 
give rise to the sex cells or in some forms to buds (Braem, ’08) are 
practically absent in the coenosare, and yet as experiment shows 
the coenosareal tissue will give rise to the restitution masses. 
After their sudden and violent separation the coenosarcal cells are 


BEHAVIOR OF DISSOCIATED CELLS 283 


spheroidal and without vacuoles, whereas in the hydroid itself 
they are in some measure vacuolated and provided with out- 
growths which probably all interconnect. This change in appear- 
ance is perhaps correlated with a change in physiological state 
whereby they pass as a result of the shock and isolation into an 
indifferent condition, in which condition they unite to build up 
the restitution mass. The mass differentiates like a planula, the 
outer stratum becoming the ectoderm, while the inner mass gives 
rise to an entoderm and material that is used up as food. After 
the several kinds of cells unite to form the restitution mass, they 
undergo changes and it becomes impossible to trace them. It 
is conceivable that some degree of specification persists and that 
elements derived originally from the ectoderm migrate in the mass 
until they reach the region of the surface, while the elements 
derived from the entoderm shift towards the interior of the mass. 
However, I find no facts in support of this idea, and it seems to 
me extremely improbable. - 

I conceive then that in the formation of the hydroid restitution 
masses the phenomenon is essentially one of regressive differentia- 
tion. The cells of the hydroid layers perhaps as a result of isola- 
tion, perhaps in part as a result of the shock, pass quickly into a 
simplified indifferent state. The restitution masses of sponges 
are so like those of hydroids that in them too regressive differen- 
tiation of tissue elements probably plays a great part. 


2 


It may be anticipated that there are other animals besides 
sponges and hydroids in which the somatic cells when forcibly 
disjoined, will fuse and give rise to totipotent regenerative mate- 
rial. Perhaps where the body cells in general have not this power, 
it may still be possessed by comparatively undifferentiated amoebo- 
cytes in such forms as ascidians. And it is possible that the sex 
cells in some animals before they have progressed too far in their 
special differentiation, may possess this power. From _ these 
standpoints a few observations were made on the common aleyon- 
arian Leptogorgia, and on the immature gonads of the starfish, 
Asterias. 


284 H. V. WILSON 


For Leptogorgia it was shown that when short pieces of the 
colony are pressed through cloth as before, or simply squeezed 
under water with forceps, small lumps of tissue and isolated cells 
are pressed out. An active fusion goes on between all these, with 
display of amoeboid phenomena. Masses are thus formed which 
become more or less spheroidal and acquire a smooth surface. 
By bringing together the smaller ones, bodies of considerable size 
up to and over 1 mm. in diameter may be produced. These 
bodies remained alive in laboratory dishes for many days, but 
exhibited no metamorphosis. 

In the case of Asterias, immature gonads about an inch long 
were cut up and pressed through gauze. Abundant cells and cell 
masses were thus obtained, which fused actively forming reticu- 
lated plates and massive lumps. ‘These soon acquired a smooth 
surface. They remained alive in laboratory dishes for a couple 
of days, but underwent no further change. 

Experience in the handling of regenerative sponge masses sug- 
gests that possibly in the case of the aleyonarian the fusion masses 
might metamorphose if transferred to the harbor water. It 
seems incredible that anything in the shape of an individual could 
come from the fused germ cells of Asterias, however probable 
this might be in the case of a coelenterate. On the other hand, 
where the mass of cells is unable to give activity to the regenera- 
tive power when removed from the body fluids to water, it might 
conceivably display regenerative power, in some degree, if replaced 
in a proper body. Perhaps in some animals it might give rise 
to a bud individual, or under other conditions become merged into 
the tissues of the locality, or die gradually in some process of 
absorption, or be extruded as something foreign, or finally pass 
into one of the categories of tumors. From this point of view, an 
experiment was started to determine what the Leptogorgia balls, 
above described, would do if replaced in the Leptogorgia body. 
The experiment which must be regarded as no more than a tenta- 
tive one is described farther on in the paper. 


BEHAVIOR OF DISSOCIATED CELLS 285 


3 

Some facts concerning the regressive differentiation of cells 
have long been familiar. The choanocytes of fresh water sponges 
for instance, it is known since the time of Lieberkiihn’s investiga- 
tions (’56), give up during the winter their distinctive features and 
assume the character of mesenchyme elements. Metschnikoff 
observed the same phenomenon in marine sponges that were kept 
in foul water (79). The flagellated chambers in both these 
cases again developed with the return of proper conditions, but 
there is no reason to suppose that the same cells that once were 
choanocytes again became transformed into such elements. It 
is possible, of course, that this is so, in which case the de-speciali- 
zation of the choanocytes is partial and temporary, and a complete 
return to the indifferent (totipotent) state of the typical amoebo- 
cyte is not made. On the other hand it is certainly possible that 
the choanocytes do retrogress into this state, from which they 
ontogenetically arise in the gemmule development of the spongil- 
lidae and at any rate in a percentage of individuals in some cases 
of larval metamorphosis (Wilson, 94; Evans, ’99). A third possi- 
bility, that they are used up as food material, must also be con- 
sidered, until more intensive investigation settles the question. 

The idea of a thorough going de-specialization of cells into 
indifferent elements underlay the older account of the origin of 
gemmules in the spongillidae, according to which the gemmule is 
made up of the transformed cells, of all kinds, located in a region 
of the sponge body. More precise investigations have shown that 
the spongillid gemmule and some other asexual reproductive 
bodies in sponges (‘Tethya buds, e. g., Maas, ’01) arise as a conge- 
ries of similar mesenchyme cells. The origin of these cells espe- 
cially in cases where the sponge body degenerates, is a matter of 
physiological interest, and deserves to be better known. It is 
not impossible, as I have said (’07b, p. 252) that they are “groups 
of amoebocytes which are in part recruited from transformed 
collar cells and other tissue cells, such as pinacocytes (flat cells of 
canal walls), that have undergone regressive differentiation into an 
unspecialized amoeboid condition.” Maas (710, p. 124) more 
recently advances the same idea. 


286 H. V. WILSON 


In recent times the question of the de-specialization of cells 
has become a prominent one. Some pathologists as is well known 
believe this occurrence to be at the bottom of tumor formation. 
And in regenerative phenomena of many kinds it has been shown 
that cells lose their specific features, become less specialized, and 
re-acquire a capacity for new differentiation. Schultz reviews 
numerous cases in his essay on reduction (’08). In none of the 
instances, however, which he records as occurring in animals is 
there an obvious assumption of the totipotent character. Possi- 
bly this occurs in certain tumors and in the case of Moniezia 
reported by Child (’06). In Moniezia muscle cells lose their 
fibrillae and become spermatogonia, and it may be that this 
behavior is to be interpreted as a return to the totipotent charac- 
ter followed by a subsequent differentiation into sex cells. In 
the well known case first described by Driesch (02) and again 
studied by Schultz (’07) where the excised branchial sac of asci- 
dians is remodelled into the condition of a stolon bud, which then 
transforms into a small ascidian, the cellular metamorphosis 
stops short in its backward path before the condition of totip- 
otent regenerative tissue is reached. In the reduction of starv- 
ing hydras studied by Schultz (’06), the hydra body is greatly 
simplified, becoming a mouthless spheroidal sac. But in this 
sae the two layers, ectoderm and entoderm, persist, and although 
some of the cells assume an embryonic appearance, there is no 
obvious metamorphosis of cells into totipotent regenerative tissue. 
Still one striking feature of this case of reduction, is the great 
development of sex cells (male) which increase in number and 
ripen while the body in general dwindles, and it is conceivable 
that here, as in Moniezia, we have the transformation of cells 
into indifferent elements which then differentiate into male germ 
cells. In the reduction of hydra, the spheroids eventually degener- 
erate and die. They do not undergo a regenerative awakening 
as in the case of the ascidian branchial sac. And this in turn 
may be due to the differentiation of the indifferent elements, as 
fast as they are formed, into germ cells. 

In a ease of reduction in the coelenterates which is not recorded 
by Schultz, there would appear to be extensive de-specializa- 
tion of cells. I refer to the observations of Perkins on the larva 


BEHAV{OR OF DISSOCIATED CELLS 287 


of the medusa, Gonionema (’02). Perkins observed that the hy- 
dra-like larvae of Gonionema after they had been for some time 
in an aquarium retracted their tentacles and became transformed 
into shapeless masses which had the power to throw out pseudo- _ 
podia and ereep about. In such plasmodial bodies the original 
differentiation into layers and even cells appeared to be lost, while 
the nuclei remained visible scattered unevenly through the sub- 
stance. These bodies continued alive for two months and under- 
went repeated fission until owing to their diminution in size it 
became impossible to follow their history. It is possible that 
under favorable conditions these masses would have behaved 
like the restitution bodies of Pennaria and Eudendrium, and once 
more have developed the normal structure of the species. 

Regressive differentiation of cells undoubtedly occurs in sponges 
when owing to confinement or abnormal chemical environment the 
body gradually breaks down into masses of regenerative material. 
I have recently (’11b) touched upon the question as to whether 
in these instances the de-specialized cells persist as part of the 
regenerative tissue. The latest accounts (Maas, 710; Miiller, 
‘11b) indicate that while the choanocytes are absorbed, other 
de-specialized cells persist. 


4 


Eugen Schultz in a series of papers (04, ’06, ’07,’08) discusses 
a variety of phenomena which he groups under the head of ‘re- 
duction.’ Under this head he classifies such simplifications of 
structure as the normal cyclical loss of gonads and associated 
organs in Planaria (Curtis, ’02), the loss of organs and cellular 
changes in starving planarians, the changes occurring in starving 
hydras that have been referred to, and the remodelling of the 
ascidian branchial sac into the condition of a stolon bud. While 
it is not clear, as Schultz admits, that all of these simplification 
processes should be regarded as belonging in one category, a con- 
sideration of them and of other cases leads Schultz to conceive 
of a phase in the life history of organisms which may or may not 


288 H. V. WILSON 


oecur according to circumstances. To this phase he applies the 
term, reduction, and he conceives of it as an inverse process to 
ontogeny. Under unfavorable circumstances a simplification proc- 
ess may set in, in the course of which differentiations that were 
acquired during ontogeny are given up in inverse order. Such 
an orderly retrogressive course of events may, like ontogeny, be 
disturbed by coenogeneticadaptations. In this process, the organ- 
ism acts as a whole, that is the course of reduction is determined 
not by a struggle for existence between the different kinds of 
cells, but rather the needs of the entire organism dictate what 
structures shall be sacrificed. The behavior of the organism in 
reduction is thus looked on as an adaptive response. What there 
is of new and old in these ideas, is sufficiently stated in Schultz’s 
essay (’08). 

Schultz’s theory of a reduction process may be applied to the 
behavior of sponges which give up their organization under the 
influence of confinement (Wilson, ’07a; Miiller, ’11b) or abnormal 
chemical environment (Maas, ’06). In the monaxonid Stylo- 
tella, for instance, I have found (’07a) that the oscula and the 
bulk of the pores close, much of the canal system is suppressed, 
the skeletal arrangement is simplified, and the flagellated cham- 
bers for the most part broken up into their constituent cells which 
become scattered in the mesenchyme. At any time the sponge 
may be made to reassume its normal differentiation on transfer 
to better conditions. The passage of the body into this simplified 
condition, which is essentially like the winter state of some spon- 
gillidae (Weltner, 93), is obviously an adaptive response on the 
part of the whole sponge, which thus protects itself against the 
bad water of the aquarium, as the spongillid does against extreme 
cold. Moreover this simplified state is very similar to a stage in 
the metamorphosis of such a mass of totipotent regenerative 
tissue as a sponge gemmule. In them both we have a simple 
flat epithelial covering layer and an internal mass, which in the 
case of the gemmule is composed of indifferent amoebocytes and 
in the case of the ‘reduced’ sponge is largely so composed. Or 
again it is much like the ‘pupal’ stage (just after fixation) of those 
larvae of silicious sponges in which amoeboid cells split up to 


BEHAVIOR OF DISSOCIATED CELLS 289 


form the choanocytes.2. The stage then may fairly be looked on as 
repeating an embryonic stage that actually occurs, whether or no 
it be a coenogenetic modification of the more typical (palingenetic) 
course of development. 

In the later history of this behavior of sponges in confinement, 
etc., the sponge ceases to act as an individual. It breaks up either 
with (Stylotella) or without (Calearea, Spongillidae) considerable 
death into numerous masses of totipotent tissue. If we are to 
apply the reduction theory here we must assume that owing to 
coenogenetic adaptation, the sponge body does not continue to 
dwindle and simplify itself until it reaches the condition of an 
egg, or an individual heap of blastomeres, or a single gemmule- 
like mass of totipotent cells. This would entail a tremendous 
loss of substance and altogether is a process which it is foolish 
to contemplate as a possibility. In the backward development of 
the sponge, coenogenesis prevents a too strict adherence to the 


2Maas in his recent study (’10) of the cellular changes involved in the break- 
ing down of a sponge body into masses of regenerative tissue, views the phenomena 
from a standpoint not strictly that of the reduction-theory. This is not the 
place for a discussion of Maas’ standpoint, but I may mention that he debates 
the question as to whether a fundamental similarity exists between late stages 
in the reduction of sponges and the ‘pupal’ stage in ontogeny and decides there is 
none, because in the former the interior is chiefly made up of amoebocytes and in 
the latter of immigrated ectodermal cells destined to become directly transformed 
into choanocytes. There are still a good many unsolved problems concerning 
the behavior of the layers in sponge ontogeny, and among these, if we accept the 
current view that the choanocytes typically arise from the ciliated covering cells 
of the larva, is the relation between the typical pupae and thosein which the inte- 
rior is chiefly filled with amoebocytes, or as I have called them ‘formative cells’ 
(94), which divide up and form the choanocytes. My account (’94), it seems to me, 
established the fact that such pupae are exceedingly common in certain species 
of monaxonida, so common that I regarded them as typical. There is no doubt 
that in my work of that time the bulk of the sponges reared came from pupae of 
this sort. Evans in studying spongillas several years later found (’99) that the 
choanoeytes not infrequently arose in this way, as I and writers before me had de- 
seribed in detail. The occurrence of such pupae cannot be passed over as patho- 
logical, and if we keep them in view instead of the type which is densely filled with 
small cells, the similarity between the pupal stage and the reduced sponge is 
obvious. Maas has some comments on larvae of this type in Sycons (10, p. 105), 
and is inclined to believe that they are to be looked on as individuals in which the 
amoebocytes have actually incorporated and digested the immigrated ectodermal 
elements. 


290 H. V. WILSON 


path over which the organism has come, by setting into activity 
a process of division. The sponge body in the simplified condi- 
tion above described may conceivably divide up again and again 
until very small masses result which then go over the last steps 
of reduction, transforming themselves each into a single embryo- 
like heap of totipotent cells. Or the division may stop when the 
masses resulting therefrom are still of considerable size. Or a 
division of far-reaching extent may take place which consists in 
the breaking of all ties, anatomical and physiological, between 
the component cells. Such a process would result in the virtual 
dissolution of the sponge body into an enormous number of 
independent cells, and these, or such as could, might then go 
through the last phases of reduction becoming totipotent. In the 
actual behavior of the sponges all these varieties of the division 
process are associated together in a complex and over-lapping 
fashion. 

When the division process leads to the production of independ- 
ent cells, these wander about on or through the old skeleton 
and are attracted to one another or to masses of reduced tissue, 
and so unite to build up or aid in building up such masses. This 
is a useful habit to such cells. In the body of an animal, since by 
hypothesis they are totipotent, they might conceivably through 
growth and division give rise to a group large enough to transform 
into an organism of that species. But out of the body or cut 
off from the possibility of growth, their only hope for life would 
lie in fusion. Only so could they give rise to organisms. ‘The 
habit of fusion displayed by these elements being useful may have 
been aequired. More probably it has been preserved as an inher- 
itance from early ancestors. 

In the retrograde differentiation of organisms it is conceivable 
that there might be a difference in the path followed according 
as to whether the individual had its origin in an asexual mass or 
in an egg. Schultz in passing suggests (’07) that possibly some 
difference might be found between the reduction process of asci- 
dians derived from eggs and buds. Following out the logic of the 
theory we should expect in the one case to find stages resembling 
the free larva and egg, in the other case no such stages. But 


BEHAVIOR OF DISSOCIATED CELLS 291 


such a difference seems one too good to be hoped for—coenogenesis 
would surely not allow the oozooid to go back through the com- 
plications of tadpole larva and egg development. Even with 
lower forms, sponges and coelenterates, it is incredible that the 
path of reduction would ever lead through the stage of the ciliated 
larva to an egg, which by virtue of its egg nature, in contrast 
with that of a simple totipotent cell, would again in its upward 
differentiation develop into a ciliated larva! It is indeed highly 
probable that coenogenesis would usually take care, in animals 
capable of asexual development that reduction should lead to the 
formation of bud-like or in other cases gemmule-like anlages 
rather than to the production of eggs and sperm. The idea is 
that the bud with its quick development or the gemmule-like 
anlage made up of totipotent regenerative cells, unhampered by 
a tendency to retrace the path of phylogenesis and free to develop 
at once into an organism, would present a great advantage over 
an egg to an animal seeking, so to speak, to restore itself. 

It may be seen then that with some use of the coenogenesis 
idea it is possible to view the whole set of retrogressive changes 
undergone by sponges in confinement or under abnormal chem- 
ical conditions as reduction phenomena. Schultz’s theory it 
seems to me enables us to get a better picture of the facts as a 
whole than is possible without it. Miiller evidently entertains 
the same opinion, since he classifies (‘11b) the changes in the 
spongillidae under the caption of reduction. 


o 


In the preceding section, it has been shown that the gradual 
breaking down of a sponge body into small masses and eventually, 
in part, into independent and indifferent cells may be thought 
of as a case of reduction, as a return to an embryonic condition. 
When we compare with such phenomena the forcible breaking 
down of a sponge or hydroid into elements which can live inde- 
pendently for a time and which unite to build up new organisms, 
it would seem that we have in these elements the same end result 
that is reached in the comparatively slow process of reduction. 


292 H. V. WILSON 


The mechanical division of the flesh and the attendant stimuli 
(shock) apparently, in the case of the pressed out tissue, bring 
on the final stage at once. 

The same facts may be viewed from a somewhat different stand- 
point if we confine ourselves within the narrower limits of phys- 
iology. Child has just published an interesting essay (11), the 
kernel of whichis theidea that when a part of an organism becomes 
physically or physiologically isolated, there is a tendency for it to 
develop into a new whole, provided it has the power of complete 
regulation or in other words, is made up of totipotent material. 
(By physiological isolation, Child means the condition in which 
the part is cut off, through one cause or another, from the exchange 
of stimuli which maintain that mass of material as an individual 
organism.) Such a statement well covers the facts of the pressing 
experiments and the reduction of sponges, provided, (1) we 
remember that the cells as they exist in the stem of the hydroid, 
for instance, are not at the time totipotent, but only become so 
by retracing their development; and provided, (2) we employ the 
category of tropisms for the union of the dissociated cells to form 
masses. 

Finally it may be of interest to consider the question, does 
any phylogenetic significance attach to the power of an organism 
to break up into small bits of indifferent protoplasm which then 
recombine? Did the early ancestors of existing animals practise 
such habits, which in modified form have proved useful and so 
have been retained? It is at least possible that the early organ- 
isms were amorphous masses of protoplasm. Assuming this, it 
is plain how useful the power of easy, quick division might be, 
from how many dangerous situations an organism might escape 
which had the power to break up into parts, these retreating from 
the situation independently, leaving perhaps dead or dying tissue— 
a “sauve qui peut’ proceeding in short. But when we consider 
the difficulties besetting the life of very small plasmodial masses 
of this kind,’ it is also plain how useful would be the power to 
fuse with one another. We may then conceive of these early 


57 have found that the very small plasmodial masses of sponges are at a disad- 
vantage (’07, p: 257). 


BEHAVIOR OF DISSOCIATED CELLS 293 


ancestors as creatures amorphous in shape and inconstant in 
size. According to local conditions and needs of the moment, 
the body divided or fused with a corresponding mass. The 
behavior of the totipotent cells which we have been considering 
may, for speculative purposes, be considered as a survival of such 
primitive habits. 


EUDENDRIUM. RESTITUTION FROM DISSOCIATED CELLS 


Species used. The common Eudendrium of Beaufort harbor 
is E. carneum Clarke (Mem. Bost. Soc. Nat. Hist., III, no. 4, p. 
137; Nutting, 01, p. 333). The most striking characteristic of 
the species concerns the arrangement of the male gonophores, 
which are ‘four or five chambered, borne in a verticil around the 
body of aborted hydranths which are themselves joined to pedicels 
bearing ordinary hydranths, the two being thus borne in pairs 
symmetrically disposed on the branches” (Nutting). 

Experiment July 9. A clean male colony was chopped up with 
scissors into pieces about 3 mm. long and a mass of such pieces 
pressed through gauze in the usual way: the mass was laid on a 
square of gauze, and the latter folded to make a sac, which was 
immersed in a watch glass of sea water and squeezed repeatedly 
with fine forceps. The hydroid flesh streams through the pores 
of the gauze and falls on the bottom as a fine sediment. A little 
is sucked up with a pipette and examined on a slide. It consists 
of isolated cells and minute cell masses with some larger frag- 
ments. The latter are picked out. Fusion is observed to go 
on under the microscope. The tissue is now transferred to watch 
glasses of fresh sea water in which it is shaken towards the center 
with the purpose of facilitating the fusion process, and the watch 
glasses are immersed in large bowls of water. Within a few hours 
time fusion in the watch glasses leads to the formation of irregu- 
larly lobed flattened masses, varying from a fraction of a milli- 
meter to about 5 mm. in diameter with a thickness of 1 mm. or 
less. Such masses exhibit slow changes of shape. All the masses 
of any considerable size are later in the day carefully picked out 
with a large pipette and transferred to fresh sea water. On sub- 


294 H. V. WILSON 


sequent days they are transferred in this fashion three times a 
day. An effort is made to keep them clean and free of the 
bottom. On the day after their formation the surface of all these 
masses is smooth and a thin perisare is to be seen round many of 
them. 

On July 13, some of the masses have begun to transform. In 
fig. 1 is shown one of the smaller from which an outgrowth has 
sprouted. The perisare round the outgrowth, c s., is thinner than 
round the original mass, and in the outgrowth the ectoderm and 
entoderm are clearly distinguishable in the living object. The 
base of the body representing the original mass appears in life 
uniformly opaque. A number of the flattened larger masses 
are in the condition shown in fig. 2, which represents a part of a 
mass 6 mm. long by 3 mm. wide, of an irregular shape, more or 
less lobed, and partially subdivided into spheroidal nodules, 
a, b, c, each with its own perisare. A good deal of the mass is 
dead. One of the spheroidal nodules, c, has sprouted, and in the 
outgrowth, c s, the ectoderm and entoderm are distinct. In 
such a mass the living parts are easily distinguished by their 
bright color and sharp contour. The perisarc is distinguishable 
round some parts of the mass that are dead, but in other such 
places it cannot be made out. In the case of such large masses 
possibly the perisare does not form round the whole mass. The 
nodule or lobe from which an outgrowth has sprouted is often 
not cut off from the general mass by perisare. This is true of the 
masses shown in figs. 2and 4. In other cases, however, the nodule 
that has sprouted is completely surrounded by perisare and is 
merely imbedded in the general mass (fig. 3). 

Some of the more promising masses alive on July 13 were iso- 
lated. Among these was the mass shown in fig. 4. The condi- 
tion of this mass on the next day is shown in fig. 5. During the 
twenty-four hours that have elapsed since fig. 4 was made, out- 
growth ¢ s 1 has died, and the general mass from which the out- 
growths project is dead. Outgrowth cs 2 has however grown 
and bifurcated. One subdivision, a, adheres to the glass like 
an ordinary hydrorhiza, the other subdivision, b, projecting 
up in the water and now ending in a hydranth. The latter 


Vig. 1 Eudendrium. Restitution mass four days old. cs, coenosareal out- 
growth; ec, ectoderm; en, entoderm, op, perisarc of original mass; p, perisare of 
outgrowth; s, space between original mass and its perisare. 150. 

Fig. 2. Eudendrium. Part of large restitution mass four days old. a, b, 
spheroidal nodules of living tissue with perisarc; c, similar nodule with coeno- 
sarcal outgrowth, cs; d. t., dead tissue; other lettering as before. 90. 

Fig. 8 Eudendrium. Nodule of living tissue n, with coenosarcal outgrowth, 
cs. Nodule was part of a large restitution mass. Other lettering as before. 
x 150. 

Fig. 4 Eudendrium. Part of large restitution mass, rs, with two coenosarcal 
outgrowths, cs.1, cs.2. Other lettering as before. 90. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. Il, NO. 3 


295 


296 H. V. WILSON 


has not yet acquired the characteristics of the species, although 
considerable differentiation has gone on. Nettle cells in the 
enlarged tentacular ends can be made out. The ectoderm of ¢s 2 
has contracted away from the perisare, thus assuming the condi- 
tion common in the adult. In the entodermie cavity of the 
coenosare a movement of small spheroidal particles goes on con- 
stantly. Such particles are perhaps remains of the interior of 
the original mass, serving now as food material. The specimen 
is preserved in the condition drawn (fig. 5). 

Two of the masses that were isolated on July 13, at that time 
in the condition of fig. 4, have developed completely formed 
hydranths on July 15. One of them is shown in fig. 6. The 
general mass, 7s, from which the outgrowth has sprouted, is 
for the most part dead, although it still includes some spheroidal 
nodules, n, of bright orange color that are obviously alive. The 
original outgrowth, ¢s., 1s a hydrorhiza, and its apex has appar- 
ently died, for here there is a regeneration point, rp. A branch, 
a, ascends in the water and bears the hydranth. This has the 
characteristic shape and size, the long slender tentacles, large 
hypostome, and bright orange color of the normal adult polyp. 
The other mass is also largely made up of dead tissue, but an 
outgrowth has survived and become a hydrorhiza bearing two 
fully developed hydranths like that of fig. 6. Both masses now 
preserved. 

Summing up, it may be said that the larger masses of this 
experiment, of the general character shown in fig. 2, gave rise to a 
number of coenosarcal outgrowths like those of figs. 2 and 4. 
In all, about three dozen such were obtained from July 13 to 
July 20. A good many were preserved as soon as this state was 
reached. From the others that were kept for further develop- 
ment, three masses gave four hydranths as already recorded. My 
experience with sponges suggests that exposure to the natural 
water of the harbor would give a higher rate of survival and 
transformation. 

Along with the larger plasmodial masses of this experiment, 
very many small lumps were formed, 200 to 300 in diameter, 
which became spheroidal and in about one day after their forma- 


on mass of fig. 4, twenty-four hours later. 
dead. Outgrowth, cs.2, has bifurcated, 


Fig. 5 Eudendrium. Restituti 
General mass, rs, and outgrowth, ¢s.1, now 
and one branch, }, has developed a hydranth. 90. 

Fig. 6 Eudendrium. Restitution mass, 7s, with coenosarcal outgrowth, cs, 


the latter bearing a vertical branch, a, which ends ina hydranth. d. t., dead tis- 
sue; n, nodule of living tissue; rp, regeneration point. Other lettering as before. 


x 90. 
297 


298 H. V. WILSON 


tion secreted a perisare. These small masses adhered so firmly 
to the glass that the water could be drained off when the glass 
was changed to a fresh bowl. I had expected that many of them 
would transform. As a matter of fact only half a dozen devel- 
oped outgrowths, and these were short and in no case gave rise 
to a hydranth. These masses remained alive for days. Perhaps 
their failure to produce hydranths was due to the small amount of 
material making up the mass. 

Experiment July 14.4. Colonies chopped up and pressed through 
gauze 4 p.m. The tissue was distributed with the pipette in 
watch glasses, on slides and covers, and all were immersed in 
large bowls of water. The tissue from the start had a bad color, 
not orange but a dirty brown. The colonies I used were probably 
not clean enough, and perhaps too much Perophora and other 
organisms infesting the hydroid got in the cultures. All of these 
preparations died before the tissue even reached the state of 
smooth compact masses. 

Experiment July 15. Colonies, some with male, some with 
female gonophores, others without gonophores, were pressed out 
at 4 p.m. The tissue is handled in essentially the way described 
for Exp. July 9, and at 7 p.m. fusion has led to the formation of 
flattened plasmodial masses and lumps of all sizes. The color of 
the masses at this time is good, reddish-orange, and the surfaces 
are not far from smooth. But on the following day much of the 
tissue is dead. Many of the nodules and lobes composing the 
larger masses show, however, a thin perisarec. The masses includ- 
ing nodules of live tissue are transferred to fresh sea water. The 
masses gradually died or when nodules remained alive they 
exhibited no change, continuing in a dormant condition. In 
this and most of the subsequent experiments I feel satisfied that 
the cause of failure lay partly in the fact that too much tissue was 
pressed out. This made the first steps in the handling slow and 
the aggregations of tissue too large. 

Experiment July 18. A clean colony was pressed out in the 
usual way in watch glasses, the cells settling on the bottom. 

*The records of several very unsuccessful experiments are given in the expecta- 
tion that they may serve to point out features in the method of treatment which 


are to be avoided 


BEHAVIOR OF DISSOCIATED CELLS 299 


These are allowed to stand about fifteen minutes, during which 
time they are shaken to the center a couple of times. The 
glasses are then transferred to bowls of water. Finer sediment 
floats off, but the coarser clings to the bottom. After ten min- 
utes the glasses are removed to fresh bowls. Again some sedi- 
ment is lost, but not much. The peripheral sediment is now 
gently dislodged with the pipette, and heaped up towards the 
center of the glass. After an hour the tissue is sucked up with a 
large pipette, an effort being made not to break it up more than 
is necessary, and deposited on bottom of fresh bowls in heaps 
+ to 2 mm. thick and about 5 mm. in diameter. The tissue was 
later on again transferred in the same way. The technique of 
this experiment proved wrong, for on the next day most of the 
tissue was dead. Probably there was too much handling, and 
the heaps made were too massive. 

Experiment July 19. Colony pressed out 11:30 a.m. Tissue 
collected in watch glasses and shaken to center after ten minutes. 
Glasses transferred to bowls 12 m., and again to fresh bowls 
1 p.m. Glasses now removed from bowl and with pipette the 
tissue is transferred to fresh watch glasses of water and there 
deposited in heaps about 5 mm. in diameter and } mm. thick. 
The material is now coarsely lumpy, showing that fusion has gone 
on, and the heaps are therefore porous, not dense and compact. 
Glasses transferred to fresh bowls at 1:20 p.m. gently so that the 
tissue aggregations are not disturbed. Again so transferred at 
3 p.m. 

The coherence and slow contraction of the heaps of tissue is 
shown by the fact that at 5 p.m. they have curled up slightly at 
the edge and are now free or nearly free from the bottom. In 
this condition the heap or cake can be pushed with the pipette 
over the bottom. Only the thinner heaps have this amount of 
coherence—to behave in this way they should not be over } mm 
thick. Glasses transferred to fresh sea water at 5 pm. and at 
ips: 

On the next day the result was disappointing. Most of the 
material was dead. Some of the smaller cakes were however 
alive or alive in places, and had developed a perisare. One of 


300 H. V. WILSON 


these is shown in fig. 7. Several such were preserved for sections. 
Again | attribute the failure, comparatively speaking, of this 
experiment to too much handling and to the fact that heaps of 
too large a size were made. It must be borne in mind that 
when any considerable part of the tissue dies during the first 
day, the surviving masses having been infected have a poor 
chance for further growth. 

Experiment July 22. Colonies pressed out 3:10 p.m Only 
tops of clean colonies were used and female gonophores were 
mostly excluded. Tissue collected in watch glasses and trans- 
ferred to bowls, about as before. At 7 p.m. tissue was coarsely 
lumpy, showing that fusion had gone on. It was transferred to 
bowls of fresh sea water and deposited in areas 7 mm. in diameter 
and 4 mm. thick. Tissue nearly all dead the next day. 

Experiment July 23. Tops of clean male colonies were pressed 
out at 3:30 p.m. At 7 pm. tissue was sucked up with pipette 
and distributed in fresh bowls to form thin areas about 10 mm.in 
diameter. These areas soon become transformed into reticular 
sheets having considerable coherency. They become free or 
nearly free from the bottom, and some curl up slightly at the edge 

indications of contraction. Another set of preparations were 
made at the same time and treated in same way, from the tops of 
female colonies. This tissue too went so far as to form reticular 
sheets of considerable coherency. But on the following day the 
tissue was practically all dead. 

Experiment July 25. Male and female colonies were pressed 
out at about 2 p.m. The tissue collected from each colony 
was kept by itself. One-half hour after preparation, the tissue 
which had been squeezed out in watch glasses, was shaken to the 
center of the glass, and the glasses were transferred to bowls. 
The glasses were transferred to fresh sea water at intervals of 
two hours until 9 a. m. on the next day. Five hours after prep- 
aration the tissue had united to form thin sheets, which soon 
began to crack evidently owing to a process of contraction. 
These sheets of tissue however went no further in development and 
on the next day were dead. 


BEHAVIOR OF DISSOCIATED CELLS 301 


In the peripheral region of each watch glass numerous small 
masses of tissue, a fraction of a millimeter, formed. These are 
alive on July 26 and show a smooth contour. They are still 
alive on July 27 and now have an obvious perisare. Many of 
them cling to the bottom, but many are free and are held together 
in loose, thin, flat aggregations by débris and the perisare of dead 
tissue. On July 28 and 29 most of these masses are still alive and 
a dozen have sent out coenosarcal outgrowths. Several are 
shown in fig. 8. A little group of masses all interconnected is also 
shown in fig. 9. 


G@ 


Fig. 7 Eudendrium, Restitution mass twenty-four hours old. a, b, ¢, 
lobes of living tissue; d, ¢.. dead tissue; p, perisare. X 150. 


Experiment July 27. A elean male colony pressed out at 1:45 
p-m. in a watch glass. Tissue was shaken to center after a few 
minutes, and the water in watch glass was renewed after the 
tissue had once more settled on the bottom. The tissue was then 
distributed with a large pipette over the bottom of two watch 
glasses and shaken to center. In each watch glass a central 
collection of tissue is thus formed in the shape of a thin area 10 
to 15 mm. in diameter. Outside of this are scattered many small 
lumps, | mm. and less in diameter. The glasses are transferred 
to bowls 3:30 p.m. — Four such cultures were made. 


302 H. V. WILSON 


At 5 p.m. coherent cakes of open, reticulated texture have 
been formed. These are free or nearly free from the bottom. 
They are transferred with a large pipette to fresh bowls of water. 
At 7 p.m. they are again transferred, and several are cut in pieces. 
They have a good color and are obviously alive. On the next 
day many of the cakes are still alive and have formed a perisare 
over much of their surface. Where the perisare has formed, the 
included tissue is uniformly opaque and by reflected light appears 
distinetly orange. The formation of the perisare cuts out, I 
think, such masses as pieces of gonophores, tentacles, and hy- 
dranths, from the undifferentiated mass, round which it forms. 
Gonophore and tentacle fragments are to be sure sometimes 
partially surrounded by the growing (fusing) masses of tissue, 
from which they may be seen projecting. But probably a con- 
siderable part of each such mass dies, and with it the fragments 
of gonophores, ete. 

On July 29 most of the tissue is dead or dying. Bacteria and 
infusoria are abundant. Nodules of live tissue surrounded by 
perisare oecur imbedded in the general mass. In some of the 
disintegrating cakes the arrangement of the perisare is much 
plainer than it was when the whole cake was alive. It may now 
be seen that the perisare was seereted round compact nodules, 
lobes, and anastomosing cords. All such are intricately combined 
with débris and tissue that never secreted perisare to form the 
general mass. 

All the larger masses soon died. But a good many of the small 
lumps that lay in the outer part of the watch glass, away from the 
central cake, were still alive on August 1. Several of them by 
this time had sprouted coenosareal outgrowths, and were essen- 
tially like fig. 1, although in some instances the mass had given 
rise to two opposite outgrowths. 

The practical problem in handling this tissue seems to be to 
get masses large enough to provide the necessary amount of mate- 
rial for hydranth development, and yet thin and reticular enough 
to expose all parts of the mass to the water. Small masses, a 
fraction of a millimeter in diameter, are much easier to keep alive 
than large masses. 


BEHAVIOR OF DISSOCIATED CELLS 505 


Fig. 8 Eudendrium. Group of restitution masses three days old; two still 
spheroidal; one with a coenosarcal outgrowth, cs.; one with two such outgrowths. 
op, perisare of original mass; s, space between original mass and its perisarc. 
Other lettering as before. > 150. 

Fig. 9 Eudendrium. Three restitution masses interconnected, four days old; 
two masses with coenosarcal outgrowths, es;7s, central part of original restitution 
mass. Other lettering as before. 90. 


Experiment August 1. Colony pressed out at 3:30 p.m. Tissue 
in comparatively small amount is collected in watch glasses and 
allowed to settle. Water is gently drawn off and fresh added 
without disturbing the layer of ‘sediment’ that clings tothe bottom. 
Glasses transferred to bowls after thirty minutes: transferred 
again 6 p.m. On the following day all the larger masses dead or 
dying. Many of the very small masses are also dead. Still 
there are very many small masses, a fraction of a millimeter in 
diameter, that are alive. These are more or less spheroidal and 
covered with thin perisarc, some attached to bottom of glass, 


304 H. V. WILSON 


some to cover glasses that had been put in the watch glasses. On 
August 3 many of these masses are still alive, surrounded by peri- 
sare, but they have not thrown out coenosareal processes. 

In this experiment a small amount of tissue was pressed out. 
and until the details in the method of treatment are more precisely 
marked out, this is certainly a safe step. It will be noticed that 
the tissue was left 7m situ where it was first deposited on the bot- 
tom of the glass. The results indicate that this is not a good 
method for the larger masses. And yet it seems desirable for 
the tissue to establish some connection with the bottom, and this 
it will not do if disturbed and dislodged too much. The experi- 
ment records show clearly what a great influenceapparently slight 
differences in the treatment had on the vitality of the fusion 
masses, and how much in the dark I remained as to what details 
were good, and what bad. 

Experiment August 2, a. Colony pressed out at 4:15 p.m. 
Tissue collected in a watch glass, shaken gently to the center, and 
water changed several times, each time the tissue being stirred up 
considerably by the pipette current. Glass transferred to bowl 5 
p.m.; transferred again 6 p.m. Tissue does not cling to the 
bottom, as it does when left undisturbed where it first settled. 
At 7. p.m. the tissue coheres sufficiently for pieces 2 to 3 mm. wide 
to be sucked up with large pipette. Other smaller pieces about 
1 mm. wide are sucked up. All pieces are thin, about 4 mm. 
thick. These pieces, forty-five in number, are scattered over 
the bottom of three bowls. 

On August 3 at 9 a.m. the masses are alive and of good color. 
Some are free, some slightly attached to the glass. The latter are 
freed and all are gently transferred with large pipette to fresh 
bowls of water. They resemble the Pennaria mass shown in fig. 
14. Examination shows that the perisare has formed over parts 
of many plates, but in other places while the surface of the plate 
is smooth, no perisare can be seen. In still other places the con- 
tour is rough, the periphery here consisting of rounded cells. Bac- 
teria are present, here and there in swarms, but not much of the 
tissue is dead. 


BEHAVIOR OF DISSOCIATED CELLS 305 


On August 7 a good deal of the tissue is now dead. But a 
large number of the pieces include lobes and nodules of living 
tissue surrounded by perisare. Two of the plate-like masses 
show each a coenosarcal outgrowth of considerable length, about 
like those of fig. 4. On August 8 two other masses show each a 
similar outgrowth. On August 9 another mass has developed an 
outgrowth, which soon becomes sickly, losing its well marked 
layers and developing in the interior numerous dark masses. On 
August 11 two other pieces show each a coenosarcal outgrowth. 
These outgrowths are horizontal and creeping and each bears a 
vertical branch. Both are sickly as is shown by the fact that the 
layers are not everywhere uniformly differentiated, but in places 
appear to be breaking up, while in other places the tissue is 
densely concentrated. 

The method practised in this experiment is evidently good and 
yet too much of the tissue dies leaving the surviving masses slow 
to transform. In the hope of stimulating these masses they 
were given a liberal supply of pure oxygen on August 6, but with 
no discoverable results. 

While an effort was made in this and the other experiments to 
pick out pieces large enough to be distinguished with the eye, 
some pieces of gonophores and tentacle fragments remain in the 
cultures. Some of these are incorporated by the plasmodial masses. 
Many others undoubtedly die without being incorporated. 
Another structure too deserves mention as being occasionally 
present in the cultures. In cutting up the hydroid, many 
hydranths are snipped off at the very base. Some of these get in 
the cultures and escape notice. I have found that when such 
hydranths are isolated a process of reduction takes place analo- 
gous to that described by Schultz for hydra (’06), the hydranth 
gradually becoming in the course of a few days a mouthless 
spheroidal body. Apparently such a process goes on in the 
pressed out cultures more rapidly, for oceasionally spheroidal 
bodies are seen quite like the reduced hydranths just referred 
to. I have moreover several times found such a body embraced 
by a plasmodial mass. Where bodies like gonophores, tentacle 
fragments, and reduced hydranths (or the bodies that look like 


506 H V. WILSON 


such) become surrounded by the plasmodial tissue, it is a question 
what becomes of them. As already said, the formation of the 
perisarc, [ am inclined to think, cuts out such bodies from the 
comparatively homogeneous material round which it forms. 
Pictures are sometimes had which indicate that possibly such 
bodies are attacked by the plasmodial tissue, the latter invading 
and absorbing them. Again a stray fragment of stem perisare 
from the colony used may get into the cultures, and if some of the 
coenosare has been left inside, it may form a regeneration knob 
at one end of the piece. I have seen a piece or two of this kind. 
Such a fragment might very well regenerate a hydranth in the 
midst of the plasmodial masses. But fragments of this sort are 
easily distinguished from the plasmodial masses or nodules of 
the latter. 

Experiment August 2, b. A colony was pressed out and the 
tissue allowed to settle on a cover glass immersed in a watch glass. 
In transferring, the cover glass was always kept in the watch 
glass, and thus the tissue was not directly exposed to the air. 
The material settling on the cover soon transformed itself into a 
multitude of small, more or less spheroidal masses, a fraction of ¢ 
millimeter in diameter. They went so far as to form a perisare. 
A large number of them died about one day after preparation, 
but many remained alive for days; were still alive on August 9. 
They had not sent out coenosareal outgrowths, but as was learned 
later from sections the originally solid mass in several cases, 
perhaps in all, had developed into a sac, the wall of which was 
made up of ectoderm and entoderm layers. 

In the same way on August 2 tissue was allowed to settle over 
the bottom of a watch glass, forming a very thin deposit. Great 
numbers of small lumps formed which had the same history as the 
above. It is quite possible of course that a few of these lumps sent 
out coenosarcal outgrowths, but that such escaped notice. 

Summary. Eleven experiments with Hudendrium were made. 
In all experiments fusion led to the formation of plasmodial 
masses. In eight experiments an extensive formation of perisare 
took place. In five experiments numerous small masses with 
perisare, which remained alive for days, were formed; in three of 


BEHAVIOR OF DISSOCIATED CELLS 307 


these experiments several of the spheroidal masses sent out 
coenosareal outgrowths. In two experiments a considerable 
number of coenosareal outgrowths were obtained from nodules 
of tissue which had remained alive in comparatively large flat- 
tened plasmodial masses, and in one experiment these outgrowths 
gave rise to hydranths, four in number. 

The first experiment, July 9, was much the most successful. 
In those that followed a common error undoubtedly lay in endeay- 
oring to handle too much tissue. But over and above that the 
water grew warmer with the advancing season and the Euden- 
drium colonies perhaps became more abundantly infested with 
protozoan ectoparasites. 

The technique in general of these experiments is especially 
faulty in that, (1) it allows parts which must die, tentacle and gon- 
ophore fragments, to get into the cultures; and, (2) it subjects 
masses of tissue that evidently need the best environment to the 
harmful influences of quiet water in a laboratory dish. Small 
gauze floats kept at the top of a running aquarium were used, but 
with no success. Very probably if the cultures were placed out- 
side in the harbor water, they would do better. As to methods, 
the following may be added to what has already been said: 

Only clean colonies or parts of colonies should be used. If the 
whole hydroid is used I believe that colonies without gonophores 
are the best, and those with female gonophores the worst. Stem 
tissue would perhaps be better than that of the whole colony. 
Care should be taken to allow the cells and small lumps to cohere, 
and not to break up the cohering tissue at first more than is neces- 
sary. It is well to get the tissue in comparatively small pieces a 
few hours after preparation, and in fresh dishes away from the 
original surface of attachment. If the masses of tissue a few hours 
after preparation appear soft and pasty, they will probably not 
live. They should show a good color, absence of the characteris- 
tic color indicating presence of dirt or infesting animals. The 
gauze (silk bolting cloth), usually used runs 50 meshes to 25 mim. 
A cloth running about 75 meshes to 25 mm. was also used. The 
sea water was well aérated and filtered. An effort was made to 
pick out from the cultures all coarser particles, such as pieces of 


30S H. V. WILSON 


hydranths and gonophores, that could be seen with the eve. The 
tissue was usually pressed out and kept in solid watch glasses with 
a cavity 50 mm. in diameter and 10 mm. deep. These were im- 
mersed in crystallization dishes 200 to 250 mm. in diameter or in 
finger bowls of about 120 mm. diameter. Dishes, instruments, 
and gauze were thoroughly cleaned, but were not. sterilized. 
Possibly sterilization in hot water would be advantageous. I 
lay emphasis on the technique of the experiments in the hope that 
it may be improved by others. With a more certain technique 
this method of growing hydroids ought to lend itself to the pro- 
duction of hydrids, as I have suggested (’O07b, ’11b) for sponges. 


Histological study of the restitution masses of Hudendriwm 


Observation record, July 27. At 4:30 p.m. a drop of the Euden- 
drium tissue was squeezed directly from the gauze sae on to a 
slide. A supported cover was put on and slide examined at once. 
The fluid contained quantities of separate cells. In addition 
there were present a few small masses each consisting of several 
cells. All the cells were about spheroidal, but they varied a 
great deal in size. Some contained abundant pigment granules, 
others a few, and others were quite transparent. Four types 
could be distinguished, all of which were abundantly represented, 
but plenty of transitional cells connecting these types were also 
present. The types are shown in fig. 10, a, 6b, c, d. Cell a is 
well filled with pigment granules which appear brownish red by 
transmitted light. Cell 6 contains similar granules, but they are 
few in number. Cells or particles ¢ and d are transparent and either 
without pigment or show only a faint granule or two. The slide 
was kept in a moist chamber and examined at intervals for four 
hours. The formation of numerous minute masses each consist- 
ing of a few cells was observed. These grew large through fusion 
with one another from hour to hour. At 8 p.m. a few plasmodial 
masses of considerable size were present, the largest of which 
measured about 300u x 50u. This was a thin flattened, sheet-like 
mass having an irregular outline. Its general body was opaque, 
but the peripheral part was thinner and here it could be seen that 
all of the four types of cells entered into the composition of the 


BEHAVIOR OF DISSOCIATED CELLS 309 


mass. At this time, 8 p.m., numerous small plasmodial masses 
grading down to aggregates of a few cells were also present. In 
many of these too the four types of cells could be distinguished. 
Finally in the preparation at this time there still remained quan- 
tities of free cells. In this preparation there were no bits of ten- 
tacles, gonophores, or foreign particles. 


Ci 
igh X aN 


an 
Fig. 10 Eudendrium. Elements of the pressed out tissue. > 700. 
Fig. 11 Eudendrium. From a section through a lobe of the mass shown in 
fig. 7. cen, enidoblast; f. c., free cell; p, perisare. X 1200. 


Results from study of sections. The plasmodial mass shown in 
fig. 7 was about twenty-four hours old when preserved. A part of 
it had already died, but there were three large lobes of living tissue 
surrounded by an obvious perisare. These lobes were found to 
agree in structure. Part of a section through one of them is 
shown in fig. 11. In the interior of the lobe there are numerous 
cells which seem to be free, that is the body is well defined all 
round. These cells vary in size; the nuclei are relatively large with 


310 H. V. WILSON 


abundant nucleoplasm and usually with a conspicuous nucleolus; 
the cytoplasm as a rule shows vacuoles and solid inclusions in 
vacuolar spaces. Similar cells are met with which are not sharply 
delimited all round, but only on one side (cell a in fig. 11); on the 
other side, the cell shading off into the syneytial reticulum. In 
such a case, I take it, we have a mass which has broken away from 
the general syneytium on one side, the protoplasm on this side 
condensing to form a film of exoplasm. Cnidoblast cells, ¢ 7., 
with included nematocysts are also common in the interior of the 
lobe. Between the cells or groups of cells the protoplasm exists 
as a vague reticulum, the vacuolar spaces in which are of all sizes. 
Scattered in the reticulum are nuclei. The external stratum of the 
lobe in some places is not markedly different from the interior. 
In other places it shows smaller cells and more nematocysts than 
the interior. No doubt some of the nematocysts have been 
carried over from the parent hydroid; possibly others are new for- 
mations. The external structure of the lobe is in most places 
continuous with the interior, but here and there it is separated 
as in the next mass to be described. The dead part of the mass 
(fig. 7) consists of a loose granular stuff including some nuclei and 
nematocysts. The live and dead tissue are not sharply separated, 
but grade into each other. 

Two other plasmodial masses about twenty-four hours old were 
sectioned. These masses were irregular bodies of the same 
general character as the one shown in fig. 7. They were however 
entirely alive, and surrounded everywhere by a distinct perisare. 
They proved to be essentially alike in internal structure. Part 
of a section through one is shown in fig. 12... The body is solid and 
an outer stratum is almost everywhere separated from an inner 
mass by a vaguely delimited cleft-like space, s. The outer stra- 
tum is chiefly composed of comparatively smooth cells, forming 
four or five layers, and of enidoblasts. The cells have large nuclei 
and are in general well defined. There are places however where 
one can only find nuclei lying in a vaguely reticular protoplasmic 
matrix The inner mass is a complex syneytium containing 
abundant large nuclei and vacuoles with inclusions. Numerous 
enidoblasts are seattered through it, and well defined ordinary 


BEHAVIOR OF DISSOCIATED CELLS Silt 


A 


2. 


Sa ye pee PS 
Die eee 


Vig. 12. Eudendrium. From a section through a restitution mass about 
twenty-four hours old. cn, enidoblast; f. c., free cell; p, perisare; p. /., protoplas- 
mie film; s, space separating an outer stratum, o. s., from the inner mass, f. ™”. 
x 1200. 

Fig. 13 Eudendrium. Section of the restitution mass, with one coenosareal 
outgrowth, shown in fig. 8. Section strikes the original mass and does not include 
the outgrowth. y, yolk mass. Other lettering as before. > 350. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO, 3 


aH LP H. V. WILSON 


cells are found in it here and there. Protoplasmic films, p /, 
are common which mark off cells or areas on one side while on the 
other side the protoplasmic area has no distinct boundary. 

The masses shown in fig. 8, two spheroidal and two with out- 
erowths, were sectioned. These bodies when preserved were 
three days old. In fig. 13 is shown a section through the dilated 
body of one of the masses which had a coenosareal outgrowth. 
All the bodies proved to be in essentially the same condition as 
far as the differentiation of layers is concerned. In them all an 
ectoderm and entoderm are distinctly differentiated. The two 
layers are separated by a cleft-like space, there being no distinct 
supporting lamella. In the small spheroidal masses and in the 
dilated portions (representing the original shape) of those with 
outgrowths, there is a central yolk that is still continuous with the 
entoderm in spots. The yolk mass does not extend into the out- 
erowths. The perisare is laminated and in the sections almost 
everywhere widely separated from the ectoderm. The ectoderm 
is composed of small cells, probably all interconnected, forming 
in places one layer, in other places two or three layers. Small 
nematocysts and stages in development of these are common, 
and other inclusions also are present in the ectoderm. The 
entoderm consists of a single layer of large closely packed cells 
varying from a more or less cubical to a somewhat flattened shape. 
These cells are uninucleate, the eytoplasm more or less vacuolated 
and sometimes containing small (developing) nematocysts and 
other inclusions. The central yolk, y, is a granular mass in places 
composed of small spheres of varying size. In it small spheroidal 
vesicles containing one or two deeply staining granules are 
common, doubtless representing swollen and degenerating nuclei. 
A few small nematocysts are also found in the yolk. The yolk 
mass although in general separate from the entoderm is perfectly 
continuous with this layer in spots. 

While my observations on the histological strueture of the 
restitution masses, both in Eudendrium and Pennaria, are frag- 
mentary, they are nevertheless definite. From them it would 
seem that in Eudendrium the solid aggregate formed by the fusion 
of the isolated cells passes into the condition of a syneytium which 


BEHAVIOR OF DISSOCIATED CELLS 313 
includes partially or perhaps completely free cells (fig. 11). An 
outer stratum in which cell bodies are well differentiated, and 
which is several layers deep, now becomes marked off from an 
inner mass (fig. 12). The outer layer probably represents the 
ectoderm, while the inner mass represents a yolk-entoderm, which 
subsequently splits into the definitive entoderm and the yolk. 
Finally ectoderm, entoderm and a central yolk mass appear (fig. 
13), as in the development of a coelenterate planula, Manicina 
for instance (Wilson, ’SS8), or Eudendrium itself (Hargitt, 04). 


PENNARIA. RESTITUTION FROM DISSOCIATED CELLS 


Species used. The species used was Pennaria tiarella MceCrady 
(Proc. Elhott Soc., vol. 1, no. 1, p. 153; Nutting, ’01, p. 337). 
Pale specimens with light colored ova and deeply colored speci- 
mens with orange ova were abundant together on the floats 
round the laboratory wharf at Beaufort during August. Both 
pale and colored forms liberated medusae at dusk, about 7 p.m. 
The forms appear to represent merely the extremes in ‘a range of 
color variation (for the varieties at Woods Hole mde Hargitt, 00). 

Experiment July 26. vigorous and clean colony was pressed 
out in the usual way at 4:25 p.m. Quantities of cells of various 
kinds, especially spheroidal granular cells with more or less pig- 
ment, came through; also small cell groups, and bits of tentacles. 
Fusion of the cells and cell aggregates begins at once and proceeds 
rapidly. In ten minutes time a mass 500ux 100u has been 
formed, in a watch glass kept under the microscope, practically 
from isolated cells. Such masses change their shape slowly and 
fuse with one another. The tissue which was pressed out in a 
watch glass was shaken to the center at 4:35 p.m. At 4:53 it has 
formed a thin coherent cake. This is now sucked up in piaces 
with the pipette, and so broken into pieces about 51min. in diameter 
which are transferred to a bowl of water. 

At 6 p.m. the tissue lies on the bottom of the bowl in the 
shape of thin, somewhat reticular sheets. These are freed from 
the bottom (they had already begun to curl up round the edge) 
with small pipette and are transferred with a large pipette to a 


314 H. V. WILSON 


fresh bowl. In so doing the sheets are of course broken where- 
upon the peripheral parts turn white and disintegrate quickly. 
But the body of the piece remains alive, keeps its color (a reddish 
tinge), contracts and soon has a comparatively smooth surface 
once more. Fraements accidentally broken from the sheets are 


Fig. 14 Pennaria. Restitution mass, four days old. Photograph. ™& 50 


also transferred, and these also quickly ‘heal.’ All the masses 
continue to contract, and by 8 p.m. many of them have a massive 
shape, although some at this time are still sheet-like. They all 
have a smooth surface, and the majority of them are in the neigh- 
borhood of 1 mm. in diameter. One of the largest is shown in 
fig. 14. 


BEHAVIOR OF DISSOCIATED CELLS yl lPa) 


On July 27 at 9 a.m. the masses are surrounded by a distinct 
perisare and with some exceptions are still alive or include con- 
siderable live tissue. The color of the live tissue is pink. The 
smaller masses are of compact shape, spheroidal or ovoidal, and 
are alive throughout. The larger masses are of a somewhat 
lobular shape, and while the projecting lobules are alive, a consid- 
erable part of the body of the mass is dead or dying. Several are 
now preserved. Sections confirmed what f have just said as to the 
distribution of the live tissue. In fig. 20 a section through one of 
the smaller masses is figured, and it may be seen that the whole 
mass was alive, and while in general still solid had begun to ditfer- 
entiate the ectoderm and entoderm layers. The masses at this 
stage are soft and burst easily on rough handling. The water 
was henceforth renewed with a siphon. 

On July 29 two of the masses have developed outgrowths. A 
photograph of one of them is shown in fig. 15, and the other is 
represented by a photograph (fig. 16) and a camera sketch (fig. 
21), the latter made from the living object. The original mass in 
both cases was spheroidal, and the thick perisarc, 0. p., which 
surrounded it, still persists. The mass, as sections of similar 
bodies show, has differentiated into ectoderm and entoderm layers 
which surround a central cavity containing the remains of yolk 
material. The body shown in fig. 21 has developed one long 
outgrowth ¢ and two short ones, a and 6, just protruding at the 
opposite end from the original perisare. In the long outgrowth 
the ectoderm and entoderm are thicker than elsewhere. The peri- 
sare over this outgrowth is noticeably thinner than that over the 
original mass, and the ectoderm in a part of the outgrowth has 
contracted away from the perisare, remaining connected with it 
by strands, ect. s. after the fashion characteristic of the adult 
hydroid. The original mass, too, it may be seen has contracted 
away from the perisare and has materially changed its once globu- 
lar shape. The other mass, fig. 15, which has developed only a 
single outgrowth, represents a slightly earlier stage than figs. 
16 and 21. In it the original mass has contracted away from the 
perisare, 0. p., but remains connected with it by strands of ecto- 
derm. In the outgrowth however the ectoderm has not yet 


316 H. V. WILSON 


1.5, 16 
Fig. 15 Pennaria. Restitution mass three days old. a, coenosarcal out- 
growth; op, perisare of original mass Photograph Xx 50 
Fig. 16 Pennaria Restitution mass three days old. op, perisare of original 
mass; p, perisare of long coenosarcal outgrowth. Photograph x 50 


begun to separate from the perisarcal covering. Both these 
masses including the outgrowths are firmly adherent to the bot- 
tom of the vessel. 

On July 29 a large number of the masses are dead. Only the 
smaller ones together with two or three of the larger survive, and 
the latter have been injured and evidently are in bad condition. 
Injury often comes in changing the water, the siphon setting up a 
current which strains the bodies of the larger masses especially 
since these are attached to the bottom only throughout a part of 
their extent. The small masses are more perfectly attached to 
the glass. It is clear that if one wishes to grow hydroids in this 
way it is better to produce comparatively small masses instead of 
large ones such as that shown in fig. 14. 


BEHAVIOR OF DISSOCIATED CELLS 


17 18 
Fig. 17 Pennaria. Restitution mass, five days old, with two outgrowths, 
one branched op, perisare of original mass Photograph x 5O 
Fig. 18 Pennaria Restitution mass, two days old, metamorphosed, with 
hydranth. op, perisarc of original mass. Photograph x 50 


On July 30 one of the masses has developed three outgrowths, 
rach about like the long outgrowth in fig. 16. The mass itself 
and two of the outgrowths adhere to the bottom, while the third 
rises obliquely in the water. On July 31 only one of these out- 
erowths remains adherent to the bottom; the other two rise 
obliquely in the water. The extremities of the latter have now the 
character of knobs, reddish brown in color and resting upon 
lighter colored stalks. This mass continues to develop and on 
August 1 has reached the condition shown in fig. 19. The ascend- 
ing outgrowths now bear hydranths, each with the lower filamen- 
tous tentacles and the upper short capitate tentacles characteristic 
of this hydroid. The thick perisare, 0. p., marks out the size of 
the original mass from which the outgrowths sprouted. The out- 
erowth x is the one that remained adherent to the glass. In the 


318 H. V. WILSON 


Fig. 19 Pennaria. Restitution mass six days old, completely metamorphosed, 
with developed hydranths. op, perisare of original mass; x, perisarc of outgrowth 


adherent to glass. Photograph. % 50. 


living body the perisare ended in a closed rounded extremity 
and contained a coenosareal prolongation extending throughout 
its length. When the body was pried from the glass, the coeno- 
sareal prolongation retracted and in the photograph it appears 
very short. 

Experiment August 3. A clean vigorous colony about 5 inches 
high is selected and only stem material (coenosare) is used. All 
lateral branches are cut off, also the base and tip of the main 
stems. The latter are then cut into pieces about 3 mm. long and 


BEHAVIOR OF DISSOCIATED CELLS 319 


Fig. 20 Pennaria. Section of restitution mass, seventeen hours old. ec 
ectoderm; p, perisarc; y. en., yolk entoderm. X 150. 

Fig. 21 Pennaria. Same mass as in fig. 16. a, 6, c, coenosareal outgrowths; 
are of original mass; p, perisare of long coeno- 


ec, ectoderm; en, entoderm; op, peris 
sarcal outgrowth. 90. 

Fig. 22 Pennaria. Median section of restitution mass two days old. a, b,c, 
short coenosarcal outgrowths; ec, ectoderm; en, entoderm; op, perisare of original 
mass; y, yolk material. > 150. 


320 H. V. WILSON 


these are pressed through gauze in the usual way. The tissue 
thus obtained is pure coenosareal tissue broken up into separate 
cells and minute aggregations of cells. The tissue is pressed out 
at 8:10 p.m. Fusion goes on rapidly, and in twenty minutes 
round the edge of the collection of tissue small bars and plates, | 
to 3. mm. long, have been formed. At 4:50 p.m. the material 
exists as a thin cake about 4mm. thick, of considerable coherency, 
vet of loose reticular texture. This les on the bottom scarcely 
adherent to the glass. It is cut with scalpel and needle into pieces 
about 4 mm. in diameter. Nine such pieces are prepared and 
together with some much smaller fragments are transferred with a 
pipette to a fresh bowl of water. 

At 7 p.m. all the pieces have contracted considerably. The 
plate-like pieces have begun to curl up at the edge and the smallest 
fragments have already assumed a compact, massive shape. The 
color is a light pink. On the next morning (August 4) all the 
masses are still alive and have secreted a distinct perisarc. The 
smaller are spheroidal or ovoidal in shape, the larger of an irregu- 
larly massive shape, measuring up toa length of 3mm. They are 
all adherent to the glass. The water is now changed by siphoning, 
and some of the larger masses rupture. They rupture very easily 
and it is clear they are too large to thrive. After rupturing, a 
mass quickly acquires once more a smooth surface within the 
perisare. 

On August 5, two of the masses have partially transformed. 
One is shown in section in fig. 22. This mass was originally 
spheroidal and in the living state, including the perisare, meas- 
ured about 4 mm. in diameter. It was firmly attached to the 
elass, and at 9 a.m. August 5 was still spheroidal. At 2 p.m. 
it was observed to be triangular. The triangular character be- 
‘ame more marked during the afternoon, and it was plain that 
the mass was sending out three outgrowths, a, b,c. The body was 
preserved at 7 p.m. Seetions showed that the originally solid 
plasmodial body had differentiated the ectodermal and entodermal 
layers. 

The other mass on August 5 had developed farther. A photo- 
graph of this mass is shown in fig. 18, and a camera sketch made 


BEHAVIOR OF DISSOCIATED CELLS yA 


from the living object in fig. 23. The original mass was spheroi- 
dal; its outline is indicated by the thick perisare, 0. p. Two short 
outgrowths, a and 6, had developed, and these together with the 
original mass adhered to the glass. A third outgrowth c¢ ascended 
in the water and had transformed into a hydranth bearing whorls 
of short stubby tentacles. The ectoderm and entoderm had 
developed throughout the mass, and the ectoderm of the tentacles 
and of the short outgrowths included abundant nettle cells. The 
hydranth while under the miscroscope, in a watch glass, was 
frequently active, bending from side to side. The size of the 
gastric cavity varied with the contraction state but during most 
of the time was large. Shortly after transferring the mass from 
the breeding dish to the watch glass and just after fig. 23 was made 
a quantity of granular material was twice ejected from the mouth, 
after which the polyp contracted considerably. The water was 
then changed, and the polyp returned to about the condition 
shown in the figure. It was then preserved (4 p.m.). The 
tentacles of the hydranth arranged in three whorls, all look alike, 
and are of a more or less globular shape. The upper whorls 
doubtless represent the capitate tentacles of the adult. Possibly 
the tentacles of the lower whorl elongate and develop into the 
filamentous tentacles. In the development of the egg polyp, 
Hargitt (O00, p. 400) finds that the filamentous tentacles appear 
first, the capitate somewhat later In the restitution polyp 
shown in fig. 23 there may have been a slight difference in the 
time of appearance of the several whorls. 

On August 6 another small mass has developed an outgrowth 
and resembles fig. 15. On the next day it is dead. By this time, 
August 7, many of the masses including all the larger ones are 
dead, but some survive and of these three have developed each an 
outgrowth. One of them is shown in fig. 24, the others are sub- 
stantially like it. In all three the coelenterate layers have re-ap- 
peared; the original mass has contracted away from the perisarc 
with which it remains connected by ectodermic strands. While 
the mass shown in fig. 24 was under the miscroscope, these strands 
retracted into the body. The three partially transformed masses 
are now preserved. 


aZZ H. V. WILSON 


23 24 


Fig. 23) Pennaria. Samemassasinfig.18. a,b, short coenosarcal outgrowths; 


c, outgrowth that has become a hydranth; ec, ectoderm; en, entoderm; m, 
mouth; op, perisare of original mass; ¢, tentacle. X 90. 
Fig. 24 Pennaria. Restitution mass three days old, with outgrowth. Let- 


tering as before. X 90. 


On August 8 only four of the masses of this experiment are alive. 
Two are still spheroidal. One has developed an outgrowth and 
is substantially like fig. 24. The other is shown in fig. 17; it has 
given rise to two outgrowths at opposite poles, one of which has 
developed a lateral branch. The original mass has contracted 
away from the perisarc, remaining connected with it by ectoder- 
mic strands in the usual way. On the perisare algae have settled 
and these appear in the photograph as rounded spots. In all 
of these bodies, even in the spheroidal masses, sections showed that 
the ectoderm and entoderm had developed, and a gastral cavity 
containing the remnants of a yolk mass was present. The bodies 
were preserved August 8. 

Summary. Two experiments were made. In the first both 
stems and hydranths were used. In the second only stem tissue 
was used. A large number of solid plasmodial masses were 
obtained, and these within a day uniformly secreted a distinct 
perisare. In the first experiment all the larger masses gradually 
died. Three masses } to 3 mm. in diameter, differentiated the 


BEHAVIOR OF DISSOCIATED CELLS 323 


coelenterate layers and developed coenosarcal outgrowths. One 
of these masses went further and developed perfect hydranths. 
In the second experiment also the larger masses died. Ten smaller 
masses for the most part about 3 mm. in diameter, the largest 
reaching a diameter of :‘) mm., differentiated the ectoderm and 
entoderm layers. Of these, eight developed coenosarcal out- 
growths, and of the eight one mass produced an actively motile 
hydranth with whorls of tentacles. 

Comparison with egg development. For the purpose of compari- 
son with the restitution hydroids my assistant, Mr. O. W. Hyman, 
reared Pennaria from the egg. As Hargitt (00) states, the eggs 
vary considerably in size, the planulas in size and shape both. 
Several planulas were measured and it was found that they 
ranged in length from about :> mm. to 1 mm., while the cross 
diameter was about 7 mm. It will be seen that these planulas 
were not so far removed in bulk from the restitution masses that 
transformed, although neither in the case of the planulas nor in 
that of the restitution masses was there any uniformity of size. 
On the other hand the hydranths produced by metamorphosing 
planulas are fairly constant in size, and they agree in this respect 
with those produced by the restitution mass shown in fig. 19. 


Histological study of the restitution masses of Pennaria 


Sections of the Pennaria stem show that the entoderm is made 
up of a single layer of large columnar cells tapering towards the 
base and measuring about 30u by l5u. The cells contain very 
many large spheroidal granules that stain pale blue with haema- 
toxylin. The ectoderm contains an abundance of nettle cells, 
large and small. In each layer the elements are freely inter- 
connected to such a degree that in many regions at least. the 
structure is that of a reticular syneytium. 

In order to study the composition of the pressed out tissue | 
squeezed pieces of stem in a watch glass of water so that the 
squeezed out tissue fell on an immersed cover glass. After five 
minutes strong formalin was added. Much of the tissue thus 
fixed adheres to the cover, and this when mounted on a slide in 
water gives clear pictures. In such a preparation, fig. 25, there 


324 H. V. WILSON 


Pig. 25 Pennaria. Elements of coenosarecal tissue. From a preparation 
fixed five minutes after tissue was pressed out x GOO. 


are quantities of large granular cells, many of which are spheroidal 
as a, other with pseudopodia as c, some with a larger vacuole as 
b. The granules are generally scattered all through the cell, 
but as in d there may be some clear protoplasm. These cells 
exist separately but also in small aggregations as d and e. The 
cells and contained granules are of about the same size as those 
seen in sections of the stem entoderm and the cells evidently 
represent the entodermic elements. In the preparation again 
are numbers of cnidoblasts of various sizes with included 
nettle cells, /. Other elements, doubtless also eetodermic, are 
finely granular pale cells varying a good deal in size,.g, to- 
gether with small groups of such cells. When the stem piece is 
pressed, numbers of cells must be ruptured, and the preparation 
contains quantities of free granules, b, doubtless derived in large 
part from the entoderm cells. These are transparent, and for the 
most part spheroidal, although often irregular in shape. They 
seem to stick together and even to fuse. Droplets of translucent 
substance smaller than the entodermie granules and varying in 
size down to the vanishing point are common. It seems probable 
that some of this material spoken of as droplets and granules 
represents minute fragments of protoplasm that have rounded off. 
And the question is worth formulating, although I can not answer 
it: are such minute granular or drop-like bodies, representing por- 


BEHAVIOR OF DISSOCIATED CELLS 325 


tions of broken down cells, incorporated in the restitution mass as 
it grows? Again in such a preparation one finds some masses, 7, 
composed of finely granular material, with cell boundaries here 
and there vaguely showing, and sometimes with included nettle 
cells. These are probably lumps of ectoderm. 

If such a preparation be made and examined alive in sea water, 
the same elements are observed. Formalin removes the color 
from the granules in the entoderm cells, but in the living prepara- 
tion it may be seen that they have in general an orange tint, 
ranging from yellow to reddish, although some are colorless. The 
entoderm cells execute slow amoeboid changes of shape. 

When the stem pieces are pressed through gauze and the tissue 
is at once examined alive in a drop of sea water, it is found to con- 
sist of the same elements described above. With this treatment 
more entoderm cells seem to be ruptured and there are fewer 
groups of cells. The coenosare is almost entirely broken up into 
the elements a, /, g, h, of fig. 25. When the entire hydroid is cut 
up and pressed through gauze, again the same elements are found 
if the tissue is examined at once, although possibly more aggre- 
gates of cells occur than when stem tissue alone is used. 

If the drop of live tissue pressed out through gauze, or squeezed 
out without using gauze, be kept under observation, it may be 
seen that small masses are soon formed which include entoderm 
cells, enidoblasts, and the pale cells that probably are of ecto- 
dermic origin. As these masses grow in size 1t becomes impossible, 
owing to their opacity, to study their composition while alive 

As stated already, fusion between the cells of the pressed 
out stem tissue goes on so rapidly that in twenty minutes times 
small bars and plates, 1 to 3 mm. long, can be drawn off with a 
pipette. Some of these were preserved and sectioned and it 
could be seen that such masses were solid bodies of fairly uniform 
structure showing no stratification into incipient layers. The 
superficial part does not differ from the interior. The structure 
throughout is that of a cellular syncytium, that is in certain regions 
no cell boundaries can be seen, the protoplasm here appearing 
as a syncytial mass containing scattered nuclei, while in other 
places cell boundaries are visible. Even where cells are marked 


326 H. V. WILSON 


out it is probable that they are interconnected with one another 
and the rest of the mass by protoplasmic strands. The cells 
that are marked out vary in size and shape. Here and there cells, 
usually in groups, may be recognized by the contained granules 
as the original entoderm cells. Only a small fraction of the mass 
however is now made up of such cells, and yet the entodermic 
elements composed a very large part of the tissue when the fusion 
masses began to form. It is plain then that the entoderm cells 
after fusion to form, or rather help form, the plasmodial masses, 
undergo a transformation which effectually precludes us from 
recognizing them later. The large cnidoblasts formed a conspicu- 
ous set of elements in the tissue when fusion began, and these are 
to be seen in very considerable numbers scattered throughout 
the plasmodial mass at the stage under examination (twenty min- 
utes old). A comparison between the sections of this and later 
stages indicates that the bulk of the nematocysts carried over 
from the parent gradually disappear during the development of 
the plasmodial mass. If this is so, it is a question of some inter- 
est what becomes of the enidoblast cell itself? Does it share in the 
formation of the regenerative tissue? The point is worthy of 
special study, including as it does the idea of the de-specialization 
of a highly differentiated element. 

The protoplasm of the cell and syncytial areas in this stage is for 
the most part finely vacuolated so as to present a reticular appear- 
ance to a high power. The nuclei are in general large and con- 
tain abundant nucleoplasm. The mass at this time seems to 
have no surfacé film apart and distinet from the superficial syney- 
tial and cell areas. Finally in connection with this stage it may be 
said that owing to the transformation which the entoderm cells 
undergo after fusion, it does not seem hopeful to attack from 
purely histological evidence the question as to whether ectodermic 
and entodermic elements become segregated, the ones on the out- 
side, the others in the interior of the mass. There is of course 
always a possibility that this occurs, but it seems remote. 

Somewhat older plasmodial masses formed by the fusion of 
stem tissue pressed through gauze were studied. These were 
preserved 1 hour after fusion had begun, and were considerably 


BEHAVIOR OF DISSOCIATED CELLS 327 


larger than the mass just described. Sections show however that 
they have essentially the same structure. Cells or protoplasmic 
areas with the entoderm granules are still recognizable here and 
there. Perhaps most of such cells lie in the interior but some are 
found at the surface. This is not a point of importance for the 
question as to the possible segregation of ectodermic and entoder- 
mic elements, since as I have already explained a very large part of 
the entodermic material can no longer be recognized as such. It 
is clear that many of the areas or elements of the plasmodial mass 
now without granules, and with a vacuolated protoplasm, must 
have been formed from entoderm cells. 

The solid plasmodial mass does not long remain unstratified. 
As already said the masses uniformly secrete a perisare within 
about one day after fusion begins, and this in itself is probably 
evidence that the superficial layer has assumed something of an 
epithelial character. Masses preserved July 27, about seven- 
teen hours old, were sectioned. Some of them were healthy and 
alive all through, and fig. 20 represents a section through such an 
one. In this body an outer layer, ec., the ectoderm, has separated 
from an inner mass, y. en., the yolk entoderm. The latter as 
later stages show gives rise to the definitive layer of entoderm and 
a central yolk mass, as in the planula development of Pennaria 
(Hargitt, 00, 04). Both ectoderm and yolk entoderm are cellular 
syneytia in which free elements or apparently free elements are 
included. In both the ectoderm and yolk entoderm, some of the 
large nematocysts carried over from the parent, are present. On 
one side of the body it will be seen the differentiation into layers 
has not yet been carried out, the ectodermal region here shading 
off into the inner mass. The isolated and irregular cavitiesin the 
yolk entoderm doubtless represent the beginnings of the gastric 
cavity. Inothermassesof the same lot preserved at the same time 
the shape was lobular, and only the projecting lobules were alive, 
while the more central part of the body was dead or dying. See- 
tions showed that in the lobules the layers were present, in about 
the same stage of differentiation as in fig. 20. 

Restitution masses from Experiment August 3 that were pre- 
served nineteen hours after fusion began were sectioned. These 
proved to be in about the condition shown in fig. 20. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY. VOL. Il, NO. 3 


328 H. V. WILSON 


A mass two days old, from Experiment August 3, was sectioned, 
and a median section is represented in fig. 22. The mass was orig- 
inally spheroidal, but gave rise to three short outgrowths a, ), c. 
In this body the ectoderm and entoderm are well differentiated. 
The yolk entoderm of the earlier stage has obviously given rise 
to the entoderm, en., and to more centrally located yolk material, 
y, and the latter has been nearly absorbed. In places the ento- 
derm is still continuous with yolk elements. The best sections 
show that the entoderm is still a reticular syncytium, but the cell 
bodies are distinctly outlined in places. In the most distinct 
regions they have a columnar shape as in the adult. The ento- 
derm is now well stored with the spheroidal granules found in the 
adult cells. The ectoderm also appears to be in reality a reticular 
syncytium, but cell bodies are clearly differentiated and regularly 
arranged in the regions of the outgrowths. They have here an 
elongated, columnar shape. Some large nematocysts, apparently 
such as were carried over from the parent, are present in the ecto- 
derm. A very few such nematocysts are found in the entoderm, 
and these seem to be in a phase of dissolution. 

Other masses from the same experiment (August 3) that were 
four and five days old were sectioned, and the results may be 
briefly given. While the mass is still spheroidal and before it 
has developed outgrowths, it may differentiate an ectoderm, ento- 
derm and central yolk. By the time a well defined layer of ento- 
derm is present, the yolk mass is small in amount and consists of 
scattered spheres or small groups of spheres. In such spheroidal 
masses the ectoderm and entoderm have the character of reticular 
syneytia. As outgrowths develop, both ectoderm and entoderm 
assume the character of columnar epithelia, especially in the out- 
growths themselves. 


LEPTOGORGIA. FUSION OF DISSOCIATED CELLS 


The species used was Leptogorgia virgulata, the ‘sea feather’ 
that is common, especially under piers, in Beaufort harbor. Some 
introductory experiments were made under my direction by my 
assistant, Mr. O. W. Hyman. He established the fact that when 
Leptogorgia is cut into small pieces, and these pressed in gauze sacs 


BEHAVIOR OF DISSOCIATED CELLS 329 


in the usual way, the tissue is broken up into small masses and 
separate cells. By the subsequent union of such masses and cells, 
smooth balls up to and over 1 mm. in diameter are formed. ‘These 
remain alive for days in laboratory dishes, but do not transform. 

A record of two subsequent experiments is here given. 

Experiment August 9. Pieces 4 to 5 mm. long are cut from the 
upper end of a yellow colony. The horny axis occupies about 
one-fourth the total diameter. These pieces are simply squeezed 
with forceps in a watch glass of water. The tissue exudes freely 
from the cut ends. Much of it is stringy. With pipette it is 
dispersed and so broken up. It is then shaken to the center of 
the watch glass and the glass transferred to a bowl of water, 11:15 
a.m. 

Some of the tissue is now examined under a supported cover. 
It is made up as follows: (1) Ciliated cords and masses, vary- 
ing in size from large to minute are abundantly present. 
These are doubtless pieces of mesenterial filaments with mesen- 
terial tissue. (2) Motionless masses of loosely packed cells, 
also varying considerably in size are abundant. (3) Isolated 
cells and groups of a few cells are abundant (fig. 26). In fig. 26, 
a represents a characteristic small cell group made up of a few 
spheroidal cells with sharp outlines, some full of highly refractive 
granules, some merely containing a good many such. Similar 
granules form a compact mass at one end of the cell group, but this 
mass lacks a bounding pellicle. Separate spheroidal granular 
cells, b, resembling the constituents of a are common, and they 
may have pseudopodia. There is an abundance of small spheroi- 
dal masses of glassy protoplasm, c, the larger with one or two 
granules. Finally there are plenty of isolated granules, d, such 
as are found in the granular cells. In the category c the smallest 
elements must surely be fragments of cells or bits of intercellular 
connectives that have rounded off. 

The preparation under the microscope was watched for about 
an hour, and it was observed that fusion took place involving all 
the classes of constituents above enumerated. At 12 m. several of 
the larger ciliated masses were motionless, the surface bearing 
instead of cilia numerous small pseudopodia and transitional 
stages from cilium to pseudopod. 


330 H. V. WILSON 


The tissue left in the bowl was examined at intervals. The 
masses. grew evidently through fusion with one another and 
through incorporation of the granular cells and other elements. 
A typical mass at 2 p.m. is shown in fig. 27. The mass is full 
of granules like those of fig. 26, and in it the outlines of some sphe- 
roidal cells, like fig. 26 b, are distinguishable. Round it are similar 
granular cells and very many small masses of glassy protoplasm 
some with a few granules, some without any, ranging in size from 
mere points up nearly to the diameter of the granular cells. 

By 9 p.m. the process of fusion had gone so far that spheroidal 
masses with smooth surface, from about } mm. in diameter down- 
wards, were present. A number of these were now picked out and 
transferred to fresh sea water. In the formation of these masses 
about ten hours transpired. It is evident that during this time 
some regressive differentiation of the fusing lumps of tissue to a 
simpler condition took place. What the character of these 
changes were and how uniform they were, it would be interesting 
to know. 

The further history of the masses is briefly as follows. On 
August 10 they were alive. Atthistime they are perfectly opaque, 
and the smooth surface shows abundant fine flagellum-like 
pseudopods. Many of them are surrounded by a deposit of whit- 
ish material. This on examination proves to be made up of 
spheroidal cells of various sizes and degrees of granulation, which 
evidently have been slowly given off from the mass. This giving 
off of cells continues during the next day. It is probably evidence 
of a bad condition of the mass, and it is noteworthy that some of 
the masses do not exhibit it. 

All of these bodies were preserved on August 11, and several 
were later sectioned. While still alive something could be learned 
of their histological structure by gently crushing them under a 
cover glass. On doing so some of the contents streams out in the 
shape of small spheroidal masses. The larger of these are like the 
granular cells (6) of fig. 26. From such they range down to mi- 
nute particles of glass-like protoplasm just large enough to be seen 
at a magnification of 600. Of the intermediate sizes some are 
full of granules, others contain one or a few, while still others are 


BEHAVIOR OF DISSOCIATED CELLS 331 


Fig. 26 Leptogorgia. Cell aggregate and free elements of pressed out tissue. 
xX 1200. 
Fig. 27 Leptogorgia. Fusion mass two hours fifteen minutes old. > 150. 


without granules and glass-like. The granules in all these bodies 
are mostly of one size and yellowish. Such an examination by 
no means necessarily implies that in the natural condition the 
mass is composed of spheroidal cells. Rather, I assume that, on 
crushing, the cell bodies are separated and, where such exist, the 
intercellular strands are broken. A quick contraction would then 
make all the protoplasmic masses spheroidal. A body that has 
been slightly erushed under a cover glass in this way may heal. 
One such was kept for two hours in a moist chamber, and at the 
end of this time the body healed perfectly and was once more 
surrounded, as it originally was, by a smooth surface pellicle. 
Sections showed that these bodies did not have a uniform com- 
position. In one (ball 1) the structure was as follows: There is a 
surface film but sections give nothing definite as to its composi- 
tion. The interior is solid and shows no stratification into layers 
(ectoderm and entoderm). In many regions one finds protoplasm 
studded with nuclei but no cell boundaries can be made out. 
In other places there are small cells, rounded or angular that are 
very closely packed. In still other places while the tissue is 
compact the cells are slightly separated by unstained substance, 
probably fluid. In such places the cell bodies are distinctly 
outlined and intercellular strands of protoplasm are freely present. 


3a2 H. V. WILSON 


In a few small areas there is a scanty accumulation of mesogloeal 
jelly imbedded in which are some strands and small masses of 
nucleated protoplasm which are freely interconnected. The 
jelly stains blue with haemalum (or haematoxylin) and with a 2 
mm, objective appears homogeneous. These several observations 
show that the mass is a syneytium in different parts of which cell 
bodies are differentiated in various degrees. 


Fig. 28 Leptogorgia. Median section of a fusion mass two days old. Layers 
interpreted as ectoderm, ec, and entoderm, en, have developed. ms, mesogloeal 


jelly. > 350. 


Another mass (ball 2) has the structure shown in fig. 28. | There 
is a surface film which appears as a mere line. The general mass 
consists of the syncytial cellular tissue described for ball 1, but 
round half of the body two layers, apparently the ectoderm and 
entoderm, are differentiated. These layers are distinetly differ- 
entiated, although there is no mesogloeal jelly between them. At 
about the middle of the body they fade away into the general 
mass. The entoderm consists of more or less columnar elements, 


BEHAVIOR OF DISSOCIATED CELLS BaD 


the ectoderm of irregular cell bodies separated by a good deal of 
fluid and freely interconnected. Near the center of this ball is a 
considerable collection of mesogloeal jelly, of the same character 
as that deseribed for ball 1. 

In still another case (ball 3) while a part of the body resembles 
ball 1, at one end of the body the structure is that shown in fig. 
29. We find here an accumulation of finely granular, lightly 
stained material, cg, which is quite different from mesogloeal jelly 
and is probably a fluid that has coagulated in the fixation process. 


Fig. 29. Leptogorgia. Part of section through a fusion mass two days old, 
cg, coagulum; s.c., superficial cells. > 1200. 


The same ball contains some mesogloeal jelly near the center. In 
the region round the coagulated fluid the cells are loosely packed. 
They are more or less rounded and intercellular connections are 
practically absent. Vacuoles are common in these cells and a con- 
spicuous nucleolus is frequently to be seen. In this region the 
surface layer is formed of flattened cells, s.c. In this ball there is 
no differentiation of ectoderm and entoderm layers. 

Interpreting the results of this study of sections, it seems prob- 
able that ball 1 represents an earlier stage, and ball 2 a later one 
in which the coelenterate layers have begun to differentiate. 


334 H. V. WILSON 


In ball 8 the condition shown in fig. 29 is perhaps to be correlated 
with the gradual extrusion of rounded granular cells which goes 
on in the case of some of the masses during life, as already recorded. 
Such a condition is one, perhaps, into which the dense syncytial 
cellular structure passes when the struggle for life is going against 
the body. It may of course only be a mortuary change, viz., a 
step in a process of gradual dying. 

The results of this experiment, while very inconclusive, suggest 
that the bodies formed by the fusion of cell masses and cells would 
regenerate into new individuals if placed under good conditions, 
possibly hung out in gauze bags in a part of the harbor where 
the current is good. 

BHeperiment August 10. Small pieces of a Leptogorgia colony 
were pressed out through gauze at 4.30 p.m. The tissue that 
streams through the gauze is finer than that obtained in the experi- 
ment just recorded. It is made up of the elements shown in 
fig. 26 and of small opaque lumps of tissue, mostly spheroidal and 
many of them ciliated, which are commonly three or four times the 
diameter of one of the granular cells (fig. 26 6). The living tissue 
and the spicules are separated as far as possible. 

Fusion goes on and by 7 p.m. masses of irregular shape are 
present. These are transferred to fresh sea water. The next 
morning a number of smooth balls have been formed, some of 
which have incorporated spicules. These are kept for a couple of 
days during which they show no external signs of differentiation. 
Sections showed that these balls had essentially the same structure 
as those of the preceding experiment. 

Experiment to test the regenerative powers of a fusion mass when 
inserled in the body of the parent species. August 10. Six of the 
Leptogorgia fusion masses produced in the experiment of August 
9 were inserted in the parent species in the following way. A piece 
of an orange colored Leptogorgia, about five inches long, was slit 
lengthwise down to the horny axis. The slit so made was pushed 
open and the fusion masses dropped in with a pipette in a row. 
Ties were then made round the piece of Leptogorgia closing up the 
covering layer of polyps over or partly over the fusion masses. 

On August 11, 10 a.m., the ties are removed. The slit has not 
healed but the edges have curled in. The whitish fusion masses 


BEHAVIOR OF DISSOCIATED CELLS aT) 


may be seen in the slit. They have fused with one another in 
some degree, the number of masses now being four. The piece 
of Leptogorgia looks healthy; the polyps are well expanded. 
At 7 p.m. the whole preparation is preserved. The questions 
are: Have the fusion masses undergone any histological differen- 
tiation? Have they established union with the Leptogorgia? 

Sections through one of the masses showed that the body had 
grown deeper into the slit and had established connection with the 
Leptogorgia on one side of the slit. This connection included per- 
feet continuity with the entodermie lining of a coelenteric cavity 
which had been laid open, and also with the entoderm of several 
small coenenchymal canals in the neighborhood. The whole 
fusion mass is solid and somewhat club-shaped at its outer end 
where it shows a stratification into an outer stratum and an inner 
core. The thickness of the outer stratum is considerable includ- 
ing several layers of cells. The other masses did not penetrate 
so deep into the slit. They were found to be in continuity with 
the superficial layer of the Leptogorgia, but had not established 
connection with the interior of the latter. 

The indication from this experiment, which was merely meant as 
a tentative one, is that the fusion masses if allowed to grow would 
have become part of the Leptogorgia colony. 


ASTERIAS. FUSION OF THE DISSOCIATED CELLS OF THE 
IMMATURE GONAD 

Experiment August 5. Gonads 25 mm. long of the common 
starfish, Asterias arenicola, were cut into pieces about 5 mm. long 
and these were pressed through gauze at 12m. Abundant sepa- 
rate cells and small cell aggregates stream through the gauze to- 
gether with some larger pieces of gonad. The latter are picked 
out, and the remaining material is shaken to the center of watch 
glass. A drop of the material is now examined under the micro- 
scope. Many of the cells whether free or combined in small aggre- 
gates are coarsely granular. Both cells and aggregates show 
fine pseudopodia. The cells and the aggregates quickly combine 
and in a few minutes the field of the microscope presents the 
appearance shown in fig. 830. There are numerous small masses, 
such as a, which have been formed by the fusion of cell aggregates 


¥510) H. V. WILSON 


and separate cells; and there is an abundance of free elements. 
Among the latter coarsely granular cells like 6, with and without 
pseudopodia, are conspicuous. There are also many clear glass- 
like cells (¢) ranging down to bits just visible. Free granules 
resembling those of the granular cells are abundant. The masses 
(a) make the impression of being aggregations of the granular 
cells (b), but doubtless other elements enter into their composition. 
Abundant fine pseudopodia, occasionally branched, cover the sur- 
face of such masses. 

By | p.m. the tissue in the watch glass has combined to form 
a thin and extensive reticular plate, produced by the gradual 


a 


Fig. 30 Asterias. Small fusion mass and free clements of pressed out gonad 
tissue. a X 600. b and e X 1200. 


fusion of masses of many shapes. The reticulum in general is 
attached, though feebly, to the glass, but pieces 1 to 2 mm. 
wide have been broken off and are free. At 7 p.m. the whole 
reticulum is broken up into pieces of about this size, and all 
transferred to fresh sea water. On the following day a number of 
such pieces had contracted into smooth, massive bodies. But 
all pieces died in a day or two. 


BEHAVIOR OF DISSOCIATED CELLS Bad 


Experiment August 16. Gonads 20 mm. long were used. The 
gonads were cut into pieces and these simply teased up with needles 
in a watch glass of sea water. The pieces are thus broken into 
small masses, cells, and fragments, essentially like those shown in 
fig. 30. Fusion commences at once as in the former experiment, 
the isolated cells and the masses both throwing out pseudopodia. 
The granular cells in particular are observed to become intercon- 
nected by delicate and complex pseudopodial networks. The 
formation of pseudopodia by the granular cells may go on to such 
an extent that the granular substance of the cell body almost 
disappears in the network of pseudopodial strands. This e peri- 
ment was not carried farther. 

Sections showed that the gonads used in these experiments were 
in the indifferent stage. The germinal epithelium lining the 
follicle is more than one layer deep, and many of the nuclei are 
large and rich in chromatin. The epithelium has proliferated to 
such an extent that the lumen is nearly filled with cells. In 
sections these are compressed, with rather vaguely granular 
eytoplasm and nuclei which are smaller than in the lining cells. 
When the living gonad is slightly pressed, these cells exude and 
appear as the spheroidal granular elements described above. 


LITERATURE CITED 


Braem, I. 1908 Die Knospung der Margeliden, ein Bindegled zwischen 
geschlechtlicher und ungeschlechtlicher Fortflanzung. Biol. Central- 
blatt, Bd. 28. 

Cuinp, C. M. 1906 The development of germ cells from differentiated somatic 
cells in Moniezia. Anat. Anzieger, Bd. 29. 
1911 Die Physiologische Isolation von Teilen des Organismus als 
auslésungsfaktor der Bildung neuer Lebewesen und der Restitution. 
Vortrige u. Aufsiitze a. Entw-Mech. d. Organismen, Heft 11. 

Curtis, W. The life history, the normal fission and the reproductive organs of 
Planaria maculata. Proc. Boston Soe. Nat. Hist., vol. 30. 

Drinscu, H. 1902 Studien wt. das Regulationsvermégen der Organismen: 
6. Die Restitutionen von Clavellina lepadiformis. Arch. f. Entw.-Mech. 
Bd. 24. 

Evans, R. 1899 The structure and metamorphosis of the larva of Spongilla 
lacustris. Quart. Journ. Mier. Sci., vol. 42. 


338 


Harairr, 


LIEBERKU 


Maas, O. 


Metscunt 


H. V. WILSON 


C.W. 1900 A contribution to the natural history and development 
of Pennaria tiarella MeCr. Amer. Naturalist, vol. 34. 

1904 The early development of Pennaria tiarella MeCr. Arch. f. Entw.- 
Mech., Bd. 18. 

HN, N. 1856 Beitrage zur Entwickelungsgeschichte der Spongillen. 
Archiv fiir Anat. u. Physiologie, J. Muller. 

1901 Die Knospungentwicklung der Tethya und ihr Vergleich mit 
der geschlechtlicher Fortpflanzung der Schwimme. Zeitschr. f. wiss. 
Zool., Bd. 70. 

1906 Ueber die Einwirkung Karbonatfr. Salzlésungen auf erwachsene 
Kalkschwiimme u. auf Entwicklungsstadien derselben. Arch. f. 
Entw.-Mech. d. Organismen, Bd. 22. 

1910 Ueber Involutionserscheinungen bei Schwiimmen u. ihre Bedeu- 
tung fiir die Auffassung des Spongienkérpers. Festschr. z. sechzig- 
sten Geburtstage Richard Hertwigs, Bd. 3. 

KOFF, E. 1879 Spongiologische Studien. Zeitschr. f. wiss. Zool., Bd. 32. 


Mituier, Kart 191la Versuche it. die Regenerationsfihigkeit der Siisswasser- 


NUTTING, 
PERKINS, 


SCHULTZ, 


WELTNER, 


schwimme. Zool. Anzeiger, Bd. 37, Nr. 3-4. 
1911b Beobachtungen tiber Reductionsvorgiinge. bei Spongilliden, 
nebst Bemerkungen zu deren diusserer Morphologie u. Biologie. Ibid, 
Nr. 5. 
C. C. 1901 The hydroids of the Woods Hole Region. Bulletin of 
the U. 8. Fish Commission for 1899. 
H. F. 1902 Degeneration phenomena in the larvae of Gonionema. 
Biol. Bulletin, vol. 3. 
E. 1904 Ueber Reduktionen: 1. Ueber Hungererscheinungen bei 
Planaria lactea. Arch. f. Entw.-Mech., Bd. 18. 
1906 Ueber Reduktionen: 2. Ueber Hungererscheinungen bei Hydra 
fusea L. ITbid., Bd. 21. 
1907 Ueber Reduktionen: 3. Die Reduktion u. Regeneration des 
abgeschnittenen Kiemenkorbes von Clavellina lepadiformis — [bid., 
Bd. 24. 
1908 Ueber umkehrbare Entwickelungsprozesse u. ihre Bedeutung fiir 
eine Theorie der Vererbung. Vortrige u. Aufsitze iiber Entwicklungs- 
mechanik d. Organismen, Heft 4. 

W. 1893 Spongillidenstudien II. Archiv f. Naturgeschichte, 
Jahrg. 1893, Bd. 1. 


Witson, H. V. 1888 On the development of Manicina areolata. Journ. of Mor- 


phology, vol. 2. 

1894 Observations on the gemmule and egg development of marine 
sponges. Ibid., vol. 9. 

1907a A new method by which sponges may be artificially reared. 
Science, vol. 25, June 7, 1907. 

1907b On some phenomena of coalescence and regeneration in sponges. 
Journ. Exper. Zool., vol. 5. 

1911la On the regenerative power of the dissociated cells in hydroids. 
Proc. Amer. Soc. Zoologists, Science, Mar. 10. 

1911b Development of sponges from dissociated tissue cells. Bulletin 
of the Bureau of Fisheries, vol. 30. 


RHYTHMS IN. THE REPRODUCTIVE ACTIVITY OF 
INFUSORIA 


LORANDE LOSS WOODRUFF anno GEORGE ALFRED BAITSELL 
Sheffield Biological Laboratory, Yale University 
THIRTEEN FIGURES 


In a study of the life history of Paramaecium caudatum by 
pedigree cultures, Calkins clearly illustrated the cycle, while the 
rhythms in the division rate were later emphasized by Woodruff 
in a study of the life history of several species of hypotrichous 
Infusoria. The fluctuations in their rate of reproduction were 
classified as follows:! 

“A rhythm is a minor periodic rise and fall of the fission rate, 
due to some unknown factor in cell metabolism, from which recov- 
ery is autonomous.” 

A eycle is a periodic rise and fall of the fission rate, extending 
over a varying number of rhythms, and ending in the extinction 
of the race unless it is ‘rejuvenated’ by conjugation or changed 
environment” (ef. fig. 1). 

Gregory, in a study of the life history of Tillina magna, stated 
that ‘The curve which represents the general vitality of the proto- 
plasm shows the normal rhythmic fluctuations observed by Wood- 
ruff.” Gregory also made an analysis of the data secured by 
Popoff in his study of the life history of Stylonychia mytilus, 
and she stated that “If the curve of Stylonychia is plotted from 
average records of five and ten day periods, it will be found to 
correspond to the curves of Paramaecium, Oxytricha and Tillina, 
each showing the rhythmic periods of high and low vitality.” 
More recent work has shown that Paramaecium aurelia may be 
bred indefinitely on a culture medium which is varied from day 
to day,*1.e., the cycle does not occur under these conditions though 


1 Woodruff (’05). 2 Cf. Woodruff (‘11a), Taf. 26, 27. 
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 4 


NOVEMBER, 1911 


339 


F (GO, ‘YNApOOAA) = ‘sauT] UayoIq [Borjtoa Aq poywusSisop oe sumyyAYI 
SNOLMBA OY} JO S}IUIT] oY, “Spoyjour ainq[no saistped JepuN vidosnjul JO a}vd UOISsYy ayy UT suUIyZAYA, puw SopOAD, OYVAYSN][T OF, “wOTy 
-BNUITYS [BIOyTyAe Aq apoAo ysay Jo pus yw ,poywuoanfey, “WNIpour oanjjno uoIsnjur Avy ‘“sporsod fipp wa} 10} posvsoav uresv ‘syeurrue 
JO SOUT] INO} OY} JO UOTSIAIP Jo 99¥A ATIVp oFvIOAV VY} Burmoys “y oiny[no ‘xvypey VyoUyAXG Jo Aaoysty OFT oy} Jo Ydvay T “By 
TO6T 


“AON pue 
f “09q ‘po 


‘idy pus £061 Avy pus 
Arne = oung pus Avy “IBN ‘qoq puv-use ‘90d JAON ‘pO “4 qsnany Apne oun “dy “Ivy P 
098 008 002 009 008 ( 


00 


oO 


OF 


RHYTHMS IN THE INFUSORIA 341 


100 ; 200 288 
June 1904 July Aug. Sept. Oct. Nov. 


Fig. 2. Graph of the life history of Gastrostyla steinii showing the average 
daily rate of division of the four lines of animals which compose the culture, again 
averaged for ten day periods. Hay infusion culture medium. To illustrate a 
case where the rhythms are apparently absent for a considerable period. Compare 
with fig. 3. (Woodruff, ’05) 


the rhythms persist undiminished; and also that the same result 
may be attained by a constant culture medium of beef extract’ 
(cf. fig. 4). 

It is obvious from these investigations that the life history of 
Infusoria in pedigree cultures comprises many minor rhythmic 
fluctuations in the fission rate from which recovery is autonomous. 
The results with beef extract as a constant medium for Para- 
maecium aurelia naturally led to an intensive study of therhythms, 
in order to determine if these also can be eliminated by a still 
more constant environment, i.e., whether they are due to minor 
variations in the environment or to unknown intracellular phe- 
nomena, as originally stated. Possible sources of variation in the 
environment which might give rise to variations in the metabol- 
ism of the cell which would become apparent as rhythms in 
the rate of reproduction are: 1) Chemical composition of the 
culture medium, 2) Quantity and quality of the bacterial flora 
of the culture medium, 3) Excretion products of the paramaecia, 
4) Mechanical stimulation during isolation, 5) Light, 6) Baro- 
metric pressure, and 7) Temperature. The data secured which 
bear on this question are given in the present paper. 


8’ Woodruff and Baitsell (711). 


GEORGE A. BAITSELL 


WOODRUFF AND 


LORANDE L. 


342 


(co, ‘YNapoo Ay) ‘potsed 
Jo}IOYS B 1Oj oFv1VAV UB Aq MOIA 0} 4YZNoIG oq uvo puv yuasoid ore Aoyy SsajoyjsoAou ‘spotsad fipp wa} IOJ PASBIIAB ST 
o}BI OY} WOYAM ouITy JO YASus] urezt0d vw doy avodde you AvuT suTYyYAYA YSnoyyye yeyy yowy ayy aVVISN][ OF, “spotsed finp 
aay 1oy atay poyjoyd vyeq °Z “BY UL UMOYS TTUT4s VTAJSOLStY) Jo AN4[Nd oUIES OY} Jo L1048TY ast] oy} Joydery ‘Sy 


JaquIoAON 1940309 Joquieydag qsnany Ange FO6T ‘une 
* 886 002 oor 
= aa 
| 
| 
| 


00 


| 
— “0 
| 


OT 


—- + — —+-—-+4 


J oe 


RHYTHMS IN THE INFUSORIA 343 


METHODS 


The animals employed in this study were taken from the pedi- 
gree culture of Paramaecium aurelia (1) which one of us‘ has had 
under daily observation for fifty-one months and which has 
attained 2500 generations, up to the present time (August 1, 
1911), under the conditions of a varied environment, without 
conjugation or artificial stimulation. From this culture a sub- 
culture was isolated line by line on October 1, 1910, at the 2012th 
generation, and carried for ten months on a constant culture 
medium of beef extract. It was then discontinued.’ The average 
daily rate of division of the four lines of this subculture (IB), 
again averaged for five day periods, was computed and the result 
is graphically shown in fig. 4. 

The experiments in regard to the rhythms were begun on June 
8, 1911, by isolating two subcultures line by line from IB at the 
2335th generation, and placing the animals in a similar manner 
on depression slides in five drops of the beef extract medium. 
This medium consisted of a 0.025 per cent solution of Liebig’s 
extract of beef. The slides were kept in small moist chambers to 
prevent evaporation. The cultures were continued by isolating 
each day an organism from each of the four lines of the respective 
cultures, and placing it in fresh medium on a sterile depression 
slide. The number of divisions during the previous twenty-four 
hours was recorded at the time of isolation and from this data the 
graphs were drawn. One of these two subcultures was placed in 
a thermostat chamber at a temperature of practically 82° F. 
(culture IB82a) and the other in a chamber at a temperature of 
practically 76° F. (culture [B76a), and maintained at this tempera- 
ture for forty days. 

A second series of two subcultures was similarly started from 
IB on June 18th, at the 2346th generation, and treated exactly 
the same as the above cultures. The cultures of this series were 
designated IB82b and IB76b, respectively. A third series of 
two subcultures was isolated in the same manner from IB on 
June 28th, at the 2355th generation, and these cultures were 


. 4 Woodruff (’11a). 
5 Wor further details of this subculture (IB) ef. Woodruff and Baitsell (’11). 


344 


gg ok 


LORANDE L. 


WOODRUFF AND GEORGE 


r=} 
— 


1 CY — 


BYuNLvesadW3l SNOISIAIG 


A. 


BAITSELL 


Part 1 


4 


345 


« 
a 


IN THE INFUSORIA 


RHYTHMS 


‘ystsiod sumyyAyI oy} ssofayjioaou ‘padoydura st 4oviyx9 Joaq Jo WMIpeul 91n4 
-]NO JURYISUOD aTqRIIMS B WAY IN990 4OU seop efoAd 9Y} YSNOYIS yey JoRJ oY} oyBAYSNI[I OF, ‘spotsod Sutpuodsesi09 
ay} 1OF (*,) eangvsoduie} WOOL oSvIOAV ayy SoAIS ydvis waddn oy, ‘spotsed Avp aay 10} padvisav urese ‘ainyj[nod sty} 
ASIdUIOD YOTYA S]J9O ,1O4SIS, JO SOUT] INO} oY} JO WOTSTATIp Jo ay¥A ATIVp oBvIOAG OY} YUaSeIdar SoyvUIpPIO oY, “WNIpeu 
ainjpNd yowryxo Joog “TT6T ‘6g A[NE UO ponurTzUOdsIp sv 41 [GUN ‘OT6T ‘T 19q099Q UO JT a1N}]Nd WIOIZ UOTYBOTSI S4I JO aUATy 
ay) Woy ‘Gy orngqnoqns ‘eIpaine wNTOsBUIBIe JO A1oysty ajeTduI0D oy} Seats ydvIs amo] oY, (Zpue syed) FSI 


GWed—V 


346 LORANDE L. WOODRUFF AND GEORGE A. BAITSELL 


designated IB82c and IB76c. There were then the following 
cultures, all of which were kept in the dark, involved in this 
experiment: 


IB—at room temperature. 
IB82a—isolated from IB and continued for forty days at 82° F. 
1B76a—isolated from IB and continued for forty days at 
IB82b—isolated from IB and continued for forty days at 82° F. 
IB76b—isolated from IB and continued for forty days at 76° F. 
IB82c—isolated from IB and continued for thirty days at 82 
IB76c—isolated from IB and continued for thirty days at 76° 


Since the entire point involved in this study depends upon the 
constancy of the environment to which the animals are subjected, 
this will be considered in detail. 

1. Chemical composition of the culture medium. The culture 
medium was made up by weighing out the proper amount of 
Liebig’s extract of beef and diluting it with distilled water. The 
solution was then put into about one hundred test tubes, plugged 
with cotton and sterilized. The medium remained sterile until 
used. Since all the culture medium which was used throughout 
the experiment was made up at one time there was no variation 
in the medium itself during the work. 

2. Quantity and quality of the bacterial flora of the culture medium. 
Paramaecium i$ an animal which depends on bacteria for its food, 
and consequently these must be supplied. Sufficient bacteria 
were ‘automatically’ transferred with the animals at the first 
isolation to provide ample food until the next isolation at the end 
of twenty-four hours. Again at this time sufficient bacteria 
were ‘automatically’ carried over with the animals to infect the 
fresh medium in which they live for the following twenty-four 
hours, and so on. The quality of the bacterial flora was initially 
the same on all the slides because all the paramaecia used to start 
the various lines were taken from the same environment when the 
experiments were begun, and it is believed that this condition was 
maintained by the cross infection of all the slides almost daily. 
This also served to eliminate variations due to infections from 
the air of the moist chambers. Obviously the number of bacteria 
on a slide varied during the twenty-four hours between isolations. 


RHYTHMS IN THE INFUSORIA 347 


But a study of the preparations showed that the paramaecia 
keep down the results of the rapid multiplication of the bacteria 
by feeding on them, so that, although there is ample food for the 
animals at all times, the variation in the bacterial content of the 
medium is not so great as would at first glance appear to be the 
case. However, the point to be emphasized is that these varia- 
tions, small as they were, were only of twenty-four hour duration 
since fresh culture medium was supplied daily. Consequently 
any effect of the slight and unavoidable variation in the quantity 
of the bacteria could result only in an intradiurnal rhythm in 
the division rate, and since the count of the generations was taken 
at twenty-four hour intervals, this variation would not appear in 
our records. Elaborateness of method is not necessarily coin- 
cident with exactness of technique, and therefore it was considered 
unnecessary to attempt to ‘sterilize’ the paramaecia and feed 
them on pure cultures of bacteria. Any effort in this direction 
has met only with partial success and has introduced compli- 
eating factors which, it is believed, would more than counter- 
balance any advantages to be gained for the problem in hand. 

3. Light. Throughout the experiments all the cultures were in 
absolute darkness except for the short time daily when the count 
was being made. This was unavoidable, but each animal was 
not exposed to the light for more than three minutes. A control 
culture carried in the light showed that light does not influence 
the rate of reproduction of paramaecium. This is in accord 
with the previous results® on the effect of light on the division 
rate of free-living Infusoria. 

4,and 5. Excretion products of paramaecia and mechanical 
stimulation during isolation. These factors may be eliminated 
because they could only give rise to an intradiurnal rhythm 
which would not appear in the data. 

6. Barometric pressure. A careful study was made of the varia- 
tions in the barometric pressure which occurred during the experi- 
ments. There was absolutely no correlation between the small 
fluctuations in pressure and the rhythms in the division rate, and 


§ Maupas (’88) and Woodruff (’05). 


348 LORANDE L. WOODRUFF AND GEORGE A. BAITSELL 


consequently it can be positively stated that this factor plays no 
part in our results. 

7. Temperature. This is the chief possible variable in the envir- 
onment which we have to consider. In the original discussion of 
rhythms it was stated that ‘‘the results serve to emphasize the 
fact that while temperature does influence the rate of multipli- 
cation, it is not the most important element among the factors 
which cause fluctuations.” A study of figs. 4 and 5 shows that 
there is a certain amount of correlation between the fluctuations 
of fission rate and of temperature, when the cultures aresubjected 
to the ordinary changes in temperature of the laboratory. 

Our experiments were carried on in a Panum thermostat,’ 
heated at one end with a gas flame (with an automatic regulator) 
and cooled at the other by a large ice chest. The thermostat 
was divided into nine chambers, grading down in temperature 
from one end to the other of the apparatus. The temperature 
was recorded in each chamber by a maximum and minimum 
registering thermometer, by a tube thermometer, and in one 
chamber also by a thermograph. Experiments were conducted 
in six of the nine compartments, but an account is given here of 
the results of the cultures at the two temperatures within the 
optimum zone for the strain of Paramaecium being used. The 
detailed data in regard to the effect of different temperatures on the 
division rate of this animal and its relation to the temperature 
coefficient of chemical reactions in general will be published later. 
We should state, however, that our results at other temperatures 
are entirely concordant with those here described. 

The temperature selected for the study of the influence of 
temperature on the rhythms were approximately 82° F. and 76° F. 
as it was found that the optimum zone for the culture included 
these points. During the fifty days that the experiments covered 
the variations in temperature did not exceed 3° F.; for the greater 
part of the time the variation was less than 1° F., and for several 
periods it was less than 0.5° F. The greatest variation noted 


7 This apparatus was constructed for this and similar studies from an improved 
model designed by Professor L. F. Rettger of the Sheffield Bacteriological Labora- 
tory of Yale University. Our thanks are due Dr. Rettger for his kind codperation. 


| 

i} 

| 
OXYTRICHA ~ B 


| 
PLEUROTRICHA = A 


PLEUROTRICHA =~ B 
} 


Dec. Jan. February March April 
1902 1903 
Fig. 5 Sections of the culture graphs of two cultures of Oxytricha fallax and 
two cultures of Pleurotricha lanceolata, together with the graph of the average 
room temperature for the same period. Averages for ten day periods. To illus- 
trate a striking instance of the apparent relation of rhythms to the fluctuations in 
temperature. (Woodruff, ’05) 


349 


350 LORANDE L. WOODRUFF AND GEORGE A. BAITSELL 


above occurred during a week of unusually hot weather when the 
sudden change was too great to be immediately compensated for 
by the automatic regulator. Great care was taken in removing 
the preparations from the chambers for the daily count and isola- 
tion. The culture medium to be used on one day had’ been put 
the day before in the proper chamber of the thermostat, and con- 
sequently the animals were transferred to fresh culture medium 
of the same temperature. Of course any effect of variations 
arising from the daily transfers could only be intradiurnal and con- 
sequently would not appear in our results. It should also be 
emphasized that the recorded temperature was that of the air 
in the thermostat, whereas the animals were in culture fluid on 
slides within the moist chambers in the thermostat. The tem- 
perature of the moist chambers obviously was still more constant 
than that of the thermostat chamber, as likewise was the liquid 
in which the organisms were living. Consequently the varia- 
tions in temperature which the animals experienced certainly 
never exceeded 3° F. throughout the experiments, and this maxi- 
mum variation occurred only at one period. Fora period of ten 
consecutive days there was no visible variation greater than 0.4° F. 
It is believed that the temperature conditions were maintained 
as nearly constant as modern apparatus and the necessities of the 
experiment allowed. 


RESULTS 


The results can be stated briefly because graphs of the rate of 
reproduction bring out the points involved far better than a 
description by words. Fig. 6, A, gives the average daily rate of 
division of the four lines of ‘sister’ cells of Paramaecium aurelia, 
series IB82a, again averaged for ten day periods, at 82°F. Band 
C show the same for series IB82b and IB82c._ Fig. 7, A, B, and C, 
shows similarly the results derived from IB76a, and IB76b, and 
IB76e. Figs. 8 and 9 give the same data averaged for five day 
periods. Fig. 10 shows the average daily rate of division, for 
five day periods, of line 1 (of the four lines) of series IB82a and 
line 1 of IB76a. Fig. 11 gives the same data for series IB82b 
and IB76b. Fig. 12 presents the average daily rate of division 


RHYTHMS IN THE INFUSORIA Sol 


A B Cc 


Fig. 6 A, Graph of the average daily rate of division at 82°F. of the four lines 
of ‘sister’ cells of Paramaecium aurelia, series IBS82a, again average for ten day 
periods. B and C, Similar graphs for series IB82b and IB82c respectively. To 
illustrate rhythms in the fission rate when the cultures are subjected to practically 
constant conditions, including temperature. 


A B & 

Fig. 7 A, Graph of the average daily rate of division at 76°F. of the four lines 
of ‘sister’ cells of Paramaecium aurelia, series IB76a, again averaged for ten day 
periods. B and C, Similar graphs for series IB76b and IB76c respectively. To 
illustrate rhythms in the fission rate when the cultures are subjected to pract ically 
constant culture conditions, including temperature. Compare with fig. 6. 


302 LORANDE L. WOODRUFF AND GEORGE A. BAITSELL 


A B Cc 


Fig.8 A, Graph of the average daily rate of division at 82°F. of the four lines of 
‘sister’ cells of Paramaecium aurelia, series IB82a, again averaged for five day 
periods. B and C., Similar graphs for series IB82b and IB82c respectively. To 
illustrate the fact that rhythms in the rate of division appear more pronounced 
under practically constant environmental conditions. Compare with the last 
ten periods of the culture subjected to room temperature changes (fig. 4). 


of the four lines of IB82a and IB76a, respectively. The vertical 
dotted lines include the ten day period during which temperature 
variations were entirely absent, or not greater than 0.4° F. Fig. 
13 gives the same results for series IB82b and IB76b during the 
ten days of most constant temperature. 

A study of these graphs of the rate of reproduction of Para- 
maecium shows that the exceptionally and practically constant 
conditions of the environment failed to diminish or eliminate the 
rhythms—but on the contrary tended to bring them out more 
clearly. The fact that the rhythms appear more pronounced 


A B Cc 
Fig.9 A, Graph of the average daily rate of division at 76°F. of the four lines of 
‘sister’ cells of Paramaecium aurelia, series IB76a, again averaged for five day 
periods. B and C., Similar graphs for series IB76b and IB76e respectively. To 
illustrate the same point as fig. 8. 


RHYTHMS IN THE INFUSORIA 353 


Fig. 10 Graph of the average daily rate of division for five day periods of line 
1 (of the four lines of ‘sister’ cells) of series IBS2a (= continuous line) and of 
line 1 (of the four lines of ‘sister’ cells) of series IB76a (= - - - - line). To 
illustrate the fact that rhythms of practically the same amplitude and character 
appear in a graph of a single line of cells as appear when four such lines are 
averaged together. Compare with fig. 8, section A, and fig. 9, section A. 


under the practically constant conditions existing during these 
experiments than they do under ordinary laboratory conditions, 
clearly suggests that they are due to a fundamental factor in 
cell phenomena and not to extraneous causes. For if they are 
due to inherent intracellular conditions, one would a priorz expect 
to find them more clearly brought out when the cell is free from 
extraneous influences. 

A study of the curves of the division rate at the two tempera- 
turesshows that temperature, as is well known, markedly influences 
the rate, but it also shows that the rhythms persist—the repro- 
ductive activity being, as it were, pitched at a higher scale, but 
its character in no wise altered. In other words, it is not sug- 


354 LORANDE L. WOODRUFF AND GEORGE A. BAITSELL 


(ees) 


Fig. 11 Graph of the average daily rate of division for five day periods of line 
1 (of the four lines of ‘sister’ cells) of series IB82b (= continuous line) and of 
line 1 (of the four lines of ‘sister’ cells) of series IB76b (= - - - - line). To 
illustrate the same points as fig. 10. Compare with fig. 8, section B, and fig. 9, 
section B. 


gested that the division rate is not largely a function of tempera- 
ture—all other conditions being equal. It is probable that the 
temperature coefficient of the mean rate of division for a period 
including several rhythms will coincide closely with that of chem- 
ical reactions in general, but it is also probable for example that 
the rate of division at the crest of a rhythm at a high temperature 
and at the bottom of a rhythm at a low temperature will give a 
coefficient higher than the theory demands. Experiments to 
determine this point are in progress. 

It should also be pointed out that the total number of divisions 
attained during a prolonged period of time is comparatively con- 
stant. For example, the number of generations attained by 
culture I during 1909 was 613 and during 1910 was 612. Of 


ud 


RHYTHMS IN THE INFUSORIA 


*SUOT}IPUOD oINyVIoduIE, yURYSUOD AT[BOTJOvAd Jo potted v BulIMp o}eA oy} UT SUOT} 
-BIIVA POYIVU OY} MOYS 0} PUR ‘UOISTATP JO 9}¥1 ATIVp OY} JO plodad [BoTdAY B aYBAASNT[L OT, “YoG°O UBY} Sse] 1M Popslodor 


SUOTIVIVA 9INYVIOdUId} oY} YIM Surmp sAvp wo} oy} opNjour soul] peqyop [BoTyA0a oy, *(EUTy 


(ouT] snonutyu0s =) BZF 


J] Set1as JO S|] 


90 


‘ 


IO4SIS, JO SOUT] ANOJ oY} JO UOISTATP JO o}vA App dBvIOA 


---- =) BOJqI jo pus 


ayy jo ydeay 


a 
G 


1 Sl 


4 


NO. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 11, 


356 LORANDE L. WOODRUFF AND GEORGE A. BAITSELL 


Big. 13 Graph of the average daily rate of division of the four lines of ‘sister 
cells of series IBS2b (= continuous line) and of IB76b (= - - - - line) for the ten 
days during which temperature variations were practically absent. To illustrate 
the same points as fig. 12. Compare with the section of fig. 12 included within the 
vertical dotted lines. 


course this very exact coincidence is an ‘accident’ but, taken with 
a considerable amount of data along the same line, it quite defi- 
nitely points to the fact that the organism has the potential for 
about a certain number of bipartitions during a long period of 
time and this number is approximately attained irrespective of 
the minor fluctuations in the rate, due to external or internal 
causes. 

In a recent paper, Jennings states that 

“Within the same line the rate is sometimes very different for a cer- 
tain period, as a week or ten days, from the rate during the rest of the 
time. This is much more evident when one inspects a table in which the 
fissions are recorded day by day. The rate in a given line is there seen 
at times to drop, remain low for perhaps ten days, then return to the 


RHYTHMS IN THE INFUSORIA 357 


original rate. In most of all these cases there are evidences of patho- 
logical conditions during these periods of lowered rate of fission. Mon- 
strosities appear, and many of the specimens die. Therefore these peri- 
ods of slower rate are not to be considered as giving the characteristic 
rate for the race when healthy. In comparing different races, the 
periods when the rate of fission is high and uniform should be compared.” 


These observations of Jennings are difficult to understand in 
view of our results with Paramaecium. His statement in regard 
to weekly variations we would, at first glance, interpret as further 
evidence of rhythms; but throughout the more than four years 
of the life of this pedigree culture, a monster has never been seen 
in any of the direct lines, and only two or three times has a single 
deformed individual been seen in the heavy stock cultures which 
have been seeded from this strain. Further, it is an unusual 
occurrence for a line in any of our experiments to die out without 
assignable cause. Therefore it is necessary to emphasize that 
whereas the statement quoted seems, at first thought, to be in 
regard to periodic fluctuations in the rate of bipartition identical 
with those we call rhythms, nevertheless these fluctuations have 
absolutely nothing in common since, according to Jennings’ 
statement, those occurring in his lines are pathological. 


CONCLUSIONS 


The results of studies on the life history of free living Infusoria 
by exact pedigree culture methods show that, when these organ- 
isms are bred on comparatively constant culture media of hay 
or other infusions, the reproductive activity shows cycles and 
rhythms. Further results show that when Paramaecium aurelia 
is bred on a varied culture medium, or on a constant medium of 
beef extract, cycles do not occur, but rhythms persist. The 
results given in the present paper show that it isnot possible by 
constant environmental conditions to eliminate the rhythms and 
to resolve the graph of the multiplication rate into an approxi- 
mately straight line. It therefore seems justifiable to conclude 
that there are inherent rhythmical changes in the phenomena of 
the cell which are brought to view still more clearly when not 


358 LORANDE L. WOODRUFF AND GEORGE A. BAITSELL 


influenced by external factors. Variations in the rhythm of divi- 
sion is well-known in the development of the metazoon egg and it 
has yet to be satisfactorily explained. Towle in a paper on the 
effects of stimuli on Paramaecium makes the following state- 
ment: “There may even prove to be rythmical changes in sensi- 
tiveness like those described by Lyon for cleaving eggs, and Scott 
for unfertilized eggs. Something of this nature is indicated by the 
fact that paramaecia from the same culture vary in sensitiveness 
from day to day.”’ Woodruff (05) wrote: ““In my work on the 
effect of chemicals on Infusoria I have found that individuals 
react differently at various times to a given stimulus and I believe 
we have the clue to these changes in sensitiveness manifested in 
the rhythms of the fission rate.”’ 

Finally, the data justify the conclusion that the cells of this 
pedigree culture of Paramaecium aurelia have the potentiality 
to perpetuate themselves indefinitely by division (under proper 
environmental conditions)—the only necessary variations in the 
rate of reproduction being normal minor periodic rises and 
falls of the fission rate, due to some unknown factor in cell 
phenomena, from which recovery is autonomous. 


LITERATURE CITED 


G.N.Carxins 1904 Studies on the life history of Protozoa. IV. Death of the 
Aseries. Conclusions. Journ. Exper. Zool., vol. 1, no. 3. 


L. H. Gregory 1909 Observations on the life history of Tillina magna. Journ. 
Exper. Zool., vol. 6, no. 3. 


H.S. Jennines anp G. T. Harairr 1910 Characteristics of the diverse races of 
Paramaecium. Journ. Morph., vol. 21, no. 4. 


BE. P. Lyon 1902 Effects of potassium cyanide and of lack of oxygen upon the 
fertilized eggs and embryos of the sea urchin. Amer. Journ. Physiol., 
vol. 8, no. 1. 


1904 Rhythms of susceptibility and of carbon dioxide production in 
cleavage. Amer. Journ. Physiol., vol. 11, no. 1. 


E. Maupas 1888 Recherches expérimentales sur la multiplication des infusoires 
ciliés. Arch. d. zool. exper. et gén., 2me ser., T. 7. 


RHYTHMS IN THE INFUSORIA 359 


M. Pororr 1907 Depression der Protozoenzelle und der Geschlechtszellen der 
Metazoen. Archiv f. Protistenkunde, R. Hertwig Festband. 


J. W. Scorr, 1903 Periods of susceptibility in the differentiation of the egg of 
Amphitrite. Biol. Bull., vol. 5, no. 1. 

E. W. Towtse 1904 A study of the effects of certain stimuli, single and com- 
bined, upon Paramaecium. Amer. Journ. Physiol., vol. 12, no. 2. 


L. L. Wooprurr 1905 An experimental study on the life history of hypotrichous 
Infusoria. Journ. Exper. Zool., vol. 2, no. 4. 


1909 Further studies on the life history of Paramaecium. Biol. Bull., 
vol. 17, no. 4. 

191la Two thousand generations of Paramaecium. Archiv f. Protis- 
tenkunde, Bd. 21, 3. 

1911b The effect of excretion products of Paramaecium on its rate of 
reproduction. Journ. Exper. Zool., vol. 10, no. 4. 


L. L. Wooprurr anv G. A. Bairsett 1911 The reproduction of Paramaecium 
aurelia in a ‘constant’ culture medium of beef extract. Journ. Exper. 
Zool., vol. 11, no. 1. 


CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSHUM OF COMPARATIVE ZOOLOGY 
AT HARVARD COLLEGE, E. L. MARK, DIRECTOR, NO. 226 


THE REACTIONS OF EARTHWORMS TO DRY AND 
TO MOIST SURFACES 


G. H. PARKER anp H. M. PARSHLEY 


The certainty with which an earthworm that is creeping over 
a partly moistened surface will avoid dry areas is well known to 
students of animal activities. It is the object of this paper to 
discuss briefly the character of this response, the location of the 
receptors concerned in it, and the nature of the stimulus. The 
work was carried out on the common dungworm, Allolobophora 
foetida (Sav.), but there is reason to believe that the results 
obtained apply equally well to most species of earth-worms. 

If a normal worm is allowed to creep over a horizontal sheet 
of filter paper that is wet with tapwater excepting for a few spots 
and if it is directed by some such stimulus as light toward one of 
these dry spots, on reaching the spot, it will usually continue to 
creep over the dry surface for a distance varying from a few seg- 
ments to half its length, stop, swing its head from side to side, 
then draw the anterior part of its body back to the moist region, 
and finally proceed to crawl in a new direction over the moist 
part of the paper. Of seventy worms put to this test only four 
failed to show the series of reactions just described. These four 
crept completely across an extensive dry area without showing 
the characteristic reaction, but a few days later three of these 
worms responded in a normal way, showing that their previous 
atypical condition was probably due to some unusual and tem- 
porary state. It is, therefore, fair to conelude that Allolobophora 
as a rule avoids dry areas. 

To ascertain the part of the worm that is stimulated by a dry 
surface, several kinds of experiments were tried. To test the 
sensibility of the posterior end of the worm, individuals were 
made to creep backward by touching their anterior ends slightly 

361 


362 G. H. PARKER AND H. M, PARSHLEY 


and, when thus creeping, were directed toward a dry surface. 
Thirteen worms tested in this way made considerable backward 
excursions over such surfaces showing, as was to have been 
expected, that the posterior end of the worm is not especially 
sensitive to dryness. Next, forty-five worms, all of which had 
been found to respond normally to a dry surface, were deprived 
of their prostomiums and in some instances of an adjacent seg- 
ment or two, and were shortly afterwards tested on filter paper. 
All crept freely over a dry surface without showing the lateral 
movements and the retraction of the head characteristic of nor- 
mal worms. The regeneration of the prostomium takes place 
in from one to two weeks according to the extent of the injury. 
This regeneration was found to restore to the worm its original 
sensitiveness to dry surfaces and enabled it to react again ina 
typical fashion. There is, therefore, every reason to believe 
that the region of the prostomium is the portion of the worm 
that is stimulated by dryness. 

The terminal surface left after the removal of the prostomium 
offers more or less of an obstacle to the ordinary locomotor move- 
ments of the worm and to avoid this feature in the experiments, 
supplementary tests were made in which the prostomium, instead 
of being removed, was anaesthetized. After some preliminary 
trials, three anaesthetics satisfactory for this purpose were found; 
they were a weak solution of chlorotone, a 35 per cent solution of 
magnesium sulphate, and a 1 per cent solution of ether, all aque- 
ous. If the anterior tip of a worm is bathed with one of these 
solutions for from one to five minutes, the prostomium is found 
to remain insensitive to a dry surface for one or more days. 
Such anaesthetized worms will creep persistently over a surface 
of dry filter paper on which, before anaesthetization, they could 
not be induced to advance more than a very short distance. Full 
recovery from the effects of the anaesthetic occurs in a day or 
two. It is, therefore, clear from this evidence also that the pros- 
tomium and possibly some of the adjacent parts of the worm are 
the receptive surfaces for this response. 

It might be supposed that the greater harshness of dry filter 
paper as contrasted with moist filter paper, instead of the simple 


REACTIONS TO DRY AND MOIST SURFACES 363 


absence of water, was the significant factor in these reactions, 
but such is not the case, for, if worms are allowed to creep on 
surfaces that remain equally rough whether they are wet or dry, 
the same reactions are observed as in the tests in which filter paper 
was used. Such surfaces as those of bricks, tiles, ete., present 
_these conditions. In trials with the moist and dry surfaces of 
bricks results similar to those got on filter paper were obtained. 
Moreover worms drew back from a dry smooth brick to creep on 
a wet rough one, and from a dry rough brick to creep on a wet 
smooth one, showing that the presence or absence of moisture, 
and not roughness or smoothness were the significant elements 
in these reactions. 

From these observations, it is quite evident that the prostomial 
region of an earthworm can be stimulated by dryness to such an 
extent as to call forth vigorous locomotor responses of a charac- 
teristic kind. A moist surface seems to be unstimulating and to 
afford merely a condition favorable for the locomotion of the 
animal. In this respect the earthworm is the reciprocal of the 
human being, for our skin is more receptive to the condition of 
wetness than to that of dryness. With us, however, the sensa- 
tion of wetness is produced in all probability by a complex of 
pressure and temperature stimuli, whereas in the earthworm the 
response to dryness is dependent very likely upon a simpler stim- 
ulus. This is apparently the selective extraction of water from 
the peripheral protoplasm of the worm, a process which is favored 
by the eapillarity of the dry surface over which the worm begins 
to ereep and is probably dependent chiefly upon evaporation 
from the surface of the worm itself. Under such circumstances 
the materials in the peripheral protoplasm of the prostomium 
must become concentrated and probably initiate stimulation by 
undergoing some such change as partial coagulation. Processes 
of this kind are not well exemplified in the outer skin of man, but 
are more nearly comparable with what occur in our mouths when 
by excessive evaporation the oral surfaces become somewhat dry. 


AN ATTEMPT TO ANALYZE THE CONSTITUTION 
OF THE CHROMOSOMES ON THE BASIS OF SEX- 
LIMITED INHERITANCE IN DROSOPHILA 


T. H. MORGAN 


From the Zoological Laboratory, Columbia University 


FOUR FIGURES—COLOR PLATE 


In several preliminary notes I have given a brief account of 
the origin of four mutations in the eye-color of the fly, Droso- 
phila ampelophila. The heredity of these eye-colors may now 
be given in full, and the bearing of the results, on sex-limited 
inheritance in general, discussed. In addition to the eye-color 
data I shall also. describe a few cases in which two other sex- 
limited characters have been studied in connection with eye- 
color, namely: short proportionate wings and yellow body-color. 
A full account of the heredity of these latter two characters will 
be reserved, however, for later publication. Here they are 
used only in so far as they give an opportunity to study the mode 
of inheritance of three sex-limited characters present in one 
individual.! 

The experiments on Drosophila have led me to two principal 
conclusions: 

First, that sex-limited inheritance is explicable on the assumption 
that one of the material factors of a sex-limited character is carried 
by the same chromosomes that carry the material factor fer femaleness. 

Seconp, that the ‘association’ of certain characters in inher- 
itance is due to the proximity in the chromosomes of the chemical 
substances (factors) that are essential for the production of those 
characters. 


1 The facts here recorded were first announced in a public lecture given in the 
Marine Biological Laboratory at Woods Hole, Mass., July 7, 1911. 
365 


366 T. H. MORGAN 


PART I 


THE HEREDITY OF RED, VERMILION (OR BRIGHT-RED), PINK, 
AND ORANGE EYES 


The eyes of the wild fly are dull red, and may be designated 
by the letter R. The bright-red eye is vermilion in color and is 
indicated by V in the tables. 

The pink eye is more translucent than the red eye, but of about 
the same general tone. It lacks the dark fleck seen in the red 
and vermilion eye when the eye is examined with a lens. This 
black fleck changes its position as the lens travels over the eye. 
The pink eye, P, is with a little experience easily distinguished 
from the other colors, especially in newly hatched flies. When 
the fly gets old the eye turns to a brown color very characteristic 
of this type of eye. 

The orange is the faintest eye color in the series. If the fly 
is very small it may be only tinged with orange. If the fly is 
large (coming from a well-fed maggot) the orange eye, O, is deep 
orange in shade; and without some experience it may be confused 
with the pink eye, especially if a mixed culture containing flies 
of different ages and sizes is examined. A little experience will 
soon make one familiar with the difference between these two 
colors. 

I do not hesitate to state that there are no intergrades between 
these eye-colors. Each color is distinct and breeds true to its 
kind. Moreover, the heterozygous flies show the dominant 
color. One ‘dose’ is indistinguishable from two doses of the color 
determiner. ' 

In making the matings and recording the numbers I have been 
assisted in the experiments with eye-color by Miss Eleth Cattell; 
and in the experiments in sex-limited inheritance for three 
factors by Miss E. M. Wallace and by Miss M. B. Abbott. I 
wish to express here my appreciation of the assistance that they 
have given. I have discussed the theoretical results of the eye- 
color inheritance with Mr. A. H. Sturtevant, as the work went on. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 367 


and this discussion has been helpful to me in finding suitable 
formulae for the data. 

Rather than defer the discussion of the interpretation of the 
results to the end of the account I will take up each case in turn. 
A few words will suffice to make clear the symbolism used. The 
red eye of the wild fly seems to contain three pigments: red, pink, 
and orange. The mutants have arisen by the loss in turn of one 
of the factors that make possible the development of the red 
color. If these three colors (or the factors that stand for them) 
are represented by the symbols R, P, and O, then the red eye is 
RPO, the pink eye is rPO, the orange eye is rpO. Obviously 
there is another combination possible, viz: the loss of the pink 
factor and the retention of R and O, giving RpO, which is the for- 
mula for the bright red or vermilion eye. The matter may be 
better expressed in another way. Should from any cause what- 
soever the factor for pink (P) drop out, vermilion (RpO) would 
appear. If, on the other hand, the red factor (R) should be lost 
from the red-eyed fly, pink would result (rPO). By crossing a 
vermilion fly with a pink one, some orange-eyed flies (rpO) 
would appear in the second (inbred) generation by recombination. 

In the formulae that follow it is always assumed that one dose 
of red or pink gives the same result as do two doses, which accords 
with the facts. . 

It is necessary to say a word in advance about sex determina- 
tion in these flies. I assume that every egg after eliminating its 
polar bodies, contains the sex chromosome, called X. Prior to 
their extrusion the egg, like all the other cells of the female, 
contains two X’s or XX. The male cells contain one X. Half 
the spermatozoa contain one X, the other half lack X. Miss 
N. M. Stevens has shown that these relations are actually present 
in Drosophila. The peculiar ‘coupling’ of X with the factor for 
pink, that runs through the formulae and gives the significant 
results connected with sex-limited inheritance, will be discussed 
later. 


368 T. H. MORGAN 


Red eye by vermilion eye 


When red-eyed females were crossed with vermilion-eyed 
males all the offspring (93 in number) were red. These inbred 
produced in the second generation red females, red males, and 
vermilion males. The result shows that vermilion is sex-limited. 
The following table gives the results and numerical data: 


Red tQ.ancs stn ont een rE O02 
Red @ by Vermilion #7 = fs 2» Fests DE. ee Oe Le) 
aoe | Vermilion MAMAS AS aes - 110 
The two classes of males taken together number 289, which is a 
close approximation to the 302 red females. The results may 
be accounted for in the following way: 


Red 9° 
Vermilion @ 


= RPOX — RPOX 
= RpOX — RpO 


F Red @ RPOXRpOX 
: Red @  RPOXRpO 
: Red @ RPOX — RpOX 
Gametes of F, Red ¢ RPOX —RpO 
RPOXRPOX Red 92 
RNGocernivon RPOXRpOX Red 9 
Bh eS RPOXRpO Red 3 
RpOXRpO Vermilion @ 


It will be noted that the red females belong to two classes, one 
pure, the other heterozygous: the red male is also heterozygous, 
while the vermilion male is pure. 

The reciprocal cross, namely, red male by vermilion female 
gives red females and vermilion males. In other words the 
daughters are like the father and the sons like the mother. This 
gives what I eall criss-cross inheritance. When these F,’s are 
inbred they give red males and females and vermilion males and 
females as shown in the next table. 


CHROMOSOMES AND SHEX-LIMITED INHERITANCE 369 


’ 


| Red 9 Bagot sree eee LOG 

coy RON NG Reduct: Aone enter er oe 

eaters uiacrnil J C 

Red & by Vermilion 9 acre ee Veunilion. 9 agi. 207 
Vermilion = 186 


The two preceding crosses are typical for all cases of sex- 
limited inheritance in Drosophila, and for some, perhaps for all, 
other cases. They may be summed up in the statement that 
where in one combination a character in the grandfather is trans- 
mitted to his grandsons alone, the reciprocal combination gives 
criss-cross inheritance. 

The number of males and of females in each class is approxi- 
mately equal in the F, generation. The results are accounted for 
as follows: 


Vermilion @ RpOX — RpOX 
Red & RPOX — RpO 


Red 2 RpOXRPOX 
Vermilion 7 RpOXRpO 


Red 9 RpOX — RPOX 


SERGI Osh Vermilion *@ RpOX — RpO 


RpOXRpOX Vermilion 9° 


F, Generation RPOXRpOX Red 9? 
RpOXRpO Vermilion & 
RPOXRpO Red & 


Red eye by pink eye 


The results of this cross have been already published (Science, 
1911), but the hypothetical explanation not given. For the sake 
of completeness the facts must be restated here. Red female 
by pink male gave red male and red female offspring. These 
inbred gave in the F; generation 3063 red males and females and 
169 pink males and females. 


370 T. H. MORGAN 


Red 9? | 
(Red 2 \ Red of“ 
\Red @ /Pink 9? | 

Pink @f 


. 8063 
Red 9 by Pinkw= ¢ 
....169 


In this case there is no sex-limited inheritance. An analysis 
of the result, based on the same formulae, gives the following: 


Red @ RPOX — RPOX 
Pink o& rPOX — rpO 


Red 9 RPOXrPOX 
Red &@ RPOXrpO 


Red 9 RPOX — rPOX 


Gametes of F; Red o& RPOX — rPOX — RpO — rpO 


RPOXrPOX 


RPOXRPOX ) 
3 red 9 
rPOXRPOX 


rPOXrPOX 1 pink 9 


RPOXRpO 

F, RPOXrpO 3 red & 
rPOXRpO 
rPOXrpO 1 pink @ 


The expectation is three times as many red females as pink 
females, and three times as many red males as pink males. 
The actual ratio is about 20 to 1, taking the two sexes together. 
Thus while all the classes,are represented, and the reds in excess 
of the pinks, they are much more numerous than expectation. 
The cause of this deficit in the pinks will be discussed later when 
other similar results can be brought forward. 

The reciprocal cross, red males and pink females, gave in the 
first generation red males and females. These produced in the 
F, generation red males and females, and pink males and females 
in the proportions shown in the next table. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE oil 


Red @ | 
_ J Rede Red af 
~ \Redi /Pink 2 | 

Pink wf 


Red & by Pink 


The analysis of the result, based on the same formulae, is as 
follows: 


Pink 9 rPOX — rPOX 
Red & RPOX — RpO 


Red @ rPOXRPOX 


Ha Red @ rPOXRpO 


Red @ rPOX — RPOX 


COURS 1) Red @ rPOX — RPOX — ipO0 — RpO 


rPOXrPOX Pink @ 
rPOXRPOX Red @ 
RPOXrPOX Red 9 
RPOXRPOX Red 9 
z rPOXrpO Pink & 
rPOXRpO ted 
RPOXrpO Red & 
RPOXRpO Red & 


In this combination also the expectation is three red females to 
one pink female and three red males to one pink male, while the 
realization is about 5 to 1 for all reds versus all pinks. 


Red eye by orange eye 


When red-eyed females are crossed with orange-eyed males 
all of the offspring have red eyes. These inbred produce red- 
eyed males and females, pink-eyed males and females, and ver- 
milion-eyed, males and orange-eyed males. No females with 
vermilion or with orange eyes appear in the second generation. 
Here two characters are, in a sense, sex-limited, although the 
parents showed only one of them, viz., the orange. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 4 


« 


Bie T. H. MORGAN 


Red OA Ae. cae eee 138 

Rede Gatien. oc 102 

, _ J Red @ } Vermilion @.... ; eee OS 

Heda ThRvAO Teneo =e aedie PYRO Rs oi -by ee 5 
Pinner ccc ote e3 

Orange ¢..... Fara eovs Pence: 5083 


The number of pinks and oranges is very small, although the 
other classes contain a fair number of offspring. The analysis 
follows: 


Red 2 RPOX — RPOX 
Orange o@ rpOX — rpO 


F, Red 9 RPOXrpOX 
Red & RPOXrpC 


Red? RPOX — rPOX — RpOX — rpOX 


ees ont Redo’ RPOX — rPOX — RpO — rpO 


RPOXRPOX | ; 
3 


RPOXrPOX red Q 
rPOXRpOX | 
rPOX1POX 1 pink @ 
RpOXRPOX | 
RpOXrPOX }; 3 red 9 
rPOXRPOX 

F, rPOXrPOX 1 pink @ 
RPOXRpO | 
RPOXrpO a8} red J 
rPOXRpO j 
rPOXrpO 1 pink @ 


RpOXRpO | 
RpOXrpO 3 vermilion 
rpOXRpO | 


rpOXrpO 1 orange o 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 373 


In this case the expectation was far ahead of the realization; 
for, while the red females to the pink females are estimated as 
3 to 1 they are as 37 to 1 in the actual count. Again the vermilion- 
eyed males should be as numerous as the red males, but they are 
not half as numerous. Thus while the formulae give the classes 
actually realized in the experiment with the sexes properly dis- 
tributed—a matter of no small complexity—yet the numerical 
results are by no means the expected ones. 

The reciprocal cross, red males by orange females, gives red 
females and vermilion males. These inbred produce eight classes 
in the second, or F2, generation. These eight classes represent, 
in fact, the whole gamut of eye-colors. 


Red @ Be jen LOS 

REGU Ue eras icra cise ctaseeye LOO 

Vermilion @ Epix estechsias 151 

Redo pyran pene ui Red oN Vermilion SAG erp OR CH acsntS o 135 
(Vermilion #7 SE Pink 9 Soonteieee bee 

Pinks Guta eee ti areeeee 

[oes OrancepOwa- ee: ae rsasves 13 

Oran gevigha nae ees mae cco 31 


In this combination the number of pinks and oranges is by no 
means so small as in the preceding case, although the other colors 
are not much more numerous than before. The analysis is as 
follows: 


Fi 


Orange 
Red 


Red 
Vermilion 


T. H. MORGAN 


9 rpOX ,— rpOX 
o RPOX — RpO 


9 rpOXRPOX 
ao rpOXRpO 


Gametes of Fy 


Red @ rpOX — RPOX — RpOX 
Vermilion o& rpOX — RpOX — RpO 


— rPOX 
— rp.O 


The expectation in the F, generation is 3 reds, 3 vermilion, 1 
The numbers realized are somewhat in this 
same ratio, except that the pinks and the oranges again run 


pink, 1 orange. 


behind their schedules. 


rpOXrpOX Orange 
rpOXRpOX Vermilion 
RPOXrpOX Red 
RPOXRpOX Red 
RpOXrpOX Vermilion 
RpOXRpOX Vermiblon 
rPOXrpOX Pink 
rPOXRpOX Red 
rpOXRpO — Vermilion 
rpOXrpO Orange 
RPOXRpO Red 
RPOXrpO Red 
RpOXRpO — Vermilion 
RpOXrpO Vermilion 
rPOXRpO Red 
rPOXrpO Pink 


A curious, and I think significant, 
relation will be observed between the sexes in the last two classes; 
for, the pink females are twice as numerous as the pink males, 
while the reverse holds for the orange-eyed flies. 
relation comes up again in the ‘pink-male by orange-female 


cross to be described later. 


The same 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 375 


This completes the crosses between red and the other colors. 
We may now take up the remaining combinations. It will be 
noted that from now on the results are about an exact duplication 
of the series just described. The results may be said to be mirror 
figures of each other which suggests the fanciful idea that the 
combinations of colors, that the tables represent, have some such 
stereometric relation. 


Vermilion eye by pink eye 


When a vermilion-eyed female is crossed with a pink-eyed 
male, all the female offspring are red, and all the male offspring 
are vermilion. These inbred produce in the second generation 
all four classes of both sexes: 


Red 9 . 110 

Red @... 62 

Vermilion @ 104 

Re ' ast Red 92 Vermilion @.. . 84 
Vermilion 2 by Pink o = eatin of » Pink 9 36 
Pink @ a 6 

Orange 9 . 36 

Orange © 17 


A deficiency in the males of every class is noticeable in this cross. 
The total of all the females is 286 and of all the males 179, nearly 
2to1. The analysis, as shown in the next table, calls of course 
for equal numbers. 


376 T. H. MORGAN 


Vermilion 9 RpOX — RpOX 
Pink o& rPOX — rpO 


Red 2 RpOXrPOX 


Fi Vermilion o& RpOXrpO 
; = Red 9 RpOX — rPOX — RPOX — rpOX 
Gametes of Fi Vermilion o& RpOX — rpOX — RpO-= — rpO 
RpOXRpOX = Vermilion 9? 
RpOXrpOX = Vermilion 2 
rPOXRpOX = Red @ 
rPOXrpOX = Pink 9° 
RPOXRpOX = Red @ 
RPOXrpOX = Red 9 
rpOXRpOX = Vermilion ? 
F, rpOXrpOX = Orange 9 
RpOXRpO = Vermilion @ 
RpOXrpO = Vermilion @ 
rPOXRpO = Red @ 
rPOXrpO = Pink @ 
RPOXRpO = Red & 
RPOXrpO = Red & 
rpOXRpO = Vermilion #7 
rpOXrpO = Orange & 


The expectation both for males and females is 3 red, 3 vermilion, 
1 pink, 1 orange. The females give approximately this result, 
while the males fall below the expectation, especially the pink 
and orange males. 

The reciprocal cross, vermilion male by pink female, gives all 
red offspring. These inbred give for the F, generation red males 
and females, vermilion males, pink males and females, orange 
males. Here again a case of double sex-limited inheritance 
occurs, and of course in the same colors, vermilion and orange, as 


before. 
Red @.. satiate) 
| Red & _ 804 
pt Gah A) {Red 9 \_J] Vermilion &@. Oot, 
Vermilion &@ by pink ° \Red ¢ / Pink? © 214 
| Pink @. 99 


Orange 84 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 377 


The numbers are relatively high in this experiment, and signifi- 
cant. The red and the vermilion males taken together give 
approximately the same number as the red females. Similarly 
the pink and the orange males, taken together, are nearly as 
numerous as the pink females. The following table shows what 
the expectation is in regard to these numbers: 


Pink @ rPOX — rPOX 
Vermilion « RpOX — RpO 


Red 2 rPOXRpOX 


IB Red @ rPOXRpO 
Red 9 rPOX — RPOX — rpOX — RpOX 

Gametes of Fi Red @ rPOX — RPOX — rpO — RpO 
rPOXT POX = Pink @ 
rPOXRPOX. = Red 9° 
RPOXrPOX = Red @ 
RPOXRPOX = Red 9 
rpOXrPOX = Pink 9 
rpOXRPOX = Red 9 
RpOXrPOX = Red @ 

F, RpOXRPOX = Red @ 
rPOXrpO = Pink ¢ 
rPOXRpO = Red & 
RPOXrpO = Red @ 
RPOXRpO = Red & 
rpOXrpO = Orange & 
rpOXRpO = Vermilion @ 
RpOXrpO = Vermilion @ 
RpOXRpO = Vermilion ~ 


The preceding analysis shows that there should be three times 
as many red females as pink females. There are, in fact, some- 
what more than three times as many. The pink males should 
be to the red males (or to the vermilion) as 1 to 3. They do not 
come up to this ratio but nearly approximate to it. Similarly for 
the orange males. In this instance, where the numbers are large, 
it is quite apparent that the expected and the realized results 
fairly agree. The failure is here again obviously due to a reduc- 
tion in the number of the pink and orange classes. 


378 T. H. MORGAN 


Vermilion eye by orange eye 


When a female with vermilion eyes is bred to a male with orange 
eyes, all of the offspring are vermilion eyed. These inbred pro- 
duce the two grand-parental colors in males and in females. 


WermiltonOn mcr =. « 909 
ae J Vermilion ? \ J Vermilion ¢..... aire LL 
a = 
MEeeuIbOn Paya Or aneerc \ Vermilion / Orange mOr Geiser 131 
Orangesicie cece: 177 


The one point to notice here is the excess of orange males over 
orange females that has occurred in all of the preceding cases 
where both classes occur. The absence of red and of pink from 
the combination is due of course to the absence of pink in both 


parents. 
Vermilion 9 RpOX — RpOX 
Orange @ rpOX — rpOX 


Vermilion 9 RpOXrpOX 


a Vermilion co’ RpOXrpOX 


Vermilion @ RpOX — rpOX 


HOE Vermilion o RpOX — rpOX — RpO — rpO 


RpOXRpOX | 
RpOXrpOX } 8 vermilion 9 
rpOXRpOX | 

F. rpOXrpOX 1 orange @ 
RpOXRpO | 
RpOXrpO 3 vermilion 
rpOXRpO | 
rpOXrpO 1 orange of 


The analysis calls for 3 vermilion females to 1 orange female. 
In fact, 53 times as many vermilion as orange females are found 
and the same disproportion, due to deficiency in orange, is found 
also in the male classes. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 379 


The reciprocal cross, vermilion male by orange female, gives 
also all vermilion offspring. These inbred give both original 
classes in both sexes. 


{ Vermilion 9.....411 


a { Vermilion Q Vermilion <. 330 

7 by Ors =h is a 
Wermslionsc) by, Orange. ¢ \ Vermilion Orange ..50 
, Orange: ........ 62 


Here a slight excess of orange males over orange females occurs. 
The relation of vermilion to orange is seen in the following analy- 
sis. 


Orange 9 rpOX — rpOX 
Vermilion « RpOX — RpO 


P Vermilion 9 rpOXRpOX 
: Vermilion &@ rpOXRpO 


Vermilion 9 rpOX — RpOX 


Gametes of Fi Vermilion o& rpOX — RpOX — rpO — RpO 
rpOXrpOX. =1 orange @ 
rpOXRpOX 
RpOXrpOX ; = 3 vermilion 9 
RpOXRpOX 

Fy 
rpOXrpO =1 orange 
rpOXRpO | 
RpOXrpO }=3 vermilion < 
RpOXRpO J 


The expectation is for three vermilion females to one orange 
female: the actual numbers are 8 to 1. The same expectation 
holds for the male while the actual numbers give 53 to 1. 


380 T. H. MORGAN 


Pink eye by orange eye 


When a pink-eyed female is paired with an orange-eyed male, 
all of the offspring are pink. These inbred produce pink males 
and females and orange males. 


ee Pinks Oneee tee. 1035 
Pink 9 by Orange @ = eee °S| Bink he eee eee heb ID 
Caste \Orangeug's) seven ser se 518 


The two classes of males taken together give almost exactly the 
same number as the females. Both pink and orange males occur 
in equal numbers. In this experiment there appeared 15 orange 
females (not given in the table). This is possibly due to further 
mutation in the hybrid or to some error. The formulae are as 
follows: 


Pink @ rPOX — rPOX 
Orange & rpOX — rpO 


F Pink @ rPOXrpOX 
Pink @ rPOXrpO 


G sof F Pink @ rPOX — rpOX 
Sate Les souel Pink @ rPOX — rpO 


rPOXrPOX Pink 9 
rpOXrPOX Pink @ 


a) 


F, rPOXrpO Pink o 
rpOXrpO Orange ¢ 


The reciprocal cross, pink males by orange females, gives the 
criss-cross Inheritance, viz., pink females and orange males. It 
is of interest to note in passing that there were 233 pink females 
and 215 orange males recorded in this F; generation, showing that 
in the direct cross the sexes of opposite colors appear in nearly 
equal numbers. In the second generation both colors in males 
and females occur. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 381 


{ Pink 9 572 

: J Pinke ) Pink @. 292 
Pink & by Orange 9 = Oradeec. |LOrance.e 319 
| Orange o.. syarate telnet 


In the F, generation a peculiar relation becomes apparent: there 
are one half as many pink males as pink females, while in the 
orange class the females are only a little more than half as numer- 
ous asthe males. Thus while the total number of females of both 
classes (884) is nearly the same as the total number of the males 
when both classes are added together (814), yet this approximate 
equality is due to the reverse ratio of the sexes in the two color 
classes. It is worth noting, too, that this relation exists in the 
same group, in which as stated above, the pink females and 
orange males existed in equal numbers in the F; generation. In 
fact, it is just these two classes that still exist in equal numbers in 
the F, generation that give the significance to these results. 
The analysis of this case is as follows: 
Orange @ rpOX — rpOX 
Pink & rPOX — rpO 


Fr Pink @ rpOXrPOX 
Z Orange @ rpOXrpO 

: Pink 9 rpOX — rPOX 
Gametes of F; Orange @ rpOX — rpO 


rpOXrpOX Orange @ 


P rPOXrpOX Pink @ 
e ; rpOXrpO Orange 
rPOXrpO Pink @ 


At present I can offer no reasonable explanation of this peculiar 
relation between color and sex as shown in this experiment. It 
appears to be related to facts to be described later in connection 
with associative inheritance, and it is probably also related to a 
change in the sex-ratio that I have recorded in another cross 
(see Proe. Soe. Exp. Biol. and Med., 1911). As I have these 
problems still under investigation I shall not discuss them further 
here. 


382 T. H. MORGAN 


Discussion of results on eye-color 


An examination of the formulae used to interpret the preceding 
results will show three points of importance. First, that the red 
factor R may be present in the male-producing sperm. It is 
present there, in fact, if the fly has either red or vermilion eyes. 

Second, the factor for pink is only present when X or the sex 
factor is present. It is absent, therefore, from all male-producing 
sperm. It is true that X may exist without the pink factor, as in 
the vermilion and orange flies that owe their peculiarity to the ab- 
sence of P. If the P is contained in X, as its connection with 
sex establishes, then its absence must be due to its loss from X. 
Consequently while X may exist without P, the latter, P, can 
occur only when X is present. 

Third, the factor for orange, O, is present in every case. It 
might, therefore, be omitted from all of the formulae without 
affecting the results, provided the absence of R and of P be 
assumed to give O. But since orange is a definite color the 
absence of red and pink can not be assumed to leave orange. 
For this reason I have always inserted it. Its location can not 
be identified because it seems never to be lost. I shall give my 
reasons later for not identifying it with the color-producer C. 

The facts here recorded for the factor P amount in my opinion 
to a demonstration that this factor is intimately associated with 
the factor for sex. All of the 58 classes found in the second filial 
generation can be accounted for on the assumption that X contains 
P, when P is present; and, as I pointed out in connection with the 
heredity of white eyes versus red eyes, sex-limited inheritance 
ean be explained by assuming that X carries red if red is pres- 
ent. In the ease of vermilion and of orange eyes pink is lost from 
X and the formulae give the classes realized. The asymmetrical 
distribution of pink follows the same law as the asymmetrical dis- 
tribution of the sex chromosomes. 

A study of the formulae also reveals the fact that in the male of 
these classes (red and pink) when pink is present in the simplex 
condition (it can not occur otherwise in the male) no interchange 
takes place between the pink element contained in the X-chromo- 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 383 


some and any other chromosome, because, as I have previously 
pointed out, the sex chromosome in the male has no mate. Con- 
sequently no such interchange of chromatic particles, as we must 
assume possible for the other chromosomes, is here possible. The 
entire scheme of sex-limited inheritance rests, as I have tried to 
show, on this simple basis. To prevent a possible misunderstand- 
ing I may point out that the behaviour of the R-factor illustrates 
how an interchange of R and no Ris possible in the male. If the 
R is contained in some other chromosome in the heterozygote it 
may interehange position with its absent condition in a corre- 
sponding position or particle in the mate of this chromosome. 

The facts here recorded for the inheritance of pink make out a 
strong case in favor of the view that sex-limited inheritance 
can be explained if we locate the factor for pink in one of the sex 
chromosomes. I have pointed out that a similar assumption 
explains the heredity of white eyes, also sex-limited. I can state 
that the same assumption will account for the inheritance of yellow 
color and for the two wing mutations that are sex-limited. More 
important still is the fact that the extremely complicated results 
that follow when two or more of these sex-limited characters are 
combined must also be explained on the same principle. It is 
this evidence that has convinced me that segregation, the key 
note to all Mendelian phenomena, is to be found in the separation, 
during the maturation of the egg and sperm, of material bodies 
(chemical substances) contained in the chromosomes. 

This conclusion need not mean that the material bodies pres- 
ent in the chromosomes are the substances out of which the 
unit-characters are built up. On the contrary all that this evi- 
dence goes to show is that the bodies represent some material 
necessary for the development of the particular character in ques- 
tion, and it is certain that other parts of the cell also contribute 
to the elaboration of the unit-character. This is the view I 
should adopt, provisionally, as the more probable. We see, in 
fact, that the red color of the eye of the wild fly is due to the col- 
laboration of at least four different factors in the cell, namely, a 
red, a pink, and an orange determiner, and a color producer. 
The pink determiner and the color producer are carried by the 


384 T. H. MORGAN 


X-chromosome, but are not otherwise related. The red factor 
is contained in some other chromosome; the orange factor we 
can not yet locate. 

Concerning the chemical nature of the three colors I have no 
facts to offer. That they may be related chemically is made 
probable by the evidence, to be given later, that they are all three 
activated by the same color-producer C. If this is admitted we 
see that similar substances may be contained in different chromo- 
somes, and the further conclusion is then near at hand that in 
some cases the same substance may be carried by more than one 
chromosome. In connection with a mode of inheritance that is 
not as yet clearly Mendelian, viz., beaded and truncated wings 
I shall examine this assumption further, but it is not needed for 
the cases here described that follow Mendel’s law for one pair of 
factors. 

In the application of the Mendelian formulae to the F, genera- 
tion it has been only too often apparent that while the formulae 
give in all cases the expected classes, yet the numerical results 
depart widely from expectation. A consideration of the facts 
will bring conviction to anyone, I think, that the numerical depar- 
tures from expectation are due to special, disturbing factors. At 
another time, when I am able to present other data for wing inher- 
itance, and for disturbances in the sex ratios, I shall take up the 
question of these irregularities more fully. Here I can only 
point out one or two possibilities. In some cases the disturbance 
can be traced directly to the principle of ‘association.’ By this 
I mean that during segregation certain factors are more likely to 
remain together than to separate, not because of any attraction 
between them, but because they lie near together in the chromo- 
somes, as will be explained more fully later. For example, when 
red 2 RPOX is crossed with pink # rPO the offspring are red ¢ 
RPOXrPOX and red « RPoXrpO. As shown by the analysis 
on page 370 there are four classes of spermatozoa possible, but if 
R and P tend to hold together? rather than interchange with r 
there will be more female-producing sperm RPOX than rPOX, 


2 In reality C and P. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 385 


and hence proportionately more reds in relation to pink than 
random or Mendelian segregation demands. The same principle 
applied to other cases will often account for the disturbances in 
the Mendelian ratios. 

It is evident, however, that other conditions also may be respon- 
sible for the irregular ratios. The fact that the low types of 
mutants—those that have lost two factors, for example—fall 
short of expectation as compared with the normal type, and the 
disturbances in the sex ratios call perhaps for a different explana- 
tion. In regard to the former it is probable either that gametes 
containing certain combinations are less likely to fertilize or be 
fertilized or that the product of such fertilization is less viable. 
Until certain work is completed that I have on hand there is no 
need to attempt to decide which of the suppositions is the correct 
one. 


386 T. H. MORGAN 


PARTE ai 


THE RELATION OF THE COLOR PRODUCER C TO THE COLOR 
DETERMINERS OF EYE COLOR 


In a preceding paper (Science, July 1910) dealing with sex- 
limited inheritance of white eyes I have shown how the results 
are explicable on the assumption that the factor for red color 
is absent from the X-chromosome in the white-eyed individuals. 
It may appear that this assumption flatly contradicts the assump- 
tion made in the preceding cases for eye-color in which it is shown 
that the factor (a factor!) for red R is present in the male-produc- 
ing sperm (when red is present at all). There is no contradiction, 
however, for it is not the color determiner R that is present in 
the X-chromosome, but the color-producer C. It is the absence 
of C from X that gives the white-eyed fly whose formula is cRPO. 
For the sake of simplicity I have not introduced this relation in 
the preceding examples. The change, if introduced, gives pre- 
cisely the same results, but adds another letter. 

I purpose now to consider the relation of this C factor to the 
eye colors. It may make the case simpler if first an example 
for white and red is given. 

When a white-eyed male is crossed to a red female the offspring 
are red. These inbred give red females (50 per cent), red males 
(25 per cent), and white males (25 per cent). The formulae are 
as follows: 


Red @ CRX — CRX 
White o cRX — cR 


Red 9 CRX = cRX 


Pi Red 7 CRX — eR 
CRXCRX = Red 9 

2 CRXcRX = Red 9 

: GRKcR| = Red @ 
CRXeR) = White a 


The converse cross, white female by red male, gives red females 
and white males. These inbred give red males and females and 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 387 
white males and females. The formulae for this case are as 
follows: 


White 9 cRX — cRX 
Red o& CRX — cR 


Red 9 cRX — CRX 


F White @ cRX —cR 
eRXcRX = White 9 
F CRXcRX = Red 9 
ecRXcR = White @ 
CRXcR = Redo 


These two examples will serve to show the method by which all 
the white-red combinations can be treated. The results are those 
that I have already published (Science, 1910, vol. 32). 


Pink eye by white eye 


The results of this combination have been already given 
(Science, 1911). Since the numerical relations were peculiar 
I repeated the experiment and obtained large numbers of individ- 
uals that furnish a better basis for interpretation. When a 
pink-eyed female is bred to a white-eyed male all the offspring 
have red eyes. These inbred produce red-, white-, and pink- 
eyed offspring in the following proportions: 


Red-eyed females............ ro A oes 0 ce ENCI 1183 
Red-eyed males............ 

White-eyed males............. 
Pink-eyed females......... 
Pink-eyed males........... 


If we interpret these results in the same terms as those used 
for white and red we get the following formulae. Since the orange 
factor is present throughout and the orange eye is not involved 
O is omitted. The white-eyed male came from red stock through 
the loss of C and his formulae is cRP + cR. 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. I1, No. 4 


a4 


388 T. H. MORGAN 
Pink 9 (CrPX: — iCrPxX 
White «@ cRPX — cR 


ted 9 CrPXcRPX 
Red & CrPXcR 


Gatcciesoe an Red @ CrPX — crPX — cRPX — CRPX 
+ e : Red @ CrPX — CRPX — cR — cr 
CrPXCrPX = Pink @ 
CrPXerPX = Pink 9? 
CrPXcRPX = Red @ 
CrPXCRPX = Red @ 
CRPXCrPX = Red 9 
CRPXerPX = Red @ 
CREXcRPX | 3= Red @ 
CRPXCRPX = Red 9 
F, cRCrPX = Red & 
ceRerPX = White @ 
eReRPX = White @ 
cRCRPX = Red & 
erCrPX = Pink ¢ 
ercrPX = White @ 
ereRPX = Whites 
erCRPX = Red & 


The expectation is six red females to two pink females or 3 to 1. 
The realization is close to expectation, there being a small deficit 
in pink females. For the males the expectation is three reds to 
one pink. The realization is 4 to 1 due to deficit in pink males. 
The expectation for white males is the sum of the pink and red 
males: the realization is not far from this number. The white 
males should be to the red males as 4:3 which is approximately 
realized. 

The reciprocal cross, white-eyed females to pink-eyed males, 
gives in the first generation red-eyed females and white-eyed 
males. These inbred gave the following colors and ratios: 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 


Red-eyed 
Red-eyed 


White-eyed 
White-eyed 


Pink-eyed 
Pink-eyed 


389 


ORME MOC PETERS py Aloo. 900, soy ie Wc Vee AG 706 

CRE ne ae tar ERA erent 747 

ORM reel eRe eat Fr elon ae er) eae S04 

Fo enchants oy oe Re CT eae, A RPTL OP ear Ptr eS ae 832 

OREN al hats ON ee ee en 8 Oe ee ee 204 

Efe PR er: cere hs entree Ae Puree tated ls 210 
the same formulae to this case gives the 


The application of 
following results: 


F 


Gametes of F; 


Red 
White 


White. 9 cRPX — cRPX 
Pink @ CrPX — cr 

Red @ cRPXCrPX 

White o« cRPXcr 

@ eRPX — CrPX — 
o& cRPX — crPX — cr 
cRPXcRPX = White 2 
cRPXerPX = White 9? 
CrPXcRPX = Red @ 
CrPXerPX = Pink ? 
CRPXcRPX = Red @ 
CRPXerPX = Red 92 
erPXcRPX = White 9 
erPXerPX = White 9? 
ceRPXer White 7 
cRPXcR White @ 
CrPXer Pink @ 
CrPXcR Red & 
CRPXer Red ¢& 
CRPXcR Red ¢& 
erPXer White @ 
erPXceR White @ 


In both males and females the color ratio 
The actual numbers are a fair approximation to this expectation. 


CRPX — crPX 


is 4 white, 3 red, 1 pink. 


390 T. H. MORGAN 


PART III 


HEREDITY OF TWO SEX-LIMITED CHARACTERS COMBINED 
WITH FOUR EYE COLOR CHARACTERS.’ 


In this experiment a male with short proportionate wings and 
white eyes—both sex-limited characters—was mated toan orange- 
eyed female with long wings. The white-eyed male was a white 
from red stock, cRPO. 

In the first generation all of the offspring had long wings like 
the mother’s; the females had red eyes and the males had ver- 
milion eyes. These were inbred and produced the second or 
F, generation that contained flies having red, vermilion, pink, 
orange, and white eyes. In each of these classes, however, the 
short-winged individuals were males, as shown in the next table: 


Red @ long wings eS eee so pega roe eget ae Ou 

Red) ict long; wintss-ceessescre aster sets sn, VERE eee See 14 
Redo short. win scree ae eee PPR Ee ec seed 6 2 66 
Vermilion 92 long wings.. Sea SaTKe t Aaa MG hunts ate) 
Mermilion "67! Long) Wiles x. ses, eee eds cceeaneds verbal ore teats .. 204 
Vermilion & short wings Bc SME RIE See peonae il) 
Pink @ long wings oe ; =e .. 88 
Pink o& long wings ; 5 : Se oes 7 
Pink @ short wings : ; , Ci ee. 
Orange 2 long wings.. as Sethe Bh enc 119 
Orange @ long wings.... ba fase eS : : erigedi) 
Orange o& short wings........... By creates ; mols eal 
White & long wings eGo: Wathadee 95 
White o short wings tops Trea : 232 2202 


It will be observed at once that the inheritance of short wings 
and of white eyes is strictly sex limited. It will also be observed 
that each eye color has been combined with short wings, but 
only of course in the male sex. The total number of short winged 
males having red, vermilion, pink, and orange eyes is 92, while 


3 See also ‘“‘The Method of Inheritance of Two Sex-Limited Characters in the 
Same Animal.’’ Proce. Soc. Exp. Biol. and Med. vol. 8, Oct. 1910. 


CHROMOSOMES AND SEX-LIMITED - INHERITANCE 391 


the number of short winged males with white eyes is twice this 
number. In this connection I wish to point out that the grand- 
father had white eyes. The possible significance of this may be 
discussed later. It is obvious, however, on any theory of chance 
elimination- of unit characters in the egg that the total number 
of males of all eye colors having short wings must be equal to 
the number of short winged white eyed males. 


The analysis of the results following the same methods as here- 
tofore is as follows: 


Long-winged, orange-eyed 9 LCrpOX — LCrpOX 
Short-winged, white-eyed o IeRPOXIcRpO 


Long-winged, red-eyed 9 LCrpOXIlcRPOX ; Fr 
Long-winged, vermilion-eyed « LCrpOXIeRpO L 


LCrpOX LerpOX 
Py LOR pO , |LeRpOX 
3 |LCrPOX $ | LerPOX 
2 = |LCRPOX = |LeRPOX 
CERCHES I % )1CrpoX & | lerpOX 
&) |ICRpOX © |1ecRpOX 
= ICrPOX | lerPOX 
ICRPOX leRPOX 
é g 8 {LCrpOX E FI a) LCRpOX 
& §, © \lcRpO Bo ® \ lerpO 


The random fertilization of these sixteen kinds of eggs by the 
two kinds of female-producing spermatozoa LCrpOX and LCR- 
pOX and their fertilization by the two kinds of male producing 
spermatozoa IeRpO and lerpO is represented in the following 
table: 


392 


Egg 


1CrpOXLCrpOX 


ICRpOXLCrpOX 


ICrPOXLCrpOX 


ICRPOXLCrpOX 


LerpOXLCrpOX 
LerpOXLCrpOX 
LerpOXLCrpOX 


LeRPOXLCrpOX 


lerpOX LCrpOX 
leRpOXLCrpOX 
lerPOXLCrpOX 
IeRPOXLCrpOX 


Egg 
LCrpOX1CrpO 

LCRpOX1CrpO 
LCrPOXICrPO 
LCRPOXI1CrpO 


ICrpOX1CrpO 
ICRpOX1CrpO 
1CrPOX1CrpO 
ICRPOX1CrpO 


LerpOX1CrpO 
LeRpOX1CrpO 
LerPOX1CrpO 
LeRPOXI1CrpO 


ICrpOX1CrpO 
ICRpOXICrpO 
1CrPOX1CrpO 
ICRPOXICrpO 


Sperm 
LCrpOXLCrpOX 
LCRpOXLCrpOX 
LCrPOXLCrpOX 
LCRPOXLCrpOX 


Sperm no 


long orange 
long vermilion 
long pink 
long red 


long orange 
long vermilion 
long pink 
long red 


long orange 
long vermilion 
long pink 

long red 


long orange 
long vermilion 
long pink 
long red 


long orange 
long vermilion 
long pink 

long red 


short orange 
short vermilion 
short pink 
short red 


long orange 
long vermilion 
long pink 
long red 


short orange 
short vermilion 
short pink 
short red 


40 10 40 10 40 40 40 140 40 40 410 +40 


¥™™A™A 40 40 10 40 


AAqQyA 


WA A™® 


AAA™A 


and 


and 


and 


and 


and 


and 


and 


and 


» MORGAN 


Sperm X 


LCRpOX 
LCRpOX 
LCRpOX 
LCRpOX 


LCRpOX = 


LCRpOX 
LCRpOX 
LCRpOX 


LCRpOX 
LCRpOX 
LCRpOX 


LCRpOX = 


LCRpOX 
LCRpOX 
LCRpOX 
LCRpOX 


long vermilion 
long vermilion 
long red 
long red 


long vermilion 
long vermilion 
long red 
long red 


long vermilion 
long vermilion 
long red 
long red 


long vermilion 
long vermilion 
long red 
long red 


Sperm no X 


IeRpO 
leRpO 
IeRpO 
IeRpO 


leRpO 
IcRpO 
leRpO 
IecRpO 


IcRpO 
IeRpO 
IceRpO 
IeRpO 


leRpO 
IeRpO 
IecRpO 
leRpO 


long vermilion 
long vermilion 
long red 
long red 


short vermilion 
short vermilion 
short red 
short red 


long white 
long white 
long white 
long white 


short white 
short white 
short white 
short white 


AAAA QAAqaa 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 


WA A ®™ 


WAAA 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 393 


Summarizing this last table we get: 


Long winged red 9...12 Long winged red @ 4 
Long winged vermilion 9...12 and Long winged vermilion o7 4 
Long winged pink @... 4 Long winged pink @ 2 
Long winged orange 92 4 Long winged orange a2 
Long winged white ~ 4 

Short winged red 1 

Short winged vermilion = 4 

Short winged pink @ a2 

Short winged orange @ ees 

Short winged white o..... 1 


There should be, obviously, as many females as males. The 
results show 1012 females and 7428 males. There is a distinct 
falling off of males. The two sexes give about 10 to 8 or 5 to 3. 

Tn each class with eye color the males should be the same in 
number in the red and the vermilion; and in the pink and the 
orange. A great variability is however realized as shown below: 


RED VERMILION PINK ORANGE | WHITE 
One MwA eC ae sciep-r)- tyre ms 14 254 7 79 95 
Short winged @ 66 10 15 1 202 


The enormous discrepancy between theory and fact shown by the 
table may well make one reject the theory as totally inadequate 
to explain the facts. Nor can one appeal to the relative viability 
of the males to help him out of the dilemma; for, the long winged 
normal males run far behind the number for the vermilion, orange, 
and white; yet the long winged red eyed males are normal for the 
species, and do not run behind under the same conditions used in 
this experiment. A closer scrutiny of the table will, however, indi- 
cate a relation that may be very significant. The great excess of 
males is found in two classes, the long winged vermilion and the short 
winged white male, and these are respectively the father (the same 
as the grandmother’s combination) and the grandfather of this PF, 
generation! 


394 T. H. MORGAN 
PART IV 
THE INHERITANCE OF THREE SEX-LIMITED CHARACTERS 


In the following four crosses three sex-limited characters are 
involved: white eyes, short wings; and yellow color. In the first 
two crosses all these characters are contained in one of the par- 
ents; in the other two crosses two of the characters are in one 
parent and one in the other. 

In earlier papers (Science, 1910 and 1911) I have described 
the main facts for inheritance of red versus white eyes, and long 
versus short (short proportionate) wings. In order to understand 
the relation of yellow to normal color, the following experiment 
may be cited. A long, red, normal (body color) female was 
crossed with along, red, yellow male. The offspring, both male 
(654) and female (705), were long, red, normal. These inbred 
gave: 


Normal @ Se cH favs Sunken ray fereyativiasct toe OP MMSE SR ASHES AT Cue TRIN: SETS 525 
Normal o.. BRR e MTG paler ats iti ciaehe Be Se, CEE yee a TE 340 
Yellow ¢& Lee id RO CTO OIE Ceti Sree eae Oe beac 194 


The reciprocal cross, viz., long, red, yellow female by long, red, 
normal male, gave females (397) with long wings, red eyes, nor- 
mal color and males (282) with long wings, red eyes, and yellow 
color. These inbred gave: 


S07 117-1 BRS Pe One arr SS I dR SA oa Ln nk A mam ie, As a 0 346 
Wormiall Gus ste he eee irs So ee ee ee cya ce Oe Sk ee 259 
MEU ons: Botton ole RETRO eron Ca aah CAL A cua ee aA REI ERIN ie Ge ea 226 
VSN » ohh Se oi ae NtS cinch ERAT I TEE SPARRO Aa mre 230 


We are now in position to take up the experiments in which three 
sex-limited characters, white eyes, short wings, and yellow color 
are involved; see Plate I. 

When a yellow, short winged, white eyed male is bred to a 
normal wild fly with normal color, long wings, and red eyes, all 
the offspring are like the mother, i.e., normal color, long wings, 
red eyes. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 395 


aoe, ‘ _ JNRL 9? = Normal red long 9 
VWs cp by NRL e — (NRL o& = Normal red long # 


These inbred have produced four main classes of males, and 
three small classes, represented by a few male flies only; the females 
are represented by but a single class, as shown below: 


Normal. Color 


RED EYES | WHITE EYES 
Long wings | Short wings Long wings Short wings 
g | fo} | ) fof g of g a 
1879 | 606 | 167 1 3 
| | 
Yellow Color 
RED EYES WHITE EYES 
Long wings Short wiugs Long wings Short wings 
= = —- = 
Q | fol 9 (ofl 2 fol g fol 
7 1438 | 96 


i 


In addition to the flies recorded in the table there were two females 
belonging to two classes, viz., one, normal, long, white female; 
and one, yellow, long, white female. I shall not hesitate to ignore 
these two cases as exceptional, due either to accidental contam- 
ination through the food, or to sporting within the stock. Omit- 
ting these two females it will be seen that all of the females fall 
into one class having normal color, long wings, and red eyes. In 
all there were 1879 of these females, as against 1022 males. 
The females are, therefore, almost twice as numerous as the males. 
The three sex-limited characters appear only in the grandsons; 
one class containing all three sex-limited characters, yellow, 
short, white (96); one class containing two sex-limited characters, 
yellow, white (143); and one class containing one sex-limited 
character, short wings (167). For the moment the other three 
classes of males may be left out of account. It will be noticed 
also that the grandfather’s combination is well represented by 96 


396 T. H. MORGAN 


individuals, while the father’s combination which is the grand- 
mother’s also is represented by the great majority of all the males 
(606). The remaining large class contains two of the grandfather’s 
characters, viz., yellow and short. Of the three small classes of 
males, one contains two of the grandfather’s characters, viz., 
white and short, and one contains one of his characters, viz., 
yellow color. 

In my first attempt to analyze this case I ignored the three small 
classes of males because I found empirically that the remaining 
classes and the females could be very simply accounted for, as 
the following formule will show. 

Normal, red long @ NRLX — NRLX 
Yellow, white short o YWSX — — 


Normal red long @ NRLXYWSX 


uy Normal red long o NRLX —— 
NRLX — NRSX — YWLX — YWSX 
Gametes of F; NRLX — 
F, Generation 

NRLXNRLX = Normal red long 9 — NRLX = Normal red. long rou 
NRLXNRSX = Normal red long 9 — NRSX = Normal red short fou 
NRLXYWLX = Normal red long 9 — YWLX = Yellow white long fou 
NRLXYWSX = Normal red long 9? — YWSX = Yellow white short @ 


This scheme meets with two serious difficulties. It calls for 
equal numbers of each kind of male while in reality one class 
is at least three times as numerous as any one of the others. If 
we tried to explain this anomaly (as in fact I think we must) 
on the basis of some sort of “association”? taking place, we still 
have to meet a more serious theoretical difficulty. It will be 
seen that only four classes of eggs are represented in the F, gen- 
eration. There are two classes of eggs containing N and R, 
and two containing Y and W, but no class containing N and W, 
and none containing Y and R. No theoretical explanation can 
admit this arbitrary treatment, for the theory on which we are 
working demands the full interchange of all of these characters— 
unless some special reason can be given for failing in this regard. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 397 


Now it will be seen that the combinations of N and R, and Y and 
W are the combinations that existed in the parents of this cross. 
To admit the foregoing scheme requires the recognition of this 
union as permanent in subsequent generations. Yet this is op- 
posed to the Mendelian treatment of the case, unless association 
is admitted as valid. 

If now we take up the case of the three small classes of males 
we find no place for them in this scheme. I see no reason for 
ignoring them, small though the classes be. This consideration 
leads me to the conclusion that instead of four classes of eggs 
in the F, generation, the possibility of eight classes must be 
admitted, but owing to the initial association of N and R, and Y 
and W, their separation only occasionally occurs. When it does, 
the small classes of males appear, and the number of individuals 
in these classes is a measure of the infrequeney with which the 
separation occurs. The scheme when fully worked out is as 


follows: 
Normal female NRLX — NRLX 
yellow white short « YWSX — —— 


F Normal 9 NRLXYWSX 
: Normal o& NRLX — 


Gametes of F; 
NRLX — NRSX — NWLX — NWSX — YWSX — YWLX — YRSX — YRLX 


NRLX = —— 
NRLXNRLX = Normal red long 2 > — NRLX = Normal red long ou 
NRLXNRSX = Normal redlong 9 — NRSX = Normal red short of 
NRLXNWLX = Normal red long 9 — NWLX = Normal white long @& 
NRLXNWSX = Normal red long 9 — NWSX = Normal white short & 
NRLXYWSX = Normal red long 2 — YWSX = Yellow white short rot 
NRLXYWLX = Normal red long 2 — YWLX = Yellow white long of 
NRLXYRSX = Normalredlong 9 — YRSX = Yellow red short fot 


YRLX = Yellow red long of 


NRLXYRLX = Normal red long ? 


It is seen that the two factors N and W (or Re) tend to hold to- 
gether. Both are contained in the sex chromosome, i.e., N and 
e are there. Both are present in the grandmother, and through 


398 T. H. MORGAN 


her carried into her son—the father of the F, generation. The 
grandmother transmits only one X to her grandson—the one in 
question. It is also seen that the two factors Y and R tend to 
hold together in the same way. On the other hand the factor 
for long and short wings seems freer to leave one X and pass to 
its partner without showing any very great tendency to associate 
with the color factors in X. 


The reciprocal cross, viz., a female with yellow color, white 
eyes, and short wings, bred to a normal male with normal color, 
red eyes, and long wings, gave females with normal color, red 
eyes, and long wings, and males with yellow color, white eyes, 
and short wings. 


J NRL ¢ = normal red long Q 


SRS ees A | YWS o& = yellow white short ~ 


These inbred gave the classes in the next tables: 


Normal Color 


RED EYES WHITE EYES 
(=e ; a. Sa Se oT, 
Long wings Short wings Long wings | Short wings 
9 a | 9g o On Ollaac OF aaecr 
| 


439 319 | 208 193 1 5 11 


Yellow Color 


RED EYES WHITE EYES 
——_ = i 
Long wings Short wings Long wings Short wings 
= | | >| wh es 
Q Si ane st || 2 foil g fof 


Here again the two pairs of grandparental characters, viz., nor- 
mal color with red eyes; and yellow with white eyes, are repre- 
sented by the eight large classes in the F, generation; while short 
and long wings are nearly equally distributed. But even here 
there are more grandchildren with normal color, red eyes, and 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 399 


long wings, than with short wings. ‘These two colors went to- 
gether with long wings in the grandfather. Conversely, the 
grandmother combined short wings with yellow color and white 
eyes, and there is an excess of short winged grandchildren ( 2 
and ¢) over long winged (and ¢). The shorter analysis is as 
follows: 


Yellow, white, short @ YWSX — YWSX 
Normal, red, long o& NRLX — —— 


Normal @ YWSXNRLX 


Bi Yellow, white, short « YWSX —— 


5 YWSX — YWLX — NRLX — NRSX 
Gametes of F; SANDS = 


YWSXYWSX = Yellow white short 9 — YWSX_ Yellow white short 
YWSXYWLX = Yellow white long 9 — YWLX Yellow white long 
YWSXNRLX = Normal red long 9 — NRLX Normal red _ long 
YWSXNRSX = Normal red short 9 — NRSX Normal red — short 


EQ Qyiay a, 


If the more extended analysis for the gametes of the female 
were used there would be four more classes of eggs, namely, 
YRSX, YRLX, NWLX, NWSX, which would give four new 
classes of females, namely, yellow, red, short; yellow, red, long; 
normal, white, long; and normal, white, short; of which the 
second, third, and fourth are represented in the table by seven, 
one and four females respectively. The extended analysis would 
also give four other classes of males, whose formulae correspond 
to those of the four new types of eggs given above in the text, of 
which two are realized and two are not. 


In the third cross, a female with normal color, white eyes, and 
short wings was bred to a male with yellow color, red eyes, and 
long wings. The female offspring had normal color, red eyes, 
and long wings, and the male offspring had normal color, white 
eyes, and short wings. 


es : is {NRL @ = Normal red long 9? 
NWS 9 by YRL g' = \NWS @ Normal white short o 


400 Wie sie 


MORGAN 


The F,’s inbred gave the classes shown in the next table: 


Normal Color 
RED EYES WHITE EYES 
Long wings | Short wings Long wings Short’ wings 
g | rot g ou 2 fof | rofl 
439 | 7 235 218 237 359 | 387 
| 
Yellow Color 
RED EYES WHITE EYES 
Long wings Short wings Long wings Short wings 
ees = - “ ed: SS 
2 fot ie) fol g ou ol ofl 
345, 210 | 4 


in the second generation there are four kinds of females and six 
kinds of males. If we recognize the union of N and W and Y 
and R in the gametes of F, (the union that was present in the 
grandparents) the expectation on the shorter analysis is as follows: 
NWSX — NWSX 
YRL — — 


Normal, red long 9 NWSXYRLX 


Bi Normal, white short «@ NWSX —— 


YRLX — YRSX 


Gametes of F, NWLX — Rath 


NWSXNWLX = Normal, white long 2 — NWLX = Normal white long rot 
NWSXNWSX = Normal, white short 9 — NWSX = Normal white short ¢& 
NWSXYRLX = Normal red long 9 — YRLX = Yellow, red, long ree 
NWSXYRSX = Normal, red short 9 — YRSX = Yellow, red short oft 


If we admit the more extended segregation, the same dispropor- 
tions appear, but the two smaller classes of males are now repre- 
sented and two classes of males do not appear at all (as in the 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 401 


realization). In regard to the couplings in these cases it is of 
great importance to notice that once more the long and short 
factors segregate without regard to the color factors, yet even 
here a remarkable fact comes to light. The two classes of males 
that exceed the others are those in which the long, red, yellow 
combination and the short, white, normal combination exist. 
These are the two combinations that were in the grandparents. 
If we assume that they more often remain in the same chromo- 
some the numerical results become apparent. 


In the fourth cross, a male with normal color, white eyes, and 
short wings was bred to a female with yellow color, red eyes, and 
long wings. The female offspring had normal color, red eyes, 
and long wings, and the males had yellow color, red eyes, and 
long wings. 


=) NRL @ = Normal red long 


INNIS) eh 3 OS YRL & = Yellow red long 


The second or F, generation is represented in the next table: 


Normal Color 


RED EYES WHITE EYES 
Long wings Short wings Long wings Short wings 
= = | = = 
g fof S| GE i ee isi g | fos! 
608 2 137 | i oho37 


Yellow Color 


RED EYES WHITE EYES 
Long wings | Short wings Long wings Short wings 
2 fof Q cof gl X2) of g of 
389 | 248 101 1 1 


Only four large classes of males are represented and it is signifi- 
cant that these are the yellow-red and the normal-white—the 
two combinations that correspond to the two grandparental com- 


402 T. H. MORGAN 


binations. These two classes oceur both in the long and short 
wings, but here again occurs the significant fact that the yellow, 
red, long are twice as frequent as the yellow, red, short. The 
former is the grandmaternal combination. Again the normal, 
white, short are nearly twice as numerous as the normal, white, 
long and it is the former combination that is characteristic of 
the grandfather. The shorter analysis follows: 


YRLX—YRLX 
NWsSxX — —— 


F Normal red long 9 YRLXNWSX 
: Yellow red long @ YRLX — 


YRLX — YRSX — NWLX — NWSX 


Gametes of F, En 


YRLXYRLX = Yellow, red,long 9 — YRLX = Yellow,red,long . & 
YRLXYRSX . = Yellow, red, long @ — YRSX = Yellow, red, short fou 
YRLXNWLX = Normal red, long 9 — NWLX = Normal, red, long ol 
YRLXNWSX = Normal red, long @ — NWSX = Normal, red, short fe 


The same objections may be urged against this scheme that have 
been given for the first short scheme. It is unnecessary to write 
out the longer scheme again in this case as the same principle 
employed in the first instance is applicable here. Of the four 
additional classes of males called for by the longer analysis, 
three appear in the results represented by any 2, 1, and 1 males 
respectively. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 403 


PART V 
CONCLUSIONS 
A THEORY TO ACCOUNT FOR ‘ ASSOCIATIVE” INHERITANCE 


In the preceding pages I have tried to show how the mechanism 
that exists in the chromosomes ean be applied to the mechanism of 
heredity, provided we deal with particles or chemical substances 
in the chromosomes rather than with the chromosomes as units. 
The evidence makes out, I believe, a very strong case in favor of 
the idea that sex-limited inheritance is.connected with the same 
physical body that determines sex, and I have not hesitated to 
identify that body with the sex chromosome. The second point 
of significance in the results is that while in the female there may 
be an interchange between homologous chromosomes, no inter- 
change takes place in the male of those factors connected with 
sex-limited inheritance. We can explain this result if these char- 
acters are contained in the single X chromosome in the male 
which alone has no mate. The third point of interest in these 
results is the necessity of assuming some combination or rather 
localization amongst some of the substances resident in the same 
chromosome. The peculiar ratios found in the second genera- 
tion find their explanation only by means of such an assumption. 
Couplings and linkages have been described before to account 
for observed ratios, notably by Bateson and his collaborators, but 
I think we see here clearly for the first time that these unions are 
not due to inherent relations, or fusions, or attractions, or corre- 
lations, or repulsions, but to juxtaposition of particles in the chro- 
mosomes. It has been shown in a considerable number of cases 
that at one stage in the process of union of homologous chromo- 
somes the members of each pair twist around each other like 
the components of a rope. Subsequently these twisted chro- 
mosomes fuse together and shorten. Later a longitudinal split 
appears in the shortened, double chromosome. This split now 
lies in one plane, i.e., it does not follow the turns of the united 
chromosomes. In consequence of the position of the new plane 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 4 


404 ; T. H. MORGAN 


of splitting or division each half, or new chromosome, must be 
made up of parts of one and parts of the other of the two original 
chromosomes that united in pairs as Janssens has shown. As a 
result of the subsequent ‘‘reduction”’ division the cells that are 
produced will contain new combinations of the materials com- 
posing the original chromosomes. If the chromosomal materials 
that represent the factors of heredity are placed lineally along 
the chromosome and in corresponding linear series in each pair 
of homologous chromosomes, random separation of these mate- 
rials will be brought about by means of the cell mechanism, 
explained above, except in those cases where the materials lie 
near together. In the former case, the usual Mendelian random. 
segregation will take place; in the latter case, groups of factors 
will tend to remain together or be assocfated in heredity. These 
latter cases correspond to those in which we find ‘‘association”’ 
to occur. It will be observed that while such associations will 
be more or less common according to the nearness of the associat- 
ing factors in the chromosome, the associations are not absolute 
for occasionally the twisting of the chromosomes will be such 
that even regions lying lineally near together will come to le 
on opposite sides of the united chromosomes. These cases repre- 
sent the small classes observed in the tabtes. In the case of the 
X-chromosomes we should expect interchanges of the postulated 
kind to occur when two X’s are present, as in the female of Droso- 
phila; but no interchanges when only one X is present as in the 
male. The experimental results. accord completely with this 
‘anticipation and afford strong evidence in favor of the view ex- 
pressed above. This assumption seems to be far simpler than 
the assumption of attractions, repulsions and complicated ratios 
that Bateson has suggested as an interpretation of similar phe- 
nomena. 

it is obviously not essential to this hypothesis of factoral inter- 
change to limit it to the particular stage of the chiasma type here 
suggested, for if a similar phenomenon occurs at any other stage 
in division the same results will follow. 


. 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 405 


DISINTEGRATION OF A SPECIES AND ITS RECONSTITUTION BY 
RECOMBINATION 


The series of mutations that have appeared in Drosophila 
can all be accounted for on the assumption of losses from the orig- 
inal germ-plasm. There is possibly one exception to this rule, 
namely, the melanitic or black mutant that may appear to add 
something to the original color. When crossed to the normal it 
gives an intermediate type in the first generation, and this fact 
also might be urged in favor of the change being in a positive 
direction. But since in the second generation the black fly 
appears as the recessive type it is not improbable that even this 
mutant may be due to loss. The most convincing proof that 
the eye and wing mutations are due to loss is found in the recon- 
stitution of the original or wild type when certain recombinations 
aré made. So many examples of this have been given in the pre- 
ceding pages that I need not go over the evidence again. Two 
points, however, may be recalled as instructive. In several com- 
binations the female alone is reconstituted. In the older term- 
inology the female is the atavist, the male is the neomorph; 
and it is always the female and never the male that shows this 
relation if either one is atavistic. The reason for this is also 
clear from the evidence—the male-producing sperm has all 


- the recessive facters in question, while the female-producing 


sperm has one ol more dominant factors. No better example 
than this one of unisexual atavism, or reconstitution, could be 
cited to show the advantage that the modern explanation of 
heredity has over the older view, where a fact of this kind would 
have seemed totally inexplicable. 

The second point of interest in this evidence is found in the 
different behavior in heredity of the original and of the reconsti- 
tuted (atavistic) type, for while the wild type-continues to breed 
true the reconstituted type splits later into its components. The 
reconstituted type may be said to be physiologically complete, 
but morphologically dismembered. For example, when a ver- 
milion male is bred to a pink female, all of the offspring are red. 
The four substances, R P O C, necessary to produce red have 


406 T. H. MORGAN 


been brought together, so that all the elements of the wild fly 
are present that collectively give red eye-color; but now instead 
of the chromosomes that carry these substances being represented 
in duplicate, some of the chromosomes lack one or the other 
substance. It is due to this that the splitting takes place in the 
formation of the germ-cell of these reconstituted types. Never- 
theless it is possible even to reconstruct the original wild forms 
by suitable combinations, such, for example, as will give in duplex 
the four factors essential to the development of red eyes, and the 
structurally reconstituted types will breed as true to type as the 
wild flies, in contrast to the atavists that are only physiologically 
reconstituted. 


THE PRESENCE AND ABSENCE THEORY IN RELATION TO THE 
. THEORY OF ASSOCIATION 


The appearance of so many mutants, due to losses, raises the 
question as to whether all new types that follow Mendel’s law may 
come under the same category. There are types, it is true, among 
domesticated forms that appear to have added something to the 
original type from which the mutants arose, but some of these 
are due to hybridization, and some may be due to losses of jnhib- 
iting factors, whose absence permits the further elaboration of 
characters already present. As yet the evidence is insufficient, 
I think, to allow any certain generalization in this regard, yet 
the evidence suffices at least to show that many or even most 
cases that follow Mendel’s law fall under this head. In so far 
as the mutants are due to losses they are explicable on the pres- 
ence and absence theory, which may seem to give some grounds 
for the universal application of this principle. There is, however, 
one possibility that we can not afford at present toignore, namely, 
the evidence of ‘‘association”’ which is clearly furnished by some 
of the crosses described in the preceding pages. I sheuld like 
to dwell a little further on this point. If I am right in explaining 
the results of those cases where two or three sex-limited charac- 
ters are involved on the grounds of the juxtaposition of substances 
(factors) in the chromosomes, it 1s only a step further to cases, 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 407 


like those of the gray mouse when the color factors for gray, 
namely, black, yellow, chocolate, and ticking, remain permanently 
associated when crossed with a mutant which contains only one 
of these same factors. Thus when a yellow bearing germ cell 
meets one bearing gray, all of the offspring are yellow. These 
yellows inbred produce only grays and yellows and not blacks and 
chocolates also, as should happen, did each of the elements in 
the original gray have for its mate the absence of its particular 
factor.t. In Drosophila the eye color is made up of three or four 
color factors. When a wild female is crossed to one of the mutants 
an orange eyed male for example, all of the offspring are red. 
In the F, generation not only red and orange, but pink and bright 
red also appear, although even here there is a stronger tendency 
for the original red to reappear more often than its products. It 
seems to me probable at least that the difference between the 
mice and the fly in this respect may be due to the closer associa- 
tion of the factors in the mice than in the fly. If so, the differ- 
ence is one of degree only and not of kind. 


THE FERTILITY OF DEFICIENT MUTATIONS 


A striking falet in regard to most if not all of these mutations in 
Drosophila is their infertility compared with the original stock 
kept under identical conditions. As I have this matter under 
investigation I wish here to touch on it very briefly, and only in 
so far as it bears on the numerical proportions of the different 
types. The pure stock of several of the new types is less produc- 
tive than the original stock, and the failure in several cases of 
the deficient types to appear in the expected ratios suggests that 
this failure is due in part to the failure in fertility or in vitality 
of the new types. Whether this is due to failure in the develop- 
ment of the egg, or of the sperm, or to failure to fertilize, or to lack 
of development of the embryo, are points requiring special study, 
but the facts are sufficiently numerous to raise the question as 
to whether a type that lacks some material present in the original 


4See Morgan, T. H. The influence of the environment and of heredity on the 
inheritance of coat color in mice. New York Academy of Science. 1911 


408 T. H. MORGAN 


stock may not in many cases lose also its full power of productiv- 
ity. It may seem improbable that the presence of some substance 
necessary for the development of a particular color in the eve 
could have any influenceon the rest of the germ, but the same sub- 
stance that is essential for eye color may be essential for the pro- 
duction of other things in the body that are not so apparent. In 
fact, I have already obtained evidence in the case of the eye 
color-producer, C, that the absence of C not only affects the eye- 
color, but other parts in the body as well. It is not impossible, 
therefore, that in certain cases the absence of a factor may have 
an important influence in one or another way on the productivity 
of the animal. I do not wish to discuss further this question until 
I can bring forward certain evidence that bears directly on it, 
but I have raised the question here, first in order to point out its 
possible bearing on the disturbance of Mendelian ratios, and 
second, in order to make clear that while I regard the evi- 
dence in favor of the mosaic inheritance of certain characters as 
established, I am not unappreciative of the fact that a simple 
factor may have a wider influence in development than appears 
when only a single character is under consideration. 


ORIGIN OF MUTATIONS THROUGH CHROMATIN LOSSES 


The mutations that have occurred in Drosophila may throw 
some light on the origin of mutations in general. If my analysis 
is correct it follows that mutations arise not through losses of 
whole chromosomes as some cytologists have hinted (for the 
individual chromosomes must be supposed to earry not one but a 
host of factors) northrough the doubling of one or of all of the chro- 
mosomes (which would only add to what is already present in 
duplex) but mutations arise through the regrouping of partic- 
ular substances carried by the chromosomes. These substances 
may be so small in amount that their absence may entirely escape 
a cytological examination. It would seem that ample opportunity 
must be present for losses of this kind, since any irregularity in 
the division of the chromatin in the germ tract would lead to 
the appearance in time of a mutant if such were viable and if the 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 409 


right combinations were brought about. If, for instance, in the 
division of the spermatogonial cells the material particles at any 
level should fail to divide when the rest of the chromosome divides, 
one of the resulting cells will be deficient in the substance in ques- 
tion, and its offspring will be correspondingly deficient. Or if 
after synapsis similar particles in homologous chromosomes should 
pass into one chromosome, instead of segregating, one of the result- 
ing cells will be deficient. The wonder is that such losses are so 
infrequent. 


THE BEARING OF THE RESULTS ON THE CONSTITUTION OF THE 
SEX CHROMOSOMES 


The experiments on Drosophila have shown that a most compli- 
cated series of facts relating to sex-limited inheritance can he 
accounted for, as I pointed out in my paper of 1910, on the assump- 
tion that one of the factors for such characters is combined with 
the sex factors, X, or more specifically (Morgan,® Wilson 1911) 
if it is contained in the accessory or sex chromosome. The ab- 
sence of this chromosome in half of the spermatozoa, and the 
impossibility of an interchange between this simple X-chromo- 
some in the male (since X has no pair in synapsis) is the signi- 
ficant feature of the explanation. What is most important, is 
the discovery that the X-chromosome contains not only one of the 
essential factors in sex determination, but also all other characters 
that are sex-limited in inheritance. 

The discovery of this relation leads us a step farther, I think, 
in the analysis of the problem of sex determination, for it 
shows that the determination of sex is only one of several (per- 
haps of a large number of) properties contained in the sex chro- 
mosomes. Only by the loss of a factor from one of the other X’s 
(in the female) is it possible to aiscover just how many factors 
are containedin X. Already four or five such losses have appeared 
in my mutations. This leads at once to the inference that it is not 
the X-chromosomes, as such, that is a factor in sea determination, 
bul only a very small part of its material. 


5 In a paper read before the American Society of Naturalists, December 29, 
1910. Published in American Naturalist, 1911. 


410 T. H. MORGAN 


I suggested in 1905 that the female sex is determined by the 
presence of more chromatin in the fertilized egg. Wilson sug- 
gested in 1906 that it is a particular chromosome that gives the 
quantitative results (or at least a more or less ‘active’ chromo- 
some). The results of the experiments dealing with sex-limited 
inheritance in Drosophila demand that we go one step further, for 
they show that it is only a small part of this chromosome that 
is involved in sex determination. 

If this is admitted we can understand how sex may be regulated 
in the same way, even when X and its mate Y appear to our rela- 
tively gross methods of measurement to be equal. The differ- 
ence in size between X and Y that gives a completely graded 
series in different species has little to do, therefore, with condi- 
tions relating to sex determination, except in so far as the initial 
loss of the sex substance contained in Y led to a decrease in size. 
1 am inclined to think that the difference in size relation between 
X and Y represents largely the loss from Y of those materials 
that play a role in sex-limited inheritance. Jf X 7s the sex chro- 
mosome, then Y is the sex-limited chromosome in a double sense. 
Its final disappearance in certain forms represents the total loss 
of all characters that can become sex-limited in inheritance. 

If it is legitimate to draw any inference from the analogy be- 
tween sex determining factors, and factors that determine other 
characters of the organism, it follows with a fair degree of plaus- 
ibility that Y lost a sex factor (contained in the three X’s of the 
species, 2 in the female, 1 in the male) in the same way that it 
has lost other factors also. If these factors are, as I have sug- 
gested, on a par, it follows that the material in question is the 
female determining factor, F. Where then is the male determin- 
ing factor? I have given my reason recently for dissenting from 
the position taken by several writers that the male condition is 
simply less X chromatin. It seems to me that if we treat the 
problem of sex determination by the same methods used for Men- 
delian characters in general, we can not Justify such a position 
but are led inevitably to the conclusion that if the X-chromosome 
contains (not is) the factor for producing a female, the factors 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 411 


for producing the male must be located in some other chromosome.® 
This interpretation I have developed briefly in a recent article’ 
in which the female factor supposed to be contained in the X- 
chromosome is represented by F and the male factor, supposed 
to be contained in some other chromosome (not in Y however 
which is ranked with the three X’s except in so far as certain fac- 
tors have been lost), is represented by M. The following scheme 
shows how the relation of the sexes on this basis and how sex is 
determined: 


Gametes of female FM-FM a XM-XM 
Gametes of male FM-M XM-M 

FP female FMFM ; XMXM 
, male FMM 2 XMM 


56 Not in X because in males having only one X (no Y) the scheme will not work 
out. 

7The application of the conception of pure lines to sex limited inheritance and 
to sexual dimorphism. The American Naturalist, Feb., 1911. 


‘ PLATE 1 
EXPLANATION OF FIGURES 


Normal red-eyed fly. 
Yellow fly with white eyes and “short proportionate’”’ (or miniature) wings. 
Short winged fly (as in fig. 2) with normal color and white eyes. 

4 Brown fly with red eyes and long wings. This fly is yellowish, but differs 
from fig. 2 in the absence of a black factor (absent also in fig. 2) as well as a 
second factor probably a yellow factor. Its formula is Br by. In the repro- 
duction the bands and tip of the abdomen are too black. They should be more 
like fig. 2 and the yellow should be more saffron. This fly was not used in these 
experiments. 


Wwe 


412 


CHROMOSOMES AND SEX-LIMITED INHERITANCE 
T. H. MORGAN PLATE 1 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. Il, NO. 4. FE. M. Wallace and M. B, Abbott, Del. 


413 


ON THE STRUCTURE, PHYSIOLOGY AND USE OF 
PHOTOGENIC ORGANS, WITH SPECIAL REF- 
ERENCE TO THE LAMPYRIDAE 


E. J. LUND 
Bruce Fellow, Johns Hopkins University 


NINE FIGURES 


Introduction........ ; Pipe iS 2 ths SRO eR 415 
Material. 22... a ne tea c ey teen tore ears , 416 
Methods......... ; 417 
Structure of the wiaatior ante organs 08: the Ls cpemidae oo 418 
Physiology.......... 5 oe Sh SoS Se 431 
A. Effects of ehemcals! ER a2 bn AEE SRNR CEM Ee Sead ores 437 
B. Observations on the ioe alization oF the photogenic process in the organs. 439 
C. Effects of rapid changes in air pressure, upon the organs......... .. 442 
DR (Controlieey sa. nieiiaeiice : aerate: Meta Booopoes Ze5 
E. Effects of temperature......... ae : Ae ro ae . 448 
. Upon the photogenic process. bu Serta ese ete ... 448 

b. Upon the photogenic Beeretion! of Cyoridina squamosa (?) and 
Cyclopina gracilis..... es AS eno a: 449 
ce: Uponiithetosmic acid reaction............--.:0.+: Asie 451 
Use of the organs......... Be dea ene oers Lei ased aoe 455 
SUMMIM Aiye eerste hres me see sett ae sara ts as Pas EPPA eR rive to eit oT Cee 457 


INTRODUCTION 


The subject matter of this investigation is presented in some 
detail, partly because the forms of Lampyridae here studied 
have not been studied before and partly because variety of mate- 
rial has been available. I have attempted to bring together the 
main results of other workers on the subject with my own in 
order that some of the points in the anatomy, physiology and use 
of these organs of light production may become more intelligible. 

The literature on the subject of animal and plant photogeny 
that has accumulated is vast. Yet the advance in our knowledge 
of the processes, structure and use of the photogenic tissues lies 
mostly within the last forty years. Often the observations by 
different authors vary and are contradictory upon points that 

415 


416 E. J. LUND 


would seem simple, and as a resuit a great number of notes and 
references are found in running over the literature that are little 
more than confusing and have added nothing to our knowledge 
of the subject. A list of the most important papers bearing 
directly upon the subject matter and others to which reference 
will be made has been appended at the end of this article. 

I am. indebted to Professor E. A. Andrews for his kindly inter- 
est in the work and for reading the manuscript, to Professor H. F. 
Nachtrieb of the University of Minnesota, and to Professor 
C. M. Child of the University of Chicago for generously placing 
laboratory facilities at my disposal during the summer seasons 
of 1909 and 1911. 

I shall use the term photogeny which has come into more gen- 
eral use of late, to designate the processes concerned in the pro- 
duction of light by living organisms, without rejecting nor accept- 
ing the term phosphorescence as a proper one; first, because the 
latter has come to stand for apparently different phenomena 
occurring under varying conditions, and second, in view of the 
fact that our former conception of the processes of oxidation has 
undergone change in recent times and is continuing to do so. 
Le Bon has employed the general term phosphorescence in making 
a perhaps serviceable division of the phenomena: (1) Phosphor- 
escence generated by light; (2) Phosphorescence independent of 
light and determined by different physical excitants, such as 
heat, friction, electricity and the X-rays; (3) Phosphorescence 
by chemical reaction; (4) Invisible phosphorescence. Accord- 
ing to this division our problem can with certainty be placed 
under the third head, for there is abundant proof that the pro- 
duction of light in animal and plant tissues is a result of some 
kind of change in the chemical constitution of the substances 
concerned in the process. 


MATERIAL 


The material was obtained during the summers of 1908-11. 
The early part of the work was done upon material collected in 
the vicinity of St. Paul, Minnesota. Other material was col- 
lected at various points in Jamaica throughout the season, 1910, 


PHOTOGENIC ORGANS 417 


during the summer session of the Johns Hopkins University 
Marine Zoological Laboratory located at Montego Bay, while 
material used in 1911 was obtained in the vicinity of Chicago. 
The time for the occurrence of the different species of Lampy- 
ridae that were studied is about the same wherever they were 
observed, lasting from about the last part of June to about the 
first part of August in Minnesota. The period of their occur- 
rence in Jamaica is considerably longer though varying with the 
altitude and various climatic conditions. The following identi- 
fied species of luminous Lampyridae have been studied especially 
as regards the anatomy of the photogenic organs. 


Minnesota specics 


Pyropyga indicta Lec. 

Lecontia (Pyractomena) lucifera Melsh. 

Photuris pennsylvanicus DeGeer. 

Photinus ardens Lec.—two varieties. 
Jamaica species 


Photinus maritimus. 
Photinus commissus E, Oliv. 
Photinus pallens Fabr. 
Photinus pantoni E, Oliv. 
Photinus ebriosus E, Oliv. 
Photinus suavis H, Oliv. 
Photuris jamaicencis E, Oliv. 
Several other forms of Photinus sp. (?). 
Studies on the physiology have been mostly upon Photuris pennsylvanicus, 
Photinus ardens, P. maritimus, P. ebriosus. P. pallens. 


METHODS 


The varying methods of preparation used by different authors, 
no doubt in part account for some of the minor differences in 
their descriptions of the structure of the organs. Most of the 
papers deal with the European species Luciola italica, Lampy- 
ris splendidula and L. noctiluea. I have not had the opportunity 
to get material of these species. However, the descriptions show 
that no essential histological differences exist between the Kuro- 
pean, American and Jamaican forms so far studied, though there 
may be lesser differences, consisting mainly in the size and posi- 
tion of the organs on the abdominal segments and the grosser 
relations of the tracheal system to the layers of the photogenic 


418 E. J. LUND 


organ. In killing and fixing, the abdomen was cut off before 
immersing in the fixing fluid. All of the more common killing and 
fixing agents were tried. Among these formalin 4 to 10 per cent, 
alcohol 80 per cent, Flemming’s fluid, osmie acid 0.5 per cent 
aqueous solution, and water at the boiling point, were the most 
useful, depending upon what part or structure the preparation was 
intended to bring out. Other methods employed will be described 
later on. All sections were stained on the slide or cover glass. A 
large number of stains were tried and as a result the following were 
found most useful: Thionin aqueous solution, Borax carmine 
and Lyons blue, and Ranviers picro carmine as general stains. 
The best of all was found to be a solution made up as follows: 
anilin blue 0.5 gram, orange G. 2.0 grams, oxalic acid 2.5 grams, 
water 100 ec. Instead of anilin blue Lyon’s blue may be used, 
though the preparations obtained with the latter were much 
inferior. Photogenic tissue fixed in boiling water for one minute 
afforded beautiful preparations. The photogenic cells of Odon- 
tosyllis pachydonta could be differentiated from among the other 
epidermal cells. Granules of the photogenic cells; fat globules 
in the fat cells, and chitin stain orange-yellow (and yellow) 
respectively; cytoplasm light blue, cell membranes where present 
dark blue, granules of the dorsal layer (urate granules of some 
authors) stain blue. To show relations of the tracheal capil- 
laries and tracheal end cells a 0.5 per cent aqueous solution of 
osmic acid was used, sections cut 1 to 15 microns thick and 
mounted to advantage without staining. By means of dissec- 
tion of the active organs and examination under the binocular 
or compound microscope in a dark room facts fundamental to 
an understanding of the photogenic process could be made out. 


STRUCTURE 


The following description of the structure of the photogenic 
organs applies to all the species of Lampyridae examined except 
when otherwise stated. The location of the organs is limited 
to the sternal plate of the fifth and sixth or part of either the fifth 
or sixth abdominal segment. Jn all the forms studied the chitin 


PHOTOGENIC ORGANS 419 


of the sternae opposite the photogenic organs is transparent and 
covered with hairs as are other parts of the abdominal ring. — Lin- 
ing the inside of the chitinous integument is the thin hypodermis, 
represented by a single layer of flattened cells with flattened 
nuclei (figs. 7and 8 H). No evidence of proliferation of the hypo- 
dermal cells and their subsequent differentiation into cells of 
the photogenic layer has been obtained from any of the prep- 
arations of the adult forms studied. DuBois has studied the 
development of the photogenic organs in Lampyris noctiluca 
and Pyrophorus noctilucus and states that the tissues are derived 
from proliferating cells of the hypodermis. If this observation 
be correct. it would preclude the photogenic cells from having a 
common origin with the cells of the fat-body as some authors 
have supposed. The fact that the granules of the cells of the 
photogenic layer stain in several respects like those of the globules 
in the cells of the fat body, affords by itself no evidence of an 
ontogenetic relationship of the fat-body and. the photogenic 
cells. In all the species the organ consists of the two usual layers, 
the dorsal or ‘‘urate cell layer’? of some authors and the lower or 
photogenic layer. The photogenic layer is completely enclosed 
by the dorsal layer, except where the former is applied to the 
hypodermis over the whole or part of the sternite depending upon 
the size of the organ. I have found no trace of any membrane 
covering the inner side of the organ as Wielowiejski (’82) states 
that he found in Lampyris splendidula. Townshend (’04) finds 
no membrane present in the American form, Photinus margin- 
ellus, and no such structure has been found in the other European 
species. 

In consideration of the theories advanced by Heineman (’86) 
and DuBois (95) to account for the control of the organs by 
means of the segmental muscles in producing increased air pres- 
sure or causing a flow of blood through the organ we should per- 
haps expect to find a greater development of the muscles related 
to them. This however is not the case. The muscular devel- 
opment in the segments bearing the photogenic organs is not more 
extended than in any of the segments anterior to them. The 
muscular apparatus of each of the fifth and sixth segments con- 


420 E. J. LUND 


sists of three main sets of muscles, one dorsi-ventral, and two 
antero-posterior sets one dorsal and one ventral (fig. 1, a, b, c). 
The dorsi-ventral set is composed of two bundles one on each side 
in each segment. These are inserted dorsally in the tergum and 
penetrating the tissue of the photogenic organ are inserted on 
the ventral side in the chitin of the sternite, appearing onthe 
exterior as a visible spot or indentation which is not luminous. 
These muscle fascicles are not in contact with the photogenic layer 


Fig. 1 Camera lucida outline of abdomen to show relations of muscles and 
photogenic organ. A, longitudinal section; B, cross section; a, longitudinal tergal 
muscles; b, longitudinal sternal muscles; c, vertical muscles or ‘vertical expira- 
tors;’ d, dorsal layer; e, photogenic layer; f, vertical tracheae. 


but separated from it by more or less of a continuation of the 
dorsal layer. The antero-posterior muscles are shown in fig. 1. 
Their insertions are on the tergal and sternal membranes. They 
serve to shorten the abdomen by telescoping the abdominal rings. 
Their position and arrangement has no definite relation to any 
single trachea or group of tracheal trunks. The tissue of the 
photogenic organ does not adhere tightly to the chitinous integ- 


PHOTOGENIC ORGANS 421 


ument, e. g., P. pennsylvanica and P. pallens. It is of a pasty 
consistency yet firm enough so that it may readily be separated 
from the integument intact by proper manipulation with a small, 
soft brush. In this way it may be studied with readiness under 
a low or high power, treated with reagents, ete. A more direct 
effect of the reagent can be observed in this way, than by simply 
immersing the whole animal or abdomen. The two layers of 
the organs vary in thickness within certain limits in different 
species, but no marked seasonal variation of either layer has 
been found. However, relative changes in the contents of the 
cells of the dorsal and photogenic layers are plainly evident (p.4°2). 
Wielowiejski (’89) came to the conclusion that cells of the layers 
in L. splendidula were distinct. Bongardt (03) found no photo- 
genic cells in the process of transformation into cells of the dor- 
sal layer in the same form. Du Bois (’95) states that the cells 
of the photogenic layer in L. noctiluca are transformed directly 
into the erystalline deposit which he states makes up the dorsal 
layer. He figures no cellular structure in the dorsal layer. Bon- 
gardt (’03) also found cells in L. noctilueca which combined some 
of the characters of both the photogenic and the dorsal layer 
cells, and says that they may in so far be considered ‘Ubergang- 
zellen.’ The figures of both these authors are too inadequate 
to enable me to make out what all the facts are. In all the 
species the active adults of which I have examined, the dorsal 
and photogenic cells (figs. 7 and 8) are different as regards nuclei, 
cytoplasm and cytoplasmic granules, and no intermediate or 
transformation stages of the cells as such of one layer into the 
other has been found. This was most clearly shown in sections 
of Photinus ardens, P. ebriosus, P. pallens, Photuris jamaicencis 
and P. pennsylvanieca fixed in osmic acid and stained in iron- 
haematoxylin followed by van Gieson’s-picric-acid-fuchsin, also 
with phospho-molybdie acid haematoxylin (fig. 7), thionin-eosin, 
and best of all by the anilin blue-orange G. stain. In these 
preparations the groups of photogenic cells show a definite pe- 
ripheral boundary (fig. 7, Z.), the cytoplasm of which is nearly 
uniform. Inside of this boundary are the granular contents. 
In some species tracheal-end-cells of the tracheal system are 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 4 


422 E. J. LUND 


found between the two layers and may readily be mistaken for 
cells which seem to be transforming from cells of one, into cells 
of the other layer, if they are not brought out by the use of osmic 
acid. Cells which seem to be in the process of transformation 
may also appear in sections which have been improperly and 
weakly stained in, e. g., picro-carmine. In some such prepara- 
tions it is almost impossible to distinguish any sharp line of separ- 
ation but other sections of the same specimen when stained with, 
e.g., the anilin-orange G. stain show a sharp line of demarcation 
between the layers. The dorsal layer consists of closely packed 
polygonal cells sharply outlined by cell walls and with a nucleus 
centrally located. The cytoplasm may contain different amounts 
of granule deposit, depending upon the physiological state of 
the photogenic organ. The cells however nearly always retain 
their original outline and size, showing that the visible difference 
in the state of the cells is only that of the amount of the product 
of metabolism. The filling up of the cells of the dorsal layer may 
proceed to such an extent that apparently all the cytoplasm 
becomes displaced, the original cell boundary and position of 
the nucleus barely remaining visible. How this comes about 
we shall see very shortly. In all the species the organs are pene- 
trated by tracheal trunks which arise from more or less irregular 
branching tracheal trunks in the body cavity or from tracheae 
which lie upon the dorsal side of the organ. These tracheae which 
pass into the organ penetrate the dorsal and photogenic layers 
in different ways. Thus in males and females of P. pennsyl- 
vanica P. pallens, P. jamaicencis and males of P. suavis and P. 
commissus the vertical tracheation is most pronounced. Taking 
all the species as a whole we find a great variation in the direc- 
tion of penetration of the tracheae through the organs. In the 
forms which have a thick photogenic organ the tracheae as 
a rule are vertical while in species which have a thinner organ 
or the photogenie layer only one cell deep very few tracheae 
penetrate the photogenic layer. In the latter forms the tracheae 
branch and spread out upon the ventral side and between the 
dorsal and photogenic layers of the organ (figs. 3 and 8). The 
illustrations are from such organs as have their tracheae verti- 


PHOTOGENIC ORGANS 423 


cally placed except fig. 8. Sections of organs with vertical or 
horizontal tracheation show the general relation of parts most 
distinctly. 

The photogenic layer consists of three different kinds of cells 
(figs. 4, 7 and 8) (a) The cells of the tracheal epithelium, (b) The 
tracheal end cells,! and (c) Densely granular photogenic cells, 
which make up the greater part of the photogenic layer though 
this may vary in different species and different specimens of the 
same species. When a horizontal section (fig. 4) is made through 
an organ with vertical tracheation thereby obtaining cross sec- 
tions of the vertical tracheae, we find the lumens of the vertical 
tracheae appearing as round openings (fig. 4, Z). Immediately 
outside of the vertical tracheal tube, crescent shaped nuclei 
may be found applied close to it. (Fig. 4 NV.) These designate 
the position of the cells of the tracheal epithelium and are the 
only cells to be considered as making up the tracheal epithel- 
ium. The cytoplasm of the’ tracheal epithelial cells is clear 
and small in amount. Outside of this structure are found 
somewhat irregularly placed nuclei of cells which are arranged 
in a concentric ring or mass around the vertical trachea and 
its epithelium. These are the tracheal end cells. They form 
a vertical hollow cylinder (figs. 3, 4, # and 7, #) around the 
vertical trachea and the length of this cylinder of cells shown 
in section in figs. 3 and 7 is equal to the length of the part 
of the vertical trachea which passes through the photogenic 
layer. If the trachea assumes an oblique direction the extent 
of the cylinder or tube of tracheal end cells is longer and 
follows the trachea throughout the whole distance which it 
passes in the photogenic layer. The distinction between the 
cells of the tracheal epithelium and the tracheal end cells has 
not been made heretofore. The tracheal end cells have often 
been considered as constituting the tracheal epithelium. Town- 
shend (’04) states that they are distinct in P. marginellus. 


‘The term transition cell has been used by Townshend (’04) and Holmgren 
(95). The term tracheal-end-cell however indicates their position, they being 
terminal to the part of the tracheae which have taenidia (Fig. 2), andsince no nuclei 
indicating the presence of a distinct tracheal epithelium is present I have preferred 
te use the older term. 


424 E. J. LUND 


Emery, and Wielowiejski consider the tracheal cells to be derived 
from the tracheal epithelium but do not state in what way they 
are to be considered as being derived. To me it seems most 
reasonable to consider the tracheal end cell to be a greatly en- 
larged terminal tracheal epithelial cell which has come to surround 
the tracheal furcation by virtue of its original position on the 
tracheal branch. The number of cells making up the tracheal 
epithelium is very small compared to the number of tracheal end 
cells of such a dorsi-ventral branch. In many instances it is 
difficult to be sure whether all the cells, the nuclei of which 
appear as in (fig. 4, #) are actually tracheal end cells, i.e., whether 
every one of these cells enclose within their cytoplasm the origin 
of the tracheoles from a tracheal branch (figs. 2 and 8), or 
whether some of these nuclei represent cells which do not surround 
any such fureation but are simply placed in between the tracheal 
end cells. In some species it is more difficult to tell than in 
others, for the reason that the number of cells making up the 
cylinder around the trachea is very much greater, e.g., P. pennsyl- 
vanica and P. ebriosus, than where the cylinder is made up of a 
single layer of cells. In order to make certain that all the cells 
of the cylinder are tracheal end cells the number of points at 
which reduction of osmic acid had taken place (90 to 100 more or 
less) in preparations which showed all the furcations to contain 
reduced osmic acid were counted and compared with the number 
of nuclei present. The results (P. pennsylvanica) showed a very 
close agreement in number. In cylinders where a single row of 
cells in section are present, e.g., P. sp. near maritimus and Pho- 
tinus sp. (?) (fig. 7) it is very evident that every cell of the tra- 
cheal cylinder is a true tracheal end cell. The tracheae which 
pass to the ventral side end here in two or more branches and 
each one of these terminate ultimately in a tracheal end cell and 
tracheoles. The latter penetrate upward into the lower side 
of the photogenic cells. In many forms branches from the 
part of the vertical trachea opposite the line of separation of the 
two layers pass out between the two layers. These terminate 
on the dorsal side of the photogenic layer as in (fig. 3, X), their 
tracheoles passing ventrally. Consequently we may have a 


PHOTOGENIC ORGANS 425 


single photogenic cell or group of cells which are in section com- 
pletely surrounded by tracheal end cells. Branching of the large 
tracheae rarely occurs in the dorsal layer, the formation of branches 
being almost entirely confined to the part passing through the 
photogenic layer. I have never found tracheal end cells with 
the exception of one or two doubtful cases in the dorsal layer nor 
have I found tracheoles penetrating the dorsal layer in any of 
the species. This points directly to the fact that the supply of 
air or oxygen to the dorsal layer has no peculiar significance for 
the functioning of its cells as is the case with the photogenic 
layer. 

The walls of the tracheae and branches are strengthened by 
taenidia as in other forms. Erect hairs situated on the inside 
of the larger tracheal trunks are often seen. I have not found 
them present in the tracheae of the photogenic layer. Shortly 
before or at the point where a tracheal branch enters the tra- 
cheal end cell its diameter is diminished to about 2. (fig. 2). 
In some species this diminution does not take place until the 
point of origin of the tracheoles is reached. The taenidia dis- 
appear at this point leaving a small tube which passes into the 
cytoplasm of the tracheal end cell. This smooth tube soon divides 
into two or more branches, the tracheoles, tracheal capillaries 
of some authors. The point of fureation is always located near 
the side of the tracheal end cell which is adjacent to the photo- 
genic cell, some times only a very thin layer of cytoplasm remain- 
ing between it and the photogenic cell. The nucleus of the tra- 
cheal end cell is more proximal than the fureation (figs 2, 7 and 
8). The diameter of the tracheole appears to be very nearly the 
same in all the species, about 1.1. Since the tracheoles can 
only be made to appear by means of the reduction of osmic acid 
and since the amount of reduction varies (fig. 2) and depends 
upon several conditions the tracheoles appear to have a greater 
diameter in some preparations than in others. The diameter 
varies-as nearly as could be determined from 1.1 to 2u. They 
are tubular. This could be seen in very thin sections ly thick; 
when light passed through, the cross section of the tracheole 
appeared as a thick black ring, with a small opening. Their 


426 E. J. LUND 


diameter is not constant throughout but diminishes slightly as 
they pass into the photogenic cell so that we have a slightly taper- 
ing tube. 

The tracheoles are easily shown to consist of a firm substance 
different from the cytoplasm as others have found in the Euro- 


Fig. 2 a, b, c, d, e, free hand drawings of tracheal end cell apparatus selected 
from different preparations showing successive degrees of osmic acid reduction 
upon the tracheoles, in the cytoplasm and distal end of trachea; n, nucleus; t, 
taenidia. 


pean forms and Townshend ’04 in the American form P. mar- 
ginellus. Tissuetreated with a n/s. of NaOH show the tracheoles 
to remain intact while the cytoplasm of the cells is dissolved away. 
The structure which remains after treatment with the hydroxide 


PHOTOGENIC ORGANS 427 


cannot so far as staining reactions are concerned be distinguished 
from the substance of the trachea and may in so far be considered 
of a chitinous nature. That the substance of the tracheole is 
different from that of the trachea by reason of the fact that osmic 
acid is reduced upon it is of no significance, for reduction is not 
always limited to the tracheole. 

Weilowiejski and Bongardt (’03) found the number of trache- 
oles in L. splendidula arising from each tracheal end cell to vary, 
there being sometimes as many as six or seven. Emery found 
uniformly two present in Luciola italica. Townshend found them 
to vary in number in Photinus marginellus though usually two 
were present. I have found two in the following forms, male 
Photuris pennsylvanicus, Photinus sp. near maritimus, P. ebrio- 
sus, P. ardens var (?). In females of Photuris pennsylvanicus I 
found three present. In some cases four seemed to be present 
in females (?) of P. pennsylvanicus, but this could not be defi- 
nitely determined because of the character of the osmic acid reduc- 
tion in these preparations. In males of Photuris pennsylvanicus, 
Photinus sp. near maritimus, P. ebriosus, and P. ardens var (?) 
the number is constant. 

Emery considered the tracheoles to pass between the photogenic 
cells in Luciola italica. Wielowiejski was of the same opinion 
for Lampyris splendidula, though later (89) he found that some- 
times the tracheoles did penetrate into the cells. Bongardt (’03) 
states that he has not found the tracheoles to enter the photogenic 
cells in any of the species which he studied. Townshend (’(04) 
states that it is ‘probable that the tracheoles in their course out- 
side the cylinders’ are intercellular, and Watase (95) also states 
that they are intercellular. Before we attempt to answer the 
question as to whether the tracheoles pass into or between the 
photogenic cells in the species studied we may consider the struc- 
ture of the photogenic cell. 

These are shown in (figs. 3, 4, 7 and 8, P). In nearly all the 
species a cell wall is absent and in some species cell boundaries 
are missing. Thus the photogenic cells may be considered to 
be masses of protoplasm in which nuclei are placed, and lying 
between the cylinders the latter formed by the tracheal end cells, 


428 E. J. LUND 


The only means of distinguishing the structural individuality 
of the photogenic cell where it does appear is the fact that densely 
granular deposits in the cytoplasm tend to group themselves 
about the region of a nucleus (fig. 4), thus leaving narrow radiat- 
ing areas or strands of cytoplasm which may in some species 
and apparently under some physiological conditions of the photo- 
genic cells appear as cell boundaries or more rarely as cell walls, 
consequently they may be spoken of as cells. The nuclei are 
placed in the central parts of the cells and vary in size and shape. 
They are larger than the tracheal end cell nuclei. In Photinus 
commissus the nuclei are narrow and appear in rows as Townshend 
figures for P. marginellus (fig. 3). In P. sp. near maritimus, for 
example, they appear rounded and lie end to end forming a chain- 
like vow (fig. 5). The cytoplasm of the photogenic cells are in 
all cases filled with granules. In Photinus sp. (?) (Jamaican) 
(fig. 7, Z) a peripheral zone of very finely granular cytoplasm 
appears. The characteristic granules of the interior are missing 
from this zone. I have found the same but narrower zone present 
in sections of Photinus ebriosus and also in some specimens of 
P. sp. near maritimus. Whether it appears as a result of partial 
exhaustion of the contents in the metabolism of the cell or whether 
it is a permanent condition I have as yet been unable to deter- 
mine. The anilin orange G. stain most clearly differentiates the 
parts of the photogenic cell. Coarse granules appear yellowish 
orange while the peripheral zone takes a dense blue stain. That 
these differences in the appearances and relations of the photo- 
genic cells in the different forms is not due to preparation is shown 
by the fact that they appear the same in sections with different 
methods of fixation and staining. 

To return to the capillaries; in spite of the fact that the above 
writers with the exception of Wielowiejski (’89) agree that the 
tracheoles are intercellular in the European forms and the Amer- 
ican form studied by Townshend, I have become convinced after 
examining a great many preparations that the tracheoles are in 
no ease limited to the cell boundaries where such are present. 
The fact that they do penetrate into the cytoplasm is clearly 
shown by the fact that cross sections of the tracheoles appear 


PHOTOGENIC ORGANS 429 


close to the nuclei of the large photogenic cells in the same focal 
plane. Furthermore in tissue where no cell boundaries are evi- 
dent we find them penetrating in all conceivable directions. 
Thus several osmic acid preparations of the photogenie organ of 
P. ebriosus show the cytoplasm of the photogenic cells pene- 
trated in all directions by an almost innumerable number of 
tracheoles. Such preparations can not but convince one of the 
importance of this greatly developed tracheolar net work for the 
functioning of the photogenic layer. 

Observations differ also as to whether or not the tracheoles 
anastomose. Emery states that they do not anastomose in 
Luciola italica, Wielowiejski found the tracheoles to anastomose 
in some cases. Bongardt (’03) found no anastomoses. Town- 
shend has figured the anastomoses in P. marginellus (figs. 5 and 
6). In horizontal sections of Photinus sp. (?) (Jamaican) I 
have found structures like that figured by Townshend in fig. 6. 
The radiating structures arising from the tracheal end cells are 
undoubtedly tracheoles. I was however unable to convince 
myself that anastomoses did occur in this case. In the photo- 
genic organs of P. ebriosus the tracheolar network is especially 
dense and in sections showing strong reduction of osmic acid a 
great many anastomoses could be plainly seen. I have found no 
branching of the tracheoles of the photogenic organ in any of 
the species examined. Anastomoses of tracheoles have been 
found by Wistinghausen, and Holmgren in certain caterpillars. 
Wistinghausen and others have found the tracheoles to be made 
up of a chitinous tube with a ‘peritracheal membrane’ which is 
continuous with the cytoplasm of the transition cells (correspond- 
ing to the tracheal end cells of photogenic organ). It is readily 
seen that anastomoses of the tracheoles renders the mechanism 
of respiration much more efficient due to the possibility of flow 
of the air through the tracheoles caused by unequal pressure on 
different parts. Since it is found that the tracheoles do anasto- 
mose in those instances where they have been made a special 
study we should perhaps expect to find them filled with air. But 
such however appears not to be the case so far as direct obser- 
vation shows. I separated the organ of P. pennsylvanica from 


430 EK. J. LUND 


the sternite by means of a very small brush and immediately 
examined them under the microscope in the dark. The ventral 
ends of the vertical tracheae appeared as glistening, white, branch- 
ing structures and were filled with air. No trace of tracheoles 
could be seen though they occur upon or very near the ventral 
surface of the photogenic layer, neither could they be seen by 
means of ordinary light. Thus it would appear that they were 
filled with liquid. From the fact that the tracheoles are found 
to anastomose it becomes easier to explain this condition as being 
a temporary rather than a permanent one for the tracheoles could 
be more readily freed from the liquid by the unequal pressure 
caused by the respiratory muscles than they could if the tracheoles 
ended blindly. Bongardt (’03) found the tracheoles to be sur- 
rounded by a continuation of the cytoplasm of the tracheal end 
cell. Wistinghausen and Holmgren found a peritracheal ‘mem- 
brane’ around the tracheoles continuous with the ‘transition 
cells’ of certain caterpillars. I have not found any distinct 
cytoplasmic membrane surrounding the tracheoles. However 
when osmic acid becomes reduced the greater the reduction the 
greater is the outside diameter of the tracheole. Furthermore in 
teased preparations the cytoplasm appears spread out between 
the tracheoles in a membrane like fashion and passes out upon 
them a short distance (fig. 2) and as others have figured. The 
reduction of the acid occurs around the fork in the cytoplasm of 
the tracheal end cell and spreads outward until the whole of the 
tracheal end cell becomes blackened (fig. 2). Now since in such 
preparations there is always a definite limit to the amount of reduc- 
tion around the tracheole it would appear that we have a limited 
region around the chitin tubule which reacts to osmic acid the 
same as the cytoplasm of the tracheal end cell, and hence is spe- 
cifically different from the adjacent cytoplasm of the photo- 
genic cell. Thus we may conclude that whatever the function or 
property of the tracheal end cell may be the thin layer correspond- 
ing to the peritracheal membrane of Wistinghausen and others 
possesses the same properties as the cytoplasm of the tracheal end cell 
so far as the osmic acid reaction indicates. The acid is not reduced 
in the chitin of the tracheae, neither does the cytoplasm of the 


PHOTOGENIC ORGANS 431 


tracheal epithelium of the photogenic organ show this specificity. 
So that whether we consider the substance of the tracheole to be 
different from the chitin of the trachease and endowed with the 
specific property of reducing osmie acid or whether we consider 
the reduction due to the specific property of the cytoplasm of 
the tracheal end cell the fact remains, that the nucleus and clear 
cytoplasm of the tracheal end cell and also the outer parts of 
the tracheole have the same property, viz., the power to reduce 
osmie acid and that this property is primarily peculiar to the tra- 
cheal end cell. 

When the cells of the organ are treated with osmic acid they 
of course are killed after a short time though the process of photo- 
genesis may go on for some time after immersing the organs. 
This is undoubtedly due to the characteristic slow penetration 
of osmie acid, for I found that the light from a solution of the 
luminous secretion of the ostracods Cypridinia squamosa(?), 
and Cyclopina gracilis was instantly extinguished when a small 
quantity of osmie acid solution was added. It is evident that 
the osmic acid, reaction is not necessarily due to, and it is believed 
will be shown not to be dependent upon, the temporary nor con- 
tinued vitality of the cytoplasm but that it is due to some spe- 
cific formed substance present in the cytoplasm of the tracheal 
end cell and the peritracheolar membrane.2 To this question we 
shall return later. 


PHYSIOLOGY 


As stated above the crystalline deposit in the cytoplasm of 
the dorsal layer cells may vary in amount. I have found that 
where the crystals appear in abundance the photogenic cells 
are shrunken and the region of the tracheal cylinders are large and 
vacuolated, while in cases where the crystalline deposit is small 
or not present the photogenic cell groups are plump and the cyto- 
plasm very densely crowded with granules. That this is not due 
to methods of fixation, staining, etc., is clearly shown by the fact 

2 It seems to me that the proper term for this membrane should be peritracheolar 


membrane rather than the misnomer ‘peritracheal membrane’ since it is peculiar 
to the tracheole rather than the trachea. 


432 E. J. LUND 


that the same differences appear in material carried through 
different processes of preparation. In extreme cases the filling 
of the dorsal layer cells has gone on to such an extent that the 
cell boundaries have almost disappeared. When the tissue is 
treated with gold chloride and formic acid the granules appear 
in the form of large crystals in the dorsal layer. These are either 
monoaxial or biaxial crystals. The granular deposit in the dor- 
sal layer under normal conditions where only alcohol and xylol 
have been used in preparation of the sections appears uniformly 
granular like it appears in fresh tissue. The same is true when 
osmic acid has been used on such organs. The granules are 
minute crystals less than 0.54 in diameter. Their crystalline 
nature is shown by their high birefringency. They are totally 
different from the granules of the photogenic cells which are 
non-crystalline and yield totally different reactions. Bongardt 
(03) and Koélliker (64) found them to be ammonium urate and 
after adding NaOH ammonia was given off. Others have found 
them to be ‘some salt’ of uric acid. When the xanthin testis 
applied to the isolated tissue, the residue after treatment with 
concentrated HNO; and ammonia vapor shows a reddish purple 
color. Upon subsequent addition of NaOH however, the color 
turns deep reddish-brown instead of bluish-violet. The tissue 
gives the test for guanin under some conditions, with an alkaline 
solution of diazobenzolsulphonic acid. It would appear from 
these results that we have here a nitrogenous compound related 
to if not identical with some of the derivatives from nucleic 
acid. The interesting point is that we would accordingly have 
nitrogen present, and perhaps phosphorus. This immediately 
suggests that we should expect the granules of the photogenic cells 
to be a nitrogenous compound since a direct relation exists between 
the amounts of the granular contents in the layers. Furthermore 
this direct relation and the actual tracing of the products of decom- 
position resulting from the process of photogenesis from their 
place of origin in the photogenic cell into the dorsal layer cells, 
is shown strikingly and conclusively in sections of the organs of 
different species (e.g., P. pennsylvanica, P. ebriosus) prepared 
by a variety of methods. The photograph (fig. 5) shows a view 


PHOTOGENIC ORGANS 433 


of a longitudinal section of the photogenic organ of P. sp. near 
maritimus. The dense white mass is the crystalline deposit in 
the dorsal layer cells, in this case the cells are completely filled 
with the minute erystals. Above this mass is shown small 
amounts of the same product among the viscera of the body 
cavity. On the ventral side shown under high power in fig. 6 
are the same products definitely located on the most peripheral 
parts of the cytoplasm of the photogenic cells, between the tra- 
cheal end cells making up the cylinder and the photogenic cell. 
It is important to note the exact limits in the location of this 
crystalline deposit none being found within the cytoplasm of 
the tracheal end cell nor in any part of the tracheal cylinder. 
The outlines of the nuclei are shown in (fig 5) appearing in chains, 
and separate in (fig. 6). They are surrounded by a layer of the 
same substance. The photographs are from specimens which 
showed the greatest amount of accumulated products of photo- 
genesis in the organ found in any of the material examined. I 
have found different amounts of the accumulated products of 
katabolism in different species and different specimens of the same 
species, so that we may find organs with only a small amount of 
deposit on the periphery of the photogenic cell masses with no 
such deposit upon the surface of the nucleus and in many cases 
if not in most cases very little if any exists in the photogenic layer. 
The degree of filling of the dorsal layer cells also corresponds to 
the amounts of the deposit upon and in the photogenic cells. 
It is important to note that the region of the nuclei is the last 
to show the presence of the waste products. This would seem to 
indicate that the chemical processes that have to do with the for- 
mation of this crystalline decomposition product can be traced 
back to the region of the nucleus. This point becomes more 
interesting when we compare the results obtained from the inves- 
tigations by R. 8. Lillie (02) and others on the properties and 
functions of the cell nucleus in metabolism, with special reference 
to its oxidative properties. From what follows it will be seen 
that the foregoing may or may not have to do with the immedi- 
ate processes of photogenesis, i.e., determine the localization of 
the origin of the light emissions from the parts of the photogenic 


434 Hy) Jie LUND. 


cell. It is to be considered as a condition where the waste prod- 
uct has not been removed from the photogenic cells, but has 
begun to accumulate. In this condition the organ is still fune- 
tional and the intensity so far as can be seen with the naked eye 
is apparently the same. From the facts concerning the rela- 
tion, In amount, of granular deposit in the dorsal and ventral 
layers it is evident that we are to consider the granules of the 
photogenic cells as at least one if not the main source from which 
the crystalline deposit is derived. Furthermore DuBois states 
that he observed in the luminous secretion of the centipede, Orya 
barbarica, a direct transformation of the granules of the photo- 
genic secretion into crystals. From this it would appear that 
Weitlaner (’09) is in error when he concludes (p. 103) that ‘‘3. 
Die harnsauren Ammoniakschéllechen Ko6llikers [crystalline de- 
posit] haben den Hauptteil am Leuchten, sie sind die Elemente 
des Leuchtens und man spricht richtiger von Leuchstoff als von 
Leucht-organen,’’ and in view of the fact that observations by 
other workers upon the forms which he studied (Lampyris splen- 
didula and L. noctiluca) and also as far as observation on other 
forms go, the dorsal layer is non-photogenic. I have found no 
traces of ight from any parts of the body cavity, and the fact 
that small amounts of the crystalline substance do occur in the 
body cavity is of course no evidence for the localization of a proc- 
ess of photogenesis in the body cavity or fat body since the sub- 
stance is soluble and in part undoubtedly dissolves in the fluids 
of the body cavity. We are not even justified in concluding that 
the origin or localization of the light in the photogenic tissue is 
exactly the parts of the photogenic cells which as shown in (fig 6) 
contain the crystalline deposit, for observation shows that light 
isnot limited in its origin to the regions where the crystalline 
deposit appears in the photogenic cell. All we can say is that the 
places in the photogenic cell where the crystalline deposit appears 
in preparations, are the points at which crystallization takes place 
or where the cytoplasm becomes most saturated with the deeompo- 
sition product. It is of vital importance to one’s understand- 
ing of the processes of cell metabolism that we do not simply 


PHOTOGENIC ORGANS 435 


assume simplicity and let this assumption cause us to overlook 
evident complexity in the processes which go on in the cell. 

The granules of the photogenic cell are uniform in size approxi- 
mately 0.4 in diameter. In fresh tissue they do not appear to 
be spherical. The fat globules in the cells of the fat body are 
from 5 to 15 times larger and perfectly spherical. Thus we see 
at once a difference between them in some of their physical 
properties. In some of the literature on the subject when an 
explanation of the process of photogenesis is attempted it is usu- 
ally stated that the process is ‘‘probably one of oxidation of a 
substance of fatty nature.’ This led me to try various micro- 
chemical tests for fats and related substances. Osmic acid has 
usually been regarded as a specific ‘stain’ for fat. Lately Sudan 
III and Scharlach R. have been used and considered to be more 
accurate tests. Osmic acid blackens the globules in the fat body 
under all conditions, so that it becomes of importance to deter- 
mine if possible the conditions under which this substance is 
and is not reduced in the tracheal end cell where no such fat deposit 
is present but instead a clear cytoplasm. To this we shall return 
later. 

In order to see what the effect would be to feed the insects 
with Sudan III as has been done by Riddle (710) in studying the 
deposition of the yolk layers in the hens egg I fed specimens of 
the various species and also Pyrophorus plagiophthalmus on sugar 
cane impregnated with Sudan III for five or six days. The ani- 
mals remained perfectly healthy and active. Such animals became 
pinkish in color showing that the Sudan III was distributed 
through the body tissues. Sections showed that the stain had 
been strongly deposited in the fat globules of the fat cells, the for- 
mer being stained a deep orange red. A faint pinkish coloration 
took place in the other tissues, which was very probably due to 
the presence of the stain in solution in the body fluids.. The 
photogenic layer showed a slightly stronger stain than the other 
tissues of the body with the exception of the fat globules in the 
fat cells. But with such a slight coloration it is impossible to 
tell whether this slightly stronger staining in the photogenic cells 
than in the body may not have been due to the smaller size and 


436 E. J. LUND 


greater number of granules of the photogenic cells, and hence it 
is difficult to say whether slightly weaker color was due to smaller 
degree of specific affinity for the stain than in the case of the fat 
globules or as simply due to the physical conditions. This at 
once indicates a difference inthe chemical nature of the fat deposit 
in the fat cells and the granules of the photogenic cells. Since 
as has been pointed out above and has also been found to be 
true by others that the waste product from the photogenic cells is 
a nitrogenous compound, we should expect nitrogen present in the 
substances used up by the photogenic cell in the processes of 
photogenesis and since many of the ordinary stains show similar 
reactions of the globules of the fat cells and the granules of the 
photogenic cells, they must be considered as having some proper- 
ties in common. This led me to compare as far as possible the 
reactions of the nitrogenized fats with the reactions of the gran- 
ules of the photogenic cells. 

Loisel (’03) has studied and compared the staining reactions of 
neutral fats with those of lecithin by a method which he recom- 
mends (q.v.) and has found marked differences in their affinities 
for many stains. By mordanting with iron-alum, staining for 
a short time and then dehydrating with acetone, the following 
results are obtained. 


STAIN PHOTOGENIC GRANULES FAT GLOBULES 


Aqueous sol. orange G.. yellowish-orange not stained 


Methyl green... bluish-green pale green 
Acid fuchsin.... deep purple very pale purple or not stained 
Gentian violet deep violet very faintly violet or not stained 
Erythrosin deep crimson not stained 

Also 
Phosphomolybdic-hae- dark blue-black | not stained, clear 

matoxylin (a myelin | 

stain) 
Sudan III (fed)...... slight or none.’ | deep orange red 


3 This agrees with results of Daddi on lecethin and fat. Archiv, ital. de Bio- 
logie, ’96. t. 26, p. 145. 


PHOTOGENIC ORGANS 437 


These staining reactions agree with those for lecithin and fat 
respectively found by Loisel and Daddi. Their differences are 
so marked in all cases that if the staining reaction is to be regarded 
as an index to the chemical nature of a tissue substance at all it 
seems that it must be regarded so in this case. Of course we 
are not justified in stating that the photogenic cell granules are 
a lecithin until further conclusive tests can be applied. Yet it 
is suggestive that a substance, the derivatives from which contain 
nitrogen so closely resemble the well known lecithins in all the 
reactions that it has so far been possible to obtain. 


Effects of certain chemicals 


The dorsal layer of the photogenic organ shows a strongly acid 
reaction while the photogenic layer has a slightly weaker acid 
reaction. This was tried several times by removing the two lay- 
ers from the transparent chitin and then turning the respective 
layers down upon the litmus paper and comparing the resulting 
colors. Heineman also found the tissues of the organ of Py¥ro- 
phorus to have an acid reaction. This of course would not neces- 
sitate that the photogenic process take place in an acid medium. 
for it is well known that different parts of the same cell and the 
same parts of a cell at different times may give different reactions 
as regards acidity and alkalinity. Watase (95, p. 115) reason- 
ing from analogy to certain other photogenic processes which do 
occur in alkaline media, states that here the process also takes 
place in an alkaline medium. However, so far, all the evidence 
we have, rests only upon analogy. Kastle and MeDermott (’10) 
have repeated and added a large number of experiments on the 
effects of chemicals on the light production of the organs of P. 
pyralis. The results of these authors and many others show con- 
clusively that oxygen is a necessary element for the process of 
photogenesis. This statement may not be the same as saying that 
the process is one of simple oxidation. J have found a strong solu- 
tion of hydrogen peroxide to have a most striking effect of increas- 
ing the light intensity. The following are notes from one experi- 
ment : 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. LI, No. 4 


438 E. J. LUND 


(A) 1. The organ of a @ P. pennsylvanica was isolated 
intact and inverted on a slide. One to two drops of strong H.O, 
was added. 

Result: A strong constant light appeared and oxygen was 
evolved. 

2. ‘The viscera of the body cavity were then treated with H.Q. 

Result: No light but evolution of oxygen. 

(B) 2. A second organ was isolated in the same way and 
placed on the slide dorsal layer up then H.O, added. 

resull: Bright, strong, constant light appeared. The dorsal 
layer was opaque to light. When one of these brightly glowing 
organs were placed in the tube and bulb apparatus (see p. 443) 
and pressure increased a slight increase in intensity of the light 
could be noted although the organs were apparently at their 
upper limit of normal intensity. When left moistened with H.O, 
in the air they remained brightly luminous for two hours. The 
results of this experiment and from the results found by Kastle 
and McDermott (’10), Bongardt (03) and others that in pure 
oxygen the light is more intense than under conditions where 
oxygen is diluted as in air, show at once that a direct relation 
exists between the intensity of the light and the oxygen content 
or pressure. Kastle and McDermott and other previous writers 
have shown that water is a necessary condition for the existence 
of the ight and that by drying the tissue of the organs after lorg 
periods (thirteen months in one case, Kastle and McDermott) 
light may be emitted from the substance upon admittance of 
water and oxygen. These and other writers have further found 
that by repeated drying and moistening with the admission of 
oxygen the process of photogenesis ceases after a time indicat- 
ing that the photogenic process is a result of a direct utilization 
of material. This is of course in perfect agreement with the 
results reported above on the origin of the nitrogenous waste 
product in the dorsal layer, and the diminution in the cell contents 
of the photogenic cells. From a study of the dried photogenic 
material we at once see as others have long ago pointed out that 
the immediate process of light production is not dependent 
upon the protoplasm of the cell but upon interactions between 
formed substances. 


PHOTOGENIC ORGANS 439 


Observations on the localization of the photogenic process in the 
organs 


While collecting at night I found that certain Jamaican species 
normally emit the light in a regular way, e.g., Photinus sp. (?) 
(a small yellow form), emits what appears to be a long sharp 
flash lasting from 2 to 23 seconds when seen at a distance of a few 
rods. This flash when examined more closely is seen to consist 
of a number of smaller flashes caused by regular, rapid and numer- 
ous changes in intensity. In some other species, e.g., P. sp. near 
maritimus the flash is similarly made up of a series of rapid changes 
in intensity but not as uniform and definite as in the above men- 
tioned form. Photinus ebriosus shows a strong uniform flash. 
These species may readily be recognized by the character of their 
flash. A further study would undoubtedly enable one to dis- 
tinguish other forms by this means. Similar observations have 
been made upon some of the American Lampyridae by McDermott 
Calo) 

Max Schultze (64-65) from his study of the reduction of osmic 
acid in the tracheal end cell suggested that the process of light 
production took place in these cells. Emery (’84) was the first 
to make a study of the localization of the light in the photogenic 
layer. He found that the light in Luciola italica was localized 
in rings which were constant in position but sometimes were 
broken up into luminous points and since he found the reduced 
osmic acid in the forks to be located here and in the tracheoles 
he was of the opinion that the region where the photogenic proc- 
ess took place was where the osmic acid was reduced. 

I have studied Photuris pennsylvanicus, Photinus ebriosus, P. 
pallens, P. sp. near maritimus, and others. When the organs are 
at the height of their intensity and when the light is not intense 
enough it is impossible to see anything with the compound micro- 
scope in the dark. But when the organs show a medium inten- 
sity the surface of the organ and the removed tissue can easily 
be studied. A view (fig. 9) under the high power, of the organs 


4 Unfortunately this form could not be identified for me. 
5 Can. Ento., vol. 42, pp. 357-363. 


440 E. J. LUND 


of, e g., P. sp. near maritimus shows the surface of the photogenic 
layer to be made up of a number of shaded, oval or round spots 
distributed uniformly in a brightly luminous field. This generally 
occurs when the animal is under constant mechanical stimula- 
tion such as very slight pressure or when the organ is removed 
and then stimulated. To identify these shaded areas camera 
lucida drawings were made of the active organ and then horizontal 
sections cut and compared. These showed that the shaded areas 
(fig 9, S) corresponded exactly to the cross sections of the vertical 
cylinders and the bright areas between them to the photogenic 
cells as in (fig. 4). Thus from direct observation we are readily 
able to locate the origin of the light in the cytoplasmic region of 
the photogenic cells. In certain favorably lighted regions on 
the organs, the yellowish green light is most intense in the region 
which would correspond approximately to the inner region of 
the zone bounding the photogenic cells (figs. 7, Z and 9, P). This 
zone 1s narrow in most forms and almost absent in some so that 
as far as could be determined the periphery of the photogenic 
cell shows under these conditions the greatest intensity. It must 
be noted here however, that in many cases sections show that 
the tracheal end cells lie in depressions in the cytoplasm of the 
photogenic cells as is shown in an extreme case in (fig. 8) so that 
we should expect to have a more or less indefinite bright zone 
about the vertical cylinder if the light is located in some definite 
part of this region. The nuclei of the photogenic cells appeared 
in bead-like rows in some places. The intensity of the light was 
no greater in the region around the nucleus where the crystalline 
deposit appears than in the other regions of the cytoplasm, thus 
the place where the waste products originate can not be consid- 
ered to be limited to the places where they crystallize out for we 
must assume that the nitrogenous waste product is at least par- 
tially formed at the points of origin of the ight. Under some con- 
ditions, often when the tissue of the organ is injured, the light 
proceeds from certain angular small spots in the form of an intense 
bright flash. This may occur a great many times at regular 
or irregular intervals. The angular spots are always constant 
in position and shape but the intensity of the flash may vary. 


PHOTOGENIC ORGANS 441 


I attempted to see if there was any regular distribution of them 
but none such could be made out definitely. During the emis- 
sion of the flashes from these points clouds of diffused light orig- 
inating from the deeper parts of the photogenic layer often passed 
over the organ. These generally proceeded from areas similar 
if not identical in outline and constant in position. The organs 
in this condition of activity appear very much like the view one 
may obtain in the spinthariscope. When the same organ was 
stimulated the total light content increased greatly the charac- 
teristic shaded areas with the light regions between them reap- 
peared, changing the general view of the field. Thus there 
appears to be two modes of photogenic activity of the organ which 
show that the mutual relation of the factors which make the 
photogenic process possible is not a constant and fixed one in all 
respects but that through some variation in these relations we may 
have the location where the photogenic reaction takes place limited 
to certain special regions under some conditions while under other 
conditions the localization may be extended in area. This is a 
‘spreading’ of the process and a result of increased stimulus. 
Structurally it is impossible to refer these angular points 
which are constant in position to anything except the region of 
the forks of the tracheoles in the very distal part of the tracheal 
end cell, as Emery did. The fact that those points which were 
visible, were placed irregularly is very likely due to the distri- 
bution of those forks which are at the ends of the branches of 
the vertical tracheae on the ventral surface for here the forks are 
scattered irregularly over the ventral surface of the photogenic 
layer (fig. 3). The fact that the greatest intensity of the light 
when such that the organ can be examined is located on the 
periphery of the photogenic cell would indicate as Emery also 
found that the primary liberation of light energy is from the 
region where the forks surrounded by their thin layer of tracheal 
end cell cytoplasm are imbedded in (fig. 8) or applied to (figs 4 
and 7) the cytoplasm of the photogenic cell. When we consider 
this to be the case the fact that upon increased stimulation, the 
cytoplasm of the photogenic cell appears to be the origin of the 
light, becomes easily explainable for through this we have innu- 


442 E. J. LUND 


merable tracheoles passing, and if the photogenic process took 
place around them the cytoplasm would be uniformly illumined 
as is found to be the case. On the other hand if we assume the 
photogenic process to actually take place uniformly throughout 
the cytoplasm of the photogenic cell we would have to find an 
explanation for the bright angular points in the less stimulated 
state of the organ. The tracheal end cells are not luminous and 
the fact that the cylinders are not totally dark is due to diffusion 
of light into them from the photogenic cell regions. For if we 
are to attribute any degree of luminosity to them (i.e., the small 
amount which they show) we should expect them to be brightly 
luminous, which is not the case. Now since these cells are not 
luminous and the peritracheal membrane around the tracheoles 
is blackened in the same way as the tracheal end cell we have no 
reason to believe that light production is localized in this mem- 
brane and not in the cytoplasm. On the contrary we should 
expect that the process of light production really takes place 
in the cytoplasm of the photogenic cell and very close to the con- 
tact surfaces between the tracheal end, and photogenic cells 
of the organ. This leads us farther to localize the angular spots 
in the cytoplasm of the photogenic cell immediately adjacent to the 
forks and tracheoles, for here we evidently have the greatest amount 
of reaction because of the presence of the fork and body of the 
tracheal end cell at this point. This agrees with all the facts 
from observation. The same explanation was arrived at from a 
detailed study of the osmic acid reaction before the above study 
of the active organs had been made. Finally when we come 
to consider the control of the organ the above conclusion would 
seem to be more in accord with the facts. 


Effects of rapid changes in pressure wpon the organ 


DuBois (’86) found that under diminished air pressure the 
light became weaker in Pyrophorus and at six hundred atmos- 
pheres the organs were intensely active. Evidently changes 
in air pressure in the tracheae normally vary only between narrow 
limits and if these changes have anything to do with the control 
of the organ we should expect to be able to affect the action of 


PHOTOGENIC ORGANS 443 


the organ by rapidly changing the air pressure in the tracheae 
of the body and photogenic organ. In order to approach this 
condition as nearly as possible a strong rubber bulb of 50 ce. 
capacity was fitted to one end of a heavy rubber tube at the other 
end of which was fitted a 2 dram vial. The specimen was then 
placed in this vial and a plug of slightly moistened cotton wool 
was inserted so as to keep the animal in the tube. The following 
are the notes taken in one experiment from among many on differ- 
ent forms. 

Photurus pennsylvanica 3. (a). The actively flashing ani- 
mal placed in the vial and pressure applied with the hand. 

Result: Flashes normally bright and strong when under atmos- 
pherie pressure. Increase of pressure increased intensity of flash 
temporarily. 

(b). I then cut off the abdomen from the same individual. 
Light became totally extinguished. On placing it in the vial, 
no light was produced at atmospheric pressure. I pressed the 
bulb and as the pressure was increased light appeared more 
intense in proportion to increase in pressure. An intense flash 
could be produced in this way though not quite as intense as 
that of the normal animal. Flashes were produced at will, at 
intervals, for twenty-five minutes. Only a very slight diminu- 
tion in the intensity of the light at the end of twenty-five minutes 
was noted. 

(ec). I then isolated the organ from the same individual com- 
pletely from all other tissue and placed it in the vial. A faint 
constant light was evident. I increased pressure and increase 
in light intensity in the whole organ took place. This could be 
repeated again and again, an intense flash being produced at will. 

(d). I then cut the same isolated organs into halves longi- 
tudinally. Effects of increase in pressure on halves was the 
same as in (¢). 

(e). LI removed photogenic layer from one-half of the organ in 
(d) and placed the tissue in the tube. A weak constant light was 
given off due very likely to stimulation of tissue. Increase of 
pressure gave same results as in (c) and (d). After a short time 
the tissue lost its power of response and light soon became extin- 
guished. 


444 E. J. LUND 


It will be noted that in this experiment the increase of pressure 
on the exterior was presumably the same as in the interior of 
the tracheal tubes so that the only apparent effect was to produce 
increased partial pressures of the oxygen and other constituents 
of the air. Continued constant pressure produced corresponding 
constant intensity. The conditions are not the same as when 
contraction of the abdominal muscles takes place while the spir- 
acles are closed, yet increase in concentration of oxygen is brought 
about in the tracheal tubes assuming that a free flow of air into 
the tubes takes place. Of this there can be no doubt, for the 
tracheae are kept distended by the taenidia. Thus it seems that 
the only explanation of these experimental facts, providing increase 
of pressure does not mechanically or otherwise stimulate the 
photogenic tissue, is that this increase in intensity is directly 
due to the increase in oxygen content in the photogenic tissues 
for as has been shown by Kastle and McDermott, and others, 
the intensity of the light is greater where the oxygen content is 
higher. It will be noted that these results are obtained where 
the total pressure is constant. Now when the pressures are 
lowered by means of suction on the rubber tube the opposite, 
when any, are produced, viz., the intensity of the light is decreased 
with decrease of pressure and often may be temporarily extin- 
guished. The same results from experiments with low pressures 
have been obtained by DuBois for Pyrophprus. In order to 
eliminate the possibility that the increase in pressure is a stimulus 
upon the cells in a mechanical sense I need only cite the experi- 
ments of Bongardt (’03, p. 31). He took fresh and dried photo- 
genic tissue of L. noctiluca (the former responded to stimuli but 
the dried tissue could not be made to respond under any condi- 
tions) placed them in a glass tube and slightly moistened the 
dried tissue which became luminous. Then the pressure in the 
tube was decreased to 10 mm. of mercury. After 243 minutes 
the light totally disappeared in both the moistened and fresh 
tissues. When air was again admitted (equivalent to increase in 
pressure or oxygen content) both the moistened and fresh tissue 
became intensely luminous. This leads us to the question of the 


PHOTOGENIC ORGANS 445 


Control of the organ 


There have been three main theories concerning the control 
of the photogenic process. First, the theory of DuBois, which 
holds that the respiratory processes are only related to the con- 
trol of the organ in so far as they serve to supply the necessary 
amount of oxygen in the same capacity as to the other tissues 
of the body. In other words that the respiratory muscular mech- 
anism does not determine when a flash shall or shall not take place, 
DuBois (’96) considered the control to be nervous in so far as 
the nervous system controlled the action of the muscles which 
are related to the photogenic organs; and that by contraction 
and relaxation of these muscles blood is made to flow ‘through 
passages’® in the organ. By this means of the control of the blood 
flow he explains “why sensorial or psychic stimuli may affect 
the production of light’? (p. 420). He notes however that the 
‘photogenic cells’ are ‘directly excitable.’ 

The second theory is that of Heineman (86) and Watase (’95) 
who hold that the control is by the respiratory muscular mechan- 
ism causing the inflow of air to the photogenic tissue. Heineman’s 
theory rests upon the fact that he found muscles passing over 
the surface of the abdominal organ of Pyrophorus under which 
passed large tracheal tubes penetrating into the photogenic tis- 
sue. When he blew through a tube which was inserted into the 
large ‘prothoracic spiracle’ (?) the light from the organs increased 
in intensity. This he considers an ‘experimentum crucis’ sup- 
porting his theory of control. The third one is that the organs 
are under direct nerve control. Max Schultze (’64) considered 
it probable that nerves connect with the tracheal end cells. 
Bongardt (03) figures and states that he found nerve fibers passing 
along the vertical tracheae and finally ending in the tracheal end 
cells of L. splendidula. If such nerves do exist in the organ, 
it becomes of vital importance in view of the results from the 
experiments with pressure, to see if their presence can be demon- 

’ These ‘passages’ which DuBois erroneously figured and considered blood 
spaces in L. noctiluea are of course nothing but the regions of the vertical cylinders 


and tracheae which sometimes do not stain readily and thus appear as if spaces 
are present when actually no such spaces exist, as Bongardt and others have shown. 


446 Ba Jd. LUND 


strated by a physiological method and if so what part they play 
in the process of photogeny in the Lampyridae. If we do find 
such a control we will have found nothing new so far as photo- 
genic organs in many other forms are concerned, for instance in 
the photogenic Pennatulidae, Panceri noted the passage of the 
wave of light over the colony of polyps, serving as an index of 
the fact that the passage of a nervous impulse precedes the pro- 
duction of light. Asto whether the result of this impulse partakes 
of the nature of a contraction in the photogenic cell we shall speak 
presently. Here we are particularly concerned with the question 
as to whether the nerve impulse ends directly in the photogenic 
tissue, and if so where it ends, or whether it ends in the parts 
of a mechanism which is external to the photogenic tissue, such 
as the muscles of the body. 

First of all it must be noted that the respiratory movements 
of the abdomen do not at all follow or correspond to the flashes 
emitted from the organ. Hence the simple respiratory move- 
ments do not account for the periodic emission of the light. In 
trying to carry out the idea of the compressibility of the volume 
of air and its translation in the tracheae of the photogenic organ 
I made repeated efforts to see if any minute movements of the 
tergae or other parts could be seen in the animal emitting flashes 
of ight under a binocular. No trace of such movement could be 
found. Hence if such a thing as control of the light flashes con- 
sists In a muscular mechanism its action has no visible external 
effects of contraction. So far as DuBois’ theory of control by 
means of blood flow is concerned, we would have to place the same 
limitations upon the mechanism, 1.e., action without any visible 
external movements of parts, as in the theory of control by respira- 
tion,for Du Boissuggests the dorso-ventral muscles as being at least 
part of the mechanism. It might however, very easily be the 
case that by the contraction of the heart, ete., we would be able 
to get some sort of a rhythmical flow of blood through the organ 
if the spaces did exist as he supposed. The results of the follow- 
ing experiment show I believe, that both of these theories do not 
account for the essential observed fact that the flashes are at irreg- 
ular intervals, are sharply defined and that we must inevitably 
admit the existence of a specific and direct control of the nervous 


PHOTOGENIC ORGANS 447 


system over the photogenic organ from a physiological standpoint. 
The notes taken from only two out of several similar experiments 
are reported here. 

Experiment I. (a). I made a ventral incision in a fresh active 
specimen on the segment anterior to the organ in order to cut the 
nerve cord. 

Effect: The organ remained dark. No spontaneous flashes 
emitted though the animal was active being able to run and fly, 
brush its antennae, ete. Mechanical stimulation caused a weak 
light to appear at the point stimulated. 

(b). A second specimen gave the same results except that a 
very faint constant light appeared and remained. Shaking the 
animal and otherwise stimulating it without stimulating the 
organ directly gave no flash. 

Experiment IT. (a). I laid open an active male Photuris 
pennsylvanicus by cutting along the sides of the posterior three 
segments, leaving the terga attached anterior and posterior. 

Effect: The flashes were strong and normally controlled as 
in fresh specimen. 

(b). Then I raised and removed the connected tergae from 
behind, also removed the viscera overlying the organ with a small 
brush thereby exposing the whole dorsal surface of the organ 
without injuring the nearby dorsal layer or anything laying 
applied close to it. 

Effect: Flashes normal, under complete control and not notice- 
ably different from fresh specimen. 

(ce). Then I made an incision anterior to the organ and close 
to it 

Effect: Light became extinguished except at points where 
it apparently had been injured by manupilation and then only 
showed weak and constant light at these points. Control abso- 
lutely lost. Mechanical stimulation of the isolated organ caused 
light to appear at point stimulated. This isolated organ when 
placed in the tube and bulb apparatus responded to increases 
and decreases in pressure by a marked increase in the intensity 
of the light over nearly the whole organ. 

In (ec) of Experiment 2 the blood flow if any was present must 
have been interfered with. The dorso-ventral muscles were torn. 


448 E. J. LUND 


The terga carrying the spiracles and valves were removed, and 
yet the flashes were normal and spontaneous and under perfect 
control. It must be noted that the last nerve ganglion, nerves 
and nerve chain were not dislocated in this experiment. Another 
fact of almost equal importance is that the photogenic tissue, 
apart from the nerve chain and other tissue, is highly irritable and 
responds locally to stimuli. Where is the muscular mechanism 
for such a response in this case? 


Effects of temperature 


Mangold (’10) has summarized the results of experiments by 
others on the effects of different temperatures upon photogeny 
(p. 345). No uniform results have been obtained among differ- 
ent observers. But from those which have been obtained it is 
clearly seen that there are optimum and maximum temperatures 
for the process. I have found from a number of experiments 
that with live animals or removed abdomens placed in a vial 
into which a thermometer was fitted and the whole immersed 
in water at 90° to 98° C. the temperature at which the light 
was extinguished varied in the case of Photuris pennsylvanica 
between 45° and 54° C. There was greatest intensity at about 
40° C. Photinus sp. near maritimus gave results varying from 
47° to 55° C. The specimens of this species were immersed with 
forceps directly into water the temperature of which was registered 
by a thermometer. With this species the temperature could be 
raised considerably above the point at which light became extin- 
guished and the light revived. This was noted again and again. 
The maximum temperature from which the light was revived was 
found to be about 84° C. The light reappearing at temperatures 
around 50° C. In all cases where the whole animal was used it 
died at about 40° to 45° C. and all control of the light was lost, 
so that in P. sp. near maritimus the light was always continu- 
ous after the specimen had been heated above a temperature of 
40° to 45° C. In raising the temperature the light in P. sp. near 
maritimus was always found to pass from the characteristic 
greenish yellow to a yellowish, thence to yellowish orange and 
finally to orange and sometimes into a reddish orange color before 


PHOTOGENIC ORGANS 449 


it became extinguished, the rapidity of these changes depending 
upon the temperature of the water. The higher the temperature 
the more rapid the changes and hence at high temperatures the 
color would seem to change directly into orange before the light 
became extinguished. At lower temperatures of the water (50° 
to 60° C.) the transition in colors was gradual. Traces of this 
same change in color was noted in Photuris pennsylvanica and 
Photinus ebriosus. When the organs were cooled and the light 
reappeared (P. sp. near maritimus) the transition in color of the 
light was in the reverse order, i.¢., orange to yellowish and thence 
to greenish yellow. This sequence in color changes becomes more 
interesting when we compare them with color changes reported 
to occur in certain marine photogenic organisms. 

These experiments upon the live animal and the severed abdo- 
men are obviously ill adapted to tell us definitely at what temper- 
ature the particular chemical process between the formed sub- 
stances which results in light takes place, for we have to take 
into account the control and subsequent loss of control of the 
organ, also the removal of the necessary supply of air or oxygen 
to the organs and the formed substances concerned, upon the 
death of the animal. There may also be other effects of temper- 
ature which have nothing to do with the essential process of pho- 
togeny itself but which indirectly determine whether it shall or 
shall not take place. Yet certain facts stand out: (1), that the 
process of photogeny is prevented from taking place at some temper- 
ature; (2), that it may be continued if the material is not heated 
above a critical temperature which prevents the possibility of further 
action when the temperature is afterward lowered. 

With the purpose of finding if possible whether the essential 
chemical process concerned in photogeny is affected by certain 
definite temperatures I performed a series of experiments upon 
the luminous secretion of Cypridina squamosa (sp.?) and Cyclo- 
pina gracilis’, two luminous salt water ostracods found upon the 
bottom in shallow water at Montego Bay. The solutions were 
made with pure filtered rain water, no distilled water being avail- 
able at the time. Solutions with sea water were also tried in 


7 Miss Mary J. Rathbun has had these and some other forms identified forme 
and I wish here to express my thanks. 


450 E. J. LUND 


other experiments. No noticeable differences in the results 
was found.’ Three or four of the ostracods were squeezed with 
a pair of forceps in the test-tube containing the rain water, 
thus causing in many cases a copious, Intensely luminous greenish 
yellow secretion to be freed in the water. When this was shaken 
a homogeneous luminous solution was obtained. The thermom- 
eter was hung in the luminous solution in the test-tube and the 
latter was immersed in a beaker of water kept at from 90° to98° C. 
The alcohol lamp was enclosed in a hood so that no light from this 
source disturbed the observations. The following are the tabu- 
lated notes on the results of one series of experiments upon the 
luminous secretion in a solution of rain water after the tempera- 
ture effects had been noted in a general way by a preceding series 
of trials. 


SOLUTION IN 


IMMERSION FILTERED RAIN RAISED TO RESULT 
WATER 
degrees 

A 60 C. Recovered to a bright lumin- 

escence upon cooling 
first B 65 Recovered to a bright lumin- 

Same solution { escence upon cooling 
|second B’ 65 Showed only a very shght 


recovery of luminescence 
upon cooling 


first C 67 Recovered to a less degree 

Same solution 4 | than solution B’ at 65° C. 
second | (CH 67 | Did not recover 

Sane nition first D | 70 | Recovered very faintly 

second IDY 7 Did not recover 
E 71 Did not recover 

F 70 Did not recover until after one 

minute and then very faintly 
G 71 Did not recover 


Solutions raised to temper- 
atures above 70° to 71° C. 
never recovered their pho- 
togenic power. 


8 Thus there would seem to be no inherent reason so far as the essentials of 
the chemical process are concerned why photogenic organisms should not be found 


in fresh water, asis the case. Perhaps the photogenic function is to be considered 
a primitive one so far as phylogeny is concerned. 


PHOTOGENIC ORGANS 451 


In all these experiments light disappeared at 50° C. and light 
reappeared again at 50° C. The latter was determined with 
solutions raised to 51°, 52° and 53° C., and then allowed to 
cool, for when the solutions were raised to 60° or 65° C. and higher, 
the temperature at which visible reappearance of the light took 
place was not so well defined and sharp, because of the low inten- 
sity and hence it was difficult to determine the exact temperature 
at which the process began again. The time rate of return of 
the light depends upon the time it takes for a solution to be low- 
ered to the temperature where light reappears, for quick cooling 
caused a quick return of the light.° Thus we see that here we 
have a luminous secretion which when isolated yields definite 
results while with the photogenic organs of the Lampyridae it is 
more difficult to determine the exact temperatures, very probably 
for the reasons given above. The important thing is that they 
show exactly similar behavior toward temperatures. 

Since the specific effect of temperature is one of the chief criteria 
for showing the existence of enzymes these results have led me 
to study further the osmic acid reaction in the tracheal end cells 
of the Lampyridae. If the reduction of osmic acid on the tra- 
cheoles and in the tracheal end cells depends upon the existence 
of a specific enzyme then when the tissues of the organ are heated 
above the temperature at which the enzyme is destroyed and 
then treated with osmic acid we should presumably obtain no 
reduction of the acid in the structures where the enzyme is located. 
If this is the case we shall have found an explanation for the reduc- 
tion of osmic acid in these cells. 

The following are the summaries of results from six experiments 
to test the effect of temperature upon the photogenic tissue with 
regard to subsequent reduction of osmie acid in the tracheoles 
and the tracheal end cells. All the experiments except Ex- 
periment 5 had controls; the latter passed through the same 


° In these experiments where intensity of light was noted the results depend 
upon the visual sense of the observer. No apparatus for measuring the intensity 
and point of disappearance of the light was at hand. But precautions were taken 
to make all the observations in a perfectly dark night and against a black back- 
ground. 


452 E. J. LUND 


processes except that of being heated. Sections were cut and 
mounted, some stained and others unstained. 

Experiment 1. Sixteen specimens of Photinus sp. near mari- 
timus were taken. 

(a). Control. From eight specimens I removed the abdomen 
and placed them directly into 0.5 per cent osmic acid. 

(b). Eight whole specimens were immersed in water at 98° 
C. for one minute. The abdomens then cut off and placed in a 
0.5 per cent solution of (OsO,) osmie acid thirteen hours (same 
length of time as in (a).) 

Result: (a). Reduction took place either in whole or greater 
part of organ in all the eight specimens. 

(b). No reduction of OsO, took place in any of the tracheoles 
of the boiled specimens. 

Experiment 2. Twelve specimens of P. ardens var.? were 
taken. 

(a). Control. Abdomens of six placed in 0.5 per cent OsOx; 
three of these were cut and mounted. 

(b). Six were placed in water at 99° C. in a dry vial for about 
one minute. Then abdomens cut off and placed in the OsO, 
solution. 

Result: (a). In the three that were cut the tracheoles showed 
strong reduction. 

(b). No trace of reduction of OsO,; in any of the tracheoles 
of the six specimens. 

Experiment 3. Four specimens of Photuris pennsylvanica were 
taken. 

(a). Control. Abdomens of two immersed in OsQ,. 

(b). Two immersed in boiling water one minute; abdomen 
cut off and placed in 0.5 per cent OsQO, thirty-six hours (same as 
(a)). 

Result: (a). Strong reduction in tracheoles and tracheal end 
cells. 

(b). No reduction in tracheoles and tracheal end cells. 

Experiment 4. Twelve specimens of P. sp. near maritimus were 
taken. 


PHOTOGENIC ORGANS 453 


(a). Control. Six were placed in OsO, directly and left in 
same time as (b). 

(b). Six were placed in water at 80° C. for three to four min- 
utes then into OsO,. 

Result: (a). Showed strong reduction in most of the speci- 
mens, and slightly less in one or two. All showed reduction. 

(b). Three showed no reduction in tracheoles, while three 
showed traces of reduction. 

Experiment 5. Two specimens of P. pennsylvanica were im- 
mersed in water at 714° C. until light became yellow, then orange 
and was finally extinguished. The light did not return upon 
cooling. The abdomens were removed and placed in the OsO, 
solution. No control. 

Result: Strong reduction took place in the tracheoles and 
tracheal end cells. 

Experiment 6. Twelve specimens of P. sp. near maritimus 
were taken. 

(a). Control. Abdomens of six placed in OsO,. 

(b). Six placed in water at 60° C. (several minutes). Some 
regenerated light slightly after immersing in the OsO;. All 
passed through yellow then orange when immersed in the water 
at 60°. 

Result: Reduction took place more or less in tracheoles of all 
specimens in both (a) and (b).!° 

From a study of the results of these experiments with temper- 
ature upon the intact organ, the luminous secretion and the effect 
of temperature upon the osmic acid reaction, it may at once be 
asked; is the reduction of osmie acid in the tracheal end cells 


‘0 Bongardt (’03, p. 11) states that he succeeded in staining the tracheoles black 
by ‘‘fixing the tissue in alcohol or sublimate-acetic, staining in borax carmine, 
decolorizing in acid alcohol washing in water twenty-four hours. placing in solu- 
tion of 1:400 OsO, over night then into strong acetic eight to ten hours, washed in 
distilled water, hardened in alcohol embedded and cut.’ Another method is also 
mentioned using gold chloride, which is more complex. From such treatment 
of the tissue it is obviously difficult to learn anything about the nature and where- 
fore of the reduction, which in such a method may be due to entirely different 


causes. 


tHE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 4 


454 E. J. LUND 


and tracheoles dependent upon the vitality of the protoplasm of 
the tracheal end cell or is it only dependent upon the existence 
of the formed substances which by themselves, as has been shown 
by drying the organs, are able to produce the phenomenon of 
photogenesis? With this comes the question, did the tempera- 
tures below that at which osmie acid reduction is prevented from 
taking place kill the protoplasm of the photogenic cells? An 
answer to the latter will furnish an answer to the former question. 
The animals are very sensitive to higher temperatures and dry- 
ness. In all the temperature experiments upon the whole animal 
the specimen always died at about 45° to 50° C. if left in for as 
long a time as one minute. Control of organ was lost though 
very often a residual glow, sometimes quite strong, remained 
after cooling if it had not been heated to temperatures which per- 
manently prevented the continuance of the light. It will be noted 
that strong reduction took place even as high as 713° C. and that 
traces of reduction were evident in Experiment 4, where the 
organs had been heated to 80° C. for three to four minutes. Fur- 
ther it is important to note that the reduction of the acid does 
not depend upon the continuance of the process of photogenesis 
while the acid is penetrating into the photogenic layer as is plainly 
shown in the reported experiments (4 and 5). Therefore all the 
conditions which are necessary for the process of photogenesis 
are not necessary for the osmie acid reduction in the tracheal end 
cell, it is only necessary that the organs be not heated above a 
certain temperature. 

It is of course impossible to prove that the protoplasm of the 
cells of the photogenic organ was killed at 713° or 80° C. as in 
Experiment 4. Another important fact to be remembered how- 
ever is that osmic acid is itself a strong killing and fixing agent. 
The longer the time that the organs are left in the solution the 
more extensive (within certain limits) is the osmic acid reduction 
in the tracheal end cell (fig. 2) now if a small amount of reduction 
has taken place around the fureation then evidently we must have 
had the penetration of the acid to the tracheal end cell and this 
it seems would be efficient in destroying its vitality. If then 
the first traces of the acid are sufficient and the time long enough 


PHOTOGENIC ORGANS 455 


for the killing of the cytoplasm then the further reduction of the 
osmie acid can not have been due to the vitality of the cyto- 
plasm but to some substance in the cytoplasm which by itself 
is able to reduce osmic acid. The temperatures at which the 
possibility of return to the conditions necessary for photogenesis 
becomes permanently destroyed in the Lampyridae, viz.. 70° to 
84° C., corresponds in general to the temperature which destroys 
the possibility of the osmic acid reduction in the tracheal end cell 
to take place. When tincture of guaiacum is applied to the tissue 
of the photogenic layer it does not turn blue, neither does it turn 
blue when hydrogen peroxide is added though the hydrogen perox- 
ide is decomposed. The power of decomposition of hydrogen 
peroxide is not limited to the photogenic organs but any of the 
body parts when placed in this reagent decompose it showing the 
presence of a ‘catalase.’ I immersed specimens for one minute 
in boiling water and tested parts of the body with hydrogen 
peroxide. No decomposition took place showing that the ‘cata- 
lase’ was destroyed. In view of these reactions of osmic acid 
we may consider it certain that the normal reduction in fresh 
photogenic tissue is due to the presence of some substance prob- 
ably of the nature of a reductase which is formed in the cytoplasm 
of the tracheal end cell and peritracheolar membrane. 


USE OF THE ORGANS 


Apart from the problems concerning the chemical-physiologica] 
mechanism of the luminous organs of the Lampyridae, Elateridae 
and numerous other forms, is the significance and ultimate use 
of these organs to their possessors. 

Osten-Sacken (’61) states that he found males of Photinus 
pyralis while themselves emitting flashes of light, to respond posi- 
tively to the light from the females, and that a short time there- 
after he found them copulating. Emery (’86) also found the 
males to respond to the females in the same manner though 
he did not observe a subsequent copulation.!: It is a matter of 

' Recently MeDermott writes me that he ‘has very definitely confirmed” 


Osten-Sacken’s observations for Photinus pyralis, and extended them to P. con- 
sanguineus and P. scintillans. 


456 E. J. LUND 


common knowledge to people of regions of the tropics where 
Pyrophorus abounds, that it responds under certain conditions 
to ordinary artificial light stimuli. I have collected P. plagio- 
phthalmus by means of a small partly enclosed incandescent light 
and by means of it have found them to react strongly at distances 
of fifty yards or more. During this reaction all their photogenic 
organs are in full activity. 

Comparison of the eyes and the photogenic organs in those 
species of Lampyridae where both males and females were ob- 
tained, showed. (a) That in all males, and to a smaller extent 
in those species, the females of which are most active and abun- 
dant, the eyes are greatly developed. (b) That in all cases the 
eyes and at the same time the photogenic organs of the males are 
larger than the corresponding organs of the female, i.e., the extent 
of development of the eyes is in direct relation to the extent of 
development of the photogenic organs and activity. In one large 
female (unidentified) which I found on the ground among some 
grass the organ was exceedingly small, the eyes were very much 
reduced in size and added to the above shortcomings the wings 
were rudimentary. The photogenic organ, however was very 
active and well controlled giving off a strong greenish yellow light. 
This becomes of further interest when we note that there exists 
a direct relation between the head ganglia and the control of the 
photogenic organs. When the head is removed from specimens 
control of the light is interfered with or lost and the spontaniety 
of the flashes is also lost. Removed abdomens which contain 
the last abdominal ganglion show no spontaneous control, usually 
the organs give off a continuous or irregular glow or else light dis- 
appears though they respond to mechanical and some other kinds 
of stimuli. 

A syllid, probably Odontosyllis pachydonta Verril, during 
certain periods, reacted to each other and to an artificial light 
stimulus in a most striking manner. The eyes of this syllid are 
unusually developed. In one ease a female—one of the sylliidae— 
filled with eggs was taken while brillantly luminous, and some of 
the eggs, which were shed readily, were fertilized with sperm 


taken from individuals caught in the tow net in the same place 


PHOTOGENIC ORGANS 457 


and at the same time. The next morning these had developed 
into late cleavage stages. Galloway and Welch (711) have found 
by direct observation that the positive response of the male and 
female of Odontosyllis oenopla to each other results in the bring- 
ing together of the eggs and sperm in the water so that here we 
have one of the first well established instances of the sexual adjunct 
significance of photogenicity in organisms. 


SUMMARY 


1. No fundamental structural difference exists in the photo- 
genic organs of the species of Lampyridae studied. The elemen- 
tary photogenic mechanisms of which the organs are made up, 
viz., the trachea, tracheoles, tracheal end and photogenic cell 
with their relation to the nervous system, taken collectively, are 
the same in all the forms. 

2. The tracheoles are tubular and have been shown to anas- 
tomose in Photinus ebriosus. They are made up of a chitinous 
substance, which is more resistant than the protoplasm to re- 
agents. Their number, where it could be definitely determined 
has been found to be constant for each tracheal end cell. The 
tracheoles are in no case limited in their course to the outside of 
the photogenic cells but penetrate into the cytoplasm of the latter. 
They are apparently filled with a liquid under such conditions 
as make observation of them possible. 

3. No distinet cytoplasmic membrane around the tracheoles 
could be demonstrated yet the reactions to osmic acid shows that 
a limited region about the tracheoles has the same specific prop- 
erty of reducing the acid as the cytoplasm of the tracheal end cell. 

4. The reduction of osmic acid upon the tracheoles and in 
the cytoplasm of the tracheal end cell is dependent upon the pres- 
ence of a substance probably of the nature of a reductase which 
shows the same properties as regards the effects of temperatures 
upon it as enzymes in general. The process of photogenesis is 
dependent upon the presence of this substance as is shown by the 
parallel effect of temperature upon the production of light. 


458 EB. J. LUND 


5. The process of photogenesis is independent of the vitality 
of the cytoplasm and is a resultant of the interactions of formed 
substances in the presence of water and oxygen. It is highly 
probable that it partakes of the nature of an oxidation but this 
has yet not been demonstrated for the photogenic process in the 
Lampyridae if in any animal. 

6. It has been shown that photogenesis is incident upon the 
utilization of a nitrogenous compound—the photogenic granules— 
giving staining reactions like those of lecithin and different from 
those of the true fats, and that this nitrogenous compound appears 
at least in part at the endof the process in the form of a nitrogenous 
waste product. This crystalline substance appears from its re- 
actions to be allied to or identical with some of the split products 
of nucleic acid. 

7. The dorsal layer cells become the repositories for the waste 
product. No direct transformation of the photogenic cells, as 
such, into cells of the dorsal layer takes place. 

8. The photogenic process is localized in and adjacent to the 
cytoplasm of the photogenie cells and especially (as far as could 
be determined) where the cytoplasm of the tracheal end cell, and 
tracheole is applied to the photogenic cell. 

9. The increase in intensity of the light resulting from increase 
in pressure is due to the increased oxygen content in the regions 
where photogenesis takes place. Changes in the oxygen content 
is not the primary means of control of the organs. 

10. The primary control of the organ is by nerves in direct 
connection with the photogenic tissue and not by an external 
respiratory muscular mechanism. The termination of nerve 
fibers in the tracheal end cells as Bongardt states he found in 
Lampyris splendidula is supported by the fact that photogenesis 
may be limited to points which can structurally only be referred 
to the tracheal end cells and that upon increased stimulation a 
phenomenon similar to a ‘spread’ of the stimulus takes place. 
Furthermore the photogenic tissue is irritable and responds locally 
to mechanical stimuli. 

11. A direct control relation exists between the photogenic 
organs and the nerve centres of the head. This is apparently 


PHOTOGENIC ORGANS 459 


correlated to the relation which exists between the degree of 
development of theeyes and the photogenic organs, 1.e., the extent 
of development of the eyes is in direct proportion to the extent 
of development of the photogenic organs. 

12. The positive response to light by some photogenic organ- 
isms results in bringing the eggs and sperm in nearer proximity 
to each other in cases where these are set free in the water before 
fertilization. This has been shown for Odontosyllis enopla by 
Galloway and Welch. This is also very probably true for Odon- 
tosyllis pachydonta (?) Verrill. In other cases the response may 
lead to copulation as has been found for Photinus pyralis by 
Osten-Sacken, and for Photinus consanguineus and P. scintillans 
by McDermott. 


Zoological Laboratory 
Johns Hopkins University, 
September 20, 1911. 


LITERATURE CITED 


Bonaarpt, J. 1903 Beitrage zur Kenntniss der Leuchtorgane einheimisher 
Lampyriden. Zeits. f. wiss Zool., Bd. 75, pp. 1-45. 

Crark, E. D. 1910 The plant oxidases. Dissertation, Columbia University. 

Dusors, R. 1886 Contribution a l’étude de la production de la lumiére par les 
étres, Les Elatérides luminaux. Bull. Soc. Zool. France, Année I, 
pp. 1-275. 
1892 Sur le méchanisme de la production de la lumiére chez l’Orya 
barbarica de Algerie. Comptes Rend. Acad. Sci. France, 1892, exvii, 
pp. 184-6. 
1895 Physiological Light. Smiths. Inst., Wash., D. C. Report for 
1895, pp. 413-431. 

Ermer, T. 1872 Bemerkungen iiber die Leuchtorgane der Lampyris splendidula. 
Archiy f. mikr. Anat., Bd. 8 p. 653. 

Emery, C. 1884 Untersuchungen iiber Luciola italica. Zeits. f. wiss. Zool., 
Bd. 40, pp. 338-355. 
1885 La luce della Luciola italica, osservata col Microscopio. Bull. 
Soc. Entomol. Ital. Ann. 17, Trim 3/4, p.351-5. Also Jour. Roy. Mier. 
Soc., 1886 p. 234. 


1886 La luce negli amori delle Luciole. Bull. Soc. Entomol. Ital., 
Ann. 18, p. 406. 


460 E. J. LUND 


Emery, C. 1886 Sur le lumiére des Lucioles (Luciola italica). Arch. Se. Phys. 
et Nat. (Genéve) (3), T. 14, no. 9, p. 272-5. 


Gatitoway, T. W., and Weicu, P. 8. 1911 Studies on a phosphorescent Bermu- 
dan annelid, Odontosyllis enopla Verrill. Trans. Am. Mier. Soc., 
January 1911, vol. 30, pp. 13. 


Heineman 1872 Uber die Leuchtorgane der in Vera Cruz vorkommenden 
Leuchtkifer. Archiv. f. mikr. Anat., Bd. 8, pp. 461-471. 


1886 Zur Anatomie und Physiologie der Leuchtorgane Mexikanischer 
Cucujos (Pyrophorus). Archiv. f. mikr. Anat., Bd. 27, pp. 296-383. 


HoitmGren 1895 Die trachealen Endverzweigungen bei den Spinndriisen der 
Lepidopterenlarven. Anat. Anz., Bd. 11, pp. 340-346. 


Ives, E. H., and Cospitentrz, W. W. 1910 Luminous efficiency of the firefly’ 
Bulletin of the Bureau of Standards, Wash., D. C. vol. 6, pp. 321-336. 


IKxastie, J. H. 1910 The oxidases and other oxygen catalysts concerned in bio- 
logical oxidations. Hygienic Laboratory, Pub. Health and Mar. Hosp. 
Serv. U. 8., Bull. No. 59. 


Koéturker, A. 1857 Uber das Leuchten der Lampyris. Verhandl. d. Wiirzb. 
phys-med. Gesell., 1857. Bd. 8, pp. 217-224. 


isc4 Uber den Bau der Leuchtorgane der Ménnchen der Lampyris 
splendidula. Sitzungsber d. Niederrh. Gesells. f. Nat. und Heilkunde. 


Le Bon, G. 1908 The evolution of forces. 


Liuuig, R.S. 1902 Oxidative properties of the cell nucleus. Am. Jour. Physiol., 
vol. 7, p. 412. 


LoiseL, M. G. 1903 Essai sur la technique microchemique comparative de la 
lecithine et des graisses neutres. C. R. de la Société de Biologie, 
vol. 55, p. 703. 


Lunp, E. J. 1911 Notes on light reactions in certain luminous organisms. 
Johns Hopkins University Circular, Feb. 


Macartney 1810 Observations upon luminous animals. Phil. Trans. vol. 
100, pp. 277-279. 


Mancoup, E. 1910 Die Production von Licht. Handbuch der vergleich. Phys- 
iol. (Winterstein), pp. 225-392. 


Mann, G. 1901 Physiological Histology, Oxford. 
Mouuisu, H. 1905 Luminosity in plants. Smiths. Rep. 1905, pp. 351-362. 


Nurrine, C. C. 1899 The utility of phosphorescence in deep sea animals. Am, 
Nat., vol. 33, pp. 792-799. 


1910 The theory of abyssal light. Proce. 7, Cong. Zool. 


Osten-Sackren, Baron. 1861 Die amerikanischen Leuchtkafer. Stettiner 
Entom. Zeitung, vol. 22, pp. 54-55. 


Prerers, A. 1905 Luminescence in etenophores. Jour. Exp. Zodl., vol. 2. 


PHOTOGENIC ORGANS 461 


Rippue, O. 1910 Studies with Sudan III in metabolism and inheritance. 


Jour. 
Hxp. Zodl., vol. 8, p. 163. 


ScuHrerner, O. and Suttivan, M. X. 1911 Concurrent oxidation and reduction 


by roots. Bot. Gaz., vol. 51, p. 273; also the same journal, vol. 51, p. 121. 
Scuuttzp, M. 1865 Zur Kenntniss der Leuchtorgane von Lampyris splen- 
didula. Archiv. f. mikr. Anat., Bd. I, p. 124. 


Starke, J. 1895 Uber Fettgranula und eine neue Eigenschaft des Osmium- 
tetroxydes. Du Bois Raymond. Archiv. f. Phys., p. 70. 


TowNsHEND, A. B. 1904 The histology of the light organs of Photinus margin- 
ellus. Am. Nat., vol. 38, pp. 127-148. 


Warasn, S. 1895 Physical basis of animal phosphorescence. Biol. Lects., 
Woods Hole. 


1898 Animal luminosity. Biol. Lectures, Woods Hole. 


WeirtLaNner, F. 1909 Etwas vom Johanneskaéferchen. Verh. Zool. Bot. 
Wein, Bd. 59, pp. 94-108. 


Ges. 

WieLtowigsski, H. R. von 1882 Zur Kenntniss der Leuchtorgane von Lampy- 
riden. Zeits. f. wiss. Zool., Bd. 37, pp. 354428. 
1889 Beitrage zur Kenntniss der Leuchtorgane der Insekten. Zool. 
Anzeig., Jahrg. 12, pp. 594-600. 

WISTINGHAUSEN 


1890 Uber Tracheenendigungen in den Serikterien der Raupen. 
Zeits. f. wiss. Zool., Bd. 49, p. 565. 


PLATE 1 


EXPLANATION OF FIGURES 


3. Photomicrograph of part of a longitudinal vertical section of photogenic 
organ of o& Photuris pennsylvanica, killed in 0.5 per cent OsO,, thirty-six hours. 
Weakly stained in Ranvier’s picro carmine; high power; D, dorsal layer showing 
cell walls and content of granular waste product; P, photogenic layer showing 
nuclei NV. and photogenic granules G.; V, vertical tracheae; F, furcations where 
osmic acid reduction first takes place; two tracheoles arise from each; S, sternite, 
Y, tracheal end cells between dorsal and photogenic layers from which tracheoles 
pass ventrally; similar ones may be found scattered over the ventral surface. 

4 Horizontal section through photogenic layer of o& Photuris pennsylvanicus; 
Lyons blue and borax carmine; L, lumen of vertical tracheae; NV, nuclei of tracheal 
epithelium; £, nuclei of tracheal end cells forming a cylinder around vertical 
trachea; P, nuclei of photogenic cells. This photograph shows the so-called ‘cell 
boundaries’ and grouping of the photogenic granules about the nuclei. 


PHOTOGENIC ORGANS PLATE 1 
E. J. LUND 


THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, No. 4 


463 


PLATE-2 


EXPLANATION OF FIGURES 


5 Photomicrograph of a vertical longitudinal section of photogenic organs of 
Photinus sp. near maritimus, taken with the aid of a ‘bulls-eye’ condenser by 
reflected light; the posterior and part of anterior photogenic organ is shown, with 
the relative distribution of the crystalline nitrogenous waste product. A, among 
the abdominal vicera; D, dorsal layer cells completely filled; P, periphery of pho- 
togenic cell; NV, deposit around the rows of nuclei of photogenic cells. Note rela- 
tive thickness of dorsal and photogenic layers in this condition of the organ. 

6 Photomicrograph of high power of photogenic and part of dorsal layer 
showing distribution of crystalline deposit. D, dorsal layer, cell outlines and posi- 
tion of nuclei barely indicated; P, photogenic cells greatly diminished in size and 
containing the waste products in the peripheral region of the photogenic cell and 
around the nuclei (cf. figs. 1, 2 and 5). 


464 


PLATE 2 


PHOTOGENIC ORGANS 


E. J. LUND 


6 
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 11, NO. 4 


165 


PLATE 3 
EXPLANATION OF FIGURES 


7 Camera lucida drawing of part of vertical section of photogenic organ of 
Photinus sp. (?) (Jamaican); fixed in boiling water, stained in phospho-molybdie 
acid haematoxylin. #, tracheal end cells forming eylinder—shown in section— 
around vertical trachea, V which terminates in branches on ventral side; 7’, tra- 
cheal trunk showing hairs; 


Z, peripheral zone of photogenic cell; P, photogenic 
granules V. nuclei of tracheal epithelium; D, dorsal layer cells; HW, hypodermis; 
C, chitin of sternite bearing hairs. 

8 Camera lucida drawing of part of cross section of photogenic organ of 
Photuris pennsylvanica; fixed in osmiec acid. The photogenic layer is one cell 
deep and the tracheal end cells are distributed over the dorsal and ventral sides 
instead of forming cylinders around the tracheae. 7, trachea from which 
branches B, arise passing into the two partly embedded tracheal end cells C; 
n, nuclei of tracheal e.c., F. fureation from which arise three tracheoles. Reduc- 
tion of the acid has not been sufficient to show the tracheoles throughout their 
entire course. A thick deposit of the reduced OsO, is shown about the fureation 
with traces of it in the cytoplasm, nucleus and region between the photogenic and 
tracheal end cell; H, nuclei of hypodermis. J, chitin; D, dorsal layer cells, note 
relative thickness of dorsal and photogenic layers; P, photogenic cells with gran- 
ules. 

9 View of portion of ventral surface of photogenic organ of Photinus sp. near 
maritimus, drawn by means of cam. luc. outline of active organ. SS, shaded area, 
corresponding to eylinder and showing only a faint light due to diffusion of hght 
from photogenic areas L. The latter correspond to region of photogenic cells. P, 
peripheral region of photogenic cells showing under some conditions a slightly 
greater intensity of light; V, nuclei of photogenic cells faintly visible as small 
shaded spots. 


466 


PHOTOGENIC ORGANS PLATE 3 
E. J. LUND 


THE JOURNAL OF EXPERIMENTAL zOOLOGY, vou. 11, No. 4 


467 


EXPERIMENTS ON DEVELOPING CHICKENS’S EGGS 


STEWART PATON 


Biological Laboratory, Princeton University 


The results of observations made upon the developing eggs of 
several species of selachians suggested the possibility of repeating 
these experiments, with certain modifications, upon the embryos 
of other vertebrates. 

The egg of scyllium canicula is admirably adapted for studying 
the primitive movements of the heart and myotom without in 
any way disturbing the normal relationship of the growing organ- 
ism, but on account of the difficulty of securing these eggs in 
large quantities a search was made for material which could be 
more easily obtained. Observations made upon several species 
of lizards, frogs and fresh water fish were for many reasons unsatis- 
factory and attention was then directed to the chicken’s egg. 

Experimenters have at various times removed the fertilized 
egg from the shell, and after detaching the embryo, have suc- 
ceeded in keeping the latter alive for some time not, however, 
exceeding a period of twelve hours. After many futile attempts, 
an Operative technique has been devised making it possible in 
the majority of instances to remove the fertilized chicken’s egg 
from the shell, place it in a glass dish containing fluid and return 
the receptacle to the incubator, when development under the 
conditions to be mentioned proceeds uninterruptedly. The egg 
freed from the shell becomes an object for observation and exper- 
iment, and not only the incidence of the primitive movements 
of the heart, but also many other interesting phenomena con- 
nected with the growth of the embryo may be observed and 
recorded. The technique employed in the operations is as 
follows: 

THE JOURNAL OF RXPERIMENTAL ZOOLOGY, vou. ll, No. 4 


469 


470) STEWART PATON 


All solutions are sterilized in the autoclave. Such a small 
quantity of fluid is lost during the process that in the majority 
of cases it is not often necessary to replace it but, if in certain 
cases it is essential, this may easily be accomplished if the fluids 
have been sterilized in graduated flasks. In order to shorten 
the operation as much as possible, and to minimize the risk of 
exposing the sterilized fluids to the air, the solutions are poured 
into dishes in which the eggs are to be placed, covered and put 
in the thermostat. The lids, which should be 5 mm. to 10 mm. 
larger than the dishes, rest upon collars of cotton held in place by 
string, and by this means free access is given to the air. Care 
should be taken that the cotton does not come into contact with 
the fluid in the dishes, and on the other hand, these collars must 
be sufficiently thick to raise the lids and give plenty of opportunity 
for the passage of air. Many embryos are killed by a deficient 
supply of oxygen. The cotton acts as a filter and prevents all 
bacteria except those within the shell from contaminating the 
fluids in the dishes. 

After the egg has been for the requisite amount of time in 
the incubator, it is removed, the shell is wiped off with 95 per cent 
or preferably 100 per cent alcohol, and with the aid of a pair of 
forceps that have been sterilized, an opening with smooth edges 
is made in it and the contents allowed to slide gently into the 
dish containing the fluid which should be of the same temperature 
as the egg. 

If the dish contains sufficient fluid the egg will quickly right 
itself so that the embryo is on top. Even slight differences of 
temperature seem to be fatal to the success of the experiment, 
and on this account, it is better to conduct the whole process of 
transferring the egg from its shell to the dish in some kind of 
warm chamber, such an one as can readily be constructed in the 
laboratory. When the egg is in the dish and covered, the proc- 
ess of development may be observed through the glass top with- 
out exposing the contents to the air. 

The earlier in development that the transfer is made, the greater 
is the chance of failure, but when the embryo has attained the 
size seen under normal conditions at about the 26th-27th hour 


EXPERIMENTS ON DEVELOPING CHICKENS’S EGGS A471 


of incubation, the operation is nearly always accomplished with- 
out serious injury to the growing organism. 

The action of a variety of fluids upon the embryo were observed 
and a brief account of some of the effects that were noted will 
now be given: 

The constituents of Ringer’s solution were tried singly and in 
combination. NaCl in varying strength from 0.5 per cent to 
2 per cent if uncombined, with other salts at once killed the 
embryo, but development although apparently taking place 
at a slower rate than normal, followed when the egg was placed 
in 0.7 per cent solution of NaCl to which 2.7 cc. of a molecular 
CaCl, solution was added. The record of all my experiments 
show that the rate of development is retarded in the sodium- 
calcium solutions, and the vitality of the embryo is also weakened. 

The extremely interesting fact in connection with this experi- 
ment is that the presence of such minute quantities of calcium 
is sufficient not only to protect the life of the embryo, but also 
to insure at the proper time the incidence of the cardiac move- 
ments. The same observation with practically similar results 
was made upon the eggs of trout. In the presence of these minute 
quantities of calcium, when combined with sodium, the regular 
and rhythmical pulsations of the heart begin, and are continued 
for several hours with increasing force and rate, but later the 
embryo dies. 

The calcium alone is not sufficient to insure the continuation 
of the developmental processes. In solutions to which KCl 
(6.3 ec. of a molecular solution) has been added, in addition to 
NaCl and CaC, growth seems to take place at a normal rate. 
The effect of MgCl, alone upon the growth of the embryo chick 
and its relations to the primitive movements of the heart has not 
been experimentally determined, but my impression is that the 
solutions containing MgCl, when not combined with CaCl, and 
NaCl are highly toxic. 

The effect of urea in strengths varying from 0.5 to 2 per cent 
either alone or combined with NaCl does not act as a stimulus 
to growth and the embryo soon dies when placed in solutions 
containing this substance. 


472 STEWART PATON 


Embryos detached from the egg and floated in any of the solu- 
tions named, live but a short time and the incidence of the prim- 
itive movements of the heart in these detached specimens is 
never observed. 

Caution should be observed in basing any deductions in regard 
to the immediate action of the various salts upon the incidence 
of the cardiac pulsations. It does not seem probable that the 
failure of the detached embryos to develop is purely the result 
of shock, but is due chiefly to the absence of nourishment supplied 
by the egg. 


BINDIXC 


==_ 7, JUN 2 2 1966 


QL The Journal of experimental 
1 zoology 

J68 

Viel! 

cop.2 \A\\ 


Biclogical 
& Medical 
Serials 


PLEASE DO NOT REMOVE 
CARDS OR SLIPS FROM THIS POCKET 


UNIVERSITY OF TORONTO LIBRARY 


Eye pant 


ht 
ere tha 


sua 


> 


2 
SS 
= 


SEES 


nes: 


mee 
oy 


ie ;