Methods and Techniques 



31 



presentation of quantitative data is inade- 

 quate and perhaps irrelevant to answer the 

 basic question presented by the material to 

 be analyzed. What is the meaning, for in- 

 stance, of the "quantitative" resvilts described 

 in percentages of positive differentiation in 

 grafts? Luther ('36), for instance, has con- 

 cluded that there is a gradient of physiolog- 

 ical activity (Aktivitdtszustand) arovmd the 

 rim of the trout blastoderm on the basis of 

 the fact that differentiation occurs in a de- 

 creasing percentage of grafts as the material 

 for grafting is removed from progressively 

 greater distances from the midline of the 

 embryonic shield. If a particular factor, or 

 group of factors, or a certain quantity of such 

 factors necessary for differentiation, char- 

 acterizes the cells near the embryonic shield 

 should not every graft from that area dif- 

 ferentiate if the experiment is adequately 

 performed? What is the meaning, in terms 

 of the functions of the grafted cells, of the 

 fact that only 84% of the grafts removed 

 from a particular region have differentiated? 

 May not the significance of these results be 

 that the grafts have been removed in different 

 ways from zones of transition, or that they 

 have been implanted under differing experi- 

 mental conditions? In other words, may not 

 the quantitative variations in such results 

 indicate variation in the techniqvie of experi- 

 mental procedure as well as variation in the 

 activity of the tissue? Too many factors, 

 which need no enumeration here, are varied 

 in even such a simple experiment as the im- 

 plantation of a graft on the yolk sac, which 

 though in some ways is simple in others is 

 drastic and crude; and the experimental 

 procedure, which is manual and therefore 

 difficult to subi'ect to critical control, is prob- 

 ably differently performed each time. 



It can be of the greatest advantage to the 

 investigator to acknowledge that quantita- 

 tive variation in his results reflects his own 

 uncertainties as well as the accomplishment 

 of his embryonic material, if he wishes to 

 improve his experimental approach both 

 from the technical and the intellectual as- 

 pect. The value of statistical treatment and 

 its advantages in connection with the en- 

 deavor to attain the maximum precision in 

 analysis are particularly great in the case 

 of embryological material where so bewil- 

 deringly many variations are inherent in the 

 material and where so many sources of error 

 confuse the methods of analysis. But statis- 

 tical results must not be interpreted as final 

 to such a degree that they mask the weak- 

 nesses of the technical procedure where these 



actually affect the interpretation of results. 

 Embryology has not yet sufficiently matured 

 towards the perfection of its methods that 

 quantification can be its only desideratum, 

 and it may well be that the necessity to im- 

 prove upon these methods represents the 

 most urgent challenge immediately confront- 

 ing us. 



TECHNIQUES OF INTERFERENCE 

 WITH THE EMBRYO 



It was Wilhelm Roux who first had the 

 insight to appreciate the inadequacy of the 

 descriptive method, no matter how precise 

 the terms in which its results are couched, 

 to demonstrate what he called the Causal- 

 nexus of events, and to formulate a program 

 designed to analyze that causal relationship 

 within sequences of events which had al- 

 ready been so clearly expressed on an in- 

 ferential basis by His (see quotation on p. 

 17, Section I). It was the simplicity of 

 Roux' first statement of his problem that 

 enabled him to try to answer his question 

 with an apparently simple experiment: 



Fast alle aber fiihrten im Weiterfolgen zu einer 

 und derselben grossen Vorfrage, zu einer Alterna- 

 tive, von welcher aus die causale Auffassung fast 

 aller Bildungsvorgdnge in zivei wesentlich ver- 

 schiedene Bahnen gelenkt wird. Dies ist die Frage: 

 1st die Entwicklung des ganzen befruchteten Eies 

 resp. einzelner Theile desselben "Selbstdifferenzie- 

 rung" dieser Gebilde resp. Theile oder das Produkt 

 von "Wechselwirkungen mit ihrer Umgebung?" 

 Eventuell, welches ist der Antheil jeder dieser 

 beiden Differenzirungsarten in jeder Entwick- 

 lungsphase des ganzen Eies und seiner einzelnen 

 Theile? 



In der Beantwortung dieser Frage liegt meiner 

 Ein sicht nach der Schliissel zur causalen Erkenntnis 

 der embryonalen Entwicklung (1885; cited from 

 Ges. Abh. 2.-14). 



The question with which he was con- 

 cerned happened to involve the relationship 

 of the part to the whole, and happened to 

 revive in a new form the old controversy 

 between preformationists and epigeneticists; 

 but this was not its main significance. His 

 great contribution from the methodological 

 point of view was that he saw his problem 

 in terms of a single alternative and in terms 

 of clearcut relationships; relationships so 

 expressed that he could alter them in what 

 was to him a simple experiment. An em- 

 bryo or an embryonic part depends for its 

 capacity to differentiate on a mutual inter- 

 action with its surroundings, or it does not; 

 remove it from its surroundings, and its reply 

 should be unequivocal. 



