Febeuaby 4, 1898.] 



SCIENCE. 



159 



goes its way and neither takes account of 

 the other. Others do not share this view. 

 They look upon the soma and the germ as 

 very intimately bound together — associated 

 in much the same way as the stolon and 

 the hydranths of a hydroid are. 





{^3 



^e 1* trv. - p 1 a. s TTi 



The germ plasm at the tip of the stolon 

 gives rise at intervals to hydranths very 

 much as the germ plasm of other animals 

 gave rise at intervals to somas. In both 

 cases the germ plasm is modifiable to a 

 limited extent by such modifications of the 

 soma as result from starvation, reduction 

 in general vigor, or the secretion of specific 

 substances affecting the germ plasm. In 

 both cases developing soma and germ may 

 be simultaneously modified by external 

 agents, so that while the developing genera- 

 tion C is being changed, future generations, 

 D, E, etc., are being potentially changed in 

 the same fashion because in the germ plasm. 

 For example, although I do not know that 

 the experiment has been tried, a dense solu- 

 tion might produce a spindling soma C and 

 a spindling stolon, so that even if the solu- 

 tion were diluted again a spindling soma D 

 would rise. By other agencies we may 

 modify the protoplasm at the tip of the 

 stolon so that it will thenceforth tend to 

 produce modified hydranths. Just as the 

 formation of the stolon and hydranth are 

 parts of one developmental process, so are 

 phylogenesis and ontogenesis parts of one 

 process. Every ontogenesis is dependent 

 upon a preceding phylogenesis and every 

 phylogenesis is dependent upon a preced- 

 ing ontogenesis. 



I have said that morphogenesis seeks to 

 explain the development of the individual 



and the race. When have we explained 

 development ? We have explained any ef- 

 fect when we know its immediate causes — 

 that is to say, the essential conditions under 

 which the eifect occurs. We seek, then, to 

 know the essential conditions under which 

 phylogenesis and ontogenesis occur. 



What general methods must we employ 

 to learn these conditions of development? 

 There ai-e two principal methods : one is 

 the method of observation of the differences 

 in development under known dissimilar 

 conditions ; the second method, more appli- 

 cable and more certain, is the method of 

 experiment. 



I may illustrate the way in which simple 

 comparative observation and observation 

 with experiment throw light upon the pro- 

 cesses of development. The simple obser- 

 vation that in the tunicate Doliolum the 

 sexual buds, detaching themselves from the 

 ventral stolon, crawl over the surface of the 

 animal to the dorsal stolon to arrange them- 

 selves there in regular order might have 

 taught us that one of the conditions direct- 

 ing individual development is response of 

 the different parts of the developing indi- 

 vidual to stimuli coming from other parts 

 of the organism. On the other hand, the 

 experiments of Driesch upon the gastrula 

 of sea-urchins enforced the fact vividly, 

 for he found that even after the mesenchyme 

 cells had been hopelessly mixed up by shak- 

 ing the gastrula they still migrated toward 

 their destined place. So, too, the observa- 

 tion of the decline of the descendants of 

 famous men might have led us to the law 

 of regression toward mediocrity as a con- 

 dition of phylogenesis just as Galton's ex- 

 periments with sweet peas did. In the 

 foregoing cases there is, however, a precis- 

 ion and decisiveness about the experimental 

 method which marks it as one to be pre- 

 ferred where applicable. In addition to 

 experiment, an allied method applicable 

 especially to the study of variation is that 



