624 



SCIENCE 



[N. S. Vol. XXXVII. No. 961 



somes Ab and aB, as shown in the upper 

 part of the diagram. If this should happen 

 only occasionally it would fully explain 

 how it is that two characters borne by the 

 same chromosome tend to remain together, 



of a series of undoubted facts; and it is 

 certainly worthy of the most attentive fur- 

 ther examination. 



The three eases that have been consid- 

 ered have led us, step by step, to the border 



B 



i 



B 



D JO 



yet may separate so as to pass into differ- 

 ent chromosomes and hence into different 

 germ-cells. As shown in the lower part of 

 the diagram, a similar explanation may be 

 extended to much larger series of charac- 

 ters, the behavior of which in detail may 

 depend upon their arrangement in the 

 threads, or on the character of the twisting. 

 On the basis of this hypothesis an attempt 

 has recently been made by Sturtevant to 

 calculate from the observed results the de- 

 gree and character of the twisting of the 

 chromosomes, and the relative position of 

 the different specific elements within them. 

 This, admittedly, is a bold venture into a 

 highly hypothetical region. Its justifica- 

 tion is the pragmatic one that it "works." 

 The hypothesis gives us the only intelli- 

 gible explanation that has yet been offered 



line of research in this field. I have not 

 hesitated, in discussing the last case, to ad- 

 vance beyond the solid ground of observed 

 fact into a debatable and hypothetical re- 

 gion ; for it is by just such venturesome ad- 

 vances that new possibilities of discovery 

 are opened. We have much to gain and 

 nothing to lose by the use of explanatory 

 hypotheses that are naturally suggested by 

 the facts and help us to formulate them for 

 analysis, so long as such hypotheses are not 

 allowed to degenerate into dogmas accepted 

 as an act of faith, but are only used as in- 

 struments for further observation and ex- 

 periment. The " chiasmatype " hypothesis 

 is no more than this; and though it is di- 

 rectly suggested by observed facts it re- 

 mains for the present unproved. The more 

 general conclusions that have been indi- 



