208 PHYSIOLOGICAL GENETICS 



of segmentation (Castle and Gregory) ought to be mentioned, 



and the whole series of let hals which act upon early development. 



Nearly related are the cases in which mutant genes influence 

 directly early pattern formation in vertebrates, though the 

 exact type of influence is not always know'n. These are the cases 

 of monstrosities in vertebrates which we have already reported 

 (see page 46), some of which demonstrate a genie action at a very 

 early time of development, probably upon the organizer directly 

 (see Lehmann, page 50). To this category belong also all such 

 lethal genes as are known to kill the embryo in very early stages, 

 probably by upsetting primary pattern formation, e.g., the cases 

 analyzed in Drosophila by Redfield (1926). 



Another group of most important facts has been derived from 

 the study of a phenomenon called homocosis. We have described 

 on page 36 the development of the mutant aristapedia in 

 Drosophila in which the last segment of the antenna is trans- 

 formed into a tarsus. We saw that this differentiation of a 

 homologous organ takes place at the time of leg differentiation 

 which occurs prior to the time of differentiation of the antenna. 

 Here, then, a mutant gene changes an embryological process 

 by shifting its initiation to a different point in time. There are 

 similar cases in Drosophila which, however, are not so clear as this 

 one. There are the mutations bithorax and tetraptera (see 

 Astauroff, 1929) in which the metathorax assumes a number of 

 characters of the mesothorax, including the formation of wings 

 instead of halteres. There cannot be any doubt that the explana- 

 tion w r ill be of the same type as for aristapedia. There is also 

 the mutant proboscipedia (Bridges and Dobzhansky, 1933). 

 Here the mouth parts undergo a number of variable changes 

 which bring their structure near to the type of biting organs of 

 other insects. The modified lobes may show resemblance to 

 the labrum, maxillary palpi, antennae, and tarsi, the homology 

 of all these organs being apparent. Again, the same type of 

 interpretation applies as in the case of aristapedia. These facts 

 are of greatest importance in linking gene action with develop- 

 mental processes. They point out that the mutant gene in 

 question does not control a localized action of a special type. It 

 controls a general, quantitative process which leads to a special 

 result, because it shifts a process of pattern formation into a 

 phase in which it ought not to be. To be more specific, at the 



