J. E. DUERDEN 191 



represent a separate mutation, a distinct unit character, even though not 

 somatically recognizable. 



It is barely conceivable that any factorial alteration does not repre- 

 sent some separate, distinct degree of change rather than a continuity. 

 The former may be ever so small, but is still a discrete amount, as in all 

 Morgan's results. On the conception entertained by Castle of changes 

 up and down of the same factor the passage from one stage to the other 

 could be held to be continuous, while on Morgan's conception it would 

 be step-like. Somatically no difference need be apparent between the 

 two, but the underlying conceptions are different and are important from 

 a Mendelian point of view. It is not the degree of somatic change which 

 matters, as Jennings rather implies, but the manner in which it has 

 been factorially effected, and all considerations seem to favour a graded 

 or step-like process as compared with a continuous one. It may of 

 course be argued that a flight of stairs becomes an inclined plane if the 

 steps are made sufficiently small and numerous, and a structure in a 

 successional, degenerative phase may appear to make its descent down 

 a plane, instead of in a graded fashion. 



Though a continuous series of stages in the reduction of the claw 

 could be procured yet it does not follow that a like continuity is followed 

 by the factorial changes. They could just as readily be discrete germinal 

 changes, and it is held that the seemingly continuous degeneration of 

 the plumes and digits is of this character. So many influences are con- 

 cerned in the somatic expression of factors that it becomes no sure guide 

 as to the nature of the factor itself where comparatively small changes 

 are concerned. Changes may be discrete and successive and yet give 

 somatic continuity so far as our observation can extend. On the con- 

 ception of every somatic change being a step the term 7'ectigradation of 

 Prof H. F. Osborn is peculiarly appropriate compared with that of con- 

 tinuous variation ; for the latter has a somatic application only, whereas 

 the former gives the factorial significance in addition. Variations in 

 any definite direction may be continued without being continuous. 



So far then as concerns the retrogressive changes in the claw on the 

 fourth toe of the ostrich we may think with Castle, and those who agree 

 with him, that it is due^ to a continuous retrogressive change in the 

 factor itself; or we may hold with Morgan and his supporters that the 

 change is due to multiple allelomorphs, perhaps of the nature of so many 

 successive losses or changes in the main factor, or to the influence of 

 multiple modifying factors influencing the expressive nature of the claw 

 factor. 



