106 BULLETIN 114, UNITED STATES NATIONAL MUSEUM. 



from a lower number of rows to a higher can not be denied, and it 

 may therefore be admitted as possible to regard getulus or JiolbrooJci 

 as ancestral. But if we consider that a maximum of 23 rows and a 

 minimum of 19 is characteristic of 5 of the 8 forms and common at 

 wldelv so^arated localities in the 3 (which, as will appear later, are 









































\ 

 \ 



\ 

 \ 

 \ 

 \ 



\ 



• 



\ 

























\ 



\ 

 s 



\ 











. 















-1 



1 

 t 



1 

 1 

 1 

 1 

 1 























/ 

 / 

 / 

 / 

 / 

 / 

 / 

 / 





















y^ 



y 



, 

















/ 

 / 

 1 



/ 

 / 



/ 























' 









































W) S 



fl s 



S iz; 

 PI " 



w 

 o 



'^ 2 



o a 



^ OQ 



03 Q 



<N (M 



evidently specialized in other structural features), it looks as if this 

 were a fundamental character of the getulus group, and as if forms 

 possessing a lower maximum and a lower minimum were distinctly 

 removed from the ancestral condition. 



A glance at the diagram (fig. 25) of extremes and averages of ven- 

 tral plates shows that the eastern forms differ decidedly in this respect. 



