ASEXUAL REPRODUCTION IN SAGARTIA 221 



Orange stripes 



The factors concerned in the deterixdnation of the number of 

 stripes apparent on a given individual at a particular time are 

 such that any mere enumeration of orange stripes in a set in- 

 cluding recently divided individuals is of highly questionable 

 significance. Nevertheless, a study of a tabulation of counts of 

 orange stripes may yield suggestive results and may be made 

 an occasion for pointing out further the effect of processes of 

 asexual reproduction upon the external appearance of specimens. 



Counts of orange stripes made upon four lots of specimens 

 collected at Woods Hole are given in table 14, and the totals 

 obtained by adding the first three are plotted in figure 42. In 

 figure 42 are given, also, data published by Davenport ('03). 

 If we look at the solid line in figure 42, representing a summary 

 of my counts of July and September (lots 1, 2 and 3), we see 

 that, beginning with twelve, the higher even numbers of stripes 

 are represented by more individuals than the odd numbers. This 

 is in harmony with the greater number of completed regenerations 

 in the classes with higher numbers of stripes. The odd numbers 

 here probably represent chiefly uncompleted regeneration. Be- 

 low twelve, the odd numbers of orange stripes are most abundant. 

 This is particularly true below eight. Among those lower num- 

 bers, recently divided specimens form probably the greater pro- 

 portion of individuals. The predominance of odd numbers 

 among these classes is in all likelihood chiefly due to those speci- 

 mens which have formed no new stripes. With the relative 

 positions of orange stripes and mesenteries of different orders 

 in mind, it is obvious that divisions in two complete endocoels 

 or two incomplete endocoels of the first grade would give usually 

 odd numbers of old stripes; divisions in one complete endoco^l 

 and one incomplete endocoel of the highest grade vrould ordi- 

 narily give even numbers of old stripes; while divisions in other 

 planes might give either even or odd numbers. The relative 

 frequency of divisions in different planes (table 9) is such as to 

 make the expected ratio of odd to even numbers of old orange 

 stripes approximately 4:3. It is possible that during the 



