ASEXUAL REPRODUCTION IN SAGARTIA 209 



pass. That these openings affect the position of the division 

 plane is possible but quite unlikely. If they are of significance 

 in the location of the fission plane, their distribution should 

 show some relation to the frequency of fission in different 

 planes. As to this, no evidence is at hand. 



The fact that cinclides are commonly located in other positions 

 as well as on the orange stripes is evidence against the idea 

 suggested by Davenport ('03, p. 143) that the presence of the 

 stripes may ''be considered as a case of warning coloration?" 



Orange stripes in fission ahd regeneration 



During the process of fission, an orange stripe lying in the 

 endocoel cut by the plane of division has been observed, in a 

 few cases, to be divided by the tearing of the column wall. After 

 the division the narrow border of orange along the edges of the 

 pieces must be promptly absorbed, for in no case observed has 

 there been found early in regeneration an orange stripe on the 

 boundary between the old tissue and the new. On the contrary, 

 nine specimens in early stages of regeneration which were killed 

 and sectioned after counting the orange stripes give clear evi- 

 dence of lack of stripes in positions almost certainly containing 

 them before the division. At the time these were killed none 

 showed any orange stripes in the newest regenerating area. Of 

 the nine individuals, six represent the paired products of three 

 divisions (nos. 11, 13 and 14, table 3) and three are unpaired 

 specimens (no. 30, table 6, a specimen represented in figure 9, 

 and one other). All of the paired and two of the unpaired 

 specimens showed division through two complete endocoels. 

 The third unpaired specimen had divided through endocoels of 

 the first incomplete grade. All divided in planes where orange 

 stripes were to be expected. Furthermore, the paired specimens 

 showed orange stripes in such numbers and positions as to 

 occupy all regions, outside of the new area, in which they nor- 

 mally occur, indicating that the orange stripes were probably 

 fully formed before the division occurred. In none of these 

 cases was there any evidence of an orange stripe marking the 



