210 DONALD WALTON DAVIS 



endocoel on the boundaiy between new and old parts. In all of 

 them the number of observed orange stripes corresponds with 

 the number of complete and first order of incomplete endocoels 

 lying wholly within the old part. Thus the specimen of which 

 a section is shown in figure 9 had three orange stripes in the 

 positions indicated diagrammatically in figure 36. The other 

 unpaired specimens showed five and seven orange stripes, 

 respectively, corresponding with the number of undisturbed 

 endocoels normally bearing stripes. Each of the pairs had three 

 orange stripes in one member and seven in the other similarly 

 situated. The indication * given by these nine specimens is 

 that an orange stripe lying in a space cut by a plane of fission 

 is lost. 



In regeneration following division through an endocoel occupied 

 by an orange stripe, a new stripe is finally developed in the 

 bounding endocoel in addition to those formed in endocoels 

 lying entirely within the new region. Were this not true, com- 

 plete endocoels lacking orange stripes would be common, 

 whereas they are actually exceedingly rarely, if ever, found. 

 The formation of new bounding mesenteries, when division 

 occurs in one of the lower grades of incomplete endocoels, has 

 already been described (p. 195), the result being that the mesen- 

 teries of a pair enclosing the boundary between new and old 

 regions are of the same grade. The bounding endocoel conse- 

 quently may be definitely designated as belonging to a certain 

 cycle. When this bounding endocoel is of the second incomplete 

 grade or of a lower order, no orange stripe is formed in it. The 

 endocoel on the boundary between new and old regions, is, 

 then, no exception to the general rule that orange stripes are 

 formed in endocoels of the complete mesenteries of the first 

 cycle, but not of lower cycles. In support of this statement I 

 may cite a number of examples. Anemone no. 82, table 7, a 

 section of which is shown as figure 17, had divided through in- 

 complete endocoels of the second grade or possibly lower. This 

 specimen, when killed, had ten equidistant orange stripes, 

 equaling the number of endocoels of the first two cycles. Evi- 

 dently there were no orange stripes in the bounding endocoels. 



