214 DONALD WALTON DAVIS 
great diversity in features other than the relation of stripes to 
mesenteries. Thus two have eight stripes; eight have ten 
stripes; twenty-three have twelve stripes; five have fourteen 
stripes; nine have sixteen stripes; one has eighteen stripes, and 
one has twenty stripes. Also thirty-six are diglyphic; twelve 
are monoglyphic, and one is triglyphic. Four, of which one is 
represented in figure 26, show no evidence of having undergone^ 
division, while the others give more or less evidence of regenera- 
tion, and a few show clearly the precise position of the plane of 
fission. Of the latter, some represent divisions in complete en- 
docoels, some in incomplete endocoels of the first grade, and 
one, of which a section is represented in figure 17, in incomplete 
endocoels of the second order. Most of the specimens are 
biradially symmetrical, but a number depart from this con- 
dition. The triglyphic specimens and diglyphic individuals with 
an uneven number of pairs of complete mesenteries cannot be 
strictly biradially symmetrical. It is clear, therefore, that in 
nearly all cases undivided specimens, or those in advanced 
stages of regeneration, have orange stripes corresponding in 
number and position with the complete endocoels and the 
highest order of incomplete endocoels. 
The number of incomplete endocoels of the first grade being 
almost invariably equal to the number of complete endocoels, 
the number of orange stripes is twice the number of pairs of 
complete mesenteries in the overwhelming majority of cases. 
Nevertheless, in any given population of this species one finds a 
large proportion of specimens showing an odd number of orange 
stripes. These are to be accounted for, in small part, on the 
basis of the irregularities in arrangement of the incomplete 
mesenteries described above (p. 211). Thus the exclusion of a 
pair of incomplete mesenteries of the first cycle from a primary 
exocoel owing to a division in an incomplete endocoel of second 
or lower grade adjacent to a pair of complete mesenteries, would 
reduce by one the number of orange stripes in the fully re- 
generated individual. Again, the unexplained duplication of 
ncomplete mesenteries of apparently first grade was shown to 
be associated with an extra orange stripe. By far the greatest 
