230 SEMI-CENTENNIAL OF TORREY BOTANICAL CLUB 
colonies and consist in the latter of a group of four or five central 
cells surrounded by, respectively, twelve or eleven peripheral cells. 
In this group again the splitting or doubling of the spinous 
projection of the cell in its incipient stages foreshadows the further 
development of this character through the whole series. The 
incised or bifid tips and the doubling of the spines appear in 
graded stages of development which suggest very strongly that 
the species have been produced as end members in series of con- 
tinuous variants. Under P. Ehrenbergii and its synonyms we 
find included by most authors forms in which the degree of lobing 
varies widely. In some forms the cells are only bluntly bifid 
(FIG. 26). In others there is every degree of inequality between 
the two points of the bifid spine, suggesting that the forms may 
have arisen from the Diactinia by the budding off of an accessory 
tooth on the main spine, or at another point on the body of the cell 
rather than by splitting the tip of the spine itself (FIG. 25). 
Nitardy’s treatment of the group recognizes the depth of 
lobing and the degree of separation of the points as the principal 
basis for distinguishing the types and his two species, with a 
variety under the first, form what it seems to me is in part at least 
a natural series. 
In his first species, P. incisum Hassall, however, Nitardy 
includes forms with the spines very unequally cleft (ғіс. 24), one 
half frequently much more strongly developed (14, pl. 5. f. 7 
and pl. 7. f. 8) along with others іп which the spines are very short 
and even blunt (’14, pl. 7. f. 6,7, and гг). There is no adequate 
evidence that all these forms could come from one mother colony. 
In the variety P. incisum var. Rota Nit. he includes a natural 
group with the spines much more strongly developed and as a 
rule quite equally bifid. 
n the second species, P. lobatum Nit., he includes what he 
regards as the handsomest forms in the whole genus, with strongly 
developed lobes deeply bifid (pl. 5. f. 4). This is plainly Braun's 
and Cooke's P. Rotula Ehrenb. The five species and four varieties 
recognized by De-Toni also show characteristic differences in the 
cell form and lobing but reliable figures are not available for 
grouping them in an evolutionary series. 
I have never seen in P. Ehrenbergii the 1 + 5 + 10 configura- 
