REPORT OF THE DIRECTOR I9IO 237 



in which the size of the dot at the supposed points of origin is made 

 commensurate with the importance or age of the group. If the 

 orad radial groups and orad interradial series were developed as 

 branches of one of the others, a still more primitive form would be 

 represented by text figure i6. As soon as each b.v. of figure i6 

 had given rise to two other b.vs. we should have a condition like that 

 shown in text figure 17. At about this stage we may suppose that 

 the initial members of the orad radial and orad interradial series 

 made their appearance. 



The possibility of a primitive or nepionic stage like that shown in 

 text figure 18, where these b.vs. may be supposed to have appeared 

 simultaneously and not at the angles but on the middle of the sutures, 

 should perhaps not be dismissed from mind. In this case new b.vs. 

 would be formed by branches, simultaneous or alternate, on each 

 side of these and also on the sutures, if the first were so formed. 

 There would be as many natural groups as there were sutures in this 

 case and we should have no difficulty in accounting for the upper 

 young members on the common sutures of the radials. A form like 

 this appears rather complex for a beginning though even here we 

 might credit acceleration with changing a primitive series of three 

 into a simultaneous group of three and credit natural selection with 

 the placing of such a group at the middle of the sutures where it 

 could function to best advantage and not interfere with new mem- 

 bers protruded at the corners. These oldest b.vs. are, however, not 

 now on the middle of the sutures, but dififerences in rate of growth 

 might have displaced them. It has seemed to the writer that the 

 evidence from this form and also that from Cleiocrinus is on the 

 whole strongly against the latter view. Utilizing a recommenda- 

 tion of the late Professor Rowland, we may keep both hypotheses in 

 mind, but for the present may give the former a position of from 80 

 to 95 points of credibility on our mental sliding scale of 100 and wait 

 for the evidence yet to come from future research. 



Grozvth lines. The plates so far discussed all show minute 

 growth lines, but these are only noticed between the ridges. On 

 finding a plate in which the canal coverings had been worn away and 

 the number of ridges therefore doubled, we might be able to use this 

 character to aid in distinguishing between the canal bottoms and the 

 true external plate depression between the canals. The canal floors 

 would show no growth lines. 



The radianal. In normal plate development the plate angles 

 occur where there is least resistance to plate extension or at the 



