Keokuk Species of Agaricocrinus. — Gordon. 259 
All the examples noticed have regularly twelve arms and oc- 
cur in the lower crinoid bed (No. 2). If as stated by the above 
authors the modifications of the base denote geological suc- 
cession we have here a continuation of the upper Burlington 
form with a slightly increased concavity. In this case it would 
seem proper to consider this form specifically distinct from the 
one to which it is now referred. On the other hand should 
these modifications be deemed varietal only, then no good 
reason would remain for retaining A. whitifieldi Hall, since 
that form is more nearly related to the typical A. americanus 
Roemer than is the one under discussion. 
An apparent exception to the rule given by Wachsmuth 
and Springer occuts in the case of A. springeri White, from 
Indiana, which is said to have the base truncate but not de- 
pressed. The type specimen was obtained from the "terrace 
drift, west bank of the Wabash, Clinton, Vermilion Co., 
Ind." Its geological horizon is therefore uncertain, and we 
are inclined to think it belongs to the upper Burlington. 
Specimens of the subexcavate form may be seen in the col- 
lection of Mr. N. K. Bunket, Mr. L. A. Cox, and the writer. 
In its normal state the A. americanus Roemer is said to 
have twelve arms, three upon each of the posterior rays and 
two upon each of the others respectively. The number of 
arms is not constant however ; additional arms frequently 
appearing upon the postero-lateral and autero-lateral rays. 
We have never observed more than two arms upon the anter- 
ior ray. In examining a number of specimens showing more 
than the normal number of arms we have been impressed 
with a certain uniformity of structure which it is difficult to 
believe to be merely accidental. Rarely or never do we find a 
division of one pair of rays only. The addition of arms on 
the postero-lateral rays is always accompanied, so far as we 
have observed, by an increase of arms in the antero-lateral 
rays. Of five examples before us, four have sixteen arms and 
one seventeen. Three have the arms regularly arranged — 
four on each postero-lateral ray and three on each of the 
antero-lateral rays. The other two are somewhat abnor- 
mal, one having two additional arms on one of the antero- 
lateral rays, making seventeen in all, while the other up- 
on one side has both additional arms placed upon the antero- 
