HETEROCRINIDAE 
95 
The evidence on this point is remarkably decisive. Among- up-wards of 400 specimens of 
Form D from the American Devonian and Lower Carboniferous in which these plates are 
visible, there is not a single case in which the left anterior radial is not completely and widely 
separated into two triangular segments, with the anterior and left posterior radials meeting 
between them. These are distributed as follows: Helderbergian i ; Oriskany i ; Onondaga 4; 
Hamilton 10; typical Lower Burlington 6; Upper Burlington 37; Keokuk 58; the remaining 
300 are from the Knobstone, or New Providence shale, equivalent to the Lower Burlington 
in part, in which at several localities the flattened cups of united radial plates and detached 
anchylosed basals occur in great numbers. 
Thus it is seen that the two structures, the axil-arm and the divided segments, arose 
together, and continued in conjunction to the end of the family. The presence of the two 
segments unconnected proclaims the axil-arm almost without exception, which is the impor- 
tant fact because nearly always in evidence. 
The axil-arm system of branching as above described is modified in the Gotland species, 
Calceocrinus pinniilatus of Bather ; here the arms are composed of two main trunks with 
cuneate brachials, from each of which a pinnule is given off at the longer side, thus forming 
two similar rows of alternating pinnules, or ramifies, after the manner of Decadocrimis. This 
is so different from all other species in the family that it will undoubtedly, as Bather has sug- 
gested, form the type of a new genus. 
The branching of the median arm is also a feature of considerable variability. Dr. Bather 
in the Crinoidea of Gotland, page 72, stated that “ the 1 . ant. arm gradually bifurcates at a 
greater distance from the cup, and eventually ceases to branch altogether.” Examination of 
the American material, with the important additions now available, shows that the modifica- 
tion in this character is not attended by any such regularity. 
In the Ordovician Form A the arm is either simple or bifurcates once at varying heights. 
In Form C, while the arm in most species as herein described is unbranched, there is one 
from the same locality in which it bifurcates three or four times. In the Carboniferous species 
the tendency in regard to the branching of this arm is the reverse of that above supposed ; 
Halysiocrinus dactylus from the Burlington, for example, branches once and then again in 
one ramus, producing 3, or perhaps 4, finials ; while in H. nodosus, the most prominent 
species from the later Keokuk, represented in the collection by numerous specimens, the 
median arm divides twice and three times, and sometimes even once more, to a maximum of 8, 
or 10, finials. Thus there is a progression in this character from a median arm simple or 
with a single bifurcation in the Ordovician, to one of three or more bifurcations in the 
Carboniferous. 
In a similar way the primary branches of the lateral arms increase in number from the 3 
of Form B to the 5 or more axil-arms of Forms C and D. 
So also in regard to the form and proportions of the flattened cup as seen from, the left 
anterior side. In Forms A and B it is usually subquadrangular, about as high as wide (pi. 28, 
figs. 5a, 14) ; it changes in many species of Form C to a trapezoidal form -wider than high, 
and widening toward the proximal margin (pi. 29, figs. Ja, 12a) ; in mature specimens of 
Form D the cup has spread out transversely until the margin at the hinge may exceed the 
height as 3.5 to 2 (pi. 30, figs. 8, 9, 10). 
Thus there were at least six definite phylogenetic modifications from the Ordovician to 
the Carboniferous, which were more or less concurrent, viz : 
1. Stem from unsymmetric lateral to median position. 
2. Number of basals from 4 to 3. 
3. Number of lateral arms from 3 to 2. 
4. Segments of 1. ant. R from connected to disconnected. 
5. Median arm from simple to many-branched. 
6. Lateral arms from equal bifurcation to axil-arm system. 
