C. L. Edwards 
285 
rosette like Fig. 25 with median latei'al bars the apposed terminal processes fused, 
a button like the above would result. In 1.5G, from Samoa, there is a marked 
tendency toward the formation of these patterns. 
But even when boles are present the ends of the rosette branches curl out and 
project freely, only very rarely forming the smooth contour often found in Types a 
and h oi H. floridana. In my statistical series 6% I'ave one bole (Fig. 26); 1 7^ 
two holes and only a fraction of 1 ' /o "'ore titan two holes (Fig. 27). By special 
searching outside of the statistical series rosettes were found with as many as six 
holes completely formed (Figs. 27 b — d). Even then thk fully develoi'EU 
PERFORATED PLATES OF H. FLORIDANA (Figs. 39 — 41), WITH A MEAN NUMBER OF 
18 — 15 HOLES, ARE DISTINCTLY DIFFERENT FROM ANY FOUND IN H. ATRA. So also 
the rosettes differ in the two species. In H. atixi they are more elongated, the 
parts more nearly at right angles to one another and the whole more delicate, 
while in H. floridana they are stellate, appear heavier, often thicker at tbe centre 
and the branches iiave swollen ends (cf. Figs. 23 — 27 d with 28 — 33). Sometimes 
rosettes like those of H. atra are found in H. floridana, as well as stages inter- 
mediate to the stellate form typical of the latter species. The pos.sessioii of 
peculiar perforated plates and in almost all cases the different kind of rosettes, 
are among the chief characters which have led me to separate and re-establish 
H. floridana as a distinct species. 
Number per sq. cm. 
One hundred of these spicules weio counted in the field (■2s(j.mm.) of a 1 in. 
ocular and a A in. objective, and thus to each sq. cm. of surface there could be 
50,000 spicules. The possible error in this calculation may be placed at 50 /^. 
As a general ride these spicules are much more crowded in H. floridana {c^. p. 263) 
but there are exceptions like 30 of that species, which has scarcely more spicules 
than the least number in H. atra. 
G. Calcareous Spicules of the Ambulacral Appendages and the 
Differentiation of Pedicels and Papillae. 
From 13 Holothurids, 204 dorsal ambulacral appendages were taken, and from 
3 specimens, 53 ventral appendages. Certain characters were not determinable, 
or absent, in some cases. 
a. Form. 
In form 18 (8'8 7^) are cylindrical, 26 (12'8 7,7 conical, but chieHy because of 
much contraction, 160 (78'4 °/^) of the dorsal ambulacral appendages were not 
determinable. From those determined it may be assumed that a considerable 
majority in the bivium are conical in H. atra, instead of cylindrical as in H. flori- 
dana. 
b. Suckers. 
Suckers are present in 31 (15"2 7o). rudimentary in 1 ("5 °/_), absent in 
121 (59-3 7o) »nd not determinable in 57 (25 7o)- 
