io6 
A general feature may be noted that in the great majority of cases the two uppermost 
predominate, and that when the number is reduced to 2 pairs or to 1 pair, these occupy the 
same space as 3 — 6 rows. The larger numbers 5 — 6 were certainly less frequent than the 
smaller numbers. 
There is not in this case any confusion between what might be counted to points and 
what to crown, for in no case did we see what in other species is called a crown. It is also 
striking that an examination of scores of anthocodiae disclosed hardly any case of intermediate 
spicules between the points. 
The supporting bundle is enormously developed. Its largest spicules may attain to a 
length of 6 mm. and project for 1.5 mm. These large spindles are densely covered with minute, 
blunt thorns, except for a length of 0.6 mm. at the projecting tip, which is smooth. Very 
marked in the large supporting-bundle spicules of the colony from Station 240 is the disappearance 
of the red colour from the proximal third of most of the largest. The same tendency is to be 
recognized in the yellow colonies. 
Other spicules. The spindles of the upper cortex are often up to 4 mm. in length. They 
are covered with large, blunt, truncate and sometimes compound warts, and they lie transversely. 
Many show an abrupt narrowing at one end. In the lower cortex and canal-walls there are 
numerous stouter and shorter spindles, exceeding 1 mm. in length and bearing crowded, rough 
tubercles often compound. Large triradiates and multi-radiates also occur, besides bracket-like 
and quite irregular forms approaching the stellate type. A few curiously starfish-like forms occur, 
and triradiates with one arm exuberantly branched. Many much smaller counterparts of the 
large types occur. There are also occasional long, narrow spindles. 
Within the hemprichi-gxowp the specimens agree best with D. gigcinteci , though this may 
not be obvious at first sight. Many of the anthocodiae show but one pair of spicules in each 
point, which suggests approximation to D. earned , D. doederleini , and D. punicea. From the 
first they are at once separated by the massive spicules, which are visible from a distance and 
lie in very regular transverse rows. This conflicts too seriously with the original description by 
Wright and Studer (p. 196), where it is said: “The stem and branches . are leathery 
and thickly packed with fine spicules. The latter are not recognisable by the unassisted eye and 
lie scattered confusedly in several layers' ’. Both from D. d.oederleini and from D. punicea our 
specimens are sufficiently separated by the enormous strength of the supporting bundle. 
Deduction : The specimens agree more closely with D. gigantea than with any other 
Glomerate known to us, and not least in showing quite an unusual degree of variability. Among 
the important features of agreement the following may be noted : 
(1) the very characteristic plump hemispherical masses closely beset with polyps; 
(2) the supporting bundle of great strength and showing many spindles of over 4 mm. (the 
maximum size mentioned for D. gigantea) and projecting for 1.5 mm.; 
(3) the large spindles of the supporting bundle densely covered with minute, blunt thorns, 
except for the last half millimetre at the projecting tip, which is smooth ; 
(4) the spindles of the upper cortex frequently reach 4 mm. in length, are covered with large, 
blunt warts, and lie transversely; 
(5) the spicules of the lower cortex and of the canal-walls ; 
