TUBARIUM OF CEPHALODISCUS HODGSONI. 51 
apparently lateral ostia are explicable on the ground that a lateral branch is in each 
of these cases in the initial stages of its growth. The openings are oval, sometimes 
circular, and average between 3 by 2 and 3°5 by 2°5 mm. across. Each opening has 
four or five spines around its margin. Five representative figures are given of the 
ostia and their spines, since the number and distribution of the spines of the test of 
Cephalodiscus will probably prove to be among the features useful for distinguishing 
the various species (figs. 16-20, plate 4). 
The spines are simple, or forked, or even trifid. The length of the free part 
varies considerably, usually within the limits 5 to 15 mm. The width varies from 
°6tolmm. The central axis of a spine is darker than the general test, and can be 
traced for some distance into the latter. The total length of a spine, including this 
embedded base, varies from 12 to 25 mm. Most of the spines arise from the edges 
of the ostia. 
The successive increments by which the spines are built up remain clearly 
visible, and a study of the fine dark lines which mark off the different strata 
shows that the forking of a spine is not due to dichotomy, but that one of the 
two limbs of the fork is the primary spine and the other one, although it may be 
equal to it in size, or even greater, has been secondarily attached to its side (see 
fig. 21, plate 4). 
In connection with the probable freedom of locomotion possessed by the 
polypides of Cephalodiscus and the bearing which this freedom has upon the mode 
of increase of the test, it may be noted here that in specimen C a few pieces of 
Flustra (Carbasea) have been caught up among the branches of the colony. In 
some places the flattened branches of the coenoecium of the polyzoon have been 
completely buried in the test, but in others the strata of the test are continued 
along one surface, or along one edge, as a thin film, or as several films, with here 
and there a spine of the usual structure. This coating of the foreign object 
extends to a distance of 20 or 30 mm. away from the nearest hollow part of the 
tubarium occupied by polypides. It seems clear from this that not only have the 
polypides the power of moving about over the surface of their own test—one 
cannot conceive how the terminal increments of the spines are put on if they 
cannot do so—but they are also able to roam about upon neighbouring foreign 
objects, which they smear over with their profuse secretion. 
Specimen C is further of interest in that in certain parts of the colony a 
stoloniferous polyzoon allied to Bowerbankia is found embedded in the test. The 
relations of the zooids of the polyzoon to the Cephalodiscus tubarium are such as to 
suggest that when they were flooded by the not yet solidified material of the test, 
they had time to dispose themselves so that their tentaculated ends were flush 
with the surface before the material hardened. Some of the zooids are normal to 
the surface of the tubarium, others are oblique, but all have their openings flush with 
the surface. 
