158 



cancelli, consists of broad, convex, porous lobes which are similar to those of fig. 9. Fig. 8 is 

 a colony in which the central zooecia of the rays are specially elongated and have a very 

 distinctly biserial arrangement. But in fig. 10, representing a specimen which agrees with fig. 1 1 

 in its uniserial, slightly prominent zooecial rays, the lobes of the ovicell are not flat-roofed, 

 as in that specimen, but are just as convex as in fig. 9. 



It does not seem easy to give any other explanation of these differences than the 

 assumption that the form of the ovicell is variable. It appears probable, however, that the 

 ovicell has to adapt its form to the space which is available. Thus the colony shown in fig. 1 1 

 has a very small projection of its zooecial rays ; while it is growing on the convex surface of 

 a cylindrical Coral-branch of less diameter than itself. Owing to the latter reason, the basal 

 surface of the colony is very concave, the edges of the Lichcnopora being reflected round the 

 Coral. This position would naturally result in some stretching of the central region of the colony, 

 and would thus tend to prevent it from increasing greatly in thickness. The slight projection 

 of the zooecial rays would operate in the same direction by not allowing much vertical room 

 for the growth of the ovicell ; and it is perhaps owing to the combination of these two conditions 

 that the roof of the ovicell is unusually flat. In fig. 8 the conditions are very different. Although 

 this colony too is growing on a cylindrical object of less diameter than its own, its basal lamina 

 is for the most part free; being only attached along a narrow line to its substratum. There is 

 thus no interference with the growth of the colony in vertical height. As the peristomes are, 

 moreover, unusually long, there is plenty of room for the ovicell to attain a greater vertical 

 height than in the other case ; and the roof is accordingly very convex, with a correspondingly 

 large perforated surface. 



In some cases there is evidence of the development of what I have previously termed 

 secondary ovicells ; that is, ovicells which are formed at a late stage in the growth of the colony. 

 Thus on slide 108. AB., one of the colonies has a normal ovicell in its central region; and, in 

 addition, a much smaller ovicell, with an ooeciostome which has developed, in a more superficial 

 position, between the outer ends of two or three of the zooecial rays. The ovicell shown in 

 fig. 9 may also be regarded as a secondary one, since it is being developed in a colony which 

 has already reached a considerable vertical thickness. 



The ooeciostome varies little in form, although its position is not constant. In fig. 8 two 

 ooeciostomes (o) are seen, at two points on the margin of the ovicells, of which it may be assumed 

 that at least two are present. Each is a short, simple tube, terminated by a nearly circular 

 ooeciopore. It is usually circular in section, but is occasionally compressed. In most cases the 

 ooeciostome is found on the central side of one of the zooecial rays, or between the central 

 ends of two of them. The number of ovicells is apparently always small (not more than two or 

 three); but the development of cancelli generally prevents the line of junction of two contiguous 

 ovicells from being visible; and the number of ovicells has to be inferred from that of the 

 ooeciostomes. 



Certain colonies are found to have reached a large size without showing any tracé of 

 an ovicell. A specimen of this kind, on slide 354. A., has a diameter of 3.2 mm. Its central 

 region is depressed, and the entire surface which is not occupied by zooecial rays is covered 



15S 



