99 



as isolated internodes, in which ovicells are seldom present ; and it seems to me unsafe, in 

 most cases, to attempt to identify them with recent species. It may, hovvever, be noted that 

 C. elongata has been recorded in the fossil condition v ). 



Among recent species the present tbrm has considerable resemblance, in its broad inter- 

 nodes, its general mode of branching, and its black joints to the well known C. denticulala, 

 from which it appears to differ in the more inflated shape of its ovicell, in the position of this 

 structure near the middle of the length of the internode and in the greater profuseness with 

 which branches are developed in the fertile internode. 



Miss A. Robertson has called attention to the importance of the mode of branching 

 in the fertile internodes '-) ; and this seems to be really a character which can be used with 

 advantage in the discrimination of the species. These internodes may be regarded as the parts 

 of the colony in which the definitive characters of the species find their fullest expression ; the 

 more proximal parts of the zoarium frequently showing characters which are nearly uniform in 

 several distinct species. 



The erect sterns of a Crisia colonv, like those of some of the more delicate branching 

 Cheilostomes to which I shall have occasion to refer in the later parts of this Report, commonly 

 arise as branches given off by adnate rootlets 3 ). The present species has been found to furnish 

 definite evidence in favour of this view. In its basal region the internodes of the erect branch 

 consist of a very small number of zooecia, and have at first a diameter which is little greater 

 than that of the rootlet. In tracing the branch upwards the internodes are found to become 

 both broader and longer; the number of their zooecia increasing to a corresponding degree. 

 It is only when growth has gone on for some time that the definitive characters of the species 

 become marked. It thus follows that young branches, or the proximal parts of older ones, have 

 an appearance entirely different from that of fully grown specimens ; and in the absence of 

 other evidence might easily be referred to different species (cf. figs 3, 5). This is of course 

 true of the genus and not merely of the present species. As I have pointed out in an earlier 

 paper *), the presence of ovicells is highly desirable, if not essential, in order to arrive at a 

 definite conclusion with regard to the determination of a specimen. It is fortunate that ovicells 

 are well developed in much of the 'Siboga' material. 



The axis of a branch is usually sinuous, the convexities following one another in regular 

 alternate order, and each one typically giving rise to a new branch. The branches, like the 

 convexities of the stem, are thus alternately arranged. The form of the internode depends to 

 a large extent on the frequency with which transverse joints are developed. This is a very 

 variable character; and it accordingly follows that the internodes themselves are also variable 

 in length and in the extent to which they produce branches. 



Fig. 1 represents a portion of a colony in which the specific characters are fully developed. 

 Joints are not numerous and the internodes are thus very long. The longer internodes have 



1) Neviani, A., 1891, "Contr. Conosc. Bri. Foss. Ital.", "Bri. postpliocenici Livorno'', Boll. soc. geol. ital., X, fase. II. p. 133 

 (sep., p. 37); and elsewhere. 



2) Robertson, A., 1910, p. 230. 



3) Cf. Waters, 1910, t. cit., p. 239. 



4) Harmer, S. F., 1891, p. 128; cf. also A. Robertson, t. cit., p. 225. 



99 



