IOI 



internodes are of a form entirely different from that of the yoimger internodes; being narrow 

 and with the peristomes hardly developed: — which may, however, indicate that the projecting 

 portions have been lost with the atrophy of the corresponding polypides. But the reference to 

 the present species is justihed by the fact that the more peripheral internodes of the same stem 

 take on a broad form and may become fertile internodes bearing ovicells of the typical character. 

 The number of zooecia in the internode remains low in this colony, even in the peripheral parts. 



The ovicell l ) of the present species has the form shown in figs i, 7. lts inflated part 

 begins very suddenly, and the most prominent region is situated near its distal end. The 

 ooeciopore is of the form found in C. dentictilata, and is a small transverse slit, visible only 

 when the ovicell is seen from its distal end (fig. 6). In a single case I have found twin-ovicells 

 (fig. 8), similar to those described by d'Orbigxy ~) in C. patagonica. In rare cases, apparently 

 where the growth is not very vigorous, the ovicell is less inflated distally, and may be simply 

 pear-shaped. 



The number of branches borne by the fertile internodes is typically three, as shown in 

 fig. i; but one of these may be suppressed; and in a single instance (fig. 4) I have found 

 four branches. 



I have referred the specimens from Stat. 77 with some doubt to the present species. 

 In slide 37. R. 1 the ovicells are delicate and pear-shaped, not specially inflated. This may be 

 due to the fact that they have been formed in parts of the erect sterns, near their origin from 

 rootlets, which are still slender and have not assumed the adult characters. The specimens in 

 question have a considerable resemblance to C. denticulata, var. gracilis, described by Busk 

 from the Philippine Islands 3 ). 



The following notes may be added with regard to the specimens, recorded above, which 

 are not from the 'Siboga' Collection. They serve, in some cases, as a justification of the synonymy, 

 where references to the Memoirs cited will be found. 



The specimens from Torres Straits agree in all essential respects, including the characters 

 of the ovicells, with the 'Siboga' material. 



The same may be said of the Oueensland specimens. The material obtained from the 

 Australian Museum consisted of unsorted fragments which were part of the remains of the 

 Collection described by Prof. Haswell. There is thus a strong presumption that the Crisia 

 belongs to the species described by that author as C. terrac-rcginac. In his account of this 

 species Haswell gives the number of the zooecia in an internode as 16 — 22, the complete 

 internode figured having 18. The occurrence of an even number of zooecia is somewhat unusual *) 

 in Crisia. Haswell further states that the branches arise from the 8 th to the i3 th zooecium 

 (i. e. from the 4 th to the 7 th of one side); and that the joints are black. It may be noted, in 



1) It is uncertain to what extent the absence of ovicells indicates that a colony is of the male sex. Miss Robertson' (1903, 

 "Embr. Crisia" Univ. Calif. Publ. Zool. I, p. 116) states that C. eburnca (= C. occidcntalis Trask, cf. Robertson, 1910, p. 239), is 

 "certainly dioecious", fhough this did not appear to be the case in all the species investigated. 



2) d'Orbigny, A., 1839, "Voy. Amér. Mérid.", T. V, 4e Partie, "Zooph.", PI. I, fig. I : — see also PI. XII, fig. 13 in my 

 paper on Crisia, already referred to. 



3) Busk, G., 1886, l.cit. 



4) In C. cribraria Osburn (1912, p. 216) describes the internodes as normally constituted of an even number of zooecia. 



IOI 



