Polyzoa Entoprocta and Ctenostomata. 27 
(3) The rhizome, although it is not always adherent, never gives rise to vertical 
branches. 
(4) The number of tentacles is always eight. 
I find these characters constant in a large series of specimens from Bengal, 
Madras, Perak and the Tale Sap. 
In eastern waters B. caudata is characteristic of estuarine tracts in which the 
water has a lower salinity than that of the open sea. I found the species abundant 
at Koh Yaw in water of a specific gravity (corrected) of from 1-004 to 1-0085. It 
occurred (often with Vidorella bengalensis) on sticks and stones. I also took a speci- 
men on a worm-eaten fragment of a wooden pier at Port Weld on the coast of Perak. 
This place is situated up a creek, some distance from the open sea (Straits of 
Malacca), but the water is probably almost if not quite as salt as that of the Straits. 
Division PALUDICELLEA . 
1911. Palludicellina, Annandale, Fann. Brit. Ltd., Freshw. Sponges, etc., p. 186. 
1913. Paludicellea, Harmer, Siboga-Exp., mon. XXVIIIa, p. 43. 
Harmer {loc. cit.) includes in this division the following families : Paludicellidae, 
Victorellidae, Arachnidiidae and Nolelhdae (= Cylindroeciidae, aud.); w^hereas I 
have hitherto included only the Paludicellidae, Victorellidae and Hislopiidae — the 
last a freshwater family referred to by Harmer only in a foot-note. He supports his 
views as to the inclusion of certain marine genera with abundant evidence and 
clears up several anatomical points hitherto obscure, in particular by means of his 
excellent figures. These show that there is practically no difference in the general 
structure of the polypide between Cylindroecium and Victorella. In fig. 19 of his 
plate iv, for example, it is quite clear that the polypide of Cylindroecium (or Nolella) 
papuensis possesses a cardiac store-chamber and a well-defined single funiculus. 
Indeed, now that this evidence on anatomy is available, the grounds on which the 
family Cylindroecidae is separated from the family Victorellidae become rather flimsy. 
It is somewhat otherwise with the Arachnidiidae, in which Harmer follows Lop- 
pens^ in placing the freshwater genus Arachnoidea, Moore. His figure of the polypide 
of Arachnidium irregulare {op. cit., pi. iii, fig. 6) shows quite clearly that there is neither 
a proventriculus, nor a spherical chamber, nor a funiculus. This, of course, does not 
rule Arachnidium out of the division — the alimentary canal is merely in a simple and 
probably primitive condition ; but it does prove that Arachnoidea is by no means 
closely related to Arachnidium. Arachnoidea I would still retain in the family 
Hislopiidae on anatomical grounds, for although its anatomy is still imperfectly 
known, it certainly possesses a spherical chamber closely resembling that of Hislopia. 
This structure is not clearly indicated either in Moore's original sketch' or in Rous- 
selet's more elaborate figure,^ but I have seen it without a doubt in specimens 
mounted by the latter author and in Hislopia the horny lining of the gizzard remains 
1 Ann. Biol. Incust. Ill, p. 150 (1908). '■^ The Tanganyika Problem, p. 296, fig. (1903). 
Proc. Zoo/.Soc. London J907 I), pi. xiv, figs. 5, 6. 
