36 University of California Publications. IZoology. 



Corymorpha palma appear similarly to those in Clava up to the 

 number of twelve at least. In embryos of Titbnlaria the first 

 tentacles to be developed are two in number and opposite. The 

 distal tentacles develop in fours. In regenerating, however, 

 both proximal and distal tentacles appear in considerable num- 

 bers simultaneously and in a peculiar manner. Possibly the 

 conditions under which regeneration takes place have determined 

 the method of regeneration in this case. There is a wide 

 discrepancy here between this method and the one employed by 

 the related Corymorpha, where the distal tentacles appear in fours 

 and as buds, not ridges. The appearance of the tentacles in 

 Hydra has been variously reported. In one case, in a Hydra 

 bud two opposite tentacles appeared simultaneously, then a single 

 dorsal one, followed by a ventral one, completing the quartette. 



Another form is worthy of note in this connection — the 

 Bimeria robusta of San Pedro. The tentacles of the budding 

 bydranths in this species appear in fives, there being usually two 

 complete whorls of five tentacles each, and another with from one 

 to three. In the adult condition the whorls are almost indistin- 

 guishable except by the habit of carrying the tentacles alternately 

 raised and depressed. I venture to suspect that the related 

 Perigonimus may develop on this plan, a view supported by 

 Allman's figure of P. repens. There maybe here a criterion for 

 the determination of larger groupings of far more moment than 

 the fixed or free habit of the gonophores. 



But it is quite impossible at present to establish certain 

 phylogenies on this basis. I shall refer more in detail to the 

 growth of tentacles in discussing the species concerned. 



The exceptional shape of the proboscis of H. milleri should 

 be emphasized, as it is of some taxonomic importance. It is 

 usually long, and especially prominent in the fertile bydranths 

 where it is set off from the body of the hydranth by a narrowed 

 base and is swollen distally. Even in cases of extreme con- 

 traction, which are common among the sterile bydranths, this 

 swollen extremity preserves its identity (PI. II, Fig. 18). In no 

 case have I seen what might be called a typically conical proboscis. 

 The shape is clearly intermediate between the conical and the 

 trumpet shaped types, the latter of which is so characteristic of 

 the Eiidendriidce . 



