17 



whatever has yet been found in any member of the division Anasca. Smitt 

 and Hincks have both given an answer to this question, but as neither of them 

 had any idea of the existence of this sac, the quite calcified, arched surface and 

 the absence of the elevated margins were for them the most important differences 

 from the other Cheilostomata. At any rate both Hincks and Smitt declare that 

 these forms cannot be traced directly from the Membraniporidae, but that their 

 origin must be sought for in the genus Membraniporella. Hincks' states in con- 

 tinuation of the above-given citation: »The passage to the old Lepralian type 

 is not through such forms or through the genus Micropora, but through Mem- 

 braniporella in which the calcareous covering is an outgrowth from the margin 

 of the cell, overarching as it were the original membranous covering*; and a 

 similar view of this form's importance as connecting link between the two men- 

 tioned divisions is expressed by Smitt, both in his work on the Scandinavian 

 Bryozoa and in »Floridan Bryozoa*^. In the last work he says: »In the above 

 described Membraniporella Agassizii we have seen one of the most evident con- 

 necting links between the Flustrine and Escharine types*. Harmer has a similar 

 view of the importance of the acanthostegous forms as connecting link between 

 the anascous and ascophorous, and he seems in a preliminary paper ^ inclined 

 to suppose that the whole division Ascophora had an acanthostegous origin, while 

 in his main work*, he supports such an origin with certainty only for Umbonula 

 verrucosa and for forms related to this species. Harmer in contrast to the two 

 authors mentioned has given moie detailed reasons for his view, which we must 

 examine into here. Harmer has in fact observed, that the membrane, which 

 originally alone represents the frontal wall in Umbonula verrucosa and U. pavo- 

 nella, gradually becomes covered by a calcareous layer arising from the posterior 

 and lateral margins of the zocecium, which itself is covered by a membrane 

 (epitheca), and he therefore compares this process with that taking place in 

 Membraniporella or Cribrilina in which the original membranous frontal wall is 

 covered by two series of hollow spines. But while in the family Cribrilinidce the 

 single spines at the outside are connected by lateral twigs they are in Umbonula 

 according to Harmer fused together into a two-layered lamina, the calcified layer 

 of which corresponds to the basal (or inner) half of the spines while the mem- 

 branous cover corresponds to the frontal (or outer-) half. The author further 

 finds points of comparison, partly in the circle of pores, which appear on the 

 margin of the calcified frontal wall in U. verrucosa and U. pavonella, and partly 

 in the radial buttresses which separate every two of such neighbouring pores and 



22, p. 128; * 103, p. 21; ' 18, p. 13; ' 19, p. 295, 331. 



