BRYOZOA. 357 



placing Acanthocladia in the immediate vicinity. My arrange- 

 ment is quite different. I find that there exists between Tham- 

 niscus and Polypora a very close chain of intermediate links. 

 Examples of T. dubius (the type of the genus) now before me, 

 clearly show that non-poriferous dissepiments occasionally con- 

 nect the branches. In cell structure there is no difference be- 

 tween T. dubius and such species of Polypora, as P. tubercu- 

 lata Prout, P. cestriensis and P. spinulifera. In species of 

 Thamniscus the branches bifurcate in precisely the same man- 

 ner, though as a rule more frequently than they do in Poly- 

 pora. Such a slight difference can, of course, scarcely be taken 

 into account as of even generic importance. Now as to the 

 presence or absence of dissepiments, we must, because of their 

 constant development in so many forms, regard them as a 

 character of considerable value. But here again the ''vanishing 

 lines of nature" come into play, giving us an inkling of geneo- 

 logical relation. For instance, in Polypora halliana Prout, we 

 have a true species of the genus, but upon comparison, it is 

 found that the species is closely related to P. maccoyana, and 

 this to P. distinct a and P. gracilis (Prout.) The last is brought 

 into close relation with Thanmiscus divaricans by an unde- 

 scribed form from the Keokuk group of Keutucky. In P. halliana 

 and P. maccoyana there are from two to six dissepiments to 

 each bifurcation, in P. distincta and P. gracilis, generally two, 

 while in T. divaricans there are rarely two, generally one, and 

 sometimes none. The same mav be said of T. ramulosa* 



* T. ramulosa belongs to another line of development tending the same way through 

 P. cestriensis and P. tubercitlata. It is my opinion, and good evidence can be brought 

 to support it, that the majority of the Carboniferous and Permian species of Thamni- 

 scs (including the type species) were developed from Devonian and early Carboniferous 

 species of Polypora, such as I have mentioned, and not from Silurian and Devonian 

 species now referred to Thamniscus. When it comes to the latter, we have but little 

 evidence to show that they were derived from Polypora, yet the inference is fair 

 that they were, and that Polypora always had within itself the tendency to assume the 

 characters of Thamniscus. If this is admitted then we can admit that Fenestella also 

 had a continual tendency to run into Polypora. Again many problems present them- 

 selves which can only be explained by supposing that the changed form reverted back 

 into the parent stem. (A Fenestella gradually assumed the characters of Polupora and 

 vice versa). Hence, in all formations from the Niagara to the Coal Measures, there were 

 intermediate forms between Polypora and Fenestella. These species, the evidence leads 

 me to believe, were more often derived from typical species of the genera than from 

 each other. 



