﻿Manchester Memoirs, Vol. xlix. (1905), No. \%. 25 



the pits in several series on the walls of the wood cells, 

 was of wide distribution in Palaeozoic times, extending to 

 Pteridospermeai, and even to some Ferns, as well as to 

 Cordaiteoe. It is now almost limited to the Araucarieae, 

 certainly an ancient group, which may reasonably be sup- 

 posed to have retained a primitive character. 



The Araucarieae, however, show clear relations to 

 the rest of the Conifers, in the organization of their 

 cones, and other characters, and it seems difficult to 

 attribute a different origin to them from that of the 

 remaining sub-orders, with the possible exception of the 

 Taxaceoe. The latter family, however, itself shows, on 

 other lines, some approximation towards Cordaiteae, 

 through the isolated type Ginkgo, as indicated by anato- 

 mical characters, and by those of both male and female 

 flowers. 



On the whole, then, although the indications are some- 

 what scanty, yet, such as they are, they appear all to point 

 to an affinity between the Conifers generally and the 

 Palaeozoic Cordaitese, and through them with the Pterido- 

 spermeae, and so in the last resort with the Fern-phylum 

 in the broadest sense. I, at least, find it impossible to 

 believe that the Coniferae are an unnatural group, and that 

 their various tribes can have been derived from totally 

 different sources. 



The final conclusion, then, to which, as I think, we 

 are led on the basis of the evidence at present open to us, 

 is that the Gy mnospermous seed-plants generally are best 

 regarded as in a broad sense, monophyletic, the whole 

 class having been ultimately derived from the Pterido- 

 spermic stock, itself allied most nearly to the Ferns. In 

 a narrower sense, however, we may still speak of the 

 probable origin of the Gymnosperms as polyphyletic, for, 

 as we have seen, there is reason to believe that their 



