184 W. I. Robinson, 



Zaphrentidpe, although an approximation to it is characteristic 

 of certain of the Cyathaxonidse, typically of Petraia, in which the 

 septa, even in adult forms, do not reach the center of the deep 

 calyx. From the study of early stages of a large number of 

 Zaphrentidse and Cyathophyllidse, it has come to be recognized 

 that the development of the young of Tetracoralla has either 

 two, four, or six septa, all of which are complete and meet at 

 the center. If an ancestral genus were deduced from the 

 ontogeny of these forms it would be characterized by two or 

 more complete septa instead of by a few rudimentary septal 

 ridges. A strict analogy with the post-larval development in 

 modern forms would add the further conception of an earlier 

 disk-shaped corallum with a few complete septa. There is 

 proof that such a condition actually existed among Tetracoralla 

 as well as among Hexacoralla. The writer examined several 

 specimens of Pachyphyllum woodmani collected by C. O. Dun- 

 bar from the Upper Devonian (Lime Creek) of Iowa, and found 

 in them two small colonies which were cemented to the convex 

 surface of fragments of large brachiopod shells. By slowly 

 etching the inner surface of the shells with dilute hydrochloric 

 acid the structure of the protocorallite in contact with the shell 

 could be seen. As indicated in Plate I, Figure 10, the whole 

 colony diverges from a minute circular attachment such as a 

 hemispherical or disk-shaped sessile protopolyp might secrete 

 and very similar to the first lime-secreting stages described by 

 Duerden (1904) for Siderastrcoa and by Mavor (1915) for 

 Agaricia. 



The writer has found in his examination of abundant material 

 of Streptelasma from various Ordovician localities that the 

 earliest growth does not suggest a stage without septa. If such 

 a one was present as that described for S. profundum and then 

 was so speedily lost in later forms that it does not appear in the 

 succeeding species, S. cornkuhim and S. rusticum, it may be 

 seriously questioned whether it has phylogenetic significance. 

 Figures 5 h and 5 c give conclusive evidence that S. cornkulum 

 has a life history like that of other Tetracoralla and not like that 

 indicated by Brown for 5. profundum. 



As far as the present data go the hollow cone described by 

 Brown as being the first stage in the skeleton of S. profundum 

 is an exceptional and improbable condition among the Tetra- 



