54 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



of the dorsal coelome. The youngest representatives have numerous small cirrus 

 sockets without sculpture, in alternating rows, concealed basals, no free dorsal margin 

 of the radials, and a fairly small radial cavity; the radial pits are shallow or obliterated. 

 Gislen observed a rosette in Semiometra impressa. He said that the recent form 

 Notocnnus may properly be included in this series, as it is undoubtedly nearly related 

 to Loriolometra, from which it differs in its shallower and wider radial coelome pits, by 

 having a relatively larger centrodorsal cavity and, in full grown individuals, by the 

 absence of sculpture on its cirrus sockets. The later forms of this series advance to- 

 ward group a of the Macrophreata (see Part 3, page 63, under Macrophreata). The 

 lettensis group with its fine radial pores corresponds as regards this character to the 

 family Asterometridae in which the radial pits are deep, but very narrow. 



The conometrid series, relatively poorly represented among species in which the 

 radial pentagon is known by the genera Amphorometra, Placometra, Conometra, and 

 Jaekelometra, is more numerously represented within the genus Glenotremites by the 

 anglesensis and pellati groups. This series is distinguished from the preceding one 

 by the absence of radial coelomic pits. Among recent forms it has its analogy in the 

 Thalassometrida, sensu stricto. 



Gislen said it is very remarkable what a large number of species of which only the 

 centrodorsal is known are found in both these series, contrasting with the conditions 

 in the solanocrinid series. In the genus Glenotremites of the solanocrinid type only 

 five species are found (the exilis group). But there are 30 species that can certainly 

 be included hi the two series now under discussion. This is due in the case last men- 

 tioned to the fact that the elements of the radial pentagon are loosely connected both 

 with each other and with the centrodorsal. In the recent Thalassometrida the corre- 

 sponding ossicles very soon fall apart when boiled in caustic potash. A feature to be 

 noticed is that in the recent Mariametrida the centrodorsal and the radial pentagon 

 are very firmly united with each other, exactly as appears to have been the case in the 

 fossil solanocrinids. 



It seemed scarcely probable to Gislen that the notocrinids could have descended 

 from any of the comasterids or solanocrinids. In the last named, to be sure, we some- 

 times find diverticula from the dorsal coelome between the centrodorsal and the 

 radial pentagon, indicated by shallow branched meandering furrows. These, how- 

 ever, are entirely different from the deep and broad perpendicularly penetrating pits 

 which occur in Loriolometra. Because of their depth they traverse almost the whole 

 of the centrodorsal even in full-grown individuals it is easy to assume their pro- 

 longation into the stem in stalked young. 



Although there are a good many similarities in regard to general features between 

 Archaeometm and Amphorometra, Gisle"n said it is not very probable that the cono- 

 metrids are derived from either of the two families first mentioned. This is made still 

 less likely by the fact that in the Upper Jurassic we meet with a form which it seems 

 might fairly well be considered a primitive ancestral form of the Thalassometrida and 

 Conometridae in general, and of Placometra in particular. 



Gislen said we therefore reach the conclusion in these two cases also that these 

 two families possibly represent one, or perhaps two, special offshoots from the 

 pentacrinid stock, which in that case have attained the comatulid type by independent 

 paths. 



