ACTINIAN BUNODEOPSIS GLOBDLIFERA. 313 



nor a histological identity of the ui^per part of the median streak of the filament with 

 tlie lower or gl andular streak proper." Embryologieal considerations also point to the 

 conclusion that a difierent origin must be assigned the apical median part of the 

 mesenterial filament in the uppermost region as compared with the whole filament 

 below. From the researches of McMurrich, H. V. Wilson (1888), Appellof (1900), and 

 myself (1899), there is good reason to suppose that the former is a downgrowth of the 

 stomodjeal (>ctoderm, according to my interpretation, to meet the upgrowing endodermal 

 filament. 



In Bunodeoims the ciliated bauds quickly disappear along with the two lateral meso- 

 glceal lobes which supported them, and the simple filament remaining is altogether different 

 from the middle lobe of the more complex filament, although it corresponds with it 

 in position, representing as it does the termination of the free edge of the mesentery. 

 The unmodified endodermal tissue, which constituted the intermediate streak of the 

 complex filament, and covered the middle lobe laterally, is likewise no longer represented 

 in the simple lobe (PI. 26. tig. 16). 



The cellular constituents along the front and sides of the simple filament, however, 

 most closely recall those of the glandular streak of the compound tilament, consisting 

 mainly of ciliated supporting cells and coarsely granular gland-cells, with the addition of 

 a few thick-walled nematocysts. The posterior part of the filament on botli sides is 

 constituted of much shorter cells, wliich seem to be all ciliated supporting cells, no 

 nematocysts nor gland-cells being recognizable. Histologically, and probably physio- 

 logically, the hinder region of the simple filament most nearly approaches the ciliated 

 bands, though no morphological identity seems possible. 



At first the mesenterial mesogloea as it enters the posterior part of tha filament remains 

 swollen as in the trilobed filament, but in the lower part of the filament it becomes 

 flattened and rapidly thins out at each side. The anterior surface of the mesogloea 

 is provided with a layer of extremely delicate muscular fibrils. 



In the lower region of the polyp the filaments and free edge of the mesenteries are 

 greatly folded and contorted, but they by no means crowd the gastro-ccelomic cavity 

 (PI. 25. fig. 4). 



No gonads were developed within any of the jiolyps examined. 



In its essential featiu'es, B. glohulifera closely resembles B. strumosa and B. antilliensis, 

 which were incompletely described in the paper (1897) "The Actiniarian Family 

 Aliciidae." An important external specific distinction is that the columnar evaginations 

 in the first are smooth and thick-walled throughout, while in the two older species they 

 are thin-walled in places, and the large nematocysts occur in restricted thickened bands 

 or tubercles. In sections this dift'erence is indicated by the practical uniformity in 

 thickness of the ectodermal evaginations of glohulifera (fig. 5), and their alternating 

 narrow and broad outline in strmnosa and antilliensis, according as nematocysts are 

 absent or present. 



SECOND SERIES. — ZOOLOGY, VOL. VIII. 46 



