82 THE VOYAGE OF H.M.S. CHALLENGER. 



Chsetoceros dicladia, n. sp. (Plate VIII. fig. 1, and Plate XIX. figs. 7 and 8.) 



Frustulis in latere zonali transverse triplo longioribus, undulatis, in series per spatium 

 medio constrictum divisis ; quae binis utrinque cornubus breviusculis subclavatis quater 

 costatis connectuntur, setarum costis spinulosis. Inter Kerguelen et Heard. 



The species of Chsetoceros in which specimens of Dicladia capreolus, Ehrenb., have 

 been repeatedly observed, was often found in gatherings made between Kerguelen and 

 Heard Islands, near the polar ice-barrier in the Antarctic Ocean. 



The frustules have their zonal sides three times as long as broad, while the transversely 

 directed bounding line is undulating, so that an oblong space with a central contraction is 

 left between the adjoining frustules. The setae, which spring from the surface of the 

 valve, are somewhat short, and are provided with somewhat club-shaped outer extremities 

 which bear four thorny ribs. 



Bacteriastrum, Shadb. 



This very remarkable genus was instituted by Shadbolt in 1860 to embrace several 

 curious small sun-like organisms, which are surrounded by a few radiating protuberances 

 of considerable length, and which are very often met with in the stomach of Mollusca. 

 Shadbolt defines his genus in the following words: 1 — "Frustules awned, united into a 

 jointed conferva-like cylindrical filament, valves discoidal, with marginal radiating awns." 



Although there is a great analogy between the present genus and CJisetoceros, I am 

 of opinion, notwithstanding the opposite view of Professor H. L. Smith, that the differ- 

 ences are sufficient to warrant the preservation of both, apart altogether from the less 

 important question already referred to of the desirability of avoiding too great multiplica- 

 tion of the species belonging to any given genus. Among these differences may be noted the 

 following : — (1.) In Bacteriastrum the valves are always perfectly round, while in Chseto- 

 ceros they are generally oval ; (2.) in the former there is a greater number of awns, which 

 always radiate round the border, than in the latter ; and (3.) in seriate CJisetocerotidse 

 the frustules constantly interlace by the alternate crossing of the awns — a character which 

 is not found in Bacteriastrum. 



Only a few species belonging to this genus are yet known, and these were diminished 

 when Lauder, 2 after observing that Bacteriastrum curvatum, Shadb., constantly ter- 

 minated the series of Bacteriastrum furcatum, Shadb., 3 united the two forms under the 

 name of Bacteriastrum varians. This important observation, however, increases the 

 difficulty of determining the exact limits of specific forms, and, although among those 



1 Pritchard, op. cit., p. 803. 



2 Trans. Micr. Soc. Loud., new series, vol. sii. p. 7, pi. iii- figs. 1-6, 1864. 



3 = Adiniscus sexfurcatus, Ehrenb., Mikrogeol., pi. xxxv. B. 4, fig. 15; Adiniscus biseptenarius, Ehrenb. ; 

 Adiniscus bisodonarius, Ehrenb., loc. cit. See also Pritchard, op. cit., p. 863, pi. vi. fig. 26 ; Trans. Micr. S<><: 

 Lond., vol. ii. pi. i. figs. 1 and 2. 



