[KEPORT.ON THE DIATOMACE.K. 93 



of plane spinous or denticulate surfaces, and on this account W. Smith, acting on the sug- 

 gestion of Thwaites, 1 separated from the genus Melosira those forms which are connected 

 in the manner last named, grouping them into the genus Orthosira. This scheme has, 

 however, not been generally adopted, the character in question not being regarded as of 

 Generic significance, nor can it ever be defended unless the jrenus Melosira should 

 ultimately become so rich in species that its separation into two parts should become 

 necessary for the facilitating of systematic work. 



Melosira sol, Ehrenb., var. nov. (Plate X. fig. 3, Plate XVII. fig. 13, and Plate XXI. 

 fig. 7.) 



Among the Melosira? contained in the Challenger collection, the present forms, which 

 were collected in the neighbourhood of Kerguelen Island, are of great elegance. The two 

 complete frustules, which are united to one another by their valval sides (Plate XXI. 

 fig. 7), although having very close affinities to Melosira sol, Ehrenb.,' 2 cannot be regarded 

 as identical with the latter. The description given by Pritchard 3 of Melosira sol, as 

 determined by Ehrenberg, corresponds with the characters of the variety now in question 

 with the exception of the fact that, in the latter, one additional feature has to be added 

 — namely, the circumstance that between the radiating lines or plicae, which run centripe- 

 tally from the periphery and quickly disappear, more minute radial lines occur and also 

 soon vanish. Although this is not clearly indicated on Plate XXI. fig. 7, owing to the 

 insufficiency of the magnifying power employed, it is shown in Plate X. fig. 3 under a 

 power of 800 diameters. 



With the comparatively imperfect instruments used by Ehrenberg, it is not surprising 

 that these fine fines escaped his observation. 



Melosira costata, Grev., var. nov. (Plate XXIII. fig. 5.) 



Among the pelagic organisms collected in the Bay of Yedo, the very elegant series 

 of frustules represented in the present figure were not unfrequently met with. That 

 these frustules belong to the genus Melosira cannot be questioned, and the presence of 

 the longitudinally disposed costules, which are contiguous to the surface of junction of two 

 adjoining valves, seems to point to their identity with Melosira costata, Grev. On 

 comparing, however, Greville's figures 4 with the present one, the following differences 

 may be noted : — (1.) The smooth areas around the different frustules of Melosira costata 

 are almost equal, and in many more restricted than in the present form, where they are 

 much larger and unequal ; and (2.) in the typical Grevillean species these smooth areas 



1 Thwaites, Ann. Nat. Hist, March 1848. 



2 = Gallionella sol, Ehrenberg, Mikrogeol., pi. xxxv. A. 22, fig. 12. This species rivals Melosira arenaria, 

 Moore, in size, Ealfs, Ann. Nat. Hist., vol. xii. pi. ix. fig. 4. 



3 Pritchard, op. cit, p. 819. * Trans. Micr. Soc. Lond., n. s., vol. xiv. p. 77, pi. viii. figs. 3-6, 18G6. 



