O’Mrara—Report on the Irish Diatomacce. 251 
succeeding authors. Ralfs and Rabenhorst, however, abandoning the 
generic distinction, have relegated the several species of Orthosira to 
the genus Melosira. Heiberg, on the contrary, recognises the distinc- 
tion of Thwaites, but includes the species of Cyclotella in the genus 
Orthosira, and establishes a new genus, Paralia, to receive the single 
species Orthosira marina (Wm. Sm.), on the ground that the frustule 
possesses an elevated keel similar to that which characterises the 
genus Lysigonium. It is a question, then, whether the distinction of 
Thwaites should be recognised, as most authors since his time have 
done, or discarded, as Rabenhorst has considered it ought to be; and 
the following observations of Pfitzer seem to supply a reasonable 
solution. Having referred to the original distinction, he adds, ‘A 
far more important distinction exists in the mode of developing Auxo- 
spores. Although the Orthosire in their mode of growth agree 
thoroughly with Melosira, have the same structure of the primordial 
cell, and the same mode of cell-division, they differ in this respect, 
that in the process of spore-formation from a single cell, the valves of 
which are pushed away from one another, the contents enveloped in a 
mucous investment come out free, and are then, without being in con- 
tact with the mother-cell, developed{into a single Auxospore in which 
the firstling cell is so situated that the plane of division crosses that of 
the mother-cell, whereas in Melosira it is parallel to it. Thwaites 
had observed this feature in the case of Orthosira aurichalcea, and 
Smith refers to the peculiarity as an important generic distinction. 
But only one species had been observed in this aspect, and so it was 
questionable whether all the Orthosire obeyed the same law. Fr. 
Schmitz has succeeded in proving this in respect to another species, 
Orthosira roeseana (Rab.), = O. spinosa (Grev.)’? Ueber Bau und 
Entwicklung der Bac., p. 184. If then the mode of developing Aux- 
ospores be regarded, as I consider it ought to be, of importance as a 
generic distinction, the conclusion is inevitable that the genus Ortho- 
sira should not be merged in Melosira, as Ralfs and Rabenhorst have 
treated it. And also, forasmuch as in those species of Cyclotella in 
which the formation of Auxospores has been noticed, the daughter-cell 
is parallel to the mother-cell, for this reason, as well as on the old 
ground of distinction, the species of Cyclotella should not with Heiberg 
be included in the genus Orthosira. 
Orthostra arenarta, (D. Moore.) Fresh water. 
Frustules very large; cell-cavity sub-spherical. Spines on juncture 
surfaces short, broad, and close. Striw, on side view punctate, 
radiate, stronger at the margin, and loosing their radiate arrangement 
as they approach the centre. Strie on front view punctate, trans- 
verse. (Pl. 26, fig. 6.) 
' Wm. Sm., B.D., Vol. ii., p. 59, Pl. lii., fig. 334. Heiberg, De 
Danske Diat., p. 31. Ralfs, in Ann. N. Hist., Vol. xii, p. 349, 
Pl. ix., fig. 4.—Melosira arenaria, Kiitz. Bac., p. 55, T. xxi., fig. 27. 
Rab. Siissw. Diat., p. 14, T. ii., fig. 5. 
R. I. A. PROC., SER. II., VOL, II., SCIENCE. 21L 
