ió 



forma typica. 



Joints stalked or sessile, cuneate to almost circular, ;i median joint occasionally elongate, 

 compressed; in stalked joints apex often raised as a small cushion, stalk long, central filaments 

 exposed (fig. 7). 



forma digitata. 



Joints sessile; lower broad, bearing above elongate, compressed, fingerlike joints, often 

 numerous and fringing upper margin of large joints (fig. 9). 



forma iindulata. 



Margin of joints undulating, thickened, recurved, grooved (fig. 10). 



This species is most closely allied to //. Tuna, from which it is often impossible to 

 sei)arate it by ontward form alone. But an examination of the internal structure shews that 

 the two species differ in the details of the central strand and the peripheral cells. 



( )n a cursory examination of a longitudinal preparation of the apex of a joint, certain 

 marking on the walls of the filaments suggest that the cavities are connected with one another 

 by means of pores, in a manner similar to that of H. macroloba and H. incrassata. By means 

 of transverse sections however and of carefully teased preparations it is seen that the walls 

 are quite intact at these points and that the fusion really takes place in the ordinary way 

 described in //. Tuna. The peculiar markings are nothing but the result of the close adhesion 

 of the peripheral cells to the central filaments and of these filaments to one another. This 

 adhesion of the filaments prevents them from falling apart easily in preparations and constitutes 

 a difference between this species and H. Tuna, where the adhesion is ver)- slight. This appear- 

 ance is figured by Professor Askenasy (1. c. tab. III fig. 8j under H. macroloba. He kindly 

 sent me the original plant for examination and it clearly falls into the H. cuneata of my 

 classification. The fusion of filaments which takes place at the apex is not however represented 

 in Professor Askenasy's figure. 



The other point of difference between H. cuneata and H. Tuna lies in the peripheral 

 cells, which in //. cuneata are connected along their side walls for \'i to ' „ their length, (fig. 

 14), whereas in H. Tuna the connection varies from 1 / 25 to Vio °f their length (fig. 6). 



In consequence of the close adherence of the peripheral cells in H. cuneata, they do not drop 

 apart in decalcified and teaseel preparations, as is the case with other species; it is therefore 

 necessary to cut transverse sections of a joint, in order to see the peripheral cells in side view. 



In examining the peripheral cells in snrface view of specimens of H. cuneata, it is well to 

 remember this connection of the side walls, for if the material be dry the side walls are often crumpled 

 and, as seen trom above, give the appearance of much greater thickness than is really the case. 



It is thus seen that the distinction between //. cuneata and H. Tuna depends only on 

 the relative amount of adhesion between the central filaments, and the extent of connection 

 between the peripheral cells. It is therefore not improbable that a series of forms may be found 

 which wil! unite the two species. Indeed, one such plant exists among the specimens of Halimeda 

 brought from Rangiroa by Professor Agassiz in the " Albatros/' . This has the somewhat thick, 



