RHIZOSOLENIA CURVATA 435 



the sub-Antarctic Zone range up to 14-5° C. in summer (Deacon, 1933, p. 206), and 

 that R. curvata has only been recorded in very small numbers at five stations where the 

 surface temperature was over 7-6° C. and never with certainty above 8-3° C, the force 

 of this argument becomes even more apparent. 



In the eastern South Pacific the sub-Antarctic Zone reaches its greatest width, 

 through the agency of known geophysical factors, and it is here that R. curvata shows its 

 greatest meridional range in terms of distance north of the Antarctic convergence. In 

 terms of latitude of course, its northerly limit is reached in other areas where the con- 

 vergence itself may lie as much as 800 miles farther north than it does in longitude 80 W. 



GENERAL BIOLOGY 



The only inaccuracy in Zacharias's original description of R. curvata (1905, pp. 120-1) 

 arises from the limitations of his material. He speaks of the species as being slender, 

 and goes on to state that all his specimens (from a single sample) had diameters ranging 

 from 51 to 57ju. and varied between 850 and 1000/x in length. Karsten (1905, p. 97) 

 gives 48-80/x as the diameter range and 572-900/^ as the range in length. In the 

 endeavour to establish the essentially sub-Antarctic habit of this species I undertook a 

 very large number of measurements by the methods described on p. 420, this method 

 of attack having been suggested by Wimpenny's recent work (1936). The variation in 

 dimensions, especially in length, was enormous, and the work was concentrated on the 

 more constant and significant diameter measurements. Even here I found an extreme 

 range of 20-135/x and a normal range of 31-1 17/x, while the mean of over 3000 measure- 

 ments was 627/x. Specimens 1500^ in length were not uncommon, and one 2000/x long 

 was measured. The grosser individuals can certainly not be described as slender, for 

 they do not reach the extremes in length very often, and are only rivalled in bulk by a 

 few warm-water members of the genus such as R. robusta Norman, and by rarer sub- 

 Antarctic species. 



Apart from this question of the proportions of the individuals, nothing can be added 

 to the early descriptions of the species by Zacharias and Karsten. Karsten 's description 

 is to be found under the synonym R. curva (1905, p. 97), and he gives admirable figures 

 of the details of the cell structure (Taf. xi, figs. 2, 2 b). An attempt has been made to 

 depict a typical individual in Plate XIV, fig. 1, while the outlines in figs. 2-4 are 

 intended to give an idea of the size range of the species. 



The typical habit of R. curvata is solitary; but since binary fission is by far the 

 commonest method of reproduction, chains of two or three individuals are by no means 

 rare. The largest number of individuals I have seen in one chain is six, and chains of 

 more than three are rare in our samples. The way in which the frustules are united in 

 catena is indicated in Plate XIV, fig. 8. Another very peculiar type of colonial attachment 

 has been rarely observed in this species, represented by fig. 9 in the plate. I have 

 termed this process "rafting". It will be seen that the individuals adhere together by 

 their long axes, and unlike the individuals in the chains, which have never been known to 

 differ from each other by more than 2/x in diameter, the individual frustules in these 



