4 i6 DISCOVERY REPORTS 



completed a circumpolar cruise, was similarly dealt with. The results, which are de- 

 scribed at length in this paper, appear to confirm the initial hypothesis that this species 

 is a valid indicator of the southward limit of the sub-Antarctic influence in the surface 

 water of the Southern Ocean. 



R. curvata Zacharias was first described by two independent workers in 1905, and 

 confusion in the synonymy has resulted. Zacharias 's description of the species, from a 

 single sample obtained some 300 miles south-west of Cape Horn, was published in July 

 of that year (Zacharias, 1905). His description is followed by the statement that it ap- 

 pears to him to be new to science, but that he puts forward this idea with "the reserva- 

 tion that it may well have been discovered in the material collected by the 'Valdivia', 

 which is even now being worked up" (pp. 120-1). This reservation was well justified, 

 for Karsten's manuscript with the description of the species under the name R. curva 

 (Karsten, 1905, p. 97, Taf. xii, fig. 2) had already been received by the editors in April. 

 I was unable to discover the exact date of Karsten's paper, and in my report of 1934 

 (p. 161) I used the name that Karsten had given. 



When it became evident that the species was of peculiar ecological interest, I was led 

 to study the subsequent literature, and found that Zacharias (1906, p. 557) claimed 

 priority for curvata. Further, the justice of this claim is acknowledged by Karsten in 

 his Atlantic Ocean report (1906, p. 164) under the subheading of R. semispina, another 

 member of the genus. It thus appears that the name R. curvata Zacharias must be 

 retained, and R. curva Karsten is to be regarded as a synonym. I have had the advantage 

 of being able to discuss the point with Mr N. I. Hendey, who has recently been engaged 

 in a systematic study of our diatom collections, and he has confirmed this view. 



Before passing to a discussion of the early records of the species, I should state that 

 throughout this paper I have followed the hydrological definitions of the surface water 

 laid down by Deacon (1933, 1937). 



Previous records of the occurrence of R. curvata have been made by Zacharias (1905, 

 1906), by Schimper in Karsten's reports on the material collected by the 'Valdivia' 

 (Karsten, 1905, 1907), by Mangin (1922) in his report on the phytoplankton collected 

 by the ' Scotia', and by Hardy (1935) and myself (1934) from material collected in the 

 course of the Discovery Committee's investigations. 



Zacharias 's first record (1905, p. 120), from 300 sea miles south-west of Cape Horn, is 

 from the typical habitat of the species, which, as I hope to show in this paper, is the more 

 southerly portion of the sub-Antarctic surface water. In 1934 (p. 161) I expressed sur- 

 prise at his finding it in such numbers so far south, but more recent work has shown that 

 the Antarctic convergence usually lies farther south in that longitude than the data then 

 available indicated. Zacharias 's second record (1906, p. 556), however, is most extra- 

 ordinary. It is from 12 N, 28 W, in the tropics between St Paul's Rocks and the Cape 

 Verde Islands. In all the 160 records of the species I have been able to get together, 

 there is only one other doubtful record of its appearance north of the sub-Antarctic 

 Zone, and I have been able to work through a considerable amount of tropical and sub- 

 tropical material obtained during our voyages to and from the Antarctic, in addition 



