216 DISCOVERY REPORTS 



Chaetoceros atlanticum var. (or phase) neapolitana (Schroder) Hustedt, previously often re- 

 garded as a distinct species, C. neapolitana Schroder. Here lumped with the type. 



Gran and Yendo (191 4) noted that this form, which they regarded as a distinct species, was most 

 often to be found in the warmer waters around Japan, while the highly variable and cosmopolitan 

 C. atlanticum Cleve was mainly confined to the colder water masses in that region. Earlier work round 

 Japan suggests that the neapolitana form might be found in the colder waters also. Though Okamura 

 (1907) identified his specimens as C. atlanticum, Gran and Yendo seem satisfied that they were the 

 neapolitana form, and doubted the localities whence they were recorded in consequence ! 



Hustedt (1927-37), Hendey (1937) and Boden (1950), the latter working on South African material, 

 have regarded neapolitana as a variety of C. atlanticum. It was Hustedt who regularized this change, 

 and both he and Hendey also dealt with the very confused synonymy of the type. 



Our material seems clearly to support the view that neapolitana can hardly be a separate species. 

 Both the type (C atlanticum) and the variety occurred together at some stations near the shelf-edge, 

 with seemingly intermediate forms, and neapolitana more frequently than type or intermediates in the 

 warmer oceanic water still farther offshore. 



The regular banded appearance of the setae, due to chloroplasts penetrating into them and disposed 

 at regular intervals, was very characteristic of the neapolitana phase, and is shown in figures of material 

 from other localities by Schroder himself (1900) and Cupp (1943) among others. However, our 

 material also showed individual colonies of both type and intermediates with a similar disposition of 

 chloroplasts. We have to remember that the movements of chloroplasts within the shell are probably 

 governed by light intensity. Hence this appearance in preserved samples may depend more upon the 

 depth at which they were captured and time interval prior to fixation than upon any greater tendency 

 towards this pattern in the neapolitanum phase, rather than in type or ' intermediate ' strains. 



Certainly we would agree with Gran and Yendo that neapolitana is a warm-water phase of the 

 species, and this is also the view of Cupp (1943) and Takano (1954) who gives excellent figures of 

 what he terms var. neapolitana and var. skeleton. 



It seems significant to me that when material from the north-eastern Atlantic only has been studied, 

 there are many examples of diatoms that show a polymorphism, seemingly so definitely related to 

 temperatures (and therefore having a distinctive space/time distribution) that systematists have 

 failed to agree upon the specific or subspecific status of the different forms. Yet, when material from 

 regions with steeper temperature gradients has been examined, these same cosmopolitan diatoms have 

 seemed to fall much more convincingly into categories of less than specific rank (often obviously linked 

 by intermediate forms) of single types. Thus these chaetocerids, and the solenoids mentioned below, 

 seem justifiably 'lumped' on the basis of our Benguela current material, derived from a typical 

 upwelling region where the temperature gradients are frequently very steep. Such abrupt variations 

 in temperature are also to be found near the northern limits of the Kuroshiwo current, whence some 

 of Gran and Yendo's material and most of Takano's was derived. 



Chaetoceros costatum Pavillard, previously widely referred to as C. adhaerens Mangin. 



This change was advocated by Hustedt (1927-37, pp. 699-700). Its propriety is not very obvious 

 from the figures he reproduced from Pavillard and from Mangin. Cupp (1943, p. 127) follows 

 Hustedt and gives better figures, that agree well with the varied South-west African samples we 

 studied. 



In our material the species was very restricted in distribution, but abundant at four autumn stations 

 inshore. Here the range of variation in dimensions coupled with the presence of mucous (or ' pecti- 

 naceous ') cushions between corners of adjoining cells, a character of Mangin's definition, in chains 

 that also displayed the intercalary bands shown by Pavillard's figure, led me to conclude that Hustedt 



