216 DISCOVERY REPORTS 



racters are thus to some extent correlated with regional distribution. The spherical forms 

 are found off the coast of South Africa, also in the Mediterranean, and are considerably 

 greater than the tubular ones in diameter. The population of the North Atlantic com- 

 prises cells whose length is usually about twice the breadth, these forms making their 

 appearance in the English Channel. Moving southwards, a considerable lengthening 

 along the principal axis (girdle view) is noticed, and in latitude 44 52' S the population 

 consists entirely of forms whose length is ten to fourteen times as much as the breadth. 

 These forms are simple hyaline tubes, having thin walls and a circlet of bristles at each 

 end ; no spines appear on the bristles. Farther south a mixture of forms is encountered 

 whose length varies from two to ten times the breadth, and minute spines are sometimes 

 observed on the bristles. 



Still farther south, in latitude 57 36' S, forms of varying length are encountered 

 bearing an inner corona of shorter bristles furnished with claws, and these are inter- 

 mixed with previously described forms. 



The Corethron population presents forms varying greatly in length, which may or 

 may not be furnished with bristles, the bristles may or may not be furnished with spines, 

 and coronas of claws may or may not be present. C. criophilum, the type of the genus, 

 which is widespread in sub-Antarctic and Antarctic waters, was described by Castracane 

 in 1886 in the Reports of the Voyage of the ' Challenger'. In 1905 Karsten described two 

 forms from Antarctic waters, C. valdiviae and C. inerme, but Mangin (191 5) held that 

 owing to the presence of so many intermediate forms, it was impossible to distinguish 

 the one from the other. Hart (1934) supported this view. Hustedt (1930) is certain that 

 C. valdiviae of Karsten is synonymous with C. criophilum of Castracane, but prefers 

 Karsten 's specific name upon the grounds that his description of the organism was more 

 accurate. For my own part, I find there is no case whatever for refusing to admit 

 Castracane's species: the description and illustrations provided are unmistakable. 



From the almost unlimited supply of material at my disposal I have no hesitation in 

 saying that the species created by Karsten are merely phases of the one species C. 

 criophilum Castracane, and that the presence or absence of rhizosolenoid markings, 

 circlets of bristles, spines, or spines with claws are of no specific significance. 



Regarding C. hystrix Hensen, I find myself in complete agreement with Ostenfeld 

 (1902) and consider this species to be synonymous with C. criophilum Castracane. 

 C. hystrix has been held to be separate from C. criophilum on account of the absence of 

 the corona of short spines bearing claws and because the length of the frustule is seldom 

 more than three or four times the diameter of the valve. But I find that the relation be- 

 tween the diameter and the length of the frustule is of no specific importance, and that 

 the intermixture of so many forms of various lengths at many stations in the more 

 southerly waters made it impossible to make any such separation. 



Hustedt (1929) states that he has observed " klammerborsten " upon the valves of 

 specimens from the Sea of Japan, and I have observed them upon specimens from the 

 South Atlantic. The rhizosolenoid markings upon the connecting zone are to be ob- 

 served upon some specimens, particularly if the material is examined when mounted dry. 



