324 DISCOVERY REPORTS 



to twenty-four frustules where it is most abundant, as already described in the notes 

 on the species, and even the typically solitary Synedra pelagica may be seen in rafts 

 when it is dividing rapidly. When we remember that the oceanic Chaetocerids of Group 

 IV and the smaller neritic species of that genus are all colonial in habit, it is obvious 

 that the property is general for the majority of the important Antarctic species at the 

 time when conditions are at their optimum. 



The minority include the important solenoid diatom Corethron criophihm, of 

 Group II. This species also has the habit of forming chains, but under different en- 

 vironmental conditions which it is possible to study in some detail. ' In the mmute 

 type phase of the far south, and the hystrixjtype intermediates so characteristic of the 

 main increase in the Northern Region, this species is solitary. As already explained m 

 the notes on the species, the large strongly silicified spinose phase of the Northern 

 Region, and the more northerly special areas, tends to give way to a population of the 

 less strongly silicified, usually spineless, inerme phase, which forms very long chains, 

 during the post-maximal decrease and in autumn. This change was almost a total one 

 in the South Georgia area in January-February 1930. 



The change over to a largely spineless, thinner walled chain form of Corethron at a 

 definite season provides an opportunity for testing any correlation that may exist 

 between type of Corethron population and differences in the environmental conditions. 

 The fact that the change accompanies the rise to maximum temperature for the year 

 might lead one to conclude that temperature alone, or perhaps temperature with 

 seasonal rhythm inherent in the organisms themselves, is its primary cause. But the 

 change also coincides with maximum depletion of nutrient salts in the medium. 

 Ahhough the depletion of phosphate may be large, it is always present in considerable 

 quantity in the Antarctic zone, and there is little likelihood of its exercising more than 

 a secondary influence. On the other hand, the depletion of silica (directly involved with 

 cell wall thickness, one of the features of the change) may be relatively enormous 

 (Clowes, 1938, p. 112), and Pearsall (1932) has shown that concentrations lower than 

 500 mg. per m.^ may affect the development of certain fresh-water diatoms. We know 

 that diatom populations can flourish at lower concentrations in the sea, but it is strongly 

 suggestive that the fall to some 300 mg. or less quoted by Clowes occurs at the time at 

 which the maximum change in form of the Corethron population has been observed. 

 In fact, it would seem that temporary shortage of silica is most likely the main cause of 

 this change, as, no doubt, it is connected with the lessening quantity of the phyto- 

 plankton as a whole. This suggestion had already been made hypothetically (Hart, 

 1934, p. 185). No analyses for silica were available at that time, but some of Cooper's 

 (1933, p. 697) observations strongly favoured such a view. 



From 1933 onwards silicate analyses were adopted as part of our routine observations, 

 and there is much support for the above hypothesis on general grounds, as Clowes 

 (1938, pp. 111-14) has already shown. In an endeavour to make a more exact test of 

 the possible correlation between silica content and the proportion of the spineless chain 

 form in the Corethron population, I have attempted a statistical analysis of the obser- 



