INTRODUCTION ^47 



evidence of a secondary autumnal increase in March 1930, with Rhizosolenia spp. dominant (Hart, 



'^OurTatlr wo/k from 1933 onwards was based on results obtained by Harvey's (1934) method. 

 I have not previously dealt with our results from the sub-Antarctic Zone, because they were relatively 

 few and widely scattered, not lending themselves to the methods of presentation adopted in dealing 

 with the Antarctic material, which was the main object of study (Hart, 1942). Although most of these 

 observations are from oceanic waters, I think it profitable to discuss them briefly here since they 

 seem to me to give a good indication of the type of seasonal cycle (of standing crop) that may be 

 expected on the shelf. In November 1936 several observations were made on the shelf and in the 

 oceanic waters just beyond the shelf edge. These showed very little phytoplankton on the shelf but 



some rich hauls offshore. . ..^u^f^.oc 



The widely scattered observations available throughout the sub-Antarctic Zone, north of 54 b 

 are shown as a scatter plot in Fig. 7. No grouping of these observations by time intervals would 

 permit calculation of means that could be plotted to show the seasonal variation effectively^ Owing o 

 the'mall numbers of observations and their wide dispersion, the differences between the monthly 

 rneans for example, would not be statistically significant. Wide dispersion is a common source of 

 rifficuity in all 'samples' of quantitative estimations of phytoplankton, whatever method is adopted^ 

 I Teerns certam that it is due to the extremely patchy distribution of these orgamsms m nature, though 

 1 meThods have limitations that may exaggerate this feature. Plotting the logarithms of the individual 

 observations instead of the actual numbers of units per m.^ overcomes another great difficulty due to 

 t^^^^^ of finding a scale on which all the observations can be shown. The observations 

 Iwn we not all obtained in the same year and, as already stated, they were scattered throughout 

 Llblntarctic Zone, but apart from the logarithmic plotting they have not been otherwise mam- 

 pulated fnty way. Moreover, nearly all the species involved have a completely circumpolar distn- 

 Kiitinn as in the Antarctic Zone to the southward. 



tZ2Z shows clearly that the only months in which hauls exceeding looo umts per m.» occurred 

 weTe Nov:ltr Dec "bl and March' Further, .he proportion of very small hauls, be ow too unrts 

 were NovemDer ^ ^^^^^^^^ ^._^.|^_^ ^^ ^^^^ postulated for he 



rnhtroctreTa^r: IfThe Ant^arctie Zone (Hart, .,4. PP- 307-8). but centred slightly earher 

 " wLr 1: Ts Manation of the low values recorded over the shelf in Novenrber .,36? The 



'Tild not fake place un^ December in our southern locality. But cond.tions in the two areas are not 



"' I " n the Antar ti 'the time of the main increase in oceanic sub-Antarct,c waters ts remark- 

 s' ^(^hirrths' of the year, as compared 7;;-^-:^ rTm r^!::: 



L:^^SytSe:rr:eitres:^^^^^^^^^ 



