354 DISCOVERY REPORTS 



that under the former low concentration, due to their not migrating so deeply, then the 

 zones might not only close up but overlap to produce a further effect of uneven distri- 

 bution, as shown in Fig. 190. 



Most likely, when formed, belts of plankton would before long be broken up, so as to 

 lose their form of parallel lines and appear as irregular patches. If parallel belts of 

 plankton retained their form for some time and the phytoplankton causing them died 

 down gradually over the area, then they would come to the surface in the form of belts 

 in the daytime, if the phytoplankton had become sufficiently reduced. Such a hypo- 

 thesis might account for the parallel belts of Salps seen in the daytime in sub-tropical 

 waters. 



It is likely that in nature the conditions will not be so hard and fast as those outlined 

 for simplicity in the foregoing propositions ; but a variety of changing conditions will 

 merge into one another producing as a rule a less sharply defined effect. 



If the hypothesis that the range of vertical migration may be determined by the 

 phytoplankton production is correct, then the foregoing propositions would give an 

 adequate explanation of the horizontal distribution of the different animals in relation 

 to the phytoplankton such as we have found. The animals which have an extensive 

 migration would show a separation from the dense phytoplankton, e.g. the macro- 

 plankton and the copepoda Group A. Those that remain longer in the surface layer and 

 do not migrate so deeply would not be so separated, e.g. copepoda Group C, and might 

 flourish in the zones of denser phytoplankton. Those which showed little or no regular 

 vertical migration and only rarely come into the surface layers would not show any 

 distributional relationship with the phytoplankton, e.g. the copepoda Group B. If a 

 separation and aggregation of animals is brought about by differences in vertical 

 migration, as we have just described, but differences due to some other factor than the 

 phytoplankton, then the phytoplankton production might be modified accordingly by 

 grazing (cf. Harvey, 1934). 



Propositions concerning the acceleration or retardation of water layers 



We will now consider the movements of an animal undergoing a more or less uniform 

 vertical migration between water layers, one of which may be accelerated or retarded in 

 relation to the other. 



In Fig. 191 let a body of water be moving from left to right, and at the point A let the 

 upper layer X increase its speed, perhaps under the influence of wind, but the layer Y 



A B 



Fig. 191. 



