VERTICAL MIGRATION RECONSIDERED 335 



independent of the larger patchy distribution also recorded. Atkins (1922) writes con- 

 cerning the herring and />H values that : 



... It must not, however, be left out of consideration that large numbers of fish themselves modify 

 the hydrogen-ion concentration by their respiration, and the water in a shoal must be less alkaline 

 than the sea water in general. These shoal fish are presumably accustomed to this somewhat stuffy 

 sea-atmosphere, and so it appears probable that if separated from the shoal a herring by making for 

 a region of lower pH would be led back to the shoal.. . . 



It seems possible that the laying down of skeletal structures of silicate or carbonate 

 would also be a counteraction to the increased alkalinity set up by photosynthesis in the 

 upper layers. 



Unfortunately in our present survey pH observations were taken at only a limited 

 number of stations. They were taken only by the 'Discovery', and since she did not 

 work the stations to the west and south of the island we have no pH observations in 

 these regions where the phytoplankton production was greatest. The differences in />H 

 values in this survey were not large ; they are tabulated in TableLXXIII arranged in order 

 of stations 1 of ascending phytoplankton values. The higher phytoplankton values and the 

 />H values of 8-34 2 and over are shown in heavy type. Whilst the differences inpH values 

 are slight, we see that the higher pH values are in general correlated with the higher 

 phytoplankton values. It must be remembered that measurements of phytoplankton 

 values by numbers of cells (and also by volume) have been shown to give only a rough 

 estimation (p. 68). It should be noted too that the stations are taken at different times 

 of the day and under different weather conditions, so that the rate of photosynthesis may 

 vary considerably at different stations. The states of weather are tabulated in Beaufort 

 notations. The high/>H values, 8-33, at the two stations of lowest phytoplankton values 

 may in part be due to the fact that the stations were worked on sunny days (shown as 

 b.c. and b. in the table) when the photosynthetic activity would be higher than on 

 cloudy days. Dividing the series of pH values into four groups of four stations each we 

 see that the average />H values for the groups are 8-29, 8-32, 8-31 and 8-34. Now very 

 much higher phytoplankton catches were taken at stations at which no pH determina- 

 tions were made. It seems reasonable to suppose that there would be higher pH values 

 on the average at these stations than at those given in the table. We have already seen 

 in the later surveys that the pH values in general had a correlation with the reduced 

 phosphate values and so with phytoplankton production. But in these surveys the 

 ranges in pH values are not great ; they are tabulated on p. 336. 



We see that not only are the ranges met with small, but that if increased alkalinity is 

 indeed an effective cause in modifying vertical migration then it would not be a particular 

 degree of alkalinity that brought it about, but a certain increase relative to the lower 



1 St. 139, where no phytoplankton observations were taken, is omitted. 



2 Unless otherwise stated the />H values quoted in this section represent the average />H values for the 

 top 50 m. of water based on samples taken at o, 10, 20, 30, 40 and 50 m. It should also be noted that the 

 pH values used in this report differ from those given in the published Station Lists in that they have been 

 corrected for temperature, salt error and the effect of pressure. I am much indebted to Mr G. E. R. Deacon 

 for kindly estimating the corrections for me. 



43-2 



