31 There is, however, another factor which appears to be of considerable importance. This ts the effect of distribution of activity in the water with depth on the activity in the plankton sample. The plankton sample is taken from 4 water layer at a depth of 200 meters to the surface. The water sam- ples are taken at discrete depths down to 150 meters. The fol- lowing uncertainties sre therefore inherent in the sampling method. First, there is not a uniform decrease of activity with depth (Figs. 8 and 9). There are, indeed, changes with depth which range from a sharp decrease from the surface to 25 meters to the other extreme where there are sharp increases at a derth of 100 metars. are taken at discrete Secondly, since the water samples intervals of 25 to 50 meters, the pre- sent data do not permit ruling out the possibility that extremes of activity in narrow regtons distort the total picture at a given station elther by being missed completely or being up- li- overemphasized. And thirdly, there {8 no guarantee that the plankton is uniformly distributed throughout the depth which is sampled by the plankton net. There is, for example, some evidence that a change in the vertical distribution of the plankton population occurs from hours of daylight to hours of darkness. The diurnal-nocturnal difference in ratio of activ- ity in surface plankton (material on filter papers) to that in the surface water is significant at stations 1-30, which for the most part show a sharp decrease in activity in the water from the surface to 25 meters, and no subsequent increase. No such significant difference was found at the remaining stations,