164 SIR W, A. HERDMAN: RESULTS OF CONTINUOUS 
and other cases of more variation, even in that same series, such. as:— 
May 25th— Ceratium furca, 6000, 2000, 8000, 1000. 
Are we entitled from this to conclude that the Peridinium is evenly 
distributed through the zone of water sampled and the Ceratium much less 
so? I doubt it. 
The Copepoda seem also to indicate in many cases a fairly even 
distribution. Sometimes they occur only in units, and yet each haul of the 
series shows a few :— 
April 3rd—Oithona similis, 8, 4, 3, 3, 5, 11 ; 
April 13th—Temora longicornis, 10, 5, 10, 10, 10 ; 
April 13th— Oithona similis, 20, 20, 20, 20, 20. 
Other cases, again, seem to indieate considerable variation in adjacent 
hauls. Which of these contradictory impressions received from an inspection 
of the results of the hauls is true to nature? If the Oithonas on April 13th 
had been very irregularly scattered through the water, is it likely that we 
could catch exactly 20 in each of five successive hauls? On the other hand, 
if they are evenly distributed, how can we account for one haul (April 6th) 
catching 40 and the next 140, or for the series on May 25th— 20, 80, 460, 
290, in the four successive hauls ? 
Some of the other common organisms of the plankton outside the above 
main groups also give conflicting evidence. The pelagie arrow-worm, Sagitta 
bipunctata, 1s present in nearly every haul in numbers varying from one to 
twenty-seven, but in some series one or two individuals are present in every 
haul, while in another series the successive hauls varied from one to eleven. 
The impression one receives from an inspection of the lists and numbers as 
_ they stand is that if on each occasion one haul only in place of four or six had 
been taken, and one had used the results of that haul to estimate the abundance 
of any one organism or group of organisms in that sea-area, one might have 
arrived at conclusions about 50 per cent. wrong in either direction. 
Is such a result of any real value as a basis for calculations as to the 
population of the sea? And is it possible that sueh numerical variations are 
compatible with the hypothesis of an even distribution of the plankton 
throughout a sea-area of constant character? The answer to such questions 
depends to some extent upon the possible range of error under the conditions 
of the experiment, and upon the possibility of allowing for that experimental 
error, and of reducing it by more refined methods of collecting and 
estimating. I feel confident that the possibility of error in the collecting 
was reduced to a minimum. There is also the possibility of error in the 
microscopic examination and estimation of the contents of the catch. This 
can only apply in the case of the more minute organisms, present in great 
abundance, such as the diatoms, which have to be estimated from counted 
samples. In the case of Copepoda and Sagitta and other larger organisms 
this source of possible error is excluded, as these are picked out from the 
