144 THE DISTRIBUTION OF THE PLANKTON [PART II 



the same construction were coupled together, lowered to the same 

 depth and drawn up again, and seven of such double catches were 

 made in the open sea in the North Atlantic. The catches were 

 compared, but it was seen that no two of them were exactly alike, 

 for when the mean of the two catches was taken as 100 it was 

 found that the average divergence of the individual catches was 

 6'8 7o- ^^^^ previously to these experiments Hensen^ had made a 

 number of experimental hauls in the Baltic by the method of 

 making a vertical haul, emptying the net and then immediately 

 lowering it again in the same spot to exactly the same depth. 

 But here again the volume of the plankton taken in each of the 

 pairs of hauls was never exactly the same, but the average diver- 

 gence of each haul from the mean of the two was 15*88 7o ''^^^ ^^^ 

 probable divergence 7*4 7o- Now while the result of these trials 

 was no doubt disappointing it was perhaps as we ought to 

 have expected. It is quite impossible to construct and use two 

 scientific instruments so that they will give exactly the same 

 results : even with the exact and highly developed apparatus and 

 methods of the chemists and physicists one never finds that the 

 results of duplicate analyses or determinations are exactly alike. 

 Experimental errors must always be encountered and one has to 

 accept these and be as certain as possible that their magnitude is 

 less than that of the variations which he is investigating. The 

 vertical net is a very complex instrument and it is not possible to 

 make two so exactly alike, or to use the same net twice in suc- 

 cession, so that the same result will be obtained in each case, even 

 when we should expect that the conditions of trial were such as to 

 lead us to expect identical results. Further the final result of the 

 fishing operation depends on the exact estimation of the amount 

 of water passing through the meshes of the net ; and this 

 estimation is based on the determination of the constants of the 

 nets, and here again errors may creep into the methods. 



But even when we allow for these experimental errors it is 

 still the case that the plankton is not distributed in an uniform 

 manner throughout the sea at any one time. We have first of all 

 divergences from an average composition in a horizontal direction, 

 that is the plankton is not usually the same at two parts of a sea 



^ Ueber d. Bestimmung d. Planktons, loc cit. See also Appendix. 



