530 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 



osmium vapour for a short time before being placed on 

 the drop of water and sediment, which is sufficient to 

 cause narcotisation. When high powers are used, it is 

 impossible to count the organisms in the complete area 

 of the drop. In this case only a fraction is counted, in 

 the following way. The glass counting plate on which 

 the drop rests is crossed by a series of parallel lines 

 running in one direction, from the observer. If the 

 number of spaces between the lines, which are covered 

 by the whole deposit when the cover glass has been 

 applied, is divided by five, that gives the spaces in which 

 the organisms should be counted in order to arrive at 

 one-fifth of the total in the catch. These spaces counted 

 should be equally distributed over the whole area, so that 

 an average can be obtained. It is best first to count the 

 organisms in one-fifth of the mass with a high power for 

 the smallest and most frequent forms, and then, under 

 lower magnification, the whole mass for the larger and 

 less frequent. 



It is well also to take another quarter of a litre of 

 the same sample from the Kriimmel bottle (fig. 4) and 

 add formaline to make a 1 to 2 per cent, solution. After 

 this liquid is partly removed from the catch by filtration 

 through very fine filter paper, the residue can be 

 centrifuged and compared with the centrifuged samples 

 of living forms. The amount of water taken for the 

 centrifugation of the latter must depend on the number 

 of organisms present. If 15 c.cm. of water are first 

 taken and centrifuged, and too many organisms are found 

 for an easy count, then it should be discarded and a 

 smaller quantity taken. 



The extremely small quantity of water taken for these 

 samples is astounding and might be considered in- 

 sufficient for two reasons : — First, that not enough animal 



