28 DISCOVERY REPORTS 



more than 2 mm. in length were picked out at this stage; but if, as sometimes happened, 

 they occurred in sufficiently large numbers to be fairly sampled by the pipette, i.e. over 

 200, they might be left in. 



Next the sample was examined for smaller but rarer organisms, which on account of 

 their small number would not be fairly sampled by the pipette. These, which may 

 include Radiolaria, Foraminifera, small Chaetognatha, Polychaeta, rare Copepoda, 

 Ostracoda, Euphausiid larvae, Pteropoda and Appendicularia, are also picked out and 

 recorded. All the organisms picked out were separated into genera and preserved in 

 separate tubes, so that they were readily accessible for reference by ourselves and for 

 subsequent examination by specialists in different groups. 



The volume of the residue, after the larger organisms had been picked out, was again 

 measured after allowing the sample to settle in a nieasuring jar for at least 24 hours. 



In some instances this residue was small enough to be analysed as a whole, but usually 

 a fraction only could be taken. The fraction for analysis was taken by means of the 

 stempel pipette and flask as in the N 50 V samples, but the fractions taken were larger. 

 This stempel method and apparatus is described by Hensen (1887, 1895). The most 

 convenient fractions for examination were 1/20, 1/40 and 1/60, and were taken by using 

 a 2-5 cc. pipette after diluting the sample to 50, 100 and 150 cc. respectively in stempel 

 flasks of corresponding sizes. The amount of dilution employed depended upon the 

 nature of the sample. Other fractions from 1/20 to 1/150 were obtained by using 5 cc. or 

 1 cc. pipettes, and smaller fractions still by halving the sample before diluting. In deter- 

 mining the size of the fraction to be examined it was decided that there must be con- 

 tained in the fraction at least 100 of the commonest animal species present. Actually in 

 practice the number of the commonest animal in the fraction examined was often more 

 than double this amount. Choice of a suitable fraction was further influenced by the 

 relative abundance of larger and smaller organisms in the sample. For example, in 

 exceptionally bulky samples where large Calanoid Copepoda and small Oithonid 

 Copepoda might occur in numbers of over 500 and 5000 respectively, the residue was 

 more accurately and speedily analysed by taking two fractions: one with a 5 cc. or 

 2-5 cc. pipette for the larger organisms, and one with a smaller pipette or greater dilu- 

 tion for the smaller organisms. When shaking up the sample immediately before sampling 

 with the pipette, particular care must be taken to get an even mixture and to counteract 

 the effects of gravity and centrifugal force. This applies to the N 50 V samples as well. 



The quantity of liquid taken up by the pipette in separating these fractions from the 

 sample is often too great for convenient handling on the special microscope slide, and 

 the excess is preferably drained away before the pipette is discharged. This is easily 

 done by allowing the pipette to stand inverted, until all the organisms have settled on 

 the roof of the plunger. The pipette is now opened slightly so that the superfluous liquid, 

 which is held back by surface tension, may be drawn away slowly, either by a light 

 pipette or blotting paper, without disturbing the settled plankton. The pipette is now 

 closed and the contents are shaken round and discharged and spread out on the slide. 



The slide is similar to that shown in Fig. 15, only larger and with every fifth line 



