METHODS USED IN THE 'WILLIAM SCORESBY' 139 



reduction in the oxygen content of the second two samples was solely due to reduction by hydrogen 

 sulphide, the result would indicate that a concentration of about o-8 cc. H 2 S/1. had been oxidized in 

 the 4 hr. which elapsed. Unfortunately we had no opportunity of making more experiments of this 

 nature, and it is not possible to say precisely how efficiently the Rideal-Stewart method accomplished 

 the task which it was set. 



TREATMENT OF THE PLANKTON SAMPLES 



The microplankton catches (N 50 V samples) were worked up by a counting technique essentially 

 similar to Hensen's method, as described by Steuer (191 1). From the whole sample, thoroughly 

 mixed by agitation in a spherical Stempel flask, subsamples were drawn off with an 0-5 ml. Stempel 

 pipette and placed in a cell on a large slide, ruled with squares of approximately i-8 mm. These can 

 almost be contained within one field of the low-power objective (-§ in.) of an ordinary compound 

 microscope. The cell was originally intended for counting centrifuged samples at sea, where a cover- 

 glass must be used. It measured approximately 10 x 24 of the squares and 0-4 mm. in depth so that 

 almost exactly 0-3 ml. of fluid could be contained in it, trapped under a large rectangular cover-glass. 

 For this work cover-glasses were not used during the early stages of the counts, but the cell was found 

 very convenient to contain the larger volume of fluid while spreading it over the face of the slide with 

 a mounted needle. When difficult rich subsamples prolonged the counting, or higher powered 

 objectives than the £ in. were needed for identification, it was found that evaporation at room tempera- 

 ture allowed one to apply a cover-glass after some 45 min. Most of the counts took more than 2 hr. 

 and continuous observation for a longer period increased personal error due to eye-strain, though 

 a brief pause sufficed to counteract this. Thus the use of a cover-slip during the later stages, to prevent 

 the preparation drying up, became an essential part of the technique. 



The rulings on the face of the slide enabled one to resort to higher magnifications, or to recapitulate 

 when necessary, without ' getting lost ', and it was thus quite practicable to work through the whole 

 subsample, using a large mechanical stage. A very abundant inshore pennate diatom, Fragilaria 

 karsteni, has the habit of growing in very long ribbon-shaped colonies, and it was found that the 

 numbers of frustules could be estimated rapidly, and with a high degree of accuracy, by measuring the 

 ribbons under low power with a micrometer eyepiece, then measuring the width of a few individual 

 frustules under high power and dividing appropriately. 



For the method as a whole to be successful the subsample must not be so thick that individual cells 

 or colonies obscure each other unduly. With large samples this was avoided by preliminary fractioniza- 

 tion and adjustment of volume, using a large (5-ml.) Stempel pipette and measuring cylinders for 

 added water. Errors were obviously increased by such a crude method, but could be reduced by 

 making separate counts at different dilutions for the more difficult stations, and pooling the results ; 

 and by such expedients as counting large or conspicuous species at normal dilutions where the density 

 of smaller dominant species necessitated further subsampling before they could be counted. When 

 the general nature of the flora had become familiar it was nearly always found possible to judge the 

 extent of fractionization and dilution needed from the settlement volume and macroscopic appearance 

 of the sample. At worst two trials enabled the necessary adjustment to be made. 



The main object of this line of work was to study the spatial distribution of the larger phytoplankton 

 organisms in relation to the hydrological features of the area. Obviously such estimations can bear 

 little direct relation to the total amount of phytoplankton present — it has long been known that there 

 are many autotrophic organisms among the minute forms that escape the finest nets — but it is claimed 

 that, so far as the diatoms and larger dinoflagellates are concerned, they show up the grosser quanti- 

 tative differences quite fairly in areas such as this, where the gradient of population density is very 



