METHODS USED IN THE "WILLIAM SCORESBY' 137 



recent refinements of the pigment extraction method successively developed by Riley (1938), Krey 

 (1939) and Richards and Thompson (1952), should certainly prove a most valuable line of investiga- 

 tion in the Benguela current area. 



Estimation of salinity and phosphorus 

 The confined laboratory accommodation on board the 'William Scoresby' made it necessary to 

 restrict as much as possible the amount of analysis carried out in the ship. Estimations of at least 

 salinity, phosphate and oxygen were desirable for the work. Of these properties the salinity is by far 

 the most stable, and unaffected by normal storage, and it was decided to keep the salinity samples for 

 later analysis on shore. This has been the practice in the past on board the 'William Scoresby'. The 

 estimations were eventually made by the Government Laboratory, London, to whom we owe our 



thanks. 



The Atkins-Deniges colorimetric method was used for the determination of the dissolved inorganic 

 phosphate-phosphorus. Recently, many new photometric systems have been described, which 

 facilitate the colour comparison and attain a high standard of accuracy, but after experimentation both 

 before the cruise and later on board the ship, it was decided that in the conditions in which we worked 

 a visual comparison would yield the most satisfactory results of the required degree of accuracy. 

 This method was adopted and we are greatly indebted to Dr J. H. Oliver for perfecting a suitable 

 instrument for the purpose. 



A Lovibond type of colorimeter was used. This had a series of thirty-three specially made glass 

 slides, with colours flashed on to them representing a range in phosphate concentration from 0-03 to 

 3-00 mg. atoms P/m. 3 The slides were calibrated by the same operator in a shore laboratory, against 

 standard phosphate solutions (KH 2 P0 4 ) made up in phosphate-free sea-water. (This latter, of salinity 

 35-oo%„, was collected for the purpose in the Gulf of Guinea.) This procedure obviated any salt-error 

 in the estimation. The sensitivity of the method varies slightly with the intensity of colour, and is at 

 its lowest, ±0-05 mg. atoms P/m. 3 , at concentrations of 3-00 mg. atoms P/m. 3 All the phosphate 

 samples were analysed within a maximum of 10 hr. after collection. 



Estimation of dissolved oxygen 

 Winkler's method was used for the determination of dissolved oxygen. The samples were stored under 

 a water-seal in the precipitated form (Mn(OH) 3 ) and all were analysed within 20 hr. of collection. 

 Standardization of the sodium thiosulphate was effected by titration with an approximately equal 

 strength (0-22 n) solution of potassium iodate. This standard, kept in a slightly alkaline solution, 

 proved very stable even in the most adverse climatic conditions. 



For normal sea-waters the Winkler method works very well, but in the presence of certain con- 

 taminants the analysis can lead to erroneous results. It was to be expected, therefore, that in the 

 vicinity of the South-west African coast, where organic production is very high, such contamination 

 might be encountered. The principal sources of interference were considered to fall into three groups: 



1. Large quantities of micro-organisms in the samples. 



2. Relatively high concentrations of nitrites. 



3. The presence of hydrogen sulphide. 



Organic matter can affect the Winkler estimation in two ways. First, direct oxidation of organic 

 matter is very rapid around pH 12-0, and this might occur during the initial alkaline stage of the 

 analysis. Secondly, the iodine when liberated is liable to be absorbed by the micro-organisms. As 

 it was not possible to analyse all the samples immediately after collection, and as they would have to 

 be stored in one or other of these two stages, it was necessary to decide which course would cause the 



