CHEMICAL RESULTS OF LAST CRUISE OF CARNEGIE 



Piacific, and of the oxygen content of the surface and 

 the subsurface water of a large part of the open Pa- 

 cific. 



In the Atlantic the Carnegie occupied only thirty- 

 four stations. Hydrogen ions and phosphates were de- 

 termined at various levels at all stations and dissolved 

 oxygen at two stations. Although the data obtained are 

 not extensive when compared with those obtained in the 

 Pacific, they constitute an important contribution to our 

 knowledge of the chemical conditions in regions not pre- 

 viously investigated, namely, the central North Atlantic 

 and the Caribbean Sea. These data became particularly 



interesting when compared with those obtained by others 

 in both the North and South Atlantic. 



On board the Carnegie . H. R. Seiwell had charge of 

 the chemical investigations in the Atlantic and from 

 Panama to Samoa in the Pacific (stations 1 to 93); J. H. 

 Paul from this point to Yokohama and thence to San 

 Francisco (stations 94 to 129); and H. W. Graham from 

 San Francisco to the termination of the cruise at Samoa 

 (stations 130 to 162). Between San Francisco and Hono- 

 lulu, E. G. Moberg of the Scripps Institution of Oceanog- 

 raphy, who was a guest investigator on board, made 

 some of the analyses, especially of oxygen. 



METHODS 



Phosphate 



For the determination of phosphate, use was made 

 of the colorlmetric method of Denlges (1920, 1921) as 

 adaptedforseawater by Atkins (1923a, 1925). This 

 method depends on the formation of an intense blue, on 

 the addition of certain reagents to a solution containing 

 phosphate, the intensity of the color being in direct re- 

 lation to the quantity of dissolved phosphate. To 100 ml 

 of the sample to be analyzed, were added 1 ml of a so- 

 lution consisting of one part of 10 per cent ammonium 

 molybdate and three parts of 50 per cent (by volume) 

 sulphuric acid, and one or two drops of stannous chlo- 

 ride solution freshly prepared from 0.1 gm of tin dis- 

 solved in 2 ml of hydrochloric acid, in the presence of a 

 small quantity of copper sulphate, and made up to 10 ml 

 with distilled water. The blue color appearing in the 

 sample was matched against that of a standard phos- 

 phate solution similarly treated. For this comparison 

 the sample and standard were placed in glas? tubes 

 about 300 mm long and about 25 mm in diameter and 

 graduated in 125 units. These tubes were viewed in a 

 colorimeter constructed to throw diffused light through 

 the columns of liquid when standing in a vertical posi- 

 tion and the levels of the liquids in the tubes were ad- 

 justed until both showed the same intensity of the blue 

 color. By using appropriate standards, the difference 

 in the scale readings of the two tubes was kept to within 

 twenty-five scale divisions and usually to much less. 



The working standards used were made up from a 

 stock solution of potassium dUiydrogen phosphate pre- 

 served with toluene. A supply of this solution, which 

 was thought sufficient to last until the Carnegie's ar- 

 rival in a home port, was prepared before the start of 

 the cruise. This supply was almost exhausted on arriv- 

 al in Japan (station 112), however, where a temporary 

 standard was made up, calibrated with the old, and used 

 for stations 113 to 129. A fresh standard, prepared at 

 the Scripps Institution of Oceanography by Moberg, was 

 used from San Francisco until the end of the cruise 

 (stations 130 to 162). The old standards were checked 

 against this standard and were found to have suffered no 

 deterioration. 



Those who are familiar with this method will real- 

 ize that the accuracy of the results, as expressed in per 

 cent, decreases with decrease in concentration of phos- 

 phate in the samples analyzed and that values of 5mgP04 

 per cubic meter or less should be regarded as indicat- 

 ing merely the order of magnitude. For "salt error" 

 no correction was applied, and consequently our results 



are comparable with nearly all the results of phosphate 

 studies of sea water reported before 1930. 



The phosphate values are expressed as milligrams 

 of PO4 per cubic meter, because at the time the data 

 were computed and tabulated this was the most common- 

 ly used method. Many of the earlier data found in the 

 literature are given as P2O5, whereas many of those 

 published later are expressed as P or, more recently, 

 as milligram atoms per kilogram or cubic meter. 



Silicate 



Silicate was determined by the method of Dienert 

 and Wandenbulcke (1923) as modified by Atkins (1923a). 

 This is a colorlmetric method, similar to that used for 

 phosphate. On the addition of certain reagents to a sil- 

 icate solution, a yellow color like that of a picric acid 

 solution is developed. In making the determination, 2 

 ml of a 10 per cent solution of ammonium molybdate 

 and 6 drops of a 50 per cent (by volume) solution of 

 sulphuric acid were added to a 100 ml sample of sea 

 water. In the same colorimeter used for the phosphate 

 determination, the sample thus treated was compared 

 with a picric acid standard of a concentration such that 

 the scale readings of the two tubes differed by less than 

 25 units. The standards were prepared as directed by 

 King and Lucas (1928), who found that a solution of 25.6 

 mg of dry picric acid per liter has a color correspond- 

 ing to that of a silicate solution equivalent to 50 mg 

 Si02 per liter. 



Samples for the determination of silicate were 

 drawn from the Nansen bottles into "citrate of magne- 

 sia" bottles and analyzed the following day. Experi- 

 ments carried out at the Scripps Institution of Oceanog- 

 raphy have shown that sea water may be stored for at 

 least this long in bottles of this type, If well seasoned, 

 without undergoing any measurable increase in the sili- 

 cate content. In order to insure that the bottles used on 

 the Carnegie would not give off silicate to the water, 

 they were filled with sea water and left standing for 

 several weeks. Nevertheless, the water stored in a few 

 of these bottles consistently gave results that obviously 

 were too high. These bottles were finally discarded and 

 the results obtained from water stored in them have 

 been marked "rejected" in the tables and have not been 

 included in the graphs. 



The results of silicate determinations are expressed 

 as mg Si02 per cubic meter and, as in the case of phos- 

 phate, correction for salt error has not been made. 



