VEGA-EXPEDITIONEKS VETENSKAPLIGA ARBETEN. 



333 



'^'^/O.ooos to the obs. numbers of sp. weight, which are higher than I.0235 



b o ^ IO.OOO.5 » » > » » « >■ 



^^ O \0.0004 » » ^ » „ >, » 



° ^ iO.0003 » >. » » »• « » 



^ .^ "0.0003 >. » >■ » > » " 

 b o !^ I 



o - S< 0.0002 * - » ' » » , 



The f^reatest discrepancy between the corrections (=0.00012) 

 is at 1.0240 and I think the computed spec. weights can be 

 considered to be true within O.0002. This ambiguity causes 

 an uncertainty of O.02 p. C. in the calculated amount of salt. 

 These corrections, however, are strictly applicable only in 

 case the original observation was made within those limits of 

 temperature, where the correction from the temperature of the 



water-sample in sltii to -j-^o C is less than O.oooi and conse- 

 quently can be omitted. The observations taken on board 

 the Vega genérally fiilfiU this condition, with the exception of 

 the determinations in the winter months, which were executed 

 in the cabin at a temperature of about + 10° or + 11° C. I 

 have recorded these observations without any corrections at 

 all. Happily the titrations of chlorine executed on board the 

 Vega are sufficient to give one an idea of the saltness of these 

 water-samples. 



From the reduced spec. weights at 7— p C the percentage 

 of salt in the water-samples is computed by means of the 

 coefficient 123.8. I deduced this number from the analyses of 

 the following 4 water-samples. 



