42 University of California Publications in Zoology [Vol.15 



measured the specific gravity with a hydrometer of total immereion 

 (corresponding in accuracy with the pj'cnometer) while the other 

 measured the chlorine b.y Mohr's titrimetric method. From an exam- 

 ination of the results Nansen (1906, p. 11) concluded that, while the 

 chlorine content of the North Atlantic ea.st of Greenland was but 

 slightly less than normal, that from the Barents Sea exceeded the 

 normal by an amount which, allowing for the errors in measurement, 

 makes the specific gravity computed from it 0.000025 too high. 



Our own observations bear out. in a general way. this concli^sion 

 of Nansen. A series of samples taken from San Diego Bay as well as 

 the ocean were carefully titrated and also tested for specific gravity 

 with a pyenometer. A comparison of the results (see p. 33) shows 

 that, while the average difference in the values obtained by the two 

 methods is only 0.000025. differences of as much as 0.00015 frequently 

 occur. Furthermore, repeated titrations of the "normal sea-water" 

 rarely gave results differing, in specific gravity, from the true value 

 by more than 0.00004. thus indicating that the larger differences may 

 have been partly due to a variation in the proportion of the salts in 

 the water of this region. However, many more similar observations 

 must be made before such variation can be proven. 



Nevertheless it is certain that, while the relation between chlorine 

 content and salinity may be so nearly the same in many regions that 

 the difference can not be detected by the usual methods, there are 

 regions where the difference is very apparent. Therefore, until the 

 magnitude of this difference is accurately known for a given region, it 

 would undoubtedly be wiser to determine both the chlorine content and 

 specific gravity directly than to depend iipon a reduction formula 

 which may not apply with sufficient accuracy. If this were done and 

 the observations repeated whenever the values obtained by both 

 methods differed in specific gravity by more than 0.00003. one could 

 readily decide whether or not the difference was due to an accidental 

 error in measurement or to a variation in the composition of the sea- 

 water. Such a comparative series would certainly be of increasing 

 value as the work progressed. 



Since it has not been feasible for the Seripps Institution to make 

 such an extensive series of comparisons, the direct method of measur- 

 ing specific gravity was finally adopted, because : 



1. In plankton investigations a knowledge of variations in salinity 

 is of fundamental importance, and it is more advisable to measure it 

 bv the most direct method available than to measure the chlorine con- 



