146 



so frequently selected that no mistake could be made in connectina- 

 them by means of a smooth curve. A table was' then constructed 

 for each division of the siide-wire scale based on the data of the 

 curve. 



The work of testing 60 to SO samples and locating a similar number 

 of points along the curve of salinities vei*sus slide-wire readings 

 brought out some interesting facts in connection with the operation 

 of the apparatus. 



It was soon observed by readings of the apparatus that the tem- 

 perature factor played an important part in the test of a sample and 

 it was necessary tliat each sample be given plenty of time to permit 

 its temperature to come to the same as that of the bath. In order 

 to determine how long it took for a test sample to assume the bath 

 temperature, 25° C. the following slide-wire readings were taken 

 of a sample Avhich was '22° C, when placed in the X cell. 



It took approximately 12 minutes after placing the sample in the 

 test call to obtain an accurate scale reading. 



A sample was then heated to approximately 25° C, tliat of the bath, 

 and then placed in cell A' of the apparatus. 



It can be seen from the foregoing that in order to save timt- 

 it is well to bring the samples of sea water desii-ed tested near 

 the temperature of the bath, and even then a sample which dif- 

 fers by 2° C. should be allowed at least 12 minutes to come to rest 

 befoi-e testing. A sample near the bath temperature of 25° C. may 

 be accurately tested from 5 to 8 minutes after placing it in X cell. 

 In one instance where the sainide placed in A' cell was 3° colder 

 than the bath temperature, the )ninimum point as heard by the 

 phones gave slide-wire reading of 260. But this point crept con- 

 tinuously along tlie scale to 280 and 290 in less than 1 minute in- 

 terval and finally came to rest at 380. 70 points farther along on the 



