70 DR W. S. BRUCE, MR A. KING, AND MR D. W. WILTON ON THE 



Hence the density of the sample at t' is 



(W + w) 



(186-8437 + 0-00567t' + 0'00887R) 



Thus if this particular hydrometer (No. 25) floats at division 31*8 of the stem with 



an added load of 5'80 grams, and the temperature of the water be 21 *2° C, the density 



at 21°*2 C. is 



186-2944 + 5-80 



186-8427 + 0-00567 x 21-2 + 0-00887 x 31-8 



= 1-02589. 



Hydrometer 14. — This hydrometer was used in the testing of samples 200 to 578. 

 Its weight in vacuo is 181*0189 grams. The volume at 0° C. of the body of the 

 hydrometer up to the lowest division of the stem is 181*5471 c.c. The rate of 

 expansion per degree C. of the body is 0*00624 c.c. The volume per mm. of the stem 

 is 0*009046 c.c. 



As in the case of hydrometer 25, these constants were determined in Edinburgh 

 before the start of the Expedition. 



Thus in determinations of density made with this instrument, the density of- the 



sample is found by the formula 



W + w 



(181-5471 + 0-00624t' + 0-009046R) ' 



W, w, R, and t' having the meanings assigned above. 



Mr Wilton never took less than four separate readings of the stem by successive 

 addition of weights so as to increase the amount of immersion. In most cases, 

 however, he took considerably more, generally about eight. From each of these 

 readings an independent value for the density could be calculated ; but as this would 

 have involved much laborious calculation, he calculated the density from the mean 

 effective weight and the mean stem-reading. The temperature of the sample during 

 experiment (t') was read before immersing the hydrometer in the liquid and after 

 removing it, and the mean of these two readings was taken as the " temperature 

 during experiment." This was used in the calculation of the density by 'means of 

 the formula given above. The two readings of temperature never differed by more 

 than 1 °*0 C, but generally varied by very much less.* It was very exceptional to 

 have so large a difference as 1°*0 C, and in those cases the accuracy of the density 

 determination may be open to doubt. In the tables the column T' gives the 

 temperature of the air at the time when the determination of density was being 

 made, and the column t' gives the temperature of the sample of sea-water whose 

 density is being determined. The nearer these temperatures are to each other, the 

 less alteration in the temperature of the sample during the experiment is to be 

 expected. 



* Mr Buchanan would not be satisfied in his chemical work with an experiment in which the temperature of the 

 solution during experiment varied by more than - 01° C, but under the circumstances and conditions such a degree 

 of accuracy as this could not be expected. 



