(37-2 



DR HUGH ROBERT MILL ON THE 



In order farther to test the action of the hydrometer by means of the double readings, 

 we may consider the amount of depression produced by the addition of the same weight 

 in different cases. The examples are taken from the first group of 760 determinations, 

 in order to take advantage of their slightly greater accuracy due to more favourable 

 conditions of observation. 



When weight No. 7 was added to the hydrometer it produced a mean depression of 

 3G"8 mm. of stem. It weighs 0'3312 grammes, and the difference of volume immersed 

 by it due to variations in temperature and salinity was shown by calculation and experi- 

 ment to be less than half a millimetre in water ranging from 1"020 to 1'026 in density 

 for 10° of temperature. This weight was employed in 305 cases with the results stated 

 in Table XV. 



Table XV. — Variation in Hydrometer Headings — I. 



Discrepancy in Density Determinations. 

 (Units of Fifth Decimal place.) 



Number of Times Weight 7 Produced Depressions of — 



35-5 



36-0 



36'5 



37-0 



37-5 



38-0 



38-5 



390 



^ 5, . 

 5 to 7, . 



Total, 



12 



2 



1 



41 



4 

 13 



57 



15 



84 

 3 



42 

 3 

 2 



11 

 3 



10 



2 



12 



44 



74 



102 



47 



14 



10 



2 



Weight No. 6, weighing 0"5184 grammes, and producing a mean depression of 57'7 

 millimetres, was employed 293 times with the results shown in Table XVI., arranged 

 according to the value of the density discrepancies in the double readings. 



Table XVI. — Variation in Hydrometer Readings — II. 



Discrepancy in Density Deter- 

 minations. 

 (Units of Fifth Decimal place.) 



Number of Times Weight 6 produced Depressions of — 



56-0 



56-5 



57-0 



57-5 



58-0 



58-5 



59-0 



59-5 



60-0 



^5, . 

 5 to 7, 

 ^7, 



Total, 



1 

 8 



1 

 18 



2 

 27 



3 



2 

 57 



5 

 27 

 61 



12 

 27 

 10 



13 



1 

 4 



3 



4 



5 

 1 

 1 



9 



19 



29 



62 



93 



49 



18 



7 



7 



Notwithstanding their short range these tables show plainly that when the 



