g TESTING MISCELLANEOUS SUPPLIES. 



as the oil gives values to which this correction does not apply. If the correc- 

 iion is applied to such values, the results obtained for 15.5 C. will contain 

 large errors. This is well illustrated by the following example : 



The specific gravity of a sample of oil was carefully determined at two tem- 

 peratures, with the following results : 



(1) Specific gravity at j 



(2) Specific gravity at r^r =0.9298. 



If the temperature correction is applied to the second value, the result is 

 as follows: 0.9298+9.5X0.00064=0.93588 (specific gravity at 15.5 C.), an error 

 far too large to be admissible. Now, from the known density of water at vary- 



ing temperatures," the specific gravity at ^^ ? be calculated from equa- 

 tion 2, as follows: 



If to this value tha correction is applied the result is 0.92795+9.5X0.00064= 

 0.93403, a value differing less than two in the fourth place from that actually 

 determined, which is sufficiently accurate for most commercial work. 



It is a simple matter to determine the specific gravity at any selected tem- 

 perature as compared with water at 15.5 C., but to do so most easily it is 

 necessary to properly calibrate the pyknometer or plummet. The coefficient of 

 cubical expansion of glass varies somewhat, and, while this correction should 

 be applied when necessary in calibrating, care should always be taken to apply 

 the correction for only a very few degrees. The Bureau of Standards in cali- 

 brating volumetric apparatus assumes the cubical coefficient of expansion of 

 glass per degree centigrade as equal to 0.000025, and if this value is always 

 used, on determinations within a very few degrees of the desired temperature 

 no appreciable error will be made. Since it is not safe to assume that the 

 cubical coefficient of expansion of glass is the same for all temperatures, appa- 

 nitiis should be calibrated at a temperature very near that at which the deter- 

 mination is to be made. 



2. Westphal Balance. 



Tho \\Ystphnl balance, on account of its convenience, is frequently used in 

 makiiiK HjMTiHc-Knivity determinations, but it is difficult to determine the 

 errors for this instrument. The weights can be easily tested, but the divisions 

 <>M the arm are not easily tested, and both may be somewhat incorrect. It is 

 I "-MIT not to use a Westphal balance, but a plummet may be used on an ana- 

 lytical i>:i lance, which, when properly calibrated, is a very useful piece of appa- 

 ratus, though the temperature is rather hard to control. 



3. Vacuum Weighing. 



It is not necessary to reduce weighings to vacuum, since the corrections 

 entering in both numerator and denominator would not make any appreciable 

 difference in the result. It should be understood that both in calibrating 

 the ni'punitiiR and in making the specific-gravity determination all weights 

 are made in air. 



Laiulholt. Bornstein, and Meyerhofer, Physikalische Chemische Tabellen, 

 pages 37 to ;$!: Bui-ran of Standards, Circular No. 19, page 27. 



