32 



spirit of the Act, and hence in any official scientific verifications 

 this mode is adopted. The tare of the vessel to be examined, 

 together with the glass plate, is taken, and this weight, to- 

 gether with the weight of water it should contain, is placed in 

 one pan of the scales, and in the other the measure filled with 

 water. The errors due to temperature and barometric pres- 

 sure are observed, and balanced bj means of weights in either 

 pan, according as the error is plus or minus. If correct, the 

 balance will be in equilibrium. As a difference of 1° Fahr. 

 between 62° and 70° will occasion an error of from four to six 

 grains in the gallon, or thirty-two to forty-eight in the bushel,, 

 the importance of observing the temperature will be evident. 



The reason for specifying brass weights by the Act will be 

 evident after the observations on specific gravity of weights,, 

 and it is done to fix the conditions ; but as the specific gravity 

 of water is very different to that of brass, it appears to me that 

 consistency demands the same scientific accuracy in this point 

 as in others, especially when it is remembered that as fine, or 

 finer, corrections are applied which are due to other circum- 

 stances ; and even the weighing of the platinum pound weight, 

 which is so little affected by the buoyancy of the air on account 

 of its great specific gravity, has to be reduced to vacuo. 



Captain Kater, in his researches, takes the weight of brass 

 as eight times that of water ; the effect, therefore, of the 

 buoyancy of the air upon the brass will be one-eighth of that 

 upon the water, thus lessening the effect on the water by about 

 one-eighth of the whole quantity.* The calculation to give the 

 correction is 7T ^j x 1 x 70,000 grains = 2*46 grains = the 

 gain in the weight of water due to a depression of one inch of 

 the barometer in one gallon. Since Kater' s inquiries further 

 researches have been made by other scientists, especially by 

 Regnalt, whose results are accepted throughout Europe as 

 authoritative. It is true that this correction may be made 

 now in official comparisons, but I can find no account of it, 

 and I omitted to ask Mr. Chaney if it were observed when I 

 saw him. 



With reference to the expansion of gases under varying 

 conditions something has to be said further on under the head 

 of gas measurement. The question of saturation somewhat 

 complicates the matter, which would otherwise be simple and 

 satisfactory ; but with reference to water and the expansion of 

 liquids generally there is little satisfaction. The expansion of 

 water by heat is, as we know, subject to the, as yet, unaccount- 

 able phenomenon that from 32° to 39° it contracts, and from 



* Water is taken as 831 times the weight of air; this, multiplied by 

 thirty inches barometer for the pressure of air, gives 24,930. 



