614 



miscellanp:ous papers. 



mass as a cubic decimeter of pure water at maximum density, and the 

 experimental determination of that mass was made by tiudiug the differ- 

 ence of weight in air and in water of a hollow brass cylinder 

 whose exterior dimensions at a temperature of 17.6o C. were, height 

 = 2.437G72 decimeters, diameter = 2.428368 decimeters, volume 

 = 11.2900054 cubic decimeters. The difference of weight in question 

 was first measured in terms of certain brass weights, by the aid of 

 which the platinum kilogram of the archives was subsequently con- 

 structed, special care being taken to apply the corrections necessary 

 to reduce all the weighings to what they would have been if made in a 

 vacuum.* 



The best results hitherto obtained for the weight of a cubic decimeter 

 of water, expressed in terms of the kilogram of the archives, are as 

 follows : t 



Date. 



1795.... 

 1797.. I 

 1821.. S 

 1825.... 

 1830.... 

 1841..:. 



Coiiutrv- 



France . . 



England 



Sweden . 

 Austria . 

 Russia .. 



Observer. 



Lef evre-Gineau 



Slinckburgh and Kater 



Berzelius, Svanberg, and Akermann 



Stampfer 



Kupfter 



Mean - 



Weigbt of a 

 cubic decime- 

 ter of water 

 at 4° C. 



Grams. 



1000. 000 



1000. 480 



1000. 296 

 999. 053 

 999. 989 



1000. 084 



These results show the extreme diflficulty of determining the exact 

 mass of a given volume of water. The discordance between the dift'er- 

 ent observers amounts to more than one part in a thousand, while good 

 weighings are exact to one part in eight or ten millions. Without 

 doubt two weights can be compared at least a thousand times more 

 accurately than either of them can be reproduced by weighing a speci- 

 fied volume of water, and for that reason the kilogram, like the En- 

 glish pound, can now be regarded only as an arbitrary standard of 

 which copies must be taken by direct comparison. As already stated, 

 the kilogram is equivalent to 15432.34874 English troy grains, or about 

 2 pounds 3 ounces avoirdupois. 



In consequence of the circumstance that the mass of a body is not 

 affected either by temperature or flexure, weighing is an easier process 

 than measuring; but in order to obtain precise results many precau- 

 tions are necessary. Imagine a balance with a block of wood tied to 

 its right-hand pan and accurately counterpoised by lead weights in its 

 left-hand pan. If with things so arranged the balance were immersed 

 in water the equilibrium would be instantly destroyed, and to restore 



' Base du Systeme M6trique, T. 3, pp. o74-'5. 



t This table Uas been deduced from the data given l#y Professor Miller iu 44, p. 760. 



