OF ZINC AND ANTIMONY. 357 



Scheerer * was tried, but so great accuracy could not be obtained with it as with a 

 bottle. f On an average, about ten grammes of crystals J were taken for each determi- 

 nation, and the bottle used was capable of containing about the same weight of water. 

 The cnstals, coarsely powdered, were introduced into the bottle, covered with water, 

 and, on account of the action of SbZug on hot water, the entangled air was removed by 

 an air-pump. The bottle was then filled with water, and, after the stopper had been 

 introduced, suspended in a large beaker of water, the temperature of which was very 

 slightly higher than that of the room. In contact with it was placed the bulb of a 

 centigrade thermometer, graduated to tenths of a degree. ^\Tien an equilibrium had 

 been established in the temperature, the bottle was removed from the beaker, wiped 

 diy, and weighed. In calculating the specific gravity the weight of the water was cor- 

 rected for the temperature so that the unit is in all cases distilled water at 4° C. A 

 similar correction could not be made for the temperature of the substance, as the co- 

 efficients of expansion of the crystals are not known ; but as the maximum difference 

 between the temperatures in the different determinations was not over 10° C, this cor- 

 rection would only very slightly aff'ect the relative results. The mean temperature was 

 about 15" C. In order to show that very accurate results can be obtained by delicate 

 manipulation with a specific-gravity bottle, I will subjoin the numbers obtained in the 

 determination of the specific gravity of antimony and zinc. 



* Poggendorf, Annalen, Vol. LXVII. p. 120. 



t A specific-gravity bottle for delicate experiments should be made with a thick rim, ground square at the 

 top, and the glass stopper should be so fitted to the neck as not to leave a channel between the two in which 

 water can collect. 



t The very great liability to error which the use of a small amount of substance, in a specific gravity de- 

 termination, necessarily involves, does not seem to be appreciated by many experimenters, and it may there- 

 fore be of use to add a very simple mathematical statement of the process. Let M = weight of substance at 

 4° C. and m := weight of bottle filled with water at same temperature. Place x = the weight of bottle, sub- 

 stance, and water, as this is the weight on which the accuracy of each determination depends, since 

 the weight m is the same for all, and is the mean of a large number of observations ; then Sp. Gr. = 



u' 

 ^ u ; 8 u = — S .r. Here 6 u represents the amount of error produced in the specific gravity 



M 



{M-\-m)—x M 



40 



by making an error of Sx in the weight. Suppose u = 6.32, then S k = ^^ S ar, so that, if forty grammes of 



the substances are used, an error of one millegramme in the weight x will produce an error of one one-thou- 

 sandth in the specific gravity. Suppose, however, only one gramme is taken, then the same error in the 

 weight X will cause an error of four onc-hundredths in the specific gravity. From this it appears, that, 

 where only a limited amount of a substance is at command, it is best to unite it all in one careful experiment, 

 rather than to distribute it through several ; for it must be remembered, that, of two experiments in which the 

 liability to error is as one to four, the relative value is not as these numbers, but as their squares. 



