VARIATIONS OF SPP:CIFIC (GRAVITY. 263 



which are too soft — lead, I'or example — have too little cohesion, and 

 are torn in attempting to draw them into wires. 



In the process of "wire drawing- the metals are then in fact com- 

 pressed, and in accordance with what has been said, a certain dim- 

 inution of density is to he expected if the pressure is high enough, 

 but there are no certain grounds to predict what might happen in 

 further continuation of the process. 



1 had formerly found advantage in the use of the ]\vknometer, in- 

 stead of the Archimedean method of determining specific gravity, 

 but had found an improvement in the latter when the supporting 

 wire is carefully platinized. By the friendly advice of Dr. Ch. Ed. 

 Guillaume and Dr. P. Cha]jpius, both of the Bureau International 

 des Poids et Mesures in Pavilion de Bivteuil in Sevres, who have 

 been good enough to impart to me and my colleagiu^ the approved 

 method employed at that institution, we have been able to avoid com- 

 pletely the error formerly encountered, and have employed the 

 method of displacement, with the same success we had previously 

 obtained by the pyknometer. 



I give here only a snuill part of the results obtained, and choose 

 for the purpose of illustration the experiments with i)latinum. It is 

 not onl}^ a duty but a privilege to acknowledge in this connection the 

 uncommon goodness of the firm of W. C. Heriius, in Ilanau, who 

 have not only furnished me liberally witli the costly matei-ial. but 

 also, to avoid error, have separately drawn all the wires iviiuired for 

 the research. 



Three parallel series of experiments were carried out, as follows: 

 From a single bar of the llertius connnercial platinum pieces lying 

 side by side were cut off, and three cylinders, each about 6 mm. in 

 diameter and 35 mm. long, were turned to a uniform weight of 32 

 grams, then sent to us in Basel, where we determined their specific 

 gravity. We then returned the samples to Hanau, where by rolling 

 and hammering they were reduced to a diameter of 3 and length of 

 55 mm., suitable for wire drawing. The specific gravity of these 

 samples was again measured in Jiasel. Returning them to Hanau, 

 they were drawn successively to wires of 1, O.T, and 0.4 nun. diameter. 

 After each drawing complete determination of the specific gravities 

 of these wires was made in Basel, and they were then before further 

 drawing softened by heating to incandescence for three" minutes. 

 But before being further drawn — the wires were always cold 

 drawn — the density of these softened wires was determined, so that 

 we made 3 times 3 determinations on hard wire and the sam(> number 

 on soft wire." 

 The results which I now pi-oceed to give aiv the mean of tliree re- 

 ef Here follow in the original article full details of a single deterniiiiation, 

 from which the. extreme care and ac^-uracy of the work is ajiparent. 



