SOLID ELEMENTS, AND THEIR PERIODIC RELATIONS. 57 



fore its specific gravity was 12.16. The platinum weighed 166.534 grams; 

 its loss of weight suspended in water at 20 was 7.793 grams ; therefore 

 its specific gravity was 21.37. These values, which have not been reduced 

 to the vacuum standard or to 4,* are all near the accepted values, corre- 

 sponding to the pure metals. In the following table the true densities, 

 corrected to the vacuum standard, are given. 



The thallium contained a trace of lead, but since the compressibility of 

 thallium is not far from that of lead, this trace could have produced no 

 essential effect upon the results. The weight of the bar in air was 89.458 

 grams, and its loss of weight in water was 7.540 grams. Therefore its 

 specific gravity at 20 was 11.86. 



The iron was a piece of wrought iron, commercially pure. Manganese 

 was obtained from two sources, one, a sample from a German firm 

 intended to be pure, and another, undoubtedly purer, kindly given us by 

 the Roessler & Hasslacher Chemical Company of New York. To this 

 firm also we are much indebted for their interest in our investigation and 



o 



for their gifts of material. The two samples of manganese had slightly 

 different densities, but were not very different in their behavior. 



Of chromium two samples were used, one specimen prepared by Gold- 

 schmidt's method a sample of the very earliest product of this method 

 shortly after its invention the other, a much purer sample, furnished by 

 the Roessler & Hasslacher Company and stated by them to contain 99 

 per cent of chromium. Between these two samples there was a very great 

 difference in compressibility, the first sample being far less compressible. 

 We are inclined to think that this first sample contained a silicide of 

 chromium or some other less compressible compound, and therefore accept 

 the later result with chromium of known purity as the more accurate. 



Nickel was placed in the jacket in the form of the small cubes in which 

 the pure material may be procured in commerce. These cubes are per- 

 meated in every direction with fine pores or channels, and upon applying 

 pressure to the mercury in which they are immersed, this mercury is 

 driven into the innermost recesses of the lumps, giving false values for the 

 compressibilities. Various devices were tried in the effort to stop these 

 channels, but finally, the simple expedient of hammering each lump heavily 

 on all sides proved to be very efficient and prevented the entrance of any 

 appreciable amounts of mercury into the interior of the metal. 



The molybdenum used in the experiments was a sample also kindly fur- 

 nished by the Roessler & Hasslacher Company. Its purity was stated by 

 them to be 98 per cent. 



*This reduction causes so small a change in the values as to have no important 

 effect in the present work. The ur.corrected values are given here for the sake of 

 more ready comparison with the work of others, which is usually not corrected. 



