SOLID ELEMENTS, AND THEIR PERIODIC RELATIONS. 63 



each of these experiments is essentially at least a three-fold repetition of 

 the same process, and those experiments which were duplicated gave a yet 

 higher degree of certainty. 



The study of the curves shows that changing pressure produces prac- 

 tically no effect upon the compressibility of these comparatively incom- 

 pressible substances. Such change as may occur is beyond the accuracy 

 of the method, and can only be determined when larger quantities are used. 

 This is in accord with Amagat's experience with glass, which is much 

 more compressible than most of the metals. 



There are only a few older results with which to compare these, but all 

 which have been found are higher than those given in the table on pp. 61, 

 <>2. Regnault measured the compressibility of copper, and Amagat meas- 

 ured this quantity also, as well as the compressibilities of steel and lead. 

 The more recent work of Lussana* included tin, lead, cadmium, zinc, and 

 copper ; it was executed by an ingenious method, and yielded results much 

 higher even than Amagat's. It is interesting and significant to note that in 

 spite of the great discrepancy in the absolute values, the relative values of 

 the five metals as given by Lussana correspond approximately with the 

 relative values recorded above. Lussana places the five elements in the 

 following order, beginning with most compressible, and ending with the 

 least: Lead, cadmium, tin, zinc, and copper. The results in the table 

 above fall in the same order of succession. Hence, although the absolute 

 values obtained by Lussana were not far from twice as great as ours, the 

 relative values do not contradict ours. Because relative values rather 

 than absolute ones were chiefly sought in the present research, the present 

 results are published with confidence that they will be of use even if time 

 should show that Lussana's scanty figures are more absolutely accurate. 

 It may be added that careful scrutiny has not yet been able to discover 

 any plausible reason for supposing that our low values are less than the 

 truth. The divergence of the results gives interesting evidence of the 

 great difficulty in determining the compressibilities of very incompressible 

 substances. 



COMPARISON OF RESULTS. 



Before summing up the results in the large table and those previously 

 found concerning the elementary substances, it may be illuminating to 

 compare these magnitudes with other well-known compressibilities. The 

 compressibility of mercury is somewhat less than one-tenth of that of 

 water, which is usually considered as a rather incompressible substance. 

 A few of these compressibilities are not far from that of mercury, espe- 



*Regnault and Amagat, Landolt-Bornstein Tabellen (2d Edition), p. 278 (1894). 

 Lussana, II Nuovo Cimento [5], 4, 371 (1902) ; 5, 153 (1903) ; 7,355 (1904). The latter 

 recent work became known to us only after the completion of our own work. 



