247 
Matthiesien, and gave a mean sp. gr. of 2 47. The remainder were 
inferior samples, and probably contained zinc and other impurities. 
Professor Miller has kindly examined for me the specific gravity 
of a good specimen of graphite silicon in his possession (not analyzed), 
and found it to be 2*337. 
Deville gives as the specific gravity of boron diamond 2*68. 
The crystalline form of the boron diamond is the same as that of 
the carbon diamond, and similar relations seem to exist between the 
specific gravity and atomic weights. The atomic weight of boron is 
7* 2, viewing its oxide as corresponding to carbonic acid in composition. 
VT2 = 2-683. sp. gr. = 2-68. 
But the same relation would not appear to hold for silicon, which 
does not affect the like tendency to crystallize in the same forms 
as carbon and boron, although the relations between the numbers in 
its case also are in the same direction, and not devoid of simplicity. 
The atomic weight of silicon is 14*2. 
Si \Zl4*2 = 2*42 sp r gr. of diamond silicon, = 2*46 to 2*48. 
Si V28-4 = 2*30 sp. gr. of graphite silicon, =2*33. 
The differences exhibited in this case from the similar forms of 
Carbon and Boron are not sufficiently marked to throw doubt upon 
the relations as being due to some unexplained law. As an arith- 
metical probability, indeed, the discordance lessens the value of the 
testimony in the previous cases. But our chemical knowledge of 
the manner in which Silicon doubles and quadruples itself in the 
silicates, to unite with the same quantity of base, gives support to 
the idea that its atomic weights may be different in the various 
forms of the separate element. 
When we consider how much we multiply the errors of experi- 
ment in raising the observed specific gravities to the second, third, 
and fourth powers, it remains scarcely possible that the’ simple 
relations between them and the atomic weights, in the cases which 
I have pointed out, can be due to chance. I have purposely avoided 
any speculation as to the bearing which these relations may have 
on the molecular arrangement of the particles of the elements in 
their various forms, as I desire, in the first place, to submit the 
testimony on which the relations themselves are founded to the 
consideration of chemists. 
