104 M. C. Lea on Numerical Relations existing 
Viewed in this light the above series approaches very near to 
exactness, and may probably be wholly so, for the equivalent ¢ - 
i Dumas’ recent re 
copper is by no means positively known. 
searches were too unconcordant to permit him to offer a positive 
opinion. It will be shown further on that these metals, as well 
as many others here grouped together by relations depending 
upon a number approximating to 445, are also united by analo 
gies of atomic volume. 
The second number in the above series, 152°5, does not cor 
respond with the equivalent of any known element, and it, like 
the number 164, which occurs twice in the nitrogen seri 
Platinum, 
ridium, 9 
Palladium, 53-3 Osmium, 99°6 
The difference between the equivalents of platinum and pallad = 
jum is 45°4, the rest vary a little more. a 
The elements, carbon, boron and silicium, are united by the 
we take Gerhardié 
lage for carbon €=12, the sum of the equivalents of thes? 
number 44 in the following manner: if 
ree substances amounts to 44 exactly 
Carbon (€), 12 
Boron, il? 
Silicium, 21> 
With respect to the metals of the earths and alkaline car — 
magnesium and calcium have been considered elsewhere, , 
tly examined 
ee erbium and terbium, and the cerite metals have 
fact not yet been determin 
considered, but that it extends also to them becomes : 
* As determined by Deville » by analyses of bromid of boron. — 
, the Se of the three first ! 
not yet 
