62 On the Sinicture of [Jan. 



volume of hydrogen weighs G'Q'AH. If, on the other hand, the 

 determination of Gay Lussac {Mem d' Arcueil, ii. 285) be not an 

 error of the press, it would make the volume of hydrogen to weigh 

 7*65. It is only in organic nature that we become acquainted with 

 other degrees of oxidation of hydrogen besides water. In ammonia 

 6 volumes of hydrogen are combined with 1 volume of oxygen : 

 in oxalic acid 1 volume of hydrogen is combined with 18 volumes 

 of oxygen. 



(To he continued.) 



Article VI. 



On the StriLchire of Crystals from Spherical Particles of Matter. 



By Mr. N. I. Larkin. 



(To Dr. Thomson.) 



The handsome manner in which you Introduced my former 

 communication encourages me to trouble you with a few properties 

 1 have since discovered peculiar to the tetrahedral pile of spheres. 



That a rhomboidal dodecahedron may be formed from a cube by 

 application of quadrangular pyramids to each face is known. I 

 attempted to construct it of spheres on that principle, and found 

 that a cube of three spheres on each edge was capable of being 

 converted into a dodecahedron by applying at each face a quad- 

 rangular pyramid of five spheres ; so that each of the four spheres 

 composing the base may stand against the side of the cube instead 

 of being opposite to the angles : the result was unexpected, for it 

 proved, when completed, to be an octohedron with a single sphere 

 on each face, agreeing in that general property with the cube 

 described in my former paper: the difference is in the dimensions 

 of the octohedrons. In the cube the octohedron has two spheres 

 on each edge : in the rhomboidal dodecahedron it has five. 



I have likewise succeded in forming the twenty-four sided 

 figure common to the white garnet, from a rhomboidal dodecahe- 

 dron, which was constructed from an octohedron having thirteen 

 spheres on each edge; it proved to be a canted cube with a single 

 sphere on each of the triangular faces, and a peculiar kind of 

 quadrangular pyramid on each of its square faces. 



The following properties of those figures in their first rudiments, 

 wliich have been hitherto formed of spheres, according to the 

 tetrahedral structure, may be worth reciting. The tetrahedron, 

 octohedron, and rhomboid, are the most simple, having no other 

 spheres than those which form their solid angles; the first two are 

 the only ones that can properly be called simple, and have nothing 

 apparently in common but their faces ; the third may be considered 

 as two of the first joined base to base, or as one of the second with 



