48 FORMATION OF SHELLS OF ANIMALS, ETC., 
And as the same reasoning would apply to the molecules 
on the other side of the line c, c’, a quadrilateral area 
would be enclosed corresponding to the superficies of a 
primitive crystal. And extending this reasoning to similar 
groups situated on all sides of the one first supposed, a 
solid quadrilateral form would result corresponding to 
that of a primitive crystal. Though all crystals thus 
formed would be quadrilateral, yet their exact shape will 
vary according to the size of the angle a, c, s. If this 
angle be less than 45°, then, as the three angles of 
every triangle are equal to two right angles, c, a, c’ will 
be greater than a right angle, and the form of the crystal 
represented by ac, 4c, will be rhomboidal or lozenge- 
shaped. But if the angle a, c, B be exactly 45°, then the 
angle c, a, B will be a right angle, and all the angles of 
the figure a c, a’ c will be right angles, and therefore the 
crystal represented by this figure will be rectangular. 
Now the molecules at the point 3B, being impelled in 
opposite directions and with an equal force, will be kept 
in their place without the aid of gravity, and therefore 
here, where the sine of the angle a, c, B vanishes, the 
attraction acting upon the molecules will be at zero. 
But as the force of attraction required to balance that of 
repulsion estimated from the point B increases as the sine 
of the angle a, c, B, the density of these crystals, being as 
the attractive force holding their molecules together, will 
be least at the centre and greatest at the surface. This 
inference agrees with several recorded facts connected with 
the polarization of light by crystals. As, during the 
formation of the primitive crystals of different crystallizable 
substances, or during the formation of those of the same 
substance under different circumstances, different quan- 
tities or intensities of electricity would doubtless be evolved, 
