FORM OF MINERALS.	57
	infinite number of lines, parallel to the ideal axis of the figure. The same irregularity carried to a greater extent, frequently causes certain faces required for the symmetry of the form, altogether to disappear. Again, some crystals do not fill the space marked out by their outline, holes and vacancies being left in the faces, occasionally to such an extent that they seem little more than mere skeletons. This appearance is very common on crystals produced ar­tificially, as in common salt, alum, bismuth, silver, &c. A perfect crystal can only be produced when, during its for­mation, it is completely isolated, so as to have full room to expand on every side. Hence the most perfect crystals have been originally imbedded singly in some uniform rock mass. Next to them in perfection are forms that grow singly on the surface of some mass of similar or distinct composition, especially when the point of adherence is small. An incompleteness of form, or at least a difficulty in determining it, arises from the minuteness of some crys­tals, or from their contracted dimensions in certain direc­tions. Thus some appear mere tabular or lamellar planes, while others run out into acicular, needle-shaped, or capil­lary crystals. Amid all these modifications of the general form of the crystal, of the condition and aspect of its indi­vidual faces, or of its linear dimensions, one important ele­ment, the angular measurement, remains constant. In some monoaxial crystals, indeed, increase of temperature produces an unequal expansion in different directions, slightly chang­ing the relative inclination of the faces, but so small as to be scarcely perceptible in common measurements, and hence producing no ambiguity. More important are the angular changes which in many species accompany slight changes in chemical composition, particularly in the relative propor­tions of certain isomorphous elements. But notwithstand­ing these limitations, the great truth of the permanence of 33