6. _Color and Streak._—The colors of minerals are of two kinds,—_essential_ and _non-essential_. By the essential color in any case we mean the color of the mineral itself in its purest state. The non-essential colors, on the other hand, are chiefly the colors of the impurities contained in the minerals. Metallic minerals, which are always opaque, usually have essential colors; but vitreous minerals, which are always more or less transparent, often have non-essential colors. The explanation is this: In opaque minerals we can only see the impurities immediately on the surface, and these are, as a rule, not enough to affect its color; but in diaphanous minerals we look _into_ the specimen and see impurities below the surface, and thus bring into view, in many cases, sufficient impurity so that its color drowns that of the mineral. To prove this we have only to take any mineral (serpentine is a good example in our series) having a non-essential color, and make it opaque by pulverizing it or abrading its surface, when the non-essential color, the color of the impurity, immediately disappears; just as water, yellow with suspended clay, becomes white when whipped into foam, and thus made opaque. What we understand by the _streak_ of a mineral is its essential color, the color of its powder; and it is so called because the powder is most readily observed by scratching the surface of the mineral, and thereby pulverizing a minute portion of it. The streak and hardness are thus determined at the same time. The streak of soft minerals is easily determined by rubbing them on any white surface of suitable hardness, as paper, porcelain, or Arkansas stone. ESSENTIAL AND ACCESSORY MINERALS.—Lithologists, regarding minerals as constituents of rocks, divide them into two great classes: the _essential_ and the _accessory_. The essential constituents of a rock are those minerals which are essential to the definition of the rock. For example, we cannot properly define granite without naming quartz and orthoclase; hence these are essential constituents of granite; and if either of these minerals were removed from granite it would not be granite any longer, but some other rock. But other minerals, like tourmaline and garnet, may be indifferently present or absent; it is granite still; hence they are merely accidental or accessory constituents. They determine the different _varieties_ of granite, while the essential minerals make the _species_. This classification, of course, is not absolute, for in many cases the same mineral forms an essential constituent of one rock and an accessory constituent of another. Thus, quartz is essential in granite, but accessory in diorite.