The tubercles are outgrowths of the body-walls; they are either smooth, warty, or spiny, as in many caterpillars. While the armature of insects is of little morphological importance, it is evidently of great biological importance, the welfare or even the life of the insect depending upon it; and it varies in each species of insect, especially in Diptera, where the position of even a single seta characterizes the species. [Illustration: FIG. 218.—Section through an antennal pectination of _Saturnia carpini_: _a_, hypodermis, formative cells of the hairs (_c_); _d_, cuticula; _e_, trachea.—After Semper. ] [Illustration: FIG. 219.—Flattened hairs from the lateral tufts of larva of _Gastropacha americana_: _A_, three from the lateral tuft of _Heteropacha rileyana_. ] The mode of development of the hairs was first described by Semper. In the pectination of the antenna of _Saturnia carpini_ he observed that the hairs arise, like the scales of the wings, from large round formative-cells lying in the cavity, which send out through the hypodermis and cuticle a long slender process which finally becomes the hair (Fig. 218). Tactile hairs are those setæ arising over nerve cells or nerve terminations and will be discussed under the organs of sense. [Illustration: FIG. 220.—The same in _G. quercifolia_: _a_, a small hair ending in two minute processes. ] =Scales.=—In very rare cases the hairs of caterpillars (Fig. 219) are flattened and scale-like, and this passage in the same insect of cylindrical hairs into flattened scale-like ones, shows that the scales are only modified hairs. Also, as we shall see farther on, Semper has proved that their mode of origin is identical. While true scales are characteristic of Synaptera (Thysanura and Colembola), as well as Lepidoptera and Trichoptera, they also occur in the Psocidæ (Amphientomum), in many Coleoptera (Curculionidæ, Cleridæ, Ptinidæ, Dermestidæ, Byrrhidæ, Scarabæidæ, Elateridæ, and Cerambycidæ), and in the Culicidæ, and a few other Diptera, though they are especially characteristic of the Lepidoptera, not a species of this great order being known to be entirely destitute of them. [Illustration: FIG. 221.—Flattened and spinulated hairs of tufts of larva of _Acronycta hastulifera_. ] [Illustration: FIG. 222.—Scales from dorsal tuft, on second thoracic segment of larva of _Gastropacha quercifolia_. ] The scales vary much in shape, but are more or less tile-like, attached to the surface of the body or wing by a short slender pedicel, and are more loosely connected with the integument than the hairs, which are thicker at the base or insertion than beyond.