Still more remarkable and obscure in function are the vesicles found near the appendages; we figure a pair after Oudemans, showing them in the exserted state. In the retracted state the outer portion of the vesicles is withdrawn into the basal part _P_ (Fig. 95), so that the vesicles are then only just visible, being {189}concealed by the ventral plate. The abdominal appendage is not retractile. In _Machilis_ there are twenty-two of these vesicles, arranged either two or four on one ventral plate of the hind body. They are also present in _Campodea_, where there are six pairs. They are usually said to be absent in _Japyx_ and in _Lepisma_, but Haase shows[119] that _Japyx_ possess a pair placed behind the second ventral plate of the abdomen. The vesicles appear to be exserted by the entrance of blood into them, and to be retracted by muscular agency. Much difference of opinion prevails as to their function; it appears probable that they may be respiratory, as suggested by Oudemans. The scales found on the bodies of the Ectotrophous Thysanura may be looked on as modified hairs, and are essentially similar to those of the Lepidoptera, and they drop off as readily as do those of the Lepidoptera. Stummer-Traunfels, who has recently published[120] the results of his researches on the mouth-organs of Thysanura and Collembola, confirms the division of the Thysanura into Entotrophi and Ectotrophi, and considers that the Collembola agree with the former group. The German author therefore proposes to divide our Aptera, not into Thysanura and Collembola, but into Ectognathi and Entognathi, the former group consisting of Machilidae and Lepismidae, the latter of Campodeidae, Japygidae and the various families of Collembola. We think it far more natural, however, to retain the older division into Thysanura and Collembola. [Illustration: FIG. 95.—Abdominal appendage and exsertile vesicles of _Machilis_. _A_, appendage; _V_, vesicles protruded; _P_, basal portion; _R_, muscles, × 70.] COLLEMBOLA. The sub-order Collembola, which we have defined on p. 182, consists of small Insects, many of which possess the capacity of leaping, or springing suddenly, and when disturbed or alarmed naturally make use of this means of escaping. Their leaps, however, appear to be made quite at random, and very frequently do {190}not have the result of taking the creature into concealment, and in such circumstances they may be rapidly and frequently repeated until the Insect feels itself, as we may suppose, in a position of safety. Three families may be very readily distinguished, viz. (1) Lipuridae, in which no leaping apparatus is present; (2) Poduridae, a leaping apparatus exists near the extremity of the abdomen; the body is subcylindric and evidently segmented; (3) Smynthuridae, a leaping apparatus exists: the body is sub-globular with comparatively large head and abdomen, the intervening thoracic region being small; the segmentation of the body is obscure.