Still farther out, below low-water mark, where one must wade, can be seen the beautiful branchial crowns of various Sabellids protruding from their tubes; but care is necessary on approaching these worms, as eyes are, in many cases, present on the branchiae and a shadow is readily perceived; then the brightly-coloured tuft disappears, and only a piece of sandy or muddy cylindrical tubing remains to tell where the _Sabella_ has withdrawn. In order to obtain the worms one must dig quickly and deeply before they have been disturbed; for the tube is of considerable length, and the inhabitant withdraws to the bottom of it. Some of these soft-skinned worms have the power of boring into hard rocks,[343] though by what means they do so is {287}uncertain.[344] _Polydora ciliata_ makes a tube of mud projecting from the mouth of U-shaped galleries in chalk, limestone, shells, and even shale; it has no hard jaws or other structures sufficient to account for the holes, but it is possible that the specially strong chaetae on the sixth segment may be of some use in this work. Other lithodomous worms are _Sabella saxicava_ and _Dodecaceria concharum_, which is a common little borer, forming galleries in oyster-shells, etc. The TUBES formed by these Polychaetes are very varied in constitution.[345] In some cases a mucus, which hardens to form a firm protective envelope, is secreted from special parts (_e.g._ the ventral gland shields of Terebellids and Sabelliformia), or from the greater part of the general surface of the body; in other cases the secretion serves to stick together particles of mud or sand, or shelly fragments, so as to form a more or less cylindrical tube (rarely branched), which is lined internally by the hardened "mucus," having the appearance of silk. [Illustration: FIG. 151.—_Clymene ebiensis_ in its tube (_t_) (from _Règne Animal_). _a_, Anterior, _p_, posterior end, which is, however, injured.] But the process of tube-making is not a simple one, for in many cases, at least, the worms exhibit definite powers of choice. Thus some species of _Sabella_ choose only the very finest particles of mud; _Terebella conchilega_ chooses fragments of shell and grains of sand; _Onuphis conchylega_ employs small stones more or less of a size; _Sabellaria_ makes use only of sand grains. Whilst some worms, like _Terebella_, _Nicomache_, and others, make a very irregular tube, _Pectinaria_ builds a most remarkably neat house, open at each end, which it carries about with it, the narrow end uppermost (Fig. 152); the grains of sand are nearly all of the same size and only one layer in thickness, embedded in abundant "mucus," and with the outer surface quite smooth.