[Illustration: FIG. 139.—_Aphrodite._ Foot, × 2. _a_, Elytron; _b_, notopodium; _c_, neuropodium; _d_, neuropodial cirrus; _n_, aciculum; 1, iridescent bristles; 2, stiff chaetae; 3, felt.] GILLS.—We have already seen that several different organs, _e.g._ the palps in Sabelliformia, the prostomial tentacles of Chlorhaemidae, and the notopodial cirri of sundry other Polychaetes, may take on a respiratory function. There are, however, certain "gills" developed either on the parapodium itself or elsewhere on the body which it is difficult to homologise. Such are the _retractile gills_ on the parapodia of the Glyceridae (Fig. 136, C); those of _Dasybranchus_, near the abdominal neuropodia; those of _Mastobranchus_, near the notopodia. _Nephthys_ has a sickle-shaped gill on the under surface of the notopodium. The long gill filaments at the posterior end of _Sternaspis_, again, are only doubtfully interpreted as the dorsal cirri of some of the posterior segments. {269}Since primitively the whole skin of the worm is respiratory, any part of the skin may become more or less specialised for this function, and chiefly, of course, on the more actively moving parapodia. The blood-vessels constituting the essential part of the "gill" may make use of any already existing outgrowth (such as a cirrus or a tentacle), or may push the body-wall out on their own account. INTERNAL ANATOMY. Probably those organs which have the greatest effect in modifying the shape of the body are the SEPTA, for we find in the long, free-swimming worms that these are regularly present throughout the body, and external "segmentation" of the body is well marked. In burrowing and tubicolous forms the septa are frequently incompletely developed, or more or fewer may be absent; and the body becomes less distinctly segmented externally, tends to vary greatly in diameter during movement, or becomes plumper. With the disappearance of the septa there is also a diminution in the number of NEPHRIDIA, as in _Arenicola_, with only six pairs. Further, there is frequently a dimorphism of these organs; instead of all of them serving equally as excretory organs and as genital ducts, some of the most anterior in the Sabelliformia and Terebelliformia become greatly enlarged, and take on practically the whole of the former function; whilst more or fewer of the posterior nephridia dwindle in size, and become genital ducts. The absence of septa allows a free communication between the successive segments, and thus a freer flow of coelomic fluid for the distension of the anterior end of the worm during burrowing.