GUIs of Arenicola 55 



ages, may present differences as great as those between crotchets 

 from different species but of about the same age. But soon after 

 the worm has settled down to its littoral habit the crotchets assume 

 the form characteristic of the species, and, henceforward, change 

 only gradually and in a definite manner (see p. 49). 



Among the caudate species the crotchets of A. pusilla are most 

 readily recognised by reason of the full rounded curve of their post- 

 rostral region, giving the free end of the crotchet the appearance of 

 a swan's head and beak. The crotchets of A. marina, assimilis and 

 glacialis, which present no dilatation of their post-rostral portion, 

 are so closely similar that their aid cannot be invoked in separating 

 these species from one another, but they may be of use in dis- 

 tinguishing this trio from the remaining species. The crotchets of 

 A. cristata are intermediate, in the form of their post-rostral region, 

 between those of A. marina and pusilla, and those of A. loveni are 

 intermediate between those of A. marina and cristata. 



GILLS. 



The gills are paired hollow outgrowths of the body- wall, situated 

 immediately posterior to the notopodia in certain segments. They 

 do not arise, in A. marina, until after the full number of chaetiferous 

 segments, and a considerable number about thirty of tail-segments 

 have been developed. The first indication of the formation of gills 

 is seen in the penultimate and two or three preceding chaetiferous 

 segments, in which, behind each notopodium, the segmental blood- 

 vessels form a well-marked loop, over which there arises a slight 

 elevation of the body-wall the incipient gill (cf. PI. X, Fig. 30). 

 The gills of these segments become successively conical, digitiform 

 and branched, and, meanwhile, gills make their appearance on the 

 last and on the more anterior segments until the full number 

 (thirteen pairs) has been attained. The gills are, from their earliest 

 stage of formation, respiratory structures ; there is no evidence that 

 they are modified cirri. 



The gills of A. cristata arise in a similar manner and sequence ; 

 they first make their appearance in young examples, about 5 mm. 

 in length, which have the full number (seventeen) of chaetiferous 

 and about twenty tail-segments. 



In A. ecaudata also, the gills do not arise until the worm has 

 almost attained its full number of segments. When the worn. 

 possesses about sixty fully formed segments, the formation of gill, Q 



