ON AMPHILINA PARAaONOPORA 59 



anterior end of the worm to form a thick-walled invaginable 

 conical bulb (Pis. 3, 4, figs. 1, fc, and 16). In transverse section 

 (PI. 4, fig. 18) it can be seen that the cavity of the retracted 

 bulb is star-shaped in outline, and in longitudinal sections of 

 the protruded proboscis it is seen that the portion of thickened 

 wall at the very extremity of the worm forms a kind of terminal 

 cushion (PI. 4, fig. 16, tt). The lateral walls of this bulb 

 resemble the ordinary body-wall in general histological structure, 

 but there are three very striking and significant differences, 

 and had one of these been observed by previous writers, the 

 true function of the proboscis and the huge so-called ' retractor ' 

 muscle attached to it would have been obvious. The first 

 difference is the total absence of circular and longitudinal 

 muscle-layers, the second is the presence of very distinct large 

 radial muscle-fibres (PI. 4, fig. 16, rtm) which extend up to 

 the base of the cuticle (PI. 4, fig. 19), and the third the minute 

 but well-marked serration of the cuticle (ser) which covers 

 the greater part of the outer surface of the protruded proboscis 

 (PI. 4, figs. 16, 16 a) but not the terminal thickening (tt), 

 which is altogether devoid of a cuticle. This serration of the 

 proboscis cuticle has not been previously observed. With 

 regard to the second difference mentioned, the radial fibres 

 (rtm) which originate immediately under the serrated cuticle 

 at first run longitudinally, i. e. parallel with the cuticle (PL 4, 

 fig. 16, a), but soon bend nearly at right angles ^ and run direct 

 to the surface of the big muscle (bm) shown in PL 4, fig. 16 

 as occupying the axis of the proboscis and then l)end again and 

 run parallel to the surface of the big muscle posteriorly, and 

 at the hind end of the proboscis these muscle-fibres diverge, 

 extend posteriorly in the general parenchyma of the central 

 core of the body for a considerable distance, and finally 

 apparently become continuous with the parenchyma. It is 

 difficult to determine how far these fibres extend posteriorly 

 because they are not easily distinguishable from the ordinary 



^ Some anteriorly attached fibres do not bend but run longitudinally 

 in the proboscis wall, crossing the greater number of fibres in so doing. 

 They do not form a layer of longitudinal muscles. 



