ON THE GENERAL MORPHOLOGY OF THE MYXINOID FISHES. 729 



arteries were given off— the anterior two emerging to the left of the median linea and 

 the third to the right. In front of and behind the muscle the aorta emerges to become 

 opposed to the notochord, i.e. its normal position. Ventrally the muscle extends so far 

 as to cover the whole of the last gill and the greater part of the fifth, and also the last 

 three ductus branchiales externi. Its ventral margin may be quite regular (cp. fig. 13) 

 or very ragged (J. Muller evidently saw this latter condition, and cp. the Chart, fig. 2), 

 and the whole of this margin may be connected by a thin ligamentous sheet with the 

 body-wall, either exactly opposite the dorsal border of the obliquus or the ventral border 

 of the parietalis, or the posterior portion of the muscle may extend even further down- 

 wards so that the sheet is attached behind to the body-wall opposite the outer edge of 

 the rectus. 



Dorsally, the sheet-like portion of the constrictor extends much further forwards 

 than ventrally — reaching as far as the second gill. Opposite the fourth gill it 

 bifurcates, sending forward on each side a narrowing muscular limb (cp. fig. 3, c.b. c'.), 

 separated by a median space from that of the other side. This limb, as shown in the 

 figure, was in this specimen separated from the main portion of the sheet by a 

 ligamentous partition, but this was certainly not the case in other individuals. In one 

 example the two limbs were connected by a dorsal transverse muscular bridge, itself 

 with a median longitudinal partition, and the whole anterior portion of the muscle 

 was modified. The anterior limb above passed forwards and slightly downwards, 

 becoming gradually narrowed, until over the second gill it merged into a gradually 

 narrowing ligament which curved upwards and forwards to be inserted into the 

 superficial internal fascia of the parietalis at the side of the notochord. This ligament, 

 which is certainly of exceptional occurrence, had a curious connection with the 

 posterior fibres of the third portion of the constrictor pharyngis, these fibres arising 

 from the ligament, and curving forwards over the first gill to reach their insertion on 

 the copulo-copularis (cp. the constrictor pharyngis). 



The arrangement just described seems to indicate that the third portion of the 

 constrictor pharyngis, if not the first two, corresponds to a modified loop (or loops) of 

 the constrictor branchiarum. If so, they would represent the loops in connection with 

 the anterior or lost gills, the first adult gill being otherwise without a loop. We must, 

 however, not forget that the last loop of all, i.e. the one immediately in front of the 

 ductus oesophago-cutaneus (figs. 2, 3, 13, 6), is also unconnected with a gill in the 

 adult. This view is strikingly borne out in most specimens, as there is no break 

 whatever between the two muscles (cp. the Chart, fig. 2), the only difference being 

 that instead of the fibres of the constrictor branchiarum passing on to the parietalis, 

 they meet each other in the middle line over the gut, as already figured and described 

 for Myxine by J. Muller. Further support is to be found in the condition in 

 Bdellostoma, for the third division of the constrictor pharyngis of Myxine unquestion- 

 ably includes J. Muller's first branchial loop of Bdellostoma. 



The dorsal anterior limb of the constrictor sends downwards two muscular loops 



