ON THE GENERAL MORPHOLOGY OF THE MYXINOID FISHES. 737 



demarcation, into the posterior section of the constrictor pharyngis {cp'".). I have 

 taken the boundary to be where the first loop merges into the anterior limb, but this 

 is an arbitrary distinction. 



First Loop (1). — On the left side this is somewhat similar to the condition repre- 

 sented in fig. 3, but on the right it did not reach the longitudinalis linguae (I. lg.), but 

 fanned out on the dorso-lateral surface of the external wall of the second pleural sack 

 in front of its efferent gill duct. A small bundle was sent backwards on to the root 

 of the gill duct. Hence, on the left only the ventral fibres were present, and on the 

 right only the non-ventral, the two sides together representing the fibres which may 

 be present on each side, as shown in fig. 3. 



Second Loop (#). — Takes no part in forming the definitive ventral longitudinal 

 tract (7), but terminates freely in front in the fatty areolar tissue slightly below and 

 to the side of the cardiac aorta at about the level of the origin of the second afferent 

 gill duct. This loop despatches an anterior slip to the second efferent gill duct, 

 another anterior and smaller one at the same level to the fatty areolar tissue external 

 to this duct, and a third postero-ventral one to the first efferent gill duct. On the 

 right side the loop terminated in front mostly on the inner surface of the first efferent 

 gill duct. There seems little doubt that the ventral forking of the loop corresponds 

 to the anterior and posterior fibres contributed to the ventral longitudinal tract. 



Third Loop (3). — Inserted by a very narrow anterior and a very wide posterior 

 bundle into the superficial internal fascia of the parietalis, not far from its ventral border 

 and in front of the fourth gill sack. 



Fourth Loop (4). — Inserted in front into the superficial internal fascia of the parie- 

 talis in the same way and at the same level as the third loop. 



Behind the fourth loop are seen the five large processes into which the ventral border 

 of the posterior external sheet of the constrictor is prolonged. These are mostly 

 inserted into the internal fascia at the junction of the ventral margin of the parietalis 

 with the obliquus. The insertions of the third and fourth loops just described continue 

 this series forwards. 



Ventral Longitudinal Tract (7, 7'). — The sections throw considerable light on the 

 nature of this tract. The postero-ventrally coursing bundles, forming part of loops 

 3 and 4 and the whole of loop 5 (3, 4, 5), correspond unquestionably in every detail 

 with the ventral bundles previously described from dissections, of which the third and 

 fourth may contribute to the ventral tract, and the fifth usually does not. In the 

 sections the ventral bundles of loops 3 and 4, after a typical course, combined to form 

 a separate or anterior section of the ventral longitudinal tract (7), which, passing 

 forwards immediately ventral to the first efferent gill duct, terminates in front on the 

 obliquus muscle not far from the lateral margin of the rectus. The greater part of this 

 section, i.e. between the points of entry of the loops, is formed of fibres from loop 3 

 passing backwards and of loop 4 passing forwards. Both loops, on entering the section, 

 despatch fibres forwards and backwards. Posteriorly the section splits, leaving a lacuna 



