96 PROFESSOR J. C. EWART AND MR J. C. MITCHELL ON THE 



with the dorsal part, i.e., with what he calls the anterior part of the dorsal longitudinal 

 canal. The dorsal portion of the hyomandilmlar he represents accurately enough as con- 

 nected with the infra-orbital — his dorsal longitudinal canal. He makes no reference to 

 the innervation of any of the cranial canals, and, like Monro, considers the hyomandibular 

 division of the facial as a branch of the trigeminal. 



Garman, after pointing out that " the manner in which the Batoidei became possessed 

 of the ' pleurals ' is a question of considerable interest," states that the " clue to the 

 solution of the problem is to be seen in Chlamydoselachus " (5). He considers that the 

 dorsal portion (pleural canal) was derived from the spiracular, and the ventral loop (sub- 

 pleural canal) from the jugular. Elsewhere he says, " No doubt the pleural originated 

 as a branch of the orbital." Why Garman should suppose the " pleurals " originated 

 from the "orbital" (i.e., the first portion of the infra-orbital) canal, it is difficult to under- 

 stand. Some of the tubules of the infra-orbital open into the dorsal extension, and in 

 some cases, e.g. JR. ocellata, this dorsal portion is disconnected from the ventral portion. 

 But, on the other hand, both the ventral loop and the long dorsal extension are supplied 

 by the hyomandibular nerve ; and when the canals are examined in embryo skate, and 

 in the less specialised members of the Batoidei group, it becomes evident that the ventral 

 loop (subpleural) has been formed by a backward extension of the hyomandibular canal, 

 which turned upon itself (in some cases before reaching the level of the first branchial 

 cleft), and grew forwards to penetrate the snout and then proceed backwards along the 

 margin of the pectoral fin, thus forming, first, the ventral, and afterwards the dorsal 

 loop. 



IV. The Lateral Canal. — This canal has been frequently figured, and more or less 

 accurately described. Hitherto, it has usually been looked upon as beginning in the skate 

 at or uear the temporal commissure. The commissure has been said to consist of two canals 

 (the aurals), and the part between the end of the commissure and the supra-orbital canal 

 has been described as a separate canal (the occipital.) But, as the commissure and the 

 whole of the longitudinal canal, including its branches, from the beginning of the supra- 

 orbital to the tip of the tail, is innervated by one nerve, viz., the lateralis division of the 

 vagus, we shall consider the lateral canal as including in addition to the trunk portion, 

 the commissural portion and the part immediately in front of it. 



The first precommissural portion (occipital of Garman ; lp., fig. 6) runs backwards 

 and inwards to the outer end of the temporal commissure. It has the same diameter as 

 the cranial part of the supra-orbital, and lies for the most part in a groove in the cranial 

 cartilage. 



The commissure (lc., fig. 6), formed by the union of two branches, one from each 

 lateral canal, runs right across the cranium, immediately behind the auditory pores. It 

 lies immediately beneath the skin and firmly adheres to it. Neither of those portions 

 give off any tubules, nor do they open to the exterior by means of pores. 



The first part (L., fig. 6) of the lateral canal of the trunk, after running for a short 

 distance directly backwards, forms a sigmoid curve, and reaches the supra-scapular 



