546 H R POLLARD, 



Comparison with the PI a cod er mi. Since the mucous 

 canals of fish run in the dermal bones it is evident that their course 

 may with favourable material l)e traced in fossil forms. Grooves have 

 been depicted by Pander in his work on the Placodermi which can 

 only have been occupied by mucous canals. The comparison of these 

 grooves with the tubes of the Siluroids yields results which I hope 

 to show have special value in determining the phylogeny of the Tele- 

 ostei. More than 20 years since, Prof. Huxley following up the 

 work of Pander came to the conclusion that the Placodermi possessed 

 undoubted affinities with the Siluroids and especially that Coccosteus 

 was a near ally of Clarias. I propose to consider how far this con- 

 clusion is borne out by the distribution of the lateral line system. 



I have ventured to reproduce a figure from Pander (Fig. 7) of 

 Coccosteus with the mucous canals marked in red. In Coccosteus the 

 infraorbital canal runs in the suborbital bone in the posterior part of 

 which a groove passes from the main stem (5 in Fig. 7) backwards 

 and downwards towards where one must consider the articulation of 

 the lower jaw to have been placed. This stem appears to 

 correspond with branch no. 5 in Clarias. The infraorbital 

 trunk then passes upwards into the postorbital bone in which it gives 

 off a complete trunk (6 in Fig. 7) which proceeds backwards in the 

 postorbital crossing into the bone marked y by Huxley, his supra- 

 temporal in Clarias, which I take to be the preopercular notwith- 

 standing its occurrence as a true cranial bone. This groove cor- 

 responds to the branch 6 of Clarias. 



In the preopercular it joins another groove representing in Cocc- 

 osteus the opercular canal of other forms. One part of this groove 

 passes upwards and backwards exactly as in Clarias into the squa- 

 mosal bone, near the posterior and upper corner of which the groove 

 curves backwards and downwards and passing into the suprascapular 

 probably continued as the body line. The possibility is however open 

 that the groove in the suprascapular may represent the branch to 

 pore 7 in Clarias ' ). At the posterior angle of the squamosal a short 

 groove runs upwards and inwards. 



Returning to the infraorbital canal it may be seen that from trunk 

 G in the postorbital the groove runs upwards into what one must con- 



1) Cf. a recent paper by Traquair in: Ann. and Mag. Nat. Hist, 

 (fi), 5, 1890. A "supratemporal" commissure is present in Coccosteus 

 1 havo observed traces of a corresponding one in Synodontis [Sept. !)2.] 



