jQ KINSHIPS OF OS TEA CODE RMS 



no instance has a trace of endoskeletal parts been ob- 

 served. 



The more that is determined of the structural characters 

 of Ostracoderms, the less is it possible to accept the 

 views as to their affinities with forms other than "fishes," 

 either (Cope) as to their permanent larval-ascidian char- 

 acters, or (Patten) as to their relationships with arachnids. 

 Their general kinship is certainly to the fishes. Accord- 

 ing to Smith Woodward, the markings appearing on the 

 visceral surface of head tests indicate the presence of 

 gill pouches ; in some forms clearly marked furrows sug- 

 gest the possession of vertical semicircular canals ; fish-like 

 sense organs occur (Fig. 77)\. and their derm plates, in 

 their cancellated and bone-like characters, cannot well be 

 likened to the exoskeletal parts of invertebrates. 



The lamprey-like form, Palceospondylus gimni, Traquair 

 (Fig. 73), in the Lower Devonian is by many looked upon 

 as the actual solution of the Cyclostome, and even of the 

 Ostracoderm puzzle. This interesting fossil was discov- 

 ered by Dr. Marcus Gunn, in the Lower Old Red Sand- 

 stone of Caithness, and was described in several papers by 

 Traquair {Trans. Edhi. Soc, 1 892-1 894). It is of very 

 small size, commonly of about an inch in length, but is 

 admirably preserved (Fig. 73). There can be no doubt 

 that Palaeospondylus possessed a ring-like mouth sur- 

 rounded by barbels like those of a Myxinoid, and that it 

 lacked paired fins. But as a Cyclostome it must have 

 highly specialized, having the same relation to the more 

 primitive Cyclostomes of its day, as had the minute Acan- 

 thodians (p. 81) to the existing sharks. It had thus a 

 remarkably large caudal fin with elaborately bifurcating 

 supports ; it had evolved stout, ring-like vertebrae, even in 

 the caudal region, which had developed stout neural proc- 



