GEOLOGICAL SURVEY IN SILURIAN ROCKS OF SOUTH OF SCOTLAND. 859 



branchial openings have as yet been seen or described in that genus. But if we consider 

 the row of eight openings, apparently branchial, which in Birkenia are arranged in an 

 oblique line on the side, at the junction of head and body (see text figure 3), we are 

 struck by the very interesting fact that Rohon has already figured two rows, right and 

 left, of similar openings on the ventral surface of the Osteostracan genus Tremataspis 

 (xxvii. p. 70 and xxix. p. 8), though the number of openings here is only six on each 

 side. Now, in Tremataspis the head is broad and depressed, while in Birkenia the 

 creature, as it is always found lying on its side, must have been more laterally com- 

 pressed, and consequently the position of the branchial openings must have been in 

 the two genera relatively as in the skate and in the shark. 



For these reasons I have often felt inclined to refer the Birkeniiclse to the 

 Osteostraci, but there are two serious difficulties in the way of this idea. The first 

 is the absence of the cranial buckler with orbits on the top, which is so prominent a 

 feature in all known Osteostraci. The second difficulty is the utter absence of any 

 microscopic proof, as the apparent substance of the scutes of Birkenia, as preserved 

 in the schists of such localities as Birkenhead Burn, shows in sections under the 

 microscope absolutely no structure whatever, except a very faint fibrillation or 

 striation. 



But whatever may be the position which increased knowledge may afterwards assign 

 to Birkenia and Lasanius, for the present they are best placed in an Order by them- 

 selves, for which, I repeat, no place can be found in the system unless we admit it to 

 the " Ostracodermi." 



Conclusion. 



The fossil fishes from the Silurian rocks of the South of Scotland, described in the 

 preceding pages, constitute eight species, which are all new to science. They may be 

 arranged in five genera, four of which are also new ; the remaining one, Thelodus,. 

 having been named by Agassiz in 1831 from detached scales occurring in the Ludlow 

 Bone Bed. The following classification of the species has been adopted : — 



Sub-class— OSTRACODERMI. 

 Order — Heterostraci. 



Family — Calolepidw. 



1. Tlielodus Scoticus, Traq., — Ludlow and Downtonian horizons. 



2. ,, planus, Traq., — Ludlow. 



3. Lanarkia horrida, Traq., — Downtonian. 



4. „ spinosa, Traq., — Downtonian. 



5. ,, spinulosa, Traq., — Downtonian. 



VOL. XXXIX. PART III. (NO. 32). 6 Q 



