338 Proceedings of the Royal Society of Edinburgh. [Sess. 
zoology and anatomy as Agassiz should have based his classification of 
fishes upon characters so trivial as the mere external aspect of their scales, 
or that he should have distinguished many of the families into which he 
divided the order of the Ganoids by characters equally superficial.” He 
maintained that it could not stand the test of anatomical inquiry. 
The detailed researches which enabled Dr Traquair to give the death- 
blow to this classification are embodied in two classic memoirs, viz. his 
“ Monograph on the Ganoid Fishes of the British Carboniferous Forma- 
tions : Part I., Palseoniscidse,” published by the Palseontographical Society 
in 1877, and his memoir “ On the Structure and Affinities of the Platy- 
somidse,” which appeared in the Transactions of this Society in 1879. Until 
the publication of these researches, the Palseoniscidse and the Platysomidm 
had been compared with the existing North American bony pike, Lepidosteus. 
But Dr Traquair showed that the affinities of Palceoniscus, as indicated 
by the skeleton, point most strongly not to Lepidosteus, to which its 
angular scales and fulcrated fins give it superficial resemblance, but to 
Polyodon. Hence it followed that this sub-family ought to be grouped 
with the Acipenseroidei. In short, they were closely related to the 
modern sturgeons. He further showed that the Platysomidae were merely 
a specialised offshoot from the Palseoniscidse. Traquair founded his con- 
clusions on the fundamental characters of the skeleton, which since that 
time has been recognised as the proper basis of a scientific classification. 
Another series of researches of extreme zoological interest, showing 
Dr Traquair’s powers as an original investigator, is associated with the new 
fish fauna found by the Geological Survey in the Downtonian rocks of the 
south of Scotland. Previous to this discovery the family of the Coelolepidae 
of Pander, represented by the genus Thelodus, was known only by scattered 
scales in strata of like age in England and other parts of the world. From 
their shagreen-like scales the Coelolepidce were considered to be sharks. 
But the beautiful specimens of Thelodus and Lanarkia — two genera of the 
Coelolepidse — in the Survey collection led him to place them with the 
Ostracodermi and in the order Heterostraci, of which the only family 
previously recognised was that of the Pteraspidse. He enlarged the order 
of the Heterostraci, and included in it four families, the Coelolepidse, the 
Psammosteidse, the Drepanasidse, and the Pteraspidse. He pointed out that 
the armour plates of the last three of these families had been formed 
by the fusion of the Coelolepid scales with each other and with hard tissue 
developed in a deeper layer of the skin. 
The geological aspects of Dr Traquair’s researches are well illustrated in 
his paper on “ The Distribution of Fossil Fish Remains in the Carboniferous 
