154 
JOURNAL OF MAMMALOGY 
are from an area which has remained free from ice during the whole 
Pleistocene. This gradual evolution may be observed in another 
important character; in the older forms the protocone of the cheekteeth 
is somewhat detached from the main body of the tooth and becomes 
more adpressed in more recent ones ; the anterior lobe of the protocone 
is on the whole rather long and has been described as very characteristic 
of E. c. przewalskii by Lydekker;^^ it appears to be constantly so in the 
living and the more recent fossil forms, but in teeth which have been 
described from the lower Pleistocene of central Europe it is sometimes 
much smaller than the posterior lobe (from which it is generally clearly 
separated), though this is far from being constant, even in teeth from 
the same locality. The oldest tooth known (Bawdsey) is almost in- 
distinguishable in this particular character from the most recent types. 
It is this type, with a rather heavy head and six lumbar vertebrae, 
that has transmitted its principal characters to modern heavy races 
such as the “Pinzgauer” and ‘‘Belgian” draught horses. Their far 
superior size compared to that of the wild “tarpan” is most certainly 
the product of conscious breeding for certain purposes which has 
taken place in historical times and even within a rather short period. 
It has frequently been supposed that the most heavy types are derived 
from the middle sized horse of the ‘‘robustiis” type, but this is impossible 
as that form had long ceased to exist at the time when horses were 
first domesticated. On the contrary all the available evidence afforded 
by fossil and prehistoric remains points to the fact that the most an- 
cient domestic horses were of a type closely resembling the wild tarpan 
of Mongolia in their rather light build and somewhat clumsy head. 
Professor Ewart believed that the “Celtic” pony is of another type 
than the tarpan. It has been already shown by Lydekker^^ that the 
skull characters adduced by him are not very reliable, and it has also 
been shown above that slender limbs are found in E. stenonis as well 
as in the more recent forms of the caballus type. It would further 
appear probable that the presence of only five lumbar vertebras in 
some of the skeletons examined by Professor Ewart is due to infusion 
of southern blood I have ascertained^^ that the Lofoten pony pre- 
served in the Bergen Museum and regarded as of pure Celtic origin 
2® Guide to the specimens of the Horse family, p. 14, fig. 16A (1907). 
21 “Guide,” p. 15. 
22 Ridgeway, The Origin and Influence of the Thoroughbred Horse, pp. 419-420 
(1905); and Lydekker, The Horse and its relatives, p. 121 (1912). 
22 With the kind help of Professor Brinkmann. 
