14 Proceedings of lioyal Society of Edinburgh. [sess. 
not be difficult to account for tlie remarkable difference in the 
characters between the Tarpan of Gmelin and Pallas and the 
Tarpan of the Moscow Zoological Garden. It is hardly necessary 
to point out that were we to find in one and the same herd adult 
horses with an erect or semi-erect mane and a short bushy tail, 
and others with a long flowing mane and a full tail, long 
enough to reach the ground, we should hesitate, even if they 
happened to be of the same colour, to regard them as intimately 
related ; and if in addition some of them retained, while others 
had completely lost, the hind chestnuts, we should unhesitatingly 
look upon them as belonging to two different varieties, if not 
different species. Such differences, coupled with a want of agree- 
ment in the number of the lumbar vertebrae or in the ribs, would 
make in favour of adopting the view that the herd in question 
consisted of the feral descendants of domesticated horses, or had 
resulted from the intercrossing of a true wild horse with members 
of one or more domesticated varieties. 
Seeing that herds of mouse-dun wild horses no longer occur in 
Europe, and have not during recent years been met with in even 
the most remote parts of Central Asia, it might perhaps be as- 
sumed that the Tarpan’s place in Nature must for ever remain a 
mystery. 
This was the conclusion I arrived at when my attention was 
first directed to the subject. But having ascertained that, by 
crossing carefully selected forms, remote types are sometimes 
restored in all their original purity, I thought it worth while to 
make some experiments. In the case of pigeons, by mixing the 
blood of two well-marked breeds (such as the owl and archangel 
breeds of fanciers) and crossing the mongrels with a white fantail, 
I at once obtained birds which closely resembled Columba livia , 
the recognised wild ancestor of all the tame pigeons. 
Bearing this and like experiments with zebras, dogs, rabbits, etc. in 
mind, I selected for my Tarpan experiments a mouse-dun Shetland 
pony mare, which seemed to me to be a blend of at least three 
varieties — in its head (PI. I. 1) it suggests the wild horse; 1 in 
1 Tlie wild horse ( E . prejvalsTcii ), like Grevy’s zebra (E. grevyi), has a very 
long head, the distance between the eye (inner canthus) and the nostril being 
decidedly longer than in a forest horse eight inches higher at the withers, and 
