They drew, sculpted and painted rhinos, mammoths, giant deer and lions, but they also produced illustrations of less exotic beasts, like owls, mustelids and rabbits. Comparisons made with living animals and fossils reveal that these depictions are, on the whole, biologically accurate and often result from informed observation. The greatest concentration of cave art occurs in southern France and northern Spain where horses and bison are the most frequently depicted animals. And today, I’d like to talk about those ancient depictions of horses. They’re fascinating, and seemingly reveal a weird and unexpected amount of variation relative to what we might assume.


The horses concerned are members of the so-called caballoid or caballid or caballine group – the complex of populations that are known variously as Equus caballus (domestic horses and their feral derivatives), E. ferus (tarpans or forest horses) and E. przewalksii (Przewalski’s horses or takhis). Ancient populations of this complex have been given a few other names, including E. lambei and E. lunensis. I will definitely be avoiding any discussion here of caballoid phylogeny or taxonomy since it’s quite messy. The emphasis on horses in Palaeolithic art most likely reflects both their familiarity and their importance in hunting, though their appearance at some places might represent their use as fertility symbols or religious icons. It should be noted that a few Palaeolithic horse images have been identified as depictions of asses or other species but I won’t be discussing those here.

Permafrost horses and what they tell us

Horses depicted in cave art are generally stocky, mostly tan or yellowish with a white belly, and usually shown with a stiff, dark mane*. They thus resemble Przewalski’s horse of modern Mongolia. Corroboratory evidence that Ice Age horses of some populations looked like this comes not only from living wild caballoids but also from the Selerikan horse (or Selerikan pony), a Pleistocene stallion preserved in Siberian permafrost, discovered in 1968, and extensively described in works largely unknown in the west (Guthrie 1990, Ukraintseva 2013).

* It should be noted that not all ancient horses were like this - we have evidence that some Pleistocene horses in North America (and maybe elsewhere) had long, flowing manes. More on that another time. It doesn't affect the mostly European horses discussed here.

The Selerikan horse was initially discovered, entirely by chance, thanks to gold-mining operations: a tunnel happened to coincide with the area of its burial, and its hindlegs protruded from the ceiling. These were used for holding cables and hanging lanterns but were eventually deemed inconvenient so were removed (via the use of explosive powder) and thrown away. Officials at the Siberian Academy of Sciences eventually found out about this discovery and were able to retrieve both the discarded legs and the rest of the body, though the head was not preserved and was presumably removed by predators while the horse was mired (Guthrie 1990). Anyway, its preserved coat shows that it was tan or coffee-coloured in life with a coal black mane and tail and a dark dorsal stripe. Long hairs are present right at the base of its tail, as is typical for caballoids but not for other horse groups.

The Selerikan horse is not unique: several other Siberian permafrost horses are known. Because they tend to be discovered by people busy exploiting the ground for profit, they typically do not get retained for study. What was described as a ‘white horse’ carcass was discovered in the bank of the Yana River in 1878, and a horse containing a full-term embryo was found at Sana Creek in 1950 (some of this specimen, but only some, was saved).

Permafrost caballoid remains have also been recovered from Alaska. The Titaluk River specimen is significant in preserving a good hoof, which is of the broad sort typical for caballoids (other horse groups have narrower hooves). The hoof is especially interesting in having a heavily worn leading edge (suggestive of use in regular digging), but flaring, overgrown lateral edges indicating both a rapid rate of hoof growth and a relatively sedentary lifestyle that did not involve copious walking on abrasive terrain (Guthrie & Stoker 1990). The amount of overgrowth is in fact so marked that it might be assumed to be abnormal were it not for the otherwise normal rate of wear elsewhere, and the normal length of the hoof overall. Guthrie & Stoker (1990) suggested that it is a ‘winter hoof’ and that those lateral areas of excess growth would have broken away during the spring – similar patterns are seen in domestic horses with untrimmed hooves. If so, it supports the idea that horses were able to remain sedentary on the Alaskan North Slope during the Pleistocene, a discovery which is consistent with the idea that the land surface, vegetation and pattern of wind and snowfall here was different from that present during the Holocene.

It isn't yet known how widespread hooves of this sort were among Pleistocene horses. I haven't seen them mentioned in discussions of the Selerikan horse.

Dark faces, beards, stripes and more

Several forest- and plain-dwelling horses related to Przewalski’s horse roamed prehistoric Europe, and differences depicted among those horses shown in cave art probably represent these varieties. Historically, experts on Palaeolithic art put quite some effort into recognising different forms or ‘subspecies’ among these images, the end result being proposals that up to 37 distinct horse forms can be recognised. However, it became obvious as these schemes became ever more elaborate that the vast number of variations seen in pigmentation, proportions, mane form, head shape and so on more reflect the whims, style and abilities of artists more than actual variation in the horses themselves (Bahn & Vertut 1999). But read on.


Some paintings illustrate the horses in their winter coat, with long hair around the jawline and hoofs (Guthrie 2005). A very prominent, dark shaggy beard is shown in a horse at Zubialde Cave in Spain. However, some of the images at Zubialde are definitely recent fakes and there is consequently doubt about various of the remaining images. It has been suggested that these ‘winter’ forms might be preferentially depicted in those places where the art was produced during cold phases (example: the Cave of Niaux, in south-west France), with ‘warm phase’ horses being prevalent (and ‘winter’ ones being absent) in places where the art was made during warm phases (Bahn & Vertut 1999).

The neck regions and faces of horses are sometimes shown as being especially dark relative to the rest of the animal. This is the case in the famous Horse Panel at Chauvet and in some of the Spanish horses depicted at Ekain. Several horses at these locations also reveal a prominent ‘M’-shaped pattern on the side of the body. This seems to represent the line of demarcation between the pale belly and the tan or reddish upper body, the white of the belly extending onto the side of the chest and flank in inverted U-shaped patches.

Sometimes, stripes across the shoulders and/or neck base are shown. This is the case in several of the Lascaux horses – the so-called Chinese horses – where the stripes are thick, fully black and far more prominent that the transverse stripes seen on any modern wild caballoids. These prominent stripes are also present in some (maybe all: it’s hard to say) of the Ekain horses, meaning that we have transverse shoulder stripes depicted together with the prominent lateral ‘M’ [Ekain horse panel shown above: photo by Xabier Eskisabel]. Stripes are depicted on enough occasions that they seem to be a genuine anatomical feature of some European horse populations, though their size and intensity might have been exaggerated by the artists. And it gets better, since some of the Ekain horses also reveal fine, slender stripes on the legs as well as traces of a dorsal midline stripe. These animals must have been quite elaborate – similar overall to Przewalski’s horses, but more elaborately marked. They have that lateral 'M', and shoulder stripes, and leg stripes, and a dorsal, midline stripe and a stiff, erect, dark mane.

Leopard-spotted Pleistocene horses?

And on that note about elaborate appearances... a long-standing mystery is the presence at the Cave of Pech-Merle de Cabrerets, France, of a group of heavily spotted horses that have dark heads and necks. These are the so-called ‘Dappled horses of Pech-Merle’. Such a colour scheme is unknown in wild living horses so it’s often been assumed that the spotting is some sort of artistic decision. Further support for this view comes from the fact that other animal illustrations in the same cave are also shown with a spotted coat. However, Pruvost et al. (2011) found genetic data indicating that some Palaeolithic horses were ‘leopard spotted’ just like some modern domestic breeds, most famously the Appaloosa. They therefore argued that those ancient dappled horse images were accurate after all. The idea that some Pleistocene horses really were so boldly patterned is pretty incredible, but then so is the presence of striping on zebras.

I like the idea of boldly patterned Pleistocene horses and don’t regard it as especially outlandish. However, I admit to remaining somewhat sceptical. As just mentioned, we see evidence that Palaeolithic artists were sometimes putting random spotting on all sorts of things (even on otherwise blank bits of cave wall) and were not necessarily intending those spots to represent zoological accuracy. Furthermore, even if the genes associated with a spotted coat have been discovered in fossil horse DNA, it doesn’t necessarily follow that there’s a direct link between those genes and what’s seen in some cave art.

The majority of Palaeolithic art depicts caballoid horses that looked much like – if not exactly like – the wild horses that survive today on the Asian steppes. And, thanks to the Selerikan horse, we have direct evidence that some Pleistocene horses did indeed look like this. But good, consistent cave art indicates that horses of some populations (subspecies or species?) were more elaborate in appearance, with stripes here and here, and tidily demarcated flank and belly colouring. Were some Pleistocene horses even more elaborate, with spotted coats and heads and necks far darker than those of their bodies? These are exciting concepts that make for a more vivid view of the Pleistocene world, but are they accurate? Maybe, maybe not.

For previous Tet Zoo articles on horses and Pleistocene megafauna, see..

Refs - -

Bahn, P. G. & Vertut, J. 1997. Journey Through the Ice Age. Weidenfeld & Nicolson, London.

Guthrie, R. D. 1990. Frozen Fauna of the Mammoth Steppe: the Story of Blue Babe. The University of Chicago Press, Chicago and London.

- . 2005. The Nature of Paleolithic Art. The University of Chicago Press, Chicago and London.

- . & Stoker, S. 1990. Paleoecological significance of mummified remains of Pleistocene horses from the North Slope of the Brooks Range, Alaska. Arctic 43, 267-274.

Pruvost, M., Bellone, R., Benecke, N., Sandoval-Castellanos, E., Cieslak, M., Kuznetsova, T., Morales-Muñiz, A., O’Connor, T., Reissmann, M., Hofreiter, M. & Ludwig, A. 2011. Genotypes of predomestic horses match phenotypes painted in Paleolithic works of cave art. Proceedings of the National Academy of Sciences 108, 18627-18630.

Ukraintseva, V. V. 2013. The Selerikan horse. In Ukraintseva, V. V. (ed) Mammoths and the Environment. Cambridge, Cambridge University Press, pp. 87-105.