Giddy up! Live horses return at last to the Roman Circus. And what’s more – they were small!

Friday 25th of October was a day I’d been looking forward to for some time. In the event, it was an even better day than I had imagined it would be. At long last, there were horses at the circus centre – two in fact, a Caspian horse called Star and a Dartmoor pony by the name of Guinevere. Just to see the horses running round the starting gates was a real joy. This must have been the first time horses have done this for about 1,700 years. The animals didn’t come to race though. They came to take part in two experiments, both aimed at working out the size of the horses that raced in the circus all those years ago.

The special guests arrive. Left to right: Francesca Tompkins (who drove the horse transporter), Guinevere (the Dartmoor pony), Jane Cadman (owner of the horses), and Star (the Caspian horse).

The size of the horses that raced in Roman circuses is a bit of an issue generally. Watch any of the Ben-Hur films for example or the Time Team special shot at the Roman circus and you will see strapping horses hammering round the racetrack. But look at Roman period images of charioteers as archaeologists do and you will see very small ones. Which is right? Were the horses big or were they small? Or could it even be that the horses really were big but were deliberately downsized by the Roman artists so as not to dominate their imagery?

From the Colchester circus site, we already have evidence pointing to the past presence there of small horses – animals in the range of 10 to 12 hands. In other words, they would have been about 44 inches (1.1 m) high, measured to the tops of their shoulders. This evidence comes from two remarkable finds. One of them is a horse’s pedal bone (the main bone in a hoof). This was lying on the floor of one of the entrances which led to the circus seating and on to the arena. As hoofs go, this is tiny. In fact, it is the sort of size that would match a horse in the 10 to 12 hands range.

The other find is a broken horse bit. In its original condition, this would been about 4.5 inches across, a size suited for a small horse, perhaps in the order of 11 to 12 hands in height.

However, although the hoof bone and the horse bit both independently point to the same size of small horse, we do need more proof. More evidence to the same effect would make for a more convincing case and this what we were seeking on that Friday. 

Experiment 1: what is the largest size of horse that would fit the starting gates?

Circus visitors who are used to working with horses sometimes comment that the starting gates look too narrow to have accommodated four horses in a row (which was the normal arrangement when raced with a chariot). To this remark, we usually reply rather lamely, “Well the horses were small”. So the plan was to test this explanation and see if we could we fit four small horses in a row in one of the gates. If successful, we would capture the moment in a photograph to provide the proof that we needed.

We had hoped to have had at least four horses for the trial but, in the event, there were two (a Caspian horse and a Dartmoor pony). However, all was well and the horses performed brilliantly. Both animals belong to long-established breeds, one British, the other from Iran. Although very small and pony sized, Caspians are thinner and have the conformation of horses which is why they are sometimes referred to as miniature horses. They are thought by some to have been a forerunner of the much larger Arabian horse and they have a reputation for being good at pulling carts and chariots as well as jumping. The ancestry of the Dartmoor pony could be as old since there is some evidence of domesticated ponies living on the Dartmoor as far back as 1500 BC.

So what was the outcome of the first experiment? Well, we have our photograph, we have the evidence, and we know that four horses in the 10-12 hands range plus a chariot could indeed have fitted into any of our starting gates. We also found that, had the horses been a little bigger, the fit would have been tight and the animals would have been at risk of injury at the moment when the team took off.

With just two horses to work with, a bit of photographic trickery was needed to get our picture. Three photographs were involved in this process. The first was of the two horses standing together in the centre of one the starting gates. The second featured one of those horses next to the left-hand side of it and the third showed the other one next to the other side. I suppose the resulting single photograph risks being regarded as a bit suspect but I’m sure that one day it will be verified by another photograph, this showing four horses genuinely standing in the gate all at the same time!

A tight but comfortable fit. Four horses in a line on the site of one of the starting gates. (The bases of the walls are reconstructed in such a way that the remains of the gate are preserved intact underneath.)
The excavated remains of two of the starting gates in 2009 with a charioteer of the green team in his quadriga in one of them and a charioteer of the blue team in his quadriga in the other. (The charioteers and quadrigas were kindly drawn by Peter Froste shortly after the excavation so that they could be digitally added to the photograph.)

Experiment 2: what can Roman imagery tell us about the height of the horses?

There are many images surviving from the Roman world which show charioteers and their horses together. The overall impression from these images is that the horses depicted in them were small compared to their charioteers. Our second experiment involved recreating one of these scenes so that the height of the charioteer could be estimated in relation to the height of a real horse which, in this case, is 12 hands high.

As the starting point, we took a well-known mosaic pavement From Baccano in Italy showing four very similar panels each featuring a charioteer and his horse. Since each charioteer stands next to his horse, height comparisons between man and animal are made easy. For our experiment, we chose the panel with the charioteer from the red team.



To in effect introduce a real horse of known height into the mosaic picture, people of different heights each played the part of the charioteer and stood next to the Caspian horse, reins in hand. Each one then gently raised the animal’s head so that the mouth of the animal was level with top of our charioteer’s shoulder, just as it is in the mosaic.

Testing… Testing… Testing… Left to right: Robin Mathieson – much too tall; Sam Rowley – a little too tall; Jane Cadman – not quite as tall.

It became apparent that the Caspian and our charioteer’s horse are very different in shape. Their overall heights agree reasonably well as do the lower parts, namely the lengths of their legs and the distances of their underbellies above ground level. But their necks and bodies are different, the former being too long and the latter too short. So what do we do now? Well, the composition of the mosaic picture focuses on the close relationship between the shoulder of the charioteer and the head of his horse thus strongly suggesting that, if anything in the picture is an accurate representation of reality, then this is likely to be it and we can use what we see here to link and scale the mosaicist’s image with our photograph.

The result, which is shown below, suggests that, if the horse on the mosaic had been 12 hands high as is the case with the Caspian, then our charioteer would have been 5 foot 4 inches tall (or perhaps a little more) which, on the face of it, seems to be a good result. Had his horse been any taller, then so too would have been our charioteer. In fact, for every extra hand in height, our charioteer would have been about 5.25 inches taller. Thus if the Caspian had been 13 hands high, our charioteer would have been about 5 foot 9 inches tall. But then again, had the Caspian horse’s neck been more upright, more like the neck of the horse on the mosaic, then despite being 12 hands, she would have still been able to reach our charioteer’s shoulder.

A = 64 inches (ie 5 foot 4 inches). D = Caspian’ withers = 48 inches (ie 12 hands).

Of course, you might ask, “Rather than align the Caspian’s mouth with the shoulder of our charioteer, why not simply align the withers of the mosaic charioteer’s horse with that of our Caspian?” The trouble is that, if we do this, then our charioteer would end up being about 6 foot 3 inches which, of course, would have been possible, but in reality highly unlikely!

So what’s the answer?

So this brings us back to the original question, “How big were the circus horses?” Well, we now have various pointers to the answer, namely the small hoof bone, the small horse bit, and the limited width of the starting gates. All of this evidence is consistent with horses in or closely around the 10 to 12 hands mark. The image of the red charioteer and his horse on the mosaic from the villa at Baccano has its problems but, if nothing else, it appears that the horse in it, if not of this order of height, is very unlikely to have been much taller than 13 hands or so.

Of course, the horses would all have varied in height as indeed would have the charioteers who raced them. Small charioteers could have had larger horses than normal and large charioteers the opposite. Nevertheless, all the evidence we have so far points to a likely average height of 10 to 12 hands, refinable at a push to between 10.5 and 11.5 hands.

We are much indebted to Jane Cadman, an Essex-based equine podiatrist, for bringing her two horses (Star and Guinevere) to the circus and gamely taking part in the experiments, to Francesca Tompkins for driving the transporter and helping with the horses on the day, and to Dreda Randall (of the Caspian Horse Society) for her interest and support.
The photograph of the mosaic of the four charioteers from the Villa dei Severi at Baccano is reproduced courtesy of the VRoma Project ( The others, including the one of the blue charioteer panel, were taken by the Colchester Archaeological Trust.