The Horse Collar and the Harness That Never Was

Read the episode (Narration script)

Welcome to Origins. In this program we explore the technological and scientific roots of the modern world. Let's begin.

In 1910, a retired French cavalry officer named Richard Lefebvre des Noëttes hitched two horses to a light two-wheeled cart. The harness was his own reconstruction of an ancient one, pieced together from old carvings and paintings. A soft band ran around each horse's neck and tied up to a yoke on its back. When the horses leaned into the load, the band slid against the windpipe.

By his measurements, two horses in this rig could haul only about half a ton. A single horse in a modern padded collar could manage roughly three times that.

From this test, the commandant drew a conclusion about the whole ancient world. If Greek and Roman horses were half strangled whenever they pulled, then the heavy hauling of antiquity must have fallen to human beings. Weak harness, he argued, helped explain slavery. And the thing that rescued Europe was a medieval invention: the horse collar.

The collar was real, and it mattered. The harness in that cart was the problem.

Start with the animal. For thousands of years, the standard way to make an animal pull was the yoke, a wooden beam laid across a pair of oxen. Oxen suit it. Their bodies take the push at the base of the neck and the shoulders. A horse is built differently. Its windpipe and the big veins of its neck run close to the surface, just where a strap would want to ride up. Yet a horse moves faster than an ox, and can deliver roughly half again as much work per second. The prize was obvious. The puzzle was where to put the force.

Different places answered that puzzle in different ways, and China answered it early. A lacquered box from the state of Chu, painted in the fourth century before the common era, shows a bar resting across a horse's chest rather than its throat, with straps running back to the cart. By the Han dynasty, tomb reliefs and stamped bricks show breast straps, a band across the chest that lets the horse push with its body. Han carriages often ran with a single horse between two shafts.

Then came the collar itself: a stiff, padded ring that sits on the shoulders, with the pulling lines fixed low on either side. Historians place it in China around the fifth century. A mural in the Buddhist caves at Dunhuang, painted in the late four hundreds, appears to show one. A later Dunhuang wall painting, celebrating the general Zhang Yichao's victory in the mid eight hundreds, shows a collar plainly. Somewhere along the long trade routes, the idea travelled west. By around the tenth and eleventh centuries, collared horses were spreading across European fields.

That journey would make a fine story by itself. What made the collar famous was the commandant.

Lefebvre des Noëttes was an experimenter by temperament. As a supply captain in 1896, he had tried using the covers of army wagons as ferry boats during manoeuvres, and he started researching how horses were shod. An accident ended his military career. He then studied for two years at the École des Chartes, France's school for archivists, and turned his attention to ancient pictures of horses at work.

His book on animal power appeared in 1924. A larger edition followed in 1931, now subtitled "a contribution to the history of slavery," with a preface by Jérôme Carcopino, one of France's leading historians of Rome. The idea travelled fast. In 1933, Leo Löwenthal, writing in the Frankfurt School's journal of social research, used it to argue that slavery ended out of economic necessity rather than moral awakening. In 1962, the American historian Lynn White wrote in his widely read book on medieval technology that a team of horses could pull only about a thousand pounds with the old yoke harness. The great historian of Chinese science, Joseph Needham, built on the same work. For decades, the strangled Roman horse sat in textbooks as settled fact.

Some people who knew horses found it hard to believe. Romans raced chariots for miles at a full gallop. Somehow the horses kept breathing.

The decisive test came from a horseman. Jean Spruytte was a former French army officer and riding teacher who built working replicas of ancient chariots and harness, and drove them. In 1977 he published experiments on the harness systems of Egypt, Greece and Rome, worked out with the American scholar of ancient horses Mary Littauer, who later translated the book into English.

His finding was blunt. Lefebvre des Noëttes had fused two different harnesses into a third that nobody had ever used. In the main Greek and Roman system, a forked yoke sat just behind the horse's withers, the high point of the back, and a strap across the chest carried the pull. Cut to fit, that strap stayed low and clear of the throat. Spruytte's pair, in a correctly modelled harness, comfortably drew about 975 kilograms, cart and load included. That is close to a ton, about double the commandant's figure.

Since then, the picture of ancient haulage has kept shifting. The Belgian archaeologist Georges Raepsaet studied Gallo-Roman harness and farm vehicles in detail, and in 1991 the French historian Marie-Claire Amouretti titled an article on the affair "the prestige of a scientific error." In 2017, Gail Brownrigg and Joost Crouwel argued that curved wooden plaques held by a metal bow, seen on Roman reliefs and on Trajan's Column, may be an early ancestor of the collar's hames, the rigid frame that later carried the traces.

What, then, did the collar really change? On the road, perhaps less than its reputation suggests. Some historians who compare the evidence find medieval carters hauling around a ton, much like Roman teams. Its larger gift was in the field. With the pull fixed low on the shoulders, a horse could drag a plough through heavy ground, work that in Roman times mostly fell to oxen. In parts of Europe, horses slowly began to share the plough with oxen, and in many places, oxen held on for centuries.

The commandant was right to put a harness on a horse and see what happened. Asking the animal is a good instinct. But an experiment can only test the object in front of it. His horses were real, his measurements were honest, and the harness came from his own reading of the pictures. Humans keep remaking the tools that tie us to other animals, and then telling stories about what those tools did. Sometimes the story needs to be hitched up and tested too.

Thanks for listening to Origins.