History of Collar Harnessing in Source Pictures

All pictures are stored both in a small ('thumbnail') and in a full size version in order to save downloading times. Click the thumbnail to view the larger version. Some thumbnails are actually clippings from the larger versions.

The Roman Collar in Shoulder Traction

Roman, about 0 AD CST#1: A Roman horse at a wheat mill around 1 AD. The horse seems to bear some type of collar because a thicknes is depicted around the neck base. The collar fits over the shoulders. Whether the traction force goes to the breast base or the shoulders depends on the position of the invisible traction point. It only makes a difference if the mill needs heavy force. If so, a wider walking circle would lower the force again. It is easier and more efficient for a horse to walk in a wider cirlce than in a narrow one.

Roman Mule Collars, 22/23 AD CST#2: Roman mules in collars on a coin. It is a sesterce of Emperor Tiberius of 22/23 AD. This is, at the mid 1999 state of research, the earliest representation of the typical Roman collar. See CMi#1 to CMi#3 for details of this collar design. Here the iron core seems padded and the fastening straps below the wood pads are visible too. It is very much like CNT#1 but in shoulder traction.

Roman Mule Collars CST#3: Roman mules in collars on an undated coin. The collars are covered by some decoration pieces or a heavy padding.

Roman Mule Collars CST#4: Roman mules in collars on a coin. Undated, somewhat abraded.

Roman Mule Collars, 37/41 AD CST#5: Roman mules in collars on a coin. A sesterce of Emperor Caligula of 37/41 AD.

Collars on Roman mules, c. 96 AD CST#6: Collars on Roman mules pulling a light travelling wagon. The inscription, perhaps Memoriae Domitiane, indicates a time of c. 96 AD. The collar looks two-parted to the front of the throat. These are the fastening straps more clear depicted in CST#8, 9 or CNT#1. The girth strap is also visible over the body. A bump on the top of the horse's back, betwen the collar and the girth, seems to be caused by a damage. It created a dent below too. It is no indication of a yoke.

Roman, before 100 AD CST#7: Mule in collar at a vallus, a Roman "harvesting machine" around 100 AD. Several damaged relief depictions of it were found. Click the picture to see the composition of two such fragments showing the whole front part of a vallus in operation. This device is also mentioned by Roman writers.
Vallus The whole vallus in a reconstruction based on the reliefs from different sites mainly in France. But instead of the Roman collar the artist wrongly placed an MA ring collar over the mule.

Roman Horse Collars, c. 106 AD CST#8: Roman army collar from the column of Emperor Trajan, c. 106 AD. Two pairs of mules hauling torsion gun carts -- a Roman field artillery system. The relief of the lower mule is somewhat demaged. But the above one is the most detailed representation of a Roman collar in position of shoulder traction.
We see an iron collar bow, only slightly padded, that ends in the wooden pad. We know these wood pieces from CMi#3. In the front are the fastening straps we know from CNT#1. Unlike CNT#1 here the two straps seemed to be fixed together by a front band. The rein close to the collar supports the idea that it was guided through the upper part of the collar like shown in CST#9.
Trajan Column Section The context of the above picture. The carts are depicted during march. The guns and carts are visible in much detail.

Roman Horse Collars, c. 106 AD CST#9: Roman army collar from the column of Emperor Trajan, c. 106 AD. This is one of two mules hauling a gun cart again. But here the collar is on a higher neck position unlike the two other depictions above. The many details make an error on the artist's side less likely. A soldier at the right seems to be moving the wheel of the cart. This at first suggests some assistence of the mules offroad in difficult terrain. But the base of the wheel is well visible - removing the idea of a bad road. Instead the moment we see is probably the unharnessing of the mules. The fasten straps (in front under the neck) connected the wooden frontpads at the lower sides of the collar bow. The release of this straps allowed the rotation of the wood pads and this the release from the harness. If it was a fail safe design, the straps could not be released as long as the system was under tension on the shoulders at the neck base. To release the tension and then the strap the soldier pushes the cart pole forward by turning the wheel. So the collar at this depicted moment is not in hauling position but up the neck.
Trajan Column Section This interpretation is supported by the context of the above picture. This is the whole scene. The artillery cart arrives in the lower left at a city wall. Here Roman soldiers already stationed some torsion guns at the wall of the city. Another such device is in a well protected position in front of the wall. The newly arrived unit is just unharnessed to the left of the other at the site where the position building is still in work.

Roman Collar, c. 274 AD, Roman Car Pole, c. 106 AD CST#10: Roman collar in a relief by Aurelian, 274 AD. Although a somewhat abraded relief, the depicted collar seems of different design then that in CST#8 170 years before. And it is a rare depiction of the collar's backside. The reins are fed through the top part of the collar. The pole of the car is visible in front of the wheels, too. It is clear that the pole is attached to the side of the collar like in CMi#1 and not to the top. The collar is positioned at the upper shoulder part at the base the neck.
roman collar, c. 274 ad Same relief section from another view. To the left is a mule with the same collar harness but in a condition less well preserved.

Roman Collar on a Mule, Ostia 350 AD CST#11: A Roman collar applied to a single mule by a double-poled, 4-wheeled wagon. Black and white street mosaic at the Roman seaport of Ostia, c. 350 AD. That image is a scan out of Needham (1965). A detailed video sequence of the same mosaic, done in the 1980s, is also available to me. By comparing the two, the black streak atop the collar is clearly a photographic artefact by Needham or his source. There is nothing like it in the real mosaic. But the depicted back top of the collar is just one line of mosaic stones, as confirmed by the video.
Needham, Science, Vol. 4, Pt. II (1965) By the girth strap Needham assumed this picture was a "throat-and-girth" harness type. But the girth strap is today considered a necessity for all kinds of collar harness too. It allows the horse to break the wagon and keep it distant from its rear. It has the same function as in the "throat-and-girth" harness. Only the back top line identifies this harness as collar type.

(Unlike in Needham's sketch here, in his several sketches of collar harness types (p. 320, 328), Needham always neglected the girth strap. Nevertheless the girth strap is visible in his photographs of today's collars.)

Ostia Relief A clearer picture of the Ostia relief taken from Brancards et Transport (1993). It's only 30 % the size of Needham's reproduction but shows more details about the mosaic structure. The structure at the back of the mule is clearly no artefact but a depiction of the top of a Roman collar.

HTML by Erik Möller
Content by Seneca