The Vallus, a 2000-year-old Gallo-Roman Harvesting Machine

Bas-relief of the vallus from Buzenol-Montauban, now in the Musée Gaumais in Virton, Belgium

Malagne, the Archeoparc of Rochefort, in Belgium (between Brussels and Luxemburg city), is an interpretation centre for rural Gallo-Roman civilization especially designed to highlight an important villa active in the early centuries of the Christian era in Northern Gaul. Malagne’s special mission is heritage protection, scientific research and mediation of this listed site and archaeological experimentation plays a central role in our activities. Hence, the years-long intensive work on reconstituting and effectively using the vallus, the Gallo-Roman harvester, in situ was carried out in partnership with Prof. Georges Rapsaet of the Free University of Brussels (Université Libre de Bruxelles) [see references].

Study of written and iconographic sources was the first step in our inquiries. Pliny the Elder in his Natural History wrote pithily in the first century BCE that “in the great land-holdings of Gaul, large valli mounted on two wheels with an edge equipped with teeth are pushed across the fields, harnessed to a draft animal in such a way that the ears fall into the vallus.”

Reconstruction of single yoke used at Malagne from the Pforzheim well discovery

In the fourth century, the agronomist Palladius provided a more detailed description in his Treatise on Agriculture (Opus agriculturae).

Sculpted funerary bas-reliefs from Buzenol-Montauban, Arlon, Trier, Koblenz and the agricultural calendar on the Porte de Mars in Reims all represent fragmentary images of the harvester, thus completing written documents and provide us with a general idea of what the vallus must have looked like. The harvester is made up of a body with a comb in front in which cereal grain stems are caught and the ears broken off to fall into the box-like gathering case of the implement, which is wheeled and pushed by an animal harnessed to the shafts – just as we would push a wheelbarrow. It took two people to work the implement, a driver called the bubulcus working from behind to steer the harvester and a conpulsor in front to push the ears of grain – which might build up and choke the teeth – into the collecting box.

Malagne summer 2015, first trial run harnessing Capucine to the vallus

As early as 1960, several archaeological experiments had already been carried out in attempts to understand how this 2000-year-old machine worked. Each trial concentrated on one or another of the problems involved – for the sort of cereal grain crop, for the choice of draft animal, the proportions of the harvester or the (misguided) use of a horse collar that only appeared in the medieval period. After all this trial and error, many lessons were learned, and the more holistic issue of draft and especially the problem of understanding the harnessing system were taken into consideration.

Malagne, Autumn Festival, 27 September 2015: Capucine’s first public demonstration with the vallus

This was the main thrust of the research carried out in Malagne by Prof. Raepsaet, and the team was assisted by the discovery of a second-century single yoke in a well in Pforzheim in Germany that was used as an experimental model. It is made of a wooden transversal piece in the middle of an arched device that fits to the shape of the animal’s withers, plus two independent discs sitting before the shoulder blades. Several cinches held the single yoke and discs in place. This device was attached with ropes to the double shafts and the harvester.

We decided to choose a European donkey as draft animal, and our Marius played his role with flair in pushing the harvester. He was led into the shafts, head towards the box, the yoke on his neck and then attached to the shafts with ropes. Our hands-on research went on for over two years in order to work out the best results. The role of the bubulcus or oxdriver guiding the animal and of the compulsor, became clearer. The vallus produces the best yields when applied to easy-breaking ears such as spelt, emmer or einkorn wheat.

Malagne, summer 2015, our first experiments with Capucine harvesting spelt

These long years of experimentation have been a success and shown that in a few hours the harvester could reap a crop that would have required a full day with sickles. Please note that the implement presented here is nonetheless but one of the solutions possible. Our young lady donkey, Capucine, has taken over from Marius now, so our research is ongoing…. Of course, Malagne provides the public with opportunities to see Capucine and the vallus at work, so please visit us at

Françoise Fontaine, Director, Malagne, Archeoparc de Rochefort, Belgium.

For further information, please see the following references:

Raepsaet, Georges. Attelages antiques, jougs et jouguets. Approches ethno-technologiques. Etudes d’archéologie 9, CReA-Patrimoine) de l’Université Libre de Bruxelles, 2016, 189 pp.

Raepsaet, Georges. La moissoneuse gallo-romaine au fil de l’histoire. Une icône, révélateur épistémologique au cœur de la technologie romaine. Etudes d’archéologie 19. Brussels : CReA-Patrimoine, 2022, 194 pp.

Nys, R., Bonato, S., Limbrée, C. Le vallus: moissonneuse gallo-romaine, Malagne la Gallo-Romaine, Rochefort, 2010, 34 pp. (Collection Vi@Malagne #1 :