Tire choices in Roman chariot racing

Bela I. Sandor
Formal chariot racing was a sophisticated and popular sport for over 1800 years, from
Etruria in the 6th c. B.C. down to the fall of Constantinople, and the races held in a large
number of circuses and hippodromes imply that huge numbers of racing chariots were
made over the course of those centuries. It may therefore be thought surprising that no
racing machine has been found, but the dearth of such hardware is plausible given the
perishable wood and leather components of the lightweight vehicles and the desirability
of recycling the metal parts. In this situation a particular artifact must be accorded special
significance. It is a hand-sized bronze model of a Roman racing biga, known as the Tiber
model because found in the river. Dated to the 1st-2nd c. A.D., it is now on display in
the British Museum (GR 1894.10-30.1, Bronze 2694). With this model as our guide, all the
major dimensions of Roman racing chariots have been reasonably well determined; fur-
ther, several technical aspects (some obvious, others quite subtle) of actual racing chariots
can be established from it.1 Among the subtle details of the model, one feature is espe-
cially intriguing in view of the remarkably realistic work of its maker, who was clearly
knowledgeable in matters large and small of vehicular racing. As first mentioned to me by
J. Swaddling of the British Museum during our latest study of the model (November 2014),
an unusual tire configuration is apparently represented on this all-bronze model: the right
wheel has a slightly raised rim, as if to indicate a thin iron tire, but the left wheel lacks this
feature (fig. 1). This asymmetrical arrangement is not only curious, it also implies extra
work and expense. What, then, could be the reason for it?
One possibility for having apparently only one tire on the Tiber model — which was
probably a toy for a rich individual; the emperor Nero, an avid racer, was said to play with
toy chariots2 — is that two different castings were used in the toy’s production, and either
one could have been used in the assembly, in random selection from a box.3 However, two

Fig. 1a. Bronze model of a Roman racing chariot in the Fig. 1b. Right wheel of the model in fig. 1a,
British Museum (author; copyright Trustees of the British evidently representing a thin iron tire (author;
Museum). copyright Trustees of the British Museum).

1 B. I. Sandor, “The genesis and performance characteristics of Roman chariots,” JRA 25 (2012)
475-85.
2 Suet., Nero 22.1.
3 Robert Hurford, pers. comm. (August 2015); see his www.chariotmaker.com
 Tire choices in Roman chariot racing 439

identical wheels would have been easier to produce for toys and real chariots alike. Fur-
thermore, a high-ranking customer would probably demand for his sophisticated toy an
authentic representation of reality, if indeed that was one tire only.
Physics-based analysis, and analogies to modern automobile races held on oval tracks,
combine to suggest an answer, one that makes immediate sense to engineers with experi-
ence in the mechanical sciences.4 This explanation can be introduced to non-engineers by
reviewing the quintessential requirements for the optimization of racing wheels. A great
improvement in wheel design and construction became possible with the advent of iron
tires and the invention of heat-shrinked hoops on wooden fellies. The hoops provided for
wheel consolidation, a matter of constant concern for wheels using wooden components.
Iron tires also had good resistance to wear, and they stiffened the wooden rims, reducing
the harsh washboard effect that can be a problem even on smooth roads.5 Iron tires are still
in use even in advanced countries, mainly in farm vehicles. The great technical advantages
of iron tires become blurred, however, when one considers racing chariots. The problem is
twofold: mass (weight) and mass distribution are difficult to optimize, and in every con-
figuration some undesirable side effects are encountered. The basic problem is that iron
is more than 10 times denser than wood or leather (the materials of ancient tires). Even
thin iron hoops adversely affect the acceleration and sustained speed with any available
horsepower. For a given diameter of a wheel any extra weight is detrimental, by increasing
the wheels’ impression, and thus the frictional resistance to rolling, on the track. In other
words, a large wheel diameter is advantageous for running on a soft track (with less sink-
ing), but the extra mass results in lower linear (Force = mass x acceleration) and rotational
(non-linear) accelerations, which occur at the start of the race and after every turn. To
judge by their fine efforts in vehicle optimization when compared to any other chariot in
any other ancient civilization, Roman engineers definitely understood these fundamental
requirements for reducing weight and rotational inertias.
The numerous compromises involved in designing a good wheel imply that perfec-
tion cannot ever be reached. For racing purposes, iron tires require particularly complex
decision-making. For a sophisticated racing chariot, a range of tire choices worth consider-
ing would be:
(a) no tire of any kind on the basic rim;
(b) wood or leather tires on each wheel, but no iron tires;
(c) iron tire on each wheel;
(d) one iron tire: (d1) on the left wheel only, (d2) on the right wheel only.
Choice (a) seems to be the most primitive; it can be seen on the Florence chariot, the earliest
high-performance machine, one of the 8 Tut-class chariots.6 R. Hurford believes, not unrea-
sonably, that this chariot originally had rawhide tires.7 In either case, this would have been
a fine racing chariot, but probably only in informal competitions. These wheels are in fact
among the lightest, with low rotational inertias, but with low durability, thereby requiring
frequent and expensive repairs.

4 If one mentions to engineers the possibility of using only one iron tire on a racing chariot, a
typical response will be: “Of course, for left turns only, it is best to have only one tire, and it
should be on the right wheel.”
5 B. I. Sandor, “The rise and decline of the Tutankhamun-class chariot,” OJA 23 (2004) 153-75.
6 M. A. Littauer and J. H. Crouwel, Chariots and related equipment from the tomb of Tutankhamun
(Oxford 1985).
7 Pers. comm. (Nov. 2014).
440 B. Sandor

Choice (b) is of anything but iron. The best choice for racing should be wooden tires
attached by rawhide straps, metal clamps, or/and glue, although all these are complex, dif-
ficult to manage, and of questionable reliability.8 Such tires would provide the desirable
stiff rim, no significant wheel consolidation, and possibly less weight and lower rotational
inertia than in the case of iron tires, depending on the depth of the wooden tire layer. On
the whole, this wooden tire system could be a winner, assuming that it would survive the
severe sideways forces that are pres-
ent during high-speed turns. A good
view of some of the crucial issues is
given by the Lyon circus mosaic where
two crashes are in progress. One crash
(fig. 2) is clearly caused by a broken
wheel at the end of a turn around the
barrier (a place of severe loading),
with no evidence present of an iron
tire.9 Here is an illustration of a racing
team taking a calculated risk with the
lightest reasonable wheel, one of poor
structural integrity, ‘going for broke’
in the hope of winning the prize, but Fig. 2. Lyon circus mosaic, detail of crash (author; courtesy
losing everything. Musée Gallo-Romain de Lyon-Fourvière.

Fig. 3. Schematic of weight transfer in fast-turning vehicles. “Indy 500” is the common name of the Indianapolis
500-mile automobile race. The “Indy ref.” article reveals many important but subtle details of how purpose-
built, asymmetric tire sets are best for overall racing performance. Based on physics, an optimized racing
chariot must have a similarly asymmetric tire system, which ancient engineers had to know. These features
are not noticeable on any vehicle or photograph by casual observers, and not readily found in a Google search.

8 There were several ways to obtain the desired “python effect” for wheel consolidation. Rawhide
tires, shrinking during the drying process, provided consolidation and a small degree of shock
and vibration isolation, such as a moccasin would provide for the foot, but could last for only
c.10-20 km of slow travel on good roads. Hurford, pers. comm. and see www.chariotmaker.com
9 B. I. Sandor, “Chariots’ inner dynamics: springs and rotational inertias,” in A. Veldmeijer and
S. Ikram (edd.), Chasing chariots (Proc. 1st Int. Chariot Conference, Cairo 2012) (Leiden 2013)
217-28.
 Tire choices in Roman chariot racing 441

Choice (c) is the most conservative approach: both wheels are fully protected with iron
hoops. Such a chariot is likely to finish the race without any wheel failure, but it is not very
likely to win against more daring competitors who employ lighter wheels.
Choice (d) offers an intriguing compromise: protect with iron the wheel which is going
to be the most stressed, but not waste iron on the other wheel. Of the two possibilities,
only choice (d2) makes sense, because the right wheel is by far the most stressed by the
recurring counterclockwise turns. For Roman chariots and modern racing cars alike, this is
caused by the weight transfer to the right wheel(s) during left turns (fig. 3). Today’s riders
in cars and trucks often feel this kind of weight transfer on fast turns.
We should also consider the advantages and disadvantages of using iron nave (hub)
hoops, necessary to prevent catastrophic splitting of the hubs. Since these bands have simi-
lar but much smaller effects to those of tires in terms of the additional weight and weight
distribution, we will not discuss them in detail here, but it is good to remember that in
racing even minute changes in weight could dramatically alter the dynamic performance.
Thus nave hoops should remain part of the overall analysis.
In spite of the clear technical advantages of a single-tire configuration, it must be
understood that a racing chariot with an iron tire on the right wheel only (d2) would not
necessarily produce a winning machine, because any racing event involves an infinite
number of controllable and uncontrollable variables (e.g., the diet of the horses, air tem-
perature, etc.), but it makes full sense in engineering thinking, ancient or modern. It is the
epitome of rational compromises in a complex system, elegantly balancing the need for
safety, structural durability, and desire for winning. This one-tire configuration is supe-
rior in a circus race to those where both wheels have iron tires, but it could lose a race to a
vehicle without any iron tires if that arrives at the finishing line without structural failure.

Fig. 4. Schematic of the relative merits of different tire systems in ancient racing.

The relative probabilities of winning with a variety of vehicle and tire configurations
is illustrated schematically (by crude but educated estimations) in fig. 4. The point of this
simplistic chart is that the probability of winning is always less than 100% for any vehicle;
442 B. Sandor

one with no iron tire(s) has a roughly 50/50 chance of winning or of failing entirely; one
with two iron tires has a lower chance of winning than the one with no iron; a special case
is the extremely heavy Ben-Hur (movie) chariot with two iron tires; while the Monteleone
chariot10 (not a dedicated racing chariot) could win only if all the others failed some way.
The chariot with only one iron tire (on the right side) has the best chance of winning if the
one with no iron fails.
What did the Romans know about such complex probabilities? Probably quite a lot,
but it was based on their own judgements. Three groups of people had the most vested
interest in dealing with racing on a probabilistic basis: team owners, drivers, and bettors.
The grounds for their choices included the keen observation that without any iron on
the wheels the right wheel was failing often and predominantly, while both wheels hav-
ing iron tires tended to be safe but was seldom a winning combination. Each of the three
major configurations probably resulted in some wins and some losses (fig. 4), which in
turn would have fueled a raging debate within all racing teams regarding the wisdom of
selecting any particular design. Eventually, some saw the wisdom of favoring a racing
chariot with an iron tire on the right wheel only (d2), with the best set of compromises, as
providing the highest probability of earning the most prize money when averaged over
many years of racing.
More archaeological finds of relevant models (and indeed of actual chariots) would
be most welcome in order to solidify our understanding of the issues discussed here. In
the meantime, the analytical approach is on firmer ground and can guide us in what to
look for,11 whether in evaluating new finds or in examining representations in works of
art (which often do not contain reliable details). As the finest available representation of a
Roman racing chariot, the Tiber model gives us a glimpse into the Romans’ probabilistic
thinking for winning races and bets.
sandor@engr.wisc.edu Dept. of Engineering Physics, University of Wisconsin, Madison

10 A. Emiliozzi, “The Monteleone chariot,” MetMusJ 46 (2011) 9-132.

11 The value of an analytically-based approach was demonstrated in the discovery of Neptune:
astronomers evaluating perturbations in the orbit of Uranus pinpointed where to look in the
sky to find a previously unknown object.