104 THE ROMAN EMPIRE
existing economy, and in particular by the prevailing agricultural system.
The effi ciency and productivity of ancient agriculture is to be judged with
reference to its ability to maintain ancient societies over time.
On yields, it hardly needs saying – the agronomists all make this obvious
observation – that there must have been enormous variation from one area
and terrain to another in the Mediterrenean region. This is easily illustrated
for large units and for the modern world from contemporary data.^25 Average
wheat yields in the third decade of this century in kilograms per hectare (kg/
ha) at a sowing rate (e.g.) of 135kg/ha range from 1,710 or a little less than
thirteen- fold in Egypt to 269 or two- fold in Cyrenaica. Elsewhere in north
Africa, Tunisia registered 400 and Algeria 540, or three- fold and four- fold
respectively. Italy including Sicily had an average yield of 1,200, a little
under nine- fold, and Greece 620, or about four- and-a- half-fold. This should
warn us against generalizing from the whole of the Mediterranean (with or
without Egypt, which practised irrigation agriculture) or the whole of the
Roman empire. But generalizing about yields for Italy or Greece or the north
African provinces is just as dubious a practice. The contrast visible in the
modern data between the extraordinarily high yields reported from the
wadi- valleys and alluvial fans of the interior of Tripolitania or southern
Numidia and the modest returns from the dry- farming belt in north Africa
would be mirrored in the ancient sources, if we had a full complement of
evidence; as it is we have no yield fi gures from antiquity with which to
contrast a number of notices of yields evidently achieved through fl ood- zone
practices, which both ancient writers and modern commentators have found
incredible.^26
Leaving aside this data from north Africa, the ancient literary sources do
contribute a few scraps relevant to Italy and Sicily. There are three main
texts. Varro ( Rust. 1.44.1), commenting on the diversity of yields from one
district and soil to another, says that one place might yield ten- fold and
another fi fteen- fold from the same seed, as in some parts of Etruria.
Columella states (3.3.4) that at least in the greater part of Italy, a four- fold
yield in cereals ( frumenta, not, we note, in wheat, triticum ) was rare,
implying that the yield usually fell below this level. Cicero in the Verrines
(2.3.112) gives for the territory of Leontini in Sicily the sowing rate of six
modii per iugerum, or a little over 160 kg/ha, a wheat yield in a good year
of eight- fold, a little over 1,300 kg/ha; and another yield in an excellent year
of ten- fold, or about 1,625 kg/ha.
The comments of Varro and Columella are both very brief. Varro is
treating legumes and cereals together in a chapter dealing with sowing. He
gives specifi c sowing rates for beans, wheat, barley and emmer ( far ), adding
the caveat that they should be varied according to locality and richness of
soil. Then follows the illustration from Etruria. This reads like an authentic
piece of information drawn from a good source, and it suggests that relatively
high yields were a fact of life in Etruria. It is signifi cant that the lower of the
two yields cited is still high, when a more striking contrast would have