The Economic History of Byzantium From

origins of wheat produced on the state’s demesnes and destined for consumption in Constantinople. These granaries were located on the northern coast of Asia Minor (Aminsos, Amastris), in particular Bithynia (Kios, Panormos, Nikomedeia, and Pegai), Thrace (Herakleia) and Bulgaria (Philippopolis), Smyrna, and Cyprus (Paphos).117

Barley, which could be made into bread like most grain, grew everywhere because it is more hardy; a spring barley called leptitis is apparently mentioned only in the Geoponika.118

Millet (kenchros), also edible but apparently not much valued at the time, is a spring cereal. It is mentioned in the eleventh century in lists of exemptions and in the twelfth by Anna Komnene in connection with the region of Dyrrachion, and in Bulgaria.119 Finally, it should be noted that the cultivation of two other cereals, oats and rye, spread during the Middle Ages. Rye (briza) was unknown to ancient Greece and does not feature in the Geoponika, but it had been grown in the West since the early Middle Ages; by the thirteenth century it was also being cultivated in Chalkidike. It was used

for making bread.120

Oats (brome) had been no more than weeds in ancient Greece but are noted as fodder for sheep in the Geoponika and were grown for grain in the eleventh century, according to the list of exemptions. Oats have turned up in southern Italy on archaeological sites of a later period. They were certainly intended for animal consumption and may have played a part in feeding the army’s horses.121

Cultivated legumes, which corresponded to dry vegetables (ospria), are often mentioned without further detail in the texts and seem on the whole to have been the same as those in the Geoponika: lentils ( phake), ers (robin), peas (pissos), broad beans (kyamos), calavances (phasoulos), chick peas (erebinthos), and vetches (bikion);122 yellow lentils (phaba) are mentioned in the proto-Byzantine period and were grown on the western side of Asia Minor in the eleventh century.123 Lupine (thermos, then loupinos) occurred as a human comestible during the proto-Byzantine period and is still mentioned in the

117J.-C. Cheynet, “Un aspect du ravitaillement du Constantinople aux Xe–XIe s.,” Studies in Byzantine Sigillography 6 (1999): 1–26.

118Koukoules, Bi´o", 5:21–22; Teall, “Grain Supply,” 99–100; Geoponika, 3.3.12.

119Teall, “Grain Supply,” 99; for lists of exemptions, cf. for instance Lavra, 1: no. 48; Anna Comnene, Alexiade, ed. B. Leib, 3 vols. (Paris, 1945), 3:94; J. Darrouze`s, “Deux lettres de Gre´goire Antiochos ´ecrites de Bulgarie vers 1173,” BSl 23 (1962): 280 (millet bread).

120M. P. Ruas, “Les plantes exploite´es en France au Moyen Age d’apre`s les semences arche´ologiques,” in Plantes et cultures nouvelles: En Europe occidentale au Moyen Age et `a l’epoque moderne (Auch, 1992), 9–35; Xe`ropotamou, no. 9; Koukoules, Bı´os, 5:23.

121Jarde´, Ce´re´ales, 4; Geoponika, 18.2.6; Teall, “Grain Supply,” 99; Koukoules, Bi´o", 5:259; for France, cf. Ruas, “Les plantes,” 22; for exemption lists, see for instance Lavra, 1: no. 48; for southern Italy, see Martin, La Pouille, 331 n. 10.

122Lentils and ers, Xe`ropotamou, no. 9; peas, Kaplan, Les hommes et la terre, 30 n. 44; beans and calavances, Koukoules, Bi´o", 5:97; chick peas, ibid., 98; vetch, in Ottoman registers from the 15th century. For southern Italy: beans, chick peas, and “legumina,” Martin, La Pouille, 331. See also Koder, Gemu¨se.

123 Patlagean, Pauvrete´ ´economique, 39; Patmos, 2: no. 50 (phaba). MM, 4:202 (1192: phabata). Yellow

¨ ˆ lentils were grown in Bithynia during the 15th century: O. L. Barkan and E. Meric¸li, Hu¨davendigar

Livasi Tahrir Defterleri, vol. 1 (Ankara, 1988).

252 JACQUES LEFORT

Middle Ages.124 Legumes were cultivated in gardens, and at least some kinds were mostly grown in fields, as noted below.

Vegetables There were many kinds of vegetables, at least in the market-gardening suburbs around the large towns. On the basis of book 12 of the Geoponika, especially chapter 1, which reveals which vegetables were planted “under the climate of Constantinople,” and using many other sources, J. Koder has listed nearly one hundred vegetables, fresh or dried, from parsley to carrots, that were grown in the Byzantine Empire; some have nowadays been abandoned.125 The range was presumably often less varied in the countryside.

Industrial Crops Some plants were grown for industrial use, especially for textiles. Flax (linos, linarion) is mentioned in relation to its purchase at a set price in exemption lists for the eleventh century. It was produced in Macedonia, possibly Bulgaria, in Asia Minor, in Apulia, and in Calabria; oil was also extracted from it.126 Hemp (kannabis) cultivation was practiced more in Campania than in Apulia, and also in Chalkidike in the fourteenth century.127 During the period under consideration, cotton (bambax) was cultivated in Crete and probably also in Cyprus.128

Livestock In addition to the animals that immediately come to mind—such as horses and donkeys, mules (which did not require shoeing),129 bovines, including buffalo, goats, sheep, and pigs, as well as poultry130 and bees (mentioned above)—there were also camels, mentioned, for instance, in a novel of Nikephoros II Phokas as part of the excessive wealth acquired by some monasteries, probably in Asia Minor.131 The numerical importance of each species varied according to region, with sheep certainly the most numerous.132

Farming Techniques and Production The following paragraphs summarize the little that is known about farming techniques in Byzantium and about the possible yields of two

124 Geoponika, 2.39; Patlagean, Pauvrete´ ´economique, 39; Theophanes, Chronographia, ed. C. de Boor, 2 vols. (Leipzig, 1883–85; repr. Hildesheim, 1963), 1:419 (hereafter Theophanes) (loupinos); Koukoules, Bios, 5:98; for a reference to lupine in Campania in 1137, see Martin, “Travail agricole,” 118.

125Koder, Gemu¨se; idem, “Fresh Vegetables for the Capital,” in Constantinople and Its Hinterland, ed. C. Mango and G. Dagron (Cambridge, 1995), 49–56.

126Lavra, 1: no. 48; Koder, Eparchenbuch, 9.1; 9.6; Iviron, 3: no. 58; E. Patlagean, “Byzance et les marche´s du grand commerce,” in Mercati e mercanti nell’alto medioevo: L’area euroasiatica e L’area mediterranea (Spoleto, 1993), 599; Martin, La Pouille, 332; J.-M. Martin and G. Noye´, “Les campagnes de l’Italie me´ridionale byzantine (Xe–XIe sie`cles),” Me´lRome Moyen Age 101 (1989): 580; ODB, s.v. “Linen.”

127Geoponika, 2.40.2; Lefort and Martin, “Organisation,” 20 n. 31; Lefort, Chalcidique occidentale, index, s.vv. “lin” “chanvre,” “coton.”

128Malamut, Les ˆıles, 2:390; MM 6:96 (in Crete, 1118).

129Teall, “Agricultural Tradition,” 52.

130Chickens were included in the kaniskia that landowners owed to officials; ODB, s.v. “Fowl, domestic.”

131Svoronos, Novelles, 157.

132I referred above to the composition of the livestock that peasants owned at the beginning of the 14th century; the composition of some domanial flocks in the 11th century will be discussed below.

important lines of production, wine and cereals. Note, too, that the production of olive oil varied according to the age of the olive trees—according to my hypothesis, approximately 2 or 3 liters per tree.133 Olivier de Serres provides data on the production and income from mulberry trees in southern France in the sixteenth century: twenty to twenty-five trees yielded 10 quintaux of leaves, which yielded 5 to 6 livres of silk, worth 10 or 12 ´ecus.134

Cultivation took up limited areas mainly located on fluvial terraces, in the hilly zones between the slopes and the plains, which were, at the time, poorly drained, at least in areas where this geographical profile predominated. This agricultural space proved adequate for a long time; when more land was needed, it was enlarged, although this involved considerable clearance work. Once this had been done, the empire never needed to import foodstuffs. On the contrary, it exported them in the twelfth century.135

Gardens Although gardens are not always mentioned, even in the most precise descriptions of peasant properties, presumably most farms included one, since vegetables formed an indispensable part of a family’s nourishment. The size of gardens that have been listed (an average of 0.2–0.4 modios in several villages in Macedonia according to praktika from the beginning of the 14th century) was adequate, going by an average central European person’s consumption in the nineteenth century, which corresponded to 40 m2 (0.04 modios) of horticultural produce, including potatoes.136 The garden often lay close to the house for obvious reasons, being the plot that required the most work and manure, and, since houses were generally placed near sources of water, it could also be watered. When a garden was located at a considerable distance from the dwelling, as was sometimes the case, this was to benefit from irrigation, particularly, as noted above, along the diversion canals that brought water from the streams to the mills; precise agreements about sharing water were drawn up when the owner of the garden was not the owner of the mill.137

Large towns were surrounded by market-gardening suburbs, as was the case with Constantinople and, according to John Kaminiates, with Thessalonike from the tenth century on. These parcels of land often belonged to the powerful, in which case they

133For the cultivation of olive and fruit trees, see Geoponika, 9 and 10; for vines and olives, see A. I. Pini, “Vite e olive nell’alto medioevo,” in L’ambiante vegetale nell’alto medioevo (Spoleto, 1990), 329–70;

La production de vin et d’huile en Me´diterrane´e ( BCH, suppl. 26) (Athens-Paris, 1993); on oil presses, M. C. Amouretti et al., “A propos du pressoir `a huile,” Me´l Rome Antiquite´ 96 (1984): 379–421; ODB, s.v. “Olive Press.” Around 1970, in the Languedoc, production levels varied between 5 and 8 kg per tree between the fifth and fourteenth years, and between 10 and 30 kg after the fifteenth year, depending on variety and whether the crop was dry or irrigated; cf. R. Loussert and G. Brousse, L’olivier (Paris, 1978), 444; in the same region today, between 5 and 6 kg of olives are required to make a liter of oil. On the cultivation of olive trees in the 13th through 15th centuries, see Laiou, “Agrarian Economy,” 353.

134Lieutaghi, Livre des arbres, 870. On sericulture, cf. Guillou, “La soie,” 76–78.

135A. Kazhdan, “Two Notes on Byzantine Demography of the Eleventh and Twelfth Centuries,” ByzF 8 (1982): 120.

136Koder, Gemu¨se, 69.

137Iviron, 1: nos. 9 (10th century) and 30 (11th century); Lefort, Chalcidique occidentale, 136–37 (14th century).

254 JACQUES LEFORT

would have been cultivated by tenant gardeners. West of Thessalonike, the monastery of Iveron held some “fields” near the Golden Gate, one of which was already being exploited as a garden by the beginning of the twelfth century; it covered an area measuring six modioi, was called ta Keporeia, and included two wells and two cisterns. To the west and southeast of the town, the presence of tightly packed parcels of land suggests intensive use of the soil linked to the proximity of the urban market.138 The produce of the Iveron gardens near the Golden Gate is known in the fifteenth century: cabbage, leeks, carrots, garlic and onions, lettuce, cucumbers, pumpkins, and melons were all grown on separate plots.139 Such a regimented horticultural landscape may, perhaps, not have been as clearly defined in the countryside; vines and fruit-bearing trees were often associated with growing vegetables, as was emphasized by the practice of calling certain plots kepampelon or kepoperibolion.140

Meadows were certainly less rare on estate lands than on peasant holdings. As very valuable parcels of land, they were nearly always classified by the fisc as “first-quality land”; they could be fairly vast and were located in the most humid spots. Frequently irrigated and occasionally drained, they were cultivated for scything.141 In the Geoponika, vetch, alfalfa, lupine, and, as noted above, oats were grown for fodder.142 Hay was stored in barns, especially in places with no winter pasturage.143

Vineyards In 985 the monastery of Iveron gave the Athonite community a large vineyard of 30 plinthia (90 modioi) situated near Hierissos in eastern Chalkidike, which seems to have been quite exceptional.144 Generally speaking, vineyards tended to be small, on the order of one modios or a bit larger. They were most commonly held by small farmers. Of those vineyards properly belonging to a domain, some were exploited directly, though many more were rented out to peasants who did not necessarily live on the estate.145 In Macedonia, at the beginning of the fourteenth century, peasants in several villages owned on average 5 modioi of vineyards. By the end of the Byzantine period, viticulture represented perhaps 16% of the area under cultivation in some regions, which comes quite close to the 21% level achieved in Greece in 1860 and never exceeded since then.146 The area of land dedicated to vines may have been less previously, but these data show how important a role viticul-

138Kaplan, Les hommes et la terre, 63–65; John Kaminiates, De expugnatione Thessalonicae, ed. G. Bo¨hlig (Berlin, 1973), 5–6; J.-M. Spieser, Thessalonique et ses monuments du IVe au VIe sie`cle (Athens–Paris, 1984), 16; Iviron, 2: no. 52; Lefort and Martin, “Organisation,” 21 and n. 34.

139Iviron, 4: no. 97; cf. Laiou, “Agrarian Economy,” 352ff.

140On Latium, see Toubert, Structures, 1:210–14; see also Jardins et vergers en Europe occidentale, VIIIe–XVIIIe sie`cles (Auch, 1987).

141Ge´ome´tries, § 53, 54, 66, 150, 151, 202; Lavra, 1: no. 56.

142Geoponika, 2.39.4 (lupine), 3.1.14 (alfalfa), 3.6.7 (vetch); for meadows, see also Kaplan, Les hommes et la terre, 75–76.

143Iviron, 2: no. 52 (chortobolon).

144Iviron, 1: no. 7.

145There is an allusion to dues on rented vines (ampelopakton) in some 12th-century documents: in 1152 near Strumica, Iviron, 3: no. 56; in 1163 in Thessaly, C. Astruc, “Un document ine´dit de 1163 sur l’e´veˆche´ thessalien de Stagi,” BCH 83 (1959): 215.

146On the 19th century, S. Kourakou-Dragona, “ HJ ejxe´lixh tou' eJllhnikou' ajmpelw´ na,” in IstoriJ´a tou' eJllhnikou' krasiou'(Athens, 1992), 118.

ture played in the peasant economy and point to a very commercialized wine production.

What may be gleaned about viticultural techniques points, on the whole, to continuity with the practices of the previous period. In the tenth century, a letter addressed to the xenodochos of Pylai evokes the grape harvest and wine making in a manner that may owe much to the Geoponika, or alternatively, may point to the permanence of viticultural practices.147 Such permanence may well be the case with regard to the practice of digging out the whole plot prior to planting the vines, which is alluded to in the proto-Byzantine period. However, it was apparently less common than the simple and cheaper practice of digging ditches in which to plant vinestock.148 Although the digging-out practice is well attested during the Middle Ages under the term kylisma, it does not follow that this was always practiced.149 In any case, irrespective of the processes employed to produce the vine plants, the occasion for long passages in the Geoponika, it is clear that the best ones were selected with the aim of improving the quality and quantity of the produce.150 As in the proto-Byzantine period, vines were manured.151 The tasks that generally had to be carried out on the plot after planting, especially pruning, hoeing, harrowing, staking, and tying,152 had surely not changed. Three methods of training vines are mentioned in the Geoponika: rampant vines, low vines trained up stakes, and vines trained up trees. Evidence for all three is found in the period under consideration.153

Training vines along stakes was probably the most common method on plots where vines were the sole crop; according to an instruction from the fisc, the stocks were planted at intervals of 0.7 m, in regular rows, which could constitute units of measurement and made it easier to count vine stems.154 Since vines exhaust the soil and benefit from replanting, viticulture involved distinguishing between functional microplots, evidence for which is found in both the Geoponika and medieval documents: the new vines (phyteia, neophyton),155 the replanted vines, and old vines (palaiampela) that would one day be renewed.156 Nurseries, which were probably linked to the ancient domanial economy, are the only feature for which there is not much evidence.157 The diversity

147Theodore Daphnopates, Correspondance, ed. Darrouze`s and L. G. Westerink (Paris, 1978), 209.

148Geoponika, 5.19.1.

149Ge´ome´tries, §49 and 180 (in the theme of Thrace), 73 (Thrace and Macedonia), 191 and 193 (Bithynia); Iviron, 3: no. 67; MM 2:507 (1401). The term ampelotopion probably also designates a plot of land that has been dug over for planting vines.

150For instance Geoponika, 5.8.1.

151Ibid., 3.15.4; 5.16.8–9; Ge´ome´tries, § 28.

152Ge´ome´tries, § 49.

153Geoponika 4.1.11; 5.2.14; 5.2.15. Rampant vine (chamampelon), S. Cusa, I diplomi greci ed arabi di Sicilia pubblicati nel testo originale, tradotti ed illustrati (Palermo, 1868–82), 2:636; Darrouze`s, “Deux lettres,” 279; climbing vine (anadendras), see for instance Patmos, 2: no. 52.

154Ge´ome´tries, § 280; Xe`ropotamou, no. 16 (traphoi).

155Koukoules, Bi´o", 5:281.

156Cf. some of the microtoponyms at Radolibos from the beginning of the 12th century (Iviron, 2: no. 53): Chersampela, Palaiampela, Phyteia, Epano Phyteia, despotikh` futei´a.

157Psellos, (J. F. Boissonade, Anecdota graeca [Paris, 1829; repr. 1962], 1:245, does indeed mention the tree nursery (phytorion) that features in the Geoponika.

256 JACQUES LEFORT

of microplots, as implied by wine-growing techniques, explains in part why plots that were solely used for growing vines, and were normally enclosed, did not always form continuous land units and were not always clearly separated from the fields; at least, there was no such clear separation at Radolibos at the beginning of the twelfth century.158 Winepresses were often set up close to vineyards. I. Papangelos may have found remains of such presses in Chalkidike.159

In sufficiently humid regions, vines were frequently grown alongside trees on plots of land close to the dwellings; in some cases, the vines were simply grown beside the trees; in others it was a case of vines trained up trees as in the Geoponika, but the landscape they formed had changed completely since then. On the vast ideal estates of the proto-Byzantine period, where vineyards flourished over the plains, vines simply grew up poplar trees set 7 m apart in regular rows; fruit trees or crops could be grown in the gaps.160 In the medieval period, on the other hand, vines were supported by fruit trees, which were called dendra hypoklema for this reason, in little vine orchards (ampeloperibolia), suggesting intensive exploitation of the soil within a smallholding context,161 as in the case of the other types of plots, mentioned above, which combined a variety of crops.

Nothing is known about the yield, though a recent hypothesis suggests that it could amount to 25 hectoliters of wine per hectare in Chalkidike,162 or ca. 25 measures (metra) per modios,163 which is not inconceivable, although the only information available for the same region suggests a volume of yield that was twice as low.164 In any case, a grower who cultivated more than 2 modioi of vines would have produced more wine than required for consumption by his household.

Trees and specialized crops These included olive groves, which, like orchards165 (mentioned above in connection with smallholdings), were farmed directly on estates. Some, on the southern side of Mount Athos, comprised hundreds of olive trees.166 There were

158In 7th-century Galatia, the Vie de The´odore de Syke´ˆon, ed. A.-J. Festugie`re (Brussels, 1970), chap. 115, does, however, allude to the village vineyard. Other examples could be cited.

159I. Papangelos, ““Ampelo" kai` oi«no" sth`n mesaiwnikh` Calkidikh´,” in IstoriJ´a tou' eJllhnikou' krasiou'

(as above, note 146), 231–37. It is not always possible to distinguish between wine presses and olive presses.

160Geoponika, 4.1.

161Patmos, 2: no. 50. On viticulture, cf. Kaplan, Les hommes et la terre, 71–72; Martin, La Pouille, 340–43.

162Papangelos, ““Ampelos,” 224.

163Concerning the content of the metron, see Schilbach, Byzantinische Metrologie, 112–13.

164Iviron, 2: no. 42: at Athos, in 1080, a vineyard worth 100 nomismata (possibly 15 modioi in size) produced 124 measures of wine. In Castile during the 18th century, yield varied greatly according to place and year, and may have been in the order of 13 hl/ha; however, Castilian yields are low; cf. A. Huetz de Lemps, “Vignobles et vins de Castille du XVIe au XVIIIe sie`cles,” in Le vigneron, la viticulture et la vinification en Europe occidentale, au Moyen Age et a` l’e´poque moderne (Auch, 1991), 168–69. In France during the modern age, average yields seem to be in the order of 20 hl/ha: M. Lachiver, “La viticulture franc¸aise `a l’e´poque moderne,” ibid., 215.

165Cf. for instance Iviron, 1: no. 9.

166Xe´nophon, no. 1 (1089; 300 olive trees); Iviron, 2: no. 52; Patmos, 2: no. 50; Harvey, Economic Expansion, 145.

also banks of reed beds, reeds being used to plait ropes and weave baskets,167 as well as most of the industrial crops mentioned above, including mulberry trees.

Fields These were generally squat, rectangular parcels,168 often set in open landscapes, but varying greatly in size. In one specific case, the partially preserved cadaster of Radolibos, dating from the beginning of the twelfth century, includes a description of 979 fields, which measured on average 2.2 modioi.169 In Chalkidike, forty-six fields were sold at the beginning of the fourteenth century that measured on average 5.4 modioi.170 In other places, fields could be much larger; in Capitanata, a recently colonized region, the fields could sometimes assume very complex shapes and often measured more than a hectare (ca. 10 modioi);171 in Melanoudion in Caria at the end of the eleventh century, in a region dedicated to growing cereals, fifteen parcels measured on average 47 modioi.172 These figures point to the existence in some areas of tightly knit plots, all the more so when soil occupancy was long established, resulting in fields being divided up among heirs. In places where the plots were much more loosely knit, the explanation may lie in geographical or historical conditions that now elude us. Some fields lay within the village (esochorapha), but most of them were some distance away. Because peasants tended to cultivate the loamier fields closest to their dwellings before tackling more demanding soils, there was a tendency for cereal-growing areas to develop. Threshing floors were sometimes out in the countryside, close to the fields, and often appear to have belonged to one particular farmer.173

After harvest and before they were plowed again, fields were often turned over to pasture, which served to manure them.174 Perhaps this method of fertilization was sufficient; Pliny thought that wheat needed less fertilizer than barley, and could even do without.175 Theophrastus and Pliny refer to beans and the Geoponika to lupine as sources of green fertilizer;176 to my knowledge, no text confirms the use of green fertilizer in this way during the Middle Ages, although this is not inconceivable. In any case, yields did increase, though of course slowly, as a result of selective sowing, an ancient practice that is mentioned by the Roman agronomists and in the Geoponika, whose advice was adopted by Psellos when writing on the subject. There is no reason why Byzantine peasants should not also have selected their sowing seed, though this process

167Geoponika, 3.2.2; Cecaumeni Strategicon, ed. B. Wassilievsky and V. Jernstedt (St. Petersburg, 1896; repr. 1965), 36 (hereafter Kekaumenos, Strategikon); Iviron, 2: no. 52.

168For central Italy, see Toubert, Structures, 1:283–84, 294–96; for Macedonia, Lefort, “Radolibos,” 212–13.

169Iviron, 2: no. 53.

170Xe`ropotamou, no. 16.

171Lefort and Martin, “Organisation,” 21.

172Patmos, 2: no. 50.

173Iviron, 2: no. 53; Kaplan, Les hommes et la terre, 59–60. For harvests, see the literary text mentioned above, Daphnopates, Correspondance, 209.

174Farmer’s Law, sec. 27; Iviron, 1: no. 9.

175Pliny the Elder, Naturalis historia, 18, p. 53: “land that has not been manured will be sown with wheat rather than with barley” (after Theophrastus); Teall, “Grain Supply,” 129.

176Theophrastus, Historia plantarum, 8.9.1; Pliny, Naturalis historia, 18, p. 30; Geoponika, 2.39.6.

258 JACQUES LEFORT

occurs automatically to some extent.177 Furthermore, yields could hold their own and even increase when the fallow land was planted with legumes, some of which served to increase its fertility, as the authors of antiquity had previously pointed out.178

Not much data is available on cereal-growing methods in Byzantium. Practices must have varied according to region and period. References to extreme cases in some literary texts lead one to deduce that growing cereals was a precarious and unproductive process or, alternatively, a miraculous one. Little can be drawn from information of this kind. Documentary evidence points to the likelihood that land commonly lay fallow every second year, at the end of the period under consideration, north of the Aegean, in Macedonia at least. This was certainly the case in Chalkidike in the thirteenth century, since Theodore Skaranos sowed 103 modioi of seed179 over his 270 modioi of land, part of which, therefore, was not directly farmed. Consequently, by late November, he was probably planning on sowing spring wheat in fallow land as a catch crop, and following the same process with the millet and ers that he harvested.180 A two-year rotation also seems to have been followed at Radolibos at the beginning of the twelfth century, albeit for negative reasons, since the tiny size of the farms would not have allowed land to be left fallow for more than one year and even suggests that part of the fallow land was farmed. Crop rotation, involving wheat and legumes or wheat and spring barley, with the second sowing on fallow land, is very ancient and is mentioned in the Geoponika.181 This is not to say that catch crops were current practice nor continuously engaged in through the centuries, but that they remained a possibility within the crop system when the population increased, one that was certainly used before the thirteenth century. Pliny had recourse to Virgil when he clearly stated the relationship between lack of space and the practice of catch crops: “Virgil advises allowing the fields to rest every second year—if the size of the farm permits it, this is certainly very useful; if it is not possible, one must sow wheat in a field where lupine or vetch or beans or any other plant that enriches the soil has been harvested.”182

Both archaeology and the textual sources allow one to deduce the importance of legumes and their role in the crop system; they were cultivated in Syria during the proto-Byzantine period; vetches appear to have formed part of the field crops in the seventh-century Negev.183 The purchase of dried vegetables at a fixed price is mentioned in the eleventh-century lists of exemptions, which suggests that these were not simply garden produce limited to consumption by the family, as has been asserted, but

177Geoponika, 2.16.1–3; Psellos, Boissonade, Anecdota graeca, 1:242; J. R. Harland, Les plantes cultive´es et l’homme (Paris, 1987), 172.

178Pliny, Naturalis historia, 18, p. 50; Geoponika, 2.12.2; Toubert, Structures, 1:250.

179Allowing for 1 modios of wheat (12.8 kg) being sown over 1 modios of land (about 1,000 m2): Ge´ome´tries, p. 216. This was close-set sowing; see p. 254 below.

180J. Lefort, “Une exploitation de taille moyenne au XIIIe sie`cle en Chalcidique,” in Afiej´rwma sto`n Ni´ko Sborw'no(as above, note 1) 1:368, 370.

181Jarde´, Ce´re´ales, 85–86: Geoponika, 3.6–7 (grain/legumes), 2.12.2 (legumes/wheat), 3.3.12 (wheat/ spring barley); on crop rotation in medieval Latium, Toubert, Structures, 1:246–50.

182Pliny, Naturalis historia, 18, p. 50.

183Tate, “Les campagnes,” 71; Patlagean, Pauvrete´ ´economique, 247.

that they had a place in the cereal crop cycle.184 This was the case in Latium in the first half of the tenth century;185 in Campania, calavances, broad beans, and chick peas have been attested since the first half of the eleventh century, whereas in Apulia, the introduction of legumes into the crop rotation apparently dates only from the beginning of the twelfth century.186 In Macedonia and Bithynia, Ottoman tax registers from the fifteenth century show that legumes were an integral part of the agricultural cycle.187 Though these facts apply only to specific regions and periods, the important point is that they reveal the existence of practices that had long been known.

Generally speaking, while the system of agriculture was certainly traditional, it was inherently capable of improvement, up to a point. Land clearance was an altogether different matter, and there is reason to think that the decision to bring part of a forest or marshland into cultivation was taken only when every possibility of improvement had been exhausted, when it was no longer enough simply to farm the available fields as intensively as possible.

There is no direct information about yields, which would in any case have varied greatly from year to year. In Greece, the wheat harvest in 1921 yielded on average 6.6 quintals per hectare; it varied from 4.9 in Chios to 11.5 in Arcadia, and was 9.8 in Macedonia. Barley yielded rather more, 7.1 quintals per hectare, as in the case of mixed wheat and barley crops, which yielded 8.3.188 In 1938, in eastern Anatolia, a dry crop close to the Murat River yielded 6.3 quintals of wheat and only 4 of barley per ha.189 These proportions were certainly not surpassed by Byzantine agriculture, nor indeed achieved unless as an exception. Current research into the Ottoman tax registers shows that in Bithynia at the beginning of the sixteenth century yields were very diverse locally, depending on soil, altitude, and exposure, and that on average they often oscillated between 4 and 6 quintals per hectare.190 With regard to the thirteenth century, we may deduce from the testament of Skaranos that barley yielded 5.3 grains for one, this is, approximately 5.4 quintals per hectare.191 Somewhat insecure calculations concerning Radolibos at the beginning of the twelfth century suggest that the minimum yield from cereal crops was 5.1 grains for 1, that is, ca. 5.3 quintals per hectare.192 These calculations have been carried out with respect to an average holding, comprising one ox and 25 modioi of land, supposing that the holding had a balanced budget. The calculations take into account the necessary sowings, the probable consumption, and dues in cereals as known to us; at the same time, it has been assumed

184Cf. for instance Lavra, 1: no. 48.

185Toubert, Structures, 1:248.

186Martin, “Travail agricole,” 118; idem, La Pouille, 336.

187Cf. examples in Lefort, Chalcidique occidentale; for Bithynia, Barkan and Meric¸li, Hu¨davendigaˆr.

188Jarde´, Ce´re´ales, 203–5. On mixed crops cf. above, note 111.

189Kaplan, Les hommes et la terre, 82.

190According to data collated by Barkan and Meric¸li, Hu¨davendigaˆr.

191Lefort, “Une exploitation,” 369. Yield in quintals per ha yield per grain 0.128 (quintal sown per modios) the number of modioi per ha. We have seen (cf. note 51) that one modios seems to have amounted to 1,250 m2 in Macedonia.

192Lefort, Radolibos, 222.

260 JACQUES LEFORT

rather arbitrarily, though not implausibly, that one-eighth of the fallow land was used to grow spring cereals or legumes—a hypothesis suggested by N. Kondov (for the 14th century) in an article published in 1974, which has perhaps not received sufficient attention.193

Kondov considers unlikely the low estimate put forward by Svoronos in his Le Cadastre de The`bes (3 grains for 1) and has proposed a less pessimistic estimate for the medieval Balkans, relying primarily on agronomic arguments, especially with regard to our knowledge of the normal density of sown grain during the Byzantine period: a modios of wheat sown over one modios of land corresponds to an intensive sowing, as advised in the Geoponika,194 of more than 300 grains to the square meter. This sowing is characteristic of intensive farming. Kondov considers it probably correct to assume an average yield of 4.2–5.2 grains for one in the case of cereal crops in fourteenth-century Macedonia, adding that this rate of production, the higher estimate at any rate, would have resulted in a marketable surplus.195 Indeed, yields from cereal crops were probably less poor than has long been stated, and recently too, on the basis of certain estimates by Svoronos: 3 or 3.5 for one.196

As noted earlier, Byzantine cereal crops were grown within a smallholding context. The little that is known about them—fields worked with plows, the probable existence of biannual fallow land and of catch crops, and the introduction of new plants—sug- gests that agricultural practice in the Middle Ages was not less elaborate than in the proto-Byzantine period. Some seventh-century papyri refer to very contrasting yields, as one might expect: 4–5 grains for one in one case but, probably in a more favorable area or during a good year, 8–9 for one in another.197 Similar yields are obviously no less likely in the Middle Ages, the more so in that, until the twelfth century, very often only the best land was put to the plow. In the most fertile regions, average yields of slightly more than 5 quintals to the hectare were apparently plausible in the twelfth century. Grain was stored in dug-out silos198 or in lofts.199

The picture presented by scholars in the past, of an extensive cereal production spread over huge areas in a routine and unproductive manner, has played an impor-

¨

193 N. Kondov, “Uber den wahrscheinlichen Weizenertrag auf der Balkanhalbinsel im Mittelalter,” EtBalk 10 (1974): 97–109.

194 Geoponika, 2.20; Kondov, “Weizenertrag,” 108.

195 Kondov, “Weizenertrag,” 108–9.

196 Harvey, Economic Expansion, 180; Kaplan, Les hommes et la terre, 82. Svoronos, “Structures ´economiques,” uses a yield of 1:3.5 (p. 57 and n. 32); however, several of his evaluations are rightly based on a yield of 1:5 (cf. p. 58). According to our calculations, set out below, the only probable conclusion is that an average yield of 1:3.5 for second-quality land, with no catch crops, is not credible when set alongside the high fiscal levies that we have accepted.

197 Patlagean, Pauvrete´ ´economique, 246–48. In 6th-century Lycia, a 1:5 yield was considered very

198Ioannis Cinnami Epitome rerum ab Ioanne et Alexio Comnenis gestarum ed. A. Meineke (Bonn, 1836), 106 (hereafter Kinnamos); D. Z. Sophianos, ”Osio" Louka'", oJ bi´o" tou' oJsi´ou Louka' tou' Steiriw´ tou

(Athens, 1989), 217; Kaplan, Les hommes et la terre, 125.

199Iviron, 2: no. 52.