| No. of slaves per household | No. of households | Percentage of households |
| 1 | 1,143 | 74 |
| 2 | 330 | 21 |
| 3 | 57 | 4 |
| 4–6 | 12 | 1 |
Source: ASG, CdSG, N.185,01009.
Evidence from wills indicates a similar pattern in Venice. The famous traveler Marco Polo manumitted one slave in his will in 1324, and the Venetian painter Nicoletto Semitecolo also manumitted one slave in 1386.22 Panthaleo Iustiniano, a procurator of S. Marco, manumitted three slaves in 1393.23 Madalutia, the wife of Bernardo Aymo, manumitted two slaves in 1410.24 Andrea Barbarigo, a fifteenth-century patrician known for frugality, had two slave women in his household and one rented out.25 Giosafat Barbaro manumitted one slave in 1493.26 At the upper end of the spectrum, the apothecary Nascimbene de Ferraria mentioned six slaves in his will, but only four were present in his household.27
The number of slaves per household in the Mamluk sultanate ranged more widely. ‘Abd al-Laṭīf ibn ‘Abd al-Muḥsin al-Subkī boasted that he had gone through more than a thousand slave women, but elite civilian households more normally counted their slaves in the dozens.28 At lower social levels and after slave retinues shrank in the late fourteenth century, Mamluk civilian households resembled Italian ones. The households represented in a cache of fourteenth-century documents from Jerusalem had no more than four slaves.29 Ibn Ṭawq, a professional witness (shāhid) in Syria, owned two slave women.30 In Damascus, the estate of a private secretary of the sultan (kātib sirr) included five slaves, and a judge’s estate had four.31 A study of Syrian amirs in the late fifteenth and early sixteenth centuries found that none had more than one umm walad.32
Military households tended to have more slaves. Each amir commanded a unit composed of both mamluks and free horsemen.33 The size of the unit depended on the amir’s rank: an amīr mī’a muqaddam alf could have one hundred mamluks, an amīr arba’īn or amīr ṭablakhāna could have forty, an amīr ‘ashara could have ten or twenty, and an amīr khamsa could have five. Since the Mamluk army was theoretically composed of twenty-four amirs of one hundred, forty amirs ṭablakhāna, twenty amirs of twenty, fifty amirs of ten, and thirty amirs of five, there should have been a total of 26,650 mamluks in the army, both enslaved and manumitted.34 Each amir registered his mamluks with the diwān al-jaysh, which administered the army, and was not supposed to purchase more mamluks than his rank allowed. In practice, powerful amirs could accumulate 150 to 1,000 mamluks.35 Although a large mamluk corps was prestigious, amirs with too many mamluks risked the sultan’s retaliation if they appeared to challenge his power.36
In addition to mamluks, middle-ranking military households might have thirty or forty domestic slaves, while prominent amirs had two or three hundred.37 The number of slave concubines varied. Tankiz, governor of Damascus in the early fourteenth century, had nine, the amir Qawṣūn had sixty, and the amir Bashtak had eighty.38 Taghrī Birdī, governor of Damascus and father of the chronicler Yūsuf ibn Taghrī Birdī, had eight slave mothers (ummuhāt awlād) as well as a group of concubines who had previously belonged to Sultan Barqūq.39 Contemporary estimates of the total slave population of a single military household are rare, but Sunqur, governor of Bahnasā, owned sixty slave concubines, thirty additional slave women, and fifty mamluks at the time of his death in 1335.40
Table 2. Number of Mamluks per Sultan
Sources: Ibn Taghrī Birdī, Al-Manhal al-ṣāfī, no. 717; al-Yunīnī, Dhayl mirāt, 3:250; Ibn Iyās, Bidāʾiʿ al-zuhūr, 1:1:361; Simeonis, Itinerarium, 72–73, 79; Frescobaldi, “Pilgrimage,” 47; Gucci, “Pilgrimage,” 100; Lannoy, OEuvres, 116; “Piloti, Traité,” 54; Adorno, Itinéraire, 188; Ghistele, Voyage, 31; Breydenbach, Sanctarum peregrinationum, fol. 85r; Fabri, Evagatorium, 18:25; Harff, Pilgrimage, 106–7, 124.
The largest Mamluk household was naturally that of the sultan. The sultan’s mamluk corps (al-mamālīk al-sulṭāniyya) consisted of three groups: mamluks purchased by the reigning sultan (al-mushtarawāt or al-julbān), mamluks inherited by the reigning sultan from previous sultans (al-qarāniṣa), and mamluks inherited or confiscated from amirs who had died or lost favor (al-sayfiyya).41 The citadel of Cairo had twelve barracks (ṭibāq) with a capacity of one thousand mamluks each.42 However, contemporary estimates of the size of the sultan’s mamluk corps varied widely, as shown in Table 2. The sultan’s mamluk corps were subject to high turnover: several sultans purged the qarāniṣa and the sayfiyya to consolidate their power. The most notable purges were those of al-Nāṣir Muḥammad in 1310 and al-Nāṣir Faraj in 1411–1412.43 Their effects on the slave trade remain to be investigated.
In addition, the sultan’s household might include anywhere from forty to twelve hundred women, enslaved and free, as well as six hundred eunuchs and an unknown number of domestics.44 A sultan’s chief wife might have her own retinue of up to one thousand slave women.45 Like other military households, the sultan’s household tended to be smaller in the fifteenth century, and fifteenth-century sultans were more likely to be monogamous.46
Figure 2. Median Age of Slaves Sold in Genoa. Black points are based on more data (at least ten sales per year) than gray points.
(ASG, CdSG, N.185,00624, N.185,00625; ASG, Not. Ant. 172, 236–39, 253, 258, 265, 273, 286–87, 292, 363, 366–67, 379–82, 396–405, 449, 685, 719, 768; ASG, Notai ignoti, b.xxiii; Balard, “Remarques”; Balard, La Romanie; Cibrario, Della schiavitù; Amia, Schiavitù; Delort, “Quelques précisions”; Epstein, Speaking; Ferretto, “Codice diplomatico”; Gioffrè, Il mercato; Heers, Gênes; Tardy, Sklavenhandel; Tria, “La schiavitù”; Verlinden, “Esclavage et ethnographie”; Williams, “Commercial Revolution.”)
A third approach to slave demography is to consider the balance of age, gender, and origin. The origins of the Mediterranean slave population are discussed in Chapters 2 and 5. Age was included in Italian slave sale contracts, but notaries habitually rounded slaves’ ages to the nearest multiple of two or five.47 The median age of slaves sold in Genoa and Venice was normally between fifteen and twenty-five years old, as shown in Figures 2 and 3. Girls in their early teens seem to have been the most desirable for domestic work. When Cataruccia Dolfin asked her cousin in Alexandria for a slave, she requested a girl twelve years or older because “you can better use them in this age as you want.”48 Guglielmo Querini also preferred slaves between twelve and fifteen years old for domestic service.49 Francesco Datini thought that a girl between six and ten would learn his ways more quickly and provide him with better service.50 Further analysis shows that the slave women for sale tended to be a few years older than men and that the average age of slaves from the Black Sea increased after 1460, when exporting them became more difficult.51