New insights into early medieval Islamic cuisine: Organic residue analysis of pottery from rural and urban Sicily

Abstract

Sicily, during the 9th-12th century AD, thrived politically, economically, and culturally under Islamic political rule and the capital of Palermo stood as a cultural and political centre in the Mediterranean Islamic world. However, to what extent the lifeways of the people that experienced these regimes were impacted during this time is not well understood, particularly those from lesser studied rural contexts. This paper presents the first organic residue analysis of 134 cooking pots and other domestic containers dating to the 9th -12th century in order to gain new insights into the culinary practices during this significant period. Ceramics from three sites in the urban capital of Palermo and from the rural town of Casale San Pietro were analysed and compared. The multi-faceted organic residue analysis identified a range of commodities including animal products, vegetables, beeswax, pine and fruit products in the ceramics, with a complex mixing of resources observed in many cases, across all four sites and ceramic forms. Alongside the identification of commodities and how they were combined, new light has been shed on the patterning of resource use between these sites. The identification of dairy products in calcite wares from the rural site of Casale San Pietro and the absence of dairy in ceramics from the urban centre of Palermo presents interesting questions regarding the role of rural sites in food consumption and production in Islamic Sicily. This is the first time organic residue analysis of ceramics has been used to explore foodways in a medieval multi-faith society and offers new pathways to the understanding of pottery use and resources that were prepared, consumed and combined, reflecting cuisine in different socio-economic environments within the pluralistic population of medieval Sicily.

Fig 1. Map of Sicily showing the location of Palermo and Casale San Pietro.

Colours represent the sites with ceramics used in this study where pink = CSP, yellow = GA, blue = PB and orange = CLESP. The main road between Palermo and Agrigento from North to South is marked on the map.

Fig 2. Plots of fatty acid stable isotope values obtained from individual vessels from Sicilian Islamic pottery.

a) Plot of δ¹³C_16:0 against δ¹³C_18:0. Ranges (68% confidence) of 269 modern authentic reference products are shown, D (Ruminant dairy), RF (Ruminant adipose), P (Porcine), M (Marine), and FW (Fresh water). These references are published elsewhere [47] b) Plot of Δ¹³C against δ¹³C_16:0. values <-3.3‰ are typically associated with D (Ruminant dairy), values between -3.3‰ and -1.0 ‰ are associated with R (Ruminant adipose) and above -1.0‰ can be considered as NR (Non- ruminant) [45].

Fig 3. Kernel density estimate of Δ¹³C values.

Bandwidth = 0.5. Δ¹³C values <-3.3‰ are typically associated with D (Ruminant dairy), values between -3.3‰ and -1.0 ‰ are associated with R (Ruminant adipose) and above -1.0‰ can be considered as NR (Non- ruminant) [45].

Fig 4. TIC chromatograms of pottery extracts.

a) AE chromatogram of sample CSP_25 that yielded δ¹³C values within the range of non-ruminant products, b) AE chromatogram of sample CLESP_29 that yielded δ¹³C values within the range of non-ruminant products (porcine) and shows the TAG distribution profile associated with this sample after SE, c) AE chromatogram of sample GA_34 that yielded δ¹³C values within the range of ruminant products adipose products and shows the TAG distribution profile associated with this sample after SE, d) AE chromatogram of sample CLESP_26 that yielded δ¹³C values within the range of products and shows the TAG distribution profile associated with this sample after SE. Parentheses indicate the δ¹³C_16:0 and δ¹³C_18:0 fatty acid values.

Fig 5. Plots of TAG information based on HT-GC data of individual pottery samples from Palermo sites (CSP, GA and PB) and Castronovo (CLESP) with interpretations based on fatty acid isotope values.

The dispersion factor (DF) and average carbon number (M) were calculated using statistical equations outlined by Mirabaud et al. [52]. Shapes represent ruminant dairy (●) based on Δ¹³C values <-3.3‰, ruminant adipose (◆) based on Δ¹³C values between -3.3‰ and -1.0 ‰ and non-ruminant (■) based on Δ¹³C above -1.0‰ [45].

Fig 6. TIC chromatograms of extracts typical of a variety of products identified in these ceramics.

a) AE chromatogram of sample GA_10 showing the presence of C₂₉₍₁₅₎ ketone that indicates the presumed presence of Brassica [9]. b) AE chromatogram of sample PB_26 showing the presence of C₃₁ ketone (hentriacontane-16-one) that indicates the presumed presence of leek (Allium porrum) in the ceramic samples [–13]. c) AE chromatogram of CLESP_58 showing the presence of C₂₉₍₁₀₎ ketone (nonacosane-10-one) that indicates the presumed presence of (Foeniculum vulgare) (Fennel) [14]. d) AE chromatogram of CLESP_12 indicating plant oil by a C_18:1/C_18:0 >2 and the presence of C_18:2 [2, 15]. e) Chromatogram of sample CSP_2 showing the presence of malic acid after acid butylation. f) SE chromatogram of sample CSP_C5 indicating the presence of Pinaceae biomarkers: retene, methyl-dehydroabietic acid (Methyl-DHA), dehydroabietic acid (DHA), 7-oxo-dehydroabietic acid (7-oxo-DHA). g) AE chromatogram of sample CSP 4 typical of beeswax as well as animal fat and pine products. h) SE chromatogram of CSP 4 showing distribution of alkanes and alcohols typical of beeswax products alongside HT-GC of ion 257 showing the distribution of WE. Internal standards alkane C₃₄ (IS1) and C₃₆ (IS2) are shown.

Fig 7. Malic acid yields and % tartaric acid (TA).

a) Proportions of tartaric acid in various plants and plant products [17] (S1 Table); b) Proportions of tartaric acid in CLESP and Palermo cooking pots, plotted against the amount of malic acid extracted. % TA = tartaric acid/(tartaric + malic acid) [17]. C₂₉₍₁₀₎ ketone (nonacosane-10-one) indicates the presumed presence of broad-leaved sermountain (Laserpitium latifolium) [81] or (Foeniculum vulgare) (Fennel) [82]. C₂₉₍₁₅₎ ketone indicates the presumed presence of Brassica [77]. C₃₁ ketone (hentriacontane-16-one) indicates the presumed presence of leek (Allium porrum) in the ceramic samples [–80, 91].

Fig 8

Summary figure of organic substances identified in pottery vessels from (a) CSP, (b) GA, (c) PB and (d) CLESP. Identification criteria of different commodities (ruminant dairy, ruminant adipose, non-ruminant, unidentified plant, plant oils, plant wax, beeswax, pine products, fruit products and grape products) are outlined in the text. Those not analysed for fruit acids are shown (/). Non-ruminant, ruminant and dairy were assigned based on Δ¹³C values. In one case both ruminant and dairy were identified based on clear dairy TAG distribution. Where P is noted in non-ruminant this refers to samples tentatively assigned to porcine. Specific taxonomy of plant waxes is indicated brassica (1), leek (2) and broadleaf sternum or fennel (3). In pine products pitch is indicated by (*). Evidence of heating was presumed in the presence of ketones C31, C33 and C35 [78] and/or APAAs [112, 113]. Vessel drawings used as examples based on actual vessels from CSP and PB [33, 34].