Author's personal copy Veget Hist Archaeobot DOI 10.1007/s00334-011-0310-6 ORIGINAL ARTICLE Assessing past agrobiodiversity of Prunus avium L. (Rosaceae): a morphometric approach focussed on the stones from the archaeological site Hôtel-Dieu (16th century, Tours, France) Pauline Burger • Jean-Frederic Terral • Marie-Pierre Ruas • Sarah Ivorra • Sandrine Picq Received: 18 October 2010 / Accepted: 20 July 2011  Springer-Verlag 2011 Abstract Abundant and diverse Prunus fruitstone remains from cherries, plums, sloes, peaches, etc. are frequently recovered from archaeological waterlogged contexts such as wells, latrines, lake dwellings etc. in Europe. The distinction between most of the Prunus species, based on traditional morphological characters of the fruit stones, is usually not problematic. However the discrimination between P. avium L., P. cerasus L. and related cherry species, based on classical criteria alone, often turns out to be ambiguous because of the increasing number of varieties which have been bred since Roman times. By combining geometric and traditional morphometrical approaches, the overall variation in shape and size of stones from French and Swiss excavations dating from the 1st century to the 16th century A.D. were assessed. Among these important archaeobotanical data, the detailed examination of 100 waterlogged stones from the 16th Communicated by M. Latałowa. Electronic supplementary material The online version of this article (doi:10.1007/s00334-011-0310-6) contains supplementary material, which is available to authorized users. P. Burger (&)
J.-F. Terral
S. Ivorra
S. Picq Centre de Bio-Archéologie et d’Ecologie (CBAE) (UMR 5059 CNRS/Université Montpellier 2/EPHE), Equipe Ressources Biologiques, Sociétés, Biodiversité, Institut de Botanique, 163 rue Auguste Broussonet, 34090 Montpellier, France e-mail: burgerpauline@gmail.com J.-F. Terral Université Montpellier 2, Place Eugène Bataillon, 34095 Montpellier cedex 5, France M.-P. Ruas Archéozoologie, Archéobotanique-Sociétés, Pratiques et Environnements (AASPE) (UMR 7209 CNRS/MNHN), Muséum National d’Histoire Naturelle, 55 rue Buffon, 75231 Paris Cedex 05, France century Hôtel-Dieu cesspit at Tours, France, revealed that the morphological diversity is structured into two distinct morphotypes which diverge mainly according to geometrical features. Finally, the comparison between morphological features of these well-preserved archaeological stones and modern reference material including P. avium, P. cerasus and P. 9 gondouinii, suggests that these two morphotypes, which have been initially attributed to P. avium (long stones) and P. avium/cerasus (rounded stones) according to traditional morphological parameters, would correspond to two different cultivated varieties, both belonging to Prunus avium. Results presented in this work constitute new and preliminary data obtained during the development of this project that throw light on morphological variability and biosystematic aspects. Keywords Agrobiodiversity
Archaeobiology
Cultivars
Morphometrics
Morphotypes
Prunus avium Introduction Most Prunus species are deciduous and often spiny trees and shrubs with fruits as drupes, usually with a juicy fruit flesh (Hanelt 1997; Rehder 1940). The Prunus genus which comprises more than 400 species (Rosaceae family; Linnaeus 1753), is widely distributed in the northern hemisphere with many wild and cultivated representatives in Europe. It is an economically and ecologically important group with many cultivated species, notably P. dulcis (Mill) D.A.Webb (almond), P. armeniaca L. (apricot), P. persica (L.) Batsch (peach), and P. domestica L. l.s. (plums, in the wide sense). Despite the geographical range and the popularity of this genus, the evolution history and taxonomy of Prunus remain unclear. Progenitors of many 123 Author's personal copy Veget Hist Archaeobot species and cultivars are often hypothesized, but not definitively identified. Prunus avium L. (sweet and wild cherries), P. cerasus L. (sour, dwarf or morello cherry), P. fruticosa (ground cherry) and P. 9 gondouinii (duke cherry), along with related cherry species and their interspecific hybrids constitute the Eucerasus section, part of the Cerasus (Mill.) Focke subgenus of the genus Prunus (Santi and Lemoine 1990; Thorne 1992). This classification was defined based on morphological criteria by Rehder (1947) and later by Krüsmann (1978), and has been further confirmed by chloroplast DNA variation analysis (Badenes and Parfitt 1995). The ancestors of the modern P. avium (sweet cherries) seem to have originated around the Caspian and Black Seas, from where they have slowly spread, a phenomenon initially driven by birds, hence the species name P. avium (Dirlewanger et al. 2007). Requiring warm and dry summers, but adequate rainfall or irrigation during the growing season, the natural range of P. avium covers the European temperate regions from the southeastern part of Russia to the northern part of Spain (Hedrick 1915). According to archaeobotanical data, the wild P. avium seems to have been collected by Mesolithic hunter-gatherers in southern France, but very few charred stones have been found (Vaquer and Ruas 2009). Most of the other mentions from Mesolithic sites in northern Europe seem doubtful (Bakels 1991). The fruit are more frequently recorded from the Neolithic and Bronze Ages 5500–4000 B.C. (Hedrick 1915; Marshall 1954; Bertsch and Bertsch 1949, Out 2009). It is suggested that sweet cherries may either be indigenous in southern Europe as small isolated populations, or may have been introduced in these areas since the Neolithic. The ground cherries, from the most cold hardy cherry species P. fruticosa, are widespread over the major part of central Europe, Siberia and northern Asia (Hedrick 1915). Based on fossil evidence, P. cerasus seems to have originated around the Caspian Sea, from an area very similar to that of sweet cherries, and it appears to be native to northwest and central Europe (Dirlewanger et al. 2009; Watkins 1995). The hybrid origin of P. cerasus was first suggested by Olden and Nybon (1968) on the basis of morphological and biochemical evidence. Continuous variations between the P. fruticosa and P. avium characteristics were observed in sour cherry throughout its geographical range. P. cerasus seems to be closer to P. avium in western Europe, whereas it is most closely related to P. fruticosa in eastern Europe (Hillig and Iezzoni 1988; Krahl et al. 1991). Many genetic studies have confirmed the hybrid origin of P. cerasus and have shown P. fruticosa and P. avium as its progenitors (Hancock and Iezzoni 1988; Santi and Lemoine 1990; Schuster and Schreiber 2000). During the early Middle Ages, cherries and plums were 123 highly regarded. The establishment of the cherry orchard (‘‘Kirschgarten’’) tradition in northern and central Europe would be linked to the increasing medieval taste for these flavoursome fruits (Kroll 2007; Kroll and Willerding 2004). P. cerasus was well represented in these orchards, alongside P. mahaleb, P. fruticosa and P. avium in the early Slavonic stronghold of Mikulčice in Moravia (Opravil 1998) and in several medieval sites located in eastern central Europe (Kroll 2007; Kroll and Willerding 2004; Medović 2004). The duke cherries from the fertile P. 9 gondouinii Rehd. species, previously known as P. acida Dum, Cerasus regalis or Prunus avium ssp. regalis, seem to be intermediate between P. avium and P. cerasus (Dirlewanger et al. 2009; Faust and Suranyi 1997; Saunier and Claverie 2001). Based on the analysis of 75 AFLP markers, P. avium and P. cerasus were established as the progenitors of P. 9 gondouinii Rehd. (Tavaud 2002; Tavaud et al. 2004): duke cherries would result from the pollination of sour cherry by unreduced gametes of sweet cherry (Iezzoni et al. 1990). Since antiquity in Europe, archaeobotanical data have recorded a significant increase in fruit species diversity (Ruas 1992, 1996; van Zeist 1991). In his Natural History, Pliny the Elder (1st c. A.D.) mentioned notably nine varieties of sweet cherries planted in Rome (André 1981), suggesting that several cherry varieties had already been imported and probably cultivated at this time in the Roman southern provinces, such as the Narbonnaise in France (Ruas et al. 2006). From this period, a morphological diversification of Prunus stones was noticed which is related to, after written sources, an increasing number of fruit varieties (Amigues 2002; André 1981; Raspail 1838; Quellier 2003). In order to identify more precisely the fruit remains found in archaeological excavations and particularly to distinguish varieties among these taxa, various typological analyses based on macroscopic characters of plum and sloe stones have been proposed over time by several authors (Baas 1974; Röder 1940; Rybin 1936; Werneck 1958). In 1978, Behre developed a typological analysis of stones by defining a number of ‘Prunus Formenkreise’ or types during his analysis of the P. domestica L. stones from Haithabu, Germany, 9th–11th centuries A.D. These ‘Prunus Formenkreise’ were defined according to stone size, surface sculpture and shape (finding expression in the indices), and were then related to modern varieties. This method has been applied by various archaeobotanists to stones from numerous European excavations and used notably by van Zeist and Woldring (2000) to characterise P. domestica L. endocarps from the late- and post-medieval occupation deposits in the town centre of Groningen, The Netherlands. Based on physical and morphological characteristics, the authors distinguished 13 different types of plum, some related to recent varieties. Author's personal copy Veget Hist Archaeobot At the same time, Kroll (1978) proposed criteria to distinguish P. avium and P. cerasus stones from the medieval and modern waterlogged material from Lübeck, Germany. The overlapping of quantitative size and shape features did not allow identification of all stones except those with extreme phenotypes. More recently, Kroll (2007) showed that archaeological P. avium and P. cerasus could be easily differentiated on the basis of traditional morphological features such as surface sculpture, dimensions, etc., provided that the morphological features of the fruitstones were observable, despite a long period of deposition. The study was carried out with success for the Haithabu excavation where the early Middle Age material was identified as P. cerasus. For later periods, the author considers that identification of cherry stones becomes more problematic because of the disappearance of the morphological features. Indeed, we have noticed that physical characteristics such as the cherry stone surface sculpture and hilum tend to disappear due to erosion through time and because of the type of deposit, for example acidity in the cesspit. So the efficiency of this method is often limited, particularly in the case of cherry stones, which are often classified under the generic term of P. avium/cerasus (Baas 1951). These taphonomic constraints add to the problems related to interspecific hybridization and cultivation practices which tend to increase the morphological diversity. In addition, the very local aspect of varietal improvement, the probable disappearance of ancient varieties, and the dissimilarity between these ancient varieties and the modern ones, all complicate cherry species identification even more. Difficulties in circumscribing species due to lack of diagnostic characters are also known from the genus Malus where widespread crossability, introgression and cultivation may blur taxonomic boundaries (Dickson et al. 1991). Recently, in their study of Prunus stones from the Roman vicus Tasgetium (Eschenz, Switzerland), Pollmann et al. (2005) obtained promising results by using ancient DNA to answer archaeobotanical issues. Such archaeobotanical research is supported by the development of several regional centres for conservation of ancient cultivars in Europe (Chauvet 1999; Körber-Grohne 1996). According to this scientific context, the present work aims to apply to Prunus avium/cerasus stones an analytical method previously successfully used in the study of Olea europaea L. (olive) stones (Terral et al. 2004; Newton et al. 2006) in order to: – – Evaluate the range of stone shape variation shown by the whole of the bioarchaeological material at our disposal. Distinguish morphotypes on the basis of the model of the 16th century Hôtel-Dieu archaeobotanical material from Tours, France. – – Interpret these results in term of agrobiodiversity by comparison of archaeo-diversity with a collection of modern stones from P. avium, P. cerasus and P. 9 gondouinii varieties. Open new perspectives in the research of the history of the varietal inheritance of the cherry. Materials and methods The cherry fruit is a drupe consisting of an exocarp, a thick and fleshy mesocarp and a hard and woody stone (endocarp) surrounding and protecting a single seed. The stone is a structure with bilateral symmetry consisting of two pseudo-valves deriving from a single carpel, sutured on the margin (Figs. 1, 2). Archaeological stones The material analyzed is based on 717 well-preserved stones originating from Roman and medieval archaeological features dating from the 1st to the 16th century A.D., and from ten French and three Swiss sites (Table 1). Among this collection of cherry stones from various excavations, we have particularly examined material from the cesspit of the 16th century site of the Hôtel-Dieu at Tours, France. It comprises 100 stones of Prunus spp. among many other waterlogged remains of food waste from a hospital. A substantial variability of stone morphologies according to shape, size and hilum structure had already been noticed during the archaeobotanical study, suggesting the existence of two mixed populations, one Fig. 1 Two different morphological types of cherry stones recognized within the Hôtel-Dieu archaeological material, 16th century, Tours, France; scale bar 5 mm 123 Author's personal copy Veget Hist Archaeobot Fig. 2 Analytical processing for quantifying the morphological structure of a cherry stone photographed in ventral view and morphometric (shape and size) attributed possibly to P. avium (the longer ones), the other possibly to P. cerasus or P. avium/cerasus (the shorter and rounder ones) (Ruas, unpublished) (Fig. 1). Comparative reference material To gain insight into the significant identification characters, the present study used reference material composed of 542 stones from various cherry varieties. Thus, we have collected stones from both old and current varieties of P. avium (N = 419), P. cerasus (N = 98) and P. 9 gondouinii (N = 25) obtained from INRA (Institut National de Recherche Agronomique, Bourran, France) (Table 2). P. fruticosa stones were not included in the study as they cannot be easily distinguished from wild types of P. avium (Olden and Nybom 1968). Morphometrical analyses Size analysis corresponds usually to discrete measurements—distances between defined points (landmarks), and 123 angles or ratios—but the overall object shape is not really measured (Rohlf and Marcus 1993). Considering only simple measurements, this approach is not nearly as powerful as it could be, since it does not take into account the geometrical relationships among these data. To solve this limitation, we have combined shape morphometry with traditional measurements, as has been done previously for other fruits such as olive (Terral et al. 2004) and Prunus L. section Prunus (Nielsen and Olrik 2001; Depypere et al. 2007, 2009). Shape morphometrical methods are effective in capturing information about the three-dimensional shape of biological objects and in testing for differences in shapes within and among samples of organisms. These powerful statistical procedures were applied in exploratory studies in taxonomy and evolution (Rohlf and Marcus 1993). In this work, each stone was photographed in ventral view with a digital camera (60 mm f/2.8D lens) placed at the fixed distance of 35 cm from the stones. Photographs of stones were first cut out, and then converted to black-andwhite before being resized (Fig. 2). This procedure was carried out with R software (R Development Core Team Author's personal copy Veget Hist Archaeobot Table 1 Archaeological material analysed Archaeological sites, town (department), context N Dating century A.D. Excavation supervisor Archaeobotanist France Place d’Assas, Nı̂mes (Gard), well 42 1st F. Conche N. Rovira Rue Ste Catherine, Vannes (Morbihan), well 41 1st A. Triste M.-P. Ruas BK14050, Biesheim-Kunheim, Haut-Rhin, cesspit 32 1st–2nd M. Reddé P. Vandorpe BK14064, Biesheim-Kunheim, Haut-Rhin, cesspit 44 1st–2nd M. Reddé P. Vandorpe BK14104, Biesheim-Kunheim, Haut-Rhin, cesspit 31 1st–2nd M. Reddé P. Vandorpe La Roquette, Cavillargues (Gard), well Rue des veaux, Strasbourg (Bas-Rhin), Ill river banks 29 44 4th–5th 10th–12th B. and H. Petitot, S. Alix M. Werlé L. Bouby C. Schaal Charavines, Colletière (Isère), handcraft shops 41 11th E. Verdel, M. Colardelle K. Lundström-Baudais/C. Schaal Charavines, Colletière (Isère), extension area 50 11th E. Verdel, M. Colardelle K. Lundström-Baudais/C. Schaal Charavines, Colletière (Isère), stalling area 45 11th E. Verdel, M. Colardelle K. Lundström-Baudais/C. Schaal Place Métézeau, Dreux (Eure-et-Loir), cesspit 21 12th P. Dupont M.-P. Ruas Rue du rempart Etampes (Essonne), unknown 24 12th X. Peixoto M.-F. Sellami Place Métézeau, Dreux (Eure-et-Loir), cesspit Place de la cathédrale, Tours (Indre-et-Loire), cesspit 24 16th P. Dupont M.-P. Ruas 100 16th A.-M. Jouquand M.-P. Ruas Switzerland Oberwinterthur, (Zürich canton), cesspit OWKW76 23 1st P. Vandorpe Oberwinterthur, (Zürich canton), cesspit OWKW78 35 1st P. Vandorpe Oberwinterthur, (Zürich canton), cesspit OWKW78(1) 25 1st P. Vandorpe Schoffelgasse Zürich, (Zurich canton), cesspit 34 13th M. Kühn Hallwyl Castle, Seengen, (Argovie canton), ditch 32 14th–15th M. Kühn 2005) using the functions proposed by Claude (2008) (R functions are indicated in italic). Geometrical analysis of the stones was carried out following the procedure developed by Terral et al. (2004). In ventral view, the two external half outlines between two homologous landmarks, the base of stone and the apex, are obtained with the Conte() function and were similarly adjusted in an orthonormed basis and standardized by size with the BooksteinM() function (Baseline superimposition) (Bookstein 1991). Then, a least-squared third-degree polynomial curve was fitted to each half outline defined by 20 equally spaced points (Fig. 2). Finally, two third degree polynomial equations described the external geometrical structure of the stone. The eight quantitative parameters (four by half outline) obtained may be easily used as variables in multivariate statistical analyses. In addition, length (mm), thickness (mm) and area (mm2) were measured in ventral view using computerized image analysis systems on digitized photographs (Fig. 2). Statistical analyses In order to evaluate the range of shape diversity in the P. avium/cerasus archaeological material at our disposal for this study, a standardized Principal Component Analysis (PCA) was carried out on 717 stones and eight quantitative parameters presented previously (four by polynomial equations/two equations by stone). Morphological variation in the archaeological material shown by PCA was then analysed by testing correlation between the size features measured and new coordinates of cherry stones in the PCA space. The combination of results from shape and size analyses allowed us to test a possible allometric phenomenon between shape and size which are a priori two independent parameters. Especially for the Hôtel-Dieu archaeological stones, PCA results have been thought of as revealing the internal structure of the data. Finally, a discriminant analysis (DA) was carried out on 542 stones from a non-exhaustive collection of material from modern varieties belonging to P. avium, P. cerasus and P. 9 gondouinii, eight quantitative variables (the eight shape parameters from the polynomial morphometrical analysis) and one qualitative parameter with three distinct modalities, corresponding to the allocation of stones to a Prunus species. This multivariate statistical analysis aimed to test morphological differentiation between fruitstones of modern Prunus species and finally, to provide tangible elements for identifying the Hôtel-Dieu archaeological stones. 123 Author's personal copy Veget Hist Archaeobot Table 2 Prunus species, varietal denomination, origin and number of modern stones analysed as reference material (in brackets, plant material— Prunus species or variety—used as rootstock) Species Cultivated variety N Prunus avium Alex III 10 Argot 13 Black Star (Prunus mahaleb) 19 Brooks (Edabriz) 17 Early Bigi 10 Early Star 18 Europepice 93-17 (MM14) 17 Europepice 94-04 (MM14) 29 Firmred 10 Giant Red (MM14) 28 Grace Star 10 Grace Star (Prunus mahaleb) 21 Lalastar 10 Lodi 20 Masdel Kabel Panaro 1 Sweet Early (Prunus mahaleb) 10 21 Penny 10 Ruby 10 Santina 10 Simcoe Probla Skeena (MM14) Sumcoro 10 20 Tieton 10 V3648 12 V3868 20 Victor Griotte du Nord From different clonal individuals Results The 2-dimensional plot (PCA1-2) explaining 93.2% of the total inertia shows that archaeological stones are mainly distributed along the first factorial axis (PCA1), expressing 75.1% of the total morphological variability. This axis appears to differentiate ovate (x \ 0) from obovate stones 123 7 10 5 Haut-Rhin acide 9 Olivet Hâtive 5 Olivet Tardive 6 Toulennea a 8 Sweet Early Panaro 1 Reine Hortense Prunus 9 gondouinii 9 10 Vanda All the accessions come from INRA, Bordeaux, France 7 Sumbigo Sumele Prunus cerasus 3 Sandra Rose 2 71 Cerise Cure 4 Griotte de Provence 7 Gros Guin de Cœur 7 Impératrice Eugénie 7 (x [ 0). The second axis (PCA2), orthogonal to the PCA1, explains 18.1% of the variability and contributes to distinguish asymmetrical elliptic stones (y [ 0) and rather symmetrical elliptic/oval stones (y \ 0) (Fig. 3). In any case, as no clear structure in morphological diversity was highlighted, we have hypothesized that the first axis expresses shape variation induced by size differences, and Author's personal copy Veget Hist Archaeobot independently the second axis allows discrimination of the stones according mainly to shape features. In order to test this hypothesis, a correlation analysis between each morphometric descriptor was carried out (Table 3). The results show that the existence of correlations among size variables emphasizes the limits of traditional morphometry to characterize stones and to discriminate different statistical populations, or Prunus morphotypes. In addition, size is generally influenced by environmental factors such as climatic and edaphic parameters, or human practices in the case of cultivated plants. In addition, as the correlation between traditional characters and PCA1 is significant, the first PCA axis reflects a pattern of size variations, from small (x \ 0) to large stones (x [ 0). Independently, even if a slight correlation between PCA2 and ‘length’ was noticed, the second axis revealed shape differences (Table 3). Thus, these results underline an allometry phenomenon, thus a relationship between size and shape, however almost non-existent on the second axis of the PCA, which explains mainly shape variations. After evaluation of the overall range of morphological diversity in the archaeological material, geometrical variability of the archaeological stones from the Hôtel-Dieu site (Tours, France) was precisely examined. Coordinates of these stones in the first axis of PCA appeared relatively homogeneous and normally distributed (Shapiro–Wilk normality test: W = 0.99, P-value = 0.87), while those concerning PCA2 were not (Fig. 4a). Their frequency distribution, whose bimodality was tested, implies two significant distinct populations of stones which may correspond to two different shape morphotypes (Fig. 4b). The existence of Table 3 Tests of linear correlation between stone morphometrical descriptors: size measurements (length, thickness and area) and geometrical features (new coordinates in the two first dimensions of PCA); R = Pearson correlation coefficient Tested correlations Length–thickness R 0.34 P-value Significance \0.0001 *** Length–area 0.65 \0.0001 *** PCA1–length 0.35 \0.0001 *** PCA1–thickness \0.0001 *** 0.13 \0.001 *** PCA2–length -0.10 0.007 PCA2–thickness -0.05 0.21 ns 0.06 0.10 ns PCA1–area PCA2–area -0.23 * n.s. not significant * Significant; *** highly significant these two morphotypes attributable to PCA2 was validated by an analysis of variance (ANOVA) carried out on the new coordinates (PCA1 and PCA2) of stones in the morphological space (PCA biplot 1-2) (Table 4). In order to find out the systematic status of these two Prunus forms, we have then performed a discriminant analysis (DA) on the reference material (542 modern stones) presented in Table 2. The overall discriminant power computed by the discriminant analysis is equal to 86.5% in which 93.2% and 77.6% of stones belonging to P. avium and P. cerasus respectively are well-differentiated from their relatives. However, the discrimination rate of P. 9 gondouinii stones is weak, only equal to 8%. Fig. 3 PCA analysis biplot 1-2 showing the overall morphological variation of 717 cherry stones from 13 archaeological sites 123 Author's personal copy Veget Hist Archaeobot Fig. 5 Discriminant analysis biplot 1-2 showing: a a pattern of shape c variation in Prunus spp.; b the location in morphological space, of the Hôtel-Dieu archaeological stones allocated to Prunus avium presents shared or intermediate morphological features as shown by previous studies (Dirlewanger et al. 2009; Faust and Suranyi 1997; Saunier and Claverie 2001). When compared to this preliminary discriminant model as additional samples, the Hôtel-Dieu archaeological stones were assigned to a Prunus species. The probability (P) that archaeological stones belong to a species was calculated using the Mahalanobis distance between stones and each group centroid in the morphological space defined by the discriminant analysis. We have considered the allocation reliable and very accurate if P [ 0.80. In this case, 91 archaeological stones were attributed to P. avium. Moreover, five stones were allocated to P. avium with a probability ranging between 0.7 and 0.8, and four were not classified (P \ 0.7) (Fig. 5). Finally, according to results from the Hôtel-Dieu archaeological material structured in two different morphotypes (Figs. 4, 5), we suggest that these distinct morphotypes correspond to two different cultivated varieties of P. avium. Discussion Fig. 4 a Bimodal distribution of the Hôtel-Dieu stones in relation to PCA2 scores. Results from the normality test (Shapiro–wilk test) are presented for each population/morphotype identified. b The two distinct morphotypes highlighted in the Hôtel-Dieu archaeological material, emphasized within the overall morphological diversity shown by PCA1-2 biplot The DA biplot 1-2 illustrates this pattern. It shows that canonical scores of the first axis, explaining 94.7% of the total variance, contribute highly to discriminate the three Prunus species (Fig. 5). The second axis (5.3% of the total variance) structures the morphological diversity according to shape differences, following most probably an analogous trend shown by the PCA carried out on archaeological stones. Morphological variability of P. 9 gondouinii appears to be significantly higher than that of its ancestors. Compared to its parental relatives, P. 9 gondouinii 123 By comparing stone size, surface sculpture and shape of archaeological P. domestica L. stones with modern varieties, Behre (1978) began morphometric studies on archaeobotanical material. Traditional stone parameters (dimension, ratio and surface sculpture description) were used to characterise archaeological cherry stones. But for more recent historical periods, it appears that these characteristics are not sufficient to differentiate cherry species because of the overlapping ranges of stone size and shape parameters (Kroll 1978), probably resulting from centuries of cultivation and hybridization. Most archaeobotanists consequently expressed reservation about this type of analysis and they continued to use the term of P. avium/cerasus for archaeobotanical fruitstones. In this context, the originality and the innovation of our study is to apply morphometrical methods combining geometric and traditional approaches to assess the overall variation in shape and size of cherry stones, as previously achieved successfully for other fruit species of interest such as Olea europaea L. (Terral et al. 2004) and Vitis vinifera L. (Terral et al. 2010). The co-occurrence of two Prunus avium varieties in Tours is not surprising given the technical improvement of fruit cultivation during the Renaissance. Indeed, in northern France, cherries became highly marketed products. They were intensively cultivated in orchards and not only in Author's personal copy Veget Hist Archaeobot 123 Author's personal copy Veget Hist Archaeobot Table 4 Results of testing the division of the Hôtel-Dieu Prunus stones into two distinct statistical populations (or morphotypes) using analysis of variance (ANOVA) applied to principal components (PCA1 and PCA2) of the PCA carried out on shape descriptors Tested principal component Statistical populations POP1 Mean PCA1 PCA2 0.714 -0.3 Analysis of variance (ANOVA) POP2 Wilk’s Lambda F-test DF SD Mean SD 0.466 0.581 0.592 0.986 1.420 1, 103 0.124 0.062 0.146 0.380 167.990 1, 103 P-value 0.236 n.s. \0.001* POP statistical populations (see Fig. 4), SD Standard deviation at P = 0.05, DF degree of freedom, n.s. not significant * Highly significant gardens of the elite as a result of curiosity (Quellier 2003). In the current state of research, results highlight the fact that the morphometrical method used in this work allows us to rule out the doubt remaining about the characterization of the stones in the cesspit. According to a rental dated from 1587, the Hôtel-Dieu hospital was occupied both by religious personnel of the institution and people affected by various diseases, notably leprosy (Jouquand et al. 1996). The social position of these occupants could be related to the consumption of two different cherry varieties. The mention of a garden in the hospital suggests that these fruits and other plants were probably produced locally (Jouquand et al. 1996). Nevertheless, these results should be interpreted with caution, as the relevance of the use of stones from modern cherry cultivars as reference material for identification of ancient stones from archaeological contexts could be biased by hybridisation and introgression. We cannot exclude that these stones may correspond to P. fruticosa as the morphology of its stones overlaps with wild P. avium ones. Moreover, in the present state of research, it is impossible to identify the origin of these cherries. Were these varieties cultivated in the Tours region? Did the cherries come from more distant places to be finally marketed in Tours? To be more accurate, this study should be completed by the analysis of archaeological stones from different historical periods and different geographical origins. We should also use our method to confirm or re-evaluate the nature of archaeological fruitstones already attributed to a particular species based on traditional parameters. These promising preliminary results open new and interesting perspectives on the assessment of cherry agrobiodiversity at different taxonomic levels (species, subspecies and variety) and on the understanding of its cultivation and consumption history in Europe. Conclusions The morphometric study presented in this work constitutes an innovative archaeobiological contribution intended to characterize past cherry agrobiodiversity in methodological, 123 taxonomical, bio-archaeological and historical perspectives. The combination of geometrical morphometry (baseline superimposition method) and traditional measurements brings clues to define archaeological morphotypes and to link them to current species or forms. In the future, we need to increase our data corpus of the available reference and archaeological material. Study of a larger number of reference stones from wild forms of cherry, diverse varieties and various origins and of archaeological stones from an increasing number of excavation sites would probably provide important results. We particularly need to include archaeological cherry stones from northwestern and eastern European sites and notably ancient P. cerasus stones. We will also improve the methodological approach by integrating analysis of stones performed in different orientations, using powerful methods such as the Elliptic Fourier Transforms (EFT) and 3D analysis. Finally, such a study opens new and interesting perspectives into the understanding of the biogeographical and evolution history of the cherries. In addition, it will be fascinating to reveal if different varieties were consumed by different social categories of people since the Roman period in the provinces, and to discuss the role of these varieties in marking social distinction. Indeed, the economic and social status of cherries has changed in European societies between classical and medieval times. These fruits played an important social role in the medieval elite diet regime (Grieco 1996) before becoming a more common fruit during the later centuries (Quellier 2003). Acknowledgments We wish to thank all the archaeobotanists (Bouby L, Derreumaux M, Hallavant C, Kühn M, Rovira N, Schaal C, Sellami M-F, Vandorpe P, Woldring H, Zech-Matterne V), the excavation supervisors (Demolon P, Dupont P, Jouquant A-M, Plumier J, Réddé M, Triste A, Verdel E and Colardelle M) and all the people who collected and sent us stones from archaeological excavation sites in France, Switzerland and Belgium. We also are grateful to Stéphanie Mariette (UREF—Unité de recherche sur les espèces fruitières, Villenave d’Ornon, France) and to Evelyne Leterme (Domaine de Barolle—Montesquieu, France) for their kind cooperation in the collection of reference stones respectively from the INRA and the CVRA orchards. Finally, we would like to thank Emma Passmore who helped us to improve our English. We also would like to thank the two anonymous referees for their helpful comments and suggestions. This Author's personal copy Veget Hist Archaeobot project was supported by a grant of the CNRS ANR program FRUCTIMEDHIS ‘‘Plant cultivation improvements and new foodstuffs during Antiquity and the Middle Ages: cross-perceptions and readings of the history of fruit trees in the Mediterranean French areas’’. References Amigues S (2002) Études de botanique antique. (Mémoires de l’Académie des Inscriptions et Belles-Lettres 25) de Boccard XV Paris André J (1981) L’alimentation et la cuisine à Rome. Les Belles Lettres. Collection d’Etudes Anciennes, Guillaume Budé, Paris Baas J (1951) Die Obstarten aus der Zeit des Römerkastells Saalburg/ Taunus. Saalburg-Jahrbuch 10:14–28 Baas J (1974) Kultur- und Wildpflanzenreste aus einem römischen Brunnen von Rottweil-Altstadt. FBW 1:373–413 Badenes ML, Parfitt DE (1995) Phylogenetic relationships of cultivated Prunus species from an analysis of chloroplast DNA variation. Theor Appl Genet 90:1,035–1,041 Bakels CC (1991) Western Continental Europe. In: Van Zeist W, Wasylikowa K, Behre K-E (eds) Progress in Old World palaeoethnobotany. Balkema, Rotterdam, pp 279–298 Behre K-E (1978) Formenkreise von Prunus domestica L. von der Wikingerzeit bis in die frühe Neuzeit nach Fruchtsteinen aus Haithabu und Alt-Schleswig. Ber Deutsch Bot Ges 91:161–179 Bertsch K, Bertsch F (1949) Geschichte unserer Kulturpflanze, 2nd edn. Wissenschaftliche Verlagsgesellschaft, Stuttgart Bookstein FL (1991) Morphometric tools for landmark data. Geometry and biology. Cambridge University Press, Cambridge Chauvet M (1999) Le patrimoine fruitier. Hier, aujourd’hui, demain. Actes du colloque de la Ferté Bernard, 15–17 octobre 1998. AFCEV, diffusion INRA Gap Claude J (2008) Morphometrics with R. Springer, Paris Depypere L, Chaerle P, Vander Mijnsbrugge K, Goetghebeur P (2007) Stony endocarp dimension and shape variation in Prunus section Prunus. Ann Bot 100:1,585–1,597 Depypere L, Chaerle P, Breyne P, Vander Mijnsbrugge K, Goetghebeur P (2009) A combined morphometric and AFLP based diversity study challenges the taxonomy of the European members of the complex Prunus L. section Prunus. Plant Syst Evol 279:219–231 Dickson EE, Kresovich S, Weeden NF (1991) Isozymes in North America Malus (Rosaceae): hybridization and species differentiation. Syst Bot 16:363–375 Dirlewanger E, Claverie J, Wünsch A, Iezzoni AF (2007) 3. Cherry. In: Kole C (ed) Fruits and nuts (Genome mapping and molecular breeding in plants 4). Springer, Berlin, pp 103–118 Dirlewanger E, Claverie J, Iezzoni AF, Wünsch A (2009) 14. Sweet and sour cherries: linkage maps, QTL detection and marker assisted selection. In: Folta KM, Gardiner SE (eds) Genetics and genomics of Rosaceae, R. Jorgensen Series Ed. Springer, New York, pp 291–313 Faust M, Suranyi D (1997) Origin and dissemination of cherry. Hort Rev 19:263–317 Grieco A (1996) Alimentation et classes sociales à la fin du Moyen Age et à la Renaissance. In: Flandrin JL, Montanari M (eds) Histoire de l’alimentation. Fayard, Paris, pp 479–490 Hancock AM, Iezzoni AF (1988) Malate dehydrogenase isozyme patterns in seven Prunus species. Hort Sci 23:381–383 Hanelt P (1997) European wild relatives of Prunus fruit crops. Bocconea 7:401–408 Hedrick UP (1915) Cherries of New York. J. B. Lyon, Albany, New York Hillig KW, Iezzoni AF (1988) Multivariate analysis of a sour cherry germplasm collection. J Am Soc Hort Sci 113:928–934 Iezzoni AF, Schmidt H, Albertini A (1990) Cherries (Prunus). In: Moore JN, Ballington JR Jr (eds) Genetic resources of temperate fruit and nut crops, vol 1. I.S.H.S, Wageningen, The Netherlands, pp 111–173 Jouquand A-M, Morin S, Champagne F, Rodier X, Husi P, Wittman A (1996) La fouille des abords de la cathédrale de Tours. Tours, place François Sicard. Document final de synthèse de sauvetage programmé. Orléans, SRA Centre Körber-Grohne U (1996) Pflaumen, Kirschpflaumen, Schlehen. Heutige Pflanzen und ihre Geschichte seit der Frühzeit. Theiss, Stuttgart Krahl KH, Lansari A, Iezzoni AF (1991) Morphological variation within a sour cherry collection. Euphytica 52:47–55 Kroll H (1978) Kirschfunde aus dem 13./14. bis 16. Jahrhundert aus des Lübecker Innenstadt. Ber Deutsch Bot Ges 91:181–185 Kroll H (2007) Die Dornenhecke, der Kirschgarten und der Hafen von Haithabu. Weiteres zum Steinobst aus Haithabu. (Das archäologische Fundmaterial 8, Ber Ausgrabungen Haithabu 36). Wacholtz, Neumünster, pp 313–329 Kroll H, Willerding U (2004) Die Pflanzenfunde von Starigard/ Oldenburg. In: Hafner A, Müller-Wille M (eds) Starigard/Oldenburg. Hauptburg der Slawen in Wagrien, V Naturwissenschaftliche Beiträge. (Offa-Bücher 82). Wacholtz, Neumünster, pp 135–184 Krüsmann G (1978) Manual of cultivated broadleaved trees and shrubs, vol 3. Batsford, London Linnaeus C (1753) Species Plantarum, 1st edn. Holmiae, Stockholm Marshall RE (1954) Cherries and cherry products. Economic crops, vol 5. Interscience, New York Medović A (2004) Zum Ackerbau in der Lausitz vor tausend Jahren. Der Massenfund verkohlten Getreides aus dem slawischen Burgwall unter dem Hof des Barockschlosses von Groß Lübbenau. In: Hafner A, Müller-Wille M (eds) Starigard/Oldenburg. Hauptburg der Slawen in Wagrien, V Naturwissenschaftliche Beiträge. (Offa-Bücher 82). Wacholtz, Neumünster, pp 185–236 Newton C, Terral JF, Ivorra S (2006) The Egyptian olive (Olea europaea subsp. europaea) in the later first millennium BC: origins and history using the morphometric analysis of olive stones. Antiquity 80:405–414 Nielsen J, Olrik DC (2001) A morphometric analysis of Prunus spinosa, P. domestica ssp. insititia, and their putative hybrids in Denmark. Nord J Bot 21:349–363 Olden EJ, Nybon N (1968) On the origin of Prunus cerasus L. Hereditas 59:327–345 Opravil E (1998) Zusammenfassende Übersicht der Ergebnisse von Analysen der Makroreste pflanzlicher Herkunft aus Mikulčice. In: Polaček L (ed) Studien zum Burgwall von Mikulčice 3. Archeologického ústavu AV ČR, Brno, pp 327–356 Out WA (2009) Sowing the seed? Human impact and plant subsistence in Dutch wetlands during the late Mesolithic and early and middle Neolithic (5500–3400 cal B.C.) Diss Universiteit Leiden (Academic Studies Leiden University 18) Pollmann B, Jacomet S, Schlumbaum A (2005) Morphological and genetic studies of waterlogged Prunus species from the Roman vicus Tasgetium (Eschenz, Switzerland). J Arch Sci 32: 1,471–1,480 Quellier F (2003) Des fruits et des hommes. L’arboriculture fruitière en Île-de-France (vers 1600-vers 1800), Rennes, Collection Histoire. Presses Universitaires de Rennes Raspail FV (1838) Cours d’agriculture et d’économie rurale à l’usage des écoles primaries. Paris, Hachette R Development Core Team (2005) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL: http://www. R-project.org 123 Author's personal copy Veget Hist Archaeobot Rehder A (1940) Manual of cultivated trees and shrubs, 2nd edn. Macmillan, New York Rehder A (1947) Manual of cultivated trees and shrubs hardy in North America, exclusive of the subtropical and warmer temperate regions, 2nd edn. Macmillan, New York Röder K (1940) Sortenkundliche Untersuchungen an Prunus domestica L. Kühn-Archiv 54B:1–133 Rohlf FJ, Marcus LF (1993) A revolution in morphometrics. Tree 8:129–132 Ruas MP (1992) The archaeobotanical record of cultivated and collected plants of economic importance from medieval sites in France. In: Pals J-P, Buurman J, van der Veen M (eds) Festschrift for Professor Van Zeist, review of palaeobotany and palynology 73. Elsevier, Amsterdam, pp 301–314 Ruas MP (1996) Eléments pour une histoire de la fructiculture en France au Moyen Age: données archéobotaniques de l’Antiquité au XVIIe siècle. In: Colardelle M (ed) L’Homme et la nature au Moyen Age, actes du Ve Congrès international de la Société d’Archéologie Médiévale (Grenoble 6–9 octobre 1993). Errance, Paris, pp 92–105 Ruas MP, Bouby L, Pradat B (2006) Les restes de fruits dans les dépôts archéologiques du Midi de la France (Ve-XVIe siècle). In : Ruas M-P (ed) Cahier spécial fructiculture, Cultures des fruits et lieux des cultures de l’Antiquité, du Moyen Âge et de l’époque Moderne. Des savoirs en pratique, des mots et des images, actes du séminaire de Toulouse (31 mars-1er avril 2005), Archéologie du Midi Médiéval 23–24:145–193 Rybin WA (1936) Spontane und experimentelle erzeugte Bastarde zwischen Schwarzdorn und Kirschplaume und das Abstammungsproblem der Kulturpflaume. Planta 25:22–58 Santi F, Lemoine M (1990) Genetic markers for Prunus avium L. 2. Clonal identifications and discrimination from P cerasus and P cerasus 9 P avium. Ann Sci For 47:219–227 Saunier R, Claverie J (2001) Le cerisier: evolution de la culture en France et dans le monde. Point sur les variétés, les porte-greffes. Le fruit belge 490:50–62 Schuster M, Schreiber H (2000) Genome investigation in sour cherry, P. cerasus L. Acta Hort 538:375–379 Tavaud M (2002) Diversité génétique du cerisier doux (Prunus avium L.) sur son aire de répartition : Comparaison avec ses espèces 123 apparentées (P. cerasus et P. 9 gondouini) et son compartiment sauvage. Doctoral Thesis, ENSAM INRA, Montpellier (France) Tavaud M, Zanetto A, David JL, Laigret F, Dirlewanger E (2004) Genetic relationship between diploid and allotetraploid cherry species (Prunus avium, Prunus 9 gondouini and Prunus cerasus). Heredity 93:631–638 Terral JF, Alonso N, Buxo R, Chatti N, Fabre L, Fiorentino G, Marinval P, Perez G, Pradat B, Alibert P (2004) Historical Biogeography of olive domestication (Olea europaea L.) as revealed by geometrical morphometry applied to biological and archaeological material. J Biogeogr 31:63–77 Terral JF, Tabard E, Bouby L, Ivorra S, Pastor T, Figueiral I, Picq S, Chevance J-B, Jung C, Fabre L, Tardy C, Compan M, Bacilieri R, Lacombe T, This P (2010) Evolution and history of grapevine (Vitis vinifera) under domestication: new morphometric perspectives to understand seed domestication syndrome and reveal origins of ancient European cultivars. Ann Bot 105:443–455 Thorne RT (1992) Classification and geography of the flowering plants. Bot Rev 58:225–348 Van Zeist W (1991) Economic aspects. In: Van Zeist W, Wasylikowa K, Behre K-E (eds) Progress in Old World palaeoethnobotany. Balkema, Rotterdam, pp 109–130 Van Zeist W, Woldring H (2000) Plum (Prunus domestica L.) varieties in late and post-medieval Groningen: the archaeobotanical evidence. Palaeohistoria 39(40):563–576 Vaquer J, Ruas MP (2009) La grotte de l’Abeurador, FélinesMinervois (Hérault): occupations humaines et environnement du Tardiglaciaire à l’Holocène. In: De Méditerranée et d’ailleurs…, Mélanges offerts à Jean Guilaine. Archives d’Ecologie Préhistorique, Toulouse, pp 761–792 Watkins R (1995) Cherry, plum, peach, apricot and almond Prunus spp. (Rosaceae). In: Smartt J, Simmonds N (eds) Evolution of crop plants. Longman, New York, pp 423–428 Werneck HL (1958) Die Formenkreise der bodenständigen Pflaumen in Oberösterreich. Ihre Bedeutung für die Systematik und Wirtschaft der Gegenwart. Mitteilungen der Höheren Bundeslehr- und Versuchsanstalten für Wein-, Obst- und Gartenbau (Klosterneuburg). Obst und Garten (Serie B) 8:59–82