Sampling, preparation and treatment of modern reference material
A total of 53 modern wood samples were analyzed. The modern reference samples were collected in 2014 during the annual archaeological field mission, from 36 individuals (Table S1). For some trees, two wood samples of different diameters were collected in order to take into account anatomical variability within individual.
The collected individuals showed different conditions of growth described in the introduction section and detailed in the Table 1. With the agreement of the Tifigit inhabitants and local authorities, wood sampling was achieved but samples were not collected from trunks, to avoid injuring trees of major symbolic, ecological and economic importance. Only section samples with perfect axial symmetry were retained to avoid any impact of biomechanical constraints (formation of reaction wood) on wood characters.
Once collected, the samples were air-dried during a month at the laboratory. The samples were separately wrapped in tin foil and buried in the sand and then charred without oxygen, at 450 °C for 15 to 20 min depending on the size of the sample. As a result, samples were enriched in carbon (content > 90%)20,26, reached their maximal shrinkage27, and thus are considered to become morphologically comparable to charcoal produced in medieval fires27,28,29,30,31. The minimum and the maximum diameter of wood samples were measured (mm) using a digital measuring calliper before and after carbonization. The diameter used in the following analyses is the mean of the two measurements carried out before carbonization.
Archaeological material
Twenty archaeological charcoal fragments of Argan tree identified during a previous analysis session13 were included in this study (Table S2). All the Argan charcoal fragments were collected in the medieval archaeological deposits of Îgîlîz13. They come from various contexts, for the most part from living units, and belong to the period of highest human activity at the site, between the late 11th and early thirteenth centuries.
Quantitative eco-anatomical analysis of wood applied to the Argan tree
The approach consists in measuring constitutive elements of wood and aims to understand variations according to intrinsic (inferred by the branch diameter mainly age of tree18, linked to the existence of growth rings that are often difficult to distinguish) and environmental parameters affecting the cambial activity and thus, rate of growth and wood development20,28,29,30. This high resolution analysis of wood structure, particularly of conductive elements, allows addressing numerous questions that have been successfully solved in the case of the olive tree and the grapevine, such as phenology, ecology, climate, impact of human activities and agricultural practices20,24,25,31,32,33.
Argania spinosa wood is diffuse-porous with a dendritic and diagonal arrangement of vessel elements in transversal section34. The axial parenchyma bands are in tangential alignment and composed of multicellular strands. In radial alignment, woody rays are 1–3 cells wide and of heterocellular composition (Fig. 6).
Wood anatomical features of the Argan tree (in blue) and measured anatomical characters (in red).
To apply a quantitative eco-anatomy approach to the Argan tree, both modern charred samples and archaeological charcoal are broken manually in transverse anatomical section. The following wood constitutive elements and anatomical characters related to sap conduction and reserve storage are observed and measured under a reflected-light microscope connected with an image analysis system (DFC300 FX Leica camera and LAS Leica software) (Fig. 6): (1) vessel density (DVS—number of vessels / mm2), (2) vessel surface area (SVS, µm2), (3) ray density (DRA—number of rays / mm2), (4) axial parenchyma density (DPA, number of bands / mm2), (5) Density of wood fenestrated zones bordered on one side by the radial alignment of axial parenchyma cells and on the other side, tangentially, by rays (DWF—number of fenestrated zones / mm2).
These anatomical features were measured several times (see ‘Statistical analyses’ section) following radial lines from the cambium inwards the sample and crossing a small number of growth rings (i.e. functional rings from a sap conduction point of view). Moreover, the hydraulic conductivity or vascular conductivity (CD) was assessed using the following formula: CD = (SVS/π)2/DVS (after32,35,36,37). Finally, the ratio ‘Conductive surface / total wood area’ (SC) was calculated.
Statistical analyses
In order to determine the number of measurements required for an optimal assessment of anatomical features, a rarefaction method was carried out from the analysis of test wood samples. For each one, repeated measurements of anatomical characters (Surface vessel area (SVS), Density of vessels (DVS), Ray density (DRA), Axial parenchyma density (DPA) and Density of wood fenestrated zones (DWF)) were performed following the aforementioned method and the cumulative mean value was then calculated for each character20,29. For each test sample and anatomical character, the number of measurements of a character required for an optimal assessment was quantified as the minimum number of measurements required to stabilize the mean value (rarefaction curve or cumulative mean curve).
Furthermore, different measurement sessions were carried out with the aim of testing possible errors and reproducibility of measurements taken by one or various observers, respectively. The data sets produced were tested using the PCA performed to evaluate the Argan anatomical variability. The ARG8-2 sample was used as test sample. In addition to the initial measurements. The ARG8-2 sample was analyzed 4 times: twice by one operator (ARG8-2 (1-OP1) and ARG8-2 (2-OP1)) and twice by another (ARG8-2 (3-OP2) and ARG8-2 (3-OP2)) at different times. The additional data were incorporated into the PCA as additional individuals for comparison with initial anatomical features of ARG8-2.
After showing that measurement errors have no impact on the validity of results and the measurements are reproducible, quantitative eco-anatomical data were processed using a principal component analysis (PCA) in order to evaluate anatomical plasticity in the reference modern material, to appreciate relationships between characters and wood sample caliber and to confront archaeological data to the reference model. PCA was applied on 53 reference modern samples and 7 quantitative variables (anatomical characters) to: (1) validate the hypothesis that there is a significant relationship between the size of the branch and anatomy, as previously demonstrated by analyses of wood development and structure18,20,38 and dendrochronology39; (2) identify the anatomical characters most affected by the age of the branch and, in that case, model the ‘anatomical characters—caliber of the branch’ relationship; (3) develop predictive model that might estimate the minimum branch caliber from eco-anatomical data of archaeological charcoal.
Finally, data from analysis of the 20 archaeological charcoal fragments were included in PCA as additional statistical samples. They do not contribute to the development of the reference model, but are compared to the modern reference samples in order to infer the ecological conditions under which Argan trees grew during the Middle Ages.
Source: Ecology - nature.com
