Grain and annulus diameter as criteria for distinguishing pollen grains of cereals from wild grasses

The main criterion for distinguishing morphologically between pollen from wild grasses and that of cultivated grasses is biometric. In Western Europe, the biometric threshold generally used is 45 μm for grain diameter and 8 μm for annulus diameter, a threshold above which the pollen is attributed to cereals. However, for the seaboard, these criteria are not satisfactory because indigenous grass species have large sized pollen grains (> 45 μm and > 8 μm). When we realize palynological studies to answer questions about anthropisation or early beginnings of agricultural activities, it is important to clearly differentiate pollen grains of littoral wild grasses from cultivated grasses. In the present study, we propose a new biostatistical approach to advance the discrimination of the two types mentioned above. In this matter, we used 35 species of wild grasses from the Armorican littoral (France) and 4 species of Cerealia type grasses. For each sample, diameter and annulus sizes of 300 grains were measured to calculate the mean values. Technics of observations and measurements were identical to the ones utilized to analyse fossil samples. Statistical analysis allowed us to show a biometric threshold coinciding to a combination of two values of diameter: 47 μm for grain diameter and 11 μm for annulus diameter. The pertinence of this pair of threshold values was tested a posteriori using assessment criteria such as specificity, sensitivity, positive and negative predictive values. The threshold 47-11 μm appeared to be a discriminant pair, placing a minimum of wild grasses into the Cerealia type group while retaining the maximum of cereals in this same group. Furthermore, this biostatistical approach could be applied more generally on different corpus of grasses.