Optimization of a method of pollen counting adapted to honeybee diet

Honeybee diet surveys need accurate knowledge of pollen supplies. Our studies concern the flora contribution in different landscape environments based on pollen loads collected in our experimental apiaries throughout the year, and require a large number of time consuming microscopic analyses. Knowledge of bee pollen diet focuses first on identifying the major resources that are related to land use, especially in cropping systems. Rare species are not necessarily sought in this process. While referring to previous studies, we aimed to reduce the counted pollen grains in order to reduce the time spent on each sample. For this, counts on whole transects were compared with alternative methods of counting by sub-sampling, using the Louveaux palynological method (1978). For each pellet sample, a drop from a homogenized suspension was dried, degreased and mounted onto a slide in gelatinized glycerine. The counting was achieved using a digital camera fitted to a microscope. The study was carried out in two phases. Phase 1 : Test of slides homogeneity, carried out on a 10-slide set with one replication. For each slide we compared the exhaustive count of each taxa present in the whole transect to those present in each quarter. Phase 2 : Search for an alternative counting method. Work was carried out on the first 6 samples with replications studied in the previous phase retaining exactly the same transects but counting only one microscopic field in three, then four, and finally one in five. Data analysis was carried out using the Chi-square test on species > 5%. Phase1 showed a good homogeneity of the slides in particular for the 2nd and 3rd quarters of the transect. In phase2, the comparison of alternative counting methods to the full transect method verified the linearity of the counting results. Suitable results were given by the reading mode "one in three". The accuracy of this method in particular for samples containing big pollen grains, e.g. Zea mais, distributed differently on the slide was improved. This task consisted of a general review of their number along the transect in order to establish the exact proportion. In conclusion we propose a method of palynological analysis which is adapted to the predominant species visited by honeybees for their pollen diet.