Abstract/Summary

Grasses are abundant in the British Isles, and their pollen contributes significantly to exposure to airborne allergens. This thesis addresses issues relating to the production of pollen, the phenology of pollen production and the changing phenology of selected British grasses. Published pollen dimension data for some common grasses was reviewed and pollen grains were measured and the dimensions found to be comparable. Correlation between anther length and pollen size was significant. Better understanding of the relationship between pollen production and grain size might contribute to estimates of airborne pollen loads. In order to make new estimates of pollen production, a counting approach using automated image-based methods was established. Counts were made per anther for 10 grass species found on the University of Reading Whiteknights Campus, and spikelet and floret counts were used to estimate pollen production per inflorescence for these species. The counts generated here were compared to the ranges of published counts and within-species variation was highlighted. Phenological characteristics of 9 common grass species found on Whiteknights Campus are presented following weekly recordings of phenological stages according to a modified BBCH General Scale at 10 sampling sites throughout spring and summer of 2016. Phenological data indicating the timing of anthesis alongside estimates of abundance of the species and the pollen production data generated for this study were used to make a pollen exposure appraisal. The possible value of these exposure estimates is considered. Increasingly, herbarium specimens are contributing insights into phenological change through time and providing a better understanding of the impact of climate on flowering times. Flowering time data was collected from 898 specimens representing 15 common grasses from the British Isles. A method estimating and applying an adjusted first flowering date was applied, and relationships between climate parameters and these adjusted first flowering dates were sought using a General Linear Model. First Flowering Date was significantly affected in 9 of the study species.