Constancy
Of the 150 pollen balls we analyzed, 21 out of 150 bees (14%) were caught on Clarkia flowers that were different from the majority of the Clarkia pollen found in their pollen balls, indicating at least some amount of pollinator inconstancy (Figure 3). The contents of the pollen balls, however, indicate that pollinators were often constant. This includes a caveat: the final measurement of RRA with no cutoff returned all four species of Clarkia in 100% of the samples, indicating the unlikely result that all pollinators are not only inconstant, but visited all four Clarkia species - even when collected in communities containing fewer than four Clarkiaspecies. Not only is this result unlikely based on the biology of the system, but it is exceedingly rare that any quantitative analysis with relative read abundance would use raw read count in the analysis. As such, the rest of our results will compare RRA with 5% (RRA5) and 10% (RRA10) cutoffs with qAMPseq.
In the large majority of samples, pollen balls contained only one species of Clarkia pollen. Estimates of single-species pollen balls varied among methods, with 66% (RRA5), 74% (RRA10), or 76% (qAMPseq) of samples containing only one Clarkia species. This indicated a striking level of constancy, emphasized by the fact that 70% of bees we sampled were captured in multi-species Clarkiacommunities. Furthermore, the Pearson’s Chi-squared test comparing the number of bees from communities with one to four floweringClarkia species versus the number of pollen samples with one to four flowering Clarkia species was significant (X2 (3)=77.05, p<0.001), confirming that even in diverse Clarkia communities, bees were constant (Figure 4, Panel B; Table 2).
Despite overall pollinator constancy, a quarter (qAMPseq, 24%; RRA10, 26%) to a third (RRA5, 33%) of the bees were inconstant, often carrying two species of Clarkia pollen. The most common multi-species combination in pollen balls was C. cylindrica andC. unguiculata (Figure 3). No pollen balls contained all four species of Clarkia . Bees carrying C. speciosa pollen - the most behaviorally specialized of the Clarkia (James 2020) - tended to only carry C. speciosa pollen: when present, C. speciosa was the only species of pollen in the pollen ball in 12/14 cases. Bees carrying the other three species of Clarkia often carried mixtures of the three (Figure 3).
Relative abundance measurements of Clarkia in pollen balls were largely the same between qAMPseq, RRA5, and RRA10, and most similar between qAMPseq and RRA10 measurement methods (Table 2 and Figure 4; Panel B).