Pollen and Pollinators
With this study, we have contributed to the growing body of literature
that uses molecular methods to understand pollen use by bees (Mitchell
et al 2009, Galimberti 2014, Bell et al. 2017, Galliot et al 2017, Lucas
et al. 2018, Bell et al. 2019), and in the process provided some
evidence for heterospecific pollen transfer in flowering plant
communities (Mitchell et al. 2009, Arceo-Gomez et al. 2019a, b). That
said, pollen balls in our study were typically comprised of only one
species of Clarkia , which indicates that heterospecific pollen
transfer may not be as common in the Clarkia system as in other
flowering plant systems. The bees that did carry heterospecific pollen,Hesperapis regularis and Lasioglossum sp. , were the two
bee taxa commonly shared between C. cylindrica, C. unguiculataand C. xantiana , which suggests that pollinator inconstancy
introduces heterospecific pollen between these three species ofClarkia (Figures 3 and 4, panel B).
There were two important results revealed by our pollen analysis that
would not have been available using metabarcoding or visitation
observations alone. First, though C. xantiana was preferred by
pollinators, C. cylindrica was carried with greater total
representation in Hesperapis and Lasioglossum(Dialictus) sp. pollen balls (Figure 5). This result is likely
due to the number of C. cylindrica in Clarkia communities-C. cylindrica has the highest average floral abundance of all the
species (Figure S3). Therefore, while its pollen was found in high
abundance on many bees (as in Figure 5), it is not carried in proportion
to its relative abundance in Clarkia communities (shown in Figure
4). Consequently, C. cylindrica is avoided relative to its
abundance, suggesting the pollinator behaviors in this system might
limit its competitive dominance. Second, we were able to resolve
differences in preference between Lasioglossum taxa. Because we
collected and sacrificed the bee specimens in this study, we could
identify Lasioglossum with higher resolution than observing
visitation without sampling, and show that the taxa exhibit differences
in their relationships to Clarkia : the putative specialist on theClarkia genus, Lasioglossum pullilabre (Moeller 2004,
Eckhart et al. 2006), carries C. cylindrica with higher
frequency, whereas the likely generalist, L. (Dialictus) sp. ,
carries C. xantiana with higher frequency (Figure 5).
With this study, we were also able to better delineate theClarkia use by rare pollinators in our dataset. The rare
pollinators, Apis mellifera , Bombus sp. , andMegachile sp. (rare to the dataset, but not rare in the
ecosystem; Moeller 2005; Eckhart 2006, Singh 2013) all carried more
species of Clarkia pollen in their pollen balls than they had
been observed visiting. The specialization of the pollinatorClarkia visitation network was higher than that of the pollen-use
network, indicating that pollinators use more diverse resources than the
plants we caught them on. Given that sampling effort is a perennial
issue in network analyses, we suggest that pollen networks like this one
and others (for example: Alarcón 2010, Galliot et al. 2017, Lucas et al.
2018) are a potential means to understand plant-pollinator relationships
when sampling effort is constrained. We say this with caution, however:
pollen analysis data should complement, not supplant, well-designed
sampling methods. For example, it is highly likely that Apis
mellifera and Bombus sp. use all four Clarkia pollens,
rather than just two per species (Singh 2013), but our sample size of a
few bees per species makes that impossible to say with certainty.
Other patterns in pollen use were similar to what we expected based on
previous observations of pollinator behavior. The two Clarkiaspecies most often found in multi-species pollen balls, C.
cylindrica and C. unguiculata , have been shown to occur together
with higher frequency than any other Clarkia species pair in this
system, and exhibit pollinator-mediated character displacement in floral
traits (Eisen and Geber 2018). Given the frequency with which they occur
together in pollen balls, it’s possible that character displacement in
the floral traits of these two species could be driven by the
competitive effects of heterospecific pollen transfer. Furthermore,Diadasia angusticeps bees carried pollen balls of single-species
composition (C. speciosa ), which corresponds to previous
observations that the species is behaviorally more specialized onC. speciosa (Singh 2013, James 2020).