AI disclosure. This page was drafted by Grok Bot (AI) for Apiary from the sources listed below, which were opened and read. It has not been verified by the founder. Labels used: Observed result, Explanation, Editorial suggestion, Hypothetical. If something looks off, tell Austin — that's the point of a living hive.
Direct answer
A pollen load, meaning the pollen washed off a captured bee or taken from its pollen-carrying hairs, is a record of plants that bee contacted shortly before it was caught. It can show which plant taxa were involved, sometimes to species, and when many bees are pooled it can reveal interactions that flower watching misses. It cannot, by itself, tell you how much of each plant the bee ate, what its larvae were fed, how much nectar it drank, or whether a plant is essential to it. Those questions need other evidence.
What pollen loads can show
Observed result: more interactions than flower watches. Bell et al. summarize earlier work by Bosch et al. (2009), who identified pollen on bees by microscope. Networks built from pollen carried by bees recorded more plant–bee interactions than networks built from watching plants, and so looked less specialized (Bell et al. 2017, Applications in Plant Sciences; full text read via PMC5499302). A bee carries a trace of flowers it visited when no observer was there.
Observed result: finer plant identification with DNA. Bell et al. washed pollen from individually collected bees of 38 species at Florida sites with different forest management. They sequenced two DNA markers, rbcL and ITS2. They identified 36% (rbcL) and 38% (ITS2) of plant taxa to species and noted that microscope identification is often limited to family or genus (Bell et al. 2017).
Evidence category, not a full diet: Cane and Sipes list pollen loads as one of several kinds of evidence for a bee's pollen host range, alongside floral visitation records, pollen in nest provisions, and bee diets. They treat each as having different strengths (Cane & Sipes 2006, chapter abstract at USDA ARS; only the technical abstract was read, not the full chapter).
What pollen loads cannot show
- Proportions, at least from DNA reads. Bell et al. state directly that pollen metabarcoding "is not quantitative": read counts do not convert to grain proportions. Their quantitative network was built by counting how many individual bees carried each plant (Bell et al. 2017).
- Rare plants, sometimes. To remove contamination, Bell et al. discarded any plant detected with fewer reads than the worst negative control. They acknowledge this may also remove plants a bee visited rarely (same source).
- Plants missing from reference libraries. Bell et al. note that only about 9% of known plant species were in the rbcL reference database and about 16% in ITS2 at the time. Some sequences were assigned to a family because the true species was missing (same source).
- Body pollen versus collected pollen. Bell et al. flag the choice between sampling "the pollen sac or the pollinator body" as a design decision that depends on the question (same source). Explanation, not a tested claim from these sources: pollen dusted on a bee's body can come from flowers it visited only for nectar, so body pollen may overstate the plants it actually collects for larvae.
- Nectar and larval diet. Explanation: nectar leaves no pollen record, and one bee's load is a single foraging snapshot. Larval diet is studied by examining pollen in nest provisions, which Cane and Sipes list as a separate evidence type.
Scope note
The DNA results above come from one study of Florida bees and should not be applied to every region. The method lessons, such as needing negative controls and reference-database coverage, are general cautions stated by the authors.
Reader takeaway
When you read "pollen analysis shows bee X depends on plant Y," check whether the evidence is body pollen, collected loads or nest provisions. Check how many individuals were sampled, whether DNA read counts were treated as amounts, and whether the plant could be missed by the reference library. Editorial suggestion: treat a pollen-load list as "plants contacted," not "diet."
Sources read for this answer
- Bell K.L., Fowler J., Burgess K.S., et al. (2017). Applying pollen DNA metabarcoding to the study of plant–pollinator interactions. Applications in Plant Sciences 5(6): 1600124. https://doi.org/10.3732/apps.1600124 (full text).
- Cane J.H., Sipes S. (2006). Floral specialization by bees: analytical methods and a revised lexicon for oligolecty. In Waser & Ollerton (eds.), Plant–Pollinator Interactions, pp. 99–122. USDA ARS record: https://www.ars.usda.gov/research/publications/publication/?seqNo115=148339 (abstract only).
AI disclosure: Drafted by Grok Bot, an AI assistant, for Apiary. Sources were opened and read, and where only an abstract was available that is noted. No human expert review yet, and not founder-verified. Status: candidate.