Table 3.

Description of behavioural and morphological effect traits

Effect trait no.	Name	Description	Correlation (t-test) with pollen stigmal deposition
1	Body length	Body length is related to both inter tegular distance and body mass69,70 and is inter-correlated with a wide range of functional characteristics40, including foraging range in bees41.	t7 = 4.78** (r = 0.85)
2	Mean time on flower	The mean amount of time (seconds) spent foraging on an oilseed rape floret. Data from Woodcock et al.18 but augmented with unpublished data.	t7 = −1.13 NS
3–4	Nectar or pollen foraging	The probability during a foraging event the pollinator will forage for nectar (trait 3) or pollen (trait 4). Data from Woodcock et al.18, but augmented with unpublished data.	t7 = −0.71 NS t7 = −1.31 NS
5	Stigmal contact	The probability that stigmal contact will be made when foraging. Data from Woodcock et al.18, but augmented with unpublished data.	t7 = 2.61* (r = 0.70)
6	Dry pollen on body	The probability of presence of free dry pollen anywhere on the individual. Data from Woodcock et al.18 but augmented with unpublished data.	t11 = −1.31 NS
7	Hairiness index	Hairiness affects pollen grain deposition on stigmas37 and, in bees, is used to detect electromagnetic fields emitted by flowers as pollination cues38. For each species, body parts that contact oilseed stigmas (head, thorax, sternum, abdomen underside, femora, tibiae and meta-tarsus (legs assessed separately)) were scored as: (1) coarse setae or extremely short hairs; (2) short (c. basal tibiae 1 diameter) but dense hairs (>50 mm2); (3) long (>basal tibiae 1 diameter) dense (>50 mm2) hairs. This score was summed and given as a percentage of the maximum score of 24.	t7 = 2.44* (r = 0.67)
8	Mouthpart type	The length of the tongue used to collect nectar affects host plant specialisation, and is defined as either long, medium or short58. A separate category is listed for insects with chewing mouthparts.	NA
9–13	Specific pollen collecting structures.	The presence of setae specifically used to collect pollen, listed by Michener57 as the basitarsal scopa (trait 9), femoral corbicula (trait 10), strict tibial corbicula (trait 11), propodeal corbicula (trait 12) or abdominal corbicula (trait 13). Note these structures are associated with bees, however, their absence will affect the pollen carrying capacity and thus likelihood of pollen stigmal transfer of other species (e.g., for flies).	NA
14	Pollen carried only in the crop	Pollen carried only in the crop and, as such, not available for pollination57. As above, these structures are associated with bees, however, their absence may affect the likelihood of pollen contacting plant stigmas for other pollinating groups.	NA
15	Corbicula pollen moist	Pollen in corbicula storage structures may be either dry or moistened. Moistened pollen is less freely available for deposition onto plant stigmas57. As above, these structures are associated with bees, however, their absence may affect the likelihood of pollen contacting plant stigmas for other pollinating groups.	NA

These were derived for each pollinator species or functional type of bee (N = 44), other Hymenoptera (N = 1), butterflies (N = 1) and flies (N = 11). To confirm the importance of these traits as predictors of pollination success (and so identify effect traits for assessment of the mass ratio hypothesis) they were correlated with a small sub-set of species where pollen stigmal deposition rates had been quantified^20,68 (Supplementary Methods). The significance of these correlations is shown. For some effect traits there was insufficient range in the trait characteristic to provide a correlation (indicated by NA)