1981] 
Tepedino & Parker — Flower-Visiting Insects 
323 
and observed distributions were tested for significance using the x 2 
test (Zar 1974). 
The distributions of more abundant species across sunflower 
heads was compared with the balance of the foraging assemblage as 
follows: each individual recorded was assigned to one of two 
mutually exclusive categories, according to whether it foraged alone 
or with at least one other insect (irrespective of species) on the 
inflorescence. A chi-square test of independence was used to com- 
pare each species represented by >10 individuals with the balance of 
the assemblage. 
Results 
Bees were the predominant visitors to sunflowers; we recorded 15 
species in 5 families (Appendix). The species were similar to that 
reported previously by Parker (1981) for the same study site. Onion 
visitors included many species of wasps and flies that did not forage 
on sunflowers. In contrast to sunflowers, there were more non-bee 
than bee visitors to onions. 
For all sunflower censuses the distribution of total insects across 
flower heads did not differ significantly from a Poisson distribution, 
i.e., insects appeared to be foraging independently of other insects. 
The coefficients of dispersion were mostly around 1.0. There was no 
tendency for C.D.’s to be greater or less than 1; for 8 censuses C.D. 
was >1.0 and for 6 censuses C.D. <1.0. (Table 1). 
Only 2 of 7 censuses of onions deviated significantly from a 
random distribution (Table 1). Both deviations occurred on the 
same day and were in the direction of under-dispersion; more heads 
with single visitors were recorded than expected. There was a 
general tendency for insects visitors to be under-dispersed on 
onions; in all tests C.D. > 1.0. 
There was no indication that any particular species foraged other 
than randomly, with respect to other occupants of sunflower heads. 
The results of 34 comparisons of the distribution of individuals of 
abundant species with the balance of foragers for the single and 
joint foraging categories are shown in Table 2. Only one comparison 
yielded significant results; another closely approached significance 
(7/31 Peredovik, AM, Halictus ligatus , P. = 0.051). It is likely that 
these two instances were due to chance. 
