1976] Waddington — Foraging Patterns of Halictid Bees 
113 
are employed to determine the foraging patterns of sweat bees 
(Halictidae) gathering pollen and nectar from bindweed (Convolvu- 
lus arvensis) flowers. 
Methods 
All observations were made at dense floral arrays of bindweed 
(Convolvulus arvensis) on Campus West, The University of Kansas, 
Lawrence. Several acres of bindweed surrounded the test areas. 
Observations were made of the three prime bee visitors to the bind- 
weed flowers: Agapostemon texanus, Augochlorella striata, and 
Lasioglossum (Dialictus) sp., all sweat bees of the family Halictidae. 
The species of Lasioglossum were difficult to discriminate in the 
field, so it is possible that data for more than one species of Lasio- 
glossum were pooled under the designation Lasioglossum sp. 
Honeybees, bumblebees, butterflies, and moths were also observed 
visiting bindweed flowers for food. 
The following methods were used to determine the temporal pat- 
terns of the bees within the bindweed patch. Eleven test quadrats, 
each 1m 2 (1 X lm), were delimited by string in the bindweed popu- 
lation. The number of halictids (individual species were not dis- 
tinguished) on flowers within quadrats 1-3 on June 30, 1974, and 
quadrats 4-7 on July 2, was recorded every two minutes from about 
0830 to 1030 h. Counts of each of the three halictid species were 
made every 3 minutes at quadrats 8-11 on July 3, again between 
0830 and 1030 h. At the end of each day’s observations the number 
of flowers within each quadrat was recorded. 
The spatial patterns of halictid bees were tested employing the 
following methods. A 12 X 6m rectangular study area, near the 
edge of several acres of bindweed, was delimited by string. A ran- 
dom numbers table (Rohlf and Sokal, 1969) was used to determine 
the coordinates of the southwest corner of 32 0.25m 2 (0.5 X 0.5m) 
sample quadrats within the study area. Each of the quadrats was 
delimited with string in late afternoon on July 4. At 0915, 0955, and 
1045 h on July 5 the number of bees within each of the 32 quadrats 
was recorded. 
The chi-square test of goodness of fit was employed to test the null 
hypothesis that the temporal and spatial data approximate Poisson 
distributions, the expectation if the bees are dispersed at random. 
The G-test was applied to 2 * 2 tables to test whether the presence 
and absence of species pairs at bindweed are dependent on each other. 
