This is an unformatted, plain-text file of the article starting on page 398 of the September 1996 issue of Florida Entomologist (vol. 79, no. 3). It is put on WWW only to facilitate automated indexing and retrieval. Tables are scrambled and illustrations are omitted.
View, read, or make a reprint of this item in its original format, including tables and illustrations.
Go to homepage of Florida Entomologist on WWW for more information about this service.
TEMPORAL AND SPATIAL VARIATION IN THE FORAGING BEHAVIOR
OF HONEY BEES (HYMENOPTERA: APIDAE) AT CHINESE VIOLETS
Ethel M. Villalobos1 and Todd E. Shelly2
1Department of Biology
2Hawaiian Evolutionary Biology Program
Abstract
This study describes temporal and spatial variation
in the foraging behavior of honey bees, Apis mellifera L., at
the Chinese violet, Asystasia gangetica (L.) T. Anderson, in Hawaii.
Workers made 3 types of visits: “out” (O-) visits involving
nectar robbing through corollar slits made by carpenter bees (Xylocopa
sonorina (Smith), “in-upright” (IR-) visits involving
upright entry along the base of the corolla, and “in-upside
down” (ID-) visits involving spiral, upside-down entry along
the top of the corolla. In general, individual workers displayed
only 1 tactic over 10 successive flower visits and over successive
days. Nectar-robbing workers visited more flowers per min but
spent less time per flower than workers making IR- or ID-visits.
Bees making O- or IR-visits carried similar nectar loads but only
very small amounts of pollen, whereas the reverse was true for
bees making ID-visits. O- and IR-visits were made throughout the
day, but ID-visits were observed only in the morning. Based on
inter-site comparisons, the incidence of nectar robbing was influenced
by the local density of carpenter bees which made the perforations
used by nectar robbing honey bee workers.
Key Words: Apis mellifera, nectar robbing, pollination,
Hawaii
Resumen
Este estudio describe la variación temporal
y espacial del comportamiento forrajero de la abeja de miel, Apis
mellifera L., en la violeta china, Asystasia gangetica (L.) T.
Anderson, en Hawaii. Las obreras hicieron 3 tipos de visitas:
Las visitas “por fuera” (F-) incluyeron el robo de néctar
a través de las incisiones hechas previamente a la corola
por las abejas carpinteras, Xylocopa sonorina (Smith), las visistas
“dentro-por arriba” (DA-) incluyeron entradas por encima
de la flor, a lo largo de la base de la corola, y las visitas
“dentro-volteadas” (DV-) incluyeron entradas en espirtal,
volteadas a lo largo del extremo de la corola. En general, las
obreras individuales desarrollaron una sola táctica cada
10 visitas sucesivas a las flores, en 10 días sucesivos.
Las obreras ladronas de néctar visitaron más flores
por minuto pero pasaron menos tiempo por flor que las obreras
que efectuaban visitas de los tipos DA- y DV-. Las abejas haciendo
visitas de los tipos F- y DA- llevaron cargas de néctar
similares pero sólo pequeñas cantidades de polen,
mientras que lo inverso ocurrió en abejas haciendo visitas
DV-. Las visitas F- y DA- fueron hechas a lo largo de todo el
día, pero las visitas DV- fueron observadas solamente en
la mañana. Basada en comparaciones inter-sitios, la incidencia
del robo de néctar fue influenciada por la densidad local
de las abejas carpinteras, las cuales hacen las perforaciones
usadas por las obreras ladronas de la abeja de miel.
Research on the foraging biology of the honey bee,
Apis mellifera L., has focused primarily on colony performance,
and comparatively little attention has been given to the food-collecting
behavior of individual workers (Schmid-Hempel 1991). Available
data reveal that workers visiting the same plant species typically
specialize on either nectar or pollen collection. In some instances,
this distinction is absolute, and individual workers gather either
nectar or pollen exclusively (McGregor et al. 1959; Free 1960a,b).
Often, however, workers may collect both nectar and pollen, although
they may concentrate primarily on one of these resources. For
example, among honey bees visiting native cotton, Gossypium thurberi,
40% collected pollen only and 60% concentrated on nectar collection
but gathered pollen passively as well (Buchmann & Shipman
1990; see also Fell 1986; Weaver 1956).
Nectar foragers may also display variable foraging
modes, either obtaining nectar “legitimately” (thereby
serving as pollinators) or via nectar theft or robbing (thus avoiding
contact with pollen-bearing anthers, sensu Inouye 1980). Though
reports of nectar theft and robbing are common for honey bees
(Burrill 1925; Hawkins 1961; Helms 1970; Benedek et al. 1973;
Barrows 1980), to our knowledge only 1 study presented information
on worker specialization regarding legitimate vs. illegitimate
nectar collection. In observing honey bees visit hairy vetch (Vicia
villosa Roth.), Weaver (1956) found that workers reached the nectary
by either entering the flower or inserting their tongue through
the petals at the base of the corolla. Observations of individual
workers over 2-10 consecutive floral visits revealed a high degree
of specialization for 1 tactic or the other, and only 10 of the
300 individuals observed used a mixture of foraging tactics.
The present study describes temporal and spatial
variation in the foraging behavior of honey bees at the Chinese
violet, Asystasia gangetica (L.) T. Anderson (Acanthaceae), in
Hawaii. Earlier observations (Barrows 1980; Gerling 1982) recorded
Xylocopa sonorina (Smith) as a primary nectar robber (actually
perforating flowers) and A. mellifera as a secondary nectar robber
(using existing perforations) of this plant species (sensu Inouye
1980). Here, we describe diurnal variation in the incidence of
3 types of floral visits-2 involving corollar entry and 1 involving
nectar robbing-and compare body sizes, nectar and pollen loads,
rates of floral visitation, and floral handling times among honey
bee workers displaying the different foraging tactics. Workers
were also individually marked to assess between-day consistency
in foraging behavior. Finally, we compared the incidence of nectar
robbing at several sites differing in honey bee, carpenter bee,
and floral densities.
Materials and Methods
Study Site and Species
A. gangetica, a procumbent or scandent perennial
herb, was introduced to Hawaii around 1925 and is now widespread
in disturbed lowland areas (Wagner et al. 1990). Flowers are pentamerous,
zygomorphic, and vary in color from pale blue to purple (or less
commonly, white or yellowish). Corollas vary from approximately
25-35 mm in length (Barrows 1980). Both A. mellifera and X. sonorina
were introduced to Hawaii in the late 1800s (Lieftinck 1956; Barrows
1980).
Most of the field work was conducted during May-June
1993 in a small untended lot (100 m2) on the campus of Chaminade
University, Honolulu, Hawaii. All observations were made under
sunny or only slightly overcast skies, and air temperatures ranged
between 25-31°C at 15 cm above ground. A. gangetica was the
most abundant flower within the study area. Patches of A. gangetica
covered approximately 1/2 the lot, and the rest of the area was
covered by various grasses or bare rock surfaces. A. mellifera
and X. sonorina were the only species of bees observed at the
study site.
Censuses of Foraging Bees
We recorded the numbers of bees foraging in 4 quadrats
located within representative patches of A. gangetica during morning
(0800-1000 hours) and afternoon (1400-1600 hours) periods on 6
d. Each 1 m2 quadrat was separated by a minimum of 3 m. On a given
census day, 2-3 counts were made during the morning and afternoon
study periods, respectively, and successive counts were separated
by a minimum of 25 min. Also, on each census day we recorded flower
abundance in the 4 quadrats and collected 100 flowers haphazardly
from the study site at 1000 hours and examined them in the laboratory
for corollar slits or perforations.
As described below, nearly all the honey bees that
were observed displayed only 1 type of foraging tactic. Thus,
in addition to the estimates of absolute density provided by the
quadrat censuses, we recorded the relative density of the different
types of foragers by walking a 10 m transect through the study
area and noting the behavior (based on a single floral visit)
of all honey bees within 1 m of the line. Three surveys were made
during morning and afternoon periods, respectively, on 5 d, and
successive surveys were separated by at least 30 min.
Behavioral Observations
The type of foraging behavior displayed by individual
honey bees was noted during observations of consecutive floral
visits. Initially, we included only data for bees observed to
visit 25 or more flowers. However, because of the invariant nature
of foraging behavior, additional observations were limited to
10 consecutive floral visits per individual honey bee. Morning
(0800-1000 hours) and afternoon (1400-1600 hours) observations
of foraging behavior were made on 10 d, and during any given morning
or afternoon we observed only 10-15 sequences of floral visits
to reduce the likelihood of observing the same individual more
than once. Also, we attempted to select focal bees randomly by
initiating observations on individuals in flight (i.e., whose
foraging tactic was unknown to us). Though emphasis was on honey
bees, behavioral observations were also made for carpenter bees
(2-6 sequences of floral visits were observed during a given morning
or afternoon period on 5 d).
Three types of floral visits were distinguished:
“out” (O) visits during which bees did not enter the
corolla but probed the base of the corolla from the outside (i.e.,
nectar robbing), “in-upright” (IR) visits during which
bees landed on the lower petals and entered the corolla in an
upright position, and “in-upside down” (ID) visits during
which bees landed on the lower petals in an upright position but
then entered the flower while making a spiral path to the top
of the corolla, thus assuming an upside-down position. [Perforations
were absent on previously bagged flowers visited exclusively by
honey bees, and we conclude (as did Barrows 1980) that honey bees
used existing slits and did not perforate flowers themselves.]
For approximately 1/2 the bees observed, we also recorded the
duration of each floral visit (to the nearest 0.1 s with a stopwatch),
the total duration of the observations to obtain rates of flower
visitation. Also, for bees making either type of in-visit we recorded
the number of flowers they approached (operationally defined as
1-3 s hovering immediately in front of a flower) but did not land
on and the number of flowers they landed on but did not enter.
Measurement of Nectar and Pollen Loads
Nectar and pollen loads were estimated for a sample
of honey bees observed displaying each of the different foraging
tactics. Bees were observed over 10 consecutive floral visits
between 0800-1000 hours over 3 consecutive mornings, captured
with an aerial net, and immediately transported to the laboratory
for dissection. Nectar in the crop was drawn into a 10 ul glass
capillary tube. Pollen was scraped off the hindlegs using a dissecting
pin, dried at 45°C for 24 h, and weighed to the nearest 0.1
mg on a Mettler AE163 balance. Because the display of nectar robbing
may be related to bee size (e.g., Barrow & Pickard 1984),
we also measured forewings of these bees (specifically, the length
of the medial cross vein) to the nearest 0.1 mm using a dissecting
microscope equipped with a disc micrometer.
Between-day Consistency in Individual Behavior
To determine whether individual workers displayed
the same foraging tactic over successive days, we uniquely marked
individuals with different color combinations of enamel paint
on their thorax. Bees were observed for 10 floral visits, captured
with an aerial net, and cooled in an ice chest for marking. Following
marking, bees were placed in full sunlight, and after 1-2 min
of warming the bees took flight and often resumed foraging activity.
A total of 70 bees (approximately equal numbers of O-, IR-, and
ID-foragers) was marked between 0830-1100 hours in a single morning,
and over the next 7 d we monitored the site between 0900-1000
hours and observed the foraging behavior of marked individuals
for 10 consecutive floral visits. Additionally, we checked for
within-day consistency in behavior by observing marked workers
between 1400-1500 hours on the first 2 d following marking.
Inter-site Variation in Foraging Behavior
Honey bee foraging tactics were monitored concurrently
at 3 other locations, all of which were infrequently mowed lawns
within 5 km of the main study site. A. gangetica was the dominant
flowering plant at each of these additional sites, and honey bees
and carpenter bees were the only floral visitors observed.
Following the above protocol, we recorded bee abundance
in 4 quadrats (each 1 m2) at each of these sites. At a given site,
3 censuses were made between 0800-1000 hours on 4-5 d. In addition,
we walked a transect line (10-15 m) at each site and recorded
the number of honey bees making 0-, IR-, and ID-visits. At a given
site, 4-6 transects were made during a given morning over 4-5
d. We also made 1 count of flowers in the quadrats per census
day collected 100 flowers at 1000 hours and examined them in the
laboratory for corollar slits or perforations.
Results
Bee and Flower Abundance
Both carpenter bees and honey bees were more abundant
in the morning than in the afternoon. Over a total of 60 censuses
made in the 4 quadrats at the Chaminade site, the density of carpenter
bees averaged 0.33 per m2 in the morning (SE=0.07; range:0-2)
but only 0.08 per m2 in the afternoon (SE=0.04; range: 0-2). Similarly,
the average density of honey bees was 2.08 per m2 in the morning
(SE=0.14; range: 0-5) compared to 1.15 per m2 in the afternoon.
The mean density of A. gangetica flowers was 31.2 per m2 (SE=2.2;
range=23-37; n=6 d × 4 plots=24 censuses). Flower density
was relatively constant over the study period and did not differ
significantly over the different census days (c2=3.9; P > 0.05;
Kruskal-Wallis test with c2 approximation).
Foraging Behavior of Carpenter Bees
Carpenter bees made O-visits exclusively (n=21 and
10 sequences of 10 floral visits for morning and afternoon periods,
respectively). Individuals typically approached a flower frontally,
landed on the petals, and then walked over the flower to the base
of the corolla. Carpenter bees moved rapidly between flowers,
spending little time at each. Averaged over all timed observations,
carpenter bees visited 16 flowers per min (SE=0.9; range: 12-24)
and spent 1.5 s at each flower (SE=0.1; range=0.3-3.4; n=150).
Temporal Variation in Honey Bee Foraging Tactics
During almost all of the floral visitation sequences
observed (215/224=96%), individual honey bees made only 1 type
of floral visit. However, the incidence of the different types
of foragers varied through the day. In the morning, we observed
approximately equal numbers of O-foragers (n=60) and I-foragers
(I=IR and ID combined; n=57), but in the afternoon nearly 2/3
(59/98) of the bees observed made O-visits. Even more striking,
IR- and ID-foragers were equally common during the morning (29
and 28 observations, respectively), but all I-foragers observed
in the afternoon were making IR-visits exclusively. The few bees
(n=9) that displayed mixed foraging tactics were all observed
in the morning and were making a combination of IR- and ID-visits.
Data from the transects showed the same temporal
pattern in foraging behavior. On average, we noted 20 honey bees
per morning transect (SE=0.9; range: 16-25; n=15) of which 11
(55%), 5 (25%), and 4 (20%) were making O-, IR-, and ID-visits,
respectively. Consistent with the census data from the quadrats,
fewer honey bees were present in the afternoon transects (x=12;
SE=0.7; range: 6-19; n=15). Of these, an average of 8 individuals
(67%) were making O-visits, and 4 individuals were making IR-visits.
No honey bees were seen to make ID-visits during any afternoon
transect.
Behavioral Differences Among Honey Bee Foraging Tactics
Rates of floral visitation differed significantly
among the different foraging modes (H=54.1; P < 0.001; Kruskal-Wallis
test; no significant difference was found between morning and
afternoon visitation rates for bees making O- or IR-visits, and
data for each of these tactics were pooled over the 2 periods;
Table 1). Floral visits per min were similar between bees making
IR- and ID-visits (P > 0.05), but these rates were significantly
lower than that observed for O-foragers (P < 0.05 for both
tests; Dunn’s rank sum multiple-comparison test following Daniel
1990). Though IR- and ID-foragers had similar floral entry rates,
bees making IR-visits both landed on a higher proportion of flowers
(after approaching) and entered a higher proportion of flowers
(after landing) than bees making ID-visits (U=558 and 660, respectively;
P < 0.001 in both cases; Mann-Whitney test; Table 1).
Bees utilizing different foraging tactics also exhibited
different handling times at individual flowers (H=50.1; P <
0.001; Kruskal-Wallis test; handling times did not differ between
morning and afternoon for bees making O- or IR-visits, and data
for each of these tactics were pooled over the 2 periods; Table
1). Handling times were similar between bees making IR-and ID-visits
(P < 0.05), but these individuals spent nearly twice as much
time at a given flower as O-foragers (P < 0.05 in both cases;
Dunn’s rank sum multiple comparison test following Daniel 1990).
Table 1. Comparisons among bees making O-, IR-, or
ID-visits, respectively. Average values (se) are presented. Sample
sizes for behavioral data were: O-54, IR-36, and ID-15 individual
bees observed over 10 consecutive floral visits. For nectar and
pollen loads and wing length, 15 bees from each foraging tactic
were examined.
Foraging Tactic
Nectar and Pollen Loads
Both nectar (H=7.0) and pollen (H=24.1) loads varied
with foraging tactic (P < 0.05 in both cases; Kruskal-Wallis
test; Table 1). Bees making O-visits had significantly larger
nectar loads than either IR- or ID-foragers, and IR-foragers,
in turn, carried more nectar than ID-foragers (P < 0.05 in
all cases; Dunn’s rank sum multiple-comparison test following
Daniel 1990). Approximately 1/2 of the O- and IR-foragers had
at least 8 ul of nectar, whereas none of the ID-foragers carried
more than 5 ul. Conversely, ID-foragers had significantly larger
pollen loads than O- or IR-foragers (P < 0.05 in both cases;
Dunn’s rank sum multiple-comparison test following Daniel 1990;
Table 1). All ID-foragers had pollen loads exceeding 1 mg (dry
weight), whereas none of the O-foragers and only 27% (4/15) of
the IR-foragers carried measurable amounts of pollen.
Bees displaying different foraging tactics did not
differ significantly in body size (H=0.9; P > 0.05; Kruskal-Wallis
test; Table 1). Also, body size was unrelated to the volume of
nectar loads (rs=0.21; data pooled over O- and IR-foragers) or
to the dry weight of pollen loads (rs=0.11; ID-foragers only;
P > 0.05 in both cases; Spearman’s rank correlation).
Between-Day Consistency in Individual Behavior
Nearly all of the re-sighted individuals displayed
a single foraging tactic across days. Of the 70 bees marked, 22
were seen on 1 or more mornings after marking (x=4.3 d; range=2-6);
re-sighting probabilities were similar among the 3 foraging types
(29%, 35%, and 34% for O-, IR-, and ID-foragers, respectively).
Of the 22 re-sighted bees, only 2 (2/22=9%) were seen displaying
2 different foraging modes. In the first case, the worker made
ID-visits for 2 d and then O-visits over the next 4 d, and in
the second, the worker made IR-visits on 3 d and then O-visits
on 2 d. Over the 2 d when we made morning and afternoon observations,
a total of 11 different individuals (all O- or IR-foragers) was
seen during both periods of the same day, and all displayed the
same foraging behavior at the 2 times.
Inter-site Variation in Foraging Behavior
The relative occurrence of the different foraging
tactics varied considerably among the 4 sites (Fig. 1). Most notably,
O-visits represented over 50% of all visits at the Wilson and
Chaminade sites but were relatively rare at Manoa and completely
absent at Makiki. Conversely, IR-visits were much more common
at Makiki and Manoa than Wilson or Chaminade. The incidence of
ID-visits varied only between 11%-21% among the sites.
Though only a few sites were monitored, the incidence
of O-visits appeared to be correlated with the abundance of carpenter
bees and consequently the frequency of perforated flowers (Table
2). At Makiki, where O-visits were absent, carpenter bees were
never recorded from the quadrats, and less than 10% of the flowers
had holes or slits. In contrast, carpenter bees and perforated
flowers were most common at Wilson and Chaminade, where high proportions
of honey bees made O-visits. The density of honey bees was also
higher at these latter sites, but since only carpenter bees (and
not honey bees) perforated flowers it seems unlikely that the
occurrence of O-visits was directly related to the abundance of
conspecific foragers (Table 2). Flower density also seemed unimportant
as the Makiki site had a flower density nearly 50% higher than
the Chaminade site (Table 2).
Table 2. Bee and flower density and frequency of
perforated flowers at the 4 study sites. Average values (se) are
provided. See text for sampling methods.
Site
Discussion
As commonly reported for honey bees (Seeley 1985;
Winston 1987), workers at our study sites showed specialization
for nectar vs. pollen collection. IR- and ID-foragers gathered
both resources, but the former concentrated on nectar and the
latter on pollen; workers making O-visits collected nectar exclusively.
Also, with respect to nectar harvesting, workers either collected
it legitimately (IR- and ID-visits) or via robbing through corollar
slits made by carpenter bees (O-visits). Consistency in nectar-collecting
tactics was evident both within individual trips (see also Weaver
1956) and between trips on the same day or on different days (see
Higashi et al. 1988 for similar data on bumble bees).
Nectar robbing appeared to be an opportunistic strategy
whose incidence was correlated with the abundance of carpenter
bees and corollar perforations. However, it is not known why particular
individuals specialized on 1 tactic over another. In the bumble
bee B. pratorum L., smaller workers were nectar robbers exclusively,
whereas larger workers both robbed nectar and collected pollen
(Barrow & Pickard 1984; see also Brian 1952). In our study,
however, foraging behavior was unrelated to worker size. It also
seems unlikely that worker age was important as other studies
(Ribbands 1952; Free 1963) found no consistent age-related pattern
in honey bee foraging behavior. Based on the observations of Weaver
(1956), learning was perhaps an important determinant of foraging
specialization, and workers may have become “fixed”
to a particular foraging mode that was initially successful. Alternatively,
recent studies (Calderone & Page 1988; Robinson & Page
1989) have demonstrated a genetic component to task specialization
in A. mellifera, and it is possible that inter-individual variation
in foraging behavior reflects underlying genetic differences.
In addition to inter-individual variation in behavior,
several behavioral differences were evident among the 3 foraging
tactics. First, though abundance of all foragers declined in the
afternoon, ID-workers were completely absent. Whether ID-foragers
remained in the hive or visited another food source is unknown,
but based on morning and afternoon sightings of the same individuals,
it is unlikely that ID-workers switch to O- or IR-visits at A.
gangetica in the afternoon. Second, rates of floral visitation
varied with foraging mode: O-foragers visited more flowers per
min and spent less time per flower than IR- or ID-foragers, which
had similar movement rates and handling times. Travel distances
between successive floral visits may have also varied with foraging
tactic (Zimmerman 1982a), but this hypothesis remains untested.
Finally, the degree of floral discrimination appeared to differ
between I-tactics. Based on lower probabilities of landing after
approach and floral entry after landing, ID-foragers were apparently
more selective than IR-foragers. This observation suggests that
visual signals associated with pollen availability (Zimmerman
1982b) were more conspicuous and/or reliable than possible visual
(Thorp et al. 1975) or odor (Heinrich 1979) cues associated with
nectar levels. It is unlikely that scent-marking by previous honey
bee visitors was responsible (Wetherwax 1986; Giurfa & Nunez
1992), because unless IR- and ID-foragers leave different odors,
a worker would be unable to assess which type of floral visits
was made previously.
Description of the different foraging modes raises
many questions regarding the economics of food collection. For
example, how does the availability of nectar and pollen change
through the day? How does the rate of nectar intake compare between
workers making O- and IR-visits, respectively? How does this rate
vary through the day for these 2 tactics? How does the rate of
pollen collection compare between workers making IR- and ID-visits,
respectively? Is floral rejection by IR- and ID-workers actually
indicative of depleted resources? We intend to address these and
related questions in future studies.
Acknowledgments
We thank Nani Wong for identifying the plant and
the students of Biology 110, Spring 1993, Chaminade University,
for their enthusiastic assistance in the field. Caryn Ihori, Emma
Shelly, and Meredith Whitney assisted with marking and measuring
honey bees.
References Cited
Barrow, D. A., and R. S. Pickard. 1984. Size-related
selection of food plants by bumblebees. Ecol. Entomol. 9: 369-373.
Barrows, E. M. 1980. Robbing of exotic plants by
introduced carpenter and honey bees in Hawaii, with comparative
notes. Biotropica 12: 23-29.
Benedek, P., L. Bank, and J. Komlodi. 1973. Behaviour
of wild bees (Hymenoptera: Apoidea) on hairy vetch (Vicia villosa
Roth.) flowers. Z. Ang. Entomol. 74: 80-85.
Brian, A. D. 1952. Division of labour and foraging
in Bombus agrorum Fabricius. J. Anim. Ecol. 21: 223-240.
Buchmann, S. L., and C. W. Shipman. 1990. Pollen
harvesting rates for Apis mellifera L. on Gossypium (Malvaceae)
flowers. J. Kansas Entomol. Soc. 63: 92-100.
Burrill, A. C. 1925. Honey bees follow wood bees
for nectar. Science 62: 134.
Calderone, N. W., and R. E. Page. 1988. Genotypic
variability in age polyethism and task specialization in the honey
bee, Apis mellifera (Hymenoptera: Apidae). Behav. Ecol. Sociobiol.
22: 17-25.
Daniel, W. W. 1990. Applied Nonparametric Statistics.
PWS-Kent, Boston, Mass.
Fell, R. D. 1986. Foraging behaviors of Apis mellifera
L. and Bombus spp. on oilseed sunflower (Helianthus annuus L.).
J. Kansas Entomol. Soc. 59: 72-81.
Free, J. B. 1960a. The pollination of fruit trees.
Bee World 41: 141-151, 169-186.
Free, J. B. 1960b. The behaviour of honeybees visiting
flowers of fruit trees. J. Anim. Ecol. 29: 385-395.
Free, J. B. 1963. The flower constancy of honeybees.
J. Anim. Ecol. 32: 119-131.
Gerling, D. 1982. Nesting biology and flower relationships
of Xylocopa sonorina Smith in Hawaii (Hymenoptera: Anthophoridae).
Pan-Pacific. Entomol. 58: 336-351.
Giurfa, M., and J. A. Nunez. 1992. Honeybees mark
with scent and reject recently visited flowers. Oecologia 89:
113-117.
Hawkins, R. P. 1961. Observations on the pollination
of red clover by bees. I. The yield of seed in relation to the
numbers and kinds of pollinators. Ann. Appl. Biol. 49: 55-65.
Heinrich, B. 1979. Resource heterogeneity and patterns
of movement in foraging bumblebees. Oecologia 40: 235-246.
Helms, C. W. 1970. Bees and blueberries. Canadian
J. Zool. 48: 185.
Higashi, S., M. Ohara, H. Arai, and K. Matsuo. 1988.
Robber-like pollinators: overwintered queen bumblebees foraging
on Corydalis ambigua. Ecol. Entomol. 13: 411-418.
Inouye, D. W. 1980. The terminology of floral larceny.
Ecology 61: 1251-1253.
Lieftinck, M. A. 1956. Revision of the carpenter
bees (Xylocopa Latreille) of the Moluccan Islands, with notes
on other Indo-Australian species. Tijdschr. Entomol. 99: 55-73.
McGregor, S. E., S. M. Alcorn, E. B. Kurtz, and G.
D. Butler. 1959. Bee visitors to saguaro flowers. J. Econ. Entomol.
52: 1002-1004.
Ribbands, C. R. 1952. Division of labour in the honeybee
community. Proc. Roy. Soc. London (B) 140: 132-143.
Robinson, G. E., and R. E. Page. 1989. Genetic determination
of nectar foraging, pollen foraging, and nest-site scouting in
honey bee colonies. Behav. Ecol. Sociobiol. 24: 317-323.
Schmid-Hempel, P. 1991. The ergonomics of worker
behavior in social Hymenoptera. Adv. Stud. Behav. 20: 87-134.
Seeley, T. D. 1985. Honeybee ecology. Princeton Univ.
Press, Princeton, NJ.
Thorp, R. N., D. L. Briggs, J. R. Estes, and E. H.
Erikson. 1975. Nectar fluorescence under ultraviolet irradiation.
Science 189: 476- 478.
Wagner, W. L., D. R. Herbst, and S. H. Sohmer. 1990.
Manual of the flowering plants of Hawaii. Bishop Museum, Honolulu,
Hawaii.
Weaver, N. 1956. The foraging behavior of honeybees
on hairy vetch. I. Foraging methods and learning to forage. Insectes
Soc. 3: 537-549.
Wetherwax, P. B. 1986. Why do honeybees reject certain
flowers? Oecologia 69: 567-570.
Winston, M. L. 1987. The biology of the honey bee.
Harvard University Press, Cambridge, MA.
Zimmerman, M. 1982a. The effect of nectar production
on neighborhood size. Oecologia 52: 104-108.
Zimmerman, M. 1982b. Optimal foraging: random movement
by pollen collecting bumblebees. Oecologia 53: 394-398.
Fig. 1. Average frequencies of the different foraging
tactics at the 4 study sites. Average numbers (SE) of honey bees
per transect (n=15-21) were: Makiki 12.0 (0.6); Manoa 12.2 (0.4);
Wilson 26.4 (1.1); Chaminade 20.0 (0.9).
Chaminade University
Honolulu, Hawaii, 96816
University of Hawaii
Honolulu, HI 96822
O
IR
ID
Flowers/Min
10.1 (0.7)
6.9 (0.4)
5.9 (0.4)
% Landings
95 (1.6)
73 (3.9)
% Entries
83 (1.5)
58 (3.5)
Time/Flower (s)
4.3 (0.2)
7.6 (0.4)
7.5 (0.5)
Medial Cross Vein (mm)
1.31 (0.01)
1.29 (0.02)
1.31 (0.01)
Nectar Load (ul)
9.7 (1.6)
7.8 (1.3)
3.3 (0.6)
Pollen Load (mg)
0.0 (0)
1.1 (0.5)
5.1 (0.7)
Number/m2
Perforated Flowers (%)
A. mell.
X. son.
A. gan.
Makiki
0.8 (0.1)
0.0
42 (2.2)
8 (1.2)
Manoa
0.6 (0.1)
0.1 (0.05)
25 (1.9)
65 (3.0)
Wilson
2.6 (0.2)
0.4 (0.2)
56 (2.9)
86 (3.8)
Chaminade
2.1 (0.1)
0.3 (0.1)
31 (2.2)
89 (2.2)