J. exp. Biol.

114, 1-15 (1985) 1

Printed in Great Britain © The Company of Biologists Limited 1985

TEMPERATURE REGULATION OF HONEY BEES

(APIS MELLIFERA) FORAGING IN THE SONORAN DESERT
BY PAUL D. COOPER*, WILLIAM M. SCHAFFER
Department of Ecology and Evolutionary Biology, University of Arizona, Tucson,
Arizona 85721, U.SA.
AND STEPHEN L. BUCHMANN
USDA-ARS Carl Hayden Bee Research Center, Tucson, Arizona 85719
and Department of Ecology and Evolutionary Biology, University of Arizona,
Tucson, Arizona 85721, U.SA.

Accepted 26 June 1984

SUMMARY

A heat budget for foraging honey bees (Apis mellifera L.) indicated that
at 30-35 °C all bees are in positive heat balance during flight. Observations
of honey bees returning to their hives at high ambient temperatures support
the conjecture that honey bees regulate head and thorax temperatures at
high T a by regurgitating droplets of honey stomach contents which are then
evaporated. The proportion of returning bees with a droplet on the tongue
increased with increasing shade temperature (T s ), from essentially no bees
at 20 °C to 40% of returning bees at 40 °C. Pollen foragers carry relatively
little fluid during the hottest periods, and pollen foraging decreased at high
ambient temperatures. Thoracic temperatures of pollen collectors are sig-
nificantly higher than thoracic temperatures of water and nectar gatherers
at 40°C (46-13 vs 44°C). Additionally, water and nectar foragers with
extruded droplets have slightly cooler heads and thoraces (38-94 and
43-22°C) than bees not extruding droplets (40-28 and 44-18°C). Wing-
loading and thoracic temperatures of bees are inversely correlated at high
ambient temperatures (35 °C) and this is probably caused by a higher
propensity of heavier bees to extrude fluid, thus reducing thoracic tem-
perature.

INTRODUCTION

Honey bees have been observed to forage at high ambient temperatures and
under intense solar radiation characteristic of summer in the Sonoran desert of
southwest North America (Heinrich, 1975). Heinrich (1979a, l9S0a,b) reported
that honey bees regulated thoracic temperature above 25 °C ambient and that
evaporation of water from the mouth is used at high ambient temperatures (T a =
46°C) to prevent overheating. However in a previous study of foraging patterns of
bees on Agave schottii in the Sonoran desert, bees were observed to stop foraging at

•Present address: Department of Zoology, La Trobe University, Bundoora, Victoria, Australia 3083.
Key words: Temperature regulation, honey bees, Apis mellifera, evaporative cooling, heat balance, deserts.
2 P . D . C O O P E R , W . M . S C H A F F E R AND S. L . B U C H M A N N

ambient temperatures of 35 °C. Although Heinrich demonstrated that honey bees

could evaporate water for cooling and did so in laboratory conditions, his field
studies of honey bee body temperatures were performed at ambient temperatures
. well below temperatures when evaporative cooling might be necessary (Heinrich,
19796). Convective cooling also may account for a relatively large proportion of heat
loss as demonstrated for other flying insects (Church, 1960; Casey, 1980; Chappell,
1982), obviating high rates of water loss associated with evaporative cooling.
We undertook this study to determine what effect desert ambient temperatures
had on bee body temperatures and whether the relationship between climate and
body temperature might limit the ability of individual bees to forage. We proceeded
by constructing a heat budget for foraging honey bees, then testing our budget in
the field by measuring body temperatures of returning and departing foraging
honey bees over a wide range of ambient temperatures. In addition, we examined
the effect of evaporative cooling and wing-loading on body temperatures of return-
ing foragers, as well as measuring temperatures of pollen foragers. Finally, we
attempt to integrate our field and laboratory data to determine the environmental
limits to foraging by individual honey bees.

METHODS AND MATERIALS

Laboratory studies
Honey bees (Apis mellifera caucasica) were captured in the field, anaesthetized
and frozen with dry ice and returned to the laboratory. After defrosting, bees were
weighed on a Sartorius balance to the nearest 10~4 g, then the head, thorax and
abdomen were separated and each part weighed individually. Linear dimensions
were measured to the nearest 0-2 mm using calipers, while wing areas were
determined from tracing wings on graph paper and either weighing the paper or
counting the number of square millimeters contained. Surface areas for body parts
were calculated using appropriate geometrical equations.
Oxygen consumption (Vo2) was measured by flying bees in an airtight 4-1 glass
jar for at least 5 min continuously. A 50 cm3 air sample was removed after the flight
period with a syringe and injected into an oxygen analyser (Applied Electro-
chemistry S-3A) at a constant rate of 25 cm3 min" 1 using a perfusion pump. An
initial reference gas sample was handled similarly. Both gas samples were injected
through a combination of Drierite® and Ascarite® before analysis to remove water
vapour and CO2. Oxygen consumption rate during hovering flight was then
calculated for STP conditions using standard equations for closed respirometry
systems. Cooling curves for honey bees were determined on freshly killed bees; bees
were killed by injecting 1—5 pil ethyl acetate into the thorax at the junction of the
thorax and pedicel. Rate of cooling for bees in the absence of forced convection was
measured by heating the bees up to 50°C in an oven, then rapidly placing them into
a 1000 cm3 Erlenmeyer flask immersed in a constant temperature (25 °C) water
bath. The temperatures of the bee thoraces were monitored constantly using a 44-ga
thermocouple connected to a Bailey Bat-12 thermocouple thermometer which was
connected to a channelyser and chart recorder (Bailey Instruments, Inc.). The jar
 Temperature regulation of honey bees 3
temperature was recorded simultaneously by scanning with the channelyser every
2s.
Thoracic conductance (Cth) of individual bees was determined at various wind
speeds by implanting 44-ga thermocouples in the thorax. The bees were suspended
in the centre of a laminar flow wind tunnel by glueing the wings to toothpicks which
were anchored in clay. The clay was secured to the sides of the tunnel. A
microscope light was used to heat the bee until thoracic temperature (T,h) was at
least 50°C. The heat source was then switched off and the wind tunnel fan engaged,
while continuously recording thoracic and air temperatures (cooling rates were not
measured until wind speed stabilized). A similar method was employed for
measuring head conductance (Ch), although some error may result in the cooling
coefficient since heat is lost at a rate proportional to the difference between the
equilibrium head temperature and the temperature of the head at any time (Hegel &
Casey, 1982). Abdominal conductance (Cab) at different wind speeds was cal-
culated from the following equation:

Cab(v) = Cab(0) x C th (v)/C th (0),

where C(0) is the conductance of the respective body part in the absence of forced
convection and C(v) the conductance at different wind speeds.

Field work
Honey bees were captured in the field, held between thumb and finger using a
latex glove, and a 40-ga thermocouple implanted in the tip of a 30-ga needle was
quickly inserted into either the head, thorax or abdomen. The thermocouple was
•connected to a Bailey Bat-12 thermocouple thermometer and the temperature of the
various body parts was recorded within 3 s. Only one temperature measurement was
made per bee to avoid problems with cooling during the measurement period.
Shade temperature was recorded at various intervals, usually every lOmin, by using
the same probe and holding it in the shade approximately 1 m above soil surface.
Cosine-corrected short-wave radiation was measured with a Li-Cor 200S
pyranometer. Ground-reflected radiation was measured occasionally by inverting
the radiation sensor over open ground. Wind speed was measured using a hot ball
anemometer (Chappell, 1982), which was previously calibrated in a wind tunnel.
Both wind speed and radiation were mqnitored continuously with the sensors
located 2 m above ground and recorded on a Hewlett-Packard 7100B dual channel
recorder.

Heat budget calculations

Heat budgets of honey bees at various temperatures were calculated using
equation 1 (Bartholomew, 1977):

Q = S + R + AE + C + M, (1)

where Q is total heat flux, S is solar radiation, R is thermal radiation into and away
4 P. D. COOPER, W. M. SCHAFFER AND S. L. BUCHMANN
from the bee, A is the latent heat of evaporation, E is the rate of evaporation, C is
convectional exchange with the air and M is the metabolic rate of the honey bee.
Net changes in heat (Q = 0) were assumed to result in either changes in bee body
temperature or variations in the rate of evaporation.
Solar radiation (S | ) was determined from the pyranometer measurements, with
reflected ground radiation (S ] ) estimated using an albedo of 25 % (based upon the
several measurements with the pyranometer inverted). Incoming long wave
radiation (L j ) was estimated using the formula of Swinbank (1963) where
L | =<5Ta6 with <5=5-31Xl(T 13 Wm~ 2o K~ 6 and T a is air temperature. Out-
going long wave radiation (L f ) was estimated using the Stefan-Boltzman relation,
L | = aT s 4 , with T s the ground surface temperature and o= 5-67X 10~8 W
m °K~4. The air and the corresponding surface temperatures used in calculations
were 20, 27; 30, 42; and 40, 55°C.
Absorptivity of honey bees was assumed to be 0-91 (Willmer & Unwin, 1981)
while surface areas exposed to radiation varied with radiation source. With solar
radiation (S | ) only 25 % of the surface area was assumed to be irradiated (as in a
sphere, Monteith, 1973), while the other forms of radiation (S \ , L | and L f )
were assumed to involve 50% of the surface area (Kenagy & Stevenson, 1982). Bee
radiative heat loss was calculated using the Stefan-Boltzman relation, assuming an
emissivity of 0-97 with T s equal to bee surface temperature, which was assumed to
be air temperature for flying bees (Mahoney & King, 1977).
Metabolic heat gain was estimated from oxygen consumption measurements,
assuming that 80% of energy produced was in the form of heat (Hegel & Casey,
1982). Normal flight evaporative water loss (cuticular plus spiracular loss) was
assumed to be approximately 10% of the metabolic rate (Heinrich, 1975; Nicolson
& Louw, 1982). Convective heat exchange was estimated by multiplying the various
body part conductances by the gradients for heat loss (Tt,—Ta). Tb was the average
temperature measured for arriving bees - either head, thorax or abdomen. T a was
assumed to be the mid-point of the T a range in which the measurements were made.
As a result, convective heat loss may be overestimated as conductance is not equal to
the convective coefficient, but may be somewhat greater (Kingsolver & Moffatt,
1982).
Budget error analysis
Our heat budget can be criticized because of the inherent error in convective heat
loss associated with our use of the conductance, rather than the true convective
coefficient. The error associated with our method may result in an overestimate of
convective cooling but presumably this is a small error. However our assumption
that bee surface temperature equals ambient temperature may be invalid, as Cena &
Clark (1972) found that thoracic surface temperatures of bees returning to the hive
could be as high as 20°C above ambient at low ambient temperatures. Our estimate
of radiational heat loss could be underestimated by 25 % at 20°C.
Our calculation for heat production was determined from the energy cost
associated with hovering flight. However, the power needed for flight decreases as
forward flight speed increases (Pennycuick, 1972). If bees fly at the minimum
power for transport, metabolic heat production during flight is probably lower than
 Temperature regulation of honey bees 5
the values in our budget. Since we do not know the change in metabolism with
increasing flight speed though, we cannot calculate the error associated with this
part of our budget. Changes in radiant heat gain will vary with time of day. If we
assume that T a and T s are constant (which is unlikely), we can calculate the
variation in heat gain dependent only on the change in short-wave radiation. Thus
at T a = 30°C and T s = 42°C, bees will gain 0-111 W and 0-146 W when short-wave
radiation (S) is 600 and 1200 Wm~ 2 , respectively. Short-wave radiant energy was
approximately 850 Wm~ 2 for our calculated value of 0-121 W under the same T a
and T s conditions, and our maximum error is 20%.
At low temperatures, then, we may be underestimating heat loss, but at high
temperatures when T t h = T a , our heat budget error is approximately 10%.

RESULTS

Morphometric measurements of returning foragers are given in Table 1. Mean

body mass of foragers was 100-8 mg with the thorax accounting for about 40% of
the total mass (range 33-6-48-5 mg). Abdominal mass tended to be larger but was
also more variable (ranging between 17-9 and 76-5 mg) as a result of differences in
the nectar or water load. The abdomen has the largest surface area, corresponding
to its large mass; the thorax and head being smaller progressively. Wing area was
relatively constant, but wing loading varied as a result of differences in abdominal
mass.
Mean oxygen consumption rate of flying bees carrying pollen was 84-65 ±
12-60 ml O z g " ^ " 1 (JV=8) compared to 73-03 ± 8-85 ml C^g" 1 h" 1 (A'= 8) for
nectar carriers. However these values were not significantly different from each other
so for purposes of determining heat production during flight the two groups were
combined to yield 78-84 ± 7-59mlO 2 g~ 1 h~'. These values are similar to previously
reported metabolic rates (Heinrich, 19806; Withers, 1981).
Thoracic conductance of honeybees varied between 6-3xl0~ 4 W o C~' in still air
and 5xl0~ 3 W°C~ I at a wind speed of 5-1 ms~'. The increase in conductance was
roughly linear with the square root of wind speed (Fig. 1). Head conductance also
increased with wind speed, although the rate of increase was less than that of the
thorax (Fig. 1).

Table 1. Honey bee morphometrics

Body section Units X s.u. A"
m
Body mass ( g) 100-80 20-20 96
Head mass (m|) 9-97 1-36 96
Head surface area 2-46X10"5 0-43X10"5 13
Thorac c mass (mg) 40-70 2-90 96
Thorac c surface area 4-15X10"5 0-29X10"5 15
Abdom nal mass (mg) 46-68 17-44 96
Abdom nal surface area (m2) 9-OOxlO"5 l-60xl0" s 14
Wing a ea (cm2) 0-36 0-04 76
Wing load (N m" 2 ) 28-27 610 76

Bees used in measurements were all returning forage

 P. D. COOPER, W. M. SCHAFFER AND S. L. BUCHMANN

Fig. 1. Thoracic and head conductance of honey bees in relation to the square root of wind
velocity. Means ± 2 standard errors. Points are fitted by straight line least squares regression. Cth
= l-93V 1/2 +1-40 ( ^ = 0-97), C h = O81V 1/2 +0-34 (>-2 = 0-95).

Table 2. Heat budget for Apis mellifera/Tymg at 3 m s ' under clear sky conditions
Temperature (°C) Gain Loss

20 Metabolism 0037 Radiant 0063

Radiation 0 099 Convection 0-125
Evaporation 0-004
Total 0136 0192

30 Metabolism 0037 Radiant 0-072

Radiation 0121 Convection 0-064
Evaporation 0-004
Total 0158 0140

40 Metabolism 0037 Radiant 0-085

Radiation 0143 Convection 0009
Evaporation 0086*
Total 0180 0180

•Water loss rate equal to 0035 mg H 2 O s ' or 127-4 % body mass h"
All values are for 100-mg bee and are expressed in W.
Heat loss by evaporation at 40 °C is determined by subtraction.

Heat budgets of flying honey bees were constructed using body temperatures of
bees measured in the field and the appropriate ambient temperatures, thoracic and
head conductance measured at different wind speeds, the amount of radiation
present under clear sky conditions and the heat production associated with
metabolism of flying bees determined in the laboratory (Table 2). With increasing
temperature, radiative heat gain and loss increase, and convective heat loss
decreases. At the lower temperatures there is a net heat efflux, while evaporative
 Temperature regulation of honey bees 7
cooling must occur at the highest temperatures if T b is kept within normal limits.
By iterating this approach at different flight speeds, a family of curves can be
constructed relating net heat exchange to ambient temperature (Fig. 2). At slow
flight speeds, bees will gain heat at all temperatures, but as flight speed increases,
bees begin to gain heat only at progressively higher temperatures.

Field studies
Temperatures of head, thorax and abdomen of both departing and arriving bees
are shown in relation to local shade temperature in Fig. 3. Temperatures of
departing bees are relatively independent of shade temperatures, but arriving bee
temperatures have a sigmoidal relation with increasing shade temperature. Below
40 °C, temperatures of both heads and thoraces of arriving and departing bees are
maintained above T a , while abdominal temperatures of arriving bees are nearly
always close to ambient temperatures.
To determine the effect of various types of foraging on bee body temperatures,
we divided arriving bees into three groups according to their activity immediately
before capture for temperature measurement: (1) bees not carrying pollen and with
no fluid observed on the proboscis (NPNF), (2) bees not carrying pollen and with
fluid observed on the proboscis (NPF), and (3) bees carrying pollen but no fluid
regurgitation observed (PNF). Results of these experiments are given in Table 3.
Two trends are apparent, (1) bees with fluid extruded are cooler than bees without
fluid, and (2) pollen carriers tend to be the hottest bees by 2-3°C at the highest
temperatures. Additionally, as shade temperature increased, an increasing pro-

Fig. 2. Calculated net heat exchange (not including evaporative water loss) of honey bees in flight
at various velocities plotted against ambient temperature under clear sky conditions. Heat gain
would occur at T a = 30°C for bees flying at 3 m s " ' and at 33°C for bees flying at 5'1 m s " 1 .
8 P. D . C O O P E R , W. M. S C H A F F E R AND' S. L . BUCHMANN

portion of nectar foragers were observed to have fluid on the proboscis, while few
pollen carriers ever had fluid droplets (Fig. 4). Stomach content analysis of 19
pollen foragers performed at the time of temperature measurements yielded an
average crop content of 1'3/u.l, with 15 of the 19 bees having completely empty
crops, compared with a mean content of 79fi\ in fluid-carrying bees. The
proportion of foragers with pollen also decreased at the higher temperatures
(Fig. 4).
Higher thoracic temperatures with higher wing loads would be expected if
metabolic rate was elevated to offset the higher lift requirement. However, as wing
load increased, thoracic temperature decreased (Fig. 5) at ambient temperatures
above 35 °C. This may be explained by assuming that the heaviest bees (those with
the higher wing loads) are more likely to use evaporative cooling. The mean weight
of droplet-carrying bees is 107-9 ±3-0mg (JV=23) compared to 98-7 ± 2-5 mg
(iV= 74) (P< 0-10) for foraging bees not carrying droplets.
Foraging bees have an alternative method for controlling body temperature
within normal ranges. They can cease flying and allow themselves to lose heat
passively. This method typically results in thoracic temperature decreases of about
2°C (42-7 ± 0-2 vs 40-8 ± 0-4°C) within 15 s at ambient temperatures of 36-37°C.

DISCUSSION
Thermoregulation in flying insects has been investigated with respect to effects of
metabolic rate, body size, behaviour and orientation and the possible advantages of
high body temperatures during foraging (May, 1979; Kammer, 1981; Casey, 1981;
Willmer, 1982a). Heinrich (1979c) integrated many of these aspects of thermo-
regulation in presenting his overview of bumble-bee ecology. Recently, several
studies have included the effect of microclimate on body temperature in flying
insects in the field (Chappell, 1982; Kingsolver & Moffat, 1982; Willmer, 1982fe,c).
The studies on Xylocopa californica (Chappell, 1982) and Colias eurytheme (King-
solver & Moffat, 1982) indicated that heat budgets calculated from laboratory
models can be used to predict temperatures of insects flying in the field. Our model
has allowed us to estimate the relative importance of convective and evaporative
cooling for honey bees flying at high ambient temperatures commonly occurring in
the desert (Table 3).

Foraging bees
Low ambient temperature flight
The model of heat transfer indicates that honey bees will have a net heat loss
during flight at low T a . The field data support this prediction as bees returning to
the hive at ambient temperatures below 27°C are cooler than departing bees.
Continuous flight at ambient temperatures below 20°C will result in thoracic

Fig. 3. Body temperatures of Apis mellifera caucasica arriving and departing from a standard
Langstroth hive plotted against air temperature in the shade. (A) Head temperatures, T h ; (B)
thoracic temperatures, T t h ; (C) abdominal temperatures, T a b J . Means ± confidence interval.
Numbers of bees in parentheses. O, arriving bees; • , departing bees.
 Temperature regulation of honey bees

35

30

50

40

45

40

20

10 15 20 25 30 35 40 45 50

Fig. 3
 Table 3. Influence of pollen and regurgitation on temperatures of arriving Caucasian honey bees
NPNF* NPF* PNF*
X 95 % CI X 95 % CI X 95 % CI n
o
Head
o
20-25 51 33-69f 0-6 0 — 27 34-90f 0-5
25-30 21 36-93 0-8 0 20 37-04 0-5
30-34 9 39-81 1-3 1 4010 11 40-19 0-7
34-38 18 39-72f 0-8 1 42 30f — 17 4105 0-8
38-42 73 40-28t 0-4 24 38-94f 0-7 3 4203 6-2

Thorax
IP
20-25 47 37-71 0-7 0 35 37-73 0-5 n
25-30 26 40-66 0-8 0 22 40-69 0-7 x
30-34 31 42-02f 0-7 3 41-73 4-2 19 43-95f 0-5
34-38 120 42-69J: 0-3 11 41-80J: 11 17 44-16-j- 0-5
38-42 50 44-18f 0-6 39 43-22f 0-7 7 4613f l-V

AbdomeVi
20-25 41 29-25 0-6 0 24 28-90 0-5
25-30 25 32-82 0-8 0 _ 18 31-85 0-6
30-34 14 36-39 10 0 — — 11 37-43 0-6
34-38 12 37-51f 10 1 42-40 — 19 38-96t 0-9
38-42 67 39-60t 0-6 25 38-21t 1-0 3 41-30 2-4
w
Statistical differences relative to NPNF temperatures. c
Mean and 95 % confidence interval. n
• N P N F , no pollen, no fluid; NPF, no pollen, fluid on proboscis; PNF, pollen, no fluid.
x
2
t-P<005; I 0 >
2
2
 Temperature regulation of honey bees 11

0-50 •

20

Fig. 4. Proportion of returning honey bees extruding a droplet of fluid on the tongue. Proportion
increases with air temperature in the shade (T s ) for nectar and water collectors (hollow circles).
Pollen gatherers (solid circles) almost never regurgitated fluid. Proportion of bees returning with
pollen (triangles) declines for T 3 exceeding 38°C. Values determined from pooling data for 5°
intervals between 20 and 30°C, 8° intervals between 30 and 38°C and the 4° interval between 38
and 42°C. Lines fitted by eye.

50
o
0
° ° <6 . g> 0
0 " — Q —-———
o 0 0
40 o 0

30 •

20 30 40
Wing load (N m" 2 )
Fig. 5. Relation between thoracic temperature and wing-loading for bees flying at temperatures
between 35 and 40°C under clear sky conditions. Line fitted by least squares regression T, h =
47-22-0-123 (WL), r = 0-415, jV= 40. Slope significantly different from zero (t = 2-81, df = 38,
P<0-0l).

temperatures dropping below the minimum temperature of 33 °C necessary to

produce lift (Esch, 1976). Bees leaving the hive with a thoracic temperature of 35 °C
would be able to fly only for 35-40 s before thoracic temperatures fell 5°C. Slower
flight would reduce the rate at which bee temperature falls below the critical
temperature for flight (Fig. 2). Esch (1976) suggests that honey bees can change
the body angle for flight and could sacrifice flight velocity for additional lift and thus
fly at lower thoracic temperatures, although we never observed this, and indeed the
12 P. D . C O O P E R , W. M. S C H A F F E R A N D S. L. BUCHMANN

increased metabolism associated with slower flight speed (Pennycuick, 1972) may
preclude this possibility. More reasonably, bees could alternate between flying and
warming-up as Heinrich (197%) suggested, permitting foraging at lower tem-
peratures.

High ambient temperature flight

We have observed foraging activity at ambient temperatures above 40 °C.
Although a high rate of evaporative water loss is predicted for body temperature
maintenance (127% body weight h" 1 ) at 40°C from our model, foraging does not
appear to be limited at 40°C. However, the rate at which heat can be lost by
evaporative cooling may be limited either by availability of fluid or by the ability of
bees to extrude fluid. Increased cloud cover and lower energy gain from direct solar
radiation would reduce heat input rate and thus lessen the required rate of
evaporative cooling, thereby extending foraging time.
Ambient temperatures above 40°C curtail pollen collection (Fig. 4). Pollen
carriers have the highest temperatures measured for returning foragers. Since bees
carrying pollen do not have significantly higher flight metabolic rates than the other
foragers, it is reasonable to conclude that the differences in body temperature result
from the lack of fluid available for evaporative cooling.
Bees can tolerate body temperatures of 50 °C for short periods without lethal
effects (Lensky, 1964), and we have measured thoracic temperatures of pollen
carriers exceeding 50°C. Without fluid available, bees could fly for 45 s at
T a = 40 °C before body temperature increased from 40 to 50 °C. Assuming a flight
speed of 5 m s " , and a 'beeline' flight, a foraging radius of only 225 m would be
possible. Evaporation of l - 58mgH2O over the 45-s flight would prevent any
increase in body temperature. Returning foragers in June 1981 carried 8-10 mg of
nectar with a concentration of between 35 and 40 % total dissolved solids (un-
published data). Thus, even in these conditions, 3-4mg H2O were available for
evaporative cooling, permitting a further 120 s of flight time before body tem-
perature began increasing.
Bees collecting pollen at high ambient temperatures are more susceptible to
increased body temperatures since they do not carry nectar during the hottest
periods of summer. The observed higher temperatures of pollen gatherers may also
reflect increased wing-loading as a result of heavier payloads (mean mass of pollen
collectors was 110-1 mg compared to 95-8mg for other foragers in June) and
possibly increased parasite drag associated with forward flight and the location of
pollen on the legs.
Wing-loading does appear to influence thoracic temperature in nectar- and water-
collecting bees, but bees with the highest wing-loads have the lowest thoracic
temperatures, instead of the expected higher temperatures (Heinrich, 1979c). The
lower thoracic temperatures can be explained if bees with heavier nectar or water
loads exhibit a greater proclivity to cool evaporatively, and the relation between
body weight and regurgitation is consistent with this hypothesis. The relation
between thoracic temperature and body weight is in contrast to the measurements
of Heinrich (19796) for A.m. adansonii. His measurements were done at relatively
 Temperature regulation of honey bees 13

Relative bee mass

Fig. 6. Relationship between net heat flux, relative bee mass (loaded mass/unloaded mass) and
ambient temperature for honey bees flying in the desert under clear sky conditions. Honey bees
with no load represented by vertical dashed line through point e of abscissa. Heat balance (gain =
loss) represented by horizontal dashed line. Angled lines represent isoclines of temperature, with
ambient temperature increasing from right to left. Highest temperature isocline bisects intersection
of heat balance and empty bee weight lines indicating that bee would always be gaining heat when
flying in these conditions.

low temperatures when honeybees would be losing heat during flight and this might
prevent observation of any effects of foraging load on thoracic temperatures.
Foraging for nectar and water does not appear to be limited by normal tem-
peratures in the Sonoran desert. Pollen foraging may be limited at high T a .
However availability of pollen may also be low during the hottest periods of the day,
as most anthers dehisce either in the morning or evening. If bees overheat, they can
quickly land and cool off in favourable microclimates which occur on plants (see
above). Thus 'resting bees' discussed in a previous paper (Schaffer et al. 1979) may
actually be overheated, rather than out of energy. Although bees began flying again
when given a 25 % sugar solution, they might have done this if only water had been
imbibed.
Body weight, heat balance and ambient temperature can be tentatively integrated
in Fig. 6. The ambient temperature isoclines increase in temperature from right to
left. The critical temperature is the ambient temperature when empty bees (no
fluids in honey stomach) are gaining heat during flight (point where dotted lines
intersect). At lower temperatures, unladen bees actually lose heat during flight, as
indicated by the intersection of temperature isoclines at progressively more negative
points on the empty bee line. Bees loaded with nectar or water are able to fly and
maintain body temperature by evaporative cooling at temperatures when empty
bees of the same mass begin to increase body temperature.
PollenTcarriers are represented on the graph as empty bees. By eliminating the
possibility of evaporative cooling, any increase in weight would automatically result
in an increase in net heat gain. This response is what we believe is observed in the
increased body temperatures of pollen collectors.
This study is consistent with the previous suggestion (Schaffer et al. 1979) that
14 P. D. C O O P E R , W. M. S C H A F F E R AND S. L . BUCHMANN

honey bees stop foraging on Agave schottii above 35 °C, not as a result of individual
thermoregulatory problems, but rather in response to the increased requirements
for fluid necessary for hive thermoregulation. Our calculations indicate that a single
bee could continue to forage at profit, with respect to both energy and fluid, on the
agaves. Possibly the response of hive bees to bees bringing in relatively small loads
is decreased at high temperature. If bees switch to foraging on flowers with either
more dilute nectar or which contain greater quantities of nectar, the rate of fluid
and/or energy brought into the hive would be increased. Studies on artificially
heated hives and colonies indicate that at high temperatures both nectar foraging
and water foraging increase (unpublished data), suggesting that if nectar is available
in large quantities, foraging would be stimulated to increase the rate at which it
enters the hive. Thus the reduced foraging on Agave schottii at high temperatures
may reflect the colony requirements for resources, and not a simple individual bee
thermoregulatory response.

This work was supported by NSF grant DEB-80-03783. We would like to thank
M. Glass, M. Hardy, R. Sumner and D. Vleck for assistance in various aspects of
the work. We especially appreciated the maintenance of the experimental hives
performed by C. Shipman and USDA-ARS Carl Hayden Bee Research Laboratory
for the use of their facilities. We thank M. D. Buchmann for the illustrations in
Fig. 3. The manuscript was improved thanks to the efforts of P. Adams, T. Casey,
M. Chappell, E. Edney, B. Heinrich and two anonymous reviewers.

REFERENCES
BARTHOLOMEW, G. A. (1977). Body temperature and energy metabolism. In Animal Physiology: Principles
and Adaptations, 3rd ed., (ed. M. S. Gordon), pp. 364-449. New York, London: MacMillan.
CASEY, T. M. (1980). Flight energetics and heat exchange of gypsy moths in relation to air temperature.
jf. exp. Biol. 88, 133-145.
CASEY, T. M. (1981). Behavioral mechanisms of thermoregulation. In Insect Thermoregulation, (ed. B.
Heinrich), pp. 79-114. New York: John Wiley & Sons.
CENA, K. & CLARK, J. A. (1972). Effect of solar radiation on temperatures of working honey bees. Nature,
Land. 236, 222-223,
CHAPPELL, M. A. (1982). Temperature regulation of carpenter bees (Xylocopa califomica) foraging in the
Colorado Desert of Southern California. Phys Zool. 55, 267-280.
CHURCH, N. S. (1960). Heat loss and the body temperatures of flying insects. II. Heat conduction within the
body and its loss by radiation and convection. J . exp. Biol. 37, 186-212.
ESCH, H. (1976). Body temperature and flight performance of honey bees in a servo-mechanically controlled
wind tunnel. J. comp. Phys. 109, 265-277.
HEGEL, J. R. & CASEY, T. M. (1982). Thermoregulation and control of head temperature in the sphinx moth,
Manduca sexta.J. exp. Biol. 101, 1-15.
HEINRICH, B. (1975). Thermoregulation and flight energetics of desert insects. In Environmental Physiology
of Desert Organisms, (ed. N. Hadley), pp. 90-105. Stroudsberg, Pennsylvania: Dowden, Hutchinson &
Ross Inc.
HEINRICH, B. (1979a). Keeping a cool head: honeybee thermoregulation. Science, NY. 205, 1269-1271.
HEINRICH, B. (19796). Thermoregulation of African and European honeybees during foraging, attack, and
hive exits and returns. J. exp. Biol. 80, 217-229.
HEINRICH, B. (1979C). Bumblebee Economics. Cambridge, Mass.: Harvard University Press.
HEINRICH, B. (1980a). Mechanisms of body-temperature regulation in honey bees, Apis mellifera. I.
Regulation of head temperature. J. exp. Biol. 85, 61-72.
HEINRICH, B. (19806). Mechanisms of body-temperature regulation in honey bees, Apis mellifera. II.
Regulation of thoracic temperature at high air temperatures. .7. exp. Biol. 85, 73—87.
KAMMER, A. E. (1981). Physiological mechanisms of thermoregulation. In Insect Thermoregulation, (ed. B.
Heinrich), pp. 115-158. New York: John Wiley & Sons.
 Temperature regulation of honey bees 15
KENAGY, G. ] . & STEVENSON, R. D. (1982). Role of body temperature in the seasonality of daily activity in
tenebrionid beetles of eastern Washington. Ecology 63, 1491—1503.
KINGSOLVER, J. G. & MOFFATT, R. B. (1982). Thermoregulation and the determinants of heat transfer in
Colias butterflies. Oecologia 53, 27-33.
LENSKY, Y. (1964). Resistance des Abeilles (Apis mellifica L. var. Ligustka) a des Temperatures £lev£es.
Insectes Soc. 11, 293-300.
MAHONEY, S. A. & KING, J. R. (1977). The use of the equivalent black-body temperature in the thermal
energetics of small birds. J. therm. Biol. 2, 115-120.
MAY, M. (1979). Insect thermoregulation. A. Rev. Ent. 24, 313-349.
MONTEITH, J. L. (1973). Principles of Environmental Physics. New York: American Elsevier Publishing Co.,
Inc.
NICOLSON, S. W. & Louw, G. N. (1982). Simultaneous measurement of evaporative water loss, oxygen
consumption, and thoracic temperature during flight in a carpenter bee. J. exp. Zool. 222, 287-296.
PENNYCUICK, C. J. (1972). Animal Flight. London: Edward Arnold Ltd.
SCHAFFER, W. M., JENSEN, D. B., HOBBS, D . E., GUREVITCH, J., T o o o , J. & SCHAI-I-ER, M. V. (1979).
Competition, foraging energetics, and the cost of sociality in three species of bees. Ecology 60, 976-987.
SWINBANK, W. C. (1963). Long wave radiation from clear skies. Q. jfl R. met. Soc. 89, 339^348.
WILLMER, P. G. (1982a). Microclimate and the environmental physiology of insects. Adv. Insect Phvs. 16,
1-57.
WILLMER, P. G. (19826). Thermoregulatory mechanisms in Sarcophaga. Oecologia 53, 382—385.
WILLMER, P. G. (1982c). Hygrothermal determinants of insect activity patterns: the Diptera of water-lily
leaves. Ecol. Ent. 7, 221-231.
WILLMER, P. G. & UNWIN, D. M. (1981). Field analyses of insect heat budgets: reflectance, size and heating
rates. Oecologia 50, 250-255.
WITHERS, P. C. (1981). The effects of ambient of air pressure on oxygen consumption of resting and hovering
honeybees. J. comp. Physiol. 141B, 433-437.