AMER. ZOOL., 28:829-844 (1988)
Estimates of Daily Energy Expenditure in Birds: The Time-Energy
Budget as an Integrator of Laboratory and Field Studies1
DAVID L. GOLDSTEIN
Department of Biological Sciences, Wright State University,
Dayton, Ohio 45435
SYNOPSIS. Measures of energy expenditure by free-living birds can provide quantitative
tests of a number of ecological theories, regarding such diverse phenomena as foraging
strategies, resource competition, or parental investment. Our confidence in these tests
rests heavily on the confidence we have in the estimated rates of energy expenditure. The
most common approach to obtaining such estimates is the construction of time-energy
budgets, in which the durations of an animal's daily activities are multiplied by the respective energy costs of the activities, and these costs are summed. Our knowledge of the
energy costs of activities, particularly locomotion, has greatly advanced in recent years,
as has the ability to adequately assess thermoregulatory costs. Comparisons between timeenergy budgets and direct measures of energy expenditure obtained using doubly labeled
water indicate that time-energy budgets can yield accurate estimates of energy expenditure. However, this is likely to be achieved only under fairly rigorous conditions in which
resting costs, activity costs, and thermoregulatory costs are all well described. Evidence is
accumulating to suggest that, under some conditions, energy expenditure by birds reaches
a maximum sustainable level, at which point it is limited by the physiological capacities
to ingest and assimilate energy. Under these conditions, behavioral responses to changing
physical environments and resource availability may be critical to the maintenance of
energy balance.
energy expenditure by free-living birds
The pattern of use and acquisition of have been applied to such diverse phenomenergy by animals is a critical determinant ena as annual patterns of energy expenof evolutionary fitness. Some energy costs, diture (Wijnandts, 1984; Masman, 1986,
such as basal metabolism and perhaps ther- and others), parental investment in mating
moregulation, are obligatory and must be systems involving monogamy {e.g., Westerbalanced by energy intake. Other activities terp and Bryant, 1984), polyandry (Maxalso require energy expenditure at differ- son and Oring, 1980), and helpers at the
ent levels, but the pattern of expenditure nest (Reyer and Westerterp, 1985), foron these activities may be moreflexibleand aging strategies (Gill and Wolf, 1975; Pyke,
1979), and others.
responsive to the daily needs of the animal.
The commonest way by which daily
Finally, the energy which is acquired in
excess of that needed to sustain obligatory energy expenditure (DEE) is estimated for
and activity costs is available for storage birds is construction of time-energy budand production. Thus, the balance between gets. The daily pattern of activities is
energy acquisition and expenditure is of recorded in the field, and the durations of
fundamental importance in determining the various activities are multiplied by their
the capacities of animals to survive the respective energy costs to calculate total
stresses of the physical environment, grow, energy expenditure. Alternatively, energy
and reproduce. As such, studies of field expenditure can be measured directly using
energetics are able to address problems in doubly labeled water. The confidence we
a number of areas of ecology. Studies of have in using these data as quantitative tests
of ecological theories rests on the reliability of the estimates of DEE. In this paper
I will discuss the techniques which are used
in the laboratory and field to estimate these
1
From the Symposium on Energetics and Animal energy costs, and will then discuss the
Behavior presented at the Annual Meeting of the implications of these approaches for the
American Society of Zoologists, 27-30 December
reliability of time-energy budgets.
1986, at Nashville, Tennessee.
INTRODUCTION
829
830
DAVID L. GOLDSTEIN
METHODS OF ESTIMATING ENERGY
EXPENDITURE BY FREE-LIVING BIRDS
Three primary methods have been used
to estimate the daily energy expenditure
of free-living birds. These are time-energy
budgets, doubly labeled water, and analysis
of energy consumption and assimilation.
In time-energy budget analyses, a compendium is made of the durations of activities performed by birds in the field. These
durations are then multiplied by the
respective energy costs of the various activities, and the energy costs are summed.
Energy expenditure for thermoregulation
and production must be added to this. The
advantage of the time-energy budget is that
it analyses two important resources, time
and energy. It also requires a minimum of
expense and equipment in the field. The
disadvantage is that accurate time-energy
budgets require an extensive knowledge of
the activity costs and thermal biology of
the animal.
The doubly labeled water technique
measures the rate of CO2 production of an
animal by comparing the turnover rates of
isotopes of oxygen (typically 18O) and
hydrogen (tritium or deuterium). An animal is captured, injected with isotopes, and
released. Some time later the animal is
recaptured. The difference in isotope levels
in blood samples taken just before release
and just after recapture is used to calculate
the rate of CO 2 production. The calculated
rate is an integrated one for the entire
interval between the blood samples. The
advantage of this technique is that it is
direct and quite accurate (within 7-8% of
simultaneous direct measures of CO2 production in laboratory validation studies
[summarized in Masman, 1986]). Disadvantages include potentially high financial
costs and the need to recapture the study
animal. A knowledge of the composition
of the diet is required to convert from CO2
production to energy equivalents. Finally,
doubly labeled water provides a single integrated measure of energy expenditure;
additional observations and assumptions
are required to address questions about the
energetics of specific behaviors or behavioral patterns.
In certain situations, it is possible to
quantitatively collect the excreta produced
by free-living birds. Together with a
knowledge of the diet and the digestive
assimilation efficiency, one can calculate the
rate of energy ingestion and expenditure.
This technique is attractive because it
relates energy expenditure directly to food
intake. However, it can be used only in
special circumstances where quantitative
collection of excreta over known time
intervals is possible. This approach has been
successfully employed just a few times
(Moss, 1973; Wijnandts, 1984).
Of these three techniques, time-energy
budget analysis requires the greatest
knowledge of an animal's behavior and
physiology and the greatest number of
assumptions. However, because of its relatively small expense and general accessibility, this approach to estimating daily
energy expenditure has been used much
more frequently than the other two.
Because of this, and because the timeenergy budget is the only approach which
explicitly incorporates behavioral observations (this is a symposium on behavior
and energetics), most of the following discussion will relate to an analysis of timeenergy budgets.
METHODS OF ESTIMATING THE
ENERGY COSTS OF ACTIVITIES
Measurement in the laboratory
Laboratory measurements of the energy
costs of avian activities have all employed
the same basic technique—measurement
of the oxygen consumption of an animal
engaged in steady-state activity. Recently,
the protocol for calculating the oxygen
consumption of animals in non-steady-state
activity (oxygen concentration in the measured air stream not at equilibrium) has
been described (Bartholomew etal., 1981).
This protocol should provide the means to
measure in the laboratory the energy costs
of activities which are often not sustained
for long periods (such as preening, feeding,
or perhaps singing). However, the technique has not yet been applied to the measurement of avian activity costs.
TIME AND ENERGY BUDGETS OF FREE-LIVING BIRDS
831
Measurement of oxygen consumption
in the field
In some unusual circumstances, it is possible to apply laboratory-type measurement of oxygen consumption directly to
active animals in natural environments.
This has recently been accomplished with
free-living hummingbirds fed from feeders
which were also designed as respirometry
masks (Bartholomew and Lighton, 1986).
Doubly labeled water
DAILY ENERGY
EXPENDITURE
(kJ/j)
4
8
12
16
20
24
HOURS FLYING/24 HOURS
FIG. 1. Daily energy expenditure (measured with
doubly labeled water) as a function of the time spent
in flight in sooty terns (Sterna fuscata). The slope of
this regression has the units of kj/g hr, and represents
the metabolic cost of flight. This analysis assumes that
other costs besides flight (such as costs of thermoregulation, digestion, comfort movements, etc.) do not
differ between flying and non-flying birds. Redrawn,
with permission, from Flint and Nagy (1984).
As described above, doubly labeled water
provides a measure of energy expenditure
integrated over the entire interval between
the initial and final blood samples. The
interval required for sufficient change in
the levels of isotopes in the blood is typically one to several days; hence, this technique is not well suited to direct measure- because of the inherent inaccuracies in both
ment of the cost of individual activities. the doubly labeled water measurement and
However, under certain circumstances data the estimated durations and energy costs
acquired using doubly labeled water can of the other activity and thermoreguiatory
be analyzed to calculate activity costs.
components. This approach is therefore
For activities with high energy costs, likely to be most applicable to activities
measurement periods during which the which occupy a relatively large portion of
activity occurs for a long duration should the total energy budget. This technique is
yield a higher total CO2 production than basically that used to estimate the energy
those during which the activity occurs for cost of foraging by Gambel's quail (Goldjust a short time. One can calculate a stein and Nagy, 1985) and has also been
regression relating total energy consump- used to measure flight costs in the Eurotion (CO2 production) to the duration of pean robin (Tatner and Bryant, 1986) and
the particular activity (Fig. 1). The slope the barn swallow (Turner, 1983).
of this regression represents the increment
in energy expenditure per unit time {i.e., Estimates of activity costs based on
the energy cost of the activity). This tech- measurement of heart rate
nique has been used to calculate the energy
The oxygen consumption of an animal
cost of flight in several species of birds is equal to the product of the cardiac out(Hails, 1979; Turner, 1983; Flint and Nagy, put (ml blood/min) and the difference in
1984; Masman, 1986; Tatner and Bryant, oxygen content between arterial and ve1986) and the energy cost of swimming in nous blood (ml O 2 /ml blood); cardiac outjackass penguins (Nagy et al, 1984).
put in turn is equal to stroke volume (ml
In a second approach, it is possible that blood/beat) times heart rate (beats/min).
one can estimate with reasonable accuracy Hence, heart rate is one variable which can
the energy costs of all components of a time change in response to a changing demand
budget except for one. These costs can be for oxygen delivery to the tissues (i.e., a
subtracted from the total energy consump- changing oxygen consumption). Using
tion measured with doubly labeled water, heart rate as an index to metabolic rate is
and the difference provides an estimate of attractive because it can be monitored from
the energy cost of the final activity cate- a distance, and changes in heart rate can
gory. This estimate is likely to have a rel- be detected over very short time intervals.
atively great uncertainty associated with it,
Despite these attractions, the use of heart
832
DAVID L. GOLDSTEIN
Callipepla gambelii
mean mass 140.9 g
O 1981
A 1982
S3
the loss of stored energy. This protocol has
been used several times to estimate the
energy cost of flight (e.g., Nisbet, 1963;
Dolnik and Gavrilov, 1973).
A
ENERGY COSTS OF SPECIFIC ACTIVITIES
*
Energy costs have been measured for a
variety of activities in birds spanning a wide
range of body masses. To compare data
o O
o o O
between species, the energy cost of activity
solid symbols - gular flutter
is usually expressed as a multiple of some
other
standardized cost, such as average
40
45
50
30
35
daily metabolic rate, basal metabolic rate,
or standard metabolic rate. I will express
FIG. 2. Resting metabolic rate (RMR) as a function these data as multiples of basal metabolic
of ambient temperature in Gambel's quail (Callipepla
gambelii) captured from the same population during rate (BMR, the metabolic rate measured
the summers of two different years. Note that RMR on resting post-absorptive birds at night)
was significantly higher in 1982 than in 1981, and or resting metabolic rate (RMR, measured
that in both years it was significantly lower than the during the day). It is worth noting, howallometrically predicted value. From "Resource uti- ever, that BMR and RMR can vary between
lization by desert quail: Time and energy, food and
water" by D. L. Goldstein and K. A. Nagy, Ecology, seasons and populations (Fig. 2); it is not
1985, 66, 378-387. Copyright 1985 by Ecology. clear whether, in a particular species, activReprinted by permission.
ity costs are really multiplicative above these
changing resting costs, or whether they are
instead additive.
rate as an indicator of energy metabolism
has two serious limitations (reviewed by Walking and running
Johnson and Gessaman, 1973). First, stroke
Measurement of the metabolic cost of
volume and arterial-venous oxygen differ- terrestrial locomotion in birds has been
ence can also change to enhance oxygen accomplished in a number of laboratory
delivery. Second, the relationship between studies, representing more than fifteen difheart rate and metabolic rate appears to ferent species from ten orders. Certain patbe highly idiosyncratic; the relationship terns emerge from these data. Birds do not
that holds for one animal may not hold for show any gait transitions as do mammals.
another, or even for that same individual The metabolic cost of running increases
on a different day (see Johnson and Ges- linearly as a function of speed, and the
saman [1973] for summary of data). Both parameters in this linear relationship (the
of these problems imply that heart rate may slope and the Y-intercept) vary inversely in
not provide an unambiguous index to oxy- magnitude with body mass. Thus, for birds
gen consumption. Johnson and Gessaman which are well adapted to terrestrial loco(1973) concluded that the relationship motion (including, among the birds which
between heart rate and metabolic rate must have been studied, galliforms, the rhea and
be rigorously quantified in captive birds ostrich, a tinamou, a plover, and a roadbefore it can be used as an index to free- runner), it is possible to predict the metaliving metabolic rate in those same indi- bolic cost of terrestrial locomotion based
viduals. This has since been accomplished on the body mass of the animal (Mb, in kg)
in just one study (Wooley and Owen, 1978). and the running speed (s, in m/sec). Taylor
et al. (1982) have calculated a predictive
equation which incorporates both of these
Fat loss
If a single activity is sustained for a rel- variables:
atively long duration without feeding, then
it can be possible to estimate the energy Vo2(ml/kg s)
= 0.279Mb-°-246 0.566Mb-°285-s (1)
expenditure during that time period from
3
TIME AND ENERGY BUDGETS OF FREE-LIVING BIRDS
833
It is clear, however, that not all birds are
2.5-j
geometrically similar, and not all birds are
2.0
equally well adapted for movement on land.
Several different types of avian terrestrial
1.5
locomotion have now been studied which
(ml/kg
s)
do not fit the mathematical relationship
1.0
described above. The marabou stork (Leptoptilos crumeniferus) is a long-legged walker,
0.5
and its cost of transport is much greater
than predicted in comparison to a galli0 0.03 0.06 0.09 0.12 0.15
form of similar mass (the slope of the metSPEED (m/s>
abolic rate/running speed regression is
eight times greater for the stork than for FIG. 3. The measured energy costs of hopping (open
a similar-sized turkey [Fedak et al, 1974; symbols and upper line), and the predicted (from Taylor et al., 1982) costs of running at the same speeds
Bamford and Maloiy, 1980]).
(solid symbols and lower line), for the 19-g European
Waddling birds, including penguins and robin (Erithacus rubecula, circles), the 57-g dipper (Cina goose, also have higher costs of transport clus cinclus, triangles), and the 28-g white-crowned
than do similar-sized galliforms (Fedak et sparrow (Zonotrichia leucophrys, solid lines). Hopping
a consistently higher energy cost than running.
al, 1974, Pinshow et al., 1977). These has
Data from Paladino and King (1984), Bryant et al.
species obviously move relatively awk- (1985), and Tamer and Bryant (1986).
wardly on land, and the "cost of waddling"
is apparently the price paid for efficiency
in other forms of locomotion (aerial or sources. First, allometric analysis reveals a
underwater flight or surface swimming).
systematic difference between flight costs
The energy cost of terrestrial hopping, measured in the laboratory (using wind
a form of locomotion used particularly by tunnels) and those measured in the field
many passerine birds, also consistently (using doubly labeled water or estimates of
exceeds the cost of walking (Fig. 3).
fat loss during flight)—wind tunnel studies
Terrestrial locomotion in the field is yield flight costs fully 50% greater than
more complex than running on a constant- field studies (Masman, 1986). This probaspeed horizontal treadmill. Slopes and bly results from both the aerodynamics of
evenness of the terrain, as well as speed, the wind tunnels and the extra load imposed
can vary continually in natural situations. on birds by the measuring equipment
One study (discussed below, Foraging) has (masks and tubing). Second, morphologiexamined the energetics of terrestrial for- cal adaptations, such as reduced wing loadaging in the field. No other studies of the ing and high aspect ratio, contribute to
energetics of natural terrestrial locomo- lower costs of flight in some species. These
adaptations are particularly evident in
tion are available.
highly aerial birds, such as swifts, hirunFlight
dines, and sooty terns, and flight in these
Studies of energetics have paid more species costs just 3-5 times BMR (Hails,
attention to flight than to any other avian 1979; Turner, 1983; Flint and Nagy, 1984).
activity. According to a recent review Some of this lower cost may result from
(Masman and Klaasen, 1987), the energy mode of flight as well as from morphologcost of flight has been estimated for 58 ical adaptation. For example, at higher
wind speeds sooty terns were observed to
species from eight orders.
As expected, the energy cost of flight increase the proportion of time spent glidincreases with increasing body mass (Fig. ing (Flint and Nagy, 1984). Nevertheless,
4). Over a broad range of body masses, this it seems clear that the highly aerial species
cost for most species averages approxi- incur lower flight costs than do other birds.
In an effort to integrate the effects of
mately 11 times BMR. However, a wide
body
mass and morphology into a single
variation in flight costs exists for any given
predictive
equation, Masman (1986) calbody mass. This variation has two primary
834
DAVID L. GOLDSTEIN
respectively, than the cost of sustained flapping flight (Table 1).
° • / /
ENERGY
10EXPENDITURE
DURING
FLIGHT
(watt)
j "
O
•
10
100
BODY MASS (g)
1000
FIG. 4. The energy cost of flight derived from laboratory wind tunnel studies (open symbols) and field
studies (solid symbols). Data for highly aerial species
(swifts, hirundines, and the sooty tern) are indicated
by triangles. Redrawn, with permission, from Masman (1986). Regression lines are included only to
show differences between data sets; see the text (eq.
2) for a predictive equation relating flight costs to
body mass. See Masman and Klaasen (1987) for a
complete listing of references.
culated a regression equation relating flight
costs (Mf) to body mass (m, g) wingspan (b,
cm) and wing area (S, cm2). Only data from
free-flying birds were used. The resulting
equation—
M f = 1 7 . 3 6 m 1 013b-4.236S1.928
g)
—explains 84% of the variation in flight
costs. Masman (1986) notes that the last
two terms in this equation approximate the
inverse square of the aspect ratio [(b2/S)~2],
and hence this morphological parameter
may be of particular importance in determining flight costs.
The exponents in this empirically derived
equation describing the cost of flight differ
significantly from a theoretically derived
equation based on the same variables
(Greenewalt, 1975; Masman, 1986). Theoretically predicted costs may also differ
significantly from individual empirical values (Flint and Nagy, 1984; Tatner and
Bryant, 1986). Hence, the empirically
derived relationship described by eq. 2
appears at this time to hold the greatest
promise for predicting the cost of sustained
flapping flight in unstudied species of birds.
Finally, the energy costs of gliding or of
short duration, short distance, low speed
flights may be substantially lower or higher,
Swimming
In free-swimming black ducks (Anas rubripes), telemetry of heart rate provided an
estimated cost of surface swimming equal
to 2.2 times the resting rate (Wooley and
Owen, 1978). This was also the minimum
cost of transport for mallards (Anas platyrhynchos) swimming in a laboratory water
tunnel (measured as oxygen consumption;
Prange and Schmidt-Nielsen, 1970); the
cost for mallards increased up to 4.1 times
RMR at higher speeds. The single measurement for the cost of underwater swimming by a bird, measured using doubly
labeled water, was 9.8 times the resting
value (Nagy et al., 1984). This high cost
more resembles the energy cost of aerial
flight, as might be expected from the flightlike underwater swimming of penguins.
Foraging
Knowledge of the energy cost of acquiring and handling food can be important
for testing theories of optimal foraging as
well as for construction of energy budgets
in general. However, only a limited number of studies have addressed this issue.
For aerial foragers such as swifts and
swallows, costs of flight (discussed above)
include the cost of capturing food. This
pertains also to studies of the cost of hovering flight in hummingbirds (Pearson,
1950; Bartholomew and Lighton, 1986;
and others).
Only one study has addressed the energetics of terrestrial foraging. In this study
of Gambel's quail (Callipepla gambelii), timeenergy budgets were compared with simultaneous measurements of energy expenditure obtained using doubly labeled water
(Goldstein and Nagy, 1985). To construct
the time-energy budget, the energy cost of
foraging was assigned a metabolic cost
based on studies of the cost of transport in
other galliforms (Fedak et al., 1974). The
foraging quail alternately or simultaneously walked and fed, and the terrain
over which they travelled of course was not
as regular as a laboratory treadmill. Thus,
it was necessary in this situation to select a
TIME AND ENERGY BUDGETS OF FREE-LIVING BIRDS
835
TABLE 1. Energy costs of avian activities.
Activity
Flight
Aerial species
Range of
costs assumed
by authors
constructing
time-energy
budgets
Measured cost (xBMR)
2.7-5.7
see text
6.8
Other birds, sustained flight
-11
7.5-15.2
European robin, short flights
Gliding
see text
23
2
1.3-7
Terrestrial locomotion
varies with speed and
form of locomotion;
see text
Perch
Rest
Alert
1.0
Preen
1.6-2.3
Eat
1.7-2.2
Sing/call
Bathe
2.9
2.9
1.5-5
1.25-2.6
1.9-2.1
walking speed from which to calculate
energy costs in the field. The speed chosen
was that equivalent to the normal walking
(non-feeding) speed of the birds in the field;
the cost was 7 kj/hr, or approximately 3.5
times the resting metabolic rate (note,
though, that RMR in these birds was only
50-70% of the predicted value). The generally close agreement (average within
11%) between estimates of daily energy
expenditure based on energy budgets and
doubly labeled water suggests that the estimated cost of foraging was in fact close to
the real cost. Because of the differences
between natural foraging and treadmill
running, it is difficult to make any conclusions from this study about the separate
costs of walking and feeding in the field.
More studies of this type will certainly be
needed before we can have a more general
understanding of the costs of terrestrial
locomotion and foraging in free-living
birds.
1.3-2.5
2.5
1.7-2.8
2-10
Source (measured cost)
Hails (1979, review), Turner
(1983), Flint and Nagy (1984)
Masman (1986, review)
Tatner and Bryant (1986)
Baudinette and Schmidt-Nielsen
(1974)
Taylor (1982, review), Paladino
and King (1984), Bryant et al.
(1985), Tatner and Bryant
(1986)
Wooley and Owen (1978), Weathers et al. (1984), Buttemer et al.
(1986)
Wooley and Owen (1978), Weathers et al. (1984), Buttemer et al.
(1986)
Wooley and Owen (1978), Weathers et al. (1984), Buttemer et al.
(1986)
Wooley and Owen (1978)
Wooley and Owen (1978)
Three studies have also examined the
energy cost of eating. For the black duck
eating natural vegetation, the budgerygah
{Mellopsittacus undulatus) eating seeds
(Weathers et al., 1984), and the loggerhead
shrike {Lanius ludovidanus) eating strips of
chicken (Buttemer et al, 1986), feeding
entailed an energy expenditure approximately twice the basal level (Table 1).
Preening
The energy cost of preening, measured
in a laboratory metabolism chamber, is
approximately twice the BMR in two species
(Table 1). In free-living black ducks, this
cost has been estimated as 1.6 x BMR
(Wooley and Owen, 1978).
Standing and perching
For many birds, perching is the position
assumed at rest. However, the energy
expenditure while perching is not necessarily the resting metabolic rate; perching
836
DAVID L. GOLDSTEIN
bo
in
c
s
12
-a
c
3
n
c
'53 S5
a oo
boo
3—
i|
I
CO O
a
• 5-
Swi
1
c
be.
a
Other activities
co in
UII
1
birds may be alert and wary, and as a result
have heightened metabolic rates. In both
budgerygahs and shrikes in the laboratory,
and in black ducks in the field, alert birds
have metabolic rates approximately twice
the basal or resting value (Table 1). Other
postural changes could also entail energy
costs: the black duck while standing has a
metabolic rate estimated as 1.6 times the
resting rate.
0 q
01 oo
• £
•c
Sing
m o
to
Oi
1
to.
actiive day)
lime bu ets wi thin
s
a
ii
§•
1
to
o
to
d
iri
CO
to ai
CO
in
<N
c
t'
o o
in
a> r»
CO
CO
^*
(all
Wall
M
o o
m <j>
CO
o
fcp
o o
eed
co in
d in
CO CO
a
tan
CM in
:g
CO CO
a
erch
o o
• * to
uj
Rest
Q-
g
a
CO
00
u
ot o
to
to
to
S
O
nd ph;
nnu al cycl
V
"
o
PATTERNS OF ACTIVITY IN THE
FIELD—THE TIME BUDGET
Contributions of various activities to
the time budget
CO
r~ o>
u.
Llert
"i
o
d d
ire i
Fly
§
lion
rial lion
•S
OS CM
The energy cost of shuffling back and
forth on a perch is approximately 2.3 times
BMR for the budgerygah (Weathers et al.,
1984). Black ducks expend approximately
3 times the resting metabolic rate while
calling, bathing, or wing flapping (Wooley
and Owen, 1978). In general, several lowlevel activities entail energy expenditures
approximately 2-3 times the resting level
(Table 1). These scattered data can at least
provide a starting point for estimating
energy costs for unmeasured activities and
unstudied species.
1-°..
O4
c
.5 SL
? E
••?
8 .
i 3 ii
1
w i
U «
u
_3
•-
a.,
The time budgets of different species of
birds are complex sums of times spent in
many different activities. However, two
activities predominate in most time budgets—resting/perching, and activities
associated with feeding. The times allocated to both of these activities may vary
significantly between portions of the annual
cycle or between seasons (Table 2).
Resting, perching, or alert perching
often constitute the majority of a time budget (although the energy costs of resting
and alert perching may differ significantly,
these states are often not differentiated in
time-budget analyses). Mockingbirds spend
90% or more of their active day perching
throughout the breeding cycle (Biedenweg, 1983); this activity also predominates
in the time budgets of several other small
insectivores (Table 2; Holmes et al., 1979;
Ettinger and King, 1980). The proportion
of time spent perching is consistently
reduced during the period of raising nest-
837
TIME AND ENERGY BUDGETS OF FREE-LIVING BIRDS
lings as compared with the pre-breeding
phase (paired Mest, n = 14, / = —2.7, P <
0.02; see Appendix la for references).
Interestingly, an inter-species comparison reveals that during the non-breeding
season the proportion of time spent perching is negatively correlated with the length
of the active day (Fig. 5; n = 20, r = —0.55,
P < 0.01; see Appendix lb for references).
Thus, during shorter days, birds allocate a
greater proportion of the day to resting,
and the actual length of time spent perching varies little with a changing active day
length. The linear regression describing
Figure 5 predicts that a bird with an active
day of 8 hr will perch for 7.4 hr of that
time (92.5%), and a bird with a 16 hr active
day will perch for 8.5 hr (53.1 %). This correlation is rather weak, describing only 30%
of the variation in the data. However, the
finding is consistent with the suggestion
that individual species conserve resting time
in the face of seasonal and geographic variation in activity budgets (Herbers, 1981).
For other species, foraging predominates the time budget (Table 2). For example, yellow-billed magpies {Pica nuttalli)
spend an annual average of 56% of the
active day feeding (Verbeek, 1972). In contrast to the small insectivores mentioned
above, marsh wrens (Telmatodytes palustris)
feed for more than 50% of the active day
for most of the year, and for more than
93% of the day while raising nestlings (Verner, 1965). Semipalmated sandpipers (Calidris pusilla) feed only 30% of the day while
attending their young, but for 95% of the
day while flocking in preparation for
migration (Ashkenazie and Safriel, 1979).
The proportion of time spent flying is of
interest because of the high energy cost of
this activity. The proportion of time spent
flying varies widely between species. As
expected, it is generally highest in species
which are aerial feeders (e.g., 65.7% of the
active day in purple martins (Utter and
LeFebvre, 1973); 82.2% in the sand martin
and 76.4% in the barn swallow [Turner,
1983]). For these species, however, the
energy cost of flight is relatively low. In
other species, flight typically constitutes less
that 10% of the day, and tends to occupy
more time while rearing nestlings than
IOC75
% OF DAY
SPENT
50PERCHING
25
8
12
16
20
ACTIVE DAY LENGTH (h)
FIG. 5. The percent of the active day spent perching
as a function of active day length in 20 species. Y =
131.7 - 4 . 9 X , n = 20,r 2 = 0.30,P< 0.05. The actual
time spent perching varies little at different day lengths
(see text). Most data are for the period immediately
preceding the reproductive season, but some are for
other times of year (winter or post-reproductive).
There is no correlation in these data between the
active day length and the body mass of the bird. See
Appendix lb for references.
during the non-breeding season (paired
Mest, n = 14, t = 1.8, 0.05 < P < 0.1; see
Appendix la for references).
Other activities, such as preening,
aggression, or courtship, tend to occupy
smaller proportions (less than 5%) of avian
activity cycles. Some exceptions to this generalization have been recorded, however.
Nest building occupies 13% of the male
marsh wren's day during part of the breeding season (Verner, 1965). And singing may
occupy 50% or more of the active day in
some species during the early portion of
the breeding season (Verner, 1965; Schartz
and Zimmerman, 1971).
The effect of body size on the time budget
Small animals have high mass-specific
metabolic rates, and hence must consume
a relatively greater amount of energy per
day than larger animals. One might therefore expect some clear relations between
the proportions of times spent in different
activities and the body mass of the bird.
Such trends are difficult to discern.
Walsberg (1983) noted that small birds
tend to spend a greater proportion of time
in flight than large birds. His analysis (using
log-transformed data) used the median
proportion of time in flight for species for
which more than one value was available.
The pattern is less consistent if we use data
for specific portions of the annual cycle.
During the non-breeding season the pro-
838
DAVID L. GOLDSTEIN
portion of time in flight is only very weakly
correlated with body mass (n = 23, r =
— 0.2, P > 0.3 for untransformed data;
r = - 0 . 1 2 , P > 0.5 for log-transformed
data; see Appendix lb for references). In
contrast, during the breeding season the
two variables are related (n = 23, r = —0.56,
P < 0.01 for log-transformed data; see
Appendix lc for references).
It does seem to be the case that the birds
which spend the greatest proportion of the
day perching/resting are the smaller
species (see above). Nevertheless, no significant correlation exists between the proportion of time spent resting and body mass
(non-breeding season: n = 25, r = 0.19,
P > 0.35 for untransformed data; r = 0.16,
P > 0.4 for log-transformed data; breeding
season: n = 24, r = 0.24, P > 0.2 for
untransformed data; r = 0.18, P > 0.4 for
log transformed data; see Appendices lb
and 1 c for references for non-breeding and
breeding seasons, respectively).
Thus, in contrast to expectations based
on the allometry of energy requirements
(see, e.g., Calder, 1984), body mass appears
not to exert a consistent influence on time
budget allocation. Five-gram hummingbirds may feed for less than 10% of the day
(Stiles, 1971), whereas 1-kg ducks may feed
for more than 70% of the day (Dwyer,
1975). Rather, time budget allocation
appears to be much more strongly determined by the ecology of the animal, such
as the availability and energy content of
food (Wolf and Hainsworth, 1971; Kaminski and Prince, 1981), the portion of the
annual cycle (Table 2), and the physical
environment (temperature, rain, etc.; see
Maxson and Oring, 1980; Bryant and Westerterp, 1983; Goldstein, 1984; Masman,
1986). This variation in time budget allocation undoubtedly is reflected in variation
in energy expenditure. For example, Walsberg (1983) noted that despite the general
increase in daily energy expenditure (DEE)
with body size, a large variation in DEE
exists at any given mass. Careful study of
similar-sized birds with different patterns
of time allocation (such as the small insectivores discussed above) could yield valuable insights into the tradeoffs between
expenditure of time and energy (as has been
done for sit-and-wait vs. actively foraging
lizards [Anderson and Karasov, 1981],
whose different patterns of activity resemble those of the different insectivorous
birds).
INTEGRATION OF TIME BUDGETS AND
ACTIVITY COSTS—THE ENERGY BUDGET
Daily energy expenditure by free-living
birds can be calculated by multiplying the
durations of daily activities (hr/day) by
their respective energy costs (kj/hr). The
activities which contribute most to the
energy budget are those which occur for
the greatest proportions of time and those
with the greatest energy costs. This suggests that resting, foraging, and flight are
likely to contribute most to avian energy
budgets. Consistent with this, Walsberg
(1983) has noted that "obligatory" costs
(those associated with basal and thermoregulatory demands, and largely accrued
during periods of rest) typically constitute
40-60% of the total energy budget.
Unfortunately, the actual proportions of
energy budgets which are accounted for
by "obligatory" costs and by activity are
for the most part difficult to calculate. This
is primarily because energy costs other than
those associated with activity, particularly
costs of basal metabolism and thermoregulation, are often not accurately known.
Basal metabolic rates are often calculated
from allometric predictions, rather than
actually measured. Adequate methods for
assessing thermoregulatory demands (Bakken and Gates, 1975; Robinson^al., 1976)
have become available and widely used only
recently and are not incorporated into most
published time-energy budgets. Also, the
interaction between thermoregulatory and
activity costs is poorly understood (see
Walsberg, 1983, for discussion) and may
be a complex function of temperature and
intensity of activity (Paladino and King,
1984).
Hence, there are numerous sources of
error in constructing time-energy budgets.
These include 1) errors in measuring the
duration of activities, 2) errors associated
with the energy costs assigned to activities,
and 3) errors in assessing the energy costs
not associated with activity, particularly
TIME AND ENERGY BUDGETS OF FREE-LIVING BIRDS
basal and thermoregulatory costs. The
error inherent in the final energy budget
is a cumulative product of these several
independent sources (Travis, 1982). However, the actual magnitude of each of these
errors usually is not accurately known. Two
techniques have been used to assess the
uncertainty associated with energy budgets—comparison with doubly labeled
water measurements, and sensitivity analyses.
839
lead to significant errors in estimating total
energy expenditure. For loggerhead
shrikes, energy budgets constructed using
metabolic rates measured on the same birds
differed very little from simultaneous measures of DEE using doubly labeled water.
In contrast, substitution into the timeenergy budget of metabolic data from a
different population of shrikes, differing
by only 12% in thermal conductance, produced a 22% increase (inaccuracy) in the
resulting energy budget. This occurred
Comparisons between time-energy budgets
because the shrikes spent most of their time
(TEB) and doubly labeled water (DLW)
at temperatures below thermoneutrality
measurements
(Weathers et al, 1984). It is likely that
energy
costs of activities also vary within
The doubly labeled water method provides a direct and accurate measure of daily and between populations.
energy expenditure against which energy
Second, it is essential to use a robust
budgets can be compared. Simultaneous meteorological model. The important
measurements using these two approaches effects of wind and solar radiation on avian
have now been made both in the field metabolic rates are now well documented.
(Utter, 1971 [in Weathers et al, 1984]; Ignoring these effects in time-energy budTurner, 1983; Williams and Nagy, 1984; gets can produce significant inaccuracies in
Bryant et al., 1985; Goldstein and Nagy, the estimate of DEE. For example, ignor1985; Masman, 1986) and in aviaries ing the effects of wind on metabolic rate
(Weathers and Nagy, 1980; Weathers et al, resulted in a 15% underestimate of DEE
1984; Buttemer et al, 1986). These studies in aviary-housed budgerygahs (Buttemer et
permit the evaluation of two questions al, 1986). Accurate continuous assessment
(Weathers et al, 1984; Masman 1986). First, of the microclimates occupied by free-livare time-energy budgets accurate {i.e., do ing birds is a much greater challenge, parthey, on average, compare favorably with ticularly for mobile species. Despite this,
simultaneous DLW measurements)? Sec- this approach has proven successful in sevond, are time-energy budgets sensitive eral studies (Mugaas and King, 1981; Bieenough to evaluate individual variation in denweg, 1983; Goldstein, 1984;Stalmaster
energy expenditure {i.e., do TEB and DLW and Gessaman, 1984; Masman, 1986).
estimates of daily energy expenditure
Of the six studies in which it is possible
[DEE] correlate closely)?
to examine the correlation between DLW
In general, the agreement (on average) and TEB estimates of DEE, three yield sigbetween DLW and TEB estimates of DEE nificant correlations (Weathers^al, 1984;
is good (within 10% for most of the stud- Buttemer «/ al, 1986; Masman ?/ al, 1986).
ies). Thus, it appears that time-energy bud- Of the three for which no significant corgets can yield accurate estimates of DEE. relation is found, two failed to use robust
However, the best agreement occurs only meteorological models (Weathers and
when certain rigorous qualifications are Nagy, 1980; Williams and Nagy, 1984); the
met. First, it is important to use measured lack of correlation in the third study may
metabolic data from the population under have resulted from the limited range of
study. Basal metabolic rate and thermal DEE's measured on Gambel's quail (Goldconductance often differ significantly from stein and Nagy, 1985). Goldstein and Nagy
allometrically predicted values; they can (1985) also noted that inaccuracies in inalso vary between seasons and populations dividual time-energy budgets can be
(Fig. 2; Hudson and Kimzey, 1966; accounted for in part by individual variaWijnandts, 1984). Even subtle inaccuracies tion in BMR or RMR. For Gambel's quail,
in assigning thermoregulatory costs can the coefficient of variation for measures of
840
DAVID L. GOLDSTEIN
RMR (approx. 15%) exceeded that for the
measures of DEE (6%). Thus, it is indeed
possible for time-energy budgets to track
individual variation in energy expenditure,
but rigorous analyses are needed to achieve
this sensitivity. And of course, it is necessary that time budgets be constructed for
individual birds; individual activity patterns can vary considerably (Rijnsdorp et
al., 1981), but following individuals for
complete time budgets can be difficult (see
below).
Sensitivity analyses
In a sensitivity analysis of a time-energy
budget one assesses the effect of varying
the input values (such as the cost of flight)
on the resultant energy budget. Sensitivity
analyses could be accomplished for any of
the sources of error listed above (activity
durations, activity costs, non-activity costs).
A major uncertainty in constructing timeenergy budgets could be the inaccuracy in
measured durations of activities. This error
has two sources. First, it is often difficult
or impractical to follow individual birds for
entire days. Consequently, time budgets are
often calculated from averages of observations on several birds, and often for
observation periods which are assumed to
be representative of the entire day. Second, there is some inaccuracy inherent in
the actual timing of activities as they are
observed; this is likely to vary with the
method of recording, the difficulty in
observing the bird, and the narrowness or
breadth of the activity categories. The
errors inherent in various methods of sampling behavior are well documented (e.g.,
Altmann, 1974). Nevertheless, this source
of error has been neglected in the literature on avian energy budgets. Only a small
number of studies present time budgets for
individual birds (Hainsworth, 1977; Walsberg, 1978; Williams and Nagy, 1984;
Goldstein and Nagy, 1985), and, to my
knowledge, no study presents any estimate
of confidence limits on the time budget (as
could be done, for example, by having more
than one observer independently construct
time budgets for the same birds). No sensitivity analyses have been published which
examine the error associated with inaccurate time budgets.
The inaccuracy in assigning energy costs
to activities has received a bit more attention. The range of activities and species for
which energy costs have been estimated is
quite limited. Thus, for most studies it is
necessary to estimate the energy costs of
activities (Table 1). Typically, sensitivity
analyses examine the effect of a 10-20%
inaccuracy in these estimates. As expected,
energy budgets are most sensitive to errors
in the costs of high-energy activities, such
as locomotion (Walsberg, 1978; Turner,
1983; Bryant et al., 1985), or high-duration
activities, such as perching or feeding
(Ettinger and King, 1980; Biedenweg,
1983; Gauthier et al, 1984). It is worth
noting that the actual inaccuracy could be
considerably larger than 25%. For example, the cost of alert perching is approximately twice the resting metabolic rate,
whereas quiet perching could require little
more than the resting energy expenditure.
These two states may not be easy to distinguish in the field, and the consequences to
an energy budget could be significant.
As discussed above, errors in estimating
thermoregulatory costs may also be important. Not all time-energy budgets account
for the influence of ambient temperature
on metabolic expenditure (see Walsberg,
1983, for a review of these), and far fewer
account for other meteorological variables
(e.g., solar radiation, wind). Furthermore,
as noted above, the interaction of thermoregulatory and activity costs is poorly
understood. Thermoregulation has been
estimated to comprise as much as 20% of
the total energy budget (Mahoney, 1976,
cited in Ettinger and King, 1980). As
expected, energy budgets for species which
spend much of their day resting are more
sensitive to variation in these costs than
those which spend most of the day active
(Ettinger and King, 1980; Turner, 1983).
However, even this conclusion is premised
on the assumption that activity costs substitute for thermoregulatory costs during
activity.
SOME GENERAL CONCLUSIONS ABOUT
ENERGY EXPENDITURE BY
FREE-LIVING BIRDS
Numerous estimates of avian daily energy
expenditure have now been published, and
841
TIME AND ENERGY BUDGETS OF FREE-LIVING BIRDS
these have been subjected to allometric
analyses several times. The most recent of
these (Walsberg, 1983) concluded that the
exponent relating DEE to body mass (m,
g) was lower (DEE <* m061) than the exponent relating BMR to body mass (BMR oc
mo.7-o.8) This implies that smaller species
have field metabolic rates (DEE) which are
a greater multiple of BMR than do larger
species. Walsberg attributed this finding to
the observation that small birds spend
greater proportions of their active day in
an energy-intensive activity, flight. However, the data set analysed by Walsberg
contained many time-energy budget estimates of DEE which did not accurately
account for thermoregulatory costs. Meteorological influences on DEE are likely to
be more pronounced for small animals than
large ones, and this could influence the
relation between DEE and body mass. To
examine this, I analysed the relationship
between DEE and body mass for species
which have been studied either with doubly labeled water or with analysis of excretion and assimilation rates. These studies
should include accurate assessments of
energy expenditure for thermoregulation.
The resulting mass dependence of DEE
(DEE oc m0-53, Fig. 6) is indeed less steep
than suggested by Walsberg's (1983) analysis, but the two regression lines do not
differ significantly in slope (P < 0.05,
ANCOVA) and cross at a body mass of
approximately 100 g. Whether this result
implies any consistent error in estimating
thermoregulatory or other costs is difficult
to say.
Over a broad scale, the relation between
DEE and body mass is not different for
birds feeding nestlings than for other birds
(Fig. 6). However, this probably reflects
the range of inter-specific variation in DEE,
as well as the condensed logarithmic scale.
For a given species, DEE does vary seasonally. Walsberg (1983) tabulated data for
variation of DEE through the reproductive
cycle for eight species. Energy expenditure
for female birds typically peaks during egg
synthesis, whereas this peak varies for
males. Energy expenditure typically reaches
a minimum for males during incubation,
but this period varies for females.
Drent and Daan (1980) suggested that
I O.OOO-i
DAILY ENERGY
EXPENDITURE
<kJ)
100
1000
10.000
BODY MASS (9)
FIG. 6. Daily energy expenditure as a function of
body mass in birds not engaged in breeding activities
(solid symbols) and those feeding nestlings (open symbols). The solid line represents the bestfitleast squares
linear regression of all data points (Log DEE = 1.29 +
0.55 log mass, n = 21, r2 = 0.91, P < 0.01). The
regressions for the two individual data sets (breeding
and non-breeding) are not statistically different from
the regression for all points or from each other
(ANCOVA, P > 0.05). All data were collected by
direct means (either doubly labeled water or measures
of energy excretion and assimilation rates). Nineteen
species are represented; data for male and female
purple martins (Progne subis) and house martins (Delichon urbica) are plotted separately. See Appendix Id
for references.
for several species DEE reaches a maximum of approximately 4 x BMR, and that
this may represent a general maximum sustainable rate of energy expenditure for
birds. This ceiling is similar to the maximum rate of energy expenditure by birds
in a variety of other energetically stressful
situations (Kirkwood, 1983). Together,
these findings suggest that DEE may be
limited by the physiological capacities of
the bird to acquire or assimilate energy.
Studies of the digestive physiology of hummingbirds suggest that, for these birds,
movement of sugars through and across
the alimentary tract may be the rate limiting step in energy expenditure (Karasov
et al., 1986). Such constraints on energy
processing have already been incorporated
into various models of optimal foraging
(Schoener, 1983; Belovsky, 1986). During
such periods of maximal energy expenditure, behavioral responses to changes in
weather or resource availability are likely
to be critical in balancing energy budgets.
Evans (1976), for example, notes that during times of particular cold, the costs of
foraging by sandpipers may exceed the
energy gained therefrom, and the birds
842
DAVID L. GOLDSTEIN
actually forego foraging on such days. Further quantitative analyses of time and
energy use during such potentially stressful
portions of the annual cycle should be an
important tool for testing hypotheses about
interactions between behavior and energetics.
King (1974) concluded his review of time
and energy budgets with a plea for, among
other things, improved micrometeorological modeling, data denning the coupling
of energy expenditure to the microclimate,
and data describing energy costs of activities. This plea has been vigorously
answered in the ensuing 12 years. Of
course, gaps persist in these areas, particularly in denning the energy costs of various activities. However, the tools now exist
for constructing rigorous time-energy
budgets. If possible, validation with doubly
labeled water can only strengthen the confidence in these estimates of DEE. Applications of these techniques to a range of
species, and especially to individual species
across changing seasons, habitats, and
physical conditions, will likely reveal novel
and important interactions between behavior, fitness, and the physical and biotic
environments.
consumption during hover-feeding in free-ranging Anna hummingbirds. J. Exp. Biol. 123:191199.
Bartholomew, G. A., D. Vleck, and C. M. Vleck.
1981. Instantaneous measurements of oxygen
consumption during pre-flight warm-up and postflight cooling in sphingid and saturniid moths. J.
Exp. Biol. 90:17-32.
Baudinette, R. V. and K. Schmidt-Nielsen. 1974.
Energy cost of gliding flight in herring gulls.
Nature 248:83-84.
Belovsky, G. E. 1986. Optimal foraging and community structure: Implications for a guild of generalist grassland herbivores. Oecologia 70:35-52.
Bernstein, N. P. and S.J. Maxson. 1985. Reproductive energetics of blue-eyed shags in Antarctica.
Wilson Bull. 97:450-462.
Biedenweg, D. W. 1983. Time and energy budgets
of the mockingbird (Mimus polyglotlos) during the
breeding season. Auk 100:149-160.
Bryant, D. M., C. J. Hails, and R. Prys-Jones. 1985.
Energy expenditure by free-living dippers (Cinclus cinclus) in winter. Condor 87:177-186.
Bryant, D. M. and K. R. Westerterp. 1983. Shortterm variability in energy turnover by breeding
house martins Delichon urbica: A study using doubly-labeled water. J. Anim. Ecol. 52:525-543.
Burger, A. E. 1981. Time budgets, energy needs,
and kleptoparasitism in breeding lesser sheathbills (Chionis minor). Condor 83:106-112.
Buttemer, W. A., A. M. Hayworth, W. W. Weathers,
and K. A. Nagy. 1986. Time-budget estimates
of avian energy expenditure: Physiological and
meteorological considerations. Physiol. Zool. 59:
131-149.
Calder, W. A., III. 1984. Size, function, and life history.
ACKNOWLEDGMENTS
The author was supported during the
preparation of this manuscript by NRSA
fellowship 1 F32 AM07501-01 Al and by
NSF grant PCM-83-16262 to Eldon J.
Braun.
Harvard University Press, Cambridge, Massachusetts.
Costa, D. P., P. Dann, and W. Disher. 1986. Energy
requirements of free ranging little penguin,
Eudyptula minor. Comp. Biochem. and Physiol.
85A:135-138.
Dolnik, V. R. and V. M. Gavrilov. 1973. Energy
metabolism during flight of some passerines. In
B. E. Bykovskii (ed.), Bird migrations, ecological and
REFERENCES
Altmann.J. 1974. Observational study of behavior:
Sampling methods. Behaviour 49:227-267.
Anderson, R. A. and W. H. Karasov. 1981. Contrasts
in energy intake and expenditure in sit-and-wait
and widely foraging lizards. Oecologia 49:67-72.
Ashkenazie, S. and U. Safriel. 1979. Time-energy
budget of the semipalmated sandpiper Calidris
pusilla at Barrow, Alaska. Ecology 60:783-799.
Bakken, G. S. and D. M. Gates. 1975. Heat transfer
analysis of animals: Some implications for field
ecology, physiology, and evolution. In D. M. Gates
and R. B. Schmerl (eds.), Perspectives in biophysical
ecology, pp. 255-290. Springer, New York.
Bamford, O. S. and G. M. O. Maloiy. 1980. Energy
metabolism and heart rate during treadmill exercise in the Marabou stork. J. Appl. Physiol. 49:
491-496.
Bartholomew, G. A. and J. Lighton. 1986. Oxygen
physiologicalfeatures, pp. 288-296. Halstead Press,
New York.
Drent, R. H.andS. Daan. 1980. The prudent parent:
Energetic adjustments in avian breeding. Ardea
68:225-252.
Dwyer, T. J. 1975. Time budgets of breeding gadwalls. Wilson Bulletin 87:335-343.
Ettinger, A. O. andj. R. King. 1980. Time and energy
budgets of the willow flycatcher (Empidonax traillii) during the breeding season. Auk 97:533-546.
Evans, P. R. 1976. Energy balance and optimal foraging strategies in shorebirds: Some implications
for their distributions and movements in the nonbreeding season. Ardea 64:117-139.
Fedak, M. A., B. Pinshow, and K. Schmidt-Nielsen.
1974. Energy cost of bipedal running. Am. J.
Physiol. 227:1038-1044.
Flint, E. N. and K. A. Nagy. 1984. Flight energetics
of free-living sooty terns. Auk 101:288-294.
Gauthier, G.,J. Bedard, and Y. Bedard. 1984. Com-
TIME AND ENERGY BUDGETS OF FREE-LIVING BIRDS
parison of daily energy expenditure of greater
snow geese between two habitats. Can. J. Zool.
62:1304-1307.
Gill, F. B. and L. L. Wolf. 1975. Economics of feeding territoriality in the golden-winged sunbird.
Ecology 56:333-345.
Goldstein, D. L. 1984. The thermal environment
and its constraint on activity of desert quail in
summer. Auk 101:542-550.
Goldstein, D. L. and K. A. Nagy. 1985. Resource
utilization by desert quail: Time and energy, food
and water. Ecology 66:378-387.
Greenewalt, C. H. 1975. The flight of birds. Trans.
Am. Phil. Soc. 65:1-67.
Hails, C.J. 1979. A comparison of flight energetics
in hirundines and other birds Comp. Biochem.
Physiol. 63A:581-585.
Hails, C.J. 1984. The breeding biology of the Pacific
swallow Hirundo tahitica in Malaysia. Ibis 126:198—
211.
Hainsworth, F. R. 1977. Foraging efficiency and
parental care in Colibri coruscans. Condor 79:6975.
Herbers,J. M. 1981. Time resources and laziness in
animals. Oecologia 49:252-262.
Holmes, R. T., C. P. Black, and T. W. Sherry. 1979.
Comparative population bioenergetics of three
insectivorous passerines in a deciduous forest.
Condor 81:9-20.
Hudson, J. W. and S. L. Kimzey. 1966. Temperature
regulation and metabolic rhythms in populations
of the house sparrow, Passer domesticus. Comp.
Biochem. Physiol. 17:203-217.
Johnson, S. F. and J. A. Gessaman. 1973. An evaluation of heart rate as an indirect monitor of freeliving energy metabolism. In). A. Gessaman (ed.),
843
Masman, D. and M. Klaasen. 1987. Energy expenditure during free flight in trained and free-living
kestrels, Falco tinnunculus. Auk 104:603-616.
Maxson, S. J. and L. W. Oring. 1980. Breeding season time and energy budgets of the polyandrous
spotted sandpiper. Behaviour 74:200-263.
Moss, R. 1973. The digestion and intake of winter
foods by wild ptarmigan in Alaska. Condor 75:
293-300.
Mugaas.J. N. andj. R. King. 1981. Annual variation
of daily expenditure by the black-billed magpie.
Stud. Avian Biol. Number 5, 1-78.
Nagy, K. A., W. R. Siegfried, and R. P. Wilson. 1984.
Energy utilization by free-ranging jackass penguins, Spheniscus demersus. Ecology 65:1648-1655.
Nisbet, J. C. T. 1963. Weight loss during migration
II: A review of other estimates. Bird Banding 34:
139-159.
Obst, B. S., K. A. Nagy, and R. E. Ricklefs. 1987.
Energy utilization by Wilson's storm petrel
{Oceanites oceanicus). Physiol. Zool. 60:200-210.
Paladino, F. V. and J. R. King. 1984. Thermoregulation and oxygen consumption during terrestrial locomotion by white-crowned sparrows
Zonolrichia leucophrys gambelii. Physiol. Zool. 57:
226-236.
Pearson, O. P. 1950. The metabolism of hummingbirds. Condor 52:145-152.
Pearson, O. P. 1954. The daily energy requirements
of a wild Anna hummingbird. Condor 56:317322.
Pinshow, B., M. A. Fedak, and K. Schmidt-Nielsen.
1977. Terrestrial locomotion in penguins: It costs
more to waddle. Science 195:592-594.
Prange, H. D. and K. Schmidt-Nielsen. 1970. The
metabolic cost of swimming in ducks. J. Exp. Biol.
Ecological energetics of homeotherms, pp. 4 4 - 5 4 . Utah
53:763-777.
University Press, Logan.
Pyke, G. H. 1979. The economics of territory size
and time budget in the golden-winged sunbird.
Kaminski, R. M. and H. H. Prince. 1981. Dabbling
Am. Nat. 114:131-145.
duck activity and foraging responses to aquatic
invertebrates. Auk 98:115-126.
Roby, D. D. and R. E. Ricklefs. 1987. Energy expenditure in adult least auklets and diving petrels
Karasov, W. H., D. Phon.J. M. Diamond, and F. L.
during the chick-rearing period. Physiol. Zool.
Carpenter. 1986. Food passage and intestinal
60:661-678.
nutrient absorption in hummingbirds. Auk 103:
453-464.
Reyer, H.-U. and K. Westerterp. 1985. Parental
energy expenditure: A proximate cause of helper
King, J. R. 1974. Seasonal allocation of time and
recruitment in the pied kingfisher (Ceryle rudis).
energy resources in birds. In R. A. Paynter, Jr.
Behav. Ecol. Sociobiol. 17:363-369.
(ed.), Avian energetics, pp. 4-85. Nuttall Ornithological Club, Cambridge, Massachusetts.
Rijnsdorp, A., S. Daan, and C. Dijkstra. 1981. Hunting in the Kestrel, Falco tinnunculus, and the adapKirkwood.J. K. 1983. A limit to metabolisable energy
tive significance of daily habits. Oecologia 50:
intake in mammals and birds. Comp. Biochem.
391-406.
Physiol. 75A:l-3.
Robinson, D. E., G. S. Campbell, andj. R. King. 1976.
Koplin.J. R., M. W. Collopy, A. R. Bammann, and
An evaluation of heat exchange in small birds. J.
H. Levenson. 1980. Energetics of two wintering
Comp. Physiol. 105:153-166.
raptors. Auk 97:795-806.
Levenson, H. 1980. Time and activity budget of Schartz, R. L. andj. L. Zimmerman. 1971. The time
and energy budget of the male dickcissel (Spiza
ospreys nesting in northern California. Condor
atnericana). Condor 73:65—76.
81:364-369.
Lundberg, P. 1985. Time-budgeting by starlings Schoener, T. W. 1983. Simple models of optimal
feeding-territory size: A reconciliation. Am. Nat.
Sturnus vulgaris: Time minimizing, energy maximizing and the annual cycle organization. Oeco121:608-629.
logia 67:331-337.
Siegfried, W. R., A. E. Burger, and P. G. H. Frost.
1976. Energy requirements for breeding in the
Masman, D. 1986. The annual cycle of the kestrel,
maccoa duck. Ardea 64:171-191.
Falco tinnunculus. Drukkerij van Denderen B.V.,
Groningen, The Netherlands.
Soltz, R. L. 1984. Time and energy budgets of the
844
DAVID L. GOLDSTEIN
red-tailed hawk in southern California. Southwest. Nat. 29:149-156.
Stalmaster, M. V. and J. A. Gessaman. 1984. Ecological energetics and foraging behavior of overwintering bald eagles. Ecol. Monogr. 54:407-428.
Stiles, F. G. 1971. Time, energy, and territoriality
of the Anna hummingbird (Calypte anna). Science
173:818-821.
Tarboton, W. R. 1978. Hunting and the energy budget of the black-shouldered kite. Condor 80:8891.
Tatner, P. and D. M. Bryant. 1986. Flight cost of a
small passerine measured using doubly labeled
water: Implications for energetic studies. Auk 103:
169-180.
Taylor, C. R., N. C. Heglund, and G. M. O. Maloiy.
1982. Energetics and mechanics of terrestrial
locomotion. J. Exper. Biol. 97:1-21.
Travis, J. 1982. A method for the statistical analysis
of time-energy budgets. Ecology 63:19-25.
Turner, A. K. 1983. Time and energy constraints
on the brood size of swallows, Hirundo rustica,
and sand martins, Riparia riparia. Oecologia 59:
331-338.
Utter.J. M. andE. A. LeFebvre. 1973. Daily energy
expenditure of purple martins (Progne subis) during the breeding season: Estimates using D2O18
and time budget methods. Ecology 54:597-603.
Verbeek, N. A. M. 1972. Daily and annual time budget of the yellow-billed magpie. Auk 89:567-582.
Verner,J. 1965. Time budget of the male long-billed
marsh wren during the breeding season. Condor
67:125-129.
Wakely,J. S. 1978. Activity budgets, energy expenditures, and energy intakes of nesting ferruginous hawks. Auk 95:667-676.
Walsberg, G. E. 1978. Brood size and the use of time
and energy by the phainopepla. Ecology 59:147153.
Walsberg, G. E. 1983. Avian ecological energetics.
In D. S. Farner, J. R. King, and K. C. Parkes
(eds.), Avian biology, Vol. 7, pp. 161-220. Academic Press, New York.
Weathers, W. W., W. A. Buttemer, A. M. Hayworth,
and K. A. Nagy. 1984 An evaluation of timebudget estimates of daily energy expenditure in
birds. Auk. 101:459-472.
Weathers, W. W. and K. A. Nagy. 1980. Simultaneous doubly labeled water ( ! HH"O) and timebudget estimates of daily energy expenditure in
Phainopepla nilens. Auk 97:861-867.
Westerterp, K. R. and D. M. Bryant. 1984. Energetics of free existence in swallows and martins
(Hirundinidae) during breeding: A comparative
study using doubly labeled water. Oecologia 62:
376-381.
Wijnandts, H. 1984. Ecological energetics of the longeared owl (Asio otus). Ardea 72:1-92.
Williams, J. B. and K. A. Nagy. 1984. Daily energy
expenditure of savannah sparrows: Comparison
of time-energy budget and doubly-labeled water
estimates. Auk 101:221-229.
Wolf, L. L. and F. R. Hainsworth. 1971. Time and
energy budgets of territorial hummingbirds.
Ecology 52:980-988.
Wooley, J. B. and R. B. Owen. 1978. Energy costs
of activity and daily energy expenditure in the
black duck. J. Wildl. Manage. 42:739-745.
APPENDIX 1
REFERENCES FOR INTRA- AND
INTER-SPECIES COMPARISONS OF
TIME AND ENERGY BUDGETS
Intraspecific comparison of time budgets,
breeding vs. non-breeding seasons
Ashkenazie and Safriel, 1979; Bernstein
and Maxson, 1985; Biedenweg, 1983;
Ettinger and King, 1980; Hails, 1984;
Holmes et al., 1979; Levenson, 1980;
Lundberg, 1985; Masman, 1986; Maxson
andOring, 1980; Mugaas and King, 1981;
Soltz, 1984.
Time budgets, non-breeding season
Ashkenazie and Safriel, 1979; Bernstein
and Maxson, 1985; Biedenweg, 1983;
Bryant et al., 1985; Dwyer, 1975; Ettinger
and King, 1980; Goldstein and Nagy, 1985;
Hails, 1984; Holmes et al., 1979; Koplin
et al., 1980; Levenson, 1980; Lundberg,
1985; Masman, 1986; Maxson and Oring,
1980; Mugaas and King, 1981; Pearson,
1954; Siegfried et al, 1976; Soltz, 1984;
Stalmaster and Gessaman, 1984; Tarboton, 1978; Wolf and Hainsworth, 1971.
Time budgets, feeding nestlings
Ashkenazie and Safriel, 1979; Bernstein
and Maxson, 1985; Biedenweg, 1983;
Burger, 1981; Ettinger and King, 1980;
Hails, 1984; Hainsworth, 1977; Holmes et
al., 1979; Levenson, 1980; Lundberg,
1985; Masman, 1986; Maxson and Oring,
1980; Mugaas and King, 1981; Schartz and
Zimmerman, 1971; Soltz, 1984; Turner,
1983; Utter and LeFebvre, 1973; Verbeek, 1972; Wakely, 1978; Walsberg, 1978;
Williams and Nagy, 1984.
Daily energy expenditure measured bydirect methods
Bryant et al., 1985; Costa et al., 1986;
Flint and Nagy, 1984; Goldstein and Nagy,
1985; Moss, 1973; Obst etal, 1987; Reyer
and Westerterp, 1985; Roby and Ricklefs,
1986; Turner, 1983; Utter and LeFebvre,
1973; Westerterp and Bryant, 1984;
Wijnandts, 1984; Williams and Nag)-, 1980.