Hummingbirds obtain most of their dietary calories from floral nectar ingested during hovering flight. Despite the importance of dietary sugar to hummingbird metabolism, the turnover of newly ingested carbon in the pool of actively metabolized substrates has not been adequately characterized in hovering hummingbirds. By combining respirometry with stable carbon isotope analysis of respired breath, we show that in rufous (Selasphorus rufus) and Anna's (Calypte anna) hummingbirds at high foraging frequencies, utilization of newly ingested sugars increased over a period of 30–45 min until it accounted for virtually 100% of the fuel oxidized. This newly ingested sugar disappears from the actively metabolized pool of substrates over a similar time course. These results demonstrate that turnover of carbon in the pool of actively metabolized substrates is rapid; carbon from ingested sucrose is available for oxidation for 30–45 min before being cleared. By monitoring expired CO2 for the appearance and disappearance of the signature characteristic of newly ingested sugar and then calculating energy budgets using video recordings of hummingbird activity, we estimated the proportion of recently ingested sugar used to fuel ongoing metabolism as well as the proportion devoted to energy storage. Consistent differences between species in the percentage of ingested cane sugar oxidized during the 2 h experimental periods suggest that individuals of each species adopted energy intake patterns appropriate to their needs. This approach provides a means by which to examine the partitioning of dietary carbon intake between energy expenditure and storage using non-invasive, field-compatible techniques.

Hummingbirds have served as exceptional model organisms for the study of foraging energetics due to their high visibility and cooperative nature in the wild, ease of use in the laboratory, high rates of energy intake and expenditure, and because most of the calories they ingest come from simple sugars in floral nectar. As hovering hummingbirds transition from a fasted to a fed state, they rapidly switch from oxidizing fatty acids to oxidizing carbohydrates (Suarez et al.,1990; Welch et al.,2006). Using broad-tailed hummingbirds Selasphorus platycercus, Welch et al. (Welch et al., 2006) found that recently ingested sugars can provide virtually all the fuel required for energy metabolism during repeated bouts of foraging. An important question concerns the rate of turnover of recently ingested sugars in the pool of actively metabolized substrates under such ecologically relevant circumstances. It is possible that ingested sugars remain as part of a pool of actively metabolized substrates for an extended period and that more recently ingested sugar molecules simply mix with this pool. Alternatively, ingested sugars may remain in the pool of actively metabolized substrates for a brief period of time, being rapidly replaced by newly ingested sugar molecules as foraging proceeds. Carleton et al.(Carleton et al., 2006)previously estimated the turnover rate of carbon in endogenous stores of energy (fat) over a period of several days, but this work was not designed to examine carbon turnover at the shorter time scales relevant to birds engaged in repeated foraging bouts. The study presented here expands on these previous studies and quantifies the turnover rate of ingested sugar in the pool of actively metabolized substrates in actively foraging rufous (Selasphorus rufus) and Anna's (Calypte anna) hummingbirds.

Recent work with broad-tailed hummingbirds S. platycercus(Welch et al., 2006) has shown the feasibility of determining the contribution of ingested fuel to the fueling of oxidative metabolism during hovering flight. The methodology employed in this and previous studies(Carleton et al., 2006; Carleton et al., 2004; Welch et al., 2006) takes advantage of the difference in stable isotopic signature of carbon in sugar derived from two distinct plant sources. Sugar from sugar beets displays a carbon stable isotope ratio particular to plants with a C3 photosynthetic pathway, while sugar from sugar cane displays a distinct carbon stable isotope ratio particular to plants with a C4 photosynthetic pathway. By varying the availability of sugar from each source and then tracking the isotopic composition of the CO2 expired by the birds, it is possible to determine the contribution of dietary sugar to oxidative metabolism. These studies benefit from the ability to conduct repeated measurements over time,and may be conducted under natural field conditions. Because hummingbirds absorb nearly all of the sugar they ingest(Karasov et al., 1986; McWhorter et al., 2006), sugar not oxidized is reserved for energy storage by conversion to glycogen or fat. Thus, by monitoring sugar intake, and by tracking the use of ingested sugar via stable isotope analysis of expired CO2, it is possible to estimate rates of net energy storage non-invasively using the same individual.

Thus, our goals in the studies reported here were: first, to determine the timing and extent to which Anna's and rufous hummingbirds support hovering flight with newly ingested sugars. We hypothesize that the ability to rely primarily on recently ingested sugar to fuel oxidative metabolism during flight is a trait common to hummingbirds. If so, then Anna's and rufous hummingbirds should oxidize recently ingested sugars as quickly and extensively as broad-tailed hummingbirds(Welch et al., 2006). Second,we wished to determine the rate of turnover of newly ingested sugar within the pool of actively metabolized substrates, while mimicking conditions experienced by foraging wild hummingbirds. We hypothesize that recently ingested sugars are removed from the pool of actively metabolized substrates approximately as quickly as they are incorporated. Third, we wished to evaluate the rate of net energy gain by Anna's and rufous hummingbirds by combining stable isotope tracking of carbon in expired CO2 with monitoring of nectar intake and energy expenditure.

We report δ13C on a per mil (‰) basis relative to the international carbon standard, Vienna Pee Dee Belemnite (VPDB), where:
\[\ {\delta}^{13}\mathrm{C}=\frac{(^{13}\mathrm{C}{/}^{12}\mathrm{C})_{\mathrm{sample}}-(^{13}\mathrm{C}{/}^{12}\mathrm{C})_{\mathrm{standard}}}{(^{13}\mathrm{C}{/}^{12}\mathrm{C})_{\mathrm{standard}}}{\times}10^{3}.\]
All solid and gas samples were submitted to the University of California,Santa Barbara Marine Science Institute Analytical Laboratory for analysis of 13C/12C ratios by mass spectrometry.

This facility utilizes a Roboprep-CN stable isotope ratio mass spectrometer(Europa Scientific, Crew, UK) equipped with an autosampler for introduction of gas samples into the continuous flow combustion chamber.

Animal care and experimental design

All hummingbirds were captured with a modified Hall trap(Russell and Russell, 2001). Individual Selasphorus rufus Gmelin 1788 (body mass at start of experiment=3.4±0.2 g; N=4, 2 male/2 female) were captured in Inyo, Mono and Santa Barbara Counties in California, USA. Individual Calypte anna Lesson 1829 (body mass at start of experiment=4.8±0.8 g; N=3, 2 male/1 female) were captured in Santa Barbara County, California, USA. Captive hummingbirds were housed at the UCSB Aviary in individual outdoor, wire-mesh enclosures measuring 1.8 m tall by 0.6 m wide by 2.4 m long. Birds were fed ad libitum on a 13% (w/v)solution of Nektar-Plus (Guenter Enderle, Tarpon Springs, FL, USA)supplemented with beet sugar (5% w/v). The δ13C value of the maintenance diet was –25.84±0.11‰ (N=10; Table 1). Birds were subjected to natural photophase and ambient weather conditions. Capture, housing and experimental protocols were approved by the University of California, Santa Barbara Institutional Animal Care and Use Committee (Protocol 672).

Data collection was conducted in an enclosure measuring 0.92 m wide×0.54 m high ×0.51 m deep, in the laboratory at an average temperature of 24.0±0.3°C. The only perch available to the hummingbird within the cage was placed on top of a balance, which was monitored in order to determine bird mass. Data collection took place between August and October of 2006 between 06:00 h and 11:00 h. Prior to each experiment, the selected hummingbird was fasted overnight to ensure that it would be oxidizing primarily fat at the beginning of the period of data collection (Suarez et al.,1990; Welch et al.,2006).

Experimental design and data collection were based largely on those reported in Welch et al. (Welch et al.,2006). Following the overnight fast of approximately 8–10 h,hummingbirds were provided with a disposable 20 ml syringe containing a solution of cane (C4 photosynthetic pathway) sugar (20% w/v). Theδ 13C value of this solution was–11.63±0.11‰ (N=10; Table 1) and it was available to the bird for approximately 1 h. The syringe containing the cane sugar solution was weighed to the nearest 0.0001 g immediately prior to and following the period that it was available to the bird. The difference in mass before and after the period it was available to the bird was taken as the mass of the solution ingested. As the density of solid sucrose is 1.587 g cm–3, a 20% w/v sucrose solution is equal to an 18.6% w/w solution. Given that the specific density of an 18.6% w/w sucrose solution is 1.07677 (Horwitz, 1975), this means that the mass of sucrose ingested (Sucingest; in g) may be determined by the following equation:
\[\ \mathrm{Suc}_{\mathrm{ingest}}=(\mathrm{Sol}_{\mathrm{ingest}}{/}{\rho}_{\mathrm{sol}}){\times}0.186,\]
where Solingest is the mass of cane sugar solution ingested by the bird (in g), ρsol is the specific gravity of an 18.6% w/w sucrose solution and 0.186 is the proportion (w/w) of the solution that is sucrose.

Immediately after removing and weighing the cane sugar solution, the hummingbird was provided with a 20% w/v beet sucrose solution. Theδ 13C value of the beet sucrose solution was–24.02±0.11‰ (N=10; Table 1). This solution was available to the hummingbird for approximately one additional hour, such that the total time allotted to this experiment for each individual was 2 h.


Hummingbirds had to hover to feed, inserting their head into a plastic tube extending from the front of the feeder. This tube was derived from a disposable 30 ml syringe and, except for the front opening, was airtight. Halfway along its length, plastic tubing was attached to the mask, allowing incurrent air to be drawn through the mask and delivered to respirometry equipment. Air first passed through a column of Drierite™ (W. A. Hammond Drierite, Xenia, OH, USA) to scrub water vapor before entering the carbon dioxide analyzer (CA-2A, Sable Systems International, Las Vegas, NV, USA). After leaving the carbon dioxide analyzer, air passed through a Drierite™–Ascarite™–Drierite™ column (Ascarite II, Arthur H. Thomas, Philadelphia, PA, USA), to scrub any carbon dioxide and additional water from the line, and then into the oxygen analyzer (FOXBOX,Sable Systems International). Air flow was maintained by a mechanism internal to the FOXBOX (thus, after the removal of water vapor) at a rate of 500 ml min–1. An infrared emitter and receiver were placed on either side of the front edge of the mask such that the infrared beam was disrupted by the presence of the hummingbird's head in the mask. By determining the length of time the infrared emitter was occluded, we were able to resolve the duration of any feeding event (and subsequent gas analysis event). The signal from the infrared receiver, along with data from the carbon dioxide analyzer,oxygen analyzer and balance, was reported to a notebook computer for recording via connection to a Universal Interface II (Sable Systems International). Data were recorded at 0.05 s intervals for 2 h using Expedata software (v. 1.0.17, Sable Systems International).

Immediately before data collection, the oxygen analyzer was calibrated with well-mixed ambient air drawn through the mask in the absence of a hummingbird. The carbon dioxide analyzer was calibrated with CO2-free nitrogen gas (zero gas) and 0.5% CO2 in nitrogen gas (Praxair, Danbury, CT,USA). In each case, tubing was removed directly downstream of the mask and held inside a small reservoir into which flowed the calibration gas at a rate in excess of the flow rate of air pulled through the respirometry system.

STP-corrected oxygen depletion and carbon dioxide enrichment associated with each feeding event were determined after first correcting by subtracting baseline values (determined as the linear extrapolation of points directly before and after the feeding event in question). These baseline-corrected data were then converted to ml of gas by application of standard equations(Withers, 1977). Determination of absolute rates of oxygen consumption and carbon dioxide production was not possible during this experiment because subsampling of incurrent air was attempted in each case (see below). However, as subsampling did not discriminate between oxygen and carbon dioxide, relative volumes (ml) of oxygen and carbon dioxide respired by the bird were determined. These were obtained by integrating the gas depletion or enrichment peak over time (min)and used to calculate respiratory quotient (RQ).

For the purpose of estimating metabolic rate of hovering hummingbirds during this experiment, complementary measurements of O2 and CO2 during hover-feeding were obtained for all individuals. These measurements were taken on the same day approximately 2 h after the experiment described above. Flow rate was held at 1200 ml min–1 and no expired breath subsamples were taken (see below). Otherwise, the methodology adopted during this complementary data collection period was identical to that described above.

Collection and analysis of expired CO2

Expired CO2 was collected while hummingbirds were hover-feeding at the respirometry mask by drawing air from the incurrent airline approximately halfway between the mask and the carbon dioxide analyzer via a 60 ml syringe (Welch et al., 2006). These samples contained both ambient CO2 as well as CO2 expired by the hummingbird. Thus, in order to estimateδ 13C of respired breath(δ13Cbreath) we used a two-part concentration-dependent mixing model adapted from Phillips and Koch(Phillips and Koch, 2002),such that:
\[\ {\delta}^{13}\mathrm{C}_{\mathrm{breath}}=[{\delta}^{13}\mathrm{C}_{\mathrm{ambient}}(f_{\mathrm{a}})+{\delta}^{13}\mathrm{C}_{\mathrm{sample}}]{/}1-f_{\mathrm{a}},\]
where δ13Csample is δ13C of air collected in the syringe. δ13Cambient is averageδ 13C of air collected at three points during the 2-h experimental period (one within the first 15 min, one near the halfway point,and one within 20 min of the end of the 2 h period) in the same manner as above when a hummingbird was not present at the mask. fais the fraction of CO2 in the gas sample from ambient air. Ambient[CO2] (p.p.m.) was determined using the carbon dioxide analyzer immediately before a feeding bout. [CO2] (p.p.m.) of the air sample was determined during stable isotope analysis by the University of California,Santa Barbara Marine Science Analytical Lab. Immediately following collection,contents of the 60 ml syringe were injected into pre-evacuated 12 ml Exetainer vials (Labco Limited, Buckinghamshire, UK) until a positive pressure was achieved. Samples were stored at room temperature for as long as 5 days before submission for analysis. All data associated with individual feeding events having δ13Cbreath values for which the CO2 concentration of the sample was not at least twice the CO2 concentration of the ambient air were excluded from further analysis.

Time and energy budgets

The activity of hummingbirds was recorded on videotape during the entirety of the 2-h experimental period. The recording period was divided into 2 min blocks for further analysis, with the first block beginning when the hummingbird first fed from the suspended feeder. Hummingbird activity was classified as either hovering/flying or perching. The proportion of each 2 min block devoted to either activity was quantified.

Energy expenditure (in ml O2) during each 2 min block was determined by multiplying the amount of time spent either hovering/flying or perching by metabolic rates associated with each activity. As described above,hovering metabolic rate (in ml O2 g–1h–1) was estimated via measurement of mass-specific oxygen consumption rate for each hummingbird during complimentary experiments. The relatively small size of the experimental enclosure greatly constrained the forward flight speed of the hummingbirds. Estimates of the oxygen consumption rate in small hummingbirds as a function of flight speed indicate a relatively flat relationship at low flight speeds, suggesting metabolic rate during hovering is equal to metabolic rate during forward flight in this range(Berger and Hart, 1972). As a result, we assume the metabolic cost of low-speed forward flight within the enclosure to be equal to the cost of hovering. We ignore the energetic costs of acceleration and deceleration, as good estimates of these do not exist. Estimates of mass-specific oxygen consumption rate (in ml O2g–1 h–1) during perching for both C. anna and S. rufus were taken from Lasiewski's seminal work(Lasiewski, 1963). These mass-specific oxygen consumption rates were multiplied by an estimate of the hummingbirds mass (as described above) during the feeding event closest in time to each 2 min period to obtain total metabolic rates (MRblock;in ml O2 h–1) for each activity for that period.

As described above, estimates of the respiratory quotient were obtained for each feeding event. Assuming hummingbirds oxidize primarily fat and/or carbohydrate (Suarez et al.,1990; Welch et al.,2006), total energy expenditure during each 2 min period(Eblock; in J) can be estimated as:
\[\ E_{\mathrm{block}}=\{[(1-\mathrm{RQ}){/}0.29][h_{\mathrm{O}_{2}(\mathrm{fat})}]+[(1-\mathrm{RQ}){/}0.29][h_{\mathrm{O}_{2}(\mathrm{carb})}]\}{\times}\{t_{\mathrm{hov}}[\mathrm{MR}_{\mathrm{block}(\mathrm{hov})}{/}3600]+t_{\mathrm{perch}}[\mathrm{MR}_{\mathrm{block}(\mathrm{perch})}{/}3600]\},\]
where RQ is the respiratory quotient for the feeding event closest to that 2 min block (constrained to be between 0.71 and 1.0), hO2(fat) is the thermal equivalent of oxygen exchange when fat is the metabolic substrate (19.8 J ml–1)(Brouwer, 1957), hO2(carb) is the thermal equivalent of oxygen exchange when carbohydrates are the metabolic substrate (21.1 J ml–1) (Brouwer,1957), thov and tperch are the duration of time spent hovering/flying and perching (in s), respectively,during that 2 min block, and MRblock(hov) and MRblock(perch) are the rates of oxygen consumption (in ml O2 h–1) for hovering/flying and perching,respectively, during that 2 min block.

Determination of cane sugar oxidation rate

A non-linear function was fitted to δ13Cbreathvalues during the first hour (feeding events for which cane sugar solution was available) and separately to δ13Cbreath values during the second hour (feeding events for which beet sugar solution was available). Thus, instantaneous estimates ofδ 13Cbreath values were possible. We assume that the incorporation of carbon into expired CO2 can be approximated by single-compartment, first-order kinetics(Carleton et al., 2006). The non-linear fitting formula is:
\[\ {\delta}^{13}\mathrm{C}_{\mathrm{breath}}(t)={\delta}^{13}\mathrm{C}_{\mathrm{breath}}({\infty})+[{\delta}^{13}\mathrm{C}_{\mathrm{breath}}(0)-{\delta}^{13}\mathrm{C}_{\mathrm{breath}}({\infty})]\mathrm{e}^{-kt},\]
where δ13C(t) is the isotope composition of the carbon in expired CO2 at time t13Cbreath(0) is the estimated initial isotope composition of the carbon in expired CO213Cbreath(∞) is the asymptotic equilibrium isotope composition of the carbon in expired CO2 and k is the fractional rate of isotope incorporation into the pool of expired CO2 (Carleton et al.,2006; Carleton and Martínez del Rio, 2005; O'Brien et al., 2000). The subscript `i' (for incorporation) will be used to indicate the application of each of these variables to the period of the experiment during which cane sugar solution is available. The subscript `d' (for disappearance) will be used to indicate the application of each of these variables to the period of the experiment during which beet sugar solution is available. For each 2 min block for which time budgets were estimated an averageδ 13Cbreath value was estimated by solving Eqn 5 with time (t) equal to the median value for that 2 min block (in min).
This δ13Cbreath value provides a means of estimating the proportion of expired CO2 derived from oxidation of exogenous carbohydrate (Carleton et al.,2006; Welch et al.,2006). Specifically, the fraction of expired CO2derived from oxidation of cane sugar (fexo) during any 2 min block was estimated as:
\[\ f_{\mathrm{exo}}=({\delta}^{13}\mathrm{C}_{\mathrm{breath}}-{\delta}^{13}\mathrm{C}_{\mathrm{C}3}){/}({\delta}^{13}\mathrm{C}_{\mathrm{C}4}-{\delta}^{13}\mathrm{C}_{\mathrm{C}3}),\]
where δ13CC4 is the δ13C value of the cane sugar solution and δ13CC3 is theδ 13C value of endogenous fuels [estimated asδ 13Cbreath(0) from Eqn 6], during the first hour of the experiment, and δ13CC3 is theδ 13C value of the beet sugar solution during the second hour of the experiment.
For each mol sucrose oxidized, 12 mol O2 are consumed (2×6 mol O2 per unit hexose). Thus, the amount of cane sugar oxidised(Mcane; in μmol) during each 2 min period may be estimated as:
\[\ M_{\mathrm{cane}}=[f_{\mathrm{exo}}(Met_{\mathrm{block}}){\times}10^{6}]{/}12,\]
where Metblock is the amount of O2 (in mol)consumed during that 2 min block.

RQ values during the first feeding event following a fast were near 0.71 in both rufous and Anna's hummingbirds (S. rufus: 0.74±0.01, N=2; C. anna: 0.76±0.02, N=3), indicating that the birds relied primarily on the oxidation of fatty acids to fuel hovering flight. After 40 min of access to food, average RQ values for each bird were approximately 1.0 and remained near this value for the remainder of the experiment in both species (S. rufus: 1.01±0.01, N=4; C. anna: 1.01±0.01, N=3), indicating that the birds had switched to fuelling hovering flight exclusively with carbohydrates.

During the first feeding bout following the fast,δ 13Cbreath values were near theδ 13C value of the maintenance diet containing beet sugar and increased over the first hour towards the δ13C value of cane sugar in the experimental diet (Fig. 1A, Table 2).δ 13Cbreath averaged–27.02±1.36‰ (N=2) in S. rufus, and–27.36±1.15‰ (N=3) in C. anna during the first feeding event. For both species, this value was more negative than, but not significantly different from, the δ13C value of the maintenance diet (S. rufus: t1=–1.2299, P=0.4346; C. anna: t2=–2.2889, P=0.1493). The fractional rate of isotopic incorporation into the pool of expired CO2 (ki) varied extensively between individuals. ki averaged 7.1±7.6%(min–1; range 0.7–16.2%; N=4) in S. rufus. ki averaged 4.5±3.7% (min–1; range 0.2–7.1%; N=3) in C. anna. During a period of availability of cane sugar solution, whenδ 13Cbreath values had reached a plateau (the period beginning 40 min after the first feeding on cane sucrose), the proportion of metabolism (i.e. CO2) fuelled by dietary cane sugar (fexo) approached 100%, similar to results shown previously in S. platycercus(Welch et al., 2006). fexo averaged 0.83±0.18 (range 0.61–1.01; N=4) for S. rufus during this steady state period of feeding on cane sugar (Fig. 1A, Table 2). fexo averaged 0.81±0.31 (range 0.46–1.03; N=3) for C. anna during this steady state period of feeding on cane sucrose (Fig. 1A, Table 2).

When the diet was switched from cane sugar back to beet sugar (the second hour of the experiment), the decrease inδ 13Cbreath over time mirrored the increase inδ 13Cbreath seen during the previous period of cane sugar feeding. There was less variability between individuals in the fractional rate of disappearance of labelled carbon in expired CO2compared to 13C enrichment curves observed during the previous hour(Fig. 1B). The fractional rate of isotopic disappearance from the pool of expired CO2(kd) averaged 10.9±2.9% (min–1;range 9.2–13.5; N=4) in S. rufus. kdaveraged 5.7±0.4% (min–1; range 5.3–5.9; N=3) in C. anna. During the period of beet sucrose availability when δ13Cbreath values had reached a plateau (the period beginning at least 40 min after the first feeding on beet sucrose) δ13Cbreath values neared theδ 13C signature of the beet sucrose solution.δ 13Cbreath averaged–23.97±0.54‰ (N=4) for S. rufus and–23.08±0.59‰ (N=3) for C. anna. These values are not significantly different from the δ13C signature of the beet sucrose solution (S. rufus: t3=0.1955, P=0.8575; C. anna: t2=2.7620, P=0.1099).

RQ and δ13Cbreath values were highly significantly correlated during the period of cane sugar availability in both rufous and Anna's hummingbirds (data pooled by species; Fig. 2; S. rufus: r20=0.9494, P<0.0001; C. anna: r17=0.9065; P<0.0001), suggesting that newly ingested sugars were fuelling metabolism during this period. By contrast,there was no significant correlation between RQ andδ 13Cbreath values for either species during the period of beet sugar availability (data pooled by species; S. rufus: r36=–0.1880, P=0.2722; C. anna: r29=0.3294, P=0.0810). As RQ remained near 1.0 during the entire period of beet sugar availability, no correlation would be expected.

Hummingbirds engaged in flight for varying amounts of time over the approximately 2-h period of video recordings beginning with their first hover-feeding event (Fig. 3, Table 3). S. rufusspent an average of 6.6±3.4% (range: 3.0–9.6%; N=4) of the time hovering or flying. C. anna spent an average of 9.6±7.4% (range: 5.3–18.1%; N=3) of the time hovering or flying. Consequently, hummingbirds expended varying amounts of energy during the approximately 2-h long period (Table 3). S. rufus expended an average of 1551±259 J(range: 1202–1820 J; N=4), while C. anna expended an average of 2132±748 J (range: 1297–2737 J; N=3).

Hummingbirds also ingested variable total amounts of cane sugar solution(Table 3). S. rufusingested an average of 0.601±0.224 ml (N=4) of cane sugar solution, equivalent to 350.9±131.0 μmol (N=4) of sucrose. C. anna ingested an average of 0.347±0.278 ml (N=3)of cane sugar solution, equal to 352.2±142.6 μmol (N=3) of sucrose.

The amount of ingested cane sugar oxidized by individual hummingbirds over the 2-h experimental period varied widely(Table 3). S. rufusoxidized an average of 109.5±34.3 μmol (N=4) while C. anna oxidized an average of 160.5±73.4 μmol (N=3) of the sucrose they ingested. Interestingly, there seemed to be correspondence between the amount of cane sugar each hummingbird ingested and the amount of sucrose oxidized from these meals during the 2-h experimental period(Fig. 3, Table 3). S. rufusoxidized an average 31.7±2.7% while C. anna oxidized an average of 45.5±5.9% of the sucrose they ingested in the form of cane sugar. These average values were significantly different(F1,5=17.7556, P=0.0084).

Consistent with results obtained previously in rufous(Suarez et al., 1990) and broad-tailed hummingbirds (Welch et al.,2006), RQ values displayed by rufous and Anna's hummingbirds during hovering flight rapidly rose from values near 0.71–1.0 as birds transitioned from a fasted to a fed state. These results indicate that hovering hummingbirds rely largely on fatty acid oxidation to fuel hovering flight when fasted, but switch to and rely almost exclusively on carbohydrate oxidation during repeated foraging (Suarez et al., 1990; Welch et al.,2006).

Initial δ13Cbreath values indicate that hummingbirds were oxidizing endogenous energy stores derived from the maintenance diet. Respiratory quotients (RQ) associated with these initial feeding events averaged 0.74±0.01 for S. rufus and 0.76±0.02 for C. anna, implicating fat as the primary metabolic fuel. Average initial δ13Cbreath values were 1.18 and 1.52‰ lower than the δ13C value of the maintenance diet (S. rufus and C. anna, respectively). This small discrepancy is based, in part, on the fractionation that occurs as sugars are converted into stored fat, resulting in a relative depletion of 13C (DeNiro and Epstein,1977). The magnitude of difference between the initialδ 13Cbreath values and the δ13C value of the maintenance diet is likely to be less than the actual degree of fractionation that occurs during fat synthesis from sugars as the initial RQ values are slightly greater than 0.71, indicating some contribution of carbohydrate (which would not be subject to the same fractionation) to the fuelling of hovering metabolism.

As hummingbirds continued to feed on the cane sugar solution,δ 13C values rose towards theδ 13Cbreath value of the cane sugar solution and,in several individuals, actually reached this value. The increase in hovering RQ values during the period of cane sugar availability in parallel with the rise in δ13Cbreath values indicates that the source of carbohydrates being oxidized was almost exclusively dietary. That is, the rise in RQ was due almost entirely to the recruitment of newly ingested sugars into the pool of actively metabolized substrates. RQ values displayed during the subsequent period of beet sugar availability remained near 1.0, indicating a continuing reliance upon carbohydrate oxidation,despite the fact that δ13Cbreath values declined towards the δ13C value of the beet sugar solution. As both cane and beet sugars consist of sucrose molecules, indistinguishable except via stable isotope analysis, no change in RQ would be expected as hummingbirds transitioned from reliance on one dietary sugar to the other. Thus, there was no expected significant correlation between RQ andδ 13Cbreath values during this period, and none was observed.

These results indicate that rufous and Anna's hummingbirds possess a capacity for the rapid and extensive use of recently ingested sugar in fuelling ongoing metabolism, suggesting convergence of physiological traits with other nectarivorous hovering animals such as bees and sphingid moths(Blatt and Roces, 2001; O'Brien, 1999; Welch et al., 2006). In support of our initial hypothesis, these results are strikingly similar to those described in broad-tailed hummingbirds(Welch et al., 2006),indicating that such physiological capacities are likely widespread among small hummingbirds.

The fact that δ13Cbreath values quickly declined and approached the δ13C value of beet sugar once hummingbirds were given access to this food source further supports our hypothesis that these animals make use primarily of the most recently ingested sugars when involved in steady-state foraging. As indicated byδ 13Cbreath values(Fig. 1B), hummingbirds engaged in steady state foraging were no longer relying upon oxidation of cane sugar to support hovering metabolism after approximately 30 min of feeding on beet sugar. By comparison, humans exercising at approximately 45% of their maximal O2 were observed to be still oxidizing glucose ingested more than 200 min earlier at a significant rate (Krzentowski et al.,1984). When individuals ingested glucose, rested, and then exercised at approximately 45% of their maximal O2, the ingested fuel remained available to the pool of actively metabolized substrates for an even greater period of time (Jandrain et al., 1984). The more rapid turnover of ingested sugars in the pool of actively metabolized substrates in hummingbirds is consistent with their small size and high mass-specific metabolic rates(Suarez, 1992).

Studies revealing net energy gain or loss in hummingbirds have traditionally relied on the monitoring of body mass over a period of several hours to several weeks (e.g. Calder et al.,1990; Carpenter et al.,1993; Gass et al.,1999). However, studies over shorter time-scales face problems associated with smaller mass changes due to fuel storage and utilization, as well as variation in mass due to dietary water intake, and water loss. With few exceptions (e.g. Gass et al.,1999), not much can be learned when mass change is near or equal to zero.

Other methods for determining the fate of ingested carbon in birds are available. Tissues can be sampled to characterize their carbon stable isotopic signature in relation to the signature of the diet (e.g. Hobson et al., 2005; O'Brien et al., 2000; Sydeman et al., 1997; Wolf and Martínez del Rio,2000). However, these techniques require invasive sampling that,in the case of hummingbirds, would likely be fatal and non-repeatable. On the other hand, biological 13C-NMR spectroscopy for monitoring of fuel storage and metabolism requires that animals be restrained. As a result, the techniques described here are uniquely suited to the study of energy turnover in foraging hummingbirds.

Because recently ingested sugars appear and then disappear from the pool of actively oxidized substrates (as indicated by the appearance/disappearance of a characteristic δ13C signature from expired CO2),and because nearly all of ingested sugars are absorbed by the hummingbird while little is lost in excreta (Karasov et al., 1986; McWhorter et al., 2006), it is likely that sugars not immediately oxidized to support ongoing metabolic needs are stored. Although some carbohydrate is stored in the form of glycogen, it is likely that most of the excess dietary carbon is stored as fat (Carpenter et al.,1993; Odum et al.,1961; Suarez et al.,1990). By quantifying the amount of a given sugar (with a distinct isotopic signature) ingested and monitoring its rate of utilization via a combination of respirometry and stable isotope analysis, it is possible to determine whether sugar molecules are oxidized or stored.

Despite widely varying rates of activity, energy expenditure and rates of cane sucrose ingestion across individuals, the proportion of ingested cane sugar that was oxidized remained relatively constant within each species(Table 3). This means that within species, each individual stored the same fraction of ingested sugar despite variation in total intake. This intriguing result implies that, within species, there is relatively precise matching between each individual's rate of energy expenditure and its rate of energy intake and storage. This adds further support for the suggestion that hummingbirds possess an accurate means of matching energy intake rate to energy demand(Gass et al., 1999).

On average, Anna's hummingbirds oxidized a significantly greater proportion of ingested cane sugar than rufous hummingbirds during the 2-h experimental period (F1,5=17.7556, P=0.0084). One possible explanation for the difference in the proportion of ingested energy that is oxidized as opposed to reserved for energy storage between these species lies in the differences in their life histories. The Anna's hummingbirds collected for this study were taken from a resident population at the University of California, Santa Barbara campus. Anna's hummingbirds, particularly those inhabiting coastal areas of southern and central California, tend to stay in place in August through October (Russell,1996) (K. Welch, personal observation), the period when our experiments were conducted. Then, they disperse after breeding in late spring and summer. On the other hand, rufous hummingbirds undergo one of the most impressive annual migrations of any animal, with some individuals migrating upwards of 6000 km from breeding grounds as far north as Alaska to wintering grounds in central Mexico (Calder,1987; Phillips,1975). These flights are interrupted by refuelling stops to replenish fat stores (Carpenter et al.,1983). Hiebert (Hiebert,1993) noted that captive rufous hummingbirds maintained a higher average daily body mass during periods of the year corresponding to their southward migration compared to non-migratory times. This period of elevated body mass (mid-August through November) encompasses the period when our experiments were conducted. In contrast, Calder et al.(Calder et al., 1990) noted that resident territorial broad-tailed hummingbirds appeared to restrain food intake so as to maintain a lower body mass, presumably facilitating aerial agility. Anna's hummingbirds (adult males in particular) are territorially aggressive and may derive similar benefits from restraining mass gain during the majority of the foraging period. So, the possibility exists that rufous hummingbirds oxidized, on average, a smaller percentage of the cane sugar they ingested compared to Anna's hummingbirds, in part because of the seasonal predisposition to fat deposition. A potential application of the techniques described here is to test this hypothesis using a variety of hummingbird species with disparate life history characteristics. In contrast with other methods, the range of possibilities is considerably broadened given that the procedures can be carried out in the field.

  • O2

    rate of oxygen consumption (ml O2 g–1h–1)

  • CO2

    rate of carbon dioxide production (ml CO2g–1 h–1)

  • RQ respiratory quotient


  • E

    energy expenditure (J)

  • fexo

    fraction of expired CO2 derived from oxidation of cane sugar

  • t

    time (s)

  • M

    amount of metabolic substrated oxidised (μmol)

  • Met

    amount of oxygen consumed (mol)

  • Mb

    body mass (g)

  • MR

    metabolic rate (ml O2 h–1)

  • γ13C

    isotopic 13C/12C ratio referenced to international standard (Carleton et al.,2006)

  • VPDB

    Vienna Pee Dee Belemnite C standard

Charlotte Guebels analyzed all video recordings and calculated time budgets for each hummingbird. We thank Andrea Hochevar, Andrea Wisniewski, William Talbot Bowen, Nicole Boyd, and Samantha Levinson for help in capturing and/or caring for hummingbirds. Dan Day assisted in the preparation of solid samples for submission to the UCSB MSI Analytical Lab. We thank François Péronnet, Bradley Hartman Bakken and Carlos Martínez del Rio,who provided helpful insight and discussion regarding these data. We thank Georges Paradis and the staff at the UCSB MSI Analytical Lab for their work in analyzing all stable isotope samples. We also thank John Lighton and Sable Systems International, Inc. for equipment and technical support, and Cyril Johnson for fabrication of the IR-detector circuitry used in this study This work was supported by NSF grant IOB 0517694 to R.K.S.

Berger, M. and Hart, J. S. (
). Die Atmung beim Kolibri Amazilia fimbriata wahrend des Schwirrfluges bei verschiedenen Umgebungstemperaturen.
J. Comp. Physiol.
Blatt, J. and Roces, F. (
). Haemolymph sugar levels in foraging honeybees (Apis mellifera carnica):dependence on metabolic rate and in vivo measurement of maximal rates of trehalose synthesis.
J. Exp. Biol.
Brouwer, E. (
). On simple formulae for calculating the heat expenditure and the quantities of carbohydrate and fat oxidized in metabolism of men and animals, from gaseous exchange (oxygen intake and carbonic acid output) and Urine-N.
Acta Physiol. Pharmacol. Neerl.
Calder, W. A., III (
). Southbound through Colorado: migration of rufous hummingbirds.
Nat. Geogr. Res.
Calder, W. A., Calder, L. L. and Frazier, T. D.(
). The hummingbird's restraint: a natural model for weight control.
Carleton, S. A. and Martínez del Rio, C.(
). The effect of cold-induced increased metabolic rate on the rate of C-13 and N-15 incorporation in house sparrows (Passer domesticus).
Carleton, S. A., Wolf, B. O. and Martínez del Rio, C.(
). Keeling plots for hummingbirds: a method to estimate carbon isotope ratios of respired CO2 in small vertebrates.
Carleton, S. A., Bakken, B. H. and Martínez del Rio,C. (
). Metabolic substrate use and the turnover of endogenous energy reserves in broad-tailed hummingbirds (Selasphorus platycercus).
J. Exp. Biol.
Carpenter, F. L., Paton, D. C. and Hixon, M. A.(
). Weight gain and adjustment of feeding territory size in migrant hummingbirds.
Proc. Natl. Acad. Sci. USA
Carpenter, F. L., Hixon, M. A., Beuchat, C. A., Russell, R. W. and Paton, D. C. (
). Biphasic mass gain in migrant hummingbirds: body composition changes, torpor, and ecological significance.
DeNiro, M. J. and Epstein, S. (
). Mechanism of carbon isotope fractionation associated with lipid synthesis.
Gass, C. L., Romich, M. T. and Suarez, R. K.(
). Energetics of hummingbird foraging at low ambient temperature.
Can. J. Zool.
Hiebert, S. M. (
). Seasonal changes in body mass and use of torpor in a migratory hummingbird.
Hobson, K. A., Thompson, J. E., Evans, M. R. and Boyd, S.(
). Tracing nutrient allocation to reproduction in Barrow's Goldeneye.
J. Wildl. Manage.
Horwitz, W. (
Official Methods of Analysis of the Association of Official Analytical Chemistry
. Washington, DC: The Association of Official Analytical Chemistry.
Jandrain, B., Krzentowski, G., Pirnay, F., Mosora, F., Lacroix,M., Luyckx, A. and Lefebvre, P. (
). Metabolic availability of glucose ingested 3 h before prolonged exercise in humans.
J. Appl. Physiol.
Karasov, W. H., Phan, D., Diamond, J. M. and Carpenter, F. L. (
). Food passage and intestinal nutrient absorption in hummingbirds.
Krzentowski, G., Jandrain, B., Pirnay, F., Mosora, F., Lacroix,M., Luyckx, A. S. and Lefebvre, P. J. (
). Availability of glucose given orally during exercise.
J. Appl. Physiol.
Lasiewski, R. C. (
). Oxygen consumption of torpid, resting, active and flying hummingbirds.
Physiol. Zool.
McWhorter, T. J., Bakken, B. H., Karasov, W. H. and Martínez del Rio, C. (
). Hummingbirds rely on both paracellular and carrier-mediated intestinal glucose absorption to fuel high metabolism.
Biol. Lett.
O'Brien, D. M. (
). Fuel use in flight and its dependence on nectar feeding in the hawkmoth Amphion floridensis.
J. Exp. Biol.
O'Brien, D. M., Schrag, D. P. and Martínez del Rio,C. (
). Allocation to reproduction in a hawkmoth: a quantitative analysis using stable carbon isotopes.
Odum, E. P., Connell, C. E. and Stoddard, H. L.(
). Flight energy and estimated flight range ranges of some migratory birds.
Phillips, A. R. (
). Migrations of Allens and other hummingbirds.
Phillips, D. L. and Koch, P. L. (
). Incorporating concentration dependence in stable isotope mixing models.
Russell, S. M. (
). Anna's Hummingbird. In
The Birds of North America
. Vol.
(ed. F. Gill), pp.
-23. Philadelphia: The Academy of Natural Sciences of Philadelphia.
Russell, S. M. and Russell, R. O. (
The North American Banders' Manual for Banding Hummingbirds
. Point Reyes Station, CA: North American Banding Council.
Suarez, R. K. (
). Ecological implications of metabolic biochemistry.
Suarez, R. K., Lighton, J. R. B., Moyes, C. D., Brown, G. S.,Gass, C. L. and Hockachka, P. W. (
). Fuel selection in rufous hummingbirds: ecological implications of metabolic biochemistry.
Proc. Natl. Acad. Sci. USA
Sydeman, W. J., Hobson, K. A., Pyle, P. and McLaren, E. B.(
). Trophic relationships among seabirds in central California: combined stable isotope and conventional dietary approach.
Welch, K. C., Bakken, B. H., Martínez del Rio, C. and Suarez, R. K. (
). Hummingbirds fuel hovering flight with newly ingested sugar.
Physiol. Biochem. Zool.
Withers, P. C. (
). Measurement of VO2, VCO2, and evaporative water loss with a flow-through mask.
J. Appl. Physiol.
Wolf, B. O. and Martínez del Rio, C.(
). Use of saguaro fruit by white-winged doves: isotopic evidence of a tight ecological association.