Interplay Between The Energetics Of Foraging And

Interplay between the energetics of foraging and thermoregulatory costs in the green-backed fire crown hummingbird Sephanoides sephaniodes Abstract In response to the interplay between variation in food quality and energetic demands, the foraging Behavior of captive green-backed fire crown hummingbirds Sephanoides sephaniodes was studied. Hummingbirds were exposed to two temperatures (25 Grados Celsius vs. 15grados Celsius), two food qualities (0.5 vs. 0.75 m sucrose solutions), and two costs of feeding (birds were provided with feeders with and without a perch). Food selection and consumption were measured, as well as time budgets and metabolic rate while feeding. We predicted that when given a choice, birds would minimize the cost of feeding by selecting feeders with a perch and with a high sugar concentration. However, rather than increasing energy consumption when energy availability was low and thermoregulatory demands were high, hummingbirds remained perched. They reduced feeding and spent most of their time perching. Our results identify a novel behavioural and Physiological strategy in hummingbirds. These birds seem to shift their foraging behavior depending on Thermoregulatory and feeding costs. When these costs are high, rather than matching them with increased energy consumption, hummingbirds reduce energy costs by reducing activity. They seemed to adopt the following strategy: when food quality was high and thermoregulatory demands were low, they adopted a high-expense lifestyle. In contrast, when thermoregulatory costs were high, they adopted an energy conserving strategy even when food quality was high. We hypothesize that limitations imposed by Physiological processes may explain why animals do not forage during all available time and why under Some circumstances they choose foraging behaviours with lower rates of net energy gain.

INTRODUCTION Animals can face unpredictability in food availability and quality. In addition, daily energy demands may change unpredictably. For example, in endotherms, a sudden cold spell can lead to greatly increased thermoregulatory expenses (Dawson & O'Connor, 1996). To match supply to demands, animals must show behavioural and physiological flexibility. Because hummingbirds are small and have an expensive foraging mode, they are good models to investigate how animals respond to variation in food quality and energetic demands (P. J. Miller, 1996). The small size of hummingbirds Makes them sensitive to variation in energy availability and demands (Hainsworth, 1981; L. L. Wolf & Hainsworth, 1983; Calder, 1984;Mc Nab, 1988). In the wild, hummingbirds often face short-term fluctuations in environmental temperature and energy availability (Cornet et al., 1979; Pyke & Waser, 1981; Tamm, 1989). The challenge posed by these fluctuations

is accompanied by extremely high feeding costs. The cost of hovering in feeding hummingbirds can reach the upper metabolic limits of vertebrates (Bartholomew & Lighton, 1986; Suarez et al., 1990; Suarez, 1998; Winter, 1998; Chai & Dudley, 1999). To satisfy their energetic demands while feeding on nectar, a relatively dilute sugar solution (Baker & Baker, 1973; Baker, 1975), hummingbirds must consume relatively enormous amounts of food. Hummingbirds feeding on dilute nectar can ingest nearly three times their body mass in nectar per day (Powers & Nagy, 1988; Gass, Romich & Suarez, 1999; Mc Worther & Martinez del Rio, 1999). When modelling the dynamic energy and mass Budgets in biological systems, Kooijman (2000) identified the main factors that determine energy acquisition as body mass, food availability and food type, and temperature. Following Kooijman (2000), we investigated the interplay between foraging energetic, the costs of thermoregulation, and the feeding decisions of captive green-backed fire crown hummingbirds Sephanoides sephaniodes (Trochilidae).

The cost of feeding was varied by providing hummingbirds with feeders with and without a perch. Sucrose concentration was varied to manipulate food quality. To probe the effect of thermoregulatory costs on feeding, environmental temperature was varied. To provide a mechanistic foundation for the behavioural responses of the hummingbirds in our study, their time budgets while feeding were also monitored. Because the maintenance of high and constant body temperature is costly for small endotherms (Johnston & Bennett, 1996), we predicted that in the cold, the increased thermoregulatory demands would lead to increased energy consumption. Concomitantly, because thermoregulatory costs and basal rates of energy expenditure in small birds typically account for nearly 40-60% of total daily energy expenditure (B. A. Wolf, Wooden & Walsberg, 2000), we also predicted that increased energetic feeding costs would lead to increased feeding rates, supporting the compensatory feeding hypothesis (Mc Worther & Martinez del Rio, 1999). In fact, R. S. Miller (1985) pointed out that hovering is not necessarily the preferred mode of feeding in hummingbirds. Because oxygen consumption increased 2.7- fold when S. sephaniodes fed while perched, and 6.8-fold when they fed hovering (M. J. Fernandez, M. V. Lopez- Calleja & F. Bozinovic, pers. comm.) consequently, we expected that when given a choice, hummingbirds would select feeders with perches and with a high sugar concentration. We also anticipated that at high temperatures the benefits conferred by selecting a low cost, high return strategy would be reduced. Thus, we expected harsh conditions to intensify the need to make wise feeding choices. Our design and questions are ecologically relevant: the green-backed fire crown is a migratory hummingbird that winters in the semi-arid and Mediterranean environments of central Chile. In central Chile, these hummingbirds feed primarily on the nectar of Sophora macrocarpa (Papilonaceae), Tristerix aphylus, T. tetrandrus (Lorantaceae), Lobelia polyphylla (Lobeliaceae), Porlieria chilensis (Zigofilaceae) and the introduced Eucalyptus globulus (Myrtaceae). The sugar concentration in the nectar of these flowers ranges from 0.25 to 1.21 m (Belmonte, 1988; Smith-Ramirez, 1993; Fraga, Ruffini & Grigera, 1997; Smith-Ramirez & Armesto, 1998). Thus, birds can encounter food of varying quality while foraging. Winters in central Chile can be cold and wet (di Castri & Hajek, 1976). Thus, birds can be faced with cold spells in which thermoregulatory demands increase.

METHODS Field observations, birds capture and maintenance Non-reproductive birds (mean body mass-sd = 5.7- 0.4 g) were captured with mist nets in central Chile (33grados 17'S, 71grados 11'W) during June-July. Birds were maintained in individual 30x30x30 cm cages for 3 days, after which those birds that maintained a constant body mass (e.g. did not loss weight) were transferred to a 60x60x60 cm cages in the aviary. Birds were maintained at 25 grados C under a 12:12 light:dark cycle and fed a 0.6 m sucrose solution, fruit-flies Drosophila melanogaster, and water ad libitum. Between experiments, a vitamin and protein supplement was added to the sucrose solution (Vimiprotein-L1, Rhone-Poulenc Rorer, S.A., Santiago, Chile, 0.3 g/50 ml of solution). A total of 21 individuals were observed and studied. The study was conducted during Austral winter-spring 1999 And autumn 2000.

Preference trials Preference trials were conducted between 08:00 and 12:00. Variations were made to temperature (15 Grados C and 25 Grados C), difference in sugar concentration (0.75 and 0.5 m sucrose), and the presence (P+) or absence (P7) of a perch in each of the available feeders. Each trial consisted of 1 bird in a cage with 2 feeders. Feeders were offered in all possible combinations of temperature, sugar concentration and perch availability in a fully factorial design. At the 2 experimental temperatures (15 Grados C and 25 Grados C) birds spend 2.7 and 1.5 times more energy in thermoregulation, respectively, than at their thermo neutral zone (28-33 Grados C; Lopez-Calleja & Bozinovic, 1995). Both sugar concentrations were selected coinciding with previous feeding experiments (Lopez- Calleja, 1999). Preference trials were determined in the 60x60X60 cm cages by providing each of 8 birds 2 feeders (with the different options). The feeders were at the same distance (c. 55 cm) from a single perch and 20 cm apart. Feeders outside the cage were used for dripping control. Preference between feeding situation through volumetric measurement of food consumption, was determined twice daily by measuring the amount of food consumed from a graduated plastic feeder (. 0.1 ml). To minimize Position effects, feeders were switched randomly between experiments. Food consumption, body mass balance and time budget to asses the effects of sugar concentration and ambient temperature on feeding strategy 8 birds were acclimated for 7 days at 25 Grados C, then each bird was sequentially measured in each of 4 treatments (0.75/with-perch; 0.50/ with-perch; 0.75/without-perch; 0.50/without-perch). Birds were then acclimated to 15 8C and the experimental Protocol was repeated. Feeding frequency was monitored for 1 h (from 10:00 to 11:00) using 2 video-recorders (Sony CCD-TR413 NTSC, Sony Corporation, Japan) placed in front of the

experimental cages. These recorders were connected to a TV-VHS (Sony Corporation, Japan) which allowed us to record the animal’s behaviour without disturbing them. To minimize position effects, feeders were switched randomly between experiments. To determined the mass balance, birds were weighed daily at 07:30 and mass balance was calculated as the difference in weight between consecutive days. Statistics Statistical analyses were performed using STATISTICA (1997) statistical package for Windows 95. The effect of treatments was tested by a multi-way ANOVA test followed by the a posteriori Tukey test. Results are reported as mean - 1 sd,n= number of individuals. RESULTS Food preferences with different ambient temperatures Comparisons involving feeders with different sugar concentrations but with the same perch availability presumably provide information about food preferences. When ambient temperature was high (i.e. low thermoregulatory energy requirements), captive green backed Fire crown hummingbirds significantly preferred the high sugar concentration regardless of whether the feeder had a perch (F1, 44 = 38.79, P <0.0001, see Fig. 1a) except in the case of 0.5 m/with-perch vs 0.75 m/without perch. Comparisons involving the same sugar concentration presumably provide information about diet selection based on perch availability. When offered food with lower sugar concentration, hummingbirds significantly chose to feed on feeders with perch (F1,44 = 29.17, P< 0.0001; Fig. 1b). When presented with the higher sugar concentration they showed no significant preference between feeders with or without a perch (F1,44 =0.166, P <0.690; Fig. 1b). The preferences of hummingbirds for feeders with different perch availability depended on the quality of the available food. When birds were provided with feeders containing low sugar concentration without perches and feeders containing high sugar concentration with perches, they preferentially fed in feeders with a perch and high sugar concentration (F1,44 =6.65, P= 0.01; Fig. 1a). When they were offered an option between a feeder with a perch containing low sugar concentration and a feeder without a perch containing high sugar concentration, they did not show any significant preference (F1,44 =1.75, P= 0.18; Fig. 1a). Hummingbirds subjected to low ambient temperature significantly preferred the high sugar concentration regardless of perch availability (F1,44 =31.65, P < 0.001; Fig. 2a), but not for the treatment 0.5 m/with-perch vs 0.75 m/without-perch. When offered only low sugar concentration, birds significantly preferred the feeder with the perch, and presumably a lower metabolic expenditure while feeding (Fig. 2b). Contrary to the results of the high ambient temperature treatment, birds showed significant preferences for feeders with perches when presented with the high sugar concentration (F1,44 =6.77, P < 0.01; Fig. 2b). Mean consumption was higher in birds that could use a perch but this difference was not significant. As before, the preferences of birds for feeders with different perch availability depended on the quality of the available food. In addition, when they were offered a choice between a feeder with a perch containing low sugar concentration and a feeder without a perch containing high sugar concentration, they showed no significant preferences (F1,44 = 1.471, P =0.232; Fig. 2a).Food consumption, body mass balance and time budget Food consumption (ml/min) decreased significantly with increasing sugar concentration and ambient temperature, and was also significantly affected by the interaction between

sugar concentration and ambient temperature (Table 1a, Fig. 3a). Energy intake (kJ/min) was not affected by sugar concentration but only by ambient temperature (i.e. thermoregulatory energy requirements; see Table 1b, Fig. 3b). Contrary to our expectations, energy intake was lower when hummingbirds were kept at the low ambient temperature. Table 2. ANOVA testing for the influence of perch availability (P), sugar concentration (S), and ambient temperature (Ta) on a) the frequency of visits to the feeder and b) resting time (min) in Sephanoides sephaniodes.

Regarding time budgets, multi-way ANOVA indicated that the frequency of feeder visits was higher at high ambient temperature (Fig. 4, Table 2a). Resting time was also higher at low ambient temperature and affected by the interaction between sugar concentration and perch availability (Fig. 4, Table 2b). In hummingbirds at low ambient temperature resting time was 47% higher in comparison with birds at high ambient temperature, independently of food quality or perch availability. Also, the frequency of feeder visits was 31% lower in birds at low ambient temperature. As expected, a negative linear relationship was observed between resting time and the frequency of feeder visits (y= 35.05 70.40 x; r =0.44, Syx = 11.69, P = 0.0013). This behaviour may allow hummingbirds to maintain body mass balance under different energy challenges, as supported by the lack of statistically significant effects of treatments (perch availability: F1,27 = 0.101, P= 0.753, ambient temperature F1,27 =0.212, P = 0.649 and sugar concentration F1,27 = 0.126, P= 0.726) on daily mass balance (g/day). DISCUSSION Classic frameworks dealing with foraging behavioural ecology, including patch use as well as diet selection, state that these behaviours depend on the ecological context in which foraging takes place (Stephens & Krebs, 1986; Brown, 1988, 1999). Nevertheless, the physiological capacities of animals, in combination with the energetic cost of feeding and the availability, structure and chemical properties of food can also have an important, albeit relatively unstudied effect on foraging choice (Caraco et al., 1990; Martinez del Rio, 1990; Torres-Contreras & Bozinovic, 1997; Bautista et al., 1998; Bozinovic & Vaquez, 1999). Theoretically, when an animal minimizes the time spent to meet its energetic/ nutritional requirements or maximizes the energy obtained, it is considered to be maximizing its fitness (e.g. Schoener, 1971; Hixon & Carpenter, 1988; Dukas, 1998). Here we discuss the results of our experiments which are designed to help us better understand the interplay between foraging energetics, the costs of thermoregulation, and the feeding decisions of captive green-backed fire crown hummingbirds. Our study examined how the feeding preferences of hummingbirds vary with thermoregulatory costs, feeding expenditure, and food quality. Hummingbirds increased energy intake (kJ/min) while minimizing feeding costs. This behaviour is emphasized when they are confronted with higher thermoregulatory costs. When ambient temperature was low hummingbirds minimized their feeding activity despite the availability of more concentrated food (Figs 1b & 2b). It seems that, depending on environmental and physiological factors, S. sephaniodes would have decreased their net energy intake by choosing the perchless feeders. This behaviour allows hummingbirds great flexibility in the management of their energy budgets and mass balance. The puzzling reduction in energy intake at the lower temperature occurs even though hummingbirds were perched without feeding for a longer time in comparison to birds under high ambient temperatures. Comparisons

involving time budgets and the use of feeders as a function of the cost of feeding and food quality presumably provide information about how hummingbirds select and consume food and manage their time. It was observed that hummingbirds at low ambient temperatures did not increase energy intake, but rather reduced the frequency of feeder visits and spent most of their time perching. A similar behaviour was observed by Bautista et al. (1998) in starlings. Hummingbirds did not increase their energy intake when subjected to high thermoregulatory costs, despite the opportunity to reduce feeding costs by perching. Why do the green-backed fire crown hummingbirds spend most of their time resting instead of increasing energy intake? Thermoregulatory costs account for 40-60% of the total daily energy expenditure of small birds (B. A. Wolf & Walsberg, 1996). Also, ambient temperature affects not only thermoregulatory costs, but May also affect the viscosity of the solution and therefore the resulting costs of nectar procurement. For example, Kingsolver & Daniel (1983) demonstrated that thermoregulatory costs may relate not only to the necessity of keeping the homeothermy at a particular ambient temperature, but also to the necessity of heating cool nectar to body temperature once it is ingested. Nevertheless, our experiments do not allow us to test if the additional cost of warming ingested food is negligible or not. Our previous unpublished measurements of oxygen consumption during feeding revealed a significant increase in metabolic rate during both perch and hover feeding in comparison to resting. The rate of metabolism during resting increased 2.7-fold during perch feeding, and 6.8-fold during hover feeding (see Suarez et al., 1990). Lopez-Calleja & Bozinovic (1995) reported that the maximum thermoregulatory metabolic rate of S. sephaniodes is 26.5 ml O2/g h. A comparison of the thermoregulatory costs at 15 Grados C plus the cost of feeding (With or without a perch) revealed the importance of adopting a perch-feeding strategy. At 15 Grados C the thermoregulatory cost of a 5.7 g hummingbird is c. 0.99 kJ/h. The cost of hover feeding is 3.04 kJ/h while the cost of perch feeding is 1.2 kJ/h (Fernandez et al. unpubl. data). Consequently, everything else being equal, the total energy expenditure (cost of thermoregulation plus the cost feeding) may decrease from c. 4.0 kJ/h to 2.2 kJ/h if hummingbirds adopt a perch-feeding strategy. Thus, given the variability in ambient temperature, in energy content of food and in perch availability, these hummingbirds are often approaching their metabolic ceiling when foraging, determining their behavioural decisions about foraging and time use. Thus, both the rates of food intake and the available feeding time must be taken into account when interpreting potential constraints on energy budgets (see Kvist & Lindstrom, 2000). Alternatively, the rate of digestion and/or nutrient transport across gut membranes may constrain the rate of energy intake at a maximal level as hypothesized by Diamond et al. (1986) and Karasov et al. (1986). These authors suggested that hummingbirds are energy maximizers, but that their digestive processing time determines feeding behaviour, explaining why hummingbirds spend most of their time perching. Although our experiments do not allow us to directly test the digestive constraint hypothesis; Lopez-Calleja, Bozinovic & Martinez del Rio (1997) documented that S. sephaniodes consumed significantly higher volumes of nectar early in the morning than later in the afternoon without compromising digestive efficiency. McWhorter & Martinez del Rio (1999) suggested that the rate of water processing by the kidneys of hummingbirds might impose a constraint on energy intake. These authors hypothesized that because hummingbirds feed on diets that are dilute aqueous solutions, their feeding time as well as energy assimilation may be constrained by excess water elimination. In addition, at low ambient temperature there is an additional water excess as a result of decreased evaporative water loss (Calder, 1984). These physiological mechanisms may also explain why hummingbirds spend a long time resting when feeding on diets with low sugar concentrations.

One major goal in physiological ecology is to understand the factors that might impose a metabolic ceiling to the energy and time budgets of an animal. As pointed out by Hammond & Diamond (1997), the energy budget of an organism may be limited by the amount of food available in the environment and/or by its own design. Changes in time budgets and in the expression of physiological and behavioural mechanisms of energy conservation indicate the existence of a metabolic ceiling in hummingbirds. Recently, McWhorter & Martinez del Rio (2000), McWhorter & Lopez-Calleja (2000) examined the factors that might impose a constraint or might affect the energy and time budgets. McWhorter & Martinez del Rio (2000) asked if gut function limits hummingbird food intake and therefore influences their energy budgets. These authors found that when hummingbirds were faced with higher energetic demands acutely, they were unable to increase energy assimilation to meet these demands, suggesting a central limitation to their energy budgets. On the other hand, Lopez-Calleja (1999) suggested that the peripheral organs (muscles) of S. sephaniodes could not transform energy into work and heat quickly enough to meet demands when birds were chronically exposed to low temperatures (and thus higher energetic demands). That hummingbirds limit foraging activity regardless of food availability, suggests that the concomitant energetic demands of thermoregulation and flight costs may exceed the energetic capacity of the muscle. Physiological capacities therefore influence food consumption and thus the energy budgets of hummingbirds. Finally, McWhorter & Lopez-Calleja (2000) proposed that the relative importance of central vs peripheral limitations changes based on the temporal and spatial conditions experienced by the animal. That is, both central and peripheral limitations are important influences on the energy budget of hummingbirds and therefore affect their foraging ecology. In addition, our previous laboratory studies of torpor in Chilean green-backed fire crown hummingbirds (Lopez-Calleja et al., 1997) demonstrated that body temperature is lowered only when the hummingbird is energetically stressed. Thus, they can reduce energy expenditure to compensate for low food availability and/or quality and for high thermoregulatory and feeding costs (see also Tiebout, 1991, 1992, 1993). Our results identify a novel behavioural and physiological strategy in hummingbirds; these animals seem to shift their foraging strategy depending on thermoregulatory and feeding costs. When these are high, rather than matching them with increased energy consumption, they reduce energy costs by reducing activity. Hummingbirds seemed to adopt the following strategy: when food quality was high and thermoregulatory demands were low, they adopted a high-expense lifestyle. In contrast, when thermoregulatory costs were high, they adopted an energy-conserving strategy even when food quality was high. We hypothesize that limitations imposed by physiological processes may explain why animals are not capable of foraging during all the time available, and why under some circumstances they choose foraging behaviours with lower rates of net energy gain (Bozinovic & Martinez del Rio, 1996; Bozinovic & Vasquez, 1999; Bozinovic et al., 2000). In addition, Sandlin (2000) demonstrated that intraspecific interactions such as competition may also determine foraging decision, i.e. selection of high or low food quality. All these factors indicate that any particular behaviour represents an integrated response to the biotic and abiotic environment and the physiological state of the animal. The physiological constraints to the foraging ecology of animals are probably extremely important in determining when and how a food patch should be used and a diet selected (Tamm, 1989), which in addition to inter- and intraspecific interactions (Sandlin, 2000) would seem to be major components of the energy budget of free-living hummingbirds.

Acknowledgements This work was funded by a FONDAP 1051-0001 grant to F. Bozinovic. We are extremely grateful to Lee Gass, Fabian Jaksic, Carlos Martinez del Rio, and Todd McWhorter, Roberto Nespolo and Alvaro Palma for useful comments and their generous help. All experiments reported in this article were conducted according to the Chilean current law and under permit SAG-1485 of Servicio Agricola y Ganadero, Chile.