University of Groningen Avian Adaptation Along an Aridity Gradient Tieleman, Bernadine Irene

University of Groningen

Avian adaptation along an aridity gradient Tieleman, Bernadine Irene

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2002

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Tieleman, B. I. (2002). Avian adaptation along an aridity gradient: Physiology, behavior, and life history. s.n.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 27-09-2021 CHAPTER 8 Energy and water budgets of larks in a life history perspective: is parental effort related to environmental aridity?

B. Irene Tieleman, Joseph B. Williams, and G. Henk Visser Submitted to Ecology ABSTRACT We compared physiological, demographic and ecological variables of larks to gain insights into life history variation along an aridity gradient, incorporating phylogenetic relationships in analyses when appropriate. Quantifying field metabolic rate (FMR) and water flux (WF) of parents feeding nestlings as measures of parental effort, we found that parental FMR and WF were lower by 24-39% and 39-61%, respectively, in larks from arid environments compared with species from mesic areas. Water and energy requirements of 6-8 day old nestlings were reduced in desert species. Nestling growth rate, clutch size and number of clutches decreased with increasing aridity, and nest predation rates increased with increasing aridity. We combined FMR and WF of parents and chicks, energy and water accumulated during growth, and brood size to establish energy and water budgets of parent-brood units. Parent-offspring energy budgets equaled 101 kJ d-1 for Bar-tailed Desert Lark, 265 kJ d-1 for Hoopoe Lark, 162 kJ d-1 for Dunn’s Lark, 389 kJ d-1 for Skylark, and 345 kJ d-1 for Woodlark, a 28% reduction in the desert species when taking into account mass differences. Family unit water fluxes were 23.8 g d-1 for Bar-tailed Desert Lark, 48.6 g d-1 for Hoopoe Lark, 37.5 g d-1 for Dunn’s Lark, 101.4 g d-1 for Skylark, 82.8 g d-1 for Woodlark. Parent-brood units of arid-zone species used 28-50% less water per gram mass than species from mesic areas. These results support the hypothesis that decreasing food and water availability favor lower energy and water requirements of parents and young, reduced growth rates, and smaller clutch sizes with increasing aridity. The decrease in parental effort with increasing aridity might reflect a lower fitness value of a single brood for arid-zone species than for larks from mesic habitats, suggesting that the probability of adult survival is higher in arid than in mesic areas.

ABSTRACT Introduction A central tenet of life history theory is that current reproductive investment is traded off against residual reproductive value (Williams 1966; Stearns 1992; Roff 1993). This trade-off, or cost of reproduction, is fundamental in predicting the optimal life history in a variety of environments. The difficulty of obtaining direct measures of fitness costs to demonstrate a cost of reproduction has stimu- lated investigators to find quantifiable currencies that are related to fitness. A frequently used measure of current reproductive investment is parental effort, the proportion of available resources devoted to reproduction as opposed to growth and maintenance (Reznick 1985), that can be expressed in terms of energy, assuming that life history trade-offs are the result of energy allocation (Drent & Daan 1980; Bryant 1988). Because separating resources allocated to reproduction and to maintenance is problematic in field studies on birds, many studies use total daily energy expenditure as a measure of parental effort (Bryant 1988; Weathers & Sullivan 1989; Tinbergen & Verhulst 2000). In this study we use parental energy and water expenditure in the field as proxy to quantify parental effort, and we relate these variables to basal metabolic rate and total evaporative water loss from laboratory studies. In addition, we investigate variation in clutch size as an independent measure of parental effort. Whereas clutch size is ultimately evaluated in light of parental fitness allocation to current and future reproduction (Perrins & Moss 1975; Boyce & Perrins 1987), it may also be viewed as the outcome of a trade-off between growth rate and nutrient requirements of the young, and depend on environmental conditions such as food availability and risk of nest predation. Lack (1968) argued that nestling mortality due to predation could be reduced by shortening the nestling period, but that this would require faster growth and result in higher energy demands of the young. He suggested that growth rates are a compromise between food availability and risk of predation. Environments with a higher predation risk should select for faster growing young, and thereby force parents to raise fewer young per nesting attempt. In contrast with this prediction however, high FIELD nest predation, slow growth, and small broods occur together in the tropics, THE whereas low nest predation, fast growth and large broods coincide in temperate IN zones (Skutch 1966; Ricklefs 1979). Adding knowledge of the relationships between LARKS

growth rate, clutch size and nest predation over an environmental continuum of OF increasing aridity may provide new insights into the effect of environmental factors on demography and physiology. BUDGETS Deserts are characterized by high ambient temperatures (T ), unpredictable, low

a TER A rainfall and reduced primary productivity, resulting in limited food and water W availability for their inhabitants. One might expect that birds exposed to these AND conditions require specific physiological and behavioral adaptations that permit survival and reproduction (Serventy 1971; Dawson 1984; Williams & Tieleman ENERGY 167 2001). Low food and water availability could constrain energy and water intake, and the thermal environment may limit time available for foraging and force birds to minimize activity during the middle part of the day (Williams et al. 1999; Tieleman & Williams 2002a). In such an environment, natural selection potentially favors individuals with low rates of energy expenditure and water loss (Louw & Seely 1982; Williams & Tieleman 2001). However, during the reproductive season, parents not only provide food and water for themselves, but also for their offspring, and may need to elevate their own energy and water requirements in order to produce young. Whereas the amount of energy and water that can be invested per brood may be determined by the time available for foraging and the availability of food and water, the number of chicks that can be reared per brood depends also on the daily energy and water requirements per young. Reductions in these requirements could be accomplished by reducing nestling metabolism, growth rate, and evaporative and excretory water losses (Klaassen & Drent 1991). To date, few studies have investigated how energy and water are allocated to different components of the energy and water balance of a parent- brood complex in different environments. Early work on metabolism and water flux did not show any general physiological differences between desert and non-desert species (Bartholomew & Cade 1963; Dawson & Schmidt-Nielsen 1964; Serventy 1971; Dawson 1984). Subsequent studies, typically on single species of desert birds, have reported low basal metabolic rate (BMR), total evaporative water loss (TEWL) (Dawson & Bennett 1973; Weathers 1979; Arad & Marder 1982; Withers & Williams 1990), and low field metabolic rate (FMR) (Nagy 1987) and water flux (WF) (Nagy & Peterson 1988). More recently, across species comparisons between desert and non-desert species supported the hypothesis that arid-zone birds have on average lower BMR, FMR (Tieleman & Williams 2000) and TEWL (Williams 1996), also when taking into account phylogenetic relatedness among species. Results for comparisons of field water flux were equivocal, with conventional analysis showing differences between desert and non-desert species, but independent contrast analysis not (Tieleman & Williams 2000). Broad-scale interspecific comparisons of metabolism and water loss have the inherent problem that species differ not only in habitat but also in phylogenetic background, diet and behavior. Restricting comparative analyses to a small group of closely related species occurring in different environments provides the oppor- tunity for a more detailed examination of physiological adaptations while limiting complications due to dissimilar lifestyles or evolutionary history (Coddington 1988; Bennett 1988; Price 1991; Leroi 1994). The lark family (Alaudidae) has representatives living in environments ranging from hyperarid deserts to mesic grasslands (Cramp 1988; Pätzold 1994). Because all larks are ground-foraging birds that eat similar foods, a mixture of insects and seeds, beha- 168 vior and diet are not confounding factors in our analyses. In addition, knowledge of the phylogeny of the lark family based on molecular evidence (Tieleman et al. 2002b) allows us to select species with phylogenetic relatedness in mind. This family provides an appropriate model to study physiological and behavioral adaptation to the environment (Williams & Tieleman 2000; Tieleman & Williams 2002b; Tieleman et al. 2002c; Tieleman et al. 2003). Among larks, BMR and TEWL decrease along a gradient of increasing aridity (Tieleman et al. 2002b; Tieleman et al. 2002c). The variation in BMR and TEWL can not be explained by phylogeny (Tieleman et al. 2002b), or be attributed to acclimatization to temperature, food availability or day length, and is likely to have a genetic component (Tieleman et al. 2003). These laboratory measurements gain evolutionary significance if one finds consistent patterns in data collected in the field, where natural selection operates on a combination of physiology and behavior. In this study we compare physiological, demographic and ecological variables of larks to gain insights into the connection between life history variation and physiology along an aridity gradient. We quantify FMR and water flux of parents feeding nestlings as a measure of parental effort, and investigate if nestling FMR, water flux, and growth rate vary with environmental aridity. This information allows construction of energy and water budgets for the parent-offspring complex, and provides an integrative perspective of a series of components on which natural selection might act. We test the hypothesis that energy and water requirements of the parent-offspring complex are reduced in deserts, and that parental effort decreases with increasing aridity. In addition, we measure clutch size, nestling growth rate and nest predation risk to explore how this set of variables varies with decreasing food availability in the environment. We predict that arid environments where primary productivity is low provide less nest cover, resulting in higher nest predation rates, and less food, increasing clutch size and growth rate, and potentially overriding the effect of high predation risk favoring increased growth rate (Lack 1968). FIELD THE IN Methods

Study areas LARKS We studied Hoopoe Larks (Alaemon alaudipes), Dunn’s Larks (Eremalauda dunni), OF Bar-tailed Desert Larks (Ammomanes cincturus) and Black-crowned Finchlarks (Eremopterix nigriceps) from April to June 2001 in Mahazat as-Sayd, a nature BUDGETS TER A

reserve in the Arabian Desert (N 22°15', E 41°50'). Characterized by hot and dry W summers, the Arabian Desert is classified as an arid inland desert, similar to large AND parts of the Sahara (Meigs 1953). The flat gravel plains in Mahazat are intersec-

ted by wadis and dominated by a sparse vegetation of perennial grasses, including ENERGY

169 Stipagrostis sp, Panicum turgidum and Lasurius scindicus, and small acacia trees Acacia sp.. Yearly rainfall averages 90 ± 76 mm (SD) in Mahazat, but varies lar- gely with some years receiving less than 35 mm and others more than 140 mm

(National Wildlife Research Center, unpublished data). Records for Ta show nighttime lows of about 5 °C and daytime highs of 25 °C for January, and mini- ma of 28 °C and maxima up to around 49 °C for June/July (National Wildlife Research Center, unpublished data). During the breeding season average daily

Tas vary from below 30 °C at the beginning of April, to around 35 °C in the first

week of June, while maximum Tas increase from about 40 °C to 48 °C (Tieleman & Williams 2002a). Depending on rainfall, the breeding season can start in March and end in June, but in drought years birds do not breed at all. Skylarks (Alauda arvensis) and Woodlarks (Lullula arborea) were studied from April to June 2002 in Aekingerzand, a nature reserve in the Netherlands (N 52°52', E 06°20'). Aekingerzand is covered with low vegetation dominated by heather (Calluna vulgaris, Erica sp.), grasses (Festuca sp., Molinia caerulea) and scattered trees. Free standing water is available year round in lakes and ponds. Average yearly rainfall in this area is 773 mm. Mean minimum and maximum temperatures vary from -0.8 °C and 4.4 °C in January to 11.3 °C and 21.4 °C in July (Koninklijk Nederlands Meteorologisch Instituut). During the breeding

season average daily Tas increase from 7.5 °C in April to 14.4 °C in June, while

maximum Tas increase from 12.2 °C to 19.4 °C (Koninklijk Nederlands Meteorologisch Instituut). Woodlarks and Skylarks in the Netherlands breed from April until July.

Larks All species in this study nest on the ground in open cup nests (Cramp 1988). Dunn’s Larks, Bar-tailed Desert Larks and Black-crowned Finchlarks usually nest under small bushes or grasses. Hoopoe Larks choose more exposed nest sites, often adjacent to or on top of small bushes or grass clumps. Skylarks and Woodlarks construct nests that are typically concealed in heather or in grass. In the desert species males and females both incubate and share brooding duties during the nestling phase, but in the Skylarks and Woodlarks only the female incubates and broods nestlings. In all species both parents feed the nestlings.

Doubly labeled water Measurements of water flux and FMR were obtained using the doubly labeled water (DLW) technique, in which the rate of decline of 2H in the body water pool provides a measure of water flux (Nagy & Costa 1980), and the loss rates of 2 18 both H and O yield an estimate of CO2 production (Lifson & McClintock 1966; Nagy 1980; Speakman 1997). We mist-netted birds, injected them with a 1:2 mixture of 99.9 atom % 2H and 95.5 atom % 18O using a 100 or 250 µl

170 Hamilton syringe. The injection volume equaled 4.3 µl per gram mass. We weig- hed birds with a Pesola spring balance that had been calibrated against a Mettler analytical balance. After a 1 hour equilibration period (Williams & Nagy 1984a; Williams 1985), a 80-100 µl sample of blood (initial) was removed from the bra- chial vein, and birds were banded and released. After about 24 or 48 hours we recaptured birds, took a second blood sample (final), measured body mass and released them. We obtained blood samples of 3 uninjected individuals per species to determine background levels of isotopes. Isotope ratios of 2H/1H and 18O/16O were determined in duplicate (initial) or triplicate (final) for each sample at the Center for Isotope Research, University of Groningen (Visser & Schekkerman 1999). The coefficient of variation of the duplicate or triplicate measurements was generally less than 2%. We calculated water influx with equation 3 of Nagy and Costa (1980), and corrected for isotope fractionation effects assuming an evaporative water loss of 25% and a fractionation factor of 0.941 (equation 7.6 (Speakman 1997; Visser et al. 2000b)).

Total body water was estimated from isotope dilution. Rates of CO2-production were calculated with equation 7.17 of Speakman (1997). Validation studies on adult birds have shown that estimates of water flux using isotopes of hydrogen are usually within ±10% of values obtained by standard laboratory methods (Nagy & Costa 1980), and estimates of CO2 production from DLW are within 8-10% (Williams & Nagy 1984a; Speakman 1997). In growing birds the possibility exists that isotopes are differentially incorporated into growing tissues leading to errors in estimates of CO2 production (Williams & Nagy 1985b; Klaassen et al. 1989). However, three validation studies on growing chicks suggest that the errors in the estimates of CO2 production are in the same range as for adult birds (Klaassen et al. 1989; Visser & Schekkerman 1999; Visser et al. 2000a). We measured metabolism and water flux of 6-8 day old nestlings after growth had slowed, and metabolism and water flux approached a maximum. All nestlings increased mass during the measurement period.

CO2 production can be converted to energy expenditure when the composition of the diet is known (Gessaman & Nagy 1988; Weathers & Sullivan 1989). We FIELD assumed that seeds contain 13.5% protein, 5.1% lipid, and 81.4% carbohydrate THE

(MacMillen 1990) and that insects contain 62.0% protein, 14.9% lipid, and IN 15.0% carbohydrate (Williams & Prints 1986). We assumed for all species of LARKS

larks that adults during the breeding season consume a diet of 90% insects and OF -1 10% seeds, and calculated a conversion factor of 24.16 kJ l CO2 based on standard conversion factors for protein, fat and carbohydrate metabolism (Gessaman BUDGETS & Nagy 1988). For nestlings, that consume a diet of exclusively insects, this fac- TER A

-1 W tor equaled 24.39 kJ l CO2. AND Nest monitoring

We visited nests with eggs or young at 1-6 day intervals to determine nest survi- ENERGY val. When all eggs or nestlings disappeared we assumed that predation was the 171 cause. Calculations of nest mortality were made using the Mayfield-method (Mayfield 1975; Johnson 1979) for the total nest period, including laying, incubation and nestling phase. Daily survival rate was calculated as DSR = 1 - daily nest mortality.

Environmental aridity

We calculated an aridity index as Q = P/((Tmax + Tmin)(Tmax - Tmin)) * 1000, where

P is average annual precipitation (mm), Tmax is the mean maximum temperature

of the hottest month (°C) and Tmin is the mean minimum temperature of the col- dest month (°C) (Emberger 1955; Tieleman et al. 2002b). Although perhaps intuitively not straightforward, this index has been empirically derived to descri- be primary productivity in arid and semi-arid areas (Emberger 1955). This index is low in hot, dry deserts and high in cool, wet areas. Because Q increases rapid- ly when environments become more mesic, we avoided unequal weighing of data on mesic species by using log Q in our analyses (Tieleman et al. 2002b). Climate data for the geographical regions of the larks in this study have been reported in Tieleman et al. (2002b).

Statistical analysis and phylogenetic effect We used General Linear Models procedures in SPSS 10.0 for analysis of vari- ance (ANOVA) and covariance (ANCOVA) to investigate differences in FMR and WF among species. We always tested the two-way interaction terms and removed them from the model when insignificant. After one-way ANOVAs we used the Tukey test for multiple comparisons (Zar 1996). In comparisons among species ANOVA has proven useful because it takes into account intraspecific variation. However, this approach has been criticized because each species is treated as independent, whereas phylogenetic relatedness may cause non-independence among species (Felsenstein 1985a). Regression analysis with average values for species ignore intraspecific variation, but can take into account phylogenetic relationships when appropriate (Felsenstein 1985a; Garland et al. 1992). In addition to our analyses using ANOVA, we per- formed regression analysis using species averages taking into account phylogeny when appropriate. For these analyses we added data on FMR and water flux of the Dune Lark (Williams 2001) to our data set and correlated these variables with environmental aridity. We used the phylogeny of larks from Tieleman et al. (2002b), and placed the Thekla Lark, for which the phylogenetic relationships were not established, as sister species to the Crested Lark. To evaluate whether a phylogenetic effect [sensu Grafen (1989) and Harvey and Pagel (1991)] exists among the larks in this study, we used the test for serial independence (TFSI) to determine if there was a significant positive autocorrelation for mass-corrected field metabolic rate, mass-corrected water flux, growth rate,

172 clutch size, number of clutches and daily nest survival (Reeve & Abouheif 1999; Abouheif 1999). In each simulation the topology was randomly rotated 2000 times per iteration and the original data were shuffled 2000 times in order to provide the null hypothesis sampling distribution (Reeve & Abouheif 1999). The test for serial independence is more suitable for smaller data sets than other phylogenetic autocorrelation methods (Cheverud et al. 1985; Gittleman & Kot 1990; Martins & Hansen 1996; Abouheif 1999). If no phylogenetic effect exists, then incorporating phylogeny in statistical methods would be unnecessary (Gittleman & Kot 1990; Björklund 1997; Abouheif 1999). If a phylogenetic effect does exist, this may be attributable to phylogenetic constraint or to ecological factors and corrections for phylogenetic relationships may or may not be appropriate (Westoby et al. 1995). For the cases where we found a phylogenetic effect, we provided results of phylogenetic independent contrast analysis using the Phylogenetic Diversity Analysis Package (Garland et al. 1992) in addition to results from conventional statistics. Averages are reported ± 1 SD unless noted otherwise.

Results Field metabolic rate of parents We compared FMR of adults feeding 5-8 day old nestlings of six species of larks from arid and mesic environments (Figure 1C, Table 1). To take into account differences in body mass we calculated mass-corrected FMR as FMR divided by mass0.879. The exponent 0.879 was the common slope for all species determined by ANCOVA with log FMR as dependent variable, species as fixed factor and log mass as covariate in a model without interaction term (log mass F1, 51 = 19.48, P < 0.0001), after verifying that the interaction had no significant effect on log

FMR (species x log mass F5, 46 = 0.27, P = 0.93). Mass-corrected FMR differed significantly between species (F4, 52 = 31.81, P < 0.0001), and a Tukey test indi- cated that one homogeneous subset consisted of Bar-tailed Desert Lark, Hoopoe Lark, Dunn’s Lark and Black-crowned Finchlark (all species pairs, P > 0.51), another of Woodlark and Skylark (P = 1.00) (Figure 1A). Mass-corrected FMR FIELD of the arid-zone species was 24-39% lower than of the mesic larks. Because the THE IN number of nestlings per nest varied (Table 1), we calculated parental energy expenditure per chick by dividing mass-corrected FMR by the number of chicks LARKS OF the parent was feeding, and found no significant differences between species (F5,

52 = 0.96, P = 0.45) (Figure 1B). A relative measure of parental workload commonly used for comparisons across BUDGETS TER A

species is the ratio of FMR and basal metabolic rate (BMR) (Drent & Daan W 1980; Daan et al. 1990). We have determined BMR for all species of larks inclu- AND ded in this study, except the Bar-tailed Desert Lark (Tieleman et al. 2002b;

Tieleman et al. 2002c; Tieleman et al. 2003). From these studies, we used the ENERGY

173 n 2 2 10 7 8 7 SD 1.41 0.71 1.16 0.49 0.53 0.82 ents of six species larks. 2.7 3.5 4.0 2.0 2.5 2.7 #chicks per nest Nests e fed by par 11 11 11 3 2 20 n FMR/WF 2.86 3.05 3.19 0.99 0.68 2.33 SD verage A WF (g/day) 6.4 4.9 8.2 13.1 23.5 16.2 7.41 1.61 8.09 13.26 15.54 10.36 SD ds, and average number of chicks per nest that wer ent bir verage A FMR (kJ/day) 34.9 34.3 41.3 72.2 82.2 101.5 3 20 13 12 12 2 n 1.78 1.80 6.30 3.80 1.32 0.04 SD verage 16.5 21.5 39.8 33.6 27.0 13.6 A Mass (g) t Lark owned Finchlark -tailed Deser oodlark able 1. Body mass, field metabolic rate, water flux of par Bar Black-cr Dunn's Lark Hoopoe Lark Skylark W Species 174 T FIELD THE IN LARKS OF Figure 1. Field metabolic rate (average ± SE) of parents (a, b, c) and nestlings (d, e, f) of six species of larks taking into account mass-differences (a, d), expressed per nestling (b, e), and BUDGETS expressed per brood (c, f). Metabolizable energy intake of the entire brood, Ebrood (1f, see text) TER A is the sum of FMR (dark grey) and tissue production (light grey). Letters indicate homogeneous W

subsets based on Tukey tests (significance level P = 0.05). AND ENERGY

175 average BMR for Black-crowned Finchlark, Dunn's Lark, Woodlark and Skylark, species in which we found no significant relationship between log BMR and log mass with regression analysis. For Hoopoe Larks we estimated BMR based on body mass using the regression equation log BMR = 0.436 + 0.674 log mass 2 (SEslope = 0.226, r = 0.43, df = 13, P = 0.01). For the Bar-tailed Desert Lark we estimated BMR using data of the Desert Lark (Ammomanes deserti), a sister species that occurs in habitats of similar aridity (Tieleman et al. 2002b). Because Bar-tailed Desert Larks are about 23% smaller than Desert Larks, we used BMR per gram body mass of the Desert Lark to estimate BMR for the Bar-tailed Desert Larks. Parental FMR varied among species from 1.7 to 2.2 times BMR (Figure 3A). A Tukey test showed significant differences only of Bar-tailed Desert Lark compared with Woodlark and Dunn's Lark (F5, 52 = 6.61, P < 0.0001).

Water flux of parents For comparison of water flux among species, we calculated mass-corrected values as water flux/mass0.730 (Figure 2C, Table 2). The exponent 0.730 was the common slope for all species in an ANCOVA with log water flux as dependent

variable (log mass F1, 51 = 7.06, P = 0.01), after removing the insignificant inter-

action term from the model (F5, 46 = 0.16, P = 0.98). Mass-corrected water flux

differed significantly among species (F5, 52 = 33.65, P < 0.0001), and a Tukey test showed identical subsets of arid-zone larks and mesic-zone species as for mass- corrected FMR (Figure 2A). Mass-corrected water flux for arid-zone larks feeding nestlings was 39-61% lower than for larks from mesic areas. To compare the water flux of parents per chick, we divided mass-corrected water flux by the number of nestlings that each parent fed. Parental water flux per chick was similar among most species (Tukey all P > 0.05), with only a significantly lower mass- corrected water flux per chick in Bar-tailed Desert Larks compared with Skylarks

(ANOVA species F5, 52 = 3.25, P = 0.013, Tukey P < 0.05) (Figure 2B). Analogous to the ratio FMR/BMR, we calculated the ratio of water flux and total evaporative water loss (TEWL) as an index of parental water requirements relative to a physiological minimum of evaporation. Using data of TEWL per species from previous laboratory studies (Tieleman et al. 2002b; Tieleman et al. 2002c; Tieleman et al. 2003), we followed the same procedure as outlined above for

TABLE 2. Field metabolic rate and water flux of 6-8 day old nestlings of five species of larks.

Mass (g) FMR (kJ/day) WF (g/day) Species Average SD n Average SD Average SD N chicks

Bar-tailed Desert Lark 12.2 1 15.0 6.7 1 Dunn's Lark 13.0 1.58 4 22.7 3.86 7.1 1.58 4 Hoopoe Lark 20.4 2.34 13 31.1 4.98 6.7 1.40 13 Skylark 21.1 1.48 6 45.5 8.28 14.7 3.53 6 Woodlark 17.5 1.41 10 36.1 3.95 11.6 1.45 10 176 FIELD THE IN Figure 2. Water flux (average ± SE) of parents (a, b, c) and nestlings (d, e, f) of six species of larks taking into account mass-differences (a, d), expressed per nestling (b, e), and expressed per LARKS brood (c, f). Water requirements of the entire brood are the sum of water flux (dark grey) and OF water accumulated in tissue (light grey). Letters indicate homogeneous subsets based on Tukey tests (significance level P = 0.05). BUDGETS TER A W AND ENERGY

177 estimating BMR. Water flux varied among species (F4, 45 = 13.56, P < 0.0001), and was about 6.7 x TEWL in the mesic species and about 4.3 x TEWL in the arid-zone larks (Figure 3C).

Differences in parental FMR and water flux between sexes? To explore if males and females worked equally hard we used ANCOVA with log FMR or log water flux as dependent variable, species and sex as fixed factors and log mass as covariate. Neither log FMR nor log water flux differed significantly

between the sexes (log FMR F2, 49 = 2.71, P = 0.08; log water flux F2, 49 = 0.33, P = 0.72).

Field metabolic rate of nestlings We compared energy expenditure of 6-8 day old nestlings among five species of larks from arid and mesic environments (Figure 1E, Table 2). At 6-8 days nestling energy expenditure reaches an asymptote (Williams 2001) and therefore

Figure 3. Parental FMR (a) and water flux (b) expressed in multiples of BMR and TEWL, respectively, per brood for six species of larks. Letters indicate homogeneous subsets based on 178 Tukey tests (significance level P = 0.05). enables calculation of peak energy demand of the brood. In an ANCOVA with log FMR as dependent variable, species as fixed factor and log mass as covariate, the interaction between species and log mass was insignificant (F3, 25 = 0.97, P =

0.42). We found no significant influence of log mass (F1, 28 = 0.47, P = 0.50) but a significant effect of species on log FMR (F4, 28 = 10.73, P < 0.0001). Therefore, we removed mass from the model and investigated differences in FMR between species, excluding the Bar-tailed Desert Lark data from analysis due to low sample size. Skylark chicks had a higher FMR than nestlings of Woodlark and Hoopoe Lark, that were indistinguishable from each other, and Dunn’s Lark chicks had a lower FMR than the previous three species (Figure 1E). Mass-specific FMR (kJ -1 -1 day g ) of nestlings also differed among species (F4, 29 = 7.64, P < 0.0001) and was 16-43% lower in the arid-zone species than in the mesic larks (Figure 1D).

Water flux of nestlings For a comparison of water requirements of 6-8 day old nestlings (Figure 2E, Table 2) we tested the effect of species and log mass on log water flux with an ANCO-

VA and found no significant effects of the interaction (F3, 25 = 0.56, P = 0.65) or log mass (F1, 28 = 0.05, P = 0.82), but a significant effect of species (F3, 28 = 23.88, P < 0.0001). After removing mass from the analysis, we used a Tukey test with water flux as dependent variable to investigate differences between species, excluding Bar-tailed Desert Lark data from analysis due to low sample size. Nestlings of Dunn’s Larks and Hoopoe Larks had indistinguishable water fluxes, that were lower than the value for Woodlark chicks. Skylark young had higher water fluxes than nestlings of the other three species (Figure 2E). Mass-specific -1 -1 water flux (g day g ) differed significantly among species (F4, 29 = 17.78, P < 0.0001). Hoopoe Lark nestlings had a lower mass-specific water flux than the other three larks (Figure 2D).

Nestling growth rates We measured body mass of known-age nestlings to determine their growth rates

(Figure 4). Growth is commonly described by a logistic curve of the form W(t) FIELD THE = A / (1 + exp (-K (t - ti)), where W(t) (in g) is the weight at age t (day), A is IN the asymptote of the growth curve (g), K is the growth rate constant (day-1), and

ti is the inflexion point or age at maximal growth rate (day) (Ricklefs 1979). LARKS

Some chicks of the desert species did not grow and starved to death (Figure 4). OF These data were excluded when we constructed growth curves (Table 3).

To explore if growth rates are correlated with aridity, we added data for Dune BUDGETS TER

Larks from the Namib Desert (Williams 2001) and for Desert Larks and Crested A W Larks from the Negev Desert (Shkedy & Safriel 1992b) to our results (Table 4, AND Figure 5A). A regression model (r2 = 0.60, n = 9, P = 0.065) with growth rate constant as dependent variable showed a decrease in growth rate with increasing ENERGY aridity (slope ± SE = 0.115 ± 0.044, t = 2.63, P = 0.04), but no significant effect 179 of adult body mass (slope ± SE = -0.005 ± 0.003, t = 1.84, P = 0.12). Phylogeny was not a significant factor affecting growth rate (TFSI, P = 0.14).

Energy and water budgets of the parent-brood complex A budget of energy requirements for parents and brood combined can be com-

posed as Efam = 2 x FMRpar + Ebrood, where FMRpar is parental FMR (Figure 1c,

Table 1) and Ebrood is the metabolizable energy requirements of the nestlings

(Figure 1f). Ebrood can be calculated as n x (FMRnestling + ETnestling), where n is the

number of nestlings per nest (Table 1), FMRnestling is FMR per nestling (Figure 1e,

Figure 4. The relationship between body mass and age of nestlings, and the growth constant K (±

180 SE) for six species of larks. Day 0 is hatching day. TABLE 3. Logistic growth curve variables for growing nestlings of six species of larks. The logistic function is W(t) = A / (1 + exp (-K (t - ti)), where W(t) is the weight at age t, A is the asymptote of the growth curve, K is the growth rate constant, and ti is the inflexion point or age at maximal growth rate. The 95% confidence interval around K was calculated as K ± tdf x SE.

2 Species A ti K SE K 95% CI K r df

Bar-tailed Desert Lark 13.79 3.42 0.52 0.077 0.36 - 0.68 0.92 30 Black-crowned Finchlark 10.58 2.73 0.62 0.079 0.46 - 0.78 0.95 24 Dunn's Lark 15.81 3.63 0.50 0.056 0.39 - 0.61 0.90 71 Hoopoe Lark 26.12 4.93 0.41 0.038 0.33 - 0.49 0.91 105 Skylark 24.92 2.97 0.61 0.027 0.56 - 0.66 0.97 93 Woodlark 21.01 3.67 0.55 0.024 0.50 - 0.60 0.98 87

Table 2), and ETnestling is energy accumulated in new tissue. ETnestling can be estimated from the increase in wet mass per day (Table 3) and the energy density of wet tissue given by kJ/g wet mass = 3.51 + 4.82 x u, where u is the proportion of adult mass attained (Weathers 1996). Absolute Efam averaged 101 kJ/d for Bar- tailed Desert Lark, 265 kJ/d for Hoopoe Lark, 162 kJ/d for Dunn’s Lark, 389 kJ/d for Skylark, and 345 kJ/d for Woodlark. Expressed per gram of family to take into account mass-differences between species, these values equaled 1.98 kJ d-1 g-1, 1.96 kJ d-1 g-1 2.08 kJ d-1 g-1, 2.75 kJ d-1 g-1, and 2.79 kJ d-1 g-1, respectively, or about 28% lower in the desert species compared with the mesic-zone larks.

A family's water budget can be constructed as Wfam = 2 x WFpar + WFbrood, where

WFpar is parental water flux (Figure 2c, Table 1), and WFbrood is the water flux of the nestlings calculated as n x (WFnestling + WTnestling), with n the number of chicks per nest (Table 1), WFnestling the water flux per nestling (Figure 2e, Table

2), and WTnestling the amount of water accumulated in new tissue. WTnestling can be estimated from the increase in wet mass per day (Table 3) and the assumption that a 6-8 day old chick has a body water content of 69.4%, the average based on the dilution spaces of the chicks injected with doubly labeled water. The -1 -1 absolute Wfam equaled 23.8 g d for Bar-tailed Desert Lark, 48.6 g d for Hoopoe Lark, 37.5 g d-1 for Dunn’s Lark, 101.4 g d-1 for Skylark, 82.8 g d-1 for Woodlark. FIELD

Accounting for mass-differences among species, Wfam expressed per gram of fami- THE -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 ly averaged 0.46 g d g , 0.36 g d g , 0.48 g d g , 0.72 g d g , 0.67 g d g , IN respectively. Families of the desert larks used 28-50% less water per gram than species from mesic areas. LARKS OF

Field metabolic rate and water flux along an aridity gradient: conventional and phylogenetic analyses BUDGETS TER

We reanalyzed FMR and WF of Dune Lark adults and nestlings (Williams 2001), A W correcting for fractionation effects, and found for adults (27.0 ± 1.81 g, n = 11) AND an average FMR of 49.0 ± 17.05 kJ d-1 and WF of 5.99 ± 1.10 g d-1, and for nestlings at day 8 (mass 15.0 g) a FMR of 25.7 kJ d-1 and a WF of 5.29 g d-1 based on ENERGY logistic equations. Adding these data to the results of the present study we regres- 181 , ence vival rate (DSR 3, 5, 6 5, 6 7, 8, 9 2 2 3, 5 3, 4, 5 3 2 Refer 2 3, 4 1 2 2 16 13 26 n 21 29 93 ), and daily sur -1 0.0097 0.0206 0.0230 SE DSR 0.0290 0.0332 0.0188 SD, n) and number (year ez and Manrique (1992). 7. Maclean (1970b). 8. Lloyd ± 0.9379 0.9380 0.9491 0.9862 0.9636 0.9412 0.9618 0.8968 DSR 0.8638 0.9340 0.8162 0.8016 2 3.5 2.5 2 2 2 1 1 1 1 1 ez (1997). 6. Suár # Clutches 12 14 n 30 21 24 68 anes and Suár 0.900 0.730 SD 0.679 0.768 0.741 0.723 owth constant K, clutch size (average e. 3.24 3.60 2.58 3.92 4.07 3.50 4.15 4.20 Clutch size 2.57 3.24 3.11 1.90 2.88 2.99 a owth constant 0.612 0.553 0.473 Gr 0.622 0.520 0.540 0.307 0.498 0.409 t, aridity = 2.05. om this study and literatur 2.59 2.59 2.60 3.20 3.20 2.59 2.24 2.26 Aridity 2.24 1.78 2.05 1.76 1.78 1.78 . 3. Cramp (1988). 4. Shkedy and Safriel (1992b). 5. Y om Tieleman et al.(2002b)), nestling gr om Negev Deser t Lark SE, n) for 14 larks fr t-toed Lark ± owned Finchlark t Lark t-toed Lark owth constant fr -tailed Deser ested Lark oodlark Gr able 4. Aridity index (fr Thekla Lark Spike-heeled Lark Skylark W Lesser Shor 1. Boyer (1988). 2.This study (1999). 9. Lloyd (1998). Calandra Lark Shor T Species a Cr average Deser Black-cr Dune Lark 182 Dunn's Lark Hoopoe Lark Bar sed species-average values of FMR and WF against an index describing the aridity of the environment (Emberger 1955; Tieleman et al. 2002b), using both conventional regression and phylogenetic independent contrast analysis when appropriate. Lower values of this index indicate more extreme aridity. Conventional analysis showed that mass-corrected FMR and mass-corrected WF of adults decreased significantly with increasing aridity (FMR in kJ day-1 g-0.879, r2 = 0.99, slope ± SE = 1.23 ± 0.054, n = 7, P < 0.0001; WF in g day-1 g-0.730, r2 = 0.81, slope ± SE = 0.60 ± 0.128, n = 7, P = 0.006). For nestlings, mass-specific FMR and mass-specific WF also decreased with increasing aridity, but the latter was not significant (FMR in kJ day-1 g-1, r2 = 0.71, slope ± SE = 0.394 ± 0.127, n = 6, P = 0.037; WF in g day-1 g-1, r2 = 0.64, slope ± SE = 0.169 ± 0.064, n = 6, P = 0.057). Using the TFSI we found no significant phylogenetic effect in the data for mass-corrected FMR (P = 0.12) and mass-corrected WF of adults (P = 0.06), but significant phylogenetic autocorrelation in the data for mass-specific FMR (P = 0.047) and mass-specific WF of nestlings (P = 0.006). The latter two results and the marginal insignificance of adult water flux prompted us to additionally analyze these data using phylogenetic independent contrast analysis. This analysis showed that adult mass-corrected WF decreased significantly with increasing aridity, confirming the results of the conventional analysis (contrasts water flux r2 = 0.71, slope ± SE = 0.565 ± 0.162, df = 5, P = 0.017). Contrast analysis of the nestling data revealed trends of decreasing mass-specific FMR and WF with increasing aridity, identical to the conventional analysis, but insignificant (contrasts FMR r2 = 0.60, slope ± SE = 0.335 ± 0.136, df = 4, P = 0.07; contrasts WF r2 = 0.71, slope ± SE = 0.145 ± 0.071, df = 4, P = 0.11).

Clutch size and number of larks along an aridity gradient We combined information on clutch size and number of clutches per breeding season of larks from the Arabian Desert and the Netherlands with data from eight other lark species from the literature to test if clutch size and number of clutches varied with aridity (Table 4). When aridity increased, clutch size decreased (Figure 5B, slope ± SE = 0.746 ± 0.333, P = 0.045, r2 = 0.30, n = 14) and FIELD THE

number of clutches per season decreased (Figure 5C, slope ± SE = 1.39 ± 0.181, IN P < 0.0001, r2 = 0.87, n = 11). On average larks from hyperarid deserts have one clutch per year with 2.8 eggs, whereas species from mesic areas lay three clutches LARKS OF of 3.9 eggs per year. We determined if phylogeny might confound our analyses and found no significant phylogenetic autocorrelation for clutch size (TFSI, P = 0.06), but a significant autocorrelation for number of clutches (TFSI, P = 0.001). BUDGETS TER A

Therefore, we also analyzed the correlation between aridity and number of W

clutches using phylogenetic independent contrasts and found a significant AND regression confirming the results of the conventional analysis (slope ± SE = 1.37 2 ± 0.271, r = 0.74, t = 5.05, df = 7, P = 0.001). ENERGY 183 Nest predation along an aridity gradient We collated daily survival rates of nests for six species from the literature in addition to our own data on six species of larks from the Arabian Desert and the Netherlands (Table 4, Figure 5D). There was no significant phylogenetic effect in the daily survival rates (TFSI, P = 0.08). Using regression analysis we found that daily survival rate significantly increased with decreasing aridity (slope ± SE = 0.0969 ± 0.021, P = 0.001, r2 = 0.68, n = 12). Because the total nest period, the sum of laying, incubation and nestling-phase, is similar among all larks, daily survival rate directly reflects the probability that a nest survives until fledging. If we assume a total nest period of 24 days and use the regression equation to determine average daily survival rate, the probability that a lark nest in a hyperarid desert survives is about 2%, compared with 87% for a nest in a mesic habitat.

Discussion Field metabolic rate of parents Results of this study support the hypothesis of reduced FMR in desert birds and correspond with differences in BMR previously found among the same species of larks from arid and mesic environments (Tieleman et al. 2002b; Tieleman et al. 2002c). FMR was below allometric predictions based on body mass by 50% for Bar-tailed Desert lark, 44% for Hoopoe Larks, 50% for Dunn’s Larks, 43% for

Figure 5. Growth constant K (A), clutch size (B), number of clutches (C) and daily nest survival 184 rate (D) of larks along an aridity gradient. Data and sources are listed in Table 4. Black-crowned Finchlarks, 11% for Skylarks, and 16% for Woodlarks (Tieleman & Williams 2000). The association between levels of BMR and FMR has been attributed to selection for the size of the metabolic machinery required to main- tain levels of energy expenditure during the period when parents care for nestlings, the suggested time of peak energy demand (Daan et al. 1990). However, when variation in body mass is taken into account, organ sizes do not differ among Hoopoe Larks, Dunn’s Larks, Skylarks and Woodlarks (Tieleman et al. 2003). The differences in BMR and FMR among these species may be related to variation in tissue-specific metabolic rates of various organs rather than on their size, an avenue for future work. Although the reduced FMR in desert larks could be partly attributable to low levels of BMR, the physiological mechanisms that link BMR and FMR are poorly understood (Ricklefs 1996; Ricklefs et al. 1996), and behavioral differences may also play a role. Larks in the desert are inactive during a long period of the middle part of the day, when it is too hot to forage (Tieleman & Williams 2002a), whereas the mesic-zone larks are active throughout the day. The significance of the ratio between FMR and BMR in birds and mammals has been heavily debated since Drent and Daan (1980) first suggested that the optimal working capacity of birds tending broods is around 4 times BMR (Weathers & Sullivan 1989; Daan et al. 1990; Ricklefs 1996). The larks in this study had lower ratios, between 1.7 and 2.2 times BMR (Figure 3A), and similar to other ground-foraging passerines (Weathers & Sullivan 1989). In spite of this fairly constant ratio across larks from arid and mesic environments, workload when defined as the absolute power output in energy per unit time is 24-39% lower in the arid-zone larks. The amount of work that these species do is perhaps not the direct result of physiological constraints, but appears limited by the amount of time available for activity. The combinations of hard working birds with a high BMR and less hard working species with a low BMR result in identical ratios that would unjustly suggest identical parental efforts. Absolute levels of FMR and BMR placed in an ecological context might provide a more instructive measure FIELD of parental effort assuming that they are closely related to fitness-components THE such as survival, via potential effects of body condition, immune system and IN aging, and reproductive success, through brood size and feeding rates. LARKS

Water flux of parents OF Water flux was 39-61% lower in the arid-zone larks than in those from mesic areas, when taking into account variation in body mass (Figure 2A). Compared BUDGETS TER A

with allometric predictions water flux was reduced by 41% in Bar-tailed Desert W larks, 36% in Hoopoe Larks, 37% in Dunn’s Larks, and 48% in Black-crowned AND Finchlarks, and increased by 30% in Skylarks and 5% in Woodlarks (Tieleman

& Williams 2000). These results correspond with differences in TEWL pre- ENERGY viously found among the same species of larks from arid and mesic environments 185 and are consistent with the hypothesis of reduced water flux in desert birds (Tieleman et al. 2002b; Tieleman et al. 2002c). The association between TEWL and water flux is likely to result from both physiological adjustments, including increased skin resistance to evaporative water loss (Tieleman & Williams 2002b), and behavioral adaptations, including selection of favorable microcli- mates and reduced activity during hot hours (Williams et al. 1999; Tieleman & Williams 2002a). Larks in the desert spend a significant period of the hot part of the day at the nest, providing shade or even cooling the nestlings (Personal observation). This behavior presumably requires a significant amount of water for evaporative cooling and makes the low water flux of these species more remarkable. One might predict that the proportions of evaporative and excretory water loss differ markedly between birds from arid and mesic environments, with larks from the latter group losing larger quantities of surplus water by excre- tion. Larks in the desert have lower water fluxes and TEWL than species from mesic areas. The ratio WF/TEWL is 36% lower in the desert species, despite their

exposure to higher Tas and concurrent higher requirements for evaporative cooling. The more frugal water economy of larks in the desert may be interpreted as the result of selection in an environment where water is a limiting factor. Alternatively, the higher ratio of WF/TEWL might indicate the overabundance of water in temperate zones, and lack of selection in environments where water may be a waste product.

Energy and water requirements of nestlings The energy and water demands placed upon the parents by the brood depend on the energy and water requirements per nestling, and on the number of chicks in the nest. One might expect that nestlings with low energy and water requirements are favored by natural selection in environments where food and water are in short supply. In support of this idea, we found that mass-specific FMR was 16- 43% lower in 6-8 day old nestlings of Bar-tailed Desert Lark, Hoopoe Lark and Dunn’s Lark than in chicks of Skylark and Woodlark (Figure 1D). Although water flux per nestling was reduced in Hoopoe Larks, Bar-tailed Desert Larks and Dunn’s Larks (Figure 2E), mass-specific values were only significantly lower in Hoopoe Lark chicks, compared with young of Skylarks and Woodlarks (Figure 2D). In addition, the growth rates of larks in the desert were reduced and con- tributed to lowering the daily energy and water requirements of nestlings in these arid environments (Figure 5A).

Parental effort per nestling Larks from deserts expend less energy and lose less water while rearing young than larks from temperate areas, but because their broods were smaller losses per

186 chick were the same (Figure 1B, Figure 2B). Parental energy expenditure and water loss per chick may be viewed as an index of parental effort because it provides a total measure of the amount of energy and water required by a parent to provide resources to a nestling. However, it does not separate costs of parental self maintenance from costs of work that is specifically carried out to raise the young and therefore does not provide a prediction of the energy and water costs to rear an additional chick. Still, it is noteworthy that despite lower energy and water requirements per chick, parental energy and water expenditure to supply a chick’s requirements is as high in deserts as in mesic areas. Parents in deserts may be constrained by low water availability that reduces their evaporative cooling capacity and forces them to minimize activity during the hot part of the day. The resulting narrow window of time available to provision the brood leads to a smaller total amount of food that can be gathered per day. Therefore parental self- maintenance costs, although low per time unit in deserts, contribute a large proportion to the daily parental energy and water expenditure per chick.

Growth rate, clutch size and nest predation: Lack's dilemma With increasing aridity nest predation increased, growth rate decreased and clutch size decreased. This result does not support Lack’s (1968) prediction that growth rates should decrease with increasing risk of predation. We propose that food availability may have a larger selective influence on growth rate, overriding the effect of nest mortality. Low food availability should select for reduced over- all daily energy requirements of the brood, a prediction that is confirmed by the combination of reduced metabolism and water loss per nestling, low growth rate and small clutch sizes in arid environments. The combination of high nest predation, slow growth and small broods is also found in birds from the tropics (Skutch 1966; Ricklefs 1979). Although food may not be intuitively assumed to be a limiting factor on reproduction in the tropics, the reduced daily energy requirements per nestling (Weathers 1992) in addition to their slow growth rate, potentially also indicate selection for lower food requirements of nestlings. The more influential role of food availability than mortality on nestling growth rate does not contradict Lack’s original idea that growth rate might be a com- FIELD THE

promise between food supply and mortality, but shifts the emphasis of his subse- IN quent prediction from mortality to food supply. Along an aridity gradient food supply and mortality are correlated, and distinguishing the effect of each factor LARKS OF separately is not possible. To test if growth rate is the outcome of a trade-off between food supply and nest mortality would require an experimental approach. BUDGETS TER

Energy and water budgets of the parent-brood complex: a life history perspective A W Compared with temperate environments, the daily energy and water budgets of a family of larks in the desert, where food and water are scarce, are markedly AND lower not only when expressed in absolute terms but also per gram of family mass ENERGY

(Figure 1F, 2F). The frugal use of energy and water in arid-zone families are the 187 result of a combination of low parental and nestling FMR and water flux, slow growth, and small broods. Although in proximate terms the reduced energy and water budgets can be intuitively understood in light of the environmental conditions, our insights at the ultimate level are less complete. A central prediction of the evolutionary theory of life histories is that parental investment should vary directly with the fitness of current offspring and inversely with adult survival. If parental effort as measured by energy expenditure, water loss and clutch size is correlated with parental investment, i.e. the fitness consequence of the parental effort (Trivers 1972; Daan & Tinbergen 1997; Tinbergen & Verhulst 2000), we would conclude that the fitness value of a single brood is lower for an arid-zone species than for a lark from mesic habitats and predict that an adult lark in the desert has a higher probability of survival. Insights into the physiological, ecological and environmental factors affecting adult survival might provide the key to understanding how evolution has fashioned these distinct sets of physiological, behavioral and demographic variables in arid and mesic environments.

Acknowledgments We thank Abdulrahman Khoja, Patrick Paillat, Stéphane Ostrowski and the other staff at the National Wildlife Research Center, Taif, Saudi Arabia, for logistic support throughout this study. Wildlife research programs at the NWRC are possible through the generous support of HRH Prince Saud al Faisal and under guidance of A. Abuzinada of the National Commission of Wildlife Conservation and Development, Saudi Arabia. Vince Schuler and Riek van Noordwijk helped with field work in Saudi Arabia. We are grateful to Wouter de Vlieger and Staatsbosbeheer for permission to work at the Aekingerzand, to Berthe Verstappen for conducting the isotope analyses so quickly, and to Serge Daan for commenting on an earlier draft. Financial support for this study was made available by the Schuurman Schimmel van Outeren Foundation, the Schure Beijerinck Popping Foundation, the National Wildlife Research Center, and the National Science Foundation.

188 FIELD THE IN LARKS OF BUDGETS TER A W AND ENERGY

189